Tetraquinane Total Synthesis: (−)−Crinipellin A

            In addition to steroids, the cyclopentane ring system has been discussed here in the context of ‘privileged’ molecular scaffolds for drug discovery research and development programs. Aside from favorable drug-like properties and diverse bioactivity, cyclopentanoid-based natural products such as the linear triquinanes (e.g. coriolin), are very interesting synthetic targets from the standpoint of architectural molecular complexity. The crinipellins were first isolated in 1979 from a mushroom called Crinipellis stipitaria and are the only known tetraquinane natural products. Enone-containing crinipellins (e.g. crinipellin A) exhibit potent antibacterial and anticancer biological activities. Researchers from Hee-Yoon Lee’s laboratory in Daejeon, Korea have disclosed the first total synthesis of (-)-crinipellin A starting from an optically active cyclopentane (1). The new route features a cascading sequence of tandem [3+2] cycloaddition reactions that proceed in a single operation to forge the tetracyclic core in a highly stereocontrolled fashion. Lee’s approach parallels the ‘two-phase’ strategy for terpenoid total synthesis pioneered by Phil Baran at the Scripps Research Institute. Biomimetic two-phase terpene synthesis entails an intial ‘cyclase’ phase to assemble the carbocyclic skeleton followed by a series of controlled oxidative transformations (i.e. the ‘oxidase’ phase). Along similar lines, Lee’s group first constructs the requisite tetraquinane framework (2) and then adjusts the oxidation state of five carbon centers (designated above with red circles) to complete the total synthesis of crinipellin A.

            Lee designed an intricate diazo-allene system (shown below in brackets) as a precursor to a highly reactive trimethylene-methane (TMM) diyl species, poised to undergo a cascading sequence of stereocontrolled cycloaddition reactions. A p-toluenesulfonylhydrazone (5) serves as a latent diazo group, revealed upon anion formation under basic conditions. The key step in the synthesis of 5 is an iron-catalyzed S_N2′-type reaction of the epoxyalkyne 3 with an acetal-containing Grignard reagent. This reaction produces the allene 4 as an inseparable (and inconsequential) 1:1 mixture of diastereomers. Three subsequent functional group interconversions afford the requisite cycloaddition substrate 5.

            Upon exposure to sodium hydride in reluxing toluene solution, 5 undergoes intramolecular [3+2] cycloaddition to fashion an intermediary heterocyclic 5,5-fused bicyclic system. Extrusion of nitrogen then gives rise to the putative TMM diyl diradical, which engages the pendant exocyclic olefin in a second cycloaddition reaction to furnish the desired tetraquinane 2 with complete stereocontrol. According to the authors, the relative stereochemical outcome is apparently governed by conformational constraints within the tether that place the bulky silyl ether in a pseudoequatorial position in the TMM diyl cycloaddition transition state. In what some might refer to as a biomimetic ‘oxidase’ phase, the advanced intermediate 2 is finally elaborated into (-)-crinipellin A by a twelve-step sequence of controlled oxidative transformations. Lee’s group also confirmed the absolute stereochemistry of the natural product through asymmetric total synthesis.

Contemporary Methods for Stereoselective Construction of the Cholestane-type Steroid Side Chain

            Naturally occurring sterols are typically substituted at carbon position 17 of the cyclopentenophenanthrene D-ring with a functional group that, to an extent, defines the structural subclass to which a given steroid belongs. For example, limonoids generally possess a heteroaromatic furan ring at C-17, whereas cardenolides bear a butenolide-lactone system. In the 1940s, a significant amount of effort was invested in the degradation of the spirocyclic eastern side chain functionality of abundant plant sapogenins, into the two carbon units present in the corticosteroids and other bioactive pregnane derivatives. Moreover, nearly all of the classical anabolic steroids have a b-hydroxy substituent at C-17. However, in contrast to these noteworthy examples, a vast array of more lipophilic sterols contain a modified iso-octyl side chain appended to the 17-position of the D-ring with either (R) or (S) stereochemistry at C-20. These are the cholestane-type steroids, which include important neurotrophic compounds (see below) as well as vitamin D and the bile acids. Their branching side chain is an essential pharmacophoric feature in various biological structure-activity relationship (SAR) studies, yet relativity little synthetic attention has been paid to the stereocontrolled construction of this motif.

(A) PC12 cells exhibit no neurite outgrowth. (B) Sprouting in PC12 cells is indicative of chemically-induced neurite outgrowth. From: Danishefsky et al J. Org. Chem. 2005, 70, 9849 – 9856.

            In the case of the polyhydroxylated sterol, NGA0187, a side-chain truncated structural analogue (structure shown above) was synthesized by Danishefsky’s group and evaluated for neurotrophic activity as compared to that of a synthetic sample of the natural product. The side-chain truncated analogue did not promote neurite outgrowth in PC-12 cells to any appreciable extent, while NGA0187 was active in the assay (269% outgrowth rel. to a DMSO control) at a concentration of 30 micromolar. Aplykurodinone-1 is a degraded marine steroid that contains the precise side-chain corresponding to that of the biogenetic cholesterol precursor, lanosterol. Finally, the molecular framework of withanolide A features an oxidatively cyclized iso-octyl side chain that contributes to the natural product’s potent neuritogenic properties, exhibited in human SH-SY5Y neuroblastoma cells.

            Establishment of the stereogenic methyl configuration at C-20 with a high level of control remains a vexing problem in steroid total and partial synthesis. Danishefsky has referred to this as the ‘C-20 problem,’ stating that “the challenge is that of correlating the configuration of the presumably ‘freely rotating’ C-20 with the resident stereochemistry of the polycyclic domain.” In this post, we will highlight modern strategies that attempt to address the ‘C-20 problem,’ as we examine selected case studies involving cholestane side chain assembly. For the interested and motivated reader, a recent Review article covers this topic in a more comprehensive fashion.

            In 2005, Danishefsky’s group reported the partial synthesis of NGA0187 starting from adrenosterone. Conjugate addition of the organocuprate (Gilman-type reagent) derived from the vinyl iodide 4 to the oxidatively functionalized D-ring enone 5 establishes the configurational relationship between stereogenic carbons 17 and 20 in a highly controlled fashion. The aldehyde precursor to 4 was obtained as a single enantiomer by oxidative cleavage of the side chain of commercially available ergosterol. The advanced intermediate 6, bearing a strategically vital triethylsilyl ether (TES) protecting group at C-11, was suitable for subsequent elaboration to NGA0187.        

              OSW-1 is a bioactive cholestane glycoside that contains a ketone group within its iso-octyl side chain. A related synthetic strategy (shown below), involving, in this case, conjugate addition of an alkoxyvinyl cuprate to an analogous D-ring enone (9), was employed by Zhendong Jin’s group in the course of their semisynthesis of the exceptional cytostatic agent, OSW-1.

            A captivating alternate approach to the synthesis of OSW-1 described in 2008 hinges upon Masayoshi Tsubuki’s recognition that a 4-methylthienyl heteroaryl system can serve as a latent isopentyl group, comprising a portion of the cholestane side chain. The thiophene is easily incorporated into the OSW-1 side chain by a [2,3]-sigmatropic rearrangement of the C-16 O-alkylated ether precursor 12, affording 13 as an inconsequential diastereomeric mixture. In the final step of Tsubuki’s synthesis, reductive desulfurization of 14 with W-2 Raney Nickel under an atmosphere of hydrogen furnished OSW-1 with outstanding efficiency.

            Gademann and co-workers have disclosed a stereocontrolled preparation of withanolide A (discussed here) from a readily available steroid precursor. Their expedient partial synthesis (highlighted below) proceeds in only 13 total steps and involves minimal use of protecting groups. The synthesis is initiated by a highly diastereoselective addition of lithiated 1,3-dithiane to the C20-keto group of 15. Gademann and co-workers do not comment on the stereochemical outcome of this conversion except to say that the product is known. A 1978 Review article by Jerzy Wicha provides a detailed rationale for the addition of sterically bulky nucleophiles to the 20-ketone of the steroid system, a reaction with a rich history dating back to Woodward’s total synthesis of cholesterol (JACS 1951). This stereochemical result is best explained by ‘steric approach’ control, which favors attack of the carbonyl from ‘outside of the molecule,’ or, in other words, from the C16 side. The observed diastereoselectivity, dictated by the conformational transition state model (TS-I) shown below, generally predominates for the addition of bulky nucleophiles to C17 non-hydroxylated (i.e. 17a-H) steroidal 20-ketones. The dithiane 16 was converted into withanolide A in good overall yield.

            In Danishefsky’s total synthesis of racemic aplykurodinone-1, the requisite tricyclic hydrindanone core (18) is assembled in 15 steps from the Danishefsky diene (17). Next, Lewis acid-mediated conjugate addition of a cuprate derived from 19 was accomplished with a high degree of facial selectivity to provide the trisubstituted alkene intermediate 20. In a critical operation, the homogenous double bond hydrogenation of 20 using the Crabtree iridium catalyst occurs with catalyst approach from the less hindered face of the olefin, resulting in the desired stereochemical outcome at the newly formed C-20 stereocenter (steroid numbering). Alternate conditions were also developed that provide exclusively the C-20-epi isomer of aplykurodinone-1. The advanced intermediate 21 could be easily advanced to the racemic natural product target by a short three-step sequence.

            Finally, a new chiral auxiliary-based approach to the stereocontrolled establishment of the natural cholestane C-20 configuration was developed very recently by Douglas Covey’s laboratory, in the course of their partial synthesis of a ‘smoothened’ (18,19-di-nor-) cholesterol analogue starting from a derivative of testosterone. The strategy is conceptually straightforward and somewhat linear in nature (6 total steps required for conversion of 22 à 25), but should provide general access to the desired stereochemical outcome at C-20 across a broad range steroidal substrates. The unique smoothened cholesterol derivative described by Covey, lacking angular methyl groups within the steroid core framework, will be used in biochemical and biophysical studies comparing its abilities to condense and stabilize lipid membranes with those of natural cholesterol.