The Roush group in sunny Jupiter, Florida, along with some associates from the (not-so-sunny) University of Michigan, have published an epic synthesis of (+)-tedanolide and (+)-13-deoxytedanolide. This synthesis almost certainly had its share of emotional ups and downs, and several snags were overcome to finally get at the natural product and its deoxygenated analogue. Kudos to the researchers for maintaining their sanity and really putting their noses to the grindstone on this one.
They envisioned using macrolactonization as a key step, but observed competitive cyclization at the C13 hydroxyl group. Good regioselectivity was achieved by replacing Yamaguchi conditions with more or less standard esterification conditions (DMAP, DCC, DMAP-HCl), but they took a tough hit on yield under these conditions and had to look for alternatives. Noticing at this point in the synthesis that they had a hydroxyl group at C15, where the natural product has a ketone, they attempted to oxidize C15 before conducting the macrolactonization. Alas, the six-membered hemiketal of this would-be ketone with the C11 hydroxyl turned out to be the sole product of any kind of oxidation. Luckily, 33a could be taken on to (+)-13-deoxytedanolide via dehydroxylation, C15 oxidation (now possible because the hemiketal is disfavored without the C13 hydroxyl), and macrolactonization.
But what about the natural product with the C13 hydroxyl? Here’s where a stroke of genius came into play: the diastereomer of 32 with the configuration of C15 switched can only cyclize to put the C16 fragment (that big chunk on the left) axial, and the authors imagined that this would severely destabilize the cyclized form. They had used an aldol reaction to forge the C12-C13 bond, and so had to back up to a new aldehyde intermediate, 49. This was actually the second aldehyde they tried; the first contained the desired Alloc protecting group (see 32 and 33 above), but they couldn’t get an Evans auxiliary off without chewing up the Alloc group too! So, they went for the TBDPS-protected aldehyde…and long story short, this didn’t work either. Combining this with aldol partner 5 led to the desired product, but TES removal led to opening of the epoxide! Ack!Why not try to install the epoxide as a final step from the corresponding alkene? As it happened, this strategy worked, but getting to the desired unsaturated aldehyde was somewhat non-trivial. They coupled this new aldehyde, 52, with 5, then reached (+)-tedanolide by Yamaguchi macrolactonization, C15 oxidation, and epoxidation with mCBPA.Yet another testament to the value of persistence in organic synthesis. Two thumbs up!