For the Love of SAM: Engineering Methyltransferase Catalysis Via S-Adenosylmethionine Analogues

For the Love of SAM: Engineering Methyltransferase Catalysis via S-Adenosylmethionine Analogues

by Kalli Corinne Catcott

B.S. in Chemistry/Biochemistry, University of California, San Diego

A dissertation submitted to

The Faculty of the College of Science of Northeastern University in partial fulfillment of the requirements for the degree of Doctor of Philosophy

9 December 2016

Dissertation directed by

Zhaohui Sunny Zhou Professor of Chemistry and Chemical Biology Faculty Fellow of the Barnett Institute of Chemical and Biological Analysis

Dedication

To my wonderful wife: the only Sam I want to spend the rest of my life with

Acknowledgements

Science is a field which grows continuously with ever expanding frontiers. […] Any particular advance has been preceded by the contributions of those from many lands who have set firm foundations for further developments. […] Further, science is a collaborative effort. The combined results of several people working together is often much more effective than could be that of an individual scientist working alone. —Prof. John Bardeen, Nobel Laureate1

I certainly have many people to thank for both laying the foundations for my work and for helping me along the way.

First, I am grateful to Professor Zhaohui Sunny Zhou for taking me into his lab and giving me incredible freedom to operate. Sunny puts the philosophy in Doctor of Philosophy; and, our conversations always leave me thinking about more than just methyltransferases.

Drs. Richard Duclos, Jason Guo, and Jared Auclair have been key experts in guiding my research and helping me acquire important data along the way. Additionally, Dr. Wanlu Qu’s work on AdoVin was fundamental in propelling forward my own.

My labmates, Shanshan Liu and Kevin Moulton, have been key advisors and uncomplaining sounding boards throughout my graduate studies. I look forward to working for one of you someday. I also must thank the ambitious undergraduate researchers I’ve worked with: Michael Pablo, Dillon Cleary, Paige Dickson, Lenny Negrón, and Diego Arévalo.

Our collaboration with Professor Vicki Wysocki and Jing Yan analyzing AdoVin by native mass spectrometry has been both fruitful and fulfilling. I feel lucky for having the chance to work with them.

iii

The gracious members of my committee—Professors Jeffrey Agar, George O’Doherty,

Penny Beuning, and Dr. Nicholas Yoder—have markedly improved the work I have done. I have such respect for your experience and views. This dissertation is much better for your insights.

Thank you to ImmunoGen, especially Nick, who allowed me to start at Northeastern while still working and were supportive when I transitioned to Northeastern full time.

Many thanks to the Northeastern University chemistry community, who have often loaned/given me supplies/equipment/expertise. My best efforts would have been stymied if not for your generosity.

My excellent parents gave me an outstanding foundation from which to build and have continued their support my whole life. I am privileged to be your daughter.

My son, Gram, has only witnessed that last year and a half of my efforts, but has contributed immeasurably to my joy. Thank you for brightening every day.

Finally, I must dedicate this work to my incredible wife, whose support—emotional, financial, and moral—has been paramount in my achievements. I would be lost without you.

Abstract of Dissertation Enzymes are exquisite catalysts—selective and specialized. Yet, enzymes are frequently

engineered to alter the substrates used or reaction catalyzed. Naturally-occurring catalytic promiscuity in enzymes was first described almost a hundred years ago. More recently, these properties have been exploited to expand the biocatalytic space available for synthetic chemistry and biological probes. Methyltransferases represent a large and diverse class of enzymes whose catalytic promiscuity has only been marginally explored. A vast majority of methyltransferases utilize S-adenosylmethionine (AdoMet, SAM) as a source of methyl for the transfer reaction.

Detailed here are two AdoMet analogues, Se-adenosylselenohomocysteine selenoxide (SeAHO) and S-adenosylvinthionine (AdoVin), which can be used to induce catalytic promiscuity in their respective enzyme systems.

SeAHO is an AdoMet analogue, wherein the methylsulfonium is replaced with a selenoxide. Here, SeAHO is shown to alter the activity of catechol-O-methyltransferase

(COMT), converting it to a putative oxidoreductase. The synthesis and characterization of

SeAHO is also described. Unlike the sulfur counterpart, S-adenosylhomocysteine sulfoxide

(SAHO), SeAHO is readily reduced by thiols, cysteine and glutathione; and, no reduction is observed in the presence of thioethers.

In AdoVin, the methylsulfonium is replaced with vinyl sulfonium. AdoVin is utilized as a thiopurine-S-methyltransferase (TPMT) substrate to form bisubstrate adducts, conferring putative ligase activity. Detailed here are routes of purification and characterization of AdoVin and its adducts. AdoVin has also been used as a probe in a newly described framework,

IsoLAIT, which utilizes native mass spectrometry to identify enzyme-substrate pairs in cellular contexts. v

Table of Contents

Dedication ...... ii

Acknowledgements ...... iii

Abstract of Dissertation ...... v

Table of Contents ...... vi

List of Figures ...... xi

List of Tables ...... xv

List of Equations ...... xvi

List of Symbols and Abbreviations...... xvii

Introduction ...... 1

Catalytic Promiscuity ...... 2

Methyltransferases ...... 4

AdoMet Analogues ...... 5

1 Se-Adenosylselenohomocysteine Selenoxide ...... 8

1.1 Introduction ...... 10

1.2 Preparation ...... 11

1.3 Characterization ...... 14

1.3.a NMR ...... 14

1.3.b IR Spectroscopy ...... 17

1.3.c Mass Spectrometry ...... 19

1.4 Stability ...... 26

1.4.a NMR Time Course ...... 26

1.4.b Aqueous Buffers at Various pH ...... 28

1.5 Reactivity ...... 32

1.5.a Reactivity with Biological Thiols and Thioethers ...... 33

1.5.b Reactivity with Whole Proteins ...... 35

1.6 Conclusions ...... 41

1.7 Experimental Procedures ...... 42

1.7.a NMR ...... 42

1.7.b IR Spectroscopy ...... 42

1.7.c Mass Spectrometry ...... 42

1.7.d Reverse Phase HPLC ...... 42

1.7.e Methyltransferase-Catalyzed Reduction and Bottom-Up MS ...... 43

2 Engineering Methyltransferase Activity: Conversion to an Oxidoreductase ...... 45

2.1 Introduction ...... 46

2.2 Catechol-O-Methyltransferase ...... 46

2.3 Oxidation of Epinephrine ...... 47

vii

2.4 Oxidation of Other Catechols ...... 56

2.5 Conclusions ...... 59

2.6 Experimental Details ...... 60

2.6.a Preparation of Enzymes ...... 60

2.6.b Preparation of SeAHO ...... 61

2.6.c Enzyme-Mediated Reactions ...... 61

3 S-Adenosylvinthionine Purification and Analysis ...... 64

3.1 Introduction ...... 68

3.2 Purification of AdoVin ...... 69

3.3 Characterization ...... 71

3.3.a High Resolution Mass Spectrometry ...... 72

3.3.b NMR ...... 74

3.4 Degradation ...... 76

3.5 Binding ...... 80

3.5.a AdoMet and AdoVin TPMT-Binding Competition ...... 80

3.5.b AdoVin-TNB Adduct Binding with TPMT ...... 81

3.6 Adduct Formation in Human Cell Lysate ...... 87

3.7 Conclusions ...... 88

3.8 Experimental Details ...... 89

viii

3.8.a Chemoenzymatic Synthesis of AdoVin ...... 89

3.8.b Formation of AdoVin-TNB Adduct and TPMT•Adduct Complex ...... 90

3.8.c SCX-HPLC ...... 90

3.8.d High Resolution Mass Spectrometry ...... 90

3.8.e NMR ...... 91

3.8.f LC-MS ...... 91

3.8.g AdoVin and AdoMet Binding ...... 91

3.8.h Molecular Models ...... 92

3.8.i HeLa Cell Lysate Screening ...... 92

4 Identifying Enzyme-Substrate Pairs with AdoVin and Native MS ...... 94

4.1 Introduction ...... 97

4.2 IsoLAIT Keys ...... 97

4.3 Native Mass Spectrometry of In Vitro Samples ...... 99

4.4 Native Mass Spectrometry of Ex Vivo Samples ...... 105

4.5 Conclusion ...... 112

4.6 Experimental Details ...... 113

4.6.a Preparation of S-adenosylvinthionine (AdoVin) ...... 113

4.6.b In Vitro Reactions ...... 113

4.6.c Ex Vivo Reactions ...... 114

4.6.d Sample Preparation ...... 114

4.6.e Native Mass Spectrometry ...... 115

4.6.f Tandem Mass Spectrometry ...... 115

5 Conclusions and Future Directions ...... 117

5.1 Se-Adenosylselenohomocysteine Selenoxide ...... 117

5.1.a O-Methyltransferases ...... 117

5.1.b S-Methyltransferases ...... 117

5.1.c N-Methyltransferases ...... 118

5.1.d C-Methyltransferases ...... 119

5.2 S-Adenosylvinthionine ...... 119

5.2.a Thiol-S-Methyltransferase ...... 119

5.2.b MET7A ...... 120

5.2.c Aziridines ...... 121

5.2.d Molecular Models ...... 122

References ...... 123

List of Figures

Figure 0.1 Overview of Directed Evolution ...... 3

Figure 0.2 Enzyme-Catalyzed Methylation ...... 5

Figure 0.3 Selected AdoMet Analogues with Enzyme Activity ...... 6

Figure 0.4 Structure of AdoMet Analogues, SeAHO and AdoVin ...... 6

Figure 1.1 Structure of AdoMet, SAHO, SeAHO, and SeAH ...... 11

Figure 1.2 Synthesis of SeAH ...... 13

1 Figure 1.3 H NMR of SeAHO in 97:3 CD3CO2D:30% Aqueous H2O2 ...... 15

Figure 1.4 COSY NMR of SeAHO in 97:3 CD3CO2D:30% Aqueous H2O2 ...... 16

Figure 1.5 IR Spectrum of SeAHO and SeAH ...... 17

Figure 1.6 Zoomed in IR Spectrum of SeAHO and SeAH ...... 18

Figure 1.7 IR Spectrum of SeAHO and SAHO ...... 18

Figure 1.8 SeAHO Fragments Observed by Mass Spectrometry ...... 20

Figure 1.9 LCQ-MS Spectrum of SeAHO ...... 21

Figure 1.10 XIC from the LCQ-MS of Peaks in the SeAHO Sample ...... 22

Figure 1.11 MS-MS Spectra of SeAHO and Dehydration Product ...... 23

Figure 1.12 MALDI-ToF Spectrum of SeAHO ...... 24

Figure 1.13 MS-MS Spectra of MALDI-ToF of SeAHO ...... 25

Figure 1.14 Time Course NMR of SeAHO ...... 27

Figure 1.15 Stability of SeAHO in Aqueous Buffers at Various pH Values ...... 28

Figure 1.16 1H NMR of SeAHO at Various pH values ...... 29

Figure 1.17 COSY of SeAHO ...... 30

Figure 1.18 Hydration and Racemization of SeAHO...... 32

Figure 1.19 SeAHO—Reduced By Thiols, Not By Thioethers ...... 34

Figure 1.20 SAHO—Not Reduced by Thiols ...... 35

Figure 1.21 SeAHO—Reduced in the Presence of Some Proteins ...... 36

Figure 1.22 Reactivity of SeAHO with Various Amino Acids ...... 38

Figure 1.23 MALDI-MS Spectra of SeAHO-Treated COMT ...... 40

Figure 2.1 Catechol-O-Methyltransferase Reaction ...... 47

Figure 2.2 Oxidation of Epinephrine ...... 48

Figure 2.3 UV Spectra of Epinephrine and Adrenochrome ...... 48

Figure 2.4 Proposed Epinephrine Oxidation Mechanism ...... 49

Figure 2.5 Adrenochrome Retention Time and Standard Curve ...... 50

Figure 2.6 Background Adrenochrome Formation ...... 51

Figure 2.7 Increase in Adrenochrome at Various COMT Concentrations ...... 52

Figure 2.8 Rate of Adrenochrome Formation vs COMT Concentration ...... 53

Figure 2.9 Background Oxidation of Epinephrine ± COMT ...... 53

Figure 2.10 Adrenochrome-Catalyzed Oxidation ...... 54

Figure 2.11 COMT Active Site ...... 55

Figure 2.12 Structures of Catechol, Dopamine, and Tolcapone ...... 56

Figure 2.13 Oxidation of Dopamine, Catechol, and Epinephrine ...... 57

Figure 2.14 COMT-Catalyzed Methylation Reactions ...... 58

Figure 2.15 Tolcapone Oxidation Reaction ...... 59

Figure 3.1 Synthesis and Reactivity of AdoVin ...... 64

Figure 3.2 TPMT-Bound and Free Ligand Fractions ...... 66

Figure 3.3 Change in TNB Concentration Over Time ...... 67

Figure 3.4 HPLC Chromatograms of Ex Vivo Reactions ...... 68

xii

Figure 3.5 SCX-HPLC Chromatogram of AdoMet and AdoVin ...... 70

Figure 3.6 Separation of AdoVin by SCX in Various Mobile Phases ...... 71

Figure 3.7 Mass Spectrum of AdoVin...... 72

Figure 3.8 High Resolution Mass Spectrum of AdoVin ...... 73

Figure 3.9 1H NMR of Purified AdoVin and Commercial AdoMet ...... 75

Figure 3.10 1H NMR of Purified AdoVin ...... 76

Figure 3.11 Stability of AdoVin ...... 77

Figure 3.12 AdoVin Degradation Pathways...... 78

Figure 3.13 XIC and Mass Spectra of AdoVin Degradation ...... 79

Figure 3.14 TPMT Binding of AdoMet vs AdoVin ...... 81

Figure 3.15 AdoVin-TNB Adduct Dissociation with Guanidinium ...... 82

Figure 3.16 Guanidinium Induced Dissociation of AdoVin-Adduct from TPMT ...... 83

Figure 3.17 AdoVin-TNB Adduct Dissociation with AdoMet or AdoHcy ...... 84

Figure 3.18 AdoVin-TNB Adduct Competing with AdoMet and AdoHcy ...... 85

Figure 3.19 Crystal Structure of TPMT with AdoHcy Bound ...... 86

Figure 3.20 Mobiligram of TPMT•Adduct Complex ...... 86

Figure 3.21 XIC and MS of AdoVin-TNB Adduct in HeLa Lysates ...... 88

Figure 4.1 Conventional Catalytic Cycle ...... 94

Figure 4.2 IsoLAIT Framework ...... 96

Figure 4.3 TPMT-Catalyzed Reactions with AdoMet and AdoVin ...... 98

Figure 4.4 S-Adenosylvinthionine Probe Preparation ...... 99

Figure 4.5 Native MS and Tandem MS of In Vivo Complex ...... 101

Figure 4.6 CID-MS of TNB-AdoVin Adduct ...... 103

Figure 4.7 Quasi MS3 of Unlabeled TNB-AdoVin Adduct ...... 104

xiii

Figure 4.8 Native MS and Tandem MS of Ex Vivo Complex ...... 106

Figure 4.9 Mass Spectrum of Apo-TPMT ...... 108

Figure 4.10 SID of [TPMT•Adduct] Complex ...... 109

Figure 4.11 MS and Tandem MS of [TPMT•AdoMet] Complex ...... 110

Figure 4.12 MS and Tandem MS of [TPMT•AdoHcy] Complex ...... 111

Figure 5.1 TNB Oxidation with SeAHO ...... 118

Figure 5.2 MET7A Homology Model Aligned with As(III) Methyltransferase ...... 121

Figure 5.3 Aziridinoadenosines ...... 122

xiv

List of Tables

Table 0.1 Enzyme Commission Subclasses ...... 1

Table 1.1 1H Shifts of SAHO and SeAHO in Deuterated Acetic Acid ...... 15

Table 1.2 13C Shifts SeAHO in Deuterated Acetic Acid ...... 16

Table 3.1 SCX-HPLC Buffer Systems ...... 70

List of Equations

Equation 1 Dissociation Constant ...... 83

xvi

List of Symbols and Abbreviations

°C degrees celsius 13C carbon 13 15N nitrogen 15 1H NMR proton nuclear magnetic resonance

α alpha Å angstrom A absorbance A adenosine ACS American Chemical Society AdoHcy S-adenosylhomocysteine AdoMet S-adenosylmethionine AdoVin S-adenosylvinthionine AMBA 2-amino-5-mercaptobenzoic acid Ar aromatic As arsenic AS3MT arsenic (III) methyltransferase ATP adenosine triphosphate AUC area under the curve

β beta B base BEH ethylene bridged hybrid BSA bovine serum albumin

γ gamma C18 octadecyl carbon chain CAS Chemical Abstracts Service xvii

CD3CO2D deuterated acetic acid C-H carbon-hydrogen single bond

CH3CN acetonitrile

CH3CO2NH4 ammonium acetate

CH3COOH acetic acid CHCA α-cyano-4-hydroxycinnamic acid CID collision induced dissociation cm-1 reciprocal centimeters, wavenumber cm-1M-1 reciprocal centimeters reciprocal molar, unit of molar absorptivity CoA coenzyme A COMT catechol-O-methyltransferase COSY correlation spectroscopy CXXC cysteine-amino acid-amino acid-cysteine sequence motif CXXS cysteine-amino acid-amino acid-serine sequence motif

δ delta, chemical shift D deuterium

D2O deuterated water Da dalton DARTS drug affinity responsive target stability DMSO dimethyl sulfoxide DNA deoxyribonucleic acid DTT dithiotheritol

ε extinction coefficient, molar absorptivity E. coli Escherichia coli EC Enzyme Commission ESI electrospray ionization eV electronvolt

xviii

FDA Food and Drug Administration FT-ICR Fourier transform ion cyclotron resonance FTIR Fourier transform infrared spectroscopy

γ gamma Gdn guanidinium GSH glutathione GXGXG glycine-amino acid-glycine-amino acid-glycine sequence motif

H2O water

H2O2 hydrogen peroxide HCl hydrochloric acid His-Tag polyhistidine-tag HPLC high performance liquid chromatography hr-1 reciprocal hour HRMS high resolution mass spectrometry

IMAC immobilized metal affinity chromatography IPTG isopropyl β-D-1-thiogalactopyranoside IR infrared IsoLAIT Isotope-Labeled, Activity-based Identification and Tracking

K lysine

K3PO4 potassium phosphate tribasic kapp apparent turnover rate kcat turnover rate

Kd dissociation constant KCl potassium chloride kDa kilodalton

KH2PO4 potassium phosphate monobasic kV kilovolt xix

LB Broth Luria-Bertani broth, lysogeny broth LC-MS liquid chromatography-mass spectrometry

µL microliter µM micromolar µm micron, micrometer M molar m/s meters per second m/z mass over charge M+ molecular ion [M+H]+ protonated molecular ion MALDI matrix-assisted laser desorption/ionization MAT L-methionine S-adenosyltransferase mbar millibar mg/mL milligram per milliliter Mg2+ magnesium (II) ion

MgCl2 magnesium (II) chloride MHz megahertz min minute min-1 reciprocal minute mL/min milliliter per minute mM millimolar mm millimeter mRNA messenger ribonucleic acid MS mass spectrometry MS2, MS-MS tandem mass spectrometry MS3 tandem mass spectrometry-mass spectrometry MTAN 5’-methylthioadenosine nucleosidase MTase methyltransferase MWCO molecular weight cut off

Na (s) sodium metal NaCl sodium chloride NAD+ nicotinamide adenine dinucleotide NADP+ nicotinamide adenine dinucleotide phosphate NaOH sodium hydroxide

NH3 ammonia

NH4Cl ammonium chloride

NiSO4 nickel (II) sulfate nm nanometer nM nanomolar nmol nanomole NMR nuclear magnetic resonance Nu: nucleophile

PCR polymerase chain reaction pH negative logarithm of the hydrogen ion concentration PN part number ppm parts per million PRMT protein arginine methyltransferase

Rf retention factor RMSD root-mean-square deviation RNA ribonucleic acid RPLC reverse phase liquid chromatography R-SH thiol R-S-R thioether

S=O sulfur-oxygen double bond SAH S-adenosylhomocysteine SAHO S-adenosylhomocysteine sulfoxide

xxi

SAM S-adenosylmethionine SCX strong cation exchange SDS-PAGE sodium dodecyl sulfate polyacrylamide gel electrophoresis Se=O selenium-oxygen double bond SeAH Se-adenosylselenohomocysteine SeAHO Se-adenosylselenohomocysteine selenoxide SID surface induced dissociation SPROX stability of proteins from rates of oxidation

T0 time zero TCA trichloroacetic acid TCEP tris(2-carboxyethyl)phosphine TFA trifluoroacetic acid TLC thin layer chromatography TMT thiol-S-methyltransferase TNB 2-nitro-5-thiobenzoic acid ToF time of flight TPMT thiopurine-S-methyltransferase

UV ultraviolet v/v volume per volume xg g-force XIC extracted ion chromatogram

Y tyrosine z charge

xxii

Introduction

The Enzyme Commission has categorized enzymes into six major subclasses (Table 0.1)

based on the type of reaction catalyzed2,3. The focus of the classification is on the reaction being

catalyzed, not necessarily the enzyme doing the catalysis. Thus, enzymes from different

organisms, which catalyze the same reaction, would be given the same set of EC numbers.

Though this system is quite effective, it does not take into account enzymes that catalyze more

than one type of reaction. In fact, decades before this system was proposed, such catalytic-

promiscuity had been reported in pyruvate decarboxylase, which is able to catalyze dehydration

between acetaldehyde and benzaldehyde in addition to its namesake decarboxylation, both lyase-

type reactions4.

Table 0.1 | Enzyme Commission Subclasses Class Name Type of Reaction A + BH  AH + B 1 Oxidoreductases A + BO  AO 2 Transferases A + B-L  A-L + B

3 Hydrolases AB + H2O  AOH + BH 4 Lyases A-X-Y-B  AB + XY 5 Isomerases ABC  BCA 6 Ligases A + B  AB

Enzymes are incredibly alluring catalysts. They have high regio-, stereo-, and chemoselectivity. They function at ambient temperatures, in aqueous media, and can often be easily separated from the desired products. Expanding the catalytic space accessible with enzymes has been an active area of research for decades and has a wide range of applications5-7.

Enzymes are engineered for substrate selectivity with some frequency8-11. Less frequent,

however, is engineering an enzyme’s catalytic reactivity. That is, altering the type of reactions an 1

enzyme performs—possibly altering its Enzyme Commission subclass. Further, there are few

examples of utilizing an enzyme with a new substrate to alter the type of reaction catalyzed and

change the enzyme subclass.

The expansion of enzymes’ catalytic range is an appealing proposition: maintain regio-

and stereo-selectivity, but catalyze new or different, useful reactions. Many enzyme classes rely on bisubstrate interactions, including several which utilize a common cosubstrate:

methyltransferases and AdoMet; kinases and ATP; acetyltransferases and acetyl Co-A;

dehydrogenases and NAD+/NADP+. Substrate analogues can be used as tools for expanding the

catalytic activities of certain enzymes and interrogating an enzyme’s activity and specificity.

Catalytic Promiscuity

Catalytic promiscuity is the ability of an enzyme to catalyze a different type of reaction than that to which it has naturally evolved12-15. This differs from substrate promiscuity in which

an enzyme may perform its native reaction with a wide range of substrates. Catalytic promiscuity

has been an important factor in evolutionary diversification of enzymes12. However, compared to

investigations into substrate promiscuity, there are fewer reported examples of purposeful

enzyme engineering that results in altering the enzyme class16-27.

Prof. Frances Arnold’s group at CalTech has produced a dazzling array of new catalytic

activities from natural oxidoreductases cytochrome P450s—cyclopropanation16,17,20, sulfimidation18, aziridination19, amination21, and N-H insertion26. This catalytic promiscuity is

afforded by the rational “tuning” of the metal cofactor in combination with high throughput

screening using directed evolution28. In the native enzyme, the iron catalytic core is bound to the

enzyme through a cysteine. Modifying this linkage to a serine or histidine changes the redox

potential of the enzyme and allows access to different chemical spaces. These catalytic activities can then be coaxed from the enzyme through the use of multiple rounds of directed evolution

(Figure 0.1).

Figure 0.1 | Overview of Directed Evolution Starting with a known enzyme, a gene library of mutants can be created through error prone PCR and DNA shuffling. These genes can then be expressed and screened for desired catalytic activity. The selected mutants can undergo additional rounds of evolution to further refine desired traits.

Many purposeful efforts toward inducing catalytic promiscuity have centered on altering cofactors, such as active-site metals and ions29-35. Engineering enzymes through site-directed

mutagenesis has also been used to alter the catalysis types of enzymes36-42. More recently,

directed-evolution has gained traction as the key tool in expanding an enzyme’s catalytic space43-

47. Lipases, hydroxylases, and oxidoreductases have been major classes of enzymes engineered for catalytic promiscuity12-15,48. Though not as widely researched, methyltransferases are an appealing class of enzymes for engineering catalysis. They often utilize common cosubstrate S- adenosylmethionine (AdoMet, SAM) as a source of methyl for group transfer. Altering this common substrate is a tempting avenue for inducing catalytic promiscuity.

Methyltransferases

The methyltransferase family of enzymes (EC 2.1.1.x) is a large and varied class, catalyzing the methylation of a wide range of substrates49 (Figure 0.2). Targets of methylation include small molecules, proteins, RNA, and DNA50. Over 300 AdoMet-dependent methyltransferases have been identified51,52, many of which have unknown natural targets. More recently, methyltransferases have been acknowledged for having great potential as useful biocatalysts53-55.

Figure 0.2 | Enzyme-Catalyzed Methylation Methyltransferases catalyze the transfer of a methyl group from AdoMet to a nucleophile producing a methylated product and AdoHcy. Nucleophiles can be C, N, S, O, As or halide atoms found in a wide range of biomolecules.

AdoMet Analogues

Previously, various AdoMet analogues, which alter the sulfonium methyl group, have

been synthesized and employed as substrates for methyltransferases55-58 (Figure 0.3). Several

analogues59-61 have been designed with downstream, bioorthogonal reactions in mind; and,

others have applications in increasing the diversity of methyltransferase catalyzed products58,62.

There are also previously described analogues that use selenium in the place of sulfur57,63. In the case of a selenium propargyl analogue59, the selenonium improves the stability of the analogue

as compared to the corresponding sulfonium. Almost all of these, however, maintain the methyltransferase’s catalytic specificity as a transferase.

Figure 0.3 | Selected AdoMet Analogues with Enzyme Activity

Herein are described two AdoMet analogues, which can confer substrate-induced

catalytic promiscuity to methyltransfserases—Se-adenosylselenohomocysteine selenoxide

(SeAHO) and S-adenosylvinthionine (AdoVin). The selenoxide analogue of AdoMet, SeAHO, is shown here to oxidize catechol-O-methyltransferase substrates in an enzyme-dependent manner.

The vinyl sulfonium analogue, AdoVin, forms a bisubstrate adduct with thiophenols when catalyzed by thiopurine-S-methyltransferase.

Figure 0.4 | Structure of AdoMet Analogues, SeAHO and AdoVin

The sulfoxide analogue of AdoMet, S-adenosylhomocysteine sulfoxide (SAHO) has not

yet been detected in vivo, however, is a known methyltransferase inhibitor in vitro 64-68.The

selenium analogue of SAHO, SeAHO, has also not been detected from biological samples. With a redox potential distinct from sulfur69, it is expected that SeAHO may have altered reactivities

and characteristics when compared to SAHO. Chapter 1 describes its preparation and

characterization and Chapter 2 details its use as a substrate by catechol-O-methyltransferase, conferring putative oxidoreductase activity on the enzyme.

Unlike other alkyl-sulfonium analogues of AdoMet, the vinylsulfonium analogue,

AdoVin, does not undergo methyltransferase-catalyzed group transfer. Instead, thiopurine-S- methyltransferase catalyzes the formation of a bisubstrate adduct between AdoVin and a thiophenol substrate. There is some precedent for methyltransferase-catalyzed adduct formation70-72; however, the phenomenon had not previously been described with S-

methyltransferases. In Chapter 3, the activity and characterization of AdoVin and its bisubstrate

adducts are described. Chapter 4 outlines a platform for utilizing probes, such as AdoVin, for

detecting and identifying enzyme-substrate pairs from cellular contexts using native mass

spectrometry.

Chapter 1 Se-Adenosylselenohomocysteine Selenoxide

Adapted with permission from:

R.I. Duclos, D.C. Cleary, K.C. Catcott, Z.S. Zhou. “Synthesis and characterization of Se- adenosyl-L-selenohomocysteine selenoxide.” Journal of Sulfur Chemistry. 36(2), 135- 144, (2015). http://dx.doi.org/10.1080/17415993.2014.979173

R.I. Duclos synthesized the compounds, performed NMR analysis, collected and analyzed IR spectra, collected and analyzed MS spectra, collected and analyzed RPLC data, and wrote the manuscript. D.C. Cleary synthesized the compounds, collected and analyzed IR spectra, and reviewed the manuscript. K.C. Catcott developed aqueous preparations for the compounds, collected and analyzed IR spectra, collected and analyzed MS spectra, collected and analyzed RPLC data, performed and analyzed the redox reactions, and reviewed the manuscript.

Z.S. Zhou was principal investigator for this project and oversaw manuscript preparation.

1 Se-Adenosylselenohomocysteine Selenoxide

Selenium is an essential micronutrient for all animals and many other living organisms73-

82. However, a high level of selenium is toxic, thus, selenium metabolites should be maintained

within a fairly narrow concentration range of adequacy for the biosynthesis of the over 25 human

selenoproteins to balance deficiency and toxicity78,83. Organoselenium metabolites are only

present in trace amounts78,84, relative to the well-known sulfur analogues that include the amino

acids L-methionine and L-cysteine85, the biological methyl donor S-adenosyl-L-methionine54,86-88

and the byproduct of methylation S-adenosyl-L-homocysteine88-90. Oxidations of S-

adenosylhomocysteine are reported to give the sulfoxide SAHO64,68,89,91,92 and the corresponding

sulfone64,68,93. These have not been detected as metabolites in vivo, but as close structural analogs

of AdoMet, these analogues are methyltransferase inhibitors in vitro. As examples, sulfoxide

SAHO is an inhibitor of catechol-O-methyltransferase64,65 phenylethanolamine-N- methyltransferase64, histamine-N-methyltransferase64, protein II methyltransferase66, viral

mRNA methyltransferases67 and Escherichia coli cyclopropane fatty acid synthase68.

The selenium analogue of SAHO, Se-adenosylselenohomocysteine selenoxide (SeAHO),

has not yet been detected from biological samples, nor has its synthesis and properties been reported. With a redox potential distinct from sulfur69, it is expected that SeAHO may have some

altered reactivity and activities when compared to SAHO. Herein is described the preparation and characterization of SeAHO.

1.1 Introduction

Se-Adenosylselenohomocysteine selenoxide (SeAHO) was prepared from Se-

adenosylselenohomocysteine (SeAH), which was synthesized from selenomethionine by a

method that did not require any extractions or column chromatography. SeAHO was characterized by melting point, infrared spectroscopy (IR), mass spectrometry (MS), and nuclear magnetic resonance (NMR). SeAHO was stable in buffered aqueous environments with no evidence of glycosidic hydrolysis or electrocyclic eliminations over a wide (3–12) pH range at ambient temperature.

SeAHO is quite distinct from its sulfoxide analogue (SAHO): SeAHO undergoes hydration at the larger more polarizable selenium; it is racemized at the selenium center at low pH; and it is readily reduced by biological thiols—glutathione (GSH) and cysteine.

Though reduced by thiols, SeAHO is not reduced by thioethers, such as methionine, or several other amino acid sidechains. However, SeAHO is reduced in the presence of some methyltransferases, though concurrent oxidation of these enzymes was not detected.

(a) (b)

Figure 1.1 | Structure of AdoMet, SAHO, SeAHO, and SeAH (a) S-Adenosyl-L-methionine (AdoMet, SAM) is the common methyl donor used by methyltransferases in biological systems. (b) S-Adenosylhomocysteine sulfoxide (SAHO) is the sulfoxide analogue of AdoMet and a known methyltransferase inhibitor. (c) Se-adenosylselenohomocysteine selenoxide (SeAHO) is the selenoxide analogue of AdoMet. (d) Se- adenosylselenohomocysteine (SeAH) is the selenide analogue of AdoMet and the product of SeAHO reduction.

1.2 Preparation

Se-Adenosylselenohomocysteine selenoxide (SeAHO, Figure 1.1(c)) was prepared from

Se-adenosylselenohomocysteine (SeAH, Figure 1.1(d)), which was synthesized as reported94 by

Dr. Richard Duclos and Dillon Cleary. Briefly, SeAH was prepared by a variation of the

previous literature via L-selenohomocysteine59,95. Ammonia was added by condensation using a

dry ice condenser from a cylinder of anhydrous ammonia to a round bottom flask that was 11

prepared with an inert atmosphere and L-selenomethionine in a −80 °C dry ice/acetone bath.

This reaction was stirred magnetically and small pieces of sodium metal were added until the reaction remained blue—about three molar equivalents. The reaction was stirred in darkness for an additional 1 hour, and then ammonium chloride was added slowly to neutralize any sodium amide present. The reaction was then removed from the dry ice/acetone bath and nitrogen was blown over the stirred mixture to remove solvent. The residue dried under vacuum to give L- selenohomocysteine56,59,96 as a white solid that was dissolved in water. To this solution was

added 5’-chloro-5’-deoxyadenosine97,98. To this mixture was then added 10% aqueous sodium

hydroxide solution and additional water. This reaction mixture was stirred magnetically for 1.5

hours at 80 °C. The solution remained cloudy for 15 minutes before becoming homogeneous.

Acetic acid was then added dropwise until the solution pH was between 5 and 6. Solvent was

removed under vacuum to give a yellowish white solid. This solid was dissolved in an

ethanol/toluene solution, the solvents removed by rotary evaporation, and the residue dried under

vacuum to remove all moisture. Two hot filtration/recrystallizations from methanol gave 24%

overall yield of SeAH as a gray amorphous solid. Trace amounts of residual methanol were

removed by dissolving SeAH at a concentration of 5 mg/mL in deionized water (pH 6), freezing

with dry ice, and lyophilizing to give a white solid.

From the selenide SeAH, the selenoxide SeAHO is easily prepared. Initially, solid SeAH

was dissolved in neat acetic acid producing a cloudy, grey solution. Hydrogen peroxide was then

added to about a 45-times molar excess. Within 30 seconds, the solution turned a clear,

homogenous, colorless solution. Complete conversion to the selenoxide occurred within 5

minutes as monitored by TLC, which was carried out on plastic-backed silica gel 60, PE SIL G

Whatman Plates UV254. In 12:1:3, isopropanol:water:acetic acid, SeAHO had an Rf of 0.27; and

in 12:5:3, isopropanol:water:acetic acid, it had an Rf of 0.45. SeAHO was frozen at -80 °C then placed on dry ice and put under vacuum for an hour to remove the solvent. The melting point was determined to be 115-120 °C.

Figure 1.2 | Synthesis of SeAH The S-methyl group of commercially available selenomethionine was cleaved by ammonia and sodium metal, converting it to selenohomocysteine. After neutralization with ammonium chloride, selenohomocysteine was then combined with 5’-Cl-5’-deoxyadenosine in aqueous sodium hydroxide at 80 °C to produce Se-adeosylselenohomocysteine, which was then neutralized with acetic acid.

The oxidation was also performed in aqueous solution by dissolving SeAH in deionized

water or 50 mM potassium phosphate buffer at pH 3, pH 7, or pH 12. For low and neutral pH

solutions, a combination of sonication and vortexing was used over 30 minutes to dissolve

SeAH. In the basic solution, SeAH dissolved instantly and produced a clear solution. Then 5-

times molar excess hydrogen peroxide was added. The reaction was mixed by inversion over 10

minutes.

For applications that required excess hydrogen peroxide to be removed, catalase (EC

1.11.1.6) was used. A catalase stock solution was prepared (Sigma PN: C1345) at 10 mg/mL in

50 mM Tris buffer at pH 8. About 1 µL of the catalase stock solution was added to the SeAHO

reaction. The solution was briefly vortexed and left at room temperature until the solution

stopped bubbling—about 2 minutes. The catalase was then removed by ultrafiltration by

centrifugation through a 10 kDa MWCO membrane. The SeAHO concentration was then

-1 -1 determined by A260nm using an extinction coefficient of 15,400 cm M (for the adenine moiety).

SeAHO prepared in this manner was stored frozen at -80 °C. Under these conditions, SeAHO

shows no significant degradation for at least six months.

1.3 Characterization

1.3.a NMR

NMR spectra were recorded on a Bruker 700 MHz spectrometer (details in Section 1.7.a).

1 The H NMR of SeAHO in 600:9:1, deuterated acetic acid:H2O:H2O2, an organic acidic environment near the pKa’s of the carboxylic acid group and the adenine residue, was clean and

assignable (Figure 1.3, Table 1.1). The selenoxide has a distinctly different conformation than

the corresponding sulfoxide SAHO as evidenced by the dramatic downfield shift of the α-proton

in the NMR. For SAHO prepared in parallel to SeAHO, hydrogen peroxide oxidation in acetic

acid solution according to a literature report68 gave nearly equal ratios of the R and S sulfur- epimers at the newly chiral sulfur center for the product SAHO. These two diastereomers had distinctly different resonances in the proton NMR for the diastereotopic α, γ, 2’, 3’, 4’, 5’, and for one aromatic resonance.

1 Figure 1.3 | H NMR of SeAHO in 97:3 CD3CO2D:30% Aqueous H2O2

Table 1.1 | 1H Shifts of SAHO and SeAHO in Deuterated Acetic Acid SAHO is present as a racemic mixture at the S=O. SeAHO contains residual

aqueous H2O2.

Proton assignment SAHO in CD3CO2D SeAHO in 97:3 CD3CO2D:30% aqueous H2O2 α 4.15, 4.17 4.58-4.67 βa 2.49-2.58 2.74-2.83 βb 2.44-2.53 2.60-2.69 γa 3.28-3.37, 3.28-3.37 3.84-3.95 γb 3.19-3.25, 3.28-3.37 3.84-3.95 1’ 6.19 6.20 2’ 4.88, 4.93 4.94 3’ 4.64, 4.67 4.72 4’ 4.57-4.62, 4.60-4.65 4.58-4.67 5a’ 3.54, 3.61 3.84-3.95 5b’ 3.54, 3.55 3.65 Ar 8.42 8.45 Ar 8.43, 8.46 8.48

The 13C NMR was recorded from the same SeAHO sample after 1H data were acquired

(Figure 1.4, Table 1.2). Some hydrolysis of the C1’-adenine glycosidic bond was observed.

Figure 1.4 | COSY NMR of SeAHO in 97:3 CD3CO2D:30% Aqueous H2O2

Table 1.2 | 13C Shifts SeAHO in Deuterated Acetic Acid

Carbon assignment SeAHO in 97:3 CD3CO2D:30% aqueous H2O2 β 32.9 γ 44.6 5’ 49.1 3’ 73.9 2’ 74.5 4’, α 80.7 1’ 91.0 Ar 142.7 Ar 149 The carbon assignments of 2’, 3’, 4’, and α carbons are tentative due to hydrolysis of the C1’-adenine glycosidic bond during the course of the experiment.

1.3.b IR Spectroscopy

SeAHO prepared from acetic acid then lyophilized was tested by IR spectroscopy (Figure

1.5) on a Bruker Alpha-P FT IR as detailed in Section 1.7.b. Some carbonyl stretch was observed at 1777 cm−1 in the IR spectrum (Figure 1.6(a)), characteristic of α-amino carboxylic acids at low pH99. SeAHO also showed characteristic Se=O stretching bands100 in the IR spectrum at 847 and

878 cm−1 (Figure 1.6(b)). In the sulfoxide, SAHO, the peaks corresponding to characteristic of sulfoxide stretching91,92 are at 975 and 998 cm−1. Additionally, the sulfoxide carbonyl stretch is not shifted to as high a frequency as the selenoxide (Figure 1.7).

Figure 1.5 | IR Spectrum of SeAHO and SeAH Full IR spectrum of SeAHO (red, top) and SeAH (blue, bottom). Note the SeAHO carbonyl peak at 1777cm-1 where thec orrespondins SeAH band is likely obscured by peaks at lower frequncies; and the SeAHO bands at 878 and 847 cm-1 due to selenoxide stretch, not observed in SeAH.

(a) (b)

Figure 1.6 | Zoomed in IR Spectrum of SeAHO and SeAH IR spectrum of SeAHO (red, top) and SeAH (blue, bottom) zoomed in on (b) SeAHO carbonyl peak at 1777cm-1 where the corresponding SeAH band is likely obscured by peaks at lower frequncies; and, (c) SeAHO bands at 878 and 847 cm-1 due to selenoxide stretch, not observed in SeAH.

Figure 1.7 | IR Spectrum of SeAHO and SAHO As compared to the sulfoxide (blue), the shift of the carbonyl band (1777 cm-1) is much more pronounced in SeAHO. Additionally, note the presence of the selenoxide stretch bands at 878 and 847 cm-1. Sulfoxide stretch bands are present ~1000 cm-1.

1.3.c Mass Spectrometry

The mass spectra of SeAHO were obtained by liquid chromatography mass spectrometry on a quadrupole ion trap (LCQ-MS), in an acidic environment (acetonitrile/water/0.1% formic acid), as well as by matrix-assisted laser desorption ionization time of flight mass spectrometry

(MALDI-ToF) (details in Section 1.7.c). The base peak seen in all MS methods (m/z 431) could be several possible cyclic or elimination products of SeAHO (Figure 1.8(e)).

(a) (b)

(c)

NH2 (d) N N

H2C O N N

HO OH adenosyl cation + m/z 250.09 M (e)

NH2 NH2 N N N N HO2C Se HO2C Se O N N N O N NH2 NH2 HO OH HO OH

NH NH 2 HO C 2 O 2 NH O N N N N Se Se H2N O N N O N N

HO OH SeAHO dehydration product HO OH + m/z 431.06 M Figure 1.8 | SeAHO Fragments Observed by Mass Spectrometry

The molecular ion of SeAHO (m/z 449, Figure 1.8(a)) was observed by the LCQ-MS

(Figure 1.9). The dehydration product (m/z 431, Figure 1.8(e)) was also observed. The extracted

ion chromatograms for these two peaks overlap at ~2.2 minutes (Figure 1.10) with the

dehydration product also eluting at ~4.77 minutes (Figure 1.9, Figure 1.10). This suggests that

the dehydration product (Figure 1.8(e)) is both a degradation product of SeAHO as well as a product caused by in source fragmentation.

m/z

Figure 1.9 | LCQ-MS Spectrum of SeAHO SeAHO (m/z 449) and the dehydration product (m/z 431) are both readily seen by LCQ-MS.

Figure 1.10 | XIC from the LCQ-MS of Peaks in the SeAHO Sample The dehydration product (m/z 431) appears to coelute with SeAHO (m/z 449) suggesting some in source dehydration. The hydrated compound (m/z 467) is also present.

Collision induced dissociation (CID) was used to collect tandem MS data for both

SeAHO and the dehydration product (Figure 1.11). The adenosyl cation is easily observed from

the elimination product precursor, but not seen in the SeAHO precursor. It appears that the Se-

C5’ linkage is thus weaker in the dehydration product than in SeAHO.

Figure 1.11 | MS-MS Spectra of SeAHO and Dehydration Product MS-MS data for CID of precursor ions m/z 431 (dehydration product, top) and m/z 449 (SeAHO, bottom). The adenosyl cation is easily observed from the dehydration product precursor, but not seen in the SeAHO precursor. It appears that the Se-C5’ linkage is weaker in the dehydration product than in SeAHO.

In the MALDI-MS spectrum (Figure 1.12), the molecular ion of SeAHO (m/z 449, Figure

1.1(a)), as well as the hydrated ion (m/z 467, Figure 1.8(b)), and dehydration product (m/z 431,

Figure 1.8(e)) are all present. The CID tandem MS fragmentations of SeAHO (m/z 449) and the hydrate (m/z 467) were again distinct from the fragmentation of the dehydration product (m/z

431) (Figure 1.13). The dehydration product again shows strong fragmentation across the Se-C5’ bond producing the adenosyl cation (m/z 250, Figure 1.8(c)) as well as cleavage of adenine (m/z

136, Figure 1.8(d)). Fragmentation of SeAHO and the hydration product produces a selenium containing ion at m/z 395 as well as a fragment at m/z 256 and the adenosyl cation (m/z 250).

m/z

Figure 1.12 | MALDI-ToF Spectrum of SeAHO The SeAHO molecular ion (m/z 449) is present as well as the dehydration product (m/z 431).

(a)

(b)

(c)

Figure 1.13 | MS-MS Spectra of MALDI-ToF of SeAHO Collision induced dissociation was performed on various precursor ions. (a) SeAHO precursor (m/z 449), (b) hydrated precursor (m/z 467), (c) dehydration product precursor (m/z 431).

1.4 Stability

1.4.a NMR Time Course

In order to test the stability of SeAHO, we ran a brief 1H NMR time course (details in

Section 1.7.a), analyzing the spectra at 10 minutes, 40 minutes, and 2 hours after oxidation of

SeAH was performed in acetic acid. The C1’-adenine glycosidic linkage was found to be labile and underwent some hydrolysis (Figure 1.14). No selenoxide elimination was observed, likely due to the lack of aromaticity of the adenosine group101,102.

(a)

6 6 5 5 4 4 3 3 2 (b) Ch 0.20 SAH_RD.009.001.1r.esp

0.15

0.10 Normalized IntensityNormalized

0.05

6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 Chemical Shift (ppm)

0.13

0.12

0.11

0.10

0.09

0.08

0.07

0.06 Normalized Intensity Normalized 0.05

0.04

0.03

0.02

0.01

6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 Chemical Shift (ppm)

Figure 1.14 | Time Course NMR of SeAHO NMR spectra of SeAHO (a) at 10 minutes after preparation, (b) at 40 minutes after preparation, and (c) at 2 hours after preparation.

1.4.b Aqueous Buffers at Various pH

Characterization of SeAHO in an aqueous environment was of greater biological significance than in acetic acid. The SeAHO was completely stable to hydrolysis of the C1’- adenine glycosidic bond in 50 mM potassium phosphate buffered aqueous solutions for at least 3 hour at ambient temperature over the wide pH range of 3–12 as no elimination or other degradation products were seen by HPLC (Figure 1.15) or proton NMR (Figure 1.16). Reverse phase HPLC was run on an Agilent 1100 using an Alltech Econosil C18 (4.6 x 250 mm) and a gradient described in Section 1.7.d.

(a)

(b)

(c)

(d)

Figure 1.15 | Stability of SeAHO in Aqueous Buffers at Various pH Values C-18 RPLC traces at 275 nm of SeAHO in 50 mM potassium phosphate at (a) pH 3, (b) pH 7, and (c) pH 12. (d) Adenine standard. Very little hydrolysis of the glycosidic linkage is observed at any pH.

rd-seaho-ph3.003.001.1r.esp (a) M02(s) Water M03(m) 8.37

0.025

0.020

M04(m) 4.91

0.015 8.36 6.13 Normalized Intensity Normalized 4.91

0.010 4.90

0.005

0 0.90 0.58 0.37 1.00

8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 Chemical Shift (ppm) rd-seaho-ph7.004.001.1r.esp (b) M03(dd)

J(M02)=9.82 Hz

Water M01(m)

M05(t) 8.21 4.87 0.05 4.91

M04(m) 8.28 8.20 0.04 M06(m) 6.08 6.07

4.92 M07(m) M08(m) 0.03 8.29 Normalized Intensity Normalized

4.53 M10(m) 3.84 6.07 6.06 4.49 3.82 4.54 0.02 3.83 M09(m) 4.57 3.67 3.77 3.65 3.65 3.85 4.36 3.81 2.33 2.36 4.48 2.35 2.36 3.45 2.32 2.37 4.36 0.01 3.63 4.59 2.54 4.64 2.59 3.22 3.17 3.17 2.31 2.52 2.52

0 0.03 0.32 0.73 0.93 1.00 0.06 1.70 2.42 2.43 0.45 0.01 0.00 0.40 1.09 0.75 1.09 0.24 0.34 0.31 0.21 0.23 0.300.34 0.34 1.35

8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 Chemical Shift (ppm)

rd-seaho-ph12.003.001.1r.espM01(s) (c) M02(s) Water

M04(s) M06(s) M08(dd) M03(s) M05(d) M07(m) M09(dd) M10(m) 8.22 0.0045 8.29

0.0040

0.0035 6.08 0.0030 8.32 6.07 0.0025 Normalized Intensity Normalized 4.80 4.84

0.0020 4.94 4.47 0.0015 3.71

0.0010 3.70 3.73 3.65 3.64 3.66 3.67 2.12 1.98 1.97 2.13 2.11 2.11 0.0005 1.99

0 0.41 0.53 0.40 0.55 1.00 0.54 0.49 0.43 0.33 1.61

8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 Chemical Shift (ppm) Figure 1.16 | 1H NMR of SeAHO at Various pH values

SeAHO in D2O phosphate buffer at (a) pH 3 (6:4 mix), (b) pH 7 (3:3:2:2 mix), and (c) pH 12 (6:4 mix).

2.5

3.0

3.5

4.0

4.5 F1 Chemical Shift (ppm) 5.0

5.5

6.0

6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 F2 Chemical Shift (ppm) Figure 1.17 | COSY of SeAHO

SeAHO in D2O phosphate buffer pH 7 present in a 3:3:2:2 mix. With 40% of the material showing α-proton shifted downfield, and likely a 50:50 mixture of selenoxide and a hydrate.

1 H NMR (details in Section 1.7.a) was carried out at various pH values in D2O phosphate buffered solutions (Figure 1.16). SeAHO generally appeared to be a 60:40 mixture in 50 mM phosphate buffers of D2O at pH values of 3 and 7. High resolution NMR at pH 7 (Figure 1.17)

showed that 40% of the material had the α-proton shifted downfield, and that SeAHO was also a

50:50 mixture, likely to be a mix of selenoxide and a hydrate, analogous to the reported data for

selenomethionine selenoxide103-105. SeAHO was presumably mostly a hydrate at pH 3 and mostly

in the selenoxide form at pH 12 where only a very small amount of decomposition was observed.

The coordination of the α-amino acid moieties with the selenoxide functional group of

selenomethionine selenoxide in aqueous solutions have already been proposed at acidic105, neutral104,105, and basic103,105 pH values. The protonation, hydration, and racemization of the

selenoxide functional group at low pH are also well known106,107. These data correlate well with

our NMR data for SeAHO in the acidic organic (Figure 1.3) and in the phosphate buffered

aqueous environments at pH values 3, 7, and 12 (Figure 1.16). The characteristic coordination of

the selenoxide can be intermolecular or intramolecular (Figure 1.18). The interaction of the α-

amino acid moiety of SeAHO with the selenoxide/hydrated group results in deshielding of the α- proton completely in deuterated acetic acid (Figure 1.3, Figure 1.18(c)) and to the extent of about

40% in neutral buffered aqueous solutions (Figure 1.16).

Partial protonation or a conformational difference in the adenosine moiety may account for the two sets of aromatic and glycosidic protons seen in pH 3 D2O phosphate buffer (Figure

1.16(a)) that were not seen in deuterated acetic acid solution by proton NMR (Figure 1.3).

(a) (e)

NH2 NH2

HO2C HO2C OH Se Se O OH

adenosine adenosine

(b) (c) (d) O O O OH H OH H OH H H NH2 O NH2 O NH2 Se Se Se H H H O O O adenosine adenosine adenosine

Figure 1.18 | Hydration and Racemization of SeAHO SeAHO must have intermolecular or intramolecular (as shown) coordination of the acidic protons of the ammonium (as shown) or carboxylic acid groups with the selenoxide oxygen, and can exist in the selenoxide (a) or a hydrate (e) form in acidic or neutral aqueous solution.

1.5 Reactivity

In our efforts to understand the reactivity of SeAHO, we researched the reactions of

related compounds. The redox biochemistry of methionine is well understood103,108,109.

Selenomethionine is also easily oxidized with biological oxidants such as hydrogen peroxide to

give a mixture of selenoxide selenomethionine and a hydrate in the neutral pH range103-105. NMR data are pH dependent, and only single compounds are seen at low pH103,110 and at high pH103,105.

These data are consistent with studies of other selenoxides106,107. Selenomethionine selenoxide is

homogeneous by HPLC and stable at ambient temperature111,112. Homocysteine113,

selenohomocysteine96,114, cysteine109,115-118, and selenocysteine69,109 also undergo oxidations.

Interestingly, very little is known about the biochemical activity and redox chemistry119 of

SeAH56,59,80,81,120, which is structurally related to its methyl analogue selenomethionine78,84,103-105 and to its corresponding sulfur analogue S-adenosylhomocysteine88,89.

1.5.a Reactivity with Biological Thiols and Thioethers

SeAHO was readily reduced back to SeAH at ambient temperature by glutathione (GSH) and cysteine, but not by thioethers methionine or S-adenosylhomocysteine (SAH) as observed by

C18 HPLC (Figure 1.19, Section 1.7.d) with absorbance monitored at 260 nm. SAHO was not reduced under these biological conditions with glutathione (GSH) or cysteine (Figure 1.20). As reported121,122, the reduction of sulfoxides generally requires more forcing conditions.

Reduction of the selenoxide by thiols is not unexpected. The redox potentials of generic selenoxides and disulfides suggest that in a mixed system, the selenoxide would tend to be reduced to the selenide and the thiols oxidized to the disulfide69.

(a)

(b)

(c)

(d)

Figure 1.19 | SeAHO—Reduced By Thiols, Not By Thioethers C18 HPLC chromatograms at 260 nm. (a) SeAHO standard ~2.3 minutes. (b) SeAHO treated with 50-times molar excess of cysteine completely reducing it to SeAH (~6.5 minutes). (c) SeAHO treated with 50-times molar excess of methionine, showing no reduction. (d) SeAHO treated with one molar equivalent of S-adenosylhomocysteine (SAH, ~5.2 minutes), showing no reduction the selenoxide.

(a)

(b)

Figure 1.20 | SAHO—Not Reduced by Thiols C18 HPLC chromatograms at 260 nm of SAHO (~2.3 minutes) treated with 50-times molar excess of (a) glutathione and (b) cysteine and monitored at 260 nm. No reduction was observed.

1.5.b Reactivity with Whole Proteins

The reactivity of SeAHO with whole proteins was also of interest. As a potential methyltransferase substrate or inhibitor, we tested incubating SeAHO with two methyltransferases and two other proteins and monitored for reduction. SeAHO was incubated separately with catechol-O-methyltransferase (COMT), thiopurine-S-methyltransferase (TPMT), bovine serum albumin (BSA), and lysozyme at a 10-to-1 molar ratio for 1 hour at room temperature. Each reaction was then centrifuged using a 10 kDa MWCO ultrafiltration device to

remove the protein. The flow through containing any small molecules was then tested by HPLC using the parameters described in Section 1.7.d.

SeAHO was partially reduced by both COMT and TPMT, and to a lesser extent by BSA.

No reduction was seen with lysozyme (Figure 1.21). The reaction with BSA is likely due to its free cysteine123,124, as SeAHO is reduced by cysteine in vitro.

(a)

(b)

(c)

(d)

Figure 1.21 | SeAHO—Reduced in the Presence of Some Proteins C18 HPLC chromatograms at 260 nm of SeAHO incubated with (a) catechol-O- methyltransferase (COMT), (b) thiopurine-S-methyltransferase (TPMT), (c) bovine serum albumin (BSA), and (d) lysozyme. SeAHO (~2.2 minutes) was partially reduced to SeAH (~6.0 minutes) in the presence of COMT, TPMT, and BSA. No reduction was observed in the presence of lysozyme. 36

To investigate further, SeAHO was incubated with additional amino acids and analogues to see if other chemical groups (i.e., primary amines) were capable of reducing the selenoxide.

Histidine, lysine, aspartic acid, tryptophan, and tyrosinamide were incubated with SeAHO in 10- fold molar excess for 1 hour. Reactions were then analyzed by HPLC (details Section 1.7.d). No reduction of SeAHO was observed under any of these conditions (Figure 1.22).

(a)

(b)

(c)

(d)

(e)

Figure 1.22 | Reactivity of SeAHO with Various Amino Acids Reverse phase HPLC chromatograms at 260nm of SeAHO (~2.2 minutes) treated with 10-fold molar excess of (a) histidine, (b) lysine, (c) aspartic acid, (d) tryptophan, and (e) tyrosinamide. No reduction was observed. Small peak observed at ~6.1 minutes was present in SeAHO control.

It is possible that though free amino acids may not reduce SeAHO in vitro, in the context of the whole protein, reduction may occur, along with the concurrent oxidation of a protein residue. The next step taken was to treat COMT and TPMT with SeAHO, then digest the proteins and look for oxidation via mass spectrometry (method details Section 1.7.e). Upon analyzing the spectra, it was found that no significant difference existed between the SeAHO treated proteins and the control samples (Figure 1.23).

Figure 1.23 | MALDI-MS Spectra of SeAHO-Treated COMT COMT was treated with 10-fold molar excess of SeAHO (top) or AdoMet (bottom), then trypsin digested and run on MALDI-MS. Fragment Y130-K144 (m/z 1897) appears unoxidized in both samples. Across the entire spectra, no clear oxidation peaks were observed in the SeAHO treated sample.

The fact that SeAHO is reduced in the presence of some methyltransferases, however the

methyltransferases themselves do not appear to be oxidized, suggested that the selenoxide may

be reduced due to some side reaction catalyzed by the enzyme. This could take the form of H2O2 formation when SeAHO binds to the enzyme active site. However, peroxide was not detected in samples of SeAHO incubated with COMT using starch iodide paper (Whatman PN: 2651-500), nor by monitoring gas evolution using catalase (EC 1.11.1.6).

1.6 Conclusions

SeAHO can be prepared in aqueous buffered solution at a variety of pH values from

SeAH. Minimal cleavage is seen at the glycosidic linkage in aqueous environments, though some

dehydration reaction may occur as observed by MS. Unlike the equivalent sulfoxide, SeAHO is

readily reduced by thiols, cysteine and glutathione, and no reduction is observed in the presence

of thioethers.

SeAHO was also found to be reduced in the presence of methyltransferases, COMT and

TPMT. However, oxidation of the enzymes was not observed by bottom-up MS techniques.

The reaction mechanism of many methyltransferases is essentially proximity—even

mutation of catalytic bases often do not destroy enzyme activity125,126. The substrates—AdoMet

and the appropriate nucleophile—are bound close together in the proper orientation and shape to

promote methyl transfer127,128. Given that COMT appears to catalyze reduction of SeAHO, it was

of interest to further investigate potential COMT-catalyzed oxidation of a substrate using

SeAHO.

1.7 Experimental Procedures

1.7.a NMR

NMR spectra were generously recorded by Dr. Jason J. Guo on a Bruker 700 MHz spectrometer. Samples were held at 4 °C during collection.

1.7.b IR Spectroscopy

IR spectra were obtained on a Bruker Alpha-P FT IR with OPUS software.

1.7.c Mass Spectrometry

The LCQ-MS spectra were acquired on a Finnigan LCQ ESI ion trap MS with an Agilent

1100 front end using an Apollo C18 column (5 μm, 4.6 x 150 mm), UV detection at 275 nm and the gradient described in Section 1.7.d.

MALDI-MS was done on an Applied Biosystems SCIEX 5800 TOF/TOF using a matrix of α-cyano-4-hydroxycinnamic acid and trifluoroacetic acid.

1.7.d Reverse Phase HPLC

Reverse phase HPLC was run on an Agilent 1100 using a Grace C18 column (4.6 x 150 mm) unless stated otherwise. The gradient was run at 1 mL/min starting at 98% H2O with 0.1% formic acid, 2% acetonitrile with 0.1% formic acid for the first 2 minutes, increasing to 40% acetonitrile with 0.1% formic acid by 22 minutes and 90% acetonitrile with 0.1% formic acid by

27 minutes. The gradient held at 90% acetonitrile with 0.1% formic acid for 5 minutes then ramped back down to 2% acetonitrile with 0.1% formic acid over 30 seconds. The gradient held at 98% H2O with 0.1% formic acid, 2% acetonitrile with 0.1% formic acid for the rest of the run.

1.7.e Methyltransferase-Catalyzed Reduction and Bottom-Up MS

TPMT and COMT each were incubated with a 10-fold molar excess of SeAHO or

AdoMet, as a control. These were incubated overnight at 37 °C overnight. The reactions were then buffer exchanged into a 50 mM Tris buffer at pH 7.9 using repeated ultrafiltration and dilution with a 10 kDa ultrafiltration membrane. This was done until no small molecule could be detected in the flow through by monitoring absorbance at 260 nm.

About 7.5 µg of each enzyme was then digested by trypsin using Promega's immobilized trypsin spin column (PN: V9012), following their method. Briefly, the resin was rinsed with 50 mM ammonium bicarbonate. Then the protein—diluted in 40% acetonitrile and 50 mM ammonium bicarbonate—was applied to the column and incubated for 30 minutes at room temperature. Then the peptide fragments were recovered from the column using 40:60 acetonitrile : ammonium bicarbonate with 0.2% TFA.

The fragments were then acidified with TFA to pH<3 and purified using EMD-

Millipore's C18 ZipTips (PN: ZTC18M096), following their method. Briefly, the tips were washed with acetonitrile, 0.1% TFA in 50% acetonitrile, and aqueous 0.1% TFA. The peptide solution was applied to the tips. Tips were washed with aqueous 0.1% TFA and then eluted with

0.1% TFA in 70% acetonitrile. These elutions were then concentrated using a speedvac at moderate temperature (30-40 °C) until sample sizes had reduced to 1-2 µL. They were then combined 1:1 with α-cyano-4-hydroxycinnamic acid (CHCA) matrix and spotted on MALDI target plate (~1-2 µL /spot).

MALDI-MS was done on an Applied Biosystems SCIEX 5800 TOF/TOF. For each sample, multiple scans were acquired to build up the signal.

Chapter 2: Engineering Methyltransferase Activity: Conversion to an Oxidoreductase

Adapted from:

K.C. Catcott, Z.S. Zhou. “Methyltransferase-Mediated Oxidation Using Selenoxide Analogue of S-Adenosylmethionine.” In preparation.

K.C. Catcott conceived of, performed, and analyzed the experiments. Z.S. Zhou was principal investigator for this project and oversaw manuscript preparation

2 Engineering Methyltransferase Activity: Conversion to an

Oxidoreductase

Enzymes are engineered for substrate selectivity with some frequency8-11; however, less

frequent are examples of enzyme engineering that results in altering the enzyme class16-27.

Notably, the Arnold group at CalTech has done extensive work using directed evolution to coax an amazing array of catalytic diversity from cytochrome P450s: adding cyclopropanation16,17,20, sulfimidation18, aziridination19, amination21, and N-H insertion26 functionality to the enzymes.

The methyltransferase family is a large and diverse group of enzymes with hundreds of

unique enzymes currently identified127,129. The vast majority are classified as S-

adenosylmethionine (AdoMet, SAM) dependent using AdoMet as the methyl source for these

reactions49,127. Methyltransferases catalyze the transfer of a methyl group to a nucleophile—C, S,

N, O, As, and, even halides—producing the methylated product and S-adenosylhomocysteine

(AdoHcy, SAH); these targets of methylation can be small molecules, proteins, DNA, and

RNA49,53,127,129.

Clearly, methyltransferases are a class of enzymes that provide great potential as useful

biocatalysts53-55. Previously, various AdoMet analogues, which alter the sulfonium methyl group,

have been synthesized and employed as substrates for methyltransferases55-57. Some of these

groups lend themselves to downstream, bioorthogonal reactions, such as propargyl59 ketone60, or azido61 analogues; and, others increase the diversity of products possible, such as ethyl and

propyl analogues62. There are also previous examples of using selenium in the place of sulfur in

AdoMet analogues57,63—most notably, the synthesis and use of a selenium propargyl analogue59,

where the selenonium greatly enhances the stability of the molecule over the corresponding

propargyl sulfonium.

2.1 Introduction

Methyltransferases are a broad class of enzymes that catalyze methyl-transfer from S- adenosylmethionine (AdoMet, SAM) to a nucleophile. AdoMet analogues can be used by these enzymes to alter the transfer reaction. Analogue Se-adenosylselenohomocysteine selenoxide

(SeAHO) is shown here to act with catechol-O-methyltransferase (COMT) to oxidize catechol substrates of the enzyme: that is, it has putative oxidoreductase activity. SeAHO opens up the possibility of utilizing the diversity and specificity of the methyltransferase family in a new class of reactions for biocatalysis.

2.2 Catechol-O-Methyltransferase

Catechol-O-methyltransferase (COMT) is a widely-studied, Mg2+-dependent, small

molecule methyltransferase, which methylates several neurotransmitters in vivo as well as some o-benzenediol-containing xenobiotics130-133. COMT is known to be inhibited by methylation

product S-adenosylhomocysteine (AdoHcy, SAH) and by the sulfoxide analogue of AdoMet,

SAHO65. COMT is perhaps most famous for its role in the metabolism of L-dopa134,135, a drug

used to treat Parkinson’s disease. COMT inhibitors, tolcapone and encapone, have been

approved by the FDA to be used in combination with a dopamine-analogue therapy136-139.

The soluble form of human catechol-O-methyltransferase (COMT, EC 2.1.1.6) with an

N-terminal histidine tag was purified from a transfected E. coli cell line previously reported140.

Details of the method can be found in Section 2.6.a.

Figure 2.1 | Catechol-O-Methyltransferase Reaction Catechol-O-methyltransferase (COMT) catalyzes the methylation of catechols to the corresponding guaiacol.

2.3 Oxidation of Epinephrine

The initial COMT substrate chosen for testing SeAHO activity was epinephrine

(adrenaline). Epinephrine has a known oxidation product adrenochrome (Figure 2.2)141-146.

Additionally, the UV absorbance spectra of epinephrine and adrenochrome are distinct, with epinephrine having an absorbance maximum at 265 nm and adrenochrome having two maxima at 310 nm and 485 nm (Figure 2.3). It was posited that COMT would bind and orient SeAHO and epinephrine to mediate oxidation of the catechol moiety to an orthoquinone (Figure 2.4).

This product quickly cyclizes to adrenochrome145,147-149.

Figure 2.2 | Oxidation of Epinephrine Epinephrine is oxidized to an orthoquinone, which cyclizes to adrenochrome.

Figure 2.3 | UV Spectra of Epinephrine and Adrenochrome Epinephrine (blue) and its oxidation product adrenochrome (red) have distinct UV spectra.

Figure 2.4 | Proposed Epinephrine Oxidation Mechanism (a) In the presence of COMT, SeAHO forms a peroxide with deprotonated epinephrine followed by loss of water and formation of the orthoquinone, which (b) undergoes base-catalyzed cyclization. This cyclization could also be catalyzed by another equivalent of SeAHO.

Adrenochrome was purchased from Sigma (PN: A5752) and a standard was prepared in

-1 -1 144 water (ε480nm 4020 cm M ) . It was used to confirm retention time (~4.6 minutes) on reverse phase HPLC (method described in Section 1.7.d), as well as establish a standard curve relating concentration to area under the curve (AUC) in the HPLC chromatograms (Figure 2.5).

(a) (b)

Figure 2.5 | Adrenochrome Retention Time and Standard Curve Adrenochrome standards were tested on reverse phase HPLC and detected by

A480nm to determine (a) retention time, and (b) the correlation between the area under the curve and injection amount.

Epinephrine will autoxidize to form adrenochrome upon standing142,143. To minimize this, reactions were run at 4 °C. Still, the background reaction was detectable (Figure 2.6), and subtracted from the enzyme containing reactions in order to quantify only enzyme catalyzed adrenochrome formation.

Figure 2.6 | Background Adrenochrome Formation Reaction of epinephrine and SeAHO in the absence of COMT. Adrenochrome concentration was determined by reverse phase HPLC.

To test the COMT catalyzed oxidation of epinephrine using SeAHO, reactions containing

0.5 mM epinephrine, 0.5 mM SeAHO (preparation described in Section 2.6.b), and various

concentrations of COMT (0, 6.25, 12.5, or 25.0 µM) were set up in the manner described in

Section 2.6.c and containing 8 µM adenosylhomocysteine nucleosidase (MTAN, EC 3.2.2.9)—

an enzyme which cleaves SAH and SeAH to eliminate product inhibition150,151. Adrenochrome concentration was determined by reverse phase HPLC using absorbance at 480 nm.

COMT was found to mediate the oxidation of epinephrine to adrenochrome with SeAHO in a concentration-dependent manner (Figure 2.7). The initial rate of adrenochrome formation was determined for each reaction condition and found to have a linear relationship to the concentration of COMT (Figure 2.8). Epinephrine was not oxidized by COMT without SeAHO to a greater extent than epinephrine alone (Figure 2.9). By the final time point (42.25 hours), possible adrenochrome-catalyzed oxidation142,143 obscured the enzyme-mediated effects

observed (Figure 2.10). The enzyme-mediated oxidation appeared to be more than one-turnover,

-1 however the kcat was incredibly slow under the conditions tested at ~0.4 hr . Though COMT- catalyzed methylation is also quite slow (~24 min-1)133,152 with methyl-transfer being the rate limiting step153,154.

Figure 2.7 | Increase in Adrenochrome at Various COMT Concentrations COMT-mediated oxidation of epinephrine by SeAHO is concentration dependent.

Figure 2.8 | Rate of Adrenochrome Formation vs COMT Concentration Linear relationship between initial rate of adrenochrome formation and COMT concentration.

Figure 2.9 | Background Oxidation of Epinephrine ± COMT

Figure 2.10 | Adrenochrome-Catalyzed Oxidation Enzyme-mediated reaction obscured by adrenochrome-catalyzed oxidation at later time points. Adrenochrome concentration was determined by reverse phase HPLC.

COMT-mediated oxidation of epinephrine proceeds without the divalent cation, Mg2+.

As mentioned previously, COMT-catalyzed methylation, using AdoMet, requires Mg2+ ions for activity130-133. Mg2+ sits in the COMT active site and is key in coordinating and orientating the

catechol for alkylation133. For our purposes, Mg2+ was omitted since it has also been commonly used as a catalyst to study the non-enzymatic oxidation of catechols147,155,156.

There are several possible mechanisms for the reaction between SeAHO and epinephrine

(Figure 2.4). A proposed first step in COMT methylation is deprotonation of one of the catechol hydroxyls by a lysine residue (Figure 2.11), though site-specific mutagenesis does not destroy enzyme activity126,133,157,158. This deprotonation is a possible first step for epinephrine oxidation, followed by formation of a peroxide between the selenoxide and the deprotonated hydroxyl group, loss of water, and base-catalyzed cyclization of the orthoquinone (Figure 2.4)143,146,148,149.

It is also possible that the cyclization is caused by another equivalent of SeAHO and loss of water.

Figure 2.11 | COMT Active Site Crystal structure of COMT with AdoMet, Mg2+ and dinitrocatechol bound (PDB = 3BWM)159. AdoMet (green) is coordinated in the COMT active site through hydrogen bonding between Glu90 and the ribose hydroxyl groups, as well as through interactions with the adenine and amino acid moieties. The catechol (blue) is coordinated to the Mg2+ ion (pink and grey) that is bound by Asn170 and Asp141 in a pocket defined by Leu198. Lys144 is thought to act as a catalytic base used for catechol deprotonation.

As described in Chapter 1, SeAHO is reduced to SeAH in the presence of COMT, though

oxidized COMT was not observed by bottom-up mass spectrometry (Section 1.5.b). COMT

contains the known oxidoreductase motif CXXS160 as well as the CXXC motif, which is commonly found in proteins with redox reactivity conferred by reversible disulfide bond 55

formation161-164, though mass spectrometric studies of COMT have not revealed any intramolecular disulfides165. It is possible that COMT is mediates oxidation of epinephrine by

SeAHO not through active site binding and orientation as with transmethylation, but through one of these potentially redox-active motifs.

2.4 Oxidation of Other Catechols

In addition to epinephrine, it was of interest to test whether other COMT substrates were amenable to COMT-mediated oxidation using SeAHO. Catechol and dopamine, known COMT substrates130, were tested along with nonsubstrate tolcapone, an FDA approved COMT inhibitor136,138 (Figure 2.12).

Reactions were set up as described in Section 2.6.c containing 1 mM catechol, dopamine, or epinephrine, 1 mM SeAHO, and 0 or 25 µM COMT in 50 mM Tris, 0.5 M NaCl at pH 7.9.

Methylation reactions using 1 mM AdoMet with 100 µM MgCl2 in the place of SeAHO were set up in parallel. Reactions were analyzed for oxidation or methylation of the substrates by reverse phase HPLC.

(a) (b) (c)

Figure 2.12 | Structures of Catechol, Dopamine, and Tolcapone (a) Catechol and (b) dopamine are known COMT substrates. (c) Tolcapone is an FDA approved COMT inhibitor.

Oxidation products were observed to a much greater extent in COMT containing reactions than in the control reactions (Figure 2.13). Background oxidation was again seen in the epinephrine reactions as well as the catechol reactions. No background oxidation was seen for dopamine in the observed time frame.

(a) (b) (c)

Figure 2.13 | Oxidation of Dopamine, Catechol, and Epinephrine Oxidation peaks of COMT substrates—(a) dopamine, (b) catechol, and (c) epinephrine—reacted with SeAHO and COMT (red) or without COMT (green) and monitored by reverse phase HPLC chromatograms at 480 nm. Reactions at T0 are in blue.

(a)

(b)

(c)

(d)

(e)

(f)

Figure 2.14 | COMT-Catalyzed Methylation Reactions Reverse phase HPLC chromatograms monitored at 280 nm. COMT catalyzes the methylation of (a) catechol (~5.9 minutes) to (b) guaiacol (~15.6 minutes); (c) dopamine (~3.2 minutes) to (d) O-methyldopamine (~6.9 minutes); and (e) epinephrine (~2.6 mintues) to (f) metanephrine (~3.5 minutes). AdoMet (~2.6 minutes) and AdoHcy (~5.4 minutes) are also observed.

Tolcapone was also tested as a substrate for COMT catalyzed oxidation. Reactions were

set up as described in Section 2.6.c containing 1 mM tolcapone, 1 mM SeAHO, and 0 or 25 µM

COMT. Reactions were analyzed for oxidation by reverse phase HPLC. As expected, no oxidation products of tolcapone were observed (Figure 2.15).

(a)

(b)

Figure 2.15 | Tolcapone Oxidation Reaction Reverse phase HPLC trace at 325 nm of (a) tolcapone incubated with SeAHO and COMT overnight shows no oxidation products when compared to (b) tolcapone standard.

2.5 Conclusions

Here, we have described using SeAHO as a substrate for COMT as a way of modifying

the activity of the enzyme. Instead of a methyl transfer reaction, COMT mediated oxidation of its

substrates. The methyltransferase class is ripe for exploitation in the biocatalyst field. Their diverse but specific substrate pool encompasses small molecules, proteins, and nucleic acids.

Utilizing these enzymes in new ways can unlock a wider range of biocatalytic reactions.

2.6 Experimental Details

Reagents of ACS grade or better were obtained from either Sigma or Fisher unless otherwise noted. High performance liquid chromatography (HPLC) was done on an Agilent 1100

HPLC and with a diode array detector and analyzed using ChemStation (version B.03.02).

2.6.a Preparation of Enzymes

Adenosylhomocysteine nucleosidase (MTAN, EC 3.2.2.9) and the soluble form of human catechol-O-methyltransferase (COMT, EC 2.1.1.6) with an N-terminal histidine tag was purified

from a transfected E. coli cell line previously reported140,150,151. Cell stock was grown in LB

Broth (Fisher PN: BP1426-2) until the logarithmic growth phase was reached as monitored by

A600nm. Then, vector transcription was induced by adding isopropyl β-D-1-thiogalactopyranoside

(IPTG) (Fisher PN: BP1755-10) to 0.4 mM. After 8 to 16 hours of induction, the cells were

pelleted and washed four times in 50 mM Tris, 500 mM NaCl pH 7.9. After the final

resuspension, the cells were lysed by sonication then centrifuged under 8000 xg for 45 minutes at

4 °C to remove cell debris.

COMT was then purified from the supernatant using an immobilized-metal affinity-

chromatography (IMAC) column (GE PN:17-0920) using their method. Briefly, the column was

rinsed with deionized water and charged with a 0.1 M NiSO4 solution, then rinsed again with deionized water. The column was rinsed with a binding buffer (20 mM sodium phosphate, 0.5 M sodium chloride and 20 mM imidazole pH 7.4), then the supernatant was loaded. The column was again rinsed with binding buffer, followed by wash buffer (20 mM sodium phosphate, 0.5 M sodium chloride and 100 mM imidazole pH 7.4). Fractions were eluted with elution buffer (20 mM sodium phosphate, 0.5 M sodium chloride and 500 mM imidazole pH 7.4).

Fractions were then assessed by sodium dodecyl sulfate polyacrylamide gel

electrophoresis (SDS-PAGE). Selected fractions were then pooled and buffer exchanged into 50 mM Tris, 0.5 M NaCl pH 8 with at least four rounds of centrifugal concentration using a 10 kDa

MWCO ultrafiltration membrane with dilution in the final buffer. After the final concentration

step, glycerol was added to 10% (v/v) as a cryoprotectant. The concentration of COMT was

determined by absorbance at 280 nm using an extinction coefficient of 24,000 M-1cm-1. COMT was then stored frozen at -80 °C prior to use.

2.6.b Preparation of SeAHO

SeAHO was prepared from Se-adenosylselenohomocysteine (SeAH). SeAH was dissolved in 50 mM K3PO4 pH 12, treated with 5 molar excess H2O2, and reacted at room temperature for 10 minutes. Excess peroxide was removed by adding a small amount of catalase

(1 µL of 10 mg/mL in 50 mM Tris pH 8) and left until the solution stopped bubbling. Catalase was then removed by ultrafiltration through a 10 kDa MWCO membrane. The SeAHO concentration was then determined by absorbance at 260 nm using an extinction coefficient of

15,400 cm-1M-1 and conversion was determined by reverse phase HPLC as described in Section

1.7.d. SeAHO prepared in this manner was stored frozen at -80 °C prior to use.

2.6.c Enzyme-Mediated Reactions

Unless otherwise stated, reactions were run in 50 mM Tris, 0.5 M NaCl at pH 7.9. Upon combining the reaction components, reactions were vortexed briefly and incubated at 4 °C to minimize the background reaction. Aliquots were then removed at the noted time points and proteins were precipitated with the addition of trichloroacetic acid (TCA) to 10% (v/v). The reactions were centrifuged at 10,000 xg for 10 minutes at 4 °C and the supernatant was frozen at

-80 °C. Supernatants were analyzed for product formation by reverse phase HPLC using the parameters described in Section 1.7.d, with samples being thawed immediately prior to injection

Chapter 3: S-Adenosylvinthionine Purification and Analysis

single turnover

Adapted with permission from:

W. Qu, K.C. Catcott, K. Zhang, S. Liu, J.J. Guo, J. Ma, M. Pablo, J. Glick, Y. Xiu, N. Kenton, X. Ma, R.I. Duclos, Z.S. Zhou. “Capturing Unknown Substrates via in situ Formation of Tightly Bound Bisubstrate Adducts: S-Adenosyl-vinthionine as a Functional Probe for AdoMet-Dependent Methyltransferases.” J. Am. Chem. Soc. 138(9), 2877-2880, (2016). http://dx.doi.org/10.1021/jacs.5b05950

W. Qu designed performed experiments, analyzed data, and wrote the manuscript. K.C.

Catcott completed binding and degradation studies, prepared material for NMR and HRMS analysis, and revised the manuscript. K. Zhang performed the kinetics study. S. Liu performed binding studies and reviewed the manuscript. J.J. Guo acquired and analyzed NMR spectra. J.

Ma performed binding studies. M. Pablo purified AdoVin. J. Glick analyzed mass spectra. Y.

Xiu and N. Kenton synthesized and characterized vinthionine. X. Ma acquired NMR spectra.

R.I. Duclos synthesized and characterized 2-amino-5-mercaptobenzoic acid. Z.S. Zhou was principal investigator for this project and oversaw manuscript preparation.

3 S-Adenosylvinthionine Purification and Analysis

The work in the chapter builds on the dissertation of Dr. Wanlu Qu, parts of which were

published in the Journal of the American Chemical Society as cited above and include Figures

3.2, 3.2, 3.3, 3.4, 3.7, and 3.11.

Dr. Qu’s work revolved around utilizing an AdoMet analogue, S-adenosylvinthionine

(AdoVin), to capture methyltransferase substrates through the formation of bisubstrate adducts

(Figure 3.1).

Figure 3.1 | Synthesis and Reactivity of AdoVin Formation of AdoVin from vinthionine and ATP catalyzed by MAT. TPMT- catalyzed in situ formation of bisubstrate adduct between AdoVin and thiol substrates.

Conceptually, the interaction between an enzyme and its substrates or products is transient. Bisubstrate adducts, however, should bind significantly more tightly with the enzyme than either substrate alone due to the synergistic binding interactions, thereby resulting in a more persistent complex. This venerable concept has been explored—albeit in a few limited cases—to identify unknown enzymes and substrates, i.e., the formation of kinase-substrate complex via

ATP-based cross-linker166,167. Weinhold, Rajski, Thompson and others developed bisubstrate-

adduct inhibitors for DNA and protein arginine methyltransferases (PRMTs) via AdoMet

analogues with 5’-aziridinyl adenylates71,72,168-171, but to our knowledge, neither the tight binding

between the adducts and methyltransferases nor its application in identifying of unknown substrates was discussed.

The approach utilized with S-adenosylvinthionine, a probe in which a vinyl sulfonium replaces the methyl sulfonium in AdoMet, is to form a bisubstrate adduct via an addition reaction to the vinyl sulfonium (a Michael-type acceptor). AdoVin and the nucleophilic substrate form a covalent tight-binding adduct.

To examine the utility of AdoVin, thiopurine methyltransferase (TPMT, EC 2.1.1.67) with a broad specificity toward aromatic thiols was used60,172. Reduced Ellman’s reagent (TNB),

o-bromophenol, and m-bromophenol—known substrates of TPMT—all reacted with AdoVin and formed stable adducts, as confirmed by HPLC-UV-Vis and mass spectrometric analysis.

Conversely, non-substrates toward AdoMet (e.g., substituted phenols) did not react with AdoVin.

Next, tight binding of the resulting bisubstrate-adduct to the methyltransferase was

investigated. Free ligands and enzyme-adduct complex were separated via either ultrafiltration or

immobilized metal ion affinity chromatography (the recombinant TPMT contained a hexa- histidine-tag). Under both conditions, the AdoVin adducts were observed only in the enzyme

complex (Figure 3.2), indicating markedly tight binding between AdoVin adducts and the enzyme.

Figure 3.2 | TPMT-Bound and Free Ligand Fractions Reverse phase HPLC chromatograms at 325 nm of the TPMT complex and free ligand fraction separated via immobilized metal ion affinity chromatography (IMAC). The AdoVin-TNB adduct was observed in the TPMT fraction, but not in free ligand fraction.

Adduct formation is both time dependent (first-order kinetics) and enzyme concentration

dependent (Figure 3.3), consistent with the proposed mechanism that involves a rapid initial

binding of the thiol substrate and AdoVin with TPMT and the subsequent addition reaction. The

-1 reaction is first-order for adduct formation with AdoVin (kapp = 0.33 ± 0.12 min ), while the kcat of transmethylation with AdoMet (13.6 ± 0.4 min-1)60. Only a single turnover formation of the bisubstrate-adduct was observed, again consistent with the markedly enhanced binding affinity of the bisubstrate-adduct compared to the individual components.

Figure 3.3 | Change in TNB Concentration Over Time Changes in substrate concentration and TNB absorbance at 411 nm from the formation of adduct catalyzed by TPMT at various concentrations.

AdoVin was incubated with the crude cell lysate of E. coli that expressed recombinant

human TPMT, but no adduct was detected. As a positive control, TNB was added to the crude

cell lysate, and the corresponding adduct was detected, and again, only in the enzyme complex indicating tight binding under physiological conditions as well (Figure 3.4).

Unexpectedly, in E. coli lysates, aside from the AdoVin-TNB adduct, another adduct was

detected in the TPMT complex, but only when TNB was added to the cell lysates, suggesting this

unknown peak was derived from TNB (Figure 3.4). Based on the UV-Vis spectrum of the

unknown adduct, the mass change (-30 Da) and fragmentation pattern of isotopic labeled

adduct173-175, we postulated that the nitro-group in TNB and adduct was reduced to an amine,

which could be catalyzed by any of four nitroreductases existing in E. coli176. This was confirmed by the authentic amino thiol (2-amino-5-mercaptobenzoic acid, AMBA) and the corresponding adduct with AdoVin. It is worth noting that this amino thiol (AMBA) had not

been reported as a substrate of TPMT but was confirmed as a substrate toward AdoMet.

Altogether, this serendipitous finding underscores the utility of our approach in directly identifying enzyme substrates, even unknowns.

Figure 3.4 | HPLC Chromatograms of Ex Vivo Reactions HPLC chromatograms (260 nm) of (a) ex vivo reaction of crude cell lysate and (b) the captured AdoVin-TNB and AdoVin-AMBA adducts from isolated TPMT complex, illustrating affinity enrichment.

3.1 Introduction

With AdoVin’s function as an S-methyltransferase probe documented, it was of great interest to further characterize the probe and understand its functionality. A purification method was developed for AdoVin, which was then used to isolate material for characterization by high- resolution mass spectrometry and NMR. Additionally, its major degradation products were determined by LC-MS. Binding and competition studies were used to detail the strength of the

interaction between AdoVin, AdoVin-adducts, and TPMT. Finally, AdoVin was used to capture

exogenous TNB from mammalian cell lysates.

3.2 Purification of AdoVin

Since AdoVin was prepared enzymatically (details in Section 3.8.a), it was of interest to

develop a purification method that could separate AdoVin from excess reagents. Purification of

AdoMet through strong cation exchange (SCX) chromatography has been reported177,178. Thus,

AdoVin was prepared (Section 3.8.a) and proteins were removed by acid precipitation. The mixture was then separated by SCX-HPLC as detailed in Section 3.8.c using buffer system 1

(Table 3.1). The sulfonium containing AdoVin was well resolved from other compounds present in the sample and had a similar retention time as the AdoMet standard (Figure 3.5).

Though the initial separation method was acceptable, it was of interest to determine what impact the mobile phase had on separation. AdoVin was prepared as before and separated by

SCX-HPLC (Section 3.8.c) using the buffer systems outlined in Table 3.1. Though retention times for AdoVin changed slightly with the various buffer systems, each one tested afforded sufficient separation to be used for purification (Figure 3.6).

(a) AdoMet (Sigma A2408)

AdoMet

(b) AdoVin synthesis reaction

AdoVin

Figure 3.5 | SCX-HPLC Chromatogram of AdoMet and AdoVin SCX-HPLC chromatogram at 260 nm of (a) AdoMet purchased from Sigma (PN: A2408) and (b)AdoVin which was prepared enzymatically. AdoVin is sufficiently resolved for purification by this method.

Table 3.1 | SCX-HPLC Buffer Systems Mobile Phase A Mobile Phase B 10 mM KH PO , pH 3 10 mM KH PO + 0.5 M NaCl, pH 3 1 2 4 2 4 + 20% Acetonitrile + 20% Acetonitrile 2 10 mM KH2PO4, pH 3 10 mM KH2PO4 + 0.5 M NaCl, pH 3 3 10 mM Ammonium Acetate, pH 3 0.5 M Ammonium Acetate, pH 3

(a)

(b)

(c)

Figure 3.6 | Separation of AdoVin by SCX in Various Mobile Phases SCX-HPLC chromatograms of AdoVin monitored at 260 nm using (a) 10 mM

KH2PO4, pH 3 + 20% CH3CN / 10 mM KH2PO4, 0.5 M NaCl, pH 3 + 20% CH3CN; (b) 10 mM KH2PO4, pH 3 / 10 mM KH2PO4, 0.5 M NaCl, pH 3; or (c) 10 mM CH3CO2NH4, pH 3 / 0.5 M CH3CO2NH4, pH 3

3.3 Characterization

Though the functionality of AdoVin was thoroughly documented, it was of interest to better characterize the compound. Thus experiments were planned to further document the structure and stability of AdoVin, as well as the interaction between any AdoVin-adducts and their enzymes.

3.3.a High Resolution Mass Spectrometry

Though AdoVin had been observed by LC-MS (Figure 3.7), analysis by high resolution mass spectrometry (HRMS) was desired. AdoVin was prepared enzymatically from ATP and vinthionine as described in Section 3.8.a. After removal of MAT by acid precipitation, a high- resolution mass spectrum was acquired for the protein depleted AdoVin reaction mixture as described in Section 3.8.d.

Figure 3.7 | Mass Spectrum of AdoVin AdoVin (M+; expected m/z 411.10, observed m/z 411.13; mass difference 0.03 Da)

The HRMS spectrum of AdoVin (Figure 3.8) showed the monoisotopic peak (m/z

411.1444, M+) matched the expected peak (m/z 411.1445) very well (observed error -0.24 ppm).

The fine isotopic structure was also observed in the +1 Da peak.

Intens. No FA AdoVin 100uM_000001.d: +MS x107 1+ (a) 411.1443

2.0

1.5 AdoVin Observed Mass

1.0

0.5 1+ 412.1477

411.8739 411.9829 0.0 x107 NoFAAdoVin₁₀₀uM₀00001.d:C₁₆H₂₃N₆O₅S+, M, 411.14452 1+ (b) 411.1445

2.0 AdoVin Predicted Mass

1.5

1.0

0.5 1+ 412.1479

0.0 411.0 411.2 411.4 411.6 411.8 412.0 412.2 m/z

Intens. No FA AdoVin 100uM_000001.d: +MS Intens . 1+ No FA AdoVin 100uM_000001.d: +MS 7 6 x10 1+ x10 412.1477 411.1444 4 (c) 2.0

3 1.5 AdoVin Observed Mass

2 1.0

1 0.5

412.1417

0.0 0 x107 NoFAAdoVin₁₀₀uM₀00001.d:C₁₆H₂₃N₆O₅S+, M, 411.14452 x106 NoFAAdoVin₁₀₀uM₀00001.d:C₁₆H₂₃N₆O₅S+, M, 411.14452 1+ 1+ 411.1445 412.1479

(d) 2.0

1.5

AdoVin Predicted2 Mass 1.0

1 0.5 1+ 412.1422

0.0 0 411.138 411.140 411.142 411.144 411.146 411.148 411.150 411.152 411.154 m/z 412.135 412.140 412.145 412.150 412.155 412.160 m/z

Figure 3.8 | High Resolution Mass Spectrum of AdoVin (a,c) High-resolution mass spectrum of AdoVin; monoisotopic m/z 411.1444 (M+ of AdoVin expected m/z 411.1445; observed error -0.24 ppm). Spectrum is an average of 24 scans. (b,d) Predicted mass spectrum of AdoVin. (c,d, left) Zoom-in of monoisotopic peak. (c,d, right) Zoom-in of the +1 Da peak that shows fine isotopic structure.

3.3.b NMR

In addition to HRMS, we acquired an NMR spectrum of AdoVin. AdoVin was prepared as described in Section 3.8.a, the purified with SCX-HPLC as described in Section 3.8.c. Purified material was then frozen at -80 °C and lyophilized. The sample was reconstituted in D2O and the

1H- NMR spectrum was acquired as detailed in section 3.8.e.

The spectrum of AdoVin was compared to that of commercially available AdoMet

(Sigma PN: A2408) (Figure 3.9) and agrees with published data88,179. Water suppression in the

AdoVin sample was achieved through excitation sculpting180. Additional material was prepared with four cycles of lyophilization and reconstitution in D2O to remove residual H2O from the sample. The 1H- NMR spectrum of this sample had improved resolution around 4.8 ppm (Figure

3.10).

AdoVin AdoMet

Figure 3.9 | 1H NMR of Purified AdoVin and Commercial AdoMet 1 (Top) H-NMR spectrum of AdoVin in H2O/D2O (95:5). Region of water suppression is marked with the bracket. (Bottom) 1H-NMR spectrum of AdoMet- toluenesulfonate salt (Sigma PN: A2408). The peaks due to the toluenesulfonate are at 7.3 and 7.6 ppm.

Figure 3.10 | 1H NMR of Purified AdoVin

3.4 Degradation

AdoVin was found to have a half-life similar to AdoMet (Figure 3.11, ~15 hours)181. It was of interest to understand the key degradation pathways of AdoVin, which were expected to be similar to AdoMet181—intramolecular nucleophilic attack to form homoserine lactone and 5’-

(vinylthio)adenosine and hydrolysis to form S-pentosylvinthionine and adenine (Figure 3.12).

Figure 3.11 | Stability of AdoVin AdoVin with (■) 0 mM, (▲) 1 mM, or (●) 10 mM DTT. AdoVin was synthesized enzymatically and the protein was removed by ultracentrifugation with 30 kDa MWCO filter. The solutions containing 50 μM AdoVin with and without DTT (1 mM or 10 mM) were incubated at 37 °C. Aliquots of the reactions were analyzed by SCX- HPLC monitored at 260 nm.

AdoVin was prepared enzymatically from ATP and vinthionine as described in Section

3.8.a. After removal of MAT by acid precipitation, samples were incubated at 37 °C for 0, 20, or

144 hours then stored at -80 °C prior to analysis by LC-MS as described in section 3.8.f.

Figure 3.12 | AdoVin Degradation Pathways AdoVin undergoes (top) intramolecular nucleophilic attack to form homoserine lactone and 5’-(vinylthio)adenosine or (bottom) is hydrolyzed to form S- pentosylvinthionine and adenine.

Under the conditions tested, AdoVin was partially degraded by the 20 hour time point.

From this sample, AdoVin (m/z 411), homoserine lactone (m/z 102), 5’-(vinylthio)adenosine (m/z

310) and adenine (m/z 136) were all readily observed (Figure 3.13). This suggests that AdoVin degrades in a manner similar to AdoMet. For AdoMet at pH 7.5, the relative rates of intramolecular nucleophilic attack and hydrolysis are similar with hydrolysis being slightly slower181. Given the similarity between the half-life and degradation pathways of AdoVin and

AdoMet, it is expected that this relationship would also be true for AdoVin.

(a)

(b)

(c)

(d)

Figure 3.13 | XIC and Mass Spectra of AdoVin Degradation (a) AdoVin (m/z 411), (b) homoserine lactone (m/z 102), (c) 5’- (vinylthio)adenosine (m/z 310) and (d) adenine (m/z 136) were all present after AdoVin was incubated at 37 °C for 20 hours.

3.5 Binding

While it was shown that the AdoVin‐TNB adduct does not dissociate from TPMT after extensive washing through ultrafiltration membranes nor after His‐Tag purification (Figure 3.2), questions remained regarding the manner and strength of the interaction between the adduct and the enzyme. How does the interaction between AdoVin and TPMT compare to that of the native substrate, AdoMet? Can we quantify the binding interaction between the adduct and the enzyme?

3.5.a AdoMet and AdoVin TPMT-Binding Competition

In order to compare the binding interactions of TPMT with AdoVin and TPMT with

AdoMet, a competition experiment was run as described in Section 3.8.g. The unbound small molecule fraction and the TPMT-bound small molecule fraction were analyzed by SCX-HPLC as described in Section 3.8.c (Figure 3.14).

AdoMet binds TPMT much more strongly than AdoVin and is enriched in the protein fraction by a factor of 300. While AdoMet binds TPMT much more strongly, the reaction with

AdoVin is one turnover. In ex vivo experiments, where endogenous AdoMet is expected to be in micromolar concentrations, adduct formation with AdoVin still proceeds. Thus, under experimental conditions, AdoVin adequately binds TPMT for adduct formation, even when other substrates are present

(a) Free Ligand Fraction

AdoMet AdoVin (b) TPMT Containing Fraction

Figure 3.14 | TPMT Binding of AdoMet vs AdoVin SCX-HPLC chromatogram at 260 nm of (a) free ligand and (b) TPMT-bound fractions from a reaction containing TPMT with AdoMet and AdoVin in molar excess.

3.5.b AdoVin-TNB Adduct Binding with TPMT

The next aim was to quantify the strength of the interaction between the AdoVin-TNB adduct and TPMT. The first experiment looked at adduct dissociation from TPMT at various concentration of guanidinium. From those results, a pseudo‐Kd was extrapolated. The

TPMT•AdoVin-TNB complex was prepared without MTAN as described in Section 3.8.b.

Several reactions were set up containing 0, 1.8, 2.4, 3.0, 3.6, or 5.4 M guanidinium hydrochloride and incubated at 37 °C for 1 hour. Small molecules were then separated from the protein fraction using ultrafiltration with a 10 kDa MWCO filter and analyzed by SCX-HPLC as described in Section 3.8.c (Figure 3.15).

0 M Gdn

5.4 M Gdn

1.8 M Gdn

2.4 M Gdn

3.0 M Gdn

3.6 M Gdn

Figure 3.15 | AdoVin-TNB Adduct Dissociation with Guanidinium SCX-HPLC chromatograms at 260 nm of AdoVin-TNB adduct dissociation from TPMT at various guanidinium concentrations

To calculate a pseudo-Kd, the AUC of the adduct peak in the 5.4 M guanidinium sample was defined as 100% adduct dissociation and the 0 M guanidinium sample was defined as 0%.

Then, for each concentration tested, % free adduct was determined. Given this and knowing the starting TPMT•Adduct concentration, the concentrations of the free adduct, apo-TPMT, and

TPMT•Adduct could be determined for each sample. From here, a pseudo-Kd was be calculated at each concentration of guanidinium using Equation 1182,183.

Equation 1 | Dissociation Constant Ligandx [Enzyme] K = d [Enzyme•Ligand]

The pseudo-Kd values were then plotted on a log scale vs. the guanidinium concentrations

used (Figure 3.16). By extrapolating from this dataset, a pseudo-Kd at 0 M guanidinium was determined to be 0.55 µM. The actual Kd was expected to be considerably smaller than this.

Figure 3.16 | Guanidinium Induced Dissociation of AdoVin-Adduct from TPMT

Pseudo-Kd’s were calculated for TPMT•Adduct complexes at various guanidinium concentrations and used to extrapolate a pseudo-Kd for the complex without guanidinium.

We realized the limitation of this guanidinium method and designed a binding

competition experiment to obtain a more accurate value. TPMT•AdoVin-TNB complex was

prepared without MTAN as described in Section 3.8.b. The complex was then combined with

370-times molar excess of AdoMet or AdoHcy and incubated at 37 °C for 1 hour. Small

molecules were then separated from the protein fraction using ultrafiltration with a 10 kDa

MWCO filter. Proteins were precipitated from the retentate using 10% (v/v) TCA and pelleted

by centrifugation. The resulting small molecule fractions were analyzed by SCX-HPLC as

described in section 3.8.c (Figure 3.17).

Figure 3.17 | AdoVin-TNB Adduct Dissociation with AdoMet or AdoHcy SCX-LC chromatograms at 260 nm of AdoVin-TNB Adduct dissociation from TPMT at with high concentrations of AdoMet or AdoHcy

By comparing the AUCs for each ligand in the samples, we can determine the percentage

of each that was bound or free at equilibrium and how much adduct dissociated from TPMT

(Figure 3.18). Based on the starting concentrations of each species and the reported Kd of

AdoMet (7 µM) and AdoHcy (0.75 µM)184, Equation 1 can be used to determine the

concentration of apo-TPMT at equilibrium and thus the Kd of the AdoVin-TNB adduct. Based on competition with AdoMet, the Kd for the AdoVin-TNB adduct was 1 nM; on competition with

AdoHcy, 0.15 nM. Based on these results, we reported the Kd for the AdoVin-TNB adduct at <1 nM.

Figure 3.18 | AdoVin-TNB Adduct Competing with AdoMet and AdoHcy The TPMT•AdoVin-TNB complex was subjected to vast molar excess of AdoMet and AdoHcy to compete off the adduct. Even with high concentrations of competing ligands, more than 99% of the adduct remains bound to TPMT.

A bisubstrate adduct should bind significantly more tightly with the enzyme than either substrate alone due to the synergistic binding interactions; this is the premise of multisubstrate adduct inhibition185-190. Indeed, quantifying the adduct binding confirmed the robustness of the

interaction we had previously observed. Additionally, the crystal structure of TPMT with

AdoHcy bound191 reveals the active site to sit in a deep narrow pocket (Figure 3.19). It is possible that the synergistic binding effects are augmented by steric hindrance, trapping the bulky adduct within the protein’s folds. Ion mobility spectra that were collected as part of

Chapter 4, show that even as TPMT is unfolding in the gas phase, some adduct remains bound

(Figure 3.20). 85

(a) (b)

Figure 3.19 | Crystal Structure of TPMT with AdoHcy Bound (a) Ribbon diagram of TPMT (grey) with AdoHcy (red) bound (PDB = 2BZG)191. (b) Molecular surface of TPMT with AdoHcy (red) bound. Solvent accessible cavity is highlighted.

Figure 3.20 | Mobiligram of TPMT•Adduct Complex [TPMT•AdoVin-TNB] complex was analyzed by native mass spectrometry then the 10+ precursor (m/z 3084) was subjected to collision induced dissociation (CID) at 500 eV. The resulting species were analyzed by ion mobility.

3.6 Adduct Formation in Human Cell Lysate

Though it was determined that AdoVin could be enzymatically conjugated to exogenous

substrates, it was speculated as to whether there were endogenous substrates present in

mammalian cells that could also be trapped in this way. AdoVin was tested with TPMT for

endogenous and exogenous (TNB) substrate capture in HeLa cell lysate as detailed in Section

3.8.i. The samples were then run on LC-MS as described in Section 3.8.f.

In the sample that contained AdoVin and TPMT, but no TNB, the LC chromatogram revealed no new compound with absorbance at 260 nm. Additionally, there were no adduct peaks observed in the mass spectrum. The sample containing TNB, however, did form AdoVin-TNB adduct (m/z 610), which was seen in the mass spectrum (Figure 3.21). It is unsurprising that no adduct was found from the sample without TNB as TPMT has no reported endogenous substrate and may be important in detoxification184,192.

(a) HeLa ex vivo AdoVin-TNB adduct formation XIC of m/z 610.2

(b)

AdoVin-TNB adduct

Figure 3.21 | XIC and MS of AdoVin-TNB Adduct in HeLa Lysates Extracted ion chromatogram (a) of m/z 610.2 (AdoVin-TNB adduct) in the small molecule fraction of the TPMT catalyzed reaction in HeLa lysate. The mass spectrum (b) of the AdoVin-TNB adduct from this reaction.

3.7 Conclusions

AdoVin has been shown as an effective substrate of TPMT, conferring putative ligase activity on the enzyme. The AdoMet analogue can be similarly purified using SCX chromatography and shares similar degradation pathways. Though AdoVin itself does not bind to TPMT very tightly when compared to native substrates, the bisubstrate adduct is a very tight 88

binder, with a Kd in the low nanomolar range. This is likely due to the synergistic binding observed for bisubstrate adducts, as well as steric hindrance involved with the bulky adduct leaving the enzyme active site. Further applications utilizing AdoVin can be envisioned, as

AdoVin can be synthesized in vivo when vinthionine is supplemented175,193-196. Moreover, the formation of bisubstrate adducts may also have broad utility in detection of enzyme-substrate pairs utilizing whole protein detection methods.

3.8 Experimental Details

Reagents of ACS grade or better were obtained from either Sigma or Fisher unless otherwise noted. Recombinant histidine-tagged archaeon Methanococcus jannaschii L- methionine S-adenosyltransferase (S-adenosyl-methionine synthetase, MAT, EC 2.5.1.6)90,193-199, histidine-tagged human thiopurine-S-methyltransferase (TPMT, EC 2.1.1.67) and histidine- tagged E. coli 5'-methylthioadenosine/S-adenosyl-homocysteine nucleosidase (MTAN, EC

3.2.2.9) were purified as reported150,199,200. High performance liquid chromatography (HPLC) was done on an Agilent 1100 HPLC and with a diode array detector and analyzed using

ChemStation (version B.03.02).

3.8.a Chemoenzymatic Synthesis of AdoVin

S-Adenosylvinthionine (CAS 83768-89-2) was prepared chemoenzymatically by combining 50 mM potassium phosphate (pH 8.0), 5 mM KCl, 2.5 mM MgCl2, 1 mM ATP and

500 μM vinthionine with 50 μM MAT and incubated at 37 °C for 2 – 6 hours. The concentration

-1 -1 151 of AdoVin was determined using ε260 nm = 15,400 M cm based on the value for AdoMet .

3.8.b Formation of AdoVin-TNB Adduct and TPMT•Adduct Complex

AdoVin synthesis was prepared as described (Section 3.8.a) and included 480 μM TNB,

2 mM TCEP, 83 μM TPMT and 1.75 μM MTAN. The reaction incubated at 37 °C for 3 – 4

-1 -1 151 hours. The concentration of TNB was determined using ε411 nm = 13,600 M cm .

3.8.c SCX-HPLC

Strong cation exchange chromotography (SCX) was performed on an Agilent 1100

HPLC using ChemStation (version B.03.02). The column was a PolySULFOETHYL A (4.6 x

100 mm, 5 µm, 200 Å) from PolyLC. Unless otherwise noted, mobile phase A was 10 mM potassium phosphate at pH 3 and mobile phase B was 10 mM potassium phosphate, 0.5 M NaCl at pH 3. With a flow rate of 0.5 mL/min, the gradient was started at 100% A for the first 5 minutes; then ramped up to 100% B over the next 15 minutes and was held there for 5 minutes

The gradient then ramped down to 100% A over 1 minute and was held there for 4 more minutes.

3.8.d High Resolution Mass Spectrometry

High resolution mass spectrometry was generously acquired by Nicholas Schmitt on a

Bruker Solarix XR ESI-FT-ICR. A 1:1 water/acetonitrile solution contain 100 µM AdoVin was electrosprayed at a flow rate of 2 µL/min. Spectrum is an average of 24 scans. Acquisition of the spectrum was performed using ftms Control in the Compass 1.5 software and the expected mass spectrum was simulated using DataAnlysis (version 4.2).

3.8.e NMR

NMR spectra were generously recorded by Dr. Jason J. Guo and Xiaoyu Ma on a Bruker

Avance II 400 MHz NMR spectrometer and analyzed with Bruker Topspin (version 2.1). Water suppression was achieved through excitation sculpting180.

3.8.f LC-MS

LC-MS data were obtained using an H-Class Acquity UPLC system coupled to a Xevo

G2-S Q-ToF mass spectrometer (Waters Corp, Milford, MA). Liquid chromatography was performed on a BEH-C18, 2.1 mm x 150 mm column, with pore size of 1.7 µm (Waters Corp,

Milford, MA). Mobile phase A consisted of 0.1% formic acid (v/v) in HPLC grade water and mobile phase B consisted of 0.1% formic acid (v/v) in HPLC grade acetonitrile with a flow rate at 0.61 mL/min. A gradient was applied by starting at 2% mobile phase B increasing to 30% mobile phase B over 6.8 minutes, then increasing to 100% mobile phase B over 2.7 minutes, holding at 100% mobile phase B for 3.4 minutes, and finally decreasing to 2% mobile phase B over 1 minute and holding there for 6.1 minutes. After liquid chromatography, samples were introduced via an electrospray ion source in-line with the Xevo G2-S Q-ToF. Data were processed manually using Waters UNIFI 1.7.1 software.

3.8.g AdoVin and AdoMet Binding

TPMT was incubated with 75-times molar excess of AdoMet and AdoVin at 37 °C for 2 hours. The free ligand fraction was separated from the TPMT-containing fraction ultrafiltration with a 10 kDa MWCO filter. Protein was removed from the retentate by adding TCA to 10%

(v/v) followed by pelleting with centrifugation. The resulting supernatant along with the free ligand fraction were analyzed by SCX-HPLC.

3.8.h Molecular Models

Files were downloaded from the Protein Data Bank (http://www.rcsb.org/pdb) and

visualized using Yasara (version 15.11.18)201.

3.8.i HeLa Cell Lysate Screening

In 10% HeLa cell lysate (AbNova PN: L013V2), 1 mM vinthionine, 4 mM ATP, 2.5 mM

MgCl2, 5 mM KCl, 200 µM MAT, 0 or 0.4 mM TNB, 100 µM TPMT, and 2 mM TCEP were combined in 50 mM Tris at pH 8 and incubated at 37 °C for 6 hours. Proteins were then precipitated with 10% trichloroacetic acid (v/v) and pelleted by centrifugation at 10,000 xg for

10 minutes. The remaining small molecule fraction was then run on LC-MS as described in

Section 3.8.f.

Chapter 4: Identifying Enzyme-Substrate Pairs with AdoVin and Native MS

Adapted from:

K.C. Catcott, J. Yan, W. Qu, V.H. Wysocki, Z.S. Zhou. “Identifying Unknown Enzyme- Substrate Pairs from Cells with Native Mass Spectrometry.” Submitted.

K.C. Catcott performed sample preparations, analyzed mass spectra, and wrote the manuscript. J. Yan collected and analyzed the mass spectra, wrote the methods, and revised the manuscript. W. Qu performed sample preparation, wrote the methods, and revised the manuscript. V.H. Wysocki and Z.S. Zhou directed experimental design and oversaw manuscript preparation.

4 Identifying Enzyme-Substrate Pairs with AdoVin and Native MS

In the complex cellular milieu, understanding which enzyme catalyzes which reaction is paramount for deciphering biology and disease. There are still many enzymes whose exact functions or substrates remain unknown. Conversely, we know the nature of some biotransformations, but not precisely the responsible enzymes or proteoforms202 (specific genetic and splicing variations including any post-translational modifications). For example, more than

80 methyltransferases exist in humans127; and while histone substrates are well characterized, non-histone substrates remain poorly understood.

Figure 4.1 | Conventional Catalytic Cycle An enzyme binds two substrates to create a ternary enzyme•substrate•substrate complex. After catalysis, the enzyme•product•product complex dissociates, products are released, and the apo-enzyme continues on to substrate binding. The transient enzyme-substrate interactions often do not have sufficient affinity to be analyzed by mass spectrometry.

Methods for label-free identification of enzyme substrates are limited. Previously

described efforts have included performing high-throughput screening on a broad range of

substrates203,204 or performing substrate screening in silico through molecular docking programs205. Another reported tactic is to create enzyme mutants that lack catalytic activity, but

maintain substrate binding, then trap and identify substrates206. Additionally, drug discovery

efforts utilize ‘drug affinity responsive target stability’ (DARTS)207 and ‘stability of proteins

from rates of oxidation’ (SPROX)208 methodologies as label free techniques for confirming small molecule-protein interactions, though they usually start with the ligand and screen for a target.

For enzyme families with common substrates, such as S-adenosylmethionine (AdoMet,

SAM) for methyltransferases or adenosine triphosphate (ATP) for kinases, tagging through radiolabeled substrates or bioorthogonal analogues has been employed59,60,209. However, this

methodology fails to identify specific enzyme-substrate pairs, particularly from cellular contexts

as multiple substrates can be tagged by multiple enzymes with no clear path for deconvolution.

Identifying enzyme-substrate pairs is hampered by the fact that their interactions are

inherently transient. In a conventional catalytic cycle (Figure 4.1), the enzyme binds the

substrates, converts to products, and releases them. With few exceptions210-213, ternary

complexes are generally weakly bound, so cannot be directly observed by mass spectrometry.

However, non-catalytic complexes are observed by native mass spectrometry (MS). Enhancing

the enzyme-substrate affinity is one key to successful detection.

Towards this end, we envision a new enzyme-substrate pair detection platform, dubbed

IsoLAIT (Isotope-Labeled, Activity-based Identification and Tracking). IsoLAIT links solution-

phase activity to gas-phase detection by using a probe to capture, and native MS to identify,

enzyme-substrate pairs from a single sample run with minimal sample preparation (Figure 4.2).

for such complexes; th dissociates the enzyme from adduct. The probe’s isotopi enzyme•substrate-probe complex can be analyzed formation between the probe and the enzyme’s The IsoLAIT platform uses a substrate analogue as Figure 4.2 | IsoLAIT Framework and enzyme from the complex. components involved. Further analysis can be used to el ereby, unknowns can be identified without by native mass spectrometry. Tandem spectrometry

native substrate. The resulting tightly bound ucidate the structures and sequences of the substrates ucidate the structures and sequences of substrates a probe. An enzyme catalyzes bisubstrate adduct c flag is easily identified, and serves as a reporter a priori knowledge of the mass of the knowledge of the mass

4.1 Introduction

Here we demonstrate the IsoLAIT framework with AdoMet analogue, S-

adenosylvinthionine (AdoVin). As a probe, AdoVin creates a tightly bound [enzyme•substrate- probe] complex with thiopurine-S-methyltransferase (TPMT) and its substrates. Then, this persistent complex is identified by native mass spectrometry from the in vitro and ex vivo samples without separation.

4.2 IsoLAIT Keys

IsoLAIT has three keys: first, a probe that preserves the interaction between the enzyme and substrate for analysis; next, a detection method that retains enzyme-substrate interactions; and finally, a unique isotopic-flag that clearly identifies the complex, without a priori knowledge of its constituents.

First, the IsoLAIT probe forms an activity-based bisubstrate adduct, which has sufficient affinity to the enzyme such that the [enzyme•substrate-probe] complex remains intact throughout subsequent analysis. Additionally, high affinity and low turnover enriches the

[enzyme•substrate-probe] complex over the transient catalytic complex.

Second, the complex is analyzed by native mass spectrometry, which maintains protein complexation in the gas phase. Augmented with nanoelectrospray ionization, native mass spectrometry has emerged as a powerful way to characterize protein-ligand binding213-215. Upon dissociation, the enzyme and substrate-probe are separated and can be further interrogated, elucidating the identity and structure of each component. An added benefit, the complex is

detected in multiple charge states, providing redundancy in complex systems where peaks overlap and helping to deduce the enzyme mass.

Third, the unique isotopic-flag conferred by the probe is used as a telltale-sign for the

[enzyme•substrate-probe] complex. Overcoming the limitation of typical workflows that require

pre-defined masses, we are able to identify unknown substrates and enzyme proteoforms with

unforeseen modifications.

Figure 4.3 | TPMT-Catalyzed Reactions with AdoMet and AdoVin (a) Common methyl donor, S-adenosylmethionine (AdoMet, SAM), is used by thiopurine-S-methyltransferase (TPMT) to methylate aromatic thiols, such as 2- nitro-5-thiobenzoic acid (TNB) and 2-amino-5-mercaptobenzoic acid (AMBA), and produces S-adenosylhomocysteine (AdoHcy, SAH). (b) S- Adenosylvinthionine (AdoVin), an AdoMet analogue and methyltransferase probe, forms a covalent bond with TNB or AMBA under TPMT catalysis and results in a substrate-probe adduct that tightly binds to the enzyme.

Herein, the IsoLAIT framework is demonstrated with methyltransferases, a large family

of enzymes with diverse substrates. In transmethylation, methyl from common donor AdoMet is

enzymatically transferred to a nucleophilic substrate; the enzyme thiopurine-S-methyltransferase

(TPMT, EC 2.1.1.67) methylates thiophenols (Figure 4.3). IsoLAIT was demonstrated here with

TPMT and used AdoVin197,216 as a probe.

4.3 Native Mass Spectrometry of In Vitro Samples

We posited that TPMT would catalyze the formation of a substrate-probe adduct186

(Figure 4.3); indeed adduct formation between the substrate’s thiol and the vinyl sulfonium in

AdoVin was confirmed216. However, it was not clear whether such an [enzyme•substrate-probe] complex would bind sufficiently to survive native mass spectrometry. To investigate, we prepared in vitro samples as described in section 4.6.b (Figure 4.4).

Figure 4.4 | S-Adenosylvinthionine Probe Preparation S-Adenosylvinthionine (AdoVin) is synthesized enzymatically using S- adeonsylmethionine synthetase (MAT, EC 2.5.1.6). Vinthionine (1 mM) is combined with adenosine triphosphate (ATP) (2 mM) in the presence of KCl (10

mM) and MgCl2 (5 mM) in tris buffered saline (pH 7.9). MAT (25 µM) is added to start the reaction, which is then incubated at 37°C for 4 hours. ATP can be labeled with stable isotopes 13C and 15N at the highlighted positions to produce AdoVin with a +15 Da mass shift.

For the in vitro samples, the isotopically labeled probe was reacted with the TNB

substrate and the TPMT enzyme to form the [TPMT•TNB-AdoVin] complex. Whole cell

samples were prepared using unlabeled AdoVin probe, which, along with the TNB substrate, was

added to cell lysate from a TPMT-expressing E. coli strain to form the complex.

For the IsoLAIT platform, minimal sample preparation is required. The reaction mixtures

were first exchanged into a volatile buffer near physiological pH (e.g., ammonium acetate adjusted to pH 8.0 with ammonium hydroxide) to maintain native conformations and enhance

ionization, and then directly infused into the mass spectrometer as described in sections 4.6.d-

4.6.e (Figure 4.5).

100

Figure 4.5 | Native MS and Tandem MS of In Vivo Complex (a) In vitro [TPMT•TNB-AdoVin] complex and apo-TPMT were detected in multiple charge states using native mass spectrometry. In this simple system, the bound and apo- forms of TPMT were easily distinguished (inset). (b) Collision induced dissociation (CID) at a collision energy of 500 eV of the 10+ charged precursor ion of the [TPMT•TNB-AdoVin] complex (*) in (a). CID resulted in some apo-TPMT with the 10+ and 9+ charge states and the TNB-AdoVin adduct (highlighted and inset). Doublets observed in the TNB-AdoVin spectra are due to a mixture of natural and isotopically labeled AdoVin (+15 Da).

101

From the in vitro samples, the [enzyme•substrate-probe] complex and apo-enzyme were readily detected and also resolved from each other (Figure 4.5(a), inset). Tandem mass spectrometry (e.g., collision induced dissociation (CID) at a collision energy of 500 eV) resulted in the apo-enzyme and the substrate-probe adduct (e.g., TNB-AdoVin) (Figure 4.5(b)). The latter was easily spotted from its signature 15 Da doublet imparted by the isotopically labeled AdoVin probe (Figure 4.5(b), inset).

Further fragmentation of the TNB-AdoVin adduct was achieved by using higher collision energies for CID (1200 eV, Figure 4.6). The resulting fragments confirmed the identity of the

TNB substrate. Alternatively, quasi-MS3, with ion mobility substituting for the second analysis stage, provided another route for fragmentation (Figure 4.7).

102

Figure 4.6 | CID-MS of TNB-AdoVin Adduct CID tandem mass spectrum of precursor ion m/z 3084 (z=10) of the [TPMT•TNB-AdoVin] adduct at a collision energy of 1200 eV, which results in dissociation and fragmentation of the bisubstrate adduct (m/z 50 – 650). The TNB-AdoVin adduct is identified intact with the 15 Da stable isotope labeling present in the adenine group (m/z 610, z=1; m/z 625, z=1). Additional fragments are identified, some containing the adenine isotope flag (indicated by Δ).

103

Figure 4.7 | Quasi MS3 of Unlabeled TNB-AdoVin Adduct (a) Mass spectrum of in vitro [TPMT•TNB-AdoVin] complex (*) at m/z 3089 (z=10). (b) Tandem mass spectrum after CID at a collision energy of 500 eV of [TPMT•TNB-AdoVin] complex (precursor ion m/z 3089), resulting in apo-TPMT singly charged TNB-AdoVin adduct (highlighted) at m/z 610. (c) Quasi MS3 spectrum of CID at a collision energy of 20 eV of TNB-AdoVin adduct (precursor ion m/z 610, z=1), resulting in various adduct fragments.

104

4.4 Native Mass Spectrometry of Ex Vivo Samples

It was of interest to test the IsoLAIT platform on a more complex matrix. To investigate,

we prepared ex vivo (i.e., E. coli cell lysate) samples (Section 4.6.c, Figure 4.4).

As compared to in vitro results, it is not clear whether both adduct-bound and apo-forms

of the enzyme from the ex vivo sample were present (Figure 4.8(a), inset). This could be due to peak overlap, a common issue for the analysis of complex biological samples, as well as overall signal strength.

105

Figure 4.8 | Native MS and Tandem MS of Ex Vivo Complex (a) [TPMT•AMBA-AdoVin] complex was identified in multiple charge states from whole cells without purification. In this complex matrix, bound and apo- forms of TPMT were not resolved (inset). (b) CID at a collision energy of 500 eV of the 10+ charged precursor ion of the [TPMT•AMBA-AdoVin] complex (*) in (a). After collision, multiple charge states for the apo-enzyme were observed along with the AMBA-AdoVin adduct (highlighted and inset). In this case, isotopically labeled probe was not used, resulting in a single peak.

106

It is worth noting, the mass charge ratio for the ex vivo sample differs from that of the in vitro sample. Tandem mass spectrometry of the ex vivo samples revealed the substrate-probe adduct to be modified as seen in Chapter 3. The nitro group of the TNB substrate was reduced to an amine (2-amino-5-mercaptobenzoic acid or AMBA) in the cell lysate, likely due to endogenous nitroreductases176, and the corresponding amine-containing substrate-probe adduct

(i.e., AMBA-AdoVin) was formed (Figure 4.8(b)). Conventional workflows would likely miss

this species as the mass changes from in vitro to ex vivo samples; however, IsoLAIT is not bound

by predefined masses, but screens for mass patterns imparted by the probe and thus will identify

substrates no matter their modifications.

Equally interesting is the ability to identify the functional enzyme proteoform. For

tandem mass spectra, identifying the apo- and adduct-bound-enzyme peaks was straightforward.

Upon closer analysis of apo-TPMT, the observed and theoretical masses differed—the enzyme underwent methionine cleavage when expressed in E. coli (Figure 4.9). These results highlight

IsoLAIT’s utility in identifying an unknown, biologically-relevant, active proteoform.

107

Figure 4.9 | Mass Spectrum of Apo-TPMT CID at a collision energy of 1350 eV of precursor ion m/z 3407 (z=9, highlighted in blue) dissociates salt adducts and results in TPMT (m/z 3358, z=9) with a mass of 30,214 Da. The theoretical mass given the enzyme sequence is 30,343 Da, suggesting N-terminal methionine removal

Complex dissociation using surface induce dissociation (SID) instead of CID yielded similar results (Figure 4.10). Though redundant in the case of the monomeric TPMT enzyme, as interest in characterizing multimers or protein complexes increases215, SID’s utility is significant217,218.

108

Figure 4.10 | SID of [TPMT•Adduct] Complex (a) After initial separation of an in vitro sample, surface induced dissociation (SID) tandem mass spectrum of precursor ion m/z 3084 (z=10, from Fig. 2a) of the [TPMT•TNB-AdoVin] adduct (star) at a collision energy of 500 eV. After dissociation, apo- (open circles) and adduct-bound (closed circles) enzyme are observed in multiple charge states. The singly-charged TNB-AdoVin adduct (triangle) is observed with the 15 Da pairing due to stable isotope labeling (m/z 610 and 625). (b) After initial separation of an in vitro sample, SID tandem mass spectrum of precursor ion m/z 3084 (z=10, from Fig. 2a) of the [TPMT•TNB- AdoVin] adduct at a collision energy of 1000 eV, which results in increased dissociation and fragmentation of the bisubstrate adduct. The TNB-AdoVin adduct is identified intact with the 15 Da stable isotope labeling (m/z 610 and m/z 625). Additional fragments are identified, some containing the adenine isotope flag (indicated by Δ). (c) After initial separation of an ex vivo sample, SID tandem mass spectrum of precursor ion m/z 3080 (z=10, from Fig. 2c) of the [TPMT•AMBA-AdoVin] adduct (star) at a collision energy of 500 eV. After dissociation, apo- (open circles) and adduct-bound (closed circles) enzyme are observed in multiple charge states. The singly-charged AMBA-AdoVin adduct (triangle) is observed at m/z 580.

109

Throughout screening, we attempted to identify the enzyme-substrate complex without the use of a probe. Though TPMT was observed in complex with the common substrate

(AdoMet) and byproduct S-adenosylhomocysteine (AdoHcy, SAH) (Figure 4.11, Figure 4.12), we did not observe the TPMT enzyme bound to either TNB or AMBA substrates without the use of the AdoVin probe, again highlighting the transient nature of substrate-enzyme interactions.

Figure 4.11 | MS and Tandem MS of [TPMT•AdoMet] Complex (a) Mass spectrum of [TPMT•AdoMet] complex (highlighted) at m/z 3096 (z=10) formed by combining 4 µM TPMT + 40 µM AdoMet in vitro. (b) Collision induced dissocation (CID) tandem mass spectrum of [TPMT•AdoMet] complex from (a) (precursor ion m/z 3096) at a collision energy of 300 eV. Singly charged AdoMet (m/z 399) and TPMT in multiple charge states (z=9 and 10) were observed.

110

Figure 4.12 | MS and Tandem MS of [TPMT•AdoHcy] Complex (a) Mass spectrum of [TPMT•AdoHcy] complex (highlighted) at m/z 3061 (z=10). (b) Tandem mass spectrum after CID at a collision energy of 300 eV of [TPMT•AdoHcy] complex (precursor ion m/z 3061). TPMT in multiple charge states (z=9 and 10) and (c) singly charged AdoHcy (m/z 385) were observed.

Through the given examples, IsoLAIT demonstrates its robustness as a general platform, succeeding in contexts where conventional workflows may fail. Peak overlap in complex matrices is overcome through using the probe as a flag in the tandem mass spectra. Importantly, this method requires no prior knowledge of the specific enzyme or substrate for identification.

The unexpected alteration of substrate observed in the ex vivo samples perfectly illustrate both the challenges associated with biological systems and the utility of our methods. Had the ex vivo screening been based on the in vitro results, we would have completely missed the

111

[enzyme•substrate-probe] complex from ex vivo samples due to modifications to both the

enzyme and substrate.

4.5 Conclusion

Demonstrated here is a streamlined version of IsoLAIT. The framework could be

augmented by including other compatible separation techniques providing even greater detail.

Upstream sample purification, inline liquid chromatography methods, and downstream

separation, such as ion mobility, which we have successfully used, could further enhance the

method.

The IsoLAIT framework solves two key challenges in detecting enzyme-substrate pairs:

first, how to overcome the transient nature of enzyme-substrate interactions, and second, how to

identify components from a complex mixture with unexpected modifications. The former—

transient interaction—is overcome by the combination of a probe that perpetuates enzyme-

substrate interactions and native mass spectrometry that retains binding in the gas phase. The

latter—identification—is accomplished by using an activity-based probe and monitoring the

telltale isotopic pattern.

This appears to be the first reported example of mass spectrometric identification of enzyme-substrate complexes from the cellular milieu, but it need not be the last. In this case, we used a TPMT specific probe (i.e., AdoVin), but IsoLAIT probes can be tailored for other systems. More promiscuous probes could be used to screen for enzyme-substrate pairs more broadly. For example, aziridinoadenosines have been shown to work as activity-based probes for a range of methyltransferases and form stable, tight-binding bisubstrate adducts70,72. Previous studies using chemical tagging and shotgun proteomics may be amenable for IsoLAIT 112

adaptation219. IsoLAIT can be broadly applicable and similarly successful for other enzyme systems, particularly those catalyzing group transfer and with multiple substrates, such as glycosyltransferases and kinases.

4.6 Experimental Details

All chemicals with reagent purity or above were purchased from Sigma (St. Louis, MO) and Fisher (Pittsburgh, PA) unless otherwise noted. Immobilized metal ion affinity chromatography (IMAC) was performed on HisTrap HP columns (GE Healthcare Life Sciences

PN: 17-52447-01). Ultrafiltration was carried out using filters with 10 kDa MWCO (EMD

Millipore PN: UFC5010).

4.6.a Preparation of S-adenosylvinthionine (AdoVin)

S-Adenosylvinthionine (CAS 83768-89-2) was prepared enzymatically using S-adenosylmethionine synthetase (MAT, EC 2.5.1.6) using 5 mM of both labeled (+15 Da, Sigma-Aldrich

PN: 645702) and unlabeled ATP (Sigma-Aldrich PN: A2386) along with 1 mM vinthionine. The reaction proceeded in a 50 mM potassium phosphate buffer (pH 8.0), 5 mM KCl, 2.5 mM

MgCl2, and was initiated with 50 μM S-adenosylmethionine synthetase (MAT, EC 2.5.1.6) and

incubated at 37 °C. After 2-4 hours incubation, this mixture was used as the in situ probe.

AdoVin has a stability similar to S-adenosylmethionine (AdoMet, SAM), thus the samples were used immediately and without freezing.

4.6.b In Vitro Reactions

For the in vitro samples, His-tagged TPMT was grown in transfected E. coli and purified using an IMAC column151. Adduct formation was performed in 50 mM potassium phosphate (pH 113

8.0). The reaction solution contained the in situ AdoVin probe at 1 mM, 500 μM 2-nitro-5-

thiobenzoic acid (TNB), 2 mM tris(2-carboxyethyl)phosphine (TCEP), 3.5 μM S-adenosyl-

homocysteine nucleosidase (MTAN, EC 3.2.2.9) and 100 μM TPMT. The reaction was incubated at 37 °C for 3 hours.

4.6.c Ex Vivo Reactions

For ex vivo reactions, E. coli transformed with a TPMT-expressing plasmid were grown at 37 °C in LB broth for 12 hours, then washed three times with lysis buffer (50 mM potassium phosphate, 0.5 M NaCl, pH 8.0). Cells were lysed via sonication on ice, and the cell debris and unbroken cells were removed by centrifugation (8000 xg, 4 °C, and 60 minutes). To the supernatant, exogenous reagents were added: in situ AdoVin probe at 1 mM, 500 μM TNB, 2 mM TCEP, 3.5 μM MTAN. The reaction was allowed to react at 37 °C for 3 hours.

4.6.d Sample Preparation

Prior to analysis by MS, all samples were buffer exchanged into 20 mM ammonium acetate (adjusted to pH 8.0 with ammonium hydroxide) using at least ten cycles of concentration and dilution in a 10 kDa MWCO centrifuge, ultrafiltration concentrator. Samples were then frozen at -80 °C and thawed immediately prior to analysis. Note: in the case where freezing denatures or unfolds the proteins, samples should be stored at 4 °C or analyzed immediately.

Further sample dilution was done in 20 mM ammonium acetate (pH 8). The concentration of ex vivo TPMT•AMBA-AdoVin was estimated at 2 µM by SDS-PAGE. The concentration of the in vitro TPMT•TNB-AdoVin was 10 µM as determined by A280nm using the extinction coefficient

-1 -1 for TPMT (ε280nm = 39,420 cm M ).

114

4.6.e Native Mass Spectrometry

The nanoelectrospray experiments were performed on a Synapt G2S HDMS (Waters

Corp., Wilmslow, UK) with a customized surface-induced dissociation (SID) device installed before the ion mobility cell as previously described218. Each sample was filled into a glass capillary pulled using a Sutter Instruments P-97 micropipette puller (Novato, CA) and electrically connected to high voltage with a platinum wire. The nanoelectrospray source was at a voltage of 1.2-1.5 kV. The sampling cone voltage was set to 20 V and the source offset voltage was set to 20 V to avoid source activation of the complex. Other instrument conditions were 5 x

10-3 mbar for the source pressure, 2.0 mL/min gas flow rate to the trap cell, 120 mL/min gas ﬂow

to the helium cell and 60 mL/min gas flow to the ion mobility cell. The ion mobility wave

velocity was 200 m/s and the wave height was 16 V. The TOF analyzer pressure was 1.2 x 10-6 mbar.

4.6.f Tandem Mass Spectrometry

Tandem mass spectrometry experiments were performed via dissociation of the selected

ions with collision induced dissociation (CID) and SID. CID experiments were conducted with a

trap gas flow rate of 4.0 mL/min and SID was conducted with a trap gas flow rate of 2.0 mL/min.

The CID and SID (MS2) experiments were conducted in the trap travelling wave ion guide region before the ion mobility cell. The acceleration voltage in CID and SID was obtained as described previously217,218. The collision energy in eV was calculated via multiplying the acceleration voltage by the charge state of the precursor ion. The fragment ions generated from

MS2 experiment were separated in the ion mobility cell. The quasi MS3 experiment was conducted by selection in the quadrupole, activation in the trap CID cell and separation in the ion

mobility cell and activation in the transfer travelling wave ion guide. 115

Chapter 5: Conclusions and Future Directions

Homology model of MET7A (red) aligned with As (III) S-adenosylmethionine methyltransferase (blue, PDB: 4FS8)220 Class I methyltransferase motif128, GXGXG, highlighted (green and yellow)

116

5 Conclusions and Future Directions

Descriptions of the AdoMet analogues contained in the previous chapters have included some of their applications; SeAHO as a tool for biocatalysis and biosynthesis, and AdoVin as a probe for studying enzyme-substrate pairs. Many potential uses, however, have yet to be described and it is our hope that some of these may be explored in future projects.

5.1 Se-Adenosylselenohomocysteine Selenoxide

Chapter 2 described how the selenoxide analogue of AdoMet, Se-

adenosylselenohomocysteine selenoxide (SeAHO), could be used to alter the catalytic activity of

catechol-O-methyltransferase (COMT) resulting in putative oxidoreductase activity.

Methyltransferases constitute a large and diverse enzyme family127,129, and their use as biocatalysts is already an active area of research54,221. It is our hope that the scope of SeAHO application could be expanded and its usefulness increased

5.1.a O-Methyltransferases

Given the success seen with COMT, the possibility of utilizing SeAHO with other O-

methyltransferases is promising. There are many small molecule O-methyltransferases described

as methylation plays a key role in many biosynthetic routes54,221; applications here would be intriguing.

5.1.b S-Methyltransferases

Few S-methyltransferases have been described51,52. It is unlikely that SeAHO would be an effective substrate for enzymes targeting free thiols. The reduction potentials of organic

117

selenium and sulfur compounds are such that, in a system containing a mix of selenoxides and

thiols, formation of the selenide and disulfide will be favored69. Indeed, experiments that tested thiopurine-S-methyltransferase (TPMT) mediating substrate oxidations utilizing SeAHO found that the oxidation of TPMT substrate TNB by SeAHO proceed faster in the absence of the enzyme (Figure 5.1). It would be interesting to investigate enzymes targeting thioethers, such as thioether-S-methyltransferase222, however, as we found that SeAHO is not readily reduced by thioethers in vitro.

Figure 5.1 | TNB Oxidation with SeAHO TNB is more quickly oxidized by SeAHO in the absence of TPMT (red) than with TPMT present (blue).

5.1.c N-Methyltransferases

Many N-methyltransferases target lysines, arginines, and nucleosides127. Most interesting, perhaps, would be testing nucleoside N-methyltransferases for catalysis with SeAHO to achieve site specific oxidative damage of DNA and RNA.

118

5.1.d C-Methyltransferases

Using SeAHO in conjunction with a methyltransferase to target carbons is perhaps the most appealing application of this compound. The potential to provide C-H functionalization in a regio- and stereo-selective way could be a powerful tool to add to the synthetic chemistry

toolbox223,224. Over forty C-methyltransferases have been described and their substrates range

from fatty-acid chains225 to nucleotides226-230, amino acids231,232 to metal-chelating macrocycles233.

5.2 S-Adenosylvinthionine

Chapters 3 & 4 described the AdoMet analogue S-adenosylvinthionine (AdoVin) and its application in the native mass spectrometry platform, IsoLAIT. As demonstrated, AdoVin conferred ligase activities onto thiopurine-S-methyltransferase (TPMT). The thiol-reactive

AdoVin could be used to interrogate other S-methyltransferases, such as thiol-S- methyltransferase (TMT, EC 2.1.1.9)234.

Other AdoMet analogues have conferred ligase activities onto methyltransferases70-

72,168,235-237, it would be interesting to use these in the IsoLAIT platform to screen for

methyltransferase-substrate pairs. Beyond methyltransferases, we hope that the IsoLAIT

platform can be adapted to screen for enzyme-substrate pairs in different enzyme families and for

the development of additional enzyme probes.

5.2.a Thiol-S-Methyltransferase

Thiol-S-methyltransferase (TMT, EC 2.1.1.9)234, is an enzyme of unknown sequence with wide tissue distribution238 whose activity and substrate specificity has been noted for

119

decades234,239-242. Its activity is found to be primarily membrane associated and can be observed in red blood cell membrane fractions234,239,242. TMT methylates some aliphatic thiols and has been implicated in drug metabolism239,240,243.

It is expected that AdoVin would be an effective substrate for TMT, similarly to its compatibility with TPMT. Used in conjunction with the IsoLAIT framework, AdoVin could be a useful tool in identifying TMT from whole cell samples.

5.2.b MET7A

Given what is known about the activity and cellular location of TMT, some preliminary work has been done to determine potential candidates, with MET7A being the most promising. A search of the Red Blood Cell Collection244—a database of the red blood cell proteome— revealed MET7A as the only membrane associated protein with unverified activity which was identified as a putative methyltransferase245,246.

From this starting point, we completed homology modeling on the expected AdoMet- binding domain of MET7A using Yasara (described in Section 5.2.d). Four out of the top five hits were verified methyltransferases. When we compare the homology model of MET7A to the known methyltransferase, As (III) S-adenosylmethionine methyltransferase (AS3MT, EC

2.1.1.137 )220, the RMSD between the two is 0.616 Å. Perhaps more importantly, the GXGXG

motifs present in most Class I methyltransferases128, show very good alignment (Figure 5.2).

120

Figure 5.2 | MET7A Homology Model Aligned with As(III) Methyltransferase MET7A homology model (red) shows good alignment with the crystal structure of known methyltransferase As (III) methyltransferase (AS3MT, blue). The regions highlighted in yellow (MET7A) and green (AS3MT) show the GXGXG motif common to Class I methyltransferases.

It is our hope that continued experimentation on the MET7A protein will confirm it as an

AdoMet-dependent methyltransferase and determine its target substrates.

5.2.c Aziridines

IsoLAIT was shown here using AdoVin as an S-methyltransferase specific probe, but we would like to see the platform used with other systems. Aziridinoadenosines (Figure 5.3) are compatible with many methyltransferases—including N, C methyltransferases—and form stable, tight-binding bisubstrate adducts70-72. A highly active electrophile, aziridine derivatives would

121

likely be an excellent tool for broad screening of methyltransferase-substrate pairs in complex

mixtures.

Nucleophile Nucleophile

R NH2 R NH2 N N N N HO2C N HO2C N O N N MTase O N N NH2 NH2

HO OH HO OH

Aziridine AdoMet Analogue Bisubstrate Adduct Figure 5.3 | Aziridinoadenosines AdoMet analogues, aziridinoadenosines, are active with a range of methyltransferases and form stable, tight-binding adducts.

5.2.d Molecular Models

Homology modeling was graciously done by Lisa Ngu using Yasara. Files were downloaded from the Protein Data Bank (http://www.rcsb.org/pdb) and visualized using Yasara

(version 15.11.18)201. Alignment was completed using the built-in MUSTANG module247.

122

References

1 Bardeen, J. John Bardeen - Banquet Speech, <> (Date Accessed: 11 November 2016).

2 [IUPAC-IUB Commission on Biochemical Nomenclature (CBN). Nomenclature of multiple enzyme types. Recommendations 1971]. Hoppe Seylers Z Physiol Chem 353, 852-854, (1972).

3 Moss, G. P. Enzyme Nomenclature, (Date Accessed: 11 November 2016).

4 Neuberg, C. & Hirsch, J. Uber ein Kohlenstoffketten knupfendes ferment (Carboligase). Biochem. Z. 115, 282-310, (1921).

5 Schoemaker, H. E., Mink, D. & Wubbolts, M. G. Dispelling the myths--biocatalysis in industrial synthesis. Science 299, 1694-1697, (2003). http://dx.doi.org/10.1126/science.1079237

6 Rubin-Pitel, S. B. & Zhao, H. Recent advances in biocatalysis by directed enzyme evolution. Comb Chem High Throughput Screen 9, 247-257, (2006).

7 Zhao, H., Chockalingam, K. & Chen, Z. Directed evolution of enzymes and pathways for industrial biocatalysis. Curr Opin Biotechnol 13, 104-110, (2002).

8 Savile, C. K. et al. Biocatalytic asymmetric synthesis of chiral amines from ketones applied to sitagliptin manufacture. Science 329, 305-309, (2010). http://dx.doi.org/10.1126/science.1188934

9 Khoury, G. A. et al. Computational design of Candida boidinii xylose reductase for altered cofactor specificity. Protein Sci 18, 2125-2138, (2009). http://dx.doi.org/10.1002/pro.227

10 Nair, S. K. & van der Donk, W. A. Structure and mechanism of enzymes involved in biosynthesis and breakdown of the phosphonates fosfomycin, dehydrophos, and phosphinothricin. Arch Biochem Biophys 505, 13-21, (2011). http://dx.doi.org/10.1016/j.abb.2010.09.012

11 Antikainen, N. M. & Martin, S. F. Altering protein specificity: techniques and applications. Bioorganic & medicinal chemistry 13, 2701-2716, (2005). http://dx.doi.org/10.1016/j.bmc.2005.01.059

12 O'Brien, P. J. & Herschlag, D. Catalytic promiscuity and the evolution of new enzymatic activities. Chemistry & biology 6, R91-R105, (1999). http://dx.doi.org/10.1016/S1074- 5521(99)80033-7 123

13 Gupta, R. D. Recent advances in enzyme promiscuity. Sustainable Chemical Processes 4, 2, (2016). http://dx.doi.org/10.1186/s40508-016-0046-9

14 Arora, B., Mukherjee, J. & Gupta, M. N. Enzyme promiscuity: using the dark side of enzyme specificity in white biotechnology. Sustainable Chemical Processes 2, 25, (2014). http://dx.doi.org/10.1186/s40508-014-0025-y

15 Bornscheuer, U. T. & Kazlauskas, R. J. Catalytic promiscuity in biocatalysis: using old enzymes to form new bonds and follow new pathways. Angewandte Chemie (International ed. in English) 43, 6032-6040, (2004). http://dx.doi.org/10.1002/anie.200460416

16 Coelho, P. S., Brustad, E. M., Kannan, A. & Arnold, F. H. Olefin cyclopropanation via carbene transfer catalyzed by engineered cytochrome P450 enzymes. Science 339, 307- 310, (2013). http://dx.doi.org/10.1126/science.1231434

17 Coelho, P. S. et al. A serine-substituted P450 catalyzes highly efficient carbene transfer to olefins in vivo. Nat Chem Biol 9, 485-487, (2013). http://dx.doi.org/10.1038/nchembio.1278

18 Farwell, C. C., McIntosh, J. A., Hyster, T. K., Wang, Z. J. & Arnold, F. H. Enantioselective imidation of sulfides via enzyme-catalyzed intermolecular nitrogen- atom transfer. J Am Chem Soc 136, 8766-8771, (2014). http://dx.doi.org/10.1021/ja503593n

19 Farwell, C. C., Zhang, R. K., McIntosh, J. A., Hyster, T. K. & Arnold, F. H. Enantioselective Enzyme-Catalyzed Aziridination Enabled by Active-Site Evolution of a Cytochrome P450. ACS Cent Sci 1, 89-93, (2015). http://dx.doi.org/10.1021/acscentsci.5b00056

20 Heel, T., McIntosh, J. A., Dodani, S. C., Meyerowitz, J. T. & Arnold, F. H. Non-natural olefin cyclopropanation catalyzed by diverse cytochrome P450s and other hemoproteins. Chembiochem 15, 2556-2562, (2014). http://dx.doi.org/10.1002/cbic.201402286

21 Hyster, T. K., Farwell, C. C., Buller, A. R., McIntosh, J. A. & Arnold, F. H. Enzyme- controlled nitrogen-atom transfer enables regiodivergent C-H amination. J Am Chem Soc 136, 15505-15508, (2014). http://dx.doi.org/10.1021/ja509308v

22 McIntosh, J. A. et al. Enantioselective intramolecular C-H amination catalyzed by engineered cytochrome P450 enzymes in vitro and in vivo. Angewandte Chemie (International ed. in English) 52, 9309-9312, (2013). http://dx.doi.org/10.1002/anie.201304401

23 Bordeaux, M., Tyagi, V. & Fasan, R. Highly diastereoselective and enantioselective olefin cyclopropanation using engineered myoglobin-based catalysts. Angewandte

124

Chemie (International ed. in English) 54, 1744-1748, (2015). http://dx.doi.org/10.1002/anie.201409928

24 Singh, R., Bordeaux, M. & Fasan, R. P450-catalyzed intramolecular sp3 C-H amination with arylsulfonyl azide substrates. ACS Catal 4, 546-552, (2014). http://dx.doi.org/10.1021/cs400893n

25 Siegel, J. B. et al. Computational design of an enzyme catalyst for a stereoselective bimolecular Diels-Alder reaction. Science 329, 309-313, (2010). http://dx.doi.org/10.1126/science.1190239

26 Wang, Z. J., Peck, N. E., Renata, H. & Arnold, F. H. Cytochrome P450-Catalyzed Insertion of Carbenoids into N-H Bonds. Chem Sci 5, 598-601, (2014). http://dx.doi.org/10.1039/C3SC52535J

27 Hyster, T. K., Knorr, L., Ward, T. R. & Rovis, T. Biotinylated Rh(III) complexes in engineered streptavidin for accelerated asymmetric C-H activation. Science 338, 500-503, (2012). http://dx.doi.org/10.1126/science.1226132

28 Arnold, F. H. The nature of chemical innovation: new enzymes by evolution. Q Rev Biophys 48, 404-410, (2015). http://dx.doi.org/10.1017/S003358351500013X

29 Yamamura, K. & Kaiser, E. T. Studies on the oxidase activity of copper(II) carboxypeptidase A. Journal of the Chemical Society, Chemical Communications, 830- 831, (1976). http://dx.doi.org/10.1039/C39760000830

30 Bakker, M., van Rantwijk, F. & Sheldon, R. A. Metal substitution in thermolysin: Catalytic properties of tungstate thermolysin in sulfoxidation with H2O2. Can J Chem 80, 622-625, (2002). http://dx.doi.org/10.1139/V02-082

31 Neuwald, A. F. An unexpected structural relationship between integral membrane phosphatases and soluble haloperoxidases. Protein Science 6, 1764-1767, (1997).

32 Littlechild, J., Garcia-Rodriguez, E., Dalby, A. & Isupov, M. Structural and functional comparisons between vanadium haloperoxidase and acid phosphatase enzymes. J Mol Recognit 15, 291-296, (2002). http://dx.doi.org/10.1002/jmr.590

33 Hemrika, W., Renirie, R., Dekker, H. L., Barnett, P. & Wever, R. From phosphatases to vanadium peroxidases: A similar architecture of the active site. P Natl Acad Sci USA 94, 2145-2149, (1997). http://dx.doi.org/DOI 10.1073/pnas.94.6.2145

34 Tanaka, N., Dumay, V., Liao, Q. N., Lange, A. J. & Wever, R. Bromoperoxidase activity of vanadate-substituted acid phosphatases from Shigella flexneri and Salmonella enterica ser. typhimurium. European Journal of Biochemistry 269, 2162-2167, (2002). http://dx.doi.org/10.1046/j.1432-1033.2002.02871.x

125

35 van de Velde, F., Konemann, L., van Rantwijk, F. & Sheldon, R. A. The rational design of semisynthetic peroxidases. Biotechnol Bioeng 67, 87-96, (2000).

36 Bell, I. M., Fisher, M. L., Wu, Z. P. & Hilvert, D. Kinetic studies on the peroxidase activity of selenosubtilisin. Biochemistry 32, 3754-3762, (1993).

37 Broun, P., Shanklin, J., Whittle, E. & Somerville, C. Catalytic plasticity of fatty acid modification enzymes underlying chemical diversity of plant lipids. Science 282, 1315- 1317, (1998).

38 Broadwater, J. A., Whittle, E. & Shanklin, J. Desaturation and hydroxylation. Residues 148 and 324 of Arabidopsis FAD2, in addition to substrate chain length, exert a major influence in partitioning of catalytic specificity. The Journal of biological chemistry 277, 15613-15620, (2002). http://dx.doi.org/10.1074/jbc.M200231200

39 Schmidt, D. M. et al. Evolutionary potential of (beta/alpha)8-barrels: functional promiscuity produced by single substitutions in the enolase superfamily. Biochemistry 42, 8387-8393, (2003). http://dx.doi.org/10.1021/bi034769a

40 Nilsson, L. O., Gustafsson, A. & Mannervik, B. Redesign of substrate-selectivity determining modules of glutathione transferase A1-1 installs high catalytic efficiency with toxic alkenal products of lipid peroxidation. Proc Natl Acad Sci U S A 97, 9408- 9412, (2000). http://dx.doi.org/10.1073/pnas.150084897

41 Pettersson, P. L., Johansson, A. S. & Mannervik, B. Transmutation of human glutathione transferase A2-2 with peroxidase activity into an efficient steroid isomerase. The Journal of biological chemistry 277, 30019-30022, (2002). http://dx.doi.org/10.1074/jbc.M204485200

42 Nixon, A. E., Firestine, S. M., Salinas, F. G. & Benkovic, S. J. Rational design of a scytalone dehydratase-like enzyme using a structurally homologous protein scaffold. P Natl Acad Sci USA 96, 3568-3571, (1999). http://dx.doi.org/10.1073/pnas.96.7.3568

43 Goldsmith, M. & Tawfik, D. S. Directed enzyme evolution: beyond the low-hanging fruit. Curr Opin Struct Biol 22, 406-412, (2012). http://dx.doi.org/10.1016/j.sbi.2012.03.010

44 Fasan, R., Meharenna, Y. T., Snow, C. D., Poulos, T. L. & Arnold, F. H. Evolutionary history of a specialized p450 propane monooxygenase. J Mol Biol 383, 1069-1080, (2008). http://dx.doi.org/10.1016/j.jmb.2008.06.060

45 Koch, D. J., Chen, M. M., van Beilen, J. B. & Arnold, F. H. In vivo evolution of butane oxidation by terminal alkane hydroxylases AlkB and CYP153A6. Appl Environ Microbiol 75, 337-344, (2009). http://dx.doi.org/10.1128/AEM.01758-08

126

46 Brustad, E. M. & Arnold, F. H. Optimizing non-natural protein function with directed evolution. Current opinion in chemical biology 15, 201-210, (2011). http://dx.doi.org/10.1016/j.cbpa.2010.11.020

47 Esvelt, K. M., Carlson, J. C. & Liu, D. R. A system for the continuous directed evolution of biomolecules. Nature 472, 499-503, (2011). http://dx.doi.org/10.1038/nature09929

48 Hult, K. & Berglund, P. Enzyme promiscuity: mechanism and applications. Trends Biotechnol 25, 231-238, (2007). http://dx.doi.org/10.1016/j.tibtech.2007.03.002

49 Clarke, S. G. & Banfield, K. in Homocysteine in Health and Disease (eds Ralph Carmel & Donald W. Jacobsen) Ch. 7, 63-78 (University Press, 2001).

50 Schubert, H. L., Blumenthal, R. M. & Cheng, X. in The Enzymes Vol. Volume 24 (eds G. Clarke Steven & Tamanoi Fuyuhiko) 3-28 (Academic Press, 2006). http://dx.doi.org/10.1016/S1874-6047(06)80003-X.

51 Chang, A. et al. BRENDA in 2015: exciting developments in its 25th year of existence. Nucleic Acids Res 43, D439-446, (2015). http://dx.doi.org/10.1093/nar/gku1068

52 Schomburg, D. BRENDA: The Comprehensive Enzyme Information System, (2016).

53 Klimasauskas, S. & Weinhold, E. A new tool for biotechnology: AdoMet-dependent methyltransferases. Trends Biotechnol 25, 99-104, (2007). http://dx.doi.org/10.1016/j.tibtech.2007.01.006

54 Struck, A. W., Thompson, M. L., Wong, L. S. & Micklefield, J. S-adenosyl-methionine- dependent methyltransferases: highly versatile enzymes in biocatalysis, biosynthesis and other biotechnological applications. Chembiochem 13, 2642-2655, (2012). http://dx.doi.org/10.1002/cbic.201200556

55 Zhang, J. & Zheng, Y. G. SAM/SAH Analogs as Versatile Tools for SAM-Dependent Methyltransferases. ACS Chem Biol 11, 583-597, (2016). http://dx.doi.org/10.1021/acschembio.5b00812

56 Bothwell, I. R. & Luo, M. Large-scale, protection-free synthesis of Se-adenosyl-L- selenomethionine analogues and their application as cofactor surrogates of methyltransferases. Organic letters 16, 3056-3059, (2014). http://dx.doi.org/10.1021/ol501169y

57 Huber, T. D., Johnson, B. R., Zhang, J. & Thorson, J. S. AdoMet analog synthesis and utilization: current state of the art. Curr Opin Biotechnol 42, 189-197, (2016). http://dx.doi.org/10.1016/j.copbio.2016.07.005

58 Thomsen, M., Vogensen, S. B., Buchardt, J., Burkart, M. D. & Clausen, R. P. Chemoenzymatic synthesis and in situ application of S-adenosyl-L-methionine analogs.

127

Organic & biomolecular chemistry 11, 7606-7610, (2013). http://dx.doi.org/10.1039/c3ob41702f

59 Willnow, S., Martin, M., Luscher, B. & Weinhold, E. A selenium-based click AdoMet analogue for versatile substrate labeling with wild-type protein methyltransferases. Chembiochem 13, 1167-1173, (2012). http://dx.doi.org/10.1002/cbic.201100781

60 Lee, B. W. et al. Enzyme-catalyzed transfer of a ketone group from an S- adenosylmethionine analogue: a tool for the functional analysis of methyltransferases. J Am Chem Soc 132, 3642-3643, (2010). http://dx.doi.org/10.1021/ja908995p

61 Islam, K. et al. Bioorthogonal profiling of protein methylation using azido derivative of S-adenosyl-L-methionine. J Am Chem Soc 134, 5909-5915, (2012). http://dx.doi.org/10.1021/ja2118333

62 Dalhoff, C., Lukinavicius, G., Klimasauskas, S. & Weinhold, E. Direct transfer of extended groups from synthetic cofactors by DNA methyltransferases. Nat Chem Biol 2, 31-32, (2006). http://dx.doi.org/10.1038/nchembio754

63 Bothwell, I. R. et al. Se-adenosyl-L-selenomethionine cofactor analogue as a reporter of protein methylation. J Am Chem Soc 134, 14905-14912, (2012). http://dx.doi.org/10.1021/ja304782r

64 Borchardt, R. T. & Wu, Y. S. Potential inhibitors of S-adenosylmethionine-dependent methyltransferases. 3. Modifications of the sugar portion of S-adenosylhomocysteine. J Med Chem 18, 300-304, (1975).

65 Coward, J. K., D'Urso-Scott, M. & Sweet, W. D. Inhibition of catechol-O- methyltransferase by S-adenosylhomocysteine and S-adenosylhomocysteine sulfoxide, a potential transition-state analog. Biochemical pharmacology 21, 1200-1203, (1972).

66 Gillet, L., Looze, Y., Deconinck, M. & Leonis, J. Binding capacities of various analogues of S-adenosyl-L-homocysteine to protein methyltransferase II from human erythrocytes. Experientia 35, 1007-1009, (1979).

67 Pugh, C. S. & Borchardt, R. T. Effects of S-adenosylhomocysteine analogues on vaccinia viral messenger ribonucleic acid synthesis and methylation. Biochemistry 21, 1535-1541, (1982).

68 Guerard, C., Breard, M., Courtois, F., Drujon, T. & Ploux, O. Synthesis and evaluation of analogues of S-adenosyl-L-methionine, as inhibitors of the E. coli cyclopropane fatty acid synthase. Bioorganic & medicinal chemistry letters 14, 1661-1664, (2004). http://dx.doi.org/10.1016/j.bmcl.2004.01.051

69 Michio, I. & Kenta, A. From Sulfur to Selenium. A New Research Arena in Chemical Biology and Biological Chemistry. Current Chemical Biology 7, 2-24, (2013). http://dx.doi.org/10.2174/2212796811307010002

128

70 Weller, R. L. & Rajski, S. R. Design, synthesis, and preliminary biological evaluation of a DNA methyltransferase-directed alkylating agent. Chembiochem 7, 243-245, (2006). http://dx.doi.org/10.1002/cbic.200500362

71 Osborne, T., Roska, R. L., Rajski, S. R. & Thompson, P. R. In situ generation of a bisubstrate analogue for protein arginine methyltransferase 1. J Am Chem Soc 130, 4574- 4575, (2008). http://dx.doi.org/10.1021/ja077104v

72 Comstock, L. R. & Rajski, S. R. Conversion of DNA methyltransferases into azidonucleosidyl transferases via synthetic cofactors. Nucleic Acids Res 33, 1644-1652, (2005). http://dx.doi.org/10.1093/nar/gki306

73 Mudd, S. H. & Cantoni, G. L. Selenomethionine in enzymatic transmethylations. Nature 180, 1052, (1957).

74 Wrobel, K., Wrobel, K. & Caruso, J. A. Selenium speciation in low molecular weight fraction of Se-enriched yeasts by HPLC-ICP-MS: detection of selenoadenosylmethionine. J Anal Atom Spectrom 17, 1048-1054, (2002). http://dx.doi.org/10.1039/b200920j

75 Thomson, C. D. Assessment of requirements for selenium and adequacy of selenium status: a review. Eur J Clin Nutr 58, 391-402, (2004). http://dx.doi.org/10.1038/sj.ejcn.1601800

76 Tinggi, U. in Reviews in Food and Nutrition Toxicity, Volume 4 Resource Management (CRC Press, 2005). http://dx.doi.org/10.1201/9781420037524.ch2.

77 Reilly, C. Selenium in Food and Health. (Blackie Academic & Professional, 1996).

78 Gammelgaard, B., Gabel-Jensen, C., Sturup, S. & Hansen, H. R. Complementary use of molecular and element-specific mass spectrometry for identification of selenium compounds related to human selenium metabolism. Anal Bioanal Chem 390, 1691-1706, (2008). http://dx.doi.org/10.1007/s00216-007-1788-8

79 Lenz, M. & Lens, P. N. The essential toxin: the changing perception of selenium in environmental sciences. Sci Total Environ 407, 3620-3633, (2009). http://dx.doi.org/10.1016/j.scitotenv.2008.07.056

80 Ogra, Y. et al. Speciation of selenomethionine metabolites in wheat germ extract. Metallomics 1, 78-86, (2009). http://dx.doi.org/10.1039/B813118J

81 Rao, Y. et al. Mapping of selenium metabolic pathway in yeast by liquid chromatography-Orbitrap mass spectrometry. Anal Chem 82, 8121-8130, (2010). http://dx.doi.org/10.1021/ac1011798

129

82 Rayman, M. P. Selenium and human health. Lancet 379, 1256-1268, (2012). http://dx.doi.org/10.1016/S0140-6736(11)61452-9

83 Hatfield, D. L., Tsuji, P. A., Carlson, B. A. & Gladyshev, V. N. Selenium and selenocysteine: roles in cancer, health, and development. Trends Biochem Sci 39, 112- 120, (2014). http://dx.doi.org/10.1016/j.tibs.2013.12.007

84 B'Hymer, C. & Caruso, J. A. Selenium speciation analysis using inductively coupled plasma-mass spectrometry. J Chromatogr A 1114, 1-20, (2006). http://dx.doi.org/10.1016/j.chroma.2006.02.063

85 Finkelstein, J. D. Methionine metabolism in mammals. J Nutr Biochem 1, 228-237, (1990).

86 Loenen, W. A. S-adenosylmethionine: jack of all trades and master of everything? Biochem Soc Trans 34, 330-333, (2006). http://dx.doi.org/10.1042/BST20060330

87 Chiang, P. K. et al. S-Adenosylmethionine and methylation. FASEB J 10, 471-480, (1996).

88 Stolowitz, M. L. & Minch, M. J. S-Adenosyl-L-Methionine and S-Adenosyl-L- Homocysteine, an Nmr-Study. J Am Chem Soc 103, 6015-6019, (1981). http://dx.doi.org/10.1021/ja00410a004

89 Duerre, J. A. Preparation and properties of S-adenosyl-L-homocysteine, S-adenosyl-L- homocysteine sulfoxide and S-ribosyl-L-homocysteine. Arch Biochem Biophys 96, 70-76, (1962).

90 Biastoff, S., Teuber, M., Zhou, Z. S. & Drager, B. Colorimetric activity measurement of a recombinant putrescine N-methyltransferase from Datura stramonium. Planta Med 72, 1136-1141, (2006). http://dx.doi.org/10.1055/s-2006-947191

91 Duerre, J. A., Salisbury, L. & Miller, C. H. Preparation and characterization of sulfoxides of S-adenosyl-l-homocysteine and S-ribosyl-l-homocysteine. Analytical Biochemistry 35, 505-514, (1970). http://dx.doi.org/10.1016/0003-2697(70)90213-7

92 Rekunova, V. N., Rudakova, I. P. & Yurkevich, A. M. Synthesis of potential inhibitors of transmethylases. Tetrahedron Letters 14, 2811-2814, (1973). http://dx.doi.org/10.1016/S0040-4039(01)96057-5

93 Hickey, S. F. & Hammond, M. C. Structure-guided design of fluorescent S- adenosylmethionine analogs for a high-throughput screen to target SAM-I riboswitch RNAs. Chemistry & biology 21, 345-356, (2014). http://dx.doi.org/10.1016/j.chembiol.2014.01.004

130

94 Duclos, R. I., Jr., Cleary, D. C., Catcott, K. C. & Zhou, Z. S. Synthesis and characterization of Se-adenosyl-L-selenohomocysteine selenoxide. J Sulphur Chem 36, 135-144, (2015). http://dx.doi.org/10.1080/17415993.2014.979173

95 Ramalingam, K. & Woodard, R. W. A convenient synthesis of s-(5′-deoxy-5′-adenosyl)- (±) -2-methylhomocysteine. Tetrahedron Letters 26, 1135-1136, (1985). http://dx.doi.org/10.1016/S0040-4039(00)98415-6

96 Zhou, Z. S., Smith, A. E. & Matthews, R. G. L-Selenohomocysteine: one-step synthesis from L-selenomethionine and kinetic analysis as substrate for methionine synthases. Bioorganic & medicinal chemistry letters 10, 2471-2475, (2000).

97 Scovill, J. P., Thigpen, D. L. & Lemley, P. V. A Convenient Method for the Synthesis and Raney Nickel Desulfurization of 5′-deoxy-5′ -methylthioadenosine. Phosphorus, Sulfur, and Silicon and the Related Elements 85, 149-152, (1993). http://dx.doi.org/10.1080/10426509308038193

98 Robins, M. J., Hansske, F., Wnuk, S. F. & Kanai, T. Nucleic acid related compounds. 66. Improved syntheses of 5′-chloro-5′-deoxy- and 5′-S-aryl(or alkyl)-5′-thionucleosides. Canadian Journal of Chemistry 69, 1468-1474, (1991). http://dx.doi.org/10.1139/v91- 217

99 Orlov, I. G., Markin, V. S., Moiseev, Y. V. & Khurgin, U. I. Infrared spectra of amino acids and peptides. Chemistry of Natural Compounds 3, 163-166, (1967). http://dx.doi.org/10.1007/BF00567992

100 Paetzold, R., Lindner, U., Bochmann, G. & Reich, P. Dimethyl-und diathylselenoxide sowie ihre oxoniumsalze. darstellung, eigenschaften und schwingungsspektren. Zeits Anorgan Allge Chemie. 352, 295-308, (1967).

101 Reich, H. J., Renga, J. M. & Reich, I. L. Organoselenium Chemistry - Conversion of Ketones to Enones by Selenoxide Syn Elimination. J Am Chem Soc 97, 5434-5447, (1975). http://dx.doi.org/DOI 10.1021/ja00852a019

102 Katritzky, A. R., Meth-Cohn, O., Gilchrist, T. L. & Rees, C. W. Comprehensive Organic Functional Group Transformations. (Pergamon, 1995).

103 Zainal, H. A., Wolf, W. R. & Waters, R. M. An NMR spectroscopic investigation of the oxidation reactions of DL-selenomethionine. J Chem Technol Biot 72, 38-44, (1998). http://dx.doi.org/10.1002/(SICI)1097-4660(199805)72:1<38::AID-JCTB863>3.0.CO;2-R

104 Block, E. et al. Allium chemistry: synthesis, natural occurrence, biological activity, and chemistry of Se-alk(en)ylselenocysteines and their gamma-glutamyl derivatives and oxidation products. Journal of agricultural and food chemistry 49, 458-470, (2001).

131

105 Ritchey, J. A., Davis, B. M., Pleban, P. A. & Bayse, C. A. Experimental and theoretical evidence for cyclic selenurane formation during selenomethionine oxidation. Organic & biomolecular chemistry 3, 4337-4342, (2005). http://dx.doi.org/10.1039/b513238j

106 Davis, F. A. & Reddy, R. T. Asymmetric oxidation of simple selenides to selenoxides in high enantiopurity. Stereochemical aspects of the allyl selenoxide/allyl selenenate rearrangement. The Journal of Organic Chemistry 57, 2599-2606, (1992). http://dx.doi.org/10.1021/jo00035a014

107 Kurose, N., Takahashi, T. & Koizumi, T. First synthesis of optically pure selenuranes and stereoselective alkaline hydrolysis. Their application to asymmetric [2,3] sigmatropic rearrangement of allylic selenoxides. Tetrahedron 53, 12115-12129, (1997). http://dx.doi.org/10.1016/S0040-4020(97)00546-2

108 Schoneich, C. Redox processes of methionine relevant to beta-amyloid oxidation and Alzheimer's disease. Arch Biochem Biophys 397, 370-376, (2002). http://dx.doi.org/10.1006/abbi.2001.2621

109 Jacob, C., Giles, G. I., Giles, N. M. & Sies, H. Sulfur and selenium: the role of oxidation state in protein structure and function. Angewandte Chemie (International ed. in English) 42, 4742-4758, (2003). http://dx.doi.org/10.1002/anie.200300573

110 Isab, A. A. A 1H NMR study of the reaction of gold(III) with DL-seleno-methionine in aqueous solution. Inorganica Chimica Acta 80, L3-L4, (1983). http://dx.doi.org/10.1016/S0020-1693(00)91228-3

111 Uden, P. C. et al. Analytical selenoamino acid studies by chromatography with interfaced atomic mass spectrometry and atomic emission spectral detection. Fresenius' Journal of Analytical Chemistry 362, 447-456, (1998). http://dx.doi.org/10.1007/s002160051105

112 Gammelgaard, B., Cornett, C., Olsen, J., Bendahl, L. & Hansen, S. H. Combination of LC-ICP-MS, LC-MS and NMR for investigation of the oxidative degradation of selenomethionine. Talanta 59, 1165-1171, (2003). http://dx.doi.org/10.1016/s0039- 9140(03)00026-2

113 Cooper, A. J. & Meister, A. Enzymatic oxidation of L-homocysteine. Arch Biochem Biophys 239, 556-566, (1985).

114 Bănică, A., Culeţu, A. & Bănică, F.-G. Electrochemical and EQCM investigation of l- selenomethionine in adsorbed state at gold electrodes. Journal of Electroanalytical Chemistry 599, 100-110, (2007). http://dx.doi.org/10.1016/j.jelechem.2006.09.023

115 Harman, L. S., Mottley, C. & Mason, R. P. Free radical metabolites of L-cysteine oxidation. The Journal of biological chemistry 259, 5606-5611, (1984).

132

116 Reddie, K. G. & Carroll, K. S. Expanding the functional diversity of proteins through cysteine oxidation. Current opinion in chemical biology 12, 746-754, (2008). http://dx.doi.org/10.1016/j.cbpa.2008.07.028

117 Devarie-Baez, N. O., Zhang, D., Li, S., Whorton, A. R. & Xian, M. Direct methods for detection of protein S-nitrosylation. Methods (San Diego, Calif.) 62, 171-176, (2013). http://dx.doi.org/10.1016/j.ymeth.2013.04.018

118 Sasaki, E. et al. Co-opting sulphur-carrier proteins from primary metabolic pathways for 2-thiosugar biosynthesis. Nature 510, 427-431, (2014). http://dx.doi.org/10.1038/nature13256

119 De Silva, V., Woznichak, M. M., Burns, K. L., Grant, K. B. & May, S. W. Selenium redox cycling in the protective effects of organoselenides against oxidant-induced DNA damage. J Am Chem Soc 126, 2409-2413, (2004). http://dx.doi.org/10.1021/ja037294j

120 Iwig, D. F. & Booker, S. J. Insight into the polar reactivity of the onium chalcogen analogues of S-adenosyl-L-methionine. Biochemistry 43, 13496-13509, (2004). http://dx.doi.org/10.1021/bi048693+

121 Wallace, T. J. & Mahon, J. J. Reactions of Thiols with Sulfoxides. III. Catalysis by Acids and Bases. The Journal of Organic Chemistry 30, 1502-1506, (1965). http://dx.doi.org/10.1021/jo01016a039

122 Firouzabadi, H. & Jamalian, A. Reduction of oxygenated organosulfur compounds. Journal of Sulfur Chemistry 29, 53-97, (2008). http://dx.doi.org/10.1080/17415990701684776

123 Hirayama, K., Akashi, S., Furuya, M. & Fukuhara, K. Rapid confirmation and revision of the primary structure of bovine serum albumin by ESIMS and Frit-FAB LC/MS. Biochem Biophys Res Commun 173, 639-646, (1990).

124 Paris, G., Kraszewski, S., Ramseyer, C. & Enescu, M. About the structural role of disulfide bridges in serum albumins: evidence from protein simulated unfolding. Biopolymers 97, 889-898, (2012). http://dx.doi.org/10.1002/bip.22096

125 Takata, Y. et al. Catalytic mechanism of glycine N-methyltransferase. Biochemistry 42, 8394-8402, (2003). http://dx.doi.org/10.1021/bi034245a

126 Du, X. et al. A catechol-O-methyltransferase that is essential for auditory function in mice and humans. P Natl Acad Sci USA 105, 14609-14614, (2008). http://dx.doi.org/10.1073/pnas.0807219105

127 Tamanoi, F. & Clarke, S. G. The Enzymes: Protein Methyltransferases. (Elsevier Science, 2011).

133

128 Schubert, H. L., Blumenthal, R. M. & Cheng, X. Many paths to methyltransfer: a chronicle of convergence. Trends Biochem Sci 28, 329-335, (2003). http://dx.doi.org/10.1016/S0968-0004(03)00090-2

129 Boriack-Sjodin, P. A. & Swinger, K. K. Protein Methyltransferases: A Distinct, Diverse, and Dynamic Family of Enzymes. Biochemistry 55, 1557-1569, (2016). http://dx.doi.org/10.1021/acs.biochem.5b01129

130 Axelrod, J. & Tomchick, R. Enzymatic O-methylation of epinephrine and other catechols. The Journal of biological chemistry 233, 702-705, (1958).

131 Senoh, S., Tokuyama, Y. & Witkop, B. The Role of Cations in Non-enzymatic and Enzymatic O-Methylations of Catechol Derivatives. J Am Chem Soc 84, 1719-1724, (1962). http://dx.doi.org/10.1021/ja00868a045

132 Guldberg, H. C. & Marsden, C. A. Catechol-O-Methyl Transferase: Pharmacological Aspects and Physiological Role. Pharmacological Reviews 27, 135, (1975).

133 Vidgren, J., Svensson, L. A. & Liljas, A. Crystal structure of catechol O- methyltransferase. Nature 368, 354-358, (1994).

134 Espinoza, S., Manago, F., Leo, D., Sotnikova, T. D. & Gainetdinov, R. R. Role of catechol-O-methyltransferase (COMT)-dependent processes in Parkinson's disease and L-DOPA treatment. CNS Neurol Disord Drug Targets 11, 251-263, (2012).

135 Scheggia, D., Sannino, S., Scattoni, M. L. & Papaleo, F. COMT as a drug target for cognitive functions and dysfunctions. CNS Neurol Disord Drug Targets 11, 209-221, (2012).

136 Adler, C. H., Singer, C., O'Brien, C. & et al. RAndomized, placebo-controlled study of tolcapone in patients with fluctuating parkinson disease treated with levodopa-carbidopa. Archives of Neurology 55, 1089-1095, (1998). http://dx.doi.org/10.1001/archneur.55.8.1089

137 Lees, A. J. Evidence-based efficacy comparison of tolcapone and entacapone as adjunctive therapy in Parkinson's disease. CNS Neurosci Ther 14, 83-93, (2008). http://dx.doi.org/10.1111/j.1527-3458.2007.00035.x

138 Prada, M. D., Borgulya, J., Napolitano, A. & Zürcher, G. Improved Therapy of Parkinson's Disease with Tolcapone, a Central and Peripheral COMT Inhibitor with an S- Adenosyl-L-Methionine-Sparing Effect. Clinical Neuropharmacology 17, (1994).

139 Piccini, P. et al. The catechol-O-methyltransferase (COMT) inhibitor entacapone enhances the pharmacokinetic and clinical response to Sinemet CR in Parkinson's disease. J Neurol Neurosurg Psychiatry 68, 589-594, (2000).

134

140 Cotton, N. J., Stoddard, B. & Parson, W. W. Oxidative inhibition of human soluble catechol-O-methyltransferase. The Journal of biological chemistry 279, 23710-23718, (2004). http://dx.doi.org/10.1074/jbc.M401086200

141 Valerino, D. M. & McCormack, J. J. Xanthine oxidase-mediated oxidation of epinephrine. Biochemical pharmacology 20, 47-55, (1971).

142 Misra, H. P. & Fridovich, I. The Role of Superoxide Anion in the Autoxidation of Epinephrine and a Simple Assay for Superoxide Dismutase. Journal of Biological Chemistry 247, 3170-3175, (1972).

143 Trautner, E. M. & Bradley, T. R. The Early Stages of the Oxidation of Adrenaline in Dilute Solution. Australian Journal of Biological Sciences 4, 303-343, (1951).

144 Green, D. E. & Richter, D. Adrenaline and adrenochrome. Biochem J 31, 596-616, (1937).

145 Heacock, R. A. & Powell, W. S. Adrenochrome and related compounds. Progress in medicinal chemistry 9, 275-339, (1972).

146 Heacock, R. A. The Chemistry Of Adrenochrome And Related Compounds. Chemical Reviews 59, 181-237, (1959). http://dx.doi.org/10.1021/cr50026a001

147 Axelrod, J. Enzymic Oxidation of Epinephrine to Adrenochrome by the Salivary Gland. Biochim Biophys Acta 85, 247-254, (1964).

148 Matthews, S. B., Henderson, A. H. & Campbell, A. K. The adrenochrome pathway: the major route for adrenalin catabolism by polymorphonuclear leucocytes. J Mol Cell Cardiol 17, 339-348, (1985).

149 Garnayak, S. & Patel, S. Oxidation of Epinephrine to Adrenochrome by Cetyltrimethylammonium Dichromate: A Mechanistic Study. Ind Eng Chem Res 53, 12249-12256, (2014). http://dx.doi.org/10.1021/ie500037x

150 Lee, J. E., Cornell, K. A., Riscoe, M. K. & Howell, P. L. Expression, purification, crystallization and preliminary X-ray analysis of Escherichia coli 5'- methylthioadenosine/S-adenosylhomocysteine nucleosidase. Acta Crystallogr D Biol Crystallogr 57, 150-152, (2001).

151 Cannon, L. M., Butler, F. N., Wan, W. & Zhou, Z. S. A stereospecific colorimetric assay for (S,S)-adenosylmethionine quantification based on thiopurine methyltransferase- catalyzed thiol methylation. Anal Biochem 308, 358-363, (2002).

152 Schultz, E. & Nissinen, E. Inhibition of Rat-Liver and Duodenum Soluble Catechol- Ortho-Methyltransferase by a Tight-Binding Inhibitor or-462. Biochemical pharmacology 38, 3953-3956, (1989). http://dx.doi.org/Doi 10.1016/0006-2952(89)90673-4

135

153 Hegazi, M. F., Borchardt, R. T. & Schowen, R. L. SN2-like transition for methyl transfer catalyzed by catechol-O-methyltransferase. J Am Chem Soc 98, 3048-3049, (1976). http://dx.doi.org/10.1021/ja00426a079

154 Hegazi, M. F., Borchardt, R. T. & Schowen, R. L. .alpha.-Deuterium and carbon-13 isotope effects for methyl transfer catalyzed by catechol O-methyltransferase. SN2-like transition state. J Am Chem Soc 101, 4359-4365, (1979). http://dx.doi.org/10.1021/ja00509a052

155 Kalyanaraman, B., Premovic, P. I. & Sealy, R. C. Semiquinone anion radicals from addition of amino acids, peptides, and proteins to quinones derived from oxidation of catechols and catecholamines. An ESR spin stabilization study. Journal of Biological Chemistry 262, 11080-11087, (1987).

156 Nikolic, G. M., Premović, P. I. & Nikolić, R. S. Spectrophotometric Study of Catechol Oxidation by Aerial O2 in Alkaline Aqueous Solutions Containing Mg(II) Ions. Spectroscopy Letters 31, 327-333, (1998). http://dx.doi.org/10.1080/00387019808003257

157 Ehler, A., Benz, J., Schlatter, D. & Rudolph, M. G. Mapping the conformational space accessible to catechol-O-methyltransferase. Acta Crystallographica Section D D70, 2163-2174, (2014). http://dx.doi.org/10.1107/S1399004714012917

158 Rivett, A. J. & Roth, J. A. Kinetic studies on the O-methylation of dopamine by human brain membrane-bound catechol O-methyltransferase. Biochemistry 21, 1740-1742, (1982).

159 Rutherford, K., Alphandery, E., McMillan, A., Daggett, V. & Parson, W. W. The V108M mutation decreases the structural stability of catechol O-methyltransferase. Bba-Proteins Proteom 1784, 1098-1105, (2008). http://dx.doi.org/10.1016/j.bbapap.2008.04.006

160 Fomenko, D. E. & Gladyshev, V. N. CxxS: Fold-independent redox motif revealed by genome-wide searches for thiol/disulfide oxidoreductase function. Protein Science 11, 2285-2296, (2002). http://dx.doi.org/10.1110/ps.0218302

161 Fomenko, D. E. & Gladyshev, V. N. Identity and functions of CxxC-derived motifs. Biochemistry 42, 11214-11225, (2003). http://dx.doi.org/10.1021/bi034459s

162 Arner, E. S. & Holmgren, A. Physiological functions of thioredoxin and thioredoxin reductase. Eur J Biochem 267, 6102-6109, (2000).

163 Woycechowsky, K. J. & Raines, R. T. Native disulfide bond formation in proteins. Current opinion in chemical biology 4, 533-539, (2000).

164 Martin, J. L. Thioredoxin--a fold for all reasons. Structure 3, 245-250, (1995).

165 Vilbois, F. et al. Mass spectrometric analysis of human soluble catechol O- methyltransferase expressed in Escherichia coli. Identification of a product of ribosomal

136

frameshifting and of reactive cysteines involved in S-adenosyl-L-methionine binding. Eur J Biochem 222, 377-386, (1994).

166 Suwal, S. & Pflum, M. K. Phosphorylation-dependent kinase-substrate cross-linking. Angewandte Chemie (International ed. in English) 49, 1627-1630, (2010). http://dx.doi.org/10.1002/anie.200905244

167 Suwal, S., Senevirathne, C., Garre, S. & Pflum, M. K. Structural analysis of ATP analogues compatible with kinase-catalyzed labeling. Bioconjug Chem 23, 2386-2391, (2012). http://dx.doi.org/10.1021/bc300404s

168 Zhang, C., Weller, R. L., Thorson, J. S. & Rajski, S. R. Natural product diversification using a non-natural cofactor analogue of S-adenosyl-L-methionine. J Am Chem Soc 128, 2760-2761, (2006). http://dx.doi.org/10.1021/ja056231t

169 Pljevaljcic, G., Pignot, M. & Weinhold, E. Design of a new fluorescent cofactor for DNA methyltransferases and sequence-specific labeling of DNA. J Am Chem Soc 125, 3486- 3492, (2003). http://dx.doi.org/10.1021/ja021106s

170 Gottfried, A. & Weinhold, E. Sequence-specific covalent labelling of DNA. Biochem Soc Trans 39, 623-628, (2011). http://dx.doi.org/10.1042/BST0390623

171 Peters, W. et al. Enzymatic site-specific functionalization of protein methyltransferase substrates with alkynes for click labeling. Angewandte Chemie (International ed. in English) 49, 5170-5173, (2010). http://dx.doi.org/10.1002/anie.201001240

172 Wang, L. & Weinshilboum, R. Thiopurine S-methyltransferase pharmacogenetics: insights, challenges and future directions. Oncogene 25, 1629-1638, (2006). http://dx.doi.org/10.1038/sj.onc.1209372

173 Liu, M. et al. Protein isoaspartate methyltransferase-mediated 18O-labeling of isoaspartic acid for mass spectrometry analysis. Anal Chem 84, 1056-1062, (2012). http://dx.doi.org/10.1021/ac202652z

174 Liu, M. et al. Discovery of undefined protein cross-linking chemistry: a comprehensive methodology utilizing 18O-labeling and mass spectrometry. Anal Chem 85, 5900-5908, (2013). http://dx.doi.org/10.1021/ac400666p

175 Wan, W., Zhao, G., Al-Saad, K., Siems, W. F. & Zhou, Z. S. Rapid screening for S- adenosylmethionine-dependent methylation products by enzyme-transferred isotope patterns analysis. Rapid communications in mass spectrometry : RCM 18, 319-324, (2004). http://dx.doi.org/10.1002/rcm.1335

176 Spain, J. C. Biodegradation of nitroaromatic compounds. Annual review of microbiology 49, 523-555, (1995). http://dx.doi.org/10.1146/annurev.mi.49.100195.002515

137

177 Xu, Z., Chen, Y., Shen, W. & Cen, P. Adsorption and desorption behaviors of S- adenosyl-L-methionine in a fixed-bed ion-exchange column. Korean Journal of Chemical Engineering 23, 108-112, (2006). http://dx.doi.org/10.1007/BF02705700

178 Matos, J. R. & Wong, C.-H. S-adenosylmethionine: Stability and stabilization. Bioorganic Chemistry 15, 71-80, (1987). http://dx.doi.org/10.1016/0045-2068(87)90008- 3

179 Hanna, G. M. NMR regulatory analysis: determination and characterization of S- adenosyl-L-methionine in dietary supplements. Pharmazie 59, 251-256, (2004).

180 Hwang, T. L. & Shaka, A. J. Water Suppression That Works - Excitation Sculpting Using Arbitrary Wave-Forms and Pulsed-Field Gradients. J Magn Reson Ser A 112, 275-279, (1995). http://dx.doi.org/DOI 10.1006/jmra.1995.1047

181 Hoffman, J. L. Chromatographic analysis of the chiral and covalent instability of S- adenosyl-L-methionine. Biochemistry 25, 4444-4449, (1986).

182 Cornish-Bowden, A. Fundamentals of enzyme kinetics. (Butterworths, 1979).

183 Klotz, I. M. Introduction to biomolecular energetics : including ligand-receptor interactions. (Academic Press, 1986).

184 Woodson, L. C. & Weinshilboum, R. M. Human kidney thiopurine methyltransferase. Purification and biochemical properties. Biochemical pharmacology 32, 819-826, (1983). http://dx.doi.org/10.1016/0006-2952(83)90582-8

185 Coward, J. K. & Pegg, A. E. Specific multisubstrate adduct inhibitors of aminopropyltransferases and their effect on polyamine biosynthesis in cultured cells. Advances in enzyme regulation 26, 107-113, (1987).

186 Anderson, G. L., Bussolotti, D. L. & Coward, J. K. Synthesis and evaluation of some stable multisubstrate adducts as inhibitors of catechol O-methyltransferase. J Med Chem 24, 1271-1277, (1981).

187 Yau, E. K. & Coward, J. K. Synthesis of Complex 6'-Alkynyl-6'-Dethia Nucleoside Analogs of S-Adenosylhomocysteine as Potential Inhibitors of Methyltransferases. J Org Chem 55, 3147-3158, (1990). http://dx.doi.org/10.1021/Jo00297a034

188 Broom, A. D. Rational design of enzyme inhibitors: multisubstrate analogue inhibitors. J Med Chem 32, 2-7, (1989).

189 Le Calvez, P. B., Scott, C. J. & Migaud, M. E. Multisubstrate adduct inhibitors: drug design and biological tools. Journal of enzyme inhibition and medicinal chemistry 24, 1291-1318, (2009). http://dx.doi.org/10.3109/14756360902843809

138

190 Gower, C. M., Chang, M. E. & Maly, D. J. Bivalent inhibitors of protein kinases. Critical reviews in biochemistry and molecular biology 49, 102-115, (2014). http://dx.doi.org/10.3109/10409238.2013.875513

191 Wu, H. et al. Structural basis of allele variation of human thiopurine-S-methyltransferase. Proteins 67, 198-208, (2007). http://dx.doi.org/10.1002/prot.21272

192 Weinshilboum, R. M., Otterness, D. M. & Szumlanski, C. L. Methylation pharmacogenetics: catechol O-methyltransferase, thiopurine methyltransferase, and histamine N-methyltransferase. Annu Rev Pharmacol Toxicol 39, 19-52, (1999). http://dx.doi.org/10.1146/annurev.pharmtox.39.1.19

193 Leopold, W. R., Miller, J. A. & Miller, E. C. S--vinyl homocysteine, an analog of ethionine that is highly mutagenic for S. typhimurium TA100. Biochem Biophys Res Commun 88, 395-401, (1979).

194 Kim, H. J., Balcezak, T. J., Nathin, S. J., McMullen, H. F. & Hansen, D. E. The use of a spectrophotometric assay to study the interaction of S-adenosylmethionine synthetase with methionine analogues. Anal Biochem 207, 68-72, (1992).

195 Sohn, Y. W., Lee, G. H., Liem, A. & Miller, J. A. Activation of H-ras oncogenes in male B6C3F1 mouse liver tumors induced by vinthionine or 2-chloroethyl-methyl sulfide. Carcinogenesis 17, 1361-1364, (1996).

196 Leopold, W. R., Miller, J. A. & Miller, E. C. Comparison of some carcinogenic, mutagenic, and biochemical properties of S-vinylhomocysteine and ethionine. Cancer Res 42, 4364-4374, (1982).

197 Zhao, G. & Zhou, Z. S. Vinyl sulfonium as novel proteolytic enzyme inhibitor. Bioorganic & medicinal chemistry letters 11, 2331-2335, (2001).

198 Singh, S. et al. Facile chemoenzymatic strategies for the synthesis and utilization of S- adenosyl-(L)-methionine analogues. Angewandte Chemie (International ed. in English) 53, 3965-3969, (2014). http://dx.doi.org/10.1002/anie.201308272

199 Graham, D. E., Bock, C. L., Schalk-Hihi, C., Lu, Z. J. & Markham, G. D. Identification of a highly diverged class of S-adenosylmethionine synthetases in the archaea. The Journal of biological chemistry 275, 4055-4059, (2000).

200 Peng, Y. et al. Structural basis of substrate recognition in thiopurine s-methyltransferase. Biochemistry 47, 6216-6225, (2008). http://dx.doi.org/10.1021/bi800102x

201 Krieger, E. & Vriend, G. YASARA View - molecular graphics for all devices - from smartphones to workstations. Bioinformatics 30, 2981-2982, (2014). http://dx.doi.org/10.1093/bioinformatics/btu426

139

202 Smith, L. M., Kelleher, N. L. & Consortium for Top Down, P. Proteoform: a single term describing protein complexity. Nat Methods 10, 186-187, (2013). http://dx.doi.org/10.1038/nmeth.2369

203 Kuhn, M. L., Majorek, K. A., Minor, W. & Anderson, W. F. Broad-substrate screen as a tool to identify substrates for bacterial Gcn5-related N-acetyltransferases with unknown substrate specificity. Protein Sci 22, 222-230, (2013). http://dx.doi.org/10.1002/pro.2199

204 Levy, D. et al. A proteomic approach for the identification of novel lysine methyltransferase substrates. Epigenetics Chromatin 4, 19, (2011). http://dx.doi.org/10.1186/1756-8935-4-19

205 Xu, T., Zhang, L., Wang, X., Wei, D. & Li, T. Structure-based substrate screening for an enzyme. BMC Bioinformatics 10, 257, (2009). http://dx.doi.org/10.1186/1471-2105-10- 257

206 Flint, A. J., Tiganis, T., Barford, D. & Tonks, N. K. Development of "substrate-trapping" mutants to identify physiological substrates of protein tyrosine phosphatases. Proc Natl Acad Sci U S A 94, 1680-1685, (1997).

207 Lomenick, B. et al. Target identification using drug affinity responsive target stability (DARTS). Proc Natl Acad Sci U S A 106, 21984-21989, (2009). http://dx.doi.org/10.1073/pnas.0910040106

208 West, G. M., Tang, L. & Fitzgerald, M. C. Thermodynamic analysis of protein stability and ligand binding using a chemical modification- and mass spectrometry-based strategy. Anal Chem 80, 4175-4185, (2008). http://dx.doi.org/10.1021/ac702610a

209 Jiang, H., Kim, J. H., Frizzell, K. M., Kraus, W. L. & Lin, H. Clickable NAD analogues for labeling substrate proteins of poly(ADP-ribose) polymerases. J Am Chem Soc 132, 9363-9372, (2010). http://dx.doi.org/10.1021/ja101588r

210 Yu, Y., Kirkup, C. E., Pi, N. & Leary, J. A. Characterization of noncovalent protein- ligand complexes and associated enzyme intermediates of GlcNAc-6-O-sulfotransferase by electrospray ionization FT-ICR mass spectrometry. J Am Soc Mass Spectrom 15, 1400-1407, (2004). http://dx.doi.org/10.1016/j.jasms.2004.06.002

211 Yin, S., Xie, Y. & Loo, J. A. Mass spectrometry of protein-ligand complexes: enhanced gas-phase stability of ribonuclease-nucleotide complexes. J Am Soc Mass Spectrom 19, 1199-1208, (2008). http://dx.doi.org/10.1016/j.jasms.2008.05.012

212 Ganem, B., Li, Y. T. & Henion, J. D. Observation of noncovalent enzyme-substrate and enzyme-product complexes by ion-spray mass spectrometry. J Am Chem Soc 113, 7818- 7819, (1991). http://dx.doi.org/10.1021/ja00020a085

213 Wiseman, J. M., Takats, Z., Gologan, B., Davisson, V. J. & Cooks, R. G. Direct characterization of enzyme-substrate complexes by using electrosonic spray ionization

140

mass spectrometry. Angewandte Chemie (International ed. in English) 44, 913-916, (2005). http://dx.doi.org/10.1002/anie.200461672

214 Sun, J., Kitova, E. N., Wang, W. & Klassen, J. S. Method for distinguishing specific from nonspecific protein-ligand complexes in nanoelectrospray ionization mass spectrometry. Anal Chem 78, 3010-3018, (2006). http://dx.doi.org/10.1021/ac0522005

215 Catherman, A. D., Skinner, O. S. & Kelleher, N. L. Top Down proteomics: facts and perspectives. Biochem Biophys Res Commun 445, 683-693, (2014). http://dx.doi.org/10.1016/j.bbrc.2014.02.041

216 Qu, W. et al. Capturing Unknown Substrates via in Situ Formation of Tightly Bound Bisubstrate Adducts: S-Adenosyl-vinthionine as a Functional Probe for AdoMet- Dependent Methyltransferases. J Am Chem Soc 138, 2877-2880, (2016). http://dx.doi.org/10.1021/jacs.5b05950

217 Zhou, M., Jones, C. M. & Wysocki, V. H. Dissecting the large noncovalent protein complex GroEL with surface-induced dissociation and ion mobility-mass spectrometry. Anal Chem 85, 8262-8267, (2013). http://dx.doi.org/10.1021/ac401497c

218 Zhou, M., Dagan, S. & Wysocki, V. H. Protein subunits released by surface collisions of noncovalent complexes: nativelike compact structures revealed by ion mobility mass spectrometry. Angewandte Chemie (International ed. in English) 51, 4336-4339, (2012). http://dx.doi.org/10.1002/anie.201108700

219 Dugan, A. et al. Discovery of Enzymatic Targets of Transcriptional Activators via in Vivo Covalent Chemical Capture. J Am Chem Soc 138, 12629-12635, (2016). http://dx.doi.org/10.1021/jacs.6b07680

220 Ajees, A. A., Marapakala, K., Packianathan, C., Sankaran, B. & Rosen, B. P. Structure of an As(III) S-adenosylmethionine methyltransferase: insights into the mechanism of arsenic biotransformation. Biochemistry 51, 5476-5485, (2012). http://dx.doi.org/10.1021/bi3004632

221 Wessjohann, L., Dippe, M., Tengg, M. & Gruber-Khadjawi, M. in Cascade Biocatalysis 393-426 (Wiley-VCH Verlag GmbH & Co. KGaA, 2014). http://dx.doi.org/10.1002/9783527682492.ch18.

222 Mozier, N. M., McConnell, K. P. & Hoffman, J. L. S-adenosyl-L-methionine:thioether S- methyltransferase, a new enzyme in sulfur and selenium metabolism. The Journal of biological chemistry 263, 4527-4531, (1988).

223 Shul'pin, G. B. Selectivity enhancement in functionalization of C-H bonds: A review. Organic & biomolecular chemistry 8, 4217-4228, (2010). http://dx.doi.org/10.1039/c004223d

141

224 Wencel-Delord, J. & Glorius, F. C-H bond activation enables the rapid construction and late-stage diversification of functional molecules. Nat Chem 5, 369-375, (2013). http://dx.doi.org/10.1038/nchem.1607

225 Akamatsu, Y. & Law, J. H. Enzymatic alkylenation of phospholipid fatty acid chains by extracts of Mycobacterium phlei. The Journal of biological chemistry 245, 701-708, (1970).

226 Greenberg, R. & Dudock, B. S. Bacterial tRNA methyltransferases. Methods Enzymol 59, 190-203, (1979).

227 Kealey, J. T., Gu, X. & Santi, D. V. Enzymatic mechanism of tRNA (m5U54)methyltransferase. Biochimie 76, 1133-1142, (1994).

228 Gu, X., Ofengand, J. & Santi, D. V. In vitro methylation of Escherichia coli 16S rRNA by tRNA (m5U54)-methyltransferase. Biochemistry 33, 2255-2261, (1994).

229 Pfeifer, G. P., Grunwald, S., Boehm, T. L. & Drahovsky, D. Isolation and characterization of DNA cytosine 5-methyltransferase from human placenta. Biochim Biophys Acta 740, 323-330, (1983).

230 Roy, P. H. & Weissbach, A. DNA methylase from HeLa cell nuclei. Nucleic Acids Res 2, 1669-1684, (1975).

231 Tang, M. C., Fu, C. Y. & Tang, G. L. Characterization of SfmD as a Heme peroxidase that catalyzes the regioselective hydroxylation of 3-methyltyrosine to 3-hydroxy-5- methyltyrosine in saframycin A biosynthesis. The Journal of biological chemistry 287, 5112-5121, (2012). http://dx.doi.org/10.1074/jbc.M111.306316

232 Frenzel, T., Zhou, P. & Floss, H. G. Formation of 2-methyltryptophan in the biosynthesis of thiostrepton: isolation of S-adenosylmethionine:tryptophan 2-methyltransferase. Arch Biochem Biophys 278, 35-40, (1990).

233 Moore, S. J. et al. Characterization of the enzyme CbiH60 involved in anaerobic ring contraction of the cobalamin (vitamin B12) biosynthetic pathway. The Journal of biological chemistry 288, 297-305, (2013). http://dx.doi.org/10.1074/jbc.M112.422535

234 Weinshilboum, R. M., Sladek, S. & Klumpp, S. Human erythrocyte thiol methyltransferase: radiochemical microassay and biochemical properties. Clin Chim Acta 97, 59-71, (1979).

235 Comstock, L. R. & Rajski, S. R. Methyltransferase-directed DNA strand scission. J Am Chem Soc 127, 14136-14137, (2005). http://dx.doi.org/10.1021/ja054128y

236 Hymbaugh Bergman, S. J. & Comstock, L. R. N-mustard analogs of S-adenosyl-L- methionine as biochemical probes of protein arginine methylation. Bioorganic & medicinal chemistry 23, 5050-5055, (2015). http://dx.doi.org/10.1016/j.bmc.2015.05.001

142

237 Weller, R. L. & Rajski, S. R. DNA methyltransferase-moderated click chemistry. Organic letters 7, 2141-2144, (2005). http://dx.doi.org/10.1021/ol0504749

238 Pacifici, G. M., Santerini, S., Giuliani, L. & Rane, A. Thiol methyltransferase in humans: development and tissue distribution. Dev Pharmacol Ther 17, 8-15, (1991).

239 Keith, R. A., Jardine, I., Kerremans, A. & Weinshilboum, R. M. Human erythrocyte membrane thiol methyltransferase. S-methylation of captopril, N-acetylcysteine, and 7 alpha-thio-spirolactone. Drug Metab Dispos 12, 717-724, (1984).

240 Weisiger, R. A. & Jakoby, W. B. Thiol S-methyltransferase. Methods Enzymol 77, 257- 262, (1981).

241 Weisiger, R. A. & Jakoby, W. B. Thiol S-methyltransferase from rat liver. Arch Biochem Biophys 196, 631-637, (1979).

242 Drummer, O. H., Miach, P. & Jarrott, B. S-methylation of captopril. Demonstration of captopril thiol methyltransferase activity in human erythrocytes and enzyme distribution in rat tissues. Biochemical pharmacology 32, 1557-1562, (1983).

243 Keith, R. A., Otterness, D. M., Kerremans, A. L. & Weinshilboum, R. M. S-Methylation of D- and L-penicillamine by human erythrocyte membrane thiol methyltransferase. Drug Metab Dispos 13, 669-676, (1985).

244 Sakardi, B. et al. Red Blood Cell Collection. rbcc.hegelab.org, (2013).

245 Goodman, S. R., Daescu, O., Kakhniashvili, D. G. & Zivanic, M. The proteomics and interactomics of human erythrocytes. Exp Biol Med (Maywood) 238, 509-518, (2013). http://dx.doi.org/10.1177/1535370213488474

246 D'Alessandro, A., Righetti, P. G. & Zolla, L. The red blood cell proteome and interactome: an update. J Proteome Res 9, 144-163, (2010). http://dx.doi.org/10.1021/pr900831f

247 Konagurthu, A. S., Whisstock, J. C., Stuckey, P. J. & Lesk, A. M. MUSTANG: a multiple structural alignment algorithm. Proteins 64, 559-574, (2006). http://dx.doi.org/10.1002/prot.20921

143

“All sorts of things can happen when you’re open to new ideas and playing around with things.”

Stephanie Kwolek Inventor of DuPont™ Kevlar®

144