Palladium-Catalyzed Decarbonylative Dehydration for the Synthesis of α-Vinyl Carbonyl Compounds and Total Synthesis of (−)-Aspewentin A, B, and C

Abstract

The direct α-vinylation of carbonyl compounds that forms a quaternary stereocenter is a challenging transformation. We discovered that δ-oxocarboxylic acids can serve as masked vinyl compounds and be unveiled by palladium-catalyzed decarbonylative dehydration. The carboxylic acids are readily available through enantioselective acrylate addition or asymmetric allylic alkylation. A variety of α-vinyl quaternary carbonyl compounds are obtained in good yields, and an application in the first enantioselective total synthesis of (−)-aspewentin A, B, and C is demonstrated.

An all-carbon quaternary center bearing an ethylene substituent is a common structural motif in many natural products (see Figure 1).^[1] An important approach to the construction of this unit is the α-vinylation of carbonyl compounds, and two general methods have been developed. One is the direct coupling of an enolate nucleophile with a vinyl electrophile such as an alkenyl ether,^[2] vinyl bromide,^[3] or acetylene itself.^[4,5] Although this approach can be extended to alkenylations as well, and asymmetric versions are known,^[3a,3c,4c] the scope of the enolate nucleophile is generally limited to 1,3-dicarbonyl compounds^[4] or those with only one enolizable position.^[2,3,5b] A second tactic involves addition of the enolate nucleophile to a vinyl surrogate such as vinyl sulfoxide,^[6] (phenylseleno)acetaldehyde,^[7] or ethylene oxide,^[8] followed by elimination. However, there are few reports of stereoselective additions that form the quaternary stereocenter,^[9] and none are catalytic or enantioselective. Due to these constraints, even the simplest 2-methyl-2-vinylcyclohexanone (7) is not known as a single enantiomer in the literature.

Figure 1.

Natural products with ethylene substituents.

We envisioned an alternative approach to access α-vinyl carbonyl compounds, by employing a decarboxylative elimination of δ-oxocarboxylic acids (Scheme 1). These acids may be prepared by numerous methods, including addition of an enolate nucleophile to an acrylate acceptor^[10] and palladium-catalyzed allylic alkylation.^[11] Importantly, these two methods allow for the enantioselective construction of the requisite quaternary stereocenter.

Scheme 1.

Decarboxylative approach to vinylation.

Recently, we reported on the palladium-catalyzed decarbonylative dehydration of fatty acids to form terminal olefins.^[12] Due to the importance of α-vinyl carbonyl compounds and the challenges in their preparation, we became interested in applying our decarbonylative dehydration chemistry as an alternative strategy to α-vinylation. Since the carboxylic acid, which bears a quaternary center two atoms away from the reactive carboxyl group, is more hindered than a simple fatty acid, we expected that the reaction conditions would need to be tuned for this particular class of substrates. Additionally, from a practical standpoint, our previous studies were typically conducted on ~5 g fatty acid substrate without solvent, under vacuum distillation conditions. Thus, a smaller scale alternative for implementation on laboratory scale and in the context of multistep organic synthesis would need to be developed.

At the outset of our investigation, we prepared carboxylic acid 8a and subjected it to modified palladium-catalyzed decarbonylative dehydration conditions, when slightly higher loadings of catalyst, ligand, and additive were employed (Scheme 2). We were pleased to isolate vinyl cyclopentanone 9a in 67% yield.^[13] This result demonstrated that steric bulk at the quaternary center does not significantly retard the reaction, but proximal functionality (e.g. the ketone) could alter the reaction pathway.

Scheme 2.

Decarbonylative dehydration of δ-oxocarboxylic acid 8a.

With this exciting initial result in hand, we proceeded to investigate the scope of the reaction (Table 1). First, we synthesized (R)-3-(1-methyl-2-oxocyclohexyl)propanoic acid (8b) by enantioselective d’Angelo Michael addition,^[10] and subjected it to decarbonylative dehydration (entry 2). We were delighted to obtain the desired product, (R)-2-methyl-2-vinylcyclohexanone (ent-7), in 60% yield and 92% ee. Likewise, 2-ethyl-2-vinylcyclohexanone (9c) was prepared in a similar fashion. Carboxylic acids bearing allyl or 2-methallyl substituents, which can be prepared via palladium-catalyzed allylic alkylation,^[11] also underwent decarbonylative dehydration smoothly to provide the corresponding 2-allyl-2-vinyl-substituted cyclohexanones 9d and 9e (entries 4 and 5), the latter in 92% ee from enantioenriched acid 8e. It is worth noting that double bond isomerization in the allyl moiety is negligible for 9d and does not occur at all for 9e. Acylic keto acid 8f is converted to acyclic ketone 9f in good yield (entry 6). Aside from keto acid substrates, we examined acids bearing other types of carbonyl functionalities (entries 7–10), and found that α-vinyl ester 9g, lactam 9h, and aldehyde 9i can all be prepared in good yields. More complex scaffolds such as acid 8j, obtained by oxidative cleavage of testosterone,^[14] also undergo the reaction to provide vinylated tricycle 9j (entry 11). While the reaction can be carried out in the absence of a solvent at a fairly large scale (5 mmol, entries 1–7), we found that for smaller scale synthesis it is more convenient to use N-methylpyrrolidinone (NMP) as solvent along with slightly modified conditions (entries 8–11).^[15]

Table 1.

Decarbonylative dehydration of δ-oxocarboxylic acids.

^[a]5 mmol scale.

^[b]Isolated as a 95:5 mixture of desired product and internal olefin isomer.

^[c]Condition: 0.5 mmol substrate (1 equiv), benzoic anhydride (1.2 equiv), PdCl₂(nbd) (1 mol%), Xantphos (1.2 mol%), NMP (0.25 mL), 1 atm N₂, 132 °C, 3 h.

To further demonstrate the utility of our decarbonylative dehydration approach to vinylation, we embarked on a total synthesis of aspewentin B (1, Figure 1), a norditerpene natural product isolated from Aspergillus wentii.^{[16, 17]} This terpenoid contains an α-vinyl quaternary cyclohexanone scaffold, and is therefore ideally suited for our chemistry. Retrosynthetically, we envisioned that the vinyl group could be formed by decarbonylative dehydration of δ-oxocarboxylic acid 11, which might be obtained by elaboration of allyl ketone 12.^[18] The quaternary stereocenter would be set by palladium-catalyzed enantioselective allylic alkylation of tricyclic ketone 13, which could be constructed from known aryl bromide 14.^[19]

We commenced our total synthesis by copper-catalyzed coupling of a Grignard reagent derived from aryl bromide 14 with ethyl 4-iodobutyrate (Scheme 4).^[20] α-Methylation of the coupled product produces ester 15, which was saponified and then cyclized to form tricyclic ketone 13. The ketone was converted to the corresponding allyl enol carbonate (16), and subjected to palladium-catalyzed enantioselective decarboxylative allylic alkylation to afford allyl ketone 12 in nearly quantitative yield and 94% ee. Interestingly, the allylic alkylation reaction proceeds efficiently with low palladium catalyst loading, employing a new catalytic protocol recently developed by our group.^[21] Only 2.7 mg Pd(OAc)₂ is needed for a reaction of 1.42 g starting material (16) to deliver 1.25 g allyl ketone product (12). Hydroboration/oxidation of the terminal olefin of 12, followed by further oxidation, delivers carboxylic acid 11 in 73% yield.^[18] Gratifyingly, palladium-catalyzed decarbonylative dehydration^[22] furnishes α-vinyl ketone 17 in 93% yield.^[23] Removal of the O-methyl group provides (−)-aspewentin B (1) in 78% yield.^[24] Thus, (−)-aspewentin B (1) was synthesized in 9 steps and 25% overall yield from known starting materials. With 1 in hand, two other natural products of the aspewentin family can also be accessed. Reduction of the ketone moiety of 1 furnishes (−)-aspewentin A (18) in 85% yield.^[25] Oxidation of the phenol in 18 affords (−)-aspewentin C (19) in 13% yield and (+)-10-epi-aspewentin C (20) in 7% yield, over 2 steps.^[26]

Scheme 4.

Total synthesis of (−)-aspewentin A, B, and C.

In summary, we have developed a new approach to access α-vinyl quaternary carbonyl compounds via palladium-catalyzed decarbonylative dehydration of δ-oxocarboxylic acids. A variety of acids with different scaffolds and functional groups can be transformed into the corresponding α-vinyl carbonyl compounds without erosion of the stereointegrity of the neighboring quaternary center. We have also applied the method to the first enantioselective total synthesis of (−)-aspewentin A, B, and C. Further applications of this transformation in natural product synthesis are currently ongoing in our lab and will be reported in due course.

Experimental Section

See Supporting Information for experimental details.

Scheme 3.

Retrosynthetic analysis of (−)-aspewentin B.