A Practical Solid‐Phase Synthesis of Glu7‐Phalloidin and Entry into Fluorescent F‐Actin‐Binding Reagents

Simple synthesis of a potent probe: An efficient solid-phase synthesis of Glu7-phalloidin (the resin-bound cyclic peptide shown) features the one-pot I2-mediated deprotection of S-tritylcysteine and formation of the Cys–Trp thioether bridge. A fluorescent conjugate of the resulting phalloidin analogue shows specific staining of F-actin (filamentous actin) in fixed mammalian epithelial cells (see fluorescence micrograph). Fmoc=9-fluorenylmethoxycarbonyl. Phalloidin (1) is a potent actin-binding toxin whose chemistry and bioactivity have been studied since the early 1900s.1 Phalloidin binds with high affinity to filamentous actin (F-actin) and lowers the critical concentration of actin polymerization in solution.2 It has been used extensively to study actin dynamics in vitro, and fluorescent analogues of phalloidin serve as highly specific reagents for microscopic visualization of the actin cytoskeleton.3 The natural source of phalloidin, Amanita phalloides, is a mushroom that lives in a complex ecological relationship with associated flora and is widely considered to be uncultivable.1 Thus, an efficient synthetic route to phalloidin would be highly desirable as an alternative source for this reagent and as an entry into related cyclic peptides. Although there have been a number of syntheses of phalloidin analogues both in solution4–7 and on a solid phase,8 no synthetic route has been published with yields significant enough to provide this reagent in practical quantities. The yields reported for these syntheses, which relied on the preparation of relatively complex building blocks in solution, ranged from 0.54 to 1.3 %.6, 8 Herein we report an efficient and practical solid-phase synthesis of Glu7-phalloidin (2) in 50 % overall yield from simple starting materials. Derivatization of the Glu7 side chain yielded a fluorescent analogue that stains F-actin in fixed cells at a concentration comparable to that of commercial phalloidin-based probes.1 Phalloidin is a bicyclic heptapeptide which contains an unusual bridging thioether linkage between the Cys and Trp residues. The natural product contains four common L-amino acid residues, a D-threonine residue, an unusual γ,δ-dihydroxy-L-leucine residue, and the rare cis epimer of 4-hydroxy-L-proline. Structure–activity studies have shown that the γ,δ-dihydroxy-L-leucine side chain is not essential for actin binding.7, 9 We therefore replaced this residue with glutamic acid to provide the derivative 2 with both a handle for linkage to the solid phase and a site for fluorophore attachment. The cis-4-hydroxy-L-proline residue was prepared according to published methods.8, 10 Central to any total synthesis of phalloidin and its derivatives is the formation of the transannular thioether bridge between the Cys and Trp residues. This transformation has been effected in previous syntheses by the treatment of a cysteine-derived sulfenyl chloride with a suitably protected tryptophan derivative,8, 11, 12 or through the attack of a cysteine sulfhydryl group on an oxidatively activated tryptophan indole.13, 14 Although these syntheses provided efficient access to the Trp–Cys thioether adduct in solution, there has been no report of the direct thionation of a Trp indole on the solid phase. Our approach to the formation of the critical thioether bridge was inspired by a side reaction reported to occur during solution-phase I2-mediated deprotection of S-tritylcysteine (Cys(Trt)) in peptides containing tryptophan.15 The minor product resulted from thioether formation between the Cys and Trp residues, which occurred presumably through the attack of the tryptophan indole on the sulfenyl iodide intermediate that forms upon trityl deprotection.16 By using peptides based on the sequence Cys(Trt)-Glyn-Trp, the same authors showed that I2 treatment in dilute solution led to thioether formation that was more efficient than disulfide dimerization for n>3. When we applied similar conditions to the solid-phase synthesis of model peptides based on the thioether-containing sequence of phalloidin, the only observed products were the desired thioether and the dimer resulting from on-resin intermolecular disulfide formation. For the sequence H2N-Cys-Pro-Ala-Trp-OH, at a loading value of 0.1 mmol g−1, cyclization was twice as efficient as dimerization (Scheme 1). When the L-proline residue was replaced with TIPS-protected cis-4-hydroxy-L-proline, the ratio of the thioether to the disulfide product increased to 6.6:1. These results encouraged us to undertake a solid-phase synthesis of Glu7-phalloidin with an I2-mediated cyclization as the key step to form the thioether bridge. Assessment of solid-phase I2-mediated cyclization as a strategy for the construction of the thioether bridge in phalloidin derivatives by using tetrapeptide model systems. The thioether/disulfide product ratios are 2:1 (for Fmoc-NH-Cys-Pro-Ala-Trp-OH) and 6.6:1 (for Fmoc-NH-Cys-Hyp(OTIPS)-Ala-Trp-OH). DMF=N,N-dimethylformamide, Fmoc=9-fluorenylmethoxycarbonyl, TFA=trifluoroacetic acid, TIPS=triisopropylsilyl. To generate the appropriate peptide precursor, the Glu7 residue was protected at its C terminus as an allyl ester and linked through its side chain to 2-chlorotrityl polystyrene resin. Elongation to give the heptapeptide was carried out by using standard Fmoc chemistry (Scheme 2), and after removal of the N-terminal Fmoc group17 and C-terminal allyl ester, the peptide backbone was cyclized by treatment with diphenylphosphorylazide (DPPA). Following cleavage of the product from the resin, HPLC–MS analysis showed that the macrolactamization had proceeded efficiently. No cyclodimer or higher oligomers were observed. Solid-phase peptide synthesis (SPPS) of Glu7-phalloidin. DBU=1,8-diazabicyclo[5.4.0]undec-7-ene, DIPEA=N,N′-diisopropylethylamine, NMM=N-methylmaleimide. When the resin-bound cyclic peptide 4 was treated with I2 in DMF, complete conversion into the thioether was observed, with no intermolecular disulfide dimer detected. Upon cleavage from the resin with 1 % trifluoroacetic acid/CH2Cl2, the crude, fully protected cyclic product was found to be more than 90 % pure by LC–MS; no higher oligomers were present. Removal of the protecting groups on the D-Thr and 4-hydroxy-L-proline side chains by sequential treatment with TFA/CH2Cl2 (1:1) and 50 % HF–pyridine in THF, and purification by reversed-phase HPLC, yielded a single major product whose circular dichroism (CD) and 1H NMR spectra8 were consistent with Glu7-phalloidin (see the Supporting Information for complete characterization). The overall yield of the purified material was 50 % on the basis of the initial resin loading. Synthetic phallotoxins can exist as two isolatable atropisomers.1 The synthetic route reported herein provides exclusive access to the natural atropisomer, as determined by comparison of the CD spectrum of Glu7-phalloidin with that of the authentic natural product. Furthermore, the distinctive upfield chemical shift of the Ala5 methyl group is diagnostic of the natural atropisomer because of the proximity of this group to the anisotropy field of the tryptophan indole ring.8 Tetramethylrhodamine cadaverine was conjugated to compound 2 by using PyBOP/HOBt in DMF to yield the fluorescent adduct 5 (Scheme 3). Cultured mammalian epithelial (BS-C-1) cells were fixed and treated with 5 at 1 μg mL−1 by following standard protocols (see the Supporting Information). Fluorescence microscopy revealed the F-actin staining pattern typical of commercial phalloidin conjugates (Figure 1). Actin-filament staining was completely abolished when the fixed cells were pretreated with natural phalloidin, thus demonstrating the specificity of 5 for F-actin. a) Fluorescence micrographs of a) BS-C-1 cells stained with the tetramethylrhodamine–phalloidin derivative 5 and b) cells pretreated with natural phalloidin (1) prior to the addition of 5; inset shows a phase-contrast image of the same field. Conversion of Glu7-phalloidin (2) into a rhodamine derivative. HOBt=1-hydroxy-1H-benzotriazole, PyBOP=benzotriazol-1-yloxytripyrrolidinophosphonium hexafluorophosphate. In summary, we have developed a simple synthesis of Glu7-phalloidin and its derivatives, including a fluorescent bioactive probe that is as effective as natural phalloidin conjugates in staining F-actin in fixed cells. This route will generate sufficient amounts of fluorescently labeled phalloidin to perform high-throughput image-based screens for compounds that affect actin morphology, and will allow us to make extensive modifications to the phalloidin scaffold for future structure–activity studies. Supporting information for this article is available on the WWW under http://www.wiley-vch.de/contents/jc_2002/2007/z700017_s.pdf or from the author. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.