Total Synthesis of (±)-Aplykurodinone-1: Traceless Stereochemical Guidance

Abstract

The total synthesis of the highly degraded steroidal natural product, aplykurodinone-1 (1), has been accomplished. Key features include a one-flask hydrolysis/retro-aldol/iodolactonization sequence to excise the C₈ hydroxymethylene functionality with retention of stereochemistry, and the stereoselective installation of the C₁₃ methyl group through hydrogenation with homogeneous catalyst.

Of all of the classes of natural products, steroids have perhaps had the most enduring role in prompting new ideas in total synthesis.¹ A not insignificant body of organic chemistry has been discovered in the context of pursuits directed to the total synthesis of steroids. Apparently, this interest continues to the present day.²

One of the vexing problems in steroid total synthesis is that of exercising control of the configuration at C₂₀.³ The challenge is that of correlating the configuration of the presumably “freely rotating” C₂₀, with the resident stereochemistry of the polycyclic domain. Though ingenious solutions to the “C₂₀ problem” have been advanced, it is still an interesting and thought provoking matter.

It was in this setting that we took particular note of the aplykurodines, isolated from marine mollusks of the genus Aplysia.⁴ This family can well be viewed as a rare class of highly degraded marine steroids, where C₁₃ in the aplykurodine setting corresponds to C₂₀ in wild-type steroids. Also of note is the cis fusion of the hydrindane sector in contrast to the trans C-D fusion of hydrindanes, which is so prevalent in fully mature steroids.⁵

We note that various aplykurodines exhibit cytotoxic activity against a range of human cancer cell lines.⁶ Notwithstanding the potential elements of chemistry and medicinal chemistry-based points of interest, surprisingly little attention at the level of synthesis has been directed to the aplykurodines. Our laboratory took particular note of a recently isolated member of the family, aplykurodinone-1 (1).⁷ Following its isolation from the sea hare Synphonota geographica in 2005, the structure of aplykurodinone-1 (1) was elucidated through a combination of spectroscopic methods, X-ray crystallography, and chemical correlation. As shown in Scheme 1, 1 was found to possess a cis-fused C-D ring with epimeric C₈ (steroidal numbering) and an unsaturated side chain (as compared to cholesterol).

Scheme 1.

Synthetic Strategy Toward Aplykurodinone-1 (1).

On the basis of considerations advanced above, a total synthesis effort directed to this “degraded” steroid was undertaken. Below is described the inaugural total synthesis of (±)-1, resulting from this initiative. As will be seen, the program potentially encompassed several chemistry-level initiatives of broader heuristic value.

Using ideas stemming from analysis by pattern recognition,⁸ we envisioned that, in principle, the hypothetical hydrindanone 4 contains an array of coordinated functionality, which could well enable access to the eventual target system, 1. The plan envisioned halolactonization as the instrumentality for fusing the γ-lactone to the cis-fused hydrindanone, thereby creating a ladder-like topography. Accordingly, we needed to gain secure control over the configurational relationship between strategic carbons 3 and 7 in 1 (vide infra). It occurred to us that an otherwise unnecessary quaternary function at C₈ could orchestrate the required relationship of the angular methyl group at C₇ with a soon-to-be-installed acetic acid moiety at C₃ (see structures 5 and 6). Iodolactonization leading to the γ-lactone could then follow in the wake of excision of the now traceless C₁ fragment.

Given the general line of thinking adumbrated above, and a long-term proclivity for Diels-Alder reactions, we hoped to advance along such lines. In the event, it was discouraging to find that attempted cycloaddition between 2⁹ and 3¹⁰ failed under various thermal and Lewis acid-mediated conditions. However, reaction of 2 with methyllithium generated an operational lithium enolate,¹¹ which did undergo effective cycloaddition with 3, thereby providing a 73% yield of 4. It is well recognized that realization of the required cycloaddition does not, in and of itself, inform as to the limiting, mechanistic, extremi (two sequential Michael additions or an anionically mediated Diels-Alder reaction).¹²

The enone was regioselectively masked as a thioketal.¹³ The C₉ ketone was then reduced, as shown, and the resultant alcohol protected as a MOM ether. Upon reduction of the ester and removal of the dithiane ketal,^13a,14 compound 10 was in hand. The primary alcohol was next converted to its diazoacetyl derivative, 11. As shown in Scheme 2, intramolecular cyclopropanation onto the electron-deficient enone afforded the activated cyclopropane, 5, though only in moderate yield (40%).¹⁵

Scheme 2.

Synthesis of 6 En Route to Aplykurodinone-1 (1).

^aKey: (a) MeLi, DME/THF, −50 °C; TFA, CHCl₃, reflux, 73%; (b) HSCH₂CH₂SH, AcOH, p-TSA, 74%; (c) NaBH₄, CH₂Cl₂/MeOH, −78 °C→ −40 °C; (d) MOMCl, iPr₂NEt, CH₂Cl₂, 84% over 2 steps; (e) DIBAL-H, CH₂Cl₂, −78 °C; (f) Tl(NO₃)₃, MeOH/THF/H₂O, 95% over 2 steps; (g) TsNHN=CHCOCl, PhNMe₂, Et₃N, CH₂Cl₂, 0 °C, 88%; (h) bis-(N-t-butylsalicylaldiminato) copper (II), PhMe, reflux, 40%; (i) NaBH₄, CeCl₃•7H₂O; (j) I₂, PPh₃, imidazole, PhMe, reflux, 78% over 2 steps; (k) SmI₂, THF/MeOH, 85%; (l) HCl, THF/H₂O; (m) DMP, CH₂Cl₂, 80% over 2 steps.

The next phase entailed the opening of the cyclopropane ring, with concomitant installation of unsaturation at C₄–C₅. Toward this end, the C₅ ketone was converted to an iodide through a two-step sequence, commencing with Luche reduction,¹⁶ which provided a 2:1 (β:α) epimeric mixture of alcohols.¹⁷ These compounds were readily converted to iodide 12. In the event, upon exposure of 12 to SmI₂,¹⁸ it readily suffered the hoped-for reductive cyclopropane opening, to deliver lactone 13 in good yield (85%). The latter was converted to 6 in a straightforward fashion, as shown.

The otherwise extraneous C₈ oxymethylene functionality had indeed been exploited as a very useful C₃ stereochemical linchpin (vide supra). Having served its function in mediating the stereoselective installation of the C₃ center, the C₁ fragment was cleaved via a hydrolysis–retro-aldol sequence following exposure of 6 to the action of K₂CO₃ in water at 100 °C. Subsequent iodolactonization¹⁹ of the resultant carboxylic acid (14) was accomplished in the same flask, to provide 15 in 75% yield from 6. The re-emergence of the cis-junction was confirmed through NOESY analysis of the methyl ester derivative of 14 and by X-ray diffraction of the subsequent deiodinated intermediate, 16.²⁰ Finally, Saegusa oxidation²¹ of 16 afforded 7, presenting the tricyclic core of aplykurodinone-1, in the sense shown.

With compound 7 in hand, the synthesis entered its terminal phase, i.e. installation of the C₁₁ side chain. A stepwise strategy was set in place. As outlined in Scheme 4, BF₃-mediated conjugate addition of the cuprate derived from vinyl bromide 17²² was accomplished with a high degree of facial selectivity (10:1 dr) to provide intermediate 18. Given the difficulties surrounding the heterogeneous hydrogenation of the C₂₀–C₂₂ olefin as a means of installing the C₂₀ stereocenter of trans-fused mature steroids,²³ it was hoped that homogeneous hydrogenation protocols may be more productive.²⁴ Thus, following exposure to Crabtree catalyst,²⁵ under an atmospheric pressure of H₂, the tri-substituted olefin was reduced in a diastereoselective fashion (>5:1) to furnish compound 19, which bears all of the stereocenters of aplykurodinone-1 (1). A single crystal of this advanced intermediate was obtained, and X-ray crystallographic analysis was found to support the proposed structure. Elaboration of the side chain was accomplished in a straightforward fashion, as shown. Thus, cleavage of the TBS group²⁶ and subsequent oxidation²⁷ delivered aldehyde 20, which readily underwent modified Julia olefination²⁸ to furnish the target compound, aplykurodinone-1 (1). Its spectroscopic properties were the same as those reported for the isolated natural product. In the course of these synthetic studies, we also investigated a different route for the emplacement of the C₁₁ side chain. The fully elaborated dienyl fragment could be appended to the core (7) in a single step. It was hoped that the intermediate thus obtained could then be converted, in one step, to aplykurodinone-1 (1), through regio- and stereo-selective hydrogenation. In the event, the conjugate addition of 21²⁹ to 7 was accomplished, to provide 22 in moderate yield. Upon exposure to Wilkinson’s catalyst³⁰ under atmospheric H₂, the 13-epi-aplykurodinone-1 isomer (23) was selectively formed (>6:1 dr). Perhaps in each of the two hydrogenation transition states (18→19; 22→23), the double bond projects from the β-face of the molecular backbone. The bulky catalyst then approaches the olefin from the unhindered of the olefin, to provide the observed stereochemical outcome.³¹

Scheme 4.

^aKey: (a) 17, t-BuLi, CuCN, BF₃•OEt₂, −78 °C→ −98 °C, 73%, 10:1 dr; (b) Crabtree catalyst, H₂, CH₂Cl₂, 75% conversion, 50% yield (pure); (c) HF, CH₃CN/THF; (d) DMP, CH₂Cl₂, 87% over two steps; (e) THF, LiHMDS, −78 °C, 68%.

In summary, the total synthesis of aplykurodinone-1 (1) has been accomplished. A key feature of the effort involved an anionically-mediated cycloaddition of a metallo-enolate derived from 2 with 3 (see formation of 4). The seemingly extraneous oxymethyl C₁ function at C₈, used to govern the configurational relationship between C₃ and C₇. The C₈ function is ultimately excised. The stereochemistry at C₁₃ is managed with good stereoselectivity based on suitable order of introduction of H and Me to this carbon.

Scheme 3.

^aKey: (a) K₂CO₃, H₂O, 100 °C; (b) NIS, CH₂Cl₂, 75%; (c) Ra-Ni, CH₂Cl₂/EtOH, 90%; (d) TMSOTf, Et₃N, CH₂Cl₂, 0 °C; Pd(OAc)₂, CH₃CN, 76% (90% brsm).

Scheme 5.

^aKey: (a) 17, t-BuLi, CuCN, Et₂O, −78 °C, 51%; (b) Wilkinson catalyst, H₂, benzene, 67%.