ABSTRACT
The example chosen for this is the synthesis of the polyketide natural product (+)-discodermolide discovered in the marine sponge Discodermia dissoluta. Since the compound showed promise as an anticancer agent, it was required to develop a scaled-up total synthesis to provide enough quantities of this material for clinical trials. The natural source amount is 7 mg from 434 g of frozen sponge or 0.002% w/w [1]. With a molecular weight of 593 g/mol, this works out to an E-factor of 62,000. This is a minimum estimate since it does not include the solvent demand required in the extraction process. In 2000, there were two total syntheses reported in the literature, which we refer to in this compilation as the Smith G2 [1,2] and Paterson G1 [3,4] plans. The Tanimoto similarity index for the two plans with respect to the target bonds selected in the plans is 42%. Novartis [5-8] decided to choose the Smith G2 plan for scale-up and not surprisingly both plans had a 93% Tanimoto similarity index. By contrast, the Paterson G1-Novartis pair has a 58% similarity in the target bonds made. The data presented in Table 11.1 show that this was not a wise choice since even at the level of kernel metrics (reaction yield and atom economy) the Paterson G1 plan produced six times less waste than the Smith G2 plan. The decision made by Novartis was done without the bene“t of a thorough metrics evaluation of the synthesis performance. Moreover, probably because of scale-up issues, the Novartis plan turned out to produce 1.5 times more waste than the Smith G2 plan. Since that time, Smith produced two more synthesis generations with the fourth announced in 2005 [9] producing 79.1 kg waste per mole, a 4.5-fold improvement over the Smith G2 plan. In 2004, Paterson also improved his strategy to 44.3 kg waste per mole in Paterson G3 [10] and this represents so far the most material ef“cient plan at the kernel metrics level. This example illustrates to process chemists contemplating a scaled-up plan for a given target from a set of published procedures, that a well thought out and executed metrics analysis can go a long way in making good decisions about which plans to choose. The automatic consequence of selecting and carrying out material ef“cient and cost effective plans is the achievement of greener syntheses.
