Date of Award


Degree Name

Doctor of Philosophy




Chu-Young Kim


Polycyclic polyether natural products have received much attention due to their diverse biological activities, ranging from extreme toxicity to therapeutic properties, including antimicrobial, antifungal, and anticancer activities. They are featured by multiple cyclic ether groups that vary in the number, size, and arrangement. Ionophore polyethers contain multiple tetrahydrofuran and tetrahydropyran rings connected by either spiroketal systems or carbon-carbon single bonds. Over the past 40 years, significant progress has been made in deciphering polyether biosynThesis pathways. Remarkably, all members of the polycyclic polyether family are thought to be generated via a common biosynthetic scheme. In 1983, Cane, Celmer, and Westly proposed a biosynthetic model for producing the stereospecific ether rings of polycyclic polyethers. It postulates that linear polyene intermediates undergo a cascade of enzyme-catalyzed reactions to form the polyether rings, including stereospecific epoxidation and epoxide ring opening reactions.

In 2001, the biosynthetic gene cluster for polyether monensin was isolated from Streptomyces cinnamonensis. It contains the genes of a flavin-dependent epoxidase MonCI and two epoxide hydrolases MonBI and MonBII. Gene deletion and feeding experiments showed that MonCI catalyzes all three stereoselective epoxidations for the intermediate premonensin, while MonBI and MonBII are responsible for the following epoxide ring opening reactions. In 2008, the gene cluster responsible for the biosynThesis of lasalocid was identified from Streptomyces lasaliesis. It also includes the genes of a flavin-dependent epoxidase Lsd18, and an epoxide hydrolase Lsd19. Similar to MonCI, Lsd18 transforms two olefins in the intermediate prelasalocid into two epoxides in a stereoselective manner. Also similar to MonBI/MonBII, Lsd19 catalyzes epoxide ring opening reactions, generating the final product lasalocid with a THF-THP construction. So far, the mechanism of epoxide ring opening reactions has been extensively studied well. The Lsd19-substrate/product complex crystal structures revealed how energetically favored 5-exo cyclization and energetically unfavored 6-endo cyclization are catalyzed. Also, gene mutagenesis and the crystal structure determination of MonBI/MonBII further explained the synergistic effects and allosteric regulation between the epoxide hydrolases. In contrast, as the first step of the polyether ring construction, the mechanism of stereoselective epoxidation is poorly understood due to the lack of epoxidases' molecular structures.

My research project is focused on determining the molecular mechanism of stereospecific epoxidation catalyzed by MonCI and Lsd18. Recombinant MonCI and Lsd18 was successfully expressed in E. coli BL21(DE3) and purified using standard chromatography techniques. These proteins were shown to be active by in vitro enzyme assays that included the extracted proteins MonCI/Lsd18, substrates, FAD, and reducing agent nicotinamide adenine dinucleotide (NADH) or nicotinamide adenine dinucleotide phosphate (NADPH). The epoxide products were detected by either GC-MS or LC-MS. We solved the MonCI crystal structure at 1.90 Å resolution and the Lsd18 ligand-free and ligand-bound complex crystal structures at 1.54 and 1.85 Å, respectively by X-ray crystallography. MonCI and Lsd18 share high similarity in both sequence (identity = 47.8%) and three-dimensional structure (RMSD = 1.63). Both MonCI and Lsd18 structures contain a fused FAD- and substrate-binding pocket with three openings. Presumably, one for substrate, one for solvent, and one for FAD reduction/NAD(P)H binding. Since they share high structural similarity to the p-hydroxybenzoate hydroxylase, especially the FAD-binding domain, the mechanism of FAD reduction and oxidation process is expected to be similar. The catalytic cycle consists of five steps. First step is FAD reduction. When the isoalloxazine ring of oxidized FAD moves out from the pocket (“out” conformation), it is exposed to the solvent and accepts electrons from NAD(P)H. After reduction, FAD moves back to the pocket (“in” conformation) and transforms into the intermediate C4a-(hydro)peroxyflavin by reacting with molecular oxygen. Then C4a- (hydro)peroxyflavin donates one hydroperoxyl oxygen atom to the substrate (oxidation). The final step is accomplished at the end by FAD eliminating one water molecule and turning back to the oxidized status. To investigate how epoxidases that are involved in the polyether biosynthesis achieve stereoselectivity, we conducted the docking and modeling work on MonCI and Lsd18. A molecular dynamic simulation was also performed on MonCI with its native substrate premonensin. Our findings suggest that the stereoselectivity of epoxidation is determined by the unique preorganization of the substrate-binding pocket which permits only one face of the alkene to approach the reactive C4a-(hydro)peroxyflavin. Several key residues at the active site that play an important role in stereoselectivity have been identified and mutated to further test our hypothesis.

Overall, our crystallographic and computational studies have provided important molecular insights into how the stereoselective epoxidation is achieved by the flavin-dependent monooxygenases MonCI and Lsd18. Due to the high sequence similarity, all epoxidases that are involved the biosynThesis of polyether natural products could are likely to employ the same catalytic mechanism. Our results are expected to help bioengineering efforts for bioproduction of polyether natural product analogs for drug research.




Received from ProQuest

File Size

160 pages

File Format


Rights Holder

Qian Wang

Included in

Biochemistry Commons