Synthesis of a 19-Residue Peptide with Alamethicin-Like Activity (Solid-Phase Peptide Synthesis/Lipid Bilayer Membranes/Voltage-Dependent Conductances) BALTHASAR F

Proc. Nati. Acad. Sci. USA Vol. 74, No. 1, pp. 115-119, January 1977 Biochemistry

Synthesis of a 19-residue peptide with alamethicin-like activity (solid-phase peptide synthesis/lipid bilayer membranes/voltage-dependent conductances) BALTHASAR F. GISIN*, SHIGERU KOBAYASHI*, AND JAMES E. HALO * Department of Biochemistry, The Rockefeller University, New York, N.Y. 10021; and t Department of Physiology and Pharmacology, Duke University Medical Center, Durham, North Carolina 27710 Communicated by R. B. Merrifield, October 18,1976

ABSTRACT This paper describes the chemical synthesis ALA-o, on the other hand, the same side group is engaged in of a compound with voltage-gating characteristics similar to an amide bond with phenylalaninol and the imino function of those observed in nerve membranes. For alamethicin (ALA), a Pro2 is blocked by an Ac-Aib residue. Gln'9 again is identical natural antibiotic that induces such properties in lipid bilayer membranes, there are two proposed structures, one a cyclic and in both sequences and provides the a-carboxyl group that makes the other an open chain peptide. The open chain sequence ALA a monobasic acid. (ALA-o) proposed by Martin and Williams [(1976) Biochem. J. In view of the significance ALA has gained as an agent that 153, 181-1901 was synthesized by stepwise solid-phase con- imparts nerve-like electrical properties upon artificial mem- densation of four fragments prepared by solid-phase synthesis. branes, its detailed molecular mechanism of action is of con- The product, purified to homogeneity, was not identical with siderable interest. However, the same mechanism would almost the main component of natural ALA. Nevertheless, in lipid bilayer membranes the exponential dependence of conductance certainly not be compatible with both of the proposed primary on voltage and the dependence of conductance on a high power structures. Chemical synthesis seemed to be the appropriate of the peptide concentration were qualitatively similar for approach to decide which structure, if either, exhibits electrical ALA-o and for natural ALA. Like ALA, ALA-o showed the activity. characteristics of a channel-former, although the single-channel The available evidence for and against ALA-c and ALA-o conductances were less well defined for the synthetic com- indicated the latter more probably ALA pound. This work establishes that a cyclic structure is not a represented for two necessary condition for a peptide to induce voltage-dependent reasons. First, in partial hydrolysates of ALA, a peptide com- conductances in membranes and that ALA-o possesses all the prising -Glu(y-Pro)-, the cycle-forming link, was not found structural elements required for such an activity. (19). Second, the presence of blocking groups on the Glul' y-carboxyl (phenylalaninol) and the Pro2 imino group (Ac-Aib) Alamethicin (ALA), a peptide with antibiotic activity against was demonstrated (21, 22, 17), thus precluding a direct covalent certain bacteria, is found in the culture broth of Trichoderma link between these two residues. On these grounds, the synthesis viride (2, 3). Because of its amphiphilic character, ALA shows of ALA-o rather than ALA-c was undertaken first. This effort pronounced surface activity and adsorbs strongly to biological and the evaluation of the product is the subject of the current and artificial membranes. As a consequence, it causes lysis of preliminary contribution (a detailed report for publication human erythrocytes and damages the outer membrane of elsewhere is in preparation). Ehrlich ascites tumor cells (4). Moreover, ALA induces aggre- gation and fusion of lecithin vesicles (5). Synthesis The peptide also affects the electrical properties of planar The synthesis of ALA-o was performed by the solid-phase artificial lipid membranes in a unique and specific way: minute method (23). A fragment condensation approach rather than concentrations of ALA induce voltage-dependent conductances a stepwise synthesis was chosen for the following reasons. First, that mimic the action potential phenomena observed in nerve due to the sterically hindered amino group of Aib, some of the membranes. These effects were observed by Mueller and Rudin coupling reactions could have failed to go to completion. Had (6) and have been studied also by other groups (7-9). It is now that occurred at each of the eight Aib residues in a stepwise generally believed that several ALA molecules aggregate in the approach, one would have been confronted in the end with the membrane to form ion-conducting channels (8, 10-12). difficult task of separating the target peptide from a large Despite a number of physicochemical (13, 14), conforma- number of deletion peptides similar in size and other properties. tional (15-18), and chemical (17-22) studies, the primary Fragments, on the other hand, could be purified individually structure of ALA remains unknown. Currently, there are two before their stepwise assembly on the resin. In this way, ho- proposed sequences: a cyclic (ALA-c) and an open chain mogeneity of the sequences covered by each fragment could (ALA-o) peptide structure (Fig. 1). The former, proposed by be assured and the purification of the crude final product would Payne et al. (19), has been confirmed by Ovchinnikov et al. be greatly simplified. Second, the uncertainty of the exact (20), while the latter, proposed by Martin and Williams (21, 22), structures of the N-terminal and COOH-terminal portions of found the support of the Jung group (17). The two proposals the peptide demanded a strategy that would allow the prepa- agree on a sequence of 15 consecutive residues (3-17 according ration of additional sequences, including ALA-c, without re- to ALA-o numbering). In ALA-c the y-carboxyl group of Glu'8 quiring resynthesis of the whole molecule. And third, analogues is in amide linkage with Pro2 to form a cyclic structure. In designed to furnish information on the mode of action of ALA could be prepared efficiently by combining the already avail- Abbreviations: (cf. ref. 1) ALA, alamethicin; ALA-o and ALA-c, open able fragments with new fragments and cyclic ALA, respectively; Aib, a-aminoisobutyric acid; Boc-, t- containing, e.g., single butoxycarbonyl; DCC, dicyclohexylcarbodiimide; Ac, acetyl; HOBT, amino acid substitutions. N-hydroxybenzotriazole; Bzl, benzyl; PE, phosphatidylethanolamine; Based on these considerations, four protected fragments of GMO, glycerol monooleate; Phol, L-phenylalaninol (phenylalanine ALA-o were synthesized (Figs. 1 and 2): the tetrapeptide 1-4, with -CH20H instead of -COOH); Res (resin), CH2-polystyrene-co- Ac-Aib-Pro-Aib-Ala-O-Res (A), the tetrapeptide 5-8, Boc- 1%-divinylbenzene. Aib-Ala-Gln-Aib-O-Res (B), the nonapeptide 9-17, Boc-Val- 115 Downloaded by guest on September 23, 2021 116 Biochemistry: Gisin et al. Proc. Natl. Acad. Sc. USA 74 (1977)

(A) PRo-AIB-ALA-AIB-ALA-GLN-AIB-VAL-AIB-GLY-LEu-AIB-PRO-VAL-AIB-AIB-GLU-GLN-OH (2) (3) (4) (17) (18) (19)

PHOL (B) AC-AIB-PRO-AIB-ALA-AIB-ALA-GLN-AIB-VAL-AIB-GLY-LEU-AIB-PRO-VAL-AIB-AIB-GLU-GLN-OH 1 2 3 4 5 8 9 17 18 19

(C) i j I E I FIG. 1. Proposed sequences for alamethicin. Aib is a-aminoisobutyric acid; Phol, L-phenylalaninol. (A) the cyclic structure (ALA-c) by Payne et al. (19) and (B) the open chain structure (ALA-o) by Martin and Williams (21, 22). (C) Fragments used in the synthesis of ALA-o. Aib-Gly-Leu-Aib-Pro-Val-Aib-Aib-O-Res (C), and the dipep- gel permeation chromatography (Ac-A-B-C-OH, ALA-o) using tide Boc-Glu(Phol)-Gln-O-Res (D). These fragments were a Sephadex G-25 column (1.6 X 90 cm) in ethanol/water (1:1, prepared starting with chloromethylated (23, 24) polysty- vol/vol) in connection with a differential refractometer. The rene-co-1%-divinylbenzene resin (Res) to which the Boc-pro- peaks superimposed with the trace obtained from quantitative tected COOH-terminal amino acid was esterified with the radioactivity analysis of the fractions. Homogeneity of each cesium salt procedure (25). Trifluoroacetic acid (50%) in fragment and ALA-o was established by thin-layer chroma- methylene chloride was used for deprotection and diisopro- tography on silica gel G plates. One single, ninhydrin-negative pylethylamine (5%) in the same solvent was used for neutral- iodine/tolidine-positive spot was observed in each case. ization. Coupling was with 1.5 equivalents each of Boc-amino Condensation of the fragments (Fig. 3) was effected by acid and dicyclohexylcarbodiimide (DCC) (26) in methylene preactivating the carboxyl component with DCC and HOBT chloride for four hours followed by another 16-hr period of (1:1.5:3.0, eq/eq/eq) in methylene chloride/N,N-dimethyl- coupling with fresh reagents. This protocol was normally suf- formamide at 0-0 for 1 hr and 1 hr at room temperature prior ficient to give coupling yields greater than 99% as determined to addition to the resin-bound amine component in dimethyl- with picrate (27). However, either when the carboxyl compo- formamide. The reaction was allowed to proceed at room nent was Boc-Aib or when the amine component ended in an temperature and was monitored by measuring the radioactivity Aib residue, longer reaction times were necessary. In these cases, of the solution phase. The molar equivalents used, coupling coupling was repeated until the picrate test indicated a yield greater than 98%. In order to suppress nitrile formation the glutamine residue of fragment B was coupled in the presence AIB PRO AIB ALA of N-hydroxybenzotriazole (HOBT) (28). Model experiments using infrared spectroscopy showed indeed that under the BOC-! O-RES conditions used this side reaction did not occur. The low chemical reactivity of Aib is also reflected in the low BOC +-OH H + O-RES color value it yields when run on an amino acid analyzer under standard conditions (ref, 18, 19; B. F. Gisin, unpublished data). BOC -OH H 1- - O-RES This drawback was overcome by using tritium-labeled Aib* throughout the present work. Consequently, radioisotopic BOC - OH H - O-RES analysis could be used to determine Aib in hydrolysates, in in- tact peptides, or in some cases in peptide-resins. The accuracy BOC O-RES was comparable to that obtained from an amino acid analyzer for the other amino acids. The radioactive label was also very H n_^,- useful in the tracing of chromatograms. Thus, the ninhydrin- negative N-acetylated peptides were readily distinguished from AC20 the unlabeled reagents and side-products on thin-layer chromatograms as well as in the effluent from columns. Table 1 shows the amino acid composition of hydrolysates Ac-AiB-PRO-AIB-ALA-O-RES of peptide-resins and of the cleaved and purified fragments. Cleavage from the resin was with HF (30) for 20-60 min at 00. In each instance, this procedure gave rise to some ninhydrin- | HF positive side products. They were removed by treatment of the crude material in aqueous (pH 4-6) solution with Dowex 50W X 2 (SOaH-form). Because Ac-A-B-C-OH and ALA-o were not AC-AIB-PRO-AIB-ALA-OH sufficiently soluble in water, a mixture of ethanol and water (1:1, vol/vol) was used in these cases. The peptides were puri- Ac- -OH fied further by crystallization (Ac-A-OH, Ac-A-B-OH) or by FIG. 2. Scheme of the synthesis of fragment A. The fragments Boc-B-O-Res (Boc-Aib-Ala-Gln-Aib-O-Res), Boc-C-O-Res (Boc- t a-Amino [methyl-3H]isobutyric acid was purchased from ICN Iso- Val-Aib-Gly-Leu-Aib-Pro-Val-Aib-Aib-O-Res), and Boc-D-O-Res tope & Nuclear Division, Irvine, Calif., and diluted with unlabeled [Boc-Glu(Phol)-Gln-0-Res] were prepared in the same stepwise Aib to approximately 4000 cpm/,umol. manner as shown for Ac-A-O-Res. Downloaded by guest on September 23, 2021 Biochemistry: Gisin et al. Proc. Natl. Acad. Sci. USA 74 (1977) 117 Table 1. Amino acid ratiosa of synthetic fragments, ALA-o, and natural alamethicin Compound Glu Pro Gly Ala Val Leu Aib Phol b

Ac-A-O-Res 1.01 (1) 1.00 (1) 1.99 (2) Boc-B-O-Res 1.05 (1) 1.07 (1) 1.88 (2) Boc-C-O-Res 1.09 (1) 1.03 (1) 1.72 (2) 1.05 (1) 4.09 (4) Ac-A-OH 1.01 (1) 1.05 (1) 1.94 (2) Ac-A-B-OH 0.97 (1) 1.03 (1) 1.98 (2) 4.02 (4) Ac-A-B-C-OH 0.93 (1) 2.08 (2) 1.09 (1) 1.96 (2) 1.96 (2) 1.06 (1) 7.91 (8) Synthetic ALA-o 2.98 (3) 2.06 (2) 0.94 (1) 1.87 (2) 2.03 (2) 1.06 (1) 8.10 (8)C 0.96 (1) Natural ALAd 2.98 (3) 2.12 (2) 0.94 (1) 1.68 (2) 2.13 (2) 1.02 (1) 8.15 (8)c 0.98 (1) Numbers in parentheses are the integral values. a Samples were hydrolyzed with concentrated HC/propionic acid 1:1 (vol/vol) at 132 + 2° for 4-8 hr (cf. ref. 29). Aib was determined by radioisotope analysis of the hydrolysates unless noted otherwise. The other amino acids were determined on an amino acid analyzer. b Determined on amino acid analyzer using pH 10.0 buffer. 'Determined on amino acid analyzer. d U-22324 8831-CEM-93.3, The Upjohn Co., Kalamazoo, Mich.

times, and yields of the cleaved purified products based on the methods were as described by Eisenberg et al. (7) and Hall limiting reactant were as follows: 1.0 Ac-A-OH + 1.5 H-B- (10). O-Res, 40 hr, 71%; 1.0 Ac-A-B-OH + 1.3 H-C-O-Res, 17 hr, The conductance-voltage curve of ALA-o shows the expo- 33%; 1.0 Ac-A-B-C-OH + 11.0 H-D-O-Res, 17 ht- 69%. nential dependence of conductance on voltage characteristic for natural ALA (Fig. 4). In membranes formed from solutions Characterization of product of bacterial phosphatidylethanolamine(PE)(SupelcoBellefonte, Synthetic ALA-o (Ac-A-B-C-D-OH) was found to consist of one Pa.), and glycerol monooleate (GMO) (Sigma, St. Louis, Mo.), component by chromatography on Sephadex G-25 in etha- in ratios from 4:1 to 1:1 (by weight), a change of 5.7 h 0.5 mV nol/water (1:1, vol/vol) and on Sephadex LH-20 in methanol. induces an e-fold change in conductance for ALA-o. Under the Thin-layer chromatography in six different solvent systems same conditions, 4.7 + 0.3 mV induces an e-fold conductance showed one spot, staining yellow-white in the iodine/tolidine for natural ALA. Thus, the natural ALA curve is somewhat reaction (1-butanol/acetic acid/pyridine/water, 15:10:3:2, RF steeper than that of ALA-o. = 0.56; 1-butanol/acetic acid/water/ethyl acetate, 1:1:1:1, RF The conductance-voltage curves of Fig. 4 are taken at ap- = 0.77; ethyl acetate/pyridine/water/acetic acid, 30:10:5:3, proximately the same voltage range for both substances, but the RF = 0.28; chloroform/methanol/water, 65:24:4, RF = 0.43; concentration of natural ALA is 0.32 jag/ml and that of ALA-o 1-propanol/water, 7:3, RF = 0.60; 1-butanol/acetic acid/water, is 6.4 jsg/ml. ALA-o must therefore be present in a 20-fold 4:1:1, RF = 0.56). Amino acid ratios (Table 1) were in agree- higher concentration than natural ALA to induce the same ment with the expected values. conductance at the same voltage. This result led us to consider Thin-layer chromatography showed readily that synthetic the possibility that the active principle in the synthetic material ALA-o was not identical with the main component of natural might be a 5% contaminant with the structure of natural ALA. ALA.§ In all systems where a different RF was observed the For that reason, our ability to detect natural alamethicin de- natural product migrated faster, indicating that the synthetic liberately introduced into the purified synthetic material was product is more polar. Titration of the two compounds (3 mM tested. It was found that thin-layer chromatography could in 30% ethanol/water) revealed that they differed markedly detect less than 1% of natural ALA mixed with ALA-o and we also in the pK of their acidic group. Thus, for natural ALA an apparent pKa of 5.2 was measured, while the corresponding + value for synthetic ALA-o was 4.4. Under the same conditions Ac- AJ-OH H-Jj -O-RES the acidities of the model compounds Boc-Glu(OH)OBzl and 1. DCC-HOBT Boc-Glu(OBzl)-OH were determined. The former ('y-carboxyl 2. HF group) had a pKa' of 5.1 and the latter (a-carboxyl group) a PKa' of 4.4. It appears, therefore, that in natural ALA the carboxyl group may be in the -y- rather that in the a-position that was Ac-I A - B A-BH+ H- -O-RES proposed in previous structural work (19-22). Whether this is the only difference between natural ALA and synthetic ALA-o 1 DCC-HOBT I may be resolved as more COOH-terminal variants become available. 2. HF Activity of synthetic ALA-o in black lipid films Ac-[ A - B - C I-OH + H- X-O-RES We report here a preliminary characterization of the electrical activity of ALA-o in black lipid films. The data were obtained 1. DCC-HOBT in planar lipid films formed in 1 M KC1 solutions using the 2. HF technique described by Montal and Mueller (31). Additional data, to be reported elsewhere, were also obtained in films - - - formed by the technique described by Mueller et al. (32). The Ac- A B C Ii-OH - § The sample of natural ALA used was obtained from G. B. Whitfield ALA o of Upjohn Co. In addition to the main spot, it contained two minor FIG. 3. Scheme of the solid-phase condensation of four fragments components by thin-layer chromatography. to give ALA-o. See text for experimental details. Downloaded by guest on September 23, 2021 118 Biochemistry: Gisin et al. Proc. Natl. Acad. Sci. USA 74 (1977)

-~10-7

10- 1 30 40 50 Voltage (mV) FIG. 4. Conductance-voltage curves for ALA-o and natural ALA Concentraton (g/ri) (1 mho = 1 siemens). The conductance increases exponentially with voltage for both ALA (e) and ALA-o (A). The membranes (compo- FIG. 5. Voltage to attain a conductance of 10 nanomhos as a sition PE:GMO, 2:1 by weight) were formed by the Montal-Mueller function of ALA-o and ALA concentrations. The ALA-o curve is technique (31). Solutions were 1 M KCl. ALA concentration was 0.32 parallel to the ALA curve and displaced from it by 100 mV on the jsg/ml. ALA-o concentration was, 6.4 jig/ml. The same slopes are ob- voltage axis or a factor of 20 on the concentration axis. the concen- served for a range of PE:GMO from 4:1 to 1:1. The conductance scale tration scale is logarithmic and the voltage scale, linear. Membrane is logarithmic and the voltage scale, linear. composition and salt are the same as for Fig. 4. concluded that the activity of the synthetic sample was not due Fig. 5 shows the voltage at which a conductance of 10 na- to a contaminant identical with natural ALA. It is very probable nomhos is achieved for various concentrations of ALA-o and that the mechanism of conductance requires the molecule to natural ALA. The curves are parallel and displaced from one enter the hydrophobic interior of the membrane, an event more another along the concentration axis by a factor of 20 or along likely for nonpolar than for polar molecules. The reduced ac- the voltage axis by 100 mV. In both cases an e-fold increase in tivity of ALA-o is therefore consistent with the greater polarity concentration decreases the voltage necessary to give a con- that was observed when it was compared with natural ALA by ductance of 10 nanomhos by about 34 mV. Combining this thin-layer chromatography. datum with the steepness of the conductance-voltage curve

...... I~~~~~~~~; F 85mV 0.3 MKC - ALA 210e~

I- 3

Current nAI

4 k.- T r T"' T --T -- - -- 99 mrV a) M KCi ALA- o

0.6

6 3 0 0.5 10 Current {nA) FIG. 6. (a) Single channel characteristics of natural alamethicin. The histogram shows the number of times a given current occurred during a sampling period of 32 sec at a rate of 10,000 samples per second. The oscilloscope photograph shows several sweeps taken during the generation of the histogram. Note that the levels in the photograph are well defined and sharp and that the peaks in the histogram are narrow and well defined. Salt and membrane are as in Fig. 4; voltage is clamped at 85 mV. Oscillograph scales: current, 0.10 nA/division; time, 5 msec/division. Figures over peaks are conductances in nanomhos. ALA concentration, 40 ng/ml. (b) Single channel characteristics of ALA-o. Sampling rate and period as in (a). Note that the current trace in the oscilloscope photograph is much noisier than that for ALA and that the levels and transitions are poorly defined. In the histogram note that the current has almost as great a probability of being between the poorly defined peaks as on a peak. Peaks, such as they are, are much broader than for ALA. Voltage is 99 mV. Oscillograph scales: current 0.10 nA/division; time, 5 msec/division. Figures and fiducial marks over peaks indicate the corresponding conductances in nanomhos. ALA-o concentration, 800 ng/ml. Downloaded by guest on September 23, 2021 Biochemistry: Gisin et al. Proc. Natl. Acad. Sci. USA 74 (1977) 119

allows computation of the dependence of the conductance of This work was supported by Grants HL-17961 and, in part, HL- a constant voltage on the concentration (7). For natural alam- 12157 from the U.S. Public Health Service. ethicin the dependence is on the 7 ± 0.7 power of the concentration and for ALA-o it is on the 6 ± 0.6 power. The power 1. IUPAC-IUB Commission on Biochemical Nomenclature (1971) dependence is a function of the membrane type and composi- J. Biol. Chem. 247,977-983. tion and these results are for PE/GMO 2:1 membranes of the 2. Meyer, C. E. & Reusser, F. (1967) Experientia 23, 85-86. type described by Montal and Mueller (31). Natural alamethicin 3. Reusser, F. (1967) J. Biol. Chem. 242, 243-247. and ALA-o thus exhibit almost the same power dependence in 4. Jung, G., Dubischar, N., Leibfritz, D., Ottnad, M., Probst, H. & PE/GMO 2:1 membranes. Stumpf, Ch. (1975) in Peptides 1974, ed. Wolman, Y. (John Like ALA, ALA-o showed current fluctuations characteristic Wiley & Sons, New York), pp. 345-354. for a channel former. However, the channel character- 5. Lau, A. L. Y. & Chan, S. (1974) Biochemistry 13,4942-4948. single 6. Mueller, P. & Rudin, D. 0. (1968) Nature 217,713-719. istics were considerably different for the two compounds. While 7. Eisenberg, M., Hall, J. E. & Mead, C. A. (1973) J. Membr. Biol. natural ALA showed sharp, well-defined transitions between 14, 143-176. various levels (7, 8, 10, 12), ALA-o exhibited fuzzy, poorly e- 8. Boheim, G. (1974) J. Membr. Biol. 19,277-303. fined levels. Fig. 6 gives a comparison of single channel dat_ 9. Boheim, G. & Hall, J. E. (1975) Biochim. Biophys. Acta 389, for natural ALA and ALA-o under comparable conditions. The ? 436-443. oscilloscope face photograph (Fig. 6a) shows current fluctua- 10: Hall, J. E. (1975) Biophys. J. 15, 934-939. tions for several successive sweeps during single channel activity 11. Bauman, G. & Mueller, P. (1974) J. Supramol. Struct. 2, 538- of natural ALA. A number of inter-level transitions appear, all 557. of which are sharp and well defined and last for less than a 12. Gordon, L. G. M. & Haydon, D. (1975) Phil. Trans. R. Soc. millisecond. By contrast, Fig. 6b shows a London Ser. B 270,433-447. similar photograph 13. Chapman, D., Cherry, R. J., Flint, E. G., Hauser, H., Phillips, M. taken of the current fluctuations due to ALA-o. Here the levels C., Shipley, C. G. & McMullen, A. I. (1969) Nature 224, 692- are much less well defined and transitions between levels are 694. much less sharp. Probability histograms, which show the relative 14. McMullen, A. I. & Stirrup, J. A. (1971) Biochim. Biophys. Acta likelihoods of various currents, are shown next to the oscilloscope 241,807-814. photographs for the corresponding substance (see refs. 7 or 10 15. Hauser, H., Finer, E. G. & Chapman, D. (1970) J. Mol. Biol. 53, for detailed descriptions of such histograms). Note that the 419-433. conductance average of the two histograms is about the same, 16. McMullen, A. I., Marlborough, D. I. & Bayley, P. M. (1971) FEBS but that the peaks in the natural ALA histogram are much more Lett. 16, 278-280. distinct and sharp than those in the ALA-o histogram. There 17. Jung, G., Dubischar, N. & Leibfritz, D. (1975) Eur. J. Biochem. is 54,395-409. thus a clear and graphic difference in the single-channel 18. Jung, G., K6nig, W. A., Leibfritz, D., Ooka, T., Janko, K. & Bo- characteristics exhibited by ALA-o and natural ALA, even heim, G. (1976) Biochim. Biophys. Acta 433, 164-181. though their macroscopic conductance properties are rather 19. Payne, J. W., Jakes, R. & Hartley, B. S. (1970) Biochem. J. 117, similar and differ only quantitatively. 757-766. Because of the differences in single-channel characteristics 20. Ovchinnikov, Yu. A., Kiryushkin, A. A. & Kozhevnikova, I. V. and because the single-channel patterns of ALA-o are not as (1971) Magy. Kem. Lapja 41, 2085-2099. reproducible as those of ALA, comparison of the magnitude of 21. Martin D. R. & Williams, R. J. P. (1975) Proc. 533rd Meeting, single-channel conductances for the two materials is difficult. Biochem. Soc. Trans. 3, 166-167. Nonetheless, they are very similar and are both on the order of 22. Martin, D. R. & Williams, R. J. P. (1976) Biochem. J. 153, a nanomho in 1 M KCI solutions. 181-190. 23. Merrifield, R. B. (1963) J. Am. Chem. Soc. 85,2149-2154. Conclusions 24. Gisin, B. F. & Merrifield, R. B. (1972) J. Am. Chem. Soc. 94, 6165-6170. The synthesis of an active peptide according to the open chain 25. Gisin, B. F. (1973) Hefv. Chim. Acta 56, 1476-1482. alamethicin structure of Martin and Williams (21, 22) estab- 26. Sheehan, J. C. & Hess, G. P. (1955) J. Am. Chem. Soc. 77, lishes that a cyclic structure (19) is not required for electrical 1067-1071. activity. Because of the observed similarities in their modes of 27. Gisin, B. F. (1972) Anal. Chim. Acta 58,248-249. action, it is unlikely that natural alamethicin differs radically 28. Konig, W. & Geiger, R. (1970) Chem. Ber. 103,788. from ALA-o in structure. Thus, this synthesis indicates that the 29. Scotchler, J., Lozier, R. & Robinson, A. B. (1970) J. Org. Chem. ALA-o structure possesses the basic components necessary to 35,315. induce a strongly voltage-dependent conductance. The tech- 30. Sakakibara, S. & Shimonishi, Y. (1965) Bull. Chem. Soc. Jpn. 38, niques developed in this synthesis will enable 1412. generation of a 31. Montal, M. & Mueller, P. (1972) Proc. Natl. Acad. Sci. USA 69, number of analogues, and differences between these materials 3561-3566. will materially aid attempts to understand the mechanism of 32. Mueller, P., Rudin, D. O., Tien, H. T. & Wescott, W. C. (1962) voltage-dependent conductance. Circulation 26, 1167-1170. Downloaded by guest on September 23, 2021