A Formal Synthesis of the Neem Limonoid Azadirachtin

            Azadirachtin exhibits potent anti-feedant and growth-disruptant properties against a broad spectrum of insects. In addition, its low mammalian toxicity has stimulated commercial applications of formulated azadirachtin, obtained from Neem extract, as a safe biopesticide. An oxidative rearrangement product derived from the limonoid (tetranortriterpenoid) system, azadirachtin is an architecturally complex small molecule. It contains 16 stereocenters, seven of which are tetrasubsituted carbon atoms, along with an array of sensitive functionality including a hemiacetal, tetrasubstituted epoxide and enol ether. In fact, azadirachtin is so complex that it cannot be synthesized in meaningful quantities by chemists. The single disclosure of a chemical synthesis of azadirachtin by the research group of Steven Ley (Cambridge) relies on a “relay” strategy, required 22 years of development time (and more than 35 workers) and proceeds in 0.00015% overall yield. The Cambridge effort necessitated more than 70 synthetic operations with a longest linear sequence of 48 steps. Recently, a second approach to azadirachtin was reported by Hidenori Watanabe’s laboratory at The University of Tokyo. Watanabe’s route, which is analyzed and discussed in some detail below, intersects with an advanced intermediate that was produced during the course of the Ley group’s pioneering relay synthesis.

            Azadirachtin is comprised of western decalin and eastern furo[2,3b]pyran substructural fragments connected by a congested central C8-C14 linkage. In the single completed synthesis of the natural product, the C8-C14 bond is forged by an intermolecular alkylation/Claisen rearrangement sequence (diagrammed above). Very recently, Zhen Yang’s research group (Shenzhen, China) has reported progress towards the execution of a strategy that will utilize a similar sigmatropic rearrangement reaction to eventually connect the crowded vicinal quaternary carbon atoms at C8 and C14. The new University of Tokyo approach to azadirachtin is conceptually unique. Watanabe’s synthetic route features a novel intramolecular tandem radical cyclization reaction that stitches together the B and E rings simultaneously with the C8-C14 bond already in place. This post will highlight their synthesis of an optically active eastern pyran segment as well as the completion of an asymmetric formal synthesis of azadirachtin.

            Synthesis of the eastern pyran system begins from an optically active monosubstituted cyclopentenone (structure shown above) that is first elaborated to a carboxylic acid with additional diol functionality protected as an acetal. The latent furo[2,3b]pyran system is next revealed by ozonolysis of the cyclopentene moiety which, upon exposure of the initial oxidative cleavage product to a Brönsted acid, furnishes the requisite oxacyclic system as a mixture of four inseparable diastereomers. Subsequent silyl etherification of the hydroxyl group allowed chromatographic separation of the stereoisomers and the desired one was obtained in moderate (30%) yield. The lactone was then partially reduced to a b-oriented O-methyl acetal. Use of the b-acetal in a subsequent epoxidation step was previously demonstrated by Ley and co-workers to be synthetically advantageous. The b-acetal was elaborated to a methyl ketone in six additional steps and subsequent nucleophilic addition of lithium trimethylacetylide to the ketone delivered a tertiary alcohol with modest stereoselectivity. Finally, a desilylated anomeric hydroxyl group was converted to a phenylseleno moiety, intended to serve as a radical precursor in a later key step, through the intermediacy of an anomeric mesylate. The overall route to the furo[2,3b]pyran substructure proceeds in 18 steps from the known cyclopentenone shown in the scheme above. Synthesis of the complementary western spirocyclic carbaldehyde system (shown below) was previously described by Watanabe and co-workers back in 2007 (highlighted here).

            The two hemispheric fragments of azadirachtin were joined by lithiation of the eastern terminal alkyne with capture of the newly formed carbanion by the aldehyde substituent appended to the western tetrahydrofuran ring system. This condensation reaction (depicted in the scheme above) proceeds with full stereocontrol and an exceptional level of efficiency, especially when one considers the complexity and functional group tolerance of the transformation. The nucleophilic acetylide addition is conducted successfully in the presence of an a,b-unsaturated lactone, two different acetals and a selenoether moiety. Moreover, the product of the fragment union is a somewhat sensitive propargylic acetate that serves as a latent allene system in a later stage of the synthetic route.

            The tetrasubstituted allene required for the critical tandem radical polycyclization step is then installed by S_N2' addition of a methylcopper reagent to the aforementioned propargylic acetate to provide requisite allene substrate as a single stereoisomer. All of the previously noted comments about a high degree of functional group compatibility equally apply to this challenging S_N2' methylation as well. In the highlighted transformation, heating of the allene in N,N-dimethylformamide (DMF) under free-radical promoting conditions (ⁿBu₃SnH, AIBN) generates a carbon-centered radical from the phenylseleno functionality. The transiently formed radical engages the internal carbon atom of the allene in an addition reaction that simultaneously establishes the eastern E-ring, along with a new tertiary C8 radical. The latter is then poised to add into the proximal spirocyclic enone acceptor which assembles the decalin B-ring and terminates the polycyclization cascade. The radical cyclization product possesses the full carbon framework of azadirachtin, including an intact C8-C14 bond and all of the appropriate functionality to complete the synthesis of the limonoid natural product. Ten additional steps, post-radical cyclization, were required to produce an intermediate that intersects with Steven Ley’s synthetic process. Per the previously demonstrated Ley protocol, nine additional steps are then required to elaborate the common advanced intermediate into azadirachtin. In order to fully execute the new Watanabe route, which encompasses the endgame of Ley, a longest linear sequence of 39 synthetic steps are required. Ley’s original route, while notoriously lengthy and laborious (~71 total steps), required a longest linear sequence of 48 operations. Therefore, Watanabe’s new approach does not present a great deal of improvement in step economy or overall efficiency. However, the new method for construction of the azadirachtin skeleton provides an innovative solution to ‘the C8-C14 problem’ that has tormented synthetic chemists since the full structure determination (revision) by Kraus in 1985.

            The azadirachtin case study challenges the reigning perception that ‘any molecule can be synthesized, given the appropriate time and resources.’ This is actually the wrong metric by which to judge the success of a synthetic project. At the end of a total synthesis campaign, we should always ask ourselves, ‘how does this technology compare with extraction from natural sources or microbial production by an optimized fermentation process involving something like engineered biosynthesis?’ Organic synthesis should provide some advantage, if not with regard to efficiency of production, then at least in terms of access to analogues with improved properties that could not be obtained by a semisynthetic derivatization strategy.