Totally Synthetic by Paul H. Docherty, 12 March 2007

Total Synthesis of Terpestacin

Trost

B. M. Trost, G. Dong, J. A. Vance, J. Am. Chem. Soc. 2007, 129, 4540-4541.

Continuing his application of the Pd-AAA (Pd catalyzed asymmetric allylic alkylation), Barry Trost and his group have recently completed a rather nice total synthesis of the biologically interesting macrocycle, Terpestacin. This small but complex molecule has shown angiogenesis inhibition along with inhibition of syncytia production in HIV-infected T cells. However, perhaps more importantly for us, it represents an interesting molecular target, with a highly substituted cyclopentanone ring appended to a 15-member macrocycle. Since its isolation in 1993, there have been three notable syntheses, completed by Tatsuta, Myers and Jamison, and now this one by Trost.

As can be seen from the retrosynthesis, they planned to use the Pd-AAA-Claisen strategy twice in this synthesis, conferring asymmetry to several centres, firstly in the cyclopentane:

As can be seen from the result, they were able to create a quaternary centre α to a carbonyl with high levels of enantioselectivity and pretty decent yield. The reagents used looks slightly complex, but bear in mind that they are also protecting the free alcohol from the epoxide in this reaction sequence. They then built-up complexity and functionality around the cyclopentanone, using a nice example of the Sakurai allylation, and were able to alkylate by producing the dianion of the sulfone, and then desulfonating with palladium after the alkylation.

With this material in hand, it was time for a RCM, but they must have completed this with some fear - afterall, there is quite a few olefins in the substrate! With some degree of luck, the desired cyclisation occured in around 40%, generating the desired trans olefin. They attribute their success to optimisation of reaction conditions, and use of an allylic alcohol. Removal of the PMB protection (using magnesium bromide and DMS in this case) set them up for the second Pd-AAA-Claisen, which was again very successful, with both a great yield and d.r., all but completing the carbocyclic core of the molecule.

However to finish they had to functionalise the side chain, which was no simple task, as they planned to dihydroxylate the olefin, and cleave to generate the aldehyde, which could be reduced to the natural product. However, the substrate for the dihydroxylation has five olefins, so a chemoselective reaction was required. They rationalised that the trisubstituted endocyclic olefins would be confirmationally restricted, and thus might react slower than the desired olefin, and that using an asymmetric dihydoxylation they might bias this reaction further still. Seems like a gamble to me, but what do I know? A 65% yield would suggest that I don’t know a lot, and that this synthesis by Trost is rather sweet!