Inhibition by the Channel Blockers Tetrodotoxin and Saxitoxin

The Journal of Neuroscience July 1966, 6(7): 2064-2070

3H-Batrachotoxinin-A Benzoate Binding to Voltage-Sensitive Sodium Channels: Inhibition by the Channel Blockers Tetrodotoxin and Saxitoxin

George B. Brown The Neurosciences Program and the Department of Psychiatry, The University of Alabama at Birmingham, Birmingham, Alabama 35294

The sodium channel blockers tetrodotoxin (TTX) and saxitoxin channel protein progresses through multiple closed, noncon- (STX) and the channel activator batrachotoxin (BTX) produce ducting states to an open, conducting conformation that is main- their effects by binding to separate and distinct sites on the tained only briefly before closing to a distinct inactivated state, channel protein. The fact that TTX- and STX-modified sodium different from the resting conformation (Armstrong, 198 1). The channels are blocked to sodium flux has precluded drawing any conformational state of the voltage-sensitive sodium channel direct conclusions regarding the effect of TTXSTX on BTX may also be affected by several specifically acting neurotoxins binding based on electrophysiological or 22Na flux measure- and, conversely, the binding characteristics of certain toxins are ments. Nevertheless, these sites have been presumed to be non- a function of sodium channel conformation. Because of this interacting. In this study, 3H-batrachotoxinin-A benzoate (BTX- relationship, biochemical and pharmacological study of these B), a tritiated congener of BTX, has been used to provide a ligandxhannel interactions can provide additional insight into direct assessment of these binding interactions. Equilibrium the structure and function of this important membrane-bound specific binding of 3H-BTX-B to sodium channels in vesicular protein (Albuquerque and Daly, 1976; Cahalan, 1980; Kho- preparations of mouse brain in the presence of scorpion toxin dorov, 1985; Narahashi, 1974, 1982; Rogart, 1981). was measured using a filtration assay procedure. At 25°C both At least three classes of specifically acting neurotoxins have TTX and STX inhibit 3H-BTX-B binding in a concentration- been used extensively in the study of voltage-sensitive sodium dependent and noncompetitive manner. This inhibition is mark- channels. These include the heterocyclic guanidines tetrodotox- edly temperature-dependent, being negligible at 37°C and max- in (TTX) and saxitoxin (STX), the “lipid-soluble” toxins batra- imal at WC, the lowest temperature investigated. Scatchard chotoxin (BTX), veratridine, aconitine and grayanotoxin, and analysis of BTX-B binding isotherms at 25°C in the presence the alpha-polypeptide neurotoxins from scorpion and sea anem- and absence of 1 PM TTX revealed that inhibition is due to a one. In the formalism of Catterall (1980), TTX and STX bind 3-fold decrease in the affinity of BTX-B binding with no change at type I sites at or near the mouth of the channel, causing in the number of binding sites sites (B,,). The concentration blockade of ion flux. The lipid-soluble toxins, of which BTX is dependence for TTX inhibition of both specific 3H-STX and 3H- the most potent, bind at site II, resulting in stabilization of an BTX-B binding is identical, suggesting that inhibition of 3H- open conformation of the channel. The third site, type III, binds BTX-B binding is due to a direct effect of TTXSTX binding at the polypeptide a-neurotoxins, which act by inhibition of so- their specific sodium channel site. The channel blockers did not dium channel inactivation. The binding of toxins at type II and alter the binding of scorpion toxin under these assay conditions, III sites has been shown to be positively cooperative, and an nor did BTX-B affect the binding of 3H-STX. These findings allosteric model describing this interaction has been presented support the conclusion that occupancy of the sodium channel (Catterall, 1977b). Type I and type II sites, however, have been TTX/STX binding site by the channel blockers can be accom- considered to be distinct and noninteracting based primarily on panied by a conformational perturbation of the BTX site re- the inability of BTX to affect specific binding of radiolabeled sulting in decreased BTX-B binding affinity. TTX or STX (Catterall and Morrow, 1978; Colquhoun et al., 1972). The complementary question, that of the effect of the The voltage-sensitive sodium channel is responsible for the rap- channel blockers on the binding of BTX, has not been directly id rising phase of the action potential in a variety of excitable investigated, since a suitable labeled BTX analog has not been membranes. In order to mediate this transient process, it is clear available and the actions of TTX (channel blocker) and BTX that the ionophore must be conformationally mobile, able to (channel activator) are antagonistic, thus rendering assessments assume different states. The conformational dynamics of the based on electrophysiological or flux measurements problemati- sodium channel are directly reflected in the voltage-dependent cal. gating process by which, in response to a depolarizing pulse, the Recently, Brown et al. (198 1) have synthesized a radiolabeled BTX derivative that retains full biological activity. The avail- ability of 3H-batrachotoxinin-A 20-cY-benzoate (3H-BTX-B) has Received Oct. 11, 1985; accepted Dec. 30, 1985. permitted a direct measurement of the effect of the channel This work was supported by National Institutes of Health Grant NS- 156 17. I blockers on toxin binding at site II. The experiments presented would like to thank Professor Frederick Benington for helpful advice and a careful here show that TTX and STX inhibit BTX-B binding to voltage- reading of the manuscript, Mr. Jack Johnston for tecnnical expertise, and Ms. Jackie Bowel1 for secretarial support in the preparation of the manuscript. sensitive sodium channels at temperatures less than 37°C and Correspondence should be addressed to Dr. Brown, Neurosciences Program, suggest that sites I and II are conformationally coupled. University of AJabama at Birmingham, University Station, Birmingham, AL 35294. A preliminary report of these findings has appeared (Brown Copyright 0 1986 Society for Neuroscience 0270-6474/86/072064-07$02.00/O and Johnston, 1983).

2064 The Journal of Neuroscience inhibition of Batrachotoxin Benzoate Binding by TTX 2065

Materials and Methods MICROSACS SYNAPTOSOMES Materials Leiurusquinquestriatus quinquestriatus scorpion venom and TTX were obtained in lyophilized form from Sigma Chemical Co. (St. Louis, MO). STX was kindly provided by Dr. John Daly, National Institutes of Health. Tritiated STX was prepared by tritium exchange and purified by high-voltage paper electrophoresis as described by Ritchie et al. ( 1976). The preparation used in these studies had a specific activity of 3 Ci/mmol and a radiochemical purity of 70%. Veratridine was purified from a mixture of veratrum alkaloids (Sigma Chemical Co.) essentially as described by Blount (1935). BTX-B and ‘H-BTX-B were prepared as previously reported (Brown et al., 1981). ‘H-BTX-B had a specific activity of 14 Ci/mmol and a radiochemical purity of 90%. [)H-BTX-B is currently available from New England Nuclear (Boston, MA) at a specific activity of 30-50 Ci/mmol.] All other reagents were of the highest grade commercially available. Vesicularpreparations of rodent brain Synaptosomal membrane vesicles were prepared from the brains of 6-month-old male Sprague-Dawley rats by a combination of differential and sucrose density-gradient centrifugation as described by Catterall et Figure 1. Inhibition of ‘H-BTX-B binding by TTX and STX. Micro- al. (1979). The material sedimentina at the 1.0/l .2 M interface was sacs or synaptosomes were equilibrated with 20 nM )H-BTX-B and 150 aspirated’and collected by centrifuga~on following dilution of the sus- &ml L. quinquestriatusscorpion venom at 25°C in the presence and pension in 0.32 M sucrose. The pellet was resuspended in standard absence (control) of 1 PM TTX or STX. Specifically bound label for incubation buffer (see below) at a final protein concentration of 3-4 mg/ each case, determined as the difference between total binding and bind- ml and incubated at 37°C for 20 min prior to use in the binding assays. ing in the presence of 250 PM veratridine, is plotted as the percentage In most of the experiments reported here, however, another vesicular of specific binding normalized to control binding, defined as 100%. The preparation described as “microsacs” or “synaptoneurosomes” was used data are the means + SD of 3 experiments run in triplicate. as a source of neuronal sodium channels. The basic biochemical and morphological properties of this preparation have been previously char- acterized (Creveling et al., 1980; Daly et al., 1980). The predominant was adjusted to approximately 1 nM, and unlabeled TTX or STX at a vesicular elements of the preparation appear to consist of resealed post- concentration of 1 PM was included in parallel assays for the determi- synaptic membrane with attached, resealed presynaptic endings. Pre- nation of nonspecific binding. Under these conditions, specific )H-STX vious studies have shown that the microsac preparation behaves sim- binding accounted for 95% of the total binding. ilarly to synaptosomal preparations with respect to sodium channel neurotoxin binding (e.g., Angelides and Brown, 1984). Microsacs were Data analysis prepared from Swiss Webster mouse brain cerebral cortical slices by a All data points were determined in triplicate and are presented as the modification (Creveling et al., 1983; Daly et al., 1980) of the original mean of these determinations. Unless otherwise indicated, the figures procedure of Chasin et al. (1974). Briefly, the freshly dissected tissue show the results of 1 experiment that is representative of 2 or more was homogenized in 2 volumes (wt/vol) of ice-cold standard incubation separate determinations. Curvilinear fits to the data have been approx- buffer using 10 strokes of a loose-fitting glass-glass homogenizer. The imated by eye, whereas linear fits were determined by regression anal- homogenate was diluted with 2 additional volumes of cold buffer and ysis. centrifuged at 1000 x g for 15 min at 5°C. The supematant was discarded, and the pellet was resuspended in standard incubation buffer Results by repetitive pipetting using a 9 in. Pasteur pipette. The protein concentration was adjusted to approximately 3 mg/ml. All protein deter- Inhibition of BTX-B binding by channelblockers minations were performed using the procedure of Peterson (1977) with BSA as a standard. An initial assessmentof the effect of the channel blockers TTX and STX on 3H-BTX-B binding to neuronal sodium channels Measurementof tritiated toxin binding wasmade by inclusion of thesetoxins in standardbinding assays Specific binding of ‘H-BTX-B was measured using minor modifications using mouse microsacs incubated at a temperature of 25°C. of the previously described procedure (Catterall et al., 198 1). Binding Figure 1 graphically shows that 1 PM TTX or STX reduced reactions were initiated by addition of 150 ~1 of vesicular preparation specific3H-BTX-B binding by 50-60% relative to control bind- containing 150-500 pg protein to a solution in standard incubation ing in the absenceof the channel blockers. This degreeof in- buffer of ‘H-BTX-B, 50 pg L. quinquestriatusquinquestriatus scorpion hibition remained unchanged when the concentration of the venom and various unlabeled effecters as indicated. The concentration blocker was increasedto 5 PM or when the incubation time was of 3H-BTX-B was generally 20-25 nM, and the total assay volume was extended from 60 to 120 min. Thus, under equilibrium con- 335 pl. Standard incubation buffer contained 130 mM choline chloride, 50 mM HEPES buffer adjusted to pH 7.4 with Tris base, 5.5 mM glucose, ditions, maximal inhibition of BTX-B binding by TTX or STX 0.8 mM MgSO,, 5.4 mM KCl, and 1 mg/ml BSA. Incubations were was substantial, but incomplete, suggestinga noncompetitive carried out for 60 min at the indicated temperature and were then mechanismfor the observed inhibition. The resultsof a similar terminated by addition of 3 ml ice-cold wash buffer. The tissue was experiment using rat brain synaptosomesconfirmed that the immediatelv collected on Whatman GF/C alass fiber filters and washed TTX- or STX-induced inhibition was not peculiar to the mi- 3 more times with 3 ml cold wash buffer.-The wash buffer contained crosacpreparation (Fig. 1). 163 mM choline chloride, 5 mM HEPES (pH 7.4), 1.8 mM CaCl,, 0.8 The effect of TTX and STX on BTX-B binding was readily mM MgSO,, and I mg/ml BSA. Radioactivity associated with the tissue reversible. In one series of experiments, microsacswere pre- was determined by liquid-scintillation spectroscopy of the filters sus- equilibrated with 1 PM TTX or STX in standard incubation pended in 10 ml scintillation cocktail (3a70B, RPI). Nonspecific binding buffer for 30 min at 25°C in the absence of 3H-BTX-B. Subse- was determined from parallel incubations containing 250 PM veratridine and has been subtracted from the data. Specific binding measured in quent addition of 3H-BTX-B and assayof specificbinding pro- this way was nominally 75% of the total binding. ducedthe expectedinhibition relative to control tubespre-equil- Binding of 3H-STX was measured in a manner analogous to that ibrated without TTX or STX. If, on the other hand, microsacs described for )H-BTX-B. In these assays, the concentration of ‘H-STX that had been pre-equilibrated with blocking agent were washed Vol. 6, No. 7, Jul. 1986

0 4 3 040 35 30 25 15 % TEMPERATURE Co:: Figure 3. Temperature dependence of TTX inhibition of ‘H-BTX-B binding. Specific binding of ‘H-BTX-B in mouse brain cortical microsacs in the presence (0) or absence (x) of 1 PM TTX is presented as a function of incubation temperature. All incubations were carried out for 60 min using the same tissue preparation and a constant concentration of )H-BTX-B. Temperature was controlled by immersion of the assay tubes in thermostatted water baths, and the tubes were capped to prevent loss due to evaporation.

of the plots, both in the absenceand presenceof TTX, is con- sonant with BTX-B binding to a single classof noninteracting sites. From the slopesof the lines, the apparent dissociation constants for BTX-B binding in the absenceand presenceof TTX were calculated to be 65 and 180 nM, respectively. As can be seenfrom the intercepts on the abscissa,TTX doesnot ap- 0 2 4 6 8 IO 12 14 I8 20 22 preciably affect the number of binding sites, so that inhibition BOUND (CPM 10-3;6 of 3H-BTX-B binding may be solely ascribedto the approximate 3-fold reduction in BTX-B binding affinity. Figure 2. Displacement of specific 3H-BTX-B binding in mouse brain Similar experiments were carried out usingveratridine as the cortical microsacs by BTX-B in the presence and absence of TTX. A, Microsacs were incubated with 22 nM )H-BTX-B and the indicated displacing ligand in place of unlabeled BTX-B. In this case,the concentrations of unlabeled BTX-B in the presence (Cl) or absence (0) apparent KD for veratridine binding was increased4-fold, from of 1 PM TTX for 1 hr at 25°C. Specifically bound label was measured 3.5 FM in the absenceof TTX to 14 I.IM in the presenceof TTX by filtration assay as described under Materials and Methods. The dis- (data not shown). Thus, the effect of TTX is not specific for nlacement curve for BTX-B binding is shifted to the right in the presence BTX-B binding, but extends as well to other ligandsthat bind of TTX. B, The data in A presented in the form of-a Scatchard plot. at type II siteson the sodium channel. There is little changein B,,,, but the KD for BTX-B binding, determined from the slopes of the lines, is increased from 65 nM in the absence of Temperaturedependence of TTX inhibition TTX to 180 nM in the presence of TTX. In a previous study, Creveling et al. (1983) reported that inclusion of TTX at a concentration of 1 PM in 3H-BTX-B binding by 3 repetitions of centrifugation and resuspensionin standard assaysto guineapig microsacsat 37°C in the presenceof scorpion incubation buffer prior to assayof 3H-BTX-B binding, 87% of toxin increasedspecific binding by approximately 7%. Since control binding was recovered. high-affinity 3H-BTX-B binding is dependent on the positively cooperative effect of scorpion toxin binding (Catterall et al., TTX decreasesBTX-B binding ajinity without 1981) and the binding of cY-scorpiontoxins is membrane-po- affecting the number of binding sites tential-dependent (Catterall, 1977a, b), addition of TTX was At the first level of analysis,the decreasein 3H-BTX-B binding presumed to increase BTX-B binding indirectly by stabilization in the presenceof the channel blockers may be due to either a of the membrane potential, and hence scorpion toxin binding, decreasein the affinity of BTX-B for its binding site, a loss of againstdepolarization mediated by any residual sodium ions the available number of binding sites, or both. In order to dis- presentin the vesicular preparation. The reported lack of TTX tinguish betweenthese possibilities the equilibrium binding iso- inhibition of 3H-BTX-B binding at 37°C prompted a more de- thermsfor 3H-BTX-B binding to mousebrain microsacsat 25°C tailed examination of the effect of temperatureon TTX/BTX-B were determined in the presenceand absenceof 1 /IM TTX by binding interactions. displacementassay. Specific binding for a constant concentra- Specific binding of )H-BTX-B to mousemicrosacs in the pres- tion of 3H-BTX-B was measuredas a function of increasing ence and absenceof 1 PM TTX was measuredas a function of concentrations of unlabeled BTX-B. The resulting inhibition temperature in the range 18-40°C. The data in Figure 3 indicate curves (Fig. 2A) reveal a substantial shift to the right for 3H- that, as previously reported, TTX does not inhibit ‘H-BTX-B BTX-B displacementin the presenceof 1 PM TTX. The apparent binding at 37°C. However, inhibition is already marked at 35°C K,,5 values are 85 and 200 nM in the absenceand presenceof and increasesapproximately linearly as the temperatureis low- TTX, respectively. For further analysis,bound and free amounts ered to 18°C.The slopecorresponds to 5% reduction in specific of BTX-B were calculated for each concentration point, and the 3H-BTX-B bindingPC. Specific binding in the absenceof TTX values usedto construct Scatchardplots (Fig. 2B). The linearity is, in contrast, temperature-insensitive in the range 25-40°C. The Journal of Neuroscience Inhibition of Batrachotoxin Benzoate Binding by TTX 2067

\. , ~OO,b~-- j ‘3 F --- 7% 10-7 2 5 10-6 CONCENTRATION OF INHIBITOR CM) CON~OENT~ATK~N OF ‘“i~~2R~~~T50~~~ CM) Figure 5. Comparisonof the concentrationdependence of TTX in- Figure 4. Dependenceof TTX and STX inhibition of specific3H- hibition of )H-STX and ‘H-BTX-B bindingto mousebrain cortical BTX-B bindingon concentration.Mouse brain cortical microsacs were microsacs.Microsacs from the sametissue preparation were incubated incubatedwith a constantconcentration of ‘H-BTX-B (20 nM) andthe with a constantconcentration of either‘H-STX (1 nM, 0) or ‘H-BTX-B indicatedconcentrations of eitherTTX (0) or STX (0) for 1 hr at 25°C. (20 nM, 0) andthe indicatedconcentrations of unlabeledTTX for 1 hr Specificallybound label was then assayedas described under Materials at 25°C.Specific binding was determined by filtration assayas described and Methods.The effectof TTX or STX at a concentrationof 1 PM underMaterials and Methods.The resultsare expressedas percentage wasdefined as the maximalinhibition, and the data are plottedas the of maximalinhibition, definedas the inhibition in the presenceof 1 PM percentageof this maximalinhibition vs concentrationof the inhibitor. unlabeledTTX. The resultingdisplacement curves are superposable, givinga K, 5value of approximately20 PM for TTX inhibition of both ‘H-STX and ‘H-BTX-B binding. The decreasein specificbinding of 3H-BTX-B between 25 and 18°C in the absenceof TTX is expected as a consequenceof decreasedscorpion toxin binding affinity at lower temperatures, then measuredas a function of the concentration of unlabeled with a concomitant reduction of the positively cooperative effect TTX. The data, normalized to the percentageof maximum on 3H-BTX-B binding (Brown et al., 1981; Catterall, 1977a, b). inhibition to facilitate comparison, are presentedas displacement curves in Figure 5. The 2 curves are essentiallysuper- Concentration dependence of TTX and STX inhibition posable,consistent with the conclusion that TTX inhibition of If inhibition of 3H-BTX-B binding by TTX and STX is mediated ‘H-BTX-B binding is a result of TTX binding to type I siteson by interaction of the channel blockers at type I sites on the the voltage-sensitivesodium channel. sodiumchannel, the concentration dependenceofthis inhibition should follow the appropriate Langmuir absorption isotherm TTX does not aflect scorpion toxin enhancement of ‘H-BTX-B for that binding. In order to examine this possibility, specific binding 3H-BTX-B binding to mousecerebral cortex microsacswas mea- In all of the experiments reported here, L. quinquestiatus quin- suredat 25°Cas a function of both STX and TTX concentration questriatus scorpion venom at a concentration of 150 &ml was in the range of 3-1000 nM. The data are presentedin Figure 4 included in the assaysto enhancethe binding affinity of ‘H- in the form of inhibition curves from which estimatescan be BTX-B. To assessthe possibility that the TTX-induced decrease made of the concentration displacing 50% of the total displace- in BTX-B binding affinity is secondary to a direct effect on able binding (IQ. The apparent IC,, values are 10 and 20 nM scorpion toxin binding, dose-responsecurves for scorpion toxin for STX and TTX, respectively. For comparison, the apparent enhancementof 3H-BTX-B binding in the presenceand absence equilibrium dissociation constants for STX and TTX binding of 1 PM TTX were compared. The resultsdemonstrate that the to rat brain sodium channelsare approximately 2 and 15 nM, concentrationdependence of scorpion toxin enhancementof ‘H- respectively (Catterall et al., 1979). Thus, the concentration de- BTX-B binding is not altered by TTX (Fig. 6). It is unlikely, pendenceof STX and TTX inhibition of 3H-BTX-B binding is therefore, that the mechanismof TTX inhibition involves mod- in the range appropriate for binding of these ligands at their ification of the cooperative interaction between BTX-B and known sodium channel site. scorpion toxin binding. This conclusionis further supportedby In theseexperiments it was helpful to use a concentration of the finding that, in a variety of tissues,binding of the purified tissue (1.5 mg protein/ml) such that an adequate specific to 1Z51-labeleda-scorpion toxin from L. quinquestriatus quinques- nonspecific binding ratio for ‘H-BTX-B could be maintained triatus to voltage-sensitivesodium channelsis unaffectedby the even when 60% of specific binding was blocked. Under these channelblockers (Catterall, 1977a;Ray and Catterall, 1978;Ray conditions the concentration of ‘H-BTX-B binding sitesin the et al., 1978).It shouldbe noted, however, that Siemanand Vogel assaywas found to be 3.6 nM. Although the site density for ‘H- (1983) recently reported on the ability of TTX to block effects STX binding was not determined for this tissue preparation, of purified ar-scorpiontoxins from Buthus tamulus on sodium previous studies(Catterall et al., 1981) using rat brain synap- channel inactivation in single nerve fibers of Xenopus laevis. tosomessuggest that there may be 2-3 times as many STX sites This blocking action was observed only if TTX was added si- as BTX-B sites. Thus, the concentration of STX sites in the multaneously with or prior to addition of scorpion toxin to the present experiments is at least as high as, and perhapsseveral incubation medium. Similar findings have been describedear- times higher than, the dissociation constant for STX binding. lier for the effect of TTX on the actions of scorpion toxin I from As discussedby Chang et al. (1975), this circumstanceleads to Androctonus australis Hector on crayfish neuromuscularjunc- artifactually high values for the apparent dissociation constant tion and rat brain synaptosomes(Romey et al., 1976). (in this case,the IC,,) when the data are treated as in Figure 4. In order to circumvent this problem, 2 displacementbinding Discussion assayswere carried out in parallel at 25°C usingthe sametissue Characterization of the binding to voltage-sensitive sodium preparation and identical conditions for both, except that in one channelsfor a growing number of specificallyacting neurotoxins assay‘H-STX (1 nM) replaced‘H-BTX-B (20 nM) as the labeled has provided a valuable perspective contributing to the devel- ligand. The specificbinding of both 3H-STX and 3H-BTX-B was opment of a molecularunderstanding of the structure and func- Brown Vol. 6, No. 7, Jul. 1986

Table 1. Modifiers of BTX-B binding

Compound (or class) Effect Reference a-Scorpion toxin, sea Increase affinity Catterall et al. (198 1) anemone toxin Local anesthetics Decrease affinity Creveling et al. (1983), Postma and Catterall (1984) Diphenylhydantoin and Decrease affinity Willow and Catterall @ ,k/ carbamazepine (1982) 0 ‘02 05 IO 2 5 IO 20 50 100 CONCENTRATION OF SCORPION VENOM (,ug/ml) (anticonvulsants) or-Cyan0 pyrethroid Increase affinity Brown and Olsen Figure6. Effect of TTX on the concentration dependence for scorpion insecticides (1984) toxin enhancement of ‘H-BTX-B binding. Mouse brain cortical microsacs were incubated with 3H-BTX-B (25 nM) and the indicated concen- TTX and STX Decrease affinity This study trations of L. quinquestriutusscorpion venom in the presence (0) or absence (0) of 1 FM TTX for 1 hr at 25°C. Scorpion venom was added from a concentrated stock solution prepared by gentle stirring of the under no condition was any effect of BTX-B on 3H-STX binding lyophilized venom in distilled water (2 mg/ml) at 5°C for 60 min, fol- discernible, even using concentrations of BTX-B as high as 5 lowed by high-speed centrifugation to remove insoluble material. Spe- PM (data not shown). On the other hand, both TTX and STX cific binding of 3H-BTX-B was measured by filtration assay as described inhibit 3H-BTX-B binding to neuronal sodium channelsin a under Materials and Methods. Since maximal binding of )H-BTX-B in concentration- and temperature-dependentmanner. Since the the presence of TTX is less than that in the absence of TTX, the data have been normalized to a percentage of the maximum enhancement concentration dependenceof TTX inhibition of 3H-STX bind- produced by scorpion venom in either the presence or absence of 1 PM ing parallelsexactly the concentration dependenceof inhibition TTX. In each case, maximal enhancement was defined as the difference of 3H-BTX-B binding (Fig. 5) it appearsthat inhibition is me- between specific ‘H-BTX-B binding in the absence of scorpion venom diated by interaction of the channel blockers with their normal (=O) and that in the presence of scorpion venom at a concentration of type I site on the sodium channel. 150 pg/ml. At saturating concentrations of STX or TTX at 25”C, the binding data indicate the presenceof a single classof BTX-B sitescharacterized by a 3-fold lower affinity for BTX-B binding tion of the sodium channel. The demonstration that 3H-BTX- (Fig. 2B). There is no evidence for a population of sodium B, an active congenerof batrachotoxin, interacts selectively at channels that are not so modified. This observation requires the batrachotoxin binding site, or site II, of the sodium channel some comment, since the vesicular preparations used in these (Brown et al., 1981; Catterall et al., 1981) provided a mechanism studiescould possibly contain a percentageof “inside out” ves- to make a direct assessmentof effectsof other sodium channel icles. Sodium channelsin thesevesicles could still bind the lipid- neurotoxins and ligandson toxin binding at site II. A noteworthy solubleBTX derivative but not TTX or STX, sincetype I sites observation from thesestudies is that a surprisingly large num- are externally located and the channel blockers, being highly ber of unrelated compounds, binding at distinct sites on the charged,cannot penetratethe membrane(Narahashi et al., 1966). voltage-sensitivesodium channel, are capableof modifying the Experimentally, however, BTX-B binding to such “inside out” binding of 3H-BTX-B. A list of these compounds (or classof vesicles would not be measuredwith the filtration assayused compounds)and their effect on 3H-BTX-B binding is presented in thesestudies. Sodium channelsin “inside out” vesicleswould in Table 1. One of theseinteractions, the positively cooperative also be inaccessibleto binding of a-scorpion toxin and, as a effect of cy-scorpiontoxins and anemonetoxin, has had the prac- consequence,would exhibit lo- to 15-fold lower affinity for tical impact of allowing convenient measurementof 3H-BTX-B BTX-B binding than in “right side out” vesicles, where the binding by virtue of the toxin-induced increasein affinity (Cat- positively cooperative interaction betweenBTX-B and scorpion terall et al., 1981). toxin binding is manifest. Further, the off-rate for BTX-B bind- In the face of suchextensive interactions betweenthe binding ing in the absenceof scorpiontoxin is quite rapid and specifically of 3H-BTX-B and other sodium channel ligands, it seemsrea- bound ligand is lost from the tissueduring filtration and washing sonableto suggestthat binding site II is in a domain that exhibits (Brown et al., 1981). These 2 factors combine to ensurethat considerableconformational flexibility and that is conforma- specific‘H-BTX-B binding measuredin thesestudies essentially tionally coupled, directly or allosterically, with other more re- reflects binding only to those sodium channels that are also mote regions of the channel protein(s). A similar impression accessibleto TTX and STX. might be reachedby consideringthe effectsof BTX on sodium As mentioned above, TTX and STX are positively charged channelfunction. Virtually all parametersconstituting an elec- at physiological pH by virtue of the highly basic guanidinium trophysiological description of the channel are altered in the moiety common to both structures. BTX and its derivatives presenceof BTX. Inactivation is blocked, the voltage depen- exist as mixtures of positively charged and neutral speciesat denceof activation is shifted to more negative potentials, the pH 7.4, representingtitration of the homomorpholino ring ni- conductanceof the channelis decreased,and the selectivity filter trogen. Recent evidence suggeststhat it is the charged species becomesless discriminating to positively chargedions (seeRho- of BTX-B that binds preferentially at the sodium channel site dorov, 1985, for review). (Angelidesand Brown, 1984; Brown and Daly, 1981). If binding Resultsof the experiments presentedhere provide the first of theseligands were to occur such that the positive chargesof evidence for an interaction between STX or TTX binding at BTX-B and the channelblocker were situated within the Debye site I and 3H-BTX-B binding at site II on the voltage-sensitive length, then electrostatic repulsion might be a possiblemech- sodium channel.These 2 siteshave been consideredto be non- anism for TTX/STX inhibition of BTX-B binding. However, interacting basedprimarily on the inability of BTX to affect the sinceBTX-B hasno effect on TTX/STX binding, this possibility binding of labeledSTX or TTX. Indeed, in the present studies, -nay be safely discarded. The Journal of Neuroscience Inhibition of Batrachotoxin Benzoate Binding by TTX 2069

Table 2. Thermodynamic parameters for BTX-B binding in the presence and absence of TTX at 25°C

&Z/m01 l/.&M TTX $!Zl/mol) de9 Absent -9.8 0 32.9 Present -9.2 9.95 64.3

151 gain in AS0 indicates an increasein the number of equivalent 0.0032 0.0033 0.0034 energy states available to products relative to reactants. The I/T magnitude of these changesinduced by inclusion of TTX is Figure 7. Van? Hoff analysisof the temperaturedependence for ‘H- striking. To place this in perspective, consider that, were it not BTX-B binding in the presence of 1 PM TTX. The logarithms of the for the compensatory increasein ASO,the unfavorable enthalpy association constants, KA, for specific )H-BTX-B binding to mouse brain changewould reduce BTX-B binding affinity at 25°C to zero. cortical microsacs in the presence of 1 PM TTX at representative tem- Consideringthe substantialchanges in both binding interactions peratures of 25, 30, and 37°C plotted against the inverse of the tem- and “randomness” characterizing BTX-B binding in the pres- perature in degrees Kelvin. The linearity of the plot suggests the absence ence of TTX, significant conformational differencesare likely of a membrane phase transition or significant differences in the heat to be involved. The most straightforward hypothesis is that capacities of reactants and products in this temperature range. The binding of the channel blockers at type I sites of the sodium standard enthalpy change for the binding reaction can therefore be calculated from the slope of the line (-m’YR), yielding AH0 = 9.95 kcal/ channel is accompanied by a conformational perturbation in mol. the type II binding domain, reflected in decreasedbinding affinity for the “lipid-soluble” neurotoxins. Whatever the mechanism, it is clear that type I and type II sitescan no longer be As shown in Figure 3, binding of 3H-BTX-B is highly tem- consideredto be noninteracting. perature-dependent in the presenceof TTX, whereas in the The interesting relationship between TTX/STX and BTX-B absenceof TTX the KD for ‘H-BTX-B binding is constant in binding disclosedby the experiments presented here under- the range 25 to 37°C. Analysis of the equilibrium thermody- scoresthe view that further information regarding the spatial namic parametersunderlying thesediffering behaviors can pro- relationship between type I and type II sites on the voltage- vide additional insight into the mechanism of TTX/BTX-B sensitive sodium channel will be very useful in developing a binding interactions. The standard free energy of associationis molecularunderstanding of structure-function relationshipsfor relatedto the equilibrium associationconstant by the expression this protein. Two relevant parametersare (1) the distance sep- AGo = -RT In KA (1) arating type I and type II sites, and (2) identification of the BTX-B binding protein(s). With regard to the first parameter, where R is the gasconstant and T is the temperaturein degrees Angelides and his colleagueshave applied techniques of fluo- Kelvin. Further, the standardfree energydefines the differences rescenceresonance energy transfer to assessthe distancebetween in standardenthalpy, AHo, and entropy, ASO,between reactants a number of specifically bound sodium channel neurotoxins and products by the relationship (Angelidesand Brown, 1984; Angelidesand Nutter, 1983, 1984; AGo = AH0 - TASO (2) Darbon and Angelides, 1984). There is currently no information regardingthe distance separatingSTX/TTX and BTX or BTX The standardenthalpy can be calculated from the temperature derivatives bound to the sodium channel. Concerningpoint (2), dependenceof the associationconstant expressedas a van? Hoff isolation and characterization of the TTX/STX and a-scorpion plot of In KA vs l/T. The slopeof the resultant line is -AHo/ toxin binding proteins of neuronal sodium channelshave been R. (For more complete discussionof theseconsiderations, see achieved (reviewed in Catterall, 1984), but no comparabledata Weiland and Molinoff, 1981.) To assessthe standard enthalpy concerningthe BTX binding protein(s)are available. Sincesolu- changefor BTX-B binding in the presenceof TTX, data from bilization of the sodium channel resultsin lossof BTX-B bind- Figure 3 have been used to construct the van’t Hoff plot in ing, progresson this front will require covalent labelingwith a Figure 7, yielding AH0 = 9.95 kcal/mol. Thus, in the presence suitable derivative in situ prior to solubilization. Current ex- of TTX, BTX-B binding is strongly endothermic. In the absence periments in this laboratory are directed toward the develop- of TTX, however, AH0 = 0, since BTX-B binding is tempera- ment and application of fluorescent and photoactivatable de- ture-insensitive in the temperature range under discussion.In rivatives of BTX that will be usefulin theseareas of investigation. this case,high-affinity binding is driven solely by a large positive entropy change. Now, knowing the association constants for References BTX-B binding at 25°C in the absenceand presenceof TTX Albuquerque, E. X., and J. W. Daly (1976) Batrachotoxin, a selective (15.2 x lo6 and 5.6 x 106 M-I, respectively, from Fig. 2) and probe for channels modulating sodium conductances in electrogenic the standard enthalpies calculated above, all relevant thermo- membranes. In The Specificity and Action of Animal, Bacterial and dynamic parameterscomparing BTX-B binding at 25°C in the Plant Toxins: Receptors and Recognition, Ser. B, Vol. 1, P. Cuatre- absenceand presenceof TTX can be determined from appro- casas, ed., pp. 297-338, Chapman and Hall, London. priate substitutionsin equations (1) and (2). These parameters Angelides, K. J., and G. B. Brown (1984) Fluorescence resonance are collected in Table 2. Brief inspection shows that, in the energy transfer on the voltage-dependent sodium channel. Spatial relationship and site coupling between the batrachotoxin and L.eiurus presenceof TTX, the positive enthalpy changeis largely coun- quinquestriatus quinquestriatus a-scorpion toxin receptors. J. Biol. teracted by an increasein the entropy. Relative to binding in Chem. 259: 6117-6126. the absenceof TTX, the entropy changeis approximately dou- Angelides, K. J., and T. J. Nutter (1983) Mapping the molecular struc- bled. The large positive AH0 suggeststhat in the presenceof ture of the voltage-dependent sodium channel. Distances between the TTX there is a net loss of favorable dipole-dipole, hydrogen tetrodotoxin and Leiurus quinquestriatus quinquestriatus scorpion toxin bond, or ionic interactions on binding of BTX-B, whereasthe receptors. J. Biol. Chem. 258: 1958-1967. 2070 Brown Vol. 6, No. 7, Jul. 1986

Angelides, K. J., and T. J. Nutter (1984) Molecular and cellular map- Creveling, C. R., E. T. McNeal, J. W. Daly, and G. B. Brown (1983) ping of the voltage-dependent Na+ channel. Biophys. J. 45: 3 l-34. Batrachotoxin-induced depolarization and [3H]batrachotoxinin-A 20- Armstrong, C. M. (198 1) Sodium channels and gating currents. Phys- a-benzoate binding in a vesicular preparation from guinea pig cerebral iol. Rev. 61: 644-683. cortex. Inhibition by local anesthetics. Mol. Pharmacol. 23: 350-358. Blount, B. K. (1935) The veratrine alkaloids, I and II. J. Chem. Sot. Daly, J. W., E. T. McNeal, C. Partington, M. Neuwirth, and C. R. 145:‘122-125. ’ Creveling (1980) Accumulations of cyclic AMP in adenine-labeled Brown, G. B., and J. W. Daly (198 1) Interaction ofbatrachotoxinin-A cell-free preparations from guinea pig cerebral cortex: Role of alpha- benzoate with voltaae-sensitive sodium channels. The effects of DH. adrenergic and H,-histamine@ receptors. J. Neurochem. 35: 326- Cell Mol. Neurobioi 1: 36 l-37 1. 337. Brown, G. B., and J. A. Johnston (1983) Tetrodotoxin and saxitoxin Darbon, H., and K. J. Angelides (1984) Structural mapping of the displacement of batrachotoxinin-A benzoate bound to voltage-sen- voltage-dependent sodium channel. Distance between the tetrodo- sitive sodium channels. Sot. Neurosci. Abstr. 9: 675. toxin and Centruroides sr@&4s s&iius II beta-scorpion toxin recep- Brown, G. B., and R. W. Olsen (1984) Batrachotoxinin-A benzoate tors. J. Biol. Chem. 259:.6074-6084. binding as an index of pyrethroid interaction at Na+ channels. Sot. Khodorov. B. I. (1985) Batrachotoxin as a tool to studv voltaae-sen- Neurosci. Abstr. 10: 865. sitive sodium channels of excitable membranes. Prog. Biophyi. Mol. Brown, G. B., S. C. Tieszen, J. W. Daly, J. E. Wamick, and E. X. Biol. 45: 57-148. Albuquerque (198 1) Batrachotoxinin-A 20-oc-benzoate: A new ra- Narahashi, T. (1974) Chemicals as tools in the study of excitable dioactive ligand for voltage-sensitive sodium channels. Cell. Mol. membranes. Physiol. Rev. 55: 8 13-889. Neurobiol. 1: 19-40. Narahashi, T. (1982) Pharmacology of ionic channels in excitable Cahalan, M. (1980) Molecular properties of sodium channels in ex- membranes. Adv. Pharmacol. Ther. 3: 3-l 8. citable membranes. In The Cell Surface and Neuronal Function, C. Narahashi, T., N. C. Anderson, and J. W. Moore (1966) Tetrodotoxin W. Cotman, G. Poste, and G. L. Nicholson, eds., pp. l-47, Elsevier/ does not block excitation from inside the nerve membrane. Science North Holland, Amsterdam. 153: 765-767. Catterall, W. A. (1977a) Membrane potential-dependent binding of Peterson, G. L. (1977) A simplification of the protein assay method scorpion toxin to the action potential Na+ ionophore. J. Biol. Chem. of Lowry et al. which is more generally applicable. Anal. Biochem. 2.52; 8660-8668. 83: 346-356. Catterall, W. A. (1977b) Activation of the action potential Na ion- Postma, S. W., and W. A. Catterall (1984) Inhibition of binding of ophore by neurotoxins. An allosteric model. J. Biol. Chem. 252: [3H]batrachotoxinin-A 20-alpha-benzoate to sodium channels by lo- 8669-8676. cal anesthetics. Mol. Pharmacol. 25: 2 19-227. Catterall, W. A. (1980) Neurotoxins that act on voltage-sensitive so- Ray, R., and W. A. Catterall (1978) Membrane potential-dependent dium channels in excitable membranes. Annu. Rev. Pharmacol. Tox- binding of scorpion toxin to the action potential ionophore. Studies icol. 20: 15-43. with a 3(4-hydroxy 3-[‘ZsI]iodophenyl) propionyl derivative. J. Neu- Catterall, W. A. (1984) The molecular basis of neuronal excitability. rochem. 31: 397-407. Science 223: 653-66 1. Ray, R., C. S. Morrow, and W. A. Catterall (1978) Binding ofscorpion Catterall, W. A., and C. S. Morrow (1978) Binding of saxitoxin to toxin to receptor sites associated with voltage-sensitive sodium chan- electricallv excitable neuroblastoma cells. Proc. Natl. Acad. Sci. USA nels in synaptic nerve ending particles. J. Biol. Chem. 253: 7307- 74: 21 l-215. 7313. Catterall, W. A., C. S. Morrow, and R. P. Hartshome (1979) Neu- Ritchie, J. M., R. B. Rogart, and G. R. Strichartz (1976) A new method rotoxin binding to receptor sites associated with voltage-sensitive for labelling saxitoxin and its binding to non-myelinated nerve fibers sodium channels in intact, lysed, and detergent solubilized brain of the rabbit vagus, lobster walking leg, and gartish olfactory nerves. membranes. J. Biol. Chem. 2j4: 11379-l 1387. J. Physiol. (Lond.) 261: 477-494. Catterall. W. A.. C. S. Morrow. J. W. Dalv. and G. B. Brown (198 1) Rogart, R. (198 1) Sodium channels in nerve and muscle membrane. Binding of batrachotoxinin-A 20-cu-benzoate to a receptor site asso: Annu. Rev. Physiol. 43: 71 l-725. ciated with sodium channels in synaptic nerve-ending particles. J. Romey, G., J. P. Abita, R. Chicheportiche, H. Rochat, and M. Laz- Biol. Chem. 256: 8922-8927. dun&i (1976) Scorpion neurotoxin. Mode of action on neuromus- Chang, K.-J., S. Jacobs, and P. Cuatrecasas (1975) Quantitative aspects cular junctions and synaptosomes. Biochim. Biophys. Acta 448: 607- of hormone-receptor interactions of high affinity. Effect of receptor 619. concentration and measurement of dissociation constants of labeled Sieman, D., and W. Vogel (1983) Tetrodotoxin interferes with the and unlabeled hormone. Biochim. Biophys. Acta 406: 294-303. reaction of scorpion toxin (&thus tamulus) at the sodium channel Chasin, M., F. Mamrak, and S. G. Saminiego (1974) Preparation and of excitable membrane. Pfluegers Arch. 397: 306-3 11. properties of a cell-free, hormonally responsive adenylate cyclase from Weiland, G. A., and P. B. Molinoff (198 1) Quantitative analysis of guinea pig brain. J. Neurochem. 22: 103 l-1038. drug-receptor interactions: I. Determination of kinetic and equilib- Colquhoun, D., R. Henderson, and J. M. Ritchie (1972) The binding rium properties. Life Sci. 29: 3 13-330. of labelled tetrodotoxin to non-myelinated nerve fibers. J. Physiol. Willow, M., and W. A. Catterall (1982) Inhibition of binding of (Lond.) 227: 95-126. [3H]batrachotoxinin-A 20-alpha-benzoate to sodium channels by the Creveling, C. R., E. T. McNeal, D. H. McCulloh, and J. W. Daly (1980) anticonvulsant drugs diphenylhydantoin and carbamazepine. Mol. Membrane potentials in cell-free preparations from guinea pig cere- Pharmacol. 22: 627-635. bral cortex: Effect of depolarizing agents and cyclic nucleotides. J. Neurochem. 35: 922-932.