<<

Applications of resonance Raman to the study of bioinorganic macromolecules

DISSERTATION

Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the Graduate School of The Ohio State University

By

Pearson T. Maugeri, B.S.

Biophysics Graduate Program

The Ohio State University

2017

Dissertation Committee: Professor Hannah S. Shafaat, Advisor Professor James A. Cowan Professor Terry L. Gustafson Professor Marcos Sotomayor © Copyright by Pearson T. Maugeri 2017 Abstract

Transition metals are absolutely critical for the existence and maintenance of biologi- cal systems. Many cellular processes, such as metabolism, the detoxifying of harmful chemicals, gene regulation, and extracellular signaling rely on the unique properties that metal cofactors provide. One very important group of metalloproteins is the ferritin-like superfamily. The ferritin-like superfamily contains proteins that have a large array of functions, ranging from the storage and transport of essential nutri- ents such as iron and oxygen (ferritin and hemerythrin) to the 2-electron oxidation of the stable C-H bonds found in hydrocarbons (di-iron hydroxylases). Of particular interest are the ribonucleotide reductases, enzymes that catalyze the de novo syn- thesis of deoxyribonucleotides from ribonucleotides. Additionally, different classes of ribonucleotide reductases use dissimilar sets of metals to achieve similar end results. In 2009, a novel protein group called the R2-like ligand-binding oxidases, or R2lox, was discovered. The function of R2lox is not currently known, although it does do 2-electron chemistry, as seen by a novel Tyr-Val linkage that forms during the acti- vation of oxygen. This new class of proteins exhibits sequence similarity to the R2 subunit of the ribonucleotide reductases and is reminiscent of bacterial multicom- ponent monooxygenses (BMMs) in both the bound ligand and also its hypothesized chemistry. It has been shown that this protein is upregulated in the virulent strain of Mycobacterium tuberculosis, and therefore it is of great interest to uncover the role that this protein plays in its various hosts.

ii There are many ways to study bioinorganic systems; however, one of the most ver- satile and sensitive techniques is spectroscopy. Many different regimes are able to be plied to furthering the understanding of bioinorganic systems. One of the most useful types of spectroscopy in a chemist’s toolbox is vibrational spectroscopy, due to the opportunity that it affords to extract structural information. One typeof vibrational spectroscopy is Raman , which uses inelastically scattered pho- tons to report on the vibrational modes of a . Furthermore, the resonance Raman phenomenon can and has been used to great effect in bioinorganic systems by tuning the Raman excitation beam to an electronic absorption, enhancing the signal and making resonance a very useful tool for the study of bioinorganic structure and function. This thesis presents two main projects that are different in character. The first is the development of a custom variable wavelength resonance Raman experimental sys- tem. A brief background of resonance is presented before outlining the layout of the experimental setup. Collaborations with other research groups were also conducted, and the custom resonance Raman system was used to probe a wide range of bioinorganic systems. The other section of the thesis details experiments on R2lox. One chapter outlines the investigation of an unusual phenomenon where the optical properties of R2lox change drastically upon irradiation with . The other chapter discusses setting up R2lox to be probed via transient absorption so that electron transfer constants may be extracted.

iii For Charlene, my dearest love

iv Acknowledgments

No man is an island, Entire of itself, Every man is a piece of the continent, A part of the main.

The above excerpt from John Donne’s 1623 work, Meditations XVII: Devotions upon Emergent Occasions, is an absolutely spot on summary of what my academic experience has been up to this point. I could not have done any of this without the firm and unyielding support that I have received from friends, family and fellow academics. I am truly not an island. First, I would like to acknowledge my advisor, Dr. Hannah Shafaat. She has provided an excellent model of scientific excellence to emulate. I would like to thank her for giving me the opportunity to work in her group and to assemble the custom resonance Raman system, which is detailed in chapter 3. I would also like to thank Dr. Dongping Zhong, in whose lab I gained a lot of my fundamental understanding of ultrafast systems as well as biological spectroscopy. In addition, thank you to my thesis committee, who sacrificed their time to give me good feedback on this document as well as on my research projects. Each of you have helped me grow as a scientist and I am very grateful for your mentorship. I would also like to thank my collaborators on the projects discussed in this work. Thank you to Dr. Cowan, Dr. Jingwei Li, and Stephen Pearson for collaborating with me on the project in chapter 6. To Dr. Thomas Rauchfuss and Michaela Carlson, v thank you for your collaboration on the project described in chapter 7. Thank you to Dr. Kyle Lancaster, Dr. John Caranto, and Avery Vilbert for the collaboration on the project described in chapter 8. Special thanks to Dr. Martin Högbom, Dr. Julia Griese, and Dr. Rui Branca for all of your help in running experiments in the study described in chapter 4, including providing mutants, as well as all of your helpful discussions regarding R2lox. Thank you to all for your contributions to this work. I would also like to thank my labmates in the Shafaat research group. Each and every one of you has affected my academic life for the better and I would not be here if it wasn’t for your influence. Thank you Tasha, for helping me to integrate intothis lab and for providing me with someone to bounce ideas off and for watching my dog on the rare occasions I got out of town. Thank you Jeff for all the discussions about Game of Thrones, and for all your help with the resonance Raman system. Thank you Cami for all your help with troubleshooting the labeling conditions in Chapter 5 and for our discussions on the pros and cons of OSU football. Thank you Shelby for helping me with my inorganic chemistry knowledge, especially with point groups, and for all our geeky conversations about video games. I would also like to thank Dr. Michael Stevenson, who provided me with an invaluable viewpoint of the academic process though the lens of a postdoctoral researcher. We will always have Akron. Thank you Sean for being a good TA partner when we taught CHEM 4310. You made teaching that class much more enjoyable. Thank you Effie for your help in all of the R2lox purification and experiments as well as for being a good officemate the past year despite my, shall we say, less than orderly desk. I will look back on all of our discussions about how finicky R2lox is with fondness. I have been blessed beyond measure to have been afforded the opportunity to work and grow scientifically with all of you during my time here. I would also like to thank my undergraduate professors at Harding University

vi who nurtured my love for science. Thank you Dr. Stewart for teaching me how to think outside of the box and solve puzzles in organic synthesis. I have applied that skill to so many other things here at OSU. Thank you Dr. Matlock for taking time out of your day to speak with me about biochemistry and to help me transition to thinking in a more biophysical context. Finally, I would like to thank Dr. Murray, who was so patient with me when I was fumbling around in his upper-level physics courses and had no idea what I was doing. Thank you all for showing me that one can simultaneously be a scientist and a man of God. Finally, I would like to extend my boundless appreciation to family, who has listened to my joys and my frustrations for these last several years. They have always been a source of unwavering support and I cannot thank them enough for everything they have done. Finally, thank you to my amazing wife, Charlene. I would not have finished if it wasn’t for you. You have been by my side this whole time andIamso thankful that were always ready to listen when I had bad news and celebrate with me when I had good news. I would like to thank each and every person who has helped me in my academic journey, but there is not enough room in this document for that. Again, I am in no way an island and the work described herein, while it has my name on it, would simply not have been possible without a massive amount of assistance and support from an incalculable number of sources.

vii Vita

February 17, 1990 ...... Born, Gainesville, FL

May 2008 ...... Dacula High School, Dacula, GA

May 2012 ...... B.S. in Biochemistry and Molecular Bi- ology, Harding University, Searcy, AR September 2012 to August 2014 ...... Graduate Teaching Associate, Depart- ment of Chemistry and Biochemistry and Department of Physics, The Ohio State University September 2014 to August 2015 ...... CMBP Fellow, Cellular, Molecular, and Biochemical Sciences Training Pro- gram, The Ohio State University September 2015 to present ...... Graduate Teaching Associate, Depart- ment of Chemistry and Biochemistry

Publications

Slater, J. W., Marguet, S. C., Cirino, S. L., Maugeri, P. T. & Shafaat, H. S. Ex- perimental and DFT Investigations Reveal the Influence of the Outer Coordination Sphere on the Vibrational Spectra of Nickel-Substituted Rubredoxin, a Model Hydro- genase Enzyme. Inorganic Chemistry 56, 3926-3938 (2017).

Miller, E. K., Trivelas, N. E., Maugeri, P. T., Blaesi, E. J. & Shafaat, H. S. Time- Resolved Investigations of Heterobimetallic Cofactor Assembly in R2lox Reveal Dis- tinct Mn/Fe Intermediates. Biochemistry 56, 3369-3379 (2017).

Maugeri, P. T., Griese, J. J., Branca, R. M., Miller, E. K., Smith, Z. R., Högbom, M. & Shafaat, H. S. Driving protein conformation changes with light: Photoinduced structural rearrangement in a heterobimetallic oxidase. In preparation (2017). viii Presentations

“Driving differential metal geometry formation using light” Maugeri, P. T., Miller,E. K., Smith, Z. R. & Shafaat, Biophysics Seminar Series, Columbus, OH (Sep 2017)

“Driving protein conformational changes with light: Evidence for photoinduced struc- tural rearrangement in a heterobimetallic Mn/Fe oxidase” Maugeri, P. T., Miller, E. K. & Shafaat, H. S., 2017 CMBP/CRB Annual Symposium, Columbus, OH (May 2017)

“Driving protein conformational changes with light: Evidence for photoinduced struc- tural rearrangement in a heterobimetallic Mn/Fe oxidase” Maugeri, P. T., Miller, E. K. & Shafaat, H. S., 2017 IGP Annual Symposium, Columbus, OH (May 2017)

“Driving protein conformational changes with light: Evidence for photoinduced struc- tural rearrangement in a heterobimetallic Mn/Fe oxidase” Maugeri, P. T., Miller, E. K. & Shafaat, H. S., 2016 Ohio Inorganic Weekend, Akron, OH (Nov 2016)

“Using resonance Raman spectroscopy to probe cofactor assembly of R2lox, a novel heterobimetallic oxidase” Maugeri, P. T., Miller, E. K., Trivelas, N. E. & Shafaat, H. S., 2016 CMBP/CRB Annual Symposium, Columbus, OH (May 2016)

“Using resonance Raman spectroscopy to probe cofactor assembly of R2lox, a novel heterobimetallic oxidase” Maugeri, P. T., Miller, E. K., Trivelas, N. E. & Shafaat, H. S., 2016 IGP Annual Symposium, Columbus, OH (May 2016)

“Using resonance Raman spectroscopy to probe cofactor assembly of R2lox, a novel heterobimetallic oxidase” Maugeri, P. T., Miller, E. K., Trivelas, N. E. & Shafaat, H. S., 251st American Chemical Society National Meeting, San Diego, CA (Apr 2016)

“Using resonance Raman spectroscopy to probe cofactor assembly of R2lox, a novel heterobimetallic oxidase” Maugeri, P. T., Miller, E. K. & Shafaat, H. S., 2015 Ohio Inorganic Weekend, Bowling Green, OH (Nov 2015)

“Using resonance Raman spectroscopy to probe a novel Mn/Fe oxidase” Maugeri, P. T., Miller, E. K., Trivelas, N. E. & Shafaat, H. S., CMBP Seminar Series, Columbus, OH (Nov 2015)

ix “Using resonance Raman spectroscopy to elucidate metal incorporation of R2lox, a novel heterobimetallic oxidase” Maugeri, P. T., Miller, E. K., Trivelas, N. E. & Shafaat, H. S., 2015 CMBP/CRB Annual Symposium, Columbus, OH ( May 2015)

“Using resonance Raman spectroscopy to elucidate metal incorporation of R2lox, a novel heterobimetallic oxidase”Maugeri, P. T., Miller, E. K., Trivelas, N. E. & Shafaat, H. S., 2015 IGP Annual Symposium, Columbus, OH ( May 2015)

Fields of Study

Major Field: Biophysics

x Contents

Page Abstract ...... ii Dedication ...... iv Acknowledgments ...... v Vita ...... viii List of Figures ...... xiv List of Tables ...... xvii

Chapters

1 Introduction ...... 1 1.1 Metals are important in biological systems ...... 1 1.2 The ferritin-like superfamily ...... 4 1.3 Class I ribonucleotide reductases: A case study of differential metal- binding ...... 7 1.4 R2-like ligand-binding oxidase: R2 or BMM? ...... 11 1.5 Spectroscopy: A suite of versatile techniques to study bioinorganic systems ...... 14

2 Resonance Raman Scattering: A theoretical treatment ...... 23 2.1 Overview of Resonance Raman Scattering ...... 23 2.2 Overview of Incoherent Light Scattering and the Polarizability Tensor 25 2.3 Molecular Vibrations ...... 30 2.4 Classical Description of Rayleigh and Raman Scattering ...... 33 2.5 QM Treatment of Rayleigh and Raman Scattering ...... 36 2.6 The Resonance Effect ...... 39 2.7 The Utility of Resonance Raman Scattering in Bioinorganic Spectroscopy 43

3 Development of variable wavelength Resonance Raman system . 45 3.1 Introduction ...... 45 3.1.1 Laser pulses and nonlinear ...... 46 3.1.2 Sample excitation and collection optics ...... 57 3.1.3 Raman light separation and measurement ...... 61 3.1.4 System details ...... 64 xi 4 Driving protein structure changes with light ...... 71 4.1 Introduction ...... 71 4.2 Methods & Materials ...... 75 4.2.1 Protein production and purification ...... 75 4.2.2 Sample preparation ...... 76 4.2.3 Iron quantification assay ...... 77 4.2.4 Resonance Raman spectroscopy ...... 77 4.2.5 Quantitative photochemical experiments ...... 78 4.3 Results and Discussion ...... 80 4.3.1 Exposure to light drastically changes the optical spectrum of R2lox ...... 80 4.3.2 Resonance Raman spectroscopy of photoconverted R2lox . . . 83 4.3.3 Spectroscopic results suggest R2lox undergoes photoinduced rearrangement to yield tyrosinate coordinations ...... 85 4.3.4 Y162 is the coordinating tyrosinate ligand ...... 86 4.3.5 Mechanism of photoconversion process in R2lox ...... 89 4.4 Conclusion ...... 92

5 Probing electron transfer kinetics in R2lox via RuII-labeled Cys . 93 5.1 Introduction ...... 93 5.2 Materials & Methods ...... 101 5.2.1 Synthesis of RuII complex ...... 101 5.2.2 Identification and preparation of R2lox Cys mutants . . . . . 101 5.2.3 Preparation of labeled protein samples ...... 103 5.2.4 Luminescence and TCSPC measurements ...... 103 5.3 Results and Discussion ...... 104 5.3.1 Electron transfer calculations using the Pathways model . . . 104 5.3.2 Labeling of R2lox with RuII ...... 108 5.3.3 Luminescence and TCSPC measurements ...... 109 5.4 Conclusions & Future directions ...... 116

6 Applications of resonance Raman spectroscopy: Probing [2Fe-2X] (X=S,Se) glutathione clusters ...... 117 6.1 Introduction ...... 117 6.2 Materials and Methods ...... 120 6.2.1 Sample identity and preparation ...... 120 6.2.2 Resonance Raman spectroscopy ...... 121 6.3 Results and Discussion ...... 123 6.4 Conclusion ...... 125

7 Applications of resonance Raman spectroscopy: The search for a terminal hydride in synthetic FeFe hydrogenase mimics ...... 126 7.1 Introduction ...... 126 7.2 Materials and Methods ...... 128 xii 7.2.1 Sample identitiy ...... 128 7.2.2 Resonance Raman spectroscopy ...... 129 7.3 Results and Discussion ...... 131 7.4 Conclusion ...... 135

8 Applications of resonance Raman spectroscopy: Illuminating the structure and reactivity of a unique ruffled heme in cytochrome P460 ...... 136 8.1 Introduction ...... 136 8.2 Materials and Methods ...... 139 8.2.1 Sample preparation and identity ...... 139 8.2.2 Resonance Raman spectroscopy ...... 139 8.3 Results and Discussion ...... 142 8.4 Conclusion ...... 149

9 Conclusion ...... 150

Bibliography ...... 153

xiii List of Figures

Figure Page

1.1 Phylogenetic tree of the ferritin-like superfamily ...... 5 1.2 Proposed electron transfer pathway in R2a ...... 8 1.3 Comparison of active sites of the class I RNRs ...... 10 1.4 Structure and metal center of Mn/Fe R2lox ...... 12 1.5 and corresponding energy ranges . . . . . 17 1.6 Energy splitting of spins in an increasing magnetic field ...... 18

2.1 diagram comparing vibrational spectroscopy techniques . 25 2.2 Stokes vector representaions of common polarizations of light . . . . . 29 2.3 Angular distribution of scattered light under different incident polar- ization conditions ...... 30 2.4 Energy level diagram comparing vibrational spectroscopy techniques . 42

3.1 Excitation and decay pathways for a system interacting with electro- magnetic radiation ...... 49 3.2 Comparison of bandwidth and pulse duration for CW amd pulsed laser systems ...... 51 3.3 Indices of refraction for ordinary and extraordinary axes of KDP . . . 57 3.4 Increase in collection efficiency with low f-number lens ...... 59 3.5 Labeled image of typical ...... 63 3.6 of Ti3+ ions ...... 65 3.7 Spectrum of toluene/acetonitrile Raman standard ...... 68 3.8 Resonance Raman experimental system ...... 69

4.1 Structure of Mn/Fe R2lox from G. kaustophilus ...... 74 4.2 UV/vis and action spectra for Fe/Fe and Mn/Fe R2lox ...... 81 4.3 Dark control for the photoconversion of R2lox ...... 82 4.4 Resonance Raman spectra of photoconverted R2lox: isotopic labeling and resonance Raman excitation profiles ...... 84 4.5 Resonance Raman spectra of 18O-exchanged R2lox ...... 87

xiv 4.6 Resonance Raman spectra of apo- and photoconverted Fe/Fe and Mn/Fe R2lox ...... 87 4.7 UV/vis absorption spectra of Tyr mutants in close proximity to Fe center in R2lox ...... 89 4.8 Proposed reaction mechanism of R2lox photoconversion ...... 90

5.1 Scheme of inner sphere electron transfer ...... 94 5.2 Nonadiabatic potential energy curves for electron transfer ...... 95 5.3 Multiple step electron transfer reaction scheme ...... 97 5.4 Comparison of single step and multistep ET barriers ...... 98 5.5 Modified Latimer diagram ofII Ru laser flash experiments ...... 99 5.6 Latimer diagram of RuII laser flash-quench experiments ...... 100 5.7 Molecular structure of the RuII label ...... 102 5.8 Cysteine mutant positions for labeling of R2lox ...... 105 5.9 Mechanism for labeling of cysteine residue with RuII label ...... 107 5.10 UV/vis spectrum of ruthenated A67C R2lox ...... 109 5.11 Sample metallations of RuII-labeled R2lox ...... 110 5.12 Luminescence measurements of RuII-labeled R2lox ...... 111 5.13 Time-correlated single counting traces of RuII-labeled A67C R2lox ...... 114 5.14 Time-correlated single photon counting traces of RuII-labeled apo- and Fe/Fe R2lox Cys mutants ...... 115

6.1 Examples of iron-sulfur clusters found in nature ...... 118 6.2 Sample absorbance spectra of various iron-sulfur clusters ...... 119 6.3 Chemical structure of FeS and FeSe cluster samples ...... 121 6.4 UV/vis absorbance spectra of FeS and FeSe cluster samples . . . . . 122 6.5 Resonance Raman spectra of FeS and FeSe glutathione clusters . . . 124

7.1 Active site structures of NiFe, FeFe and Fe-only hydrogenases . . . . 127 7.2 Chemical structures of the synthetic hydrogenase mimics ...... 129 7.3 UV/vis absorbance spectrum terminal-H hydrogenase mimic . . . . . 130 7.4 Resonance Raman spectrum of the non-bridging precursor compound at 457.9 nm excitation ...... 131 7.5 Resonance Raman spectrum of the bridging hydride and deuteride compounds at 457.9 nm excitation ...... 132 7.6 Resonance Raman spectrum of the terminal hydride and deuteride compounds and 457.9 nm excitation ...... 133 7.7 Resonance Raman spectrum of compounds I−V at 364 nm excitation 134

8.1 Crystal structure of hydroxylamine oxidoreductase from N. europaea 138 8.2 Crystal structure of cytochrome P460 from N. europaea ...... 139 8.3 Proposed mechanism of cytochrome P460 ...... 140 8.4 UV/vis absorbance spectrum of cytochrome P460 ...... 141

xv 8.5 Resonance Raman spectra of cytochrome P460 with and without hy- droxylamine ...... 143 8.6 Resonance Raman spectra of cytochrome P460 complexed to 15N and 14N−labeled hydroxylamine under aerobic and anerobic conditions . . 144 8.7 Resonance Raman spectra of cytochrome P460 with 20x DEA/NO and Angeli’s salt under aerobic and anaerobic conditions ...... 145 8.8 Resonance Raman spectra comparing DEA/NO incubated samples with cytochrome P460 without ligand ...... 146 8.9 High− resonance Raman spectra comparing cytochrome P460 with {FeNO}6 and {FeNO}7 cytochrome P460 ...... 147

xvi List of Tables

Table Page

3.1 Phase matching conditions for various nonlinear processes ...... 55 3.2 Optical elements used in resonance Raman experimental system . . . 70

4.1 Extinction coefficients of purple acid phosphatases from various organ- isms ...... 85 4.2 Quantum efficiencies for various R2lox mutants ...... 88

5.1 Electron coupling strengths for selected cysteine residues of R2lox . . 106 5.2 Emission spectra data for RuII-labeled R2lox cysteine mutants . . . . 111 5.3 Luminescence lifetimes for RuII-labeled R2lox cysteine mutants . . . 113

8.1 Resonance Raman samples for cytochrome P460 experiments . . . . . 141

xvii Chapter 1 Introduction

1.1 Metals are important in biological systems

Biological systems are diverse and very complex. The that comprise the inner workings of the cell have many different functions, such as maintaining struc- tural integrity, interacting with outside stimuli, and generating energy to power all the cellular components. For the most part, biological molecules are comprised of the common p-block, nonmetallic elements that show up in any organic chemistry course: carbon, oxygen, nitrogen, hydrogen, sulfur, phosphorus, etc. These building blocks impart important characteristics to biological molecules such as polarity and defined 3-dimensional structures due to a range of electronegativities as well asversa- tile bond formation capabilities.1 While these are crucial in biological structure and function, they alone do not possess the physical characteristics that are required to carry out all the workings of the cell. In addition to those common p-block ele- ments, metals also play an important role in the maintenance of the cell.2 Sodium and potassium are crucial for transmitting electrical signals up and down the axon of a nerve cell, in addition to carrying the signals that contract muscles.3–6 Calcium is a member of many important cellular signaling pathways, such as in the contraction of cardiac myocytes.7–11 Magnesium is a very important component of DNA struc- ture, in addition to acting as a cofactor in many important enzymatic systems.12–15

1 Additionally, metals also play a very important role in the global stabilization of the cellular environment as well as in the regulation of the osmotic pressure that the cells experience.16,17 These are only a few exceptions of all the functions that these elements play in the continual maintenance of life. Within the larger group of metals, there exists a subset called transition metals. These elements are in the d-block region of the periodic table and possess many special characteristics that their compatriots in the s-block do not. One is the ability to participate in redox reactions at biologically-relevant redox potentials.18 That makes transition metals integral parts of biological reactions that require the movement of electrons from one chemical group to another. Most of the transition metals that occur naturally in biology, with a few exceptions, are in the first row of the d-block.19 These are well-known elements, such as manganese, iron, nickel, copper, and zinc. By incorporating these elements, biological macromolecules, namely proteins, can access redox states not normally allowed by traditional organic frameworks.18,19 Another related property of transition metals in biology is to act as electron reservoirs to store energy for use in biologically relevant reactions.20 Finally, transition metals can form more bonds than traditional organic building blocks due to their increased size and d-orbitals, allowing for more complex active site geometries.21 The number of important metabolic reactions that depend on transition metals for their function is immense. The two reactions by which all complex life receives energy from the sun, photosynthesis and cellular respiration, involve metals at the heart of their mechanism. Photosynthesis is the process by which electrons are extracted from to provide energy for the fixation of CO2 into sugars to nourish the organism. One of the main players in the photosynthetic process is photosystem II (PSII). PSII is a large transmembrane enzyme that absorbs light at around 680 nm to cycle its reaction center through four distinct states.22–24 The reaction center,

2 the oxygen-evolving complex (OEC), is a Mn cubane structure with a Ca in one of the corners and a dangling Mn off to one side.25 When a photon is collected by the light-harvesting complex, a tyrosine radical is formed via electron transfer, which in turn oxidizes the OEC.26–28 The OEC is the heart of PSII and where the 4 electrons and 4 protons are extracted from water, forming molecular oxygen in the process.25,29 Cellular respiration is the process by which ATP is produced from the oxidation of sugars. There are many complex reactions that make up the process of cellular respiration.30 One of these reactions is oxidative phosphorylation, which transfers electrons from the reduced cofactors NADH and FADH2 which are produced earlier in the process, to molecular oxygen.31 This process creates water and releases energy that is used to make ATP, the energy currency of the cell.32 There are four complexes, aptly named complexes I-IV, that participate in this transfer, all of which contain transition metal cofactors such as heme units and iron-sulfur clusters.33–38 Transition metal complexes are also involved in many other reactions that extract energy from available raw materials. Many archaea and anaerobic bacteria carry out

2- - reactions that make use of alternate electron acceptors, such as SO4 , NO3 and CO2,

39–41 all of which require metalloenzymes. The CO2-reducing reactions are of interest, considering the current global climate situation.42 Methanogens and acetogens can

catalyze the reduction of CO2 to CH4 and acetate, respectively, to drive the pro- duction of ATP.41 This is done using a combination of enzymes, with active sites containing nickel, iron, and sulfur.43–45 These reactions are quite unfavorable thermo- dynamically, however, through coupling to more favorable reactions, the reduction of

41 CO2 can occur under biological conditions. Transition metals in redox states such as FeIII and MnIV are also used as terminal electron acceptors themselves by some microorganisms.46 When considering the metabolisms of aerobic and anaerobic or- ganisms, it is safe to say that a large percentage of organisms receive their energy

3 from reactions that are catalyzed by metalloenzymes. Metalloenzymes do more than catalyze energy-producing reactions. They can also store and transport needed nutrients that cells need to survive. Well-known metalloproteins that function in this way are hemoglobin and myoglobin, two heme- containing O2 transport proteins that are introduced in many undergraduate bio- chemistry courses as a case study for cooperativity.47–49 Another nutrient-regulating metalloprotein is ferritin. Ferritin is responsible for the regulation of iron in the cell.50 It forms a nanocage around the iron to sequester it and protect the cell from the iron-catalyzed production of highly oxidizing radicals via the Fenton process.51,52 Ferritin is also involved in several other processes, such as immune response, stress re- sponse, and apoptosis.53–55 Ferritin is found in nearly every organism on the planet.56 There are many other examples of transition metals in biology, such as the use of iron nanocrystals by magnetotactic bacteria to sense the force lines of the Earth’s magnetic field and travel along them.57 Again, this is by no means a comprehensive overview of the crucial role transition metals play in the genesis and continuation of life; rather, it is meant as a series of examples illustrating the numerous amount of functions that transition metals play in cellular systems.

1.2 The ferritin-like superfamily

Within the overarching classification of metalloproteins, there exists a large group called the ferritin-like superfamily.58 The ferritin-like superfamily shares a common bundled helix secondary structure as well as a similar di-metal core, which is usu- ally di-iron (but not always, which shall be expanded upon later in this chapter).58,59 Members of the ferritin-like superfamily have a wide array of functions. Ferritin, as mentioned above, sequesters and regulates the levels of free iron in the cell. Addi- tionally, some types of ferritin possess ferrioxidase activity at the di-iron core and are

4 capable of oxidizing FeII to FeIII.60

Figure 1.1: The figure above shows the phylogenetic tree of the ferritin-like superfam- ily, separated by secondary structure motif. The figure was adapted from reference [59].

Rubrerythrins are thought to reduce harmful hydrogen peroxide molecules and lower the oxidative stress of the cell.61,62 There are many types of rubrerythrins in nature, and it is hypothesized that the other members of the ferritin-like superfamily originated from a simple 2-helical rubrerythrin chain, owing to the sequence similarity between the N- and C-terminal pairs of helices of the erythrin subunit of this group of enzymes.58,63 Mn catalases catalyze the disproportionation reaction of hydrogen per- oxide, forming water and oxygen.64 They are the first exception to the di-iron active site that was mentioned above. Mn catalases can either be based on a heme active site or a Mn active site.65,66 Another enzyme group in this family, the type 2 fatty

5 acid desaturases, uses an external ferredoxin electron transfer partner to introduce a double bond into a fatty acid carbon chain.58,67 This subfamily of proteins is thought to be the most similar to the next two groups: the di-iron hydroxylases and ribonu- cleotide reductases.58,68 Di-iron hydroxylases perform O- insertion reactions on C-H bonds in hydrocarbons.69 Methane monooxygenase (MMO), which catalyzes the formation of methanol from methane and oxygen, resides in this group.70–72 MMO is of particular interest due to its ability to break the exceptionally strong (∼105 kcal/mol) C-H bond in methane.73–75 This group also contains enzymes that hydrox- ylate phenols and other aromatic functional groups.76–78 Monooxygenases are being investigated for their possible integration into technology, as they have a wide range of substrates.79 Lastly, the ribonucleotide reductases are a very important class of en- zyme, capable of the de novo synthesis of deoxyribonucleotides from the one-electron reduction of ribonucleotides.80 This reaction is very important to all life on Earth, so this enzyme and others related to it are quite universal across domains of life.80 Ribonucleotide reductases exist as heterodimers, with R1 and R2 subunits.81 The R2 subunit is where the oxidizing equivalent is generated, where it then is transferred to the R1 subunit via hole transfer. The R1 subunit contains the binding site for the ribonucleotide and the location of the enzymatic reaction.82 There are three classes of ribonucleotide reductase: class I, class II, and class III.83 Class I ribonucleotide reduc- tases use a di-metal active site to generate the oxidizing equivalent and will be visited later in this chapter, as they are the only ones of the three classes that are a part of the ferritin-like superfamily.58,84,85 Class II ribonucleotide reductases use a cobalamin cofactor for catalytic activity, whereas class III makes use of a [4Fe-4S] cluster teth- ered to a S-adenosylmethionine (SAM) group.81,86–91 Class I ribonucleotide reductases are the most widespread of the three classes, with the others localized primarily in archaea and eubacteria.80

6 1.3 Class I ribonucleotide reductases: A case study of differential metal-binding

There are three subclasses in the class I ribonucleotide reductases: class Ia, class Ib, and the recently discovered class Ic.85 Class Ia is the longest known and most well-studied of the three classes. It makes use of an FeIII/FeIII/Y• active site to generate the thiyl radical at the substrate binding site.92–95 The catalytic cycle begins with the apoprotein binding 2 FeII atoms, which are now coordinated by mainly carboxylate (Asp and Glu) and imidazole (His) ligands.80 This active site binding

96–98 motif is used throughout the class I ribonucleotide reductases. When O2 binds, the active site follows a series of steps that generates a FeIII/FeIII metal center, also known as intermediate X, which subsequently generates a rather stable radical on Tyr122 via hydrogen atom transfer.99–101 This stable tyrosyl radical has been confirmed by the unique UV/vis absorbance signature at 410 nm as well as by EPR.102–104 The active site is also now bridged by a µ-oxo ligand and the binding mode of Glu238, one of the bridging carboxylates, is changed.80 Tyr122 is at a significant distance from the active site cysteine on the R1 subunit (∼35 Å).80,105 Thus, the radical must be propagated through several proton-coupled electron transfer steps, crossing over the R2→R1 gap by tunneling between R2-Tyr356 and R1-Tyr731 until it reaches Cys439 on the R1 subunit, where the reaction occurs.80,106,107 Many studies have focused on this mechanism, and as a result, it is quite understood at this point, along with how to reprogram it to achieve differential activity.108–112 The class Ib RNRs are less well-understood. They operate in a similar way to the class Ia enzymes in that a tyrosine is the carrier of the oxidizing equivalent from the metal site to the cysteine on the R1 subunit. However, there are differences. While class Ib enzymes can function with an FeIII/FeIII site, the enzymatic reaction is much more efficient with a MnIII/MnIII core.97 Additionally, instead of dioxygen reacting directly with the metal 7 site to produce the active form such as in class Ia RNR, an intermediate reductase called NrdI reduces dioxygen to superoxide (O2•−), which goes on to produce the active form of the dimanganese cofactor.97,113

Figure 1.2: The proposed electron transfer pathway in class Ia ribonucleotide reduc- tase from Escherichia coli. The role of W48 is uncertain. The vertical line represents the boundary between the R2 and R1 subunits. Figure taken from [106].

The di-iron form of class Ib RNR enzymes however, only requires dioxygen to function.114 Both class Ia and class Ib RNR enzymes are found in Escherichia coli, where class Ia is the primary enzyme during typical aerobic conditions.97,113 It is cur- rently not understood what role class Ib RNR plays in E. coli, but it does seem to be produced in higher quantities during periods of oxidative stress.115,116 However, class Ib RNR enzymes do play a primary role in other organisms such as some members of the Bacillus genus and other eubacteria.83,114,117 The third subclass of ribonucleotide reductases is the class Ic RNRs. What sepa- rates this class from the others is the presence of a heterobimetallic Mn/Fe active site, 8 as opposed to a Fe/Fe or a Mn/Mn site.118 In 2007, a novel ribonucleotide reductase from Chlamydia trachomatis was reported that demonstrated RNR activity but had a phenylalanine in the place of the stable tyrosyl radical from the class Ia and class Ib RNRs.98,119 For some time, it was thought that an FeIV/FeIII intermediate would be able to transfer the radical directly, without the assistance of the Tyr122 radical.120,121 Further investigations, however, revealed that a Mn/Fe active site is much more ac- tive than the previously shown Fe/Fe site.118,122 In fact, a unique MnIV/FeIII active site was shown to generate the propagating radical in the place of the stable tyrosyl radical from the other subclasses.118 The only other previously known enzymatically- active, heterobimetallic Mn/Fe protein was the purple acid phosphatase found in Ipomoea batatas (the sweet potato).123 Furthermore, it was also shown that R2c pro- ceeds through a MnIV/FeIV intermediate on the path to the active MnIV/FeIII state.124 These unprecedented high-valent Mn/Fe active sites became an area of intense study. Other than the lack of the nearby Tyr, the electron transfer pathway between the Mn/Fe center and the thiyl radical is fairly similar to that of class Ia RNRs. Many of the intermediate steps are tyrosyl radicals, with Tyr331 and Tyr991 mediating the transfer across the R2-R1 boundary.125 In fact, by trapping the active site radical with an azide-modified ribonucleotide, a direct measurement of the ET transfer distance was made via double electron-electron resonance (DEER) spectroscopy and found to be ∼43 Å.125–127 Surveying these three subclasses of class I ribonucleotide reductases provides a unique view into what metals are needed for certain reactions. Here, there are three different systems, each with different metals, yet all three catalyze the denovosyn- thesis of deoxyribonucleotides. For a long time, it was thought that ribonucleotide reductases only contained di-iron cofactors. However, recent studies have shown that this is not the case. Class Ib RNR enzymes can make use of a homobimetallic core of

9 Figure 1.3: Comparison of the different active sites of the class I ribonucleotide re- ductases. From top to bottom: 1MXR (R2a), 3N37 (R2b), and 4M1I (R2c). R2a is in its oxidized form, whereas R2b and R2c are both in their reduced forms.

10 either Fe or Mn, with a different oxidant generating the hole depending on the metals used. On the other hand, in the case of the class Ic enzymes, while it could undergo a single turnover with a di-iron active site, it was found to be catalytic when allowed to bind both Mn and Fe. This is not the only example of nature utilizing multiple sets of metallic active sites to catalyze the same reaction. For example, superoxide dismutase (SOD), which protects the cell from oxidative stress by catalyzing the dis- mutation of superoxide to dioxygen or hydrogen peroxide, can contain NiIII, FeIII, or MnIII in its active site.128,129 Nature makes use of this theme time and time again depending on certain biological parameters and availability of raw materials to the organism. These observations set up the next protein to be introduced, which is the primary system of focus in this work.

1.4 R2-like ligand-binding oxidase: R2 or BMM?

In the previous section, the active site structure and function of the class I ribonu- cleotide reductases were discussed as well as the difference between the identities of the metal sites and how it pertains to the generation of the oxidizing equivalent that is transferred to the R1 subunit. Bioinformatic studies on the sequences of the class Ic ribonucleotide reductases revealed 2 subgroups.130 The first group contained the recently identified C. trachomatis R2, whereas the second group contained Rv0233.130 This second group was shown to be missing the C-terminal Tyr that mediates the electron transfer between the R2 and R1 subunits in the class I RNRs.130 Rv0233 is a hypothesized virulence factor based on a 7x upregulation in the virulent H37rv strain of Mycobacterium tuberculosis compared to the non-virulent BCG strain.131 Addi- tionally, Rv0233 was shown to not possess any RNR activity.130 Due to similarity of Rv0233 to R2c in sequence and structure, Rv0233 has been denoted as an R2-like ligand binding oxidase, or R2lox. This nomenclature will be used for the remainder

11 of this work.

Figure 1.4: Crystal structure of Mn/Fe R2lox. The inset shows a blown-up image of the Mn/Fe metal center. (PDB:4HR0)

The crystal structure of R2lox shows several unusual features. Firstly, the active site is similar in structure to the R2c active site. ICP-SFMS analysis along with the detection of K-edge emission lines showed an active site that had a 1:1 ratio of Mn/Fe.130 Further studies on the metal-binding properties of a homolog of R2lox from Geobacillus kaustophilus (GkR2lox) show that it can also support a di-iron active site.132 BLAST analysis shows that GkR2lox possesses 41% sequence identity to R2lox from Mycobacterium tuberculosis. Additionally, the active site binding residues are similar to the bacterial monooxygenases (BMMs) and R2 subunits.98,130,132 This suggests that the active site is flexible in its structure depending on its state.133,134 This flexibility is evident in the crystal structures of apoR2lox and both aerobically and anaerobically reconstituted R2lox, as well as the Fe/Fe active sites.132,135 Metal

12 loading experiments in GkR2lox have shown a sequence- and site-dependent metal binding scheme.132 The experiments showed that site 2 heavily favors the binding of FeII, while site 1 can bind either FeII or MnII. Crystal structures of the apo form of the protein reveal an ordered site 2 which is able to be resolved in the electron density map, and a very disordered site 1, lending credence to the possibility of FeII binding first in site 2, which then would impose enough structure on site 1 forII eitherFe or MnII to bind.132 R2lox participates in 2-electron chemistry, similar to that of the BMMs. The met- als bind in the +II oxidation state. Then oxygen binds, which brings the metals to an MIV/FeIV oxidation state. At the end of the oxygen activation reaction, the metals are in a MIII/FeIII state.136 The intermediate steps are currently unknown, but it is an active area of research.137 R2lox can be thought of as a “bridging” species between the R2 subunits and the BMMs because it shares aspects of both groups. In R2lox, an R2-conserved phenylalanine is replaced with an alanine in position 171, which allows for the active site to be accessed by an exogenous long-chain fatty acid.130 This fatty acid is positioned in a way that is reminiscent of the BMMs. spectrometry stud- ies have shown this fatty acid to be a distribution of hydroxylated fatty acids, with the greatest contributions from C-16 and C-18 species.135 Another unusual structural feature of R2lox is the presence of an unprecedented Tyr-Val crosslink in the sec- ondary coordination sphere.130 This Tyr-Val crosslink is hypothesized to provide the 2 electrons that bring the active site to its finalIII M /FeIII resting state.132 The resting state of R2lox is a MIII/FeIII oxidation state, similar to R2a introduced above. This resting state is bridged by a hydroxo ligand, making the resting state an intermediately antiferromagnetically coupled system.136 This is in contrast to R2c, which has a bridging oxo ligand and is a strongly antiferromagnetically coupled sys- tem.136 The oxo ligand provides more superexchange pathways between the two metal

13 centers, increasing the coupling strength.136,138 This means that the FeIII/FeIII has an S=0 state and is therefore EPR-silent. Furthermore, R2lox has also been shown to convert its active site structure upon exposure to light to a more R2c-like structure, with the deprotonation of the bridging hydroxo ligand.136 Currently, the function of R2lox is unknown. As mentioned above, it is closely related structurally to both BMMs and R2. While its structure is reminiscent of the R2 subunit, it performs 2-electron chemistry and seems to be laid out in a way as to invite comparison to the BMMs. If it does possess a function similar to that of the BMMs, then it does so in a more simplified context, without multiple components, and thus its mechanism of action will be of much interest to the scientific community. There are many studies currently being conducted on this system to answer impor- tant questions about its metal-binding properties, oxygen activation mechanism, and overall function.

1.5 Spectroscopy: A suite of versatile techniques to study bioinorganic systems

In any field of science, it is critical to have multiple ways of conducting measurements on systems of interest to extract the most information possible. There are many ways to probe bioinorganic systems. There are of course the traditional protein-based experimental techniques such as and x-ray crystallography.139,140 Bioinformatic analysis is a powerful technique that has been used in recent years to great effect.59 Another relatively new technique is the use of computers to calculate characteristics of biological systems in silico based either on Newton’s laws of motion or solving the Schrödinger equation.141,142 However, the presence of metal cofactors affords additional means of measuring the structure and function of bioinorganic systems. One technique that can be utilized within a bioinorganic framework is

14 electrochemistry. By measuring the current of electrons that flow in and out of the metalloprotein under varying conditions, many inferences can be made about the system and its activity.143,144 Additionally, metalloproteins are especially suited for x-ray crystallography studies due to the heavy atom in the active site, which can be utilized to solve the phase problem of x-ray structure determination.145 While all the above methods are very useful for the study of bioinorganic systems, one of the most versatile and powerful techniques used in scientific measurement is spectroscopy. Spectroscopy is the measurement of the interaction of matter with an electromagnetic field. This interaction is visualized in the form of a spectrum, which conveys the amount of the matter-field interaction with respect to the energy ofthe electromagnetic radiation. Spectroscopy is a useful technique because it can report on molecular identity very quickly and yield information that would be borderline impos- sible any other way. However, like any real-world application of theory, spectroscopy is not perfect. The measured signal is oftentimes very low, and too much light can damage the system. Therefore, long acquisition times and large amounts of sample are needed to be able to make any meaningful observations. Despite these limitations, spectroscopy is a powerful technique that can be applied to many systems. Spectroscopy as a discipline did not really gain its footing until the advent of . One of the central pillars of quantum mechanics is that only a certain number of discrete energy states are allowed by a system. This contrasts with classical theory, which allows for the system to occupy any one of an infinite amount of states. This resolution of the so-called “ catastrophe”, where the energy of blackbody radiation of an object approached infinity (Equation 1.1) as the emitted wavelength approached the UV region, came in 1901 with the introduction of discrete energy packets.146 These packets were only allowed to have energy values of hν, where h is Planck’s constant and ν is the frequency of the energy packet.146 This allowed

15 Planck to derive the correct form of the spectral radiance of a blackbody per solid angle (Equation 1.2).146

2ν2k T B (T ) = B (1.1) ν c2

2hν3 1 Bν(ν, T ) = 2 hν (1.2) c k T e B − 1 After this advancement, in addition to many others, the field of spectroscopy de- veloped rapidly. Now different energies interacting with the system under study were visualized as quantized energy levels. The energy of electromagnetic radiation used depends on what aspect of the system is measured. Figure 1.5 shows the electro- magnetic spectrum and what each energy region corresponds to inside a molecular system. The boundaries between each type of electromagnetic radiation are fuzzy and the ranges discussed in the following paragraphs are approximate. The lowest energy radiation, called ultra, super, or extremely low frequency (ULF, SLF, ELF) radiation, has ranging from ∼1–100 Hz, which corresponds to a wavelength of ∼3000 km to 300 Mm.147 This regime of radiation does not interact with biological matter in a context useful to bioinorganic spectroscopy and will not be discussed further. The next lowest energy radiation are radio . This comprises the region of the spectrum from ∼kHz-GHz, with of ∼0.3 m-300 km. The lower region of the spectrum is still not that useful for bioinorganic spectroscopy; however, the higher energy regime is very useful. All subatomic particles have spin. These subatomic particles can be combined together in nuclei in a manner whereby the overall nucleus also exhibits spin. This nuclear spin has the capacity to interact with magnetic fields. By placing the system in a magnetic field, an energy gap arises from the biasing of the nuclear spin with respect to the external field. This energy gap is linear with respect to the magnetic field strength and corresponds to radio

16 Figure 1.5: The figure above is a depiction of the electromagnetic spectrum accom- panied with approximate energy ranges.

17 frequencies in the ∼100 MHz-1 GHz range.148

Figure 1.6: The figure above shows graphically the split in energy that arises from the interaction of a non-zero spin system and an increasing magnetic field.

These frequencies can be used to excite nuclear spins. The excited nuclear spins then decay back to the equilibrium state. This decay is measured to extract useful information about the environment in which the spins reside. This is the origin of nuclear magnetic resonance (NMR) spectroscopy. NMR is a very important tech- nique because it allows for accurate measurement of structure and dynamics on the molecular level with little to no damage to the system.149–151 A caveat, nuclei must have a non-zero spin to be NMR-active.152

18 radiation is on the same energy scale as rotational energy levels, from ∼1-100 GHz, corresponding to wavelengths of ∼1-100 mm. Many experiments have been conducted on ultracold molecular systems to measure their rotational spec- tra and extract valuable constants.153–155 However, microwave frequencies can also correspond to electron spin levels. Analogous to NMR, when an electron is placed in an external magnetic field, their spin orientation is biased and can be probed via microwave radiation.156 This gives rise to electron paramagnetic resonance (EPR) spectroscopy, which can yield a wealth of information about the local geometric and electronic structure around the electron spin.157–159 Like NMR, there is a caveat; the system is required to be paramagnetic, meaning that it has at least 1 unpaired elec- tron.156 Although it is common for EPR experiments to be run at X-band (∼10 GHz) frequencies, there are many examples of EPR experiments conducted on bioinorganic systems at Q-band (∼30-50 GHz) and W-band (∼70-110 GHz) frequencies as well as even higher regimes.160–163 The next region of the electromagnetic spectrum is infrared radiation, which spans from ∼1-100 THz (3-300 ￿m). Within this broad classification, there are several sub- regions. The first is terahertz (THz) spectroscopy. THz spectroscopy is a very useful technique for monitoring collective motions of aqueous solvent around a protein.164,165 This is because it covers the frequency range from ∼1-10 THz. The time scales for oscillation of radiation in this regime are picoseconds, which correspond with water dynamics.166,167 The next several regions are far-, mid-, and near-infrared. This re- gion ranges from ∼10-100 THz. Here it becomes useful to begin referring to energies in wavenumbers (cm-1). The infrared region spans from ∼30-3000 cm-1. This is the region that interacts with molecular vibrations. All particles are constantly vibrating, even at absolute zero. This is due to the Heisenberg uncertainty principle, and quan- tum fluctuations at near-zero temperatures is an active area of research168 today. –170

19 Vibrational spectroscopy is very useful to chemists. By monitoring the vibrations of molecules, information about their structure as well as their environment can be gathered.171,172 In bioinorganic chemistry, many of the metal-centered vibrations are difficult to pick up with Fourier transform infrared (FT-IR) spectroscopy due torea- sons discussed in Chapter 2. However, some bioinorganic systems have the capability to form metal-ligand bonds to CO and CN, which have very strong FT-IR signatures and can report on many characteristics of the metal center.173,174 The interaction of UV and visible light with matter causes the valence electrons to change energy levels. This region of the electromagnetic spectrum ranges from ∼0.4-100 PHz (3-750 nm). The lower energy regime is the , which is typically reported in nanometers. The range of visible light is ∼400-700 nm. The absorbance of this regime of light can report on the identity of the sample as well as the populations of different species. Bioinorganic systems have an advantage over non-metal-containing biological systems in the visible range in that it is common for them to be colored due to the absorption of visible light by the metal cofactor. Protein scaffolds typically only absorb in the UV portion of the spectrum, due toaromatic side chains such as tyrosine and tryptophan.175 In addition to absorbance, fluorescence spectroscopy, which yields information on the environment of the fluorescent molecule and its emission lifetime, is also commonly conducted with visible and UV light.176–178 The more energetic region of the UV/vis spectrum of light begins to bleed into the x-ray region and is classified as ionizing radiation. X-ray radiation was discovered in 1896 by Wilhelm Röntgen, who gave them the name x-rays, after an unknown quantity from , due to their mysterious nature.179 X-rays occupy the 0.1-10 EHz portion of the spectrum, which corresponds to wavelengths of ∼30-3000 pm and are highly ionizing. While UV and visible light affects the energy levels of the valence electrons, x-ray radiation causes a change inthe

20 energy levels of the core electrons. X-ray absorbance and fluorescence can yield infor- mation on the identity of the metal cofactor and its coordination environment.180,181 Finally, gamma rays are the highest region of the electromagnetic spectrum, ranging in energies from 10 EHz and beyond. These rays interact with the energy levels of nuclei. This is distinct from the nuclear spin orientation, which is probed with NMR. Mössbauer spectroscopy is a very sensitive technique that makes use of gamma pho- tons at 14.4 keV (about 860 pm) to probe the environment around an 57Fe nucleus.182 Vibrational spectroscopy holds a very important place in the field of chemistry. Based on the vibrational energy levels of the molecule, many inferences can be made about its structure and function. As mentioned above, FT-IR makes use of infrared light to probe vibrations via absorption. Two-dimensional infrared absorbance spec- troscopy can be used to correlate vibrations and identify vibrations that are coupled together in the molecule.183,184 FT-IR is not the only vibrational technique available, however. Other types of vibrational spectroscopy can report on molecular vibrations without using IR light, each with their own pros and cons. These include Raman spectroscopy, nuclear resonance vibrational spectroscopy (NRVS), and vibrational coherence spectroscopy (VCS).185–188 In Raman spectroscopy, inelastically scattered are used to report on the vibrational frequencies of the sample.185 Of spe- cial interest is resonance Raman spectroscopy, which uses the visible absorbance of the sample to enhance Raman scattering and enable the study of samples of smaller concentrations.189 This work described herein follows 2 paths: a development of resonance Raman spectroscopic instrumentation and how it can be used to solve problems in the field of bioinorganic chemistry, as well as experiments that further the understanding of R2lox. In chapter 2, a theoretical treatment of resonance Raman spectroscopy is given, as well as how it emerges from the Raman scattering phenomenon and how

21 it reports on vibrations. Chapter 3 describes the resonance Raman instrumentation that was developed for studying bioinorganic systems. Chapter 4 outlines a series of experiments that were done on R2lox to investigate an interesting phenomenon where its optical characteristics where dramatically changed upon exposure to light. Chapter 5 describes work done to lay the foundation to study electron transfer in R2lox by using ruthenium modification and pump-quench experiments. Chapters 6,7 & 8 give details about collaborations with other research groups using the resonance Raman experimental system described in chapter 3 and demonstrating the wide range of applications that resonance Raman spectroscopy has in the field of bioinorganic chemistry.

22 Chapter 2 Resonance Raman Scattering: A theoretical treatment

2.1 Overview of Resonance Raman Scattering

As introduced in Chapter 1, resonance Raman spectroscopy is a powerful technique for the study of bioinorganic systems. Raman scattering was discovered in 1928 by C. V. Raman when he noticed how the wavelength of light changed when viewed through a material.190,191 Briefly, Raman scattering probes the vibrational frequen- cies of a material by quantifying the difference, or ”Raman shift” of the wavelength between incoming and inelastically scattered photons. By measuring the vibrational frequencies of the molecule, inferences about its structure can be made. Thus, Raman spectroscopy has found a niche in forensics, medical diagnosis, and materials science, as well as in basic chemical and physics research.185 While powerful, this technique is inherently a low yield process, with only about 1 in every 1,000,000 photons undergo- ing Raman scattering, while the rest undergo , where the incoming and scattered light have the same wavelength.192 The invention of the laser in 1960 allowed Raman spectroscopy to be taken to new heights as a method of chemical identification.193 allowed researchers to use much higher photon fluxes, thus making the measurement of inelastically scattered photons less difficult and allowing Raman spectroscopy to grow in utility.

23 As discussed above, Raman spectroscopy measures the vibrational frequencies of a system via inelastically scattered photons. This is in contrast to IR spectroscopy, which measures vibrational frequencies via the absorption of infrared photons. The difference between these techniques is shown in Figure 2.1. These two spectroscopic techniques give different information about the system of interest. This is dueto the different selection rules for each technique. There are some advantages ofRa- man spectroscopy over IR for biological systems. The first reason is that biological macromolecules exist in an aqueous environment, which has very large IR bands that can obscure signal from the system, due to the strong dependence of IR absorption intensities on the molecular dipole. In contrast, the of Raman scattered light depends on the polarizability, which means that the Raman bands of water are not as large and will not obscure the system of interest. Additionally, Raman spectroscopy can also detect vibrations that do not have a particularly strong dipole. The latter reason is a more practical one. Raman spectroscopy uses UV or visible light to probe vibrations, whereas IR spectroscopy relies on the infrared portion of the electromagnetic spectrum. IR light is intrinsically more difficult to manipulate than UV or visible light. In addition, there is a wider variety of optical devices that can be applied to the visible portion of the spectrum. There are further reasons that Raman scattering is beneficial for the purposes of this study which will be discussed in a later section. In the following sections, the theory of Raman scattering will be discussed. In section 2.2 incoherent light scattering will be introduced. Next, in section 2.3 the basics of molecular vibrations will be discussed which will lay the groundwork for the introduction of vibrational spectroscopy. The two following sections, 2.4 and 2.5, will discuss the theory of Raman scattering in a classical and quantum mechanical framework, respectively. Section 2.6 will introduce the resonance Raman effect and

24 Figure 2.1: The figure above compares different types of vibrational spectroscopy. From left to right: infrared absorption, Rayleigh, Stokes Raman, and anti-Stokes Raman.

finally, section 2.7 will elaborate on the utility of resonance Raman scattering within the field of bioinorganic spectroscopy.

2.2 Overview of Incoherent Light Scattering and the Polarizability Tensor

Before the theory of Raman scattering can be developed, it is important to introduce incoherent light scattering. Both Raman and Rayleigh scattering are incoherent pro- cesses. When radiation is incoherently scattered, the photons are out of phase with each other. Thus, the waves interfere with each other and cause a lower signal to be obtained. In coherent light scattering, all the photons are in phase with each other. The basic model of light scattering that will be employed here is this: a plane wave that represents the incoming light wave of intensity I interacts with the molecular electron cloud and scatters off with a total power of P. Because the light waveis an oscillating electromagnetic field, it can interact with the molecule and induce the formation of a time-dependent dipole moment in the molecule. This induced dipole

25 moment can be written as a summation of linear and nonlinear terms,

µind = µL + µNL (2.1)

where

µL = µ(1) (2.2)

and

µNL = µ(2) + µ(3) + ... + µ(n) (2.3)

The linear term consists of a rank-2 tensor that relates to the linear contribution of the induced dipole to the electric field as follows:

(1) · µρ = αρσ Eσ (2.4)

The nonlinear term is, as the name suggests, nonlinearly dependent on the electric field. The nonlinear contribution can be written as follows,

1 1 . µNL = β : E E + γ . E E E + ... (2.5) ρ 2 ρστ σ τ 6 ρστδ σ τ δ The higher order terms will be revisited in section 3.1.1. For now, because the nonlinear terms are much lower in magnitude than the linear term, we will only concern ourselves with the latter in the development of the theory of Raman scat- tering. The rank-2 tensor in the linear polarization term is called the polarizability tensor, and it is the primary determinant of the intensity of the Raman scattering of a molecule. The polarizability tensor describes how easily the electron cloud can be distorted. Molecules that have large polarizability tensors hold their electrons more loosely than those with small polarizability tensors. Usually, large molecules exhibit

26 large polarizabilities because their outer electrons are not held as tightly due to the nuclear screening effect of the inner electrons. Let us now take a closer look at the polarizability tensor. Being a rank-2 tensor, it has the form of a 3x3 matrix, which can interact with an electric field in the x-, y-, and z-axes, like shown below.  

αxx αxy αxz     αρσ = α α α  (2.6)  yx yy yz

αzx αzy αzz For a randomly oriented molecule, each vector of the electric field induces the propa- gation of an induced dipole along the 3 directions. For example, the x-component of the electric field, Ex, creates the following polarization fields.

µx = αxxEx

µy = αyxEx (2.7)

µz = αzxEx

This can be repeated for all three components of the electric field. It was stated above that the induced dipole depends on the electric field linearly, and since each component of the electric field can induce the propagation of the dipole in any ofthe axes, then the polarization vectors µx, µy, and µz can be written as

µx = αxxEx + αxyEy + αxzEz

µy = αyxEx + αyyEy + αyzEz (2.8)

µz = αzxEx + αzyEy + αzzEz

27 The above system of equations can be written in a less cumbersome way,

∑ µρ = αρσEσ (2.9) σ where ρ runs over the molecular polarizations and σ enumerates the components of the electric field. The relationship between I and P is linear, as shown by the equation below.

P = σsI (2.10) where I is the incident irradiance and P is the scattered power integrated over all space. In this equation, the proportionality constant, σs, is called the scattering cross- section. It is related to the square of the polarizability tensor, shown by the following equation.194

8πe4 ∑ σ = ω4 |(α ) |2 (2.11) fi 9c4 L ρσ fi ρ,σ The scattering cross-section is reported in units of m2, or in barns (b), where one

−28 2 2 barn is equal to 10 m , or 100 fm . The typical values of σs for Raman scattering are on the order of 10 mb or 10−30 m2.195 Additionally, light scattering has a very strong polarization dependence. The polarization state of an electromagnetic wave can be represented by Stokes vectors.196 The elements of these column vectors (I, Q, U, and V) represent various polarizations of light. Figure 2.2 shows the Stokes vectors for various common polarizations. During a scattering event, the incident polarization state can be transformed into the scattered polarization state via the scattering matrix, a 4x4 matrix where the matrix elements contain information about how the electromagnetic wave interacts with the particle. Equation 2.12 shows the projection of the incident vector in the

28 Figure 2.2: The above figure shows example Stokes vectors for common polarizations.

polarization space to the scattered vector.

       Is  S11 S12 S13 S14  Ii              Qs S21 S22 S23 S24 Qi         =     (2.12) U  S S S S  U   s   31 32 33 34  i 

Vs S41 S42 S43 S44 Vi The matrix elements are calculated from the scattering coefficients, which are themselves calculated from the transformation of the solution of the vector wave equation in spherical coordinates. This derivation is beyond the scope of this work, the complete derivation may be found in chapters 3-5 of reference [196]. The result is represented in the polar plot in Figure 2.3. The distribution is in the scattering plane, which is the plane produced from the vector of the incident light and the detected scatter. The directionality is measured relative to the incident vector. The trace shows the scattering distribution for light polarized parallel to the scattering plane, 29 Figure 2.3: The polar plot above shows the angular distribution of scattered light under different incident polarization conditions. The red trace is when the incident light is polarized parallel to the scattering plane, is polarized perpendicular, and purple is unpolarized.

the blue for light polarized perpendicular to the scattering plane, and the purple trace is for unpolarized incident light. As seen in Figure 2.3 there is a node when collecting at 90°relative to the incident vector when the incoming light is polarized parallel to the scattering plane. This will be very important to consider in chapter 3 when the experimental system is described.

2.3 Molecular Vibrations

Molecules are constantly in motion due to being surrounded by energy. A large part of this energy manifests in vibrations of the nuclei in the molecule. Molecular vibrations are periodic in nature, with frequencies ranging from ∼1013 to 1014 Hz.

30 This is equivalent to a period on the order of 10 fs. Energetically, this is around ∼300 to ∼3000 cm-1. Vibrations can be more or less energetic depending on the identity of the atoms involved. Molecules have a total of 3N−6 vibrations, where N is the number of atoms that make up the molecule and the subtracted modes take into account translational and rotational motion.197 The exception to this is molecules where the atoms are in a linear geometry, in which case the number of vibrations is equal to 3N−5 due to the absence of a unique rotation about the molecular axis. It is important to discuss molecular vibrations when laying the groundwork for Raman scattering, because the observed frequency spectrum of the Raman scattered light directly reports on the vibrations in the sample. The simplest way to model a is by using a spring. This models the periodic molecular vibration using Hooke’s law, which states that the restoring force of a spring is related to the displacement by a proportionality constant k. This results in the following relation,

∂U F (r) = = −kr (2.13) ∂r where U is the potential energy in the spring, and can be written as

1 U(r) = kr2 (2.14) 2 This potential energy surface is a function of displacement and is a parabola. If one writes down the time-independent Schrödinger equation using this surface as the potential,

[ ] h￿2 1 − ∇ + kr2 Ψ = EΨ (2.15) 2m 2

31 When the above expression is solved, eigenstates of the form

( ) 1 E = n + hω￿ (2.16) n 2 are calculated. At the ground state, n=0, and thus the energy of the molecular

1 vibration is 2hω￿ . This means that even at 0 K, molecules still vibrate at some frequency that is non-zero, even when all molecular motion as stopped. By modeling the molecular vibration as a spring, one can also calculate the fre- quency of motion by

√ k ω = (2.17) µ where µ is the reduced mass and is calculated by

m m µ = 1 2 (2.18) m1 + m2 This becomes a powerful approximation because now the prediction of the vibrational shift can be calculated when one or more isotopes are introduced to the molecule. This is exceptionally useful in vibrational spectroscopic techniques because hypothe- sized intermediates can be experimentally observed using vibrational spectroscopy in conjunction with isotopic labeling. Finally, while the harmonic approximation is useful because it is simple to solve, it does not model real molecules perfectly. At higher vibrational energy levels, the harmonic approximation begins to fall apart. This is because in the harmonic ap- proximation, all of the energy levels are exactlyhω ￿ apart. In reality, the spacing between vibrational energy levels begins to decrease in a nonlinear fashion. This is due to the of the molecular bond.198 To better model this behavior,

32 other potentials such as the (shown below) have been used.

( )2 −β(r−r0) U(r) = De 1 − e (2.19)

In the equation above, De is the well depth, β controls the width of the well, and r0 is the equilibrium bond distance. The solutions of the Schrödinger equation using the Morse potential as the energy surface are more complex than the simple harmonic oscillator, but they model real molecular vibrations more accurately.

2.4 Classical Description of Rayleigh and Raman Scattering

In this section, the theory of Raman scattering will be presented in the framework of classical physics. It is useful to begin here, even though the classical theory cannot fully explain all aspects of the Raman scattering phenomenon. This section will begin with the linear polarization term introduced in 2.2, and derive equations that describe the Rayleigh and Raman scattering processes, while discussing properties of Raman scattering. Qualitatively, Raman scattering occurs when an incoming photon collides with a molecule and distorts the electron cloud, inducing the formation of a dipole. The interaction can be modeled as an ”absorption” where the electron is promoted to a high-lying vibrational state and then relaxes. Most of the time, the photon is elastically scattered, causing it to depart with the same energy it originally had, and the electron returns to its previous energy level. However, occasionally, a photon is inelastically scattered, meaning it will have a different energy when it departs, and the electron settles to an energetic state that is not equal to where it started. As mentioned above, the rank-2 polarizability tensor governs the Raman scattering process. It linearly relates the electric field, E to the induced dipole created by the

33 impinging electric field. The tensor, α, may be written as a Taylor expansion around the equilibrium bond position,

( ) ( ) ∑ ∂α 1 ∑ ∂2α α = (α ) + ρσ q + ρσ q q + ··· (2.20) ρσ ρσ eq ∂q i 2 ∂q q i j i i eq i,j i j eq where qi denotes the coordinate along the with frequency νi. In order to simplify the derivation, all terms after the first derivative term will be discarded. Additionally, only one vibration will be considered here, so the summation may also be dropped. Thus, the expression may be written as,

( ) ∂αρσ αρσ = (αρσ)eq + qi (2.21) ∂qi eq

Next, the displacement, qi will be written in a time-dependent manner, due to the temporal interaction with E. To simplify this expression, the molecular bond will be approximated as a simple harmonic oscillator, which at low temperatures is a valid approximation. Thus, qi now takes the form of

qi = qi,eq cos(2πνit + ϕi) (2.22) where ϕi is a time-independent phase factor. Plugging this into equation 2.21 yields the following expression:

( ) ∂αρσ αρσ = (αρσ)eq + qi,eq cos(2πνit + ϕi) (2.23) ∂qi eq Now, a time-dependent electric field will be introduced, which represents the incoming field of frequency νE

E = E0 cos(2πνEt) (2.24)

The electric field, when combined with Equations 2.9 and 2.23, (again, dropping the

34 sum on the condition only one vibrational mode is present) gives an E-dependent expression for the molecular polarization.

( ) (1) ∂αρσ µ = αeqE0 cos(2πνEt + E0qi,eq cos(2πνit + ϕi)cos(2πνEt) (2.25) ∂qi eq

By substituting in the trigonometric identity,

1 cosA cosB = [cos(A + B) + cos(A − B)] (2.26) 2 the molecular dipole term may be rewritten as a sum of polarizations,

(1) (1) (1) (1) µ = µ (2πνE) + µ (2π(νE + νi)) + µ (2π(νE − νi)) (2.27)

where

(1) Ray µ (2πνE) = µ cos(2πνEt)

(1) Ram µ (2π(νE + νi)) = µ cos(2π(νE + νi)t) (2.28)

(1) Ram µ (2π(νE − νi)) = µ cos(2π(νE − νi)t)

The cosine functions above give the molecular dipole frequency, and the µRay and µRam terms are the polarizations that give rise to Rayleigh and Raman scattering, respectively. The second equation is known as anti-Stokes Raman scattering, due to the scattered photon possessing more energy than the incident photon. On the other hand, Stokes Raman scattering occurs when the scattered photon is less energetic than the impinging photon. Stokes Raman scattering is more intense than anti-Stokes due to the larger population of molecules in the ground vs excited vibrational states. As mentioned earlier in the section, infrared absorption and Raman spectroscopy complement each other quite well, despite the fact that they both measure vibra-

35 tional modes. Firstly, the selection rules for IR and Raman spectroscopy are dif- ferent. The intensity of IR signal depends on the derivative of dipole with respect to displacement of the mode. Thus, only modes that exhibit a change in dipole are IR-active. This leads to a more general statement that only molecules which possess permanent dipoles are IR active. This is in contrast to Raman spectroscopy, which depends on the derivative of the polarizability with respect to the displacement of the normal mode. Thus, vibrations that exhibit no change in polarizability are not Raman active. This leads to a difference in modes that are detected by IR vs Ra- man spectroscopy. Usually, molecules that have strong dipoles, such as water, are good infrared absorbers, and molecules that have large electron clouds and are easily polarized exhibit strong Raman scattering.

2.5 QM Treatment of Rayleigh and Raman Scat- tering

Now that the classical theory of Raman scattering has been presented, it will now be treated quantum mechanically. Looking at this phenomenon through the lens of quantum mechanics is important because Raman scattering deals with electrons on the molecular level, which are small enough that quantum effects can manifest. Thus, to correctly model Raman scattering, the system must be treated quantum mechanically. In addition, utilizing QM will also provide a smooth transition into an explanation of the resonance Raman effect. In this section, the history of the develop- ment of the Kramers-Heisenberg-Dirac (KHD) relation will be discussed. From there, the sum-over-states (SS) and time-dependent (TD) representations of the polarizability tensor will be discussed, along with the strengths and weaknesses of each representation. In 1925, Kramers and Heisenberg postulated the dispersion relation that describes

36 the scattering of a photon by an electron in an atomic environment.199 This relation was very important, because it allowed for a ”negative absorption” (stimulated emis- sion), as well as predicting inelastic scattering of light, which would later be observed by C. V. Raman.190,191,200 Shortly thereafter, it was derived by P.A.M. Dirac by mod- eling the atom in a quantized electric field and applying second-order time-dependent perturbation theory.201,202 Dirac also used the KHD dispersion relation to calculate the Einstein A and B coefficients for the absorption and emission of light, further validating its viability as a theoretical tool.202 Suppose a molecule is resting in the ground vibronic state I. A photon then in- teracts with the molecule and scatters off, leaving the molecule in vibronic state F. Just as in the classical treatment, the electric field of the photon induces a dipole in the molecule. The strength of the dipole and the scattered light is dependent on the polarizability of the molecule. The KHD dispersion relation gives the form of the polarizability tensor, (αρσ)fi, in terms of the initial and final wavefunctions Ψi and

Ψf ,

∑ ⟨ΨF | µˆρ | ΨN ⟩⟨ΨN | µˆσ | ΨI ⟩ ⟨ΨF | µˆσ | ΨN ⟩⟨ΨN | µˆρ | ΨI ⟩ (αρσ)fi = + (2.29) ENI − EL − iΓ ENF + EL − iΓ N where EL is the incident photon energy, |ΨI ⟩, |ΨN ⟩, and |ΨF ⟩ are the initial, interme- diate, and final vibronic states. ENI and ENF are the differences in energy between the intermediate and initial states, and between the intermediate and final states, respectively. The transition dipole operator is represented by µˆ and is summed over the indices ρ and σ. Finally, Γ is the intrinsic linewidth of the intermediate state N that arises from the energy-time form of the Heisenberg uncertainty principle.194,200 The polarizability tensor here is imaginary, however, it is squared to give a real value for the Raman scattering cross-section.194

37 Equation 2.29 is known as the ”sum over states” (SS) form of the polarizability tensor. To calculate the Raman scattered intensities, the above expression is summed over all the intermediate states N. This causes a problem for most molecules due to a very large number of intermediate vibronic states as the system size grows. Addition- ally, the eigenstates grow difficult to calculate as more anharmonicity is introduced.200 As such, the SS form of the polarizability tensor is computationally difficult for large systems. Thus, an alternative method for the calculation of α was developed from time-dependent perturbation theory.200 This new form of the polarizability tensor was termed the time-dependent (TD) theory of Raman scattering. Briefly, the TD form makes use of the time evolution operator to propagate the system forward in time,

iHt Ψn(x, t) = Ψ(x, 0)ne h￿ (2.30) where Ψ(0) is the wavefunction at time zero, and H is the Hamiltonian operator for the system of interest.203 This occurs due to the ground state wavepacket |i⟩ undergoing a vertical transition to another potential surface and then propagating across according to Equation 2.30. This formalism greatly reduces the complexity of the Raman scattering cross section compared to the sum-over-states formalism. It is important to note that the TD method does not mean that the measurement itself is that of a time-dependent variable. The TD formalism is mathematically equivalent to the SS method and is a steady state picture of the light scattering process. Rather, the time part of the TD formalism refers to the wavepacket on the higher energy surface that propagates in the time domain.194

38 2.6 The Resonance Effect

As mentioned above, Raman scattering is intrinsically a very inefficient process. For every 106 photons that strike a sample, only about 1/1000 of them are Rayleigh scattered photons, and only 1/1000 of those are Raman scattered.204 As such, it is not as useful for molecules that cannot exist at a high number density or exhibit a high enough scattering efficiency. For instance, proteins are usually stable only until about 0.001 M, and even then can degrade when exposed to high levels of laser radiation. Contrast this with toluene, which is very polarizable due to the aromatic π electron system, but also can achieve a concentration of 9.4 M in a pure sample. Thus, if one wishes to use Raman scattering to probe a protein system, a more efficient way of obtaining the data is required. This application is wherein the utility of the resonance Raman effect lies. The resonance Raman effect occurs when the incoming laser radiation is closeor identical to an electronic transition of the sample. Under these conditions, the electron is promoted to an excited state surface instead of a high-lying vibrational state. This causes vibronic coupling between the ground and excited states, which in turn causes an enhancement in the intensity of scattered light, up to 106 for some modes.205 Throughout this section, the quantum mechanical Raman equations introduced in section 2.5 will be discussed within the context of resonance. Quantum mechanics is crucial for a theoretical treatment of the resonance Raman effect. Thus, the presentation of the resonance Raman effect will begin withEquation 2.29, the KHD dispersion relation. As mentioned before, resonance Raman scatter- ing occurs when the frequency impinging electromagnetic field matches an allowed electronic transition of the target system. As the frequency of the impinging photon approaches an allowed electronic transition, the first term in Equation 2.29, known as the ”resonance term”, begins to dominate the polarizability. This can be expressed 39 below in the following limit:

∑ ⟨Ψ | µˆ | Ψ ⟩⟨Ψ | µˆ | Ψ ⟩ ⟨Ψ | µˆ | Ψ ⟩⟨Ψ | µˆ | Ψ ⟩ lim F ρ N N σ I + F σ N N ρ I → EL ENI ENI − EL − iΓ ENF + EL − iΓ N (2.31) ∑ ⟨Ψ | µˆ | Ψ ⟩⟨Ψ | µˆ | Ψ ⟩ = F ρ N N σ I ENI − EL − iΓ N Equation 2.29 may be further processed by invoking perturbation theory under the condition of resonance.206 The complete analysis is outside the scope of this chapter, but in short, the electronic Hamiltonian is perturbed away from the equilibrium position in coordinate space. The Hamiltonian is related to the nuclear coordinates through the Born-Oppenheimer approximation.207 After this treatment, as well as further simplifications such as only taking one electronically excited into account as well as restricting the sum over states to one vibrational mode, the polarizability tensor may be written as a sum of terms206,208

(αρσ)fi = A + B + C + D (2.32)

The terms involve different couplings between ground and excited states. TheCand D terms here are very small, and will not be considered further in this section.206 The A-term is a consequence of the Condon approximation, which states that the electronic transition moment operator is evaluated at the ground state equilibrium geometry.209 This can be written as

0 µge = µge. (2.33)

The A-term enhancement is given by the following equation,

∑ ⟨f | n⟩⟨n | i⟩ A = (µ0 ) (µ0 ) (2.34) ge ρ eg σ ϵ − ϵ + ϵ − E − iΓ n n i 0 L

40 where |i⟩, |n⟩, and |f⟩ are the initial, intermediate, and final vibrational states and

0 0 (µge)ρ and (µeg)σ are the ρ-th and σ-th components of the electronic transition mo- ment between the ground and excited states at the equilibrium nuclear coordinate Q0. The electronic transition moment is outside of the sum because a set of coordinates may be chosen so that only one element of the electronic transition moment is nonzero and the summation over the indices of the tensor can be eliminated. The terms in the denominator represent the energy levels of the resonance Raman scattering process; ϵi and ϵn are the energies of the initial and intermediate states relative to the zero point energy level of the ground and excited states, respectively, and ϵ0 is the separation of the lowest vibrational levels of the ground and excited states. The levels are depicted graphically in figure 2.4. The A-term is non-zero if and only if every term in the numerator is non-zero. This means that the electronic transition moment components must be an apprecia- ble magnitude for resonance Raman scattering to occur. Additionally, the overlap integral must also be non-zero, which means that the vibrational states must be non- orthogonal to each other. This can only occur when the mode in question is totally symmetric because there must be displacement of the equilibrium position in the excited state relative to the ground state.210 The B-term is more complicated than the A-term and gives the electronic coupling between two electronic excited states, labeled e and s. It is smaller in magnitude than the A-term. The B-term is given by the following equation,

k ∑ 0 0 0 0 H (µ )ρ(µ )σ ⟨f | Qk | n⟩⟨n | i⟩ + (µ )ρ(µ )σ ⟨f | n⟩⟨n | Qk | i⟩ B = es ge gs gs ge E − E ϵ − ϵ + ϵ − E − iΓ e s n n i 0 L (2.35)

41 Figure 2.4: The figure above shows presents a model for the Stokes resonance Raman scattering of an incoming (blue) photon. The energy levels are labeled as they are referred to in Equation 2.34 Here, |g⟩ and |e⟩ refer to the ground and excited electronic states, respectively .

where

⟨ ( ) ⟩ ˆ k ∂H Hes = e s (2.36) ∂Qk 0 is the vibronic coupling operator and is a consequence of the breakdown of the Born- Oppenheimer approximation. B-term resonance Raman enhancement is the result of going beyond the Condon approximation, and expressing the transition dipole moment in a Taylor series in molecular normal coordinates209

( ) 3∑N−6 0 ∂µge µge = µge + Qk + ... (2.37) ∂Qk k=1 0 The B-term allows for forbidden electronic transitions to ”intensity borrow” from an

42 allowed one.209 The closer in energy these two electronic transitions are, the higher the B-term enhancement, as seen from the denominator in Equation 2.35. Additionally, the B-term allows for enhancement of non-totally symmetric modes, which are prohib- ited by A-term enhancement.205,206,208 This can be readily observed in heme systems. The vibrations coupled into the Q-band excitations (∼500-550 nm) can couple into the A-term vibrations that are enhanced by Soret band (∼400 nm) excitation. This will be discussed further in Chapter 8.

2.7 The Utility of Resonance Raman Scattering in Bioinorganic Spectroscopy

Biological systems that utilize metals to perform their function are of great inter- est due to their importance in many processes that are integral to life. Arguably the most crucial reactions on Earth, photosynthesis and respiration, use metals to shuttle electrons around to accomplish the goal of converting available chemicals into useful energy for the continuation of multicellular life. With the advent of quan- tum mechanics in the 20th century, spectroscopy has become a very powerful tool in measuring chemical structure and change. Thus, using spectroscopy to measure the reactions that are central to life on Earth is of great interest to the scientific com- munity. Specifically, vibrational spectroscopy is very useful in the study of chemical systems, as vibrations directly report on molecular structure. As mentioned in sec- tion 2.1, Raman spectroscopy is useful for the study of biological macromolecules due to its use of UV/visible light as opposed to infrared for IR absorption. Addition- ally, due to differing selection rules, the detection of molecular vibrations via Raman spectroscopy is not as hampered by the presence of water as IR absorption is. Fur- thermore, as outlined in section 2.6, most proteins are not stable at concentrations around 0.001 M, making conventional Raman very difficult. Resonance Raman can

43 alleviate this problem, however, in the field of bioinorganic spectroscopy, resonance Raman becomes even more useful as a technique. Proteins are very large molecules, ranging in size from ∼1000 to millions of Da. They contain many atoms, so the number of normal modes is very large, in addition to a very large density of states that makes identifying individual modes impossible. However, as resonance Raman spectroscopy enhances modes that are coupled into an electronic transition, the metal centers of bioinorganic systems can be isolated from the rest of the protein scaffold by means of the visible absorption peaks present. Due to the large enhancement of these peaks, the rest of the protein is effectively invisible to the photon detector, allowing for a site-selective probe of the active site. Resonance Raman spectroscopy has been used in many studies involving metal centers in proteins and protein mimics. These studies include work on the structure of nonheme diiron centers, the structure of iron-sulfur cluster proteins, and heme- containing proteins, among others.211–218 Additionally, resonance Raman has been used extensively to study the mechanism by which bioinorganic processes take place, notably within oxygen activation reactions of non-heme iron proteins.75,111,219 In con- clusion, Raman scattering in a very powerful spectroscopic technique that can report on the molecular vibrations and thus the structure of a molecule of interest. Further chapters will discuss the instrumentation used to measure Raman scattered light.

44 Chapter 3 Development of variable wavelength Resonance Raman system

3.1 Introduction

As mentioned in chapter 2, Raman scattering is a very inefficient process, as about only 1 in 1,000,000 photons are Raman scattered.192 The original Raman experiments made use of high intensity mercury lamps, but the signal was very low.190,191 With the advent of the laser in the early 1960’s, coherent light was available to do Raman scattering experiments.193 This led to a massive increase in the utility of Raman spectroscopy as a probe of molecular vibrations due to an increase in measurable signal. There are many types of Raman experiments, ranging from black box set ups utilizing a single continuous wave laser to multiple beam experiments that make use of femtosecond pulses of light to measure vibrations in the excited state.220–223 The system used in this work is laid out in the following way. First, a laser is needed, due to the massive gain in efficiency of Raman scattering that coherent light provides. This laser, to satisfy the resonance criteria, must be tuned to an excitation band of the sample. The laser is directed onto a sample at a geometry that prevents

45 the laser beam from being directly measured. The scattered light is collected and focused into a spectrograph, with the Rayleigh light filtered out along the way. The spectrograph resolves the light by frequency and directs it towards the optical imaging device. Historically, the optical imaging device that was used was a and a monochromator was used to step through the region of interest.224 Recently, as technology has improved, CCD cameras have been used to image whole regions of the spectrum at once, thus cutting down on data acquisition time.225 There are many commercial Raman systems available, but many of them are black box systems, akin to a tabletop FT-IR instrument. A common wavelength is 785 nm, which is produced from a GaAlAs semiconductor. This wavelength is beneficial because it eliminates a large amount of sample fluorescence while remaining a high enough frequency to give a good Raman signal, due to the ν4 dependence of Raman scattering.226 1064 nm laser systems benefit from a further decrease in fluorescence of some systems, although scattering is ∼3x lower.227 Other common wavelengths are 532 nm and 514.5 nm, along with many others.224,228–230 However, because resonance Raman utilizes the resonance effect, wavelength tunability of a single system is desired to yield the most versatility. To meet this goal of versatile measurement of many bioinorganic systems, a custom resonance Raman system was developed. In the following sections, each component of the experimental setup will be explained in greater detail, as well as how it fits into the landscape of the past and current technology.

3.1.1 Laser pulses and

The first part of any modern Raman system, resonance or otherwise, is a coherent light source, namely a laser. As stated before, lasers are essential for resonance Raman experiments of biological systems due to the increase in Raman scattering

46 efficiency arising from the coherent nature of laser radiation. This helps tooffsetthe low number density that biological systems possess. Here coherence refers to both spatial and temporal coherence. Spatial coherence is a measure of how well a wave correlates with itself at different spatial points along its path.231 Light with a high degree of spatial coherence exhibits a strong phase relationship at its wave front.231 Temporal coherence is how well a wave maintains a phase relationship with itself at different points in time.232,233 Single frequency lasers can generate light with strong temporal coherence, as the wave propagates in a sinusoidal fashion. In contrast, a light bulb emits light in different phases and frequencies and does not exhibit spatial or temporal coherence. Coherence is a fundamental property of a laser that makes it uniquely suited to the measurement of molecular systems. When Raman scattering occurs, it is the interaction of the electromagnetic field with the electron cloud ofthe molecule that produces the scattered light.206 If the incoming electromagnetic fields have the same phase, then the scattering will be more efficient. Throughout this section, the basic theory behind the creation and propagation of laser light will be discussed, as well as how it pertains to the system used in this work. The word “laser” stands for “light amplification by the stimulated emission of radiation”. In basic terms, laser light amplifies itself through the phenomenon of stimulated emission of a gain material. This gain material can vary depending on the desired wavelengths, which are determined by the excitation and decay pathways of the gain material.230,234 Stimulated emission is the emission of a quantum of light by a particle that is stimulated by an incoming photon. The created quantum of light has the same characteristics (frequency, phase, polarization, wavevector orientation, etc.) as the original photon.235 This can be thought of as the opposite of absorption, where a particle is promoted to a higher excited state by an incoming photon. The other common decay pathway is spontaneous emission, in which a particle in an excited

47 state decays with no outside influence. The coefficients which govern absorption, spontaneous and stimulated emission are called the Einstein A and B coefficients.235 These pathways are shown in Figure 3.1. Through the property of detailed balance, which states that in equilibrium the number of particles leaving a quantum state must equal the number of particles entering the state, the conditions that are essential for lasing can be described. Consider a two-level system, with upper and lower states, u and l, separated in energy by νul. Thus,

NuAul + NuBulu(ν) = NlBluu(ν) (3.1)

where Nu and Nl are the number of particles in the upper and lower states, respec-

tively, Aul,Bul, and Blu, are the Einstein coefficients for spontaneous emission, stim- ulated emission, and absorption respectively, and u(ν) is the photon energy density

at νul which can be expressed by the following equation,

8πhη3ν3 u(ν) = hν (3.2) c3(e kB T − 1) which was Planck’s solution to the ultraviolet catastrophe that puzzled scientists at the turn of the 20th century. By applying some straightforward algebra, the ratio between stimulated and spontaneous emission from the upper level can be calculated,

Bulu(ν) 1 = hν (3.3) A k T ul e B − 1 This expression will be equal to one if the exponent term is ln(2) or about 0.693. For a wavelength of 450 nm, this corresponds to a temperature of about 46,000 K. This is the reasoning behind why blackbody emission blue shifts as temperature increases. Therefore, in most situations, spontaneous emission far outweighs stimulated emis- sion. However, in a laser, this cannot be the case, as stimulated emission is at the

48 Figure 3.1: The figure above shows the various decay and excitation pathways when a system interactions with an incoming photon.

very core of how a laser operates. For a laser to amplify light, the ratio of stimulated to spontaneous emission must be much greater than one. This is accomplished by a population inversion of the gain material. At equilibrium, there exists a distribution of states. The probability

236 that a particle exists in state i is dependent on the energy of that state, Ei. This is called the Boltzmann distribution. Since lasing wavelengths depend on the electronic energy level of the gain medium, which typically is several orders of magnitude higher than kBT at ambient temperatures, there are very few particles in the excited state. For lasing to occur, the gain medium needs to be in a population inversion, which means that most of the emitting particles need to be in the upper excited state.235 This will not occur without some sort of external pumping mechanism that inputs

49 energy into the system. Electrical pumping can be used for gas and semiconductor lasers, where electrons are used to create a population inversion by colliding with the gain medium.235 Solid state and dye lasers benefit from optical pumping, where either a high intensity lamp or an external laser provides energy in the form of photons to create the population inversion.235 When a population inversion is attained, the process of stimulated emission causes an exponential increase in the amount of photons throughout the length of the laser crystal. This can be shown mathematically from the following equation,

( ) − gu σul[Nu g Nl]z I(z) = I0e l (3.4) where I(z) is intensity inside the gain material at length z, I0 is the initial intensity, gu and gl are the degeneracies of the upper and lower levels, σul is the stimulated emission cross section, and z is the length of the medium the laser is traveling through.235 Since

σul and z are always positive, if

gu Nu > Nl (3.5) gl then the intensity of the laser will increase exponentially as a function of medium length. As seen above, the mode that gains the majority, keeps the majority due to the self-amplification process.235 Short pulses can be created by adding plane waves at frequencies around the center frequency so that they constructively interfere at one point and destructively interfere everywhere else in the time domain.237 The more plane waves that are added, the shorter the pulse. This large bandwidth that ultrafast pulses contain can be very useful for certain measurements.238 This can be seen in Figure 3.2. Additionally, ultrafast pulses are also useful in nonlinear optics, which will be discussed later. Since nonlinear processes are dependent on the peak intensity

50 Figure 3.2: The figure above compares CW (top) and pulsed (bottom) lasers. The left hand column shows the plane wave each type of pulse contains. The middle column is a representation of the electric field of each pulse, and the final columnis the bandwidth.

of the pulse, ultrafast pulses can drive nonlinear processes with very high efficiency. However, ultrafast lasers are not appropriate for steady state resonance Raman experiments, due to their high peak power and large bandwidth, which can mask the relatively small bandwidth of Raman scattered light. On the opposite end of the pulse duration spectrum, there are continuous wave (CW) lasers. These lasers have one frequency and thus, effectively zero bandwidth. This makes them ideal for steady state Raman experiments because their small band- width allows for the resolution of Raman scattered light. Additionally, CW lasers do not have a peak power, because there is no pulse. However, a downside to CW lasers is that to efficiently drive nonlinear processes, either a prohibitive amount of poweror very specialized optical equipment is needed.239–241 Another downside to CW lasers is the lack of wavelength tuning. To combine a narrow bandwidth with the ability to efficiently drive nonlinear

51 processes, a quasi−CW laser can be used.242 Quasi−CW lasers use pulses that are on the ps timescale to reach intensities that can drive nonlinear processes while keeping bandwidth around 1−20 cm-1.242 This is useful because quasi−CW lasers will not mask the Raman scattered light and the peak pulse power is low enough that the sample is not destroyed before a sufficient amount of data is collected. Additionally, quasi−CW lasers have enough peak power to drive nonlinear processes, albeit not as efficiently as shorter pulses.242–245 This means that a large range of wavelengths can be accessed by a single laser, which is immensely useful for resonance Raman experiments. As mentioned above, the true advantage of the quasi−CW laser system as an excitation source for resonance Raman experiments is the ability to drive nonlinear optical processes while maintaining both a low bandwidth and peak intensity. A nonlinear optical process can occur when the polarization response of a material is not linear with respect to the electric field.246 In Chapter 2, the field-induced molecular dipole was given as,

(1) · µρ = αρσ Eσ (3.6)

where αρσ is the polarizability tensor of the molecule. In this case, the induced dipole was linear in the electric field. However, there are also higher order effects thatcan be induced,

(2) 1 µρ = βρστ : EσEτ 2 (3.7) 1 . µ(3) = γ . E E E ρ 6 ρστδ σ τ δ In this section, the molecular picture introduced in chapter 2 will be replaced by that

52 of a materials-focused picture,

(1) P = ϵ0χ · E

(2) (2) P = ϵ0χ : EE (3.8) . (3) (3) . P = ϵ0χ . EEE

where P is the polarization, ϵ0 is the vacuum permittivity, χ is the electric suscep- tibility, and E is the impinging electric field. Since the polarization response of the material can depend nonlinearly on the electric field, many interesting and useful pro- cesses arise that can give rise to a wide variety of effects. The nonlinear dependence of polarization on electric field originates from the shape of the potential well where the electrons reside.247 When acted upon by an external sinusoidal field, the electrons oscillate in a non-sinusoidal fashion, due to the anharmonic potential. This gives a nonlinear response of the material which can yield very useful effects. Here, second harmonic generation will be used as a case study. Second harmonic generation (SHG) is a 2nd order process that induces the formation of a wave that is double the frequency of the original wave.246 Assume two electric fields of frequencies

nd ω1 and ω2 are incident on a medium that has a 2 order polarization response. The total incoming electric field can be modeled by the two electric fields and their complex conjugates,

− ∗ E iω1t E iωt E1(t) = 1e + 1 e (3.9) − ∗ E iω2t E iωt E2(t) = 2e + 2 e

53 Together, the impinging electric field is a linear sum of the two incident fields,

E(t) = E1(t) + E2(t) (3.10) − ∗ − ∗ E iω1t E iωt E iω2t E iωt E(t) = 1e + 1 e + 2e + 2 e where E is the envelope function and the complex exponential is the modulation within the envelope function. When the above expression is substituted into the polarization equation for a 2nd order process, the following expression is the result,

(2) (2) P = ϵ0χ : EE

− ∗ − ∗ (2) (2) E iω1t E iωt E iω2t E iωt 2 P = ϵ0χ :( 1e + 1 e + 2e + 2 e ) (3.11) (2) (2) 2 −i(2ω1)t 2 −i(2ω1)t −i(ω1+ω2)t P = ϵ0χ :(|E1| e + |E2| e + E1E2e

∗ − − E E i(ω1 ω2)t | |2 | |2 + 1 2 e + E1 + E2 + c.c)

The terms that contain 2ω1 and 2ω2 are the second harmonics and ω1 +ω2 and ω1 −ω2 are sum frequency and difference frequency generation, respectively. When ω1 = ω2, these terms become SHG and optical rectification terms, or zero-frequency terms, respectively.248 These effects and more happen every time the laser interacts withthe polarizable medium, however, the many other higher-order effects are much weaker and are not observable under typical conditions, similar to the higher order effects from Chapter 2. Additionally, there are other conditions that need to be met to efficiently drive a nonlinear process. One of these conditions, and probably one of the most important condition, is phase matching. Phase matching occurs when the refractive index for the fundamental wave matches the refractive index for the second harmonic wave.249 This can take various forms, depending on the nonlinear process of interest. Table 3.1 has some examples of phase matching conditions matched up with the corresponding

54 nonlinear process. The wavevector k can be converted to the index of refraction by the following expression,250

ω k = n(ω) (3.12) c The index of refraction can be either measured experimentally, or calculated by the Sellmeier equation.251 In this work, the crystals that are used are uniaxial, meaning that there are 2 unique indices of refraction along the 3 principle axes.252 The axis

that has the unique the index of refraction (ne) is called the extraordinary axis and

252 the other two are the ordinary axes (no). The optic axis is defined as the path which the light propagates through the crystal.

Table 3.1: Phase matching conditions for various nonlinear processes

Phase matching Identity Nonlinear process conditions Sum frequency ω + ω = ω k + k − k = 0 generation (SFG) 1 2 3 1 2 3 Second harmonic 2ω = ω 2k − k = 0 generation (SHG) 1 3 1 3 Difference frequency ω − ω = ω k − k − k = 0 generation (DFG) 1 2 3 1 2 3 4-wave mixing ω − ω + ω = ω k1 − k2 + k3 − k4 = 0 (self-diffraction) 1 2 3 4

The extraordinary axis and one of the ordinary axes are perpendicular to the optic axis. For critical phase matching, which is used here, phase matching can be achieved by changing the angle of the crystal with respect to the incoming beam according to Equation 3.13,

( ) ( ) 1 cos θ 2 sin θ 2 2 = + (3.13) ne(θ) no ne 55 2 where ne(θ) is the index of refraction of the extraordinary axis at the angle speci- fied.253. For example, consider a KDP (potassium diphosphate) crystal with 1064 nm light impinging upon it. The indices of refraction are given in Figure 3.3. Because the wavevector of the second harmonic wave is twice the magnitude of the wave vector of

the fundamental, and the phase matching condition for SHG is 2k1-k2=0, then

( ) ω 2ω 2 n(ω) = n(2ω) c c (3.14) n(ω) = n(2ω)

The indices of refraction must be the same for the fundamental and second harmonic for phase matching to occur. To satisfy this requirement, the 1064 nm and 532 nm phase velocities must be equal,

c n = (3.15) v where c is the speed of light in a vacuum and v is the phase velocity of light in a medium. Thus, because longer wavelengths typically have higher phase velocities through a material due to the wavelength dependence of the refractive index, the fundamental must propagate through the crystal on the ordinary axis and the second harmonic will exit the crystal perpendicular to the fundamental, along the extraordi- nary axis.251 Equation 3.13 is then solved and found to be an angle of approximately 36 degrees. Additionally, polarization is very important for nonlinear optics, as the electric fields inside the crystal must be aligned with the indices of refraction. Inthe

example above, the two incident electric fields must be aligned with no and the second

harmonic will be aligned with the ne axis. The above example is for type I critical phase matching. There are other types of critical phase matching, as well as noncrit- ical phase matching, which does not require an angle adjustment.254 Rather, because

56 Figure 3.3: The figure above shows the ordinary and extraordinary indices of refrac- tion for KDP at 1064 and 532 nm. The figure on the right shows a side view ofthe crystal and indicates the optic axis.

the index of refraction is temperature dependent, the temperature of the crystal can be increased to achieve phase matching conditions.250 This method of phase matching was not used in this work and will not be discussed further.

3.1.2 Sample excitation and collection optics

After the frequency of the laser light is doubled, it must be incident onto the sample. This is a very crucial part of the resonance Raman experimental set up. Many systems use geometries such that the laser light passes through the sample and the Raman scattered light is collected at a right angle.224 This is probably the most common sample geometry due to its simplicity. Another example is a back-scattering geometry, where the laser is incident on the sample at a 45° angle and the Raman scattered light is collected.255 Because this method does not rely on transmission of the laser light through the sample, it is very useful if one wishes to conduct experiments on frozen samples or highly absorbing ones to avoid self-absorption.256,257 As seen in figure 2.3, 57 when light is polarized parallel to the scattering plane (in this system the scattering plane is formed by the incident light vector and the collection vector perpendicular to it) that there is a node at 90° relative to the incident vector. This polarization of light is called s-polarized, whereas the polarization normal to the scattering plane is p-polarized. It is important to keep this in mind when designing Raman scattering detection systems. Another important consideration for Raman experiments is the amount of laser irradiation incident on the sample. Consider a 0.001 M sample in a 90° scattering geometry with a 1 cm path length. This leads to ∼1015 molecules in the beam path. If a 10 mW beam of laser light at 450 nm of diameter 1 mm is incident on the sample, then there are ∼1016 photons striking the sample every second; additionally,

-1 -1 15 assuming that ϵ450=5000 M cm , this leads to ∼10 photons absorbed, which is about 1 photon for every molecule per second. For sensitive samples, this can cause degradation quite quickly. This becomes even more important when one considers that resonance Raman experiments are conducted under conditions of resonance, and thus even more energy is being absorbed by the molecule and must be dissipated. Additionally, many Raman experiments are conducted by focusing the beam onto the sample, increasing the photon flux per unit area even more. Thus, many systems use means of reducing the amount of time each molecule spends in the path of the laser beam.189 This is accomplished by flowing a stream of sample past the beam or rotating the sample holder so that the beam is constantly hitting another area of the sample.75,258,259 After the Raman scattering event occurs, the scattered light is collected and fo- cused into the measurement device. This needs to be done as efficiently as possible, due to the small amount of Raman scattered light. Many set ups use a low f-number collimating lens to accomplish the efficient gathering of light.75,260 The f-number of a

58 Figure 3.4: The figure above shows how using a low f-number lens can inrease the efficiency of scattered light collection.

lens is defined in the following way,

F f = (3.16) D where f is the f-number, F is the focal length of the lens, and D is the effective diameter of the lens. In the case of signal collection, D is the diameter of the lens, but when F is calculated with respect to laser light, the diameter of the laser beam is used if the laser beam has a smaller diameter than the focusing lens. A low f-number lens is beneficial because then the lens can be put in close prox- imity to the sample and simultaneously have a large diameter. With a f-number of 2, a 2 inch lens would have to be placed 4 inches away to optimal collimation of light. However, a 2-inch lens of f=1 could be placed only 2 inches away and col- lect an increased number of Raman scattered photons. When using lenses, however, the problem of is encountered. Chromatic aberration is when a lens cannot focus all incoming wavelengths to a single point, or in this case, all

59 wavelengths originating from a source will not be collimated properly.261 Rather, the light will be convergent or divergent. This is due to the wavelength dependence of the index of refraction.251 To combat chromatic dispersion, an aspheric lens may be used. An aspheric lens does not have a spherical profile; instead it possesses a more complicated surface profile of the mathematical form,

2 ∑m r 2n z(r) = √ + A2nr (3.17) r2 − n=2 R(1 + 1 (1 + κ) R2 ) where z is the height of the lens along the profile, r is the radial coordinate, Ris the radius of curvature,κ is the conic constant (different values give different conic sections), and the summation term at the end allows for higher order corrections.262 An aspheric lens can eliminate or drastically reduce chromatic aberration in the col- lection of Raman scattered light and thus increase the detection efficiency. Another way of collecting the Raman scattered light is via fully reflective methods, by utilizing parabolic mirrors to collect and collimate the light.259,263 Reflective optics have the benefit of being frequency-agnostic over the regimes that bioinorganic resonance Ra- man is concerned with. However, parabolic mirrors are much more difficult to align for this system and thus were not used for light collection. After the scattered light is detected, the Rayleigh light must be separated from the Raman scattered light, due to the much higher levels of Rayleigh scattered light. Rayleigh scattered light can mask Raman scattered light at low Raman shifts, which are important in monitoring vibrations that involve metals. The separation of Rayleigh and Raman scattered light can be accomplished in several ways. One of these methods is by using a lowpass or notch filter. Lowpass or notch filters will letall wavelengths through that have a lower frequency than the specified wavelength (low- pass) or block only one specific wavelength, with a certain bandwidth (notch).189,264 For anti-Stokes Raman, a highpass or notch filter must be used instead. This method

60 of separation is simple and effective at removing Rayleigh scattering from the data collection. However, each filter is only effective at a certain wavelength and forreso- nance Raman, it is desirable to operate at many wavelengths. An alternate method is by separating the wavelengths in space and selecting only the Raman light to be passed into the spectrograph. This can be accomplished by a dispersive optical el- ement, such as a .265 This solution is usable over a large wavelength range, however, it is more difficult to set up and maintain.

3.1.3 Raman light separation and measurement

After the Raman scattered light is collected and separated from the Rayleigh light, it is measured. In the original Raman experiments, this was accomplished with a spec- troscope.190,191 Later experiments utilized a monochromator which stepped through the range of interest and a photodiode which measured the Raman intensity at each step.189,224 After camera technologies become more advanced, charge-coupled device (CCD) cameras were used in conjunction with to image whole parts of the Raman spectrum at once.189,225 In this section, the spectrograph and CCD camera will be discussed in further detail. Spectrometers are tools that separate light in the frequency domain so that it may be imaged by some measurement device, in this case a CCD camera. Spectrographs utilize dispersive elements to separate the light into its constitutive elements, such as or gratings. Gratings tend to be more effective than prisms. This is due tothe wavelength dependence of light through glass, which makes adjustments with a prism more difficult, as well as the absorption of light by the glass itself. Transmission optics can also temporally disperse optical signals as well, although that is not a concern with this system as the resonance Raman measurements are done in a steady-state time regime. Additionally, gratings can more efficiently separate light over a shorter

61 distance than prisms, making them ideal for use in compact spectrographs.266 Grat- ings are ruled, meaning that they are manufactured via etching grooves periodically via a mechanical process, or holographic, meaning that they are made from exposing a photoactive layer to a laser-generated inference pattern.267,268 Ruled gratings are more efficient, but holographic gratings are more uniform and have less stray light. A diffraction grating is a surface than is formed into a periodic structure that reflects light and separates it into frequency components(Figure 3.5). This behavior is described by the aptly named grating equation,

d sin θm = mλ (3.18) where d is the periodic distance of the grating, θm is the angle of scattering relative to the normal vector of the grating, m is an integer and represents the order of diffraction, and λ is the wavelength of light.269 This above equation can also be written as

d[sin θm + sin θi] sin θm = mλ (3.19)

269 where θi is the angle of incident light relative to the normal vector of the grating. Gratings disperse light because of the constructive and destructive interference due to the small path length differences that occur when light interacts with the periodic grating surface.269 These small path length differences create phase differences. When these phase differences are equal to λ/2, then a spot of minimum intensity occurs, and when there are equal to λ a spot of maximum intensity occurs. The incident light is separated into frequency components based on the wavelength dependence of the grating equation. The diffraction order arises from a light wave of wavelength λ possessing multiple spots of maximum intensity due to the periodic nature of light.269 Light of wavelength λ will diffract in the first order at the same angle that lightof wavelength λ/2 will diffract in the second order. Also, note that at zero order, the

62 Figure 3.5: The figure above shows a typical diffraction grating, with values from equation 3.19 labeled on the figure.

grating behaves as a mirror and light is merely reflected, with no diffraction. After the light is separated into frequency components, it needs to be imaged to produce the Raman spectrum. The early days of Raman spectroscopy used phos- phorescent screens.190,191 As technology progressed, monochromators were used that isolated narrow bandwidths of light; the monochromator was stepped through the de- sired range to build the Raman spectrum.224 Recently, charge coupled device (CCD) cameras are used to image the entire spectrum at one time.189 This has allowed for a greater throughput than measuring the intensity of one wavelength at a time as well as reducing the time that a sample is in front of the laser beam.

63 CCD cameras operate as a shift register.270 Each pixel has a semiconductor which is photoactive. When light strikes this region, an electron is produced via the photo- electric effect.271 The electron is kept separate from the positive hole that is produced by a potential field on the other side of an insulator. The pixel is then instructed by a control circuit to shift the electron to the next pixel, which shifts its electron to the next pixel, and so on.270 The electrons are collected at the end of the chip are converted to a voltage signal and this is read by a computer and a Raman spectrum is produced. CCD cameras are generally used for their ability to image whole regions of a spectrum simultaneously, as mentioned earlier in the chapter. However, CCD cameras do suffer from some disadvantages. Because of the size limit of CCD cameras, theycan only image relatively small amounts of the spectrum at a time. Additionally, the pixels of CCD cameras have a lower signal-to-noise than a photodiode, although many new developments have increased the imaging capability of CCD cameras. Additionally, due to the readout time required to operate CCD cameras, they cannot detect signals as quickly as other measurement devices, although that is not of concern in this application. CCD cameras are typically more efficient at lower temperatures due to temperature fluctuations causing unwanted electrons to mask the measurement electrons. This is called the dark current. Thus, many CCD cameras need to be cooled to allow for measurements that require a high signal-to-noise ratio. This can

be accomplished via liquid N2 cooling, cryo coolers, or thermoelectric junctions (the Peltier effect).272,273

3.1.4 System details

In this work, the quasi−CW system that was used was a 10 ps Ti:Sapphire laser purchased from Spectra Physics. The ps laser was pumped by 20.5 W of 532 nm

64 Figure 3.6: The figure above depicts TiIII fluorescence. The green arrows represent the optical pumping excitation and the red arrows are the fluorescence of the TiIII ions. Ti3+ is very effective as a tunable laser source for reasons given235 inref[ ].

light produced by a diode pumped solid state (DPSS) laser. The Ti:Sapphire laser is capable of producing wavelengths ranging from 700 nm up to 1000 nm, originating from the broad fluorescence spectrum of theIII Ti ions embedded in the sapphire matrix (Figure 3.6.) The wavelength is controlled by a birefringent plate in the cavity that allows certain wavelengths through depending on the incidence angle of the laser light and the pulse duration is set by a Gires-Tournois interferometer. The laser is capable of outputting up to 4.2 W of 800 nm light. After exiting the cavity, the laser beam is reflected at a 90° angle by a Pellin-Broca prism, which disperses any back-scattered light off of optical elements so that it doesn’t enter back into the laser cavity. This light was then directed onto a 2 mm BBO (barium bo- rate) crystal from Eksma Optics, which was type I critically phase-matched with the

65 incoming beam to produce the second harmonic of the laser light, with wavelengths ranging from 350 to 500 nm. Because of the smaller range of the SHG light compared to the fundamental, there is a gap in frequency coverage between 500 nm and 700 nm. To bridge this gap, the resonance Raman system also includes a CW Kr/Ar ion laser capable of running experiments at 514.5, 568.2 nm, and 647.1 nm. The Ti:Sapphire possesses the ability to extend into the UV-B and UV-C region through 3rd har- monic (THG) and 4th harmonic (FHG) generation.244 Even lower wavelengths can be reached, although due to the angular limit of BBO with critical phase-matching, to go lower than 206 nm requires a sum of the 3rd and 4th harmonic waves.245 The light is focused and recollimated by a pair of 1 in, 33 mm plano-convex lenses and then a plano-cylindrical lens is used to adjust the beam profile of the second harmonic. The second harmonic is separated from the fundamental by a second Pellin-Broca prism. The Pellin-Broca directs the second harmonic to a dichroic long pass mirror which rids the beam of the remaining fundamental light. The light is directed at a 90° angle and then periscoped down using a pair of aluminum UV- enhanced mirrors. As discussed above, it is important to consider polarization. The fundamental laser beam is s-polarized (parallel to the scattering plane) and the second harmonic is p-polarized (perpendicular to the scattering plane and parallel to the laser table). Thus, the mirrors must be aligned in a way as not to change the polarization of the second harmonic. For a further discussion, see sections 2.2 and 3.1.2. Table 3.2 shows the optical components used in the experimental system. The excitation beam was directed towards the sample in several geometries, de- pending on the application. Many room temperature samples were taken at a 90° geometry where the Raman scattering was collected at a perpendicular angle from the excitation beam. Samples at 77 K were collected at a back-scattering geometry, where the beam was directed to the back of the sample and the scattered light was

66 collected at 45° relative to the excitation beam. The light was collimated by an as- pheric lens and focused into the spectrograph. The spectrograph was single-grating, which allowed for high throughput data collection. Various gratings ranging from 1200 grooves/mm to 3600 grooves/mm were used depending on the wavelength used and desired resolution. The light was imaged onto a back-illuminated CCD camera cooled to -75 °C by the Peltier effect. The CCD was connected to the data acquisition machine and the measurements were read out to proprietary software that was bundled with the CCD camera. To make use of the spectra that were collected, a calibrated vector of values was needed for the frequencies that the data was collected at. Before every experiment, a 50/50 toluene-acetonitrile mixture by volume was run. The reasons for this standard were two-fold. The first reason was to calibrate the spectrograph. Many of the measure- ments done were very sensitive and resolution on the order of 1-5 cm-1 was desired. Therefore, the software needed to be calibrated in wavenumber space instead of wavelength space. This was done in an external data analysis software package. Frequency is the inverse of wavelength so using known toluene-acetonitrile peaks, a calibration polynomial was generated from a Taylor expansion of 1/x and applied to the experimental data to calibrate the to within 1 cm-1. The standard also served as a way to maximize the Raman signal prior to collecting data. This allowed for reproducible day-to-day signal collection. Figure 3.7 shows a sample spectrum of the toluene-acetonitrile standard. The development of this system and experimental procedures has been a major component of the work done during the course of graduate study. Table 3.2 shows the optical components used and Figure 3.8 describes how they are assembled. This system has shown quite a large range of versatility in its ability to measure the vibrational modes of multiple bioinorganic systems. Chapter 4 describes experiments

67 Figure 3.7: Example spectrum of a 50% v/v solution of toluene and acetonitrile used for the calibration of the spectrometer for resonance Raman scattering.

conducted on a unique photoinduced structural change of R2lox using this system and Chapters 6-8 demonstrate its use on bioinorganic systems outside of our research group.

68 69

Figure 3.8: The figure above shows the Resonance Raman experimental system described in this chapter. This is atop-down view, while the insets are viewed from the side. The numbers on each optic correspond to Table (ref) which contains their relevant information. Table 3.2: Optical elements used in resonance Raman experimental system

Number Optical element Manufacturer Part number 1 Pellin-Broca prism Thorlabs ADBU-20 2 25.4 mm UVFS plano-convex spherical lens Newport Corporation SPX014 3 6mm x 6mm x 2mm BBO crystal Eksma Optics BBO-605H 4 25.4 mm UVFS plano-convex spherical lens Newport Corporation SPX014 5 25.4 mm Plano-convex cylindrical lens Thorlabs LJ4667RM 6 Pellin-Broca prism Thorlabs ADBU-20 7 1” Inherited long pass dichroic mirror N/A N/A 8 25.4 mm UV-enhanced aluminum mirror Thorlabs PF10-03-F01 70 9 25.4 mm UV-enhanced aluminum mirror Thorlabs PF10-03-F01 10 25.4 mm UV-enhanced aluminum mirror Thorlabs PF10-03-F01 11 25.4 mm UV-enhanced aluminum mirror Thorlabs PF10-03-F01 12 25.4 mm UV-enhanced aluminum mirror Thorlabs PF10-03-F01 13 25.4 mm UVFS plano-convex spherical lens Thorlabs LA4380 14 25.4 mm UV-enhanced aluminum mirror Thorlabs PF10-03-F01 15 25.4 mm UV-enhanced aluminum mirror Thorlabs PF10-03-F01 16 25.4 mm UV-enhanced aluminum mirror Thorlabs PF10-03-F01 17 12.7 mm UVFS plano-convex spherical lens Thorlabs LA4327 18 25.4 mm UV-enhanced aluminum mirror Thorlabs PF10-03-F01 19 50.8 mm UVFS Aspheric lens Edmund Optics 67-266 20 50.8 mm UVFS plano-convex spherical lens CVI Optics PLCX-50.8-103.0-UV Chapter 4 Driving protein structure changes with light

4.1 Introduction

The high levels of complexity seen in nature can in part be attributed to the advent of natural systems that harness solar energy to drive biological processes. Light-driven proteins are responsible for many of the key reactions underlying multicellular life on this planet. These systems include the light-harvesting complexes of green plants and cyanobacteria that drive photosynthesis, UV-resistance proteins that shield DNA from photodamage, light-sensing receptors such as rhodopsin and G-protein coupled ion channels, FAD-containing BLUF domains to drive DNA repair and circadian rhythms, and photoactive yellow protein (PYP) to control phototactic responses, among others.26,221,274–281 The light-responsive components of these proteins are typi- cally made up of organic cofactors, though the active sites of metalloproteins can also be driven by light. One example of this is seen in the generation of the Ni-L state, now recognized as important for catalysis, in the [NiFe] hydrogenase.174 However, photochemical efficiencies are typically low for biologically relevant first-row transi- tion metals due to rapid deactivation of the excited state.282–286 These limitations also hamper use of these earth-abundant elements for solar energy applications.287–289 Thus, identifying molecular mechanisms by which the photochemical efficiencies of

71 first-row transition metals can be increased is relevant both to understanding biolog- ical systems that feature these cofactors as well as to alternative energy research. Two examples of efficient, photoinduced reactions that utilize a first-row transition metal are the photo-Fenton and photodecarboxylation reactions.290–292 The former process relies on an FeIII-OH moiety, either in solution or as part of a nanoparti- cle assembly (e.g., rust), that is homolytically cleaved upon photoexcitation to form FeII and a hydroxyl radical (OH•). The hydroxyl radical has a very high reduction potential (∼1.9 V vs NHE) and can oxidize a wide variety of organic compounds.293 Photodecarboxylation reactions between an FeIII-carboxylate moiety are also observed to occur in small model complexes as well as siderophores, with relevance for iron re- lease and cycling in marine environments.294–298 Both sets of reactions are driven by violet (400–430 nm) and UV (250–400 nm) light with quantum efficiencies that can approach values as high as 3%, depending on the reaction conditions.292,294 Due to the low costs and minimal toxicity of iron along with the potential to initiate this chem- istry using sunlight, these processes are under investigation for use in environmental remediation processes.299 However, early reports indicate that substrate specificity and product selectivity remain low.292,300 As such, the ability to access this chemistry in a controlled fashion is of great interest. Biological systems are known to impart high levels of selectivity due to the precise primary and secondary sphere control present within the active site of an enzyme. For example, the cytochrome P450 family has the ability to catalyze a vast array of selective oxidation reactions, each dependent on a specific substrate and protein- derived redox partner.301 In nature, these reactions are thermally or chemically ac- tivated and thus lack the temporal and spatial control that a light-driven process would confer. Proteins are biocompatible and function in aqueous solutions, making them the ultimate “green” catalyst. Combining the favorable aspects of photo-Fenton

72 or photodecarboxylation chemistry with the advantages of a protein scaffold would be valuable. However, there are few proteins with the appropriate electronic and geometric configuration to carry out these light-driven processes. The R2-like ligand-binding oxidases (R2lox) represent a new class of metallopro- tein that, in the presence of manganese, spontaneously assemble a heterobimetallic Mn/Fe active site.130,132,136 In the absence of Mn, R2lox can also bind two Fe ions to generate a standard diiron active site, potentially rendering the protein cambialis- tic.135 This protein is a member of the large ferritin-like superfamily of proteins, with a mixed carboxylate-histidine bimetallic active site. More specifically, R2lox shares structural similarities both with the R2 subunit of class 1c ribonucleotide reductases (RNR), a heterobimetallic Mn/Fe enzyme that acts as a one-electron oxidant to initi- ate DNA synthesis, as well as the bacterial multicomponent monooxygenases, which perform two-electron oxidation reactions on a diverse array of substrates. The in vivo function of R2lox remains unclear, as the protein does not show RNR activity and instead generates a unique intramolecular tyrosine-valine crosslink between Y162 and V72. This crosslink forms only upon metal binding and oxygen activation and reflects a two-electron oxidation of the β-carbon of V72. The role of this crosslink has been postulated to be structural, poising the active site in the proper configuration for subsequent catalysis, though this remains under investigation. Following assembly and activation, R2lox adopts a MnIII/FeIII resting state with two carboxylate bridges and a single “hydroxo” (µ-OH) ligand between the two metals (Figure 4.1). This protonation state has been characterized using EPR and X-ray spectroscopy, as it is unique among this class of proteins; related diiron carboxylate proteins typ- ically feature at least one deprotonated “oxo” (µ-O) ligand between the two met- als.21,132,136 This difference may be relevant to the chemistry performed by thepro- tein, as the protonation state of the bridging ligand plays a key role in tuning the

73 reduction potential of the active site.302 Further, this proton may also participate in photoinduced chemical processes, as the metal cluster now closely resembles those implicated in iron-mediated photo-Fenton and photodecarboxylation reactions. In fact, an earlier study revealed cryogenic, visible-light photolysis of R2lox generates a species that exhibits EPR spectral parameters consistent with intramolecular depro- tonation of the hydroxide bridge.136 Upon warming, the resting state is recovered.

Figure 4.1: The crystal structure of Mn/Fe R2lox (PDB: 4HR0) The inset shows the metal center of R2lox and highlights several of its unique structural characteristics.

In this work, we have expanded upon these initial observations. We find that photoexcitation of aerobically assembled R2lox at room temperature generates a new species with distinct spectral properties. This state features a broad in the visible region of the spectrum and resonance Raman spectra with distinct high frequency modes from 1000–1600 cm-1. The optical and vibrational features of this new species are characteristic of proteins with direct metal-tyrosinate coordination, such as the purple acid phosphatases.303 Here, we present multiple lines of spec-

74 troscopic and photochemical evidence for photoinduced rearrangement of the R2lox active site and propose a reaction mechanism that resembles the photodecarboxyla- tion processes found in siderophores. Understanding the parameters governing the efficiency of this transformation has implications for developing molecular systems containing first-row transition metals that can effectively participate in controlled chemical reactions.

4.2 Methods & Materials

All materials were obtained from Fisher Scientific or VWR unless otherwise stated. Optical elements were purchased from Newport Corporation or Thorlabs unless oth- erwise noted. All solutions were prepared using deionized water (ELGA Flex II).

4.2.1 Protein production and purification

R2lox was produced and purified as previously reported.132,137 Briefly, a 300 mL culture of E. coli strain DE3* (New England Biolabs, Cambridge, MA), containing the plasmid encoding the Geobacillus kaustophilus R2lox I gene was shaken at 200 rpm in TB media (Formedium, Norfolk, United Kingdom) at 37 °C for 15 hours. The cells were then diluted into 1 L of fresh culture and shaken at 200 rpm at 37 °C until reaching an OD of ∼1.0. Cell growth was arrested by storage at 4 °C for up to 6 hours. Prior to induction, an aliquot of EDTA at pH 7.0 was added to the cell culture to a final concentration of 0.5 mM and shaken for 10 min to eliminate any free metals from the culture media. Protein expression was induced by addition of an aliquot of IPTG (GoldBio, St. Louis, MO) to a final concentration of 0.5 mM to the cell culture. Protein expression was allowed to continue for 15 h at 18 °C while shaking at 200 rpm. Finally, cells were collected by centrifugation at 5900 x g and 4 °C. Pellets were frozen at -80 °C until purification of R2lox.

75 Cells were lysed by sonication on ice for 2 min in 15 sec increments followed by 1.75 min rest periods at a power level of 25% (Model Q700, Qsonica, Newtown, CT). The lysate was then centrifuged at 39,000 x g for 30 min followed by a 10 minute heat treatment at 60 °C and a subsequent centrifugation. The lysate was then loaded in 5-10 mL aliquots onto 5 mL Ni-NTA columns (McLab) equilibrated with a solution containing 25 mM HEPES, 20 mM imidazole, 300 mM NaCl, and 0.5 mM EDTA at pH 7.0. The column was first washed with 25 mL of a solution containing 25mM HEPES, 40 mM imidazole, 300 mM NaCl, and 0.5 mM EDTA at pH 7.0 and 50 mL of a solution containing 25 mM HEPES, 40 mM imidazole, and 300 mM NaCl at pH 7.0 to remove weakly bound proteins. Pure apo-R2lox was eluted with 25 mL of a solution containing 25 mM HEPES, 300 mM NaCl, and 250 mM imidazole at pH 7.0. Samples were buffer-exchanged into 25 mM HEPES, 50 mM NaCl at pH 7.0using a stirred cell concentrator equipped with a 3 kDa MWCO filter (Amicon, Millipore Sigma, Billerica, MA). The protein was concentrated to a final concentration of ∼1 mM, flash-frozen in liquid nitrogen, and stored in aliquots at -80 °C until use.Protein concentrations were determined using previously published extinction coefficients (ϵ280 = 47757 M-1cm-1).136 Protein purity was verified using SDS-PAGE. Point mutations were introduced into the plasmid encoding full-length Geobacillus kaustophilus R2loxI by site-directed mutagenesis using the QuikChange Lightning kit (Agilent) and verified by DNA sequencing. All point mutants were produced and purified in metal-free form according to the same protocol as wild-type R2lox.

4.2.2 Sample preparation

To prepare metallated samples, apoR2lox was diluted to a concentration of 0.1 mM in 25 mM HEPES, 50 mM NaCl at pH 7.0. Protein was then rapidly mixed with either 1 equivalent each of MnII and FeII or 2 equivalents of FeII, both prepared in

76 deionized water, to generate Mn/Fe or Fe/Fe R2lox, respectively. Metallation was monitored using optical spectroscopy over the timescale of 1 hour. Deuterated samples

were prepared in buffers made with D2O (99.9%, Cambridge Isotope Laboratories, Cambridge, MA). Protein samples were allowed to equilibrate in deuterated buffers at 4 °C for 16 hrs prior to metallation to promote complete exchange of protons. After metallation, holoR2lox was run through a HiTrap desalting column (GE Amersham, Stockholm, SE) to separate the protein from any unbound metal ions. R2lox was photoconverted using the 457.9 nm line (P = 30 mW) on a Kr/Ar mixed gas laser (Coherent Innova Spectrum I-70C).

4.2.3 Iron quantification assay

The amount of iron bound was quantified using 2,2-bipyridine as a colorimetric indi- cator.304 To denature the protein and reduce the bound metals, R2lox was shaken at 200 rpm in solution with 143 mM sodium dithionite and 0.107 mM of 2,2-bipyridine at 37 °C for 90 min. The mixture was centrifuged at 17,900 x g for 30 min to remove pre- cipitate, and the supernatant was analyzed by UV/vis . The concentration of iron was determined by monitoring the absorbance of [FeII(bpy)3]2+

-1 -1 305 at 522 nm (ϵ522=8650 M cm ). Metallation percentages were determined by com- parison to the initial protein concentration.

4.2.4 Resonance Raman spectroscopy

Metallated protein samples were concentrated to ∼1 mM and placed in flame-sealed 0.8 mm ID borosilicate capillaries (Kimble). Laser excitation wavelengths of 457.9 nm, 488.0 nm, 514.5 nm, and 568.2 nm were generated by a Kr/Ar laser (Coherent Innova Spectrum 70-C). The appropriate interference filters (Semrock, MaxLine, Rochester, NY) were used to remove plasma lines. The resonance Raman setup was similar to

77 that described previously.306 Briefly, the beam was focused onto the sample at a135° backscattering angle using a f/4 parabolic focusing mirror. The scattered light was collimated by an f/0.8 UV-fused silica aspheric lens (Edmund Optics, Barrington, NJ), focused onto the 100 µm slit of a f/4.6 single-grating spectrograph equipped with a 1200 gr/mm grating blazed at 500 nm (Princeton Instruments Isoplane 320), and imaged onto a Peltier-cooled CCD (Princeton Instruments Pixis 100B). Rayleigh scattering was rejected by the appropriate edge filter (Semrock RazorEdge). A 50/50 (%v/v) mixture of toluene and acetonitrile was used to calibrate the spectrograph to within 1 cm-1. Relative resonance Raman enhancement profiles were generated by normalizing all intensities to the non-resonant phenylalanine peak at 1005 cm-1. Spectra were exported and analyzed in Igor Pro 6 (Wavemetrics, Portland, OR).

4.2.5 Quantitative photochemical experiments

Photochemical action spectra were obtained with 200 µL samples of 20 ￿M Mn/Fe and Fe/Fe R2lox within 10 mm cuvettes (SF-Q-10, Starna Cells Inc, Atascadero, CA). The samples were irradiated for a fixed amount of time within a Fluoromax-4 fluorometer (JY Horiba, Kyoto, Japan) at 25 °C using a bandwidth of 10nm.Mea- surements were made in 40-nm increments from 300 to 700 nm. Absorbance spectra were measured before and after irradiation. Fresh samples were used for each wave- length. The difference in absorption at 550 nm (Fe/Fe R2lox) or 540 nm(Mn/Fe R2lox) due to photolysis was normalized to the integrated lamp intensity over the irradiation range at each excitation wavelength. The extinction coefficient of the pho- toconverted product was obtained by irradiating the sample with 402 nm light until the absorption features of the photoproduct were no longer changing. To determine the concentration of protein present, an iron quantification assay was performed on the photoproduct. For the Fe/Fe R2lox samples, homobimetallic cofactor formation

78 was assumed. From this assumption, 80  3% of the protein present was metallated. For Mn/Fe R2lox, the concentration of metallated cofactor present was estimated from the iron quantification assay by assuming a similar fraction of metals bound ( 80%) as the homobimetallic cofactor. A potassium ferrioxalate actinometer was used to quantify the flux of the fluorom- eter lamp for the quantum yield experiments. The actinometry protocol was adapted from references [307] and [308].307,308 Briefly, a solution of 0.15 M potassium ferriox- alate trihydrate (Strem Chemicals, Inc., Boston, MA) in 0.05 M H2SO4 was prepared along with a 0.2% solution of 1,10-phenanthroline (Sigma Aldrich) in 1.64 M sodium acetate and 0.5 M H2SO4. A 0.2 mL sample of the potassium ferrioxalate solution was irradiated in the fluorometer at 402 nm (3 nm bandwidth). Following irradia- tion, 0.4 mL of the phenanthroline solution was added to the potassium ferrioxalate solution. The solution was allowed to develop in the dark for 10 min at 25 °C, after which time the absorbance at 510 nm was measured (ϵ=1.1x104 M-1cm-1). The flux

II 2+ of the fluorometer lamp was determined based on the concentration of[Fe (phen)3] using equation 4.1.307

A ν ν I = 510 2 3 (4.1) ϵ510Φ402ν1dt The measurement was repeated in triplicate for 3 different irradiation times (10, 20, and 30 seconds). To quantify the photoconversion efficiencies, a sample of20 µM Fe/Fe or Mn/Fe R2lox was irradiated in the fluorometer at 402 nm (3 nm bandwidth) for an appropriate amount of time. The absorption spectrum was measured following irradiation, and the concentration of the photoproduct was found using the previously determined extinction coefficient. To calculate the quantum yield of the process, the rate of formation of photoconverted R2lox was divided by the photon flux and fraction of photons absorbed (eqs 4.2 and 4.3). New samples were used for each independent

79 measurement.

1 d[photoproduct] Φprotein = (4.2) fmI dt

− − (1 − 10 A402,i ) + (1 − 10 A402,f ) f = (4.3) m 2

4.3 Results and Discussion

4.3.1 Exposure to light drastically changes the optical spec- trum of R2lox

Upon UV and violet photoexcitation, a broad absorption band centered around 550 nm appears in the optical spectrum of R2lox (Figure 4.2). Similar spectra and timescales of conversion are seen for both the Fe/Fe and Mn/Fe cofactors. How- ever, the extinction coefficients of the new species are metal dependent, with the maximum values estimated at 2600 M-1cm-1 (550 nm) for Fe/Fe R2lox and only 1200 M-1cm-1 (540 nm) for Mn/Fe R2lox. These values were calculated assuming compara- ble levels of metallation (∼80%) for both cofactors and quantitative conversion after 4 hrs, which may underestimate the extinction coefficient. Metallated protein stored in the dark did not undergo this change (Figure 4.3). To quantify the wavelength dependence of this transformation, the action spec- trum was measured (Figure 4.2, right hand axis). Overall, the photoconversion ef- ficiencies as a function of wavelength follow the trace of the absorption spectrumof metallated R2lox, suggesting this process derives from the protein itself and is not simply an artifact of irradiation. Both Mn/Fe and Fe/Fe cofactors exhibit low effi- ciencies in the violet region of the spectrum (∼400 nm), likely due to excitation into the shoulder of the absorption feature at 465 nm, with higher conversion efficiencies

80 seen in the near-UV from 300−360 nm.

Figure 4.2: The figure above shows the UV/vis spectrum of Fe/Fe (top) andMn/Fe (bottom) R2lox before (gray) and after (black) irradiation. The points show the measured of each metallated variant. The inset shows the complete conversion of to the photoconverted species. The time constants for Fe/Fe and Mn/Fe are 55.3  1.3 min and 67.5  9.32 min respectively.

Irradiation wavelengths longer than 460 nm do not induce this transition in either cofactor. Using actinometry, the quantum yields for this process at 402 nm were found to be 2.3  0.090% for Fe/Fe R2lox and 3.1  0.26% for Mn/Fe R2lox. That these values are quite similar indicates that the metal in site 1 exerts only a minor effect on

81 this process, and thus the process is likely to be Fe-centered. The quantum efficiencies for conversion are relatively high for a protein-based photochemical process mediated by a first-row transition metal; for comparison, the Ni-C to Ni-L conversion, inwhich a proton is transferred from nickel to a sulfur thiolate, has an estimated quantum yield of 0.07%.309

Figure 4.3: The figure above shows Fe/Fe (top) and Mn/Fe (bottom) before (gray) and after (black) being left in the dark for 2 weeks.

82 4.3.2 Resonance Raman spectroscopy of photoconverted R2lox

Further characterization of the photoconverted R2lox samples was carried out using resonance Raman spectroscopy, with excitation into the new visible absorption band (Figure 4.4). Interestingly, the Fe/Fe R2lox samples exhibit intense vibrational bands at 1170 cm-1, 1290 cm-1, 1500 cm-1, and 1600 cm-1, which are much higher in energy than typically observed in non-heme diiron enzymes. Additional bands are observed at 590 cm-1 and 800 cm-1, with similar spectra obtained for both the Fe/Fe and Mn/Fe cofactors. To explore the degree of coupling between the active site vibrations and nearby proton motion, solvent isotope-dependent experiments were performed by exchanging R2lox into D2O prior to metallation. While only minor changes are seen in the high-frequency region of the spectrum, a pronounced shift can be resolved in the difference spectra of both photoconverted Mn/Fe and Fe/Fe R2lox for the bandat 586 cm-1, with differential peak shifts of +14-1 cm and +12 cm-1. Attempts to obtain resonance Raman spectra of R2lox prior to photoconversion were unsuccessful, as even low photon fluxes caused this change to occur; unfortunately, the amounts of sample that would be required for a true single-pass experiment are prohibitive. Resonance Raman excitation profiles were measured using multiple lines from a mixed-gas Kr/Ar laser across the visible absorption band to assess whether the observed vibrations were coupled to a single electronic transition.310 The spectra were normalized to the non-resonant phenylalanine peak at 1004 cm-1 to correct for slight variations in protein concentration. The relative peak heights of the five most intense bands, as indicated on the spectrum, were plotted as a function of excitation wavelength. As seen in the Figure 4.4 inset, the relative intensities of all bands track with the optical transition, indicating that the resonance Raman spectrum derives from a single species with excitation into one electronic transition.

83 Figure 4.4: Resonance Raman spectra of photoconverted Fe/Fe (top) and Mn/Fe (bottom) R2lox (λex = 457.9 nm, Pex = 25 mW, 298 K) prepared in H2O (blue) and D2O (red). Shown in black are the difference spectra (H2O-D2O). (Inset) Resonance Raman excitation profiles for key vibrational modes overlaid with the absorption spectra of photoconverted Fe/Fe (left) and Mn/Fe (right) R2lox. The peaks repre- sented in the resonance Raman excitation profile are tagged in the rR spectra with the respective symbols.

84 4.3.3 Spectroscopic results suggest R2lox undergoes photoin- duced rearrangement to yield tyrosinate coordinations

The spectra of photoconverted R2lox bear close resemblance to the purple acid phos- phatases (PAPs) and related metal-tyrosinate proteins and model compounds.303,311 These species feature phenolate-to-FeIII LMCT bands centered around 550 nm, which confer the characteristic purple associated with these enzymes. The phenolate ligands in these systems are typically bound at the FeIII in a terminal fashion. The striking similarity in optical spectra between the two cofactors suggests that the dominant ion contributing to the LMCT transition is the FeIII center in site 2. The observed variation in extinction coefficient may be attributed to differences in thesite 1 metal identity, though the various metallation states of PAPs (FeIIIFeIII, FeIIIZnII, and FeIIIMnII) do not show significant variation in their extinction coefficients (Table 1).123,312,313 Additionally, small-molecule MIII-phenolate compounds show an increase in transition energy and decrease in extinction coefficient whenIII Mn is used rather than FeIII for a given ligand.314,315These differences are unpronounced in the different metallation states of photoconverted R2lox.

Table 4.1: Extinction coefficients of purple acid phosphatases from various organisms

Species of Origin Metal centers ϵ (M-1cm-1) Sus scrofa (pig) FeIIIFeIII 3100312 Phaseolus vulgaris (red kidney bean) FeIIIZnII 3360313 Ipomoea batatas (sweet potato) FeIIIMnII 3207123

Additional support for this interpretation is found in the resonance Raman spectra of photoconverted R2lox, which share a number of similarities with those of purple acid phosphatases and metal-phenolate complexes. Typical RR bands of purple acid

85 phosphatase include a low frequency mode between 580-600 cm-1 that is dominated by iron-oxygen stretching motions. Additionally, these spectra feature four distinctive high-frequency bands reflecting the in-plane tyrosine C-H bending motion (Y9a), C-O stretch (Y7a’), and ring-stretching modes (Y19a and Y8b) between 1150 cm-1 and 1600 cm-1.316 The frequencies of these bands do not change significantly upon deutera- tion, suggesting there are no exchangeable protons coupled into these vibrations. The notable change observed in the Fe-O stretching mode of the photoconverted Mn/Fe R2lox variant upon deuteration, mentioned above, suggests increased coupling of this

18 mode to nearby protons. R2lox was also incubated overnight in OH2 prior to photol- ysis to see if the exchangeable µ-OH ligand is coupled into the tyrosinate vibrations. However, no isotope-dependent bands are observed that would be consistent with a µ-OH bridge between the two metals (Figure 4.5). The peaks described above are absent in the apo-R2lox Raman spectrum (Figure 4.6).

4.3.4 Y162 is the coordinating tyrosinate ligand

As R2lox features the aforementioned tyrosine-valine crosslink near the active site, it was hypothesized that this moiety may be responsible for the light-induced transi- tion seen. Thus, photolysis experiments were also performed on variants possessing mutated crosslink residues, including the V72A, V72L, and Y162F proteins. These mutants are unlikely to form the ether crosslink upon metal binding and oxygen acti- vation due to decreased stability of the resultant alkyl radical that would be formed. However, both V72A and V72L undergo photoconversion upon irradiation with ef- ficiencies that are indistinguishable from the native protein (Table 4.2), suggesting crosslink formation has no impact on the photoconversion process. On the other hand, the Y162F mutant assembles to give a similar MnIII/FeIII resting state species (Figure 4.7) but does not generate a purple sample even after prolonged irradiation,

86 18 Figure 4.5: Resonance Raman spectra of OH2 exchange of Fe/Fe R2lox prior to photolysis at 25 mW of 514.5 nm excitation. The spectrum does not change, indicating that the bridging ligand is not involved in the tyrosinate-Fe vibrations.

Figure 4.6: Resonance Raman spectra of apoR2lox (black), photoconverted Fe/Fe R2lox (blue), and photoconverted Mn/Fe R2lox (red). The proposed tyrosinate peaks are not present in the apoR2lox spectrum.

87 suggesting that Y162 is the tyrosine residue that coordinates to the metal center in the photoconverted species. In the WT R2lox crystal structure, the Y162 phenolic oxygen is located only ∼5 Å from the Fe center.

Table 4.2: Quantum efficiencies for various R2lox mutants

WT V72A V72L Y175F Y162F Mn/Fe 3.1  0.26 3.4  1.3 3.6  0.18 0.19  0.019 0 Fe/Fe 2.3  0.90 2.9  0.38 2.7  0.21 0.19  0.021 0

We also considered whether multiple species could be contributing to the observed spectroscopic signals. Like the related RNR and BMM enzymes, R2lox contains a number of tyrosine residues near the active site.106,317,318 Tyr175, which is slightly farther away from the Fe center compared to Y162 and participates in hydrogen bonding interactions with the Mn-bound water ligand and/or the coordinating E202, was also explored as a candidate for photoinduced modification. Upon reconstitution with metals under aerobic conditions, the Y175F mutant develops an optical spectrum characteristic of the MnIII/FeIII oxidized protein, suggesting this residue is not critical for assembly. However, photoexcitation of Y175F with violet or near-UV light results in a purple product that is spectrally indistinguishable from photoconverted WT R2lox (Figure 4.7), with comparable yields. Thus, Y175 seems an unlikely suspect for the metal-binding residue, consistent with the greater distance between the tyrosinate oxygen and the iron center. Additionally, quantum yield analysis of the Y175F mutant showed a dramatic decrease ( 15x) in the efficiency of the purple product formation, indicating that Y175 plays a role in this process (Table 4.2).

88 Figure 4.7: UV/vis absoprtion spectra of Y162F (left) and Y175F (right). The gray traces are prior to irradiation and the black traces are after light exposure. The top set of spectra in each plot are Fe/Fe and the bottom set are Mn/Fe.

4.3.5 Mechanism of photoconversion process in R2lox

The above experiments are accompanied by a suite of crystal structures and tan- dem mass spectrometry experiments showing that upon light exposure, E69 is de- carboxylated. Furthermore, EPR on both dark and light exposed Mn/Fe R2lox a lack of change between the initial and final state, indicating that the final prod- uct is a MIII/FeIII state with a similar primary coordination sphere. These experi- ments were not my work and thus are not included in this chapter. Synthesizing the above data, Figure 4.8 shows a proposed mechanism for the light-induced conversion, which resembles those invoked for photoinduced decarboxylation reactions of FeIII- OOCR species. These processes have been identified inIII Fe -carboxylate moieties in siderophores, playing an important role in mediating iron release from chelating compounds as well as in model compounds and atmospheric processes.295,298 Briefly, LMCT excitation of the metal-carboxylate species in the resting state of R2lox gener- ates an FeII-carboxylate radical species. This would occur due to the unique electronic properties of the hydroxyl-bridged FeIII. When this carboxylate radical rearranges, 89 Figure 4.8: A proposed mechanism of the conversion process that R2lox undergoes when exposed to light.

CO2 is lost, which results in the conversion of the glutamate to homoalanine (alanine with a methyl group). The regeneration of the sp3 carbon on the γ-carbon of E69, now homoalanine, requires a H+ and e- from a nearby source, which in this case is the metal center, forming a MIII/µ-O/FeIV complex. The experiments show that homoalanine is formed and not vinyl-glycine, thus the necessity for an extra H+ and e- rather than a dehydrogenation of the radical species. The pentacoordinate FeIII species then promotes the homolysis of the Tyr-Val crosslink and forms a bond with the newly formed tyrosyl radical, forming a FeIV-tyrosinate in the process. Here the mechanism bifurcates, with the MIII/FeIV center resonating to MIV/FeIII, which a preference for the latter state. In the former state, the teritary radical on V72 donates a H+ and e- to the MIII/FeIV center, regenerating the resting state MIII/µ-OH/FeIII center and forming a double bond between the α- and β-carbon on V72. Meanwhile, the favored resonance state (MIV/FeIII) proceeds to abstract a H+ (via the aquo ligand on the site 1 metal) and e- from nearby Y175, which abstracts a H+ and e- from a

90 nearby residue (most likely Y176) which propagates the hole out to the solvent. thus regenerating the resting state. Additionally, Y175 itself could also take a H+ and e- from water due to its partial solvent exposure. This would explain the sharp decrease in the quantum yield of photoconversion in the Y175F mutant. As mentioned before, the final MIII/FeIII-tyrosinate species is the source of the purple color that is formed. The exceptionally high quantum yields seen for this process are also reminiscent of those previously noted for photodriven decarboxylation of FeIII-EDTA and FeIII-NTA compounds.292,319,320 This hypothesis is consistent with the observations made in this study and rep- resents the first example of light-driven cofactor rearrangement in a member ofthe ferritin-like superfamily. R2lox is relatively unique among this class of proteins in that the bridging oxygen-derived ligand in the resting state of the protein is pro- tonated; most well-characterized members of this family feature deprotonated µ-O ligands.70,302,321 This proton may be critical for the photochemistry observed, localiz- ing the LMCT excited states onto the carboxylate ligand rather than the bridging oxo moiety.322 This process resembles that seen in iron-bound siderophores, in which the carboxylate-centered excitation drives the decarboxylation process. Another unique aspect of this reaction is that a single protein species, as characterized crystallo- graphically and through mass spectrometry, is formed after excitation, suggesting controlled reactivity that is localized to only one of the carboxylate ligands.300,323–325 Photoinduced iron-mediated decarboxylation reactions have also been used to gener- ate new carbon-carbon bonds through radical coupling processes, an underexplored area of research for enzymatic catalysis. These initial studies on R2lox may provide a foundation for engineering metalloproteins to perform controlled light-activated de- carboxylation reactions, which may ultimately lead to formation of unique chemical species.

91 4.4 Conclusion

It has been shown that R2lox, a member of the ferritin-like superfamily capable of binding both an Fe/Fe and a heterobimetallic Mn/Fe cofactor, undergoes a photoin- duced transition to generate a new state of the protein. The optical and resonance Raman spectra of this distinct species suggest the rearranged active site features a MnIII/FeIII or FeIII/FeIII cofactor with a bound phenolate ligand. Photochemical in- vestigations reveal that the photoconversion is initiated by UV and violet light with remarkably high efficiency for an Fe-mediated process. A mechanism consistent with these data as well as collaborator experiments has been proposed. Further studies are underway to probe the structure, mechanism, and potential utility of this photo- conversion process.

92 Chapter 5 Probing electron transfer kinetics in R2lox via RuII-labeled Cys

5.1 Introduction

Electron transfer (ET) is an integral process for the maintenance and continuation of life. Some of the most crucial reactions on Earth require electron transfer, such as photosynthesis and cellular respiration, the two reactions by which all multicellular organisms on earth ultimately extract energy from raw materials.32,326,327 Addition- ally, many of the difficult reactions that are performed by nature, including thetwo mentioned above, reactions that humanity wishes to mimic due to the outstanding efficiency and minimal environmental impact that they offer, require electron transfer to occur.328–330 Due to the central place electron transfer holds in the natural world, it has been the subject of much study over the past century.331 There are two types of electron transfer in molecular systems: inner sphere and outer sphere electron trans- fer. Inner sphere electron transfer is the transfer of an electron between two covalently linked groups.332,333 A simple model of inner sphere electron transfer is shown in Fig- ure 5.1. The first step is the collision of the donor and the acceptor to form acomplex, with a strong interaction between the donor and acceptor molecules. Next, the elec-

93 Figure 5.1: The figure above shows the steps for a single-step inner sphere electron transfer from a donor (D) and an acceptor (A).

tron is transferred from the donor to the acceptor, and finally the complex dissociates, except now with a charge separation caused by the electron transfer. A classic exam-

2+ 2+ ple of inner sphere ET is the reduction of [Co(NH3)5Cl] by [Cr(H2O)6] , where the transfer of the electron is mediated by a bridging chloride anion.334 The products of

2+ 2+ II this reaction are [Co(NH3)5H2O] and [Cr(H2O)5Cl] , where the Co has been con- verted into CoIII. Inner sphere electron transfer is inhibited by large bulky groups, which prevents the formation of the strong covalent interaction required for inner sphere electron transfer.335 Thus, in biological systems, outer sphere electron transfer is favored, which is the transfer of an electron between two chemical groups that are not directly bonded together.336 In outer sphere electron transfer, the electron tunnels to traverse the distance between the donor and the acceptor.318,337 Outer sphere electron transfer rates can be described semi-classically by Marcus theory, shown in Equation 5.1.106,333

( ) 3 (∆G0+λ)2 4π 2 − 4λkB T kET = 2 HABe (5.1) h λkBT

Here, kET is the rate of electron transfer, h is Planck’s constant, kB is the Boltz- mann constant, λ is the reorganization energy, ∆G0 is the free energy change of the reaction, and HAB is the electronic coupling strength. The theory must be semi- classical because electrons are quantum mechanical particles and cannot be described by pure classical theory. Marcus theory is also noteworthy for combining thermo-

0 dynamics (∆G ) with kinetics (kET). The parameters of the equation above can be

94 Figure 5.2: Intersecting parabolas showing the reorganization energy and free energy change of a nonadiabatic electron transfer reaction.

visualized in Figure 5.2. The vertical axis of the figure represents the energy of the system, and the horizontal axis represents the system coordinate.331 This is different from a typical reaction coordinate, where one degree of freedom is allowed to remain unconstrained while the rest are fixed in place. The system coordinate here represents the total state of the system due to the solvent dependence of the electron transfer process.338,339 The free energy of the reaction, which is the difference in energy of the donor and acceptor in their respective equilibrium geometries, is represented by ∆G0. For

95 the reaction to proceed spontaneously, ∆G0 must be less than 0. The reorganization energy (λ) is defined as the energy difference between the donor and acceptor atthe donor equilibrium geometry.339 Reorganization energy is a very important factor in the rate of electron transfer because it effectively measures the degree of geometric change a system must undergo to accommodate the electron transfer event. As seen from Equation 5.1, when ∆G0=λ in magnitude, the electron transfer rate is solely de-

pendent on HAB, the coupling energy between the two states, which can be described in Equation 5.2.340

− 1 − 0 0 2 β(r r ) HAB(r) = HABe (5.2)

0 0 Here, HAB is the electronic coupling at the close distance r . Equation 5.1 holds

true in the nonadiabatic limit, which is when HAB has a small value when compared with the reorganization energy.340 In the nonadiabatic limit, the donor and the ac- ceptor mostly retain their identity. Figure 5.2 shows the energy surfaces under the

nonadiabatic limit. However, when the magnitude of HAB is large when compared to

λ, then the adiabatic limit is reached. The value of HAB in protein systems is typically significantly less than λ, and so electron transfer in proteins is in the nonadiabatic limit and Equations 5.1 and 5.2 hold true. By combining Equations 5.1 and 5.2, it is seen that electron transfer rate exhibits an exponential dependence on distance between the donor and the acceptor. The parameter β is the distance decay constant and represents how fast the rate decays as a function of distance. Large values of β correlate to lower transmission probability through the potential barrier. Studies have shown that β takes on values of 0.8−1.2 Å-1 in donor-acceptor complexes as well as in a rigid matrix of donors and accep- tors.341–343 Proteins have been found to exhibit a rather uniform β of 1.4 Å, which is similar to a frozen organic glass.344 Many biological applications of electron transfer

96 Figure 5.3: The above scheme shows a 2-step electron transfer reaction between a donor and two acceptors.

must occur with the donor and acceptor very far apart. To circumvent this issue, long-range electron transfer in proteins is a multi-step process.317 Consider the simple system presented in Figure 5.3. The donor transfers an electron to acceptor 1. Accep- tor 1 will in turn transfer an electron to acceptor 2. In addition to this process being viewed as a movement of negative charge from left to right, it can also be thought of as the migration of a positive “hole” from right to left. By invoking multi-step electron transfer, biological systems can shuttle electrons across their entirety. Due to the complex environment inside a protein matrix, a new model is needed to analyze the barrier that electrons experience during the transfer process. The left side of Figure 5.4 is adequate for a single step through a medium. However, in proteins, the electrons can be transferred via covalent bonds, hydrogen bonds, or through space.345 This model is known as the Pathways model, and it invokes a simplification of the medium that the electron travels through in a protein to make the calculation more tractable; the reality is much more complex. A more appropriate model is shown on the right side of Figure 5.4, where different barriers are multiplied together to arrive

97 Figure 5.4: The left side of the figure represents a single step electron transfer through a medium. The right side is a model of multi-step electron transfer, where the electron sees a distribution of barriers.

at the total barrier that the electron experiences while traveling from the donor to the acceptor. Equation 5.3 shows the functional form of the heterogeneous barrier that an electron will experience inside a protein environment.346,347

∏NC ∏NH ∏NS ∼ C H S HAB ϵi ϵj ϵk i=2 j=1 k=1

C ϵ = 0.6 (5.3) ( ) 2 ϵH = ϵC e−1.7(r−2.8) 1 ϵS = ϵC e−1.7(r−1.4) 2 There are many chemical groups located in protein scaffolds that are capable of electron transfer. One such group are the iron-sulfur clusters, which will be discussed further in Chapter 6. Amino acid side chains can also participate in multi-step elec- tron transfer in proteins.337,348,349 The most common are tryptophan and tyrosine, however, cysteine, methionine, and histidine have also been implicated in biological electron transfer.37,106,350,351 Electron transfer rates can occur on a wide range of time scales.331 To probe this

98 Figure 5.5: A modified Latimer diagram that details laser flash experiments using the natural lifetime of the RuII excited state.

phenomenon, laser pulse-triggered RuII labels are used340,345,352 Figure 5.5 shows a diagram that outlines how this process is used in the laboratory setting. The resting state RuII complex is excited by a laser pulse, exciting it to a *RuII state. The excitation of the RuII complex alters the redox chemistry, allowing it to act as a donor or acceptor in an electron transfer reaction. The emission of the excited RuII complex can be used to monitor the transition between the excited and ground states, and acts as a probe of electron transfer kinet- ics. This method, however, is only viable when the lifetime of the RuII state directly overlaps with the electron transfer rate of the system.352 Previous studies have mea- sured the lifetimes of RuII-modified proteins and found them to be on the order of 102−103 ns.352,353 For electron transfer rates much slower than this, a reductive (or oxidative) intermolecular quencher is used (Figure 5.6). The quencher generates the same ruthenium intermediates as the example before, but on much longer timescales, allowing for the measurement of electron transfer in the 10 µs−sec time domain.352 A combination of these two experimental procedures allows for a very wide range of accessible time domains for electron transfer measurements.352

99 Figure 5.6: Diagram showing the excitation of Ru(II) and the subsequent redox pro- cesses that ensue via intermolecular quenching. The left side of the diagram shows reductive quechning, whereas the right side is oxidative quenching.

To extract the parameters in Equations 5.1 and 5.2, several experimental parame- ters must be adjusted. By adjusting the distance between the donor and the acceptor, the distance decay parameter (β) can be calculated. Different RuII complexes have different redox potentials. Thus, the driving force can be varied by attaching different RuII compounds to the protein and measuring the electron transfer rate. This allows for the determination of the reorganization energy (λ). Many proteins in the ferritin-like superfamily rely on electron transfer to perform their chemical function. To date, no studies have been conducted on RuII-modified non-heme diiron proteins to measure their electron transfer rates. Developing a sys- tem by which electron transfer rates may be measured in R2lox via a pumped flash- quench experiment would be very beneficial for future research on electron transfer pathways in non-heme iron proteins. Additionally, there is a dearth of information available for electron transfer in Mn/Fe systems, making these experiments important in other ways as well. This chapter outlines the identification of appropriate placementII ofRu complexes

100 using the Pathways model presented above, the preparation of each RuII-labeled mu- tant protein, and initial experiments on the RuII-modified complex. Here, the ground- work is laid for the study of the electron transfer through the protein matrix of R2lox.

5.2 Materials & Methods

5.2.1 Synthesis of RuII complex

The synthesis of the Ru(II) complex was prepared as reported in the literature.354,355 Briefly, 0.028 g of 5,6-epoxy-5,6-dihydro-[1,10] phenanthroline was mixed with 0.06g of cis-dichlorobis(2,2-bipyridyl)Ru(II) in 10 mL of a 75/25 v/v EtOH/water solution and refluxed in a foil-wrapped pressure vial at 100 °C for 3 hrs, where thereaction changed from purple to orange. After letting it cool, the EtOH was evaporated by

a stream of N2. Next, the solution was cooled on ice and then crashed out with a saturated solution of NH4PF6 in water. The precipitate was vacuum filtered and allowed to dry overnight in the dark. The product was confirmed via ESI with am/z of 305.41. Figure 5.7 shows the molecular structure of the product.

5.2.2 Identification and preparation of R2lox Cys mutants

This compound above has been reported to specifically label at cysteine residues via the epoxy-phenanthroline ligand.355 Wild-type R2lox does not have any cysteines present in its structure, making the wild-type protein a good template on which to build cysteine mutants for RuII labeling. Mutants were identified by using the Pathways plugin (version 1.2) in VMD (version 1.9.3) on a Linux machine running Ubuntu 16.04.356 The 4HR0 crystal structure was used to model R2lox. Since RuII is the ultimate electron donor and it will be attached via a cysteine in vitro, the donor in the program was set to the closest carbon to the cysteine (either the γ-carbon, β-carbon (Ala mutants), or the α-carbon (Gly mutants)). Ala and Gly mutants

101 Figure 5.7: Molecular structure of the RuII label used in this work

were adjusted by a factor of 0.6 and 0.36, respectively, to account for the extra carbon, as 0.6 is the decay factor for traveling through a covalent bond. The acceptor was set as either the Mn or the Fe, and the results were averaged to get the final score that was used to identify the proper Cys placement. Here, the 4AZU crystal structure of azurin was used as a benchmark to effectively compare coupling strengths because azurin is a very well characterized system for electron transfer.352 Mutants that were within an order of magnitude of the azurin value were deemed to have good coupling, whereas mutants that were greater than 3 orders of magnitude less had poor coupling. A variety of mutation loci were identified and prepared using site-directed mutagenesis. Primers were designed using SerialCloner 2.6 and bought from Sigma-Aldrich. PCR was done using an OpenPCR thermocycler. Afterwards, DPN1 was used to eliminate leftover template DNA from the mix. The PCR product was then transformed into DH5￿ cells and the mutation was confirmed via sequencing (Genewiz). The sequence-confirmed plasmid was then transformed into DH5￿ and BL21-DE3* cells and cell stocks were prepared. The purification of each mutant was

102 done according to established protocols for R2lox discussed in section 4.2.1.137

5.2.3 Preparation of labeled protein samples

ApoR2lox was thawed and incubated with 5x DTT for 10 min. Then the protein was passed through a HiTrap desalting column to remove the DTT. The apoR2lox was diluted to 25 µM and incubated with either 50 µM (A67C, D122C, G230C) or 100 µM (L74C) RuII label for ∼45 hrs at 37 °C in 50 mM Tris pH 8.5, 50 mM NaCl, 1% DMF in volumes ranging from 5-40 mL. After the reaction was complete, the solution was spun down at 9,800 xg for 20 min to remove any precipitated protein from solution. The supernatant was then concentrated down to ∼1 mL via spin centrifugation and the labeled protein was separated from the excess label by a PD-10 size exclusion column equilibrated with 25 mM HEPES, 50 mM NaCl, pH 7. The purity of the mixture was then confirmed by UV/vis spectroscopy using the system of equations in Equation 5.4. Efficiencies of 90-100% were attained.

5.2.4 Luminescence and TCSPC measurements

Samples were metallated as discussed in section 4.2.2.137 The metallated protein was then separated from excess metals using a HiTrap desalting column. The samples were prepared by dilution with nitrogen-sparged 25 mM HEPES pH 7 and 50 mM NaCl in a 1 cm fluorescence cuvette (Starna Cells). The absorbance at 451 nmwas between 0.05 and 0.1 for optimal measurements. The samples were blanketed with

II N2, stoppered, and sealed with parafilm to prevent oxygen from quenching theRu excited state. Luminescence was measured using a FluoroMax-4 (Horiba Scientific) using an excitation wavelength of 450 nm and measuring the emission intensity from 475 to 800 nm. Time-correlated single photon counting (TCSPC) traces were obtained using an EPL-445 laser source at 444.4 nm with a pulse width of 84.4 ps and a mini-τ

103 detector (Edinburgh Instruments). Data was imported into and visualized using Igor Pro 7. Lifetime traces were fit to a simple multi-exponential function to determine the decay constants.

5.3 Results and Discussion

5.3.1 Electron transfer calculations using the Pathways model

As discussed above, the Pathways model was used to computationally identify the best places to make mutations for the RuII label. Figure 5.8 shows the locations that were determined to be well exposed to solvent for optimal labeling as well as having good coupling to the metal center. As mentioned before, CuII azurin was used as a benchmark to determine what a good coupling constant would give in the Pathways

II program. The K122H model labeled with Ru (bpy)2(Im)(His) has a HAB value of 0.16 cm-1.357 When this model was put into the Pathways software the resulting net decay was 2.06 x 10-4 with a distance of 18.49 Å. Values on this order of magnitude and larger were considered good mutants to give favorable coupling between the RuII and the metal center. Table 5.1 shows the mutants that were made based on this analysis, in addition to the calculated electron transfer path distances and calculated net decay factor. Additionally, the change for each decay relative to the azurin K122H benchmark value is also given. The values here cover 6 orders of magnitude. The size of R2lox prohibits efficient coupling aside from positions that are relatively closeto the metal center. The positions that showed a high degree of coupling were only one or two amino acids away from the metal center. This does not preclude the possibility of electron transfer in R2lox, however, it does mean that the number of spots that are available for labeling is lower than in a smaller system. As seen from Figure 6, R2lox is a mostly α−helical protein. In addition to the small amount of space available for labeling, only 1 out of every 4 amino acids is solvent facing and will label efficiently.

104 Figure 5.8: Cysteine mutants identified by the Pathways program and made using site-directed mutagenesis techniques. The scale to the right is the log of the electronic coupling calculated by the Pathways program, where red is the lowest coupling and blue is the greatest. PDB:4HR0

105 Table 5.1: Electron coupling strengths for selected cysteine residues of R2lox

Relative log (Net decay Mutant Metal Distance (Å) 10 decay factor) factor Mn 18.17 -3.895 61.8% A67C Fe 22.82 -4.891 6.24% Mn 18.48 -4.175 9.36% L74C Fe 20.03 -5.050 4.33% Mn 22.75 -5.434 1.79% T96C Fe 22.77 -5.447 1.73% Mn 19.64 -3.946 55.0% H100C Fe 19.66 -3.996 49.0% Mn 30.96 -8.250 0.00273% D122C Fe 33.25 -9.128 0.000361% Mn 19.38 -5.380 2.02% D196C Fe 24.43 -5.863 0.665% Mn 14.83 -4.086 39.8% G203C Fe 12.31 -3.038 445% Mn 37.38 -8.951 0.000547% R269C Fe 35.27 -8.992 0.000494%

106 Figure 5.9: Reaction mechanism of the labeling of a Cys residue by the Ru label.

107 5.3.2 Labeling of R2lox with RuII

The structure of the RuII label is shown in Figure 5.7. The epoxide moiety on the 1,10-phenanthroline ligand is attacked by a thiolate generated from the deprotonation of the mutated cysteine. The product of the reaction is a mixture of the hydroxy- lated and ring-closed product (Figure 5.9). It has been shown that the addition of a mild base, such as potassium carbonate, can drive the reaction to the more ther- modynamically favorable aromatized species.355 Figure 5.10 shows a sample UV/vis absorption spectrum of RuIIA67C R2lox. The extinction coefficient at 280 nm of the Ru label was found to be 59,800 M-1cm-1, so the combined extinction coefficient of the Ru-R2lox system is ∼107,500 M-1cm-1. The literature value for the extinction coefficient at 451 nm of the Ru label is 13,300 M-1cm-1.355 Using these values, the ratio of Ru/R2lox can be calculated from the absorption spectrum via the following system of equations.

A280 = ϵ280,R2loxc280,R2loxl + ϵ280,Ruc280,Rul (5.4)

A451 = ϵ451,Ruc451,Rul

As mentioned above, there were anywhere from 0.90-1 Ru labels per R2lox molecule. Attempts made to separate any adventitiously bound RuII label did not yield any higher purity of labeled protein. After the protein was labeled with the RuII molecule, it was metallated. The metallation process was conducted exactly as described before in section 4.2.2 and in the literature.137 Buffer concentrations of 25 mM were enough to protect the protein from the decrease in pH upon the addition of FeII and MnII except for the D122C mutant, which required 100 mM HEPES to remain in solution. This possibly has to do with the fact that the other mutants that were labeled (A67C,L74C, and G203C)

108 Figure 5.10: .The figure above shows sample UV/vis spectraII ofRu -labeled R2lox (solid), unlabeled A67C R2lox (dashed), and free RuII label (dotted).

are all in positions that were either aliphatic or had no side chain, thus adding a +2 charge to the protein. However, since D122 is acidic, and deprotonated at pH 7 then a -1 charge is being replaced with a +2 charge, leading to a greater destabilization of the protein when the pH is dropped. Figure 5.11 shows the difference absorption of labeled Mn/Fe and Fe/Fe R2lox during the metallation.

5.3.3 Luminescence and TCSPC measurements

After the labeled R2lox was generated, several photophysical experiments were con- ducted to ascertain the effectiveness of this platform for electron transfer studies of R2lox. Luminescence spectroscopy was used to measure qualitative quenching of the RuII label when in a metallated system, as well as the energy gap between the triplet state and ground state. This can give insight into how the label is behaving on the protein scaffold with and withoutIII Fe and/or MnIII bound. Figure 5.12 shows the

109 Figure 5.11: .The figure above shows sample metallationsII ofRu -labeled A67C R2lox. On the left is Fe/Fe and the right is Mn/Fe.

luminescence spectra of apo, Fe/Fe, and Mn/Fe A67C, along with the apo and Fe/Fe luminescence spectra for L74C, D122C, and G203C mutants. Table 5.2 shows the emission maximums for each mutant. All of the mutants showed a decrease in the intensity of emissive radiation upon metal binding. Furthermore, when correlated with the traces in Figures 5.13 and 5.14, it can be seen that the mutants with the largest degree of quenching in the temporal domain also have the largest shift in their luminescence spectra. A67C in particular is interesting due to the emission maximum shifting in separate directions depending on the metal. The energy gap law states that a molecule that has a larger energy gap between its excited and ground states has a larger quantum yield of emission due the

110 Figure 5.12: Luminescence measurements of labeled R2lox mutants. The black traces are apoR2lox, the multiple red traces are Fe/Fe R2lox, and blue is Mn/Fe R2lox.

Table 5.2: Emission spectra data for RuII-labeled R2lox cysteine mutants

Luminescence Peak relative to Mutant Metal peak (nm) apo (nm) apo 606 0 A67C Fe 608 +2 Mn 604 -2 apo 612 0 L74C Fe 615 +3 apo 609 0 D122C Fe 608 -1 apo 608 0 G203C Fe 608 0

111 decrease in the denominator of equation 5.5,

k Φ = r (5.5) kr + knr

where kr is the rate constant for emissive decay, and knr is the rate constant for non-emissive decays.358,359 This is due to the increase in potential energy surface overlap as the energy gap decreases, leading to an increase in knr and thus an overall lower quantum yield of emission.360,361 The Fe/Fe variant follows the energy gap law, whereas the Mn/Fe variant does not follow the energy gap law. Along with A67C, the Fe/Fe variant of L74C follows the energy gap law. Neither D122C nor G203C exhibit much quenching. These observations correlate with the Pathways calculations. G203C, however, was expected have a higher degree of coupling compared to the other mutants due to its proximity to the metal center. In addition to luminescence, TCSPC experiments were also conducted on RuII- labeled R2lox. TCSPC stands for time-correlated single photon counting and is used to measure the lifetimes of light-emitting molecules. A laser pulse excites the sample in a periodic fashion, and snapshots of the emission are made along the temporal domain until a certain threshold of counts is achieved. The decay trace is an ex- ponential decay. Figures 5.13 and 5.14 show the TCSPC traces of various labeling locations and metallated forms of R2lox, and Table 5.3 tabulates the lifetimes from the various mutants. All of the traces here have lifetimes around 1 µs or less, which means that the sample was well degassed, and there was no triplet quenching from oxygen. Also according to the tabulated data, all of the metallated variants of R2lox have shorter lifetimes than the apo forms. This indicates that there is some sort of quenching that arises from the metal cofactors. Using the data from Table 5.1, one might expect G203C to be the highest-quenching mutant, and for D122C to have barely

112 Table 5.3: Luminescence lifetimes for RuII-labeled R2lox cysteine mutants

Mutant Metal τ1 (ns) τ2 (ns) τ3 (ns) apo 14.8  1.27 271  4.46 1071  8.53 A67C Fe 12.6  0.879 159  2.77 730  3.29 Mn 12.7  0.99 231  3.43 906  6.45 apo 13.2  1.42 224  4.4 802  4.62 L74C Fe 13.4  0.997 184  2.18 707  4.8 apo 16.7  2.18 170  4.39 1008  3.14 D122C Fe 9.29  1.03 176  3.96 871  3.15 apo 13.4  1.16 215  5.03 850  4.34 G203C Fe 8.50  0.77 149  3.58 731  2.4

any quenching at all. In fact, G203C seems to have the lowest amount of quenching in the metallated form, along with D122C, which is expected. This correlates well with luminescence data discussed above. Taking a closer look at D122C and G203C, τ1 seems to decrease by ∼35-45%, which is more than the L74C and A67C mutants. Also interestingly, the second lifetime does not change significantly in D122C, whereas in G203C, it does change. This trend flips in the third lifetime for these 2 mutants. An unexpected surprise was the degree of quenching observed in the L74C mutant. This mutant was expected to have intermediate coupling, on the order of 10-5, which was about an order of magnitude lower than the azurin benchmark. However, it seemed to exhibit a large degree of quenching between apo- and Fe/Fe R2lox, on par with the A67C mutant, which was expected to exhibit the strongest coupling based on the Pathways calculations. As opposed to the other mutants, A67C does exhibit expected behavior in that it shows a relatively large amount of quenching in the metallated forms. The Pathways calculations showed a similar degree of coupling as the azurin benchmark. It also had the greatest degree of quenching in the medium to long timescales of all of the mutants. Thus, it was chosen for Mn/Fe experiments as well. The quenching of the

113 Figure 5.13: .The figure above shows multiple TCSPC tracesII forRu -labeled apo (black), Fe/Fe (red), and Mn/Fe (blue) A67C R2lox. The inset shows the experiments on a 500 ns timescale, with a temporal resolution of 0.244 ns.

114 Figure 5.14: The figure compares the TCSPC traces of various mutants in bothapo- (black) and Fe/Fe (red) R2lox. The traces are constructed from experiments done on two separate time scales (500 ns and 5 µs).

115 Mn/Fe form isn’t quite as drastic as the Fe/Fe form; it is on par with the quenching that the other mutants exhibited in their Fe/Fe forms. This is to be expected, as the decay factor is ∼10x smaller for coupling to FeIII as opposed to MnIII in this mutant. A67 is very close to E69, a coordinating glutamate of the site 1 (MnIII) metal.

5.4 Conclusions & Future directions

Electron transfer is the engine that drives life. Ruthenium-labeled proteins have proven to be a very good way to study biological electron transfer reactions. There have not yet been forays into running flash-quench experiments with ruthenated non- heme di-iron enzymes nor in Mn/Fe systems. R2lox represents a unique case, due to the divide in its structure (R2c) and hypothesized function (BMMs), which is dis- cussed at length in section 1.4. This chapter has dealt with identifying, preparing, and screening various mutants to use in the study of electron transfer reactions in R2lox. Future directions of this work include conducting transient absorption mea- surements on these labeled mutants to extract rate constants and begin to measure the various parameters of Marcus electron transfer theory.

116 Chapter 6 Applications of resonance Raman spectroscopy: Probing [2Fe-2X] (X=S,Se) glutathione clusters

6.1 Introduction

Iron-sulfur clusters are crucial in biological systems across all domains of life.362 They are involved in many cellular processes, which include iron trafficking and gene expres- sion. In fact, improper iron-sulfur cluster biogenesis has been implicated in human dis- eases.363 Iron-sulfur clusters are also integral in many electron transfer pathways.364,365 Cellular respiration, one of the most important reactions on earth for maintaining multicellular life, relies on iron-sulfur proteins to shuttle electrons through the path- way. Additionally, many complex enzymes utilize iron-sulfur clusters within accessory proteins or within themselves to act as transportation for electrons to the active site, where they are used to perform chemically difficult reactions. Iron-sulfur clusters can take on many forms, as seen in Figure 6.1. The most common are the [2Fe-2S] and [4Fe-4S] clusters. However, various other forms of iron-sulfur clusters have been found in nature, such as the medial [3Fe-4S] cluster in [NiFe] hydrogenase and the [8Fe-7S] P-cluster found in nitrogenase.366 Most iron- sulfur clusters can carry a single electron, however the large P-cluster in nitrogenase

117 can carry two electrons.367

Figure 6.1: Examples of different iron-cluster forms found in nature. [2Fe-2S] and [4Fe-4S] (a and b) clusters are the most common, whereas [3Fe-4S] (c) clusters are more rare, and [8Fe-7S] (d) clusters have only been identified in nitrogenase.

There is an experimental precedent for probing the activity of iron-sulfur cluster- containing proteins by substituting selenium into the cluster in the place of sulfur. This has been shown to alter the behavior of the iron-sulfur clusters, but not si- lence it.368 Selenium substitution has also been used to probe the stability of Fe-X (X=S,Se) clusters, both in experimental and theoretical contexts.369,370 Selenium- substituted clusters have also proven to be a useful tool to probe the spin state and electronic coupling in iron-sulfur clusters by magnetic resonance experiments.371,372 Finally, selenium substitution is useful for studying the assembly of iron-sulfur clus- ters. Additionally, due to importance of iron-sulfur clusters in the field of bioinorganic

118 chemistry, iron-sulfur proteins were some of the first systems that were studied by resonance Raman spectroscopy. Many early resonance Raman studies were conducted on ferredoxins and adrenodoxins.214,215,373

Figure 6.2: The figure above shows sample UV/vis absorbance spectra of various [2Fe-2S] and [4Fe-4S] clusters found in nature. Figure was taken from reference [224].

The vibrations associated with iron-sulfur clusters are very low in energy, around 200−400 cm-1, due to the relatively large masses of the constituent atoms. Thus, resonance Raman spectroscopy of iron-sulfur clusters needs to be conducted at low temperatures so that the v=0 state may be fully populated. Resonance Raman exper- iments can yield very useful structural information about iron-sulfur clusters, such as the oxidation state and structural identity ([2Fe-2S], [4Fe-4S], etc). The absorption

119 bands of iron-sulfur clusters, which originate from S→Fe LMCT’s, are very broad, as seen in figure 6.2 giving a wide selection of wavelengths from which to select.224 However, this can lead to lower scattering intensities which arise from interference from a high density of electronic transitions.374 Recently, Qi et al demonstrated that [2Fe-2S] clusters can be coordinated by glu- tathione (GSH) under physiological conditions.375 Various spectroscopic and electro-

chemical techniques indicated that the product was indeed a [2Fe-2S](GS)4 species. Additionally, it was shown that the iron-sulfur cluster assembly protein, ISU, can form this product in the presence of glutathione. Glutathione is a common antioxi- dant present in the cell. It is a tripeptide that consists of a Glu, Cys, and Gly, with a γ−peptide bond between the carboxylate sidechain of the Glu and the amino group of the Cys. Glutathione protects the cell against many harmful chemicals, such as reactive oxygen species, by reducing them.376 It is then in turn reduced by an enzyme called glutathione reductase.377,378 The purpose of this following study was to use res-

onance Raman to characterize [2Fe-2Se](GS)4 and compare it against [2Fe-2S](GS)4. The structural data gleaned from the resonance Raman experiments could be very beneficial in the wider context of studying the formation of iron-selenium clustersin a biological context.

6.2 Materials and Methods

6.2.1 Sample identity and preparation

The samples were prepared by members of the Cowan research group at Ohio State University. The sample preparation method the protocol described previously, ex- cept selenium was used in the place of sulfur for the [2Fe-2Se] compound.375 The compounds were dissolved in HEPES buffer, placed in EPR tubes, and frozen in liq-

uid N2. Figure 6.3 shows the chemical structure of the samples that were studied. 120 The absorbance spectrum was supplied by the Cowan research group and is shown in Figure 6.4.

Figure 6.3: The figure shows the chemical structures of the [2Fe-2S] and [2Fe-Se] clusters that were studied using resonance Raman spectroscopy.The dotted lines in the GS structure show the divisions into the component amino acids.

6.2.2 Resonance Raman spectroscopy

The sample was held in a finger dewar (Wilmad WG-816-Q), surrounded by liquid

N2 to maintain a low temperature. The laser excitation at 457.9 nm was from a Coherent I-70C Spectrum Kr/Ar mixed gas laser. The laser was incident onto the sample in a backscattering geometry, and scattered light was collected by a 55 mm FD f/1.2 mounted compound lens (Canon) and focused into a single grating, 320 mm spectrograph (Princeton Instruments) by a 50 mm, f/4 achromatic lens (Newport).

121 Figure 6.4: The figure shows the UV/vis absorbance spectra of the [2Fe-2S] and [2Fe-Se] clusters.

Rayleigh light was filtered out of the collected light by an edge filter coated forthe appropriate laser excitation wavelength (Semrock). The Raman scattered light was separated by a single 1800 gr/mm holographic grating and imaged onto a Peltier- cooled CCD detector (Princeton Instruments). Data was collected in 20 min intervals for 2 hours, and the resulting spectra were monitored for photodegradation. The spectra were calibrated within 1 cm-1 by a 50/50 mixture by volume of toluene and acetonitrile. Data was exported and analyzed in Igor Pro 6 (Wavemetrics Inc.).

122 6.3 Results and Discussion

The resonance Raman spectra of the GSH clusters is shown in Figure 6.5. As men- tioned above, the peaks are localized in the low frequency region of the spectrum, due to the relatively large atomic weights of the Fe and S or Se atoms that are involved in the molecular vibrations. The [2Fe-2Se] cluster also has more peaks than the [2Fe-2S] cluster, as well as a greater resonance enhancement. This is probably due to the redshift of the [2Fe-2Se] absorption spectrum, shown in Figure 6.4, resulting in an increased overlap and therefore resonance enhancement. The [2Fe-2S] resonance Ra- man spectrum shows relatively weak enhancement, with stronger bands at 264, 300, and 330 cm-1 and weaker flanking peaks at 187 and 453-1 cm . This is quite different from many of the ferredoxin [2Fe-2S] clusters that have been probed via resonance Raman.379 However, some of the modes present in the GSH-complexed [2Fe-2S] cluster have been seen in other systems, most notably the 330 cm-1 peak, which has been identi- fied as a terminal Fe-Cys stretching mode.380,381 Bridging modes have been identified in [2Fe-2S] species at 280−290 cm-1 and 390−420 cm-1, both of which are absent or obscured in the GSH-complexed [2Fe-2S] cluster.381 As mentioned before, the higher intensity of the [2Fe-2Se] resonance Raman spectrum can be attributed to the larger absorbance at the selected excitation wavelength. Additionally, the [2Fe-2Se] spec- trum shows more bands than the previous spectrum. Intense bands can be seen at 209, 265, and 360 cm-1 with smaller bands located at 181, 297, and 304 cm-1. The peak at 237 cm-1 marked with an asterisk arises from matrix ice. Reported spectra of [2Fe-2Se] clusters in biological systems have shown peaks at 176, 220, 261, and 355 cm-1 when an excitation of 457.9 nm was used, which are similar in frequency to some of the peaks observed in this study.381 Chalcogenic substitution of S for Se should shift the bridging vibrations, however, it should have little to no effect on the 123 Figure 6.5: Resonance Raman spectra of [2Fe-2S] and [2Fe-2Se] glutathione clusters. The data was collected at 15 mW of 457.9 nm light for 2 hrs.

124 terminal Fe-Cys stretches. Fe-Cys terminal stretching modes have been identified at 270−290 cm-1 and 320−360 cm-1.215 These values match up with the peaks in the [2Fe-2Se] spectrum at 265, 297, and 360 cm-1. These values for the terminal Fe-Cys stretching modes are weaker than what is seen in literature, which would lead to a lower complexing strength seen in the clusters when compared to literature values of [2Fe-2S] proteins.215,381 This may arise from the hypothesized activity of GSH as a trafficking agent for iron-sulfur clusters ina biological context. The lowered complex binding strength may be beneficial when the GSH cluster is required to deposit the iron-sulfur cluster in the desired protein scaffold.

6.4 Conclusion

Iron-sulfur clusters are crucial for many biological processes. Thus, the trafficking and assembly of iron-sulfur clusters is very important. An effective way of studying the assembly and activity of iron-sulfur clusters is through the chalcogenic substitution of sulfur for selenium. Previously, glutathione (GSH) has been shown to act as a trafficking molecule in addition to its well-studied and characterized reducing prop- erties. Resonance Raman spectra of GSH-complexed [2Fe-2S] and [2Fe-2Se] clusters were measured. The resonance Raman spectra were comparable to published spectra of [2Fe-2S] and [2Fe-2Se] clusters. Furthermore, a lowered Fe-Cys terminal stretching frequency indicates a weaker bond than that observed in iron-sulfur cluster containing proteins. This has implications for the hypothesized iron-sulfur trafficking function of GSH.

125 Chapter 7 Applications of resonance Raman spectroscopy: The search for a terminal hydride in synthetic FeFe hydrogenase mimics

7.1 Introduction

The global levels of CO2 are rising, due to the prevalent use of fossil fuels for energy sources.42 To slow this accumulation of carbon dioxide, a clean energy source that does not produce harmful byproducts is needed. Hydrogen is a clean burning fuel that has a very high energy density and only produces water as the result of its oxidation.382,383 Thus, it is of great scientific and industrial importance to be able to make hydrogen in an efficient manner so that it may be used as a clean sourceof energy. The development of a catalyst that can cheaply produce hydrogen has been a goal of scientific research for many decades. Many 2nd and 3rd row transition metal complexes have been shown to produce hydrogen very efficiently.384–387 However, these elements are rare and expensive and are often toxic. Thus, it is important to create hydrogen production catalysts that use 1st row transition metals, as they are cheaper and not as toxic. This is more challenging than using 2nd and 3rd row transitions

126 metals, as the complexes are more difficult to synthesize. However, some groups have been able to incorporate 1st row transition metals, such as nickel and cobalt, in complexes that can produce hydrogen in an aqueous environment.388,389

Figure 7.1: Active site structures of [NiFe] (a), [FeFe] (b), and [Fe]-only (c) hydroge- nases.

Another strategy that has been employed by many groups is to mimic what nature has developed. Hydrogenases are a class of enzyme that can reduce protons to make

390 hydrogen, or oxidize H2 to make protons and electrons. They are very efficient at this reaction, exhibiting turnover frequencies on the order of 103−104 s-1.391–393 There are three main classes of hydrogenases, organized by the metals present in the active site: [NiFe], [FeFe], and [Fe]-only hydrogenases.394 In terms of active site structure, the [NiFe] and [FeFe] variants are more similar while the [Fe] hydrogenase active site has a much different geometry, as seen in Figure 7.1, and will not be discussed further in

127 this chapter.395,396 Additionally, the [NiFe] and [FeFe] hydrogenases possess different numbers of iron-sulfur clusters., which facilitate electron transfer from the surface to the active site of the protein.328 Although they are similar in active site structure and function, the [NiFe] and [FeFe] hydrogenases have varying degrees of activity. The [NiFe] hydrogenase tends to be a more active hydrogen oxidizer, whereas the

397,398 [FeFe] hydrogenase is more biased towards the production of H2 from protons. Uncovering the mechanism by which [FeFe] hydrogenases carry out their function is

very important for the production of H2 using environmentally practices. Studying [FeFe] hydrogenases are difficult because they are very large enzymes with many subunits that require post-translational assembly. Additionally, they are

extremely O2 intolerant. Thus, many synthetic active site mimics the have been syn- thesized to facilitate the study of the [FeFe] hydrogenase mechanism, which is still not completely understood.399–402 Recently, efforts have gone into refining synthetic models that possess the bridging amino-dithiolate that has been shown to be present within the [FeFe] hydrogenase active site, in addition to stabilizing the terminal hy- dride form that occurs during catalysis in the native enzyme.394,403,404 In 2014, Huynh et al published work that investigated these structures computationally.405 The work described in this section is the result of experiments done on [FeFe] hydrogenase ac- tive site models in an attempt to observe a terminal hydride species via resonance Raman spectroscopy.

7.2 Materials and Methods

7.2.1 Sample identitiy

The samples were received from the Rauchfuss group. The identity of the compounds is shown in Figure 7.2. The samples were synthesized, dissolved in dichloromethane, and then frozen and shipped on dry ice to maintain sample integrity. The chemical

128 structures of the samples that were studied is shown in Figure 7.2, along with the identifying Roman numeral. A sample UV/vis spectrum of the terminal hydride species (IV), provided by the Rauchfuss research group, is shown in Figure 7.3.

Figure 7.2: Chemical structures of the synthetic hydrogenase mimics. The Roman numeral will be referred to for the remainder of the chapter.

7.2.2 Resonance Raman spectroscopy

The sample was held in a finger dewar (Wilmad WG-816-Q), surrounded by liquid N2 to maintain a low temperature. For the 364 nm excitation, the 728 nm fundamental was generated from a 10 ps Ti:Sapphire (Spectra Physics) and doubled using a criti- cally phase matched BBO crystal (Eksma Optics). The laser excitation at 457.9 nm was from a Coherent I-70C Spectrum Kr/Ar mixed gas laser. The laser was incident onto the sample in a backscattering geometry, and scattered light was collected by a

129 Figure 7.3: The figure shows a sample UV/vis absorbance spectrum of the hydroge- nase mimics.

55 mm FD f/1.2 mounted compound lens (Canon) and focused into a single grating, 320 mm spectrograph (Princeton Instruments) by a 50 mm, f/4 achromatic lens (New- port). Rayleigh light was filtered out of the collected light by an edge filter coated for the appropriate laser excitation wavelength (Semrock). The Raman scattered light was separated by a single 1800 gr/mm holographic grating and imaged onto a Peltier-cooled CCD detector (Princeton Instruments). Data was collected in 20 min intervals for 2 hours, and the resulting spectra were monitored for photodegradation. The spectra were calibrated within 1 cm-1 by a 50/50 solution by volume of toluene and acetonitrile. Data was exported and analyzed in Igor Pro 6 (Wavemetrics Inc.).

130 7.3 Results and Discussion

The resonance Raman spectrum of the non-bridging compound at 457.9 nm is shown below in Figure 7.4. This wavelength, although not as strongly absorbed as more energetic frequencies by the compounds under investigation as seen in Figure 7.3, was the first wavelength used because it is a lower energy and will not damagethe sample as much as UV light would.

Figure 7.4: Resonance Raman spectrum of I. The spectrum was acquired using 10 mW of 457.9 nm cw light for 2 hrs.

There is a high density of states located in the 350−600 cm-1 and 1000−1200 cm-1 frequency ranges. The lower frequency peaks can be ascribed to metal-ligand centered vibrations, whereas the peaks present in the intermediate range as well as the peak at ∼1600 cm-1 are likely aromatic ring deformations from the dppv ligand. Additionally, the strong peak at 1860 cm-1 is a CO stretching vibration. This spectrum will be a useful comparison for the bridging and terminal hydride compounds shown next. Figure 7.5 shows the resonance Raman spectrum of the bridging hydride(deuteride) 131 species when excited with 457.9 nm light. The spectrum shows a high density of peaks in the low region of the spectrum, as well as the 1000 cm-1 and 1600 cm-1 phenyl stretches.

Figure 7.5: Resonance Raman spectra of II and III. The spectra were acquired using 10 mW of 457.9 nm cw light for 2 hrs.

However, the CO peak at 1860 cm-1 was not present in the bridging compounds. Additionally, the intensity of the peaks in the low frequency range was lower than in the unbridged precursor compound. The bridging hydride species was not observed. Previous Raman studies of synthetic FeFe hydrogenase models have given bridging hydride frequencies of around 1220 cm-1 with a shift to 891 and 1009 cm-1 upon deuteration.406 No such peak was observed here. The study above used an excitation source of 785 nm, which is far outside the absorption of the system under study. The higher frequency photons used here may have contributed to the disappearance of the bridging hydride due to photodisassociation The next set of compounds that was investigated was the terminal hydride and

132 deuteride compounds (IV and V). Iron-hydride stretches have been measured for synthetic FeFe hydrogenase mimics by FT-IR at 1844 cm-1.406,407

Figure 7.6: Resonance Raman spectra of IV and V. The spectra were acquired using 10 mW of 457.9 nm cw light for 2 hrs.

Simple isotope replacement analysis puts the iron-deuteride stretch at 1315 cm-1. However, no such peaks were observed in the resonance Raman spectrum collected at 457.9 nm excitation. This is probably due to the reported instability of the terminal hydride species when exposed to laser radiation.406 Figure 7.6 shows the terminal hydride compounds when excited at 457.9 nm. Compared to the precursor compound the terminal compounds have a higher number of CO stretching frequencies. This is most likely due to the breaking of the symmetry of the molecule by adding the hydride. These experiments were repeated at an excitation wavelength of 364 nm. This was done to harness the higher absorbance in the near-UV regime of the UV/vis spectrum of the compounds. However, the downside is that near-UV photons are more energetic

133 than their visible counterparts, and have a higher probability of causing damage to the system under study. Figure 7.7 shows the spectra excited at 364 nm. The spectra collected at 364 nm look quite different from those collected at 457.9 nm. The density of peaks in the low frequency range is much lower compared to the spectra shown before. Most of the peaks in the mid frequency range have disappeared as well, as have the CO stretches in the 1800 – 1950 cm-1 range. This disappearance of peaks may be attributed to the high energy photons used in this experiment, as well as destruction of the sample due to oxidation because these experiments were done later than the ones shown before.

Figure 7.7: Resonance Raman spectra of I−V. The spectra were acquired using 10 mW of 364 nm for 2 hrs.

134 7.4 Conclusion

Hydrogen is a good source of clean energy that would be greatly beneficial if it were able to be harnessed. Hydrogenases are very active enzymes that catalyze the

reduction of proteins to form H2, as well as the oxidation of H2 to form protons. [FeFe] hydrogenases are the most active hydrogen producers found in nature. Studying

[FeFe] hydrogenase is often very difficult, due to2 itsO sensitivity and complex in vivo maturation pathways. Thus, model compounds are frequently utilized to study this complex enzyme. Model compounds were synthesized that were thought to stabilize the terminal hydride species and resonance Raman spectroscopy was used to try to identify the iron-hydride stretching frequencies. No hydride peaks were seen in either the bridging or the terminal hydride compounds, possibly due to the relatively high frequency of the laser radiation used in this study. If this study were to be conducted in the future, it would be beneficial to use non-resonant Raman to avoid the high- energy photons in the UV region of the spectrum. As stated before, the Ti:Sapphire laser has the capability to reach 785 nm, which is found in many commercial black- box Raman spectrometers. However, care will need to be taken to attenuate the laser intensity due to the high photon fluxes the fundamental Ti:Sapphire is capable of emitting.

135 Chapter 8 Applications of resonance Raman spectroscopy: Illuminating the structure and reactivity of a unique ruffled heme in cytochrome P460

8.1 Introduction

The nitrogen cycle is very important for the continuation of life on Earth, be- cause every organism on the planet uptakes and releases nitrogen as part of its metabolism.408,409 The very blueprint of life, DNA, depends on the utilization of nitrogen for the formation of the base−pairing sequences that encode for the ma- chinery of the cell. However, gaseous nitrogen, although it is the largest source of nitrogen on the planet, is unusable by almost every organism on the planet due to the strength of the triple bond.410 Thus, for nitrogen to be useful in a biological context, it needs to be converted to a form that can be utilized. Molecular nitrogen is converted to ammonia by nitrogen−fixing bacteria using an enzyme called nitrogenase.329 The ammonia is then either utilized, or is oxidized by nitrifying bacteria to nitrites and

408,411 nitrates. From there, the NOx molecules are either used or converted back to atmospheric nitrogen by denitrifying bacteria.412 This is a very simple overview of a

136 very complex system called the nitrogen cycle.413 Understanding how these pathways function not only has implications for basic science research, but there are also industrial and environment impacts. Nitrous ox- ide (N2O) is a toxic gas that has been shown to be a very effective ozone depletion

414 415–417 agent. A major source of N2O gas is from nitrifying bacteria Thus, by under- standing the mechanisms by which ammonia is converted into nitrite as well as the nature of side product formation, the levels of N2O can be addressed. Throughout the nitrogen cycle, there are many very complex enzymes that are re- quired to carry out the reactions. Nitrifying bacteria use a pair of enzymes called am- monia monooxygenase (AMO) and hydroxylamine oxidoreductase (HAO) to produce

- 408,418,419 NO2 from ammonia. AMO is a membrane−bound Cu protein, while HAO is a cytoplasmic protein that is a trimer, with each monomer containing 8 hemes.420 Crystallographic evidence of HAO from the nitrifying bacterium Nitrosomas europaea shows that one of the eight hemes on each monomer of HAO is a “ruffled” heme, where covalent linking has distorted the canonically planar heme geometry (Figure 8.1), and is the loci of catalysis.421 This unique heme has been named P460 due to the absorbance of its Soret band when in the ferrous state.422,423 The heme ruffling has been shown to be the resultof a covalent link to Y467 within the protein scaffold.421 It is thought that the ruffling of the heme P460 is what causes this particular catalytic activity in HAO.424 Resonance Raman spectroscopy of the heme P460 chromophore from N. europaea HAO has been conducted.425 The spectra showed a unique heme cofactor that is very distinct from the spectra of traditional heme that were collected previously.217,425 This can be attributed to the ruffling phenomenon, which breaks the symmetry of the planar heme and is the source for the increased number of peaks.425 Heme P460 has also been identified in a monoheme cytochrome from N. europaea

137 Figure 8.1: The crystal structure of HAO and from N. europaea. The inset structure shows the ruffled heme active site bound to the substrate, as well as the covalently bound Tyr. (PDB:4N4O)

called cytochrome P460. Instead of a covalent link to a tyrosine, like in HAO, the source of the ruffling is a bond between a nearby lysine and the13’−meso carbon of the heme ring, forcing to carbon to adopt an sp3 configuration (Figure 8.2).426 Cytochrome P460, like HAO, has been shown to catalyze the oxidation of hydrox- ylamine.427 Thus, cytochrome P460 can be utilized as a model to study the P460 cofactor in HAO. As mentioned above, HAO has been shown to produce NO from hydroxylamine,

which is then oxidized to NO2. If cytochrome P460 is to be used as a model, then more must be known about its reaction mechanism. Figure 8.3 shows a proposed mechanism for the production of NO from hydroxy- lamine by cytochrome P460. The goal of the work described below is to use resonance Raman spectroscopy to inform on the structure of the cofactor under different con- ditions to gain a more complete understanding of the reaction mechanism of heme P460 in the oxidation of hydroxylamine to NO. 138 Figure 8.2: The crystal structure of cytochrome P460 and from N. europaea. The inset structure shows the ruffled heme active site bound to phosphate as well asthe covalently linked lysine. (PDB:2JE3)

8.2 Materials and Methods

8.2.1 Sample preparation and identity

The samples were received from the Lancaster group at Cornell University. The identity of the samples is shown in Table 8.1. The samples were prepared and then frozen and sent in a dry shipper to maintain sample integrity. Room temperature samples were also sent for UV/vis absorbance spectroscopy (Figure 8.4) to determine the optimal wavelengths for the resonance Raman experiments. DEA/NONOate and Angeli’s salt were used as NO• and NO− donors, respectively.

8.2.2 Resonance Raman spectroscopy

The sample was held in a finger dewar (Wilmad− WG 816−Q), surrounded by liquid

N2 to maintain low temperatures. The laser excitations at 457.9 nm and 568.2 nm were from a Coherent I−70C Spectrum Kr/Ar mixed gas laser. The laser (20 mW) was incident onto the sample in a backscattering geometry, and scattered light was

139 Figure 8.3: A proposed mechanism for the reaction of cytochrome P460 and hydrox- ylamine.

140 Table 8.1: Resonance Raman samples for cytochrome P460 experiments

Sample Number Sample Identity

1 250 µM cytochrome P460 + O2 2 250 µM cytochrome P460 + O2 + 100x hydroxylamine 15 3 250 µM cytochrome P460 + 100x N-hydroxylamine + O2 4 250 µM cytochrome P460 + 100x hydroxylamine 5 250 µM cytochrome P460 + 100x 15N-hydroxylamine 6 250 µM cytochrome P460 + O2 + 20x Angeli’s salt 7 250 µM cytochrome P460 + 20x Angeli’s salt 8 250 µM cytochrome P460 + O2 + 20x DEA/NONOate 9 250 µM cytochrome P460 + 20x DEA/NONOate

Figure 8.4: The figure above shows the UV/vis absorbance spectrum for cytochrome P460 as well as labeling the excitation wavelengths used in the resonance Raman experiments.

collected by a 55 mm FD f/1.2 mounted compound lens (Canon) and focused into a single grating, 320 mm spectrograph (Princeton Instruments) by a 50 mm, f/4 achromatic lens (Newport). Rayleigh light was filtered out of the collected light by

141 an edge filter coated for the appropriate laser excitation wavelength (Semrock). The Raman scattered light was separated by a single 1800 gr/mm holographic grating and imaged onto a Peltier−cooled CCD detector (Princeton Instruments). Data was collected in 20 min intervals for 2 hours, and the resulting spectra were monitored for photodegradation. The spectra were calibrated within 1 cm-1 by a 50/50 mixture solution by volume of toluene and acetonitrile. Data was exported and analyzed in Igor Pro 6 (Wavemetrics Inc.).

8.3 Results and Discussion

The resonance Raman spectrum of cytochrome P460 is shown in Figure 8.5 (black trace). As mentioned above, a covalent linkage of K70 to the 13’−meso carbon of the ring is the source of the unique geometry in this cofactor. Thus, the symmetry of the heme ring is broken which gives rise to the multitude of peaks observed in the spectrum. The peaks are very intense, due to the polarizability of the conjugated sys- tem of the heme ring. Like in traditional heme resonance Raman spectra, the peaks in the 1200−1600 cm-1 can be assigned to skeletal stretches of the porphyrin ring.217 The lower frequency bands from 200−700 cm-1 arise from Fe−centered vibrations, in- cluding any metal−ligand stretches that occur. In the bottom trace, hydroxylamine has been added to the protein. As seen in the gray trace in Figure 8.5, the peaks decreased in intensity as well as number. This can be attributed to electron density moving away from the Fe−heme bonds and into the Fe−hydroxylamine bond, de- creasing the enhancement of the prior vibrations. The high frequency bands change as well, with the most prominent being the intense 1625 cm-1 peak that is not present in the protein−only sample. Cytochrome P460 was also complexed with 15N−labeled hydroxylamine to identify the vibrational modes that involve hydroxylamine. Figure 8.6 shows the cytochrome

142 Figure 8.5: Resonance Raman spectrum of cytochrome P460 with (gray) and without (black) hydroxylamine. The spectra were collected using 20 mW of 457.9 nm light for 2 hrs.

P460 complexed to 15N−hydroxylamine as well as with 14N−hydroxylamine under aerobic and anaerobic conditions. Interestingly, the anaerobic spectra do not differ from each other whatsoever. This is noteworthy because the aerobic spectra do differ slightly, most notably in the 400−600 cm-1 region, which is where Fe−N stretches have been identified in heme systems.428 However, none of the spectra here resemble the cytochrome P460−only spectrum from Figure 8.5. Therefore, the hydroxylamine must be interacting in some way, although anaerobic system does not exhibit any change. The next set of samples were investigated to ascertain the mechanism by which cytochrome P460 oxidizes hydroxylamine. To understand what intermediate the re- action proceeds through, two different compounds were introduced to cytochrome P460: diethylamine NONOate (DEA/NO), a NO• donor, and Angeli’s salt, a ni- troxyl (NO−) donor.429–433 If the mechanism proceeds via a NO• intermediate, then the resonance Raman spectrum with DEA/NO will match the hydroxylamine prod- uct spectrum. Likewise, if the spectrum with Angeli’s salt matches, the mechanism

143 Figure 8.6: Resonance Raman spectrum of cytochrome P460 complexed to 15N (red) and 14N−labeled (blue) hydroxylamine under aerobic and anerobic conditions. The top set is under aerobic conditions and the bottom set is under anaerobic conditions. Spectra were collected using 20 mW of 457.9 nm light for 2 hrs.

proceeds through a nitroxyl anion intermediate. The spectra will be referred to here- after in the Enemark−Feltham notation, which notates the metal−nitrosyl complex by the number of electrons in the d−orbital of the metal and the π∗ orbital of the NO• ligand.434 The reason for this is that NO is a non−innocent ligand, which means that it can donate electrons into the d−orbitals of the metal, thus casting the oxidation state of the metal and the ligand in doubt.435 Here, FeIII has 5 d−electrons and NO has a single electron in its π∗ orbital, so an FeIII-NO complex in this case will be denoted as {FeNO}6. Likewise, the nitroxyl anion has 2 electrons in its π∗ orbital, so an FeIII-NO− complex is denoted as {FeNO}7. The resonance Raman spectra of {FeNO}6 and {FeNO}7 cytochrome P460 under aerobic and anaerobic conditions is shown in Figure 8.7. The spectra here look very different from those in Figures 8.5 and 8.6, which is unsurprising. The spectra have the same basic structure, however, the peaks are more intense than in the previous spectra. The biggest changes are localized in the low frequency part of the spectrum, from ∼200−600 cm-1. The {FeNO}7 spectra particularly have the greatest amount of change, with very intense peaks present in this region, as well as throughout the spec-

144 Figure 8.7: Resonance Raman spectra of cytochrome P460 with 20x DEA/NO (red/black) and Angeli’s salt (blue/green). under aerobic (blue/red) and anaero- bic (green/black) conditions. Spectra were collected using 20 mW of 457.9 nm light for 2 hrs.

tra. Neither of the {FeNO}7 spectra look very much like the hydroxylamine−bound cytochrome P460 spectrum, which seems to eliminate the nitroxyl radical as an in- termediate in the mechanism. The {FeNO}6 samples have less intense bands than the {FeNO}7 spectra. In fact, as shown in Figure 8.8, the {FeNO}6 cytochrome P460 spectra look remarkably similar to the cytochrome P460 spectrum. The low frequency peaks are shifted slightly relative to each other and there is a rather intense peak in the aerobic {FeNO}6 spectrum at 230 cm-1. However, the medium to high frequency (∼750−1600 cm-1) range is almost identical, with slight variations in intensity. This is interesting, because according to the resonance Raman data, cytochrome P460 either does not interact with NO• or the reaction finished before the samples were frozen and measured. The first reason is unlikely, as NO• has been shown to interact with heme systems.436,437 Therefore, according to the data, perhaps cytochrome P460 interacts with NO• and completes its reaction very quickly. Additionally, higher frequency resonance Raman spectra were measured to ob-

145 Figure 8.8: Resonance Raman spectra of cytochrome P460 with 20x DEA/NO un- der aerobic (red) and anaerobic (blue) conditions compared with cytochrome P460 without ligand present. Spectra were collected using 20 mW of 457.9 nm light for 2 hrs.

serve any possible nitric oxide stretching frequencies. This could lead to greater insights about the structural identity of the hydroxylamine−complexed cytochrome P460, as well as the {FeNO}6 and {FeNO}7 species. Nitric oxide stretching frequen- cies depending greatly on the oxidation state of the NO molecule, because any extra electrons in the π∗ orbital of the NO molecule decreases the bond order and conse- quently the vibrational frequency. Free NO−, NO•, and NO+ stretching frequencies are 1284, 1876, and 2345 cm-1, respectively.438 The vibrations of heme−nitrosyl com- plexes have been measured extensively in the literature. NO stretches of {FeNO}6 and {FeNO}7 nitrosylheme species have been measured and found to be ∼1900−1920 cm-1 and ∼1620−1670 cm-1, respectively.439,440 Lower energy excitation was also introduced in this stage of the experiments. This is due to the alternative resonance enhancement that arises from Q−band excitation compared to Soret band excitation. Soret band excitation leads to the enhancement of A−type Raman bands, mentioned in section 2.6 and are totally symmetric.441 Like-

146 wise, Q−band excitation can lead to B−type Raman bands, which are non−totally symmetric.441 Furthermore, it has been observed that Soret band excitation may not enhance the NO stretches of a nitrosylheme complex, perhaps due to the lack of orbital conjugation between the heme and the nitric oxide.439 Therefore, the 568.2 nm laser line was also employed to measure the N−O stretching frequency. The excitation wavelength is denoted on the UV/vis absorbance spectrum in Figure 8.4.

Figure 8.9: Resonance Raman spectra comparing the high−frequency region of cy- tochrome P460, {FeNO}6, and {FeNO}7. The top spectrum of each pair was collected at 457.9 nm and the bottom spectrum was collected at 568.2 nm Spectra were col- lected using 20 mW of light for 2 hrs.

The resonance Raman spectra of the high frequency measurements is shown in Figure 8.9. Unsurprisingly, the 568.2 nm excitation gave a much different spectrum

147 than the 457.9 nm excitation. As mentioned previously, the NO stretching frequencies for {FeNO}6 and {FeNO}7 nitrosylheme species are in the 1900−1920 and 1620−1670 cm-1 ranges. However, as observed in the spectra above, there are no vibrations in this range. The only peak that may be a N−O stretch in the small band in the 457.9 nm excitation of the {FeNO}7 species at 1673 cm-1. This may be a NO• species; however, similar peaks were observed in the resting state cytochrome P460. Also, the stretching frequency for the {FeNO}7 species should be lower, at about 1620 cm-1. For decades, HAO was thought to catalyze the production of nitrite from hydrox- ylamine directly.442,443 That is, the reaction mechanism was thought to proceed from

- hydroxylamine straight to NO2 with the loss of 4 electrons. Recent studies have

- 444 showed that cytochrome P460 does in fact produce NO2 in the presence of O2. However, current work has shown that HAO catalyzes the oxidation of hydroxylamine

445 - to nitric oxide (NO) with a loss of 3 electrons. The NO is then oxidized to NO2 by

a reaction with O2, which explains the participation of O2 in the reaction observed previously.443 This functionality was also found to be present in cytochrome P460

under oxidizing conditions, with N2O being formed under anaerobic conditions. The data above, when considered together, does not support the original hypoth-

- • esis that cytochrome P460 makes NO2 . Rather, the addition of NO to form the {FeNO}6 species looks remarkably similar to cytochrome P460 with no added hy- droxylamine, indicating that it is likely that the timescale of the P460-NO• reaction is faster than the timescale of the sample preparation because NO• does interact with heme systems. The {FeNO}7 samples are unique in nature (Figure 8.7) and do not resemble any other sample in this study. Thus, it is probable that the hydoxylamine reaction proceeds through an {FeNO}6 intermediate and likely forms NO•, as opposed

- 444,445 to NO2 as thought previous to the aforementioned studies.

148 8.4 Conclusion

The nitrogen cycle is a very important part of the global network. It is a very com- plicated process; there are many complex enzymes that perform difficult chemistry to convert nitrogen into usable forms. Hydroxylamine oxidoreductase (HAO) is a mul- tiheme enzyme that can catalyze the production of nitric oxide from hydroxylamine using an unusual “ruffled” heme active site. The heme gains its unique geometry from a covalent link to a nearby Tyr to the porphyrin ring itself. This ruffled heme has also been identified in cytochrome P460 in N. europaea, which has also been shown to catalyze the oxidation of hydroxylamine. This makes it a good model to use for studying the active site of HAO. Resonance Raman spectroscopy was used to study the structure of cytochrome P460, as well as its hydroxylamine bound form. Addi- tionally, reagents were added which induce the formation of {FeNO}6 and {FeNO}7 species to approximate the hypothesized intermediates of the hydroxylamine oxida- tion reaction. Finally, high frequency measurements were conducted to attempt to measure the N−O stretches that would result during the mechanism. The resonance Raman study revealed many complicated spectra, with very intense bands that were difficult to analyze. The results overall were inconclusive, and further studies into this mechanism by using complementary techniques would be very advantageous.

149 Chapter 9 Conclusion

Bioinorganic chemistry is a field that has many applications. From energy produc- ing reactions to medical applications, nature has optimized many different reactions that are driven by transition metals. There are many ways in which to probe these bioinorganic systems to learn more about them to apply that knowledge in improv- ing our own technology. Of all the various ways by which to study these systems, spectroscopy remains the most versatile and powerful. Spectroscopy is sensitive and can yield a lot of information about the system of interest. Vibrational spectroscopy specifically holds a prominent spot in the field of chemistry due to its relationship with chemical structure. There are many kinds of vibrational spectroscopy; one of the most useful is Raman spectroscopy. Raman spectroscopy can be extended into resonance Raman spectroscopy, which uses the resonance Raman effect and parleys it into a site-specific probe of metal-centered vibrations in bioinorganic systems. This thesis details the development of a custom resonance Raman spectroscopic system and its use on a variety of bioinorganic systems. This system has yet to reach its full potential, with plenty of room to extend into the UV range to probe protein-based vibrations as well. Another branch of the work described herein is the study of a novel Mn/Fe protein, R2lox. R2lox is part of the ferritin-like superfamily, a very large group of proteins

150 with a wide range of functionalities. R2lox itself is similar in primary structure to the R2 subunits of the class I ribonucleotide reductases, which is where the radical is generated to reduce ribonucleotides for the de novo synthesis of deoxyribonucleotides. However, other aspects about R2lox lend to its function, which is currently unknown, being aligned with the BMMs (bacterial multicomponent monooxygenases) which can hydroxylate a host of hydrocarbon chemical species. If R2lox does exhibit this function, then it would be very interesting to the bioinorganic community, due to its ability to do very complex chemistry using a relatively simple scaffold. This thesis outlines work that probed a very unusual metal center structural change that was induced via irradiation by light, as well as laying the groundwork for the study of ET reactions in this system via ruthenium modification. In Chapter 1, the stage was set. The two prongs of the research conducted during graduate study were introduced and developed. In Chapter 2, Raman scattering was discussed, starting with incoherent light scattering and proceeding through to the quantum mechanical treatment of Raman scattering and the origin of the resonance Raman effect. In Chapter 3, the development of the custom resonance Raman system was discussed, as well as how it fits into past and current technology. The benefits of the system are highlighted as well as what modifications will be next to make it more effective. Chapter 4 outlines the discovery and subsequent characterization of a light- induced structural rearrangement of the primary coordination sphere of R2lox. It was found that it possessed remarkably similar characteristics to photodecarboxyla- tion chemistry, and may be a first step into engineering photochemical processes into metalloprotein scaffolds. Chapter 5 discusses the field of electron transfer and lays the groundwork for the study of ET reactions in R2lox. Cysteine mutants were made so the distance dependence of the electron transfer rate can be probed. Addition- ally, labeling conditions were optimized for R2lox for future experiments. Chapters

151 6−8 were collaborations with other research groups and exhibit the versatility of the custom resonance Raman instrumentation that was discussed in Chapter 3. This is the conclusion of this thesis. There are many things remaining to be studied about R2lox, including its role in the cell. Also, the custom resonance Raman instrumentation has the potential to be used for many more projects in the future to probe the metal-centered vibrations of bioinorganic systems.

152 Bibliography

1. Nelson, D. L. & Cox, M. M. Lehninger Principles of Biochemistry 5th editio (W. H. Freeman & Co., New York, NY, 2008).

2. Cowan, J. A. Inorganic Biochemistry: An Introduction 2nd editio (Wiley-VCH Inc, New York, NY, 1997).

3. Doyle, D. A. The Structure of the Potassium Channel: Molecular Basis of K+ Conduction and Selectivity. Science 280, 69–77 (1998).

4. Kofuji, P. & Newman, E. Potassium buffering in the central nervous system. Neuroscience 129, 1043–1054 (2004).

5. Morth, J. P. et al. Crystal structure of the sodium–potassium pump. Nature 450, 1043–1049 (2007).

6. Rose, C. & Chatton, J.-Y. Astrocyte sodium signaling and neuro-metabolic coupling in the brain. Neuroscience 323, 121–134 (2016).

7. Fearnley, C. J., Roderick, H. L. & Bootman, M. D. Calcium Signaling in Cardiac Myocytes. Cold Spring Harbor Perspectives in Biology 3, 1–20 (2011).

8. Cheung, W. Y. Calmodulin Plays a Pivotal Role in Cellular Regulation. Science 207, 19–27 (1980).

153 9. Yatani, A., Codina, J., Imoto, Y., Reeves, J. P. & Birnbaumer, L. A G Protein Directly Regulates Mammalian Cardiac Calcium Channels. Science 238, 1288– 1292 (1987).

10. Berridge, M. J., Lipp, P. & Bootman, M. D. The versatility and universality of calcium signaling. Nature Reviews Molecular Cell Biology 1, 11–21 (2000).

11. Chin, D. & Means, A. R. Calmodulin: a prototypical calcium sensor. Trends in Cell Biology 10, 322–328 (2000).

12. Hartwig, A. Role of magnesium in genomic stability. Mutation Research/Fundamental and Molecular Mechanisms of Mutagenesis 475, 113–121 (2001).

13. Owczarzy, R., Moreira, B. G., You, Y., Behlke, M. A. & Walder, J. A. Predicting Stability of DNA Duplexes in Solutions Containing Magnesium and Monovalent Cations. Biochemistry 47, 5336–5353 (2008).

14. Cowan, J. Structural and catalytic chemistry of magnesium dependent en- zymes. Biometals 15, 225–235 (2002).

15. Sissi, C. & Palumbo, M. Effects of magnesium and related divalent metal ions in topoisomerase structure and function. Nucleic Acids Research 37, 702–711 (2009).

16. Beauchamp, D. L. & Khajehpour, M. Studying salt effects on protein stability using ribonuclease t1 as a model system. Biophysical Chemistry 161, 29–38 (2012).

17. Scott Cayley, D., Guttman, H. J. & Thomas Record, M. Biophysical charac- terization of changes in amounts and activity of Escherichia coli cell and com- partment water and turgor pressure in response to osmotic stress. Biophysical Journal 78, 1748–1764 (2000).

154 18. Valdez, C. E., Smith, Q. A., Nechay, M. R. & Alexandrova, A. N. Mysteries of Metals in Metalloenzymes. Accounts of Chemical Research 47, 3110–3117 (2014).

19. Lin, Y.-W. Rational design of metalloenzymes: From single to multiple active sites. Coordination Chemistry Reviews 336, 1–27 (2017).

20. Holm, R. H., Kennepohl, P. & Solomon, E. I. Structural and Functional Aspects of Metal Sites in Biology. Chemical Reviews 96, 2239–2314 (1996).

21. Solomon, E. I. et al. Geometric and electronic structure/function correlations in non-heme iron enzymes. Chemical reviews 100, 235–350 (2000).

22. Joliot, P. Period-four oscillations of the flash-induced oxygen formation in pho- tosynthesis. Photosynthesis Research 76, 65–72 (2003).

23. Zouni, A. et al. Crystal structure of photosystem II from Synechococcus elon- gatus at 3.8 Å resolution. Nature 409, 739–743 (2001).

24. Mino, H. & Kawamori, A. EPR studies of the water oxidizing complex in the

S1 and the higher S states: the manganese cluster and YZ radical. Biochimica et Biophysica Acta (BBA) - Bioenergetics 1503, 112–122 (2001).

25. Umena, Y., Kawakami, K., Shen, J.-R. & Kamiya, N. Crystal structure of oxygen-evolving photosystem II at a resolution of 1.9 Å. Nature 473, 55–60 (2011).

26. Kühlbrandt, W., Wang, D. N. & Fujiyoshi, Y. Atomic model of plant light- harvesting complex by electron crystallography. Nature 367, 614–621 (1994).

27. Hofbauer, W. et al. Photosystem II single crystals studied by EPR spectroscopy

• at 94 GHz: The tyrosine radical Y D. Proceedings of the National Academy of Sciences 98, 6623–6628 (2001).

155 28. Vasil’ev, S., Orth, P., Zouni, A., Owens, T. G. & Bruce, D. Excited-state dy- namics in photosystem II: Insights from the x-ray crystal structure. Proceedings of the National Academy of Sciences 98, 8602–8607 (2001).

29. Cox, N., Pantazis, D. A., Neese, F. & Lubitz, W. Biological Water Oxidation. Accounts of Chemical Research 46, 1588–1596 (2013).

30. Hatefi, Y. The Mitochondrial Electron Transport and Oxidative Phosphoryla- tion System. Annual Review of Biochemistry 54, 1015–1069 (1985).

31. Boyer, P. D. et al. Oxidative phosphorylation and photophosphorylation. An- nual review of biochemistry 46, 955–966 (1977).

32. Iwata, S., Ostermeier, C., Ludwig, B. & Michel, H. Structure at 2.8 Å resolution of cytochrome c oxidase from Paracoccus denitrificans. Nature 376, 660–669 (1995).

33. Senkler, J., Senkler, M. & Braun, H.-P. Structure and function of complex I in animals and plants - a comparative view. Physiologia Plantarum 161, 6–15 (2017).

34. Huang, S. & Millar, A. H. Succinate dehydrogenase: the complex roles of a simple enzyme. Current Opinion in Plant Biology 16, 344–349 (2013).

35. Xia, D. et al. Structural analysis of cytochrome bc1 complexes: Implications to the mechanism of function. Biochimica et Biophysica Acta (BBA) - Bioener- getics 1827, 1278–1294 (2013).

36. Cooley, J. W. Protein conformational changes involved in the cytochrome bc1 complex catalytic cycle. BBA - Bioenergetics 1827, 1340–1345 (2013).

37. Tan, M.-L., Balabin, I. & Onuchic, J. N. Dynamics of Electron Transfer Path- ways in Cytochrome c Oxidase. Biophysical Journal 86, 1813–1819 (2004).

156 38. Lucas, M. F., Rousseau, D. L. & Guallar, V. Electron transfer pathways in cytochrome c oxidase. Biochimica et Biophysica Acta (BBA) - Bioenergetics 1807, 1305–1313 (2011).

39. Hao, O. J., Chen, J. M., Huang, L. & Buglass, R. L. Sulfate‐reducing bacteria. Critical Reviews in Environmental Science and Technology 26, 155–187 (1996).

40. Jia, R., Yang, D., Xu, D. & Gu, T. Electron transfer mediators accelerated the microbiologically influence corrosion against carbon steel by nitrate reducing Pseudomonas aeruginosa biofilm. Bioelectrochemistry 118, 38–46 (2017).

41. Ragsdale, S. W. Nickel and the carbon cycle. Journal of Inorganic Biochemistry 101, 1657–1666 (2007).

42. Solomon, S., Plattner, G.-K., Knutti, R. & Friedlingstein, P. Irreversible climate change due to carbon dioxide emissions. Proceedings of the National Academy of Sciences 106, 1704–1709 (2009).

43. Ragsdale, S. W. & Kumar, M. Nickel-Containing Carbon Monoxide Dehydro- genase / Acetyl-CoA Synthase (1996).

44. Harrop, T. C. & Mascharak, P. K. Structural and spectroscopic models of the A-cluster of acetyl coenzyme a synthase/carbon monoxide dehydrogenase: Nature’s Monsanto acetic acid catalyst. Coordination Chemistry Reviews 249, 3007–3024 (2005).

45. Evans, D. Chemistry relating to the nickel enzymes CODH and ACS. Coordi- nation Chemistry Reviews 249, 1582–1595 (2005).

46. Richter, K., Schicklberger, M. & Gescher, J. Dissimilatory Reduction of Ex- tracellular Electron Acceptors in Anaerobic Respiration. Applied and Environ- mental Microbiology 78, 913–921 (2012).

157 47. Gros, G., Wittenberg, B. A. & Jue, T. Myoglobin’s old and new clothes: from molecular structure to function in living cells. Journal of Experimental Biology 213, 2713–2725 (2010).

48. Schechter, A. N. Hemoglobin research and the origins of molecular medicine. Blood 112, 3927–3938 (2008).

49. Yuan, Y., Tam, M. F., Simplaceanu, V. & Ho, C. New Look at Hemoglobin Allostery. Chemical Reviews 115, 1702–1724 (2015).

50. Honarmand Ebrahimi, K., Hagedoorn, P.-L. & Hagen, W. R. Unity in the Bio- chemistry of the Iron-Storage Proteins Ferritin and Bacterioferritin. Chemical Reviews 115, 295–326 (2015).

51. Watt, R. K. The many faces of the octahedral ferritin protein. BioMetals 24, 489–500 (2011).

52. Chiancone, E. & Ceci, P. The multifaceted capacity of Dps proteins to combat bacterial stress conditions: Detoxification of iron and hydrogen peroxide and DNA binding. Biochimica et Biophysica Acta (BBA) - General Subjects 1800, 798–805 (2010).

53. Ong, S. T., Shan Ho, J. Z., Ho, B. & Ding, J. L. Iron-withholding strategy in innate immunity. Immunobiology 211, 295–314 (2006).

54. Larade, K. Accumulation and translation of ferritin heavy chain transcripts following anoxia exposure in a marine invertebrate. Journal of Experimental Biology 207, 1353–1360 (2004).

55. Latunde-Dada, G. O. Ferroptosis: Role of lipid peroxidation, iron and fer- ritinophagy. Biochimica et Biophysica Acta (BBA) - General Subjects 1861, 1893–1900 (2017).

158 56. Theil, E. Ferritin: Structure, Gene Regulation, And Cellular Function In An- imals, Plants, And Microorganisms. Annual Review of Biochemistry 56, 289– 315 (1987).

57. Yan, L. et al. Magnetotactic bacteria, magnetosomes and their application. Microbiological Research 167, 507–519 (2012).

58. Andrews, S. C. The Ferritin-like superfamily: Evolution of the biological iron storeman from a rubrerythrin-like ancestor. Biochimica et Biophysica Acta (BBA) - General Subjects 1800, 691–705 (2010).

59. Lundin, D., Poole, a. M., Sjoberg, B.-M. & Hogbom, M. Use of Structural Phy- logenetic Networks for Classification of the Ferritin-like Superfamily. Journal of Biological Chemistry 287, 20565–20575 (2012).

60. Pfaffen, S. et al. A Diatom Ferritin Optimized for Iron Oxidation but Not Iron Storage. Journal of Biological Chemistry 290, 28416–28427 (2015).

61. Coulter, E. D. et al. Rubrerythrin-catalyzed substrate oxidation by dioxygen and hydrogen peroxide. Inorganica Chimica Acta 297, 231–241 (2000).

62. Kurtz, D. M. Avoiding high-valent iron intermediates: Superoxide reductase and rubrerythrin. Journal of Inorganic Biochemistry 100, 679–693 (2006).

63. Gomes, C. M., Le Gall, J., Xavier, A. V. & Teixeira, M. Could a Diiron- Containing Four-Helix-Bundle Protein Have Been a Primitive Oxygen Reduc- tase? ChemBioChem 2, 583–587 (2001).

64. Barynin, V. V. et al. Crystal Structure of Manganese Catalase from Lacto- bacillus plantarum. Structure 9, 725–738 (2001).

65. Murthy, M. R. N. et al. The structure and heme environment of beef liver catalase at 2.5 Å resolution. Acta Crystallographica Section A Foundations of Crystallography 37, C29–C29 (1981). 159 66. Whittaker, J. W. Non-heme manganese catalase – The ‘other’ catalase. Archives of Biochemistry and Biophysics 525, 111–120 (2012).

67. Shanklin, J., Guy, J. E., Mishra, G. & Lindqvist, Y. Desaturases: Emerging Models for Understanding Functional Diversification of Diiron-containing En- zymes. Journal of Biological Chemistry 284, 18559–18563 (2009).

68. Lindqvist, Y., Huang, W., Schneider, G. & Shanklin, J. Crystal structure of ￿9 stearoyl-acyl carrier protein desaturase from castor seed and its relationship to other di-iron proteins. The EMBO journal 15, 4081–92 (1996).

69. Wang, V. C.-C. et al. Alkane Oxidation: Methane Monooxygenases, Related Enzymes, and Their Biomimetics. Chemical Reviews 117, 8574–8621 (2017).

70. Baik, M. H., Newcomb, M., Friesner, R. a. & Lippard, S. J. Mechanistic stud- ies on the hydroxylation of methane by methane monooxygenase. Chemical Reviews 103, 2385–2419 (2003).

71. Shu, L. et al. An Fe2IVO2 Diamond Core Structure for the Key Intermediate Q of Methane Monooxygenase. Science 275, 515–518 (1997).

72. Lovell, T., Han, W.-G., Liu, T. & Noodleman, L. A Structural Model for the High-Valent Intermediate Q of Methane Monooxygenase from Broken- Symmetry Density Functional and Electrostatics Calculations. Journal of the American Chemical Society 124, 5890–5894 (2002).

73. Han, W.-G. & Noodleman, L. Structural model studies for the high-valent in- termediate Q of methane monooxygenase from broken-symmetry density func- tional calculations. Inorganica Chimica Acta 361, 973–986 (2008).

74. Tinberg, C. E. & Lippard, S. J. Revisiting the mechanism of dioxygen activation in soluble methane monooxygenase from M. capsulatus (Bath): Evidence for a

160 multi-step, proton-dependent reaction pathway. Biochemistry 48, 12145–12158 (2009).

75. Banerjee, R., Proshlyakov, Y., Lipscomb, J. D. & Proshlyakov, D. A. Structure of the key species in the enzymatic oxidation of methane to methanol. Nature 518, 431–434 (2015).

76. Buhler, B., Witholt, B., Hauer, B. & Schmid, A. Characterization and Ap- plication of Xylene Monooxygenase for Multistep Biocatalysis. Applied and Environmental Microbiology 68, 560–568 (2002).

77. Mitchell, K. H., Rogge, C. E., Gierahn, T. & Fox, B. G. Insight into the mech- anism of aromatic hydroxylation by toluene 4-monooxygenase by use of specif- ically deuterated toluene and p-xylene. Proceedings of the National Academy of Sciences 100, 3784–3789 (2003).

78. Acheson, J. F., Bailey, L. J., Brunold, T. C. & Fox, B. G. In-crystal reaction cycle of a toluene-bound diiron hydroxylase. Nature 544, 191–195 (2017).

79. Torres Pazmiño, D., Winkler, M., Glieder, A. & Fraaije, M. Monooxygenases as biocatalysts: Classification, mechanistic aspects and biotechnological appli- cations. Journal of Biotechnology 146, 9–24 (2010).

80. Kolberg, M., Strand, K. R., Graff, P. & Kristoffer Andersson, K. Structure, function, and mechanism of ribonucleotide reductases. Biochimica et Biophysica Acta (BBA) - Proteins and Proteomics 1699, 1–34 (2004).

81. Licht, S. S., Lawrence, C. C. & Stubbe, J. Class II Ribonucleotide Reductases Catalyze Carbon−Cobalt Bond Reformation on Every Turnover. Journal of the American Chemical Society 121, 7463–7468 (1999).

161 82. Yee, C. S., Chang, Ge, J., Nocera, D. G. & Stubbe, J. 2,3-Difluorotyrosine at Position 356 of Ribonucleotide Reductase R2: A Probe of Long-Range Proton- Coupled Electron Transfer. Journal of the American Chemical Society 125, 10506–10507 (2003).

83. Tomter, A. B. et al. Ribonucleotide reductase class I with different radical generating clusters. Coordination Chemistry Reviews 257, 3–26 (2013).

84. Logan, D. T. Closing the circle on ribonucleotide reductases. Nature Structural & Molecular Biology 18, 251–253 (2011).

85. Huang, M., Parker, M. J. & Stubbe, J. Choosing the Right Metal: Case Studies of Class I Ribonucleotide Reductases. Journal of Biological Chemistry 289, 28104–28111 (2014).

86. Eliasson, R., Pontis, E., Jordan, A. & Reichard, P. Allosteric Control of Three B 12 -dependent (Class II) Ribonucleotide Reductases. Journal of Biological Chemistry 274, 7182–7189 (1999).

87. Larsson, K.-M., Logan, D. T. & Nordlund, P. Structural Basis for Adenosyl- cobalamin Activation in AdoCbl-Dependent Ribonucleotide Reductases. ACS Chemical Biology 5, 933–942 (2010).

88. Mulliez, E., Ollagnier, S. & Fontecave, M. in Iron Metabolism 161–176 (Wiley- VCH Verlag GmbH, Weinheim, Germany, 1999).

89. Andersson, J., Bodevin, S., Westman, M., Sahlin, M. & Sjöberg, B.-M. Two Active Site Asparagines Are Essential for the Reaction Mechanism of the Class III Anaerobic Ribonucleotide Reductase from Bacteriophage T4. Journal of Biological Chemistry 276, 40457–40463 (2001).

162 90. Masalha, M., Borovok, I., Schreiber, R., Aharonowitz, Y. & Cohen, G. Analysis of Transcription of the Aerobic Class Ib and Anaerobic Class III Ribonucleotide Reductase Genes in Response to Oxygen. Journal of Bacteriology 183, 7260–7272 (2001).

91. Wei, Y. et al. A Chemically Competent Thiosulfuranyl Radical on the Es- cherichia coli Class III Ribonucleotide Reductase. Journal of the American Chemical Society 136, 9001–9013 (2014).

92. Ravi, N., Bollinger, J. M., Huynh, B. H., Stubbe, J. & Edmondson, D. E. Mech- anism of Assembly of the Tyrosyl Radical-Diiron(III) Cofactor of E. Coli Ri- bonucleotide Reductase: 1. Moessbauer Characterization of the Diferric Radical Precursor. Journal of the American Chemical Society 116, 8007–8014 (1994).

93. 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 . Journal of the American Chemical Society 116, 8015–8023 (1994).

94. Bollinger, J. M. et al. Mechanism of Assembly of the Tyrosyl Radical-Diiron(III) Cofactor of E. coli Ribonucleotide Reductase. 3. Kinetics of the Limiting Fe2+ Reaction by Optical, EPR, and Mossbauer Spectroscopies. Journal of the Amer- ican Chemical Society 116, 8024–8032 (1994).

95. Fontecave, M. & Gerez, C. Tyrosyl radicals and ribonucleotide reductase. Meth- ods in enzymology 348, 21–30 (2002).

96. Högbom, M. et al. Displacement of the tyrosyl radical cofactor in ribonucleotide reductase obtained by single-crystal high-field EPR and 1.4-Å x-ray data. Pro- ceedings of the National Academy of Sciences of the United States of America 100, 3209–3214 (2003).

163 97. Boal, A. K., Cotruvo, J. A., Stubbe, J. & Rosenzweig, A. C. Structural Basis for Activation of Class Ib Ribonucleotide Reductase. Science 329, 1526–1530 (2010).

98. Hogbom, M. The Radical Site in Chlamydial Ribonucleotide Reductase Defines a New R2 Subclass. Science 305, 245–248 (2004).

99. Hoganson, C. W., Sahlin, M., Sjöberg, B.-M. & Babcock, G. T. Electron magnetic resonance of the tyrosyl radical in ribonucleotide reductase from Escherichia coli. Journal of the American Chemical Society 118, 4672–4679 (1996).

100. Siegbahn, P. E. M., Eriksson, L., Himo, F. & Pavlov, M. Hydrogen atom trans- fer in ribonucleotide reductase (RNR). Journal of Physical Chemistry B 102, 10622–10629 (1998).

101. Gräslund, A. & Ehrenberg, A. Tyrosyl radical, diiron center and enzyme mech- anism in ribonucleotide reductase. Applied Magnetic Resonance 31, 447–455 (2007).

102. Petersson, L., Graslund, A., Ehrenberg, A., Sjöberg, B. M. & Reichard, P. The iron center in ribonucleotide reductase from Escherichia coli. Journal of Biological Chemistry 255, 6706–6712 (1980).

103. Sjöberg, B. M., Reichard, P., Graslund, A. & Ehrenberg, A. The Tyrosine Free Radical in Ribonucleotide Reductase from Escherichia coli. Journal of Biological Chemistry 253, 6863–6865 (1978).

104. Bleifuss, G. et al. Tryptophan and tyrosine radicals in ribonucleotide reductase: A comparative high-field EPR study at 94 GHz. Biochemistry 40, 15362–15368 (2001).

164 105. Uhlin, U. & Eklund, H. Structure of ribonucleotide reductase protein R1. Na- ture 370, 533–539 (1994).

106. Warren, J. J., Ener, M. E., Vlček, A., Winkler, J. R. & Gray, H. B. Electron hopping through proteins. Coordination Chemistry Reviews 256, 2478–2487 (2012).

107. Minnihan, E. C., Nocera, D. G. & Stubbe, J. Reversible, Long-Range Radical Transfer in E. coli Class Ia Ribonucleotide Reductase. Accounts of Chemical Research 46, 2524–2535 (2013).

108. Sturgeon, B. E. et al. Reconsideration of X, the Diiron Intermediate Formed during Cofactor Assembly in E. coli Ribonucleotide Reductase. Journal of the American Chemical Society 118, 7551–7557 (1996).

109. Ormö, M. et al. Engineering of the iron site in ribonucleotide reductase to a self-hydroxylating monooxygenase. The Journal of biological chemistry 267, 8711–4 (1992).

110. Persson, B. O. et al. Iron ligand mutants in protein R2 of Escherichia coli ribonucleotide reductase. Journal of Biological Inorganic Chemistry 1, 247– 256 (1996).

111. Moënne-Loccoz, P., Baldwin, J., Ley, B. A., Loehr, T. M. & Bollinger, J. M. O2 activation by non-heme diiron proteins: Identification of a symmetric µ- 1,2-peroxide in a mutant of ribonucleotide reductase. Biochemistry 37, 14659– 14663 (1998).

112. Baldwin, J. et al. Rational reprogramming of the R2 subunit of Escherichia coli ribonucleotide reductase into a self-hydroxylating monooxygenase. Journal of the American Chemical Society 123, 7017–7030 (2001).

165 113. Cotruvo, J. A. & Stubbe, J. An Active Dimanganese(III)−Tyrosyl Radical Cofactor in Escherichia coli Class Ib Ribonucleotide Reductase. Biochemistry 49, 1297–1309 (2010).

114. Lofstad, M., Gudim, I., Hammerstad, M., Røhr, Å. K. & Hersleth, H.-P. Acti- vation of the Class Ib Ribonucleotide Reductase by a Flavodoxin Reductase in Bacillus cereus. Biochemistry 55, 4998–5001 (2016).

115. Monje-Casas, F., Jurado, J., Prieto-Álamo, M.-J., Holmgren, A. & Pueyo, C. Expression Analysis of the nrdHIEF Operon from Escherichia coli. Journal of Biological Chemistry 276, 18031–18037 (2001).

116. Gon, S., Faulkner, M. J. & Beckwith, J. In Vivo Requirement for Glutaredox- ins and Thioredoxins in the Reduction of the Ribonucleotide Reductases of Escherichia coli. Antioxidants & Redox Signaling 8, 735–742 (2006).

117. Zhang, Y. & Stubbe, J. Bacillus subtilis Class Ib Ribonucleotide Reductase Is a Dimanganese(III)-Tyrosyl Radical Enzyme. Biochemistry 50, 5615–5623 (2011).

118. Jiang, W. et al. A Manganese(IV)/Iron(III) Cofactor in Chlamydia trachomatis Ribonucleotide Reductase. Science 316, 1188–1191 (2007).

119. Roshick, C., Iliffe-Lee, E. R. & McClarty, G. Cloning and Characterization of Ribonucleotide Reductase from Chlamydia trachomatis. Journal of Biological Chemistry 275, 38111–38119 (2000).

120. Voevodskaya, N., Lendzian, F. & Gräslund, A. A stable FeIII–FeIV replace- ment of tyrosyl radical in a class I ribonucleotide reductase. Biochemical and Biophysical Research Communications 330, 1213–1216 (2005).

166 121. Voevodskaya, N. et al. Chlamydial ribonucleotide reductase: Tyrosyl radical function in catalysis replaced by the FeIII-FeIV cluster. Proceedings of the Na- tional Academy of Sciences 103, 9850–9854 (2006).

122. Voevodskaya, N., Lendzian, F., Ehrenberg, A. & Gräslund, A. High catalytic activity achieved with a mixed manganese-iron site in protein R2 of Chlamydia ribonucleotide reductase. FEBS Letters 581, 3351–3355 (2007).

123. Schenk, G. et al. A Purple Acid Phosphatase from Sweet Potato Contains an Antiferromagnetically Coupled Binuclear Fe-Mn Center. Journal of Biological Chemistry 276, 19084–19088 (2001).

124. Jiang, W., Hoffart, L. M., Krebs, C. & Bollinger, J. M. A Manganese(IV)/Iron(IV) Intermediate in Assembly of the Manganese(IV)/Iron(III) Cofactor of Chlamy- dia trachomatis Ribonucleotide Reductase. Biochemistry 46, 8709–8716 (2007).

125. Livada, J. et al. Direct Measurement of the Radical Translocation Distance in the Class I Ribonucleotide Reductase from Chlamydia trachomatis. The Journal of Physical Chemistry B 119, 13777–13784 (2015).

126. Eriksson, L. A. Sulfinylimine Radical in Azido-CDP- and Azido-UDP-Inhibited Ribonucleotide Reductase. Journal of the American Chemical Society 120, 8051–8054 (1998).

127. Fritscher, J. et al. Structure of the Nitrogen-Centered Radical Formed during Inactivation of E. coli Ribonucleotide Reductase by 2‘-Azido-2‘-deoxyuridine- 5‘-diphosphate: Trapping of the 3‘-Ketonucleotide. Journal of the American Chemical Society 127, 7729–7738 (2005).

128. Meier, B., Parak, F., Desideri, A. & Rotilio, G. Comparative stability studies on the iron and manganese forms of the cambialistic superoxide dismutase from Propionibacterium shermanii. FEBS Letters 414, 122–124 (1997).

167 129. Aguirre, J. D. & Culotta, V. C. Battles with Iron: Manganese in Oxidative Stress Protection. Journal of Biological Chemistry 287, 13541–13548 (2012).

130. Andersson, C. S. & Högbom, M. A Mycobacterium tuberculosis ligand-binding Mn/Fe protein reveals a new cofactor in a remodeled R2-protein scaffold. Pro- ceedings of the National Academy of Sciences of the United States of America 106, 5633–5638 (2009).

131. Schmidt, F. Complementary Analysis of the Mycobacterium tuberculosis Pro- teome by Two-dimensional Electrophoresis and Isotope-coded Affinity Tag Tech- nology. Molecular & Cellular Proteomics 3, 24–42 (2003).

132. Griese, J. J. et al. Direct observation of structurally encoded metal discrimina- tion and ether bond formation in a heterodinuclear metalloprotein. Proceedings of the National Academy of Sciences 110, 17189–17194 (2013).

133. Andersson, M. E. et al. The Crystal Structure of an Azide Complex of the Di- ferrous R2 Subunit of Ribonucleotide Reductase Displays a Novel Carboxylate Shift with Important Mechanistic Implications for Diiron-Catalyzed Oxygen Activation. Journal of the American Chemical Society 121, 2346–2352 (1999).

134. Sazinsky, M. H. & Lippard, S. J. Correlating Structure with Function in Bacte- rial Multicomponent Monooxygenases and Related Diiron Proteins. Accounts of Chemical Research 39, 558–566 (2006).

135. Griese, J. J. et al. Structural basis for oxygen activation at a heterodinuclear manganese/iron cofactor. Journal of Biological Chemistry 290, 25254–25272 (2015).

136. Shafaat, H. S. et al. Electronic structural flexibility of heterobimetallic mn/fe cofactors: R2lox and R2c proteins. Journal of the American Chemical Society 136, 13399–409 (2014).

168 137. Miller, E. K., Trivelas, N. E., Maugeri, P. T., Blaesi, E. J. & Shafaat, H. S. Time-Resolved Investigations of Heterobimetallic Cofactor Assembly in R2lox Reveal Distinct Mn/Fe Intermediates. Biochemistry 56, 3369–3379 (2017).

138. Cox, N. et al. Electronic structure of a weakly antiferromagnetically coupled MnIIMnIII Model relevant to manganese proteins: A combined EPR, 55Mn- ENDOR, and DFT study. Inorganic Chemistry 50, 8238–8251 (2011).

139. Link, A. J. et al. Direct analysis of protein complexes using mass spectrometry. Nature Biotechnology 17, 676–682 (1999).

140. Meisburger, S. P. & Ando, N. Correlated Motions from Crystallography beyond Diffraction. Accounts of Chemical Research 50, 580–583 (2017).

141. Scheraga, H. A., Khalili, M. & Liwo, A. Protein-Folding Dynamics: Overview of Molecular Simulation Techniques. Annual Review of Physical Chemistry 58, 57–83 (2007).

142. Kukushkin, A. K. & Jalkanen, K. J. Role of quantum chemical calculations in molecular biophysics with a historical perspective. Theoretical Chemistry Accounts 125, 121–144 (2010).

143. Zhang, J. et al. Interfacial electrochemical electron transfer in biology - Towards the level of the single molecule. FEBS Letters 586, 526–535 (2012).

144. Armstrong, F. in Metalloproteins 205–222 (CRC Press, 2015).

145. Hendrickson, W. Determination of macromolecular structures from anomalous diffraction of synchrotron radiation. Science 254, 51–58 (1991).

146. Planck, M. Ueber das Gesetz der Energieverteilung im Normalspectrum. An- nalen der Physik 309, 553–563 (1901).

147. Pilipenko, V. ULF waves on the ground and in space. Journal of Atmospheric and Terrestrial Physics 52, 1193–1209 (1990). 169 148. Mittermaier, a. New Tools Provide New Insights in NMR Studies of Protein Dynamics. Science 312, 224–228 (2006).

149. Mittermaier, A. K. & Kay, L. E. Observing biological dynamics at atomic resolution using NMR. Trends in biochemical sciences 34, 601–11 (2009).

150. Kleckner, I. R. & Foster, M. P. An introduction to NMR-based approaches for measuring protein dynamics. Biochimica et Biophysica Acta (BBA) - Proteins and Proteomics 1814, 942–968 (2011).

151. Kay, L. E. NMR studies of protein structure and dynamics. Journal of Magnetic Resonance 213, 477–491 (2011).

152. Harris, R. K., Becker, E. D., Cabral de Menezes, S. M., Goodfellow, R. & Granger, P. NMR nomenclature. Nuclear spin properties and conventions for chemical shifts(IUPAC Recommendations 2001). Pure and Applied Chemistry 73, 1795–1818 (2001).

153. Bogey, M. Microwave absorption spectroscopy in the ν3 states of OCS and N2O through energy transfer from N2*. J. Phys. B: At. Mol. Phys. 8, 1934–1938 (1975).

154. Bowman, W. C., Herbst, E. & De Lucia, F. C. Millimeter and submillimeter spectrum of NO+. The Journal of Chemical Physics 77, 4261–4262 (1982).

155. Harada, K., Tanaka, K., Tanaka, T., Nanbu, S. & Aoyagi, M. Millimeter-wave spectroscopy of the internal-rotation band of the He–HCN complex and the intermolecular potential energy surface. The Journal of Chemical Physics 117, 7041–7050 (2002).

156. Sahu, I. D., McCarrick, R. M. & Lorigan, G. A. Use of electron paramagnetic resonance to solve biochemical problems. Biochemistry 52, 5967–5984 (2013).

170 157. Rieger, P. H. Electron paramagnetic resonance studies of low-spin d5 transition metal complexes. Coordination Chemistry Reviews 135-136, 203–286 (1994).

158. Walker, F. Magnetic spectroscopic (EPR, ESEEM, Mössbauer, MCD and NMR) studies of low-spin ferriheme centers and their corresponding heme proteins. Coordination Chemistry Reviews 185-186, 471–534 (1999).

159. Hagen, W. R. High-frequency EPR of transition ion complexes and metallo- proteins. Coordination Chemistry Reviews 190-192, 209–229 (1999).

160. Wahba, H. M. et al. Structural and Biochemical Characterization of a Copper- Binding Mutant of the Organomercurial Lyase MerB: Insight into the Key Role of the Active Site Aspartic Acid in Hg–Carbon Bond Cleavage and Metal Binding Specificity. Biochemistry 55, 1070–1081 (2016).

161. Reijerse, E., Lendzian, F., Isaacson, R. & Lubitz, W. A tunable general purpose Q-band resonator for CW and pulse EPR/ENDOR experiments with large sample access and optical excitation. Journal of Magnetic Resonance 214, 237–243 (2012).

162. Rapatskiy, L. et al. Detection of the Water-Binding Sites of the Oxygen- Evolving Complex of Photosystem II Using W-Band 17 O Electron–Electron Double Resonance-Detected NMR Spectroscopy. Journal of the American Chem- ical Society 134, 16619–16634 (2012).

163. Stoll, S. et al. Hydrogen Bonding of Tryptophan Radicals Revealed by EPR at 700 GHz. Journal of the American Chemical Society 133, 18098–18101 (2011).

164. Markelz, A. G. Terahertz Dielectric Sensitivity to Biomolecular Structure and Function. IEEE Journal of Selected Topics in Quantum Electronics 14, 180– 190 (2008).

171 165. Born, B. & Havenith, M. Terahertz Dance of Proteins and Sugars with Water. Journal of Infrared, Millimeter, and Terahertz Waves, 1245–1254 (2009).

166. Falconer, R. J. & Markelz, A. G. Terahertz Spectroscopic Analysis of Peptides and Proteins. Journal of Infrared, Millimeter, and Terahertz Waves 33, 973– 988 (2012).

167. Woods, K. N. THz time scale structural rearrangements and binding modes in lysozyme-ligand interactions. Journal of Biological Physics 40, 121–137 (2014).

168. Sutcliffe, M. J. et al. Hydrogen tunnelling in enzyme-catalysed H-transfer re- actions: flavoprotein and quinoprotein systems. Philosophical Transactions of the Royal Society B: Biological Sciences 361, 1375–1386 (2006).

169. O’Ferrall, R. A. M. A pictorial representation of zero-point energy and tun- nelling contributions to primary hydrogen isotope effects. Journal of Physical Organic Chemistry 23, 572–579 (2010).

170. Layfield, J. P. & Hammes-Schiffer, S. Hydrogen Tunneling in Enzymes and Biomimetic Models. Chemical Reviews 114, 3466–3494 (2014).

171. Nishiguchi, S., Goto, Y. & Takahashi, S. Solvation and Desolvation Dynamics in Apomyoglobin Folding Monitored by Time-resolved . Journal of Molecular Biology 373, 491–502 (2007).

172. Rathod, R. et al. Side-chain specific isotopic labeling of proteins for infrared structural biology: The case of ring-D4-tyrosine isotope labeling of photoactive yellow protein. Protein Expression and Purification 85, 125–132 (2012).

173. Shafaat, H. S., Weber, K., Petrenko, T., Neese, F. & Lubitz, W. Key hydride vi- brational modes in [NiFe] hydrogenase model compounds studied by resonance Raman spectroscopy and density functional calculations. Inorganic Chemistry 51, 11787–97 (2012). 172 174. Hidalgo, R., Ash, P. A., Healy, A. J. & Vincent, K. A. Infrared Spectroscopy

During Electrocatalytic Turnover Reveals the Ni-L Active Site State During H2 Oxidation by a NiFe Hydrogenase. Angewandte Chemie International Edition 54, 7110–7113 (2015).

175. Edelhoch, H. Spectroscopic Determination of Tryptophan and Tyrosine in Pro- teins. Biochemistry 6, 1948–1954 (1967).

176. Qiu, W. et al. Protein surface hydration mapped by site-specific mutations. Proceedings of the National Academy of Sciences of the United States of America 103, 13979–13984 (2006).

177. Halle, B. & Nilsson, L. Does the dynamic report on slow protein hydration dynamics? Journal of Physical Chemistry B 113, 8210–8213 (2009).

178. Zhong, D., Pal, S. K. & Zewail, A. H. Biological water: A critique. Chemical Physics Letters 503, 1–11 (2011).

179. Röntgen, W. C. On a New Kind of Rays. Science 3, 227–231 (1896).

180. Hummer, A. A. & Rompel, A. in Biomolecular Spectroscopy: Advances from Integrating Experiments and Theory 1st, 257–305 (Elsevier Inc., 2013).

181. Schneck, E. et al. Atom-scale depth localization of biologically important chem- ical elements in molecular layers. Proceedings of the National Academy of Sci- ences 113, 9521–9526 (2016).

182. Giovanelli, R. & Orefice, A. Physical discussion of the Mössbauer effect. Physica B: Condensed Matter 293, 155–163 (2000).

183. King, J. T. & Kubarych, K. J. Site-specific coupling of hydration water and protein flexibility studied in solution with ultrafast 2D-IR spectroscopy. Journal of the American Chemical Society 134, 18705–18712 (2012).

173 184. Hill, R. E., Hunt, N. T. & Hirst, J. D. in Biomolecular Spectroscopy: Advances from Integrating Experiments and Theory 1st ed., 1–36 (Elsevier Inc., 2013).

185. Das, R. S. & Agrawal, Y. Raman spectroscopy: Recent advancements, tech- niques and applications. Vibrational Spectroscopy 57, 163–176 (2011).

186. Scheidt, W. R., Durbin, S. M. & Sage, J. T. Nuclear resonance vibrational spectroscopy–NRVS. Journal of inorganic biochemistry 99, 60–71 (2005).

187. Liebel, M., Schnedermann, C., Wende, T. & Kukura, P. Principles and Ap- plications of Broadband Impulsive Vibrational Spectroscopy. The Journal of Physical Chemistry A 119, 9506–9517 (2015).

188. Maiuri, M. et al. Low frequency dynamics of the nitrogenase MoFe protein via femtosecond pump probe spectroscopy — Observation of a candidate promot- ing vibration. Journal of Inorganic Biochemistry (2015).

189. Loehr, T. M. in Spectroscopic Methods in Bioinorganic Chemistry 5, 136–153 (1998).

190. Raman, C. V. A Change of Wave-length in Light Scattering. Nature 121, 619– 619 (1928).

191. Raman, C. V. A new radiation. Indian J. Phys. 2, 387–398 (1928).

192. Harris, D. C. & Bertolucci, M. D. Symmetry and Spectroscopy: An Introduction to Vibrational and Electronic Spectroscopy (Dover Publications, Inc., New York, NY, 1989).

193. Maiman, T. H. Stimulated Optical Radiation in . Nature 187, 493–494 (1960).

194. Myers, A. B. & Mathies, R. A. in Biological Applications of Raman Spec- troscopy: Vol. 2 - Resonance Raman Spectra of Polyenes and Aromatics (ed Spiro, T. G.) 1–58 (John Wiley & Sons, Inc, 1987). 174 195. McCreery, R. L. in Handbook of Vibrational Spectroscopy: Vol 1 (eds Chalmers, J. M. & Griffiths, P. R.) 920–932 (John Wiley & Sons, Ltd, New York,NY, 2002).

196. Bohren, C. F. & Huffman, D. R. Absorption and Scattering of Light by Small Particles (John Wiley & Sons, Inc, New York, NY, 1983).

197. Landau, L. D. & Lifshitz, E. M. Mechanics (Pergamon Press Ltd., Oxford, 1960).

198. Califano, S. in Vibrational States 266–302 (John Wiley & Sons, Ltd, 1976).

199. Kramers, H. A. & Heisenberg, W. Über die Streuung von Strahlung durch Atome. Zeitschrift für Physik 31, 681–708 (1925).

200. Lee, S.-Y. & Heller, E. J. Time-Dependent Theory of Raman Scattering. J. Chem. Phys. 71, 4777–4788 (1979).

201. Dirac, P. A. M. The Quantum Theory of Dispersion. Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences 114, 710–728 (1927).

202. Dirac, P. A. M. The Quantum Theory of the Emission and Absorption of Radiation. Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences 114, 243–265 (1927).

203. Tannor, D. J. Introduction to Quantum Mechanics: A time-dependent perspec- tive 1–13 (University Science Books, 2007).

204. Keresztury, G. in Handbook of Vibrational Spectroscopy: Vol 1 (eds Chalmers, J. M. & Griffiths, P. R.) 71–87 (John Wiley & Sons, Ltd, New York, NY, 2002).

205. McHale, J. L. in Handbook of Vibrational Spectroscopy: Vol 1 (eds Chalmers, J. M. & Griffiths, P. R.) 534–556 (John Wiley & Sons, Ltd, New York,NY, 2002). 175 206. Long, D. A. The Raman Effect: A Unified Treatment of the Theory of Raman Scattering by Molecules (John Wiley & Sons, Ltd, Chichester, UK, 2002).

207. Born, M. & Oppenheimer, R. Zur Quantentheorie der Molekeln. Annalen der Physik 389, 457–484 (1927).

208. Albrecht, A. C. On the Theory of Raman Intensities. The Journal of Chemical Physics 34, 1476 (1961).

209. McHale, J. L. Molecular Spectroscopy 1–463 (Prentice-Hall, 1999).

210. Ferraro, J. R., Nakamoto, K. & Brown, C. W. Introductory Raman Spectroscopy 1–434 (Academic Press, 2003).

211. Shiemke, A. K., Loehr, T. M. & Sanders-Loehr, J. Resonance Raman study of the µ-oxo-bridged binuclear iron center in oxyhemerythrin. Journal of the American Chemical Society 106, 4951–4956 (1984).

212. Sjoeberg, B. M., Sanders-Loehr, J. & Loehr, T. M. Identification of a hydrox- ide ligand at the iron center of ribonucleotide reductase by resonance Raman spectroscopy. Biochemistry 26, 4242–4247 (1987).

213. Fox, B. G., Shanklin, J., Ai, J., Loehr, T. M. & Sanders-Loehr, J. Resonance Raman evidence for an Fe-O-Fe center in stearoyl-ACP desaturase. Primary sequence identity with other diiron-oxo proteins. Biochemistry 33, 12776–12786 (1994).

214. Yachandra, V. K. et al. Resonance Raman spectra of spinach ferredoxin and adrenodoxin and of analog complexes. Journal of the American Chemical So- ciety 105, 6462–6469 (1983).

176 215. Han, S., Czernuszewicz, R. S., Kimura, T., Adams, M. W. W. & Spiro, T. G. Fe2S2 protein resonance Raman spectra revisited: structural variations among adrenodoxin, ferredoxin, and red paramagnetic protein. Journal of the Ameri- can Chemical Society 111, 3505–3511 (1989).

216. Benko, B. & Yu, N. T. Resonance Raman studies of nitric oxide binding to ferric and ferrous hemoproteins: detection of Fe(III)–NO stretching, Fe(III)– N–O bending, and Fe(II)–N–O bending vibrations. Proceedings of the National Academy of Sciences of the United States of America 80, 7042–6 (1983).

217. Hu, S., Morris, I. & Singh, J. Complete assignment of cytochrome c resonance Raman spectra via enzymic reconstitution with isotopically labeled hemes. Journal of the American Chemical Society 115, 12446–12458 (1993).

218. Macdonald, I. D., Sligar, S. G., Christian, J. F., Unno, M. & Champion, P. M. Identification of the Fe-O-O bending mode in oxycytochrome P450cam byRes- onance Raman spectroscopy. Journal of the American Chemical Society 121, 376–380 (1999).

219. Ho, R. Y. N. et al. Resonance Raman studies of the iron(II)-α-keto acid chro- mophore in model and enzyme complexes. Journal of the American Chemical Society 123, 5022–5029 (2001).

220. Dos Santos, I. F., Hutchinson, I., Ingley, R., Edwards, H. G. & de Faria, D. L. Amorphous iron oxides investigated by portable and bench-top Raman spec- trometers: An extraterrestrial analytical perspective. Vibrational Spectroscopy 87, 20–26 (2016).

221. McCamant, D. W., Kukura, P. & Mathies, R. A. Femtosecond Stimulated Raman Study of Excited-State Evolution in Bacteriorhodopsin. The Journal of Physical Chemistry B 109, 10449–10457 (2005).

177 222. Kukura, P., McCamant, D. W. & Mathies, R. A. Femtosecond Stimulated Ra- man Spectroscopy. Annual Review of Physical Chemistry 58, 461–488 (2007).

223. Agarwalla, B. K., Ando, H., Dorfman, K. E. & Mukamel, S. Stochastic Liou- ville equations for femtosecond stimulated Raman spectroscopy. The Journal of Chemical Physics 142, 024115 (2015).

224. Spiro, T. G. & Czernuszewicz, R. S. Resonance Raman spectroscopy of metal- loproteins. Methods in Enzymology 246, 416–460 (1995).

225. Pelletier, M. J. Raman Spectroscopy Using an Echelle Spectrograph with CCD Detection. Applied Spectroscopy 44, 1699–1705 (1990).

226. Yang, S. et al. Laser wavelength dependence of background fluorescence in Ra- man spectroscopic analysis of synovial fluid from symptomatic joints. Journal of Raman Spectroscopy 44, 1089–1095 (2013).

227. Lupoi, J. S. & Smith, E. A. Characterization of Woody and Herbaceous Biomasses Lignin Composition with 1064 nm Dispersive Multichannel Raman Spectroscopy. Applied Spectroscopy 66, 903–910 (2012).

228. Jernshøj, K. D., Hassing, S. & Olsen, L. F. A combination of and polarized resonance Raman scattering applied in the study of Arenicola Marina extracellular hemoglobin. The Journal of Chemical Physics 139, 065104 (2013).

229. Kim, K., Kim, K. L. & Shin, K. S. Visible-Light Response of 4-Aminobenzenethiol and 4,4￿-Dimercaptoazobenzene Silver Salts. The Journal of Physical Chemistry C 117, 5975–5981 (2013).

230. Kilmer, J., Iadevaia, A. & Yin, Y. Laser sources for Raman spectroscopy in Security (eds Dubinskii, M. & Post, S. G.) 8039 (2011), 803914.

178 231. Spiller, E. Coherence effects from visible light to X-rays. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, De- tectors and Associated Equipment 347, 161–169 (1994).

232. Gover, A. & Dyunin, E. Coherence Limits of Free Electron Lasers. IEEE Jour- nal of Quantum Electronics 46, 1511–1517 (2010).

233. Singh, N. & Vora, H. S. On the coherence measurement of a narrow bandwidth . Applied Physics B 110, 483–489 (2013).

234. Körber, K. & Kulik, C. in Excimer Laser Technology 33–40 (Springer-Verlag, Berlin/Heidelberg, 2005).

235. Silvast, W. T. Laser Fundamentals 2nd (Cambridge University Press, Cam- bridge, UK, 2004).

236. Pathria, R. K. Statistic Mechanics 2nd (Elsevier Ltd., Oxford, UK, 1996).

237. Diels, J.-C. & Rudolph, W. in Ultrafast Laser Pulse Phenomena 2nd, 1–60 (Academic Press, San Diego, CA, 2006).

238. Diels, J.-C. & Rudolph, W. in Ultrafast Laser Pulse Phenomena 2nd, 277–341 (Academic Press, San Diego, CA, 2006).

239. Thompson, D. E., McMullen, J. D. & Anderson, D. B. Second‐harmonic gener- ation in GaAs ”stack of plates” using high‐power CO 2 laser radiation. Applied Physics Letters 29, 113–115 (1976).

240. Kozlovsky, W. J., Nabors, C. D. & Byer, R. L. Second-harmonic generation of a continuous-wave diode-pumped Nd:YAG laser using an externally resonant cavity. Optics Letters 12, 1014 (1987).

241. Moore, J., Tomes, M., Carmon, T. & Jarrahi, M. Continuous-wave ultravio- let emission through fourth-harmonic generation in a whispering- gallery res- onator. Optics Express 19, 24139–24146 (2011). 179 242. Doig, S. J. & Prendergast, F. G. Continuously Tuneable, Quasi-Continuous- Wave Source for Ultraviolet Resonance Raman Spectroscopy. Applied Spec- troscopy 49, 247–252 (1995).

243. Zhao, X., Tengroth, C., Chen, R., Simpson, W. R. & Spiro, T. G. Time‐resolved Raman spectroscopy with a tunable ultraviolet kilohertz nanosecond laser. Journal of Raman Spectroscopy 30, 773–776 (1999).

244. Zhao, X., Chen, R., Tengroth, C. & Spiro, T. G. Solid-State Tunable kHz Ultraviolet Laser for Raman Applications. Applied Spectroscopy 53, 1200–1205 (1999).

245. Balakrishnan, G., Hu, Y., Nielsen, S. B. & Spiro, T. G. Tunable kHz Deep Ultraviolet (193–210 nm) Laser for Raman Applications. Applied Spectroscopy 59, 776–781 (2005).

246. Franken, P. A., Hill, A. E., Peters, C. W. & Weinreich, G. Generation of Optical Harmonics. Physical Review Letters 7, 118–119 (1961).

247. Huang, D. Formula for Coulomb effect on the nonlinear optical responses in quantum wells. Physical Review B 53, 13645–13650 (1996).

248. Bass, M., Franken, P. a., Ward, J. F. & Weinreich, G. Optical Rectification. Physical Review Letters 9, 446–448 (1962).

249. Fiore, A., Berger, V., Rosencher, E., Bravetti, P. & Nagle, J. Phase matching using an isotropic nonlinear optical material. Nature 391, 463–466 (1998).

250. Zhang, W., Yu, H., Wu, H. & Halasyamani, P. S. Phase-Matching in Nonlin- ear Optical Compounds: A Materials Perspective. Chemistry of Materials 29, 2655–2668 (2017).

180 251. Grace, E., Butcher, A., Monroe, J. & Nikkel, J. A. Index of refraction, Rayleigh scattering length, and Sellmeier coefficients in solid and liquid argon and xenon. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment 867, 204–208 (2017).

252. Boeuf, N. Calculating characteristics of noncollinear phase matching in uniaxial and biaxial crystals. Optical Engineering 39, 1016 (2000).

253. Shao, Z. Refractive indices for extraordinary waves in uniaxial crystals. Physical Review E 52, 1043–1048 (1995).

254. Chen, J., Zheng, Y., An, N. & Chen, X. Noncollinear third-harmonic generation with large angular acceptance by noncritical phase matching in KDP crystal. Optics Letters 40, 4484 (2015).

255. Shriver, D. F. & Dunn, J. B. The Backscattering Geometry for Raman Spec- troscopy of Colored Materials. Applied Spectroscopy 60, 319–323 (1974).

256. Czernuszewicz, R. S. & Johnson, M. K. A Simple Low-temperature Cryostat for Resonance Raman Studies of Frozen Protein Solutions. Appl. Spectrosc. 37, 297–298 (1983).

257. Hong, Z. & Asher, S. a. Dependence of Raman and Resonance Raman Inten- sities on Sample Self-Absorption. Applied Spectroscopy 69, 75–83 (2015).

258. Czernuszewicz, R. S. Closed-cycle Refrigerator Solution and Rotating Solid Sample Cells for Anaerobic Resonance Raman Spectroscopy. Applied Spec- troscopy 40, 571–573 (1986).

259. Russell, M. P., Vohník, S. & Thomas, G. J. Design and performance of an ultraviolet resonance Raman spectrometer for proteins and nucleic acids. Bio- physical journal 68, 1607–1612 (1995).

181 260. Grzyska, P. K., Appelman, E. H., Hausinger, R. P. & Proshlyakov, D. A. Insight into the mechanism of an iron dioxygenase by resolution of steps following the FeIVboxHO species. Proceedings of the National Academy of Sciences 107, 3982–3987 (2010).

261. Kempe, M. & Rudolph, W. Impact of chromatic and spherical aberration on the focusing of ultrashort light pulses by lenses. Optics letters 18, 137–9 (1993).

262. Pruss, C., Garbusi, E. & Osten, W. Testing Aspheres. Optics and Photonics News 19, 24–29 (2008).

263. Schulz, B. et al. Fully reflective deep ultraviolet to near infrared spectrometer and entrance optics for resonance Raman spectroscopy. Review of Scientific Instruments 76, 073107 (2005).

264. Flaugh, P. L., O’Donnell, S. E. & Asher, S. A. Development of a New Opti- cal Wavelength Rejection Filter: Demonstration of its Utility in Raman Spec- troscopy. Appl. Spectrosc. 38, 847–850 (1984).

265. Kaminaka, S. & Mathies, R. A. High-Throughput Large-Aperture Prism Pre- filter for Ultraviolet Resonance Raman Spectroscopy. Applied Spectroscopy 52, 469–473 (1998).

266. Strong, J. Resolving Power Limitations of Grating and Prism Spectrometers. Journal of the Optical Society of America 39, 320 (1949).

267. Chowdhuri, M. B. et al. Spectroscopic comparison between 1200grooves∕mm ruled and holographic gratings of a flat-field spectrometer and its absolute sensitivity calibration using bremsstrahlung continuum. Review of Scientific Instruments 78, 023501 (2007).

268. Labeyrie, A. & Flamand, J. Spectrographic performance of holographically made diffraction gratings. Optics Communications 1, 5–8 (1969). 182 269. Diels, J.-C. & Rudolph, W. in Ultrafast Laser Pulse Phenomena 2nd, 61–142 (Academic Press, San Diego, CA, 2006).

270. Boyle, W. S. & Smith, G. E. Charge coupled semiconductor devices. Bell Syst Tech J 49, 587–593 (1970).

271. Millikan, R. A. A Direct Determination of ”h”. Physical Review 4, 73–75 (1914).

272. Rieser, U., Krbetschek, M. R. & Stolz, W. CCD-Camera Based High Sensitivity TL/OSL-Spectrometer. Radiation Measurements 23, 523–528 (1994).

273. Kujawa, S. & Krahl, D. Performance of a low-noise CCD camera adapted to a transmission electron microscope. Ultramicroscopy 46, 395–403 (1992).

274. Wu, D. et al. Structural basis of ultraviolet-B perception by UVR8. Nature 484, 214–219 (2012).

275. Jenkins, G. I. Structure and function of the UV-B photoreceptor UVR8. Cur- rent Opinion in Structural Biology 29, 52–57 (2014).

276. Dexheimer, S. et al. Femtosecond impulsive excitation of nonstationary vibra- tional states in bacteriorhodopsin. Chemical Physics Letters 188, 61–66 (1992).

277. Luecke, H., Schobert, B., Richter, H. T., Cartailler, J. P. & Lanyi, J. K. Struc- ture of bacteriorhodopsin at 1.55 A resolution. Journal of Molecular Biology 291, 899–911 (1999).

278. Jung, A. et al. Structure of a bacterial BLUF photoreceptor: insights into blue light-mediated signal transduction. Proceedings of the National Academy of Sciences of the United States of America 102, 12350–12355 (2005).

279. Conrad, K. S., Manahan, C. C. & Crane, B. R. Photochemistry of flavoprotein light sensors. Nature chemical biology 10, 801–9 (2014).

183 280. Imamoto, Y. & Kataoka, M. Structure and photoreaction of photoactive yellow protein, a structural prototype of the PAS domain superfamily. Photochemistry and photobiology 83, 40–49 (2007).

281. Johansson, L. C. et al. Structure of a photosynthetic reaction centre determined by serial femtosecond crystallography. Nature Communications 4, 2911 (2013).

282. McCusker, J. K., Toftlund, H., Rheingold, A. L. & Hendrickson, D. N. Ligand

5 1 conformational changes affecting T2→ A1 intersystem crossing in a ferrous complex. Journal of the American Chemical Society 115, 1797–1804 (1993).

283. Monat, J. E. & McCusker, J. K. Femtosecond Excited-State Dynamics of an Iron(II) Polypyridyl Solar Cell Sensitizer Model. Journal of the American Chemical Society 122, 4092–4097 (2000).

284. Cannizzo, A. et al. Light-induced spin crossover in Fe(II)-based complexes: The full photocycle unraveled by ultrafast optical and X-ray spectroscopies. Coordination Chemistry Reviews 254, 2677–2686 (2010).

2+ 285. Auböck, G. & Chergui, M. Sub-50-fs photoinduced spin crossover in [Fe(bpy)3] . Nature Chemistry 7, 629–633 (2015).

286. Büldt, L. A., Guo, X., Vogel, R., Prescimone, A. & Wenger, O. S. A Tris(diisocyanide)

2+ chromium(0) Complex Is a Luminescent Analog of Fe(2,2￿-Bipyridine)3 . Jour- nal of the American Chemical Society 139, 985–992 (2017).

287. O’Regan, B. & Grätzel, M. A low-cost, high-efficiency solar cell based on dye- sensitized colloidal TiO2 films. Nature 353, 737–740 (1991).

288. McDaniel, N. D., Coughlin, F. J., Tinker, L. L. & Bernhard, S. Cyclometalated Iridium(III) Aquo Complexes: Efficient and Tunable Catalysts for the Homo- geneous Oxidation of Water. Journal of the American Chemical Society 130, 210–217 (2008). 184 289. Roger, I. & Symes, M. D. First row transition metal catalysts for solar-driven water oxidation produced by electrodeposition. J. Mater. Chem. A 4, 6724– 6741 (2016).

290. Haag, W. R. & Yao, C. C. D. Rate Constants for Reaction of Hydroxyl Rad- icals With Several Drinking-Water Contaminants. Environmental Science & Technology 26, 1005–1013 (1992).

291. Ruppert, G., Bauer, R., Heisler, G. & Novalic, S. Mineralization of cyclic or- ganic water contaminants by the photo-fenton reaction - Influence of structure and substituents. Chemosphere 27, 1339–1347 (1993).

292. Feng, W. & Nansheng, D. Photochemistry of hydrolytic iron (III) species and photoinduced degradation of organic compounds. A minireview. Chemosphere 41, 1137–1147 (2000).

293. Warren, J. J., Tronic, T. A. & Mayer, J. M. Thermochemistry of Proton- Coupled Electron Transfer Reagents and its Implications. Chemical Reviews 110, 6961–7001 (2010).

294. Grabo, J. E., Chrisman, M. A., Webb, L. M. & Baldwin, M. J. Photochemical reactivity of the iron(III) complex of a mixed-donor, α-hydroxy acid-containing chelate and its biological relevance to photoactive marine siderophores. Inor- ganic Chemistry 53, 5781–5787 (2014).

295. Butler, A. & Theisen, R. M. Iron(III)-siderophore coordination chemistry: Re- activity of marine siderophores. Coordination Chemistry Reviews 254, 288–296 (2010).

296. Barbeau, K., Rue, E. L., Bruland, K. W. & Butler, A. Photochemical cycling of iron in the surface ocean mediated by microbial iron(III)-binding ligands. Nature 413, 409–413 (2001).

185 297. Klipper, F. C., Carrano, C. J., Kuhn, J. U. & Butler, A. Photoreactivity of iron(III)-aerobactin: Photoproduct structure and iron(III) coordination. Inor- ganic Chemistry 45, 6028–6033 (2006).

298. Barbeau, K., Zhang, G., Live, D. H. & Butler, A. Petrobactin, a Photoreactive Siderophore Produced by the Oil-Degrading Marine Bacterium Marinobacter hydrocarbonoclasticus. Journal of the American Chemical Society 124, 378– 379 (2002).

299. Jack, R. S., Ayoko, G. A., Adebajo, M. O. & Frost, R. L. A review of iron species for visible-light photocatalytic water purification. Environmental Science and Pollution Research 22, 7439–7449 (2015).

300. Kavitha, V. & Palanivelu, K. The role of ferrous ion in Fenton and photo- Fenton processes for the degradation of phenol. Chemosphere 55, 1235–1243 (2004).

301. Lewis, D. F. & Hlavica, P. Interactions between redox partners in various cy- tochrome P450 systems: functional and structural aspects. Biochimica et Bio- physica Acta (BBA) - Bioenergetics 1460, 353–374 (2000).

302. Ray, K., Pfaff, F. F., Wang, B. & Nam, W. Status of reactive non-heme metal-oxygen intermediates in chemical and enzymatic reactions. Journal of the American Chemical Society 136, 13942–58 (2014).

303. Averill, B. A. et al. Spectroscopic and magnetic studies of the purple acid phosphatase from bovine spleen. Journal of the American Chemical Society 109, 3760–3767 (1987).

304. Moss, M. L. & Mellon, M. G. Colorimetric Determination of Iron with 2,2’- Bipyridyl and with 2,2’,2’-Terpyridyl. Industrial & Engineering Chemistry An- alytical Edition 14, 862–865 (1942).

186 305. Braterman, P. S., Song, J. I. & Peacock, R. D. Electronic absorption spectra of the iron(II) complexes of 2,2’-bipyridine, 2,2’-bipyrimidine, 1,10-phenanthroline, and 2,2’:6’,2”-terpyridine and their reduction products. Inorganic Chemistry 31, 555–559 (1992).

306. Slater, J. W., Marguet, S. C., Cirino, S. L., Maugeri, P. T. & Shafaat, H. S. Experimental and DFT Investigations Reveal the Influence of the Outer Coor- dination Sphere on the Vibrational Spectra of Nickel-Substituted Rubredoxin, a Model Hydrogenase Enzyme. Inorganic Chemistry 56, 3926–3938 (2017).

307. Murov, S. L., Carmichael, I. & Hug, G. L. Handbook of Photochemistry, 2nd edition 2nd, 298–310 (Marcel Dekker, Inc., New York, NY, 1993).

308. Ma, X. & Tian, H. Photochemistry and Photophysics. Concepts, Research, Ap- plications. By Vincenzo Balzani, Paola Ceroni and Alberto Juris. 504 (Wiley- VCH Verlag, Weinheim, Germany, 2014).

309. Fichtner, C., van Gastel, M. & Lubitz, W. Wavelength dependence of the photo- induced conversion of the Ni-C to the Ni-L redox state in the [NiFe] hydrogenase of Desulfovibrio vulgaris Miyazaki F. Physical Chemistry Chemical Physics 5, 5507 (2003).

310. Wang, X., Valverde-Aguilar, G., Weaver, M. N., Nelsen, S. F. & Zink, J. I. Res- onance raman de-enhancement caused by excited state mixed valence. Journal of Physical Chemistry A 111, 5441–5447 (2007).

311. Gaber, B. P., Sheridan, J. P., Bazer, F. W. & Roberts, R. M. Resonance Raman scattering from uteroferrin, the purple glycoprotein of the porcine uterus. The Journal of biological chemistry 254, 8340–2 (1979).

187 312. Antanaitis, B. C., Strekas, T. & Aisen, P. Characterization of pink and purple uteroferrin by resonance Raman and CD spectroscopy. Journal of Biological Chemistry 257, 3766–3770 (1982).

313. Beck, J. L. et al. Properties of a purple phosphatase from red kidney bean: a zinc-iron metalloenzyme. Biochimica et Biophysica Acta (BBA) - Protein Structure and Molecular Enzymology 869, 61–68 (1986).

314. Neves, A. et al. Synthesis, crystal structure, electrochemical, and spectroelec- trochemical properties of the new manganese (III) complex [H2BBPEN= N, N’-bis (2-. Inorganic chemistry 31, 4749–4755 (1992).

315. Setyawati, I. A., Rettig, S. J. & Orvig, C. Cationic iron(III) complex with

a hexadentate N 2,N’2,O2-aminopyridylphenolate ligand. Canadian Journal of Chemistry 77, 2033–2038 (1999).

316. Que, L. Metalloproteins With Phenolate Coordination. Coordination Chemistry Reviews 50, 73–108 (1983).

317. Gray, H. B. & Winkler, J. R. Long-range electron transfer. Proceedings of the National Academy of Sciences of the United States of America 102, 3534–9 (2005).

318. Winkler, J. R. & Gray, H. B. Long-range electron tunneling. Journal of the American Chemical Society 136, 2930–2939 (2014).

319. Carey, J. & Langford, C. Photodecomposition of Fe (III) aminopolycarboxy- lates. Can. J. Chem. 3665–3670 (1973).

320. Budac, D. & Wan, P. Photodecarboxylation: mechanism and synthetic utility. Journal of Photochemistry and Photobiology, A: Chemistry 67, 135–166 (1992).

188 321. Solomon, E. I., Light, K. M., Liu, L. V., Srnec, M. & Wong, S. D. Geometric and electronic structure contributions to function in non-heme iron enzymes. Accounts of Chemical Research 46, 2725–2739 (2013).

322. Ensing, B., Buda, F. & Baerends, E. J. Fenton-like Chemistry in Water: Oxi- dation Catalysis by Fe(III) and H 2 O 2. The Journal of Physical Chemistry A 107, 5722–5731 (2003).

323. Stadtman, E. R. & Berlett, B. S. Fenton chemistry: Amino acid oxidation. Journal of Biological Chemistry 266, 17201–17211 (1991).

324. Berlett, B. & Stadtman, E. Protein oxidation in aging, disease, and oxidative stress. The Journal of Biological Chemistry 272, 20313–20316 (1997).

325. Stadtman, E. R. & Levine, R. L. Free radical-mediated oxidation of free amino acids and amino acid residues in proteins. Amino Acids 25, 207–218 (2003).

326. Nelson, N. & Yocum, C. F. Structure and function of photosystems I and II. Annual Review of Plant Biology 57, 521–565 (2006).

327. Kaila, V. R. I., Verkhovsky, M. I. & Wikstro, M. Proton-Coupled Electron Transfer in Cytochrome Oxidase. Chemical Reviews 110, 7062–7081 (2010).

328. Fontecilla-Camps, J. C., Volbeda, A., Cavazza, C. & Nicolet, Y. Structure/function relationships of [NiFe]- and [FeFe]-hydrogenases. Chemical Reviews 107, 4273– 4303 (2007).

329. Hoffman, B. M., Lukoyanov, D., Yang, Z.-y., Dean, D. R. & Seefeldt, L.C. Mechanism of Nitrogen Fixation by Nitrogenase: The Next Stage. Chemical Reviews 114, 4041–4062 (2014).

330. Wasielewski, M. R. Photoinduced electron transfer in supramolecular systems for artificial photosynthesis. Chemical Reviews 92, 435–461 (1992).

189 331. Marcus, R. A. Electron Transfer Reactions in Chemistry: Theory and Exper- iment (Nobel Lecture). Angewandte Chemie International Edition in English 32, 1111–1121 (1993).

332. Haim, A. Role of the bridging ligand in inner-sphere electron-transfer reactions. Accounts of Chemical Research 8, 264–272 (1975).

333. Marcus, R. & Sutin, N. Electron transfers in chemistry and biology. Biochimica et Biophysica Acta (BBA) - Reviews on Bioenergetics 811, 265–322 (1985).

334. Taube, H., Myers, H. & Rich, R. L. Observations on the mechanism of electron transfer in solution. Journal of the American Chemical Society 75, 4118–4119 (1953).

335. Hubig, S. M., Rathore, R. & Kochi, J. K. Steric Control of Electron Transfer. Changeover from Outer-Sphere to Inner-Sphere Mechanisms in Arene/Quinone Redox Pairs. Journal of the American Chemical Society 121, 617–626 (1999).

336. Zusman, L. Outer-sphere electron transfer in polar solvents. Chemical Physics 49, 295–304 (1980).

337. Blumberger, J. Recent Advances in the Theory and Molecular Simulation of Biological Electron Transfer Reactions. Chemical Reviews 115, 11191–11238 (2015).

338. Marcus, R. A. Chemical and Electrochemical Electron-Transfer Theory. Annual Review of Physical Chemistry 15, 155–196 (1964).

339. Gray, H. B. & Winkler, J. R. Electron Transfer in Proteins. Annual Review of Biochemistry 65, 537–561 (1996).

340. Winkler, J. R. & Gray, H. B. Electron transfer in ruthenium-modified proteins. Chemical Reviews 92, 369–379 (1992).

190 341. Leland, B. A. et al. Picosecond fluorescence studies on intramolecular photo- chemical electron transfer in porphyrins linked to quinones at two different fixed distances. The Journal of Physical Chemistry 89, 5571–5573 (1985).

342. Closs, G. L. & Miller, J. R. Intramolecular Long-Distance Electron Transfer in Organic Molecules. Science 240, 440–447 (1988).

343. Warman, J. M., Hom, M., Paddon-Row, M. N., Oliver, A. M. & Kroon, J. (Sub)nanosecond intramolecular hole transfer through a rigid, 10-sigma-bond, saturated-hydrocarbon bridge. Chemical Physics Letters 172, 114–119 (1990).

344. Moser, C. C., Keske, J. M., Warncke, K., Farid, R. S. & Dutton, P. L. Nature of biological electron transfer. Nature 355, 796–802 (1992).

345. Beratan, D. N., Onuchic, J. N., Betts, J. N., Bowler, B. E. & Gray, H. B. Electron tunneling pathways in ruthenated proteins. Journal of the American Chemical Society 112, 7915–7921 (1990).

346. Beratan, D. N., Betts, J. N. & Onuchic, J. N. Protein electron transfer rates set by the bridging secondary and tertiary structure. Science 252, 1285–1288 (1991).

347. Beratan, D. N., Betts, J. N. & Onuchic, J. N. Tunneling pathway and redox- state-dependent electronic couplings at nearly fixed distance in electron transfer proteins. The Journal of Physical Chemistry 96, 2852–2855 (1992).

348. Aubert, C., Mathis, P., Eker, a. P. & Brettel, K. Intraprotein electron transfer between tyrosine and tryptophan in DNA photolyase from Anacystis nidulans. Proceedings of the National Academy of Sciences of the United States of America 96, 5423–5427 (1999).

191 349. Barry, B. A. et al. Proton Coupled Electron Transfer and Redox Active Ty- rosines: Structure and Function of the Tyrosyl Radicals in Ribonucleotide Re- ductase and Photosystem II. The journal of physical chemistry letters 3, 543– 554 (2012).

350. Hirasawa, M. et al. The role of aromatic and acidic amino acids in the electron transfer reaction catalyzed by spinach ferredoxin-dependent glutamate syn- thase. Biochimica et Biophysica Acta (BBA) - Bioenergetics 1363, 134–146 (1998).

351. Wang, M., Gao, J., Müller, P. & Giese, B. Electron Transfer in Peptides with Cysteine and Methionine as Relay Amino Acids. Angewandte Chemie Interna- tional Edition 48, 4232–4234 (2009).

352. Bjerrum, M. J. et al. Electron transfer in ruthenium-modified proteins. Journal of Bioenergetics and Biomembranes 27, 295–302 (1995).

353. Chang, I. J., Gray, H. B. & Winkler, J. R. High-Driving-Force Electron-Transfer in Metalloproteins - Intramolecular Oxidation of Ferrocytochrome-C by Ru(2,2’-

3+ Bpy)2(Im)(His-33) . Journal of the American Chemical Society 113, 7056– 7057 (1991).

354. Wei, H., Yin, J. & Wang, E. Bis(2,2￿-bipyridine)(5,6-epoxy-5,6-dihydro-[1,10] phenanthroline)ruthenium: Synthesis and Electrochemical and Electrochemi- luminescence Characterization. 80, 5635–5639 (2008).

355. Dwaraknath, S. et al. A facile and versatile methodology for cysteine specific labeling of proteins with octahedral polypyridyl d6 metal complexes. Journal of Inorganic Biochemistry 136, 154–160 (2014).

192 356. Balabin, I. A., Hu, X. & Beratan, D. N. Exploring biological electron transfer pathway dynamics with the Pathways Plugin for VMD. Journal of Computa- tional Chemistry 33, 906–910 (2012).

357. Langen, R. et al. Electron tunneling in proteins: coupling through a beta strand. Science 268, 1733–1735 (1995).

358. Engelman, R. & Jortner, J. The energy gap law for radiationless transitions in large molecules. 18, 145–164 (1970).

359. Bixon, M., Jortner, J., Cortes, J., Heitele, H. & Michel-Beyerle, M. E. Energy Gap Law for Nonradiative and Radiative Charge Transfer in Isolated and in Solvated Supermolecules. The Journal of Physical Chemistry 98, 7289–7299 (1994).

360. Caspar, J. V., Kober, E. M., Sullivan, B. P. & Meyer, T. J. Application of the energy gap law to the decay of charge-transfer excited states. Journal of the American Chemical Society 104, 630–632 (1982).

361. Lakowicz, J. R. Principles of fluorescence spectroscopy 3rd editio (Springer Science+ Business Media LLC, New York, NY, 2006).

362. Beinert, H. Iron-Sulfur Clusters: Nature’s Modular, Multipurpose Structures. Science 277, 653–659 (1997).

363. Ye, H. & Rouault, T. A. Human iron-sulfur cluster assembly, cellular iron homeostasis, and disease. Biochemistry 49, 4945–4956 (2010).

364. Brzóska, K., Meczyńska, S. & Kruszewski, M. Iron-sulfur cluster proteins: elec- tron transfer and beyond. Acta biochimica Polonica 53, 685–91 (2006).

365. Fontecave, M. Iron-sulfur clusters: ever-expanding roles. Nature Chemical Bi- ology 2, 171–174 (2006).

193 366. Lanzilotta, W. N., Christiansen, J., Dean, D. R. & Seefeldt, L. C. Evidence for coupled electron and proton transfer in the [8Fe-7S] cluster of nitrogenase. Biochemistry 37, 11376–11384 (1998).

367. Peters, J. W. et al. Redox-dependent structural changes in the nitrogenase P-cluster. Biochemistry 36, 1181–1187 (1997).

368. Moulis, J. M. & Meyer, J. Characterization of the selenium-substituted 2[4Fe- 4Se] ferredoxin from Clostridium pasteurianum. Biochemistry 21, 4762–4771 (1982).

369. Zheng, B., Chen, X. D., Zheng, S. L. & Holm, R. H. Selenium as a struc- tural surrogate of sulfur: Template-assisted assembly of five types of tungsten- iron-sulfur/selenium clusters and the structural fate of chalcogenide reactants. Journal of the American Chemical Society 134, 6479–6490 (2012).

370. Türker, L. & Erkoç, Ş. ￿. A theoretical study on certain iron-sulfur and iron- selenium clusters. Journal of Molecular Structure: THEOCHEM 623, 17–21 (2003).

371. Noodleman, L. Exchange coupling and resonance delocalization in reduced

+ + (Fe4S4) and (Fe4Se4) clusters. 2. A Generalized Nonlinear Model for Spin- State Energies and EPR and Hyperfine Properties. Inorganic Chemistry 30, 246–256 (1991).

372. Mouesca, J., Noodleman, L., Case, D. a. & Lamottes, B. Spin Densities and Spin Coupling in Iron-Sulfur Clusters : A New Analysis of Hyperfine Coupling Constants. Inorganic Chemistry 34, 4347–4359 (1995).

373. Backes, G. et al. The Environment of Fe4S4 Clusters in Ferredoxins and High- Potential Iron Proteins. New Information from X-ray Crystallography and Res- onance Raman Spectroscopy. J. Am. Chem. Soc 113, 2055–2064 (1991).

194 374. Spiro, T. G., Czernuszewicz, R. S. & Han, S. in Biological Applications of Raman Spectroscopy, Vol. 3 (Resonance Raman Spectra of Heme and Metallo- proteins) (ed Spiro, T. G.) 523–553 (Wiley & Sons, New York, NY, 1988).

375. Qi, W. et al. Glutathione Complexed Fe − S Centers. Journal of the American Chemical Society 134, 10745–10748 (2012).

376. Battin, E. E. & Brumaghim, J. L. Antioxidant activity of sulfur and selenium: A review of reactive oxygen species scavenging, glutathione peroxidase, and metal-binding antioxidant mechanisms. Cell Biochemistry and Biophysics 55, 1–23 (2009).

377. Deponte, M. Glutathione catalysis and the reaction mechanisms of glutathione- dependent enzymes. Biochimica et Biophysica Acta - General Subjects 1830, 3217–3266 (2013).

378. Meister, A. Glutathione metabolism and its selective modification. Journal of Biological Chemistry 263, 17205–17208 (1988).

379. Beardwood, P. & Gibson, J. F. Resonance-Raman and infrared spectra of the

2– dimeric iron–chalcogenide complexes [Fe2X2(YR)4] (X and Y=S or Se). J. Chem. Soc., Dalton Trans. 1507–1516 (1984).

380. Meyer, J., Moulis, J.-M. & Lutz, M. Structural differences between [2Fe-2S] clusters in spinach ferredoxin and in the “Red paramagnetic protein” from . A resonance Raman study. Biochemical and Biophysical Research Communica- tions 119, 828–835 (1984).

381. Meyer, J., Moulis, J.-M. & Lutz, M. Resonance Raman spectroscopy of [2Fe−2X]2+ (X=S,Se) clusters in ferredoxins. Biochimica et Biophysica Acta (BBA) - Pro- tein Structure and Molecular Enzymology 873, 108–118 (1986).

382. Turner, J. A. Sustainable Hydrogen Production. Science 305, 972–974 (2004). 195 383. Momirlan, M. & Veziroglu, T. The properties of hydrogen as fuel tomorrow in sustainable energy system for a cleaner planet. International Journal of Hydrogen Energy 30, 795–802 (2005).

384. Goldsmith, J. I., Hudson, W. R., Lowry, M. S., Anderson, T. H. & Bernhard, S. Discovery and high-throughput screening of heteroleptic iridium complexes for photoinduced hydrogen production. Journal of the American Chemical Society 127, 7502–7510 (2005).

385. Cline, E. D., Adamson, S. E. & Bernhard, S. Homogeneous catalytic system for photoinduced hydrogen production utilizing iridium and rhodium complexes. Inorganic Chemistry 47, 10378–10388 (2008).

386. Holladay, J. D., Hu, J., King, D. L. & Wang, Y. An overview of hydrogen production technologies. Catalysis Today 139, 244–260 (2009).

387. Rodríguez-Lugo, R. E. et al. A homogeneous transition metal complex for clean hydrogen production from methanol–water mixtures. Nature Chemistry 5, 342– 347 (2013).

388. Han, Y., Wu, Y., Lai, W. & Cao, R. Electrocatalytic water oxidation by a water-soluble nickel porphyrin complex at neutral ph with low overpotential. Inorganic Chemistry 54, 5604–5613 (2015).

389. Kandemir, B., Kubie, L., Guo, Y., Sheldon, B. & Bren, K. L. Hydrogen Evo- lution from Water under Aerobic Conditions Catalyzed by a Cobalt ATCUN Metallopeptide. Inorganic Chemistry 55, 1355–1357 (2016).

390. Vignais, P. M., Billoud, B. & Meyer, J. Classification and phylogeny of hydro- genases. FEMS Microbiology Reviews 25, 455–501 (2001).

196 391. Jones, A. K., Sillery, E., Albracht, S. P. J. & Armstrong, F. A. Direct compari- son of the electrocatalytic oxidation of hydrogen by an enzyme and a platinum catalyst. Chemical Communications, 866–867 (2002).

392. Vincent, K. A., Parkin, A. & Armstrong, F. A. Investigating and Exploiting the Electrocatalytic Properties of Hydrogenases. Chemical Reviews 107, 4366– 4413 (2007).

393. Madden, C. et al. Catalytic Turnover of [FeFe]-Hydrogenase Based on Single- Molecule Imaging. Journal of the American Chemical Society 134, 1577–1582 (2012).

394. Lubitz, W., Ogata, H., Rüdiger, O. & Reijerse, E. Hydrogenases. Chemical Reviews 114, 4081–4148 (2014).

395. Shima, S. et al. The Crystal Structure of [Fe]-Hydrogenase Reveals the Geom- etry of the Active Site. Science 321, 572–575 (2008).

396. Xu, T. et al. A Functional Model of [Fe]-Hydrogenase. Journal of the American Chemical Society 138, 3270–3273 (2016).

397. Shafaat, H. S., Rüdiger, O., Ogata, H. & Lubitz, W. [NiFe] hydrogenases: A common active site for hydrogen metabolism under diverse conditions. Biochim- ica et Biophysica Acta (BBA) - Bioenergetics 1827, 986–1002 (2013).

398. Peters, J. W. et al. [FeFe]- and [NiFe]-hydrogenase diversity, mechanism, and maturation. Biochimica et Biophysica Acta (BBA) - Molecular Cell Research 1853, 1350–1369 (2015).

399. Barton, B. E., Olsen, M. T. & Rauchfuss, T. B. Aza- and Oxadithiolates Are Probable Proton Relays in Functional Models for the [FeFe]-Hydrogenases. Journal of the American Chemical Society 130, 16834–16835 (2008).

197 400. Olsen, M. T., Barton, B. E. & Rauchfuss, T. B. Hydrogen activation by biomimetic diiron dithiolates. Inorganic Chemistry 48, 7507–7509 (2009).

401. Streich, D. et al. High-Turnover Photochemical Hydrogen Production Cat- alyzed by a Model Complex of the [FeFe]-Hydrogenase Active Site. Chemistry - A European Journal 16, 60–63 (2010).

402. Camara, J. M. & Rauchfuss, T. B. Combining acid–base, redox and substrate binding functionalities to give a complete model for the [FeFe]-hydrogenase. Nature Chemistry 4, 26–30 (2011).

403. Carroll, M. E., Barton, B. E., Rauchfuss, T. B. & Carroll, P. J. Synthetic models for the active site of the [FeFe]-hydrogenase: Catalytic proton reduction and the structure of the doubly protonated intermediate. Journal of the American Chemical Society 134, 18843–18852 (2012).

404. Berggren, G. et al. Biomimetic assembly and activation of [FeFe]-hydrogenases. Nature 499, 66–69 (2013).

405. Huynh, M. T., Wang, W., Rauchfuss, T. B. & Hammes-Schiffer, S. Computa- tional Investigation of [FeFe]-Hydrogenase Models: Characterization of Singly and Doubly Protonated Intermediates and Mechanistic Insights. Inorganic Chemistry 53, 10301–10311 (2014).

406. Galinato, M. G. I., Whaley, C. M., Roberts, D., Wang, P. & Lehnert, N. Favor-

+ able Protonation of the (µ-edt)[Fe2(PMe3)4(CO)2(H-terminal)] Hydrogenase Model Complex Over Its Bridging µ-H Counterpart: A Spectroscopic and DFT Study. European journal of inorganic chemistry 2011, 1147–1154 (2011).

407. Van der Vlugt, J. I., Rauchfuss, T. B., Whaley, C. M. & Wilson, S. R. Char- acterization of a Diferrous Terminal Hydride Mechanistically Relevant to the

198 Fe-Only Hydrogenases. Journal of the American Chemical Society 127, 16012– 16013 (2005).

408. Ferguson, S. J. Nitrogen cycle enzymology. Current Opinion in Chemical Bi- ology 2, 182–193 (1998).

409. Canfield, D. E., Glazer, A. N. & Falkowski, P. G. The Evolution and Futureof Earth’s Nitrogen Cycle. Science 330, 192–196 (2010).

410. MacKay, B. A. & Fryzuk, M. D. Dinitrogen Coordination Chemistry: On the Biomimetic Borderlands. Chemical Reviews 104, 385–402 (2004).

411. Francis, C. A., Beman, J. M. & Kuypers, M. M. M. New processes and players in the nitrogen cycle: the microbial ecology of anaerobic and archaeal ammonia oxidation. The ISME Journal 1, 19–27 (2007).

412. Hayatsu, M., Tago, K. & Saito, M. Various players in the nitrogen cycle: Di- versity and functions of the microorganisms involved in nitrification and deni- trification. Soil Science and Plant Nutrition 54, 33–45 (2008).

413. Gruber, N. & Galloway, J. N. An Earth-system perspective of the global nitro- gen cycle. Nature 451, 293–296 (2008).

414. Ravishankara, A. R., Daniel, J. S. & Portmann, R. W. Nitrous Oxide (N2O): The Dominant Ozone-Depleting Substance Emitted in the 21st Century. Sci- ence 326, 123–125 (2009).

415. Chandran, K., Stein, L. Y., Klotz, M. G. & van Loosdrecht, M. C. M. Nitrous oxide production by lithotrophic ammonia-oxidizing bacteria and implications for engineered nitrogen-removal systems. Biochemical Society transactions 39, 1832–7 (2011).

199 416. Kozlowski, J. A., Kits, K. D. & Stein, L. Y. Comparison of Nitrogen Oxide Metabolism among Diverse Ammonia-Oxidizing Bacteria. Frontiers in Micro- biology 7, 1–9 (2016).

417. Soares, J. R. et al. Nitrous oxide emission related to ammonia-oxidizing bacteria and mitigation options from N fertilization in a tropical soil. Scientific Reports 6, 30349 (2016).

418. Arp, D., Sayavedra-Soto, L. & Hommes, N. Molecular biology and biochemistry of ammonia oxidation by Nitrosomonas europaea. Archives of Microbiology 178, 250–255 (2002).

419. Hooper, A. B. & Terry, K. Hydroxylamine oxidoreductase of Nitrosomonas: Production of Nitric Oxide from Hydroxylamine. Biochimica et Biophysica Acta (BBA) - Enzymology 571, 12–20 (1979).

420. Lipscomb, J. D. & Hooper, A. B. Resolution of multiple heme centers of hy- droxylamine oxidoreductase from Nitrosomonas. 1. Electron paramagnetic res- onance spectroscopy. Biochemistry 21, 3965–3972 (1982).

421. Cedervall, P., Hooper, A. B. & Wilmot, C. M. Structural Studies of Hydrox- ylamine Oxidoreductase Reveal a Unique Heme Cofactor and a Previously Unidentified Interaction Partner. Biochemistry 52, 6211–6218 (2013).

422. Rees, M. & Nason, A. A P-450-like cytochrome and a soluble terminal oxi- dase identified as cytochrome O from Nitrosomonas europaea. Biochemical and Biophysical Research Communications 21, 248–256 (1965).

423. Hooper, A. B. & Nason, A. Characterization of hydroxylamine-cytochrome c reductase from the chemoautotrophs Nitrosomonas europaea and Nitrosocystis oceanus. Journal of Biological Chemistry 240, 4044–57 (1965).

200 424. Arciero, D. M. & Hooper, A. B. Consideration of a phlorin structure for haem P -460 of hydroxylamine oxidoreductase and its implications regarding reaction mechanism. Biochemical Society Transactions 26, 385–389 (1998).

425. Andersson, K. K., Babcock, G. T. & Hooper, A. B. P460 of hydroxylamine ox- idoreductase of Nitrosomonas europaea: Soret resonance Raman evidence for a novel heme-like structure. Biochemical and biophysical research communica- tions 174, 358–63 (1991).

426. Pearson, A. R. et al. The crystal structure of cytochrome P460 of Nitrosomonas europaea reveals a novel cytochrome fold and heme - Protein cross-link. Bio- chemistry 46, 8340–8349 (2007).

427. Bergmann, D. J. & Hooper, A. B. Cytochrome P460 of Nitrosomonas europaea. Formation of the heme-lysine cross-link in a heterologous host and mutagenic conversion to a non-cross-linked cytochrome c ’. European Journal of Biochem- istry 270, 1935–1941 (2003).

428. Desbois, A., Lutz, M. & Banerjee, R. Low-frequency vibrations in resonance Raman spectra of horse heart myoglobin. Iron-ligand and iron-nitrogen vibra- tional modes. Biochemistry 18, 1510–1518 (1979).

429. Keefer, L. K., Nims, R. W., Davies, K. M. & Wink, D. A. “NONOates” (1- substituted diazen-1-ium-1,2-diolates) as nitric oxide donors: Convenient nitric oxide dosage forms. Methods in Enzymology 268, 281–293 (1996).

430. Maragos, C. M. et al. Complexes of •NO with nucleophiles as agents for the controlled biological release of nitric oxide. Vasorelaxant effects. Journal of Medicinal Chemistry 34, 3242–3247 (1991).

201 431. Tamagnini, F. et al. Nitric oxide-dependent long-term depression but not endo- cannabinoid-mediated long-term potentiation is crucial for visual recognition memory. The Journal of Physiology 591, 3963–3979 (2013).

432. Hope, H. & Sequeira, M. R. Angeli’s salt. Crystal structure of sodium triox-

odinitrate(II) monohydrate, Na2N2O3•H2O. Inorganic Chemistry 12, 286–288 (1973).

433. Amatore, C., Arbault, S., Ducrocq, C., Hu, S. & Tapsoba, I. Angeli’s Salt

(Na2N2O3) is a Precursor of HNO and NO: a Voltammetric Study of the Re- active Intermediates Released by Angeli’s Salt Decomposition. ChemMedChem 2, 898–903 (2007).

434. Enemark, J. & Feltham, R. Principles of structure, bonding, and reactivity for metal nitrosyl complexes. Coordination Chemistry Reviews 13, 339–406 (1974).

435. Kaim, W. & Schwederski, B. Non-innocent ligands in bioinorganic chemistry— An overview. Coordination Chemistry Reviews 254, 1580–1588 (2010).

436. Stong, J. D., Burke, J. M., Daly, P., Wright, P. & Spiro, T. G. Resonance Raman spectra of nitrosyl heme proteins and of porphyrin analogs. Journal of the American Chemical Society 102, 5815–5819 (1980).

437. Maes, E. M., Walker, F. a., Montfort, W. R. & Czernuszewicz, R. S. Resonance Raman spectroscopic study of nitrophorin 1, a nitric oxide-binding heme pro- tein from Rhodnius prolixus, and its nitrosyl and cyano adducts. Journal of the American Chemical Society 123, 11664–11672 (2001).

438. Kim, E. K. & Kochi, J. K. Charge-transfer structures of aromatic electron donor-acceptor complexes leading to electron transfer with the electrophilic nitrosonium cation. Journal of the American Chemical Society 113, 4962–4974 (1991).

202 439. Tomita, T., Haruta, N., Aki, M., Kitagawa, T. & Ikeda-Saito, M. UV resonance Raman detection of a ligand vibration on ferric nitrosyl heme proteins. J. Am. Chem. Soc. 123, 2666–2667 (2001).

440. Tomita, T., Hirota, S., Ogura, T., Olson, J. S. & Kitagawa, T. Resonance Raman Investigation of Fe−N−O Structure of Nitrosylheme in Myoglobin and Its Mutants. The Journal of Physical Chemistry B 103, 7044–7054 (1999).

441. Spiro, T. G. & Strekas, T. C. Resonance Raman spectra of heme proteins. Effects of oxidation and spin state. Journal of the American Chemical Society 96, 338–345 (1974).

442. Hendrich, M. P. et al. The Active Site of Hydroxylamine Oxidoreductase from Nitrosomonas: Evidence for a New Metal Cluster in Enzymes. Journal of the American Chemical Society 116, 11961–11968 (1994).

443. Hooper, A. B. & Balny, C. Reaction of oxygen with hydroxylamine oxidore- ductase of Nitrosomonas. FEBS Letters 144, 299–303 (1982).

444. Caranto, J. D., Vilbert, A. C. & Lancaster, K. M. Nitrosomonas europaea cy- tochrome P460 is a direct link between nitrification and nitrous oxide emission. Proceedings of the National Academy of Sciences 113, 14704–14709 (2016).

445. Caranto, J. D. & Lancaster, K. M. Nitric oxide is an obligate bacterial nitri- fication intermediate produced by hydroxylamine oxidoreductase. Proceedings of the National Academy of Sciences 114, 8217–8222 (2017).

203