University of Florida Thesis Or Dissertation Formatting

INVESTIGATIONS INTO THE ENZYMATIC MECHANISM OF BACILLUS SUBTILIS OXALATE DECARBOXYLASE: AN ELECTRON PARAMAGNETIC RESONANCE APPROACH

UMAR TARIQ TWAHIR

A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

2015

To my parents, uncle, aunt and beloved wife

ACKNOWLEDGMENTS

I would like to start with thanking God for providing me every opportunity, blessing and success in my life. Throughout my life there were many moments of uncertainty, not knowing what the next step would be for me. However, the one consistent thing in my life from the beginning was God, and I always knew everything would be as it should, needing only from me to give my all in everything I set out to do.

During my time here at the University of Florida, the success and memories were due to a collection of individuals that included mentors, colleagues, collaborators, friends and family. First and foremost I would like to extend my deepest appreciations for my Ph.D. advisor, Professor Alexander Angerhofer. With your help, support and encouragement, I have grown into the scientist that I am today. You have always gone above and beyond to ensure that I was afforded all necessary opportunities for the advancement of my research and career. You truly are the model I look to when reflecting on the prospective future of my own career. You have encouraged me to present my research at multiple local, national and even international meetings, expanding my outlook, and for all of this I am truly grateful.

I would like to also thank my committee members: Professors Gail Fanucci,

David Wei, Benjamin Smith, and Yuri Sautin. I know that you each have multiple commitments and serve on many committees; however, you have taken the time to assist in my academic pursuits, and for that I am extremely thankful. I have also had the opportunity to take courses and collaborate with many of you and this has made the experience even more valuable.

There has been a multitude of people that have assisted me throughout the years during my research, including a number of graduate and undergraduate students.

I would like to acknowledge my lab mate Justin Goodsell, for all his helpful conversations, our learning pow-wows in my office, all the cheese fry runs and coffee breaks. You have truly been a great friend through this process and extremely helpful with your brain full of facts. I have to also thank Cory Lee, and Laura Molina, undergraduate researchers, for their hard work and diligence in the lab that led to progression of the research. I would like to thank one of my really great friends, Corey

Stedwell, for always being there when needed. We have shared many great moments together, both professionally and recreationally. We also had the opportunity to collaborate and publish a paper together. He has always served as a great confidant, and provided some great guidance from his experiences to help me throughout my graduate career. I would like to thank Matt Burg, a friend from day one of our graduate program together and one that will last, for his help and insight into all the biochemical techniques I annoyed him to teach me. He has been a great source of information for me throughout my career, and even harbored the start a new collaboration. I would like to also thank Adam Cismesia (lunchbox) for making sure there was always coffee and snacks every day to keep us going.

I would like to thank Andrew Ozarowski, located at the National High Magnetic

Field Laboratory (NHMFL), for his help and assistance in my many visits to the Mag

Lab. I would like to also thank Todd Prox, and Brian Smith for all their assistance in the design and construction of all and any necessary components essential to my research.

Behind all of my successes stands the team of people who have help mold me into the man I am today. I would like to give a special thanks to my parents, Mohamed and Bibi

Twahir, and my uncle and aunt, Shaffieq Chace and Rajdai Singh for their unwavering

support through my entire education. I would like to thank them for making sure they did all they could for me to leave me short of nothing. I will forever be thankful to my wife

Saudia Ally for always standing by me through this process, supporting me and keeping me grounded. She has been patient with me through all my moods, and crazy ideas, always keeping firm her willingness to follow me down any path I chose. Funding for this work was provided by the National Science Foundation under Grant CHE-1213440 and the National Institute of Health Instrumentation Grant NIH S10 RR031603. The high- field EPR spectra were recorded at the NHMFL, which is funded by the NSF through the

Cooperative Agreement no. DMR-1157490, the State of Florida and the U.S. DOE.

TABLE OF CONTENTS

page

ACKNOWLEDGMENTS ...... 4

LIST OF TABLES ...... 10

LIST OF FIGURES ...... 11

LIST OF SCHEMES ...... 15

LIST OF ABBREVIATIONS ...... 16

ABSTRACT ...... 18

CHAPTER

1 BACKGROUND ...... 20

Introduction ...... 20 Oxalic Acid ...... 20 Bacillus subtilis Oxalate Decarboxylase Structure and Reactivity ...... 20 Research Objectives and Overview ...... 28

2 ELECTRON PARAMAGNETIC RESONANCE ...... 30

Electron Paramagnetic Resonance for Studying Bacillus subtilis Oxalate Decarboxylase ...... 30 Spin Hamiltonian ...... 31 Electron and Nuclear Zeeman Effect ...... 32 Hyperfine Interaction - Nuclear Quadrupole Interaction ...... 34 Identification of Spin Trapped Adducts via the Hyperfine Interaction ...... 37 Zero Field Splitting-Application to Mn(II) ...... 39 High Field/Frequency EPR ...... 42 Parallel Mode EPR-Application to Mn (III) S=2 ...... 45

3 IDENTIFICATION OF SUPEROXIDE PRODUCTION UNDER TURNOVER CONDITIONS FOR Bacillus subtilis OXALATE DECARBOXYLASE ...... 48

Introduction ...... 48 EPR Spin-Trapping ...... 50 Experimental ...... 51 Materials ...... 51 Electron Paramagnetic Resonance ...... 52 Mass Spectrometry ...... 53 Hydrogen Peroxide Fluorescence Assay ...... 54 Results ...... 54

Spin-trapping of Radicals Produced from Wild-Type OxDC ...... 54 Spin trapping of Radicals Produced with T165V ...... 60 Quantitative Analysis of Radical Production ...... 64 Discussion ...... 66

4 REDOX CYCLING OF THE METAL CENTERS OF Bacillus subtilis OXALATE DECARBOXYLASE: IDENTIFICATION OF Mn(III) ...... 73

Introduction ...... 73 Experimental ...... 76 Materials ...... 76 Electron Paramagnetic Resonance Studies ...... 76 Enzyme Oxidation Experiments ...... 77 Results ...... 77 X-band Redox Cycling ...... 77 High-Field/Frequency Redox Cycling ...... 85 Discussion ...... 88

5 IMMOBILIZATION OF Bacillus subtilis OXALATE DECARBOXYLASE ON A ZN-IMAC RESIN ...... 94

Introduction ...... 94 Review of Previous EPR pH Dependent Structures ...... 96 Experimental ...... 98 Materials ...... 98 Resin Preparation and Enzyme Immobilization ...... 99 Enzyme Kinetic Assays ...... 100 Microscopy ...... 100 Multi-Field/Frequency Electron Paramagnetic Resonance ...... 100 Results ...... 101 X-Band EPR and Enzyme Kinetics ...... 101 Multi-Field/Frequency Electron Paramagnetic Resonance ...... 104 Effects of Freeze-Thaw Cycles on Resin and Immobilized Enzyme ...... 110 Discussion ...... 116

6 CONCLUSIONS AND FUTURE OUTLOOK ...... 122

Introduction-Preliminary X-Ray Structural Analogs ...... 122 Protein Crystallography ...... 123 Material and Methods ...... 124 Preliminary Data ...... 125 Long Range Electron Transfer Mechanism ...... 128 Conclusion ...... 131

APPENDIX

A EXPRESSION AND PURIFICATION OF Bacillus subtilis OXALATE DECARBOXYLASE ...... 134

B X-RAY DIFFRACTION DATA COLLECTION, PROCESSING AND REFINEMENT PARAMETERS ...... 136

LIST OF REFERENCES ...... 138

BIOGRAPHICAL SKETCH ...... 151

LIST OF TABLES

Table page

3-1 Spectral parameters of BMPO-radical adducts for WT-OxDC, including hyperfine coupling, g-value, and individual spectral weights based on the simulated spectra...... 56

3-2 Spectral parameters of BMPO-radical adducts for T165V OxDC mutant, including hyperfine coupling, g-value, and individual spectral weights based off simulated spectra...... 61

5-1 Site specific zero field splitting parameters of Mn (II) sites in WT-OxDC ...... 97

5-2 Michaelis-Menten kinetics of free and immobilized OxDC...... 103

5-3 Total protein on IMAC resin after initial loading and subsequent freeze-thaw cycles...... 112

5-4 Michaelis-Menten kinetics of free and immobilized OxDC after freeze-thaw cycling...... 115

6-1 Forward and reverse primer sequences to generate W96F, W274F, and W96F/W274F OxDC site directed mutants...... 129

B-1 Data collection and refinement statistics on (LS-CAT) beam-line 21-ID-F for pH 4.6 OxDC crystal...... 136

B-2 Data collection and refinement statistics on R-AXIS VII++ for pH 4.6 OxDC crystal...... 137

LIST OF FIGURES

Figure page

1-1 Crystal structure of Bacillus subtilis oxalate decarboxylase with the manganese centers represented as gray spheres...... 21

1-2 A) N-terminal active site with formate and one water molecule bound and B) C-terminal active site two water molecules coordinated to the metal with manganese represented by purple spheres...... 22

1-3 Crystal structure of wild-type oxalate decarboxylase of the trimer unit formed. Zoomed portion of the structure shows the tryptophan dimer (W96 and W274) that sits at the interface of the monomer units...... 27

2-1 Representation of the Electron Zeeman Interaction of S = 1/2 system, placed in a magnetic field ...... 33

2-2 A) Energy level diagram depicting the splitting that arises for a system with S = 1/2 and I = 1 ...... 35

2-3 Simulated spectra of DMPO spin adducts formed with superoxide (blue trace), hydroxyl (orange trace), and methyl (yellow trace) radicals...... 38

2-4 Simulated spectra of Mn(II) and respective energy level diagrams with allowed and forbidden transitions shown in red and grey vertical lines respectively...... 41

2-5 Simulated spectrum of Mn(II) with spectral parameters of 400 GHz microwave frequency, 5K temperature, 249 MHz hyperfine coupling, g = 2.001, and zero-field splitting values of D = 0.1 cm-1 and E/D = 0.3...... 43

2-6 Simulated spectra of a tyrosyl radical at 20K ...... 44

2-7 Energy level diagram for Mn(III) with magnetic parameters of S = 2, g = 2.00, A = 140 MHz, D = -2.38 cm-1, and E/D = 0.13. A) Excitation along the z-axis (parallel mode), B) with excitation along the y-axis (perpendicular mode)...... 47

3-1 X-band EPR spectra of the BMPO-radical adducts produced from the reactions between oxalate and WT-OxDC in 50 mM citrate buffer pH 4.0 (black) ...... 55

3-2 Time-Resolved X-band EPR spectra of the BMPO-radical adducts produced from the reactions between oxalate and WT-OxDC in 50 mM citrate buffer pH 4.0 ...... 57

3-3 ESI-Q-MS analysis of BMPO-superoxide reaction mixture ...... 58

3-4 ESI-Q-MS analysis of control mixture, with enzyme omitted ...... 59

3-5 X-band EPR spectra of the BMPO-radical adducts produced from the reactions between oxalate and OxDC mutant T165V in 50 mM citrate buffer pH 4.0 (black) ...... 61

3-6 Time-Resolved X-band EPR spectra of the BMPO-radical adducts produced from the reactions between oxalate and OxDC mutant T165V in 50 mM citrate buffer pH 4.0 ...... 62

3-7 ESI-Q-MS analysis of T165V mutant BMPO-superoxide reaction mixture ...... 63

3-8 EuTc hydrogen peroxide fluorescence assay. Standard concentrations of H2O2 were prepared () and mixed 1:1 with EuTc working solution ...... 64

3-9 From left to right for WT (yellow) and T165V (gray): concentrations of the •– •– BMPO radical adduct for the CO2 and the O2 radical ...... 65

4-1 Redox cycling of (25 mg/mL) wild-type OxDC in 50 mM succinate, 500 mM NaCl poised at pH 4.2 using hexachloroiridate and ascorbate as oxidizing and reducing agents respectively ...... 78

4-2 Redox cycling of (25 mg/mL) wild-type OxDC in 50 mM succinate, 500 mM NaCl poised at pH 4.2 using hexachloroiridate and ascorbate as oxidizing and reducing agents respectively ...... 79

4-3 Redox cycling of (25 mg/mL) wild-type OxDC in 50 mM phosphate, 500 mM NaCl poised at pH 8.0 using hexachloroiridate and ascorbate as oxidizing and reducing agents, respectively ...... 80

4-4 Redox cycling of (25 mg/mL) wild-type OxDC in 50 mM phosphate, 500 mM NaCl poised at pH 8.0 using hexachloroiridate and ascorbate as oxidizing and reducing agents, respectively ...... 81

4-5 Experimental and simulated parallel mode spectra for WT-OxDC pH 4.2. A) Mn (II) prior to oxidation (blue trace) ...... 82

4-6 Temperature dependence between 5K to 90K of Mn(III) generated in oxidized wild-type OxDC (25 mg/ml) in 50 mM succinate, 500 mM NaCl poised at pH 4.2 using hexachloroiridate as oxidizing agent ...... 83

4-7 Temperature dependent simulations of Mn(III) generated in oxidized wild- type OxDC at 5K (blue), 10K (red), 20K (orange), 50K (purple) and 90K (green) for confirmation of the sign of D ...... 84

4-8 Redox titration of (25 mg/mL) wild-type OxDC in 50 mM succinate, 500 mM NaCl poised at pH 4.2 using hexachloroiridate and ascorbate as oxidizing and reducing agents, respectively at 406.4 GHz HF-EPR ...... 86

4-9 Redox titration of (25 mg/mL) wild-type OxDC in 50 mM phosphate, 500 mM NaCl poised at pH 8.0 using hexachloroiridate and ascorbate as oxidizing and reducing agents, respectively at 406.4 GHz HF-EPR...... 86

4-10 Tyrosyl radical generated upon redox titration of (25 mg/mL) wild-type OxDC in 50 mM phosphate, 500 mM NaCl poised at pH 8.0 using 32 mM hexachloroiridate at 406.4 GHz ...... 87

5-1 Free (black trace) and Immobilized (red trace) WT-OxDC at 5 K...... 102

5-2 Comparison of Michaelis-Menten kinetic analysis of decarboxylase activity of free OxDC at pH 4.2 (blue) and pH 5.5 (red) as well as IMAC-bound OxDC at pH 4.2 (green) and pH 5.5 (cyan) ...... 103

5-3 The 406.4 GHz HF-EPR spectra at 3K of WT-OxDC free enzyme at pH 4.06 (red), immobilized at pH 3.88 (black) and simulation (blue) based on site A previously determined magnetic parameters shown in Table 5-1 ...... 105

5-4 406.4 GHz HF-EPR at 3 K of free enzyme at pH 8.42 (red) and immobilized at pH 8.50 (black) ...... 106

5-5 The 406.4 GHz HF-EPR spectra at 20K and two pHs of WT-OxDC free enzyme A), and immobilized B) and simulation based on site H (blue) previously determined magnetic parameters shown in Table 5-1 ...... 107

5-6 pH dependence of the 406.4 GHz HF-EPR spectra at 20K of A) free enzyme, B) immobilized ...... 108

5-7 The 406.4 GHz HF-EPR at 20K spectra of WT-OxDC free enzyme (black), immobilized (red) and respective simulations for site L (blue) based on previously determined magnetic parameters shown in Table 5-1 ...... 110

5-8 Free and Immobilized WT-OxDC and the effects of freezing and thawing the resin bound enzyme ...... 111

2 2 5-9 Magnetic (B1 ) and electric (E1 ) fields across a TE102 cavity, with the rectangular border representing the walls of the resonator...... 113

5-10 Comparison of fresh, unfrozen, Zn-loaded resin observed under A) 20x and C) 4x magnification, and OxDC-loaded resin exposed to 6 freeze-thaw-wash cycles observed under B) 20x and D) 4x magnification ...... 114

5-11 A pH titration of WT-OxDC immobilized on a Zn-IMAC resin using a poly- buffer posed at desired pH ...... 115

6-1 Crystal structure of Bacillus subtilis oxalate decarboxylase hexameric structure with one trimer unit represented in orange and the other in blue at pH 4.6 at 2.1 Å with the manganese centers represented as purple spheres .. 125

6-2 N-terminal active site of the pH 4.6 OxDC structure showing manganese (purple sphere) coordinated by His95, His 97, His 140, and Glu101, and acetate ...... 126

6-3 A) N-terminal active site of the pH 4.6 OxDC structure overlaid with the sequence SENST 161-165 from PDB: 1J58 (orange) and 1UW8 (blue) with the glutamate residue represented as sticks for all three structures ...... 127

6-4 Overlay of PDB: 1J58 with MTSL spin label at sites V222 (N-terminal, blue label) and V243 (C-terminal, magenta label)...... 130

LIST OF SCHEMES

Scheme page

1-1 A) Decarboxylase reaction and B) oxalate oxidase reaction...... 23

1-2 Proposed mechanism for the decarboxylation mechanism in OxDC ...... 24

2-1 Reaction of 5,5-dimethyl-1-pyrroline N-oxide (DMPO) with radical species RX. The radical species binds across the double bond of the pyrroline ring adjacent to the nitrogen, generating a persistent radical...... 37

3-1 A) Reaction that occurs between 5,5-dimethyl-pyrroline N-oxide (DMPO) spin-trap and superoxide ...... 51

4-1 Proposed reaction mechanism of Oxalate Oxidase. Mono-protonated oxalate binds to Mn, followed by the binding of dioxygen and generation of superoxide and Mn(III) ...... 75

LIST OF ABBREVIATIONS

B. sub Bacillus subtilis

BMPO 5-tert-butoxycarbonyl-5-methyl-1-pyrroline-N-oxide

DMPO 5,5-dimethyl-1-pyrroline-N-oxide

E. coli Escherichia coli eff Effective

EMR Electron Magnetic Resonance

ENDOR Electron Nuclear Double Resonance

EPR Electron Paramagnetic Resonance

ESEEM Electron Spin Echo Envelope Modulation

ESR Electron Spin Resonance

GHz Gigahertz

Glu Glutamate

His Histidine

Iso Isotropic

KIE Kinetic Isotope Effect

LRET Long Range Electron Transfer

MHz Megahertz mM mili Molar mT mili Tesla

NMR Nuclear Magnetic Resonance

OxDC Oxalate Decarboxylase

OxOx Oxalate Oxidase

PBN N-tert-butyl-α-phenylnitrone

PCET Proton Coupled Electron Transfer

POBN α-(4-Pyridyl N-oxide)-N-tert-butylnitrone

QM Quantum Mechanics

SOC Spin Orbit Coupling

Trp Tryptophan

WT Wild-Type

ZFI Zero-Filed Interaction

ZFS Zero-Field Splitting

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy

INVESTIGATIONS INTO THE ENZYMATIC MECHANISM OF BACILLUS SUBTILIS OXALATE DECARBOXYLASE: AN ELECTRON PARAMAGNETIC RESONANCE APPROACH

Umar Twahir

December 2015

Chair: Alexander Angerhofer Major: Chemistry

Oxalate Decarboxylase from Bacillus subtilis is a Mn-dependent enzyme that catalyzes the degradation of oxalate to carbon dioxide and formate during 99.8% of turnovers, while the other 0.2% leads to 2 equivalents of carbon dioxide and hydrogen peroxide. OxDC is a member of the bicupin superfamily of enzymes, with a mononuclear Mn(II) coordinated in each of its two domains, with the N-terminal manganese designated as the site of catalysis, with no definitive role assigned to the C- terminal manganese. The proposed mechanism employs oxygen as a co-catalyst necessary to drive the initial oxidation step in the catalytic cycle, suggesting Mn(III) is present while superoxide is generated. Mn(III) has also been seen in related enzymes, where the Mn ion cycles through oxidation states during turnover, suggesting that OxDC might follow a similar trend.

Superoxide production was studied utilizing both EPR spin trapping and mass spectrometry. The superoxide and previously identified carbon dioxide anion radical were spin-trapped simultaneously suggesting distinct sites of production. The source of superoxide production is likely due to dioxygen binding at one of the manganese sites,

oxidizing the metal center to Mn(III). In order to directly observe Mn(III) chemical oxidation of the enzyme was carried out, and studied using parallel-mode X-band EPR and Mn(II) oxidation using High-Field/Frequency EPR. The observed Mn(III) EPR signature was then also observed in as-prepared enzyme, and exhibits a strong pH dependency. The pH dependent studies were conducted using an enzyme immobilization technique that allows for exposure of multiple buffer conditions on a single sample. OxDC is only active at low pH where the majority of the Mn(III) signal was observed.

The combination of these results suggests that at least one metal center cycles between the +2 and +3 oxidation state, and oxygen binding is at a site different than N- terminal manganese center. A possible site for oxygen binding is the C-terminal manganese, which would then require the mechanism to propagate via a long-range electron transfer. A tryptophan dimer pair involved in a π-π stacking motif at the monomer interfaces within a single trimer unit may facilitate the electron transfer mechanism.

CHAPTER 1 BACKGROUND

Introduction

Oxalic Acid

Oxalic acid, H2C2O4, is the most common naturally occurring dicarboxylic acid and has shown toxicity to most mammals and leads to diverse physiological and industrial issues [1]. It is a relatively strong acid with pKa’s at 1.2 and 4.2. Oxalic acid has the ability to form both soluble and insoluble salts with metals. Calcium oxalate is an important sparingly soluble compound at room temperature and neutral pH. Several biosynthetic pathways have been proposed for the production of oxalic acid, including the oxidation of glycolate and glyoxylate by glycolate oxidase, as well as the action of isocitrate lyase on isocitrate [2-5]. Approximately 60% of urinary tract stones contain calcium oxalate [6]. Excessive consumption and elevated concentrations of oxalic acid have shown to cause urinary tract stones, hyperoxaluria, hypocalcaemia, cardiomyopathy, cardiac conductance, and mortality [6-13]. In addition to crop fungal infections, oxalic acid scaling in the wood and paper industry have presented major issues requiring remediation [14-19].

Bacillus subtilis Oxalate Decarboxylase Structure and Reactivity

Oxalate decarboxylase was first identified approximately 50 years ago and was initially studied and found to be expressed in a variety of fungi including, Collybia velutipes, Coriolus hersutus, Sclerotinia sclerotiorum, Coriolus versicolor, Myrothecium verrucaria, Aspergillus niger and Agaricus bisporus [20-26]. Due to low levels of expression in fungi and the lack of a recombinant overexpression system, attempts to carry out studies difficult. Identification of bacterial oxalate decarboxylase isolated from

Bacillus subtilis, provided a system that could be exploited to produce larger quantities of active recombinant protein [27].

A B

Figure 1-1. Crystal structure of Bacillus subtilis oxalate decarboxylase with the manganese centers represented as gray spheres A) Hexameric structure with one trimer unit represented in orange and the other in blue B) Side-view of the hexamer C) Monomer unit. Structures were generated using PDB file 1J58 in MacPyMol.

Oxalate decarboxylase (OxDC) is a member of the cupin superfamily of proteins

[28-30], characterized by a series of conserved residues that form β-barrels to support the binding of a range of metallo-cofactors [31-33]. High-resolution X-ray crystallography of OxDC provided evidence that the monomer is composed of two β-barrel domains, each possessing a manganese ion [28, 30]. Electron paramagnetic resonance

confirmed the resting oxidation state of the metal centers to be +2 [34-36]. The enzyme exists as a homo-hexamer comprised of a dimer of stacked trimers as shown in Figure

1-1 A and B, with one trimer in blue and the other in orange and the monomer shown in

Figure 1-1 C. A side view of the hexameric structure shown in Figure 1-1 B, illustrates the alpha helical arms of one trimer unit protruding into the neighboring trimer unit facilitating the stacked structure. The inter-manganese distance in a single monomer unit is 26 Å, while the inter-manganese distance between the N- and C-terminal sites of neighboring monomers within a single trimer unit is 21 Å. The manganese ions are coordinated in a pseudo-octahedral environment in both domains by 4 conserved residues. The N-terminal Mn is coordinated by His95, His 97, His 140, and Glu101

(Figure 1-2 A) and in the C-terminal domain the residues are His273, His275, His319, and Glu280, as shown in Figure 1-2 B.

A B

Figure 1-2. A) N-terminal active site with formate and one water molecule bound and B) C-terminal active site two water molecules coordinated to the metal with manganese represented by purple spheres.

The N-terminal site has been designated as the active site, since it has been shown to also coordinate formate, which is one product of catalysis, with the second open valence coordinated by water. When formate is not present, the N-terminal manganese has an additional water molecule coordinated. The C-terminal manganese has been shown to coordinate either one or two water molecules [28, 30].

OxDC shows an impressive enzymatic rate enhancement of 2.51013 at an optimum pH of 4.2 [37]. A very interesting but poorly understood aspect of this enzyme is the apparent bifurcation of its chemistry. Wild-type (WT) OxDC primarily acts as a decarboxylase to produce carbon dioxide and formate (99.8% of all turnovers), while

OxDC acts as an oxidase in about 0.2% of all turnovers producing hydrogen peroxide and carbon dioxide (Schemes 1A and B) [38, 39].

Scheme 1-1. A) Decarboxylase reaction and B) oxalate oxidase reaction.

Although the primary reaction catalyzed by OxDC is a redox-neutral disproportionation reaction and does not consume dioxygen, it requires dioxygen for turnover as a co-catalyst [21, 26, 38-40]. Current mechanistic proposals suggest that dioxygen is bound to one of the two Mn ions and acts as a transient electron sink to destabilize the carbon-carbon bond in oxalate resulting in a superoxide radical. Recent

EPR spin trapping experiments combined with mass spectrometry provided evidence for superoxide production during catalysis [41]. After decarboxylation has taken place, superoxide acts as an electron source to reduce the resulting carbon dioxide anion radical as shown in Scheme 1-2 [42]. This mechanism requires oxygen to cycle through its 1-electron reduced state as superoxide which would likely be protonated in the pH

range in which the enzyme is active [43]. Alternative proposals suggest one of the Mn ions undergoing a redox cycle between its +2 and +3 oxidation states allowing the associated dioxygen to remain as a hydroperoxyl radical throughout the reaction [29,

44, 45].

A B

D C

Scheme 1-2. Proposed mechanism for the decarboxylation mechanism in OxDC. A) Mono-protonated oxalate and dioxygen bind to the manganese center, generating Mn(III). B) Glutamate-162 abstracts a proton from oxalate leading to a PCET, followed by heterolytic cleavage of oxalate and release of carbon dioxide. C) The remaining carbon dioxide anion radical then re-abstracts a proton from glutamate generating D) bound formate [42].

The order of binding of dioxygen and oxalate to the metal center remains unclear. In any case upon binding of mono-protonated oxalate in a mono-dentate fashion, it is proposed that the metal center is oxidized to Mn(III). Glutamate-162 is believed to act as a transient base, abstracting a proton from the oxalate mono-anion, generating an oxalate radical intermediate, immediately followed by an electron transfer and the heterolytic cleavage of the unreactive carbon-carbon bond of oxalate.

Kinetic isotope experiments (KIE) provided evidence for the proton coupled electron transfer (PCET) in the first step of the mechanism [40]. More recently QM calculations have attempted to shed light on the radical mechanism of oxalate degradation, demonstrating the overall lower energy for the cationic and neutral radical intermediates [46]. More so, it is suggested that the removal of an electron to form a cation intermediate is sufficient to cause the cleavage of the carbon-carbon bond, even without the simultaneous loss of a proton. Nevertheless, the cleavage of the carbon- carbon bond leads to the release of carbon dioxide, the first product of catalysis, leaving a carbon dioxide anion (formyl) radical bound to the manganese center. Previous EPR spin trapping experiments provided evidence that this radical is produced and can be released in solution [42].The radical center can then re-abstract a proton from glutamate-162, now acting as a transient acid, generating a formate bound to the manganese center, which can then be released in solution allowing for another round of catalysis. The re-protonation of the radical bound intermediate is thought to be the point in the mechanism which dictates whether the enzyme will act as a decarboxylase or oxidase [29, 30, 47]. If the carbon dioxide anion radical is not re-protonated, it will react with the bound superoxide and generate a peroxycarbonate intermediate. The intermediate will then spontaneously decay to produce the second equivalent of carbon dioxide and hydrogen peroxide.

In an effort to better understand the bifurcation that occurs in the mechanism, it is instructive to compare oxalate decarboxylase to oxalate oxidase, and consider both structural and mechanistic similarities and differences. Oxalate oxidase occurs in both the mono and bi-cupin forms of the enzyme. Their quaternary structure is similar to

oxalate decarboxylase and are thought to have an evolutionary relationship [33, 48].

Both OxDC and OxOx enzymes require manganese, dioxygen, and oxalate for catalysis and both belong to the cupin superfamily [38]. The same four conserved residues coordinating the metal center in OxDC are also conserved in OxOx, 3 histidine’s and 1 glutamate [28-30, 49].

Sequence alignment studies of both Bacillus subtilis OxDC and Ceriporiopsis subvermispora OxOx identified a major structural difference between these two enzymes. The existence of an N-terminal active site flexible lid in OxDC, consisting of a serine-glutamate-arginine-serine-threonine sequence at positions 161-165, was found in

OxDC versus aspartate-alanine (serine)-serine-asparagine-glutamine in OxOx [47].

Site-directed mutagenesis of this lid region can modulate OxDC activity, providing a switch controlling the oxidase activity, leading to the ability to convert OxDC into an oxidase enzyme [47]. For decarboxylation to occur the presence of a protonating group, tentatively assigned to be glutamate-162 in OxDC, appears to be necessary and an analogous amino acid is non-existent in OxOx [44, 47]. This lid has also been shown to act as a gate to a solvent channel that leads from the surface of the protein to the active site, allowing the entry and exit of substrate and product molecules.

One major unknown that still exists is the role the C-terminal manganese during catalysis. Due to their structural similarity as shown in Figure 1-2, it was initially thought that the C-terminal site was the active site due to the presence of the Glu-333 thought to act as the necessary transient acid/base during catalysis. Mutation of this site to alanine led to a factor of 25 decrease in the decarboxylase activity, indicating a critical role in catalysis. After the formate was shown to interact with the N-terminal manganese, and

the presence of the glutamate residue of the lid gating this site, the role of the C- terminal manganese again was unknown. Later work suggested that this site may act as secondary active site, or may simply play a structural role [30, 50]. More recent EPR investigations of the pH dependent studies of OxDC revealed that there are at least 6 different sets of magnetic parameters of manganese, 4 of which belong to the C- terminal site, suggesting that this site plays more than a structural role, because of the variability it exhibits [36].

Figure 1-3. Crystal structure of wild-type oxalate decarboxylase of the trimer unit formed. Zoomed portion of the structure shows the tryptophan dimer (W96 and W274) that sits at the interface of the monomer units, connecting the N- and C-terminal manganese sites of neighboring monomers within a single trimer.

As mentioned above, the distance between manganese centers of neighboring monomers within a trimer is shorter than the distance within a single monomer. The site of oxygen binding in OxDC is still unknown, and could potentially be the C-terminal

manganese. This would then require an electron shuttling pathway between the N-and

C-terminal manganese. Furthermore, there is a tryptophan dimer pair derived from W96 and W274 that sits between the N- and C-terminal manganese sites of neighboring monomers. The W96/W274 dimer drastically reduces the distance to about 9 to 10Å electrons would have to “hop” when moving between the two metal sites. With the shorter distance between the two redox centers, there is precedence for tunneling mechanisms in other proteins involving tryptophan residues which allows for a long- range electron transfer mechanism (LRET) [51-54]. Identification of the oxygen-binding site within the enzyme is critical to determine if the enzyme mechanism is dependent on a LRET.

Research Objectives and Overview

The overarching goals of this research are motivated by the bifurcated mechanism that exists for OxDC and its ability to carry out two distinct yet related chemical reactions. The goal of this introduction is to provide sufficient background of the enzymatic system of interest. This research will employ a variety of techniques including bioinorganic chemistry, enzyme kinetics, and molecular spectroscopy with a focus on Electron Paramagnetic Resonance (EPR). Chapter 2 will provide a theoretical framework for the employment of EPR and how it is utilized to study oxalate decarboxylase. Chapter 3 describes EPR spin-trapping experiments carried out to identify radicals produced during enzyme turnover. Chapter 4 focuses on the oxidation state of both metal centers in the enzyme, their ability to be redox cycled, and the implication on the overall enzyme mechanism. Chapter 5 describes an immobilization technique for his-tagged proteins applicable to low temperature EPR studies. Chapter 6 will provide discussion of the current and future directions of the project, including

protein crystallography and development of new site-directed mutants for EPR studies.

Also, a brief summary of the results from each chapter and a discussion of their implications will be given.

CHAPTER 2 ELECTRON PARAMAGNETIC RESONANCE

Electron Paramagnetic Resonance for Studying Bacillus subtilis Oxalate Decarboxylase

As described in Chapter 1, Bacillus subtilis oxalate decarboxylase is a member of the cupin superfamily of proteins, which is characterized by a series of conserved residues that form β-barrels that facilitate the incorporation of metallo-cofactors necessary for catalysis [28, 30, 32, 33, 55]. OxDC has been previously shown to require manganese for catalysis, incorporating a mononuclear manganese in each of the cupin folds [50]. Previous EPR studies have shown that the metal incorporated is manganese and the resting oxidation state is +2 [34, 36]. Manganese is a redox active transition metal that is incorporated into a variety of proteins and used as a cofactor during catalysis. EPR can provide access to both structural and electronic structure information for transition metals coordinated in proteins, as well as information about the local environment of the metal center. EPR can also be utilized in the identification of transient and long-lived radical species.

OxDC has also been shown to exhibit a radical mediated mechanism, producing superoxide and carbon dioxide anion radicals during catalysis [28, 35, 42, 56]. As such,

EPR can provide a wealth of information concerning the overall mechanistic scheme of

OxDC, exploiting the metal centers as intrinsic probes before, during and after catalysis.

In combination with low temperature EPR and EPR spin trapping, radicals generated can be studied, providing information about concentrations, location, and kinetics. The goal of this chapter is to introduce EPR and provide a theoretical framework for the experiments and results presented in this dissertation.

Spin Hamiltonian

Electron paramagnetic resonance, also referred to as electron spin resonance

(ESR), or electron magnetic resonance (EMR), is utilized in the study of paramagnetic species. A paramagnetic sample can be an ion or a molecule that contains unpaired electrons and leads to the generation of a permanent magnetic moment. Upon the application of a magnetic field the randomly organized magnetic dipoles, considered as a macroscopic system comprised of an ensemble of spins, orients either parallel or anti- parallel to the magnetic field, generating an overall net magnetization. The effects of the local magnetic and electronic environments on the electron spins can be observed and converted to electronic and structural information in the region of the unpaired spin. The energies of the ground state paramagnetic species is reflected in the static spin

Hamiltonian provided in Equation 2-1 [57, 58].

퐻 = 퐻퐸푍 + 퐻푍퐹푆 + 퐻퐻퐹 + 퐻푁푍 + 퐻푁푄 + 퐻푁푁 (2-1)

The effective spin Hamiltonian includes information concerning the total effective spin S, and its interaction with local nuclei bearing spin in its vicinity. The effective spin

Hamiltonian is comprised of an electron Zeeman interaction, HEZ, zero-field splitting,

HZFS, hyperfine interaction, HHF, nuclear Zeeman interaction, HNZ, nuclear quadrupole,

HNQ, and nuclear-nuclear interaction, HNN. The nuclear-nuclear interaction is usually ignored as its contribution is most often negligible compared to the remaining terms.

The effective spin Hamiltonian contains only the spin portion of the total Hamiltonian utilizing the S and/or I spin operators. With appropriate approximations, this sufficiently describes the paramagnetic species examined in this work. Each term will be discussed below in accordance with how it is used to describe the systems mentioned above.

Electron and Nuclear Zeeman Effect

In the simplest system with the total electron spin S = 1/2, the available spin quantum numbers are ms = ± 1/2, describing two potential spin states. At zero-field these two spin states are degenerate, therefore indistinguishable. In an externally applied magnetic field, the degeneracy is lifted due to alignment of the spin either parallel (ms = + 1/2) or anti-parallel (ms = - 1/2) with the magnetic field. The separation in the sub-levels is directly proportional to the strength of the magnetic field as shown in

Figure 2-1 A, and is referred to as the electron Zeeman effect [58]. The Zeeman term in the Hamiltonian is represented in Equation 2-2.

퐻퐸푍 = 훽푒퐵0𝑔푆 (2-2)

The electron Zeeman term includes the Bohr magneton, βe, magnetic field, B0,

Zeeman factor (g-factor), g, which is a proportionality constant that relates the gyromagnetic ratio to the magnetic moment, and total spin, S. The g-factor for a non- interacting electron is a scalar value, ge = 2.0023193. However, in an environment with non-spherical symmetry it is represented by a tensor resulting in three principal axes, gx, gy, and gz. There are three types of systems that are distinguished by the g-tensor components. In the isotropic case all g-tensor components are equal. When gx = gy ≠ gZ the system is axial; and when gx ≠ gy ≠ gz the g-tensor is rhombic. The ge for the free electron is well known and any deviation from this value arises from spin-orbit coupling

(SOC) leading to a geff, represented as g above. This effect arises from coupling of admixtures of ground and excited states of the orbital momentum to the angular momentum. This leads to the addition of another term to the Zeeman term known as

HLS = λLS, where L is the orbital angular momentum, S is the spin angular momentum

and λ representing the spin-orbit coupling interaction. Depending on the energy gap between the ground and excited states, variations from ge are observed, with smaller energy gaps leading to larger deviations and vice versa.

Figure 2-1. Representation of the Electron Zeeman Interaction of S = 1/2 system, placed in a magnetic field. A) Upon increasing the magnetic field, the separation in the ms sublevels is further increased proportional to the strength of the increasing magnetic field. B) Upon interaction with the appropriate electromagnetic radiation matching the energy separation in the ms sublevels, absorptions are observed.

In the basic EPR experiment, the sample is irradiated at a constant frequency; taken in this case to be 9.85GHz, and the magnetic field is swept. Once the splitting in the energy levels (ΔE) matches that of the microwave energy, the resonance condition

(ΔE = hν = βe B g) is met, and an absorption is observed as shown in the Figure 2-1 B at

351mT. The signal is generally recorded as a first derivative due to field modulation of the signal which is detected phase-sensitively.

The nuclear Zeeman effect describes the coupling of nuclear spins to the external magnetic field [58]. The nuclear Zeeman term (Equation 2-3) includes the nuclear magneton, βn, magnetic field, B0, nuclear Zeeman factor (g-factor), gn, and total nuclear spin, I.

퐻푁푍 = −𝑔푛훽푛퐵0퐼 (2-3)

Because the magnitudes of the nuclear magnetic moments are much smaller than those of the unpaired electrons (i.e. nuclear Zeeman interaction is 1/658 of the electron

Zeeman interaction for a proton), the nuclear Zeeman effect, HNZ exhibits much smaller perturbations on the EPR spectrum.

Hyperfine Interaction - Nuclear Quadrupole Interaction

The hyperfine term describes the coupling of the nuclear magnetic moment and the local magnetic field at the nucleus generated by electron magnetization as described by Equation 2-4 [58-60].

퐻퐻퐹 = 푆̃ 푨 퐼 (2-4)

The hyperfine term includes the vector representation of the electron (S) and nuclear (I) spins, whose coupling is described by the tensor, A. The hyperfine interaction can arise from coupling to the nucleus of the paramagnetic ion itself and to neighboring magnetic nuclei. In the simplest case, where couplings are isotropic, the interaction can be described by a scalar value, aiso. This most commonly arises from couplings of the electron spin with the nuclear spin on the same center. In the case where the unpaired electron is coupling to neighboring nuclei through space, A is represented as a traceless tensor with terms Axx, Ayy, and Azz.

The hyperfine interaction leads to additional splitting in the Zeeman energy levels, giving rise to additional splittings in the spectrum, as shown in Figure 2-2 A. The observed splittings can be calculated using the equation 2nI + 1, where n is the number of equivalent nuclei, and I is the nuclear spin. In the case of a system with S = 1/2, and I

=1, as in the case of an electron in the presence of nitrogen-14 (14N), a three line spectrum would be observed as in the Figure 2-2 B.

A B

Figure 2-2. A) Energy level diagram depicting the splitting that arises for a system with S = 1/2 and I = 1. The red lines represent the allowed transitions. B) Example EPR spectra that arises from a system with S = 1/2 and I = 1, with an isotropic hyperfine coupling constant.

The hyperfine tensor (A) can be expressed as a sum of the isotropic Fermi contact interaction (HF) and an electron-nuclear dipole-dipole coupling term (HDD). The

Fermi interaction is the nonzero probability to find the electron in the nucleus for an s- orbital due to the radial wave functions. In Equation 2-5 and 2-6, aiso represents the isotropic coupling term, and includes the magnetic moment, electron and nuclear

Zeeman factor, Bohr and nuclear magneton and the local electron spin density at the

2 nucleus, |휓0(0)| .

퐻퐹 = 푎푖푠표 푆̃ 퐼 (2-5)

2 휇 푎 = 0 𝑔 훽 𝑔 훽 |휓 (0)|2 (2-6) 푖푠표 3 ℏ 푒 푒 푛 푛 0

The dipolar term of the hyperfine interaction is described as shown in Equation 2-7. The dipolar term is orientation and distant dependent and couples the electron and nuclear spins via the traceless and symmetric tensor T shown in Equation 2-8. The elements of the tensor incorporate the distance between the unpaired spin and the local nuclei, r, magnetic moment, electron and nuclear Zeeman factor, Bohr and nuclear magneton and the ground state wave function of the system, 휓0.

퐻퐷퐷 = 푆̃ 푻 퐼 (2-7)

휇 3푟 푟 −훿 푟2 푇 = 0 𝑔 훽 𝑔 훽 〈휓 | 푖 푗 푖푗 | 휓 〉 (2-8) 푖푗 4휋ℏ 푒 푒 푛 푛 0 푟5 0

Nuclei with I ≥ 1, such as 14N and 2H, also have a quadrupole moment, arising from a non-spherical charge distribution around the nucleus. The Hamiltonian term that describes this effect is given in Equation 2-9 [61].

퐻푁푄 = 퐼̃ 푷 퐼 (2-9)

In this form the, the nuclear quadrupole is represented by the traceless tensor P, which is expanded in Equation 2-10.

푒2푞푄 퐻 = 푃 퐼2 + 푃 퐼2 + 푃 퐼2 = [(3퐼2 − 퐼(퐼 + 1)2) + 휂(퐼2 − 퐼2)] (2-10) 푁푄 푋 푋 푌 푌 푍 푍 4퐼(2퐼−1)ℏ 푍 푋 푌

The nuclear quadrupolar Hamiltonian describes the electrical charge distribution about the atom in IX, IY, and IZ, and includes the electric field gradient, eq, the quadrupole moment, Q, and the asymmetry parameter, η. Quadrupolar nuclei lead to small second- order effects, leading to shifts in the allowed transitions, and the appearance of forbidden transitions. Advanced EPR techniques such as Electron Spin Echo Envelope

Modulation (ESEEM) and Electron Nuclear Double Resonance (ENDOR), can be used

to measure quadrupolar couplings directly, which are observed as first order perturbations in these spectra.

Identification of Spin Trapped Adducts via the Hyperfine Interaction

The hyperfine interaction provides a sensitive method for the identification of local nuclei in the vicinity of the unpaired electron of interest. As such, EPR can be utilized for the identification of free radicals produced during enzyme mechanisms. The major difficulty in observing free radicals is their extremely short lifetimes and reactivity.

One commonly used methodology for the identification of free radicals is spin-trapping

[62, 63]. In order to extend the lifetime of such radical species, a cyclic or linear nitrone

(N-oxide of an imine) can be added to the reaction mixture that reacts with the radicals present. Some spin trapping agents are designed with high selectivity, while others react with a broader range of free radicals. One example of a spin trap with low selectivity is 5,5-dimethyl-1-pyrroline N-oxide (DMPO). DMPO is a commonly used spin- trap in the study of oxygen-, nitrogen-, sulfur- and carbon-centered radicals. An example of a reaction with DMPO is shown in Scheme 2-1.

Scheme 2-1. Reaction of 5,5-dimethyl-1-pyrroline N-oxide (DMPO) with radical species RX. The radical species binds across the double bond of the pyrroline ring adjacent to the nitrogen, generating a persistent radical.

The persistent radical that is formed is stabilized by the bulky side groups that protect the nitroxide from reacting. The persistent radical formed has a half-life of for example of t1/2 = 45 seconds for superoxide, versus the diffusion-controlled half-life of the radical

in free solution. Characteristic hyperfine coupling constants are expected for the persistent radical formed when the radical reacts with the spin trap. In the case of

DMPO, three examples are provided in Figure 2-3.

Figure 2-3. Simulated spectra of DMPO spin adducts formed with superoxide (blue trace), hydroxyl (orange trace), and methyl (yellow trace) radicals.

The hyperfine couplings for the spectra reflected in Figure 2-3 with the β and γ positions

β γ labeled, are aN = 39.79 MHz, aH = 31.95 MHz, and aH = 3.36 MHz for the superoxide

β adduct; aN = 41.78 MHz, and aH = 41.78 MHz for the hydroxyl adduct; and aN = 45.96

β MHz, and aH = 65.30 MHz for the methyl adduct [62, 64]. The differences observed in the hyperfine coupling are due to variations in the spin-density delocalization that results based on the electron withdrawing and donating nature of the radical species, as well as the possible conformational dynamics of the bound radical.

In some cases multiple radical species may be produced during catalysis, increasing the difficulty of making an assignment. When the radical species are

centered on different atoms, such as a mixture of the superoxide and methyl radicals, the assignment becomes simpler, due to the large difference in the hyperfine coupling of the β-hydrogen. When the radical species are centered on the same atom, such as multiple carbon-centered radicals, isotopic labeling can assist in identification, as further splittings in the spectrum will be observed [42].

Zero Field Splitting-Application to Mn(II)

Thus far, systems with S = 1/2 have been discussed. When S > 1/2, as in the case of biradicals, triplet states, and transition metals, the dipole-dipole interaction between electrons removes the ground state degeneracy in the ms sublevels at zero magnetic field. If each spin is regarded as an individual dipole, the dipole-dipole interaction is repulsive, leading to the lowest energy state where the two electrons are farther apart. The Hamiltonian that describes the zero-field splitting (zfs) is shown in

Equation 2-11, where D is the traceless and symmetric zero-field interaction tensor [58,

61].

퐻푍퐹푆 = 푆̃ 푫 푆 (2-11)

In the case of a transition metal such as high-spin manganese (II), the total spin is 5/2, and the zero-field interaction can have large perturbations on the spectra observed.

When manganese is not coordinated to any ligands, it exists in spherical symmetry.

Upon binding of ligands, the electronic structure of the manganese center changes based on the ligand type and symmetry, known as the ligand field theory, and dictates the zero-field interaction. As such, the zero-field interaction can be described by the components of its principal axes as shown in Equation 2-12.

2 2 2 퐻푍퐹푆 = 퐷푋푆푋 + 퐷푌푆푌 + 퐷푍푆푍 (2-12)

The zero-field interaction, D, can be treated in the same manner as the g and A tensor were treated. When D > 0, there are two primary situations, depending on whether the system exhibits axial or rhombic symmetry. This becomes more apparent when the

Hamiltonian is expressed in terms of axial and rhombic zero-field splitting terms, D and

E, respectively as shown in Equation 2-13, 2-14, and 2-15.

1 퐻 = 퐷 [푆2 − 푆(푆 + 1)] + 퐸(푆2 − 푆2) (2-13) 푍퐹푆 푍 3 푋 푌

3퐷 퐷 = 푍 (2-14) 2

(퐷 −퐷 ) 퐸 = 푋 푌 (2-15) 2

When D ≠ 0 and E = 0 the tensor has axial symmetry. When the sample exhibits

D ≠ 0 and E ≠ 0, the tensor has rhombic symmetry. The degree to which the tensor deviates from axial symmetry (referred to as “rhombicity”) is commonly represented as η

= E/D, where η is between 0 and 1/3. The zero-field interaction is extremely useful in biological EPR, specifically when metal ions are critical to catalysis. In combination with the g and A tensor, information can be garnered about the local environment and structure of the metal center, including local perturbations on the metal center during catalysis. This proves even more useful when protein crystallography and nuclear magnetic resonance (NMR) are not applicable.

A major difficulty arises when the magnitude of the zero-field interaction is much larger than that of the electronic Zeeman energy, leading to extremely complicated spectra. Under some circumstances the zero-field interaction is so large that the spectra can be broadened beyond detection.

A B

Figure 2-4. Simulated spectra of Mn(II) and respective energy level diagrams with allowed and forbidden transitions shown in red and grey vertical lines respectively with spectral parameters of 9.65 GHz microwave frequency, 5K temperature, 249 MHz hyperfine coupling, g = 2.001, and zero-field splitting values of A) D = 0 cm-1 and E/D = 0, B) D = 0.1 cm-1 and E/D = 0, C) D = 0.1 cm-1 and E/D = 0.3.

Examples of Mn(II) spectra and their respective energy level diagrams, that also include allowed and forbidden transitions, are provided in Figure 2-4. All spectra were generated utilizing magnetic parameters of g = 2.001, and A = 249 MHz at 9.65 GHz microwave frequency and 5K temperature. In order to illustrate the effect of the zero- field interaction on the manganese spectra and concomitant energy levels, the D and

E/D parameters were varied with parameters of (A) D = 0 cm-1 and E/D = 0, (B) D = 0.1 cm-1 and E/D = 0, (C) D = 0.1 cm-1 and E/D = 0.3. It can be clearly seen that in the

absence of zero-field splitting (Figure 2-4 A) a sextet of lines is observed due to the

55Mn nuclear spin of 5/2. Transitions from the higher spin manifolds are not observed because all sublevels occur at similar geff values, as shown in the energy levels diagram in Figure 2-4 A. Upon introduction of an axial zero field splitting of 0.1 cm-1, a shift at zero-field is observed from 0 GHz in the energy level diagram, and the geff of the allowed transitions are shifted from g ≈ 2.0, resulting in the spectrum provided in Figure

2-4 B. Upon increasing the rhombicity, (Figure 2-4 C) the spectral feature with the greatest intensity appears at approximately 10 mT, very different from the other two cases presented. Cases like this make it difficult to gain any information from the EPR spectra, and other methodologies may be required. One option is to carry out experiments at higher fields and frequencies where the Zeeman energy dominates the magnetic parameters, allowing for identification of species with larger zero-field interactions.

High Field/Frequency EPR

In the past two decades much effort has been exerted to extend the EPR experiment to higher fields and frequencies [65, 66]. With the advent of newer microwave and magnet technology, more improvements will be possible. Nevertheless, the current technologies have provided the ability to carry out routine experiments in magnetic fields up to 14.5 Tesla. In many systems, for example the manganese system in Figure 2-4 C, the large zfs, as compared to the electronic Zeeman energy, causes the zfs to dominate the spectrum. Looking back at the static Hamiltonian for the spin system, the field dependent parameters are the electron and nuclear Zeeman terms, while the hyperfine interaction, zero-field interaction and nuclear quadrupole are field independent. By increasing the field, the magnitude of the Zeeman energies become

larger, while other interactions remain unchanged. In the case of the manganese (II) system described above with zero-field splitting parameters D = 0.1 cm-1 and E/D = 0.3, the spectrum at 9.65 GHz does not provide much information. However, when the same system is studied at 400 GHz, more spectral features are observed, as shown in Figure

2-5.

Figure 2-5. Simulated spectrum of Mn(II) with spectral parameters of 400 GHz microwave frequency, 5K temperature, 249 MHz hyperfine coupling, g = 2.001, and zero-field splitting values of D = 0.1 cm-1 and E/D = 0.3.

The spectrum provided in Figure 2-5 resolves transitions between all six spin manifolds for high spin manganese (II), with transitions between the |±5/2> |±3/2> manifolds at approximately 13850 mT and 14750 mT, transitions between the |±3/2> |±1/2> at approximately 14100 mT and 14500 mT, and transitions between the |-1/2> |-1/2> spin manifolds at approximately 14250 mT.

In some instances the difficulty in analyzing EPR spectra are not always due to large zero-field splittings, and can result in overlapping signals due to multiple species with similar g-tensors or an anisotropic g-tensor comprised of very similar g-values. One such situation is the identification of amino acid radicals generated during catalysis.

A B

Figure 2-6. Simulated spectra of a tyrosyl radical at 20K with spectral parameters, AX = 30 MHz, AY = 26 MHz, AZ = 20 MHz hyperfine coupling, gx = 2.0068, gy = 2.0039, gz = 2.0018 and a microwave frequency of A) 9.65 GHz and B) 400 GHz.

The tyrosine radical shown in the Figure 2-6 A plays a major role in photosystem II, and is also generated during catalysis of OxDC. The g-tensor and A-tensor for the tyrosyl radical has been simulated as gx = 2.0068, gy = 2.0039, gz = 2.0018 and AX = 30 MHz,

AY = 26 MHz, AZ = 20 MHz [67, 68]. In Figure 2-6 A, the spectrum simulated for the tyrosyl radical at 9.65 GHz, while (Figure 2-6 B) shows the spectrum simulated at 400

GHz. At X-band frequencies (9.65 GHz), it would be extremely difficult to identify the source of the signal observed. However, at 400 GHz the g-tensor is fully resolved, and can be used for identification of the radical species. Figures 2-4, 2-5, and 2-6 demonstrate two major benefits in moving to higher fields and frequencies. However, these are not the only benefits. Other advantages include higher sensitivity, higher orientation selectivity, and increased low-temperature electron spin polarization.

Parallel Mode EPR-Application to Mn (III) S=2

EPR proves to be a sensitive technique in the study of biomolecules, specifically metalloenzymes. However, integer spin species referred to as non-Kramer’s ions, such as Mn(III), tend to have extremely large fine structure parameters compared to the

Zeeman energy and are not typically observable. Another factor that complicates the matter is that non-Kramer’s sublevels are split at zero-field and further split quadratically, unlike half-integer spin species (Kramer’s ions), which split linearly in the presence of an increasing magnetic field. The resonance condition changes to

(ℎ휐)2 = (𝑔̃훽퐵)2 + ∆2 (2-16) where the 𝑔̃ is an angle-dependent g-value, and ∆ represents the energy difference at zero field between pairs of spin levels. Due to the large splittings that arise between the energy levels, resonances are not typically observed for such species at X and Q-band frequencies. Most efforts have been focused on the utilization of High-Field EPR to observe signals from such species. However, due to experimental limitations preventing the use of resonators at higher frequencies, larger sample quantities and concentrations than are routinely feasible may be required for traditional EPR experiments. Another methodology for studying integer spin species is introduction of the microwave polarization parallel to the applied field, rather than the typical perpendicular excitation utilized in studying half-integer spin species [61]. The Hamiltonian that describes a spin system with strong axial splitting and a rhombic component for S = 2, and with polarization in the z plane is

1 퐻 = 퐷 {푆2 − 푆(푆 + 1)} + 퐸(푆2 − 푆2) + 𝑔 훽퐵푆 cos 휃 + 𝑔 훽퐵푆 sin 휃 . (2-17) 푍 3 푥 푦 ∥ 푍 ⊥ 푧

When the axial splitting term D is much larger than the Zeeman energy, the energy levels for the system are

푚푠 | ± 2 > 푊±2 = +2퐷 ± 2𝑔∥훽퐵푐표푠휃 (2-18)

푚푠 | ± 1 > 푊1 = −퐷 ± 𝑔∥훽퐵푐표푠휃 (2-19)

푚푠 |0 > 푊0 = −2퐷 (2-20)

In the case where E = 0, there are no allowed transitions in either doublet in first order, however state mixing occurs between ±1 and 0 states and result in a form |+1,0,-1> .

This leads to extremely weak transitions for any  ≠ 0 due to second-order Zeeman effects. When E ≠ 0, an admixture of states occurs for both the ±1 and ±2 spin manifolds. When  ≠ 0 transitions can be seen to arise in both doublets in the form of

2 2 ½ W± = ±½{(𝑔̃||βHcos) + ∆ } , where the ±2 manifold has 𝑔̃|| = 4𝑔|| and a transition

2 energy of Δ2 = 12E /(W2-W0), and the ±1 manifold has 𝑔̃|| = 2𝑔|| and a transition energy of Δ1 = 6E. The spin manifold is treated as a new class of basis sets with symmetric and anti-symmetric sets of |2s>, |2a>, |1s> and |1a> , which are linear combinations of the

1 1 type |2푠 > = {| + 2 > +| − 2 >} and |2푠 > = {| + 2 > −| − 2 >}. Using these in the √2 √2

Hamiltonian in Equation 2-17 and the angular dependence, allowed transitions arise due to admixtures of the states [61, 69]. Species of this nature typically have very weakly allowed transitions in conventional EPR. However parallel mode EPR allows for transitions between ms = ±2 as shown in Figure 2-7 A, due to the admixture of states that arises when the system is rhombic. If the B1 field polarization is held along the z- axis, parallel to the magnetic field, resonances are seen to arise at a geff ≈ 8 for a Mn(III) species having magnetic parameters of D = -2.38 cm-1, E/D = 0.13, S = 2, g = 2.00, and

A = 140 MHz. When the B1 field polarization is held at any angle perpendicular to the z- axis in the x-y plane no resonances are observed as shown in Figure 2-7 B.

A B

Figure 2-7. Energy level diagram for Mn(III) with magnetic parameters of S = 2, g = 2.00, A = 140 MHz, D = -2.38 cm-1, and E/D = 0.13. A) Excitation along the z- axis (parallel mode), B) with excitation along the y-axis (perpendicular mode).

Vertical red lines mark allowed transitions that arise in parallel mode between the ms =

±2 sublevels. The multiple transitions arise due to the hyperfine coupling of the 55Mn nuclear spin of I = 5/2. The redox cycling of the metal centers in OxDC has been proposed to occur between Mn(II) and Mn(III). Parallel mode EPR is used to further study the presence of Mn(III) in OxDC and its role in catalysis.

CHAPTER 3 IDENTIFICATION OF SUPEROXIDE PRODUCTION UNDER TURNOVER CONDITIONS FOR Bacillus subtilis OXALATE DECARBOXYLASE

Introduction

The major reaction catalyzed by Bacillus subtilis oxalate decarboxylase is a redox-neutral disproportionation reaction, which does not consume dioxygen, but, requires it for turnover as a cofactor [26, 38-40, 70]. Current mechanistic proposals suggest that dioxygen is bound to the N-terminal Mn ion and acts as a transient electron sink to generate superoxide, and lead to the destabilization of the carbon-carbon bond in oxalate. After decarboxylation has taken place, superoxide acts as an electron source to reduce the resulting carbon dioxide anion radical [42]. This mechanism requires oxygen to cycle through its 1-electron reduced state as superoxide which would likely be protonated in the pH range (3.9 – 6.0) in which the enzyme is active [43]. Alternative proposals suggest that one of the Mn ions undergoes a redox cycle between its +2 and

+3 oxidation states, allowing the associated dioxygen to remain a hydroperoxyl radical throughout the catalytic cycle [29, 44, 45].

OxDC has also been shown to exhibit oxalate oxidation for 0.2% of all turnovers, resulting in the production of hydrogen peroxide [39]. It is then relevant to consider the structurally related enzyme oxalate oxidase (OxOx) [49], which may be evolutionarily related to OxDC [33, 48]. Sequence alignment studies of both Bacillus subtilis OxDC and Ceriporiopsis subvermispora OxOx identified a major structural difference between these two enzymes in the existence of an N-terminal active site flexible lid in OxDC, which is not present in OxOx. The lid consists of a Serine-Glutamate-Arginine-Serine-

Threonine sequence at positions 161-165, gating entry and exit of substrate and product molecules [47]. Site-directed mutagenesis in the lid region has been shown to

enhance the apparent oxidase activity of OxDC, while also decreasing its decarboxylase activity. For decarboxylation to occur, the presence of a protonating group is suggested to be essential, and has been tentatively assigned to be glutamate-

162 in OxDC [44, 47]. The absence of such a group in OxOx is suggested to lead to a peroxycarbonate intermediate, which decays under acidic conditions to form hydrogen peroxide and carbon dioxide [29, 30, 47, 71]. However, direct experimental evidence for this peroxycarbonate intermediate is still missing, and other mechanistic proposals exist in the literature [72].

Results from the spin-trapping experiments on the T165V OxDC mutant suggested that the destabilization of the closed conformation of this lid region leads to increased loss of the intermediate carbon dioxide radical anion into solution, where it can react under diffusion-controlled rates with dioxygen to produce superoxide and

•– eventually hydrogen peroxide [42]. Even in WT OxDC the CO2 radical intermediate can be trapped by appropriate spin traps, suggesting that the 0.2% oxidase activity is an apparent activity due to the enzyme’s ability to stabilize the intermediate species compared to an inter-protein mechanism [42, 47]. Spin trapping of OxOx under turnover

•– conditions also yields a CO2 radical adduct with both PBN and DMPO spin traps.

•– In any case, when OxDC loses its intermediate CO2 radical anion the initial oxidation step has already taken place, and the protein is now fixed with an electron on either one of its Mn ions or the dioxygen. This would lead to enzyme inhibition unless the enzyme can rid itself of the reducing equivalent, for example by losing the superoxide anion or its protonated counterpart, the hydroperoxyl radical. It would then be expected to find evidence for superoxide in the solution under turnover conditions.

•– Experimental detection of O2 at low pH is made difficult by the acid-catalyzed disproportionation of superoxide to hydrogen peroxide and oxygen, which occurs at very fast rates [73]. In order to further investigate the production of superoxide and any additional radical species, EPR spin trapping and mass spectrometry were employed to study the WT and T165V variant OxDC.

EPR Spin-Trapping

Free radicals, due to their innate reactivity result in extremely short half-lives, making their identification difficult. One method to facilitate the identification of these species of interest is spin trapping. In order to extend the half-life of the free radical, a nitrone based compound is introduced into the reaction to react with the radical and form a persistent nitroxide radical.

A common spin trap used for identification of superoxide is 5,5-dimethyl-pyrroline

N-oxide (DMPO) (Scheme 3-1 A) [74]. However, DMPO tends to react rather slowly with superoxide, making it necessary to utilize high spin trap concentrations to outcompete the disproportionation reaction [74]. Moreover, DMPO decays quickly with a half-life of less than one minute at pH 7 [75], and its superoxide adduct can decay into a hydroxyl adduct, necessitating cumbersome control experiments and causing possible misidentification of radical species [76, 77]. To avoid these problems the spin trap 5-tert- butoxycarbonyl 5-methyl-1-pyrroline N-oxide (BMPO) (Scheme 3-1 B) was used in the present study. It features an extended half-life of the persistent superoxide adduct of t1/2

= 23 minutes, facilitated by its bulky side groups protecting the nitroxide radical as shown in Scheme 3-1 B [64, 78]. Radicals can be identified based on their observed splitting patterns that arise from the nitroxide’s hyperfine splitting, which varies based on the radical and nitrone pair being investigated. In the case of OxDC, there will be a

•– • mixture of O2 and HO2 with the equilibrium favoring the hydroperoxyl radical at the reaction pH 4.0 [63, 79].

Scheme 3-1. A) Reaction that occurs between 5,5-dimethyl-pyrroline N-oxide (DMPO) spin-trap and superoxide. Superoxide binds adjacent to the nitrogen producing a persistent radical with a half-life of 45 seconds. B) Reaction of 5- tert-butoxycarbonyl 5-methyl-1-pyrroline N-oxide (BMPO) spin-trap with superoxide producing a persistent radical with a half-life of 23 minutes facilitated by its bulky tert-butoxycarbonyl side group.

It is critical to confirm the identification of the trapped radical species more definitively than identification based on the hyperfine splitting for the observed system.

Previous studies have utilized spin trapping in combination with mass spectrometry to confirm the identity of spin-trapped radical adducts with both DMPO and BMPO spin traps [80-85].

Experimental

Materials

The following chemicals were purchased from Fisher Scientific (ACS Grade): potassium acetate, glycerol, citric acid, ethylenediaminetetraacetic acid (EDTA), and potassium phosphate. Potassium oxalate monohydrate and diethethylenetriamine-

pentaacetic acid (DTPA) were purchased from Sigma-Aldrich. 5-tert-butoxycarbonyl 5- methyl-1-pyrroline N-oxide (BMPO) was obtained from Applied Bioanalytical Labs

(Bradenton, FL). All solutions were prepared in distilled water generated by a Thermo

Scientific Barnstead Nanopure Model 7134 for EPR ,and in HPLC grade water obtained from Fisher Scientific for mass spectrometry. Stock solutions of 1 M acetate or citrate buffer were used to buffer the solutions at pH 4.0 at a final buffer concentration of 50 mM. Potassium oxalate stock solutions were prepared at 0.5 M and the pH was adjusted to match that of the pH of the buffering solution. A 2 M BMPO stock solution was also prepared, along with a 4 mM stock of DTPA. Enzyme expression and purification procedures can be found in Appendix A.

Electron Paramagnetic Resonance

Experiments were performed on a Bruker ELEXSYS E580 CW/Pulsed or a

Bruker ELEXSYS-II E500 CW X-band spectrometer equipped with a super high-Q cavity (ER 4123SHQE). Reactions were carried out in mini-Eppendorf tubes with the following final concentrations: 50 mM potassium acetate or citric acid buffer adjusted to pH 4.0, 50 mM potassium oxalate pH adjusted to match the buffer pH, 20 M DTPA,

100 mM BMPO. Enzyme OxDC was added last to the reaction mixture, resulting in a final concentration of approximately 5 mg/mL. The total reaction volume was 100 L and also contained 35 L distilled water generated by a Thermo Scientific Barnstead

Nanopure Model 7134 and 20 L glycerol. Glycerol was added to prevent precipitation of the enzyme upon lowering the pH. The solution was mixed using a vortexer (Fisher

Scientific Deluxe Vortex Mixer) and then immediately transferred to a quartz capillary

(13 mm IDOD), which was sealed at the bottom with Cha-seal (Kimble Chase Life

Sciences, Vineland NJ). All spectra were collected at room temperature, with the following instrumental parameters: 100 kHz modulation frequency, 1 G modulation amplitude, 20.48 ms time constant, 40.96 ms conversion time, 10 dB microwave attenuation, 60 dB receiver gain, and 1024 data points per spectrum. Simulation of the

EPR spectra was carried out using the EasySpin toolbox for MATLAB [67]. Individual spectra for the carbon dioxide radical anion adduct and the hydroperoxyl adduct were fitted first to the spectra where they dominate. This was followed by a weighted fit of both components which provided the concentrations after comparison to a calibration curve using 4-hydroxy TEMPO as the calibration standard. All experiments were performed in triplicate for error analysis.

Mass Spectrometry

All experiments were performed on an Advion Expression compact mass spectrometer (Advion, Ithaca, NY). The instrument is equipped with an electrospray ionization (ESI) source, single quadrupole mass analyzer, and an electron multiplier detector. Solutions were prepared similar to those for EPR except no glycerol was used, as its low vapor pressure dramatically reduces electrospray nebulization efficiency, thus decreasing ionization efficiency. Before injection into the mass spectrometer, the enzyme was precipitated from solution by mixing with acetonitrile, followed by centrifugation to pellet unwanted protein [68]. The supernatant was diluted in acetonitrile/water/formic acid (50:50:1) to a final volume of 500 μL. The solution was infused into the ESI source at a flow rate of 20 μL min-1, nebulized with nitrogen gas, and introduced into the mass spectrometer using positive mode electrospray ionization.

Ion source conditions were optimized to promote high signal intensity of the protonated

radical-bound spin trap. Due to the inherently low concentrations of analytes in this study, mass spectra were recorded at 1 Hz and were allowed to average for 5 minutes prior to analysis by Mass Express™ (Advion, Ithaca, NY). The samples used for the MS experiments were tested for the presence of the radical adduct by EPR before injection.

Hydrogen Peroxide Fluorescence Assay

The Europium tetracycline (EuTc) Hydrogen Peroxide assay kit was obtained from Active Motif (Carlsbad, CA) [86]. All experiments were carried out with a Horiba

Jobin Yvon FluoroMax-3. Reaction mixtures were prepared, run and quenched in a 1.5 mL eppendorf and then transferred to a 5 mm square cuvette. The sample temperature during the reaction and measurement was 25o C. The assay was carried out as specified by Active Motif. The EuTc powder sample was used as provided with no further purification and dissolved in 100 mL of DI-H2O. A standard curve was prepared using 30% hydrogen peroxide adjusted to final concentrations of 200 μM, 160 μM, 100

μM, 40 μM, 20 μM, and 10 μM. Enzyme reaction mixtures were prepared in 100 μL total volume as follows: 25 μL 0.2 M acetate buffer pH 4.0, 10 μL 0.5 M potassium oxalate pH 4.0, 44 μL DI-H2O, and 1 μL enzyme. The reactions were allowed to proceed for 3 min and were then quenched with 900 μL 0.1 M HEPES buffer pH 7.0. The reaction product was mixed in a 1:1 volume ratio with the prepared EuTc solution and allowed to incubate for 10 minutes prior to the fluorescence measurement. Samples were excited at a λexcitation of 400 nm, and measured at λemission of 617 nm with a bandwidth of 10 nm.

Results

Spin-trapping of Radicals Produced from Wild-Type OxDC

Spin trapping studies were carried out on wild type B. subtilis OxDC in an effort to identify free radicals produced during the enzymatic mechanism. It has been

previously shown that a carbon dioxide anion radical is produced during turnover [42]. It has been proposed that superoxide is generated; however, no previous evidence has been provided for this species. Identifying these short-lived radicals directly by EPR is difficult, but spin-traps greatly extend the lifetime of these radicals, thereby facilitating their identification. Scheme 3-1 B depicts the reaction that occurs when hydroperoxyl reacts with BMPO, forming a persistent nitroxide radical adduct. The reaction is carried out at an optimum pH of 4.0 for OxDC, below the pKa of superoxide (4.88), which shifts the equilibrium of superoxide to favor its protonated form, the hydroperoxyl radical.

Figure 3-1. X-band EPR spectra of the BMPO-radical adducts produced from the reactions between oxalate and WT-OxDC in 50 mM citrate buffer pH 4.0 (black). The orange and magenta traces represent simulations of the superoxide and carbon dioxide anion radical adduct, respectively. Blue: sum of the simulations of the two spin adducts. Reprinted from Free Radical Biology and Medicine, Twahir,U.T., Stedwell,C.N., Lee,C.T., Richards,N.G.J., Polfer,N.C., Angerhofer,A., 80, 59-66, 2015, with permission from Elsevier.

The resulting X-band EPR spectrum is shown in the black trace in Figure 3-1.

Control experiments were also carried out where individual reaction components were omitted to ensure that signal observed indeed arose from the reaction catalyzed by

OxDC. It is apparent that contributions are present from more than one species. BMPO radical adducts for superoxide usually show two slightly different spectra due to the presence of two diastereomers, arising from conformational flexibility of the bound radical [64]. An additional contribution from a carbon based radical adduct was also observed and was necessary to produce a satisfactory simulation (magenta trace in

Figure 3-1) of the spectrum. Table 3-1 shows the magnetic parameters used in the simulation together with the corresponding spectral weights.

Table 3-1. Spectral parameters of BMPO-radical adducts for WT-OxDC, including hyperfine coupling, g-value, and individual spectral weights based on the simulated spectra. H N Spin-Trapped g-value a (mT) a (mT) Spectral Weight Adduct (Relative Percentage)

[BMPO-OOH] 2.005 1.21 1.35 61 (Conformer I)

[BMPO-OOH] 2.005 0.96 1.32 29 (Conformer II)

-  [BMPO-CO2 ] 2.005 1.74 1.48 10

•– The CO2 adduct was expected and had been previously reported as being derived from oxalate [42]. Unexpectedly, the major contribution to the spectrum was the hydroperoxyl adduct with a combined spectral weight of 90% for the two diastereomers, while the carbon dioxide anion radical contributed only 10% according to the simulated spectra. The hydroperoxyl adduct had evaded detection in our earlier work, because the PBN spin-trap is not a very efficient trap for superoxide [87]. Simulations were

initially performed with literature values for the expected radical adducts [64], followed by an iterative fitting approach producing the converged values reported here. The hyperfine coupling constants for the hydroperoxyl (HOO•) adduct matched well with those in the literature, suggesting that HOO• is produced during the enzymatic catalysis as a transient intermediate [64, 78].

Figure 3-2. Time-Resolved X-band EPR spectra of the BMPO-radical adducts produced from the reactions between oxalate and WT-OxDC in 50 mM citrate buffer pH 4.0. EPR Parameters: 10 dB attenuation, 100 kHZ modulation frequency, 1 G modulation amplitude, 20.48 ms time constant, 40.96 conversion time, 60 dB receiver gain and 1024 points. Reprinted from Free Radical Biology and Medicine, Twahir,U.T., Stedwell,C.N., Lee,C.T., Richards,N.G.J., Polfer,N.C., Angerhofer,A., 80, 59-66, 2015, with permission from Elsevier.

In order to confirm that the radical adduct was indeed derived from superoxide, mass spectrometry experiments were carried out on the reaction mixture.

It is critical to carry out secondary identification of spin trapped adducts, as spin trapping can be misleading due to the innate reactivity of the species under investigation.

Time resolved spin trapping was conducted on wild type to approximate the lifetime of the trapped species (Figure 3-2), in order to confirm the maximum time frame for mass spectrometry experiments. Figure 3-2 shows that the spin-trapped adducts can be seen beyond 30 minutes, well in excess of the time frame for a typical mass spectral analysis.

Figure 3-3. ESI-Q-MS analysis of BMPO-superoxide reaction mixture. Mass spectral features include: [BMPO+H]+ (m/z 200), [BMPO+Na]+ (m/z 222), + + • + [BMPO+HO2+H] (m/z 233), [BMPO+K] (m/z 238), [BMPO+HCO2 +H] (m/z 245). The inset gives an expanded view of the m/z range for the radical- bound spin trap. Spectra were normalized to the intensity of the protonated spin trap ([BMPO+H]+). The potassiated spin trap (m/z 238) is approximately 625 times more intense and off the scale. Reprinted from Free Radical Biology and Medicine, Twahir,U.T., Stedwell,C.N., Lee,C.T., Richards,N.G.J., Polfer,N.C., Angerhofer,A., 80, 59-66, 2015, with permission from Elsevier.

One key aspect to also keep in mind is that the quenching of the radical EPR spectrum due to either a reduction or oxidation does not necessarily indicate loss of the spin adduct. The hydroxylamine version of the spin trapped adduct, where the oxygen of the

nitroso moiety is protonated, is still a suitable candidate for mass spectral analysis. The

MS analysis (Figure 3-3) of the enzyme reaction mixture in the presence of BMPO provides confirmation that superoxide is produced during turnover.

Figure 3-4. ESI-Q-MS analysis of control mixture, with enzyme omitted. Mass spectral features include: [BMPO+H]+ (m/z 200), [BMPO+K]+ (m/z 238). No features are seen for radical bound spin-trapped masses as previously identified. Reprinted from Free Radical Biology and Medicine, Twahir,U.T., Stedwell,C.N., Lee,C.T., Richards,N.G.J., Polfer,N.C., Angerhofer,A., 80, 59- 66, 2015, with permission from Elsevier.

In Figure 3-3, peaks at m/z 200, 222, 233, 238, and 245 represent the protonated spin trap, sodiated spin trap, superoxide bound radical adduct, and potassiated spin trap, and a protonated form of the carbon dioxide anion radical adduct, respectively.

From the zoomed portion of the figure, indicated by the red box, it is evident that there is a peak at the expected mass-to-charge ratio for the BMPO-superoxide radical adduct

(m/z 233). It is also clear from the mass spectrum that BMPO favors metallation, given

that both sodium and potassium bound spin traps are present. The peak at m/z 245 also provides evidence for the second trapped radical, identified to be the carbon dioxide anion radical.

Control experiments were carried where enzyme was omitted from the reaction mixture. The masses representative of the spin trapped adducts are no longer seen as shown in Figure 3-3.The metal bound forms of the spin trap (m/z 222 and m/z 238) are also drastically reduced due to the lack of salts that originate from the enzyme preparation. The mass spectral analysis in combination with the EPR experiments provides clear evidence that both the carbon dioxide anion radical and superoxide are produced and released from the protein during the enzymatic degradation of oxalate by

OxDC.

Spin trapping of Radicals Produced with T165V

It has been previously shown that the lid mutants T165V and SENS161-

•– 164DASN exhibit increased CO2 yields by spin trapping experiments using α-phenyl N- tertiary-butyl nitrone (PBN) and α-(4-Pyridyl N-oxide)-N-tert-butylnitrone (POBN) [42,

•– 47]. If the observed superoxide arises only from the reaction of escaped CO2 radical anion with dissolved dioxygen, the trapping rates of superoxide should also increase in these mutants. Spin trapping was performed on the T165V variant of OxDC in the presence of BMPO. The experimental results, together with theoretical simulations, are shown in Figure 3-5, and the magnetic parameters used, together with the corresponding spectral weights, are represented in Table 3-2.

The spectrum for spin trapping carried out with T165V OxDC mutant is dominated by the carbon dioxide anion radical. The carbon-based radical adduct dominating the spectra was confirmed to be the carbon dioxide adduct by comparison

•– with a synthetically produced CO2 radical in the presence of BMPO. The synthetic

•– CO2 was generated by the Fenton reaction in the presence of formic acid, as previously described [42].

Table 3-2. Spectral parameters of BMPO-radical adducts for T165V OxDC mutant, including hyperfine coupling, g-value, and individual spectral weights based off simulated spectra. H N Spin-Trapped g-value a (mT) a (mT) Spectral Weight Adduct (Relative Percentage)

[BMPO-OOH] 2.005 1.21 1.36 24 (Conformer I)

[BMPO-OOH] 2.005 0.95 1.32 10 (Conformer II)

-  [BMPO-CO2 ] 2.005 1.75 1.48 66

Figure 3-5. X-band EPR spectra of the BMPO-radical adducts produced from the reactions between oxalate and OxDC mutant T165V in 50 mM citrate buffer pH 4.0 (black). The orange and magenta spectra represent simulations of the superoxide and carbon dioxide anion radical adduct, respectively. Blue: sum of the simulations of the two spin adducts. Reprinted from Free Radical Biology and Medicine, Twahir,U.T., Stedwell,C.N., Lee,C.T., Richards,N.G.J., Polfer,N.C., Angerhofer,A., 80, 59-66, 2015, with permission from Elsevier.

In order to confirm that the radical adduct formed was indeed derived from superoxide, mass spectrometric and control experiments like those for the wild-type enzyme were carried out on the reaction mixture. Similar to the experiments on the wild- type enzyme, time-resolved spin trapping was conducted on T165V to identify the lifetime of the trapped species (Figure 3-6), in order to confirm the maximum time frame for mass spectrometry experiments as done for the WT enzyme. It is again seen that the lifetimes for both species exceed the time frame for mass spectral analysis.

Figure 3-6. Time-Resolved X-band EPR spectra of the BMPO-radical adducts produced from the reactions between oxalate and OxDC mutant T165V in 50 mM citrate buffer pH 4.0. EPR Parameters: 10 dB attenuation, 100 kHz modulation frequency, 1 G modulation amplitude, 20.48 ms time constant, 40.96 conversion time, 60 dB receiver gain and 1024 points. Reprinted from Free Radical Biology and Medicine, Twahir,U.T., Stedwell,C.N., Lee,C.T., Richards,N.G.J., Polfer,N.C., Angerhofer,A., 80, 59-66, 2015, with permission from Elsevier.

As expected, the mass spectrum of the T165V variant reaction mixture is similar to that of the wild-type protein in the presence of BMPO providing confirmation that superoxide is produced during turnover. In Figure 3-7, the assignment of the peaks at

m/z 200, 222, 233, 238, and 245 represent the protonated spin trap, sodiated spin trap, superoxide bound radical adduct, potassiated spin trap, and a protonated form of the carbon dioxide anion radical adduct, respectively. From the zoomed portion of the figure, indicated by the red box, it is evident that there is a peak at the expected mass- to-charge ratio for the BMPO-superoxide radical adduct (m/z 233) and the carbon dioxide radical anion (m/z 245).

Figure 3-7. ESI-Q-MS analysis of T165V mutant BMPO-superoxide reaction mixture. Mass spectral features include: [BMPO+H]+ (m/z 200), [BMPO+Na]+ (m/z + + • + 222), [BMPO+HO2+H] (m/z 233), [BMPO+K] (m/z 238), [BMPO+HCO2 +H] (m/z 245). Inset gives an expanded view of the m/z range circa the radical- bound spin trap. Spectra were normalized to the intensity of the protonated spin trap ([BMPO+H]+). Reprinted from Free Radical Biology and Medicine, Twahir,U.T., Stedwell,C.N., Lee,C.T., Richards,N.G.J., Polfer,N.C., Angerhofer,A., 80, 59-66, 2015, with permission from Elsevier.

By qualitative comparison of the experimental data for the WT and T165V variant of

OxDC, it is not possible to discern whether the source of observed superoxide is solely due to release from the protein during turnover or is a side product due to the reaction

of released carbon dioxide anion radical with dissolved oxygen. In order to address this question quantitation of the radical production as well as hydrogen peroxide production is essential.

Quantitative Analysis of Radical Production

Hydrogen peroxide quantitation was carried out in order to assess the relative amounts produced by both the WT and T165V mutant enzyme.

Figure 3-8. EuTc hydrogen peroxide fluorescence assay. Standard concentrations of H2O2 were prepared () and mixed 1:1 with EuTc working solution. Enzyme reaction mixtures were prepared in 100 μL total volume: 25 μL 0.2 M acetate buffer pH 4.0, 10 μL 0.5 M potassium oxalate pH 4.0, 44 μL DI-H2O, and 1 μL enzyme. The reactions were allowed to run for 3 min and were then quenched with 0.1 M HEPES buffer pH 7.0. It is seen that the wild-type form of the enzyme produces slightly higher concentrations of H2O2 compared to the T165V mutant, which was expected since it had been previously shown to have reduced activity. Reprinted from Free Radical Biology and Medicine, Twahir,U.T., Stedwell,C.N., Lee,C.T., Richards,N.G.J., Polfer,N.C., Angerhofer,A., 80, 59-66, 2015, with permission from Elsevier.

The fate of superoxide in solution is spontaneous disproportionation to hydrogen peroxide and oxygen, such quantitation of hydrogen peroxide production under similar

conditions would provide another methodology for comparison of superoxide production. Europium tetracycline binds to H2O2 in a 1:1 fashion, leading to an enhanced fluorescent yield [86]. Experiments were conducted with a λexcitation of 400 nm, and the fluorescent yield was monitored at a λemission of 617 nm. These experiments showed that under the reaction conditions WT OxDC produced 74.6 ± 14 μM H2O2, while T165V produced 55.9 ± 5.3 μM H2O2 (Figure 3-8).

Figure 3-9. From left to right for WT (yellow) and T165V (gray): concentrations of the •– •– BMPO radical adduct for the CO2 and the O2 radical in reactions carried out as described in the text, concentrations of H2O2 as determined by the EuTc assay, and published values for the decarboxylase and oxidase activities in terms of observed kcat/KM Reprinted from Free Radical Biology and Medicine, Twahir,U.T., Stedwell,C.N., Lee,C.T., Richards,N.G.J., Polfer,N.C., Angerhofer,A., 80, 59-66, 2015, with permission from Elsevier.

These results can be considered as somewhat expected, because T165V has been previously noted as having decreased kinetics. Results of quantitative analysis of

hydroperoxyl radical, carbon dioxide radical, and hydrogen peroxide are shown in

Figure 3-9 for comparison, and are rationalized as described below.

Figure 3-9 shows from left to right the concentration of carbon dioxide anion radical, superoxide, and hydrogen peroxide produced during catalysis, followed by the decarboxylase and oxidase activities for WT (yellow) and T165V (grey) OxDC

•– respectively. The rate of CO2 radical release from the protein is approximately ten times higher in T165V than in WT. Comparing the first and fourth column pair of Figure

3-9, an inverse relationship is seen, with the T165V mutant exhibiting a ten times lower decarboxylase activity, indicating that loss of this intermediate is associated with the decrease in decarboxylase activity. A similar inverse relationship is seen between the second and sixth column pair in Figure 3-9 for the hydroperoxyl adduct yields and the

OxOx activities of the two strains, albeit the differences between the two strains is smaller and the error bars are generally larger because oxidase activity is still about two orders of magnitude slower than decarboxylase activity. The fifth and sixth column pair represents the same data set, where the 6th column pair has been multiplied by a factor of 100 as a visual aid. When using the EuTc fluorescence assay Figure 3-8 and third column pair in Figure 3-9, a slightly smaller yield of H2O2 production is observed in

T165V (56 ± 5 μM compared to 75 ±14 μM in WT). This is still consistent with the spin trapping experiments, because the EuTc assay measures total yield of H2O2 while the spin trapping experiment only determines the trapping yield of escaped superoxide that was generated within the protein.

Discussion

The combination of EPR spin trapping and mass spectrometry experiments has demonstrated the presence of superoxide/hydroperoxyl radical production during

catalysis by OxDC. This is the first experimental report of a superoxide/hydroperoxyl- derived radical adduct observed in either OxDC or OxOx. These experiments also confirm prior observation of the carbon dioxide radical anion intermediate during turnover. These experiments were performed at pH 4, where superoxide dismutates with essentially diffusion-controlled kinetics [79]. The trapping rate of superoxide with

BMPO is only 0.24 M-1s-1 at pH 7.4 [88]. However, trapping yields with many cyclic nitrones increase dramatically in the presence of hydroperoxyl radical at low pH [89], possibly also with superoxide when the spin trap can be protonated [90], and greatly facilitates the identification of the radical species. Previous attempts had been unsuccessful, possibly due to the much lower protein concentrations attainable, as well as the diffusion controlled disproportionation of superoxide at lower pH.

In order to better compare the relative yields of the two radical adducts, quantitative EPR was performed for the WT and the T165V mutant under the same experimental conditions. A principal component analysis was performed utilizing the simulated EPR spectra for the WT and the T165V mutant. The contributing components to the spectra (superoxide and carbon dioxide anion radical) were then weighted relative to their individual spectral contributions. Double integration of the simulated components provided concentrations for radical yields relative to a calibration curve using 4-hydroxy TEMPO as the calibration standard. All experiments were performed in triplicate for error analysis. Quantitative experiments yielded concentrations of 1.4 ± 0.7

•– • μM, and 9.9 ± 1.8 μM for CO2 and HO2 , respectively, for the wild-type enzyme; 11 ±

•– • 1.0 μM, and 3.4 ± 1.3 μM for CO2 and HO2 , respectively, for the T165V mutant. It should be noted that quantitative EPR typically carries uncertainties of up to 5-10%

resulting in the large error bars [42]. The yield of superoxide adduct is approximately

•– 60% lower in the mutant, while the yield of the CO2 radical adduct is suppressed in WT by about a factor of ten compared to T165V. If the majority of the superoxide radical

•– were generated in the solution by loss of the CO2 radical from the protein followed by subsequent reaction with dioxygen, its yield should follow the same observed relative

•– ratio exhibited by the CO2 radical in the two strains. Moreover, if the production of the

• HO2 radical were to strictly follow the decarboxylase activity of the enzyme, one would expect a ten times lower yield in T165V [63]. The oxidase activity of T165V has been previously found to be a factor of three higher than that of the WT enzyme [63]. It is

• more relevant to relate the production of HO2 with the oxidase rather than the decarboxylase activity of the enzyme. Earlier reports on oxalate oxidase did not show evidence of the hydroperoxyl adduct upon trapping with DMPO [42] suggesting that the oxidase activity in OxOx relies on a mechanism different from that of the residual oxidase activity in OxDC WT or mutants.

The kcat/KM for decarboxylase activity in the T165V mutant is approximately ten times lower than in WT as shown in column four of Figure 3-9 [64]. The decrease in decarboxylase activity is attributed to the impeded motion of the lid gating the N- terminal active site, leading to stabilization of the open conformation and allowing for an

•– •– increased release of the CO2 radical into solution [64]. The intermediate CO2 radical can thus diffuse away from the active site at a much higher rate in the mutant and is therefore not available for conversion to formate. The loss of this intermediate is clearly linked to the decreased decarboxylase activity.

In order to evaluate if there is an increase in superoxide generation upon excess

•– release of the CO2 , the hydrogen peroxide production was investigated via an assay utilizing europium tetracycline. Measurement of the hydrogen peroxide production under turnover conditions for both the WT and T165V mutant produced a similar trend, showing a higher production of hydrogen peroxide for the wild-type protein (Figure 3-8 and Figure 3-9), with H2O2 production in T165V of 56 ± 5 μM compared to 75 ±14 μM in

WT. Comparison of column two and three of Figure 3-9 shows that the hydrogen peroxide concentration is much more similar than the superoxide concentrations for the

WT and T165V mutant. This provides evidence that superoxide/hydroperoxyl is produced both intra- and extra-protein, and the release of the carbon dioxide anion radical from OxDC may explain the “oxidase” activity exhibited by the assay. The intra- protein superoxide production would occur via an oxidase mechanism, for example the one suggested for oxalate oxidase by Whittaker [43], with subsequent release of hydroperoxyl radicals into solution. The extra-protein superoxide production would occur

•– through the reaction of the CO2 radical with dissolved dioxygen [64]. In the absence of spin trap this ‘extra-protein’ superoxide will dismutate to oxygen and H2O2 which is seen in the EuTc assay as well as in previously conducted oxygen consumption experiments

[91]. It disguises the true OxOx activity in these assays by allowing for additional production of hydrogen peroxide and additional consumption of dioxygen.

These results indicate that the oxidase activity displayed by OxDC may solely be due to its inability to completely control the radical chemistry occurring at the active site, leading to the release of the carbon dioxide anion radical and the observed secondary mechanistic pathway. This may help the protein to remove the excess electrons from

•– oxalate when it acts as an oxidase. However, the uncontrollable release of the CO2 radical masks the ability to directly measure any true oxidase activity exhibited by the enzyme. At issue is whether superoxide is released into the solution under turnover conditions from the active site, from a site different from the active site, or is simply generated outside the protein. BMPO is too bulky to fit into the active site, if this is the site of generation, to react with a closely held superoxide radical. All tentative mechanisms for OxDC discussed in the literature to date suggest oxalate activation by one-electron transfer as the first step with dioxygen being the driving oxidative force leading to superoxide or hydroperoxyl bound to the N-terminal manganese [44]. If this takes place in the active site, a peroxycarbonate species would logically form, leading to oxidase activity which could explain the 0.2% oxidase activity in OxDC [29]. Opaleye et al. suggested a similar mechanism for oxalate oxidase, although the peroxo compound suggested there still includes the oxalate [92]. More recently, Whittaker et al. favored an oxidase mechanism for OxOx that bypasses the peroxycarbonate intermediate and generates superoxide or hydroperoxyl through inner or outer sphere electron transfer based on competition experiments with superoxide dismutase [72]. This mechanism allows for the loss of hydroperoxyl from the protein and is consistent with our findings of a hydroperoxyl adduct to BMPO. Release of the carbon dioxide radical anion into solution through a loss of control of this intermediate by the protein will automatically lead to superoxide production. This could also explain the 0.2% oxidase activity observed in WT OxDC, not as a separate enzymatic pathway, but rather as the result of loss of control over the decarboxylase pathway [42]. This raises the question of the fate

•– of the enzyme when it loses the CO2 intermediate, since at that point it contains an

extra electron, most likely on a superoxide radical bound to one of the Mn ions. The simplest explanation would be an additional 1-electron transfer leading to Mn(III) and hydrogen peroxide in or somewhere near the active site with the needed extra proton being donated by the solution [72].

More intriguing is the ability to observe both superoxide/hydroperoxyl and the carbon dioxide anion radical at the same time, and in unequal quantities. The unequal production of radicals observed in the wild type and T165V mutant enzymes suggests that the two radicals are produced at different locations in the protein and via separate mechanisms as outlined above. The previously proposed mechanisms for OxOx are based on both the mono and bi-cupin forms of the enzyme. In the case of the mono- cupin enzyme, a single manganese site exists for each monomer, and does not provide an obvious secondary site for oxygen binding, as in OxDC. If oxygen and substrate were binding at the same site in OxOx, this would lead to the proposed peroxycarbonate intermediate and the inability to observe the superoxide intermediate.

Due to the high structural similarity between the N- and C-terminal manganese sites, it was initially proposed that catalysis occurred at the C-terminal site. It was later shown that the N-terminal site is where catalysis occurs through a series of site-directed mutants, EPR, and X-ray crystallographic studies. However, the C-terminal site was then assumed to serve a structural role, as variants that led to the removal of the C- terminal manganese produced inactive protein and the disruption of the protein quaternary structure. Binding of oxygen at the C-terminal site would provide the necessary reductive and oxidative equivalents necessary for catalysis, and would also separate the sites of superoxide/hydroperoxyl and carbon dioxide anion radical

production. The necessary electrons to facilitate catalysis would then be shuttled between the N- and C-terminal manganese via a tryptophan dimer pair that sits approximately equidistant between the two sites.

CHAPTER 4 REDOX CYCLING OF THE METAL CENTERS OF Bacillus subtilis OXALATE DECARBOXYLASE: IDENTIFICATION OF Mn(III)

Introduction

Oxalate decarboxylase has been shown to require manganese for catalysis [50].

The enzyme displays a pH-dependent catalytic behavior between pH 4 and 6, exhibiting its highest activity at pH 4, thought to be linked to its substrate specificity [37]. The enzyme shows a strong dependence on di-oxygen as a cofactor during catalysis, but O2 is consumed only when the enzyme acts as an oxidase [93]. It has been proposed that oxygen drives the initial oxidation step in the catalytic cycle, generating a Mn(II) or

Mn(III) bound superoxide [29, 44, 45].

Superoxide production was studied utilizing both spin trapping EPR and mass spectrometry as described in Chapter 3 [41]. The enzymatic mechanism of OxDC has been proposed to function in one of two manners. Initially mono-protonated oxalate, and dioxygen bind mono-dentate to the manganese center at the N-terminus, which has been assumed to be the active site. Co-crystallization with formate, which is one product of catalysis, was shown to bind at the N-terminus. Also site-directed mutants at this site leads to catalytic consequences [28, 29].

- Binding of HO2CCO2 and O2 is immediately followed by a proton coupled electron transfer (PCET) in which glutamate-162, tentatively assigned as the necessary transient acid/base, can abstract a proton from oxalate, while oxygen acts as an electron sink forming a metal bound superoxide, leading to the destabilization of the carbon-carbon bond of oxalate as shown in Scheme 1-2. After the release of carbon dioxide, a carbon dioxide anion radical is left bound to the manganese. Superoxide can then participate in the reduction of this radical, followed by the abstraction of a proton

from glutamate-162 to generate formate. Throughout this mechanistic scheme, the oxidation state of the manganese remains unchanged [21, 26, 38, 40, 93].

The second proposed mechanism by which catalysis occurs is via an oxidation of the metal center. Mono-protonated oxalate is still presumed to bind mono-dentate to the

N-terminus manganese center, followed by the binding of oxygen and leading to the generation of a superoxide bound Mn(III). Glutamate-162 can then abstract a proton from oxalate leading to the aforementioned PCET by forming a Mn(II) bound oxalate radical intermediate, followed by heterolytic cleavage of the carbon-carbon bond of oxalate and the release of carbon dioxide. The remaining carbon dioxide radical anion is then reduced by the metal and a proton is abstracted from glumate-162, forming a formate-bound Mn(III) [29, 44, 45]. The second mechanism requires the redox cycling between the +2 and +3 oxidation states of manganese. The presence of Mn(III) is precedented in related enzymes, such as oxalate oxidase, superoxide dismutase, extradiol catechol dioxygenase, and manganese peroxidase [72, 94-99], suggesting that

OxDC may follow a similar trend.

Oxalate oxidase is a closely related enzyme to OxDC, having many of the same requirements for catalysis, such as metal dependence, oxygen dependence and oxalate

[38]. It has been previously shown that the active form of OxOx is Mn(III) [72]. It has been proposed that OxOx follows a mechanism very similar to that of OxDC. However, it does not contain an analogous proton acceptor/donor such as glutamate-162. This then leads to the formation of a peroxycarbonate intermediate that rapidly decomposes into carbon dioxide and hydrogen peroxide [38, 71, 92, 100, 101]. An example of this mechanism is shown in Scheme 4-1.

Scheme 4-1. Proposed reaction mechanism of Oxalate Oxidase. Mono-protonated oxalate binds to Mn, followed by the binding of dioxygen and generation of superoxide and Mn(III). After loss of carbon dioxide, a bound formyl radical intermediate is formed, followed by reaction with bound superoxide generating a peroxycarbonate intermediate. The peroxycarbonate intermediate then decomposes to another equivalent of carbon dioxide and hydrogen peroxide [102].

A secondary proposed mechanism does not require dioxygen bound at the active site metal, where transient oxidation of the metal occurs leading to the production of a hydroperoxyl radical and finally hydrogen peroxide [100, 103]. The production of superoxide/hydroperoxyl radicals has been shown to occur under turnover conditions for OxDC suggesting that, at the minimum, a transient Mn(III) species is generated [41].

Herein, the first EPR spectroscopic evidence is provided for the presence of Mn(III) in wild-type oxalate decarboxylase. The major difficulty in studying integer spin species, such as high-spin Mn(III) with an S = 2 total spin (non-Kramer’s system) is the extremely large zero-field splittings leading to difficulties to observe transitions at X-band frequencies. However, applying the microwave field oriented parallel to the magnetic field allows for the excitation and observation of some transitions of the integer spin species. Parallel polarization exhibits different selection rules, allowing for transitions

between successively close sublevels, such as Ms = ± 2 [61, 69, 104]. Herein, parallel mode X-band EPR and multi-frequency EPR is used to characterize the redox cycling of the metal centers in WT-OxDC, as well at the generation of the stable Mn(III) species.

Experimental

Materials

The following chemicals were purchased from Fisher Scientific (Pittsburg PA,

ACS Grade) and used as received without further purification: phosphoric acid, glacial acetic acid, sodium chloride, sodium phosphate, sodium acetate, sodium hydroxide, tris(hydroxymethyl)aminomethane (Tris), bis-tris, piperazine, and succinate. Zinc sulfate was purchased from Sigma Aldrich (St. Louis, MO). Uncharged Profinity IMAC resin was purchased from Bio-Rad Laboratories (Hercules, CA). All solutions were prepared utilizing 18MΩ∙cm de-ionized water generated by a Thermo Scientific Barnstead

Nanopure model 7134.

Electron Paramagnetic Resonance Studies

Experiments were performed on a Bruker ELEXSYS E580 CW/ Pulsed or a

Bruker ELEXSYS-II E500 CW X-band spectrometer equipped Dual Mode Cavity (Bruker

ER 4116DM). Experimental conditions were typically: 100 kHz modulation frequency, 10

G modulation amplitude, 0.63 mW microwave power, and 5 K temperature. High field/frequency measurements were collected on a variable frequency/field broadband transmission spectrometer at 406.4 GHz in a field ranging from 13.9 T to 14.9 T, 50 kHz modulation frequency, 1 or 25 G modulation amplitude, 0.2 or 2 mT/s sweep rate, and in a temperature range of 3 to 20 K. Simulations of experimental spectra were carried out using the EasySpin toolbox for MATLAB™ [105].

Enzyme Oxidation Experiments

Enzyme expression and purification procedures can be found in Appendix A.

Previous attempts to observe Mn(III) in OxDC were unsuccessful (unpublished data) following a protocol used from generation of Mn(III) in oxalate oxidase [72]. In this report they utilized sodium meta-periodate a two-electron oxidizing agent to oxidize metal centers from Mn(II) to Mn(IV) and then ascorbic acid was used to reduce Mn(IV) to

Mn(III). The buffer conditions for oxidation in OxDC were optimized based on buffers suitable for electrochemical oxidation and reduction experiments. The protein is typically stored in amine-based buffers, which would interfere with oxidation experiments, as they will also be oxidized during redox cycling experiments. Therefore, after expression and purification of the enzyme it was exchanged into a buffer consisting of 50 mM succinate, 500 mM NaCl, pH 4.2, or 50 mM potassium phosphate, 500 mM NaCl, pH

8.0. Prior to exchanging the protein into the succinate buffer, it was necessary to lower the pH, and this was carried out in a poly-buffer containing 50 mM tris, 50 mM bis-tris,

50 mM piperazine, 50 mM succinate, 500 mM NaCl, and 20% glycerol. The protein was dialyzed into the buffer overnight and then the pH adjusted in 1 pH unit decrements to the desired pH. Once the final pH was reached, the sample was then dialyzed into its final buffer. Stock solutions of 0.1 M hexachloroiridate and ascorbate were used in 2 to

10 μL additions for the oxidation and reduction experiments, respectively.

Results

X-band Redox Cycling

The metal centers in OxDC have been extensively studied by EPR [34-36, 106].

It has been suggested that oxidation of at least one metal center is necessary during

catalysis. In order to generate Mn(III) in OxDC, potassium hexachloroiridate was used as an oxidant and any changes were monitored by low-temperature EPR.

Hexachloroiridate (IV) is a known 1-electron oxidizing agent, and would facilitate the oxidation from Mn(II) to Mn(III) [107]. Figure 4-1 shows the redox cycling of OxDC at pH 4.2 in a mixed solution of 50 mM succinate and 500 mM sodium chloride.

Figure 4-1. Redox cycling of (25 mg/mL) wild-type OxDC in 50 mM succinate, 500 mM NaCl poised at pH 4.2 using hexachloroiridate and ascorbate as oxidizing and reducing agents respectively. Perpendicular mode spectra show WT-OxDC unadulterated (black trace), after addition of 8 mM hexachloroiridate (orange trace), after subsequent addition of 8 mM ascorbate (blue trace). EPR Parameters: 9.618 GHz microwave frequency, 100 kHz modulation frequency, 10 G modulation amplitude, 0.63 mW microwave power.

The manganese spectrum prior to oxidation is shown in the black trace. Upon addition of 8 mM hexachloroiridate, the majority of the signal has been diminished. After addition of 8 mM ascorbate to the oxidized sample, the original signal for the enzyme is restored,

albeit with lower intensity, most likely due to the dilution of the sample after addition of the oxidant and reductant.

It is evident that the manganese oxidation state is affected by the addition of oxidant, and to confirm the oxidation state, parallel mode EPR was utilized. The most likely species generated is Mn(III), but conventional X-band EPR lacks the ability to observe signals generated by such species due to the excessively large zero-field splittings. Parallel mode experiments were carried out simultaneously with the perpendicular mode experiments, and are shown in Figure 4-2.

Initially, a signal with a geff ≈ 5.2 arising from Mn(II) is seen in the black trace.

Figure 4-2. Redox cycling of (25 mg/mL) wild-type OxDC in 50 mM succinate, 500 mM NaCl poised at pH 4.2 using hexachloroiridate and ascorbate as oxidizing and reducing agents respectively. Parallel mode spectra (right panel) show WT- OxDC unadulterated (black trace), after addition of 8 mM hexachloroiridate (orange trace), after subsequent addition of 8 mM ascorbate (blue trace). EPR Parameters: 9.331 GHz microwave frequency, 100 kHz modulation frequency, 10 G modulation amplitude, 0.63 mW microwave power.

After oxidation, the signal is greatly reduced and a new signal is generated with a geff ≈

8.3 indicative of the generation of Mn(III). Similar spectra have been observed in other enzymatic systems such as Mn-superoxide dismutase, photosystem II, Mn substituted

2,3-dioxygenase, as well as synthetic systems such as Mn(III) Salen, Mn(III)-peroxo complexes, and Mn-hydroxo complexes amongst many others [94, 108-114].

Figure 4-3. Redox cycling of (25 mg/mL) wild-type OxDC in 50 mM phosphate, 500 mM NaCl poised at pH 8.0 using hexachloroiridate and ascorbate as oxidizing and reducing agents, respectively. Perpendicular mode spectra show WT-OxDC unadulterated (black trace), after addition of 8 mM hexachloroiridate (orange trace), after subsequent addition of 8 mM ascorbate (blue trace). EPR Parameters: 9.618 GHz microwave frequency, 100 kHz modulation frequency, 10 G modulation amplitude, 0.63 mW microwave power.

Upon reduction of the oxidized enzyme, the Mn(III) signal is lost, and the original Mn(II) signal is restored. Interestingly, if closer attention is given to the geff ≈ 8.3 region of the un-oxidized enzyme signal, there is a signal with similar fine structure as the observed

Mn(III), indicating the presence of Mn(III) prior to the addition of oxidant. This would indicate that the resting state of the manganese is not purely Mn(II) as previously shown experimentally.

The same experiment was then repeated at pH 8.0, outside the range in which

OxDC exhibits activity, as shown in Figure 4-3. Again, the black trace is the spectrum of the un-oxidized wild-type enzyme poised at pH 8.0 in 50 mM potassium phosphate, and

500 mM sodium chloride.

Figure 4-4. Redox cycling of (25 mg/mL) wild-type OxDC in 50 mM phosphate, 500 mM NaCl poised at pH 8.0 using hexachloroiridate and ascorbate as oxidizing and reducing agents, respectively. Parallel mode spectra show WT-OxDC unadulterated (black trace), after addition of 8 mM hexachloroiridate (orange trace), after subsequent addition of 8 mM ascorbate (blue trace). EPR Parameters: 9.331 GHz microwave frequency, 100 kHz modulation frequency, 10 G modulation amplitude, 0.63 mW microwave power.

A different trend is seen in both the perpendicular and parallel mode spectra. The perpendicular mode spectrum shows that after oxidation, not only is the Mn(II) signal

depleted, but there is also the growth of an amino acid radical as well. This is likely a tyrosyl radical, that has been previously identified to appear upon exposure of enzyme to oxalate [35]. The signal centered about 1600 G is proposed to represent a less structured pentacoordinated Mn(II) system. The loss of this signal could be due to the formation of a more structured heaxcoordinated system similar to that observed at g ≈ 2 prior to oxidation, adding to the overall signal observed after reduction. Another possibility is that there was manganese again present in a higher oxidation state that has been converted to Mn(II) after treatment with ascorbate.

Nevertheless, the observed spectrum for the redox-cycled sample resembles that of the untreated sample, implying that the enhanced manganese signal is not caused by the loss of manganese from the enzyme or by some source other than the enzyme itself. The parallel mode spectra (Figure 4-4) also show that the Mn(II) signal is decreased by the addition of oxidant.

A B

Figure 4-5. Experimental and simulated parallel mode spectra for WT-OxDC pH 4.2. A) Mn (II) prior to oxidation (blue trace). Simulation (orange trace) of the signal yields magnetic parameters of g = 2.001, A = 249 MHz, |D| = 0.0447 cm-1, E/D = 0.17. B) Mn (III) after oxidation with 8 mM hexachloroiridate (blue trace). Simulation (orange trace) of the signal yields magnetic parameters of g = 2.00, A = 140 MHz, D = -2.38 cm-1, E/D = 0.13.

Also, a signal grows in at a similar field position as seen for the low pH redox cycling, although with no resolved structure. It should also be mentioned that the fine structure seen at low pH in the un-oxidized sample around 800 G is not present in the high pH spectrum. The same increase in Mn(II) signal is again observed after reduction of the oxidized sample. Simulations of the low pH parallel mode spectra are shown in Figure

4-5.

Figure 4-6. Temperature dependence between 5K to 90K of Mn(III) generated in oxidized wild-type OxDC (25 mg/ml) in 50 mM succinate, 500 mM NaCl poised at pH 4.2 using hexachloroiridate as oxidizing agent. EPR Parameters: 9.406 GHz microwave frequency, 100 kHz modulation frequency, 10 G modulation amplitude, 2 mW microwave power. Experiments were conducted under non-saturating conditions.

The Mn(II) signal (Figure 4-5 A) was found to have magnetic parameters of g =

2.001, A = 239.5MHz (88.85 G), |D| = 0.0447 cm-1 (1.34 GHz), E/D = 0.17, which are

very similar to those previously assigned to the N-terminal manganese [36]. Simulations of the Mn (III) signal provided magnetic parameters of g = 2.00, A = 140 MHz (49.95 G),

D = -2.38 cm-1 (71.2 GHz), E/D = 0.13, which are similar to those of MnSOD and PSII, with the hyperfine well within the range expected for mononuclear Mn(III) [94, 109]. The sign of D was assigned based on the variable temperature EPR, which shows a steady decrease in signal intensity as a function of increasing temperature (Figure 4-6). The temperature dependence trend was then simulated using a positive or negative D to identify the sign as shown in Figure 4-7.

A B

Figure 4-7. Temperature dependent simulations of Mn(III) generated in oxidized wild- type OxDC at 5K (blue), 10K (red), 20K (orange), 50K (purple) and 90K (green) for confirmation of the sign of D. Simulation parameters: A) g = 2.00, A = 140 MHz, D = -2.38 cm-1D = +2.38 cm-1, B) E/D = 0.13.

The Mn(III) signal arises from transitions that occur between the Ms = ±2 manifolds. The decrease in the signal intensity as a function of increasing temperature

2 is indicative of a negative axial zero-field splitting, placing the electron in the dz orbital with the Ms = ±2 manifolds lying lowest in energy. The symmetry expected for Mn(III), a high spin S = 2 complex with 5D electronic ground state and a large negative axial splitting term, is indicative of either a five or six coordinate Mn(III). OxDC has been shown to have its metal centers in both five and six coordinate geometry. The metal

binding site in both the C- and N-terminal site is comprised of 4 conserved amino acid residues that include 3 histidines and 1 glutamate. The other two open valences have been shown to coordinate either one or two water molecules, at the C-terminal site, and two water molecules or 1 water molecule and formate at the N-terminal site. The previous crystal structure data and the simulation of the Mn(III) signal complement one another, providing a good description of the oxidized metal center.

High-Field/Frequency Redox Cycling

The major difficulty faced in studies focused on the metal centers in OxDC is the wide range of zero-field splittings that arise from the two metal centers [36]. Due to the large and variable zero-field splitting of the C-terminal manganese of 1.5 ≤ |D| ≤ 11

GHz, the majority of the signal usually observed at X-band frequencies arises solely from the N-terminal manganese. Thus, it is difficult to confirm that both manganese centers become oxidized, or the rate at which they are oxidized relative to each other.

In order to address the N/C-terminal in the same experiment, the redox cycling was then repeated at 406.4 GHz as shown in Figure 4-8. At low pH, two sets of sextets can be seen, one from each metal site in OxDC. Upon successive additions of oxidant, it is clearly observed that both signals decay at a similar rate, and are effectively completely oxidized, with some residual Mn(II) also reflected in the X-band experiments.

Upon reduction of the treated enzyme, the Mn(II) signal is then recovered, although slightly increased and with shifted relative intensities.

When the experiment was repeated at pH 8.0 (Figure 4-9), only one site is shown to be affiliated with the N-terminal manganese, due to the broadening of the signal affiliated with the C-terminal site [36]. Nevertheless, a similar trend is observed in the low pH oxidation, as expected from the X-band experiments.

Figure 4-8. Redox titration of (25 mg/mL) wild-type OxDC in 50 mM succinate, 500 mM NaCl poised at pH 4.2 using hexachloroiridate and ascorbate as oxidizing and reducing agents, respectively at 406.4 GHz HF-EPR. Instrumental parameters: 50 kHz modulation frequency, 1 G modulation amplitude, 0.2 mT/s sweep rate and at 20K.

Figure 4-9. Redox titration of (25 mg/mL) wild-type OxDC in 50 mM phosphate, 500 mM NaCl poised at pH 8.0 using hexachloroiridate and ascorbate as oxidizing and reducing agents, respectively at 406.4 GHz HF-EPR. Instrumental parameters: 50 kHz modulation frequency, 1 G modulation amplitude, 0.2 mT/s sweep rate and at 20K.

Again, as function of oxidant concentration, a radical is observed after the addition of 16 mM iridate. Upon reduction of the treated sample, the Mn(II) signal is again observed, albeit with some additional shoulders appearing at the low field edge of the first two lines in the sextet, more than likely due to some remaining oxidized radical species.

Figure 4-10. Tyrosyl radical generated upon redox titration of (25 mg/mL) wild-type OxDC in 50 mM phosphate, 500 mM NaCl poised at pH 8.0 using 32 mM hexachloroiridate at 406.4 GHz HF-EPR (blue trace) and simulation (orange trace). Instrumental parameters: 50 kHz modulation frequency, 1 G modulation amplitude, 0.2 mT/s sweep rate and at 20K. Spectral parameters based on simulation were found to be gx = 2.00680, gy = 2.00394, gz = 2.00179, Ax = 30 MHz, Ay = 26 MHz, Az = 20 MHz, requiring gx strain of 0.001 FWHM.

An expanded spectrum of the radical is shown in Figure 4-10. The g-anisotropy is resolved due to the high field/frequency of the experiment, in which the Zeeman energy becomes very large, allowing the tentative identification as a Tyrosyl radical. Simulation of the generated radical provided magnetic parameters of gx = 2.00680, gy = 2.00394, gz

= 2.00179, Ax = 30 MHz, Ay = 26 MHz, Az = 20 MHz, which match well with those in the

literature [67, 68]. Figure 4-2 showed fine structure that could be assigned to Mn(III) at low pH, but Figure 4-4 there was none identified in the high pH parallel mode spectra.

The generation of Mn(III) as a function of lowering pH may indicate that the pH dependence of the enzyme may be related to the availability of Mn(III). The claim for the specific pH dependence exhibited by OxDC was presumed to occur due to the substrate pKa. Oxalic has two pKa’s at 1.2 and 4.2. Reports show the enzyme has its highest catalytic activity in the range of pH 3.9 to 4.2, predicating the requirement for mono-protonated oxalate.

Discussion

Through a combination of multi-frequency/field and parallel mode EPR experiments, the presence of Mn(III) in Bacillus subtilis oxalate decarboxylase has been demonstrated. This is the first spectroscopic evidence for the presence of Mn(III) in wild type OxDC. In order to directly observe Mn(III) in OxDC, the metal centers were chemically oxidized. Redox cycling of OxDC with hexachloroiridate and ascorbate demonstrated that both metal centers can be reversibly oxidized at similar rates, albeit

Mn(III) was clearly observed only at low pH. Previously, it has been suggested that the pH dependence between 4 and 6 exhibited by the enzyme was due to the requirement for the mono-anion of oxalate. However, OxDC exhibits activity above the second pKa

(4.2) of oxalate, suggesting that the measured activity, albeit lower with increased pH, above pH 4.2 would occur with the di-anion of oxalate. It was shown in Figure 4-2 prior to and after oxidation that there is fine structure that could be assigned to Mn(III) at low pH, but a broad unresolved feature was observed in the high pH parallel mode spectra after oxidation (Figure 4-4).

The observed Mn(III) in parallel mode X-band EPR arises from transitions between the closely space ms = ± 2 sublevels. Simulations of the Mn(III) signal provided magnetic parameters of g = 2.00, A = 140 MHz (49.95 G), D = -2.38 cm-1 (71.2 GHz),

E/D = 0.13, where the sign of D was confirmed via variable temperature experiments, indicative of a five or six coordinate Mn(III) system. The Mn(II) system observed in parallel mode measurements, which serves as another source of following the oxidation of the metal center, occurs from double quantum transitions, and simulation of this spectrum provided magnetic parameters of g = 2.001, A = 239.5MHz (88.85 G), |D| =

0.0447 cm-1 (1.34 GHz), E/D = 0.17, very similar to the parameters previously assigned to the N-terminal manganese site. It is also noteworthy to mention that no clearly resolved fine structure indicative of Mn(III) is observed at high pH prior to oxidation, however, this does not prove the absence of Mn(III). The generation of Mn(III) as function of lower pH in the native enzyme may indicate that the pH dependence of the enzyme could be related to the availability of Mn(III).

The metal centers in OxDC have been shown to exhibit a multitude of pH dependent manganese structures [36]. The N-terminal site has been assigned the role as the active site and presents two pH dependent structures, while the C-terminal site exhibits 4 pH dependent structures and has been assigned to possibly provide only a structural role in the protein [30]. The clear observation of Mn(III) only at low pH would seems to also play a role in the overall catalysis, as its presence in the un-treated sample is observed only in the pH range in which the enzyme exhibits activity.

Secondly, the protein itself shows evidence of oxidation based on the presence of a tyrosyl radical. Identification of this radical species is tentatively assigned via the

resolved g-tensor at 406.4 GHz. Such a radical has been previously observed during turnover at pH 5.2 [35], and may be linked to an increased oxidative environment after multiple turnovers.

It is well understood that the ligand environment of the metal center dictates its electronic properties, and proteins typically tune the redox properties as well as oxidation state to carry out the desired mechanism [115]. The pH dependent Mn(II) structures exhibited by OxDC imply that the enzyme undergoes a range of conformational changes that also play a major role and facilitate catalysis within the dictated pH range. It would seem reasonable to assume that protein tunes the redox couple between Mn(II)/Mn(III) to provide a more favorable oxidation by oxygen which has a reduction potential of -0.33V [116, 117]. It has been shown that the redox couple of Mn in a variety of superoxide dismutase’s have been identified in the range of 290mV to 625mV versus the reduction of Mn(III) to Mn(II) in solution at +1.51V [118, 119]. The use of hexachloroiridate provides an upper bound of the reduction potential of the manganese centers in OxDC at +0.9V; however, it can be assumed that it will be much lower. High pH redox cycle of the enzyme also demonstrates that the manganese centers are oxidized; however, there is no clearly observed resolved Mn(III) structure.

The signal that arises in parallel mode at high pH could possibly be identified as Mn(III) with a much larger anisotropy in its magnetic parameters leading to unresolved spectra, or may be due to a system of much larger axial zero-field splitting. Magnetic parameters for the high pH signal could not be ascertained due to the inability to reliably simulate the species.

Still unclear was whether both metal sites are susceptible to oxidation and if so are their oxidation potentials similar. Due to the large and variable zfs of the C-terminal manganese, it is not possible to confirm if it is oxidized under the same experimental conditions as the N-terminal manganese at X-band. The majority of the signal observed at X-band arises from the N-terminal metal center. In order to directly observe if both the

C- and N-terminal manganese are oxidized, it was necessary to carry out experiments at 406.4 GHz, where the Zeeman energy is much larger than the zfs, allowing for direct observation of the C-terminal manganese. It was critical to carry these experiments out at lower pH (4.2), where the zfs of the C-terminal is much smaller (D ≈ 1.5 GHz). At pH

4.2 the observed spectrum arises from transitions between the ms = ± 1/2 spin manifold.

The central 6-line spectrum is split into two sets of sextets with the lower field component of the main sextet arising from the C-terminal site, and the higher field component from the N-terminal site. These experiments (Figure 4-8) confirmed that both metal centers are oxidized at similar potentials, albeit the N-terminal site is oxidized slightly faster. The oxidation of the N-terminal site faster, may indicate the accumulation of Mn(III) at this site more readily as this may be required for the breakdown of oxalate. At higher pH (8.0), the C-terminal manganese has a zfs of D ≈

11 GHz, making direct observation difficult and a single sextet is observed (Figure 4-9), representative of the N-terminal manganese. Unfortunately, due to the low sensitivity of the spectrometer used for the high-field studies, and the high dielectric loss of water, it was not possible to observe transitions affiliated with Mn(III) at any pH. These experiments however, did reveal an amino acid radical after oxidation (Figure 4-10), with a fully resolved g-tensor allowing for tentative identification as a tyrosyl radical

based on comparison with the literature. Previously, the accumulation of a tyrosyl radical was observed in OxDC during catalysis at pH 5.2 [35].

Previously the N-terminal manganese has been shown to be the only metal center directly involved in catalysis, and would only require redox cycling at this site.

Oxidation of both metal centers in OxDC and the structural similarities between the manganese ions would suggest that either site is suitable for oxygen binding. Oxygen binding at the C-terminal site would reduce the possibility for the radical intermediates to interact. This is further supported with spin trapping studies outlined in Chapter 3. The distance between manganese centers intra-monomer is 26 Å, while the distance inter- monomer between C- and N-termini manganese of neighboring monomer units within a single trimer is 21 Å. These distances are typically too long to facilitate a long-range electron transfer. However, the existence of a tryptophan dimer pair (tryptophan-96/274) that is involved in a pi-pi stacking motif between monomers drastically reduces the distances the electron needs to jump to less than 10 Å [36]. The tryptophans are also connected along the backbone of the protein to directly coordinating histidines at the manganese centers providing a direct path for electrons to transfer between metal centers. Upon binding of oxygen at the C-terminal manganese, the metal center is oxidized forming superoxide/hydroperoxyl radical and Mn(III). The mechanism then proceeds as previously described with oxalate binding at the N-terminal manganese with the C-terminal manganese providing the necessary reducing and oxidizing equivalents.

Crystal structures of a putative oxalate decarboxylase from Thermotoga maritime presented oxalate bound in a bi-dentate fashion to the manganese center [120]. If

oxalate binds in the same fashion in B. subtilis OxDC, this would leave no open valences at the N-terminal site for oxygen binding and would require a secondary binding site. A series of inhibitors of OxDC have been identified, with nitric oxide (NO) a dioxygen mimetic, acting as an uncompetitive inhibitor. Studies conducted at X-band including a known nitric oxide releaser did not show any interaction of NO with the N- terminal manganese site [121]. This would further compel the argument that oxygen is binding at the C-terminal manganese.

CHAPTER 5 IMMOBILIZATION OF Bacillus subtilis OXALATE DECARBOXYLASE ON A ZN-IMAC RESIN

Introduction

Electron paramagnetic resonance has proven to be an exceptional technique for studies on OxDC. The manganese sites act as intrinsic probes for following the enzymatic catalysis, and the ability to study production of radical intermediates.

Tremendous effort, both biochemically and spectroscopically has been put forth to provide insights into the enzymatic mechanism [21, 22, 27, 29, 35, 36, 38, 41, 42, 44,

91, 93, 106, 121, 122]. Utilization of an E. coli overexpression system has provided a source for preparing high yields of the enzyme. Nevertheless, the high consumption in carrying out spectroscopic studies has been a bottleneck. Previous EPR studies alluded to the need for protein concentrations upwards of 20 mg/mL to sufficiently study all the spectroscopically relevant forms of the enzyme with sufficient signal-to-noise ratio [106].

In an effort to provide a solution for this high demand, an immobilization protocol of the enzyme utilizing immobilized metal affinity chromatography (IMAC) was developed. Not only can immobilization of the enzyme on a resin reduce consumption, it also provides a pseudo ‘reusable’ sample. A single sample can be exposed to multiple buffer conditions or small molecules by simply equilibrating the resin with a buffer with the desired condition. This technique in principle can be applied to any protein system that utilizes terminal histidine-tags.

It is quite common to utilize IMAC resins in the purification of proteins, by incorporating affinity tags (e.g.) such as a poly(His) tag [123] at either the N- or C- terminus of the protein. The technique is also amenable to systems without a poly(His) tag, such as natural metal binding proteins, antibodies, and phosphopeptides [124]. A

variety of metals have been utilized in IMAC, including Cu (II), Ni (II), Zn (II), Co (II), and

Fe (III). Due to their characteristics as Lewis-acids (electron-acceptors), these metal ions coordinate to nitrogen, sulfur and oxygen, leading to metal chelates by one or more histidines of the poly(His) tag [125]. Surface accessible histidines, outside of the poly(His) tag, can also take part in binding to the metal centers, as well as Glu, Asp,

Arg, Lys, Tyr, Cys, and Met. Other materials have also been utilized for the immobilization of proteins, such as CNBr Sepharose 4B, Fractogel® EMD Azlactone,

Fracto-gel® EMD Epoxy, and Eupergit® C, through covalent linkages by taking advantage of the variability in the amino acid side chains [126, 127]. The added benefit of working with IMAC resins for the purpose of purification or immobilization is the wide variety of additives, salts, solvents, buffers, detergents while wide pH ranges can be tolerated. Industrially, IMAC resins are exploited because scale-up is straight forward, requiring only an increase in the resin bed size [124].

Immobilization of oxalate-degrading enzymes has been used for the purposes of bedside clinical testing for oxalate levels, and for bioremediation of excess oxalate industrially [128-130]. OxDC has also been studied immobilized on Eupergit® C [131]. In that study, the effects of pH, temperature and kinetic parameters were compared for solubilized and immobilized protein. Immobilization on Eupergit® C is expected to lead to multiple attachment sites on the protein with the various side chains of the protein amino acids. Lin et al. reported that an increase in temperature stability of OxDC up to

70 oC, compared to 55 oC for the free enzyme. Both forms of the enzyme showed similar pH dependent activity profiles as a function of increasing pH, although the immobilized enzyme exhibited approximately 27% loss in activity and a 37% increase in

Km. The decrease in catalytic efficiency is attributed to structural deformations of the protein when it is adsorbed on the resin.

The encoded His6-tag of OxDC was exploited to immobilize the enzyme on a Zn-

IMAC resin to be used for further EPR studies [132]. Zinc was chosen as the support due to its diamagnetic (i.e., EPR silent) nature. Utilization of the His-tag for immobilization is precedented by the role it plays in the large-scale purification of proteins [125, 133-135]. Previous studies have also utilized immobilized proteins and model systems to study the effects of immobilization by EPR [126, 136-140]. In an effort to qualify the utilization of OxDC bound to a Zn-IMAC resin, the effects of immobilization were studied by EPR and biochemical analysis. A common experiment carried out when studying the metal centers in OxDC and its variants is the pH dependence of the manganese structure. OxDC has been shown to exhibit a wide variability in its observed

EPR spectra due to the perturbation of the manganese center arising from pH- dependent conformational changes the enzymes experiences. An extensive study of high field/frequency EPR pH dependence was previously carried out for the wild type enzyme and will be utilized herein to identify any perturbations the resin may have on the metal active sites [34, 36, 106]. The results speak to constraints in the conformational sampling of the protein bound to a resin.

Review of Previous EPR pH Dependent Structures

Previous studies utilizing Electron Paramagnetic Resonance, has shown that the resting oxidation state of the majority of Mn ions is +2 [34, 36, 93]. Chapter 4 provided evidence for the redox cycling of the manganese centers between the +2 and +3 oxidation states. The major difficulty that has been faced in further understanding the enzyme is the electronic structure of the individual manganese ions. Initial multi-

frequency experiments identified two observable manganese sites having zero-field splitting (ZFS) parameters of 1200 and 2700 MHz and an E/D of 0.21 and 0.25 at pH

6.0, referred to as Site 1 and 2, respectively [34]. Initially, the site having the larger zfs was associated with the N-terminal active site manganese, as perturbations of this site were seen upon introduction of acetate and formate. This was later shown to be incorrect and the changes in the zfs were actually found to be related to the pH change that occurred upon the introduction of acetate into the sample. Subsequent pH- dependent multi-frequency EPR revealed that there are in fact more than two structures affiliated with the two manganese sites [36]. Tabares et al. provided evidence that there are two pH-dependent forms of the N-terminal manganese, Site A and B, and there are five pH-dependent forms of the C-terminal manganese site, Site H,M,X,L, and L2. That work also showed that the site with the larger zfs was associated with C-terminal manganese, instead of the previous assignment to the N-terminal Mn proposed by

Angerhofer et al.

Table 5-1. Site specific zero field splitting parameters of Mn (II) sites in WT-OxDC Site Species g (iso) A (MHz) D (MHz) E (MHz) Ref. N-term 1 2.00094 250 2700 675 [34] C-term 2 2.00087 253 1200 252 [34] N-term A 2.00088 252 -1350 230 [36] N-term B 2.00077 253 -1110 300 [36] C-term H 2.00080 250 10730 1700 [36] C-term X 2.00080 251 1400 340 [36] C-term M 2.00080 251 -1500 450 [36] C-term L 2.00086 251 4170 720 [36] C-term L2 2.00078 252 5060 250 [36]

The site speciation nomenclature represents the pertinent pH regime: low pH (4.0) for sites A, L and L2; high pH (8.5) for sites B and H; and intermediate pH (5.0-7.0) for sites

M and X. Table 5-1 lists all the previously experimentally measured magnetic parameters.

Later DFT and mutagenesis studies by Campomanes et al. confirmed the assignment made by Tabares et al. [106]. This was accomplished with a W132F OxDC mutant that led to second shell effects on the N-terminal manganese, leaving the C- terminal manganese unaffected. One common theme among all these studies was the need for large volumes of highly concentrated enzyme in order to identify all the subtle changes in the spectroscopic data. This is further complicated by dilutions in the sample incurred when changing the pH of the solution with addition of base, leading to a decrease in signal intensity and making the observation of already weak signals even more difficult. In order to avoid high consumption of proteins, this study attempted to immobilize OxDC on a resin that would provide highly concentrated sample, and allow exposure of one sample to multiple conditions. Effective immobilization would allow for the studies to be carried out on one sample, for example in the presence of multiple pH’s and buffer conditions, as well as small molecules (inhibitors and substrate analogs), with no successive dilutions of the sample and concomitant decrease in the observed EPR signal.

Experimental

Materials

The following chemicals were purchased from Fisher Scientific (Pittsburg PA,

ACS Grade) and used as received without further purification: phosphoric acid, glacial acetic acid, sodium chloride, sodium phosphate, sodium acetate, sodium hydroxide, and tris(hydroxymethyl)aminomethane (Tris). Zinc sulfate was purchased from Sigma

Aldrich (St. Louis, MO). Uncharged Profinity IMAC resin was purchased from Bio-Rad

Laboratories (Hercules, CA). All solutions were prepared utilizing 18MΩ∙cm de-ionized water generated by a Thermo Scientific Barnstead Nanopure model 7134.

Resin Preparation and Enzyme Immobilization

The enzyme expression and purification procedure can be found in Appendix A.

IMAC columns were custom designed of Kel-F (thermoplastic chlorofluropolymer) to serve as cryogenic EPR sample containers (4x5 mm IDxOD, 5.46 cm length for X-band and 6.1x7.3 mm IDxOD, 3.3 cm length for high field). A disk of polypropylene filter paper

(5 micron particle size, Typar 3609L, Midwest Filtration LLC) with a diameter of the ID was fit tightly into the bottom of each column. Resin preparation was carried out following the ProfinityTM IMAC Resin Manual. Uncharged resin (250 µL of 50/50 v/v) and solvent were added to the column, washed for 15 minutes with 1% acetic acid, 0.12 M phosphoric acid for cleaning, followed by 10 column volumes of DI water. Further washing for another 15 minutes with 2 M NaCl removed ionic contaminants followed by rinsing with 10 column volumes of DI water. Ten column volumes of binding buffer (50 mM sodium phosphate, 0.3 mM NaCl at pH 8.0) was then flowed through the column, followed by 10 column volumes of 50 mM sodium acetate (0.3 M NaCl, pH 4.0) to prepare for metal binding. Zn(II) ions were loaded onto the column by applying 5 column volumes of 0.3 M ZnSO4, followed by 5 column volumes of 50 mM sodium acetate, 0.3

M NaCl, pH 4.0 and 10 column volumes of DI water to rinse. Finally, the column was equilibrated with starting buffer (50 mM Tris-HCl, 500 mM NaCl).

WT OxDC was loaded onto the column by passing 400 µL of 25-40mg/mL free enzyme solution through the column. The eluent was collected and passed through the column at least three more times to capture as much His6-tagged OxDC as possible.

The column was then re-equilibrated with starting buffer, leaving it ready for use.

Enzyme Kinetic Assays

The Michaelis-Menten parameters of the decarboxylase activity of free and resin- bound OxDC were determined through an end-point assay measuring the production of formate, as previously described [30, 93, 106]. Protein-loaded IMAC resin (125 µL washed and centrifuged before re-suspension to remove any un-bound enzyme) was mixed with 875 μL of starting buffer and constantly agitated to prevent sedimentation.

Reactions were initiated in a 25°C water bath by adding 10 μL of the slurry (or 1.5 μL of free WT OxDC for the control reaction) to 99 μL buffered oxalate solutions (acetate buffer at either pH 4.2 or pH 5.5). Because low pH was expected to diminish the protein binding capacity of the resin due to increasing protonation of histidine (pKa ≈ 6), test were carried out at two pH values, 4.2 and 5.5. Similar reactions were conducted at both pH values using free WT OxDC as the control.

Microscopy

A 50-100 µL aliquot of resin slurry was placed on a microscope slide and covered with a glass coverslip. A Zeiss PrimoVert microscope with various magnification levels ranging from 4-40x was used to observe the resin beads before and after freeze-thaw cycles.

Multi-Field/Frequency Electron Paramagnetic Resonance

Experiments were performed on a Bruker ELEXSYS E580 CW/Pulsed or a Bruker

ELEXSYS-II E500 CW X-band equipped with an Oxford ESR900 helium flow cryostat using a Dual Mode Cavity (Bruker ER 4116DM). The bottom 10 mm of the sample column was carefully placed in the center of the resonator each time an experiment is performed. To do this reproducibly the collar sealing the Kel-F rod that holds the sample column was never removed from it. Since the collar attaches to the resonator’s sample

100

stack, the sample itself was always held at the same vertical position. Correct sample placement was also visually inspected through the resonator window. Experimental conditions were typically: 100 kHz modulation frequency, 10 G modulation amplitude,

0.63 mW microwave power, and temperature set to 5 K. High field/frequency measurements were carried out on a variable frequency/field broadband transmission spectrometer [141] at 406.4 GHz in a field ranging from 13.9 T to 14.9 T, 50 kHz modulation frequency, 1 or 25 G modulation amplitude and 0.2 or 2 mT/s sweep rates for high-resolution narrow sweeps or low-resolution wide sweeps, respectively, in a temperature range between 3 and 20 K. Simulated spectra were generated using the

EasySpin toolbox in MATLAB[105].

Results

X-Band EPR and Enzyme Kinetics

Initial studies of the immobilized OxDC were carried out at X-band to characterize the free versus immobilized forms of the enzyme, as well as Michaelis-

Menten kinetics, to identify any effect on the enzyme activity after immobilization. Figure

5-1 compares the X-band EPR spectrum of the free enzyme (black trace) and immobilized enzyme (red trace) recorded at 5 K. The two spectra in Figure 5-1 look very similar. All EPR peaks seen in the free enzyme are reproduced in the spectrum of the immobilized sample with small variations in relative intensity, possibly due to different relative enzyme concentration and metal incorporation.

As mentioned above, a coupled assay is used to quantify the total amount of formate produced after a specified time via the decarboxylase pathway of OxDC.

Samples of the free enzyme and IMAC bound enzyme were studied at two pH values,

101

4.2 and 5.5. The enzyme exhibits optimal activity at a pH centered about 4, with lower activities up to pH 6.

Figure 5-1. Free (black trace) and Immobilized (red trace) WT-OxDC at 5 K. Instrumental parameters: 100 kHz modulation frequency, 10 G modulation amplitude, and 0.63 mW microwave power. Reprinted from Biochemistry and Biophysics Reports, Twahir,U.T., Molina,L., Ozarowski, A., Angerhofer, A., 4, 98-103, 2015, with permission under a creative commons license from ScienceDirect.

The higher pH in this study was chosen based on the average pKa (6.0) of histidines, the coordinating ligands to metal scaffold in the resin. Typically, purification procedures involving poly(His) tags are carried out under slightly basic conditions to account for the pKa of histidine. Experiments at pH 5.5 involve a mixed protonation state of the His-tag, promoting a higher population of bound protein as compared at pH 4.0. All kinetic parameters are reported in Table 5-2, derived from Michaelis-Menten curves in Figure

5-2.

102

The Michaelis-Menten data suggest that there is essentially no effect on the binding of the substrate, according to the extremely similar KM values. At both pH points,

Table 5-2. Michaelis-Menten kinetics of free and immobilized OxDC. Reprinted from Biochemistry and Biophysics Reports, Twahir,U.T., Molina,L., Ozarowski, A., Angerhofer, A., 4, 98-103, 2015, with permission under a creative commons license from ScienceDirect. Free OxDC Immobilized Free OxDC Immobilized (4.2) OxDC (4.2) (5.5) OxDC (5.5) Enzyme 7.2 ± 0.27 19.1 ± 0.4 7.2 ± 0.2 19.1 ± 0.4 Concentration [μM]

Vmax [mM/s] 1.13 ± 0.09 1.08 ± 0.05 0.076 ± 0.005 0.054 ± 0.007

Vmax [U/mg] 215 ± 17 77 ± 4 14.4 ± 0.09 3.9 ± 0.5

KM [mM] 12 ± 3 16 ± 2 7 ± 1 8 ± 3 -1 kcat [s ] 158 ± 13 56 ± 3 10.6 ± 0.7 2.8 ± 0.3 -1 -1 kcat/KM [s M ] 13000 ± 3000 3600 ± 500 1600 ± 300 400 ± 100

Figure 5-2. Comparison of Michaelis-Menten kinetic analysis of decarboxylase activity of free OxDC at pH 4.2 (blue) and pH 5.5 (red) as well as IMAC-bound OxDC at pH 4.2 (green) and pH 5.5 (cyan). Reprinted from Biochemistry and Biophysics Reports, Twahir,U.T., Molina,L., Ozarowski, A., Angerhofer, A., 4, 98-103, 2015, with permission under a creative commons license from ScienceDirect.

103

according to kcat/KM, there is a 4-fold decrease in overall activity for the IMAC bound enzyme. However, the maximal rate of the enzyme is almost unchanged. The decreased kcat/KM may be attributed to the steric crowding that is caused by the immobilization. It becomes more difficult for the substrate to diffuse to the active site of

OxDC; however, upon entering the active site binding is unaffected (similar KM). This approach would be an excellent methodology for studying small molecule interactions, such as those of inhibitors with active site metal, as the binding is unaffected, although the rate is slowed. These initial EPR and kinetic experiments reflect only perturbations at the N-terminal site, assuming the C-terminal site is not directly involved in catalysis.

The high field/frequency experiments described in the next section report on both metal sites and provide a more comprehensive understanding of the overall effects.

Multi-Field/Frequency Electron Paramagnetic Resonance

After successful immobilization of the enzyme on an IMAC resin and determination of the catalytic efficiency, it was critical to identify any potential spectroscopic perturbations the immobilization may have on the active site metals. In order to accomplish this, experiments were carried out at 406.4 GHz, and 14 T. Due to the range of zfs parameters that exist for the enzyme metal sites, it is not possible to address all of the possible pH dependent sites within the enzyme at X-band frequencies.

Experiments were carried out on both the IMAC bound and free enzyme. The site speciation identification of Tabares et al. shown in Table 5-1 is used to qualify the signals observed in this experiment. The N-terminal site shows two distinct pH dependent structures Site A and B. These sites were identified at 3K, at which the higher order spin manifolds become more populated due to the Boltzmann distribution,

104

thus allowing for observation of the |±5/2>|±3/2> transitions. The major benefit of conducting such experiments is the ability to directly attain the value and sign of D and

E. The low pH form of the N-terminal site can be seen in Figure 5-3.

There are three discernable features in the free enzyme at pH 4.06 (red trace): the central feature located at about 14.52 T is representative of the residual population of ms = ± 1/2 spin manifolds, a low field feature centered at 14.32 T, and a higher field split feature centered about 14.6 T.

Figure 5-3. The 406.4 GHz HF-EPR spectra at 3K of WT-OxDC free enzyme at pH 4.06 (red), immobilized at pH 3.88 (black) and simulation (blue) based on site A previously determined magnetic parameters shown in Table 5-1. Spectral Parameters: 50 kHz modulation frequency, 25 G modulation amplitude, 2 mT/s sweep rate. Reprinted from Biochemistry and Biophysics Reports, Twahir,U.T., Molina,L., Ozarowski, A., Angerhofer, A., 4, 98-103, 2015, with permission under a creative commons license from ScienceDirect.

The same features are also observed for the IMAC bound enzyme at pH 3.88, although the high field split feature is slightly broadened due to a slightly more anisotropic distribution of its zfs. For reference, the spectral simulation for Site A (blue trace) is also included. It was necessary to include approximately 25% D/E-strain in the simulated

105

spectra to account for the spectral width of the features. These results suggest that the

N-terminal manganese low pH form is overall unaffected by the immobilization process.

The high pH form of the enzyme was also studied, as this is useful in studying binding interactions of solute and small molecules, such as substrates and inhibitors, under non-turnover conditions. Experiments were conducted at a pH of 8.42 and 8.50 for the free (red trace) and immobilized (black trace) enzyme, respectively, as shown in

Figure 5-4. For reference the spectral simulation for Site B (blue trace) is also included.

As expected, there are subtle changes between the high and low pH spectra for the N- terminal site, such as shifting of the low field line to higher fields (14.33 T to 14.36 T) due to the decrease in the absolute value of D. The high field feature also loses its low field counterpart observed in the low pH spectra.

Figure 5-4. 406.4 GHz HF-EPR at 3 K of free enzyme at pH 8.42 (red) and immobilized at pH 8.50 (black). The simulation (blue) is based on parameters of site B by Tabares et al. [15] Instrumental parameters: 50 kHz modulation frequency, 25 G modulation amplitude, and 2 mT/s sweep rate. Simulation parameters: g = 2.00077, A = 253 MHz, D = -1100 MHz, E = 300 MHz. Reprinted from Biochemistry and Biophysics Reports, Twahir,U.T., Molina,L., Ozarowski, A., Angerhofer, A., 4, 98-103, 2015, with permission under a creative commons license from ScienceDirect.

106

Experiments were then carried out to investigate the high pH C-terminal site. The zfs of this site is so large that it becomes difficult to observe transitions between higher spin manifolds due to the spread of its resonances across a very large field range.

However, the |+1/2>|-1/2> transitions exhibit large second-order effects that are visible at the low field side of the main sextet at 20K in OxDC (Figure 5-5) [36]. These resemble the structure seen in the spectra of MnSOD [142]. In order to identify the high pH form of this site, experiments were carried out in the same fashion as those at 3K, except at an elevated temperature of 20K where the ms = ± 1/2 spin manifolds show much higher population (Figure 5-5). It can be seen from Figure 5-5 that site H is observed in both the free and IMAC bound forms of the enzyme. The feature is much more profound in the IMAC bound enzyme due to the enhanced signal-to-noise.

A B

Figure 5-5. The 406.4 GHz HF-EPR spectra at 20K and two pHs of WT-OxDC free enzyme A), and immobilized B) and simulation based on site H (blue) previously determined magnetic parameters shown in Table 5-1. Spectral Parameters: 50 kHz modulation frequency, 25 G modulation amplitude, 2 mT/s sweep rate. Reprinted from Biochemistry and Biophysics Reports, Twahir,U.T., Molina,L., Ozarowski, A., Angerhofer, A., 4, 98-103, 2015, with permission under a creative commons license from ScienceDirect.

The shoulders that appear on the low and high field sides of the main sextet can be

107

attributed to the |±3/2>  |±1/2> transition between higher spin manifolds. A simulation of site H is also included in the Figure 5-5 for comparison.

To identify the remaining sites of the C-terminal manganese, experiments were conducted at 20K over a much narrower field range (Figure 5-6). The appearance of the mid-pH species (M and X) leads to the appearance of a secondary sextet of lines on the low field side of the high pH lines. At high pH a single species is seen to exist, which is the same species seen in Figure 5-2, however, now focusing on the transitions that arise from the ms = ± 1/2 spin manifolds. As the pH is lowered in the free enzyme to

7.56 a second set of resonances begin to appear as shoulders on the low field portion of the high pH sextet.

A B

Figure 5-6. pH dependence of the 406.4 GHz HF-EPR spectra at 20K of A) free enzyme, B) immobilized. Simulations for sites M and X are shown at the bottom. Instrumental parameters: 50 kHz modulation frequency, 1 G modulation amplitude, and 0.2 mT/s sweep rate. Reprinted from Biochemistry and Biophysics Reports, Twahir,U.T., Molina,L., Ozarowski, A., Angerhofer, A., 4, 98-103, 2015, with permission under a creative commons license from ScienceDirect.

Once the pH is further lowered in the range of 4-6, the presence of a total of 4 species becomes evident. Site M and X are the mid pH C-terminal species that become prominent in the pH 6.60 and 5.39 spectra. Simulations of these sites have been included based on the values reported in Table 5-1. As the pH is lowered to 4.06 a third

108

shoulder can be seen at the most extreme low field region, which is the appearance of site A. Similar behavior involving the appearance of a secondary species has also been observed by Angerhofer et al. when experiments of WT-OxDC were carried in a pH range of 5.2 to 6.0 [34].

When the analogous experiment was carried out on the IMAC bound enzyme

(Figure 5-6 B), a slightly different trend was observed. At high pH only a single species is seen, indicative of site B. As the pH is lowered the intermediate sites are not observed until pH 5.22 where they appear as small shoulders on the low field side of the existing sextet. The features become more pronounced as the pH is further lowered.

Tabares et al. previously constructed pH speciation curves, which suggested that the appearance of the intermediate pH species occurred at pHs as high as 7.5 and could exist at pHs as low just below 4. The appearance of site X occurs first at higher pH and equal amounts of site X and M exist around pH 6. Simulations of sites X and M are included in Figure 5-6 for visual aid.

The trend in Figure 5-6 is the same as that observed in the free enzyme pH dependence. However, this indicates that the site X is largely suppressed in the IMAC bound enzyme. This could be due to two scenarios: the experimental parameters suppress site X to some extent in both forms of the enzyme, but more so in the IMAC bound form; or upon binding of the enzyme on the resin, this secondary intermediate pH site is restricted and site M is preferred under these conditions. Nevertheless, the intermediate pH species is observed in both forms of the enzyme in the pH regions where the enzyme is active.

109

The lowest pH form of the C-terminal site (site L) is barely seen as shoulders on the edges of the main sextet as shown in Figure 5-7.

Figure 5-7. The 406.4 GHz HF-EPR at 20K spectra of WT-OxDC free enzyme (black), immobilized (red) and respective simulations for site L (blue) based on previously determined magnetic parameters shown in Table 5-1. Reprinted from Biochemistry and Biophysics Reports, Twahir,U.T., Molina,L., Ozarowski, A., Angerhofer, A., 4, 98-103, 2015, with permission under a creative commons license from ScienceDirect.

The most intense feature of this species is observed on the low field edge of the high field line of the main sextet. Again we see here that the free enzyme shows a more pronounced shoulder than that of the IMAC bound enzyme, but the feature is observed in both forms.

Effects of Freeze-Thaw Cycles on Resin and Immobilized Enzyme

Immobilization of enzyme on an IMAC resin allows for the exposure of a single sample to multiple buffer conditions. In the case of OxDC, the focus of the EPR experiments is on the manganese centers, and as such the majority of experimentation is carried out at liquid helium temperatures. Therefore, it is of interest to understand the effects of freezing and thawing on the bound enzyme, as well as on the resin itself.

110

Figure 5-8 shows free OxDC and the immobilized enzyme and the effect of freezing and thawing on the resin-bound enzyme. Initial binding of the enzyme to the Zn-IMAC resin demonstrates that the overall signal intensity of the resin bound enzyme and the free enzyme in solution is approximately the same.

Figure 5-8. Free and Immobilized WT-OxDC and the effects of freezing and thawing the resin bound enzyme. Respective concentrations of immobilize enzyme reflected in Table 5-3. Spectral Parameters: 100 kHz modulation frequency, 10 G modulation amplitude, 0.63 mW microwave power, and temperature set to 5 K. Reprinted from Biochemistry and Biophysics Reports, Twahir,U.T., Molina,L., Ozarowski, A., Angerhofer, A., 4, 98-103, 2015, with permission under a creative commons license from ScienceDirect.

Between each freeze/thaw cycle, 10 column volumes of starting buffer (50mM

Tris, 500mM NaCl, pH 8.5) was flowed through the column and then refrozen in liquid nitrogen prior to the next experiment. After the second freeze/thaw, a slight decrease is observed in the manganese intensity, immediately followed by an approximate 2-fold increase in the overall signal intensity that is shown to persist up to five freeze-thaw cycles. It is noteworthy to mention that at high pH the zfs parameters for the C-terminal manganese are so large that the bulk signal observed at X-band frequencies

111

corresponds to the N-terminal site. With every freeze-thaw cycle the flow through was collected and monitored utilizing a Bradford assay for protein that was eluted from the resin as reported in Table 5-3. The EPR data suggests that the apparent protein concentration is increasing according the signature manganese spectra of OxDC as a function of freeze/thaw cycles. However, protein concentration assays showed that with each freeze-thaw cycle protein is eluted from the column.

Table 5-3. Total protein on IMAC resin after initial loading and subsequent freeze-thaw cycles. Reprinted from Biochemistry and Biophysics Reports, Twahir,U.T., Molina,L., Ozarowski, A., Angerhofer, A., 4, 98-103, 2015, with permission under a creative commons license from ScienceDirect. Freeze-thaw cycles IMAC-bound OxDC (mg) 0 13.7 1 10.9 2 9.9 3 8.9 4 8.4 5 7.7 6 7.0

There are a few explanations for this observation which all involve one main concept, the “sweet spot” of the resonator. EPR resonators are designed such that the electric field component of the microwave is minimized and the magnetic field component is maximized at the center of the resonator where the sample is placed.

Thus, there is a gradient for the optimal position (the sweet spot) in the center of this delicate balance between the electric and magnetic held within the resonator as shown in Figure 5-9. If during freeze/thaw cycles the resin is expanded and contracted followed by additional column washes, protein could migrate through the resin from areas not optimally positioned in the resonator (i.e., towards the top of the resin bed) to a region that is more suitable and enhances the EPR response. The second possibility for

112

enhanced signals after freeze-thaw cycles is the breakdown and migration of smaller beads or fraction with their bound protein, into the more sensitive portion of the cavity.

2 2 Figure 5-9. Magnetic (B1 ) and electric (E1 ) fields across a TE102 cavity, with the rectangular border representing the walls of the resonator. The magnetic field is maximized in the center of the cavity, such that sample will be placed in a valley between the electric components. Optimal sensitivity is achieved with sample placement in the B1 field and as far away from the E1 field. Overlay of the B1 and E1 fields would show an optimal positioning of the sample [143].

In order to assess the effect of repeated freeze-thaw cycles on the resin, samples were put through repeated freeze/thaw treatment and studied under a microscope as shown in Figure 5-10. Fresh IMAC Profinity resin kept at 4°C consists of spherical polymeric UNOsphere beads of various sizes ranging from 51-63 µm (as described by the ProfinityTM user manual). In Figure 5-10 (A) and (C), micrographs of the resin beads are shown under 20x, and 4x magnification. Prior to freezing the beads exhibit a consistent spherical shape. After several cycles, the beads show significant fracturing as seen in Figures 5-10 (B) (black arrows) and 5-10 (D). Many of the large beads are broken into smaller, misshapen pieces, and could potentially lead to higher surface area and tighter packing on the column. The fracturing and breaking of the resin is expected due to the fragility of the polystyrene beads typically used for protein

113

purification resins. The breakdown of the polymer beads, and the observed increase in the observed EPR signal, would be explained by the second situation described above.

What remained unknown was the effect on the enzyme activity after exposing the enzyme bound resin to multiple cycles and freezing and thawing. Experiments were then carried out on a sample that had been exposed to 10- freeze thaw cycles, and reported in Table 5-4 with the initially immobilized enzyme kinetics data for comparison.

After 10 freeze-thaw cycles it can be seen that there is another approximate 4-fold decrease in activity with no effective change in KM at pH 4.2.

Figure 5-10. Comparison of fresh, unfrozen, Zn-loaded resin observed under A) 20x and C) 4x magnification, and OxDC-loaded resin exposed to 6 freeze-thaw-wash cycles observed under B) 20x and D) 4x magnification. Reprinted from Biochemistry and Biophysics Reports, Twahir,U.T., Molina,L., Ozarowski, A., Angerhofer, A., 4, 98-103, 2015, with permission under a creative commons license from ScienceDirect.

These results carry both negative and positive results in that the enzyme has suffered from a large loss in overall efficiency; however, at the same time the results

114

also provide evidence that the enzyme is still active after exposure to 100 column volumes of buffer and 10 freeze-thaw cycles.

Table 5-4. Michaelis-Menten kinetics of free and immobilized OxDC after freeze-thaw cycling. Free OxDC Immobilized Immobilized (4.2) OxDC (4.2) OxDC after 10 Freeze-Thaw Cycles (4.2) Enzyme 7.2 ± 0.2 19.1 ± 0.4 10 ± 1 Concentration [μM]

Vmax [mM/s] 1.13 ± 0.09 1.08 ± 0.05 0.12 ± 0.03

Vmax [U/mg] 215 ± 7 77 ± 4 0.05 ± 0.01

KM [mM] 12 ± 3 16 ± 2 15 ± 7 -1 kcat [s ] 158 ± 13 56 ± 3 12 ± 3 -1 -1 kcat/KM [s M ] 13000 ± 3000 3600 ± 500 800 ± 400

Figure 5-11. A pH titration of WT-OxDC immobilized on a Zn-IMAC resin using a poly- buffer posed at desired pH. A) Parallel mode X-band spectra of immobilized WT OxDC at different pH. EPR parameters: 9.384 GHz microwave frequency, 100 kHz modulation frequency, 10 G modulation amplitude, 0.63 mW microwave power, 5K. B) Relative intensity of the low-field line of the g ≈ 8.8 Mn(III) signal relative to the low field line of the g ≈ 5 Mn(II) signal as measured in the same spectrum.

A pH dependent study was carried out on the immobilized enzyme. A single sample was exposed to multiple buffer conditions in a very short period of time

(approximately 5 hours). In order to vary the pH of the resin bound enzyme, a poly

115

buffer containing 50 mM tris, 50 mM bis-tris, 50 mM piperazine, 50 mM succinate, and

500 mM NaCl, was used to vary the pH from 8.5 to 4.0. Parallel mode spectra of the pH dependence are shown in Figure 5-11. Upon addition of the poly buffer at pH 8.5 to the sample, some subtle changes to the Mn(II) spectrum are observed. The relative intensity of the low-field lines of Mn(III) vs. Mn(II) is shown in Figure 5-11 B and indicates a steady increase in Mn(III) concentration relative to the Mn(II) site we are observing at g ≈ 5.

Discussion

In an effort to provide a solution for the high protein consumption of enzyme during spectroscopic investigations, an immobilization of oxalate decarboxylase was carried out, taking advantage of the intrinsic His-tag used for purification of the enzyme.

IMAC resins have shown great promise in their ability to bind high concentrations of proteins, but primarily for the purpose of purification. Initial low-temperature X-band

EPR spectra (Figure 5-1) of free and immobilized OxDC both show very similar spectra with slight differences in intensity likely due to differences in the concentration. Strong transitions are observed near g ≈ 2 and a weaker one near g ≈ 4 (half-field), which are split by hyperfine interaction with the 55Mn nucleus with a coupling strength of approximately 90 G [35, 93]. The half field signal might be due to the xy1 transition [144] of a C-terminal Mn(II) site with a fine structure parameter |D| of the order of 4 GHz [36].

The immobilized enzyme also still maintains approximately 25% activity as compared to the WT enzyme. The immobilized enzyme exhibits the similar binding affinity as that of the free enzyme with KM’s of 16 mM (free) and 12 mM (immobilized). These results suggest that immobilization affects only the overall activity, which could be explained by

116

steric crowding when enzyme is attached to the resin. However, once substrate diffuses close enough, overall kinetics are unaffected.

One major goal of the immobilization of OxDC is to study the multiple manganese structures exhibited by the enzyme as a function of pH. In order to address the range of manganese structures present in OxDC as a function of pH, the study was further carried out at 406.4 GHz. Previously multiple studies were conducted in an effort to identify all manganese species within the enzyme, which were observed in different pH environments as shown in Table 5-1 [34, 36, 106]. These studies were used as qualifiers for the immobilized enzyme. All the pH dependent forms of the enzyme were observed in the immobilized enzyme with some small differences. When studying species such as manganese in proteins, a wide range of zfs is observed. The higher spin manifolds and transitions between them are more sensitive to the zfs, and allow for these magnetic parameters to be easily observed as first order effects on the spectrum.

Sites A and B, the low and high pH forms of the N-terminal site, were studied at

3K where transitions between the |±5/2>|±3/2> are the prominent features of the spectrum. Site A (Figure 5-3), the low pH form of the N-terminal manganese showed some broadening of its high field lines, indicative of an increase in the anisotropy of the zero field splitting. The spectrum matched well with that of the free enzyme indicating it was overall unaffected. The spectrum of site B (Figure 5-4), the high pH conformation of the N-terminal site, also matched well with the spectrum of the free enzyme.

The pH dependent sites of the C-terminal Mn appear as perturbations of the transitions that occur between the ±1/2 spin manifolds. Identification of these sites was carried out at 20K where the ms = ±1/2 are more populated. Site H, the high pH

117

conformation of the C-terminal site, exhibits an extremely large zero field splitting of D =

10.7 GHz, leading to a broadening of its resonances across the magnetic field.

However, at 20K signals associated with this site are observed as second order effects on the low field side of the transitions occurring between the ±1/2 spin manifolds (Figure

5-5). These transitions are observed in both the free and immobilized enzyme, albeit with lower intensity in the free enzyme. As the pH is lowered in the range of 7.5 to 4.5, the zfs of the C-terminal site is lowered to D = 1.4 GHz (site X) and -1.5 GHz (site M).

Site X is suppressed in the immobilized enzyme; however a secondary set of resonances is shown to appear as a lower field sextet in Figure 5-6 A for the free enzyme. Site X is expected to appear in a pH range of 5.5 to 7.5. This only slightly overlaps with the active pH range of OxDC, and would only slightly affect the overall activity observed. Site M appears in the immobilized enzyme when the pH is lowered in the range of pH 5 and lower. Site M is also observed in the free enzyme. Site L the low pH conformation of the C-terminal site is also observed in both the free and immobilized enzyme as low field shoulders on the main sextet, showing that the majority of the expected sites are still observed after immobilization.

Initial immobilization indicated that there was no benefit in terms of signal intensity via EPR versus the free enzyme at X-band. But after multiple freeze-thaw events increased signal intensity was observed, as shown in Figure 5-8. There are two possible explanations for this: migration of protein through the resin from areas outside of the sensitive portion of the resonator to more sensitive area or tighter packing of the resin after multiple freeze thaw cycles, leading to same movement of protein into the more sensitive portion of the resonator. Addition of glycerol to the buffer, typically added

118

to prevent aggregation of protein in solution, showed no major increase in signal. This removes the possibility that the initial decreased signal intensity was due to a spin-spin interaction from neighboring manganese sites. Microscopic examination of the resin bed before and after multiple freeze-thaw events revealed that the beads in the resin begin to fracture, and the morphology of the beads begins to change. Disruption of the large beads into a collection of fragmented beads allows for the migration of the smaller portions with protein attached further down the column, leading to an increased effective concentration of the resin, which is in line with the second proposal. This effect is seen to occur even though concomitant measurements of the total protein concentration of the resin showed a decrease as a function of successive freeze-thaw cycles, but with a large concentration of the protein still retained (Table 5-3).

After seeing that the resin begins to fracture with concomitant loss of protein, the question remained what happens to the activity of the protein. The purpose of these experiments was to study the sample under multiple conditions. In the manner carried out here, this would consistently include multiple freeze-thaw cycles. Therefore the activity of the protein was again measured after a sample had been exposed to 10 freeze-thaw cycles. Again the protein was shown to exhibit four-fold decrease in activity, with no observable effects on KM. Even though another loss in activity was observed, this experiment also shows that after exhaustive freeze-thaw cycles the enzyme still retains activity.

As an example of the potential uses of the immobilized enzyme, a pH dependent study was carried out at X-band, and changes in the observed spectra were followed with parallel mode EPR. Chapter 4 details the initial observations of Mn(III) in OxDC.

119

The parallel mode EPR spectra in Figure 5-11 shows that initially at pH 8.5 a single sextet of lines is observed recognized as Mn(II). As the pH is lowered into the active range of the enzyme, a second sextet of lines is observed centered about 800 G. This signal is recognized as Mn(III) in OxDC. Previously, in order to clearly observe Mn(III) in

OxDC it was necessary to chemically oxidize the metal centers. However, in this case simply lowering the pH generated Mn(III). This further suggests that Mn(III) may in fact be the active form of the metal centers in OxDC under turn-over conditions. These results provide an example of an application of the IMAC resin in low-temperature EPR studies. If these experiments were carried out in solution on free enzyme, the sample would suffer from a dramatic decrease in overall signal due to the dilution incurred by changes in the pH. These results provide a major contribution to knowledge of the overall mechanism of OxDC, providing further evidence for the pH dependent behavior of the manganese centers, in particular, the effect of pH on the oxidation state of the metal centers.

Immobilization of OxDC on an IMAC resin presents a new possibility for studying this enzyme. Immobilization allows for extensive studies to be carried out, without the high consumption of protein, with the additional benefit of enabling studies on a single sample under multiple conditions. After immobilization, the overall activity is shown to suffer, but even after multiple freeze-thaw cycles retains activity, with but one of the spectroscopically observable manganese sites unaffected.

One of the major drawbacks encountered with studies of enzymes in the presence of potential ligands of interest, is the difficulty in removing any unwanted substrates. This procedure may include extensive dialysis for an extended period of

120

time to recover unadulterated enzyme. In the case of the enzyme bound to a resin, simply equilibrating the resin with the desired buffer system allows for exposure of one sample to multiple conditions in a much shorter period of time. This methodology can easily be extended to any enzyme system that utilizes a non-cleavable affinity tag.

121

CHAPTER 6 CONCLUSIONS AND FUTURE OUTLOOK

Introduction-Preliminary X-Ray Structural Analogs

The enzymatic mechanism of Bacillus subtilis oxalate decarboxylase has been the subject of considerable research, due to the enzymes possible applications medicinally and industrially. Current mechanistic schemes suggest that the N-terminal manganese site is the location of catalysis. Upon binding of oxygen and oxalate to the manganese site, degradation of oxalate occurs via a radical mediated mechanism.

Oxygen is proposed to bind to the metal site, where it abstracts an electron from the manganese center, generating superoxide and Mn(III) as shown in Chapters 3 and 4. In

Chapter 5 enzyme immobilization technique showed that the presence of Mn(III) exhibits pH dependent behavior. Nevertheless, there are still many unanswered questions concerning the overall mechanistic scheme. EPR-spin trapping experiments suggest that the site of superoxide production is not the same as the site of catalysis

[41]. If this is in fact the case, it begs the question, where does oxygen bind in the protein. There are three possible cases for oxygen interaction with the protein: (1) oxygen binds at the N-terminus and protonation of the carbon dioxide anion radical is much faster than the generation of a peroxycarbonate intermediate; (2) oxygen binds at the C-terminal manganese and the mechanism occurs via a long-range electron transfer, (3) oxygen binds at some distal site generating superoxide via an amino-acid radical.

Site-directed mutagenesis and pH-dependent EPR experiments suggest that the

C-terminal manganese plays a critical role in catalysis [36, 50]. Oxidation of the metal centers in OxDC shows that both metal sites are susceptible to oxidation with very

122

similar redox potentials, which would suggest that oxygen could bind to either site.

Catalysis could then occur via a LRET mechanism utilizing a pair of stacking tryptophans between neighboring monomer units in a single trimer, as shown in Figure

1-3. Previously unpublished data identified a series of competitive (bicarbonate), non- competitive (thiocyanate and azide), and un-competitive (nitric oxide and nitrite) inhibitors for OxDC. EPR experiments which included methylamine hexamethylene methylamine (MAHMA) NONOate, a known NO releaser, suggested that NO does interact with the N-terminal manganese [121]. It can be assumed that bicarbonate binds at the N-terminal manganese site, however, the binding site for the other inhibitors are not as obvious. Advanced EPR techniques such as ESEEM and ENDOR would be useful in identifying the location of these small molecules if they are within proximity of the N-terminal of the metal centers up to 9 Å away [145]. Another methodology for identifying the binding sites and interaction of the small molecules with the protein is through protein crystallography, previously carried out with OxDC and OxOx [30, 92].

Protein Crystallography

Previous crystal structures of OxDC have shed light on its structure and reactivity

[28, 30]. The identity of the active site of OxDC was confirmed by co-crystallization of formate, a product of catalysis, at the N-terminal manganese. Addition of the small molecule inhibitors to the protein prior to crystallization, may determine if the C-terminal manganese is directly involved in catalysis, and uncover any additional binding sites in the protein. In this work, initial crystal screens provided three optimized conditions, two at pH 7.5 very similar to those previously used, and one at pH 4.6 closer to the active pH of the enzyme. To our knowledge this is the first report of OxDC crystallization within its active pH range.

123

Material and Methods

OxDC was expressed and purified as described in Appendix A and was then further purified with a HiTrapQ HP anion exchange column (column size of 5mL,

GEHealthcare, AKTA FPLC System) and HiLoad 16/60 SuperDex-200 gel filtration chromatography (GE Healthcare). Initial crystallization of the native His6-tagged OxDC was determined using the sparse matrix screen, Crystal Screen HT, with 2-4 mg/ml of purified protein. Optimized conditions were found to be: (A) 0.2M MgCl2 or 0.2M CaCl2,

0.1M sodium HEPES pH 7.5, and 30% v/v PEG 400; (B) 0.1M sodium acetate trihydrate pH 4.6, 0.2M sodium chloride, and 30% v/v (+/-)-2-methyl-2,4-pentanediol. Crystals at pH 7.5 grew in 4-7 days, while crystals at pH 4.6 took up to 3 months to grow using the hanging drop method at room temperature. Crystals were mounted on nylon

CryoLoops, and precooled in liquid nitrogen prior to data collection. Data were collected on an in house R-AXIS VII++ image plate with a Rigaku generator operated at 50 kV

(current) and 100 mA (voltage), equipped with a dual image detector of size 300 mm, and at the Argonne National Lab Advanced Photon Source (APS) on the Life Sciences

CAT (LS-CAT) beam-line 21-ID-F using a Rayonix MX-300 CCD detector. All measurements were performed at 100 K. Prior to data collection on APS, crystals were soaked in the mother liquor with 10 mM oxalate. Raw data were processed using

XDS/XSCALE, to a resolution of 2.1 Å and 1.45 Å for the pH 4.6 crystals on the R-AXIS

VII++ and APS instruments, respectively [146]. Molecular replacement was carried out using PDB:1J58 as a starting model in phaser [147]. Preliminary refinement is being carried out in phenix-refine in PHENIX 1.7.1, with model building conducted in COOT

[148-150]. Details of the data collection and preliminary refinement values are provided in Appendix B, Tables B-1 and B-2.

124

Preliminary Data

Previous crystallization of Bacillus subtilis OxDC has shown that the enzyme exists in hexameric form. However, the structures previously presented were poised at a pH of 8.0 or higher, outside of the active pH range of the enzyme (4.0 - 6.0).

Independent screening for crystallization conditions identified protein crystals that formed at pH 4.6, well within the active pH range. Initial refinement suggests the enzyme retains the same quaternary structure exhibited in the high pH crystal structures as shown in Figure 6-1.

Figure 6-1. Crystal structure of Bacillus subtilis oxalate decarboxylase hexameric structure with one trimer unit represented in orange and the other in blue at pH 4.6 at 2.1 Å with the manganese centers represented as purple spheres.

The low pH crystal also revealed that acetate binds monodentate to the N-terminal manganese center as shown in Figure 6-2. This is expected, as previous kinetic measurements in the presence of citrate, succinate, and acetate as buffers showed that

125

acetate played an inhibitory role [91]. The interaction of acetate is also logical due to its structural similarity to oxalate. The larger sizes of citrate and succinate would possibly prevent them from entering the active site.

Figure 6-2. N-terminal active site of the pH 4.6 OxDC structure showing manganese (purple sphere) coordinated by His95, His 97, His 140, and Glu101, and acetate.

The N-terminal active site of OxDC is gated by a series of amino acids SENST

161-165. This loop has been shown to exist in two conformations, causing glutamate-

162 to orient towards or away from the manganese center. Preliminary refinement of the structure with acetate bound to the N-terminal metal center revealed a third conformation of the glutamate residue located between the two previous states as shown in Figure 6-3. Glutamate-162 is thought to act as a transient acid/base during catalysis, abstracting a proton in the initial step of catalysis leading to a proton coupled electron transfer. In Figure 6-3, the two previous conformations of glutamate-162 in the closed (blue, PDB: 1UW8) and open conformation (orange, PDB: 1J58) are overlaid on the low pH crystal structure glutamate in magenta. The closed conformation of glutamate clashes with bound acetate molecule, with the open conformation too far

126

away to react. However, the half-open structure identifies a third glutamate conformation that could possibly represent the active conformation.

Figure 6-3. A) N-terminal active site of the pH 4.6 OxDC structure overlaid with the sequence SENST 161-165 from PDB: 1J58 (orange) and 1UW8 (blue) with the glutamate residue represented as sticks for all three structures, and glutamate for the low pH structure in magenta. B) N-terminal active site of the pH 4.6 OxDC structure at 1.45 Å, with glutamate-162 (orange) represented as sticks.

Data collected on APS for the pH 4.6 OxDC crystals, provided a structure with a resolution of 1.45 Å. Preliminary refinement indicated that the structure included two

127

conformations for the loop gating the active site. Prior to data collection the pH 4.6

OxDC crystal was soaked in the mother liquor that also included 10 mM oxalate, with the hope that substrate would co-crystallize, or allow for identification of intermediate catalytic structures. At this point no densities matching that of oxalate or expected bound intermediates have been observed in the vicinity of the metal center. The two conformations resemble that of the previously identified open conformation and the now

‘half-open’ structure (magenta glutamate Figure 6-3 A) identified in the 2.1 Å structure. It can be inferred that structure is an active form of the enzyme, due to the motion in the active site loop previously shown to exist only in one conformation or the other.

Further optimization of the low pH crystallization is currently under way to crystallize the protein in alternative buffers, such as citrate and succinate, to remove the possibility of buffer molecules interacting with either of the metal sites. Co-crystallization of any of the known un-competitive or non-competitive inhibitors at both pH 7.5 or 4.6 would identify where and how these small molecules interact with the enzyme. This would provide insights into possible locations for oxygen binding sites, and any additional binding sites in the protein.

Long Range Electron Transfer Mechanism

Identifying the site of oxygen binding is extremely important in rationalizing future mechanistic schemes for OxDC, especially to understand the bifurcation in the mechanism. EPR spin-trapping experiments shown in Chapter 3 would indicate that oxygen does not bind at the N-terminal site, and due to the structural similarities in the

N and C-terminal metal sites, the C-terminal site presents a reasonable alternative.

However, the distance between the metal centers intra-monomer (26 Å) are too far to facilitate an electron transfer reaction. The inter-monomer distance (21 Å) between

128

neighboring monomer units in the trimer is shorter, but electron transfer is still unlikely.

A tryptophan dimer pair (Figure 1-3) that lies between monomer units could facilitate a long-range electron transfer reaction, shortening the distance for electron hops to 8-10

Å. In order to test the importance of this dimer, site-directed mutants W96F, W274F, and W96F/W274F, provide conserved mutations that should only slightly perturb the dimer stacking. Preliminary refinement of pH 4.6 OxDC crystals indicates that the quaternary structure, identified at high pH is still present at low pH. Primer sequences designed to carry out these mutations have been made (see Table 6-1).

Table 6-1. Forward and reverse primer sequences to generate W96F, W274F, and W96F/W274F OxDC site directed mutants. Primer Sequence W96F FWD Primer 5’ – AAACCGATCAttcCACCGTCAGC – 3’ W96F REV Primer 5’ – CAGGCGGAACGAGCATAT – 3’ W274F FWD Primer 5’ – AGAACTGCACttcCACCCGAATACCC – 3’ W274F REV Primer 5’ – CTCATGGCGCCGGGTTCT – 3’

Primers were designed using the NEBaseChanger online interactive software, and will be carried out using NEB Q5® Site-Directed Mutagenesis Kit.

Chapter 4 provided evidence for the possibility that the metal centers in OxDC can be redox-cycled, and Mn(III) has been proposed as the active oxidation state for the metal centers. EPR experiments conducted at 406 GHz confirm that both sites are susceptible to oxidation, however, does not indicate if this is catalytically relevant. It is expected that one site would be redox-cycled, while the other would remain in one oxidation state.

One methodology for identifying which metal site, if any, is redox-cycled during catalysis is through site-directed spin-labeling. Spin labels are nitroxides that can be site selectively placed on proteins, which can be used to gain information about local and

129

global dynamics, accessibility, and distance measurements. In this case, spin labels can be placed in the vicinity of one or both metal centers (Figure 6-4) and the relaxation rates of the spin labels can be measured at cryogenic temperatures under turnover conditions.

Figure 6-4. Overlay of PDB: 1J58 with MTSL spin label at sites V222 (N-terminal, blue label) and V243 (C-terminal, magenta label).

The sites V222 (N-terminal, blue label), and V243 (C-terminal, magenta label) have been chosen as examples for possible incorporation of the S-(1-oxyl-2,2,5,5- tetramethyl-2,5-dihydro-1H-pyrrol-3-yl)methyl methanesulfonothioate (MTSL) spin label due to their proximity to the individual metal sites [151, 152]. The most probable rotamers of the spin label as shown in Figure 6-4 were found using the open source package MMM for MATLAB™ [153].

Two relaxation pathways, spin-lattice relaxation, T1, and spin-spin relaxation, T2, govern the spin system. The T1 is affiliated with the dissipation of energy to its surroundings through thermal vibrations in the lattice, while the T2 is governed the interaction of local spins through dipolar and exchange interactions [58, 61]. Depending

130

on the metal center and its T1, the relaxation rate of a spin label in its vicinity will be affected due to paramagnetic relaxation enhancement [154, 155]. If one or both of the metal centers are changing oxidation state the relaxation enhancement on the spin label is expected to change. In order to observe if in fact a change occurs, the T1 of the spin label can be measured by an inversion recovery or saturation recovery experiment prior to introduction of substrate. Upon inclusion of substrate in the reaction mixture, the reaction can be quenched at different time points through rapid freezing, and the T1 of the spin label can be measured, and any changes can be identified.

Conclusion

The enzymatic mechanism of Bacillus subtilis oxalate decarboxylase catalyzes the heterolytic cleavage of the facile carbon-carbon bond of oxalic acid via a radical mediated mechanism. Dioxygen is proposed to bind to the N-terminal manganese center to generate Mn(III), the active oxidation state of the metal center. The order of binding of substrate and oxygen is still unclear. Also, no proof for the binding geometry of substrate has yet been provided. The fate of dioxygen appears to be superoxide, which was shown in Chapter 3 to be generated and released from the enzyme during catalysis. This was additionally supported with quantitative experiments following the production of hydrogen peroxide, which is generated by dismutation of superoxide.

Spin-trapping experiments with the T165V OxDC mutant, known to produce much higher concentrations of carbon dioxide anion radical were studied showing that there is increased production of hydrogen peroxide, but still less than that of the wild-type enzyme. If superoxide were solely linked to the release of the carbon dioxide anion radical in solution, which reacts with dioxygen in solution, it is expected that superoxide production in the T165V mutant is higher than for WT enzyme. Furthermore, both

131

radical species are generated, superoxide and the carbon dioxide anion radical, and are also released can be trapped simultaneously. This would further suggest that the radicals are generated at two separate sites, not as previously proposed at solely the N- terminal manganese site.

The source of superoxide generation is likely on one or the other manganese sites in the enzyme. Upon superoxide generation, the metal center would then be oxidized to Mn(III). In order to identify if Mn(III) can be stabilized in OxDC, chemical oxidation of the metal centers was carried out using hexachloroiridate. Through multi- frequency/field, and parallel-mode EPR, Mn(III) was observed and characterized at pH

4.2. At high pH a parallel mode signal was also observed at the same field position of the low pH Mn(III) signal, however, with no discernable fine structure to conclusively identify the source of the signal. More so, a radical species was generated during the high pH oxidation studies, more than likely a tyrosyl radical previously observed under turnover conditions. High field/frequency experiments show that the N- and C-terminal metal centers were oxidized at potentials, with the N-terminal Mn being oxidized at a slightly faster rate.

In an effort to circumvent the large consumption of enzyme during these spectroscopic studies, an immobilization technique has been developed utilizing the encoded His6-tag used for purification. This methodology was applied to OxDC to study the pH dependence of the manganese centers and followed with parallel-mode X-band

EPR. Mn(III) was shown to exhibit a pH dependent behavior, appearing only in the active pH range of the enzyme, 4.0 – 6.0. The pH dependent behavior of Mn(III) would

132

suggest that the activity may be linked to its presence versus the charge-state of the substrate.

These series of experiments suggest that either metal center is a suitable binding site for dioxygen. If oxygen binds at the C-terminal manganese site, the mechanism will then dependent on a long-range electron transfer. The identification of the oxygen binding site, and kinetic measurements of the proposed site-directed mutants will assist in determining if both sites are directly involved in catalysis, and if so what their individual roles are. These studies may also shed light on how the enzyme controls whether it acts as an oxidase or decarboxylase, or if the observation of oxidase activity is due to the enzyme’s inability to completely control its radical intermediates.

133

APPENDIX A EXPRESSION AND PURIFICATION OF Bacillus subtilis OXALATE DECARBOXYLASE

A purified pET31a vector containing either wild type or mutant DNA was transformed into competent BL21 (DE3) cells from Novagen and grown on a Luria broth agarose (LBA) gel overnight at 37°C. From the plate, a single colony was selected and grown overnight in 50 mL of low-salt Luria Broth (LB) media containing 5 mg ampicillin

(Fisher Scientific) at 37°C. Four 2 L flasks containing 475 mL LB media and 500 L of

100 mg mL-1 ampicillin were then inoculated with 6 mL of culture. The flasks were shaken at a constant temperature of 37°C until the solutions reached an optical density of 0.5 at 600 nm. The bacteria were then heat shocked at 42°C for fifteen minutes, followed by the addition of isopropyl β-D-1-thiogalactopyranoside (IPTG) and MnCl2

(final concentrations of 1 mM and 5 mM, respectively) to each flask and were grown for

4 hours at 37°C. The cells were harvested and kept overnight at -80°C.

Cells were then thawed and suspended in 75 mL lysis buffer (50 mM Tris-Cl, 0.5

M NaCl, 10 mM imidazole, 10 M MnCl2 at pH 7.5). The resuspended cellular solution was then sonicated (Sonic Dismembrator Model 705, Fisher Scientific) for five 20- second cycles with 30-second intervals between cycles. Cellular debris was removed by centrifugation (12000g 20 min 4°C). Lysate was then collected and loaded on a Ni-

NTA (Qiagen, Hilden, Germany) column prepared by equilibration using a wash buffer

(50 mM potassium phosphate 0.5 M NaCl, 20 mM imidazole at pH 8.5) prior to loading of lysate. Protein was eluted using an elution buffer containing 50 mM potassium phosphate, 0.5 M NaCl, 250 mM imidazole at pH 8.5.

134

The fractions containing enzyme, identified by gel electrophoresis, were pooled and placed in a snakeskin dialysis unit and a series of dialyses were carried out to transfer the protein into storage buffer (50 mM Tris-Cl, 0.5 M NaCl pH 8.5), as well as to remove excess imidazole. Residual imidazole in the enzyme preparation leads to aggregation in subsequent concentration steps. To remove dissolved metals from the preparation, Chelex 100 resin (Bio-Rad, Hercules CA) was added to the enzyme after the serial dialysis steps. The solution was shaken for approximately one hour following removal of the resin. The enzyme solution was then concentrated using Amicon

Centriprep YM-30 (EMD Millipore, Billerica, MA). Concentrated enzyme samples

(approximately 40 mg/mL) were stored as 200 L aliquots in Eppendorf tubes at -80°C until used for experiments. Enzyme activity was determined using an endpoint-stopped assay, coupling the formation of formate to the reduction of nicotinamide adenine dinucleotide (NAD+) [39, 50]. An additional assay was performed to confirm activity prior to experimentation, using the oxidation of o-phenyldiamine to 2,3-diaminophenazine, which gives a pale yellow color indicative of active enzyme [39]. Metals analysis was also carried out with ICP-MS at the Center for Applied Isotope Studies at the University of Georgia. These results typically indicated 1.4 to 1.8 Mn per subunit.

135

APPENDIX B X-RAY DIFFRACTION DATA COLLECTION, PROCESSING AND REFINEMENT PARAMETERS

Table B-1. Data collection and refinement statistics on (LS-CAT) beam-line 21-ID-F for pH 4.6 OxDC crystal.

OxDC pH 4.6 APS 0.978700 Wavelength (Å) 39.16 - 1.451 (1.502 - 1.451) Resolution range (Å) I 1 2 1 Space group 79.574 156.657 98.266 90 105.65 90 Unit cell (Å) 203407 (20135) Unique reflections 99.81 (99.00) Completeness (%) 8.85 (2.40) Mean I/sigma(I) 11.30 Wilson B-factor 0.1198 (0.1731) R-work 0.1554 (0.2359) R-free 10873 Number of non-hydrogen atoms 9611 macromolecules 30 ligands 1232 water 1136 Protein residues 0.015 RMS(bonds) 1.52 RMS(angles) 98 Ramachandran favored (%) 0.17 Ramachandran outliers (%) 6.32 Clashscore 16.80 Average B-factor 14.90 macromolecules 20.70 ligands 31.20 solvent Statistics for the highest-resolution shell are shown in parentheses.

136

Table B-2. Data collection and refinement statistics on R-AXIS VII++ for pH 4.6 OxDC crystal.

OxDC pH 4.6-R-AXIS VII++ 1.541800 Wavelength (Ã…) 28.96 - 2.101 (2.176 - 2.101) Resolution range (Ã…) I 1 2 1 Space group 79.552 156.452 98.204 90 105.75 90 Unit cell 66915 (6585) Unique reflections 99.71 (98.76) Completeness (%) 12.73 (4.59) Mean I/sigma(I) 18.40 Wilson B-factor 0.1507 (0.1653) R-work 0.1951 (0.2290) R-free 9691 Number of non-hydrogen atoms 8966 macromolecules 19 ligands 706 water 1132 Protein residues 0.009 RMS(bonds) 1.16 RMS(angles) 98 Ramachandran favored (%) 0 Ramachandran outliers (%) 3.67 Clashscore 19.60 Average B-factor 18.90 macromolecules 30.40 ligands 29.30 solvent Statistics for the highest-resolution shell are shown in parentheses.

137

LIST OF REFERENCES

[1] Miller, C.; Kennington, L.; Cooney, R.; Kohjimoto, Y.; Cao, L. C.; Honeyman, T.; Pullman, J.; Jonassen, J.; Scheid, C. Oxalate Toxicity in Renal Epithelial Cells: Characteristics of Apoptosis and Necrosis. Toxicol. Appl. Pharm. 162:132-141; 2000.

[2] Giachetti, E.; Pinzauti, G.; Bonaccorsi, R.; Vanni, P. Isocitrate lyase from Pinus pinea. Eur. J. Biochem. 172:85-91; 1988.

[3] Pinzauti, G.; Giachetti, E.; Camici, G.; Manao, G.; Cappugi, G.; Vanni, P. An isocitrate lyase of higher plants: Analysis and comparison of some molecular properties. Arch. Biochem. Biophys. 244:85-93; 1986.

[4] Richardson, K. E.; Tolbert, N. E. Oxidation of Glyoxylic Acid to Oxalic Acid by Glycolic Acid Oxidase. J. Biol. Chem. 236:1280-1284; 1961.

[5] Seal, S. N.; Sen, S. P. The photosynthetic production of oxalic acid in Oxalis corniculata. Plant Cell Physiol. 11:119-128; 1970.

[6] Frassetto, L., Kohlstadt,I. Treatment and Prevention of Kidney Stones: An Update. Am. Fam. Physician. 84:1234-1242; 2011.

[7] Elder, T. D.; Wyngaarden, J. B. The Biosysnthesis and Turnover of Oxalate in Normal and Hyperoxaluric Subjects. J. Clin. Invest. 39:1337-1344; 1960.

[8] Hodgkinson, A. Oxalic acid in biology and medicine. Academic Press; 1977.

[9] Noonan, S. C. S. G. P. Oxalate content of foods and its effect on humans. Asia Pac. J. Clin. Nutr. 8:64-74; 1999.

[10] Rahman, M. M.; Abdullah, R. B.; Wan Khadijah, W. E. A review of oxalate poisoning in domestic animals: tolerance and performance aspects. J. Anim. Physiol. Anim. Nutr. 97:605-614; 2013.

[11] Sidhu, H., Schmidt, M. E., Cornelius, J. G., Thamilselvan, S., Khan, S. R., Hesse, A., Peck, A. B. Direct correlation between hyperoxaluria/oxalate stone disease and the absence of the gastrointestinal tract-dwelling bacterium Oxalobacter formigenes: possible prevention by gut recolonization or enzyme replacement therapy. J. Am. Soc. Nephrol. Suppl. 14:S334-340; 1999.

[12] Sidhu, H.; Allison, M.; Peck, A. B. Identification and Classification of Oxalobacter formigenes Strains by Using Oligonucleotide Probes and Primers. J. Clin. Microbiol. 35:350-353; 1997.

[13] Libert, B.; Franceschi, V. R. Oxalate in crop plants. J. Agric. Food. Chem. 35:926-938; 1987.

138

[14] Chakraborty, N.; Ghosh, R.; Ghosh, S.; Narula, K.; Tayal, R.; Datta, A.; Chakraborty, S. Reduction of Oxalate Levels in Tomato Fruit and Consequent Metabolic Remodeling Following Overexpression of a Fungal Oxalate Decarboxylase. Plant Physiol. 162:364-378; 2013.

[15] Dias, B. B. A.; Cunha, W. G.; Morais, L. S.; Vianna, G. R.; Rech, E. L.; de Capdeville, G.; Aragão, F. J. L. Expression of an oxalate decarboxylase gene from Flammulina sp. in transgenic lettuce (Lactuca sativa) plants and resistance to Sclerotinia sclerotiorum. Plant Physiol. 55:187-193; 2006.

[16] Hart, P. W., Rudie, A.W. Mineral scale management, Part I. Case studies. TAPPI Journal 5:22-27; 2006.

[17] Mäkelä, M. R. H., K.; Lundell, T. K. Oxalate decarboxylase: biotechnological update and prevalence of the enzyme in filamentous fungi. Appl. Microbiol. Biotechnol. 87:801; 2010.

[18] Rudie, A. W., Hart, P.W. Mineral Scale Management Part II. Fundamental Chemistry. TAPPI Journal 5:17-23; 2006.

[19] Winestrand, S.; Gandla, M. L.; Hong, F.; Chen, Q. Z.; Jönsson, L. J. Oxalate decarboxylase of Trametes versicolor: biochemical characterization and performance in bleaching filtrates from the pulp and paper industry. J. Chem. Technol. Biotechnol. 87:1600-1606; 2012.

[20] Dutton, M. V.; Kathiara, M.; Gallagher, I. M.; Evans, C. S. Purification and characterization of oxalate decarboxylase from Coriolus versicolor. FEMS Microbiol. Lett. 116:321-325; 1994.

[21] Emiliani, E.; Bekes, P. Enzymatic Oxalate Decarboxylation in Aspergillus niger. Arch. Biochem. Biophys. 105:488-493; 1964.

[22] Emiliani, E.; Riera, B. Enzymatic Oxalate Decarboxylation in Aspergillus Niger. 2. Hydrpgen Peroxide Fomration and Other Characteristics of Oxalate Decarboxylase. Biochim. Biophys. Acta 167:414; 1968.

[23] Kathiara, M.; Wood, D. A.; Evans, C. S. Detection and partial characterization of oxalate decarboxylase from Agaricus bisporus. Mycol. Res. 104:345-350; 2000.

[24] Lillehoj, E. B.; Smith, F. G. An Oxalic Acid Decarboxylase of Myrothecium Verrucaria. Arch. Biochem. Biophys. 109:216; 1965.

[25] Magro, P.; Marciano, P.; Di Lenna, P. Enzymatic oxalate decarboxylation in isolates of Sclerotinia sclerotiorum. FEMS MIicrobiol. Lett. 49:49-52; 1988.

[26] Shimazono, H. Oxalic Acid Decarboxylase, A new Enyme from the Mycelium of Wood Destroying Fungi. J. Biochem. 42:321-340; 1955.

139

[27] Tanner, A.; Bornemann, S. Bacillus subtilis YvrK is an acid-induced oxalate decarboxylase. J. Bacteriol. 182:5271-5273; 2000.

[28] Anand, R.; Dorrestein, P.; Kinsland, C.; Begley, T.; Ealick, S. Structure of oxalate decarboxylase from Bacillus subtilis at 1.75 angstrom resolution. Biochemistry 41:7659- 7669; 2002.

[29] Just, V. J.; Burrell, M. R.; Bowater, L.; McRobbie, I.; Stevenson, C. E. M.; Lawson, D. M.; Bornemann, S. The identity of the active site of oxalate decarboxylase and the importance of the stability of active-site lid conformations. Biochem. J. 407:397- 406; 2007.

[30] Just, V. J.; Stevenson, C. E. M.; Bowater, L.; Tanner, A.; Lawson, D. M.; Bornemann, S. A Closed Conformation of Bacillus subtilis Oxalate Decarboxylase OxdC Provides Evidence for the True Identity of the Active Site. J. Biol. Chem. 279:19867- 19874; 2004.

[31] Dunwell, J. Cupins: A new superfamily of functionally diverse proteins that include germins and plant storage proteins. Biontechnol. Genet. Eng. 15:1-32; 1998.

[32] Dunwell, J.; Purvis, A.; Khuri, S. Cupins: the most functionally diverse protein superfamily? Phytochemistry 65:7-17; 2004.

[33] Khuri, S.; Bakker, F. T.; Dunwell, J. M. Phylogeny, Function, and Evolution of the Cupins, a Structurally Conserved, Functionally Diverse Superfamily of Proteins. Mol. Biol. Evol. 18:593-605; 2001.

[34] Angerhofer, A.; Moomaw, E. W.; García-Rubio, I.; Ozarowski, A.; Krzystek, J.; Weber, R. T.; Richards, N. G. J. Multifrequency EPR Studies on the Mn(II) Centers of Oxalate Decarboxylase. J. Phys. Chem. B 111:5043-5046; 2007.

[35] Chang, C.; Svedruzic, D.; Ozarowski, A.; Walker, L.; Yeagle, G.; Britt, R.; Angerhofer, A.; Richards, N. EPR spectroscopic characterization of the manganese center and a free radical in the oxalate decarboxylase reaction - Identification of a tyrosyl radical during turnover. J. Biol. Chem. 279:52840-52849; 2004.

[36] Tabares, L.; Gatjens, J.; Hureau, C.; Burrell, M.; Bowater, L.; Pecoraro, V.; Bornemann, S.; Un, S. pH-Dependent Structures of the Manganese Binding Sites in Oxalate Decarboxylase as Revealed by High-Field Electron Paramagnetic Resonance. J. Phys. Chem. B 113:9016-9025; 2009.

[37] Wolfenden, R.; Lewis, C. A.; Yuan, Y. Kinetic Challenges Facing Oxalate, Malonate, Acetoacetate, and Oxaloacetate Decarboxylases. J. Am. Chem. Soc. 133:5683-5685; 2011.

[38] Svedružić, D. J., S.; Toyota, C. G.; Reinhardt, L. A.; Ricagno, S.; Linqvist, Y.; Richards, N. G. J. The enzymes of oxalate metabolism: unexpected structures and mechanisms. Arch. Biochem. Biophys. 433:176-192; 2005.

140

[39] Tanner, A.; Bowater, L.; Fairhurst, S. A.; Bornemann, S. Oxalate decarboxylase requires manganese and dioxygen for activity - Overexpression and characterization of Bacillus subtilis YvrK and YoaN. J. Biol. Chem. 276:43627; 2001.

[40] Reinhardt, L. A.; Svedružić, D.; Chang, C. H.; Cleland, W. W.; Richards, N. G. J. Heavy Atom Isotope Effects on the Reaction Catalyzed by the Oxalate Decarboxylase from Bacillus subtilis. J. Am. Chem. Soc. 125:1244-1252; 2003.

[41] Twahir, U. T.; Stedwell, C. N.; Lee, C. T.; Richards, N. G. J.; Polfer, N. C.; Angerhofer, A. Observation of superoxide production during catalysis of Bacillus subtilis oxalate decarboxylase at pH 4. Free Rad. Biol. Med. 80:59-66; 2015.

[42] Imaram, W.; Saylor, B.; Centonze, C.; Richards, N.; Angerhofer, A. EPR spin trapping of an oxalate-derived free radical in the oxalate decarboxylase reaction. Free Rad. Biol. Med. 50:1009-1015; 2011.

[43] Bielski, B. H. J. Reevaluation of the Spectral and Kinetic Properties of HO2 and O2- Free Radicals. Photochem. Photobiol. 28:645-649; 1978.

[44] Svedružić, D. L., Y.; Reinhardt, L. A.; Wroclawska, E.; Cleland, W. W.; Richards, N. G. J. . Investigating the roles of putative active site residues in the oxalate decarboxylase from Bacillus subtilis. Arch. Biochem. Biophys. 464:36-47; 2007.

[45] Begley, T. P.; Ealick, S. E. Enzymatic reactions involving novel mechanisms of carbanion stabilization. Curr. Opin. Chem. Biol. 8:508-515; 2004.

[46] Molt, R. W.; Lecher, A. M.; Clark, T.; Bartlett, R. J.; Richards, N. G. J. Facile Csp2–Csp2 Bond Cleavage in Oxalic Acid-Derived Radicals. J. Am. Chem. Soc. 137:3248-3252; 2015.

[47] Burrell, M.; Just, V.; Bowater, L.; Fairhurst, S.; Requena, L.; Lawson, D.; Bornemann, S. Oxalate decarboxylase and oxalate oxidase activities can be interchanged with a specificity switch of up to 282 000 by mutating an active site lid. Biochemistry 46:12327-12336; 2007.

[48] Dunwell, J. M.; Gane, P. J. Microbial Relatives of Seed Storage Proteins: Conservation of Motifs in a Functionally Diverse Superfamily of Enzymes. J. Mol. Evol. 46:147-154; 1998.

[49] Woo, E.-J.; Dunwell, J. M.; Goodenough, P. W.; Marvier, A. C.; Pickersgill, R. W. Germin is a manganese containing homohexamer with oxalate oxidase and superoxide dismutase activities. Nat. Struct. Mol. Biol. 7:1036-1040; 2000.

[50] Moomaw, E.; Angerhofer, A.; Moussatche, P.; Ozarowski, A.; García-Rubio, I.; Richards, N. Metal Dependence of Oxalate Decarboxylase Activity. Biochemistry 48:6116-6125; 2009.

141

[51] Brettel, K.; Byrdin, M. Reaction mechanisms of DNA photolyase. Curr. Opin. Struct. Biol. 20:693-701; 2010.

[52] Dempsey, J. L.; Winkler, J. R.; Gray, H. B. Proton-Coupled Electron Flow in Protein Redox Machines. Chem. Rev. 110:7024-7039; 2010.

[53] Onuchic, J. N.; Beratan, D. N.; Winkler, J. R.; Gray, H. B. Pathway Analysis of Protein Electron-Transfer Reactions. Annu. Rev. Biophys. Biomol. Struct. 21:349-377; 1992.

[54] Winkler, J. R. J. Electron tunneling pathways in proteins. Curr. Opin. Chem. Biol. 4:192-198; 2000.

[55] Dunwell, J.; Culham, A.; Carter, C.; Sosa-Aguirre, C.; Goodenough, P. Evolution of functional diversity in the cupin superfamily. Trends. Biochem. Sci. 26:740-746; 2001.

[56] Jonsson, S.; Svedruzic, D.; Wroclawska, E.; Chang, C.; Richards, N. Structure and mechanism of enzymes mediating oxalate metabolism. Biochemistry 42:8617- 8617; 2003.

[57] Abragam, A.; Pryce, M.-H. L. Theory of the nuclear hyperfine structure of paramagnetic resonance spectra in crystals. Proc. Roy. Soc. A: The Royal Society; 1951: 135-153.

[58] Schweiger, A.; Jeschke, G. Principles of pulse electron paramagnetic resonance. Oxford University Press; 2001.

[59] Atherton, N. Principles of Electron Paramagnetic Resonance. Ellis Horwood Limited, Chichester, England; 1993.

[60] McGarvey, B. R. The isotropic hyperfine interaction. J. Phys. Chem. 71:51-66; 1967.

[61] Abragam, A.; Bleaney, B. Electron Paramagnetic Resonance of Transition Ions. Dover Publishing; 1986.

[62] Eaton, S. S.; Eaton, G. R.; Berliner, L. J. Biomedical EPR-part A: Free radicals, metals, medicine and physiology. Springer Science & Business Media; 2004.

[63] Rosen, G. M., Britigan, Bradley E., Halpern, Howard J., Pou, Sovitj Free Radicals: Biology And Detection By Spin Trapping. 198 Madison Avenue, New York, New York 10016: Oxford University Press, Inc. ; 1999.

[64] Zhao, H.; Joseph, J.; Zhang, H.; Karoui, H.; Kalyanaraman, B. Synthesis and Biochemical Applications of a Solid Cyclic Nitrone Spin Trap: a Relatively Superior Trap for Detecting Superoxide Anions and Glutathiyl Radicals. Free Rad. Biol. Med. 31:599- 606; 2001.

142

[65] Savitsky, A.; Möbius, K. High-field EPR. Photosynth. Res. 102:311-333; 2009.

[66] Telser, J.; Krzystek, J.; Ozarowski, A. High-frequency and high-field electron paramagnetic resonance (HFEPR): a new spectroscopic tool for bioinorganic chemistry. J. Biol. Inorg. Chem. 19:297-318; 2014.

[67] Dorlet, P.; Seibold, S. A.; Babcock, G. T.; Gerfen, G. J.; Smith, W. L.; Tsai, A.-l.; Un, S. High-Field EPR Study of Tyrosyl Radicals in Prostaglandin H2 Synthase-1†. Biochemistry 41:6107-6114; 2002.

[68] Gerfen, G. J.; Bellew, B. F.; Un, S.; Bollinger, J. M.; Stubbe, J.; Griffin, R. G.; Singel, D. J. High-frequency (139.5 GHz) EPR spectroscopy of the tyrosyl radical in Escherichia coli ribonucleotide reductase. J. Am. Chem. Soc. 115:6420-6421; 1993.

[69] Hendrich, M. P.; Debrunner, P. G. Integer-spin electron paramagnetic resonance of iron proteins. Biophys. J. 56:489-506; 1989.

[70] Emiliani, E.; Bekes, P. Enzymatic oxalate decarboxylation in Aspergillus niger. Arch. Biochem. Biophys. 105:488-493; 1964.

[71] Escutia, M. R.; Bowater, L.; Edwards, A.; Bottrill, A. R.; Burrell, M. R.; Polanco, R.; Vicuña, R.; Bornemann, S. Cloning and Sequencing of Two Ceriporiopsis subvermispora Bicupin Oxalate Oxidase Allelic Isoforms: Implications for the Reaction Specificity of Oxalate Oxidases and Decarboxylases. Appl. Environ. Microbiol. 71:3608- 3616; 2005.

[72] Whittaker, M. M.; Pan, H.-Y.; Yukl, E. T.; Whittaker, J. W. Burst Kinetics and Redox Transformations of the Active Site Manganese Ion in Oxalate Oxidase: Implications for the Catalytic Mechanism. J. Biol. Chem. 282:7011-7023; 2007.

[73] Bielski, B. H. J.; Allen, A. O. Mechanism of the Disproportionation of Superoxide Radicals. J. Phys. Chem. 81:1048-1050; 1977.

[74] Finkelstein, E.; Rosen, G. M.; Rauckman, E. J. Spin Trapping. Kinetics of the Reaction of Superoxide and Hydroxyl Radicals with Nitrones. J. Am. Chem. Soc. 102:4994-4999; 1980.

[75] Buettner, G. R.; Oberley, L. W. Considerations in the Spin Trapping of Superoxide and Hydroxyl Radical in Aqueous Systems Using 5,5-Dimethyl-1-pyrroline- 1-oxide. Biochem. Biophys. Res. Commun. 83:69-74; 1978.

[76] Villamena, F. A., Zweier, J.L. Detection of Reactive Oxygen and Nitrogen Species by EPR Spin Trapping. Antioxid. Redox Signaling 6:619-629; 2004.

[77] Buettner, G. R.; Mason, R. P. [9] Spin-trapping methods for detecting superoxide and hydroxyl free radicals in vitro and in vivo. In: Lester Packer, A. N. G., ed. Methods in Enzymology: Academic Press; 1990: 127-133.

143

[78] Khan, N.; Wilmot, C. M.; Rosen, G. M.; Demidenko, E.; Sun, J.; Joseph, J.; Hara, J.; Kalyanaraman, B.; Swartz, H. M. Spin traps: in vitro toxicity and stability of radical adducts. Free Rad. Biol. Med. 34:1473-1481; 2003.

[79] Bielski, B. H. J.; Cabelli, D. E.; Arudi, R. L.; Ross, A. B. Reactivity of HO2/O2- Radicals in Aqueous Solution. J. Phys. Chem. Ref. Data 14:1041-1100; 1985.

[80] Iwahashi, H.; Parker, C. E.; Mason, R. P.; Tomer, K. B. Combined Liquid Chromatography/Electron Paramagnetic Resonance Spectrometry/Electrospray Ionization Mass Spectrometry for Radical Identification. Anal. Chem. 64:2244-2252; 1992.

[81] Yue Qian, S.; Kadiiska, M. B.; Guo, Q.; Mason, R. P. A novel protocol to identify and quantify all spin trapped free radicals from in vitro/in vivo interaction of HO and DMSO: LC/ESR, LC/MS, and dual spin trapping combinations. Free Rad. Biol. Med. 38:125-135; 2005.

[82] Domingues, P.; Domingues, M. R. M.; Amado, F. M. L.; Ferrer-Correia, A. J. Detection and characterization of hydroxyl radical adducts by mass spectrometry. J. Am. Soc. Mass. Spectrom. 12:1214-1219; 2001.

[83] Imaram, W.; Gersch, C.; Kim, K. M.; Johnson, R. J.; Henderson, G. N.; Angerhofer, A. Radicals in the reaction between peroxynitrite and uric acid identified by electron spin resonance spectroscopy and liquid chromatography mass spectrometry. Free Rad. Biol. Med. 49:275-281; 2010.

[84] Guo, Q.; Qian, S. Y.; Mason, R. P. Separation and identification of DMPO adducts of oxygen-centered radicals formed from organic hydroperoxides by HPLC- ESR, ESI-MS and MS/MS. J. Am. Soc. Mass Spectrom. 14:862-871; 2003.

[85] Huang, C.-H.; Shan, G.-Q.; Mao, L.; Kalyanaraman, B.; Qin, H.; Ren, F.-R.; Zhu, B.-Z. The first purification and unequivocal characterization of the radical form of the carbon-centered quinone ketoxy radical adduct. Chem. Commun. 49:6436-6438; 2013.

[86] Wolfbeis, O. S.; Dürkop, A.; Wu, M.; Lin, Z. A Europium‐Ion‐Based Luminescent Sensing Probe for Hydrogen Peroxide. Angew. Chem. Int. Ed. Engl. 41:4495-4498; 2002.

[87] Vina, J.; Tsez, G. Glutathione Metabolism and Psychological Functions. CRC Press; 1986.

[88] Keszler, A.; Kalyanaraman, B.; Hogg, N. Comparative investigation of superoxide trapping by cyclic nitrone spin traps: the use of singular value decomposition and multiple linear regression analysis. Free Rad. Biol. Med. 35:1149-1157; 2003.

[89] Allouch, A.; Lauricella, R. P.; Tuccio, B. N. Effect of pH on superoxide/hydroperoxyl radical trapping by nitrones: an EPR/kinetic study. Mol. Phys. 105:2017-2024; 2007.

144

[90] Villamena, F. A.; Merle, J. K.; Hadad, C. M.; Zweier, J. L. Superoxide Radical Anion Adduct of 5,5-Dimethyl-1-pyrroline N-Oxide (DMPO). 1. The Thermodynamics of Formation and Its Acidity. J. Phys. Chem. A 109:6083-6088; 2005.

[91] Molina, L. M. G., T. J.; Twahir, U. T.; Moomaw, E. W.; and Angerhofer, A. . Kinetic Studies of Oxalate Decarboxylase and Oxalate Oxidase using a Luminescent Oxygen Sensor. J. Biochem. Tech. 5:826-831; 2014.

[92] Opaleye, O.; Rose, R.-S.; Whittaker, M. M.; Woo, E.-J.; Whittaker, J. W.; Pickersgill, R. W. Structural and Spectroscopic Studies Shed Light on the Mechanism of Oxalate Oxidase. J. Biol. Chem. 281:6428-6433; 2006.

[93] Tanner, A.; Bowater, L.; Fairhurst, S.; Bornemann, S. Oxalate decarboxylase requires manganese and dioxygen for activity - Overexpression and characterization of Bacillus subtilis YvrK and YoaN. J. Biol. Chem. 276:43627-43634; 2001.

[94] Campbell, K. A.; Yikilmaz, E.; Grant, C. V.; Gregor, W.; Miller, A.-F.; Britt, R. D. Parallel Polarization EPR Characterization of the Mn(III) Center of Oxidized Manganese Superoxide Dismutase. J. Am. Chem. Soc. 121:4714-4715; 1999.

[95] Emerson, J.; Wagner, M.; Reynolds, M.; Que, L., Jr.; Sadowsky, M.; Wackett, L. The role of histidine 200 in MndD, the Mn(II)-dependent 3,4-dihydroxyphenylacetate 2,3-dioxygenase from Arthrobacter globiformis CM-2, a site-directed mutagenesis study. J. Biol. Inorg. Chem. 10:751-760; 2005.

[96] Jackson, T. A.; Yikilmaz, E.; Miller, A.-F.; Brunold, T. C. Spectroscopic and Computational Study of a Non-Heme Iron {Fe−NO}7 System: Exploring the Geometric and Electronic Structures of the Nitrosyl Adduct of Iron Superoxide Dismutase. J. Am. Chem. Soc. 125:8348-8363; 2003.

[97] Sundaramoorthy, M.; Youngs, H. L.; Gold, M. H.; Poulos, T. L. High-Resolution Crystal Structure of Manganese Peroxidase: Substrate and Inhibitor Complexes. Biochemistry 44:6463-6470; 2005.

[98] Van Aken, B.; Agathos, S. Implication of manganese (III), oxalate, and oxygen in the degradation of nitroaromatic compounds by manganese peroxidase (MnP). Appl. Microbiol. Biotechnol. 58:345-351; 2002.

[99] Whiting, A. K.; Boldt, Y. R.; Hendrich, M. P.; Wackett, L. P.; Que, L. Manganese(II)-Dependent Extradiol-Cleaving Catechol Dioxygenase from Arthrobacter globiformis CM-2. Biochemistry 35:160-170; 1996.

[100] Whittaker, M. M. W., J. W. Characterization of recombinant barley oxalate oxidase expressed by Pichia pastoris. J. Biol. Inorg. Chem. 7:136-145; 2002.

145

[101] Moomaw, E. W.; Hoffer, E.; Moussatche, P.; Salerno, J. C.; Grant, M.; Immelman, B.; Uberto, R.; Ozarowski, A.; Angerhofer, A. Kinetic and Spectroscopic Studies of Bicupin Oxalate Oxidase and Putative Active Site Mutants. PLoS ONE 8:e57933; 2013.

[102] Moussatche, P.; Angerhofer, A.; Imaram, W.; Hoffer, E.; Uberto, K.; Brooks, C.; Bruce, C.; Sledge, D.; Richards, N. G. J.; Moomaw, E. W. Characterization of Ceriporiopsis subvermispora bicupin oxalate oxidase expressed in Pichia pastoris. Arch. Biochem. Biophys. 509:100-107; 2011.

[103] Borowski, T.; Bassan, A.; Richards, N. G. J.; Siegbahn, P. E. M. Catalytic Reaction Mechanism of Oxalate Oxidase (Germin). A Hybrid DFT Study. J. Chem. Theory Comput. 1:686-693; 2005.

[104] Weil, J. A.; Bolton, J. R. Electron Paramagnetic Resonance Spectroscopy: Elementary Theory and Applications, Second Edition. Wiley-Interscience. ; 2007.

[105] Stoll, S.; Schweiger, A. EasySpin, a comprehensive software package for spectral simulation and analysis in EPR. J. Magn. Res. 178:42-55; 2006.

[106] Campomanes, P.; Kellett, W. F.; Easthon, L. M.; Ozarowski, A.; Allen, K. N.; Angerhofer, A.; Rothlisberger, U.; Richards, N. G. J. Assigning the EPR Fine Structure Parameters of the Mn(II) Centers in Bacillus subtilis Oxalate Decarboxylase by Site- Directed Mutagenesis and DFT/MM Calculations. J. Am. Chem. Soc. 136:2313-2323; 2014.

[107] Lüthen, H.; Böttger, M. Hexachloroiridate IV as an Electron Acceptor for a Plasmalemma Redox System in Maize Roots. Plant Physiol. 86:1044-1047; 1988.

[108] Britt, R. D.; Peloquin, J. M.; Campbell, K. A. Pulsed and Parallel-Polarization EPR Characterization of the Photosystem II Oxygen-Evolving Complex. Annu. Rev. Biophys. Biomol. Struct. 29:463-495; 2000.

[109] Campbell, K. A.; Force, D. A.; Nixon, P. J.; Dole, F.; Diner, B. A.; Britt, R. D. Dual-Mode EPR Detects the Initial Intermediate in Photoassembly of the Photosystem II Mn Cluster: The Influence of Amino Acid Residue 170 of the D1 Polypeptide on Mn Coordination. J. Am. Chem. Soc. 122:3754-3761; 2000.

[110] Campbell, K. A.; Lashley, M. R.; Wyatt, J. K.; Nantz, M. H.; Britt, R. D. Dual- Mode EPR Study of Mn(III) Salen and the Mn(III) Salen-Catalyzed Epoxidation of cis-β- Methylstyrene. J. Am. Chem. Soc. 123:5710-5719; 2001.

[111] El Ghachtouli, S.; Vincent Ching, H. Y.; Lassalle-Kaiser, B.; Guillot, R.; Leto, D. F.; Chattopadhyay, S.; Jackson, T. A.; Dorlet, P.; Anxolabehere-Mallart, E. Electrochemical formation of MnIII-peroxo complexes supported by pentadentate amino pyridine and imidazole ligands. Chem. Commun. 49:5696-5698; 2013.

146

[112] Groni, S.; Blain, G.; Guillot, R.; Policar, C.; Anxolabéhère-Mallart, E. Reactivity of MnII with Superoxide. Evidence for a [MnIIIOO]+ Unit by Low-Temperature Spectroscopies. Inorg. Chem. 46:1951-1953; 2007.

[113] Gupta, R.; Taguchi, T.; Borovik, A. S.; Hendrich, M. P. Characterization of Monomeric MnII/III/IV–Hydroxo Complexes from X- and Q-Band Dual Mode Electron Paramagnetic Resonance (EPR) Spectroscopy. Inorg. Chem. 52:12568-12575; 2013.

[114] Hayden, J.; Farquhar, E.; Que, L.; Lipscomb, J.; Hendrich, M. NO binding to Mn- substituted homoprotocatechuate 2,3-dioxygenase: relationship to O2 reactivity. J. Biol. Inorg. Chem. 18:717-728; 2013.

[115] Lippard, S. J.; Berg, J. M. Principles of bioinorganic chemistry. New York, USA; 1994.

[116] Ilan, Y. A.; Czapski, G.; Meisel, D. The one-electron transfer redox potentials of free radicals. I. The oxygen/superoxide system. Biochim. Biophys. Acta 430:209-224; 1976.

[117] Jones, R. D.; Summerville, D. A.; Basolo, F. Synthetic oxygen carriers related to biological systems. Chem. Rev. 79:139-179; 1979.

[118] Kálmán, L.; Williams, J.; Allen, J. Energetics for oxidation of a bound manganese cofactor in modified bacterial reaction centers. Biochemistry 50:3310-3320; 2011.

[119] Miller, A.-F. Superoxide dismutases: ancient enzymes and new insights. FEBS Lett. 586:585-595; 2012.

[120] Schwarzenbacher, R.; von Delft, F.; Jaroszewski, L.; Abdubek, P.; Ambing, E.; Biorac, T.; Brinen, L. S.; Canaves, J. M.; Cambell, J.; Chiu, H. J. Crystal structure of a putative oxalate decarboxylase (TM1287) from Thermotoga maritima at 1.95 Å resolution. Proteins 56:392-395; 2004.

[121] Moral, M.; Tu, C.; Imaram, W.; Angerhofer, A.; Silverman, D.; Richards, N. Nitric oxide reversibly inhibits Bacillus subtilis oxalate decarboxylase. Chem. Commun. 47:3111-3113; 2011.

[122] Chang, C. H.; Richards, N. G. J. Intrinsic carbon-carbon bond reactivity at the manganese center of oxalate decarboxylase from density functional theory. J. Chem. Theory Comput. 1:994; 2005.

[123] Hochuli, E.; Bannwarth, W.; Döbeli, H.; Gentz, R.; Stüber, D. Genetic approach to facilitate purification of recombinant proteins with a novel metal chelate adsorbent. Nat. Biotechnol. 6.11:1321-1325; 1988.

[124] Cheung, R. C. F.; Wong, J. H.; Ng, T. B. Immobilized metal ion affinity chromatography: a review on its applications. Appl. Microbiol. Biotechnol. 96:1411- 1420; 2012.

147

[125] Gaberc-Porekar, V.; Menart, V. Perspectives of immobilized-metal affinity chromatography. J. Biochem. Biophys. Methods. 49:335-360; 2001.

[126] Chikere, A. C.; Galunsky, B.; Schünemann, V.; Kasche, V. Stability of immobilized soybean lipoxygenases: influence of coupling conditions on the ionization state of the active site Fe. Enzyme Microb. Technol. 28:168-175; 2001.

[127] Jack, G. Immunoaffinity chromatography. Mol. Biotechnol. 1:59-86; 1994.

[128] Fiorito, P. A.; Córdoba de Torresi, S. I. Optimized multilayer oxalate biosensor. Talanta 62:649-654; 2004.

[129] Malpass, C. A.; Millsap, K. W.; Sidhu, H.; Gower, L. B. Immobilization of an oxalate-degrading enzyme on silicone elastomer. J. Biomed. Mater. Res. 63:822-829; 2002.

[130] Raghavan, K. G.; Tarachand, U. Degradation of oxalate in rats implanted with immobilized oxalate oxidase. FEBS Lett. 195:101-105; 1986.

[131] Lin, R.; Wu, R.; Huang, X.; Xie, T. Immobilization of Oxalate Decarboxylase to Eupergit and Properties of the Immobilized Enzyme. Prep. Biochem. Biotechn. 41:154- 165; 2011.

[132] Twahir, U.; Molina, L.; Ozarowski, A.; Angerhofer, A. Immobilization of bacillus subtilis oxalate decarboxylase on a Zn-IMAC resin. Biochem. Biophys. Rep.; 2015.

[133] Block, H.; Maertens, B.; Spriestersbach, A.; Brinker, N.; Kubicek, J.; Fabis, R.; Labahn, J.; Schäfer, F. Chapter 27 Immobilized-Metal Affinity Chromatography (IMAC): A Review. In: Richard, R. B.; Murray, P. D., eds. Methods Enzymol.: Academic Press; 2009: 439-473.

[134] Gaberc-Porekar, V.; Menart, V. Potential for Using Histidine Tags in Purification of Proteins at Large Scale. Chem. Eng. Technol. 28:1306-1314; 2005.

[135] Porath, J. Immobilized metal ion affinity chromatography. Protein Express. Purif. 3:263-281; 1992.

[136] Bhardwaj, A.; Lee, J.; Glauner, K.; Ganapathi, S.; Bhattacharyya, D.; Butterfield, D. A. Biofunctional membranes: an EPR study of active site structure and stability of papain non-covalently immobilized on the surface of modified poly(ether) sulfone membranes through the avidin-biotin linkage. J. Membr. Sci. 119:241-252; 1996.

[137] Butterfield, D. A.; Bhattacharyya, D.; Daunert, S.; Bachas, L. Catalytic biofunctional membranes containing site-specifically immobilized enzyme arrays: a review. J. Membr. Sci. 181:29-37; 2001.

148

[138] Chen, Y.; Pasquinelli, R.; Ataai, M.; Koepsel, R. R.; Kortes, R. A.; Shepherd, R. E. Coordination of Two High-Affinity Hexamer Peptides to Copper(II) and Palladium(II) Models of the Peptide−Metal Chelation Site on IMAC Resins. Inorg. Chem. 39:1180- 1186; 2000.

[139] Garcia-Galan, C.; Berenguer-Murcia, Á.; Fernandez-Lafuente, R.; Rodrigues, R. C. Potential of Different Enzyme Immobilization Strategies to Improve Enzyme Performance. Adv. Synth. Catal. 353:2885-2904; 2011.

[140] Marg, G. A.; Millhauser, G. L.; Skerker, P. S.; Clark, D. S. Application of EPR Methods in Studies of Immobilized Enzyme Systemsa. Ann. N. Y. Acad. Sci. 469:253- 258; 1986.

[141] Ozarowski, A.; Zvyagin, S. A.; Reiff, W. M.; Telser, J.; Brunel, L.-C.; Krzystek, J. High-Frequency and -Field EPR of a Pseudo-octahedral Complex of High-Spin Fe(II): Bis(2,2‘-bi-2-thiazoline)bis(isothiocyanato)iron(II). J. Am. Chem. Soc. 126:6574-6575; 2004.

[142] Un, S.; Tabares, L. C.; Cortez, N.; Hiraoka, B. Y.; Yamakura, F. Manganese(II) Zero-Field Interaction in Cambialistic and Manganese Superoxide Dismutases and Its Relationship to the Structure of the Metal Binding Site. J. Am. Chem. Soc. 126:2720- 2726; 2004.

[143] Weber, R. T.; Jiang, J.; Barr, D. P. EMX user's manual.

[144] DeVore, T. C.; Van Zee, R. J.; Weltner, W. High spin molecules: ESR of MnF and MnF2 at 4°K. J. Chem. Phys. 68:3522-3527; 1978.

[145] Carmieli, R.; Papo, N.; Zimmermann, H.; Potapov, A.; Shai, Y.; Goldfarb, D. Utilizing ESEEM spectroscopy to locate the position of specific regions of membrane- active peptides within model membranes. Biophys. J . 90:492-505; 2006.

[146] Kabsch, W. Xds. Acta Crystallogr. Sect. D 66:125-132; 2010.

[147] McCoy, A. J.; Grosse-Kunstleve, R. W.; Adams, P. D.; Winn, M. D.; Storoni, L. C.; Read, R. J. Phaser crystallographic software. J. Appl. Crystallogr. 40:658-674; 2007.

[148] Adams, P. D.; Afonine, P. V.; Bunkóczi, G.; Chen, V. B.; Davis, I. W.; Echols, N.; Headd, J. J.; Hung, L.-W.; Kapral, G. J.; Grosse-Kunstleve, R. W. PHENIX: a comprehensive Python-based system for macromolecular structure solution. Acta Crystallogr. Sect. D 66:213-221; 2010.

[149] Afonine, P. V.; Grosse-Kunstleve, R. W.; Echols, N.; Headd, J. J.; Moriarty, N. W.; Mustyakimov, M.; Terwilliger, T. C.; Urzhumtsev, A.; Zwart, P. H.; Adams, P. D. Towards automated crystallographic structure refinement with phenix. refine. Acta Crystallogr. Sect. D 68:352-367; 2012.

149

[150] Emsley, P.; Cowtan, K. Coot: model-building tools for molecular graphics. Acta Crystallogr. Sect. D 60:2126-2132; 2004.

[151] Berliner, L. J.; Grunwald, J.; Hankovszky, H. O.; Hideg, K. A novel reversible thiol-specific spin label: papain active site labeling and inhibition. Anal. Biochem. 119:450-455; 1982.

[152] Kenyon, G. L.; Bruice, T. W. Novel sulfhydryl reagents. Method. Enzymol. 47:407; 1977.

[153] Polyhach, Y.; Bordignon, E.; Jeschke, G. Rotamer libraries of spin labelled cysteines for protein studies. Phys. Chem. Chem. Phys. 13:2356-2366; 2011.

[154] Livshits, V.; Dzikovski, B.; Marsh, D. Mechanism of relaxation enhancement of spin labels in membranes by paramagnetic ion salts: dependence on 3d and 4f ions and on the anions. J. Magn. Res. 148:221-237; 2001.

[155] Schütte, H.; Flossdorf, J.; Sahm, H.; Kula, M.-R. Purification and Properties of Formaldehyde Dehydrogenase and Formate Dehydrogenase from Candida boidinii. Eur. J. Biochem. 62:151-160; 1976.

150

BIOGRAPHICAL SKETCH

Umar Tariq Twahir, son of Mohammed Kamal Twahir and Bibi Shakira Twahir, elder bother to Aisha and Fabiya Twahir grew up in Bronx, New York, where he completed his primary and secondary education in 2005. Umar started his undergraduate career in chemistry at the City College of New York, but shortly after transferred to Georgia State University. He earned his Bachelor of Science in Chemistry in December of 2009. He then went on to also complete a Master of Science in

Chemistry under the supervision of Professor Stuart Allison, and completed this degree in May of 2011.

In August of 2011, Umar went on to start his graduate career at the University of

Florida, pursuing a Doctor of Philosophy in Physical Chemistry under the advisement of

Professor Alexander Angerhofer. His research entailed the utilization of Electron

Paramagnetic Resonance to shed light onto the enzymatic mechanism catalyzed by

Bacillus subtilis Oxalate Decarboxylase. During his time at the University of Florida he was awarded The Center for Condensed Matter Science for CCMS Graduate Student

Summer Fellowship (2012) and the Early Career Physical Chemistry Award (2012). He was also the recipient of multiple graduate student travel and poster awards while attending local and international meetings. Umar received his Ph.D. in Chemistry in the fall of 2015.

151