Asymmetric Synthesis of the Torgov Diene: Application to Total Synthesis of Estrone

Igor Torgov (1912 - 2007)

            Igor Vladimirovich Torgov was a Soviet chemist who served as the head of steroid chemistry at the Institute of Natural Compounds Chemistry of the Soviet Academy of Sciences (currently the Shemyakin Institute of Bioorganic Chemistry). Dr. Torgov’s famous reaction is a carbocyclization of an achiral diketone (discussed in detail below) that provided an early pathway for the total chemical synthesis of steroid hormones, and is still used for the industrial-scale production of certain steroids at the factories of Schering AG in Berlin, Germany. The product of the Torgov cyclization, the Torgov diene, is a steroidal D^8,14 dienone which serves as a synthetic precursor to a range of estrol and estradiol derivatives including estrone. Torgov’s synthetic methodology provides expedient access to racemic steroids. However, since its original disclosure in 1963, the development of an enantioselective variant of the Torgov cyclization has remained elusive. Torgov and co-workers were eventually able access enantiopure steroid intermediates by resorting to an enantioselective reduction of a key intermediate, mediated by S. cerevisiae yeast culture. But the use of fermenting yeast reduction suffers from drawbacks that include large reaction volumes and difficulties encountered during extraction of water-soluble products. Additionally, in the late 1990s, workers at Schering reported studies toward the development of a catalytic asymmetric version of the Torgov cyclization, but high selectivity and turnover numbers were not achieved. Consequently, enantioselectivity in the Torgov process for steroid production has been considered a classic unsolved problem for organic chemists.

            Very recently, Benjamin List’s laboratory at the Max Planck Institute disclosed a report describing a catalytic asymmetric version of the Torgov cyclization (see above, conversion of 1 à ‘Torgov’s diene’). Mechanistically, the Torgov cyclization involves (1) isomerization of intermediate 1 to the endocyclic D^8,9 isomer A (2) Prins-type cation-olefin cyclization (3) deprotonation of the ensuing carbocation to give the alcohols B and, finally, (4) isomerization and dehydration of B, which furnishes the intact Torgov diene. List and co-workers recognized that the chirogenic, stereo-determining step is likely the cyclization of intermediate A and that this transformation might be catalyzed by an enantiopure Brønsted acid. Upon screening of a range of acids, a novel chiral disulfonimide (DSI) (shown above, see box) emerged that was uniquely able to deliver the requisite diene at low temperature and with high enantioselectivity. A gram-scale Torgov cyclization of intermediate 1, catalyzed by the unique pentafluorothio- and nitro-containing DSI, proceeded in high yield to furnish Torgov’s diene with an enantiomeric ratio of 96.5:3.5. A single recrystallization provided essentially optically pure material (>99.9:0.1 e.r.). Moreover, the precious DSI could be recovered from the reaction in 88% yield. List’s group then repeated a two-step procedure that was previously developed by E. J. Corey’s laboratory to diastereoselectively reduce Torgov’s diene to estrone methyl ether (3). Methyl ether deprotection yielded fully synthetic (+)-estrone, thus completing the shortest route to the bioactive steroid reported to date.

            List’s achievement in the enantioselective chemical synthesis of estrone is rivaled in its conciseness only by the work of E. J. Corey’s group, which has previously disclosed no less than three (!!!) distinct enantioselective processes. The first, reported in 2004 (chemistry not shown), uses a catalytic enantioselective Diels-Alder reaction as the key step. The second, outlined in the scheme above, proceeds through the intermediacy of Torgov’s diene. In this work, the achiral diketone 1 is first reduced enantioselectively using oxazaborolidine catalysis in combination with the reductant catecholborane. A single recrystallization of the product from ethyl acetate-hexane affords highly enantioenriched 4 (99% ee). Torgov cyclization of mono-ketone 4 then proceeds efficiently in methanolic hydrochloric acid and the ensuing optically active dienol 5 is oxidized to Torgov’s diene by the IBX reagent. Finally, the same two-step reductive sequence from above, borrowed by List and co-workers, efficiently converts Torgov’s diene into estrone methyl ether. An alternate enantioselective total synthesis of Torgov’s diene, reported by Corey’s group in 2008, is discussed here. We should note that, these days, (+)-estrone is commercially available from Sigma-Aldrich for about $1.75/gram.

Taccalonolides: Microtubule-Stabilizing Steroids with in vivo Antitumor Efficacy

                  Microtubules are noncovalent polymers of a- and b- tubulin heterodimers, assembled in a filamentous tube-shaped structure. During mitosis, microtubules form the mitotic spindle that transports daughter chromosomes to opposing poles of the dividing cell. Given their important role in cell division, microtubules have been an important target for anticancer drug discovery. Microtubules must exist in a dynamic state, growing and shortening by the reversible association and dissociation of a- and b-tubulin heterodimers at both ends. Disruption of microtubule dynamics by small-molecule microtubule inhibitors prevents cell cycle progression and inhibits mitosis. Microtubule inhibitors can be stabilizing or destabilizing agents. For example, the anti-cancer drug taxol (see Figure above, boxed structure), the first microtubule stabilizer identified, operates by promoting polymerization and increasing microtubule polymer mass in cells. Destabilizing agents such as the vinca alkaloids depolymerize microtubules and decrease polymer mass. Both stabilization and destabilization disrupt microtubule dynamics, eventually resulting in apoptotic cell death. Because tumors are dependent upon continuous mitotic division in order to grow and metastasize, this pharmacological effect is particularly detrimental to various forms of cancer. Although many types of cancer respond initially to microtubule inhibitor treatment, complete remission is rarely achieved. This is often due to the emergence of multiple drug resistance (MDR), wherein tumors can effectively decrease the intracellular concentration of drug by overexpression of membrane efflux pumps such as P-glycoprotein (Pgp). Thus, new clinically efficacious microtubule stabilizers capable of evading MDR are in high demand.

            In 2003, the taccalonolides, a relatively new class of plant-derived microtubule stabilizers, were isolated from a variety of Tacca species by Susan Mooberry’s group in San Antonio. The taccalonolides are highly oxygenated hexacyclic steroid lactones that, similar to other microtubule inhibitors, increase the density of cellular microtubules, interrupt mitotic progression and, consequentially, promote apoptosis of cancer cells. Importantly, taccalonolides A and E were shown to be potent and effective antitumor agents in vivo, with the ability to circumvent multiple mechanisms of MDR. Moreover, these antimitotic steroid lactones cause microtubule-stabilizing effects through a unique binding site not shared by other known microtubule stabilizers, and via an entirely new mechanism of action, which is not currently well understood.

            The most potent of the antiproliferative taccalonolides are, without exception, epoxidized across carbons C22 and C23 on the a-face of the molecule. Treatment of naturally occurring taccalonolides that contain a C22/23-enol ester double bond with dimethyldioxirane (DMDO) cleanly installs the activity-promoting oxirane functionality, leading to dramatic increases in antiproliferative potency in HeLa cells. For example, epoxidation of the newly isolated taccalonolide T (1) results in the generation of a semisynthetic derivative (2) with a 744-fold increase in potency, relative to 1. The subnanomolar IC₅₀ value exhibited by T-epoxide (2) compares favorably with the IC₅₀ of paclitaxel (taxol) in this assay. T-epoxide was also shown to cause interphase microtubule bundling in HeLa cells, as determined by immunofluorescence visualization (images shown above). The stereoselectivity of the DMDO-mediated epoxidation may be due to steric shielding of the b-face by the axial C27-methyl substituent, highlighted in the scheme above.

Xenograft data from: Mooberry et al J. Med. Chem. 2014, 57, 6141 – 6149.

            The chemically-modified taccalonolide analog T-epoxide (2) also exhibited some in vivo antitumor efficacy in a xenograft model of breast cancer. Individual doses of 0.25 mg/kg of 2 were administered twice in the first week, with no additional doses given due to an average 10% body weight loss observed on day seven. Tumor growth was completely inhibited in the first week and significant antitumor effects were sustained for an additional week (original data from J. Med. Chem. 2014 manuscript reproduced above). In spite of an apparently narrow therapeutic window, the ability of a relatively low dose of compound 2 (total dose of 0.5 mg/kg) to produce antitumor effects highlights the exceptional potency of this exciting new microtubule-stabilizing steroid lactone.