The Catalytic Enantioselective Total Synthesis of (+)-Liphagal
This tetracyclic meroterpene has had quite a bit of attention from the synthetic community since its isolation back in 2006 from a Caribbean sponge. Along with other well known natural products such as resveratrol, staurosporine, and wortmannin it's an inhibitor of phosphatidylinositol 3-kinase (PI3K), an enzyme involved in numerous diseases. The fact that those all have wikipedia pages is quite telling. It's also got some cytotoxicity towards various cancer cell lines, but I suspect the reason people keep making it is the unique and interesting 6-7-5-6 ring system.
Somewhat uncommonly, the isolation and total synthesis were disclosed in the same paper by Andersen and coworkers, and subsequent (quite similar) total syntheses were recently reported by Baldwin and Alvarez-Manzaneda. There's also a handful of other studies (see the paper for references), including an oh-so-close formal synthesis by Mehta and coworkers, lacking only the final double demethlation with BI3 (eek!) reported by Andersen. I can understand why they gave that a miss. Andersen suggested two possible ways that nature might make this compound from a farnesyl arene: either by formation of a decalin system, followed by ring expansion; or by first forming the furan, then a different cationic cascade cyclisation to give the 6-7 system directly. Baldwin's synthesis was based around the first conjecture (he additionally suggested the intermediacy of an ortho-quinone methide in the decalin expansion sequence), whereas Mehta and Anderson mimicked the second.
Enter the Stoltz group. They're not so interested in biomimicry, as that's been pretty well done already, but they chose liphagal to showcase their palladium-catalysed enantioselective decarboxylative alkylation methodology. You may have also seen it in the group's formal total synthesis of (+)-hamigeran B in an Org. Lett. paper at the start of the year. I saw Stoltz talk about this a couple of years ago, and it's really come on since then. Here it's used straightaway as the source of asymmetry for the whole synthesis. Unfortunately, looking at the SI it's a glovebox job, and they only report the reaction on a couple of hundred mgs, although I assume it can be scaled somewhat as it's such an early step.
Thus, having already set liphagal's C-11 quat. centre in the first step, the group proceeded to elaborate the ketone into a tricyclic cyclobutene using some pretty well established chemistry, before bolting on the veratrol ring with a neat palladium mediated arylation in the microwave. This gave access to the ring expansion precursor as a single diastereoisomer. Treatment of this (well, a racemic version) with BF3•OEt2 did effect ring expansion, and also formed a bridged byproduct by a Cargill rearrangement. In the quite likely event you've got no idea what that is (I didn't), I've drawn out the structure below so you can try and guess the mechanism (it's given in the paper).
Aluminium trichloride didn't form any of the Cargill byproduct, but did give Friedel-Crafts byproducts and again only a poor yield of the expanded product. The group therefore decided to attempt to deactivate the veratrol ring (and provide a handle for furan formation) by bromination. Remarkably, they found that this was possible even in the presence of the cyclobutene, using nothing more exotic than Br2 in CHCl3, giving a respectable yield of 65%.
It was found that the brominated compound could be rearranged without the need for any Lewis acid, simply by heating a solution of it in ortho-Dichlorobenzene in the microwave. Next, the enone double bond was removed by hydrogenation with Adams' catalyst and epimerisation, methylation and reduction were carried out. The dihydrobenzofuran ring was forged by cyclisation of an alkoxide onto the aryne formed upon treatment of the bromide with excess LDA. The crucial hydrogenation used to form the trans 6-7 worked perfectly under standard conditions to give only the desired product in essentially quantitative yield, presumably due to large groups present on the α-face. Oxidation of the dihydrobenzofuran was unexpectedly difficult, with DDQ giving overoxidation. Nitrosonium tetrafluoroborate (apparently a hydride abstractor) gave much better results, and a Bouveault-type formylation followed by demethylation delivered the natural product. Nice work!