The Pennsylvania State University the Graduate School Eberly College of Science

The Pennsylvania State University The Graduate School Eberly College of Science

MECHANISTIC DISSECTION OF TAURINE α-KETOGLUTARATE DIOXYGENASE (TauD): A MODEL α-KETOGLUTARATE DIOXYGENASE

Submitted in Partial Fulfillment of the Requirements for the Degree of

Doctor of Philosophy

December 2005

Thesis of John C. Price was reviewed and approved∗ by the following:

Joseph M. Bollinger, Jr. Associate Professor of Biochemistry and Molecular Biology and Associate Professor of Chemistry Thesis Co-advisor Co-chair of Committee

Carsten Krebs Assistant Professor of Biochemistry and Molecular Biology and Assistant Professor of Chemistry Thesis Co-advisor Co-chair of Committee

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

Craig E. Cameron Paul Berg Professor of Biochemistry and Molecular Biology

Michael T. Green Assistant Professor of Chemistry

Robert A. Schlegel Professor of Biochemistry and Molecular Biology Head of the Department of Biochemistry and Molecular Biology

∗ Signatures are on file in the Graduate School

iii

Abstract

The oxidizing power of O2 is the basis for the function of the respiratory cycle and is used in many biosynthetic processes. Harnessing these oxidizing equivalents requires levels of precaution and precise control, even so preventing and repairing the damage due to reduced oxygen species is a constant effort in aerobic organisms. The

Fe(II)•α-ketoglutarate dioxygenase enzymes harness the oxidizing equivalents of O2 in a mechanism which requires the decarboxylation of α-ketoglutarate to create a highly oxidized Fe center. These enzymes are very effective at specific two electron oxidations of unactivated carbon atoms, and occupy key positions in a surprisingly large number of biological systems.

This study describes the mechanistic dissection of a model Fe(II)•α-ketoglutarate dioxygenase, taurine dioxygenase (TauD). Contributions made in the study of this system have experimentally expanded our understanding of these enzymes and verified theories postulated more than 20 years ago. Chapter 1 introduces general features of the enzyme family, and summarizes the literature describing what was known prior to this study regarding the mechanism of action for these enzymes. Chapter 1 also introduces several members of the family, using these systems to describe the utility of the family and the importance of understanding their mechanism of action. Chapter 2 presents all the work done in kinetically characterizing the model system (TauD) describing formation of the reactive enzyme complex and detection of intermediates that accumulate within the catalytic cycle. Chapter 3 delineates the use of sophisticated spectroscopy and a synthetic substrate isotopomer to chemically and structurally characterize the two accumulating intermediates. Chapter 4 explores the effect of substrate binding and the unproductive reaction(s) that occur in the absence of substrate. Chapter 5 contains studies from a variety of alternative substrates which offer interesting, yet not fully developed insights into the delicate balance that allows the active site to activate oxygen and successfully utilize the resulting oxidizing equivalents so effectively and specifically.

Table of Contents List of Figures…………………………………………………………………………...viii List of Tables………………………………………………………………………...…xvii Acknowledgements………………………………………………………………….xviii Chapter 1 Introduction………………………….………………………………...1

General features of the Fe(II)•αKG dioxygenase family…………….….1 HIFα hydroxylases……………………………………………………….2

AlkB……………………………………………………………………...3 Clavaminate synthase……………………………………………………4 SyrB2…………………………………………………………………….4 Prolyl-4-hydroxylase...... 5 Consensus mechanism for the Fe(II)•αKG dioxygenase family………..6 TauD……………………………………………………………………..9 References……………………………………………………………….24

Chapter 2 Kinetic dissection of the catalytic mechanism of taurine•α-ketoglutarate dioxygenase……………………………..………………………………..31 Abstract………………………………….……………………………….32 Introduction……………………………….……………………………...33 Materials and Methods…………………………………………………...34

Construction of TauD overexpression system……………..………...35 Overexpression of TauD………………………...…………...………35 Purification of TauD…………………………………………..……..36 Fe(II) titration…………………………………………………….….37 Determination of steady state sulfite production rate.…………….....37

Determination of steady state CO2 production rate……………….....38

Kinetics of CO2 in a single turnover…………………………………39 Mössbauer spectroscopy and data analysis………………………..…40

Stopped-flow experiments and simulations……………………….....40 Preparation of Mössbauer samples………………………………..…41 Results…………………………………………………………………....42 Determination of optimum Fe/TauD ratio and substrate concentrations...... 42 Changes at the active site upon binding of substrate (Mössbauer)…..42 Kinetics of substrate binding………………………………………...43 Stopped-flow evidence for two intermediates…………………….....45 Trapping J and its’ characterization by Mössbauer...…….………….45 Simulation of stopped-flow and steady state rate……………………46 Nature of M…………………………………………………………..47

Kinetics of CO2 evolution in a single turnover………………………49

[O2] dependence on the formation of J………………………………50 Substrate binding rates……………………………………………….52 Explanation of slow reformation for quaternary complex….………..53 Discussion………………………………………………………………..54 Formation of reactive TauD complex………………………………..54

Kinetic description of the reaction with O2…………………………..55 Identity of novel Fe intermediate J…………………………………..58 References………………………………………………………………..83 Chapter 3: Characterization of two accumulating intermediates in the catalytic cycle of

taurine•α-ketoglutarate dioxygenase…………………………………….88 Abstract…………………………………………………………………..91 Introduction………………………………………………………………92 Materials and methods…………………………………………………..93

2 Synthesis of 1,1-[ H]2-taurine (D-taurine)………………………...…93

High-field Mössbauer spectroscopy and analysis…………………..96 EPR spectroscopy……………………………………………………96

Cryoreduction by low temperature γ-radiolysis……………………..96 Results……………………………………………………………………96

Transient state CO2 evolution from D-taurine containing complex…96 Identification of J as hydroxylating intermediate……………………97

2- Uncoupling of CO2 and SO3 production……………………………98 Freeze-quench Mössbauer…………………………………………...99 High-field Mössbauer………………………………………………100 Cryoreduction of J to Fe(III)……………………………………….100 Structural characterization of J……………………………………..102 Nature of rate limiting step…………………………………………103 Discussion………………………………………………………………104 Reference……………………………………………………………….122

Chapter 4: Probing the mechanism of “untriggered” O2 activation by the TauD•Fe(II)•αKG ternary complex……………………………………126 Abstract…………………………………………………………………126 Introduction……………………………………………………………..127

Materials and methods………………………………………………….129 Results…………………………………………………………………..129 Stopped flow evidence for intermediates…………………………...130 Trapping Fe intermediate and characterization by Mössbauer…..…131 Discussion………………………………………………………………131 References………………………………………………………………141

Chapter 5: Use of substrate analogues in the dissection of the taurine•α-ketoglutarate dioxygenase reaction……………………………………………………143

vii

Abstract…………………………………………………………………143 Introduction……………………………………………………………..143 Materials and Methods………………………………………………….145

Synthesis of 1,1-[F]2-taurine (F-taurine)……………………………145 Synthesis of N-oxalylglycine……………………………………….148 Synthesis of 4-Oxo-4-thiocarboxy butyric acid…………………….148

Formation of TauD•Fe(III)•αKG•taurine•NO complex……………150 Use of Peracetic acid as an O atom donor………………………….151 Double mixing stopped-flow absorbance experiments……………..151 Results and Discussion…………………………………………………151

1,1-[F]2-taurine (F-taurine) in TauD quaternary complex………….151 4-oxo-4-thiocarboxy-butyric acid as an analog of αKG……………157 N-oxalylglycine as an analog of the αKG………………………….161

Peracetic acid as an O-atom donor to form J6……………………...163

NO as an O2 analog…………………………………………………165 References………………………………………………………………202 Appendix: Spectroscopic characterization of synthetic routes………………………....208

viii

List of Figures

Scheme 1-1: Generalized reaction of the Fe(II)•αKG dioxygenase family of enzymes……………………………………………………………………………14

Scheme 1-2: Generalized concensus mechanism of the Fe(II)•αKG dioxygenase family……………………………………………………………………………15 Scheme 2-1: The current working hypothesis for the TauD chemical mechanism……65 Scheme 2-2: Kinetic mechanism used to simulate the kinetics of the first intermediate determined……………………………………………………………………………66 Scheme 2-3: Kinetic mechanisms used to simulate the data in Figure 2-10……………67 Scheme 3-1: Mapping of the minimal kinetic mechanism for TauD onto the chemical mechanism………………………………………………………………………….107 Scheme 3-2: Kinetic mechanism used to simulate the kinetics of the first intermediate incorporating the kinetic isotope effect……………………………………………108

2 Scheme 3-3: Synthetic route to 1,1-[ H]2-taurine (D-taurine)………………………..109

14 2- Scheme 3-4: Uncoupling of CO2 evolution from SO3 production………………...110 Scheme 5-1: Synthesis of kinetic mechanism with the chemical mechanism………..168

Scheme 5-2: Synthetic route to 1,1-[F]2-taurine (F-taurine)………………………….169 Scheme 5-3: Synthetic route to N-oxalylglycine (NOG)……………………………..170

Scheme 5-4: Synthetic route to 4-Oxo-4-thiocarboxy-butyric acid (S-αKG)………….171 Figure 1-1: The coordination geometry of the Fe changes upon binding of the substrate……………………………………………………………………………16 Figure 1-2: Jellyroll structural motif…………………………………………………….17 Figure 1-3: Regulation of the HIF………………………………………………………18 Figure 1-4: Repair of 1-methyladenine lesion in DNA mediated by AlkB…………19 Figure 1-5: Clavaminate synthase catalyzes three independent reactions…………… 20 Figure 1-6: SyrB2 catalyzes the specific chlorination of a threonine residue…………..21

Figure 1-7: TauD oxidatively catabolizes the small molecule taurine…….…………….22 Figure 1-8: Stereo specific hydroxylation by prolyl-4-hydroxylase…………………….23 Figure 2-1: Absorption spectra acquired during titration of TauD with Fe(II)………….68 Figure 2-2: Mössbauer spectra of freeze-quench TauD samples………………………..69

Figure 2-3: Absorbance-versus-time traces monitoring binding of αKG……………….71

Figure 2-4: Stopped-flow spectra of the O2 reaction……………………………………72 Figure 2-5: Kinetic traces from the reaction of Figure 2-4……………………………...73 Figure 2-6: Comparison of the concentrations of the first intermediate determined by Mössbauer with simulation………………………………………………………….74 Figure 2-7: Analysis to determine the Mössbauer spectrum of M……...... 75 Figure 2-8: Comparison of the relative amounts of M and the quaternary complex by stopped-flow absorption and Mössbauer……………………………………………77

14 14 Figure 2-9: Kinetics of production of CO2 from 1-[ C]-αKG in single-turnover……78

Figure 2-10: Absorbance-versus-time traces (318 nm) to measure [O2] dependence…79 Figure 2-11: Absorbance-versus-time traces (318 nm) to measure [protein] dependence…………………………………………………………………………...80

Figure 2-12: Absorbance-versus-time traces after mixing at 5 °C of TauD•Fe(II)

complex with O2, αKG and taurine………………………………………………….81 Figure 2-13: Inhibition of (A) regeneration of the quaternary complex by taurine in a

single turnover……………………………………………………………………….82

14 14 Figure 3-1: Kinetics of production of CO2 from 1-[ C]-αKG in single-turnover from

the TauD•Fe(II)•αKG•D-taurine reaction with O2…...... 111 Figure 3-2: Kinetic difference spectra from the reaction of the D-taurine quaternary

complex with O2……………………………………………………………………112 Figure 3-3: Kinetic difference spectra from the reaction of the H-taurine quaternary

complex with O2……………………………………………………………………113

Figure 3-4: Kinetics of, and C1-deuterium KIE, of the quaternary

TauD•Fe(II)•αKG•taurine complex with limiting O2………………………………114 Figure 3-5: Freeze-quench Mössbauer spectra Kinetics of, and C1-deuterium KIE, of

the quaternary TauD•Fe(II)•αKG•taurine complex with limiting O2………………115 Figure 3-6: 4.2 K, weak-field Mössbauer spectra of samples prepared by using the

synthetic D-taurine analog………………………………………………………….117 Figure 3-7: Mössbauer spectra of TauD samples recorded at 4.2 K in an 8-T magnetic

field applied parallel to the γ-beam…………………………………………………118 Figure 3-8: 4.2-K/40-mT Mössbauer spectra of a 20-ms sample recorded (A) before and (B) after 1.4 Mrad of cryoreduction...... ………………………………………..119 Figure 3-9: 77-K X-band EPR spectra of a 20-ms sample recorded before and after cryoreduction……………………………………………………………………….120 Figure 3-10: Effect of solvent viscosity on the steady-state rates of hydroxylation of H- taurine (circles) and D-taurine (squares)………………………...... 121 Figure 4-1: Active site of taurine dioxygenase……………………………………...134

14 Figure 4-2: Transient kinetics of CO2 production from TauD•Fe(II)•αKG………135 Figure 4-3: Kinetic difference UV-visible absorption spectra from the reaction of

TauD•Fe(II)•αKG with O2…………………………………………………………136 Figure 4-4: Kinetic traces from the reaction of Figure 4-3………………………137

Figure 4-5: Absorbance-versus-time traces (318 nm) after mixing ternary complex

against varying concentrations of O2……………………………………………….138 Figure 4-6: Freeze-quench Mössbauer spectra of TauD•Fe(II)•αKG complex after

reaction with O2…………………………………………………………………….139 Figure 5-1: Comparison of the spectra the TauD•Fe(II)•αKG•taurine and either

taurine or 1,1-[F]2-taurine…………………………………………………………..172

14 14 Figure 5-2: Effect of F-taurine on kinetics of CO2 production from 1-[ C]-αKG in single-turnover……………………………………………………………………173

Figure 5-3: UV-visible absorption spectra following the reaction of

TauD•Fe(II)•αKG•F-taurine quaternary complex with O2………………………..174 Figure 5-4: Kinetic traces from the reaction of Figure 5-3…………………………175

Figure 5-5: Absorbance-versus-time traces (318 nm) measuring the [O2] dependence TauD•Fe(II)•αKG•F-taurine quaternary complex…………………………………176 Figure 5-6: Absorbance-versus-time traces (318 nm) measuring the [protien]

dependence TauD•Fe(II)•αKG•F-taurine quaternary complex…………………177 Figure 5-7: Absorbance-vs-time traces for the reduction of a pre-formed population of J by ascorbate………………………………..……………………………………178 Figure 5-8: Mössbauer spectra of the samples from the time dependent reaction of the

TauD•Fe(II)•αKG•F-taurine with O2……………………………………………..179 Figure 5-9: Model of TauD active site with F-taurine bound……………………….181

Figure 5-10: Absorption spectra of titration of TauD with S-αKG…………………..182 Figure 5-11: Titration curves monitoring the development of the 540 nm MLCT band

during the titration of S-αKG………………………………………………………183 Figure 5-12: UV-visible absorption spectra recorded of reaction between TauD•Fe(II)

•S-αKG•taurine and O2…………………………………………………………….184 Figure 5-13: Absorbance-versus-time traces (318 nm) measuring the [O2] dependence

of TauD•Fe(II) •S-αKG•taurine complex…………………………………………185 Figure 5-14: Freeze-quench Mössbauer spectra TauD•Fe(II)•S-αKG•taurine complex

reaction with O2…………………………………………………………………..186 Figure 5-15: Absorbance-versus-time traces (318 nm) following the reaction of various

S-αKG complexes with O2…………………………………………………………188 Figure 5-16: Absorption spectra acquired during titration of TauD with NOG in the

absence of O2……………………………………………………………………….189 Figure 5-17: Absorption spectra acquired during competitive binding titration

converting αKG complex to NOG complex……………………………………..…190

xii

Figure 5-18: U V - v i s i b l e a b s o r p t i o n d a ta recorded during reaction of

TauD•Fe(II)•NOG•taurine with O2-saturated buffer……………………………….191 Figure 5-19: Freeze-quench Mössbauer spectra of TauD•Fe(II)•NOG•taurine complex

after reaction with O2………………………………………………………………192 Figure 5-20: UV-visible absorption spectra (pathlength = 1 cm) recorded after mixing at 5 °C of quaternary complex with buffered peracetic acid……………………….194 Figure 5-21: Absorbance-versus-time traces (318 nm) after mixing

TauD•Fe(II)•αKG•taurine with varying concentrations of peracetic acid…………195 Figure 5-22: UV-visible absorption spectra (pathlength = 1 cm) recorded after mixing at 5 °C of binary complex with buffered peracetic acid……………………………196 Figure 5-23: Freeze-quench Mössbauer spectra of samples from the reaction of TauD quaternary complex with peracetic acid……………………………………………197 Figure 5-24: 4.2 K / 40 mT Mössbauer spectra of the NO-TauD complex recorded

before (A) and after (B) γ-irradiation at 77 K………………………………………198 Figure 5-25: 4.2 K Mössbauer spectra of the TauD {FeNO}8 complex recorded in several magnetic fields……………………………………………………………………..200

1 Figure A-1: H-NMR spectrum (300 MHz, 22°C) in D2O of 2-aminoethanesulfonic acid (taurine) purchased commercially (product of Scheme 3-3 step A)…………..209

13 Figure A-2: C-NMR spectrum (75 MHz, 22°C) in D2O of 2-aminoethanesulfonic acid (taurine) purchased commercially (product of Scheme 3-3 step A)…………..210

1 Figure A-3: H-NMR spectrum (300 MHz, 22°C) in CDCl3 of phenoxysulfonyl-acetic acid phenyl ester (product of Scheme 3-3 step A)………………………………….211

13 Figure A-4: C-NMR spectrum (75 MHz, 22°C) in CDCl3 of phenoxysulfonyl-acetic acid phenyl ester (product of Scheme 3-3 step A)………………………………….212 Figure A-5: High-resolution mass spectrum (electrospray positive ion mode) of phenoxysulfonyl-acetic acid phenyl ester (product of Scheme 3-3 step A)………213

xiii

1 Figure A-6: H-NMR spectrum (300 MHz, 22°C) in CDCl3 of carbamoyl- methanesulfonic acid phenyl ester (Scheme 3-3, product of step B)………………214

13 Figure A-7: C-NMR spectrum (75 MHz, 22°C) in CDCl3 of carbamoyl- methanesulfonic acid phenyl ester (Scheme 3-3, product of step B)………………215 Figure A-8: Infra-red spectrum (on NaCl plate) of carbamoyl-methanesulfonic acid phenyl ester (Scheme 3-3, product of step B)………………………………………216 Figure A-9: Mass spectrum (APCI, positive ion mode) of carbamoyl-methanesulfonic acid phenyl ester (Scheme 3-3, product of step B)…………………………………217

1 Figure A-10: H-NMR spectrum (300 MHz, 22°C) in CDCl3 of 2-aminoethane-1-sulfonic acid phenyl ester (Scheme 3-3, product of step C)…………………………………218 Figure A-11: Expanded view of Figure A-10…………………………………………..219

13 Figure A-12: C-NMR spectrum (75 MHz, 22°C) in CDCl3 of the 2-aminoethane-1- sulfonic acid phenyl ester (Scheme 3-3, product of step C)………………………..220 Figure A-13: Infra-red spectrum (on NaCl plate) of the 2-aminoethane-1-sulfonic acid phenyl ester (Scheme 3-3, product of step C)………………………………………221 Figure A-14: Mass spectrum (APCI, positive ion mode) of 2-aminoethane-1-sulfonic acid phenyl ester (Scheme 3-3, product of step C)………………………………………222

1 2 Figure A-15: H-NMR spectrum (300 MHz, 22°C) in CDCl3 of 1,1-[ H]2-2-aminoethanesulfonic acid phenyl ester (Scheme 3-3, product of step D)……….………...223

13 2 Figure A-16: C-NMR spectrum (75 MHz, 22°C) in CDCl3 of 1,1-[ H]2-2-aminoethanesulfonic acid phenyl ester (Scheme 3-3, product of step D)…………………224

2 Figure A-17: Mass spectrum (APCI, positive ion mode) of 1,1-[ H]2-2-aminoethanesulfonic acid phenyl ester (Scheme 3-3, product of step D)…………………225

1 2 Figure A-18: H-NMR spectrum (300 MHz, 22°C) in D2O of 1,1-[ H]2-2-aminoethanesulfonic acid (Scheme 3-3, product of step E)………………………………226

13 2 Figure A-19: C-NMR spectrum (300 MHz, 22°C) in D2O of 1,1-[ H]2-2-aminoethanesulfonic acid (Scheme 3-3, product of step E)………………………………227

xiv

Figure A-20: Expanded view of Figure A-19…………………………………………228

Figure A-21: 1H-NMR spectrum (300 MHz, 22°C) of the difluoro-phenoxysulfonyl- acetic acid (Scheme 5-2, product of step A)………………………………………..229

Figure A-22: Proton decoupled 13C-NMR spectrum (75 MHz, 22°C) of the difluoro- phenoxysulfonyl-acetic acid (Scheme 5-2, product of step A)……………………..230 Figure A-23: Mass spectrum of the difluoro-phenoxysulfonyl-acetic acid (Scheme 5-2, product of step A) as a mono-anionic dimer………………………………………..231

Figure A-24: 19F spectrum at 75 MHZ, 22°C of the difluoro-phenoxysulfonyl-acetic acid (Scheme 5-2, product of step A)……………………………………………………232

Figure A-25: 1H-NMR spectrum at 300 MHZ 22°C of the carbamoyl difluoro- methanesulfonic acid phenyl ester (Scheme 5-2, product of step B)………………233

13 Figure A-26: Proton decoupled C-NMR spectrum (75 MHz, 22°C, CDCl3) carbamoyl difluoro-methanesulfonic acid phenyl ester (Scheme 5-2, product of step B)……..234 Figure A-27: Mass spectrum of carbamoyl difluoro-methanesulfonic acid phenyl ester (Scheme 5-2, product of step B)……………………………………………………235 Figure A-28: Infrared spectrum of carbamoyl difluoro-methanesulfonic acid phenyl ester (Scheme 5-2, product of step B)……………………………………………………236

Figure A-29: 19F spectrum at 75 MHZ, 22°C of carbamoyl difluoro-methanesulfonic acid phenyl ester (Scheme 5-2, product of step B)………………………………………237

Figure A-30: 1H-NMR spectrum (300 MHz, 22°C) of the 1,1-difluoro-2-aminoethane-1- sulfonic acid phenyl ester (Scheme 5-2, product of step C)………………………..238

Figure A-31: 13C-NMR spectrum (75 MHZ, 22°C) of the 1,1-difluoro-2-aminoethane-1- sulfonic acid phenyl ester (Scheme 5-2, product of step C)………………………..239 Figure A-32: Mass spectrum (APCI, positive mode) of the 1,1-difluoro-2-aminoethane-1- sulfonic acid phenyl ester (Scheme 5-2, product of step C)………………….…….240 Figure A-33: Infrared spectrum of 1,1-difluoro-2-aminoethane-1-sulfonic acid phenyl ester (Scheme 5-2, product of step C)………………………………………………241

Figure A-34: 1H-NMR spectrum (300 MHZ, 22°C) of 1,1-difluoro-2-aminoethane-1- sulfonic acid (Scheme 5-2, product of step D, F-taurine)…………………………..242

Figure A-35: Proton decoupled 13C-NMR spectrum (75 MHz, 22°C) of 1,1-difluoro-2- aminoethane-1-sulfonic acid (Scheme 5-2, product of step D)…………………….243 Figure A-36: Mass spectrum (APCI, positive mode) of the 1,1-difluoro-2-aminoethane-1- sulfonic acid (Scheme 5-2, product of step D)……………………………………..244

Figure A-37: 19F spectrum at 75 MHZ, 22°C of the difluoro-phenoxysulfonyl-acetic acid, F-taurine (Scheme 5-2, product of step D)…………………………………………245

1 Figure A-38: H-NMR spectrum (300 MHZ, 22°C) in D2O of 2-oxopentane-1,5-dioic acid (α-ketoglutarate, αKG)………………………………………………………..246

1 Figure A-39: H-NMR spectrum (300 MHZ, 22°C) in D2O of N-oxalylglycine (Scheme 5-3, product)………………………………………………………………………...247

13 Figure A-40: C-NMR spectrum (75 MHZ, 22°C) in D2O of N-oxalylglycine (Scheme 5-3, product)………………………………………………………………………248 Figure A-41: Mass spectrum (electrospray ionization in negative ion mode) of N- oxalylglycine (Scheme 5-3, product)……………………………………………….249

1 Figure A-42: H-NMR spectrum (300 MHZ, 22°C) in CDCl3 of Boc-Glu(OtBu)-OH (Scheme 5-4, starting material)……………………………………………………..250

1 Figure A-43: H-NMR spectrum (300 MHZ, 22°C) in CDCl3 of Boc-Glu(OtBu)-SH (Scheme 5-4, step A product)………………………………………………………251 Figure A-44: Mass spectrum (electrospray ionization in positive ion mode) of Boc- Glu(OtBu)-SH (Scheme 5-4, step A product)………………………………………252

1 Figure A-45: H-NMR spectrum (300 MHZ, 22°C) in D2O of 4-amino-4-thiocarboxy butyric acid (Scheme 5-4, step B product)………………………………………….253 Figure A-46: Mass spectrum (electrospray ionization in positive ion mode) of 4-amino-4- thiocarboxy butyric acid (Scheme 5-4, step B product)……………………………254

xvi

1 Figure A-47: H-NMR spectrum (300 MHZ, 22°C) in D2O of 4-oxo-4-thiocarboxy butyric acid (Scheme 5-4, step C product)………………………………………….255 Figure A-48: Mass spectrum (electrospray ionization in negative ion mode) of 4-oxo-4- thiocarboxy butyric acid (Scheme 5-4, step C product)……………………………256

xvii

List of Tables

Table 1-1: Representative members of the Fe(II)•αKG dioxygenase family………….13 Table 2-1: Relative amounts of TauD•Fe(II)•αKG•taurine, J, and the second intermediate, determined by Mössbauer spectroscopy……………………………..64 Table 5-1: Mössbauer parameters of {FeNO}7 species……………………………….201

xviii

Acknowledgements My time here has been so rushed, but trying to stop the progression of time is impossible. As much as I would like take the time to sit back and savor the experiences of the past 4 years, my various responsibilities insist that I continue to move. As this epoch draws to a close, I recognize that in my time here I have grown tremendously. This is in large part because of my mentors and peers. Marty, I am in awe of your ability to walk (often run ☺) into a situation totally unprepared and yet have a total command of the scientific principles at work. I strive to emulate your ability to approach a problem scientifically; I believe the grant described you as a laser. Carsten, it has been a pleasure to be your first student (± ½). Watching you and Marty has also taught me about the importance of controlling personal pride and insecurities in the scientific community and in a partnership. Thank you to all the members of the lab, especially Eric, Bhramara, Danny, Lana, Gang, Gretchen, and Lee. My family situation has made me kind of an odd duck in the lab. Rarely taking part in the evening get togethers, yet I have always felt a real spirit of companionship and support. Good luck to the new generation Gretchen, Lee, Ying-Hui, and Wei, you are in good hands. Without a doubt the single most important influence in my life is my wife Jessica. Oh, if I could only attain the same level of competence and confidence in my area of responsibility as you have achieved in yours. Thank you for your kind and loving heart and your example of forgiveness. You have shown me that life is a miracle, more than the sum of its parts. I have seen you work wonders with people whom I had resigned to resent and dislike for the rest of my life. You are the bravest person I know and my greatest hope is that I can help you build the kind of life you truly deserve.

Chapter 1: Introduction

The Fe(II)•αKG dependent dioxygenase family of enzymes perform an amazing variety of reactions including hydroxylations, desaturations, ring formation and ring expansions (see Table 1-1). These enzymic functions result in the ability of biological systems to sense oxygen, repair DNA/RNA, synthesize connective tissue and antibiotics, and catabolize small molecules. The malfunction of members of the family has been identified in connection with Zellweger syndrome (defect in peroxizome assembly), rhizomelic chondrodysplasia punctata (disruption of peroxizome targeting), Refsum disease (inability to digest phytalic acid, a major component of chlorophyll), Ehlers- Danlos syndrome (skin fragility, joint hyper-mobility), scurvy (generalized connective tissue problems including hemorrhages and tooth loss), and various ischemic diseases (Hausinger 2004). The serious nature of these disease states has spurred the intensive study of this enzyme family for more than 30 years.

General Features of the Fe(II)•αKG dependent dioxygenase family. Considering the diverse functions and the variety of enzymatic activities seen within the enzyme family, the generalized reaction is quite simple (see Scheme 1-1). All of these enzymes utilize Fe(II), the co-substrate 2-oxo-pentane-1,5-dioic acid or α-ketoglutarate (αKG), and oxygen (O2). The target carbon is oxidized, resulting in either the hydroxylation or the desaturation of the substrate, concomitant with the decarboxylation of αKG and the breaking of the O2 double bond. Two electrons for the O-O bond cleavage come from αKG, and two come from the substrate. In the hydroxylations both electrons come from the same carbon. Desaturation requires that one electron comes from each of the carbons

in the double bond. In all cases, one atom from O2 is incorporated into succinate, while the other may be transferred to substrate or reduced to water, depending on the enzymatic activity (Cardinale et al. 1971; Baldwin et al. 1991). The active site of these enzymes consists of a single Fe(II) atom coordinated facially by three amino acids, two histidine 2

ligands and an aspartic acid (Figure 1-1), leaving three Fe coordination sites available for substrate binding. The ligands are provided by an HX(D/E)XnH sequence, where X can be any amino acid. This motif is highly conserved throughout the family (Aravind et al. 2001). The Fe(II) binding site is buried within a “jellyroll” structural motif. This structure is a hallmark of the cupin superfamily and consists of two, four-stranded, β sheets (Schofield et al. 2004). The active site is positioned within the two β sheets, which cup the active site like the palms of a pair of hands (see Figure 1-2). In the

Fe(II)•αKG dioxygenases the HXD/EXnH is in or near the second and seventh strands of the cupin fold. Another common feature among many of the Fe(II)•αKG-dependent dioxygenases includes a large section of random coil which caps the active site. This cap is mobile isolating the active site from the environment after the substrate has bound (O'Brien et al. 2003). Although the general mechanism and active site structure are very similar throughout the family the large variation in size and morphology of the substrates necessitates major changes in the binding site, surface of the “jellyroll”, and on the active site cap. As mentioned, the presence of one (or more) jellyroll structural motif(s) is a defining feature of the cupin superfamily. Within this superfamily, there is a large group of transcription factors known as the jumonji proteins. Many of these jumonji

transcription factors posses both the “jelly roll” and the HXD/EXnH metal binding sequence, leading to the hypothesis that many of these transcription factors are members of the Fe(II)•αKG-dependent dioxygenase family (Schofield et al. 2004).

HIFα hydroxylases. The HIFα hydroxylases are examples of how transcription can be modulated using the generalities of structure and chemistry found in the

Fe(II)•αKG-dependent dioxygenase family. Their function is to regulate the hypoxic response in mammals. The hypoxia inducible factor (HIF) is a transcriptional activator involved in initiating the genetic response to low oxygen concentration in the body. It plays a role in inducing the transcription of a variety of genes to ameliorate the hypoxic 3

conditions, processes such as vascular growth and glucose metabolism (Ivan et al. 2001; Jaakkola et al. 2001). HIF is constantly being synthesized, and under normal oxygen

concentrations its α subunit is immediately targeted by two independent types of Fe(II)•αKG dioxygenases. HIFα prolyl-4-hydroxylase targets a specific proline residue (Pro564), hydroxylating it and thereby promoting proteolysis that is also dependent on von Hipple-Lindau tumor suppressor protein (Ivan et al. 2001; Jaakkola et al. 2001).

HIFα asparaginyl hydroxylase hydroxylates a specific asparagine residue (Asn851). Thereby blocking the association of p300, which acts as a co-activator with HIF (Lando et al. 2002). This step inactivates the transcriptional activation complex (see Figure 1-3).

It is a beautifully dynamic arrangement in which [O2] directly controls the genetic response in an array of biological functions. The constant flux of biosynthesis versus catabolism for this single transcription factor is energy intensive, yet ensures a fast and metered response to the O2 levels in the cell. AlkB. Another member of the Fe(II)•αKG dioxygenase family, AlkB, plays a different yet similarly critical role in maintaining the integrity of the genetic code in a host of organisms from bacteria to humans. DNA repair is a process that is integral to the survival of an organism. Alkylating agents, such as methyl-methanesulfonate, create lesions in DNA which can have mutagenic and even lethal effects. AlkB is the second gene product from an operon that was found to rescue E. coli from agents that alkylate the DNA (Kataoka et al. 1983). AlkB was shown to be specific for N-alkylated bases. Expression of the human hABH gene, an AlkB homologue, in E. coli mutants sensitive to alkylating agents was shown to protect the bacteria from methyl-methanesulfonate. This suggests that the proteins are structurally and functionally conserved from bacteria to human (Wei et al. 1996). Sequence analysis suggested that AlkB contains a cupin fold and a HX(D/E)XnH metal binding sequence and may belong to the Fe(II)•αKG dioxygenase family (Aravind et al. 2001). Work with the purified protein confirmed that

AlkB uses the same general chemical scheme seen throughout the Fe(II)•aKG 4

dioxygenase family (see Figure 1-4), to repair N-methylated DNA. The oxidative

decarboxylation of αKG was shown to de-methylate DNA concurrent with the formation of formaldehyde (Trewick et al. 2002). N-alkylated bases within RNA, and single- or double-stranded DNA can serve as a substrate for AlkB. Although the exact search and repair mechanism is unknown, chemical crosslinking studies suggest that the enzyme scans along the DNA for N-methylated bases. These damaged bases are less effective at forming hydrogen-bond pairs with the opposite strand and can be brought into the AlkB active site for demethylation (Mishina et al. 2003). This repair mechanism is quite novel. It is catalytic and efficient, avoiding the previously characterized methods of DNA repair which involved the excision of the damaged base and the re-synthesis of that portion of the DNA. Clavaminate synthase. Clavaminate synthase is one of the most versatile

members of the Fe(II)•αKG dioxygenase family. Streptomyces clavuligerus uses clavaminate synthase in the production of an important β-lactamase inhibitor, clavulanic acid. Amazingly, three steps, first a hydroxylation, then a cyclization, and finally a desaturation during the synthesis of clavulanic acid are all catalyzed by the same enzyme, clavaminate synthase (see Figure 1-5). Each of these steps requires that the Fe site rebind another molecule of αKG and the substrate, which, for the final step is the product of the previous reaction (Busby et al. 1995). In each step, the O-O bond of O2 is broken and one oxygen atom is incorporated into succinate. The first reaction incorporates the second oxygen atom derived from O2 into the product. The second and third reactions result in this second oxygen atom being reduced to water (Busby et al. 1996). It is unknown how the enzyme is able to catalyze such mechanistically different yet highly specific reactions. It was experimentally confirmed that the ligands to the Fe center remain the same in each of the reactions (Khaleeli et al. 2000), leading to the belief that the changes are influenced by the substrate itself. The facial triad allows an amazing versatility in the reaction mechanism, while maintaining enzyme specificity. 5

SyrB2. A further display of chemical versatility is seen in the reaction of another

Fe(II)•αKG dioxygenase, SyrB2 from pseudomonas syringae. This bacterium synthesizes the phytotoxin syringomycin E from components that include a modified threonine residue (Guenzi et al. 1998). A key modification in the synthesis is the specific chlorination of the threonine C4 position (see Figure 1-6). The substrate threonine is covalently attached to another protein, SyrB1. SyrB2 transiently docks with SyrB1, modifying the threonine, and then releasing it. Upon rebinding of αKG, SyrB2 is competent to chlorinate again (Vaillancourt et al. 2005). The surprising ability to catalytically incorporate chlorine requires that the chlorine be positioned to rebound onto the substrate after hydrogen atom abstraction has taken place (vide infra). Even more surprising is that no hydroxylation product was detected in the limiting chloride experiment. The crystal structure reveals that chloride is directly coordinated to the Fe center (C. L. Drennon unpublished data). This occurs due to a modification of the Fe chelating facial triad of amino acids. The carboxylate residue has been changed to an alanine. This creates an open coordination site and the necessity for a supplementary negative charge to maintain the charge balance within the active site. Binding of chloride in the open coordination site maintains the charge balance in the active site and presumably positions the chloride for transfer to the substrate. The question is still open though as to how the protein is able to differentiate between chlorination and hydroxylation to quench the substrate radical (vide infra). This is a striking example of how innovative nature can be in utilizing the general features of the Fe(II)•αKG dioxygenase family.

Prolyl-4-hydroxylase The study of the Fe(II)•αKG dioxygenase family really began in the early 1960’s with the investigation of collagen biosynthesis. Huge advances were made in the understanding of collagen as a molecule, and were followed closely by the biochemical characterization of the enzyme prolyl-4-hydroxylase (P4H). To a large degree, the mechanical strength and elasticity of the collagen triple helix (and hence most 6

biological connective tissue) is due to the hydrogen bonding interactions of a post- translationally modified proline residue within a –XPG- (where X can be any amino acid) peptide recognition sequence. P4H is responsible for the specific C4 hydroxylation of

proline (see Figure 1-3) and in vertebrates is active as an α2β2 tetramer (~ 240 KD) in the lumen of the endoplasmic reticulum (Myllyharju 2003). Using a cell free homogenate

Peterkofsky and co-workers showed the necessity for atmospheric O2 and the substrate specificity of P4H for proline within the collagen peptide, rather than free proline in solution (Peterkofsky et al. 1963). Studies with these cell-free extracts also showed that the hydroxylation mechanism was stereo-specific, always hydroxylating C4 trans (on the opposite side of the ring) to the C1 carboxylate (see Figure 1-8). The reaction was shown to be energetic enough to break a C-F bond, yet could be inhibited by the electron withdrawing effects of a cis fluorine on the proline C4 carbon (Gottlieb et al. 1965). Subsequent studies on the purified enzyme showed that, although the high energy oxidation of proline is reminiscent of the reactions catalyzed by the cytochromes P450, P4H contains no heme. They also established that P4H shows an absolute requirement

for Fe(II), αKG, and O2 for activity (Hutton et al. 1966; Kivirikko et al. 1967). A study by Rhoads and Udenfriend demonstrated that the substrate hydroxylation requires the

decarboxylation of αKG (Rhoads et al. 1968). They also noted that CO2 is evolved even in the absence of substrate (uncoupling of the reaction). Cardindale and co-workers showed further that one atom of oxygen from the gaseous O2 substrate is incorporated into succinate (the product of αKG decarboxylation), while the other is found in the new C4 hydroxyl group of the modified proline (Cardinale et al. 1971). Later Tuderman and co-workers demonstrated that while an exogenous reductant like ascorbate is necessary for long-term activity, P4H can turn over as many as 30 times with out any external reductants (Tuderman et al. 1977). They showed further that the use of small superoxide scavengers, such as nitroblue tetrazolium, can competitively inhibit the production of 7

hydroxyl-proline. These key experiments, along with a host of others, spurred the proposal of a legion of possible mechanisms.

Consensus mechanism for the Fe(II)•αKG dioxygenase family. In the early 1980’s, Hanauske-Abel and Günzler proposed a mechanism that accommodated all of the data in the literature. This mechanism employed strong chemical theory to propose plausible intermediates that rationalize the chemistry, but for which there was no direct evidence (Hanauske-Abel et al. 1982). Knowing from the literature that decarboxylation

and cleavage of the O2 bond occur prior to hydroxylation of substrate, their proposed mechanism started with a 6-coordinate intermediate in which αKG chelates the metal in a bidentate fashion and oxygen coordinates to the metal to form an end-on Fe(III)- superoxide complex (see Scheme 1-2 species I). The absence of a heme also required that the Fe center be bound by multiple unidentified protein residues. They reasoned that a bi-dentate binding mode for αKG decreased the electron density of the C2 carbonyl

carbon, making it an ideal target for SN2 type nucleophilic attack. The second intermediate (Scheme 1-2, species II), a bicyclic species formed after attack by the coordinated superoxide, decays upon decarboxylation of αKG. This decarboxylation yields two electrons, which are donated into the anti-bonding orbital of the of the O-O bond. These reducing equivalents break the O-O bond and allow the formation of succinate, all prior to the hydroxylation of substrate, as required by the literature. The

breaking of the O2 bond requires 4 electrons, so the Fe remains in a highly oxidized Fe(IV) state after decarboxylation. In the absence of substrate an exogenous reductant could regenerate the resting Fe(II) state, as required by the literature. They hypothesized that, in the presence of substrate, the ferryl (Fe(IV)=O) center (Scheme 1-2, species III) would oxidize the C4 carbon of the target proline via homolytic cleavage of the C4 carbon-hydrogen bond, transferring a hydrogen atom to the oxygen. This would result in the formation of a substrate radical (Scheme 1-2, species IV) which could rapidly recombine with the Fe-OH to reduce the Fe to Fe(II) and give the final product. 8

Although no data within the P4H literature supported the formation of these intermediates, the mechanism explained the accumulated experimental findings. The authors also cited extensive precedent for the individual steps. Similar reactivity had been seen in other

systems, such as the observation that O2 species catalyze the decarboxylation of αKG (Jefford et al. 1976; San Filippo et al. 1976), and the extensive literature on the cytochrome P450 enzymes, which employ a ferryl species in the hydroxylation of unactivated carbon atoms (Chang et al. 1979; Gustafsson et al. 1979). The basic tenets of the Hanauske-Abel proposal have been adopted as a consensus mechanism for all members of the family (Solomon et al. 2000; Costas et al. 2004; Hausinger 2004). The substrate binding steps leading to the first transient intermediate of the consensus mechanism have been investigated experimentally. The binary Fe•enzyme complex has an octahedral Fe center, with HX(D/E)XnH motif binding the Fe as a facial triad of amino acids and the remaining coordination sites filled by waters (Pavel et al. 1998; Vålegard et al. 1998). Extensive evidence supports the bi-dentate chelating mode of the αKG (see Figure 1-1). Model complexes and experiments with high concentrations of purified enzymes (Pavel et al. 1998; Ho et al. 2001) have shown that

the interaction between Fe(II) and αKG elicits the formation of a weak metal-to-ligand charge-transfer band (MLCT) between 400 and 600 nm, giving these ternary

Fe•enzyme•αKG complexes a lavender color. Magnetic circular diochroism (MCD) (Pavel et al. 1998), resonance Raman (Ho et al. 2001), and X-ray crystallography (Zhang et al. 2000; Vålegard et al. 2004) experiments have shown that this complex is a 6-

coordinate Fe center. The His2carboxylate amino acid motif is again facially coordinating, with αKG in the equatorial plane and an axial water molecule (see Figure 1-1). Addition of the target substrate elicits a subtle change at the Fe site. The substrate does not coordinate directly to the metal, but the formation of this

Fe•enzyme•αKG•substrate quaternary complex causes the coordinating water molecule to be expelled, forming a 5-coordinate square pyramidal Fe site. This complex has also 9

been extensively investigated and MCD (Zhou et al. 1998), resonance Raman (Ho et al. 2001), and crystallography (Zhang et al. 2000; O'Brien et al. 2003) support the conclusion that this complex now has an open coordination site on Fe, where presumably,

the final reactant, O2, can bind to form the first intermediate of the Hanauske- Abel/Günzler mechanism. Studies on P4H (Wu et al. 1999), and claviminate synthase (Baldwin et al. 1991), offered support for the proposed intermediate state, species IV. Both groups used specially designed substrates that could indicate whether a ligand based radical forms during the catalytic cycle. An in depth biochemical study of the catalytic cycle requires large amounts of protein, and saturating concentrations of substrate. Many of the systems discussed during this introductory chapter present technical difficulties because of complexity required to prepare the reactive complex. Mammalian proteins are notoriously difficult to express and purify in large quantities. Several of these systems also require substrates that are large and/or very difficult to obtain, and/or unstable.

Taurine α-ketoglutarate dioxygenase. Taurine α-ketoglutarate dioxygenase (TauD) displays none of these challenges to detailed mechanistic characterization. It is an endogenous E. coli protein, used in the catabolism of a common, and simple mammalian waste product, 2-aminoethanesulfonic acid (taurine), and is expressed in response to sulfur starvation (Eichhorn et al. 1997). This enzyme displays what may be described as the classic Fe(II)•αKG dioxygenase activity (see Figure 1-7). The His2Asp coordinated Fe center is buried within the cupin fold. The bidentate binding of αKG to the Fe elicits the MLCT band with maximal absorbance at 530 nm (ε = 180 M-1cm-1). Binding of taurine induces the formation of a 5-coordinate square pyramidal complex (Ho et al. 2001), resulting in a diagnostic 10 nm blue shift of the MLCT band to 520 nm

(ε = 200 M-1cm-1). Multiple biochemical studies have been carried out on TauD, facilitated by the large quantities of protein that an E. coli based expression system can provide. This has allowed Hausinger and co-workers to investigate the mechanism of the 10

O2 reaction. Their studies confirmed that the 520 nm absorbance band decayed upon

addition of O2, indicating that the TauD•Fe(II)•αKG•taurine complex was destroyed by decarboxylation of αKG (Ryle et al. 1999). Oddly, they did not see the reappearance of the 520 nm absorbance after the reaction, and had no hypothesis as to why this complex

never reformed. It may indicate that the products were competing for αKG binding sites under the conditions of their experiment.

TauD was selected as an ideal member of the Fe(II)•αKG dioxygenase family for mechanistic studies. As an endogenous E. coli protein it is easily produced in large quantities. The studies presented in the subsequent chapters focus wholly upon the TauD system. The mechanistic insights provided by these studies are likely to be generally

applicable to the entire Fe(II)•αKG dioxygenase family, although this notion remains to be experimentally verified. The results of subsequent chapters can be briefly summarized. Chapter 2 entails

the careful kinetic dissection of the reaction of the TauD•Fe(II)•αKG•taurine quaternary

complex with O2. Under single turnover conditions this reaction can be described by a three-step minimal kinetic mechanism. This mechanism assumes (because of the experimental conditions) that a preformed quaternary complex (i.e. that the substrates taurine and αKG are bound) is sitting in an anaerobic environment, waiting for binding of

O2. Addition of O2 initiates the reaction, and we observe the formation of a transient oxidized Fe species. This species (given the acronym J) forms with a true second order

rate constant (i.e. first order in [O2] and [quaternary complex]), and decays to a second intermediate M. The formation of the quaternary complex is studied under transient state conditions and the rate limiting step of the reaction is assigned as the dissociation of products. In this chapter, M is firmly established as a high spin Fe(II) product complex of ambiguous composition. Chapter 2, does not state as definitively the identity of J. The chemical nature and oxidation state of J are not well defined within the context of chapter 2. Therefore, chapter 2 also discusses the various possible identities of J, within 11

the consensus mechanism, and some of the literature references that make these assignments plausible. Chapter 3 presents the careful chemical and spectroscopic characterization of J. Here the deuterium isotopomer of taurine is used to show that J is the hydroxylating species and confirm that M is a product complex. High field Mössbauer studies show that J is a high spin, S = 2, Fe(IV) species. Resonance Raman by the group of Robert Hausinger and XAS studies by our group are discussed which indicate that J is a Fe(IV)=O species.

Chapter 4 investigates the decoupling of CO2 production and product production. Since the early studies with P4H, it has been observed that these enzymes will produce

CO2 (albeit at a slower rate) without the presence of substrate. Low level decoupling of

2- 2 CO2 and SO3 production was also detected in the reaction of the 1,1-[ H]2-taurine containing quaternary complex with O2. Production of CO2 indicates that a J analog is formed within the active site. The uncontrolled production of reactive intermediates can be deleterious to the enzyme. Chapter 4 discusses one of the strategies TauD and other members of the family employ to quench these oxidizing intermediates. Hausinger and co-workers showed that specific aromatic amino acids in the active site become hydroxylated and activity can be rescued with the return of Fe(II) to the active site. Chapter 5 contains the preliminary results of studies involving a number of substrate analogues. These analogues were designed to target specific intermediates in the catalytic cycle. None of them is as successful in their design as the deuterium substituted taurine analog discussed in chapter 3. However, these studies do offer interesting clues about the balance of forces within the active site, and are documented here to provide a basis for the design of more successful analogues in the future. All components of the reaction have been altered except for the protein itself. Studies involving protein variants are being carried out by other students within the Bollinger/Krebs group. 12

In summary this thesis describes the most in-depth mechanistic study to date of an

Fe(II)•αKG dioxygenase. Two accumulating intermediates are characterized and placed within the catalytic cycle. Strong evidence suggests that the basic elements of this study

may be applied broadly to the entire Fe(II)•αKG dioxygenase family.

Enzyme Biological function: Chemical activity: Reference:

prolyl-4-hydroxylase collagen/connective tissue synthesis hydroxylation (Myllyharju 2003)

HIFα prolyl-4-hydroxylase oxygen sensing hydroxylation (Ivan et al. 2001)

AlkB dealkylation of DNA/RNA hydroxylation / elimination (Chen et al. 1994)

clavaminate synthase antibiotic synthesis ring formation / desaturation (Lloyd et al. 1999)

SyrB2 antibiotic synthesis chlorination (Vaillancourt et al. 2005) taurine:α-ketoglutarate taurine catabolism hydroxylation / elimination (Eichhorn et al. 1997) dioxygenase

Table 1-1: Representative members of the Fe(II)•αKG dioxygenase family

-O O- O- R H R OH -O ++ ++ O2 O O CO2 O `R H `R H O O

Scheme 1-1: Generalized reaction of the Fe(II)•αKG dioxygenase family of enzymes

OH OH R 2 - 2 - OOC OOC HO2 2+ L1 + α-KG O L1 + substrate L Fe Fe2+ O 2+ 1 HO L L Fe 2 2 O O 2 O O L2 - 2 H2 O - 1 H2 O L L 3 3 L3 binary complex ternary complex quaternary complex

+ O - products 2

R` R` R HO -O R - - OOC OOC . O 2+ L1 O Fe O 3+ L1 L2 Fe O O O L2 CO2 L3 L3 V I

R` R` R`

. R R -O R OH - - - - O OOC O OOC L OOC 3+ 1 4+ L1 O 4+ L1 O Fe O Fe Fe L L O O L O 2 O 2 2 CO2 CO2 L3 L3 L3 IV III II

Scheme 1-2: Generalized concensus mechanism of the Fe(II)•αKG dioxygenase family, based on the intermediate states proposed by Hanauske-Abel and Günzler (Hanauske-

Abel et al. 1982). 16

Figure 1-1: The coordination geometry of the Fe changes upon binding of the substrate.

Initially the Fe is 6-coordinate with a bound water molecule (left). Binding of N-α- acetyl-L-arginine “triggers” the formation of a 5-coordinate Fe site (right). Structures are adapted from Clavaminate synthase (CAS) structures deposited to the protein data base (1DS1 (left) and 1DRY (right)) both from the work of Zhang and co-workers (Zhang et al. 2000). Analogous active sites are found in all members of the family.

Figure 1-2: Jellyroll structural motif holds the Fe binding motif within a protective pocket. Structure adapted from TauD OTJ1 (O'Brien et al. 2003).

Figure 1-3: Regulation of the HIF by the Fe(II):αKG dependent prolyl and asparaginyl hydroxylases. Taken with permission from Schofield and Ratcliffe Nature Reviews Mol. Cell Biol. 2004, 5, 343-354.

+ - NH3 COO

N CH3 N 7 5 1 O 9 3 ++2 N N -OOC O R

AlkB. Fe(II)

+ COO- NH3

N N 7 5 1 CH2 ++CO2 + 9 3 - O N N O O R

Figure 1-4: Repair of 1-methyladenine lesion in DNA mediated by AlkB. AlkB has been shown to reverse damage due to any 1-alkylpurine and 3-alkylpyrimidine in vivo (Delaney et al. 2004). 20

Figure 1-5: Clavaminate synthase catalyzes three independent reactions during the

synthesis of clavulinic acid. Each step requires O2 and αKG, yet the reaction of substrate A results in hydroxylation, B a cyclization, while C is desaturated. The

arginine side chain of B is converted to a primary amine in B1 through the action of a separate enzyme. The red oxygen atom shows the final destination of the O atom from the O-O bond cleavage in each step. Modifications due to clavaminate synthase are highlighted in yellow.

HO COO-

+ O O - H3N ++2 + Cl

-OOC O S SyrB1

SyrB2. Fe(II) Cl COO- HO

+ CO2 + + H O + O 2 H3N -O O S SyrB1

Figure 1-6: SyrB2 catalyzes the specific chlorination of a threonine residue, prior to the incorporation of this residue into the phytotoxin syringomycin E.

O COO-

-OOS

++O2

-OOC O + NH3

O TauD . Fe(II) S COO- HO O- O

- + OOS

+ CO2 + O OH -O O + NH3 + NH3

Figure 1-7: TauD oxidatively catabolizes the small molecule 2-aminoethanesulfonic acid (taurine). Hydroxylation of the C1 carbon liberates sulfite from the organosulfonate which can be metabolized as a sulfur source. 23

1 R - O COO

2 3 RN ++O2 5 4 -OOC O

P4H . Fe(II)

1 R - O COO

2 3 R N ++CO2 5 4 -O O OH

Figure 1-8: Stereo specific hydroxylation of a collagen proline residue by prolyl-4-

hydroxylase

References

Aravind, L. and Koonin, E. V. (2001). "The DNA-repair protein AlkB, EGL-9, and

leprecan define new families of 2-oxoglutarate- and iron-dependent

dioxygenases." Genome biol. 2(3): RESEARCH0007.

Baldwin, J. E., Adlington, R. M., Crouch, N. P., Keeping, J. W., Leppard, S. W., Pitlik, J.,

Schofield, C. J., Sobey, W. J. and Wood, M. E. (1991). "Studies on the

mechanism of deacetoxy-deacetylcephalosporin C synthase using cyclopropyl

substituted cephalosporin probes." J. Chem. Soc. Chem. Commun.(11): 768-70.

Baldwin, J. E., Adlington, R. M., Crouch, N. P., Schofield, C. J., Turner, N. J. and Aplin,

R. T. (1991). "Cephalosporin biosynthesis: a branched pathway sensitive to an

isotope effect." Tetrahedron 47(47): 9881-9900.

Busby, R. W., Chang, M. D. T., Busby, R. C., Wimp, J. and Townsend, C. A. (1995).

"Expression and purification of two isoenzymes of clavaminate synthase and

initial characterization of the iron binding site. General error analysis in

polymerase chain reaction amplification." J. Biol. Chem. 270(9): 4262-9.

Busby, R. W. and Townsend, C. A. (1996). "A single monomeric iron center in

clavaminate synthase catalyzes three nonsuccessive oxidative transformations."

Bioorg. & Med. Chem. 4(7): 1059-1064.

Cardinale, G. J., Rhoads, R. E. and Udenfriend, S. (1971). "Simultaneous incorporation

of oxygen-18 into succinate and hydroxyproline catalyzed by collagen proline

hydroxylase." Biochem. and Biophys. Res. Comm. 43(3): 537-43. 25

Chang, C. K. and Kuo, M.-S. (1979). "Reaction of iron(III) porphyrins and iodosoxylene.

The active oxene complex of cytochrome P-450." J. Am. Chem. Soc. 101(12):

3413-15.

Chen, B. J., Carroll, P. and Samson, L. (1994). "The Escherichia coli AlkB protein

protects human cells against alkylation-induced toxicity." J. Bacteriol. 176(20):

6255-61.

Costas, M., Mehn, M. P., Jensen, M. P. and Que, L., Jr. (2004). "Dioxygen activation at

mononuclear nonheme iron active sites: enzymes, models, and intermediates."

Chem. Rev. 104(2): 939-986.

Delaney, J. C. and Essigmann, J. M. (2004). "Mutagenesis, genotoxicity, and repair of 1-

methyladenine, 3-alkylcytosines, 1-methylguanine, and 3-methylthymine in alkB

Escherichia coli." Proc. Nat. Acad. Sci. U.S.A. 101(39): 14051-14056.

Eichhorn, E., Van Der Ploeg, J. R., Kertesz, M. A. and Leisinger, T. (1997).

"Characterization of α-ketoglutarate-dependent taurine dioxygenase from

Escherichia coli." J. Biol. Chem. 272(37): 23031-23036.

Gottlieb, A. A., Fujita, Y., Udenfriend, S. and Witkop, B. (1965). "Incorporation of cis-

and trans-4-fluoro-L-prolines into proteins and hydroxylation of the trans isomer

during collagen biosynthesis." Biochemistry 4(11): 2507-13.

Guenzi, E., Galli, G., Grgurina, I., Gross, D. C. and Grandi, G. (1998). "Characterization

of the syringomycin synthetase gene cluster. A link between prokaryotic and

eukaryotic peptide synthetases." J. Biol. Chem. 273(49): 32857-32863. 26

Gustafsson, J. A., Rondahl, L. and Bergman, J. (1979). "Iodosylbenzene derivatives as

oxygen donors in cytochrome P-450 catalyzed steroid hydroxylations."

Biochemistry 18(5): 865-70.

Hanauske-Abel, H. M. and Günzler, V. (1982). "A stereochemical concept for the

catalytic mechanism of prolylhydroxylase. Applicability to classification and

design of inhibitors." J. Theor. Biol. 94(2): 421-55.

Hausinger, R. P. (2004). "Fe(II)/α-ketoglutarate-dependent hydroxylases and related

enzymes." Crit. Rev. Biochem. Mol. Biol. 39(1): 21-68.

Ho, R. Y. N., Mehn, M. P., Hegg, E. L., Liu, A., Ryle, M. J., Hausinger, R. P. and Que,

L., Jr. (2001). "Resonance Raman studies of the iron(II)-alpha-keto acid

chromophore in model and enzyme complexes." J. Am. Chem. Soc. 123(21):

5022-5029.

Hutton, J. J., Jr., Tappel, A. L. and Udenfriend, S. (1966). "Requirements for a-

ketoglutarate, ferrous ion, and ascorbate by collagen proline hydroxylase."

Biochem. and Biophys. Res. Comm. 24(2): 179-84.

Ivan, M., Kondo, K., Yang, H., Kim, W., Valiando, J., Ohh, M., Salic, A., Asara, J. M.,

Lane, W. S. and Kaelin, W. G., Jr. (2001). "HIFα targeted for VHL-mediated

destruction by proline hydroxylation: Implications for O2 sensing." Science

292(5516): 464-468.

Jaakkola, P., Mole, D. R., Tian, Y.-M., Wilson, M. I., Gielbert, J., Gaskell, S. J., von

Kriegsheim, A., Hebestreit, H. F., Mukherji, M., Schofield, C. J., Maxwell, P. H.,

Pugh, C. W. and Ratcliffe, P. J. (2001). "Targeting of HIFα to the von Hippel- 27

Lindau ubiquitylation complex by O2-regulated prolyl hydroxylation." Science

292(5516): 468-472.

Jefford, C. W., Boschung, A. F., Bolsman, T. A., Moriatry, R. M. and Melnick, B. (1976).

"Reaction of singlet oxygen with alpha-ketocarboxylic acids. Oxidative

decarboxylation and peroxyacid formation." J. Am. Chem. Soc. 98(4): 1017-8.

Kataoka, H., Yamamoto, Y. and Sekiguchi, M. (1983). "A new gene (alkB) of

Escherichia coli that controls sensitivity to methyl methanesulfonate." J. Bacteriol.

153(3): 1301-7.

Khaleeli, N., Busby, R. W. and Townsend, C. A. (2000). "Site-directed mutagenesis and

biochemical analysis of the endogenous ligands in the ferrous active site of

clavaminate synthase. The His-3 variant of the 2-His-1-carboxylate model."

Biochemistry 39(29): 8666-8673.

Kivirikko, K. I. and Prockop, D. J. (1967). "Enzymic hydroxylation of proline and lysine

in protocollagen." Proc. Nat. Acad. Sci. U.S.A. 57(3): 782-9.

Lando, D., Peet, D. J., Whelan, D. A., Gorman, J. J. and Whitelaw, M. L. (2002).

"Asparagine hydroxylation of the HIF transactivation domain: A hypoxic switch."

Science 295(5556): 858-861.

Lloyd, M. D., Merritt, K. D., Lee, V., Sewell, T. J., Wha-Son, B., Baldwin, J. E.,

Schofield, C. J., Elson, S. W., Baggaley, K. H. and Nicholson, N. H. (1999).

"Product-substrate engineering by bacteria: studies on clavaminate synthase, a

trifunctional dioxygenase." Tetrahedron 55(33): 10201-10220. 28

Mishina, Y. and He, C. (2003). "Probing the Structure and Function of the Escherichia

coli DNA Alkylation Repair AlkB Protein through Chemical Cross-Linking." J.

Am. Chem. Soc. 125(29): 8730-8731.

Myllyharju, J. (2003). "Prolyl 4-hydroxylases, the key enzymes of collagen

biosynthesis." Matrix Biology 22(1): 15-24.

O'Brien, J. R., Schuller, D. J., Yang, V. S., Dillard, B. D. and Lanzilotta, W. N. (2003).

"Substrate-induced conformational changes in Escherichia coli taurine/α-

Ketoglutarate dioxygenase and insight into the oligomeric structure."

Biochemistry 42(19): 5547-5554.

Pavel, E. G., Zhou, J., Busby, R. W., Gunsior, M., Townsend, C. A. and Solomon, E. I.

(1998). "Circular Dichroism and Magnetic Circular Dichroism Spectroscopic

Studies of the Non-Heme Ferrous Active Site in Clavaminate Synthase and Its

Interaction with α−Ketoglutarate Cosubstrate." J. Am. Chem. Soc. 120(4): 743-

753.

Peterkofsky, B. and Udenfriend, S. (1963). "Conversion of proline to collagen

hydroxyproline in a cell-free system from chick embryo." J. Biol. Chem. 238(12):

3966-77.

Rhoads, R. E. and Udenfriend, S. (1968). "Decarboxylation of α-ketoglutarate coupled to

collagen proline hydroxylase." Proc. Nat. Acad. Sci. U.S.A. 60(4): 1473-8.

Ryle, M. J., Padmakumar, R. and Hausinger, R. P. (1999). "Stopped-flow kinetic analysis

of Escherichia coli taurine/α-ketoglutarate dioxygenase: Interactions with α-

ketoglutarate, taurine, and oxygen." Biochemistry 38(46): 15278-15286. 29

San Filippo, J., Jr., Chern, C.-I. and Valentine, J. S. (1976). "Oxidative cleavage of a-keto,

a-hydroxy, and a-halo ketones, esters, and carboxylic acids by superoxide." J. Org.

Chem. 41(6): 1077-8.

Schofield, C. J. and Ratcliffe, P. J. (2004). "Oxygen sensing by HIF hydroxylases." Nat.

Rev. Mol. Cell Biol. 5(5): 343-354.

Solomon, E. I., Brunold, T. C., Davis, M. I., Kemsley, J. N., Lee, S. K., Lehnert, N.,

Neese, F., Skulan, A. J., Yang, Y.-S. and Zhou, J. (2000). "Geometric and

electronic structure/function correlations in non-heme iron enzymes." Chem. Rev.

100: 235-349.

Trewick, S. C., Henshaw, T. F., Hausinger, R. P., Lindahl, T. and Sedgwick, B. (2002).

"Oxidative demethylation by Escherichia coli AlkB directly reverts DNA base

damage." Nature 419(6903): 174-178.

Tuderman, L., Myllyla, R. and Kivirikko, K. I. (1977). "Mechanism of the prolyl

hydroxylase reaction. 1. Role of co-substrates." Euro. J. Biochem. 80(2): 341-8.

Vaillancourt, F. H., Yin, J. and Walsh, C. T. (2005). "SyrB2 in syringomycin E

biosynthesis is a nonheme FeII a-ketoglutarate- and O2-dependent halogenase."

Proc. Nat. Acad. Sci. U.S.A. 102(29): 10111-10116.

Vålegard, K., Terwisscha van Scheltinga, A. C., Dubus, A., Ranghino, G., Oester, L. M.,

Hajdu, J. and Andersson, I. (2004). "The structural basis of cephalosporin

formation in a mononuclear ferrous enzyme." Nature Struct. & Mol. Biol. 11(1):

95-101.

Vålegard, K., van Scheltinga, A. C., Lloyd, M. D., Hara, T., Ramaswamy, S., Perrakis, A.,

Thompson, A., Lee, H. J., Baldwin, J. E., Schofield, C. J., Hajdu, J. and 30

Andersson, I. (1998). "Structure of a cephalosporin synthase." Nature 394(6695):

805-9.

Wei, Y.-F., Carter, K. C., Wang, R.-P. and Shell, B. K. (1996). "Molecular cloning and

functional analysis of a human cDNA encoding an Escherichia coli AlkB

homolog, a protein involved in DNA alkylation damage repair." Nucl. Ac. Res.

24(5): 931-7.

Wu, M., Moon, H.-S., Begley, T. P., Myllyharju, J. and Kivirikko, K. I. (1999).

"Mechanism-Based Inactivation of the Human Prolyl-4-hydroxylase by 5-

Oxaproline-Containing Peptides: Evidence for a Prolyl Radical Intermediate." J.

Am. Chem. Soc. 121(3): 587-588.

Zhang, Z., Ren, J., Stammers, D. K., Baldwin, J. E., Harlos, K. and Schofield, C. J.

(2000). "Structural origins of the selectivity of the trifunctional oxygenase

clavaminic acid synthase." Nature Struct. Biol. 7(2): 127-133.

Zhou, J., Gunsior, M., Bachmann, B. O., Townsend, C. A. and Solomon, E. I. (1998).

"Substrate Binding to the a-Ketoglutarate-Dependent Non-Heme Iron Enzyme

Clavaminate Synthase 2: Coupling Mechanism of Oxidative Decarboxylation and

Hydroxylation." J. Am. Chem. Soc. 120(51): 13539-13540.

Chapter 2: Kinetic dissection of the catalytic mechanism of taurine•α-ketoglutarate dioxygenase

Portions of this chapter are published in:

John C. Price, Eric W. Barr, Bhramara Tirupati, J. Martin Bollinger, Jr., and Carsten Krebs, 2003, “The First Direct Characterization of a High Valent Iron Intermediate in the

Reaction of an α-Ketoglutarate-Dependent Dioxygenase, an Fe(IV) Complex in Taurine/α-Ketoglutarate Dioxygenase (TauD) from Escherichia coli,” Biochemistry, 42, 7497-7508.

John C. Price, Eric W. Barr, Lee M. Hoffart, Carsten Krebs, and J. Martin Bollinger, Jr.,

2005, “Kinetic Dissection of the Catalytic Mechanism of Taurine:α-Ketoglutarate Dioxygenase (TauD) from Escherichia coli”, Biochemistry, 44, 8138-8147.

Abstract:

The Fe(II)- and α-ketoglutarate (αKG) -dependent dioxygenases have diverse roles in biology. A few examples of include the synthesis of collagen and fatty acids, sensing of oxygen in mammals, in nutrient acquisition and antibiotic synthesis in microbes, and in repair of alkylated DNA in both. A consensus mechanism for these

enzymes, involving (1) addition of O2 to a square pyramidal, (His)2(Asp/Glu)-facially- coordinated Fe(II) center to which αKG is also bound via its C-1 carboxylate and ketone oxygen atoms; (2) attack of the uncoordinated oxygen of the bound O2 on the ketone carbonyl of αKG to form a bicyclic Fe(IV)-peroxyhemiketal complex; (3) decarboxylation of this complex concomitantly with formation of an oxo-ferryl (Fe(IV)=O) intermediate; (4) abstraction of an H-atom from the substrate by the Fe(IV)=O complex; and (5) generation of product reduction of Fe via an “oxygen rebound”, has repeatedly been proposed, but none of the postulated intermediates occurring after addition of O2 has ever been detected. This work tests the concensus mechanism for the specific case of taurine•α- ketoglutarate dioxygenase (TauD), and presents direct evidence for the accumulation of two intermediates in the reaction of TauD with oxygen. A combination of rapid kinetic experiments (stopped-flow and freeze-quench), and spectroscopic methods (UV-visable absorption and Mössbauer), demonstrate that the first intermediate to accumulate contains an oxidized iron center. This intermediate forms with second-order kinetics

(first order in O2 and TauD quaternary complex) concomitant with the formation of CO2. The second kinetically resolved intermediate decays during the rate limiting step and is inferred to be a TauD•product(s) complex.

Introduction The study of collagen biosynthesis and subsequent realization that prolyl hydroxylase (a key enzyme in the biosynthetic pathway) is responsible in large part for the strength and elasticity of the collagen triple helix soon spurred the accumulation of a large body of research (Kivirikko et al. 1967). Hanauske-Abel et al used sound chemical reasoning to formulate a chemical mechanism to explain the prolyl hydroxylase activity (Hanauske-Abel et al. 1982). A key feature of their proposal was the formation of an Fe(IV)=O as the hydroxylating species (see Scheme 2-1), similar to the highly oxidized

Fe•porphyrin complex theorized to occur in the reaction of the cytochrome P450 class of hydroxylases. Although there was no direct evidence for the proposed mechanism, it explained current experimental literature while maintaining chemical logic. The basic tenants of the mechanism can be applied broadly to all members of the Fe(II)•αKG dioxygenase family (see Scheme 2-1). Later work on deacetoxy-deacetylcephalosporin C synthase (Baldwin et al. 1991), and prolyl-4-hydroxylase (Wu et al. 1999) supported the hypothesis that the chemistry of this enzyme family proceeds through a radical mechanism. This mechanism provided a basis of understanding, not only of prolyl-4- hydroxylase but also of all members of the Fe(II)/α-ketoglutarate dependent dioxygenase family. Members of this family of enzymes play key roles in many critical biological processes (Hausinger 2004), and understanding their mechanisms of action will give insight into many disease states that arise due to the malfunction of enzymes within the family. The in-depth study of enzyme mechanisms has in several cases led to radical reevaluations of the proposed chemistry for numerous enzymes. Therefore, the only way to truly understand the catalytic mechanism of an enzyme is through the direct detection and spectroscopic characterization of intermediate species. This can be accomplished through the use of stopped-flow and freeze-quench experiments, which allow the reaction

34 of an enzyme to be dissected kinetically during the course of a single turnover. Reaction intermediates that accumulate can then be characterized and studied via various specialized spectroscopies during the course of a single turnover.

Taurine•α-ketoglutarate dioxygenase (TauD) has been shown to be a member of the non-heme Fe(II)•αKG dependent dioxygenase family (Eichhorn et al. 1997). As an endogenous E. coli protein, a well established methodology already exists for the overexpression of the protein as well as the production of protein variants. The substrate for TauD is a stable, commercially available small molecule, 2-amino-ethane sulfonate (taurine). Although TauD itself plays no medically relevant role, it is proposed that a general mechanism is employed by all members of the family. Thus the mechanistic principles we define in the study of TauD can be applied to the more medicinally significant members of the family. In view of these points, TauD is an ideal model system for the study of this enzyme family, and was chosen as the focus for this study. MATERIALS AND METHODS Materials. Culture media components (yeast extract and tryptone) were purchased from Marcor Development Corporation (Hackensack, NJ). Isopropyl-β-D- thiogalactopyranoside (IPTG) was purchased from Biosynth International (Naperville, IL). Ampicillin was purchased from IBI (Shelton, CT). 5,5’-dithio-bis-(2-nitrobenzoic acid) (Ellman’s reagent) was purchased from Pierce (Rockford, IL). Glycerol, ammonium sulfate, sodium chloride, ferrous ammonium sulfate, and sodium hydroxide were purchased from EM Science (Gibbstown, NJ). Trizma (Tris) base, 2-amino-ethane-

1-sulfonic acid (taurine), 2-oxoglutarate (αKG), sodium sulfite, imidazole, poly(ethyleneimine), ethylenediaminetetraacetic acid (EDTA), sodium hydrosulfite (sodium dithionite), 2-methylbutane, and L-glutamate oxidase from streptomyces were purchased from Sigma Corp. (St. Louis, MO). Trichloroacetic acid and sulfuric acid were purchased from Fisher Scientific (Pittsburgh, PA). Dithiothreitol (DTT) was purchased from United States Biochemical (Cleveland, OH). DEAE-Sepharose FF resin

was purchased from Pharmacia (Piscataway, NJ). 57Fe metal was purchased from

Advanced Materials and Technologies (New York, NY). 1-[14C]-αKG was purchased from New England Nuclear (Boston, MA). All reagents used for synthesis were purchased from Aldrich Chemical Corp. (St. Louis, MO) and, unless otherwise noted, were used without further purification. Silica gel for the column chromatography was purchased from Sorbent Technologies (Atlanta, GA), Dowex 50W-X8 was purchased from Bio-Rad laboratories (Richmond, CA). O-[2H]-t-butanol (99% isotopic enrichment)

2 and [ H]5-ammonium hydroxide (99% enrichment) were purchased from Cambridge Isotope Laboratories (Andover, MA). Trimethylsilyl fluorosulfonyldifluoroacetate was purchased from William R. Doblier, Jr. (Dept. of Chem, University of Florida). Boc- Glu(Otbu)-OH was purchased from Bachem (King of Prussia, PA) Dimethyl oxalylglycine was purchased from Frontier Scientific (Logan, Utah). Construction of TauD over-expression strain by Bhramara Tirupati. An 878 basepair DNA fragment containing the tauD gene was amplified by PCR by using a suspension of Escherichia coli strain JM109 as template and primers, "forward" (5’- GGAGAAGTCATATGAGTGAACG-3’) and "reverse" (5’- CCGTCCACTCTCGAGTTACCCC-3’). The PCR fragment was restricted with NdeI (site underlined in "forward") and XhoI (underlined in "reverse") and ligated with pET22b that had been restricted with the same endonucleases. The ligation solution was

used to transform E. coli DH5α to ampicillin resistance. The entire TauD coding sequence from a plasmid (pTauD) isolated from one transformant was verified. This plasmid was then used to transform the expression strain, E. coli BL21(DE3). Over-expression of TauD. BL21(DE3)/pTauD cells were grown in enriched medium (35 g/L tryptone, 20 g/L yeast extract, 5 g/L NaCl, brought to pH 7.3 with NaOH and supplemented with 150 mg/L ampicillin) while shaking vigorously in Fernbach flasks at 37 °C to an optical density at 600 nm of 0.6-1. Cultures were cooled on ice for 15 min and then induced to over-express the TauD gene by addition of IPTG to 0.2 mM.

After induction, cultures were shaken at 21 °C for an additional 16 hr. The cells were then harvested at 4 °C by centrifugation at 10,000 x g for 10 min. A typical yield was 16 g of wet cell mass per L of culture. The cell paste was either used immediately for protein purification or was frozen in liquid nitrogen and stored at -80 °C prior to being used. Purification of TauD. All purification steps were carried out at 4 °C. In a representative purification, 107 g of cell paste was re-suspended in 500 mL of buffer A (50 mM Tris•HCl, pH 7.6, containing 10% (v/v) glycerol). The cells were lysed by a single passage through a French pressure cell at 16,000 psi. Following centrifugation of the lysate at 16,000 x g for 10 min, 0.05 volume equivalents of a 10% solution of poly(ethyleneimine) in buffer A was added to the supernatant slowly with continuous stirring. The resulting suspension was centrifuged at 16,000 x g for 10 min. The supernatant was brought to 60% of saturation in ammonium sulfate by slow addition of the solid with continuous stirring. The resulting suspension was centrifuged at 16,000 x g for 10 min, and the pellet (29.1 g) was re-dissolved in buffer A (108 mL) containing 1 mM sodium EDTA and 1 mM dithiothreitol. The protein solution was then dialyzed twice against 4 L of this buffer. The dialysate was loaded onto a 1 L DEAE-Sepharose FF column in buffer A. The column was washed first with 1 L of buffer A and then with a 750 mL linear gradient of 0-0.05 M NaCl in buffer A. The protein was eluted with a 2.25 L linear gradient of 0.05-0.2 M NaCl in buffer A. Fractions containing TauD were pooled and concentrated to ~ 3 mM in an Amicon Diaflow stirred pneumatic concentrator with a YM-30 membrane. The protein was dialyzed twice against 4 L of buffer A, flash- frozen in liquid nitrogen and stored at –80 °C. The protein was estimated to be greater than 95% pure by denaturing polyacrylamide gel electrophoresis and coomassie staining. The concentration of TauD was determined by absorbance at 280 nm by using the molar absorptivity of the protein (45,600 M-1cm-1) calculated according to the method of Gill and von Hippel (Gill et al. 1990).

Titration of TauD with Fe(II) to determine maximum stoichiometry of binding. Samples of concentrated, pure apo TauD were dialyzed against 50 mM Tris•HCl, pH 7.6, to remove the glycerol from the protein solution. Oxygen was removed from the protein by placing it into a vacuum flask and exchanging the head-space of the flask with argon on a vacuum/gas manifold. After ~ 30 cycles of gentle evacuation and refilling with argon over a period of ~ 1 h, during which time the protein was gently stirred on ice, the flask was sealed and kept at 0 °C overnight to allow for equilibration of the solution and gas phases. In a glove box (MBraun; Peabody, MA) with a nitrogen atmosphere,

buffered, O2-free solutions of taurine, αKG, and ferrous ammonium sulfate were prepared from dry stocks. Aliquots of the taurine and αKG stock solutions were added to the TauD solution (see legend to Figure 2-1 for concentrations). The protein solution was centrifuged for 2 min at the maximum speed in an Eppendorf Minispin microcentrifuge and placed in a cuvette, which was then sealed with a rubber septum. The cuvette was removed from the glovebox and this solution was used to "blank" the HP8453 diode array spectrophotometer. An aliquot of the buffered ferrous ammonium sulfate solution was added to the protein solution through the rubber septum. The solution was gently mixed and its absorption spectrum acquired. Additional aliquots were added until the absorbance at 520 nm no longer increased, indicating that saturation had been achieved. Determination of the steady-state rate for sulfite production. The procedure of Eichhorn, et al. (Eichhorn et al. 1997), which utilizes colorimetric quantitation of sulfite by its reaction with Ellman's reagent, was modified slightly for determination of TauD activity at 5 °C and pH 7.6, the reaction conditions employed in the stopped-flow and freeze-quench experiments. To each 0.5 mL aliquot of air-saturated assay solution

containing 50 mM Tris•HCl, pH 7.6, 5 mM taurine, and 5 mM αKG was added 10 µL of a solution containing 50 µM TauD and 1 µL of 100 mM ferrous ammonium sulfate in 5

mN H2SO4. The reaction was allowed to proceed at 5 °C for times from 30 to 90 s and was then terminated by addition of 150 µL of 0.5 M trichloroacetic acid. Within 30 s of

termination, 400 µL of the solution was neutralized with 100 µL of 1 M Trizma base. Within another 30 s, 40 µL of a 2.5 mM solution of Ellman's reagent in 50 mM Tris•HCl, pH 7.6, was added. After a ~ 15 min incubation to allow for reaction of the sulfite and

Ellman's reagent, the absorbance at 411 nm (A411) was determined. The quantity of sulfite produced by TauD in each assay solution was determined by comparison of its

A411 to those of standard samples for which known quantities of added sodium sulfite were reacted with Ellman's reagent.

14 Determination of steady-state rate of CO2 production. CO2 was measured by a procedure modified from that described by Jones and co-workers (Jones et al). In a 5 ºC cold room, 10 µL of enzyme mix (7 µM TauD , 0.5 mM Fe(II), and 50 mM Tris, pH 7.6), pre-loaded with iron 30 s prior to initiating the reaction, was added to 40 µL of assay mix (6.34 mM [1-14C]−αKG (0.788 Ci/mole), 6.25 mM 1H-taurine or 2H-taurine, 50 mM Tris•HCl, pH 7.6). The reaction was allowed to proceed in the cap of a 1.5 mL microfuge tube that was placed at the bottom of a hermetically sealed 20 mL glass scintillation vial, which also contained a flat-bottomed, cylindrical, 0.8 cm x 3.5 cm tube. The tube contained the CO2 trap (a 1.5 x 1.5 cm piece of filter paper that had been soaked in 2 M NaOH and blotted partially dry). After a reaction time of between 10 and 80 s, the

reaction was quenched and CO2 liberated by the addition of 1 mL 2 M HCl through the rubber septum cap sealing the top of the scintillation vial. The liberated CO2 was allowed to bind to the filter paper at room temperature for a period of approximately 90 min. The septum was removed from the vial, and the filter paper was removed from the flat- bottomed tube. The filter paper was placed into a new uncapped scintillation vial and left

° 14 to dry overnight at 37 C. Scintillation fluid was added and the CO2 counted in a LKB Wallac 1217 Rackbeta liquid scintillation counter. The total counts from the assay mix were measured by placing a known amount of the assay mix directly onto a piece of NaOH-treated filter paper, which was then dried and counted as above.

Kinetics of CO2 production in a single turnover. The kinetics of production of

14 14 CO2 from 1-[ C]-αKG in a single turnover were determined by chemical quenched- flow experiments. Oxygen was removed from concentrated stocks of pure TauD as described for the titration of TauD with Fe(II). In a glovebox (MBraun; Peabody, MA) containing a nitrogen atmosphere, buffered, O2-free solutions of αKG, taurine, and ferrous ammonium sulfate were prepared from dry stocks. Aliquots of these stocks were added to the enzyme to achieve the concentrations listed in the legend to Figure 2-9. 1-

[14C]-αKG, supplied in 10% ethanol in water, was added directly to the protein solution in an aliquot small enough to ensure that the final ethanol concentration was less than 0.02 %. The protein solution was centrifuged for 2 min at the maximum speed in an Eppendorf Minispin microcentrifuge and then loaded into one syringe of an Update Instruments System 1000 chemical/freeze quench apparatus. Air-saturated buffer and the quench solution (either 2 M HCl or 4 M NaOH) were loaded into separate syringes. In

the experiment, actuation of the ram drive caused the protein and O2-saturated buffer to be mixed at 5 °C and to pass through a reaction hose of length appropriate to give the desired reaction time. Upon emerging from the reaction hose, this solution was mixed with the quench solution and then injected into a 15 mL scintillation vial. Within 30 seconds, a CO2 trap (a 1 mm square piece of filter paper that had been soaked in 2 M NaOH, blotted dry, and then placed in a smaller plastic scintillation vial) was placed inside, the entire system was sealed with a rubber septum, and 1.2 mL 2 M HCl was injected through the septum. The entire system was gently shaken for 90 min at room

14 temperature to allow for trapping of the CO2. The filter paper was removed and dried overnight at 37 °C in a fresh scintillation vial. An 8 mL aliquot of scintillation fluid was added to each of the two vials containing the original quenched reaction mixture and the dried CO2 trap. Radioactivity present in each vial was quantified in a Perkin-Elmer 1217 Rackbeta scintillation counter. To account for variation in the volume of sample delivered in the quenched-flow "shots," the radioactivity in the volatile (trap) fraction

14 (corresponding to CO2) was divided by the sum of the radioactivity in the volatile and non-volatile (sample) fractions to obtain a fractional conversion. Mössbauer spectroscopy and data analysis. Mössbauer spectra were recorded on spectrometers from WEB research (Edina, MN) operating in the constant acceleration mode in transmission geometry. Spectra were recorded with the temperature of the sample maintained at 4.2 K. The sample was kept inside an SVT-400 dewar from Janis

(Wilmington, MA), and a magnetic field of 40 mT was applied parallel to the γ-beam. The quoted isomer shifts are relative to the centroid of the spectrum of a metallic foil of

α-Fe at room temperature. Data deconvolution and analysis was performed by using the program WMOSS from WEB research. Lee M. Hoffart used the Mathematica software package from Wolfram Research Inc. (Champaign, IL) to deconvolute the contributions from the individual Fe(II) species, quaternary complex and M shown in Figure 2-8. Stopped-flow absorption experiments and kinetic simulations. Stopped-flow absorption experiments were carried out at 5 °C in an Applied Photophysics (Surrey, U.K.) SX.18MV apparatus equipped with a diode array detector and housed in the MBraun anoxic chamber. A pathlength of either 1 cm or 0.2 cm and an integration time of 1.28 ms were used. O2-free stocks of the TauD•Fe(II)•αKG ternary or TauD•Fe(II)•αKG•taurine quaternary complex were prepared as described above for the Fe(II) titration experiment, with the exception that ferrous ammonium sulfate was added to the protein in a single aliquot in the glove box prior to centrifugation and loading into the stopped-flow apparatus. The TauD complex solution was mixed in the stopped-flow apparatus with an equal volume of either buffer A that had been allowed to reach equilibrium at 0 °C with a gas phase of 1.05 atm O2 (Figures 2-4A and 2-5) or O2-free buffer A (Figure 2-4B). In the former experiment, a final O2 concentration of 1 mM was assumed to result (Hitchman 1978). For the series of experiments examining the dependence of formation of J on [O2], the O2-free TauD•Fe(II)•αKG•taurine complex

was mixed in the stopped-flow apparatus with buffer solutions containing varying [O2],

which were prepared by mixing appropriate volumes of O2-saturated and O2-free buffer solutions. The stopped-flow mixing ratio was either 1:1 or 2:5 (protein:O2-buffer). For experiments investigating binding of substrates, the mixing ratio was 1:1. In some experiments with small absorbance changes, a photomultiplier detector was used in place of the diode array. Experimental details, including final reactant concentrations, are given in the figure legends. Kinetic simulations were performed with the program KinTekSim (KinTek Corporation; State College, PA). Preparation of Mössbauer samples. The (non-freeze-quenched) samples used to acquire Mössbauer spectra for the TauD•Fe(II) binary and TauD•Fe(II)•αKG•taurine quaternary complexes (Figure 2-2, spectra A and B) were prepared as described above, with the exception that the iron stock contained 57Fe(II) and was prepared as previously

57 described (Bollinger et al. 1994) by dissolution of Fe(0) in O2-free 2 N H2SO4. Prior to its addition to the protein, the final, acidic 57Fe(II) stock (50 mM Fe(II) in 100 mM

+ remaining H ) was mixed with 0.5 equivalent volumes of O2-free 1 M Tris•HCl, pH 7.6, in order to avoid transient acidification of the protein solution. After the samples were fully constituted (see legend to Figure 2-2 for concentrations), they were transferred to Mössbauer cells, sealed in vials, removed from the glovebox, frozen by immersion of the vials in liquid nitrogen, removed from the vials, and stored in liquid nitrogen. The apparatus and procedure for preparation of freeze-quenched Mössbauer samples have been described (Ravi et al. 1994). The TauD•57Fe(II)•αKG•taurine complex was prepared in the glovebox as above, loaded into a syringe, and removed from the glovebox. By actuation of the drive motor, this solution was mixed at 5 °C in a

volume ratio of 1:2 with O2-saturated buffer A. In preparation of samples for cryoreduction, glycerol (20% w/v) was present in both protein and O2 solutions. The resulting reaction solution was passed through an "aging hose" of length appropriate to give the desired reaction time, and the reaction was terminated by injection of the solution into the cold (-150 °C) 2-methylbutane cryosolvent. The total reaction time was

calculated as the sum of the known time for transit through the aging hose and the "quench time," the time elapsed between injection of the reaction solution into the cryosolvent and its cooling to a temperature below which no further reaction occurs. For the latter quantity, an estimate of 10 ms was obtained from earlier work on the ribonucleotide reductase system (Baldwin et al. 2000) and was assumed in this study.

Results Determination of optimum Fe(II)/TauD ratio and reactant concentrations. As a prelude to the kinetic/spectroscopic experiments investigating the mechanism of TauD, equilibrium binding measurements similar to those previously reported by Ryle, et al. (Ryle et al. 1999) were carried out to determine the optimum Fe(II)/TauD stoichiometry.

Consistent with the previous work, the visible chromophore (λmax = 520 nm, ε520 = 200 M-1cm-1) of the TauD•Fe(II)•αKG•taurine quaternary complex was seen to develop

4 -1 -1 rapidly (> 3 x 10 M sec , vida infra) as Fe(II) was added in the absence of O2 to a solution containing the enzyme (typically 0.5-1 mM) and its two substrates (5 mM each) (Figure 2-1). Plots of the background-corrected (protein absorption removed) absorbance at 520 nm as a function of the concentration of Fe(II) (inset to Figure 2-1) could be analyzed to give an apparent dissociation constant (Kd) for the complex of 6 ± 4 µM and a maximum complex stoichiometry of 0.87 ± 0.05 Fe(II)/TauD (means ± standard deviations from six independent trials). Fe(II)/TauD ratios of 0.80 to 0.85 were selected for the stopped-flow absorption and freeze-quench Mössbauer experiments. Similar titration experiments (not shown) established, in agreement with results reported by Ryle, et al. (Ryle et al. 1999), that an αKG concentration of 5 mM is effectively saturating (KD = 30 ± 10 µM). Changes at the Fe(II) site upon substrate binding monitored by Mössbauer spectroscopy. Also as a prelude to the mechanistic experiments, changes at the Fe(II) center upon conversion of the TauD•Fe(II) binary complex to the

TauD•Fe(II)•αKG•taurine quaternary complex were monitored by Mössbauer spectroscopy. The spectrum of the binary complex, recorded at 4.2 K in a 40-mT

magnetic field applied parallel to the γ-beam, is shown in Figure 2-2A. It can be simulated as a broad quadrupole doublet with parameters (isomer shift, δ, of 1.27 ± 0.05 mm/s; quadrupole splitting, ∆EQ, of 3.06 ± 0.05 mm/s) typical of high-spin Fe(II). Upon addition of αKG and taurine (Figure 2-2, spectrum B), the Mössbauer parameters change significantly (δ = 1.16 ± 0.05 mm/s, ∆EQ = 2.76 ± 0.05 mm/s). The reduction in isomer shift upon binding of the substrates has been shown to occur in other mononuclear non- heme Fe(II)-enzymes (Arciero et al. 1983; Wolgel et al. 1993) and is consistent with the conversion of a six-coordinate Fe(II) site to a five-coordinate, square-pyramidal site. Kinetics of substrate binding. The 530 nm chromophore in the TauD system develops due to a metal-to-ligand charge transfer band between Fe(II) and αKG. This charge transfer band requires direct bonding interaction between αKG and the Fe(II) in the active site. It is sensitive to the presence of substrate, and blue shifts to 520 nm upon addition of taurine. Thus we can assess rates of binding for αKG and taurine. To measure the kinetics for αKG binding to the TauD•Fe(II) complex in the presence and absence of taurine, the development of the Fe(II)-to-αKG charge-transfer transition at 520-530 nm was monitored. As shown in Figure 2-3, binding of αKG is multiphasic, both when taurine is absent and when taurine is present in the αKG solution (i.e., when the two substrates are mixed with the enzyme simultaneously). With increasing [αKG], the fraction of the reaction that is complete in the dead-time of the instrument (1.3 ms) increases. The shape and λmax of the spectrum that develops in this fastest phase are indistinguishable from those characterizing subsequent changes. A phase with apparent first-order rate-constant of > 100 s-1 and observed amplitude that decreases with increasing [αKG] is then observed. A second observed phase with smaller amplitude and apparent first-order rate constant > 20 s-1 is observed. Finally, a phase that is too slow (k

-1 -1 < 0.5 s ) to be on the catalytic pathway (kcat = 1.3 s ) is also observed at all

concentrations. The change in the final equilibrium absorbance values is indicative of the change in KD for the quaternary complex. This change is transient-state evidence for the demonstrated synergistic binding of the two substrates (Ryle et al. 1999), which reduces

the apparent KD for αKG from ~ 300 µM in the absence of taurine to ~ 6 µM in the presence of 5 mM taurine. Stopped-flow absorption evidence for two intermediates (J, M) in the TauD reaction. The time-dependent absorption spectra acquired after initiation of the reaction at 5 °C by mixing of the TauD•Fe(II)•αKG•taurine quaternary complex with O2-saturated buffer provide clear evidence for the accumulation of two distinct intermediate states

prior to the regeneration of the quaternary complex following consumption of O2. To illustrate this conclusion, the essentially invariant spectra from a control experiment in which the quaternary complex was mixed with O2-free buffer (Figure 2-4B) have been subtracted from the experimental, time-dependent spectra of the complete reaction (Figure 2-4A). The control establishes that the changes observed in the complete reaction

are associated with reaction of the complex with O2, and the subtraction removes absorption from the protein (including the reactant complex) and contaminants so as to reveal these changes more clearly. An additional control experiment was conducted to confirm that, as reported by Ryle, et al. (Ryle et al. 1999), absorbance changes are smaller and are observed only on a much longer time scale when the ternary

TauD•Fe(II)•αKG complex (i.e., with taurine omitted) is mixed with O2 (Figure 2-5B). The first intermediate state (J) in the complete reaction is characterized by an absorption feature centered at 318 nm (Figure 2-4C, red trace). It develops to its maximum extent within the first 20-25 ms and then decays within a reaction time of ~ 600 ms (Figure 2- 5A, circular points). The difference spectrum corresponding to the time of maximal accumulation of the first intermediate state exhibits, in addition to the positive feature at 318 nm, a less intense negative feature at lower energy attributable to loss of the charge- transfer band of the reactant complex (Figure 2-4C, red trace). The minimum of this

difference spectrum is, however, red-shifted by approximately 65 nm from the peak of the reactant complex that has decayed (inset to Figure 2-4C). This shift is attributable to

(1) significant absorption by the first intermediate in the region of the λmax of the starting complex and (2) the more precipitous "trailing off" toward lower energy of the intermediate's spectrum in comparison with that of the starting complex. Indeed, J is approximately isosbestic with the starting complex at its λmax of 520 nm, and, consequently, the kinetic trace at this wavelength exhibits a lag phase during the accumulation of the intermediate (Figure 2-5A, squares). This coincidence may (in addition to differences in the reaction conditions employed) explain why evidence for the intermediate was not previously noted. As decay of J proceeds, the difference spectrum comes to reflect more closely the negative feature at 520 nm attributable to loss of the starting quaternary complex (Figure 2-4C, green trace in main figure and inset). This observation indicates that the second intermediate state (M) has less absorption than J in this spectral region. Finally, M, which has essentially no absorbance in the UV-visible region, decays as the starting quaternary complex is regenerated and its 520 nm absorption returns, resulting in a nearly featureless difference spectrum (Figure 2-4C, black trace). Trapping of a novel Fe intermediate (J) and its characterization by Mössbauer spectroscopy. The nature of the intermediate states was probed by rapid freeze-quench Mössbauer spectroscopy. The 4.2-K / 40-mT spectrum of a sample that was rapidly frozen ("freeze-quenched") after a reaction time of 20 ms, near the time of maximum absorbance at 318 nm in the stopped-flow experiments, exhibits a contribution from the high-spin Fe(II) site and an additional prominent peak at 0.75 mm/s (Figure 2-2, spectrum C). The deconvolution of this spectrum (solid line plotted over spectrum C in Figure 2-2) reveals that it is composed of 50 ± 3 % of the spectrum of the quaternary

complex (spectrum B) and 46 ± 3 % of a quadrupole doublet with δ = 0.31 ± 0.03 mm/s

and ∆EQ = 0.88 ± 0.03 mm/s (solid line plotted above spectrum C). Spectra of samples in

which the reaction was quenched after 80 ms, 200 ms, and 5 min (Figure 2-2, spectra D, E, and F) can be analyzed as superpositions of the contributions from multiple high-spin

Fe(II) species and the δ = 0.31 species. They reveal that the absorption of the new species decreases with increasing time (27 ± 3 % after 80 ms, 13 ± 3 % after 200 ms, and zero after 5 min). Thus, the Mössbauer spectra unambiguously demonstrate the accumulation of a novel Fe intermediate in the reaction of the TauD•Fe(II)•αKG•taurine complex with O2. Simulation of the stopped-flow kinetic data and prediction of the steady-state turnover rate. The kinetic traces from two diagnostic wavelengths: 318 nm (circles, lefthand axis), where the first intermediate dominates; and 520 nm (squares, righthand axis), where the reactant complex and the first intermediate both absorb, are shown in Figure 2-5. These traces have initial and final absorbance values of zero because, as in Figure 2-3C, the absorption from the reactant complex has been removed by subtraction. The traces can be simulated according to Scheme 2-2. The parameters of the simulations

are the concentrations of the quaternary complex and O2, the rate constants for the three steps of the reaction sequence, and the changes in molar absorptivity (∆ε) from those of the reactant complex for the two intermediates at the three wavelengths. These parameters are summarized in Scheme 2-2, along with estimates from the titration experiments of the molar absorptivities of the reactant complex at the three wavelengths. The quantities of the first intermediate determined in the freeze-quench Mössbauer experiments provide crucial constraints for this simulation. With rate constants consistent with the stopped-flow data, the simulation over-predicts the quantities of the intermediate if it is assumed that the concentration of the starting quaternary complex is equal to the known concentration of the limiting constituent (Fe(II) or TauD, depending on whether the Fe(II)/TauD ratio was less than or greater than the complex stoichiometry of 0.9 determined in the titration experiments). This observation suggests that a fraction of the starting complex either is unreactive toward O2 or reacts with markedly different

(slower) kinetics. Heterogeneity in reactivity is commonly observed in "single-turnover" experiments with enzymes and seems to be observed almost invariably in Fe-dependent oxidases and oxygenases (see, for example, references (Liu et al. 1995; Baldwin et al. 2000)). Best agreement between the simulation and data is obtained when it is assumed that the concentration of the reactive complex is 0.72 times the expected concentration.1 With this assumption and the parameters given in Scheme 2-2, the stopped-flow kinetic traces can be simulated well (compare solid lines and data points in Figure 2-4A; see below for discussion of the imperfect agreement between the A520 traces at long reaction time), and the predicted concentrations of the J (solid line in Figure 2-6) match almost precisely with those determined by Mössbauer (data points in Figure 2-6).

The steady-state turnover number predicted by Scheme 2-2 (with the O2 concentration expected of an air-saturated solution, the condition under which our activity assays have been carried out) is 1.3 s-1. Here, it has again been assumed that 0.72 equivalents of the enzyme is reactive. The fact that the measured value of 1.3 ± 0.2 s-1 (mean ± standard deviation for 4 independent trials) agrees well with the predicted value establishes that all three observed steps are kinetically competent to be constituents of the hydroxylation pathway. The nature of the second intermediate state (M) and its conversion to the reactant complex. The minimal kinetic mechanism predicts that M should accumulate to a measurable degree. Decay of the Mössbauer signature of the first intermediate species J is associated with the return of the broad quadrupole doublet component(s) attributable to

1 As an alternative to heterogeneity in the reactant, the assumption of reversibility in formation of the first intermediate can correct the over-prediction of the quantities of J as a function of time. This reversibility would have important implications for the identity of the intermediate, variation of the concentrations of the reactants (the quaternary complex and O2), demonstrates (vide infra) that this is not the case.

high spin Fe(II) species, thereby clearly demonstrating that M is a high-spin Fe(II) species. However, at reaction times of 0.2 – 10 seconds, the Fe(II) features are not all attributable to the starting quaternary complex, implying that the spectrum of M is partially resolved from that of the reactant complex. To resolve the spectrum and kinetics of M, two series of samples (time courses) were analyzed. The formation time course plots the formation of M, while the decay time course follows the decay of M and re-formation of the quaternary complex. For both time courses, simulation of the kinetics of J confirmed that, as in the previously described experiment, only ~ 70 % of the Fe(II)

present in the sample was competent to react rapidly with O2 to form J. The results of the analyses of the two time courses are summarized in Table 2-1. The experimental Mössbauer spectrum for each time point was analyzed as a linear combination of the experimentally derived reference spectra for the quaternary complex, J, and M. The reference spectrum for J was generated from the spectrum of the 20 ms sample after removing the features emanating from the quaternary complex. Similarly the reference spectrum for M was generated from the spectrum of the 200 ms sample by removing spectral features of J and the quaternary complex. The relative amount of J was easily determined from the well-defined high-energy line. The relative amount of the quaternary complex was estimated from the kinetics of M in combination with the observed amount of unreactive quaternary complex. Summation of reference spectra in the proportions given in Table 2-1 gave best agreement (minimum sum squared error) with the experimental spectra. Due to significant overlap of the reference spectrum of M and quaternary complex it is not possible to determine whether the features that we attribute to the second intermediate arise from one species or several species with similar Mössbauer parameters. It is clear that the spectrum of this species is distinct from that of the quaternary complex. Moreover, these features are distinct from those of the binary

TauD•Fe(II) and ternary TauD•Fe(II)•αKG complexes, because the high-energy lines of these complexes are at even higher energies than that of the quaternary complex. This is

illustrated in Figure 2-7 D-F, in which the reference spectrum of M (hash marks) is compared to spectra of the quaternary, ternary, and binary complexes, respectively. The fast substrate binding rates and the absence of the diagnostic metal to ligand charge

transfer band, which will be observed if αKG is chelating the Fe site, suggest that M is a TauD•Fe(II)•product complex. As mentioned previously the data will not allow differentiation between one of several possible product complexes. Determination of the contribution for the quadrupole doublet associated with M, during a single turnover reaction, shows the decay of this species experiences a similar lag to that seen in the 520 nm rebound of quaternary complex. The kinetic data from the Mössbauer analyses were compared directly to stopped-flow data from an experiment with similar reaction conditions (see Figure 2-8). At 480 nm, the reactant quaternary complex and J are approximately isosbestic, whereas M is essentially transparent. Thus, the decay and rebound of absorbance at this wavelength should correlate with the accumulation and disappearance, respectively, of the second intermediate M. At 600 nm, quaternary complex still has a strong absorbency while the contribution due to J is

minimized and M is essentially transparent. Thus the A600-versus-time trace should directly reflect disappearance and re-accumulation of the quaternary complex. Comparison of the titration data (Figure 2-1) and the extinction coefficients from the simulation of the stopped flow results, allows estimation of the extinction coefficients for

-1 -1 -1 -1 all three species at both wavelengths: ε480 = 185 M cm and ε600 = 130 M cm for the

-1 -1 -1 -1 -1 -1 quaternary complex, ε480 = 185 M cm and ε600 = 30 M cm for J, and ε480 = 0 M cm

-1 -1 and ε600 = 0 M cm for M. The concentrations of the species can be obtained from Mössbauer by correlating the relative amounts with the known total 57Fe concentration. Comparison of the results from the two methods is shown in Figure 2-8 for M (panel A) and the quaternary complex (panel B). The agreement is satisfactory and corroborates that the partially resolved Mössbauer features do, in fact, arise from M.

Kinetics of decarboxylation of αKG in a single turnover. To assess the timing of decarboxylation of αKG relative to formation of J and, thus, to determine whether the C1

14 carboxylate ligand is present in the intermediate, the kinetics of CO2 production in a single turnover were defined in chemical quenched-flow experiments with 1-[14C]-αKG.

Given the KD of αKG in the presence of 5 mM taurine (20 ± 19 µM) (Ryle et al. 1999), the concentration of binary TauD•Fe(II) complex (1 mM) was sufficiently high to assure

that the αKG was 98% saturated with enzyme. Acid, base, and denaturant quenching were tested. Acid and base quenching yielded kinetic data in mutual agreement and fitting the compiled data as an exponential rise gave an apparent first-order rate constant of 42 s-1 (Figure 2-9). The non-zero y-intercept of approximately 8% of the quantity of

14 CO2 at completion is attributable to ~ 5 µM contaminating O2 (which typically remains following the procedure employed to remove O2). With the concentration of O2 employed (0.26 mM), this apparent first-order rate constant corresponds to a second- order rate constant of 1.6 x 105 M-1s-1, in excellent agreement with the value obtained in

simulation of the stopped-flow absorption data ((1.5 ± 0.2) x 105 M-1s-1, see Scheme 2-1). These results suggest that decarboxylation occurs during formation of J. To assess the possibility that decarboxylation occurs after formation of J but both acid and base quenching lead to cleavage of the intact C1-C2 bond in this complex, attempts were made to quench with organic solvents (THF) and denaturants (5.5 M urea, 4 M guanidinium•HCl). Unfortunately, none of these treatments terminated the reaction

14 sufficiently rapidly to obtain accurate kinetics of CO2 production. Thus, the possibility that both acid and base quenching of J induce cleavage of the αKG C1-C2 bond cannot be excluded. The simplest interpretation of the data, however, is that decarboxylation occurs prior to or concomitantly with formation of J.

Dependence of the kinetics of J on concentration of O2. The dependence on [O2] was interrogated specifically, with the quaternary complex as the limiting reactant and a maximum O2 concentration of 1.4 mM (Hitchman 1978). As predicted, A318-versus-time

traces (Figure 2-10) show a clear dependence on [O2] in the formation phase. Using the equation for a two-exponential rise-fall to fit these data, which is not rigorously valid due to the facts that pseudo-first-order conditions are not maintained and the rise and fall phases are not kinetically well-resolved at low [O2], shows that the dependence on [O2] is, in fact, approximately linear (inset to Figure 2-10A). To evaluate the data rigorously and quantitatively, traces were simulated (solid lines in Figure 2-10A) according to the 3-step kinetic mechanism. Agreement is good after the first ~ 15 ms. The deviation at shorter reaction times may be a result of a prolonged mixing artifact. Indeed, the first few milliseconds of data show a steep decay phase that is certainly a mixing artifact. If this artifact is not fully extinguished until ~ 15 ms, then the deviation of the experimental traces from the theoretical traces would be explained. Alternatively, the deviation at early time could reflect accumulation of an absorbing precursor to J. Chemical logic indicates that these precursors must form, and in cytochrome P450, which uses a covalently bound heme iron to perform very similar chemistry, the peroxo bound form was stable enough to be characterized by X-ray crystallography (Schlichting et al. 2000).

To assess the possibility that a precursor accumulates, the A318 traces were simulated according to an expanded kinetic mechanism that includes a dissociable initial complex with O2 as a precursor to J (lines in Figure 2-10B; parameter values are given in Scheme 2-3). As must be the case, the expanded mechanism allows for better agreement in the early regions of the traces due to absorbance from the precursor. However, the approximately first-order dependence of the traces on [O2] at longer times effectively sets upper limits both on the quantity of the precursor that can accumulate (~ 0.3 equiv) and on the time of maximal accumulation (~ 1 ms). The minimal and fast accumulation of such a species would explain why it was not detected in the freeze-quench Mössbauer experiments, in which a quench time of 10 ms was estimated and J was observed to have reached or approached its maximum level at the shortest accessible reaction time.

The dependence of the kinetics of J on the concentration of the

TauD•Fe(II)•αKG•taurine complex was also explored. In this case, an upper limit on the accessible concentration of the excess reactant (the quaternary complex) of ~ 1 mM was imposed by declining mixing efficiency. Viscosity increases with the increasing protein concentration, and mixing of the more concentrated protein solutions with the O2- containing buffer resulted in an incrementally greater mixing artifact. Nevertheless, the expected dependence was seen after accounting for the mixing artifact (Figure 2-11). The same simulation parameters were used to verify the kinetics of the reactions for the excess quaternary complex (solid lines in Figure 2-11) as were used in the simulation of the [O2] experiments (see Figure 2-10). The comparison of simulation with experimental data shows that the mixing artifact which gives the appearance of additional absorbance does not change the kinetics of the system. The maximum absorbance occurs precisely at the time point predicted by the simulation. If the mixing were more efficient the simulations would match the data precisely. Verification of substrate binding rates. To distinguish whether the kinetic

complexity in αKG binding arises from the presence of multiple conformational states (parallel binding processes) or from a multi-step (sequential) binding process, an experiment was performed in which the binary TauD•Fe(II) complex was mixed

simultaneously with αKG, taurine and O2 and the kinetics of formation of J were defined by monitoring development of its absorption feature at 318 nm. As shown in Figure 2-12, formation of J is only slightly slower in this reaction than in mixing of the pre-formed

quaternary complex with O2 (compare circles to black trace, which is a simulation for the reaction of pre-formed quaternary complex with O2). Results of a control experiment in which taurine was omitted (Figure 2-12, squares) establish that both substrates must bind for rapid O2 activation, i.e. that the quaternary complex must form prior to O2 activation. Simulation of the experimental trace implies that the requirement for αKG and taurine binding imposes a lag in formation of J characterized by an apparent first-order rate

constant of at least 150 s-1 (red trace), corresponding to a second order rate of at least

4 -1 -1 3x10 M s . By contrast, if a step involved in formation of the O2-reactive complex were to be primarily rate-limiting (2.5 s-1), formation of J would be markedly slower and its accumulation suppressed upon initiation of the reaction by mixing simultaneously with all three substrates (Figure 2-12, blue trace).

Explanation for Previously Observed Slow Phase of TauD•Fe(II)•αKG Complex Regeneration (experiments performed by Eric Barr). The three-step minimal kinetic mechanism successfully simulates the data for formation and decay of J under all experimental conditions used in this study. The three-step kinetic mechanism does not successfully rationalize the slow phase in regeneration of the 520 nm charge-transfer band of the TauD•Fe(II)•αKG•taurine complex following depletion of O2 (when limiting with respect to the other substrates) and decay of the putative Fe(II)-containing intermediate. This phase is slower than the steady-state rate constant under these conditions (i.e. 5 °C and pH = 7.6) and therefore too slow to be on the catalytic pathway, providing argument that it must be attributable to conditions employed in the stopped- flow experiments but not the steady-state determinations of kcat (e.g., the much greater concentration of enzyme). Accumulation of high concentrations of products is a likely

cause. The evidence for a preferred or obligatory binding order of αKG and then taurine would suggest that product release probably occurs in the order of 1-hydroxytaurine followed by succinate. Interference with this sequence upon trapping of the TauD•Fe(II)•succinate complex by taurine (which was present in excess in the stopped- flow experiments) might then explain the previously observed slow phase in regeneration of the quaternary complex. This hypothesis makes two predictions that can be tested. First, the slow phase should become less pronounced with diminishing concentration of taurine. Figure 2-13A shows that this prediction is borne out. Second, taurine might become inhibitory at high concentrations. Indeed, taurine inhibition is observed and is

most pronounced at higher solvent viscosity (Figure 2-13B), when release of product (ordered substrate binding experiments would suggest succinate) is retarded (vide supra).

Discussion

Formation of the reactive TauD•Fe(II)•αKG•taurine complex. With the investigation of the dynamics for formation of the reactive TauD•Fe(II)•αKG•taurine complex, the effort was not only to understand the formation of quaternary complex but also to evaluate whether formation of the quaternary complex is the rate limiting step of the reaction. The demonstration of [αKG] dependence for absorbance at completion (i.e. at equilibrium) in the absence of taurine (Figure 2-3A), but independence in the presence of taurine (Figure 2-3B) is a transient state observation of the previously demonstrated synergistic binding of the two substrates (Ryle et al. 1999). The presence of taurine reduces the apparent KD for αKG from ~ 300 µM in the ternary complex (i.e. the absence of taurine), to ~ 30 µM in the quaternary complex (i.e. saturating taurine concentrations).

Multiple kinetic phases are observed during the binding of αKG (Figure 2-3). Two possible mechanisms would account for the observation of several phases in these experiments. (1) It could indicate that several sequential steps occur during αKG binding, or (2) it may suggest the presence of multiple conformational states that bind αKG with different kinetics, affinities, or both. The presence of multiple conformational states of TauD has previously been proposed (Ryle et al. 1999; O'Brien et al. 2003). The

Fe(II)•αKG dioxygenase family also displays a variability in the binding mode of αKG (Hausinger 2004). The bi-dentate binding of αKG is observed two planes, with the C1 carboxylate opposite to either of the two histidine ligands. Regardless of this complexity, at the highest [αKG] examined, at least 95 % of the total change in absorbance occurs in the two fastest phases, which are both much faster than the rate-limiting regeneration of

the reactive TauD•Fe(II)•αKG•taurine quaternary complex. Experiments in which the

TauD•Fe(II)•αKG ternary complex was mixed with taurine revealed that the slight shift

in λmax of the Fe(II)-to-αKG transition (from 530 nm to 520 nm) associated with taurine binding is complete in the dead-time of the instrument, also ruling out taurine binding as rate-limiting. When taurine was present with the TauD•Fe(II) complex prior to mixing

with αKG in the stopped-flow apparatus, binding of αKG was significantly slowed (Figure 2-3B, lower three traces). Previous x-ray crystallographic studies showing that TauD assumes a “closed” position upon binding of taurine to the TauD•Fe(II) complex (O'Brien et al. 2003) provide a simple explanation for this observation. The ordered addition of substrates, while an important mechanistic detail, does not seem to affect the

multiple phases seen during the binding of αKG. These phases are observed both in the absence (Figure 2-3A) and the presence (Figure 2-3B, upper traces) of taurine. The question whether these phases represent multiple conformational states of the protein or a

series of steps during the binding of αKG must be addressed. Also if these phases represent a series of steps during the formation of quaternary complex, it is important to establish which phase represents creation of the reactive complex. The > 150 s-1 lag in the activation of oxygen observed when substrates are required to bind prior to O2 activation (Figure 2-12), suggests that the kinetic complexity in αKG binding arises from parallel processes involving different states of the enzyme. The three different phases seen during the binding of αKG may reflect heterogeneity in the protein, or multiple chelating modes for the αKG. Variable states of the enzyme may also account for the heterogeneity seen in the reaction of the quaternary complex with oxygen accounting for the ~ 30 % of the bound Fe which is catalytically inactive.

Importantly, the observation of a > 150 s-1 lag during the formation of the quaternary complex eliminates the substrate binding steps as candidates for the rate-limiting step. Kinetic description of the TauD reaction. When compared to Scheme 2-1 which invokes eight different states in the catalytic cycle, Scheme 2-2 seems very simplistic. This three state mechanism is a kinetic description of the reaction. Its assumptions and

simplifications are all tested experimentally and clear evidence is presented for the deviation of the simulation from experiment during the reformation of the quaternary complex. The bimolecular nature of the step in which the first intermediate is formed has been verified by explicit variation of reactant concentrations. Variation of both oxygen and quaternary complex show a first order effect on the rate of formation for J. In the synthesis of the TauD kinetic and chemical mechanisms presented in Scheme 2-1, the

318-nm-absorbing species J forms from the TauD•Fe(II)•αKG•taurine complex and O2. It is obvious that only species I can truly form with second order kinetics. If, however, species I forms with unfavorable kinetics (i.e. slower formation and faster decay), it would not accumulate and we would observe the following species, II, form with a first order dependence on oxygen. Similarly any series of intermediates wherein the decay is faster than the formation do not accumulate. Due to the second order addition of oxygen, a continual increase in the rate is possible as [O2] is increased. This means that the

kinetically masked species would accumulate if the [O2] could be increased far enough, unfortunately the [O2] in aqueous solution is limited to ~ 2 mM. Thus, the kinetic mechanism predicts that formation of J is kinetically first-order in the quaternary complex and first-order in O2, but does not restrict the possible identities of J. The kinetic mechanism considers all three kinetically resolved steps as irreversible. The likelihood that this assumption is correct is dependent on the placements of the intermediate states within the TauD cycle. Clearly, decarboxylation of

αKG and O-O bond cleavage are expected to be irreversible, but earlier steps may not be.

The combination of a fast CO2 production and the absence of accumulating intermediates at high oxygen concentrations argues that decarboxylation occurs either through the decay of a kinetically masked intermediate which forms prior to the accumulation of J or that the harsh nature of the quench facilitates this decarboxylation from a complex that would not normally yield CO2 until its decay. The radical recombination is also a step

where in the assumption of irreversibility could be made with some confidence. Therefore assignment of irreversibility if true may be an important clue for the integration of the kinetic and chemical mechanisms. Scheme 2-2 is not able to predict the reformation phase of the 520 nm trace. This is in part because the scheme takes no account of the fact that significant concentrations of the products, succinate, 1-hydroxytaurine (or its breakdown products) and CO2, are produced through the course of the reaction. The time-dependent accumulation of these products complicates the kinetics of recovery of the reactant complex in the last phase of the reaction.

Approximation of the 1-10 s portion of the A520 kinetic trace as a first order process allows a rate constant of 0.4 s-1 to be estimated. This rate constant is too slow to be part of a catalytic cycle operating at 1.3 s-1, the rate constant under initial velocity conditions (i.e., in the absence of products). Among all the kinetic phases for all three

wavelengths, it is the recovery phase of A520 that should, according to Scheme 2-2, be most diagnostic of the conversion of the second intermediate into the reactant quaternary complex. Therefore, the disagreement between the simulation and data probably arises from an over-simplification relative to this step that is inherent in Scheme 2-2. Multiple phases occur during the rebound of the 520 nm absorbing quaternary complex. The initial rebound rate matches the simulation. Later time points show that some inhibition of the rebound is taking place. High concentrations of taurine (Figure 2-13) exaggerate the inhibition and account in large part for the deviation from the predicted rebound rate in the ~ 1-10 s regime of the A520-versus-time trace. Thus, inhibition of succinate release by the out-of-sequence binding of taurine is at least partly responsible for the slow phase of quaternary complex regeneration. In spite of the demonstrated retardation of αKG binding associated with prior addition of taurine to the TauD•Fe(II) complex (Figure 2- 3B, lower three traces), frustration of the preferred substrate binding order is unlikely to contribute to the slow phase or to inhibition by taurine, as αKG binding is not

significantly slowed when the enzyme is simultaneously exposed to the two substrates (Figure 2-3B, upper three traces). This inhibition also solidifies the identification of M as a product complex. Trapping of the product complex by taurine slows the conversion of M to quaternary complex (Figure 2-8). Attempts to more precisely model the rebound rate by inclusion of taurine inhibition in the decay of M to the quaternary complex are insufficient to fully reconcile the slow lag. The inability to model the lag with simple substrate inhibition, indicates that multiple levels of inhibition occur. The observation of increased inhibition at higher viscosity (Figure 2-13B) indicates that the high concentration of protein (a viscosogen) may play a role in exaggerating this inhibition. Simple product inhibition is an inadequate answer, as additional succinate seems to affect the rebound to a minor degree. These points argue that multiple factors such as conformational states of the protein, hydration of the active site and the presence of the labile product 1-hydroxy-taurine may contribute to the inhibition. The ability of Scheme 2-2 to accurately predict the steady-state rate provides support for this conclusion.

Identity of the novel oxidized TauD•Fe intermediate J. Within the context of chapter 2 the identity of J is not well defined. The experiments presented in chapter 3 make clear several important chemical and structural details that allow definitive assignment of J within the catalytic cycle. Here we attempt to differentiate between the possible identities of J and give literature precedent for the possible assignments.

The isomer shift of the novel intermediate (δ = 0.31 mm/s) is at the lower end of the range typical of a high spin Fe(III) species (δ = 0.3 to 0.6 mm/s), and unusually large for an Fe(IV) species (δ = -0.1 to 0.3 mm/s) (Greenwood et al. 1971). Borovic and coworkers reported two model complexes that have remarkably similar Mössbauer parameters. An Fe(III)=O model complex where significant hydrogen bonding allows for

stabilization of the Fe=O unit, has the Mössbauer parameters (δ = 0.33 mm/s, ∆EQ=0.92 mm/s) (MacBeth et al.). They also reported the protonated form, a Fe(III)-OH complex, which is also stabilized by H-bonds (δ = 0.30 mm/s, ∆EQ=0.71 mm/s). Interestingly, they

proposed that the Fe(III)-OH complex forms from the Fe(IV)=O complex by abstracting a hydrogen atom from solvent or 9, 10-dihydroanthracene (which was also in solution). This process is analogous to the conversion of species III to species IV (Scheme 2-1). Evidence for a Fe(III) moiety similar to species I is found in the reaction cycle of the DNA cleaving agent bleomycin. This protein forms an activated low spin (S = 1/2) Fe(III)-OOH complex which has been extensively characterized by Mossbauer (Burger et al. 1983), mass spectroscopy (Sam et al. 1994), and MCD spectroscopy (Neese et al. 2000). Activated bleomycin forms prior to the cleavage of the DNA. This Fe(III)-OOH may not be competent to attack the ribose ring and multiple theories incorporate the formation of an Fe(IV)=O moiety, after O-O bond cleavage, which may act as the functional oxidant. The isomer shift of the TauD intermediate is within a range described by several Fe(III) sites, parameters for the low spin site of activated bleomycin are 0.16 mm/s (Burger et al. 1983), and high-spin Fe(III) sites (e.g., the Fe(III) form of tyrosine hydroxylase (Meyer-Klaucke et al. 1996) exhibit an isomer shift of 0.56 mm/s). The literature offers several examples of Fe(III) species with Mössbauer parameters very similar to those for J. The majority of known Fe(IV) species, including the many heme Fe(IV)=O intermediates (see (Debrunner 1989) and references cited therein) and several inorganic heme and non-heme models (e.g.,(Simonneaux et al. 1982),(Grapperhaus et al. 2000),(Rohde et al. 2003),(Lim et al. 2003)) that have been characterized have isomer shifts less than 0.14 mm/s. All of these species contain low-spin (S = 1) Fe(IV) sites. High-spin Fe(IV) sites have, however, been demonstrated or proposed in several protein and inorganic dinuclear Fe complexes. In one model system, a high-spin Fe(IV)=O site

2- has been proposed to form by isomerization of an Fe(III)-(O )2-Fe(IV) species to a putative Fe(III)-O2--Fe(IV)=O unit (Zheng et al. 2000). The Fe(IV) site in this complex has an isomer shift of 0.10 mm/s, toward the upper end of the range characterizing the low-spin Fe(IV) complexes but still considerably less than that of the TauD intermediate.

Significantly larger isomer shifts have been observed for the high-spin Fe(IV) site of the

formally Fe(IV)Fe(III) complex X in ribonucleotide reductase protein R2 (δ = 0.26 mm/s, (Sturgeon et al. 1996)) and the two (putatively high-spin) Fe(IV) sites of the formally

Fe(IV)Fe(IV) complex Q in methane monooxygenase (δ between 0.14 mm/s and 0.21 mm/s for its two sites (Liu et al. 1995; Shu et al. 1997; Shu et al. 1997). These two protein diiron complexes (especially the former) provide encouragement that an isomer shift as high as 0.31 mm/s may be possible if J is a high-spin mononuclear Fe(IV) species. One factor that could contribute to the unusually high isomer shift of the TauD intermediate is electron donation by the ligands, which would confer Fe(III) character to the high-valent site. This notion has previously been advanced to explain the observation that the isomer shifts of the Fe(IV) sites in the aforementioned protein-bound diiron intermediates are invariably greater than those of the most directly comparable inorganic models for them. For example, the formally Fe(IV) site in the protein intermediate, cluster X from ribonucleotide reductase, has an isomer shift (0.26 mm/s) 0.16 mm/s greater than the Fe(IV) site in the putative Fe(III)-O2--Fe(IV)=O model complex with which it shares the same number of valence electrons. The Fe(IV) site of an Fe(III)Fe(IV) model with exclusively carboxylate coordination (as opposed to the pyridine and amine coordination of the first model complex) has a significantly larger isomer shift (0.19 mm/s) (Lee et al. 2002), suggesting that one simple factor that can influence the capacity of the coordination sphere to confer partial Fe(III) character and thereby increase isomer shift is the chemical nature or charge of the ligands. Perhaps the Fe ligands of the TauD intermediate are particularly well-suited to partially absorb the iron high-valency by strong electron donation. Density functional theory calculations suggest that very little electron density is donated from the ligands, rather that the unique isomer shift is influenced by the geometry of the active site proposed in chapter 3. The assignment of a formal oxidation state for J in chapter 3 restricts the possible positions for this species in the catalytic cycle. Within the context of chapter 2 the

oxidation state is poorly defined. The assignment of J as a Fe(III) species allows for evaluation of individual intermediates and would indicate either that J is a species involved in the activation of oxygen such as species I or II, or that J is the post hydroxylation species IV. Assignment of J as a Fe(IV) center would similarly allow resonance forms of I or II, or would also allow assignment as species III which is similar to the proposed oxidant of bleomycin and cytochrome P450. Moieties similar to species I have been shown to accumulate as a Fe(III)-OOH species in activated bleomycin, (Solomon et al. 2000) and in the heme oxygenase cytochrome P450 (Schlichting et al. 2000). Computational studies by Solomon and co-

workers suggest that this initial Fe(II)-O2 adduct would best be described as a high-spin Fe(III) (S = 5/2) antiferromagnetically coupled to a superoxide radical anion (S = 1/2) to yield an S = 2 ground state (Brown et al. 1995). The fast rate of CO2 production argues against assignment of J as species I, although the [O2] dependence studies (Figure 2-10) indicate that small amounts of a precursor to J (i.e. species I) may accumulate. Conceivable resonance formulations for species II would include both Fe(IV)- peroxo and Fe(III)-superoxo contributors. It is not obligatory that an intermediate of this

structure occur in the conversion of the initial Fe(II)-O2 adduct to the presumptive Fe(IV)=O hydroxylating species: the attack of the uncoordinated oxygen on the αKG carbonyl and decarboxylation could be concerted rather than the sequential mechanism shown in Scheme 2-1. However, precedent for this structural motif is found in the coordination chemistry of rhodium (Barbaro et al. 1992; Dutta et al. 2000), iridium (Barbaro et al. 1991; Barbaro et al. 1992), and cobalt (Barbaro et al. 1992). In particular, Rh and Ir complexes, in which the metal is in the +III oxidation state and is facially coordinated by the peroxide and a bidentate quinone hemiketal anion, have been structurally characterized (Barbaro et al. 1991; Dutta et al. 2000). Analogous Fe complexes have been proposed as intermediates in the reaction cycles of several mononuclear non-heme-Fe enzymes (Que et al. 1996; Solomon et al. 2000; Ho et al.

2001). The rate of decarboxylation argues that J cannot be species II. If it were the rate of decarboxylation would correlate with the decay of J. Thus the fast rate of decarboxylation indicates that J is more likely species III. The caveat to this analysis is that the attempted quench of the reaction using denaturants like guanidine or organics like tetrahydrofuran were unable to stop the reaction. So, although the assignment of J as species II is not favored, it remains a possibility because the harsh nature of the quench

might actually facilitate CO2 production from a bicyclic moiety such as II. The chemical characterization of J in chapter 3 allows this ambiguity to be clarified. Structural analogs give some precedent for the placement of a Fe(III) resonance form for both species I and II. The assignment of an Fe(III) J as the hydroxylating intermediate III is again simply a argument over the predominate resonance structure. The highly electronegative oxygen atom would be more electron poor in the formally Fe(III) species. Knowing that four electrons are required for the O-O bond cleavage, two electrons must be donated from the Fe center. Species III is favored for the placement of a Fe(IV) species, here we would expect that CO2 should be produced concurrent with or prior to its formation, and the extensive study of cytochrome P450 suggests that this species would be an effective H atom abstracting unit. Species IV is a likely candidate for a Fe(III) species. Here the hydroxylating Fe(IV)=O has abstracted the substrate hydrogen atom reducing it to a Fe(III)-OH precisely as predicted for the Borovic model (MacBeth et al. 2000). The lifetime of J is surprisingly long compared to what one would expect for a state containing a substrate radical, which in similar systems has been measured to be ~ 50 picoseconds (Auclair et al. 2002). Radical recombination to quench the unstabilized substrate radical would likely be a facile reaction resulting in the reduction of the Fe site to Fe(II), species V. Still the Fe(III)-OH low isomer shift of 0.31 mm/s published by Borovik is very interesting. The effect of an organic radical (S = 1/2) anti-ferromagnetically coupled to the high spin

Fe(III) center (S = 5/2) would give an integer spin species which would result in a quadrupole doublet Mössbauer spectrum similar that observed for J. The oxidation and chemical identity of the second kinetically resolved intermediate M is more easily defined, within the context of this chapter. The oxidation state is well within recognized limits for a high spin Fe(II) complex. The multitude of

Fe(II) species that occur between the radical recombination (IV) and the reaction with O2 (I) are all possibilities for this species. Decay of M to the quaternary complex defines the rate limiting step of the reaction. The fast substrate binding rates indicate that the rate- limiting step of the reaction is associated with product dissociation. The ordered addition of substrates and trapping of the product complex by taurine suggest that M is a succinate complex where 1-hydroxy taurine or its degradation products are removed allowing taurine an opportunity to bind. In conclusion, we have obtained kinetic and spectroscopic evidence for two

intermediates in the reaction of taurine/αKG dioxygenase, TauD, with dioxygen. Importantly, J is the first reported oxidized intermediate for any of the Fe(II)/αKG- dependent enzymes. These intermediates are kinetically competent to be on the taurine hydroxylation pathway. Chemical and spectroscopic characterization of the intermediate J as presented in chapter 2 allows for the definitive assignment of its oxidation state. Kinetic studies to test for temporal correlation between its formation or decay and transformation of an isotopic derivative of taurine allows placement of the species within the TauD catalytic cycle. The study begins the methodical dissection of the TauD reaction. These findings in combination with results in subsequent chapters provide powerful insight into the mechanism of oxygen activation and substrate modification by

the αKG-dependent dioxygenases.

Time (s) TauD•Fe(II)•αKG•taurine J Fe(II) Intermediate formation time course 0.02 48 ± 3 44 ± 3 8 ± 3 0.08 29 ± 6 26 ± 3 42 ± 6 0.2 35 ± 10 11 ± 3 54 ± 10 decay time course 0.02 43 ± 3 48 ± 3 9 ± 3 0.42 53 ± 8 5 ± 3 42 ± 8 1 59 ± 6 < 3 38 ± 6 3 65 ± 5 < 3 32 ± 5 6 69 ± 4 < 3 28 ± 4 8 74 ± 4 < 3 23 ± 4 10 73 ± 4 < 3 24 ± 4

Table 2-1: Relative amounts, given as a percentage of the total iron concentration, of

TauD•Fe(II)•αKG•taurine, J, and the second intermediate, determined by Mössbauer spectroscopy

+ HN3 -

OH OH SO3 2 - 2

OOC OOC- HO2 2+ His99 + α-KG O His99 + taurine His Fe Fe2+ O 2+ 99 HO Asp Asp Fe 2 101 O O 101 O O Asp101 > 3 × 104 M-1 s -1 > 3 × 104 M-1 s -1 His His 255 255 His255 binary complex ternary complex quaternary complex

5 -1 -1 - products + 3 H2 O (1.5 ± 0.2) × 10 M s + O2

+ HN3 + HN3 SO -

3 - -

HO O SO3 -

OOC OOC- . O 2+ His99 O Fe O 3+ His99 Asp101 Fe O O O Asp101 CO2 His255 His255 V I

+ + + HN3 HN3 HN3 - - - -

. SO3 SO3 O SO3

- - - OH - O OOC O OOC OOC 3+ His99 His O His99 O Fe O Fe4+ 99 Fe4+ Asp Asp O O Asp O 101 O 101 101 CO2 CO2 His255 His255 His255 IV III II

Scheme 2-1: The consensus mechanism (Hanauske-Abel et al. 1982) adapted to portray the current working hypothesis for the TauD chemical mechanism. The rate constants refer to the reaction at 5 °C.

TauD Fe(II) αKG Taurine -1 -1 ε318 = 115 M cm -1 -1 ε520 = 200 M cm

-1 5 -1 -1 2 ± 0.5 s O2 (1.5 ± 0.2) M s

M J -1 -1 -1 -1 ε318 = 0 M cm ε318 = 1,615 M cm ε = 0 M-1 cm-1 ε = 170 M-1 cm-1 520 -1 520 13 ± 2 s

Scheme 2-2: Kinetic mechanism used to simulate the stopped-flow data and the kinetics of the first intermediate determined by Mössbauer.

TauD Fe(II) αKG Taurine TauD Fe(II) αKG Taurine -1 -1 -1 -1 ε318 = 0 M cm ε318 = 0 M cm

O2 -1 -1 O2 -1 -1 k = 150,000 M s kfor =1,000,000 M s

-1 kback = 2,200 s

O2 adduct -1 -1 ε318 = 1,500 M cm

k = 415 s-1

J J -1 -1 -1 -1 ε318 = 1,680 M cm ε318 = 1,640 M cm

k = 15 s-1 k = 14.7 s-1

Fe(II) intermediate Fe(II) intermediate -1 -1 -1 -1 ε318 = 0 M cm ε318 = 0 M cm

Scheme 2-3: Kinetic mechanisms used to simulate the data presented in Figure 2-10, the three step mechanism (A) for panel A and the four step mechanism (B) for panel B. The product release step (decay of the Fe(II) intermediate) is omitted from these models because it has almost no effect on the simulations.

0.3 0.16 520

0.2 A

520 0 A 01 0.1 [Fe(II)] mM

0 400 500 600 700 800 Wavelength (nm)

Figure 2-1: Absorption spectra acquired during titration at room temperature of TauD

(0.88 mM) with Fe(II) in the absence of O2 and the presence of 5 mM αKG and 5 mM taurine. The sample prior to the first addition of Fe(II) (containing all other components) was used as spectral reference. The spectra shown correspond to 1.2 mM (green line), 0.86 mM (red line), 0.63 mM (blue line), 0.41 mM (black line with triangular symbols), 0.19 mM (black line with circular symbols) and 0 mM (black line) Fe(II). The points in the inset depict the absorbance at 520 nm (background corrected to the line defined by the absorbances at 318 and 800 nm) as a function of concentration Fe(II). The solid line is a fit of the quadratic equation for binding to the data. The fits from all (6) such titrations gave a mean (± standard deviation) Fe(II)/TauD stoichiometry of 0.87 (± 0.05) and a mean (± standard deviation) dissociation constant of 6 (± 4) µM.

Figure 2-2: Mössbauer spectra of TauD samples recorded at 4.2 K in a 40-mT

magnetic field applied parallel to the γ-beam. A and B are spectra of the binary TauD•Fe(II) complex and the quaternary TauD•Fe(II)•αKG•taurine complex, respectively. These samples contained 2.25 mM TauD, 1.09 mM

57Fe(II) and either (A) 0 or (B) 5 mM αKG and taurine. C through F are spectra of freeze-quenched samples from the reaction of the quaternary

TauD•Fe(II)•αKG•taurine complex with O2. The concentrations after mixing were 1.5 mM TauD, 1.5 mM 57Fe(II), 5 mM αKG, 5 mM taurine and 1.3 mM

O2. The reaction times were (C) 20 ms, (D) 80 ms, (E) 200 ms, and (F) 5 min. The solid line above spectrum C is a theoretical simulation using parameters quoted in the text of the spectrum of the intermediate. The simulated spectrum has been normalized to 46% of the area of spectrum C. The solid line overlaid with spectrum C is the summation of the simulated spectrum at

this percentage with the spectrum of the quaternary TauD•Fe(II)•αKG•taurine complex (50% of B).

0.06 A 0.06 B 530 520 A A ∆

0.03 ∆ 0.03

0 0 0.01 0.1 1 10 0.01 0.1 1 10 Time (s) Time (s)

Figure 2-3: Absorbance-versus-time traces monitoring binding of αKG in the absence (A) and presence (B) of taurine. In A, TauD containing 0.69 equivalents

of Fe(II) was mixed at 5 °C in the absence of O2 with solutions of αKG to give final concentrations of 0.49 mM TauD and 1.5 mM (solid trace), 3 mM (dotted trace), 6 mM (dashed trace), 9 mM (solid trace with open

circles), or 15 mM (solid trace with open squares) αKG. In B, both sets of traces are from experiments in which TauD containing 0.69 equivalents of

Fe(II) was mixed with a solution of αKG to give final concentrations of 0.49 mM TauD and 3 mM (solid traces), 6 mM (dotted traces), or 9 mM

(dashed traces) αKG. For the lower three traces, taurine was present in the TauD•Fe(II) solution such that [taurine] was 5 mM after mixing. In

the upper three traces, taurine was present in the αKG solution, also giving 5 mM taurine after mixing.

A 0.9 B 0.9

0.6 0.6

0.3 0.3 Absorbance Absorbance

0 0 300 400 500 600 700 300 400 500 600 700 Wavelength (nm) Wavelength (nm)

0.15 C 0

0.1 Absorbance -0.02 0.05 500 600 Absorbance Wavelength (nm) 0

300 400 500 600 700 Wavelength (nm)

Figure 2-4: UV-visible absorption spectra acquired after equal-volume mixing at 5 °C

of a solution of 2.4 mM TauD, 2.2 mM Fe(II), 10 mM αKG, and 10 mM

taurine in 50 mM Tris•HCl, pH 7.6, with (A) O2-saturated buffer or (B)

O2-free buffer. Panel C shows the subtraction of the spectra in B from those in A. The spectra were recorded 20 ms (red lines), 68 ms (blue lines), 210 ms (green lines), and 10 s (black lines) after mixing.

0 0 0.16 0.16 A A 520 520 318 318 A A

0 -0.02 0 -0.02

0.01 0.1 1 10 0.01 0.1 1 10 Time (s) Time (s)

Figure 2-5: Kinetic traces from (A) the reaction of Figure 2-4A and (B) the reaction of

the ternary TauD•Fe(II)•αKG complex (in the absence of taurine) with O2- saturated buffer under the same conditions. The symbols depict the experimental absorbances at 318 nm (red circular points, left axis), and 520 nm (blue square points, right axis). The solid lines in A are simulations of the kinetic data according to the parameters summarized in Scheme 2-2 and the text.

0.6

0.4

0.2 [Intermediate] (mM)

0 00.10.20.3 Time (s)

Figure 2-6: Comparison of the concentrations of the first intermediate determined by analysis of spectra C-F of Figure 2-2 (data points) with those predicted by Scheme 2-2 (solid line).

Figure 2-7: Analysis to determine the Mössbauer spectrum of M (A-C) and

comparison to the spectra of the TauD•Fe(II)•αKG•taurine, TauD•Fe(II)•αKG, and TauD•Fe(II) complexes (D-F, respectively). A shows the spectrum of a sample prepared by reacting the

TauD•Fe(II)•αKG•taurine complex with O2 at 5 °C for 200 ms (hash marks). Concentrations after mixing were 1.5 mM TauD, 1.5 mM 57Fe, 5

mM αKG, 5 mM taurine, and 1.3 mM O2. The solid line is a reference spectrum of J, scaled to 11 % of the total absorption. Removal of the contribution of J to the spectrum in A results in the hash-marked spectrum

in B. The solid line is the spectrum of the TauD•Fe(II)•αKG•taurine complex, scaled to the same intensity. Removal of 35 % of the spectrum

of the TauD•Fe(II)•αKG•taurine complex results in the reference spectrum for the second intermediate, displayed as hash marks in C. The

solid line overlaid is a quadrupole doublet with the parameters given in the text and Table 2-1. In D-F, the hash marked spectrum in each case is the reference spectrum of the second intermediate. The solid line is the

spectrum of the TauD•Fe(II)•αKG•taurine complex (D), the TauD•Fe(II)•αKG complex (E), or the TauD•Fe(II) complex (F) scaled to the same intensity. All spectra were collected at 4.2 K in a 40-mT magnetic field.

A B

0.1 40 70% Fe -0.04 480 % Fe 600 -A 0.05 A 35 -0.08

0 0 0 0.01 0.1 1 10 0.01 0.1 1 10 Time (s) Time (s)

Figure 2-8: Time dependence of the relative amounts of the second intermediate (A)

and the TauD•Fe(II)•αKG•taurine complex (B) obtained from analysis of stopped-flow absorption (left axis) and Mössbauer (right axis) spectroscopic experiments. The negative of the absorbance changes

monitored at 480 nm (-∆A480, green line in A) correlates with the concentration of the second intermediate. The absorbance change at 600

nm (A600, red line in B) correlates with the concentration of the TauD•Fe(II)•αKG•taurine complex. Relative amounts obtained from analysis of the Mössbauer data are shown as circles (decay time course) and triangles (formation time course) in A and B. The error bars reflect the

results obtained from analyses using reference spectra for the second intermediate, which were generated by removing either more (45 %) or less (25 %) of the quaternary complex than its estimated amount (35 %). For both plots, the scales were determined as described in the text. The final concentrations in the stopped-flow experiment were 2 mM TauD, 1.4

mM Fe(II), 5 mM αKG, 5 mM taurine, 0.9 mM O2, and those in the Mössbauer experiment are given in the legend of Figure 2-7.

2 CO

14 0.9

0.6

0.3

Fraction of total total Fraction of 0 0 80 160 240 320 Time (ms)

14 14 Figure 2-9: Kinetics of production of CO2 from 1-[ C]-αKG in single-turnover, chemical-quenched-flow experiments in which the

TauD•Fe(II)•αKG•taurine complex in buffer A was mixed at 5 °C with

O2-containing buffer A. The concentrations after mixing were: 1.1 mM TauD, 1 mM Fe(II), 0.06 mM αKG, 5 mM taurine. The circular points are

average values from experiments using either 4 N NaOH or 2 N H2SO4 as the quench solution. The error bars are the standard deviations for each reaction time in 5 independent trials. The solid line is a fit of the equation for an exponential increase and corresponds to k = 42 s-1. From this value

and the O2 concentration of 0.26 mM, a second-order rate constant of 1.6 x 105 M-1s-1 is calculated.

A B 0.12 150 0.12 obs k 318 0 318 A 01 A ∆ [O ] mM ∆ 0.06 2 0.06

0 0 0.01 0.1 0.01 0.1 Time (s) Time (s)

Figure 2-10: Absorbance-versus-time traces (318 nm) after mixing at 5 °C of an O2- free solution containing 0.72 mM TauD, 0.5 mM Fe(II), 0.1 mM αKG, and 4.28 mM taurine in buffer A (50 mM Tris•HCl, pH 7.6) with an equal

volume of buffer A containing varying concentrations of O2. The symbols

are the data from experiments with final O2 concentrations of 0.15 mM (squares), 0.3 mM (triangles), 0.6 mM (circles) and 0.95 mM (crosses). The solid lines are from simulations that are described in the text and that employed the parameters shown in Table 2-1. Panel A shows simulations according to the three-step minimal kinetic mechanism (Scheme 2-2) and panel B shows simulations based on a four-step mechanism that includes a

dissociable initial O2 complex as a precursor to J.

0.18 318

A 0.12 ∆

0.06

0 0.001 0.01 0.1 Time (s)

Figure 2-11: Absorbance-versus-time traces (318 nm) after mixing at 5 °C of a one

volume of 0.98 mM O2 containing buffer A (50 mM Tris•HCl, pH 7.6)

with an 2.5 volume equivalents of protein solution containing varying

amounts of quaternary complex. The symbols are the data from

experiments with final quaternary complex concentrations of 0.5 mM

(black squares), 1 mM (green triangles), and 2 mM (blue circles). The

solid lines are from simulations that are described in the text and that

employed the parameters shown in Scheme 2-3.

0.16 318 A ∆ 0.08

0 0.01 0.1 1 Time (s)

Figure 2-12: Absorbance-versus-time traces after mixing at 5 °C of TauD•Fe(II)

complex with O2, αKG and taurine. Final concentrations after mixing

were 0.5 mM TauD•Fe(II), 0.2 mM O2, 5 mM αKG and either 0 mM taurine (squares) or 5 mM taurine (circles). The black line is a simulation according to the three-step kinetic mechanism (Scheme 2-2), which assumes pre-formed reactive quaternary complex. The blue line is a simulation according to an expanded mechanism that places a 3 s-1 step for

quaternary complex formation prior to O2 addition. The red line is a simulation according to this same expanded mechanism, but with an effective rate constant of 150 s-1 for formation of the quaternary complex.

A 0 B 1 ) -1 520 A ∆ v/[E] (s v/[E] -0.02

0 0.01 0.1 1 10 100 024 Time (s) Taurine (mM)

Figure 2-13: Inhibition of (A) regeneration of the quaternary

TauD•Fe(II)•αKG•taurine complex after a single turnover and (B) catalysis in the steady state by increasing concentrations of taurine. A:

∆A520-versus-time traces after mixing at 5 °C of a solution containing 1.2 mM TauD, 1 mM Fe(II), 10 mM αKG, and 2 mM (black), 6 mM (red), 10 mM (blue), or 40 mM taurine (green) with an equal volume of air-

saturated buffer A. B: Inhibition of steady-state turnover at 5 °C by high concentrations of taurine at RV = 1.6 (circles) and RV = 4.8 (squares). The assay conditions were the same as in Figure 2-4, except that the

taurine concentration was varied as indicated.

References

Arciero, D. M., Lipscomb, J. D., Huynh, B. H., Kent, T. A. and Münck, E. (1983).

"EPR and Moessbauer studies of protocatechuate 4,5-dioxygenase.

Characterization of a new ferrous ion environment." J. Biol. Chem. 258(24):

14981-14989.

Auclair, K., Hu, Z., Little, D. M., Ortiz de Montellano, P. R. and Groves, J. T. (2002).

"Revisiting the Mechanism of P450 Enzymes with the Radical Clocks

Norcarane and Spiro[2,5]octane." J. Am. Chem. Soc. 124(21): 6020-6027.

Baldwin, J., Krebs, C., Ley, B. A., Edmondson, D. E., Huynh, B. H. and Bollinger, J.

M., Jr. (2000). "Mechanism of Rapid Electron Transfer during Oxygen

Activation in the R2 Subunit of Escherichia coli Ribonucleotide Reductase. 1.

Evidence for a Transient Tryptophan Radical." J. Am. Chem. Soc. 122(49):

12195-12206.

Baldwin, J. E., Adlington, R. M., Crouch, N. P., Keeping, J. W., Leppard, S. W.,

Pitlik, J., Schofield, C. J., Sobey, W. J. and Wood, M. E. (1991). "Studies on

the mechanism of deacetoxy-deacetylcephalosporin C synthase using

cyclopropyl substituted cephalosporin probes." J. Chem. Soc. Chem.

Commun.(11): 768-70.

Barbaro, P., Bianchini, C., Linn, K., Mealli, C., Meli, A., Vizza, F., Laschi, F. and

Zanello, P. (1992). "Dioxygen uptake and transfer by cobalt(III), rhodium(III)

and iridium(III) catecholate complexes." Inorg. Chim. Acta 198-200: 31-56.

Barbaro, P., Bianchini, C., Mealli, C. and Meli, A. (1991). "Synthetic models for

catechol 1,2-dioxygenases. Interception of a metal catecholate-dioxygen

adduct." J. Am. Chem. Soc. 113(8): 3181-3183.

Bollinger, J. M., Jr., Tong, W. H., Ravi, N., Huynh, B. H., Edmonson, D. E. and

Stubbe, J. (1994). "Mechanism of Assembly of the Tyrosyl Radical-Diiron(III)

Cofactor of E. coli Ribonucleotide Reductase. 2. Kinetics of The Excess Fe2+

Reaction by Optical, EPR, and Moessbauer Spectroscopies." J. Am. Chem.

Soc. 116(18): 8015-23.

Brown, C. A., Pavlosky, M. A., Westre, T. E., Zhang, Y., Hedman, B., Hodgson, K.

O. and Solomon, E. I. (1995). "Spectroscopic and Theoretical Description of

the Electronic Structure of S = 3/2 Iron-Nitrosyl Complexes and Their

Relation to O2 Activation by Non-Heme Iron Enzyme Active Sites." J. Am.

Chem. Soc. 117(2): 715-732.

Burger, R. M., Kent, T. A., Horwitz, S. B., Münck, E. and Peisach, J. (1983).

"Moessbauer study of iron bleomycin and its activation intermediates." J. Biol.

Chem. 258(3): 1559-1564.

Debrunner, P. G. (1989). "Mössbauer spectroscopy of iron porphyrins." Phys.

Bioinorg. Chem. Ser. 4(Iron Porphyrins, Pt. 3): 137-234.

Dutta, S., Peng, S.-M. and Bhattacharya, S. (2000). "Ligand Control on Molecular

Oxygen Activation by Rhodium Quinone Complexes." Inorg. Chem. 39(10):

2231-2234.

Eichhorn, E., Van Der Ploeg, J. R., Kertesz, M. A. and Leisinger, T. (1997).

"Characterization of α-ketoglutarate-dependent taurine dioxygenase from

Escherichia coli." J. Biol. Chem. 272(37): 23031-23036.

Gill, S. C. and von Hippel, P. H. (1990). "Calculation of protein extinction

coefficients from amino acid sequence data [Erratum to document cited in

CA111(25):228358u]." Analyt. Biochem. 189(2): 283.

Grapperhaus, C. A., Mienert, B., Bill, E., Weyhermüller, T. and Wieghardt, K. (2000).

"Mononuclear (Nitrido)iron(V) and (Oxo)iron(IV) Complexes via Photolysis

of [(cyclam-acetato)FeIII(N3)]+ and Ozonolysis of [(cyclam-

acetato)FeIII(O3SCF3)]+ in Water/Acetone Mixtures." Inorg. Chem. 39(23):

5306-5317.

Greenwood, N. N. and Gibb, T. C. (1971). Moessbauer Spectroscopy. New York,

Wiley.

Hanauske-Abel, H. M. and Günzler, V. (1982). "A stereochemical concept for the

catalytic mechanism of prolylhydroxylase. Applicability to classification and

design of inhibitors." J. Theor. Biol. 94(2): 421-55.

Hausinger, R. P. (2004). "Fe(II)/α-ketoglutarate-dependent hydroxylases and related

enzymes." Crit. Rev. Biochem. Mol. Biol. 39(1): 21-68.

Hitchman, M. L. (1978). Chemical Analysis, Vol. 49: Measurement of Dissolved

Oxygen. New York, Wiley.

Ho, R. Y. N., Mehn, M. P., Hegg, E. L., Liu, A., Ryle, M. J., Hausinger, R. P. and

Que, L., Jr. (2001). "Resonance Raman studies of the iron(II)-alpha-keto acid

chromophore in model and enzyme complexes." J. Am. Chem. Soc. 123(21):

5022-5029.

Jones, M. E., Kavipurapu, P. R. and Traut, T. W. (1978). "Orotate

phosphoribosyltransferase: orotidylate decarboxylase (Ehrlich ascites cell)."

Meth. Enzymology 51: 155-67.

Kivirikko, K. I. and Prockop, D. J. (1967). "Enzymic hydroxylation of proline and

lysine in protocollagen." Proc. Nat. Acad. Sci. U.S.A. 57(3): 782-9.

Lee, D., Pierce, B., Krebs, C., Hendrich, M. P., Huynh, B. H. and Lippard, S. J.

(2002). "Functional Mimic of Dioxygen-Activating Centers in Non-Heme

Diiron Enzymes: Mechanistic Implications of Paramagnetic Intermediates in

the Reactions between Diiron(II) Complexes and Dioxygen." J. Am. Chem.

Soc. 124(15): 3993-4007.

Lim, M. H., Rohde, J.-U., Stubna, A., Bukowski, M. R., Ho, R. Y., Münck, E., Nam,

W. and Que, L., Jr. (2003). "An FeIV=O complex of a tetradentate tripodal

nonheme ligand." Proc. Nat. Acad. Sci. USA: accepted.

Liu, K. E., Valentine, A. M., Wang, D., Huynh, B. H., Edmondson, D. E., Salifoglou,

A. and Lippard, S. J. (1995). "Kinetic and spectroscopic characterization of

intermediates and component interactions in reactions of methane

monooxygenase from Methylococcus capsulatus (Bath)." J. Am. Chem. Soc.

117(41): 10174-10185.

MacBeth, C. E., Golombek, A. P., Young, V. G., Jr., Yang, C., Kuczera, K., Hendrich,

M. P. and Borovik, A. S. (2000). "O2 activation by nonheme iron complexes:

A monomeric Fe(III)-Oxo complex derived from O2." Science 289(5481):

938-41.

Meyer-Klaucke, W., Winkler, H., Schünemann, V., Trautwein, A. X., Nolting, H. F.

and Haavik, J. (1996). "Mössbauer, electron paramagnetic resonance, and x-

ray absorption fine-structure studies of the iron environment in recombinant

human tyrosine hydroxylase." Eur. J. Biochem. 241(2): 432-439.

Neese, F., Zaleski, J. M., Zaleski, K. L. and Solomon, E. I. (2000). "Electronic

Structure of Activated Bleomycin: Oxygen Intermediates in Heme versus

Non-Heme Iron." Journal of the American Chemical Society 122(47): 11703-

11724.

O'Brien, J. R., Schuller, D. J., Yang, V. S., Dillard, B. D. and Lanzilotta, W. N.

(2003). "Substrate-induced conformational changes in Escherichia coli

taurine/α-Ketoglutarate dioxygenase and insight into the oligomeric

structure." Biochemistry 42(19): 5547-5554.

Que, L., Jr. and Ho, R. Y. N. (1996). "Dioxygen activation by enzymes with

mononuclear non-heme iron active sites." Chem. Rev. 96(7): 2607-2624.

Ravi, N., Bollinger, J. M., Jr., Huynh, B. H., Stubbe, J. and Edmondson, D. E. (1994).

"Mechanism of Assembly of the Tyrosyl Radical-Diiron(III) Cofactor of E.

Coli Ribonucleotide Reductase: 1. Moessbauer Characterization of the

Diferric Radical Precursor." J. Am. Chem. Soc. 116(18): 8007-8014.

Rohde, J.-U., In, J.-H., Lim, M. H., Brennessel, W. W., Bukowski, M. R., Stubna, A.,

Münck, E., Nam, W. and Que, L., Jr. (2003). "Crystallographic and

Spectroscopic Characterization of a Nonheme Fe(IV)=O complex." Science

299(5609): 1037-1039.

Ryle, M. J., Padmakumar, R. and Hausinger, R. P. (1999). "Stopped-flow kinetic

analysis of Escherichia coli taurine/α-ketoglutarate dioxygenase: Interactions

with α-ketoglutarate, taurine, and oxygen." Biochemistry 38(46): 15278-

15286.

Sam, J. W., Tang, X.-J. and Peisach, J. (1994). "Electrospray Mass Spectrometry of

Iron Bleomycin: Demonstration That Activated Bleomycin Is a Ferric

Peroxide Complex." J. Amer. Chem. Soc. 116(12): 5250-6.

Schlichting, I., Berendzen, J., Chu, K., Stock, A. M., Maves, S. A., Benson, D. E.,

Sweet, R. M., Ringe, D., Petsko, G. A. and Sligar, S. G. (2000). "The catalytic

pathway of cytochrome P450cam at atomic resolution." Science 287(5458):

1615-1622.

Shu, L., Nesheim, J. C., Kauffmann, K., Münck, E., Lipscomb, J. D. and Que, L., Jr.

(1997). "An Fe2IVO2 diamond core structure for the key intermediate Q of

methane monooxygenase." Science 275(5299): 515-518.

Shu, L., Nesheim, J. C., Kauffmann, K., Münck, E., Lipscomb, J. D. and Que, L., Jr.

(1997). "An Fe(IV)2O2 diamond core structure for the key intermediate Q of

methane monooxygenase." Science 275(5299): 515-518.

Simonneaux, G., Scholz, W. F., Reed, C. A. and Lang, G. (1982). "Mössbauer spectra

of unstable iron porphyrins. Models for compound II of peroxidase." Biochim.

Biophys. Acta 716(1): 1-7.

Solomon, E. I., Brunold, T. C., Davis, M. I., Kemsley, J. N., Lee, S. K., Lehnert, N.,

Neese, F., Skulan, A. J., Yang, Y.-S. and Zhou, J. (2000). "Geometric and

electronic structure/function correlations in non-heme iron enzymes." Chem.

Rev. 100: 235-349.

Sturgeon, B. E., Burdi, D., Chen, S., Huynh, B.-H., Edmondson, D. E., Stubbe, J. and

Hoffman, B. M. (1996). "Reconsideration of X, the diiron intermediate

formed during cofactor assembly in E. coli ribonucleotide reductase." J. Am.

Chem. Soc. 118(32): 7551-7557.

Wolgel, S. A., Dege, J. E., Perkins-Olson, P. E., Juarez-Garcia, C. H., Crawford, R.

L., Münck, E. and Lipscomb, J. D. (1993). "Purification and characterization

of protocatechuate 2,3-dioxygenase from Bacillus macerans: a new extradiol

catecholic dioxygenase." J. Bacteriol. 175(14): 4414-4426.

Wu, M., Moon, H.-S., Begley, T. P., Myllyharju, J. and Kivirikko, K. I. (1999).

"Mechanism-Based Inactivation of the Human Prolyl-4-hydroxylase by 5-

Oxaproline-Containing Peptides: Evidence for a Prolyl Radical Intermediate."

J. Am. Chem. Soc. 121(3): 587-588.

Zheng, H., Yoo, S. J., Münck, E. and Que, L., Jr. (2000). "The Flexible Fe2(µ-O)2

Diamond Core: A Terminal Iron(IV)-Oxo Species Generated from the

Oxidation of a Bis(µ-oxo)diiron(III) Complex." J. Am. Chem. Soc. 122(15):

3789-3790.

Chapter 3: Characterization of two accumulating intermediates in the catalytic cycle of taurine•α-ketoglutarate dioxygenase

Portions of this chapter are published in:

John C. Price, Eric W. Barr, Timothy E. Glass, Carsten Krebs and J. Martin Bollinger, Jr., 2003, “Evidence for Hydrogen Abstraction from C1 of Taurine by the High-Spin

Fe(IV) Intermediate Detected during Oxygen Activation by Taurine:α-Ketoglutarate Dioxygenase (TauD) ,” J. Am. Chem. Soc., 43, 13008-13009.

John C. Price, Eric W. Barr, Bhramara Tirupati, J. Martin Bollinger, Jr., and Carsten Krebs, 2003, “The First Direct Characterization of a High Valent Iron Intermediate in the

Reaction of an α-Ketoglutarate-Dependent Dioxygenase: a Fe(IV) Complex in Taurine/α-Ketoglutarate Dioxygenase (TauD) from Escherichia coli,” Biochemistry, 42, 7497-7508.

John C. Price, Eric W. Barr, Lee M. Hoffart, Carsten Krebs, and J. Martin Bollinger, Jr.,

2005, “Kinetic Dissection of the Catalytic Mechanism of Taurine:α-Ketoglutarate Dioxygenase (TauD) from Escherichia coli”, Biochemistry, 44, 8138-8147.

Abstract In this work, we demonstrate that substitution of the C1 hydrogen atoms of taurine with deuterium slows decay of the oxidized iron intermediate, J, by a factor of ~ 37 without affecting the rates of the two other resolved steps of the minimal mechanism. This allows a greater percentage of the total iron content to accumulate as the species J (Scheme 3-1). A kinetic isotope effect (KIE) of this magnitude requires that decay of the intermediate occur concomitantly with, or be tightly kinetically coupled to, cleavage of the C1-H bond and suggests that hydrogen tunneling contributes to the transfer of the hydrogen atom. Further spectroscopic characterization reveals the identity of this first kinetically resolved intermediate, J, as a formally Fe(IV), high spin, S = 2 complex. An independent resonance Raman study by another group indicated that J contains a Fe(IV)-oxo unit. This was confirmed in an XAS study by our group, in which a short, 1.62 Å, interaction between the Fe-center of J and one of its ligands was detected. This short interaction strongly corroborated the presence of the Fe(IV)-oxo group. The combined results prove that J is a novel Fe(IV)=O complex, that is either the hydroxylating intermediate or rapidly and reversibly converts to the hydroxylating species. Chemical logic dictates that the former possibility is more likely.

2 Solvent viscosity studies in combination with the 1,1-[ H]2-taurine analog confirm that the identity of the rate limiting step of the reaction is product release. This also corroborates that the second kinetically resolved intermediate is a product complex.

Introduction In the study of TauD, the previous chapter gave strong evidence for the accumulation of two intermediates in the reaction with oxygen. Although transient state techniques allow the range of possible identities to be narrowed, conclusive assignment of the positions of these species in the catalytic cycle was lacking.

One of the strengths in using the TauD system as a model for the αKG dependent dioxygenase family is that the substrate is small enough to allow analogues to be synthesized. The most successful analog employed in the study of the TauD system is

2 1,1-[ H]2-taurine (D-taurine). In studying the transfer of atoms during a chemical reaction the use of isotopomers is an extremely useful tool. The substitution of a protium with deuterium doubles the mass and causes a very large chemically inert mass perturbation. A kinetically resolved step involving transfer of the atom in question will show a very diagnostic change in the rate upon isotopic substitution (Westheimer 1961). The work of Wu and co-workers (Wu et al. 1999) indicated that a radical intermediate is

employed in the Fe(II)•αKG dioxygenase family. This is in agreement with the mechanism of Hanauske-abel and co-workers (Hanauske-Abel et al. 1982) who postulated that a hydrogen atom is abstracted and a substrate radical is formed. Because of this necessity for H-atom removal from the substrate we would expect a reduction in the rate of hydroxylation upon deuterium substitution of the taurine C1 protons. This slowing of the transfer rate would favor accumulation of the C1-H cleaving intermediate,

perhaps allowing its direct detection. The rate of CO2 production presented in chapter 2 leads us to believe that J is either species III or IV from scheme 3-1. A change in the rate of formation for the 318 nm band will be evidence that J is IV. If we see a change in the rate of decay for the 318 nm band we have evidence that J is III.

Materials and Methods

Materials. Chapter one gives a comprehensive list of the chemicals and materials used for experiments described in this chapter. Methods. The preparation and quality control procedures for the TauD used during the series of experiments described in this chapter is as described in chapter one. The stopped-flow and freeze-quench experiments described in this chapter and the analysis of these data were carried out in the same manner as described in chapter one.

2 Synthesis of 1,1-[ H]2-taurine. Deuterium-substituted taurine was synthesized from chlorosulfonylacetyl chloride according to Scheme 3-3. Conversion to the phenoxysulfonyl-acetic acid phenyl ester (step A) was accomplished by a procedure similar to that of Hoogenboom and co-workers. (Hoogenboom et al. 1959). Starting material (4.0 g, 23 mmol) was added to phenol (4.7 g, 50 mmol) under an argon atmosphere. Toluene (3 mL) was added to the mixture, followed by refluxing for 12 h. The toluene was removed under vacuum and the product was purified by chromatography on silica. After adsorption of the crude product, the column was flushed with 1 volume of hexane to remove any residual toluene, and the product was eluted with dichloromethane/hexane (60 % dichloromethane by volume). After removal of the solvent under vacuum, the product was obtained as a cream colored solid (5.4 g, 19

1 mmol). H-NMR (22° C, 300 MHz, CDCl3): 4.37 ppm singlet (2H), 7.15 ppm multiplet

13 (2H), 7.36 ppm multiplet (8H). C-NMR (22° C, 75 MHz, CDCl3): 54.4 ppm singlet, 121.5 ppm singlet, 122.6 ppm singlet, 127.2 ppm singlet, 128.3 ppm singlet, 130.1 ppm singlet, 130.6 ppm singlet, 149.7 ppm singlet, 150.5 ppm singlet, 160.7 ppm singlet. MS: m/z = 293 (MH+). Phenoxysulfonyl-acetic acid phenyl ester was converted to carbamoyl- methanesulfonic acid phenyl ester (Scheme 3-3, step B) by the procedure of

Hoogenboom and co-workers (Hoogenboom et al. 1959). Ammonia (~ 30 mL) was condensed from the gas phase onto the phenoxysulfonyl-acetic acid phenyl ester (4.0 g,

14 mmol), and the mixture was stirred overnight at -35 °C. Ethanol (~ 10 mL of 100 %) was added to the solution, which was refluxed for 30 min and then adsorbed onto silica. The product was purified by column chromatography with 1:1 (v:v) ethylacetate/hexane as the mobile phase. After removal of solvent under vacuum, carbamoyl-methanesulfonic acid phenyl ester was obtained as a white solid (2.8 g, 13 mmol). 1H-NMR (22° C, 300

MHz, CDCl3): 4.2 ppm singlet (2H), 5.8 ppm singlet (1H), 6.6 ppm singlet (1H), 7.4 ppm

13 multiplet (5H). C-NMR (22° C, 75 MHz, CDCl3): 55.2 ppm singlet, 122.6 ppm singlet, 128.4 ppm singlet, 130.7 ppm singlet, 149.5 ppm singlet, 161.9 ppm singlet. MS: m/z = 216 (MH+). Carbamoyl-methanesulfonic acid phenyl ester was reduced to 2-aminoethane-1- sulfonic acid phenyl ester (Scheme 3-3, step C) with a borane-methyl sulfide complex. Carbamoyl-methanesulfonic acid phenyl ester (2.6 g, 12 mmol) was dissolved under an argon atmosphere in 200 mL tetrahydrofuran, which had been distilled from sodium and benzophenone. This solution was cooled to 0 °C, and borane-dimethyl sulfide (20 mL, 40 mmol) was dripped in as a 2 M solution in tetrahydrofuran. This solution was then heated to reflux, stirred for two hours, and allowed to cool. Ethylacetate (100 mL) was added, and the solution was washed with saturated aqueous solutions of ammonium chloride and then sodium chloride. The organic phase was dried over anhydrous magnesium sulfate and reduced to a clear viscous oil under vacuum. This oil was then stirred in diethylamine a minimum of 4 h, after which the diethylamine was removed under vacuum. The resulting white solid was purified by column chromatography on silica using 2 % methanol in dichloromethane as the mobile phase. The product (1.2 g, 6.0 mmol) was collected as a clear oil after removal of solvents under vacuum. 1H-NMR

(22° C, 300 MHz, CDCl3): 1.4 ppm singlet (2H), 3.4 ppm multiplet (4H), 7.3 ppm

13 multiplet (5H). C-NMR: (22° C, 75 MHz, CDCl3) 37.4 ppm singlet, 54.1 ppm singlet,

122.5 ppm singlet, 127.8 ppm singlet, 130.4 ppm singlet, 149.5 ppm singlet. MS: m/z = 202 (MH+). Exchange of the C1 protia of 2-aminoethane-1-sulfonic acid phenyl ester for deuteria was effected by treatment with sodium t-butoxide in O-[2H]-t-butanol (Scheme 3-3, step D). 2-Aminoethane-1-sulfonic acid phenyl ester (0.50 g, 2.5 mmol) was dissolved in 30 mL of O-[2H]-t-butanol (99 % isotopic enrichment) under an argon atmosphere. Sodium t-butoxide (0.096 g, 1.0 mmol) was added to this solution, which

2 was stirred at 60 °C overnight. H2O (60 mL, 99.9 % isotopic enrichment) was added to quench the exchange reaction. This mixture was extracted with dichloromethane, and the t-butanol layer was reduced under vacuum to a pale yellow, viscous oil. This material was carried forward without further purification. The contaminant contributing the pale yellow color could be removed by column chromatography on silica gel using 2 %

2 methanol in dichloromethane as mobile phase. After removal of the solvent, 1,1-[ H]2-2- amino-ethanesulfonic acid phenyl ester (0.30 g, 1.5 mmol) was obtained as a clear,

1 viscous oil. H-NMR (22° C, 300 MHz, CDCl3): 1.4 ppm singlet (2H), 3.3 ppm singlet

13 (2H), 7.3 ppm multiplet (5H). C-NMR (22° C, 75 MHz, CDCl3): 37.4 ppm singlet, 122.5 ppm singlet, 127.8 ppm singlet, 130.4 ppm singlet, 149.5 ppm singlet. MS: m/z = 204 (MH+).

2 1,1-[ H]2-2-aminoethane-1-sulfonic acid phenyl ester was converted to the free acid via hydrogenolysis of the benzyl carbon-oxygen bond (Scheme 3-3, step E) in a procedure similar to that described by Sturm et al. (Sturm et al. 1983). The ester (0.30 g, 1.5 mmol) was dissolved in 12 mL tetrahydrofuran, which had been distilled from

2 sodium and benzophenone. To this solution was added 1.8 mL [ H]5-ammonium

2 hydroxide (26 % in H2O) and 0.15 g 10 % palladium on carbon (Pd/C) catalyst. This mixture was then capped and stirred for 3 h under just greater than 1 atm H2(g). After filtration, the solution was extracted with dichloromethane. Lyophilization of the

2 aqueous layer afforded 1,1-[ H]2-2-aminoethane-1-sulfonic acid (0.15 g, 1.0 mmol) as a

1 fine white powder. H-NMR (22° C, 300 MHz, D2O): 3.05 ppm triplet (2H), 3.3 ppm

13 triplet (2H). C-NMR (22° C, 75 MHz, D2O): 47.7 ppm singlet, 35.7 ppm singlet. HRMS: m/z = 126.0198 (M-). High Field Mössbauer Spectroscopy and Data Analysis. For high-field spectra, the sample was kept inside a 12SVT dewar (Janis), which houses a superconducting magnet that allows for application of variable magnetic fields between 0 and 8T oriented

parallel to the γ-beam. The isomer shifts quoted are relative to the centroid of the spectrum of a metallic foil of α-Fe at room temperature. Data analysis was performed using the program WMOSS from WEB research. EPR spectroscopy. EPR spectra were recorded on an ESP300 spectrometer from Bruker (Billerica, MA) equipped with a ER 041 MR Microwave Bridge and a 4102ST X- band Resonator from Bruker. The temperature was maintained at 77 K by immersion of the sample in liquid nitrogen in a "finger dewar."

Cryoreduction by low-temperature γ-radiolysis. Freeze-quenched 20-ms samples suitable for EPR and Mössbauer spectroscopy were irradiated in the γ-irradiation facility of the Breazeale nuclear reactor at Penn State University using a 60Co-source (activity 26

krad/h). A total dose of ca. 1.4 Mrad was delivered. During the γ-irradiation the temperature of the samples was maintained at 77 K by immersion in liquid nitrogen.

Results

2 Activation of oxygen by the 1,1-[ H]2-taurine (D-taurine) containing TauD•Fe(II)•αKG•taurine quaternary complex. Chemical quench-flow experiments

allowed the timing of CO2 release to be determined for the reaction of

2 TauD•Fe(II)•αKG•1,1-[ H]2-taurine with O2. Quenching with 4 M sodium hydroxide,

-1 gave an observed rate of 40 s (Figure 3-1). At the experimental concentration of O2 (0.266 mM), this corresponds to a second order rate constant of 1.5 x 105 M-1s-1. This

1 agrees well with the rate constant for CO2 production in the reaction with the 1,1-[ H]2-

taurine (H-taurine) substrate, indicating that O2 activation is unimpeded by the substitution with deuterium.

Identification of J as the hydroxylating intermediate. The reaction of O2 and the quaternary TauD•Fe(II)•αKG•D-taurine complex results in two key perturbations of the time dependent UV-visible absorption spectra. First, the transient, positive feature at 318 nm ascribed to the intermediate J shows an increased amplitude and a later time of maximum absorbance (Figure 3-2). Second, the transient minimum seen at 520 nm associated with loss of the αKG charge-transfer band is no longer apparent (Figure 3-2 inset). In the reaction using H-taurine the tailing absorbance of the 318 nm absorbance band, due to J, masks the decay of the 520 nm band until J has decayed to form M. Now the decay of M is faster than its formation. Because M does not accumulate, the decay of the quaternary complex is now kinetically masked at 520 nm. The decay at 600 nm is still apparent (see inset to Figure 3-2). As was demonstrated in chapter 2 (see Figure 2-8), at 600 nm the absorbency of J is only ~ 23 % of that for the quaternary complex. Thus 600 nm is a better indicator than 520 nm of the transient state concentration of quaternary

complex. In agreement with the CO2 production data, the absorbance at 600 nm shows that the metal-to-ligand charge transfer (MLCT) band is destroyed during the reaction with O2.

As a control for the D-taurine studies, the reaction of O2 and the quaternary

1 TauD•Fe(II)•αKG•1,1-[ H]2-taurine complex was monitored. The quaternary complex

1 was formed with synthetic 1,1-[ H]2-taurine (H-taurine) obtained by a procedure mimicking that used to prepare the D-taurine. These results agree fully with the results obtained when the quaternary complex contained commercial taurine. Difference spectra at short times (less than 0.030 s) in the otherwise-identical reaction employing the synthetic D-taurine are very similar and are also dominated by the positive 318-nm feature J (Figure 3-2). However, direct comparison of the kinetics of this feature shows that it persists for much longer in the D-taurine reaction (Figure 3-3). The amplitude

(open circles and squares in Figure 3-4) is nearly twice as great (under these conditions) and the decay phase much slower with D-taurine (squares). These stopped-flow data, as well as all other stopped-flow and freeze-quench kinetic data from experiments carried out under different reaction conditions (see below), can be accounted for by the same kinetic scheme simply by adjustment of the rate constant for decay of the J from (13 ± 2) s-1 for H-taurine to (0.35 ± 0.05) s-1 for D-taurine (see Scheme 3-2), setting limits of 28 < kH/kD < 50 on the kinetic isotope effect (KIE), if error limits and uncoupling (vide infra, Scheme 3-4) are taken into account.

2- Steady state uncoupling of CO2 and SO3 production. The steady-state rates of decarboxylation of αKG and release of sulfite from taurine (the ultimate result of C1 hydroxylation) were determined for both H-taurine and D-taurine. With the former, a kcat of 1.3 ± 0.2 s-1 was obtained, irrespective of which product was quantified (Scheme 2-4). As previously noted, this value agrees well with that predicted by Scheme 2-2, taking into account (1) that our preparations of TauD bind 0.87 equivalents of Fe(II) and (2) that

apparently only 72% of the incorporated Fe(II) is competent to activate O2. The agreement of the two assays establishes that the reaction is, within the uncertainty of our measurements, completely coupled under these conditions with H-taurine. With D-

-1 taurine, a value of 0.11 ± 0.02 s was obtained for CO2 production, whereas the rate constant for sulfite production was 0.07 ± 0.01 s-1. Slower sulfite production suggests that the D-taurine reaction is partially uncoupled as a result of the very large KIE. This uncoupling could explain why the observed steady-state rate constant for decarboxylation is less than the value of (0.19 ± 0.04) s-1 predicted by Scheme 2-2: uncoupled decay may produce an off-pathway form of the enzyme (e.g., an Fe(III)-containing form by one- electron reduction of the Fe(IV) intermediate) that may be slow to re-enter the catalytic cycle. The occurrence of uncoupling would also imply that the measured KIE

underestimates the true KIE on C1-H bond cleavage, because kobs for decay of the Fe(IV) complex in the D-taurine reaction would represent the sum of the rate constants for

coupled and uncoupled decay, whereas it is likely that only the former involves C1-H cleavage. Thus, the true KIE could be as large as 58 (37 x 0.11/0.07), or larger if error limits are considered. Freeze-quench Mössbauer The freeze-quench Mössbauer method was used to confirm that the observed effect on decay of the 318-nm absorption is indeed a reflection of a large KIE on decay of the intermediate J. Samples were prepared by rapidly freezing the reaction solution after 0.021 s, 0.084 s, and 0.16 s for the H-taurine reaction and after 0.084 s, 2.0 s, and 5.0 s for the otherwise-identical D-taurine reaction. Weak- field (40 mT applied parallel to the γ-beam), 4.2-K spectra of these samples are superpositions of the quadrupole doublet of the oxidized species J and the (at least two) incompletely resolved doublets characteristic of Fe(II)-containing complexes (probably due to heterogeneity in the quaternary complex) (Figure 3-5). The concentrations of the intermediate J determined from its contribution to each spectrum (filled circles and squares in Figure 3-4) confirm the existence of a large KIE on decay of the intermediate. For example, at 0.084 s, the contribution of J has already decayed to approximately half of its maximum value in the H-taurine reaction (Figure 3-5, spectrum A) but remains at its (greater) maximum value in the D-taurine reaction (spectrum B), due to its much slower decay. For both reactions, the kinetics of J agree well with those predicted by Scheme 2-2 (solid traces in Figure 3-4). Under optimized reaction conditions (i.e., excess

O2, rather than the limiting-O2 conditions employed for the above kinetic experiments), the increased lifetime of the complex with D-taurine permits its accumulation to ~ 80% of the iron in the sample (Figure 3-6, spectrum C). This augmentation of lifetime and purity proved to be extremely useful for spectroscopic characterization of the complex. High Field Mössbauer. Further insight into the electronic structure of the novel intermediate J was obtained by determination of its Mössbauer spectrum in a magnetic field of 8 T applied parallel to the γ-beam. The spectrum of the intermediate (Figure 3-7, spectrum B) was obtained by subtracting the experimental spectrum of the reactant

100

complex (solid line in Figure 3-7, spectrum A) from that of the 20-ms freeze-quenched sample (hatched data points in Figure 3-7, spectrum A), which contains the maximum concentration of J. The spectrum of the reactant complex was scaled to an intensity of 50% (its contribution to the spectrum of the 20 ms sample) prior to this subtraction. The spectrum of the intermediate thus obtained provides detailed insight into the nature of the electronic ground state of the intermediate. The fact that the intermediate gives rise to a quadrupole doublet in a weak (40 mT) magnetic field but displays hyperfine interactions in a strong externally applied field is typical of species with integer-spin electronic states. Importantly, the 8-T spectrum is a sharp, six-line spectrum with peak separations that reflect an effective magnetic field of ca. 24 T at the 57Fe nucleus. The effective magnetic field is a consequence of an internal magnetic field of ~ 32 T oriented antiparallel to the externally applied 8-T field. The internal magnetic field is given by the expression

〈S〉⋅A/gNβN, in which 〈S〉 is the expectation value of the electronic spin and A/gNβN is the

57 hyperfine tensor for the Fe nucleus. A/gNβN is dominated by the Fermi contact term, which yields an isotropic value of ~ 22 T. From the magnitude of the internal magnetic field, Bint = 32 T, we estimate 〈S〉 as 1.5 at 8 T. This spin expectation value unambiguously requires that the intermediate have an electronic ground state with S ≥ 2. The reasonable assumption that the TauD intermediate is mononuclear and the absence of an EPR signal in the g = 4.3 region (vide infra) imply that the intermediate has S = 2. This spin state implies that J have formal oxidation state of +IV. Cryoreduction of J yields an Fe(III) complex. The high-field Mössbauer spectra unambiguously demonstrate that the novel intermediate is paramagnetic and has an integer-spin ground-state of S ≥ 2. As noted above and elaborated in the Discussion, these properties and its Mössbauer parameters strongly suggest that the Fe center of the intermediate, J, is best described as high-spin Fe(IV). In order to further corroborate this assignment, we sought to test the prediction that one-electron reduction of the intermediate should produce a high-spin Fe(III) complex. To achieve reduction, we

101

employed γ-radiolysis at low temperature (cryoreduction). In this technique a glycerol- containing sample is exposed to high-energy radiation (e.g. from a 60Co source) while the sample is immersed in liquid nitrogen. This procedure generates mobile electrons, which can reduce transition metal ions and clusters in metalloproteins (e.g., see (Davydov et al. 1994) and references cited therein). During cryoreduction the sample is kept at low temperature (77 K), this will minimize structural changes that may occur upon reduction. This technique has been used, for example, to demonstrate that the formally Fe(IV/IV) cluster Q from methane monooxygenase could be reduced by one electron to yield an

Fe(III/IV) species, Qx (Valentine et al. 1998). In analogy, we expected that the high-spin Fe(IV) site of the TauD intermediate could be cryoreduced to a high-spin Fe(III) site. The Mössbauer spectrum of a glycerol-containing sample in which the quaternary

TauD•Fe(II)•αKG•taurine complex was reacted with O2 for 20 ms prior to freeze- quenching (Figure 3-8, spectrum A) is essentially identical to the spectrum shown in Figure 3-5, spectrum C and is explained as a superposition of 46% of the novel intermediate and 54% of the quaternary TauD•Fe(II)•αKG•taurine complex. After γ- irradiation of this sample at 77 K for 50 h with a total dose of 1.4 Mrad, the two quadrupole doublets arising from the quaternary complex and the intermediate are still visible, but a broad, magnetically split component is also detectable (Figure 3-8, spectrum B). The effect of the γ-irradiation continues to accumulate as the dosage is increased to 4.4 Mrad (Figure 3-8, spectrum C), with further reduction of the quadrupole doublet due to J and the increase of the magnetically split component (indicated by blue triangles). The intensity of the doublet associated with the intermediate is diminished by 30% (from 46% of the total intensity before reduction to 14% after reduction) by the procedure. Thus, the total cryoreduction yield under these conditions is ca. 71 % (32 / 45). Importantly, the contribution from the quaternary complex (54% of the total intensity) is unchanged, implying that only the Fe site of the intermediate is susceptible to cryoreduction. The decrease in the absorption of the quadrupole doublet associated with

102

the intermediate is accompanied by the appearance of a broad, magnetically split component extending from –8 mm/s to + 8 mm/s. Although these features are weak and not very well resolved, they are reminiscent of a high-spin Fe(III) center in the slow relaxation limit, as is expected for the one-electron reduced state of the high-spin Fe(IV)- containing intermediate. To obtain further evidence that the cryoreduced species contains a high-spin

Fe(III) site, we monitored the effect of γ-irradiation also by EPR spectroscopy. No EPR signals are detected in the spectrum recorded before cryoreduction (Figure 3-9 top), as is expected for a sample containing integer spin Fe-centers (both intermediates and the quaternary complex have an S = 2 ground state). After irradiation, several EPR-active species are observed (Figure 3-9 bottom). In addition to an intense signal at g = 2 originating from organic radicals generated during γ-irratiation. There is a pronounced, quasi-isotropic signal at g = 4.2, which is typical of high-spin Fe(III) sites with nearly rhombic zero field splitting parameters (E/D ~ 0.33). Taken together, the Mössbauer and EPR data provide strong evidence for the conversion of J to a high-spin Fe(III) species by cryoreduction. This demonstration is, in turn, strong evidence for assignment of the formal oxidation state of the intermediate as +IV. Structural characterization of J. An independent study by Proshlyakov and co- workers at Michigan State University used continuous-flow resonance Raman spectroscopy to investigate presence of a Fe-oxo bond in J (Proshlyakov et al. 2004). Their study showed that a vibrational band at 821 cm-1 shifted to 787 cm-1 if the reaction

18 was initiated with O2. Extensive precedent ((Nakamoto 2002); (Kitagawa et al. 1994); (Rohde et al. 2003)) allows this O-isotope sensitive resonance band to be assigned as a Fe=O with confidence. They also noted a second oxygen sensitive resonance band

-1 16 18 583/555 cm for O2 and O2 respectively, that was tentatively assigned as a stretching frequency for the iron-oxygen bond of a Fe-peroxo moiety. The authors speculated that if this band does indicate a peroxo moiety it could belong to a precursor to J such as

103

species I (see Scheme 3-1). Subsequent experiments show that this resonance feature is unaffected by longer aging times, indicating that this vibration is an artifact of the experimental conditions (Proshlyakov, unpublished data), and not an intermediate species. The structural assignment of J as an Fe(IV)=O was further refined and reinforced by an rapid freeze-quench x-ray absorption experiment. Here in collaboration with Pamela Riggs-Gelasco and co-workers (Riggs-Gelasco et al. 2004) we showed that J has a Fe K-edge with energy and pre-edge feature similar to other well characterized Fe(IV)=O complexes (Rohde et al. 2004). They also detected a N/O scatterer 1.62 Å from the Fe center, which was assigned as an oxo ligand and ascribed to a Fe(IV)=O moiety based on literature precedent (Penner-Hahn et al. 1986; Shu et al. 1997; Costas et al. 2004). Nature of the Rate-Limiting Step: Solvent Viscosity Effects. The large C1 deuterium kinetic isotope effect (kH/kD ~ 35) on decay of J and absence of accumulation of an Fe complex between J and the second intermediate establish that taurine hydroxylation and decay of J are kinetically correlated (with a rate constant of 13 ± 2 s-1 which Eric Barr has confirmed directly via chemical-quench experiments monitoring sulfite) and, therefore, that chemical steps also are not rate-limiting. Thus, only steps involved in release of products remain as viable candidates. Release of products could involve simple dissociation, a protein conformational change required for dissociation, an Fe-ligand-substitution step required for dissociation of a coordinated product (e.g., succinate, as shown in Scheme 3-1), or some combination of these processes. Variation of solvent viscosity has been used extensively to provide evidence that product release is rate-limiting in enzyme reactions (Caldwell et al. 1991; Adams et al. 1992; Cole et al. 1994). The idea is that diffusion of products away from the active site is less rapid at

higher viscosity, leading to diminution of kcat when product dissociation is rate-limiting but not when chemical steps are rate-limiting. Of the possible constituents of product release in the TauD reaction, simple dissociation is expected to be most sensitive (directly

104

proportional) to the relative solvent viscosity (RV). Ligand substitution within the active site should be insensitive, and a conformational change might or might not be slower at

higher RV. The dependence of kcat on solvent viscosity was examined, with glycerol used as viscosogen, to determine whether the rate-limiting step is diffusional in nature.

Indeed, kcat decreases with increasing solvent viscosity (Figure 3-10, circular points and blue line). A control experiment was performed with D-taurine as substrate. Previous results have shown that the large deuterium kinetic isotope effect on hydroxylation renders the chemical step rate-limiting for this substrate, which, in turn, should eliminate the solvent viscosity effect. As expected, the dependence of kcat for hydroxylation of D- taurine on RV (Figure 3-10, square points and red line) is much less pronounced (< 20 % of the effect for H-taurine). These data support the deduction that steps involved in product release limit kcat for hydroxylation of H-taurine. However, because kcat is not strictly proportional to RV, it is likely that the product release step is complex, involving both simple dissociation and one or more additional step that is insensitive (or less sensitive) to RV (e.g., ligand dissociation or a conformational change). Discussion Within the context of chapter chapter it was not possible to define the identity of the novel intermediate J. The results presented in this chapter show that J is the Fe(IV)=O hydroxylating species. The observation of the large kinetic isotope upon decay of J is clear evidence that decay of J is directly linked to the hydrogen atom abstraction. Multiple arguments provide strong evidence that the novel intermediate contains a formally high-spin Fe(IV) center. (1) The Mössbauer spectra show that the intermediate has an integer-spin ground state with S ≥ 2. Since TauD is a mononuclear Fe enzyme, we conclude that it has an S = 2 ground state. (2) Cryoreduction of the TauD intermediate is accompanied by the formation of a high-spin Fe(III) species. Therefore, the intermediate must have a formal oxidation state greater than +III, i.e. +IV. (3) The EXAFS and resonance Raman spectroscopic features are reminiscent of those seen for Fe(IV)=O

105

complexes in heme systems. This formal designation of the oxidation state does not conclusively rule out the possibility of a resonance structure, in which an electron is transferred from a coordinating ligand (e.g. a histidine to the Fe). The possibility exists that J may be best described as a Fe(III)=O coupled to a histidine radical. This would allow for resonance stabilization, by delocalizing the oxidizing equivalents over a larger number of atoms. The cryoreduction experiment would not differentiate between the reduction of a Fe(IV)=O site, and the quenching of the histidyl radical in the coupled Fe(III)=O site. The oxidation of a histidine ligand requires that the histidine deprotonate.

In other protein systems, the pKa of the ε amine in histidine, with a nominal pKa of 6.05, has been observed to vary over a range of pKa = 3 to 9 (Tishmack et al. 1997; Edgcomb et al. 2002), where the influence of counter ions and proton acceptors has been cited as

the main influence over this range. Direct coordination of the ε amine to the Fe(IV) may allow deprotonation of the δ amine with a pKa similar to that quoted for the first deprotonation (ε amine) in solution. Formation of a histidyl radical although quite feasible considering the literature regarding the variability of histidine pKa’s within the protein seems unfavorable here because of the absence of a reasonable proton acceptor. DFT calculations also indicate that the ligands do not contribute a large amount of electron density to the Fe(IV)=O moiety. Placement of the two intermediates into the overall picture of the catalytic cycle is now possible. The identification of J as the hydroxylating species and M as a product complex indicates that the hydrogen atom abstraction probably occurs at 13 s-1. Using a substrate analog Begley and co-workers (Wu et al. 1999) gave evidence that the mechanism of prolyl-4-hydroxylase includes a radical intermediate. So the decay must be several times faster than the formation in order for the substrate based radical intermediate (species III from Scheme 3-1) not to accumulate. Indeed Auclair and co- workers (Auclair et al. 2002) show, in the reaction of cytochrome P450, the lifetime of a

106

substrate radical species similar to IV is ~ 50 picoseconds. This indicates that the rate constant for the rebound process may be on the order of 1 x 1010 s-1. The inability of our studies to show evidence for intermediates prior to J indicates only that the accumulation of these intermediates is kinetically unfavorable (i.e. decay is faster than formation). The evidence presented by Proshlyakov and co-workers (Proshlyakov et al. 2004) that a Fe-peroxo moiety may accumulate at cryogenic temperatures appears to be an artifact of the experimental conditions. These findings indicate that postulated intermediates in the chemical mechanism will be detectable only upon perturbation of the kinetics by changes to the reactants themselves, protein mutagenesis, the use of substrate analogues, or some combination of these. The chemical logic requiring decarboxylation and O-O bond cleavage prior to the formation of J provides a basis for further exploration employing perturbations of the active site complex.

107

+ HN3 - OH OH SO3 2 - 2 - OOC OOC HO2 2+ His99 + α-KG O His99 + taurine His Fe Fe2+ O 2+ 99 HO Asp O Asp Fe 2 101 4 -1 -1 O 101 O O Asp101 > 3 × 10 M s > 3 × 104 M-1 s -1 His His 255 255 His255 binary complex ternary complex quaternary complex

+ 3 H O 5 -1 -1 2 -1 (1.5 ± 0.2) × 10 M s + O2 - products (2.5 ± 0.5) s

+ HN3 + HN3 SO - 3 - - HO O SO3 - - OOC OOC . O 2+ His99 O Fe O 3+ His99 Asp101 Fe O O O Asp101 CO2 His255 His255 M I

fast fast

+ + + HN3 HN3 HN3 - . SO -1 SO - - SO - 3 kH = (13 ± 2) s 3 O 3 OH - - - -1 - O fast OOC O OOC His kD = (0.35 ± 0.05) s OOC 3+ 99 4+ His99 O 4+ His99 O Fe O Fe Fe Asp Asp O O Asp O 101 O 101 101 CO2 CO2 His255 His255 His255 IV J II

Scheme 3-1: Mapping of the minimal kinetic mechanism for TauD presented in reference (Price et al. 2003) onto the current working hypothesis for the chemical mechanism. The assignment structures and of rate constants to individual steps relies on results from these references (Price et al. 2003), (Price et al. 2003), (Price et al. 2005), (Proshlyakov et al. 2004), and (Riggs- Gelasco et al. 2004). The rate constants refer to the reaction at 5 °C.

108

TauD Fe(II) αKG Taurine -1 -1 ε318 = 115 M cm -1 -1 ε520 = 200 M cm -1 -1 ε700 = 40 M cm -1 5 -1 -1 2 ± 0.5 s O2 (1.5 ± 0.2) M s

M -1 J -1 -1 kH = 13 ± 2 s -1 -1 ε318 = 0 M cm ε318 = 1,615 M cm -1 -1 -1 -1 ε520 = 0 M cm -1 ε520 = 170 M cm -1 -1 kD = 0.35 ± 0.5 s -1 -1 ε700 = 0 M cm ε700 = 0 M cm KIE = kH/kD = 37

Scheme 3-2: Kinetic mechanism used to simulate the stopped-flow data and the kinetics of the first intermediate determined by Mössbauer.

109

O O O D + + D + Ph NH NH3 NH3 Cl 3 O NH2 D D OOS OOS OOS OOS OOS OOS AB CDE Cl O O O O O- Ph Ph Ph Ph

2 Scheme 3-3: Synthetic route to 1,1-[ H]2-taurine (D-taurine).

110

2- RSO3

2- DOD RSO3

DD Fe(II) O

Fe(IV)

2- RSO3 - e DD OH

Fe(III)

H-Taurine (s-1) D-Taurine (s-1)

14 CO2 1.3 ± 0.2 0.11± 0.2

2- SO3 1.3 ± 0.2 0.072 ± 0.005

14 2- Scheme 3-4: Uncoupling of CO2 evolution from SO3 production indicates that the decay of J follows an unproductive (lower) pathway. The steady state rates for the

14 2- CO2 and SO3 production in the reactions using either H-taurine or D-taurine are tabulated. These data show that ~ 35 % of the decay of J occurs along a pathway that will not result in the hydroxylation of taurine when D-taurine is in the active site.

111

CO 0.9 14

0.6

0.3

Fraction of total total Fraction of 0 0 100 200 300 Time (ms)

14 14 Figure 3-1: Kinetics of production of CO2 from 1-[ C]-αKG in single-turnover, chemical-quenched-flow experiments in which the TauD•Fe(II)•αKG•D-

taurine complex in buffer A was mixed at 5 °C with O2-containing buffer A. The concentrations after mixing were: 1.1 mM TauD, 1 mM Fe(II), 0.06 mM

αKG, 5 mM taurine. The circular points are values from experiments using 4 N NaOH. The solid line is a fit of the equation for an exponential increase

-1 and corresponds to k = 40 s . From this value and the O2 concentration of 0.26 mM, a second-order rate constant of 1.5 x 105 M-1s-1 is calculated.

112

0 0.3

0.2 Absorbance 0.1 -0.03

Absorbance 500 600 Wavelength (nm) 0

300 400 500 600 700 Wavelength (nm)

Figure 3-2: Kinetic difference spectra (pathlength = 1 cm) recorded after mixing at 5 °C,

of a solution of 2 mM TauD, 1.8 mM Fe2+, 10 mM αKG and 10 mM synthetic D-taurine in 50 mM Tris buffer, pH 7.6, with a solution of air-

saturated (0.4 mM O2) buffer in a 2:5 volume ratio. The spectra were recorded 0.028 s (red line), 0.21 s (blue line), 2.6 s (green line), 10 s (brown line) and 50 sec (black line) after mixing.

113

0 0.3

0.2 Absorbance

0.1 -0.03

Absorbance 500 600 Wavelength (nm) 0

300 400 500 600 700 Wavelength (nm)

Figure 3-3: Kinetic difference spectra (pathlength = 1 cm) recorded after mixing at 5 °C,

of a solution of 2 mM TauD, 1.8 mM Fe2+, 10 mM αKG and 10 mM synthetic H-taurine in 50 mM Tris buffer, pH 7.6, with a solution of air-

saturated buffer (0.4 mM O2) in a 2:5 volume ratio. The spectra were recorded at 0.028 s (red line), 0.21 s (blue line), 2.6 s (green line), 10 s (brown line) after mixing.

114

0.3 318 A

∆ 0.15

0 0.01 0.1 1 Time (s)

Figure 3-4: Kinetics of, and C1-deuterium KIE, on the reaction at 5 °C of the quaternary

TauD•Fe(II)•αKG•taurine complex with limiting O2 monitored by stopped-flow absorption. A solution of the quaternary complex (in 50 mM Tris buffer, pH 7.6)

containing 2.0 mM TauD, 1.8 mM Fe(II), 17 mM αKG and 17 mM synthetic H- taurine (circles) or D-taurine (squares) was mixed with air-saturated buffer in a volume ratio of 2:5. The pathlength was 1 cm. The solid lines are simulations according to Scheme 3-2 and assumptions (0.87 equiv Fe(II) bound, 72% of complex active) indicated by simulations in Chapter 2. The initial concentrations

of O2 were calculated according to (Hitchman 1978).

115

0.0 A 2.0

0.0 B 2.0 ) %

( 0.0

2.0 C

0.0 ORPTION ORPTION

S D

B 2.0 A

0.0

2.0 E

0.0 F 2.0

-2 0 2 VELOCITY (mm/s)

Figure 3-5: Mössbauer spectra (recorded at 4.2K in a 40-mT magnetic field applied parallel to the γ beam) of samples prepared by freeze-quench of the reaction at 5 °C of a solution of the quaternary TauD•Fe(II)•αKG•taurine complex

with O2-saturated buffer (1:1 volume ratio). The concentrations after mixing were: 2 mM TauD, 1.8 mM Fe, 5 mM αKG, and 5 mM taurine.

116

Samples for spectra A, B, and C contained D-taurine in the quaternary complex and were quenched 0.084 ms, 2.0 s, and 5.0 s, respectively, after mixing. Samples for spectra D, E, and F contained synthetic H-taurine in the quaternary complex and were quenched 0.021 s, 0.084 s, and 0.16 s, respectively, after mixing. The solid lines are theoretical simulations according to the reported parameters of the spectrum of the intermediate J

(δ = 0.31 mm/s, ∆EQ = 0.88 mm/s). They are plotted at 46%, 23%, 12%, 40%, 22%, and 14% of the total intensity of spectra A through F, respectively.

117

0.5%

-2 0 2 Velocity (mm/s)

Figure 3-6: 4.2 K, weak-field Mössbauer spectra of samples prepared by using the

synthetic D-taurine analog. The top spectrum is a quaternary complex sample frozen aerobically. The bottom spectrum is of a sample frozen 0.11 s after mixing

a solution of 5.11 mM TauD, 3.63 mM 57Fe(II), 12.3 mM αKG, and 12.3 mM D-

taurine with O2-saturated buffer in a volume ratio of 1:2. The red line is the

theoretical spectrum of the Fe(IV) intermediate (δ = 0.31 mm/s, ∆EQ = 0.88 mm/s) scaled to 80% of the total iron absorption.

118

Figure 3-7: Mössbauer spectra of TauD samples recorded at 4.2 K in an 8-T magnetic

field applied parallel to the γ-beam. The spectrum of the 20-ms sample is shown in A. Overlaid as a solid line is the experimental 4.2-K/8-T spectrum of the quaternary complex, scaled to 50% of the total absorption of the spectrum of the 20-ms sample. The reference spectrum for the TauD intermediate is shown in B. It is obtained by removing 50% of the spectrum of the quaternary complex from the spectrum of the 20-ms sample. The solid line overlaid with the experimental data in B is a theoretical simulation using

the parameters δ = 0.31 mm/s, ∆EQ = 0.88 mm/s, A/gNβN = (-20 T, -20 T, -15 T) , E/D ≈ 0, D ≈ +15 cm-1.

119

Figure 3-8: 4.2-K/40-mT Mössbauer spectra of a 20-ms sample recorded (A) before and (B) after 1.4 Mrad of cryoreduction (as described in the text). Spectrum C shows the continued cryoreduction to a yield of 4.4 Mrad. In each spectrum the contribution from the quadrupole doublet associated with the quaternary complex (green) remains the same, while the contribution from the quadrupole doublet of J (red) decreases. This decrease correlates with the increase of a Fe(III) species (blue).

120

Figure 3-9: 77-K X-band EPR spectra of a 20-ms sample recorded before (top) and after (bottom) cryoreduction. Experimental parameters were: frequency, 9.51 GHz; power, 2 mW; modulation amplitude, 10 G; modulation frequency, 100 kHz.

121

100

60 % Activity 40

246 Relative Viscosity

Figure 3-10: Effect of solvent viscosity on the steady-state rates of hydroxylation of H-

taurine (circles) and D-taurine (squares). Activity was measured at 5 °C in buffer A with 5 µM TauD, 1 mM Fe(II), 1 mM αKG, 0.5 mM taurine, and

ambient O2 (~ 380 µM). Glycerol was used to vary the relative viscosity (RV) from 1.6 (0% w/w glycerol) to 6.4 (38% w/w glycerol) according to the reference ((Weast 1984)).

122

References

Adams, J. A. and Taylor, S. S. (1992). "Energetic limits of phosphotransfer in the

catalytic subunit of cAMP-dependent protein kinase as measured by viscosity

experiments." Biochemistry 31(36): 8516-22.

Auclair, K., Hu, Z., Little, D. M., Ortiz de Montellano, P. R. and Groves, J. T. (2002).

"Revisiting the Mechanism of P450 Enzymes with the Radical Clocks Norcarane

and Spiro[2,5]octane." J. Am. Chem. Soc. 124(21): 6020-6027.

Caldwell, S. R., Newcomb, J. R., Schlecht, K. A. and Raushel, F. M. (1991). "Limits of

diffusion in the hydrolysis of substrates by the phosphotriesterase from

Pseudomonas diminuta." Biochemistry 30(30): 7438-44.

Cole, P. A., Burn, P., Takacs, B. and Walsh, C. T. (1994). "Evaluation of the catalytic

mechanism of recombinant human Csk (C-terminal Src kinase) using nucleotide

analogs and viscosity effects." J. Biol. Chem. 269(49): 30880-7.

Costas, M., Mehn, M. P., Jensen, M. P. and Que, L., Jr. (2004). "Dioxygen activation at

mononuclear nonheme iron active sites: enzymes, models, and intermediates."

Chem. Rev. 104(2): 939-986.

Davydov, R., Kuprin, S., Gräslund, A. and Ehrenberg, A. (1994). "Electron Paramagnetic

Resonance Study of the Mixed-Valent Diiron Center in Escherichia coli

Ribonucleotide Reductase Produced by Reduction of Radical-Free Protein R2 at

77 K." J. Am. Chem. Soc. 116(24): 11120-8.

Edgcomb, S. P. and Murphy, K. P. (2002). "Variability in the pKa of histidine side-chains

correlates with burial within proteins." Proteins: Structure, Function, and Genetics

49(1): 1-6.

123

Hanauske-Abel, H. M. and Günzler, V. (1982). "A stereochemical concept for the

catalytic mechanism of prolylhydroxylase. Applicability to classification and

design of inhibitors." J. Theor. Biol. 94(2): 421-55.

Hitchman, M. L. (1978). Chemical Analysis, Vol. 49: Measurement of Dissolved Oxygen.

New York, Wiley.

Hoogenboom, B. E., Abbott, R., Locatell, L., Jr. and Hinman, R. L. (1959). "Chemistry

of the 1,2,4-thiadiazine ring system. II. New synthesis of 1,2,4,2H-thiadiazine-

3,5(4H,6H)-dione 1,1-dioxide." J. Org. Chem. 24: 1983-6.

Kitagawa, T. and Mizutani, Y. (1994). "Resonance Raman spectra of highly oxidized

metalloporphyrins and heme proteins." Coord. Chem. Rev. 135/136: 685-735.

Nakamoto, K. (2002). "Resonance Raman spectra and biological significance of high-

valent iron(IV,V) porphyrins." Coordination Chemistry Reviews 226(1-2): 153-

165.

Penner-Hahn, J. E., Smith Eble, K., McMurry, T. J., Renner, M., Balch, A. L., Groves, J.

T., Dawson, J. H. and Hodgson, K. O. (1986). "Structural characterization of

horseradish peroxidase using EXAFS spectroscopy. Evidence for Fe = O ligation

in compounds I and II." J. Am. Chem. Soc. 108(24): 7819-25.

Price, J. C., Barr, E. W., Glass, T. E., Krebs, C. and Bollinger, J. M., Jr. (2003).

"Evidence for hydrogen abstraction from C1 of taurine by the high-spin Fe(IV)

intermediate detected during oxygen activation by taurine:α-ketoglutarate

dioxygenase (TauD)." J. Am. Chem. Soc. 125(43): 13008-13009.

124

Price, J. C., Barr, E. W., Hoffart, L. M., Krebs, C. and Bollinger, J. M., Jr. (2005).

"Kinetic Dissection of the Catalytic Mechanism of Taurine:α-Ketoglutarate

Dioxygenase (TauD) from Escherichia coli." Biochemistry 44(22): 8138-8147.

Price, J. C., Barr, E. W., Tirupati, B., Bollinger, J. M., Jr. and Krebs, C. (2003). "The first

direct characterization of a high-valent iron intermediate in the reaction of an α-

ketoglutarate-dependent dioxygenase: A high-spin Fe(IV) complex in taurine/α-

ketoglutarate dioxygenase (TauD) from Escherichia coli." Biochemistry 42(24):

7497-7508.

Proshlyakov, D. A., Henshaw, T. F., Monterosso, G. R., Ryle, M. J. and Hausinger, R. P.

(2004). "Direct detection of oxygen intermediates in the non-heme Fe enzyme

taurine/α-ketoglutarate dioxygenase." J. Am. Chem. Soc. 126(4): 1022-1023.

Riggs-Gelasco, P. J., Price, J. C., Guyer, R. B., Brehm, J. H., Barr, E. W., Bollinger, J. M.,

Jr. and Krebs, C. (2004). "EXAFS spectroscopic evidence for an Fe=O Unit in the

Fe(IV) intermediate observed during oxygen activation by taurine:α-ketoglutarate

dioxygenase." J. Am. Chem. Soc. 126(26): 8108-8109.

Rohde, J.-U., In, J.-H., Lim, M. H., Brennessel, W. W., Bukowski, M. R., Stubna, A.,

Münck, E., Nam, W. and Que, L., Jr. (2003). "Crystallographic and Spectroscopic

Characterization of a Nonheme Fe(IV)=O complex." Science 299(5609): 1037-

1039.

Rohde, J.-U., Torelli, S., Shan, X., Lim, M. H., Klinker, E. J., Kaizer, J., Chen, K., Nam,

W. and Que, L., Jr. (2004). "Structural Insights into Nonheme

Alkylperoxoiron(III) and Oxoiron(IV) Intermediates by X-ray Absorption

Spectroscopy." Journal of the American Chemical Society 126(51): 16750-16761.

125

Shu, L., Nesheim, J. C., Kauffmann, K., Münck, E., Lipscomb, J. D. and Que, L., Jr.

(1997). "An Fe2IVO2 diamond core structure for the key intermediate Q of

methane monooxygenase." Science 275(5299): 515-518.

Sturm, K., Muschaweck, R. and Hropot, M. (1983). "5-Sulfamoylorthanilic acids, a

sulfonamide series with salidiuretic activity." J. Med. Chem. 26(8): 1174-87.

Tishmack, P. A., Bashford, D., Harms, E. and van Etten, R. L. (1997). "Use of 1H NMR

Spectroscopy and Computer Simulations To Analyze Histidine pKa Changes in a

Protein Tyrosine Phosphatase: Experimental and Theoretical Determination of

Electrostatic Properties in a Small Protein." Biochemistry 36(39): 11984-11994.

Valentine, A. M., Tavares, P., Pereira, A. S., Davydov, R., Krebs, C., Hoffman, B. M.,

Edmondson, D. E., Huynh, B. H. and Lippard, S. J. (1998). "Characterization of a

Mixed-Valent Fe(III)Fe(IV) Form of Intermediate Q in the Reaction Cycle of

Soluble Methane Monooxygenase, an Analog of Intermediate X in

Ribonucleotide Reductase R2 Assembly." J. Am. Chem. Soc. 120(9): 2190-2191.

Weast, R. C., Astle, Melvin J.,Beyer, William H. (1984). CRC Handbook of Chemistry

and Physics. Boca Raton, Florida, CRC Press Inc.

Westheimer, F. H. (1961). "The magnitude of the primary kinetic isotope effect for

compounds of hydrogen and deuterium." Chem. Rev. 61: 265-73.

Wu, M., Moon, H.-S., Begley, T. P., Myllyharju, J. and Kivirikko, K. I. (1999).

"Mechanism-Based Inactivation of the Human Prolyl-4-hydroxylase by 5-

Oxaproline-Containing Peptides: Evidence for a Prolyl Radical Intermediate." J.

Am. Chem. Soc. 121(3): 587-588.

126

Chapter 4: Probing the mechanism of “untriggered” O2 activation by the TauD•Fe(II)•αKG ternary complex

Abstract

The addition of taurine to the ternary TauD•αKG•Fe(II) complex enables the formation of the five coordinate Fe(II) site which will react rapidly with O2. The ternary complex, although coordinatively saturated, also shows reactivity towards oxygen. This reaction is much slower than the rapid activation of oxygen and subsequent chemistry that occurs when taurine is present. Studies by Ryle and co-workers (Ryle et al. 2003) showed that this aberrant pathway results in hydroxylation of the adjacent tyrosine 73 (Tyr73). They hypothesis that the Fe(IV)=O intermediate, J, forms and in absence of the substrate abstracts a H-atom from the tyrosine. The positioning of this tyrosine adjacent to the active site allows the Fe(IV)=O intermediate to be quenched in a controlled manner. This pathway has been referred to as a salvage mechanism. It appears to be the favored pathway for quenching the oxidant, ensuring that crucial components of the protein don’t become chemically altered. Here, data is presented showing that the Fe(IV)=O species, J, does not accumulate without the substrate present. Rather, we show accumulation of a high-spin Fe(III) species, probably the product of the hydrogen atom abstraction that forms the tyrosyl radical characterized by Ryle and co-workers (Ryle et al. 2003). This species absorbs strongly in the same region of the UV-visible spectrum as the Fe(IV)=O. These results are a pointed example of the necessity for characterization of species by multiple methods when studying a perturbed reaction mechanism. In this case, the stopped-flow absorption spectra appear very similar to those of J, but freeze-quench Mössbauer indicates that the only accumulating intermediate is a Fe(III) species. A recent paper on TauD variants (Grzyska et al. 2005) uses only the UV absorbance to track transient species, which they designate as emanating from analogues of J. According to the results

127

presented here, such a designation is tenuous at best. They may observe J, an uncoupled radical species similar to the Fe(III)•Y73• described by Ryle and co-workers, or something else entirely. These results also set the stage for understanding the background reactivity of newly discovered members of the Fe(II)•αKG dioxygenase

family. The evolution of CO2 without the presence of substrate, and the effect on of this aberrant reactivity on the protein, is a complication that must be accounted for in the study of systems where the natural substrates are not known or where the study involves the use of substrate analogues. Introduction In enzyme systems employing highly reactive intermediates, the possibility exists that the enzyme will damage and/or inactivate itself through misdirected reactions of these intermediates. The control of intermediates to avoid this has been termed “negative catalysis” (Retey 1990). High energy intermediates can decay in any number of ways. The specificity and efficiency of the enzyme is derived by minimizing the energy of the intended chemical route, and maximizing the energies of alternative reaction pathways. In TauD, for example, the Fe(IV)=O could cleave the protein backbone by a reaction similar to the cleavage of sulfite from the taurine substrate. Controlling the formation and reactivity of this moiety is then important to maintaining a functional enzyme. The presence of the natural substrate, taurine, in the TauD active site plays an important role in the formation and the control or “negative catalysis” of J. O’Brien and co-workers showed that taurine induces a conformational change in the protein that “caps” the active site with a large segment of random coil (O'Brien et al. 2003). Similar conformational changes have been demonstrated in other members of the family and this isolation of the active site from the environment may be a part of the “negative catalysis” (Hewitson et al. 2005). A second mechanism that protects the protein from aberrant chemistry of high energy intermediate(s) is the invocation of a trigger for the initiation of the reaction. The presence of the substrate taurine plays an important role in triggering the activation of

128

oxygen in TauD. It has been proposed that the open coordination site on iron which is present in the five coordinate square pyramidal geometry of the quaternary complex but absent in the octahedral geometry of the ternary complex is necessary for oxygen binding and the initiation of chemistry (Ho et al. 2001). This control of reactivity protects against the formation of the highly oxidizing Fe(IV)=O in the absence of the substrate. The triggering of oxygen activation by the presence of substrate is general to most oxygenases (Que et al. 1996).

Despite these considerations, it has been noted widely that CO2 can be produced without the formation of product, in members of the Fe(II)•αKG dioxygenase family.

Indeed CO2 production may proceed for multiple turnovers without the presence of substrate. Hausinger and co-workers have shown that the ternary complexes of two members of the family, TfdA (Liu et al. 2001), and TauD, will activate oxygen (Ryle et al. 2003). Upon exposure of the anaerobic TauD ternary complex to oxygen they observed a transient tyrosyl radical that decayed as a catechol complex formed. These studies showed that self hydroxylation of the enzyme resulted from this unproductive turnover of

αKG, and in the case of TauD a transient tyrosyl radical, Tyr73• was observed. This process is slow, with the tyrosyl radical formation (which is not the rate limiting step) occurring at an observed rate constant of ~ 500 M-1s-1. This chemistry results in an inactive form of the enzyme that can be rescued via an exogenous reductant such as dithionite and/or removal of the presumably oxidized Fe(III) center (Liu et al. 2001). They further showed that the tyrosyl radical is located on Tyr73, which is positioned adjacent to the binding site of taurine in the TauD active site (see Figure 4-1), and that the oxygen atom in the catecholate product of the quenched tyrosyl radical is derived from water. This reactivity was explained as a “salvage mechanism”, the hypothesis being that the activation of oxygen although not favored, may proceed to decarboxylate αKG. The production of CO2 suggests a stepwise, irreversible, cascade towards the formation of J.

129

The highly reactive nature of J requires that it be controlled or quenched in a controlled manner. Thus, according to the hypothesis, the placement of Tyr73 (see Figure 4-1) is to allow the reactive intermediate to be quenched in a manner that will allow the enzyme to continue to function. It has long been taught in basic biochemistry textbooks that prolyl hydroxylase cannot maintain full reactivity without the presence of a reductant. Although the chemical equation is balanced, without the addition of the electrons such an endogenous reductant would provide (Stryer et al. 1996). In the context of recovery from aberrant reactivity, the need for ascorbate in the prevention of connective tissue diseases such as scurvy can be rationalized as the maintenance of prolyl hydroxylase activity and subsequent collagen biosynthesis. Although the biochemical details are not established,

the presence of a similar salvage mechanism in the other members of the Fe(II)•αKG dependent dioxygenase family is a possibility. The in vitro study of this family of proteins is made difficult by the large substrates that many of these proteins target, establishing a model for the action of these proteins in ”negative catalysis mode” and the pathway of the salvage mechanism is of use for understanding the action of these proteins with substrate analogues.

Materials and Methods Materials. Chapter one gives a comprehensive list of the chemicals and materials used for experiments described in this chapter. Methods. The preparation and quality control procedures for the TauD used during the series of experiments described in this chapter are as described in Chapter one. The stopped-flow and freeze-quench and chemical quench experiments described in this chapter and the analysis of these data were carried out as described in Chapter one. Results

14 Kinetics of CO2 Production in a Single Turnover. The kinetics of CO2 production were assessed in a chemical-quenched flow experiment, in which the ternary

130

complex TauD•Fe(II)•αKG was mixed with air saturated buffer at 5 °C as described in

the methods section of Chapter 2. The rate constant of CO2 production is dramatically reduced from (1.5 ± 0.2) x 105 M-1s-1 in the presence of taurine, to 5 x 102 M-1s-1 without taurine (see Figure 4-2). This slow rate of CO2 production supports the idea that the activation of oxygen is not favored by the ternary complex. Crystallographic studies of clavaminate synthase (Zhang et al. 2000), monitoring the effect of substrate show a coordinating water moves further away from the Fe center in the presence of substrate. These structures indicate that a coordinating water molecule is forced off of the Fe site. In depth MCD studies (Zhou et al. 1998) support that there is a well defined coordination change from octahedral to square pyramidal triggered by the presence of substrate. These studies, along with similar studies in other systems, support the idea of that the slow activation of oxygen in the taurine free reaction is due to the steric obstacle of a water

molecule coordinated to the Fe center impeding the coordination of the O2 molecule. Stopped-flow absorption evidence for intermediates in the reaction of

TauD•Fe(II)•αKG reaction. Reaction of the ternary complex with oxygen shows features (see Figure 4-3) similar to those observed in the reaction of the quaternary complex with

O2, which are described in Chapter 2 (see Figure 2-4). There are, however, several important perturbations from the reaction with taurine. The absorbance at 318 nm develops much less rapidly (~ 0.1 s-1 ; see Figure 4-4), and the minimum in decay of the 520 nm complex actually precedes the maximum of the 318 nm band. As reported by Ryle and co-workers (Ryle et al. 2003), a transient tyrosyl radical is evident by the accumulating absorbance band at 408 nm (see inset to Figure 4-3). The formation rate constant for this 408 nm species matches that for the 318 nm accumulation at ~ 0.1 s-1.

-1 The tyrosyl radical decays completely (kdecay ~ 0.05 s ), but absorbance at 318 nm is still at its maximum, indicating that the product of tyrosine oxidation may absorb strongly in this region. The accumulation of the 318 nm absorbance shows a strong oxygen dependence (see Figure 4-5). Fitting the traces as the sum of three exponential processes

131

shows that, as expected, the rate of formation for this strongly absorbing product species

increases with increasing [O2]. The second phase of the absorbance profile shows an

approximately first order dependence on [O2]. At higher O2 concentrations, the second phase of the reaction is kinetically resolved and shows what may be the accumulation of a transient species. The apparent rate constant of increase deviates from an ideal first- order dependence in that it appears to approach a non-zero intercept (see inset to Figure 4-5). This same effect is seen in the study of the quaternary complex and is an indication

of the incomplete kinetic resolution of the O2-dependent and O2-independent steps. Trapping of a novel Fe intermediate and characterization by Mössbauer spectroscopy. Freeze-quench Mössbauer samples were made at time points that, according to the rate for CO2 production observed in the reaction of the ternary complex

with O2, should allow the accumulation and decay of the reactive intermediate responsible for H-atom abstraction to be observed. The 0 ms sample was made as a resting-state sample in the glove box from the protein used in the freeze-quench experiment and agrees well with spectra of the ternary complex obtained it the past with a

δ = ##, ∆EQ = ## (see Figure 4-6). J does not accumulate to any measurable degree. Rather, a new oxidized species is observed. This species is best described as high spin ferric species (see Figure 4-5B) indicating that H atom abstraction and the consequent reduction of J has already occurred. Discussion As in the investigation of the reaction of the quaternary complex in Chapter 2 and

Chapter 3, the rate of CO2 production is an indirect measure for the rate of oxygen activation. Without the substrate, taurine, present in the active site, the CO2 production rate is slowed by a factor of 300. The rate of tyrosyl radical formation may also indicative of O2 activation, showing an effect on its rate of formation with different concentrations of oxygen (Ryle et al. 2003). Interestingly Hausinger and co-workers showed previously that the rate of tyrosyl radical formation shows some effect when

132

different concentrations of O2 are used. The formation of the tyrosyl radical may offer an opportunity to investigate the activation of oxygen in absence of the substrate trigger. If

the rate of formation for this species is first order in [O2] but the rate saturates for [protein] that indicates this species may be the first to accumulate, but a protein conformational change or diffusion of water from the active site is rate limiting and occurs prior to O2 activation. This study focused on Fe based intermediates rather on the tyrosyl radical.

The first order dependence of the transient absorption at 318 nm seen in the [O2] dependence suggests that J may accumulate at sufficiently high [O2]. If the slow

activation of O2 is due to the presence of coordinated waters sterically blocking the O2 from accessing the Fe(II) atom, allowing activation only when the water diffuses away from the active site, we would expect that the rate of formation for J will be first order in

[protein] but the rate will be mixed-order (saturating) in [O2]. The accumulation of absorbance at 318 nm has previously been diagnostic for the presence of the oxidizing intermediate, J. In this series of experiments the absorbance still reflects when Fe(II) oxidation occurs, but it appears that the Fe(III) product contributes to absorbance at 318 nm. This analysis also explains why the minimum for the 520 nm metal-to-ligand charge transfer band (MLCT) is now seen prior to the maximum of the 318 nm band. The oxidizing moiety analogous to J is formed upon decarboxylation, which destroys the MLCT band observed at 520 nm. The absorption due to Fe(III) complex masks the decay of absorbance at 318 nm that would normally occur upon the decay of J. Thus the exponential rise in absorbance at 318 nm is caused by the seamless exchange of two species, J and the Fe(III) complex. Although Ryle and co-workers investigated the identity of the tyrosyl radical there is no hard kinetic data tracing the formation and decay of this intermediate. A freeze-quench EPR time course help may elucidate the identity of the two species contributing to the 318 nm absorbance. EPR is very sensitive to ferric species, and a freeze-quench EPR time course may be able to distinguish between multiple Fe(III) species.

133

The previous chapters show that J is a highly oxidizing intermediate that can be stablized in the active site for several seconds without damaging the protein. Although the salvage mechanism proposed by Ryle, et al. requires that the electron for reduction of J comes from Tyr73, this electron transfer must be fast relative to the slow rate for oxygen activation because J does not accumulate to a detectable level. The EPR studies of Ryle and co-workers, and the observation of the transient 408 nm absorbance in the stopped-flow experiments indicate that the quenching of the tyrosyl radical is likely the rate limiting step of the reaction which produces L-DOPA73. A primary strategy in the “negative catalysis” of TauD is the isolation of the active site when taurine is present. The relatively facile oxidation of Tyr73 requires that the tyrosine have an available proton acceptor. It may be that the oxidation of taurine, when available, is preferred over the oxidation of Tyr73 because of the exclusion of water and other proton acceptors necessary to reduce the reduction potential of Tyr73 within the active site. This allows Tyr73 to be available as a quenching site if J forms without substrate present, yet prohibits Tyr73 from interfering with the normal catalytic cycle.

134

Figure 4-1: Active site of taurine dioxygenase adapted from OTJ1 (O'Brien et al. 2003). Distances given in angstroms (Å). Protons (white) were added artificially using the program Weblab ViewerPro.

135

2 CO

14 0.9

0.6

0.3

Fraction of total Fraction total of 0 0 50 100 150 Time (s)

14 14 Figure 4-2: Kinetics of production of CO2 from 1-[ C]-αKG in single-turnover, chemical-quenched-flow experiments in which the TauD•Fe(II)•αKG

complex in buffer A was mixed at 5 °C with O2-containing buffer A. The concentrations after mixing were: 1.1 mM TauD, 1 mM Fe(II), 0.06 mM αKG. The solid line is a fit of the equation for an exponential increase and

-1 corresponds to k = 0.03 s . From this value and the O2 concentration of 0.26 mM, a second-order rate constant of 115 M-1s-1 is calculated.

136

0.009 0.7 408 A

0.35 0

Absorbance 110100 Time (s) 0 300 400 500 600 700 Wavelength (nm)

Figure 4-3: Kinetic difference UV-visible absorption spectra (pathlength = 1 cm) recorded after mixing at 5 °C, of a solution of 0.423 mM TauD, 0.338 mM

2+ Fe , 5 mM αKG in 50 mM Tris buffer, pH 7.6, with a solution of O2-

saturated (0.6 mM O2) buffer in a 1:1 volume ratio. The spectra were recorded at 0.020 s (red line), 2 s (blue line), 8 s (green line), 15 s (tan line) and 100 sec (black line) after mixing. Absorbance-vs-time trace of tyrosyl radical formation and decay (inset) was generated using a 3 point dropline correction (398 nm and 418 nm) for the absorbance at 408 nm.

137

0 0.6 A 520 318 A

-0.02 0

1 10 100 Time (s)

Figure 4-4: Kinetic traces from the reaction of Figure 4-3. The symbols depict the experimental absorbances at 318 nm (red circular points, left axis), 520 nm (blue square points, right axis).

138

0.04 0.18 obs k 318

A 0.12 0 ∆ 00.7[O ] mM 2 0.06

0 1 10 100 1000 Time (s)

Figure 4-5: Absorbance-versus-time traces (318 nm) after mixing at 5 °C of an O2-free solution containing 0.72 mM TauD, 0.5 mM Fe(II), 0.1 mM αKG, (50 mM Tris•HCl, pH 7.6) with an equal volume of buffer A containing varying

concentrations of O2. The symbols are the data from experiments with final

O2 concentrations of 0.15 mM (black squares), 0.3 mM (green triangles), and 0.6 mM (blue circles). The solid lines are from fits that are described in the text. The inset is a plot of the observed rate constant for the second phase of the absorbance profile verses oxygen concentration.

139

0 ms 2 %

250 ms 0.5 %

5 s 0.5 %

50 s 0.5 %

-2 0 2 Velocity (mm/s)

50 s 0.5 %

-8 -4 0 4 8 Velocity (mm/s)

Figure 4-6: Mössbauer spectra of TauD samples recorded at 4.2 K in a 40-mT

magnetic field applied parallel to the γ-beam. These samples contained 1.6

140

mM TauD, 1.3 mM 57Fe(II) and 6 mM αKG after mixing. TauD•Fe(II)•αKG

complex was mixed in a 1:2 ratio with O2 saturated buffer. The solid line above the 50 s spectrum is a theoretical simulation using parameters of J. The theory fits poorly, the wide scale spectrum of the 50 s sample (B) illustrates better the identity of the new species as high spin ferric.

141

References

Grzyska, P. K., Ryle, M. J., Monterosso, G. R., Liu, J., Ballou, D. P. and Hausinger, R. P.

(2005). "Steady-state and transient kinetic analyses of taurine/α-ketoglutarate

dioxygenase: effects of oxygen concentration, alternative sulfonates, and active-

site variants on the Fe(IV)-oxo intermediate." Biochemistry 44(10): 3845-3855.

Hewitson, K. S., Granatino, N., Welford, R. W. D., McDonough, M. A. and Schofield, C.

J. (2005). "Oxidation by 2-oxoglutarate oxygenases: non-haem iron systems in

catalysis and signalling." Philosoph. trans. Series A, Math., phys., and engin.

sciences 363(1829): 807-28; discussion 1035-40.

Ho, R. Y. N., Mehn, M. P., Hegg, E. L., Liu, A., Ryle, M. J., Hausinger, R. P. and Que,

L., Jr. (2001). "Resonance Raman studies of the iron(II)-alpha-keto acid

chromophore in model and enzyme complexes." J. Am. Chem. Soc. 123(21):

5022-5029.

Liu, A., Ho, R. Y., Que, L., Jr., Ryle, M. J., Phinney, B. S. and Hausinger, R. P. (2001).

"Alternative reactivity of an alpha-ketoglutarate-dependent iron(II) oxygenase:

enzyme self-hydroxylation." J. Am. Chem. Soc. 123(21): 5126-7.

O'Brien, J. R., Schuller, D. J., Yang, V. S., Dillard, B. D. and Lanzilotta, W. N. (2003).

"Substrate-induced conformational changes in Escherichia coli taurine/α-

Ketoglutarate dioxygenase and insight into the oligomeric structure."

Biochemistry 42(19): 5547-5554.

Que, L., Jr. and Ho, R. Y. N. (1996). "Dioxygen activation by enzymes with mononuclear

non-heme iron active sites." Chem. Rev. 96(7): 2607-2624.

142

Retey, J. (1990). "Reaction selectivity of enzymes through negative catalysis or how

enzymes work with highly-reactive intermediates." Angewandte Chemie 102(4):

373-9.

Ryle, M. J., Liu, A., Muthukumaran, R. B., Ho, R. Y. N., Koehntop, K. D., McCracken,

J., Que, L., Jr. and Hausinger, R. P. (2003). "O2- and α-Ketoglutarate-Dependent

Tyrosyl Radical Formation in TauD, an α-Keto Acid-Dependent Non-Heme Iron

Dioxygenase." Biochemistry 42: 1854-1862.

Stryer, L. and Editor (1996). Biochemistry, 4th Revised Edition.

Zhang, Z., Ren, J., Stammers, D. K., Baldwin, J. E., Harlos, K. and Schofield, C. J.

(2000). "Structural origins of the selectivity of the trifunctional oxygenase

clavaminic acid synthase." Nature Struct. Biol. 7(2): 127-133.

Zhou, J., Gunsior, M., Bachmann, B. O., Townsend, C. A. and Solomon, E. I. (1998).

"Substrate Binding to the α-Ketoglutarate-Dependent Non-Heme Iron Enzyme

Clavaminate Synthase 2: Coupling Mechanism of Oxidative Decarboxylation and

Hydroxylation." J. Am. Chem. Soc. 120(51): 13539-13540.

143

Chapter 5: Use of substrate analogues in the dissection of the taurine•α-ketoglutarate dioxygenase reaction

Abstract The work presented in Chapters 2 and 3 showed that two intermediates

accumulate during the reaction of the TauD quaternary complex with O2. Each of the components in the reaction of TauD•Fe(II)•αKG•taurine complex with O2 could in principle, be manipulated to alter the reactivity. Using the chemical logic of the proposed catalytic mechanism (see Scheme 5-1), analogues were designed in an attempt to either; (1) slow the reaction at a known point in the cycle, (2) allow accumulation of otherwise kinetically masked intermediates, or (3) build stable analogues of reactive species. Here work is presented for several substrate analogues, the 1,1-[F]2-taurine derivative of taurine, the thiocarboxylate and amido derivatives of αKG, nitric oxide as a O2 analog, and per-acetic acid as an O-atom donor. While each of these analogues offers intriguing clues about the kinetics and relative reactivities of intermediates within the catalytic cycle, none has yet provided as complete a story as has the use of the deuterated taurine analogue in Chapter 3. Introduction

Before mutagenesis of proteins became the common biochemical tool it is today, the use of chemical analogues was one of the most effective methods for studying enzyme function and mechanism, and in certain cases remains today the method of choice. In the previous chapters the use of a 14C labeled αKG analog was used to follow the fate of an individual carbon during the reaction, and specific substitution of the C1 carbon of taurine with deuterium offered valuable mechanistic insight into the catalytic cycle. With similar goals in mind, further substrate analogues were designed.

144

Prolyl-4-hydroxylase was shown to dehalogenate trans-4-fluoroprolyl peptides, but cis-4-fluoroprolyl peptides inhibited the reaction (Gottlieb et al. 1965). This indicated that the hydroxylation was sensitive to a loss of electron density caused by the highly electronegative cis-fluorine, but was fully reactive enough to break a fluorine carbon bond. In an effort to slow the decay of J to an even greater degree than seen with the D-taurine analogue, an analog was synthesized which incorporated fluorines at C1,

1,1-[F]2-taurine. Our results indicate that TauD is not able to dehalogenate the substrate, and, surprisingly, the activation of oxygen is slowed by this substitution. Although our study of the reaction with H-taurine and D-taurine indicate that the substrate does not participate chemically in the catalytic cycle until after the formation of J, substitution of the C1 hydrogens with fluorine substantially slows the activation of oxygen, and formation of J. The oxygen-dependence experiments presented in Chapter 2 show that chemical precursors to J occurring after the addition of oxygen, which chemical logic indicates should occur, are kinetically masked. The use of N-oxalylglycine (NOG) as αKG analog might favor the accumulation of a peroxo moiety (species I, Scheme 5-1). In NOG the C2 is incorporated into an amide by the substitution of C3 with nitrogen, this will allow delocalization of nitrogens free lone pair electrons into the carbonyl. By making the carbonyl carbon more electron rich it will be a poorer electrophile for the attack by the

Fe-peroxo moiety (species I, Scheme 5-1). The use of another αKG analogue, 4-Oxo-4- thiocarboxy-butyric acid, should still allow nucleophilic attack, yet decarboxylation might be slowed or stopped by the addition of the heavy, soft sulfur atom. This may favor the accumulation of a kinetically masked intermediate such as species II. Accumulation of these species would be interesting in its own right. Observation of additional intermediates would be a very important verification of DFT calculations. For example, calculations indicate that species I should be formulated as a high spin (S = 5/2) Fe(III) antiferromagnetically coupled to a superoxide radical (S = 1/2) (Brown et al.

145

1995). Ultimately, spectroscopic characterization of intermediates is the only way to confirm the hypothetical reaction mechanism, and deliniate how these enzymes accomplish their remarkable chemistry. The oxygen analogue nitric oxide and O-atom donor peracetic acid were chosen because there is a wealth of information in the published literature regarding the chemistry of these chemicals with Fe. The diatomic NO molecule forms an adduct with the Fe(II) centers which is described as a {FeNO7} species (Enemark et al. 1974). In many non-heme Fe-enzymes (Arciero et al. 1985; Orville et al. 1992)these species are are found to have a (S = 3/2) ground state where the Fe(III) (S = 5/2) is covalently bound to NO-1 (S = 1). This makes these systems amenable for characterization by paramagnetic resonance methods (EPR, ENDOR, ESEEM). O-atom donors such as peracetic acid have been used to initiate the formation of Fe(IV)=O intermediates similar to J in a number of heme systems (Schunemann et al. 2002; van den Berg et al. 2004). Using peracetic acid

to initiate the reaction may bypass the decarboxylation of αKG, and allow the formation of oxidized J analog with a new 6-coordinate geometry. Materials and Methods Materials. Chapter 2 gives a comprehensive list of the chemicals and materials used for experiments described in this chapter. Methods. The preparation and quality control procedures for the TauD protein are described in Chapter 2. The stopped-flow, freeze-quench and chemical quenched-flow experiments described in this chapter and the analysis of these data were carried out in the same manner as described in Chapter 2.

Synthesis of 1,1-[F]2-taurine. Trimethylsilyl fluoro-sulfonyldifluoroacetate was converted to difluoro-phenoxysulfonyl-acetic acid (Scheme 5-2, step A) in a procedure similar to that described by Subramanian and co-workers (Subramanian et al. 1973). Under an argon atmosphere, phenol (1.5 g, 16 mmol) was dissolved in diethyl ether (78 mL) which had been freshly distilled from sodium and benzophenone. This solution was

146

then cooled to 0 °C. Sodium hydride (0.83 g, 17 mM in a 60 % mineral oil dispersion) was then added to the cooled solution. The starting material (2.0 g, 8.03 mmol), a clear liquid of undetermined density, was added slowly with using a syringe of known mass through a septum to the stirring solution. This solution was subsequently refluxed 5 hours, after which the solution was allowed to cool to room temperature. The room temperature solution was quenched with water (100 mL). The pH of this solution was adjusted to ~9. The solution was then extracted with 3 volumes of diethyl ether. The aqueous solution was then acidified to a pH of ~1 and extracted again with 3 volumes of diethyl ether. The diethyl ether was then removed from the acidic extract and the product, difluoro-phenoxysulfonyl acetic acid, was obtained as a clear oil (1.8 g, 7.2 mmol). 1H-

13 NMR (22° C, 300 MHz, CDCl3): 7.4 ppm multiplet (5H), ~9.5 ppm singlet (1H). C-

NMR (22° C, 75 MHz, CDCl3): 113.1 ppm broad triplet, 121.9 ppm singlet, 128.5 ppm singlet, 130.5 ppm singlet, 150.3 ppm singlet, 160.1 ppm narrow triplet. 19F-NMR (22° C,

75 MHz, CDCl3) -105.2 singlet, MS: m/z = 503 (2 x 252 –H, dimer anion). Difluoro-phenoxysulfonyl acetic acid was converted to carbamoyl-difluoro- methanesulfonic acid phenyl ester (Scheme 5-2, step B). Under an argon atmosphere, difluoro-phenoxysulfonyl acetic acid (1.8 g, 7.14 mmol) was added to methylene chloride (78 mL) which had been freshly distilled from sodium and benzophenone. Anhydrous dimethyl formamide (1.9 mL) was added to the solution. This solution was allowed to

stir as it cooled to 0 °C in an ice bath. Oxalyl chloride (7.24 g, 5 mL, 57.1 mmol) was then added. The solution was stirred at room temperature for 3 hours, after which time the methylene chloride was removed under vacuum, yielding a pale yellowish solid.

Under an argon atmosphere this solid was cooled to 0 °C in an ice bath. Ammonia- saturated (~ 9 M) methanol (40 mL) was then added to the solid. The resulting solution was allowed to stir at room temperature for 8 – 16 hours. Removal of the solvent under vacuum yielded a white solid. Stirring the white solid in ethyl acetate (~50 ml), allowed the solid to be filtered from the crude product mixture. After removal of the ethyl acetate

147

under vacuum, the product was purified by column chromatograph with 20 % ethyl acetate/hexane as the elution solvent. After removal of the solvent under vacuum, carbamoyl difluoro-methanesulfonic acid phenyl ester (1.1 g, 4.37 mmol) was obtained as

1 a white solid. H-NMR (22° C, 300 MHz, CDCl3): 6.3 ppm doublet (2H), 7.4 ppm

13 multiplet (5H). C-NMR (22° C, 75 MHz, CDCl3): 113.8 ppm broad triplet, 122.1 ppm singlet, 128.5 ppm singlet, 130.5 ppm singlet, 150.3 ppm singlet, 158.7 ppm triplet. 19F-

- NMR (22° C, 75 MHz, CDCl3) -107.1 singlet, MS: m/z = 250 (M ). Carbamoyl difluoro-methanesulfonic acid phenyl ester was converted to 1,1- difluoro-2-aminoethane-1-sulfonic acid phenyl ester (Scheme 5-2, step C) using a borane-methyl sulfide complex. Carbamoyl-methanesulfonic acid phenyl ester (2.0 g, 8.0 mmol) was dissolved under an argon atmosphere in 200 mL tetrahydrofuran, which had been distilled from sodium and benzophenone. This solution was cooled to 0 °C, and borane-dimethyl sulfide (16 mL, 32 mmol) was dripped in as a 2 M solution in tetrahydrofuran. This solution was then heated to reflux, stirred for two hours, and allowed to cool. Ethylacetate (80 mL) was added, and the solution was washed with saturated aqueous solutions of ammonium chloride and then sodium chloride. The organic phase was dried over anhydrous magnesium sulfate and reduced to a clear viscous oil under vacuum. This oil was then stirred in diethylamine a minimum of 4 h, after which the diethylamine was removed under vacuum. The resulting white solid was purified by column chromatography on silica using 2 % methanol in dichloromethane as the mobile phase. The product (1.20 g, 5.06 mmol) was collected as a clear oil after

1 removal of solvents under vacuum. H-NMR (22° C, 300 MHz, CDCl3): 1.65 ppm singlet (2H), 3.45 ppm triplet (2H), 7.35 ppm multiplet (5H). 13C-NMR: (22° C, 75 MHz,

CDCl3) 44.4 ppm triplet, 122.1 ppm singlet, 122.8 ppm broad triplet, 128.2 ppm singlet, 130.4 ppm singlet, 150.2 ppm singlet. MS: m/z = 238 (MH+). 1,1-difluoro-2-aminoethane-1-sulfonic acid phenyl ester was converted to the free acid via hydrogenolysis of the of the benzyl carbon-oxygen bond (Scheme 5-2, step D) in

148

a procedure similar to that described by Sturm et al.(Sturm et al. 1983). The ester (0.35 g, 1.5 mmol) was dissolved in 12 mL tetrahydrofuran, which had been distilled from sodium and benzophenone. To this solution was added 2 mL ammonium hydroxide (25 %) and 0.15 g 10 % palladium on carbon (Pd/C) catalyst. This mixture was then capped

and stirred for 3 h under just greater than 1 atm H2(g). After filtration, the solution was extracted with dichloromethane. Lyophilization of the aqueous layer afforded 1,1- difluoro-2-aminoethane-1-sulfonic acid (0.16 g, 1.0 mmol) as a fine white powder. 1H-

13 NMR (22° C, 300 MHz, D2O): 5.5 ppm triplet (2H). C-NMR (22° C, 75 MHz, D2O): 40.6 ppm triplet, 118.5 ppm broad triplet. HRMS: m/z = 159.9890 (M-). Synthesis of N-oxalylglycine (NOG) (Scheme 5-3). Dimethyl oxalylglycine (0.3 g, 1.7 mmol) was added to stirring water (2 mL). After 20 min, 1 M sodium hydroxide was added (3.8 mL, 3.8 mmol) this solution was stirred 2 hours at room temperature ( ~ 20

°C). The basic solution was then added to a column (~50 g gel) of the acid form Dowex 50w-x8 cation exchange resin, and eluted with water until the pH rose to approximately 4. The 1.5 mL fractions were dried overnight in a Savant speedvac concentrator. Fractions with solid material were checked for product via NMR. Product containing fractions were pooled and lyophilized to afford a fine white powder. 1H-NMR (22° C, 300 MHz,

13 D2O): 3.90 ppm singlet (2H). C-NMR (22° C, 75 MHz, D2O): 41.48 ppm singlet, 161.67 ppm singlet, 162.87 ppm singlet, 172.97. MS: m/z = 146 (M-). Synthesis of 4-Oxo-4-thiocarboxy butyric acid. Boc-Glu(OtBu)-OH was converted (Scheme 5-4, step A) to Boc-Glu(Otbu)-SH in a procedure similar to that of Xia et al. (Xia et al. 2001). Under an argon atmosphere, the Boc-Glu(OtBu)-OH was dissolved in diethyl ether and cooled to -15 °C. N-methyl morpholine (0.34 g, 0.36 mL, 3.3 mmol) was added with a syringe to the stirring solution. Isobutyl chloroformate (0.45 g, 0.43 mL, 3.3 mmol) was slowly added via a syringe to the stirring solution. Hydrogen sulfide gas was then bubbled through the cold stirring solution for 20 min, after which the solution was allowed to stir at room temperature for 1 h. After stirring, water (20 mL)

149

was added to the solution. The resulting mixture was then extracted with methylene chloride. After removal of the solvent under vacuum, the product was collected as a

1 white solid (0.97 g, 3.2 mmol). H-NMR (22° C, 300 MHz, D2O): 1.45 ppm singlet (18H), 1.90 ppm multiplet (1H), 2.10 ppm multiplet (1H), 2.31 ppm multiplet (2H), 2.70 ppm singlet (1H), 3.10 ppm singlet (1H), 3.95 ppm singlet (1H), 4.30 ppm multiplet (1H),

13 5.35 ppm doublet (1H). C-NMR (22° C, 75 MHz, D2O): # ppm singlet, # ppm singlet. MS: m/z = 146 (M-). Boc-Glu(Otbu)-SH was deprotected (Scheme 5-4, step B) to form the glutamic acid analog, 4-amino-4-thiocarboxy butyric acid. The Boc-Glu(Otbu)-SH (0.23g, 0.66 mmol) was dissolved in methylene chloride (6 mL). This solution was then cooled to

0°C. Trifluoroacetic acid (99 % HPlC grade, 2ml, 26 mmol) was added to the cooled solution, which was allowed to stir at room temperature for 5 hours. The solution was quenched by removing the solvent and the trifluoroacetic acid under vacuum. Dowex 1x8-200 anion exchange resin in the formate bound form was used to separate the product and undeprotected material. The reaction mixture was added to the top of the Dowex column, and the column was eluted with a step gradient of formic acid: 30 mL at 10 mM formic acid, 30 mL at 100 mM formic acid, and 30 mL at 1 M formic acid. Individual fractions were assayed for product using a Waters ZQ mass spectrometer operating in the ESI-positive ion mode. Fractions containing product were pooled and product (0.06 g, 0.37 mmol) was collected as a sticky white solid after removal of the

1 solvent under vacuum. H-NMR (22° C, 300 MHz, CDCl3): 2.05 ppm multiplet (2H),

13 2.39 ppm multiplet (2H), 3.85 ppm multiplet (1H). C-NMR (22° C, 75 MHz, D2O): # ppm singlet, # ppm singlet. MS: m/z = 162 (M-). 4-amino-4-thiocarboxy butyric acid was converted (Scheme 5-4, step C) to 4-oxo- 4-thiocarboxy butyric acid via enzymatic catalysis using glutamate oxidase. Triethylamine carbonic acid was used as a buffering system for the enzymatic catalysis. 4-amino-4-thiocarboxy butyric acid was dissolved in triethylamine solution (100 mM, 9.4

150

mL), CO2 gas was then allowed to bubble through the solution until the pH reached 7. Catalase (15 µL, ~500 units) was added to this solution. This mixture was then used to dissolve the glutamate oxidase (5 units as ammonium sulfate precipitate). The flask containing this solution was then capped and purged with O2 gas. O2 saturated water (4 mL) was then added to the stirring solution, which was allowed to react at room temperature overnight. The reaction mixture was centrifuged at 14,000 rpm in an Eppendorph Minispin Plus to separate particulate matter, and subsequently filtered through a 3 KD molecular weight cut off Amicon microcon in the same centrifuge. The filtered solution was then lyophilized to dryness, resulting in a yellow film. This was redisolved in 0.5 mL water and the product was purified chromatography on a column (~50 g gel) of the acid form of Dowex 50w-x8 cation exchange resin. Water was used to develop the column until the pH rose to approximately 5. The individual fractions were assayed for product using a ESI Waters ZQ mass spectrometer operating in the negative ion mode. Individual fractions were dried under vacuum to yield the product 4-oxo-4-

1 thiocarboxy butyric acid as transparent golden oil. H-NMR (22° C, 300 MHz, D2O): 3.35 ppm multiplet (2H), 4.45 ppm multiplet (2H), 3.59 ppm multiplet (1H). 13C-NMR

(22° C, 75 MHz, D2O): # ppm singlet, # ppm singlet. MS: m/z = 161 (M-). Formation of TauD•Fe(III)•αKG•taurine•NO complex. After formation of the

quaternary complex in the absence of O2 as described in Chapter 2 for the stopped-flow and freeze-quench experiments, the solution of quaternary complex was placed in a small round-bottom flask fitted with an air-tight Teflon valve. This flask was then sealed and removed from the MBraun anaerobic chamber. The same vacuum system used to place the protein solution under an argon atmosphere was now configured to deliver nitric oxide gas. The head-space of the protein flask was exchanged a single time, refilling with ~ 800 psi of nitric oxide following gentle evacuation. This flask was then returned to the anaerobic chamber, where the sample was transferred too spectroscopy cells.

151

These sample containers were then removed from the anaerobic chamber and frozen in liquid nitrogen. Use of Peracetic acid as an O atom donor. Peracetic acid decays to form a gas, either oxygen or ozone, if left in buffer at room temperature. Therefore new stock solution was prepared immediately before each experiment. Peracetic acid is available commercially as a concentrated (5 M) solution in a 4 M acetic acid solution. Consequently, the initial 100 mM stock solutions were made by dilution in 1 M Trisma base. This solution was subsequently diluted using Buffer A (50 mM Tris, 10 % glycerol

pH=7.6 at 5 °C) to make solutions of appropriate concentration. The experiments were all performed at 5 °C. Mixing was 1:1, anaerobic protein solution against peracetic acid solution, in both the stopped-flow and the freeze quench experiments. Concentrations of reactants can be found in the appropriate legend. Double mixing stopped-flow absorbance experiments. The double-mix stopped- flow absorption experiments were carried out at 5 °C in an Applied Photophysics (Surrey, U.K.) SX.18MV apparatus equipped with a diode array detector and housed in the MBraun anoxic chamber. A pathlength of 1 cm was used. The solution of TauD complex was mixed in the stopped-flow apparatus with an equal volume of either buffer

A that had been allowed to reach equilibrium at 21 °C with a gas phase of 1.05 atm O2

(giving 0.6 mM O2 (Hitchman 1978) after mixing, Figure 5-7). This solution was aged prior mixing in a 1:1 volume ratio with the third reactant and injection into the observation cell. This second mix was with O2-free buffer A in the control (Figure 5-7 black trace), 3 mM ascorbic acid (green trace) or 20 mM ascorbic acid (Figure 5-7 blue trace). Experimental details, including final reactant concentrations, are given in the figure legends. Results and Discussion

1,1-[F]2-taurine as a substrate in the TauD reaction.. The substitution of the C1 hydrogens with fluorine in taurine changes the amine pKa from 9 to 6.5 as assessed by

152

titration. Despite the pKa difference, the titration of iron into the solution of TauD, 1,1-

[F]2-taurine (F-taurine), and αKG shows that the Kd for iron is not changed to a measurable degree by the use of F-taurine in the quaternary complex. F-taurine elicits the same 530 nm to 520 nm shift seen in the presence of taurine in the active site (Figure 5-1). Mössbauer spectra confirm that the Fe site experiences the same changes in isomer

shift and ∆EQ (Figure 5-8) when F-taurine is used to form the quaternary complex as are seen with the use of the unmodified substrate. Magnetic circular dichroism studies performed by Micheal Neidig, of the Solomon group at Stanford University, indicate that the transition energies and magnetic field dependencies of the resting state quaternary complex containing either taurine or F-taurine are indistinguishable. Prolyl-4-hydroxylase was shown to convert trans-4-fluoroprolyl peptides to the trans-4-hydroxyprolyl isomer (Takeuchi et al. 1969; Takeuchi et al. 1969; Uitto et al. 1977). Assuming that hydroxylation at C1 would result in the formation of sulfite, it appears that TauD, unlike prolyl hydroxylase, is not competent to break the C-F bond and dehalogenate its substrate. Under steady state reaction conditions F-taurine behaves as a

competitive inhibitor, with a KI of 90 µM very similar to the quoted Kd of 20±19 µM for the binding of taurine. Consistent with the inability to dehalogenate, a one equivalent

14 burst of CO2 is produced in the steady state, suggesting that the inhibition occurs after formation of J. Transient-state experiments show that the inhibition is more complex than initially expected. Chemical quenched-flow experiments monitoring the evolution of the

3 CO2 show that the rate constant is reduced to ~1.5 x 10 (Figure 5-2), two orders of magnitude less than the rate constants seen with D-taurine (Figure 3-1) or H-taurine

(Figure 2-9). This rate constant is still three fold faster than that observed for CO2 production in the reaction of the ternary complex with O2 (i.e. in the absence of a taurine

analog, see Figure 4-2). Figure 5-2 shows that the F-taurine complex reacts in an [O2] dependent manner in a regime were the lower limit is faster than the reaction of the

153

ternary complex (Figure 5-2, black trace), and the upper limit is slower than the taurine quaternary complex (Figure 5-2, red trace). This indicates the “triggering” effect of taurine in the active site is partly counteracted by another interaction presumably related to the fluorine substitution of C1. Crystallographic studies of TauD (Elkins et al. 2002; O'Brien et al. 2003) have shown that taurine binds such that the closest atom (the proton of C1, the target of the hydroxylation) is ~ 3.5 Å away from the Fe(II) site. This second

sphere substrate binding is a common theme within the Fe(II)•αKG dependent dioxygenase family (Hausinger 2004): the co-substrate which will be modified (in this case taurine) does not bind directly to the Fe(II) center. Although fluorine is electronically quite different from hydrogen, it is nearly isosteric with an atomic radius of 0.7 Å compared to 0.5 Å for hydrogen (see Figure 5-9). The C-F bond length is calculated to be 1.4 Å almost 0.5 Å longer than the C-H bond. Thus, the combination of sterics and electrostatics certainly causes a perturbation to the active site. These changes

seem to be most influential in the formation of the initial O2 adduct, it is surprising that

these changes can effect a 100 fold reduction in the rate-constant for O2 activation. Stopped-flow absorption spectroscopy of the single turnover reaction shows the accumulation of a 318 nm band similar to that seen in the reaction of the normal quaternary complex with O2 (see Figure 5-3). There are two important perturbations. (1)

In agreement with the transient CO2 data, the rate constant for the increase in absorption at 318 nm is much smaller. Also, as anticipated in the design of the substrate analog, the decay is slower (Figure 5-4). (2) In contrast to the reaction using D-taurine (Figure 3-2), decay of the 520 nm band is not totally kinetically masked by the 318 nm band (inset Figure 5-3). It appears as though the 318 nm absorbance band does not trail so far into the visible region as was observed in the reactions of O2 of complexes formed with H- taurine and the D-taurine analog. Although the full decay of the 520 nm band is not observed, a minimum absorbance is seen prior to the maximum absorbance time of the 318 nm band. Mössbauer spectra indicate that the absorbance is in part due to the

154

accumulation of the intermediate J (vide infra) exactly as seen in the reaction with the

taurine and D-taurine quaternary complexes with O2. Seeing this reversed relation between the maximum at 318 nm and the minimum at 520 nm may be an indication of absorbance contributions from at least two different processes, similar to the reaction of the ternary complex with O2 described in Chapter 4.

Addition of oxygen and formation of J with concomitant decarboxylation of CO2, progresses (although slowly) forms through the same mechanism as observed in the normal reaction cycle. Formation of J is followed by unproductive decay, perhaps through oxidation of Tyr73. Although stopped-flow absorbance experiments give no evidence for accumulation of tyrosyl radical, as seen in the reaction of the ternary complex with O2 (Chapter 4) the formation of tyrosyl radical may be kinetically masked.

The [O2] and [TauD•Fe(II)•αKG•F-taurine] dependence of A318, was measured better understand the effect of fluorine substitution on the kinetics of the reaction. The accumulation phase of each experiment shows that the reaction between quaternary complex and O2 proceeds with a normal concentration dependence (Figures 5- 5 and 5-6), but the decay of this species depends more drastically on which reactant is limiting. When oxygen is in excess, the intermediate is quickly quenched with kinetics unaffected by the concentration of oxygen. The absorbance at 318 nm returns to the initial value (Figure 5-5A), although the baseline is sloped due to the reaction of excess

O2 with free Fe. When O2 is limiting (i.e. the quaternary complex is in excess) the lifetime of the absorbance is greatly extended and decays again in a concentration independent manner (Figure 5-6A). The sole difference between the two experiments is the rate of decay. The concentration-dependent rate of accumulation at 318 nm is the same in the different experiments, but the rate of decay is slower in the limiting O2 experiment (Figure 3-4).

The observed rate constants (kobs) for the first phase are linear in response to concentration (Figures 5-5B and 5-6B) and give a second order rate constant of ~ 3 x 103

155

-1 -1 M s . This rate constant agrees well with that obtained for CO2 production. The plot of kobs -versus-[O2] (Figure 5-5B) shows a zero intercept because the rates of formation and of decay are so similar that mixing of the two phases is observed. This effect was also

observed in the [O2] dependence experiments presented in Chapter 2 (Figure 2-10).

The varying-[O2] experiment uses limiting αKG to maintain single turnover conditions. The fast [O2]-independent decay seen in this experiment may be due to reduction of J by free Fe(II), present because αKG is unavailable to ensure tight binding of Fe within the TauD active site. The effect of reductants on the decay of the 318-nm absorbance was investigated in double-mix stopped-flow experiments. The

TauD•Fe(II)•αKG•F-taurine quaternary complex was mixed with O2, and J was allowed to accumulate at 5 °C for 6 seconds (the time of maximal absorbance seen in Figure 5- 6A). This pre-formed J was then mixed with either anaerobic buffer A (Figure 5-7, black trace), or ascorbic acid (Figure 5-7, 3 mM green trace and 20 mM blue trace). A fast, concentration-dependent reduction in absorbance at 318 nm was observed. This fast decay implies that the Fe site is accessible to exogenous reductants, and that the quenching of J by a reductant such as free Fe(II) (because of the unknown Kd of Fe(II) in

the binary complex) may be responsible for the faster decay of J in the O2-dependence experiments. The reduction of J by free Fe in solution would result in two equivalents of Fe(III). In control experiments the oxidation of solution Fe(II) results in absorbance at 318 nm, thus it is not obvious how the decay can be completed so quickly with no evidence for a second equivalent of Fe(III). In absence of other options reduction by Fe(II) in solution is the favored mechanism. Freeze-quench Mössbauer experiments were used to confirm that the transient absorbance at 318 nm is indeed due to the accumulation of J (see Figure 5-8). J accumulates to a maximum of approximately 29 % of the total Fe in the sample in ~ 0.2 s.

This tmax corresponds well with the kinetics of the 520 nm band (see Figure 5-4) and with the expectation that J should reach maximum concentration before the minimum at 520

156

nm is reached, before 1.5 seconds in Figure 5-4. J absorbs significantly out to 600 nm, so it must decay prior to this time. This observation supports the idea that a second species with significant absorption at 318 nm accumulates (Figure 5-4). Although there is no evidence for the oxidation of Tyr73, the formation of Fe(III) through the “salvage

pathway” accommodates the accumulated CO2, stopped-flow data, and Mössbauer data.

A significant deviation from the CO2 and stopped-flow data is apparent in the 20 ms time-point (Figure 5-8, spectrum A). The 20-ms time-point shows that 12 % of the total Fe is already oxidized to Fe(IV). This surprising result was confirmed with a different protein preparation in a second freeze-quench experiment. In the stopped-flow data, no significant absorbance develops in the first 20 ms, which would show 30 % of the absorbance seen at 200 ms in order to agree with the Mössbauer. The kinetics of CO2 production corroborate the stopped-flow data in suggesting that there should not be a significant accumulation of Fe(IV) at 20 ms. This discrepancy between the early time points of the chemical quenched-flow and stopped-flow data those of the freeze-quench data may be tied to the larger question of how the modification of a second sphere substrate could affect the kinetics of oxygen activation at the iron center. The fluorine bond length is 0.5 Å longer as well as being ~ 0.2 Å larger than a hydrogen, so a steric effect may contribute. There is also a change in electronegativity between a fluorine and hydrogen. The small dipole moment of the H- C bonds will result in a partial positive charge on the hydrogen, whereas the direction of the dipole in the F-C bond will be opposite and the magnitude larger (see Figure 5-9). This change in the electric field gradient around the Fe may cause the observed reduction in the rate for the activation of oxygen. Such a field may destabilize a negatively charged Fe(III)-OO•- (species I of Scheme 5-1) adduct relative to an uncharged moiety like species II. This may effect the binding equilibrium of oxygen inducing a higher KD for

O2. Some combination of the steric and electronic effects may explain the data.

157

In an attempt to test this hypothesis that the electronic effects are an important perturbation, the effect of F-taurine on the TauD NO adduct (vide infra) was measured. We reasoned that the NO adduct of the quaternary complex should experience the same perturbations from the F-taurine substitution as does the O2 adduct, particularly after the former has been cryo-reduced to the iso-electronic state. If a field due to the C-F dipole moment disfavors the development of electron density on the diatomic ligand, Mössbauer spectroscopy may detect an effect on the isomer shift of the NO adduct, particularly after cryo-reduction. Unfortunately this experiment failed to reveal an effect. Although Mössbauer is sensitive to changes of geometry and oxidation state, the change due to fluorine substitution in the second sphere (if it affects the Fe site at all) may simply be too subtle to measure. In conclusion, every measurement made of a stable species (quaternary complex, , and NO adduct) or a transient species (J) incorporating F-taurine, indicates that F-taurine mimics taurine precisely. TauD is not able to abstract fluorine and produce sulfite. J decays through another mechanism. Possibly the salvage mechanism described in Chapter 4, although the absence of tyrosyl radical and the lack of absorbance at 600 nm for a dihydroxy-phenylalanine species (as observed by Ryle and co-workers) argues against it. F-taurine slows the rate constant for decay of J, as anticipated. Surprisingly, the F-taurine analog also affects the rate of oxygen activation, slowing it by two orders of magnitude. These experiments illustrate unexpected subtlety in substrate triggering of O2 activation by the Fe(II) and αKG dependent enzymes. Use of 4-oxo-4-thiocarboxy-butyric acid as an analog of the co-substrate αKG. The absorption-monitored titration of 4-oxo-4-thio-carboxy-butyric acid (S-αKG) into the solution of TauD, iron, and taurine shows several prominent perturbations from the behavior with αKG. The chromophore due to the metal-to-ligand charge transfer (MLCT) transition (Figure 5-10) is red shifted approximately 20 nm in comparison to the complex formed with αKG, from 520 to 540 nm. The MLCT band is also weaker with a molar

158

absorptivity of approximately 50 M-1cm-1, only 20 % of the value for the αKG complex. The titration curve indicates that the binding affinity of S-αKG is effected by the presence of taurine, similar to the affect described in Chapter 2, changing the S-αKG binding constant from 3 mM to 0.3 mM in the presence of taurine (Figure 5-11). Mössbauer does not reveal a difference in the environment of the Fe(II) upon addition of taurine reporting that both the ternary and quaternary complexes are composed of two ferrous species with the parameters ∆EQ = 2.65 mm/s, δ = 1.12 mm/s for the major (~ 60

1 %) species and ∆EQ = 2.59 mm/s, δ = 1.36 mm/s for the minor species . The rate of steady-state turnover is affected by the use of S-αKG, giving 0.2 s-1, approximately twenty percent of the rate with αKG under these conditions (1.3 s-1). The transient-state reaction also shows perturbations of the 318 nm band. Absorbance at 318 nm accumulates quickly, as in the reaction with the normal substrate (Figure 5-12). Although the total accumulation of 318 nm absorbance is quite small compared to the reaction with αKG, the observation of the expected dependence on [O2] indicates that J is the first intermediate to accumulate (Figure 5-13). The small absorbance change results in very noisy data when collected on the photo diode array. If it is assumed that, after

elimination of COS (the sulfur substituted CO2 analog) the Fe(IV) oxidizing unit is identical to J (i.e. same extinction coefficient), then it can be concluded that there is very little accumulation of J, or that only a small portion of the total quaternary complex reacts with O2. Simulation of the traces as arising from a single-turnover reaction with a

-1 -1 molar absorption coefficient for J (ε318 = 1,500 M cm ), and a concentration of reactive complex of approximately 20 % of the total Fe concentration, reproduces the

1 Speciation of multiple Fe(II) species within a broad doublet is very difficult to ascertain. A minimum of two doublets contribute to both the ternary and quaternary complexes. The quoted parameters allow for the simulation of the doublet, oddly the best fit for the quadrupole doublet requires 66 percent of the first species and 50 percent of the second.

159

experimental data well. These data are similar to those from the [O2] experiments in Chapter 2 (Figure 2-10), except here no attempt was made to limit the reaction to a single turnover. Decay of the 540 nm species correlates with the increase in 318 nm absorbance, but, although substrates are in excess and are available for reformation of the quaternary complex, there is no indication that the quaternary complex reforms (inset to Figure 5-12). This is evidently the reason that no secondary turnover occurs. The control reaction of the ternary TauD•Fe(II)•S-αKG complex showed no observable changes in the UV- visible region on a timescale considered catalytically relevant (see Figure 5-15). Interestingly the negative absorption feature that develops due to the destruction of the quaternary complex now shows a minimum close to 520 nm, very similar to the feature expected in the reaction of the normal quaternary complex (inset to Figure 5-12). Freeze-quench Mössbauer data indicate that J does accumulate upon reaction with

O2, but to a very low extent (Figure 5-14). A maximum of only 8 ± 3 percent is seen during the time course. The 20 ms and 50 ms time points contain the same percentage of J. This may indicate that the evolution of COS and therefore the formation of J are slow while decay of J is relatively fast. Simulation according to the same mechanism used in

the [O2]-dependence experiments suggests that 20 ms is the time of maximum accumulation and the 50 ms sample should contain 80 percent of this maximum concentration. Such a small change would be within the uncertainty of the method. Seeing that the 540 nm complex does not reform during the stopped flow experiment may indicate that the release of product is more severely inhibited with the presence of COS. This would indicate that a large concentration of the second intermediate M may accumulate. Unfortunately, this species would contain Fe(II) and the addition of another Fe(II) species to the already broad quadrupole doublet of the starting material can not be stated with any certainty. To differentiate between kinetic and electronic (extinction coefficient) differences in the J analog and the possibility that only a small portion of the quaternary complex

160

reacts, the D-taurine analog was used to allow a greater accumulation of the Fe(IV) intermediate. Slowing the rate of decay for the Fe(IV) moiety increased the maximal absorbance at 318 nm. This greater value is consistent with the idea of 20 % reactivity

that was established in the [O2] dependence simulations (Figure 5-15). In summary the TauD•Fe(II)•S-αKG•taurine quaternary complex forms a chromophore at 540 nm, and the extinction coefficient for this complex is ~ 50 M-1cm-1, approximately 20 % of the value for a αKG containing complex. The stopped-flow and freeze-quench Mössbauer data percent indicate that J only accumulates, maximally, to 20% of the total Fe concentration. There are at least two possible explanations for these observations. 1) It is possible that this reaction could reflect a low level contamination by unthiolated αKG. The NMR data does not detect any contamination by unthiolated αKG. Mass spectroscopy, which is much more sensitive, gives some indication that a contaminant may be present (see Appendix for the relevant spectra). In most experiments the final concentration of S-αKG was 6 mM. If in the stopped flow 20 percent of the Fe is reactive due to a contaminant, this would necessitate that > 0.1 mM oxo-αKG (approximately 2 percent of the S-αKG concentration) be in solution. If this contaminant

(Kd = 50 µM) is bound more tightly than the S-αKG (Kd = 300 µM) then the reactivity we observe could be due to this 2 percent unthiolated αKG. Detection of COS with a coupled gas chromatograph mass spectrometer (GCMS) would show that S-αKG will allow reactivity but would not negate the possibility of a 2% contamination. If no COS were detected it would imply that the only reactivity is due to this contaminant. 2) The asymmetry of the terminal carboxylate may introduce new variables into the reaction. If the large electron rich sulfur atom binds to the Fe(II), the charge transfer from the metal may be repressed. This absorbance band is though to be the donation of electron density from the electron rich Fe atom into the empty π orbitals of the ligand. This may make the charge transfer band either non existent or at a very different energy and wavelength. Reactivity of the Fe(II) with oxygen may be altered as well. Binding of the Fe(II) with

161

the oxygen atom would more nearly approximate the normal binding mode and perhaps diminish the effect of the sulfur. If the binding mode with sulfur directly chelated to the metal is preferred by 4:1 then oxygen would be bound in 20 % of the complex. DFT calculations, MCD, or EXAFS may be able to differentiate between the two possible explanations.

Use of N-oxalylglycine as an analog of the co-substrate αKG. N-oxalyglycine (NOG) is a known inhibitor of prolyl-4-hydroxylase (McDonough et al. 2005). NOG chelates Fe(II) in solution more effectively than αKG does, so the order of addition in forming the TauD complex is a necessary consideration with this analogue. Addition of NOG to the TauD•Fe and taurine solution helped ensure that the quaternary complex would form instead of an equilibrium of solution complexed Fe(II)•NOGn and TauD complex. The chromophore of the NOG quaternary complex is quite different from that

associated with the αKG complex. It has an absorbance maximum at 383 nm, and a molar absorptivity of ~ 190 M-1cm-1 (see Figure 5-16). Competitive binding experiments

involving the titration of NOG into the pre-formed αKG containing quaternary complex

confirmed that the binding constant for NOG is comparable to that for αKG, with a Kd of ~ 30 uM (see Figure 5-17). The TauD•Fe•NOG•taurine complex exhibits a quadrupole doublet (see Figure 19) with parameters typical of high spin Fe(II) (δ = 1.23 mm/s, and

∆EQ = 2.42 mm/s). These Mössbauer spectra indicate that the electronics of the NOG quaternary complex are different from those of the αKG complex.

NOG is an efficient inhibitor in the steady state, with a KI of 0.05 mM. Data from stopped-flow absorption experiments indicates that the reaction does not progress to the point of forming the oxidized Fe species, J. There is no apparent change spectrum for the first ~ 30 ms (see Figure 5-18B). At ~ 30 ms the entire spectrum shows a fast and significant absorbance increase of ~ 0.5 absorbance units. This is followed by the slow accumulation of absorbance in the UV regime of the spectrum, accompanied by a decay of the other portions of the spectrum to the initial baseline values. These spectra were

162

reproduced over multiple shots and in separate experiments. Thus it appears that 30 ms

after O2 is introduced into the system, a reaction occurs, which gives rise to an extremely broad absorption feature. The decay of these absorbance features is associated with the accumulation of a band with an absorbance maximum at ~318 nm. This band appears very similar kinetically and spectroscopically to the oxidation of the protein-free control

(i.e. oxidation of the Fe(II)•NOG solution complex). Freeze-quench Mössbauer data confirm that the Fe center does not become oxidized upon mixing with oxygen. A very rapid change is observed after mixing that

may represent the formation of a complex with oxygen. A new Fe(II) species forms (δ =

1.14, and ∆EQ = 2.7 mm/s; see Figure 5-19), approximately 30 % of the total Fe converts to this new species. This species is quite stable, remaining unchanged even in samples of

extended reaction time (e.g. 900 ms). The [O2] dependence experiments presented in chapter 2 (see Figure 2-10) indicate that the equilibrium for the initial oxygen species

may favor O2 in solution. Higher [O2] should shift this equilibrium, however an attempt

to use high pressure gaseous O2 to increase the population of this new species was unsuccessful. The new ferrous species observed in the reaction of NOG with oxygen may arise from an Fe(II)-oxo species similar to I (see Scheme 5-1). The study of bleomycin (Burger et al. 1983; Neese et al. 2000) and computational studies (Schenk et al. 2004) have indicated that these Fe-O2 species are best described as Fe(III) superoxide complexes. Although this system is likely high spin (in contrast to the low spin bleomycin complex), density functional theory (DFT) calculations may be able to ascertain whether the geometry of the active site in combination with the electron rich NOG ligand would permit such a species to form and be stable. The slow increase in absorbance at 318 nm is likely to arise from reaction of O2 with Fe(II)•NOGn complex in solution. It may also represent the action of excess oxygen in oxidizing the newly formed

Fe(II) intermediate. The [O2]-dependence of this species should be investigated in

163

stopped-flow experiments similar to those in chapter 2 (see Figure 2-10). If there is a

concentration dependence this would be strong evidence that this species is an Fe-O2 adduct. Magnetic Mössbauer would also be useful in the comparison of this new Fe(II)

8 species with the cryo-reduced {FeNO } (vide infra) analog of the Fe-O2 complex. Use of Peracetic acid as an O-atom donor for the formation of an analog of J. The chemical oxidant peracetic acid is commonly used as an O-atom donor to generate high valent iron species of heme and non-heme models, and in the heme-centered enzyme systems such as the cytochromes P450, and peroxidases (Schunemann et al. 2002; van den Berg et al. 2004). This chemical oxidant can be used at much higher concentrations than are accessible with buffered O2 solutions. Furatachi and co-workers also demonstrated reversible O-O bond cleavage in a chelating peracid moiety (Furutachi et al. 2005). The idea behind using peracetic acid as an O-atom donor as the oxidant is to initiate formation of a Fe(IV)=O intermediate. Bypassing the necessity for decarboxylation it may allow formation of a 6-coordinate J analog (J6). If J6 forms, the direct comparison of the resulting 6 coordinate Fe(IV)=O species with J would further our understanding of the structure and electronic configuration of the Fe center. The formation of J6 with in the quaternary complex should initiate hydroxylation of the substrate. The hydroxylation of taurine is detected through the reaction of Ellmann’s reagent (5,5`- Dithiobis(2-nitrobenzoic acid)) with the breakdown product sulfite. Ellmann’s reagent detects reduced sulfur compounds through the formation of a disulfide bond which produces a vivid yellow by product (5-thio 2-nitrobenzoic acid). The production of sulfite was never verified. High concentrations of protein are necessary since only a single turnover is expected. Although TauD is estimated >95 % pure and contains no cysteine to contribute to a background signal, the ~ 5 % contamination contributes a considerable background making the results of this experiment ambiguous.

164

Stopped-flow absorption spectra show features very similar to those that develop

in the reaction of the ternary complex with O2. Development of absorbance at 318 nm is very slow (see Figure 5-20). There is a slight decay at 520 nm seen in the 0.2 s spectrum. Later in the reaction, increasing absorbance centered at 600 nm likely reflects formation of the catechol species characterized by Ryle and co-workers (Ryle et al. 2003). Experiments in which [peracetic acid] was varied reveal that the rate constant for accumulation of absorbance at 318 nm is almost independent of [peracetic acid] (see Figure 5-21). This may be an indication that the [peracetic acid] is much lower than calculated. Or this may be an indication that the formation of the first oxidized intermediate is slower than its decay at the concentrations investigated here, and only at the highest experimental concentration do we begin to see an effect on the formation rate. This implies that the second order rate of formation has bypassed the rate of decay. The higher concentrations of peracid may be deleterious for the enzyme. At 6 mM final [peracetic acid] absorbance features at 408 nm consistent with those seen previously for tyrosyl radical were detected. With taurine present in the active site, the hydroxylation of

substrate should be preferred the route for decay of J6 rather than the hydroxylation of Tyr73. TauD contains multiple tyrosines and the accumulating tyrosyl radical was assumed to be an indication of deleterious side reactions. The 5-coordinate quaternary complex is optimized for chemistry with oxygen and reacts almost four orders of magnitude faster than the 6-coordinate ternary complex. When the reaction is initiated by peracetic acid, the distinction between 5-coordinate and 6-coordinate is no longer important. In fact the oxidation (as assessed by the absorbance at 318 nm) of the binary complex is faster than the oxidation of the quaternary complex (see Figure 5-22), and an absorption feature at 318 nm develops to a greater extent. The

presence of taurine seems to have very little effect. It is the presence of αKG that slows the reaction rate relative to that of the binary complex. This may be an indication that

αKG lies along the path that peracetic acid takes to the Fe center and is a steric obstacle

165

to the O-atom donor. A second interesting change occurs in the reaction with the binary complex. An absorption feature with maximum at 518 nm forms, concurrent with the formation of the 318 nm band (inset to Figure 5-22). If, similar to the report of Furutachi and co-workers (Furutachi et al. 2005) the peracid is chelating the Fe center as an αKG analog, a metal-to-ligand charge transfer band similar to that observed with αKG could be responsible for this 518 nm band. The control reaction of peracetic acid with Fe(II) in solution is slightly slower than reaction with the quaternary complex. The control reaction also does not show any sign of a chromophore at 518 nm. Freeze-quench Mössbauer was used to investigate the reaction of both the quaternary complex and the binary complex with peracetic acid. Interestingly, while the reaction with quaternary complex shows that 12 percent of the total Fe accumulates as an Fe(IV) with the parameters of J (Figure 5-23), there is no accumulation of Fe(IV) in the reaction with the binary complex. This Fe(IV) species may be J6 or it may reflect that two equivalents of peracetic acid react with the active site, one with the Fe center and

another with the coordinated αKG forming J. In comparison to the octahedral heme

systems we would expect the 6-coordinate J6 to be low spin and the precedent set by the low spin Fe(IV) (S = 1) heme systems indicates that the isomer shift (δ) should become smaller. In conclusion, peracetic acid functions poorly as an oxidant in the TauD system. Formation of Fe(IV) occurs when the quaternary complex is oxidized. This formation is slow and appears to be unchanged from the O2- initiated reaction. The observation of an absorbance band at 518 nm suggests that peracetic acid can be used as an oxidizing αKG analog, but the decomposition of this species by an unknown mechanism does not allow observation of high valent Fe species. Use of higher concentrations may induce accumulation of these species. The observation of tyrosyl radical indicates that the protein itself may be a target for peracetic acid oxidation at high concentrations.

166

Use of NO as an O2 analog for the formation of TauD•Fe(III)•αKG•taurine•NO complex. NO has been used successfully as an O2 analog in a number of non-heme iron systems(Arciero et al. 1985; Orville et al. 1992). The reaction of the TauD quaternary complex with nitric oxide (NO) gas could be tracked visually. The lavender color of the quaternary complex converts to a vivid yellow color, and, after extended reaction times (5 minutes), a yellow precipitate begins to form. The soluble yellow complex is the desired product. The precipitate suggests over-oxidation of the complex, probably through the reaction with a second equivalent of NO. The reaction is concentration (i.e. pressure) dependent. The time necessary for transfer of the NO containing flask into the anoxic chamber and transfer of the sample containers out to be frozen is ~ 5 min. An NO pressure of ~800 torr was found to react slowly enough to allow the necessary time to complete the sample transfer and freezing. The 4.2 K/40 mT Mössbauer spectrum of a sample of the TauD-NO adduct is shown in Figure 5-24. The deconvolution of this spectrum reveals that it contains 80 % of the NO-adduct and 20 % of the quaternary starting complex. The magnetic Mössbauer spectra (see Table 5-1 for parameters) of this sample indicate that, as reported (Arciero et al. 1983), the complex has an 3/2 ground state (see Figure 5-24). This state arises from coupling between the S = 5/2 Fe(III) iron site with the bound (S = 1) NO- molecule, the one electron reduced form of the NO• reactant (Schenk et al. 2004). Structurally very similar to a bound O2, the complex has

one less electron than the corresponding O2 complex (species I from Scheme 5-1).

NO is a convenient O2 analog because of its stability and its diagnostic spectroscopic features. The cryo-reduction of the NO adduct allows formation of a complex that is iso-electronic with the initial Fe(II)-O2 adduct. In the Mössbauer spectrum, this cryo-reduced species is observable as a new quadrupole doublet (δ = 1.07 mm/s and ∆EQ = 2.45 mm/s). This large increase in δ (0.69 ppm / 1.07 ppm) indicates that the cryo-reduction is predominately a metal centered reduction (Neese 2002). The spectrum of this species is a quadrupole doublet at low magnetic fields (40 mT), but at

167

high magnetic fields, a split magnetic spectrum is observed (see Figure 5-25). This observation is diagnostic of paramagnetic species with integer-spin ground states. The magnitude of the magnetic splitting is determined by the internal magnetic field. These data unambiguously demonstrate that the TauD {FeNO}8 (Enemark et al. 1974) species has an integer-spin ground state and that it is paramagnetic. The large number of unknown parameters preclude their unambiguous determination. Importantly, the spectra can be analyzed by assuming either an S =1 or S = 2 electronic ground state. In other words it is not possible to determine the spin ground state from the spectra unambiguously. The S = 1 ground state seems likely as (S = 2) high spin Fe(II) (which is paramagnetic) will likely antiferromagnetically couple to the (S = 1) reduced NO ligand. Alternatively, the NO- ligand can have a diamagnetic (S = 0) ground state, resulting in a

S = 2 ground state for the {FeNO}8 species. Computational studies have shown that the angle of the NO can effect the spin state of this ligand, at smaller Fe-N-O angles the spin state can change from S = 1 to S = 0 (Serres et al. 2004). Thus without further information the spin ground state can not be determined. It is impossible at this time to tell how accurately this {FeNO}8 complex models

the initial Fe-O2 complex (species I in Scheme 5-1). It raises the possibility that the latter may be a high spin Fe(II) complex similar to the species seen in the experiments incorporating NOG in the quaternary complex. It is important to remember though, the lessons learned from the studies with the F-taurine and S-αKG analogues which indicate that the electronic state of the Fe center and its environment within the active site of the protein are carefully modulated to optimize the reaction with oxygen. These subtle effects may be disrupted due to the incorporation of either NOG or NO.

168

+ 3 H O 5 -1 -1 2 -1 (1.5 ± 0.2) × 10 M s + O2 - products (2.5 ± 0.5) s

+ HN3 + HN3 SO - 3 - - HO O SO3 - - OOC OOC . O 2+ His99 O Fe O 3+ His99 Asp101 Fe O O O Asp101 CO2 His255 His255 M I

fast fast

Scheme 5-1: Mapping of the minimal kinetic mechanism for TauD presented in reference (Price et al. 2003) onto the current working hypothesis for the chemical mechanism. The assignment of structures and of rate constants to individual steps relies on results from these references (Price et al. 2005), (Price et al. 2003), (Price et al. 2005), (Proshlyakov et al. 2004), and (Riggs- Gelasco et al. 2004). The rate constants refer to the reaction at 5 °C.

169

O O O F F F F Si + F NH + NH3 3 F O F OH F NH2 F F OOS OOS OOS OOS O S O AB C D- F O O O O Ph Ph Ph

Scheme 5-2: Synthetic route to 1,1-[F]2-taurine (F-taurine).

170

O O O O O HO HN HN

OO OOH

Scheme 5-3: Synthetic route to N-oxalylglycine (NOG).

171

O O O O

O O O O NH2 O HS HS NH NH HO HS ABC

OOH OOH OO OO

Scheme 5-4: Synthetic route to 4-Oxo-4-thiocarboxy-butyric acid (S-αKG).

172

0.15

0.1 Absorbance 0.05

0 500 600 700 Wavelength (nm)

Figure 5-1: Comparison of the spectra for the TauD•Fe(II)•αKG ternary complex (blue), the TauD•Fe(II)•αKG•taurine quaternary complex (black) and the

TauD•Fe(II)•αKG•1,1-[F]2-taurine quaternary complex (red) at saturating

concentrations of αKG and either taurine or 1,1-[F]2-taurine.

173

2 CO

14 0.9

0.6

0.3

Fraction of total total Fraction of 0 0.001 0.01 0.1 1 10 Time (s)

14 14 Figure 5-2: Kinetics of CO2 production from 1-[ C]-αKG in single-turnover, chemical- quenched flow experiments where TauD•Fe(II)•αKG•taurine (red), TauD•Fe(II)•αKG•F-taurine (green and blue) or TauD•Fe(II)•αKG complex

(black) was reacted with O2. The concentrations after mixing were 1.1 mM TauD, 1 mM Fe(II), 0.06 mM αKG, 5 mM or 0 mM taurine analog anaerobic complex

was then mixed at 5°C in a 1:2 volume ratio against O2-containing Buffer A that

was either saturated with air (0.26 mM final O2, red, blue, and black points) or

saturated with O2 (1.26 mM final O2, green). The points represent values from experiments using 4 N NaOH as the quench solution. The solid line is a fit of the

equation for an exponential increase. The red (taurine) and black (no taurine) points of the control experiments give rates that agree well with values reported in Chapters 2 and 4. The reaction of the F-taurine complex show an intermediate

-1 -1 rate and corresponds to a kobs = 0.38 s (black) and k = 2.15 s (blue). At the

-1 -1 relevant O2 concentration, a second order rate constant of 1,500 M s and 1,700 M-1s-1 which agrees well the rate of 3000 M-1s-1 determined in stopped-flow absorption experiments.

174

0.004 0.2

0.1 Absorbance -0.004

Absorbance 500 600 Wavelength (nm) 0

300 400 500 600 700 Wavelength (nm)

Figure 5-3: UV-visible absorption spectra (pathlength = 1 cm) recorded after mixing at

5 °C, of a solution of 2 mM TauD, 1.8 mM Fe2+, 10 mM αKG and 10 mM synthetic F-taurine in 50 mM Tris buffer, pH 7.6, with a solution of air-

saturated buffer (0.4 mM O2) in a 2:5 volume ratio. The spectra were recorded at 0.020 s (red line), 0.5 s (blue line), 7 s (green line), 50 s (brown line), and 100 s (black line) after mixing.

175

A B 0.2 0 0 0.6 A A 520 520 318 318 A A

-0.01 -0.02 0 0

0.01 0.1 1 10 100 0.01 0.1 1 10 100 Time (s) Time (s)

Figure 5-4: Kinetic traces from (A) the reaction of Figure 5-3 and (B) the reaction of the

ternary TauD•Fe(II)•αKG complex (in the absence of taurine) with O2- saturated buffer under the same conditions. The symbols depict the experimental absorbances at 318 nm (circular points, left axis) and 520 nm (square points, right axis).

176

A B

0.06 3 318 obs A k ∆ 0.03 1.5

0 0 0.01 0.1 1 10 0 0.2 0.4 0.6 0.8 1 [O ] mM Time (s) 2

Figure 5-5: (A) Absorbance-versus-time traces (318 nm) after mixing at 5 °C of an O2- free solution containing 0.8 mM TauD, 0.5 mM Fe(II), 0.1 mM αKG, and 4 mM F-taurine in buffer A with an equal volume of buffer A containing

varying concentrations of O2, 0.15 mM (black), 0.3 mM (red), 0.6 mM (blue), 0.95 mM (green). (B) The observed rate constants for the development of 318 nm absorbance plotted against oxygen concentration gives a second order rate constant of 3,000 M-1s-1.

177

A B 0.09 2 318 A obs ∆ 0.045 k 1

0 0 0.01 0.1 1 10 100 1000 0 0.2 0.4 0.6 0.8 1 [O ] mM Time (s) 2

Figure 5-6: (A) Absorbance-versus-time traces (318 nm) after mixing at 5 °C of an O2-

containing solution of Buffer A (0.098 mM O2) with an equal volume of TauD solution containing varying concentrations of quaternary complex. In each

experiment F-taurine and αKG were maintained at 4 mM final concentration, TauD was maintained at 125 % of the [Fe(II)] which was the limiting reagent for the quaternary complex at 0.2 mM (blue), 0.4 mM (red), 0.6 mM (green), and 0.8 mM (black). (B) The observed rate constants for the development of 318 nm absorbance plotted against oxygen concentration gives a second order rate constant of 2,500 M-1s-1.

178

0.34 318 A ∆ 0.17

0.01 0.1 1 10 100 Time (s)

Figure 5-7: Absorbance-vs-time traces for the reduction of a pre-formed population of J

by ascorbate. A solution consisting of TauD (0.9 mM), Fe(II) (0.7 mM), αKG

(5 mM), F-taurine (4 mM), and O2 (0.6 mM) were allowed to react for 6

seconds, and this solution was then mixed in a 1:1 volume ratio with either O2 free buffer A (black trace), 3 mM ascorbate (green trace), or 20 mM ascorbic (blue trace).

179

A 0.5%

B 0.5%

C 0.5%

D 0.5%

E 0.5%

-2 0 2 Velocity (mm/s)

Figure 5-8: Mössbauer spectra of the samples from the time dependent reaction of the

TauD•Fe(II)•αKG•F-taurine with O2. Spectra were recorded at 4.2 K in a 40-mT magnetic field applied parallel to the γ-beam. These samples contained 1.6 mM

57 TauD, 1.3 mM Fe(II) 4 mM F-taurine and 5 mM αKG and 1.3 mM O2. The reaction times were A) 20 ms containing 12 % Fe(IV) B) 60 ms containing 19 % Fe(IV) C) 200 ms containing 29 % Fe(IV) D) 700 ms containing 20 % Fe(IV) E)

180

2 sec containing 16 % Fe(IV). The solid line above spectrum C is a theoretical simulation of the spectrum of the intermediate using parameters quoted in the text. The simulated spectrum is plotted at an intensity corresponding to 29 % of the area of spectrum C.

181

Figure 5-9: Active site of taurine:α-ketoglutarate dioxygenase adapted from OTJ1 (O'Brien et al. 2003). Fluorines (shown in cyan) have been artificially substituted for the original C1 hydrogens using the program Weblab ViewerPro.

182

0.03 520 A

0 500 600 700 Wavelength (nm)

Figure 5-10: Absorption spectra acquired during titration at room temperature of TauD

(1.2 mM) with S-αKG in the absence of O2 and the presence of 1 mM Fe(II) and 5 mM taurine. The sample prior to the first addition of S-αKG (containing all other components) was used as spectral reference. The spectra shown correspond to 5 mM (green line), 3.1 mM (red line), 1.4 mM (blue line), 0.47 mM (black line

with circular symbols) and 0 mM (black line) S-αKG.

183

0.03 540 A ∆

0 0123 [S-αKG] (mM)

Figure 5-11: Absorbance vrs concentration curves monitoring the development of the

540 nm metal-to-ligand charge transfer band during the titration of S-αKG into a solution of 1.2 mM TauD, 1 mM Fe(II), in the absence of taurine (blue points) and in the presence of 5 mM taurine (red points). The solid blue line is a fit of the hyperbolic equation, and the red line is a fit of the quadratic equation for binding to the data which gives a dissociation constant of ~ 3 mM and 0.3 mM, respectively.

184

0 0.1 Absorbance -0.018

Absorbance 500 600 Wavelength (nm) 0

300 400 500 600 700 Wavelength (nm)

Figure 5-12: UV-visible absorption spectra (pathlength = 1 cm) recorded after mixing

at 5 °C of a solution of 1.3 mM TauD, 1 mM Fe2+, 12 mM S-αKG and 10 mM

taurine in 50 mM Tris buffer, pH 7.6, with a solution of O2-saturated buffer (1.2

mM O2) in a 1:1 volume ratio. The spectra were recorded 0.020 s (red line), 0.1 s (blue line), 0.5 s (green line), 1 s (brown line), and 10 s (black line) after mixing.

185

0.12 318 A

∆ 0.06

0 0.01 0.1 Time (s)

Figure 5-13: Absorbance-versus-time traces (318 nm) after mixing at 5 °C of an O2-free solution containing 0.64 mM TauD, 0.5 mM Fe(II), 6 mM S-αKG, and 4 mM taurine in buffer A (50 mM Tris•HCl, pH 7.6) with an equal volume of buffer A

containing varying concentrations of O2. The symbols are the data from

experiments with final O2 concentrations of 0.15 mM (squares), 0.3 mM (triangles), 0.6 mM (circles) and 0.95 mM (crosses). The solid lines are from simulations that are described in the text and that employed the same parameters were used for the reaction of the normal quaternary complex (Chapter 2). A concentration of 0.1 mM (20 %) active quaternary complex was assumed for the simulation.

186

Figure 5-14: Mössbauer spectra of TauD samples recorded at 4.2 K in a 40-mT

magnetic field applied parallel to the γ-beam. The 0 ms spectrum is that of the quaternary TauD•Fe(II)•S-αKG•taurine complex. These samples contained 1.6 mM TauD, 1.3 mM 57Fe(II), 6 mM S-αKG and 4 mM taurine. The 0 ms sample was not prepared by freeze-quench but was prepared in the MBraun glove box. The remaining spectra are of freeze-quenched samples from the reaction of the

quaternary TauD•Fe(II)•S-αKG•taurine complex with O2. The concentrations

187

after mixing were the same as for the 0 ms sample and 1.3 mM O2. The solid line above the 20 and 50 ms spectra is a theoretical simulation of the spectrum of J according to the parameters quoted in the text. The simulation above each spectrum is plotted at an intensity corresponding to 8 % of the total Fe intensity.

188

0.24 318

A 0.12 ∆

0 0.01 0.1 1 Time (s)

Figure 5-15: Absorbance-versus-time traces (318 nm) for the reaction at 5 °C of a O2

saturated buffer A (0.6 mM O2) with an O2-free solution containing 0.65 mM TauD, 0.5 mM Fe(II), and either 6 mM S-αKG (black trace), 6 mM S-αKG and 4 mM taurine (blue trace), 6 mM S-αKG and 4 mM D-taurine (red trace) or 6 mM αKG and 4 mM taurine (green trace).

189

0.2 0.07 520 A

0.1 0 01

bsorbance [Fe(II)] mM A

400 500 600 Wavelength (nm)

Figure 5-16: Absorption spectra acquired during titration at room temperature of TauD

(1.2 mM) with NOG in the absence of O2 and the presence of 1 mM Fe(II) and 5 mM taurine. The sample prior to the first addition of S-αKG (containing all other components) was used as spectral reference. The spectra shown correspond to 1.45 mM (red line), 0.75 mM (blue line), 0.5 mM (green line), 0.25 mM (black line with circular symbols) and 0.1 mM (black line) NOG. The points in the inset depict the absorbance at 380 nm (background corrected to the line defined by the absorbances at 320 and 440 nm) as a function of the concentration of NOG. The solid line is a fit of the quadratic equation for binding to the data, it gives a dissociation constant of 0.025 mM.

190

0.4

0.3

0.2

Absorbance 0.1

0 400 500 600 700 800 Wavelength (nm)

Figure 5-17: Absorption spectra acquired during titration at room temperature of TauD

(1.2 mM) with NOG in the absence of O2 and the presence of 1 mM Fe(II), 2 mM αKG and 5 mM taurine. The sample prior to the addition of Fe(II) (containing TauD, αKG, and taurine) was used as spectral reference. The spectra shown correspond to 0 mM NOG (red line) and 24 mM NOG (blue line) with graded additions in between (thin black lines).

191

0.06 0.1 Absorbance Absorbance

0 0 300 400 500 600 700 0.01 0.1 1 10 Wavelength (nm) Time (s)

Figure 5-18: UV-visible absorption data (pathlength = 1 cm) recorded after mixing at 5 °C, of a solution of 1.3 mM TauD, 1 mM Fe2+, 2.6 mM NOG and 10 mM taurine

in 50 mM Tris buffer, pH 7.6, with O2-saturated buffer (1.2 mM O2) in a 1:1 volume ratio. A) Kinetic difference spectra were recorded at 0.020 s (red line), 0.05 s (blue line), 0.1 s (green line), 0.2 s (yellow line), and 1.5 s (black line) after mixing. B) Absorbance-versus-time traces from the reaction in A) at 318 nm (black), 340 nm (green), 380 nm (red), 425 nm (blue), and 470 nm (black squares).

192

1.0

-2 0 2 Velocity (mm/s) Figure 5-19: Mössbauer spectra of TauD samples recorded at 4.2 K in a 40-mT

magnetic field applied parallel to the γ-beam. Top spectrum is of the quaternary TauD•Fe(II)•NOG•taurine complex. These samples contained 1.6 mM TauD, 1.3

mM 57Fe(II) and 6 mM αKG and 4 mM taurine. The sample for the top spectrum was not prepared by freeze quench but was prepared in the MBraun glove box. The bottom spectrum is of a freeze-quenched sample from the reaction of the

quaternary TauD•Fe(II)•NOG•taurine complex with O2. The concentrations after mixing were 1.5 mM TauD, 1.5 mM 57Fe(II), 5 mM αKG, 5 mM taurine and 1.3

mM O2. The blue line overlaying the top spectrum is a simulation of the spectrum

193 for the NOG quaternary complex according to the parameters quoted in the text. The red line overlaid on the bottom spectrum is composed of two species: 68 % quaternary complex and 32 % of the new ferrous species. This spectrum was fully formed in the 20 ms sample and was unchanged upon extended aging times.

194

0.038

0.2 Absorbance

0.1 0 500 600 Absorbance Wavelength (nm)

0 300 400 500 600 700 Wavelength (nm)

Figure 5-20: UV-visible absorption spectra (pathlength = 1 cm) recorded after mixing at 5 °C of quaternary complex with buffered peracetic acid in a 1:1 volume ratio.

The final concentrations were 0.65 mM TauD, 0.5 mM Fe2+, 5 mM αKG and 5 mM taurine in 50 mM Tris buffer, pH 7.6, which reacted with buffered peracetic acid (3 mM final concentration). The spectra were recorded 0.020 s (red line), 0.2 s (blue line), 1 s (green line), 70 s (brown line), and 1000 s (black line) after mixing.

195

0.5 318 A ∆ 0.25

0 0.01 0.1 1 10 100 Time (s)

Figure 5-21: Absorbance-versus-time traces (318 nm) after mixing at 5 °C of an O2-free solution containing 1.3 mM TauD, 1 mM Fe2+, 10 mM αKG and 10 mM taurine in buffer A (50 mM Tris•HCl, pH 7.6) with an equal volume of buffer A containing varying concentrations of peracetic acid. The symbols are the data from experiments with final peracetic acid concentrations of 0.75 mM (green trace), 1.5 mM (blue trace), 3 mM (red trace) and 6 mM (black trace).

196

0.4 0.07

0.2 Absorbance

0 500 600 Absorbance Wavelength (nm)

0 300 400 500 600 700 Wavelength (nm)

Figure 5-22: UV-visible absorption spectra (pathlength = 1 cm) recorded after mixing at 5 °C of binary complex with buffered peracetic acid in a 1:1 volume ratio. The

final concentrations were 0.65 mM TauD, 0.5 mM Fe2+ in 50 mM Tris buffer, pH 7.6, which reacted with buffered peracetic acid (3 mM final concentration). The spectra were recorded 0.020 s (red line), 0.2 s (blue line), 1 s (green line), 70 s (brown line), and 1000 s (black line) after mixing.

197

1% 20 ms 20 ms 1%

1% 100 ms

1% 1 s 100 ms 1%

0.5% 30 s

-8 -4 0 4 8 -2 0 2 Velocity (mm/s) Velocity (mm/s)

Figure 5-23: Mössbauer spectra of samples from the reaction of TauD quaternary complex with peracetic acid, recorded at 4.2 K in a 40-mT magnetic field applied

parallel to the γ-beam. These samples were all made by freeze-quench at the reaction times indicated, and contained final concentrations of 1.6 mM TauD, 1.3

mM 57Fe(II) and 5 mM αKG, 5 mM taurine, and 4 mM peracetic acid. The spectra on the left were recorded in the wide 9R scale and show that the final product of the reaction contains Fe(III). The spectra on the right are of the same freeze-quench samples recorded in a narrower 3R scale. The solid lines plotted above the right spectra are a theoretical simulations of the spectrum of J using parameters quoted in the text, they are plotted at intensities corresponding to a maximum accumulation of 12 % J.

198

0.0

A 1.0

0.0 B 1.0

ABSORPTION (%) 2.0

-1.0

0.0 C 1.0

-8 -4 0 4 8 VELOCITY (mm/s)

Figure 5-24: 4.2 K / 40 mT Mössbauer spectra of the NO-TauD complex recorded

before (A) and after (B) γ-irradiation at 77 K (total dose 4.4 Mrad). The difference

spectrum (C) illustrates the change that occurs upon reduction. The solid line

(overlay of spectrum C) was generated by assuming that 31 % of the NO-TauD

complex are converted to a quadrupole doublet (δ = 1.07 mm/s and ∆EQ = 2.42

199 mm/s, 31 %). The spectrum of the NO-TauD complex (B) was obtained by subtracting 20 % of the quaternary complex from the top spectrum.

200

0.0 A 0.2

0.4

0.0

0.2 B

ABSORPTION (%) 0.4

0.0

0.2 C 0.4

-8 -4 0 4 8 VELOCITY (mm/s)

Figure 5-25: 4.2 K Mössbauer spectra of the TauD {FeNO}8 complex recorded in

magnetic fields of (A) 40 mT, (B) 4 T, and (C) 8T oriented parallel to the γ-beam.

The spectra were generated by adding the contribution of the TauD {FeNO}7

adduct to the difference spectra obtained from the data recorded before and after

cryoreduction. The solid lines are simulations computed using the parameters

given in Table 5-1 with T; Euler angles rotating the EFG into the coordinate

frame of the ZFS tensor α, β, γ = (90, 90, 90)°.

201

7 Table 5-1: Mössbauer parameters of {FeNO} species with Stotal = 3/2 ground states (these species contain high spin Fe(III) - sites (SFe = 5/2), which are antiferromagnetically coupled to NO (SNO = 1). D, E/D, and Aii/gNβN values are given for the S = 3/2 ground state.

Parameter IPNS-NO IPNS-ACV- [Fe(EDTA)(NO)] cis-[(cyclam- [L’Fe(N3)2(NO)] TauD:αKG: NO ac) Fe(NO)]+ taurine:NO D (cm-1) 14 14 12 12.6 13.6 11.8 E/D 0.015 0.035 0.020 < 0.01 0 0.02 δ (mm/s) 0.75 0.65 0.66 0.64 0.62 0.69 ∆EQ (mm/s) -1.0 -1.2 -1.72 -1.78 -1.28 -1.70 η 0.1 1.0 0 0.4 0.2 0.8 Axx/gNβN (T) -25.4 -22.5 -23.2 -20.0 -21.3 -24.3 Ayy/gNβN (T) -24.0 -20.3 -23.2 -24.8 -20.8 -24.7 Azz/gNβN (T) -32.7 -24.7 -31.2 -23.4 -25.4 -31.5 α, β. γ a 0, 8, 0 10, 20, 0 0, 10, 0 70, 0, 20 ref (Orville et (Orville et (Orville et al.)0} (Hauser et al. (Hauser et al. This work al.)0} al.)0} 2000)1} 2000)1}

a Euler angles rotating the electric field gradient (EFG) into the coordinate frame of the zero field splitting (ZFS) tensor.

202

References

Arciero, D. M., Lipscomb, J. D., Huynh, B. H., Kent, T. A. and Münck, E. (1983). "EPR

and Moessbauer studies of protocatechuate 4,5-dioxygenase. Characterization of

a new ferrous ion environment." J. Biol. Chem. 258(24): 14981-14989.

Arciero, D. M., Orville, A. M. and Lipscomb, J. D. (1985). "[17O]water and nitric oxide

binding by protocatechuate 4,5-dioxygenase and catechol 2,3-dioxygenase.

Evidence for binding of exogenous ligands to the active site iron of extradiol

dioxygenases." J. Biol. Chem. 260(26): 14035-44.

Brown, C. A., Pavlosky, M. A., Westre, T. E., Zhang, Y., Hedman, B., Hodgson, K. O.

and Solomon, E. I. (1995). "Spectroscopic and Theoretical Description of the

Electronic Structure of S = 3/2 Iron-Nitrosyl Complexes and Their Relation to O2

Activation by Non-Heme Iron Enzyme Active Sites." J. Am. Chem. Soc. 117(2):

715-732.

Burger, R. M., Kent, T. A., Horwitz, S. B., Münck, E. and Peisach, J. (1983).

"Moessbauer study of iron bleomycin and its activation intermediates." J. Biol.

Chem. 258(3): 1559-1564.

Elkins, J. M., Ryle, M. J., Clifton, I. J., Hotopp, J. C. D., Lloyd, J. S., Burzlaff, N. I.,

Baldwin, J. E., Hausinger, R. P. and Roach, P. L. (2002). "X-ray crystal structure

of Escherichia coli taurine/.alpha.-ketoglutarate dioxygenase complexed to

ferrous iron and substrates." Biochemistry 41(16): 5185-5192.

Enemark, J. H. and Feltham, R. D. (1974). "Principles of structure, bonding, and

reactivity for metal nitrosyl complexes." Coord. Chem. Rev. 13(4): 339-406.

203

Furutachi, H., Hashimoto, K., Nagatomo, S., Endo, T., Fujinami, S., Watanabe, Y.,

Kitagawa, T. and Suzuki, M. (2005). "Reversible O-O Bond Cleavage and

Formation of a Peroxo Moiety of a Peroxocarbonate Ligand Mediated by an

Iron(III) Complex." Journal of the American Chemical Society 127(13): 4550-

4551.

Gottlieb, A. A., Fujita, Y., Udenfriend, S. and Witkop, B. (1965). "Incorporation of cis-

and trans-4-fluoro-L-prolines into proteins and hydroxylation of the trans isomer

during collagen biosynthesis." Biochemistry 4(11): 2507-13.

Hauser, C., Glaser, T., Bill, E., Weyhermueller, T. and Wieghardt, K. (2000). "The

Electronic Structures of an Isostructural Series of Octahedral nitrosyl-iron

Complexes {Fe-NO}6,7,8 Elucidated by Moessbauer Spectroscopy." Journal of

the American Chemical Society 122(18): 4352-4365.

Hausinger, R. P. (2004). "Fe(II)/α-ketoglutarate-dependent hydroxylases and related

enzymes." Crit. Rev. Biochem. Mol. Biol. 39(1): 21-68.

Hitchman, M. L. (1978). Chemical Analysis, Vol. 49: Measurement of Dissolved Oxygen.

New York, Wiley.

McDonough, M. A., McNeill, L. A., Tilliet, M., Papamicaeel, C. A., Chen, Q.-Y., Banerji,

B., Hewitson, K. S. and Schofield, C. J. (2005). "Selective Inhibition of Factor

Inhibiting Hypoxia-Inducible Factor." J. Amer. Chem. Soc. 127(21): 7680-7681.

Neese, F. (2002). "Prediction and interpretation of the 57Fe isomer shift in Mossbauer

spectra by density functional theory." Inorganica Chimica Acta 337(Complete):

181-192.

204

Neese, F., Zaleski, J. M., Zaleski, K. L. and Solomon, E. I. (2000). "Electronic Structure

of Activated Bleomycin: Oxygen Intermediates in Heme versus Non-Heme Iron."

Journal of the American Chemical Society 122(47): 11703-11724.

O'Brien, J. R., Schuller, D. J., Yang, V. S., Dillard, B. D. and Lanzilotta, W. N. (2003).

"Substrate-induced conformational changes in Escherichia coli taurine/α-

Ketoglutarate dioxygenase and insight into the oligomeric structure."

Biochemistry 42(19): 5547-5554.

Orville, A. M., Chen, V. J., Kriauciunas, A., Harpel, M. R., Fox, B. G., Munck, E. and

Lipscomb, J. D. (1992). "Thiolate ligation of the active site Fe2+ of isopenicillin

N synthase derives from substrate rather than endogenous cysteine: spectroscopic

studies of site-specific Cys----Ser mutated enzymes." Biochemistry 31(19): 4602-

12.