Novel Reactions Catalyzed by Ferritin-Like Diiron-Carboxylate

The Pennsylvania State University

The Graduate School

Eberly College of Science

NOVEL REACTIONS CATALYZED BY

FERRITIN-LIKE DIIRON-CARBOXYLATE

ENZYMES: CHEMISTRY, KINETICS AND

MECHANISMS

A Dissertation in

Biochemistry and Molecular Biology

Ning Li

Submitted in Partial Fulfillment

of the Requirements for the Degreee of

Doctor of Philosophy

August 2012

DOCTORAL COMMITTEE PAGE

The dissertation of Ning Li was reviewed and approved* by the following:

Joseph Martin Bollinger, Jr. Professor of Chemistry and Professor of Biochemistry and Molecular Biology Dissertation Adviser Chair of Committee

Carsten Krebs Professor of Chemistry and Professor of Biochemistry and Molecular Biology Dissertation Adviser Co-Chair of Committee

Squire J. Booker Associate Professor of Chemistry and Associate Professor of Biochemistry and Molecular Biology

Ming Tien Professor of Biochemistry and Molecular Biology

Wayne Curtis Professor of Chemical Engineering

Scott Selleck Professor and Head of Department of Biochemistry and Molecular Biology

*Signatures are on file in the Graduate School.

Abstract

The activation of molecular oxygen plays an important role in biological oxidation reactions. Various enzymes employ transition metals to facilitate this chemical process.

The ferritin-like dimetal-carboxylate oxidases or oxygenases, which include soluble methane monooxygenase (sMMO), toluene/o-xylene monooxygenase, phenol hydroxylase (PhOH), alkene monooxygenase (AMO), plant soluble fatty acyl–acyl carrier protein desaturase, and the β2 (R2) subunit of class Ia ribonucleotide reductase

(RNR), as well as N-oxygenase AurF and cyanobacterial aldehyde decarbonylase

(cAD), activate O2 at a carboxylate-bridged non-heme dimetal center. Such enzymes have been extensively investigated due to their extremely interesting and complicated chemistry, as well as their potential commercial usage. These enzymes usually function in multicomponent complexes including a reductase, a scaffold subunit, and a ferritin-like oxygenase/oxidase. Diiron is most commonly employed by these oxygenases/oxidases as the metal cofactor. A general reaction cycle of the diiron cofactor involves: 1) the reduction of the diferric form to diferrous by the reductase component, in which electrons are transferred from NADH or NADPH; 2) activation of molecular oxygen by the diferrous cofactor, forming some type of peroxo-diferric intermediate; 3) oxidation of the substrate as the peroxo species is converted to diferric form, ready to be reduced to start another catalytic cycle. Manganese has also been reported to replace one or both of the iron atoms of the metal cofactors, most notably in the β2 subunits of class Ib and Ic ribonucleotide reductases. The class Ic β2 subunit utilizes a heterodinuclear Mn/Fe cofactor while the β2 subunit of class 1b ribonucleotide reductase employs a dimanganese cofactor. Due to the presence of the transition metals, these enzymes usually possess rich spectroscopic features, allowing

iii multiple spectroscopic methods such as Mössbauer spectroscopy, electron paramagnetic resonance (EPR) spectroscopy, UV/visible absorption spectrophotometry, electron-nuclear double resonance (ENDOR) spectroscopy,

Raman spectroscopy, etc. to be used to study the reactions. Due to the speed at which most of the reactions catalyzed by these dimetal enzymes occur, the millisecond to second time scale is required, thereby necessitating fast quench or real time measurement; therefore, rapid-mix techniques are usually adjoined to spectroscopic methods to meet this requirement.

In this thesis, the reaction mechanism and intermediates of dinuclear ferritin-like oxygenase/oxidase are summarized and the progress I have made during my Ph.D. study on AurF and cADs is narrated in detail.

TABLE OF CONTENTS

TABLE OF CONTENTS ...... v LIST OF FIGURES ...... viiii LIST OF TABLES ...... xiii LIST OF SCHEMES ...... xiii

Chapter 1 Current understanding of ferritin-like diiron-carboxylate enzymes: chemistry, kinetics and mechanisms ...... 1 1.1 Ferritin-like dimetal-carboxylate proteins ...... 2 1.2 Peroxo intermediate species ...... 4 1.3 Soluble methane monooxygenase (sMMO) ...... 9 1.4 Ribonucleotide reductase ...... 12 1.4.1 Class I RNR ...... 12 1.4.2 Class Ia RNR ...... 13 1.4.3 Class Ic RNR ...... 14 1.4.4 Class Ib RNR ...... 16 1.5 Plant soluble Stearoyl-Acyl Carrier Protein Δ9 Desaturase ...... 20 1.6 Toluene/o-xylene monooxygenase (ToMO) ...... 21 1.7 AurF ...... 23 1.7.1 The function of AurF and the debate on its metal cofactor ...... 23 1.7.2 Diiron AurF and the peroxo intermediate discovered by our group ...... 27 1.7.3 The discovery of 4 electron oxidation of 4-hydroxylaminobenzoate ...... 30 1.7.4 The study of 4-hydrazinobenzoate analog reactivity ...... 32 1.8 Cyanobacterial aldehyde decarbonylase (cAD) ...... 36 1.8.1 The discovery of formate as the co-product ...... 36 1.8.2 The discovery of oxygen dependence ...... 40 1.8.3 The discrepancy between oxygen dependent and independent mechanism ...... 43 1.8.4 Transient kinetics study and metal cofactors ...... 46 1.8.5 Ongoing study of AD ...... 49 1.9 References ...... 51 Chapter 2 The mechanism of the reaction of AurF with its substrate analogue, 4-hydrazinobenzoate ...... 64 Abstract ...... 65

Introduction ...... 66 Materials and Methods ...... 69 Materials ...... 69 Stopped-flow experiments ...... 69 HPLC and LC/MS analysis ...... 69 Results ...... 70

III/III III/III Analysis of the reaction of AurF (Fe2 and peroxo-Fe2 -AurF) with Ar-NHNH2 by Stopped-Flow Absorption (SF-Abs) spectrophotometry...... 70 Study of the Reaction Kinetics and Stoichiometry...... 70

III/III Evaluation of Diiron Products in Reaction of diferric and Peroxo-Fe2 -AurF with

Ar-NHNH2 by Mössbauer Spectroscopy...... 71

Verification of Catalytic Oxidation of Ar-NHNH2 by as-isolated AurF...... 73 Evidence for the reaction product by isotopic labeling LC/MS assay...... 74 Discussion...... 75 References ...... 95 Appendix A Four-electron oxidation of p-hydroxylaminobenzoate to p-nitrobenzoate by a peroxodiferric complex in AurF from Streptomyces thioluteus ...... 98 Abstract ...... 99 Introduction ...... 99 Results ...... 100

III/III Testing for a Reaction Between Peroxo-Fe2 -AurF and Ar-NHOH by Stopped-Flow Absorption (SF-Abs) Experiments...... 100

III/III Evaluation of Di-iron Products in Reaction of Peroxo-Fe2 -AurF with Ar-NHOH by Mössbauer Spectroscopy...... 100

II/II Verification of Catalytic Oxidation of Ar-NHOH by Fe2 -AurF...... 102 III/III Testing for Reduction of µ-oxo-Fe2 -AurF by Ar-NHOH...... 102 III/III Re-evalution of the Diiron Products from the Reaction of Peroxo-Fe2 -AurF with

Limiting Ar-NH2...... 102 Discussion ...... 103 Experimental Procedures ...... 104 Acknowledgments ...... 104 References ...... 104 SUPPORTING INFORMATION ...... 105

Appendix B Detection of formate, rather than carbon monoxide, as the stoichiometric coproduct in conversion of fatty aldehydes to alkanes by a cyanobacterial aldehyde decarbonylase ...... 111 ABSTRACT ...... 112 INTRODUCTION, RESULTS, AND DISCUSSION...... 112 ACKOWLEDGEMENT ...... 115 REFERENCES ...... 115 SUPPORTING INFORMATION ...... 116 Appendix C Conversion of fatty aldehydes to alka(e)nes and formate by a cyanobacterial aldehyde decarbonylase: cryptic redox by an unusual dimetal oxygenase ...... 129 ABSTRACT ...... 130 INTRODUCTION, RESULTS AND DISCUSSION...... 130 ACKNOWLEDGEMENT ...... 133 REFERENCES ...... 133 SUPPORTING INFORMATION ...... 134 Appendix D Evidence for Only Oxygenative Cleavage of Aldehydes to Alk(a/e)nes and Formate by Cyanobacterial “Aldehyde Decarbonylase” ...... 136 ABSTRACT ...... 139 EXPERIMENTAL PROCEDURES ...... 144 RESULTS ...... 147 Development of a Direct LC/MS Assay for Formate...... 147

Demonstration of the O2 Requirement for Formate Production by the Np and Pm ADOs...... 147

18 Assessment of the Origin of the New O-atom in Formate by Use of O2...... 148 Testing Short-chain Aldehydes as ADO Substrates...... 149 Quantification of n-Hexane Produced by the ADO-Catalyzed Conversion of n-Heptanal...... 150

18 Measurement of the Exchange Rate of the Aldehyde Carbonyl of Octanal with H2 O. 150 18 Test for Trace Hydrolytic Cleavage of n-Heptanal by Analysis of Reactions in H2 O. ... 151 Test for a Stoichiometric or Catalytic Role of the Reducing System...... 152 DISCUSSION ...... 153 REFERENCES ...... 155 SCHEME AND FIGURE LEGENDS ...... 159

vii

SUPPORTING MATERIAL ...... 169 Appendix E Summary of unpublished data ...... 178

LIST OF FIGURES

LIST OF FIGURES ...... viii LIST OF SCHEMES ...... xiii Chapter 1 Current understanding of ferritin-like diiron-carboxylate enzymes: chemistry, kinetics and mechanisms ...... 1 Figure 1.1: Representative crystal structures of ferritin and ferritin-like dimetal-carboxylate proteins...... 3 Figure 1.2: Active site structures of Pm AD, AurF and E. coli R2...... 7

Figure 1.3: Possible idealized geometries of (hydro)peroxo-Fe2(III/III) complexes...... 8 Chapter 2 The mechanism of the reaction of AurF with its substrate analogue, 4-hydrazinobenzoate ...... 64 Figure 1A. SF-Abs experiments to demonstrate the reduction of as-isolated AurF by

Ar-NHNH2 and the oxidation of Ar-NHNH2 reduced AurF with oxygen...... 81 III/III Figure 1B. SF-Abs experiments to monitor the reaction of peroxo-Fe2 -AurF with

Ar-NHNH2 ...... 82 III/III Figure 2. Kinetics of the reaction of peroxo-Fe2 -AurF with Ar-NHNH2...... 83

Figure 3. Kinetics of the reaction of as-isolated AurF with Ar-NHNH2 ...... 84 III/III Figure 4. Dependence of the formation of peroxo-Fe2 -AurF on the concentration of

O2 ...... 85 Figure 5. 4.2-K/53-mT Mössbauer spectra of samples in which either as-isolated AurF III/III III/III (Fe2 -AurF) or peroxo-Fe2 -AurF was reacted with Ar-NHNH2...... 86 Figure 6. Reversed-phase high performance liquid chromatography (RP-HPLC) of the small molecule reactants and products in the reaction of as-isolated AurF with Ar-NHNH2...... 88 Figure 7. LC/MS chromatogram demonstrating the derivatized benzoate products of

Ar-NHNH2 oxidation catalyzed by as-isolated AurF in the presence of O2...... 89 Figure S1 ...... 90 Figure S2 ...... 91 Figure S3 Selected ion monitoring chromatogram demonstrating the production of

Ar-NO2 from Ar-NHNH2 catalyzed by AurF ...... 92

viii

Figure S4 Selected ion monitoring chromatogram demonstrating the incorporation of

oxygen atoms to Ar-NO2 from Ar-NHNH2 oxidation catalyzed by AurF ...... 93 Figure S5. Selected ion monitoring chromatogram demonstrating the product with m/z =

297 from Ar-NHNH2 oxidation catalyzed by AurF ...... 94 Appendix A Four-electron oxidation of p-hydroxylaminobenzoate to p-nitrobenzoate by a peroxodiferric complex in AurF from Streptomyces thioluteus ...... 98 Figure 1. Sequential-mixing SF-Abs experiments to monitor the reaction of III/III peroxo-Fe2 -AurF with Ar-NHOH...... 100 III/III Figure 2. 4.2-K/53-mT Mössbauer spectra of samples in which peroxo-Fe2 -AurF was

reacted with Ar-NHOH or Ar-NH2...... 101 Figure 3. Reversed-phase high performance liquid chromatography (RP-HPLC) of the small II/II molecule reactants and products following incubation of Fe2 -AurF with excess Ar-NHOH

and O2...... 102

III/III Figure S1. 4.2-K∕53-mT Mössbauer reference spectrum of peroxo-Fe2 -AurF...... 107 Figure S2. 4.2-K∕53-mT Mössbauer spectra of samples from the reaction of III/III peroxo-Fe2 -AurF with Ar-NHOH in the presence of limiting O2...... 108 Figure S3. 4.2-K∕53-mT Mössbauer spectra of samples from the reaction of III/III peroxo-Fe2 -AurF with Ar-NHOH in the presence of excess O2...... 109 Figure S4. Reversed-phase high performance liquid chromatographic (RP-HPLC) analysis of the synthetic Ar-NHOH upon its incubation in 100 mM HEPES buffer (pH 7.5)...... 109 Figure S5. 4.2-K∕53-mT Mössbauer spectra of samples from the treatment of as-isolated III/III μ-oxo-Fe2 -AurF with Ar-NHOH...... 110 1 Figure S6. 360 MHz H-NMR spectrum of 5 mg crude Ar-NHOH dissolved in 1 g DMSO-d6...... 110

1 Figure S7. 360 MHz H-NMR spectrum of 5 mg crude Ar-NHOH dissolved in 1 g DMSO-d6, 2 to which 100 μL H2O was added and allowed to reach equilibrium for 24 h...... 110 Appendix B Detection of formate, rather than carbon monoxide, as the stoichiometric coproduct in conversion of fatty aldehydes to alkanes by a cyanobacterial aldehyde decarbonylase ...... 111 Figure 1. Reconstructed mass spectra showing formate production in reactions of Np AD...... 113 Figure S1. GC-MS analysis of the decarbonylation of n-octadecanal to heptadecane by Np AD and confirmation of the requirement for the reducing system...... 122

Figure S2. GC-MS total ion chromatograms demonstrating the dependence of n-C17H36 yield on time (A) and Np AD concentration (B) in the reaction with R-13CHO...... 123 Figure S3. Myoglobin/UV-absorption assay for detection of CO...... 124 Figure S4. LC-MS calibration curves for formate, 1-[13C]formate, and propionate...... 125

Figure S5. Dependence of formate yield on reaction time (compare front to back traces) and Np AD concentration (compare different color traces within a group)...... 126 Figure S6. Analysis of deuterium incorporation into heptadecane upon reaction of Np AD ...... 127 Appendix C Conversion of fatty aldehydes to alka(e)nes and formate by a cyanobacterial aldehyde decarbonylase: cryptic redox by an unusual dimetal oxygenase ...... 129

Figure 1. Reconstructed mass spectra illustrating the catalytic requirement for O2 and the 18 18 incorporation of O from O2 into the formate product in the Np AD reaction...... 131 Figure S1. Determination of the ratio of NADPH oxidized to formate produced during the Np AD reaction...... 135 Appendix D Evidence for Only Oxygenative Cleavage of Aldehydes to Alk(a/e)nes and Formate by Cyanobacterial “Aldehyde Decarbonylase” ...... 136 Figure 1. LC/MS detection of formate produced enzymatically by Pm ADO and Np ADO...... 159

18 Figure 2. LC/MS detection of formate produced in O2-tracer experiments by Np ADO and Pm ADO...... 159 Figure 3. Time-dependence of, and requirements for, the Np ADO-catalyzed production of formate...... 159 Figure 4. Determination of the kinetics of exchange of the carbonyl oxygen of n-[13C]-octanal with solvent by 13C-NMR-spectroscopy...... 159

18 18 Figure 5. LC/MS O2- and H2 O-isotope-tracer experiments to determine the origin of the O-atom incorporated into formate by Np ADO and Pm ADO...... 160 Figure 6. NADH:formate stoichiometry of the Np ADO-catalyzed production of formate from n-heptanal...... 160 Figure S1. Single-ion-monitoring (SIM) chromatograms showing the proportionality between the peak intensity (area) and the concentration of the formate analyte ...... 171 Figure S2. Plot of peak intensity (area) versus the concentration of the formate analyte172 Figure S3. 1H-NMR spectrum of n-1-[13C]-octadecanal substrate...... 173 Figure S4. 13C-NMR spectrum of n-1-[13C]-octadecanal substrate...... 174 Figure S5. 1H-NMR spectrum of n-1-[13C]-octanal substrate...... 175 Figure S6. 13C-NMR spectrum of n-1-[13C]-octanal substrate...... 176 Figure S7. 360 MHz 1H-NMR spectrum of n-1-[2H]-heptanal substrate...... 177 Appendix E Summary of unpublished data ...... 178

III/III Figure 1. Kinetic traces showing the reaction of peroxo-Fe2 -AurF with 4-mercaptobenzoate...... 179

III/III Figure 2. Kinetic traces showing the reaction of peroxo-Fe2 -AurF with 4-mercaptobenzoate...... 180 Figure 3. Stopped flow experiment demonstrating the binding of 4-hydroxybenzoate to III/III Fe2 -AurF...... 181 Figure 4. EPR Spectra showing the redox reaction of rv0233 ortholog of Saccharopolyspora erythraea (SRV)...... 183 Figure 5. Stopped flow experiment demonstrating the substrate triggered oxygen activation catalyzed by Np AD...... 184 Figure 6. Kinetic traces showing the substrate dependence of oxygen activation catalyzed by Np AD...... 185 Figure 7. Kinetic traces showing the substrate concentration dependence of oxygen activation catalyzed by Np AD...... 186

II/II Figure 8. 4.2-K/53-mT Mössbauer spectra of samples in which Fe2 -Np AD was reacted with O2...... 187 III/III Figure 9. LC/MS analysis of the formate produced by reacting peroxo-Fe2 -Np AD intermediate with MeOPMS/NADH...... 188 Figure 10. Kinetic traces demonstrating the oxidation of MeOPMS by Np AD...... 189 Figure 11. Kinetic traces demonstrating the effect of as isolated Np AD on the oxidation of NADH by O2...... 190 Figure 12. Stopped flow experiment demonstrating the reduction of cyanobacterial ferredoxin by dithionite...... 191 Figure 13. Stopped flow experiment demonstrating the reduction of as isolated Np AD by dithionite-reduced ferredoxin...... 192 Figure 14. Stopped flow experiment demonstrating the reduction of as isolated Np AD by NADH-reduced MeOPMS...... 193 Figure 15. LC/MS chromatograms demonstrating the formate production from aldehyde substrates catalyzed by as isolated Np AD in the presence of H2O2...... 194 Figure 16. Activity of Np AD and its variants in the presence of N/F/FR or N/PMS reducing system...... 195 Figure 17. Diagram showing the constructed plasmid for recombining Synechocystis sp. PCC6803 AD (sll0208) and AAR (sll0209) genes to the pAQ1 plasmid of Synechococcus sp. PCC7002...... 196 Figure 18. Coomassie-blue stained 12% SDS–PAGE analysis of whole cell proteins of Synechococcus sp. PCC7002, wild type or 6803 AD over-expressing strains...... 197 Figure 19. GC/MS chromatogram showing the detection of heptadecane in the Synechococcus sp. PCC7002 strain that expresses 6803 AD and AAR genes...... 198 Figure 20. 13C-formate yield in the presence of different ferredoxins...... 199

LIST OF TABLES

Appendix B Detection of formate, rather than carbon monoxide, as the stoichiometric coproduct in conversion of fatty aldehydes to alkanes by a cyanobacterial aldehyde decarbonylase ...... 111 Table 1. Concentrations of formate and heptadecane measured in Np AD reactions under two different sets of conditions...... 113 Appendix D Evidence for Only Oxygenative Cleavage of Aldehydes to Alk(a/e)nes and Formate by Cyanobacterial “Aldehyde Decarbonylase” ...... 136

Table S1. Yields of formate from C7, C8 and C10 saturated n-aldehyde substrates in Np ADO reactions...... 170 Appendix E Summary of unpublished data ...... 178 Table 1. A list of the AurF or AurF-related variants constructed...... 182 Table 2. Affect of 6803 cell extract (soluble portion) on the yield of formate product of Np AD reaction...... 200 Table 3. Affect of metals on the yield of formate product of Np AD reaction...... 201

xii

LIST OF SCHEMES

Chapter 1 Current understanding of ferritin-like diiron-carboxylate enzymes: chemistry, kinetics and mechanisms ...... 1

Scheme 1.1: Diverse reaction pathways of the peroxo-Fe2(III/III) intermediates in reactions of the ferritin-like diiron-carboxylate proteins...... 6 Scheme 1.2: Proposed sMMO reaction mechanism...... 11

Scheme 1.3: Detailed reaction mechanism of E. coli β2 and Ct β2...... 18

Scheme 1.4: Proposed reaction mechanism of β2 subunit of class Ib RNR...... 19 Scheme 1.5: AurF reaction and proposed mechanisms...... 26 Scheme 1.6: Proposed reaction mechanism of 4-hydrazinobenzoate with AurF...... 35 Scheme 1.7: Three possible reactions catalyzed by cAD...... 39 Scheme 1.8: Proposed mechanisms for formate production emphasizing on oxygen source...... 42 Scheme 1.9: Proposed reaction mechanism of cAD...... 50 Chapter 2 The mechanism of the reaction of AurF with its substrate analogue, 4-hydrazinobenzoate ...... 64 Scheme 1 ...... 79 Scheme 2 ...... 80 Appendix A Four-electron oxidation of p-hydroxylaminobenzoate to p-nitrobenzoate by a peroxodiferric complex in AurF from Streptomyces thioluteus ...... 98 Scheme 1. Reactions catalyzed by AurF...... 99

Scheme 2. Proposed mechanism of the four-electron oxidation of Ar-NHOH to Ar-NO2 by III/III peroxo-Fe2 -AurF...... 103

Scheme S1. Proposed pathways for the conversion of Ar-NH2 to Ar-NO2 by AurF...... 106 Appendix B Detection of formate, rather than carbon monoxide, as the stoichiometric coproduct in conversion of fatty aldehydes to alkanes by a cyanobacterial aldehyde decarbonylase ...... 111 Scheme 1. Three possible outcomes of the Np AD reaction...... 112 Appendix C Conversion of fatty aldehydes to alka(e)nes and formate by a cyanobacterial aldehyde decarbonylase: cryptic redox by an unusual dimetal oxygenase ...... 129 Scheme 1. Two Alternative Explanations for the Similarity of Np AD to Di-iron Oxidases and Oxygen-ases and its Requirement for a Reducing System to Promote an Apparently Hydrolytic Reaction...... 131

xiii

Scheme 2. Hypothetical Mechanism for the Np AD Reaction...... 132 Appendix D Evidence for Only Oxygenative Cleavage of Aldehydes to Alk(a/e)nes and Formate by Cyanobacterial “Aldehyde Decarbonylase” ...... 136 Scheme 1. Predicted origins of the oxygen atoms in the formate co-product generated in the oxygenative (A) and hydrolytic (B) reactions purportedly catalyzed by cyanobacterial “ADs”...... 159

xiv

Chapter 1

Current understanding of ferritin-like diiron-carboxylate enzymes: chemistry, kinetics and mechanisms

1.1 Ferritin-like dimetal-carboxylate proteins

Many enzymes harbor two closely positioned transition metals to serve as electrostatic catalysts [1]. One unifying feature distinguishing one main group of these enzymes is that a carboxylate group bridges the two transition metals at the active site with histidine, glutamate and/or aspartate residues providing additional coordinating ligands. The dimetal center is tuned to activate molecular oxygen and perform a large variety of redox reactions. Examples of these enzyme include R2 (β2) subunit of class

Ia ribonucleotide reductase [2-3], methane monooxygenase [4], stearoyl-acyl carrier protein Δ9-desaturase [5], toluene/o-xylene monooxygenase [6], phenol hydroxylase

[7], alkene hydroxylase [8], N-oxygenase AurF [9] and cyanobacterial aldehyde decarbonylase (cAD) [10]. Structurally, they resemble the fold of ferritin and therefore they are termed ferritin-like dimetal-carboxylate proteins (some representative structures are shown in figure 1.1).

Figure 1.1: Representative crystal structures of ferritin and ferritin-like dimetal-carboxylate proteins. Data acquired from RCSB protein data bank.

1.2 Peroxo intermediate species

The most commonly present transition metal in the dimetal-carboxylate enzymes is iron. Diiron-carboxylate enzymes activate molecular oxygen by diferrous (Fe2(II/II)) metal cofactor which is reduced from the resting diferric (Fe2(III/III)) state to form a peroxo complex [11]. Although the structures of the cofactors of these ferritin-like diiron-carboxylate enzymes are similar, the spectroscopic properties of the peroxo complexes generated from each enzyme are different (Scheme 1.1). For example, wild-type toluene monooxygenase forms a transparent peroxo intermediate [12-13].

An isoleucine to tryptophan mutation (I100W) would convert the transparent peroxo intermediate to a blue complex (a band centered ~700 nm) [13]. Soluble methane monooxygenase forms a peroxo intermediate, which absorbs at 450 nm and 720 nm, and is converted to another intermediate Q that is responsible for the hydroxylation of methane [14]. In my research work, AurF and cAD are studied and two new peroxo intermediates are discovered, which expands the repertoire of known diiron peroxo complexes in the literature. AurF forms a rusty-colored long-lived peroxo-Fe2(III/III) intermediate, which absorbs at ~500 nm [15]. The presence of a third histidine which is coordinated to one iron at the diiron center might play a role in the stability of peroxo-Fe2(III/III)-AurF complex. cAD peroxo-Fe2(III/III) has been observed recently and has a yellow color (absorbs at ~450 nm). Both AurF and cAD peroxo intermediates exhibit site resolution (distinguishable feature of two irons in the peroxo complexes) shown by Mössbauer spectroscopy analysis. In summary, it is likely that evolutionally these enzymes finely tune the seemingly similar milieu of the diiron cofactor to form various peroxo intermediates in order to carry out specific reactions.

The structure of the dimetal cofactor of E. coli β2, AurF and cAD are demonstrated in

4 figure 1.2 and the possible geometries of (hydro)peroxo-Fe2(III/III) complexes are summarized in figure 1.3.

Scheme 1.1: Diverse reaction pathways of the peroxo-Fe2(III/III) intermediates in reactions of the ferritin-like diiron-carboxylate proteins. Adapted from ref [16].

Pm AD AurF

Alignment of Ec R2 three proteins

Figure 1.2: Active site structures of Pm AD, AurF and E. coli R2. Data acquired from

RCSB protein data bank.

Figure 1.3: Possible idealized geometries of (hydro)peroxo-Fe2(III/III) complexes.

Adapted from AurF NSF grant proposal.

1.3 Soluble methane monooxygenase (sMMO)

Methanotrophic bacteria use methane as the sole carbon and energy source for their growth [17]. The first key step of methane metabolism is the hydroxylation of methane to methanol, which requires the breaking of a very strong C-H bond (104 kcal/mol). This reaction is catalyzed by methane monooxygenase (MMO) [18-23].

Two types of MMO have been discovered in methanotrophs, particulate MMO

(pMMO), which is bound to the membrane; and soluble MMO (sMMO), which exists in the cytoplasm of the cell. sMMO from both Methylococcus capsulatus (Mc) and

Methylosinus trichosporium (Mt) have been studied extensively to understand this methane hydroxylation reaction [14]. The active site of sMMO is a carboxylate-bridged diiron center [23-24]. sMMO has three components: a hydroxylase protein MMOH, a reductase protein

MMOR, and a regulatory protein MMOB [23-24]. The oxidation/hydroxylation reaction is catalyzed by MMOH, a 251-kDa component with an α2β2γ2 configuration

[4]. The α subunit has the 4-helix bundle ferritin-like protein architecture that coordinates the diiron cofactor. The MMOR component is a 38.5-kD protein containing one [2Fe-2S] cluster and one FAD cofactor, which transfers electron from

NADH to MMOH [25-27]. MMOB component functions by facilitating the structural change during the redox reaction process and the hydroxylation rate and regioselectivity is affected by the addition of MMOB component in the sMMO complex [24-25, 28].

For sMMO, the activation step is the most complicated one. Although the complete in vivo reaction requires MMOH, MMOB and MMOR, as isolated diferric MMOH itself can be reduced chemically, to the diiron(II/II) form, which can then activate oxygen in

9 the presence of MMOB [20, 25]. Fe2(II/III)-MMOH can be produced by cryoreduction of frozen diiron(III/III)-MMOH via γ-irradiation, yet it is not reactive with oxygen and therefore is not relevant to the study of the enzyme’s mechanism

* [29]. The first intermediate is a peroxo-Fe2(III/III) complex designated P , which has distinct optical spectroscopic features at 420 nm and 720 nm [30]. The intermediate

P* is then converted to Hperoxo (P) with similar optical features to P*, facilitated by proton transfer [31]. P* is only detected in the Mt sMMO but not in the Mc ortholog

[30]. P is potent to oxidize electron rich substrates such as propylene [31-33]. The geometric structure of P* and P are not yet known. In the absence of any substrates, P can rapidly via an O-O bond cleavage convert to the intermediate Q which is responsible for methane hydroxylation and which is believed to have a diamond core structure [20, 34-35]. Q is easily monitored spectroscopically due to its bands at 350 nm and 420 nm [20]. The linear dependence of the decay rate of Q on methane and other hydrocarbon substrates indicates a weak substrate binding [20, 31, 34, 36]. In the absence of substrates, Q decays into the diferric form by acquiring two electrons and two protons. Another intermediate, termed Q*, with an absorption band at ~455 nm is proposed to exist along the pathway of the decay of Q [30]. The spontaneous conversion of sMMO Hperoxo to a high-valent Q may be the reason for its ability to oxidize the strong C-H bond of methane. The proposed reaction for sMMO is summarized in Scheme 1.2.

Scheme 1.2: Proposed sMMO reaction mechanism. Adapted from ref [24].

1.4 Ribonucleotide reductase

1.4.1 Class I RNR

Ribonucleotide reductase (RNR) catalyzes the formation of deoxyribonucleotides from ribonucleotides by replacing the 2’OH group by a hydrogen. The enzyme provides the building blocks for de novo DNA synthesis. The chemically challenging reductive reaction, the catalysis step, is achieved by generating a transient cysteine thiyl radical (C•) at the active site. This C• is regenerated after 3’H abstraction on the ribose ring, making the reaction catalytic [37-38]. Three different strategies have evolved to generate this C•, and RNRs have been classified according to the metallocofactors required in this activation step [39-41]. All class I RNRs are composed of two subunits: R1 (α) and R2 (β) and each of them is a homodimer (α2 and β2) [39]. The α2 subunit binds to the substrates and allosteric effectors, whereas the β2 subunit belongs to the ferritin-like dimetal-carboxylate protein family and houses a dimetal unit (identical or different) to facilitate the activation of the reduction reaction [39, 41]. They are further categorized into three subclasses: class Ia β2 employs diiron(III/III)-Y• as a cofactor to generate the C• of the α subunit; class Ib β2 employs a dimaganese(III/III)-Y• cofactor, instead of diiron, and class Ic β2 uses a heterodinuclear Mn(IV)/Fe(III) without a tyrosyl radical [41-44].

1.4.2 Class Ia RNR

Among all three sub classes of RNRs, class I metallocofactor formation is best understood for the Ia enzymes from E. coli and mouse [45]. The tyrosyl radical of E. coli β2 (Ec β2) was first detected by UV/vis absorption spectrophotometry and electron paramagnetic resonance (EPR) spectroscopy in 1970s [46-47]. Following that discovery, the assembly of E. coli, Saccharomyces cerevisiae and mouse

β2-Fe2(III/III)-Y• has been studied extensively by an array of biophysical methods such as stopped flow ultraviolet/visible (SF UV/vis) spectroscopy, rapid freeze-quench (RFQ) EPR, electron-nuclear double resonance (ENDOR), Mössbauer, and extended X-ray absorption fine structure (EXAFS) spectroscopies [45]. In vitro, the apo Ec β2 loaded with Fe(II) reacts with O2 rapidly, forming a peroxo-Fe2(III/III) complex similar to the Hperoxo (P) in sMMO. This intermediate was revealed by studying the variants of Ec β2 and wild-type mouse (Mus musculus) β2 [48-50]. A near surface tryptophan (W48) reduces the peroxo-Fe2(III/III) complex to a mixed-valent

•+ µ-oxo-Fe2(III/IV) species termed X and forms a W48 cation radical (W48 ) [51].

One more electron that may come from Y356 through a Y• intermediate rapidly reduces the W48•+ to W48 [52-53]. X has an S = 1/2 ground state and a nearly isotropic g = 2 EPR signal [54-55]. X oxidizes the proximal Y122 to a tyrosyl radical

(Y122•) through a proton couple electron transfer (PCET) process and itself is reduced to the product µ-oxo-Fe2(III/III) cluster. The half life of the

µ-oxo-Fe2(III/III)-Y122• exhibits organism dependency (e.g. 20 min and 4 days in human and E. coli β2s, respectively). These are significantly longer than the half life of Y• in solution, which is less than 1 ms [41, 43]. The reaction mechanism of Ec

RNR is summarized in Scheme 1.3.

1.4.3 Class Ic RNR

McClarty and co-workers first cloned, expressed and purified the α2 and β2 of C. trachomatis (Ct) RNR [56]. Amino acid sequence alignments of the Ct RNR subunits with those of other class I RNRs revealed that, despite conservation of most of the catalytically critical residues, two key residues differ: a) the aspartate ligand to Fe1

(D84 in Ec β2) is substituted by glutamate (E89 in Ct β2), and b) the radical-harboring tyrosine residue (Y122 in Ec β2) is replaced by a non-oxidizable residue, phenylalanine (F127 in Ct β2). The RNR from Ct is catalytically active and the same

D to E and Y to F substitutions are observed in the β2 subunits from other organisms such as Tropheryma whipplei. These RNRs lacking the tyrosine residue required for cofactor formation have been classified as class Ic RNR [57]. Our research group discovered that the Ct β2 has maximum in vitro activity in the presence of stoichiometric amounts of both Mn and Fe. EPR and Mossbauer spectroscopic data showed that the radical initiation by Ct β2 involves a novel stable Mn(IV)/Fe(III) cofactor, instead of a Fe2(III/III)-Y• as in class Ia β2 [58]. Essentially the Ct β2 Mn(IV) replaces the function of the Y• to initiate radical chemistry in class Ia β2s catalysis

[58]. The oxidation states of the Mn and Fe of this cofactor were assigned by combining EPR and Mössbauer analysis. Dithionite chemically reduces the EPR silent Mn(IV)/Fe(III) cofactor to the EPR-active Mn(III)/Fe(III) state. Treatment of the active holoenzyme with the substrate analog,

2'-azido-2'-deoxyadenosine-5'-diphosphate (N3-ADP), leads to the formation of the

EPR-active Mn(III)/Fe(III) state and the formation of a nitrogen-centered radical. This evidence proved that the Mn(IV)/Fe(III) form of the Ct β2 is indeed the active form

[58]. Later, a Mn(IV)/Fe(IV) intermediate in cofactor assembly was also discovered

14 and spectroscopically characterized by Jiang, et al.. The complex forms upon reaction of the Mn(II)/Fe(II) Ct β2 with O2. Antiferromagnetically coupled Mn(IV) (SMn = 3/2) and Fe(IV) (SMn = 2) result in a S = 1/2 electronic ground state and g = 2 EPR signal

[58]. This enzyme provides the very first example of Mn/Fe redox cofactor in any enzyme. The mechanism and intermediates of class Ic RNR are summarized in scheme 1.3.

1.4.4 Class Ib RNR

E. coli contains two class I RNRs that are expressed under different aerobic conditions. The class Ib RNR, composed of NrdE (α) and NrdF (β), is expressed when the bacterium is oxidatively stressed and grown under conditions of iron limitation

[59-62]. Similar to class Ia RNR, in class Ib RNR, the nucleotide reduction occurs at the homodimeric α2 subunit and the metallocofactor containing homodimeric β2 subunit initiates the reaction. Cotruvo, et al. recently reported that the class Ib β2 subunit contains a dimanganese(III/III)-Y• instead of diiron(III/III)-Y• (in class Ia β2) as the cofactor [63]. This conclusion is also supported by studying endogenous E. coli class Ib RNR by turning off all five known Fe uptake pathways in vivo [45]. Although in E. coli, the class Ib RNR is expressed under iron limitation and oxidative stress, class Ib RNR is the only RNR for some prokaryotes such as Staphylococcus aureus, and Bacillus anthracis [64]. Usually, in addition to the NrdE (α) and NrdF (β) subunits, NrdI and NrdH are also encoded on the same operon [59]. NrdI, a flavodoxin with novel redox properties of E1 = -264 +/- 17 mV and E2 = -255 +/- 17 mV (redox potentials for the transfer of each electron respectively) plays a key role in the biosynthesis of E. coli class Ib RNR [65]. This NrdI is also suggested to be essential in other organisms such as Streptococcus pyogenes and B. subtilis [66-67].

Studies show that the reduced form of NrdI (denoted as NrdIhq) reduces oxygen to

- HOO or H2O2, which is then channeled via the NrdI/NrdF complex to the dimetal site to oxidize the dimanganese(II/II) cluster and generate the Y• [63]. A proposed mechanism (Scheme 1.4) invokes that two HOO- being produced to oxidize dimanganese(II/II) complex through dimanganese(III/III) intermediate to form dimanganese(IV/IV) species [64]. One electron is provided by a nearby tryptophan

(Trp31) to generate the Mn(IV)/Mn(III) intermediate similar to class Ia X, and the complex oxidizes the nearby Y to form the active dimanganese(III/III)-Y• cofactor and complete the activation process [64].

Scheme 1.3: Detailed reaction mechanism of E. coli β2 and Ct β2. Adapted from ref

[68].

Scheme 1.4: Proposed reaction mechanism of β2 subunit of class Ib RNR. Adapted from ref [63].

1.5 Plant soluble Stearoyl-Acyl Carrier Protein Δ9 Desaturase

Stearoyl-Acyl Carrier protein Δ9 desaturase (Δ9 D) is a ferritin-like diiron-carboxylate enzyme. It functions to insert a cis double bond between C-9 and C-10 of stearoyl-acyl carrier protein to produce oley-ACP, a key intermediate in the biosynthesis of unsaturated cellular lipid. The enzyme is highly stereo- and regioselective [69-71]. Similar to sMMO, the complete catalytic process of Δ9 D requires multiple components including: Δ9 D, ferredoxin, ferredoxin reductase,

NADPH and acyl-ACPs (substrates) [71]. The substrate-Δ9 D interaction has been investigated by binding competition assay using fluoresceinyl-ACP and unlabeled acyl-ACP substrates [70]. The results indicate a strong chain length preference of

18:0-ACP compared to 16:0-ACP. This interesting observation is also shown in my study of substrate dependent oxygen activation catalyzed by aldehyde decarbonylase

(AD) which will be described later in detail. Mossbauer and EXAFS data show that as

9 isolated Δ D contains ~70% μ-oxo-diion(III/III) species (δ = 0.54 mm/s, ΔEQ = 1.53 mm/s) and ~21% μ-hydroxo-diion(III/III) species (δ = 0.49 mm/s, ΔEQ = 0.72 mm/s) which is similar to as-isolated AurF (vide supra) [72]. As-isolated Δ9 D can be chemically reduced with sodium dithionite to diferrous Δ9 D, which can activate molecular oxygen in the presence of acyl-ACP to form a stable bright-blue complex

(an absorption band centered at ~700 nm) [72]. Mössbauer and resonance Raman studies revealed that this complex is a μ-1, 2-peroxo-Fe2(III/III) species, termed

9 9 peroxo-Δ D [73]. This peroxo-Δ D is stable (t1/2 for decay ~30 min at room temperature) and decays to diiron(II/II) form without double bond insertion into the added substrate [73].

1.6 Toluene/o-xylene monooxygenase (ToMO)

Toluene/o-xylene monooxygenase (ToMO) catalyzes the oxidation of toluene to a variety of products. The ferritin-like diiron-carboxylate hydroxylase component

ToMO (ToMOH), of this multicomponent complex is responsible for the hydroxylation reaction [74]. An intermediate in the O2 activation sequence was first discovered in a variant in which an isoleucine in the active site pocket was replaced with a tryptophan (I100W) [13]. When chemically reduced diiron(II/II) ToMOH was reacted with O2 in the presence of its coupling component, ToMOD, a transient band centered at ~500 nm was observed by stopped-flow absorption spectroscopy.

Freeze-quench EPR and Mössbauer studies showed that it is a Fe2(III/IV)-W• complex, similar to that observed previously in Ec RNR β2 [13]. A peroxo-Fe2(III/III) precursor to the Fe2(III/IV)-W• complex was trapped in the reaction of

ToMOH-I100W [75]. Soon after that, wild-type ToMOH was reported to form a peroxo-Fe2(III/III) intermediate which is competent to convert substrate to product

[75]. The lack of optical absorption is one of the features of this intermediate [75].

Interestingly, by comparing the 3 dimensional structures of ToMOH and sMMO, a threonine 201 (T201) residue was identified as potentially important in control of the nature of the peroxide complex that forms in ToMOH. This residue was substituted with serine by site-directed mutagenesis, which altered the optically transparent peroxo-Fe2(III/III) intermediate in favor of a complex possessing an absorption band at around 700 nm region [76]. Later, more T201 variants, such as T201G and T201C, were studied, and they also exhibited optical absorption near 700 nm, allowing the kinetics to be dissected by stopped-flow absorption experiments [77]. The fact that substitution of key residues can alter the spectroscopic property of the

21 peroxo-Fe2(III/III) intermediates inspired us to attempt a similar strategy to tune the peroxo-Fe2(III/III) intermediates in the reactions of AurF and cAD (vide infra).

1.7 AurF

1.7.1 The function of AurF and the debate on its metal cofactor

The N-oxygenase AurF from Streptomyces thioluteus converts 4-aminobenzoate

(Ar-NH2) to 4-nitrobenzoate (Ar-NO2), a building block for the biosynthesis of an antibiotic aureothin (scheme 1.5 A,B) [78-79]. It was proposed that this net six-electron oxidation proceeds via three consecutive, two-electron oxidation, through

4-hydroxylaminobenzoate (Ar-NHOH) and 4-nitrosobenzoate (Ar-NO) intermediates, requiring one equivalent of O2 and two exogenous reducing equivalents for each step

(scheme 1.5 C) [80]. Zhao’s group identified the EX28-37DEXXH pattern from the protein sequence of AurF and suggested that AurF may belong to the ferritin-like dimetal oxygenase enzyme family [9]. They reported EPR spectra resembling either mixed-valent diiron(II/III) or class Ic β2 Mn(III)/Fe(III) species while analyzing as-isolated AurF purified from Luria-Bertani medium [9]. Expression of AurF in E. coli grown in minimum medium supplemented with an overabundance of iron or manganese can alter the metals residing in the dimetal center, and the activity assay suggested that iron supplementation led to the production of Ar-NO2 whereas manganese resulted in no detectable activity [80]. Zhao’s group therefore claimed that

AurF is a diiron enzyme, and they explained that the incorporation of Mn into the dimetal center as a result of the relatively high levels of Mn in Luria-Bertani medium

18 18 [9]. O2 isotopic labeling activity assay showed 100% one O atom incorporation into the final production, Ar-NO2 [9]. To rationalize this observation, they proposed the dehydrogenation of 4-hydroxylaminobenzoate (Ar-NHOH) to 4-nitrosobenzoate

(Ar-NO) followed by oxidation to Ar-NO2 by AurF [9, 80]. Zhao’s work established that: AurF is a ferritin-like diiron-carboxylate oxygenase; the metal content can be

23 manipulated by expressing AurF in E. coli grown in minimal medium with various metal supplement; and a Mn/Fe EPR signal is detected from the as isolated AurF.

Hertwick et al studied the chemo and regioselectivity of the AurF reaction [81]. Their data suggested that the reactive substrate requires the amino group being at the para position; ortho- or meta- amino group is not reactive. AurF can oxidize the substrate analogs when the carboxyl group is replaced by SO3H- and CH3COOH-, but not CH3- and OCH3-. This suggests that the carboxyl group may serve as an anchor for substrate binding because some flexibility is allowed but a certain resemblance has to be maintained [81]. Substitution with hydroxyl or methyl groups on the benzene ring

(ortho-, meta-) allows reaction to proceed, which is important for mechanistic study

(vide infra) [81]. Hertwick also reported that immobilized AurF was catalytically active with H2O2, as oxidizing co-substrate [81]. This so-called peroxide shunt, well-known for other iron oxygenases, permits AurF reaction to be carried out in the absences of a reducing system [81]. Indeed, the natural reducing system for AurF still has not been identified. The crystal structure reported by Hertwick’s group indicated that AurF is indeed a ferritin-like dimetal-carboxylate protein. Hertwick claimed that

AurF employs a dimanganese cofactor, as they observed manganese in the crystal of

AurF [81]. Activity assays were also performed by using hydrogen peroxide. The problem with this observation is that the AurF contains both iron and manganese and it is not certain that the activity was caused by either individual metal or both together, as the authors themselves reported in the paper [81].

At that time, Dr. Jiang from our group had just discovered the class Ic Ct RNR β2 employing the heterodinuclear Mn/Fe cofactor to activate O2 in [58, 82]. The EPR spectrum of the Mn(III)/Fe(III) species of Ct β2 is similar to the one reported by

Zhao’s group in AurF [58, 83]. Our group therefore proposed that AurF might employ this Mn/Fe cofactor to activate oxygen, and perform its N-oxygenation reaction [84]. I started to work towards this direction for the first two years of my Ph.D. study and developed a method to manipulate the metal content of the enzyme produced in E. coli. EPR spectra showed that sodium dithionite-reduced Mn/Fe-AurF is able to react with O2, and the spectra are different in the absence or presence of substrate. The problem with this study is that metal chelation and reconstitution did not work with

AurF, and pure heterodinuclear Mn/Fe-AurF was not feasible to produce. Fortunately, the Mn/Fe cofactor is EPR active. To obtain a pure Mn/Fe-AurF would be the key to test the idea that the heterodinuclear form of the protein might also be active.

Scheme 1.5: AurF reaction and proposed mechanisms. Adapted and modified from ref

[9, 85].

1.7.2 Diiron AurF and the peroxo intermediate discovered by our group

Soon after we proposed the hypothesis above, Zhao’s group reported that diiron AurF is competent to convert 4-aminobenzoate (Ar-NH2) to 4-nitrobenzoate (Ar-NO2) in vitro [80]. The diiron-AurF was obtained by expressing AurF in E. coli grown in minimal medium supplemented with Fe(NH4)2(SO4)2, as they claimed that they were not able to make apo-AurF by metal chelation and therefore unable to perform metal dependence analysis by reloading metals back into AurF [80]. Their ICP-MS analysis showed a 2 to 1 ratio of iron to AurF, indicating that the diiron-AurF was fully loaded by iron. Phenazine methylsulfate (PMS)/ascorbate can support the turnover in vitro, and parallel-mode EPR spectra showed the reduction of diiron AurF by ascorbate and ferromagnetic coupling of the two Fe(II) ions to give a g = 16 EPR signal [80]. With surrogate ferredoxin, ferredoxin reductase from Anabaena sp. PCC 7119 and NADPH

(F/FR/N) reducing system or phenazine methylsulfate/NADH (PMS/N) reducing system, Zhao’s group reported the steady state rate constants for oxidizing

-1 4-aminobenzoate: for F/FR/N, kcat = 6.21 ± 0.52 min , Km = 5.24 ± 0.64 µM, and

-1 -1 -1 kcat/Km = 1.21 ± 0.31 min µM at 20 °C or for PMS/N, kcat = 5.04 ± 0.22 min , Km =

-1 -1 8.89 ± 0.87 µM, and kcat/Km = 0.57 ± 0.03 min µM at 20 °C [80]. They also reported the crystal structures of AurF with the diiron cofactor and with or without product bound. By LC/MS, they detected 4-hydroxylamine (Ar-NHOH) and

4-nitrosobenzoate (Ar-NO) as putative intermediates [80]. Although some of these data we now know were not accurate, they clearly showed that diiron AurF is capable of converting Ar-NH2 to Ar-NO2 product. Thus, Dr. Korboukh in our group started to study the reaction mechanism of diiron AurF in parallel with my studies on

Mn/Fe-AurF.

Preparation of diiron AurF for analysis by Mössbauer spectroscopy was done by expressing the protein in E. coli grown in minimal medium supplemented with

57 FeSO4. The UV/vis absorption spectrum of as-isolated AurF exhibits an absorption band at 360 nm similar to other µ-oxo diiron proteins. This band has been attributed to an oxo-to-iron charge transfer transition [15]. The 4.2-K/53-mT Mössbauer spectrum of as-isolated AurF can be simulated as two quadruple doublets with high spin ferric iron parameters of δ1 = 0.54 mm/s, ΔEQ, 1 = -1.86 mm/s (78%); and δ2 = 0.48 mm/s,

ΔEQ, 2 = 0.80 mm/s (22%) [15]. We believe that the species with larger |ΔEQ1| = 1.86

III/III mm/s is the µ-oxo-Fe2 -AurF and the other species with smaller |ΔEQ2| = 0.80

III/III mm/s is probably the µ-hydroxo-Fe2 -AurF [15]. As-isolated AurF can be chemically reduced by sodium dithionite, resulting in a decrease of UV/vis absorption at 360 nm. The reduced protein exhibits a 4.2-K/53-mT Mössbauer spectrum of one major quadruple doublet with δ = 1.24 mm/s, ΔEQ = 3.06 mm/s, typical of high spin

II/II Fe2 species with oxygen and nitrogen coordination [15]. This reduction process is relatively slow (requires >30 minutes’ incubation at ~20 °C). However, this reduction

II/II by dithionite is stoichiometric, and the resulting Fe2 -AurF can activate O2 forming a new species with an absorption band centered at 500 nm [15]. Formation of this species is complete within 0.01 s at 20 °C, and the decay is slow (t1/2 ~ 7 min at room temperature) [15]. At the time, the corresponding Mössbauer sample was made by

II/II stirring the Fe2 -AurF under pure oxygen gas for 2 min at room temperature before freezing in liquid nitrogen and acquisition of the Mössbauer spectrum (We later knew that rapid mixing of the diferrous AurF with oxygen-saturated buffer resulted in

II/II III/III samples of higher quality.). In addition to 5% Fe2 and 18% Fe2 species, the

4.2-K/53-mT Mössbauer spectrum of this sample revealed the formation of a new

28 species. Its spectrum could be simulated as two quadruple doublets with parameters δ1

= 0.54 mm/s, ΔEQ, 1 = -0.66 mm/s (49%), and δ2 = 0.61 mm/s, ΔEQ, 2 = 0.35 mm/s

(33%), typical of high-spin Fe(III) [15]. The 1.5 to 1 ratio between these two similar new species was proposed as a result of the equilibrium of more than one isomeric intermediate forms, which we later knew is likely due to the site resolution of the two irons in the intermediate complex. The ratio is indeed very close to 1 to 1 in our later studies when produced with freeze-quench method [85]. Dr. Korboukh also showed that this stable intermediate can react with 4-aminobenzoate rapidly (reaction completed within 10 ms at 20 °C) which was also confirmed by Mössbauer spectroscopy [15]. On the basis of its optical and Mössbauer properties and its

III/III reactivity with Ar-NH2, we assigned this new species as a peroxo-Fe2 -AurF

III/III III/III intermediate [15]. Peroxo-Fe2 -AurF was converted to Fe2 -AurF upon oxidation of Ar-NH2, suggesting a two electron oxidation process. We also noted that

III/III one equivalent of peroxo-Fe2 -AurF is only able to convert ~0.3 equivalent of

III/III Ar-NH2 to Ar-NO2 [15]. This led us to propose that peroxo-Fe2 -AurF is responsible for the oxidation of Ar-NH2 to the final product, Ar-NO2 and that the reaction is a 6-electron oxidation. This hypothesis inspired us to perform further

III/III studies on Ar-NHOH reaction with the peroxo-Fe2 -AurF intermediate, leading to our discovery of the novel 4-elecron oxidation process [85].

1.7.3 The discovery of 4 electron oxidation of 4-hydroxylaminobenzoate

Right before Dr. Korboukh left our research group, she discovered that

III/III peroxo-Fe2 -AurF is able to oxidize the 4-hydroxylaminobenzoate (Ar-NHOH) and some portion of the peroxo species is seemingly regenerated after this process. A problem at that time was that the synthesized Ar-NHOH had ~15% 4-aminobenzoate

(Ar-NH2) impurity, and we proposed that the peroxo regeneration would occur only when reacting with Ar-NHOH (vide infra) and that the 15% Ar-NH2 impurity would prevent complete regeneration. The purification of this compound by HPLC was suggested by Dr. Booker and method development was guided by his graduate student,

Tyler Grove. Ar-NHOH reached 99% purity, which was breakpoint in the studies of

III/III peroxo-Fe2 -AurF reacting with Ar-NHOH. Using stopped-flow absorption spectroscopy and freeze-quench Mössbauer spectroscopy, we showed that

III/III peroxo-Fe2 -AurF is capable of oxidizing Ar-NHOH [85]. When limiting O2, no

III/III regeneration of peroxo species was observed when peroxo-Fe2 -AurF was reacted with Ar-NHOH. In the presence of excess O2 and stoichiometric Ar-NHOH, complete regeneration of the peroxo species was observed [85]. The presence of excess O2 and

Ar-NHOH (two equiv. relative to peroxo-AurF) abolished the reformation.

Surprisingly this Ar-NHOH intermediate is directly oxidized to 4-nitrobenzoate

(Ar-NO2), apparently without forming the previously proposed 4-nitrosobenzoate

II/II (Ar-NO) intermediate. The AurF is reduced to the Fe2 form when O2 is limiting

[85]. I also developed a HPLC and LC/MS method through which we were able to quantify the small-molecule substrate and products as well as to identify them by m/z values. This 4-electron single-step oxidation of Ar-NHOH to Ar-NO2 is

II/II unprecedented. The reduction of the peroxide complex all the way to the Fe2 state

III/III indicates that oxidation of Ar-NHOH is catalytic, as the peroxo-Fe2 -AurF is

II/II regenerated after each turnover. Indeed, our data showed that Fe2 -AurF is able to oxidize 30 equiv. Ar-NHOH in the presence of sufficient O2 without external electron source [85]. This study unveiled the complete reaction mechanism of AurF converting the amino group to the nitro group (scheme 1.5 D).

1.7.4 The study of 4-hydrazinobenzoate analog reactivity

The hydrazine Ar-NHNH2 is structurally similar to Ar-NHOH, with the hydroxyl group (-OH) replaced by an amino group (-NH2). The substitution affects the reduction potential, making hydrazine (-NHNH2) a much better reductant. According to the literature, Ar-NHNH2 is able to provide 2 electrons when oxidized to diazene, which will decompose to benzoate and N2. We thus became interested in the outcome of reacting AurF with Ar-NHNH2 in the hope that we could uncover some aspects

III/III hidden in the one-step, 4-electron oxidation of Ar-NHOH by peroxo-Fe2 -AurF.

Our studies show that, to our surprise, Ar-NHNH2 reacts rapidly with both as-isolated

III/III III/III Fe2 -AurF and peroxo-Fe2 -AurF. In the reduction of as-isolated AurF by

II/II Ar-NHNH2, Fe2 -AurF is produced and stoichiometric substrate is required to

III/III II/II completely convert Fe2 -AurF to Fe2 -AurF. This suggested that indeed two electrons are provided by one molecule of Ar-NHNH2 in this reaction. Upon reduction

III/III III/III of peroxo-Fe2 -AurF by Ar-NHNH2, Fe2 -AurF is accumulated, suggested by stopped-flow experiments as well as revealed by Mössbauer spectroscopy. Excess

III/III Ar-NHNH2 in the reaction solution successively reduces the resulting Fe2 -AurF to

II/II III/III Fe2 -AurF, readily to generate another peroxo-Fe2 -AurF in the presence of O2.

The oxidation of Ar-NHNH2 by AurF is therefore catalytic in the presence of sufficient O2. We proposed that the direct oxidation product of Ar-NHNH2 by AurF is

Ar-N=NH, which will then be converted nonenzymatically to benzoate and N2 [86-91].

Indeed, benzoate is detected as the only major product in reaction solution containing

10 μM as-isolated AurF and 800 μM Ar-NHNH2 and exposed to air. The formation of benzoate requires abstraction of a hydrogen from solvent by its precursor. To further confirm our hypothesis that Ar-N=NH was produced, a similar reaction was carried

2 out in H2O enriched solution and the small-molecule products were analyzed by

LC/MS. In addition to the peak at m/z 121 for benzoate, a peak at m/z 122 was

2 2 detected, indicating the incorporation of H into the benzoate. In an 85% H2O enriched reaction solution, we detected a maximum of 33% Ar-2H, suggesting a preference of the benzoate precursor to abstract 1H versus 2H when forming the benzoate. Our studies suggested that, despite of the structural similarity between the

Ar-NHOH and Ar-NHNH2, the reaction mechanisms of AurF with these substrate and analog are different. A previously observed one-step, 4-electron oxidation of

III/III Ar-NHOH by peroxo-Fe2 -AurF is remodeled to a two-step, 2-electron oxidation

III/III of Ar-NHNH2 by peroxo-Fe2 -AurF by replacing the –OH group with a –NH2

III/III moiety, in which Fe2 -AurF is accumulated (Scheme 1.6). The actual reaction is potentially more complicated. Among the small-molecule products detected by

18 LC/MS, a peak at m/z 166 for Ar-NO2 was also detected, and in the presence of O2, the peak at m/z 170 is the only species detected for Ar-NO2. This result indicates that both oxygen atoms of the nitro group are coming from molecular oxygen. Thus, oxygenative hydroxylation cleavage of the hydrazine moiety may occur as well. The

Ar-NO2 product accounts repeatedly less than 5% of the total product, and a method to alter the reaction in favor of producing higher percentage of Ar-NO2 is required for further investigation. Our Mössbauer data also revealed a new species of

III/III μ-hydroxo-Fe2 species being produced upon Ar-NHNH2 reduction of

III/III III/III peroxo-Fe2 -AurF. It was seemingly converted to μ-oxo-Fe2 over longer incubation time spontaneously. This observation provides an evidence for the protonation step previously proposed for the oxidation of Ar-NHOH by

III/III III/III peroxo-Fe2 -AurF [85]. This μ-hydroxo-Fe2 species formed after 1 s and

33 decayed after 6 min. A finer time course is therefore required to understand this

III/III complex better. It would also be interesting to test if the similar μ-hydroxo-Fe2

III/III species is formed in the reduction of Fe2 -AurF by Ar-NHNH2. In summary, our studies provide the evidence for the reaction of AurF with Ar-NHNH2, dissect the reaction mechanism and expand the repertoire of AurF products.

Scheme 1.6: Proposed reaction mechanism of 4-hydrazinobenzoate with AurF.

1.8 Cyanobacterial aldehyde decarbonylase (cAD)

1.8.1 The discovery of formate as the co-product

Alkanes are found in various organisms including plants, mammals and insects. These inert and highly hydrophobic molecules have important functions (e.g. waterproofing in plant cell wall and signaling molecules for insects) [92-94]. The alkane production enzymes have been discovered in Pisum sativum, Podiceps nigricollis, braunii,

Sarcophaga crassipalpis, etc. [92-97]. However, little is known about the microbial alkane biosynthesis. Schirmer, et al. from LS9 Inc. discovered a pair of cyanobacterial enzymes that convert long chain fatty acids through a fatty aldehyde intermediate to alka(e)nes (scheme 1.7 A) [10]. The first enzyme is an NADPH dependent acyl-ACP reductase (AAR). The second enzyme catalyzes the removal of the C1 carbonyl group, converting the fatty aldehyde to alka(e)ne. It was reported that this process requires a reducing system such as ferredoxin, ferredoxin reductase, and NADPH (N/F/FR) [10].

The crystal structure of one of the AD orthologs was available and indicated that cADs belong to the ferritin-like dimetal-carboxylate protein family [10]. The identity of the metal cofactor was not definitely determined by Schirmer, et al.. The structure of the cADs and its requirement for a reducing system would indicate a redox reaction.

However, the conversion of aldehyde to alkane and CO as proposed by Schirmer, et al. is redox neutral. We started to work on these two enzymes to try to understand the controversy. We began with the exploration of the C1 co-product from this reaction.

Dr. Warui from Prof. Booker’s group synthesized the substrate 1-[13C]-octadecanal and obtained in vitro activity of Nostoc punctiforme (Np) AD expressed in E. coli grown in Rich Luria-Bertani medium. Np AD was shown to be the most active ortholog in vitro among the tested cADs [10]. As reported by Schirmer, et al., a

36 reducing system is required and the reaction is performed aerobically [10, 98]. Of the four possible one carbon products, carbon monoxide, formaldehyde, formate and carbon dioxide, I first tested for CO using myoglobin as that was proposed but not demonstrated by Schirmer et al. Upon binding of CO, reduced myoglobin would exhibit a shift in its Soret band from ~437 nm to ~423 nm, which makes it a very sensitive indicator. Standard spectra were generated by using a CO solution with known concentration and our myoglobin method was shown to be able to detect as little as 5 µM CO in solution [98]. In a reaction solution with ~80 µM heptadecane being produced (as quantified by GC/MS), CO was not detected, indicating that CO might not be the co-product [98]. CO2 would be the second reasonable candidate to test for since this enzyme belongs to the oxygenase/oxidase category and the reaction

14 requires a reducing system. However, in order to assay for CO2, C-labeled substrate would have been required. We therefore tested first for formate. To increase the sensitivity of its detection, formate is generally derivatized, for example, by

2-nitrophenolhydrazine in the presence of 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC) and then quantified by HPLC or LC/MS [99]. Our initial tests showed enzyme-dependent formate production. However, due to the presence of formate from the environment, 1-13C or 1-2H labeled substrate became the key to distinguish the formate produced by the cAD from the environmental contaminant. Dr.

Warui synthesized 1-[13C]-octadecanal, 1-[2H]-octadecanal and 1-[13C,

2H]-octadecanal, making the accurate measurement by LC/MS possible by providing for distinction of AD-produced formate on the basis of a shift in m/z by 1 or 2 units.

Out data unequivocally showed the production of formate during the cAD-catalyzed reaction [98]. Furthermore, by quantifying the heptadecane yield using GC/MS (done

37 by Dr. Warui) and the formate yield using LC/MS in parallel, we show that they are identical within experimental error, indicating that formate is the sole co-product formed under the employed assay conditions [98].

Scheme 1.7: Three possible reactions catalyzed by cAD. Adapted from ref [98].

1.8.2 The discovery of oxygen dependence

After establishing formate as the C1-derived co-product, we went on to study the oxygen dependence. Anaerobic reactions yielded no formate above the level of environmental contamination and 18O isotope-tracer assay indicated that one oxygen atom is incorporated from molecular oxygen into the formate product [100]. What is surprising is oxygenation without oxidation suggests a potential hydrolysis pathway.

However, our result of oxygen incorporation requires the breaking of molecular oxygen bond which makes cAD an oxygenase. In the oxygenation reaction, one of the two oxygen atoms in the formate product should come from the aldehyde substrate, which is exchangeable with the solvent. The other one should come from O2. Ideally,

18 13 in the presence of O2, peaks at m/z = 181 for derivatized C-formate and at m/z =

183 for derivatized 13C, 18O-formate should both be detected and have an equal intensity. Our result showed a 16% less intensity of m/z = 183 peak than the theoretical value. We suspected that the analysis of formate by derivatization not only removed one of the oxygen, but also might exchange the other with the solvent, causing the loss of 16% peak intensity at m/z = 183 [98, 100]. In a control reaction,

13 18 18 we derivatized the C-formate standard in a H2 O solvent and detected 20% O incorporation into the derivatized formate [100]. We therefore concluded that, to the limit of our measurement, all the formate is produced by a cryptic oxygenation reaction with one oxygen atom incorporated from O2 [100]. This result is essential to further propose the mechanism of cADs’ reaction. As a matter of fact, two possible reaction pathways were proposed depending on the requirement of molecular oxygen

(Scheme 1.7 B, C). A detailed mechanism for each individual reaction is also proposed by carefully considering the source of the oxygen atoms in the formate

40 product (Scheme 1.8).

Scheme 1.8: Proposed mechanisms for formate production emphasizing on oxygen source.

1.8.3 The discrepancy between oxygen dependent and independent mechanism

About the same time, Marsh’s group at the University of Michigan claimed that cADs can catalyze anaerobic aldehyde cleavage with production of formate and alka(e)ne reaction [101-102]. We realized how novel it would be if cAD could be functioning via two fundamentally different mechanisms when catalyzing the same conversion of aldehyde to alka(e)ne and formate. However, the attempt to repeat their results failed.

Three possibilities may explain this discrepancy: 1) Marsh’s group used Pm AD instead of Np AD (Although later on, in their 2nd publication, they claim all ADs follow the anaerobic reaction mechanism.), and different AD orthologs might function differently; 2) They used a different reducing system of phenazine methylsulfate/NADH (PMS/N), and different reducing systems may result in different reaction mechanisms; and 3) The cAD activity detected in their work was the result of

O2 contamination. We evaluated these possibilities respectively.

To test whether Pm and Np orthologs function differently, we procured the Pm AD gene codon optimized for E. coli expression, synthesized and subcloned into a pet28a vector with T7 promoter and a N-terminal His6-tag upstream of Pm AD gene. Pm AD was then expressed in E. coli grown in Rich Luria-Bertani medium and purified using

Ni-NTA affinity chromatography. Activity assays were performed and the results showed O2 dependence for the Pm AD reaction when using N/F/FR reducing system and one oxygen atom is incorporated into the formate product from O2 (validated

18 using O2 as source of O2).

A similar test is carried out with the PMS/N reducing system with both Pm AD and

Np AD. The result also showed that both enzymes require O2 to make product and that an oxygen atom from O2 is incorporated into the formate product.

We realized that our anaerobic controls always gave 5-15 µM formate, which is due to the remaining oxygen in the deoxygenated buffer or diffusion of a trace amount of O2 into the glove box. However, this is the range of product yield reported by Marsh’s group, whereas we typically generated 75-100 µM formate and heptadecane in our aerobic reaction.

It is still possible that small portion of the reaction is anaerobic and it is hard to distinguish from oxygen contamination. The key solution is to track the source of oxygen in the formate product, by directly measuring formate without having to derivatize it by a procedure that results in loss of either oxygen on it. I therefore developed a direct formate analysis method by LC/MS. The combination of this analysis method with isotope-tracer assays provided a powerful tool to study the cAD reaction. We were able to prove that one oxygen atom from molecular oxygen is incorporated into >90% of the formate product. This conclusion is true for both Np

AD and Pm AD with either PMS/N or F/FR/N reducing system. We also noticed that

Marsh’s group tested heptanal, a seven carbon aldehyde. We had been performing all our assays with the 18 carbon chain substrate, octadecanal. It turned out that aerobically, this short chain substrate reacted very fast (~16 min-1 at room temperature), which makes the study of transient kinetics feasible (vide infra).

GC/MS data confirmed the production of equal quantities of hexane, verifying that

AD catalyzes the same type of reaction when using heptanal as a substrate as when using octadecanal. In carrying out these experiments, we replaced the PMS with methoxylphenazine methylsulfate (MeOPMS), because PMS is unstable even on ice while MeOPMS is stable for hours at room temperature [103]. The definitive test for

O2-independent aldehyde cleavage was to test for production of formate with both

44 oxygens originating from solvent (Dr. Chang from our group demonstrated by NMR spectroscopy method that the aldehyde oxygen is exchangeable with the solvent

18 oxygen and this process is complete in a few minutes.). In H2 O enriched solution, double 18O labeled formate was not detectable. This result not only shows that our preparation of Np and Pm ADs are incapable of catalyzing O2-independent aldehyde cleavage but also sets the analytic “gold standard” for claiming the existence of

18 18 O2-independent reaction in the future. The reaction in H2 O must result in double O labeled formate by direct formate LC/MS analysis for this activity to be established.

1.8.4 Transient kinetics study and metal cofactors

Our proposed mechanism suggested the existence of a peroxo intermediate (Scheme

1.9). After establishing the oxygen dependency, we moved on to a transient kinetics study of the cAD reaction and the dissection of the cAD reaction mechanism. We rationalized the faster reaction with shorter chain substrate is due to the increased solubility and therefore the accessibility of cAD to the substrate. The diiron-cAD

(Fe2-cAD) is capable of catalyzing the reaction no less effectively than cAD with other metal cofactors shown by our studies. Dr. Warui developed a method to purify cAD anaerobically, and he showed that cAD, expressed in E. coli grown in minimal medium supplemented with ferrous ammonium sulfate, remained in diferrous form from anaerobic purification. The next thing we attempted was to study the rate of

II/II II/II Fe2 -cAD reacting with O2. One striking result was that Fe2 -cAD alone reacts with O2 slowly, however, in the presence of substrate (heptanal, octanal, nonanal or decanal), the activation is fast and a new species with an absorption band centered at

450 nm accumulates rapidly (in a few seconds). This new species is likely to be photo labile, as the lifetime measured when employing white light and photodiode array

(PDA) detection is less than that measured using 450 nm light and photo multiplier tube (PMT) detection. Octadecanal did not have this effect on the O2 addition, presumably due to its insolubility which resulted in its inaccessibility to cAD, even in the presence of a detergent such as triton-x100. The newer, smaller substrates

(hexanal, heptanal, octanal, nonanal and decanal) are all first dissolved in dimethylsulfoxide (DMSO), and their solubility is less problematic when compared to longer substrates (dodecanal, tetradecanal, hexadecanal and octadecanal). Substrate analogs of short-chain alcohol, alkane, and acid were tested and showed no O2

46 addition effect, indicating the requirement of an aldehyde moiety.

Kinetic traces monitoring the absorption at 450 nm clearly demonstrated the substrate

II/II dependence (± decanal) of O2 activation of Fe2 -cAD. The aldehyde moiety is essential for this process as alkane, alcohol or acid functionalities are not able to trigger O2 addition. Another interesting observation is the substrate-chain length dependence of the triggering efficacy. From the kinetic traces, a positive correlation between the carbon chain length and the intermediate formation rate was observed up to 10 carbons (decanal). Mossbauer spectroscopy was employed to analyze this intermediate, and our preliminary data indicate an accumulation of two new quadruple doublets with parameters typical of high spin Fe(III) complexes. The product yield

(formate) from the decay of this intermediate is 13% in the absence of reducing system (likely caused by electrons from nearby Fe(II)) and 66% in the presence of stoichiometric MeOPMS/N. We therefore assign this intermediate as a

III/III peroxo-Fe2 -cAD species (The two quadruple doublets likely represent the peroxo-Fe with or without binding of aldehyde.). To further study the other potential intermediate of reaction sequence, the peroxo intermediate was allowed to accumulate for 30 s before mixing with a reducing system such as reduced MeOPMS, or dithionite-reduced ferredoxin. By monitoring the absorption at 386 nm, signature of oxidized MeOPMS, we observed that the oxidation of reduced MeOPMS by

III/III peroxo-Fe2 -cAD was ~100 fold faster than the natural decay of

III/III peroxo-Fe2 -cAD. A control reaction in which the initial aging time was only 0.1s,

III/III not long enough for the peroxo-Fe2 -cAD to accumulate yielded oxidation of reduced MeOPMS at a much lesser rate (~20 fold slower). Similar sequential-mixing

SF-Abs experiments were performed with another reducing system,

47 dithionite-reduced ferredoxin. Unfortunately, the oxidized ferredoxin absorbs in the

III/III same region as the peroxo-Fe2 -cAD intermediate. As a result, the spectral changes around 450 nm were canceled out during the oxidation of ferredoxin by

III/III peroxo-Fe2 -cAD. However, the fast spectra change at 320 nm still provides some estimation of the time scale of reduction (t1/2 ~ 0.1 s). The reduction of as-isolated

III/III (Fe2 ) cAD has been studied as well. It seems that dithionite-reduced ferredoxin reduces AD one magnitude faster than NADH-reduced MeOPMS and as for

MeOPMS, the reduction rate of cAD is not much affected by the presence of substrate.

In summary, we now have the tools to study the detailed mechanism and intermediates of the reaction of cAD.

1.8.5 Ongoing study of AD

We have propose a radical mechanism for the cAD reaction, and Dr. Chang synthesized two substrate analogs designed to stabilize the alkyl radical intermediate produced during the reaction. Multiple variants were made to study the potential electron transfer pathway of cAD. A DNA fragment encoding a hexa-His tag was recombined into the C-terminal of the cAD gene on Synechocystis sp. PCC 6803 genomic DNA. The purification of natively expressed Synechocystis sp. PCC 6803

AD can hopefully facilitate the verification of the presence of a diiron cofactor in vivo.

Additional experiments to identify the biological relevant reductant for the reaction by fractionation of cyanobacterial extracts are also in progress.

Scheme 1.9: Proposed reaction mechanism of cAD. Adapted from ref [98].

1.9 References

1. Villafranca, J.J., Metal-assisted catalysis. Current Opinion in Structural

Biology, 1991. 1(5): p. 821-825.

2. Nordlund, P., B.M. Sjoberg, and H. Eklund, Three-dimensional structure of the

free radical protein of ribonucleotide reductase. Nature, 1990. 345(6276): p.

593-8.

3. Nordlund, P. and H. Eklund, Structure and function of the Escherichia coli

ribonucleotide reductase protein R2. J. Mol. Biol., 1993. 232: p. 123-164.

4. Rosenzweig, A.C., et al., Crystal structure of a bacterial non-haem iron

hydroxylase that catalyses the biological oxidation of methane. Nature, 1993.

366(6455): p. 537-43.

5. Fox, B.G., et al., Resonance Raman Evidence for an Fe-O-Fe Center in

Stearoyl-ACP Desaturase. Primary Sequence Identity with Other Diiron-Oxo

Proteins. Biochemistry, 1994. 33(43): p. 12776-86.

6. Yen, K.M., Construction of cloning cartridges for development of expression

vectors in gram-negative bacteria. Journal of Bacteriology, 1991. 173(17): p.

5328-5335.

7. Nordlund, I., J. Powlowski, and V. Shingler, Complete nucleotide sequence

and polypeptide analysis of multicomponent phenol hydroxylase from

Pseudomonas sp. strain CF600. J. Bacteriol., 1990. 172(12): p. 6826-33.

8. Small, F.J. and S.A. Ensign, Alkene monooxygenase from Xanthobacter strain

Py2. Purification and characterization of a four-component system central to

the bacterial metabolism of aliphatic alkenes. J. Biol. Chem., 1997. 272: p.

24913-24920.

9. Simurdiak, M., J. Lee, and H. Zhao, A new class of arylamine oxygenases:

evidence that p-aminobenzoate N-oxygenase (AurF) is a di-iron enzyme and

further mechanistic studies. ChemBioChem, 2006. 7(8): p. 1169-1172.

10. Schirmer, A., et al., Microbial biosynthesis of alkanes. Science, 2010.

329(5991): p. 559-562.

11. Krebs, C., J.M. Bollinger, Jr., and S.J. Booker, The "di-iron-carboxylate"

proteins: An expanding repertoire of cofactor compositions and catalytic

capabilities. Curr. Opin. Chem. Biol., 2011: p.

doi:10.1016/j.cbpa.2011.02.019.

12. Murray, L.J., et al., Characterization of the arene-oxidizing intermediate in

ToMOH as a diiron(III) species. J. Am. Chem. Soc., 2007. 129(46): p.

14500-14510.

13. Murray, L.J., et al., Dioxygen Activation at Non-Heme Diiron Centers:

Characterization of Intermediates in a Mutant Form of Toluene/o-Xylene

Monooxygenase Hydroxylase. J. Am. Chem. Soc., 2006. 128(23): p.

7458-7459.

14. Tinberg, C.E. and S.J. Lippard, Dioxygen activation in soluble methane

monooxygenase. Acc Chem Res, 2011. 44(4): p. 280-8.

15. Korboukh, V.K., et al., A long-lived, substrate-hydroxylating

peroxodiiron(III/III) intermediate in the amine oxygenase, AurF, from

Streptomyces thioluteus. J. Am. Chem. Soc., 2009. 131(38): p. 13608-13609.

16. Krebs, C., J.M. Bollinger, Jr., and S.J. Booker, Cyanobacterial alkane

biosynthesis further expands the catalytic repertoire of the ferritin-like

'di-iron-carboxylate' proteins. Curr Opin Chem Biol, 2011. 15(2): p. 291-303.

17. Hanson, R.S. and T.E. Hanson, Methanotrophic bacteria. Microbiol Rev, 1996.

60(2): p. 439-71.

18. Westerheide, L., M. Pascaly, and B. Krebs, Methane monooxygenase and its

related biomimetic models. Curr Opin Chem Biol, 2000. 4(2): p. 235-41.

19. Deeth, R.J. and H. Dalton, Methane activation by methane monooxygenase:

free radicals, Fe-C bonding, substrate dependent pathways and the role of the

regulatory protein. Journal of Biological Inorganic Chemistry, 1998. 3(3): p.

302-306.

20. Liu, K.E. and S.J. Lippard, Studies of The Soluble Methane Monooxygenase

Protein System: Structure, Component Interactions, And Hydroxylation

Mechanism, in Advances in Inorganic Chemistry, A.G. Sykes, Editor. 1995,

Academic Press. p. 263-289.

21. Feig, A.L. and S.J. Lippard, Reactions of Non-Heme Iron(II) Centers with

Dioxygen in Biology and Chemistry. Chemical Reviews, 1994. 94(3): p.

759-805.

22. M. Valentine, A. and S. J. Lippard, Principles of small molecule activation by

metalloenzymes as exemplified by the soluble methane monooxygenase from

Methylococcus capsulatus (Bath) *. Journal of the Chemical Society, Dalton

Transactions, 1997(21): p. 3925-3932.

23. Wallar, B.J. and J.D. Lipscomb, Dioxygen activation by enzymes containing

binuclear non-heme iron clusters. Chem. Rev., 1996. 96: p. 2625-2657.

24. Merkx, M., et al., Dioxygen Activation and Methane Hydroxylation by Soluble

Methane Monooxygenase: A Tale of Two Irons and Three Proteins A list of

abbreviations can be found in Section 7. Angew Chem Int Ed Engl, 2001.

40(15): p. 2782-2807.

25. Gassner, G.T. and S.J. Lippard, Component interactions in the soluble

methane monooxygenase system from Methylococcus capsulatus (Bath).

Biochemistry, 1999. 38(39): p. 12768-85.

26. Lund, J., M.P. Woodland, and H. Dalton, Electron transfer reactions in the

soluble methane monooxygenase of Methylococcus capsulatus (Bath). Eur J

Biochem, 1985. 147(2): p. 297-305.

27. Lund, J. and H. Dalton, Further characterisation of the FAD and Fe2S2 redox

centres of component C, the NADH:acceptor reductase of the soluble methane

monooxygenase of Methylococcus capsulatus (Bath). Eur J Biochem, 1985.

147(2): p. 291-6.

28. Fox, B.G., et al., Methane monooxygenase from Methylosinus trichosporium

OB3b. Purification and properties of a three-component system with high

specific activity from a type II methanotroph. J Biol Chem, 1989. 264(17): p.

10023-33.

29. Davydov, A., et al., Radiolytic Reduction of Methane Monooxygenase

Dinuclear Iron Cluster at 77 K. Journal of Biological Chemistry, 1997.

272(11): p. 7022-7026.

30. Tinberg, C.E. and S.J. Lippard, Revisiting the mechanism of dioxygen

activation in soluble methane monooxygenase from M. capsulatus (Bath):

evidence for a multi-step, proton-dependent reaction pathway. Biochemistry,

2009. 48(51): p. 12145-58.

31. Valentine, A.M., S.S. Stahl, and S.J. Lippard, Mechanistic studies of the

reaction of reduced methane monooxygenase hydroxylase with dioxygen and

substrates. J. Am. Chem. Soc., 1999. 121: p. 3876-3887.

32. Beauvais, L.G. and S.J. Lippard, Reactions of the peroxo intermediate of

soluble methane monooxygenase hydroxylase with ethers. J Am Chem Soc,

2005. 127(20): p. 7370-8.

33. Tinberg, C.E. and S.J. Lippard, Oxidation reactions performed by soluble

methane monooxygenase hydroxylase intermediates H(peroxo) and Q proceed

by distinct mechanisms. Biochemistry, 2010. 49(36): p. 7902-12.

34. Lee, S.-K., et al., A transient intermediate of the methane monooxygenase

catalytic cycle containing an FeIVFeIV cluster. J. Am. Chem. Soc., 1993. 115: p.

6450-6451.

35. Liu, K.E., et al., Spectroscopic detection of intermediates in the reaction of

dioxygen with the reduced methane monooxygenase/hydroxylase from

Methylococcus capsulatus (Bath). J. Am. Chem. Soc., 1994. 116(16): p.

7465-7466.

36. Brazeau, B.J. and J.D. Lipscomb, Kinetics and activation thermodynamics of

methane monooxygenase compound Q formation and reaction with substrates.

Biochemistry, 2000. 39(44): p. 13503-15.

37. Mao, S.S., et al., A model for the role of multiple cysteine residues involved in

ribonucleotide reduction: amazing and still confusing. Biochemistry, 1992. 31:

p. 9733-9743.

38. Licht, S., G.J. Gerfen, and J. Stubbe, Thiyl radicals in ribonucleotide

reductases. Science, 1996. 271: p. 477-481.

39. Nordlund, P. and P. Reichard, Ribonucleotide reductases. Annu. Rev.

Biochem., 2006. 75: p. 681-706.

40. Stubbe, J. and W.A. van der Donk, Protein radicals in enzyme catalysis. Chem.

Rev., 1998. 98: p. 705-762.

41. Stubbe, J., Di-iron-tyrosyl radical ribonucleotide reductases. Curr. Opin.

Chem. Biol., 2003. 7(2): p. 183-188.

42. Jiang, W., et al., A manganese(IV)/iron(III) cofactor in Chlamydia trachomatis

ribonucleotide reductase. Science, 2007. 316(5828): p. 1188-91.

43. Cotruvo, J.A. and J. Stubbe, Class I ribonucleotide reductases:

metallocofactor assembly and repair in vitro and in vivo. Annu Rev Biochem,

2011. 80: p. 733-67.

44. Boal, A.K., et al., The dimanganese(II) site in Bacillus subtilis class

Ib ribonucleotide reductase. Biochemistry, 2012.

45. Cotruvo, J.A. and J. Stubbe, Escherichia coli class Ib ribonucleotide reductase

contains a dimanganese(III)-tyrosyl radical cofactor in vivo. Biochemistry,

2011. 50(10): p. 1672-81.

46. Ehrenberg, A. and P. Reichard, Electron Spin Resonance of the

Iron-containing Protein B2 from Ribonucleotide Reductase. Journal of

Biological Chemistry, 1972. 247(11): p. 3485-3488.

47. Sjöberg, B.M., et al., The tyrosine free radical in ribonucleotide reductase

from Escherichia coli. Journal of Biological Chemistry, 1978. 253(19): p.

6863-5.

48. Baldwin, J., et al., Structural characterization of the peroxodiiron(III)

intermediate generated during oxygen activation by the W48A/D84E variant

of ribonucleotide reductase protein R2 from Escherichia coli. Biochemistry,

2003. 42: p. 13269-13279.

49. Yun, D., et al., (Mu-1,2-peroxo)diiron(III/III) complex as a precursor to the

diiron(III/IV) intermediate X in the assembly of the iron-radical cofactor of

ribonucleotide reductase from mouse. Biochemistry, 2007. 46(7): p. 1925-32.

50. Tong, W.H., et al., Mechanism of assembly of the diferric cluster-tyrosyl

radical cofactor of Escherichia coli ribonucleotide reductase from the

diferrous form of the R2 subunit. J. Am. Chem. Soc., 1996. 118(8): p.

2107-2108.

51. Shanmugam, M., et al., Identification of Protonated Oxygenic Ligands of

Ribonucleotide Reductase Intermediate X. J. Am. Chem. Soc., 2009. 131(9): p.

3370-3376.

52. Bollinger, J.M., Jr., et al., Novel diferric radical intermediate responsible for

tyrosyl radical formation in assembly of the cofactor of ribonucleotide

reductase. J. Am. Chem. Soc., 1991. 113(16): p. 6289-91.

53. Coves, J., et al., Enzymic and chemical reduction of the iron center of the

Escherichia coli ribonucleotide reductase protein R2. The role of the

C-terminus. Eur J Biochem, 1995. 233(1): p. 357-63.

54. Ravi, N., et al., Mechanism of Assembly of the Tyrosyl Radical-Diiron(III)

Cofactor of E. Coli Ribonucleotide Reductase: 1. Moessbauer

Characterization of the Diferric Radical Precursor. Journal of the American

Chemical Society, 1994. 116(18): p. 8007-8014.

55. Bollinger, J.M., et al., Mechanism of Assembly of the Tyrosyl

Radical-Diiron(III) Cofactor of E. coli Ribonucleotide Reductase. 2. Kinetics

of The Excess Fe2+ Reaction by Optical, EPR, and Moessbauer

Spectroscopies. Journal of the American Chemical Society, 1994. 116(18): p.

8015-8023.

56. Roshick, C., E.R. Iliffe-Lee, and G. McClarty, Cloning and characterization of

ribonucleotide reductase from Chlamydia trachomatis. J. Biol. Chem., 2000.

275(48): p. 38111-38119.

57. Högbom, M., et al., The radical site in Chlamydial ribonucleotide reductase

defines a new R2 subclass. Science, 2004. 305: p. 245-248.

58. Jiang, W., et al., A manganese(IV)/iron(III) cofactor in Chlamydia trachomatis

ribonucleotide reductase. Science, 2007. 316: p. 1188-1191.

59. Monje-Casas, F., et al., Expression analysis of the nrdHIEF operon from

Escherichia coli - Conditions that trigger the transcript level in vivo. J. Biol.

Chem., 2001. 276(21): p. 18031-18037.

60. McHugh, J.P., et al., Global iron-dependent gene regulation in Escherichia

coli - A new mechanism for iron homeostasis. J. Biol. Chem., 2003. 278(32): p.

29478-29486.

61. Vassinova, N. and D. Kozyrev, A method for direct cloning of Fur-regulated

genes: identification of seven new Fur-regulated loci in Escherichia coli.

Microbiology-Sgm, 2000. 146: p. 3171-3182.

62. Gon, S. and J. Beckwith, Ribonucleotide reductases: influence of environment

on synthesis and activity. Antioxid Redox Signal, 2006. 8(5-6): p. 773-80.

63. Cotruvo, J.A. and J. Stubbe, An active dimanganese(III)-tyrosyl radical

cofactor in Escherichia coli class Ib ribonucleotide reductase. Biochemistry,

2010. 49(6): p. 1297-1309.

64. Stubbe, J. and J.A. Cotruvo, Jr., Control of metallation and active cofactor

assembly in the class Ia and Ib ribonucleotide reductases: diiron or

dimanganese? Curr Opin Chem Biol, 2011. 15(2): p. 284-90.

65. Cotruvo, J.A., Jr. and J. Stubbe, NrdI, a flavodoxin involved in maintenance of

the diferric-tyrosyl radical cofactor in Escherichia coli class Ib ribonucleotide

reductase. Proc Natl Acad Sci U S A, 2008. 105(38): p. 14383-8.

66. Roca, I., et al., NrdI essentiality for class Ib ribonucleotide reduction in

Streptococcus pyogenes. J Bacteriol, 2008. 190(14): p. 4849-58.

67. Kobayashi, K., et al., Essential Bacillus subtilis genes. Proceedings of the

National Academy of Sciences, 2003. 100(8): p. 4678-4683.

68. Jiang, W., et al., Formation and function of the manganese(IV)/iron(III)

cofactor in Chlamydia trachomatis ribonucleotide reductase. Biochemistry,

2008. 47(52): p. 13736-13744.

69. Lindqvist, Y., et al., Crystal structure of delta9 stearoyl-acyl carrier protein

desaturase from castor seed and its relationship to other di-iron proteins.

EMBO J, 1996. 15(16): p. 4081-92.

70. Broadwater, J.A., et al., Peroxodiferric intermediate of stearoyl-acyl carrier

protein D9 desaturase: Oxidase reactivity during single turnover and

implications for the mechanism of desaturation. Biochemistry, 1998. 37: p.

14664-14671.

71. Behrouzian, B. and P.H. Buist, Mechanism of fatty acid desaturation: a

bioorganic perspective. Prostaglandins Leukot Essent Fatty Acids, 2003. 68(2):

p. 107-12.

72. Shu, L., et al., EXAFS and Mössbauer characterization of the Diiron(III) site

in stearoyl-acyl carrier protein Δ<sup>9–</sup> desaturase.

Journal of Biological Inorganic Chemistry, 1998. 3(4): p. 392-400.

73. Broadwater, J.A., et al., Mössbauer studies of the formation and reactivity of a

quasi-stable peroxo intermediate of stearoyl-acyl carrier protein

D9-desaturase. Biochemistry, 1999. 38(38): p. 12197-12204.

74. Pikus, J.D., et al., Recombinant toluene-4-monooxygenase: catalytic and

Mossbauer studies of the purified diiron and rieske components of a

four-protein complex. Biochemistry, 1996. 35(28): p. 9106-19.

75. Murray, L.J., et al., Dioxygen Activation at Non-Heme Diiron Centers:

Oxidation of a Proximal Residue in the I100W Variant of Toluene/o-Xylene

Monooxygenase Hydroxylase†. Biochemistry, 2007. 46(51): p. 14795-14809.

76. Song, W.J., et al., Characterization of a Peroxodiiron(III) Intermediate in the

T201S Variant of Toluene/o-Xylene Monooxygenase Hydroxylase from

Pseudomonas sp. OX1. Journal of the American Chemical Society, 2009.

131(17): p. 6074-6075.

77. Song, W.J. and S.J. Lippard, Mechanistic Studies of Reactions of

Peroxodiiron(III) Intermediates in T201 Variants of Toluene/o-Xylene

Monooxygenase Hydroxylase. Biochemistry, 2011. 50(23): p. 5391-5399.

78. Hirata, Y., et al., Structure of aureothin, a nitro compound obtained from

Streptomyces thioluteus. Tetrahedron, 1961. 14: p. 252-274.

79. He, J. and C. Hertweck, Biosynthetic origin of the rare nitroaryl moiety of the

polyketide antibiotic aureothin: Involvement of an unprecedented N-oxygenase.

J. Am. Chem. Soc., 2004. 126(12): p. 3694-3695.

80. Choi, Y.S., et al., In vitro reconstitution and crystal structure of

p-aminobenzoate N-oxygenase (AurF) involved in aureothin biosynthesis. Proc.

Natl. Acad. Sci. USA, 2008. 105(19): p. 6858-6863.

81. Winkler, R., et al., Regio- and chemoselective enzymatic N-oxygenation in vivo,

in vitro, and in flow. Angew Chem Int Ed Engl, 2006. 45(47): p. 8016-8.

82. Jiang, W., J.M. Bollinger, Jr., and C. Krebs, The active form of Chlamydia

trachomatis ribonucleotide reductase R2 protein contains a heterodinuclear

Mn(IV)/Fe(III) cluster with S = 1 ground state. J. Am. Chem. Soc., 2007.

129(24): p. 7504-7505.

83. Jiang, W., et al., A Mn(IV)/Fe(IV) Intermediate in Assembly of the

Mn(IV)/Fe(III) Cofactor of Chlamydia trachomatis Ribonucleotide Reductase.

Biochemistry, 2007. 46: p. 8709-8716.

84. Krebs, C., et al., AurF from Streptomyces thioluteus and a Possible New

Family of Manganese/Iron Oxygenases. Biochemistry, 2007. 46(37): p.

10413-10418.

85. Li, N., et al., Four-electron oxidation of p-hydroxylaminobenzoate to

p-nitrobenzoate by a peroxodiferric complex in AurF from Streptomyces

thioluteus. Proc. Natl. Acad. Sci. USA, 2010. 107(36): p. 15722–15727.

86. Misra, H.P. and I. Fridovich, The oxidation of phenylhydrazine: superoxide

and mechanism. Biochemistry, 1976. 15(3): p. 681-687.

87. Tomasi, A., et al., Detection of Free Radical Intermediates in the Oxidative

Metabolism of Carcinogenic Hydrazine Derivatives. Toxicologic Pathology,

1987. 15(2): p. 178-183.

88. Huang, P.-K.C. and E.M. Kosower, Diazenes. III. Properties of phenyldiazene.

Journal of the American Chemical Society, 1968. 90(9): p. 2367-2376.

89. Huang, P.-K.C. and E.M. Kosower, Diazenes. I. Decarboxylation of

phenyldiazene-carboxylic acid. Journal of the American Chemical Society,

1968. 90(9): p. 2354-2362.

90. Huang, P.-K.C. and E.M. Kosower, Diazenes. II. Preparation of phenyldiazene.

Journal of the American Chemical Society, 1968. 90(9): p. 2362-2367.

91. Patek, D.R. and L. Hellerman, Mitochondrial Monoamine Oxidase. Journal of

Biological Chemistry, 1974. 249(8): p. 2373-2380.

92. Cheesbrough, T.M. and P.E. Kolattukudy, Microsomal preparation from an

animal tissue catalyzes release of carbon monoxide from a fatty aldehyde to

generate an alkane. J. Biol. Chem., 1988. 263(6): p. 2738-2743.

93. Cheesbrough, T.M. and P.E. Kolattukudy, Alkane biosynthesis by

decarbonylation of aldehydes catalyzed by a particulate preparation from

Pisum sativum. Proc. Natl. Acad. Sci. USA, 1984. 81(21): p. 6613-6617.

94. Yoder, J.A., et al., Enhancement of diapausing flesh fly puparia with

additional hydrocarbons and evidence for alkane biosynthesis by a

decarbonylation mechanism. Insect Biochem. Mol. Biol., 1992. 22(3): p.

237-243.

95. Schneider-Belhaddad, F. and P.E. Kolattukudy, Solubilization, partial

purification, and characterization of a fatty aldehyde decarbonylase from a

higher plant, Pisum sativum. Arch. Biochem. Biophys., 2000. 377(2): p.

341-349.

96. Dennis, M. and P.E. Kolattukudy, A cobalt-porphyrin enzyme converts a fatty

aldehyde to a hydrocarbon and carbon monoxide. Proc. Natl. Acad. Sci. USA,

1992. 89(12): p. 5306-10.

97. Mpuru, S., et al., Mechanism of hydrocarbon biosynthesis from aldehyde in

selected insect species: Requirement for O2 and NADPH and carbonyl group

released as CO2. Insect Biochemistry and Molecular Biology, 1996. 26(2): p.

203-208.

98. Warui, D.M., et al., Detection of formate, rather than carbon monoxide, as the

stoichiometric co-product in conversion of fatty aldehydes to alkanes by a

cyanobacterial aldehyde decarbonylase. J. Am. Chem. Soc., 2011. 133: p.

3316-3319.

99. Whitteck, J.T., R.M. Cicchillo, and W.A. van der Donk, Hydroperoxylation by

hydroxyethylphosphonate dioxygenase. J. Am. Chem. Soc., 2009. 131(44): p.

16225-16232.

100. Li, N., et al., Conversion of Fatty Aldehydes to Alka(e)nes and Formate by a

Cyanobacterial Aldehyde Decarbonylase: Cryptic Redox by an Unusual

Dimetal Oxygenase. Journal of the American Chemical Society, 2011. 133(16):

p. 6158-6161.

101. Eser, B.E., et al., Oxygen-independent alkane formation by non-heme

iron-dependent cyanobacterial aldehyde decarbonylase: investigation of

kinetics and requirement for an external electron donor. Biochemistry, 2011.

50(49): p. 10743-50.

102. Das, D., et al., Oxygen-independent decarbonylation of aldehydes by

cyanobacterial aldehyde decarbonylase: a new reaction of diiron enzymes.

Angew Chem Int Ed Engl, 2011. 50(31): p. 7148-52.

103. Hisada, R. and T. Yagi, 1-Methoxy-5-methylphenazinium methyl sulfate. A

photochemically stable electron mediator between NADH and various electron

acceptors. J Biochem, 1977. 82(5): p. 1469-73.

Chapter 2

The mechanism of the reaction of AurF with its substrate analogue, 4-hydrazinobenzoate

Li et. al.

Abstract

The ferritin-like diiron-carboxylate N-oxygenase AurF from Streptomyces thioluteus is known to catalyze the conversion of 4-aminobenzoate (Ar-NH2) to 4-nitrobenzoate

(Ar-NO2) to provide a precursor for the biosynthesis of the antibiotics aureothin. This six electron oxidation of Ar-NH2 to Ar-NO2 was believed to proceed via three consecutive two-electron oxidation steps with 4-hydroxylaminobenzoate (Ar-NHOH) and 4-nitrosobenzoate (Ar-NO) as the two intermediates. We recently showed that a

III/III peroxo-Fe2 intermediate in AurF is competent to oxidize Ar-NH2 to Ar-NHOH and also to oxidize Ar-NHOH to Ar-NO2 in succession. Surprisingly, the second oxidation is an unprecedented single-step four-electron conversion, different from what had been proposed before. In this oxidation step, neither Ar-NO nor diferric

AurF species accumulate. In this article, we characterize the reaction between AurF and its Ar-NHOH substrate/intermediate analogue, 4-hydrazinobenzoate (Ar-NHNH2).

The substitution of the hydroxyl group with an amino moiety modifies both the redox potential and the electron donating capacity and completely changes the reaction trajectory by providing for trapping of a 2-electron oxidized substrate species by N2 elimination. By combining the rapid mixing stopped-flow spectrophotometry, freeze-quench Mössbauer spectrometry, and LC/MS detection of products, we elucidate the kinetics and mechanism of this novel reaction.

Introduction

The arylamine N-oxygenase, AurF from Streptomyces thioluteus converts

4-aminobenzoate (Ar-NH2) to 4-nitrobenzoate (Ar-NO2) [1-2]. The crystal structure of the enzyme revealed that AurF belongs to ferritin-like diiron-carboxylate protein family that includes β2 subunit of class I ribonucleotide reductases and the oxygenase components of bacterial multicomponent monooxygenases (BMMs) [3-5]. These ferritin-like diiron-carboxylate enzymes employ diiron cluster to activate O2. The enzyme AurF was found to have either Fe or Mn at the metal sites when expressed in

E. coli grown in Luria-Bertani medium and there was a brief controversy over the identity of the active metallocofactor [6-9]. Zhao’s group showed that the diiron form of AurF is capable of making the Ar-NO2 product from the Ar-NH2 substrate in the presence of oxygen and a reducing system such as phenazine methylsulfate/NADH

(PMS/N) or ferredoxin/ferredoxin reductase/NADPH (F/FR/N) [10]. It was initially proposed that this net six-electron oxidation proceeds via three consecutive, two-electron-oxidation step, through p-hydroxylaminobenzoate (Ar-NHOH) and p-nitrosobenzoate (Ar-NO) intermediates, and requires one equivalent of O2 and two exogenous reducing equivalents for each of the three steps [10]. Zhao, et al. claimed the detection of Ar-NHOH and Ar-NO intermediates, which is consistent with the proposed three-step oxidation sequence (Scheme 1A) [10]. The activity of diiron AurF was confirmed by the studies of its mechanism and identification of a key reactive

II/II III/III intermediate. Fe2 -AurF (obtained by chemically reducing as-isolated Fe2 -AurF with dithionite), can react with oxygen rapidly (reaction completed within 10 ms at

III/III 20 °C ) to form a peroxo-Fe2 intermediate which is responsible for the oxidation of

III/III Ar-NH2 to Ar-NHOH [11]. This peroxo-Fe2 intermediate is remarkably long-lived

66 in the absence of substrate (t1/2 = 7 min at 20 °C) but decays rapidly in the presence of stoichiometric Ar-NH2 (t1/2 = 5 ms at 20 °C), with conversion of the substrate to

III/III III/III Ar-NHOH and the cofactor to Fe2 [11]. Later, we showed that this peroxo-Fe2 intermediate is also responsible for the oxidation of Ar-NHOH to Ar-NO2.

Surprisingly, kinetic data established that the reaction is a novel, one-step,

III/III four-electron oxidation with peroxo-Fe2 -AurF being reduced all the way to

II/II Fe2 -AurF and Ar-NHOH being directly oxidized to Ar-NO2 (Scheme 1B) [12].

Thus, the stoichiometry of the conversion of Ar-NH2 to Ar-NO2 catalyzed by AurF

- + was necessarily revised to Ar-NH2 + 2O2 + 2e + 2H = Ar-NO2 + 2H2O [12]. The hydrazine Ar-NHNH2 is structurally similar to Ar-NHOH, having the hydroxyl group

(-OH) substituted by an amino group (-NH2). However, the substitution affects the reduction potential, marking hydrazine (-NHNH2) a much better reductant.

Ar-NHNH2 is able to provide 2 electrons when oxidized to diazene, which will decompose to benzoate and N2. We therefore became interested in the outcome of reacting AurF with Ar-NHNH2. In this work, we show that, surprisingly, Ar-NHNH2

III/III III/III reacts rapidly with both as-isolated Fe2 -AurF and peroxo-Fe2 -AurF. In the reduction of both forms of AurF by Ar-NHNH2, two electrons, instead of four, are

III/III provided by one molecule of Ar-NHNH2. Upon reduction of Fe2 -AurF by

II/II Ar-NHNH2, Fe2 -AurF is produced. Ar-NHNH2 also reacts with

III/III III/III peroxo-Fe2 -AurF, and Mössbauer spectra reveal the accumulation of Fe2 -AurF.

Excess Ar-NHNH2 in the reaction solution successively reduces the resulting

III/III II/II III/III Fe2 -AurF to Fe2 -AurF, readily to generate another peroxo-Fe2 -AurF in the presence of O2. The oxidation of Ar-NHNH2 by AurF is therefore catalytic in the presence of sufficient O2. Benzoate is detected as the major product in this reaction.

Our studies show that, despite of the structural similarity between the Ar-NHOH and

Ar-NHNH2, the reaction mechanisms of AurF with these substrate and analog are different. A previously one-step, 4-electron oxidation of Ar-NHOH by

III/III peroxo-Fe2 -AurF is remodeled to a two-step, 2-electron oxidation of Ar-NHNH2

III/III by peroxo-Fe2 -AurF by replacing the –OH group with a –NH2 moiety, in which

III/III Fe2 -AurF is accumulated (Scheme 2).

Materials and Methods

Materials

AurF was prepared as previously described [11] and 4-hydrazinobenzoate was procured from Alfa Aesar.

Stopped-flow experiments

Reduction of as-isolated AurF by Ar-NHNH2 was assayed by mixing 300 μM deoxygenated as-isolated AurF with an equal volume of solution containing various concentration of Ar-NHNH2 at 5 °C . The absorption change was monitored comprehensively by photodiode array detector or at 360 nm by photomultiplier

III/III amplifier. Reduction of peroxo-Fe2 -AurF was tested by first mixing 600 μM

II/II dithionite-reduced Fe2 -AurF with an equal volume of buffer solution containing

III/III stoichiometric O2 for 0.5 s, allowing for accumulation of peroxo-Fe2 -AurF, which was then mixed with solution containing various concentration of Ar-NHNH2 at 5 °C .

The absorption change was monitored comprehensively by photodiode array detector or at 500 nm and 360 nm, respectively by photomultiplier amplifier.

HPLC and LC/MS analysis

The HPLC assay was performed as previously describe [13]. LC/MS assay samples were either purified by filtration or derivatized before analyzing by LC/MS and the detailed methods was adapted from literature [14-15].

Results

III/III III/III Analysis of the reaction of AurF (Fe2 and peroxo-Fe2 -AurF) with

Ar-NHNH2 by Stopped-Flow Absorption (SF-Abs) spectrophotometry.

As previously reported, the UV/vis absorption spectrum of as-isolated AurF exhibits a

III/III band at 360 nm, similar to other enzymes that contain a μ-oxo-Fe2 cluster (Fig. 1A, grey) [11, 16]. Incubation of this form with stoichiometric Ar-NHNH2 results in an overall decrease in absorption (Fig. 1A, black). AurF pre-treated with Ar-NHNH2 in

III/III this manner reacts with O2 to generate the 500 nm shoulder of peroxo-Fe2 -AurF

III/III (Fig. 1A, red and blue) [11]. The reactivity of peroxo-Fe2 -AurF with Ar-NHNH2

II/II was then examined in a sequential-mixing experiment. Fe2 -AurF was mixed with

III/III stoichiometric O2 for 0.5 s to allow for formation of the peroxo-Fe2 intermediate

(Fig. 1B, grey to black). The solution was then mixed with either substoichiometric

III/III Ar-NHNH2 or excess Ar-NHNH2 (Fig. 1B). Treatment of peroxo-Fe2 with substoichiometric Ar-NHNH2 resulted in a spectrum similar to as-isolated AurF (Fig.

1B, black and red). Treatment with excess Ar-NHNH2 yielded an absorption spectrum similar to that of dithionite-reduced AurF (Fig. 1B, black and blue).

Study of the Reaction Kinetics and Stoichiometry.

III/III Our previous studies showed that the formation and decay of peroxo-Fe2 intermediate can be monitored by the absorption change at 500 nm; similarly the diferric AurF species can be monitored by the absorption at 360 nm [11]. As before,

III/III peroxo-Fe2 -AurF was produced in the first mix and then mixed with varying concentrations of Ar-NHNH2 (Fig. 2). The absorption change at 500 nm was then used for the determination of kobs by fitting the traces by using the equation for two parallel exponential decay processes. The inset shows the relationship between kobs of

70 the faster phase of larger amplitude and the substrate concentration. The slope gives a second order rate constant of 17.5 (± 0.2) mM-1·s -1. Similarly, a second order rate

-1 -1 III/III constant of 106 (± 3) mM ·s was obtained for the O2 addition (peroxo-Fe2 -AurF

III/III formation) (Fig. 4). Fe2 -AurF reacted with various concentration of Ar-NHNH2 anaerobically and change in absorbance at 360 nm was monitored. kobss were obtained by fitting the traces by the equation of two parallel exponential decay processes. The fast decay phase of the reaction shows the reaction between diferric AurF with

Ar-NHNH2. Fitting the kobss versus Ar-NHNH2 concentrations with a hyperbolic

-1 -1 -1 equation gives kmax = 7.6 (± 0.1) s and kmax/K0.5 = 1.3 mM ·s (Fig. 3). Partial

III/III reduction of Fe2 -AurF with substoichiometric Ar-NHNH2, tested by stopped-flow experiment indicates that stoichiometric Ar-NHNH2 is required to reduce

III/III Fe2 -AurF (Fig. S1). Change of absorbance at 360 nm is also monitored in the

III/III reduction of peroxo-Fe2 -AurF with Ar-NHNH2 (Fig. S2). The two phases of

III/III formation and decay correspond to the production of Fe2 -AurF and the conversion

III/III II/II of so produced Fe2 -AurF to Fe2 -AurF. Slight decay caused by stoichiometric

III/III Ar-NHNH2 is likely due to the possibility that the peroxo-Fe2 -AurF was lower than the theoretical amount, making the so calculated “stoichiometric Ar-NHNH2” slightly in excess. Both assays indicate that only 2 electrons of Ar-NHNH2 contribute

III/III III/III to the reduction of either Fe2 -AurF or peroxo-Fe2 -AurF.

III/III Evaluation of Diiron Products in Reaction of diferric and Peroxo-Fe2 -AurF with Ar-NHNH2 by Mössbauer Spectroscopy.

III/III Mössbauer spectroscopy was used to test the hypothesis that Fe2 -AurF AurF can oxidize Ar-NHNH2. Incubation of as-isolated AurF (Fig. 5A) with 2 equiv of

Ar-NHNH2 for 10 min resulted in a different spectrum (Fig. 5B) that can be analyzed

71 as a quadruple doublet δ = 1.23 mm/s, ΔEQ = 3.08 mm/s. The parameters are very

II/II similar to those previous reported for Fe2 -AurF generated by dithionite reduction of as-isolated AurF (δ = 1.24 mm/s, ΔEQ = 3.06 mm/s). This quadruple doublet accounts for 88% of the total intensity of the spectrum (blue reference spectrum in Fig. 5B). We estimate an uncertainty of ± 3 on this and all other percentages of total absorption area

III/III III/III given in the text. 5% of μ-oxo-Fe2 and 2% of peroxo-Fe2 were probably

II/II formed by Fe2 -AurF with O2 contamination over 10-min incubation time (Fig. 5B, green and red reference quadruple doublet respectively). As proposed above,

III/III Ar-NHNH2 can react with Fe2 -AurF as well. In order to avoid the complexity

II/II created by newly generated Fe2 -AurF reacting with excess oxygen, we reacted

II/II III/III Fe2 -AurF with substoichiometric O2 to form peroxo-Fe2 -AurF (with no excess oxygen remained) which was then mixed with substoichiometric Ar-NHNH2 or excess

II/II Ar-NHNH2 respectively. Dithionite-reduced AurF (Fe2 ) was mixed with an equal volume of buffer solution containing 0.75 equiv O2 and the mixture was allow for 0.5

III/III s to form the peroxo-Fe2 -AurF (Fig. 5C, G) which accounted for 62% and 65% of total iron intensity respectively. In one set of the freeze-quench reactions, 0.6 equiv

Ar-NHNH2 (Ar-NHNH2/O2 = 0.8, Ar-NHNH2/Fe2 = 0.6) was mixed with the

III/III peroxo-Fe2 -AurF and the reaction was freeze quenched at 10 s (Fig. 5D) and 6

III/III min (Fig. 5E). Over 10 s’ reaction, peroxo-Fe2 species decreased from 65% to

III/III II/II 15%, Fe2 -AurF increased from 7% to 30% and the Fe2 -AurF increased from 18% to 28%. One new quadruple doublet with a δ = 0.48 mm/s, ΔEQ = 0.82 mm/s appeared,

III/III presumably a μ-hydroxo-Fe2 species (very similar to the small-portion species observed in as-isolated AurF we reported before), which accounted for 18% of total

III/III iron. After 6 min’s reaction, peroxo-Fe2 -AurF further decreased to 8%,

III/III II/II Fe2 -AurF increased to 38% and the Fe2 -AurF increased to 40%. The new quadruple doublet decreased to 6%. This change clearly indicated that the

III/III peroxo-Fe2 -AurF reacted with Ar-NHNH2 and was converted in to diferric as well as diferrous. The difference spectrum F = E-C was generated to better demonstrated

III/III the change. Within 6 min, 57% of peroxo-Fe2 -AurF was converted to 37% of

III/III III/III II/II Fe2 -AurF (including both μ-oxo- and μ-hydroxo-Fe2 ) and 20% of Fe2 -AurF.

Noted that only 0.6 equiv Ar-NHNH2 was added, therefore ~17% (2 20% + 37% -

60%) of the reduction was presumably caused by the self decay of

III/III III/III peroxo-Fe2 -AurF. When we mixed the peroxo-Fe2 -AurF with 1.5 equiv

Ar-NHNH2 (Ar-NHNH2/O2 = 2, Ar-NHNH2/Fe2 = 1.5), within 1 s (Fig. 5H),

III/III III/III peroxo-Fe2 -AurF was almost completely consumed, Fe2 -AurF in this same

II/II time increased from 4% to 20% and Fe2 -AurF increased from 30% to 53%. Again,

III/III the new quadruple doublet of μ-hydroxo-Fe2 -AurF appeared, accounting for 21%

II/II of total iron. In the 8-min sample, the quadruple doublet corresponding to Fe2 -AurF accounted for 93% of the total iron (Fig. 5I), indicating that the final product of

III/III II/II reduction of peroxo-Fe2 by excess Ar-NHNH2 is Fe2 -AurF.

Verification of Catalytic Oxidation of Ar-NHNH2 by as-isolated AurF.

III/III Three demonstrated reactions: (1) reduction of Fe2 -AurF by Ar-NHNH2 to

II/II II/II III/III Fe2 -AurF; (2) reaction of Fe2 -AurF O2 to form peroxo-Fe2 -AurF; and (3)

III/III III/III reduction of peroxo-Fe2 -AurF by Ar-NHNH2 to Fe2 -AurF make up a complete potential catalytic cycle. To verify that this cycle is indeed functional, 10 µM as-isolated AurF was incubated with 800 µM Ar-NHNH2 on ice, exposed to air for 12 min (stirring to facilitate oxygen diffusion). The substrate and products were separated from the enzyme by filtration. The filtrate was then analyzed by HPLC

73 equipped with a Hamilton reverse phase PRP-1 column and a photodiode array detector. The chromatogram of a standard solution containing 400 µM each of

Ar-NHNH2, benzoate, and Ar-NO2 demonstrates resolution of these three components

(Fig. 6, black trace). Each component was injected individually to confirm its identity by both its absorption spectrum and its m/z value (LC/MS in parallel). A control reaction solution containing only 800 µM Ar-NHNH2 without AurF was prepared and analyzed by the same method. The chromatogram shows that Ar-NHNH2 is relatively stable under the assay conditions in the absence of AurF (Fig. 6, red trace). A small amount of benzoate (less than 5%) was detected in the control solution due to the self oxidation during the incubation time (Fig. 6, red trace, confirmed by LC/MS). In the complete reaction solution, all the Ar-NHNH2 was consumed and ~450 µM benzoate was detected, indicating that more than 50% of the Ar-NHNH2 was converted to benzoate. A trace amount of Ar-NO2 was detected, and its percentage of yield seems to depend on AurF concentration (Fig. S3).

Evidence for the reaction product by isotopic labeling LC/MS assay.

According to the literature, the direct 2-electron oxidation product of Ar-NHNH2,

Ar-N=NH is unstable and quickly oxidized by O2 and decomposed to benzoic acid and N2 [17-19]. It is reported that the benzoate is generated from Ar-N=NH by abstraction of a hydrogen atom from the solvent [20]. Thus we performed the reaction

2 in H2O enriched solution to indirectly verify the production of Ar-N=NH. If

Ar-N=NH is produced by Ar-NHNH2 oxidation, benzoate would be expected to decompose to with incorporation of 2H at the 4-position. To increase the detection sensitivity, the benzoate was derivatized by 2-nitrophenylhydrazine prior to LC/MS

2 analysis [14-15]. A reaction solution containing 85% H2O, 800 µM Ar-NHNH2 and

40 µM as-isolated AurF was incubated at room temperature, exposed to air for 10 min.

It was then subjected to the derivatization and analyzed by LC/MS. Derivatized benzoate has m/z = 256 in the negative ion mode. A 16% fraction with m/z = 257

(relative to the fraction with m/z = 256) is expected on the basis of the natural isotope

13 15 2 17 2 abundance of C, N, H and O. The chromatogram from the reaction in H2O shows a marked increase in the m/z = 257 portion of the benzoate (product) derivative, suggesting the production of deuterated benzoate (Fig. 7). This is consistent with our prediction that Ar-N=NH was being produced by oxidation of Ar-NHNH2 by AurF. In

2 this 85% H2O enriched reaction solution, up to 33% of the derivative has an m/z =

257 corresponding to 4-2H-benzoate (16% of m/z = 256 portion being subtracted out first) indicating the production of 4-2H-benzoate by AurF. The discrepancy between

2 2 the ratio of 85% H2O and 33% 4- H-benzoate was likely due to the kinetic isotope effect (KIE) of the hydrogen abstraction step (preference of forming 4-1H-benzoate over 4-2H-benzoate) in the solution (with an calculated H/D KIE of

0.85/0.15/0.66*0.33 ~ 11).

Discussion.

In our previous work, we identified an reactive intermediate designated as

III/III II/II peroxo-Fe2 -AurF that forms rapidly upon exposure of Fe2 -AurF to O2 in the

III/III absence of substrate [11]. This peroxo-Fe2 species is remarkably stable (t1/2 ~ 7 min at room temperature) and reacts rapidly with Ar-NH2 (t1/2 < 30 ms) converting it into Ar-NHOH [11]. Soon after, we reported that this intermediate is also capable of oxidizing Ar-NHOH via a novel, single-step 4-eletron reaction generating Ar-NO2 and

II/II Fe2 -AurF [12]. Notably, spectroscopic evidence suggested a complete absence of the expected accumulation of 2-electron oxidized intermediate state containing Ar-NO

III/III (or Ar-N(OH)2) and Fe2 -AurF. Accordingly, in this work, we investigated the reaction of the peroxo complex with Ar-NHNH2, an analog predicted to decompose rapidly following its oxidation by just two electrons. We anticipated that this reaction would provide indirect support for the intermediate of a 2-electron state in the

Ar-NHOH reaction. The reactivity of hydrazine and its derivatives has been studies extensively [18-26]. Literature suggests that hydrazine derivatives are oxidized to diazene first by a 2-electron oxidation before further decompose to N2 and the rest.

Surprisingly, our studies showed that Ar-NHNH2 is capable of reducing as-isolated

II/II AurF to Fe2 -AurF. Our studies show that substoichiometric Ar-NHNH2 only partially reduced as-isolated AurF and stoichiometric Ar-NHNH2 is able to fully

III/III II/II reduce as-isolated Fe2 -AurF to Fe2 -AurF, suggesting that two electrons are provided by one Ar-NHNH2 molecule in this reduction. Stopped-flow data showed

III/III that upon reduction of peroxo-Fe2 -AurF, two equiv Ar-NHNH2 is required to fully

II/II reduce the peroxo complex to Fe2 -AurF state, suggesting that, again, only two electrons are coming from one Ar-NHNH2 molecule (Fig. S2). Note that a small

III/III III/III II/II fraction of the peroxo-Fe2 -AurF is reduced to Fe2 -AurF and then Fe2 -AurF in the presence of only one equiv Ar-NHNH2 (Fig. S2, red trace). It is likely that

III/III peroxo-Fe2 -AurF was formed less than theoretically calculated amount, resulting in the “stoichiometric Ar-NHNH2” relatively in excess. As we predicted, shown by

Mössbauer spectroscopy, this two-electron oxidation of Ar-NHNH2 by

III/III III/III peroxo-Fe2 -AurF allows for the accumulation of Fe2 -AurF intermediate, previously not seen in the reaction with Ar-NHOH (Fig. 5). Our Mössbauer data also

III/III revealed a new species of μ-hydroxo-Fe2 being produced upon Ar-NHNH2

III/III III/III reduction of peroxo-Fe2 -AurF. It was seemingly converted to μ-oxo-Fe2 over

76 longer incubation time spontaneously. This observation provides an evidence for the protonation step previously proposed for the oxidation of Ar-NHOH by

III/III III/III peroxo-Fe2 -AurF [13]. This μ-hydroxo-Fe2 species formed after 1 s and decayed after 6 min. A finer time course is therefore required to understand this

III/III complex better. It would also be interesting to test if the similar μ-hydroxo-Fe2

III/III species is formed in the reduction of Fe2 -AurF by Ar-NHNH2.

As for the small molecule product, we carefully verified the production of benzoate and its H abstraction from solvent by LC/MS, providing an indirect evidence for the production of Ar-N=NH. The benzoate is also used as the measurement to show catalytic oxidation of Ar-NHNH2 by as-isolated AurF in the presence of O2.

In addition, about 5% of Ar-NHNH2 is converted to Ar-NO2 and isotopic tracing

18 18 experiment shows that in the presence of O2, both oxygen atoms of Ar-NO2 are O

(Fig. S4). More interestingly, the production of Ar-NO2 seems to be enzyme concentration dependent (Fig. S3). This interesting observation indicates that a small portion of Ar-NHNH2 may be hydrolyzed first and converted finally to Ar-NO2. The small fraction indicates that this process is not favored by in the oxidation of

Ar-NHNH2 by AurF. The AurF concentration dependency of Ar-NO2 production is consistent with this hypothesis as more concentrated AurF gave higher chance for

Ar-NO2 formation, whereas the production of benzoate was relatively not affected.

Currently the total product only accounts for ~60% of the substrate analogue consumed. Bimolecular reaction has been proposed for diazene compounds and one peak we detected at m/z = 297 corresponding to a 4-diazenylbenzoate dimer might be an explanation for the other 40% Ar-NHNH2 consumed (Fig. S5). By replacing the

–OH group with a –NH2 moiety, we showed that the oxidation of the substrate analog

77 is remodel from a one-step, 4-electron process to a two-step, 2-electrion sequence, and our studies provide the evidence for this reaction of AurF with Ar-NHNH2, dissect the reaction mechanism and thus expand the repertoire of AurF products.

Scheme 1

Scheme 2

Figure 1A.

SF-Abs experiments to demonstrate the reduction of as-isolated AurF by Ar-NHNH2 and the oxidation of Ar-NHNH2 reduced AurF with oxygen. A solution of as-isolated

AurF (0.30 mM in 100 mM HEPES pH 7.5 and 10% glycerol buffer; grey line) was mixed anaerobically with an equal volume of O2-free buffer containing 300 μM

Ar-NHNH2 at 6 °C for 100 s (black line). A solution of 300 μM as-isolated AurF and

1 equiv. of Ar-NHNH2 was incubated for 30 min and mixed against an equal volume of an O2-free buffer (red line) or a buffer solution containing 1.8 mM O2 (blue line) for 1 s.

Figure 1B.

III/III SF-Abs experiments to monitor the reaction of peroxo-Fe2 -AurF with Ar-NHNH2:

II/II A solution of Fe2 -AurF (0.60 mM in 100 mM HEPES pH 7.5 and 10% glycerol buffer; grey line) was mixed with an equal volume of reaction buffer containing 600

μM O2. This solution was allowed to react at 5 °C for 0.5 s to permit accumulation of

III/III peroxo-Fe2 -AurF (black line) before mixing with an equal volume of an O2-free solution containing 0.5 equiv (red line) or 2 equiv (blue line) Ar-NHNH2 for 100 s.

Figure 2.

III/III Kinetics of the reaction of peroxo-Fe2 -AurF with Ar-NHNH2: A solution of

II/II Fe2 -AurF (600 μM Fe2) was mixed with an equal volume of reaction buffer containing 600 μM O2. This solution was allowed to react at 5 ℃ for 0.5 s to permit

III/III accumulation of peroxo-Fe2 -AurF and was then mixed with an equal volume of

O2-free solution of buffer containing 0.0 mM (black), 0.3 mM (red), 0.6 mM (green),

1.2 mM (blue), 2.4 mM (orange), 4.8 mM (sky blue) or 9.6 mM (pink) Ar-NHNH2 and the absorbance at 500 nm monitored. The inset shows the apparent first-order rate constant for the fast decay phase of the reaction (obtained by fitting the equation for two parallel exponential decay processes to the data) versus [Ar-NHNH2], indicating a second-order rate-constant (slope) of 17.5 (± 0.2) mM-1 s-1.

Figure 3.

Kinetics of the reaction of as-isolated AurF with Ar-NHNH2: A solution of as-isolated

AurF (200 μM Fe2) was mixed anaerobically with an equal volume of O2-free solution of buffer containing 0.0 mM (black), 0.2 mM (red), 0.4 mM (green), 0.8 mM

(blue), 1.2 mM (orange), or 2.4 mM (sky blue) Ar-NHNH2. The inset shows the apparent first-order rate constant for the fast decay phase of the reaction (obtained by fitting the equation for two parallel exponential decay processes to the data) versus

-1 [Ar-NHNH2], indicating a first-order rate-constant of kmax = 7.6 (± 0.1) s , and

-1 -1 kmax/K0.5 = 1.3 mM ·s .

Figure 4.

III/III Dependence of the formation of peroxo-Fe2 -AurF on the concentration of O2: A

II/II solution of Fe2 -AurF (100 μM Fe2) was mixed with an equal volume of O2-free buffer (black) or buffer containing 50 μM (red), 0.1 mM (green), 0.3 mM (blue), 0.9 mM (orange), 1.8 mM (grey) O2. The inset show the apparent first-order rate constant for the reaction (obtained by fitting the equation for a single exponential growth to the

-1 data) versus [O2], indicating a second-order rate-constant (slope) of 106 (± 3) mM s-1.

Figure 5.

4.2-K/53-mT Mössbauer spectra of samples in which either as-isolated AurF

III/III III/III (Fe2 -AurF) or peroxo-Fe2 -AurF was reacted with Ar-NHNH2. Left panel: a

III/III solution of as isolated Fe2 -AurF (1.2 mM Fe2) was either hand frozen right away

(A) or after mixing with 1 equivalent volume of buffer solution containing 2.4 mM

Ar-NHNH2 (Ar-NHNH2/Fe2 = 2) and incubating for 10 min in the absence of O2 (B).

II/II Middle panel: a solution of Fe2 -AurF (1.2 mM Fe2) was mixed with 0.5 equivalent volume of buffer solution containing 1.8 mM O2 (O2/Fe2 = 0.75). This solution was

III/III allowed to react at 5 °C for 0.11 s to permit accumulation of peroxo-Fe2 -AurF. (C)

The solution was then directly freeze-quenched. (D and E) The solution was then mixed with one-sixth equivalent volume of an O2-free solution of 2.88 mM

Ar-NHNH2 (Ar-NHNH2/Fe2 = 0.6), and this solution was allowed to react for 10 s (D) or 6 min (E) prior to being freeze-quenched. F is the difference spectrum E-C. Right

II/II panel: a solution of Fe2 -AurF (1.2 mM Fe2) was mixed with 0.5 equivalent volume of buffer solution containing 1.8 mM O2 (O2/Fe2 = 0.75). This solution was allowed to

III/III react at 5 °C for 0.11 s to permit accumulation of peroxo-Fe2 -AurF. (G) The solution was then directly freeze-quenched. (H and I) The solution was then mixed with one-sixth equivalent volume of an O2-free solution of 7.2 mM Ar-NHNH2

(Ar-NHNH2/Fe2 = 1.5), and this solution was allowed to react for 1 s (H) or 8 min (I) prior to being freeze-quenched. The red, blue, green and purple lines illustrate the

III/III II/II fractional contributions of the reference spectra of peroxo-Fe2 -AurF, Fe2 -AurF,

III/III III/III µ-oxo-Fe2 -AurF and µ-hydroxo-Fe2 -AurF, respectively, to the experimental spectrum, as described in the text.

Figure 6.

Reversed-phase high performance liquid chromatography (RP-HPLC) of the small molecule reactants and products in the reaction of as-isolated AurF with Ar-NHNH2.

As-isolated AurF (10 µM) was incubated with 800 µM Ar-NHNH2 in the presence of excess O2 for 12 min at 0 °C. Small molecules were separated from the enzyme and analyzed as described in the methods section (blue). A control experiment was carried out under identical conditions, except for omission of AurF (red). A solution containing 400 µM each of Ar-NHNH2, benzoate, and Ar-NO2 was also analyzed

(black).

Figure 7.

LC/MS chromatogram demonstrating the derivatized benzoate products of Ar-NHNH2 oxidation catalyzed by as-isolated AurF in the presence of O2. A reaction solution containing 800 µM Ar-NHNH2 and 10 µM as isolated AurF was incubated at room temperature for 10 min in the presence of excess O2. Small molecules were separated from the enzyme by filtration before being derivatized as described in the methods section. Filtrates were analyzed by LC/MS (red). A similar sample of reaction solution containing 85% 2H2O, 800 µM Ar-NHNH2 and 40 µM as-isolated AurF was prepared and analyzed (blue).

Figure S1

Kinetic traces after mixing as-isolated AurF (0.3 mM in 100 mM HEPES pH 7.5 and 10% glycerol buffer) at 5 °C with an equal volume of O2-free buffer containing 0 mM (black trace), 0.1 mM (red trace), 0.3 mM (green trace) or 0.6 mM (blue trace)

Ar-NHNH2.

Figure S2

Kinetic traces after mixing DT-reduced AurF (0.6 mM in 100 mM HEPES pH 7.5 and 10% glycerol buffer) at 5 °C with an equal volume of buffer solution containing

0.6 mM O2, allowing the reaction to proceed for 0.5 s, and then mixing the resultant solution with an equal volume of O2-free buffer containing 0 mM (black trace), 0.3 mM (red trace), 0.6 mM (green trace) or 1.2 mM (blue trace) Ar-NHNH2.

Figure S3

Selected ion monitoring chromatogram demonstrating the production of Ar-NO2 from

Ar-NHNH2 catalyzed by AurF: Reaction solutions (in 100 mM HEPES pH 7.5 and 10% glycerol buffer) containing 0.8 mM Ar-NHNH2 and 1 μM (purple), 2 μM (green), 1

μM (blue), or 1 μM (red) as-isolated AurF were incubated on at 0 °C in the presence of excess O2 for 10 min, supplied with 50 μM 2,3,5,6-2H-4-nitrobenzoate as the internal standard, and filtrated. The filtrates were analyzed by LC/MS directly as described in the methods section.

Figure S4

Selected ion monitoring chromatogram demonstrating the incorporation of oxygen atoms to Ar-NO2 from Ar-NHNH2 oxidation catalyzed by AurF: Reaction solutions

(in 100 mM HEPES pH 7.5 and 10% glycerol buffer) containing 0.8 mM Ar-NHNH2 and 10 μM as-isolated AurF was incubated on at 0 °C in the presence of either excess

16O2 or excess 18O2 for 10 min and filtrated. The filtrates were analyzed by LC/MS directly as described in the methods section. A similar control sample leaving out

AurF was prepared (in the presence of 16O2) and analyzed.

Figure S5.

Selected ion monitoring chromatogram demonstrating the product with m/z = 297 from Ar-NHNH2 oxidation catalyzed by AurF: Reaction solutions (in 100 mM

HEPES pH 7.5 and 10% glycerol buffer) containing 0.8 mM Ar-NHNH2 and 10 μM as-isolated AurF was incubated on at 0 °C in the presence of either excess O2 for 10 min and filtrated. The filtrates were analyzed by LC/MS directly as described in the methods section. A similar control sample leaving out AurF was prepared and analyzed.

References

1. Hirata, Y., et al., Structure of aureothin, a nitro compound obtained from

Streptomyces thioluteus. Tetrahedron, 1961. 14: p. 252-274.

2. He, J. and C. Hertweck, Biosynthetic origin of the rare nitroaryl moiety of the

polyketide antibiotic aureothin: Involvement of an unprecedented N-oxygenase.

J. Am. Chem. Soc., 2004. 126(12): p. 3694-3695.

3. Wallar, B.J. and J.D. Lipscomb, Dioxygen activation by enzymes containing

binuclear non-heme iron clusters. Chem. Rev., 1996. 96(7): p. 2625-2657.

4. Merkx, M., et al., Dioxygen activation and methane hydroxylation by soluble

methane monooxygenase: a tale of two irons and three proteins. Angew.

Chem., Int. Ed., 2001. 40(15): p. 2782-2807.

5. Nordlund, P. and P. Reichard, Ribonucleotide reductases. Annu. Rev.

Biochem., 2006. 75: p. 681-706.

6. Winkler, R. and C. Hertweck, Sequential enzymatic oxidation of aminoarenes

to nitroarenes via hydroxylamines. Angew. Chem., Int. Ed., 2005. 44(26): p.

4083-4087.

7. Simurdiak, M., J. Lee, and H. Zhao, A new class of arylamine oxygenases:

evidence that p-aminobenzoate N-oxygenase (AurF) is a di-iron enzyme and

further mechanistic studies. ChemBioChem, 2006. 7(8): p. 1169-1172.

8. Winkler, R., et al., Regio- and chemoselective enzymatic N-oxygenation in vivo,

in vitro and in flow. Angew. Chem., Int. Ed., 2006. 45(47): p. 8016-8018.

9. Krebs, C., et al., AurF from Streptomyces thioluteus and a possible new family

of manganese/iron oxygenases. Biochemistry, 2007. 46(37): p. 10413-10418.

10. Choi, Y.S., et al., In vitro reconstitution and crystal structure of

p-aminobenzoate N-oxygenase (AurF) involved in aureothin biosynthesis. Proc.

Natl. Acad. Sci. U. S. A., 2008. 105(19): p. 6858-6863.

11. Korboukh, V.K., et al., A long-lived, substrate-hydroxylating

peroxodiiron(III/III) intermediate in the amine oxygenase, AurF, from

Streptomyces thioluteus. J. Am. Chem. Soc., 2009. 131(38): p. 13608-13609.

12. Li, N., et al., Four-electron oxidation of p-hydroxylaminobenzoate to

p-nitrobenzoate by a peroxodiferric complex in AurF from Streptomyces

thioluteus. Proc Natl Acad Sci U S A, 2010. 107(36): p. 15722-7.

13. Li, N., et al., Four-electron oxidation of p-hydroxylaminobenzoate to

p-nitrobenzoate by a peroxodiferric complex in AurF from Streptomyces

thioluteus. Proc. Natl. Acad. Sci. USA, 2010. 107(36): p. 15722–15727.

14. Whitteck, J.T., R.M. Cicchillo, and W.A. van der Donk, Hydroperoxylation by

hydroxyethylphosphonate dioxygenase. J. Am. Chem. Soc., 2009. 131(44): p.

16225-16232.

15. Warui, D.M., et al., Detection of formate, rather than carbon monoxide, as the

stoichiometric co-product in conversion of fatty aldehydes to alkanes by a

cyanobacterial aldehyde decarbonylase. J. Am. Chem. Soc., 2011. 133: p.

3316-3319.

16. Brown, C.A., et al., Spectroscopic and Electronic Structure Studies of

met-Hemerythrin Model Complexes: A Description of the Ferric-Oxo Dimer

Bond. Inorganic Chemistry, 1995. 34(3): p. 688-717.

17. Itano, H.A., Phenyldiimide, hemoglobin, and Heinz bodies. Proc Natl Acad Sci

U S A, 1970. 67(2): p. 485-92.

18. Misra, H.P. and I. Fridovich, The oxidation of phenylhydrazine: superoxide

and mechanism. Biochemistry, 1976. 15(3): p. 681-687.

19. Tomasi, A., et al., Detection of Free Radical Intermediates in the Oxidative

Metabolism of Carcinogenic Hydrazine Derivatives. Toxicologic Pathology,

1987. 15(2): p. 178-183.

20. Huang, P.-K.C. and E.M. Kosower, Diazenes. III. Properties of phenyldiazene.

Journal of the American Chemical Society, 1968. 90(9): p. 2367-2376.

21. Patek, D.R. and L. Hellerman, Mitochondrial Monoamine Oxidase. Journal of

Biological Chemistry, 1974. 249(8): p. 2373-2380.

22. Sinha, B.K., Enzymatic activation of hydrazine derivatives. A spin-trapping

study. Journal of Biological Chemistry, 1983. 258(2): p. 796-801.

23. Kalyanaraman, B. and B.K. Sinha, Free radical-mediated activation of

hydrazine derivatives. Environ Health Perspect, 1985. 64: p. 179-84.

24. Hill, H.A.O. and P.J. Thornalley, The effect of spin traps on

phenylhydrazine-induced haemolysis. Biochimica et Biophysica Acta (BBA) -

Molecular Cell Research, 1983. 762(1): p. 44-51.

25. Huang, P.-K.C. and E.M. Kosower, Diazenes. I. Decarboxylation of

phenyldiazene-carboxylic acid. Journal of the American Chemical Society,

1968. 90(9): p. 2354-2362.

26. Huang, P.-K.C. and E.M. Kosower, Diazenes. II. Preparation of phenyldiazene.

Journal of the American Chemical Society, 1968. 90(9): p. 2362-2367.

Appendix A

Four-electron oxidation of p-hydroxylaminobenzoate to p-nitrobenzoate by a peroxodiferric complex in AurF from Streptomyces thioluteus

Li et al., Proc Natl Acad Sci U S A. 2010;107(36):15722-7

Abstract

Introduction

Scheme 1. Reactions catalyzed by AurF.

Results

III/III Testing for a Reaction Between Peroxo-Fe2 -AurF and Ar-NHOH by Stopped-Flow

Absorption (SF-Abs) Experiments.

III/III Evaluation of Di-iron Products in Reaction of Peroxo-Fe2 -AurF with Ar-NHOH by

Mössbauer Spectroscopy.

III/III Figure 1. Sequential-mixing SF-Abs experiments to monitor the reaction of peroxo-Fe2 -AurF with Ar-NHOH.

100

III/III Figure 2. 4.2-K/53-mT Mössbauer spectra of samples in which peroxo-Fe2 -AurF was reacted with Ar-NHOH or Ar-NH2.

101

II/II Verification of Catalytic Oxidation of Ar-NHOH by Fe2 -AurF.

III/III Testing for Reduction of µ-oxo-Fe2 -AurF by Ar-NHOH.

III/III Re-evalution of the Diiron Products from the Reaction of Peroxo-Fe2 -AurF with Limiting

Ar-NH2.

Figure 3. Reversed-phase high performance liquid chromatography (RP-HPLC) of the small II/II molecule reactants and products following incubation of Fe2 -AurF with excess Ar-NHOH and O2.

102

Discussion

Scheme 2. Proposed mechanism of the four-electron oxidation of Ar-NHOH to Ar-NO2 by

III/III peroxo-Fe2 -AurF.

103

Experimental Procedures

Acknowledgments

References

104

SUPPORTING INFORMATION

Four-electron oxidation of p-hydroxylaminobenzoate to

p-nitrobenzoate by a peroxodiferric complex in AurF from

Streptomyces thioluteus.

Ning Lia,1, Victoria Korneeva Korboukha,b,1, Carsten Krebsa,b,2, and J. Martin Bollinger, Jr.a,b,2

105

Scheme S1. Proposed pathways for the conversion of Ar-NH2 to Ar-NO2 by AurF.

106

III∕III Figure S1. 4.2-K∕53-mT Mössbauer reference spectrum of peroxo-Fe2 -AurF.

107

III∕III Figure S2. 4.2-K∕53-mT Mössbauer spectra of samples from the reaction of peroxo-Fe2 -AurF with Ar-NHOH in the presence of limiting O2.

108

III∕III Figure S3. 4.2-K∕53-mT Mössbauer spectra of samples from the reaction of peroxo-Fe2 -AurF with Ar-NHOH in the presence of excess O2. Figure S4. Reversed-phase high performance liquid chromatographic (RP-HPLC) analysis of the synthetic Ar-NHOH upon its incubation in 100 mM HEPES buffer (pH 7.5).

109

III∕ Figure S5. 4.2-K∕53-mT Mössbauer spectra of samples from the treatment of as-isolated μ-oxo-Fe2 III-AurF with Ar-NHOH. 1 Figure S6. 360 MHz H-NMR spectrum of 5 mg crude Ar-NHOH dissolved in 1 g DMSO-d6. 1 Figure S7. 360 MHz H-NMR spectrum of 5 mg crude Ar-NHOH dissolved in 1 g DMSO-d6, to which 100 2 μL H2O was added and allowed to reach equilibrium for 24 h.

110

Appendix B

Detection of formate, rather than carbon monoxide, as the stoichiometric coproduct in conversion of fatty aldehydes to alkanes by a cyanobacterial aldehyde decarbonylase

Warui et. al., (Li co-first) J Am Chem Soc. 2011;133(10):3316-9

111

Scheme 1. Three possible outcomes of the Np AD reaction.

ABSTRACT

INTRODUCTION, RESULTS, AND DISCUSSION

112

Table 1. Concentrations of formate and heptadecane measured in Np AD reactions under two different sets of conditions.

Figure 1. Reconstructed mass spectra showing formate production in reactions of Np AD.

113

114

ACKOWLEDGEMENT

REFERENCES

115

SUPPORTING INFORMATION

Detection of formate, rather than carbon monoxide, as the

stoichiometric coproduct in conversion of fatty aldehydes to

alkanes by a cyanobacterial aldehyde decarbonylase.

Douglas M. Warui†§, Ning Li‡§, Hanne Nørgaard†, Carsten Krebs*†‡, J. Martin

Bollinger, Jr.*†‡, and Squire J. Booker*†‡

116

Materials and Methods

Materials. NADPH, spinach ferredoxin and ferredoxin reductase, tris-(2-carboxyethyl) phosphine hydrochloride (TCEP), phenylmethanesulfonyl fluoride (PMSF), lysozyme, ribonuclease, n-octadecanol, heptadecane, bovine myoglobin, sodium dithionite, 13 2 2-nitrophenylhydrazine (NPH), [ C]formate (sodium salt), deuterium oxide ( H2O), lithium aluminum hydride (LAH), and lithium aluminum deuteride were all obtained from Sigma–Aldrich (St. Louis, MO). N-(2-Hydroxyethyl)piperazine-N’-2-ethane sulfonic acid (HEPES) was purchased from Fisher Scientific (Pittsburgh, PA), and imidazole was purchased from J. T. Baker Chemical Co (Phillipsburg, NJ). Pyridinium chlorochromate was obtained from TCI America (Portland, OR) and N-(3-dimethylaminopropyl)-N’-ethylcarbodiimide (EDC) was from Chem-Impex (Wood Dale, IL), while isopropyl -D-thiogalactopyranoside (IPTG) was obtained from Gold BioTechnology (St. Louis, MO). [1-13C]-Octadecanoic acid and heptadecane-d36 were obtained from Cambridge Isotope Laboratories, Inc (Andover, MA), while nickel nitrilotriacetic acid resin (Ni-NTA) was obtained from Qiagen (Valencia, CA). Sephadex G-25 resin was obtained from GE-Biosciences (Piscataway, NJ), while all other buffers and chemicals were of the highest grade available.

General Methods. UV–visible absorption spectra were recorded on a Cary 50 spectrophotometer from Varian (Walnut Creek, CA) or an Agilent (Foster City, CA) 8453 diode-array spectrophotometer, both using their associated software for operating the instrument and processing the data. Sonic disruption of E. coli cell suspensions was conducted with a 550 sonic dismembrator from Fisher Scientific using a horn containing a 1/2 inch tip.

Plasmid design, cell growth, and protein purification. The DNA sequence that encodes Nostoc punctiforme (Np) aldehyde decarbonylase (AD) was codon-optimized for over-expression in E. coli, synthesized, and inserted into the NdeI and EcoRI restriction sites of expression vector pET-28a by GeneArt (Regensburg, Germany). This plasmid construct, which places the gene under the control of a T7 promoter, allows overproduction of the protein containing an N-terminal hexahistidine tag separated from its native start codon by a spacer of 10 amino acids. The codon-optimized gene sequence is shown below, which was verified at the Penn State

117

University Molecular Core Sequencing Facility. The resulting plasmid, designated pNpADwt, was used to transform E. coli BL21 (DE3) (Invitrogen; Carlsbad, CA) for protein production. Protein overproduction and purification followed procedures similar to those reported by Schirmer et al.,1 with minor differences as noted. Protein expression was carried out in shake flasks at 37 C in Luria–Bertani (LB) media supplemented with 50 µg/mL kanamycin, and was induced at an OD600nm of 0.8 by addition of IPTG to a final concentration of 1 mM. The culture was shaken at 37 C for 3 additional hours before being harvesting by centrifugation at 5000  g for 15 min. Cell paste was frozen in liquid nitrogen and stored at -80 C until needed.

All purification steps were performed at 4 ºC unless noted otherwise. In a typical purification, 30 g cell paste was resuspended at room temperature in 80 mL lysis buffer [20 mM sodium phosphate buffer (pH 8.0), 150 mM NaCl, 10 mM imidazole and 1 mM TCEP]. Lysozyme and ribonuclease were added to final concentrations of 0.2 mg/mL and the suspension incubated on ice for an additional 30 min. PMSF, dissolved in a minimal volume of ethanol, was then added to a final concentration of 1 mM, upon which the cells were subjected to four 1 min bursts of sonic disruption at a setting of 7 while on ice, with 8 min of cooling following each sonic burst. The sonicated cell lysate was then centrifuged at 10,000  g for 30 min at 4 C to remove cellular debris. The cleared cell lysate was mixed with Ni-NTA slurry and incubated on ice for 60 min with occasional gentle shaking, then loaded into a column. The resin was washed with wash buffer [50 mM sodium phosphate buffer (pH 8.0), 150 mM NaCl, 40 mM imidazole and 1 mM TCEP] before the protein was eluted with elution buffer [50 mM sodium phosphate buffer (pH 8.0), 150 mM NaCl, 250 mM imidazole and 1 mM TCEP]. The purified protein was exchanged into gel-filtration buffer [50 mM Tris-HCl (pH 7.2), 50 mM NaCl, 10% glycerol and 1 mM TCEP] by size-exclusion chromatography on Sephadex G-25. Protein concentration was determined by UV absorption at 280 nm using a calculated molar absorptivity of 22,920 M-1 cm-1 (http://ca.expasy.org). From 30 g of cell paste, ~800 mg of protein was obtained. It was shown by SDS–PAGE (12%) to be >98% pure (see below).

Codon-optimized gene sequence of Nostoc punctiforme aldehyde decarbonylase. 5’-ATGCAGCAGCTGACCGATCAGAGCAAAGAACTGGATTTTAAAAGCGAAACCTATA AAGATGCATATAGCCGTATTAATGCAATTGTTATTGAAGGTGAACAGGAAGCACATGA AAATTATATTACCCTGGCACAGCTGCTGCCGGAAAGCCATGATGAACTGATTCGTCTG AGCAAAATGGAAAGCCGTCATAAAAAAGGTTTTGAAGCATGTGGTCGTAATCTGGCA 118

Coomassie-blue stained 12% SDS–PAGE analysis of hexahistidine-tagged Np AD (28 kDa). Lane 1, pellet from cell lysis; lane 2, soluble fraction of lysate; lane 3, molecular mass markers; lane 4, wash fraction; lane 5, final purified Np AD. GTTACCCCGGATCTGCAGTTTGCAAAAGAATTTTTTAGCGGTCTGCATCAGAATTTTC AGACCGCAGCAGCAGAAGGTAAAGTTGTTACCTGTCTGCTGATTCAGAGCCTGATTAT TGAATGTTTTGCAATTGCAGCATATAATATTTATATTCCGGTTGCAGATGATTTTGCACG TAAAATTACCGAAGGTGTTGTTAAAGAAGAATATAGCCATCTGAATTTTGGTGAAGTT TGGCTGAAAGAACATTTTGCAGAAAGCAAAGCAGAACTGGAACTGGCAAATCGTCA GAATCTGCCGATTGTTTGGAAAATGCTGAATCAGGTTGAAGGTGATGCACATACCATG GCAATGGAAAAAGATGCACTGGTTGAAGATTTTATGATTCAGTATGGTGAAGCACTGA GCAATATTGGTTTTAGCACCCGTGATATTATGCGTCTGAGCGCATATGGTCTGATTGGT GCATAA-3’

Synthesis of n-octadecanal substrate. The synthesis of n-octadecanal (R-CHO, where R = 1 n-C17H35) was carried out as described by Schirmer et al. n-Octadecanol (1 g) was stirred overnight at room temperature with 0.4 grams of pyridinium chlorochromate in 75 mL dichloromethane. Solvent was removed from the reaction by rotary evaporation to dryness, and the resulting solute was redissolved in hexane and filtered through a Whatman filter paper. The solvent was again removed by rotary evaporation, and the solute was resuspended in a minimal volume of hexane before loading onto a gravity-fed silica column equilibrated in hexane. The column was washed with two column-volumes of hexane before the product was eluted by an 8:1 mixture of hexane:ethyl acetate. Fractions that contained the octadecanal product were

119 identified by thin-layer chromatography (TLC) (vanillin stain) and pooled and dried down before being dissolved in ethyl acetate. The octadecanal product was >95% pure as determined by GC-MS (see below).

Synthesis of isotopically labeled n-octadecanal substrates (R-13CHO and R-13C2HO). [1-13C]-n-Octadecanoic acid was first reduced to the corresponding primary alcohol using lithium aluminium hydride (LAH). [1-13C]-n-Octadecanoic acid (1 g) was dissolved in 20 mL tetrahydrofuran (THF). LAH (1 g) was added slowly to the solution, which was then stirred for 2 h at room temperature (22 °C). Solvent was removed by rotary evaporation, and the resulting mixture was resuspended in dichloromethane (CH2Cl2) and filtered. The filtrate was dried down, redissolved in CH2Cl2, and then dried down again. The resulting mixture, which contained the alcohol product, was resuspended in 75 mL CH2Cl2. Pyridinium chlorochromate (1 g) was added to oxidize the alcohol to the aldehyde. The reaction was allowed to proceed overnight at room temperature while stirring, and the octadecanal product was purified as described above. The purified product was determined to be >95 % pure by GC-MS. [1-13C, 1-2H]-n-Octadecanal was synthesized similarly, except that lithium aluminum deuteride was used to reduce the acid to the alcohol.

Assays. Most assays (see amendments in figure legends) were conducted at room temperature in a final volume of 100 µL, and contained the following components: 50 µM Np AD, 500 µM octadecanal, 1 mM NADPH, 50 µg/mL ferredoxin, 50 mU/mL spinach ferredoxin reductase, 0.1 % triton X-100, 100 mM HEPES buffer, pH 7.4, 10 % glycerol. For heptadecane analysis, assays were quenched and extracted by addition of an equal volume of ethyl acetate subsequent to the addition of heptadecane-d36 as an internal standard. Extracted samples were analyzed on a Shimadzu GC-17A gas chromatograph connected to a Shimadzu GCMS-QP500 mass spectrometer using a HP-5-MS 30 m column (ID: 0.25 narrow bore; film: 0.25 µm) (Agilent Technologies; Foster City, CA). The inlet and oven temperatures were set to 320 and 80 °C, respectively. Upon injection, the oven temperature was maintained at 80 °C for 5 min before being ramped up to 320 °C at 35 °C/min. Upon reaching 320 °C, this temperature was held for an additional 5 min. Total ion chromatograms were generated by scanning from 50-500 m/z. Under conditions of single ion monitoring (SIM), m/z values of 71 and 82 were monitored.

120

Formate analysis was conducted according to a literature procedure2 as amended below. Reactions (100 µL) were quenched by addition of 10 µL of 1 mM propionic acid (internal standard) dissolved in a 1:1 solution of pyridine:HCl. EDC (10 µL of a 0.29 M solution) and NPH (10 µL of a 0.12 M solution in 250 mM HCl) were added, and the solution was incubated at 60 °C for 15 min. Precipitated protein was pelleted by centrifugation at 13,00 rpm, and a 10 µL aliquot of the supernatant was analyzed by LC-MS (Waters LC and ZQ mass detector) using a

Hamilton PRP-1 analytical column. An isocratic mobile phase (2:1 methanol:H2O with 0.05% acetic acid) was used for separation with SIM monitoring at m/z values of 180 (formate), 181 (13C-formate), 182 (13C,2H-formate), and 208 (propionic acid). Standard curves (Figure S4) were generated by adding known amounts of formate, 13C-formate, and propionic acid to reaction buffer and subjected the solution to the above analysis.

Quantification of CO was conducted by assessing its ability to bind to the Fe(II) site of dithionite-reduced myoglobin, shifting its Soret band from 434 nm to 423 nm as described by Iizuka et al.3 Assays, as described above, were incubated in septum-sealed vials for designated reaction times, after which a solution of bovine myoglobin and sodium dithionite was injected to give final concentrations of 10 µM and 20 mM, respectively. The solution was allowed to stand for 10 min before the vial was opened and the absorption spectrum recorded. For comparison to the effects of known quantities of CO, samples were generated by injecting varying amounts of CO-saturated water into assay mixtures and incubating them for 3 h under assay conditions before addition of myoglobin and dithionite.

121

Figure S1. GC-MS analysis of the decarbonylation of n-octadecanal to heptadecane by Np AD and confirmation of the requirement for the reducing system. As shown from the results, Np AD decarbonylates n-octadecanal to produce heptadecane (red trace) only in the presence of the complete reducing system (ferredoxin, ferredoxin reductase and NADPH), and much less or no product is observed when any of the components of the reducing system is omitted from the reaction.

122

Figure S2. GC-MS total ion chromatograms demonstrating the dependence of n-C17H36 yield on time (A) and Np AD concentration (B) in the reaction with R-13CHO. A) For four separate 100 µL samples containing 1 mM NADPH, 50 µg/mL ferredoxin, 50 mU/mL ferredoxin reductase, 13 2 0.1% triton x-100, 100 mM HEPES (pH 7.4), 10% glycerol, 500 µM R- CHO, 50 µM n-C17 H36 and 100 µM Np AD, the reaction was either terminated immediately after addition of enzyme by adding 50 µL ethyl acetate (black) or incubated at 37 °C for 15 min (green), 30 min (blue) or120 min (red) before quenching and analysis by GC/MS. (B) Three 100 µL reactions containing 0 µM Np AD (black), 50 µM Np AD (blue) or 200 µM Np AD (red) with 1 mM NADPH, 50 µg/mL ferredoxin, 50 mU/mL ferredoxin reductase, 0.1% triton x-100, 100 mM HEPES (pH 7.4), 13 2 10% glycerol, 500 µM R- CHO, 50 µM n-C17 H36 were incubated at 37 °C overnight (14 hours) followed by extraction with 50 µL ethyl acetate and GC/MS analysis of the supernatant.

123

Figure S3. Myoglobin/UV-absorption assay for detection of CO. Two assay solutions containing 50 µM Np AD, 1 mM NADPH, 50 µg/mL ferredoxin, 50 mU/mL ferredoxin reductase, 0.2% triton x-100, 100 mM HEPES (pH 7.4), 10% glycerol, and 500 µM R-13CHO were either assayed for CO immediately (black) after mixing or assayed after incubation at room temperature for 3 hours (red). Spectra corresponding to known amounts of CO were generated by injecting 0 (blue), 5 (green), 10 (purple), 20 (orange), 40 (light blue), 80 (hot pink), 160 (dark green) µM (final concentration) CO saturated water, respectively, into the reactions (all lacking NADPH), which were then incubated at room temperature for 3 hours before assaying for CO.

124

Figure S4. LC-MS calibration curves for formate, 1-[13C]-formate, and propionate. Standard curves were generated by mixing different known concentrations (0, 20, 50, 100 and 200 µM) of - 13 13 - - authentic formate (HCO2 ), 1-[ C]-formate (H CO2 ) and propionate (C2H5CO2 ) in buffer solutions. The 2-phenylhydrazide derivatives of each were analyzed by LC-MS as described above.

125

Figure S5. Dependence of formate yield on reaction time (compare front to back traces) and Np AD concentration (compare different color traces within a group). Six 100 µL reactions containing 0 µM Np AD (grey and black), 50 µM Np AD (green and blue) or 100 µM Np AD (purple and red), each with 1 mM NADPH, 50 µg/mL ferredoxin, 50 mU/mL ferredoxin reductase, 0.1% triton x-100, 100 mM HEPES (pH 7.4), 10% glycerol, and 500 µM R-13CHO were incubated at 37 °C for either 1 h or 4 h (indicated in the figure). Analysis for formyl-2-nitrophenylhydrazide by LC-MS with detection by SIM (m/z = 181) was carried out as described above.

126

Figure S6. Analysis of deuterium incorporation into heptadecane upon reaction of Np AD in 2 H2O (A) and 78% H2O (B). Two separate 100 µL reactions containing 1 mM NADPH, 50 µg/mL ferredoxin, 50 mU/mL ferredoxin reductase, 0.1% triton x-100, 100 mM HEPES (pH 7.4), 13 2 10% glycerol, 500 µM R- CHO, 50 µM n-C17 H36 and 50 µM Np AD were prepared in H2O (A) 2 or 78% H2O (B). Reactions were incubated at room temperature for 13 h and then extracted with 50 µL ethyl acetate. Alkane products were analyzed by GC/MS with detection by SIM. 2 + 2 Fragment ions corresponding to m/z values 82 (n-C5 H11 ; black) of the n-C17 H36 internal + 2 + 2 standard, 71 (n-C5H11 ; red) of n-C17H36, and 72 (n-C5 HH10 ; blue) of n-C17 HH35 were monitored.

127

References

(1) Schirmer, A., Rude, M. A., Li, X., Popova, E., del Cardayre, S. B. (2010) Science, 329, 559–562 (2) Whitteck, J. T., Cicchillo, R. M., van der Donk, W. A. (2009) J. Am. Chem. Soc., 131, 16225–16232 (3) Iizuka, T., Yamamoto, H., Kotani, M. Yonetani, T. (1974) Biochim., Biophys., Acta, 371, 126–139

128

Appendix C

Conversion of fatty aldehydes to alka(e)nes and formate by a cyanobacterial aldehyde decarbonylase: cryptic redox by an unusual dimetal oxygenase

Li et. al., J Am Chem Soc. 2011;133(16):6158-61

129

ABSTRACT

INTRODUCTION, RESULTS AND DISCUSSION

130

Scheme 1. Two Alternative Explanations for the Similarity of Np AD to Di-iron Oxidases and

Oxygen-ases and its Requirement for a Reducing System to Promote an Apparently Hydrolytic Reaction.

Figure 1. Reconstructed mass spectra illustrating the catalytic requirement for O2 and the incorporation 18 18 of O from O2 into the formate product in the Np AD reaction.

131

Scheme 2. Hypothetical Mechanism for the Np AD Reaction.

132

ACKNOWLEDGEMENT

REFERENCES

133

SUPPORTING INFORMATION

Conversion of Fatty Aldehydes to Alka(e)nes and Formate by a

Cyanobacterial Aldehyde Decarbonylase: Cryptic Redox by an

Unusual Dimetal Oxygenase

Ning Li†, Hanne Nørgaard‡, Douglas M. Warui‡, Squire J. Booker*†‡, Carsten

Krebs*†‡, and J. Martin Bollinger, Jr.*†‡

134

Figure S1. Determination of the ratio of NADPH oxidized to formate produced during the Np AD reaction. The main panel shows spectra after 10-fold dilution of the reaction solution at the indicated times, and the inset shows the quantities of 13C-formate produced (circles) and NADPH consumed (triangles; calculated according to the molar absorptivity of NADP at 340 nm of 6.22 mM-1cm-1) at each reaction time. The reaction contained, in a final volume of 0.40 mL, 0.050 mM Np AD, 0.50 mM 1-[13C]-octadecanal, 0.10 mg/ml each of spinach ferredoxin and ferredoxin reductase (from Sigma-Aldrich), and 2.0 mM NADPH in air-saturated 100 mM HEPES buffer, pH 7.4, containing 0.2% triton X-100 detergent. Immediately after constitution of the reaction solution, an aliquot was diluted by 10-fold with buffer and the absorption spectrum of this solution recorded; a second aliquot was immediately subjected to the 2NPH coupling procedure (which terminates the reaction), and the formate was later quantified by LC-MS. Thereafter, the reaction was allowed to proceed at 23 °C for 1 h (green) or 2 h (purple) before aliquots were similarly analyzed. For NADPH, the quantities shown represent the decrease relative to the first time point.

135

Appendix D

Evidence for Only Oxygenative Cleavage of Aldehydes to Alk(a/e)nes and Formate by Cyanobacterial “Aldehyde Decarbonylase”

Li, et. al.

136 Evidence for Only Oxygenative Cleavage of

Aldehydes to Alk(a/e)nes and Formate by

Cyanobacterial “Aldehyde Decarbonylase”†

Ning Li,a Wei-chen Chang,b Douglas M. Warui,b Squire J. Booker,a,b,* Carsten Krebs,a,b,* J. Martin

Bollinger, Jr.a,b,*

Departments of aBiochemistry and Molecular Biology and of bChemistry, The Pennsylvania State

University, University Park, Pennsylvania 16802

†This work was supported by the National Science Foundation (MCB-1122079 to CK, SJB, and JMB).

AUTHOR EMAIL ADDRESS [email protected], [email protected], [email protected]

RECEIVED DATE

Please send correspondence to: J. Martin Bollinger, Jr. Carsten Krebs Squire J. Booker Department of Chemistry Department of Chemistry Department of Chemistry 336 Chemistry Building 332 Chemistry Building 302 Chemistry Building University Park, PA 16802 University Park, PA 16802 University Park, PA 16802 Phone: 814-863-5707 Phone: 814-865-6089 Phone: 814-865-8793 Fax: 814-865-2927 Fax: 814-865-2927 Fax: 814-865-2927

137 1Abbreviations used: 2NPH, 2-nitrophenylhydrazide; ACP, acyl carrier protein; AD, aldehyde decarbonylase; ADO, aldehyde-deformylating oxygenase; amu, atomic mass unit; DMSO, dimethylsulfoxide; Ec, Escherichia coli; GC, gas chromatography; IPTG, isopropyl-β-D1-thiogalactopyranoside; LB, Luria-Bertani; LC, liquid chromatography; MeOPMS,

1-methoxy-N-methylphenazine methosulfate; MS, mass spectrometry; N/F/FR, reducing system comprising NADPH, ferredoxin, and ferredoxin reductase from spinach; NMR, nuclear magnetic resonance; N/PMS, reducing system comprising NADH and N-methyl-phenazinium methylsulfate; Np,

Nostoc punctiforme; Pm, Prochlorococcus marinus; PMS, phenazine methosulfate or

N-methyl-phenazinium methylsulfate; SIM, single ion monitoring.

2This particular substrate isotopomer was selected as the synthetically simplest isotopic perturbation that would afford mass-resolution of enzymatically produced formate from the “environmental” formate contaminant. Our previous work established that the C1 hydrogen of the aldehyde substrate is fully retained in the formate product, giving the formate an m/z of 46, equivalent to that produced from a

1-[13C]-labeled aldehyde.

138 ABSTRACT

Cyanobacterial “aldehyde decarbonylases” (ADs) catalyze conversion of Cn fatty aldehydes to formate

– (HCO2 ) and the corresponding Cn-1 alk(a/e)nes. Previous studies of the Nostoc punctiforme (Np) AD produced in Escherichia coli (Ec) showed that this apparently hydrolytic reaction is actually a cryptically redox oxygenation process, in which one atom from O2 is incorporated into formate and a protein-based reducing system (NADPH, ferredoxin, and ferredoxin reductase; N/F/FR) provides all four electrons needed for complete reduction of O2. Two subsequent publications by Marsh and co-workers reported that their Ec-expressed Np and Prochlorococcus marinus (Pm) AD preparations transform aldehydes to the same products more rapidly by an O2-independent, truly hydrolytic process, which they suggested to proceed by transient substrate reduction with obligatory participation by the reducing system (they used a chemical system, NADH and phenazine methosulfate; N/PMS). To resolve this discrepancy, we re-examined our preparations of both AD orthologs by a combination of (i) activity assays in the presence

18 18 and absence of O2 and (ii) O2 and H2 O isotope-tracer experiments with direct mass-spectrometric

– detection of the HCO2 product. For multiple combinations of AD ortholog (Np and Pm), reducing system

(protein-based and chemical), and substrate (n-heptanal and n-octadecanal), our preparations strictly require O2 for activity and do not support detectable hydrolytic formate production, despite having catalytic activities similar to or greater than those reported by Marsh and co-workers. Our results, especially of the 18O-tracer experiments, suggest that the activity observed by Marsh and co-workers could have arisen from contaminating O2 in their assays. The definitive reaffirmation of the oxygenative nature of the reaction implies that the enzyme, initially designated as aldehyde decarbonylase when the

C1-derived co-product was thought to be carbon monoxide rather than formate, should be re-designated as “aldehyde-deformylating oxygenase” (ADO).

139 A recently discovered orthologous group of ferritin-like non-heme dimetal-carboxylate enzymes from cyanobacteria catalyzes the second of two enzymatic steps through which fatty acids linked to acyl carrier protein (ACP)1 are converted to diesel-fuel alk(a/e)nes (1). This pathway has been touted as a potential basis for a sunlight-driven, carbon-neutral bioprocess to renewable, fungible fuels (1-15).

Schirmer, et al. identified the cyanobacterial genes encoding these enzymes and demonstrated the ability of the Nostoc punctiforme (Np) ortholog, isolated after over-expression in Escherichia coli (Ec), to convert n-octadecanal (which is produced by the first enzyme in the pathway, acyl-ACP reductase) to heptadecane (R-H) in vitro (1). They suggested that the other product might be carbon monoxide, and therefore named the enzyme “aldehyde decarbonylase” (AD) (1). We subsequently showed that the

– C1-derived co-product is actually formate (HCO2 ) (16). Conversion of a Cn aldehyde to the corresponding Cn-1 alk(a/e)ne and formate is an apparently redox-neutral, formally hydrolytic outcome, but the structure of the Prochlorococcus marinus (Pm) ortholog, which revealed the enzymes’ similarity to other ferritin-like diiron-carboxylate oxidases and oxygenases (1), and the dependence of in vitro activity on the presence of a reducing system (NADPH, ferredoxin, and ferredoxin reductase; N/F/FR) analogous to those employed by such oxidases/oxygenases (17, 18) hinted that the enzyme could be an oxygenase. We subsequently confirmed this possibility by showing that O2 is also required for activity

– and that the O-atom incorporated into the HCO2 product originates from O2, implying a reaction stoichiometry of four electrons from the reducing system per aldehyde cleaved and establishing the reaction as an unusual oxygenation process (Scheme 1A) (19). This conclusion, which is definitively reaffirmed by the results presented below, implies that the enzyme is more aptly designated as

“aldehyde-deformylating oxygenase” (ADO), a nomenclature that we hereby adopt.

Shortly after our studies were published, Marsh and co-workers reported that their preparations of the Np and Pm ADOs catalyze cleavage of aldehydes to alk(a/e)nes and formate in the presence of a reducing system but in the absence of O2 (Scheme 1B) (20, 21). Their work raised the stunning possibility that the same enzyme might catalyze two fundamentally different reactions, the first involving reductive activation of O2 and its subsequent attack on the aldehyde carbonyl in a manner distinct from, but clearly

140 related to, mechanisms of other well-studied diiron oxidases/oxygenases, and the second, for which they

II/II posited a mechanism involving transient one-electron reduction of the substrate carbonyl by the Fe2 cofactor and formation of an organometallic formyl-FeII intermediate, bearing little resemblance to any reaction previously attributed to a member of this well-studied enzyme family (20). Apart from this provocative central claim, Marsh and co-workers reported several other interesting observations. First, the chemical reducing system that they employed, NADH and phenazine methosulfate (N/PMS), reportedly supported greater activity than the N/F/FR protein-based reducing system used by both Schirmer, et al. (1) and us (16, 19), suggesting that the extremely modest in vitro turnover rates achieved in previous studies could have resulted from sluggish electron delivery to the ADO cofactor by the heterologous (spinach) ferredoxin. Second, they reported that the phenazine of the reducing system binds tightly to the enzyme, an observation hinting that heterocyclic redox cofactors might have a role in the reaction as it occurs in vivo (20). Finally, they noted that linear aldehydes as short as C7 are good substrates (20), an observation potentially resolving the challenge to mechanistic analysis presented by the very low solubility of longer-chain fatty aldehyde substrates.

The central claim of the Marsh work, aldehyde cleavage in the absence of O2, is inconsistent with our own published observations. Most importantly, we consistently detected much less product (< 20 µM) when O2 was intentionally removed prior to constitution of the reactions than when reactions were constituted with air-saturated solutions, which, under optimized conditions, yielded > 100 µM products

(19). Product yields in the O2-depleted reactions were invariably similar to the levels of residual O2 typically remaining after our standard deoxygenation procedure (22-26), consistent with a stoichiometric requirement for O2. By contrast, Marsh and co-workers reported that, in their experiments, inclusion of O2 actually diminished product yield (20). It is noteworthy that, in both of their studies, enzyme concentrations and product yields depicted in figures were generally very modest (tens of µM) and similar to our product yields in O2-depleted reactions. The low product yields and inconsistency with our published observations combined to raise concerns about the Marsh conclusion of O2-independent turnover by the ADs.

141 Isotope-tracer experiments to assess the origin of the O-atom incorporated into the formate product represent the ultimate arbiter for the nature of the ADO reaction (compare Scheme 1, reactions A and B). For example, in proving that the Np ADO can cleave its substrate by an oxygenative process, we

18 carried out the reaction under an atmosphere of O2, converted the formate to its 2-nitrophenylhydrazide

(2NPH) derivative (which extrudes one of the two O-atoms of formate), and showed that the formyl-2NPH derivative had ~ 34% 18O (compared to a content of 50% expected for production of entirely HC16O18O–) (19). Although shortcomings in the analysis – specifically the use of a chemical-coupling procedure that removes one of the two atoms being isotopically traced and partially exchanges the other with solvent (as a result of abortive derivatization events) – prevented a firm conclusion as to whether any hydrolytic cleavage occurs, the conclusion that oxygenative formate production occurs was entirely unambiguous (19). Analogously, following their initial claim of hydrolytic

18 activity (20), Marsh and co-workers carried out the reaction in H2 O in an attempt to verify the proposed solvent origin of the incorporated O-atom in their reactions (21). In this case, interpretation was further complicated by the potential for exchange of the substrate carbonyl O-atom with solvent 18O.

Theoretically, complete solvent exchange prior to enzyme-mediated formate production by the putative hydrolytic pathway would have resulted in the amide carbonyl in the formyl-2NPH analyte having precisely the same 18O/16O isotopic composition as the solvent [which, unfortunately, they did not report, thereby preventing critical evaluation of their conclusions (21)]. By contrast, in concluding that the observed result [61-64% 18O in the formyl-2NPH carbonyl (note that this is almost identical to the percentage of solvent-derived O-atoms in our prior analysis)] was consistent with a solvent origin for the incorporated O-atom, they invoked very slow exchange of the substrate carbonyl O-atom with solvent (40% in 2 h), which they reported to have observed experimentally but presented without documentation (21).

Importantly, if one were to invoke the opposite assumption of fast solvent exchange, as is commonly observed for aldehyde carbonyl oxygen atoms (27), then their result would imply precisely the opposite

16 conclusion – that the incorporated O-atom was derived from residual atmospheric O2. It is unclear why

18 Marsh and co-workers did not pre-incubate the aldehyde substrate in the H2 O solvent to permit complete

142 exchange prior to initiation of the ADO reaction and thereby eliminate this ambiguity. In our view, the combination of (i) the shortcomings in the execution and reporting of the crucial isotope-tracer experiment (the omission of solvent isotopic composition, the failure to pre-exchange the substrate with the solvent, the surprising and undocumented claim of very slow carbonyl-oxygen solvent exchange, and, most importantly, the use of an analytical procedure that removes one of the two O-atoms being traced and partially exchanges the other) and (ii) our prior, unequivocal demonstration of formate production by the oxygenative pathway cast serious doubt on the claim of hydrolytic activity and made it imperative to seek more definitive analysis.

We reasoned that mass-spectrometric (MS) analysis of the formate product itself, rather than its

2NPH derivative, would provide for a very sensitive isotopic probe for the proposed hydrolytic activity.

18 Following complete exchange of the substrate O-atom with H2 O solvent, any hydrolytic cleavage would

18 18 – produce HC O O whereas any oxygenative activity occurring in a reaction intended to be O2-free but

16 18 16 – containing residual atmospheric O2 would produce HC O O . Thus, the absence or presence of a signal for the M+4 product (where M is the mass of the formate with natural abundance of oxygen isotopes) in a

18 reaction carried out in H2 O should be rigorously dispositive regarding the existence of hydrolytic activity. It appears that direct MS analysis of formate has, in the past, proven difficult, undoubtedly owing to its prevalence in the environment. However, our previous studies showed that, with the substrate labeled with 13C or 2H at C1, the MS signal arising from ADO-generated formate is resolved from that arising from the "environmental" formate (16). In the present case, mass resolution was anticipated to be enhanced even further by the addition of 2 or 4 amu from the 18O in the HC18O16O– or HC18O18O– product, respectively.

In this work, we report results of the 18O-tracer experiments with direct MS analysis of the formate product that we have forecast above, in combination with simpler determinations of product yields in O2-replete and O2-depleted reactions, which show unequivocally that our ADO preparations simply do not support detectable O2-independent, hydrolytic cleavage, irrespective of the ADO ortholog

(Np or Pm), the reducing system (N/F/FR or N/PMS), or the substrate (n-heptanal or n-octadecanal)

143 employed. In addition, aided by the important technical advances of Marsh and co-workers (21), including the N/PMS reducing system (which we improved upon slightly by use of a further modified phenazine) and the more soluble, shorter-chain substrate (n-heptanal) on which they reported, we demonstrate turnover numbers (~ 0.3 s-1) approaching the 100-101 s-1 regime typical of diiron oxidases/oxygenases (25, 28-33), thus setting the stage for transient-state kinetic and spectroscopic dissection of the reaction mechanism.

EXPERIMENTAL PROCEDURES

Materials. Phenazine methosulfate (>90%), 2-Nitrophenylhydrazine (97%),

1-Methoxy-5-methylphenazinium methylsulfate (>95%), pyridine (>99.9%), spinach ferredoxin, spinach ferredoxin reductase, NADH (>97%), and NADPH (>97%) were purchased from Sigma-Aldrich.

1-(3-Dimethylaminopropyl)-3-ethylcarbodiimide (98%) and pyridinium dichlorochromate were

18 18 18 18 purchased from Alfa Aesar. O2 gas (99% O) was purchased from ICON Isotopes. H2 O (>97% H2 O),

13 13 13 2 2 1-[ C]-stearic acid, n-1-[ C]-octanoic acid, [ C]-formate, and B H3 (98% H, 1 M in THF) were purchased from Cambridge Isotope Laboratories.

Synthesis of substrates. n-1-[13C]-octadecanal was synthesized as described by Schirmer, et al. (1).

13 1-[ C]-stearic acid (1.0 g, 3.5 mmol) was reduced with BH3 (1M in THF solution; 4.2 mL, 4.2 mmol) in

THF (40 mL) and subsequently treated with pyridinium dichlorochromate (1.98 g, 5.3 mmol) in methylene chloride (40 mL). n-1-[13C]-octanal was prepared in the analogous manner from n-1-[13C]-octanoic acid. n-1-[2H]-heptanal was synthesized in identical fashion from n-1-heptanoic acid

2 2 using B H3 as reductant (1M B H3 solution in THF). Analytical thin layer chromatography (TLC) was carried out on pre-coated TLC aluminum plates (silica gel, grade 60, F254, 0.25 mm layer thickness) from

EMD Chemicals. Flash column chromatography was performed on silica gel (230-400 mesh, grade 60) obtained from Sorbent Technologies. Products were characterized by 1H and 13C nuclear magnetic resonance (NMR) spectroscopy (see the Supporting Information). NMR spectra were recorded on a

Brüker 300, 360 or 400 MHz spectrometer at the nuclear magnetic resonance facility of the Department

144 of Chemistry, the Pennsylvania State University. Chemical shifts (δ in ppm) were determined from the known signals of solvent (CDCl3 or d6-DMSO), and coupling constants are given in Hertz (Hz).

Preparation of Proteins. The DNA sequence that encodes Pm ADO was codon-optimized for over-expression in Ec, synthesized, and inserted into the NdeI and BamHI restriction sites of expression vector pET-28a by GeneArt (Regensburg, Germany). This plasmid construct, which places the gene under the control of a T7 promoter, allows for over-production of the protein containing an N-terminal His6-tag separated from its native start codon by a spacer of 10 amino acids. The resulting plasmid, designated pPmADOwt, was used to transform Ec BL21 (DE3) (Invitrogen; Carlsbad, CA) for protein production.

Protein overproduction and purification followed procedures similar to those reported by Schirmer, et al.

(1), with minor differences, as noted. Protein expression was carried out in shake flasks at 37 °C in

Luria–Bertani (LB) medium supplemented with 50 μg/mL kanamycin and was induced at an OD600nm of

0.8 by addition of isopropyl β-D-1-thiogalactopyranoside (IPTG) to a final concentration of 0.25 mM. The culture was shaken at 37 °C for 4 additional hours before being harvesting by centrifugation at 6,000 rpm for 20 min. Cell paste was frozen in liquid nitrogen and stored at -80 °C until it was used. All purification steps were performed according to the procedure reported previously (16, 19). Np ADO was prepared as described in previous work (16, 19).

Aldehyde Cleavage Reactions. All ADO reactions were carried out in 100 mM HEPES buffer, pH

7.5, containing 10% glycerol. Standard reactions with the n-octadecanal substrate and N/PMS reducing system contained 0.1 mM ADO, 0.5 mM n-1-[13C]-octadecanal, 0.3 mM PMS, 6 mM NADH, and 0.2% triton x-100. Reactions with the n-octadecanal substrate and N/F/FR reducing system contained 0.1 mM

ADO, 0.5 mM n-1-[13C]-octadecanal, 2 mM NADPH, 100 µg/mL ferredoxin, 100 mU/mL spinach ferredoxin reductase, and 0.2% triton x-100. Reactions with the n-heptanal substrate contained 0.1 mM

ADO, 16 mM n-1-[2H]-heptanal, 0.3 mM MeOPMS, and 6 mM NADH (with no detergent).

Analysis for Formate by Liquid Chromatography/Mass Spectrometry (LC/MS). Reaction samples for direct formate analysis were filtered through Amicon spin filters with a molecular weight cutoff of 10 kDa and stored at 4 °C before LC/MS analysis. A 3-µL aliquot of each sample was injected onto an

145 Agilent QQQ 6410 LC/MS spectrometer equipped with an Agilent HILIC analytical column. The isocratic mobile phase contained 80/20 (v/v) acetonitrile/20 mM aqueous ammonium acetate. The flow rate was 0.1 mL/min. The negative-ion detection mode was used with single ion monitoring (SIM) at m/z

1 12 16 – 2 12 16 – 1 13 16 – 2 12 16 18 – 1 13 16 18 – values of 45 ( H C O2 ), 46 ( H C O2 or H C O2 ), 48 ( H C O O or H C O O ), and 50

2 12 18 – 1 13 18 – ( H C O2 or H C O2 ). The sample preparation and analysis for formate by derivatization to formyl-2NPH were the same as described in previous work (16).

Analysis for Hexane by Gas Chromatography/Mass Spectrometry (GC/MS). A solution containing

0.1 mM ADO, 4 mM n-heptanal, 0.3 mM MeOPMS, 6 mM NADH, 50 mM HEPES buffer (pH 7.5), and

10 % glycerol in a final volume of 700 μL was allowed to react in a sealed vessel for 5 min. Gas from the head space of the sample was analyzed on a Shimadzu GCMS-QP2010S interfaced with a

Shimadzu-5MS 30 m column (ID: 0.25 narrow bore; film: 0.25 µm). The inlet and oven temperature were set to 250 °C and 40 °C, respectively. Upon injection, the oven temperature was held at 40 °C for 3 min and then ramped up to 120 °C at 10 °C /min. Total ion chromatograms were generated by scanning the range m/z = 30-3,000.

18 18 O2 and H2 O Isotope-Tracer Experiments. Standard reaction solutions (see above) were prepared in an MBraun anoxic chamber from solutions rendered O2-free on a vacuum-argon line. Samples

16 18 were exposed to either O2 or O2 (~720 torr) as described previously (34) and incubated for sufficient time for the reaction to reach completion. The resultant solution was either filtered anaerobically (for direct formate analysis) or derivatized according to the standard procedure (see above) before storage or

18 18 analysis. In the H2 O isotope tracer experiments, the reaction solution was prepared with O2-free H2 O.

The final 18O content of the reaction solution was ~85%.

Determination of the Rate of Exchange of the Aldehyde Carbonyl of Octanal with 18O from

13 18 Solvent. Reaction mixtures containing 15 mM n-1-[ C]-octanal in 0.75 mL H2 O with 0.06 mL d6-dimethylsulfoxide (d6-DMSO) added as internal standard were prepared and immediately subjected to

13C NMR analysis on a Bruker 850 MHz spectrometer at 298 K. A time-dependent shift of the carbonyl

13C signal from δ = 211.053 to δ = 211.005 (upfield shift by 10.2 Hz) signifies exchange of 16O with 18O

146 (35, 36).

Determination of the NADH:Formate Reaction Stoichiometry. Reaction solutions contained 0.1 mM ADO, 16 mM n-1-[2H]-heptanal, 0.3 mM MeOPMS, and varying [NADH]. They were allowed to react in air for 10 min at room temperature. They were then subject to the 2NPH derivatization reaction and LC/MS analysis.

RESULTS

Development of a Direct LC/MS Assay for Formate. As noted above, the most rigorously dispositive difference between the reaction pathway reported by Marsh and co-workers (20, 21) and that previously demonstrated by us (19) is the origin (H2O or O2, respectively) of the new O-atom incorporated into the formate product (Scheme 1). To assess this origin, a direct LC/MS assay for formate, rather than its 2NPH derivative, was sought. The conditions of the LC/MS analysis are provided in the

Methods section, and Figure S1 of the Supporting Information shows that the peak area in the single-ion

13 – chromatogram at m/z = 46, corresponding to the H CO2 analyte, increases with the concentration of

13 – 13 H CO2 injected. Thus, the formate product from a n-1-[ C]-aldehyde substrate can readily be detected at tens of µM concentrations by this assay. Comparison of peak areas obtained by analysis of ADO reactions to the standard curve generated from Figure S1 provides for quantification of the formate product, but, in the many experiments performed as part of this study, we found that the experimental uncertainty associated with quantification of the product by the direct method is somewhat larger than that associated with the 2NPH-derivatization method. We therefore carried out both procedures in this study: the direct analysis for purposes of detection and isotopic analysis of the formate, and the

2NPH-derivatization procedure for its accurate quantification (when required).

Demonstration of the O2 Requirement for Formate Production by the Np and Pm ADOs.

Detecting both the 2NPH derivative of formate and its alkane co-product, we previously demonstrated that the activity of the Np ADO supported by the N/F/FR reducing system requires the presence of O2.

When the first paper by Marsh and co-workers reporting O2-independent ADO activity (of the Pm

147 ortholog supported by the N/PMS reducing system) was published, we formulated three most likely explanations for the apparent discrepancy between our observations and theirs: (1) different catalytic capabilities of the different ADO orthologs; (2) different catalytic capabilities supported by the different reducing systems; and (3) O2 contamination in their assays. With the direct LC/MS assay in hand, we simultaneously evaluated possibilities 1 and 2 by testing for formate production from

13 n-1-[ C]-octadecanal, in the presence and absence of O2, by both orthologs in the presence of Marsh’s

N/PMS reducing system (Figure 1). Following a 1-h incubation in O2-depleted solution, neither reaction

13 – gave a m/z = 46 peak for H CO2 (gray traces in panels A and B) that was significantly enhanced relative to the corresponding peak from an identical sample quenched at the shortest possible reaction time (black traces). By contrast, the same 1-h incubation in the presence of ambient O2 resulted in an intense m/z = 46

13 – peak (red traces) at the same retention time as for the authentic H CO2 standard (dashed blue trace in panel A). These results establish that our preparations of both ADO orthologs require O2 to produce formate, weighing against explanations 1 and 2. As they were both carried out in the presence of natural

18 abundance O2 (0.2 % O), it is not surprising that neither reaction gave a significant peak at m/z = 48 corresponding to 1H13C18O16O– (rear traces in both panels).

18 Assessment of the Origin of the New O-atom in Formate by Use of O2. Our previous work employing the 2NPH-derivatization assay established that the O-atom incorporated into the formate product upon aldehyde cleavage by the Np ADO supported by the N/F/FR reducing system arises from O2 in the majority of reaction events. To assess the origin of the formate O-atom with other combinations of

ADO ortholog and reducing system, and to verify more quantitatively its origin in the Np ADO–N/F/FR

18 reaction, we applied the direct LC-MS formate analysis to reactions carried out under O2 (Figure 2). The

O2-depleted reactions of both Np ADO (panels A and C) and Pm ADO (panels B and D) gave negligible

1 13 16 – peaks at m/z = 46 for H C O2 (black traces), regardless of whether the N/F/FR reducing system (A and

B) or the N/PMS reducing system (C and D) was employed. This observation is consistent with the results

18 of Figure 1. For all four combinations, the reactions to which O2 was subsequently added after removal

1 13 18 16 – of atmospheric O2 (red traces) gave peaks at m/z = 48 for H C O O2 that were much more intense than

148 1 13 16 – the corresponding peaks at m/z = 46 for H C O2 . These results demonstrate that, for all four combinations of ADO ortholog and reducing system, O2 rather than H2O is the source of the O-atom incorporated into the formate product in the vast majority of reaction events. The minor peaks at m/z = 46

16 in panels C and D reflect a minor atmospheric O2 contaminant in these reactions rather than a trace level of hydrolytic formate production, as established below.

Testing Short-chain Aldehydes as ADO Substrates. The small peaks at m/z = 46 in the red traces of Figure 2, panels C and D, potentially suggest the production of a small but detectable quantity of

16 formate by a truly hydrolytic pathway, which would result in incorporation of an O-atom from H2 O

18 rather than O2. Alternatively, these peaks could reflect contamination of these reactions by

16 environmental O2 (equivalent to explanation 3 for the discrepancy between our results and those of

Marsh and co-workers). To distinguish between these two possible interpretations and provide for the most sensitive detection of even trace hydrolytic formate production, we sought to carry out reactions in

18 1 13 18 – H2 O and analyze for the presence of any H C O2 , which would result from exchange of the aldehyde carbonyl O-atom of the substrate with solvent (27) prior to C1–C2-bond cleavage and incorporation of the second O-atom from solvent in the ADO-mediated hydrolytic event (Scheme 1B). Proper interpretation of the results of such an experiment requires knowledge of the extent of the carbonyl-solvent exchange occurring prior to turnover. We therefore sought to employ a substrate that would be amenable to the direct determination of the carbonyl-solvent exchange kinetics by 13C NMR spectroscopy. The report by

Marsh and co-workers that linear aldehydes as short as n-heptanal are substrates for the ADOs (20, 21) inspired us to test various saturated linear aldehydes (n-heptanal, n-octanal, and n-decanal), and, indeed, all were found to be active ADO substrates (Table S1).

To verify that the catalytic activity of the ADO operating on even the shortest substrate, n-heptanal, has the same general characteristics as ADO-mediated cleavage of the more physiologically relevant n-octadecanal, we carried out reactions with the Np ADO and the chemical reducing system [in this case, 1-methoxy-N-methylphenazine methosulfate (MeOPMS) was used in place of the PMS used by

Marsh and co-workers, because the former compound reportedly exhibits greater photolytic stability (37),

149 a property desirable for mechanistic analysis by stopped-flow absorption experiments] at two different

ADO concentrations and also with serial omission of a single reaction component to define the requirements for activity (Figure 3). The complete enzyme reactions rapidly generated formate (black and gray symbols and fit lines), and, although only a very short (if any) steady state was observed (the reason for this characteristic is not known), the initial rates of formate production showed the expected correlation to ADO concentration (compare black to gray trace). Fits of the equation for an exponential

“burst” and linear rise phases to the progress curves gave initial rates (v/[ADO]) of 0.27 ± 0.03 s-1, much greater than any previously reported ADO turnover number (20). This result hints at the value of the short substrates in mechanistic analysis. Upon omission of the ADO enzyme (purple), NADH (green),

MeOPMS (red), or O2 (blue), minimal formate was produced. Importantly, after a 4 min unproductive incubation in the absence of O2, opening of the sealed reaction vessel to the air (blue arrow) led to commencement of formate production, further confirming the requirement for O2. The data in Figure 3 suggest that turnover of the shorter, more soluble substrate, n-heptanal, has the same requirements as turnover of the physiologically relevant longer fatty aldehydes.

Quantification of n-Hexane Produced by the ADO-Catalyzed Conversion of n-Heptanal. As additional evidence for the mechanistic correspondence between ADO-mediated cleavage of n-heptanal and turnover of the longer fatty aldehydes, the stoichiometry of n-hexane to formate was determined.

GC/MS analysis of the head space from triplicate, sealed ADO reactions (described in the Methods section) permitted both the detection of the n-hexane co-product, and, by integration of the peak of the total-ion chromatogram and comparison to an external standard curve, its quantification. The triplicate reactions produced 322 ± 35 µM formate and 362 ± 40 µM hexane (mean ± standard deviation), verifying the expected 1:1 reaction stoichiometry of the two co-products.

18 Measurement of the Exchange Rate of the Aldehyde Carbonyl of Octanal with H2 O. The published interpretation of the solvent-18O-tracer experiment carried out by Marsh and co-workers in an effort to verify the expected solvent origin of the O-atom incorporated into the formate product relied on their report that the substrate carbonyl oxygen exchanges with solvent very slowly on the timescale of the

150 ADO reaction. Indeed, the assumption of rapid exchange would have led them to exactly the opposite conclusion: that the origin of the incorporated O-atom must have been contaminating O2. With the knowledge that short-chain aldehydes are viable ADO substrates, we used 13C-NMR spectroscopy to determine the rate of carbonyl-solvent exchange directly (Figure 4). Incubation of the specifically labeled

18 13 substrate in H2 O resulted in a time-dependent change in the chemical shift of the 1- C nucleus from

211.053 ppm to 211.005 ppm, reflecting exchange of the 16O in the carbonyl group for 18O (35, 36). This shift was nearly complete after 6 min (top spectrum) and had a half-life of approximately 1 min (second spectrum from the bottom). These spectra demonstrate that pre-incubation of a short-chain n-aldehyde

18 substrate in H2 O for ≥ 6 min permits the substrate to attain the O-isotopic composition of the solvent.

18 Test for Trace Hydrolytic Cleavage of n-Heptanal by Analysis of Reactions in H2 O. With the kinetics of carbonyl-solvent exchange elucidated, the definitive test for hydrolytic formate production forecast above could now be properly executed and interpreted (Figure 5). The n-1-[2H]-heptanal

2 18 substrate was pre-incubated in either natural O-abundance H2O (panel A) or 85% O H2O (panel B) for several minutes to ensure that the carbonyl O-atom would achieve the isotopic composition of the solvent, and the reaction solution was then rendered complete by the addition of the Np ADO and the N/MeOPMS chemical reducing system. In either solvent, negligible formate was detected in reactions from which O2

16 16 had been removed and not added back (black traces in both panels). In H2 O solvent under O2 gas

2 12 16 – (panel A, red traces), an intense peak at m/z = 46 was detected ( H C O2 ), and no significant peak at

2 12 18 16 – 2 12 18 – 16 either m/z = 48 ( H C O O ) or m/z = 50 ( H C O2 ) was seen, as expected. In H2 O solvent under

18 2 12 18 16 – O2 gas (panel A, blue traces), the major peak shifted to m/z = 48 ( H C O O ), reflecting incorporation

18 of a single O-atom from O2 into the formate product. Interestingly, a small but significant peak at m/z =

50 reflects the production of a minor quantity of formate with both O-atoms originating from O2

2 12 18 – ( H C O2 ). This product could result from (1) an abortive event that results in incorporation of one

O-atom from O2 into the aldehyde substrate without achieving C1–C2 fragmentation and (2) subsequent, successful aldehyde cleavage occurring either without substrate release to solvent (upon which the 18O incorporated during the abortive event could exchange) or before carbonyl-solvent exchange can occur.

151 18 16 This rare double-incorporation event may hold useful mechanistic clues. In H2 O solvent under O2 gas

(panel B, red traces), the intensities of the major peak at m/z = 48 and the minor peak at m/z = 46,

2 12 16 18 – 2 12 16 – corresponding to H C O O and H C O2 , respectively, precisely mirror the solvent isotopic composition of 85% 18O and 15% 16O. Most importantly, no significant peak at m/z = 50, corresponding to

2 12 18 – the H C O2 that should have been produced in 72% of hydrolytic events (the fraction expected to result from 0.85 18O from exchange and 0.85 18O from hydrolysis), could be detected (red arrow in panel

18 18 B). Only when the reaction was carried out in both H2 O and O2 did the m/z = 50 peak become prevalent (indeed, predominant), and, under these conditions, the ratio of the m/z = 50 and m/z = 48 peaks again reflected the solvent 18O:16O composition, implying a vastly predominant outcome of precisely one

O-atom of formate from solvent (by exchange) and precisely one from O2 (by oxygenative cleavage). In other words, our preparations simply do not support detectable O2-independent, hydrolytic formate production.

Test for a Stoichiometric or Catalytic Role of the Reducing System. The final piece of evidence cited by Marsh and co-workers in support of O2-independent, hydrolytic aldehyde cleavage by ADO was their observation that NADH is not consumed during the reaction (21). The absence of a stoichiometric requirement for electrons (4 e-/turnover) is incompatible with the oxygenative pathway (Scheme 1A) with which our published data and the extensive evidence presented above are uniquely consistent. To assess the Marsh report of a catalytic rather than stoichiometric role for the reducing system, reactions were run to completion (verified by examination of different reaction times) with limiting and varying [NADH]

(Figure 6). The yield of formate was found to increase linearly with increasing [NADH]. The slope of the line gives 0.25 ± 0.03 formate/NADH. Given that NADH is a two-electron donor, this experimental stoichiometry corresponds to 8 e-/formate, twice the theoretical ratio of 4 e-/formate predicted by Scheme

1A. This deviation most likely reflects the “uncoupling” of NADH oxidation from formate production in roughly half of reaction events, a result that is not surprising in view of the non-physiological nature of the reducing system and its inherent reactivity to O2 even in the absence of ADO.

152 DISCUSSION

The strict, stoichiometric O2 and NADH requirements of both the Np and Pm ADO orthologs in the presence of either the N/F/FR protein-based reducing system or the N/(MeO)PMS chemical reducing system, together with the O-isotope labeling pattern of the formate product, firmly establish that our ADO preparations simply do not support hydrolytic aldehyde cleavage. In particular, the absence of any

2 12 18 – 18 detectable H C O2 product in reactions run in H2 O is stark evidence that one oxygen atom of the formate invariably derives from O2. Whereas it would remain formally possible that Marsh and co-workers had obtained a different enzyme form (e.g., possessing a different metallocofactor or some post-translational modification) having distinct catalytic capabilities, it appears that, while our manuscript was being written, these authors also came to recognize that their report of O2-independent ADO activity is, in fact, incorrect (N. Marsh, personal communication). Thus, the nature of the reaction is now firmly established as oxygenative aldehyde cleavage to formate and alk(a/e)ne, providing argument that the designation “aldehyde-deformylating oxygenase” (ADO) is more appropriate than “aldehyde decarbonylase” (AD).

Despite the error of its main conclusion, the work of Marsh and co-workers nevertheless provided valuable technical tools from which we have profited in this study (20, 21). For example, we concur that the N/(MeO)PMS chemical reducing system supports greater activity than the N/F/FR protein-based system originally reported by Schirmer, et al. (1) and adopted by us in our first two studies (16, 19).

Indeed, the coupling efficiency supported by this reducing system (~ 50%) is surprisingly high and much greater than the value that we previously measured for the heterologous protein-based system. As importantly, their discovery that ADO accepts much shorter, more soluble aldehydes is an extremely useful advance. Our past experience in dissecting the mechanisms of O2-utilizing metalloenzymes by direct detection and characterization of intermediate states has taught us that the most generally informative approach involves rapid mixing of the complex of the reduced enzyme form and its substrate(s) with O2 to initiate rapid formation of the intermediate complexes. In these experiments, it is desirable to saturate the enzyme by its substrate. For the case of ADO, the physiologically relevant

153 long-chain n-aldehyde substrates are only sparingly soluble, and we have experienced difficulty in achieving saturation at accessible substrate concentrations. In addition, steady-state turnover rates thus far achieved with these long substrates are far too modest (< 1 min-1) to have inspired hope that reactive intermediates might be induced to accumulate. Shorter aldehydes such as n-heptanal are more soluble and, indeed, we find that n-heptanal supports an initial rate of ~ 0.3 s-1. And, although 0.3 s-1 is still a relatively modest rate constant compared to those exhibited by many oxygenases (25, 28-33), it approaches the realm of 100-102 s-1 that, in our experience, signifies that the selected reaction conditions may support intermediate accumulation (23).

SUPPORTING INFORMATION AVAILABLE

Figures depicting LC/MS analysis of the 13C-labeled formate standard, 1H- and 13C-NMR spectra of n-1-[13C]-octadecanal, n-1-[13C]-octanal, and n-1-[2H]-heptanal, and a table comparing the

ADO-catalyzed production of formate from n-heptanal, n-octanal, and n-decanal. This material is available free of charge at the journal website http://pubs.acs.org.

CONFLICT OF INTEREST

The authors declare a financial interest in LS9, Incorporated, which seeks to use the ADOs for production of diesel fuel in bacteria.

154 REFERENCES

1. Schirmer, A., Rude, M. A., Li, X., Popova, E., and del Cardayre, S. B. (2010) Microbial

biosynthesis of alkanes, Science 329, 559-562.

2. Ducat, D. C., Way, J. C., and Silver, P. A. (2011) Engineering cyanobacteria to generate

high-value products, Trends Biotechnol. 29, 95-103.

3. Schirmer, A., Rude, M., and Brubaker, S. (2009) Methods and compositions far producing

hydrocarbons, p patent application WO/2009/140696, LS9.

4. Service, R. F. (2009) ExxonMobil fuels Venter’s efforts to run vehicles on algae-based oil,

Science 325, 379.

5. Reppas, N. B., and Ridley, C. P. (2010) Methods and compositions for the recombinant

biosynthesis of n-alkanes, p patent 7794969, Joule Unlimited Inc., US.

6. Keasling, J. D. (2010) Manufacturing Molecules Through Metabolic Engineering, Science 330,

1355-1358.

7. Radakovits, R., Jinkerson, R. E., Darzins, A., and Posewitz, M. C. (2010) Genetic Engineering of

Algae for Enhanced Biofuel Production, Eukaryotic Cell 9, 486-501.

8. Du, J., Shao, Z. Y., and Zhao, H. M. (2011) Engineering microbial factories for synthesis of

value-added products, J. Ind. Microbiol. Biotechnol. 38, 873-890.

9. Hess, W. R. (2011) Cyanobacterial genomics for ecology and biotechnology, Curr. Opin.

Microbiol. 14, 608-614.

10. Malcata, F. X. (2011) Microalgae and biofuels: A promising partnership?, Trends Biotechnol. 29,

542-549.

11. Patra, T., Manna, S., and Maiti, D. (2011) Metal-Mediated Deformylation Reactions: Synthetic

and Biological Avenues, Angew. Chem.-Int. Edit. 50, 12140-12142.

12. Robertson, D. E., Jacobson, S. A., Morgan, F., Berry, D., Church, G. M., and Afeyan, N. B. (2011)

A new dawn for industrial photosynthesis, Photosynth. Res. 107, 269-277.

13. Zhang, F. Z., Rodriguez, S., and Keasling, J. D. (2011) Metabolic engineering of microbial

155 pathways for advanced biofuels production, Curr. Opin. Biotechnol. 22, 775-783.

14. Hallenbeck, P. (2012) Microbial Technologies in Advanced Biofuels Production, Springer, New

York.

15. Serrano-Ruiz, J. C., Ramos-Fernandez, E. V., and Sepulveda-Escribano, A. (2012) From biodiesel

and bioethanol to liquid hydrocarbon fuels: new hydrotreating and advanced microbial

technologies, Energy Environ. Sci. 5, 5638-5652.

16. Warui, D. M., Li, N., Nørgaard, H., Krebs, C., Bollinger, J. M., Jr., and Booker, S. J. (2011)

Detection of formate, rather than carbon monoxide, as the stoichiometric co-product in

conversion of fatty aldehydes to alkanes by a cyanobacterial aldehyde decarbonylase, J. Am.

Chem. Soc. 133, 3316-3319.

17. Merkx, M., Kopp, D. A., Sazinsky, M. H., Blazyk, J. L., Müller, J., and Lippard, S. J. (2001)

Dioxygen activation and methane hydroxylation by soluble methane monooxygenase: a tale of

two irons and three proteins, Angew. Chem. Int. Ed. 40, 2782-2807.

18. Fox, B. G., Lyle, K. S., and Rogge, C. E. (2004) Reactions of the diiron enzyme stearoyl-acyl

carrier protein desaturase, Acc. Chem. Res. 37, 421-429.

19. Li, N., Nørgaard, H., Warui, D. M., Booker, S. J., Krebs, C., and Bollinger, J. M., Jr. (2011)

Conversion of Fatty Aldehydes to Alka(e)nes and Formate by a Cyanobacterial Aldehyde

Decarbonylase: Cryptic Redox by an Unusual Dimetal Oxygenase, J. Am. Chem. Soc. 133,

6158-6161.

20. Das, D., Eser, B. E., Han, J., Sciore, A., and Marsh, E. N. G. (2011) Oxygen-Independent

Decarbonylation of Aldehydes by Cyanobacterial Aldehyde Decarbonylase: A New Reaction of

Diiron Enzymes, Angew. Chem.-Int. Edit. 50, 7148-7152.

21. Eser, B. E., Das, D., Han, J., Jones, P. R., and Marsh, E. N. G. (2011) Oxygen-Independent

Alkane Formation by Non-Heme Iron-Dependent Cyanobacterial Aldehyde Decarbonylase:

Investigation of Kinetics and Requirement for an External Electron Donor, Biochemistry 50,

10743-10750.

156 22. Krebs, C., Price, J. C., Baldwin, J., Saleh, L., Green, M. T., and Bollinger, J. M., Jr. (2005) Rapid

freeze-quench 57Fe Mössbauer spectroscopy: Monitoring changes of an iron-containing active site

during a biochemical reaction, Inorg. Chem. 44, 742-757.

23. Bollinger, J. M., Jr., and Krebs, C. (2006) Stalking intermediates in oxygen activation by iron

enzymes: Motivation and method, J. Inorg. Biochem. 100, 586-605.

24. Krebs, C., Galonić Fujimori, D., Walsh, C. T., and Bollinger, J. M., Jr. (2007) Non-heme

Fe(IV)-oxo intermediates, Acc. Chem. Res. 40, 484-492.

25. Jiang, W., Hoffart, L. M., Krebs, C., and Bollinger, J. M., Jr. (2007) A manganese(IV)/iron(IV)

intermediate in assembly of the manganese(IV)/iron(III) cofactor of Chlamydia trachomatis

ribonucleotide reductase, Biochemistry 46, 8709-8716.

26. Li, N., Korboukh, V. K., Krebs, C., and Bollinger, J. M., Jr. (2010) Four-electron oxidation of

p-hydroxylaminobenzoate to p-nitrobenzoate by a peroxodiferric complex in AurF from

Streptomyces thioluteus, Proc. Natl. Acad. Sci. USA 107, 15722–15727.

27. Carey, F. A., and Sundberg, R. J. (2007) Advanced Organic Chemistry: Part A Structure and

Mechanisms, 5th ed., Springer, New York.

28. Baldwin, J., Krebs, C., Ley, B. A., Edmondson, D. E., Huynh, B. H., and Bollinger, J. M., Jr.

(2000) Mechanism of rapid electron transfer during oxygen activation in the R2 subunit of

Escherichia coli ribonucleotide reductase. 1. Evidence for a transient tryptophan radical, J. Am.

Chem. Soc. 122, 12195-12206.

29. Price, J. C., Barr, E. W., Tirupati, B., Bollinger, J. M., Jr., and Krebs, C. (2003) The first direct

characterization of a high-valent iron intermediate in the reaction of an -ketoglutarate-dependent

dioxygenase: a high-spin Fe(IV) complex in taurine/-ketoglutarate dioxygenase (TauD) from

Escherichia coli, Biochemistry 42, 7497-7508.

30. Hoffart, L. M., Barr, E. W., Guyer, R. B., Bollinger, J. M., Jr., and Krebs, C. (2006) A unified

mechanism for the -ketoglutarate-dependent dioxygenases: Evidence for a C-H-cleaving Fe(IV)

157 complex in prolyl-4-hydroxylase, Proc. Natl. Acad. Sci., U. S. A. 103, 14738-14743.

31. Xing, G., Diao, Y. H., Hoffart, L. M., Barr, E. W., Prabhu, K. S., Arner, R. J., Reddy, C. C., Krebs,

C., and Bollinger, J. M., Jr. (2006) Evidence for C-H cleavage by an iron-superoxide complex in

the glycol cleavage reaction catalyzed by myo-inositol oxygenase, Proc. Natl. Acad. Sci. USA 103,

6130-6135.

32. Galonić, D. P., Barr, E. W., Walsh, C. T., Bollinger, J. M., Jr., and Krebs, C. (2007) Two

interconverting Fe(IV) intermediates in aliphatic chlorination by the halogenase CytC3, Nature

Chem. Biol. 3, 113-116.

33. Korboukh, V. K., Li, N., Barr, E. W., Bollinger, J. M., Jr., and Krebs, C. (2009) A long-lived,

substrate-hydroxylating peroxodiiron(III/III) intermediate in the amine oxygenase, AurF, from

Streptomyces thioluteus, J. Am. Chem. Soc. 131, 13608-13609.

34. Matthews, M. L., Neumann, C. S., Miles, L. A., Grove, T. L., Booker, S. J., Krebs, C., Walsh, C.

T., and Bollinger, J. M., Jr. (2009) Substrate positioning controls the partition between

halogenation and hydroxylation in the aliphatic halogenase, SyrB2, Proc. Natl. Acad. Sci. U. S. A.

106, 17723-17728.

35. Jameson, C. J. (1977) Isotope shift in NMR, J. Chem. Phys. 66, 4983-4988.

36. Risley, J. M., and Van Etten, R. L. (1979) O-18 Isotope shift upon C-13 NMR spectra and its

application to the study of oxygen-exchange kinetics, J. Am. Chem. Soc. 101, 252-253.

37. Hisada, R., and Yagi, T. (1977) 1-Methoxy-5-methylphenazinium methyl sulfate. A

photochemically stable electron mediator between NADH and various electron acceptors, J.

Biochem. 82, 1469-1473.

158 SCHEME AND FIGURE LEGENDS

Scheme 1. Predicted origins of the oxygen atoms in the formate co-product generated in the oxygenative

(A) and hydrolytic (B) reactions purportedly catalyzed by cyanobacterial “ADs”. Non-enzymatic exchange of the carbonyl O-atom of the aldehyde substrate with solvent is depicted on the left.

Figure 1. LC/MS detection of formate produced enzymatically by Pm ADO and Np ADO. Three identical standard reaction solutions with N/PMS reducing system and n-1-[13C]-octadecanal substrate were prepared anareobically for both Np ADO (A) and Pm ADO (B) respectively. They were filtered and measured by LC/MS right away (black line), or after a 1-h incubation at 22 °C in the absence (grey line)

13 - or presence of O2 (red line). An authentic H COO solution was analzyed similarly as a standard (blue dotted line). The traces are the single-ion-monitoring (SIM) LC/MS chromatograms at the indicated values of m/z.

18 Figure 2. LC/MS detection of formate produced in O2-tracer experiments by Np ADO and Pm ADO.

Standard reaction solutions with the N/F/FR (A and B) or N/PMS (C and D) reducing system and n-1-[13C]-octadecanal substrate were prepared anaerobically for both Np ADO (A and C) and Pm ADO (B and D). Solutions were incubated at 22 °C for 1 h either anaerobically (black line) or after introduction of

18 O2 (red line), filtered, and assayed directly for formate by LC/MS. The traces are the single-ion-monitoring (SIM) LC/MS chromatograms at the indicated values of m/z.

Figure 3. Time-dependence of, and requirements for, the Np ADO-catalyzed production of formate.

Standard reaction solutions with n-1-[2H]-heptanal substrate were prepared as described in Materials and

Methods, and, as indicated, the [Np ADO] was varied or a single reaction component was omitted.

Aliquots were removed after varied times of incubation at 22 °C, derivatized, and analyzed for formyl-2NPH by LC/MS.

Figure 4. Determination of the kinetics of exchange of the carbonyl oxygen of n-[13C]-octanal with solvent by 13C-NMR-spectroscopy. n-1-[13C]-octanal (natural abundance of oxygen, >99% 16O) was

18 dissolved in 700 μL of H2 O with 60 μL of d6-DMSO as the internal standard. Spectra were required at

159 varying time of incubation at ambient temperature (~ 22 °C). The exchange of 16O from the C1 position of the aldehyde with solvent 18O results in the indicated change in the chemical shift.

18 18 Figure 5. LC/MS O2- and H2 O-isotope-tracer experiments to determine the origin of the O-atom incorporated into formate by Np ADO and Pm ADO. Standard reaction solutions (see Materials and

Methods) of 500 µL total volume with n-1-[2H]-heptanal substrate and the N/MeOPMS reducing system

16 18 were prepared anaerobically with H2 O (A) or H2 O (B). All reactions were incubated for 5 min at 22 °C,

16 one set in the absence of O2 (black line), the second under ~ 1 atm of natural-abundance (>99% O) O2

18 (red line), and the third under ~ 1 atm of O2 (> 99% isotopic enrichment) (blue line). The reactions were filtered and subjected to direct LC/MS analysis for formate. The traces are the single-ion-monitoring

(SIM) LC/MS chromatograms at the indicated values of m/z.

Figure 6. NADH:formate stoichiometry of the Np ADO-catalyzed production of formate from n-heptanal.

Reactions were carried out at 22 °C in 100 mM HEPES buffer (pH 7.5) for 5 min. They contained 0.3 mM MeOPMS, 0.2% triton x-100, 4 mM n-1-[2H]-heptanal, the indicated concentration of Np ADO

[0.033 mM (red circles), 0.1 mM (blue circles), or 0.2 mM (green circles)], and varying concentration of

NADH (values shown on abscissa). The solutions were subjected to the derivatization procedure and analyzed for formyl-2NPH by LC/MS.

160

Li, et al. Scheme 1 (column width)

161

Li, et al. Figure 1 (page width)

162

Li, et al. Figure 2 (page width)

163

Li, et al. Figure 3 (column width)

164

Li, et al. Figure 4 (column width)

165

Li, et al. Figure 5 (page width)

166

Li, et al. Figure 6 (column width)

167 For table of contents use only

Title: Evidence for Only Oxygenative Cleavage of Aldehydes to Alk(a/e)nes and Formate by

Cyanobacterial “Aldehyde Decarbonylase”

Authors: Ning Li, Wei-chen Chang, Douglas M. Warui, Squire J. Booker, Carsten Krebs, J. Martin

Bollinger, Jr.

168 SUPPORTING MATERIAL

Evidence for Only Oxygenative Cleavage of

Aldehydes to Alk(a/e)nes and Formate by

Cyanobacterial “Aldehyde Decarbonylase”†

Ning Li,a Wei-chen Chang,b Douglas M. Warui,b Squire J. Booker,a,b,* Carsten Krebs,a,b,* J. Martin

Bollinger, Jr.a,b,*

Departments of aBiochemistry and Molecular Biology and of bChemistry, The Pennsylvania State

University, University Park, Pennsylvania 16802

169 Table S1. Yields of formate from C7, C8 and C10 saturated n-aldehyde substrates in Np ADO reactions. .

Substrate 1-[2H]-heptanal 1-[2H]-octanal 1-[2H]-decanal

[2H]-formate yield 0.126 0.119 0.108

(mM)

* Reaction condition: 16 mM substrate, 0.02 mM Np ADO, 0.3 mM MeOPMS, 12 mM NADH, 2 min at 37 °C, in 100 mM HEPES buffer, 10% glycerol, pH 7.5.

170

Figure S1. Single-ion-monitoring (SIM) chromatograms showing the proportionality between the peak intensity (area) and the concentration of the formate analyte [m/z = 46: 13C-formate].

171

Figure S2. Plot of peak intensity (area) versus the concentration of the formate analyte [m/z = 46:

13C-formate] in the chromatograms from Figure S1.

172

Figure S3. 1H-NMR spectrum of n-1-[13C]-octadecanal substrate.

173

Figure S4. 13C-NMR spectrum of n-1-[13C]-octadecanal substrate.

174

Figure S5. 1H-NMR spectrum of n-1-[13C]-octanal substrate.

175

Figure S6. 13C-NMR spectrum of n-1-[13C]-octanal substrate.

176

Figure S7. 360 MHz 1H-NMR spectrum of n-1-[2H]-heptanal substrate.

177

Appendix E

Summary of unpublished data

178

III/III Figure 1. Kinetic traces showing the reaction of peroxo-Fe2 -AurF with II/II 4-mercaptobenzoate. An O2-free reaction solution containing 400 μM Fe2 -AurF was mixed with a buffer solution containing 1.8 mM O2 for 0.1 s, allowing the peroxo intermediate to accumulate, before mixing with an O2-free buffer solution containing various concentration of 4-mercaptobenoate at 5 °C. Absorption change at 500 nm was monitored.

179

III/III Figure 2. Kinetic traces showing the reaction of peroxo-Fe2 -AurF with II/II 4-mercaptobenzoate. An O2-free reaction solution containing 400 μM Fe2 -AurF was mixed with a buffer solution containing 1.8 mM O2 for 0.1 s, allowing the peroxo intermediate to accumulate, before mixing with an O2-free buffer solution containing various concentration of 4-methylmercaptobenoate at 5 °C. Absorption change at 500 nm was monitored.

180

Figure 3. Stopped flow experiment demonstrating the binding of 4-hydroxybenzoate III/III III/III to Fe2 -AurF. 50 μM Fe2 -AurF was mixed with 200 μM 4-hydroxybenzoate at 19 °C and spectra were collected by photodiode array spectrophotometer.

181

AurF V97W/S I98W/Y (ToMOH I100W) AurF A40W N43W (E.coli R2 W48) AurF E101D (E.coli D84E) AurF H223A/I/N/Q/K AurF T100S/W (ToMOH T201S) AurF A195W/S Ct R2 I223H (AurF H223) E.coli R2 I234H (AurF H223)

Table 1. A list of the AurF or AurF-related variants constructed.

182

Figure 4. EPR Spectra showing the redox reaction of rv0233 ortholog of Saccharopolyspora erythraea (SRV). As isolated SRV was either analyzed by EPR directly (green), or treated with dithionite in the presence of methylviologen (red), or reduced and re-oxidized before analysis (blue).

183

Figure 5. Stopped flow experiment demonstrating the substrate triggered oxygen II/II activation catalyzed by Np AD. 0.2 mM Fe2 -Np AD, in the absence (left panel) or presence (right panel) of 1 mM decanal was reacted with a buffer solution containing 1.8 mM O2 at 5 °C and the spectra were collected by photodiode array spectrophotometer.

184

Figure 6. Kinetic traces showing the substrate dependence of oxygen activation II/II catalyzed by Np AD. Left panel: 0.2 mM Fe2 -Np AD, in the absence (black line) or presence (red line) of 4 mM decanal was reacted with a buffer solution containing 1.8 mM O2 at 5 °C and the absorption change at 450 nm was monitored. Right panel: 0.2 mM II/II Fe2 -Np AD, in the absence of any substrate or analog (black line) or in the presence of either 4 mM nonane (blue line), 1-nonanol (green line), octanoic acid (purple line), or nonanal (red line) was reacted with a buffer solution containing 1.8 mM O2 at 5 °C and the absorption change at 450 nm was monitored.

185

Figure 7. Kinetic traces showing the substrate concentration dependence of oxygen II/II activation catalyzed by Np AD. 0.2 mM Fe2 -Np AD, in the absence (red line) or presence of decanal of 0.5 mM (orange line), 1 mM (dark green line), 2 mM (light green line), or 4 mM (blue line) was reacted with a buffer solution containing 1.8 mM O2 at 5 °C and the absorption change at 450 nm was monitored.

186

II/II Figure 8. 4.2-K/53-mT Mössbauer spectra of samples in which Fe2 -Np AD was II/II reacted with O2. A solution of Fe2 -Np AD (1.2 mM Fe2) with 6 mM decanal was either hand frozen right away (A) or reacted with 1 equivalent volume of buffer solution containing 1.8 mM O2 (O2/Fe2 = 1.5) for 30 s before being freeze-quenched (B). A III/III reference spectrum of the peroxo-Fe2 -Np AD intermediate (C) with two simulated quadruple doublets (δ1 = 0.54 mm/s, ΔEQ1 = 1.22 mm/s (red); δ2 = 0.48 mm/s, ΔEQ2 = 0.50 II/II III/III mm/s (blue)) was generated by subtracting Fe2 -Np AD and Fe2 -Np AD components from B.

187

III/III Figure 9. LC/MS analysis of the formate produced by reacting peroxo-Fe2 -Np II/II AD intermediate with MeOPMS/NADH. 0.9 mM Fe2 -Np AD with 8 mM 2 1-[ H]-decanal was mixed with 2-volume of buffer solution containing 1.8 mM O2 (O2/Fe2 = 4) for 30 s, derivatized right away or reacting with various concentration of MeOPMS/NADH before derivatization. A slope of 0.9 was obtained by fitting the first two data points.

188

Figure 10. Kinetic traces demonstrating the oxidation of MeOPMS by Np AD. Black III/III trace: decay of peroxo-Fe2 -Np AD in the absence of MeOPMS. Red trace: oxidation III/III of NADH-reduced MeOPMS by preformed peroxo-Fe2 -Np AD. Blue trace: oxidation II/II of NADH-reduced MeOPMS by a solution containing Fe2 -Np AD, decanal and O2.

189

Figure 11. Kinetic traces demonstrating the effect of as isolated Np AD on the oxidation of NADH by O2. The oxidation of NADH was monitored by change of absorption at 340 nm in the presence of various concentration of as isolated Np AD and ambient O2 at room temperature.

190

Figure 12. Stopped flow experiment demonstrating the reduction of cyanobacterial ferredoxin by dithionite. An O2-free buffer solution containing 0.2 mM ferredoxin was mixed with an equal volume of O2-free buffer solution containing 0.18 mM dithionite at 5 °C and the data were obtained by photodiode array spectrophotometer.

191

Figure 13. Stopped flow experiment demonstrating the reduction of as isolated Np AD by dithionite-reduced ferredoxin. An O2-free buffer solution containing 0.2 mM ferredoxin was mixed with an equal volume of O2-free buffer solution containing 0.18 mM dithionite at 5 °C for 500 s allowing complete reduction before mixing with stoichiometric as isolated Np AD. The data were obtained by photodiode array spectrophotometer.

192

Figure 14. Stopped flow experiment demonstrating the reduction of as isolated Np AD by NADH-reduced MeOPMS. An O2-free buffer solution containing 0.2 mM as isolated Np AD was mixed with an equal volume of O2-free buffer solution containing 0.1 mM NADH and 0.1 mM MeOPMS at 5 °C and the data were obtained by photodiode array spectrophotometer.

193

Figure 15. LC/MS chromatograms demonstrating the formate production from aldehyde substrates catalyzed by as isolated Np AD in the presence of H2O2. Reaction solutions containing 0.1 mM as isolated Np AD, 6% H2O2 and 16 mM aldehyde were incubated for 5 min at room temperature before being filtrated and analyzed for formate by LC/MS.

194

Figure 16. Activity of Np AD and its variants in the presence of N/F/FR or N/PMS reducing system. Reaction solutions containing 0.1 mM enzyme, 0.5 mM 1-[13C]-octadecanal, 0.2% tx-100 and either N/F/FR (blue) or N/PMS (red) reducing system were incubated at room temperature for 1 h before being derivatized and analyzed for formate. 1, wild type Np AD; 2, Np AD Y22F; 3, Np AD C71A; 4, Np AD C107A; 5, Np AD C117A; 6, Np AD Y123F; 7, Np AD M60A.

195

Figure 17. Diagram showing the constructed plasmid for recombining Synechocystis sp. PCC6803 AD (sll0208) and AAR (sll0209) genes to the pAQ1 plasmid of Synechococcus sp. PCC7002.

196

Figure 18. Coomassie-blue stained 12% SDS–PAGE analysis of whole cell proteins of Synechococcus sp. PCC7002, wild type or 6803 AD over-expressing strains. The gene encoding Synechocystis sp. PCC6803 AD (~ 26 kDa) was recombined into pAQ1 plasmid of Synechococcus sp. PCC7002. Whole cell samples of eight 6803 AD over-expressing strains (1-8) and one wild type strain (wt) were analyzed by coomassie-blue stained SDS-PAGE gel (The very left lane is protein molecular weight marker.).

197

Figure 19. GC/MS chromatogram showing the detection of heptadecane in the Synechococcus sp. PCC7002 strain that expresses 6803 AD and AAR genes. Black: perdeuterated heptadecane and heptadecane standard. Blue: wild type Synechococcus sp. PCC7002 sample. Red: 6803 AD/AAR expressing Synechococcus sp. PCC7002 strain sample.

198

Figure 20. 13C-formate yield in the presence of different ferredoxins. Reactions containing 0.1 mM Np AD, 0.5 mM 1-[13C]-octadecanal, 0.2% tx-100 and various reducing systems were incubated at room temperature for 1 h before being derivatized and analyzed for formate. Reaction 1-8 were using different Synechocystis sp. PCC6803 ferredoxins (all paired with 6803 ferredoxin reductase slr1643). Reaction 9 was using spinach ferredoxin and spinach ferredoxin reductase. NADPH was provided for ferredoxin/ferredoxin reductase reducing system. Reaction 10 was using MeOPMS/NADH reducing system.

199

Np AD 6803 cell lysate MeOPMS NADH 1-[2H]-decanal 1-[2H]formate 0.1 mM (supernatant) 0.3 mM 6 mM 8 mM μM + - + + + 273 + + + + + 442 - + + + + 2 + + - + + 104 + + - - + 81 + 3kDa filter + + + 2 filtrate + 3kDa filter + + + 221 retained + 10kDa filter + + + 2 filtrate + 10kDa filter + + + 217 retained + 30kDa filter + + + 1 filtrate + 30kDa filter + + + 244 retained

Table 2. Affect of 6803 cell extract (soluble portion) on the yield of formate product of Np AD reaction.

200

Np ADa Metal(s) used for MeOPMS NADH 1 -[2H]-decanal 1-[2H]formate 0.1 mM reconstitution (1 eq. ea.) 0.3 mM 6 mM 8 mM μM + - + + + 104 + Cu/Fe + + + 106 + Mn/Fe + + + 290 + Ni/Fe + + + 388 + Zn/Fe + + + 159 + Fe/Fe + + + 287 a This Np AD enzyme was purified from E. coli cells grown in rich LB medium and treated by reductive chelation prior to these assays.

Table 3. Affect of metals on the yield of formate product of Np AD reaction.

201 Ning Li

[email protected] 814 321 3898

Education June 2007 - summer 2012 Ph.D. degree in biochemistry, microbiology and molecular biology, The Pennsylvania State University August 2005 – May 2007 Graduate studies in the College of Medicine, University of Florida August 2001 – June 2005 Bachelor degree in Life Sciences, NanKai University, Tianjin, China

Publications In print Korneeva Korboukh, V.; Li, N.; Barr, E. W.; Bollinger, J. M., Jr.; Krebs, C. “A Long-Lived, Substrate-Hydroxylating Peroxodiiron(III/III) Intermediate in the Amine Oxygenase, AurF, from Streptomyces thioluteus” J. Am. Chem. Soc., 2009, 131, 13608-13609.

Li, N.; Korneeva Korboukh, V.; Krebs, C.; Bollinger, J. M., Jr. “Four-electron oxidation of p-hydroxylaminobenzoate to p-nitrobenzoate by a peroxodiferric complex in AurF from Streptomyces thioluteus" Proc. Natl. Acad. Sci. U S A., 2010, 107, 15722-15727.

Warui, D. M.; Li, N. (co-first); Nørgaard, H.; Krebs, C.; Bollinger, J. M, Jr. ; Booker, S. J. “Detection of formate, rather than carbon monoxide, as the stoichiometric co-product in conversion of fatty aldehydes to alkanes by a cyanobacterial aldehyde decarbonylase” J. Am. Chem. Soc. 2011, 133, 3316-9.

Li, N.; Nørgaard, H.; Warui, D. M.; Booker, S. J.; Krebs, C.; Bollinger, J. M, Jr. “Conversion of Fatty Aldehydes to Alka(e)nes and Formate by a Cyanobacterial Aldehyde Decarbonylase: Cryptic Redox by an Unusual Di-metal Oxygenase” J. Am. Chem. Soc. 2011, 133, 6158-61.

In preparation Li, N.; Chang, WC.; Warui, D. M.; Booker, S. J.; Krebs, C.; Bollinger, J. M, Jr. “Evidence for Only Oxygenative Cleavage of Aldehydes to Alk(a/e)nes and Formate by Cyanobacterial “Aldehyde Decarbonylase”

Li, N.; Bollinger, J. M., Jr.; Krebs, C. “Mechanism of oxidation of the substrate analog, para-hydrazinobenzoate, by the N-oxygenase, AurF, from Streptomyces thioluteus”

Activities and Awards 2011 Richard L. and Norma L. McCarl Award 2007 - 2008 Graduate Student Fellowship for Graduate Studies 2008-2009 President of the Penn State Chinese Friendship Association, Penn State University 2006-2007 Vice President of the Chinese Friendship Association, University of Florida 2006 Graduate Student Senator, College of Medicine, University of Florida 2003 Chairman of the Student Union in the College of Life Sciences, Nankai University

Curriculum Vitae Ning Li