Large rings are central to drug discovery, and an attractive way of generating them is through macrocyclic ring-closing metathesis (MRCM),[1],[2] even when stereoisomeric mixtures are generated (often) or the ring does not contain an alkene unit, indicating the considerable power of MRCM. However, when it comes to rings with a trisubstituted olefin, MRCM is typically inefficient and there is currently no reliable way for controlling stereochemistry.[3],[4] If for some reason MRCM happens to selectively deliver an E- or a Z-trisubstituted alkene,[5],[6],[7] it could be the undesired isomer,[8],[9] and several additional and costly steps will be needed to reverse stereochemistry. The state-of-the-art is especially perilous when ring formation must occur late-stage in a multistep sequence with a substrate that is rather precious. These are some of the key and longstanding problems that we set out to address. We chose to concentrate our efforts on the dolabelide family of natural products, which is comprised of four macrocyclic trisubstituted alkenes that are active against cervical cancer (Fig. 1). The total syntheses of two members of this family had been reported. In both cases, a late-stage MRCM was used to access the macrocyclic olefin, but only as equal stereoisomeric mixtures.[3],[4] The desired E-isomers were isolated in just 20–30% yield after chromatography.
We chose to approach the problem from two angles. On one front, we focused on transformations that afford sparsely substituted macrocycles with minimal entropic support, and on another front, we set out to design and perform a new route leading to dolabelide C with a late-stage catalytic stereoretentive MRCM being the crucial event.
Not long after we began, we faced an unexpected complication (Fig. 2). Even though the energy difference between the competing transition states was predicted to be high in a stereoretentive process, our model transformation (1 to 2, Fig. 2) afforded a »1:1 mixture of alkene isomers. After some debate and several control experiments, we discovered that the culprit was a small amount of E-butene byproduct, capable of causing pre-metathesis isomerization of the trisubstituted alkene. To circumvent this, we ran the transformation under mild vacuum, and indeed, at 100 Torr, selectivity improved dramatically to 95:5 E:Z. The process was still mildly efficient, however, significant homocoupling occurred at the disubstituted olefin terminus, an event probably facilitated at higher concentration caused by solvent removal under reduced pressure. To counter homocoupling and lower the odds of E-2-butene encountering the diene, we further diluted the solution. Also, we surmised that the rate of the desired intramolecular process would be unaffected. The next few experiments validated our analysis: at 1.0 mM concentration, MRCM delivered 2 in >98:2 E:Z selectivity (Fig. 2). With the optimal MRCM condition secured, the methodological studies proceeded smoothly. Without the need for much entropic support, a wide range of 12- to 22-membered ring of E- and Z-trisubstituted macrocyclic alkenes within a lactone, lactam, or carbocycle was efficiently synthesized in high stereoisomeric purity (typically, >95%).
But would the approach pass the ultimate test? Might we be able to use it to perform a late-stage MRCM en route to dolabelide C (Fig. 3)? It took us 36 steps to reach the needed substrate (3). The stage was finally set! Excitingly, the MRCM delivered trisubstituted alkene 4 in 66% yield as a single isomer (>98:2 E:Z), allowing us to complete the synthesis in 2.0% overall yield, a seven-fold improvement compared to what was previously reported.[3]
Getting an unbiased diene to cyclize to one isomer is challenging, but forcing a bias substrate to cyclize to the unfavored isomer is an entirely different ballgame. This challenge was tempting and important. A macrocycle with a Z versus an E alkene has a different contour and can exhibit different affinity for the same biological receptor or associate with an entirely different set of targets. Might MRCM be used to shape-shift a macrocycle? We chose to probe the case of fluvirucin B1 that our group had synthesized long ago by a MRCM that, unusually, afforded the Z isomer, a preference largely owing to the substrate’s conformational bias (Fig. 4).[10] We were excited to see that treatment of secondary amide 5a afforded the E-6a with appreciable selectivity (23:77 Z:E). Would a more conformationally flexible tertiary amide lend itself more readily to the demands of a stereoretentive transformation? This was indeed the case, as with a slightly different Mo complex, we were able to convert 5b to E-6b in 8:92 Z:E selectivity (Fig. 4).
As it turned, at the end, we did get what we wanted, courtesy of catalytic stereoretentive MRCM!
Filippo Romiti, Assistant Professor, University of Texas at Dallas
Filippo Romiti recently completed his postdoctoral studies with Prof. Amir H. Hoveyda in Boston and Strasbourg and will begin his independent career in June 2022 as an assistant professor in the Department of Chemistry and Biochemistry at the University of Texas at Dallas (https://labs.utdallas.edu/romiti/). His research interests are the design and development of efficient, broadly applicable, and selective transformations that may be used to generate complex organic molecules that are of particular interest in drug design and discovery.
[1]. Hoveyda, A. H. & Zhugralin, A. R. The remarkable metal catalysed olefin metathesis. Nature 450, 243–251 (2007).
[2]. Hughes, D., Wheeler, P. & Ene, D. Olefin metathesis in drug discovery and development – examples from recent patent literature. Org. Process Res. Dev. 21, 1938–1962 (2017).
[3]. Hanson, P. R. et al. Total synthesis of dolabelide C: a phosphate-mediated approach. J. Org. Chem. 76, 4358–4370 (2011).
[4]. Park, P. K., O’Malley, S. J., Schmidt, D. R. & Leighton, J. L. Total synthesis of dolabelide D. J. Am. Chem. Soc. 128, 2796– 2797 (2006).
[5]. Nicolaou, K. C., Montagnon, T., Vassilikogiannakis, G. & Mathison, C. J. N. The total synthesis of coleophomones B, C, and D. J. Am. Chem. Soc. 127, 8872–8888 (2005).
[6]. Anketell, M. J., Sharrock, T. M. & Paterson, I. A unified total synthesis of the actinallolides, a family of anti-trypanosomal macrolides. Angew. Chem. Int. Ed. 59, 1572–1576 (2020).
[7]. Wasser, P. & Altmann, K.-H. An RCM-based total synthesis of the antibiotic disciformycin B. Angew. Chem. Int. Ed. 59, 17393–17397 (2020).
[8]. Smith, III, A. B., Mesaros, E. F. & Meyer, E. A. Total synthesis of (–)-kendomycin exploiting a Petasis–Ferrier rearrangement/ring-closing metathesis synthetic strategy. J. Am. Chem. Soc. 127, 6948–6949 (2005).
[9]. Toelle, N., Weinstabl, H., Gaich, T. & Mulzer, J. Light-mediated total synthesis of 17-deoxyprovidencin. Angew. Chem. Int. Ed. 53, 3859–3862 (2014).
[10]. Houri, A. F., Xu, Z., Cogan, D. A. & Hoveyda, A. H. Cascade catalysis in synthesis. An enantioselective route to Sch 38516 (and fluvirucin B1) aglycon macrolactam. J. Am. Chem. Soc. 117, 2943–2944 (1995).
Please sign in or register for FREE
If you are a registered user on Research Communities by Springer Nature, please sign in