Stereospecific Synthesis of Substituted cis-Hydrindan-5-ones and Their Regiospecific Enolization and Functionalization: Synthetic Intermediates for Reserpine

An approach to reserpine (1) is described which is based on an anionic oxy-Cope rearrangement to establish three of the five contiguous asymmetric centers of the E ring and a regiospecific enol silylation coupled with hydroboration-oxidation to introduce the last two centers. The key intermediate 8b for the anionic oxy-Cope rearrangement is prepared from the readily available 7,7-dimethoxynorbornenone 6 in 50% overall yield by addition of propargyl alcohol dianion, hydride reduction, and TBS formation. Rearrangement of 8b followed by hydrogenation affords the desired 4-[(silyIoxy)methyl]-l,l-dimethoxytetrahydroindan-6-one 10 in which all the protons were cis in nearly 80% yield. Force-field calculations on the two regioisomeric methyl enol ethers derived from 10 indicate a strong preference (96:4 at 25 "C) for enolization toward the ring juncture. Indeed treatment of 10 under thermodynamic conditions (BMDA, Me,SiCI) produced a single (>95%) enol ether assigned the desired A'' (reserpine numbering) structure 11 by a combination of spectroscopic and chemical means. Conversion of 11 into its TBS enol ether 15 followed by hydroboration-oxidation introduced the last two asymmetric centers in an E-ring precursor 5 of reserpine (93% for the last step). Finally the unusual base-catalyzed hydrolysis of the dimethyl ketal which occurs in this process is rationalized by strong intramolecular complexation of a methoxyl group with the boron atom leading to ultimate displacement of this methoxyl group with hydroxide. Reserpine (l), a Rauwolfia alkaloid with good antihypertensive properties, has been synthesized several times since Woodward's classical synthesis in the late 1950s.* This first synthesis prepared the E-ring derivative 2a in which the five contiguous asymmetric centers were correctly established and then coupled it to 6methoxytryptamine to give, after reduction, the lactam 3a and thence reserpine in several further steps. Pearlman3 prepared an analogous E-ring intermediate 2b from a different precursor and then followed the Woodward route to reserpine. In Wender's ~yn thes i s ,~ the cyclic amine 4 was prepared by a completely different approach and taken onto reserpine. We have begun an approach to reserpine in which the E-ring target molecule is of lower functionality, namely, the aldehyde tosylate 2c, which on reaction with 6-methoxytryptamine would give directly the desired immonium salt necessary for cyclization to re~erp ine .~ We now report the anionic oxy-Cope rearrangement of the readily available substrate 8b to produce a cis-hydrindenone with three of the five contiguous asymmetric centers established and the regiospecific enolization of a derivative of this hydrindenone which permits the introduction of the final two centers with the correct stereochemistry. In this manner, the hydrindanone 5, with the correct stereochemistry a t the five contiguous asymmetric centers, has been prepared in eight or nine steps in good overall yield (12-14%). In our earlier work on the rearrangement of substituted allylic alcohols derived from the bicyclic enone 6, we showed that the stereochemistry of the tertiary hydrogens about the periphery of the cyclohexanone system was all cis due to the geometric demands of the transition state.6 We reasoned that the use of a simple trans-3-alkoxypropenyl-substituted alcohol, e.g., 8b, would permit the easy preparation of the hydrindenone 9 with the all-cis arrangement of the three contiguous asymmetric centers. This was put into practice as follows: Reaction of the enone 6 (Scheme 'Deceased September 20, 1984 ti &e 1, Ar= 3.4.5-(MeO)&H2