A New Total Synthesis of Ouabagenin

            Masayuki Inoue’s laboratory at the University of Tokyo has been publishing some really great synthetic work over the last few years. For example, their review of convergent total syntheses of complex terpenoids that appeared earlier this year in Chemical Reviews serves as an excellent overview of modern academic campaigns directed toward the synthesis of important bioactive targets including limonoids, cardenolides and taxanes. In 2014, the group disclosed an impressive 22-step synthesis of ryanodol (structure shown above), a botanical insecticide with a stereochemically complex polycyclic molecular architecture. Ryanodol is interesting from an agrochemical perspective because it exhibits highly selective toxicity to insects and its biological mechanism of action may be unique from that of its better-known pyrrole-2-carboxylate ester, ryanodine.

            Earlier this year in the journal, Chemical Science, Inoue and co-workers reported a new de novo synthesis of the classical steroidal target, ouabagenin. Ouabagenin and its C3-O-glycosylated parent structure, ouabain, are high-affinity ligands for the membrane-bound sodium pump, Na⁺/K⁺-ATPase. The chemistry and biology of ouabain and related cardenolides has been reviewed [here] and discussed previously at this site. Only one total synthesis of ouabagenin (by Deslongchamps’ group) predates that of Inoue and required ~41 steps. Inoue’s new total synthesis is slightly shorter and can be considered more modular, due to its highly convergent nature. Three critical building blocks were assembled by Inoue et al in the course of their total synthesis campaign. These fragments corresponded to the functionalized cardiotonic steroid butenolide (lactone moiety), A/B decalin and D-ring substructures. The C-ring was fashioned at a late stage by a stereocontrolled aldol reaction that simultaneously establishes the stereochemical configurations at C8, C13 and C14. For a more detailed analysis of Inoue’s overall strategy for cardenolide synthesis, see here.

            Construction of the highly oxygenated A/B-ring fragment, along with identification of a suitable protecting group strategy for its polyhydroxylated array of functionality, is the most challenging aspect of a strategy for ouabain total synthesis. Synthesis of Inoue’s decalin substructure begins with a Diels-Alder reaction between (R)-perillaldehyde and Rawal’s diene. Further elaboration of the resultant cis-decalin provides the bicyclic dienone shown in the Scheme above. A challenging stereoselective ketone reduction was required to establish a directing group at C3 that would later deliver functionality to steroid carbons 1 and 5, exclusively from the top face. The C3-b-alcohol was obtained with moderate stereoselectivity (d.r. 3:1) by a ketone reduction with a chiral hydride reagent prepared from lithium aluminum hydride. Subsequent exposure of the resultant triene to mCPBA then induced a unique triple epoxidation, wherein the olefins adjacent to C3 were epoxidized exclusively from the b-face due to anchimeric assistance. Next, an oxidation/reduction sequence selectively opened the two epoxides proximal to the C3 ketone to give a diol intermediate. The requisite C1/C3/C5-triol was obtained using DIBAL-H, which reduced the C3 ketone from the bottom face and fragmented the remaining epoxide in a regioselective fashion. The three cis-hydroxy groups were then cleverly protected as an orthoester. Finally, the tertiary alcohol appended to the eventual steroid B-ring was converted to an a,b-unsaturated carbonyl by a three-step sequence involving dehydration, oxidative cleavage of the exocyclic olefin and dehydrogenation. Desilylation with fluoride completed the 15-step synthesis of the ornate A/B decalin fragment.

            The A/B decalin substructure was stereoselectively functionalized at C9 by implementation of a 6-exo radical cyclization protocol that ultimately furnished the critical aldol substrate depicted above. Closure of the steroid C-ring by means of an aldol reaction is well-precedented. One such example that is highlighted in Inoue’s review of terpenoid total synthesis is from Deslongshamps’ 2008 total synthesis of (-)-ouabain (aldol transformation outlined above). Inoue’s aldol approach is somewhat more ambitious in that it attempts to control three stereocenters in a single operation due to the meso nature of the D-ring substructure. A desymmetrization of this nature could theoretically generate up to eight distinct diastereomeric products. However, perhaps due to subtle stereoelectronic properties imparted by the acetal-containing linker of Inoue’s substrate, the base-mediated aldol cyclization, in this example, delivered the desired stereoisomer as the major product (65%), along with a minor 13,14a-diastereomer (8%). The major aldol product was next deoxygenated at C7 by a standard three-step sequence.

            The Inoue group’s endgame parallels that of Baran’s recent partial synthesis of ouabagenin to a certain extent (comparison delineated in the Scheme above). However, the relative simplicity and straightforward nature of Inoue’s approach, in my opinion, renders it preferable (and perhaps more scalable). For example, in both cases, the butenolide lactone is appended to C17 by means of an optimized Stille coupling. However, the Stille strategy necessitates the non-trivial task of subsequently installing the correct configuration at C17 by reducing the D^16,17 olefin from the concave a-face. For this, Baran resorted to screening various superbases to isomerize a tetrasubstituted olefin that was, itself, generated by a cobalt boride-mediated reduction of the dienoate-Stille product. Ultimately, Barton’s base (BTMG) was demonstrated to deliver the correctly configured 17b-enoate with moderate stereoselectivity (d.r. 3:1). Global deprotection with methanolic HCl then furnished semisynthetic ouabagenin. Inoue’s solution to the C17 problem is far more intuitive. Exhaustive silylation of their Stille product produces an advanced intermediate that is capable of blocking the approach of reducing agents from the b-face. In this case, a simple palladium on carbon-catalyzed hydrogenation of the silylated penultimate precursor, followed by global deprotection, is sufficient to secure fully synthetic ouabagenin. To date, a total of three laboratories have successfully generated synthetic ouabagenin, since the isolation of ouabain by Arnaud 27 years ago in 1988. This timeline clearly serves as a testament to the three-dimensional complexity of highly oxygenated steroids such as ouabain. Deslongchamps, Inoue and Baran [see also here] have provided interesting solutions to the ouabain problem, each with varying degrees of practicality.