Design, Synthesis, and Functional Expression of a Gene for Charybdotoxin, a Peptide Blocker of K+ Channels CHUL-SEUNG PARK, SHARON F

Proc. Nati. Acad. Sci. USA Vol. 88, pp. 2046-2050, March 1991 Biochemistry Design, synthesis, and functional expression of a gene for charybdotoxin, a peptide blocker of K+ channels CHUL-SEUNG PARK, SHARON F. HAUSDORFF, AND CHRISTOPHER MILLER Howard Hughes Medical Institute, Graduate Department of Biochemistry, Brandeis University, Waltham, MA 02254 Communicated by Clay M. Armstrong, December 6, 1990

ABSTRACT A gene encoding charybdotoxin (CTX), a K+ We have designed a synthetic gene for CTX and have channel blocker from scorpion venom, was designed, synthe- achieved its high-level expression in Escherichia coli by using sized, and expressed as a cleavable fusion protein inEscherichia a cleavable fusion protein strategy. The following three in coli. A sequence-specific protease, factor Xa, was used to cleave vitro posttranslational processing steps were required to yield the fusion protein and thus release the toxin peptide. The fully functional CTX: (i) proteolytic cleavage of the CTX recombinant toxin was purified, oxidized to form disulfide coding sequence, (ii) oxidation ofthe six cysteine residues to bonds, and treated to form N-terminal pyroglutamate. Recom- form three disulfide bonds, and (iii) formation of N-terminal binant CTX is identical to the native venom CTX with respect pyroglutamate. This system will permit a close mechanistic to high-performance liquid chromatography mobility, amino study of the toxin-K+ channel interaction by a combination acid composition, and N-terminal modification. With single of reconstitution of single K+ channels and site-directed Ca2+-activated K+ channels as an assay system, recombinant mutagenesis of CTX. CTX shows blocking and dissociation kinetics identical to the native venom toxin. The synthetic gene and high-level expression of functionally active CTX make it possible to study the MATERIALS AND METHODS fundamental mechanism of the toxin-ion channel interaction. Bacterial Strains and Plasmids. E. coli DH1 was used for plasmid propagation and BL21(DE3) was used for the expres- Charybdotoxin (CTX) is a basic peptide originally isolated sion ofthe gene 9-X,-CTX fusion protein. E. coli BL21(DE3) from the venom of the scorpion Leiurus quinquestriatus as a is a A lysogen of BL21 wherein the prophage contains the high-affinity blocker of the high-conductance Ca2l-activated RNA polymerase gene ofT7 bacteriophage under the control K+ channel from skeletal muscle (1, 2). This peptide also of the lacUV5 promoter (21). Plasmid pSR9, a vector for inhibits Ca2+-activated K+ channels found in neurons (3-5), producing fusion proteins with the gene 9 protein of T7 kidney (6), and erythrocytes (7, 8); in addition, CTX blocks bacteriophage, was obtained from K. M. Blumenthal (22). voltage-activated K+ channels from a variety of sources (5, Design, Synthesis, and Construction ofa CTX Gene. A DNA 7, 9-12). In all cases, CTX inhibits in the nanomolar con- sequence encoding the CTX peptide was designed to contain centration range by binding to the external part of the K+ the maximum number ofunique restriction enzyme sites. The channel. Inhibition of the channel is known to be due to the gene was constructed from four overlapping oligonucleotide physical occlusion of the channel's conduction pathway by pairs, each of approximately 45 base pairs. Each oligonucle- the toxin (13-15). Because the channel-toxin interaction is otide was chemically synthesized on an Applied Biosystems mechanistically understood, CTX has proven to be an infor- model 380A DNA synthesizer, gel-purified, and phosphoryl- mative of the external "mouth" of ated. The eight oligonucleotides were annealed and ligated probe K+ channels using T4 DNA ligase. The synthetic CTX gene, which is (16-18). flanked by Sal I and HindIII sites, was inserted into pSR9 for The amino acid sequence of CTX, determined by peptide the gene 9-Xa-CTX construction pCSP105. After cloning, sequencing (19), is shown in Fig. 1. The toxin is a 37-residue the sequence of the synthetic CTX gene was confirmed by polypeptide of 4.3 kDa, with eight positively charged resi- DNA sequencing of both strands. General rules and con- dues, two negative charges, three disulfide bonds, and a straints for producing synthetic genes were followed (23). blocked N terminus formed by a pyroglutamate residue. The Expression and Purification of CTX Fusion Protein. E. coli three-dimensional structure of the toxin is known (20). Built BL21(DE3) cells harboring pCSP105 were grown at 37°C in on a foundation of a three-strand antiparallel f3-sheet, the LB medium in the presence ofampicillin (100 ,ug/ml). Fusion molecule is roughly ellipsoidal, with major and minor axes of protein synthesis was induced at late logarithmic phase 2.5 nm and 1.5 nm, respectively. The functional groups show (OD650, 0.8 to 1.0) by the addition of 0.5 mM isopropyl a remarkable spatial segregation on the surface of the toxin; B-D-thiogalactoside (IPTG). Cells were harvested 2-3 hr one face of the molecule is strongly polar, whereas most of later, washed, and resuspended in 50 ml of 10 mM Tris HCI, the hydrophobic groups project off the opposite face. pH 8.0/50 mM NaCl/2 mM Na2EDTA. Lysozyme was added Electrophysiological studies have yielded information to 2 mg/ml and, after 20 min on ice, 2-mercaptoethanol, about the mechanism of toxin block, but it has been difficult phenylmethylsulfonyl fluoride, pepstatin, and leupeptin were to probe the toxin biochemically. Purification ofCTX from its added to final concentrations of 10 mM, 1 mM, 1 ,M, and 1 native source is costly and yields only a small amount of AM, respectively. After a brief sonication, the extract was protein. In addition, because oflack ofspecificity ofchemical clarified by a centrifugation at 27,000 x g for 30 min. All of modifications, information on the functions of specific resi- the following steps were performed at 4°C. Nucleic acids dues is nonexistent. To overcome this problem, we have were precipitated by the slow addition of 0.1 vol of 30% developed a system to produce large quantities ofgenetically (wt/vol) streptomycin sulfate and subsequent centrifugation. manipulable CTX. The fusion protein was precipitated by the slow addition of solid ammonium sulfate to 50% saturation. After centrifuga- The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" Abbreviations: CTX, charybdotoxin; IPTG, isopropyl -D- in accordance with 18 U.S.C. §1734 solely to indicate this fact. thiogalactoside. 2046 Downloaded by guest on September 29, 2021 Biochemistry: Park et al. Proc. Natl. Acad. Sci. USA 88 (1991) 2047

Mael Avrll planar bilayers. Lipids, membrane preparation, native venom CTX preparation, and planar bilayer methods were as described (2). Measurements of CTX blocking and dissociation kinetics of single K+ channels were carried out by observing at least 100 discrete blocking events caused by Ampr addition of CTX to the external side of the bilayer. In each experiment, a single channel was incorporated into the planar bilayer by addition of plasma membrane vesicles under osmotic gradients; further channel insertion was suppressed by adjusting the salt concentration to equimolar. The final solutions on the two sides of the membrane were as follows: the internal solution contained 10 mM Hepes (pH 7.4), 150 mM KCl, and 30 ,uM CaC12 and the external solution con- Ori tained 10 mM Hepes (pH 7.4), 150 mM KCI, 0.1 mM EGTA, bovine serum albumin (30 pug/ml), and the desired concentration of CTX. Gel Electrophoresis, Immunoblotting, and Analytical Meth- ods. Fusion protein and factor Xa digestion products were Sall 1 analyzed by SDS/polyacrylamide gel electrophoresis on 12% GTC GAC GGA TCC ATC GAA GGT CGT CAA TT1 ACC AAT GTT TCT TGT ACC ACT Val Asp Gly Ser Ik -C& jay AM Gin Phe Thr Asn Val Ser Cys Thr Thr gels and Coomassie brilliant blue staining. Immunoblot analysis was done by electroblotting proteins from the SDS gel 10 20 TCT AAG GAA TGT TGG TCC GTT TGT CAA CGT CTG CAT AAC ACC AGC CGC GGT onto nitrocellulose filters and treating them with total rabbit Ser Lys Glu Cys Trp Ser Val Cys Gin Arg Leu His Ann Ir Ser Aig Gly antiserum against native venom CTX. Amino acid analysis 30 37 HindIlI was done using a Waters Pico-Tag system. The molar ex- AAA TGC ATG AAC AAA AAA TGT CGT TGT TAC TCC TAG GAA TTC CAA GCI T Lys Cys Met Asn Lys Lys Cys Ag Cys Tyr Ser End tinction coefficients of both the native and the recombinant CTX were calculated based on two amino acid analysis FIG. 1. Design of synthetic gene for CTX and its expression in E. results. The peptide sequencing of the N-terminal uncyclized coli. (Upper) The expression vector of gene 9-Xa-CTX fusion and cyclized recombinant CTX was performed using an protein is shown. Vector pCSP105 was constructed by insertion of a Applied Biosystems model 475A protein sequencer. Before synthetic CTX gene and flanking sequences into Sal I and HindIII peptide sequencing, cysteine residues of native and the sites of pSR9. CTX coding region is hatched, Xa recognition se- N-terminal cyclized recombinant CTX were reduced with quence is solid, and the last 12 bases of T7 gene 9 are stippled in the with iodoaceta- plasmid map. Ampr, ampicillin resistance; Ori, origin. (Lower) DNA dithiothreitol and carboxamidomethylated sequence of synthetic gene and corresponding amino acid sequence mide, and the peptide was deblocked by pyroglutamate are shown. CTX coding sequence is in bold (from Gln-1 to Ser-37) aminopeptidase (sequencing grade, Boehringer Mannheim) and factor Xa recognition sequence is underlined (J-_hj-QIY-Arg). (19).

tion, the precipitate was dissolved in 10 ml of 50 mM Tris-HCI, pH 7.0/50 mM NaCI/5 mM 2-mercaptoethanol and RESULTS dialyzed against 2 liters ofthe same buffer. The dialysate was Design and Construction of CTX Gene Expression System. fractionated on a DEAE-cellulose (2.5 x 12 cm, DE52, The overall strategy for expressing recombinant CTX in E. Whatman) column with the same buffer by using a 0.05 M-0.5 coli is indicated in Fig. 1. The peptide was produced as a M NaCI linear gradient (total volume, 300 ml). Fusion protein fusion protein in which the CTX coding sequence was fused fractions were pooled, concentrated, stored at -20°C. onto the C-terminal portion of the T7 gene 9 protein, as Fusion Protein Cleavage by Factor X. The DEAE-cellulose- described by Howell and Blumenthal (22). The tetrapeptide purified fusion protein pool was dialyzed against 2 liters of50 sequence Ile-Glu-Gly-Arg, which is recognized by the "re- mM Tris HCI, pH 8.3/150 mM NaCl/0.5 mM 2-mercapto- striction protease" blood coagulation factor Xa, was posi- ethanol. After dialysis, 3 mM CaC12 was added. CTX was tioned immediately upstream from the CTX sequence. The cleaved from the fusion protein by factor Xa (restriction N-terminal residue of native CTX is pyroglutamate, which protease factor Xa, Boehringer Mannheim, 10 ,ug/mg of may be formed by nonenzymatic cyclization of N-terminal fusion protein) at room temperature for up to 36 hr. The glutamine. A translation termination codon was inserted at digestion mixture was loaded on a Mono-S FPLC column the end of the CTX coding sequence. equilibrated with 30 mM sodium phosphate (pH 7.5). Material The nucleotide sequence encoding the 37 residues of CTX was eluted with a total of 20 ml of a 0-0.6 M NaCl linear was designed to maximize the number of unique restriction gradient. sites. Nine restriction sites in the CTX coding region and four N-Terminal Cyclization and Purification of Recombinant sites in the flanking region were inserted into the gene to CTX. To cyclize the N-terminal glutamine, 5% (vol/vol) facilitate future cassette-mutagenesis studies. Eight chemi- acetic acid was added to the uncyclized CTX fraction from cally synthesized oligonucleotides were individually gel- the Mono-S column. After incubation at 65°C for the appro- purified and phosphorylated, and the 5' end groups were priate time (usually 2 hr), the material was rechromato- confirmed by 5' end analysis. The oligonucleotide duplexes graphed on Mono-S by FPLC as above. CTX with N-terminal were annealed and ligated in a single reaction. The full-length pyroglutamate, which elutes earlier than uncyclized CTX, gene was inserted into the Sal I and HindIII sites ofthe fusion was further purified by reverse-phase HPLC on a C18 column vector pSR9, downstream from the T7 promoter. The final Vydac, 4.6 x 250 mm, S,u). A strong-cation-exchange HPLC construct pCSP105 was propagated in E. coli DH1. column (The Nest Group, Southport, MA, polysulfoethyl Expression and Purification of Recombinant CTX. The gene aspartamide, 4.6 x 200 mm, 5,u) was used to confirm the 9-CTX fusion protein was expressed in E. coli BL21(DE3), the purity of the final recombinant CTX. chromosome of which carries a lactose-inducible T7 RNA Planar Bilayer Assay of Native and Recombinant CTX. The polymerase gene. Bacteria freshly transformed with pCSP105 channel blocking activity of both native and recombinant were induced in late logarithmic phase by addition ofIPTG. The CTX was measured on single high-conductance Ca2+- soluble fusion protein was partially purified by ammonium activated K+ channels from rat skeletal muscle inserted into sulfate precipitation and DEAE-cellulose column chromatog- Downloaded by guest on September 29, 2021 2048 Biochemistry: Park et A Proc. Natl. Acad. Sci. USA 88 (1991) A B 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 9

- 0Q) \ \ _u _ 1 * _ .-- () \ of.-i -- o'!1c

u.gI3 |_...... _| | .. . FIG. 2. Expression and cleavage of gene 9-Xa-CTX fusion protein. Fusion protein gene 9-Xa-CTX was expressed in E. coli BL21(DE3) by IPTG induction. The fusion protein was purified with ammonium sulfate precipitation and DEAE-cellulose column chromatography. CTX was liberated from the fusion protein by factor Xa. Coomassie-stained gel (A) and immunoblot (B) of Laemmli 12% polyacrylamide gel of uninduced cell-free extract ofE. coli carrying pCSP105 (lane 1); total cell-free extract induced with IPTG for 2 hr (lane 2); soluble fraction (lane 3); DEAE-cellulose pool offusion protein at 0-hr factor Xa digestion (lane 4); 3-, 8-, 15-, and 24-hr factor Xa digestion (lanes 5-8, respectively); reverse-phase HPLC purified fully processed recombinant CTX (lane 9 of immunoblot). Molecular mass markers are indicated at the left at 84, 47, 33, 24, 16, and 6.2 kDa. raphy. After chromatography, the fusion protein was greater to pyroglutamate at 1 hr and was fully cyclized in 4 hr. When than 60% pure as judged by Coomassie staining of SDS/ the acid-treated recombinant CTX was chromatographed on polyacrylamide gels (Fig. 2A). The yield of fusion protein was a Mono-S column, a new peak was eluted at the position of approximately 30 mg/liter of culture. native CTX, as shown in Fig. 3A. This shift in elution time is The partially purified fusion protein was digested with a a result of the single charge difference between the cyclized sequence-specific protease, blood coagulation factor Xa. and uncyclized forms of CTX. Over 24 hr, a major fusion protein band at 42 kDa was Since the cyclized and oxidized CTX was fully active, as gradually split into two products, the gene 9 product portion we show below, we infer that the proper disulfide bonds were of the fusion protein at 37 kDa and a small protein at 5 kDa, formed during this treatment. The oxidized and cyclized as seen on Coomassie-stained gels and immunoblots (Fig. 2). recombinant CTX was purified to complete homogeneity on Minor bands in the immunoblot are likely to be nonspecific C18 reverse-phase HPLC (Fig. 3B). The final yield of recom- bacterial proteins cross-reacting with anti-CTX antiserum. binant CTX was approximately 1 mg/liter of culture. The recombinant CTX was further purified by cation- Chemical Characterization of Recombinant CTX. The exchange chromatography, where a major peak was eluted at amino acid compositions of native and recombinant CTX, approximately 360 mM NaCl. This is slightly later than the shown in Table 1, were identical within experimental error elution ofnative venom CTX (330 mM NaCl), the N terminus and were as expected from the sequence ofthe synthetic CTX ofwhich is blocked by pyroglutamate. Peptide sequencing of gene. The 280-nm extinction coefficients ofthe native and the the recombinant material confirmed that the first five resi- recombinant CTX were essentially identical (9000 and 9200 dues were, as expected, Gln-Phe-Thr-Asn-Val. M-1 cm-l, respectively). To characterize the recombinant With its six cysteines in reduced form, CTX is inactive as CTX further, several cycles ofpeptide sequencing were done a channel-blocker (24), and so it is necessary to form the three from its N-terminal end. Without deblocking by pyrogluta- disulfide bonds properly after cleavage of the reduced CTX mate aminopeptidase, no significant sequencing reaction peptide. We have found that inclusion of a low concentration occurred. However, after denaturation, carboxamidometh- (0.3-0.5 mM) of 2-mercaptoethanol in the factor Xa cleavage ylation, and cleavage of the N-terminal pyroglutamate, the reaction mixture led to complete disulfide formation during CTX could be sequenced from the second amino acid resi- this reaction (data not shown). As we demonstrate below, the due, phenylalanine. correct disulfide bonds, yielding fully active material, are Functional Activity of Recombinant CTX. The activity of formed under these conditions. CTX was examined at high resolution by the block of single To cyclize the N terminus of recombinant CTX to pyro- Ca2+-activated K+ channels (2, 23). In this assay, shown in glutamate, we treated the purified toxin with 5% acetic acid Fig. 4A, single Ca2+-activated K+ channels were reconsti- for 2 hr at 650C. The N-terminal glutamine was half-converted tuted into planar lipid bilayer membranes, and discrete block- A B - 1.0 M NaCI | - 100 % B FIG. 3. Chromatographic profiles of recombinant CTX. (A) Mono-S FPLC profile of N-terminal uncyclized (upper curve) and cyclized r 50 CTX (lower curve). The N-terminal glutamine of recombinant CTX was cyclized in 5% acetic acid for 2 hr at -L 65°C. (B) Reverse-phase HPLC profiles of recombinant CTX. The N-terminal cyclized CTX was injected on a C18 column equilibrated with 0.1% trifluoroacetic acid and eluted with a linear gradient of 0-30% (vol/vol) ac- 0 10 20 30 40 O 10 20 30 40 etonitrile with 0.1% trifluoroacetic Time (min) Time (min) acid over 40 min at 1 ml/min. Downloaded by guest on September 29, 2021 Biochemistry: Park et al. Proc. Nati. Acad. Sci. USA 88 (1991) 2049 Table 1. Amino acid composition of native and A recombinant CTX Nafiv Venm C1X, 2SnM Recombinant Expected from

- Amino acid Native CTX CTX sequence r~~~~I- F'! _- _ Asp + Asn 2.8 2.9 3 Glu + Gln 2.9 3.0 3 Ser 5.0 5.1 5 Reobinant CIX (Cydiae), 2amM Gly 1.5 1.0 1 L I His 0.5 0.5 1 I1 r: I r Ir _ Arg 3.0 3.3 3 £ Thr 3.9 4.4 4 E Ala 0.3 0.1 0 £: 1 Pro 0.2 <0.1 0 cbn-tCIX (U_1ic ), 7snm Tyr 0.9 0.9 1 u - tF---~-B --- -ff Val 1.9 2.0 2 Met 0.8 0.7 1 Cys ND ND 6 TMC - l - o - P lie 0.2 <0.1 0 Leu 1.1 1.1 1 Phe 1.0 1.0 1 Lys 3.4 3.6 4 0 v. CTX B C Trp ND ND 1 * r. CTX(cycl.) a r. CTX(uncycl.) Total 37 1.0001 ND, not determined. ing events, due to binding of single molecules of CTX to the channel, were observed electrically. Statistical distributions D 0.100.

of channel dwell-times in the blocked and unblocked states 0 may be calculated from many such blocking events, and toxin association and dissociation rates may be readily measured (1, 2, 23). The raw data traces of Fig. 4A illustrate these CTX-induced blocking events with preparations of toxin 0.010 purified from scorpion venom, fully processed recombinant 0.005 0.005 1 0 5 10 15 20 0 10 20 30 40 toxin, and uncyclized recombinant toxin at 25 nM, 25 nM, Time(sec) Time(sec) and 75 nM, respectively, in the external medium. It is clear that fully processed recombinant toxin behaved similarly to FIG. 4. Channel blocking characteristics of native and recombi- native venom-purified CTX. The uncyclized peptide was a nant CTX. Discrete blocking events on single Ca2+-activated K+ much poorer inhibitor, for two reasons. (i) The blocks channels by native and recombinant CTX (N-terminal cyclized and induced by uncyclized recombinant CTX were less frequent uncyclized) were observed in planar lipid bilayers (A). The cumu- than with the other two preparations of CTX. (ii) The lative probability distribution of unblocked events (B) and blocked blocking events were approximately 5-fold shorter-lived with events (C) are shown. Exponential lifetime of blocking events is as the uncyclized material. follows: 15.5 sec for native CTX, 16.4 sec for cyclized recombinant These qualitative conclusions are supported by a quanti- CTX, 3.1 sec for uncyclized recombinant CTX. Exponential lifetime ofunblocking events is as follows: 6.1 sec for native CTX, 6.2 sec for tative analysis of the dwell-time distributions for unblocked cyclized recombinant CTX, 9.5 sec for uncyclized recombinant and blocked states (Fig. 4 B and C). All these distributions CTX. For each experiment 35 nM CTX was added to the external were monoexponential, as demanded for block by a homo- side of channel and 120 blocks were observed. Holding voltage was geneous preparation of toxin. Both on-rates and off-rates of +35 mV. v, Native; r, recombinant. native and recombinant CTX were identical (Table 2), with a dissociation constant of 12-15 nM. The uncyclized CTX indistinguishable. Due to difficulties in expressing the CTX preparation displayed much weaker channel-blocking activ- gene directly in bacteria, we adopted a fusion protein strat- ity, with an apparent dissociation constant of 110 nM. The egy. In contrast to the high-level expression of the CTX similarity in detailed function ofrecombinant and native CTX fusion protein reported here, the original construct of CTX strongly implies the identity in structure of these peptides. containing an ompA secretion sequence on its N-terminal end showed no detectable protein expression despite adequate DISCUSSION levels of mRNA (data not shown). We do not know whether the poor expression of this construct was caused by toxicity We designed and expressed a synthetic gene for CTX based of CTX to host cells, by possible nonsense mutations as on its primary amino acid sequence. The toxin protein was overexpressed as a cleavable fusion protein with the gene 9 Table 2. Kinetic parameters for block by native and product of phage T7, liberated by a sequence-specific pro- recombinant CTX tease, and further processed to the functionally active form of the protein. The recombinant CTX was similar to native Native Recombinant CTX chemically and functionally. To verify its chemical CTX CTX identity, we performed amino acid analysis, partial amino Dissociation constant, nM 14.0 ± 0.2 14.2 ± 0.1 acid sequencing, and enzymatic digestion of N-terminal On rate (k0d), M-1sec-1 (x106) 4.6 ± 0.3 4.5 ± 0.2 pyroglutamate on both native and recombinant toxins. The Off rate (kRff), sec-1 0.067 ± 0.002 0.064 ± 0.003 channel blocking activities of both toxins were measured on In each experiment, holding voltage was +35 mV and channel single channels in planar bilayers. All ofthe results, including open probability was adjusted between 0.3 and 0.5. Four experiments HPLC profiles, show that recombinant and native CTX are were performed for both toxins and data are mean ± SEM. Downloaded by guest on September 29, 2021 2050 Biochemistry: Park et al. Proc. Natl. Acad. Sci. USA 88 (1991)

observed in other systems (22), or by other unknown factors. We are grateful to Dr. Dan Oprian for help and advice throughout Aside from stable high-level expression, the gene 9 fusion this work and to Dr. Kenneth Blumenthal for a kind gift ofthe fusion protein has two other advantages: the high solubility of the protein vector pSR9 and advice on its use. This work was supported fusion protein and the absence of cysteine residues in the by National Institutes of Health Grant GM-31768. fusion vector coding sequence. 1. Miller, C., Moczydlowski, E., Latorre, R. & Phillips, M. (1985) After release of CTX from the fusion protein, two addi- Nature (London) 313, 316-318. tional steps were required for functional activity: formation 2. Anderson, C., MacKinnon, R., Smith, C. & Miller, C. (1988) J. ofdisulfide bonds and cyclization ofthe N-terminal glutamine Gen. Physiol. 91, 317-333. to pyroglutamate. Under our preparation conditions (with a 3. Hermann, A. & Erxleben, C. (1987) J. Gen. Physiol. 90, 27-47. low concentration of mercaptoethanol present), oxidation of 4. Reinhart, P. H., Chung, S. & Levitan, I. B. (1989) Neuron 2, sulfhydryl groups occurred during the factor Xa digestion. It 1031-1041. may seem paradoxical that addition of a thiol should promote 5. Schneider, M. J., Rogowski, R. S., Krueger, B. K. & disulfide formation; however, dissolved 02 will rapidly oxi- Blaustein, M. P. (1989) FEBS Lett. 250, 433-436. dize the low amount of can 6. Guggino, S. E., Guggino, W. B., Green, N. 4 Saktor, B. (1987) mercaptoethanol, which then Am. J. Physiol. 252, C128-C137. efficiently react with protein to form disulfide bonds (25). 7. Wolff, D., Cecchi, X., Spalvins, A. & Canessa, M. (1988) J. After at least 24 hr of incubation in this condition, no free Membr. Biol. 106, 243-252. sulfhydryl group was detected. Without N-terminal cycliza- 8. Castle, N. A. & Strong, P. N. (1986) FEBS Lett. 209, 117-121. tion, the oxidized CTX showed poor affinity for the Ca2+- 9. MacKinnon, R., Reinhart, P. H. & White, M. M. (1988) Neu- activated K+ channel (apparent Kd = 110 nM). Since the ron 1, 997-1001. NMR-determined structure of the peptide shows the N-ter- 10. Sands, S. B., Lewis, R. S. & Cahalan, M. D. (1989) J. Gen. minal pyroglutamate to be lying upon the hydrophobic face of Physiol. 93, 1061-1074. the molecule, positive charge at a free N terminus might 11. Schweitz, H., Stansfeld, C. E., Bidard, J.-N., Fagni, L., Maes, plausibly disrupt a hydrophobic interaction between the toxin P. & Lazdunskj, M. (1989) FEBS Lett. 250, 519-522. 12. Stuhmer, W., Ruppersberg, J. P., Schroter, K. H., Sakmann, molecule and the mouth of the channel (20). More detailed B., Stocker, M., Giese, K. P., Perschke, A., Baumann, A. & information on the solution structure of recombinant CTX Pongs, 0. (1989) EMBO J. 8, 3235-3244. should be obtained from an NMR study. Preliminary one- 13. MacKinnon, R. & Miller, C. (1988) J. Gen. Physiol. 91, dimensional NMR spectra show virtually identical patterns 335-349. for recombinant CTX and native toxin (data not shown). 14. Miller, C. (1988) Neuron 1, 1003-1006. The high-level expression of functionally active CTX, a 15. MacKinnon, R. & Miller, C. (1989) Biochemistry 28, 8087- K+-channel blocker, gives us not only a biochemical tool to 8092. study the interaction between CTX and K+-channels but also 16. MacKinnon, R. & Miller, C. (1989) Science 245, 1382-1385. an of the of the cleavable 17. MacKinnon, R., Heginbotham, L. & Abramson, T. (1990) example utility fusion protein Neuron 5, 767-771. strategy for expressing small proteins of less than 50 %imino 18. MacKinnon, R. & Yellen, G. (1990) Science 250, 276-279. acid residues. In addition, because of the experimenter's 19. Gimenez-Gallego, G., Navia, M. A., Reuben, J. P., Katz, ability to control many variables, including oxidation and G. M., Kaczorowski, G. J. & Garcia, M. L. (1988) Proc. Natl. N-terminal cyclization, this could be a useful model system Acad. Sci. USA 85, 3329-3333. for studying the folding process in small proteins. In the past, 20. Massefski, W., Redfield, A., Hare, D. R. & Miller, C. (1990) we have not been able to investigate in molecular detail the Science 249, 521-524. interaction between CTX and K+ channels because of the 21. Studier, F. W. & Moffatt, B. A. (1986) J. Mol. Biol. 189, limited amounts of peptide available from scorpion venom 113-130. 22. Howell, M. L. & Blumenthal, K. M. (1989) J. Biol. Chem. 264, and difficulties in obtaining clean reactions of protein- 15268-15273. modifying reagents with this peptide. By using our expression 23. Oprian, D. D., Molday, R. S., Kaufman, R. J. & Khorana, system in combination with site-directed alterations in the H. G. (1987) Proc. Nat!. Acad. Sci. USA 84, 8874-8878. peptide sequence, the molecular basis ofCTX recognition by 24. Smith, C. & Miller, C. (1986) J. Biol. Chem. 261, 14607-14613. and inhibition of the Ca2+-activated K+ channel may be 25. Scopes, R. K. (1987) Protein Purification-Principles and attacked. Practice (Springer, New York), 2nd Ed., pp. 248-249. Downloaded by guest on September 29, 2021