Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

The brevetoxins are multi-ring polyether neurotoxins produced by marine dinoflagellates (Ptychodiscus brevis = Gymnodinium breve) found in Florida waters (Baden, 1988) and more recently in New Zealand waters (Ishida et al., 1995). A human disease known as neurotoxic shellfish poisoning has been ascribed to the consumption of shellfish contaminated with these toxins (Baden, 1988).

Brevetoxins possess a molecular structure consisting of 11 transfused rings, 23 stereocenters and an overall linear low-energy conformation. The total synthesis of brevetoxin B in its naturally occurring form has been reported recently, as have shorter molecules that contain all of the salient features of the natural toxin, except they are shorter in length (Nicolaou et al., 1994, 1995). The brevetoxins have been used as model ligands for a more serious neurointoxication phenomenon known as ciguatera. Ciguatera is a human disease associated with the consumption of reef fishes that have accumulated the structurally related marine polyether toxins known as ciguatoxins (Lewis and Holmes, 1993).

Both the brevetoxins and ciguatoxins exert their neurotoxicity by altering the gating and sodium ion permeability of voltage-sensitive Na⁺ channels in excitable membranes (Baden, 1989; Wu and Narahashi, 1988). Numerous studies have investigated the effects of purified toxins on ion channel binding and electrophysiology, and it is known that both toxin groups bind specifically to site 5 associated with domain IV of the voltage-gated sodium channel alpha subunit (Poli et al., 1986; Rein et al., 1994; Trainer et al., 1991, 1994). It is believed, based on various lines of evidence, that these polyether toxins orientate across the plasma membrane, parallel to selected helices in the alpha subunit of the channel (Gawley et al., 1995; Trainer et al., 1994). Orientation of the brevetoxins is said to be "head-down" with the A-ring lactone facing the cell interior and the "tail" end of the molecule facing outward. These hypothetical features are supported by photoaffinity probe covalent modification (Trainer et al., 1991, 1994) and preliminary electrophysiological studies (Gawley et al., 1995).

As part of our continuing efforts to describe the effects of the brevetoxins on voltage-sensitive Na⁺ channels at the molecular level, the effects of natural toxins on single Na⁺ channels were examined under patch-clamp conditions. Appropriate brevetoxin derivatives were then specifically designed and synthesized in our laboratories, based on results of computer molecular modeling and specific radioreceptor protocols, and examined electrophysiologically. A preliminary report of some of these results has been presented (Jeglitsch et al., 1994).

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Preparation. Sensory neurons from rat nodose ganglia were isolated as described previously (Baden et al., 1994). The nodose ganglia were dissected from neonatal rats (1-2 weeks old), sacrificed by intraperitoneal injection of sodium pentobarbital and incubated in a Krebs' solution containing collagenase (0.6 mg·ml¹, type 2; Worthington Biochemical Corp., Freehold, NJ) and protease (0.4 mg·ml¹, Sigma Chemical Co., St. Louis, MO) for 1 hr at 37°C. The ganglia were transferred to a sterile culture medium (Dulbecco's Modified Eagle's Medium, 10% (v/v) fetal calf serum, 100 U·ml¹ penicillin, 0.1 mg·ml¹ streptomycin) and mechanically dissociated by a pair of forceps. The tissue fragments were triturated with a sterile fine-bore Pasteur pipette, and the dissociated neurons were plated onto laminin-coated 18-mm glass coverslips. The dissociated cells were incubated at 37°C in a 95% air, 5% CO₂ atmosphere, and electrophysiological recordings were made from neurons that had been maintained in tissue culture 36 to 72 hr. Isolated nodose ganglion neurons studied had diameters of 15 to 25 µm.

Electrophysiological recording. Single Na⁺ channel currents were recorded from cell-attached membrane patches by the patch-clamp technique (Hamill et al., 1981). Patch pipettes were fabricated from thick-walled borosilicate glass (GC150F, Clark Electromedical Instruments, Reading, UK) and the tip was fire-polished and coated with Sylgard (Dow Corning, Midland, MI) to reduce electrode capacitance. Pipettes had tip resistances of 10 to 15 megohm when filled with pipette solution. Patch electrodes sealed against the cell membrane typically had seal resistances of 5 to 10 gigaohm. Voltage steps were applied with pulse protocols generated by a PC computer (Dell 486D/50 MHz) and pClamp programs (Axon Instruments Inc., Foster City, CA). The transmembrane potential was initially held at 100 mV (the measured resting membrane potential was ~50 mV) before depolarizing voltage steps were applied. Single-channel currents were recorded with an Axopatch 200A patch-clamp amplifier (Axon Instruments Inc., Foster City, CA), filtered at 2 kHz (3 dB, 4-pole low-pass Bessel filter) and sampled at 10 kHz with an A/D converter (TL-1 Labmaster DMA interface) and computer with pClamp programs. Linear leak and capacitative currents were subtracted from current traces with records that contained no channel openings during the depolarizing pulse. Steady-state single-channel currents were recorded on videotape with an A/D recorder adaptor (PCM-2, Medical Systems, Greenvale, NY) before computer analysis.

Solutions and toxins. The composition of the physiological salt solution was: 140 mM NaCl, 3 mM KCl, 0.6 mM MgCl₂, 2.5 mM CaCl₂, 7.7 mM glucose, 10 mM histidine, adjusted to pH 7.4 with NaOH. The extracellular solution used for whole-cell recordings contained: 50 mM NaCl, 3 mM MgCl₂, 0.1 mM CdCl₂, 3 mM KCl, 90 mM TEA-Cl, 7.7 mM glucose, 10 mM HEPES adjusted to pH 7.4 with TEA-OH. In a series of experiments, whole-cell recordings were made in the presence of 1 µM TTX added to the bath solution. The intracellular pipette solution for whole-cell recordings contained: 10 mM NaF, 90 mM CsCl, 2 mM Mg₂ATP, 10 mM EGTA, 7.7 mM glucose, 10 mM HEPES adjusted to pH 7.2 with CsOH. The internal pipette solution contained a high Cs⁺ concentration and the external bath solution contained TEA to block K⁺ channels. To suppress depolarization-activated Ca⁺⁺ currents, EGTA and F were used to buffer intracellular Ca⁺⁺, and extracellular Ca⁺⁺ was replaced by Mg⁺⁺. Ca⁺⁺ currents were further suppressed by the addition of 100 µM Cd⁺⁺ to the external solution. Cd⁺⁺ was used at low concentrations (100 µM) because Cd⁺⁺ has been reported to reduce the TTX-resistant Na⁺ current by 50% at concentrations >500 µM (Ikeda et al., 1986).

1

View larger version (17K): [in this window] [in a new window]   Fig. 1.   The molecular structures of brevetoxin B and derivatives that bind to "site 5" of the voltage-gated sodium channel. The structures of brevetoxin-3 (PbTx-3), PbTx-6 and the derivatization of PbTx-3 to produce the derivatives, 2,3-dihydro-PbTx-3 A-ring diol (compound 3), 2,3,27,28,41,43-hexahydro-PbTx-3 (compound 2) and 2,3,41,43-tetrahydro-PbTx-3 (compound 1) used in this study are shown.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Preliminary voltage-clamp experiments were carried out on neonatal rat nodose ganglion cells to determine the TTX sensitivity of the whole-cell Na⁺ current. The average resting membrane potential measured in physiological salt solution was 51.1 ± 1.0 mV (n = 15). In the absence of TTX, inward Na⁺ currents activate at potentials more depolarized than 50 mV, reach a maximum amplitude at approximately 20 mV and reversed near +40 mV. In the presence of 1 µM TTX, no current was obtained upon step depolarization in 10 mV increments from 90 mV to +50 mV (n = 15, data not shown). The effect of TTX was reversible, with the inward Na⁺ current reappearing upon washout of TTX.

Isolation of single Na⁺ channel currents in cell-attached membrane patches. To separate single Na⁺ channel currents from K⁺, Ca⁺⁺ and Cl currents, specific pharmacological agents were added exclusively to the pipette solution. K⁺ currents were inhibited by the addition of Ba⁺⁺ and TEA, and Cl channels were blocked by adding 100 µM DIDS to the pipette solution. Ca⁺⁺ currents were suppressed by substituting Mg⁺⁺ for Ca⁺⁺ in the bath solution.

Activation and inactivation kinetics of PbTx-3-modified and unmodified Na⁺ channels. In the absence of PbTx-3, no single-channel activity was observed in membrane patches under steady-state conditions; that is, during maintained depolarization (data not shown). Under control conditions, voltage-dependent Na⁺ channels can only be observed following step depolarizations applied to the membrane patch. Brief (5 ms), single-channel openings were observed immediately after a depolarizing step (fig. 2A). Maintained depolarization at voltages more positive than 30 mV inactivated the Na⁺ channels and no channel openings were observed.

View larger version (29K): [in this window] [in a new window]   Fig. 2.   Depolarization-activated single sodium channel currents in the absence (A) and presence (B) of 100 nM PbTx-3. (A, B) Typical consecutive traces (200 ms duration) of unitary Na+ currents (downward deflections) obtained on depolarization to 30 mV from 100 mV (resting membrane potential  50 mV) in the absence (Control) and presence of 100 nM PbTx-3. The voltage jump protocol is shown above the current traces. Pipette and bath solutions are as described under "Materials and Methods." (C, D) Ensemble average of summarized Na+ inward current obtained in the absence (Control; C) and presence of 100 nM PbTx-3 (D) by averaging 40 consecutive traces as shown in A and B, respectively. The dashed line indicates the zero-current level.

View larger version (13K): [in this window] [in a new window]   Fig. 3.   Open probability of the voltage-dependent Na+ channel in the absence (Control) and presence of 100 nM PbTx-3. (A) The normalized total open time of the Na+ channel calculated at different membrane potentials in the absence (Control, ) and presence of 100 nM PbTx-3 (). Each data point represents mean ± S.E.M. calculated from three experiments. (B) The normalized product of the number of channels and channel open probability (NPopen) ± S.E. of the unmodified () and PbTx-modified () Na+ channel calculated from the data shown in A. The Na+ channel activation curves obtained in the absence (Control) and presence of 100 nM PbTx-3 are fit by a Boltzmann distribution with midpoints of 20 mV and 30 mV and slope parameters of 5.2 and 4.2, respectively.

Single Na⁺ channel events observed under steady-state conditions. PbTx-3 has been reported to inhibit the inactivation of voltage-sensitive Na⁺ channels (Baden, 1989; Wu and Narahashi, 1988). The modified Na⁺ channel can be studied not only with voltage jumps but also under steady-state conditions (see fig. 4A). Inhibition of Na⁺ channel inactivation by PbTx-3 can be observed during maintained depolarization at 50 mV for several minutes (see fig. 5 in Baden et al., 1994). PbTx-3 stabilizes the open conformation whereby channel openings can be observed during the entire period. The modified channel switches between the closed and open conformations and does not appear to inactivate. Although the Na⁺ channel activity remains unchanged during the course of the recording, the unitary currents were inhibited if the patch pipette was backfilled with 1 µM TTX (see above).

View larger version (26K): [in this window] [in a new window]   Fig. 4.   Subconductance levels of PbTx-3-modified Na+ channels. (A) Steady-state recordings of single Na+ channel events obtained at membrane potentials of 20 mV, 40 mV and 60 mV in the presence of 500 nM PbTx-3. Two distinct conductance (open) states (level I, II) are observed at each potential. The closed state is indicated by (C). (B) Unitary Na+ current-voltage relationship. Mean current amplitudes ± S.E. of the two conductance states obtained from three different cell-attached membrane patches in the presence of 500 nM PbTx-3. The slope conductance for level I is 10.7 ± .8 pS and 21.2 ± .6 pS for level II. The extrapolated reversal (zero-current) potential is approximately +60 mV. (C) Amplitude histogram of unitary Na+ currents obtained from a 1-min steady-state recording in the presence of 500 nM PbTx-3. Holding potential, 50 mV.

Sublevel gating and slope conductances. In the absence of PbTx-3, the brief Na⁺ channel openings observed on step depolarization of cell-attached patches had a single-channel conductance of ~21 pS. PbTx-3, unlike other known gating modifiers of the voltage-dependent Na⁺ channel, is able to stabilize more than one distinct conductance level (Schreibmayer and Jeglitsch, 1992). Figure 4A shows PbTx-3-induced sublevel behavior recorded under steady-state conditions at different membrane potentials. The slope conductances determined for the two conductance levels were 10.7 ± .8 pS for level I and 21.2 ± .6 pS for level II (n = 3, fig. 4B). The single-channel conductance (_Na) for level II is twice that of level I. The extrapolated reversal potentials obtained for the unitary current-voltage relationships in the absence and presence of PbTx-3 were similar (~+60 mV) and as predicted by the Nernst equation for a Na⁺-selective electrode, which suggests that the ionic selectivity of the Na⁺ channel is not modified by PbTx-3.

Apparent open time of the unmodified and PbTx-3-modified Na⁺ channel. PbTx-3 not only inhibits Na⁺ channel inactivation and shifts the activation curve but it also increases the open-channel probability. The increase in Na⁺ channel open probability is caused by a decrease in the closed time between individual openings on depolarization and an increase in the single-channel open time as shown in figure 5A. Under control conditions, only brief openings are observed immediately after step depolarizations. In the presence of PbTx-3, Na⁺ channel openings are observed during maintained depolarization and individual channel openings have significantly longer open times than those of the unmodified Na⁺ channel. PbTx-3 appears to stabilize the open (conducting) conformation in a concentration-dependent manner. The highest channel activity was observed in the presence of 500 nM PbTx-3, whereas 3 nM PbTx-3 had no effect on Na⁺ channel activity (data not shown). The mean apparent open time (_o) was increased in the presence of both 30 and 500 nM PbTx-3 with a maximum increase in _o of ~80% in the presence of 500 nM PbTx-3 (fig. 5B). The increase in _o was statistically significant at each membrane potential from 40 to 0 mV and saturated at PbTx-3 concentrations 500 nM. In the presence of PbTx-3, Na⁺ channel openings are obtained at more negative potentials than control conditions, because of the shift in the activation curve.

View larger version (21K): [in this window] [in a new window]   Fig. 5.   Voltage dependence of Na+ channel open time in the absence and presence of PbTx-3. (A) Unmodified (control) and modified (30 nM and 500 nM PbTx-3) single Na+ channel currents evoked by step depolarizations to 40 mV (lower current traces) and 10 mV (upper current traces) from a holding potential of 100 mV in cell-attached membrane patches. (B) The mean apparent open time ( o) obtained at different membrane potentials in the absence (Control, ) and presence of 30 nM () and 500 nM () PbTx-3. Each data point represents mean ± S.E. calculated from at least three experiments.

Effects and toxicity of various PbTx-3 derivatives. To examine the effects of modification of the proposed active sites of the PbTx-3 molecule, the actions of the following PbTx-3 derivatives on single Na⁺ channel activity were determined and compared with the naturally occurring toxin. 2,3-Dihydro-PbTx-3 A-ring diol (compound 3) was previously examined at various concentrations on single Na⁺ channels (Baden et al., 1994). In the presence of 100 nM compound 3, only brief openings occurred immediately after the step depolarization, similar to those observed under control conditions (see fig. 2A). Within the concentration range examined, compound 3 (30-500 nM) had no effect on Na⁺ channel activity. Raising the concentration of compound 3 to 1 µM caused a slight increase in Na⁺ channel activity (n = 3; fig. 6A), whereas 10 µM compound 3 produced a substantial increase in Na⁺ channel open probability (see fig. 4, Baden et al., 1994). The mean apparent open time was increased significantly at all voltages compared with control, and to the same degree as that observed with 100 nM PbTx-3 (fig. 2B). Compound 3 was approximately 100-fold less potent than PbTx-3 and concentrations >10 µM did not further increase the mean apparent open time. Under steady-state conditions, multiple subconductance levels of the open Na⁺ channel were observed in the presence 10 µM compound 3 similar to those reported previously (data not shown; see Gawley et al., 1995).

View larger version (22K): [in this window] [in a new window]   Fig. 6.   Voltage dependence of Na+ channel open time in the presence of 2,3-dihydro-PbTx-3 A-ring diol (compound 3) and 2,3,27,28,41,43-hexahydro-PbTx-3 (compound 2). (A) Modified (1 µM compound 3, 1 µM compound 2) single Na+ channel currents recorded from cell-attached membrane patches upon step depolarizations to 40 mV (lower current traces) and 10 mV (upper current traces) from a holding potential of 100 mV. (B) The mean apparent open time ( o) obtained at different membrane potentials in the absence (Control, ) and presence of 1 µM compound 3 () and 1 µM compound 2 (). The dashed line indicates  o values obtained in the presence of 500 nM PbTx-3. Each data point represents mean ± S.E. calculated from at least three experiments.

View larger version (22K): [in this window] [in a new window]   Fig. 7.   Voltage dependence of Na+ channel open time in the presence of PbTx-6 and 2,3,41,43-tetrahydro-PbTx-3 (compound 1). (A) Modified (500 nM PbTx-6, 500 nM compound 1) single Na+ channel currents recorded from cell-attached membrane patches in response to step depolarizations to 40 mV (lower current traces) and 10 mV (upper current traces) from a holding potential of 100 mV. (B) The mean apparent open time ( o) obtained at different membrane potentials in the absence (Control, ) and presence of 500 nM PbTx-6 () and 500 nM compound 1 (). The dashed line indicates  o values obtained in the presence of 500 nM PbTx-3. Each point represents mean ± S.E. calculated from at least three experiments.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

The effects of brevetoxin-3 and its derivatives were studied in TTX-sensitive sodium channels in rat nodose sensory neurons. Previous studies of neonatal and adult vertebrate sensory neurons have described at least two different types of Na⁺ currents according to their sensitivity to TTX: TTX-sensitive (TTX_s) and TTX-resistant (TTX_r). The two types of Na⁺ currents exhibit different kinetics: one with fast kinetics that is blocked by TTX at nanomolar concentrations and the other exhibiting slower kinetics that is TTX-resistant (Campbell, 1993; Elliott and Elliott, 1993; Ikeda et al., 1986). Voltage-clamp studies in adult rat nodose ganglion cells (average diameter, 35 µm) similarly reveal both TTX-sensitive and TTX-resistant Na⁺ currents. In the presence of 3 to 15 µM TTX, the TTX-sensitive Na⁺ channels are blocked completely, whereas most of the TTX-resistant inward current persists (Ikeda et al., 1986). In contrast to neurons obtained from adult rats, neurons from the nodose ganglia of neonatal rats appear to express only one type of Na⁺ channel which is blocked reversibly by 1 µM TTX. Complete block of Na⁺ current at this relatively low TTX concentration indicates that the Na⁺ inward current is caused exclusively by TTX-sensitive Na⁺ channels.

The observed effects of PbTx-3 on voltage-dependent Na⁺ channels in rat nodose ganglion neurons can be summarized as follows: 1) a shift of activation by 10 mV to more negative membrane potentials, 2) an inhibition of Na⁺ channel inactivation at all membrane potentials whereby single Na⁺ channel activity is observed during maintained depolarization, 3) an approximately 2-fold increase of the mean apparent channel open time at all membrane potentials and 4) the appearance of subconductance states with two distinct conductance levels detected (10.7 pS and 21.2 pS). These effects were observed from 30 to 500 nM PbTx-3. All effects of PbTx-3 observed in rat nodose neurons are consistent with those observed on Na⁺ channels in rat cardiac myocytes (Schreibmayer and Jeglitsch, 1992) but differ from those observed in neuroblastoma cells (Sheridan and Adler, 1989). The lack of effect of PbTx-3 on Na⁺ channel inactivation observed in neuroblastoma cells may partly be caused by the short-duration voltage pulses used to examine the kinetics of Na⁺ current inactivation in whole-cell and single-channel experiments.

All PbTx-3 derivatives examined caused a similar shift in activation to more negative membrane potentials, albeit at higher concentrations than found with PbTx-3. Similarly, inhibition of Na⁺ channel inactivation was observed for all PbTx-3 derivatives but to a lesser extent than that observed with PbTx-3. The mean apparent open times were increased by 30 to 50% compared with control by all PbTx-3 derivatives, but to a lesser degree than by PbTx-3. The effect of PbTx-3 derivatives on mean channel open time appears to be related to their binding affinity for the Na⁺ channel, whereby those derivatives with a low binding affinity exhibited a smaller effect (Rein et al., 1994). Subconductance states of the Na⁺ channel were also observed in the presence of PbTx-3 derivatives as reported previously in rat nodose neurons (Gawley et al., 1995). In the presence of PbTx-3, only two conductance levels were observed, whereas with 2,3-dihydro-PbTx-3 A-ring diol (compound 3), a minimum of five conductance levels were observed (Gawley et al., 1995).

The brevetoxin molecule has been shown to bind to receptor site 5 with the tail end of the molecule located near the S5-S6 extracellular loop of domain IV of the Na⁺ channel alpha subunit (Trainer et al., 1991, 1994). The finding that all derivatives of brevetoxin examined to date shift the voltage dependence of activation to more negative potentials indicates that the oxygen-rich nature of the brevetoxin backbone interacts with the channel in a manner that stabilizes the open configuration.

According to the model of the voltage-sensitive Na⁺ channel by Sato and Matsumoto (1995), all four extracellular loops formed by S5-S6 of all domains have to slide into the core pore formed by S2-S4 segments for full activation of the channel. In this case, the A-ring end of the PbTx-3 molecule may reach far enough into the membrane to interact with the charges of the inactivation gate and likely lead to inhibition of Na⁺ channel inactivation. In the presence of PbTx-3 and its derivatives, the open conformation is stabilized and the induction of subconductance states is observed in the toxin-modified Na⁺ channel. The induction of subconductance states may be caused by the character of the A-ring and its proximity to the inactivation loop of the channel. This is consistent with results obtained with the truncated brevetoxin-B possessing all the "essential" chemical elements except length (20.4 Å compared with approximately 30 Å) (Nicolaou et al., 1994) and the five conductance states induced by the 2,3-dihydro-PbTx-3 derivative (Gawley et al., 1995). The multiple subconductance states may be caused by the interaction of the "head" end of the toxin with the inactivation particle on the inside face of the channel associated with the loop between domains III and IV, which results in "partially open" (or partially closed) channels by differential physical occlusion. The potential for this physical misalignment of the inactivation particle is supported partly by computer modeling, which indicates a fixed angle and direction of orientation of the A-ring lactone in all closed ring toxins examined, and a general free range of movement for the "arms" of any A-ring-cleaved toxin (Rein et al., 1994). The mechanism(s) by which brevetoxins and derivatives generate the subconductance states remains to be elucidated, however.

The induction of longer mean apparent open times is a characteristic of all natural brevetoxins and synthetic derivatives, except the 1-deoxy-PbTx-3 derivative, which indicates that the carbonyl oxygen may be important for maintaining channels in the open configuration (Gawley et al., 1995). Given that voltage-dependent Na⁺ channels have been proposed to activate and inactivate based on changing dipole pairs on adjacent -helices of channel domains (Catterall, 1992; Patlak, 1991), the effect of 1-deoxy-PbTx-3 toxin (which does not induce longer mean open times) is consistent with the importance of dipole interactions in brevetoxin stabilizing channels in the open configuration.

The overall length of the brevetoxin molecule may be the controlling factor in inhibition of inactivation once Na⁺ channels have been activated. Again, results obtained with the 1-deoxy-PbTx-3 derivative indicate that an intact carbonyl moiety is not necessary for activity, whereas without the proper length (e.g., truncated brevetoxin), the presence of an intact A-ring is not sufficient to inhibit Na⁺ channel inactivation (Gawley et al., 1995). The results of the present study have identified some of the salient features associated with polyether toxin modification of voltage-sensitive Na⁺ channel gating and provide useful insights into the factors that modulate sodium channel function.