Synthesis of α-Conotoxin SI, a Bicyclic Tridecapeptide Amide with Two Disulfide Bridges: Illustration of Novel Protection Schemes and Oxidation Strategies

Several routes are described and compared for the synthesis of α-conotoxin SI, a tridecapeptide amide of sequence H-Ile-Cys-Cys-Asn-Pro-Ala-Cys-Gly-Pro-Lys-Tyr-Ser-Cys-NH₂, with bicyclic disulfides connecting Cys² with Cys⁷, and Cys³ with Cys¹³. The linear sequence was assembled smoothly on tris(alkoxy)benzylamide (PAL) supports, using stepwise Fmoc solid-phase chemistry. Side-chain protection of Cys was provided by S-2,4,6-trimethoxybenzyl (Tmob) at all four positions or by suitable pairwise combinations of S-Tmob and S-acetamidomethyl (Acm). Acidolytic cleavage/deprotection of these peptide-resins with trifluoroacetic acid (TFA)-CH₂Cl₂-Et₃SiH-H₂O-anisole (95:4:0.5:0.5:0.5), at 25 °C for 2 h, gave the corresponding peptide amides in high yields (90–97%); those Cys residues originally blocked by S'-Tmob were converted to the free sulfhydryls, whereas Cys(Acm) residues remained blocked. The fully deprotected linear tetrasulfhydryl conotoxin was oxidized successfully in dilute pH 7.5 solution in the presence of 1% (v/v) DMSO, at 25 °C for 7 h, providing monomeric bicyclic peptide in an overall yield of 39%. Other simultaneous procedures gave poor yields and/or extensive oligomers and polymers, in part due to solubility problems. Further solution synthesis strategies relied on sequential disulfide pairing as dictated by the original protection scheme. A monocyclic, bis(Acm) intermediate, with the smaller loop (Cys² with Cys⁷) already closed by the solution DMSO oxidation procedure, was cyclized further with thallium tris(trifluoroacetate) [Tl(tfa)₃; 1.2 equiv] in TFA-anisole (19:1) at 4 °C for 18 h. The respective oxidation yields were 62 and 65%, and the final overall yield of monomeric conotoxin reflecting cleavage from the support and the two orthogonal oxidation steps was 38%. Alternatively, the same chemical steps and solution conditions were used to test the strategy of forming the larger loop (Cys³ with Cys¹³) first; individual oxidation steps each proceeded in ∼50% yield, and conotoxin was obtained in an overall yield of 23%. Simultaneous or orthogonal disulfide bond formation was also examined while the peptide remained anchored to the polymeric support. The most promising results started with the appropriate resin-bound precursor, which was treated with TFA-CH₂Cl₂-Et₃SiH-H₂O-anisole (7:92:0.5:0.5:0.5), 2×15 min at 25 °C, to remove selectively S-Tmob groups from Cys² and Cys⁷ without significant loss of chains from the support. The resultant resin-bound dithiol was oxidized with 35 mM CCl₄-Et₃N (2 equiv each) in N-methylpyrrolidinone (NMP) for 4 h at 20 °C, to provide a resin-bound monocyclic bis(Acm) intermediate, which was oxidized further with Tl(tfa)₃ (2 equiv) in DMF-anisole (19:1) for 18 h at 4 °C. After acidolytic cleavage, overall yields of monomeric conotoxin were as high as 14%. Additional oligomeric material is apparently incorporated as a network which is retained on the support, accounting for the reduced yields (25–50%, respectively with polystyrene (PS) and polyethylene glycol-polystyrene (PEG-PS) graft supports) observed after cleavage of the peptidyl-PAL anchoring linkage. The polymer-supported routes are amenable to further optimization and may offer several advantages over solution methods.