Toward the Asymmetric De Novo Synthesis of Lanostanes: Construction of 7,11-dideoxy-Δ⁵-lucidadone H

Abstract

Efforts to establish an asymmetric entry to hexanorlanostanes has resulted in a concise synthesis of 7,11-dideoxy-Δ⁵-lucidadone H from epichlorohydrin. Exploiting metallacycle-mediated annulative cross-coupling (to establish a functionalized hydrindane) and stereoselective formation of the steroidal C9–C10 bond to establish a stereodefined 9-alkyl estrane, fourteen subsequent steps have been established to generate a hexanorlanostane system. Key additional steps include formal inversion of the C13 quaternary center, oxidative dearomatization/group-selective Wagner–Meerwein rearrangement and Lewis acid-mediated semi-Pinacol rearrangement.

Graphical Abstract

Lanostanes are a large class of triterpenoid natural products, members of which have been isolated from a wide variety of organisms and have been described as possessing a diverse array of biological activities.¹ Prominent structural features of lanostanes include four quaternary centers (two of which are vicinal), substitution at C17, and varied sites of oxidation as well as π-unsaturation within each of their four fused rings (Figure 1A). Examples include C27-variants [e.g., lanosterol (1)] as well as a numerous hexanorlanostanes that possess a simple two-carbon side chain at C17 (Figure 1B). In comparison to less structurally complex steroid hormones,² lanostanes stand out as synthetically challenging targets that have rarely been addressed by de novo asymmetric synthesis. The first synthesis of a lanostane was reported by R. B. Woodward over sixty years ago, defining a means to access lanosterol (1) semisynthetically from cholesterol.³ Despite substantial advances in organic chemistry that have appeared since the first reported synthesis of lanosterol, lanostanes are typically still prepared through semisynthesis rather than de novo synthesis⁴ – an approach that comes with substantial limitations based on the challenge associated with selective functionalization of complex natural product starting materials, and is generally limited to the natural enantiomeric series.^5,6 With this as background, efforts were directed at establishing a modern and concise approach to the de novo asymmetric synthesis of lanostanes that has the potential to fuel future medicinal exploration in a way that is not possible or practical with existing chemical technology. Here, we describe recent success in this program, having achieved the asymmetric synthesis of a highly oxygenated and polyunsaturated norlanostane, 7,11-dideoxy-Δ⁵-lucidadone H (9; Figure 1C). The approach taken is based on recent advances in this laboratory and embraces the synthetic ent-estrane 10 as a key intermediate, proceeding by way of a route that includes formal inversion of the C13 quaternary center and two alkyl shifts to establish the C10 and C14 quaternary centers. Several of these transformations have played important roles in other natural product synthesis efforts conducted in these laboratories, and their successful deployment here speaks to the growing generality of the synthetic approach to tetracyclic terpenoid systems.

Figure 1.

Introduction to lanostanes and an approach to the asymmetric synthesis of a hexanorlanostane (9) from an ent-estrane (10).

For context, established asymmetric syntheses of lanostane-based systems have been inspired, in part, by biomimetic polyene cyclization chemistry (Figure 2A).^7,8 As illustrated in Figure 2B, Corey’s approach to lanostenol (2) and Kobayashi’s strategy for fomitellic acid B (3) use stereoselective polyene cyclization chemistry to establish the lanostane “AB” ring system.^9,10 Notably, these key biomimetic approaches employ complex substrates that contain a functionalized “CD” ring system (trans-fused hydrindane) – in one case prepared from vitamin D,⁹ and in the other prepared by a nineteen-step sequence of chemical reactions starting from the Wieland–Miescher ketone.¹⁰

Figure 2.

Approaches to the asymmetric synthesis of lanostanes and development of a retrosynthetic strategy for norlanostanes.

The approach to lanostane synthesis described here is decidedly not biomimetic, and conceptually related to our recent studies aimed at establishing a unified strategy for the asymmetric de novo construction of estranes (15), androstanes (16), pregnanes (17), and cucurbitanes (18).¹¹ As illustrated in Figure 2C, these previous asymmetric syntheses diverge from a complex tetracyclic intermediate (III) that is easily accessible in an enantiospecific fashion from epichlorohydrin (vide infra).

As illustrated in Figure 2D, our retrosynthetic strategy to assemble a hexanorlanostane system (9) was based on late-stage conversion of a polyunsaturated C15-methylated pregnane 19 to a lanostane system by Lewis acid-mediated semi-Pinacol rearrangement.¹² This structurally complex intermediate was thought to be accessible from the A-ring aromatic precursor 20 through a sequence of reactions featuring oxidative dearomatization and group selective Wagner–Meerwein rearrangement to reposition the methyl group at C9 of 20 to C10 of 19.¹³ The substituted tetracycle 20 was then reasoned to be accessible from the now familiar ent-estrane 10 through a sequence of reactions to accomplish formal inversion of the C13 quaternary center¹⁴ followed by stereoselective introduction of the C15 and C17 substituents. Finally, 10 was known to be readily available by stepwise coupling of epichlorohydrin (21) with 2-propenyl-magnesium bromide (22), the lithium acetylide derived from 23 and TMS-propyne (24).^11c

As depicted in Figure 3, synthesis began from the known tetracyclic enone 26 – a species employed in our recent synthesis of the marine pregnene (+)-03219A that is accessible in just five steps from enyne 25.^11c Annulative cross-coupling with TMS-propyne [Ti(Oi-Pr)₄, n-BuLi, PhMe], followed by tandem protodesilylation and double asymmetric Friedel–Crafts cyclization [(S)-Binol, SnCl₄, CH₂Cl₂] smoothly converted enyne 25 to the stereodefined estrane 10.^11c–f In this way, as we have reported in other synthesis campaigns,^11c–f the steroidal tetracycle was established with exquisite levels of stereoselection, affording the C9,C13-anti-isomer 10 as the only observed tetracyclic product in a combined 51% isolated yield. With ample quantities of 10 in hand, an established three-step sequence was used to formally invert the C13 quaternary center and establish the D-ring enone-containing product 26.^11f Here, ozonolytic cleavage of the C8–C14 alkene followed by dehydration of the resulting β-hydroxy ketone provided an intermediate suitably functionalized to participate in an intramolecular aldol condensation that forges the new C8–C14 double bond present in 26.^11f Notably, the mechanism of the ring-forming reaction ensures a clean formal stereochemical inversion of C13 from tetracycle 10 to the product 26.

Figure 3.

From a simple chiral enyne to a polyunsaturated C15 Me-substituted pregnane.

With the differentially substituted dienone 26 in hand, site-selective conjugate addition to C15 was easily accomplished by exposure to MeMgCl/CuI in THF, delivering the C15 β-methyl-substituted product 27 in 72% isolated yield. Next, a two-step procedure was employed to install a simple C17 side chain (Wittig olefination followed by hydroboration), delivering 20 in 76%.¹⁵ Finally, conversion to the C15-methylated and polyunsaturated pregnane 28 was accomplished through demethylation of the anisole followed by a PIDA-mediated oxidative dearomatization and group selective Wagner–Meerwein rearrangement.^11c,16 In this fashion, the C9 β-methyl group in 20 was repositioned to C10 in 28, and a C14–C15 tetrasubstituted alkene was established.

With ample quantities of 28 available in just ten steps from enyne 25, effort was directed at converting this intermediate to the targeted hexanorlanostane. As illustrated in Figure 4, site and stereoselective hydroxy-directed epoxidation of the C14–C15 alkene [VO(acac)₂, t-BuOOH, CH₂Cl₂]¹⁷ was followed by silylation of the secondary alcohol (TESCl, imidazole, CH₂Cl₂) and deconjugative alkylation of the C3 ketone (KOt-Bu, t-BuOH, MeI) to deliver the vinyl epoxide 19. Gratifyingly, treatment of 19 with BF₃•OEt₂ resulted in an efficient and stereospecific rearrangement to the hexanorlanostane 29 that conveniently occurred along with desilylation. The facility of this rearrangement was certainly impacted by the presence of the C8–C9 alkene, enabling Lewis acid activation to result in a fully substituted allylic cation spanning steroidal carbons 9, 8 and 14.¹² Finally, reduction of the C1–C2 alkene by the action of Wilkinson’s catalyst in the presence of H₂ was followed by oxidation¹⁸ of the remaining secondary alcohol to deliver the hexanorlanostane product 9 in 75% isolated yield over the final two-step sequence.

Figure 4.

Conversion of the substituted pregnane 28 to the hexanorlanostane 7,11-dideoxy-Δ⁵-lucidadone H (9).

While several lanostane natural products are known that possess a C5–C6 alkene,¹⁹ this position is often saturated. As such, attention was briefly directed toward stereoselective reduction of the C5–C6 alkene. It was appreciated that methods have been described for the stereoselective reduction of related carbocyclic species (Figure 5A),^20,21 yet simple Pd-catalyzed hydrogenation of 29 selectively delivered the C5β isomer 34 in 83% isolated yield (Figure 5B).²² Increasing the pressure of H₂ to 88 psi did not impact stereoselection in the hydrogenation, again providing stereoselective access to the cis-fused product (9 → 35). Further, metal-catalyzed hydrogen atom transfer [Mn(dpm)₃, PhSiH₃, t-BuOOH, i-PrOH, CH₂Cl₂],²³ or hydrogenation of hexanorlanostane 36 in the presence of Adam’s catalyst delivered cis-fused products 34 and 37, both isolated from a complex mixture of products.

Figure 5.

Reactions performed to reduce the C5–C6 alkene resulted in cis-fused products.

While future studies will focus on establishing a means to access 5α-norlanostane systems, our efforts have established a concise de novo asymmetric synthesis of hexanorlanostanes (e.g., 29) that proceeds in as few as fourteen steps from a simple chiral enyne (25). While offering a conceptually unique approach to deal with the numerous structural challenges presented by lanostane targets, these efforts build on a foundation of early success aimed at establishing a conceptually unified approach to many structurally distinct tetracyclic triterpenoid systems (e.g., estranes, androstanes, pregnanes, and cucurbitanes). Studies directed at applying aspects of the chemistry described here in other synthesis campaigns are ongoing.