Empowering Synthesis of Complex Natural Products (original) (raw)

2019, Chemistry: A European Journal

Despite their previously reported synthesis, 17 Baran and coworkers wanted to enhance the efficiency and scalability of their process to allow for further biological investigations (Scheme 1). Learning from previous experience, they identified the stereoselective formation of the imidazoline ring as a critical step for improvement, which they envisioned through a "chloro-guanidylation" via stereoselective chloronium ion formation with subsequent stereospecific displacement by the guanidine to form the desired scaffold. Towards this goal, the synthesis commenced with in situ TMS protection of bis-allylic alcohol 3 employing bis(trimethylsilyl)-acetamide, followed by an optimized stoichiometric Pauson−Khand reaction with complex 4 to yield cyclopentenone 5 in 45% yield on 6.17 g scale. Interestingly, it was found that polyols and NMO displayed a synergistic effect to enhance the yield; however, despite efforts to prepare a reagent containing both of these features, the yield did not exceed 49%. Subjecting 5 to a two-step sequence of Luche reduction and Appel reaction furnished tri-chloride 6, setting the stage for a Zn/In-mediated Barbier-type allylation to yield alcohol 7 on a 5.60 g scale. Double nucleophilic azidation with subsequent Boc deprotection and guanidylation produced the desired chloro-guanidylation precursor 9. Through optimization, it was shown that subjecting crude mixtures of allylic guanidine 9 to chloro-guanidylation gave a marked increase in yield. After careful analysis of the reaction mixture, it was found that byproducts of the guanidylation reaction were producing a potent chlorination reagent in situ, leading to the development of commercially available Palau'chlor (not shown). With the ability to synthesize imidazoline 10 on a 4.60 g scale, Baran and coworkers began the endgame of the synthesis with TFA deprotection and cyanamide-mediated cyclization leading to cyclic guanidine 11. Closure of the final ring of the natural product scaffold was achieved via DMDO epoxidation with concomitant epoxide opening. The remaining alcohol was introduced through silver(II)-picolinate oxidation to afford tetracycle 12 on a 1.60 g scale. Reduction of the azide functionalities with Adams' catalyst and one-pot amidation provided a mixture axinellamines A (1) and B (2) on gram scale. This practical synthesis allowed for the identification of more bacterial pathogens that could be treated with these natural products, adding to the arsenal of broad-spectrum antibiotic scaffolds. 2.1.2 Leiodermatolide (Fürstner, 2014). While the scales of syntheses are often restricted by availability of materials and reagents, step-count, or overall efficiency, there are a few examples of the where the final amount produced was limited due to safety concerns regarding the high toxicity of the target natural product. This is exemplified by Fürstner's highly convergent synthesis of leiodermatolide 13 (Scheme 2), a potent cytotoxic spongederived marine natural product with GI50 values ≤3 nM across multiple cell lines, including those that are efflux active. 18 Through their early studies, Fürstner and coworkers observed that the tubulin disruption caused by 13 did not proceed through direct tubulin binding, and they instead hypothesized that the mode of action was centrosome declustering. This unorthodox antineoplastic activity required additional study and the authors decided their previously reported synthesis from 2012 could not furnish the necessary quantities needed for biological assays. 19