A new and efficient method for genetically manipulating the chemical structure of natural products, a long-established source of drug leads, has been developed, and its success shown in the modification of a polyketide natural product that might provide the basis for the development of potent anticancer agents. This new method, which is described in Chemistry and Biology, addresses a major limitation of natural products as leads — the difficulty of incorporating synthetic modifications, owing to their complex structures — and thereby facilitates the optimization of their pharmacokinetic and pharmacodynamic properties.

Polyketides are a large family of natural products that are constructed from acyl-coenzyme A monomers. Geldanamycin is one such polyketide that targets the chaperone protein HSP90, which is overproduced in several types of human cancer. HSP90 chaperones immature kinases, which are important components of signal-transduction pathways, many of which are dysregulated in cancer cells. These immature kinases are rapidly degraded in the presence of geldanamycin, and the subsequent reduction in mature kinases can result in apoptosis and cell death. Geldanamycin might therefore provide an ideal starting point for the generation of anticancer agents to target this pathway.

Several synthetic geldanamycin analogues, including 17-AAG, which is currently undergoing clinical evaluation, have been produced by manipulating the chemically reactive groups of this natural product. However, the modification of the inert groups of this molecule, which might allow further optimization of its pharmacological properties, has until now not been explored.

Geldanamycin is made in Streptomyces hygroscopicus by polyketide synthases (PKSs), which are structured in a modular fashion. PKS modules catalyse the step-wise elongation of a polyketide chain, each module being responsible for the incorporation of one acyl group monomer in the final structure. Patel et al. developed three approaches (double crossover using bacterial conjugation, double crossover using phage, and gene complementation using bacterial conjugation) to manipulate the inert groups of geldanamycin-related molecules by substituting one of the catalytic domains — the acyl transferase domain — at several positions on the PKS modules with those that would lead to the incorporation of different acyl group monomers. This led to the efficient production of unique geldanamycin analogues that would be very difficult to produce through conventional chemical modification.

In developing this method, the authors generated a geldanamycin analogue, KOSN1559, which binds to HSP90 with a fourfold greater affinity than that of 17-AAG. This analogue also lacked the quinone moiety that is believed to lead to hepatotoxicity of 17-AAG. This work demonstrates the success of a method that could be used to develop more potent and safer analogues of geldanamycin with improved cellular uptake while maintaining the enhanced HSP90-binding affinity through chemical modification.