As we started the work on the 3D development compound we wondered if this compound was produced by the sponge or, as is usually the case, by a symbiotic bacterium enjoying its creature comforts. Somehow I missed it but it seems that it is the latter which produces the compound. The group from Harbor Branch indeed managed to isolate and identify a bacterium living on the sponge belonging to the genus Entotheonella. I assume, in my ignorance, that this microbe is responsible for the production of discodermolide. Just why a bacterium would wish to produce such a compound is a fascinating question, host defence perhaps, discodermolide is quite a toxic compound, or something else?

So how does the bacterium produce such a complicated structure? Presumably it employs enzymes, a polyketide synthase (PKS). The complex polyketides they synthesize are incredibly diverse in their activities. Each polyketide antibiotic has evolved to bind a molecular target in competing organisms to impede their growth. The classic example is the antibacterial erythromycin A.

There is a distinct similarity between a discodermolide fragment and a piece of erythromycin:

A company called Kosan Biosciences (now owned by Bristol-Myers Squibb) claimed to be able to make any erythromycin derivative you might care to need by using PKS technology and several patents and publications back up this claim. Indeed if you ferment a complete PKS mutant with the thioester (1) out pops 6-deoxyerythronilide B (2):

Compounds such as (2) can be easily converted to intermediates in the Smith synthesis of discodermolide. The first step along this route is a Bayer-Villiger of (2) to give a di-ester (3) or enzymatic hydroxylation to produce (4).

Chemical cleavage of either compound produces intermediates containing the discodermolide stereochemistry and meet up with compounds described in published routes.

This is certainly an interesting approach but a biotechnological route needs to produce the final compound rather than fragments requiring significant chemical manipulation to reach the final molecule. Unfortunately this makes this approach no better or probably worse than a totally synthetic route. With the failure of discodermolide in phase I trials I doubt that this approach will be developed. However, with the current advances in biotechnology, fermentation and isolation technology it certainly could be a viable alternative.