MECHANISTIC INVESTIGATION OF THE DIVERSE IRON AND 2-OXOGLUTARATE-DEPENDENT OXYGENASE SUPERFAMILY
Open Access
- Author:
- Dunham, Noah Paul
- Graduate Program:
- Biochemistry, Microbiology, and Molecular Biology
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- August 17, 2018
- Committee Members:
- Amie Kathleen Boal, Dissertation Advisor/Co-Advisor
Joseph M Bollinger Jr., Committee Chair/Co-Chair
Carsten Krebs, Committee Member
Squire J Booker, Committee Member
Howard M Salis, Outside Member - Keywords:
- enzymology
metalloenzymes
oxygenase
crystallography
structural biology
C-H activation - Abstract:
- Members of the Fe(II)- and 2-oxoglutarate (2OG)-dependent (Fe/2OG) oxygenase superfamily are ubiquitous in both prokaryotes and eukaryotes and participate in essential biological functions, such as oxygen sensing, DNA repair, epigenetic regulation, collagen synthesis, and the synthesis of myriad antibiotic and natural product compounds. Activation of aliphatic C–H bond of the primary substrate is achieved by iron-mediated reduction of O2 and subsequent cleavage of 2OG to CO2 and succinate to generate the highly-reactive Fe(IV)–oxo (ferryl) complex, an intermediate that initiates the radical chemistry by hydrogen atom transfer (HAT) in a diverse set of chemical transformations. In most reactions, the four-electron reduction of O2 is coupled to two-electron oxidations of both 2OG and the primary substrate, regenerating the Fe(II) form of the cofactor for subsequent turnovers. Although most Fe/2OG enzymes characterized to date are hydroxylases, a growing number of examples are being described that catalyze halogenation, desaturation, epimerization, C–O/C bond formation, endoperoxidation, epoxidation, and the formation of ethylene. Currently, there is great interest in developing biocatalysts that can functionalize C–H bonds regio- and stereoselectively under mild, aqueous conditions. The overarching goal of the research described in this thesis is to understand the detailed mechanisms by which these enzymes direct their reactions to the intended outcomes. The expectation is that this understanding will contribute greatly to future biocatalyst design. In Chapter 1, the current understanding of mechanisms employed by Fe/2OG oxygenases is summarized, spanning details of the original experiments with the hydroxylases to outlining mechanistic possibilities in the more recently discovered C–C bonding-forming enzymes. The chapter concludes by introducing the vanadyl as a structural mimic of the fleeting ferryl intermediate, which has potential to be used in all Fe/2OG systems to gain insight into the disposition of the cofactor and substrate during the C–H activation step. Chapter 2 describes the structural characterization of an Fe/2OG hydroxylase reaction cycle. In this study, our original goal was to capture the ferryl complex using x-ray crystallography by initiating the reaction of the L-arginine hydroxylase, VioC, in crystallo. VioC proved a good candidate for such trials as it reproducibly formed crystals that diffracted x-rays to high resolution (< 2.0 Å). Although the ferryl complex ultimately proved elusive, we were successful in solving multiple high-resolution crystal structures depicting different states of the reaction cycle. In addition to product complexes with 2OG and succinate, we also solved a structure that bore electron density consistent with the peroxysuccinate intermediate. This previously hypothesized complex is one of two conceivable intermediates that result from O2 addition before O–O bond scission and ferryl formation occur. We also used VioC to yield the first crystal structure of an Fe/2OG enzyme with the vanadyl ion to mimic the reactive ferryl complex. To validate our new method, the resulting structural coordinates were computationally optimized after replacing vanadium with iron and Mössbauer parameters were calculators in silico. These parameters agreed well with those obtained experimentally, providing convincing evidence that the vanadyl can indeed provide an accurate structural representation of the ferryl complex during the C–H activation step. In Chapter 3, we investigate the mechanism of Fe/2OG-catalyzed desaturation by using the native L-arginine desaturase, NapI, and the VioC-catalyzed non-native desaturation of L-homoarginine as model systems. This chapter describes the first detailed mechanistic study of this reaction outcome. By solving a panel of high resolution x-ray crystal structures of both enzymes, performing the accompanying pulsed EPR experiments, and synthesizing substrates specifically labeled with deuterium to analyze their effects on product outcome, we determined that NapI and VioC desaturate their substrates via distinct mechanisms. We demonstrate that the mechanism frequently cited in the literature is viable, but likely not utilized by most of the native Fe/2OG desaturases. Additionally, the VioC crystal structure with vanadyl and L-homoarginine solved in this project provided further support for the vanadyl ion as an invaluable mechanistic tool. In this case, the vanadyl complex revealed crucial differences in substrate conformation that rationalized other experimental observations that the substrate complex alone could not. In Chapter 4, we investigate the reaction outcome of VioC with its substrate enantiomer, D-arginine. Structure analysis by x-ray crystallography suggested the enzyme will likely target a new carbon center. As predicted, we show that VioC promotes deamination of the α-carbon, instead of C3 hydroxylation as in the reaction with the native substrate, to yield 2-oxoarginine as the sole product. By simple product identification, the mechanism initially appeared to abide by the canonical hydroxylation pathway followed by spontaneous breakdown of the resulting intermediate, analogous to the Fe/2OG DNA and histone demethylases. Through a series of oxygen labeling studies, however, we demonstrated that the origin of the inserted oxygen atom is not O2, but rather H2O, suggesting a simple hydroxylation mechanism is not operant. Further investigation led us to conclude that the reaction proceeds by a desaturation across the αC–N bond, strikingly similar to the desaturation mechanism employed by NapI. The resulting iminium is then hydrolyzed by H2O to form 2-oxoarginine. Chapter 5 summarizes the current understanding of reactions catalyzed by clavaminic acid synthase (CAS) and our initial steps toward answering the heretofore unresolved mechanistic questions. CAS is a remarkable Fe/2OG enzyme that catalyzes hydroxylation, oxacyclization, and desaturation via three uncoupled reactions in the biosynthesis of clavaminic acid, a precursor to clavulanic acid and other potentially beneficial natural products. We first present preliminary x-ray crystal structures of the CAS 3 (CS3) isoform and confirmed reactivity with the native hydroxylation substrate. We also examined the reaction of CS3 with a non-native substrate that has been previously shown to yield a desaturated product, similar to the non-native desaturation of L-homoarginine by VioC. In conclusion, we discuss our future plans to characterize the oxacyclization reaction, and both the native and non-native desaturation reactions. In Chapter 6, we detail our mechanistic investigation into the endoperoxidation of fumitremorgin B catalyzed by FtmOx1, one of the only Fe/2OG enzymes known to facilitate this difficult transformation. In a recently-published study, the authors proposed a novel mechanism of catalysis: FtmOx1 utilizes an active site tyrosine residue to mediate the HAT from the substrate to the ferryl complex, allowing the substrate to reside farther away from the cofactor to effectively suppress a competing hydroxylation pathway. Our own study, however, suggests that this mechanism is incorrect. We clearly demonstrate that the hydroxylation pathway does indeed compete with endoperoxidation by tracking isotopes of oxygen into a specific breakdown product. We further show that the ratio of products resulting from the endoperoxidation and hydroxylation pathways can be biased by tuning the availability of O2 in the reaction. Moreover, we characterize an important intermediate during the first turnover of the reaction as a tyrosyl radical; however, we show that the formation of the species occurs after the endoperoxide ring is formed, suggesting it does not mediate the initial HAT step. Understanding the methods Nature employs to form endoperoxides is of great value, as many endoperoxide-bearing natural products show promise to positively affect human health.