Total Synthesis of Acanthodoral Using a Rearrangement Strategy

We present the second total synthesis of (±)-acanthodoral, a sesquiterpenoid derived from the marine nudibranch Acanthodoris nanaimoensis. Our approach involves a concise three-step transformation from a previously reported compound, resulting in the formation of a less strained precursor of the bicyclo[3.1.1]heptane core and both all-carbon quaternary stereocenters characteristic of the natural product. Notably, this synthetic route incorporates two pivotal steps: a Sm(II)-induced 1,2-rearrangement and a semipinacol rearrangement.

I n 1984, Andersen and co-workers isolated the sesquiterpe- noid acanthodoral (1) along with its two congeners, nanaimoal (2) and isoacanthodoral (3), from the dorid nudibranch Acanthodoris nanaimoensis (Scheme 1a). 1 This sesquiterpenoid mixture was found to have antibiotic activity against Bacillus subtilis and Staphylococcus aureus as well as antifungal activity against Pythiam ultimum and Rhizoctonia solani.1c All three natural products showcased unique carbon structures, 1 with the acanthodorane skeleton being particularly noteworthy for synthetic chemists due to its highly strained bicyclo[3.1.1]heptaneframework.In a retrosynthetic sense, this bicyclic scaffold is planned to arise from the less strained ketone 6 in an already known sequence from the synthesis of Koreeda and co-workers. 2 To access the cis-decalin core of ketone 6 and the desired stereoconfiguration of its two allcarbon quaternary centers, we envisioned making the semipinacol rearrangement with its stereospecific nature and inherent functional group interconversion.This led us back to methylene alcohol 7, which can be traced back to cyclobutano ketone 8 through a 1,2-rearrangement.Ultimately, cyclobutano ketone 8 is expected to result from a photochemical [2 + 2] cycloaddition.
The synthesis toward enone 9 starts with commercially available carboxylic acid 10.Reduction of 10 3 followed by Appel reaction of the corresponding primary alcohol 11 (not shown) readily afforded bromide 12 in 94% yield over two steps. 4Bromide 12 was then exploited as its Grignard reagent to open isobutylene oxide, giving tertiary alcohol 13 in 73% yield. 5 After treatment with poly(phosphoric acid) and cyclodehydration of 13, anisole 14 was obtained in 70% yield alongside minor quantities of its ortho regioisomer. 6Birch reduction 7 and subsequent acidic hydrolysis of crude diene 15 afforded literature-known enone 9 in 73% yield over two steps (Scheme 2).
With enone 9 in hand, we turned our attention to the photochemical [2 + 2] cycloaddition (Scheme 3).Irradiation of enone 9 with condensed allene in methanol at −77 °C readily delivered cyclobutano ketone 8 in 79% yield with a dr of 95:5. 8As anticipated, nuclear Overhauser effects of 8 confirmed the [2 + 2] photocycloaddition to result in the more favored trans-decalin configuration.To set the stage for the semipinacol rearrangement, cyclobutano ketone 8 first needs to be converted to bicyclo[3.2.1]methylene alcohol 7 in a 1,2rearrangement.This was put into practice using a method published by Nagaoka and co-workers in 2017. 8In a refluxing mixture of SmI 2 , tetra-n-butylammonium bromide, and HMPA in THF, cyclopropane species II is generated via ketyl−olefin cyclization and fragmentation, of this cyclopropane moiety delivers the bicyclo[3.2.1]heptane core of methylene alcohol 7 in 79% yield (Scheme 3).
To our delight, subjecting 7 to a refluxing mixture of hydrochloric acid in methanol 9 sufficiently initiated the envisioned type II semipinacol rearrangement. 10Here, the formation of tertiary carbocation III can be conceived, which indicates the antiperiplanar alignment of the shifting bond to the unoccupied p orbital, resulting in the desired semipinacol rearrangement pathway to give tricyclic ketone 6 in 76% yield (Scheme 4).
With α-diazo ketone 18 in hand, irradiation in methanol at −78 °C induced the crucial Wolff rearrangement and afforded methyl ester 19 in 89% yield.Eventually, the reduction of 19 with lithium aluminum hydride and subsequent Dess−Martin oxidation 12 of the resulting primary alcohol 20 yielded acanthodoral (1) in 76% yield over two steps.
Since we strived for a direct way to the natural product and a shorter endgame, we investigated the reduction of the Wolff rearrangement's intermediary ketene.Here THF was the solvent of choice, as it combines a prolonged lifetime of ketene IV and compatibility with reducing agents (Scheme 6).To minimize the risk of overreduction, the use of stoichiometrically defined agents was tested first: While Red-Al gave a mixture of tricyclic primary alcohol 20 and 1, DIBAL-H delivered 1 as a single product but only in traces (both

Organic Letters
determined via NMR).On the other hand, the superstoichiometric use of lithium aluminum hydride reliably afforded tricyclic primary alcohol 20 in 40% yield (for details, see the Supporting Information).The reduction of ketene IV therefore proved possible, but due to low yields of the natural product or mixtures with primary alcohol 20, the endgame by Koreeda and co-workers remains more feasible.
Since the isolation of the pure individual natural products was challenging due to their high volatility and small quantities, 1 and its congeners were isolated as the corresponding (pbromophenyl)urethane derivatives, which also facilitated their structure elucidation. 1 In order to compare the synthesized 1 to the authentic isolated sample, alcohol 20 was treated with pbromophenyl isocyanate in chloroform at 60 °C to afford carbamate 21 in 65% yield (Scheme 7).Eventually, the NMR spectroscopic data of the synthetic material proved to be in agreement with those published by Andersen and coworkers.1b Lastly, enough material was synthesized to investigate the antimicrobial activity of 1.Here, primary alcohol 20 and methyl ester 19 were also examined, since they feature the same characteristic bicyclo[3.1.1]heptanecore.All three compounds were tested against a selection of bacteria and fungi: Saccharomyces pombe, Mucor hiemalis, Candida albicans, Bacillus subtilis, and Staphylococcus aureus (see the Supporting Information).The antibiotic activity against B. subtilis and S. aureus for the isolated sesquiterpenoid mixture (acanthodoral (1), nanaimoal (2), and isoacanthodoral (3)) 1c was reproduced.Using synthetic 1, MICs of 106.7 μg/mL were determined against both strains.However, the lability and oxidation sensitivity of 1, which is readily oxidized to the corresponding carboxylic acid, could lead to an underestimation of its activity.Notably, the aldehyde function was not crucial for antifungal activity, as indicated by the equal or higher potency that alcohol 20 and methyl ester 19 exhibited against S. pombe and M. hiemalis.Of the three, alcohol 20 proved to be the most potent analog, inhibiting the fungi S. pombe and M. hiemalis as well as the bacteria B. subtilis and S. aureus with MICs of 26.6, 13.3, 26.6, and 26.6 μg/mL, respectively.
In conclusion, we accomplished the total synthesis of acanthodoral (1) in eight steps (12% yield) from literatureknown enone 9 and in 14 steps (4.8% yield) from commercially available carboxylic acid 10.The semipinacol rearrangement was used to take advantage of the transition from the more easily accessible trans-decalin to the soughtafter cis-decalin of ketone 6.This provides rapid access (three steps) to the natural product's two quaternary centers embedded in a less strained precursor of the bicyclo[3.2.1]heptane core.From enone 9, only two steps�a [2 + 2] photocycloaddition and Sm(II)-induced 1,2-rearrangement� were needed to pave the way for the semipinacol rearrangement.With the direct reduction of the ketene generated by the Wolff rearrangement, a shorter variant of Koreeda's endgame was explored.However, the endgame reported by Koreeda and co-workers 2 was eventually followed and remained unchanged.Furthermore, biological assays demonstrate the antifungal and antibacterial activities of the bicyclo[3.1.1]heptanes,which are not dependent on the aldehyde function of 1.