Secondary metabolites from plants, fungi and bacteria offer abundant and varied complex bioactive chemicals. Among these, polyketides have become one of the most important sources of therapeutic compounds, including antibiotics (such as erythromycin and tetracycline), anticancers (such as doxorubicin and epothilone B), immunosuppressants (rapamycin), anticholesterols (such as lovastatin), and antioxidants (such as resveratrol). The polyketides derive their potency from their considerable structural-, stereo- and functional group complexity. This complexity is wrought into the molecules through the cooperation of multiple enzymatic domains functioning together in a highly orchestrated assembly line. The medical value of polyketides has drawn much attention to the research of polyketide synthases (PKSs) with the objective of engineering these systems to produce tailor-made compounds. Researchers have met with many successes and more failures in their endeavors to exert complete control over the biosynthetic outcome of these pathways. The chief limitation is that the sequence-structure-function relationship of proteins is still poorly understood. This is true at all levels, including protein folding, substrate specificity, catalytic activity, and protein-protein interactions.

The subjects of this dissertation are enzymes from a group of PKSs called iterative PKSs, which produce aromatic polyketides such as bikaverin, aflatoxin, doxorubicin and actinorhodin. There are two types of polyketide synthases that fall within this group: the fungal iterative type I non-reducing PKS (NR-PKS) and the bacterial type II PKS. Among PKSs, these two types remain the least understood, due to the instability of the intermediates produced along their biosynthetic pathway. Unlike other types of PKSs, NR-PKSs generate a long carbon chain with ketone groups present on every other carbon. This arrangement is highly prone to undergo intra- and inter-molecular aldol/Claisen condensations, which makes it extremely difficult to study the native enzyme-substrate interactions in these systems. In this thesis, a suite of novel polyketone mimetics that maximally mimic the steric and electron structure of the unstable natural intermediate while being stable enough to be used as a probe in experimental studies are used to reveal the details of enzyme-substrate interactions of the product template domain from fungal NR-PKS and the reductase enzyme from bacterial type II PKS.