INTRODUCTION

The high toxicity of brevetoxins (PbTx-n),¹ lipid-soluble polyether marine toxins produced by the ``red tides'' dinoflagellate Ptychodiscus brevis, to human and fish was found to be due to their high affinity binding to specific receptor site (site 5) on voltage-sensitive sodium channels (1). Brevetoxin binding results in membrane depolarization due to a shift of the channel activation to more negative membrane potentials and inhibition of normal inactivation (2, 3). Sodium channels are composed of about 2000 amino acids organized in four repeated homologous domains (I-IV) each consisting of six putative transmembrane segments (S1-S6) (4). Partial localization of receptor site 5 has been suggested recently, using photolabeled derivative of PbTx-3 and site-directed antibody mapping, to be in the region of interaction of transmembrane segments S6 and S5 of homology domains I and IV, respectively, of rat brain sodium channels (5).

Since 1981, the binding of different brevetoxins and the allosteric interaction between receptor site 5 and other neurotoxins' receptors on sodium channels have been extensively studied using neuroblastoma cells and rat brain synaptosomes (6, 7, 8, 9, 10, 11, 12, 13). Scorpion -toxins were shown to inhibit inactivation of sodium channels, and their binding was demonstrated to be allosterically enhanced by alkaloid toxins (such as batrachotoxin and veratridine) (4) but not by brevetoxins. PbTx-1 and PbTx-2 were shown not to affect the binding of the scorpion -toxin Lqq V (6, 7, 9).

In contrast to the previous results, we have recently demonstrated that PbTx-1, the most active brevetoxin analog (1), induces a strong negative allosteric modulation on the binding of another scorpion -toxin, AaH II, on rat brain sodium channels (14). The binding affinity of AaH II to rat brain synaptosomes was shown to be the highest among all scorpion -toxins (K_d = 0.2-0.3 nM) (14, 15), about 10-fold higher than that of Lqq V (K_d = 2.2-8 nM) (16). The structural differences between these two pharmacologically similar scorpion -toxins, both inhibiting sodium channel inactivation and binding to receptor site 3 in a voltage-dependent manner on mammalian sodium channels (15, 16, 17) may account for the significant difference in binding affinity. In addition, the various brevetoxins were shown to have different toxicity and differentially affect the binding of [³H]batrachotoxin derivative (1, 11), related to their carbon backbone structure (1, 12).

These considerations led us to attribute the differences between ours and preceding results concerning the previously undetectable strong allosteric modulation by PbTx-1 on the scorpion -toxin binding to: 1) structural differences between the two scorpion -toxins used, AaH II in our case and Lqq V in the others, suggesting that interaction between brevetoxin and scorpion -toxin site may be, at least in part, toxin specific; and 2) the inherent structural differences in the brevetoxin analogs (1, 12) used in ours and previous studies (14). In the present work we examined these possibilities and found that none can account for the observed discrepancy, suggesting that still other, unexpected reasons should exist. We found that the routine presence of TTX may account for that discrepancy.

EXPERIMENTAL PROCEDURES

Scorpion toxins AaH II and Lqq V were purified as described previously (27). Brevetoxins (PbTx-1, -2, -3, and -9) and tetrodotoxin were from Latoxan (Rosans, France). Carrier-free Na¹²⁵I was from Amersham Corp. All other chemicals were of analytical grade. Filters for binding assays were glass fiber GF/C (Whatman, United Kingdom) preincubated in 0.3% polyethyleneimine (Sigma).

Rat brain synaptosomes were prepared from adult albino Wistar rats (about 300 g, laboratory bred), according to Ref. 26. All buffers contained a mixture of proteinase inhibitors composed of 50 µg/ml phenylmethylsulfonyl fluoride, 1 µM pepstatin A, 1 mM iodoacetamide, and 1 mM 1,10-phenanthroline. Membrane protein concentration was determined using a Bio-Rad protein assay, with bovine serum albumin as standard.

AaH II and Lqq V were radioiodinated by lactoperoxidase and monoiodotoxins were purified as described previously (14). The concentration of the radiolabeled toxins was determined according to the specific activity of the ¹²⁵I corresponding to 2424 dpm/fmol monoiodotoxin. Equilibrium competition and saturation assays were performed using increasing concentrations of the unlabeled toxin in the presence of a constant low concentration of the radioactive toxin. Standard binding medium composition was: (in mM): choline chloride 140, CaCl₂ 1.8, KCl 5.4, MgSO₄ 0.8, HEPES 25, glucose 10, bovine serum albumin 2 mg/ml, pH 7.4. Wash buffer composition was (in mM): choline chloride 140, CaCl₂ 1.8, KCl 5.4, MgSO₄ 0.8, HEPES 25, bovine serum albumin 5 mg/ml, pH 7.4.

Rat brain synaptosomes (0.1-0.15 mg of protein/ml) were suspended in 0.15 ml of binding buffer, containing ¹²⁵I-AaH II or ¹²⁵I-Lqq V. After incubation for 30 min at 37 °C, the reaction mixture was diluted with 2 ml of ice-cold wash buffer and filtered through GF/C under vacuum. Filters were rapidly washed with an additional 2 × 2 ml of buffer. Nonspecific toxin binding was determined in the presence of 0.2 µM unlabeled AaH II or 1.6 µM Lqq V, respectively, and consist typically of 15-20% of total binding for ¹²⁵I-AaH II or 30-40% for ¹²⁵I-Lqq V. Equilibrium saturation or competition experiments were analyzed by the iterative computer program LIGAND (Elsevier Biosoft). Each experiment was performed at least three times.

RESULTS

We have shown previously that brevetoxin (PbTx-1) inhibition on AaH II binding results from a 3.6-fold reduction in AaH II binding affinity (with no significant effect on receptor capacity), due to increasing the dissociation rate constant, consistent with a negative allosteric interaction between brevetoxin receptor site 5 and scorpion -toxin receptor site 3 (14). Fig. 1 demonstrates that both Lqq V and AaH II binding are significantly inhibited by PbTx-1. Lqq V-specific binding is similarly inhibited, but at apparent higher PbTx-1 concentration (Fig. 1). This result excluded the explanation in possibility 1 (see Introduction), since both scorpion -toxins binding is shown to be negatively modulated by brevetoxin, although with some quantitative differences.

Brevetoxin PbTx-1 allosterically inhibits the binding of scorpion -toxins AaH II and Lqq V to rat brain sodium channels.

Different brevetoxin analogs, PbTx-1, PbTx-2, PbTx-3 and PbTx-9, are shown in Fig. 2 to inhibit both scorpion -toxins binding; PbTx-1 being the most active one. As suggested in Fig. 1, the binding inhibition of the two toxins revealed significant quantitative differences (Fig. 2). To reveal possible structural basis for the above quantitative differences, we compared structural models of the two scorpion -toxins.

Brevetoxin analogs inhibit the binding of the two scorpion -toxins in quantitatively different manner.

AaH II and Lqq V share 63% of amino acid identity (72% similarity). Computer modeling discloses that the positively charged residues, previously suggested to participate in the receptor binding interaction (18, 19) are differentially located on the toxins' surface (data not shown). The presence of Lys-60 in Lqq V as well as the shift in position of arginine 54 of Lqq V as compared to arginine 56 in AaH II, suggested to be exposed on the active surface of the toxins, may contribute to the quantitative differences in toxin's binding modulation by brevetoxins. The recent identification of a negatively charged residue on sodium channels, which forms one of the recognition sites of scorpion -toxin receptor site 3 on homology domain IV of sodium channel (20), supports this notion. Thus, although the structural differences between the two toxins may contribute to the observed quantitative differences, they cannot account for the discrepancy between our results and the previous one.

Careful inspection of the former binding conditions revealed that TTX was routinely added to binding media used for scorpion -toxin binding modulation. TTX is most commonly used as specific blocker of sodium current in many studies and has been considered to plug the channel pore from the extracellular side, without any significant alteration of the sodium channel structure (21, 22). We examined the possible effect of TTX on the allosteric modulation of scorpion -toxin binding by brevetoxin. Surprisingly, our results (Fig. 3) elucidate an unexpected increase in AaH II binding by TTX, which is further enhanced by brevetoxin, despite the inhibitory effect of brevetoxin alone.

Tetrodotoxin reverses the inhibition of scorpion -toxin AaH II binding caused by brevetoxin PbTx-1.

As has been demonstrated previously (15, 16), TTX alone increases -scorpion binding (Fig. 3A) due to about 1.3-fold increase in the number of receptor sites with no change in affinity (16). Fig. 3A shows that TTX is able to abolish the inhibitory effect of PbTx-1 in a concentration-dependent manner. At 1 µM TTX, the concentration routinely used previously in the binding media (6, 7, 9), no significant effect of brevetoxin may be detected on the scorpion -toxin binding. This indicates that TTX induces a conformational change on the sodium channel, resulting in reversing the negative allosteric modulation by PbTx-1 on scorpion -toxin binding. Moreover, the increase in AaH II binding by TTX is enhanced by brevetoxin. This enhancement may not be explained by an additive effect. In the presence of brevetoxin, TTX increases AaH II binding by about 2-fold, as compared to about 1.3-fold increase by TTX alone (Fig. 3A). The enhancement of AaH II binding by TTX (with or without brevetoxin) is evident only at the resting membrane potential; at depolarized conditions (by high external potassium concentration; Ref. 16), TTX is not able to negate the decrease in scorpion -toxin binding (data not shown).

Scatchard analysis of AaH II binding reveals that the reversal of brevetoxin's negative effect on scorpion -toxin binding by TTX is due to a 2-fold increase in AaH II binding affinity (Fig. 3B). Brevetoxin at half saturating concentration (30 nM) increases the K_d of AaH II about 2-fold, in accordance with our previous results (14), and the presence of TTX decreases the K_d back to its original value. Thus, in the presence of both TTX and PbTx-1, the binding affinity of AaH II is about its value with no effectors (Fig. 3B; Ref. 14). This suggests that TTX in the presence of brevetoxin induces an allosteric effect on receptor site 3 that reverses or prevents the decrease in affinity induced by brevetoxin alone, resulting in scorpion -toxin binding around control level.

DISCUSSION

Our results indicate that TTX, the most used and studied sodium channels blocker, should be considered as an allosteric modifier of sodium channels, and new allosteric interactions are elucidated between receptor sites 1, 5, and 3, using scorpion -toxin as probe. The increase in scorpion -toxin binding by TTX (Refs. 15 and 16; Fig. 3) has been attributed previously to its block of sodium channel conductance and prevention of possible depolarization (known to reduce the scorpion -toxins binding, 21, 22). In contrast to previous interpretations (15, 16), we suggest that this enhancement may be attributed to a direct allosteric effect of TTX on receptor site 3, resulting in shifting more receptors to the high affinity state for scorpion -toxin, since all experiments have been performed in sodium-free media (Refs. 15 and 16; Fig. 3). The reversal of brevetoxin inhibition by TTX further supports its allosteric effect on receptor site 3.

These results indicate also the presence of allosteric interaction between TTX and brevetoxin receptor site, that are mainly detected at the region of scorpion -toxin receptor site. Previously, enhancement of [³H]saxitoxin binding by brevetoxin was reported in neuroblastoma cells (6) but not in rat brain synaptosomes (9). Tritiated brevetoxin binding was slightly enhanced (5-10%) in the presence of higher concentrations of saxitoxin or TTX (8). Together, previous and present results suggest the presence of allosteric interactions between TTX and brevetoxin receptor sites. However, such interaction may result in very limited conformational changes, as detected by the limited changes in binding at receptor sites 1 and 5.

Our present results may indicate that the binding of TTX and brevetoxin to their distinct receptor sites may induce additional, previously undetectable conformational changes on the sodium channel protein. Such conformational changes could not been detected by the study of TTX and PbTx-1 alone, suggesting that they may not be directly related to the function of each toxin. Rather, other conformational changes resulting from the binding of these toxins, may be detected using scorpion -toxin as sensitive probe for subtle conformational alterations in a different region comprising receptor site 3 on sodium channels. The significant reversal of brevetoxin-induced binding inhibition of AaH II by TTX is comparable with that previously observed by veratridine (14). Both sodium channel effectors (TTX and veratridine) could enhance the scorpion -toxin binding more efficiently in the presence of the negative modulator, PbTx-1 (Fig. 3; Ref. 14). Brevetoxins have been shown to allosterically enhance the binding of alkaloid toxins (9, 11). Our results suggest that allosteric interactions between TTX and PbTx-1 receptor sites are also present, but expressed more significantly in an indirect conformational changes sensed by scorpion -toxin binding. These may suggest an intimate dynamic relationship between at least some recognition sites comprising receptor site 3 and those of receptor sites 1 and 5.

This study may pave the way to a new insight into the dynamic conformational changes induced on the channel protein upon neurotoxin binding and action, revealing dynamic relationship between the pore region (receptor site 1) (23, 24), the transmembrane regions of segments S5 and S6 (receptor site 5) (5), and extracellular amino acid loops (receptor site 3) (20, 25) of the different domains of sodium channels. Considering the molecular mapping of these receptor sites (5, 20, 23, 24, 25), our study may contribute to the elucidation of the structural basis underlying the dynamic conformational changes responsible for the function of sodium channels.