Purification of charybdotoxin, a specific inhibitor of the high-conductance Ca2+-activated K+ channel.

Charybdotoxin is a high-affinity specific inhibitor of the high-conductance Ca2+-activated K+ channel found in the plasma membranes of many vertebrate cell types. Using Ca2+-activated K+ channels reconstituted into planar lipid bilayer membranes as an assay, we have purified the toxin from the venom of the scorpion Leiurus quinquestriatus by a two-step procedure involving chromatofocusing on SP-Sephadex, followed by reversed-phase high-performance liquid chromatography. Charybdotoxin is shown to be a highly basic protein with a mass of 10 kDa. Under our standard assay conditions, the purified toxin inhibits the Ca2+-activated K+ channel with an apparent dissociation constant of 3.5 nM. The protein is unusually stable, with inhibitory potency being insensitive to boiling or exposure to organic solvents. The toxin's activity is sensitive to chymotrypsin treatment and to acylation of lysine groups. The protein may be radioiodinated without loss of activity.

stability will make it an excellent biochemical probe of this channel.
In this study, we seek to purify CTX to homogeneity, to characterize its fundamental biochemical properties, and to assess its suitability as a probe in future efforts to isolate the Ca2+-activated K+ channel. By utilizing an assay for CTX based on the inhibition of Ca2+-activated K+ channels reconstituted into planar bilayer membranes, we demonstrate a rapid two-step procedure yielding pure, active CTX. We show this to be a highly basic protein of about 10 kDa molecular mass, of exceptional chemical stability, and with sites which may be radioiodinated without loss of inhibitory activity.

Fractionation of Scorpion Venom
Venom of L. quinqwstriutus was obtained as a lyophilized powder from Latoxan Scorpion Farm, Rosans, France, and was stored at -20 "C. A 200-mg sample of venom was dissolved in 35 ml of buffer A (40 mM NaCI, 10 mM sodium borate, 10 mM Na2C03, pH 9.0), and the undissolved mucoid material centrifuged at 5000 X g for 10 min. The supernatant was saved and the pellet extracted twice in about 10 ml of buffer A. The combined supernatant was loaded onto a 5-ml column of SP-Sephadex equilibrated with buffer A. The column was washed with 35 ml of buffer A until the A M of the eluate fell to below 0.08. A linear gradient (80 ml total volume) of buffer A to buffer B (20 mM NaC1, 10 mM Na2C03, 10 mM Na3P04, pH 12.0) was then applied to the column, and the protein concentration and pH of the eluent were followed. The peak of CTX activity was found to elute at Active fractions from the SP-Sephadex column were neutralized by addition of acetic acid, and were further fractionated by reversedphase HPLC, using a 150 A pore C8 column. Sample (about 200 pg of protein) was applied to the column, which was then washed with 5 ml of buffer C (41 mM acetic acid, 9 mM sodium acetate). A linear gradient (15 ml total volume) of buffer C to buffer D (50 mM acetic acid in 50% methanol) was run at 1 ml/min, and A M was monitored. The major CTX peak eluting at 25-30% methanol was in most cases repurified by a second pass through the reversed-phase column. Purified CTX was neutralized with Na2HP04, and methanol was removed by evaporation under a stream of N, gas. The stock solution of CTX was stored frozen at -20 "C in glass tubes. pH 10. [8][9][10][11].0.

Assay of CTX
Activity of CTX was assayed by the ability of this protein to block the Ca2+-activated K+ channel of skeletal muscle (7). Single Ca2+activated K+ channels from rat skeletal muscle transverse tubule membranes prepared as described (13) were incorporated into planar lipid bilayers formed from solutions of 21 mM 1-palmitoyl, 2-oleoyl phosphatidylethanolamine, 9 mM 1-palmitoyl, 2-oleoyl phosphatidylcholine (Avanti Polar Lipids, Birmingham, AL). Planar bilayers were formed with 150 mM KCl, 10 mM MOPS, 50 p~ CaCl,, 5 mM KOH as the "internal" solution, and 10 mM MOPS, 0.2 mM EGTA, 5 mM KOH as the "external" solution. Transverse tubule vesicles (5 pg/ml) were added to the internal solution with stirring, and transbilayer current was measured at a holding voltage in the range 0-35 mV (external solution ground). As soon as a single channel inserted into the bilayer, as shown by the appearance of unitary fluctuations in current of about 15 PA, further insertion was suppressed by adding 14607 150 mM NaCl to the external solution. Bovine serum albumin (50-100 pg/ml) was also added to the external solution to eliminate nonspecific binding of CTX to the chamber walls. With this method, the Ca2+-activated K' channels always incorporated with the CTXsensitive side facing the side of the membrane opposite to that on which the vesicles are added (7,14). A detailed description of the planar bilayer technique has been reported previously (15).
After channel insertion, a control record was recorded at a holding voltage of 20-40 mV to ensure that no "blocking" events were detected. A sample of CTX-containing solution was then added, and the subsequent channel record was collected on videotape for 10-20 min until 20-50 blocking events had occurred. The probability of being blocked #as measured directly from this record as the fraction of time spent in the blocked state.
Polyacrylamide Gel Electrophoresis SDS-polyacrylamide gels containing 6 M urea were employed, according to Swank and Munkries (16), to resolve the venom proteins, most of which are of low molecular masses. Gels were run under reducing conditions or nonreducing conditions. Reducing Conditions-Protein was reduced by boiling for 10 min in 10 mM P-mercaptoethanol, 1% SDS, 10 mM H3P04 adjusted to pH 6.8 with Tris base, or by heating 20 min at 70 "C in the above buffer containing 8 M urea in addition. Gels were cast at 12.5% acrylamide with a bis/acrylamide ratio of 1:lO in 0.1% SDS, 6 M urea, 0.1 M H3PO4, adjusted to pH 6.8 with Tris base, casting, gels were preelectrophoresed 3-4 h in 0.1 M Tris phosphate, 10 mM 8-mercaptopropionate, pH 6.8. Samples (2-5 pg per lane) were then loaded, and gels were run 16 h at room temperature at 20-30 mA. Gels were silver-stained as described by Merril et al. (17). Nonreducing Conditions-For some experiments, it was necessary to preserve CTX activity during electrophoresis for subsequent assay. Samples were prepared as above, except that @-mercaptoethanol was omitted. Gels were cast as above, but were not pre-electrophoresed with P-mercaptoproprionate. Each sample (-20 pg) was loaded on two lanes of the gel, one to be used for staining to locate the position of the band, and one for elution and assay of toxin. After electrophoresis as above, protein bands were identified by staining with Coomassie Brilliant Blue R-250. Parallel unstained bands were cut from the gel, and protein was eluted overnight at room temperature into 1 ml of 10 mM Tris-HC1, pH 9. Concentration of eluted protein was determined by the fluorescamine method (18), and CTX inhibition of channels was assayed in planar bilayers as described above. Controls demonstrated no contribution to the channel-blocking activity from acrylamide, SDS, or urea.

Iodination of CTX
Charybdotoxin was iodinated by the lactoperoxidase method (19). Reaction conditions were 50 mM Napi, 4 mCi/ml NalZ5I, 1 unit/ml lactoperoxidase, 100 p~ H202, 5 pM CTX, pH 7.2. After 15 min at room temperature, another aliquot of lactoperoxidase and Hz02 was added. After another 15 min, CTX was separated from the other reagents by SP-Sephadex cation exchange chromatography as described above.
Low specific activity iodinations and nonradioactive iodinations were carried out under similar conditions. Low specific activity iodinations were done in 0.5 mM NaI, 50 pCi/ml NalZ5I. Nonradioactive reactions were carried out with 0.5-2.0 mM NaI only. Ca2'-activated K+ channel-blocking activity of iodo-CTX was measured in planar bilayer assays with nonradioactive, iodinated CTX samples by the assay technique previously described.
Autoradiography was carried out on unstained gels. Gels were dried and developed with Kodak X-Omat AR film for 2-12 days at -70 "C.
In some experiments dried gels were cut into separate lanes and sliced at 1-cm intervals. Slices were counted for 1 or 2 min in a Packard ycounter to measure "'1 content.
Amino Acid Analysis Samples were prepared for analysis as follows. An aliquot of protein pended in 0.5 ml of 6 N HCl with 5 p l of phenol and 5 pl of 8-(-50 pg) was dried in a vacuum centrifuge. The sample was resusmercaptoethanol. After flushing with N2 gas, samples were evacuated, freeze-thawed, sealed, and heated to 110 "C for 20 h.
Performic acid oxidation was done to determine cysteine content.
A freshly made oxidant solution (9 volumes of formic acid + 1 volume of 30% H202) was incubated for 2 h at 27°C in the dark. The solution was then placed in -20 "C in the dark for 15 min, and 100 p1 was added to 5 nmol of dried CTX. The oxidation reaction was allowed to proceed for 1 h at -20 "C in the dark. The reaction products were dried under Nz gas. 100 pl of water was added, and the mixture was dried under vacuum to remove any remaining oxidants. Samples were then hydrolyzed and analyzed for amino acid content. Quantitative amino acid analysis was carried out commercially (Sequemat, Watertown, MA) by standard HPLC ion exchange techniques, using ninhydrin detection.

N-terminal Analysis
The CTX N-terminal amino acid was identified by the dansyl chloride method of Gray (20) with some modifications. Approximately 1 nmol of CTX was added to 30 pl of pyridine and 10 pl of 0.2 M NaHC03 and dried in a vacuum centrifuge. 10 pl of water and 5 pl of 0.5% dansyl chloride in acetone were added, and the reaction was incubated at 37 "C for 30 min. The products were evaporated to dryness in a vacuum centrifuge, and 0.5 ml of 6 N HCl were added. The tube was sealed under vacuum, freeze-thawed on dry ice twice under vacuum, sealed, and heated at 110 "C for 16 h. The hydrolysate was dried in a heated vacuum centrifuge, dissolved in 50 pl of water, and redried. The dried products were dissolved in 5 pl of 50% pyridine and analyzed by thin-layer chromatograpy on polyamide (21). Known dansylated amino acid standards were prepared by the method of Gray (20).

Determination of Mass Extinction Coefficient
It is both theoretically expected and empirically established (22) that the mass concentration of protein in solution (g/liter), can he accurately estimated from the absorbance at 205 and 280 nm. We used this method to estimate the mass-based extinction coefficient at 280 nm for pure CTX. A sample of CTX (-10 nmol) was adsorbed to a 0.2-ml SP-Sephadex column equilibrated in 0.1 M Na2S04, 10 mM NaPi, pH 7.0, and this was then extensively washed with the equilibration buffer to remove all traces of acetate and C1-, both of which severely interfere with the 205 nm absorbance. The CTX was eluted by washing the column with 0.5 ml of the above SO, buffer, adjusted to pH 11.8 with NaOH. The eluate was neutralized to pH 7.0 with H2S04, and the absorbance of this solution was measured against appropriate blanks in a 1-cm cuvette at 280 nm, and, after 10-40-fold dilutions, at 205 nm. The protein concentration, P, in g/ liter, is: This expression for mass concentration of protein has been found to be invariant to within +15%, regardless of the particular amino acid composition of the protein. The mass extinction coefficient, C '~~O , is then dm = Am/P.

Inhibition of Ca2+-actiuated K' Channels by Leiurus
Venom-The venom of L. quinquestriatus contains a potent inhibitor of the high-conductance Ca2+-activated K+ channel (7). It is easy to observe this inhibition at the molecular level by reconstituting single channels into planar bilayers and adding venom to the solution equivalent to the externally facing side of the channel protein. In Fig. LA we illustrate the effect of the venom. In the absence of venom, the channel opens and closes randomly on a time scale of milliseconds. Addition of venom to the external solution induces the appearance of nonconducting events which are long (1-10 S) relative to normal channel-closing events. The appearance of this venom-induced "blocked" state becomes more frequent as venom concentration is increased, but the average duration of the blocked state is independent of venom concentration (7), as would be expected for a bimolecular process in which each blocking event represents the binding of a single molecule of a toxin in the crude venom with the channel.
We can use this inhibitory effect to define a "blocking activity" by which we can quantify the agent in the venom giving rise to channel inhibition, For a simple bimolecular inhibition process, the observed probability of the channel's ient two-step method for purification of CTX from the crude scorpion venom. We first exploit the unusually basic nature (') of the protein by employing SP-Sephadex cation exchange where [CTX] represents the concentration of inhibitor in the assay medium, and Kd its apparent association constant. This blocking probability is measured directly from the singlechannel record. Accordingly, we define the blocking activity as: (2) If the inhibitor residing in the crude venom interacts with the channel in a bimolecular fashion, i.e. according to the assumptions underlying Equation 1, then this blocking activity should be linearly dependent on the venom concentration in the assay. Fig. 1B shows that this is in fact the case.
We are therefore in a position to deveIop a linear assay which we may use to purify the channel inhibitor. The procedure is as follows. A single channel is incorporated into a planar bilayer membrane, and then an aliquot of a test fraction of venom is added to the external side of the bilayer, to a final protein concentration, c. The blocking activity is measured as defined above, and a specific activity is defined as: specific activity = activity/c.
In this way, we may quantitatively follow the channel inhibitor through a purification procedure.
Charybdotoxin Purification-We have developed a conven-chromatography and pH gradient elution. A final purification by reversed-phase HPLC is then employed. Both of these steps can be used because of the exceptional chemical stability of CTX, which may be exposed to extreme ranges of pH and to solvents without loss of activity, as will be documented below. Fig. 2A shows the protein elution profile from an SP-Sephadex column run in a chromatofocusing mode, i.e. at low ionic strength and increasing pH. Under relatively basic conditions (pH 9.0), over 90% of the protein in the crude venom runs through the cation exchange column. The protein which remains on the column is then eluted by steadily increasing the pH of the eluate, and hence neutralizing the proteins according to their isoelectric points. Channel-blocking activity is found only in the last protein peak, which elutes at pH 10.8-11.0. This is consistent with the fact that purified CTX runs off an isoelectric focusing gel utilizing a pH gradient of 8.0-10.8 (data not shown), indicating that the isoelectric point of this protein is greater than 10.8.
The SP-Sephadex fractions containing CTX activity are applied to a reversed-phase C8 column (Fig. 2B). The inhibitor is eluted as the major protein peak at about 25% methanol. Two minor peaks can also be seen, one just leading and one just lagging the major peak; these two minor peaks can be entirely removed by collecting only the central part of the

TABLE I Chqybdotorin purification
Purification of CTX as illustrated in Fig. 2. For this experiment, 200 mg of dry venom was weighed out and extracted as described under "Materials and Methods," to yield the "crude venom." Protein weights reported were determined by absorbance at 280 nm (1 mg/ ml = 1 absorbance unit), and therefore should be considered only approximate. major peak, as may be shown by rechromatography on the reversed-phase column (data not shown). The efficiency of this purification scheme is summarized in Table I. The chromatofocusing step recovers 90% of the total blocking activity, and the complete procedure recovers 80% of the total activity, with an approximately 300-fold purification over the water-extractable part of crude venom. We have performed this procedure numerous times on several different batches of L. quinquestriutus venom, and have found both the yield and purification profile to be reproducible. Given the extinction coefficient and molecular mass of CTX (see below), we can estimate that this protein accounts for about 0.13% of the crude scorpion venom, and that about 1 mg of pure CTX can be recovered from 1 g of dried venom.
Previous work (7) suggested that CTX is a low molecular mass protein. We examined the mobility of purified CTX and other venom protein fractions with SDS-polyacrylamide gel electrophoresis, using highly cross-linked gels containing 6 M urea (16). The purification of CTX can be visualized on these gels (Fig. 3) with the pure product showing only a single band, even on highly overloaded gels. Reducing gels give an apparent molecular mass for CTX of about 6 kDa. We also show the protein mobility pattern on nonreducing SDS/urea gels; again, the purified material moves as a single band, and the blocking activity is preserved after cutting this band out of the gel and eluting the protein from the polyacrylamide matrix, as is shown in the figure.
Additional evidence of protein purity is provided by Nterminal amino acid analysis. After reacting the protein with dansyl chloride at pH 7.5 and hydrolyzing it, we detect only a single dansylated residue by thin-layer chromatography on polyamide plates (21). This result indicates the presence of a single species of N terminus in the protein sample. By using three different solvent systems (1.5% formic acid; benzene/ acetic acid, 9:l; and ethylacetate/acetic acid/methanol, 20:1:1), we unambiguously identified the N terminus as valine (data not shown). This result is consistent with a preliminary sequence analysis (carried out by Dr. Clive Slaughter, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY), in which valine was identified in the first cycle of Edman degradations.

Amino Acid Analysis and Determination of Molecular
Mass-We analyzed the amino acid composition of the purified CTX sample following hydrolysis in 6 N HCI, under conditions where amino acid degradation is minimized. We find ( Table 11) that most amino acids appear in integral multiples of alanine. This result further validates the purity of the sample, and it allows us to estimate a minimum molecular mass. Assuming that CTX contains a single alanine, the molecular mass is 9.2 kDa. Amino acid analysis also allows us to determine the molar extinction coefficient of CTX, again based on the assumption that the protein contains a single alanine residue. By measur-  Determined in a separate experiment by performic acid oxidation as described under "Materials and Methods."

TABLE 111
Determination of molecular mass and extinction coefficients of C T X Three different methods were used to estimate values for molecular mass or extinction Coefficient in order to allow CTX concentrations to be estimated from measured A280 of the purified material. Amino acid analysis is from Table I1 and gel electrophoresis from Fig. 3.
Mass extinction coefficient, c', (280 nm) was determined as described in the text, and molecular mass was calculated by dividing this by the molar extinction coefficient, e, derived from amino acid analysis.

Method
Molecular mass Amino acid analysis 9.3 17,500 Gel electrophoresis 6 Mass extinction 11.5 1.5 ing the alanine content of a sample of known 280 nm absorbance, we find a molar extinction coefficient (at 280 nm) of 17,500 M-' cm", Table 111). Thus, we can now convert the absorbance at 280 nm of a purified CTX sample directly into molarity.
A final method for determining the molecular mass of CTX is t o combine the above value of molar extinction coefficient with an independently measured value of mass-based extinction coefficient. Molecular mass is given simply as the ratio of these two extinction coefficients. The mass-based extinction coefficient can be measured within 15% from the absorbance at 205 nm, corrected for the absorbance at 280 nm (22); this gives a n empirical measure of the concentration of peptide bonds in solution, and has been shown to be invariant for proteins of widely differing amino acid compositions. In Table  I11 we estimate a molecular mass for CTX of 11.5 kDa using this method. Table I11 also summarizes our three independent determinations of molecular mass.

Properties of Purified
Charybdotoxin-Charybdotoxin is a n exceptionally stable protein, whose activity is maintained under conditions considered harsh for most proteins. In Table   IV we demonstrate that CTX activity is present after boiling and after overnight incubation at p H 2, at pH 12, in 0.1% SDS and 6 M urea, or in 50% methanol. It is this stability  1-3 p~) was treated under the following conditions: overnight incubation in SDS (O.l%)/urea (6 M), 4 "C; CTX boiled 10 min in 100 mM KCl, 10 mM MOPS, pH 7, with or without 10 mM dithiothreitol; methanol incubation overnight at 4 'C in 50% methanol, 50 mM sodium acetate/acetic acid, pH 4; incubation at pH 12 overnight at 4 "C in 100 mM KCl, 10 mM Nap,; incubation at pH 2 in 100 mM KCl, 10 mM HC1, overnight at 4 "C; iodination as described under "Materials and Methods," yielding a stoichiometry of I/CTX of 1.4 in this experiment; chymotrypsin reaction at 25 pg/ml chymotrypsin in 100 mM KCl, 10 mM Napi, pH 7, for 30 min at room temperature; and citraconylation carried out at 70 p~ citraconic anhydride, 50 mM NaCl, 20 mM Napi, pH 8, 30 min, 20 "C. Activity is reported relative to control activity calculated as the mean of 12 determinations of CTX activity measured in 10 separate experiments. Each determination of activity is the average of two to five separate exDeriments. Channel-blocking activity was measured as for Fig. 1, but using CTX purified to homogeneity and of known molarity. Activity is defined as described in text.
which enables us to purify the protein so straightforwardly. The activity is totally lost upon chymotrypsin treatment, by boiling in the presence of dithiothreitol, and by exposure to lysine-blocking agents such as citraconic anhydride. The fact that inhibition produced by citraconic anhydride can be fully reversed by incubation at pH 4 for several hours (data not shown) argues that the anhydride inhibition operates by acylation of an amino group critical for blocking activity.
For our long-range purposes, we wish most urgently to know the affinity of CTX for the Ca2+-activated K' channel.
The present purification allows us to determine this accurately. According to Equations 1 and 2, the apparent Kd is measured from the proportionality between blocking activity and CTX concentration, as illustrated in Fig. 4. Under our assay conditions, we find an apparent dissociation constant of 3.5 nM, 3-5-fold lower than we had previously estimated using impure material (7).
Iodination of CTX-Since CTX contains 2 tyrosine residues, we attempted to iodinate the toxin. After trying several methods, we settled upon the lactoperoxidase reaction as the most efficient and reproducible in this case. In order to use radioiodinated CTX as a ligand for the Ca2+-activated K+ channel, it is necessary to demonstrate that the sample is radiopure, that the radioactive sample is 'T-CTX, and that iodo-CTX retains the ability to bind to the channel. The first point is illustrated in Fig. 5. Lane A contains the products of the lactoperoxidase iodination reaction after removal of lZ5I-. Clearly, the sample contains impurities, most of which are not protein (data not shown). After subjecting the sample to SP-Sephadex cation exchange chromatography (as described previously for unlabeled toxin), the sample is rendered essentially radiopure ( l a n e B ) . Fig. 5 also illustrates the second point. Nonreducing gels were run with iodo-CTX and unlabeled CTX. Gels were sliced into lanes and each lane was sliced into 1-cm pieces, which were then counted in a y-counter. Parallel unlabeled CTX samples were eluted overnight. Protein yields and channelblocking activity from these eluted samples were determined. The figure shows that CTX activity and radioactivity comigrate in this system.
We measured iodo-CTX activity by reacting under conditions known to radiolabel the toxin (stoichiometry of 1.4 I-/ CTX), but using nonradioactive NaI. The reaction products were assayed for their ability to block Ca2+-activated K+ channels (Table IV). Table IV shows that iodo-CTX retains full channel-blocking activity.

DISCUSSION
This report shows that it is possible to purify charybdotoxin from the 50-60 proteins making up L. quinquestriutus venom. This inhibitor of Ca2+-activated K+ channels is a minor component of the venom, constituting approximately 0.1% of the total venom protein, by weight.

M I G R A T I O N D I S T A N C E . c m
FIG. 5. Migration of iodo-CTX and CTX activity on polyacrylamide gels. Nonreducing gels were run with unlabeled CTX and iodo-CTX before and after SP-Sephadex chromatography of the lactoperoxidase iodination reaction products. The autoradiogram was produced by a I-week exposure at -70 "C. Iodination reaction was stopped by passage over a 1-ml Dowex I-X8 (acetate) column to remove lZ5I-. Material eluting through the column, containing all the CTX-blocking activity, was examined in lane A. This impure 1251-CTX was then purified on a 1-ml SP-Sephadex column as described under "Materials and Methods" and applied to lane B. Lanes containing radiopure iodo-CTX and unmodified CTX were sliced into 1cm pieces and counted for lz5I content ( 1 3 ) or eluted for assay of blocking activity (*).
purified CTX on the reversed-phase column, we estimate that this preparation is at least 95% pure, and this conclusion is supported by the single amino acid identified by end-group analysis and the appearance of a single band on SDS/ureapolyacrylamide electrophoresis gels.
The molecular mass of this material has been determined by three independent methods, which do not all agree (Table  111). Mobility on SDS-urea gels under reducing conditions yields a molecular mass of about 6 kDa, whereas the amino acid analysis implies a minimum molecular mass of 9.2 kDa. For small proteins, SDS-gel electrophoresis is a notoriously unreliable method for determining molecular mass, and we tend to discount the low value obtained. This conclusion is further strengthened by the comparison of molar uersus mass extinction coefficients (Table 111), which implies a molecular mass of 11.5 kDa. We have previously reported that CTX activity is retained inside of 2-kDa cutoff dialysis tubing, but readily permeates 10-kDa cutoff tubing (7); this fact argues that the true molecular mass is equal to, and is not an integral multiple of, the minimum molecular mass estimated above. Accordingly, we tentatively assign a molecular mass of 10 kDa to this protein. A refinement of this value awaits the complete sequencing of the protein, which is now only partially accomplished.
Several lines of evidence demonstrate that CTX is a basic protein.
Of the approximately 75 amino acid residues in this protein, 8 are lysine and 6 arginine, whereas at most 10 groups are acidic. The molecule is retained on cation exchange columns under basic conditions which elute most other venom proteins. The isoelectric point cannot be measured with existing techniques, as the CTX molecule is not focused on a gel pH gradient of 8-10.8, but the elution profile of the chromatofocusing column (Fig. 2 4 ) argues that the isoelectric point of CTX is near 11.0. The protein contains critical disulfide bonds, since activity is lost by treatment with dithiothreitol. These disulfides probably contribute substantially to the unusual stability of the venom protein; it is known that venom proteins from buthid scorpions commonly contain several intramolecular disulfide linkages (23).
The exceptionally stable CTX molecule appears to be an excellent choice of high-affinity ligand to employ in a biochemical assay for the high-conductance Ca2+-activated K+ channel. Indeed, CTX is the only inhibitor displaying nanomolar affinity for this channel. We are now in a position to compare the present measurements of CTX inhibition of single channels with direct binding of IZ5I-CTX to membrane vesicles containing this channel and to commence efforts to purify the channel protein itself.