Tetrodotoxin-insensitive Sodium Channels BINDING OF POLYPEPTIDE NEUROTOXINS IN PRIMARY CULTURES OF RAT MUSCLE CELLS*

(Received for publication, December 8,1980, and in revised form, February 24, 1981)

John C. Lawrence$ and William A. Catterall From the Department of Pharmacology, University of Washington, Seattle, Washington 98195

"he binding of 1261-labeled derivatives of scorpion toxin and sea anemone toxin to tetrodotoxin-insensi- tive sodium channels in cultured rat muscle cells has been studied. Specific binding of '261-labeled scorpion toxin and '2SI-IabeIed sea anemone toxin was each blocked by either native scorpion toxin or native sea anemone toxin. for block of binding by several polypeptide toxins was closely correlated with Ko.e for enhancement of sodium channel activation in rat mus- cle cells. These results directly demonstrate binding of sea anemone toxin and scorpion toxin to a common receptor site on the sodium channel. Binding of both '2aI-Jabeled toxin derivatives is enhanced by the alka- loids aconitine and batrachotoxin due to a decrease in KO for polypeptide toxin. Enhancement of polypeptide toxin binding by aconitine and batrachotoxin is pre- cisely correlated with persistent activation of sodium channels by the alkaloid toxins consistent with the conclusion that there is allosteric coupling between receptor sites for alkaloid and polypeptide toxins on the sodium channel. The binding of both 126Elabeled scorpion toxin and '2SI-labeled sea anemone toxin is reduced by depolarization due to a voltage-dependent increase in KD. Scorpion toxin binding is more voltage- sensitive than sea anemone toxin binding. Our results directly demonstrate voltage-dependent binding of both scorpion toxin and sea anemone toxin to a com- mon receptor site on the sodium channel and introduce the lZ6I-labeled polypeptide toxin derivatives as specific binding probes of tetrodotoxin-insensitive sodium channels in cultured muscle cells.

Voltage sensitive Na' channels can be divided into two classes by the concentrations of tetrodotoxin and saxitoxin required to inhibit ion flux. The channels found in cultured rat skeletal muscle and denervated mammalian skeletal mus- cle are about 100-fold less sensitive to inhibition by tetrodo- toxin than those found in nerve and innervated muscle (1-5). In the preceding manuscript, the interactions of different neurotoxins with the tetrodotoxin-insensitive Na'channels of cultured skeletal muscle cells were studied using measure- ments of Na+ flux (6). The results presented indicate that these channels, like tetrodotoxin-sensitive channels, possess

' This work was supported by a research grant from the Muscular Dystrophy Association to W. A. C. and by a research fellowship from the Muscular Dystrophy Association to J. C. L. The costs of publi- cation of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertise- rnent" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

$ Present address, Department of Pharmacology, Washington Uni- versity, St. Louis, MO.

three separate receptor sites for neurotoxins. Qualitatively, the interactions of neurotoxins with tetrodotoxin-sensitive and -insensitive Na' channels are quite similar, suggesting that the two classes of Na+ channels share a large degree of structural homology (6).

Studies of neurotoxin action using measurements of ion flux have proven to be a valuable first step in elucidating the similarities and differences between tetrodotoxin-sensitive and -insensitive Na' channels. Nevertheless, changes in ion flux induced by neurotoxins are only an indirect index of neuro- toxin binding and additional information can be obtained by direct measurements of the binding of radiolabeled ligands to sites associated with voltage-sensitive Na' channels. Of the three classes of neurotoxins that act on sodium channels, only the polypeptide toxins are active on tetrodotoxin-insensitive sodium channels at concentrations lower than IO-' M (6). These toxins are, therefore, the best candidates for develop- ment, of ligand binding assays for tetrodotoxin-insensitive sodium channels.

Binding of scorpion toxin to tetrodotoxin-sensitive sodium channels has been studied in neuroblastoma cells (7-9), syn- aptosomes (10, 11), adult skeletal muscle (129, and cultured heart cells (13). Specific binding is inhibited by depolarization and enhanced by alkaloid toxins (7, 8, 10, 12, 13). Binding is also blocked by sea anemone toxin I1 leading to t,he conclusion that these two polypeptides share a common receptor site (9, 14). Binding of sea allemone toxin has been studied less frequently because it has relatively low affinity for tetrodo- toxin-sensitive sodium channels in nerve. Binding of anemone toxin I from Anemonia sulcata to sodium channels in crayfish nerve is reduced by depolarization and enhanced by veratri- dine (15). Binding of A. sulcatu toxin I1 to rat diaphragm was shown to be blocked by both unlabeled sea anemone toxin and scorpion toxin (21). In contrast, binding of A. sulcata toxin SI to rat brain synaptosomes is unaffected by alkaloid toxins, depolarization, and scorpion toxin (16), raising some question concerning the conclusion that scorpion toxin and sea anemone toxin bind to a common receptor site in a voltage- dependent manner (9, 14). In this report, we describe and compare the binding of '251-labeled derivatives of both scor- pion toxin and sea anemone toxin to tetrodotoxin-insensitive sodium channels in cultured rat muscle cells.

EXPERIMENTAL PROCEDURES

MateriaZs-All supplies were obtained from the sowces indicated in the preceding report (6) .

Primary Culture of Rat Skeletal Muscle-Skeletal muscle (1-2 g) from the forelimbs of 20-day embryonic rats was minced and incu- bated at 36 OC in 20 ml of Ca2+- and Mg2+-free phosphate-buffered saline containing 0.25% trypsin. After 1 h, 20 ml of growth medium (10% heat-inactivated horse serum, 5% fetal bovine serum, 50 units/

6223

6224 Sodium Channels in Rat Muscle Cells

ml of penicillin G, and 10 p g / d of streptomycin in DME medium') were added and the suspension was centrifuged at 200 x g for 10 min. The pelleted cells were suspended in growth medium and 0.5 ml containing 250,000 cells was added to collagen-coated wells of plastic multiwell tissue culture dishes (Costar, 1.6-cm diameter wells). Cells were incubated at 36 "C in a humidified atmosphere of 10% Con and 90% air. After 4 days, the medium was replaced by growth medium (1 d) containing 0.2 p c i / d of C3H]1eucine and 10 p~ D-arabinofur- anosykytosine to prevent fibroblast overgrowth. This medium was replaced after 4 days with growth medium containing 0.2 pCi/ml of [3H]leucine. Cells were used in experiments 10-14 days after plating. Virtually all of the muscle cells were present as multinucleated myotubes, and by visual observation, contaminating fibroblasts ac- counted for less than 10% of the cells present in the culture.

Measurements of 22Na+ Uptake-Incubations with neurotoxins in sodium-free medium and assays of "Na+ uptake in the presence of 10 m~ Na' were performed exactly as described in the preceding report (6).

Preparation of '2sI-La6eZed Derivatives of Scorpion Toxin and Sea Anemone Toxin IZ-Scorpion m~no['~~I]iodotoxin was prepared by lactoperoxidase-catalyzed iodination and purified by ion exchange chromatography as described previously (8).

Sea anemone toxin I1 was labeled with In5I in a similar reaction. Toxin I1 has no t-yrosine but has been labeled on histidine residues using chloramine-T as a catalyst with retention of biological activity (16, 2 1 ) . Lactoperoxidase will catalyze iodination of histidine in addi- tion to tyrosine although the KIM for histidine iodination is relatively high (17). We have used lactoperoxidase to prepare iodinated deriv- atives of sea anemone toxin.

Sea anemone toxin I1 (20 nmol) was lyophilized in a polypropylene tube. NaIz5I (5 mCi) was added to the dry toxin in 30 pl of 0.1 M potassium phosphate, pH 7.4, and 0.67 mM NaI. The reaction was then initiated by addition of 2 pl of 10 mM Hz02 and 2 pl of 0.1 mg/ml of lactoperoxidase (Sigma) in 5 mM potassium phosphate, pH 5.5. These additions were repeated three times at 15-min intervals. The reaction was stopped by addition of 5 pl of 250 mM NaNa followed by 180 pl of 10 mM ammonium acetate, pH 6, and 1 mg/ml of bovine serum albumin. The reaction mixture was adsorbed to a column (20 X 0.6 cm) of sulfopropyl (SP)-Sephadex C-25 equilibrated with 10 mM ammonium acetate, pH 6, and 1 mg/ml of bovine serum albumin. Unreacted Iz5I was removed by washing the column with 10 ml of 10 mM ammonium acetate, pH 6, and 1 mg/ml of bovine serum albumin. The labeled toxin derivatives were then eluted with a linear gradient from 100-400 mM ammonium acetate, pH 6, and 1 mg/ml of bovine serum albumin.

A typical column profile is dustrated in Fig. 1. Two peaks of biological activity are observed centered at fractions 58 and 78. Under these conditions, native sea anemone toxin I1 migrates in the peak centered at fraction 78. The iodinated derivative migrating in the peak centered at fraction 58 was used in the experiments described in this report. The number of moles of lZ5I per mol of toxin has not been accurately determined for this labeled derivative. If it is assumed that the derivative is monosubstituted, its biological activity per mol of '251-toxin is 50% greater than native toxin. Since the toxin derivative is used only at tracer concentrations to monitor occupancy of receptor sites by native sea anemone toxin I1 and competitors, the level of substitution of the derivative does not affect the conclusions reached. The results are reported as femtomoles of Iz5I bound/mg of cell protein or as per cent of control binding. '2sI-labeled sea anemone toxin I1 ('"I-AsTX 11) concentrations are presented in terms of moles of toxin-bound '*'I per liter.

Binding of '2511-ScTX and '2SI-AsTX-Cells were incubated at 36 OC with '2511-S~T~ or Iz5I-AsTX I1 for 1 h in a medium normally containing 156 mM Tris (adjusted at 36 "C to pH 7.4 with HCl), 5.4 mM KCl, 0.8 mM MgS04, 1.8 mM CaC12, 50 mM Hepes (adjusted at 36 "C to pH 7.4 with Tris), and 1 mg/ml of bovine serum albumin. The ionic composition of this medium was varied in some experiments by substitution of the Tris with osmotically equivalent amounts of NaCl or KC1. The binding reactions were terminated by washing the cells with medium at 0 "C in the absence of neurotoxins. Cells that were incubated with Iz5I-AsTX I1 were washed four times in 15 s with 3 ml of medium containing 156 mM Tris (pH adjusted to 7.4 at 0 "C with HCl), 50 mM Hepes (pH adjusted to 7.4 at 0 "C with Tris), 0.8

' The abbreviations used are: DME medium, Dulbecco's modified Eagle's medium; '2sIII-ScTX, scorpion m~no['~~I]iodotoxin; Iz5I-AsTX 11, '251-labeled Anemonca sdccata toxin 11; Hepes, 4-(2-hydroxyethyl)- 1-piperazineethanesulfonic acid.

- N N f + c 400 - -16 5 X d

300 - - -12 3

F - 8 b s

5 200- 5

A loo - - 4 -

' 0 20 40

U 3 o o w 40

Fraction Nurnbr FIG. 1. Separation of 'ZKI-labeled sea anemone toxin I1 deriv-

atives by ion exchange chromatography. Sea anemone toxin I1 was labeled with '"I by lactoperoxidase-catalyzed iodination as de- scribed under "Experimental Procedures" and the quenched reaction mixture was adsorbed to a column (0.6 X 20 cm) of SP-Sephadex C- 25 equilibrated with 10 mM ammonium acetate, pH 6, and 1 mg/ml of bovine serum albumin. Unreacted I2'I was removed by washing the column with 10 mI of 10 mM ammonium acetate, pH 6, and 1 mg/ml of bovine serum albumin. The labeled toxin derivatives were then eluted with a 50-ml linear gradient from 100-400 mM ammonium acetate, pH 6, and 1 mg/ml of bovine serum albumin. Fractions of 0.5 ml were collected. I2'I counts per minute were determined on an aliquot of each fraction (0). Biological activity of each fraction was tested in ion flux experiments with cultured muscle cells as described in the accompanying report (6). The enhancement of veratridine- stimulated "Na+ uptake is plotted after subtraction of 22Naf uptake in the presence of veratridine alone (A).

mM MgS04, and 1.8 mM CaCL Cells that were incubated with scorpion toxin were also washed four times in 15 s. The medium for the f i s t two washes contained: 156 mM Tris (adjusted at 0 "C to pH 7.4 with HCl), 5.4 mM KC1, 0.8 mM MgSO,, 1.8 mM CaC12, 50 mM Hepes (adjusted at 0 "C to pH 7.4 with Tris), and 1 mg/ml of bovine serum albumin. The final two washes were performed with medium of the same composition as was used to terminate the sea anemone toxin-binding reaction. Cells were suspended in 0.4 N NaOH, Iz5I was measured using a y counter, and 'H was determined using a liquid scintillation counter after correct,ing for spillover of IZsI. The amount of protein in representative cultures was determined and the amount of radioactivity from [3H]leucine present in each culture was used to correct for varying amounts of protein recovered (6). Specific binding is calculated as total binding minus nonsaturable binding observed in the presence of a saturating amount of unlabeled toxin (2 p ~ ) . Specific binding comprised 75% of total binding of "'I-AsTX 11. Because the KD for 'Z611,-ScTX binding was quite variable (see "Results"), specific binding ranged from 20-75% of total binding.

RESULTS

The results presented in the preceding report suggest that scorpion toxin and sea anemone toxin I1 interact with a common polypeptide neurotoxin binding site associated with tetrodotoxin-insensitive Naf channels of L5 cells (6), In pre- liminary experiments, we were unable to measure specific binding of ['251]-labeled derivatives of scorpion toxin or sea anemone toxin I1 using these cells. Therefore, the possibility of using primary cultures of rat skeletal muscle as a source of cells for binding studies was explored. Cells grown in primary culture were found to possess tetrodotoxin-insensitive Na' channels that are activated by the same neurotoxins that increase "Na+ uptake in L5 cells.2 Moreover, these cells were

J. C. Lawrence and W. A. Catterall, unpublished observations.

Sodium Channels in Rat Muscle Cells 6225

found to have two distinct advantages over the L5 cells with respect to their use in measurements of polypeptide toxin binding. First, it was possible to obtain a greater density of myotubes in primary culture, thus increasing the number of toxin binding sites per culture. Second, the apparent affinities of scorpion toxin and sea anemone toxin I1 were 3-fold greater than in L5 cells. This characteristic is demonstrated by the experiment shown in Fig. 2A. In this experiment, cells were incubated with increasing concentrations of scorpion toxin and sea anemone toxin I1 for 10 min in Na-free medium. Uptake of "Na+ was then assayed in medium containing 200 p~ veratridine without the polypeptide neurotoxins. The con- centrations of sea anemone toxin I1 and scorpion toxin re- quired to produce half-maximal effects on enhancing veratri- dine action were 5 X IO-' M and 2 X lo-' M, respectively. The higher affinity of scorpion toxin and sea anemone toxin for

10-8 107 10

POLYPEPTIDE TOXIN (MI

POLYPEPTIDE TOXIN (MI

5.0

F 4.0 1

v 3.0 - - X I- 8 2.0

a 9 Y 1.0

0

POLYPEPTIDE TOXIN (M)

their receptor sites in primary rat muscle cells should result in a higher ratio of specific to nonspecific binding.

While these values of K 0 . 5 for ion flux stimulation are useful for comparison with the Koa for scorpion toxin and sea ane- mone toxin I1 inhibition of ['251]-labeled polypeptide toxin binding (Fig. 2, b and c ) , they should not be taken as a measure of the true dissociation constants. For example, as shown in Fig. 2a, the effect of scorpion toxin decreases at high concentrations. Thus, the scorpion toxin concentration curve presented is probably the result of the interaction of an effect producing activation with an effect which produces inhibition. We are uncertain of the cause of the inhibitory effect but its presence causes an underestimate of the true KO in ion flux experiments. A second problem with using measurements of ion flux to estimate affinities of scorpion toxin and sea ane- mone toxin I1 is the requirement for an alkaloid toxin in the ion flux experiment. Polypeptide neurotoxins enhance alkaloid toxin activation of sodium channels but do not activate chan- nels themselves (14, 18, 19). Thus, polypeptide toxin action cannot be assessed directly in ion flux experiments.

Using primary cultures of rat skeletal muscle cells, it was possible to measure saturable binding of both '2511-S~TX and '251-AsTX 11. In the experiment presented in Fig. 2b, cells were incubated for 1 h with 0.5 nM '251~-S~TX and increasing concentrations of either unlabeled scorpion toxin or sea ane- mone toxin 11. '2511-S~TX was displaced by scorpion toxin with a KO.& of approximately 2.7 X lo-' M, and, in this experiment, essentially the same K0.5 was obtained with sea anemone toxin 11. Since tracer concentrations of labeled toxins were used (1 X 10"O M to 5 X 10"O M), these values of K0.5 provide good approximations of KD for toxin binding. The displacement curve obtained with sea anemone toxin I1 was flatter than that observed with scorpion toxin. The concentration-effect curves of sea anemone toxin I1 with respect to enhancing veratridine stimulation of "Na+ uptake are similarly not as steep as those obtained with scorpion toxin (Fig. 2a, Ref. 6). From the data of Fig. 2b, apparent Hill coefficients of 0.87 and 0.55 are derived for scorpion toxin and sea anemone toxin 11, respectively. The fact that the KO values for scorpion toxin and sea anemone toxin I1 binding are comparable to the K0.5

v 2 u T t o x i n action on "Na+ uptake (Fig. 2a) is evidence that the bikg-meSiirerebrepresents binding to a site asso- ciated with the Na+ channel.

Both unlabeled scorpion toxin and sea anemone toxin I1 decreased the specific binding of 1 X 10"O M Iz5I-AsTX I1 (Fig. 2c). In this experiment, the KD for sea anemone toxin I1 was approximately 8 X lo-' M and that for scorpion toxin was 3 X W 7 M. The Hill coefficients for block of '251-AsTX I1 binding are near 1.0 for all toxins since the low concentrations of labeled toxin used will occupy only high affinity receptor sites. Increasing concentrations of sea anemone toxins I and I11 also decreased the amount of '251-AsT~ I1 bound. Much

FIG. 2. Effect of increasing concentration of scorpion toxin and sea anemone toxin I1 on "Na+uptake and the saturable binding of 'Z611-S~TX and '2SI-AsTX 11. a, Cultured skeletal muscle cells were incubated at 36 "C with medium containing 5.4 mM KC1 and increasing concentrations of scorpion toxin (0) or sea anemone toxin I1 (A). After 10 min, the medium was aspirated and "Na' uptake was measured after 5 s in the absence of polypeptide toxins in medium containing 200 PM veratridine. b, Cells were incubated at 36 "C for 1 h with 5 X 10"' M '2511-S~TX in medium containing 5.4 mM KC1 and increasing concentrations of scorpion toxin (0) or sea anemone toxin I1 (A). c, Cultured rat skeletal muscle cells were incubated at 36 "C for 1 h with 1 X 10"' M Iz5I-AsTX I1 in medium containing 5.4 mM KC1 and increasing concentrations of scorpion toxin (A), sea anemone toxin I (O), sea anemone toxin I1 (O), and sea anemone toxin I11 (V). The amounts of Iz5I-labeled toxins specifically bound were determined as described under "Experimental Proce- dures."

6226 Sodium Channels in Rat Muscle Cells

higher concentrations of toxins I and I11 (compared to sea anemone toxin 11) were required to inhibit binding. Likewise, high concentrations of these two toxins are required to en- hance veratridine-stimulated "Na+ uptake.'

By comparing the results presented in Fig. 2, b and c, it can be seen that 10-fold higher concentrations of scorpion toxin were required to inhibit binding of '=I-labeled sea anemone toxin 11 than were required to inhibit binding of ''51-labeled scorpion toxin. This large apparent discrepancy is due to the choice of experiments for presentation. In the scorpion toxin binding experiments presented in this report, the KO for unlabeled scorpion toxin ranges from 2.7 X lo-@ M (Fig. 26) to 1.4 X lo" M (Fig. 46). We have found that when the KO for scorpion toxin is greater than 2 X M, the amount of lZ5I- labeled scorpion toxin specifically bound is less than 20% of the total amount of toxin bound. Under these conditions, it is very difficult to accurately measure specific binding. Never- theless, in several experiments the KO for scorpion toxin has been estimated to be between 2 X 10" M and 5 X M.' These estimates are within the range of scorpion toxin con- centrations that produce half-maximal inhibition of sea ane- mone toxin binding (Fig. 2c). For presentation, we have se- lected experiments in which the KD for scorpion toxin inhibi- tion of 12511-S~TX binding is in the lower part of the range since the ratio of specific to nonspecific binding of '251-labeled scorpion toxin is greater in those experiments. We believe the variation of KD values is due to variation of average membrane potential of the primary muscle cells from experiment to experiment since scorpion toxin binding is steeply voltage dependent (see below).

Both scorpion toxin and sea anemone toxin I1 increase the apparent affinities of alkaloid neurotoxins in cultured skeletal muscle cells (6). Therefore, the effects of batrachotoxin and aconitine on the binding of "51-labeled scorpion toxin and sea anemone toxin I1 were investigated. In the experiments shown in Fig. 3, cells were incubated with increasing concentrations of batrachotoxin and aconitine. The effects of the two alkaloid toxins on increasing "Na+ uptake (Fig. 3a) essentially paral- leled their actions on increasing the binding of 5 X 10"' IZ5I1- ScTX (Fig. 3b) and 1 X 10"' M '"I-AsTX (Fig. 3c). The reduction in "Na+ uptake and toxin binding as the concentra- tion of aconitine approaches M (A) was variable in differ- ent experiments. We believe that this represents a nonspecific membrane disruptive effect of this lipid soluble toxin at high concentration. In these experiments, cells were also incubated with 1 PM batrachotoxin plus increasing concentrations of aconitine. Results from previous experiments in a variety of systems indicate that batrachotoxin and aconitine interact with a common receptor site and that batrachotoxin is a full agonist whereas aconitine is a partial agonist (19, 20). There- fore, additivity of the actions of the two alkaloid neurotoxins is not predicted. Rather, as increasing concentrations of acon- itine are added, batrachotoxin should be displaced from the receptor site and the net effect diminished. The results pre-

FIG. 3. Stimulation of "Na+ uptake and binding of '2611-ScTX and '251-AsTX I1 by batrachotoxin and aconitine. a, Cultured muscle cells were incubated at 36 "C in medium containing 135 mM KC1 and increasing concentrations of batrachotoxin (8) or aconitine (A). Cells were also incubated with 1 p~ batrachotoxin plus increasing concentrations of aconitine (a). After 1 h, the medium was aspirated and **Na+ uptake was measured after 15 s in medium containing 5.4 mM KC1 and the same concentrations of neurotoxins. b and c, Muscle cells were incubated at 36 "C with 5 X lo-'" M "'II-SCTX (b ) or 1 X IO"' M "'I-AsTX I1 (c ) in medium containing 5.4 mM KC1 and 20 PM tetrodotoxin. Increasing concentrations of batrachotoxin (0) or acon-

aconitine (a) were added to this medium. The results presented are itine (A) and 1 p~ batrachotoxin plus increasing concentrations of

the amounts of '251-labeled neurotoxins specifically bound after 1 h.

sented demonstrate that increasing concentrations of aconi- tine promote parallel decreases in the effects of batrachotoxin on increasing "Na+ uptake (Fig. 3a) and on increasing the binding of both lZ5I~-ScTX and lZ5I-ATX 11 (Fig. 3, b and c) . These results strongly support the conclusion that batracho- toxin and aconitine interact with a common receptor site that is coupled allosterically to the polypeptide toxin binding site(s). In addition, these results provide further strong evi- dence that the binding of lZ5I-labeled scorpion toxin and sea anemone toxin measured in our experiments represents bind- ing to receptor sites associated with sodium channels.

The effect of batrachotoxin on enhancing the binding of the polypeptide neurotoxins was investigated further. In the ex-

80 a

Batrachotoxin of Aconitine [MI

- rn

-. Batrachotoxin or Aconitine IM1

Batrachotoxin or Aconitine IMI

periment shown in Fig. 4a, cells were incubated with 5 X 10"' M '*'I1-S~TX and increasing concentrations of unlabeled scor- pion toxin in the absence and presence of 1 PM batrachotoxin. In the absence of unlabeled scorpion toxin, batrachotoxin increased binding of '''I-labeled scorpion toxin 2.2-fold (from 2.5 fmol/mg to 5.4 fmol/mg). The KD for scorpion toxin inhibition of labeled toxin binding was decreased 2.3-fold by batrachotoxin (from 1.4 X to 0.6 X M). Thus, the effect of batrachotoxin on increasing scorpion toxin binding can be explained by an action of the alkaloid neurotoxin to increase the affinity of scorpion toxin for its receptor. Likewise, batrachotoxin increases the affinity of sea anemone toxin I1 for its receptor. As shown in Fig. 4b, batrachotoxin increased the binding of 5 X lo-" M "'I-AsTX I1 from 0.8 fmol/mg to 2.0 fmol/mg while decreasing the KD for sea anemone toxin I1 from 9 X M to 3 X w 9 .

Results were presented in the preceding manuscript which indicated that the effects of scorpion toxin and sea anemone toxin I1 on *'Na uptake were dependent upon membrane

SCORPION TOXIN (MI

i

FIG. 4. Batrachotoxin decreases KO for scorpion toxin and sea anemone toxin 11. Cultured skeletal muscle cells were incubated in the absence (0) or presence (A) of 1 p~ batrachotoxin in medium containing 5.4 mM KC1 and 20 p~ tetrodotoxin. a, The results are presented as percentages of inhibition of the binding of 5 X 10"" M IZ5I1-ScTX produced by increasing concentrations of unlabeled scor- pion toxin after a 1-h incubation. The amount of 'Z511-S~TX bound in the absence of unlabeled toxin was 2.5 fmol/mg in the absence of batrachotoxin and 5.4 fmol/mg in the presence of the alkaloid. b, The results are presented as percentages of inhibition of the binding of 5 X lo-" M 'ZsII-A~TX I1 produced by increasing concentrations of unlabeled sea anemone toxin I1 after a I-h incubation. Without unlabeled toxin, 0.8 fmol/mg of '2sII-AsTX I1 was bound in the absence of batrachotoxin and 2.0 fmol/mg were bound in the presence of the alkaloid neurotoxin.

SCORPION TOXIN (M)

,I >

l -4,; 0 ' A 1

0 lolo lo9 IOE d lo* SEA ANEMONE TOXIN II (MI

FIG. 5. Effect of depolarization with KC1 on the binding of scorpion toxin and sea anemone toxin 11. Cultured skeletal mus- cle cells were incubated at 36 "C in media containing 5 mM KC1 (O), 25 mM KC1 (A), or 135 mM KC1 (0). a, The results are presented as percentages of inhibition of the binding of 5 X 10"" M 1z511-S~TX produced by increasing concentrations of unlabeled scorpion toxin after a 1-h incubation. In the absence of unlabeled scorpion toxin, the amount of '2511-S~TX bound in medium containing 5 mM KC1 was 0.5 fmol/mg and that bound in medium containing 25 mM KC1 was 0.3 fmol/mg. b, The inhibition of binding of '251-AsTX I1 by increasing concentrations of sea anemone toxin I1 was determined after a 1-h incubation. In the absence of unlabeled toxin, the amount of sea anemone toxin I1 bound in the different media were as follows: 5 mM KCl, 0.6 fmol/mg; 25 mM KCl, 0.4 fmol/mg; and 135 mM KCI, 1.3 fmol/mg.

potential. Values for K0.S for the two toxins progressively increased when cells were incubated with increasing concen- trations of K' (6). To directly investigate the possibility that depolarization decreases the affinities of polypeptide neuro- toxins for their receptor site, binding of 5 X 10"" M "511-ScTX (Fig. 5a) or 5 X lo-'' M '"I-AsTX (Fig. 5b) was measured in media containing different concentrations of K' and increas- ing concentrations of the unlabeled polypeptide neurotoxins. Increasing the concentration of K'in the medium from 5 to 25 mM decreased the amount of '25111-S~TX bound in the absence of unlabeled toxin from 0.5 fmol/mg to 0.3 fmol/mg. This decrease was accompanied by an increase in the K D of scorpion toxin from 7 X M to 2.4 X 10" M. In other experiments, cells have been incubated in medium containing 135 mM KC1 in which the membrane potential of cultured skeletal muscle cells is essentially 0 mV (6). No specific binding of scorpion toxin was detected under these conditions. We believe the most likely reason to be that the affinity is decreased to such an extent when cells are completely depolarized that it is

6228 Sodium Channels in Rut Muscle Cells

SEA ANEMONE TOXIN II (M)

FIG. 6. Gramicidin A in the presence of Na' increases the KD for sea anemone toxin 11. Cultured skeletal muscle cells were incubated at 36 "C with 1 X lo-" M ''"I-AsTX I1 in the absence (0) and presence (A) of 10 pg/ml of gramicidin A. The medium contained 130 mM NaCl and increasing concentrations of unlabeled sea anemone toxin 11. The results are presented as percentages of the inhibition of binding of labeled toxin by sea anemone toxin I1 after 1 h. In the absence of unlabeled sea anemone toxin 11, the amount of '*'I-AsTX 11 bound was 2.0 fmol/mg without gramicidin A and 1.2 fmol/mg in the presence of gramicidin A.

technically impossible to measure specific binding of the toxin (See Fig. 12 of Ref. 6). The affinity of sea anemone toxin I1 for its receptor was progressively decreased as the concentration of K+ in the medium was increased (Fig. 5b). In medium containing 5 mM K', the Kn for sea anemone toxin I1 was about 8 X lo-' M. Increasing the concentration of K' in the medium to 25 mM or 135 mM resulted in an increase in the KO of sea anemone toxin I1 to 1.9 X IO-' M or 2.8 X M, respectively.

The results obtained after incubating cells in media con- taining different concentrations of K' are consistent with the interpretation that binding of the polypeptide neurotoxins is dependent upon membrane potential. However, these findings do not exclude the possibility that changes in the ionic com- position of the medium actually caused the observed changes in toxin affinity. We have, therefore, measured binding under conditions that promote depolarization without changing the concentration of K' in the medium.

In the presence of Na+, the cation selective ionophore, gramicidin, depolarizes electrically excitable cells. In the ex- periment shown in Fig. 6, cells were incubated in the absence and presence of 10 pg/ml of gramicidin A in medium contain- ing 130 mM NaCI. Under these conditions, gramicidin A decreased the binding of 5 X lo-" M '251-AsTX I1 from 2.0 fmol/mg to 1.2 fmol/mg and increased the KD for sea anemone toxin I1 from 1.2 X lo-* M to 2.8 X M. Gramicidin A (10 pg/ml) was without effect on '2sI-A~TX I1 binding when cells were incubated in Na+-free medium.2

The results obtained with gramicidin A support the hypoth- esis that sea anemone toxin binding is sensitive to changes in membrane potential. Additional support for this hypothesis was derived from studies of the effect of veratridine and Na' on sea anemone toxin binding. Veratridine promotes persist- a t activation of voltage-sensitive Na' channels in electrically excitable cells and, thus, in the presence of Na+ causes depo- larization. In medium containing 130 mM NaC1, veratridine (200 p ~ ) decreased the binding of 5 X 10"' M 12SI-A~TX I1 from 2.2 fmol/mg to 0.6 fmol/mg. This decrease is due to depolarization since inhibition of Na' influx with tetrodotoxin (50 p ~ ) abolished the inhibitory effect of veratridine. In the presence of tetrodotoxin, veratridine actually increased the binding of '251-labeled sea anemone toxin to 2.5 fmol/mg. This is consistent with the enhancement of binding produced by

the other alkaloid neurotoxins (batrachotoxin and aconitine) when incubations were performed in Na+-free medium (Figs. 3 and 4 ) .

DISCUSSION

Our results demonstrate specific binding of both 1Z51,-S~TX and 1251-AsTX I1 to receptor sites associated with tetrodo- toxin-insensitive sodium channels in cultured rat muscle cells and show that these two toxins bind in a voltage-dependent manner to a common receptor site. Specificity of binding of both toxin probes is supported by three experiments. (i) The KO for competitive displacement of binding of the labeled toxins by both unlabeled toxins correlates with values of for enhancement of 22Na+ influx by the toxins (Fig. 2). (ii) The enhancement of binding of both L251-labeled toxin probes by alkaloid toxins is exactly correlated with the activation of sodium channels by alkaloid toxins (Fig. 3) as expected from the allosteric interactions between receptor sites for these two classes of toxins. (iii) The voltage dependence of binding of the 12sI-labeled toxins (Fig. 5) closely parallels the voltage dependence of toxin action described in the preceding report (6). These results leave little doubt that the binding measured represents interaction with a specific receptor site associated with the sodium channel.

The experiments described show directly that scorpion toxin and sea anemone toxin bind to a common class of receptor sites. The binding of both toxins is enhanced by alkaloid toxins and reduced by depolarization. Specific scor- pion toxin binding is completely prevented by sea anemone toxin and specific sea anemone toxin binding is completely blocked by scorpion toxin. These data, together with the common mechanism of action of the two toxins (6,14), provide strong evidence that they act at a common site by a similar mechanism. Their binding and action are not identical, how- ever. Concentration-effect curves and binding curves for scor- pion toxin are always hyperbolic, whereas similar curves for sea anemone toxin are shallower than hyperbolic, indicating a heterogeneous binding affinity or a complex mechanism of binding and action. The data of Fig. 2b show that scorpion toxin binds to the polypeptide toxin receptor site with homo- geneous affinity, whereas sea anemone toxin distinguishes multiple classes of polypeptide toxin receptor sites or induces negatively cooperative interactions not observed with scorpion toxin. A detailed analysis of kinetic and equilibrium properties of sea anemone toxin I1 binding will be the subject of another report.

Voltage dependence of binding is observed with both scor- pion toxin and sea anemone toxin I1 as previously inferred from ion flux studies (6). Depolarization of cells with Na' plus gramicidin A or Na' plus veratridine provides strong support for the conclusion that binding is dependent on membrane potential and not on K+ per se. Quantitative differences are observed in the voltage dependence of the KD for scorpion toxin and sea anemone toxin I1 measured in competitive displacement experiments. With scorpion toxin, the K D is increased 3.5-fold by depolarization with 25 mM K+. Depolar- ization with 135 mM K+ reduces specific scorpion toxin binding to undetectably low levels, suggesting a substantially larger increase in Kn. In contrast, with sea anemone toxin 11, the Ku is increased 3-fold by 25 mM KC but 135 mM K' has o d y a slightly greater effect. These results with direct binding mea- surements correlate closely with ion flux data in which the voltage dependence of toxin action was measured (6).

Previous studies of scorpion toxin binding to high affinity receptor sites on tetrodotoxin-sensitive sodium channels in neuroblastoma cells (7-9), synaptosomes (10, 111, heart cells (13), and adult skeletal muscle (12) have shown that specific

Sodium Channels in Rat Muscle Cells 6229

binding is enhanced by alkaloid toxins and reduced by sea anemone toxin I1 and by membrane depolarization. These results led to the conclusion that scorpion toxin and sea anemone toxin 11 bind at a common receptor site in a voltage- dependent manner (9, 14). In contrast, direct measurements of sea anemone toxin I1 binding to low affinity receptor sites in synaptosomes did not detect effects of alkaloid toxins, scorpion toxin, or membrane depolarization on binding (16). These results reopened the question of the site and voltage dependence of sea anemone toxin I1 binding. Our data show that when specific binding of sea anemone toxin I1 to high affinity receptor sites on tetrodotoxin-insensitive sodium channels is measured, the expected allosteric enhancement of binding by alkaloid toxins, competitive displacement by scor- pion toxin, and block of binding by depolarization are ob- served. Since each of these characteristics of sea anemone toxin I1 binding and action is observed in ion flux studies with synaptosomes (20), it is surprising that they are not observed in direct binding studies.

A major objective of these studies was to develop a toxin binding method applicab1e.t.o tetrodotoxin-insensitive sodium channels and to compare toxin binding at the polypeptide toxin receptor site of tetrodotoxin-sensitive and -insensitive sodium channels. As in our ion flux studies (6), we have found great similarity in toxin action on these two classes of sodium channels. While the toxin binding specificity of the polypep- tide toxin receptor site is altered such that sea anemone toxin I1 binds with higher affinity, the allosteric interactions and voltage dependence of binding are nearly identical with pre- vious work on tetrodotoxin-sensitive sodiumn channels. These results provide further support for our working hypothesis that tetrodotoxin-sensitive and -insensitive sodium channels have substantial structural and functional homology.

Acknowledgment-We thank Dr. Laszlo Beress for providing sam- ples of Anemonia sulcata toxins I, 11, and 111.

REFERENCES

1. Harris, J. B., and Thesleff, S. (1971) Acta Physiol. Scand. 81,

2. Kidokoro Y., Heinemann, S., Schubert, D., Brandt, B. L., and Klier, F. G. (1975) Cold Spring Harbor Symp. Quant. Biol. 40,

3. Sastre, A., and Podleski, T. R. (1976) Proc. Natl. Acad. Sci. U. s. 4. Catterall, W. A. (1976) Biochem. Biophys Res. Commun. 68, 136-

5. Stallcup, W. B., and Cohn, M. (1976) Exp. Cell Res. 98,227-284 6. Lawrence, J. C., and Catterall, W. A. (1981) J. Biol. Chem. 256,

7. Catterall, W. A., Ray, R., and Morrow, C. S. (1976) Proc. Natl.

8. Catterall, W. A. (1977) J. Biol. Chem. 252,8660-8668 9. Couraud, F., Rochat, H., and Lissitzky, S. (1978) Biochem. Bio-

10. Rav, R., Morrow, C. S., and Catterall, W. A. (1978) J. Biol. Chem.

383-388

373-388

A. 73,1355-1359

142

6213-6222

Acad. Sei. U. S. A. 73,2682-2686

phys. Res. Commun. 83, 1525-1530

253, 7307-7313 11. Jouver. E.. Martins-Montat. N.. Couraud. F.. and Rochat, H.

(1978) Biochem. Biophys. Res.'Commun.'85,'377-382 12. Catterall, W. A. (1979) J. Gen. Physiol. 74,375-392 13. Couraud, F., Rochat, H., and Lissitzky, S. (1980) Biochemistry 19,

457-462 14. Catterall, W. A., and Beress, L. (1978) J. Biol. Chem. 253, 7393-

7396 15. Stengelin, S., and Hucho, F. (1980) Hoppe-Seyler's Z. Physiol.

Chem. 361,577-585 16. Vincent, J . P., Balerna, M., Barhanin, J., Fosset, M., and Lazdun-

ski, M. (1980) Proc. Natl. Acad. Sci. U. S. A. 77, 1646- 1650

17. Morrison, M., and Bayse, G. (1973) in Oxidases and Related Redox Systems (King, T., Mason, H., and Morrison, M., eds) Vol. 1, pp. 375-388, University Park Press, Baltimore

18. Catterall, W. A. (1976) J. Biol. Chem. 251, 5528-5536 19. Catterall, W. A. (1977) J. Biol. Chem. 252, 8669-8676 20. Tamkun, M., and C a t t e d , W. A. (1981) Mol. Pharmacol. 19,

21. Habermann, E., and Beress, L. (1979) Naunyn-Schmiedeberg's 78-86

Arch. Pharmacol. 309, 165-170