Proc. NatL Acad. Sci. USAVol. 78, No. 2, pp. 1245-1249, February 1981Neurobiology

Na'-channel-associated scorpion toxin receptor sites as probes forneuronal evolution in vivo and in vitro

(scorpion toxin binding affinity/central nervous system neurons and glia/cell cultures/differentiation/tetanus toxin binding cells)

YOHEVED BERWALD-NETTER*, NICOLE MARTIN-MOUTOTt, ANNETTE KOULAKOFF*, ANDFRANQOIS COURAUDt*Biochimie Cellulaire, College de France, 75231 Paris Cedex 05, France; and tBiochimie, Institut National de la Sante et de la Recherche M.dicale U-172,Facult6 de M6decine Nord, 13326 Marseille Cedex 3, France

Communicated by Roger Y. Stanier, October 9, 1980

ABSTRACT Purified neurotoxin II of the scorpion Androc-tonus australis Hector (ScTx) has previously been shown to bindspecifically to the Na+-ionophore-associated, voltage-sensitive,receptor sites of excitable cells. We have conducted binding stud-ies, using high-specific-activity '5I-labeled ScTx, to detect andquantify the Na+-channel receptors on cells of the developing fetalmouse brain. In vivo, the onset of detectable specific binding isat 12 fetal days. The rate of receptor appearance is initially slowbut increases sharply as of the 16th a o- mouse ontogenesis. Themean number of receptors at 12 and 19 days is 120 and 20,000 percell, respectively (i.e., 0.5 and 80 per square micrometer). Whencorrected for the fraction of cell population corresponding to pu-tative neuroblasts and neurons, identified by immunofluores-cence as tetanus toxin binding cells, these values are, respectively,1040 and 33,900 ScTx receptors per tetanus toxin binding cell or4.2 and 136 per square micrometer. At all stages, the toxin bindsto a single class of noninteracting sites; Kd = 0.1-0.5 nM. Similarfindings in terms of ScTx-receptor properties and quantitativeevolution were obtained in vitro. Specific '5I-labeled ScTx bind-ing to brain cells grown in culture is consistently correlated withthe presence of tetanus toxin binding cells. In cultures of centralnervous system glia without neurons, only nonspecific low-levelScTx binding was detected. These results suggest that the high-affinity scorpion toxin receptors may be used as quantitative mark-ers of neuronal differentiation.

The major common feature of neurons is their capacity to gen-erate and propagate action potentials in response to chemicalor electrical stimulation of their membrane. This response de-pends on the presence of specific ionic permeabilities as partof the signal transduction mechanisms in biological membranes.The voltage-sensitive sodium channel, a quasi-ubiquitous con-stituent of excitable cells, has been shown by electrophysiolog-ical techniques and by measurement of ionic fluxes (1, 2) topossess at least three functionally distinct components that bearbinding sites for neurotoxins (3); these sites can be monitoredby the use of labeled neurotoxins that bind with high affinityand modify the channel function in a specific way (4). The de-velopmental timetable of appearance and the evolution pattern,during mammalian central nervous system ontogenesis, ofthese neurotoxin receptors and of their associated functions areas yet unknown. Several electrophysiological studies of fetalcells in culture (5-12), although performed at relatively latestages, have shown that cultured cells often develop in vitro asthey do in vivo. Electrophysiological analysis of cells at earlydevelopmental stages is, however, hampered by their noto-riously small size, great fragility, and ambiguous morphologicaldifferentiation (13). The probing of developing ion channels byspecific binding and flux studies seems thus an approach ofchoice.

Neurotoxins purified from the venoms of African scorpionsare small basic proteins (14) that modify the action potential ofnerve (4, 15). They have been shown to bind specifically to ex-citable cultured neuroblastoma cells (16, 17), to rat brain syn-aptosomes (18, 19) and to cultured fetal chicken cardiac muscle(20) in a membrane-potential-dependent fashion (4). Toxinbinding is correlated with modification of the action potentialresponse (21, 22) and with ionic flux changes (16, 20, 23), allof which can be explained by the blocking or considerable slow-down of the Na'-channel inactivation mechanism. Thus, studyof the appearance and evolution of scorpion toxin (ScTx) re-ceptor sites during central nervous system ontogenesis can pro-vide information on the evolution of the more complex structurethat is the Na' channel.

In the work reported here, the quantitative evolution of ScTxreceptor sites was explored by binding assays, on cells of fetalmouse brains, as a function of age in vivo or time in vitro.

EXPERIMENTAL PROCEDURESMaterials. The ScTx used was toxin II from the venom of

Androctonus australis Hector purified according to Miranda etaL (24). 125I-AaHII ("2I-labeled ScTx) was prepared by the lac-toperoxidase method as described (25), with a specific activityof 700-1000 Ci/mmol (1 Ci = 3.7 x 10'° becquerels). Bovineserum albumin (fraction V), Hepes, polyornithine, polylysine,and lactoperoxidase were from Sigma. Cytosine arabinoside wasfrom Calbiochem; the chemicals for buffer solutions were fromMerck and carrier-free '25I (13-17 Ci/mg) was from Amersham.Cell culture medium and serum were from GIBCO, and cultureplates were from Falcon or Corning. Mice were random-bredSwiss, supplied by Iffa-Credo, Fresnes, France; they weremated overnight, and the postmating day was counted as fetalday 1.

Cell Suspensions. Pregnant mice were sacrificed by cervicaldislocation, the embryos were removed aseptically and rapidlydecapitated, and the brains were dissected out into Hepes (10mM) buffered Eagle's minimal essential medium. Cell disso-ciation was done without the use of enzymes or chelatingagents. The meninges were carefully removed, and the wholebrains were rinsed 2-3 times with Ca2`-, Mg2+-free phos-phate-buffered saline containing glucose (5 mg/ml) and incu-bated in this solution at 4°C for 60-90 min. At the end of in-cubation, the buffer was replaced by one containing 1% fetalcalf serum or bovine serum albumin at 1 mg/ml. The dissocia-tion was by gentle mechanical disruption of cell-cell contactsby passing the tissue several times through flame-polished pas-teur pipettes of decreasing bore. The dissociated cells wereseparated from residual aggregates by sedimentation and de-

Abbreviations: ScTx, scorpion toxin; TBC, tetanus toxin binding cells.

1245

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertise-nent" in accordance with 18 U. S. C. §1734 solely to indicate this fact.

1246 Neurobiology: Berwald-Netter et al

cantation. This procedure was repeated until essentially all ofthe tissue had been disaggregated (i.e., three-four times). Allwas done at 0-40C, and cells were collected by centrifugation.Cell counts were done by using a conventional Neubauer he-mocytometer, and cell viability was assessed by trypan blueexclusion. This procedure yielded >90% single cells havingviable counts of 83 5%, 70 ± 10%, and 65 ± 10% for cellsfrom 11- to 14-day, 15- or 16-day, and 17- to 19-day fetal brains,respectively. Cell diameters were measured by using a gradedocular eyepiece standardized according to a micrometricreference.Cultures of Dissociated Brain Cells. Cells were cultured in

Eagle's medium (alpha modification; GIBCO) supplementedwith glucose at 4 mg/ml, penicillin at 50 units/ml, gentamycinat 12.5 ,Ag/ml, and 7.5% fetal calf serum. Culture plates usedwere coated with rat tail collagen and polylysine or polyorni-thine (2 tug/ml of H20 for 1 hr). Dissociated cells were sus-

pended in culture medium and seeded at 2 x 106 (viable cellsper milliliter) per 35-mm culture plate. The next day, 1 ml ofculture medium was added, and this was replaced by 1 ml offresh medium every 3-4 days thereafter. For cultures of neu-

rons, the cells were derived from brains of 15-day mouse em-

bryos. Overgrowth of glial cells was suppressed by addition ofcytosine arabinoside at 1 ,ug/ml 48 hr after plating.

Cultures of glia without neurons were obtained by dissocia-tion of brains of newborn mice, excluding the late-developingdeep nuclei and cerebellum (26, 27). For dissociation, the tissuewas subjected to mild treatment by trypsin/collagenase, fol-lowed by pipette trituration as above. The culture medium wasthe same as above. Cells were seeded at 0.5 x 106 viable cellsper35-mm culture plate without any additional treatment. Con-fluent cell monolayers were obtained within 1-2 weeks of plat-ing. In addition to astrocytes, the cultures contained numerousciliated ependymal cells localized in well-circumscribed areasand a few oligodendrocytes. The lack of significant numbers ofneurons was assessed by phase-contrast microscopy and by im-munofluorescence assays showing a virtual absence of tetanustoxin-binding cells (TBC) (28, 29).

"25I-Labeled ScTx Binding Assays. Cells were rinsed in a

Na'-free binding incubation buffer containing 140 mM cholinechloride, 5.4mM KCl, 1.8mM CaCl2, 0.8mM MgSO4, 5.0mMglucose, 25 mM Hepes (pH adjusted to 7.3 with Tris base), andbovine serum albumin at 1 mg/ml. '"I-Labeled and native ScTxdilutions were made in the same buffer. Binding to dissociatedbrain cells in suspension was done in plastic Rhesus tubes infinal volumes of 100 and 500 ,ul, respectively, for 1-2 x 106viable cells of 11- to 17-day and 1 X 106 cells of 18- or 19-dayembryos, to ensure equilibrium binding. After incubation for30 min at 37°C, the cells were rinsed three times with coldbuffer and pelleted by centrifugation, and the bound "I was

measured without further treatment in a Kontron y-scintillationcounter. For binding assays on cultured cells, the rinsed cul-tures were covered with 1 ml of prewarmed, properly dilutedtoxin, sealed, incubated for 30 min in a 37°C water bath, thenrinsed three times in cold buffer, and dissolved in 0.1 M NaOH,and the bound radioactivity was measured. All assays were donein duplicate, and samples of starting solutions and of unboundresidual radioactivity were counted for each dilution.

RESULTS

Evolution in Vivo. As the binding of ScTx to Na+-channel-associated receptor sites is voltage dependent, the standard as-say for it requires viable cell preparations that have intactplasma cell membranes. Thus, a nonenzymatic cell dissociationmethod was developed that provides a reproducibly high yield

10 1. . 0.5 1.0 1. 20

oD 1.0 0'" 0.4-

"0.2

0.5 1.0 1.5 0.5 1.0 1.5 2.0

C D

c~3 4.0-8.0-

"0~~~~~~~~~~~~~~~~~~~~~.F) 1. Bn 0io 6c0dCe l3. 0

1-40.5 0.2-

1- 2.0- 4.0-

'0 9 87 2 46 8

ing increasing concentrations of 126I-labeled ScTx in the absence (a,total binding) or the presence (0, nonspecific binding) of excess native

ScTx (0.1 M). Results of representative experiments are given for

cells derived from 12- (A), 13- (B), 1S(C), and 17- (D) day embryos.

(E) Dependence of specific binding on cell integrity. Cells of 15-day

embryo brains were killed by three cycles of freezing and thawing inliquid N2, resulting in superposable curves of total and nonspecifically

bound radioactivity. (F) Displacement ofo251-labeled ScTx (0.2 nM)

binding to cells of 16-day embryos by increasing concentrations of un-labeled toxin.fom 12I-Labeled ScTx bound without native toxin; B.binding in presence of the indicated concentration of native toxin.

ScTx in units of -log M. (G) Scatchard plot of binding data given inD.

of isolated, viable, fetal brain cells. Binding assays were per-

formed on cell suspensions derived from 11- to 19-day fetalmouse brains.

No specific binding was detectable at <12 days. The results

of representative experiments performed on dissociated brain

cells of 12-, 16-, and 17-day fetal mice are shown in Fig. 1. The

saturable hyperbola shapes of thebudI-labeled ScTx bindingcurves, as well as the derived linear Scatchard plots, indicatethat the toxin binds to a single class of noninteracting sites. Themean numbers of receptor sites per cell increased with fetal age(Fig. 2). This increase was initially slow; between the 13th and

the 15th fetal day, the mean number of receptors per cell in-creased from 500 to 800. Subsequently, the rate of receptor

appearance increased sharply, reaching its maximum at 17-18

days of gestation. At 19 days, the mean number of ScTx recep-

tors per brain cell is wf20,000.It seems a reasonable assumption, albeit not directly proven,

that the cells in the central nervous system that express ScTxreceptors are neuroblasts and neurons, the percentages ofwhich are developmentally regulated. Thus, a more meaningful

estimation of ScTx-receptor evolution could be obtained by tak-

ing into account the neuronal cell compartment at various stages

of ontogenesis. Tetanus toxin has been shown to label cultured

neurons of peripheral and central nervous tissue selectively(28-32). We have explored the evolution of TBC in the centralnervous system of the mouse and shown that these cells are

detectable as of the 10th gestation day; i.e., they precede by

A

Proc. Nad Acad. Sci. USA 78 (1981)

Proc. Natd Acad. Sci. USA 78 (1981) 1247

60 r

50 F

40 F

e-goE- 30 F

20

10

0

30 *0

25 4,

20 v

&4

15 x

10 9:;a

5 Mc,3

11 13 15 17 19

Age (ofembryo), days

FIG. 2. Increase in abundance of ScTx binding sites on mouse fetalbrain cells during in vivo evolution. A, Number of ScTx binding sitesper cell was determined from Scatchard plots done after binding assaysas in Fig. 1G; e, age-dependent evolution of ScTx receptors calculatedper TBC; o, frequency of TBC as determined by immunofluorescence.

2 days the appearance of specific ScTx binding (29). The evolu-tion curve of TBC is given in Fig. 2, together with data on theage-dependent ScTx binding either calculated per cell or cor-

rected per TBC. This correction, as shown also in Table 1, no-

tably reduces the ratio between the minimum and the maximumvalues of ScTx receptor densities obtained. Nevertheless, thegeneral pattern of receptor evolution is preserved.To estimate the receptor density per surface area unit, cell-

diameter measurements were done on 100 viable cells each of13- and 19-day dissociated fetal brain preparations. At eitherage, the smallest cells had a diameter of 6.5 jim and the largestones had diameters of 12.0 Itm. The mean cell diameters were

practically identical-i.e., respectively, 9.3 1.1 ,um and 9.3± 1.4 ,um. Thus, if one takes 9 ,um as the average cell diameterand assumes a homogeneous receptor distribution at the surfacemembrane, the mean receptor density per cell increases from0.5 site per square micrometer at 12 days to 80 sites per squaremicrometer at 19 days (Table 1); when calculated per TBC,these values are 4.2 and 136, respectively.

The affinity of ScTx for its receptor was found to vary littleat the different ontogenetic stages. The dissociation constant,K,, was 0.12-0.46, with a tendency for higher mean values(lower affinities) in cells of older embryos (see Table 1).

All of these binding data are given for, and indeed reflect,toxin interaction with receptors on viable cells only. The bindingaffinity of ScTx to its receptors, being voltage dependent, is toolow in damaged or dead cells that are depolarized. On suchcells, no specific "2I-labeled ScTx fixation could be seen, thecurve of total binding being nonsaturable and perfectly super-

posable with the curve of nonspecific binding (see Fig. 1E).Furthermore, cell membrane depolarization, either by an in-crease in extracellular [K+] or by addition of veratridine intoNa+-containing incubation medium, resulted in a concentra-tion-dependent inhibition of specific ScTx binding (data notshown). Thus the properties of ScTx receptors in the developingbrain do not differ from those found on electrically excitableneurons and on synaptosomal preparations described before.

Evolution in Vitro. 125I-Labeled ScTx binds specifically tofetal mouse brain cells grown in culture, and the number ofbinding sites increases significantly with time in vitro (Fig. 3).The estimated mean number of receptor sites per 9-day cul-tured cell is close to that of 19-day fetal brain TBC; i.e., 35,000+ 10,000. The Scatchard plot of specific binding (Fig. 3E)shows the involvement of a single class of receptor sites witha dissociation constant of 0.5 nM.

Experiments on binding displacement (competition) of theiodinated toxin by increasing concentrations of native ScTxshow that the Kd is =0.2 nM (Fig. 3D); i.e., very close to theKd for "2I-labeled ScTx in cultured cells (Fig. 3A) as well as tothat determined for freshly dissociated fetal brain cells (see Ta-ble 1).

The progressive increase in the "2I-labeled ScTx bindingcapacity as a function of time in culture could be attributedeither to an evolution of receptors having a higher affinity or

to a quantitative increase in the number of receptors per cul-ture. The dissociation constants do not vary significantly duringthe course of the culture period studied. This fact gives cre-

dence to the notion that the observed increase in I'I-labeledtoxin fixation with time in vitro (Fig. 3C) is the consequenceof an increase in the number of receptor sites.Membrane depolarization by 145 mM K+ leads to >90% loss

of specific toxin binding. In this respect, the properties ofNa+-channel-associated ScTx receptors evolving in vitro are notdifferent from those evolving in vivo.

t Bezwald-Netter, Y., Martin, N., Koulakoff, A. & Couraud, F. (1980)13th FEBS Meeting, August 24-29, 1980, Jerusalem, p. 116 (abstr.).

§ Bezwald-Netter, Y. Bizzini, B., Couraud, F., Koulakoff, A. & Martin-Moutot, N. (1980) 4th European Neuroscience Meeting, September16-19, 1980, Brighton, U. K., p. 312 (abstr.).

Table 1. Properties of ScTx binding sites on fetal mouse brain cells

Age of embryo, Sites Sites Density,t Density,*days Kd,* nM per cell,*t no. per TBC,t no. sites per ,m2 sites per inM2 of TBC

12 0.16 120 1,040 0.5 4.213 0.12 0.06 500 200 2,170 870 2.0 0.8 8.7 3.515 0.19 ± 0.09 800 100 2,105 ± 265 3.2 ± 0.4 8.4 1.116 0.40 ± 0.10 2,500 5,555 10.0 22.217 0.40 ± 0.10 5,350 ± 150 10,700 ± 300 21.4 ± 0.6 42.8 ± 1.218 0.46 + 0.14 13,700 ± 1,300 24,920 ± 2,380 54.8 ± 5.2 99.7 ± 9.519 0.22 20,000 33,900 80 136

* Values ± SDs were determined from computed fits of Scatchard plots of 1-3 independent experiments.t Mean, calculated for all viable cells.t Mean, calculated for TBC.

Neurobiology: Berwald-Netter et aL

1248 Neurobiology: Berwald-Netter et aL

0oo 10

50 ~ml04 ~~~50 100 5-

X 0.1 0.5 1.0 1.5 2.0 3.0 0.1 0.5 1.0 1.5 2.0I251-Labeled ScTx, nM

C D

50-~~ ~ ~ ~ 6Kd= 2X10"oM

'40-

~30-4

202

10 1

A A

0 2 4 6 8 10 0 -10 -9 -8 -7Time (in culture), days ScTx, log M

FIG. 3. Binding of ScTx to in vitro cultured fetal mouse brain cells.Cells inA, C, andD were derived from 15-day mouse embryos and cul-tured under conditions allowing for neuronal survival and differentia-tion. The culture consisted primarily (-90%) of well-differentiatedneurons and their fibers. Estimated number of cells per culture was2-3 x 106. (A) ScTx binding to cells cultured for 9 days. Cells were

incubated for 30 min at 37TC with 1 ml of Na'-free medium containingincreasing concentrations of 125I-labeled SCTx in the absence (A) orpresence (A) of 0.1 1uM native toxin. (E) Scatchard plot of specific bind-ing. (B) Results of a similar experiment on cultures of glial cells with-out visible neurons. Note lack of specific ScTx binding. (C) Quanti-tative evolution of ScTx receptors as a function of time in culture; eachpoint represents the mean of a triplicate assay of 125I-labeled ScTx(0.2 nM) binding. (D) Results of a displacement experiment on a 9-dayculture. Cells were incubated with 0.1 nM '25I-labeled ScTx withoutor with increasing concentrations of native ScTx. K0.* = 0.3 nM.

DISCUSSIONWe have demonstrated the presence and studied the quanti-tative evolution of ScTx binding sites in cells derived from nor-

mal fetal mouse brains, developing in vivo or grown in vitro.These sites, which are a known component of the action-po-tential Na' channels, become specifically detectable on the12th day of gestation. This signifies that the potential-depen-dent Na' channel, or at least a part of its structure, is expressedon the surface membrane of cells at this stage of ontogenesis.With advancing fetal age or time in culture, there is an increasein the mean ScTx binding capacity onto central nervous system-derived cells. The number of ScTx receptors, as well as theirrate of appearance, increases slowly between the 12th and the16th fetal day and rapidly thereafter. There are no changes inthe ScTx receptor properties: in all cases, binding kinetics showthe involvement of a single class of high-affinity noninteractingsites.

Quantitative binding assays do not allow a direct identifica-tion of the ScTx-receptor-bearing cells. Thus, the interpretationat the cellular level of our data raises several questions. Thecentral issues are (i) the phenotypic identity of the cells bearingScTx receptors; (ii) whether the increase in ScTx binding ca-

pacity is due to a higher receptor density per cell, or to a higherratio of receptor bearing cells, or to both; (iii) whether thehigher receptor density per cell reflects a bonafide increase indensity per surface membrane unit area or a greater cell surface

area; (iv) the distribution of the receptors; and last but not least(v) the correlation between ScTx binding and the presence ofactive Na' channels.

Several lines of evidence support the hypothesis that the cellsbearing high-affinity ScTx receptors are most likely neuroblastsor neurons. First, glia in enucleated optic tracts of Necturus sp.are reported to lack saxitoxin receptors and do not show elec-trophysiological responses characteristic of the type of Na'channels found in excitable cells (33). Our data show that thereare no specific high-affinity ScTx binding sites in cultures ofcentral nervous system glia without neurons, in which we iden-tified the three principal nonneuronal cell types of nervous tis-sue; i.e., astrocytes, oligodendrocytes, and ependymal cells. Onthe other hand, whenever we observed saturable high-affinityScTx binding onto central nervous system-derived cells, phe-notypically neuronal TBC were present. These findings arecompatible with the view that the cells binding either markermay belong to the same subpopulation of brain cells. Second,previous studies of neuronal cells have shown identities in ScTxconcentration dependence between the saturable high-affinityScTx-binding component and the activation of 'Na' influx inthe presence of veratridine (34) or the enhancement of elec-trophysiological response of neuroblastoma cells (22). Third,other recent reports have suggested the presence of "Na' chan-nels" on unexcitable cells, such as glial and fibroblastic cell lines(35, 36). These reports, however, do not yet offer evidence per-mitting one to consider the implication of the action potentialchannels that bind ScTx with high affinity. Our studies on dif-ferent cell lines and primary cultures (neurons, glia, fibroblasts,or myogenic cells) showed that only neuronal cells bind ScTxspecifically in the nanomolar range.

In interpreting the observed increase in ScTx-receptor abun-dance, it is necessary to recall that in vivo the mean size of cellsdoes not change during the fetal period considered. Therefore,it seems probable that the sharp increase in ScTx-binding thatoccurs during the later stages of prenatal development corre-sponds to a neuronal maturation-related increase in receptordensity per cell and per surface area unit. This phase is char-acterized by a progressive accumulation of postmitotic neuro-blasts and neurons (26, 27) and, in our data, by a slower increasein percentage TBC. The higher ScTx-receptor density notedmay thus be an interesting indication of a qualitative change inthe functioning of neurons and correspond to a phase whenmany start emitting action potentials. Indeed, the ScTx-recep-tor density we determined on brain cells at 16 fetal days or lateris very close to, or higher than, the estimated density of Na'channels on excitable nonmyelinated nerves (ref. 37 and ref-erences therein) or on mouse neuroblastoma cells (4). As far asthe early stages in vivo are concerned, a period during whichextensive division of neuronal precursors occurs (26, 27), otherexplanations of the data may apply. One may assume that allthe ScTx binding cells possess relatively low numbers of bindingsites or, alternatively, that there is a small population (<1.0%)of cells with numerous receptors possibly involved in the early,spontaneous neuronal activity (38). Considering both the het-erogeneity of the cell population with respect to the area oforigin in the central nervous system and the known asynchronyin the evolution pattern of the different areas (26, 27), a co-existence of both situations is the most probable. In vitro, celldivision has been blocked by cytosine arabinoside. Therefore,the increase with time in ScTx binding per culture undoubtedlyreflects the maturation of cells present in the initial population.The process of neurite sprouting and elongation precedes byabout 48 hr the phase of high rate increase in ScTx receptorsper culture. This implies that de novo membrane synthesis andreceptor appearance are not concomitant.

Proc. Nad Acad. Sci. USA 78 (1981)

Proc. Natd Acad. Sci. USA 78 (1981) 1249

No data are yet available on the cell surface distribution ofScTx receptors. Direct answers to this, as well as to other ques-tions, will have to await studies on cells from discrete areas ofthe central nervous system, using cell separation techniquesand methods that permit direct visualization of ScTx bindingsites. The recently developed photoactivatable toxin derivatives(39) offer hope for the ultimate-purification of membrane com-ponents bearing the voltage-sensitive ScTx receptors.The problem of a correlation between ligand binding and

biological activity is not straightforward, especially at early de-velopmental stages. It has recently been reported (40) that scor-pion venom-sensitive sodium fluxes are present in cells of fetalchicken brain. However, in these experiments high concentra-tions of crude scorpion venom were used, and the results aredifficult to.interpret. It still remains to be determined whetherthe appearance of high-affinity receptor sites for ScTx is cou-pled, or at least temporally correlated, with the expression offunctional Na' channels. The maturation of cells in the in vitroculture system described offers the possibility of exploring,under controlled and experimentally variable conditions, theregulatory events involved in the expression of the differentNa'-channel components for which ligands are available. Ourdata suggest that the high-affinity ScTx receptors may representquantitative markers of neuronal differentiation and that ScTxis a potentially valuable tool for dissecting the development ofthe nervous system.We appreciate the conscientious typing of Ms. D. Baron; the labo-

ratory assistance of Ms. H. Houzet; the generous supply of purifiedscorpion toxin, without which this work could not have been done, byDrs. F. Miranda and H. Rochat; and the continued interest and en-couragement of Drs. F. Gros and S. Lissitzky. The constructive com-ments of the referees are gratefully acknowledged. This study was sup-ported by Institut National de la Sante et de la Recherche MedicaleGrant ATP 81.79.113, the Centre National de la Recherche Scienti-fique, and the Lique Nationale Franoaise Centre le Cancer.

1. Catterall, W. A. & Nirenberg, M. (1973) Proc. NatL Acad. Sci. USA70, 3759-3763.

2. Palfrey, C. & Littauer, U. Z. (1976) Biochem. Biophys. Res. Com-mun. 72, 209-215.

3. Catterall, W. A. (1977)J. BioL Chem. 252, 8669-8676.4. Catterall, W. A. (1980) Annu. Rev. PharmacoL ToxicoL 20, 15-43.5. Peacock, J. H., Nelson, P. G. & Goldstone, M. W (1973) Dev. BioL

30, 137-152.6. Nelson, P. G. & Peacock, J. H. (1973) Brain Res. 61, 163-174.7. Fukuda, J. & Kameyama, M. (1979) Nature (London) 279,

546-548.8. Matsuda, Y., Yoshida, S. & Yonezawa, T. (1978) Brain Res. 154,

69-82.9. Dichter, M. A. & Fischbach, G. D. (1977)J. PhysioL (London) 267,

281-298.10. Chalazonitis, A., Greene, L. A. & Nirenberg, M. (1974) Brain Res.

68, 235-252.11. Godfrey, E. W., Nelson, P. G., Schrier, B. K., Breuer, A. C. &

Ramson, B. R. (1975) Brain Res. 90, 1-21.

12. Crain, S. M. (1976) Neurophysiologic Studies in Tissue Culture(Raven, New York), pp. 131-152.

13. Spitzer, N. C. (1979) Annu. Rev. Neurosci. 2, 363-397.14. Rochat, H., Bernard, P. & Couraud, F. (1979) Adv. Cytophar-

macol 3, 325-334.15. Romey, G., Chicheportiche, R., Lazdunski, M., Rochat, H.,

Miranda, F. & Lissitzky, S. (1975) Biochem. Biophys. Res. Com-mun. 64, 115-121.

16. Catterall, W. A., Ray, R. & Morrow, C. S. (1976) Proc. Natl Acad.Sci. USA 73, 2682-2686.

17. Couraud, F., Rochat, H. & Lissitzky, S. (1978) Biochem. Biophys.Res. Commun. 83, 1525-1530.

18. Jover, E., Martin-Moutot, N., Couraud, F. & Rochat, H. (1978)Biochem. Biophys. Res. Commun. 85, 377-382.

19. Ray, R., Morrow, C. S. & Catterall, W. A. (1978) J. Biol Chem.253, 7307-7313.

20. Couraud, F., Rochat, H. & Lissitzky, S. (1980) Biochemistry 19,457-462.

21. Bernard, P., Couraud, F. & Lissitzky, S. (1977) Biochem. Biophys.Res. Commun. 77, 782-788.

22. Bernard, P. & Couraud, F. (1979) Biochim. Biophys. Acta 533,154-168.

23. Jover, E., Martin-Moutot, N., Couraud, F. & Rochat, H. (1980)Biochemistry 19, 463-467.

24. Miranda, F., Kopeyan, C., Rochat, H., Rochat, C. & Lissitzky,S. (1970) Eur. J. Biochem. 16, 514-523.

25. Rochat, H., Tessier, M., Miranda, F. & Lissitzky, S. (1977) AnalBiochem. 82, 532-548.

26. Altman, J. (1967) in The Neurosciences: A Study Program, eds.Quarton, G. C., Melnechuk, T. & Schmitt, F. 0. (RockefellerUniv. Press, New York), pp. 723-743.

27. Angevine, J. B. (1970) in The Neurosciences Second Study Pro-gram, ed. Schmitt, F. 0. (Rockefeller Univ. Press, New York),pp. 62-72.

28. Raff, M. C., Fields, K. L., Hakomori, S., Mirsky, R., Pruss, R.M. & Winter, J. (1980) Brain Res. 174, 283-308.

29. Berwald-Netter, Y., Bizzini, B., Couraud, F. & Koulakoff, A.(1981) Workshop on Monoclonal Antibodies to Neural Antigens(Cold Spring Harbor Laboratory, Cold Spring Harbor, NY), inpress.

30. Dimpfel, W., Huang, R. T. C. & Habermann, E. (1977) J. Neu-rochem. 29, 329-334.

31. Mirsky, R., Wendon, L. M. B., Black, P., Stolkin, C. & Bray, D.(1978) Brain Res. 148, 251-259.

32. Fields, K. L., Brockes, J. P., Mirsky, R. & Wendon, L. M. B.(1978) Cell 14, 43-51.

33. Tang, C. M., Strichartz, G. R. & Orkand, R. K. (1979) J. Gen.Physiol 74, 629-642.

34. Catterall, W. A. (1977)J. Biol Chem. 252, 8660-8668.35. Munson, R., Westermark, B. & Glaser, L. (1979) Proc. Natl Acad.

Sci. USA 76, 6425-6429.36. Pouyssegur, J., Jacques, Y. & Lazdunski, M. (1980) Nature (Lon-

don) 286, 162-164.37. Ritchie, J. M. & Rogart, R. B. (1977) Proc. NatL Acad. Sci. USA

74, 211-215.38. Changeux, J. P. (1979) in The Neurosciences Fourth Study Pro-

gram (MIT Press, Cambridge, MA), pp. 749-778.39. Beneski, D. A. & Catterall, W. A. (1980) Proc. Natl Acad. Sci. USA

77, 639-643.40. Pado, C. H., Munson, P., Glaser, L. & Gottlieb, D. I. (1980) Brain

Res. 185, 187-191.

Neurobiology: Berwald-Netter et aL