Building Blocks for Steroid Total Synthesis

          The Wieland-Miescher ketone (3) is a bicyclic enedione that contains the A/B ring substructure of the steroidal carbon skeleton. As such, this well-known building block has been utilized in the total synthesis of numerous terpenoids and steroids, most notably in the pursuit of an industrial process for the preparation of contraceptives and other medicinally important steroids. The versatile steroid synthon is named for Karl Miescher and Peter Wieland, industrial chemists from Ciba Geigy who first prepared the bicyclic diketone in racemic form. A historically significant enantioselective synthesis of the Wieland-Miescher ketone and the so-called Hajos-Parrish ketone (4) proceeds via the intermediacy of an enamine derived from L-proline. This process is often referred to as the Hajos-Parrish-Eder-Sauer-Wiechert (H-P-E-S-W) reaction, named after its principal investigators from Hoffmann-La Roche and Schering AG. In 2004, Houk and Clemente invoked a cyclic transition state (depicted below) to account for the observed stereoselectivity, in which an intramolecular hydrogen bond facilitates proton transfer from proline’s carboxylic acid to the developing alkoxide. The critical role of hydrogen bonding in the stereocontrol of the transformation is reinforced by early reports of reduced enantioselectivity (27-83% enantiomeric excess or ee) when the reaction is run in alcoholic (protic) solvents.

          The proline-catalyzed synthesis of the Hajos-Parrish ketone proceeds with excellent enantioselectivity. However, the Wieland-Miescher ketone, under identical conditions, is obtained in only 70% enantiomeric excess (ee) and requires multiple recrystallizations at low temperature before acceptable levels of optical purity are achieved. With this shortcoming in mind, Luo and co-workers have very recently developed a structurally simple organocatalyst for the enantioselective synthesis of 3 (and analogues thereof) under solvent-free conditions and on gram scale. A significant preparative advantage of this method is its relatively short reaction time: It was found that addition of a second weak acid (meta-nitrobenzoic acid) to the reaction leads to rate enhancement with full conversion observed in twelve hours. Based on analogy to previous reports by the same authors, a plausible transition structure for the H-P-E-S-W transformation mediated by a chiral diamine catalyst is shown below in brackets (right side). The new protocol is very practical and should prove generally useful for organic chemists.

          The stereocontrolled synthesis of an interesting series of 17-b-aryl hydrindanes (e.g. 7 & 8) was also disclosed this year in JACS by the laboratory of Glenn Micalizio. The key step is a titanium-mediated cross-coupling of an internal alkyne (5) with a 4-hydroxy-1,6-enyne (6) to fashion a highly substituted and stereodefined dihydroindane (7). This complex carbocycle, similar to the Hajos-Parrish ketone, represents a valuable chiral building block, orthogonally protected and potentially suitable for elaboration into an oxygenated steroidal natural product such as ouabain or batrachotoxin.

          Mechanistically, the authors provide evidence to suggest that the overall metal-centered [2 + 2 + 2] annulation occurs by initial formation of a metallocyclopentadiene, followed by a [4 + 2] cycloaddition with the tethered olefin and finally cycloreversion to the substituted cyclohexadiene product. In addition, the diastereoselective conversion of intermediate 7 into the 11a-hydroxy derivative 8 by means of a 3-step protodesilylation/hydroboration sequence is also demonstrated. The metallacycle-mediated hydroindane synthetic methodology of Micalizio rapidly assembles a densely functionalized steroidal C/D ring substructure and may be of broad utility in the context of future total synthesis studies. For example, many interesting limonoid natural products substituted with a (hetero)aryl system appended at C17 and oxygenated functionality at C16 have been described, yet relatively few have been synthesized by organic chemists to date. These and other terpenoids, including cortistatin A, may become more accessible by implementation of the technologies described above.

Chemical Synthesis of the Cardiotonic Steroid Glycosides

If you're interested in the organic chemistry and pharmacology associated with structually complex and medicinally important natural products such as ouabain, digoxin, and batrachotoxin, please have a look at this review article, published recently in Chemistry - A European Journal.

Enantiomeric (ent-)Steroids and Bile Acids: Total Synthesis and Biomedical Research Applications

          Steroids such as bile acids (e.g. 2-3) elicit many important pharmacological effects through specific binding interactions with various nuclear receptors and GPCRs. Alternatively, other biological phenomena are mediated by the well-known detergent properties of bile acids. Cholesterol (the enantiomer of 1) is similar in the sense that some of its associated properties are affected by direct binding with biomolecules. On the other hand, due to its lipophilic physical characteristics, cholesterol also stabilizes mammalian membrane bilayers and mediates membrane fluidity. It is sometimes difficult for biochemists to delineate and assess the relative contribution of receptor-mediated properties of steroids separately from the detergent or lipid properties. With this goal in mind, organic chemists have synthesized the unnatural enantiomeric forms of both cholesterol (1) and certain bile acids such as chenodeoxycholic acid (2). Complementary pairs of enantiomers have identical physical properties but have markedly different three-dimensional structures due to opposite stereochemical configurations at asymmetric carbon positions. Enantiomers, therefore, almost always exhibit differential binding affinities with their protein targets due to the chiral three-dimensional environment of proteinogenic ligand binding pockets. Unnatural enantiomeric steroid derivatives can only be obtained by total (de novo) synthesis.

          The synthetic strategies used by chemists to access the enantiomeric steroids 1 and 2 will be the primary topic of the current post. However, the important potential therapeutic applications of the related semisynthetic bile acid 3 developed by Intercept Pharmaceuticals are worth mentioning in brief. INT-777 (3) acts in the periphery (outside of the enterohepatic system) through activation of TGR5, a GPCR that is also activated by chenodeoxycholic acid. TGR5 is implicated in a number of liver and metabolic diseases including obesity and type II diabetes. INT-777 exhibits relatively potent TGR5 agonist activity (EC₅₀ ~ 800 nM, 166% efficacy) and is able to induce the release of glucagon-like protein 1 (GLP-1) in enteroendocrine cells, an attractive in vitro property for a potential diabetes treatment. More recently, a new semisynthetic analogue of avicholic acid was disclosed by the same group with comparable TGR5 potency (EC₅₀ 650 nM) along with weak agonist activity at the farnesoid X nuclear receptor (FXR). Agonism of FXR prevents the accumulation of toxic concentrations of certain bile acids in cells and therefore may also be therapeutically useful. FXR activation promotes bile acid excretion and represses bile acid synthesis and import.

          The enantiomer of natural cholesterol, ent-cholesterol (1), was first prepared by Rychnovsky and co-workers in 1992. Their basic strategy was to synthesize 1 via the intermediacy of ent-testosterone (9) by employing a modified version of the well-known synthetic steroid technologies invented at Hoffmann-La Roche in the 1970’s and 80’s. The cholesterol C17 side chain was then appended to the steroid nucleus at a relatively late stage (cf. 9 à 1). The ‘unnatural’ ring junction configuration of the Hajos-Parrish ketone was obtained by executing the intramolecular aldol (H-P-E-S-W) reaction under enamine catalysis conditions mediated by D-proline. The initial stereogenic center (pro-C13) was used to control all of the remaining stereocenters of 1. Enone 4 was elaborated to the exocyclic enone 6 by treatment with the Stiles’ MMC reagent and then hydrogenation, followed by a decarboxylative aldol reaction with formaldehyde. Subsequent Robinson annulation between 6 and the b-keto ester according to the Hoffmann-La Roche protocol fashioned the tricyclic intermediate 7. Next, a brilliant implementation of the single-electron transfer (SET) reductive alkylation procedure of Stork installed the C19 angular methyl group in a stereoselective fashion as depicted in the bracketed structure above. Finally, exposure of 8 to Bronsted acidic conditions completed the first synthesis of ent-testosterone (9). Nine additional operations were required to elaborate the C17 side chain and establish the A/B ring fusion stereochemistry of ent-cholesterol (1).

          Covey and co-workers disclosed an alternate synthesis of 1 in 2002 that adhered to an ‘east to west’ approach that constructed the steroidal A-ring in the final stages of the total synthesis. The racemic bicyclic b-keto ester system of 10, previously described in 1997 by He et al., was ring-opened with an in situ-generated organocuprate to establish the intact C17 cholesterol side chain appended to a suitably functionalized steroid D-ring. The racemic cyclopentanone 11 was resolved by transesterification with an enantiopure alcohol and then further elaborated (à 12) by an intramolecular aldol cyclization/dehydration sequence. Robinson annulation of 13 followed by application Stork’s reductive alkylation chemistry (similar to above) provided the advanced intermediate 15. In this series of transformations, no serious complications imparted by the lipophilic C17 side chain were noted. Finally, reduction of the dienol acetate 16 afforded synthetic 1 along with a small amount (10%) of its corresponding C3 epimer.

          The laboratory of Douglas Covey at the Washington University School of Medicine has also described the total synthesis of several unnatural enantiomeric bile acids along with characterization of some of their physical properties and receptor interaction profiles. In their synthesis of ent-chenodeoxycholic acid (2), a key transformation, subsequent to oxygenation of C7 (9 à 18), was the stereocontrolled ene reaction of (Z)-olefin 19 with methyl propiolate to provide the diene 20 in moderate yield (47%). Hydrogenation of 20 from the b-face followed by base-mediated hydrolysis then completed the synthesis of 2. Not surprisingly, the unnatural bile acid enantiomer 2, due to its disparate three-dimensional structure (relative to naturally occurring steroids), was inactive at the TGR5 G protein-coupled receptor discussed above.
NOTE: Thanks to reader Bob Hanson for pointing out several structural errors within this post. Those errors have all been corrected. -BHH

The First Semisynthesis of Hippuristanol by the Laboratory of Biao Yu

          We have recently examined an elegant partial synthesis of the antiproliferative steroid natural product hippuristanol that was disclosed by the laboratory of Pierre Deslongchamps in 2010. An alternate synthetic synthetic route, starting from hydrocortisone, was described by Biao Yu (Shanghai Institute of Organic Chemistry) and co-workers in 2009. Hydrocortisone is a very judicious precursor to the target compound as it is relatively inexpensive and contains pre-installed oxygenated functionality at postitions C3, C11, and C20. Yu’s basic strategy is to elaborate the steroidal core structure and then to introduce the eastern bicyclic spiroketal appendage at a later stage via a stereocontrolled nucleophilic addition of an organometallic species to the methyl ketone at C17.

          Hydrocortisone is first converted in several steps to a doubly benzoyl-protected 20-oxo-21-hydroxy derivative. In this preliminatry sequence, the A/B trans-fused ring junction stereochemistry is established with a dissolving metal reduction of the hydrocortisone enone system and the requisite axial 3alpha-ol is secured by a subsequent reduction of the corresponding C3 ketone with a sterically bulky hydride source (K-selectride). Next, deoxygenation of C21 is accomplished in three steps, followed by a semicarbazide-promoted elimination of water that installs a unit of unsaturation within the steroidal D-ring. Regio- and stereoselective hydration of the delta^16,17 olefin is then achieved via the intermediacy of a bromohydrin that undergoes reductive dehalogenation under free-radical conditions to afford the 16beta-hydroxy pregnane advanced intermediate.

In the endgame, the lithiated dihydrofuran nucleophile adds to the beta-hydroxy C20-ketone to generate an initial adduct (shown in brackets) that is immediately subjected to Bronsted acid-mediated spiroketalization followed by deprotection of the benzoyl group with lithium aluminum hydride (LAH). Because the dihydrofuran is racemic, the maximum overall yield of this sequence is 50%. In light of this, a quite acceptable level of efficiency (43%) is observed. This is largely due to complete chelation control of the newly formed stereogenic position at C20 (For a discussion of this Cram chelate-controlled stereochemical outcome in a related system, see this post). Unfortunately, due to anomeric stability (For a discussion of the relevant anomeric effects pertaining to this transformation, see this post), C22-epi-hippuristanol is thermodynamically favored compared to the spiroketal configuration of the natural product. Happily, exposure of C22-epi-hippuristanol to a catalytic amount of a sulfonic acid in an aprotic medium effects epimerization of the spiroketal to provide synthetic hippuristanol. A very similar stereochemical phenomenon was observed by Deslongchamps et al. in the course of their partial synthesis of this structurally captivating steroidal natural product. 

The semisynthetic preparation of hippuristanol by Biao Yu’s group is notable for its relative ease of synthesis and expediency. The lack of stereocontrol in the construction of the eastern bicyclic spiroketal appendage can be viewed as a shortcoming from the perspective of synthetic efficiency. However, from a medicinal chemistry viewpoint, the route provides access to stereochemical variation in the eastern hemispheric structural domain of hippuristanol. In vitro characterization of a series of analogues, produced in the course of the synthetic campaign, has revealed some very interesting structure-activity relationships (SARs) regarding bioactivity. For example, it was shown that the spiro configuration of the C22 position and the presence of the geminal dialkyl substitution on the steroidal F-ring are critical to the antiproliferative activity of hippuristanol. Notably, the configuration and presence of the C24 methyl group (adjacent to the geminal di-methyl) does not contribute significantly to the biological activity of hippuristanol and may represent a useful ‘point of diversity’ that could be modulated in a medchem optimization effort to improve the potency and/or physical properties of a potential anticancer drug candidate.