Convenient Access to a Strained Bicyclic Enone: A Concise and Improved Formal Synthesis of Ineleganolide

Abstract

The norcembranoid and cembranoid diterpenoids represent an intriguing class of natural products isolated from marine sources. Their chemical synthesis has been a challenging and exciting field of research over the past two decades, owing largely to their structural complexity. We recently disclosed a total synthesis of a member of this class, ineleganolide, in a 23 step longest linear sequence. In search of a shorter, more efficient route, we have devised a new strategy for the synthesis of a key bicyclic enone. Disclosed herein is our improved synthesis of this strained intermediate, completing the formal synthesis of ineleganolide in only 14 steps, thereby shortening our previous synthesis by 9 steps.

Dedicated to Prof. Thomas Maimone on his receipt of the Tetrahedron Young Investigator Award 2024.

Graphical Abstract

To create your abstract, type over the instructions in the template box below.

Fonts or abstract dimensions should not be changed or altered.

Historically, natural products have played an integral part in the advancement of synthetic organic chemistry and drug discovery over many decades.¹ The field of total synthesis continues to inspire new reaction development to this day and has shown its impact in the wider fields of chemical and biological sciences.² In this regard, the polycyclic furanobutenolide-derived cembranoid and norcembranoid diterpenoids have gained a considerable amount of attention from the synthetic community.³ Isolated from marine sources, these natural products exhibit interesting bioactivity, often in the form of cytotoxicity and/or neurotoxicity.⁴ Structurally, the cembranoid and norcembranoid diterpenoids entail very congested, polycyclic carbon frameworks, with intricate and irregular oxidation patterns. These features can be seen in the structures of horiolide (1) or verrillin (2), that have yet to be successfully synthesized by means of synthetic organic chemistry (Figure 1).^5,6

Figure 1.

Representative members of the complex norcembranoid and cembranoid diterpenoid natural products.

In recent years however, several members of this class have been the subject of successful synthesis campaigns, in turn fueling the development of new reaction methodology.⁷ As such, our group has had a long-lasting interest to synthesize ineleganolide (3), with only recent success in accomplishing its total synthesis in 2023.⁸ Structurally, ineleganolide possesses a unique [5–7-6]-tricyclic carbon core, with a bridging 5-membered lactone and β-keto tetrahydrofuran. Due to its rigid framework, ineleganolide (3) presents a significant challenge in retrosynthetic planning and bond disconnection. The groups of Pattenden and Wood have successfully demonstrated biomimetic strategies toward the synthesis of 3, as demonstrated in a semi-synthesis and the first total synthesis, respectively.^9,10 Our group’s initial approach to 3 was thwarted by the inability to install the bridging ether and β-keto tetrahydrofuran, faced with the challenge of overcoming unfavorable strain via this disconnection.¹¹ Learning from these previous efforts, our successful synthesis of 3 was achieved utilizing the bicyclic enone 4 and carboxylic acid 5 as our key intermediates (Figure 2, A). In a convergent fashion, we are able to esterify these two fragments and construct the seven-membered ring at a late stage to finish the total synthesis of 3 in a concise seven step sequence. This sequence included unique chemical transformations such as a Michael-addition and aldol cascade, an O₂-mediated allylic oxidation, and samarium diiodide induced semi-pinacol shift. Despite the late-stage efficiency in accessing our target, we realized that of the 23 steps in our synthesis, over half were dedicated to accessing enone 4 as starting material. While carboxylic acid 5 was previously known and easily accessible from (S)-norcarvone;^11,12 synthesizing intermediate 4 proved to be rather challenging initially. In this report however, we showcase a new route to access enone 4 in only seven steps and two protecting group manipulations. This represents a significant improvement from the previous 16 step sequence, enabling formal total synthesis of 3 that is increasingly step-efficient and prospective for delivering material for further biological testing.¹³

Figure 2.

(A) Recently completed total synthesis of 3 from starting materials 4 and 5. (B) Previous and new strategy for the synthesis of bicylic enone 4 (PG = protecting group).

One of the challenges in constructing building block 4 is its highly strained structure. Presumably, the planar conformation of the enone required for π-orbital overlap induces significant strain into the fused [5–5] ring system. The shorter C–O bond lengths of the embedded tetrahydrofuran moiety serve to further intensify this effect. In our hands, enone 4 often behaved as the chemical equivalent of a wound-up spring. As an exceptional Michael-acceptor, we observed 1,4-addition with numerous nucleophiles or rapid decomposition, even under mild reaction conditions. However, when the allylic alcohol was protected as a bulky silyl ether, the compounds stability was significantly improved. In early investigations we had set out with several different approaches that proved unsuccessful. Thus, synthesis of this intermediate required careful orchestration to generate such a strained and reactive compound.

Our initial route to 4 built on previous reports, taking advantage of (R)-linalool as chiral starting material. We employed a protocol popularized by the Maimone group to construct the cyclopentene, then disconnecting the C–O bond as the final step.¹⁴ Although this route allowed us to introduce chirality at the tertiary alcohol early in the sequence, the excessive use of protecting groups and redox manipulations led to a lengthy sequence of 16 steps to get to compound 4. In looking for a better route, we considered how we could simultaneously construct the enone and β-keto tetrahydrofuran. We imagined that retrosynthetic disconnection of the C-sp²/C-sp² bond through an intramolecular 1,2-addition could lead to a more direct synthesis of 4. We hypothesized that collapse of the tetrahedral intermediate would provide a sufficient thermodynamic driving force to induce the strain of the enone moiety. Ultimately this proved to be a successful strategy, reducing the length of our route to only seven steps from commercial starting materials, employing only two protecting group (PG) manipulations.

Our sequence was initiated in targeting the intermediate tertiary alcohol 9, which was previously known in the literature.^7,15 Starting from furfuryl alcohol, we can readily display cyclopentenone 6 via the known Piancatelli-rearrangement (Scheme 1).¹⁶ Due to the ease of access, we set out with the racemic mixture of compound 6 to investigate the forward route. The allylic alcohol of 6 can be converted to the silyl ether 7. Subsequent α-iodination of the enone occurred in satisfactory yield to give vinyl iodide 8.¹⁷ Next, we preformed 1,2-addition using methyl magnesium bromide in good yield. Under cryogenic conditions, slowly warming the reaction mixture, we could achieve an excellent diastereoselectivity of 16:1, favoring the syndiol 9. With the tertiary alcohol in place, we sought to alkylate using methyl chloroacetate. Deprotonation with sodium hydride was successful, isolating compound 10 in 21% yield, with 69% recovered starting material. In trying to improve this yield, we evaluated numerous solvents and different conditions, such as K₂CO₃/acetonitrile, potassium hydride/18-crown-6 or methyl bromoacetate as electrophile for example. Unfortunately, most of these conditions led to no product formation and complete decomposition of the starting material. It is noteworthy to point out that we had initially synthesized the analogous vinyl bromide 12 (see SI), where lithium-halogen exchange and 1,2-addition occurred to give product 11 successfully, yet in a diminished 18% yield (Scheme 2). Assuming that more rapidly occurring lithium-halogen exchange could boost product formation, we had turned to vinyl iodide 10. Indeed, we saw a dramatic increase in product formation, as our desired enone 11 was isolated in 81% yield. Deprotecting with our previously developed conditions,⁸ we can now access the racemic building block 4 in only seven steps from commercial starting materials.

Scheme 1.

New synthetic route to access byclic enone 4 from cyclopentenone 6.

Scheme 2.

Preliminary cyclization attempt of vinyl bromide 12.

In providing enantioenriched material, the new route benefits greatly from starting with the common cyclopentenone 6. There are several reports known to access the chiral allylic alcohol (S)-6 (Figure 3). Highlighting two of the potential reactions to achieve asymmetric induction, Noyori and co-workers reported the desymmetrizing 1,2-reduction of the corresponding dicarbonyl to isolate (S)-6 in a 60% yield and 94% ee.¹⁸ As an alternative, the enzymatic kinetic resolution of the racemic ester gives (S)-6 in >99% ee and 44% yield.¹⁹ This readily accessible chiral starting material could be converted to chiral enone (–)-4 in a stereoselective fashion, providing enantioenriched material.

Figure 3.

(A) Opportunity for the stereoselective synthesis of chiral enone (–)-4. (B) Two potential pathways to access enantioenriched (S)-6.

In conclusion we have developed a new route to access the highly strained enone 4, resulting in a significant step count reduction and concise formal synthesis of ineleganolide in only 14 steps. Employing a lithium-halogen exchange and high-yielding intramolecular 1,2-addition, we showcased how efficient disconnections and thoughtful introduction of strain can significantly reduce the synthetic load of a total synthesis. We hope this report benefits synthetic organic chemists in the field of oxidized terpene synthesis and contributes to synthetic logic of small, but ever so sophisticated, organic building blocks.