DISCOVERY OF KEY INTERMEDIATES FOR RADICAL INITIATION IN PFL-AE by Elizabeth Claire McDaniel A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Chemistry MONTANA STATE UNIVERSITY Bozeman, Montana January 2021 ©COPYRIGHT by Elizabeth Claire McDaniel 2020 All Rights Reserved ii DEDICATION To my mom who taught me how to read, write, and cook. To my dad, who keeps me looking at science in new ways. And to my big brother Chris, who always knows how to cheer me up. Thank you. iii ACKNOWLEDGEMENTS I would like to thank my Ph.D. advisor Dr. Joan Broderick for her training and patience working with me. Thanks to my committee members Dr. Jennifer DuBois, Dr. Martin Lawrence, and Dr. Robert Szilagyi. A special thank you to Dr. Will Broderick for his assistance and insights. Thank you to Dr. Szilagyi and Dr. Eric Shepard for their encouragement and for teaching me EPR spectroscopy. Thank you to our collaborators at Northwestern University, Dr. Brian Hoffman and his lab members Dr. Hao Yang, and Richard Jodts for conducting the ENDOR and RFQ experiments. Thank you so much to Dr. Doreen Brown who always went above and beyond to help and support me. Thank you to Broderick Lab members past and present for their insights on my projects as well as the good times we had in and out of the lab. To my friends and family, thank you for everything you have done over the years. Thank you for always being there for me. iv TABLE OF CONTENTS 1. INTRODUCTION ...........................................................................................................1 [Fe-S] Clusters in Biology ...............................................................................................1 Function of [Fe-S] Clusters in Biological Systems ..................................................2 Diversity in the Radical SAM Enzyme Superfamily ......................................................4 RS Enzyme Structure and Characterization .............................................................5 Radical Initiation Pathway .......................................................................................6 PFL-AE and PFL as a Model RS Enzyme System .........................................................8 PFL-AE ....................................................................................................................9 PFL .........................................................................................................................15 RS Enzyme Mechanism ................................................................................................18 Discovery and Characterization of Ω .....................................................................18 Research Objectives ......................................................................................................24 Determining the Ubiquity and Further Characterization of Ω ..............................24 Capturing and Characterizing the 5′-dAdo• ..........................................................25 Further Studies of RS Enzyme Intermediates and their Mechanisms of Formation ..............................................................................................................26 References .....................................................................................................................28 2. GENERAL METHODOLOGY .....................................................................................37 Expression and Purification of PFL-AE ........................................................................37 Protein and Iron Quantification .....................................................................................40 Expression and Purification of PFL ..............................................................................41 Expression and Purification of OspD ............................................................................45 Reconstitution of OspD..........................................................................................47 Growth and Preparation of SAM Synthetase ................................................................48 SAM Synthesis and Purification ...................................................................................48 anSAM Synthesis and Purification .........................................................................50 Electron Paramagnetic Resonance Spectroscopy (EPR) ...............................................51 Preparation of RS Enzyme Intermediate Samples .........................................................55 RFQ Samples .........................................................................................................56 Hand Quench (HQ) Photolysis Samples ................................................................57 References .....................................................................................................................59 3. PARADIGM SHIFT FOR RADICAL S-ADENOSYL-ʟ-METHIONINE REACTIONS: THE ORGANOMETALLIC INTERMEDIATE Ω IS CENTRAL TO CATALYSIS GENERAL METHODOLOGY ...................................61 Contribution of Authors and Co-Authors ......................................................................61 v TABLE OF CONTENTS CONTINUED Manuscript Information .................................................................................................64 Abstract .........................................................................................................................65 Introduction ...................................................................................................................66 Is Ω the Result of Protein Conformational Rearrangements during Assembly of the PFL-AE/SAM/PFL Ternary Complex? .............................................67 Is Ω Mechanistically Formed throughout the RS Superfamily? ...................................68 Detailed Structure of Ω? ................................................................................................70 Mechanism: How does Ω Form, and Then Liberate 5′-dAdo•? ....................................72 References .....................................................................................................................76 Supplementary Information ...........................................................................................79 Materials and Methods ...........................................................................................79 HydG Preparations .................................................................................................79 Lysine 2,3-aminomutase Preparations ...................................................................80 OspD and OspA (substrate) Preparations ..............................................................80 PFL-AE and PFL (substrate) Preparations ............................................................81 PoyD and PoyA (substrate) Preparations ...............................................................82 Anaerobic RNR-AE and RNR (substrate) Preparations ........................................83 Spore Photoproduct Lyase (SPL) Preparations......................................................85 SAM Preparation ...................................................................................................86 RFQ Sample Preparation .......................................................................................87 Rapid Freeze-Quench Experiments .......................................................................88 EPR and ENDOR Measurements ..........................................................................89 EPR of Isotopically labeled Ω with PFL-AE .........................................................90 EPR and ENDOR Ω prepared with 1/2H SAM .......................................................90 Supplemental Figures ....................................................................................................91 Supplementary References ..........................................................................................100 4. THE ELUSIVE 5′DEOXYADENOSYL RADICAL: CAPTURED AND CHARACTERIZED BY EPR AND ENDOR SPECTROSCOPIES ..........................102 Contributions of Authors and Co-Authors ..................................................................102 Manuscript Information Page ......................................................................................104 Abstract .......................................................................................................................105 Introduction .................................................................................................................106 Experimental Methods ................................................................................................108 Materials ..............................................................................................................108 Protein and SAM Preparation ..............................................................................109 Photolysis .............................................................................................................109 EPR and ENDOR Measurements ........................................................................110 vi TABLE OF CONTENTS CONTINUED DFT Calculations .................................................................................................111 Results .........................................................................................................................112 Photolysis of PFL-AE/[4Fe-4S]+/SAM ...............................................................112 Confirmation of Radical Species Using SAM Isotopologs .................................114 Structure of 5′-dAdo• ...........................................................................................117 5′-dAdo• in the Active Site ..................................................................................120 Discussion ...................................................................................................................122 Conclusion ...................................................................................................................124 References ...................................................................................................................126 Supplementary Information .........................................................................................133 Supplementary Methods ......................................................................................133 Supplementary Text .............................................................................................133 Has the fluid-solution spectrum of 5′-Ado• been seen in an FT-EPR study of photolyzed B12? ...................................................................................133 Is the EPR spectrum of a radical generated from Geobacillus thermodenitrificans (Gt) spore photoproduct lyase (SPL) compatible with assignment to 5′-dAdo•? ..........................................................................135 DFT calculations ......................................................................................137 Supplementary Figures ................................................................................................139 Supplementary References ..........................................................................................142 5. FURTHER STUDIES INTO THE MECHANISM OF INTERMEDIATE FORMATION IN RADICAL S-ADENOSYL-L-METHIONINE (SAM) ENZYMES USING PFL-AE AND AN ANHYDROUS ANALOG OF SAM ..............................143 Contribution of Authors and Co-Authors ....................................................................143 Abstract .......................................................................................................................145 Introduction .................................................................................................................146 Experimental Methods ................................................................................................152 Materials ..............................................................................................................152 Protein and SAM preparation ..............................................................................152 PFL-AE Activity Assays......................................................................................152 Photolysis Samples ..............................................................................................153 Rapid Freeze-Quench Samples ............................................................................153 Rapid Freeze-Quench Experiments .....................................................................153 EPR Measurements ..............................................................................................154 Results ........................................................................................................................155 PFL-AE Uses anSAM to Activate PFL ...............................................................155 An Ω-like Intermediate forms in the Presence of anSAM ...................................156 Photolysis of PFL-AE and anSAM ......................................................................158 vii TABLE OF CONTENTS CONTINUED Discussion ..................................................................................................................166 References ..................................................................................................................169 6. CONCLUSION ............................................................................................................173 References .................................................................................................................182 REFERENCES ................................................................................................................185 viii LIST OF TABLES Table Page 1.1 Examples of RS Enzymes in all Six Kingdoms of Life .....................................5 2.1 Recipes for RFQ Samples with Two Different Reductive Sources .................56 3.S1 Overview of radical SAM enzymes examined in this study ..........................99 4.1 Hyperfine Tensor (MHz) of 5′-dAdo• from the Experimental Plus DFT- Computed values ..............................................................................................119 4.S1 Table of hyperfine tensors (MHz) for 5′-dAdo• from current experiment plus DFT-computed compared to the radical generated from SPL .........................136 4.S2 Comparison of hyperfine couplings calculated with B3LYP and BP86 .....137 ix LIST OF FIGURES Figure Page 1.1. Common [Fe-S] clusters in biology ..................................................................1 1.2. Homolytic cleavage of SAM ............................................................................5 1.3. Representation of the TIM barrels in three RS enzymes ..................................7 1.4. Alternate uses of SAM as a cofactor or a cosubstrate ......................................8 1.5. The crystal structure of PFL-AE with SAM bound ........................................10 1.6. EPR spectra of the [4Fe-4S]+ cluster of PFL-AE ...........................................12 1.7. ENDOR spectra of PFL-AE bound to isotopically labeled SAM...................13 1.8. PFL-AE follows single turnover kinetics in the generation of Gly• ...............14 1.9. PFL is activated by PFL-AE ...........................................................................16 1.10. PFL enzyme mechanism ...............................................................................17 1.11. EPR spectra of the Ω signal and predicted 5′-dAdo• signal .........................19 1.12. The discovery and characterization of Ω ......................................................22 1.13. Two potential mechanisms of radical initiation by RS enzyme family ........23 1.14. A diagram representing the homolytic cleavage of SAM and anSAM ........27 2.1. SDS PAGE gels of the Expression of PFL-AE and PFL ................................37 2.2. Representative chromatogram for PFL-AE purification.................................38 2.3. Representative chromatogram for PFL purification .......................................43 2.4. SDS PAGE gels of purified PFL ....................................................................44 2.5. SDS PAGE gels showing the expression and purification of OspD ...............45 x LIST OF FIGURES CONTINUED 2.6. Method and representative chromatogram of SAM purification ....................49 2.7. An unpaired electron results in an EPR absorption spectrum .........................52 2.8. Hyperfine coupling on a carbon centered radical ...........................................53 2.9. The nuclear environment surrounding an unpaired electron ..........................54 3.1. Mixing conditions for PFL-AE/PFL/SAM tertiary complex ..........................68 3.2. Reactions catalyzed by the radical SAM enzymes studied in this work.........68 3.3. Normalized EPR spectra of Ω .........................................................................71 3.4. 1H ENDOR of Ω .............................................................................................72 3.5. Pathway for liberating 5′-dAdo• for H atom abstraction ................................73 3.S1. The previously accepted mechanism for radical SAM enzymes ..................91 3.S2. EPR spectra of Ω as freeze-quenched at 500 ms ..........................................92 3.S3. Normalized EPR spectra of Ω with all the radical SAM enzymes examined in this paper .......................................................................................................93 3.S4. Normalized representative EPR spectra of [4Fe-4S]+ and ([4Fe-4S]++SAM) cluster for PFL-AE in comparison with spectra of freeze-quench samples ......94 3.S5. EPR spectrum of product Gly radical formed upon annealing RNR-AE Ω as indicated, overlaid with spectrum of hand-quenched enzyme with radical ......95 3.S6. EPR spectrum of Ω generated with 57Fe and 56Fe ........................................96 3.S7. X-band EPR spectra of Ω generated with labeled SAMs .............................97 3.S8. GHz CW 15N/14N ENDOR of Ω ...................................................................98 4.1. Comparison of Adenosylcobalamin/coenzyme B12 and S-adenosylmethionine bound to a [4Fe−4S] cluster ............................................................................108 xi LIST OF FIGURES CONTINUED 4.2. X-band EPR spectra of the ([4Fe−4S]++SAM) PFL-AE complex before and after photolysis ................................................................................................113 4.3. X-band EPR spectra and simulations of 5′-dAdo• ........................................116 4.4. DFT models of 5′-dAdo• ...............................................................................121 4.5. Cartoon illustrating the proposed movements of 5′-dAdo• upon 5-C(5′) bond cleavage ...........................................................................................................122 4.S1. X-band EPR spectra of 5′-dAdo• generated from photoreduced and DT reduced samples ..............................................................................................133 4.S2. Simulations of X-band FT-EPR spectra (absorption-display) for a radical tumbling in solution; small coupling omitted ..................................................135 4.S3. Comparison of the X-band EPR spectrum of 5′-dAdo• and the reported radical generated during reaction of SPL with simulation ..............................136 4.S4. DFT calculated energy surface for rotation about the 4C′-5C′ bond ..........138 4.S5. Q-band EPR spectra and simulations of 5′-dAdo• ......................................139 4.S6. 35 GHz CW stochastic 13C ENDOR of 5′-dAdo• generated with [adenosyl- 13C ,15 10 N5]-SAM ..............................................................................................140 4.S7. X-band EPR spectra of 5′-dAdo• generated with labeled SAMs ...............141 5.1. SAM as a cofactor or a cosubstrate...............................................................147 5.2. Two Proposed mechanisms for radical initiation in RS enzymes.................149 5.3. A side by side comparison of SAM and anSAM after reductive cleavage ...151 5.4. Two graphs showing the activity of PFL-AE with SAM or anSAM ............155 5.5. The structures and EPR spectra of Ω and anΩ .............................................157 5.6. [4Fe-4S]+ cluster of PFL-AE before and after addition of anSAM ..............158 xii LIST OF FIGURES CONTINUED 5.7. EPR spectra before and after photolysis .......................................................159 5.8. EPR spectra of the anAdo• before and after 18 hr at 77 K ...........................161 5.9. Deconvolution of the radical signal ..............................................................162 5.10. A second deconvolution of the radical signal .............................................163 5.11. Simulations of the anAdo• in PFL-AE .......................................................164 5.12. EPR spectra of the anAdo• at 40 K and 77 K .............................................164 5.13. EPR spectra of the anAdo• at 75 K varying the microwave power ............164 5.14. A schematic of anSAM cleavage and comparison of signal intensities .....165 xiii NOMENCLATURE [4Fe-4S] 4 iron-4 sulfur cluster 5′-dAdo• 5′-deoxyadenosyl radical 5′-dAdoH 5′-deoxyadenosine AA flame atomic absorption AMP ampicillin anAdo• 3′,4′-anhydroadenosyl radical anAdoH 3′,4′-anhydroadenosine anSAM S-3′,4′-anhydroadenosyl-L-methionine ATP adenosine triphosphate CV column volume CW continuous wave DMSO dimethyl sulfoxide DT dithionite DTT dithiothreitol E. coli Escherichia coli EDTA ethylenediaminetetraacetic acid ENDOR electron nuclear double resonance spectroscopy EPR electron paramagnetic resonance spectroscopy FAS ferrous ammonium sulfate FPLC fast protein liquid chromatography HEPES 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid HQ hand quench IPTG Isopropyl β-d-1-thiogalactopyranoside KAN kanamycin LAM lysine 2,3-aminomutase LB medium Lysogen broth ꞵME ꞵ-mercaptoethanol or 2-mercaptoethanol MWCO molecular weight cut off Ni-NTA resin Nickel-Nitrilotriacetic acid resin O.D.600 optical density at 600 nm O/N over night oxytet oxytetracycline PFL pyruvate formate-lyase PFL-AE pyruvate formate-lyase activating enzyme PMSF phenylmethylsulfonyl fluoride QMA quaternary methylammonium RB round bottom RFQ rapid freeze quench rpm rotations per minute RS radical SAM xiv NOMENCLATURE CONTINUED SAM S-adenosyl-ʟ-methionine SDS PAGE sodium dodecyl sulphate-polyacrylamide gel electrophoresis SPEC spectinomycin TB medium terrific broth medium TLC thin-layer chromatography Tris 2-Amino-2-(hydroxymethyl)-1,3-propanediol UV-VIS ultraviolet-visible xv ABSTRACT Members of the radical S-adenosyl-L-methionine (SAM) enzyme superfamily utilize a [4Fe-4S] cluster and the small molecule, SAM, to generate methionine and the 5′deoxyadenosyl radical (5′-dAdo•). Once formed, the 5′-dAdo• abstracts a hydrogen from substrate allowing for the catalyzation of a wide array of chemistry such as DNA repair, hydrogenase maturation, and anaerobic glucose metabolism. Originally, the 5′- dAdo• was thought to form directly through homolytic cleavage of the S-C5′ bond on SAM. In 2016, this mechanism was called into question when a catalytically relevant organometallic intermediate (Ω) was discovered in pyruvate formate-lyase activating enzyme (PFL-AE). This intermediate consisted of a 5′-dAdo moiety bound to the unique iron on the PFL-AE [4Fe-4S] cluster through an Fe-C5′ bond. The work shown in this thesis provides novel insights into the RS enzyme mechanism considering the newly discovered Ω species. Using rapid freeze quench (RFQ) in conjunction with electron paramagnetic resonance (EPR) spectroscopy, Ω formation was observed in seven RS enzymes representing the totality of superfamily reaction types. Inspired by the idea that the Fe-C5′ bond in Ω could undergo photoinitiated homolysis, a unique procedure was developed to generate and capture the long elusive 5′-dAdo• through cryogenic photolysis of reduced PFL-AE and SAM. Isotopic labeling of SAM along with EPR spectroscopy confirmed definitely that this was the long sought after 5′-dAdo•. To better understand RS enzyme bond specificity and the order of intermediate formation, an analogue of SAM, S-3′4′-anhydroadenosyl-L-methionine (anSAM), was employed in RFQ and cryogenic photolysis experiments. By using anSAM, it was shown that the bond cleavage specificity of PFL-AE can changed under appropriate conditions and provided evidence that Ω forms first in the radical initiation pathway of RS enzymes. These results have greatly increased our understanding of the RS enzyme mechanism and will help future work designed to utilize the incredible enzymatic potential of this diverse superfamily. 1 CHAPTER 1 INTRODUCTION [Fe-S] Clusters in Biology Iron (Fe) is one of the oldest and most abundant metals on this planet (1). Consequently, it is not surprising that Fe became widely incorporated into living systems such as proteins (2). Protein bound Fe frequently appears in the form of an iron sulfur ([Fe- S]) cluster where Fe ions in the +2 or +3 oxidation state are bound to sulfides (S-2) (3). Despite indications that [Fe-S] proteins may be evolutionarily ancient, they were not discovered until the 1960s as a new form of spectroscopy, electron paramagnetic resonance (EPR) spectroscopy, grew in prevalence (4, 5). The first characterized Fe-S containing proteins were ferredoxins from plant and bacterial systems (4, 6-8). These ferredoxin Figure 1.1 Common types of [Fe-S] clusters in biology. Left to right, representations of a [2Fe-2S] cluster (PDB file 1X0G), [3Fe-4S] cluster (PDB file 3RGW), and a [4Fe-4S] cluster (PDB file 3CB8). Irons are shown in rust, sulfides are represented in yellow, and cysteine ligands from the protein are shown in grey. 2 proteins were identified as having functionally important [Fe-S] clusters by their distinct EPR signatures (4, 9, 10). [Fe-S] clusters commonly exist in biology in [2Fe-2S], [3Fe-4S], or [4Fe-4S] forms (Figure 1.1) although more complex clusters, such as those found in hydrogenase and nitrogenase enzymes, are also known (4, 11). [Fe-S] clusters are generally bound to the enzyme through cysteine or histidine coordination to the irons of the cluster. Less commonly, alternative cluster ligands including serine, aspartic acid, or glutamic acid are observed (3, 4, 12). The sulfide ions of the cluster are not covalently bound to the enzyme. Function of [Fe-S] clusters in Biological Systems For many years, electron transfer was thought to be the sole function of [Fe-S] clusters as their redox properties are well suited for this purpose (9). [Fe-S] clusters adopt a wide range of redox potentials within the protein environment (3). Clusters found in the two [4Fe-4S] cluster ferredoxin protein, clostridial Fd, can have redox potentials as low as -700 mV (13). Meanwhile, in [4Fe-4S] cluster high-potential iron proteins, redox potentials as high as +400 mV have been observed (3, 14). [Fe-S] clusters can easily switch between oxidations states enabling the cluster to gain or lose an electron without dissociating (4, 9). For the three most common [Fe-S] clusters mentioned earlier, the typical oxidation states are 2+ and 1+ for the [2Fe-2S] cluster, 1+ and 0 for the [3Fe-4S] cluster, and 3+, 2+, and 1+ for a [4Fe-4S] cluster (9). In the [2Fe-2S]2+ cluster, two Fe3+ ions are antiferromagnetically coupled. The addition of an electron generates the reduced (+1) state resulting from a Fe2+-Fe3+ pair. An electron can 3 also be delocalized across an Fe-Fe pair. For example, the oxidized [3Fe-4S]1+ cluster contains three Fe3+ ions while the reduced [3Fe-4S]0 cluster has one Fe3+ and an electron delocalized Fe2.5+-Fe2.5+ pair (9, 15). The [4Fe-4S] cluster is the most complex of the three and contains two antiferromagnetically coupled Fe-Fe pairs (16). The 3+, 2+, and 1+ cluster states each contain a Fe2.5+-Fe2.5+ pair coupled to a Fe3+-Fe3+, Fe2.5+-Fe2.5+, or Fe2+- Fe2+ pair respectively (9, 16, 17). In addition to electron transfer, [Fe-S] clusters are known to play important roles in the regulation for gene expression (4). The bacterial transcription factor fumarate nitrate reduction (FNR) serves as an oxygen sensor in E. coli. When the FNR protein is exposed to oxygen, its [4Fe-4S]2+ cluster is converted to a [2Fe-2S]2+ cluster. This cluster degradation initiates the switch from aerobic to anaerobic metabolism (18). The iron- sulfur-cluster regulator (IscR) contains a [2Fe-2S] cluster that plays a role in the regulation of the iscRSUA operon used for Fe-S cluster assembly in bacteria and mitochondria. The loss of the IscR [2Fe-2S] cluster, usually caused by oxidative stress, results in the repression of the iscRSUA operon (19). SoxR, another transcription factor that responds to O - 2 stress in E. coli (4, 20). The change in oxidation state of the SoxR [2Fe-2S] cluster severs as a switch between the protein active and repressed states (21). [Fe-S] clusters have been shown to hold critical structural roles in certain proteins including the DNA repair enzymes MutY and endonuclease III (22). The [4Fe-4S] cluster in MutY positions key catalytic residues that enable the protein to bind to damaged DNA (23, 24). Once the DNA is bound, MutY can excise the damaged nitrogenous base (25). Similarly, endonuclease III removes damaged pyrimidines from DNA. Its [4Fe-4S] cluster 4 is reported to be involved in positioning basic residues on the enzyme to create an optimized surface for DNA binding (26). In ATP-dependent DNA helicases, XPD and FancJ, [Fe-S] clusters have been proposed to provide structural stability that enables the enzymes to separate duplex nucleic acids into single strands (27, 28). Several enzymes have been found to rely on [Fe-S] clusters for enzyme catalysis. Aconitase is an enzyme that contains a [4Fe-4S] cluster and facilitates the isomerization of citrate to isocitrate (4). One of the irons in the [4Fe-4S] cluster is not bound to the enzyme and can coordinate to the hydroxyl and carboxylate groups on citrate. This coordination allows aconitase to hold the molecule and remove a hydroxyl group during isomerization (29, 30). Nitrite reductase enzymes (NiRs) use a [4Fe-4S] cluster as a “reservoir of reducing equivalents” to enable the six electron reduction of nitrite to ammonia (31, 32). Lastly, in one of the largest enzymatic superfamilies, the radical S-adenosyl-L-methionine (SAM) enzyme family, a [4Fe-4S] cluster holds the vital role of radical generator (33, 34). Diversity in the Radical SAM Enzyme Superfamily The radical SAM (RS) enzyme family utilizes a [4Fe-4S] cluster as a catalytic cofactor (33, 35). Since their classification in 2000 by Heidi Sofia, more than 100,000 protein sequences have been identified as belonging to the RS enzyme superfamily (36- 39). These enzymes have been found in all six kingdoms of life and are involved in catalyzing reactions in a wide array of metabolic pathways as shown in table 1.1. RS enzymes are known to catalyze DNA repair, protein radical generation, hydrogenase maturation, epimerization reactions, isomerization reactions, and much more (33, 40, 41). 5 Despite their variety of biological functions, all RS enzymes utilize a [4Fe-4S] cluster in conjunction with the small molecule SAM to initiate radical reactions via the highly reactive 5′-deoxyadenosyl radical (5′-dAdo•) (Figure 1.2) (36, 42). Enzyme Function Kingdom Pyrrolysine biosynthesis PylB (43) Archaebacteria protein Anaerobic glucose PFL-AE (42, 44) Eubacteria metabolism VIPERIN (45) Virus inhibition Fungus [FeFe] hydrogenase HydG (46) Protista maturase Molybdenum cofactor MoaA (47) Plantae biosynthesis SPL (33) DNA repair Animalia Table 1.1. Examples of RS enzymes in all six kingdoms of life and their functions. Abbreviations used are pyrrolysine (Pyl), pyruvate formate-lyase activating enzyme (PFL-AE), virus inhibitory protein, endoplasmic reticulum-associated, IFN-inducible (VIPERIN), and spore photoproduct lyase (SPL). RS Enzyme Structure and Characterization The common fold of RS enzymes is a full or partial triosephosphate isomerase Figure 1.2. SAM is homolytically cleaved to form methionine and the 5′-dAdo•. When the [4Fe-4S]2+ is reduced to the +1 state, an inner sphere electron transfer occurs between the cluster and the sulfonium of SAM cleaving the S-C5′ bond. 6 (TIM) barrel that houses an active site within the protein (Figure 1.3) (48-53). The completeness of the RS enzyme TIM barrel correlates with the size of the substrate with full (ꞵ/α)8 TIM barrels found in enzymes that act on peptides or small molecules, and partial TIM barrels with more open conformations tending to be found in enzymes catalyzing reactions on larger substrates such as proteins (48, 54). TIM barrels are often easily accessible by solvent; most structurally characterized radical SAM enzymes contain protein elements such as flexible loops that likely serve to close off the active site when substrate is present (53). All RS enzymes contain at least one [4Fe-4S] cluster in the protein active site. Several enzymes have additional auxiliary clusters whose precise function is generally unknown (55). The RS [4Fe-4S] cluster is commonly coordinated by three cysteines in a CX3CX2C motif binding to three of the four irons of the cluster (36). Non-canonical exceptions such as ThiC, which utilizes a CX2CX4C motif, and HmdB, coordinating its cluster through a CX5CX2C motif are known (56, 57). The unique iron of the cluster is the one remaining iron of the [4Fe-4S] cluster that is not coordinated by a protein cysteinal residue (58). SAM binds to this unique iron through its methionine and carboxylate moieties, providing close contact and orbital overlap between the SAM sulfonium sulfur and the [Fe-S] cluster, which is thought to be critical for catalysis (35, 59-61). The binding conformation of SAM is stabilized by the conserved motifs GGE and GXIXGX2E which coordinate the methionine and adenine moieties respectively (48, 50). Radical Initiation Pathway 7 RS enzymes use SAM either as a cofactor or, more commonly, a cosubstrate as shown in Figure 1.4 (62). When SAM is used as a cofactor, figure 1.4 left, the S-C5′ bond is homolytically cleaved to generate the 5′-dAdo• and methionine. The 5′-dAdo• reacts with substrate, and ultimately a product radical abstracts an H• from 5′-dAdoH to re- generate the 5′-dAdo•, which recombines with the methionine, with loss of an electron to the [4Fe-4S]2+ cluster, to regenerate SAM bound to the [4Fe-4S]+ cluster. When SAM serves as a cosubstrate, figure 1.4 right, the 5′-dAdo• and methionine are generated as mentioned before. However, after hydrogen abstraction, the methionine and 5′-dAdoH must leave the active site while being replace by a new SAM molecule to enter the next catalytic cycle (62). Figure 1.3. Representations of the TIM barrels present in three different RS enzymes. PFL-AE (left, PDB 3CB8) has a mostly open active site with a crescent (ꞵ/α)6 barrel. LAM (middle, PDB 2A5H) has a complete (ꞵ/α)8 barrel. MoaA (right, PDB 1TV7) contains an incomplete (ꞵ/α)6 TIM barrel with a lateral opening and an auxiliary cluster shown at the opposite end of the active site. Iron atoms are shown in rust and sulfur atoms are shown in yellow. 8 PFL-AE and PFL as a Model RS Enzyme System Pyruvate formate-lyase activating enzyme (PFL-AE) is one of the earliest characterized RS enzymes and a founding member of the RS enzyme subgroup, the glycyl radical enzyme activating enzymes (GRE-AE) (36, 63, 64). GRE-AE proteins activate enzyme substrates by abstracting a hydrogen from a glycine residue to generate a glycyl radical (Gly•) (65). The PFL-AE substrate, pyruvate formate-lyase (PFL), plays a critical role in anaerobic glucose metabolism by converting pyruvate and CoA to formate and Figure 1.4. Alternate uses of SAM as a cofactor or a cosubstrate in RS enzymes. When SAM is a cofactor, left, the S-C5′ bond is cleaved to generate the 5′-dAdo• and methionine (Met). The C5′ radical abstracts a hydrogen from substrate forming substrate radical (S•) and 5′-dAdoH which is converted back to 5′-dAdo• when a product radical (P•) abstracts a H5′. The 5′-dAdo• will then reform with methionine creating SAM bound to the reduced cluster that can be used for subsequent radical initiation. When SAM is a cosubstrate, right, the 5′-dAdo• and Met are formed. The C5′ radical removes a hydrogen from substrate forming 5′- dAdoH which must fall out of the active site with Met before a new SAM molecule enters and radical generation can continue. Iron is shown in rust and sulfur shown in yellow. 9 acetyl-CoA (42). Due to ease of over expression in E. coli cells and relative stability, PFL- AE and PFL have become models for their respective enzyme classes (65). PFL-AE PFL-AE was first discovered when Knappe et al. found that PFL activity depended on the presence of an additional uncharacterized enzyme termed the activating enzyme (66). This 28 kDa activating enzyme was shown to require exogenous iron and the small molecule SAM to activate PFL (64, 67, 68). Later work by Broderick et al. utilized absorption spectroscopy, variable temperature magnetic circular dichroism, EPR, and resonance Raman spectroscopies to show that anaerobically prepared PFL-AE contained a mix of [4Fe-4S]2+ and [2Fe-2S]2+ clusters (64). After dithionite (DT) reduction, however, 10 only the [4Fe-4S]2+ clusters remained. The reduction of the [4Fe-4S]2+ in the presence of SAM lead to a +1 state which was proposed to be the active form of the PFL-AE cluster (64). Later studies using Mössbauer spectroscopy to examine the non-reduced, anaerobically isolated PFL-AE revealed that the enzyme could contain a mixture of four different [Fe-S] clusters. The predominant form was the cuboidal [3Fe-4S]+ cluster which accounted for 66% of the total Fe. The remaining Fe was identified to be in [2Fe-2S]2+ Figure 1.5. The crystal structure of PFL-AE (PDB 3CB8) with SAM bound to the cluster. SAM is shown in black with nitrogens in blue, oxygens in bright red, sulfur in yellow, and iron in rust. 11 clusters (12%) [4Fe-4S]2+ clusters (8%) and a linear [3Fe-4S]+ cluster (10%) with the remaining iron existing as exogenous iron(II). Reduction of PFL-AE with DT resulted in 66% [4Fe-4S]2+ and 12% [4Fe-4S]+. The resulting conclusion was that the [4Fe-4S] cluster was the catalytically relevant cluster in PFL-AE (69). With 245 residues, PFL-AE is one of the smallest characterized RS enzymes (33, 54). Its structure is simple, as shown in figure 1.5, with a single domain consisting of a partial (α/ꞵ)6 TIM barrel and a flexible loop that serves to close off the active site when substrate is present (50). PFL-AE utilizes the canonical CX3CX2C motif to ligate a [4Fe- 12 4S] cluster and has no auxiliary clusters (50). When PFL-AE is photoreduced, a process in which Tris serves as sacrificial electron donor in the presence of 5-deazariboflavin and light, the reduced [4Fe-4S]+ cluster gives rise to a rhombic EPR signal (g = 2.02, 1.94, 1.88) (Figure 1.6a) (70). In the early 2000s, it was shown that PFL-AE photoreduced in the presence of SAM showed a nearly axial signal (g = 2.01, 1.88, 1.87) (Figure 1.6b). This Figure 1.6. EPR spectra of the reduced [4Fe-4S]+ cluster of PFL-AE A) PFL-AE (0.7 mM) photoreduced for 1 hr using 5-deazariboflavin as photoreductant. B) PFL-AE (0.78 mM) photoreduced as in (A) with two molar equivalents of SAM added. EPR parameters were : T = 12 K, power microwave power = 20 µW, microwave frequency = (A) 9.483 GHz or (B) 9.476 GHz, modulation amplitude = (A) 8.231 G or (B) 9.571 G. Reprinted with permission from (35). 13 perturbation in the EPR signals suggested potential binding interactions between the PFL- AE cluster and SAM (35, 70). To probe whether SAM binds directly to the [4Fe-4S] cluster in PFL-AE, electron nuclear double resonance (ENDOR) spectroscopy, a method that probes interactions resulting from the coupling of unpaired electrons with NMR active nuclei, was used (35, 59, 71). 17O-carboxylato and 15N-amino-labeled SAMs were added to reduced PFL-AE and the complex was subjected to ENDOR spectroscopy. Samples generated with 17O- carboxylato-labeled SAM showed a broad, asymmetric feature at 12.2 MHz congruent with the hyperfine coupling due to 17O coordinated directly to an iron of the [4Fe-4S]+ cluster Figure 1.7. ENDOR spectroscopic characterization of the PFL-AE/SAM complex. (A) Data obtained when 17O labeled SAM was bound to the reduced [4Fe-4S]+ cluster in PFL-AE. (B) Data obtained when either natural abundance SAM (containing 14N at the amino moiety) or 15N labeled SAM was bound to the reduced [4Fe-4S]+ cluster in PFL-AE. Conditions: T = 2 K, νMW = 34.9 GHz, rf pulse length = 60 µS. For (A): Davies ENDOR, MW pulse lengths = 80, 40, 80 ns, number of averaged transients at each point: 17O = 288, unlabeled = 200. For (B): 15N-labeled, Davies ENDOR: MW pulse lengths = 80, 40, 80 ns, number of averaged transients: 15N-labeled = 80, unlabeled = 624. Figure adapted from (59). 14 (Figure 1.7a). In the presence of 15N-amino-SAM, the signal due to natural abundance 14N- SAM was gone, and a new signal appeared arising from the 15N coordinated to the [4Fe- 4S]+ cluster (Figure 1.7b). Taken together, the ENDOR data showed that SAM binds through its amine and carboxyl moieties to the unique iron on the PFL-AE [4Fe-4S] cluster (59). In a series of elegant experiments conducted by Henshaw et al., it was shown that photoreduced PFL-AE could catalyze a single turnover, activating one molar equivalent of PFL (Figure 1.8). Photoreduction of PFL-AE using 5-deazariboflavin and light provided Figure 1.8. PFL-AE catalyzed single turnover activation of PFL. A) X-band EPR spectra of reduced PFL- AE/SAM in the absence of PFL. PFL-AE was reduced with 5-deazariboflavin for the indicated lengths of time in min. B) X-band EPR spectra of the Gly• generated by the addition of PFL to reduced PFL- AE/SAM. EPR conditions were T = 12 K (A) or 60 K (B), microwave frequency = 9.48 GHz, microwave power = 2 mW (A) or 20 μW (B), and modulation amplitude 10.084 G (A) or 5.054 G (B). Spin quantitation of the EPR signals in (A) and (B) with additional time lengths are shown in (C). The spin quantitation points are plotted against illumination time of the sample. Reprinted with permission from (72). 15 more complete reduction of the [4Fe-4S] cluster (85% after 60 min) than previously observed using DT (<20%) (Figure 1.8c) (42, 69, 72). More importantly, photoreduction allows for the removal of excess reductant by putting samples in the dark, which facilitated the experiments described here. PFL-AE samples were reduced for 0, 1, 2, 5, 10, and 30 min then combined with either SAM (figure 1.8a) or SAM and PFL (figure 1.8b) in the dark and observed with EPR spectroscopy. Samples to which only SAM was added show a [4Fe-4S]+ EPR signal, while those to which SAM and PFL were added show a glycyl radical (Gly•) signal; spin quantitation of the [4Fe-4S]+/SAM spectra and the Gly• spectra revealed a 1:1 ratio between the two. These results provided the first direct evidence for radical SAM enzymes that the [4Fe-4S]+ cluster provides the electron necessary for reductive SAM cleavage (72). PFL PFL, the substrate for PFL-AE, was first characterized as a glycyl radical enzyme (GRE) in 1992 when it was discovered that the active form of the enzyme (PFLactive) contained a radical on its glycine 734 residue (73). By using deuterated PFL (2-2H-glycine labeled PFL) in conjunction with proton nuclear magnetic resonance (1H NMR) and mass spectroscopy, Frey et al. were able to show the Gly734 hydrogen is abstracted directly by the 5′-dAdo• generated by SAM cleavage (74). To investigate the stereospecificity of the Gly734 H atom abstraction, peptides that mimic the protein’s Gly734 containing finger loop region, residues 730-741, were developed (74-76). A modified PFL activation assay that examined the formation of 5′-dAdoH over time was used to determine if the 16 Figure 1.9. PFL is activated by PFL-AE generating a radical on the Gly734residue. When PFLinactive is in the presence of PFL-AE (PDB 3CB8), the radical loop moves out in order to expose Gly734to the PFL-AE active site. The pro-S hydrogen is then abstracted generating the Gly734• and PFLactive. Once the Gly734• is in the PFL active site, it can initiate the conversion of pyruvate and CoA to formate and acetyl-CoA through an internal radical initiation. The radical loop is shown in pink, Gly734• is denoted by a red sphere. PFL was generated by modifying the PDB file 2PFL. Figure is adapted from (61). synthesized peptides could promote 5′-dAdoH formation (63, 74). Assays that used a peptide where the Gly734 residue was replaced with L-alanine showed no increase in 5′- dAdoH production over a 20 min period. However, when the Gly734 was replaced with D- alanine, the amount of 5′-dAdoH generated over 20 min was greater than that observed in assays with the non-mutated peptide. These assay results showed that the abstraction of a hydrogen from Gly734 is stereospecific for the pro-S hydrogen (74). PFL is a homodimer comprised of two 85 kDa subunits. The primary fold in each monomer is an antiparallel (ꞵ/α)10 barrel which makes up the enzyme active site. Within the ꞵ/α barrel are the active site residues Gly734, Cys418, and Cys419. The Gly734 and Cys418/Cys419 residues are located on opposing finger loops with a Cα-Cα distance of 17 Figure 1.10. The PFL mechanism to convert pyruvate and CoA to formate and acetyl-CoA. The Gly734• abstracts a hydrogen from Cys418 initiating the catalytic cycle which proceeds though a ping-pong mechanism. The Cys418• is regenerated in the last step allowing another cycle to commence. Figure adapted from (77, 80). about 4.8 Å between Gly734 and Cys419 (77). The PFL monomer is proposed to have two conformational states (78). The fist is a closed conformation where the Gly734 residue is located 8 Å from the surface of the protein (77, 78). The second is an open conformation where, in the presence of PFL-AE, the C-terminal radical domain containing Gly734 undergoes a rotation rendering Gly734 in a more solvent exposed location. The involvement of this open conformation is supported by hydrogen/deuterium exchange experiments and circular dichroism spectroscopy of the PFL-AE/PFL complex (78). Considering the two proposed conformational states of PFL, a simplified description of PFL activation by PFL-AE can be given as follows. Inactive PFL (PFLinactive) is in a closed conformational state until it encounters PFL-AE. In the presence of PFL-AE, PFLinactive shifts into an open conformation and the pro-S hydrogen is abstracted from Gly734 by the 5′-dAdo•. Once the Gly734• is generated, PFLactive shifts back into a closed conformation and the PFL catalytic cycle can begin (Figure 1.9) (78). It was proposed that PFL is able to catalyze the conversion of pyruvate and CoA to formate and acetyl-CoA through a ping-pong mechanism shown in Figure 1.10 (78, 79). 18 The Gly• generates a cysteinyl radical on the Cys418 residue located on the opposite side of the PFL active site. The cysteinyl radical then generates an oxygen-based radical on pyruvate through a nucleophilic attack of the C2′. The C1-C2 bond on the pyruvate radical homolytically cleaves to generate a carbonyl radical that abstracts a hydrogen from Cys419 generating another cysteinyl radical and formate. The newly formed formate leaves the active site and CoA enters. CoA is then converted to a sulfur-based radical by nucleophilic attack from the Cys419 radical. Lastly, the CoA radical reacts with the Cys418 bound carbonyl generating acetyl-CoA and regenerating the Cys418• allowing the catalytic cycle to repeat (80). RS Enzyme Mechanism The radical initiation mechanism originally proposed for RS enzymes involved the direct formation of the 5′-dAdo• upon SAM cleavage (Figure 1.2). In this mechanism an external electron, for example from flavodoxin, reduces the [4Fe-4S] cluster from the +2 to the +1 state (81). The [4Fe-4S]+ cluster then facilitates an inner sphere electron transfer from the cluster to the sulfonium ion of SAM. This causes homolytic cleavage of the S-C5′ bond of SAM, resulting in formation of methionine and free 5′-dAdo• (33, 36, 82). The 5′- dAdo• is highly reactive and therefore able to serve as a hydrogen atom abstractor from a variety of substrates (83). Discovery and Characterization of Ω 19 The original RS enzyme mechanism was called into question when work with PFL- AE showed the existence of an intermediate formed during RS enzyme initiation (84, 85). While conducting rapid freeze quench (RFQ) experiments designed to capture reaction intermediates, a species was trapped that exhibited an unusual EPR signal. This species, dubbed Ω, produced an EPR signal that reached its maximum intensity when reduced PFL- AE was rapidly mixed with PFL and SAM and quenched at 500 ms (Figure 1.11a). It was confirmed that Ω was not the 5′-dAdo• as its EPR signal lacked the predicted hyperfine coupling. The organic radical 5′-dAdo• would show an isotropic signal with a g value near 2.0. The spectrum would also show well-resolved splittings that resulted from hyperfine couplings between the unpaired electron in the C5′ 2p π orbital and the two equivalent protons from 1H2C(5′) (Figure 1.11b). The Ω signal, however, was axial and had g-values Figure 1.11. Figure adapted from (84). EPR spectra of the Ω signal and the predicted 5′-dAdo• signal. A) the Ω signal generated after rapidly mixing reduced PFL-AE and (PFL/ SAM) and quenching at 500 ms. EPR parameters were 12 K, microwave frequency = 9.23 GHz, microwave power = 1 mW, 100 KHz, and modulation amplitude = 8 G. B) Simulation (EasySpin) of the 5′-dAdo• with an axial g-tensor with hyperfine interactions from the two 1H from a potential 1H2C(5′). Euler angles ꞵ = 30° and 150°. 20 of g║ = 2.035, and g⊥ = 2.004 (84) (Figure 1.11a). EPR spectroscopic analysis showed that Ω formed in between the oxidation of the [4Fe-4S]+ cluster to the EPR silent +2 state and before the formation of the Gly•. As the [4Fe-4S]+ cluster signal (g = 2.01, 1.88, 1.87) disappeared from the spectrum, the Ω signal increased (35, 84). When samples were quenched at longer time points such as 750 ms or 1 s, the Gly• signal became more prevalent while the Ω signal became weaker (Figure 1.12a) (73, 84). These EPR spectra demonstrate that Ω is an intermediate along the pathway to generate active PFL. To test if Ω is catalytically competent, samples quenched at 500 ms were cryo-annealed up to 220 K (Figure 1.12b). With the increase of temperature, the Ω signal would decrease proportionally to the increase in Gly• signal. Thus, Ω was shown to be a true enzyme intermediate. Furthermore, Ω was positioned in such a way as to allow H atom extraction from Gly734 even at cryogenic temperatures (84). To see if Ω involved the paramagnetic PFL-AE cluster, the intermediate was generated with 57Fe-enriched PFL-AE. In this form of PFL-AE, the [4Fe-4S] cluster is comprised of isotopically labeled 57Fe and sulfide (69). When 57Fe-enriched PFL-AE was used to generate Ω, the resulting CW ENDOR signal was nearly identical to the signal observed for reduced 57Fe-enriched PFL-AE (Figure 1.12c). This data showed that the spin ½ which gave rise to the Ω EPR signal dwelt on the [4Fe-4S] cluster of PFL-AE (16, 84). To see if the [4Fe-4S] cluster was coupled to a fragment of SAM, Ω was generated with isotopically labeled, [13C 15 10, N5]-adenosine-SAM and observed using ENDOR spectroscopy. The resulting continuous wave ENDOR spectra revealed strong coupling between the reduced cluster and a 13C nucleus (Figure 1.12d). Mims pulse technique 21 revealed significantly weaker coupling arising from another 13C positioned further from the [4Fe-4S]+ cluster. Similar experiments were conducted using Ω generated from [methyl-13C]-SAM. The ENDOR spectrum showed significantly weaker 13C coupling than when the adenosyl moiety was 13C-labeled (figure 1.12e). Thus it was inferred that Ω involves the adenosine and not the methionine moiety of SAM (84). The isotopic labeling and ENDOR studies showed the Ω EPR signal was the result of a C from the 5′-dAdo fragment covalently bound to the paramagnetic [4Fe-4S] cluster. Although the ENDOR spectra showing 13C coupling to the [4Fe-4S] cluster was generated with uniformly labeled 13C-adenosyl-SAM, the 13C5′ is the only logical carbon to bind with the unique cluster iron. The weaker 13C coupling observed was assigned to the 13C4′ of 5′- dAdo fragment. The Ω intermediate is structurally similar to the adenosine cobalamin (AdoCbl) cofactor utilized by B12 enzymes. In B12 enzymes, radical catalysis is initiated through the cleavage of the Co(III)-[5′-C]-deoxyadenosyl bond which releases the 5′-dAdo• (86, 87). This provides a remarkable similarity between the two enzyme families as their enzymatic 22 Figure 1.12. The discovery and characterization of Ω. A) X-band spectra at 12 K of reduced PFL-AE mixed with (PF/SAM) and quenched at various times showed on the figure. Maximum Ω signal is observed at 500 ms; longer quench times show the appearance of the Gly• signal. B) (Top) PFL-AE and (PFL/SAM) quenched at 500 ms and progressively annealed to higher temperature for the given times. The Ω signal at 40 K shows rapid relaxation and the amplitude for the Gly• at 12 K is diminished. Residual intensities at both temperatures were subtracted out in the shown spectra. (Bottom) Populations of Ω and Gly• derived from EPR spectra. Percentages relative to the final Gly• concentration (Sum) taken at 220 K. EPR conditions: microwave frequency = 9.23 GHz, microwave power = 1 mW, 100-kHz, modulation amplitude = (A) 13 G, (B) 8 G, T = (A) 12 K, (B) 12 K and 40 K. C) 57Fe CW ENDOR for 57Fe-enriched photoreduced PFL-AE and Ω. (Top) 57Fe-enriched (red) and natural-abundance (gray) RFQ samples observed with CW ENDOR spectroscopy. (Bottom) randomly hopped stochastic and frequency sweep CW ENDOR for 57Fe-enriched reduced PFL-AE (16). Conditions for spectra: microwave frequency = 35.45 GHz and 35.07 GHz for RFQ and 57Fe-enriched reduced PFL-AE, respectively; microwave power = 1 mW; 100-kHz modulation amplitude = 1.3 G; rf sweep rate = 1 MHz/s; stochastic CW ENDOR cycle, rf-on = 3 ms, rf-off = 1 ms; sample collection time = 3 ms; and T = 2 K. D) 13C CW ENDOR for [adenosyl-13C10] SAM. The green dashed lines denote the simulation best matched to the axial hyperfine tensor. Simulation parameters: aiso = 9.4 MHz, 2T = 5.3 MHz, and b = 90°. Conditions: microwave frequency = 35.39 GHz, microwave power = 1 mW, 100-kHz modulation amplitude = 1.3 G, rf sweep rate = 1 MHz/s, and T = 2 K. (Inset) Mims ENDOR spectrum. Conditions: microwave frequency = 35.20 GHz; MW pulse length, (π /2) = 50 ns; t = 500 ns; and T = 2 K. E) ) Mims ENDOR spectrum from [methyl-13C] SAM. Conditions: microwave frequency = 35.08 GHz; MW pulse length, (π/2) = 50 ns; t = 500 ns; and T = 2 K. Figures and descriptions reprinted with permission from (84). 23 pathways are both shown to include an organometallic intermediate comprise of a metal- C5′ bond that is cleaved to generate 5′-dAdo• (84, 86, 87). The discovery of Ω prompted the question of what is its mechanism of formation? It is possible that the 5′-dAdo• forms first through reductive cleavage of the S-C5′ bond of SAM. The C5′ of the 5′-dAdo• then adds to the unique Fe of the [4Fe-4S] cluster generating Ω (Figure 1.13a). Alternatively, Ω may form through concerted reductive cleavage initiated by interactions between the unique iron and the sulfonium on SAM. The S-C5′ bond would be broken and the Fe-C5′ bond would immediately form producing Ω (figure 1.13b). In both mechanisms, the Fe-C5′ of Ω would have to be cleaved to generate the 5′-dAdo• (84). Figure 1.13. Two potential mechanisms of radical initiation by the RS enzyme family. Path A where the S-5′C bond is cleaved to generate the 5′-dAdo• and methionine. The 5′C then binds to the unique iron generating Ω which is cleaved to generate the 5′-dAdo• and methionine. Path B shows a direct formation of Ω, followed by the Fe-C5′ bond cleavage which generates 5′-dAdo• and methionine. 24 Research Objectives The goals of this research were inspired by the discovery of Ω and its possible role in the RS enzyme mechanism. My first goal was to determine if Ω was specific to PFL-AE or if it was a ubiquitous intermediate in the RS enzyme superfamily. Second, as this intermediate contains the 5′-dAdo moiety of SAM, I wanted to see if Ω could be used to generate and capture the elusive 5′-dAdo•. Since my first two goals were focused on the isolation of intermediates Ω and the 5′-dAdo•, my final goal was to establish a method for better understanding the process of intermediate formation. In order to accomplish this goal, I needed to determine if an analog of SAM, S-3′,4′-anhydroadenosyl-L-methionine (anSAM), could serve as a cosubstrate for PFL-AE. ENDOR spectroscopy and RFQ studies necessary for generating Ω were conducted in collaboration with the Brian Hoffman Lab at Northwestern University. Determining the Ubiquity and Further Characterization of Ω My first objective was to determine if Ω is ubiquitous to the RS enzyme superfamily. Although PFL-AE is often used as a prototype for RS enzymes, it does have some unique characteristics. As previously mentioned, the Gly734 residue is located 8 Å away from the protein surface forcing PFL to undergo a significant conformational change before it can be activated by PFL-AE (78). It is possible that this structural re-arrangement makes Ω formation necessary only for the PFL-AE/PFL complex. Perhaps Ω forms to hold the highly reactive 5′-dAdo• until PFL is in position. Alternatively, Ω could be associated with the GRE-AE subclass or only RS enzymes which use SAM as a cosubstrate. It is also 25 possible Ω is a key intermediate in the RS enzyme mechanism and is universal to the superfamily. To confirm that Ω is a key intermediate in the RS enzyme superfamily, Ω would need to be detected in a variety of RS enzymes. Due to the enormity of the RS enzyme superfamily, it was not freezable to test for Ω formation in each of the 113,775 characterized members (39). Instead, seven enzymes that represent the major subgroups of classified RS enzymes were selected. The subgroups included a second GRE-AE in addition to PFL-AE, enzymes with small molecule substrates, epimerase enzymes, two enzymes that used SAM as a cofactor, and an RS enzyme with an auxiliary cluster. RFQ experiments were conducted in the same manner for all enzymes: the reduced enzyme was rapidly mixed with (substrate/SAM) and quenched at 500 ms. The samples were then observed using EPR spectroscopy to determine if Ω had formed. Previous work from the Broderick lab showed that Ω was formed by the C5′ of the 5′-dAdo fragment binding to the unique iron of the [4Fe-4S] cluster (84). I wanted to know if Ω also contained the methionine fragment of SAM bound to the unique iron. To determine if the amino nitrogen on the methionine fragment coordinated to the [4Fe-4S] cluster, 15N-met-SAM was used to generate Ω and the intermediate was observed with ENDOR spectroscopy. Capturing and Characterizing the 5′-dAdo• My second goal was to see if Ω could be used to generate the 5′-dAdo• in a spectroscopically observable manner. In B12 enzymes, light has been used to initiate 26 cleavage of the Co(III)-C5′ bond forming the 5′-dAdo• (88, 89). As Ω and the B12 intermediate are structurally alike, we theorized a similar process could be used to cleave the Fe-C5′ bond in Ω (84). Because sulfonium centered molecules have photolytic properties, a control experiment was designed using reduced protein and SAM without substrate present (90, 91). The samples were photolyzed with 450 nm light to cleave the S- C5′ bond and kept at cryogenic temperatures to prevent the carbon radical from reacting with neighboring protein residues. After photolysis, the samples were observed using EPR spectroscopy. Photolysis samples were also made with isotopically labeled SAMs to carefully determine if the generated radical was the elusive 5′-dAdo•. Further Studies of RS Enzyme Intermediates and their Mechanisms of Formation An effective way to study the mechanism of formation of the highly reactive 5′- dAdo• is by using an allylic analog of SAM, anSAM. When anSAM is cleaved, methionine and a 3′,4′-anhydroadenosyl radical (anAdo•) are formed (Figure 1.14). Due to allylic stabilization, the anAdo• is longer lived than the 5′-dAdo• and has been characterized using EPR spectroscopy (92, 93). Additionally, anSAM was found to be a true cofactor for the RS enzyme lysine 2,3, aminomutase (LAM) allowing for the conversion of α to ꞵ-lysine (92). Although anSAM has served as a cofactor for LAM, it has never been used as a cosubstrate for PFL-AE (92). I wanted to confirm that this SAM analog could function as a cosubstrate for PFL-AE by preforming coupled activity assays and generating analogs of the RS enzyme intermediates Ω and 5′-dAdo•. RFQ experiments with PFL- 27 Figure 1.14. A diagram representing the homolytic cleavage of SAM, left, and anSAM, right. The unpaired electron on the anAdo• is depicted as being delocalization across the 5′, 4′, and 3′ carbons. AE/PFL/anSAM provided valuable information into the mechanism of Ω formation. Photolysis studies of reduced PFL-AE and anSAM supported recent work from the Broderick Lab suggesting the SAM ribose ring conformation in the RS enzyme active site is responsible for the regioselectivity of cryogenic photoinduced reductive S-C bond cleavage in SAM. 28 References 1. Tan L, Chi-Lung Y. Abundance of chemical elements in the earth's crust and its major tectonic units. International Geology Review. 1970;12(7):778-86. 2. Maret W. The metals in the biological periodic system of the elements: concepts and conjectures. International journal of molecular sciences. 2016;17(1):66. 3. Meyer J. Iron–sulfur protein folds, iron–sulfur chemistry, and evolution. JBIC Journal of Biological Inorganic Chemistry. 2008;13(2):157-70. 4. Lill R. Function and biogenesis of iron–sulphur proteins. Nature. 2009;460(7257):831. 5. Ahmad R, Kuppusamy P. Theory, instrumentation, and applications of electron paramagnetic resonance oximetry. Chemical reviews. 2010;110(5):3212-36. 6. Mortenson L, Valentine R, Carnahan J. An electron transport factor from Clostridiumpasteurianum. Biochemical and biophysical research communications. 1962;7(6):448-52. 7. Valentine R, Mortenson L, Carnahan J. The hydrogenase system of Clostridium pasteurianum. J biol Chem. 1963;238:1141-4. 8. Malkin R, Rabinowitz JC. The reconstitution of clostridial ferredoxin. Biochemical and biophysical research communications. 1966;23(6):822-7. 9. Beinert H, Holm RH, Münck E. Iron-sulfur clusters: nature's modular, multipurpose structures. Science. 1997;277(5326):653-9. 10. Beinert H, Sands RH. Studies on succinic and DPNH dehydrogenase preparations by paramagnetic resonance (EPR) spectroscopy1960. 11. Peters JW. Structure and mechanism of iron-only hydrogenases. Current opinion in structural biology. 1999;9(6):670-6. 12. Crain AV, Duschene KS, Peters JW, Broderick JB. Iron-Sulfur Cluster Proteins, Fe/S-S-adenosylmethionine Enzymes and Hydrogenases. In: Kretsinger RH, Uversky VN, Permyakov EA, editors. Encyclopedia of Metalloproteins. New York, NY: Springer New York; 2013. p. 1034-44. 13. Beinert H, Meyer J, Lill R. Iron–sulfur proteins. Encyclopedia of biological chemistry. Amsterdam: Elsevier; 2004. p. 482-9. 29 14. Moore G. Handbook of metalloproteins-Edited by A. Messerschmidt, R. Huber, T. Poulos and K. Wieghardt, John Wiley & Sons, New York, 2001, 2 volumes. ISBN 0-471 62743 7;@ $645. Journal of Organometallic Chemistry. 2002;1(659):203-4. 15. Banci L, Camponeschi F, Ciofi-Baffoni S, Piccioli M. The NMR contribution to protein–protein networking in Fe–S protein maturation. JBIC Journal of Biological Inorganic Chemistry. 2018;23(4):665-85. 16. Cutsail III GE, Telser J, Hoffman BM. Advanced paramagnetic resonance spectroscopies of iron–sulfur proteins: Electron nuclear double resonance (ENDOR) and electron spin echo envelope modulation (ESEEM). Biochimica et Biophysica Acta (BBA)- Molecular Cell Research. 2015;1853(6):1370-94. 17. Wang X-B, Niu S, Yang X, Ibrahim SK, Pickett CJ, Ichiye T, Wang L-S. Probing the intrinsic electronic structure of the cubane [4fe− 4s] cluster: Nature's favorite cluster for electron transfer and storage. Journal of the American Chemical Society. 2003;125(46):14072-81. 18. Crack J, Green J, Thomson AJ. Mechanism of oxygen sensing by the bacterial transcription factor fumarate-nitrate reduction (FNR). Journal of Biological Chemistry. 2004;279(10):9278-86. 19. Schwartz CJ, Giel JL, Patschkowski T, Luther C, Ruzicka FJ, Beinert H, Kiley PJ. IscR, an Fe-S cluster-containing transcription factor, represses expression of Escherichia coli genes encoding Fe-S cluster assembly proteins. Proceedings of the National Academy of Sciences. 2001;98(26):14895-900. 20. Hidalgo E, Demple B. An iron‐sulfur center essential for transcriptional activation by the redox‐sensing SoxR protein. The EMBO Journal. 1994;13(1):138-46. 21. Pomposiello PJ, Demple B. Redox-operated genetic switches: the SoxR and OxyR transcription factors. Trends in biotechnology. 2001;19(3):109-14. 22. Johnson DC, Dean DR, Smith AD, Johnson MK. Structure, function, and formation of biological iron-sulfur clusters. Annu Rev Biochem. 2005;74:247-81. 23. Guan Y, Manuel RC, Arvai AS, Parikh SS, Mol CD, Miller JH, Lloyd RS, Tainer JA. MutY catalytic core, mutant and bound adenine structures define specificity for DNA repair enzyme superfamily. Nature structural biology. 1998;5(12):1058-64. 24. Porello SL, Cannon MJ, David SS. A substrate recognition role for the [4Fe-4S] 2+ cluster of the DNA repair glycosylase MutY. Biochemistry. 1998;37(18):6465-75. 30 25. Lindahl T. DNA Glycosylases in DNA Repair. In: Simic MG, Grossman L, Upton AC, Bergtold DS, editors. Mechanisms of DNA Damage and Repair: Implications for Carcinogenesis and Risk Assessment. Boston, MA: Springer US; 1986. p. 335-40. 26. Kuo C-F, McRee DE, Fisher CL, O'Handley SF, Cunningham RP, Tainer JA. Atomic structure of the DNA repair [4Fe-4S] enzyme endonuclease III. Science. 1992;258(5081):434-40. 27. Rudolf J, Makrantoni V, Ingledew WJ, Stark MJ, White MF. The DNA repair helicases XPD and FancJ have essential iron-sulfur domains. Molecular cell. 2006;23(6):801-8. 28. Jankowsky E, Fairman ME. RNA helicases—one fold for many functions. Current opinion in structural biology. 2007;17(3):316-24. 29. Dai S, Friemann R, Glauser DA, Bourquin F, Manieri W, Schürmann P, Eklund H. Structural snapshots along the reaction pathway of ferredoxin–thioredoxin reductase. Nature. 2007;448(7149):92. 30. Castro L, Tórtora Vn, Mansilla S, Radi R. Aconitases: Non-redox iron–sulfur proteins sensitive to reactive species. Accounts of chemical research. 2019;52(9):2609-19. 31. Crane BR, Siegel LM, Getzoff ED. Sulfite reductase structure at 1.6 Å: evolution and catalysis for reduction of inorganic anions. Science. 1995;270(5233):59-67. 32. Sulc F, Farmer PJ. Bioinorganic Chemistry of the HNO Ligand. The Smallest Biomolecules: Diatomics and their Interactions with Heme Proteins: Elsevier; 2008. p. 429-62. 33. Broderick JB, Duffus BR, Duschene KS, Shepard EM. Radical S- adenosylmethionine enzymes. Chemical reviews. 2014;114(8):4229-317. 34. Frey PA, Magnusson OT. S-Adenosylmethionine: a wolf in sheep's clothing, or a rich man's adenosylcobalamin? Chemical reviews. 2003;103(6):2129-48. 35. Walsby CJ, Hong W, Broderick WE, Cheek J, Ortillo D, Broderick JB, Hoffman BM. Electron-nuclear double resonance spectroscopic evidence that S-adenosylmethionine binds in contact with the catalytically active [4Fe-4S]+ cluster of pyruvate formate-lyase activating enzyme. Journal of the American Chemical Society. 2002;124(12):3143-51. 36. Sofia HJ, Chen G, Hetzler BG, Reyes-Spindola JF, Miller NE. Radical SAM, a novel protein superfamily linking unresolved steps in familiar biosynthetic pathways with radical mechanisms: functional characterization using new analysis and information visualization methods. Nucleic acids research. 2001;29(5):1097-106. 31 37. Landgraf BJ, McCarthy EL, Booker SJ. Radical S-adenosylmethionine enzymes in human health and disease. Annual review of biochemistry. 2016;85:485-514. 38. Akiva E, Brown S, Almonacid DE, Barber 2nd AE, Custer AF, Hicks MA, Huang CC, Lauck F, Mashiyama ST, Meng EC. The structure–function linkage database. Nucleic acids research. 2013;42(D1):D521-D30. 39. Babbitt PC. Structure Function Linkage Database [Online Database]. University of California, San Francisco The Regents of the University of California; 2016 [cited 2020]. 40. Freeman MF, Gurgui C, Helf MJ, Morinaka BI, Uria AR, Oldham NJ, Sahl H-G, Matsunaga S, Piel J. Metagenome mining reveals polytheonamides as posttranslationally modified ribosomal peptides. Science. 2012;338(6105):387-90. 41. Vagstad AL, Kuranaga T, Püntener S, Pattabiraman VR, Bode JW, Piel J. Introduction of d‐Amino Acids in Minimalistic Peptide Substrates by an S‐Adenosyl‐l‐ Methionine Radical Epimerase. Angewandte Chemie International Edition. 2019;58(8):2246-50. 42. Knappe J, Neugebauer FA, Blaschkowski HP, Gänzler M. Post-translational activation introduces a free radical into pyruvate formate-lyase. Proceedings of the National Academy of Sciences. 1984;81(5):1332-5. 43. Quitterer F, List A, Eisenreich W, Bacher A, Groll M. Crystal structure of methylornithine synthase (PylB): insights into the pyrrolysine biosynthesis. Angewandte Chemie International Edition. 2012;51(6):1339-42. 44. Knappe J, Schmitt T. A novel reaction of S-adenosyl-L-methionine correlated with the activation of pyruvate formate-lyase. Biochemical and biophysical research communications. 1976;71(4):1110-7. 45. Chakravarti A, Selvadurai K, Shahoei R, Lee H, Fatma S, Tajkhorshid E, Huang RH. Reconstitution and substrate specificity for isopentenyl pyrophosphate of the antiviral radical SAM enzyme viperin. Journal of Biological Chemistry. 2018;293(36):14122-33. 46. Melis A, Seibert M, Happe T. Genomics of green algal hydrogen research. Photosynthesis research. 2004;82(3):277-88. 47. Mehta AP, Abdelwahed SH, Begley TP. Molybdopterin biosynthesis— Mechanistic studies on a novel MoaA catalyzed insertion of a purine carbon into the ribose of GTP. Biochimica et Biophysica Acta (BBA)-Proteins and Proteomics. 2015;1854(9):1073-7. 48. Shisler KA, Broderick JB. Emerging themes in radical SAM chemistry. Current opinion in structural biology. 2012;22(6):701-10. 32 49. Wierenga R. The TIM‐barrel fold: a versatile framework for efficient enzymes. FEBS letters. 2001;492(3):193-8. 50. Vey JL, Yang J, Li M, Broderick WE, Broderick JB, Drennan CL. Structural basis for glycyl radical formation by pyruvate formate-lyase activating enzyme. Proceedings of the National Academy of Sciences. 2008;105(42):16137-41. 51. Lepore BW, Ruzicka FJ, Frey PA, Ringe D. The x-ray crystal structure of lysine- 2, 3-aminomutase from Clostridium subterminale. Proceedings of the National Academy of Sciences. 2005;102(39):13819-24. 52. Hänzelmann P, Schindelin H. Crystal structure of the S-adenosylmethionine- dependent enzyme MoaA and its implications for molybdenum cofactor deficiency in humans. Proceedings of the National Academy of Sciences. 2004;101(35):12870-5. 53. Vey JL, Drennan CL. Structural insights into radical generation by the radical SAM superfamily. Chemical Reviews. 2011;111(4):2487-506. 54. Nicolet Y, Drennan CL. AdoMet radical proteins—from structure to evolution— alignment of divergent protein sequences reveals strong secondary structure element conservation. Nucleic acids research. 2004;32(13):4015-25. 55. Lanz ND, Booker SJ. Identification and function of auxiliary iron–sulfur clusters in radical SAM enzymes. Biochimica et Biophysica Acta (BBA)-Proteins and Proteomics. 2012;1824(11):1196-212. 56. McGlynn SE, Boyd ES, Shepard EM, Lange RK, Gerlach R, Broderick JB, Peters JW. Identification and characterization of a novel member of the radical AdoMet enzyme superfamily and implications for the biosynthesis of the Hmd hydrogenase active site cofactor. Journal of bacteriology. 2010;192(2):595-8. 57. Chatterjee A, Li Y, Zhang Y, Grove TL, Lee M, Krebs C, Booker SJ, Begley TP, Ealick SE. Reconstitution of ThiC in thiamine pyrimidine biosynthesis expands the radical SAM superfamily. Nature chemical biology. 2008;4(12):758. 58. Külzer R, Pils T, Kappl R, Hüttermann J, Knappe J. Reconstitution and Characterization of the Polynuclear Iron-Sulfur Cluster in Pyruvate Formate-lyase- activating Enzyme MOLECULAR PROPERTIES OF THE HOLOENZYME FORM. Journal of Biological Chemistry. 1998;273(9):4897-903. 59. Walsby CJ, Ortillo D, Broderick WE, Broderick JB, Hoffman BM. An anchoring role for FeS clusters: Chelation of the amino acid moiety of S-adenosylmethionine to the unique iron site of the [4Fe− 4S] cluster of pyruvate formate-lyase activating enzyme. Journal of the American Chemical Society. 2002;124(38):11270-1. 33 60. Krebs C, Broderick WE, Henshaw TF, Broderick JB, Huynh BH. Coordination of adenosylmethionine to a unique iron site of the [4Fe-4S] of pyruvate formate-lyase activating enzyme: A Mössbauer spectroscopic study. Journal of the American Chemical Society. 2002;124(6):912-3. 61. Dey A, Peng Y, Broderick WE, Hedman B, Hodgson KO, Broderick JB, Solomon EI. S K-edge XAS and DFT calculations on SAM dependent pyruvate formate-lyase activating enzyme: nature of interaction between the Fe4S4 cluster and SAM and its role in reactivity. Journal of the American Chemical Society. 2011;133(46):18656-62. 62. Duschene KS, Veneziano SE, Silver SC, Broderick JB. Control of radical chemistry in the AdoMet radical enzymes. Current opinion in chemical biology. 2009;13(1):74-83. 63. Conradt H, Hohmann-Berger M, Hohmann H-P, Blaschkowski HP, Knappe J. Pyruvate formate-lyase (inactive form) and pyruvate formate-lyase activating enzyme of Escherichia coli: isolation and structural properties. Archives of biochemistry and biophysics. 1984;228(1):133-42. 64. Broderick JB, Duderstadt RE, Fernandez DC, Wojtuszewski K, Henshaw TF, Johnson MK. Pyruvate formate-lyase activating enzyme is an iron− sulfur protein. Journal of the American Chemical Society. 1997;119(31):7396-7. 65. Shisler KA, Broderick JB. Glycyl radical activating enzymes: structure, mechanism, and substrate interactions. Archives of biochemistry and biophysics. 2014;546:64-71. 66. Knappe J, Schacht J, Möckel W, Höpner T, Vetter Jr H, Edenharder R. Pyruvate formate‐lyase reaction in Escherichia coli: The enzymatic system converting and inactive form of the lyase into the catalytically active enzyme. European Journal of Biochemistry. 1969;11(2):316-27. 67. Knappe J, Blaschkowski H-P, Edenharder R. Enzyme-dependent activation of pyruvate formate-lyase of Escherichia coli. Metabolic Interconversion of Enzymes: Springer; 1972. p. 319-29. 68. Rödel W, Plaga W, Frank R, Knappe J. Primary structures of Escherichia coli pyruvate formate‐lyase and pyruvate‐formate‐lyase‐activating enzyme deduced from the DNA nucleotide sequences. European journal of biochemistry. 1988;177(1):153-8. 69. Krebs C, Henshaw TF, Cheek J, Huynh BH, Broderick JB. Conversion of 3Fe-4S to 4Fe-4S clusters in native pyruvate formate-lyase activating enzyme: Mössbauer characterization and implications for mechanism. Journal of the American Chemical Society. 2000;122(50):12497-506. 34 70. Walsby CJ, Ortillo D, Yang J, Nnyepi MR, Broderick WE, Hoffman BM, Broderick JB. Spectroscopic approaches to elucidating novel iron−sulfur chemistry in the “radical-SAM” protein superfamily. Inorganic Chemistry. 2005;44(4):727-41. 71. Gemperle C, Schweiger A. Pulsed electron-nuclear double resonance methodology. Chemical reviews. 1991;91(7):1481-505. 72. Henshaw TF, Cheek J, Broderick JB. The [4Fe-4S]1+ cluster of pyruvate formate- lyase activating enzyme generates the glycyl radical on pyruvate formate-lyase: EPR- detected single turnover. Journal of the American Chemical Society. 2000;122(34):8331- 2. 73. Wagner A, Frey M, Neugebauer FA, Schäfer W, Knappe J. The free radical in pyruvate formate-lyase is located on glycine-734. Proceedings of the National Academy of Sciences. 1992;89(3):996-1000. 74. Frey M, Rothe M, Wagner A, Knappe J. Adenosylmethionine-dependent synthesis of the glycyl radical in pyruvate formate-lyase by abstraction of the glycine C-2 pro-S hydrogen atom. Studies of [2H] glycine-substituted enzyme and peptides homologous to the glycine 734 site. Journal of Biological Chemistry. 1994;269(17):12432-7. 75. Frank R, Gausepohl H. Modern methods in protein chemistry. In: Tschesche H, editor. New York: Walter de Gruyter; 1988. p. 42-60. 76. Castro B, Dormoy J, Evin G, Selve C. Peptide coupling reagents. Part VII. Mechanism of the formation of active esters of hydroxybenzotriazole in the reaction of carboxylate ions on the BOP reagent for peptide coupling. A comparison with Itoh's reagent. J Chem Res(S). 1977;182. 77. Becker A, Fritz-Wolf K, Kabsch W, Knappe J, Schultz S, Wagner AV. Structure and mechanism of the glycyl radical enzyme pyruvate formate-lyase. Nature Structural & Molecular Biology. 1999;6(10):969. 78. Peng Y, Veneziano SE, Gillispie GD, Broderick JB. Pyruvate formate-lyase, evidence for an open conformation favored in the presence of its activating enzyme. Journal of Biological Chemistry. 2010;285(35):27224-31. 79. Himo F, Eriksson LA. Catalytic mechanism of pyruvate formate-lyase (PFL). A theoretical study. Journal of the American Chemical Society. 1998;120(44):11449-55. 80. Byer AS, McDaniel EC, Impano S, Broderick WE, Broderick JB. Mechanistic Studies of Radical SAM Enzymes: Pyruvate Formate-Lyase Activating Enzyme and Lysine 2, 3-Aminomutase Case Studies. Methods in enzymology: Elsevier; 2018. p. 269- 318. 35 81. Crain AV, Broderick JB. Flavodoxin cofactor binding induces structural changes that are required for protein–protein interactions with NADP+ oxidoreductase and pyruvate formate-lyase activating enzyme. Biochimica et Biophysica Acta (BBA)-Proteins and Proteomics. 2013;1834(12):2512-9. 82. Frey P. Lysine 2, 3-aminomutase: is adenosylmethionine a poor man's adenosylcobalamin? The FASEB journal. 1993;7(8):662-70. 83. Frey PA, Hegeman AD, Ruzicka FJ. The radical SAM superfamily. Critical reviews in biochemistry and molecular biology. 2008;43(1):63-88. 84. Horitani M, Shisler K, Broderick WE, Hutcheson RU, Duschene KS, Marts AR, Hoffman BM, Broderick JB. Radical SAM catalysis via an organometallic intermediate with an Fe–[5′-C]-deoxyadenosyl bond. Science. 2016;352(6287):822-5. 85. Broderick WE, Hoffman BM, Broderick JB. Mechanism of radical initiation in the radical S-adenosyl-l-methionine superfamily. Accounts of chemical research. 2018;51(11):2611-9. 86. Banerjee R. Radical carbon skeleton rearrangements: catalysis by coenzyme B12- dependent mutases. Chemical Reviews. 2003;103(6):2083-94. 87. Brown KL. Chemistry and enzymology of vitamin B12. Chemical reviews. 2005;105(6):2075-150. 88. Lowe DJ, Joblin KN, Cardin DJ. A precursor in the photolytic cleavage of the Co (III) C bond of coenzyme B-12 unobservable by electron paramagnetic resonance. Biochimica et Biophysica Acta (BBA)-General Subjects. 1978;539(3):398-401. 89. Yoder LM, Cole AG, Walker LA, Sension RJ. Time-resolved spectroscopic studies of B12 coenzymes: influence of solvent on the photolysis of adenosylcobalamin. The Journal of Physical Chemistry B. 2001;105(48):12180-8. 90. Hogenkamp H. A cyclic nucleoside derived from coenzyme B12. Journal of Biological Chemistry. 1963;238(1):477-80. 91. Matsuda A, Muneyama K, Nishida T, Sato T, Ueda T. A new synthesis of 5′-deoxy- 8, 5′-cyclo-adenosine and-inosine: conformationally-fixed purine nucleosides (nucleosides and nucleotides XVI). Nucleic acids research. 1976;3(12):3349-58. 92. Magnusson OT, Reed GH, Frey PA. Spectroscopic evidence for the participation of an allylic analogue of the 5‘-deoxyadenosyl radical in the reaction of lysine 2, 3- aminomutase. Journal of the American Chemical Society. 1999;121(41):9764-5. 36 93. Magnusson OT, Reed GH, Frey PA. Characterization of an allylic analogue of the 5‘-deoxyadenosyl radical: An intermediate in the reaction of lysine 2, 3-aminomutase. Biochemistry. 2001;40(26):7773-82. 37 CHAPTER 2 GENERAL METHODOLOGY Expression and Purification of PFL-AE The Escherichia coli (E. coli) PFL-AE gene, pflA, was inserted into the pCAL-n- EK plasmid and transformed into BL21(DE3)pLysS cells. Two colonies from the plated transformation were used to inoculate two overnight cultures containing 50 mL lysogeny broth (LB) media and 50 µg/mL ampicillin (AMP). The growths were incubated at 37 ℃ with 200 rpm agitation in a shaker incubator (New Brunswick Scientific I 26). From the O/N cultures, 10 mL was used to inoculate each of six growths containing 1.5 L of warmed LB media + 50 µg/mL of AMP in 2.8 L baffled Fernbach flasks. The culture was grown at 37 ℃ and 200 rpm agitation until the cells reached an optical density at 600 nm (O.D.600) of ~0.3. At that time, 5 g/L of glucose was added to each flask. When the cells reached an Figure 2.1. SDS PAGE gels of the growth of PFL-AE and PFL. A) samples of PFL-AE before and after growth was induced with IPTG. B) three gels showing the molecular weight ladder, PFL sample before induction, and PFL sample after induction. All pre and post-inductions samples were resuspended in 50 and 100 µL SDS-PAGE sample buffer respectively. 38 O.D.600 of 0.6-0.8, they were induced with 0.25 mM isopropyl β-D-1- thiogalactopyranoside (IPTG) and a pre-induction sample was taken by spinning down 1 mL of culture, removing the supernantant, and storing at -80 ℃ until needed. Ferrous ammonium sulfate (FAS) and L-cysteine were added to a final concentration of 0.2 mM at the time of induction and the cultures continued to grow at a reduced temperature of 30 ℃. To maintain a pH between 7.0 and 7.2, the optimal range for E. coli growth, 6.0 M NH4OH was added as needed (1). After two hours, a secondary addition of FAS and L-cysteine was Figure 2.2. Representative spectra for PFL-AE purification. A) Clarified lysate is first run over 1 L Superdex-75 prep grade resin with 50 mM Tris, pH 7.5, 100 mM KCl, 2.0 mM DTT buffer. Blue box indicates collected fractions. Inset shows color of collected fractions. B) Repeated purification with collected fractions from initial purification of PFL-AE. Blue box denotes A fractions and green boxes denote B fractions. Inset indicates color of combined A fractions. 39 added to a final concentration of 0.4 mM. Once the pH no longer showed dramatic change over a 30 min period, usually 4-5 hr after induction, a few drops of antifoam were added, and the cells were left to sparge at 4 ℃ with N2 gas overnight (16-18 hr). The following morning a post-induction sample was taken, prepared as described for the pre-induction sample, to determine if the protein had expressed (Figure 2.1a). The cells were harvested by centrifugation at 6,000 rpm and 4 ℃ for 10 min, flash frozen in liquid nitrogen, and stored at -80 ℃ until purification. The purification of PFL-AE was conducted in an anaerobic vinyl chamber (Coy Laboratory Products, Inc.) at room temperature. The cell pellet was added to prepared PFL- AE lysis buffer (50 mM Tris, pH 7.5, 100 mM KCl, 10 mM MgCl2, 10% w/v glycerol with the addition of 2 mM DTT, 1 mM phenylmethylsulfonyl fluoride (PMSF), 0.4 mg/mL lysozyme, 1% w/v triton x-100 and trace DNAse RNAse) in a ratio of 1 g pellet to 1-2 mL lysis buffer. The pellet was thawed in the prepared lysis buffer, broken down using a flat sided spatula, then homogenized by pulling the lysate up into a 30 mL syringe affixed with an 18 gauge needle and gently adding the lysate back into a beaker until no longer viscus. The E. coli lysate was spun down at 18,000 rpm for 1 hr at 4 ℃ and the supernatant or clarified lysate was collected. If the supernatant appeared cloudy, a 0.45 µm syringe filter was used to gently filter the solution. The clarified lysate was decanted into a 50 mL superloop to be loaded onto the purification column. Using a liquid chromatography (FPLC) system, the clarified lysate was loaded onto a Waters AP-5 column, Superdex-75 prep grade resin, equilibrated with gel filtration buffer (50 mM Tris, pH 7.5, 100 mM KCl, 2.0 mM DTT) at 3 mL/min. The protein eluted from 40 the column after running 550-600 mL of gel filtration buffer as shown in figure 2.2a. Due to the dark color of the protein, the progression of PFL-AE elution could be monitored by observing the dark brown band moving down the column as well as by an FPLC, UV-vis detector at 280 nm. Fractions that showed significant color were collected and concentrated with 10K MWCO Millipore Amicon Ultra centrifugal concentrators (Figure 2.2a insert), flash frozen in liquid nitrogen, and stored at -80 ℃ until needed. To further purify the protein, the above purification procedure was followed with minor changes. Factions from previous purification were thawed in an anaerobic chamber and loaded onto the Superdex-75 column as before. Smaller fractions of protein, 5 mL, were collected and their absorbances at 426 nm and 280 nm were measured on a UV-vis spectrometer. Samples with the highest absorbance 426/280 nm ratio, >0.170, were pooled as A fractions and samples with the second highest ratios, 0.16-0.17, comprised the B fractions (Figure 2.2b). A and B fractions were concentrated with 10K MWCO centrifugal concentrators and flash frozen in liquid nitrogen. Purified PFL-AE was stored at -80 ℃ until needed. PFL-AE B fractions from multiple purifications could be combined and run over the Superdex-75 column again to obtain protein with a higher Fe/protein ratio. Protein concentration and iron quantity were determined using the Bradford Assay and flame atomic absorption (AA) spectroscopy. Protein and Iron Quantification The quantification of purified protein was accomplished using the Bradford Assay (2). Bovine albumin was used to generate standards with concentrations 0-6 µg/mL. Bio- 41 Rad protein assay dye reagent concentrate was applied to dye the standards and protein samples. Duplicates of three different dilutions were generated for the sample protein and their absorbances were measured at 595 nm. In the case of PFL-AE, a correction factor of 0.67 was multiplied by the average mg/mL of protein yielding a more accurate PFL-AE concentration (3). To calculate the average number of iron atoms present per protein, AA spectroscopy was used (4). A standard curve was generated with 0.0-2.0 ppm dilutions from a 1000 ppm iron standard from Ricca Chemical Company. Two replicates of a single protein sample dilution were used to determine the concentration of iron in said sample. The iron concentration was then divided by the protein concentration generating the iron number. Expression and Purification of PFL PFL is a non-iron containing protein and can therefore be grown and purified aerobically. The PFL E. coli gene, pflB, was coded in the pKK223-3 plasmid and expresses in BL21(DE3)pLysS cells. Two O/N cultures were prepared as described for PFL-AE. Three 1.5 L growths of LB media + 50 µg/mL AMP in 2.8 L Fernbach flasks were each inoculated with 15 mL of O/N. The cells were grown at 37 ℃ and 200 rpm agitation until they reached an O.D.600 between 0.8-1.0. The growth was then induced with 0.2 mM IPTG and a pre-induction sample was taken. The cultures continued to grow overnight (16-18 hr) at 30 ℃ with 200 rpm agitation. The following morning a post-induction samples was taken (Figure 2.1b) and antifoam was added to each flask. The cell pellet was harvested by 42 centrifugation at 6,000 rpm and 4 ℃ for 10 min then flash frozen in liquid nitrogen and store at -80 ℃ until purification. PFL containing cells were added to prepared PFL lysis buffer (20 mM Tris, pH 7.2, 10 mM MgCl2, 5% w/v glycerol, 1% w/v triton X-100 with the addition of 1mM PMSF, 0.16 mg/mL lysozyme, and trace amounts of DNAse and RNAse) in a 1 g cell pellet to 1- 2 mL lysis buffer ratio. The cells were lysed and spun down following the same procedure as PFL-AE. The remaining procedure was conducted at 4 ℃. The clarified lysate was added to a 50 mL super loop and loaded at 3 mL/min onto a Waters AP-5 300 mm column containing Acell Plus QMA resin equilibrated with no salt buffer A (20 mM Tris, pH 7.2). Using a previously written FPLC purification program, the protein was washed with 300 mL of no salt buffer A before beginning a 900 mL linear gradient from 0% to 100% high salt buffer (20 mM Tris, pH 7.2, 500 mM NaCl). Lastly, 300 mL of high salt buffer were run over the column. Protein began to elute at 50% high salt buffer (750 mL) as shown in figure 2.3a. Eluted protein was collected in 10 mL fractions. An impurity around 40 kDa elutes with PFL. To determine which fractions contained the purest PFL, aliquots of the 43 fractions were run on a 10% acrylamide SDS PAGE gel (Figure 2.3b).The cleanest PFL fractions were concentrated using 30K MWCO Millipore Amicon Ultra centrifugal concentrators. Protein was the flash frozen in liquid nitrogen and stored at -80 ℃ until needed. Figure 2.3. Representation of a PFL purification. Collected fractions are denoted by the black box. A) Chromatograph of PFL purified on a 600 mL QMA anion exchange column. Buffer A is 20 mM Tris, pH 7.2 and buffer B is 20 mM Tris, pH 7.2, 500 mM NaCl. B) A representative SDS PAGE gel of a PFL purification collected from the QMA anion exchange column. Molecular weight later is noted by an L. C) Chromatagram of PFL purified over a 10 mL hydrophobic phenyl sepharose column. Buffer A is 20 mM Tris, pH 7.2, 1.0 M NH4OH and buffer B is 20 mM Tris, pH 7.2. 44 In order to have a higher grade of pure protein as required for generating Ω samples, PFL could be further purified. Fractions collected from the QMA column purification were concentrated to 20-30 mL. Using the same 30K MWCO Millipore Amicon Ultra centrifugal concentrators, PFL was buffer exchanged into Buffer B (20 mM Tris, pH 7.2, 1.0 M (NH4)2SO4) by concentrating the protein down, adding 5x Buffer B, then re- concentrating the protein. This process was repeated three times. To remove any precipitated protein, PFL in Buffer B was filtered using a 45 μm syringe filter. The protein was then added to a 50 mL superloop and loaded at 1 mL/min onto a HighLoad High Performance 16/10 phenyl sepharose column equilibrated with Buffer B. Using a previously written FPLC program, the protein was washed with 50 mL Buffer B. Then, a 50 mL linear gradient from 100% Buffer B to 100% no salt buffer was run to elute the protein. Finally, 50 mL of no salt buffer A was run over the column. Protein began to elute around 75 mL, 50% no salt buffer A, and 2 mL fractions were collected (Figure 2.3c). Aliquots of each fraction were run on a 10% acrylamide SDS PAGE gel to Figure 2.4. SDS PAGE gels of purified PFL. A) PFL protein purified on the QMA column and PS column (left). On the right, PFL purified on the QMA column only. B) A diluted sample of purified PFL purified on QMA and PS columns. PFL protein band shown in black box. 45 determine which fractions contained the purest protein. The purest fractions were concentrated as above, flash frozen in liquid nitrogen, and stored at -80 ℃ until needed. Protein was quantified using Bradford method previously described. Expression and Purification of OspD OspD DNA plasmid, pCDF-6xH-OspD, gifted by the Jörn Piel lab, was transformed into BL21(DE3)ΔiscR cells. O/N cultures were made by adding a colony from the transformation to each of 2 x 50 mL LB media with 50 µg/mL spectinomycin (SPEC) and 30 µg/mL kanamycin (KAN). The O/N cultures were grown for ~18 hr at 37 ℃ with 200 rpm agitation. In the meantime, 6 x 1 L of terrific broth (TB) media with phosphate buffer (12 g/L tryptone, 24 g/L yeast extract, 55.0 mM glycerol, 72.0 mM K2HPO4, and 17.0 mM KH2PO4) in 2.8 L Fernbach flasks was prepared. Growth was inoculated with 10 A B Figure 2.5. SDS PAGE gels showing the expression and purification of OspD. A) a gel showing the expression of OspD before and after the addition of IPTG. B) the purification of OspD on a 25 mL Ni-NTA column. The percentages shown are the imidazole concentration in HEPES buffer. In both gels, the OspD protein is highlighted with a black box. 46 mL of O/N culture and grown at 37 ℃ with 200 rpm agitation until an O.D.600 of about 1.0. Flasks were removed from the incubator and set in ice baths until they reached 16 ℃. The growth was induced with 1 mM IPTG and 0.3 mM FAS was added. A pre-induction sample was taken and the cells grew at 16 ℃ with 200 rpm agitation for 12 hours. An additional aliquot of FAS was added five hours after induction, total FAS concentration 0.6 mM. After 12 hours, a post-induction sample was taken and the cells were harvested by centrifugation, flash frozen in liquid nitrogen, and stored at -80 ℃ until purification. An SDS PAGE gel was run to determine if the protein expressed (figure 2.5a). The purification for OspD was performed anaerobically in a Coy chamber at room temperature. The lysis buffer was prepared by adding 2 mM DTT, 0.6 mg/mL PMSF, 0.5 mg/mL lysozyme, trace RNAse and DNAse, 1% w/v triton X-100, and 15 mM MgCl2 to HEPES Buffer (50 mM HEPES, pH 8.0, 150 mM KCl, 10% w/v glycerol). Frozen cell pellet was added to the lysis buffer, 1 g cell pellet to 1.5-2 mL of lysis buffer. The cells were lysed as previously described for PFL-AE and PFL. The lysis was spun down at 17,000 rpm, 4 ℃, for 45 min. The clarified lysate was filtered through a 0.45 μm syringe filter then through a 0.2 μm syringe filter before being loaded onto a gravity flow, 25 mL Ni NTA column equilibrated with Buffer A (50 mM HEPES, pH 8.0, 150 mM KCl, 10% w/v glycerol, 1 mM imidazole). The column was washed with 25 CVs of Buffer A to which 2 mM DTT had been added. The protein was eluted with 2 CV washes containing 5%, 10%, 25%, 50%, 75%, and 100% Buffer B (50 mM HEPES, pH 8.0, 150 mM KCl, 10% w/v glycerol, 100 mM imidazole, 2 mM DTT). Aliquots of each wash were run on a 12% acrylamide SDS PAGE gel to determine which elution contained the most protein (figure 47 2.5b). Protein from 25%, 50%, and 75% washes were combined and concentrated before being run over a 75 mL Dextran Desalting column, Thermo Fisher 5K MWCO, equilibrated with 2 CVs of HEPES buffer. The desalted protein was collected and concentrated before being flash frozen in liquid nitrogen and stored at -80 ℃ until reconstitution. Protein concentration and iron number were determined using a Bradford assay and AA spectroscopy respectively. OspD rarely purified with fully loaded protein, so an inorganic reconstitution was often necessary. Reconstitution of OspD The reconstitution of OspD was conducted anaerobically in a Coy chamber at both room temperature and 4 ℃. OspD, final concentration 100 µM and 2 mM DTT was added to HEPES Buffer (50 mM HEPES, pH 8.0, 150 mM KCl, 10% w/v glycerol, and 2 mM DTT) and the solution was stirred at low speed for 10 min. A 6-fold excess of Fe3+ and S2+ were added to the solution over a 2-hour period by slowly adding a 10 mM solution of FeCl3 then a 50 mM solution of Na2S. After the final concentrations of FeCl3 and Na2S were 600 µM, the solution stirred for 2 hrs and went from a light brown color to dark, coffee brown. The reconstituted solution was aliquoted into 1.5 mL Eppendorf tubes and spun for 5 min at 13,000 rpm to pellet any precipitated protein. The supernatant was run over a 75 mL Dextran Desalting column, Thermo Fisher 5K MWCO, equilibrated with 2 CVs of HEPES Buffer. The protein was collected and concentrated before being flash frozen in liquid nitrogen and stored at -80 ℃. The protein concentration and iron number were calculated as before using the Bradford assay and AA spectroscopy. 48 Growth and Preparation of SAM Synthetase The SAM synthetase strain and growth procedure were donating by the Markham lab (5-7). The SAM synthetase gene on a pBR322 plasmid derivative (pk8) that contained the metK (SAM synthesis) gene was transformed into DM22 E. coli cells. A colony from the transformation was used to inoculate each of two 50 mL O/N cultures containing LB media and 10 µg/mL oxytetracycline (oxytet). O/N cultures were grown overnight (16-18 hr) at 37 ℃ with 200 rpm agitation. The following morning, five 2.8 L Fernbach flasks with 700 mL LB + 10 µg/mL oxytet were inoculated with 10 mL of O/N culture. The cells were grown at 37 ℃, 200 rpm agitation for 12-14 hrs. The cells were then harvested by centrifugation, 6,000 rpm, 4℃, 10 min, and flash frozen in liquid nitrogen. Cells were kept at -80 ℃ until lysed. SAM synthetase remains active in the cell lysate and does not require purification to effectively synthesize SAM. Cells were thawed and lysed in 100 mM Tris-HCl pH 8.0 buffer, (1-2 mL buffer per gram of cells). The lysis was homogenized in the same manner as PFL-AE, with an 8 G needle and 30 mL syringe. Lysis was then run through a microfluidizer (Unitronics LM10) at 15,000 psi. Small aliquots, 1 mL, were flash frozen and stored at -80 ℃ until needed for SAM synthesis. SAM Synthesis and Purification SAM was synthesized by combining 100 mM Tris, pH 8.0 buffer with 50.0 mM KCl, 55.7 mM MgCl2•6H2O, 900 μM ethylenediaminetetraacetic acid (EDTA), 14.4 mM 49 adenosine triphosphate (ATP), 12.1 mM L-methionine, 1.14 M ꞵ-mercaptoethanol, 2-5 U inorganic phosphatase, and 1 mL SAM crude lysate in a 20 mL scintillation vial. The reaction was covered in aluminum foil to decrease SAM degradation and left for 16-18 hr with low stirring at room temperature. The progression of the reaction was monitored by a TLC plate (mobile phase butanol, acetic acid, water, and 2 M formic acid in a 8:2:2:1 ratio). The reaction was quenched with 1 mL of 1.0 M HCl to enable SAM binding to the Source 15S cation exchange column. The reaction was spun for 30 min, 18,000 rpm, 4 ℃ and the supernatant collected. SAM purification was conducted aerobically at 4 ℃. A 20 mL, Source15S cation exchange column was prepared by running 40 mL of water (Buffer A), 20 mL of 1.0 M HCl (Buffer B), then 40 mL of Buffer A at 4 mL/min or slower. Half of the SAM reaction (~5 mL) was loaded onto the column through a 50 mL superloop at 1 mL/min. The FPLC program detailed in figure 2.6a was run at 2 mL/min. SAM began to elute around 60% Buffer B. The SAM peak was collected in a 250 mL round bottom (RB) flask and rotovapped to ~5 mL. SAM was stored at 4 ℃ while the second half of the solution was Figure 2.6. Method for the purification of SAM. A) Program details for the purifications of SAM on a 10 mL Source 15S cation exchange column. Buffer A is ultrapure H2O and buffer B is 1.0 M HCl. B) Representative chromatogram of a SAM purification. The black box denotes collected SAM peak. 50 purified (Figure 2.4b). The SAM elutions were combined and rotovapped down to a volume less than 10 mL. To further remove buffer from the SAM, the solution was transferred to a 50 mL RB and rotovapped down to an oily film before being stored at -80 ºC until resuspension. The resuspension of SAM was conducted in an anaerobic chamber. SAM was brought up in roughly 300 µL of 100 mM Tris, pH 8.0 buffer. The desired pH, 7.0-7.5, was then obtained by slowly adding small volumes of 1M NaOH and 1 M HCl. SAM was aliquoted into screw cap vials, flash frozen in liquid nitrogen, and stored at -80 ℃ until needed. The concentration was determined by measuring the absorbance of multiple dilutions of SAM at 260 nm and using the molar extinction coefficient of 16,000 M-1cm-1 (8, 9). AnSAM Synthesis and Purification AnSAM was synthesized in the same manner as SAM with a few modifications. In a scintillation vial, 50 mM HEPPS, pH 8.0 buffer, 20 mM MgCl2•6H2O, 1.9 mM anhydroadenosine triphosphate (anATP), 3.4 mM L-methionine, 0.25 U inorganic pyrophosphate, and 1 mL of SAM synthetase crude lysate were combined. A scintillation vial was covered with aluminum foil and let stir at low speed and room temperature for 3- 6 hr. The pH was assessed and adjusted to 8.0 by adding small amounts of 1 M HCl every hour. The reaction progression was monitored using a TLC plate (mobile phase butanol, acetic acid, water, and 2 M formic acid in a 8:2:2:1 ratio). The anSAM reaction was quench with 1 mL of 1 M HCl and spun down for 30 min at 4 ℃ and 16,000 rpm. AnSAM was 51 purified and re-hydrated in the same manner as SAM. The concentration of anSAM was determined using UV-vis spectroscopy and the molar extinction coefficient of 15,400 M-1 cm-1 (5). Electron Paramagnetic Resonance Spectroscopy (EPR) EPR is a powerful technique used to observe [Fe-S] clusters (10). Their characteristic electronic structure that arises from the interactions of individual, high spin iron centers are easily identifiable. The foundation of EPR is the energy differences associated with the Zeeman effect (11). When an unpaired electron is subjected to a magnetic field, its magnetic moment (Ms) can either align with or against the magnetic field thus creating the parallel, Ms = -1/2, or antiparallel, Ms = +1/2, spin state (12). The difference in energy of these spin states is denoted as ΔE. The relationship between ΔE and the magnetic field gives the basic equation of EPR (Equation 1) ΔE = hυ = gꞵB (1) where g is the g-factor, ꞵ is the Bohr magneton (9.274x10-24 J·T-1) , and B is the magnetic field. The g-factor for an unpaired electron in an isotropic environment is 2.002 which serves as a standard. Deviation from the ideal value provides the basis for identifying the nature and chemical environment of other unpaired electrons (13). 52 To induce a change in energy between the two electron spin states, either the magnetic field can be held constant and the microwave frequency scanned or vise-versa, which is more practical (11, 14). The most common EPR instruments operate at either X- band or Q-band microwave frequencies that are held constant at roughly 9.75 or 34.0 GHz, respectively. Other less commonly used frequencies are L, S, and W-bands at 1.1 GHz, 3.0 GHz, and 94.0 GHz respectively (15). Once the magnetic field induced energy difference matches the energy of the fixed microwave radiation, a resonance conditions is achieved and absorption of the electromagnetic radiation occurs as shown in Figure 2.7. Given the Figure 2.7. Figure adapted from (13). An unpaired electron subject to a magnetic moment can align either spin up, M = +1/2, or spin down, M = -1/2. The resulting absorption is measured by the EPR instrument and reported as the first derivative. 53 low intensity of these events, the first derivative of the absorption is plotted, resulting in the EPR spectrum (11, 13). An unpaired electron is susceptible to the magnetic field generated by other surrounding electrons and nuclei. If the nucleus has a non-zero magnetic moment, therefore being NMR active, it will cause the Zeeman levels to split. The nuclear magnetic moment (I) caused splitting of the EPR spectral features are also referred to as hyperfine couplings, which are essential in identifying the atomic environments of the unpaired electron (figure 2.8) (13, 16). The number of new EPR features can be calculated following Equation 2, where N is the number of identical nuclei and I is the nuclear spin (13). Figure 2.8. Figure adapted from (16). The hyperfine coupling resulting from three identical nuclei (N=3) on a carbon (I(13C)=1/2) centered radical. 54 2NI+1 (2) In biology, most unpaired electrons exist in anisotropic environments where the coupling (J) between electron spin (S) and spin orbital (L) causes a change in the spectrum shape (J = L ± S) (11). Given that the spin quantum state has Sx, Sy, and Sz components, each paramagnetic center has its own principal axis with unique gx, gy, and gz-values. The exact assignment of each component is only possible for crystalline samples; however, their statistical average in a randomly distributed sample still gives rise to three unique g- values, often labeled as g1, g2, and g3. The variability of these g-values determines which of the three characteristic shapes the spectra will take: isotropic (g1 = g2 = g3), axial (g1 = Figure 2.9. Figure adapted from (16). The nuclear environment surrounding an unpaired electron will cause the absorption to be either isotropic, axial, or rhombic. The characteristic 1st derivatives are shown with g- values by their corresponding peaks. 55 g2 ≠ g3), or rhombic (g1 ≠ g2 ≠ g3). For an axial signal, the g-values are often denoted g- parallel (g||) and g-perpendicular (g⊥) with g⊥ resulting from the two-equivalent g-values (Figure 2.9) (13, 17). Preparation of RS Enzyme Intermediate Samples All RS enzyme intermediate samples were kept as anaerobic as possible by conducting much of their preparation in anaerobic chambers. Samples were prepared at room temperature and stored in liquid nitrogen cooled dewars until they could be observed spectroscopically. 56 RFQ Samples Solutions of PFL-AE and (PFL + SAM) were combined as shown in table 2.1. Photoreductions were conducted following previously described methods (18). PFL was “photoreduced” for 30 min to scrub any oxygen from the sample and SAM was added afterward to prevent degradation of the small molecule while being exposed to intense light. Table 2.1. Recipes for RFQ samples with two different reductants. A and B show the concentrations of enzyme and substrate solutions reduced with 5-deazariboflavin. C and D show concentrations of enzyme and substrate solutions reduced with DT. 57 In an anaerobic chamber, reduced PFL-AE and (PFL/SAM) samples were loaded into individual syringes equipped with substrate loops. To ensure complete sample transfer into the RFQ mixing chamber, 100 mM DT solution was placed on either side of the protein sample. Pockets of N2 atmosphere from the anaerobic chamber were established between the protein and the DT solution to prevent sample contamination. Additionally, the DT solution helped protect the protein samples from atmospheric oxygen when they were transferred from the anaerobic chamber to the RFQ instrument (18). As described by Horitani et al., all RFQ experiments were conducted with a System 100 apparatus (Update Instrument). Reduced PFL-AE was rapidly mixed with (PFL + SAM) and quenched at either 500 ms or 100 ms. To improve the efficiency of the rapid quench, the mixed solution was sprayed onto two rotating wheels made from either copper or aluminum. The frozen powder was packed into a Q-band tube and stored at 77 K until run (18). Ω samples were observed using X-band continuous wave (CW) EPR spectroscopy on a Bruker ESP 300 spectrometer with an Oxford Instruments ESR 910 continuous helium flow cryostat. The spectroscopic parameters were 40 K, 9.37 GHz, 1.99 mW microwave power, 100 kHz modulation, and 8 G modulation amplitude (19). Hand Quench (HQ) Photolysis Samples HQ samples were reduced by photoreduction or DT. Photoreduction samples containing 550 µM PFL-AE, 1.0 mM DTT, 200 μM 5-deazariboflavin dissolved in DMSO in a 50 mM Tris, pH 7.5, 100 mM KCl buffer were reduced as described previously (18). 58 DT reduction samples were comprised of 550 µM PFL-AE and 3 mM DT in the same Tris buffer. PFL-AE and DT would be combined and left for at least 3 min to ensure maximum cluster reduction. Regardless of the reducing agent, 5.5 mM SAM was added after reduction and before the samples were transferred to Q-band EPR tubes. To reduce oxygen exposure, the samples were frozen in liquid nitrogen within the anaerobic chamber before being brought out and stored at 77 K. HQ samples were photolyzed in situ with a 450 nm Thorlabs diode laser. The samples were kept at 12 K and photolyzed for 1 hr before being observed by X and Q-band CW EPR spectroscopy. X-band EPR spectroscopy was conducted on the same instrument described for observing Ω samples. Q-band EPR spectroscopy was preformed using a Bruker EMX spectrometer equipped with an Oxford Instruments Mercury iTC continuous helium flow cryostat. Typical parameters for observing the radical species were 12 or 40 K, 9.38 GHz for X-band and 34.0 GHZ for Q-band spectroscopy, 1 mW microwave power, 100 kHz modulation, and 10 G modulation amplitude (20). 59 References 1. Davey K. Modelling the combined effect of temperature and pH on the rate coefficient for bacterial growth. International journal of food microbiology. 1994;23(3- 4):295-303. 2. Bradford MM. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Analytical biochemistry. 1976;72(1-2):248-54. 3. Broderick JB, Duderstadt RE, Fernandez DC, Wojtuszewski K, Henshaw TF, Johnson MK. Pyruvate formate-lyase activating enzyme is an iron− sulfur protein. Journal of the American Chemical Society. 1997;119(31):7396-7. 4. Fish W. Rapid colorimetric micromethod for the quantitation of complexed iron in biological samples in Methods in Enzymology (Riordan, JF and Vallee, BL, eds.) Vol. 158A. Academic Press, San Diego; 1988. 5. Byer AS, McDaniel EC, Impano S, Broderick WE, Broderick JB. Mechanistic Studies of Radical SAM Enzymes: Pyruvate Formate-Lyase Activating Enzyme and Lysine 2, 3-Aminomutase Case Studies. Methods in enzymology: Elsevier; 2018. p. 269- 318. 6. Markham GD, Hafner E, Tabor CW, Tabor H. S-Adenosylmethionine synthetase from Escherichia coli. Journal of Biological Chemistry. 1980;255(19):9082-92. 7. Markham GD, DeParasis J, Gatmaitan J. The sequence of metK, the structural gene for S-adenosylmethionine synthetase in Escherichia coli. Journal of Biological Chemistry. 1984;259(23):14505-7. 8. Gross A, Geresh S, Whitesides GM. Enzymatic synthesis ofS-adenosyl-l- methionine from l-methionine and ATP. Applied biochemistry and biotechnology. 1983;8(5):415-22. 9. Shapiro SK, Ehninger DJ. Methods for the analysis and preparation of adenosylmethionine and adenosylhomocysteine. Analytical biochemistry. 1966;15(2):323- 33. 10. Hagen WR. EPR spectroscopy of iron—sulfur proteins. Advances in inorganic chemistry: Elsevier; 1992. p. 165-222. 11. Weil JA, Bolton JR. Electron paramagnetic resonance: elementary theory and practical applications: John Wiley & Sons; 2007. 60 12. Abragam A, Bleaney B. Electron paramagnetic resonance of transition ions: OUP Oxford; 2012. 13. Petasis DT, Hendrich MP. Quantitative interpretation of multifrequency multimode EPR spectra of metal containing proteins, enzymes, and biomimetic complexes. Methods in enzymology: Elsevier; 2015. p. 171-208. 14. Weber RT, Jiang J, Barr DP. EMX user’s manual. Bruker Instruments, Billerica. 1998. 15. Schauer DA, Iwasaki A, Romanyukha AA, Swartz HM, Onori S. Electron paramagnetic resonance (EPR) in medical dosimetry. Radiation measurements. 2006;41:S117-S23. 16. Baldansuren A. Small Ag clusters supported on an LTA zeolite investigated by CW and pulse EPR spectroscopy, XAS and SQUID magnetometry. OPUS: University of Stuttgart; 2009. 17. Que Jr L. Physical methods in bioinorganic chemistry. University Science, Sausilito. 2000. 18. Horitani M, Shisler K, Broderick WE, Hutcheson RU, Duschene KS, Marts AR, Hoffman BM, Broderick JB. Radical SAM catalysis via an organometallic intermediate with an Fe–[5′-C]-deoxyadenosyl bond. Science. 2016;352(6287):822-5. 19. Byer AS, Yang H, McDaniel EC, Kathiresan V, Impano S, Pagnier A, Watts H, Denler C, Vagstad AL, Piel Jr. Paradigm shift for radical S-adenosyl-L-methionine reactions: The organometallic intermediate Ω is central to catalysis. Journal of the American Chemical Society. 2018;140(28):8634-8. 20. Yang H, McDaniel EC, Impano S, Byer AS, Jodts RJ, Yokoyama K, Broderick WE, Broderick JB, Hoffman BM. The elusive 5′-deoxyadenosyl radical: captured and characterized by electron paramagnetic resonance and electron nuclear double resonance spectroscopies. Journal of the American Chemical Society. 2019;141(30):12139-46. 61 CHAPTER 3 PARADIGM SHIFT FOR RADICAL S-ADENOSYL-L-METHIONINE REACTIONS: THE ORGANOMETALLIC INTERMEDIATE Ω IS CENTRAL TO CATALYSIS Contribution of Authors and Co-Authors Manuscript in Chapter 3 Author: Amanda S. Byer Contributions: Synthesized labeled SAMs, grew and purified LAM and HydG, assisted in preparation of RFQ samples at Northwestern University, prepared manuscript and generated figures. Author: Hao Yang Contributions: Collected and analyzed EPR and ENDOR results at Northwestern University, conducted RFQ experiments, prepared manuscript and generated figures. Co-Author: Elizabeth C. McDaniel Contributions: Synthesized unlabeled SAM, grew and purified PFL-AE, PFL, and assisted with growth and purification of OspD, assisted in preparation of RFQ samples at Northwestern University. Co-Author: Venkatesan Kathiresan Contributions: Assisted in operating the RFQ instrument. Co-Author: Stella Impano Contributions: Grew and purified HydG and assisted with growth and purification of OspD. Co-Author: Adrien Pagnier Contributions: Grew and purified SPL. 62 Contribution of Authors and Co-Authors Continued Co-Author: Hope Watts Contributions: Grew and purified PoyD. Co-Author: Carly Denler Contributions: Assisted in growth and purification of LAM. Co-Author: Anna L. Vagstad Contributions: Grew and purified OspA and PoyA, provided DNA for OspD and PoyD. Co-Author: Jörn Piel Contributions: Provided insight into OspD and PoyD generation and function. Co-Author: Kaitlin S. Duschene Contributions: Synthesized labeled SAMs, growth and purification of RNR-AE and RNR. Co-Author: Eric M. Shepard Contributions: Provided insight into Ω structure and formation. Co-Author: Thomas P. Shields Contributions: Provided [5′-13C]-SAM and [5′-D2-ado]-SAM. Co-Author: Lincoln G. Scott Contributions: Provided [5′-13C]-SAM and [5′-D2-ado]-SAM. Co-Author: Edward A. Lilla Contributions: Provided [5′-13C]-SAM and [5′-D2-ado]-SAM. 63 Contribution of Authors and Co-Authors Continued Co-Author: Kenichi Yokoyama Contributions: Provided [5′-13C]-SAM and [5′-D2-ado]-SAM. Co-Author: William E. Broderick Contribution: Assisted with manuscript preparation and interpretation of EPR and ENDOR data. Co-Author: Brian M. Hoffman Contribution: Assisted with manuscript preparation and interpretation of EPR and ENDOR data. Co-Author: Joan B. Broderick Contribution: Assisted with manuscript preparation and interpretation of EPR and ENDOR data. 64 Manuscript Information Amanda S. Byer, Hao Yang, Elizabeth C. McDaniel, Venkatesan Kathiresan, Stella Impano, Adrien Pagnier, Hope Watts, Carly Denler, Anna L. Vagstad, Jörn Piel, Kaitlin S. Duschene, Eric M. Shepard, Thomas P Shields, Lincoln G. Scott, Edward A. Lilla, Kenichi Yokoyama, William E. Broderick, Brian M. Hoffman, and Joan B. Broderick. Journal of American Chemical Society Status of Manuscript ____Prepared for submission to a peer-reviewed journal ____ Officially submitted to a peer-review journal ____ Accepted by a peer-reviewed journal X Published in a peer-reviewed journal Publisher ACS Publications Issue #140 DOI 10.1021/jacs8b04061 65 PARADIGM SHIFT FOR RADICAL S-ADENOSYL-L-METHIONINE REACTIONS: THE ORGANOMETALLIC INTERMEDIATE Ω IS CENTRAL TO CATALYSIS Abstract Radical S-adenosyl-ʟ-methionine (SAM) enzymes comprise a vast superfamily catalyzing diverse reactions essential to all life through homolytic SAM cleavage to liberate the highly reactive 5′-deoxyadenosyl radical (5′-dAdo•). Our recent observation of a catalytically competent organometallic intermediate Ω that forms during reaction of the radical SAM (RS) enzyme pyruvate formate-lyase activating-enzyme (PFL-AE) was therefore quite surprising, and led to the question of its broad relevance in the superfamily. We now show that Ω in PFL-AE forms as an intermediate under a variety of mixing order conditions, suggesting it is central to catalysis in this enzyme. We further demonstrate that Ω forms in a suite of RS enzymes chosen to span the totality of superfamily reaction types, implicating Ω as essential in catalysis across the RS superfamily. Finally, EPR and electron nuclear double resonance spectroscopy establish that Ω involves an Fe−C5′ bond between 5′-dAdo• and the [4Fe−4S] cluster. An analogous organometallic bond is found in the well- known adenosylcobalamin (coenzyme B12) cofactor used to initiate radical reactions via a 5′-dAdo• intermediate. Liberation of a reactive 5′-dAdo• intermediate via homolytic metal−carbon bond cleavage thus appears to be similar for Ω and coenzyme B12. However, coenzyme B12 is involved in enzymes catalyzing only a small number (∼12) of distinct reactions, whereas the RS superfamily has more than 100 000 distinct sequences and over 66 80 reaction types characterized to date. The appearance of Ω across the RS superfamily therefore dramatically enlarges the sphere of bio-organometallic chemistry in Nature. Introduction Radical S-adenosyl-L-methionine (SAM) enzymes comprise a vast superfamily, catalyzing diverse reactions essential to all life through homolytic SAM cleavage to liberate the highly reactive 5′-deoxyadenosyl radical (5′-dAdo•).1−3 In the consensus mechanism for radical SAM (RS) enzymes, electron-transfer to the sulfonium center of SAM from a reduced active-site [4Fe−4S] cluster causes reductive cleavage of the S−C(5′) bond to directly liberate 5′-dAdo• for H atom abstraction from substrate (Figure S1).4−8 However, this mechanism of radical initiation was put in question by the report of an organometallic reaction intermediate, denoted Ω, in catalysis by the RS pyruvate formate−lyase activating enzyme (PFL-AE).9 This intermediate, which has a carbon of 5′- dAdo• bonded to the unique Fe of the [4Fe−4S] cluster, formed subsequent to SAM cleavage, and 5′-dAdo• was only liberated through homolysis of the Fe−C bond of Ω.9 Here rapid freeze-quench (RFQ) EPR/ENDOR studies of a suite of enzymes, selected to collectively represent the broad range of RS superfamily reactions, implicate this organometallic intermediate as central to radical initiation across the RS superfamily. This leads us to propose a paradigm shift for radical initiation in these enzymes, that, with accompanying insights into the Ω structure determination, mechanistically unifies the RS 67 and adenosylcobalamin (coenzyme B12) enzymes: both involve homolysis of a metal-5′- deoxyadenosyl bond to liberate 5′-dAdo• for initiation of radical chemistry. 1. Is Ω the Result of Protein Conformational Rearrangements during Assembly of the PFL-AE/SAM/PFL Ternary Complex? The observation of Ω formation during rapid freeze-quench (RFQ) after mixing reduced PFL-AE with the two substrates, (PFL + SAM), 9 raised the possibility that Ω was perhaps a means by which to store and control the nascent 5′-dAdo• during the complex conformational changes required for positioning of the target H atom as PFL-AE binds SAM and its 170 kDa substrate protein PFL.10−12 To examine this possibility, we employed three different mixing conditions to RFQ trap intermediate states of the PFL-AE/SAM/PFL reaction (Figure 1). In all cases, reactions were RFQ trapped at 500 ms, as our previous work showed that Ω formation was maximal at this time.9 We repeated our original protocol,9 rapid mixing of reduced PFL-AE with a solution that contained both the SAM and PFL substrates; in that case, both substrates must properly bind to PFL-AE prior to reaction. We then examined two other mixing conditions. In one, a solution of reduced (PFL-AE + SAM) was mixed with PFL; in this case, the SAM substrate is prebound but PFL has to bind after mixing. Finally, we prepared reduced (PFL-AE + PFL), thus preforming the PFL- AE/PFL complex with its attendant rearrangement of both proteins, and rapid-mixed with SAM. Figure 1 shows that in all three cases, Ω is formed with its characteristic EPR axial signal, (g|| = 2.035, g⊥ = 2.004).9 We conclude that in the PFL-AE catalyzed reaction, an extreme case of active-site and target protein rearrangement,11,12 Ω is a central intermediate that is formed regardless of the details/order of mixing. 68 Figure 3.1. Top, premixed (PFL + SAM), RFQ Figure 3.2. Reactions catalyzed by the with PFL-AE. Middle, (PFL-AE + SAM) RFQ radical SAM enzymes studied in this with PFL. Bottom, (PFL-AE + PFL) + SAM; work. Glycyl radical enzyme activating the sli ght increase in g suggests a slightly || enzyme (GRE-AE) refers to both PFL-AE different conformation of Ω. Feature to low and RNR-AE. field of Ω signal in bottom spectrum due to Cu2+ contamination from Cu wheels used for freezing in RFQ apparatus. Conditions: Freeze-quenched, 500 ms; frequency, 9.374 GHz (top), 9.374 GHz (middle), 9.375 GHz (bottom); modulation, 10 G; T = 40 K. Samples cryoannealed at 150 K to remove a small overlapping signal, see Figure S2. 2. Is Ω Mechanistically Formed throughout the RS Superfamily? To test the intermediacy of Ω broadly across the RS superfamily, we freeze-quench trapped intermediates in the reactions of PFL-AE and six additional canonical RS enzymes that 69 3 perform diverse enzymatic functions representative of the entire superfamily (Figure 2). First, these enzymes span the two major RS subclasses by including representatives that use SAM as a cosubstrate (PFL-AE, RNR-AE, HydG, PoyD and OspD) and those that use 3 13 SAM as a cofactor (LAM and SPL). Of the additional enzymes, two (RNR-AE and 14 SPL ) catalyze reactions on macromolecular substrates (anaerobic ribonucleotide 15 reductase and DNA, respectively), two act on peptide substrates (PoyD and OspD), two 16,17 catalyze reactions of small molecule substrates (tyrosine for HydG and lysine for 18 19 LAM ), and finally one involves a second iron−sulfur cluster in catalysis (HydG). As described in Supporting Information, samples of all these diverse RS representatives were expressed in Escherichia coli, purified under anoxic conditions, subjected to Fe/S cluster reconstitution where necessary, photoreduced to generate the 1+ catalytically active [4Fe−4S] cluster, rapid-mixed with a solution containing SAM and the appropriate substrate, and quenched 500 ms after mixing, followed by brief annealing at 150 K to remove a small contaminating signal (Figure S3). The resulting EPR spectra 1+ show that in every instance the reactant [4Fe−4S] cluster signal has been completely replaced by the organometallic intermediate, Ω (Figures 3, S3, S4). Moreover, we have initiated experiments to monitor how Ω converts to product upon annealing, as previously 9 shown for PFL-AE, using representative enzymes from each of the two major RS subclasses, those using SAM as cosubstrate (RNR-AE, HydG) and as cofactor (LAM), and preliminary indications are that this occurs as anticipated (see Figure S5). 70 The Ω spectra show slight variations from enzyme to enzyme in the shape of the g|| feature of this axial signal (Figure 3); these differences likely arise from small variations in the conformation of the intermediate, as seen in the Ω signal for PFL-AE/PFL generated under different mixing conditions (Figure 1). 3. Detailed Structure of Ω. With Ω now identified as a ubiquitous RS intermediate, we carried out EPR and ENDOR measurements to refine the determination 57 of its structure. Originally, Ω was identified as involving the [4Fe−4S] cluster by its Fe 9 57 ENDOR response. The Fe line-broadening of the Ω EPR signal (Figure S6), whose 57 simulation requires inclusion of hyperfine interactions with multiple cluster Fe ions, further confirms that the spin of omega is carried by the [4Fe−4S] cluster. Incorporation of 9 1 the 5′-dAdo fragment of SAM in Ω is confirmed here by observation of a loss of H ENDOR signals when Ω is prepared with uniformly labeled [D8-ado]-SAM ([adenosyl- 2,8-D2-1′,2′,3′,4′,5′,5′′-D6]-SAM) (Figure 4). The use of specifically labeled [5′,5′′-D2- ado]-SAM produces the same loss of 1H ENDOR signal as seen for Ω made with uniformly [D8-ado]-SAM (Figure 4), unambiguously identifying the C5′ carbon of 5′-dAdo• as 1 1 forming the Fe−C bond in Ω. Note that both the H coupling, A( H) ∼ 7−8 MHz, which 2 also causes a distinct reduction in EPR line-width upon H replacement (Figure S7), and 13 13 13 the previously observed C couplings from uniformly C-labeled SAM, aiso( C) ∼ 9 9 14/15 MHz are far too small to arise from an isolated 5′-dAdo•. Finally, N-Met-SAM give 71 14/15 N ENDOR signals characteristic of direct 14 coordination to the Fe, A( N) ≈ 4 MHz (Figure S8), confirming the retention of methionine coordination at the unique iron. Together, these EPR/ENDOR observations leave no doubt that the Ω EPR signal arises from an organometallic complex that contains a bond between a cluster Fe and the C5′ carbon of 5′-dAdo•, with the methionine fragment of SAM anchored to the cluster via amino coordination, as shown in Figure 3.3. Normalized EPR spectra of Ω Chart 1 (left). The g-values of Ω follow the formed in RS reactions freeze-quenched at 500 ms, taken after annealing 1 min at 150 K; pattern of a [4Fe−4S]3+ cluster, g > g ≳ 2,20 spectra before annealing, Figure S3. RFQ || ⊥ mixing condition: (substrate + SAM) + RS enzyme, freeze-quenched 500 ms after mixing. which suggests a formal description of Conditions: frequency, 9.375 GHz; modulation amplitude, 10 G; T = 40 K. the intermediate as an [4Fe−4S]3+ cluster whose unique Fe is bound to the C5′- Chart 1 adenosyl carbanion (Chart 1). The resulting structure and reactivity of Ω exhibit intriguing similarity to adenosylcobalamin (AdoCbl, coenzyme B12), which has a bond from the C5′ carbon of a deoxyadenosyl moiety to the 72 cobalt of cobalamin, Chart 1 (right), and undergoes homolytic Co−C bond cleavage to generate 5′-dAdo• to abstract an H• from substrate. 4. Mechanism: How does Ω Form, Figure 3.4. 35 GHz CW 1H ENDOR at g = 2.0134 and Then Liberate 5′-dAdo•? Formation of Ω for PFL-AE/ PFL with (black) 1H-SAM; (blue-dashed) [D -ado]-SAM ([adenosyl- 2,8-D - 8 2 of Ω might be indirect, (a) via reductive 1′,2′,3′,4′,5′,5′′-D ]-SAM); (red) [5′,5′′-D2-ado]- 6 SAM ([adenosyl-5′,5′′-D -SAM]). Mixing con- 2 cleavage of SAM followed by combination ditions: (PFL + SAM) + PFL-AE. ENDOR conditions: microwave frequency, 35.05 GHz; modulation amplitude, 0.5 G, +0.75 MHz/s; T = 2 of 5′-dAdo• with the unique Fe of the K. 2+ [4Fe−4S] cluster (Figure 5, path 1); or it may occur in a single concerted step (Figure 5, path 2), via either (b) direct nucleophilic 1+ attack of the unique [4Fe−4S] cluster Fe at the 5′-C of SAM; or (c) concerted reductive cleavage/Ω formation initiated by interaction of the Fe with the SAM sulfur. The active site geometry of canonical RS enzymes places the S−C(5′) bond of SAM trans to the Fe−S 21 interaction (Figure 5). This geometry is not conducive to nucleophilic attack of the unique Fe at the 5′-C of SAM to produce Ω, thus disfavoring mechanism (b), but it would permit 22,23 reductive cleavage routes (a) and (c) to form Ω (Figure 5). However, these two pathways themselves pose the perplexing question: why does 5′-dAdo• move toward the cluster and bind to the unique Fe to form Ω upon cleavage of the S−C bond, rather than directly moving away and attacking the substrate? Given the active-site geometry, we suggest that Ω formation is a direct mechanistic consequence of SAM activation for 73 reductive cleavage by either routes (a) or (c): in both cases the SAM sulfur must migrate, in the transition state, toward the unique Fe of the [4Fe−4S] cluster, and when the S−C(5′) bond breaks through reductive cleavage, the 5′-dAdo• fragment need only continue along this trajectory to interact with the unique iron and form Ω. In a structural contrast, Lin and co-workers recently reported that in the noncanonical RS enzyme Dph2, the active-site Figure 3.5. Pathways for liberating 5′-dAdo• for H atom abstraction through formation of catalytically competent Ω upon SAM cleavage. Ω may be formed via reductive SAM cleavage then recombination of the 5′-dAdo radical with the [4Fe−4S]2+ cluster (pathway 1), or directly (pathway 2) via concerted reductive cleavage/Fe−C bond formation or nucleophilic attack of the unique iron on the 5′-carbon. In addition, at present it cannot be excluded that a minority of the generated 5′-dAdo radical “escapes” to attack substrate directly, without Ω formation. architecture is set up for direct nucleophilic attack of the unique iron on the Cγ(met) to form a kinetically competent organometallic intermediate with an Fe−Cγ(met) bond.24 These considerations suggest why both Dph2 and canonical RS enzymes generate their key radical intermediates through prior formation of organometallic species, Ω in the case of canonical RS enzymes. How is the Fe−C(5′) bond of Ω then activated to homolytically liberate 5′-dAdo• for reaction with substrate? This issue applies equally to the Co−C(5′) bond in the organometallic AdoCbl Co−C(5′) of B12-radical enzymes, and this question in fact 74 highlights not only the similarities but also the differences between AdoCbl and Ω. The AdoCbl cofactor is a stable, isolable compound, and requires significant activation for Co−C(5′) bond homolysis.25,26 In contrast, Ω is a true reactive intermediate, and in PFL- AE, Ω undergoes facile Fe−C bond cleavage even at low temperature (∼170 K).9 The novel organometallic intermediate Ω, whose structure is shown in Chart 1, forms in a suite of enzymes that represent the broad range of superfamily reactions, regardless of whether the enzymes consume SAM as cosubstrate or reuse SAM as cofactor, and independent of the size and complexity of the target substrate for H atom abstraction. This indicates that Ω is a mechanistically central feature for radical initiation throughout the superfamily. In the long-held mechanism of radical initiation (Figure S1), reductive cleavage of SAM directly liberates 5′-dAdo•, which then abstracts a H atom from substrate.3 Our results reveal a sharply different picture: liberation of the 5′-dAdo• for substrate H atom abstraction proceeds through the initial formation of Ω, followed by homolytic cleavage of the Fe−C(5′) bond to liberate 5′-dAdo• for reaction (Figure 5). This is a paradigm shift for the mechanism of radical initiation by enzymes of the RS superfamily. The identification of Ω as integral to the mechanism of RS enzymes clearly unifies the RS and B12-dependent radical- forming enzymes, despite the differences in reactivity of AdoCbl and Ω: both utilize the cleavage of a metal−carbon bond (Chart 1) to liberate the reactive 5′-dAdo• for H atom abstraction from substrate. Our findings thus complete the integration of the two enzyme classes, as first contemplated by both Knappe27 and 29 Frey.28 Beyond that, given the vast reach of the RS superfamily across all domains of life, 75 these results dramatically enlarge the scope and importance of bioorganometallic chemistry in Nature. 76 References 1. Sofia, H. J.; Chen, G.; Hetzler, B. G.; Reyes-Spindola, J. F.; Miller, N. E. Nucleic Acids Res. 2001, 29, 1097−1106. 2. Frey, P. A.; Hegeman, A. D.; Ruzicka, F. J. Crit. Rev. Biochem. Mol. Biol. 2008, 43, 63−88. 3. Broderick, J. B.; Duffus, B. R.; Duschene, K. S.; Shepard, E. M. Chem. Rev. 2014, 114, 4229−4317. 4. Frey, M.; Rothe, M.; Wagner, A. F. V.; Knappe, J. J. Biol. Chem. 1994, 269, 12432−12437. 5. Moss, M.; Frey, P. A. J. Biol. Chem. 1987, 262, 14859−14862. 6. Cheek, J.; Broderick, J. B. J. Am. Chem. Soc. 2002, 124, 2860− 2861. 7. Magnusson, O. T.; Reed, G. H.; Frey, P. A. J. Am. Chem. Soc. 1999, 121, 9764−9765. 8. Horitani, M.; Byer, A. S.; Shisler, K. A.; Chandra, T.; Broderick, J. B.; Hoffman, B. M. J. Am. Chem. Soc. 2015, 137, 7111−7121. 9. Horitani, M.; Shisler, K. A.; Broderick, W. E.; Hutcheson, R. U.; Duschene, K. S.; Marts, A. R.; Hoffman, B. M.; Broderick, J. B. Science 2016, 352, 822−825. 10. Becker, A.; Fritz-Wolf, K.; Kabsch, W.; Knappe, J.; Schultz, S.; Wagner, A. F. V. Nat. Struct. Biol. 1999, 6, 969−975. 11. Vey, J. L.; Yang, J.; Li, M.; Broderick, W. E.; Broderick, J. B.; Drennan, C. L. Proc. Natl. Acad. Sci. U. S. A. 2008, 105, 16137− 16141. 12. Peng, Y.; Veneziano, S. E.; Gillispie, G. D.; Broderick, J. B. J. Biol. Chem. 2010, 285, 27224−27231. 13. Sun, X.; Eliasson, R.; Pontis, E.; Andersson, J.; Buist, G.; Sjöberg, B.-M.; Reichard, P. J. Biol. Chem. 1995, 270, 2443−2446. 14. Rebeil, R.; Sun, Y.; Chooback, L.; Pedraza-Reyes, M.; Kinsland, C.; Begley, T. P.; Nicholson, W. L. J. Bacteriol. 1998, 180, 4879−4885. (15) Morinaka, B. I.; 77 Vagstad, A. L.; Helf, M. J.; Gugger, M.; Kegler, C.; Freeman, M. F.; Bode, H. B.; Piel, J. Angew. Chem., Int. Ed. 2014, 53, 8503−8507. 15. Shepard, E. M.; Duffus, B. R.; McGlynn, S. E.; Challand, M. R.; Swanson, K. D.; Roach, P. L.; George, S. J.; Cramer, S. P.; Peters, J. W.; Broderick, J. B. J. Am. Chem. Soc. 2010, 132, 9247−9249. 16. Driesener, R. C.; Challand, M. R.; McGlynn, S. E.; Shepard, E. M.; Boyd, E. S.; Broderick, J. B.; Peters, J. W.; Roach, P. L. Angew. Chem., Int. Ed. 2010, 49, 1687−90. 17. Chirpich, T. P.; Zappia, V.; Costilow, R. N.; Barker, H. A. J. Biol. Chem. 1970, 245, 1778−1789. 18. Driesener, R. C.; Duffus, B. R.; Shepard, E. M.; Bruzas, I. R.; Duschene, K. S.; Coleman, N. J.-R.; Marrison, A. P. G.; Salvadori, E.; Kay, C. W. M.; Peters, J. W.; Broderick, J. B.; Roach, P. L. Biochemistry 2013, 52, 8696−8707. 19. Cutsail, G. E., III; Telser, J.; Hoffman, B. M. Biochim. Biophys. Acta, Mol. Cell Res. 2015, 1853, 1370−1394. 20. Vey, J. L.; Drennan, C. L. Chem. Rev. 2011, 111, 2487−2506. (22) Kampmeier, J. A. Biochemistry 2010, 49, 10770−2. 21. Nicolet, Y.; Amara, P.; Mouesca, J.-M.; Fontecilla-Camps, J. C. Proc. Natl. Acad. Sci. U. S. A. 2009, 106, 14867−14871. 22. Dong, M.; Kathiresan, V.; Fenwick, M. K.; Torelli, A. T.; Zhang, Y.; Caranto, J.; Dzikovski, B.; Sharma, A.; Lancaster, K. M.; Freed, J. H.; Ealick, S. E.; Hoffman, B. M.; Lin, H. Science 2018, 359, 1247−1250. 23. Banerjee, R. Chem. Rev. 2003, 103, 2083−94. 24. Brown, K. L. Chem. Rev. 2005, 105, 2075−2149. 25. Knappe,J.;Neugebauer,F.A.;Blaschkowski,H.P.;Gan̈zler,M. Proc. Natl. Acad. Sci. U. S. A. 1984, 81, 1332−1335. 26. Frey, P. A. FASEB J. 1993, 7, 662−670. 78 27. Akiva, E.; Brown, S.; Almonacid, D. E.; Barber, A. E.; Custer, A. F.; Hicks, M. A.; Huang, C. C.; Lauck, F.; Mashiyama, S. T.; Meng, E. C.; Mischel, D.; Morris, J. H.; Ojha, S.; Schnoes, A. M.; Stryke, D.; Yunes, J. M.; Ferrin, T. E.; Holliday, G. L.; Babbitt, P. C. Nucleic Acids Res. 2014, 42, D521−D530. 79 Supplementary Information Materials and Methods [13C ,15 10 N5]-Adenosine 5′-triphosphate sodium salt and adenosine-2,8-D2- 1′,2′,3′,4′,5′,5″-D6-5′-triphosphate sodium salt solutions were purchased from MilliporeSigma/ISOTEC® Stable Isotopes. [15N]-Methionine was purchased from Cambridge Isotope Labs, Inc. (96 - 98%, NLM 752). Iron-57 (93-95%) was purchased form Cambridge Isotope Labs, Inc. (FLM-1812PK). All unlabeled adenosine 5′- triphosphate and L-methionine were purchased from MilliporeSigma. HydG Preparations [FeFe]-hydrogenase maturase HydG preparations were performed as described previously,1 with minor modifications. Briefly, the HydG gene (hydG from Clostridium acetobutylicum) cloned into a pCDF DUETTM-1 vector allowed for expression of HydG with an N-terminal His6 tag in BL21-(∆ISCR) E. coli cells. After inoculation from overnight starter cultures, cell culture growth in phosphate buffered LB media (9 L in six 2.8 L Fernbach flasks), supplemented with kanamycin (30 μg/mL) and streptomycin (50 μg/mL) antibiotics, glucose (5 g/L), and iron in the form of ferric ammonium citrate (4.2 g total in 9 L) and maintained at 37 ̊C with 200 rpm shaking, exhibited good over-expression of HydG upon induction with IPTG (0.5 mM) at an OD600 of 0.5 AU via comparatively high protein concentrations on SDS-PAGE. After transfer to 4°C, this culture was sparged overnight (14 - 16 h) with N2(g) until cell pellets (~30 - 40 g) were harvested the next day and stored at -80°C. To isolate HydG with minimal cluster loss, all lysis and purification steps were performed in an anaerobic Coy vinyl glove box (Coy Laboratories, Grass Lake, MI). At a ratio of ~2 mL of buffer per 1 g of cell pellet, lysis was performed at 4 ̊C in Buffer A (50 mM HEPES, 250 mM KCl, 5% glycerol, 10 mM imidazole, pH 8.0) in the presence of lysozyme (180 μg/mL), phenylmethylsulfonyl floride (PMSF, 180 μg/mL final, added in 1 mL MeOH), DNase (50 μg/mL), RNase (50 μg/mL), 0.5% TritonTM X-100 (v/v), and MgCl2 (2 mg/mL). After centrifugation (38,000 x g, 60 min., 4 ̊C), the HydG clarified lysate was purified via FPLC using a step gradient into Buffer B (50 mM 4-(2- hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES), 250 mM KCl, 5% glycerol, 500 mM imidazole, pH 8.0) on a Ni2+-affinity column (5 mL). The purified enzyme solution was buffer exchanged into Buffer C (50 mM HEPES, 50 mM KCl, 5% glycerol, pH 8.0) before reconstitution. To increase [4Fe-4S] cluster content in HydG, an iron-sulfur cluster reconstitution was performed at 4 C̊ as described elsewhere.2 Briefly, purified HydG (100 μM) was incubated with DTT (5 mM) for 5 minutes in Buffer C, before sodium sulfide (0.6 mM final concentration, from a 50 mM stock) and ferric chloride (0.6 mM final concentration, from a 10 mM stock) were added, respectively, over the course of an hour. After a 3.5 hour 80 incubation at 4 ̊C, the reconstitution solution was centrifuged (26,000 x g, 10 min., 4 ̊C), gel filtrated, and concentrated to yield a final iron content of 7.8 ± 0.1 irons/protein. EPR spectroscopy confirmed [4Fe-4S]+ cluster presence in photoreduced enzyme. Lysine 2,3-aminomutase Preparations Lysine 2,3-aminomutase (LAM) preparations were performed as described previously2 to yield a final iron content of 3.9 ± 0.1 irons/protein. EPR spectroscopy confirmed [4Fe-4S]+ cluster presence in photoreduced enzyme. OspD and OspA (substrate) Preparations The gene encoding OspD (OSCI_3660007 from Oscillatoria sp. PCC 6506) cloned into a modified pCDF DUETTM-1 vector allowed for expression of OspD protein with a TEV protease cleavable N-terminal His6 tag in BL21-(∆ISCR) E. coli cells. After inoculation from overnight starter cultures, cell culture growth in terrific broth media (1.2% tryptone, 2.4% yeast extract, 0.5% glycerol and 89 mM phosphate; 6 L in six 2.8 L Fernbach flasks) was supplemented with spectinomycin (50 μg/mL) and kanamycin (34 μg/mL) antibiotics and maintained at 37 ̊C with 200 rpm shaking. At an OD600 of 1.2, the cultures were chilled on ice for 15 minutes and supplemented with ferrous ammonium sulfate (FAS, 0.25 mM), induced with IPTG (1 mM), and incubated for over 24 h at 16 ̊C, with shaking at 220 rpm and N2(g) sparge until cell pellets (~40 g) were harvested the next day and stored at -80 °C. To isolate OspD with minimal cluster loss, all lysis and purification steps were performed in an anaerobic Coy vinyl glove box. At a ratio of ~2 mL of buffer per 1 g of cell pellet, lysis at ~21 C̊ in Buffer A (50 mM HEPES, 150 mM KCl, 10% glycerol (w/v), pH 8.0) in the presence of imidazole (1 mM), fresh DTT (2 mM, added 5 min prior to lysis), MgCl2 (2.5 mg/mL), lysozyme (0.18 mg/mL), TritonTM X-100 (1% w/v), PMSF (2.1 mM), and DNase and RNase (0.18 mg/mL each). After centrifugation (38,000 x g, 60 min., 4 ̊C), the OspD clarified lysate was purified via gravity- flow on a Ni-NTA agarose column (5 mL) pre-equilibrated with Buffer A with the following procedure: 1) a 10 CV wash step with Buffer A, 2) a 10 CV wash with 15% Buffer B (50 mM HEPES, 150 mM KCl, 10 % glycerol (w/v), 100 mM imidiazole, pH 8.0), 3) elution of pure OspD in 75% Buffer B. Pure fractions (confirmed by SDS-PAGE) were concentrated, gel filtered (Sephadex G25 resin column, 75 mL) in Buffer A, and re-concentrated with Amicon 10 kDa MWCO spin filters. To increase [4Fe-4S] cluster content in OspD, an iron-sulfur cluster reconstitution was performed at 4 ̊C as described elsewhere.2 Briefly, purified OspD (100 μM) was incubated with DTT (5 mM) for 5 minutes in Buffer A, before ferric chloride (0.6 mM final concentration, from a 10 mM stock) and sodium sulfide (0.6 mM final concentration, from a 50 mM stock) were added, respectively, over the course of an hour. After a 2 hour 81 incubation at 4 ̊C, the reconstitution solution was centrifuged (26,000 x g, 10 min., 4 ̊C), gel filtrated, and concentrated to yield a final iron content of ~4.0 ± 0.2 irons/protein. EPR spectroscopy confirmed [4Fe-4S]+ cluster presence in photoreduced enzyme. The gene encoding OspA (OSCI_3660009 from Oscillatoria sp. PCC 6506) in pET28 was used to produce the peptide substrate OspA with a thrombin-cleavable N- terminal His6 tag, which was prepared as described previously.3 If the lysis and purification steps were performed aerobically, the OspA protein was made anaerobic through several Schlenk line nitrogen/vacuum cycles. PFL-AE and PFL (substrate) Preparations PFL-AE preparations were performed as described previously2, 4-6 with minor modifications. Briefly, the PFL-AE gene (pflA from E. coli) cloned into a pCAL-n-EK vector allowed for expression of PFL-AE protein without a tag in BL21(DE3)pLysS E. coli cells. After inoculation from overnight starter cultures, cell culture growth in LB media (9 L in six 2.8 L Fernbach flasks) was supplemented with ampicillin (50 μg/mL) antibiotic and maintained at 37 ̊C with 200 rpm shaking. At an OD600 of ~0.3, the cell culture was supplemented with glucose (5 g/L); at an OD600 of ~0.8, the cell culture induced with IPTG (0.25 mM) and supplemented with FAS (0.20 mM) and L-cysteine (0.20 mM), wherein the temperature was reduced to 30°C and the pH maintained (every 30 min.) between pH 7.2- 7.5 with NH4OH (6 M) or HCl (6 M) as necessary, until the pH no longer showed a significant change (~4 - 5 h). Additional FAS and L-cysteine were added (0.40 mM, final), and the culture was transferred to 4 ̊C and sparged overnight (14 - 16 h) with N2(g) until cell pellets (~50 - 60 g) were harvested the next day and stored at -80 °C. To isolate PFL-AE with minimal cluster loss, all lysis and purification steps were performed in an anaerobic Coy vinyl glove box. At a ratio of ~1- 1.5 mL of buffer per 1 g of cell pellet, lysis was performed at 4 ̊C in buffer (50 mM Tris, 100 mM NaCl, 5% w/v glycerol, 1% TritonTM X-100, 10 mM MgCl2, 1.0 mM dithiothreitol (DTT), pH 7.5) in the presence of lysozyme (0.32 mg/mL), PMSF (2 mM), DNase I and RNase A (0.01 mg/mL). After centrifugation (38,000 x g, 60 min., 4 ̊C), the PFL-AE clarified lysate was purified as previously described4-6 except a lower concentration of NaCl was used in the purification buffer (50 mM Tris, 100 mM NaCl, 5% w/v glycerol, 1.0 mM dithiothreitol (DTT), pH 7.5). As the final cluster content determined by atomic absorbance spectroscopy was 3.9 ± 0.1 or 4.0 ± 0.1 Fe/protein, iron-sulfur cluster reconstitution was not needed. EPR spectroscopy confirmed [4Fe-4S]+ cluster presence in photoreduced enzyme. PFL preparations were performed as described previously2, 4-6 with minor modifications. The PFL gene (pfl from E. coli) cloned into the pKK plasmid vector allowed for expression of PFL protein without a tag in BL21(DE3)pLysS E. coli cells. PFL was grown, lysed, and purified as described previously4, 6 with the following exceptions: 1) during growth, induction with IPTG (0.2 mM) was immediately followed by a temperature drop to 30 ̊C and 14 – 16 hours of additional incubation, 2) during lysis and purification, DTT was not present in buffers (Buffer A: 20 mM Tris, pH 7.2; Buffer B: 20 mM Tris, 500 82 mM NaCl, pH 7.2; Buffer C: 20 mM Tris, 1 M (NH4)2SO4, pH 7.2). If the lysis and purification steps were performed aerobically, the PFL protein was made anaerobic through several Schlenk line nitrogen/vacuum cycles. PoyD and PoyA (substrate) Preparations The gene encoding PoyD (poyD from Candidatus Entotheonella factor TSY1) cloned into a modified pCDF DUETTM-1 vector allowed for expression of PoyD protein with a TEV protease cleavable N-terminal His6 tag in BL21-(∆ISCR) E. coli cells. After inoculation from overnight starter cultures, cell culture growth in terrific broth media (1.2% tryptone, 2.4% yeast extract, 0.5% glycerol and 89 mM phosphate; 6 L in six 2.8 L Fernbach flasks) was supplemented with spectinomycin (50 μg/mL) and kanamycin (34 μg/mL) antibiotics and maintained at 37 ̊C with 200 rpm shaking. At an OD600 of 1.2, the cultures were chilled on ice for 15 minutes and supplemented with FAS (0.25 mM), induced with IPTG (1 mM), and incubated for over 24 h at 16 ̊C, with shaking at 220 rpm and N2(g) sparge until cell pellets (~40 g) were harvested the next day and stored at -80°C. To isolate PoyD with minimal cluster loss, all lysis and purification steps were performed in an anaerobic Coy vinyl glove box. At a ratio of ~2 mL of buffer per 1 g of cell pellet, lysis at ~21 ̊C in Buffer A (50 mM HEPES, 150 mM KCl, 10 % glycerol (w/v), pH 8.0) in the presence of imidazole (1 mM), fresh DTT (2 mM, added 5 min prior to lysis), TritonTM X-100 (1% w/v), PMSF (180 μg/mL final, added in 1 mL MeOH), DNase and RNase (0.18 mg/mL each) was performed via sonication (15 sec. pulses, 59 sec. rest, 60% amplitude, 5 min. total pulse time). After centrifugation (38,000 x g; 60 min.; 4 C̊ ), the PoyD clarified lysate was purified via FPLC using a HisTrapTM column (5 mL) pre- equilibrated with Buffer A with the following procedure: 1) a 10 CV wash step with Buffer A, 2) a 10 CV wash with 25% Buffer B (50 mM HEPES, 150 mM KCl, 10 % glycerol (w/v), 100 mM imidiazole, pH 8.0), 3) elution of pure PoyD in 100% Buffer B. Pure fractions (confirmed by SDS-PAGE) were gel filtered (Sephadex G25 resin column, 75 mL) in Buffer A and concentrated with Amicon 15 kDa MWCO spin filters. To increase [4Fe-4S] cluster content in PoyD, an iron-sulfur cluster reconstitution was performed at 4 ̊C as described elsewhere.2 Briefly, purified PoyD (100 μM) was incubated with DTT (5 mM) for 5 minutes in Buffer A, before ferric chloride (0.6 mM final concentration, from a 10 mM stock) and sodium sulfide (0.6 mM final concentration, from a 50 mM stock) were added, respectively, over the course of an hour. After a 2 hour incubation at 4 C̊ , the reconstitution solution was centrifuged (26,000 x g, 10 min., 4 ̊C), gel filtrated, and concentrated to yield a final iron content of 3.6 ± 0.1 irons/protein. EPR spectroscopy confirmed [4Fe-4S]+ cluster presence in photoreduced enzyme. The gene encoding PoyA (poyA25 gene from Candidatus Entotheonella factor TSY1) cloned into a modified vector allowed for expression of PoyA protein with a TEV protease cleavable fusion with an N-terminally His6-tagged maltose binding protein (MBP) in BL21(DE3) E. coli cells. Though full length PoyA cannot be expressed without co- expression of PoyD epimerase and has very limited unepimerized production as an MBP 83 fusion, the truncation variant PoyA25 can be produced in good yields.7 After inoculation from overnight starter cultures, cell culture growth in LB (or TB) media (1 L a 2.8 L Fernbach flask) was supplemented with kanamycin (25 μg/mL) antibiotic and maintained at 37 C̊ with shaking (~220 rpm) to an OD600 of 0.7 – 1.0 for the LB media or an OD600 of 1.5 – 2.0 in the TB media. The cultures were chilled to 16 – 20 ̊C, induced with IPTG (1 mM), and incubated overnight at 16 – 20 ̊C while shaking (~220 rpm). To harvest the cells, the cultures were centrifuged (12,000 x g, 10 min., 4 ̊C) and the wet cell pellets were immediately frozen in liquid nitrogen and stored in -80 ̊C. As PoyA does not contain an FeS cluster, all lysis and purification steps were performed aerobically. To cleave the MBP solubility tag, purification of PoyA required a double Ni-affinity purification with a TEV cleavage step in between. Isolation of PoyA started with cell lysis by sonication (10x 10 sec on, 10 sec off, 70% amplitude 6.4 mm probe QSonica Q700 sonicator) in 4 mL per gram cell pellet cold Buffer P (100 mM NaPO4, 300 mM NaCl, 10% glycerol, 5 mM imidazole, pH 8.0) supplemented with protease inhibitor cocktail (cOmplete Protease Inhibitor Cocktail tablets, Roche) and 1 mg/mL lysozyme treatment for 1 h prior to sonication. After sonication, the cell lysate was centrifuged (27,000 x g, 30 min., 4 ̊C) and the clarified lysate supernatant was bound to cOmplete His-tag Purification Resin (Roche) for 1 h before being transferred to a gravity flow column and treated with the following wash steps: 1) 4x 5 CV Buffer P; 2) 1x 5 CV Buffer P with 20 mM imidazole; and 3) 3x 1 CV Buffer P with 250 mM imidazole to elute PoyA. Fractions with the His6-MBP-PoyA identified by SDS-PAGE were pooled, concentrated (Amicon Ultra 10 kDa MWCO filter units), and buffer exchanged into Buffer T (50 mM Tris, 0.5 mM EDTA, 1 mM DTT, pH 8.0) using a PD-10 BioRad desalting column (load 2.5 mL, elute 3.5 mL). TEV cleavage at room temperature overnight with a ratio of 1:100 of His6TEV- protease to His6-MBP-PoyA (expressed and purified as previously described8) removed the His6-MBP tag, as monitored by SDS-PAGE. Once complete, this cleavage reaction was supplemented with NaCl (150 mM), centrifuged to clarify any precipitant that formed, and the supernatant was passed over a cOmplete His Purification resin twice to isolate PoyA. The flowthrough and first wash (1 CV Buffer T) were pooled and concentrated to give relatively pure PoyA, which was then supplemented with 10% glycerol and flash frozen in liquid nitrogen prior to storage at -80 C̊ . As this substrate protein doesn’t coordinate FeS clusters, reconstitution was not needed. If the lysis and purification steps were performed aerobically, the PoyA protein was made anaerobic through several Schlenk line nitrogen/vacuum cycles. Anaerobic RNR-AE and RNR (substrate) Preparations From E. coli, the gene encoding RNR-AE was isolated using primers 5′-TAC ATA TGA ATT ATC ATC AGT ACT ATC CTG TCG-3′ (forward) and 5′-CTA AGC TTT CAT CGC AAA TGA ACC ACC -3′ (reverse) which imparted restriction enzyme cleavage sites for NdeI and HindIII restriction enzymes. This RNR-AE gene (nrdG from E. coli) cloned into a pCAL-n-EK vector allowed for expression of the 17 kDa RNR-AE protein 84 without a tag in BL21(DE3)pLysS E. coli cells. After inoculation from overnight starter cultures, cell culture growth in minimal media (9 L in a 14 L fermenter) supplemented with ampicillin (50 μg/mL) and chloramphenicol (50 μg/mL) antibiotic and FAS (0.3 mM) and maintained at 37 ̊C with 200 rpm shaking exhibited good over- expression. Upon induction at OD600 of 0.5 with IPTG (0.5 mM), the temperature was dropped to 25 C̊ and 2.5 hours later to 4 ̊C, wherein this culture was sparged overnight (14 - 16 h) with N2(g) until cell pellets (~20 g) were harvested the next day and stored at -80°C. To isolate RNR-AE with minimal cluster loss, all lysis and purification steps were performed in an anaerobic Coy vinyl glove box. At a ratio of ~2 mL of buffer per 1 g of cell pellet, lysis at 4 ̊C in buffer (50 mM Tris, 200 mM NaCl, 5% glycerol, pH 8.5, 1% TritonTM X-100, 10 mM MgCl2) was supplemented with PMSF (1 mM), lysozyme (8 mg/40 mL), and DNase and RNase (0.1 mg per 40 mL). After 15 minutes on ice, the cells were chemically lysed for 1 h, with stirring. The lysate solution was centrifuged (38,000 x g, 30 min., 4 C̊ ) and the supernatant was subjected to two ammonium sulfate cuts (0 – 20% ammonium sulfate and 20 – 60% ammonium sulfate). After centrifugation of the last cut (38,000 x g, 30 min., 4 ̊C), the RNR-AE was found in the pellet which was solubilized on ice in Buffer R (50 mM Tris, 200 mM NaCl, 1 mM DTT, pH 8.5) and loaded at 3 mL/min via a chilled Superloop (50 mL) onto a Superdex 75 AP5 column (~800 mL column volume). The RNR-AE eluted between 575 mL and 750 mL and the pooled fractions were concentrated (Amicon YM-10 spin filter) and reloaded onto the same column that had been equilibrated in Buffer R. The RNR-AE fractions with the best iron content as illustrated by the A426 : A280 ratio (~0.160 – 0.150) eluted between 700 mL and 720 mL and were pooled and concentrated (on ice). To increase [4Fe-4S] cluster content in RNR-AE, an iron-sulfur cluster reconstitution was performed at 4 ̊C as described elsewhere.2 Briefly, purified RNR-AE (100 μM) was incubated with DTT (5 mM) for 20 minutes in buffer (100 mM Tris, 200 mM NaCl, pH 8.5), before sodium sulfide (0.5 mM final concentration, from a 50 mM stock) and ferric chloride (0.5 mM final concentration, from a 10 mM stock) were added, respectively, over the course of two hours. After a 2.5 hour incubation at 4 ̊C, the reconstitution solution was centrifuged (38,000 x g, 10 min., 4 ̊C), gel filtrated, and concentrated to yield a final iron content of 3.5 ± 0.2 irons/protein. EPR spectroscopy confirmed [4Fe-4S]+ cluster presence in photoreduced enzyme. From E. coli, the RNR gene was isolated using primers 5′-TAC ATA TGA CAC CGC ATG TGA TGA AAC CAG ACG-3′ (forward) and 5′-CTA AGC TTT TAA CCT ATC TGC CCA TTC CCC AAA -3′ (reverse) which imparted restriction enzyme cleavage sites for NdeI and HindIII restriction enzymes. The RNR gene (nrdD from E. coli) cloned into a pCAL-n-EK vector allowed for expression of RNR protein without a tag in BL21(DE3)pLysS E. coli cells. After inoculation from overnight starter cultures, cell culture growth, in LB media (4.5 L in three 2.8 L Fernbach flasks) supplemented with ampicillin (50 μg/mL) and maintained at 37 ̊C with 250 rpm shaking, exhibited good over- expression of RNR upon induction with IPTG (1mM) at an OD600 of 0.5 AU via comparatively high protein concentrations on SDS-PAGE. After induction, cultures were 85 grown with shaking for an additional 3 hours; the cells were centrifuged (12,000 x g, 10 min., 4 ̊C) and the cell pellet (~25 g) was flash frozen in liquid nitrogen and stored at -80 ̊C. As RNR does not contain FeS cluster content, all lysis and purification steps were performed aerobically. At a ratio of ~2 mL of buffer per 1 g of cell pellet, lysis in buffer (20 mM Tris, 1 mM DTT, 10 mM MgCl2, 1% TritonTM X-100, 5% glycerol, pH 7.8) in the presence of PMSF (1 mM), DNase and RNase (0.5 mg each). After centrifugation (38,000 x g, 30 min., 4 ̊C), the supernatant was loaded onto an Accell PlusTM QMA Ion exchange AP5 column (~600 mL) that had been equilibrated with Buffer A (20 mM Tris, 1 mM DTT, pH 7.8) at 3 mL/min. The purification steps included an 300 mL isocratic step with Buffer A, followed by a linear gradient from 0% into 100% Buffer B (20 mM Tris, 500 mM NaCl, 1 mM DTT, pH 7.8) over 900 mL, and ended with a 300 mL isocratic step in 100% Buffer B; all steps were performed at 5 mL/min. Fractions collected between ~1000 and 1500 mL were pooled and concentrated (Amicon YM-80 spin filters) and buffer exchanged into Buffer C (20 mM Tris, 1 mM DTT, 1M (NH4)2(SO4), pH 7.8). The protein solution was briefly centrifuged and, using a superloop (50 mL), the supernatant was loaded at 1 mL/min onto a Phenyl Sepharose HR 16/10 column (~20 mL column volume) equilibrated with Buffer C. This second stage of the RNR purification had three steps: 1) equilibration in Buffer C for 50 mL, 2) linear gradient from 0% to 100% Buffer A (20 mM Tris, 1 mM DTT, pH 7.8) over 50 mL and 3) a final wash step with Buffer A (50 mL); all steps were performed at 1 mL/min. The RNR protein that eluted between ~90 and 105 mL was collected and reapplied to the same Phenyl Sepharose column that had been equilibrated into Buffer C and the identical purification procedure was repeated. The RNR that eluted between ~90 and 105 mL from this final purification was concentrated. As this substrate protein doesn’t coordinate FeS clusters, reconstitution was not needed. If the lysis and purification steps were performed aerobically, the RNR protein was made anaerobic through several Schlenk line nitrogen/vacuum cycles. Spore Photoproduct Lyase (SPL) Preparations SPL preparations were performed as described previously9 with minor modifications. Briefly, the splB gene from Clostridium acetobutylicum, cloned into a pET14b vector allowed for expression of SPL protein with a N-terminal His6 tag in Tuner(DE3)-pLysS E. coli cells. After inoculation from overnight starter cultures, cell culture growth in phosphate buffered LB media (9 L in 2.8 L Fernback flasks) supplemented with chloramphenicol (34 μg/mL) and ampicillin (100 μg/mL) antibiotic and FAS (0.3 mM) and maintained at 37 ̊C with 200 rpm shaking exhibited good over- expression of SPL upon induction with IPTG (1 mM) at an OD600 of 0.8 AU. Grown for 3 more hours before transfer to 4°C, this culture was then sparged overnight (14 - 16 h) with N2(g) until the cell pellet (~35 g) was harvested the next day and stored at -80°C. To isolate SPL with minimal cluster loss, all lysis and purification steps were performed in an anaerobic Coy vinyl glove box. At a ratio of ~2 mL of buffer per 1 g of cell pellet, lysis was performed at 4 ̊C in buffer (20 mM sodium phosphate pH 7.5, 350 86 mM NaCl, 5% glycerol, 10 mM imidazole) in the presence of 1% TritonTM X-100 (w/v), MgCl2 (10 mM), PMSF (1 mM), lysozyme (0.5 mg/cell g), DNase I and RNase A (< 1 mg/cell g). After centrifugation (38,000 x g, 60 min., 4 C̊ ), the SPL clarified lysate was purified via FPLC on two 1 mL in-line HisTrap column with the following step gradient: 1) 0% Buffer B (20 mM sodium phosphate pH 7.5, 350 mM NaCl, 5% glycerol, 500 mM imidizole) wash for 5 CV, 2) 5% Buffer B for 5 CV, 3) 50% Buffer B to elute SPL. After brief centrifugation, the enzyme solution was gel filtrated (Sephadex G-25 resin column, 75 mL) and immediately reconstituted to increase [4Fe-4S] cluster content in SPL; this iron-sulfur cluster reconstitution was performed at 4 C̊ as described elsewhere. Briefly, purified SPL (204 μM) was incubated with DTT (5 mM) for 5 minutes in buffer (20 mM sodium phosphate pH 7.5, 350 mM NaCl, 5% glycerol), before ferrous ammonium sulfate (1.2 mM final concentration, from a 10 mM stock) and sodium sulfide (1.2 mM final concentration, from a 10 mM stock) were added, respectively, over the course of an hour. After a 2-hour incubation at 21 ̊C, the reconstitution solution was centrifuged, desalted (Sephadex G-25 resin column, 75 mL) into buffer (20 mM sodium phosphate pH 7.5, 350 mM NaCl, 5% glycerol), and concentrated (Amicon 30 kDa MWCO centrifugation filters) to yield a final iron content of 3.2 ± 0.2 irons/protein. EPR spectroscopy confirmed [4Fe- 4S]+ cluster presence in the photoreduced enzyme. The R-spore photoproduct (R-SP) was prepared as previously described.9-12 SAM Preparation Unlabeled SAM and the following labeled SAMs ([13C ,15 10 N5-Ado]-SAM, [adenosyl-2,8-D2-1′,2′,3′,4′,5′,5″-D6]-SAM, and [15N-(amino)]-SAM) were synthesized from L-methionine and ATP (labeled as needed to achieve the indicated labeled SAM) and purified as previously described.13-14 Lyophilized SAM was reconstituted in a degassed 100 mM Tris, pH 7.0 – 8.1 buffer and brought up to a pH between 7.0 and 8.0 in an anaerobic (<1 ppm O2) environment. For preparation of 5′-labeled SAMs (including [5′,5″-D2]-SAM), synthesis from methionine and the corresponding labeled ATPs were according to previously published methods;15 specifically 5′-labeled ATP ([5′,5″-D2]-ATP) was synthesized from the corresponding [5′,5″-D2]-ribose based on established procedures.16 For a 10 mL reaction, 8 mL of SAM synthetase prepared as described previously15 was mixed with 1 mL of 130 mM ATP (pH 7.50) 1 mL of 250 mM L-methionine (pH 7.5), and 10 μL of inorganic pyrophosphatase (1 unit/ μL). The reaction mixture was incubated for 1 h at 30°C in a water bath and the progress was monitored using an HPLC. The reaction was quenched by filtration through a 1 kDa MWCO filter. The resulting small molecule fraction was diluted with ten volumes of 0.2 M sodium acetate pH 4.0, and loaded onto an Amberlite CG-50 column (50 mL, Ammonium form). The column was washed with 500 mL of 0.2 M sodium acetate and eluted by a linear gradient (125 x 125 mL)) of 0 - 0.5 M hydrochloric acid. The fractions containing SAM were identified based on HPLC, pooled, and adjusted to pH 6.0 using ammonium hydroxide. The samples were then lyophilized and the residue was 87 dissolved in water. The resulting solution was then loaded to a Biogel P2 column (20 mL) and SAM was eluted using water. Typically, SAM eluted after 2-3 column volumes. The fractions containing SAM were then combined and lyophilized. RFQ Sample Preparation For all RFQ experiments, the same sample preparation was carried out in an anaerobic Coy chamber. Generically, a solution of each radical SAM enzyme in the appropriate buffer was placed in an EPR or NMR tube and photoreduced by illumination with a 500 W halogen lamp for 1 h in an ice water bath; if the radical SAM enzyme substrate was a protein (PFL-AE and RNR-AE) or peptide (OspA and PoyA), solutions of these substrates were simultaneously photoreduced employing the identical method as for the enzyme. If the enzyme was not prepared in a Tris buffer, then some Tris buffer (50 mM, final) was added as a sacrificial electron donor. After photoreduction, and except where specifically stated otherwise, the radical SAM enzyme solution was loaded into one syringe, while SAM (1 - 5 mM) was added to the substrate solution and loaded into the second syringe. RFQ was performed as soon as possible after photoreduction completion. In general, the RFQ sample preparation for each enzyme was nearly identical, however the details for each radical SAM enzyme can be found in the following paragraphs. A HydG stock solution containing 220 μM HydG, 200 μM 5-deazariboflavin, 50 mM Tris, and 5 mM DTT in buffer (50 mM HEPES, 300 mM KCl, 5% glycerol, pH 8.0) was photoreduced in conjunction with a separate substrate mixture containing 2 mM tyrosine, 50 - 200 μM 5-deazariboflavin, and 50 mM Tris in the same buffer (50 mM HEPES, 300 mM KCl, 5% glycerol, pH 8.0). After photoreduction, SAM (5.5 mM) was added to the tyrosine solution, unless otherwise noted, and this set-up yielded a sample with the following concentrations: 110 μM HydG, 1 mM tyrosine, 125 - 200 μM 5- deazariboflavin, 2.5 mM DTT, 2.25 mM SAM. A LAM stock solution containing 1.2 mM LAM, 10 mM cysteine, 1 mM ferrous ammonium sulfate, 15% glycerol, 1.2 mM pyridoxal 5′-phosphate (PLP) was incubated at 37 ̊C for 4 h prior to photoreduction in 42 mM EPPS, pH 8.0 buffer. This stock solution was combined with 200 μM 5-deazariboflavin and photoreduced for one hour. A separate substrate solution containing 40 mM L- lysine, 50 - 200 μM 5-deazariboflavin was photoreduced for at least 15 minutes. After photoreduction, SAM (5.5 mM) was added to the substrate solution, unless otherwise noted. This set-up yielded a sample with the following concentrations: 450 μM LAM, 20 mM L-lysine, 125 - 200 μM 5-deazariboflavin, 2.25 mM SAM in buffer (40 mM EPPS, pH 8.0). A OspD stock solution containing 200 μM OspD, 200 μM 5-deazariboflavin, 50 mM Tris, 10 mM DTT in buffer (50 mM HEPES, 150 mM KCl, 10% glycerol, pH 8.0) was photoreduced in conjunction with a separate OspA mixture containing 300 μM OspA, 100 - 200 μM 5-deazariboflavin, 50 mM Tris, and 10 mM DTT in the same buffer. After photoreduction, SAM (3 mM) was added to the OspA solution, unless otherwise noted, to 88 yield a sample with the following concentrations: 100 μM OspD, 150 μM OspA, 150 - 200 μM 5-deazariboflavin, 50 mM Tris, 10 mM DTT, 1.5 mM SAM. A PFL-AE stock solution containing 550 μM PFL-AE, 200 μM 5-deazariboflavin and 1 mM DTT in buffer (100 mM Tris, 200 mM KCl, pH 7.6) was photoreduced in conjunction with a separate PFL mixture containing 770 μM PFL, 50 - 200 μM 5- deazariboflavin, 10 mM oxamate, and 1 mM DTT in the same buffer (50 mM Tris, 100 mM KCl, pH 7.6). After photoreduction, SAM (5.5 mM) is added to the PFL solution, unless otherwise noted. The protein concentrations used were such as to achieve a ratio of PFL-AE: PFL of 1:1.4 after mixing. This set-up yielded a sample with the following concentrations: 275 μM PFL-AE, 385 μM PFL, 125 - 200 μM 5-deazariboflavin, 5 mM oxamate, 1 mM DTT, 2.25 mM SAM in buffer (100 mM Tris, 200 mM KCl, pH 7.6). A PoyD stock solution containing 200 μM PoyD, 200 μM 5-deazariboflavin, 50 mM Tris, 10 mM DTT in buffer (50 mM HEPES, 150 mM KCl, 10% glycerol, pH 8.0) was photoreduced in conjunction with a separate PoyA25 mixture containing 300 μM PoyA25, 100 - 200 μM 5-deazariboflavin, 50 mM Tris, and 10 mM DTT in the same buffer. After photoreduction, SAM (3 mM) was added to the PoyA25 solution, unless otherwise noted, to yield a sample with the following concentrations: 100 μM PoyD, 150 μM PoyA25, 150 - 200 μM 5-deazariboflavin, 10 mM DTT, 1.5 mM SAM in buffer. A RNR-AE stock solution containing 500 μM RNR-AE, 200 μM 5-deazariboflavin, and 1 mM DTT in buffer (100 mM Tris, 200 mM NaCl, pH 8.5) was photoreduced in conjunction with a separate RNR mixture containing 500 μM RNR, 200 μM 5- deazariboflavin, 10 mM sodium formate, and 1 mM DTT in the same buffer (100 mM Tris, 200 mM NaCl, pH 8.5). After photoreduction, SAM (5.5 mM) was added to the RNR solution, unless otherwise noted, and this set-up yielded a sample with the following concentrations: 250 μM RNR-AE, 250 μM RNR, 200 μM 5-deazariboflavin, 1 mM DTT, 2.25 mM SAM in buffer (100 mM Tris, 200 mM NaCl, pH 8.5). A SPL stock solution containing 550 μM SPL, 200 μM 5-deazariboflavin, and 5 mM DTT in buffer (20 mM sodium phosphate, 350 mM NaCl, 5% glycerol, pH 7.5) was photoreduced. A separate substrate mixture containing 770 μM SP substrate, 10 mM DTT, and SAM (5.5. mM) was prepared in the same buffer. After photoreduction, SAM (5.5 mM) was added to the substrate solution, unless otherwise noted to yield a sample with the following concentrations: 275 μM SPL, 385 μM R-SP, 100 μM 5-deazariboflavin, 5.5 mM DTT, 2.25 mM SAM. Rapid Freeze-Quench Experiments Rapid freeze-quench (RFQ) experiments were performed with a System 100 apparatus from Update Instrument. In general, each RS enzyme, photoreduced in the presence of 5-deazariboflavin to generate the catalytically relevant [4Fe-4S]+ cluster, was loaded into one loop while the corresponding substrate and SAM (substrate+SAM) mixture was loaded into the other loop, with the aim of achieving a post-mixing ratio between 1:20 and 1:1 of RS enzyme: substrate. However, two additional mixing 89 procedures were also examined with PFL-AE: 1) (PFL-AE+SAM) + PFL, and 2) (PFL- AE+PFL) + SAM. In all cases, samples were loaded in an anaerobic chamber and, despite the RFQ instrument location outside of an anaerobic environment, strict anaerobic technique was maintained for all samples. Both enzyme loop and substrate loop tubing(s) were each connected to respective syringes that had been washed several times with a dithionite solution (100 mM, in water) and then dried with N2 gas before loading. Each enzyme or substrate solution was preceded and followed by a small plug of N2 gas and a dithionite solution (100 mM, in water) to create the following set-up: dithionite solution, N2 gas, sample (enzyme or substrate/SAM solution), second N2 gas, second dithionite solution. The N2 gas was introduced in the loops between the dithionite and the sample solutions to prevent dithionite contamination of the samples; in addition to minimizing the likelihood of potential protein oxidation, the dithionite solution helped to ensure all sample solutions exited the loop into the mixing chamber and onto the copper wheels. The mixture was quenched by spraying onto two rotating copper wheels cooled to liquid nitrogen temperatures as previously described17-19 after 500 ms mixing times, though a range of times between 250 ms and 1 sec were tested. The frozen powder was collected in a funnel and packed into precision Q-band tubes (2.5 mm OD) for EPR and ENDOR analysis. EPR and ENDOR Measurements X-band CW EPR spectroscopy was conducted on a Bruker ESP 300 spectrometer equipped with an Oxford Instruments ESR 910 continuous helium flow cryostat. Typical experimental parameters were at 40 K, 9.37 GHz, 1.99 mW microwave power, 100 kHz modulation, and 8 G modulation amplitude. EPR simulation were performed with QPOW.20-21 The 35 GHz CW ENDOR measurements employed 100 kHz field modulation and dispersion mode detection under rapid passage conditions at 2 K. 1H CW ENDOR spectra employed broadening of the RF to 100 kHz to improve signal-to-noise.22 For a single molecular orientation and for nuclei with nuclear spin of I = 1/2 (1H, 15N), the ENDOR transitions for the ms = ±1/2 electron manifolds are observed, to first order, at frequencies, where nn is the nuclear Larmor frequency, and A is the orientation-dependent hyperfine coupling. For I ≥ 1 (14N, I = 1), the two ENDOR lines are further split by the orientation- dependent nuclear quadrupole coupling (P) into 2I lines given by equation: 1 𝐴 I = : 𝜐± = |𝜈𝑛 ± | (1) 2 2 𝐴 3𝑃(2𝑀𝐼)−1 𝐼 ≥ 1: 𝜈± = |𝜈𝑛 ± ± ( )| (2) 2 2 were P is a parameter that characterizes the coupling (see ref,23) 90 EPR of isotopically labeled Ω with PFL-AE 56Fe and 57Fe PFL-AE Fig. S3 (top panel) shows the EPR spectrum of Ω with 57Fe labeled PFL-AE and non-labeled PFL-AE. The EPR spectrum of 57Fe-Ω shows the most distinguishable broadening between g|| and g⫠ in comparison to 56Fe-Ω. The EPR spectrum of 56Fe-Ω is best simulated with g|| = 2.035, g⫠ = 2.004, and EPR line-width [40, 26, 26] MHz for the unresolved hyperfine coupling. The EPR spectrum of 57Fe-Ω is simulated with the same spin Hamiltonian as 56Fe-Ω except with varied 57Fe hyperfine coupling to account for the broadening pattern. As shown in Fig. S3 (bottom panel), a singly coupled 57Fe nuclei with hyperfine coupling A = [30, 35, 30] MHz marginally matches the broadening, while the simulation of two equivalents of A(57Fe) = [30, 35, 30] MHz, or one large coupled A(57Fe) = [30, 35, 30] MHz in addition to two small coupled A(57Fe) ~ [20, 20, 20] MHz closely match the broadening pattern; although the latter two simulations do not distinguish each other. The simulation shows more than one 57Fe nuclei is required to reproduce the broadening pattern, which suggests the spin density resides within a multi-iron cluster, namely [4Fe-4S]3+ of PFL-AE consistent with our previous 57Fe ENDOR analysis. EPR and ENDOR of Ω prepared with 1/2H SAM The loss of 1H ENDOR signals from Ω upon deuteration of SAM is described in the main text. As a parallel measurement, Fig. S4 shows the EPR spectrum of Ω with 1H, [D8-ado]-SAM ([adenosyl-2,8-D2-1′,2′,3′,4′,5′,5″-D6]-SAM), and [5′,5″-D2-ado]-SAM ([adenosyl-5′,5″-D2-SAM]). The EPR spectrum of [D8-ado]-SAM and [5′,5″- D2-ado]- SAM display roughly the same line-width at g|| = 2.035, and both are noticeably narrower than 1H-SAM, which indicates the Ω Fe spin center specifically interact with the 5′-H/D nuclei of SAM. This confirms the presence of the direct Fe-(5′-C) bond as discussed in main text. 91 Supplemental Figures Fig. S3.1 The previously accepted mechanism for RS enzymes, involving reductive cleavage of SAM to generate methionine and the 5′-dAdo•, followed directly by H atom abstraction from substrate. Here, the substrate from which the H atom is abstracted is represented by “R” (red). The [4Fe-4S] cluster is shown as spheres (yellow, sulfur; rust, iron) 92 Fig. S3.2 EPR spectra of Ω as freeze-quenched at 500 ms. Top, pre- mixed PFL and SAM solution, is rapidly mixed and freeze-quenched with PFL-AE solution. Bottom, pre-mixed PFL-AE and PFL solution, is rapidly mixed and freeze-quenched with SAM solution. EPR conditions: microwave frequency, 9.374 GHz (top), 9.375 GHz (bottom); modulation amplitude, 10 G; T = 40 K. 93 Fig. S3.3 Normalized EPR spectra of Ω with all the radical SAM enzymes examined. Left, as freeze- quenched at 500 ms; Right, after annealing 1 min at 150 K. RFQ mixing condition: pre-mixed substrate (PFL, tyrosine, PoyA, RNR, OspA, α-lysine, and R-SP) and SAM solution, is rapidly mixed and freeze- quenched with radical SAM enzyme solution (PFL-AE, HydG, PoyD, RNR-AE, OspD, LAM, and SPL, respectively). In addition to the Ω signal, variable amounts of a second, free-radical-like signal, which is lost during brief (1 min) annealing at ~150 K, can be seen in each of these enzymes; the origin of this second signal has yet to be identified. As noted in main text, HydG, and to a lesser extent LAM, spectra contain a feature from Cu(II) introduced by Cu freezing wheels. EPR conditions: microwave frequency, 9.375 GHz; modulation amplitude, 10 G; T = 40 K. 94 Fig S3.4. Normalized representative EPR spectra of [4Fe-4S]+ and ([4Fe-4S]++SAM) cluster for PFL-AE in comparison with spectra of freeze-quenched samples, which show that the formation of Ω accompanies the complete loss of the cluster signal for each of the enzymes studied: PFL-AE, HydG, PoyD, RNR-AE, OspD, LAM, and SPL. Note, the ‘derivative- shaped’ feature to low field of the Ω signal, ~3250 G, is from Cu contaminant derived from the Cu wheels on which the RFQ samples are frozen. EPR conditions: microwave frequency, 9.375 GHz; modulation amplitude, 10 G; T = 12 K. 95 Fig. S3.5 EPR spectrum of product Gly radical formed upon annealing RNR-AE Ω as indicated, overlaid with spectrum of hand-quenched enzyme with radical. The Gly radical in RNR-AE has a considerably longer T1/slower relaxation than that in PFL-AE, and as a result the hyperfine-split doublet in both the annealed RNR-AE RFQ sample and hand-quenched sample are highly saturated at any usable power at 40 K, unlike the reported spectra for PFL-AE;4 at 77 K, where the unsaturated signal with resolved hyperfine doublet from the hand-quenched sample is seen, the signal from the RFQ/annealed sample is too weak to study. As in ref 4, residual signal from Ω has been subtracted. It was not possible to determine whether the extremely sharp signal with low integrated intensity, which overlays the glycyl radical signal, is present before annealing or forms as a minority byproduct during annealing. EPR conditions: microwave frequency, 9.375 GHz; power, 2 mW; modulation amplitude, 10 G. 96 Fig. S3.6 Top, the EPR spectra of Ω formed with 57Fe labeled PFL- AE (red); 56Fe PFL-AE (black) and simulation (black dash), g|| = 2.035, g⫠ = 2.004, LW|| = 40 MHz, LW⫠ = 26 MHz. Bottom, spectrum of Ω -57Fe) (red) and simulations with varied 57Fe hyperfine couplings as indicated. RFQ mixing condition: pre-mixed PFL with SAM solution is rapidly mixed with PFL-AE solution and freeze- quenched. EPR conditions: microwave frequency, 9.375 GHz; modulation amplitude, 10 G; T = 12 K. As noted in main text, this figure and the simulation confirm that the spin of W is localized on the Fe-S cluster. In response to a reviewer question, we note that no attempt is made to precisely fit the signal, which likely contains a slight distribution in parameters originating in a slight spread in geometries. The origin of the small, variable signal primarily to high field of W, which is not seen in any of the other samples of any of the enzymes (eg. Fig 3) is not known. 97 Fig. S3.7 X-band EPR spectra of Ω for PFL-AE/PFL with 1H-SAM (black), [D8-ado]- SAM ([adenosyl-2,8-D2-1′,2′,3′,4′,5′,5″-D6]-SAM, blue), and [5′,5″-D2-ado]-SAM ([adenosyl-5′,5″-D2-SAM], red), showing the distinct narrowing of the g|| feature upon uniform and specific deuteration. The apparent shift of that feature upon deuteration is a consequence of different microwave frequencies. RFQ mixing condition: pre-mixed PFL with SAM solution is rapidly mixed with PFL-AE solution and freeze-quenched. EPR conditions: microwave frequency 9.373 GHz (1H SAM), 9.375 GHz ([D8-ado]-SAM), 9.375 GHz ([5′,5″-D2-ado]-SAM); modulation amplitude, 10 G; T = 40K. 98 Fig. S3.8. 35 GHz CW 15N/14N ENDOR of Ω. 15N EDNOR shows A(15N) ~ 5.6 MHz corresponding to the hyperfine coupling of the “anchor” amine group of adenosyl methionine. Based on the nuclear g factor ratio g (14N)/g (15N) = An(14N)/An(15 n n N) = |0.71|, A(14N) ~ 4 MHz is calculated which is further split by quadrupole coupling, denoted 3P (see equation 2), as marked by the goalpost. RFQ mixing condition: pre-mixed substrates, PFL and SAM, solution, is rapidly mixed and freezequenched with PFL-AE solution. ENDOR conditions: microwave frequency, 35.042 GHz; scan rate, 0.5 MHz/s; scan direction, reverse; T = 2 K. 99 Table S3.1. Overview of radical SAM enzymes examined in this study. Iron content is reported as irons/monomer. 100 Supplementary References 1. Duffus, B. R.; Ghose, S.; Peters, J. W.; Broderick, J. B. J. Am. Chem. Soc. 2014, 136, 13086- 13089. 2. Byer, A. S.; McDaniel, E. C.; Impano, S.; Broderick, W. E.; Broderick, J. B. Methods Enzymol. 2018, in press. 3. Morinaka, B. I.; Vagstad, A. L.; Helf, M. J.; Gugger, M.; Kegler, C.; Freeman, M. F.; Bode, H. B.; Piel, J. Angew. Chem. Int. Ed. 2014, 53, 8503-8507. 4. Horitani, M.; Shisler, K. A.; Broderick, W. E.; Hutcheson, R. U.; Duschene, K. S.; Marts, A. R.; Hoffman, B. M.; Broderick, J. B. Science 2016, 352, 822-825. 5. Broderick, J. B.; Henshaw, T. F.; Cheek, J.; Wojtuszewski, K.; Smith, S. R.; Trojan, M. R.; McGhan, R. M.; Kopf, A.; Kibbey, M.; Broderick, W. E. Biochem. Biophys. Res. Comm. 2000, 269, 451-456. 6. Nnyepi, M. R.; Peng, Y.; Broderick, J. B. Arch. Biochem. Biophys. 2007, 459, 1-9. 7. Freeman, M. F.; Gurgui, C.; Helf, M. J.; Morinaka, B. I.; Uria, A. R.; Oldham, N. J.; Sahl, H.-G.; Matsunaga, S.; Piel, J. Science 2012, 338, 387-390. 8. Tropea, J. E.; Cherry, S.; Waugh, D. S. Methods Mol. Biol. 2009, 498, 297-307. 9. Silver, S. C.; Chandra, T.; Zilinskas, E.; Ghose, S.; Broderick, W. E.; Broderick, J. B. J. Biol. Inorg. Chem. 2010, 15, 943-955. 10. Chandra, T.; Broderick, W. E.; Broderick, J. B. Nucleosides, Nucleotides, & Nucleic Acids 2009, 28, 1016-1029. 11. Chandra, T.; Silver, S. C.; Zilinskas, E.; Shepard, E. M.; Broderick, W. E.; Broderick, J. B. J. Am. Chem. Soc. 2009, 131, 2420-2421. 12. Chandra, T.; Broderick, W. E.; Broderick, J. B. Nucleosides, Nucleotides, & Nucleic Acids 2010, 29, 132-143. 13. Walsby, C. J.; Hong, W.; Broderick, W. E.; Cheek, J.; Ortillo, D.; Broderick, J. B.; Hoffman, B. M. J. Am. Chem. Soc. 2002, 124, 3143-3151. 14. Walsby, C. J.; Ortillo, D.; Broderick, W. E.; Broderick, J. B.; Hoffman, B. M. J. Am. Chem. Soc. 2002, 124, 11270-11271. 15. Iwig, D. F.; Booker, S. J. Biochemistry 2004, 43, 13496-13509. 16. Scott, L. G.; Geierstanger, B. H.; Williamson, J. R.; Hennig, M. J. Am. Chem. Soc. 2004, 126, 11776-11777. 101 17. Kim, S. H.; Perera, R.; Hager, L. P.; Dawson, J. H.; Hoffman, B. M. J. Am. Chem. Soc. 2006, 128, 5598-5599. 18. Lin, Y.; Gerfen, G. J.; Rousseau, D. L.; Yeh, S.-R. Analytical Chemistry 2003, 75, 5381-5386. 19. Aitha, M.; Moller, A. J.; Sahu, I. D.; Horitani, M.; Tierney, D. L.; Crowder, M. W. J. Inorg. Biochem. 2016, 156, 35-39. 20. Nilges, M. J. Electron Paramagnetic Resonance Studies of Low Symmetry Nickel(i) and Molybdenum(v) Complexes. Part I: Computer Simulation of Electron Paramagnetic Resonance Spectra. Part II: Electron Paramagnetic Resonance Studies of Nickel(i) Triphenylphosphine Complexes. Part Iii: Electron Paramagnetic Resonance Studies of a Molybdenum(v) Thiocyanate Complex. Ph.D. Thesis, University of Illinois at Urbana-Champaign, Urbana-Champaign, IL, 1979. 21. Maurice, A. M. Acquisition of Anisotropic Information by Computational Analysis of Isotropic EPR Spectra. Ph.D. Thesis, University of Illinois at Urbana- Champaign, Urbana-Champaign, IL, 1982. 22. Hoffman, B. M.; DeRose, V. J.; Ong, J. L.; Davoust, C. E. J. Magn. Reson. 1994, 110, 52-57. 23. DeRose, V. J.; Hoffman, B. M. Methods Enzymol. 1995, 246, 554-589. 102 CHAPTER FOUR THE ELUSIVE 5′-DEOXYADENOSYL RADICAL: CAPTURED AND CHARACTERIZED BY EPR AND ENDOR SPECTROSCOPIES Contribution of Authors and Co-Authors Manuscript in Chapter 4 Author: Hao Yang Contributions: Ran RFQ quenches, EPR, ENDOR, and photolysis at Northwestern University, assisted in interpreting EPR and ENDOR data, and generated figures for manuscript. Co-Author: Elizabeth C. McDaniel Contributions: Grew and purified PFL-AE, made RFQ and hand quench samples at Northwestern University and MSU, synthesized labeled and unlabeled SAM, Co-Author: Stella Impano Contributions: Helped prepare hand quench samples and RFQ samples at Northwestern University, synthesized unlabeled SAM. Co-Author: Amanda S. Byer Contributions: Synthesized labeled SAM. Co-Author: Richard J. Jodts Contributions: Assisted with simulations. Co-Author: Kenichi Yokoyama Contributions: Provided insight into the structure of 5′-dAdo radical. 103 Contribution of Authors and Co-Authors Continued Co-Author: William E. Broderick Contributions: Provided insight into the structure of the 5′-dAdo radical, interpretation of EPR data, and assisted with manuscript preparation. Co-Author: Joan B. Broderick Contributions: Assisted with manuscript preparation and interpretation of EPR and ENDOR data. Co-Author: Brian M. Hoffman Contribution: Assisted with manuscript preparation and interpretation of EPR and ENDOR data. 104 Manuscript Information Page Hao Yang, Elizabeth C. McDaniel, Stella Impano, Amanda S. Byer, Richard J. Jodts, Kenichi Yokoyama, William E. Broderick, Joan B. Broderick, Brian M. Hoffman Journal of the American Chemical Society Status of Manuscript: _____ Prepared for submission to a peer-reviewed journal _____ Officially submitted to a peer-review journal _____ Accepted by a peer-reviewed journal __X__ Published in a peer-reviewed journal Publisher ACS Publications Issue #141 DOI 10.1021/jacs.9b05926 105 THE ELUSIVE 5′-DEOXYADENOSYL RADICAL: CAPTURED AND CHARACTERIZED BY EPR AND ENDOR SPECTROSCOPIES Abstract The 5′-deoxyadenosyl radical (5′-dAdo•) abstracts a substrate H atom as the first step in radical-based transformations catalyzed by adenosylcobalamin- dependent and radical S-adenosyl-L-methionine (RS) enzymes. Notwithstanding its central biological role, 5′-dAdo• has eluded characterization despite efforts spanning more than a half- century. Here, we report generation of 5′-dAdo• in a RS enzyme active site at 12 K using a novel approach involving cryogenic photoinduced electron transfer from the [4Fe−4S]+ cluster to the coordinated S-adenosylmethionine (SAM) to induce homolytic S−C5′ bond cleavage. We unequivocally reveal the structure of this long-sought radical species through the use of electron paramagnetic resonance (EPR) and electron nuclear double resonance (ENDOR) spectroscopies with isotopic labeling, complemented by density-functional computations: a planar C5′ (2pπ) radical (∼70% spin occupancy); the C5′(H)2 plane is rotated by ∼37° (experiment)/39° (DFT) relative to the C5′−C4′−(C4′−H) plane, placing a C5′−H antiperiplanar to the ribose-ring oxygen, which helps stabilize the radical against elimination of the 4′−H. The agreement between φ from experiment and in vacuo DFT indicates that the conformation is intrinsic to 5-dAdo• environment. 106 Introduction The discovery by Barker et al. in 1958 of a cobalamin involved in catalyzing an isomerization reaction,1 and its subsequent identification as adenosylcobalamin (coenzyme B12 (Figure 1)) by Lenhert and Hodgkin in 1961,2 set the stage for decades of study focused on the biochemical reactions of this remarkable organometallic cofactor. We now know that coenzyme B12 serves as a reversible radical generator, undergoing Co−C bond homolysis to give a 5′-deoxyadenosyl radical (5′-dAdo·) that initiates chemistry by H atom abstraction from a substrate.3 The involvement of 5′-dAdo· has been shown primarily through label transfer studies, for example, by observing the incorporation of substrate deuterons into adenosylcobalamin during catalysis.4−6 Important insights into its behavior have been attained through kinetic and spectroscopic investigations in which the Co−C5′ bond of B12 is photolytically cleaved, after which the 5′-dAdo· radical can either reform that bond or proceed to react with substrate.7,8 However, strenuous efforts over many years9 to trap and fully characterize the 5′-dAdo· formed from B12 have been unavailing: the radical has simply been too reactive to trap. The significance of the 5′-dAdo· radical in biology with the recognition that S- adenosyl- L-methionine (SAM) plays a role analogous to that of coenzyme B12 in initiating radical reactions.10,11 This role for SAM was first suggested for lysine 2,3-aminomutase12,13 and pyruvate formate-lyase activating enzyme (PFL-AE),14,15 but it is now implicated throughout the vast and diverse radical SAM (RS) superfamily.16−19 Until recently, the accepted mechanism for the RS enzymes involved SAM coordination to the unique iron of 107 the [4Fe−4S] cluster in the active site; electron transfer from the reduced [4Fe−4S]+ cluster to SAM promotes reductive cleavage of SAM to give methionine and 5′-dAdo·, which initiates chemistry via substrate H atom abstraction.16,20−22 The recent discovery of the catalytically competent organometallic intermediate Ω formed during the RS reaction provides a new twist on the mechanism by implicating a coenzyme B12-like organometallic intermediate with an Fe−C5′ bond that undergoes homolytic bond cleavage to release 5′- dAdo·.20,23,24 Insights into the characteristics of the 5′-dAdo· radical intermediate in RS enzymes have been achieved via the SAM analog S-3′,4′-anhydroadenosyl-L-methionine (anSAM), which generates an allylically stabilized 5′-dAdo· analog that is amenable to characterization,25,26 but the 5′-dAdo· radical is so extraordinarily reactive that so far attempts to generate and characterize this radical itself have not been productive. For example, photolysis of coenzyme B12 in frozen solution appears to produce 5′-dAdo·, but spin−spin interactions between the Co(II; S = 1/2) corrinoid and the S = 1/2 radical result in a broad, poorly defined electron paramagnetic resonance (EPR) signal that is not amenable to analysis.27 A sharp fluid-solution FT-EPR spectrum acquired upon photolysis of B12 and thought to arise from 5′-dAdo· is instead the result of rapid radical rearrangement rather than 5′-dAdo· itself (see SI Text and Figure S2).28 An organic radical that accumulates during the reaction of the RS enzyme spore photoproduct lyase with SAM at 40 °C also has been proposed to be 5′-dAdo·,29 but we show in the Supporting Information that this radical too is not 5′-dAdo· (see SI Text, Figure S3, and Table S1). 108 Here, after more than half a century of efforts by numerous investigators to trap and characterize the biologically central 5′-dAdo· radical, we demonstrate the formation of 5′- dAdo· in a RS enzyme active site by using a novel approach involving photoinitiated Figure 4.1. Adenosylcobalamin/coenzyme B12 (left) and S-adenosylmethionine bound to a [4Fe−4S] cluster in the active site of RS enzymes (modeled at right) both serve as precursors to the 5′-dAdo· radical intermediate (center). electron transfer from the reduced [4Fe−4S]+ cluster to the coordinated SAM. The resulting 5′-dAdo· is definitively identified through the use of isotopically labeled SAM combined with EPR and electron nuclear double resonance (ENDOR) spectroscopy, with its structure analyzed using density functional theory (DFT) computation. Experimental Methods Materials 109 [Methyl-13C]-L-methionine, 15N-L-methionine, and 2,8-D2-1′,2′,3′,4′,5′,5″-D6- adenosine 5′-triphosphate salt solution were purchased from Cambridge Isotope Laboratories, Inc. [13C 15 10, N5]-Adenosine 5′-triphosphate sodium salt solution was purchased from Sigma, and 3,3,4,4-L-methionine-d4 was obtained from CDN Isotopes. Protein and SAM Preparation PFL-AE, unlabeled SAM, and labeled SAMs were prepared as previously reported.30 Photolysis. PFL-AE was prepared for photolysis by reducing the enzyme and adding labeled or unlabeled SAM. Photoreduced PFL-AE was prepared as previously described30 with minor alterations. PFL- AE (0.55 mM), dithiothreitol (DTT, 1.0 mM), and deazariboflavin (200 μM, dissolved in DMSO) were combined in buffer (50 mM Tris, 100 mM KCl, pH 7.5) in an anaerobic COY chamber. The sample was irradiated by a 500 W halogen lamp for 1 h while the ice water bath. Labeled or unlabeled SAM (5.5 mM) was then added. The samples were then transferred to EPR tubes and frozen in liquid nitrogen inside the COY chamber. Dithionite reductions were performed similarly as follows: PFL-AE (0.55 mM), DTT (1.0 mM), and sodium dithionite (3.0 mM) were combined in an anaerobic COY chamber and allowed to sit 3 min before SAM (5.5 mM) was added. The samples were flash frozen as described above and stored in liquid nitrogen. Photolysis Photolysis was carried out in situ using a 450 nm Thorlabs diode laser. The time course of 12 K photolysis of (PFL-AE1++SAM) complex was monitored during intracavity 110 photolysis in an X-band EPR spectrometer. The in situ photolysis of the (PFL-AE1++SAM) complex prereduced to the [4Fe−4S]+ state using either deazariboflavin or dithionite gives identical EPR spectra, Figure S1. The in situ photolyzed samples were subjected to X- and Q-band CW EPR and Q-band ENDOR measurements as described below. EPR and ENDOR Measurements X-band continuous wave (CW) EPR spectroscopy was conducted on a Bruker ESP 300 spectrometer equipped with an Oxford Instruments ESR 910, while Q-band CW EPR spectroscopy was conducted on a Bruker EMX spectrometer equipped with an Oxford Instruments Mercury iTC continuous helium flow cryostat. Typical experimental parameters were at 12 and 40 K, 9.38 or 34.0 GHz, and 10 G modulation amplitude. EPR simulations were performed with the EasySpin5.2.23 program operating in Matlab.31 35 GHz CW and pulse ENDOR spectroscopic data were collected on spectrometers, described previously,32−34 that are equipped with liquid helium immersion dewars for measurements at 2 K. The CW measurements employed 100 kHz field modulation and dispersion- mode detection under rapid passage conditions. 1H CW ENDOR spectra employed bandwidth broadening of the RF to 100 kHz to improve signal- to-noise.35 1H CW ENDOR spectra were collected using the stochastic-field modulation detected ENDOR sequence36 to improve ENDOR line shape. For a single molecular orientation and for nuclei with a nuclear spin of I = 1/2 (13C), the ENDOR transitions for the ms = ±1/2 electron manifolds are observed to be first order as shown in the equation: Α 𝜈± = |𝜐𝑛 ± | (1) 2 111 where υn is the nuclear Larmor frequency and A is the orientation-dependent hyperfine coupling. When υn > A/2, the pattern is a doublet split by A and centered at υn; when υn < A/2, the pattern is a doublet centered at A/2 and split by 2υn. DFT Calculations All DFT computations of 5′-dAdo· were performed in ORCA 4.0.1.37 Initial coordinates for the 5′-dAdo fragment were taken from two structures: PFL-AE (2.77 Å resolution) with SAM bound to the [4Fe−4S] cluster (PDB 3CB8)38 and a 1.15 Å resolution crystal structure of a SAM-dependent methyltransferase (PDB 3DLC). All protein residues, water, and other molecules were removed, and the adenine was replaced with a methyl group. Hydrogens were added appropriately to the structure with the unbiased starting geometry of the 5′-CH2 hydrogens being in a pseudotetrahedral conformation. Both geometry optimizations and single-point calculations used the spin unrestricted B3LYP39−41 hybrid functional and the Ahlrichs’ valence triple-ξ with a polarization function basis set.42 The molecular orbitals were visualized as Gaussian cubes and an isosurface of 0.08 au in Pymol. Hyperfine and g tensors were calculated by the coupled−perturbed self-consistent field (SCF) approach as implemented in ORCA 4.0 using the B3LYP hybrid functional and EPR-III basis43,44 in combination with the accurate spin−orbit coupling operator [RI-SOMF(1X)].45 Calculations using the BP8639 functional were carried out in parallel. The optimized structures and all energetic and spectroscopic output parameters for 5′-dAdo· are identical for the two starting points. The results presented are those from 112 the high-resolution structure. Similar results are obtained for the two functionals, as shown in Table S2. Results Photolysis of PFL-AE/[4Fe−4S]+/SAM The present study was indirectly prompted by the idea that the Fe−C5′ bond of the Ω23,24 intermediate in RS enzymes might be cleaved photolytically, similar to the Co−C5′ bond of coenzyme B12. If this were to occur, then photoinitiated homolysis of Ω would produce the diamagnetic [4Fe−4S]2+ cluster and the sought-after 5′-dAdo· radical. However, because sulfoniums have been reported to be photochemically reactive,46,47 we first carried out control experiments in which SAM and PFL-AE/[4Fe−4S]1+ were combined and photolyzed in the absence of substrate PFL, circumstances in which SAM is not enzymatically cleaved and Ω does not form. Much to our surprise, irradiation of such samples in the EPR cavity at 450 nm and 12 K results in rapid conversion of the SAM- bound [4Fe−4S]+ state of PFL-AE (Figure 2A) to one with a new free-radical signal that is partially saturated and thus poorly resolved at 12 K, Figure 2B, but becomes well-resolved at 40 K (Figure 3A). A time course of the photolysis shows the loss of the [4Fe−4S]1+ signal directly correlates with the appearance of the new radical signal (Figure 2C). The data is consistent with photoinduced electron transfer (ET) from the [4Fe−4S]+ cluster to SAM,48 providing the EPR-silent [4Fe−4S]2+ cluster and a SAM-derived organic cryo-trapped radical to study; such photoinduced ET has not previously been reported for RS enzymes. On the basis of known RS enzyme chemistry, the most likely identity of the 113 Figure 4.2. X-band EPR spectra of the ([4Fe−4S]++SAM) PFL-AE complex (A) before and (B) after photolysis at 12 K with 450 nm LED for 1 h to produce the 5′-dAdo· radical with near-complete loss of the initial complex signal. Conversion is quantitative; residual cluster signal in spectrum B is from the enzyme out of the laser beam. (C) Time course for decay of ([4Fe−4S]++SAM) (■) upon photolysis monitored at 3600 G and increase of 5′-dAdo· (●) monitored at 3360 G. The two progress curves are fit to stretchedexponential (as the result of light scattering within the “snow” samples) decay, I = exp(−[t/τ]n), and rise, I = 1 − exp(−[t/τ]n), functions with the same parameters, τ = 11(1) min and n = 0.43(2). EPR conditions: microwave frequency, 9.38 GHz; modulation, 10 G; T, 12 K. new radical species would be the 5′-dAdo· radical, resulting from reductive cleavage of SAM. However, other forms for the trapped radical are possible.49 Rearrangement to form a 4′ radical or even a cyclo-adenosine radical must be considered, and if the highly reactive 5′-dAdo· is formed, it might have attacked a nearby protein residue to generate a protein radical, even at 12 K. We therefore pursued detailed studies that unequivocally identify the structure of this radical species. The radical as trapped in the active site of PFL-AE is stable at 77 K and below but is lost upon annealing for 1 min at 150 K. Figure 3A shows that the EPR spectrum of the radical species under nonsaturating conditions (40 K) can be well simulated by incorporating 1H hyperfine interactions appropriate for trapped 5′-dAdo·, Figure 3A. The 114 simulation31 includes anisotropic hyperfine couplings to two near-equivalent α-type 1H on 5′-C plus coupling to one near-isotropic β-type 1H on 4′-C; the hyperfine tensors are reported in Table 1. The two resulting α-type 1H hyperfine tensor components, confirmed and refined as described below, have values roughly in the ratio of ∼ 1/2/3, as expected for a trigonal-planar, Cα−1H carbon radical, and indeed quantitatively agree quite well with those of the Cα−1H radical of ·CH(COOH)2 formed by irradiation of malonic acid (Table 1). In addition, it is well- established50−52 that a β-type 1H, such as the 1H−4′-C of 5′- dAdo·, gives a large, near-isotropic coupling, as observed here (Table 1). The Q-band EPR spectrum (Figure 3F) is equally well simulated with these parameters and has the additional benefit of enhancing the influence of the small free-radical g- anisotropy, which helped refine both the g and hyperfine tensors (Table 1). Of particular note as an “internal check”, the intermediate hyperfine component of an α-proton must lie parallel to the 2pπ orbital, which is in turn parallel to g ≈ g 50−54 3 and, thus, must be parallel for the two α-protons of 5′- dAdo·. This was required for the optimized simulations of the radical spectra, particularly the Q-band spectrum (Table 1), even though the tensors were determined without imposing this as a prior constraint. Confirmation of Radical Species Using SAM Isotopologs To confirm this new radical species as the 5′-dAdo· radical and to refine its 1H hyperfine interaction parameters, we carried out photolysis experiments utilizing isotopically labeled SAMs.30 Assignment of the α-type couplings to the two 5′-Cα protons is confirmed by the collapse of both the X-band (Figure 3B) and Q-band (Figure S5) spectra 115 of the radical generated with [adenosyl-5′,5″-D2]-SAM to a doublet resulting from the near- isotropic splitting from the 4′-Cβ proton. This splitting in turn is lost in both the X-band and Q-band spectra of the radical generated with the perdeuterated-Ado SAM isotopologue 116 ([adenosyl-2,8-D2-1′,2′,3′,4′,5′,5″-D6]-SAM) (Figures 3C and S5). Together, these measurements show that the radical formed by photolysis is indeed the long-sought 5′- dAdo·. Figure 4.3. A−E, X-band EPR spectra (black) and simulations (red) of 5′-dAdo·. 5′-dAdo· generated with (A) natural abundant SAM; (B) [adenosyl-5′,5″-D2]-SAM; (C) [adenosyl-2,8-D2-1′,2′,3′,4′,5′,5″- D6]-SAM; (D) 5′-dAdo· with [adenosyl-13C10, 15N5]-SAM; E, 5′-dAdo· with [adenosyl-5′- 13C]- SAM. Features associated with the minor photolysis products are revealed in spectra of isotopologues (B, C) and indicated by (*). F, Q-band EPR, 5′-dAdo· generated with natural abundant SAM; the additional line width at low field is attributable to g/A strain. Simulations: Generated with EasySpin31 using a 5′-dAdo· model with parameters listed in Table 1. EPR conditions: microwave frequency, 9.38 GHz (A−C and G) and 34 GHz (D−F); modulation, 10 G; T, 40 K. Simulations: generated with Easyspin31 using a 5′-dAdo· model with parameters listed in Table 1. G, Q-band CW 13C ENDOR of 5′-dAdo·. From 11 to 15 MHz, generated with [adenosyl-13C10, 15N5]-SAM where (▼) represents 13C Larmor frequency and the “goalpost” connecting the doublet from 1′- 13C and/or 2′- 13C, split by A = 0.8 MHz (see eq 1). From 27 to 170 MHz, generated with [adenosyl-13C10, 15N5]-SAM where (●) represents A/2 and only the high-frequency members of the doublets for 3′- 13C and 5′- 13C are seen, as indicated, separated from their respective A/2 by 13C Larmor frequency. Spectrum between 70 and 149 MHz also seen when generated from [adenosyl-5′- 13C]-SAM. It is overlaid on the signal from 1 H of C4′ and 5′ in the spectrum with natural-abundance SAM (gray line). CW ENDOR conditions: microwave frequency, 34.8 GHz; modulation, 2 G; T, 2 K. 117 This identification is further confirmed, and the characterization of 5′-dAdo· is strengthened and enriched by the examination of the radical formed with 13C isotopologues. 3C 15 10, N5]-SAM (Figure 3D), the radical X-band EPR spectrum shows additional splittings from two 13C, one with the anisotropic hyperfine tensor expected50−52 for the 5′-13Cα “sp2” carbon and one with the smaller, nearly isotropic coupling expected for the 4′-13Cβ carbon (Table 1). The assignment of the two couplings to their respective carbons is confirmed by the EPR spectrum of radical prepared from singly labeled 5′-13C SAM (Figure 3E), which shows the splittings with the anisotropic tensor assigned to 5′-13Cα (Table 1). Note that the maximum 13C coupling for a spin in a carbon 2pπ orbital must also lie along g = g ,50−52 3 e as found experimentally for 5′-13C, without prior constraint (Table 1). The Q-band ENDOR spectra for 5′-13C-SAM, Figure 3G, shows a peak centered at a frequency corresponding to approximately A3/2 of 5′-13C (Table 1). The Q-band ENDOR spectra from the radical prepared from [adenosyl-13C 15 10, N5]- SAM (Figure 3G) show not only the 5′-13C signal but also a signal from the near-isotropically coupled 3′-13C centered at a frequency determined by essentially half its isotropic coupling (Table 1), as well as a doublet with a much weaker coupling (A = 0.8 MHz) assigned to 1′-13C and/or 2′-13C; the signal from 4′-13C is hidden under the 1H ENDOR response from weakly coupled protons (Figure S6). These 13C measurements unambiguously confirm the photochemically generated radical is indeed 5′-dAdo·. Structure of 5′-dAdo· 118 Beyond the identification of 5′-dAdo·, the observed hyperfine interactions teach us about its electronic and geometric structure. The two 1Hα and the 5′-13Cα coupling tensors are in good agreement with those for the odd electron in the 2pπ orbital of the planar (sp2) carbon-centered radical of X irradiated malonic acid (Table 1).54 The match, notably in the isotropic couplings, indicates that the spin density in the 2pπ orbital on C5′ of 5′-dAdo· (ρπ) is comparable to that in malonic acid, ρπ ∼ 0.7. Of particular note, the spin- polarization-induced isotropic coupling of the 2pπ spin of planar 13C5′ would have been sharply increased by a pseudotetrahedral “doming” distortion at C5′, which would introduce 2s character, with its large isotropic coupling,51 into the odd-electron orbital. Thus, we infer that the H2−C5′−C4′ fragment is essentially planar. The isotropic coupling of a proton β to a carbon 2pπ electron spin, such as 1H−C4′, is known to obey the relationship, aiso ≈ ρπBcos2φ MHz, where B reflects the transmission of spin to 1H through hyper- conjugation, ρπB ≈ 140 MHz, and φ is the dihedral angle between the 2pπ orbital and the Cβ−H bond,51 corresponding to the angle between the [4′CH−4′C−5′C] plane and of the 2pπ orbital (normal to the plane formed by C5′ and its two H atoms). Applying this relationship to the measured 1H−C4′ isotropic coupling (Table 1) yields a value of φ ≈ 37°. These conclusions from the experimental finding are confirmed and extended by DFT computations37 for 5′- dAdo· (Table S2 and Figure S4). First, the computations yield an energy-minimized structure that reproduces the observed hyperfine couplings extremely well, Table 1, with a value for the spin density in the C5′ 2pπ orbital of ρπ = 0.7, in agreement with the experimental analysis. The computations further generate the radical’s structure (Figure 4, top), which exhibits both a rigorously planar geometry at C5′ and a 119 dihedral “twist” at the C5′−C4′ bond, as inferred above. The dihedral (twist) angle in the Table 4.1. Hyperfine Tensors (MHz) of 5′-dAdo· from the Experimenta Plus DFT-Computed Values in Parentheses,b with Reference Values for Radicals Generated from Malonic Acidc aThe signs (±) of experimental 5′-dAdo· tensors are assigned according to ref 52; the hyperfine signs for 1′- and/or 2′- 13C, 3′- 13C, and 4′- 13C are not determined. The simulations employ g = [2.0075, 2.0015, 2.000] (gave = 2.003) and tabulated hyperfine couplings. The A3 components for 1 Hαa and 1 Hαb are parallel to g3 and normal to the C5′H2 plane (see Figure 3) as seen in refs 54 and 55. For 1 Hαa, A1||g1; simulation required the 1 Hb tensor be rotated relative to that of 1 Hαa by an angle approaching α ∼120° expected for sp2 hybridized C5′ and found by DFT optimization, but because of the small g-anisotropy, this angle could be allowed to vary over the range 90° ≲ α ≲ 120°. A3 for 4′-C−1 Hβ makes an angle of ∼40° with respect to g3. The EPR simulations tightly constrain the values of the tensor components; for strongly coupled nuclei (all but 1′- and/or 2′- 13C), numerous simulations indicate that one tensor component value may vary by as much as ±5 MHz, but the sum of multiple variations must also be no more than roughly ±5 MHz. Additional EasySpin simulation parameters: EPR line width, 36 MHz (X- band); 45 MHz (Q-band). Hyperfine strain parameters employed to account for unresolved hyperfine coupling: [30, 10, 10] MHz for A−C, E, and F; [30, 10, 40] MHz for D. The simulation for D is the sum of 80% 13C 5′-dAdo· and 20% 12C, and for E, it is the sum of 85% 13C 5′-dAdo· and 15% 12C. b The hyperfine tensors in parentheses are from the DFT calculation with B3LYP/G (Supporting Information). The energy-minimized structure (Figure 4) has equivalent C5′−H bonds and 1 H hyperfine tensors, whereas the experiment gives slightly different tensors for the two. We attribute the observed inequivalence to a slight desymmetrization due to interactions in the enzyme active sit e that are absent in the computation. c1 H and 13C tensors for ·CH(COOH)2 formed by irradiation of malonic acid, taken from McConnell et al.55 and Cole and Heller,56 respectively. 1 H tensors for ·CH2(COOH) (g = [2.0042, 2.0034, 2.0020], gave = 2.0032) formed by irradiation of malonic acid; see ref 54. The somewhat greater inequivalence between the hyperfine tensors for two α-type 1 H of 5′-dAdo· and · CH2(COOH) likely reflects the influence of contacts with the surrounding residues, as does the observation that the two radicals have the same gave yet differ in their in-plane g-values, g1 and g2. 120 energy-minimized geometry φ ≈ 39° t indicated by the experimental analysis above and creates a structure with a C5′−H antiperiplanar to the oxygen of the ribose ring, which helps stabilize the radical against elimination of 4′-H, as suggested long ago.57 The agreement between φ determined experimentally for 5′-dAdo· in the active site of a frozen enzyme and φ from DFT of the radical in vacuo indicates that the dihedral angle is determined by interactions within the radical itself, not by those with its environment. 5′-dAdo· in the Active Site Expanding our focus, the use of additional SAM isotopologues gives a sense of the relationship of the SAM fragments (5′-dAdo· and Met) subsequent to homolysis. When 5′- dAdo· is prepared from [3,3,4,4-methionine-D4]-SAM, subtle changes in the shape of the EPR spectrum (Figure S7) indicate that the apparent EPR line width of the natural- abundance spectrum includes the effects of weak hyperfine couplings to protons of the methionine side chain. In contrast, there are no changes in the EPR spectrum when prepared from CD 13 3-methyl or C- methyl methionine SAM, and no weakly coupled 13C ENDOR signal is introduced by the 13C substitution (Figure S7), indicating the radical site is remote from the methyl group. These observations indicate that, upon SAM S−C5′ homolytic bond cleavage, the C5′ radical has shifted toward the methionine C3 and C4 and 121 away from the methyl, as illustrated schematically in Figure 5. Such a movement leaves Figure 4.4. DFT models of 5′-dAdo·. Top: Perspective view of the optimized structure. Adenosine is represented by a violet sphere; the isosurface plot of the calculated HOMO (yellow) uses an isodensity of 0.08 au and shows the direction of g3 normal to the C5′H2 plane. Bottom: left, conformer with a dihedral “twist” at the C5′−C4′ bond, φ = 0; right, optimized structure geometry φ = 39.3°. the spin-bearing C5′ in proximity to the methionine side chain, allowing unresolved hyperfine couplings to those hydrogens to influence the EPR spectrum, yet removed 122 enough from the methionine 13C-methyl so as to leave no detectable couplings either EPR or ENDOR spectra. Discussion The 5′-dAdo· radical is the central species responsible for H atom abstraction from the substrate in both coenzyme B12 and the large RS superfamily of enzymes, but until now, it had eluded characterization. In the present work, we generated the 5′-dAdo· radical in PFL-AE by cryogenic photoinduced electron transfer (ET) from the [4Fe−4S]1+ to SAM, causing reductive cleavage of SAM to generate the 5′-dAdo· radical, and methionine chelated to the diamagnetic [4Fe−4S]2+ cluster. Because the cluster is diamagnetic, there are no spin−spin interactions that would interfere with the characterization of the photogenerated radical, and its assignment as 5′- dAdo· Figure 4.5. Cartoon illustrating the proposed movement of 5′-dAdo· upon S−C(5′) bond cleavage, as inferred from EPR and ENDOR has been unequivocally data, based on PDB ID 3CB8;38 [−C5′(H)2] is shown as a purple ball. established through the use of 13C and 2H labels. 123 Our early ENDOR studies showed that, in the absence of PFL substrate, SAM is coordinated to the reduced PFL-AE [4Fe−4S]1+ cluster through the amino acid group of SAM with the sulfonium in close contact to the cluster (<4 Å).58,59 Thus, this photoinduced ET is analogous to the inner-sphere ET thought to occur during enzymatic catalysis,58,60 with key roles being played by configurational interaction between donor and acceptor orbitals on the sulfonium and the PFL-AE/[4Fe−4S]1+ cluster,61 conformational and electronic influences of substrate binding,62 and the mechanistic requirement of Ω formation.20,23,24 Why, then, is 5′-dAdo·, rather than Ω, produced in the current experiments? One possible answer is that Ω formation might require the conformational con- sequences of the presence of substrate. Unfortunately, we have not yet been able to test this hypothesis because adding substrate to the PFL-AE/[4Fe−4S]1+/SAM complex leads to immediate substrate turnover.60 Even when PFL-AE/[4Fe− 4S]1+, SAM, and substrate are combined in rapid-freeze quench experiments with short deadtimes, Ω is formed and 5′- dAdo· is not observed. In other words, the enzymatic turnover system appears to be designed to avoid the production of free dAdo· and rather to form the more stable 5′-dAdo· precursor Ω. As we have previously postulated,25 the highly reactive 5′- dAdo· is never free under normal catalytic conditions. An additional consideration is the nature of the process of electron transfer and SAM cleavage itself. During catalysis, RS enzymes must overcome a large potential barrier for reductive cleavage of SAM, due to the mismatch in reduction potenti between RS [4Fe−4S]1+ clusters (−400 to −600 mV)63-67 and SAM (estimated as ∼−1.8 V).68,69 The binding of SAM and substrate in the active site is thought to perturb both potentials, 124 lowering the barrier to reductive SAM cleavage.62 RS catalysis via formation of Ω likely also lowers the barrier to this cleavage through the stabilization contributed by the Fe−C5′ bond formation, thus providing a favorable alternative pathway to direct SAM cleavage by simple ET to form 5′- dAdo·.20,23,24 The photoinduced ET reported here instead uses photon energy to overcome the barrier to direct the formation of 5′- dAdo·. While understanding the precise nature of this process will require further study, prior studies of photoreduction by metal centers70 suggest that in the present study the photon excites the cluster to a ligand-to-metal charge transfer (LMCT) excited state, which quickly relaxes to a relatively long-lived [4Fe−4S]1+ ligand field excited state that is a potent and effective electron donor. In short, the energy of the incident photon, as captured in an ET-reactive excited state, is sufficient to enable the direct formation 5′-dAdo· by S−C5′ homolysis without the formation of Ω. Conclusion In summary, after more than half a century of attempts, the primary-carbon 5′- dAdo· radical has at last been captured, and its electronic and geometric structure has been characterized by multifrequency EPR, Q-band ENDOR, and DFT computation. The capture without radical rearrangement was accomplished via cryogenic (12 K), intra-EPR- cavity, photo- induced ET from the PFL-AE [4Fe−4S]1+ to cluster-bound SAM, triggering reductive cleavage of SAM to generate the 5′- dAdo· cryo-trapped radical along with the diamagnetic [4Fe− 4S]2+ cluster. The facile and efficient nature of this photoinduced ET, not previously reported for any radical SAM enzyme, raises the intriguing question of 125 whether photo- initiated reactions might play an as-yet unrecognized role in RS enzymes or other catalytic systems involving iron−sulfur clusters. Perhaps the most surprising finding about 5′-dAdo· itself is the absence of surprises: its remarkable reactivity accompanies properties that are almost precisely as foundational studies of organic radicals long ago would have predicted. 126 References 1. Barker, H. A.; Weissbach, H.; Smyth, R. D. A coenzyme containing pseudovitamin B12. Proc. Natl. Acad. Sci. U. S. A. 1958, 44, 1093−1097. 2. Lenhert, P. G.; Hodgkin, D. C. Structure of the 5,6- dimethylbenzimidazolylcobamide coenzyme. Nature 1961, 192, 937−938. 3. Halpern, J. Mechanisms of Coenzyme B12-Dependent Rearrangements. Science 1985, 227, 869−875. 4. Brown, K. L. Chemistry and Enzymology of Vitamin B12. Chem. Rev. 2005, 105, 2075−2149. 5. Banerjee, R.; Ragsdale, S. W. The many faces of vitamine B12: Catalysis by cobalamin-dependent enzymes. Annu. Rev. Biochem. 2003, 72, 209−247. 6. Frey, P. A.; Abeles, R. H. The role of the B12 coenzyme in the conversion of 1,2- propanediol to propionaldehyde. J. Biol. Chem. 1966, 241, 2732−2733. 7. Jones, A. R. The photochemistry and photobiology of vitamin B12. Photochem. Photobiol. Sci. 2017, 16, 820−834. 8. Rury, A. S.; Wiley, T. E.; Sension, R. Energy cascades, escited state dynamics, and photochemistry in cob(III)alamins and ferric porphyrins. Acc. Chem. Res. 2015, 48, 860−867. 9. Hogenkamp, H. P. C.; Barker, H. A.; Mason, H. S. An electron- spin resonance study of coenzyme B12. Arch. Biochem. Biophys. 1963, 100, 353−359. 10. Frey, P. A. Lysine 2,3-aminomutase: is adenosylmethionine a poor man’s adenosylcobalamin? FASEB J. 1993, 7, 662−670. 11. Frey, P. A.; Magnusson, O. T. S-Adenosylmethionine: A wolf in sheep’s clothing, or a rich man’s adenosylcobalamin? Chem. Rev. 2003, 103, 2129−2148. 12. Moss, M.; Frey, P. A. The role of S-adenosylmethionine in the lysine 2,3- aminomutase reaction. J. Biol. Chem. 1987, 262, 14859− 14862. 127 13. Baraniak, J.; Moss, M. L.; Frey, P. A. Lysine 2,3-aminomutase. Support for a mechanism of hydrogen transfer involving S- adenosylmethionine. J. Biol. Chem. 1989, 264, 1357−1360. 14. Knappe, J.; Schmitt, T. A novel reaction of S-adenosyl-L- methionine correlated with the activation of pyruvate formate-lyase. Biochem. Biophys. Res. Commun. 1976, 71, 1110−1117. 15. Frey, M.; Rothe, M.; Wagner, A. F. V.; Knappe, J. Adenosylmethionine-dependent synthesis of the glycyl radical in pyruvate formate-lyase by abstraction of the glycine C-2 pro-S hydrogen atom. J. Biol. Chem. 1994, 269, 12432−12437. 16. Broderick, J. B.; Duffus, B. R.; Duschene, K. S.; Shepard, E. M. Radical S- Adenosylmethionine Enzymes. Chem. Rev. 2014, 114, 4229−4317. 17. Bridwell-Rabb, J.; Grell, T. A. J.; Drennan, C. L. A rich man, poor man story of S- adebosylmethionine and cobalamin revisited. Annu. Rev. Biochem. 2018, 87, 555−584. 18. Landgraf, B. J.; McCarthy, E. L.; Booker, S. J. Radical S- adenosylmethionine enzymes in human health and disease. Annu. Rev. Biochem. 2016, 85, 485−514. 19. Holliday, G. L.; Akiva, E.; Meng, E. C.; Brown, S. D.; Calhoun, S.; Pieper, U.; Sali, A.; Booker, S. J.; Babbitt, P. C. Atlas of the radical SAM superfamily: Divergent evolution of function using a “Plug and Play” domain. Methods Enzymol. 2018, 606, 1−71. 20. Broderick, W. E.; Hoffman, B. M.; Broderick, J. B. Mechanism of Radical Initiation in the Radical S-adenosyl-L-methionine Super- family. Acc. Chem. Res. 2018, 51, 2611−2619. 21. Frey, P. A. Travels with carbon-centered radicals. 5′- deoxyadenosine and 5′- deoxyadenosine-5′-yl in radical enzymology. Acc. Chem. Res. 2014, 47, 540−549. 22. Frey, P. A.; Hegeman, A. D.; Reed, G. H. Free radical mechanisms in enzymology. Chem. Rev. 2006, 106, 3302−3316. 23. Horitani, M.; Shisler, K. A.; Broderick, W. E.; Hutcheson, R. U.; Duschene, K. S.; Marts, A. R.; Hoffman, B. M.; Broderick, J. B. Radical SAM catalysis via an organometallic intermediate with an Fe- [5′-C]-deoxyadenosyl bond. Science 2016, 352, 822−825. 128 24. Byer, A. S.; Yang, H.; McDaniel, E. C.; Kathiresan, V.; Impano, S.; Pagnier, A.; Watts, H.; Denler, C.; Vagstad, A. L.; Piel, J.; Duschene, K. S.; Shepard, E. M.; Shields, T. P.; Scott, L. G.; Lilla, E. A.; Yokoyama, K.; Broderick, W. E.; Hoffman, B. M.; Broderick, J. B. Paradigm shift for radical S-adenosyl-L-methionine reactions: The organometallic intermediate Ω is central to catalysis. J. Am. Chem. Soc. 2018, 140, 8634−8638. 25. Horitani, M.; Byer, A. S.; Shisler, K. A.; Chandra, T.; Broderick, J. B.; Hoffman, B. M. Why Nature Uses Radical SAM Enzymes so Widely: Electron Nuclear Double Resonance Studies of Lysine 2,3- Aminomutase Show the 5′-dAdo• ′′Free Radical′′ is Never Free. J. Am. Chem. Soc. 2015, 137, 7111−7121. 26. Magnusson, O. T.; Reed, G. H.; Frey, P. A. Spectroscopic Evidence for the participation of an allylic analogue of the 5′- deoxyadenosyl radical in the reaction of lysine 2,3-aminomutase. J. Am. Chem. Soc. 1999, 121, 9764−9765. 27. Ghanekar, V. D.; Lin, R. J.; Coffman, R. E.; et al. Detection by electron spin resonance of an exchange-coupled cob(II)alamin···free radical pair species generated by anaerobic photolysis of polycrystal- line adenosylcobalamin. Biochem. Biophys. Res. Commun. 1981, 101, 215−221. 28. Bussandri, A. P.; Kiarie, C. W.; Van Willigen, H. Photoinduced bond homolysis of B12 coenzymes. An FT-EPR study. Res. Chem. Intermed. 2002, 28, 697−710. 29. Heidinger, L.; Kneuttinger, A. C.; Kashiwazaki, G.; Weber, S.; Carell, T.; Schleicher, E. Direct observation of a deoxyadenosyl radical in an active enzyme environment. FEBS Lett. 2016, 590, 4489−4494. 30. Byer, A. S.; McDaniel, E. C.; Impano, S.; Broderick, W. E.; Broderick, J. B. Mechanistic Studies of Radical SAM Enzymes: Pyruvate Formate-Lyase Activating Enzyme and Lysine 2,3-Amino- mutase Case Studies. Methods Enzymol. 2018, 606, 269−318. 31. Stoll, S.; Schweiger, A. EasySpin, a comprehensive software package for spectral simulation and analysis in EPR. J. Magn. Reson. 2006, 178, 42−55. 32. Werst, M. M.; Davoust, C. E.; Hoffman, B. M. Ligand Spin Densities in Blue Copper Proteins by Q-band 1H and 14N ENDOR Spectroscopy. J. Am. Chem. Soc. 1991, 113, 1533−1538. 129 33. Zipse, H.; Artin, E.; Wnuk, S.; Lohman, G. J. S.; Martino, D.; Griffin, R. G.; Kacprzak, S.; Kaupp, M.; Hoffman, B.; Bennati, M.; Stubbe, J.; Lees, N. Structure of the nucleotide radical formed during reaction of CDP/TTP with the E441Q-a2b2 of E. coli ribonucleotide reductase. J. Am. Chem. Soc. 2009, 131, 200−211. 34. Davoust, C. E.; Doan, P. E.; Hoffman, B. M. Q-Band Pulsed Electron Spin-Echo Spectrometer and Its Application to ENDOR and ESEEM. J. Magn. Reson., Ser. A 1996, 119, 38−44. 35. Hoffman, B. M.; DeRose, V. J.; Ong, J. L.; Davoust, C. E. Sensitivity enhancement in field-modulated CW ENDOR via RF bandwidth broadening. J. Magn. Reson., Ser. A 1994, 110, 52−57. 36. Lee, H.-I.; Igarashi, R. Y.; Laryukhin, M.; Doan, P. E.; Dos Santos, P. C.; Dean, D. R.; Seefeldt, L. C.; Hoffman, B. M. An organometallic intermediate during alkyne reduction by nitrogenase. J. Am. Chem. Soc. 2004, 126, 9563−9569. 37. Neese, F. The ORCA program system. Comput. Mol. Sci. 2012, 2, 73−78. 38. Vey, J. L.; Yang, J.; Li, M.; Broderick, W. E.; Broderick, J. B.; Drennan, C. L. Structural basis for glycyl radical formation by pyruvate formate-lyase activating enzyme. Proc. Natl. Acad. Sci. U. S. A. 2008, 105, 16137−16141. 39. Becke, A. D. Density-functional exchange-energy approxima- tion with correct asymptotic behavior. Phys. Rev. A: At., Mol., Opt. Phys. 1988, 38, 3098−3100. 40. Perdew, J. P. Density-functional approximation for the correlation-energy of the inhomogeneous electron-gas. Phys. Rev. B: Condens. Matter Mater. Phys. 1986, 33, 8822−8824. 41. Lee, C.; Yang, W.; Parr, R. G. Development of the Colle- Salvetti correlation energy formula into a functional of the electron density. Phys. Rev. B: Condens. Matter Mater. Phys. 1988, 37, 785−789. 42. Schaf̈er,A.;Huber,C.;Ahlrichs,R.Fullyoptimizedcontracted gaussian basis sets of triple zeta valence quality for atoms Li to Kr. J. Chem. Phys. 1994, 100, 5829−5835. 43. Barone, V. In Recent Advances in Density Functional Methods; World Scientific: Singapore, 1995; p 287. 44. Rega, N.; Cossi, M.; Barone, V. J. Chem. Phys. 1996, 105, 11060. 130 45. Neese, F. Efficient and accurate approximations to the molecular spin-orbit coupling operator and their use in molecular g- tensor calculations. J. Chem. Phys. 2005, 122, No. 034107. 46. Dektar, J. L.; Hacker, N. P. Photochemistry of triarylsulfonium salts. J. Am. Chem. Soc. 1990, 112, 6004−6015. 47. Saeva, F. D.; Morgan, B. P.; Luss, H. R. Photochemical conversion of sulfonium salts to sulfides via a 1,3-sigmatropic rearrangement. Photogeneration of Bronsted acids. J. Org. Chem. 1985, 50, 4360−4362. 48. Mao, Z.; Liou, S.-H.; Khadka, N.; Jenney, F. E. J.; Goodin, D. B.; Seefeldt, L. C.; Adams, M. W. W.; Cramer, S. P.; Larsen, D. Cluster-dependent charge-transfer dynamics in iron-sulfur proteins. Biochemistry 2018, 57, 978−990. 49. Jost, M.; Simpson, J. H.; Drennan, C. L. The transcription factor CarH safeguards use of adenosylcobalamin as a light sensor by altering the photolysis products. Biochemistry 2015, 54, 3231−3234. 50. Weil, J. A.; Bolton, J. R.; Wertz, J. E. Electron Paramagnetic Resonance: Elementary Theory and Practical Applications; John Wiley & Sons: New York, 1994. 51. Atherton, N. M. Principles of Electron Spin Resonance; Ellis Horwood: New York, 1993. 52. Carrington, A.; McLachlan, A. D. Introduction to magnetic resonance: with applications to chemistry and chemical physics; Harper and Row: New York, 1967. 53. McConnell, H. M.; Strathdee, J. Theory of anisotropic hyperfine interactions in pi- electron radicals. Mol. Phys. 1959, 2, 129−138. 54. Horsfield, A.; Morton, J. R.; Whiffen, D. H. Electron-spin resonance of gamma- irradiated malonic acid. Mol. Phys. 1961, 4, 327− 332. 55. McConnell, H. M.; Heller, C.; Cole, T.; Fessenden, R. W. Radiation damage in organic crystals. I. CH(COOH)2 in malonic acid. J. Am. Chem. Soc. 1960, 82, 766−775. 56. Cole, T.; Heller, C. Hyperfine Splittings in the (HOOC)- C13H(COOH) Radical. J. Chem. Phys. 1961, 34, 1085−1086. 131 57. Lott, W. B.; Chagovetz, A. M.; Grissom, C. B. Alkyl radical geometry controls geminate cage recombination in alkylcobalamins. J. Am. Chem. Soc. 1995, 117, 12194−12201. 58. Walsby, C. J.; Hong, W.; Broderick, W. E.; Cheek, J.; Ortillo, D.; Broderick, J. B.; Hoffman, B. M. Electron-nuclear double resonance spectroscopic evidence that S- adenosylmethionine binds in contact with the catalytically active [4Fe-4S]+ cluster of pyruvate formate-lyase activating enzyme. J. Am. Chem. Soc. 2002, 124, 3143− 3151. 59. Walsby, C. J.; Ortillo, D.; Broderick, W. E.; Broderick, J. B.; Hoffman, B. M. An anchoring role for FeS Clusters: Chelation of the amino acid moiety of S- adenosylmethionine to the unique iron site of the [4Fe-4S] cluster of pyruvate formate-lyase activating enzyme. J. Am. Chem. Soc. 2002, 124, 11270−11271. 60. Henshaw, T. F.; Cheek, J.; Broderick, J. B. The [4Fe-4S]+ cluster of pyruvate formate-lyase activating enzyme generates the glycyl radical on pyruvate formate- lyase: EPR-detected single turnover. J. Am. Chem. Soc. 2000, 122, 8331−8332. 61. Dey, A.; Peng, Y.; Broderick, W. E.; Hedman, B.; Hodgson, K. O.; Broderick, J. B.; Solomon, E. I. S K-edge XAS and DFT Calculations on SAM Dependent Pyruvate Formate-Lyase Activating Enzyme: Nature of Interaction between the Fe4S4 Cluster and SAM and its Role in Reactivity. J. Am. Chem. Soc. 2011, 133, 18656−18662. 62. Wang, S. C.; Frey, P. A. Binding energy in the one-electron reductive cleavage of S-adenosylmethionine in lysine 2,3-amino- mutase, a radical SAM enzyme. Biochemistry 2007, 46, 12889−12895. 63. Hinckley, G. T.; Frey, P. A. Cofactor dependence of reduction potentials for [4Fe- 4S]2+/1+ in lysine 2,3-aminomutase. Biochemistry 2006, 45, 3219−3225. 64. Ugulava, N. B.; Gibney, B. R.; Jarrett, J. T. Biotin synthase contains two distinct iron-sulfur cluster binding sites: Chemical and spectroelectrochemical analysis of iron-sulfur cluster interconversions. Biochemistry 2001, 40, 8343−8351. 65. Maiocco, S. J.; Grove, T. L.; Booker, S. J.; Elliott, S. J. Electrochemical resolution of the [4Fe-4S] centers of the AdoMet radical enzyme BtrN: Evidence of proton coupling and an unusual, low-potential auxiliary cluster. J. Am. Chem. Soc. 2015, 137, 8664− 8667. 132 66. Blaszczyk, A. J.; Silakov, A.; Zhang, B.; Maiocco, S. J.; Lanz, N. D.; Kelly, W. L.; Elliott, S. J.; Krebs, C.; Booker, S. J. Spectroscopic and electrochemical characterization of the iron-sulfur and cobalamin cofactors of TsrM, an unusual radical S-adenosylmethionine methylase. J. Am. Chem. Soc. 2016, 138, 3416−3426. 67. Maiocco, S. J.; Arcinas, A. J.; Landgraf, B. J.; Lee, K.-H.; Booker, S. J.; Elliott, S. J. Transformations of the FeS clusters of the methylthiotransferases MiaB and RimO, detected by direct electro- chemistry. Biochemistry 2016, 55, 5531−5536. 68. Grimshaw, J. Electrochemistry of the sulfonium group. In Chemistry of the Sulphonium Group; Stirling, C. J. M., Ed.; Wiley: Chichester, 1981; pp 141−155. 69. Saeva, F. D.; Morgan, B. P. Mechanism of one-electron electrochemical reductive cleavage reactions of sulfonium salts. J. Am. Chem. Soc. 1984, 106, 4121−4125. 70. McCusker, J. K. Electronic structure in the transition metal block and its implications for light harvesting. Science 2019, 363, 484− 488. 133 Supplementary Information Supplementary Methods The same radical species is observed upon photolysis of the PFL-AE/[4Fe- 4S]1+/SAM complex, regardless of whether the reduced enzyme is prepared by photoreduction (Fig. S1A) or chemical reduction with dithionite (Fig S1B). Figure S4.1. X-band EPR spectra of 5′-dAdo• generated through in situ photolysis of (PFL-AE/[4Fe- 4S]1++unlabeled SAM) complex. A, PFL- AE/[4Fe-4S]1+ was prepared with 5-deazariboflavin reduction; B, PFL-AE1+ was prepared with dithionite reduction. Experimental conditions: microwave frequency, 9.38 GHz; modulation, 10G; T = 40 K. Supplementary Text Has the fluid-solution spectrum of 5′-Ado• been seen in an FT-EPR study of photolyzed B ?1 12 An FT-EPR study of the photolysis of a variety of fluid-solution alkyl cobalamins1 reported that B12 photolysis generated a sharp EPR spectrum with the largest hyperfine splitting being a 1:2:1 triplet split by the isotropic coupling from 2 equivalent protons, a = 60 MHz, modeled in Fig S2A. Such couplings were assigned by those authors to the – C5′(H)2 protons of 5′-dAdo• formed by homolysis of the Co-C5′ bond, and such isotropic couplings are indeed found here in our study of trapped 5′-dAdo•. The individual lines of the triplet exhibited additional very small splittings (a = 4.50 MHz; a = 1.65 MHz) which are unimportant to this discussion and are ignored until point (iv) below, and are omitted in the simulated spectra of Fig S2. Can such a spectrum (Fig S2A) be the isotropic spectrum of 5′-dAdo• in fluid solution? As we here demonstrate, our work shows that the answer is no: the fluid solution spectrum of 5′-dAdo• must also show a large doublet splitting from 1H-C4′, and thus must appear not as a (major) triplet, but as a doublet of such triplets. 134 We reach this conclusion by considering the set of alternative possibilities for the fluid-solution spectrum of a 5′C-dAdo• radical tumbling in fluid solution that arise from the possibly-dynamical orientation of the -C5′(H2) plane around the C5′-C4′ bond, with corresponding consequences for the C4′(1H) isotropic hyperfine coupling. The predicted alternative spectra for 5′-dAdo• are shown in Fig S2B, C for comparison with the simulation using parameters from the reported spectrum, Fig S2A. i) The -5′CH2 fragment might be freely rotating, in which case C4′(1H) would give an isotropic coupling given by the average of the β-C4′(1H) hyperfine coupling over a 2π rotation with equal probability for all orientations, a = (ρπB) = (ρπB)(1/2) ≈ (149)/2 ≈ 74.5 MHz where we have used the value a 1 iso = 90 MHz obtained from the H-C4′ hyperfine tensor we measured (Table 1) plus the most stable twist angle as derived from the DFT computations, φs = 39°, to recalibrate the ‘scale’ factor: ρπB = 149 MHz with the resulting predicted fluid-solution spectrum given in Fig S2B. (ii) The -5′CH2 fragment might be locked in the antiperiplanar structure of the frozen radical, in which case C4′(1H) would give the isotropic value for the hyperfine tensor measured here (Table 1) for 5′C-dAdo• tumbling freely in solution a = a = (ρ 2 πB) cos (φs) = 90 MHz with the resulting predicted fluid-solution spectrum given in Fig S2C. (iii) Alternatively the orientation of -5′CH2 fragment might be thermally- averaged over all twist angles, yielding an isotropic β-C4′H hyperfine coupling, a = (ρπB) but with the weighting for each angle determined for the energy surface associated with rotation of -5′CH2 about C5′-C4′. Direct integration over the DFT-calculated energy surface of Fig S4 shows that the energy-weighted average of cos2(φ) equals the value for the fixed minimum-energy structure, = cos2(390) and thus in this case the expected coupling is identical to that observed in the present frozen-solution study, a = 90 MHz, with the predicted fluid-solution spectrum likewise given by Fig S2C. (iv) Could one of the small coupling seen in the experimental spectrum actually be from C4′H of 5′C-dAdo•? This would require that the -C5’(H2) plane be locked in a conformation with an angle of φ ≈ 90˚, which in fact is unlikely, as it has long been known that an R2HCβ-Cα(H)2 center prefers the φ ≈ 0 orientation,2 and indeed we find a much smaller angle for freeze-trapped 5′-dAdo•. According to the DFT energy surface (Fig S4), which indeed would apply more cleanly to the solution radical than to the radical trapped in the PFL-AE active site, whose conformation may be influenced by the enzyme environment, at the FT-EPR temperature (T = 283K) such a conformation would have a small equilibrium population relative to that at the energy-minimum (eq S2), P(90˚)/ P(39˚) ≈ exp(-δE(90˚)/RT) ≈ 0.16 (S2) Thus, the experimentally observed spectrum is not credibly assigned to a dominant conformation with small C4′H isotropic coupling. 135 Figure S4.2. Simulations of X-band FT-EPR spectra (absorption-display) for a radical tumbling in solution; small couplings omitted. (A) As reported for the radical studied in ref 9, 2x(5′-C-1Hα) protons, aiso = 60 MHz. (B, C) Alternative possible spectra for 5′C-Ado• depending on structure/dynamics as discussed in SI text. Predicted spectra for: (B) 2x(5′- C1Hα), aiso = 60 MHz; 4′-C1Hβ, aiso = 74.5 MHz; (C) 2x(5′-C1Hα), a 1 iso = 60 MHz; 4′-C Hβ, aiso = 90 MHz. The same Lorentzian linewidth, 1 G, and isotropic g-value, g = 2.003, are used for all simulations, with microwave frequency, 9.38 GHz. Thus, under any type of fluid-solution structural/dynamic behavior for the C4′-C5′ dihedral angle/dynamics, the spectrum of 5′-dAdo• must show a C4′H coupling larger than the C5′ (H)2 couplings, which is absent in the published spectrum: as clearly shown in Fig S2, the published spectrum is not that of 5′-dAdo•. Wherein lies the problem with the FT-EPR report? We are led to conclude that the 5′-dAdo• that has escaped geminate recombination subsequent to the fluid-solution (10 C°) photolysis of B12 has undergone radical rearrangement (a well known phenomenon)3 within the 80 ns delay time between laser pulse and initiation of the FT- EPR measurement to give the species observed by FT-EPR, an example of the high reactivity/instability of 5′-dAdo•. A reexamination of this species formed with isotopic labels could be instructive. We may note that, in contrast to this issue, we believe the remainder of the results in the FT-EPR report are reliable. Is the EPR spectrum of a radical generated from Geobacillus thermodenitrificans (Gt) spore photoproduct lyase (SPL)4 compatible with assignment to 5′-dAdo•? To begin, the high stability of the reported radical is sufficient to cast doubt on its assignment to 5′- dAdo•. Whereas we find that 5′-dAdo• decays in one minute at 150 K, the reported radical accumulates at 313 K over 30 minutes. Moreover, the reported assignment was not subjected to the required confirmation by the use of SAM isotopologues, as we have done here. 136 Figure S4.3. Comparison of the X- band EPR spectrum of 5′-dAdo• from this study and the reported spectrum of the radical generated during reaction of SPL with simulation, adapted from the previous study (4). The parameters used in the published simulation are listed in Table S1. Beyond these concerns, however, direct comparison of the authentic frozen- solution spectrum of 5′-dAdo• reported here with that of the frozen solution of the radical studied in ref 4, Fig S3, shows that the two spectra are different. Moreover, there are even more dramatic differences between the reported parameters required to simulate the published spectrum and those that simulate the current spectrum of authentic 5′-dAdo•, Fig S3, Table S1 versus Table 1). As shown in our study of 5′-dAdo•, the most distinctive feature in the frozen-solution EPR spectrum of 5′-dAdo• is the anisotropic hyperfine coupling to the two near-equivalent 5′-C protons. In fact, it has long been established that the α proton for any C(2pπ)-H radical in frozen solution state must exhibit such an anisotropic hyperfine coupling.5-9 However, the 70 K EPR spectra of the radical generated with Gt-SPL C140A variant was well-simulated with isotropic coupling to two equivalent protons with Aiso = 60 MHz. Thus the properties of the reported radical are fundamentally incompatible with an assignment to 5′-Ado•: the two radicals are different. Indeed, whether the reported radical is even associated with any SAM component, such as a rearranged 5′- dAdo•, or with the enzyme itself, must first be established by using SAM isotopologues before one could even speculate on its identity. 5-dAdo• Radical from SPL A1 A2 A3 A1 A2 A3 5-C-1Ha -105(-104) -15(-28) -60(-63) -60 -60 -60 5-C-1Hb -20(-25) -95(-101) -60(-61) -60 -60 -60 4-C-1Hβ +80 (+90) +80(+90) +110(+105) +104 +104 +104 T able S4.1. Hyperfine tensors (MHz) of 5′-dAdo• from current experiment a plus DFT-computed compared to the radical generated from SPL. 137 a Signs (+/-) of experimental tensors are fixed as described in ref 9; hyperfine tensors in parentheses are from DFT calculation with B3LYP/G. DFT calculations. DFT computations were carried out as described in the main text. Hyperfine and g tensors were calculated by the coupled−perturbed self-consistent field (SCF) approach utilized as implemented by ORCA 4.0 using the B3LYP and BP86 hybrid functionals and EPR-III basis 10-11 and in combination with an accurate spin-orbit coupling operator [RI- SOMF(1X)] Comparison of hyperfine couplings calculated with B3LYP and BP86 Functionals gave = B3LYP φ = 39.4 2.0027611 BP86 φ = 48.9 gave = 2.0027584 5’-dAdo• Ax Ay Az Ax Ay Az 5’-C-1H a -61 -101 -24 -61 -100 -28 5’-C-1H b -28 -104 -63 -29 -100 -63 4’-C-1Hβ 105 90 90 65 62 101 5’-13C 2 240 3 15 218 14 4’-13C 69 86 68 70 91 70 3’-13C -37 -33 -36 -35 -33 -30 An example B3LYP input file is below. #EPR Calculation Example Input ! UKS B3LYP/G EPR-III SOMF(1X) Grid4 NoFinalGrid VeryTightSCF %scf maxiter 300 damp fac 0.85 erroff 0.005 end end * xyz 0 2 ################# * %eprnmr gtensor 1 ori -3 end %eprnmr Nuclei = all H { aiso, adip, rho } Nuclei = all C { aiso, adip, fgrad, rho } end 138 Figure S4.4. DFT calculated energy surface for rotation about the 4C′-5C′ bond. The black dots represent the calculated single point energy. The blue dotted line is the fitted curve. Using the optimized geometry from the B3LYP functional calculation, the energy surface for rotation about the 4C′-5C′ bond was obtained by varying this angle in 10º increments and carrying out a single point calculate. The resulting potential energy surface is plotted in Figure S4. These points defining the surface were fit to the two-fold rotation function, δE(φ) = a[1 – cos(2(φ – φs)] with minimum energy set at φ 0 s = 39.4 obtained from the DFT energy minimization, and the resulting function is plotted in the figure; the difference between maximum and minimum points on the curve is, 2a = 1.69 kcal/M. 139 Supplementary Figures Figure S4.5. Q-band EPR spectra (black) & simulations (red) of 5′-dAdo•. 5′-dAdo• generated with: (a) natural abundance SAM; (b), [adenosyl-5′,5″-D2]-SAM; (c) [adenosyl-2,8-D2-1′,2′,3′,4′,5′,5″-D6]-SAM; the additional linewidth of features at low field is attributable to g/A strain. A feature associated with minority photolysis products is revealed in spectrum b and indicated by (*). Conditions: EPR, microwave frequency 34 GHz (a-c); modulation, 10 G; T = 40 K. Simulations: Generated with EasySpin using a 5′-dAdo• model with parameters listed in Table 1. 140 Figure S4.6. 35 GHz CW stochastic 13C ENDOR of 5′-dAdo• generated with [adenosyl-13C ,15 10 N5]- SAM in comparison with natural abundance SAM (gray). The ENDOR response under the arrow is attributed to the partially resolved 4′-13C ENDOR response. Experimental condition: microwave frequency, 34.81 GHz; modulation amplitude, 1.6 G; sample time, 2 ms; delay time 1 ms; RF-on time, 3 ms; T = 2 K. 141 Figure S4.7. X-band EPR spectra of 5′-dAdo• generated with natural abundance SAM (A), [CD3- Methyl]-SAM (B), [13C- methyl]-SAM (C), [3,3,4,4-methioine-D4]-SAM (D), and 35 GHz Mims 13C ENDOR of 5′-dAdo• generated with [13C- methyl]-SAM (D). The asterisk in E denoted the fifth harmonics of proton larmor. EPR conditions: microwave frequency, 9.38 GHz; modulation 10 G; T = 40 K. ENDOR conditions: microwave frequency, 34.742 GHz; τ = 500 ns; T = 2 K. 142 Supplementary References 1. Bussandri, A. P.; Kiarie, C. W.; Van Willigen, H. ”Photoinduced bond homolysis of B12 coenzymes. An FT-EPR study.,” Res. Chem. Intermed. 2002, 28, 697-710. 2. Fessenden, R. W.; Schuler, R. H. ”Electron spin resonance studies of transient alkyl radicals,” J. Chem. Phys. 1963, 39, 2147. 3. Jones, A. R. ”The photochemistry and photobiology of vitamin B12,” Photochem. Photobiol. Sci. 2017, 16, 820-834. 4. Heidinger, L.; Kneuttinger, A. C.; Kashiwazaki, G.; Weber, S.; Carell, T.; Schleicher, E. ”Direct observation of a deoxyadenosyl radical in an active enzyme environment,” FEBS Lett. 2016, 590, 4489-4494. 5. McConnell, H. M.; Strathdee, J. ”Theory of anisotropic hyperfine interactions in pi-electron radicals,” Molec. Phys. 1959, 2, 129-138. 6. Horsfield, A.; Morton, J. R.; Whiffen, D. H. ”Electron-spin resonance of gamma- irradiated malonic acid,” Mol. Phys. 1961, 4, 327-332. 7. Weil, J. A.; Bolton, J. R.; Wertz, J. E., Electron Paramagnetic Resonance: Elementary Theory and Practical Applications. John Wiley & Sons: New York, 1994. 8. Atherton, N. M., Principles of Electron Spin Resonance. Ellis Horwood: New York, 1993. 9. Carrington, A.; McLachlan, A. D., Introduction to magnetic resonance: with applications to chemistry and chemical physics. Harper and Row: New York, 1967. 10. Barone, V., In Recent Advances in Density Functional Methods, World Scientific: Singapore, 1995; p 287. 11. Rega, N.; Cossi, M.; Barone, V. J. Chem. Phys. 1996, 105, 11060. 12. Neese, F. ”Efficient and accurate approximations to the molecular spin-orbit coupling operator and their use in molecular g-tensor calculations,” J. Chem. Phys. 2005, 122, 034 07. 143 CHAPTER FIVE FURTHER STUDIES INTO THE ORDER OF INTERMEDIATE FORMATION IN RADICAL S-ADENOSYL-L-METHIONINE (SAM) ENZYMES USING PFL-AE AND AN ANHYDROUS ANALOG OF SAM Contribution of Authors and Co-Authors Manuscript in Chapter 5 Author: Elizabeth C. McDaniel Contributions: Expressed and purified PFL-AE and PFL, synthesized SAM, prepared photolysis samples, prepared RFQ samples at Northwestern University, assisted in EPR data analysis, assisted in manuscript preparation. Co-Author: Hao Yang Contributions: Conducted RFQ experiments and EPR at Northwestern University, generated simulations. Co-Author: Stella Impano Contributions: Helped prepare RFQ samples at Northwestern University Co-Author: Eric M. Shepard Contributions: Conducted EPR experiment and assisted with data interpretation. Co-Author: Richard J. Jodts Contributions: Assisted in RFQ experiments at Northwestern University Co-Author: Sarah Hill Contributions: Conducted enzyme activity assays 144 Contribution of Authors and Co-Authors Continued Co-Author: Adrien Pagnier Contribution: Assisted with data interpretation. Co-Author: Maike Lundahl Contributions: Assisted with manuscript preparation and interpretation of EPR data. Co-Author: Amanda S. Byer Contributions: Synthesized anSAM Co-Author: William E. Broderick Contributions: Assisted with manuscript preparation and data interpretation. Co-Author: Brian M. Hoffman Contributions: Assisted with manuscript preparation and data interpretation. Co-Author: Joan B. Broderick: Contributions: Assisted with manuscript preparation and data interpretation. 145 CHAPTER FIVE FURTHER STUDIES INTO THE ORDER OF INTERMEDIATE FORMATION IN RADICAL S-ADENOSYL-L-METHIONINE (SAM) ENZYMES USING PFL-AE AND AN ANHYDROUS ANALOG OF SAM Abstract All radical S-adenosyl-ʟ-methionine (SAM) enzymes utilize the small molecule SAM as either a cofactor or a cosubstrate to generate an organometallic intermediate, Ω, which cleaves to form the 5′-deoxyadenosyl radical (5′-dAdo•). Both intermediates, Ω and the 5′-dAdo•, are short-lived which has made determining their mechanism of formation a challenge. Here we lay the groundwork for future study of the radical SAM (RS) enzyme mechanism. We show for the first time that an analog of SAM, S-3′,4′-anhydroadenosyl-ʟ- methionine (anSAM), can serve as a cosubstrate for pyruvate formate-lyase activating enzyme (PFL-AE). Further, blue light, cryogenic photolysis of the PFL-AE + anSAM complex initiates photoinduced electron transfer from the PFL-AE reduced cluster to anSAM, forming both the 3′,4′-anhydroadenosyl radical (anAdo•) and the methyl radical (•CH3) as observed using electron paramagnetic resonance spectroscopy. Rapid quenching the reaction of PFL-AE with anSAM and PFL reveals the formation of an organometallic intermediate species analogous to Ω, in which the 5′C of the anAdo• is bound to the unique Fe of the [4Fe-4S] cluster. This intermediate, anΩ, forms at quench times as rapid as 100 ms, indicating it must be forming prior to the anAdo•. 146 Introduction Radical S-adenosyl-ʟ-methionine (SAM) enzymes comprise one of the largest and most catalytically diverse enzyme superfamilies (1, 2). Despite catalyzing a wide range of transformations on a variety of substrates, all radical SAM (RS) enzymes utilize a site- differentiated [4Fe-4S] cluster and the small molecule SAM to initiate catalysis via reductive cleavage of SAM, generating methionine and the 5′-deoxyadenosyl radical (5′- dAdo•) as reaction intermediates (1-3). The RS mechanism proceeds through an organometallic intermediate, Ω, in which the 5′C of the adenosyl moiety is bound to the unique Fe of the [4Fe-4S] cluster. Homolytic cleavage of the Fe-C5′ bond releases the 5′- dAdo• for hydrogen abstraction. The 5′-dAdo• is highly reactive due to the location of the unpaired electron on the primary carbon, 5′C. This reactivity allows the 5′-dAdo• to break strong X-H bonds by abstracting a H atom to initiate challenging reactions (4, 5). In the catalytic reactions of RS enzymes, SAM can be used catalytically as a cofactor where SAM is regenerated with each turnover or stoichiometrically as a cosubstrate where SAM is consumed (2, 6). In all RS enzymes, the radical initiation mechanism begins with the reduced [4Fe-4S]+ cluster homolytically cleaving the 5′C-S bond through an inner sphere electron transfer (2, 3, 7). When SAM is used as a cofactor, it undergoes reversable S-C5′ bond cleavage with the 5′-dAdo• able to re-form SAM after catalysis. The 5′-dAdo• abstracts a H atom from substrate (SH) generating 5′-dAdoH and the S•. The S• will either rearrange or otherwise generate a product radical (P•) that will abstract a hydrogen from 5′-dAdoH reforming the 5′-dAdo•. The 5′-dAdo• will then 147 recombine with methionine and loss of an electron to the [4Fe-4S]2+ cluster regenerating SAM bound to the active [4Fe-4S]+ cluster (Figure 5.1 left). When SAM is used as a cosubstrate, it is consumed stoichiometrically with substrate, producing methionine and 5′- dAdoH as products. The 5′-dAdo• abstracts a H atom from SH and the 5′-dAdoH and methionine products are replaced by a new SAM molecule for the next catalytic reaction (Figure 5.1 right) (8, 9). Most characterized RS enzymes use SAM as a cosubstrate; only three are currently reported to use SAM as a cofactor: lysine 2,3-aminomutase (LAM), spore photoproduct lyase (SPL), and QueE (6, 10-12). There does not appear to be a structural difference in the RS enzyme active sites associated with SAM being used as a cofactor versus a PH e- P• Met + 5′-dAdoH SAM 5′-dAdo• [4Fe-4S]2+ + Met [4Fe-4S]2+ + SAM 5′-dAdoH [4Fe-4S]2+ + Met 5′-dAdoH s a a S AM [4Fe-4S]2+ + Met AM r co S• S cto -su as a cofa bst rate SH [4Fe-4S]+ + SAM S• SH 5′-dAdo• [4Fe-4S]2+ + Met 5′-dAdo• [4Fe-4S]2+ + Met Figure 5.1. SAM can serve as either a cofactor (left) or a cosubstrate (right). Once the [4Fe-4S]2+ cluster is reduced to the +1 state, SAM is cleaved into the 5′-dAdo• and methionine. These two parts can either reform after catalysis and SAM is used stoichiometrically or the 5′-dAdoH and methionine must leave the active site and a new SAM molecule comes in to continue enzyme activity. Methionine is abbreviated as Met. 148 cosubstrate (9). Instead, the difference in SAM usage among RS enzymes is attributed to the chemical reaction being catalyzed by the enzyme. When SAM serves as a cosubstrate, the RS enzyme substrate undergoes an irreversible change in its oxidation state as seen in pyruvate formate-lyase activating enzyme (PFL-AE) where an electron and proton are irreversibly transferred from the enzyme substrate to the 5′-dAdo• (9). When SAM is used as a cofactor, the catalyzed reactions are described as rearrangement reactions where the substrate does not undergo a net change in its oxidation state (9). This type of rearrangement reaction is best seen in LAM which catalyzes the conversion of L-α-lysine to L-ꞵ-lysine (13). Adenosyl-cobalamin (AdoCbl) enzymes are a family of enzymes that catalyze rearrangement reactions analogous to those observed in RS enzymes (2, 14). Like RS enzymes, AdoCbl enzymes release the 5′-dAdo• intermediate through homolysis of a metal-carbon, Co(III)-5′C-adenosyl, bond (6). The Co-5′C bond is relatively weak, 26 kcal/mol, allowing AdoCbl to function exclusively as a cofactor with the 5′-dAdo• and Co(II) recombining after each catalytic cycle (2, 15). 149 Integral to the RS pathways discussed above are the mechanistic details involving 5′-dAdo• and the organometallic intermediate Ω, both of which play central roles regardless of SAM serving as a cofactor or a cosubstrate (16). The mechanism of Ω formation is not fully understood. One possibility is that the 5′-dAdo• is generated first through reductive cleavage of SAM, with Ω then formed through oxidative addition of the adenosyl radical to the unique iron of the [4Fe-4S]2+ cluster to generate an adenosyl-[4Fe- 4S]3+ species. Homolytic cleavage of the 5′C-Fe bond of Ω would then generate 5′-dAdo• for reaction with substrate (Figure 5.2a). Alternatively, Ω could be formed through a concerted process in which the organometallic species is formed directly (Figure 5.2b) (16). Due to the highly reactive nature of the 5′-dAdo• and the rapidity at which Ω forms, it has been difficult to determine which pathway is correct (5). Two of the best characterized members of the RS enzyme superfamily, LAM and PFL-AE, are examples of enzymes that use SAM as a cofactor (LAM) and a cosubstrate 3+Г A [4Fe-4S]2++ 5ʹ-dAdo• Ω +Г B [4Fe-4S]++ SAM [4Fe-4S]2++ 5ʹ-dAdo• Ω Figure 5.2. Two proposed mechanisms for radical initiation. Path A, 5′-dAdo• is formed first followed by the 5′C attaching to the unique Fe generating omega. Path B, omega is formed first by 5′C-S homolysis followed immediately by combination of the 5′-dAdo• and the unique Fe. In both pathways, omega is homolytically cleaved to generate the 5′-dAdo• for H atom abstraction. Figure: Two proposed mechanism for radical propagation. Path A. 5’-dAdo is formed first followed by the 5′C attaching to the unique Fe generating omega. Path B, omega is formed first by 5’C-S homolysis followed immediately by combination of the 5’-dAdo and the unique Fe. In both pathways, omega is homolytically cleaved to generate the 5’dAdo. 150 (PFL-AE) as previously mentioned (9, 17). LAM was one of the first characterized RS enzymes and catalyzes the reversible isomerization of L-α-lysine to L-ꞵ-lysine (2, 13, 18). After LAM cleaves SAM and generates the 5′-dAdo•, the radical abstracts the ꞵ-hydrogen from lysine bound to pyridoxal 5′-phosphate. The result is 5′-dAdoH and a ꞵ-lysine radical. The ꞵ-lysine radical rearranges to an α-lysine radical which abstracts a hydrogen from 5′- dAdoH. This regenerates the 5′-dAdo• and creates the product, ꞵ-lysine. The 5′-dAdo• can then combine with methionine, with loss of an electron to the [4Fe-4S]2+ cluster, regenerating SAM and the reduced [4Fe-4S]+ cluster. Another α-lysine comes into the active site and the enzyme repeats its catalytic cycle (18, 19). PFL-AE is another founding member of the RS enzyme superfamily and belongs to a subclass of RS enzymes called glycyl radical enzyme activating enzymes (GRE-AEs) (20, 21). GRE-AEs activate their enzyme substrates by abstracting a H atom from a glycine residue generating a glycyl radical (Gly•) (22, 23). In the case of PFL-AE, the 5′-dAdo• removes the pro-S hydrogen from a Gly374 residue on pyruvate formate-lyase (PFL) to form a catalytically essential Gly374• radical (24). This process involves an impressive conformational change in PFL as the N-terminal loop containing Gly734 must move roughly 8 Å to be accessible to the PFL-AE active site (25). The activation of PFL by PFL- AE is a net oxidation, and results in conversion of one equivalent of SAM to methionine and 5′-dAdoH; that is, SAM is a cosubstrate in this reaction. The SAM analog S-3′,4′-anhydroadenosyl-ʟ-methionine (anSAM) has been used to probe reactive intermediates during RS enzyme reactions (18, 26). When functioning in a RS enzyme in place of SAM, anSAM undergoes reductive homolytic cleavage of the S- 151 C5′ bond to form methionine and an allylically stabilized anhydroadenosyl radical (anAdo•) (Figure 5.3). Previous studies by Magnusson et al. demonstrated that anSAM is a true cofactor for LAM, supporting conversion of α-lysine to ꞵ-lysine at a rate roughly 0.25% of that observed with SAM (27). The anAdo• generated during this process builds up sufficiently in LAM to allow characterization using EPR and ENDOR spectroscopy (26, 27). Studies with anSAM and LAM have allowed us to characterize the nature of the LAM active site after anAdo• is formed, where we concluded that the 5′-dAdo• is tightly controlled through van der Waals interactions. In short, “the free radical is never truly free” (18). AnSAM has been studied in detail only with the RS enzyme LAM, which uses SAM as a cofactor; little has been done to study its reaction with other RS enzymes that e- e- SAM anSAM met met 5′-dAdo• anAdo• Figure 5.3. A side by side comparison of SAM and anSAM. Once SAM is cleaved the 5′-dAdo• is generated. When anSAM is cleaved, the allylically stabilized anAdo• is generated where the radical density is delocalized between the 5′, 4′, and 3′ carbons. 152 use SAM as a cosubstrate (28). In this chapter, we report that anSAM can function in place of SAM as a cosubstrate during PFL-AE activation of PFL. Rapid freeze-quenching of this reaction at 500 ms reveals the formation of an analog of Ω, indicating that the anSAM reaction proceeds by an analogous mechanism to that of SAM (16). Further, we show that cryogenic photolysis of anSAM bound to the [4Fe-4S]+ of PFL-AE generates a mixture of anAdo• and the •CH3 due to cleavage of the S-C5′ and S-CH3 bonds of SAM, respectively (5). Our work demonstrates that anSAM serves as a functional cosubstrate for PFL-AE and, like SAM, is subject to photoinduced reductive cleavage. These results lay the groundwork for future studies into the RS enzyme mechanism. Experimental Methods Materials Adenosine 5′-triphosphate sodium salt (ATP), L-methionine, adenylate kinase, phosphocreatine, and phosphocreatine kinase, citrate synthase, malic dehydrogenase were purchased from Sigma. Sodium dithionite (DT) was purchased from Acros Organics. All other reagents and chemicals were obtained in the highest purity from commercial sources. Protein and SAM preparation PFL-AE, PFL, SAM, and anSAM were prepared following the same methods as previously reported (16, 29). PFL-AE Activity Assays Coupled enzyme activity assays were conducted in an anaerobic chamber following 153 previously described methods with minor alterations (29). The activation mix was photoreduced for one hour in a water bath kept between 25 ℃ and 30 ℃. An aliquot (5 µL) of the activation mix was placed on the inner wall of a quartz cuvette containing 895 μL of the coupling mix. The cuvette was inverted to mix the two solutions before monitoring the change in absorbance in 340 nm (resulting from NADH oxidation in the coupled assay) at room temperature (kinetic scan at 340 nm, 0.1 s scan, 1.5 min cycle). Photolysis Samples Photolysis samples were prepared as previously described with minor changes to sample concentrations (5). A sample of 187 μM PFL-AE was reduced with 2.77 mM dithionite prior to the addition of 1.6 mM anSAM. The sample was photolyzed within the cavity of a Bruker EMX X-band EPR spectrometer at 10 K for 80 min with a 450 nm diode laser (Thorlabs). Rapid Freeze-Quench Samples PFL-AE was reduced for RFQ samples by either photoreduction using 5- deazariboflavin or chemical reduction through DT. For photoreduction, samples were prepared as previously described (16). For DT samples, 550 μM PFL-AE and 3.0 mM DT were combined in 50 mM Tris, pH 7.5, 100 mM KCl buffer in an anaerobic COY chamber. Substrate samples were prepared in the same buffer and contained 770 μM PFL, 10 mM oxamate, 3.0 mM DT, and between 3.0 mM and 5.5 mM SAM or anSAM. Rapid Freeze-Quench Experiments 154 All rapid freeze-quench (RFQ) experiments were conducted using a System 100 apparatus from Update Instrument. Before loading samples, the enzyme and substrate sample loops connected to the appropriate syringes were washed with a 100 mM DT solution and dried with N2 gas. Reduced PFL-AE containing a catalytically relevant [4Fe- 4S]+ cluster and a mix of PFL and SAM or PFL and anSAM were loaded into separate loops inside an anaerobic chamber. The protein concentrations used were to attain a 1:1.4 ratio of PFL-AE to PFL after mixing. To guarantee all the protein exited the loop into the mixing chamber and onto the aluminum wheels, a DT buffer (100 mM in water) was placed on each side of the protein solution. A pocket of N2 gas was introduced into the loops between the DT buffer solution and the protein to create the following set-up: DT buffer solution, N2 gas, protein sample, second N2 gas, second DT buffer solution. DT was used to help protect the protein by reacting with any oxygen since the RFQ instrument was not housed in an anaerobic chamber. PFL-AE and (PFL + SAM or anSAM) were rapidly mixed for either 100 ms or 500 ms. The resulting mixture was quenched by spraying onto two rotating aluminum wheels and the frozen powder was collected and packed into precision Q-band tubes (2.5 mm OD). Samples were kept at 77 K until analyzed by EPR. EPR Measurements For Ω samples, a Bruker ESP 300 spectrometer equipped with an Oxford Instruments ESR 910 continuous helium flow cryostat was used to conduct X-band continuous wave EPR spectroscopy. Experimental parameters were 40 K, 1.99 mW microwave power, 100.00 kHz modulation, and 10 G modulation amplitude. 155 All other EPR spectroscopy was conducted on a Bruker EMX X-band (9.73 GHz) spectrometer equipped with an Oxford instruments helium cryostat and temperature controller. For [Fe-S] cluster signals, samples were run at 10 K, 1 mW microwave power, 100 kHz modulation, and 10 G modulation amplitude. Experimental parameters for the anAdo• were 40 K or 75 K, 106 μW-5 mW microwave power, 100.00 kHz modulation, and 2 G or 5 G modulation amplitude. Results PFL-AE uses anSAM to activate PFL Previous studies have shown that anSAM functions as a cofactor during LAM catalysis providing a specific activity 0.25% of what was observed with SAM (27). To see A B if anSAM would serve as a cosubstrate in PFL-AE, we used a coupled enzyme activity assay to compare specific activity in the presence of SAM versus anSAM (29, 30). The activity assay imitates a segment of the citric acid cycle and indirectly quantifies PFL-AE activity by coupling the activation of PFL to the reduction of NAD+, and using UV-vis A B Figure 5.4. Two graphs showing the activity of PFL-AE in the presence of SAM (A) and anSAM (B). The absorbance of NADH is plotted against the reaction time. Activity assay trials were run in triplicate and all absorbances shown were baseline corrected. 156 spectroscopy to monitor NADH catalytically formed by malic dehydrogenase. (22, 31). Each PFL-AE assay was conducted in triplicate (Figure 5.4). The specific activities for PFL-AE in the presence of SAM and anSAM were calculated yielding 120.1 ± 5.6 U/mg and 4.12 ± .32 U/mg respectively. As was the case with LAM, PFL-AE activity is much lower in the presence of anSAM, roughly 3.43% of that seen with natural SAM (26, 27). The reduced activity is expected considering the stabilized anAdo• will likely abstract the H atom from Gly734 at a much slower rate. An Ω-like intermediate forms in the presence of anSAM We have previously shown that the organometallic intermediate Ω forms during catalysis by a range of RS enzyme superfamily members (16, 32). Previous studies have shown the existence of the organometallic 3′,4′-anhydroadenosylcobalamin complex, the 157 anAdo analog of AdoCbl, in B12 enzymes (33). However, it was unclear if an Ω-like complex could form between the unique iron of the [4Fe-4S] cluster of PFL-AE and the 3′,4′-anhydroadenosyl moiety of anSAM. If the role of Ω is to hold the highly reactive 5′- dAdo•, as is speculated, it was possible that this intermediate would not form in the presence of a more stable radical (4, 16). Alternatively, since Ω is central to RS enzyme catalysis as previously stated, it would be expected that a similar organometallic intermediate would form even in the presence of the anSAM analog of SAM (32). To determine whether anSAM follows the same radical initiation pathway as SAM, RFQ techniques were utilized to see if an Ω-like intermediate could be generated in PFL- A C anΩ Ω B Magnetic Field (G) Figure 5.5. The structures of (a) anΩ and (b) Ω. “A” is used to abbreviate the adenine moiety. C) EPR spectrum of anΩ in green and Ω in red. Samples were annealed to 170 K and 150 K respectively. EPR parameters 40 K, 9.37 GHz, 1.99 mW, 10 G. Figure E. HAVE TO REMAKE BUT THIS IS THE GENERAL IDEA! Comparison of the organometallic intermediate generated with anSAM versus unlabeled SAM. (A) The structure of anΩ and (b) Ω. (C) Top, EPR spectrum of anΩ after being annealed to 170 K and Ω, bottom, after annealing to 150 K. Both samples were quenched at 100 ms. EPR parameters 40 K, modulation 100.00 kHz, microwave power 1.99 mW, modulation amplitude 10 G. 158 A AE (32). Reduced PFL-AEC was rapidly mixed with PFL and anSAM before being freeze- quenched (77 K) at 100 ms. The EPR spectrum of the resulting sample showed an axial signal with g-values g|| = 2.035, g⊥ = 2.004. This new signal, anΩ, was nearly identical to B that of Ω with the only notable difference being a lower signal to noise ratio (Figure 5.5). D It is important to note that despite the rapid quench time of 100 ms, no anAdo• was detected in the anΩ spectrum. Photolysis of PFL-AE and anSAM Magnetic Field (G) A B Red and blue-blue (x3) to scale Magnetic Field (G) + + Figure 5.6. [4Fe-4S] cluster of PFL-AE before and after the addition of anSAM. A) the [4Fe-4S] cluster showing + + minor speciation with the [3Fe-4S] cluster. Sample contained 200 µM PFL-AE and 3.0 mM DT. B) the [4Fe-4S] cluster bound to anSAM. The same sample as shown in (A) but with the addition of 1.6 mM anSAM resulting in a final PFL-AE concentration of 187 µM and 2.77 mM DT. EPR parameters were 10 K, microwave frequency = 9.37 GHz, microwave power = 1.0 mW, modulation amplitude = 10 G. Spectrum B was increase 3x for better comparison. 159 Our earlier work established a method for generating the 5′-dAdo• radical through a photoinduced electron transfer from the reduced cluster in PFL-AE to bound SAM (5). When this process was repeated with other RS enzymes, it was found that some follow PFL-AE and cleave the S-C5′ bond releasing 5′-dAdo• and others, including LAM, cleave the S-CH3 to generate a •CH3 radical (34, 35). We carried out photolysis on reduced PFL- AE and anSAM to see if we could generate the anAdo• radical. Figure Y. EPR spectra of the [4Fe-4S] cluster of PFL-AE before and after photolysis. A) top, blue, reduced PFL-AE and anSAMIn. B aontt oamn, agerereonb, riecd uCcOedY PF Lc-hAEa manbd eanr,S APMFaLft-eAr bEe inwg apsh orteodlyuzecde fdor wovietrh a ns ohoduiru. EmPR d piatrhamioenteitrse w (eDreT ) 10 K, 100.00 KHz modulation, 1 mW microwave power, 10 G modulation amplitude. B) radical signal becomes more ands aotubrasteerdv aet d40 wK. iNthew E EPPRR p arsapmeectetrrso wsceroep 4y0 Kt,o 1 0c0.o0n0 fkiHrzm, 5 mthWe mpicrreosweanvec epo woef r a2 Gr medouducleatdio n[ 4amFpeli-t4udSe].+ C A Magnetic Field (G) D B MagnetiGc Field (G) Magnetic Field (G) Figure 5.7. EPR spectra of the [4Fe-4S]+ cluster of PFL-AE before and after photolysis. A) reduced PFL- AE and anSAM. B) reduced PFL-AE and anSAM after being photolyzed for 1 hr 20 min at 10 K. EPR parameters were T = 10 K, microwave frequency = 9.37 GHz, microwave power = 1.0 mW, modulation amplitude = 10 G. C) radical signal becomes more resolved at 40 K. D) radical signal after being annealed at 75 K for thirty minutes. New EPR parameters were T = 40 K, microwave frequency = 9.37 GHz, microwave power = 5 mW, modulation amplitude = 2 G. 160 cluster (Figure 5.6a). The sample was returned to the COY chamber and anSAM was added before refreezing the sample in liquid nitrogen. After confirming the presence of bound anSAM with EPR, (Figure 5.6b and 5.7a), the sample was irradiated for an hour and twenty minutes at 450 nm in the EPR cavity at 10 K (30). The EPR spectrum of the resulting sample showed a significantly reduced [4Fe-4S]+ + anSAM cluster signal with a new radical species growing in (Figure 5.7b). At 40 K, the new radical signal was better resolved and appeared to be a mix of two overlapping radical signals (Figure 5.7c). The sample was then annealed at 75 K for 30 minutes and the EPR spectrum was recorded at 40 K. Further changes were evident in the appearance of the EPR spectrum (Figure 5.7d), indicating one of the signals may have decayed away. After being stored at 77 K in a liquid nitrogen dewar for 18 hr, the EPR spectrum of the sample recorded at 40 K did not change much (Figure 5.8). 161 In order to deconvolute the components of the EPR spectrum in figure 5.7c, the spectrum in Figure 5.8a and 5.8b were subtracted from the EPR spectrum in Figure 5.7c. The results of this subtraction is are displayed in Figures 5.9 and 5.10, which show that the component that decayed away on annealing at 75 K and 77 K has a distinct 1:3:3:1 intensity SI figure 2. EPR spectra of the anAdo• before and after sitting in the liquid nfiotruor-gleinne dpeawttearrn ocvhearrancitgehrits.t Tico pof, ba lmuee,t hsyigl nradl iacfatle (r3 b4e, i3n6g). kFeipgtu raets 755.9 K a fnodr 5th.1i0rt syh ow a minutes. Bottom, green, signal of same sample after being kept at 77 K for 18 hr. EcPoRm ppaarraismone toef rtsh wis e•rCeH 430 r Kad, i1c0al0 s.0ig0n KalH czompoadruedla ttoio tnh,a t5 o mbtWainmedi cirno ewaarlvieer psotuwdieers, w5 ith G modulation amplitude. HydG (34). These results show that one of the components in 5.7c is due to a •CH3 radical, which decays away on annealing at 75 K, which is consistent with what was observed in photolysis studies of HydG (30). A B + 18 hr Magnetic Field (G) Figure 5.8. EPR spectra of the anAdo• before and after sitting in a liquid nitrogen dewar overnight (18 hr). A) the signal after being kept at 75 K for thirty minutes. B) the signal of the same sample after being kept at 77 K for 18 hr. EPR parameters were 40 K, microwave frequency = 9.37 GHz, microwave power = 5 mW, modulation amplitude = 2 G. 162 The EPR signal that remained after annealing the sample at 75 K for 30 minutes (Figures 5.7d and 5.8a) may be due the anAdo• radical. To confirm this, simulations of the EPR signal in figure 5.8a were generated using EasySpin for comparison (Figure 5.11). The EPR parameters from previous work by Magnusson and Frey were used to generate SIM 1 (figure 5.11b) (26, 27). SIM 2 shows the 2′-H aligned nearly parallel to the pz orbital containing the unpaired electron (figure 5.11c). As a result, there is a greater coupling observed from the 2′-1H (figure 5.7d). The splitting from SIM 2 matches the splitting observed in the anAdo• signal. It appears the protein environment of PFL-AE results in D E Magnetic Field (G) Figure 5.9. Deconvolution of the radical signal observed immediately after photolysis. A) the radical observed at 40 K before annealing. B) the signal observed at 40 K after being annealed at 77 K for 18 hr. C) the methyl radical signal resulting from subtraction of signal B from signal A. EPR conditions for a and b are 40 K, 9.37 GHz, 5 mW, 2 G. Right, comparison of the EPR signals of the •CH observed in HydG (D) and PFL-AE (E). D) Previously 3 published •CH signal observed in 100 μM HydG, 3.0 mM DT, and 5.5 mM SAM photolyzed with 450 nm light 3 at 12 K. EPR parameters are 40 K, 9.38 GHz, 5 mW (34). E) •CH signal generated as described in (C). 3 163 different constraints on the anAdo• than what is observed in LAM. As was observed by Magnusson, et al. the anAdo• radical EPR signal undergoes a noticeable change in appearance when recorded at different temperatures (26). The EPR signal at 40 K displays hyperfine coupling that can be attributed to proton interactions from the two 5′H1 nuclei, the 3′H1 nucleus and the 2′H1 nucleus with the delocalized, unpaired electron (figure 5.11a and figure 5.12a). When the anAdo• EPR spectrum was recorded at 75 K, a near isotropic signal centered at g = 2.005 was visible consistent with the presence of a carbon centered radical (figure 5.12b). The signal at 75 K decreased in intensity and lost much of the hyperfine coupling that could be observed at 40 K (Figure 5.12). The loss A D B E C Magnetic Field (G) Magnetic Field (G) Figure 5.10. Deconvolution of the radical signal observed immediately after photolysis. A) the radical observed at 40 K before annealing, same as figure 5.7c. B) the signal observed at 40 K after being annealed at 75 K for 30 min. C) the methyl radical signal resulting from subtraction of signal B from signal A. EPR conditions for a and b are 40 K, 9.37 GHz, 5 mW, 2 G. D) previously published •CH3 signal as shown in figure 5.9d (34). E) •CH3 signal generated as described in (C). 164 D AnAdo• A1 A2 A3 30 min 77K anneal 11 A Simulation 1 5′-Hexo 66 24 45 5′-Hendo 54 18 36 B 2’-H 3′-H 60 21 42 2′-H 51 51 51 C Simulation 2 5′-Hexo 66 24 45 5′-Hendo 54 18 36 3′-H 60 21 42 Magnetic Field (G) 2′-H 91 91 91 Figure 1. EPR spectra of the anAdo• at 75 K varying the microwave power. Top, Figure 5.11. Simulations of the ian ngrAeedn,o si•g nianl r uPn FatL 5 -mAWE. M icddolem, ipn ared to experimental data. A) anAdo• after annealing at 77 K for 18 hr. EPRbl upe,a sHere is my quick assessmriganmal aett 2e mrsW . Bottom, red, signal ent about a posTsib le= a n4Ad0o rKadic,a lm. Thiec orroigiwnala pavraem eftreers qfroume Fnrecy'ys p a=pe r9 fo.3r t7he sGimHulatsio,n mis sihcorrt ofw EPaR vbrea dth (SIM 1). I increased the at 104 μW. EPR parameters were 75 K, hyperfine coupling of 2'-H to [91,91,91] MHz, but keep the other hyperfine parameters the same (SIM 2) matches up the experimental EPR breadth as well the peak pattern. I power = 5 mW, modulatthiinok nit iss paomssibple1l 0tih0te.u 0p0rdo KteeHi n=z em n2voir doGunlma. teBinotn )o,f 2sp fiGl-ma me ucoodlnaustltraaitioinotn a n 1Ad obraadsiceald d ifofefrefn tolyf fr opma LrAaMm thaet tceaursses d2'e-Hs acligrni bneeardly pianre l(le2l t4o t)h.e pz orbital of the unpaired electron C) simulation 2 with inacnrd egiaves tehed b ig2ga′e-rm Hcpoul ipthulindygep o. fe 2r'-fHi. nThee dcetoeruiupmllaibnegled. eDxpe)r imHenyt pshealrl gfivien tehe Tmoeren dseofinritsiv e( aMnswHer.z) of anAdo•. Both Cheers, (B) and (C) were generHaatoed using g-values [2.005, 2.003, 2.002]. --- 5 mW --- 2 mW A --- 104 μW 40 K B 75 K Magnetic Field (G) Magnetic Field (G) Figure 5.12. EPR spectra of the anAdo• Figure 5.13. EPR spectra of the anAdo• observed observed at 40 K (A) and 75 K (B). The 40 K at 75 K varying the microwave power. Top, in spectra was recorded after annealing at 75 K for green, signal run at 5 mW. Middle, in blue, signal 30 min and the intensity was reduced by half in at 2 mW. Bottom, red, signal at 104 μW. EPR order to better compare with signal (B). EPR parameters were 75 K, microwave frequency = parameters were 9.37 GHz, 5 mW, 2 G. 9.37 GHz, microwave power = 5 mW, modula- tion amplitude = 2 G. 165 of hyperfine coupling is proposed to result from a temperature dependent change in the electron-electron spin couplings between the paramagnetic [4Fe-4S]2+ cluster and the unpaired electron on the anAdo•. At temperatures between 40 K and 75 K, it is possible the paramagnetic excited states of the [4Fe-4S]2+ cluster can become appreciably populated. This leads to increased spin-spin coupling between the paramagnetic cluster and the radical complicating the anAdo• spectrum and leading to the loss of observable hyperfine coupling (26, 37). The radical signal continued to decrease in intensity as the microwave power decreased becoming almost unresolved at 106 µW (figure 5.13). The results presented here demonstrate that photolysis of the PFL-AE [4Fe- 4S]+/anSAM complex at 10 K results in a mixture of anAdo• and •CH3 radical products. Comparison of the signal intensities of the anAdo• and •CH3 radical signals of Figure 5.9 indicate that 44% anAdo• radical and 52% •CH3 are produced on photolysis (figure 5.14). These results show that cleavage of the S-CH3 bond is preferred slightly over the S-5′C bond when anSAM is photolyzed in the PFL-AE active site (figure 5.14) (38). A B Figure 5.14. A) a schematic of anSAM reductively cleaving to the anAdo• or •CH3 upon the addition of an electron. B) comparison of signal intensities of the radical species observed in the EPR signals shown in figure 5.9. Percentages were based off the double integral of the mixed radical signal. 166 Discussion RS enzymes play an essential role in catalyzing over 80 types of reactions in all domains of life (1, 2, 6). All RS enzymes use a [4Fe-4S] cluster and SAM as either a cofactor or a cosubstrate to generate the intermediates Ω and 5′-dAdo•; the latter of which is used to abstract a hydrogen from substrate (9, 16, 39). Absent from the RS enzyme mechanism are mechanistic details involving Ω and 5′-dAdo• formation. Studying the formation of these intermediates is challenging due in part to the rapid formation of Ω and the high reactivity of the 5′-dAdo•. In our work, we show that anSAM can function as a cosubstrate in PFL-AE providing a powerful tool for further studies in the RS enzyme mechanism. A pressing question regarding the RS mechanism is does omega form through a stepwise process where the 5′-dAdo• forms first then combines with the unique Fe generating Ω or does Ω form first through a concerted mechanism? Previous RFQ time course studies with reduced PFL-AE rapidly mixed with (PFL + SAM) showed Ω formation as rapid as 2 μs (5). However, these experiments were not definitive. The 5′- dAdo• is simply too reactive and it is not unreasonable to assume it could combine with the unique Fe at the μs time scale (5). The anAdo•, on the other hand, is more stable than a primary C radical (27). The spin delocalization across the 5′, 4′, and 3′ carbons in anAdo• decreases the reactivity of the C5′ radical. Additionally, the double bond introduces steric constraints to the ribose ring that prevent cyclization. Because the five-membered ring is 167 flatter than what is observed in the 5′-dAdo•, the C5′ cannot readily attack the 8C on the adenine ring (33, 40). When anSAM is used as a cofactor in LAM, the anAdo• is able to build up sufficiently to be observed with EPR (26, 27). It is interesting that when reduced PFL-AE is rapidly mixed with (PFL + anSAM) and quenched at 100 ms, the resulting EPR spectrum does not contain an anAdo• signal. If anΩ formed through a stepwise mechanism, we would expect to see traces of the anAdo• signal as not all the radical would react in 100 ms. Our RFQ experiments and EPR spectroscopy seem to indicate that anΩ is formed through a concerted mechanism where the S-C5′ bond is cleaved followed by immediate formation of the Fe-C5′ bond. Recent work from the Broderick Lab shows that photoinduced electron transfer (ET) from the [4Fe-4S]+ cluster to SAM can produce the •CH3 instead of the catalytically relevant 5′-dAdo• in certain RS enzymes (20). The SAM ribose conformation in the RS enzyme active site is proposed to result in localized Jahn Teller distortions responsible for the regioselectivity of cryogenic photoinduced reductive S-C bond cleavage. In short, the position of the SAM ribose ring will determine if the 5′-dAdo• or the •CH3 is formed (35). Photoinduced ET in PFL-AE leads to the reductive cleavage of the S-C5′ bond of SAM generating the 5′-dAdo• (5, 35). It was surprising to observe that when reduced PFL-AE and anSAM were photolyzed, the EPR spectra showed a mix of radical signals. Deconvolution of the spectrum revealed what appeared to be signals from the anAdo• and the •CH3 radicals. Comparison of the •CH3 in PFL-AE to that generated by photolysis of reduced HydG confirmed the S-CH3 bond in anSAM was cleaved during photolysis (34). 168 Simulations of the anAdo• in the PFL-AE active site showed splitting analogous to what was observed in the anAdo• EPR spectrum. We therefore concluded the •CH3 and anAdo• were generated in a sample of reduced PFL-AE and anSAM. This is likely a result of the ribose conformation in anSAM. As previously mentioned, the C4′-C3′ double bond of anSAM causes a loss of ribose ring puckering (22). As a result, the S of anSAM undergoes different Jahn Teller distortions than the S of SAM in the PFL-AE active site allowing for the photoinduced cleavage of the S-C5′ or the S-CH3 bond. Our results show anSAM serves as an effective tool to study the RS mechanism in enzymes that used SAM as a cosubstrate. PFL-AE can use anSAM as a cosubstrate to activate PFL. PFL-AE can also use anSAM to generate an Ω analog, anΩ. Photolysis of reduced PFL-AE and anSAM shows for the first time a RS enzyme that can generate the anAdo• and •CH3. AnSAM has and continues to be a valuable and versatile SAM analog. 169 References 1. Sofia HJ, Chen G, Hetzler BG, Reyes-Spindola JF, Miller NE. Radical SAM, a novel protein superfamily linking unresolved steps in familiar biosynthetic pathways with radical mechanisms: functional characterization using new analysis and information visualization methods. Nucleic acids research. 2001;29(5):1097-106. 2. Broderick JB, Duffus BR, Duschene KS, Shepard EM. Radical S- adenosylmethionine enzymes. Chemical reviews. 2014;114(8):4229-317. 3. Frey PA, Hegeman AD, Ruzicka FJ. The radical SAM superfamily. Critical reviews in biochemistry and molecular biology. 2008;43(1):63-88. 4. Brown AC, Suess DL. Reversible formation of alkyl radicals at [Fe4S4] clusters and its implications for selectivity in radical SAM enzymes. Journal of the American Chemical Society. 2020;142(33):14240-8. 5. Yang H, McDaniel EC, Impano S, Byer AS, Jodts RJ, Yokoyama K, Broderick WE, Broderick JB, Hoffman BM. The elusive 5′-deoxyadenosyl radical: captured and characterized by electron paramagnetic resonance and electron nuclear double resonance spectroscopies. Journal of the American Chemical Society. 2019;141(30):12139-46. 6. Bridwell-Rabb J, Grell TA, Drennan CL. A rich man, poor man story of s- adenosylmethionine and cobalamin revisited. Annual Review of Biochemistry. 2018;87:555-84. 7. Henshaw TF, Cheek J, Broderick JB. The [4Fe-4S]1+ cluster of pyruvate formate- lyase activating enzyme generates the glycyl radical on pyruvate formate-lyase: EPR- detected single turnover. Journal of the American Chemical Society. 2000;122(34):8331- 2. 8. Duschene KS, Veneziano SE, Silver SC, Broderick JB. Control of radical chemistry in the AdoMet radical enzymes. Current opinion in chemical biology. 2009;13(1):74-83. 9. Vey JL, Drennan CL. Structural insights into radical generation by the radical SAM superfamily. Chemical Reviews. 2011;111(4):2487-506. 10. Moss M, Frey P. Activation of lysine 2, 3-aminomutase by S-adenosylmethionine. Journal of Biological Chemistry. 1990;265(30):18112-5. 11. Cheek J, Broderick JB. Direct H atom abstraction from spore photoproduct C-6 initiates DNA repair in the reaction catalyzed by spore photoproduct lyase: evidence for a 170 reversibly generated adenosyl radical intermediate. Journal of the American Chemical Society. 2002;124(12):2860-1. 12. McCarty RM, Krebs C, Bandarian V. Spectroscopic, steady-state kinetic, and mechanistic characterization of the radical SAM enzyme QueE, which catalyzes a complex cyclization reaction in the biosynthesis of 7-deazapurines. Biochemistry. 2013;52(1):188- 98. 13. Frey P. Lysine 2, 3-aminomutase: is adenosylmethionine a poor man's adenosylcobalamin? The FASEB journal. 1993;7(8):662-70. 14. Moss M, Frey PA. The role of S-adenosylmethionine in the lysine 2, 3- aminomutase reaction. Journal of Biological Chemistry. 1987;262(31):14859-62. 15. Halpern J, Kim SH, Leung T. Cobalt-carbon bond dissociation energy of coenzyme B12. Journal of the American Chemical Society. 1984;106(26):8317-9. 16. Byer AS, Yang H, McDaniel EC, Kathiresan V, Impano S, Pagnier A, Watts H, Denler C, Vagstad AL, Piel Jr. Paradigm shift for radical S-adenosyl-L-methionine reactions: The organometallic intermediate Ω is central to catalysis. Journal of the American Chemical Society. 2018;140(28):8634-8. 17. Frey PA, Magnusson OT. S-Adenosylmethionine: a wolf in sheep's clothing, or a rich man's adenosylcobalamin? Chemical reviews. 2003;103(6):2129-48. 18. Horitani M, Byer AS, Shisler KA, Chandra T, Broderick JB, Hoffman BM. Why nature uses radical SAM enzymes so widely: Electron nuclear double resonance studies of lysine 2, 3-aminomutase show the 5′-dAdo•“free radical” is never free. Journal of the American Chemical Society. 2015;137(22):7111-21. 19. Chen D, Frey PA. Identification of lysine 346 as a functionally important residue for pyridoxal 5 ‘-phosphate binding and catalysis in lysine 2, 3-aminomutase from Bacillus subtilis. Biochemistry. 2001;40(2):596-602. 20. Knappe J, Blaschkowski H-P, Edenharder R. Enzyme-dependent activation of pyruvate formate-lyase of Escherichia coli. Metabolic Interconversion of Enzymes: Springer; 1972. p. 319-29. 21. Rödel W, Plaga W, Frank R, Knappe J. Primary structures of Escherichia coli pyruvate formate‐lyase and pyruvate‐formate‐lyase‐activating enzyme deduced from the DNA nucleotide sequences. European journal of biochemistry. 1988;177(1):153-8. 22. Conradt H, Hohmann-Berger M, Hohmann H-P, Blaschkowski HP, Knappe J. Pyruvate formate-lyase (inactive form) and pyruvate formate-lyase activating enzyme of 171 Escherichia coli: isolation and structural properties. Archives of biochemistry and biophysics. 1984;228(1):133-42. 23. Broderick JB, Duderstadt RE, Fernandez DC, Wojtuszewski K, Henshaw TF, Johnson MK. Pyruvate formate-lyase activating enzyme is an iron− sulfur protein. Journal of the American Chemical Society. 1997;119(31):7396-7. 24. Frey M, Rothe M, Wagner A, Knappe J. Adenosylmethionine-dependent synthesis of the glycyl radical in pyruvate formate-lyase by abstraction of the glycine C-2 pro-S hydrogen atom. Studies of [2H] glycine-substituted enzyme and peptides homologous to the glycine 734 site. Journal of Biological Chemistry. 1994;269(17):12432-7. 25. Peng Y, Veneziano SE, Gillispie GD, Broderick JB. Pyruvate formate-lyase, evidence for an open conformation favored in the presence of its activating enzyme. Journal of Biological Chemistry. 2010;285(35):27224-31. 26. Magnusson OT, Reed GH, Frey PA. Characterization of an allylic analogue of the 5‘-deoxyadenosyl radical: An intermediate in the reaction of lysine 2, 3-aminomutase. Biochemistry. 2001;40(26):7773-82. 27. Magnusson OT, Reed GH, Frey PA. Spectroscopic evidence for the participation of an allylic analogue of the 5‘-deoxyadenosyl radical in the reaction of lysine 2, 3- aminomutase. Journal of the American Chemical Society. 1999;121(41):9764-5. 28. Brindley AA, Zajicek R, Warren MJ, Ferguson SJ, Rigby SE. NirJ, a radical SAM family member of the d1 heme biogenesis cluster. FEBS letters. 2010;584(11):2461-6. 29. Byer AS, McDaniel EC, Impano S, Broderick WE, Broderick JB. Mechanistic Studies of Radical SAM Enzymes: Pyruvate Formate-Lyase Activating Enzyme and Lysine 2, 3-Aminomutase Case Studies. Methods in enzymology: Elsevier; 2018. p. 269- 318. 30. Shisler KA, Hutcheson RU, Horitani M, Duschene KS, Crain AV, Byer AS, Shepard EM, Rasmussen A, Yang J, Broderick WE. Monovalent cation activation of the radical SAM enzyme pyruvate formate-lyase activating enzyme. Journal of the American Chemical Society. 2017;139(34):11803-13. 31. Brush EJ, Lipsett KA, Kozarich JW. Inactivation of Escherichia coli pyruvate formate-lyase by hypophosphite: evidence for a rate-limiting phosphorus-hydrogen bond cleavage. Biochemistry. 1988;27(6):2217-22. 32. Horitani M, Shisler K, Broderick WE, Hutcheson RU, Duschene KS, Marts AR, Hoffman BM, Broderick JB. Radical SAM catalysis via an organometallic intermediate with an Fe–[5′-C]-deoxyadenosyl bond. Science. 2016;352(6287):822-5. 172 33. Magnusson OT, Frey PA. Synthesis and Characterization of 3 ‘, 4 ‘- Anhydroadenosylcobalamin: A Coenzyme B12 Analogue with Unusual Properties. Journal of the American Chemical Society. 2000;122(37):8807-13. 34. Yang H, Impano S, Shepard EM, James CD, Broderick WE, Broderick JB, Hoffman BM. Photoinduced electron transfer in a radical SAM enzyme generates an S- adenosylmethionine derived methyl radical. Journal of the American Chemical Society. 2019;141(40):16117-24. 35. Impano S, Yang H, Jodts RJ, Pagnier A, Swimley R, McDaniel EC, Shepard EM, Broderick WE, Broderick JB, Hoffman BM. Active-site controlled, Jahn-Teller enabled regioselectivity in reductive S-C bond cleavage of S-adenosylmethionine (SAM) in radical-SAM enzymes. 2020. 36. McKenzie I, Brodovitch J-C, Ghandi K, McCollum BM, Percival PW. Hyperfine coupling in methyl radical isotopomers. The Journal of Physical Chemistry A. 2007;111(42):10625-34. 37. Bertini I, Ciurli S, Luchinat C. The electronic structure of FeS centers in proteins and models a contribution to the understanding of their electron transfer properties. Iron- Sulfur Proteins Perovskites: Springer; 1995. p. 1-53. 38. Svistunenko DA, Sharpe MA, Nicholls P, Wilson MT, Cooper CE. A new method for quantitation of spin concentration by EPR spectroscopy: application to methemoglobin and metmyoglobin. Journal of Magnetic Resonance. 2000;142(2):266-75. 39. Frey PA. Radical mechanisms of enzymatic catalysis. Annual review of biochemistry. 2001;70(1):121-48. 40. Hogenkamp H. A cyclic nucleoside derived from coenzyme B12. Journal of Biological Chemistry. 1963;238(1):477-80. 173 CHAPTER SIX CONCLUSION Since their classification as a superfamily in 2001, RS enzymes have grown in prominence among researchers around the world (1, 2). Their ability to act on a variety of substrates and their essential involvement in many difficult biological actions such as DNA repair, hydrogenase maturation, and anaerobic glucose metabolism make these enzymes advantageous in both industrial and medical field applications (2, 3). All radical SAM enzymes use a [4Fe-4S] cluster and SAM to generate the 5′-dAdo• necessary to abstract a hydrogen from substrate (2). We find that if RS enzymes are to be useful in biotechnical and biomedical applications, their radical initiation mechanism needs to be better understood. The discovery and characterization of an organometallic intermediate in PFL-AE by the Broderick Lab has shown that the RS enzyme mechanism differs from what was previously hypothesized (4). Instead of a direct formation of the 5′-dAdo• and subsequent H atom abstraction from PFL, an organometallic intermediate forms prior to the liberation of 5′-dAdo• (1, 4, 5). The organometallic intermediate, Ω, forms after SAM is cleaved and before the Gly734• is generated. The discovery of the organometallic intermediate led to the question is Ω unique to PFL-AE or is it universal for the RS enzyme superfamily. To determine if Ω is generated throughout the RS enzyme superfamily, RFQ and EPR spectroscopy were used to observe Ω in multiple RS enzymes. 174 To determine if Ω was a part of the RS reaction pathway and not a sample preparation artifact, the intermediate was generated using three different mixing conditions. Reduced PFL-AE was mixed with (PFL/SAM), (reduced PFL-AE/SAM) was mixed with PFL, and (reduced PFL-AE/PFL) was mixed with SAM, creating three samples. All reactions were rapidly quenched at 500 ms and examined with EPR spectroscopy. Despite the mixing order, each sample generated the characteristic Ω signal with only minor changes to the g║ feature. These results allowed us to conclude that Ω is a central intermediate in PFL-AE and forms regardless of the order of mixing (6). Next, Ω was shown to form in a variety of RS enzymes. RFQ experiments used to generate Ω in PFL-AE were repeated with seven RS enzymes selected to represent the superfamily’s diverse enzymatic functions. The two major RS enzyme subclasses were represented with enzymes that use SAM as a cofactor (LAM and SPL) and those that use SAM as a cosubstrate (PFL-AE, RNR-AE, HydG, PoyD, and OspD). Further representing the diversity of these enzymes, HydG uses an auxiliary Fe-S cluster in catalysis, PoyD and OspD catalyze epimerization reactions, and PFL-AE and RNR-AE activate GREs. Additionally, the seven selected enzymes catalyze reactions on a variety of substrates including macromolecules, peptides, and small molecules. Each enzyme was reduced via photoreduction with 5-deazariboflavin to generate an active [4Fe-4S]+ cluster. The reduced enzyme was rapidly mixed with the appropriate substrate and SAM before being freeze- quenched at 500 ms. The frozen samples were observed using EPR spectroscopy, revealing that each enzyme produced the axial signal characteristic for Ω (4, 7). These results show 175 Ω is formed throughout a diverse group of RS enzymes and indicate that Ω is ubiquitous among the RS superfamily (6). Because Ω appears to be a ubiquitous intermediate in RS enzymes, it was important to refine the previous determination of the Ω structure. EPR and ENDOR spectroscopies were used for this purpose. When Ω was generated with 57Fe-enriched PFL-AE, the resulting EPR signal showed 57Fe line-broadening, confirming the spin of Ω must be on the [4Fe-4S] cluster. To verify if Ω retains the methionine fragment of SAM, the intermediate was generated with 14/15N-methyl-SAM and observed with ENDOR spectroscopy. When Ω was formed with unlabeled SAM, the resulting 14N ENDOR signal was characteristic for natural abundance 14N directly coordinating to Fe. In samples prepared with 15N-methyl-SAM, the ENDOR spectra showed no sign of 14N-Fe coordination. Instead, the spectra revealed a 15N ENDOR signal congruent with Fe directly coordinating to 15N. These ENDOR results show that the methionine fragment of SAM remains coordinated through its amino group to the [4Fe-4S] cluster (7). ENDOR spectroscopy was also used with [D8-ado]-SAM and [5′,5″-D2-ado]-SAM to confirm the deoxyadenosine (dAdo) fragment of SAM is incorporated in Ω. When Ω was formed with [D8-ado]-SAM there was a decrease in the 1H ENDOR signal. The same 1H ENDOR signal loss was seen in Ω samples generated with [5′,5″-D2-ado]-SAM, demonstrating that the H5′ and H5″ on the dAdo fragment of SAM were interacting with the unique Fe. It was conclude that the Fe-C bond of Ω is formed by the C5′ of the dAdo fragment of SAM (6). The Ω intermediate is structurally similar to AdoCbl used in B12 enzymes. Both intermediates cleave a M-C5′ bond to liberate the 5′-dAdo• (4, 8, 9). Photolysis studies 176 with AdoCbl showed that the Co(III)-C5′ bond could be photolytically cleaved, releasing the 5′-dAdo• (7, 10). These studies inspired the idea that photolyzing Ω could release the 5′-dAdo•. Before attempting to photolyze Ω, a control experiment was conducted where reduced PFL-AE and SAM were photolyzed in the absence of PFL (7). Because sulfoniums have been shown to be photochemically active, it was proposed that photoinitiated homolysis of the S-C5′ bond could generate the 5′-dAdo• (7, 11, 12). Anaerobically prepared samples of reduced PFL-AE and SAM were irradiated in the EPR spectrometer cavity at 12 K with 450 nm light. The resulting EPR spectrum showed a well resolved, anisotropic radical signal at 40 K. The spectrum displayed 1H hyperfine couplings typical of the Cα-1H radical with trigonal planar geometry. A photolysis time course experiment was conducted which showed the loss of the [4Fe-4S]+ + SAM signal directly correlates with the appearance of the 5′-dAdo• signal. This correlation was consistent with the idea that a photoinduced electron transfer from the [4Fe-4S]+ cluster to SAM generates the radical species (7, 13). To confirm the observed radical species was in fact the 5′-dAdo•, photolysis experiments using reduced PFL-AE and isotopically labeled SAM were carried out. When the radical species was generated using [adenosyl-5′-5″-D2]-SAM, the EPR spectrum showed a collapse of the hyperfine coupling to a doublet. Similarly, using [adenosyl-2,8- D2-1′,2′,3′,4′,5′,5″-D6]-SAM in sample preparation collapsed the EPR spectrum into a singlet. Both signals showed the hyperfine coupling in the radical spectra results from the H5′, H5″, and H4′ on the dAdo portion of SAM interacting with the unpaired electron. In 177 samples prepared with 5′-13C SAM, additional splitting in the radical spectrum was observed. This increased splitting followed the predicted pattern of a 5′-13Cα interacting with a free radical. Additionally, the spectra of samples prepared with [adenosyl- 13C 15 10, N5]-SAM showed additional hyperfine splitting resulting from interactions with the 4′-13Cꞵ. These EPR results confirm the capture of the long-elusive 5′-dAdo• (7). The observed hyperfine interactions provided information on the electric and geometric structure of 5′-dAdo•. The 5′-13Cα and two 1Hα coupling tensors agree with those of a planar sp2 carbon centered radical that houses the unpaired electron in a 2pπ orbital (14). The isotropic couplings observed in the EPR spectrum indicate the 2pπ orbital belongs to the 5′C of 5′-dAdo•. Using the measurements of the 1H-C4′ isotropic coupling and the relationship a ≈ ρ Bcos2 iso π φ MHz, the dihedral angle (φ) between 4′CH-4′C-5′C was calculated to be φ = 37° showing that the H2-C5′-C4′ fragment of the 5′-dAdo• is predominantly planar (7, 15). The structure implied by the hyperfine couplings seen in the 5′-dAdo• EPR spectra was further supported by DFT computations (16). The computations generated an energy- minimized structure that reproduced hyperfine couplings nearly identical to those seen in the 5′-dAdo• spectrum. The computed 5′-dAdo• structure exhibits a planar geometry at the C5′ as predicted by the hyperfine interactions observed in the radical EPR spectrum. Moreover, the computed structure also showed a dihedral “twist” at the C5′-C4′ bond with a dihedral angle φ = 39° which is nearly identical to what was calculated using the 1H-C4′ coupling (7). 178 Additional SAM isotopologues were used to give a sense of the relationship of the 5′-dAdo• and methionine fragments of SAM after homolysis. When the 5′-dAdo• was generated with [3,3,4,4-methionine-D4]-SAM, subtle changes in the shape of the EPR spectrum were observed which indicate weak hyperfine couplings between the protons of the methionine side chain and the C5′ radical contribute to the line width of the natural- abundance spectrum. There are no changes in the EPR spectrum when the 5′-dAdo• is prepared from CD3-methyl or 13C-methyl methionine SAM indicating that the radical is remote from the methyl group. These observations indicate that after SAM S-C5′ homolysis, the C5′ radical shifts away from the methyl and toward the methionine C3 and C4 (7). Although the structure of Ω has been studied in detail, the mechanism for its formation in RS enzymes is not fully understood. One possible mechanism involves a reductive cleavage of SAM to produce the 5′-dAdo•, which then combines with the unique Fe of the [4Fe-4S]2+ cluster to form Ω and the [4Fe-4S]3+ cluster. The Fe-C5′ bond is then cleaved, releasing 5′-dAdo• for H atom abstraction. Alternatively, Ω could form through concerted reductive cleavage of the S-C5′ bond followed by immediate formation of the Fe-C5′ bond generating Ω. Determining which mechanism is used by RS enzymes is challenging due in part to the rapid formation of Ω and the high reactivity of the 5′-dAdo• (7). My final chapter lays the groundwork for future studies into the RS mechanism using a SAM analog, anSAM, and PFL-AE. AnSAM has been shown to be a true cofactor for LAM and has been used to characterize the nature of the LAM active site after anAdo• is formed (17-19). Although 179 extensive work has been done with anSAM serving as a cofactor to LAM, it was not known if anSAM could also serve as a cosubstrate in PFL-AE. When anSAM was used in place of SAM, PFL-AE was still able to activate PFL in coupled enzyme activity assays showing that anSAM can function as a cosubstrate. The specific activity of PFL-AE was greatly decreased in the presence of anSAM a phenomenon which was also observed by Magnusson et al. when conducting activity assays with LAM and anSAM (18). RFQ experimental conditions used to generate Ω were repeated using PFL-AE, PFL, and anSAM to generate an Ω-like intermediate, anΩ. Reduced PFL-AE was rapidly mixed with (PFL/anSAM) and quenched at 100 ms. The resulting anΩ EPR spectrum showed a signal identical to that of Ω and did not show any anAdo• signal. Because the anAdo• is significantly more stable than the 5′-dAdo•, the EPR spectrum would in theory show a mix of anAdo• and anΩ signals if the anAdo• formed first then combined with the unique Fe to form anΩ (18). Therefore, it appears that anΩ forms through concerted reductive cleavage involving the homolysis of the S-C5′ bond and immediate formation of the Fe-C5′ bond (6). The anAdo• is not formed until after the Fe-C5′ bond of anΩ is cleaved. To see if the anAdo• could be generated using photolysis, anaerobically prepared, reduced PFL-AE and anSAM were irradiated at 12 K with blue light for 80 min. The resulting EPR spectrum showed a mix of radical signals that were shown to result from the anAdo• and the •CH3. These findings support recent work from the Broderick Lab which suggests SAM ribose conformations in the RS enzyme active site result in localized Jahn Teller distortions which are responsible for the regioselectivity of cryogenic photoinduced 180 reductive S-C bond cleavage (20). The C4′-C5′ double bond in anSAM causes its ribose ring to be flatter than the ribose of SAM (21). This difference in the ribose ring geometry causes the S of anSAM to undergo different Jahn Teller distortions than the S of SAM in the PFL-AE active site. This effectively removes the regioselectivity for the S-C5′ bond observed in photolyzed samples of PFL-AE and SAM and allows for the photolytic cleavage of the S-C5′ and S-CH3 bonds of anSAM. This is the first time photolysis of a RS enzyme and SAM analog has generated the anAdo• and •CH3 simultaneously. The RS enzyme superfamily continues to surprise researchers with its variety of substrates and fundamental chemistry of radical initiation. One of the earliest characterized RS enzymes was PFL-AE: a small protein that has contributed to our knowledge of the RS mechanism in significant ways. Work with PFL-AE revealed the catalytically active state of the cluster is the [4Fe-4S]+ state which becomes oxidized to [4Fe-4S]2+ after SAM is cleaved (22). ENDOR spectroscopy studies of PFL-AE indicate that the amino and carboxyl groups of SAM directly coordinate the unique Fe of the [4Fe-4S] cluster and that reduction of SAM by the [4Fe-4S]+ cluster is a result of an inner-sphere electron transfer from the reduced cluster to the sulfonium of SAM (23-25). More recently, RFQ experiments and EPR analysis revealed the presence of a catalytically competent organometallic intermediate, Ω, generated in PFL-AE (4). My work with PFL-AE indicates there is still much to be learned about RS enzymes. By repeating RFQ experiments originally used to capture Ω in PFL-AE, I have shown the intermediate is ubiquitous in seven RS enzymes which have diverse enzymatic functions representative of the entire enzyme superfamily. Using EPR and ENDOR spectroscopies has allowed Ω to be further 181 characterized (6). A unique photolysis experiment was designed to generate and capture the 5′-dAdo• in samples of reduced PFL-AE and SAM. Furthermore, the radical structure was able to be thoroughly examined using EPR and ENDOR spectroscopies (7). Lastly, anSAM was shown to be a true cofactor for PFL-AE laying the groundwork for future studies into the RS enzyme mechanism. The generation of anΩ using RFQ indicates that Ω forms before the 5′-dAdo• through a concerted mechanism. Photolysis of anSAM and PFL-AE released the anAdo• and •CH3 radicals providing further evidence that cryogenic photoinduced regioselectivities are in part due to the SAM ribose conformation in the enzyme active site (7, 20, 26). My research provides valuable insights into the RS mechanism that have greatly increases our understanding of these remarkable and diverse enzymes. 182 References 1. Sofia HJ, Chen G, Hetzler BG, Reyes-Spindola JF, Miller NE. Radical SAM, a novel protein superfamily linking unresolved steps in familiar biosynthetic pathways with radical mechanisms: functional characterization using new analysis and information visualization methods. Nucleic acids research. 2001;29(5):1097-106. 2. Broderick JB, Duffus BR, Duschene KS, Shepard EM. Radical S- adenosylmethionine enzymes. Chemical reviews. 2014;114(8):4229-317. 3. Jäger CM, Croft AK. Anaerobic radical enzymes for biotechnology. ChemBioEng Reviews. 2018;5(3):143-62. 4. Horitani M, Shisler K, Broderick WE, Hutcheson RU, Duschene KS, Marts AR, Hoffman BM, Broderick JB. Radical SAM catalysis via an organometallic intermediate with an Fe–[5′-C]-deoxyadenosyl bond. Science. 2016;352(6287):822-5. 5. Frey PA, Hegeman AD, Ruzicka FJ. The radical SAM superfamily. Critical reviews in biochemistry and molecular biology. 2008;43(1):63-88. 6. Byer AS, Yang H, McDaniel EC, Kathiresan V, Impano S, Pagnier A, Watts H, Denler C, Vagstad AL, Piel Jr. Paradigm shift for radical S-adenosyl-L-methionine reactions: The organometallic intermediate Ω is central to catalysis. Journal of the American Chemical Society. 2018;140(28):8634-8. 7. Yang H, McDaniel EC, Impano S, Byer AS, Jodts RJ, Yokoyama K, Broderick WE, Broderick JB, Hoffman BM. The elusive 5′-deoxyadenosyl radical: captured and characterized by electron paramagnetic resonance and electron nuclear double resonance spectroscopies. Journal of the American Chemical Society. 2019;141(30):12139-46. 8. Banerjee R. Radical carbon skeleton rearrangements: catalysis by coenzyme B12- dependent mutases. Chemical Reviews. 2003;103(6):2083-94. 9. Brown KL. Chemistry and enzymology of vitamin B12. Chemical reviews. 2005;105(6):2075-150. 10. Ghanekar VD, Lin RJ, Coffman R, Blakley RL. Detection by electron spin resonance of an exchange-coupled cob (II) alamin… free radical pair species generated by anaerobic photolysis of polycrystalline adenosylcobalamin. Biochemical and Biophysical Research Communications. 1981;101(1):215-21. 11. Dektar JL, Hacker NP. Photochemistry of triarylsulfonium salts. Journal of the American Chemical Society. 1990;112(16):6004-15. 183 12. Saeva FD, Morgan BP. Mechanism of one-electron electrochemical reductive cleavage reactions of sulfonium salts. Journal of the American Chemical Society. 1984;106(15):4121-5. 13. Mao Z, Liou S-H, Khadka N, Jenney Jr FE, Goodin DB, Seefeldt LC, Adams MW, Cramer SP, Larsen DS. Cluster-dependent charge-transfer dynamics in iron–sulfur proteins. Biochemistry. 2018;57(6):978-90. 14. Horsfield A, Morton J, Whiffen D. Electron spin resonance of γ-irradiated malonic acid. Molecular Physics. 1961;4(4):327-32. 15. Atherton NM. Principles of electron spin resonance: Ellis Horwood Limited; 1993. 16. Neese F. The ORCA program system. WIREs Comput Mol Sci 2: 73–78. 2012. 17. Horitani M, Byer AS, Shisler KA, Chandra T, Broderick JB, Hoffman BM. Why nature uses radical SAM enzymes so widely: Electron nuclear double resonance studies of lysine 2, 3-aminomutase show the 5′-dAdo•“free radical” is never free. Journal of the American Chemical Society. 2015;137(22):7111-21. 18. Magnusson OT, Reed GH, Frey PA. Spectroscopic evidence for the participation of an allylic analogue of the 5‘-deoxyadenosyl radical in the reaction of lysine 2, 3- aminomutase. Journal of the American Chemical Society. 1999;121(41):9764-5. 19. Magnusson OT, Reed GH, Frey PA. Characterization of an allylic analogue of the 5‘-deoxyadenosyl radical: An intermediate in the reaction of lysine 2, 3-aminomutase. Biochemistry. 2001;40(26):7773-82. 20. Impano S, Yang H, Jodts RJ, Pagnier A, Swimley R, McDaniel EC, Shepard EM, Broderick WE, Broderick JB, Hoffman BM. Active-site controlled, Jahn-Teller enabled regioselectivity in reductive S-C bond cleavage of S-adenosylmethionine (SAM) in radical-SAM enzymes. 2020. 21. Magnusson OT, Frey PA. Synthesis and Characterization of 3 ‘, 4 ‘- Anhydroadenosylcobalamin: A Coenzyme B12 Analogue with Unusual Properties. Journal of the American Chemical Society. 2000;122(37):8807-13. 22. Henshaw TF, Cheek J, Broderick JB. The [4Fe-4S]1+ cluster of pyruvate formate- lyase activating enzyme generates the glycyl radical on pyruvate formate-lyase: EPR- detected single turnover. Journal of the American Chemical Society. 2000;122(34):8331- 2. 23. Walsby CJ, Hong W, Broderick WE, Cheek J, Ortillo D, Broderick JB, Hoffman BM. Electron-nuclear double resonance spectroscopic evidence that S-adenosylmethionine 184 binds in contact with the catalytically active [4Fe-4S]+ cluster of pyruvate formate-lyase activating enzyme. Journal of the American Chemical Society. 2002;124(12):3143-51. 24. Walsby CJ, Ortillo D, Yang J, Nnyepi MR, Broderick WE, Hoffman BM, Broderick JB. Spectroscopic approaches to elucidating novel iron−sulfur chemistry in the “radical-SAM” protein superfamily. Inorganic Chemistry. 2005;44(4):727-41. 25. Walsby CJ, Ortillo D, Broderick WE, Broderick JB, Hoffman BM. An anchoring role for FeS clusters: Chelation of the amino acid moiety of S-adenosylmethionine to the unique iron site of the [4Fe− 4S] cluster of pyruvate formate-lyase activating enzyme. Journal of the American Chemical Society. 2002;124(38):11270-1. 26. Yang H, Impano S, Shepard EM, James CD, Broderick WE, Broderick JB, Hoffman BM. Photoinduced electron transfer in a radical SAM enzyme generates an S- adenosylmethionine derived methyl radical. Journal of the American Chemical Society. 2019;141(40):16117-24. 185 References Abragam A, Bleaney B. Electron paramagnetic resonance of transition ions: OUP Oxford; 2012. Ahmad R, Kuppusamy P. Theory, instrumentation, and applications of electron paramagnetic resonance oximetry. Chemical reviews. 2010;110(5):3212-36. Aitha M, Moller AJ, Sahu ID, Horitani M, Tierney DL, Crowder MW. Journal of Inorganic. Biochemistry. 2016;156:35-39. Akiva E, Brown S, Almonacid DE, Barber 2nd AE, Custer AF, Hicks MA, Huang CC, Lauck F, Mashiyama ST, Meng EC. The structure–function linkage database. Nucleic acids research. 2013;42(D1):D521-D30. Atherton NM. Principles of electron spin resonance. Ellis Horwood Limited New York; 1993. Babbitt PC. Structure Function Linkage Database [Online Database]. University of California, San Francisco The Regents of the University of California; 2016 [cited 2020]. Baldansuren A. Small Ag clusters supported on an LTA zeolite investigated by CW and pulse EPR spectroscopy, XAS and SQUID magnetometry. Ph.D. Thesis, University of Stuttgart, Stuttgart, Germany, 2009. Banci L, Camponeschi F, Ciofi-Baffoni S, Piccioli M. The NMR contribution to protein– protein networking in Fe–S protein maturation. Journal of Biological Inorganic Chemistry. 2018;23(4):665-85. Banerjee R. Radical carbon skeleton rearrangements: catalysis by coenzyme B12- dependent mutases. Chemical Reviews. 2003;103(6):2083-94. Banerjee R, Ragsdale SW. The many faces of vitamin B12: Catalysis by cobalamin- dependent enzymes. Annual Review of Biochemistry 2003;72:209−247. Baraniak J, Moss ML, Frey PA. Lysine 2,3-aminomutase: Support for a mechanism of hydrogen transfer involving S-adenosylmethionine. Journal of Biological Chemistry 1989;264:1357−1360. 186 Barker HA, Weissbach H, Smyth RD. A coenzyme containing pseudovitamin B12. Proceedings of the National Academy of Sciences of the United States. 1958;44: 1093−1097. Barone V. In Recent Advances in Density Functional Methods; World Scientific: Singapore. 1995;287. Becke AD, Density-functional exchange-energy approximation with correct asymptotic behavior. Physical Review A. 1988;38:3098−3100. Becker A, Fritz-Wolf K, Kabsch W, Knappe J, Schultz S, Wagner AV. Structure and mechanism of the glycyl radical enzyme pyruvate formate-lyase. Nature Structural & Molecular Biology. 1999;6(10):969. Beinert H, Holm RH, Münck E. Iron-sulfur clusters: nature's modular, multipurpose structures. Science. 1997;277(5326):653-9. Beinert H, Meyer J, Lill R. Iron–sulfur proteins. Encyclopedia of biological chemistry: Elsevier, Amsterdam: 2004;482-489. Beinert H, Sands RH. Studies on succinic and DPNH dehydrogenase preparations by paramagnetic resonance (EPR) spectroscopy. Biochemical and Biophysical Research Communications. 1960;3(1):41-46. Bertini I, Ciurli S, Luchinat C. The electronic structure of FeS centers in proteins and models a contribution to the understanding of their electron transfer properties. Iron-Sulfur Proteins Perovskites: Springer. 1995;1-53. Blaszczyk AJ, Silakov A, Zhang B, Maiocco SJ, Lanz ND, Kelly WL, Elliott SJ, Krebs C, Booker SJ. Spectroscopic and electrochemical characterization of the iron-sulfur and cobalamin cofactors of TsrM, an unusual radical S-adenosylmethionine methylase. Journal of the American Chemical Society. 2016;138:3416−3426. Bradford MM. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Analytical biochemistry. 1976;72(1-2):248-54. Bridwell-Rabb J, Grell TA, Drennan CL. A rich man, poor man story of s- adenosylmethionine and cobalamin revisited. Annual Review of Biochemistry. 2018;87:555-84. 187 Brindley AA, Zajicek R, Warren MJ, Ferguson SJ, Rigby SE. NirJ, a radical SAM family member of the d1 heme biogenesis cluster. FEBS letters. 2010;584(11):2461-6. Broderick JB, Duderstadt RE, Fernandez DC, Wojtuszewski K, Henshaw TF, Johnson MK. Pyruvate formate-lyase activating enzyme is an iron− sulfur protein. Journal of the American Chemical Society. 1997;119(31):7396-7. Broderick JB, Duffus BR, Duschene KS, Shepard EM. Radical S-adenosylmethionine enzymes. Chemical reviews. 2014;114(8):4229-317. Broderick WE, Hoffman BM, Broderick JB. Mechanism of radical initiation in the radical S-adenosyl-l-methionine superfamily. Accounts of chemical research. 2018;51(11):2611-9. Broderick JB, Duffus BR, Duschene KS, Shepard EM. Radical S-Adenosylmethionine Enzymes. Chemical Reviews. 2014;114:4229−4317. Broderick JB, Henshaw TF, Cheek J, Wojtuszewski K, Smith SR, Trojan MR, McGhan RM, Kopf A, Kibbey M, Broderick WE. Biochemical and Biophysical Research Communications. 2000;269:451-456. Broderick WE, Hoffman BM, Broderick JB. Mechanism of Radical Initiation in the Radical S-adenosyl-L-methionine Superfamily. Accounts of Chemical Research. 2018;51:2611−2619. Brown AC, Suess DL. Reversible formation of alkyl radicals at [Fe4S4] clusters and its implications for selectivity in radical SAM enzymes. Journal of the American Chemical Society. 2020;142(33):14240-8. Brown KL. Chemistry and enzymology of vitamin B12. Chemical reviews. 2005;105(6):2075-150. Brush EJ, Lipsett KA, Kozarich JW. Inactivation of Escherichia coli pyruvate formate- lyase by hypophosphite: evidence for a rate-limiting phosphorus-hydrogen bond cleavage. Biochemistry. 1988;27(6):2217-22. Bussandri AP, Kiarie CW, Van Willigen H. Photoinduced bond homolysis of B12 coenzymes. An FT-EPR study. Research on Chemical Intermediates. 2002;28:697−710. Byer AS, Yang H, McDaniel EC, Kathiresan V, Impano S, Pagnier A, Watts H, Denler C, Vagstad AL, Piel Jr. Paradigm shift for radical S-adenosyl-L-methionine 188 reactions: The organometallic intermediate Ω is central to catalysis. Journal of the American Chemical Society. 2018;140(28):8634-8. Byer AS, McDaniel EC, Impano S, Broderick WE, Broderick JB. Mechanistic Studies of Radical SAM Enzymes: Pyruvate Formate-Lyase Activating Enzyme and Lysine 2,3-Aminomutase Case Studies. Methods Enzymology. 2018;606:269−318. Byer AS, Yang H, McDaniel EC, Kathiresan V, Impano S, Pagnier A, Watts H, Denler C, Vagstad AL, Piel J, Duschene KS, Shepard EM, Shields TP, Scott LG, Lilla EA, Yokoyama K, Broderick WE, Hoffman BM, Broderick JB. Paradigm shift for radical S-adenosyl-L-methionine reactions: The organometallic intermediate Ω is central to catalysis. Journal of the American Chemical Society 2018;140(28):8634−8638. Carrington A, McLachlan AD. Introduction to magnetic resonance: with applications to chemistry and chemical physics; Harper and Row: New York, 1967. Castro B, Dormoy J, Evin G, Selve C. Peptide coupling reagents. Part VII. Mechanism of the formation of active esters of hydroxybenzotriazole in the reaction of carboxylate ions on the BOP reagent for peptide coupling. A comparison with Itoh's reagent. J Chem Res(S). 1977;182. Castro L, Tórtora VN, Mansilla S, Radi R. Aconitases: Non-redox iron–sulfur proteins sensitive to reactive species. Accounts of chemical research. 2019;52(9):2609-19. Chakravarti A, Selvadurai K, Shahoei R, Lee H, Fatma S, Tajkhorshid E, Huang RH. Reconstitution and substrate specificity for isopentenyl pyrophosphate of the antiviral radical SAM enzyme viperin. Journal of Biological Chemistry. 2018;293(36):14122-33. Chandra T, Broderick WE, Broderick JB. Chemoselective deprotection of triethylsilyl ethers. Nucleosides, Nucleotides, & Nucleic Acids 2009;28(11-12):1016-1029. Chandra T, Broderick WE, Broderick JB. An efficient deprotection of N- trimethylsilylethoxymethyl (SEM) groups from dinucleosides and dinucleotides. Nucleosides, Nucleotides, & Nucleic Acids 2010;29(2):132-143. Chandra T, Silver SC, Zilinskas E, Shepard EM, Broderick WE, Broderick JB. Spore photoproduct lyase catalyzes specific repair of the 5R but not the 5S spore photoproduct Journal of the American Chemical Society. 2009;131(7):2420-2421. 189 Chatterjee A, Li Y, Zhang Y, Grove TL, Lee M, Krebs C, Booker SJ, Begley TP, Ealick SE. Reconstitution of ThiC in thiamine pyrimidine biosynthesis expands the radical SAM superfamily. Nature chemical biology. 2008;4(12):758. Cheek J, Broderick JB. Direct H atom abstraction from spore photoproduct C-6 initiates DNA repair in the reaction catalyzed by spore photoproduct lyase: evidence for a reversibly generated adenosyl radical intermediate. Journal of the American Chemical Society. 2002;124(12):2860-1. Chen D, Frey PA. Identification of lysine 346 as a functionally important residue for pyridoxal 5 ‘-phosphate binding and catalysis in lysine 2, 3-aminomutase from Bacillus subtilis. Biochemistry. 2001;40(2):596-602. Chirpich TP, Zappia V, Costilow RN, Barker HA. Lysine 2,3-aminomutase purification and properties of a pyridoxal phosphate and S-adenosylmethionine-activated enzyme. Journal of Biological Chemistry. 1970;245:1778−1789. Cole T, Heller C. Hyperfine Splittings in the (HOOC)C13H(COOH) Radical. Journal of Chemical Physics. 1961;34(3):1085−1086. Conradt H, Hohmann-Berger M, Hohmann H-P, Blaschkowski HP, Knappe J. Pyruvate formate-lyase (inactive form) and pyruvate formate-lyase activating enzyme of Escherichia coli: isolation and structural properties. Archives of biochemistry and biophysics. 1984;228(1):133-42. Crack J, Green J, Thomson AJ. Mechanism of oxygen sensing by the bacterial transcription factor fumarate-nitrate reduction (FNR). Journal of Biological Chemistry. 2004;279(10):9278-86. Crain AV, Broderick JB. Flavodoxin cofactor binding induces structural changes that are required for protein–protein interactions with NADP+ oxidoreductase and pyruvate formate-lyase activating enzyme. Biochimica et Biophysica Acta (BBA)-Proteins and Proteomics. 2013;1834(12):2512-9. Crain AV, Duschene KS, Peters JW, Broderick JB. Iron-Sulfur Cluster Proteins, Fe/S-S- adenosylmethionine Enzymes and Hydrogenases. Encyclopedia of Metalloproteins: Springer New York. 2013;1034-44. Crane BR, Siegel LM, Getzoff ED. Sulfite reductase structure at 1.6 Å: evolution and catalysis for reduction of inorganic anions. Science. 1995;270(5233):59-67. 190 Cutsail III GE, Telser J, Hoffman BM. Advanced paramagnetic resonance spectroscopies of iron–sulfur proteins: Electron nuclear double resonance (ENDOR) and electron spin echo envelope modulation (ESEEM). Biochimica et Biophysica Acta (BBA)- Molecular Cell Research. 2015;1853(6):1370-94. Dai S, Friemann R, Glauser DA, Bourquin F, Manieri W, Schürmann P, Eklund H. Structural snapshots along the reaction pathway of ferredoxin–thioredoxin eductase. Nature. 2007;448(7149):92. Davey K. Modelling the combined effect of temperature and pH on the rate coefficient for bacterial growth. International journal of food microbiology. 1994;23(3- 4):295-303. Davoust CE, Doan PE, Hoffman BM. Q-Band pulsed electron spin-echo spectrometer and its application to ENDOR and ESEEM. Journal of Magnetic Resonance Series A. 1996;119:38−44. Dektar JL, Hacker NP. Photochemistry of triarylsulfonium salts. Journal of the American Chemical Society. 1990;112(16):6004-15. Dektar JL, Hacker NP. Photochemistry of triarylsulfonium salts. Journal of the American Chemical Society. 1990;112(16):6004−6015. DeRose VJ, Hoffman BM. Protein structure and mechanism studied by electron nuclear double resonance spectroscopy. Methods in Enzymology. 1995;246:554-589. Dey A, Peng Y, Broderick WE, Hedman B, Hodgson KO, Broderick JB, Solomon EI. S K-edge XAS and DFT calculations on SAM dependent pyruvate formate-lyase activating enzyme: nature of interaction between the Fe4S4 cluster and SAM and its role in reactivity. Journal of the American Chemical Society. 2011;133(46):18656-62. Dong M, Kathiresan V, Fenwick MK, Torelli AT, Zhang Y, Caranto J, Dzikovski B, Sharma A, Lancaster KM, Freed JH, Ealick SE, Hoffman BM, Lin H. Organometallic and radical intermediates reveal mechanism of diphthamide biosynthesis. Science 2018;359(6381):1247−1250. Driesener RC, Challand MR, McGlynn SE, Shepard EM, Boyd ES, Broderick JB, Peters JW, Roach PL. [FeFe]-Hydrogenase Cyanide Ligands Derived From S- Adenosylmethionine-Dependent Cleavage of Tyrosine Angewandte Chemie International Eddition. 2010;49:1687−90. 191 Driesener RC, Duffus BR, Shepard EM, Bruzas IR, Duschene KS, Coleman NJ-R, Marrison APG, Salvadori E, Kay CWM, Peters JW, Broderick JB, Roach PL. Biochemical and Kinetic Characterization of Radical S-Adenosyl-l-methionine Enzyme HydG. Biochemistry 2013;52(48):8696−8707. Duffus BR, Ghose S, Peters JW, Broderick JB. Reversible H Atom Abstraction Catalyzed by the Radical S-Adenosylmethionine Enzyme HydG. Journal of the American Chemical Society. 2014;136(38):13086-13089. Duschene KS, Veneziano SE, Silver SC, Broderick JB. Control of radical chemistry in the AdoMet radical enzymes. Current opinion in chemical biology. 2009;13(1):74-83. Fessenden RW. Schuler RH. Electron spin resonance studies of transient alkyl Radicals. Journal of Chemical Physics. 1963;39(9):2147. Fish W. Rapid colorimetric micromethod for the quantitation of complexed iron in biological samples in Methods in Enzymology (Riordan, JF and Vallee, BL, eds.) Vol. 158A. Academic Press.1988. Frank R, Gausepohl H. Modern methods in protein chemistry. In: Tschesche H, editor. New York: Walter de Gruyter. 1988;42-60. Freeman MF, Gurgui C, Helf MJ, Morinaka BI, Uria AR, Oldham NJ, Sahl H-G, Matsunaga S, Piel J. Metagenome mining reveals polytheonamides as posttranslationally modified ribosomal peptides. Science. 2012;338(6105):387-90. Frey M, Rothe M, Wagner A, Knappe J. Adenosylmethionine-dependent synthesis of the glycyl radical in pyruvate formate-lyase by abstraction of the glycine C-2 pro-S hydrogen atom. Studies of [2H] glycine-substituted enzyme and peptides homologous to the glycine 734 site. Journal of Biological Chemistry. 1994;269(17):12432-7. Frey P. Lysine 2, 3-aminomutase: is adenosylmethionine a poor man's adenosylcobalamin? The FASEB journal. 1993;7(8):662-70. Frey PA, Hegeman AD, Ruzicka FJ. The radical SAM superfamily. Critical reviews in biochemistry and molecular biology. 2008;43(1):63-88. Frey PA, Magnusson OT. S-Adenosylmethionine: a wolf in sheep's clothing, or a rich man's adenosylcobalamin? Chemical reviews. 2003;103(6):2129-48. 192 Frey PA. Radical mechanisms of enzymatic catalysis. Annual review of biochemistry. 2001;70(1):121-48. Frey M, Rothe M, Wagner AFV, Knappe J. Adenosylmethionine-dependent synthesis of the glycyl radical in pyruvate formate-lyase by abstraction of the glycine C-2 pro- S hydrogen atom. Journal of Biological Chemistry. 1994;269:12432−12437. Frey PA. Lysine 2,3‐aminomutase: is adenosylmethionine a poor man's adenosylcobalamin? Federation of American Societies for Experimental Biology. 1993;7(8):662−670. Frey PA. Travels with carbon-centered radicals. 5′-deoxyadenosine and 5′- deoxyadenosine-5′-yl in radical enzymology. Accounts of Chemical Research 2014;47(2):540−549. Frey PA, Abeles RH. The role of the B12 coenzyme in the conversion of 1,2-propanediol to propionaldehyde. Journal of Biological Chemistry. 1966;241:2732−2733. Frey PA, Hegeman AD, Reed GH. Free radical mechanisms in enzymology. Chemical Reviews. 2006;106(8):3302−3316. Frey PA, Hegeman AD, Ruzicka FJ. The radical SAM superfamily. Critical Reviews in Biochemistry and Molecular Biology. 2008;43(1):63−88. Frey PA, Magnusson OT. S-Adenosylmethionine: A wolf in sheep’s clothing, or a rich man’s adenosylcobalamin? Chemical Reviews. 2003;103(6):2129−2148. Gemperle C, Schweiger A. Pulsed electron-nuclear double resonance methodology. Chemical Reviews. 1991;91(7):1481-505. Ghanekar VD, Lin RJ, Coffman R, Blakley RL. Detection by electron spin resonance of an exchange-coupled cob (II) alamin… free radical pair species generated by anaerobic photolysis of polycrystalline adenosylcobalamin. Biochemical and Biophysical Research Communications. 1981;101(1):215-21. Grimshaw, J. Electrochemistry of the sulfonium group. In Chemistry of the Sulphonium Group; Stirling, C. J. M., Ed.; Wiley: Chichester, 1981; pp 141−155. Gross A, Geresh S, Whitesides GM. Enzymatic synthesis ofS-adenosyl-l-methionine from l-methionine and ATP. Applied biochemistry and biotechnology. 1983;8(5):415-22. 193 Guan Y, Manuel RC, Arvai AS, Parikh SS, Mol CD, Miller JH, Lloyd RS, Tainer JA. MutY catalytic core, mutant and bound adenine structures define specificity for DNA repair enzyme superfamily. Nature structural biology. 1998;5(12):1058-64. Hagen WR. EPR spectroscopy of iron—sulfur proteins. Advances in inorganic chemistry: Elsevier. 1992;165-222. Halpern J, Kim SH, Leung T. Cobalt-carbon bond dissociation energy of coenzyme B12. Journal of the American Chemical Society. 1984;106(26):8317-9. Halpern, J. Mechanisms of Coenzyme B12-Dependent Rearrangements. Science 1985;227:869−875. Hänzelmann P, Schindelin H. Crystal structure of the S-adenosylmethionine-dependent enzyme MoaA and its implications for molybdenum cofactor deficiency in humans. Proceedings of the National Academy of Sciences. 2004;101(35):12870- 5. Heidinger L, Kneuttinger AC, Kashiwazaki G, Weber S, Carell T, Schleicher E. Direct observation of a deoxyadenosyl radical in an active enzyme environment. FEBS Letters 2016;590(24):4489−4494. Henshaw TF, Cheek J, Broderick JB. The [4Fe-4S]1+ cluster of pyruvate formate-lyase activating enzyme generates the glycyl radical on pyruvate formate-lyase: EPR- detected single turnover. Journal of the American Chemical Society. 2000;122(34):8331-2. Hidalgo E, Demple B. An iron‐sulfur center essential for transcriptional activation by the redox‐sensing SoxR protein. The EMBO Journal. 1994;13(1):138-46. Himo F, Eriksson LA. Catalytic mechanism of pyruvate formate-lyase (PFL). A theoretical study. Journal of the American Chemical Society. 1998;120(44):11449-55. Hinckley GT, Frey PA. Cofactor dependence of reduction potentials for [4Fe-4S]2+/1+ in lysine 2,3-aminomutase. Biochemistry 2006;45(10):3219−3225. Hoffman BM, DeRose VJ, Ong JL, Davoust CE. Sensitivity enhancement in field- modulated CW ENDOR via RF bandwidth broadening. Journal of Magnetic Resonance, Series A 1994;110(1):52−57. 194 Hogenkamp H. A cyclic nucleoside derived from coenzyme B12. Journal of Biological Chemistry. 1963;238(1):477-80. Hogenkamp HPC, Barker HA, Mason HS. An electron- spin resonance study of coenzyme B12. Archives of Biochemistry and Biophysics. 1963;100(3):353−359. Holliday GL, Akiva E, Meng EC, Brown SD, Calhoun S, Pieper U, Sali A, Booker SJ, Babbitt PC. Atlas of the radical SAM superfamily: Divergent evolution of function using a “Plug and Play” domain. Methods in Enzymology. 2018;606:1−71. Horitani M, Byer AS, Shisler KA, Chandra T, Broderick JB, Hoffman BM. Why nature uses radical SAM enzymes so widely: Electron nuclear double resonance studies of lysine 2, 3-aminomutase show the 5′-dAdo•“free radical” is never free. Journal of the American Chemical Society. 2015;137(22):7111-21. Horitani M, Shisler K, Broderick WE, Hutcheson RU, Duschene KS, Marts AR, Hoffman BM, Broderick JB. Radical SAM catalysis via an organometallic intermediate with an Fe–[5′-C]-deoxyadenosyl bond. Science. 2016;352(6287):822-5. Horitani M, Byer AS, Shisler, KA, Chandra T, Broderick JB, Hoffman BM. Why Nature Uses Radical SAM Enzymes so Widely: Electron Nuclear Double Resonance Studies of Lysine 2,3- Aminomutase Show the 5′-dAdo• ′′Free Radical′′ is Never Free. Journal of the American Chemical Society. 2015;137(22):7111−7121. Horitani M, Shisler KA, Broderick WE, Hutcheson RU, Duschene KS, Marts AR, Hoffman BM, Broderick JB. Radical SAM catalysis via an organometallic intermediate with an Fe-[5′-C]-deoxyadenosyl bond. Science 2016;352:822−825. Horsfield A, Morton J, Whiffen D. Electron spin resonance of γ-irradiated malonic acid. Molecular Physics. 1961;4(4):327-32. Impano S, Yang H, Jodts RJ, Pagnier A, Swimley R, McDaniel EC, Shepard EM, Broderick WE, Broderick JB, Hoffman BM. Active-site controlled, Jahn-Teller enabled regioselectivity in reductive S-C bond cleavage of S-adenosylmethionine (SAM) in radical-SAM enzymes. 2020;Manuscript Accepted. Iwig DF, Booker SJ. Insight into the Polar Reactivity of the Onium Chalcogen Analogues of S-Adenosyl-l-methionine. Biochemistry. 2004;43(42):13496-13509. Jäger CM, Croft AK. Anaerobic radical enzymes for biotechnology. ChemBioEng Reviews. 2018;5(3):143-62. 195 Jankowsky E, Fairman ME. RNA helicases—one fold for many functions. Current opinion in structural biology. 2007;17(3):316-24. Johnson DC, Dean DR, Smith AD, Johnson MK. Structure, function, and formation of biological iron-sulfur clusters. Annu Rev Biochem. 2005;74:247-81. Jones AR. The photochemistry and photobiology of vitamin B12. Photochemical & Photobiological Sciences. 2017;6:820−834. Jost M, Simpson JH, Drennan CL. The transcription factor CarH safeguards use of adenosylcobalamin as a light sensor by altering the photolysis products. Biochemistry 2015;54(21):3231−3234. Kim SH, Perera R, Hager LP, Dawson JH, Hoffman BM. Journal of American Chemical Society. 2006;128(17):5598-5599. Knappe J, Blaschkowski H-P, Edenharder R. Enzyme-dependent activation of pyruvate formate-lyase of Escherichia coli. Metabolic Interconversion of Enzymes: Springer. 1972;319-29. Knappe J, Neugebauer FA, Blaschkowski HP, Gänzler M. Post-translational activation introduces a free radical into pyruvate formate-lyase. Proceedings of the National Academy of Sciences. 1984;81(5):1332-5. Knappe J, Schacht J, Möckel W, Höpner T, Vetter Jr H, Edenharder R. Pyruvate formate‐ lyase reaction in Escherichia coli: The enzymatic system converting and inactive form of the lyase into the catalytically active enzyme. European Journal of Biochemistry. 1969;11(2):316-27. Knappe J, Schmitt T. A novel reaction of S-adenosyl-L-methionine correlated with the activation of pyruvate formate-lyase. Biochemical and biophysical research communications. 1976;71(4):1110-7. Knappe J, Neugebauer FA, Blaschkowski HP, Gan̈zler M. Post-translational activation introduces a free radical into pyruvate formate-lyase. Proceeding of the National Academy of Sciences of the United States of America.1984;81(5):1332−1335. Krebs C, Broderick WE, Henshaw TF, Broderick JB, Huynh BH. Coordination of adenosylmethionine to a unique iron site of the [4Fe-4S] of pyruvate formate- lyase activating enzyme: A Mössbauer spectroscopic study. Journal of the American Chemical Society. 2002;124(6):912-3. 196 Krebs C, Henshaw TF, Cheek J, Huynh BH, Broderick JB. Conversion of 3Fe-4S to 4Fe- 4S clusters in native pyruvate formate-lyase activating enzyme: Mössbauer characterization and implications for mechanism. Journal of the American Chemical Society. 2000;122(50):12497-506. Külzer R, Pils T, Kappl R, Hüttermann J, Knappe J. Reconstitution and characterization of the polynuclear iron-sulfur cluster in pyruvate formate-lyase-activating enzyme molecular properties of the holoenzyme form. Journal of Biological Chemistry. 1998;273(9):4897-903. Kuo C-F, McRee DE, Fisher CL, O'Handley SF, Cunningham RP, Tainer JA. Atomic structure of the DNA repair [4Fe-4S] enzyme endonuclease III. Science. 1992;258(5081):434-40. Landgraf BJ, McCarthy EL, Booker SJ. Radical S-adenosylmethionine enzymes in human health and disease. Annual review of biochemistry. 2016;85:485-514. Lanz ND, Booker SJ. Identification and function of auxiliary iron–sulfur clusters in radical SAM enzymes. Biochimica et Biophysica Acta (BBA)-Proteins and Proteomics. 2012;1824(11):1196-212. Lee C, Yang W, Parr RG. Development of the Colle-Salvetti correlation energy formula into a functional of the electron density. Physical Review B. 1988;37(2):785−789. Lee H-I, Igarashi RY, Laryukhin M, Doan PE, Dos Santos PC, Dean DR, Seefeldt LC, Hoffman BM. An organometallic intermediate during alkyne reduction by nitrogenase. Journal of the American Chemical Society. 2004;126(32):9563−9569. Lenhert PG, Hodgkin DC. Structure of the 5,6-dimethylbenzimidazolylcobamide coenzyme. Nature 1961;192:937−938. Lepore BW, Ruzicka FJ, Frey PA, Ringe D. The x-ray crystal structure of lysine-2, 3- aminomutase from Clostridium subterminale. Proceedings of the National Academy of Sciences. 2005;102(39):13819-24. Lill R. Function and biogenesis of iron–sulphur proteins. Nature. 2009;460(7257):831. Lin Y, Gerfen GJ, Rousseau DL, Yeh S-R. Ultrafast microfluidic mixer and freeze- quenching device. Analytical Chemistry 2003;75(20):5381-5386. 197 Lindahl T. DNA Glycosylases in DNA Repair. In: Simic MG, Grossman L, Upton AC, Bergtold DS, editors. Mechanisms of DNA Damage and Repair: Implications for Carcinogenesis and Risk Assessment. Boston, MA: Springer US.1986;335-40. Lott WB, Chagovetz AM, Grissom CB. Alkyl radical geometry controls geminate cage recombination in alkylcobalamins. Journal of the American Chemical Society. 1995;117:12194−12201. Lowe DJ, Joblin KN, Cardin DJ. A precursor in the photolytic cleavage of the Co (III) C bond of coenzyme B-12 unobservable by electron paramagnetic resonance. Biochimica et Biophysica Acta (BBA)-General Subjects. 1978;539(3):398-401. Magnusson OT, Frey PA. Synthesis and Characterization of 3 ‘, 4 ‘- Anhydroadenosylcobalamin: A Coenzyme B12 Analogue with Unusual Properties. Journal of the American Chemical Society. 2000;122(37):8807-13. Magnusson OT, Reed GH, Frey PA. Characterization of an allylic analogue of the 5‘- deoxyadenosyl radical: An intermediate in the reaction of lysine 2, 3- aminomutase. Biochemistry. 2001;40(26):7773-82. Magnusson OT, Reed GH, Frey PA. Spectroscopic evidence for the participation of an allylic analogue of the 5‘-deoxyadenosyl radical in the reaction of lysine 2, 3- aminomutase. Journal of the American Chemical Society. 1999;121(41):9764-5. Maiocco SJ, Arcinas AJ, Landgraf BJ, Lee K-H, Booker SJ, Elliott SJ. Transformations of the FeS clusters of the methylthiotransferases MiaB and RimO, detected by direct electro-chemistry. Biochemistry 2016;55(39):5531−5536. Maiocco SJ, Grove TL, Booker SJ, Elliott SJ. Electrochemical resolution of the [4Fe-4S] centers of the AdoMet radical enzyme BtrN: Evidence of proton coupling and an unusual, low-potential auxiliary cluster. Journal of the American Chemical Society 2015;137(27):8664− 8667. Malkin R, Rabinowitz JC. The reconstitution of clostridial ferredoxin. Biochemical and biophysical research communications. 1966;23(6):822-7. Mao Z, Liou S-H, Khadka N, Jenney Jr FE, Goodin DB, Seefeldt LC, Adams MW, Cramer SP, Larsen DS. Cluster-dependent charge-transfer dynamics in iron– sulfur proteins. Biochemistry. 2018;57(6):978-90. 198 Maret W. The metals in the biological periodic system of the elements: concepts and conjectures. International journal of molecular sciences. 2016;17(1):66. Markham GD, DeParasis J, Gatmaitan J. The sequence of metK, the structural gene for S- adenosylmethionine synthetase in Escherichia coli. Journal of Biological Chemistry. 1984;259(23):14505-7. Markham GD, Hafner E, Tabor CW, Tabor H. S-Adenosylmethionine synthetase from Escherichia coli. Journal of Biological Chemistry. 1980;255(19):9082-92. Matsuda A, Muneyama K, Nishida T, Sato T, Ueda T. A new synthesis of 5′-deoxy-8, 5′- cyclo-adenosine and-inosine: conformationally-fixed purine nucleosides (nucleosides and nucleotides XVI). Nucleic acids research. 1976;3(12):3349-58. Maurice AM. Acquisition of Anisotropic Information by Computational Analysis of Isotropic EPR Spectra. Ph.D. Thesis, University of Illinois at Urbana-Champaign, Urbana-Champaign, IL, 1982. McCarty RM, Krebs C, Bandarian V. Spectroscopic, steady-state kinetic, and mechanistic characterization of the radical SAM enzyme QueE, which catalyzes a complex cyclization reaction in the biosynthesis of 7-deazapurines. Biochemistry. 2013;52(1):188-98. McConnell HM, Heller C, Cole T, Fessenden RW. Radiation damage in organic crystals. I. CH(COOH)2 in malonic acid. Journal of the American Chemical Society. 1960;82:766−775. McConnell HM, Strathdee J. Theory of anisotropic hyperfine interactions in pi-electron radicals. Molecular Physics. 1959;2:129−138. McCusker JK. Electronic structure in the transition metal block and its implications for light harvesting. Science 2019:363:484− 488. McGlynn SE, Boyd ES, Shepard EM, Lange RK, Gerlach R, Broderick JB, Peters JW. Identification and characterization of a novel member of the radical AdoMet enzyme superfamily and implications for the biosynthesis of the Hmd hydrogenase active site cofactor. Journal of bacteriology. 2010;192(2):595-8. McKenzie I, Brodovitch J-C, Ghandi K, McCollum BM, Percival PW. Hyperfine coupling in methyl radical isotopomers. The Journal of Physical Chemistry A. 2007;111(42):10625-34. 199 Mehta AP, Abdelwahed SH, Begley TP. Molybdopterin biosynthesis—Mechanistic studies on a novel MoaA catalyzed insertion of a purine carbon into the ribose of GTP. Biochimica et Biophysica Acta (BBA)-Proteins and Proteomics. 2015;1854(9):1073-7. Melis A, Seibert M, Happe T. Genomics of green algal hydrogen research. Photosynthesis research. 2004;82(3):277-88. Meyer J. Iron–sulfur protein folds, iron–sulfur chemistry, and evolution. JBIC Journal of Biological Inorganic Chemistry. 2008;13(2):157-70. Moore G. Handbook of metalloproteins, John Wiley & Sons, New York, 2001, 2 volumes. ISBN 0-471 62743 7;@ $645. Journal of Organometallic Chemistry. 2002;1(659):203-4. Morinaka BI, Vagstad AL, Helf MJ, Gugger M, Kegler C, Freeman MF, Bode HB, Piel J. Angewandte Chemie (International Edition). 2014;53(32):8503-8507. Mortenson L, Valentine R, Carnahan J. An electron transport factor from Clostridiumpasteurianum. Biochemical and biophysical research communications. 1962;7(6):448-52. Moss M, Frey P. Activation of lysine 2, 3-aminomutase by S-adenosylmethionine. Journal of Biological Chemistry. 1990;265(30):18112-5. Moss M, Frey PA. The role of S-adenosylmethionine in the lysine 2, 3-aminomutase reaction. Journal of Biological Chemistry. 1987;262(31):14859-62. Neese F. The ORCA program system. WIREs Computational Molecular Science. 2012;2(1):73–78. Neese, F. Efficient and accurate approximations to the molecular spin-orbit coupling operator and their use in molecular g-tensor calculations. Journal of Chemical Physics. 2005;122(3):034107. Nicolet Y, Drennan CL. AdoMet radical proteins—from structure to evolution— alignment of divergent protein sequences reveals strong secondary structure element conservation. Nucleic acids research. 2004;32(13):4015-25. Nicolet Y, Amara P, Mouesca J-M, Fontecilla-Camps JC. Unexpected electron transfer mechanism upon AdoMet cleavage in radical SAM proteins. Proceeding of the 200 National Academy of Sciences of the United States of America. 2009;106(35):14867−14871. Nilges MJ. Electron Paramagnetic Resonance Studies of Low Symmetry Nickel(i) and Molybdenum(v) Complexes. Ph.D. Thesis, University of Illinois at Urbana- Champaign, Urbana-Champaign, IL, 1979. Nnyepi MR, Peng Y, Broderick JB. Archives of Biochemistry and Biophysics. 2007;459(1):1-9. Peng Y, Veneziano SE, Gillispie GD, Broderick JB. Pyruvate formate-lyase, evidence for an open conformation favored in the presence of its activating enzyme. Journal of Biological Chemistry. 2010;285(35):27224-31. Perdew JP. Density-functional approximation for the correlation-energy of the inhomogeneous electron-gas. Physical Review B. 1986;33:8822−8824. Petasis DT, Hendrich MP. Quantitative interpretation of multifrequency multimode EPR spectra of metal containing proteins, enzymes, and biomimetic complexes. Methods in Enzymology. 2015;563:171-208. Peters JW. Structure and mechanism of iron-only hydrogenases. Current opinion in structural biology. 1999;9(6):670-6. Pomposiello PJ, Demple B. Redox-operated genetic switches: the SoxR and OxyR transcription factors. Trends in biotechnology. 2001;19(3):109-14. Porello SL, Cannon MJ, David SS. A substrate recognition role for the [4Fe-4S] 2+ cluster of the DNA repair glycosylase MutY. Biochemistry. 1998;37(18):6465-75. Que Jr L. Physical methods in bioinorganic chemistry spectroscopy and magnetism. University Science Books, Sausilito. 2000. Quitterer F, List A, Eisenreich W, Bacher A, Groll M. Crystal structure of methylornithine synthase (PylB): insights into the pyrrolysine biosynthesis. Angewandte Chemie International Edition. 2012;51(6):1339-42. Rebeil R, Sun Y, Chooback L, Pedraza-Reyes M, Kinsland C, Begley TP, Nicholson WLJ. Spore Photoproduct Lyase from Bacillus subtilis Spores Is a Novel Iron- Sulfur DNA Repair Enzyme Which Shares Features with Proteins such as Class III Anaerobic Ribonucleotide Reductases and Pyruvate-Formate Lyases. Journal of Bacteriology. 1998;180(18):4879−4885. 201 Morinaka BI, Vagstad AL, Helf MJ, Gugger M, Kegler C, Freeman MF, Bode HB, Piel J. Radical S‐Adenosyl Methionine Epimerases: Regioselective Introduction of Diverse D‐Amino Acid Patterns into Peptide Natural Products. Angewandte Chemie International Edition 2014;53(32):8503−8507. Rega N, Cossi M, Barone V. Development and validation of reliable quantum mechanical approaches for the study of free radicals in solution. The Journal of Chemical Physics. 1996;105(24):11060. Rödel W, Plaga W, Frank R, Knappe J. Primary structures of Escherichia coli pyruvate formate‐lyase and pyruvate‐formate‐lyase‐activating enzyme deduced from the DNA nucleotide sequences. European journal of biochemistry. 1988;177(1):153- 8. Rudolf J, Makrantoni V, Ingledew WJ, Stark MJ, White MF. The DNA repair helicases XPD and FancJ have essential iron-sulfur domains. Molecular cell. 2006;23(6):801-8. Rury AS, Wiley TE, Sension R. Energy cascades, excited state dynamics, and photochemistry in cob(III)alamins and ferric porphyrins. Accounts of Chemical Research. 2015;48(3):860−867. Saeva FD, Morgan BP. Mechanism of one-electron electrochemical reductive cleavage reactions of sulfonium salts. Journal of the American Chemical Society. 1984;106(15):4121-5. Saeva FD, Morgan BP, Luss HR. Photochemical conversion of sulfonium salts to sulfides via a 1,3-sigmatropic rearrangement. Photogeneration of Bronsted acids. Journal of Organic Chemistry. 1985;50(22)4360−4362. Schaf̈er A, Huber C, Ahlrichs R. Fully optimized contracted gaussian basis sets of triple zeta valence quality for atoms Li to Kr. Journal of Chemical Physics. 1994;100(8):5829−5835. Schauer DA, Iwasaki A, Romanyukha AA, Swartz HM, Onori S. Electron paramagnetic resonance (EPR) in medical dosimetry. Radiation measurements. 2006;41:S117- S23. Schwartz CJ, Giel JL, Patschkowski T, Luther C, Ruzicka FJ, Beinert H, Kiley PJ. IscR, an Fe-S cluster-containing transcription factor, represses expression of 202 Escherichia coli genes encoding Fe-S cluster assembly proteins. Proceedings of the National Academy of Sciences. 2001;98(26):14895-900. Scott LG, Geierstanger BH, Williamson JR, Hennig M. Enzymatic synthesis and 19F NMR studies of 2-fluoroadenine-substituted RNA. Journal of the American Chemical Society. 2004;126(38):11776-11777. Shapiro SK, Ehninger DJ. Methods for the analysis and preparation of adenosylmethionine and adenosylhomocysteine. Analytical biochemistry. 1966;15(2):323-33. Shepard EM, Duffus BR, McGlynn SE, Challand MR, Swanson KD, Roach PL, George SJ, Cramer SP, Peters JW, Broderick JB. [FeFe]-hydrogenase maturation: HydG- catalyzed synthesis of carbon monoxide. Journal of the American Chemical Society. 2010;132(27):9247−9249. Shisler KA, Broderick JB. Emerging themes in radical SAM chemistry. Current opinion in structural biology. 2012;22(6):701-10. Shisler KA, Broderick JB. Glycyl radical activating enzymes: structure, mechanism, and substrate interactions. Archives of biochemistry and biophysics. 2014;546:64-71. Shisler KA, Hutcheson RU, Horitani M, Duschene KS, Crain AV, Byer AS, Shepard EM, Rasmussen A, Yang J, Broderick WE. Monovalent cation activation of the radical SAM enzyme pyruvate formate-lyase activating enzyme. Journal of the American Chemical Society. 2017;139(34):11803-13. Silver SC, Chandra T, Zilinskas E, Ghose S, Broderick WE, Broderick JB. Complete stereospecific repair of a synthetic dinucleotide spore photoproduct by spore photoproduct lyase Journal of Biological Inorganic Chemistry. 2010;15:943-955. Sofia HJ, Chen G, Hetzler BG, Reyes-Spindola JF, Miller NE. Radical SAM, a novel protein superfamily linking unresolved steps in familiar biosynthetic pathways with radical mechanisms: functional characterization using new analysis and information visualization methods. Nucleic acids research. 2001;29(5):1097-106. Stoll S, Schweiger A. EasySpin, a comprehensive software package for spectral simulation and analysis in EPR. Journal of Magnetic Resonance. 2006;178(1):42−55. Sulc F, Farmer PJ. Bioinorganic Chemistry of the HNO Ligand. The Smallest Biomolecules: Diatomics and their Interactions with Heme Proteins: Elsevier. 2008;429-62. 203 Sun X, Eliasson R, Pontis E, Andersson J, Buist G, Sjöberg B-M, Reichard P. Generation of the Glycyl Radical of the Anaerobic Escherichia coli Ribonucleotide Reductase Requires a Specific Activating Enzyme. Journal of Biological Chemistry. 1995;270:2443−2446. Svistunenko DA, Sharpe MA, Nicholls P, Wilson MT, Cooper CE. A new method for quantitation of spin concentration by EPR spectroscopy: application to methemoglobin and metmyoglobin. Journal of Magnetic Resonance. 2000;142(2):266-75. Tan L, Chi-Lung Y. Abundance of chemical elements in the earth's crust and its major tectonic units. International Geology Review. 1970;12(7):778-86. Tropea JE, Cherry S, Waugh DS. Expression and Purification of Soluble His6-Tagged TEV Protease Methods in Molecular Biology. 2009;498:297-307. Ugulava NB, Gibney BR, Jarrett JT. Biotin synthase contains two distinct iron-sulfur cluster binding sites: Chemical and spectroelectrochemical analysis of iron-sulfur cluster interconversions. Biochemistry. 2001;40(28)8343−8351. Vagstad AL, Kuranaga T, Püntener S, Pattabiraman VR, Bode JW, Piel J. Introduction of d‐Amino Acids in Minimalistic Peptide Substrates by an S‐Adenosyl‐l‐ Methionine Radical Epimerase. Angewandte Chemie International Edition. 2019;58(8):2246-50. Valentine R, Mortenson L, Carnahan J. The hydrogenase system of Clostridium pasteurianum. J biol Chem. 1963;238:1141-4. Vey JL, Drennan CL. Structural insights into radical generation by the radical SAM superfamily. Chemical Reviews. 2011;111(4):2487-506. Vey JL, Yang J, Li M, Broderick WE, Broderick JB, Drennan CL. Structural basis for glycyl radical formation by pyruvate formate-lyase activating enzyme. Proceedings of the National Academy of Sciences. 2008;105(42):16137-41. Vey JL, Drennan CL. Structural Insights into Radical Generation by the Radical SAM Superfamily. Chemical Review. 2011;111(4):2487−2506. Kampmeier JA. Regioselectivity in the homolytic cleavage of S-adenosylmethionine. Biochemistry. 2010;49(51):10770−107702. 204 Vey JL, Yang J, Li M, Broderick WE, Broderick JB, Drennan CL. Structural basis for glycyl radical formation by pyruvate formate-lyase activating enzyme. Proceedings of the National Academy of Sciences of the United States of America. 2008;105(42):16137−16141. Wagner A, Frey M, Neugebauer FA, Schäfer W, Knappe J. The free radical in pyruvate formate-lyase is located on glycine-734. Proceedings of the National Academy of Sciences. 1992;89(3):996-1000. Walsby CJ, Hong W, Broderick WE, Cheek J, Ortillo D, Broderick JB, Hoffman BM. Electron-nuclear double resonance spectroscopic evidence that S- adenosylmethionine binds in contact with the catalytically active [4Fe-4S]+ cluster of pyruvate formate-lyase activating enzyme. Journal of the American Chemical Society. 2002;124(12):3143-51. Walsby CJ, Ortillo D, Broderick WE, Broderick JB, Hoffman BM. An anchoring role for FeS clusters: Chelation of the amino acid moiety of S-adenosylmethionine to the unique iron site of the [4Fe− 4S] cluster of pyruvate formate-lyase activating enzyme. Journal of the American Chemical Society. 2002;124(38):11270-1. Walsby CJ, Ortillo D, Yang J, Nnyepi MR, Broderick WE, Hoffman BM, Broderick JB. Spectroscopic approaches to elucidating novel iron−sulfur chemistry in the “radical-SAM” protein superfamily. Inorganic Chemistry. 2005;44(4):727-41. Wang X-B, Niu S, Yang X, Ibrahim SK, Pickett CJ, Ichiye T, Wang L-S. Probing the intrinsic electronic structure of the cubane [4fe− 4s] cluster: Nature's favorite cluster for electron transfer and storage. Journal of the American Chemical Society. 2003;125(46):14072-81. Wang SC, Frey PA. Binding energy in the one-electron reductive cleavage of S- adenosylmethionine in lysine 2,3-aminomutase, a radical SAM enzyme. Biochemistry. 2007;46(45):12889−12895. Weber RT, Jiang J, Barr DP. EMX user’s manual. Bruker Instruments, Billerica. 1998. Weil, J. A.; Bolton, J. R.; Wertz, J. E. Electron Paramagnetic Resonance: Elementary Theory and Practical Applications; John Wiley & Sons: New York, 1994. Werst MM, Davoust CE, Hoffman BM. Ligand Spin Densities in Blue Copper Proteins by Q-band 1H and 14N ENDOR Spectroscopy. Journal of the American Chemical Society. 1991;113(5):1533−1538. 205 Wierenga R. The TIM‐barrel fold: a versatile framework for efficient enzymes. FEBS letters. 2001;492(3):193-8. Yang H, Impano S, Shepard EM, James CD, Broderick WE, Broderick JB, Hoffman BM. Photoinduced electron transfer in a radical SAM enzyme generates an S- adenosylmethionine derived methyl radical. Journal of the American Chemical Society. 2019;141(40):16117-24. Yang H, McDaniel EC, Impano S, Byer AS, Jodts RJ, Yokoyama K, Broderick WE, Broderick JB, Hoffman BM. The elusive 5′-deoxyadenosyl radical: captured and characterized by electron paramagnetic resonance and electron nuclear double resonance spectroscopies. Journal of the American Chemical Society. 2019;141(30):12139-46. Yoder LM, Cole AG, Walker LA, Sension RJ. Time-resolved spectroscopic studies of B12 coenzymes: influence of solvent on the photolysis of adenosylcobalamin. The Journal of Physical Chemistry B. 2001;105(48):12180-8. Zipse H, Artin E, Wnuk S, Lohman GJS, Martino D, Griffin RG, Kacprzak S, Kaupp M, Hoffman B, Bennati M, Stubbe J, Lees N. Structure of the nucleotide radical formed during reaction of CDP/TTP with the E441Q-α2β2 of E. coli ribonucleotide reductase. Journal of the American Chemical Society. 2009;131(1);200−211.