Investigation into the Activity and Specificity of the Thioesterase II, Ketoreductase and Dehydratase Domains of Multiple Polyketide Synthases.

Abstract

Many pharmaceuticals are derived from natural products produced by assembly-line complexes found in bacteria, fungi, and plants. Assembly-line complexes, such as non-ribosomal peptide synthetases (NRPS) and type I polyketide synthases (PKSs), are composed of modules that work in succession to synthesize complex products. Each module extends and modifies an attached intermediate, then passes it to the next module in series. The modifications performed are determined by the domains that comprise each module. This dissertation focuses on two PKS domains, dehydratases and ketoreductases, and a class of stand-alone proteins, thioesterase IIs (TEIIs), that are associated with many NRPSs and PKSs. The structure and substrate specificity of RifR, the TEII of the rifamycin NRPS/PKS complex was investigated by solving its crystal structure, the first of a TEII, and by testing its activity with a variety of substrate mimics. TEIIs are predicted to remove aberrant intermediates from acyl carrier proteins (ACPs) or peptide carrier proteins (PCPs) of PKSs or NRPSs, respectively. RifR was shown to prefer acyl-ACP substrates over acyl-CoA substrate mimics, and aberrant decarboxylated acyl moieties over productive carboxylated acyl moieties. The structures of RifR in multiple crystal forms and its similarity to other thioesterases suggest RifR undergoes a conformation change during catalysis. The substrate specificity of ketoreductase domains (KRs), which are responsible for reducing the β-carbonyl of the growing polyketide intermediate, was investigated using KR-ACP didomains from the erythromycin, pikromycin and tylosin pathways. KR-ACPs were shown to be active toward non-natural substrates. KR substrate or product mimics were loaded onto the ACP portion of the KR-ACP. Some mimics competed with non-natural substrates, providing insight into the specificity of the KR domains. A working assay to detect dehydration and hydration by the dehydratase domains (DHs) of the curacin A PKS/NRPS pathway was developed. Little information is available regarding the substrate specificity of type I PKS dehydratases. While all the curacin DHs were active, dehydration of substrates occurred at different rates. A better understanding of the structure and substrate specificity of these domains and proteins will help future attempts to reengineer existing assembly-line complexes to produce newer, more potent pharmaceuticals.