The active guanidinium group of saxitoxin and neosaxitoxin identified by the effects of pH on their activities on squid axon

Specific neosaxitoxin interactions with the Na+ channel outer vestibule determined by mutant cycle analysis

The voltage-gated Na+ channel alpha-subunit consists of four homologous domains arranged circumferentially to form the pore. Several neurotoxins, including saxitoxin (STX), block the pore by binding to the outer vestibule of this permeation pathway, which is composed of four pore-forming loops (P-loops), one from each domain. Neosaxitoxin (neoSTX) is a variant of STX that differs only by having an additional hydroxyl group at the N1 position of the 1,2,3 guanidinium (N1-OH). We used this structural variant in mutant cycle experiments to determine interactions of the N1-OH and its guanidinium with the outer vestibule. NeoSTX had a higher affinity for the adult rat skeletal muscle Na+ channel (muI or Scn4a) than for STX (DeltaG approximately = 1.3 kcal/mol). Mutant cycle analysis identified groups that potentially interacted with each other. The N1 toxin site interacted most strongly with muI Asp-400 and Tyr-401. The interaction between the N1-OH of neoSTX and Tyr-401 was attractive (DeltaDeltaG = -1.3 +/- 0.1 kcal/mol), probably with formation of a hydrogen bond. A second possible attractive interaction to Asp-1532 was identified. There was repulsion between Asp-400 and the N1-OH (DeltaDeltaG = 1.4 +/- 0.1 kcal/mol), and kinetic analysis further suggested that the N1-OH was interacting negatively with Asp-400 at the transition state. Changes in pH altered the affinity of neoSTX, as would be expected if the N1-OH site were partially deprotonated. These interactions offer an explanation for most of the difference in blocking efficacy between neoSTX and STX and for the sensitivity of neoSTX to pH. Kinetic analysis suggested significant differences in coupling energies between the transition and the equilibrium, bound states. This is the first report to identify points of interaction between a channel and a non-peptide toxin. This interaction pattern was consistent with previous proposals describing the interactions of STX with the outer vestibule (Lipkind, G. M., and H. A. Fozzard. 1994. Biophys. J. 66:1-13; Penzotti, J. L., G. Lipkind, H. A. Fozzard, and S. C. Dudley, Jr. 1998. Biophys. J. 75:2647-2657).

The potencies of synthetic analogues of saxitoxin and the absolute stereoselectivity of decarbamoyl saxitoxin

The potencies of synthetic saxitoxin (+/- STX) and six of its synthetic analogues, including the enantioselectively synthesized unnatural (-)enantiomer of decarbamoyl saxitoxin (dcSTX), were measured and compared to those of natural saxitoxin [(+)STX]. The analogues, all of which were racemic (+/-) mixtures except for dcSTX, varied in the substituents at the C6 position, the carbamoyl 'moeity', and the C12 position, the hydrated ketone. The ability of the toxins to inhibit the compound action potential (AP) and to displace radiolabeled natural saxitoxin (3H-STX) from nerve membranes at equilibrium were both used as potency assays. Biological activity of both (+)- and (-)dcSTX was analyzed by the kinetics of block of single Na+ channels reconstituted in planar lipid bilayer membranes, where it was demonstrated that only (+)dcSTX had biological activity. The potency of STX analogues fell markedly as the substituent at the C6 position became smaller; Ki values from the binding competition assay (at 4 degrees C) are: (+/-)6-methanolic-STX, 5 x 10(-10) M; (+/-)6-methyl-STX, 1 x 10(-6) M; (+/-)6-dihydro-STX, 3.5 x 10(-5) M. Replacement of the ketone at the C12 position by a methylene group was accomplished in two derivatives, although both also had substituents at the C6 position. The compound (+/-)6-methyl-12-deoxy-STX was about 0.03 as potent as (+/-)6-methyl-STX and only 10(-5) as potent as racemic (+/-)STX. In synthetic compounds where the benzyloxymethyl (-CH2OCH2C6H5) substituent occurred at the C6 position, the C12-methylene derivative still displayed some binding activity (Ki = 6 x 10(-4) M). However, when the same C6 derivatized compounds also contained a 6-membered heterocyclic group (-C3H8S2-) conjugated to carbon 12, the measured binding affinity was even further decreased (Ki = 2 x 10(-3) M). The findings show that substitutions on the carbon 6 position of STX have stronger effects on STX potency than previously believed, and that the toxin may form a hydrogen bond with the sodium channel at this site. Furthermore, the total removal of oxygen from the C12 position does not completely abolish the binding activity of the molecule.

Interactions of neosaxitoxin with the sodium channel of the frog skeletal muscle fiber

Neosaxitoxin (neoSTX) differs structurally from saxitoxin (STX) in that the hydrogen on N-1 is replaced by a hydroxyl group. On single frog skeletal muscle fibers in the vaseline-gap voltage clamp, the concentrations for reducing the maximum sodium current by 50% (ED50) at pH's 6.50, 7.25, and 8.25 are, respectively, 4.9, 5.1, and 8.9 nM for STX and 1.6, 2.7, and 17.2 nM for neoSTX. The relative potencies of STX at the different pH's closely parallel the relative abundance of the protonated form of the 7,8,9 guanidinium function, but the relative potencies of neoSTX at the same pH's vary with the relative abundance of the deprotonated N-1 group. In constant-ratio mixtures of the two toxins, the observed ED50's are consistent with the notion that the two toxins compete for the same site. At pH's 6.50 and 7.25, the best agreement between observed and computed values is obtained when the efficacy term (epsilon) for either toxin is 1. At pH 8.25 the best agreement is obtained if the efficacy is 1 for STX but 0.75 for neo-STX. The marked pH dependence of the actions of neoSTX probably reflects the presence of a site in the receptor that interacts with the N-1 -OH, in addition to those interacting with the 7,8,9 guanidinium and the C-12 hydroxyl groups. Considering the three-dimensional structure of the STX and neoSTX molecules, the various site points are probably located in a fold or a crevice of the channel protein, where the extracellular orifice of the sodium channel is located.

Effects of saxitoxin analogues and ligand competition on sodium currents of squid axons

Saxitoxin (STX) and several STX analogues from dinoflagellates (genus Protogonyaulax) block sodium conductance in squid giant axons with variable potencies. Toxins, analyzed under voltage clamp, are 21-sulfosaxitoxin, 21-sulfosaxitoxin 11 alpha-hydroxysulfate, 21-sulfosaxitoxin 11 beta-hydroxysulfate, (B1, C1, C2, respectively) and gonyautoxins 2 and 3. The potency sequence for the toxins examined is STX greater than gonyautoxin 3 greater than B1 greater than C2 greater than gonyautoxin 2 much greater than C1. Guanidine, when substituted for sodium in external seawater, reduced the potency of STX to block inward current but did not affect tetrodotoxin activity. Methylguanidine also reduced the ability of STX to block outward sodium current. Inhibitory constants for guanidine and methylguanidine were 116 and 187 mM, respectively. Competition can be explained by binding at or near the toxin binding site but not by surface potential alteration.

Actions of epimers of 12-(OH)-reduced saxitoxin and of 11-(OSO3)-saxitoxin on squid axon

The actions of the 12 alpha-saxitoxinol, 12 beta-saxitoxinol and a C-12 ethylene thioketal derivative of saxitoxin, as well as those of 11 alpha-(OSO3)-saxitoxin, 11 beta-(OSO3)-saxitoxin and 11 alpha-(OH)-saxitoxin, have been examined on the isolated squid giant axon. Each of these analogues acted similarly to saxitoxin in blocking specifically the sodium channel. The relative potencies are: STX (1); 11 beta-(OSO3)-STX (gonyautoxin III) (0.42); 11 alpha-(OSO3)-STX (gonyautoxin II) (0.20); 11 alpha-(OH)-STX (0.10); 12 alpha-saxitoxinol (0.0021); 12 beta-saxitoxinol (0.0005). Thus, the presence of a bulky and negatively charged sulphate group on C-11 does not materially affect the biological activity of STX. Hydrogen bonding at the C-12 position is probably an important means of binding of STX to the membrane receptor site. The difference between the epimers of saxitoxinol suggests that the H in one of them may be geometrically better aligned than that in the other, with the hydrogen acceptor group in the receptor.

Voltage-dependent blockade of muscle Na+ channels by guanidinium toxins

Na+ channels from rat muscle plasma membrane vesicles were inserted into neutral planar phospholipid bilayers and were activated by batrachotoxin. Single channel blocking events induced by the addition of various guanidinium toxins were analyzed to derive the rates of channel-toxin association and dissociation. Blocking by tetrodotoxin, saxitoxin, and six natural saxitoxin derivatives containing sulfate or hydroxyl groups were studied. Although the binding affinities vary over 2,000-fold, all of the toxins exhibit identical voltage dependence of the blocking reactions, regardless of the toxin's net charge. The results suggest that the voltage dependence of toxin binding is due to a voltage-dependent conformational equilibrium of the toxin receptor, rather than to direct entry of the charged toxin molecule into the applied transmembrane electric field.