Total Syntheses of the Actin‐Binding Macrolides Latrunculin A, B, C, M, S and 16‐epi‐Latrunculin B

Abstract

The latrunculins are highly selective actin-binding marine natural products and as such play an important role as probe molecules for chemical biology. A short, concise and largely catalysis-based approach to this family of bioactive macrolides is presented. Specifically, the macrocyclic skeletons of the targets were forged by ring-closing alkyne metathesis (RCAM) or enyne-yne metathesis of suitable diyne or enyne-yne precursors, respectively. This transformation was best achieved with the aid of [(tBu)(Me(2)C(6)H(3))N](3)Mo (37) as precatalyst activated in situ with CH(2)Cl(2), as previously described. This catalyst system is strictly chemoselective for the triple bond and does not affect the olefinic sites of the substrates. Moreover, the molybdenum-based catalyst turned out to be broader in scope than the Schrock alkylidyne complex [(tBuO)(3)W[triple chemical bond]CCMe(3)] (38), which afforded cycloalkyne 35 in good yield but failed in closely related cases. The required metathesis precursors were assembled in a highly convergent fashion from three building blocks derived from acetoacetate, cysteine, and (+)-citronellene. The key fragment coupling can either be performed via a titanium aldol reaction or, preferentially, by a sequence involving a Horner-Wadsworth-Emmons olefination followed by a protonation/cyclization/diastereoselective hydration cascade. Iron-catalyzed C--C-bond formations were used to prepare the basic building blocks in an efficient manner. This synthesis blueprint gave access to latrunculin B (2), its naturally occurring 16-epimer 3, as well as the even more potent actin binder latrunculin A (1) in excellent overall yields. Because of the sensitivity of the 1,3-diene motif of the latter, however, the judicious choice of protecting groups and the proper phasing of their cleavage was decisive for the success of the total synthesis. Since latrunculin A and B had previously been converted into latrunculin S, C and M, respectively, formal total syntheses of these congeners have also been achieved. Finally, a previously unknown acid-catalyzed degradation pathway of these bioactive natural products is described. The cysteine-derived ketone 18, the tetrahydropyranyl segment 31 serving as the common synthesis platform for the preparation of all naturally occurring latrunculins, as well as the somewhat strained cycloalkyne 35 formed by the RCAM reaction en route to 2 were characterized by X-ray crystallography.