Inoue’s Convergent Total Synthesis of a Cardiotonic Steroid

                In the late 1940s, the cardenolide sarmentogenin (structure shown above), a close structural congener of ouabagenin, was investigated by Reichstein and Lardon as a potential plant-derived starting material for the synthesis of cortisone. At this time, the development of an efficient synthesis of the newly discovered anti-inflammatory wonder drug, beginning from a plant sterol or sapogenin, was the chemical challenge of the day (discussed here). Sarmentogenin is uniquely oxygenated at steroid position C11 and contains a C17 butenolide lactone that can be easily degraded to a ketol. The disadvantage of such a semisynthetic approach to cortisone was limited accessibility of sarmentogenin from natural sources. Sarmentogenin is mainly obtained from various Strophanthus seeds in yields ranging from 0.1 to 0.78%. Although an efficient partial synthesis of cortisone from sarmentogenin was developed, ultimately, Strophanthus seeds with high sarmentogenin content could not be cultivated cheaply enough to be of practical use as a source for commercial production of cortisone.

          Quite recently, the laboratory of Masayuki Inoue (The University of Tokyo) has described a totally synthetic approach to 19-hydroxysarmentogenin (1) starting from the readily accessible building blocks 2-4. Inoue’s synthesis of 1 is discussed below.

            To begin, the cis-decalin 2 is tethered to a dibromide derived from 3 to furnish the acetal 5. Homolytic cleavage of the carbon-bromine bond in 5 then gives rise to a carbon radical that is captured by the pendant B-ring enone double bond exclusively from the top face, which provides the requisite stereochemical outcome at steroid position C9. The resultant fused tricycle 6, obtained as an inconsequential diastereomeric mixture, is next merged into a single vinyl ether (7) upon exposure to acid. The substrate for a critical C-ring-forming aldol cyclization is obtained from 7 in two additional steps. In the event, treatment of the trione 8 with a catalytic amount of base induces chemoselective C8 enolate formation with concomitant regio- and stereoselective attack on the C14 ketone to furnish the pentacyclic product 9. The reaction correctly installs the C8, C13 and C14 stereocenters under the optimized thermodynamic conditions. Indeed, re-subjection of the minor diastereomeric product, 13,14-epi-9, to the reaction conditions produces 9, suggesting thermodynamic control over the aldol process. The stereochemical outcome of this cyclization is disparate from related intramolecular steroid C-ring-forming aldol reactions (discussed here), perhaps due to geometric constraints caused by the additional enol ether ring in 8. Seven additional operations are then required to elaborate the aldol product into the polyhydroxylated 17-oxo androstane derivative 10.

            The ketone 10 was then converted into a vinyl iodide using Barton’s method (hydrazine/I₂) and subsequent Stille coupling with the stannane 4 introduced the C17 butenolide of 11. Hydrogenation of the C16-C17 double bond from the bottom face to produce a b-oriented butenolide required shielding of the convex (top) face of the molecule with a bulky C14 silyl ether. Hydrogenation of silylated 12 proceeded from the a-face with good selectivity (d.r. 6:1) and global deprotection under acidic conditions delivered the target compound 1. The synthesis of 1 proceeds in 28 total steps from (S)-perillaldehyde, which compares quite favorably with the relatively few other de novo synthetic approaches to related cardiotonic steroids.