Synthesis of Bufadienolide Cinobufagin via Late-Stage Singlet Oxygen Oxidation/Rearrangement Approach

Abstract

This manuscript describes a concise synthesis of cinobufagin, a natural steroid of bufadienolide family, from readily available dehydroepiandrosterone (DHEA), as well as its α5-epimer derived from 3-epi-andosterone. This synthesis features expedient installation of the β17-pyrone moiety with the β14,β15-epoxide and the β16-acetoxy group using a photochemical regioselective singlet oxygen [4+2] cycloaddition followed by CoTPP-promoted in situ endoperoxide rearrangement to provide a 14β,16β-bis-epoxide in 64% yield with a 1.6:1 d.r. This β,β-bis-epoxide intermediate was subsequently subjected to a regioselective scandium(III) trifluoromethanesulfonate catalyzed House-Meinwald rearrangement to establish the β17-configuration. The synthesis of cinobufagin is achieved in 12 steps (LLS) and 7.6% overall yield, and we demonstrate that it could be used as a platform for the subsequent medicinal chemistry exploration of cinobufagin analogs such as cinobufagin α5-epimer.

Graphical Abstract

Cardiotonic steroids represent an important family of natural products found in both animals and plants.¹ Depending on the nature of the β17-heterocyclic ring, cardiotonic steroids can be classified into more frequently observed cardenolides or less abundant bufadienolides. Bufadienolides feature an α-pyrone moiety at the β17-position of the cardiotonic steroid skeleton and may or may not carry C3-glycosylation.² Bufadienolides have been found in both plants and animals and may possess significant toxicity that is used for defensive mechanisms.^2c,d Neurotoxic bufadienolides are produced by plants such as Cotyledon, Tylecodon or Kalanchoe of succulent Crassulaceae family from Southern Africa.³ These plants contain bufadienolide steroids such as cotyledosides, tyledosides, orbicusides or 3-O-acetylhellebrin for protective purposes, and the plant consumption by the livestock, and subsequently by humans, results in a cumulative neurotoxic effect (cotyledonosis or krimpsiekte). Other plant- and animal-derived bufadienolides may result in acute cardiotoxicity.

Perhaps, the most significant sources of bufadienolides for medicinal chemistry applications are toads of Bufo genus (cf. Figure 1A and 1B).⁴ ChanSu, a medicine prepared from these toads to treat various conditions including cardiac dysfunction, has been used in traditional oriental medicine for centuries.⁵ Bufalin, marinobufogenin, cinobufagin and bufotalin (cf. Figure 1A) are among the most abundant ingredients of ChanSu, and numerous studies have investigated their anticancer, anti-inflammatory and antinociceptive properties.⁴ Cinobufagin is a bufadienolide with characteristic structural features that include β14,15-epoxide and β16-acetoxy group and are often found in other bufodienolides such as cinobufotalin and cinobufoginol (cf. Figure 1B). Since its isolation in 1932 by Jensen,⁶ numerous studies have explored its medicinal properties and demonstrated its clinical significance.⁷ In addition to inhibiting Na⁺,K⁺-ATPase, cinobufagin has been shown to regulate secretion of aldosterone and cortisol and StAR gene expression as well as induce cell cycle arrest and apoptosis in tumor cells. Due to these versatile physiological effects, cinobufagin exhibits wide range of pharmacological properties that include antitumor,⁸ analgesic,⁹ cardioprotective,¹⁰ immunomodulatory,¹¹ antifibrotic,¹² antiviral,¹³ and antiprotozoal¹⁴ properties.

Figure 1.

Examples of natural bufodienolides isolated from Bufo toad venom, prior approach to bufotalin by Inoue and coworkers, and summary of this work.

Due to their high therapeutic potential, numerous studies focused on developing synthetic access to bufodienolides. Since the first reported approach to bufalin and resibufagenin from 14α-hydroxycortexolone by Sondheimer and coworkers in 1969,¹⁵ several different approaches to bufalin, resibufagenin, scillarenin and their derivatives lacking α14-hydroxylation have been reported.^16,17 However, despite these advances, the synthesis of more complex bufadienolides featuring additional oxidation remains a significant challenge. The sensitivity of β17-pyrone to oxidative, reductive, and basic conditions significantly limits the scope of the synthetic strategies that could be used to incorporate additional oxidation into the steroid skeleton. In 2020, the Inoue group reported the first approach to four different bufotalin derivatives featuring β16-oxidation (Figure 1C).¹⁸ Starting from 1 as the building block, Inoue and coworkers developed a 9-step synthesis of pyrone-containing intermediate 2, which was subjected to β14-OH-directed epoxidation followed by House-Meinwald rearrangement leading to the key intermediate 4 possessing hard-to-install β17-stereochemistry. Access to 4 enabled the completion of bufotalin and its derivatives.

Our research group has a long-standing interest in the total synthesis of steroid natural products.¹⁹ Our recent work focused on the synthesis and exploration of cardenolides oleandrin and rhodexin B featuring a β16-acetoxy group.²⁰ Similar to Inoue’s approach, these syntheses featured diastereoselective epoxidation followed by the Lewis acid-catalyzed House-Meinwald rearrangement to install the β17 stereocenter and C16 oxidation. Interested in exploring the medicinal chemistry of cinobufagin and related bufadienolides featuring β14,β15-epoxy and β16-acetoxy oxidation, we have developed an efficient synthesis of cinobufagin and demonstrated that it could be used to access related analogs with different AB-rings. Our strategy features a two-step functionalization of precursor 5, which is readily available from DHEA, to install the C17-pyrone moiety using a Stille cross-coupling (95% yield). The resultant intermediate 6 was subjected to a one-pot regioselective singlet oxygen [4+2] cycloaddition followed by rearrangement of the endoperoxide intermediate to provide bis-epoxide intermediate 8 in 64% yield (β:α = 1.6:1). Finally, regioselective activation of the C16 epoxide moiety of 8 with scandium(III) trifluoromethanesulfonate was used to establish the desired C17 stereochemistry and provide a functional handle to complete the synthesis. To the best of our knowledge, this work represents the first total synthesis of cinobufagin to date and provides a route which may potentially allow further diversification to other medicinally relevant bufadienolide natural products from a late-stage intermediate.

Our synthetic studies began with the known steroid derivative 10 that was accessed from commercially available DHEA in 57% yield over 3 steps (Scheme 1).^20a The D-ring of 10 was then modified through a known 2-step Saegusa-Ito oxidation to provide enone 5 in 62% yield. Conversion to a triflated diene was accomplished by treating the compound with 2,6-di-tert-butyl-4-methylpyridine (DTBMP) and triflic anhydride at 0 °C in DCM.²¹ The resultant vinyl triflate intermediate was used without purification as the coupling partner of the known stannane 11¹⁸ in a Stille reaction. In our studies, we found it was critical to sparge the reaction mixture with nitrogen for 1 hour at room temperature before heating to furnish diene 6 in 95% yield over 2 steps.

Scheme 1.

Synthesis of intermediate 6 from DHEA^a

^aReagents and conditions: (a) cyclohexanone (20 equiv), Al(OiPr)₃ (2.5 equiv.), PhMe, reflux, 30 min, 95% yield; (b) H₂ (1 atm), Pd/C (5 mol%), Py_(anhydr), r.t., 48 h, 80% yield, 6:1 d.r.; (c) K-selectride^® (1.5 equiv), THF, −78 °C, then add TBSCl (3 equiv), ImH (4 equiv), DMF, r.t., 16 h; 88% yield; (d) LDA, THF, −78 ºC, then TMSCl, r.t., 1 h; (e) Pd(OAc)₂, MeCN:DCM, r.t., 20 h; 62% yield, 2 steps; (f) Tf₂O (2 equiv), DTMBP (10 equiv), DCM, 0 °C to r.t., 1 h; (g) 11 (1.4 equiv), Pd(PPh₃)₄ (10 mol%), CuCl (5.0 equiv), LiCl (10 equiv), DMSO, r.t. to 60 °C, 1.5 h; 95% yield, 2 steps.

Based on the cross-coupling method above, two analogs of 6 containing a C17-pyrone and Δ^14,16 diene moiety, compounds 15 and 18, were also synthesized (cf. Scheme 2 and SI). The estrone derivative 15 was synthesized in 3 steps from (+)-methyl estrone.²² Similarly, compound 18 was generated in 5 steps from 3-epi-andosterone using the sequence developed for accessing 6 (cf. Scheme 2-SI).

Scheme 2.

Oxidation of dienes 6, 13 and 15

Access to pyrone-containing steroids 6, 15, and 18 allowed us to explore the key oxidation with photochemically generated singlet oxygen. Considering that α-pyrone is well-known to undergo [4+2] cycloaddition reactions,²³ a competitive reaction of singlet oxygen with the α-pyrone rather than the Δ^14,16 diene moiety may result in non-productive pathways leading to products 13 and 14 (Figure 2). In addition, the [4+2] cycloaddition may lead to two different diastereomers α-12 and β-12, the products of the addition to the α- and β-faces of the cyclopentadiene ring. Noteworthily, while [4+2] singlet oxygen cycloadditions have been previously explored in the context of steroid synthesis, the undesired α-face addition was primarily observed in most of the successful oxidation reactions.²⁴

Figure 2.

Potential products and side-products of ¹O₂ oxidation

Various conditions were screened for the singlet oxygen cycloaddition and endoperoxide rearrangement of (+)-methyl estrone derivative 15 (cf. Scheme 2). Initial attempts revealed that both diastereomers α- and β- of the endoperoxide were formed, but the desired β-endoperoxide was remarkably unstable at temperatures above −78 °C and rearranged to product 17.^24d,25 This led to a one-pot sequence depicted in Scheme 2 wherein the crude mixture was degassed after the singlet oxygen oxidation step, and the subsequent addition of the catalyst initiated the radical rearrangement to bis-epoxides 16. Preliminary investigations of this one-pot procedure utilized Ru(PPh₃)₃Cl₂ as the rearrangement catalyst,²⁶ which was dissolved in degassed DCM and cooled to −78 °C before addition, but these trials yielded complex mixtures of degraded material with no product. Changing the catalyst for the second step from Ru(PPh₃)₃Cl₂ to cobalt(II) tetraphenyl porphyrin (CoTPP)²⁷ yielded both diastereomers of the bis-epoxide 16 in 15 minutes at −78 °C. Switching the flood lamp white light source for a red Kessil lamp with a maximum emission closer to the maximum absorbance of methylene blue during the singlet oxygen cycloaddition shortened reaction times (2-4 hours compared to 45 minutes) but did not increase yield as it also promoted decomposition due to the large excess of singlet molecular oxygen. The oxidation of 15 resulted in 79% of 16 as the 2:1 mixture of separable α:β diastereomers. Remarkably, subjecting the 3-epi-androestrone derivative 18 to the optimized one-pot oxidation sequence resulted in enhanced β-selectivity, and the formation of 19 was observed in 52% yield with the desired β-19 being the major diastereomer (α:β = 1:1.5). Finally, applying the optimized protocol from above to the actual intermediate in the synthesis of cinobufagin (6) resulted in 64% formation of 8 containing β-8 as the major diastereomer (α:β = 1:1.6). It must be noted that the diastereomeric mixtures of 8 and 19 could be separated chromatographically or advanced forward and separated at a later stage.

The studies above provided streamlined and reproducible access to key intermediate β-8, which enabled our subsequent studies focused on the House-Meinwald rearrangement of 8 into keto-epoxide 20. We anticipated that the β16,β17-epoxide could be selectively activated in the presence of the β14,β15-epoxide moiety due to its higher propensity to form a carbocationic intermediate stabilized by pyrone. This ketone would serve as a functional handle for the completion of the synthesis through reduction, acetylation, and silyl ether deprotection to produce cinobufagin; however, treating bis-epoxide 8 with silica gel,²¹ TMSOTf and 2,6-lutidine,¹⁸ BF₃•OEt₂²⁹ or trifluoroacetic acid did not give the desired product in good yields (Scheme 1-SI). Scandium(III) trifluoromethanesulfonate (10 mol%) was also screened in anhydrous DCM with molecular sieves and was found to give complete conversion into 20 in 10 minutes at room temperature (Scheme 3). The structure of keto-epoxide 20 was confirmed by ¹H and ¹³C NMR following filtration through Celite to remove Sc(OTf)₃, but attempts at isolating the compound by column chromatography resulted in the epimerization of C17.

Scheme 3.

Synthesis of cinobufagenin, 5-epi-cinobufagenin, and analog 22^a

^aReagents and conditions: (a) Sc(OTf)₃ (10 equiv), 4 Å MS, DCM, r.t.; (b) NaBH₄, MeOH/THF(1:1), r.t.; (c) Ac₂O, DMAP, Py, r.t.; (d) HF•Py (70% w/w), THF/Py, r.t., 20 h.

Using keto-epoxide 20 without chromatography did not hinder reduction with sodium borohydride which gave a single isomer in 30 minutes at room temperature. When these conditions were applied to a mixture of keto-epoxides derived from β-8 and α-8, the resulting products could be isolated and their stereochemistry characterized using NOE spectroscopy. This analysis revealed that reduction occurs selectively from the less sterically hindered face to give a single isomer (See Supporting Information). Following work-up and concentration, the new secondary alcohol was immediately subjected to acetylation with distilled acetic anhydride and DMAP in anhydrous pyridine to give TBS-protected cinobufagin 21 in 61% yield over 3 steps. Deprotection with HF•pyridine (70% w/w) in THF at room temperature proceeded cleanly and gave cinobufagin in 94% yield (Scheme 3). Comparison of the spectroscopic and chromatographic data of the synthetic sample with an authentic standard provided a good match and confirmed the final product had been successfully synthesized in 7.6% total yield over 12 steps from DHEA. Using the synthesis of cinobufagin as the template, 5-epi-cinobufagin (22) was synthesized from the bis-epoxide intermediate β-19 in 61% yield over 4 steps.

In summary, we describe the first synthesis of bufadienolide cinobufagin in 12 steps (LLS) and 7.6% overall yield from readily available DHEA as the starting material. This synthesis features expedient installation of the β17-pyrone moiety with β14,β15-epoxide and β16-acetoxy group using a photochemical regioselective singlet oxygen [4+2] cycloaddition followed by CoTPP-promoted in situ endoperoxide rearrangements to provide β,β-bis-epoxide in 64% yield, 1.6:1 d.r. This β,β-bis-epoxide intermediate was subsequently subjected to regioselective scandium(III) trifluoromethanesulfonate catalyzed House-Meinwald rearrangement to establish the β17-configuration. These studies could be used as the platform for the preparation of cinobufagin analogs as demonstrated by synthesizing 5-epi-cinobufagin (22).³⁰

Data Availability Statement.

The data underlying this study are available in the published article and its Supporting Information.