Enantioselective Synthesis of Caprolactam and Enone Precursors to the Heterocyclic DEFG Ring System of Zoanthenol

Abstract

The enantioselective synthesis of both caprolactam and enone synthons for the DEFG ring system of zoanthenol are described. The evolution of this synthetic approach proceeds first through a synthesis using the chiral pool as a starting point. Challenges in protecting group strategy led to the modification of this approach beginning with (±)-glycidol. Ultimately, an efficient approach was developed by employing an asymmetric hetero-Diels-Alder reaction. The caprolactam building block can be converted by an interesting selective Grignard addition to the corresponding enone synthon. Addition of a model alkyne provides support for the late-stage addition of a hindered alkyne into the caprolactam building block.

Graphical Abstract

The syntheses of two synthons for the heterocyclic core of the marine alkaloid zoanthenol are described. These fragments can be generated in a divergent manner from the product of an enantioselective hetero-Diels-Alder reaction. Judicious selection of protecting groups allows control of a key Grignard reaction to form the enone synthon. The feasibility of the caprolactam as a synthon is validated.

Introduction

Zoanthenol (1) is a complex, polycyclic alkaloid belonging to the zoanthamine family of natural products.¹ These compounds exhibit a range of biological activities including anti-osteoporotic, anti-inflammatory, cytotoxic (P-388 murine leukaemia), and antibacterial activity. ² Significant synthetic efforts toward the zoanthamines have been disclosed by a number of research groups, including the syntheses of norzoanthamine and zoanthenol by Miyashita in 2004 and 2009³ and norzoanthamine by Kobayashi in 2008. ⁴ Our efforts toward the zoanthamines has focused on zoanthenol (1), which features an unusual oxidised aromatic A ring with collagen-selective anti-platelet aggregation activity. ⁵ In addition to Miyashita’s total synthesis of zoanthenol (1), both Hirama and co-workers ⁶ and our own group ⁷ have reported advanced strategies for its completion. In this communication, we disclose the enantioselective synthesis of substituted caprolactam and enone precursors, which enable two potential routes toward the synthesis of the DEFG ring system of zoanthenol.^8,9

With seven rings and nine stereocenters, zoanthenol is a densely functionalised, topographically complex target molecule. Our initial simplifying disconnection involved unravelling the heterocyclic bis-hemiaminal portion of zoanthenol (1 ⇒ 2) based on the pioneering work of the Kobayashi and Williams groups (Scheme 1).¹⁰ We envisioned that the resulting tethered side chain could be severed at either the C(8)–C(9) bond to reveal tricyclic core 3 and enone 5 or at the C(6)–C(7) bond to unveil carbocyclic core 4 and caprolactam 6.

Scheme 1.

Retrosynthesis of zoanthenol.

It was anticipated that enone 5 could be derived directly from caprolactam 6, thus allowing entry into either synthetic approach from a single DEFG-synthon (Scheme 2). Caprolactam 6 was disconnected across the amide C–N bond to reveal Weinreb amide 7. Phthalimide 7 could in turn be derived from δ-lactone 8, accessible from α,β-unsaturated lactone 9.

Scheme 2.

Retrosynthetic analysis of DEFG synthons.

Results and Discussion

Our synthetic efforts began by targeting a lactone such as 9 in enantioenriched form. Initially, we investigated an approach beginning from tri-O-acetyl-D-glucal 10. Unsaturated lactone 11 was accessed in good yield by PCC oxidation according to known methods (Scheme 3).¹¹ The extraneous acetate was removed by reduction with activated Zn dust in acetic acid followed by reconjugation upon treatment with catalytic DBU to provide 12.¹¹ Following removal of the acetate group, our initial attempts at protection with a more suitable group were unproductive because intermediates related to 12 were very sensitive to base. However, mild acidic conditions for benzyl protection ultimately were developed to provide δ-lactone 14 using trichloroacetimide 13.¹²

Scheme 3.

Synthesis of a chiral unsaturated δ-lactone

Despite the initial appeal of utilizing the chiral pool as the starting point for the synthesis of enone 14, the necessity to shuffle protecting groups prompted the exploration of a route starting from racemic glycidol 15 (Scheme 4). The enantioselective route could be secured upon completion of the racemic route from readily available (S)-glycidol.^13,14 According to literature preparations, the sequence began with benzyl protection of racemic glycidol 15, followed by nucleophilic epoxide opening with the anion of ethyl propiolate (16) to provide known alkyne 17.¹⁵ Unsaturated lactone 14 was quickly accessed through Lindlar reduction of alkyne 17, followed by cyclization upon exposure to mild acid. With suitably protected lactone 14 in hand, a highly diastereoselective cuprate addition with the Gilman reagent proceeded smoothly, yielding scaleable quantities of saturated lactone 18 as a single observed diastereomer.^13a Installation of the primary amine was accomplished through hydrogenolysis of the benzyl ether, followed by Mitsunobu reaction with phthalimide, providing the crystalline intermediate 7.

Scheme 4.

Access to Weinreb amide 7 from (±)- or (S)-glycidol.

Ultimately, we were able to access δ-lactone 9 more directly by employing the hetero-Diels-Alder catalyst developed by Jacobsen and co-workers (21, Scheme 5). ¹⁶ Thus, following reaction of diene 19, aldehyde 20, and catalyst 21, desired dihydropyran 22 was isolated in 72% yield and >99% ee and could be converted to the necessary lactone (9) using acidic pyridinium dichromate conditions. At this point, selective 1,4-addition was accomplished by treatment of 9 with Gilman’s reagent to afford 23 as a single diastereomer.¹⁷ Treatment with acidic resin induced desilylation to provide alcohol 24, and subsequent Mitsunobu reaction provided phthalimide derivative 8.¹⁸

Scheme 5.

Toward a catalytic asymmetric synthesis of synthons 5 and 6.

Chiral lactone 8 was then treated under standard conditions for Weinreb amide formation, and the intermediate alcohol was immediately trapped by addition of TBSOTf and 2,6-lutidine to yield Weinreb amide 7 (Scheme 6). Treatment of 7 with hydrazine hydrate in refluxing ethanol revealed the free primary amine, which spontaneously cyclised to form a caprolactam. Carbamate formation with Boc-anhydride provided caprolactam synthon 6. The final step in accessing enone synthon 5 was to add a single vinyl equivalent to the Boc-protected caprolactam. Thus, treatment of 6 with vinyl magnesium bromide provided isolable Grignard adduct 25. The chelation of Mg²⁺ between the Boc carbonyl and the amide carbonyl encourages addition of a single equivalent of the nucleophile, and we anticipate that a similar hydrogen-bonding event slows the collapse of hemiaminal 25. Upon standing in CHCl₃, desired enone 5 is produced.

Scheme 6.

Conversion of the δ-lactone to the ε-lactam and enone synthons.

In order to determine the feasibility of the addition of hindered alkyne 4 into caprolactam 6, a model alkyne was synthesised. In order to generate a single diastereomer of the addition product, it was necessary to generate the model alkyne as a single enantiomer. Fortunately, α-quaternary allyl ketone 26 was readily available using our asymmetric alkylation methodology¹⁹ and could be advanced to a suitable model system (Scheme 7). Thus, allyl ketone 26 was smoothly isomerised to the internal olefin, which was then ketalised to provide olefin 27. Ozonolysis with mild reductive workup allowed access to desired aldehyde 28. Treatment with the Ohira-Bestman reagent (29) induced sluggish Gilbert-Seyferth homologation to afford alkyne 30 and recovered aldehyde 28. Deprotonation of the alkyne with KHMDS and trapping with caprolactam 6 provided alkynone 31. Hydrogenation of the alkyne readily provided the final side-chain appended model product 32. This unoptimised approach provides a key proof-of-concept supporting the challenging disconnection of tethered tricycle 2 to carbocyclic core 4 and caprolactam 6.

Scheme 7.

Functionalization of a model ketone with caprolactam 6.

Conclusions

In summary, these studies constitute the synthesis of two fully-functionalised, enantiopure DEFG synthons (5 and 6) ready for late-stage coupling with our carbocyclic core structures (3 and 4). The synthetic approaches described encompass strategies beginning from a chiral glycal, racemic or enantiopure glycidol, and ultimately, a catalytic enantioselective approach employing a hetero-Diels-Alder reaction followed by a diastereoselective conjugate addition. Additionally, selective ring opening and ring closing events allow for the efficient elaboration of the key δ-lactone. The strategic choice of Boc as the amide protecting group enables selective mono-addition of vinyl magnesium bromide, and ultimately a neopentyl alkyne, into the caprolactam. Efforts to combine these synthons with carbocyclic core structures analogous to 3 and 4 are ongoing.