Handbook of Flavoproteins: Volume 2 Complex Flavoproteins, Dehydrogenases and Physical Methods 9783110298345, 9783110298284

Citation preview

Handbook of Flavoproteins Hille, Miller, Palfey (Eds.)

Handbook of Flavoproteins Oxidases, Dehydrogenases and Related Systems Volume 1 Russ Hille, Susan Miller, Bruce Palfey (Eds.) ISBN 978-3-11-026842-3 e-ISBN 978-3-11-026891-1 De Gruyter, Berlin 2013

Complex Flavoproteins, Dehydrogenases and Physical Methods Volume 2 Russ Hille, Susan Miller, Bruce Palfey (Eds.) ISBN 978-3-11-029828-4 e-ISBN 978-3-11-029834-5 De Gruyter, Berlin 2013

Available as Set ISBN 978-3-11-030089-5 e-ISBN 978-3-11-030090-1

Also of Interest Methods in Protein Biochemistry Tschesche (Ed.), 2011 ISBN 978-3-11-025233-0, e-ISBN 978-3-11-025236-1

Membrane Systems For Bioartificial Organs and Regenerative Medicine De Bartolo, Curcio, Drioli, 2013 ISBN 978-3-11-026798-3, e-ISBN 978-3-11-026801-0

Industrial Enzyme Applications An Introduction Aehle, Bornscheuer, 2014 ISBN 978-3-11-026157-8, e-ISBN 978-3-11-026397-8

Kallikrein – Related Peptidases Characterization, Regulation, and Interactions within the Protease Web Magdolen, Sommerhoff, Fritz, Schmitt (Eds.), 2012 ISBN 978-3-11-026036-6, e-ISBN 978-3-11-026037-3

Handbook of Flavoproteins Complex Flavoproteins, Dehydrogenases and Physical Methods Volume 2 Edited by Russ Hille, Susan Miller, Bruce Palfey

DE GRUYTER

Editors Prof. Russ Hille University of California Department of Biochemistry 1643 Boyce Hall Riverside, CA 92521 USA [email protected]

Prof. Susan M. Miller University of California Department of Pharmaceutical Chemistry 600 16th Street San Francisco, CA 94158 USA [email protected] Prof. Bruce Palfey University of Michigan Department of Biological Chemistry 1150 W. Medical Center Dr. Ann Arbor, MI 48109 USA [email protected]

ISBN 978-3-11-029828-4 • e-ISBN 978-3-11-029834-5 Library of Congress Cataloging-in-Publication Data A CIP catalog record for this book has been applied for at the Library of Congress. Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available in the Internet at http://dnb.dnb.de. © 2013 Walter de Gruyter GmbH, Berlin/Boston. The publisher, together with the authors and editors, has taken great pains to ensure that all information presented in this work (programs, applications, amounts, dosages, etc.) reflects the standard of knowledge at the time of publication. Despite careful manuscript preparation and proof correction, errors can nevertheless occur. Authors, editors and publisher disclaim all responsibility and for any errors or omissions or liability for the results obtained from use of the information, or parts thereof, contained in this work. The citation of registered names, trade names, trade marks, etc. in this work does not imply, even in the absence of a specific statement, that such names are exempt from laws and regulations protecting trade marks etc. and therefore free for general use. Typesetting: Compuscript, Shanon – Ireland Printing and binding: Hubert & Co. GmbH & Co. KG, Göttingen Cover image: The figure presents a cartoon representation of the human flavoenzyme dihydroorotate dehydrogenase. Kindly provided by Dr. Maria Cristina Nonato Printed on acid-free paper Printed in Germany www.degruyter.com

Preface

Since the discovery of what was initially called the Yellow Enzyme by Otto Warburg in 1933 (and subsequently renamed Old Yellow Enzyme, a name the protein still retains) flavoproteins have been identified and characterized from virtually every organism known. More recent investigations have demonstrated that, in addition to their now well-established roles in catalyzing oxidation-reduction reactions of all sorts, flavoproteins also play important roles in a broad range of biological processes, including signal transduction and photochemistry. In addition, biotechnological applications of flavoproteins have progressed significantly since the development of blood glucose tests based on the action of glucose oxidase. Finally, flavoproteins have provided the testing ground for some of the newest and most cutting edge experimental methodologies in biochemistry and biophysics, and have provided critical insight into, for example, the short-time molecular dynamics of flavoproteins and the manner in which flavoproteins activate triplet molecular oxygen for reaction with singlet molecules. Given the significant progress in our understanding of both the diversity of processes in which flavoproteins are involved and the specific manner in which they carry out their roles within (or without) the cell, it seems both appropriate and timely to provide a survey of the field in this and the accompanying volume. We have attempted, within the constraints of space and availability of the material, to compile a comprehensive pair of volumes that together accurately reflect our present understanding of how a broad spectrum of flavoproteins work. It is our hope that these volumes will prove to be useful reference sources both for workers in the field and for instructors at both the undergraduate and graduate level not only in biochemistry and biophysics, but also pharmacology, medicine and bioengineering. We wish to thank our contributors who have prepared such lucid accounts of their respective areas. We are also particularly indebted to three individuals at Walter de Gruyter: Dr. Stephanie Dawson for her enthusiasm and support for this project from the outset, Ms. Julia Lauterbach for her excellent editorial efforts that made publication of these volumes a reality, and to Ms. Sabina Dabrowski for coordinating technical production. Finally, we are of course grateful to our spouses (Kim Hille, Walter Moos and Kim Palfey) for their support and patience throughout our careers generally, and during the preparation of these volumes specifically. One of us (RH) would also like to thank the Alexander von Humboldt Foundation of Germany for its support during the editing of these volumes. Russ Hille Susan Miller Bruce Palfey

Contributing authors

David Ballou Department of Biological Chemistry University of Michigan 5301 MSRBII Ann Arbor, MI 48109, USA e-mail: [email protected] chapter 1

Emil Pai Department of Medical Biophysics University of Toronto Medical Sciences Bldg 5358 Toronto, ON M5G 1L7, Canada e-mail: [email protected] chapter 5

Barrie Entsch School of Science and Technology University of New England NSW 2351 Armidale, Australia e-mail: [email protected] chapter 1

Russ Hille Department of Biochemistry University of California, Riverside 1643 Boyce Hall Riverside, CA 92521, USA e-mail: [email protected] chapter 6

Pablo Sobrado Department of Biochemistry Virginia Tech University Blacksburg, VA 24061, USA e-mail: [email protected] chapter 2

Gary Cecchini Department of Molecular Biology VA Medical Center 4150 Clement St San Francisco, CA 94121, USA e-mail: [email protected] chapter 7

Willem van Berkel Department of Biochemistry Wageningen University Dreijenlaan 3 6703-H Wageningen, The Netherlands e-mail: [email protected] chapter 3 Jung-Ja Kim Department of Biochemistry Medical College of Wisconsin 8701 Watertown Plank Road Milwaukee, WI 53226, USA e-mail: [email protected] chapter 4 Takeshi Nishino Department of Biochemistry and Molecular Biology Nippon Medical School Bunkyo-ku 1138602 Tokyo, Japan e-mail: [email protected] chapter 5

Susan Miller Department of Pharmaceutical Chemistry University of California, San Francisco 600 16th St., S512B Genentech Hall San Francisco, CA 94158, USA e-mail: [email protected] chapter 8 Bruce Palfey Department of Biological Chemistry University of Michigan 5220E MSRB3 Ann Arbor, MI 48109, USA e-mail: [email protected] chapter 9 Robert Stanley Department of Chemistry Temple University 130 Beury Hall, 1901 N. 13th St Philadelphia, PA 19122, USA e-mail: [email protected] chapter 10

viii

Contributing authors

Holly Ellis Department of Chemistry and Biochemistry Auburn University 179 Chemistry Building Auburn, AL 36849 e-mail: [email protected] chapter 11 Carlos Gomez-Moreno Department of Biochemistry University of Zaragoza Pedro Cerbuna, 12 50009 Zaragoza, Spain e-mail: [email protected] chapter 12 Jon Stewart Department of Chemistry University of Florida 126 Sisler Hall Gainesville, FL 32611, USA e-mail: [email protected] chapter 13 Hiroshi Ishikita Graduate School of Medicine University of Kyoto Sakyo-ku 6068501 Kyoto, Japan e-mail: [email protected] chapter 14

Matthias Ullmann Department of Structural Biology and Bioinformatics University of Bayreuth Universitätsstraße 30, BGI 95447 Bayreuth, Germany e-mail: [email protected] chapter 15 Erik Schleicher Institute for Physical Chemistry University of Freiburg Albertstraße 21 79104 Freiburg, Germany e-mail: erik.schleicher@physchem .uni-freiburg.de chapter 16 Dongping Zhong Department of Physics The Ohio State University 191 West Woodruff Avenue Columbus, OH 43210, USA e-mail: [email protected] chapter 17

Table of contents

Preface ..................................................................................................................

v

1 The reaction mechanisms of Groups A and B flavoprotein monooxygenases ......................................................................................... 1.1 Introduction ........................................................................................ 1.2 Enzymes acting upon aromatic substrates – Group A........................... 1.2.1 Reactions catalyzed ................................................................... 1.2.2 Protein structures ....................................................................... 1.2.3 Detailed mechanism of PHBH ................................................... 1.2.3.1 Reductive half-reaction ................................................. 1.2.3.2 Oxidative half-reaction ................................................. 1.2.3.3 Hydroxylation chemistry .............................................. 1.2.3.4 Summary ...................................................................... 1.3 Enzymes acting upon non-aromatic substrates – Group B.................... 1.3.1 Reactions catalyzed and subclasses ........................................... 1.3.1.1 BVMOs ........................................................................ 1.3.1.2 FMOs ........................................................................... 1.3.1.3 NMOs .......................................................................... 1.3.1.4 YUCCAs ....................................................................... 1.3.2 Structural features ...................................................................... 1.4 References...........................................................................................

1 1 2 2 3 4 6 9 15 16 16 16 17 19 20 21 22 23

2 Flavin-dependent monooxygenases in siderophore biosynthesis .................. 2.1 Iron, an essential but scarce nutrient ................................................... 2.2 Siderophores ....................................................................................... 2.2.1 Siderophores are important virulence factors ............................. 2.2.2 Structural diversity of siderophores ............................................ 2.3 Flavin-dependent N-hydroxylating monooxygenases........................... 2.4 Catalytic cycle of NMOs ..................................................................... 2.4.1 Flavin reduction in NMOs ......................................................... 2.4.2 Flavin oxidation in NMOs ......................................................... 2.5 Three-dimensional structure of NMOs................................................. 2.5.1 FAD-binding domain ................................................................. 2.5.2 NADPH-binding domain ........................................................... 2.5.3 L-Ornithine-binding domain ...................................................... 2.6 The structural basis of substrate specificity in NMOs ........................... 2.7 Mechanism of stabilization of the C4a-hydroperoxyflavin by NADP+ ... 2.8 Activation of NMOs by amino acid binding ........................................ 2.9 Unusual NMOs ................................................................................... 2.10 High-throughput screening assay to identify inhibitors of NMOs ......... 2.11 Conclusions ........................................................................................ 2.12 References...........................................................................................

29 29 30 30 33 33 34 35 37 37 40 40 41 42 42 44 44 45 46 47

x

Table of contents

3 The flavin monooxygenases ......................................................................... 3.1 Introduction ........................................................................................ 3.2 Occurrence and classification ............................................................. 3.2.1 Amino acid sequence motifs...................................................... 3.2.2 DNA screening .......................................................................... 3.3 Single-component flavin monooxygenases .......................................... 3.3.1 Subclass A ................................................................................. 3.3.2 Subclass B ................................................................................. 3.3.3 Subclass C ................................................................................. 3.3.4 Subclass D................................................................................. 3.3.5 Subclass E.................................................................................. 3.3.6 Subclass F.................................................................................. 3.4 Conclusions ........................................................................................ 3.5 References........................................................................................... 4 Structure and catalytic mechanism of NADPH-cytochrome P450 oxidoreductase: a prototype of the diflavin oxidoreductase family of enzymes ............................................................... 4.1 Introduction ........................................................................................ 4.2 Properties of CYPOR flavins ................................................................ 4.3 Domain structure and function ............................................................ 4.4 Membrane binding domain (MBD) ...................................................... 4.5 FMN domain ....................................................................................... 4.6 Cytochrome P450 binding: role of the FMN domain and connecting domain .......................................................... 4.7 FAD domain........................................................................................ 4.8 Mechanism of hydride transfer ............................................................ 4.9 Interflavin electron transfer .................................................................. 4.10 FMN to heme electron transfer ............................................................ 4.11 P450 catalysis ..................................................................................... 4.12 Other CYPOR electron acceptors ........................................................ 4.13 CYPOR domain movement and control of electron transfer................. 4.14 Physiological functions of CYPOR and effects of CYPOR deficiency ... 4.15 Human CYPOR deficiency (PORD) ..................................................... 4.16 Contribution of CYPOR to inter-individual variation in human drug metabolism ................................................................................. 4.17 Unanswered questions and future directions ....................................... 4.18 References ........................................................................................... 5 The xanthine oxidoreductase enzyme family: xanthine dehydrogenase, xanthine oxidase, and aldehyde oxidase ...................................................... 5.1 Introduction ........................................................................................ 5.2 Overall structures ................................................................................ 5.3 Reaction mechanism ........................................................................... 5.4 Electron transfer from the molybdenum center to other redox-active centers ............................................................................ 5.5 Reaction of FAD with NAD+ or molecular oxygen ...............................

51 51 54 54 55 56 56 58 60 61 62 62 64 65

73 73 75 78 78 79 80 82 82 83 85 85 87 87 90 91 92 92 93

103 103 104 105 113 114

Table of contents

5.6 5.7

xi

Inhibitors of xanthine oxidoreductase .................................................. References...........................................................................................

116 120

6 Assimilatory nitrate reductase ..................................................................... 6.1 Introduction and scope........................................................................ 6.2 Enzyme structure ................................................................................. 6.3 Kinetics and mechanism...................................................................... 6.4 Post-translational regulation ................................................................ 6.5 Interconversion of sulfite oxidase and nitrate reductase activities ........ 6.6 Conclusions ........................................................................................ 6.7 References...........................................................................................

125 125 126 131 133 135 137 138

7 Succinate dehydrogenase (Complex II) and fumarate reductase.................. 7.1 History of Complex II .......................................................................... 7.2 Overview of Complex II ...................................................................... 7.3 Structure of Complex II........................................................................ 7.4 Catalytic assays ................................................................................... 7.5 Catalytic mechanism and domain movement ...................................... 7.6 Electron transfer .................................................................................. 7.7 Quinone-binding site of Complex II .................................................... 7.8 Assembly of the covalent FAD cofactor into Complex II ...................... 7.9 Concluding remarks ............................................................................ 7.10 References ...........................................................................................

141 141 143 144 146 148 151 153 155 159 159

8 Flavoprotein disulfide reductases and structurally related flavoprotein thiol/disulfide-linked oxidoreductases ......................................................... 8.1 Introduction ........................................................................................ 8.2 Group 1 FDR enzymes: classic dithiol/disulfide oxidoreductases with a single CXXXXC disulfide redox center....................................... 8.2.1 Dihydrolipoamide dehydrogenase (LipDH) .............................. 8.2.2 Glutathione reductase (GR) – two new structural studies on this classic member of the group .............................. 8.2.3 Trypanothione reductase (TryR)................................................. 8.2.4 Mycothione reductase (MycR) ................................................... 8.3 Group 2A FDR enzymes – enzymes of the Group 1 structural fold requiring an additional C-terminal Cys-based redox center .......... 8.3.1 Mercuric ion reductase (MerA) .................................................. 8.3.2 High Mr thioredoxin reductases (TrxR and TGR) ........................ 8.4 Group 2B FDR enzymes – low Mr thioredoxin reductase (TrxR) and structurally related enzymes ............................................... 8.5 Group 3 FDR enzymes – enzymes with cysteine sulfenic acid or mixed Cys-S-S-CoA redox center ......................................................... 8.6 Group 4 FDR enzymes – Group 1-fold enzymes catalyzing novel reactions .................................................................................... 8.7 Group 5 FDR enzymes – enzymes with a si side pair of Cys residues widely separated in sequence ................................................ 8.8 References...........................................................................................

165 165 169 174 176 177 177 177 180 181 183 188 192 194 196

xii

Table of contents

9 Flavoenzymes in pyrimidine metabolism.................................................... 9.1 Introduction ...................................................................................... 9.2 Pyrimidine/dihydropyrimidine interconversions ................................ 9.2.1 Overview ............................................................................... 9.2.2 Dihydroorotate dehydrogenases ............................................. 9.2.2.1 General ..................................................................... 9.2.2.2 Mechanisms of the pyrimidine half-reactions ............ 9.2.2.3 Mechanisms of the non-pyrimidine half-reactions ..... 9.2.3 Dihydrouridine synthases ....................................................... 9.2.3.1 General ..................................................................... 9.2.3.2 Pyrimidine half-reactions........................................... 9.2.3.3 Non-pyrimidine half-reaction .................................... 9.2.4 Dihydropyrimidine dehydrogenases ....................................... 9.2.4.1 General ..................................................................... 9.2.4.2 Pyrimidine half-reaction ............................................ 9.3 Methylations .................................................................................... 9.3.1 Overview ............................................................................... 9.3.2 Flavin-dependent thymidylate synthase .................................. 9.3.3 Folate/FAD-dependent methyl transferase (TrmFO) ................ 9.4 References ....................................................................................... 10 Excited state electronic structure of flavins and flavoproteins from theory and experiment ...................................................................... 10.1 Introduction ..................................................................................... 10.2 Flavin photophysics and the electronic structure of its excited states ........................................................................... 10.2.1 Moments of the charge distribution ...................................... 10.2.2 Experimental techniques for the determination of excited state electronic structure....................................... 10.3 Linear dichroism measurements of reduced anionic flavin transition dipole moments and complimentary calculations ............................ 10.4 Experimental studies of excited state electronic structure of flavins and complementary calculations ........................ 10.4.1 Oxidized flavin .................................................................... 10.4.2 Excited state structure of OYE and OYE charge transfer complex___ ....................................................... › 10.4.3 DNA photolyase and Δµ k0 .................................................... 10.4.4 Experimental results for the flavin neutral radical.................. 10.5 Computational studies on flavins ..................................................... 10.5.1 Calculations for oxidized flavins ........................................... 10.5.2 Computational results for semiquinone flavin ....................... 10.5.3 Computational studies on reduced flavins ............................ 10.6 Spectroscopic studies bearing on excited electronic states of flavins 10.6.1 Time-resolved studies of oxidized flavin ............................... 10.6.2 The triplet state of flavins ...................................................... 10.7 Photoinduced electron transfer in flavins ......................................... 10.8 Applications of flavin photochemistry ..............................................

203 203 204 204 206 206 207 210 213 213 214 214 214 214 215 215 215 216 218 221

225 225 227 227 228 229 230 230 232 234 236 238 238 240 241 243 243 244 244 245

Table of contents

xiii

10.9 Conclusions ..................................................................................... 10.10 References .......................................................................................

246 246

11 Structural properties of the alkanesulfonate monooxygenase system that dictate function ....................................................................... 11.1 Introduction ..................................................................................... 11.2 Sulfur limitation in bacterial systems ................................................ 11.3 FMN reductase of the alkanesulfonate monooxygenase system ........ 11.4 Monooxygenase enzymes of the bacterial luciferase family ............. 11.4.1 Structural properties of the bacterial luciferase family ........... 11.4.2 Structural dynamics of alkanesulfonate monooxygenase ....... 11.4.3 Active site structure in the bacterial luciferase family ............ 11.4.4 Catalytic mechanisms of the bacterial luciferase family ........ 11.4.5 Mechanistic properties of alkanesulfonate monooxygenase .. 11.5 Mechanism of flavin transfer ............................................................ 11.6 Conclusions ..................................................................................... 11.7 References .......................................................................................

255 255 256 258 259 259 261 263 265 265 269 271 271

12 Single molecule methods to study flavoproteins ........................................ 12.1 Flavoproteins and electron-transfer reactions ................................... 12.2 Bulk vs. single-molecule methods .................................................... 12.3 Single-molecule techniques for the study of biological systems ........ 12.3.1 Atomic force microscopy ...................................................... 12.3.2 Optical tweezers................................................................... 12.3.3 AFM based force spectroscopy ............................................. 12.3.3.1 The avidin-biotin complex ..................................... 12.3.3.2 Antigen-antibody interaction .................................. 12.3.3.3 Molecular interactions in transient complexes........ 12.4 Single-molecule experiments with flavoenzymes ............................. 12.4.1 Fluorescence measurements ................................................. 12.4.2 Force measurements in flavoproteins .................................... 12.5 References .......................................................................................

277 277 278 279 280 282 284 286 287 287 288 288 290 295

13 Applications of Saccharomyces pastorianus Old Yellow Enzyme to asymmetric alkene reductions ............................................................... 13.1 Identification and initial characterization ......................................... 13.1.1 History of OYE1 ................................................................... 13.1.2 OYE 1 structure and roles of key residues ............................. 13.1.2.1 Histidine 191 and Asparagine 194 ......................... 13.1.2.2 Tyrosine 196 .......................................................... 13.1.2.3 Threonine 37 ......................................................... 13.1.2.4 Tryptophan 116 ..................................................... 13.2 Substrate specificity of OYE 1 .......................................................... 13.2.1 Ketones and aldehydes ......................................................... 13.2.2 Esters .................................................................................... 13.2.3 Nitro alkenes ........................................................................ 13.3 Conclusions ..................................................................................... 13.4 References .......................................................................................

299 299 299 301 302 302 302 304 305 306 312 315 317 318

xiv

Table of contents

14 Contributions of protein environment to the reduction potentials of flavin-containing proteins ...................................................................... 14.1 Introduction ...................................................................................... 14.2 Computation of Esq/hq on the basis of the crystal structures ................. 14.3 Calculation of Esq/hq and determination of redox-linked amino acid residues .......................................................................... 14.4 Influence of the protein backbone conformation on Esq/hq .................. 14.5 Influence of the loop region near the flavin binding site on Esq/hq ....... 14.6 Influence of the FMN phosphate group on Esq/hq................................. 14.7 Conclusions ...................................................................................... 14.8 References ........................................................................................ 15 Methods based on continuum electrostatics and their application to flavoproteins – a review ........................................................................ 15.1 Introduction ...................................................................................... 15.2 The continuum electrostatic model based on the Poisson-Boltzmann equation .................................................. 15.2.1 The physical basis of the Poisson-Boltzmann equation .......... 15.2.2 Electrostatic potentials and electrostatic energies ................... 15.3 Association of flavoproteins .............................................................. 15.3.1 Electrostatic docking of flavoproteins ..................................... 15.3.2 Similarity of electrostatic potentials of proteins ...................... 15.4 Titration behavior of proteins ............................................................ 15.4.1 Microstate model ................................................................... 15.4.2 DTPA – An illustrative example ............................................. 15.4.3 Theoretical analysis of the protonation of flavoproteins ......... 15.5 Recent and upcoming developments ................................................. 15.6 References ........................................................................................

321 321 322 324 325 326 329 332 332

335 335 336 336 339 341 341 342 344 345 346 349 352 354

16 Flavoproteins and blue light reception in plants ........................................ 16.1 Introduction (light reception in plants) ............................................... 16.2 Plant phototropins ............................................................................. 16.2.1 LOV domain structure ........................................................... 16.2.2 LOV photochemistry.............................................................. 16.2.3 LOV signal propagation ......................................................... 16.3 Cryptochromes .................................................................................. 16.3.1 Cryptochrome structure ......................................................... 16.3.2 Cryptochrome photochemistry............................................... 16.3.3 Cryptochrome signal transduction ......................................... 16.4 Outlook ............................................................................................ 16.5 References ........................................................................................

361 361 363 364 365 369 371 371 374 378 380 381

17 Ultrafast dynamics of flavins and flavoproteins ......................................... 17.1 Introduction ...................................................................................... 17.2 Ultrafast dynamics of flavins ............................................................. 17.2.1 Steady-state spectroscopic properties ..................................... 17.2.2 Oxidized flavins ....................................................................

393 393 394 394 396

Table of contents

17.3

17.4

17.5

17.6 17.7

17.2.3 Anionic and neutral radical flavins ........................................ 17.2.4 Anionic and neutral fully-reduced flavins .............................. Electron transfer in model flavodoxin ................................................ 17.3.1 Experiment design, reaction scheme and probing strategy ..... 17.3.2 Femtosecond charge separation, frozen active-site configuration and critical free energies .................................. 17.3.3 Ultrafast charge recombination, vibrational quantum effect and hot ground-state cooling ................................................. 17.3.4 Photoinduced redox cycle, reaction time scales, and vibrational coupling generality........................................ Enzymatic reactions and repair photocycles in DNA photolyases ...... 17.4.1 Dynamics and mechanism of cyclobutane pyrimidine dimer repair by CPD photolyase ............................................ 17.4.1.1 Sequential splitting dynamics of the cyclobutane ring............................................ 17.4.1.2 Electron tunneling pathways and functional role of adenine moiety ............................................ 17.4.2 Dynamics and mechanism of repair of UV-induced (6-4) photoproduct by (6-4) photolyase .......................................... 17.4.2.1 Ultrafast electron and proton transfer dynamics....... 17.4.2.2 Catalytic repair photocycle...................................... Signal transduction in blue-light photoreceptors ................................ 17.5.1 Photoaddition of cysteine to flavin in phototropin.................. 17.5.2 Switching of flavin hydrogen bond in BLUF protein ............... 17.5.3 Ultrafast flavin dynamics in cryptochrome ............................. Conclusions ...................................................................................... References ........................................................................................

Index .................................................................................................................

xv

398 399 401 401 403 404 406 408 408 408 410 414 416 417 419 419 419 420 420 421 429

1 The reaction mechanisms of Groups A and B flavoprotein monooxygenases David P. Ballou and Barrie Entsch

Abstract Flavoprotein monooxygenases, found in species ranging from microorganisms to mammals, transfer one oxygen atom derived from O2 to a substrate, oxidizing it. In this chapter, we review the enzymes in Groups A and B, which accomplish all their chemistry with just one protein. The catalytic cycles of both groups are roughly similar. NAD(P)H reduces the enzyme-bound flavin, which then reacts with oxygen to form a flavin C4a-(hydro)peroxide – the key oxygenating intermediate. The terminal oxygen of the (hydro)peroxide is transferred to the substrate, leaving the hydroxyflavin, which eliminates water to form oxidized enzyme. Catalysis in both groups is strictly regulated, but in very different ways, to limit NAD(P)H oxidase activity. Group A enzymes only allow the fast reaction of NAD(P)H when the substrate to be oxygenated is bound. In contrast, Group B monooxygenases do not require substrate to be present for rapid flavin reduction, but after the flavin is reduced, NAD(P) remains bound, stabilizing the flavin (hydro) peroxide until it encounters the substrate to be oxygenated. The enzymes in Group A are aromatic hydroxylases; they add oxygen to an activated aromatic ring by electrophilic substitution. The most studied flavoprotein monooxygenase, p-hydroxybenzoate hydroxylase, belongs to this group and will be discussed in detail. The enzymes in Group B catalyze nucleophilic and electrophilic oxygenations. Their substrates include aldehydes, ketones, amines, thiols, boronates, selenides, and thioethers. Conformational changes are important for controlling catalysis in both Group A and Group B monooxygenases.

1.1 Introduction Over the past twenty years, a great many flavoprotein monooxygenases from species as diverse as microorganisms and mammals have been reported. These monooxygenases have been classified into six logical groups (A–F) based upon sequence, structure, and function by van Berkel et al. [1]. Most of the enzymes fall within Enzyme Commission classification [2] EC 1.14.13 and EC 1.14.14, which are enzymes that require an external reducing substrate (either NADH or NADPH) to reduce the flavin cofactor as part of catalysis. There are two basic structural themes found with flavin-dependent monooxygenases. In one, both the reaction with the external reductant and that with oxygen and the substrate are carried out by a single protein in which the FAD or FMN cofactor is tightly bound, whereas in the second group, reduction of the flavin (FMN or FAD) is catalyzed by a reductase, and the reduced flavin is delivered to a separate

2

1 The reaction mechanisms of Groups A and B flavoprotein monooxygenases

oxygenase that catalyzes the oxygenation reaction. These two-component oxygenases constitute groups C through F in the van Berkel et al. classification [1] and are reviewed in Chapters by Ellis and by Tu. In this Chapter, we review the enzymes in Groups A and B [1], which achieve reduction and oxygenation with just one protein. Nearly all of the enzymes in Group A are aromatic hydroxylases that are mostly found in bacteria and fungi, and they add oxygen by electrophilic substitution to an activated aromatic ring to form phenolic products. These enzymes are all related in general structure. The most studied flavoprotein monooxygenase belongs to this group – p-hydroxybenzoate hydroxylase (PHBH), EC 1.14.13.2. The enzymes in Group B add oxygen to a wide variety of aliphatic substrates, as well as to heteroatoms such as found in amines, thiols, boronates, selenides, and thioethers. In many cases the reactions involve nucleophilic substitutions carried out by proteins with structural features quite distinct from those enzymes in Group A. In recent years, there have been other reviews of the Group A enzymes; those reviews provide some alternative perspectives on this story [1,3,4].

1.2 Enzymes acting upon aromatic substrates – Group A 1.2.1 Reactions catalyzed In all of the aromatic hydroxylases, the flavin (usually FAD) remains tightly, but noncovalently, bound to the protein as a prosthetic group throughout catalysis, and thus is considered part of the protein structure. In catalysis, the flavin is reduced by NAD(P)H to initiate the reaction. Most of these enzymes require the aromatic substrate to be bound before efficient reduction can occur. The reduced flavin reacts with O2 to form a transient flavin hydroperoxide, which is the reactive species that oxygenates the aromatic ring of the substrate to form the hydroxylated product and water. The known wide range of compounds oxygenated by this class of enzymes can be found under EC 1.14.13.n [2], by referring to enzymes utilizing a flavin prosthetic group. There are roughly forty different reactions identified at present, most of which are from bacteria, but there are certainly many more, because very little research has been done upon the potential enzymes in fungi. The substrates for Group A are all aromatic compounds that are activated for electrophilic aromatic substitution by the presence of ortho- or parahydroxyl or amino substituents. The reactions catalyzed by these enzymes are most often involved in microbial pathways for the degradation of aromatic compounds (especially those derived from lignin) to provide carbon for metabolism. For example, PHBH is a component of the β-ketoadipate pathway in bacteria [5]. The substrate, p-hydroxybenzoate (pOHB), along with a small number of other compounds, induces expression of the enzymes of the pathway, and because of this induction, pOHB can be the sole carbon source for growth. In addition to degradative pathways, a small number of Group A enzymes have recently been found to participate in the biosynthesis of complex metabolites with specialized functions. For example, PhzS is integral to the formation of pyocyanin in Pseudomonas aeruginosa [6]. PhzS is particularly interesting because of suggestions (needing research) that the oxygenation reaction carried out may be mechanistically different from other Group A enzymes. Recently there have been reports of human PHBH-like structures – for example, MICALs are multidomain proteins involved in cytoskeletal rearrangements

1.2 Enzymes acting upon aromatic substrates – Group A

3

[7]. Within the protein sequence is a structure like that of PHBH, and work is underway in several laboratories to find its function.

1.2.2 Protein structures The 3-D structures of at least seven different enzymes from Group A [1] have been deposited in structural databases [8]. Most of these structures are considered variations upon that of PHBH, which was the first published structure of this family; PHBH has been extensively investigated. Therefore, we use PHBH to illustrate the common structural features of this class of enzymes. In the SCOP database [9], PHBH is classified in the Family of FAD-linked reductases, N-terminal domain, within the broad class of α- and β-proteins (with parallel β-sheet). PHBH is a homodimer, with a monomeric length of 394 amino acids. The 3-D fold of the enzyme is complex (򐂰Fig. 1.1); the N-terminal domain binds FAD, while parts of the C-terminal helical domain form the interface between the monomers. The core central domain incorporates a partly buried active site in which the isoalloxazine ring of FAD resides above pOHB (򐂰Fig. 1.1). Strands of the polypeptide interconnect the domains into an integrated structure. A striking feature of the structure is the absence of a separate domain for binding NAD(P)H. Variations on the structure of PHBH have been reported for phenol hydroxylase from a yeast [10] and for 3-hydroxybenzoate 4-hydroxylase [11]. Phenol hydroxylase has an extensive addition to the C-terminus, resulting in a polypeptide of 665 amino acids, about 270 amino acids longer than PHBH. This added polypeptide forms a fourth domain with a thioredoxin-like fold. The thioredoxin structure appears to be

Fig. 1.1: The 3-D folding of a monomer of PHBH. The structure was adapted from 1PBE [8]. The N-terminus is on the left side and develops a domain that binds the FAD (shown as a stick model) in the middle of the structure with an extended conformation. The isoalloxazine ring is in the middle-right with pOHB bound under it. The β-sheet faced with a helix on the right binds the substrate, pOHB. The back of the structure consists of a series of α-helices towards the C-terminus. The red color represents the primary area of interaction between monomers in the dimer structure.

4

1 The reaction mechanisms of Groups A and B flavoprotein monooxygenases

involved in dimer interactions in phenol hydroxylase; these interactions are completely different from those in the dimer interface of PHBH. Aside from dimer formation, no biological significance has been found for the extra domain in phenol hydroxylase and related structures. Importantly, most of the fold of the known catalytic portion of phenol hydroxylase is extremely similar to PHBH.

1.2.3 Detailed mechanism of PHBH Since 1970, a great deal has been published about PHBH, and because there are some common features across all of the Class A flavin monooxygenases (according to the classification of van Berkel et al. [1]), PHBH is a good model for Group A enzymes and for flavoprotein oxygenases in general. In practical terms, PHBH is particularly suitable as a model (note that PHBH from both P. aeruginosa and P. fluorescens are discussed; their structures differ by only two insignificant amino acid residues, and their properties are nearly identical) because PHBH is stable, easily purified, and amenable to experimental investigation. Moreover, the aromatic substrate and many substrate analogues are also stable and readily available. PHBH from P. fluorescens was one of the early examples for which high quality 3-D structures from X-ray crystallography were obtained [12,13], and PHBH from P. aeruginosa was an early candidate for sitedirected mutagenesis studies of the mechanism [14]. The first comprehensive steady-state and rapid kinetics studies of PHBH [15,16] gave important clues to mechanistic features of the enzyme. For example, even though it is a favorable exergonic reaction, the FAD in PHBH and other enzymes in Group A are only reduced very slowly unless the aromatic substrate that is to be oxygenated is present. Upon addition of pOHB and O2, the reaction proceeds rapidly and quantitatively to form product (3,4-dihydroxybenzoate [3,4DOHB]) until pOHB (if it is the limiting substrate) is totally consumed. Importantly, the 3,4DOHB product is not further oxygenated even though a second oxygenation (to form 3,4,5-trihydroxybenzoate) is thermodynamically favorable. Another property of the Class A monooxygenases that was demonstrated from the kinetics studies was that the reaction with oxygen did not occur until NADP was released from the enzyme. Thus, catalysis occurs in two parts; first, NADPH reduced the FAD rapidly, but only with pOHB bound to the enzyme; second, after NADP dissociates, the bound FADH− reacts with oxygen in the presence of bound pOHB to form product, oxidized FAD, and water. If NADPH were able to facilely reduce the FAD in the absence of substrate, the FADH− would react rapidly with O2 to form H2O2 (a toxic reactive oxygen species) and consume NADPH wastefully. Because this reaction with oxygen is even more energetically favorable than is flavin reduction, regulation of the reduction of PHBH is particularly important. Thus, PHBH has evolved to prevent the consumption of NADPH and oxygen unless pOHB is bound to the enzyme. We note that other flavoprotein oxygenases not in Group A employ a variety of strategies to limit the unwanted reaction with oxygen. Later in this Chapter is described a different strategy used by enzymes of Group B to avoid wasteful use of NAD(P)H. PHBH has a turnover number of ~45 s−1 under optimum conditions at 25ºC, implying that the rates of formation and decay of intermediates are considerably > 45 s−1. Thus, to successfully resolve and characterize transient events by stopped-flow techniques,

1.2 Enzymes acting upon aromatic substrates – Group A

5

PHBH had to be studied at 3 to 4ºC. PHBH was an excellent candidate for detailed transient kinetic analysis [17,18], and it was one of the first multiple-step enzymes for which kinetics and spectral properties of intermediates were characterized in great detail. Its organic substrates (NADPH and pOHB) change absorbance during catalysis, and FAD, as the core catalytic factor, undergoes unique absorbance and fluorescence changes. The catalytic cycle was easily divided into two separate events. In the absence of oxygen, the processes of control and catalysis of the reduction of FAD could be studied with the enzyme, NADPH, and pOHB [17]. Then, by using enzyme in complex with pOHB, reducing it anaerobically with dithionite, and then mixing it with a solution of oxygen, the oxygenation process and its control could be studied [18]. A striking feature of the overall enzymatic catalysis under optimum conditions (pH 7.5 to 8.0) is the rate-determining step as illustrated in 򐂰Fig. 1.2. During the reductive half-reaction, a charge-transfer complex between FADH− and NADP is observed beyond 700 nm (in the infrared) before NADP is released from the enzyme [16]. 򐂰Fig. 1.2 shows an enzyme-monitored turnover experiment with measurement at 790 nm. Very rapidly during the reductive half-reaction (pre-steady state conditions) a charge-transfer complex forms. The reaction settles into steady-state, and it can be seen that a substantial portion of the enzyme continues to exhibit the charge-transfer interaction described above. If the reaction is monitored at 450 nm (data not shown), the absorbance is weak, indicating that most of the flavin is reduced during turnover. Thus, the process that releases NADP from the reduced enzyme is the ratedetermining step. This slow step does not involve a chemical reaction, and thus likely involves a conformational change in the enzyme (more later). 0.02

Absorbance (790 nm)

0.015

0.01

0.005

0 0.01

0.1

1

10

Time (seconds)

Fig. 1.2: The rate-determining step in catalysis by PHBH. The traces shown were collected in an enzyme-monitored turnover experiment conducted at pH 8 and 4°C. Absorbance at 790 nm is characteristic of the charge-transfer complex between FADH− and NADP+. The red trace shows the change upon reduction of PHBH by NADPH in the presence of pOHB (no oxygen). The blue trace shows the same enzyme concentration in turnover with oxygen as the limiting substrate. Approximately 65% of the enzyme is in the charge-transfer complex during steady-state (the flat trace between 0.1 and 3 s).

6

1 The reaction mechanisms of Groups A and B flavoprotein monooxygenases

1.2.3.1 Reductive half-reaction Many flavoproteins exhibit the property of pyridine nucleotide reductases, but only a small number have a strict control mechanism (like PHBH) over this exergonic process. Also, PHBH and similar enzymes have the unusual property of having no classical binding domain for NADPH. It is these features of PHBH that have been partly responsible for much research attention. Remarkably, in PHBH, the rate constant for reduction of FAD by NADPH in the presence of pOHB is more than 105-fold greater than that in the absence of pOHB [16], yet the redox potential of FAD is almost the same with or without pOHB bound to the protein [19]. Thus, the enzyme must control the approach and orientation of NADPH to FAD through the binding of pOHB, but how? Early in the study of PHBH, it was found that several structural analogues of pOHB bound in the active site, and some stimulated the reductive half-reaction and some did not [15]. Critical to stimulating the reduction of the flavin by NADPH was the presence of a dissociable proton at the 4-position of the aromatic ring. The outstanding exception was the product (3,4DOHB), which has such a dissociable proton, but stimulated reduction only about 100-fold. This stimulation of reduction that depends on having a dissociable proton on the substrate or substrate analog is frequently observed, even when no hydroxylation is involved. Such compounds are often called effectors. If substrate analogues that stimulated reduction but did not become oxygenated [15] were present in high concentrations, the enzyme would be stimulated to catalyze the rapid formation of H2O2 and consume NAD(P)H unproductively, thus creating a lethal condition. This dual control on the use of NAD(P)H (substrate binding combined with mobile substrate proton signal) has not been demonstrated in other Group A enzymes. An explanation of the control of reduction in PHBH was realized via a combination of X-ray crystallography and site-directed mutagenesis. The high-resolution structure of wild type PHBH [13] with pOHB bound to the enzyme gave no clue to how the reductive half-reaction was catalyzed because there was no place to pack NADPH near FAD; the FAD was fully buried in the active site in a conformation we refer to as “in”. Then, in 1994, the structure of PHBH with 2,4-dihydroxybenzoate (2,4DOHB, an alternative substrate) bound in the place of pOHB showed that the enzyme could adopt a conformation with the isoalloxazine ring rotated about the ribityl side chain so that the N5 position (site of hydride transfer from NADPH) moved ~7 Å and was exposed at the surface of the protein (򐂰Fig. 1.3) [20,21]. This conformation (termed the “out” conformation) could provide a way for NADPH to approach FAD. Site directed mutagenesis studies indicated surface residues potentially involved in binding NADPH [22,23]. It was not possible to obtain a structure of wild type PHBH with NADPH bound but without pOHB, because it did not crystallize in a single, stable conformation, suggesting that it is in a dynamic equilibrium. Finally, in 2002, a structure of the R220Q variant of PHBH with NADPH bound was obtained in the absence of substrate so that virtually no reduction of the FAD occurred [24]. The R220Q variant freezes an open conformation of the enzyme that can bind NADPH and likely is involved in substrate binding. In this conformation, two domains separate and the isoalloxazine swings out slightly from the active site to open up a solvent channel for substrate (or product) to access the active site. This third conformation has been referred to as “open”. NADPH was bound in an elongated state along a cleft on the surface of PHBH without any

1.2 Enzymes acting upon aromatic substrates – Group A

7

A

B

Fig. 1.3: Selected groups in the active site of PHBH that illustrate the in and out conformations. (A) The active site formed by the complex with pOHB – the in conformation (from 1PBE [8]). The orientation shows the re face of the isoalloxazine and the hatched shading shows the solvent accessible areas. (B) The active site formed by the complex with 2,4DOHB – the out conformation (from 1DOD [8]). The orientation is identical to A, and the hatched shading again shows the solvent accessible area. The principal difference is the position of the isoalloxazine ring.

canonical nucleotide domain fold being present. This conformation did not allow the nicotinamide ring to approach FAD in this “open” conformation, but modeling FAD into the structure in the “out” conformation (see above), where it is exposed to solvent, showed that a small rotation of the nicotinamide ring allowed it to meet the isoalloxazine in an orientation that was suitable for hydride transfer [24]. Thus, we have a rational explanation of how at least three different protein conformations (“in”, “out”, and “open”) control the spatial arrangements of NADPH and FAD required to regulate the initiation of reduction. The crystal structures illustrate the complex dynamics that are necessary for the function of the enzyme, but do not help us understand what makes them occur.

8

1 The reaction mechanisms of Groups A and B flavoprotein monooxygenases P293 O –O

Y385

H O

H

O

O

H

O

H

Y201

H72

H O

H

O

H

N + NH

H

pK a 쏁 7.1

H쎵

P293 O –O O O

H

O

H

O

H

Y201

H72

H

Y385 O

H

O

H

N NH

H

H쎵 O –O O

P293 Y385 – O

H

O

H

Y201

O

H

H72

H O H

H

O

H

N N

Fig. 1.4: Diagrammatic representation of the H-bond network linking pOHB to the surface in PHBH. The arrangement of structures to form a connection of H-bonds that can exchange protons with solvent is described in detail in [46]. The figure shows how the pKa of H72 can control the ability of the network to control the interaction of the H at position 4 of pOHB. This mechanism is an integral component of both the reductive and oxidative half-reactions of PHBH. This figure is reprinted with permission from [43]. Copyright 2004 American Chemical Society.

The high-resolution structure containing pOHB (the “in” conformation) [13] showed that the 4-hydroxyl of the substrate was H-bonded in a potential network of H-bonds that connected to H72 at the surface of the protein, as illustrated diagrammatically in 򐂰Fig. 1.4. The function of these H-bond interactions was tested by specific variants: Y201F, Y385F, and H72N, in combination with structural and kinetic analysis [19,25,26]. Structural analysis showed that only the network of H-bonds was disrupted in the variants. Detailed kinetics studies of the reduction reactions (including using NADPD) showed that each substitution disrupted enzyme reduction to a different degree. Other important evidence came from disruptions to the mobility of the ribityl side chain of FAD [27,28]. When the evidence was integrated, the following model for the reductive half-reaction emerged [25,29]. When NADPH and pOHB bind to the enzyme, pOHB is transiently and rapidly deprotonated by the H-bond network (򐂰Fig. 1.4). The phenolate form of pOHB triggers a

1.2 Enzymes acting upon aromatic substrates – Group A

9

conformational rearrangement in the protein through the highly conserved loop behind the isoalloxazine ring, resulting in movement of the isoalloxazine from the “in” position to the “out” conformation. The full details of the protein conformational changes are not yet understood. The nicotinamide ring of NADPH bound on the surface then rotates to meet the isoalloxazine ring, forming a charge-transfer interaction (observed by transient absorbance at long wavelengths) that is competent for reduction. Hydride transfer occurs from the proR position of NADPH to the re face of the isoalloxazine ring [30]. After hydride transfer, the order of events is not clear. However, the FADH− rotates back into the “in” conformation, probably under the influence of the very positive electrostatic field around the interior position [31], pOHB is re-protonated from the H-bond network and the NADP dissociates from the enzyme. Note that the conformational changes occurring after reduction of FAD make up the rate-determining step in catalysis (򐂰Fig. 1.2). This control over FAD reduction by the presence of a mobile proton on pOHB is probably unique to PHBH. The structures of other related enzymes (e.g., phenol hydroxylase) do not show similar H-bond networks. We have suggested [15,25] that this feature of PHBH is an important evolutionary adaptation of the enzyme to discriminate between pOHB and p-aminobenzoate, a molecule that is essential for cellular function (for synthesis of folate). Because the aminobenzoate has no easily dissociable proton at the 4-position, it does not stimulate reduction of the FAD in PHBH and therefore, the subsequent oxygenation of this substrate to 3-hydroxy 4-aminobenzoate, a cytotoxic molecule, is avoided. The final (and perhaps most significant) feature of the control by the enzyme of the reduction of FAD is the dramatic change in the kinetics of binding of pOHB. At 4ºC, the rate constant for binding pOHB to oxidized enzyme is 5 × 106 M−1s−1, and the corresponding dissociation rate constant is 60 s−1. For the reduced enzyme (with almost the same Kd for pOHB) the corresponding values are only 1.6 × 102 M−1s−1 and 0.0034 s−1 [18]. Thus, dissociation of pOHB from the reduced enzyme does not occur during a catalytic cycle, and this feature prevents uncoupling and oxidase activity. The reduced structure with pOHB kinetically stabilizes the “in” conformation. This means that pOHB is locked in the active site for the reactions with oxygen. This active site trap appears to be a unique property of the Group A hydroxylases in general. With almost no structural changes between the oxidized and reduced enzyme forms [32], this trap must be due to the negative charge on the reduced isoalloxazine ring positioned in the “in” conformation within the strong positive electrostatic potential of the active site [31].

1.2.3.2 Oxidative half-reaction The fundamental function of hydroxylases involves the reactions with oxygen. With PHBH and related enzymes, there are at least three consecutive chemical processes during the oxidative half-reaction: reaction with oxygen to form a flavin hydroperoxide, transfer of one atom of oxygen to a substrate, leaving the hydroxyflavin, and finally, the elimination of water from the hydroxyflavin to form its starting oxidized state. This mechanism, accepted universally today, was a complete mystery in 1970. Most of our knowledge about the mechanism of the oxygen reactions of flavin-dependent hydroxylases comes from work on PHBH. A fundamental chemical property of flavins is their reactivity with oxygen [33], which can be attributed to their ability to form free radicals. Oxygen has a radical ground

10

1 The reaction mechanisms of Groups A and B flavoprotein monooxygenases

state structure, so chemical interactions involve radical species to preserve electron spin (Vol. I Chapter 10 by Jorns). The initial reaction is always the transfer of one electron from the anionic reduced flavin to oxygen to form a radical pair. The radical pair can exchange a second electron to form H2O2 and oxidized flavin, the reaction pathway channeled by flavoprotein oxidases, or the radical pair can diffuse apart, the common outcome of the accidental reaction of flavoprotein reductases with oxygen to form O2− • and flavin semiquinone. Finally, the radical pair can combine and form a covalent bond, thus creating flavin peroxide, the pathway channeled by flavoprotein hydroxylases [33]. The final pathway was only speculation in 1970, as was the location of the oxygen adduct on the isoalloxazine structure. In aqueous solutions, free flavin reactions with oxygen are even more complicated because of the development of autocatalytic kinetics due to radical recombinations [34]. The first indication of a new chemical species between flavin and oxygen in enzymatic reactions was published in 1972 [17], but it took extensive experimental work over the following years to fully resolve the kinetics and the properties of transient oxygen derivatives of FAD in the function of PHBH [18]. Subsequently, in the Massey laboratory, the same types of oxygen derivatives were also found for both melilotate hydroxylase [35] and phenol hydroxylase [36]. A typical example of the characteristic absorption spectra of flavin species in these reactions is shown in 򐂰Fig. 1.5; these spectra were deduced from stopped-flow studies of the reaction of O2 with PHBH that had been reduced in the presence of the

Absorbance

0.3

e 8.5

e 10.2 e 8.0

0.2

0.1

350

400

450

500

Wavelength (nm)

Fig. 1.5: Absorption spectra of FAD obtained when reduced PHBH in complex with pOHB was reacted with oxygen. The experiment was carried out at pH 6.55 and 2°C. The solid blue curve is of reduced enzyme and the dashed black curve of oxidized enzyme. The spectrum marked by the red circles was calculated from reaction traces recorded at multiple wavelengths as described in [37] and represents the flavin C4a-hydroperoxide formed from oxygen. The calculated spectrum marked by the green circles represents the flavin C4a-hydroxide formed from the hydroperoxide following oxygenation of pOHB. The extinction values in the previously published figure are incorrect. The correct extinction values at the spectral peaks (ε on the figure) were obtained by multiplying the original figure values by 0.9. This figure is reprinted by licence from [37]. Copyright 1989 Elsevier Ltd.

1.2 Enzymes acting upon aromatic substrates – Group A

11

substrate, pOHB. The first stable oxygen derivative of a flavin was the C4a-peroxyflavin demonstrated for bacterial luciferase ([38] and Vol. I Chapter 5 by Tu). The chemical structure of the flavin derivatives with oxygen was inferred from extensive model chemistry of flavin derivatives [39], but it was many years later that direct proof was obtained from NMR studies on the stable derivative in luciferase that the oxygen adduct of the flavin was actually a C4a-(hydro)peroxyflavin [40]. Over the last 40 years, all experimentally observed adducts between flavin and oxygen in proteins have been C4aperoxides and C4a-hydroxides, as shown in the oxygen reactions of PHBH (򐂰Fig. 1.6). To experimentally study the oxygen half-reaction of PHBH and similar enzymes, the multi-step process was studied directly by stopped-flow spectrophotometry and fluorimetry, combined with low temperatures, pH shifts, changes in solvent conditions, and mutations. Combinations of these conditions have been used to modulate the kinetics of this exergonic series of consecutive reactions, and thus define the chemistry. Spectra like those in 򐂰Fig. 1.5 were often calculated from individual kinetic traces by hand using integrated rate equations for consecutive reactions [18,35]. Today, computer software is available to rapidly deconvolute the component spectra from data consisting of rapidly scanned spectra with time. However, although these tools are convenient and impressive, it is important to use them with caution, since incorrect assumptions about the mechanistic models used will lead to incorrect outputs. In the absence of pOHB bound in the active site, reduced PHBH behaves like an oxidase, forming H2O2 in a second-order reaction with oxygen without any observed flavin derivatives [18]. However, by selecting appropriate solvent conditions, especially by including anions such as azide, the kinetics can be manipulated to better resolve intermediates. By these means, it was found that even the reaction without substrate present actually proceeded through a transient C4a-hydroperoxide [18]. It has recently been found that some oxidases also react via this pathway ( [41] and Vol. I Chapter 9 by Wongnate and Chaiyen). Note that the first electron transfer from flavin to oxygen is always found to determine the rate of formation of the flavin peroxide. The reaction of oxygen with reduced PHBH is 103-fold faster than that with free flavin [42], and some hydroxylases react even faster. With pOHB present, the reaction of reduced PHBH with oxygen also forms quantitatively the same C4a-hydroperoxide, but 10-fold faster. In specific protein environments, the kinetic barriers to reaction with oxygen are dramatically lowered (see Vol. I Chapter 10 by Jorns). With PHBH, the secondorder rate constant for the reaction of reduced PHBH with oxygen is independent of pH, consistent with the anionic reduced flavin being present in the active site at all practical pH values [43]. Group A enzymes like PHBH are different from many other hydroxylases that react preferentially with oxygen without substrate to form a flavin (hydro)peroxide that has a half-life of many seconds to minutes – such as the enzymes in Group B described below. In PHBH, pOHB in the complex with reduced enzyme helps to prevent access to solvent [13] and this barrier is essential for extending the half-life for the flavin hydroperoxide to ~1 s, which is adequate for the hydroxylation reaction [19]. From several pieces of indirect evidence, it is clear that in PHBH, the flavin C4aperoxide initially formed in the reaction with oxygen is immediately protonated to form the flavin C4a-hydroperoxide, which is an electrophile, and this process is common to Group A enzymes. The proton probably comes into the active site from a solvent channel into the re side of the flavin [44]. Only with the K297M variant of PHBH (located

O

NH

OH

– N O

H2O 쎵 3,4 DOHB 쎵 E • FAD

–O

O

N H

N

O2

–O

O

E

–O

O

B N

O

N

OH

N H OH

N

N HO –O O

N

– O

NH

OH

NH

O

O

–O

O

D

–O

O

C

N

O OH H

N – HO

N

N

O OH

N HO

N

O

NH

– O

NH

O

O

Fig. 1.6: Structural representation of the catalytic cycle of PHBH. A, Reduced flavin; B, C4a-peroxyflavin; C, C4a-flavin hydroperoxide, D, C4a-flavin alkoxide with dienone form of the product; E, C4a-flavin hydroxide. In experiments with pOHB bound to PHBH, the non-aromatic dienone intermediate is not detected because the rate constant for decomposition to 3,4DOHB is much larger than that of formation. Note that the oxygen derivatives of FAD are bound to the bridging C4a-carbon. Oxygenations of electrophilic substrates such as in Baeyer-Villiger reactions, involve the nucleophilic flavin peroxide, Species B.

pOHB 쎵 NADPH

NADP쎵

A

12 1 The reaction mechanisms of Groups A and B flavoprotein monooxygenases

1.2 Enzymes acting upon aromatic substrates – Group A

13

where it can disrupt the proposed solvent channel) has some indication been found of a protonation step slow enough that it could be measured by stopped-flow fluorescence [43]. By contrast, some enzymes in Group B specifically inhibit protonation of the peroxide [45], since, unlike the aromatic hydroxylases, the nucleophilic peroxide is the catalytically active species. The next phase in the oxygen reactions is the transfer of the distal oxygen atom of the hydroperoxide to the substrate. To understand PHBH, it is essential to consider this process with respect to the H-bond network (򐂰Fig. 1.4). As mentioned above, after the isoalloxazine has been reduced, the enzyme-bound FADH− is in the “in” conformation with pOHB bound with its 4-hydroxyl protonated, and in juxtaposition to the isoalloxazine ring for the ensuing hydroxylation reaction. After reaction of the bound FADH− with O2 to initially form the flavin C4a-peroxide, proton exchanges occur, largely facilitated by the hydrogen bond network. This leads to a deprotonation of the more acidic substrate to its nucleophilic phenolate form and protonation of the basic flavin peroxide to yield its electrophilic hydroperoxy form. The positive electrostatic field in the active site appears to be able to support approximately one negative charge, so that the overall exchange of protons preserves charge balance while setting up the active oxygen species and the substrate for the hydroxylation step. These processes are practical because the rate of proton exchange is much faster than other chemical processes occurring [46]. Measurements of the rate constants for hydroxylation as a function of pH showed a pKa of 7.1, which we have attributed to H72 at the surface of the protein at the end of the hydrogen bond network (򐂰Fig. 1.4 and [43]). A limiting rate constant at low pH for hydroxylation at ∼18 s−1 contrasted with a limiting value of ∼120 s−1 at high pH. Several lines of evidence are consistent with the hydroxylation occurring by electrophilic substitution. For example, the disruption of the H-bond network in the Y201F variant (Y201 directly bonds to pOHB, see 򐂰Fig. 1.4) decreases the rate of hydroxylation by 1000-fold [19], such that the enzyme becomes an oxidase that primarily forms H2O2. In addition, PHBH does not effectively hydroxylate p-aminobenzoate, which does not have a dissociable proton at the para position. Hydroxylation of p-aminobenzoate occurs with a rate constant of only 5 s−1, which is independent of pH [43]. At higher pH in the p-aminobenzoate reaction, the enzyme becomes more of an oxidase because these conditions cause loss of the proton at N5 of the isoalloxazine ring, promoting elimination of H2O2 [41]. These examples illustrate how hydroxylation of substrate is a competition between oxygen transfer and peroxide elimination from the flavin. PHBH may be unique in being nearly 100% coupled with its native substrate. Thus, for each mole of NADPH oxidized, one mole of 3,4DOHB is produced and one mole of O2 is consumed. If a substrate analogue that is not susceptible to electrophilic attack by the hydroperoxide is used with PHBH (such as 5-hydroxypicolinate or 6-hydroxynicotinate), then the enzyme becomes a pure NADPH oxidase [18] as mentioned earlier. With pOHB as substrate, the hydroxylation step is experimentally detected as a firstorder reaction that transforms the flavin hydroperoxide into the closely related flavin C4ahydroxide [18], the logical product of oxygen transfer to the substrate. The spectrum of this species is similar to that of the hydroperoxide, but the absorbance maximum is slightly blue-shifted (򐂰Fig. 1.5). A useful tool for studying the hydroxylation reaction has been fluorescence, because in most cases with PHBH (as well as many other flavin-dependent

14

1 The reaction mechanisms of Groups A and B flavoprotein monooxygenases

hydroxylases), the flavin hydroperoxide lacks fluorescence, but the flavin hydroxide is highly fluorescent. Thus, the appearance of fluorescence helps to identify the phase of the reaction that is due to oxygen transfer. We note that the observed single phase for hydroxylation is actually composed of two chemical processes – the measured rate of oxygen transfer to form a non-aromatic dienone product, and the very fast tautomerization to the aromatic product (򐂰Fig. 1.6), probably with the assistance of the H-bond network [43]. When 2,4-dihydroxybenzoate is substrate for PHBH, we know that the dienone intermediate is involved, because a spectral intermediate with a high extinction can be resolved [18]; after a long period of research, it was concluded that this spectral species is actually due to contributions from both the flavin C4a-hydroxide and the dienone form of the product (򐂰Fig. 1.6 and [47]). This dienone does not rapidly rearrange to the aromatic product in the active site of the enzyme. Similar kinetically stabilized dienone-like intermediates have also been observed with p-aminobenzoate as substrate for PHBH (in this case, it has extreme stability [18]) and with resorcinol as substrate for phenol hydroxylase [48]. Observation of such intermediates provides strong evidence for the chemistry of hydroxylation being an electrophilic aromatic substitution. The final tasks for flavoprotein hydroxylases are to eliminate water from the flavin and release the aromatic product into solution. For many enzymes, release of the final product is the rate-determining step in catalysis. However, for PHBH under optimum conditions (򐂰Fig. 1.2), release of NADP from the reduced enzyme flavin limits the rate of catalysis, although at pH values below 7, the release of 3,4DOHB limits turnover [43]. It was noted in early studies of PHBH that at high concentrations, substrate inhibits catalysis by forming a dead-end complex with the flavin still as the C4a-hydroxide [15]. This suggested that product is usually released from the enzyme prior to water elimination, and that after product dissociates, high concentrations of substrate can quickly replace it, thereby isolating the flavin from solvent before H2O could be released from the C4a-hydroxyflavin; the latter step is normally fast when solvent has access to the flavin [39]. Binding of substrate and release of product likely involve similar changes between the “in” and “open” conformations (see above). However, when pOHB is bound in the active site, it likely favors the “in” conformation, which prevents access of solvent and the consequent rapid dehydration of the flavin hydroxide. When reaction kinetics are studied under conditions that avoid high substrate inhibition, the rate constant for product release is dependent upon pH [43], with a pKa of 7.1, indicating again a role for the H-bond network in this phase. The rate of product release is substantially greater at high pH, implying that the 3,4DOHB phenolate is an important trigger for this process. H-bond network mutants lose this pH-dependent trigger [43]. The structure of the product complex of PHBH shows that the 3-hydroxyl of the product is H-bonded to the backbone carbonyl of P293, causing a slightly different orientation from that of the substrate complex [51]. This interaction is important because it prevents reorientation of the product in the active site and the consequent stimulation of the reduction of FAD and hydroxylation of the product. The particular interaction of product with P293 (an essential residue that interacts with substrate to promote the “out” conformation necessary for reduction of the FAD [50]) probably does not favor the “out” conformation. Instead, we surmise that the subtle changes in the protein caused by interaction with the product promote the “open” conformation, which facilitates product release. Although the crystal structure of PHBH with 3,4DOHB bound shows that the “in” conformation is formed in the crystal [51], in solution, evidence from pH dependence (above) suggests

1.2 Enzymes acting upon aromatic substrates – Group A

15

that mobilization of the 4-hydroxyl proton of 3,4DOHB through the H-bond network provides a trigger to promote formation of the “open” conformation and thus completion of catalysis [43]. Single molecule studies have shown that, in the absence of substrate, the protein structure is in rapid equilibrium between “open” and “in” conformations [52], and these dynamics are probably the reason substrate-free PHBH does not form crystals suitable for high-resolution structures.

1.2.3.3 Hydroxylation chemistry It is remarkable that PHBH and other enzymes in Group A can oxygenate aromatic rings with a flavin hydroperoxide. In the section above, some evidence is presented for a mechanism of electrophilic attack of the terminal oxygen of the flavin hydroperoxide upon the activated aromatic ring of the substrate. Perhaps the strongest evidence for this electrophilic substitution reaction comes from a study of fluorinated substrates of PHBH that are converted into orthoquinones [53] and from a Hammett relationship study using 8-substituted-FADs [54]. PHBH at 4°C and pH 9 hydroxylates pOHB with a rate constant of 120 s−1. We have described above how manipulation of the phenolic proton of pOHB in the H-bond network of PHBH can contribute an acceleration of up to 103. This contribution is not available to the same extent in other Group A enzymes such as phenol hydroxylase [55], which has no H-bond network, exhibits no pH dependence of hydroxylation, and is at least one order of magnitude less active than PHBH in the hydroxylation reaction. There must be other chemical contributions to lowering the activation energy for the hydroxylation reaction in other Group A monooxygenases, and some answers have come from studies with PHBH. The role of the strong positive electrostatic potential around the active site of PHBH in the largely solvent-free environment in the “in” conformation has been studied [31]. A series of highly conserved basic residues surround the active site with positive charges, and we have studied the effects of two mutations, K297M [31] and E49Q [49] that are not integral to substrate interactions with enzyme. The removal of a positive charge (K297M) decreased the rate constant for hydroxylation at pH 9 from 120 s−1 for wild-type to 5 s−1 for the mutant enzyme. Increasing the positive potential by removing a conserved negative charge (E49Q) had the opposite effect. The rate constant for hydroxylation became larger than that for the preceding reaction with oxygen, even at the highest concentrations of oxygen obtainable, and could not be measured [49]. It was proposed that the positive electrostatic potential stabilizes the flavin C4a-alkoxide leaving group formed during hydroxylation [43], resulting in faster oxygen transfer. Thus, it is clear that the positive electrostatic potential in the active site is a major contributor to catalysis. A number of studies of PHBH with FAD analogues chemically modified in the isoalloxazine ring have also drawn attention to the contribution of the isoalloxazine structure. For example, PHBH with 1-carba-1-deaza-FAD [56] converted the enzyme into an NADPH oxidase, even though during catalysis, substantial amounts of flavin C4ahydroperoxide were formed. The results obtained could be due to a combination of a less stable hydroperoxide and less active oxygen on the modified isoalloxazine. Analysis of a series of 8-substituted FADs bound to PHBH [54] showed that hydroxylation rates increased with more electron-withdrawing substituents, presumably by stabilization of the flavin C4a-alkoxide leaving group in hydroxylation. This work demonstrated an

16

1 The reaction mechanisms of Groups A and B flavoprotein monooxygenases

important role for the electronic properties of the isoalloxazine in hydroxylation and fully supported an electrophilic aromatic substitution mechanism. It is clear that a combination of susceptible substrate, substrate activation, protein environment (positive electrostatic field and alignment of catalytic groups, etc.), and flavin chemistry combine to produce high rates of hydroxylation in PHBH, making it an effective enzyme.

1.2.3.4 Summary PHBH (and similar enzymes) function to efficiently achieve several consecutive and diverse chemical reactions in turnover within a single polypeptide. For PHBH to be successful, we have found that it has to integrate at least four rapid conformational changes on the fly in catalysis over two active sites. Two of these conformational changes are rate determining (above). The importance of the delicate balance between conformations has been demonstrated in several studies of carefully chosen single mutations in the structure that can disrupt the dynamic properties of the protein so that its catalytic function is substantially impaired [57,58].

1.3 Enzymes acting upon non-aromatic substrates – Group B 1.3.1 Reactions catalyzed and subclasses Flavin-dependent oxygenases of Group B [1] are involved in a wide variety of both electrophilic and nucleophilic oxygenations. There are four general classes of enzymes within Group B: the Baeyer-Villiger monooxygenases (BVMOs), the flavin-containing monooxygenases (FMOs), the microbial N-hydroxylating monooxygenases (NMOs), and the YUCCA enzymes that are found in plants. By-and-large, all of the enzymes in this group use NADPH as the physiological reductant. A characteristic that distinguishes Group B flavin-dependent monooxygenases from those in Group A is that in most cases, reduction of the enzyme-bound flavin of Group B enzymes does not require substrate to be present. Thus, control of catalysis does not occur by down-regulating reduction of the flavin in the absence of substrate, as was discussed above for the enzymes in Group A. Instead, the enzyme-bound FAD readily becomes reduced by NADPH and then reacts with O2 to form a quasi-stable C4a-FAD (hydro)peroxide. Like the aromatic hydroxylases in Group A, these enzymes also utilize C4a-flavin (hydro)peroxides as activated forms of oxygen to carry out their oxygenation reactions. However, their kinetic mechanisms differ from those of Group A. For effective catalysis, after the FAD is reduced, the NADP product remains bound to the enzyme throughout the remainder of the catalytic cycle, whereas with Group A enzymes the NADP is released after the flavin is reduced and before the FADH– reacts with O2. With Group B enzymes, control of catalysis is manifest after the reaction with O2 forms the C4a-(hydro)peroxy-FAD. The bound NADP is critical for helping to maintain a quasi-stable flavin (hydro)peroxide, a general theme for Group B flavin-dependent monooxygenases; in fact, in the presence of NADP, many of the C4a-(hydro)peroxides of Group B enzymes have half-lives of several minutes at 25ºC and 2 hours at 4ºC. Without substrates present, it is these long half-lives for the C4a-FAD-(hydro)peroxides

1.3 Enzymes acting upon non-aromatic substrates – Group B NADP쎵 E FADH앥

NADPH E FADox

O2

NADPH

NADP쎵 E FADOO(H)

E FADox NADP쎵 SOH

17

S NADP쎵 E FADOH SOH

NADP쎵 E FADOO(H) S

Fig. 1.7: Schematic representation of the kinetic scheme for the Group B flavin-dependent monooxygenases. This has been called the bold mechanism in which the activated oxygen is formed before the substrate binds [4].

that prevent rapid turnover and consequent oxidase activity. This same tactic is also used by several of the two-component flavin-dependent oxygenases that are described in Chapter 11 (Ellis et al.) and Vol. I Chapter 5 by Tu. A schematic of the general mechanism for Group B oxygenases is shown in 򐂰Fig. 1.7. This has been called a bold mechanism [4] because the activated form of oxygen, the C4a-flavin (hydro)peroxide, is formed before the enzyme encounters the substrate to be oxygenated. Thus, the C4-adduct is (as Henry Kamin has suggested) like a “cocked gun” ready to go off as soon as substrate appears.

1.3.1.1 BVMOs The BVMOs catalyze the conversion of a ketone to an ester by inserting an oxygen atom between the carbonyl and the adjacent methylene group as shown in 򐂰Fig. 1.8 for cyclohexanone. This process involves a nucleophilic C4a-FAD-peroxide attack on an electrophilic carbonyl functional group to produce a tetrahedral Criegee adduct that rearranges to form the ester (or lactone) product, as shown for the cyclohexanone monooxygenase (CHMO) reaction [45,59]. This mechanism is similar to that of the classic Baeyer-

A

B

C

R N

D

R N

N HO –O O O

O

N

NH – O

N HO O

R N

O NH

O HB

R

N N H O

N

O NH

OH O

O

N

N NH

N

쎵

O

O

H2O

Fig. 1.8: Structural representation of the monooxygenation reaction brought about by the BVMO, cyclohexanone monooxygenase. (A) C(4a)-FAD peroxide attacking the electrophilic cyclohexanone, (B) Criegee intermediate that rearranges to form the 7-membered lactone, (C) C(4a)-FAD hydroxide, (D) oxidized FAD.

18

1 The reaction mechanisms of Groups A and B flavoprotein monooxygenases

Villiger reactions used in organic synthesis [60]. In 1976 CHMO from Acinetobacter NCIMB 9871 was the first BVMO to be isolated and partly characterized [61]. The first mechanistic studies were carried out in 1982 [59], the enzyme was cloned and expressed in 1988 [62], and more detailed mechanistic studies followed in 2001 [45]. These studies have formed the basis of our current knowledge about the mechanism until the present day. As elaborated below, there is active research in determining crystal structures and diverse uses of BVMOs. BVMOs are found in a wide variety of bacteria and fungi, but none have yet been found in animals or plants [63]. The BVMOs generally have broad substrate tolerances, yet they often carry out reactions with remarkable regio-selectivity and stereoselectivity. In organisms, they are involved in degradation pathways for cyclic and linear ketones, and in the biosynthesis of steroids, antibiotics, pheromones, and other chiral molecules. Many Baeyer-Villiger reactions are very useful in a wide variety of important synthetic schemes to produce precious compounds, especially for the pharmaceutical industry [63–65]. In addition to carrying out the classical Baeyer-Villiger reaction, BVMOs can also carry out oxygenations of phenyl boronic acids to yield phenols, aldehydes to acids, tertiary amines to N-oxides, and sulfur containing molecules such as thiols (often stereospecifically), sulfides, and even disulfides and dithiolanes to sulfenic acids and sulfones [64–66]. There have even been cases where BVMOs carry out epoxidations of double bonds [65]. Curiously, BVMOs that function with aromatic ketones usually do not function effectively with aliphatic ketones, and vice versa. This great diversity of reactions suggests that BVMOs are basically “vehicles” that provide reactive (hydro) peroxides to carry out both nucleophilic or electrophilic oxygenations. In contrast to the classical Baeyer-Villiger reactions, when BVMOs use sulfur-containing substrates, the C4a-adduct is most likely reacting as a hydroperoxide rather than a peroxide. Because of the great diversity and usefulness of reactions catalyzed by BVMOs, in the past 20 years there has been extensive research into using BVMOs for green biocatalytic purposes; this research has been recently reviewed extensively [63–65]. The classical Baeyer-Villiger reaction requires the use of toxic and unstable reagents such as m-chloroperbenzoic acid and frequently, halogenated solvents. Thereby producing more waste than product. The use of BVMOs avoids such reagents and makes these biocatalytic processes considerably greener, often using NADPH regenerating enzymes to also avoid the need to supply costly reagents. Considerable efforts to improve the specificity, stability, and product yields, using mutagenesis, coupled reaction schemes, hybrid enzymes, incorporation of BVMOs into the metabolic schemes of microorganisms, etc. are currently in progress [63–65]. The recognition of a common sequence motif for the enzymes [67] has encouraged data-mining that has resulted in the identification of a great many BVMOs from a wide variety of species. CHMO from Acinetobacter ATCC 9871 and phenylacetone monooxygenase (PAMO), a thermostable enzyme from Thermobifida fusca, are the BVMOs that have received the most intense mechanistic study [45,59,67–70]. The two enzymes have substantially the same overall mechanisms (򐂰Figs. 1.7 and 1.8), so the following points about CHMO have relevance to PAMO as well. Kinetic studies [45] at 4ºC have shown that NADPH binds tightly and reduces CHMO at 22 s−1 at saturating concentrations of NADPH. This rate is the same between pH 7 and 9. The resulting reduced CHMO-NADP complex reacts with O2 with a second-order rate constant of ≥ 5 x 106 M−1s−1, so fast that

1.3 Enzymes acting upon non-aromatic substrates – Group B

19

accurate measurement was not possible by standard stopped-flow methods. The first intermediate, with an absorbance peak at 366 nm, has been attributed to a C4a-peroxyFAD, as indicated in 򐂰Fig. 1.8, species A. This species reacts rapidly (110 s−1) with cyclohexanone to form the ε−caprolactone product and oxidized flavin, but with NADP still bound to the enzyme. The hydroxyflavin (C in 򐂰Fig. 1.8) is generally not observed because the rate of dehydration to form oxidized FAD prevents its accumulation. Release of NADP at ~2 s−1 is the rate determining step of the catalysis under most conditions [45] and seems to involve two steps, one a conformational change that could be related to the protein dynamics described below. If the species absorbing at 366 nm is incubated at pH 7.2, it converts at ~3 s−1 to a species absorbing at 383 nm, and this species was shown to be essentially unreactive with cyclohexanone. When the apparently protonated species absorbing at 383 nm was shifted back to pH 9 (using double-mixing stopped-flow methods), it reverted to a 366 nm species that was reactive with cyclohexanone. It was concluded that these changes in absorbance and reactivity properties were due to protonation/deprotonation events (with an apparent pKa of 8.4), with the unprotonated species being the flavin peroxide and the protonated species being the hydroperoxide. This notion is consistent with the reactive peroxide being the nucleophile to form the Criegee intermediate in the Baeyer-Villiger reaction. Although there is definitely a protonation event with an apparent pKa of 8.4 that is important for the Baeyer-Villiger reaction, it is not totally clear that this event involves protonation of the peroxide. Preliminary studies show that the 383 nm species, which is not competent for oxygenating cyclohexanone, is also not competent for oxygenating nucleophilic sulfides, contrary to what might be expected if it were a hydroperoxide (unpublished data, E. Rees and DP Ballou). Yet the 366 nm species is competent to oxygenate sulfides. Thus, this protonation event with a pKa of 8.4 might involve a protein conformational change that prevents oxygenation from occurring rather than being due to the conversion of a peroxide to a hydroperoxide. To account for the C4a-intermediate carrying out both nucleophilic and electrophilic reactions from the 366 nm species, we assume that the requisite protonation/deprotonation events are rapid.

1.3.1.2 FMOs The FMOs are reviewed in more detail in Chapter 3 by van Berkel in this volume. The first documented oxygenase of Group B was the FMO from liver microsomes discovered by Ziegler and colleagues [71] (at that time it was called a mixed function oxidase). Early detailed characterizations of FMOs from mammals (pig liver) were carried out largely by the groups of Ziegler and Ballou [72–75]. FMOs use an enzyme-bound electrophilic C4a-FAD hydroperoxide to oxygenate an incredibly broad range of carbon-bound nucleophilic nitrogens, sulfurs, or halides [76,77]. An early study characterized the differences in the mechanisms of oxygenation of sulfides by cytochrome P450s and by FMO, showing that P450 reactions are initiated by a single electron abstraction from the sulfide to the activated heme-oxygen, whereas with FMO the nucleophilic sulfide attacked the electrophilic C4a-flavin hydroperoxide [78]. FMOs can also oxygenate the highly electrophilic phenylboronic and butylboronic acids [75], suggesting that like organic peracids, FMOs (and BVMOs) are vehicles of both peroxides and hydroperoxides. FMOs functions in mammals (largely in conjunction with cytochrome

20

1 The reaction mechanisms of Groups A and B flavoprotein monooxygenases

P450s) are to participate in the metabolism of food products, including trimethylamine, as well as to detoxify xenobiotic compounds such as drugs and other substances, roles to which they are well suited due to their very broad substrate tolerances. Genome sequencing has shown that FMOs are found in all kingdoms of life, with most species having several FMOs that are often specific for given tissues and substrates [76]. Humans have five FMOs. The isozyme FMO3, which is the principle FMO for metabolizing drugs and other xenobiotics, is particularly important for oxygenating trimethylamine; mutations in its gene often lead to trimethylaminuria, the fish-odor syndrome [79–81]. Bacterial, yeast, and plant FMOs have also been isolated. The yeast enzyme is thought to mainly oxygenate thiols to their sulfenic acids, which eliminate H2O to form disulfides. This process could aid in forming disulfide bonds in proteins in the endoplasmic reticulum [82]. The general kinetic mechanism of FMOs is described in 򐂰Fig. 1.7. Curiously, the kcat for most substrates is nearly the same. This is due to the fact that the rate-determining step is usually the release of NADP from the enzyme with the concomitant dehydration of the C4a-flavin hydroxide (E–FADOH in 򐂰Fig. 1.7). In the absence of substrate, the C4a-flavin hydroperoxide species is very stable. At 4ºC and pH 7.2 it takes 2 hours for it to lose H2O2 and form FAD [74]; at higher pH values, the rate of H2O2 loss is faster, and this characteristic is in agreement with findings by Sucharitakul et al. [41] for flavoprotein oxidases. This stability made it possible to form the C4a-hydroperoxide from reduced enzyme and O2, and then react it with a series of substrates [75]. When N,N-dimethylaniline (DMA) was the substrate, the C4a-hydroperoxide peak at 377 nm shifted to the blue to 370 nm at a rate that was second order with respect to DMA concentration. The decay of this intermediate to the spectrum of oxidized flavin was independent of the substrate concentration, but was faster at higher pH, similar to the effects of pH on the decay of the C4a-hydroperoxide in the absence of substrate (the oxidase reaction) [75]. Similar results were observed with other substrates, but with different rates of forming the 370 nm species. With thiobenzamide and iodide, the oxygenation step caused spectral changes that could be followed spectrophotometrically. The rates of oxygenation of these two substrates corresponded to the rates of the blue shift of the C4a-spectrum. Thus, it was concluded that this blue shift was due to the formation of the C4a-FAD hydroxide, analogous to the reactions described above for PHBH. In this same study [75], reduced FMO was also incubated with various oxidized pyridine nucleotides and then reacted with O2. In a few cases, some C4ahydroperoxide intermediate was formed. However, none of the analogues provided the extent of stability bestowed by NADP. NHDP, which replaces the adenine of NADP with hypoxanthine, provided ~55% of the stability of NADP, whereas NAD only gave about 17% of the stability.

1.3.1.3 NMOs NMOs are found in bacteria and fungi and are reviewed in more detail in Chapter 2 of this volume. The NMOs catalyze the hydroxylation of soft nucleophiles, such as the amines in ornithine and lysine, using mechanisms similar to that of FMOs. These enzymes are involved in biosynthetic pathways, especially those for the hydroxamate-based siderophores. The scavenging of iron by bacteria is crucial

1.3 Enzymes acting upon non-aromatic substrates – Group B

21

to pathogenic bacterial virulence, and therefore, enzymes involved in siderophore biosynthesis are potential drug targets. In contrast to the broad substrate tolerance of FMOs and BVMOs, NMOs usually are very specific [83,84]. For example, ornithine monooxygenases are not capable of oxygenating lysine or arginine and lysine monooxygenase cannot oxygenate ornithine. The most thoroughly characterized NMOs are the ornithine monooxygenases from Aspergillus fumigatus (called SidA or Af-OMO) [83,85,86] and from Pseudomonas aeruginosa (PvdA) [87–89]. One of the problems with studying the OMOs is that they do not bind FAD very tightly. Therefore, only a fraction of the enzyme is populated with FAD. Af-OMO binds FAD considerably more tightly than does PvdA [88]. Reduction rates of Af-OMO or of PvdA by NADPH are about the same either in the presence or absence of ornithine, similar to the BVMOs and FMOs described above. However, in contrast to the BVMOs and FMOs, formation of the C4a-FAD hydroperoxy is faster in the presence of ornithine. The rate with Af-OMO is ~15-fold faster with ornithine present [90], while that for PvdA is ~80-fold faster [87]. Because formation of the C4a-hydroperoxide is not the rate-determining step of catalysis, this stimulation does not increase the overall rate of turnover. Thus, its activating function on the reaction with O2 is not clear. In the case of Af-OMO, arginine is an effector for Af-OMO and stimulates oxidase turnover, completely uncoupling the reaction from oxygenation and simply releasing H2O2 [90]. Although lysine is not oxygenated, it is also an effector for Af-OMO, causing NADPH turnover at about the same rate as does ornithine; it does not stimulate the reactions to the same extent as does arginine. The rate of reduction of AfOMO is 1.46 s−1 with ornithine and 20.3 s−1 with arginine present. The rate of reaction with O2 (125 µM) to form its C4a-hydroperoxide is 1.28 s−1 without substrate, 19.8 s−1 with ornithine, and 144 s−1 with arginine present [90]. In the absence of substrate, the C4a-FAD hydroperoxides of the ornithine hydroxylases are quite stable. That for Af-OMO has a half-life at pH 8, 25ºC, of 33 min. Although overall the reaction mechanisms are very similar to those of FMOs, it is clear that in contrast to FMOs, ornithine stimulates the reaction of the reduced FAD with O2.

1.3.1.4 YUCCAs The fourth subclass of the Group B monooxygenases contains the YUCCA enzymes, which are found in all plants and are involved in the biosynthesis of auxin (2-(1H-indol-3yl)acetic acid, IAA), the primary growth hormone and developmental regulator of plants [91,92]. Auxin is synthesized from tryptophan. There are 11 YUCCAs in the Arabidopsis thaliana genome, and they have functions in different developmental stages as well as in different parts of the plant [91,92]. The YUCCA enzymes are related by sequence to the FMOs and, by analogy, were previously thought to catalyze the N-oxygenation of tryptamine; it was postulated that its hydroxylated product was further oxidized to auxin [93]. Recently, however, it has been shown that YUCCA2 (and probably most or all YUCCAs) decarboxylate indole pyruvic acid (IPA) to form auxin (򐂰Fig. 1.9) [94]. The IPA substrate derives from a transamination of tryptophan by a family of amino acid transaminases [94]. The mechanism of YUCCA enzymes likely involves a nucleophilic attack of the C4a-FAD peroxide on the keto function of IPA to form a Criegee intermediate, analogously to the BVMOs. Thus, in contrast to the related FMOs, YUCCAs most likely

22

1 The reaction mechanisms of Groups A and B flavoprotein monooxygenases R

R

N

N

N HO

O NH

N

O

–O NH

O

N

NH

N H O

O O O–

O

R

N

N

N H OH O

O

O NH

쎵 CO2

O –O

– O O

–

NH O

NH

Fig. 1.9: Structural representation of the formation of auxin (indole acetic acid) by YUCCA. The flavin C4a-peroxide attacks the carbonyl of the indole pyruvate to form a Criegee intermediate (brackets). This intermediate decarboxylates, leaving the indole acetic acid and the C4a-flavin hydroxide, which loses H2O to form oxidized flavin.

use a C(4a)-FAD peroxide. Mechanistic work on YUCCAs has been sparse because of difficulties in handling these unstable enzymes. However, newly developed preparations and conditions have permitted rapid kinetics and steady-state kinetics studies in the Zhao, Ballou, and DuBois laboratories demonstrating that a C4a-FAD peroxide forms and reacts with IPA to form auxin, as shown in 򐂰Fig. 1.9 [95].

1.3.2 Structural features All of the enzymes in Group B that have been studied by X-ray crystallography have similar structures. There are published structures for two FMOs: from Schizosaccharomyces pombe [96], and from Methylophaga sp. SK1 [97,98]. However, no structures are available for the mammalian FMOs. There are structures for four BVMOs: PAMO, which was the first available structure for a Group B monooxygenase [68,99], CHMO [69,70], 2-oxo-Δ(3)-4,5,5-trimethylcyclopentenylacetyl-coenzyme A monooxygenase (OTEMO) of Pseudomonas putida ATCC 17453 [100], and mithramycin monooxygenase (MtmOIV) [101]. The last structure is not similar to the other BVMOs, and it falls better into Group A. Little is known about this enzyme, and it will not be discussed. OTEMO is unusual in that it uses FMN rather than FAD, but still has many similar characteristics to other BVMOs. There are also published structures for one NMO: PvdA, the OMO from P. aeruginosa [88,89]. There are many common features to the structures of Group B monooxygenases. They all have two Rossmann folds, one largely at the N-terminus for binding FAD and the other for binding NADP(H), whereas Group A enzymes only have a single Rossmann fold for the FAD. Therefore, the Group B monooxygenase structures are more similar to the flavoprotein disulfide reductases, such as glutathione reductase, than they are to the Group A monooxygenases. This second Rossmann fold in Group B enzymes accounts for their ability to retain the pyridine nucleotide throughout catalysis. The active site of the BVMOs and NMOs is in a cleft at the domain interface of the two Rossmann folds; in contrast, FMOs, as expected, given the wide range of substrates they can oxygenate, do not have a well-defined substrate-binding site. The Baeyer-Villiger sequence motif (FXGXXXHXXXW(P/D)), which is also nearly the same in the FMOs [67], is found to be quite distant from the active site, but part of a surface loop that connects the two

1.4 References

23

Rossmann fold domains. This linker is important for various protein dynamics required for efficient catalysis (see below). The FAD binding domains are all α/β folds like those of the glutathione oxidoreductase family. In general, the FADs are buried within the protein, although with PvdA, the FAD is more exposed, which probably accounts for its weak binding in this NMO. The BVMOs have a conserved arginine (Arg337 in PAMO) that likely is important for stabilizing the Criegee intermediate during catalysis, most likely by interacting with the carbonyl to be attacked by the C4a-peroxide [68,69]. Common to most of the Group B enzyme structures is the finding that more than one conformation exists. This undoubtedly reflects the plasticity and the protein dynamics that are important for catalysis and crucial to accommodating such a wide variety of substrates, both large and small. For example, PAMO exists in “out” and “in” forms in which the nicotinamide takes on two conformations [68], and CHMO has analogous “open” and “closed” conformations with NADP binding in two distinct conformations [70]. In one conformation of PAMO, the nicotinamide, before substrate binds, helps O2 to approach the front of the flavin to form the C4a-adduct, and it forms important H-bonds with the N5-hydrogen and the C4carbonyl of the flavin to stabilize the C4a-peroxide adduct. In CHMO it was shown that “sliding” of the nicotinamide away from the flavin opened up a well-defined substratebinding pocket [69]. This helps to explain the kinetic mechanism in which substrate binds after forming the C4a-peroxide [45,59]. Recently, it was shown that OTEMO also exhibits similar plasticity in its structure [100]. Yachnin et al. [69] have shown in CHMO how a major rotation of the nicotinamide away from the isoalloxazine ring permits the substrate to be positioned correctly to react with the C4a-flavin peroxide to form the Criegee intermediate. These motions are facilitated by the linking loop containing the signature motif (above); mutations in this loop impair activity. The structural plasticity of BVMOs and other Group B monooxygenases helps to explain how such complicated monooxygenation reactions can be catalyzed by a single enzyme. The movement of the nicotinamide of NADP is reminiscent of the movement of the isoalloxazine in PHBH that was described above. Different conformations help develop different characteristics to promote the diverse segments of the catalysis. It was postulated that CHMO, for example, employs a series of conformations that allow reduction of the flavin, formation and stabilization of the C4a-FAD peroxide by the NADP, binding of substrate, formation of the Criegee intermediate, and stereochemical selection of a given Baeyer-Villiger rearrangement [69]. A picture emerges that extensive conformational dynamics are important for many of these Group B enzymes. Future studies will undoubtedly show in greater detail how the proteins enable such a complex catalytic dance.

1.4 References [1] van Berkel WJH, Kamerbeek NM, Fraaije MW. Flavoprotein monooxygenases, a diverse class of oxidative biocatalysts. J Biotechnol 2006;124:670–89. [2] Enzyme nomenclature database. (Accessed March 20, 2012, at http://enzyme.expasy.org) [3] Entsch B, Cole LJ, Ballou DP. Protein dynamics and electrostatics in the function of p-hydroxybenzoate hydroxylase. Arch Biochem Biophys 2005;433:297–311.

24

1 The reaction mechanisms of Groups A and B flavoprotein monooxygenases

[4] Palfey BA, McDonald CA. Control of catalysis in flavin-dependent monooxygenases. Arch Biochem Biophys 2010;493:26–36. [5] Ornston LN, Parke D. The evolution of induction mechanisms in bacteria: insights derived from the study of the beta-ketoadipate pathway. Curr Top Cell Regul 1977;12:209–62. [6] Grenhagen BT, Shi K, Robinson H, et al. Crystal structure of the pyocyanin biosynthetic protein PhzS. Biochemistry 2008;47:5281–9. [7] Zucchini D, Caprini G, Pasterkamp RJ, Tedeschi G, Vanoni MA. Kinetic and spectroscopic characterization of the putative monooxygenase domain of human MICAL-1. Arch Biochem Biophys 2011;515:1–13. [8] Protein Data Bank. (Accessed March 20, 2012, at http://www.rcsb.org/pdb/home/home.do) [9] Structural Classification of Proteins. (Accessed March 20, 2012, at http://scop.mrc-lmb.cam. ac.uk/scop/) [10] Enroth C, Neujahr H, Schneider G, Lindqvist Y. The crystal structure of phenol hydroxylase in complex with FAD and phenol. Structure 1998;6:605–17. [11] Hiromoto T, Fujiwara S, Hoshkawa K, Yamaguchi H. Crystal structure of 3-hydroxybenzoate hydroxylase from Comamonas testosteroni has a large tunnel for substrate and oxygen access to the active site. J Mol Biol 2006;364:878–96. [12] Wierenga RK, de Jong RJ, Kalk KH, Hol WGJ, Drenth J. Crystal structure of p-hydroxybenzoate hydroxylase. J Mol Biol 1979;131:55–73. [13] Schreuder HA, Prick PAJ, Wierenga RK, et al. Crystal structure of the p-hydroxybenzoate hydroxylase-substrate complex refined to 1.9 Å resolution. J Mol Biol 1989;208:679–96. [14] Entsch B, Yang N, Weaich K, Scott KF. Sequence and organization of pobA, the gene coding for p-hydroxybenzoate hydroxylase, an inducible enzyme from Pseudomonas aeruginosa. Gene 1988;71:279–91. [15] Entsch B, Ballou DP, Husain M, Massey V. Catalytic mechanism of p-hydroxybenzoate hydroxylase with p-mercaptobenzoate as substrate. J Biol Chem 1976;251:7367–79. [16] Husain M, Massey V. Kinetic studies on the reaction of p-hydroxybenzoate hydroxylase; agreement of steady state and rapid reaction data. J Biol Chem 1979;254:6657–66. [17] Spector T, Massey V. p-Hydroxybenzoate hydroxylase from Pseudomonas fluorescens: evidence for an oxygenated flavin intermediate. J Biol Chem 1972;247:5632–6. [18] Entsch B, Ballou DP, Massey V. Flavin-oxygen derivatives involved in hydroxylation by p-hydroxybenzoate hydroxylase. J Biol Chem 1976;251:2550–63. [19] Entsch B, Palfey BA, Ballou DP, Massey V. Catalytic function of tyrosine residues in parahydroxybenzoate hydroxylase as determined by the study of site-directed mutants. J Biol Chem 1991;266:17341–9. [20] Gatti DL, Palfey BA, Lah MS, et al. The mobile flavin of 4-OH benzoate hydroxylase. Science 1994;266:110–4. [21] Schreuder HA, Mattevi A, Obmolova G, et al. Crystal structures of wild-type p-hydroxybenzoate hydroxylase complexed with 2,4-dihydroxybenzoate. Evidence for a proton channel and a new binding mode of the flavin ring. Biochemistry 1994;33:10161–70. [22] Eppink MH, Schreuder HA, van Berkel WJ. Interdomain binding of NADPH in p-hydroxybenzoate hydroxylase as suggested by kinetic, crystallographic and modeling studies of histidine 162 and arginine 269 variants. J Biol Chem 1998;273:21031–9. [23] Eppink MH, Overkamp KM, Schreuder HA, van Berkel WJ. Switch of coenzyme specificity of p-hydroxybenzoate hydroxylase. J Mol Biol 1999;292:87–96. [24] Wang J, Ortiz-Maldonado M, Entsch B, Massey V, Ballou D, Gatti DL. Protein and ligand dynamics in 4-hydroxybenzoate hydroxylase. Proc Nat Acad Sci USA 2002;99:608–13. [25] Palfey BA, Moran G, Entsch B, Ballou DP, Massey V. Substrate recognition by “password” in p-hydroxybenzoate hydroxylase. Biochemistry 1999;38:1153–8. [26] Lah MS, Palfey BA, Schreuder HA, Ludwig ML. Crystal structures of mutant Pseudomonas aeruginosa p-hydroxybenzoate hydroxylases: the Tyr201Phe, Tyr385Phe, and Asn300Asp variants. Biochemistry 1994;33:1555–64.

1.4 References

25

[27] van Berkel WJH, Eppink MHM, Schreuder HA. Crystal structure of p-hydroxybenzoate hydroxylase reconstituted with the modified FAD present in alcohol oxidase from methylotrophic yeasts: evidence for an arabinoflavin. Protein Science 1994;3:2245–53. [28] Palfey BA, Ballou DP, Massey V. Flavin conformational changes in the catalytic cycle of p-hydroxybenzoate hydroxylase substituted with 6-azido- and 6-aminoflavin adenine dinucleotide. Biochemistry 1997;36:15713–23. [29] Frederick KK, Palfey BA. Kinetics of proton-linked flavin conformational changes in p-hydroxybenzoate hydroxylase. Biochemistry 2005;44:13304–14. [30] Manstein DJ, Pai EF, Schopfer LM, Massey V. Absolute stereochemistry of flavins in enzymecatalyzed reactions. Biochemistry 1986;25:6807–16. [31] Moran GR, Entsch B, Palfey BA, Ballou DP. Electrostatic effects on substrate activation in parahydroxybenzoate hydroxylase: studies of the mutant lysine 297 methionine. Biochemistry 1997;36:7548–56. [32] Schreuder HA, van der Laan JM, Marjolein, MGM, et al. Crystallographic Studies on the Mechanism of p-Hydroxybenzoate Hydroxylase from Pseudomonas fluorescens. In: McCormick DB, Edmondson DE, eds. Flavins and Flavoproteins 1987. Berlin: Walter de Gruyter, 1988:527–38. [33] Massey V. Activation of molecular oxygen by flavins and flavoproteins. J Biol Chem 1994;269:22459–62. [34] Kemal C, Chan TW, Bruice TC. Reaction of 3O2 with dihydroflavins. J Amer Chem Soc 1977;99:7272–86. [35] Schopfer LM, Massey V. Kinetic and mechanistic studies on the oxidation of the melilotate hydroxylase. 2-OH-cinnamate complex by molecular oxygen. J Biol Chem 1980;255:5355–63. [36] Detmer K, Massey V. Effect of monovalent anions on the mechanism of phenol hydroxylase. J Biol Chem 1984;259:11265–72. [37] Entsch B, Ballou DP. Purification, properties, and oxygen reactivity of p-hydroxybenzoate hydroxylase from Pseudomonas aeruginosa. Biochim Biophys Acta 1989;999:313–22. [38] Hastings JW, Balny C, le Peuch C, Douzou P. Spectral properties of an oxygenated luciferaseflavin intermediate isolated by low-temperature chromatography. Proc Nat Acad Sci USA 1973;70:3468–72. [39] Ghisla S, Entsch B, Massey V, Husain M. On the structure of flavin-oxygen intermediates involved in enzymatic reactions. Eur J Biochem 1977;76:139–48. [40] Vervoort J, Muller F, Lee J, van den Berg WAM, Moonen CTW. Identifications of the true carbon-13 NMR spectrum of the stable intermediate II in bacterial luciferase. Biochemistry 1986;25:8062–7. [41] Sucharitakul J, Wongnate T, Chaiyen P. Hydrogen peroxide elimination from C4ahydroperoxyflavin in a flavoprotein oxidase occurs through a single proton transfer from flavin N5 to a peroxide leaving group. J Biol Chem 2011;286:16900–9. [42] Gadda G. Oxygen activation in flavoprotein oxidases: the importance of being positive. Biochemistry 2012;51:2662–9. [43] Ortiz-Maldonado M, Entsch B, Ballou DP. Oxygen reactions in p-hydroxybenzoate hydroxylase utilize the H-bond network during catalysis. Biochemistry 2004;43:15246–57. [44] Palfey BA, Ballou DP, Massey V. Solvent isotope effects on p-hydroxybenzoate hydroxylase. In: Yagi K., ed. Flavins and Flavoproteins 1993. Berlin: Walter de Gruyter, 1993:235–8. [45] Sheng D, Ballou DP, Massey V. Mechanistic studies of cyclohexanone monooxygenase: chemical properties of intermediates involved in catalysis. Biochemistry 2001;40:11156–67. [46] Gatti DL, Entsch B, Ballou DP, Ludwig ML. pH-Dependent structural changes in the active site of p-hydroxybenzoate hydroxylase point to the importance of proton and water movements during catalysis. Biochemistry 1996;35:567–78. [47] Schopfer LM, Wessiak A, Massey V. Interpretation of the spectra observed during oxidation of p-hydroxybenzoate hydroxylase reconstituted with modified flavins. J Biol Chem 1991;266:13080–5.

26

1 The reaction mechanisms of Groups A and B flavoprotein monooxygenases

[48] Maeda-Yorita K, Massey V. On the reaction mechanism of phenol hydroxylase. J Biol Chem 1993;268:4134–44. [49] Ortiz-Maldonado M, Cole LJ, Dumas SM, Entsch B, Ballou DP. Increased positive electrostatic potential in p-hydroxybenzoate hydroxylase accelerates hydroxylation but slows turnover. Biochemistry 2004;43:1569–79. [50] Palfey BA, Basu R, Frederick KK, Entsch B, Ballou DP. Role of protein flexibility in the catalytic cycle of p-hydroxybenzoate hydroxylase elucidated by the Pro293Ser mutant. Biochemistry 2002;41:8438–46. [51] Schreuder HA, Drenth J. Crystal structure of p-hydroxybenzoate hydroxylase complexed with its reaction product 3,4-dihydroxbenzoate. J Mol Biol 1988;199:637–48. [52] Brender JR, Dertouzos J, Ballou DP, et al. Conformational dynamics of the isoalloxazine in substrate-free p-hydroxybenzoate hydroxylase: single-molecule studies. J Am Chem Soc 2005;127:18171–8. [53] Husain M, Entsch B, Ballou DP, Massey V, Chapman PJ. Fluoride elimination from substrates in hydroxylation reactions catalyzed by p-hydroxybenzoate hydroxylase. J Biol Chem 1980;255:4189–97. [54] Ortiz-Maldonado M, Ballou DP, Massey V. Use of free energy relationships to probe the individual steps of hydroxylation of p-hydroxybenzoate hydroxylase: studies with a series of 8-substituted flavins. Biochemistry 1999;38:8124–37. [55] Xu D, Ballou DP, Massey V. Studies of the mechanism of phenol hydroxylase: Tyr289Phe, Asp54Asn, and Arg281Met. Biochemistry 2001;40:12369–78. [56] Entsch B, Husain M, Ballou DP, Massey V, Walsh C. Oxygen reactivity of p-hydroxybenzoate hydroxylase containing 1-deaza-FAD. J Biol Chem 1980;255:1420–9. [57] Cole LJ, Gatti DL, Entsch B, Ballou DP. Removal of a methyl group causes global changes in p-hydroxybenzoate hydroxylase. Biochemistry 2005;44:8047–58. [58] Cole LJ, Entsch B, Ortiz-Maldonado M, Ballou DP. Properties of p-hydroxybenzoate hydroxylase when stabilized in its open conformation. Biochemistry 2005;44:14807–17. [59] Ryerson CC, Ballou DP, Walsh CT. Mechanistic studies on cyclohexanone monooxygenase. Biochemistry 1982;21:2644–55. [60] Renz M, Meunier B. 100 Years of Baeyer-Villiger oxidations. Eur J Org Chem 1999;4:737–50. [61] Donoghue NA, Norris DB, Trudgill OW. The purification and properties of cyclohexanone monooxygenase from Nocardia globerula CL1 and Acinetobacter NCIB 9871. Eur J Biochem 1976;63:175–92. [62] Chen YC, Peoples OP, Walsh CT. Acinetobacter cyclohexanone monooxygenase: gene cloning and sequence determination. J Bacteriol 1988;170:781–9. [63] Torres Pazmino DE, Dudek HM, Fraaije MW. Baeyer-Villiger monooxygenases: recent advances and future challenges. Cur Opin Chem Biol 2010;14:138–49. [64] Leisch H, Morley K, Lau PCK. Baeyer-Villiger monooxygenases: more than just green chemistry. Chem Rev 2011;222:4165–222. [65] de Gonzalo G, Mihovilovic MD, Fraaije MW. Recent developments in the application of Baeyer-Villiger monooxygenases as biocatalysts. ChemBioChem 2010;11:2208–31. [66] Branchaud BP, Walsh CT. Functional group diversity in enzymic oxygenation reactions catalyzed by bacterial flavin-containing cyclohexanone oxygenase. J Am Chem Soc 1985;107:2153–61. [67] Fraaije MW, Kamerbeek NM, van Berkel WJH, Janssen DB. Identification of a Baeyer-Villiger monooxygenase sequence motif. FEBS Lett 2002;518:43–7. [68] Orru R, Dudek HM, Martinoli C, et al. Snapshots of enzymatic Baeyer-Villiger catalysis: oxygen activation and intermediate stabilization. J Biol Chem 2011;286:29284–91. [69] Yachnin BJ, Sprules T, McEvoy MB, Lau PC, Berghuis AM. The substrate-bound crystal structure of a Baeyer-Villiger monooxygenase exhibits a Criegee-like conformation. J Am Chem Soc 2012;134:7788–95.

1.4 References

27

[70] Mirza IA, Yachnin BJ, Wang S, et al. Crystal structures of cyclohexanone monooxygenase reveal complex domain movements and a sliding cofactor. J Am Chem Soc 2009;131: 8848–54. [71] Ziegler DM, Jollow D, Cook DE. Properties of a purified liver microsomal mixed function amine oxidase. In: Kamin, H., ed. Flavins and Flavoproteins 1969. Baltimore: University Park Press 1971:507–522. [72] Poulsen LL, Ziegler DM. The liver microsomal FAD-containing monooxygenase. Spectral characterization and kinetic studies. J Biol Chem 1979; 254:6449–55. [73] Beaty NB, Ballou DP. The reductive half-reaction of liver microsomal FAD-containing monooxygenase. J Biol Chem 1981;256:4611–8. [74] Beaty NB, Ballou DP. The oxidative half-reaction of liver microsomal FAD-containing monooxygenase. J Biol Chem 1981;256:4619–25. [75] Jones KC, Ballou DP. Reactions of the 4a-hydroperoxide of liver microsomal flavincontaining monooxygenase with nucleophilic and electrophilic substrates. J Biol Chem 1986;261:2553–9. [76] Ziegler DM. Flavin-containing monooxygenases: Catalytic mechanism and structure of FMOs. Drug Met Rev 1988;19:1–32. [77] Krueger SK, Williams DE. Mammalian flavin-containing monooxygenases: structure/ function, genetic polymorphisms and role in drug metabolism. Pharmacol Ther 2005;106: 357–387. [78] Oae S, Mikami, A Matsuura T, et al. Comparison of sulfide oxygenation mechanism for liver microsomal FAD-containing monooxygenase with that for cytochrome P-450. Biochem Biophys Res Commun 1985;131:567–73. [79] Murphy HC, Dolphin CT, Janmohamed A, et al. A novel mutation in the flavin-containing monooxygenase 3 gene, FMO3, that causes fish-odor syndrome: activity of the mutant enzyme assessed by proton NMR spectroscopy. Pharmacogenetics 2000;10:439–451. [80] Zhang J, Tran Q, Lattard V, Cashman JR. Deleterious mutations in the flavin-containing monooxygenase 3 (FMO3) gene causing trimethylaminuria. Pharmacogenetics 2003;13: 495–500. [81] Cashman JR, Zhang J. Human flavin-containing monooxygenases. Ann Rev Pharmacol Toxicol 2006;46:65–100. [82] Suh JK, Poulsen LL, Ziegler DM, Robertus JD. Molecular cloning and kinetic characterization of a flavin-containing monooxygenase from Saccharomyces cerevisiae. Arch Biochem Biophys 1996;336:268–74. [83] Chocklett SW, Sobrado P. Aspergillus fumigants SidA is a highly specific ornithine hydroxylase with bound flavin cofactor. Biochemistry 2010;49:6777–83. [84] Dick S, Marrone L, Duewel H, Beecroft M, McCourt J, Viswanatha T. Lysine N6-Hydroxylase: Stability and interaction with Ligands. J Protein Chem 1999;18:893–903. [85] Romero E, Fedkenheuer M, Chocklett SW, Qi J, Oppenheimer M, Sobrado P. Dual role of NADP(H) in the reaction of a flavin dependent N-hydroxylating monooxygenase. Biochim Biophys Acta 2012;1824:850–7. [86] Mayfield JA, Frederick RE, Streit B, Wencewicz TA, Ballou DP, DuBois JL. Comprehensive spectroscopic, steady state, and transient kinetic studies of a representative siderophoreassociated flavin monooxygenase. J Biol Chem 2010;285:30375–88. [87] Meneely KM, Barr EW, Bollinger JM, Jr, Lamb AL. Kinetic mechanism of ornithine hydroxylase (PvdA) from Pseudomonas aeruginosa: substrate triggering of O2 addition but not flavin reduction. Biochemistry 2009;48:4371–6. [88] Olucha J, Lamb AL. Mechanistic and structural studies of the N-hydroxylating flavoprotein monooxygenases. Bioorg Chem 2011;39:171–7. [89] Olucha J, Meneely KM, Chilton AS, Lamb AL. Two structures of an N-hydroxylating flavoprotein monoxygenase. J Biol Chem 2011;286:31789–98.

28

1 The reaction mechanisms of Groups A and B flavoprotein monooxygenases

[90] Frederick RE, Mayfield JS, DuBois JL. Regulated O2 activation in flavin-dependent monooxygenases. J Am Chem Soc 2011;133:12338–41. [91] Zhao Y. Auxin biosynthesis and its role in plant development. Ann Rev Plant Biol 2010;61:49–64. [92] Schlaich NL. Flavin-containing monooxygenases in plants: looking beyond detox. Trend Plant Sci 2007;12:412–8. [93] Zhao Y, Christensen S, Franhauser C, et al. A role for flavin monooxygenase-like enzymes in auxin biosynthesis. Science 2008;291:306–09. [94] Zhao Y. Auxin biosynthesis: a simple two-step pathway converts tryptophan to indole-3acetic acid in plants. Molecular Plant 2012;5:334–8. [95] Dai X, Mashigucchi K, Chen Q, et al. The biochemical mechanism of auxin biosynthesis by an Arabidopsis YUCCA flavin-containing monooxygenase. J Biol Chem 2012;288:1448–57. [96] Eswaramoorthy S, Bonanno JB, Burley SK, Swaminathan S. Mechanism of action of a flavincontaining monooxygenase. PNAS USA 2006;103:9832–7. [97] Alfieri A, Malito E, Orru R, Fraaije MW, Mattevi A. Revealing the moonlighting role of NADP in the structure of a flavin-containing monooxygenase. PNAS USA 2008;105:6572–7. [98] Cho HJ, Cho HY, Kim KJ, Kim MH, Kim SW, Kang BS. Structural and functional analysis of bacterial flavin-containing monooxygenase reveals its ping-pong mechanism. J Struct Biol 2011;175:39–48. [99] Malito E, Fraaije MW, Mattevi A. Crystal structure of a Baeyer-Villiger monooxygenase. PNAS USA 2004;101:13157–62. [100] Leisch H, Shi R, Grosse S, et al. Cloning, Baeyer-Villiger biooxidations, and structures of the camphor pathway 2-oxo-Δ(3)-4,5,5-trimethylcyclopentenylacetyl-coenzyme A monooxygenase of Pseudomonas putida ATCC 17453. Appl Environ Microbiol 2012;78:2200–12. [101] Beam MP, Bosserman MA, Noinaj N, Wehenkel M, Rohr J. Crystal structure of BaeyerVilliger Monooxygenase MtmOIV, the key enzyme of the mithramycin biosynthetic pathway. Biochemistry 2009;48:4476–87.

2 Flavin-dependent monooxygenases in siderophore biosynthesis Reeder M. Robinson and Pablo Sobrado

Abstract Microbial N-hydroxylating monooxygenases (NMOs) are a class of flavin-dependent enzymes involved in the biosynthesis of hydroxamate-containing siderophores. These flavoenzymes catalyze the NADPH- and oxygen-dependent N-hydroxylation of a narrow group of substrates that include L-ornithine, L-lysine, and primary aliphatic diamines. Siderophores are essential for microbial pathogenesis and the activities of NMOs are essential for their biosynthesis and function. To date, only a small number of NMOs have been characterized. While the biochemical properties of NMOs vary, two common characteristics have been observed. These include a high degree of substrate specificity and the stabilization of a long-lived C4a-hydroperoxyflavin intermediate. These properties ensure efficient hydroxylation of the appropriate substrate and minimal release of hydrogen peroxide, respectively. Recent mechanistic and structural studies have provided insight into the molecular mechanisms of these two traits. The high substrate specificity has been attributed to a unique “molecular ruler mechanism,” and stabilization of the C4a-hydroperoxyflavin is achieved by specific interactions with NADP+. In addition, a high-throughput screening assay that allows for identification of NMO inhibitors from small molecule libraries has been developed. This, compounded with rational drug design, holds promise for the identification of inhibitors to combat increasingly virulent microbes.

2.1 Iron, an essential but scarce nutrient Iron is an essential element required by most living organisms [1]. The need for this metal originates from its requirement as a cofactor in a wide range of biological reactions. The functional diversity of iron results from its inherent nature as a reductant or an oxidant, depending on whether it is present in the ferrous (FeII) or ferric state (FeIII), respectively. Furthermore, while these two oxidation states dominate in biological systems, when iron is bound to certain proteins the reaction between ferrous iron and molecular oxygen can lead to the formation of high-valent iron-oxo species, which catalyze a number of oxygenation reactions. One important role of iron is demonstrated by the oxygen transport protein hemoglobin [2]. Here, iron is complexed with a porphyrin ring, which allows oxygen to be transported to different tissues of the body for use in many oxygen dependent reactions. In plants, iron plays an essential role in photosynthesis. Chloroplast ferredoxin is an iron-containing protein that acts as a carrier of electrons that are produced from

30

2 Flavin-dependent monooxygenases in siderophore biosynthesis

absorbed sunlight [3]. In microorganisms, perhaps the most important iron-containing reaction is observed with nitrogenase, which catalyzes the reduction of dinitrogen to ammonia [4]. The resulting “fixed nitrogen” can be utilized in the nitrogen cycle where it is used for the biosynthesis of amino and nucleic acids, as well as other nitrogencontaining biomolecules. In vivo, free iron is toxic [5]. The reaction of iron in its ferrous or ferric form with hydrogen peroxide, termed the Fenton reaction, can lead to the generation of reactive oxygen species such as superoxide and hydroxyl radicals [6]. Effects of these oxidative by-products include lipid peroxidation, protein denaturation, and DNA strand breaks, all of which are deleterious to the cell [7]. Mammals have evolved several mechanisms to lower the concentration of free iron to minimize the level of toxic oxidative by-products. Iron-binding proteins such as transferrin, lactoferrin, and ferritin store excess iron and make it available during periods of iron limitation [8,9]. Iron sequestration in mammals also has antimicrobial activity as it limits the availability of this essential nutrient in serum [10]. This, compounded with the fact that free ferric iron forms insoluble iron-hydroxide complexes (10−18 M at pH 7), represents a fundamental problem of iron deficiency that pathogenic microbes must overcome [11]. In response, pathogens have evolved mechanisms to scavenge iron from mammalian hosts by synthesizing and secreting low-molecular weight iron chelators termed siderophores (Greek for “iron carrier”) [12].

2.2 Siderophores Siderophores provide a unique mechanism for pathogens to acquire iron where they can utilize it for their own metabolic needs in a non-toxic fashion. Siderophores competitively acquire iron from the bacterial host due to their remarkably high affinity for ferric iron. For example, the catechol siderophore enterobactin produced by the bacterium Escherichia coli possesses a Kd value for Fe3+on the order of 10−49 M [13]. This high affinity allows pathogens to proliferate by scavenging iron from hosts. Siderophores vary in their structures and denticity for iron, ranging from bi- to hexadentate [14]. Siderophores can contain a number of functional groups that chelate ferric iron, including catechols, phenols, hydroxamates, and carboxylates [15]. Examples of siderophores from different organisms are shown in 򐂰Fig. 2.1, and a list of siderophores along with their respective affinity constants for FeIII is given in 򐂰Tab. 2.1.

2.2.1 Siderophores are important virulence factors Siderophores have been shown to be linked to virulence in many human pathogens. For instance, deletion of the gene cluster involved in the biosynthesis of the siderophore anthrachelin in Bacillus anthracis resulted in both attenuated growth and virulence in macrophages and mice [25]. A similar effect was demonstrated in Pseudomonas aeruginosa where pyoverdin- and pyochelin-deficient mutants grew poorly in immunosuppressed mice. Lethality in mice was reduced from 100% in wild-type to 0% in the double mutant 48 hours post inoculation [26]. Effects on virulence were also observed in Aspergillus fumigatus and Burkholderia cepacia upon deletion of genes involved in siderophore biosynthesis [27–29]. In A. fumigatus, disruption of the biosynthesis of the

2.2 Siderophores

31

O

A OH

OH

N

H N

S

OH

S

O

B HO

N

S

N

O OH

OH O

HO

N H

O

C HO O

O

HN

N

NH O

R

N

O N

O

N H

R 쏁 (CH2)6 CH3

OH

OH HO

N

O D

NH2

HO O

OH

N

O

O NH

HO N O

O

R O R 쏁 C17–20

O NH

OH N

HN

O

O

E

O

O

N

N H

HN

O

OH OH

O

NH

HN

H N

F

OH N

O

O

O

O H

O

O

HN N

H N

O

OH O

O

OH O

O

OH

O O O O

G

HO

OH

O

N OH

OH N H

O

OH OH

NH

HO HO

(Figure continued)

32

2 Flavin-dependent monooxygenases in siderophore biosynthesis

(Figure continued)

O

I HO O

H O

NH OH

NH2

HO

HO

NH O N

HO

O

N

H

O

NH HO N

O

HO

N HN HN OH O

N H

H N

H2N

N H

N

O

N

OH

O

O

O

NH

H

H N

O OH

O

O

Fig. 2.1: The structural diversity of siderophores. (A) Yersiniabactin from Yersenia pestis. (B) Rhizobactin 1021 from Rhizobium meliloti. (C) Desferrioxamine G1 from Streptomyces spp. (D) Mycobactin T from Mycobacterium tuberculosis. (E) Ferricrocin from Aspergillus fumigatus. (F) Enterobactin from Escherichia coli. (G) Mugineic acid from Hordium vulgare. (H) Pyoverdin from Pseudomonas aeruginosa. (I) Alcaligin from Bordetella pertussis. Tab. 2.1: Siderophores and their respective affinities for FeIII. Organism

Siderophore

Alteromonas luteoviolacea Alterobactin Escherichia coli Yersinia pestis Hordium vulgare Streptomyces griseus Bordetella pertussis Pseudomonas syringae Neurospora crassa Rhizopus microsporus Azotobacter vinelandii

Enterobactin Yersiniabactin Mugineic acid Desferrioxamine E Alcaligin Pyoverdin Ferrichrome Rhizoferrin Azotobactin

Type Carboxylate/Catechol Catechol Mixed Carboxylate Hydroxamate Hydroxamate Catechol/Hydroxamate Hydroxamate Carboxylate Catechol/Hydroxamate

KFeIII (M) 10

Ref.

–53

[15]

–49

[13]

–36

[16]

–33

[17]

–32

[18]

–32

[19]

–32

[20]

–29

[21]

–25

[22]

–22

[23]

10 10

10 10 10 10 10

10 10

siderophores triacetylfusarinine C and ferricrocin resulted in a mutant strain with diminished growth in low-iron media. In mice, however, the mutant fungus was unable to establish infection, indicating the necessity for siderophores in virulence. Similarly, by targeting the biosynthesis of mycobactin through gene disruption, the rate of growth of Mycobacterium tuberculosis was significantly decreased in iron-deficient media as well as macrophage like THP-1 cells in one study, whereas in a similar study, a mycobactindeficient strain failed to grow in low-iron media [30,31].

2.3 Flavin-dependent N-hydroxylating monooxygenases

33

2.2.2 Structural diversity of siderophores Siderophores are structurally diverse molecules that can contain one, or many, functional groups important for their iron-chelating properties. Amino acids are the precursors for the functional groups and are incorporated to form mature siderophores by non-ribosomal peptide synthetases [14]. These synthetases link amino acids through thioester intermediates, and have different amino acid substrate specificities across different organisms [32,33]. The amino acid selectivity of the synthetases contributes to siderophore diversity (򐂰Fig. 2.1). Siderophores such as alcaligin from Bordetella pertussis and ferricrocin from A. fumigatus are predominately composed of a heterocyclic ring with hydroxamates that function as the chelating groups. On the other hand, siderophores such as mycobactin T from M. tuberculosis, and yersiniabactin from Yersinia pestis, contain a number of functional groups that include oxazoline/thiozoline rings, phenols, hydroxamates, and carboxylates. Because siderophores are important for virulence, the enzymes involved in siderophore biosynthesis are considered potential drug targets. In particular, microbial N-hydroxylating monooxygenases (NMOs) have emerged as ideal systems for drug development due to their central role in the biosynthesis of siderophores [26–31].

2.3 Flavin-dependent N-hydroxylating monooxygenases NMOs are flavin-dependent monooxygenases involved in the biosynthesis of hydroxamate-containing siderophores. These enzymes target the soft nucleophilic terminal amine groups of L-ornithine, L-lysine, and the primary aliphatic diamines 1, 3-diaminopropane, putrescine, and cadaverine (򐂰Fig. 2.2) [28,34–42]. NMOs are members of the Class B flavoprotein monooxygenase family, which include Baeyer-Villiger monooxygenases (BVMOs) and the bacterial and microsomal flavin-containing monooxygenases (bFMO and mFMO, respectively) [43]. Enzymes in this family are encoded by a single gene, bind flavin adenine dinucleotide (FAD), utilize NADPH as a coenzyme, keep NADP+ bound throughout the catalytic cycle, and contain two dinucleotide binding domains. Many members of the NMO family have been identified through the characterization of gene clusters involved in the biosynthesis of several hydroxamate-containing siderophores. IucD from Escherichia coli was the first NMO to be characterized. The function of IucD was determined by gene knockout studies and direct detection of hydroxylated L-lysine [44]. Similar studies to identify related NMOs were also performed in B. cepacia, M. tuberculosis, Omphalotus olearius, Ustilago maydis, Pseudomonas aeruginosa, and Aspergillus spp., among others [28,34,36,37,45,46]. Several NMOs have been specifically linked to virulence through gene knockout studies. Upon deletion of the L-ornithine monooxygenase gene, sidA, in A. fumigatus, the mutant strain was unable to produce triacetylfusarinine C and ferricrocin, and was unable to cause infection in mice [27]. Similar results were obtained when the L-ornithine monooxygenase gene, pvdA, from B. cepacia was deleted [28]. Here, the mutant bacteria were unable to establish infection in rats. These results show the function of NMOs as essential for virulence in both bacteria and fungi. Thus, NMOs clearly represent potential drug targets.

34

2 Flavin-dependent monooxygenases in siderophore biosynthesis A

NADPH O2

O 앥

O

NH2

NADP쎵 H2O

O 앥

O

Af SidA

쎵

NAD쎵 H2O

NADH O2

O

NH2

앥

O

O O

MbsG

NH3 N 6-hydroxy-L-lysine NADPH O2 NH2

NADP쎵 H2O

RhbE

1,3-diaminopropane D

NADPH O2 H2N

NH2

OH N H N-hydroxy-1,3-diaminopropane

H2N AlcA

NADPH O2 H2N

NH2 Cadaverine

H2N

NADP쎵 H2O

Putrescine E

OH

쎵

NH3 L-lysine

H2N

H N

앥

쎵

C

OH

쎵 NH3 N 5-hydroxy-L-ornithine

NH3 L-ornithine

B

N H

OH N H N-hydroxy-putrescine

NADP쎵 H2O

DesB

OH N H N-hydroxy-cadaverine

H2N

Fig. 2.2: Reactions catalyzed by NMOs. The oxygen atoms involved in the reaction are shown in red. The amino acids L-ornithine and L-lysine are N-hydroxylated by Af SidA (A) and MbsG (B), respectively. The N-hydroxylation of the aliphatic diamines 1, 3-diaminopropane, putrescine, and cadaverine are catalyzed by RhbE (C) AlcA (D), and DesB (E), respectively.

Despite the NMOs identified to date, relatively few have been characterized in detail. NMOs that have been expressed, purified, and characterized include the L-ornithine monooxygenases, P. aeruginosa pyoverdin A (Pa PvdA), A. fumigatus siderophore A (Af SidA), Streptomyces coelicolor coelichelin B (CchB), Rhizobium leguminosarum vicibactin synthase O (VbsO), and the L-lysine monooxygenases from Escherichia coli iron uptake chelate D (IucD), and Mycobacterium smegmatis G (MbsG) [35,39,47–50].

2.4 Catalytic cycle of NMOs Af SidA and Pa PvdA are the two most extensively characterized NMOs to date, and their kinetic mechanisms have been studied by steady-state and stopped-flow spectrophotometric methods [47,48,51–54]. As with other flavin-dependent monooxygenases, the catalytic cycle of microbial NMOs can be divided into two half-reactions. The first

2.4 Catalytic cycle of NMOs

35

of these is the reductive half-reaction where NADH or NADPH transfers a hydride to reduce the flavin cofactor [55]. The second is the oxidative half-reaction where reduced flavin reacts with molecular oxygen and the substrate is hydroxylated. A general kinetic mechanism of NMOs, particularly that of Af SidA, is depicted in 򐂰Fig. 2.3 [51]. The first step involves the reaction of the oxidized flavin with NADPH (򐂰Fig. 2.3B). This produces reduced flavin and NADP+, which remains bound to the enzyme, a characteristic common among Class B monooxygenases [43]. The reduced enzyme-NADP+ complex then reacts with molecular oxygen to form a C4a-hydroperoxyflavin, the hydroxylating species (򐂰Fig. 2.3C* and 2.3D). The presence of NADP+ plays a critical role in the stabilization of this intermediate. The C4a-hydroperoxyflavin is long-lived and the enzyme turns over rapidly only when substrate binds to the active site. In NMOs, substrate hydroxylation is highly specific and the enzyme predominately produces hydrogen peroxide if a non-substrate ligand binds, such as in the case of L-lysine binding to Af SidA [47,48,51–54]. Hydroxylation, involving transfer of the distal oxygen of the flavin C4a-hydroperoxide to give the C4a-hydroxide intermediate, is followed by dehydration of the flavin to produce the oxidized cofactor (򐂰Fig. 2.3F). Release of the hydroxylated product, and lastly, NADP+, completes the catalytic cycle (򐂰Fig. 2.3G). The two most important questions that have emerged from this general catalytic cycle pertain to the mechanism of stabilization of the C4a-hydroperoxyflavin and the structural characteristics that lead to the high degree of substrate selectivity. Recent mechanistic and structural information on members of the NMO family addressing these questions has recently been made available.

2.4.1 Flavin reduction in NMOs Direct measurement of the rates of flavin reduction in Af SidA and Pa PvdA show this to be the rate-limiting step in the catalytic cycle (򐂰Fig. 2.3) [51,52,54]. This is consistent with previous studies by Macheroux and coworkers where FAD analogs with IucD were used [49]. Studies with Af SidA show a biphasic reduction of the flavin with NAD(P) H [51,54]. This was originally analyzed by Dubois and coworkers as corresponding to NADPH binding followed by hydride transfer. Recent primary kinetic isotope effect (KIE) studies with NADPH resulted in KIE values for kred of 5.5 for both phases of reduction [54]. These results show that hydride transfer is rate-limiting in reduction and occurs in both phases, suggesting that NADPH is reacting with different enzyme forms. The concentration dependence of these two phases could not accurately be measured with NADPH due to its high affinity to Af SidA. In contrast, with NADH, only the fast phase was dependent on the concentration of the coenzyme. A KIE value of 4.7 was measured in this phase, suggesting that it corresponds to hydride transfer. The slow phase was concentration and isotope independent. Thus, with NADH, the first phase corresponds to flavin reduction while the second phase most likely is NAD+ release. These studies also showed that Af SidA is specific for transfer of the pro-R hydrogen at the C-4 position of the nicotinamide ring in both NADPH and NADH, an observation consistent with other flavin-dependent monooxygenases [56]. Lastly, the rates of the reductive steps in Pa PvdA and Af SidA are independent of L-ornithine [51,52]. This lack of an effect by substrate contrasts with of the behavior exhibited by p-hydroxybenzoate hydroxylase (PHBH), where substrate enhances reduction by ~1.4 × 105-fold. This highlights a major mechanistic difference among the flavin-dependent monooxygenases [57].

N O

NH

H2O

F

NADP쎵 L-orn-OH

H

N

3 s앥1

NH

N NH

H2O2 0.021 s앥1

O

N

N H OH O

R O

NADPH

R N

N

N

A Kd 쏝 1 mM

NADPH

R

G very fast

O

NH

O

N

N

N

O

O

8 s앥1

E

B

0.67 s앥1

NADP쎵 L-orn

NADP쎵

NADP쎵

R

NH

O

N N H O O HO

N

R

NH

O

L -orn D* K 쏁 2.2 mM d

N H O O HO

D

NADP L-orn

쎵

O2 1.3 쎹 105 M앥1 s앥1

N H

R

NH

Kd 쏁 680 mM

L-orn

N

O

C

R N

2.5 쎹 104 M앥1 s앥1

O2

O

N앥

N

C*

N H

N

O

N앥 NH

O

Fig. 2.3: Kinetic mechanism of Af SidA [51]. NADPH first binds to the oxidized form of the enzyme (A), which then reduces FAD to yield FADH– (B). After reduction two kinetic pathways are possible. The first and preferred pathway is where L-ornithine can bind to the NADP+-enzyme complex (C), which then reacts with molecular oxygen to form a C4a-hydroperoxyflavin (D). Alternatively, molecular oxygen can first react with FADH− to form a C4a-hydroperoxyflavin (C*) after which the binding of L-ornithine occurs (D*). After the formation of a stable C4a-hydroperoxyflavin and when L-ornithine is bound, hydroxylation occurs resulting in a C4a-hydroxyflavin (E). The flavin is then dehydrated (F) to regenerate oxidized FAD, which then allows for release of NADP+ and hydroxylated L-ornithine (G). In the absence of substrate, the C4a-hydroperoxyflavin slowly decays to hydrogen peroxide and oxidized flavin (H).

NADP쎵 L-orn-OH

NADP쎵 L-orn-OH

R N

36 2 Flavin-dependent monooxygenases in siderophore biosynthesis

2.5 Three-dimensional structure of NMOs

37

2.4.2 Flavin oxidation in NMOs In solution, the activation of molecular oxygen involves the single electron transfer from reduced flavin to oxygen, which produces a superoxide-flavin semiquinone caged radical pair that leads to formation of the C4a-peroxyflavin [58]. This reaction is relatively slow (~250 M−1 s−1) due to the spin inversion of triplet oxygen that is required to react with singlet reduced flavin. In contrast, flavoproteins can accelerate this reaction, with second-order rate constants in the range of 103–106 M−1 s−1 [59]. Following oxygen activation, the reaction with reduced flavin can proceed in various ways: a) release of the superoxide anion, b) second electron transfer followed by release of hydrogen peroxide or c) formation and stabilization of a covalent C4a-hydroperoxyflavin intermediate [58]. Flavin oxidases react with oxygen to give a flavin C4a-peroxide that breaks down directly to oxidized flavin and hydrogen peroxide, but the flavin monooxygenases must stabilize the oxygenated flavin intermediate for hydroxylation to occur. This intermediate can exist as a C4a-peroxyflavin or C4a-hydroperoxyflavin depending on whether the enzyme performs a nucleophilic or electrophilic reaction, respectively. The first electron transfer step in NMOs has not been observed to date, as the caged radical pair rapidly collapses to form the C4a-peroxyflavin [51,52]. In NMOs, the rate of activation of molecular oxygen is moderately enhanced by L-ornithine binding. Reduced Af SidA reacts with molecular oxygen on the order of 104 M−1 s−1 and 105 M−1 s−1 in the absence and presence of L-ornithine, respectively [51]. After oxygen addition a C4a-peroxyflavin is formed, however, the hydroxylation of L-ornithine requires the protonated form of this intermediate, the C4a-hydroperoxyflavin [55]. In Pa PvdA, a species at 361 nm is observed immediately after oxygen addition in the absence of L-ornithine. In the presence of L-ornithine the intermediate shifts to 376 nm [52]. The species with an absorbance peak at 361 nm is consistent with the C4a-peroxyflavin intermediate observed in cyclohexanone monooxygenase [60]. Thus, it has been suggested that this species corresponds to the C4a-peroxyflavin and the shift to 376 nm observed upon L-ornithine binding is due to protonation to form the C4a-hydroperoxyflavin [52]. In Af SidA , however, a species is observed at 382 nm immediately after oxygen addition, indicative of a C4a-hydroperoxyflavin. This species remains unchanged over a pH range of 6–10 and does not shift upon L-ornithine binding [51,61], indicating that in Af SidA, the pKa value of the C4a-hydroperoxyflavin is high (> 10). It is not clear why these two related enzymes differ in the manner in which their oxygenated flavin intermediates are generated. As mentioned previously, NADP+ remains bound after reduction where the enzymeNADP+ complex then reacts with molecular oxygen to form a stable, long-lived C4ahydroperoxyflavin. The longevity of the C4a-hydroperoxyflavin ensures hydroxylation of L-ornithine and prevents the release of hydrogen peroxide and regeneration of oxidized flavin (򐂰Fig. 2.3H). NADH can effectively reduce Af SidA; however, upon reaction with oxygen the C4a-hydroperoxyflavin intermediate is less stable [48]. These results indicate a specific role of NADP(H) in the stabilization of the C4a-hydroperoxyflavin which will be discussed in Section 2.7.

2.5 Three-dimensional structure of NMOs Determination of the three-dimensional structure of NMOs has been hampered by the lack of stable proteins that contain a tightly bound flavin cofactor [47,49]. The

38

2 Flavin-dependent monooxygenases in siderophore biosynthesis

structure of the first NMO, Pa PvdA, was only recently solved (򐂰Fig. 2.4) [62]. Conditions were obtained where the protein was co-crystallized with FAD, NADP+, and L-ornithine. Furthermore, the crystallized protein complex could be chemically reduced. Thus, the structures of the oxidized (1.9 Å) and reduced enzymes (3.03 Å) are available (PDB codes 3S5W and 3S61, respectively). The structure of Pa PvdA contains three domains, including two α/β-Rossmann-like domains for FAD- and NADPHbinding, and a small helical domain for L-ornithine-binding (򐂰Fig. 2.4) [62]. Amino acid sequence alignment among members of the NMO family predicts that all contain the three-domain architecture (򐂰Fig. 2.5). We will discuss the structure of Pa PvdA in detail; however, the analysis is also applicable to Af SidA as our group has recently solved the structure. Overall, our analysis indicates that the three-dimensional structures of these two enzymes are virtually identical1.

FAD binding domain

L-ornithine

binding

domain

FAD L-orn

NADP쎵

NADPH binding domain

Fig. 2.4: Overall structure of Pa PvdA (PDB code 3S5W). The FAD-binding domain is shown in cyan, the NADPH-binding domain in slate blue, and the L-ornithine-binding domain in salmon. FAD carbons are shown in yellow, NADP+ carbons in pink, and L-ornithine carbons in green.

1

Unpublished data (Franceschini S., Fedkenheuer F., Robinson H., Vogelaar, N., Sobrado P., and Mattevi A.)

2.5 Three-dimensional structure of NMOs

39

Fig. 2.5: Primary sequence alignment of NMOs with different substrate specificities. The monooxygenases Pa PvdA and Af SidA, the 1, 3-diaminopropane monooxygenase RhbE, the putrescine monooxygenase AlcA, the cadaverine monooxygenase DesB, and the L-lysine monooxygenase IucD are aligned. Sequence identities to Pa PvdA are 38% for Af SidA, 27% for RhbE, 26% for AlcA, 30% for DesB, and 29% for IucD. The green bar indicates the FAD-binding motif, the blue bar the NADPH binding motif, and the red bar the L/FATGY motif. Residues starred in green, blue, and red are proposed to be involved in FAD, NADPH, and L-ornithine binding, respectively, in Pa PvdA. The secondary structure of Pa PvdA (PDB code 3S5W) is threaded to the alignment. Loops represent α-helices, arrows represent β-sheets, and spaces represent loops. The alignment was performed with Multalin and ESpript. L-ornithine

40

2 Flavin-dependent monooxygenases in siderophore biosynthesis

2.5.1 FAD-binding domain The FAD-binding domain in Pa PvdA is the largest domain composed of residues 1–171, 356–396, and 405–443 [62]. The FAD-binding motif is located near the N-terminus in Pa PvdA (򐂰Fig. 2.5, green bar). This classic GXGXXG flavin-binding motif is conserved in phenylacetone monooxygenase (PAMO), cyclohexanone monooxygenase (CHMO), and the bFMO monooxygenase from Methylophaga sp. strain SK1 [63–65]. In NMOs, the last glycine is replaced by a proline (GXGXGP). It has been proposed that this amino acid change is partially responsible for the weak affinity for FAD among some members of this class of enzymes [66]. Upon solving the structure of Pa PvdA, a unique FAD binding motif in NMOs that consists of GXGXXN is observed, which originates through a shift of two amino acids to the C-terminus from the previously proposed motif. As seen in 򐂰Fig. 2.5, this motif is conserved among NMOs, although in Af SidA the asparagine is replaced by a serine. The recombinant forms of Pa PvdA and IucD are purified as inactive apo-proteins, lacking FAD. This is due to their relatively low affinities for FAD (~25 µM) [47,49]. In both cases, excess FAD is added to obtain an active, holo-enzyme. Interestingly, Af SidA and MbsG have been isolated with > 50% of FAD bound per monomer, suggesting tight binding of the FAD [48,50]. Crystallographic analysis shows that interactions between the protein and the flavin involve hydrogen bonding between the backbone and several side chains of Pa PvdA. Comparison of the FAD-binding residues with that of bFMO, which contains a tightly bound cofactor, indicates there are no significant differences, and the hydrogen bonds holding the FAD in place are almost identical (򐂰 Fig. 2.6). Olucha et al. has suggested that the weak affinity of Pa PvdA for FAD might be due to the fact that the active site is solvent exposed, as opposed to the solvent-sequestered FAD binding site in bFMO [62,65]. As mentioned above, the structure of Af SidA is almost identical to Pa PvdA1. There are no significant differences in the FAD-binding cleft in Af SidA when compared to Pa PvdA. This indicates that the tight-binding observed with Af SidA is likely due to reasons that cannot be determined by X-ray crystallographic analysis. Further work is needed to elucidate the cause of this key difference between these otherwise very similar NMOs.

2.5.2 NADPH-binding domain The NADPH-binding domain in Pa PvdA is composed of residues 170–245 and 285– 355 [66]. The second nucleotide-binding motif, GXGXXG/A, is found in this domain and is located in the middle of the polypeptide chain (򐂰Fig. 2.5, blue bar). NMOs, such as CchB, Pa PvdA, and IucD, are specific to NADPH as no turnover with NADH is observed [39,47,67]. In contrast, Af SidA and MbsG are more promiscuous as they are active with both coenzymes [48,50]. Although, Af SidA preferentially utilizes NADPH, MbsG appears to be slightly more specific for NADH. The crystal structure of Pa PvdA provides structural evidence for the specificity of NADPH over NADH. Olucha and coauthors report that the conserved Arg240 forms an ionic interaction with the phosphate group of the adenine ribose. An analogous Arg residue is found in all NMOs, suggesting that this residue might be partially responsible for the selectivity for NADPH in these enzymes (򐂰Fig. 2.5).

2.5 Three-dimensional structure of NMOs A

O N

Trp52 His53

2.83 Å

H

H H

3.16 Å

O N

2.86 Å

H

(R-group O) Gln323

2.91 Å

H

Trp52

3.20 Å

2.62 Å

3.21 Å

Pro20

H

Ser18

2.71 Å

O P OH

3.17 Å

Pro17 (backbone N)

2.55 Å

His68

O (backbone N) Gln51

O P OH

N

Ser326

O

O

N

2.70 Å

H

H

O P OH

O

O

H

O Ser21

N

OH

HO

H

Asn78

H

HO H

H

3.12 Å

NH

N

OH

HO 2.79 Å

B

Leu409 O (backbone N)

N

HO H

Gln64 (backbone N)

NH

N Thr407

2.84 Å

41

3.02 Å

O P OH

O

O

CH2

CH2 OH 2.66 Å OH 2.72 ÅAsp45 N

3.11 Å

N

O N

Lys46 (backbone N)

OH 2.42 Å OH 2.68 ÅGlu43 N

N

(backbone N) 3.06 Å (backbone O) H2N Val130

3.31 Å

Lys44 (backbone N)

N

2.83 Å

3.38 Å

(backbone O) H2N NH3

2.92 Å (backbone N) Val131

Fig. 2.6: FAD interactions among Class B monooxygenases. Interactions of FAD with residues of Pa PvdA (A) and bFMO (B).

2.5.3 L-Ornithine-binding domain

In Pa PvdA, the L-ornithine-binding domain is the smallest domain, which is composed of residues 248–285 and 398–404 [66]. It was first proposed that the L/FATGY motif was predominately involved in substrate-binding (򐂰Fig. 2.5, red bar). This hydrophobic sequence is conserved among Class B monooxygenases, but after determination of the crystal structure of Pa PvdA, the L/FATGY sequence was found to not be involved in any significant interactions with bound L-ornithine and is in fact located in the NADPH-binding domain [62]. Instead, Asn254 and Thr283, and Lys69 interact with the α-carboxylate and amino groups, respectively, of L-ornithine and provide the structural basis for the steroselectivity of L-ornithine over D-ornithine. The residues Gln64 and Asn284 provide interactions with the N5-amine of L-ornithine that hold it near the C4a of the isoalloxazine ring and in a favorable orientation for hydroxylation [62]. Amino acid sequence alignment of NMOs with different substrate specificities indicates that key residues interacting with L-ornithine in Pa PvdA are not conserved (򐂰Fig. 2.5). In particular, the amino acids Lys69, Asn254, and Asn283, which help provide specificity for L-ornithine in Pa PvdA, are conserved in Af SidA but not in all NMOs with different substrate specificities. These structural differences may be responsible for the different substrate selectivities among NMOs.

42

2 Flavin-dependent monooxygenases in siderophore biosynthesis

2.6 The structural basis of substrate specificity in NMOs As mentioned previously, NMOs are highly specific as to the hydroxylated substrate. For example, Af SidA and Pa PvdA are highly specific for the hydroxylation of L-ornithine [47,48]. Structurally similar analogs such as D-ornithine, L-lysine, L-arginine, and 1,4-diaminobutane, yield little or no hydroxylated products with these NMOs. Similar stringent selectivity has been observed with the L-lysine monooxygenases IucD and MbsG [49,50]. While both Af SidA and Pa PvdA are very specific for hydroxylation of L-ornithine, substrate analogs stimulate turnover, as reflected in NADPH oxidation, by causing a “misfiring” of the C4a-hydroperoxyflavin, which simply breaks down to hydrogen peroxide and oxidized flavin [47,48]. The most dramatic case is seen with L-lysine, a nonsubstrate effector that causes nearly complete uncoupling of the reaction. The structure of Pa PvdA complexed with L-ornithine and NADP+ provides insight into the mechanism of substrate selectivity (򐂰Fig. 2.7). As suggested by Olucha and coworkers, since 6 L-lysine is longer than L-ornithine, the N -amine may protrude too far into the active site which impedes favorable positioning and stabilization of the C4a-hydroperoxyflavin (򐂰Fig. 2.7) [62]. Thus, Pa PvdA and other NMOs may employ a structural “ruler mechanism” to select the proper substrate for hydroxylation. In contrast to the substrate selectivity of NMOs, mammalian FMOs have a broad substrate specificity as they will oxygenate a range of drugs and xenobiotics [68]. In particular, microsomal liver FMO can catalyze the hydroxylation of a number of soft nucleophiles including secondary and tertiary amines, thiols, hydrazines, sulfides, thiocarbamides, and thioamides [69–72]. Similarly, PAMO and CHMO can catalyze the oxygenation of several ketones and alkyl sulfides [73,74]. In Af SidA, the reaction of the C4a-hydroperoxyflavin with L-ornithine is concentration dependent and saturates at high concentrations of L-ornithine, indicating formation of a complex [51]. In contrast, the reactivity of the C4a-hydroperoxyflavin with bFMO has been shown to occur as a second-order reaction. Here, the intermediate has been compared to a “cocked-gun”, which is ready to fire upon contact with substrate [75]. This suggests that formation of a complex in NMOs constitutes another level of substrate selectivity ensuring that only the appropriate substrate is hydroxylated.

2.7 Mechanism of stabilization of the C4a-hydroperoxyflavin by NADP+ Insights into the mechanism by which the C4a-hydroperoxyflavin is stabilized by NADP(H) in NMOs has become evident from an examination of the three-dimensional structure of reduced Pa PvdA in complex with NADP+ as solved by Lamb and coworkers [62]. The structure shows NADP+ in a position that is not optimal for hydride transfer, but instead is positioned to stabilize the C4a-hydroperoxyflavin (򐂰Fig. 2.7A). Furthermore, the observed binding for L-ornithine positions the N5-amine in the correct location for attacking the distal oxygen of the flavin intermediate (򐂰Fig. 2.7A). This analysis is consistent with the results and the proposal of the “moonlighting” effect of NADPH in bFMO reported by Mattevi and Fraaije [65]. In this related enzyme, NADP+ is bound in the active site in a position optimal for stabilization of a C4a-hydroperoxyflavin, similar to that seen in the structure of Pa PvdA. Recently, site-directed mutagenesis of residues

2.7 Mechanism of stabilization of the C4a-hydroperoxyflavin by NADP+

43

that are expected to stabilize the binding of NADP+ produced mutant enzymes that are highly uncoupled in bFMO [76]. This supports the proposal that the observed position of NADP+ in the active site of these enzymes is important for stabilization of the C4ahydroperoxyflavin. The importance of C4a-hydroperoxyflavin stabilization by NADP+ is illustrated by the observation that chemical reduction of Af SidA with dithionite followed by exposure to oxygen in the presence of L-ornithine produces only hydrogen peroxide [51]. The crystallographically observed conformation of NADP+ in the active site of Pa PvdA shows that it creates a pocket for the formation of the C4a-hydroperoxyflavin (򐂰Fig. 2.7B), shielding it from solvent and providing hydrogen-bonding partners for stabilization. This pocket also plays a role in the selectivity for L-ornithine, since larger substrates (e.g., L-lysine) protrude too far into the substrate binding pocket and perturb the interaction of NADP+ with the C4a-hydroperoxide (򐂰Fig. 2.7C).

A

B

C

H2N

H2N

O

O H

쎵

N

R

O

N

O O NO N HR N O H2 OH H

H NH

쎵

N

앥

O

O

R O

쎵

NH3

O O OH H

N H

N HN R

O

OH O NH

N

O앥

O

O 쎵

NH3

Fig. 2.7: Stabilization of the C4a-hydroperoxyflavin in the active site of Pa PvdA (PDB code 3S61). The active site of Pa PvdA (A) where FAD carbons are shown in yellow, NADP+ carbons in pink, L-ornithine carbons in green, and side chain carbons in gray. This provides a role of stabilization for the C4a-hydroperoxyflavin by NADP+ along with the pocket (dashed circle) for the dioxygen of the intermediate defined by NADP+ and L-ornithine (B). When L-lysine binds, the amine protrudes too far into the active site to be hydroxylated and stimulates the release of hydrogen peroxide (C).

44

2 Flavin-dependent monooxygenases in siderophore biosynthesis

The exact conformational changes that occur to locate NADP+ in the position observed in the structures of all Class B monooxygenases are not known, mainly because the structures in the absence of NADP+ have been elusive. However, biochemical evidence of coenzyme-induced conformational changes has been reported in Af SidA. By monitoring the fluorescence of FAD upon binding of NAD+, an increase in flavin fluorescence is observed, while binding of NADP+ produces a decrease in fluorescence [54]. These results indicate different modes of binding of the two coenzymes in the active site that must be modulated by the 2’-phosphate of NADP+. It has also been shown that binding of NADP+ protects Af SidA from degradation by trypsin in limited proteolysis studies. This effect is not observed with L-ornithine or NAD+. Thus, it appears that NADP(H) induces conformational changes that modulate the interaction of the nicotinamide ring with the flavin cofactor. These conformational changes could either modulate the position of NADP+ or of the isoalloxazine ring of the flavin.

2.8 Activation of NMOs by amino acid binding The oxidation of reduced NMOs occurs in two distinct phases [51–53]. The first phase is the formation of a stable C4a-hydroperoxyflavin intermediate as discussed above, and the second phase is the subsequent dehydration of the C4a-hydroxyflavin to yield fully oxidized FAD. Formation of a stable C4a-hydroperoxyflavin occurs in Af SidA and Pa PvdA in the absence of substrate, but the rate of formation has been shown to increase in the presence of saturating levels of L-ornithine by ~80-fold and ~15-fold for Pa PvdA and Af SidA, respectively [47,48,51,52]. Binding of the amino acids L-citrulline and L-arginine also enhance the activity of Af SidA [53]. This effect has been observed on both reduction by NADPH and formation of the C4a-hydroperoxyflavin. L-arginine provides the highest rate enhancement with a ~10-fold increase in the rate of flavin reduction and a ~100-fold increase for the formation of the C4a-hydroperoxyflavin. Activation by L-arginine is concentration dependent and displays Michaelis-Menten kinetics with a Kactivation value for the C4a-hydroperoxyflavin of ~620 μM. It is clear that binding of L-arginine to Af SidA stabilizes the C4ahydroperoxyflavin as the enzyme becomes more coupled, changing from ~90% in the absence of L-arginine to ~100% in the presence of L-arginine. The mechanism of oxygen activation in flavoenzymes has been studied extensively [77–80]. A common trait among flavin oxidases is the presence of a positive charge near the N5 of the isoalloxazine ring. For example, the positive charges of a lysine and histidine residue in N-methyltryptophan oxidase and glucose oxidase, respectively, have been shown to activate oxygen. At saturating levels of L-arginine are present, formation of a C4a-hydroperoxyflavin in Af SidA occurs ~100-fold faster than in the absence of L-arginine [53]. Binding of L-arginine to Af SidA could play a role in oxygen activation similar to the lysine and histidine residues in flavin oxidases. Here, the positive charge of the guanidinium group in L-arginine could be involved in the activation of molecular oxygen.

2.9 Unusual NMOs M. smegmatis G (MbsG), an L-lysine monooxygenase involved in mycobactin biosynthesis (򐂰Fig. 2.2B), does not share the same mechanistic features as other members of

2.10 High-throughput screening assay to identify inhibitors of NMOs

45

the NMO family. As discussed earlier, Pa PvdA and Af SidA are highly coupled and stabilize oxygenated flavin intermediates to prevent the release of oxygen reactive species. These characteristics are not conserved in MbsG [50]. Perhaps the most interesting aspect of MbsG is the lack of control of oxygen reactivity. When substrate is absent, the enzyme functions like an NAD(P)H oxidase producing superoxide and hydrogen peroxide [50]. Once L-lysine is introduced, superoxide and hydrogen peroxide production decreases and product formation occurs, albeit at low levels. At saturating substrate concentrations MbsG is ~30% coupled, and under physiological concentrations only ~12% coupled. While superoxide and hydrogen peroxide are reactive oxygen species and seemingly detrimental, in some cases they may play a role in the expression of genes regulated by the overall oxidation state of the cell [68]. In the biosynthesis of hydroxamate-containing siderophores, hydroxylation is thought to be the first step in siderophore biosynthesis [35,41,49,81]. A recent report proposed that in the biosynthesis of mycobactin, hydroxylation of L-lysine occurs after its incorporation into the siderophore [31]. This conclusion was based upon lipidomic analysis that showed the presence of di/monodeoxymycobactin and mycobactin in media growth of wild-type M. tuberculosis. Furthermore, in an mbtG knockout strain, only dideoxymycobactin was detected. Thus, it was proposed that MbtG hydroxylation of lysine at the N6-position was the final step in mycobactin biosynthesis [31]. A recent report provided a possible explanation for the presence of dideoxymycobactin. It was determined that MbtF, a peptide synthetase involved in acylation, has high specificity for N6-hydroxy-L-lysine. However, it is also capable of acylating L-lysine, which can lead to the formation of dideoxymycobactin instead of mycobactin [82]. This could explain the accumulation of ~2% dideoxymycobactin in mbtG knockouts of M. tuberculosis compared to the wild-type. Here, mycobactin biosynthesis in the deoxy-state is still able to occur, but at a much reduced rate. These results, as well as the biochemical characterization of MbsG as an L-lysine hydroxylase, support the hydroxylation of free L-lysine in the biosynthesis of mycobactin [50].

2.10 High-throughput screening assay to identify inhibitors of NMOs As discussed in the introduction, NMOs are essential for virulence in many human pathogens, and the recent mechanistic and structural studies provide valuable information for new therapeutic strategies. In addition to rational drug design, the screening of small molecule libraries for inhibitors of NMOs could help identify new drugs. A fluorescence polarization binding assay that can monitor the binding of small molecules to the active site of NMOs has recently been developed (򐂰Fig. 2.8) [83]. This assay utilizes a fluorescently labeled ADP molecule where the ADP portion targets the NADPH binding site, while the fluorophore (TAMRA) is connected via a hydrophobic linker. The tricyclic ring of the fluorophore is postulated to stack with the isoalloxazine ring of the FAD, further increasing its affinity for NMOs. Both Af SidA and MbsG display low Kd values for the fluorescent ligand (2–6 µM). Such binding suggests that the ligand could potentially be utilized to screen for inhibitors in a number of NMOs. The fluorophore binds to the active site of Af SidA as it is displaced by NADP+ and NAD+. Furthermore, L-lysine and L-ornithine can displace the fluorophore, indicating its effectiveness for identifying analogs of either substrate. A pilot screen of 160 molecules identified sanguinarine sulfate as a weak inhibitor of Af

46

2 Flavin-dependent monooxygenases in siderophore biosynthesis

A

NH2 N H N

O

O () 5

N

O N H

() 5

O P O P O O

N

O O

N

N

O OH OH

O

COO

N

B

Signal ON

Signal OFF

Inhibitor

Fig. 2.8: The fluorescence binding assay developed for Af SidA. The fluorophore used to detect inhibitors of NMOs (A) and the cartoon depiction of the detection of inhibitors with the fluorophore shown with the crystal structure of Af SidA (B).

SidA with an IC50 value of ~500 µM [83]. These results suggest that screening of large molecule libraries will most likely identify potent and specific inhibitors of NMOs.

2.11 Conclusions NMOs constitute a large family of flavin-dependent monooxygenases involved in the biosynthesis of hydroxamate-containing siderophores. These enzymes are linked to the virulence of several human pathogens and, thus, represent potentially novel drug targets. Mechanistically, this family shares many similarities with other members of the Class B flavin monooxygenases, but with some exceptions (as with, e.g., MbsG). A unique feature of NMOs is their remarkable specificity for substrate hydroxylation, which is achieved by a “molecular ruler mechanism.” Furthermore, the exact placement of substrate along with the “moonlighting” role of NADP+ provides a well-defined pocket within which the C4a-hydroperoxyflavin is stabilized. Also, oxygen activation in NMOs is enhanced by the binding of certain amino acids. Elucidation of the

2.12 References

47

three-dimensional structures along with a developed high-throughput assay to screen for inhibitors provides a solid foundation for development of drugs that target NMOs.

Acknowledgements This work was supported by a grant from the National Science Foundation (MCB1021384 to PS).

2.12 References [1] Pierre JL, Fontecave M. Iron and activated oxygen species in biology: the basic chemistry. Biometals 1999;12:195–9. [2] Maton A, (Ed.). Human Biology and Health. Englewood Cliffs, New Jersey, USA: Prentice Hall; 1993. [3] Tagawa K, Tsujimoto HY, Arnon DI. Role of chloroplast ferredoxin in the energy conversion process of photosynthesis. Proc Natl Acad Sci USA 1963;49:567–72. [4] Kim J, Rees DC. Nitrogenase and biological nitrogen fixation. Biochemistry 1994;33:389–97. [5] Gutteridge JM, Halliwell B. Iron toxicity and oxygen radicals. Baillieres Clin Haematol 1989;2:195–256. [6] Stohs SJ, Bagchi D. Oxidative mechanisms in the toxicity of metal ions. Free Radic Biol Med 1995;18:321–36. [7] Schaible UE, Kaufmann SH. Iron and microbial infection. Nat Rev Microbiol 2004;2:946–53. [8] Eisenstein RS. Iron regulatory proteins and the molecular control of mammalian iron metabolism. Annu Rev Nutr 2000;20:627–62. [9] Ward PP, Uribe-Luna S, Conneely OM. Lactoferrin and host defense. Biochem Cell Biol 2002;80:95–102. [10] Marx JJM. Iron and infection: Competition between host and microbes for a precious element. Best Pract Res Clin Haematol 2002;15:411–26. [11] Andrews SC, Robinson AK, Rodriguez-Quinones F. Bacterial iron homeostasis. FEMS Microbiol Rev 2003;27:215–37. [12] Sandy M, Butler A. Microbial iron acquisition: Marine and terrestrial siderophores. Chem Rev 2009;109:4580–95. [13] Loomis LD, Raymond KN. Solution equilibria of enterobactin and metal-enterobactin complexes. Inorg Chem 1991;30:906–11. [14] Hider RC, Kong X. Chemistry and biology of siderophores. Nat Prod Rep 2010;27:637–57. [15] Neilands JB. Microbial iron compounds. Annu Rev Biochem 1981;50:715–31. [16] Reid RT, Live DH, Faulkner DJ, Butler A. A siderophore from a marine bacterium with an exceptional ferric ion affinity constant. Nature 1993;366:455–8. [17] Perry RD, Balbo PB, Jones HA, Fetherston JD, DeMoll E. Yersiniabactin from Yersinia pestis: biochemical characterization of the siderophore and its role in iron transport and regulation. Microbiology 1999;145(Pt 5):1181–90. [18] Murakami T, Ise K, Hayakawa M, Kamei S, Takagi SI. Stabilities of metal-complexes of mugineic acids and their specific affinities for iron(iii). Chem Lett 1989:2137–40. [19] Schwarzenbach G, Anderegg G. Metal complexes with biguanide. Pharm Acta Helv 1963;38:547–52. [20] Hou Z, Raymond KN, O’Sullivan B, Esker TW, Nishio T. A preorganized siderophore: thermodynamic and structural characterization of alcaligin and bisucaberin, microbial macrocyclic dihydroxamate chelating agents1. Inorg Chem 1998;37:6630–7. [21] Torres L, Perez-Ortin JE, Tordera V, Beltran JP. Isolation and characterization of an Fe(III)chelating compound produced by Pseudomonas syringae. Appl Environ Microbiol 1986;52:157–60.

48

2 Flavin-dependent monooxygenases in siderophore biosynthesis

[22] Wong GB, Kappel MJ, Raymond KN, Matzanke B, Winkelmann G. Coordination chemistry of microbial iron transport compounds .24. Characterization of coprogen and ferricrocin, 2 ferric hydroxamate siderophores. J Am Chem Soc 1983;105:810–5. [23] Carrano CJ, Drechsel H, Kaiser D, et al. Coordination chemistry of the carboxylate type siderophore rhizoferrin: the iron(iii) complex and its metal analogs. Inorg Chem 1996;35:6429–36. [24] Palanche T, Blanc S, Hennard C, Abdallah MA, Albrecht-Gary AM. Bacterial iron transport: coordination properties of azotobactin, the highly fluorescent siderophore of Azotobacter vinelandii. Inorg Chem 2004;43:1137–52. [25] Cendrowski S, MacArthur W, Hanna P. Bacillus anthracis requires siderophore biosynthesis for growth in macrophages and mouse virulence. Mol Microbiol 2004;51:407–17. [26] Takase H, Nitanai H, Hoshino K, Otani T. Impact of siderophore production on Pseudomonas aeruginosa infections in immunosuppressed mice. Infect Immun 2000;68:1834–9. [27] Hissen AH, Wan AN, Warwas ML, Pinto LJ, Moore MM. The Aspergillus fumigatus siderophore biosynthetic gene sidA, encoding L-ornithine N5-oxygenase, is required for virulence. Infect Immun 2005;73:5493–503. [28] Sokol PA, Darling P, Woods DE, Mahenthiralingam E, Kooi C. Role of ornibactin biosynthesis in the virulence of Burkholderia cepacia: Characterization of pvdA, the gene encoding L-ornithine N(5)-oxygenase. Infect Immun 1999;67:4443–55. [29] Schrettl M, Bignell E, Kragl C, et al. Siderophore biosynthesis but not reductive iron assimilation is essential for Aspergillus fumigatus virulence. J Exp Med 2004;200:1213–9. [30] De Voss JJ, Rutter K, Schroeder BG, Su H, Zhu Y, Barry CE, 3rd. The salicylate-derived mycobactin siderophores of Mycobacterium tuberculosis are essential for growth in macrophages. Proc Natl Acad Sci USA 2000;97:1252–7. [31] Madigan CA, Cheng TY, Layre E, et al. Lipidomic discovery of deoxysiderophores reveals a revised mycobactin biosynthesis pathway in Mycobacterium tuberculosis. Proc Natl Acad Sci USA 2012;109:1257–62. [32] Frueh DP, Arthanari H, Koglin A, et al. Dynamic thiolation-thioesterase structure of a nonribosomal peptide synthetase. Nature 2008;454:903–6. [33] Crosa JH, Walsh CT. Genetics and assembly line enzymology of siderophore biosynthesis in bacteria. Microbiol Mol Biol Rev 2002;66:223–49. [34] Visca P, Ciervo A, Orsi N. Cloning and nucleotide sequence of the pvdA gene encoding the pyoverdin biosynthetic enzyme L-ornithine N5-oxygenase in Pseudomonas aeruginosa. J Bacteriol 1994;176:1128–40. [35] Heemstra JR, Jr., Walsh CT, Sattely ES. Enzymatic tailoring of ornithine in the biosynthesis of the Rhizobium cyclic trihydroxamate siderophore vic ibactin. J Am Chem Soc 2009;131:15317–29. [36] Quadri LE, Sello J, Keating TA, Weinreb PH, Walsh CT. Identification of a Mycobacterium tuberculosis gene cluster encoding the biosynthetic enzymes for assembly of the virulenceconferring siderophore mycobactin. Chem Biol 1998;5:631–45. [37] Mei B, Budde AD, Leong SA. sid1, a gene initiating siderophore biosynthesis in Ustilago maydis: molecular characterization, regulation by iron, and role in phytopathogenicity. Proc Natl Acad Sci USA 1993;90:903–7. [38] Herrero M, de Lorenzo V, Neilands JB. Nucleotide sequence of the iucD gene of the pColVK30 aerobactin operon and topology of its product studied with phoA and lacZ gene fusions. J Bacteriol 1988;170:56–64. [39] Pohlmann V, Marahiel MA. Delta-amino group hydroxylation of L-ornithine during coelichelin biosynthesis. Org Biomol Chem 2008;6:1843–8. [40] Lynch D, O’Brien J, Welch T, et al. Genetic organization of the region encoding regulation, biosynthesis, and transport of rhizobactin 1021, a siderophore produced by Sinorhizobium meliloti. J Bacteriol 2001;183:2576–85.

2.12 References

49

[41] Kang HY, Brickman TJ, Beaumont FC, Armstrong SK. Identification and characterization of iron-regulated Bordetella pertussis alcaligin siderophore biosynthesis genes. J Bacteriol 1996;178:4877–84. [42] Barona-Gomez F, Wong U, Giannakopulos AE, Derrick PJ, Challis GL. Identification of a cluster of genes that directs desferrioxamine biosynthesis in Streptomyces coelicolor M145. J Am Chem Soc 2004;126:16282–3. [43] van Berkel WJ, Kamerbeek NM, Fraaije MW. Flavoprotein monooxygenases, a diverse class of oxidative biocatalysts. J Biotechnol 2006;124:670–89. [44] de Lorenzo V, Bindereif A, Paw BH, Neilands JB. Aerobactin biosynthesis and transport genes of plasmid ColV-K30 in Escherichia coli K-12. J Bacteriol 1986;165:570–8. [45] Welzel K, Eisfeld K, Antelo L, Anke T, Anke H. Characterization of the ferrichrome A biosynthetic gene cluster in the homobasidiomycete Omphalotus olearius. FEMS Microbiol Lett 2005;249:157–63. [46] Eisendle M, Oberegger H, Zadra I, Haas H. The siderophore system is essential for viability of Aspergillus nidulans: functional analysis of two genes encoding l-ornithine N 5-monooxygenase (sidA) and a non-ribosomal peptide synthetase (sidC). Mol Microbiol 2003;49:359–75. [47] Meneely KM, Lamb AL. Biochemical characterization of a flavin adenine dinucleotidedependent monooxygenase, ornithine hydroxylase from Pseudomonas aeruginosa, suggests a novel reaction mechanism. Biochemistry 2007;46:11930–7. [48] Chocklett SW, Sobrado P. Aspergillus fumigatus SidA is a highly specific ornithine hydroxylase with bound flavin cofactor. Biochemistry 2010;49:6777–83. [49] Macheroux P, Plattner HJ, Romaguera A, Diekmann H. FAD and substrate analogs as probes for lysine N6-hydroxylase from Escherichia coli EN 222. Eur J Biochem 1993;213:995–1002. [50] Robinson R, Sobrado P. Substrate binding modulates the activity of Mycobacterium smegmatis G, a flavin-dependent monooxygenase involved in the biosynthesis of hydroxamatecontaining siderophores. Biochemistry 2011;50:8489–96. [51] Mayfield JA, Frederick RE, Streit BR, Wencewicz TA, Ballou DP, DuBois JL. Comprehensive spectroscopic, steady state, and transient kinetic studies of a representative siderophoreassociated flavin monooxygenase. J Biol Chem 2010;285:30375–88. [52] Meneely KM, Barr EW, Bollinger JM, Jr., Lamb AL. Kinetic mechanism of ornithine hydroxylase (PvdA) from Pseudomonas aeruginosa: substrate triggering of O2 addition but not flavin reduction. Biochemistry 2009;48:4371–6. [53] Frederick RE, Mayfield JA, DuBois JL. Regulated O2 activation in flavin-dependent monooxygenases. J Am Chem Soc 2011;133:12338–41. [54] Romero E, Fedkenheuer M, Chocklett SW, Qi J, Oppenheimer M, Sobrado P. Dual role of NADP(H) in the reaction of a flavin dependent N-hydroxylating monooxygenase. Biochim Biophys Acta 2012;1824:850–7. [55] Ghisla S, Massey V. Mechanisms of flavoprotein-catalyzed reactions. Eur J Biochem 1989;181:1–17. [56] Ryerson CC, Ballou DP, Walsh C. Kinetic isotope effects in the oxidation of isotopically labeled NAD(P)H by bacterial flavoprotein monooxygenases. Biochemistry 1982;21:1144–51. [57] Husain M, Massey V. Kinetic studies on the reaction of p-hydroxybenzoate hydroxylase. Agreement of steady state and rapid reaction data. J Biol Chem 1979;254:6657–66. [58] Massey V. Activation of molecular oxygen by flavins and flavoproteins. J Biol Chem 1994;269:22459–62. [59] Palfey BA, McDonald CA. Control of catalysis in flavin-dependent monooxygenases. Arch Biochem Biophys 2010;493:26–36. [60] Sheng D, Ballou DP, Massey V. Mechanistic studies of cyclohexanone monooxygenase: chemical properties of intermediates involved in catalysis. Biochemistry 2001;40:11156–67. [61] Romero E, Avila D, Sobrado P. Effect of pH on the reductive and oxidative half-reactions of Aspergillus fumigatus siderophore A. In Flavins and Flavoproteins (Miller, S Ed) In press 2011.

50

2 Flavin-dependent monooxygenases in siderophore biosynthesis

[62] Olucha J, Meneely KM, Chilton AS, Lamb AL. Two structures of an N-hydroxylating flavoprotein monooxygenase: ornithine hydroxylase from Pseudomonas aeruginosa. J Biol Chem 2011;286:31789–98. [63] Malito E, Alfieri A, Fraaije MW, Mattevi A. Crystal structure of a Baeyer-Villiger monooxygenase. Proc Natl Acad Sci USA 2004;101:13157–62. [64] Mirza IA, Yachnin BJ, Wang S, et al. Crystal structures of cyclohexanone monooxygenase reveal complex domain movements and a sliding cofactor. J Am Chem Soc 2009;131:8848–54. [65] Alfieri A, Malito E, Orru R, Fraaije MW, Mattevi A. Revealing the moonlighting role of NADP in the structure of a flavin-containing monooxygenase. Proc Natl Acad Sci USA 2008;105:6572–7. [66] Stehr M, Diekmann H, Smau L, et al. A hydrophobic sequence motif common to N-hydroxylating enzymes. Trends Biochem Sci 1998;23:56–7. [67] Plattner HJ, Pfefferle P, Romaguera A, Waschutza S, Diekmann H. Isolation and some properties of lysine N6-hydroxylase from Escherichia coli strain EN222. Biol Met 1989;2:1–5. [68] Krueger SK, Williams DE. Mammalian flavin-containing monooxygenases: structure/function, genetic polymorphisms and role in drug metabolism. Pharmacol Ther 2005;106:357–87. [69] Ziegler DM, Jollow D, Cook DE. Flavins and flavoproteins. 1971:475–97. [70] Prough RA. The N-oxidation of alkylhydrazines catalyzed by the microsomal mixed-function amine oxidase. Arch Biochem Biophys 1973;158:442–4. [71] Paulsen LL, Hyslop RM, Ziegler DM. S-oxidation of thioureylenes catalyzed by a microsomal flavoprotein mixed-function oxidase. Biochem Pharmacol 1974;23:3431–40. [72] Paulsen LL, Ziegler DM. Microsomal mixed-function oxidase-dependent renaturation of reduced ribonuclease. Arch Biochem Biophys 1977;183:563–70. [73] de Gonzalo G, Pazmino DET, Ottolina G, Fraaije MW, Carrea G. Oxidations catalyzed by phenylacetone monooxygenase from Thermobifida fusca. Tetrahedron Asymmetry 2005;16:3077–83. [74] Branchaud BP, Walsh CT. Functional-group diversity in enzymatic oxygenation reactions catalyzed by bacterial flavin-containing cyclohexanone oxygenase. J Am Chem Soc 1985;107:2153–61. [75] Paulsen LL, Ziegler DM. The liver microsomal FAD-containing monooxygenase. Spectral characterization and kinetic studies. J Biol Chem 1979;254:6449–55. [76] Orru R, Pazmino DE, Fraaije MW, Mattevi A. Joint functions of protein residues and NADP(H) in oxygen activation by flavin-containing monooxygenase. J Biol Chem 2010;285:35021–8. [77] McDonald CA, Fagan RL, Collard F, Monnier VM, Palfey BA. Oxygen reactivity in flavoenzymes: context matters. J Am Chem Soc 2011;133:16809–11. [78] Su Q, Klinman JP. Nature of oxygen activation in glucose oxidase from Aspergillus niger: the importance of electrostatic stabilization in superoxide formation. Biochemistry 1999;38:8572–81. [79] Roth JP, Klinman JP. Catalysis of electron transfer during activation of O2 by the flavoprotein glucose oxidase. Proc Natl Acad Sci USA 2003;100:62–7. [80] Bruckner RC, Winans J, Jorns MS. Pleiotropic impact of a single lysine mutation on biosynthesis of and catalysis by N-methyltryptophan oxidase. Biochemistry 2011;50:4949–62. [81] Ge L, Seah SY. Heterologous expression, purification, and characterization of an l-ornithine N(5)-hydroxylase involved in pyoverdine siderophore biosynthesis in Pseudomonas aeruginosa. J Bacteriol 2006;188:7205–10. [82] McMahon MD, Rush JS, Thomas MG. Analyses of MbtB, MbtE, and MbtF suggest revisions to the mycobactin biosynthesis pathway in Mycobacterium tuberculosis. J Bacteriol 2012;194:2809–18. [83] Qi J, Kizjakina K, Robinson R, Tolani K, Sobrado P. A fluorescence polarization binding assay to identify inhibitors of flavin-dependent monooxygenases. Anal Biochem 2012;425:80–7.

3 The flavin monooxygenases Stefania Montersino and Willem J. H. van Berkel

Abstract Flavin monooxygenases are ubiquitous enzymes that catalyze a wide variety of regioand enantioselective oxygenation reactions via the formation of a flavin C4a-(hydro)peroxide intermediate. Based on fold and function, flavin monooxygenases can be divided into six subfamilies. Subclasses A and B comprise single-component enzymes that rely on NAD(P)H as electron donor. Subclasses C–F comprise two-component enzymes, composed of a flavin reductase and a flavin-specific monooxygenase. FAD-containing hydroxylases (subclass A) convert many (hetero)aromatic substrates ranging from monophenols and uric acids to polyketide antibiotics and antitumor agents. FAD-containing monooxygenases (subclass B) perform Baeyer-Villiger oxidation, sulfoxidation and heteroatom hydroxylation reactions. FMN-dependent TIM-barrel enzymes (subclass C) catalyze Baeyer-Villiger oxidation, hydroxylation of long-chain alkanes and nitriloacetate, oxidation and desulfurization of sulfonates, and oxidation of aldehydes coupled with generation of bioluminescence. FMN/FAD hydroxylases with an acyl-CoA dehydrogenase fold (subclass D) convert mono- and polyphenols. FAD-specific styrene monooxygenases (subclass E) oxidize styrene derivatives to the corresponding epoxides. FAD-specific halogenases (subclass F) catalyze the regioselective chlorination and bromination of activated organic molecules for the production of antibiotics, antitumor agents and other natural products. During the past few years, many new family members and several unprecedented flavin monooxygenase reactions have emerged. This review illustrates selected features, tools to retrieve novel flavin monooxygenases from genome mining, and new findings on each of the six subclasses.

3.1 Introduction Monooxygenases catalyze the insertion of a single atom of molecular oxygen into the substrate, while the other oxygen atom is reduced to water. Most monooxygenases depend on transition metal ions or organic cofactors for oxygen activation [1], but certain monooxygenases can act without the need of a cofactor [2]. This review reports on flavin-dependent monooxygenases. These enzymes are unique in employing a pure organic cofactor, ubiquitously present in every kingdom, and able to cover a wide variety of regio- and enantioselective oxygenations. Nature uses flavin monooxygenases in key biological processes ranging from light emission [3], lignin and xenobiotics degradation [4–6], antibiotics [7,8], antitumor [9], siderophore [10] and auxin biosynthesis [11], to axon guidance [12,13].

52

3 The flavin monooxygenases

Oxygen activation in flavin monooxygenases involves the (transient) stabilization of flavin C4a-(hydro)peroxide [14,15]. This species performs either a nucleophilic or electrophilic attack on the substrate. Oxygenation reactions catalyzed by flavin monooxygenases include hydroxylations, Baeyer-Villiger oxidations, sulfoxidations, epoxidations and halogenations [16] (򐂰Fig. 3.1). The peculiarity of flavin monooxygenases resides in the ability to turn molecular oxygen into a true substrate by forming a quasi-stable flavin-oxygen adduct while other oxidative flavoenzymes perform purely oxygen-consuming reactions [17–20]. Embedment in the protein active site provides an environment that strongly increases the normally poor reactivity of reduced flavin with oxygen [18,21]. The structural and chemical factors controlling oxygen reactivity of flavoenzymes are not fully understood. However, accessibility, hydrogen bonding capacity and charge distribution of the reduced flavin are important factors involved [19,22–24]. Based on fold and function, flavin monooxygenases can be divided into six subfamilies [16] (򐂰Tab. 3.1, 򐂰Fig. 3.1, 򐂰Fig. 3.2). Each subclass possesses a certain fold and performs only a subset of monooxygenation reactions. Subclass A and B enzymes rely on NAD(P)H as external electron donor, while flavin reduction and oxygenation take place in the same polypeptide chain. Two-component enzymes, on the other hand, are composed of an NAD(P)H-dependent flavin reductase and a flavin-specific monooxygenase. One of the intriguing aspects of two-component monooxygenases is the mechanism of reduced flavin transfer. Transfer from the reductase component to the monooxygenase component needs to be efficient to avoid reoxidation of the flavin by molecular oxygen. Flavin transfer can occur through free diffusion or might involve protein channeling. Every two-component system seems to have evolved a different mechanism as studied for instance in bacterial luciferase [38] and alkane sulfonate monooxygenase [39]. With several systems it was demonstrated that the transfer of reduced flavin is thermodynamically favorable because the monooxygenase component binds the reduced cofactor much better than the oxidized form [35,40]. Hydroxylation R1

Baeyer-Villiger oxidation R1

R1

R1

OH R2

R2

O R2

O O R2

N-hydroxylation

Sulfoxidation R1

R1 N R3

쎵

R2

R3

R1 O

R2

R1

R1

S O

S R2

R2 R3 쏁 Cl, I, Br, F

Halogenation

Epoxidation R1

앥

R2 N O

R1

R1

R2

R2

R3

R2

Fig. 3.1: Reactions performed by flavin monooxygenases.

3.1 Introduction

53

Tab. 3.1: Flavin monooxygenase classification. Subclass

Structural fold

Reaction

References

A

Glutathione reductase-2

Hydroxylation

[14,25–27]

B

NADP Rossmann

Baeyer-Villiger oxidation

[28,29]

N, S, P, Se, I-oxygenation

[30]

N-hydroxylation

[10]

Light emission

[3]

Baeyer-Villiger oxidation

[31]

Desulfurization, sulfoxidation

[32]

Hydroxylation

[33]

C

Tim Barrel

D

Acyl-CoA dehydrogenase

Hydroxylation

[34]

E

Glutathione reductase-2

Styrene epoxidation

[26,35–36]

F

Glutathione reductase-2

Halogenation

[37]

A

C

B

D

E

F

Fig. 3.2: Protein folds of flavin monooxygenases. (A) 4-hydroxybenzoate 3-hydroxylase (1PBE), (B) from left to right phenylacetone monooxygenase (1W4X), ornithine hydroxylase (3S5W) and bFMO (2VQ7), (C) luciferase (1LUC), (D) 4-hydroxyphenylacetate 3-hydroxylase (2JBS), (E) styrene monooxygenase (3IHM), (F) tryptophan 7-halogenase (2ARD).

54

3 The flavin monooxygenases

In this review we summarize the occurrence and classification of flavin monooxygenases. Detailed mechanistic aspects of flavin monooxygenases are discussed in other chapters of this handbook. Information about the catalytic and structural properties of flavin reductases can be found in previous reports [16,41–44].

3.2 Occurrence and classification Since the description of the first flavoprotein structures [25,45,46] and subsequent burst of DNA recombinant technology, several bioinformatics tools have been described to discriminate flavoprotein subclasses. FAD and FMN binding domains, for instance, are excellent baits for mining nature’s flavoprotein arsenal [47]. Amino acid sequence motifs and DNA screening primers have been used to explore the occurrence of flavin monooxygenases in prokaryotic and eukaryotic strains.

3.2.1 Amino acid sequence motifs The first amino acid sequence fingerprints described for flavin monooxygenases map different patches of the flavin-binding domain in contact with the cofactor. The G box fingerprint [48] identifies the dinucleotide binding βαβ-fold [49], which binds the ADP moiety of FAD, whereas the GD fingerprint is formed by residues that are in contact with the riboflavin moiety of FAD [50]. The G box and GD fingerprint are common for one-component flavin monooxygenases (subclass A and B). Single-component aromatic hydroxylases (subclass A), Baeyer-Villiger type I monooxygenases (BVMOs; subclass B) and flavin-containing monooxygenases (FMO; subclass B) possess subclass-specific fingerprints (򐂰Tab. 3.2). Single-component aromatic hydroxylases can be pooled out by the DG fingerprint, a sequence with a dual function: indirect binding of the pyrophosphate moiety of FAD and recognition of the NAD(P)H cofactor [51]. Subclass B monooxygenases contain two G box fingerprints, and can be further discriminated by the presence of a BVMO/FMO-identifying sequence motif containing residues important for catalysis (򐂰Tab. 3.2) [52]. One amino acid substitutions in the

Tab. 3.2: Amino acid fingerprints specific for flavin monooxygenases. Fingerprint

Sequencea

Subclass

Function

Reference

G box

GXGXX(G/A)

AB

FAD (ADP)

[48]

GD

TXXXXIYIIGDA(A/V)H

A

FAD (riboflavin)

[50]

DG

chhhssDGXcSXhR

A

FAD/NAD(P)H

[51]

BVMO-1

FXGXXXHXXXW[P/D]

B

catalysis

[52]

FMO

FXGXXXHXXX(Y/F)

B

catalysis

[52]

BVMO-2

[A/G]GXWXXXX[F/Y]P[G/M]XXXD

B

catalysis NADP

[53]

W box

GWXWXPL

F

active site

[37]

a

c = charged residue, h = hydrophobic residue, s = small residue, X = any residue

3.2 Occurrence and classification

55

sequence motif allow a further functional assignation, since FMOs possess a Tyr or Phe instead of Pro or Glu (򐂰Tab. 3.2). Recently, Fraaije and coworkers [53] polished the subclass B fingerprint arsenal, by defining a reliable BVMO-specific sequence motif that maps the active site and contains important residues involved in catalysis and NADP+ interaction (򐂰Tab. 3.2). Subclass A and B fingerprints, together with G box and GD fingerprints, have been used to mine the protein database resulting in the retrieval and characterization of phenylacetone monooxygenase (PAMO) from the thermostable strain Thermobifida fusca [54]. More recently, 18 flavoprotein hydroxylases [55] and 22 type I BVMOs [53] were retrieved from the Rhodococcus jostii RHA1 genome [56]. Garcia and colleagues [57] described a consensus sequence for two-component flavoproteins based on multiple sequence alignment of both the oxygenase and reductase component, but this consensus sequence is rather difficult to use for bioinformatics purposes. Sequence alignments of flavoprotein halogenases (subclass F) pointed out the presence of a halogenase conserved sequence motif that contains aromatic residues located in the active site. Dong and coworkers hypothesized that this W box motif prevents halogenases from acting as hydroxylases [37] but site-directed mutagenesis on tryptophan 7-halogenase (PrnA) excluded the suggested role [58].

3.2.2 DNA screening Screening with PCR primers represents a useful method in the quest of new enzymes. Hornung and coworkers [59] developed a pair of primers specific for flavoprotein halogenases, tailoring enzymes in the biosynthesis of natural products with antimicrobial and antitumor activity. Primer sequences were based on optimization of degenerate primers deduced from highly conserved regions of six FAD-dependent halogenases specific for diverse phenol- and pyrrole-containing metabolites (򐂰Tab. 3.3). The screening Tab. 3.3: Primers used for flavin monooxygenase screening. Primer

Nucleotide sequence

Subclass

Amino acid sequence

Reference

fw-1

5′-CGSRGSTGGG SNCRNTGG-3′

E

RGWAQW

[60]

fw-2

5′-DCSGGNTT YGAYGAYCCB-3′

E

TGFDDP

[60]

rev-1

5′- NACNGWNGT RAANGWMTT-3′

E

NSFTSV

[60]

rev-2

5′-SAGSGGSGG RTCSAKSGA-3′

E

SMDPPL

[60]

Halo-B4-FW

5′-TTCCCSCGSTAC CASATCGGSGAG-3′

F

FPRXXIGES

[59]

Halo-B7-RV

3′-CCSACCAWGAC CMWSTAGGGSG-5′

F

GWXWXXPL

[59]

56

3 The flavin monooxygenases

of an actinomycete genome collection reported several putative flavin-dependent halogenases, and more than 40 novel genes unrelated to any other type of halogenase were retrieved. Halogenase specific primers represent a powerful tool for the rapid identification of natural product producer strains even from large strain collections. Recently, a pair of primers specific for styrene monooxygenases (SMO; subclass E) has been described [60]. SMOs are rare; they have been characterized in Pseudomonas species [61–65], in a metagenomic library from contaminated soil [66], and in Rhodococcus species [36,67]. StyA2B from R. opacus 1CP represents a recently discovered SMO fusion protein, containing the monooxygenase (StyA) and reductase (StyB) components in the same polypeptide chain. Tischler and coworkers developed two pairs of primers enabling the discrimination between fused and canonical two-component SMOs [60]. The primer combination amplifies the fusion part of genes encompassing the monooxygenase component and the reductase flavin domain. In this way, they found another one-component SMO in Rhodococcus opacus MR11 and established some phylogenetic correlation between one and two-component SMOs. From this it was hypothesized that one-component SMOs evolved from two-component ancestors through convergent evolution.

3.3 Single-component flavin monooxygenases 3.3.1 Subclass A Single-component flavoprotein hydroxylases comprise important enzymes in microbial degradation of aromatic compounds, antibiotic resistance and biosynthesis of polyketides [26]. The exquisite regioselective hydroxylation in ortho or para position of the aromatic ring has no counterpart in other enzymatic systems. Originally, research focused on enzymes involved in lignin degradation [14,21,68,69]. More recently, several new flavoprotein hydroxylases involved in polyketide biosynthesis have been described. Single-component flavoprotein hydroxylases are very regioselective and display a subtle mechanism of substrate, coenzyme and oxygen recognition to avoid wasteful consumption of NAD(P)H [14]. They typically act according to an electrophilic aromatic substitution mechanism in which the oxygen-flavin adduct (C4a-hydroperoxyflavin) is the electrophile and substrates containing hydroxyl or amino activating groups act as nucleophiles [18]. Different mechanisms have evolved to ensure sufficient nucleophilicity: para-hydroxybenzoate hydroxylase (PHBH) employs a hydrogen bond network for substrate deprotonation [21], while 2-methyl-3-hydroxypyridine-5carboxylic acid oxygenase (MHPCO) selectively binds the substrate in its trianionic form, ensuring enough electron density at the oxygenated carbon [70]. Flavoprotein hydroxylases convert many (hetero)aromatic substrates ranging from monophenols and uric acids to polyketide antibiotics and antitumor agents as tetracyclines, angucyclines and indolocarbazoles [26,71]. Enzymes acting on angucyclines and indolocarbazoles are phylogenetically related (򐂰Fig. 3.3). Small differences at the amino acid level are responsible of peculiar regio- or stereoselectivity as found for 7-carboxy-K252c hydroxylase (RebC, Q8KI25 and StaC, Q06IS1 [72,73]), or UW16 12-hydroxylase (PgaE/ CabE, Q9AMJ5/Q93LY7) [74]. Aklavinone 11-hydroxylase (RdmE, Q54530) [8] and

3.3 Single-component flavin monooxygenases

Q9S 3U9

Q84H F5

Q7 X2 81

Q 93 FC 6

s ne cli cy

V5

es iv at riv de

Q5 U9 13 Q93 LY7 Q9AM J5

B8 3F 9 Q

A Co

gu

Q01911

5

Q5 5K

24 P15

Q0

5Y T

P4

An

les

94 Q1

Q

P38169

Indo

57

U8 S3 Q9

M1 C0LT

Q02301 Q54530 F1CWE3 Afl2 060 P2

326

2

ca

lo

rb

Q0

s te nz oa be ox y

38 04

724

Q59

G9

P0

HW

G3

Hy

H2 QP F8

Q9

Q93N

Q988D3

P53318

0.05

K6

10 34 _0 81 17 _0 AG CP

AG CP

es

ol az

SF

dr

do

In

3 Q0H2X 25 I K Q8 1 6I S 2 Q0 I7 27 Q

Fig. 3.3: Flavoprotein hydroxylases subclass A phylogenetic overview. Representative sequences of single-component hydroxylases were aligned with CLUSTALW [131] and a neighbor-joining tree was built and edited in FigTree (http://tree.bio.ed.ac.uk). Sequences of enzymes with known crystal structure are represented in bold. Sequence ID numbers were retrieved from the Uniprot database (http://www.uniprot.org).

tetracycline hydroxylase sequences (TetX, Q01911) [75], are poorly related (򐂰Fig. 3.3), but their 3D-structures show a close relationship with RebC and PgaE/CabE. Enzymes acting on similar compounds do not always belong to the same clade. In fact, 3-hydroxybenzoate 4-hydroxylase (3HB4H, Q05KQ5) is related to phenol hydroxylase (PHHY, P15245) by sharing most active site residues and an extra thioredoxin domain [76]. PHBH instead, outgroups quite early from the tree, with no obvious phylogenetic link to any other sequence. 2-Hydroxybenzoate 1-hydroxylase (2HB1H, P23262), 3-hydroxybenzoate 6-hydroxylase (3HB6H, Q0SFK6 and CPAG_03410) [55] and 4-hydroxybenzoate 1-hydroxylase (4HB1H, CPAG_01781) [77,78] belong to the same clade, hampering their functional assignment on the basis of sequence comparison [55].

58

3 The flavin monooxygenases

Without knowing active site geometries and careful product analysis, prediction of regioselectivity in flavoprotein hydroxylases remains challenging. The phylogenetic tree of single-component flavoprotein hydroxylases also comprises a clearly separated cluster of enzymes that act on CoA-activated substrates (򐂰Fig. 3.3). This opens interesting questions about their substrate recognition and activation. 2-Aminobenzoyl-CoA monooxygenase/reductase (ACMR, Q93FC6, Q93FB8, Q02301) [79, 80], salicyl CoA 5-hydroxylase (SalCoA5H, Q7X281), [81] and SibG (C0LTM1) belong to this subgroup. SibG is a newly characterized flavin-dependent hydroxylase converting peptidyl carrier protein-bound 3-hydroxy-4-methylanthranilic acid to the fully substituted anthranilate moiety found in sibiromycin, an antitumor antibiotic [82]. CoA activation can be viewed as another strategy to increase substrate nucleophilicity and form substrate intermediates suitable for hydroxylation.

3.3.2 Subclass B Single-component subclass B enzymes perform Baeyer-Villiger oxidation, sulfoxidation and heteroatom hydroxylation reactions. Subclass B members are NADPH-specific, contain two Rossmann folds and keep the oxidized coenzyme bound during substrate oxygenation. Three major subgroups form subclass B: Baeyer-Villiger monooxygenases (BVMOs), N-hydroxylating flavoprotein monooxygenases (NMOs) and flavin containing monooxygenases (FMOs) (򐂰Fig. 3.4). BVMOs convert ketones (or aldehydes) into esters or lactones and are widely used for the preparation of enantiopure compounds [29]. Different from subclass A hydroxylases, the flavin-oxygen adduct is deprotonated and oxygen insertion goes via nucleophilic substitution [83]. Initial research on single-component BVMOs focused on cyclohexanone monooxygenase (CHMO, Q9R2F5) [83–85], 4-hydroxyacetophenone monooxygenase (HAPMO, Q93TJ5) [52,86] and phenylacetone monooxygenase (PAMO, Q47PU3) [28,54,87]. More recently, several new (thermostable) BVMOs were described [29] and the stereopreference and substrate acceptance of selected BVMOs was improved by directed evolution [88,89] and site-directed mutagenesis [90,91]. A new generation of selfsufficient BVMOs has also been reported [92]. In these systems, the BVMO is fused to a thermostable phosphite dehydrogenase for cofactor regeneration. Among subclass B enzymes, siderophore NMO enzymes are of emerging interest, since they exhibit peculiar kinetics and structural features [10]. NMOs catalyze the hydroxylation of the side chain amino-group of ornithine or lysine or the primary amino group of putrescine. Reaction products are then incorporated into siderophores, iron chelators in plants, bacteria and fungi. Exceptionally, the N-hydroxylating enzymes can form both neutral and anionic peroxyflavin as long-lived flavin intermediates [93,94]. Ornithine hydroxylases from Pseudomonas aeruginosa (PvdA, Q51548) and Aspergillus fumigatus (SidA, Q5SE95), and lysine hydroxylases from Escherichia coli (LucD, P11295) and Mycobacterium smegmatis G (MbsG) are well characterized NMOs. As a common feature, the substrate is not required for accelerating flavin reduction by NADPH [93,94]. The presence of NADP+ ensures closure of the active site for stabilizing the oxygenated flavin during catalysis [95]. Research on MbsG revealed the production of both hydrogen peroxide and superoxide in the absence of substrate, while in the presence

3.3 Single-component flavin monooxygenases

59

Plant FMO

95 SE Q5

K4

5

4

9 12 P1

88

66

Q4

0 474

820

O05

P31 512

Human FMO

O NM

B2FR

L2

Q5 154

8

Q9SZY8

3X

FE 9H

P3

O2 3 0 2 4 VQ1 Q9S

Q8

O FM

Q

tic yo r ka Eu

P4932 6 P31513

Q93TJ5

8 Q9951 0 174 Q0

Q00730 Q5V D85

3

22

A5

A1IHE6

Q0MRG6

0 2F5 Q9R

41

U3

GA W

7P

Q8

F6

Q4

O506

0.06

Q8 2IY 8 Q1 T7 B5

38

Q9

H9

N

6

6 P9

O BVM

Fig. 3.4: Flavoprotein monooxygenases subclass B phylogenetic overview. Representative sequences of single-component BVMOs, NMOs and FMOs were aligned with CLUSTALW [131] and a neighbor-joining was built and edited in FigTree (http://tree.bio.ed.ac.uk). Sequences of enzymes with known crystal structure are represented in bold. Sequence ID numbers were retrieved from the Uniprot database (http://www.uniprot.org).

of the true substrate byproduct formation is decreased. The high oxidase activity in the absence of substrate is rather unusual for flavin monooxygenases [96]. FMOs comprise eukaryotic and prokaryotic monooxygenases able to catalyze hydroxylation of heteroatoms such as nitrogen, sulfur, phosphorus, selenium, and iodine via formation of a hydroperoxyflavin intermediate. In humans, five FMO isoforms are present that show a tissue-specific distribution. Of special interest is FMO3 (P31513), in view of the genetically-linked disease known as “fish odor syndrome” [97,98]. Like cytochrome P450, FMOs mainly convert drugs and other xenobiotics into more hydrophilic metabolites. The genome of the plant model organism Arabidopsis thaliana contains around thirty FMO genes. As found for YUCCA 1-3 (Q9SZY8, Q9SVQ1, O23024), several of these genes are deputed to auxin biosynthesis. Other plant FMOs are predicted to be involved in the biosynthesis of glucosinolates and pathogen defense [11].

60

3 The flavin monooxygenases

BVMOs, NMOs and FMOs show a similar overall fold with separate Rossmann-fold domains for FAD and NADPH. Clear differences exist in the substrate binding domains: FMOs do not possess a clear substrate binding site, BVMOs present a substrate pocket derived from a helical insertion in the NADPH domain, and the substrate binding site of NMOs is derived from inserts in the FAD-binding and NADPH-binding domains (򐂰Fig. 3.2). Differences in FAD solvent exposure may underlie the inability of some NMOs to tightly bind FAD. Structural and kinetic data on BVMOs and FMOs suggest that the pyridine nucleotide coenzyme has multiple functions in the catalytic cycle adopting various positions in the active site. After flavin reduction, NADP+ remains bound to promote oxygen activation and stabilize the anionic flavin peroxide [87,91,99]. Phylogenetic analysis of subclass B monooxygenases underlines the distinct separation among the three subfamilies, and subclade distinction among the various BVMOs (򐂰Fig. 3.4). For instance, it is interesting to see that PAMO (Q47PU3) and steroid monooxygenase (STMO, O50641) are closely related and both accept phenylacetone as a substrate [91]. Subclass B monooxygenases are known to be strictly dependent on NADPH, but Jensen and coworkers [100] recently characterized a BVMO from Stenotrophomonas maltophilia (SMFMO, B2FLR2) which can use NADH equally well. SMFMO catalyzes the oxidation of thioethers and the regioselective Baeyer-Villiger oxidation of the model substrate bicyclo[3.2.0]hept-2-en-6-one. SMFMO is structurally similar to BVMOs (and thioredoxin reductase), but phylogenetically not clustered with other BVMOs (򐂰Fig. 3.4).

3.3.3 Subclass C Subclass C comprises FMN-dependent two-component monooxygenases with a TIMbarrel fold. Subclass C enzymes catalyze a diverse group of reactions: Baeyer-Villiger oxidation, hydroxylation of long-chain alkanes and nitriloacetate, oxidation and desulfurization of sulfonates, and oxidation of aldehydes coupled with generation of bioluminescence [39]. Alkanesulfonate monooxygenase SsuD-SsuE from Escherichia coli [101] has emerged as a model enzyme to unravel the mechanism of desulfurization. Catalysis proceeds via nucleophilic attack on the sulfur group, Baeyer-Villiger rearrangement and cleavage of the carbon-sulfur bond. Recent studies highlight the importance of Arg226 as potential active site acid in SsuD catalysis [102]. The crystal structure of long-chain alkane monooxygenase (LadA) from Geobacillus thermodenitrificans NG80-2 [103] moved further the catalytic potential of flavin monooxygenases. LadA catalyzes the terminal hydroxylation of long-chain n-alkanes (C15–C36). The structure of LadA reveals a large, deep pocket in front of the C-terminal entrance of the TIM barrel sufficiently large to fit the flavin, molecular oxygen, and terminal parts of a long chain n-alkane. Biochemical studies are required to elucidate the catalytic mechanism and disclose the role of the flavin and active site residues in oxygen activation and alkane hydroxylation. RutA-RutF is an FMN-dependent two-component monooxygenase that catalyzes the oxidative ring opening of uracil to 3-ureidoacrylic acid [104]. The reaction proceeds by addition of flavin hydroperoxide to the C4 carbonyl of uracil and involves

3.3 Single-component flavin monooxygenases

61

the intermediate formation of (Z)-3-ureidoacrylic peracid. The unprecedented peroxidecatalyzed amide hydrolysis of uracil illustrates the extraordinary catalytic versatility of the flavin cofactor.

3.3.4 Subclass D Subclass D two-component monooxygenases are FMN/FAD hydroxylases with an acylCoA dehydrogenase fold [16]. Most research on these enzymes has focused on the FMN-dependent 4-hydroxyphenylacetate 3-hydroxylase (HPAH) from Acinetobacter baumanii. HPAH turned out to be a good model for studying flavin transfer [105], oxygen diffusion [23], and substrate regulation [106–108]. The aromatic substrate 4-hydroxyphenylacetate regulates HPAH catalysis through binding to the reductase component. This binding increases the rate of flavin reduction by NADH. Steady-state and rapid kinetics studies showed that dissociation of FMNH– from the reductase component limits the rate of flavin transfer to the monooxygenase component. The direction of transfer is regulated by different affinities of the two protein components for FMN in its oxidized and reduced state. Deletion mutagenesis showed that the C-terminal domain of the reductase component is an autoinhibitory domain that upon binding of the 4-hydroxyphenylacetate effector undergoes conformational changes to allow faster flavin reduction and release [109]. Residues 179–230 of the C-terminal domain were shown to be responsible for repressing the production of reduced FMN in the absence of 4-hydroxyphenylacetate. Once reduced flavin is tightly bound to the monooxygenase component, the reaction with molecular oxygen is favored. Molecular dynamics studies computed multiple oxygen diffusion paths from the surface, where dioxygen molecules may temporarily reside, to the C4a atom of the flavin in the active site [23]. The simulations suggest that dioxygen visits several niches along different paths characterized by the presence of Ala and/or Ile hydrophobic residues. Only a few paths are “complete” and competent in guiding the oxygen molecule to the re-side of the flavin. Baron and colleagues also hypothesized the presence of a preorganized oxygen cavity in proximity of the cofactor for C4a intermediate stabilization. Hydroxylation of 4-hydroxyphenylacetate in HPAH is also tightly controlled. His120 and Ser146 seem to be the key residues, allowing activation and proper orientation of the substrate for efficient hydroxylation [107,108]. In summary, as in single-component hydroxylases, substrate acts as an effector in HPAH catalysis by suppressing hydrogen peroxide formation and preventing wasteful consumption of NADH [106]. Substrate effector regulation has also been observed in two-component nitriloacetate monooxygenase [110]. Other subclass D enzymes offer an interesting outlook on synthetic applications. Valton and coworkers characterized the FMN-dependent ActVA-ActVB system from Streptomyces coelicolor, involved in biosynthesis of the antibiotic actinorhodin [33]. ActVA catalyzes the hydroxylation of the hydroquinone form of dihydrokalafungin and is also active with related pyronaphthoquinones. Rapid reaction studies showed that binding of reduced FMN to ActVA precedes its reaction with oxygen and that the rate of formation of the FMN-C4a-hydroperoxide is independent of the presence of the aromatic substrate [111]. Another interesting feature of this two-protein system is that the rate of transfer of reduced FMN from ActVB to ActVA is controlled by the release of

62

3 The flavin monooxygenases

NAD+ from ActVB. In this way, the reductase component regulates the monooxygenase activity without the need for protein-protein interactions. Evidence is accumulating that certain cofactor-independent two-component monooxygenases involved in polyketide antibiotic biosynthesis use a similar mechanism of oxygen activation as that used by two-component flavoenzymes. For SnoaW/SnoaL2 involved in nogalamycin biosynthesis it was proposed that dioxygen is activated by the anthracycline substrate itself [112]. Oxygen activation requires the initial SnoaW-mediated NADPH-dependent reduction of anthracycline into a dihydroquinone. The reduced substrate might then react with dioxygen in a similar manner to flavin. Next, protonation of the oxygenated substrate intermediate by SnoAL2 would give the hydroxylated nogalamycin product.

3.3.5 Subclass E Subclass E flavoprotein monooxygenases are relatively rare [61,63,66]. These enzymes oxidize styrene derivatives to the corresponding epoxides and provide a highly enantioselective alternative to chemical epoxidation catalysts [113]. The monooxygenase component of the styrene converting system has many structural properties in common with subclass A aromatic hydroxylases [26,35]. Recently, the first self-sufficient singlecomponent styrene monooxygenase/reductase system was discovered in a R. opacus strain [36,67]. This monoooxygenase system shows good enantioselectivity but is less efficient in substrate oxygenation. Close structural homology between the styrene monooxygenase component from Pseudomonas putida S12 and PHBH provides insight into the putative FAD and substrate binding site [35]. This information was used to design styrene monooxygenase variants with improved epoxide rates [114], different substrate profiles [115], and inversed enantiomeric preference [116].

3.3.6 Subclass F Subclass F flavoprotein monooxygenases catalyze the regioselective chlorination and bromination of activated organic molecules [117,118]. These enzymes are structurally related to subclass A aromatic hydroxylases by sharing the characteristic FADbinding scaffold of the glutathione reductase superfamily [119]. Flavin halogenases are of interest for the production of antibiotics, antitumor agents and other natural products [120]. The 3D structure of tryptophan 7-halogenase (PrnA) suggests a catalytic mechanism involving the formation of hypochlorous acid, which is guided to the substrate-binding site for the regioselective halogenation of tryptophan [37]. More recent work suggests that the chloride addition reaction requires the critical involvement of a Lys and a Glu residue [58] and provides interesting information about the structural determinants for regioselectivity control [121]. Spectroscopic studies identified a C4a-hydroperoxyflavin intermediate, which reacts with chloride anion to produce hypochlorous acid [122]. Flavin halogenases can act on tryptophan with different regioselectivity. Usually, this regioselectivity correlates with the phylogenetic clustering of the halogenases (򐂰Fig. 3.5): 4-halogenation (MibH, E21HC5), 5-halogenation (PyrH, A4DOH5), 6-halogenation (STTH, E9PI62; KtzR, A8CF74) and 7-halogenation (PrnA, P9548O; RebH,

UJ

E9

5

n pha e pto nas Try aloge 2-h

1 LIP

I1

B9Z

63

E1 B3 B2D FW 7 T7

Q8KND5

GM

83 3N Q9 76 76 O8

Q8

A7KH29

3.3 Single-component flavin monooxygenases

Z69

Q0V

Q9AL91

Q8KUG0

Q4KCZ0

E9PI62 A8C F74

2

P95480

5 HC E2I

W 3K Q9

A8CF75

0.06

Tryptophan 4-halogenase

80 E2 A1 HZ8 Q8K

0

A4

D0

H5

T 5- ryp ha to lo ph ge an na se

48

5 P9

Tryptoph 6-haloge an nase

FI4

Q54

phan Trypto genase lo a 7-h

Fig. 3.5: Flavoprotein halogenases subclass F phylogenetic overview. Representative sequences of flavoprotein halogenases were aligned with CLUSTALW [131] and a neighbor-joining tree was built and edited in FigTree (http://tree.bio.ed.ac.uk). Sequences of known crystal structures are represented in bold. Sequence ID numbers were retrieved from the Uniprot database (http://www.uniprot.org).

Q8KHZ8; KtzQ, A8CF75). Lang et al. addressed the structural basis of the regioselectivity of PrnA and PyrH [123]. It appears that His101 and Trp455 dictate the regioselectivity in PrnA while Tyr444 and Arg459 prevent the tryptophan substrate from adopting the same orientation as that seen in PyrH. Directed mutagenesis studies, supported by previous data on the role of residues in binding and catalysis [37,58], permitted the production of a mixture of halogenated compounds with the expected regioselectivity. Many flavin halogenases react with substrates that are bound via a thioester linkage to an acyl carrier protein. CmdE (Q0VZ69) catalyzes the 2-halogenation of carrierbound tryptophan for the biosynthesis of chondramide [124]. 򐂰Fig 3.5 shows that CmdE is closely related to tiacumicin halogenase (TiaM, E9LIP1). The latter enzyme catalyzes

64

3 The flavin monooxygenases

two sequential halogenation steps in the biosynthesis of tiacumicin B, a broad-spectrum antibiotic against various Gram-positive pathogenic bacteria [125]. Similar sequential dichlorination reactions on carrier-bound substrates are catalyzed by pyoluteorin halogenase (PltA, Q4KCZ0) [126] and (2,4,6-trihydroxyphenyl)-1-hexan-1-one halogenase (ChlA, Q54F14) [127]. ChlA is involved in the biosynthesis of differentiation-inducing factor 1 (DIF-1), a polyketide-derived morphogen that drives stalk cell formation in the developmental cycle of the social amoeba Dictyostelium discoideum. In order to characterize the enzymatic activity of ChlA, the enzyme was homologously expressed in Dictyostelium and the reduced flavin cofactor was provided by in situ reduction of FAD by flavin reductase SsuE from E. coli. SgcC3 from Streptomyces globisporus catalyzes the regiospecific chlorination of peptidyl-carrier-protein bound tyrosine during the biosynthesis of the potent antitumor antibiotic C-1027 [128]. A similar reaction is catalyzed by chondrochloren halogenase from Chondromyces crocatus (CndH, B9ZUJ5) [119]. The CdnH structure has a more open halogenation site than PrnA and RebH, presumably needed to accommodate the tyrosyl-protein substrate. Chloramphenicol contains an unusual dichloroacetyl moiety that is critical for its antibiotic activity. The flavin halogenase CmlS (Q9AL91) is responsible for the introduction of the dichloroacetyl group. As can be deduced from 򐂰Fig 3.5, CmlS is closely related to CndH. However, the CmlS structure shows a T-shaped tunnel leading to the active site that is blocked by the C-terminal tail of the protein [129]. Therefore, it was argued that CmlS might halogenate the free acyl group directly, or as the corresponding CoA thioester. Intriguingly, the 8α-carbon of the FAD cofactor of CmlS is covalently bound via an ester linkage to Asp277. This new type of posttranslational flavin modification discriminates CmlS from CndH and other characterized flavin halogenases. Because reduced flavin is normally delivered by a flavin reductase, the question arises how the covalently bound flavin of CmlS receives its reducing equivalents. Recently, it was reported that flavin halogenases can catalyze bipyrrole homocoupling reactions by forming C-N bonds [130]. Based on earlier biosynthesis studies of pyoluteorin and pyrrolomycin, a putative marinopyrrole biosynthetic gene cluster (mpy) from Streptomyces sp. CNQ-418 was cloned and sequenced, suggesting that monodeoxypyoluteorin is biosynthesized from L-proline and three malonate molecules via a modular polyketide synthase. Heterologous expression of the entire mpy cluster in Streptomyces coelicolor M512 resulted in the production of four dimeric marinopyrrole metabolites and the identification of three functional flavin halogenases and a pathwayspecific flavin reductase. One of the halogenases is responsible for the initial dichlorination of the pyrrolyl-S-carrier protein as previously demonstrated for the homologous PltA in pyoluteorin biosynthesis [126], while the other two halogenases are essentially involved in the atropo-selective N,C-bipyrrole homocoupling reaction.

3.4 Conclusions Flavin monooxygenases catalyze hydroxylation, Baeyer-Villiger oxidation, sulfoxidation, epoxidation or halogenation reactions by generating a reactive (hydro)peroxyflavin that facilitates splitting of the oxygen-oxygen bond. During the past few years, many

3.5 References

65

new family members and several unprecedented flavin monooxygenase reactions have emerged. This review illustrates selected features, tools to retrieve novel flavin monooxygenases from genome mining, and new findings on each of the six subclasses. Key mechanistic points related to flavin oxygen reactivity have been extensively studied in different enzymatic systems and new insights into structural features that determine the formation and stabilization of the (hydro)peroxyflavin have been obtained. Active site characteristics that differentiate oxygen-utilizing flavin monooxygenases from oxygen-consuming flavin oxidases have been summarized in a recent review [20]. Considering the large number of newly discovered flavin monooxygenases, it is clear that remaining questions about oxygen diffusion, oxygen activation and oxygen reactivity need to be assessed with more enzymes. This will provide a general picture and may offer a better explanation of the peculiarities observed with individual enzymes. One of the major unresolved questions of subclass A enzymes is their interaction with NAD(P)H. Studies with PHBH indicated that there is no space for the reduced coenzyme in the active site and that the isoalloxazine and nicotinamide moieties of the cofactors encounter each other in a more solvent exposed region [14,21,26,69]. Currently, it is far from clear how other single-component hydroxylases solve this mobility issue and how the aromatic substrates of these enzymes stimulate the rate of flavin reduction by NAD(P)H. Other longstanding questions for subclass A enzymes remain: how do flavin hydroxylases prevent uncoupling of hydroxylation and can we rationally redesign their substrate specificity? The coenzyme specificity of subclass A enzymes is another issue that needs attention. Subclass A enzymes lack a Rossmann fold for binding NAD(P)H and as a result they weakly interact with NAD(P)+. Subclass B enzymes do contain a Rossmann fold for NADP(H) binding and require the presence of NADP+ during substrate oxidation [20]. Binding of NADP+ defines the shape and geometry of the catalytic center and dramatically stabilizes the (hydro)peroxyflavin. A major challenge with these enantioselective enzymes is the development of bioinformatics tools that reliably predict their substrate specificities. Questions formulated for the oxygen reactivity of one-component enzymes can also be formulated for two-component flavin monooxygenases (subclass C–F). The latter enzymes use reduced flavin as a substrate rather than as a prosthetic group. Unresolved questions regarding subclass C–F enzymes deal with the mechanism of action (luciferases), the unprecedented reactivity (n-alkane hydroxylation), the mode of substrate recognition (CoA-activated and peptidyl-carrier bound substrates), the mode of flavin binding (CmlS), and the interaction with the redox partner (styrene monooxygenase). Collectively, it is clear that flavin monooxygenases will keep us busy to solve their intriguing properties and relationships.

3.5 References [1] Torres Pazmino DE, Winkler M, Glieder A, Fraaije MW. Monooxygenases as biocatalysts: classification, mechanistic aspects and biotechnological applications. J Biotechnol 2010;146:9–24. [2] Fetzner S, Steiner RA. Cofactor-independent oxidases and oxygenases. Appl Microbiol Biotechnol 2010;86:791–804. [3] Fisher AJ, Raushel FM, Baldwin TO, Rayment I. Three-dimensional structure of bacterial luciferase from Vibrio harveyi at 2.4 A resolution. Biochemistry 1995;34:6581–6.

66

3 The flavin monooxygenases

[4] Harwood CS, Parales RE. The beta-ketoadipate pathway and the biology of self-identity. Ann Rev Microbiol 1996;50:553–90. [5] Phale PS, Basu A, Majhi PD, Deveryshetty J, Vamsee-Krishna C, Shrivastava R. Metabolic diversity in bacterial degradation of aromatic compounds. Omics 2007;11:252–79. [6] Ahmad M, Roberts JN, Hardiman EM, Singh R, Eltis LD, Bugg TDH. Identification of DypB from Rhodococcus jostii RHA1 as a lignin peroxidase. Biochemistry 2011;50:5096–107. [7] Kallio P, Liu Z, Mäntsälä P, Niemi J, Metsä-Ketelä M. Sequential action of two flavoenzymes, PgaE and PgaM, in angucycline biosynthesis: chemoenzymatic synthesis of gaudimycin C. Chem Biol 2008;15:157–66. [8] Lindqvist Y, Koskiniemi H, Jansson A, Sandalova T, Schnell R, Liu Z, Mäntsälä P, Niemi J, Schneider G. Structural basis for substrate recognition and specificity in aklavinone-11hydroxylase from rhodomycin biosynthesis. J Mol Biol 2009;393:966–77. [9] Ryan KS, Howard-Jones AR, Hamill MJ, Elliott SJ, Walsh CT, Drennan CL. Crystallographic trapping in the rebeccamycin biosynthetic enzyme RebC. Proc Natl Acad Sci USA 2007;104:15311–6. [10] Olucha J, Lamb AL. Mechanistic and structural studies of the N-hydroxylating flavoprotein monooxygenases. Bioorg Chem 2011;39:171–7. [11] Schlaich NL. Flavin-containing monooxygenases in plants: looking beyond detox. Trends Plant Sci 2007;12:412–8. [12] Siebold C, Berrow N, Walter TS, Owens RJ, Stuart DI, Terman JR, Kolodkin AL, Pasterkamp RJ, Jones EY. High-resolution structure of the catalytic region of MICAL (molecule interacting with CasL), a multidomain flavoenzyme-signaling molecule. Proc Natl Acad Sci USA 2005;102:16836–41. [13] Hung RJ, Yazdani U, Yoon J, Wu H, Yang T, Gupta N, Huang Z, van Berkel WJ, Terman JR. Mical links semaphorins to F-actin disassembly. Nature 2010;463:823–7. [14] Palfey BA, McDonald CA. Control of catalysis in flavin-dependent monooxygenases. Arch Biochem Biophys 2010;493:26–36. [15] Vervoort J, Muller F, Lee J, Vandenberg WAM, Moonen CTW. Identifications of the True C-13 Nuclear-Magnetic-Resonance Spectrum of the Stable Intermediate-Ii in Bacterial Luciferase. Biochemistry 1986;25:8062–7. [16] van Berkel WJH, Kamerbeek NM, Fraaije MW. Flavoprotein monooxygenases, a diverse class of oxidative biocatalysts. J Biotechnol 2006;124:670–89. [17] Ghisla S, Massey V. Mechanisms of flavoprotein-catalyzed reactions. Eur J Biochem 1989;181:1–17. [18] Massey V. Activation of molecular oxygen by flavins and flavoproteins. J Biol Chem 1994;269:22459–62. [19] Baron R, McCammon JA, Mattevi A. The oxygen-binding vs. oxygen-consuming paradigm in biocatalysis: structural biology and biomolecular simulation. Curr Opin Struct Biol 2009;19:672–9. [20] Chaiyen P, Fraaije MW, Mattevi A. The enigmatic reaction of flavins with oxygen. Trends Biochem Sci 2012;37:373–80. [21] Entsch B, van Berkel WJH. Structure and mechanism of para-hydroxybenzoate hydroxylase. FASEB J 1995;9:476–83. [22] Chen L, Lyubimov AY, Brammer L, Vrielink A, Sampson NS. The binding and release of oxygen and hydrogen peroxide are directed by a hydrophobic tunnel in cholesterol oxidase. Biochemistry 2008;47:5368–77. [23] Baron R, Riley C, Chenprakhon P, Thotsaporn K, Winter RT, Alfieri A, Forneris F, van Berkel WJH, Chaiyen P, Fraaije MW, Mattevi A, McCammon JA. Multiple pathways guide oxygen diffusion into flavoenzyme active sites. Proc Natl Acad Sci USA 2009;106:10603–8. [24] Kommoju P-R, Chen Z-w, Bruckner RC, Mathews FS, Jorns MS. Probing oxygen activation sites in two flavoprotein oxidases using chloride as an oxygen surrogate. Biochemistry 2011;50:5521–34.

3.5 References

67

[25] Wierenga RK, de Jong RJ, Kalk KH, Hol WG, Drenth J. Crystal structure of p-hydroxybenzoate hydroxylase. J Mol Biol 1979;131:55–73. [26] Montersino S, Tischler D, Gassner GT, van Berkel WJH. Catalytic and structural features of flavoprotein hydroxylases and epoxidases. Adv Synth Catal 2011;353:2301–19. [27] Chaiyen P. Flavoenzymes catalyzing oxidative aromatic ring-cleavage reactions. Arch Biochem Biophys 2010;493:62–70. [28] Malito E, Alfieri A, Fraaije MW, Mattevi A. Crystal structure of a Baeyer-Villiger monooxygenase. Proc Natl Acad Sci USA 2004;101:13157–62. [29] Torres Pazmiño DE, Dudek HM, Fraaije MW. Baeyer-Villiger monooxygenases: recent advances and future challenges. Curr Opin Chem Biol 2010;14:138–44. [30] Walsh CT, Chen YCJ. Enzymic Baeyer-Villiger oxidations by flavin-dependent monooxygenases. Angew Chem 1988;27:333–43. [31] Taylor DG, Trudgill PW. Camphor revisited: studies of 2,5-diketocamphane 1,2-monooxygenase from Pseudomonas putida ATCC 17453. J Bacteriol 1986;165:489–97. [32] Eichhorn E, Davey CA, Sargent DF, Leisinger T, Richmond TJ. Crystal structure of Escherichia coli alkanesulfonate monooxygenase SsuD. J Mol Biol 2002;324:457–68. [33] Valton J, Fontecave M, Douki T, Kendrew SG, Nivière V. An aromatic hydroxylation reaction catalyzed by a two-component FMN-dependent monooxygenase. The ActVA-ActVB system from Streptomyces coelicolor. J Biol Chem 2006;281:27–35. [34] Chaiyen P, Suadee C, Wilairat P. A novel two-protein component flavoprotein hydroxylase. Eur J Biochem 2001;268:5550–61. [35] Ukaegbu UE, Kantz A, Beaton M, Gassner GT, Rosenzweig AC. Structure and ligand binding properties of the epoxidase component of styrene monooxygenase. Biochemistry 2010;49:1678–88. [36] Tischler D, Eulberg D, Lakner S, Kaschabek SR, van Berkel WJH, Schlömann M. Identification of a novel self-sufficient styrene monooxygenase from Rhodococcus opacus 1CP. J Bacteriol 2009;191:4996–5009. [37] Dong C, Flecks S, Unversucht S, Haupt C, van Pée K-H, Naismith JH. Tryptophan 7-halogenase (PrnA) structure suggests a mechanism for regioselective chlorination. Science 2005;309:2216–9. [38] Jeffers CE, Nichols JC, Tu SC. Complex formation between Vibrio harveyi luciferase and monomeric NADPH:FMN oxidoreductase. Biochemistry 2003;42:529–34. [39] Ellis HR. The FMN-dependent two-component monooxygenase systems. Arch Biochem Biophys 2010;497:1–12. [40] Valton J, Filisetti L, Fontecave M, Niviere V. A two-component flavin-dependent monooxygenase involved in actinorhodin biosynthesis in Streptomyces coelicolor. J Biol Chem 2004;279:44362–9. [41] Chung HW, Tu SC. Structure-function relationship of Vibrio harveyi NADPH-flavin oxidoreductase FRP: essential residues Lys167 and Arg15 for NADPH binding. Biochemistry 2012;51:4880–7. [42] Kirchner U, Westphal AH, Muller R, van Berkel WJ. Phenol hydroxylase from Bacillus thermoglucosidasius A7, a two-protein component monooxygenase with a dual role for FAD. J Biol Chem 2003;278:47545–53. [43] van den Heuvel RH, Westphal AH, Heck AJ, Walsh MA, Rovida S, van Berkel WJ, Mattevi A. Structural studies on flavin reductase PheA2 reveal binding of NAD in an unusual folded conformation and support novel mechanism of action. J Biol Chem 2004;279:12860–7. [44] van Lanen SG, Lin S, Horsman GP, Shen B. Characterization of SgcE6, the flavin reductase component supporting FAD-dependent halogenation and hydroxylation in the biosynthesis of the enediyne antitumor antibiotic C-1027. FEMS Microbiol Lett 2009;300:237–41. [45] Ludwig ML, Andersen RD, Mayhew SG, Massey V. The structure of a clostridial flavodoxin. I. Crystallographic characterization of the oxidized and semiquinone forms. J Biol Chem 1969;244:6047–8.

68

3 The flavin monooxygenases

[46] Schulz GE, Schirmer RH, Sachsenheimer W, Pai EF. The structure of the flavoenzyme glutathione reductase. Nature 1978;273:120–4. [47] Macheroux P, Kappes B, Ealick SE. Flavogenomics--a genomic and structural view of flavindependent proteins. FEBS J 2011;278:2625–34. [48] Wierenga RK, Terpstra P, Hol WG. Prediction of the occurrence of the ADP-binding beta alpha beta-fold in proteins, using an amino acid sequence fingerprint. J Mol Biol 1986;187:101–7. [49] Rossmann MG, Moras D, Olsen KW. Chemical and biological evolution of nucleotidebinding protein. Nature 1974;250:194–9. [50] Eggink G, Engel H, Vriend G, Terpstra P, Witholt B. Rubredoxin reductase of Pseudomonas oleovorans. Structural relationship to other flavoprotein oxidoreductases based on one NAD and two FAD fingerprints. J Mol Biol 1990;212:135–42. [51] Eppink MHM, van Berkel WJH, Schreuder HA. Identification of a novel conserved sequence motif in flavoprotein hydroxylases with a putative dual function in FAD/NAD(P)H binding. Protein Sci 1997;6:2454–8. [52] Fraaije MW, Kamerbeek NM, van Berkel WJH, Janssen DB. Identification of a Baeyer-Villiger monooxygenase sequence motif. FEBS Lett 2002;518:43–7. [53] Riebel A, Dudek HM, de Gonzalo G, Stepniak P, Rychlewski L, Fraaije MW. Expanding the set of rhodococcal Baeyer-Villiger monooxygenases by high-throughput cloning, expression and substrate screening. Appl Microbiol Biotechnol 2012;95:1479–89. [54] Fraaije MW, Wu J, Heuts DP, van Hellemond EW, Spelberg JH, Janssen DB. Discovery of a thermostable Baeyer-Villiger monooxygenase by genome mining. Appl Microbiol Biotechnol 2005;66:393–400. [55] Montersino S, van Berkel WJH. Functional annotation and characterization of 3-hydroxybenzoate 6-hydroxylase from Rhodococcus jostii RHA1. Biochim Biophys Acta 2012;1824:433–42. [56] McLeod MP, Warren RL, Hsiao WWL, Araki N, Myhre M, Fernandes C, Miyazawa D, Wong W, Lillquist AL, Wang D, Dosanjh M, Hara H, Petrescu A, Morin RD, Yang G, Stott JM, Schein JE, Shin H, Smailus D, Siddiqui AS, Marra MA, Jones SJM, Holt R, Brinkman FSL, Miyauchi K, Fukuda M, Davies JE, Mohn WW, Eltis LD. The complete genome of Rhodococcus sp. RHA1 provides insights into a catabolic powerhouse. Proc Natl Acad Sci USA 2006;103:15582–7. [57] Galán B, Diaz E, Prieto MA, Garcia JL. Functional analysis of the small component of the 4-hydroxyphenylacetate 3-monooxygenase of Escherichia coli W: a prototype of a new flavin: NAD(P)H reductase subfamily. J Bacteriol 2000;182:627–36. [58] Flecks S, Patallo EP, Zhu X, Ernyei AJ, Seifert G, Schneider A, Dong C, Naismith JH, van Pée K-H. New insights into the mechanism of enzymatic chlorination of tryptophan. Angew Chem 2008;47:9533–6. [59] Hornung A, Bertazzo M, Dziarnowski A, Schneider K, Welzel K, Wohlert S-E, Holzenkämpfer M, Nicholson GJ, Bechthold A, Süssmuth RD, Vente A, Pelzer S. A genomic screening approach to the structure-guided identification of drug candidates from natural sources. ChemBioChem 2007;8:757–66. [60] Tischler D, Gröning JAD, Kaschabek SR, Schlömann M. One-component styrene monooxygenases: an evolutionary view on a rare class of flavoproteins. Appl Biochem Biotechnol 2012;167:931–44. [61] Lin H, Qiao J, Liu Y, Wu ZL. Styrene monooxygenase from Pseudomonas sp. LQ26 catalyzes the asymmetric epoxidation of both conjugated and unconjugated alkenes. J Mol Catal B: Enzym 2010;67:236–41. [62] Marconi AM, Beltrametti F, Bestetti G, Solinas F, Ruzzi M, Galli E, Zennaro E. Cloning and characterization of styrene catabolism genes from Pseudomonas fluorescens ST. App Environ Microbiol 1996;62:121–7. [63] Mooney A, Ward PG, O’Connor KE. Microbial degradation of styrene: biochemistry, molecular genetics, and perspectives for biotechnological applications. Appl Microbiol Biotechnol 2006;72:1–10.

3.5 References

69

[64] Panke S, Wubbolts MG, Schmid A, Witholt B. Production of enantiopure styrene oxide by recombinant Escherichia coli synthesizing a two-component styrene monooxygenase. Biotechnol Bioeng 2000;69:91–100. [65] Park MS, Bae JW, Han JH, Lee EY, Lee SG, Park S. Characterization of styrene catabolic genes of Pseudomonas putida SN1 and construction of a recombinant Escherichia coli containing styrene monooxygenase gene for the production of (S)-styrene oxide. J Microbiol Biotechnol 2006;16:1032–40. [66] van Hellemond EW, Janssen DB, Fraaije MW. Discovery of a novel styrene monooxygenase originating from the metagenome. App Environ Microbiol 2007;73:5832–9. [67] Tischler D, Kermer R, Gröning JAD, Kaschabek SR, van Berkel WJH, Schlömann M. StyA1 and StyA2B from Rhodococcus opacus 1CP: a multifunctional styrene monooxygenase system. J Bacteriol 2010;192:5220–7. [68] van Berkel WJH, Müller F. Flavin-dependent monooxygenases with special reference to phydroxybenzoate hydroxylase. In: Müller F, ed. Chemistry and Biochemistry of Flavoenzymes, Vol II. Boca Raton: CRC Press USA; 1991:1–29. [69] Entsch B, Cole LJ, Ballou DP. Protein dynamics and electrostatics in the function of p-hydroxybenzoate hydroxylase. Arch Biochem Biophys 2005;433:297–311. [70] Chaiyen P, Brissette P, Ballou DP, Massey V. Thermodynamics and reduction kinetics properties of 2-methyl-3-hydroxypyridine-5-carboxylic acid oxygenase. Biochemistry 1997;36:2612–21. [71] Mohanty SK, Yu CL, Das S, Louie TM, Gakhar L, Subramanian M. Delineation of the caffeine C-8 oxidation pathway in Pseudomonas sp. strain CBB1 via characterization of a new trimethyluric acid monooxygenase and genes involved in trimethyluric acid metabolism. J Bacteriol 2012;194:3872–82. [72] Howard-Jones AR, Walsh CT. Staurosporine and rebeccamycin aglycones are assembled by the oxidative action of StaP, StaC, and RebC on chromopyrrolic acid. J Am Chem Soc 2006;128:12289–98. [73] Asamizu S, Shiro Y, Igarashi Y, Nagano S, Onaka H. Characterization and functional modification of StaC and RebC, which are involved in the pyrrole oxidation of indolocarbazole biosynthesis. Biosci Biotechnol Biochem 2011;75:2184–93. [74] Koskiniemi H, Metsä-Ketelä M, Dobritzsch D, Kallio P, Korhonen H, Mäntsälä P, Schneider G, Niemi J. Crystal structures of two aromatic hydroxylases involved in the early tailoring steps of angucycline biosynthesis. J Mol Biol 2007;372:633–48. [75] Volkers G, Palm GJ, Weiss MS, Wright GD, Hinrichs W. Structural basis for a new tetracycline resistance mechanism relying on the TetX monooxygenase. FEBS Lett 2011;585:1061–6. [76] Hiromoto T, Fujiwara S, Hosokawa K, Yamaguchi H. Crystal structure of 3-hydroxybenzoate hydroxylase from Comamonas testosteroni has a large tunnel for substrate and oxygen access to the active site. J Mol Biol 2006;364:878–96. [77] Eppink MHM, Boeren SA, Vervoort J, van Berkel WJH. Purification and properties of 4-hydroxybenzoate 1-hydroxylase (decarboxylating), a novel flavin adenine dinucleotide-dependent monooxygenase from Candida parapsilosis CBS604. J Bacteriol 1997;179:6680–7. [78] Holesova Z, Jakubkova M, Zavadiakova I, Zeman I, Tomaska L, Nosek J. Gentisate and 3-oxoadipate pathways in the yeast Candida parapsilosis: identification and functional analysis of the genes coding for 3-hydroxybenzoate 6-hydroxylase and 4-hydroxybenzoate 1-hydroxylase. Microbiology 2011;157:2152–63. [79] Hartmann S, Hultschig C, Eisenreich W, Fuchs G, Bacher A, Ghisla S. NIH shift in flavindependent monooxygenation: mechanistic studies with 2-aminobenzoyl-CoA monooxygenase/ reductase. Proc Natl Acad Sci USA 1999;96:7831–6. [80] Schühle K, Jahn M, Ghisla S, Fuchs G. Two similar gene clusters coding for enzymes of a new type of aerobic 2-aminobenzoate (anthranilate) metabolism in the bacterium Azoarcus evansii. J Bacteriol 2001;183:5268–78. [81] Ishiyama D, Vujaklija D, Davies J. Novel pathway of salicylate degradation by Streptomyces sp. strain WA46. App Environ Microbiol 2004;70:1297–306.

70

3 The flavin monooxygenases

[82] Giessen TW, Kraas FI, Marahiel MA. A four-enzyme pathway for 3,5-dihydroxy-4-methylanthranilic acid formation and incorporation into the antitumor antibiotic sibiromycin. Biochemistry 2011;50:5680–92. [83] Ryerson CC, Ballou DP, Walsh C. Mechanistic studies on cyclohexanone oxygenase. Biochemistry 1982;21:2644–55. [84] Sheng D, Ballou DP, Massey V. Mechanistic studies of cyclohexanone monooxygenase: chemical properties of intermediates involved in catalysis. Biochemistry 2001;40:11156–67. [85] Mirza IA, Yachnin BJ, Wang S, Grosse S, Bergeron Hln, Imura A, Iwaki H, Hasegawa Y, Lau PCK, Berghuis AM. Crystal structures of cyclohexanone monooxygenase reveal complex domain movements and a sliding cofactor. J Am Chem Soc 2009;131:8848–54. [86] Kamerbeek NM, Moonen MJ, van der Ven JG, van Berkel WJH, Fraaije MW, Janssen DB. 4-Hydroxyacetophenone monooxygenase from Pseudomonas fluorescens ACB: a novel flavoprotein catalyzing Baeyer-Villiger oxidation of aromatic compounds. Eur J Biochem 2001;268:2547–57. [87] Orru R, Dudek HM, Martinoli C, Torres Pazmiño DE, Royant A, Weik M, Fraaije MW, Mattevi A. Snapshots of enzymatic Baeyer-Villiger catalysis: oxygen activation and intermediate stabilization. J Biol Chem 2011;286:29284–91. [88] Reetz MT, Wu S. Laboratory evolution of robust and enantioselective Baeyer-Villiger monooxygenases for asymmetric catalysis. J Am Chem Soc 2009;131:15424–32. [89] Wu S, Acevedo JP, Reetz MT. Induced allostery in the directed evolution of an enantioselective Baeyer-Villiger monooxygenase. Proc Natl Acad Sci USA 2010;107:2775–80. [90] Dudek HM, de Gonzalo G, Pazmiño DET, Stepniak P, Wyrwicz LS, Rychlewski L, Fraaije MW. Mapping the substrate binding site of phenylacetone monooxygenase from Thermobifida fusca by mutational analysis. App Environ Microbiol 2011;77:5730–8. [91] Franceschini S, van Beek HL, Pennetta A, Martinoli C, Fraaije MW, Mattevi A. Exploring the structural basis of substrate preferences in Baeyer-Villiger monooxygenases: insight from steroid monooxygenase. J Biol Chem 2012;287:22626–34. [92] Torres Pazmiño DE, Riebel A, de Lange J, Rudroff F, Mihovilovic MD, Fraaije MW. Efficient biooxidations catalyzed by a new generation of self-sufficient Baeyer-Villiger monooxygenases. ChemBioChem 2009;10:2595–8. [93] Meneely KM, Barr EW, Bollinger JM, Lamb AL. Kinetic mechanism of ornithine hydroxylase (PvdA) from Pseudomonas aeruginosa: substrate triggering of O2 addition but not flavin reduction. Biochemistry 2009;48:4371–6. [94] Mayfield JA, Frederick RE, Streit BR, Wencewicz TA, Ballou DP, DuBois JL. Comprehensive spectroscopic, steady state, and transient kinetic studies of a representative siderophoreassociated flavin monooxygenase. J Biol Chem 2010;285:30375–88. [95] Olucha J, Meneely KM, Chilton AS, Lamb AL. Two structures of an N-hydroxylating flavoprotein monooxygenase: ornithine hydroxylase from Pseudomonas aeruginosa. J Biol Chem 2011;286:31789–98. [96] Robinson R, Sobrado P. Substrate binding modulates the activity of Mycobacterium smegmatis G, a flavin-dependent monooxygenase involved in the biosynthesis of hydroxamatecontaining siderophores. Biochemistry 2011;50:8489–96. [97] Al-Waiz M, Ayesh R, Mitchell SC, Idle JR, Smith RL. Trimethylaminuria (fish-odour syndrome): an inborn error of oxidative metabolism. Lancet 1987;1:634–5. [98] Dolphin CT, Janmohamed A, Smith RL, Shephard EA, Phillips IR. Missense mutation in flavin-containing mono-oxygenase 3 gene, FMO3, underlies fish-odour syndrome. Nat Genet 1997;17:491–4. [99] Orru R, Pazmino DET, Fraaije MW, Mattevi A. Joint functions of protein residues and NADP(H) in oxygen activation by flavin-containing monooxygenase. J Biol Chem 2010;285:35021–8. [100] Jensen CN, Cartwright J, Ward J, Hart S, Turkenburg JP, Ali ST, Allen MJ, Grogan G. A flavoprotein monooxygenase that catalyses a Baeyer-Villiger reaction and thioether oxidation using NADH as the nicotinamide cofactor. ChemBioChem 2012;13:872–8.

3.5 References

71

[101] Eichhorn E, van der Ploeg JR, Leisinger T. Characterization of a two-component alkanesulfonate monooxygenase from Escherichia coli. J Biol Chem 1999;274:26639–46. [102] Ellis HR. Mechanism for sulfur acquisition by the alkanesulfonate monooxygenase system. Bioorg Chem 2011;39:178–84. [103] Li L, Liu X, Yang W, Xu F, Wang W, Feng L, Bartlam M, Wang L, Rao Z. Crystal structure of long-chain alkane monooxygenase (LadA) in complex with coenzyme FMN: unveiling the long-chain alkane hydroxylase. J Mol Biol 2008;376:453–65. [104] Mukherjee T, Zhang Y, Abdelwahed S, Ealick SE, Begley TP. Catalysis of a flavoenzymemediated amide hydrolysis. J Am Chem Soc 2010;132:5550–1. [105] Sucharitakul J, Phongsak T, Entsch B, Svasti J, Chaiyen P, Ballou DP. Kinetics of a two-component p-hydroxyphenylacetate hydroxylase explain how reduced flavin is transferred from the reductase to the oxygenase. Biochemistry 2007;46:8611–23. [106] Sucharitakul J, Chaiyen P, Entsch B, Ballou DP. The reductase of p-hydroxyphenylacetate 3-hydroxylase from Acinetobacter baumannii requires p-hydroxyphenylacetate for effective catalysis. Biochemistry 2005;44:10434–42. [107] Ruangchan N, Tongsook C, Sucharitakul J, Chaiyen P. pH-dependent studies reveal an efficient hydroxylation mechanism of the oxygenase component of p-hydroxyphenylacetate 3-hydroxylase. J Biol Chem 2011;286:223–33. [108] Tongsook C, Sucharitakul J, Thotsaporn K, Chaiyen P. Interactions with the substrate phenolic group are essential for hydroxylation by the oxygenase component of p-hydroxyphenylacetate 3-hydroxylase. J Biol Chem 2011;286:44491–502. [109] Phongsak T, Sucharitakul J, Thotsaporn K, Oonanant W, Yuvaniyama J, Svasti J, Ballou DP, Chaiyen P. The C-terminal domain of 4-hydroxyphenylacetate 3-hydroxylase from Acinetobacter baumannii is an autoinhibitory domain. J Biol Chem 2012;287: 26213–22. [110] Uetz T, Schneider R, Snozzi M, Egli T. Purification and characterization of a two-component monooxygenase that hydroxylates nitrilotriacetate from “Chelatobacter” strain ATCC 29600. J Bacteriol 1992;174:1179–88. [111] Valton J, Mathevon C, Fontecave M, Niviere V, Ballou DP. Mechanism and regulation of the two-component FMN-dependent monooxygenase ActVA-ActVB from Streptomyces coelicolor. J Biol Chem 2008;283:10287–96. [112] Siitonen V, Blauenburg B, Kallio P, Mantsala P, Metsa-Ketela M. Discovery of a two-component monooxygenase SnoaW/SnoaL2 involved in nogalamycin biosynthesis. Chem Biol 2012;19:638–46. [113] Schmid A, Hofstetter K, Feiten HJ, Hollmann F, Witholt B. Integrated biocatalytic synthesis on gram scale: the highly enantioselective preparation of chiral oxiranes with styrene monooxygenase. Adv Synth Catal 2001;343:732–7. [114] Gursky LJ, Nikodinovic-Runic J, Feenstra KA, O’Connor KE. In vitro evolution of styrene monooxygenase from Pseudomonas putida CA-3 for improved epoxide synthesis. Appl Microbiol Biotechnol 2010;85:995–1004. [115] Qaed AA, Lin H, Tang D-F, Wu Z-L. Rational design of styrene monooxygenase mutants with altered substrate preference. Biotechnol Lett 2011;33:611–6. [116] Lin H, Tang DF, Ahmed AA, Liu Y, Wu ZL. Mutations at the putative active cavity of styrene monooxygenase: Enhanced activity and reversed enantioselectivity. J Biotechnol 2012. [117] van Pée K-H, Patallo EP. Flavin-dependent halogenases involved in secondary metabolism in bacteria. Appl Microbiol Biotechnol 2006;70:631–41. [118] Blasiak LC, Drennan CL. Structural perspective on enzymatic halogenation. Acc Chem Res 2009;42:147–55. [119] Buedenbender S, Rachid S, Muller R, Schulz GE. Structure and action of the myxobacterial chondrochloren halogenase CndH: a new variant of FAD-dependent halogenases. J Mol Biol 2009;385:520–30.

72

3 The flavin monooxygenases

[120] Butler A, Sandy M. Mechanistic considerations of halogenating enzymes. Nature 2009;460:848–54. [121] Zhu X, De Laurentis W, Leang K, Herrmann J, Ihlefeld K, van Pée K-H, Naismith JH. Structural insights into regioselectivity in the enzymatic chlorination of tryptophan. J Mol Biol 2009;391:74–85. [122] Yeh E, Blasiak LC, Koglin A, Drennan CL, Walsh CT. Chlorination by a long-lived intermediate in the mechanism of flavin-dependent halogenases. Biochemistry 2007;46:1284–92. [123] Lang A, Polnick S, Nicke T, William P, Patallo EP, Naismith JH, van Pée K-H. Changing the regioselectivity of the tryptophan 7-halogenase PrnA by site-directed mutagenesis. Angew Chem 2011;50:2951–3. [124] Rachid S, Krug D, Kunze B, Kochems I, Scharfe M, Zabriskie TM, Blocker H, Mueller R. Molecular and biochemical studies of chondramide formation - Highly cytotoxic natural products from Chondromyces crocatus Cm c5. Chem Biol 2006;13:667–81. [125] Xiao Y, Li S, Niu S, Ma L, Zhang G, Zhang H, Zhang G, Ju J, Zhang C. Characterization of tiacumicin B biosynthetic gene cluster affording diversified tiacumicin analogues and revealing a tailoring dihalogenase. J Am Chem Soc 2011;133:1092–105. [126] Dorrestein PC, Yeh E, Garneau-Tsodikova S, Kelleher NL, Walsh CT. Dichlorination of a pyrrolyl-S-carrier protein by FADH2-dependent halogenase PltA during pyoluteorin biosynthesis. Proc Natl Acad Sci USA 2005;102:13843–8. [127] Neumann CS, Walsh CT, Kay RR. A flavin-dependent halogenase catalyzes the chlorination step in the biosynthesis of Dictyostelium differentiation-inducing factor 1. Proc Natl Acad Sci USA 2010;107:5798–803. [128] Lin S, Van Lanen SG, Shen B. Regiospecific chlorination of (S)-beta-tyrosyl-S-carrier protein catalyzed by SgcC3 in the biosynthesis of the enediyne antitumor antibiotic C-1027. J Am Chem Soc 2007;129:12432–8. [129] Podzelinska K, Latimer R, Bhattacharya A, Vining LC, Zechel DL, Jia ZC. Chloramphenicol biosynthesis: the structure of CmIS, a flavin-dependent halogenase showing a covalent flavin-aspartate bond. J Mol Biol 2010;397:316–31. [130] Yamanaka K, Ryan KS, Gulder TA, Hughes CC, Moore BS. Flavoenzyme-catalyzed atroposelective N,C-bipyrrole homocoupling in marinopyrrole biosynthesis. J Am Chem Soc 2012;134:12434–7. [131] Thompson JD, Higgins DG, Gibson TJ. CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucl Acid Res 1994;22:4673–80.

4 Structure and catalytic mechanism of NADPHcytochrome P450 oxidoreductase: a prototype of the diflavin oxidoreductase family of enzymes Jung-Ja P. Kim, Anna L. Shen and Chuanwu Xia

Abstract NADPH-cytochrome P450 oxidoreductase (CYPOR) is the founding member of the diflavin oxidoreductase family of enzymes that contain two flavins, FMN and FAD, in a single peptide. The enzyme functions to transfer electrons from NADPH to an ultimate electron acceptor, most notably the heme group of cytochromes P450, via FAD and FMN. The domain structure of CYPOR is consistent with evolution of CYPOR as a fusion between two proteins related to flavodoxin and ferredoxin-NADP+ reductase, joined by a hinge and connecting domain. Evidence outlining the basis for binding of FMN, FAD, and electron acceptors is discussed, as are the catalytic steps, hydride transfer, interflavin electron transfer, and FMN to heme electron transfer. The two flavin domains undergo a series of conformational changes throughout the catalytic cycle, ranging from movements of specific amino acid residues and local loops to large scale domain movements, forming the basis for controlled transfer of electrons from NADPH to electron acceptors. Information on the physiological functions of CYPOR derived from mouse models and studies of human CYPOR polymorphisms reveal the essential roles of CYPOR and its acceptors, in particular the cytochromes P450, in metabolism of endogenous and exogenous substrates.

4.1 Introduction NADPH-cytochrome P450 oxidoreductase (CYPOR) is the first and prototypic member of the diflavin oxidoreductase family of enzymes that contain one molecule each of FAD and FMN in a single polypeptide and transfer electrons from NADPH to the ultimate electron acceptor. These enzymes perform a step-down function, transferring electrons from the two-electron donor NADPH to one-electron acceptors, with the FAD functioning as a dehydrogenase flavin and FMN as an electron carrier. Other prominent members of this family are the reductase domains of the nitric oxide synthase (NOS) isozymes (reviewed in [1–3] and the flavocytochrome P450BM3 (P450BM3) from Bacillus megaterium [4] and the flavoprotein subunits of bacterial sulfite reductase (SiR) [5], all of which transfer electrons to heme, as well as methionine synthase reductase (MSR) which reduces oxidized Cob(I)alamin [6–8] , human cancer-related novel reductase 1 (NR1) [9], pyruvate:NADP+ oxidoreductase from Euglena gracilis [10,11], and

74

4 Structure and mechanism of NADPH-cytochrome P450 oxidoreductase

reductase Tah18 protein from yeast [12]. The domain structures of these proteins are all similar to that of CYPOR, containing the FNR- and Fld-like folds, with similar functions and mechanisms of action (򐂰Figs. 4.1 and 4.2). Physiologically, CYPOR functions to transfer electrons from NADPH to a number of microsomal electron acceptors, most notably the cytochromes P450, but also heme oxygenase [13], cytochrome b5 (cyt b5) [14], squalene monooxygenase [15], and possibly indole dioxygenase [16]. In addition, a number of nonphysiological electron acceptors, including cytochrome c (cyt c), ferricyanide, menadione, and dichloroindophenol, have been useful in biochemical characterization of the enzyme. Other members of the diflavin oxidoreductase family, MSR, NOS, and P450BM3, transfer electrons to a single donor. For NOS and P450BM3, the donor is located on the same polypeptide (򐂰Fig. 4.1). In contrast, the electron acceptors for CYPOR include the multiplicity of cytochromes P450 as well as the other protein acceptors listed above. All are located in the endoplasmic reticulum, and the levels of CYPOR are substantially lower than those of its acceptors, with the ratio of CYPOR to P450 in liver endoplasmic reticulum estimated at 1:15 to 1:20 [17,18]. Although the large and diverse family of cytochromes P450 exhibits a common fold in the vicinity of the heme ligand, each also possesses unique structural features, substrate specificity, and rate-limiting catalytic steps [19]; electron transfer to all these proteins must proceed in a controlled fashion to accommodate the wide range of physiologic functions. The question arises as to how CYPOR recognizes and mediates electron transfer to this multiplicity of electron acceptors. This review describes the domain organization of CYPOR and highlights the high degree of conformational flexibility, coupled with tight control of the required conformational changes, necessary for CYPOR to precisely orchestrate electron transfer to its diverse acceptors.

FNR

Fld FMN CYPOR

P450

MBD

FMN

NR1 MSR

FMN

HEME

CaM

H

EHR

FMN

P450

BM3 NOS

FMN

FAD/NADPH H

FMN

H H

CD

FAD/NADPH

CD

FAD/NADPH

CD

FAD/NADPH

CD

FAD/NADPH

CD

FAD/NADPH

Fig. 4.1: Domain organization of CYPOR and other members of the diflavin oxidoreductase family. Fld, flavodoxin; FNR, ferredoxin-NADP+ oxidoreductase; CYPOR, NADPH-cytochrome P450 oxidoreductase; MBD, transmembrane domain; H, hinge; CD, connecting domain; BM3, Bacillus megaterium flavocytochrome P450; NR1, novel reductase 1; MSR, methionine synthase reductase, which contains an ~80 residue extended hinge region (EHR) between the FMN domain and CD; NOS, nitric oxide synthase, which has a calmodulin binding region (CaM). Note that CD consists of two different parts of the linear sequence interspersed with the FNR-like domain.

4.2 Properties of CYPOR flavins A

Fld

FNR

B

D

CYPOR

C

75

nNOS-red

E

CYPOR vs FNR/Fld

CYPOR vs nNOS-red

Fig. 4.2: Evolutionary origins of the structures of CYPOR and nNOS reductase domain (nNOS-red), shown by overlays of the ribbon structures of D. vulgaris flavodoxin (Fld) and spinach ferredoxin-NADP-oxidoreductase (FNR). (A) Structures of Fld and FNR; (B) CYPOR (34) with FMN and FAD highlighted with red sticks; (C) the nNOS reductase domain (nNOSred) [139], which contains three regulatory elements, the autoregulatory insert (AR), β-finger (BF), and the C-terminal extension (CT) shown in red; (D) overlay of the structures of Fld, FNR and CYPOR; (E) overlay of CYPOR and nNOSred.

4.2 Properties of CYPOR flavins The ability of flavins to stabilize an unpaired electron and engage in both one- and two-electron oxidation-reduction chemistry is key to their functions in electron transfer. In CYPOR, they are an essential transducer between NADPH, an obligate two-electron donor, and the heme of P450, an obligate one-electron acceptor. Furthermore, utilization of two flavins, located in separate domains, provides a mechanism for control of the kinetics of electron transfer. Both FMN and FAD cofactors can exist as the oxidized, semiquinone (one-electron reduced), and fully reduced (two-electron reduced) forms. Both the semiquinone and fully reduced forms can exist as neutral or anionic forms, with pKa’s in free solution of 8.5 and 6.5, respectively. CYPOR stabilizes the blue, neutral form of the FMN and FAD semiquinones in the pH range 6.5 to 8.5 [20,21]. In this review, the fully reduced forms are referred to as FMNH2 and FADH2, with the understanding that the protonation states of the fully reduced forms in CYPOR are in fact unknown. Those of the homologous proteins, FNR and flavodoxin, are anionic, and, thus, the fully reduced flavins in CYPOR may very well also be in the anionic forms, i.e., FADH– and FMNH–. The oxidation and protonation states of the flavins can be distinguished by their distinct visible absorption spectra, which have been invaluable in characterizing the

76

4 Structure and mechanism of NADPH-cytochrome P450 oxidoreductase

oxidation states of flavoproteins during catalysis [21,22]. Oxidized flavins have broad absorption maxima at approximately 450 and 380 nm. The neutral semiquinone forms are characterized by a broad absorbance between 500 and 700 nm, with maxima in the region between 585 and 600 nm. In CYPOR, the FMN, but not the FAD, semiquinone has a shoulder at 630 nm, which enables discrimination of the FADH• and FMNH• semiquinones and analysis of one-electron transfer reactions between FAD and FMN. The FMNH• semiquinone is stable in air for a period of days, while the FADH• semiquinone is unstable and rapidly oxidizes in air. The stability of the neutral FMNH• semiquinone is likely due to a hydrogen bond with the main chain carbonyl group of a highly conserved glycine residue (Gly141 in rat CYPOR). Formation or cleavage of this hydrogen bond upon change in the flavin oxidation state results in a significant increase in the reorganizational energy for the electron transfer reactions [23], leading to decreased reactivity of the FMNH• semiquinone and promotion of FMN shuttling between the fully reduced and semiquinone states during one-electron transfers. In contrast, this conserved glycine, found in Fld, CYPOR and NOS, is missing in P450BM3 [24,25], whose FMN semiquinone exists as an unstable anionic form having a higher reactivity than the neutral semiquinone. Consequently, the semiquinone rather than the hydroquinone is the electron donor in this enzyme. The reduction potentials of the CYPOR flavins have been determined for the rabbit [20], rat [26], and human [27,28] enzymes (򐂰Tab. 4.1). Although there are some variations in reduction potentials between species, the FAD semiquinone/reduced couple always exhibits a low reduction potential, at or near that of NADPH (–320 mV). FAD is the low-potential flavin and electron transfer proceeds from NADPH to FAD to FMN to P450 [29]. These reduction potentials have been determined for the solubilized protein in aqueous solution; the possibility that membrane lipids may influence the flavin reduction potentials is discussed below. The step-down mechanism by which electrons from the two-electron donor NADPH are transferred to a one-electron acceptor is illustrated in 򐂰Fig. 4.3. Two electrons from NADPH must enter the enzyme via hydride transfer to the FAD followed by intramolecular electron transfer to FMN. The lack of reactivity of the FMN semiquinone indicates that it is the fully reduced FMN that transfers electrons to electron acceptors and

Tab. 4.1: Redox Potentials of CYPOR Flavins. Source

pH

FMN

FAD

Ref

ox/sq

sq/red

ox/sq

sq/red

Rabbit

7.0

–110

–270

–290

–372

[20]

Human

7.0

–66

–269

–283

–382

[28]

Human

7.5

–89

–246

–328

–381

[27]

Rat

7.4

–68

–246

–325

–372

[26]

Redox potentials in mV are shown for CYPOR purified from the indicated organisms. Rat CYPOR was purified with an intact membrane binding domain, while the human and rabbit enzymes do not contain the membrane binding domain. Redox couples are: ox/sq, oxidized/ semiquinone; sq/red, semiquinone/fully-reduced.

4.2 Properties of CYPOR flavins 1e

e

77

NADPH

FMNH• FAD FMN FAD NADPH

2e

FMN FADH2

FMNH• 3e FADH2

e e

2e

FMNH2 FAD

FMNH2 4e FADH2 FMNH2 3e FADH• NADPH

FMNH• FADH•

e

2e

Fig. 4.3: Redox cycling of CYPOR flavins. Redox cycling and electron transfer through 1-3-2-1 and 2-4-3-2 reaction cycles are indicated. The heavy solid lines indicate steps common to both cycles. The 1-3-2-1 cycle is shown on the right side of the diagram, with the thin solid lines indicating steps specific to this cycle. The dashed lines indicate steps specific to the 2-4-3-2 electron transfer cycle. The priming reaction, through which reduction of fully-oxidized enzyme can enter the 1-3-2-1 cycle, is shown as dotted lines on the left side of the diagram. The gray dotted line indicates the slow oxidation of the air-stable semiquinone, FMN•/FAD.

that the fully oxidized form does not accumulate. Given these constraints, CYPOR can cycle in a 1-3-2-1 electron cycle or a 2-4-3-2 cycle, as shown in 򐂰Fig. 4.3. Electron transfer steps common to both cycles are indicated by heavy solid lines. The 1-3-2-1 catalytic cycle proceeds directly from NADPH reduction of the air-stable FMNH•/FAD semiquinone and is indicated by the thin solid lines. As the FMNH•/FAD form is the predominant form in liver microsomes [21], this 1-3-2-1 cycle is the cycle relevant to catalysis. A 2-4-3-2 catalytic cycle is also possible, however, depending on the NADPH/NADP+ ratio [22,30], and is indicated by the dashed lines. This 2-4-3-2 cycle feeds into the 1-3-2-1 cycle. For example, NADPH reduction of the fully oxidized enzyme followed by 1-electron transfer, shown on the left side of 򐂰Fig. 4.3, has been described as a “priming” reaction that can feed into the 1-3-2-1 cycle [20,21,31]. The FMNH2/FAD intermediate can formally participate in both cycles, but the very slow rate of oxidation of the FMNH•/FAD air-stable semiquinone to the fully oxidized enzyme (gray dotted line) indicates that the predominant pathway for this intermediate involves the 1-3-2-1 cycle. Although the low reduction potential of FAD, near or below that of NADPH (–320 mV), suggests that formation of the four-electron fully reduced form of the enzyme is thermodynamically unfavorable, any such fully reduced enzyme produced can also feed into the 1-3-2-1 cycle. The internal electron equilibrium within the three-electron reduced enzyme: [FADH2/FMNH•]

[FADH•/FMNH2]

lies to the right, as judged from the absorbance change at 630 nm [22,32].

78

4 Structure and mechanism of NADPH-cytochrome P450 oxidoreductase

4.3 Domain structure and function On the basis of DNA sequence homology [33], it was predicted that CYPOR likely arose from the fusion of two ancestral genes related to the flavodoxin and ferredoxin-NADP+ reductase proteins. This hypothesis has subsequently been confirmed both by site-directed mutagenesis studies confirming the structural and catalytic functions of conserved residues and by X-ray crystallography [34]. The domain organization of CYPOR is apparent from the crystal structure of CYPOR, with domains structurally related to flavodoxin and FNR (򐂰Fig. 4.2). Conservation of cofactor binding and catalytic residues is observed. Boundaries of the domains correspond to exon junctions in the gene encoding the enzyme, additional evidence that CYPOR has arisen from a gene fusion event. The three-dimensional protein structures of spinach FNR, Fld from Desulfovibrio vulgaris and rat CYPOR also strongly support a common ancestor based on the very high structural similarity between the individual domains despite their very different origin [35]. The ability to express the different domains of CPR as individual, functionally active proteins and to successfully reconstitute these domains in vitro to form a functional protein complex with NADPH-cytochrome P450 oxidoreductase activity is additional evidence that CYPOR has evolved as a result of gene fusion event [36,37]. CYPOR is anchored in the microsomal membrane by a ~56-amino acid N-terminal membrane binding domain (MBD) [38], with the catalytic functions of CYPOR residing in the soluble portion, residues 66–678 (residue numbering is based on rat CYPOR, unless otherwise noted). As shown in 򐂰Fig. 4.2, the structure of the soluble portion of CYPOR is composed of an FMN-binding domain, which is structurally similar to Fld and an FAD domain. The FAD domain consists of an FNR-like domain with binding sites for FAD and NADPH and a connecting domain (CD), composed mainly of α helices, that connects the FMN and FNR-like domains. The FMN and FAD domains are linked by a flexible hinge (residues 232–243), comprised of mostly hydrophilic residues. The two flavin isoalloxazine rings are juxtaposed to each other, making a continuous ribbon, with an ~150° bend between the planes of the two rings. The closest distance between the two rings is ~4 Å (between the two pairs of 7- and 8-methyl groups of the dimethylbenzene rings), suggesting that the two flavins communicate with each other directly through their dimethylbenzene rings [34]. The presence of a connecting domain and hinge is unique to the diflavin oxidoreductases (򐂰Fig. 4.1). Although the amino acid sequences of the connecting domains (CDs) exhibit low (< 30%) sequence homology, there is significant structural similarity among CDs of different members of the diflavin oxidoreductase family (see comparison of CYPOR and nNOS in 򐂰Fig. 4.2). Both the length and sequence of the hinge are unique for each member of this family. The hinge plays a crucial role in CYPOR’s interaction in its electron transfer partners. It is believed that the hinge and connecting domain are the source of the domain movements that control cofactor binding, interflavin electron transfer, and recognition and electron transfer to substrates (see below).

4.4 Membrane binding domain (MBD) CYPOR is anchored in the lipid bilayer of the endoplasmic reticulum and nuclear membrane by an approximately 56 amino acid membrane binding domain. The MBD

4.5 FMN domain

79

contains a 23-amino acid stretch of hydrophobic amino acids which presumably extends into the lipid bilayer, followed by a stop-transfer sequence, 45RKKKEE50, and a flexible segment susceptible to proteolytic cleavage [38,39]. Cleavage at the Lys56Ile57 bond releases the CYPOR from the microsomal membrane. The trypsin-cleaved protein is no longer able to transfer electrons to cytochrome P450, but retains activity towards other electron acceptors such as cytochrome c. Similarly, cytochrome b5 is attached to the membrane via a C-terminal membrane binding domain which is necessary for electron transfer to P450. Both passive and active roles in P450-mediated catalysis have been proposed for the MBD. Since fusion proteins, such as P450BM3 and NOS isozymes do not require the MBD for catalytic activity, the MBD may serve to localize and possibly restrict movement of CYPOR in the membrane rather than provide a specific binding site [40–42]. In this case, the precise sequence of the membrane domain would be less important than its ability to insert into the membrane. Substitution of the CYPOR membrane binding domain with that of cytochrome b5, which has a similar hydrophobicity profile but only about 20% sequence identity [43], produced a chimeric CYPOR that was able to support CYP17A-mediated P450 activity but not CYP3A4-mediated testosterone 6β-hydroxylation. Coupled with the observation that the membrane binding domain of yeast CYPOR is not required for electron transfer to CYP51 [44], it appears that the membrane binding domain may contribute to P450 recognition and binding but is likely only one of many CYPOR-P450 interactions that may vary depending on the specific P450. Finally, an interesting function for the membrane binding domain has recently been proposed by Das and Sligar [45], who have shown that the flavin reduction potentials are influenced by the composition of the lipid bilayer. The significance of these altered reduction potentials relative to catalysis has not been demonstrated, although lipid composition, including charge, has been reported to influence rates of P450 metabolism in reconstituted systems [46].

4.5 FMN domain The FMN domain, structurally similar to the bacterial flavodoxins, extends from residues 77 to 231 of rat CYPOR and consists of a five-stranded parallel β-sheet flanked by five α-helices (򐂰Fig. 4.2) with the FMN positioned at the tip of the C-terminal side of the β-sheet. Located in this domain are the binding site for the FMN prosthetic group and residues mediating binding of and electron transfer to acceptors such as cytochrome c and cytochrome P450. FMN is bound with a KD of 10−8 M and can be reversibly removed from the enzyme by high salt treatment [22,47]. In the absence of FMN, electron transfer to all acceptors, with the exception of ferricyanide, is abolished. As seen in Fld, the isoalloxazine of FMN is sandwiched between the aromatic groups Tyr140 and Tyr178, with Tyr178 coplanar with the si- face of the flavin, while Tyr140 is located on the re face at a ~60o angle to the isoalloxazine ring [34]. Mutation of Tyr178 to aspartate decreases FMN binding to undetectable levels, with an approximately 300-fold decrease in FMN binding affinity, and also disrupts FAD binding [48]. A similar decrease in FMN binding affinity is seen when the homologous residue of human CYPOR, Tyr181, is mutated to aspartate [49,50]; however FAD binding is not disrupted in the case of the human mutation. Restoration of catalytic activity by FMN demonstrates

80

4 Structure and mechanism of NADPH-cytochrome P450 oxidoreductase

that the inability to incorporate FMN is the likely basis for the NADPH-cytochrome P450 oxidoreductase deficiency (PORD) phenotype associated with this human mutation. Although mutation of Tyr140 does not affect the FMN content of the isolated variant (rat) protein, affinity for FMN is decreased seven-fold and cytochrome c reductase activity is decreased approximately five-fold. The rate of electron transfer to ferricyanide activity is identical to that seen in wild-type enzyme, suggesting that hydride transfer is not impaired. Decreased cytochrome c reductase activity as well as a requirement for preincubation with NADPH for maximum cytochrome c reductase activity suggests that Tyr140 may have a function in interflavin electron transfer [48].

4.6 Cytochrome P450 binding: role of the FMN domain and connecting domain The negatively charged surface of the FMN domain can interact with the proximal face of cytochrome P450 in the vicinity of the heme ligand, which contains a relatively electropositive conserved residues (򐂰Figs. 4.4 and 4.5) [51–54]. This region of P450 contains overlapping binding sites for CYPOR and cytochrome b5. A model of a putative P450-CYPOR complex shows the total contact area between the two molecules to be ~1500 Å2, of which 870 Å2 is located between the FMN (򐂰Fig. 4.5) domain and P450 [51]. A number of charge-pairing and van der Waals interactions have been implicated in binding of P450 to CYPOR, indicating that both electrostatic and hydrophobic interactions are necessary for the complex formation. The FMN domain bears conserved patches of acidic amino acid residues involved in the electrostatic interactions with its electron transfer partners and these interactions are specific for each electron transfer partner. Crosslinking experiments suggest that acidic 1

2

3

4

Open conformation

Closed conformation

Selection process

Specific complex (Electron transfer)

HEM E

HEME

M E

HEME HE

FMN

FA D

HEME

NADP N FM

FM N

E M HE

E M HE

FMN

HEME

E M HE

FAD

NADPH

E M HE

FAD

NA DP H

FA D

ER mebrane

Fig. 4.4: CYPOR-P450 binding and electron transfer. FMN and FAD domains are shown in golden yellow and P450 in pink. Positively and negatively charged areas are indicated by blue and red bars, respectively. Specific interactions between two domains/proteins, including hydrophobic interactions and hydrogen bonds, are shown with black sawteeth. Nucleotide binding favors formation of the closed form. Multiple P450s associate with both the open and closed forms of CYPOR. In the closed form (panel 2), cofactor binding, interflavin electron transfer, and release of cofactor occur, resulting in formation of the open form of CYPOR (panel 3). This occurs in conjunction with a selection process, resulting in formation of the specific complex, in which the CYPOR:P450 stoichiometry is 1:1. Further conformational changes including fine tuning occur to align the flavin and heme groups in a geometry optimal for electron transfer (panel 4).

4.6 Cytochrome P450 binding: role of the FMN domain and connecting domain B

A

81

C

R443(d) HEME

E208(e)

R133(e)

90°

90°

E142(d) FMN

12 Å Glu439 E93(b) R422(d)

Phe429 R122(a) K433(c)

R126(b)

HEME

FMN

E92(a)

D113(c)

Fig. 4.5: Model of a complex between P450 and CYPOR. (A) A complex of P450 and Mol A of the hinge-deletion mutant of CYPOR (see 򐂰Fig. 4.8) and an enlarged view showing the relative orientation of the FMN and heme. (B) and (C) Electrostatic surface at the interface of P450 (B) and CYPOR (C). Blue represents a positively charged surface and red indicates a negatively charged surface. Five salt-bridge pairs are shown with same letters. Glu242 makes salt bridges with both Arg422 and Arg443.

residues in the FMN domain (207Asp-Asp-Asp209 and 213Glu-Glu-Asp215) contribute to binding of cytochrome c; however, crosslinking of these residues to cytochrome P450 could not be demonstrated [55,56]. Mutagenesis studies have demonstrated the importance of Glu213 and Glu214 in electrostatic interactions with oxidized and reduced cytochrome c. The 213Glu-Glu-Asp215 cluster does not affect cytochrome P450 binding or activity, highlighting the distinct binding modes for these two substrates. However, charge-reversal of Asp208 decreases P450-dependent catalytic activity, with no effect on cytochrome c binding or activity, suggesting that this residue may have a specific role in electron transfer to P450 [57]. Chemical modification and antibody labeling experiments have also suggested that the loop between β-sheet 2 and α-helix C in CYPOR, located on the opposite face of the protein, can also contribute to P450 binding and catalysis (reviewed in [58]). Sitedirected mutagenesis of Asp113, Glu115, and Glu116 improves catalytic efficiency of cytochrome c reduction but destabilizes the CYPOR-CYP2B1 complex [59]. A variety of cytochrome P450 chemical modification studies, reviewed by Hlavica et al. [60] and Im and Waskell [53], have provided evidence implicating basic residues in the C-helix of P450 in electrostatic interactions with CYPOR and cytochrome b5. CYP2B4 site-directed mutagenesis studies have identified seven basic and hydrophobic amino acids (Arg122, Arg126, Arg133, Phe135, Met137, Lys139, and Lys433), all except Lys433 located in the mobile C-helix, as important for both cytochrome b5 and CYPOR binding [52,61]. The synthetic peptide 134DFGMGKR140, competitively inhibits P450dependent activity and crosslinking experiments have demonstrated that these basic residues form a charge pair with the peptide 425RHILAILQDCPSLRPPIDHLCELLPR450 [62], located in the connecting domain of CYPOR. Finally, in addition to electrostatic interactions, the hydrophobic amino acid residues, Val267 and Leu270 on the proximal

82

4 Structure and mechanism of NADPH-cytochrome P450 oxidoreductase

site of CYP2B4 also contribute to CYPOR recognition [63]. Although the electron transfer is presumed to occur within a 1:1 CYPOR:P450 complex [64], the presence of higher order complexes contributing to catalysis has been demonstrated [65,66]. The contribution of these higher-order complexes to catalysis is not well-understood, but it is likely that multiple P450s do bind to CYPOR in the course of catalysis. Binding is likely to involve interactions at multiple sites on the P450, including but not limited to, those described here.

4.7 FAD domain The FAD domain of CYPOR is composed of the connecting domain (CD) and the FNRlike domain which binds FAD and NADPH (򐂰Figs. 4.1 and 4.2). The FNR-like sequence consists of residues 267 to 325 and 450 to 678, interspersed with the CD (residues 244 to 266 and 326 to 450). Conserved residues necessary for FAD and NADPH binding, as well as hydride transfer, are localized in this FNR-like domain. Unlike FMN, FAD is tightly bound to the reductase with a KD < 1 nM. Removal of FAD requires treatment with a high concentration of chaotropic agent, leading to substantial polypeptide unfolding, providing further evidence for independence of the two domains [67–69]. Residues comprising the FAD binding site include 455YYSIASS461, 471 ICAVAVEY478, 488GVAT491. Although Trp677 is stacked against the re-face of the FAD, removal of this residue does not have a significant effect on FAD content; the role of this residue in catalysis is discussed below. Major determinants of FAD binding are Arg454, which stabilizes the negative charge of the FAD pyrophosphate, and Tyr456, which is positioned at a 60o angle to the si face of the isoalloxazine ring and whose phenolic hydroxyl group forms a hydrogen bond with the ribityl 4’-hydroxyl [34,70]. Removal of the hydrogen bonding interaction between Thr491 and the pyrophosphate destabilizes FAD binding to a lesser extent. The rapid and reversible dissociation of FAD under relatively mild conditions from the T491V mutant has been exploited to introduce flavin analogs into the FAD binding site [71].

4.8 Mechanism of hydride transfer CYPOR transfers the pro-R hydrogen from NADPH to FAD as a hydride. Residues essential for this hydride transfer include Ser457, Asp675, and Cys630, which are located in close proximity to the redox-active N5 of FAD and form a hydrogen bonding network that is disrupted upon binding of nicotinamide [72,73]. Replacement of these side chains with aliphatic groups decreases catalytic activites by up to three orders of magnitude. Ser457 and Asp675 interact with the nicotinamide group of NADP(H) and orient the C4 atom of the nicotinamide ring in a position for optimum hydride transfer. Cys630 is also within van der Waals distance from the nicotinamide C4 and can stabilize the carbocation formed during hydride transfer [74]. In addition, Ser457 is located ~4 Å away from the flavin N5 and on the same plane as the flavin ring, in a position to stabilize the semiquinone form of FAD, and replacement of Ser457 with alanine decreases the FAD/FADH• redox potential.

4.9 Interflavin electron transfer

83

The penultimate Trp677 residue plays an pivotal role in catalysis through control of NADP(H) binding and release [74]. Interestingly, the indole ring of Trp677 is situated at the re-face of the FAD, where the nicotinamide ring of NADPH would bind to transfer its pro-R-hydrogen as a hydride ion. Furthermore, in the structure of wild-type enzyme, while the binding site for the adenosine-pyrophosphate half of the NADP+ is clearly shown, the ribosyl-nicotinamide moiety is disordered. However, structures of CYPOR lacking the indole ring, by deletion of the two last C-terminal residues (Trp677 and Ser678), or mutation of Trp677 to glycine (Trp677Gly), reveal that the indole ring is replaced by the nicotinamide ring at the re-face of the FAD, with a tilt of ~30° between the planes of the two rings, poised to transfer the hydride ion [74]. Thus, in the wildtype protein, the indole ring of Trp677 presumably moves away from the isoalloxazine ring of FAD, allowing the nicotinamide ring to interact with the flavin so that hydride transfer can occur. In pea FNR, the homologous residue, Tyr308, is also displaced by the nicotinamide ring [75,76]. Mutagenesis and crystallographic studies have illuminated the bipartite nature of NADP(H) binding and provide an explanation of the marked preference of CYPOR and FNR for the cofactor NADPH. The primary determinant for discrimination between NADH and NADPH is the 2’-phosphate group present on NADPH but not NADH. Kinetic studies show that this 2’-phosphate of NADPH, binding as the dianion, contributes 5 kcal of binding energy through interactions with enzyme groups, with interactions with Arg597 accounting for ~3 kcal of binding energy. Lys602 and Ser596 also contribute to binding [77]. This tight binding of the 2’-phosphate is essential to compensate for the repulsive interactions between the nicotinamide and the indole ring of Trp677. When Trp677 is present, binding of the 2’-phosphate stabilizes cofactor binding sufficiently to allow the nicotinamide to displace Trp677. In the absence of Trp677, the nicotinamide can bind readily without any contribution from the 2’-phosphate and the enzyme is able to utilize NADH as the hydride donor. Furthermore, in the absence of Trp677, the enzyme is unable to displace oxidized nicotinamide after hydride transfer and catalytic efficiency with either NADH or NADPH is decreased due to rate-limiting product release [74,78,79], indicating that movement of Trp677 is required for both cofactor binding and release. These studies indicate a requirement for structural changes, in addition to Trp677 movement, in regulation of NADP(H) binding and release. While movement of Trp677 back into the nicotinamide binding site (re-face of the FAD isoalloxazine ring) displaces the nicotinamide, additional movements are necessary to disrupt the strong binding of the 2’-phosphate. Local movements of the 631GDARN635 loop (Asp632 loop), located in the FAD-binding cavity, may be coupled with Trp677 movement to allow NADPH/ NADP+ binding/release [80]. Comparison of the structure of the NADP+-bound wild-type enzyme with that of a mutant CYPOR with an engineered disulfide bond between the two flavin domains and lacking bound NADP+, shows a movement of this Asp 632 loop. Thus, Xia et al. have proposed that Asp632 loop movement, in concert with movement of Trp677, controls at least in part NADPH binding and NADP+ release (򐂰Fig. 4.6) [80].

4.9 Interflavin electron transfer CYPOR intramolecular electron transfer occurs directly from FAD to FMN. In rat and human CYPOR [34,81], the distance between the dimethylbenzene edge of the

ADP

4 Structure and mechanism of NADPH-cytochrome P450 oxidoreductase

AD P

ADP

84

3

ADP

ADP

2

FMN

1

AD P 4

5  NADP

 NADPH ADP

ADP

 FADox

 Indole ring of TRP 677

 FADred

 D632 loop

Fig. 4.6: Schematic illustration of conformational changes upon NADPH binding to and NADP+ release from CYPOR. NADPH enters oxidized CYPOR (state 1, modeled after the structure of the disulfide cross-linked mutant, which lacks bound NADP(H)). The open-closed state of CYPOR at this stage is not known, but is most likely in an open form. Stages 2–5 are most likely in the closed form. For clarity, the FMN domain is not shown. NADPH initially binds to the enzyme via its ADP-PPi moiety with the interaction between the negative charges of pyrophosphate and 2’-phosphate with the positive charges of several arginines in the binding pocket, (see text). Once NADPH binds, the Asp632 loop moves and the ribityl-nicotinamide moiety extends to search for the proper binding site for hydride transfer, keeping the ADP-PPi anchored. The indole ring of Trp677 rotates to be ready to move away from the flavin ring (state 2) and then moves away to make room for the nicotinamide ring to bind at the re-side of the isoalloxazine ring (state 3). Hydride transfer occurs, producing reduced FAD and NADP+ (state 4). It is most likely that interflavin electron transfer occurs at this stage. Once the flavin is reduced and the nicotinamide is oxidized, the nicotinamide ring moves out and the indole ring returns to the re-face of the FAD ring (state 5). The D632 loop moves back closer to where the NADP+ pyrophosphate-2′-AMP lies, causing steric hindrance as well as electrostatic repulsion, resulting in dissociation of the cofactor from CYPOR. The enzyme now returns to state 1 and the cycle repeats. Figure adapted and modified from Xia et al. [80].

isoalloxazine rings of FAD and FMN is ~4 Å, and the planes of the FAD and FMN molecules are inclined relative to each other at an angle of ~150o, an orientation that favors orbital overlap between the extended π−π systems of the flavin isoalloxazine rings. This arrangement of the flavins is expected to result in very fast and efficient interflavin ET (up to 1010 s−1 using Dutton’s ruler [82]). However, the experimentally observed electron transfer rate has been measured to be only ~50 s−1 [83,84], suggesting that electron transfer is gated by some other process. The nature of the conformational

4.10 FMN to heme electron transfer

85

movement(s) controlling the rates of interflavin as well as flavin to heme electron transfer is discussed below.

4.10 FMN to heme electron transfer The FMN domain functions both to accept electrons from the reduced FAD and to transfer those electrons on to P450. Thus, precise and specific interactions between the FMN and FAD/NADPH domains within CYPOR, and between the FMN domain and P450 are required. i.e., the FMN domain must be able to recognize both the FAD domain and P450. Movement of the FMN domain is essential for this sequential electron transfer process. The FMN domain has a very strong molecular dipole [61], which is involved in the specific docking with the heme protein. Little is known of the mechanism through which CYPOR selects one of many electron transfer partners, and it is likely that multiple protein conformations and binding sites are probed in the selection process. 򐂰Fig. 4.4 presents a model incorporating current hypotheses regarding formation of a productive CYPOR-P450 electron transfer complex. Beginning from an oligomeric complex in which multiple P450s can bind to CYPOR in a closed conformation, a selection process must occur by which one P450 binds in a more favorable conformation. This process is concurrent with NADPH binding and interflavin electron transfer, resulting in an open conformation for CYPOR (see below). Through making and breaking of both electrostatic and hydrophobic interactions, a single P450 is selected for catalysis, forming the specific complex required for electron transfer. Further conformational changes, including fine-tuning, may be needed to produce the electron transfer complex, in which the flavin and heme are positioned for electron transfer. In contrast, the mechanism of electron transfer to small molecule acceptors such as dichloroindophenol or ferricyanide presumably involves random collisions followed by electron transfer. The requirements for cytochrome c are also likely less stringent than those for P450; kinetic studies suggest the presence of more than one binding site for cytochrome c [85]. A model for a docked CYPOR-P450 complex based on the open conformation of the CYPOR hinge mutant by Hamdane et al. [51] suggests that the FMN domain interacts with the concave basic proximal face of P450. The planes of the heme and FMN are almost perpendicular to each other, and the shortest distance between the heme and flavin cofactors is about 12 Å (򐂰Fig. 4.5). However, two P450 residues, Phe429 and Glu439, lie in between the two cofactors, suggesting that these might serve to facilitate electron transfer between the FMN and heme. In the structure of the complex between the heme and FMN-binding domains of bacterial cytochrome P450BM3, the relative orientation of the two cofactors is similar to that found in the model structure of Hamdane et al., but the distance between FMN and heme is slightly longer (~18 Å) [86].

4.11 P450 catalysis A generally accepted mechanism for P450 catalysis is outlined in 򐂰Fig. 4.7. While the broad outlines of this mechanism, first proposed in 1968, are applicable to all P450 reactions (for reviews, see [87,88] and confirmed by cryocrystallography in the studies

86

4 Structure and mechanism of NADPH-cytochrome P450 oxidoreductase O (RH) 

IV

H

H2O

H

O

H2O ROH

3

H2O

Fe

Fe

Fe3

S 8

S 9

S 1

Fe3

H2O

CYPOR FMN FAD

S 2

(P450 Compound I) O2

S 7

H

(RH)

RH

Au to -o xid at io n

OH O (RH) Fe3

R

O2 O (RH) Fe3

O O (RH) Fe2

S 6

S 5

e FMN FAD CYPOR or Cyt b5

O O (RH) Fe3 S

(RH) Fe2 S 3 O O (RH) 2 Fe

4

O2

S

Fig. 4.7: Catalytic cycle for P450 catalysis. This cycle was first proposed in 1968. The cycle begins with the oxidized, substrate-free P450 (1) and proceeds in a clockwise fashion as indicated. Electrons are donated to the substrate-bound Fe3+ (2) and the oxyferrous (4) intermediates. Species (4) is shown as an equilibrium between the oxyferrous and ferric superoxide forms. Autooxidation of this species generates superoxide. Adapted and modified from [87].

of P450cam [89,90], it is likely that there are some variations, especially in rates of individual steps, between P450s [88,91]. The catalytic cycle for cytochrome P450 begins with substrate binding to oxidized P450 (1) to generate the oxidized P450 enzyme-substrate complex (2), followed by transfer of the first electron from CYPOR to generate the reduced P450 enzyme-substrate complex (3). This first electron transfer to P450 requires CYPOR. Binding of oxygen to the reduced, substrate-bound, P450 produces the oxyferrous P450 intermediate (4), which accepts a second electron from either CYPOR or cytochrome b5. The resulting reduced oxyferrous P450 intermediate (5) decays to form the ferric peroxo intermediate (6) that then protonates to form the hydroperoxo species (7). This is the precursor to the critical Fe(IV)-oxo-porphyrin cation radical, “Cytochrome P450 Compound I” intermediate (8), capable of oxidizing the substrate carbon-hydrogen bond. Species (4) can undergo oxidation, generating superoxide. Other uncoupling reactions, at steps 6–8, also abort product formation and generate peroxide or water. The rate limiting step in catalysis varies with substrate and enzyme isoform [91]. Generally, association of CYPOR with P450 is slow, t1/2 ~2 min, and kinetic studies of P450 catalysis typically utilize a preformed CYPOR-P450-substrate complex. In these preformed complexes, introduction of the first electron is relatively fast compared to overall turnover [92,93]. One major factor influencing rates of P450 catalysis is cytochrome b5. The effects of cytochrome b5 on cytochrome P450 activity are complex, with evidence for stimulation, inhibition, or lack of effect, depending on the P450 isoform, substrate, and experimental system [53]. A wide range of P450 reactions are stimulated by cytochrome b5, as evidenced by decreased in vitro metabolism by CYPs 1A, 1B, 2B, 2C, 2D, and 3A, when liver microsomal cytochrome b5 is deleted [94]. Cyt b5 is also an allosteric effector of P450 activity; this occurs in the absence of electron transfer [95,96].

4.12 Other CYPOR electron acceptors

87

Inhibition of P450 metabolism by cytochrome b5 may be explained by the fact that CYPOR and cytochrome b5 bind to overlapping sites on the flexible C-helix on the proximal side of cytochrome P450 [52]. Consistent with this hypothesis, Mn-cyt b5 has been shown to be a competitive inhibitor of reduction of ferric P450 by CYPOR [97]. Stimulation of P450 reactions by cytochrome b5 is a consequence of transfer of the second electron from cytochrome b5. Zhang et al. [71], utilizing a CYPOR protein containing 5-deaza-FAD to eliminate spectral and oxidation state changes associated with the FAD, have provided evidence that the kinetics of P450 catalysis vary depending on the source of the second electron, possibly as a result of alternate P450 conformations induced by cytochrome b5 or CYPOR binding. Transfer of the second electron from either CYPOR or cytochrome b5 to oxyferrous substrate-bound cytochrome P4502B4 leads to rapid appearance of ferric P450 and product formation, with no observable accumulation of intermediates. However, second electron transfer from CYPOR results in rapid CYPOR oxidation but delayed formation of ferric P450 and product, and accumulation of an intermediate with spectral properties similar to that of oxyferrous P450. The identity of this intermediate has not been established, but it may arise as the result of a delayed proton transfer or electron transfer step when the electron donor is CYPOR.

4.12 Other CYPOR electron acceptors Cyt b5 has a number of physiological partners, including hemoglobin, methionine synthase [98], fatty acid desaturase and elongase, and cytochrome P450. It can accept electrons from NADH-dependent cytochrome b5 reductase or NADPH-dependent CYPOR [14]. Microsomal heme oxygenases also require CYPOR as their obligatory electron donor. Unlike P450, heme oxygenase isozymes use only one substrate (heme) and the axial ligand is His, not Cys as is found in the P450s. The HO system plays a key role in the physiological catabolism of heme, yielding biliverdin, CO, and iron as the final products. Both isoforms of HO, the 33 kDa inducible HO-1 and the 36 kDa constitutive HO-2, have hydrophobic tails (~30 residues) at their C-termini that are involved in binding to the microsomal membrane. Mechanistic studies have established that the transformation of heme to biliverdin is a three-step process that consumes three molecules of oxygen and seven electrons (see Review: [90]). The potential of human CYPOR mutations to impact heme oxygenase activity has been demonstrated [99,100], but no physiological consequences have been reported to date. Squalene monoxygenase is a flavoprotein that oxidizes squalene to 2,3-oxidosqualene (squalene epoxide) in the biosynthesis of cholesterol. Unlike other flavoprotein monooxygenases, it produces an epoxide rather than a hydroxylated product. Unlike other CYPOR acceptors, electrons are transferred to an FAD prosthetic group on this acceptor. CYPOR is the major electron donor to this enzyme; however, studies in CYPOR-null livers indicate that an alternate electron donor, which is not cytochrome b5, is present [101].

4.13 CYPOR domain movement and control of electron transfer Crystal structures of various CYPOR proteins, including rat [34], human [81], yeast CYPORs [102], and their various mutant proteins [74] clearly demonstrate that the

88

4 Structure and mechanism of NADPH-cytochrome P450 oxidoreductase

enzyme molecule consists of two flavin-binding domains, and that the two cofactors are juxtaposed to each other with their dimethyl benzene rings facing one another, with the closest distance being ~4 Å. As indicated above, the relatively slow rate of interflavin electron transfer suggests some form of gating mechanism. Furthermore, although this arrangement of the two flavin domains ( “closed” conformation) is optimal for electron transfer between the two flavins, i.e., from FAD to FMN, it is incompatible with interaction of the FMN domain with P450, the physiological electron acceptor. In the closed conformation, both the FMN dimethylbenzene ring and acidic residues located in the FMN domain and shown to affect electron transfer to cytochrome P450 by mutagenesis studies [57] are not solvent–exposed and so cannot interact with P450. In addition, the crystal structure of a complex between the heme and FMN-binding domains of flavocytochrome P450 BM3 provides a structural insight into how these two domains interact with each other. In this structure, the FMN dimethylbenzene ring is oriented toward the proximal face of the heme of P450 BM3, suggesting that CYPOR must interact with P450 in a different conformation, termed “open” than the closed conformation observed in the wild-type CYPOR crystal structure [86]. There are several lines of evidence demonstrating that the two flavin domains are mobile. Superposition of the structures of wild type and various point mutant structures of rat CYPOR has shown that the relative orientation of and distance between the two flavin domains are variable, with the closest flavin-flavin distance ranging from 3.9 to 5.8 Å, suggestive of small, but significant domain movements in solution [74]. Moreover, in the crystal structure of the flavoprotein subunit of E. coli sulfite reductase, electron density for the entire FMN domain is completely disordered, again suggesting movement of the FMN domain relative to the rest of the polypeptide [103]. The most direct demonstration of a large scale domain movement and a transition from a closed to an open conformation comes from the recent crystal structures of mutant CYPOR proteins. A CYPOR variant with a four amino acid deletion in the hinge region that links the two flavin domains has been crystallized in three different extended conformations (open state), in which the distance between FAD and FMN cofactors ranges 30–60 Å (򐂰Fig. 4.8) [51]. The mutant is defective in its ability to transfer electrons from FAD to FMN. However, when FMN is reduced chemically, the mutant CYPOR is capable of reducing P450 2B4. The authors infer that a similar domain movement controlled by the hinge occurs in the wild type enzyme during its catalytic cycle, enabling the FMN domain to adopt an open conformation capable of interacting with its physiological partner, cytochrome P450. Aigrain et al. have also seen an open conformation in the crystal structure of a yeast-human chimeric CYPOR [104]. A different approach has been used by Xia et al. [80], in which an engineered disulfide linkage between the two flavin domains locks CYPOR in a closed conformation unable to interact with P450. Indeed, the mutant exhibits substantially decreased interflavin electron transfer and is essentially unable to catalyze the P450-dependent monooxygenase activity. Reduction of the disulfide linkage restores the ability of the mutant to support both interflavin electron transfer and reduction of its redox partners, consistent with domain movements being required for the FMN domain of CYPOR to interact with both the FAD domain and P450, i.e., shuttling between the two redoxactive partners. In addition, several solution studies provide evidence for large domain movements of CYPOR in catalysis. Hay et al, demonstrate, using electron-electron double resonance

4.13 CYPOR domain movement and control of electron transfer Closing-opening transition

89

Closed conformation stabilized by S-S bond S S

FAD

FAD FMN FMN Flexible hinge

Fig. 4.8: The closed and open conformations of CYPOR. Top: schematic representation; Bottom: ribbon drawing. FAD and FMN are shown with stick models in purple. Relative orientations of FMN domain observed in WT (gold) and three open conformations of the hinge-deletion mutant, Mol A (red), Mol B (blue) and Mol C (grey) are shown (left panel, see [53] for details); the engineered S-S bond (green) is shown in the closed-form mutant of CYPOR (right panel, see [80] for details).

methods, that CYPOR exists in multiple conformations in a continuum of conformational landscape that is changed by nucleotide binding [105]. Using a combination of NMR and small-angle X-ray scattering (SAXS) methods, Ellis et al. [106] have shown that the oxidized human CYPOR exists in solution as a mixture of approximately equal amounts of two conformations, one consistent with the crystal structure (closed form) and one a more extended structure which presumably is required for interaction with its ET partners (open form). In addition, the relative contributions of each conformation at equilibrium are affected by the binding of NADP(H), with the nucleotide bound form favoring the closed form. On the other hand, Vincent et al. [107] have recently employed high resolution NMR measurements with residue-specific 15N relaxation and 1H-15N residual dipolar coupling data to show that oxidized CYPOR in solution in the absence of bound nucleotide exists in a unique and predominant conformation resembling the closed conformation observed in the crystal structure. However, at present more data are accumulating for the predominance of the closed form when the nucleotide is bound. Pudney et al. [108] have demonstrated, using a combination of fluorescence resonance energy transfer and stopped flow methods, that the open and closed states of CYPOR are correlated with key steps of in the catalytic cycle, i.e., NADPH binding induces closing of CYPOR and reduction of flavins and/or NADP+ release induces opening of CYPOR. In conclusion, there is mounting evidence that CYPOR must undergo several different types of conformational changes during catalysis. Hubbard et al. have shown that, upon binding of NADPH, the C-terminus of CYPOR including the aromatic residue Trp677

90

4 Structure and mechanism of NADPH-cytochrome P450 oxidoreductase

undergoes a significant conformational change [74]. In addition, comparison of the structure of CYPOR with and without bound NADP+ suggests that the movement of the loop containing Asp632 is necessary for binding of the nicotinamide moiety of NADPH to the re-face of the FAD isoalloxazine ring. Taking these results suggesting Trp677 and Asp632 loop movements together, Xia et al. [80] have proposed a scenario for coordinated conformational changes that occurs during NADPH binding, hydride transfer and NADP+ release (򐂰Fig. 4.6). Since the NADPH-induced Trp677 movement precedes hydride transfer, these steps and the subsequent interflavin electron transfer step must occur with the enzyme in the closed conformation. This is followed by a large scale domain movement to the open conformation that is necessary for interaction with P450; this movement must be tightly coordinated with electron transfer to prevent reactions with oxygen and production of superoxide. Thus, the linkage between the Asp632 loop and large scale domain movements provides a mechanism by which electron transfer is coupled to pyridine nucleotide binding/release and the oxidation state of the enzyme. It is most likely that a similar sequence of conformational changes that occurs during the CYPOR catalysis would also take place in other members of the diflavin oxidoreductase family. However, details of the mechanism by which the large scale domain movements coordinated to movements of loops and individual amino acids remain to be established.

4.14 Physiological functions of CYPOR and effects of CYPOR deficiency Currently, 59 human P450s have been identified, of which seven are mitochondrial and two, prostacyclin synthase and thromboxane A2 synthase, are not monooxygenases. Thus 50 human P450s in the endoplasmic reticulum likely depend on CYPOR for catalysis [109,110]. Of these, 20 participate in the biosynthesis of endobiotic substrates such as steroids and eicosanoids, with another 17 for the most part involved in the metabolism of xenobiotic compounds. Remarkably, 13 are orphan enzymes [109,110], for which no function has been identified. Evidence from a number of biological systems highlights the essential cellular functions of CYPOR-dependent P450 activity, with deletion of the entire CYPOR gene being lethal in yeast and C. elegans due to impaired P450-dependent biosynthesis of, respectively, ergosterol and an as-yet unidentified lipid [111–113]. Global deletion of murine microsomal CYPOR produces multiple developmental defects and embryonic lethality, with some embryos surviving to embryonic day 13.5, but the majority dying before embryonic day 11.5. Neural tube, cardiac, eye, limb, and vascular defects are seen in homozygous null embryos, as well as a generalized failure of development, which have been ascribed to defects in cholesterol and retinoid metabolism [114,115]. The ability to delete CYPOR in a tissue-specific manner in murine models has provided further insights into the diverse physiological functions of CYPOR in metabolism of both endogenous substrates and xenobiotics. Liver-specific ablation of CYPOR gives rise to massive lipid accumulation and hepatomegaly, with a decrease in levels of serum cholesterol and triglyerides that suggests defects in regulation of hepatic lipid transport. Consistent with the central role of hepatic CYPOR in drug metabolism, liver-specific ablation of CYPOR decreases metabolism and/or clearance of xenobiotics [116–119]. In addition, the central role of the hepatic P450 system in the balance between activation versus detoxification of xenobiotics is illustrated by the contrasting

4.15 Human CYPOR deficiency (PORD)

91

effects of CYPOR ablation in specific target tissues. Thus, acetaminophen toxicity in liver as well as in extrahepatic tissues is decreased in liver-specific CYPOR null animals as a consequence of decreased hepatic production of the toxic metabolites [116]. Lung tumor multiplicity in response to the tobacco carcinogen 4-(methylnitrosamino)1-(3-pyridyl)-1-butanone (NNK) is increased in liver-specific CYPOR-null animals due to decreased hepatic clearance, but decreased in lung-specific CYPOR null animals due to decreased target tissue activation. These results both confirm the essential role of pulmonary P450-mediated metabolic activation in NNK-induced lung cancer and demonstrate the importance of the hepatic cytochrome P450 system in NNK clearance [120]. A hypomorphic mouse model, in which CYPOR expression is decreased, but not abolished, may be a useful model for human CYPOR deficiency [121]. These mice exhibit phenotypes including reduced embryonic survival, decreases in circulating cholesterol, increases in hepatic P450 expression, and female infertility with abnormal serum testosterone and progesterone levels.

4.15 Human CYPOR deficiency (PORD) Mutations in human CYPOR that significantly disrupt cholesterol biosynthesis and/or steroidogenesis result in the recently described condition, POR deficiency (PORD), characterized by disordered steroidogenesis and Antley-Bixler syndrome (ABS)-like phenotype [122,123]. Clinical findings vary greatly in PORD, ranging from severe skeletal malformations observed in ABS [124]. and congenital adrenal hyperplasia to relatively mild hormonal dysregulation, probably due to the range of catalytic activities resulting from the various mutations. In general, the most severe phenotypes are associated with disruptions in ability of CYPOR to support P450-dependent activity [110]. CYP17A1 is known to be particularly sensitive to perturbations in electron transfer, with 17,20 lyase activity favored over 17α hydroxylase activity by high molar ratios of CYPOR to CYP17A1 [125,126]. Therefore, disordered steroidogenesis is a prominent feature of PORD and distinguishes the condition from ABS due to FGFR2 mutations, where steroidogenesis is normal. PORD is also associated with congenital adrenal hyperplasia without Antley-Bixler abnormalities. The Antley-Bixler phenotype resembles phenotypes associated with defects in cholesterol biosynthesis and is the result of the requirement for CYPOR by at least two enzymes involved in cholesterol biosynthesis: squalene monooxygenase and CYP51. Defects in retinoic acid metabolism as a consequence of decreased CYP26 activity may also play a role in developmental abnormalities [115]. Since the first report of four individuals with PORD deficiency [122], numerous reports worldwide have been published describing the varying phenotypes associated with this syndrome. A total of 1957 single-nucleotide polymorphisms have been described in the human CYPOR gene (www.ncbi.nlm.nih.gov/snp), encompassing 75 missense mutations, 5 missense mutations causing premature termination, 10 frameshift/deletion/duplication mutations, and 9 splice site variants (www.cypalleles.ki.se). Mutations affecting transcription have also been identified and interpreted in terms of the structure of the CYPOR promoter [127,128]. Sequence homology and mapping of missense mutations onto the CYPOR crystal structure have allowed identification of the nature of the defect for several missense mutations. Tyr181Asp, Arg457His, Tyr459His and Val492Glu mutations result in low

92

4 Structure and mechanism of NADPH-cytochrome P450 oxidoreductase

cytochrome c and CYP17A1 activities; Tyr181Asp cause decreased affinity for FMNbinding [49,129], and Arg457His and Val492Glu cause decreased FAD-binding affinity [122,130]. The results of Tyr181Asp and Tyr459His mutations are entirely consistent with the hypothesis that these aromatic residues are required for binding of FAD and FMN [81,130]. Crystal structures of Arg457His and Val492Glu variants together with human wild type have been determined [81]. The overall three-dimensional structures of both variants are similar to wild-type enzyme; however, there are subtle but significant differences. These include local disruption of hydrogen bonding and salt bridging involving the FAD pyrophosphate moiety that leads to weaker FAD binding, unstable protein, and loss of catalytic activity, all of which can be rescued by cofactor addition. Thus, riboflavin therapy may prevent or rescue CYPOR dysfunction for patients with these mutations [81].

4.16 Contribution of CYPOR to inter-individual variation in human drug metabolism Although mutations which dramatically decrease CYPOR activity are rare, other polymorphisms, such as A503V, are quite common and there is interest in the effects of these variations on inter-individual variability in drug metabolism [131–133]. The complexity of this effort is illustrated by studies on the A503V mutant, which has a minor allele frequency of ~27% [131–133]. In view of the high frequency of this allele, several studies have attempted to assess the contribution of this mutation to inter-individual variation in drug metabolism. Variable results are reported, depending on the P450, the substrate, and the assay systems employed [134–137]. These studies highlight the requirement for evaluation of CYPOR mutants with each P450, and probably each substrate, separately. It is rapidly becoming apparent that the effects of CYPOR variants on cytochrome P450mediated metabolism require examination of each P450-CYPOR pair and possibly each substrate, further complicated by the need for a membrane environment. Studies addressing this problem have utilized a variety of systems, including reconstituted systems, bacterial expression, and expression in insect cells. Two large studies of effects of CYPOR polymorphisms on drug metabolism in human liver samples [131,132] have failed to identify dramatic in vivo effects of CYPOR polymorphisms and it has been proposed that effects of single polymorphisms are complex and dependent upon the partner P450 as well as on factors such as age, smoking, diet. More recently, De Jonge et al. [138] have shown that renal allograft recipients expressing the CYP3A5*1 variant of CYP3A5 and the common A503V variant of CYPOR require higher loading doses of tacrolimus; both polymorphisms are necessary for this effect, suggesting that the A503V variant increases CYP3A5-mediated metabolism of tacrolimus. This study again emphasizes the need to consider CYPOR polymorphisms in context of each P450.

4.17 Unanswered questions and future directions Many exciting avenues of investigation remain with respect to the catalytic mechanisms of CYPOR and other members of the diflavin reductase family. CYPOR and the corresponding reductase domains of NOS and P450BM3 function to regulate the catalytic activities associated with their respective oxygenase acceptors via diverse intra- and intermolecular regulatory mechanisms [140]. The existence of

4.18 References

93

conformational changes in the course of catalysis has been established in principle, but the specific structures of these alternate conformations and the mechanisms by which transition between conformations are regulated remain to be elucidated. The mechanism of CYPOR interaction with P450 is of particular interest. Although work to date has identified features that are likely common to interaction of CYPOR with all of its physiological partners, it is likely that there are specific interactions for each that allow CYPOR to discriminate between multiple acceptors. The demonstration that common human CYPOR variants affect substrate metabolism in an acceptor-specific manner gives impetus to studies directed toward identification of the nature of CYPOR-cytochrome P450 complexes, the mechanisms for selection of a particular electron acceptor, and the pathway(s) for electron transfer.

Acknowledgements We thank Dr. Takashi Iyanagi of the Himeji Institute of Technology, University of Hyogo, Japan, for many helpful discussions and suggestions. Parts of the work performed in the Kim and Kasper/Shen laboratories were supported by National Institutes of Health Grants, GM52682 (JJPK) and CA22484 and CA07175 (McArdle Laboratory for Cancer Research).

Abbreviations CYPOR CYPOR/POR P450/CYP cyt c cyt b5 NOS FNR Fld FMN domain CD FAD domain P450BM3 ET MS MSR ER HO PORD

NADPH-cytochrome P450 oxidoreductase CYPOR gene cytochrome P450 cytochrome c cytochrome b5 nitric oxide synthase ferredoxin-NADP+ oxidoreductase flavodoxin FMN−containing flavodoxin-like domain connecting domain FAD-containing FNR-like domain plus the connecting domain Bacillus megaterium flavocytochrome P450BM3 electron transfer methionine synthase methionine synthase reductase endoplasmic reticulum heme oxygenase NADPH-cytochrome P450 oxidoreductase deficiency

4.18 References [1] Stuehr DJ, Tejero J, Haque MM. Structural and mechanistic aspects of flavoproteins: electron transfer through the nitric oxide synthase flavoprotein domain. FEBS J 2009;276:3959–74. [2] Daff S. NO synthase: structures and mechanisms. Nitric Oxide 2010;23:1–11.

94

4 Structure and mechanism of NADPH-cytochrome P450 oxidoreductase

[3] Feng C. Mechanism of nitric oxide synthase regulation: electron transfer and interdomain interactions. Coord Chem Rev 2012;256:393–411. [4] Ruettinger RT, Wen LP, Fulco AJ. Coding nucleotide, 5’ regulatory, and deduced amino acid sequences of P-450BM-3, a single peptide cytochrome P-450:NADPH-P-450 reductase from Bacillus megaterium. J Biol Chem 1989;264:10987–95. [5] Ostrowski J, Barber MJ, Rueger DC, Miller BE, Siegel LM, Kredich NM. Characterization of the flavoprotein moieties of NADPH-sulfite reductase from Salmonella typhimurium and Escherichia coli. Physicochemical and catalytic properties, amino acid sequence deduced from DNA sequence of cysJ, and comparison with NADPH-cytochrome P-450 reductase. J Biol Chem 1989;264:15796–808. [6] Leclerc D, Wilson A, Dumas R, Gafuik C, Song D, Watkins D, Heng HH, Rommens JM, Scherer SW, Rosenblatt DS, Gravel RA. Cloning and mapping of a cDNA for methionine synthase reductase, a flavoprotein defective in patients with homocystinuria. Proc Natl Acad Sci U S A 1998;95:3059–64. [7] Olteanu H, Banerjee, R. Human methionine synthase reductase, a soluble P450 reductaselike dual flavoprotein, is sufficient for NADPH-dependent methionine synthase activation. J Biol Chem 2001;276:35558–63. [8] Wolthers KR, Scrutton NS. Protein interactions in the human methionine synthase-methionine synthase reductase complex and implications for the mechanism of enzyme reactivation. Biochemistry 2007;46:6696–709. [9] Paine MJ, Garner AP, Powell D, Sibbald J, Sales M, Pratt N, Smith T, Tew, DG, Wolf CR. Cloning and characterization of a novel human dual flavin reductase. J Biol Chem 2000;275:1471–8. [10] Inui H, Yamaji R, Saidoh H, Miyatake K, Nakano Y, Kitaoka S. Pyruvate: NADP+ oxidoreductase from Euglena gracilis: limited proteolysis of the enzyme with trypsin. Arch Biochem Biophys 1991;286:270–6. [11] Nakazawa M, Inui H, Yamaji R, Yamamoto T, Takenaka S, Ueda M, Nakano Y, Miyatake K. The origin of pyruvate: NADP+ oxidoreductase in mitochondria of Euglena gracilis. FEBS Lett 2000;479:155–6. [12] Netz DJ, StumpFig M, Dore C, Muhlenhoff U, Pierik AJ, Lill R. Tah18 transfers electrons to Dre2 in cytosolic iron-sulfur protein biogenesis. Nat Chem Biol 2010;6:758–65. [13] Schacter BA, Nelson EB, Marver HS, Masters BS. Immunochemical evidence for an association of heme oxygenase with the microsomal electron transport system. J Biol Chem 1972;247:3601–07. [14] Guengerich FP. Reduction of cytochrome b5 by NADPH-cytochrome P450 reductase. Arch Biochem Biophys 2005;440:12. [15] Ono T, Takahashi K, Odani S, Konno H, Imai, Y. Purification of squalene epoxidase from rat liver microsomes. Biochem Biophys Res Commun 1980;96:7. [16] Pearson JT, Siu S, Meininger DP, Wienkers LC, Rock DA. In vitro modulation of cytochrome P450 reductase supported indoleamine 2,3-dioxygenase activity by allosteric effectors cytochrome b(5) and methylene blue. Biochemistry 2010;49:10. [17] Petersen JA, Ebel RE, O’Keeffe DH, Matsubara T, Estabrook RW. Temperature dependence of cytochrome P-450 reduction. A model for NADPH-cytochrome P-450 reductase: cytochrome P-450 interaction. J Biol Chem 1976;251:7. [18] Shephard EA, Phillips IR, Bayney RM, Pike SF, Rabin BR. Quantification of NADPH: cytochrome P-450 reductase in liver microsomes by a specific radioimmunoassay technique. Biochem J 1983;211:8. [19] Johnson EF, Stout CD. Structural diversity of human xenobiotic-metabolizing cytochrome P450 monooxygenases. Biochem Biophys Res Commun 2005;338:331–6. [20] Iyanagi T, Makino N, Mason HS. Redox properties of the reduced nicotinamide adenine dinucleotide phosphate-cytochrome P-450 and reduced nicotinamide adenine dinucleotidecytochrome b5 reductases. Biochemistry 1974;13:1701–10.

4.18 References

95

[21] Iyanagi T, Mason HS. Some properties of hepatic reduced nicotinamide adenine dinucleotide phosphate-cytochrome c reductase. Biochemistry 1973;12:2297–308. [22] Iyanagi T, Makina R, Anan FK. Studies on the microsomal mixed-function oxidase system: mechanism of action of hepatic NADPH-cytochrome P-450 reductase. Biochemistry 1981;20:1722–30. [23] Marcus RA, Sutin N. Biochim Biophys Acta 1985;811:265–322. [24] Chen HC, Swenson RP. Effect of the insertion of a glycine residue into the loop spanning residues 536–541 on the semiquinone state and redox properties of the flavin mononucleotidebinding domain of flavocytochrome P450BM-3 from Bacillus megaterium. Biochemistry 2008;47(52):13788–99. [25] Kasim M, Chen HC, Swenson RP. Functional characterization of the re-face loop spanning residues 536–541 and its interactions with the cofactor in the flavin mononucleotide-binding domain of flavocytochrome P450 from Bacillus megaterium. Biochemistry 2009;48:5131– 41. [26] Das A, Grinkova YV, Sligar SG. Redox potential control by drug binding to cytochrome P450 3A4. J Am Chem Soc 2007;129:13778–9. [27] Brenner S, Hay S, Munro AW, Scrutton NS. Inter-flavin electron transfer in cytochrome P450 reductase - effects of solvent and pH identify hidden complexity in mechanism. FEBS J 2008;275:4540–57. [28] Munro AW, Noble MA, Robledo L, Daff SN, Chapman SK. Determination of the redox properties of human NADPH-cytochrome P450 reductase. Biochemistry 2001;40:1956–63. [29] Vermilion JL, Coon MJ. Identification of the high and low potential flavins of liver microsomal NADPH-cytochrome P-450 reductase. J Biol Chem 1978;253(24):8812–9. [30] Iyanagi T, Suzaki T, Kobayashi S. Oxidation-reduction states of pyridine nucleotide and cytochrome P-450 during mixed-function oxidation in perfused rat liver. J Biol Chem 1981;256:12933–9. [31] Yasukochi Y, Peterson JA, Masters BS. NADPH-cytochrome c (P-450) reductase. Spectrophotometric and stopped flow kinetic studies on the formation of reduced flavoprotein intermediates. J Biol Chem 1979;254:7097–104. [32] Yamamoto K, Kimura S, Shiro Y, Iyanagi T. Interflavin one-electron transfer in the inducible nitric oxide synthase reductase domain and NADPH-cytochrome P450 reductase. Arch Biochem Biophys 2005;440:65–78. [33] Porter TD, Kasper CB. NADPH-cytochrome P-450 oxidoreductase: flavin mononucleotide and flavin adenine dinucleotide domains evolved from different flavoproteins. Biochemistry 1986;25:1682–7. [34] Wang M, Roberts DL, Paschke R, Shea TM, Masters BS, Kim JJ. Three-dimensional structure of NADPH-cytochrome P450 reductase: prototype for FMN- and FAD-containing enzymes. Proc Natl Acad Sci U S A 1997;94:8411–6. [35] Karplus PA, Daniels MJ, Herriott JR. Atomic structure of ferredoxin-NADP+ reductase: prototype for a structurally novel flavoenzyme family. Science 1991;251:7. [36] Smith GC, Tew DG, Wolf CR. Dissection of NADPH-cytochrome P450 oxidoreductase into distinct functional domains. Proc Natl Acad Sci U S A 1994;91:8710–4. [37] Hodgson AV, Strobel HW. Characterization of the FAD binding domain of cytochrome P450 reductase. Arch Biochem Biophys 1996;325:99–106. [38] Black SD, Coon MJ. Structural features of liver microsomal NADPH-cytochrome P-450 reductase. Hydrophobic domain, hydrophilic domain, and connecting region. J Biol Chem 1982;257:5929–38. [39] Porter TD, Kasper CB. Coding nucleotide sequence of rat NADPH-cytochrome P-450 oxidoreductase cDNA and identification of flavin-binding domains. Proc Natl Acad Sci U S A 1985;82:973–7. [40] Taniguchi H, Imai Y, Iyanagi T, Sato R. Interaction between NADPH-cytochrome P-450 reductase and cytochrome P-450 in the membrane of phosphatidylcholine vesicles. Biochimica et biophysica acta 1979;550:341–56.

96

4 Structure and mechanism of NADPH-cytochrome P450 oxidoreductase

[41] Black S, French JS, Williams JH, Jr, Coon MJ. Biochem Biophys Res Commun 1979;91:8. [42] Cojocaru V, Balali-Mood K, Sansom MS, Wade RC. Structure and dynamics of the membranebound cytochrome P450 2C9. PLoS Comput Biol 2012;7:e1002152. [43] Gilep A, Guryev OL, Usanov SA, Estabrook RW. An enzymatically active chimeric protein containing the hydrophilic form of NADPH-cytochrome P450 reductase fused to the membrane-binding domain of cytochrome b5. Biochem Biophys Res Commun 2001:5. [44] Venkateswarlu K, Lamb DC, Kelly DE, Manning NJ, Kelly SL. The N-terminal membrane domain of yeast NADPH-cytochrome P450 (CYP) oxidoreductase is not required for catalytic activity in sterol biosynthesis or in reconstitution of CYP activity. J Biol Chem 1998;273:5. [45] Das A, Sligar SG. Modulation of the cytochrome P450 reductase redox potential by the phospholipid bilayer. Biochemistry 2009;48:9. [46] Blanck J, Smettan G, Ristau O, Ingelman-Sundberg M, Ruckpaul K. Mechanism of rate control of the NADPH-dependent reduction of cytochrome P-450 by lipids in reconstituted phospholipid vesicles. Eur J Biochem 1984;144:5. [47] Vermilion JL, Ballou DP, Massey V, Coon MJ. Separate roles for FMN and FAD in catalysis by liver microsomal NADPH-cytochrome P-450 reductase. J Biol Chem 1981;256:266–77. [48] Shen AL, Porter T, Wilson TE, Kasper CB. Structural analysis of the FMN binding domain of NADPH-cytochrome P-450 oxidoreductase by site-directed mutagenesis. J Biol Chem 1989;264(13):7584–9. [49] Marohnic CC, Panda SP, McCammon K, Rueff J, Masters BS, Kranendonk M. Human cytochrome P450 oxidoreductase deficiency caused by the Y181D mutation: molecular consequences and rescue of defect. Drug Metab Dispos 2010;38(2):332–40. [50] Nicolo C, Fluck CE, Mullis PE, Pandey AV. Restoration of mutant cytochrome P450 reductase activity by external flavin. Mol Cell Endocrinol 2010;321(2):245–52. [51] Hamdane D, Xia C, Im SC, Zhang H, Kim JJ, Waskell L. Structure and function of an NADPHcytochrome P450 oxidoreductase in an open conformation capable of reducing cytochrome P450. J Biol Chem 2009;284:11374–84. [52] Bridges A, Gruenke L, Chang YT, Vakser IA, Loew G, Waskell L. Identification of the binding site on cytochrome P450 2B4 for cytochrome b5 and cytochrome P450 reductase. J Biol Chem 1998;273:17036–49. [53] Im SC, Waskell L. The interaction of microsomal cytochrome P450 2B4 with its redox partners, cytochrome P450 reductase and cytochrome b(5). Arch Biochem Biophys 2011;507:144–53. [54] Johnson EF. The 2002 Bernard B. Brodie Award lecture: deciphering substrate recognition by drug-metabolizing cytochromes P450. Drug Metab Dispos 2003;31:1532–40. [55] Nisimoto Y. Localization of cytochrome c-binding domain on NADPH-cytochrome P-450 reductase. J Biol Chem 1986;261:14232–9. [56] Nisimoto Y, Otsuka-Murakami H. Cytochrome b5, cytochrome c, cytochrome P-450 interactions with NADPHcytochrome P-450 reductase in phospholipid vesicles. Biochemistry 1988;27:5869–76. [57] Shen AL, Kasper CB. Role of acidic residues in the interaction of NADPH-cytochrome P450 oxidoreductase with cytochrome P450 and cytochrome c. J Biol Chem 1995; 270(46):27475–80. [58] Strobel HW, Hodgson AV, Shen SJ. NADPH cytochrome P450 reductase and its structural and functional domains, in 2nd ed. Cytochrome P450: Structure, Mechamisn, and Biochemistry, ed. PR. Ortiz Montellano:1995, Plenum Press, New York. [59] Jang HH, Jamakhandi AP, Sullivan SZ, Yun CH, Hollenberg PF, Miller GP. Beta sheet 2-alpha helix C loop of cytochrome P450 reductase serves as a docking site for redox partners. Biochimica Biophysica Acta 2010;1804:9. [60] Hlavica P, Schulze J, Lewis DFV. Functional interaction of cytochrome P450 with its redox partners: a critical assessment and update of the topology of predicted contact regions. J Inorg Biochem 2003;96:19.

4.18 References

97

[61] Zhao Q, Modi S, Smith G, Paine M, McDonagh PD, Wolf CR, Tew D, Lian LY, Roberts GC, Driessen HP. Crystal structure of the FMN-binding domain of human cytochrome P450 reductase at 1.93 A resolution. Protein Sci 1999;8:298–306. [62] Bumpus NN, Hollenberg PF. Cross-linking of human cytochrome P450 2B6 to NADPH-cyto chrome P450 reductase: Identification of a potential site of interaction. J Inorg Biochem 2010;104:4. [63] Kenaan C, Zhang H, Shea EV, Hollenberg PF. Uncovering the role of hydrophobic residues in cytochrome P450-cytochrome P450 reductase interactions. Biochemistry 2011;50:3957–67. [64] Miwa GT, West SB, Huang MT, Lu AY. Studies on the association of cytochrome P-450 and NADPHcytochrome c reductase during catalysis in a reconstituted hydroxylating system. J Biol Chem 1979;254:5695–700. [65] Backes WL, Kelley RW. Organization of multiple cytochrome P450s with NADPH-cytochrome P450 reductase in membranes. Pharmacol Ther 2003;98:221–33. [66] Jamakhandi AP, Kuzmic P, Sanders DE, Miller GP. Global analysis of protein-protein interactions reveals multiple CYP2E1-reductase complexes. Biochemistry 2007;46:10192–201. [67] Kurzban GP, Strobel HW. Purification of flavin mononucleotide-dependent and flavin-adenine dinucleotide-dependent reduced nicotinamide-adenine dinucleotide phosphate-cytochrome P-450 reductase by high-performance liquid chromatography on hydroxyapatite. J Chromatogr 1986;358:296–301. [68] Narayanasami R, Nishimura JS, McMillan K, Roman LJ, Shea TM, Robida AM, Horowitz PM, Masters BS. The influence of chaotropic reagents on neuronal nitric oxide synthase and its flavoprotein module. Urea and guanidine hydrochloride stimulate NADPH-cytochrome c reductase activity of both proteins. Nitric Oxide 1997;1:39–49. [69] Narayanasami R, Horowitz PM, Masters BS. Flavin-binding and protein structural integrity studies on NADPH-cytochrome P450 reductase are consistent with the presence of distinct domains. Arch Biochem Biophys 1995;316:267–74. [70] Shen AL, Kasper CB. Differential contributions of NADPH-cytochrome P450 oxidoreductase FAD binding site residues to flavin binding and catalysis. J Biol Chem 2000;275:41087–91. [71] Zhang H, Gruenke L, Arscott D, Shen A, Kasper C, Harris DL, Glavanovich M, Johnson R, Waskell L. Determination of the rate of reduction of oxyferrous cytochrome P450 2B4 by 5-deazariboflavin adenine dinucleotide T491V cytochrome P450 reductase. Biochemistry 2003;42:6804–13. [72] Shen AL, Kasper CB. Role of Ser457 of NADPH-cytochrome P450 oxidoreductase in catalysis and control of FAD oxidation-reduction potential. Biochemistry 1996;35:9451–9. [73] Shen AL, Sem DS, Kasper CB. Mechanistic studies on the reductive half-reaction of NADPHcytochrome P450 oxidoreductase. J Biol Chem 1999;274:5391–8. [74] Hubbard PA, Shen AL, Paschke R, Kasper CB, Kim JJ. NADPH-cytochrome P450 oxidoreductase. Structural basis for hydride and electron transfer. J Biol Chem 2001;276:29163–70. [75] Deng Z, Aliverti A, Zanetti G, Arakaki AK, Ottado J, Orellano EG, Calcaterra NB, Ceccarelli EA, Carrillo N, Karplus PA. A productive NADP+ binding mode of ferredoxin-NADP+ reductase revealed by protein engineering and crystallographic studies. Nat Struct Biol 1999;6:847–53. [76] Tejero J, Perez-Dorado I, Maya C, Martinez-Julvez M, Sanz-Aparicio J, Gomez-Moreno C, Hermoso JA, Medina M. C-terminal tyrosine of ferredoxin-NADP+ reductase in hydride transfer processes with NAD(P)+/H. Biochemistry 2005;44:13477–90. [77] Sem DS, Kasper CB. Interaction with arginine 597 of NADPH-cytochrome P-450 oxidoreductase is a primary source of the uniform binding energy used to discriminate between NADPH and NADH. Biochemistry 1993;32:11548–58. [78] Gutierrez A, Doehr O, Paine M, Wolf CR, Scrutton NS, Roberts GC. Trp-676 facilitates nicotinamide coenzyme exchange in the reductive half-reaction of human cytochrome P450 reductase: properties of the soluble W676H and W676A mutant reductases. Biochemistry 2000;39:15990–9.

98

4 Structure and mechanism of NADPH-cytochrome P450 oxidoreductase

[79] Elmore CL, Porter TD. Modification of the nucleotide cofactor-binding site of cytochrome P-450 reductase to enhance turnover with NADH in vivo. J Biol Chem 2002;277:48960–4. [80] Xia C, Hamdane D, Shen AL, Choi V, Kasper CB, Pearl NM, Zhang H, Im SC, Waskell L, Kim JJ. Conformational changes of NADPH-cytochrome P450 oxidoreductase are essential for catalysis and cofactor binding. J Biol Chem 2011;286:16246–60. [81] Xia C, Panda SP, Marohnic CC, Martasek P, Masters BS, Kim JJ. Structural basis for human NADPH-cytochrome P450 oxidoreductase deficiency. Proc Natl Acad Sci U S A 2011; 108:13486–91. [82] Page CC, Moser CC, Chen X, Dutton PL. Natural engineering principles of electron tunnelling in biological oxidation-reduction. Nature 1999;402:47–52. [83] Bhattacharyya AK, Lipka JJ, Waskell L, Tollin G. Laser flash photolysis studies of the reduction kinetics of NADPH: cytochrome P-450 reductase. Biochemistry 1991;30:759–65. [84] Gutierrez A, Paine M, Wolf CR, Scrutton NS, Roberts GC. Relaxation kinetics of cytochrome P450 reductase: internal electron transfer is limited by conformational change and regulated by coenzyme binding. Biochemistry 2002;41:4626–37. [85] Sem DS, Kasper CB. Effect of ionic strength on the kinetic mechanism and relative rate limitation of steps in the model NADPH-cytochrome P450 oxidoreductase reaction with cytochrome c. Biochemistry 1995;34:12768–74. [86] Sevrioukova IF, Li H, Zhang H, Peterson JA, Poulos TL. Structure of a cytochrome P450-redox partner electron-transfer complex. Proc Natl Acad Sci U S A 1999;96:1863–8. [87] Gunsalus IC, Pederson TC, Sligar SG. Oxygenase-catalyzed biological hydroxylations. Ann Rev Biochem 1975;44:377–407. [88] Denisov IG, Makris TM, Sligar SG, Schlichting I. Structure and chemistry of cytochrome P450. Chem Rev 2005;105:2253–77. [89] Schlichting I, Chu K. Trapping intermediates in the crystal: ligand binding to myoglobin. Cur Opin Struct Biol 2000;10:744–52. [90] Kikuchi G, Yoshida T, Noguchi M. Heme oxygenase and heme degradation. Biochem Biophy Res Commun 2005;338:558–67. [91] Guengerich FP. Rate-limiting steps in cytochrome P450 catalysis. Biol Chem 2002;383: 1553–64. [92] Oprian DD, Vatsis KP, Coon MJ. Kinetics of reduction of cytochrome P-450LM4 in a reconstituted liver microsomal enzyme system. J Biol Chem 1979;254:8895–902. [93] Imai Y, Sato R, Iyanagi T. Rate-limiting step in the reconstituted microsomal drug hydroxylase system. J Biochem 1977;82:1237–46. [94] Finn RD, McLaughlin LA, Ronseaux S, Rosewell I, Houston JB, Henderson CJ, Wolf CR. Defining the in vivo role for cytochrome b5 in cytochrome P450 function through the conditional hepatic deletion of microsomal cytochrome b5. J Biol Chem 2008;283:31385–93. [95] Yamazaki H, Johnson WW, Ueng YF, Shimada T, Guengerich FP. Lack of electron transfer from cytochrome b5 in stimulation of catalytic activities of cytochrome P450 3A4. Characterization of a reconstituted cytochrome P450 3A4/NADPH-cytochrome P450 reductase system and studies with apo-cytochrome b5. J Biol Chem 1996;271:27438–44. [96] Porter TD. The roles of cytochrome b5 in cytochrome P450 reactions. J Biochem Mol Toxicol 2002;16:311–6. [97] Zhang H, Hamdane D, Im SC, Waskell L. Cytochrome b5 inhibits electron transfer from NADPH-cytochrome P450 reductase to ferric cytochrome P450 2B4. J Biol Chem 2008; 238:5217–25. [98] Chen Z, Banerjee R. Purification of soluble cytochrome b5 as a component of the reductive activation of porcine methionine synthase. J Biol Chem 1998;273:26248–55. [99] Marohnic CC, Huber WJ III, Patrick Connick J, Reed JR, McCammon K, Panda SP, Martásek P, Backes WL, Masters BS. Mutations of human cytochrome P450 reductase differentially modulate heme oxygenase-1 activity and oligomerization. Arch Biochem Biophys 2011;513(1):42–50.

4.18 References

99

[100] Pandey AV, Fluck CE, Mullis PE. Altered heme catabolism by heme oxygenase-1 caused by mutations in human NADPH cytochrome P450 reductase. Biochem Biophy Res Commun 2010;400:374–8. [101] Li L, Porter TD. Hepatic cytochrome P450 reductase-null mice reveal a second microsomal reductase for squalene monooxygenase. Arch Biochem Biophys 2007;461:9. [102] Lamb DC, Kim Y, Yermalitskaya LV, Yermalitsky VN, Lepesheva GI, Kelly SL, Waterman MR, Podust LM. A second FMN binding site in yeast NADPH-cytochrome P450 reductase suggests a mechanism of electron transfer by diflavin reductases. Structure 2006;14:51–61. [103] Gruez A, Pignol D, Zeghouf M, Coves J, Fontecave M, Ferrer JL, Fontecilla-Camps JC. Four crystal structures of the 60 kDa flavoprotein monomer of the sulfite reductase indicate a disordered flavodoxin-like module. J Mol Biol 2000;299:199–212. [104] Aigrain L, Pompon D, Morera S, Truan G. Structure of the open conformation of a functional chimeric NADPH cytochrome P450 reductase. EMBO Rep 2009;10:742–7. [105] Hay S, Brenner S, Khara B, Quinn AM, Rigby SE, Scrutton NS. Nature of the energy landscape for gated electron transfer in a dynamic redox protein. J Am Chem Soc 2010;132: 9738–45. [106] Ellis J, Gutierrez A, Barsukov IL, Huang WC, Grossmann JG, Roberts GC. Domain motion in cytochrome P450 reductase: conformational equilibria revealed by NMR and small-angle x-ray scattering. J Biol Chem 2009;284:36628–37. [107] Vincent B, Morellet N, Fatemi F, Aigrain L, Truan G, Guittet E, Lescop E. The closed and compact domain organization of the 70-kDa human cytochrome P450 reductase in its oxidized state as revealed by NMR. J Mol Biol 2012;420:296–309. [108] Pudney CR, Khara B, Johannissen LO, Scrutton NS. Coupled motions direct electrons along human microsomal P450 Chains. pLOS Biology 2011;Dec 9, e1001222. [109] Guengerich FP, Cheng Q. Orphans in the human cytochrome P450 superfamily: approaches to discovering functions and relevance in pharmacology. Pharmacol Rev 2011;63(3): 684–99. [110] Flück CE, Pandey AV. Clinical and biochemical consequences of p450 oxidoreductase deficiency. Endocrine development 2011;20:63–79. [111] Niedenthal R, Riles L, Guldener U, Klein S, Johnston M, Hegemann J. Systematic analysis of S. cerevisiae chromosome VIII genes. Yeast 1999;15:1775–96. [112] Benenati G, Penkov S, Müller-Reichert T, Entchev EV, Kurzchalia TV. Two cytochrome P450s in Caenorhabditis elegans are essential for the organization of eggshell, correct execution of meiosis and the polarization of embryo. Mechanisms of development 2009;126:382–93. [113] Rappleye CA, Tagawa A, Le Bot N, Ahringer J, Aroian RV. Involvement of fatty acid pathways and cortical interaction of the pronuclear complex in Caenorhabditis elegans embryonic polarity. BMC developmental biology 2003;Oct 3;3:8. [114] Shen AL, O’Leary KA, Kasper CB. Association of multiple developmental defects and embryonic lethality with loss of microsomal NADPH-cytochrome P450 oxidoreductase. J Biol Chem 2002;277:6536–41. [115] Otto DM, Henderson CJ, Carrie D, Davey M, Gundersen TE, Blomhoff R, Adams RH, Tickle C, Wolf CR. Identification of novel roles of the cytochrome p450 system in early embryogenesis: effects on vasculogenesis and retinoic acid homeostasis. Mol Cell Biol 2003;23:6103–16. [116] Gu J, Cui H, Behr M, Zhang L, Zhang QY, Yang W, Hinson JA, Ding X. In vivo mechanisms of tissue-selective drug toxicity: effects of liver-specific knockout of the NADPH-cytochrome P450 reductase gene on acetaminophen toxicity in kidney, lung, and nasal mucosa. Mol Pharm 2005;67:623–30. [117] Gu J, Weng Y, Zhang QY, Cui H, Behr M, Wu L, Yang W, Zhang L, Ding X. Liver-specific deletion of the NADPH-cytochrome P450 reductase gene: impact on plasma cholesterol homeostasis and the function and regulation of microsomal cytochrome P450 and heme oxygenase. J Biol Chem 2003;278:25895–901.

100

4 Structure and mechanism of NADPH-cytochrome P450 oxidoreductase

[118] Henderson CJ, Otto DM, Carrie D, Magnuson MA, McLaren AW, Rosewell I, Wolf CR. Inactivation of the hepatic cytochrome P450 system by conditional deletion of hepatic cytochrome P450 reductase. J Biol Chem 2003;278:13480–6. [119] Stiborová M, Arlt VM, Henderson CJ, Wolf CR, Kotrbová V, Moserová M, Hudecek J, Phillips DH, Frei E. Role of hepatic cytochromes P450 in bioactivation of the anticancer drug ellipticine: studies with the hepatic NADPH:cytochrome P450 reductase null mouse. Toxicol Appl Pharm 2008;226:318–27. [120] Weng Y, Fang C, Turesky RJ, Behr M, Kaminsky L, Ding X. Determination of the role of target tissue metabolism in lung carcinogenesis using conditional cytochrome P450 reductase-null mice. Cancer Res 2007;67:7825–32. [121] Wu L, Gu J, Cui H, Zhang QY, Behr M, Fang C, Weng Y, Kluetzman K, Swiatek PJ, Yang W, Kaminsky L, Ding X. Transgenic mice with a hypomorphic NADPH-cytochrome P450 reductase gene: effects on development, reproduction, and microsomal cytochrome P450. J Pharmacol Exp Ther 2005;312:35–43. [122] Flück CE, Tajima T, Pandey AV, Arlt W, Okuhara K, Verge CF, Jabs EW, Mendonça BB, Fujieda K, Miller WL. Mutant P450 oxidoreductase causes disordered steroidogenesis with and without Antley-Bixler syndrome. Nat Genet 2004;36(3):228–30. [123] Fukami MHR, Nagai T, Tanaka T, Naiki Y, Sato N, Okuyama T, Nakai H, Soneda S, Tachibana K, Matsuo N, Sato S, Homma K, Nishimura G, Hasegawa T, Ogata T. Cytochrome P450 oxidoreductase gene mutations and Antley-Bixler syndrome with abnormal genitalia and/or impaired steroidogenesis: molecular and clinical studies in 10 patients. J Clin Endocr Metab 2005;90(1):414–26. [124] Antley R, Bixler D. Trapezoidocephaly, midfacial hypoplasia and cartilage abnormalities with multiple synostoses and skeletal fractures. Birth Defects Orig Artic Ser 1975;11:397– 401. [125] Yanagibashi K, Hall PF. Role of electron transport in the regulation of the lyase activity of C21 side-chain cleavage P-450 from porcine adrenal and testicular microsomes. J Biol Chem 1986;261:8429–33. [126] Lin D, Black SM, Nagahama Y, Miller WL. Steroid 17 alpha-hydroxylase and 17,20-lyase activities of P450c17: contributions of serine106 and P450 reductase. Endocrinology 1993;132:2498–506. [127] Tee MK, Huang N, Damm I, Miller WL. Transcriptional regulation of the human P450 oxidoreductase gene: hormonal regulation and influence of promoter polymorphisms. Mol Endocrinol 2011;25(5):715–31. [128] Soneda S, Yazawa T, Fukami M, Adachi M, Mizota M, Fujieda K, Miyamoto K, Ogata T. Proximal promoter of the cytochrome P450 oxidoreductase gene: identification of microdeletions involving the untranslated exon 1 and critical function of the SP1 binding sites. J Clin Endocrinol and Metab 2011;96:E1881–7. [129] Huang N, Pandey AV, Agrawal V, Reardon W, Lapunzina PD, Mowat D, Jabs EW, Van Vliet G, Sack J, Fluck CE, Miller WL. Diversity and function of mutations in p450 oxidoreductase in patients with Antley-Bixler syndrome and disordered steroidogenesis. Am J Hum Genet 2005;76:729–49. [130] Marohnic CC, Panda SP, Martásek P, Masters BS. Diminished FAD binding in the Y459H and V492E Antley-Bixler syndrome mutants of human cytochrome P450 reductase. J Biol Chem 2006;281(47):35975–82. [131] Hart SNWS, Nakamoto K, Wesselman C, Li Y, Zhong XB. Genetic polymorphisms in cyto chrome P450 oxidoreductase influence microsomal P450-catalyzed drug metabolism. Pharmacogenetics and genomics 2008;18(1):11–24. [132] Gomes AM, Winter S, Klein K, Turpeinen M, Schaeffeler E, Schwab M, Zanger UM. Pharmacogenomics of human liver cytochrome P450 oxidoreductase: multifactorial analysis and impact on microsomal drug oxidation. Pharmacogenomics 2009;10(4):579–99.

4.18 References

101

[133] Huang N, Agrawal V, Giacomini KM, Miller WL. Genetics of P450 oxidoreductase: sequence variation in 842 individuals of four ethnicities and activities of 15 missense mutations. Proc Natl Acad Sci U S A 2008;105:1733–8. [134] Agrawal V, Huang N, Miller WL. Pharmacogenetics of P450 oxidoreductase: effect of sequence variants on activities of CYP1A2 and CYP2C19. Pharmacogenetics and genomics 2008;569–76. [135] Agrawal V, Choi JH, Giacomini KM, Miller WL. Substrate-specific modulation of CYP3A4 activity by genetic variants of cytochrome P450 oxidoreductase. Pharmacogenetics and genomics 2010;20:611–8. [136] Sandee D, Morrissey K, Agrawal V, Tam, HK, Kramer MA, Tracy TS, Giacomini KM, Miller WL. Effects of genetic variants of human P450 oxidoreductase on catalysis by CYP2D6 in vitro. Pharmacogenomics 2010;20:677–86 . [137] Subramanian M, Agrawal V, Sandee D, Tam, HK, Miller WL, Tracy TS. Effect of P450 oxidoreductase variants on the metabolism of model substrates mediated by CYP2C9.1, CYP2C9.2, and CYP2C9.3. Pharmacogenetics and genomics 2012;22:590–7. [138] de Jonge H, Metalidis C, Naesens M, Lambrechts D, Kuypers DR. The P450 oxidoreductase *28 SNP is associated with low initial tacrolimus exposure and increased dose requirements in CYP3A5-expressing renal recipients. Pharmacogenomics 2011;12:1281–91. [139] Garcin ED, Bruns CM, Lloyd SJ, Hosfield DJ, Tiso M, Gachhui R, Stuehr DJ, Tainer JA, Getzoff ED. Structural basis for isozyme-specific regulation of electron transfer in nitric-oxide synthase. J Biol Chem 2004;279:3791. [140] Iyanagi T, Xia C, Kim J-JP. NADPH-cytochrome P450 oxidoreductase: Prototypic member of the diflavin reductase family. Arch Biochem Biophys 2012; 528:72–89.

5 The xanthine oxidoreductase enzyme family: xanthine dehydrogenase, xanthine oxidase, and aldehyde oxidase Takeshi Nishino, Ken Okamoto, Bryan T. Eger and Emil F. Pai

Abstract The molybdenum-containing iron-sulfur flavoprotein enzyme family consists of xanthine dehydrogenase, xanthine oxidase, and aldehyde oxidase. This class of enzymes still poses various important questions, among them: the detailed mechanism of the unique molybdenum-dependent hydroxylation reaction, the mechanism of electron transfer among the several redox-active centers of the enzyme, and the conformational changes of the protein chain that influence the flavin environment. In addition to posing these basic enzymological problems, each of the enzymes also attracts the interest of researchers in biomedicine and pharmacology, given their importance as drug targets and in drug metabolism. This chapter focuses on more recent advances in our understanding of these macromolecules, including the reaction mechanism of hydroxylation, the conformational transition between the dehydrogenase and oxidase forms and the mechanism of inhibition of drugs that have been introduced to clinical usage.

5.1 Introduction The molybdenum-containing iron-sulfur flavoprotein enzyme family consists of xanthine dehydrogenase (XDH), xanthine oxidase (XO), and aldehyde oxidase (AO) [1,2]. XDH and XO represent two forms of the same enzyme and are commonly referred to as xanthine oxidoreductases (XOR) [3]. The protein is synthesized as the dehydrogenase form but can often be converted to the oxidase form either reversibly through formation of disulfide bonds or irreversibly by proteolysis [3]. The former process possibly forms part of physiologically relevant regulation mechanisms, while the latter one most probably represents a molecular accident since the enzyme can be isolated as the XDH form if the purification is performed appropriately so as to avoid proteolysis [4–6]. All members of the enzyme family can catalyze the hydroxylation of aldehydes but only the XORs are capable of accepting various heterocycles as substrates, including the physiologically relevant purines hypoxanthine and xanthine [1–3]. In 1902, Schardinger [7] described an aldehyde oxidase activity in bovine milk that, as we now know, was XO. More than 100 years of continuous research, including many benchmark experiments, have made this enzyme the archetype for the entire class of molybdenum-containing flavoproteins [8]. In contrast, much less is known about the AO enzymes, with work focusing mainly on extensive gene level studies [9,10] due to the difficulties in obtaining sufficient quantities of enzyme for biochemical studies [9–11]. The bacterial versions of AO do not possess the FAD-binding domain, and only the

104

5 The xanthine oxidoreductase enzyme family

eukaryotic members display sequence similarity throughout the full length of the amino acid sequence [2,12]. In this, the eukaryotic AOs are more similar to the eukaryotic XOR enzymes than to the bacterial enzymes [2,10,12]. So far, no dehydrogenase version of an AO has been found, all utilize O2 as the oxidizing substrate rather than the NAD+ used by the dehydrogenases. The specificities for reducing substrate of the three enzymes partially overlap, with each capable of hydroxylating a distinct subset of a wide range of aldehydes and aromatic heterocycles. In plants, AO enzymes participate in the biosynthesis of plant hormones [13]. The physiological substrates of mammalian AOs are not known, although their involvement in drug metabolism is well-established [14]; there is some evidence that they may be involved in the biosynthesis of retinoic acid, at least in certain tissues [15]. Structure-based sequence comparisons have identified residues in the vicinity of the active site molybdenum center of AO that differ in XOR and most probably are the determinants of substrate preference exhibited by the family members [11,16–18]. However, a glutamic acid residue thought to represent an essential catalytic base is strictly conserved in AOs and XORs, pointing to a common catalytic mechanism for all family members [18,19]. In the following discussion, we will therefore concentrate on the much better known mechanistic details of the XOR enzymes. The XOR class of enzymes still poses various important questions, e.g. the detailed mechanism of the rather unique molybdenum-dependent hydroxylation reaction [2,18] is in need of clarification as are the electron transfer mechanism within multi-redox centers [18,19], and the way protein conformational changes influence the flavin environment, determining whether reduced flavin will react with NAD+ or with molecular oxygen [20]. In addition to posing these basic enzymological problems, each of the enzymes also attracts the interest of researchers in biomedicine and pharmacology. The XORs catalyze two hydroxylation steps in the purine degradation pathway, i.e., the transformation of hypoxanthine to xanthine and of xanthine to uric acid, and thus are targets of drugs against hyperuricemia or gout [21,22]. At the same time, all enzymes of this family can form superoxide or hydrogen peroxide on reaction with O2 [23], they thus are of importance wherever the physiological and pathological effects of reactive oxygen species come into play. Finally, some important aspects of the physiological roles played by these enzymes require further clarification, e.g., the role of XO in the secretion of fat globules in breast milk [25] or the role of various organ-specific aldehyde-oxidizing isozymes [10]. As the members of this enzyme family attract rather broad interest a significant number of reviews have been published on the subject in the past. Therefore, this paper attempts to focus more on aspects of more recent advances in our understanding of these macromolecules.

5.2 Overall structures Although XDH, XO, and AO, all display similar molecular weights and share an identical complement of redox-active cofactors – two [2Fe-2S] clusters of the spinach ferredoxin variety and FAD, as well as the active site molybdenum center – the subunit composition differs in eukaryotic and prokaryotic enzymes [2]. Presently, the crystal structures of native bovine XDH and XO as well as of the recombinantly produced

5.3 Reaction mechanism

105

rat and human XDHs have been reported [20]. In all these mammalian XOR crystal structures, the protein subunits are arranged as identical dimers that display a distinct butterfly shape with the interface formed between the narrower sides of the elongated subunits [16,26,27]. The dimensions of the entire enzyme molecule are about 155 Å x 90 Å x 70 Å (򐂰Fig. 5.1). The shortest distance between atoms of cofactors from different subunits is greater than 50 Å, restricting catalytic electron flow to within a single subunit. Each monomer is composed of three subdomains. The small N-terminal domain (residues 1 to 165 in the bovine milk enzyme) harbors both two iron-sulfur centers (historically designated Fe/S I and Fe/S II on the basis of their EPR signals) and is connected to the second, FAD-containing domain (residues 226 to 531, colored yellow in 򐂰Fig. 5.1) by a long, partially disordered segment consisting of residues 166 to 225. The FAD domain is in turn connected to a third, C-terminal domain by another extended segment (residues 532 to 589), which is also partially disordered; residues 532 to 536 are not represented by electron density in the bovine crystal structure, but are partially visible in the density maps of a mutant of rat XDH [27]. The third and largest domain (residues 590 to 1332, colored blue in 򐂰Fig. 5.1) binds the Mo-pterin cofactor (Mo-pt) close to the interfaces of the Fe/S- and FAD-binding domains. Unfortunately, the recombinantly expressed proteins that have been used for crystallization lack the Mo-pt, probably due to overloading of the Mo-pt synthesis and insertion enzymes [20,27,28]. Crystal structure analysis at 1.7 Å resolution of the D428A mutant of rat XDH, a Mopt-free form, indicates that the conformation of the polypeptide chain surrounding the molybdenum cofactor is very similar to the one found in the native bovine enzyme. Even the amino acid residues in the active site assume very similar positions and orientations. The only significant difference is the absence of Mo-pt [28]. The crystal structure of Rhodobacter capsulatus XDH, has also been determined, and the overall fold of the protein is sufficiently similar to that of the bovine enzyme as to allow the use of molecular replacement techniques [29]. The bacterial enzyme is a heterotetramer, however, with the smaller A-subunit incorporating the [2Fe-2S] centers (residues A1–A153), the FAD-binding domain (residues A185–462), as well as the intervening 31 amino acid linker. The larger B-subunit binds the Mo-pt cofactor (residues B1–777). The different polypeptide composition in mammalian and bacterial enzymes notwithstanding, the overall polypeptide folds for each domain and the spatial arrangement of the four redox-active centers is very similar in all XORs (see 򐂰Fig. 5.2). Although no structure of a eukaryotic AO has yet been reported, its overall structure and cofactor location are expected to closely resemble those of XOR, due to the high degree of amino acid sequence conservation.

5.3 Reaction mechanism The oxidative hydroxylation of purines or aldehydes takes place at the molybdenum center. The reducing equivalents thus introduced into the system are transferred via the two [2Fe-2S] centers to the FAD, where reduction of the physiological electron acceptor, NAD+ or O2, occurs [1–3] (򐂰Fig. 5.1). The first and last steps in catalysis, the hydroxylation of purine or aldehyde substrates at the molybdenum center and the reoxidation at FADH2 by NAD+ or molecular oxygen, respectively, are twoelectron transfer processes. In contrast, electron transfer among the [2Fe-2S] centers

O

HCR

Glu O

OH

O(VI) S Mo S H S O CR

Glu

O

OH

O(VI) S Mo SH S O CR OH앥

2H쎵, 2e앥

Glu

O

O앥

O(VI) S Mo S S OH O CR

O2

Fig. 5.1: The subunit structure of xanthine oxidoreductase and its reaction mechanism. Left: The structure of one subunit is shown; its three domains are labeled in red, for the Fe/S domain; in yellow, for the FAD domain and in blue, for the molybdo-pterin domain. The second subunit is shown as a space-filling model in gray. Right: Cofactor arrangement and the reaction sites of substrates are shown. Bottom: Structure of molybdopterin and general mechanism of molybdenum dependent hydroxylation. R represents various nitrogen containing cyclic molecules such as hypoxanthine and xanthine, which are the physiological substrates of the enzyme, as well as aldehydes.

Glu

O앥

O(VI) S Mo S S O H

Molybdopterin

Fe/S I

Fe/S II

FAD

NAD쎵

106 5 The xanthine oxidoreductase enzyme family

5.3 Reaction mechanism

107

N-terminal domains bovine XOR rat XOR human XOR chicken XOR R.capsulatus XOR human AO D.gigas AO

----MTADELVFFVNGKKVVEKNADPETTLLAYLRRKLGLRGTKLGCGEGGCGACTVMLS ----MTADELVFFVNGKKVVEKNADPETTLLVYLRRKLGLCGTKLGCGEGGCGACTVMIS ----MTADKLVFFVNGRKVVEKNADPETTLLAYLRRKLGLSGTKLGCGEGGCGACTVMLS MAPPETGDELVFFVNGKKVVEKDVDPETTLLTYLRRKLGLCGTKLGCGEGGCGACTVMIS -------MEIAFLLNGETRRVRIEDPTQSLLELLR-AEGLTGTKEGCNEGDCGACTVMIR ---MDRASELLFYVNGRKVIEKNVDPETMLLPYLRKKLRLTGTPYGCGGGGCGACTVMIS ------MIQKVITVNG-IEQNLFVDAEALLSDVLRQQLGLTGVKVGCEQGQCGACSVILD ** * * ** * * ** * **** *

56 56 56 60 52 57 53

bovine XOR rat XOR human XOR chicken XOR R.capsulatus XOR human AO D.gigas AO

KYDRLQDKIIHFSANACLAPICTLHH-VAVTTVEGIGSTKTRLHPVQERIAKSHGSQCGF KYDRLQNKIVHFSVNACLAPICSLHH-VAVTTVEGIGNTQ-KLHPVQERIARSHGSQCGF KYDRLQNKIVHFSANACLAPICSLHH-VAVTTVEGIGSTKTRLHPVQERIAKSHGSQCGF KYDPFQKKILHHTANACLFPICALHH-VAVTTVEGIGNTKSRLHPAQERIAKSHGSQCGF ------DAAGSRAVNACLMMLPQIAG-KALRTIEGIAAPDGRLHPVQQAMIDHHGSQCGF RYNPITKRIRHHPANACLIPICSLYG-AAVTTVEGIGSTHTRIHPVQERIAKCHGTQCGF G----------KVVRACVTKMKRVADGAQITTIEGVGQPE-NLHPLQKAWVLHGGAQCGF ** * ** ** * * ****

115 114 115 119 105 116 102

bovine XOR rat XOR human XOR chicken XOR R.capsulatus XOR human AO D.gigas AO

CTPGIVMSMYTLLRNQPEPTVEEIEDAFQG--NLCRCTGYRPILQGFRTFAK CTPGIVMSMYTLLRNQPEPTVEEIENAFQG--NLCRCTGYRPILQGFRTFAK CTPGIVMSMYTLLRNQPEPTMEEIENAFQG--NLCRCTGYRPILQGFRTFAR CTPGIVMSMYTLLRNKPKPKMEDIEDAFQG--NLCRCTGYRPILEGYRTFAV CTPGFIVSMAAAHDRDRK----DYDDLLAG--NLCRCTGYAPILRAAEAAAG CTPGMVMSIYPLLRNHPEPTLDQLTDALGG--NLCRCHGYRPIIDACKTFCK CSPGFIVSAKGLLDTNADPSREDVRDWFQKHRNACRCTGYKPLVDAVMDAAA * ** * * *** ** *

165 164 165 169 152 166 155

Intermediate domains bovine XOR rat XOR human XOR chicken XOR R.capsulatus XOR human AO

-----------------------NGGCCGGNGNNPNCCMNQKKDHTVTL------SPSLF -----------------------DGGCCGGSGNNPNCCMNQTKDQTVSL------SPSLF -----------------------DGGCCGGDGNNPNCCMNQKKDHSVSL------SPSLF DSNCCGKAANGTGCCHSKGENSMNGGCCGGKANGPGCCMNE-KENVTMM------SSSLF ----------------------------------------------------------------------------------TSGCCQSKENG-VCCLDQGINGLPEFEEGSKTSPKLF

196 195 196 222

bovine XOR rat XOR human XOR chicken XOR R.capsulatus XOR human AO

NPEEFMPLDPTQEPIFPPELLRLKDVP-PKQLRFEGERVTWIQASTLKELLDLKAQHPEA NPEDFKPLDPTQEPIFPPELLRLKDTP-QKKLRFEGERVTWIQASTMEELLDLKAQHPDA KPEEFTPLDPTQEPIFPPELLRLKDTP-RKQLRFEGERVTWIQASTLKELLDLKAQHPDA DSSEFQPLDPTQEPIFPPELMTQRNKE-QKQVCFKGERVMWIQPTTLQELVALKSQYPNA PPADWLQADAAFTLAQLSSGVR-------------GQTAPAFLPETSDALADWYLAHPEA AEEEFLPLDPTQELIFPPELMIMADKQSQRTRVFGSERMMWFSPVTLKDLLEFKFKYPQA * * * * *

255 254 255 216 199 262

bovine XOR rat XOR human XOR chicken XOR R.capsulatus XOR human AO

KLVVGNTEIGIEMKFKNQLFPMIICPAWIPELNAVEHGPEGISFGAACALSSVEKTLLEA KLVVGNTEIGIEMKFKNMLFPLIVCPAWIPELNSVVHGPEGISFGASCPLSLVESVLAEE KLVVGNTEIGIEMKFKNMLFPMIVCPAWIPELNSVEHGPDGISFGAACPLSIVEKTLVDA KLVVGNTEVGIEMRLKNMLYPVILAPAWIPEMNAVQQTETGITFGAACTLSSVEEVLRKA TLIAGGTDVSLWVTKALRDLPEVAFLSHCKDLAQIRETPDGYGIGAGVTIAALRAFAEGP PVIMGNTSVGPEVKFKGVFHPGYNSPDRIEEPECCKPCIYGLTLGAGLSLAQVKDILADV * * * * **

315 314 315 341 259 322

bovine XOR rat XOR human XOR chicken XOR R.capsulatus XOR human AO

VAKLPTQKTEVFRGVLEQLRWFAGKQVKSVASLGGNIITASPISDLNPVFMASGTKLTIV IAKLPEQKTEVFRGVMEQLRWFAGKQVKSVASIGGNIITASPISDLNPVFMASGAKLTLV VAKLPAQKTEVFRGVLEQLRWFAGKQVKSVASVGGNIITASPISDLNPVFMASGAKLTLV VAELPSYKTEIFQAALEQLRWFAGPQIRNVAALGGNIMTASPISDLNPVLMASGSKLTLI HPALAG-----------LLRRFASEQVRQVATIGGNIANGSPIGDGPPALIAMGASLTLR VQKLPEEKTQMYHALLKHLGTLAGSQIRNMASLGGHIISRHPDSDLNPILAVGNCTLNLL * * * * * ** * * * * *

375 374 375 401 308 382

202

(Figure continued)

108

5 The xanthine oxidoreductase enzyme family

(Figure continued) bovine XOR rat XOR human XOR chicken XOR R.capsulatus XOR human AO

SRGTRRTVPMDHTFFPSYRKTLLGPEEILLSIEIPYSREDEFFSAFKQASRREDDIAKVT SRGTRRTVRMDHTFFPGYRKTLLRPEEILLSIEIPYSKEGEFFSAFKQASRREDDIAKVT SRGTRRTVQMDHTFFPGYRKTLLSPEEILLSIEIPYSREGEYFSAFKQASRREDDIAKVT SMEGKRTVMMDEKFFTGYRKTIVKPEEVLLSVEIPYSKEGEYFSAFKQAYRREDDIAIVT RGQERRRMPLED-FFLEYRKQDRRPGEFVESVTLPKSAPG--LRCYKLSKRFDQDISAVC SKEGKRQIPLNEQFLSKCPNADLKPQEILVSVNIPISRKWEFVSAFRQAQRQENALAIVN * * * * * * * * *

435 434 435 461 365 442

bovine XOR rat XOR human XOR chicken XOR R.capsulatus XOR human AO

CGMRVLFQPGSMQVKELALCYGGMADRTISALKTTQKQLSKFWNEKLLQDVCAGLAEELS SGMRVLFKPGTIEVQELSLCFGGMADRTISALKTTPKQLSKSWNEELLQSVCAGLAEELQ SGMRVLFKPGTTEVQELALCYGGMANRTISALKTTQRQLSKLWKEELLQDVCAGLAEELH CGMRVLFQHGTSRVQEVKLSYGGMAPTTILALKTCRELAGRDWNEKLLQDACRLLAGEMD GCLNLTLK--GSKIETARIAFGGMAGVPKRAAAFEAALIGQDFREDTIAAALPLLAQDFT SGMRVFFGEGDGIIRELCISYGGVGPATICAKNSCQKLIGRHWNEQMLDIACRLILNEVS ** * *

495 494 495 521 423 502

bovine XOR rat XOR human XOR chicken XOR R.capsulatus XOR human AO

LSPDAPGGMIEFRRTLTLSFFFKFYLTVLKKLGKDS---LAPDAPGGMVEFRRTLTLSFFFKFYLTVLQKLGRADL--LPPDAPGGMVDFRCTLTLSFFFKFYLTVLQKLGQENL--LSPSAPGGMVEFRRTLTLSFFFKFYLTVLQKLSKDQNGPPLSDMRAS-AAYRMNAAQAMALRYVRELSGEAVAVLEVMP LLGSAPGGKVEFKRTLIISFLFKFYLEVSQILKK------

531 530 531 560 462 536

C-terminal domains bovine XOR rat XOR human XOR chicken XOR R.capsulatus XOR human AO D.gigas AO

--KDKCGKLDPTYTSATLLFQKHPPANIQLFQEVPNGQSKEDTVGRPLPHLAAAMQASGE --EDMCGKLDPTFASATLLFQKDPPANVQLFQEVPKDQSEEDMVGRPLPHLAANMQASGE --EDKCGKLDPTFASATLLFQKDPPADVQLFQEVPKGQSEEDMVGRPLPHLAADMQASGE --NNLCEPVPPNYISATELFHKDPIASTQLFQEVPRGQLVEDTVGRPLVHLSAAKQACGE -----------------------------------------MSVGKPLPHDSARAHVTGQ MDPVHYPSLADKYESALEDLHSKHHCSTLKYQNIGPKQHPEDPIGHPIMHLSGVKHATGE -------------------INGKKPETDLEFKMPADG----RIWGSKYPRPTAVAKVTGT * *

589 589 590 618 19 596 192

bovine XOR rat XOR human XOR chicken XOR R.capsulatus XOR human AO D.gigas AO

AVYCDDI--PRYENELFLRLVTSTRAHAKIKSIDVSEAQKVPGFVCFLSADDIPGSNETG AVYCDDI--PRYENELSLRLVTSTRAHAKITSIDTSEAKKVPGFVCFLTAEDVPNSNATG AVYCDDI--PRYENELSLRLVTSTRAHAKIKSIDTSEAKKVPGFVCFISADDVPGSNITG AVYCDDI--PHYENELYLTLVTSTQAHAKILSIDASEAQSVPGFVCFVSAKDVPGSNITG ARYLDDL--PCPANTLHLAFGLSTEASAAITGLDLEPVRESPGVIAVFTAADLPHDNDAS AIYCDDM--PLVDQELFLTFVTSSRAHAKIVSIDLSEALSMPGVVDIMTAEHLSDVNSFC LDYGADLGLKMPAGTLHLAMVQAKVSHANIKGIDTSEALTMPGVHSVITHKDVKGKNRIT * * * * * * * ** *

647 647 648 676 77 654 252

bovine XOR rat XOR human XOR chicken XOR R.capsulatus XOR human AO D.gigas AO

LFND----------ETVFAKDTVTCVGHIIGAVVADTPEHAERAAHVVKVTYEDLP---LFND----------ETVFAKDEVTCVGHIIGAVVADTPEHAQRAARGVKITYEDLP---ICND----------ETVFAKDKVTCVGHIIGAVVADTPEHTQRAAQGVKITYEELP---IAND----------ETVFAEDVVTCVGHIIGAVIADTQEHSRRAAKAVKIKYEELK---PAPSP---------EPVLATGEVHFVGQPIFLVAATSHRAARIAARKARITYAPRP---FFTEA---------EKFLATDKVFCVGQLVCAVLADSEVQAKRAAKRVKIVYQDLEP--GLITFPTNKGDGWDRPILCDEKVFQYGDCIALVCADSEANARAAAEKVKVDLEELPAYMS * * * * **

693 693 694 722 124 702 312

bovine XOR rat XOR human XOR chicken XOR R.capsulatus XOR human AO D.gigas AO

-AIITIEDAIKNNS----FYGSELKIEKGDLKKGFSEADNVVSGELYIGGQDHFYLETHC -AIITIQDAINNNS----FYGSEIKIEKGDLKKGFSEADNVVSGELYIGGQEHFYLETNC -AIITIEDAIKNNS----FYGPELKIEKGDLKKGFSEADNVVSGEIYIGGQEHFYLETHC -PIVTIQEAIEQQS----FIKPIKRIKKGDVNKGFEESDHILEGEMHIGGQEHFYLETHC -AILTLDQALAADSR---FEGGPVIWARGDVETALAGAAHLAEGCFEIGGQEHFYLEGQA -LILTIEESIQHNS----SFKPERKLEYGNVDEAFKVVDQILEGEIHMGGQEHFYMETQS GPAAAAEDAIEIHPGTPNVYFEQPIVKGEDTGPIFASADVTVEGDFYVGRQPHMPIEPDV * * * * *

748 748 749 777 180 757 372

bovine XOR rat XOR human XOR chicken XOR R.capsulatus XOR human AO D.gigas AO

TIAIPKGEEGEMELFVSTQNAMKTQSFVAKMLGVPVNRILVRVKRMGGGFGGKETRSTLV TIAVPKGEAGEMELFVSTQNTMKTQSFVAKMLGVPDNRIVVRVKRMGGGFGGKETRSTVV TIAVPKGEAGEMELFVSTQNTMKTQSFVAKMLGVPANRIVVRVKRMGGGFGGKETRSTVV TLAVPKGEDGEMELFVSTQNLMKTQEFTASALGVPSNRIVVRVKRMGGGFGGKETRNTIL ALALP--AEGGVVIHCSSQHPSEIQHKVAHALGLAFHDVRVEMRRMGGGFGGKESQGNHL MLVVPKGEDQEMDVYVSTQFPKYIQDIVASTLKLPANKVMCHVRRVGGAFGGKVLKTGII AFAYMG-DDGKCYIHSKSIGVHLHLYMIAPGVGLEPDQLVLVANPMGGTFGYKFSP--TS * ** ** *

808 808 809 837 238 817 429

5.3 Reaction mechanism

109

bovine XOR rat XOR human XOR chicken XOR R.capsulatus XOR human AO D.gigas AO

SVAVALAAYKTGHPVRCMLDRNEDMLITGGRHPFLARYKVGFMKTGTIVALEVDHYSNAG STALALAAHKTGRPVRCMLDRDEDMLITGGRHPFLAKYKVGFMKTGTVVALEVAHFSNGG STAVALAAYKTGRPVRCMLDRDEDMLITGGRHPFLARYKVGFMKTGTVVALEVDHFSNVG TTVVAVAAFKTGRPVRCMLDRDEDMLISGGRHPFLGRYKVGFMKNGKIKSLEVSYYSNGG AIACAVAARATGRPCKMRYDRDDDMVITGKRHDFRIRYRIGADASGKLLGADFVHLARCG AAVTAFAANKHGRAVRCVLERGEDMLITGGRHPYLGKYKAGFMNDGRILALDMEHYSNAG EALVAVAAMATGRPVHLRYNYQQQQQYTGKRSPWEMNVKFAAKKDGTLLAMESDWLVDHG * ** * * * * *

868 868 869 897 298 877 489

bovine XOR rat XOR human XOR chicken XOR R.capsulatus XOR human AO D.gigas AO

NSRDLSHSIMERALFHMDNCYKIPNIRGTGRLCKTNLSSNTAFRGFGGPQALFIAENWMS NTEDLSRSIMERALFHMDNAYKIPNIRGTGRICKTNLPSNTAFRGFGGPQGMLIAEYWMS NTQDLSQSIMERALFHMDNCYKIPNIRGTGRLCKTNLPSNTAFRGFGGPQGMLIAECWMS NSADLSHGVMDRALLHLDNSYNIPNVSIMGFICKTNLSSNTAFRGFGGPQGMMIAECWMS WSADLSLPVCDRAMLHADGSYFVPALRIESHRLRTNTQSNTAFRGFGGPQGALGMERAIE ASLDESLFVIEMGLLKMDNAYKFPNLRCRGWACRTNLPSNTAFRGFGFPQAVLITESCIT PYSEFGDLLTLRGAQFIGAGYNIPNIRGLGRTVATNHVWGSAFRGYGAPQSMFASECLMD * * ** **** * ** *

928 928 929 957 358 937 549

bovine XOR rat XOR human XOR chicken XOR R.capsulatus XOR human AO D.gigas AO

EVAVTCGLPAEEVRWKNMY-------------------KEGDLTHFNQRLEGFSVPRCWD EVAITCGLPAEEVRRKNMY-------------------KEGDLTHFNQKLEGFTLPRCWD EVAVTCGMPAEEVRRKNLY-------------------KEGDLTHFNQKLEGFTLPRCWE DLARKCGLPPEEVRKINLY-------------------HEGDLTHFNQKLEGFTLRRCWD HLARGMGRDPAELRALNFYDPPERGLSAPPSPPEPIATKKTQTTHYGQEVADCVLGELVT EVAAKCGLSPEKVRIINMY-------------------KEIDQTPYKQEINAKNLIQCWR MLAEKLGMDPLELRYKNAY-------------------RPGDTNPTGQEPEVFSLPDMID * * * * * *

969 969 970 998 418 978 590

bovine XOR rat XOR human XOR chicken XOR R.capsulatus XOR human AO D.gigas AO

ECLKSSQYYARKSEVDKFNKENCWKKRGLCIIPTKFGISFTVPFLNQAGALIHVYTDGSV ECIASSQYLARKREVEKFNRENCWKKRGLCIIPTKFGISFTLPFLNQGGALVHVYTDGSV ECLASSQYHARKSEVDKFNKENCWKKRGLCIIPTKFGISFTVPFLNQAGALLHVYTDGSV ECLSSSNYHARKKLIEEFNKQNRWKKRGMCIIPTKFGISFTVPFLNQAGALVHVYTDGSV RLQKSANFTTRRAEIAAWNSTNRTLARGIALSPVKFGISFTLTHLNQAGALVQIYTDGSV ECMAMSSYSLRKVAVEKFNAENYWKKKGLAMVPLKFPVGLASRAAGQAAALVHIYLDGSV QLRPKYQAALEKAQKESTAT----HKKGVGISIGVYGSGLDGPDAS--EAWAELNADGTI * * **

1029 1029 1030 1058 478 1038 644

bovine XOR rat XOR human XOR chicken XOR R.capsulatus XOR human AO D.gigas AO

LVSHGGTEMGQGLHTKMVQVASKALK---IPISKIYISETSTNTVPNSSPTAASVSTDIY LLTHGGTEMGQGLHTKMVQVASRALK---IPTSKIHISETSTNTVPNTSPTAASASADLN LLTHGGTEMGQGLHTKMVQVASRALK---IPTSKIYISETSTNTVPNTSPTAASVSADLN LLTHGGTEMGQGLHTKMIQVASRSLG---IPTSKIYISETSTNTVPNTSPTAASVSADIN ALNHGGTEMGQGLHAKMVQVAAAVLG---IDPVQVRITATDTSKVPNTSATAASSGADMN LVTHGGIEMGQGVHTKMIQVVSRELR---MPMSNVHLRGTSTETVPNANISGGSVVADLN TVHTAWEDHGQGADIGCVGTAHEALRPMGVAPEKIKFTWPNTATTPNSGPSGGSRQQVMT *** * * ** *

1089 1089 1090 1115 535 1095 704

bovine XOR rat XOR human XOR chicken XOR R.capsulatus XOR human AO D.gigas AO

GQAVYEACQTILKRLEPFKKKNP--------------DGSWEDW-----VMAAYQDRVSL GQGVYEACQTILKRLEPFKKKKP--------------TGPWEAW-----VMDAYTSAVSL GQAVYAACQTILKRLEPYKKKNP--------------SGSWEDW-----VTAAYMDTVSL GMAVHNACQTILKRLEPIKQSNL--------------KGSWEDW-----IKTAYENCISL GMAVKDACETLRGRLAGFVAAREGCAARDVIFDAGQVQASGKSWRFAEIVAAAYMARISL GLAVKDACQTLLKRLEPIISKNP--------------KGTWKDW-----AQTAFDESINL GNAIRVACENLLKACEKP-------------------GGGYYTY-----DELKAADKPTK * **

1127 1127 1128 1156 595 1136 740

bovine XOR rat XOR human XOR chicken XOR R.capsulatus XOR human AO D.gigas AO

STTGFYRTPNLGYSFETNSGNAFHYFTYGVACSEVEIDCLTGDHKNLRTDIVMDVGSSLN SATGFYKTPNLGYSFETNSGNPFHYFSYGVACSEVEIDCLTGDHKNLRTDIVMDVGSSLN SATGFYRTPNLGYSFETNSGNPFHYFSYGVACSEVEIDCLTGDHKNLRTDIVMDVGSSLN SATGFYRIPDVGYNFETNKGKPFHYFSYGVACSEVEIDCLTGDHKNIRTDIVMDVGTSLN SATGFYATPKLSWDRLRGQGRPFLYFAYGAAITEVVIDRLTGENRILRTDILHDAGASLN SAVGYFRGYESDMNWEKGEGQPFEYFVYGAACSEVEIDCLTGDHKNIRTDIVMDVGCSIN ITGNWTASGATHCDAVTGLGKPFVVYMYGVFMAEVTVDVATGQTTVDGMTLMADLGSLCN * * ** ** * ** * * *

1187 1187 1188 1216 655 1196 800

bovine XOR rat XOR human XOR chicken XOR R.capsulatus XOR human AO D.gigas AO

PAIDIGQVEGAFVQGLGLFTLEELHYSPEGSLHTRGPSTYKIPAFGSIPTEFRVSLLRDC PAIDIGQVEGAFVQGLGLFTMEELHYSPEGSLHTRGPSTYKIPAFGSIPIEFRVSLLRDC PAIDIGQVEGAFVQGLGLFTLEELHYSPEGSLHTRGPSTYKIPAFGSIPIEFRVSLLRDC PAIDIGQIEGAFVQGIGLFTMEELRYSPEGNLYTRGPGMYKIPAFGDIPTEFYVSLLRDC PALDIGQIEGAYVQGAGWLTTEELVWDHCGRLMTHAPSTYKIPAFSDRPRIFNVALWDQP PAIDIGQIEGAFIQGMGLYTIEELNYSPQGILHTRGPDQYKIPAICDMPTELHIALLPPS QLATDGQIYGGLAQGIGLALSEDFEDIKKHATLVGAG----FPFIKQIPDKLDIVYVN-H ** * ** * * * *

1247 1247 1248 1276 715 1256 855

(Figure continued)

110

5 The xanthine oxidoreductase enzyme family

(Figure continued) bovine XOR rat XOR human XOR chicken XOR R.capsulatus XOR human AO D.gigas AO

PNKKAIYASKAVGEPPLFLGASVFFAIKDAIRAARAQHTNNNTKELFRLDSPATPEKIRN PNKRAIYASKAVGEPPLFLASSIFFAIKDAIRAARAQH-GDNAKQLFQLDSPATPEKIRN PNKKAIYASKAVGEPPLFLAASIFFAIKDAIRAARAQHTGNNVKELFRLDSPATPEKIRN PNSKAIYSSKAVGEPPLFLSASVFYAIKDAIYSAREDS---GVTEPFRLDSPATPERIRN NREETIFRSKAVGEPPFLLGISAFLALHDACAACGPHWP--------DLQAPATPEAVLA QNSNTLYSSKGLGESGVFLGCSVFFAIHDAVSAARQER---GLHGPLTLNSPLTPEKIRM PRPDGPFGASGVGELPLTSPHAAIINAIKSATGVRIYR------------LPAYPEKVLE ** * **

bovine XOR rat XOR human XOR chicken XOR R.capsulatus XOR human AO D.gigas AO

ACVDKFTTLCVTGAPGNCKPWSLRV ACVDQFTTLCVTGVPENCKSWSVRI ACVDKFTTLCVTGVPENCKPWSVRV ACVDTFTKMCPSAEPGTFKPWSVRA AVR--------RGAEGRA------ACEDKFTKMIPRDEPGSYVPWNVPI ALKA---------------------

1307 1306 1308 1333 767 1313 903

1332 1331 1333 1358 777 1338 907

Fig. 5.2: Comparison of the amino acid sequences of the domains of three mammalian XORs (bovine [80], rat [81,83], and human [82]) with those of chicken XDH [83], Rodobacter capsulatus XDH [84], and aldehyde oxidases from human [85] and Desulfovibrio gigas [86]. The amino acid residues in the N-terminal, the intermediate and the C-terminal domains are colored in red, gold and blue based on the crystal structure of bovine XOR. Others are in grey. Important residues discussed in the text are in black.

occurs in one-electron steps whereby these clusters not only serve as a convenient route for electron movement between the Mo and FAD reaction sites but also act as a modulator of the two centers, electron sinks that can receive or provide electrons to keep the centers primed in a reactive state [30,31]. The reaction mechanism of hydroxylation of hypoxanthine, xanthine, and aldehydes in general at the Mo center is shown in 򐂰Fig. 5.1. During the reaction the oxygen atom is incorporated from water [32,33]. The geometry and ligand binding of this complex are depicted in the bottom part of the figure that was updated by crystallographic analysis of the complex of XDH with the slow substrate FYX-051 [34]. Bray et al. were the first to trap intermediates in the extremely fast reaction. Using a rapid freezing method, they observed the Mo signal by EPR and showed that the hydroxylation reaction is clearly Mo-dependent [35]. In the so-called reductive half-reaction, a transient Mo(V) signal is also observed. Although the process of Mo(V) formation has been argued from various points of view, it has not been until recently that the details of this process have been elucidated as discussed in more depth in the recent review by Hille et al. [18,19]. The present model sees a hydride ion transferred from the substrate to one of the sulfur ligands of the central Mo ion, (Mo(VI)=S → Mo(IV)-SH). According to this model, Mo(V) is generated by hydride transfer followed by one-electron transfer from the Mo ion to the neighboring [2Fe-2S] cluster center. EPR spectroscopy together with work on chemical model compounds has generated various mechanistic proposals for the hydroxylation mechanism of XOR; more recently, X-ray crystallography and the analysis of point-mutants of the enzymes have established more definitive conclusions regarding the mechanistic details of the hydroxylation of purine bases. The X-ray structure of bovine XDH with bound salicylic acid (an inhibitor that competes with xanthine) provided the first clue as to how aromatic rings are wedged between the side chain rings of two phenylalanines (Phe914, Phe1009 in the bovine amino acid sequence numbering) [26]. Later, several crystal structure analyses confirmed this orientation for the purine rings [28].

5.3 Reaction mechanism

111

It is now generally accepted that XOR transfers the water exchangeable -OH and not the Mo=O ligand of the Mo-atom to the substrate [2,18] (򐂰Fig. 5.1, bottom). The free electron pairs of the oxygen, deprotonated by the neighboring glutamate residue conserved among all these enzymes [36], then attack the electrophilic carbon site of the substrate to yield the hydroxylated product. In its equatorial position, the Mo=S ligand is well placed to accept a hydride ion from the reactive carbon atom, which results in protonation to Mo-SH and reduction of the Mo-ion. The nature and binding geometry of the molybdenum ligands have been elucidated through several high-resolution X-ray crystal structures, one of them being the complex of bovine XDH and the compound FYX-051 (an XOR inhibitor introduced by Fuji Yakuhin Co.) [34]. The sulfur atom and water-derived oxygen ligands discussed above lie in the equatorial plane of the molybdenum complex. The X-ray crystal structure of this complex revealed a covalent Mo-O-C bond [34]. Protonation of the side chain of Glu1261 leads to the formation of a hydrogen bond with the N1 nitrogen atom of the substrate and facilitates the nucleophilic attack on the adjoining carbon by the oxygen atom that by loosing a proton has been transformed into a base (Mo-O−) [34]. A similar mechanistic concept seems to be at work in AO where the crystal structure of the D. gigas enzyme also shows the side chain of a glutamic acid residue placed near the Mo-OH group [35]. Mutating this glutamate renders the enzyme completely inactive [37,16]. Working with human XOR and probing the role played by charged active site residues in the activation of substrates, Yamaguchi et al. found that the enzyme’s catalytic activity decreases significantly when Glu803 and Arg881 (Arg880 in the bovine enzyme) are mutated to valine and methionine, respectively, the corresponding amino acids in human AO [16]. These mutants still possess significant aldehyde hydroxylation activity, but have lost most of their activity against purine substrates. The same authors have proposed how the substrates hypoxanthine and xanthine bind to XOR (򐂰Fig. 5.3A); hydrogen bonds formed between the substrate and several active site amino acids are the major contributors to substrate activation and facilitate the nucleophilic reaction [16]. This proposal is also consistent with the metabolic sequence of reactions in which hydroxylation of the 2-position of hypoxanthine precedes that of the 8-position. It seems highly likely that the interaction of the 2-position keto group (C=O) with Arg881 is crucial to achieving good hydroxylation rates at the 8-position. As the two mutations mentioned above do not change AO activity, such activation is obviously not needed for aldehyde substrates, where the existing keto group seems to suffice to allow proficient attack by Mo-O− on the target carbon atom [18]. Based on crystallographic work and chemical considerations, another mode of binding and following from that a different mechanism of activation has been proposed for XOR (򐂰Fig. 5.3B) [38,39]. In these models, the binding modes were obtained with nonfunctional enzyme, the desulfo form of bovine XO [38,40]. On the other hand, Nishino et al. have reported a failure to delineate the binding modes of the substrate molecules with the desulfo form of the enzyme, since the electron density seemed to represent a mixture of various orientations, presumably reflecting the very broad substrate specificity of this enzyme [20]. A QM/MM computational study has provided support for the model shown in 򐂰Fig. 5.4A [41,42]. In addition, the crystal structure of urate bound to the reduced enzyme revealed a mode of binding very similar to the one in 򐂰Fig. 5.3A [28]. Electron density connecting the C8 atom of urate to the Mo-center is bent, spanning a total of 3.5 Å. Placing the oxygen atom at the apex of the bent electron

112

5 The xanthine oxidoreductase enzyme family

A

OH B

O

C

O 쎵

Mo6

H N

앥

N NH

O

N

H N

N H

쎵

Mo6

O

앥

O

O 쎵

Mo6

NH

N H

앥

O

H C

O

R

앥

O

OH O

H N

NH2

H2N

NH2

HN

NH2

Glu880

Urate

HOH Molybdopterin

Glu1261

Fig. 5.3: Alternative models of the xanthine hydroxylation reaction. (A) Model of xanthine binding based on analyses of mutant enzymes and the crystallographically observed mode of urate binding. The hydrogen bonds to the three amino acids promote a nucleophilic reaction at the carbon atom to be hydroxylated [16]. (B) Model based on analyses of the binding mode obtained with crystals of the desulfo-form of XO. Activation of the substrate xanthine is achieved by Arg881 via accumulation of negative charge at the oxygen at the 6-position [38]. (C) The crystal structure of the urate complex of reduced XOR under anaerobic conditions, showing the reaction intermediate during hydroxylation [28]. A water molecule is present at hydrogen bond distance to N3 and N9 of xanthine.

density leads to Mo-O and C8-O distances of 2.2 Å and 1.4 Å, respectively. The latter approaches the 1.3 Å value regarded as the standard C-O distance for hydroxylated six-membered rings. A distance of 2.3 Å for the equatorial Mo-S bond is in agreement with a Mo(IV)-SH form of the sulfido ligand. The structure pictures a state of catalysis arrested at the step of oxygen transfer from the molybdenum coordination sphere to the receiving substrate carbon. Clear electron density is also observed for a water molecule at hydrogen-bond distance to the N3 and N9 atoms of xanthine. This water might well promote the tautomerization of purines thought to be essential for release of urate product, a role that had been assigned to Glu802 in the alternate binding mode model [18].

5.4 Electron transfer from the molybdenum center to other redox-active centers

2.02

113

16 K 1 mW

1.93 1.905

2.11 A

Wild-type dimer 2.05

B

C43S dimer

280

300

320

340

360

380

Magnetic field (mT)

Fig. 5.4: EPR spectra of the dithionite-reduced form of the recombinant wildtype dimer of rat/bovine XOR (top) and of the C43S mutant dimer. C43 is located at the N-terminal plant ferredoxin type cysteine motif, and mutation of this residue influences the EPR signal of Fe/SII but not of Fe/SI. (See [45] for more details.)

5.4 Electron transfer from the molybdenum center to other redox-active centers Electrons transferred to molybdenum in the reductive half-reaction are quickly re-distributed to the other redox-active centers of the enzyme [31]. Before the elucidation of any crystal structure, with the pathway taken by the electrons unclear, stoppedflow methods and steady-state kinetics had already shown that transfer in the molecule was very rapid, with both O2 and NAD+ as substrates. Lowering the temperature in the stop-flow apparatus allowed the observation of separate catalytic steps of the xanthineO2 activity. The slowest and therefore rate-determining step, turned out to be the dissociation of urate [43]. This fact has to be kept in mind in discussions of the reaction mechanism based on the various crystal structures known today. It has been known for quite some time that the two iron–sulfur centers of XOR can be distinguished on the basis of their EPR signals [1]. The Fe/S I signal displays g-values of g1,2,3 = 2.022,1.932,1.894, respectively, with line-widths and relaxation properties typical for a [2Fe-2S] cluster, while Fe/S II has g-values of g1,2,3 = 2.110,1.991,1.902, respectively, with unusually broad line widths and relaxation properties. The latter signal can therefore only be observed below 25K [44]. Site-directed mutagenesis studies employing heterologously expressed rat XOR has allowed assignment of the two distinct EPR signals to their respective clusters [45], with Fe/S I being the one in the unusual α-helical domain and Fe/S II the one in the N-terminal ferredoxin-like domain. This assignment is consistent with the observed coupling of Fe/S-I to the molybdenum center because it places this [2Fe-2S] cluster proximal to the molybdenum ion [46,47] (򐂰Fig. 5.4). This establishes the sequence of electron transfer within the enzyme molecule as Mo → Fe/ S-I → Fe/S-II → FAD as illustrated in 򐂰Fig. 5.1. The actual reaction rate of this transfer has

114

5 The xanthine oxidoreductase enzyme family

been measured by the pulse radiolysis method, which has confirmed its very rapid rate [48,49]. Electrons transferred to FAD are eventually passed on to the second substrate, either NAD+ or molecular oxygen, yielding NADH, hydrogen peroxide or superoxide.

5.5 Reaction of FAD with NAD+ or molecular oxygen Historically, XOR from mammalian sources, such as cow’s milk [50,51], has been isolated in its XO form. In contrast, purification from other organisms, such as chicken [52], insects [53] or bacteria [54] has always yielded the XDH form of the enzyme. The main difference in catalytic properties between the XDH and XO forms is the reactivity of their reduced FAD cofactor toward NAD+ and molecular oxygen [55,56]. While XDH reacts rapidly with NAD+, but slowly with oxygen, XO shows the opposite behavior. The differences in kinetic properties between XDH and XO can in part be explained by the flavin oxidation/reduction potentials, which are shown in 򐂰Tab. 5.1 and 5.2. In XDH, either in its substrate-free or NAD+-bound form, the oxidation/reduction potential of the FADH•/FADH2 couple is much lower than that of the FAD/FADH• couple, therefore the FADH• semiquinone is thermodynamically stabilized to a greater degree in XDH, than in XO [55,56]. The results from stopped-flow studies of reduced stable (non-convertible to XO) XDH from chicken liver are consistent with the assumption of FADH2 reacting with O2 to produce H2O2 and FADH• reacting with O2 to produce O2•−, in both cases regenerating oxidized FAD. In the absence of NAD+, XDH produces more O2•− per electron introduced than XO due to the higher stability of FADH• in XDH [57]. The most likely explanation for the difference in the stability of the neutral semiquinone is a major change in the protein environment of the FAD [58–61]. This change removes a strong negative charge in the vicinity of the FAD, present in XDH, upon conversion to XO, as revealed by reconstitution experiments with flavin active site probes [62,63]. Recent computational studies of the protein structure

Tab. 5.1: Steady state kinetic parameters at 25˚C. Enzyme

XO from bovine milk [31]

Xanthine-NAD+ activity

Xanthine-O2 activity kcat

Km for xanthine

Km for O2

kcat

18.3 s−1

9 µM

53 µM

–

XDH from bovine milk [5] XO from rat liver [5]

17.2 s−1

1.8 µM

46 µM

XDH from rat liver [5]

4.5 s−1

2.8 µM

260 µM

Km for xanthine

Km for NAD+

6.3 s−1

0.3 µM

7 µM

13.5 s−1

1.3 µM

8.5 µM

5.5 Reaction of FAD with NAD+ or molecular oxygen

115

Tab. 5.2: Midpoint potentials (mV) of the redox centers of bovine milk XDH and XO. XDH1

XO2

FADH•/FADH2

–410

–237

FAD/FADH•

–270

–310

Fe/S II

–235

–217

Fe/S I

–310

–310

Mo(V)/Mo(IV)

nd

–315

Mo(VI)/Mo(V)

nd

–345

1 2

In 0.1 M pyrophosphate, 0.3 mM EDTA, pH7.5, at 25˚C [59] In 0.1 M potassium N,N’-bis(2-hydroxyethyl)glycine, pH7.7, at 25˚C [88]

suggest that the negative charge is due to Asp429, as suggested by the X-ray crystal structures of XDH and XO [64]. The largest difference in the crystal structures of the bovine XDH and XO forms is the change in position of the peptide loop comprising residues Gln423-Lys433 (denoted as loop A), which passes close to the FAD; this change is very similar to that observed in the XDH to XO conversion seen in the crystal structures of mutants of the rat enzyme. In the XDH form, the side chain of bovine Asp429 (corresponding to residue 428 in the rat enzyme) comes as close as 3.6 Å to C6 of the flavin ring. The electrostatic potential of the FAD environment is drastically changed, however, once the enzyme adopts its XO conformation, which involves a major repositioning of the loop. Asp429 has now moved away from the flavin ring and is replaced by the guanidinium group of Arg426, located at a 6.3 Å distance from the nearest atom of the isoalloxazine ring (򐂰Fig. 5.5). This reversal of the electrostatic potential surrounding the redox-active part of the FAD is fully consistent with predictions based on kinetic and replacement studies of natural FAD by FAD analogs as described above. The repositioning of loop A in XO not only strongly influences the electrostatics surrounding the flavin cofactor but also blocks NAD+ from approaching the FAD ring, thereby preventing flavin-nicotinamide electron transfer. In addition, the NAD+ binding site does not appear to be fully formed in XO. This adds another obstacle for any nicotinamide dinucleotide to approach the flavin, again re-directing electron transfer to molecular oxygen [50,64]. The mechanism of conversion of mammalian XOR from XDH to XO has been thoroughly investigated in the past decade by a range of techniques, including X-ray crystal structure analyses of various mutants. The results convincingly show that the protein environment influences the reactivity of the FAD toward its two acceptor substrates NAD+ and molecular oxygen. The underlying conformational changes can be triggered by modifications far from the FAD itself [20]. A schematic representation of these causes is illustrated in 򐂰Fig. 5.5. Proteolytic nicking or disulfide formation in the long linker between FAD and Mo-pt domains, leads to loss of the short helix incorporating residues 531–535 that is part of the linker. This then triggers the disruption of a tight amino acid cluster that opens a solvent gate and induces the movement of the active site loop. Even the C-terminal peptide is involved in these conformational rearrangements [manuscript in preparation].

116

5 The xanthine oxidoreductase enzyme family A

Cys535SH

Cys535 S

Cys992SH

Cys992S

XDH

XO

or proteolysis

B

A-loop

A-loop

Trp335 Leu336 Arg334

Arg426

Arg334

Arg426

Phe549

Phe549

Ala335

XDH

XO

Closed cluster

Open cluster

C

Asp428

Arg426

Asp428 Trp335 Arg334

Arg426

Trp335

Arg334

Phe549 Ala535

Arg992

Cys992 Cys535

XDH

XO

Fig. 5.5: Schematic representation of various features involved in the conversion of XOR from XDH (left) to XO (right). (A) Proteolytic nicking or disulfide formation in the linker between two domains, which leads to the loss of helix 531–535 on the linker, (B) triggers disruption of the amino acid cluster, which in turn causes opening of the solvent gate, as well as (C) movement of the active site loop [87]. All figures show crystal structures around the FAD cofactor site of the XDH or XO forms of bovine milk xanthine oxidoreductase. FAD is shown as a yellow stick model, while the A-loop (Gln423–Lys433) is in green with important residues displayed in atom color coding.

5.6 Inhibitors of xanthine oxidoreductase Gout is a disease in which crystals of uric acid collect in tissues and joints, causing painful inflammation [22]. The factors leading to gout are not fully understood, but include high levels of uric acid in the blood. XOR is the major source of uric acid; therefore, potent XOR inhibitors that significantly lower the production of uric acid and subsequently its blood concentration should be highly effective as drugs directed

5.6 Inhibitors of xanthine oxidoreductase

117

Tab. 5.3: Inhibitors for which the mode of binding to XOR has been well examined. Allopurinol and Oxipurinol

Ki = 10–6 M

Covalent bond

Ref. [65]

Ref. [69,70]

BOF-4272

Ki = 1.2 × 10–9 M Ki’ = 9 × 10–9 M Ref. [71]

Febuxostat

Ki = 1.2 × 10–10 M Ki’ = 9 × 10–10 M Ref. [72]

Y-700

Ki = 6 × 10–10 M Ki’ = 3.2 × 10–9 M Ref. [73]

FYX-051

Ki = 5.7 × 10–9 M

Covalent bond

Ref. [74]

Ref. [34]

against gout [65]. Based on this strategy, allopurinol (򐂰Tab. 5.3) was introduced into the market as the first large-scale therapeutic drug for the disease [22]. It has been successful for more than 40 years, establishing the usefulness of XOR inhibition in general as treatment of gout. Many XOR inhibitors have been reported since [66], but until recently allopurinol has been the only clinically relevant inhibitor, likely due to lower efficacies of its potential competitors. As a housekeeping enzyme, XOR is found abundantly in a large number of tissues, and it is important that any inhibitor of XOR be sufficiently potent and its concentration in the affected tissues high enough to achieve significant systemic reduction of uric acid. Allopurinol, an isomer of hypoxanthine in which the N7 nitrogen is replaced with carbon and the C8 carbon is replaced with nitrogen (򐂰Tab. 5.3), has proven to be a more potent inhibitor than originally expected [67–69]. Massey et al. have shown that XOR converts allopurinol to oxipurinol (4,6-dihydroxypyrazolopyrimidine), which then binds very tightly to the reduced molybdenum Mo(IV). The stability of the oxipurinolMo(IV) complex is the reason for the efficacy in lowering uric acid levels in clinical usage. However, the complex is gradually re-oxidized with a half-life of 300 min at 25°C, leading to dissociation of the oxipurinol and recovery of activity [69]. The crystal structure of oxipurinol bound to reduced bovine XOR is illustrated in 򐂰Fig. 5.6A [70]. This binding mode is very similar to that seen for uric acid and shown in 򐂰Fig. 5.3.

118 A

5 The xanthine oxidoreductase enzyme family B

Glu802

Glu802

2.6 Å

2.9 Å

2.3 Å

BOF-4272

Oxipurinol 2.9 Å

2.5 Å

2.7 Å

2.9 Å

Glu1261

Thr1010

Arg880

Arg880

C

Asn768 Glu802 3.0 Å

2.8 Å

Febuxostat 2.8 Å

2.8 Å

Thr1010

Arg880

D

Asn768 Glu802

E

Asn768 Glu802

3.2 Å

2.8 Å

3.3 Å

2.9 Å

2.9 Å

Y-700 2.9 Å

2.9 Å

Thr1010

2.7 Å

FYX-051

Glu1261

Arg880

Fig. 5.6: Crystal structure of bovine milk XOR in complex with the inhibitor molecules listed in Table 5.3. (A) oxipurinol, at 2.1 Å resolution (3AMZ) [33]; (B) BOF-4272, 2.2 Å (will be deposited); (C) Febuxostat, 2.8 Å (1N5X) [72] (left side, overall view into the active site cavity, which is shown in space-filling representation; (D) Y-700, 2.0 Å (1VDV) [73]; (E) FYX-051, 1.9 Å (1V97) [34].

Once the covalent bond is broken, the interaction between oxipurinol and oxidized XOR is weak, and oxipurinol inhibits oxidized XOR with a Ki of only ~1.7 x 10–6 M [65]. The dependence of oxypurinol binding on the oxidation state of the enzyme may be the reason why allopurinol must be administered frequently (3 times a day) to assure reduction of uric acid levels in the blood. Since 1990, several pharmaceutical companies have reported very effective new XOR inhibitors [66], which have attracted great interest as candidate therapeutic agents and might bring the undisputed reign of allopurinol as the only approved drug against gout to an end [22]. Following successful clinical trials, Febuxostat has been approved for sale and clinical use in a number of countries. 򐂰Tab. 5.3 gives a list of

5.6 Inhibitors of xanthine oxidoreductase

119

inhibitors for which the mode of binding to XOR has been well examined, including by X-ray crystallography (򐂰Fig. 5.6). The features elucidated by the crystal structure of the complex of bovine milk XOR with Febuxostat account well for the tight binding of this and related inhibitors [71–74]. The inhibitory molecules fit very snugly into the channel leading to the molybdenum center, with no open space remaining after binding (򐂰Fig. 5.6C). In contrast to allopurinol, these new inhibitors are all non-covalent inhibitors. Although no direct bond to the molybdenum center is formed, these molecules all establish multiple interactions with the enzyme that include ionic and hydrogen bonds, π-π interactions between the aromatic rings and nearby phenylalanine residues, van der Waals and hydrophobic forces (򐂰Fig. 5.6C). To give an example, a hydrogen bond between the inhibitor and Asn768 of the bovine enzyme is noted. In the crystal structure, the residue’s side chain amide and the cyano group at the 3-position of the inhibitor molecule are 2.9 Å apart, indicative of a strong interaction. Although this asparagine residue is located too far from the active site for direct involvement in purine substrate recognition or catalytic activity, the cyano group of Febuxostat, and consequently the hydrogen bond interaction with Asn768, is necessary for potent enzyme inhibition [72]. The multitude of additive interactions results in the tight binding observed. The dissociation constant of Febuxostat for the active form of the enzyme, for example, is too low to be determined accurately; with the desulfo-form used as a substitute, a KD of 2 ± 0.03 × 10−9 M has been measured [72]. Steady state kinetics show a ping-pong mechanism for the catalytic reaction of XOR alternating between the Mo(VI) and Mo(IV) states. Different inhibition constants are typically observed for the two states and consequently inhibitors generally exhibit mixed-type inhibition in a Lineweaver-Burk plot [72]. Independent of the enzyme’s oxidation state, however, extremely low inhibition constants are obtained. Thus, unlike allopurinol, the inhibitory activity of the inhibitor is not much influenced by the valency of the molybdenum. This indicates that these structure-based inhibitors bind to the enzyme irrespective of changes to its oxidation state. As a result, these inhibitors, once delivered to affected tissues, bind stably to XOR throughout the whole in vivo lifetime of XOR. It is therefore expected that this type of XOR inhibitor will have the benefits of low administration frequency and low dose for the treatment of gout. Indeed, clinical studies of Febuxostat conducted in the USA and Japan support this conclusion [75–78]. FYX-051, a hybrid-type inhibitor, which also binds in the molybdenum center access channel, also acts as a slow substrate, forming a covalent complex in addition in the manner described in the section on the mechanism of hydroxylation (򐂰Fig. 5.6E). Singly bound oxygen, which is re-supplied to the molybdenum center from solvent water, forms a bond with a carbon atom of the inhibitor. In the corresponding electron density map, a bridging oxygen derived from the Mo-OH group of the oxidized enzyme can be placed at the apex of the bent electron density. In addition to this covalent interaction with the Mo ion, FYX-051 also undergoes various interactions with amino acid residues in the substrate channel, similar to those seen with the inhibitors discussed in the preceding paragraph. Given these dual properties, one would expect tight binding whether the covalent bond is formed or not. Very recently, the results of inhibition studies have been reported comparing the effects of Febuxostat on both the bovine milk and the bacterial R. capsulatus enzyme [79]. It is found that the binding affinity of the Febuxostat molecule is strikingly different for the two proteins despite the fact that they are homologous and display a high degree of similarity in both sequence and structure. While bovine XOR is strongly

120

5 The xanthine oxidoreductase enzyme family

inhibited, Febuxostat is incapable of efficiently inhibiting the bacterial enzyme. Allopurinol, however, as a mechanism-based inhibitor, is highly effective against both enzymes. In an attempt to clarify the difference in binding behavior, molecular dynamics simulations have been performed. The results are in accord with the experiments, and indicate that the mobility of several hydrophobic residues on the surface of the protein molecule exert significant influence on the inhibitory power displayed by Febuxostat against the two XOR enzymes. These results also indicate that despite more than a century of research, the broad variety of methods applied, and the fine details already elucidated, our understanding of catalysis by the members of the xanthine oxidoreductase family as well as their mode of inhibition is still far from complete and further investigations will almost certainly lead us to more exciting findings.

Acknowledgement This work was supported by Program for the Strategic Research Foundation at The Private Universities and Grant-in Aid (T.N. 24659144) for scientific research from the Japanese Ministry of Education, Science, Sports and Culture as well as grants from the Canadian Institute for Health Research (MOP-64392) and the Canada Research Chairs Program (E.F.P.).

5.7 References [1] Bray RC. Molybdenum iron-sulfur flavin hydroxylases and related enzymes. In The Enzymes XII ed. Boyer, PD. NY, USA: Academic Press, 1975:300–419. [2] Hille R. The mononuclear molybdenum enzymes. Chem Rev 1996;96:2757–816. [3] Hille R, Nishino T. Flavoprotein structure and mechanism. 4. Xanthine oxidase and xanthine dehydrogenase, Faseb J 1995;9:995–1003. [4] Waud WR, Rajagopalan KV. Purification and properties of the NAD+-dependent (type D) and O2-dependent (type O) forms of rat liver xanthine dehydrogenase. Arch Biochem Biophys 1976;172:354–64. [5] Nakamura M, Yamazaki I. Preparation of bovine milk xanthine oxidase as a dehydrogenase form. J Biochem (Tokyo) 1982;92:1279–86. [6] Saito T, Nishino T. Differences in redox and kinetic properties between NAD-dependent and O2-dependent types of rat liver xanthine dehydrogenase. J Biol Chem 1989;264:10015–22. [7] Schardinger, F. Über das Verhalten der Kuhmilch gegen Methylenblau und seine Verwendung zur Unterscheidung von ungekochter und gekochter Milch. Z. Untersuch Nahrungs Genussmittel. 1992;5:1113–21. [8] Massey V, Harris CM. Milk xanthine oxidoreductase: the first one hundred years. Biochem Soc Trans 1997;25:750–5. [9] Garattini E, Mendel R, Romao MJ, Wright R, Terao M. Mammalian molybdo-flavoenzymes, an expanding family of proteins: structure, genetics, regulation, function and pathophysiology. Biochem J 2003;372:15–32. [10] Garattini E, Fratelli M, Terao M. The mammalian aldehyde oxidase gene family. Human Genomics 2009;4:119–30. [11] Krenitsky TA. Aldehyde oxidase and xanthine oxidase–functional and evolutionary relationships. Biochem Pharma 1978;27:2763–4.

5.7 References

121

[12] Terao M, Kurosaki M, Saltini G, Demontis S, Marini M, Salmona M, Garattini E. Cloning of the cDNAs coding for two novel molybdo-flavoproteins showing high similarity with aldehyde oxidase and xanthine oxidoreductase. J Biol Chem 2000;275:30690–700. [13] Mendel RR. Cell biology of molybdenum in plants. 2011;30:1787–97. [14] Garattini E, Terao M. The role of aldehyde oxidase in drug metabolism. Expert Opin Drug Metab Toxicol 2012;8:487–503. [15] Lee MO, Manthey CL, Sladek N. Identification of mouse liver aldehyde dehydrogenases that catalyze the oxidation of retinaldehyde to retinoic acid. Biochem Pharm 1991;42:1279–85. [16] Yamaguchi Y, Matsumura T, Ichida K, Okamoto K, Nishino T. Human xanthine oxidase changes its substrate specificity to aldehyde oxidase type upon mutation of amino acid residues in the active site: roles of active site residues in binding and activation of purine substrate. J Biochem (Tokyo) 2007;141:513–24. [17] Schumann S, Terao M, Garattini E, Saggu M, Lendzian F, Hildebrandt P, Leimkühler S. Site directed mutagenesis of amino acid residues at the active site of mouse aldehyde oxidase AOX12. PloS ONE 2009;4:e5348. [18] Hille R, Nishino T, Bittner F. Molybdenum enzymes in higher organisms. Coord Chem Rev 2011;255:1179–205. [19] Hille R. Structure and function of xanthine oxidoreductase. Eur J Inorg Chem 2006: 1913–26. [20] Nishino T, Okamoto K, Eger BT, Pai EF, Nishino T. Mammalian xanthine oxidoreductasemechanism of transition from xanthine dehydrogenase to xanthine oxidase. FEBS J 2008;275:3278–89. [21] Elion, GB. The purine path to chemotherapy. Science 1989;244:41–7. [22] Pacher P, Nivorozhkin A, Szabó C. Therapeutic effects of xanthine oxidase inhibitors; renaissance half a century after the discovery of allopurinol. Pharmacol Rev 2006;58:87–114. [23] Nishino T. The conversion of xanthine dehydrogenase to xanthine oxidase and the role of the enzyme in reperfusion injury. J Biochem (Tokyo) 1994;116:1–6. [24] Berry CE, Hare JM. Xanthine oxidoreductase and cardiovascular disease: molecular mechanisms and pathophysiological implications. J Physiol 2004;555:589–606. [25] Vorbach C, Scriven A, Capecchi MR. The housekeeping gene xanthine oxidoreductase is necessary for milk fat droplet enveloping and secretion: gene sharing in the lactating mammary gland. Genes Dev 2002;16:3223–35. [26] Enroth C, Eger BT, Okamoto K, Nishino T, Nishino T, Pai EF. Crystal structures of bovine milk xanthine dehydrogenase and xanthine oxidase: structure-based mechanism of conversion. Proc Natl Acad Sci U S A 2000;97:10723–8. [27] Nishino T, Okamoto K, Kawaguchi Y, Hori H, Matsumura T, Eger BT, Pai EF, Nishino T. Mechanism of the conversion of xanthine dehydrogenase to xanthine oxidase: identification of the two cysteine disulfide bonds and crystal structure of a non-convertible rat liver xanthine dehydrogenase mutant. J Biol Chem 2005;280:24888–94. [28] Okamoto K, Kawaguchi Y, Eger BT, Pai EF, Nishino T. Crystal structures of urate bound form of xanthine oxidoreductase: Substrate orientation and structure of the key reaction intermediate. J Am Chem Soc 2010;232:17080–3. [29] Truglio JJ, Theis K, Leimkühler S, Rappa R, Rajagopalan KV, Kisker C. Crystal structures of the active and alloxanthine-inhibited forms of xanthine dehydrogenase from Rhodobacter capsulatus. Structure 2002;10:115–25. [30] Olson JS, Ballou DP, Palmer G, Massey V. Reaction of xanthine oxidase with molecular oxygen. J Biol Cem 1974;249:4350–62. [31] Olson JS, Ballou DP, Palmer G, Massey V. The mechanism of action of xanthine oxidase. J Biol Cem 1974;249:4363–82. [32] Hille R, Sprecher H. On the mechanism of action of xanthine oxidase. Evidence in support of an oxo transfer mechanism in the molybdenum-containing hydroxylases. J Biol Chem 1987; 262:10914–7.

122

5 The xanthine oxidoreductase enzyme family

[33] Xia M, Dempski R, Hille R. The reductive half-reaction of xanthine oxidase. Reaction with aldehyde substrates and identification of the catalytically labile oxygen. J Biol Chem 1999;274:3323–30. [34] Okamoto K, Matsumoto K, Hille R, Eger BT, Pai EF, Nishino T. The crystal structure of xanthine oxidoreductase during catalysis: implications for reaction mechanism and enzyme inhibition. Proc Natl Acad Sci U S A 2004;101:7931–6. [35] Tanner SJ, Bray RC, Bergmann. 13C hyperfine splitting of some molybdenum electron-paramagnetic-resonance signals from xanthine oxidase. Biochem Soc Trans 1978;6:1328–30. [36] Huber R, Hof P, Duarte RO, Moura JJ, Moura I, Liu MY, LeGall J, Hille, R, Archer M, Romao MJ. A structure-based catalytic mechanism for the xanthine oxidase family of molybdenum enzymes. Proc Natl Acad Sci U S A 1996;93:8846–51. [37] Leimkühler S, Stockert AL, Igarashi K, Nishino T, Hille R. The role of active site glutamate residues in catalysis of Rhodobacter capsulatus xanthine dehydrogenase. J Biol Chem 2004;279:40437–44. [38] Pauff JM, Zhang J, Bell CE, Hille R. Substrate orientation in xanthine oxidase: crystal structure of enzyme in reaction with 2-hydroxy-6-methylpurine. J Biol Chem 2008;283:4818–24. [39] Dietzel U, Kuper J, Doebbler JA, Schulten A, Truglio JJ, Leimkühler S, Kisker C. Mechanism of substrate and inhibitor binding of Rhodobacter capsulatus xanthine dehydrogenase. J Biol Chem 2009;284:8768–76. [40] Cao H, Pauff JM, Hille R. Substrate orientation and catalytic specificity in the action of xanthine oxidase: the sequential hydroxylation of hypoxanthine to uric acid. J Biol Chem 2010;285:28044–53. [41] Metz S, Thiel WA. Combined QM/MM study on the reductive half-reaction of xanthine oxidase: substrate orientation and mechanism. J Am Chem Soc 2009;131:14885–902. [42] Metz S, Thiel W. QM/MM studies of xanthine oxidase: variations of cofactor, substrate and active-site Glu802. J Phys Chem B 2010;114:1508–17. [43] Schopfer LM, Massey V, Nishino T. Rapid reaction studies on the reduction and oxidation of chicken liver xanthine dehydrogenase by the xanthine/urate and NAD/NADH couples. J Biol Chem 1988;263:13528–38. [44] Hille R, Hagen WR, Dunham WR. Spectroscopic studies on the iron-sulfur centers of milk xanthine oxidase. J Biol Chem 1985;260:10569–75. [45] Iwasaki T, Okamoto K, Nishino T, Mizushima J, Hori H, Nishino T. Sequence motif-specific assignment of two [2Fe-2S] clusters in rat xanthine oxidoreductase studied by site-directed mutagenesis. J Biochem (Tokyo) 2000;127:771–8. [46] Lowe DJ, Bray RC. Magnetic coupling of the molybdenum and iron-sulphur centres in xanthine oxidase and xanthine dehydrogenases. Biochem J 1978;169:471–9. [47] Coffman RE, Buettner GR. General magnetic dipolar interaction of spin-spin coupled molecular dimers. Application to an EPR spectrum of xanthine oxidase. J Phys Chem 1979;83: 2392–400. [48] Anderson RF, Hille R, Massey V. The radical chemistry of milk xanthine oxidase as studied by radiation chemistry techniques. J Biol Chem 1986;261:15870–6. [49] Kobayashi K, Miki M, Okamoto K, Nishino T. Electron transfer process in milk xanthine dehydrogenase as studied by pulse radiolysis. J Biol Chem 1993;268:24642–6. [50] Ball EG. Xanthine oxidase: purification and properties. J Biol Chem 1939;128:51–67. [51] Massey V, Brumby PE, Komai H. Studies on milk xanthine oxidase. Some spectral and kinetic properties. J Biol Chem 1969;244:1682–91. [52] Rajagopalan KV, Handler P. Purification and properties of chicken liver xanthine dehydrogenase. J Biol Chem 1967;242:4097–107. [53] Barrett D, Davidson NA. Xanthine dehydrogenase accumulation in developing Drosophila eyes. J Insect Physiol 1975;21:1447–52. [54] Leimkühler S, Kern M, Solomon PS, McEwan AG, Schwarz G, Mendel RR, Klipp W. Xanthine dehydrogenase from the phototrophic purple bacterium Rhodobacter capsulatus is more

5.7 References

[55] [56] [57] [58] [59]

[60]

[61] [62]

[63]

[64] [65] [66] [67]

[68]

[69]

[70]

[71] [72]

[73]

123

similar to its eukaryotic counterparts than to prokaryotic molybdenum enzymes. Mol Microbiol 1998;27:853–69. Saito T, Nishino T. Differences in redox and kinetic properties between NAD-dependent and O2-dependent types of rat liver xanthine dehydrogenase. J Biol Chem 1989;264:10015–22. Hunt J, Massey V. Purification and properties of milk xanthine dehydrogenase. J Biol Chem 1992;267:21479–85. Nishino T, Nishino T, Schopfer LM, Massey V. The reactivity of chicken liver xanthine dehydrogenase with molecular oxygen. J Biol Chem 1989;264:2518–27. Harris CM, Massey V. The reaction of reduced xanthine dehydrogenase with molecular oxygen. Reaction kinetics and measurement of superoxide radical. J Biol Chem 1997;272:8370–9. Hunt J, Massey V, Dunham WR, Sands RH. Redox potentials of milk xanthine dehydrogenase. Room temperature measurement of the FAD and 2Fe/2S center potentials. J Biol Chem 1993;268:18685–91. Nishino T, Nishino T. The nicotinamide adenine dinucleotide-binding site of chicken liver xanthine dehydrogenase. Evidence for alteration of the redox potential of the flavin by NAD binding or modification of the NAD-binding site and isolation of a modified peptide. J Biol Chem 1989;264:5468–73. Harris CM, Massey V. The oxidative half-reaction of xanthine dehydrogenase with NAD; reaction kinetics and steady-state mechanism. J Biol Chem 1997;272:28335–41. Massey V, Schopfer LM, Nishino T, Nishino T. Differences in protein structure of xanthine dehydrogenase and xanthine oxidase revealed by reconstitution with flavin active site probes. J Biol Chem 1989;264:10567–73. Saito T, Nishino T, Massey V. Differences in environment of FAD between NAD-dependent and O2-dependent types of rat liver xanthine dehydrogenase shown by active site probe study. J Biol Chem 1989;264:15930–5. Ishikita H, Eger BT, Okamoto K, Nishino T, Pai EF. Protein conformational gating of enzymatic activity in xanthine oxidoreductase. J Am Chem Soc 2012;134:999–1009. Elion GB, Kovensky A, Hitchings GH. Metabolic studies of allopurinol, an inhibitor of xanthine oxidase. Biochem Pharmacol 1966;15:863–80. Kumar R, Darpan, Sharma S, Singh R. Xanthine oxidase inhibitors: a patent survey. 2011;2:1071–108. Spector T, Johns DG. Oxidation of 4-hydroxypyrazolo(3,4-d)pyrimidine by xanthine oxidase, the route of electron transfer from substrate to acceptor dyes. Biochem Biophys Res Commun 1968;32(6):1039–44. Spector T, Johns DG. 4-Hydroxypyrazolo(3,4-d)pyrimidine as a substrate for xanthine oxidase: loss of conventional substrate activity with catalytic cycling of the enzyme. Biochem Biophys Res Commun 1970;38(4):583–9. Massey V, Komai H, Palmer G, Elion GB. On the mechanism of inactivation of xanthine oxidase by allopurinol and other pyrazolo[3,4-d]pyrimidines. J Biol Chem 1970;245: 2837–44. Okamoto K, Eger BT, Nishino T, Pai EF, Nishino T. Mechanism of inhibition of xanthine oxidoreductase by allopurinol: crystal structure of reduced bovine milk xanthine oxidoreductase bound with oxipurinol. Nucleos Nucleotid Nucl Acids 2008;27:888–93. Okamoto K, Nishino T. Mechanism of inhibition of xanthine oxidase with a new tight binding inhibitor. J Biol Chem 1995;270:7816–21. Okamoto K, Eger BT, Nishino T, Kondo S, Pai EF, Nishino T. An extremely potent inhibitor of xanthine oxidoreductase. Crystal structure of the enzyme-inhibitor complex and mechanism of inhibition. J Biol Chem 2003;278:1848–55. Fukunari A, Okamoto K, Nishino T, Eger BT, Pai EF, Kamezawa M, Yamada I, Kato N. Y-700 [1-[3-Cyano-4-(2,2-dimethylpropoxy) phenyl]-1H-pyrazole-4-carboxylic acid]: a potent xanthine oxidoreductase inhibitor with hepatic excretion. J Pharmacol Exp Ther 2004;311: 519–28.

124

5 The xanthine oxidoreductase enzyme family

[74] Matsumoto K, Okamoto K, Ashizawa N, Nishino T. FYX-051: a novel and potent hybrid-type inhibitor of xanthine oxidoreductase. J Pharmacol Exp Ther 2011;336:95–103. [75] Mayer MD, Khosravan R, Vernillet L, Wu JT, Joseph-Ridge N, Mulford DJ. Pharmacokinetics and pharmacodynamics of febuxostat, a new non-purine selective inhibitor of xanthine oxidase in subjects with renal impairment. Am J Ther 2005;12:22–34. [76] Becker MA, Schumacher HR, Espinoza LR, Wells AF, MacDonald P, Lloyd E, Lademacher C. The urate-lowering efficacy and safety of febuxostat in the treatment of the hyperuricemia of gout: the CONFIRMS trial. Arthritis Res Ther 2010;12(2):R63. [77] Shoji A, Yamanaka H, Kamatani N. A retrospective study of the relationship between serum urate level and recurrent attacks of gouty arthritis: evidence for reduction of recurrent gouty arthritis with antihyperuricemic therapy. Arthritis Rheum 2004;51:321–5. [78] Kamatani NFS, Hada T, Hosoya, T, Matsuzawa, Y, Ueda, T, Yamanaka, H, Kato R. Febuxostat, a novel non-purine selective inhibitor of xanthine oxidase, in a phase III placebo-controlled double-blind clinical trial in Japanese subjects with gout or hyperuricemia. Arthritis Rheum 2004;50 Supp. [79] Kikuchi H, Fujisaki H, Furuta T, Okamoto K, Leimkühler S, Nishino T. Different inhibitory potency of febuxostat towards mammalian and bacterial xanthine oxidoreductases: insight from molecular dynamics. Sci Rep 2012;2:331. [80] Berglund L, Rasmussen JT, Andersen MD, Rasmussen MS, Petersen TE. Purification of the bovine xanthine oxidoreductase from milk fat globule membranes and cloning of complementary deoxyribonucleic acid. J Dairy Sci 1996;79:198–204. [81] Amaya Y, Yamazaki K, Sato M, Noda K, Nishino T. Proteolytic conversion of xanthine dehydrogenase from the NAD-dependent type to the O2-dependent type. Amino acid sequence of rat liver xanthine dehydrogenase and identification of the cleavage sites of the enzyme protein during irreversible conversion by trypsin. J Biol Chem 1990;65:14170–5. [82] Ichida K, Amaya Y, Noda K, Minoshima S, Hosoya T, Sakai O, Shimizu N, Nishino T. Cloning of the cDNA encoding human xanthine dehydrogenase (oxidase): structural analysis of the protein and chromosomal location of the gene. Gene 1993;133:279–84. [83] Sato A, Nishino T, Noda K, Amaya Y, Nishino T. The structure of chicken liver xanthine dehydrogenase. cDNA cloning and the domain structure. J Biol Chem 1995;270:2818–26. [84] Leimkühler S, Hodson R, George GN, Rajagopalan KV. Recombinant Rhodobacter capsulatus xanthine dehydrogenase, a useful model system for the characterization of protein variants leading to xanthinuria I in humans. J Biol Chem 2003;278:20802–11. [85] Wright RM, Vaitaitis GM, Wilson CM, Repine TB, Terada L, Repine JE. cDNA cloning, characterization, and tissue-specific expression of human xanthine dehydrogenase/xanthine oxidase. Proc Natl Acad Sci USA 1993;90:10690–4. [86] Thoenes U, Flores OL, Neves A, Devreese B, Van Beeumen JJ, Huber R, Romao MJ, LeGall J, Moura JJ, Rodrigues-Pousada C. Molecular cloning and sequence analysis of the gene of the molybdenum-containing aldehyde oxido-reductase of Desulfovibrio gigas. The deduced amino acid sequence shows similarity to xanthine dehydrogenase. Eur J Biochem 1994;220: 901–10. [87] Kuwabara Y, Nishino T, Okamoto K, Matsumura T, Eger BT, Pai EF, Nishino T. Unique amino acids cluster for switching from the dehydrogenase to oxidase form of xanthine oxidoreductase. Proc Natl Acad Sci U S A 2003;100:8170–5. [88] Porras AG, Palmer G. The room temperature potentiometry of xanthine oxidase: pHdependent redox behavior of the flavin, molybdenum, and iron-sulfur centres. J Biol Chem 1982;257:11617–26.

6 Assimilatory nitrate reductase Russ Hille

Abstract Progress in our understanding of the structural and catalytic properties of the eukaryotic nitrate reductases is reviewed, with a focus on work more recent work. After a detailed discussion of the structures of the molybdenum-, heme- and flavin-containing domains of the enzyme, the state of our understanding of the chemical mechanism by which nitrate is reduced to nitrite as well as the kinetic behavior of the enzyme is considered, followed by a discussion of the manner in which the enzyme is post-translationally regulated in the course of the diurnal cycle. Finally, the structures of the active site molybdenum centers of nitrate reductase and the closely related enzyme sulfite oxidase are compared and contrasted, and efforts to convert the sulfite-oxidizing enzyme into a nitrate-reducing one are described.

6.1 Introduction and scope Nitrate reductases catalyze the reduction of nitrate to nitrite, and can be subdivided into two groups: those enzymes from eukaryotes (notably higher plants, fungi and algae) involved in the reductive assimilation of nitrate into the cell, and those from prokaryotes involved in the dissimilatory reduction of nitrate as a terminal respiratory event. These latter enzymes are members of the DMSO reductase family of molybdenumcontaining enzymes [1], which are distinguished by having two equivalents of the pyranopterin cofactor common to all mononuclear molybdenum enzymes coordinated to the metal. The dissimilatory nitrate reductases are typically membrane-bound and use the organism’s membrane-integral quinone pool as the source of reducing equivalents needed to reduce nitrate [2]. They are structurally diverse, with varied subunit and cofactor constitution. The nitrate-inducible NarGHI from Escherichia coli and other nitraterespiring organisms is perhaps the best-studied of these enzymes, with a molybdenum center (at which nitrate is reduced) and a [4Fe-4S] cluster in its cytoplasmically exposed catalytic subunit (NarG), three [4Fe-4S] clusters and a [3Fe-4S] cluster in a separate subunit (NarH), and two b-type cytochromes in a membrane-anchoring subunit (NarI, which contains the site of menaquinol oxidation) [3]. The E. coli genome encodes two other dissimilatory nitrate reductases: a constitutive NarZWY that closely resembles NarGHI; and NapAB, a periplasmic enzyme with the active site molybdenum center (which is similar to that seen in NarG) in NapA and a pair of c-type cytochromes in NapB [4]. In only a few cases (e.g., the dissimilatory reductases from Bacillus subtilis and Klebsiella pneumoniae) do these dissimilatory nitrate reductases appear to possess a flavin site. The eukaryotic assimilatory nitrate reductases are more uniform, each being constituted as a homodimer with a molybdenum center (albeit of a different type than seen

126

6 Assimilatory nitrate reductase

in the dissimilatory enzymes, with only a single equivalent of the pyranopterin cofactor present, see below), a b-type cytochrome and FAD in separately organized domains of each monomer. These enzymes catalyze the first (and rate-limiting) step in the assimilatory reduction of nitrate to ammonia in fungi, algae and higher plants, the reduction of nitrate to nitrite [5–7]. Pyridine nucleotides are the source of reducing equivalents required for catalysis by all these enzymes, those from higher plants utilizing NADH and those from fungal sources NADPH; the algal enzymes usually require NADH, although some non-specific enzymes able to utilize either NADH or NADPH have been described [5]. The present account focuses on these assimilatory nitrate reductases, summarizing in particular advances in our understanding of these enzymes in the past ten years. The reader is referred to other reviews covering earlier work on the assimilatory enzymes [5–7] and the dissimilatory enzymes [2].

6.2 Enzyme structure Nitrate reductase from a variety of higher plants has been characterized and that from Arabidopsis thaliana is typical, being a homodimer of 2 x 110 kDa [8]. Each subunit consists of a large N-terminal portion (~59 kDa) that possesses the active site molybdenum center at which nitrate is reduced to nitrite, a small central domain possessing a b-type cytochrome (14 kDa), and a C-terminal domain containing FAD and the NADH binding site (24 kDa) [9]. The structure of holo nitrate reductase and the specific orientation of the molybdenum, heme and flavin domains with respect to one another is at present unknown, but crystal structures have been reported for the molybdenum domain of the Pichia angusta enzyme [10], for bovine cytochrome b5 [11,12] and the flavin domain of the Zea mays nitrate reductase [13] (򐂰Fig. 6.1). The molybdenum domain from P. angusta bears a striking structural resemblance to the corresponding portion of both the chicken [14] and A. thaliana [15] sulfite oxidases, with molybdenum-binding and dimerization subdomains, as shown in 򐂰Fig. 6.2. The molybdenum center itself closely resembles that seen in sulfite oxidase, with a square-pyramidal LMoVIO2(S-Cys) molybdenum center in the oxidized enzyme, with one of the Mo=O groups occupying the apical position. A variety of EPR [16,17], X-ray absorption [18,19] and crystallographic studies [10,14,15] indicate that upon reduction the equatorial Mo=O becomes protonated. As with the sulfite oxidases, the equatorial Mo=O of the oxidized molybdenum coordination sphere of the P. angusta nitrate reductase faces into the solvent access channel and the substrate binding site (see below) and is generally presumed to be catalytically labile, being either transferred to (in the case of sulfite oxidase) or from (in the case of nitrate reductase) substrate. The overall direction of oxygen atom transfer for the two enzymes has been rationalized on the basis of the relative stabilities of S=O, Mo=O and N=O bonds [20]. In the overall structure of the molybdenum-containing fragment of the P. angusta nitrate reductase [10], consisting of residues 1–484 of the holoenzyme, the first 25 and last six amino acid residues are not resolved in the structure, presumably due to static or dynamic disorder (򐂰Fig. 6.2). The crystallographically resolved residues 26–37 constitute the end of an N-terminal extension that leads into the molybdenum-binding domain of the fragment, which consists of residues 38–299 and is comprised of one three- and two five-stranded β-sheets, with nine short interspersed α-helices (򐂰Fig. 6.2).

6.2 Enzyme structure

127

B A

C

Fig. 6.1: Ribbon representations of the three domains of nitrate reductase. The protein surface rendered in mesh. Cofactors are rendered in CPK coloring and metal centers rendered as spheres. (A) The molybdenum domain from Pichia angusta (2BIH), looking down the solvent access channel to the Mo center (Cys139, coordinating the molybdenum, lies to the right of the metal). As shown, the dimerization subdomain for the α 2 holoenzyme lies to the right of the molybdenumbinding subdomain. (B) The heme domain of bovine cytochrome b5 (1CYO). (C) The flavin domain from corn, with the NADH binding cleft at top, above the FAD cofactor.

An 18-residue loop (consisting of residues 141–158) with a central α-helix extends out from the central β-sheet and intercalates into the remainder of the structure, which constitutes the dimerization subdomain (residues 300–484). The dimerization domain itself is dominated by four- and six-stranded antiparallel β-sheets in an extended Greek key motif. A long loop with significant secondary structure, residues 358–385, extends from this core to one side of the molybdenum-binding portion of the protein. The pyranopterin cofactor of the molybdenum center is protected from solvent by two short loops (residues 90–93 and 197–200), and is not nearly as deeply buried (~4 Å) as seen in members of the xanthine oxidase family of molybdenum-containing enzymes. The cofactor has some 14 hydrogen-bonding interactions with the polypeptide that are highly conserved in the nitrate reductases and sulfite oxidases; the cysteine residue coordinating the molybdenum is Cys139 in the P. angusta nitrate reductase (equivalent to Cys191 in the A. thaliana nitrate reductase, Cys98 in A. thaliana sulfite oxidase and Cys185 in chicken sulfite oxidase). The substrate binding site in the P. angusta nitrate reductase fragment consists of two arginine residues, Arg89 and 144, as well as Trp158 (򐂰Fig. 6.2); all three of these residues are conserved with the sulfite oxidases. The substrate binding site also includes three residues conserved among the nitrate reductases but different in the sulfite oxidases: Met427 (a Val in the sulfite oxidases), Asn272 (a Tyr in the sulfite oxidases) and Thr425 (an Arg in the sulfite oxidases). The extent to which these residues impart substrate specificity is discussed further below. The molybdenum- and heme-containing domains of nitrate reductase are connected by a tether approximately 50 amino acid residues long (approximately residues 491 to

128

6 Assimilatory nitrate reductase

O

S•

H

S•

H H2N

O

OPl

H Asn272

Met 427

Asn272

Thr 425 Arg144

Trp158

Val 376 Tyr 241 Arg 374 Arg103

Thr 425

Cys139 Arg 89

Met 427

Cys139 Cys 98

Arg144 Trp158

Arg 89

Tyr 241 Cys 98

Arg 51

Val 376 Arg 374

Arg103 Arg 51 Trp117

Trp117

Fig. 6.2: A comparison of the structures of the molybdenum-containing domains of P. angusta nitrate reductase (left, 2BIH)) and A. thaliana sulfite oxidase (right, 1OGP). (Top) The overall protein folds of the monomer, with the molybdenum subdomains in gray and the dimer interface subdomains in blue. (Bottom) The active sites of the two enzymes, with residues involved in interacting with substrate indicated. For both active sites, the orientation to the right is rotated approximately 90° about the vertical with respect to that on the left. (Center) The disposition of the two monomers of the A. thaliana sulfite oxidase within the dimer. Also shown is the structure of the pyranopterin cofactor common to all molybdenum-containing enzymes (other than nitrogenase).

541 in the holoenzyme from A. thaliana, which will be used as the point of reference in discussing the structure of the holoenzyme). As indicated above, the heme-containing domain has significant sequence homology to the heme domain of sulfite oxidase and the mammalian cytochromes b5 (򐂰Fig. 6.3). The heme domain of both chicken [14] and human sulfite oxidase [21] consists of a 5-stranded β-sheet covering one end of a short four-helix bundle (inserted between strands 2 and 3 of the sheet); an additional α-helix at the N-terminus flanks the β-sheet opposite the four-helix bundle. Over a stretch of 59 amino acid residues, the Cα carbons of chicken heme domain and the bovine cytochrome b5 [14] have an rms deviation of 0.86 Å, reflecting a very high degree of structural homology. The heme, coordinated to His40 and 65 in the chicken enzyme, lies in the middle of the helical bundle with its two propionate side chains protruding from the end to the surface of the protein/domain. Unlike the situation found in cytochrome c, the protein surface surrounding the exposed propionates is not highly charged (other than that attributable to the propionates themselves) and is not well-conserved among the nitrate reductases, sulfite oxidases and cytochrome b5. Kisker and coworkers [21] have undertaken a comprehensive analysis of the sulfite oxidase heme domain and the crystallographically characterized cytochrome b5’s, and

6.2 Enzyme structure

129

Fig. 6.3: A comparison of the structures of the heme domains of human (left, 1MJ4) and chicken sulfite oxidases (center, 1SOX) with bovine cytochrome b5 (right, 1CYO). The histidine ligands to the heme are shown.

identified specific structural features (principally the lower solvent accessibility of the heme in sulfite oxidase and its more positive electrostatic environment) that account for the significantly higher reduction potential for the heme seen in the sulfite oxidase domain as compared to cytochromes b5. A comparison of the sequence of the heme domain of nitrate reductase with those of sulfite oxidase and cytochrome b5 suggest, particularly with respect to the several amino acid residues to the N-terminal side of the second histidine coordinating the heme iron, that the heme of nitrate reductase should have a lower potential than even the cytochrome b5. This sequence is WALYAVH64 in chicken sulfite oxidase, FEDVGH62 in bovine cytochrome b5 and EFEATH603 in A. thaliana nitrate reductase. Such a lower potential is in fact observed experimentally in the case of both spinach and Chlorella vulgaris nitrate reductases [22,23], and is expected on the basis of the physiological direction of electron transfer in nitrate reductase (from heme to molybdenum) as compared to that in sulfite oxidase (from molybdenum to heme). The heme- and FAD-containing domains are connected by a second, ~30-amino acid tether (approximately residues 621–650 in the A. thaliana protein). The flavin-binding domain itself belongs to the well-characterized family of ferredoxin:NADP+ reductaselike (FNR) flavoproteins [24], and the X-ray crystal structure [13] of the fragment from Z. mays nitrate reductase (򐂰Fig. 6.4A) consists of two-well-defined regions: an Nterminal FAD-binding subdomain (approximately residues 651–757 in the A. thaliana protein, in gray), and a C-terminal NADH-binding subdomain (approximately residues 758–917 in the A. thaliana protein, in blue). The FAD-binding subdomain consists of a six-strand antiparallel β-barrel, with the barrel distended at one end by a short α-helix inserted between the fifth and sixth β-strands so as to accommodate the isoalloxazine ring of the FAD. The FAD itself is flanked by strands 4 and 5 of the β-sheet, with its si face lying against the fourth strand and a highly conserved tyrosine residue (Y714 in the A. thaliana protein); the amino terminus of this helix interacts with the diphosphate bond of the FAD. A short linker connects to the NADH-binding subdomain, which has a central six-stranded parallel β-sheet flanked by two pairs of α-helices in the dinucleotide binding fold that is characteristic of this family of flavoproteins. As with other members of the FNR family, It has proven difficult to fully define the NADH binding site of nitrate reductase [24]. Soaking pre-grown crystals of these

130

6 Assimilatory nitrate reductase A

C

B

D

Fig. 6.4: A comparison of the FAD fragment of Z. mays nitrate reductase with ferredoxin: NADP+ reductase from yeast. The FAD- and NADH-binding portions of the fragment are indicated in gray and blue, respectively. Structures of the fragment in the absence (A) and presence (B) of ADP, with both FAD and ADP rendered in CPK. (C) perspective rotated 90° about the horizontal from that shown above, showing the putative interaction surface of the fragment that is likely presented to the heme domain of the enzyme. His48, His118 and the C-terminal Phe270, referred to in the text, are indicated, and the surface of the fragment is rendered in mesh. (D) CPK rendering of the putative interaction surface.

proteins with NAD(P)H or NAD(P)+ usually cracks the crystal, and co-crystallized complexes invariably show the pyridine nucleotide bound with the nicotinamide ring pointing away from the flavin in a configuration that is not conducive to electron transfer. A major part of the problem appears to be that the C-terminal tyrosine typically found in these enzymes (but a phenylalanine in the case of the corn nitrate reductase) stacks onto the re face of the isoalloxazine ring and must be displaced by the nicotinamide ring of the incoming pyridine nucleotide. The picture that emerges from a consideration of a number of structures with bound pyridine nucleotide or (P)ADP is that pyridine nucleotide binds with its (P)ADP moiety tightly bound in the NAD(P)H-binding subdomain, with the nicotinamide ring less tightly bound and dynamically sampling a large conformational space (in other words, just flopping around and only occasionally approaching the flavin in a manner conducive to electron transfer). In this context, the structures of enzyme complexes with ADP provide some indication of where the NADH must bind [13,25]. As shown in 򐂰Fig. 6.4B), ADP binds to the FAD fragment of corn

6.3 Kinetics and mechanism

131

nitrate reductase with its adenosine wedged between one pair of the flanking α-helices in the domain, at the end of the β-sheet facing the FAD-binding domain. Its diphosphate group extends toward the FAD-binding subdomain, and it is clear that the nicotinamide ring of NADH spans the interface of the two subdomains to interact with the re face of the flavin isoalloxazine ring. With NADH tethered to its subdomain via a comparable interaction, it is easy to envisage how its binding permits a high degree of conformational sampling of the nicotinamide ring in the complex. By analogy to the structures of various members of the FNR family in complex with their physiological partners [26], it is very likely that the heme domain interacts, if only transiently, with the flavin domain of nitrate reductase at the portion of the surface of the latter where its flavin C-8 methyl protrudes to solvent (򐂰Fig. 6.4C,D). The flavin C-8 methyl is flanked by His48 and His118 in the corn FAD fragment (equivalent to the conserved residues His698 and His768 of the A. thaliana enzyme), which along with the C-terminus of the enzyme may help position the heme domain (with possible cognate residues Asp579, Asp585 and Lys606) for electron transfer. It is likely that any such complex formed is only transient, as the heme domain must undoubtedly reorient significantly within holo nitrate reductase, once reduced, to present its redox-active face to the molybdenum domain of the enzyme.

6.3 Kinetics and mechanism In the course of turnover, NADH introduces reducing equivalents into the enzyme at the FAD in the reductive half-reaction of the overall catalytic cycle, and these are subsequently transferred via the heme to the molybdenum center, where nitrate is reduced to nitrite in the oxidative half-reaction [5–7]. Reduction of the LMoVIO2(S-Cys) core of the active site yields LMoIVO(OH)(S-Cys), consistent with X-ray absorption studies of the enzyme [18,19]. As indicated in 򐂰Fig. 6.5, it is generally believed that with each catalytic cycle the equatorial Mo=O (after reduction, protonation and displacement by nitrate) is lost to solvent, and is subsequently regenerated with oxygen derived from substrate in the oxygen atom transfer event. The basic chemistry of nitrate reduction is thus believed to be straightforward, and functionally the reverse of the oxygen atom transfer with sulfite oxidase, with nitrate serving as oxygen atom donor and the reduced molybdenum center functioning as the oxygen atom acceptor. Again, the basis for the nitrate-utilizing enzyme being a reductase and the sulfite-utilizing enzyme an oxidase has been accounted for on the basis of the relative thermodynamic stabilities of N=O, S=O and Mo=O bonds [20]. As in the case of sulfite oxidase, unfortunately, there is to date no direct experimental evidence for the direct oxygen atom transfer envisaged by this mechanism.

O S Cys VI S Mo O S 2[e앥], H쎵

O S Cys VI S Mo OH S NO3앥 HO앥

O S Cys VI S Mo O 쎵 S N O 앥

O

O S Cys VI S Mo O S NO2앥 N O

O S Cys VI S Mo O S

앥

O

Fig. 6.5: The proposed reaction mechanism for the assimilatory nitrate reductases with nitrate.

132

6 Assimilatory nitrate reductase

The general steady-state kinetics of several algal and higher plant nitrate reductases have been summarized previously [5–7]. The specific steady-state kinetic parameters for the A. thaliana enzyme are kcat = 210 s−1; Kmnitrate = 90 µM; and KmNADH = 0.8 µM at pH 7.0, 30°C [7]. More detailed rapid reaction kinetic analysis of the enzyme’s behavior has been limited by the low availability of the enzyme, although in the past decade progress has been made in the development of suitable expression systems of individual domains of nitrate reductase and, most recently, the holoenzyme. The flavin domain of nitrate reductase (sometimes referred to as the cytochrome b5 reductase fragment on the basis of its ability to transfer reducing equivalents from NADH to cytochrome b5) has a cysteine residue that is highly conserved among the nitrate reductases (but not found in other members of the FNR family of flavoproteins). This is Cys242 in the corn cytochrome c reductase fragment, Cys889 in the full-length A. thaliana enzyme and Cys833 in the P. angusta enzyme, and as indicated in 򐂰Fig. 6.4 is located in the NADH-binding subdomain near the C-terminal phenylalanine. Mutation of this residue to serine does not alter the absorption spectrum of the oxidized enzyme, and the flavin midpoint potential is increased only modestly from –287 mV to –279 mV [27]. The reaction of the flavin fragment of corn nitrate reductase, and also a mutant in which the highly conserved Cys242 is replaced with a serine residue (C242S), with NADH has been examined [27]. For both wild-type and mutant protein, the observed rate constant for the reaction exhibits a hyperbolic dependence on [NADH] – the limiting rate constant for reduction at high [NADH] for the C242S mutant is some seven-fold slower than that for the wild-type protein (68 s−1 as compared with 478 s−1) and KdNADH is larger by a factor of two (6 µM vs. 3 µM). It is thus evident that Cys242 plays a modest role in facilitating electron transfer from NADH to the flavin and only a minimal role in binding of NADH to the enzyme, its mutation to Ser reducing kred/KdNADH by a factor of just over ten. Interestingly, the product seen at the end of the reaction is not free reduced protein with FADH2, but rather an FADH2•NAD+ charge-transfer complex with a characteristic long-wavelength absorption with a maximum at 800 nm (ε = 2.7 mM−1cm−1) [27]. The reaction of a combined flavin-heme fragment of nitrate reductase from spinach (referred to as the cytochrome c reductase fragment on the basis of its ability to pass reducing equivalents on to cytochrome c rather than cytochrome b5) with NADH has also been examined [28]. As expected, the reaction exhibits more complex multiphasic kinetics. The faster, catalytically relevant kinetic processes include a fast phase whose associated spectral change reflects formation of the same FADH2•NAD+ complex as seen with the flavin fragment alone. kobs again exhibits a hyperbolic dependence on [NADH] with a kred of 560 s−1 and a Kd of 3 µM, in good agreement with the results using the (corn) flavin fragment alone (478 s−1 and 3 µM, respectively). Surprisingly, the subsequent intermediate kinetic phase, whose associated spectral change indicates that it is due to internal electron transfer from FADH2 to heme is relatively slow, 12 s−1, and appears to be rate-limited by dissociation of NAD+ from the Ered•NAD+ chargetransfer complex. The slowest phase of the reaction has been attributed to the slow intermolecular disproportionation of the two-electron reduced protein formed in the faster phases to yield one equivalent of the fully reduced 3-electron reduced fragment and one one-electron reduced fragment, which can then react rapidly with a

6.4 Post-translational regulation

133

second equivalent NADH in a process rate-limited by the intermolecular electron transfer event. The rapid reaction kinetics of holo nitrate reductase from A. thaliana has also been examined [29]. The kinetics of enzyme reduction by NADH largely confirmed the above results with the flavin-heme fragment, with a fast phase of ~700 s−1 observed at 70 µM NADH that corresponded to reduction of FADH2, and a subsequent slower phase at 28 s−1 corresponding to electron transfer on to the heme. No evidence is seen for formation of an FADH2•NAD+ charge-transfer complex with the A. thaliana enzyme, however, and its role in catalysis has therefore not been confirmed. Further complicating the issue, a significant amount of heme reduction has been reported in the course of the fast phase of the reaction. Together the two results are at least internally consistent, in that electron transfer from the flavin on to the heme would be expected to be fast absent a charge-transfer complex to retard oxidation of the flavin, but it is surprising that the complex would be so readily observed in the case of the corn [27] or spinach [28] proteins and not form at all in the A. thaliana protein. Somewhat surprisingly, the kinetics of the reoxidation of pre-reduced nitrate reductase with nitrate were not examined in this study, perhaps owing to the inevitable complications due to the fact that only ~20% of the protein possessed a molybdenum center. A series of enzyme-monitored turnover experiments were performed, however, in which enzyme at a concentration of, for example, 2 µM enzyme pre-reduced with 40 µM NADH was reacted with 90 µM nitrate. The kinetics for the approach to steady-state, as followed by the absorbance changes at 460 nm (following changes in the oxidation state of the FAD) and 557 nm (following changes in the oxidation state of the heme), are biphasic with apparent rate constants of 260-270 s−1 and 6–8 s−1, respectively. The steady-state persists for over 1.0 second, with the FAD held at 84% reduced and the heme 42% (presumably reflecting the thermodynamic distribution of reducing equivalents in the partially reduced enzyme in the steady-state). At the end of the steady-state, with the NADH presumably depleted and 50 µM nitrate remaining in solution, the enzyme slowly reoxidizes with an apparent rate constant of 0.5 s−1. This last process is too slow to support turnover, and the possibility exists that most of the observed kinetics involves the large proportion of enzyme lacking a molybdenum center (which, despite being readily reducible by NADH, would not be reoxidized by nitrate), with reducing equivalents being passed in an intermolecular process to the small population of enzyme that was functional and could react with nitrate. The implication is that the magnitude of the absorbance changes seen in the initial approach to steady-state means that the functional enzyme is essentially completely reoxidized in the steady-state, which would require that the reoxidation of enzyme by nitrate be considerably faster that its rate of reduction by NADH, and be entirely rate-limiting to catalysis.

6.4 Post-translational regulation The activity of assimilatory nitrate reductases is under tight control so as to limit the intracellular accumulation of the reactive product nitrite, and includes regulation at the transcriptional, translational and post-translational levels [5–7]. The last plays an

134

6 Assimilatory nitrate reductase

important role in lowering nitrate reductase activity at night, when photosynthetically generated reducing equivalents are not available to reduce the product nitrite on to ammonia, as nocturnal nitrite levels in the plant tissues would otherwise increase to dangerous levels [30]. This post-translational regulation of nitrate reductase involves phosphorylation at a serine residue in the hinge 1 region between the molybdenum and heme domains [31]. The modified residue has been identified as Ser543 in the spinach protein [32,33] (equivalent to Ser534 of the A. thaliana enzyme), which lies near the C-terminal end of the ~50 amino acid tether between the molybdenum- and heme-containing domains of the enzyme (with respect to the heme domain, this places the serine on the side of the domain’s β-sheet opposite the 4-helix bundle that binds the heme). Phosphorylation is carried out by any one of several specific protein kinases (calcium-dependent protein kinase, CPK [34] and the AMP-activated SNF-related kinases SNRK1 and WPK4 [35–37]). The hinge 1 region of nitrate reductase in fact has consensus recognition sequences for both CPK [34,38,39] and SNRK1 [40,41] families of kinases, which are known to have overlapping specificity [34,35,42]. Phosphorylation per se does not inhibit activity, but rather creates a recognition site that recruits a specific regulatory element, a member of the 14-3-3 family of proteins, whose binding (in the presence of Mg2+, which is required for inhibition) effectively abolishes activity [43]. Of the thirteen 14-3-3 protein isoforms encoded by the A. thaliana genome, the ω isoform has been shown to be particularly effective in binding to and inhibiting the phosphorylated form of nitrate reductase [44,45]. Recently, the steady-state kinetic properties of A. thaliana nitrate reductase in asisolated, phosphorylated and phosphorylated-(14-3-3)ω complexed forms have been investigated [45]. Of the several calcium-dependent protein kinases encoded by the A. thaliana genome, CPK-17 efficiently phosphorylated nitrate reductase at Ser534 in vitro (other secondary sites were also identified that were less effectively phosphorylated by CPK-17, including Ser34 in the N-terminal extension of the protein [45]). In conventional NADH-nitrate steady-state assays, the ω, κ and λ isoforms of 14-3-3 are found to be most effective in binding to and inhibiting phosphorylated nitrate reductase, with Ki’s of 60 and 80 nM, respectively. The as-isolated enzyme exhibits a kcat of 20 s−1, Kmnitrate of 197 µM and KmNADH of 18 µM, phosphorylated enzyme has a somewhat higher kcat and Kmnitrate of 33 s−1 and ~200 µM, respectively, while the phosphorylated enzyme complexed with 14-3-3ω has a substantially reduced kcat of 1.8 s−1 and Kmnitrate of 141 µM. An S534A mutant of the nitrate reductase is not inhibited by 14-3-3ω, underscoring the importance of the phosphorylated site in formation of the inhibitory complex. Significantly, phosphorylation and 14-3-3ω binding has no effect on turnover with NADH and cytochrome c as substrate [45]. As indicated above, cytochrome c accepts reducing equivalents from nitrate reductase at the heme site of the latter, and the results suggest that neither the reductive half-reaction nor electron transfer from the flavin to the heme are affected by 14-3-3ω binding. Similarly, 14-3-3ω binding does not influence the Ki for product nitrite, and it has been concluded that substrate binding similarly is unaffected (as suggested by the only modest changes in Kmnitrate seen in the steady-state assays). A subsequent rapid kinetic study of the system [46] has demonstrated that phosphorylation and 14-3-3ω binding slows the limiting rate constant for reoxidation of the heme site of a reduced molybdenum-heme fragment of A. thaliana nitrate reductase on reaction with saturating concentrations of nitrate by a factor of approximately ten (from

6.5 Interconversion of sulfite oxidase and nitrate reductase activities

135

310 s−1 to 35 s−1). Importantly, phosphorylation and 14-3-3ω binding has no discernible effect on the rate of reaction of the molybdenum center of the molybdenum-heme fragment with nitrate, and it has been concluded that the basis for 14-3-3ω inhibition of the phosphorylated nitrate reductase is a direct reduction in the rate of electron transfer from the heme to the molybdenum in the enzyme rather than a reduction in the rate of nitrate reduction at the molybdenum center per se. It is reasonable that this might be the case, since as indicated above the heme domain must very likely move substantially within the holoenzyme to present the face on to which the heme edge protrudes to first the flavin (to pick up a reducing equivalent) and then the molybdenum center (to transfer that reducing equivalent to the active site). Precedent for such a motion exists in the observed motion of the Rieske domain of cytochrome bc1, which must similarly interact with spatially separated sites in the protein [47].

6.5 Interconversion of sulfite oxidase and nitrate reductase activities As indicated above, the assimilatory nitrate reductases belong to the same family of molybdenum-containing enzymes as sulfite oxidase, with certain active site residues in common and others distinct between the nitrate- and sulfite-utilizing enzymes. Reviewing the substrate binding site of the P. angusta enzyme (򐂰Fig. 6.2), it includes Arg89, Arg144 and Trp158, conserved with the sulfite oxidases, as well as Met427 (a Val in the sulfite oxidases), Asn272 (a Tyr in the sulfite oxidases) and Thr425 (an Arg in the sulfite oxidases). These cognate residues and selected other conserved residues among the nitrate reductases and sulfite oxidases are summarized in 򐂰Tab. 6.1 for the enzymes from several species. It is to be noted that the position corresponding to Thr425 in the P. angusta nitrate reductase is more commonly a methionine in other nitrate reductases. While no site-directed mutagenesis work has yet been done with the molybdenum

Tab. 6.1: Homologous conserved amino acid residues among the nitrate reductases and sulfite oxidases. Included are the cysteine residues that coordinate the molybdenum in the active site (top) and those homologous to Cys242 of the corn FAD-binding fragment (cytochrome c reductase) that has been investigated by site-directed mutagenesis [27] (bottom). This domain is absent in the sulfite oxidases. P. angusta NR

A. thaliana NR

A. thaliana SO

Chicken SO

Human SO

Cys139

Cys191

Cys98

Cys185

Cys207

Arg89

Arg143

Arg51

Arg138

Arg160

Arg144

Arg196

Arg103

Arg190

Arg212

Trp158

Trp210

Trp117

Trp204

Trp226

Asn272

Asn333

Tyr241

Tyr322

Tyr343

Thr425

Asn472

Arg374

Arg450

Arg472

Met427

Met474

Val376

Val452

Val474

Cys833

Cys889

–

–

–

136

6 Assimilatory nitrate reductase

domain of the assimilatory nitrate reductases, a study has been undertaken recently attempting to modify the molybdenum domain of sulfite oxidase so as to reverse its substrate specificity [48], using the known structure of chicken sulfite oxidase and the observed sequence homologies to rationally engineer in modified activity. The platforms used for these mutagenesis studies has been the heterologous expression systems for either the chicken or human sulfite oxidase in E. coli. 򐂰Tab. 6.2 summarizes the steady-state results at pH 8.5 seen with wild-type human and chicken sulfite oxidase, along with double and triple mutants in the substrate binding site. It is evident that in both cases the double mutant, involving mutation of Tyr343/322 to Asn and Arg472/450 to Met (where the residue numbers are for the human/chicken enzyme), exhibits significantly compromised catalytic activity toward sulfite as reducing substrate, with kcat and kcat/Km reduced by factors of 20–30 and 104–106, respectively. Importantly, both double mutants exhibit a significant degree of nitrate reductase activity, with kcat 0.4–1.0 s−1 and kcat/Km ~103 M−1s−1. Interestingly, for the triple mutant that additionally included switching Val474/452 to Met, reactivity toward both sulfite and nitrate increased approximately twofold as compared to the double mutants. Unfortunately, in this study rapid-reaction experiments were limited to two singlemutants of the human sulfite oxidase, R472Q and R472M, and these examined only the reaction with sulfite. A comparison of the steady-state and rapid-reaction results illustrates the difficulty in drawing concrete thermodynamic conclusions on the basis of steady-state results alone. As illustrated in 򐂰Tab 6.3, it is evident from a detailed examination that the trends in the two data sets are not the same. While the trends in kred/Kd and kcat/Km are indeed very similar, with either mutation compromising both parameters by 10– to 40-fold; the effect on kred/Kd is a modest fourfold greater than on kcat/Km (where these parameters each reflect the slope of the hyperbolic plot of rate (constant) versus [substrate] in the low-substrate regime). On closer inspection, however, it is evident that the principal effect of each mutation is on kcat rather than Km in the steady-state

Tab. 6.2: Steady-state kinetic parameters of wild-type and various double and triple mutants of sulfite oxidase with nitrate reductase activity [48]. Superscripts refer to human or chicken enzyme. h

h

w-t

Y343N/

R472M

h

Y343N/

c

w-t

c

c

R450M

R450M/

Y322N/

R472M/ V474M

sulfite cat

k

sulfite

Km

−1

(s )

27

(µM)

8.3

sulfite

V452M

1.4

3.9

71

43,000

14,000

8.4

2.5

4.1

12,000

28

8.3 x 10

2.1 x 10

6.1 x 102

700 nm) and FAD/NAD(P)H (λmax ~650 nm) pairs. As noted qualitatively in the scheme for reduction, the Eox•NADPH species, and Ered•NADP+ when it is observed, are typically formed very rapidly and within the “deadtime” of the stopped flow instruments used to study these, which is typically ≤1.5 ms. For mercuric ion reductase, these species appear essentially simultaneously and decay together in a single exponential process to give EH2•NADP+ directly indicating they exist in rapid equilibrium and that transfer of electrons from reduced flavin to the disulfide is rate limiting in reduction [7,8]. As indicated in the scheme, transfer of electrons to the disulfide is proposed to occur by attack of the flavin C4a on the disulfide (or CysC-Hg-CysN complex described in Section 8.3 for MerA) to form a flavin C4a-S-CysC thiol adduct (򐂰Fig. 8.6, upper right), which breaks down very rapidly as it is not observed at physiological pH values during reduction of any of these enzymes. It is partially observed, however, in MerA during reduction at lower pH [9] and has been observed to form in CysN-alkylated or mutated forms of some of the enzymes including LipDH [10] and MerA [11]. In some Group 1-fold enzymes, Ered•NADP+ is not observed and Eox•NADPH decays directly to EH2•NADP+ indicating reduction of the flavin is rate limiting. In any case, as indicated by the brackets, Eox•NADPH and Ered•NADP+ are often only observed as transient intermediates in these enzymes. The final observed species during reduction depends both

172

8 FDR and structurally related flavoprotein thiol/disulfide-linked oxidoreductases Ping-pong redox cycle for reductase or dehydrogenase

Eox

FAD

2 RSH

NAD(P)H

S 쎵2e앥 앥2e앥

쎵/앥 H쎵

S FAD

EH2

RS

SR

FADH앥

앥

S

S

SH

S

NAD(P)쎵

Ternary complex dithiol dehydrogenase cycle NAD쎵 Eox

Inactive

FAD

FAD

S

FAD NAD쎵 S

S

S

S

NAD(P)H

쎵2e앥 앥2e앥

EH2 2 RSH 쎵 H쎵

쎵2e앥 앥2e앥 RS EH4

SR

NADH

NAD(P)쎵

NAD쎵

FAD

FAD

S

S앥

SH NAD(P)쎵 NAD(P)H

SH NAD쎵 NADH

NAD(P)H

NADH

FAD

FAD

앥

앥

S

SH Ternary complex disulfide reductase cycle

S

S앥

2 RSH 쎵

앥H RS

SR

Inhibited

SH

Fig. 8.5: Redox cycles for Group 1 FDR enzymes. (Top) Ping-pong reaction where each substrate undergoes oxidation/reduction with its relevant redox center and dissociates prior to the other substrate binding and reacting. This cycle, involving only Eox and EH2 forms of the enzyme, may occur in either the disulfide reductase or dithiol dehydrogenase direction. (Bottom left) Ternary complex mechanism for disulfide reductase enzymes that cycle between EH2 and an EH4 form that typically is a complex of EH2 with NAD(P)H. In this case, Eox is inactive and must be primed by reduction with one equivalent of NAD(P)H. (Bottom right) Ternary complex mechanism for dithiol dehydrogenase enzymes that cycle between the pyridine nucleotide bound complexes at the Eox and EH2 redox states. In this case, the EH4 form inhibits oxidation of the dithiol substrate.

on the final concentrations of NADP+ and NADPH in the reaction and their relative affinities for EH2. For MerA, the binding of NADPH to EH2 is favored by ~10-fold over NADP+ yielding the EH2•NADPH complex (򐂰Fig. 8.6 spectra), and the enzyme cycles through the EH2 and EH4 states of a ternary complex mechanism (򐂰Fig. 8.5) [12]. As noted in the bottom panel of 򐂰Fig. 8.6, reduction of the disulfide substrates in the Group 1 enzymes occurs by a simple thiol/disulfide interchange reaction. Studies again

8.2 Group 1 FDR enzymes

Ext. coeff. (mM –1 cm–1)

NADP쎵 H N N O

Eox Eox • NADPH/ Ered• NADP쎵 EH2 • NADP쎵 EH2 • NADPH

10 8 6

R

N H

4

Cys C S CysN S

2

O

500

600

NADP쎵

N

NH

N H S

Cys C CysN SH

Ered • NADP쎵

0 400

R

173

N

O NH

O

C4a adduct

700

Wavelength (nm) v. fast FAD

NAD(P)H

FAD

FADH v. fast

S

Eox • NAD(P)H

NAD(P)쎵

FADH “slow”

S

S

S Eox

NAD(P)쎵

앥

S

S

NAD(P)쎵

NAD(P)H

S SH

Ered • NAD(P)쎵

C4a adduct

FAD v. fast

S앥 SH EH2 • NAD(P)쎵

SR FAD 앥

SR

RSH

FAD

S

S

S

S

EH2

RSH

앥

FAD S

SR

MDS

S Eox

Fig. 8.6: Characteristic spectra and half reaction mechanisms of pyridine nucleotide complexes of Group 1-fold enzymes. (Upper left) Spectra obtained in order listed during reduction of mercuric ion reductase with ≥2 equivalents of NADPH [7,91]. (Upper right) Chemical structures of Ered•NADP+ and FAD-C4a-CysC thiol adduct. (Middle) Species corresponding to spectra in reduction mechanism [7,91]. The Eox•NADPH/Ered•NADP+ species occur only transiently during reduction, while EH2•NADP+ and EH2•NADPH are stably formed in anaerobic titrations [7,8,92]. The FAD-C4a-CysC thiol adduct is not observed during reduction at physiological pH. (Lower) Mechanism of disulfide reduction by Group 1 FDR enzymes where MDS is the mixed disulfide of the interchange cysteine thiol (CN) and a substrate thiol [2,3]. Dashed lines in structures and mechanisms indicate charge transfer interactions between electron rich donors and electron deficient acceptors.

with lipoamide dehydrogenase [13] demonstrated that the N-terminal cysteine (CysN) of the CNXXXXCC motif initiates attack on the bound disulfide substrate to form a mixed disulfide (MDS) with one substrate with concomitant release of one reduced product thiol. With this functional role, CysN is typically referred to as the interchange thiol, while CysC is typically referred to as the charge-transfer thiol because of its spectrally observable interaction with the flavin as noted above. After the MDS forms, CysC attacks CysN to reform the enzyme disulfide concomitant with reduction/release of the second product thiol. This scheme is quite general for all enzymes possessing a Group 1 fold in that CysN always takes the lead in attacking (or being attacked by) the substrate, regardless of whether the substrate is an external disulfide, a C-terminal redox center, or some other type of substrate in those enzymes with novel activities.

174

8 FDR and structurally related flavoprotein thiol/disulfide-linked oxidoreductases

Details of the mechanisms of the Group 1 enzymes have been extensively studied and reviewed previously [1,2] and the reader is referred to those reviews for more specific details. The sections below highlight some of the recent studies on these enzymes.

8.2.1 Dihydrolipoamide dehydrogenase (LipDH) Dihydrolipoamide dehydrogenase is best known as the E3 component of the pyruvate dehydrogenase, α-ketoglutarate dehydrogenase, branched chain 2-oxoacid dehydrogenase and the glycine reductase multienzyme assemblies that are vital for proper metabolism of sugars and amino acids [2,14]. Its substrate, dihydrolipoamide (򐂰Fig. 8.3), is covalently attached via an amide linkage to a lysine on the E2 component, which allows it to be alternately positioned into each of the three enzyme active sites in the complex. The disulfide form is reductively acetylated in the E1 pyruvate dehydrogenase enzyme, deacetylated in the E2 dihydrolipoamide acetyltransferase, and then reoxidized by the E3 dihydrolipoamide dehydrogenase. Extensive mechanistic and structural studies of the enzyme have been previously reviewed and the reader is referred to those for details [2]. Structural studies of human LipDH (hLipDH) both alone [15] and in complex with the pyruvate dehydrogenase complex E3 binding domain [16] have recently provided insights into the source of disease causing mutations. Mutations were found to occur in three general areas: (1) Mutation of residues in the homodimer interface likely leads to weakening of the homodimer association and loss of activity in these obligate homodimers. (2) Deletion of G101 near the surface of the lipoyl binding tunnel likely causes disruption of the binding pocket, and the P453L mutation which lies just before the essential His452 base in the disulfide active likely disrupts positioning of the His base and may also weaken binding of the FAD as the backbone carbonyl of P453 hydrogen bonds the flavin N3-H. (3) Several mutations of residues that interact with either FAD, NADH or both may disrupt binding of the cofactors. The K37E mutation provides an interesting example. The hydrophobic portion of K37 stacks against the adenine of FAD and the ammonium group hydrogen bonds with the O2’ of the AMP moiety. Thus mutation to Glu would provide less hydrophobic interactions and likely disrupt the hydrogen bonding network to the FAD. Expression of the mutant enzyme is found to lead to low protein yields and lower FAD, however, the FAD-containing enzyme is fully active. These data are consistent with a model in which FAD binding is essential for proper folding and stability of the enzyme structure. Of note, the first structure of the complex of a LipDH with the nicotinamide ring stacked in the obligate position for electron transfer against the FAD isoalloxazine ring was included in this report [15,16]. As shown in 򐂰Fig. 8.7, a single copy of the E3BD portion of the pyruvate dehydrogenase E3-binding protein associates primarily with the interface domain of one subunit but near the center of the hLipDH homodimer and the openings to the lipoyl binding tunnels [16]. The orientation of the domain effectively occludes access to one of the lipoyl binding tunnels suggesting that the enzyme utilizes only one active site unless the domain undergoes conformational changes associated with the reactions, an idea that awaits further study. In a related study, the same workers also investigated differences in association of the bcE3BD from the branched chain dehydrogenase complex with E3 since isolated complexes of this multienzyme assembly are typically devoid of any E3 subunits. Comparison of the E3BD structures alone indicates the presence

8.2 Group 1 FDR enzymes

175

Dihydrolipoamide dehydrogenase

E3BD CXXXXC

FAD

NAD(P)H Central Interface

Fig. 8.7: Structure of the human dihydrolipoamide dehydrogenase (E3) in complex with the E3 binding domain (E3BD) of the human pyruvate dehydrogenase complex (pdb: 2F5Z) [16].

of Arg118 in the bcE3BD in place of an Asn in the pyruvate dehydrogenase E3BD that would conflict with residues identified at the E3BD/E3 interface. Mutation of Arg118 to Asn allowed crystallization of the bcE3BD/E3 complex confirming this hypothesis but also showed a smaller footprint at the interface that likely contributes to the weaker binding of E3 to this complex. At present the physiological significance of these differences in binding affinity for different mulitenzyme complexes is unknown [17]. A number of studies of LipDH enzymes from bacterial systems have identified new pathways for involvement of LipDH outside of the well-characterized metabolic pathways noted above. Of note, Bryk, et al., [18] have demonstrated a role for Mycobacterium tuberculosis LipDH (LPD) in an antioxidant defense pathway. In this pathway, NADH reduction of lipoyl-domains in the E2 component of the α-ketoglutarate dehydrogenase complex is coupled to reduction of peroxides by a typical AhpC peroxiredoxin through a novel all-α-helical trimeric adapter protein, AhpD, with a CXXC motif similar to thioredoxins. Disruption of any participant in this pathway negatively impacts electron flow making it a potential target for therapeutic intervention against Mtb. As a follow-up, the same group reported the crystal structure of Mtb LPD in which they identified key conserved residues in Mtb LPD, but not conserved in eukaryotic LipDH, that are involved in regulating the peroxidase and dehydrogenase activities of the enzyme [19]. In addition Arg93 and His386 are involved in forming both open and closed conformations of the active site suggesting a role in dynamically regulating LPD function. More recently this group reported that mice infected with Mtb carrying a deletion in the LPD gene showed no evidence of disease providing further evidence for inhibition of Mtb LPD as a therapeutic target, and consequently, they have developed the first selective inhibitors of Mtb LPD based on their previous structural analysis [20]. A few bacterial LipDH enzymes have previously been reported to have their own lipoyl group appended to the N-terminus of the polypeptide chain [1]. In another recent report, such an LPD protein from Xylella fastidiosa was reported to serve as a direct

176

8 FDR and structurally related flavoprotein thiol/disulfide-linked oxidoreductases

NADH-dependent reductant of the novel hydroperoxidase Ohr from the same organism [21]. Other LPD enzymes lacking the lipoyl domain also supported reduction of the peroxidase when coupled with another lipoylated protein indicating that dihydrolipoamide serves as the immediate reductant of Ohr. With these new reports, it appears that this classic member of the FDR enzyme family can also serve important roles as a lipoamide reductase. Further mechanistic studies of the control of the two activities are of significant interest for future investigations.

8.2.2 Glutathione reductase (GR) – two new structural studies on this classic member of the group As GR is one of the most extensively characterized FDR proteins, few new studies have emerged in the last decade, but two structural studies provide interesting new insights. In the first, Berkholz and coworkers [22] present a thorough analysis of near 1 Å resolution electron density maps for human GR (GR structure used in this review). The most interesting observation from the study is evidence of compression of the distances between NADPH, FAD and the redox active cysteines in the NADPH complex. Evidence includes: (a) Observation of decreased anisotropies for the atoms of FAD, NADPH, Cys63, and Tyr197 (that lies behind NADPH) in the reduced structure compared with those in the native oxidized enzyme indicating less mobility of these atoms in the EH2•NAPDH complex. (b) Observation of shorter than normal van der Waals distances in the reduced complex vs. native oxidized structures including NADPH(C4)FAD(N5) 2.50 vs. 2.75 Å, FAD(C4a)-Cys63(Sg) 3.30 vs. 3.50 Å. (c) Observation that in the GSH-reduced EH2 structure, Cys63(Sg) moves toward FAD(N5) by 0.15 Å while FAD(N5) moves away from Cys63(Sg) towards the empty NADPH pocket by ~0.3 Å. However, when NADPH is bound in the EH2•NADPH structure, FAD(N5) moves back toward a stationary Cys63(Sg) by 0.3 Å such that the distance between Cys63(Sg) and FAD(N5) is shorter by 0.15 Å than it was in native oxidized and 0.3 Å shorter than it was in EH2. This latter observation is proposed as a possible explanation for the increased extinction coefficient for the thiolate-to-flavin charge transfer band in EH2•NADPH vs. that in EH2 (compare charge transfer bands for EH2 and EH2•NADPH in 򐂰Figs. 8.4 and 8.6). (d) Observation of distorted planarity of the nicotinamide ring with NADPH(N1) pushed toward the flavin leading to a 1,4 syn nonbonded interaction that should favor transfer of the “hydride” to FAD. In addition, the C-O bond in the ribose ring is parallel to the nicotinamide ring, which should raise the reduction potential and help regulate the flow of electrons into FAD. (e) Finally, alignment of NADPH(C4)-FAD(N5) and FAD(C4a)-Cys63(Sg) is optimal for a 1,2 addition of the hydride at N5 concomitant with formation of the C4a-thiol adduct raising the possibility that reduced FADH− is not an intermediate. This is at least consistent with the lack of evidence for any development of FADH− in transient kinetic studies of GR. Although this is both a plausible and appealing mechanism for all of these enzymes, studies of some (e.g., MerA [7,8]) do show development of partial FADH−/NADP+ charge transfer character within the mixing time of NADPH with the oxidized enzyme and prior to formation of the typical reduced EH2•NADP+ species (򐂰Fig. 8.6). A second recent study of interest reports the structure of a novel GR-type enzyme, glutathione amide reductase from Chromatium gracile that is specific for glutathione amide (GASSAG) and also NADH instead of NADPH [23]. These altered specificities appear to

8.3 Group 2A FDR enzymes

177

be conserved in Chromatiaceae species where the use of GASSAG rather than GSSG may be important in respiratory sulfide oxidation [24–26]. The structure provides new insights into the switch in specificities with NADPH to NADH being of most interest from a biotechnology perspective since NADH is much cheaper than NADPH to use in vitro but it may also be of interest for molecular pathway engineering. The main difference leading to the switch to GASSAG over GSSG specificity is replacement of an Arg that binds the GSH II carboxylate by a Glu that clearly will repel it. The authors compare the differences that have evolved here for binding NADH vs. those that were engineered into GR [27] to switch from NADPH to NADH specificity.

8.2.3 Trypanothione reductase (TryR) Trypanothione reductase occurs uniquely in trypanosomal parasites and is specific for the novel trypanosomal cellular thiol trypanothione, which is a pair of glutathione molecules connected via amide bonds to a spermidine linker (򐂰Fig. 8.3). With its unique occurrence in human pathogens, this enzyme, discovered and mechanistically characterized in the 1980s [1], has been and continues to be the target of extensive drug development studies that are beyond the scope of this review. Although some compounds have shown promise as inhibitors, a recent review noted that metabolic control analysis of trypanothione metabolism in Trypanosoma cruzi, suggests TryR may not be the best target for reversible-binding inhibitors [28]. As opposed to reversible inhibitors, subversive substrates have been more successful [29,30]. These compounds serve both as inhibitors of the normal reductase activity and as one electron acceptor substrates of the enzyme that can then react with molecular oxygen to generate superoxide radicals and regenerate the TryR inhibitor/substrate for further redox cycling that increases oxidative stress in the parasite.

8.2.4 Mycothione reductase (MycR) Actinomycetes, which include mycobacteria and streptomycetes synthesize a completely different low molecular weight thiol for cellular redox homeostasis and protection against electrophilic stressors (򐂰Fig. 8.3) [1]. This compound, 1-D-myo-inositol-2-(Nacetyl-L-cysteinyl)amido-2-deoxy-α-D-glucopyranoside, is trivially denoted as mycothiol and mycothione in its reduced and oxidized states, respectively [31]. The mycothione reductase from M. tuberculosis was identified as having sequence homology to the Group 1 FDR enzymes. It has been expressed and its basic properties characterized [1] but no structures have yet been reported.

8.3 Group 2A FDR enzymes – enzymes of the Group 1 structural fold requiring an additional C-terminal Cys-based redox center As shown in 򐂰Fig. 8.2, Group 2A FDR enzymes exhibit the same structural fold as Group 1 enzymes but have short C-terminal extensions that harbor a second essential Cys-based redox center. Current members of this group include mercuric ion reductases (MerA) (򐂰Fig. 8.8) and the high Mr thioredoxin reductases (TrxR) and related proteins (TGR) (򐂰Fig. 8.9). Although both types of proteins catalyze the NADPH-dependent

178

8 FDR and structurally related flavoprotein thiol/disulfide-linked oxidoreductases

Mercuric ion reductase catalytic core CXXC

CXXXXC

NmerA

CCAG

FAD

NAD(P)H Central Interface

Mercuric ion reductase full length model NADPH FAD

FAD

S SH

Hg

S

Hg NmerA SH N SC Hg(SR)

S

2

NADPH

NADPH RSH

NmerA S N SC Hg-SR

RSH

FAD

S

S

S

S

S

HS

HS

NmerA S N SC Hg

Hg

HS

Hg(SR)2, NADPH

Hg0

NADP

NmerA

NADP

NADP

NADP

FAD

FAD

FAD

S

S

S

SH

SH

SH

SN Hg S HS SC

NmerA

SN HS Hg HS SC

NmerA S N SC Hg

HS HS

Fig. 8.8: Structure and mechanism of full-length mercuric ion reductase, a Group 2A enzyme. (Upper) Structure of Tn501 mercuric ion reducase catalytic core (pdb: 1ZK7) [37]. (Middle) Structural model of an NmerA-Cys11-Core-Cys465 disulfide crosslinked Tn21 mutant of full-length mercuric ion reductase obtained as best fit to small angle X-ray scattering data [35]. Bars indicate N-terminal fusion of NmerA domains with a CXXC motif and extended C-terminus with CCAG motif. (Lower) Reaction mechanism for full-length mercuric ion reductase.

8.3 Group 2A FDR enzymes

179

Thioredoxin reductase 1 CXXC

CXXXXC

Trx

FAD

CSecG: mammalian

NAD(P)H Central Interface

Thioredoxin reductase 1 complex with thioredoxin

Thioredoxin glutathione reductase CXXC

Grx

CXXXXC

FAD

CSecG

NAD(P)H Central Interface

Fig. 8.9: Structures of Group 2A selenocysteine-containing thioredoxin reductases. (Upper) Rat thioredoxin reductase 1 (pdb: 3EAN) [46]. (Middle) Sec498Cys mutant of human thioredoxin reductase 1 disulfide-crosslinked via Cys32 of Cys35Ser mutant human thioredoxin (pdb: 3QFB) [44]. (Lower) Sec597Cys mutant of Schistosoma mansoni thioredoxin glutathione reductase (TGR; pdb: 2X8C) [47] showing position of N-terminal fused glutaredoxin domain with a CXXC motif and extended C-terminus with reduced CCG redox center (normally CUG).

180

8 FDR and structurally related flavoprotein thiol/disulfide-linked oxidoreductases

reduction of their oxidizing substrates, they utilize their C-terminal Cys-based center for quite different purposes. In the TrxR and TGR enzymes, the C-terminal center serves as a redox center shuttling reducing equivalents from the CNXXXXCC redox center buried in the core of the protein out to reduce the disulfide in their protein substrates bound at the surface. In contrast, in MerA enzymes, the C-terminal cysteine pair does not undergo redox cycling. Rather it is required for retrieving the incoming Hg2+ substrate from protein- or other Hg-dithiol complexes and shuttling it into the buried CNXXXXCC center where it binds allowing electrons to flow from NADPH through the flavin to reduce it. Clearly both proteins require conformational changes in their extended C-terminal domain and control of the dynamics of these processes are of significant interest. However, the differences in their functional roles for redox cycling versus metal ion binding may be modulated by different environmental factors. The following sections highlight results since the 2004 review, particularly several new structures that have become available for both types of enzymes and studies based on insights from them.

8.3.1 Mercuric ion reductase (MerA) Mercuric ion reductases (MerA) catalyze the NADPH-dependent reduction of Hg2+ as the key step in bacterial mercury resistance pathways encoded by mer operons. The enzymes function within the cytoplasm where they both scavenge Hg2+ that has leaked into the cell prior to expression of the operon proteins, and acquire Hg2+ from upstream proteins in the pathway that include an integral membrane transport protein in all pathways and organomercurial lyase (MerB) enzymes that couple with MerA to confer additional resistance to organomercurials in broad spectrum operons. Extensive mechanistic studies of MerA from the Tn501 operon, summarized in previous reviews [1,2,32], have established the role of the C-terminal cysteines in shuttling Hg2+ from the incoming Hg(SR)2 substrates to the active site, as well as the requirement for bound NADPH (i.e., the EH2•NADPH complex) for reduction of Hg2+ bound to the active site Cys pair (򐂰Fig. 8.8). As noted in the sequence bars in 򐂰Fig. 8.8, the Tn501 MerA has an additional N-terminal domain (NmerA) with a GMTCXXC motif characteristic of heavy metal binding domains. Although this domain is not completely conserved in MerA proteins, phylogenetic analysis shows the occurrence of the domain is strongly correlated with the type of cellular thiols used by the bacterial species and with other environmental factors, e.g., saltwater environments, suggesting it does play an important role in the physiological functioning of these proteins [33]. Thus like the Tn501 enzyme, all MerA proteins from operons found in γ-proteobacteria, which utilize glutathione as their cellular thiol, have a single appended NmerA domain, and we have recently turned attention toward elucidation of structure and functional properties of the NmerA domains in Tn501 and related MerA proteins from other narrow and broad spectrum operons found in γ-proteobacteria [34–38]. Although a crystal structure of MerA from a Bacillus sp. was previously reported [39], nearly all of the mechanistic work and mutagenesis studies have been performed with the Tn501 enzyme. To facilitate further structure/function analysis, a 1.6 Å crystal structure of recombinant Tn501 catalytic core lacking NmerA and its linker was obtained [37] along with NMR structures of the reduced and Hgcomplex of the NmerA domain [36] (򐂰Fig. 8.8). Kinetic studies showed the Hg-NmerA complex to be an excellent substrate for the catalytic core and that the tethered domain in intact full length protein confers a significant kinetic advantage for acquisition of Hg2+

8.3 Group 2A FDR enzymes

181

bound to proteins [37]. Co-expression of NmerA with the catalytic core increased survival of glutathione-deficient cells exposed to Hg2+ over those expressing catalytic core alone providing further evidence that NmerA plays a critical role in scavenging Hg2+ from cellular proteins and most likely from upstream mer pathway proteins. Analysis of the pKa values for the cysteines of the GMTCXXC metal binding motif in wild type and several site directed mutant NmerAs indicates that conserved Tyr and His residues are important modulators of the metal binding affinity and may also be involved in recognition or transfer of Hg2+ to partner proteins [36,38]. Recently Hong, et al., [34] showed that the NmerA domain from the pDU1358 broad spectrum MerA is at least 100-fold more efficient at removal of the Hg2+ product from its cognate MerB enzyme than either the cognate catalytic core or the cellular thiol glutathione demonstrating a clear role for this domain in providing specific protein-protein transfer of the metal ion through the pathway. Further studies investigating residues important for modulating NmerA/ MerB recognition and Hg2+ transfer are currently underway. Finally, Johs et al. recently reported successful preparation of a completely pure full-length protein from a fusion construct and results from a small angle X-ray scattering study (SAXS) of the wild type and Hg2+ and disulfide crosslinked mutant forms of the protein [35]. 򐂰Fig. 8.8 shows a model of the docking site between NmerA and the catalytic core determined as the best fit to the SAXS data for the crosslinked enzyme. Although the resolution in these studies is not sufficient to determine the nature of the interactions, the structure has provided a starting point for further structure/function investigations of factors involved in molecular recognition and in facilitating metal ion transfers.

8.3.2 High Mr thioredoxin reductases (TrxR and TGR) Thioredoxin reductases catalyze the NADPH-dependent reduction of the conserved CGPC redox active disulfide in small (~12 kDa) thioredoxin proteins (Trx). Thioredoxins play central roles as reductants for key biosynthetic processes such as reduction of ribonucleotide reductases, as well as roles in redox regulation of a wide variety of cellular processes in all forms of life [2,40]. The CGPC motif in Trx is located at the end of a helix on the surface of the protein molecule, which provides specificity for interactions with its protein partners, but makes access of its disulfide to the redox center in the thioredoxin reductases more challenging. Biological systems have evolved two different structural strategies in the thioredoxin reductases to solve this problem. The high Mr thioredoxin reductases discussed in this section utilize the Group 1 FDR fold to transfer electrons from NADPH to their buried CNXXXXCC redox center and then utilize an additional C-terminal dicysteine or cysteine/selenocysteine pair to shuttle the electrons to the surface bound thioredoxin protein. In contrast, the low Mr thioredoxin reductases discussed below in Section 8.4 place their single dicysteine redox center in a different location and utilize a conformational switching mechanism to shuttle electrons from reduced flavin to the thioredoxin protein substrate. While the latter occur widely in bacteria, yeast, fungi and plants, high Mr forms have, thus far, been found only in higher eukaroytes. As mentioned above, high Mr TrxRs exhibit diversity in the sequence of the C-terminal redox center with those from Drosophila melanogaster and Plasmodium falciparum utilizing di-cysteine centers in different sequences (CCG in D. melanogaster and CGGGKCG in P. falciparum), and all mammalian enzymes utilizing a Cys-Sec pair in a CUG sequence (U = Sec = selenocysteine). As incorporation of selenocysteine into

182

8 FDR and structurally related flavoprotein thiol/disulfide-linked oxidoreductases

proteins requires specialized cellular machinery [41], structural and mechanistic studies of the D. melanogaster and P. falciparum proteins preceded those of the mammalian forms with key aspects summarized in the 2004 review [1]. The last decade, however, has seen an explosion of studies of the mammalian enzymes as well as a realization of an even greater diversity in the structural forms of this family of enzymes that is summarized in an excellent review by Arnér [42]. Key discoveries include the presence of three distinct TrxR genes (TrxR1, TrxR2, and TrxR3) as well as a number of splice variants of at least TrxR1 and TrxR2 that provide additional functionality and/or lead to alternative cellular localization [42]. For example, splice variant 1 of TrxR1 is primarily cytoplasmic, while splice variant 2 has an appended N-terminal nuclear receptor binding domain and splice variant 3 has an appended N-terminal glutaredoxin domain as does TrxR3, which is also referred to as thioredoxin glutathione reductase (TGR) due to its additional glutathione reductase activity conferred by the glutaredoxin domain [42]. Alongside the genomic and genetic expression studies, a number of new structures of the mammalian enzymes [43–46], as well as structures of a TGR from Schistosoma mansoni [47,48], have appeared providing insights into the nature of protein-protein interactions involved in docking between the different components as well as into the dynamic rearrangements of the C-terminal regions needed for shuttling of the electrons from the buried redox center to the protein substrate. In structures of the reduced form of a U457C mutant of human TrxR1 [45] Fritz-Wolf et al. identified at least three different orientations of the C-terminal arm in different monomers in the crystal from which they proposed a model involving the concerted action of the side chains of Asn418, Asn419 and Trp407 as a “guiding bar” for sliding of the C-terminal arm between inside and outside conformations. Cheng, et al., [46] subsequently reported the structure of the first Sec-containing TrxR1 from rat (򐂰Fig. 8.9, top), and proposed the overall mechanism for electron transfers shown in 򐂰Fig. 8.10 based on the combination of High Mr thioredoxin reductase NADPH

NADPH

NADPH

FAD

FAD

FAD

앥

S

앥

S

SH

S

S

Se

Se

–Se

S

S

HS

HS

NADPH NADP쎵, Trx

SHN SHC

NADP쎵

NADP쎵

FAD

FAD

S앥

S앥

SH

Trx

Se SN HS SHC

SH

SN Trx

–Se HS

NADP쎵 FAD SN

Trx SC

S앥 SH –Se HS

SC

Fig. 8.10: General mechanism of high Mr thioredoxin reductases.

8.4 Group 2B FDR enzymes

183

their structures and previous experimental studies. Of interest in this mechanism, the Sec residue is proposed as the site of attack by the interchange thiol CN in reduction of the C-terminal cysteinyl-selenide and then to also take the lead in the nucleophilic attack on Cys32 in the C32GPC35 disulfide of thioredoxin. More recently Fritz-Wolf, et al., reported the structure of the human U457C mutant of TrxR1 as a disulfide linked complex with its thioredoxin substate [44] (򐂰Fig. 8.9, middle) that confirms the latter linkage in this mechanism. Specifically, Fritz-Wolf examined all combinations of single Cys->Ser mutants of the C32GPC35 motif in Trx with the single U457S single and double U457C,C458S mutants of TrxR1 and found that the only combination leading to formation of a crosslinked complex involves the C35S mutant of Trx and the U457C,C458S mutant of TrxR1 [44]. Comparison of the two orientations of the crosslinked structure TrxR1-Trx in 򐂰Fig. 8.9 with those of the crosslinked NmerA-MerA core in 򐂰Fig. 8.8 shows that although both dock primarily to the side of the FAD binding domains of these Group 1 structures, Trx docks at a much lower position and the C-terminal residues of the monomer reach out and down much farther to attack the Trx disulfide than appears to be necessary for the MerA C-terminus to retrieve Hg2+. Of interest in this regard, Angelucci, et al., [47,48] reported the first structures of a Sec->Cys mutant of the S. mansoni TGR (򐂰Fig. 8.9, bottom) with its fused N-terminal glutaredoxin (Grx) domain. With a much shorter linker than occurs between domains in full length MerA, the Grx is limited to a docking site near the top of the FAD-binding domain somewhat similar to that found for the crosslinked NmerA domain. Restriction to this position would then allow docking of Trx at the same site lower on the FAD binding domain and require more substantial movement of the C-terminal arm to alternatively access and reduce disulfides in the two associated protein domains [47]. Together with the structure of the E3BD:E3 complex for LipDH in 򐂰Fig. 8.7, these new structures highlight the ability of these Group 1 proteins to utilize many different contact surfaces as binding sites for substrates and scaffolding, raising the possibility that many more interactions of this type may yet be discovered.

8.4 Group 2B FDR enzymes – low Mr thioredoxin reductase (TrxR) and structurally related enzymes As noted above, low Mr thioredoxin reductases (TrxRs) are structurally organized quite differently from the Group 1 enzymes (򐂰Fig. 8.1) and utilize a completely different strategy to reduce the thioredoxin substrate. Whereas Group 1 enzymes utilize their interface domains to form obligate homodimers with active sites requiring residues from both monomers, Group 2B enzymes exemplified by low Mr TrxRs lack the interface domain and, instead uses their central domains to form homodimers with independently functioning monomer subunits. Key to this independent functioning is the location of the dicysteine redox center (CXXC in this case) in the pyridine nucleotide-binding domain rather than in the FAD-binding domain. Since FAD is needed to mediate the electron transfers between NADPH and the CXXC redox center, placement of the NADPH binding site in the same domain as the CXXC redox center effectively forces the enzyme to undergo a conformational change to present these reactants to the FAD isoalloxazine ring for alternating reduction of FAD by NADPH (FR conformation) and reoxidation by the CXXC disulfide (FO conformation). Once the CXXC dithiol has formed, the conformational

184

8 FDR and structurally related flavoprotein thiol/disulfide-linked oxidoreductases

change needed to position a new NADPH near the isoalloxazine for reduction, concominantly leads to exposure of the reduced CXXC center near the exterior surface of the protein where it can interact directly with the CGPC redox center on the surface of the Trx protein substrate. The top two panels of 򐂰Fig. 8.11 show the enzyme in these

NADPH CXXC FAD

Low M r thioredoxin reductase FO

NADPH FAD

CXXC

Low M r thioredoxin reductase FR CXXC

Trx

CXXC

FAD

NAD(P)H Central

CXXCTrxR CXXCTrx

FR complex with thioredoxin

Fig. 8.11: Structures of low Mr thioredoxin reductase from Escherichia coli. (Upper) Structure of flavin-oxidizing (FO) form of enzyme (pdb: 1TDF) [49,50] with CXXC redox center positioned for reduction by adjacent reduced FAD isoalloxazine ring. (Middle) Structure of flavin-reducing (FR) form of enzyme captured by disulfide crosslink between Cys138 of Cys135Ser mutant of TrxR and Cys32 of Cys35Ser mutant of thioredoxin (Trx) (pdb: 1F6M, with Trx chains turned off) [4] showing domain rotation and dihydronicotinamide ring of NADPH positioned for reduction of adjacent FAD isoalloxazine ring. (Lower) Same structure of FR enzyme with Trx chains visible (pdb: 1F6M) [4] showing position of the TrxR-Cys138-Trx-Cys32 disulfide-crosslinked Trx domains.

8.4 Group 2B FDR enzymes

185

two extreme conformations with the disulfide adjacent to FAD in the FO conformation [49,50] in the upper panel and the nicotinamide ring stacked against the isoalloxazine ring in the FR conformation [4,51,52] in the middle panel. The structure in the middle panel is actually that of the FR enzyme that was captured in a crosslinked complex with its Trx substrates but with the Trx domains hidden. The bottom panel shows the structure of the full complex with the Trx domains (light and dark blue) and the crosslinked disulfides visible as spheres [4,51,52]. Comparison of the FO and FR structures with the Trx domains hidden emphasizes the structural changes that occur within the TrxR domains. Focusing on the green and purple FAD and central domains, with the exception of a lengthening of the central domain helices (dark green and purple) in the FR conformation, the structures and positions of these are essentially unchanged in the two conformations [4,51,52]. Thus the interactions between these domains in the homodimer form a stable, rigid scaffold that may be essential for providing a restoring force for the rotations of the pyridine nucleotide domains relative to the rigid FAD domains. The loss of activity in monomers with mutations disrupting dimerization provided evidence for this role of rigidity in the FAD domains for proper rotational motions of the NADPH and FAD domains relative to one another [4,52]. Although not obvious from these images, the orientations of the FAD and NADPH domains in the FR conformation is much closer to the orientation of those two domains in the Group 1 enzymes, while the FO conformation is quite different as expected with the disulfide near the FAD. The upper panel in 򐂰Fig. 8.12 shows a simplified mechanistic scheme for low Mr TrxR that emphasizes the consequences of this required conformational change. Since the two redox centers in these proteins (FAD and the CXXC motif) each interact with only one of the two substrates (NADPH with FAD and Trx with the CXXC center), both of these substrate-enzyme reactions must occur in the same conformation of the enzyme, i.e., essentially at the same time in a “ternary complex” type mechanism. Reduced Trx must dissociate before the return conformational change can occur to allow transfer of electrons from FADH– to the CXXC disulfide. As can be seen in the structures, NADPH and NADP+ can bind to their binding site on the pyridine nucleotide domain when the enzyme is in the FO conformation and more likely undergo exchange at that point rather than immediately after oxidation in the FR conformation. One final thing to note about the structural organization of these enzymes (򐂰Fig. 8.11) is that the orientation of the CXXC disulfide is nearly parallel to the plane of the isoalloxazine ring rather than being distinctively perpendicular to it as it is in the Group 1 enzymes. With this orientation and similar distances between both cysteine sulfurs and the FAD C4a position, it is hard to predict which cysteine would serve as the so-called charge transfer thiol that typically forms the C4a adduct during reduction in the Group 1 enzymes. In fact, the wild type E. coli TrxR shows no evidence of a charge transfer interaction in the EH2 state with oxidized FAD and the C135XXC138 cysteines reduced. However, a charge transfer interaction is induced in the C135S mutant in high salt but not in the C138S mutant [53], which along with other evidence indicated that the more C-terminal C138 adopts the charge transfer thiol role in this protein as well. Surprisingly, Cys138 also takes the lead as the interchange thiol making the initial attack on the Trx CGPC disulfide. Key evidence for this was the fact that the crosslinked structure shown in 򐂰Fig. 8.11 could only be captured using the combination of the C135S mutant of TrxR and the C35S mutant of Trx [4]. Thus Cys138 in the low Mr TrxR behaves in

186

8 FDR and structurally related flavoprotein thiol/disulfide-linked oxidoreductases

Low Mr thioredoxins

Domain rotations

FAD

FAD SH SH

S Trx

SH SH

NAD(P)H

S 쎵 H쎵

쎵 H쎵 FADH앥 S

S

FADH앥 SH Trx

S

S

NAD(P)쎵

SH Domain rotations

FO

AhpF – alkylhydroperoxidase reductase

FR

Domain rotations

S S SH

FAD

S S

FAD

SH SH

SH SH

NAD(P)H

AhpC SH 쎵 H쎵 HS HS S AhpC

앥 H쎵 FADH앥

SH SH

S S

S

FADH앥 NAD(P)쎵

S

S

FO

Domain rotations

FR

Fig. 8.12: Redox cycles for low Mr TrxR. In these enzymes, the dithiol/disulfide redox center is located within the pyridine nucleotide binding domain and interacts with the isoalloxazine on the same side (re) as does the pyridine nucleotide. As a result, the flavin and pyridine nucleotide binding domains must undergo a rotation relative to one another for the full redox cycle to occur. In the FO orientation, the protein disulfide is positioned near the isoalloxazine allowing for oxidation of FADH−. Domain rotation to the FR orientation is then required to allow access of NADH to FAD and to allow access of the disulfide substrate to the protein dithiol redox center.

much the same way as the Sec group of the C-terminus of high Mr TrxR that appears to be the site of attack by the inner CXXXXC redox center and the takes the lead in attacking Cys32 of Trx as demonstrated by the crosslinked complex there, curious similarities in the two despite striking structural differences.

8.4 Group 2B FDR enzymes

187

Remaining questions that have become the focus for these systems include examination of factors controlling the dynamics of the conformational changes and examination of factors controlling specificity between TrxR and Trx proteins. A recent computational study examined the energetics of the conformational changes and the roles of previously identified residues involved in binding and/or catalysis [54]. Another study using protein film voltammetry examined temperature effects on the electrochemical reduction of a thermophilic TrxR with results suggesting that the flavin and disulfide potentials are within 40 mV of each other but are modulated by the conformational dynamics, which is perhaps not surprising as the environments of the redox centers might be expected to change significantly in the two conformations as at least some interactions with surrounding residues must change [55]. One additional new study has recently appeared on the closely related S. typhimurium peroxiredoxin reductase AhpF [56], which possess an N-terminal fusion of an unusual dual or double, but intertwined Trx fold with a single C129XXC132 redox center (򐂰Fig. 8.13). Electron flow in this enzyme is essentially the same as in E. coli TrxR with input of electrons to the flavin from NADH, followed by reduction of the C345XXC348

Low M r thioredoxin reductase FO AhpF FO monomer

Low M r TrxR FR-thioredoxin complex CXXC

CXXC

AhpC dimer “dual” Trx

FAD

C

NAD(P)H Central

C AhpC dimer head-to-tail

Fig. 8.13: Comparison of structure of peroxiredoxin reductase (AhpF) of the alkyl hydroperoxide reductase two-component system with E. coli thioredoxin reductase. (Upper) FO and FR homodimeric structures of E. coli thioredoxin reductase as in 򐂰Fig. 8.12. (Lower left) Structures of the Salmonella typhimurium peroxiredoxin dimer substrate (AhpC; pdb: 1YEP) [93] and a monomer of Salmonella typhimurium peroxiredoxin reductase (AhpF; pdb: 1HYU) [94] of the alkyl hydroperoxide reductase two-component system. (Lower right) Domain structure of AhpF and cartoon of head to tail disulfide bonding in AhpC dimer substrate.

188

8 FDR and structurally related flavoprotein thiol/disulfide-linked oxidoreductases

disulfide in the pyridine nucleotide domain followed by reduction of the C129XXC132 disulfide of the N-terminal Trx fusion, which then reduces the crosslinking disulfide in the protein substrate peroxiredoxin (structure in 򐂰Fig. 8.13 and mechanism in 򐂰Fig. 8.12). In this recent report, Jönsson and coworkers examined the role of each of the cysteines in the three sets of disulfides in this system and demonstrated that like E. coli TrxR, Cys348 acts both as the charge-transfer thiol and as the “interchange thiol” as it takes the lead in reducing the Trx domain disulfide. Furthermore, like Trx, Cys129 of the Trx domain CXXC motif takes the lead in attack on the peroxiredoxin disulfide [56]. Thus, properties of the structural fold likely dictate the inherent reactivity of the residues. Brief mention should be made regarding the novel low Mr thiol compound bacillithione/thiol that was recently identified in various Bacillus sp. (򐂰Fig. 8.3). Genomic analysis of genes from species that synthesize the compound has identified a putative bacillithiol reductase of the low Mr TrxR type structure, but as yet, no report of cloning or expression of the gene has appeared. This will be an interesting target for future study as bacillithione is a small molecule disulfide substrate rather than a protein substrate that has been the standard for this type of protein fold [57]. Of interest in this context, however, three new proteins have indeed been isolated that have small molecule substrates and exhibit sequence homology with low Mr TrxR [58–63]. Curiously, all of these proteins catalyze the intramolecular oxidation of dithiols to disulfides (򐂰Fig. 8.14) rather than the reduction of a disulfide. One of the proteins (DepH) utilizes NADP+ as the oxidant and thus has the normal GXGXXG type pyridine nucleotide binding motif [63]. However, the other two proteins, GliT and HlmI, both lack the motif and do not bind pyridine nucleotides at all [59–62]. Instead they have been shown to utilize molecular oxygen as the oxidizing substrate in vitro. Thus at present, these proteins appear to a novel structural form of sulfhydryl oxidases with small molecule dithiol substrates instead of the protein dithiol substrates of those involved in protein folding. A phylogenetic analysis suggests there are several classes of these proteins that will be exciting targets for future investigations [62].

8.5 Group 3 FDR enzymes – enzymes with cysteine sulfenic acid or mixed Cys-S-S-CoA redox center With the Group 3 enzymes we return to a structural fold that is more closely related to that of the Group 1 enzymes [64]. As shown in 򐂰Fig. 8.1, the Group 3 enzymes are also obligate homodimers that use a C-terminal interface domain for dimerization as well as in forming part of the binding site surrounding the Cys-based redox center. Comparison of the color-coded structures in 򐂰Fig. 8.1, however, shows that their FAD binding domains are drastically truncated such that the lower half of the homodimers has a large cleft whereas the Group 1 enzymes have extended secondary structure elements and loops that interact across the dimer interface. This may reflect the differing nature of the types of substrates utilized by these enzymes. 򐂰Fig. 8.15 highlights the other key difference structural and mechanistic difference between the Group 3 and Group 1 enzymes. Instead of the Group 1 CXXXXC motif, Group 3 enzymes have only a single cysteine in an SFXXC motif, which is located near the C4a of the isoalloxazine ring as the equivalent of the charge transfer cysteine CC in Group 1 enzymes [64]. In two of these enzymes, the Eox equivalent of the Group 1

8.5 Group 3 FDR enzymes

FAD SH

SH

S

S

FAD 쏁

S

H2O2 (GliT & Hlml)

S

S

S

FAD

FADH2

SH

S

SH

S

NADPH (DepH)

O2 (GliT & Hlml) NADP쎵 (DepH)

O

HS

O CH3

N

SN

H

O

OH

Gliotoxinred

Gliotoxinox

O

H N

O

H N

NH

NH

Hlml

O

O

HS

S S

HS Holomycinred

Holomycinox

O

O

NH

O O

N H

O

O

NH

O

O

N H O O

NH

NH NH

O

OH

H

SH

O

CH3

NS

GliT

OH

N OH

189

SH

DepH

NH

O S

HS FK228red

S FK228ox

Fig. 8.14: Non-ribosomal peptide dithiol/disulfide structures and the general pathway for their oxidation by novel FDR enzymes with sequence homology to the Group 2B low Mr thioredoxintype fold.

enzymes (i.e., where the CXXXXC redox center is a disulfide) is a cysteine sulfenic acid, Cys-S-OH. These enzymes include NADH peroxidases and NADH oxidases. The structures and mechanisms of the NADH peroxidase have been previously reviewed as have mechanistic studies of the NADH oxidases and basic mechanisms for both are summarized in 򐂰Fig. 8.16 [1,64]. The structure of an NADH oxidase from Lactobacillus

190

8 FDR and structurally related flavoprotein thiol/disulfide-linked oxidoreductases

NADH peroxidase

NADH oxidase

CoA disulfide reductase CoADR-RHD

SFXXC

FAD

C

NAD(P)H

Central Interface

RHD

Fig. 8.15: Structures of Group 3 enzymes with single Cys-X redox centers. (Upper) Enterococcus faeculis 10C1 NADH peroxidase (pdb: 1JOA) [87] with close-up of active site sulfenic acid (CysS-OH) redox center. (Upper middle) Lactobacillus sanfranciscensis NADH oxidase (pdb: 1CDU) [65] with close-up of two conformations of active site Cys-S-OH redox center. (Lower middle) Staphylococcus aureus CoA-disulfide reductase (pdb: 1YQZ) [67] with close-up of active site Cys-SCoA redox center. (Lower) Bacillus anthracis CoA-disulfide reductase-RHD (pdb: 3ICS) [68] has a novel C-terminal fusion of a rhodanese domain (RHD) shown in deep teal and cyan.

8.5 Group 3 FDR enzymes NADH peroxidase NADH FAD

H2O

앥

S

191

CoA disulfide reductase NADH

NADPH

FAD

FAD

NADPH

CoASH

FAD

앥

S OH

S

S

S CoA

HO OH H2O

H2O2

S CoA

S-CoA

S CoA CoASH

S CoA

NADH

NAD쎵

NADPH

NADP쎵

FAD

FAD

FAD

FAD

S앥

S앥

S앥

S앥

NADH NAD쎵

NAD(P)H NAD(P)쎵

NADH oxidase

CoADR-RHD

NAD쎵

NAD쎵

FADH S앥 O

FADH OH S

OH

OH

O2 NAD쎵

NADH

FADH2

FAD

S앥

S NADH H2O NAD쎵

FAD S

S

S52앥 SH앥

FAD S

CoA

S

NADH NAD쎵

FAD S앥 CoA

CoA

S앥

NAD쎵 H2O NADH

S3 SH SH

S3 SH

S

S

3NAD쎵 3SH앥 3NADH

OH

SH앥 FAD S

S

FAD CoA

S앥 S

S3H S

SH CoA

S3H S

Fig. 8.16: Mechanisms for Group 3 enzymes.

sanfranciscensis (򐂰Fig. 8.15) has been reported more recently [65]. Overall it is quite similar to the peroxidase, however the active site sulfenic acid appears to adopt two different conformations with the Cys-S-OH rotated roughly 90° and 180° relative to its orientation in the peroxidase. The authors in this manuscript [65] propose a slightly different mechanism from that shown in 򐂰Fig. 8.16. Instead of molecular oxygen reacting to form a C4a-peroxy adduct as shown in the upper right of the NADH oxidase mechanism in 򐂰Fig. 8.16, these authors propose that O2 reacts on the re face of the isoalloxazine near the bound pyridine nucleotide, and that the H2O2 then migrates to the si side where it is attacked by the Cys-thiolate to form the Cys-S-OH and release one water molecule. Studies distinguishing these mechanisms are of interest for future investigations. The second type of redox center found in the Eox forms of Group 3 fold enzymes is a mixed disulfide of the SFXXC cysteine with a molecule of Coenzyme A, Cys-S-S-CoA

192

8 FDR and structurally related flavoprotein thiol/disulfide-linked oxidoreductases

(򐂰Fig. 8.15). Previously summarized studies of these enzymes [1] indicate that CoAdisulfide is indeed a good substrate and may be the predominant cellular thiol used by organisms [66] with these proteins since they lack glutathione or any of the other thiols shown in 򐂰Fig. 8.3. The structure of the Staphyloccus aureus CoADR [67] in 򐂰Fig. 8.15 shows the overall fold is once again quite similar to the oxidase and peroxidase structures and the extended CoA chain folds into a binding pocket nestled against helices from the FAD and central domains. Mechanistic studies suggest the mechanism of this enzyme parallels that of the peroxidase (򐂰Fig. 8.16). The fourth structure shown in 򐂰Fig. 8.15 is of a more recently identified Bacillus anthracis CoADR that has a C-terminally fused rhodanese domain (RHD) [68]. So here we see yet another type of domain fusion in this extended family of FDR enzymes connecting different types of redox functionality. Rhodanese domains possess a single Cys residue and have been found to form Cys-persulfides (Cys-S-SH) that often serve as the source of sulfide that becomes incorporated into a variety of cofactors [69]. The RHD domain in the CoADR-RHD fusion protein also possesses a single Cysteine residue (C514’) that is ~25 Å away from the active site C44. Interestingly, the extended chain of the mixed disulfide CoA molecule in the Eox form of the protein extends into the C-terminal RHD domain of the other monomer at the active site interface such that it is well anchored there and non-dissociating, consistent with the enzyme’s inability to reduce CoADisulfide. Upon reduction of the enzyme by NADH, the pantetheine arm of the CoA molecule becomes mobile suggesting it can swing from the FAD domain site to the RHD domain to shuttle electrons to reduce a disulfide bound to the RHD C514’ [68]. 򐂰Fig. 8.16 shows a proposed reaction pathway for reduction of a polysulfur chain as the substrate that loads onto the RHD cysteine. This finding suggests that FDR type enzymes are then intimately involved in both the reduction of polysulfur chains and as well as their production that will be discussed below in Section 8.7.

8.6 Group 4 FDR enzymes – Group 1-fold enzymes catalyzing novel reactions This section describes the very unusual Group 1-fold enzyme 2-ketopropyl-CoM carboxylase/oxidoreductase (KPCR) from Xanthobacter Py2. The protein was identified as part of an alkene-inducible pathway that functions in aliphatic alkene metabolism and allow the organisms to grow on alkenes as their sole carbon source [70]. The pathway, which utilizes Coenzyme M (β-mercaptosulfonate) that had previously been thought to be restricted to methanogens, is found on a megaplasmid together with the CoM biosynthetic genes [71]. 򐂰Fig. 8.17 shows a comparison of the KPCR structure [72–75] with that of human glutathione reductase showing that it is clearly a Group 1-fold member. Furthermore it possesses the standard CXXXXC redox center as in all other Group 1-fold enzymes and exhibits spectra with charge transfer bands typical of EH2 and EH2•NADP+ [76]. The reaction catalyzed by the enzyme is summarized in the mechanistic scheme in 򐂰Fig. 8.17. The reaction has several novel aspects. First, the substrate attacked by the interchange thiol is a thioether rather than a disulfide. Second, this attack leads to formation of a mixed disulfide of the interchange thiol with the CoM sulfur and releases the unstable enolate of acetone that, third, reacts with a molecule of CO2 to form acetoacetate.

8.6 Group 4 FDR enzymes

193

Glutathione reductase

Ketopropyl-CoM carboxylase/oxidoreductase NADP쎵

NADP쎵 쎵

NADP

FAD

NADPH

FAD

FAD

S

S

S앥

S

S

SH

CoM SH

O

CoM S NADP쎵 FAD S앥 S

CoM S

O 앥

O2C

NADP쎵

NADP쎵

FAD

FAD

S앥 S

앥

O

CoM S

CO2

S앥 SH

O

CoM S

O C O

Fig. 8.17: Structure and mechanism of 2-ketopropyl-CoM carboxylase/oxidoreducase (KPCR). (Upper panels) Comparison of structures of Group 1 human glutathione reductase (pdb: 3DJJ) [22] and Group 4 Xanthobactor Py2 KPCR (pdb: 3Q6J) [74] showing increased packing in dithiol active site cleft. (Lower) Mechanism of KPCR-catalyzed reaction.

In the absence of CO2, the enolate becomes protonated to give acetone, which is not of much use, so the enzyme active site must exclude protons and promote the localization of CO2 in the active site to capture the incipient enolate as it forms [77]. From the structure comparison, 򐂰Fig. 8.17, it is clear that the active site cleft between the FAD/ central domains of one monomer and the interface domain of the other is much more congested in this protein than in the disulfide reductases that need to bind large disulfide substrates in the pocket near the CXXXXC redox center. Several structural studies

194

8 FDR and structurally related flavoprotein thiol/disulfide-linked oxidoreductases

reported recently identify a potential CO2 binding tunnel and the presence of hydrophobic residues in the active site that likely prevent protons from becoming available to the enolate intermediate [74,75,77–79]. Related to the use of a thiol in this enzyme to attack a thioether and release an enolate, Warlick, et al. [80] recently reported a similar reaction by a methylsulfurase involving the displacement of the enolate of 1-deoxy-Dxylulose-5-phosphate by attack of an active site cysteine on the methylthioether of the sugar. In that case, the product is simply protonated, and some other thiol must reduce the mixed disulfide as it is not a flavoprotein. This similar finding raises the possibility that perhaps more enzymes including more FDR enzymes are likely to be found that catalyze these types of ether cleavages to form enolates.

8.7 Group 5 FDR enzymes – enzymes with a si side pair of Cys residues widely separated in sequence The sulfide:quinone oxidoreductases (SQRs) are the newest and perhaps most novel members to join the FDR family [81]. H2S is a well known toxin that has gained attention more recently as a potential third gaseous signaling molecule in humans and other higher eukaryotes [82]. SQRs, which are ubiquitous in all life forms [81], decrease the toxicity of H2S by catalyzing the quinone-dependent oxidative polymerization to water soluble HS(S)nSH chains or possibly cyclic sulfurane species that have yet to be fully characterized. Having both water soluble sulfide and lipophilic quinone substrates, the enzymes are membrane associated and hence have amphiphilic surfaces. Structures of three SQRs, two from Gram negative extremophiles, hyperthermophilic Aquifex aeolicus [83] and acidophilic Acidithiobacillus ferrooxidans [84,85], and one from the thermoacidophilic archaeon Acidianus ambivalens [86], have recently been reported which exhibit overall similar monomer folds but differ in their apparent oligomerization behavior in the crystals. 򐂰Fig. 8.18 shows a single trimer of the A. aeolicus enzyme which forms dimers of trimers in the crystal associated through their large hydrophobic faces [83]. The other two enzymes dimerize in aqueous solution and in the crystals through their amphiphilic helices [85,86], which Cherney, et al., [85] noted is unlikely to be physiologically relevant. The right middle panel in 򐂰Fig. 8.18 highlights the similarities and differences of the SQR monomer to that of the classic Group 1 enzymes. Although the FAD domain is truncated relative to the Group 1 enzymes, the fold through the FAD, “NADPH” and central domains is quite similar in the two monomers. The interface domains however are strikingly different. In the GR monomer, the interface domain extends away from its own FAD-binding domain and interacts with the FAD-binding domain on the second monomer of the homodimer. In contrast, in the SQR monomer, the interface domain folds back over its own FAD-binding domain forming the membrane-associating amphiphilic helices and the entry channel for the quinone substrate to access the flavin near the si side of the pyrimidine ring (򐂰Fig. 8.18, bottom panel). The cysteine redox center also differs dramatically being composed of two cysteines widely separated in sequence, one lying at the N-terminus of an α-helix in the “NADPH” domain (C156 in A. aeolicus [83]) and one in a flexible loop in the interface domain (C347 in A. aeolicus [83]). The greater separation and flexibility in positioning of one Cys relative to the other

8.7 Group 5 FDR enzymes C

FAD

NAD(P)H

195

C

Central Int GR monomer

SQR 90°

Sulfur cluster FAD Decylubiquinone nH2S

(n 1)Q

HS (S)n

2

SH

(n 1)UQH2

Fig. 8.18: Structure of sulfide:quinone oxidoreductase (SQR) and comparison with Group 1 GR monomer. (Upper left) Structure of a crystallographic trimer of Aquifex aeloicus SQR (pdb: 3HYW) [81,83] looking at the hydrophilic face. (Middle left) Edge-view after 90° rotation of trimer. Decylubiquinone chains (sticks with blue Cs) can be seen “hanging” from the bottom face that would be oriented toward the membrane. (Lower left) SQR reaction. (Upper right) Domain structure of monomers, both color schemes are shown. (Middle right) Comparison of monomer structure of SQR and GR showing that the interface domain (magenta) of SQR folds back upon itself to cover the si side of the FAD isoalloxazine rather than extending out to form the interface with a second monomer. (Lower right) Close-up of active site showing sulfur cluster (spheres) that forms between cysteines of the redox center on the re side of the flavin.

allows room for the growing polysulfur chain in successive cycles of the catalytic reaction. Several mechanistic proposals for the roles of these two cysteines in holding the growing polysulfur chain and donating electrons to FAD have arisen from the crystallographic studies of wild type and mutant enzymes [83–86] that will require some clever mechanistic studies to sort out due to the multiple cycles of sulfide oxidation that occur before release of each polysulfur product. Although the SQR structures from different organisms are similar in overall fold, they do differ in two respects. One is the presence/absence of a nonconserved cysteine that is found in slightly different locations near [85] or covalently bonded directly [86] or via a disulfide bridge [83] to the 8aMe of the flavin isoalloxazine ring. Mutations of this cysteine lower but do not abolish the flavin reductase activity of any of the enzymes, thus its role remains unclear. The second difference in the proteins is the structure and placement of flexible “capping loops” that cover and form the sulfide entry channel on

196

8 FDR and structurally related flavoprotein thiol/disulfide-linked oxidoreductases

Fig. 8.19: Comparison of two SQR structures and their divergent capping loops. (Upper panels) Structure and close-up of Aquifex aeolicus SQR monomer (pdb: 3HYW) [81,83]. (Lower panels) Structure and close-up of Acidianus ambivalens SQR monomer (pdb: 3H8I) [86]. The interface domain in the latter structure was largely disordered. Monomers colored in green motif as in 򐂰Fig. 8.18 with the divergent capping loops shown in red in both.

the hydrophilic surface of the monomers, illustrated in 򐂰Fig. 8.19 for A. aeolicus and A. ambivalens monomers [81]. A recent bioinformatics analysis subdivides the six different subtypes [81] suggesting more differences may emerge as additional members of this novel FDR enzyme are characterized.

8.8 References [1] Argyrou A, Blanchard JS. Flavoprotein disulfide reductases: advances in chemistry and function. Prog Nucleic Acid Res Mol Biol 2004;78:89–142. [2] Williams CH, Jr. Lipoamide dehydrogenase, glutathione reductase, thioredoxin reductase, and mercuric ion reductase—a family of flavoenzyme transhydrogenases. In: Muller F, ed. Chemistry and Biochemistry of Flavoenzymes. Boca Raton: CRC Press; 1992:121–211. [3] Williams CH, Jr. Flavin-containing dehydrogenases. In: Boyer PD, ed. The Enzymes. New York: Academic Press; 1976:89–173. [4] Lennon BW, Williams CH, Jr., Ludwig ML. Twists in catalysis: alternating conformations of Escherichia coli thioredoxin reductase. Science 2000;289:1190–4. [5] Massey V, Veeger C. Biological oxidations. Ann Rev Biochem 1963;32:579–638. [6] Schulz GE, Schirmer RH, Sachsenheimer W, Pai EF. The structure of the flavoenzyme glutathione reductase. Nature 1978;273:120–4.

8.8 References

197

[7] Engst S, Miller SM. Rapid reduction of Hg(II) by mercuric ion reductase does not require the conserved C-terminal cysteine pair using HgBr2 as the substrate. Biochemistry 1998;37:11496–507. [8] Engst S, Miller SM. Rapid reduction of Hg(II) by mercuric ion reductase does not require the conserved C-terminal cysteine pair using HgBr2 as the substrate. Biochemistry 1999;38:853–4. [9] Sahlman L, Lambeir AM, Lindskog S. Rapid-scan stopped-flow studies of the pH dependence of the reaction between mercuric reductase and NADPH. Eur J of Biochem/FEBS 1986;156:479–88. [10] Thorpe C, Williams CH, Jr. Lipoamide dehydrogenase from pig heart. Pyridine nucleotide induced changes in monoalkylated two-electron reduced enzyme. Biochemistry 1981;20: 1507–13. [11] Miller SM, Massey V, Ballou D, Williams CH, Jr, Distefano MD, Moore MJ, Walsh CT. Use of a site-directed triple mutant to trap intermediates: demonstration that the flavin C(4a)-thiol adduct and reduced flavin are kinetically competent intermediates in mercuric ion reductase. Biochemistry 1990;29:2831–41. [12] Engst S, Miller SM. Alternative routes for entry of HgX2 into the active site of mercuric ion reductase depend on the nature of the X ligands. Biochemistry 1999;38:3519–29. [13] Thorpe C, Williams CH, Jr. Differential reactivity of the two active site cysteine residues generated on reduction of pig heart lipoamide dehydrogenase. J Biol Chem 1976;251:3553–7. [14] Reed LJ. Multienzyme complexes. Acc Chem Res 1974;7:40–6. [15] Brautigam CA, Chuang JL, Tomchick DR, Machius M, Chuang DT. Crystal structure of human dihydrolipoamide dehydrogenase: NAD+/NADH binding and the structural basis of diseasecausing mutations. J Mol Biol 2005;350:543–52. [16] Brautigam CA, Wynn RM, Chuang JL, Machius M, Tomchick DR, Chuang DT. Structural insight into interactions between dihydrolipoamide dehydrogenase (E3) and E3 binding protein of human pyruvate dehydrogenase complex. Structure 2006;14:611–21. [17] Brautigam CA, Wynn RM, Chuang JL, Naik MT, Young BB, Huang TH, Chuang DT. Structural and thermodynamic basis for weak interactions between dihydrolipoamide dehydrogenase and subunit-binding domain of the branched-chain alpha-ketoacid dehydrogenase complex. J Biol Chem 2011;286:23476–88. [18] Bryk R, Lima CD, Erdjument-Bromage H, Tempst P, Nathan C. Metabolic enzymes of mycobacteria linked to antioxidant defense by a thioredoxin-like protein. Science 2002;295:1073–7. [19] Rajashankar KR, Bryk R, Kniewel R, Buglino JA, Nathan CF, Lima CD. Crystal structure and functional analysis of lipoamide dehydrogenase from Mycobacterium tuberculosis. J Biol Chem 2005;280:33977–83. [20] Bryk R, Arango N, Venugopal A, Warren JD, Park YH, Patel MS, Lima CD, Nathan C. Triazaspirodimethoxybenzoyls as selective inhibitors of mycobacterial lipoamide dehydrogenase. Biochemistry 2010;49:1616–27. [21] Cussiol JR, Alegria TG, Szweda LI, Netto LE. Ohr (organic hydroperoxide resistance protein) possesses a previously undescribed activity, lipoyl-dependent peroxidase. J Biol Chem 2010;285:21943–50. [22] Berkholz DS, Faber HR, Savvides SN, Karplus PA. Catalytic cycle of human glutathione reductase near 1 Å resolution. J Mol Biol 2008;382:371–84. [23] Van Petegem F, De Vos D, Savvides S, Vergauwen B, Van Beeumen J. Understanding nicotinamide dinucleotide cofactor and substrate specificity in class I flavoprotein disulfide oxidoreductases: crystallographic analysis of a glutathione amide reductase. J Mol Biol 2007;374:883–9. [24] Bartsch RG, Newton GL, Sherrill C, Fahey RC. Glutathione amide and its perthiol in anaerobic sulfur bacteria. J Bacteriol 1996;178:4742–6. [25] Pott AS, Dahl C. Sirohaem sulfite reductase and other proteins encoded by genes at the dsr locus of Chromatium vinosum are involved in the oxidation of intracellular sulfur. Microbiology 1998;144(Pt 7):1881–94.

198

8 FDR and structurally related flavoprotein thiol/disulfide-linked oxidoreductases

[26] Vergauwen B, Pauwels F, Jacquemotte F, Meyer TE, Cusanovich MA, Bartsch RG, Van Beeumen JJ. Characterization of glutathione amide reductase from Chromatium gracile. Identification of a novel thiol peroxidase (Prx/Grx) fueled by glutathione amide redox cycling. J Biol Chem 2001;276:20890–7. [27] Scrutton NS, Berry A, Perham RN. Redesign of the coenzyme specificity of a dehydrogenase by protein engineering. Nature 1990;343:38–43. [28] Olin-Sandoval V, Moreno-Sanchez R, Saavedra E. Targeting trypanothione metabolism in trypanosomatid human parasites. Current drug targets 2010;11:1614–30. [29] Salmon-Chemin L, Buisine E, Yardley V, Kohler S, Debreu MA, Landry V, Sergheraert C, Croft SL, Krauth-Siegel RL, Davioud-Charvet E. 2- and 3-substituted 1,4-naphthoquinone derivatives as subversive substrates of trypanothione reductase and lipoamide dehydrogenase from Trypanosoma cruzi: synthesis and correlation between redox cycling activities and in vitro cytotoxicity. J Med Chem 2001;44:548–65. [30] Buchholz K, Schirmer RH, Eubel JK, Akoachere MB, Dandekar T, Becker K, Gromer S. Interactions of methylene blue with human disulfide reductases and their orthologues from Plasmodium falciparum. Antimicrob Agents Chemother 2008;52:183–91. [31] Spies HS, Steenkamp DJ. Thiols of intracellular pathogens. Identification of ovothiol A in Leishmania donovani and structural analysis of a novel thiol from Mycobacterium bovis. Eur J Biochem/FEBS 1994;224:203–13. [32] Barkay T, Miller SM, Summers AO. Bacterial mercury resistance from atoms to ecosystems. FEMS Microbiol Rev 2003;27:355–84. [33] Barkay T, Kritee K, Boyd E, Geesey G. A thermophilic bacterial origin and subsequent constraints by redox, light and salinity on the evolution of the microbial mercuric reductase. Environ Microbiol 2010;12:2904–17. [34] Hong B, Nauss R, Harwood IM, Miller SM. Direct measurement of mercury(II) removal from organomercurial lyase (MerB) by tryptophan fluorescence: NmerA domain of coevolved gamma-proteobacterial mercuric ion reductase (MerA) is more efficient than MerA catalytic core or glutathione. Biochemistry 2010;49:8187–96. [35] Johs A, Harwood IM, Parks JM, Nauss RE, Smith JC, Liang L, Miller SM. Structural characterization of intramolecular Hg(2+) transfer between flexibly linked domains of mercuric ion reductase. J Mol Biol 2011;413:639–56. [36] Ledwidge R, Hong B, Dotsch V, Miller SM. NmerA of Tn501 mercuric ion reductase: structural modulation of the pKa values of the metal binding cysteine thiols. Biochemistry 2010;49:8988–98. [37] Ledwidge R, Patel B, Dong A, Fiedler D, Falkowski M, Zelikova J, Summers AO, Pai EF, Miller SM. NmerA, the metal binding domain of mercuric ion reductase, removes Hg2+ from proteins, delivers it to the catalytic core, and protects cells under glutathione-depleted conditions. Biochemistry 2005;44:11402–16. [38] Ledwidge R, Soinski R, Miller SM. Direct monitoring of metal ion transfer between two trafficking proteins. J Am Chem Soc 2005;127:10842–3. [39] Schiering N, Kabsch W, Moore MJ, Distefano MD, Walsh CT, Pai EF. Structure of the detoxification catalyst mercuric ion reductase from Bacillus sp. strain RC607. Nature 1991;352:168–72. [40] Holmgren A, Lu J. Thioredoxin and thioredoxin reductase: current research with special reference to human disease. Biochem Biophys Res Commun 2010;396:120–4. [41] Hatfield DL, Gladyshev VN. How selenium has altered our understanding of the genetic code. Mol Cell Biol 2002;22:3565–76. [42] Arner ES. Focus on mammalian thioredoxin reductases—important selenoproteins with versatile functions. Biochim Biophys Acta 2009;1790:495–526. [43] Biterova EI, Turanov AA, Gladyshev VN, Barycki JJ. Crystal structures of oxidized and reduced mitochondrial thioredoxin reductase provide molecular details of the reaction mechanism. Proc Natl Acad Sci USA 2005;102:15018–23.

8.8 References

199

[44] Fritz-Wolf K, Kehr S, Stumpf M, Rahlfs S, Becker K. Crystal structure of the human thioredoxin reductase-thioredoxin complex. Nat Commun 2011;2:383. [45] Fritz-Wolf K, Urig S, Becker K. The structure of human thioredoxin reductase 1 provides insights into C-terminal rearrangements during catalysis. J Mol Biol 2007;370:116–27. [46] Cheng Q, Sandalova T, Lindqvist Y, Arner ES. Crystal structure and catalysis of the selenoprotein thioredoxin reductase 1. J Biol Chem 2009;284:3998–4008. [47] Angelucci F, Dimastrogiovanni D, Boumis G, Brunori M, Miele AE, Saccoccia F, Bellelli A. Mapping the catalytic cycle of Schistosoma mansoni thioredoxin glutathione reductase by X-ray crystallography. J Biol Chem 2010;285:32557–67. [48] Angelucci F, Miele AE, Boumis G, Dimastrogiovanni D, Brunori M, Bellelli A. Glutathione reductase and thioredoxin reductase at the crossroad: the structure of Schistosoma mansoni thioredoxin glutathione reductase. Proteins 2008;72:936–45. [49] Kuriyan J, Krishna TS, Wong L, Guenther B, Pahler A, Williams CH, Jr, Model P. Convergent evolution of similar function in two structurally divergent enzymes. Nature 1991;352:172–4. [50] Waksman G, Krishna TS, Williams CH, Jr., Kuriyan J. Crystal structure of Escherichia coli thioredoxin reductase refined at 2 Å resolution. Implications for a large conformational change during catalysis. J Mol Biol 1994;236:800–16. [51] Lennon BW, Williams CH, Jr., Ludwig ML. Crystal structure of reduced thioredoxin reductase from Escherichia coli: structural flexibility in the isoalloxazine ring of the flavin adenine dinucleotide cofactor. Prot Sci 1999;8:2366–79. [52] Williams CH, Arscott LD, Muller S, Lennon BW, Ludwig ML, Wang PF, Veine DM, Becker K, Schirmer RH. Thioredoxin reductase two modes of catalysis have evolved. Eur J Biochem/ FEBS 2000;267:6110–7. [53] Prongay AJ, Williams CH, Jr. Evidence for direct interaction between cysteine 138 and the flavin in thioredoxin reductase. A study using flavin analogs. J Biol Chem 1990;265:18968–75. [54] Negri A, Rodriguez-Larrea D, Marco E, Jimenez-Ruiz A, Sanchez-Ruiz JM, Gago F. Proteinprotein interactions at an enzyme-substrate interface: characterization of transient reaction intermediates throughout a full catalytic cycle of Escherichia coli thioredoxin reductase. Proteins 2010;78:36–51. [55] Hamill MJ, Chobot SE, Hernandez HH, Drennan CL, Elliott SJ. Direct electrochemical analyses of a thermophilic thioredoxin reductase: interplay between conformational change and redox chemistry. Biochemistry 2008;47:9738–46. [56] Jonsson TJ, Ellis HR, Poole LB. Cysteine reactivity and thiol-disulfide interchange pathways in AhpF and AhpC of the bacterial alkyl hydroperoxide reductase system. Biochemistry 2007;46:5709–21. [57] Helmann JD. Bacillithiol, a new player in bacterial redox homeostasis. Antioxidants & redox signaling 2011;15:123–33. [58] Cheng YQ, Yang M, Matter AM. Characterization of a gene cluster responsible for the biosynthesis of anticancer agent FK228 in Chromobacterium violaceum No. 968. Appl Environ Microb 2007;73:3460–9. [59] Li B, Forseth RR, Bowers AA, Schroeder FC, Walsh CT. A backup plan for self-protection: S-methylation of holomycin biosynthetic intermediates in Streptomyces clavuligerus. ChemBioChem 2012;13:2521–6. [60] Li B, Walsh CT. Streptomyces clavuligerus HlmI is an intramolecular disulfide-forming dithiol oxidase in holomycin biosynthesis. Biochemistry 2011;50:4615–22. [61] Scharf DH, Heinekamp T, Remme N, Hortschansky P, Brakhage AA, Hertweck C. Biosynthesis and function of gliotoxin in Aspergillus fumigatus. Appl Microbiol Biot 2012;93:467–72. [62] Scharf DH, Remme N, Heinekamp T, Hortschansky P, Brakhage AA, Hertweck C. Transannular disulfide formation in gliotoxin biosynthesis and its role in self-resistance of the human pathogen Aspergillus fumigatus. J Am Chem Soc 2010;132:10136–41. [63] Wang C, Wesener SR, Zhang H, Cheng YQ. An FAD-dependent pyridine nucleotide-disulfide oxidoreductase is involved in disulfide bond formation in FK228 anticancer depsipeptide. Chem Biol 2009;16:585–93.

200

8 FDR and structurally related flavoprotein thiol/disulfide-linked oxidoreductases

[64] Claiborne A, Mallett TC, Yeh JI, Luba J, Parsonage D. Structural, redox, and mechanistic parameters for cysteine-sulfenic acid function in catalysis and regulation. Adv Protein Chem 2001;58:215–76. [65] Lountos GT, Jiang R, Wellborn WB, Thaler TL, Bommarius AS, Orville AM. The crystal structure of NAD(P)H oxidase from Lactobacillus sanfranciscensis: insights into the conversion of O2 into two water molecules by the flavoenzyme. Biochemistry 2006;45:9648–59. [66] delCardayre SB, Stock KP, Newton GL, Fahey RC, Davies JE. Coenzyme A disulfide reductase, the primary low molecular weight disulfide reductase from Staphylococcus aureus. Purification and characterization of the native enzyme. J Biol Chem 1998;273:5744–51. [67] Mallett TC, Wallen JR, Karplus PA, Sakai H, Tsukihara T, Claiborne A. Structure of coenzyme A-disulfide reductase from Staphylococcus aureus at 1.54 Å resolution. Biochemistry 2006;45:11278–89. [68] Wallen JR, Mallett TC, Boles W, Parsonage D, Furdui CM, Karplus PA, Claiborne A. Crystal structure and catalytic properties of Bacillus anthracis CoADR-RHD: implications for flavinlinked sulfur trafficking. Biochemistry 2009;48:9650–67. [69] Mueller EG. Trafficking in persulfides: delivering sulfur in biosynthetic pathways. Nat Chem Biol 2006;2:185–94. [70] Ensign SA. Microbial metabolism of aliphatic alkenes. Biochemistry 2001;40:5845–53. [71] Krum JG, Ensign SA. Evidence that a linear megaplasmid encodes enzymes of aliphatic alkene and epoxide metabolism and coenzyme M (2-mercaptoethanesulfonate) biosynthesis in Xanthobacter strain Py2. J Bacteriol 2001;183:2172–7. [72] Jang SB, Jeong MS, Clark DD, Ensign SA, Peters JW. Crystallization and preliminary X-ray analysis of a NADPH 2-ketopropyl-coenzyme M oxidoreductase/carboxylase. Acta Crystallogr D 2001;57:445–7. [73] Nocek B, Jang SB, Jeong MS, Clark DD, Ensign SA, Peters JW. Structural basis for CO2 fixation by a novel member of the disulfide oxidoreductase family of enzymes, 2-ketopropyl-coenzyme M oxidoreductase/carboxylase. Biochemistry 2002;41:12907–13. [74] Pandey AS, Mulder DW, Ensign SA, Peters JW. Structural basis for carbon dioxide binding by 2-ketopropyl coenzyme M oxidoreductase/carboxylase. FEBS Lett 2011;585:459–64. [75] Pandey AS, Nocek B, Clark DD, Ensign SA, Peters JW. Mechanistic implications of the structure of the mixed-disulfide intermediate of the disulfide oxidoreductase, 2-ketopropyl-coenzyme M oxidoreductase/carboxylase. Biochemistry 2006;45:113–20. [76] Clark DD, Allen JR, Ensign SA. Characterization of five catalytic activities associated with the NADPH:2-ketopropyl-coenzyme M [2-(2-ketopropylthio)ethanesulfonate] oxidoreductase/ carboxylase of the Xanthobacter strain Py2 epoxide carboxylase system. Biochemistry 2000;39:1294–304. [77] Kofoed MA, Wampler DA, Pandey AS, Peters JW, Ensign SA. Roles of the redox-active disulfide and histidine residues forming a catalytic dyad in reactions catalyzed by 2-ketopropyl coenzyme M oxidoreductase/carboxylase. J Bacteriol 2011;193:4904–13. [78] Boyd JM, Clark DD, Kofoed MA, Ensign SA. Mechanism of inhibition of aliphatic epoxide carboxylation by the coenzyme M analog 2-bromoethanesulfonate. J Biol Chem 2010;285:25232–42. [79] Krishnakumar AM, Sliwa D, Endrizzi JA, Boyd ES, Ensign SA, Peters JW. Getting a handle on the role of coenzyme M in alkene metabolism. Microbiol Mol Biol Rev 2008;72: 445–56. [80] Warlick BP, Evans BS, Erb TJ, Ramagopal UA, Sriram J, Imker HJ, Sauder JM, Bonanno JB, Burley SK, Tabita FR, Almo SC, Sweedler JS, Gerlt JA. 1-methylthio-D-xylulose 5-phosphate methylsulfurylase: a novel route to 1-deoxy-D-xylulose 5-phosphate in Rhodospirillum rubrum. Biochemistry 2012;51:8324–6. [81] Marcia M, Ermler U, Peng G, Michel H. A new structure-based classification of sulfide:quinone oxidoreductases. Proteins 2010;78:1073–83.

8.8 References

201

[82] Wang R. Two’s company, three’s a crowd: can H2S be the third endogenous gaseous transmitter? FASEB J 2002;16:1792–8. [83] Marcia M, Ermler U, Peng G, Michel H. The structure of Aquifex aeolicus sulfide:quinone oxidoreductase, a basis to understand sulfide detoxification and respiration. Proc Natl Acad Sci USA 2009;106:9625–30. [84] Cherney MM, Zhang Y, James MN, Weiner JH. Structure-activity characterization of sulfide:quinone oxidoreductase variants. J Struct Biol 2012;178:319–28. [85] Cherney MM, Zhang Y, Solomonson M, Weiner JH, James MN. Crystal structure of sulfide:quinone oxidoreductase from Acidithiobacillus ferrooxidans: insights into sulfidotrophic respiration and detoxification. J Mol Biol 2010;398:292–305. [86] Brito JA, Sousa FL, Stelter M, Bandeiras TM, Vonrhein C, Teixeira M, Pereira MM, Archer M. Structural and functional insights into sulfide:quinone oxidoreductase. Biochemistry 2009;48:5613–22. [87] Yeh JI, Claiborne A, Hol WG. Structure of the native cysteine-sulfenic acid redox center of enterococcal NADH peroxidase refined at 2.8 Å resolution. Biochemistry 1996;35:9951–7. [88] The PyMOL Molecular Graphics System, version 1.4.1 Schrödinger, LLC. [89] Patterson S, Alphey MS, Jones DC, Shanks EJ, Street IP, Frearson JA, Wyatt PG, Gilbert IH, Fairlamb AH. Dihydroquinazolines as a novel class of Trypanosoma brucei trypanothione reductase inhibitors: discovery, synthesis, and characterization of their binding mode by protein crystallography. J Med Chem 2011;54:6514–30. [90] Fox BS, Walsh CT. Mercuric reductase: homology to glutathione reductase and lipoamide dehydrogenase. Iodoacetamide alkylation and sequence of the active site peptide. Biochemistry 1983;22:4082–8. [91] Sahlman L, Lambeir AM, Lindskog S, Dunford HB. The reaction between NADPH and mercuric reductase from Pseudomonas aeruginosa. J Biol Chem 1984;259:12403–8. [92] Miller SM, Moore MJ, Massey V, Williams CH, Jr, Distefano MD, Ballou DP, Walsh CT. Evidence for the participation of Cys558 and Cys559 at the active site of mercuric reductase. Biochemistry 1989;28:1194–205. [93] Parsonage D, Youngblood DS, Sarma GN, Wood ZA, Karplus PA, Poole LB. Analysis of the link between enzymatic activity and oligomeric state in AhpC, a bacterial peroxiredoxin. Biochemistry 2005;44:10583–92. [94] Wood ZA, Poole LB, Karplus PA. Structure of intact AhpF reveals a mirrored thioredoxinlike active site and implies large domain rotations during catalysis. Biochemistry 2001;40: 3900–11.

9 Flavoenzymes in pyrimidine metabolism Bruce A. Palfey

Abstract Flavoenzymes participate in the synthesis, interconversion, and degradation of pyrimidines by oxidizing dihydropyrimidines to pyrimidines, reducing pyrimidines to dihydropyrimidines, or by reductively methylating uracil moieties. Redox reactions on the carbon-carbon bond of the pyrimidine heterocycle require the action of an enzymatic acid/base adjacent to the carbonyl and hydride transfer between the heterocycle and FMN. The dihydroorotate dehydrogenases oxidize dihydroorotate to orotate using either Cys or Ser as a base; concerted and stepwise reactions have been found. Similar chemistry is performed on tRNA by the dihydrouridine synthases and on uracil and thymine by the dihydropyrimidine dehydrogenases. The similar active site structures of these enzymes suggest a common evolutionary precursor. In contrast to the similar chemistry in their pyrimidine half-reactions, their non-pyrimidine half-reactions are mechanistically diverse, involving both hydride and single-electron transfers. Flavoenzymes that methylate uracil moieties – flavin-dependent thymidylate synthases and folate/FAD-dependent methyl transferase (a tRNA-methylating enzyme), are less understood, but it is clear that they are not evolutionarily related and they do not appear to have common mechanisms.

9.1 Introduction Pyrimidines, as components of nucleotides, deoxynucleotides, or within polynucleotides, play many crucial roles in biochemistry. As monomers in DNA and RNA polymers, they help encode and execute genetic instructions. The glycosylation of proteins, the assembly of complex carbohydrates, and the synthesis of lipids and glycolipids all rely on precursors that are activated by pyrimidine nucleotide carriers. The growth, function, and responses of cells require pyrimidine derivatives. A network of metabolic reactions interconverts pyrimidine nucleotides between the many forms needed by physiology. The de novo biosynthetic pathway builds the uracil moiety of uridine-5’-monophosphate (UMP) from aspartate through seven reactions conserved throughout biology. Other pyrimidine nucleotides are derived from UMP by enzymes that convert the uracil ring into cytosine or thymine, kinases and phosphatases that add and remove phosphates, reductases that remove the 2’-hydroxyls – a host of reactions providing UTP and CTP for RNA, sugar, and lipid biosynthesis, and dTTP and dCTP for DNA synthesis. Added to this metabolic network are salvage reactions that attach pyrimidine rings to sugars, and catabolic reactions that liberate and disassemble pyrimidine rings. Pyrimidine rings within nucleic acids are also often modified – to perform specialized biological roles, or to be repaired after chemical damage. Flavin-dependent enzymes are scattered throughout this array of important metabolic reactions, linking pyrimidine metabolism to redox metabolism.

204

9 Flavoenzymes in pyrimidine metabolism

The flavoenzymes in pyrimidine metabolism act on uracil or thymine moieties, or their dihydro derivatives. This chapter will consider the substantial similarities and intriguing differences between three related families of enzymes that convert pyrimidines to dihydropyrimidines, or vice versa: The dihydroorotate dehydrogenases (DHODs), which oxidize dihydroorotate to orotate in de novo pyrimidine biosynthesis; the dihydrouridine synthases (DUSs), which reduce particular uracil moieties to dihydrouracil in almost all tRNA, for unknown biological purposes; and the dihydropyrimidine dehydrogenases (DHPDHs), which reduce uracil or thymine as a step in their catabolism. Uracil rings are also methylated by flavoenzymes – in the synthesis of dTMP by flavin-dependent thymidylate synthases (FDTSs), and in the modification of tRNA in some organisms by a folate/FAD-dependent methyl transferase (TrmFO). The mechanisms used by these methylating enzymes are less understood than those of the simpler pyrimidine/dihydropyrimidine interconversions. A last category of flavoenzymes, the DNA photolyases, photochemically repair pyrimidine-pyrimidine adducts in DNA caused by UV damage; these enzymes are not covered in this chapter.

9.2 Pyrimidine/dihydropyrimidine interconversions 9.2.1 Overview Like most flavoenzymes, the catalytic cycles of DHODs, DUSs, and DHPDHs can be divided into reductive half-reactions and oxidative half-reactions. However, it is more useful here to divide the catalytic cycles into the half-reaction involving the pyrimidine substrate and the half-reaction involving the other substrate (򐂰Scheme 9.1). In their pyrimidine half-reactions, DHODs, DUSs, and DHPDHs perform dehydrogenase/reductase reactions on carbon-carbon bonds adjacent to an activating carbonyl group, a type of reaction frequently performed by flavoenzymes (see, for example, chapter 11 in volume 1). The pyrimidine half-reaction of DHOD is its reductive half-reaction; DHO reduces the FMN prosthetic group to form the enone moiety of orotate [1]. Two C-H bonds of DHO are cleaved in the process: one is a hydride transfer to FMN, while O

R1 쏁 H, tRNA R2 쏁 H, 앥CO2앥

HN O

N

O

Pyrimidine half-reaction

R2

HN O

R1

N

R2

R1

E ox

Otherred

Ered

Otherox

Scheme 9.1: Generic catalytic scheme for enzymes that interconvert pyrimidines and dihydropyrimidines.

9.2 Pyrimidine/dihydropyrimidine interconversions

S앥 O H

HS O HN

HN

H

O

O HN

HN

H

R N

HN O

N H

H

H H

앥

N

205

R N

N

O N O

O

N H

Scheme 9.2: Carbon-hydrogen bond cleavages for enzymes that interconvert pyrimidines and dihydropyrimidines, using DHPDH as an example.

the other is a deprotonation by an enzymatic base. Some DHODs use a Cys as the base [2,3], while others use Ser [4,5]. The pyrimidine half-reaction of DUSs is the oxidative half-reaction; tRNA oxidizes the FMN prosthetic group, reducing the enone of a uracil moiety of the substrate [6] – essentially the reverse of the pyrimidine half-reaction of DHOD. Here two C-H bonds are formed on the nascent dihydrouracil, where FMN acts as a hydride donor and an active site residue – a Cys – acts as an acid. Similarly, the pyrimidine half-reaction of DHPDs is the oxidative half-reaction; DHPDHs reduce uracil or thymine to the respective dihydropyrimidines (򐂰Scheme 9.2). As with DUSs, the FMN of DHPDHs acts as a hydride donor in the physiological direction and a Cys residue acts as an acid [7]. The structures of the active sites of the DHODs, DUSs, and DHPDHs have important similarities that might be expected from their similar pyrimidine half-reactions. The pyrimidine moiety lies roughly parallel over the si-face of the isoalloxazine of FMN, with the 6-position of the pyrimidine moiety – the site of hydride transfer – in contact with the N5 of FMN [7–14]. Several residues, especially Asn residues, form hydrogen bonds to the pyrimidine moiety. These Asn residues, when present, have an approximately conserved geometry, but are absent in some of the enzymes. The base/acid, located on a loop that covers the active site, makes contact with the 5-position of the pyrimidine – the site of proton transfers – from the face opposite that in contact with FMN. The pyrimidine active sites of DHODs, DUSs, and DHPDHs likely diverged from a common ancestor; these active sites are in TIM-barrel domains which are augmented with short pieces of secondary structure, additional domains, or additional subunits. In contrast to the significant similarities of the pyrimidine half-reactions, the nonpyrimidine half-reactions of DHODs, DUSs, and DHPDH are quite diverse. The FMN of the pyrimidine site reacts directly either by single-electron transfers with ubiquinone or iron-sulfur clusters (some DHODs and the DHPDHs), or the FMN reacts directly by hydride transfer with fumarate or NADPH (other DHODs and DUSs). Intriguingly, aligning structures by the isoalloxazine of FMN aligns the sites of reaction for the nonpyrimidine reactions involving single-electron transfers near the 8-methyl edge of the flavin (򐂰Fig. 9.1), suggesting an inherent reactivity of this locus towards single-electron transfers.

206

9 Flavoenzymes in pyrimidine metabolism

Accessory domains

Pyrimidine-binding TIM barrel

Domains I–III & V

DHPDH

1B

pyrK, Fe2S2 & FAD

2

Quinone binding site

1A

DHODs

RNA binding helices DUSs

Fig. 9.1: Accessories for TIM-barrels of pyrimidine/dihydropyrimidine interconverting enzymes. The TIM-barrels that process pyrimidines are shown on the right side with a space-filling representation of FMN. In order to illustrate the similar plug-in modularity among these enzymes, the accessory domains have been displaced to the left, except for DUS. Note that the Class 1A DHODs do not have accessory domains – all chemistry occurs in the pyrimidine binding site.

9.2.2 Dihydroorotate dehydrogenases 9.2.2.1 General DHODs are found in most, if not all, organisms. Jensen grouped DHODs by amino acid sequences [5] into phylogenetic classes (򐂰Fig. 9.2). Gratifyingly, the biochemical properties of DHODs segregate according to this phylogenetic scheme. The enzymes were grouped into Class 1 and Class 2; Class 1 was further subdivided into Class 1A and Class 1B. There are large differences between each group. The Class 2 enzymes are membrane-bound monomers found in the mitochondria of all eukaryotes and in the membranes of Gram-negative bacteria. They often have hydrophobic N-terminal amino acid extensions embedded in the membrane, but sometimes they contact only the surface of the membrane by helices that form a tunnel for binding the non-pyrimidine/oxidizing substrate ubiquinone. The Class 2 enzymes use Ser as the active site base during the oxidation of DHO. In contrast, the Class 1 enzymes are soluble enzymes that use Cys as the active site base. The Class 1A enzymes are homodimers that use fumarate as their non-pyrimidine substrate. Class 1A enzymes are found in many Gram-positive bacteria and are also found in some microbial eukaryotes such as yeasts and protozoa. The Class 1B DHODs are the most complex. A core TIM-barrel dimer resembles the 1A dimer, but each of the

9.2 Pyrimidine/dihydropyrimidine interconversions

Tr yp an os om a

cr uz i

Esc he rich ia c oli

Rat tus ratt us lus cu us Mus m

Xenopus laevis

s ien sap mo Ho

Class 2 DHODs – monomers – membrane-bound – oxidized by ubiquinone – Serine as base

cei a bru osom n a p Try

aster lanog ila me h p o s Dro

La ct oc oc cu sl ac tis

Sac cho ram yce s ce revi siae

is ct la s cu oc oc ct La lis faeca ccus roco Ente

Class 1B DHODs – a2b2 heterotetramers – cytosolic – oxidized by NAD – Cysteine as base

ctis myces la Kluyvero

es onocytogen Listeria m

tilis sub us l l i c Ba

207

Class 1A DHODs – dimers – cytosolic – oxidized by fumarate – Cysteine as base

Fig. 9.2: Jensen’s phylogenetic classification of DHODs. The phylogenetic tree was based on a comparatively small number of sequences by today’s post-genomic standards, but has remained invaluable in organizing the biochemistry of DHODs. Properties such as prosthetic group content, oxidizing substrate, and active site base adhere to this classification scheme.

1B subunits binds a protein subunit (called the pyrK subunit [15]) containing an Fe2S2 cluster and an FAD prosthetic group which ultimately reduces NAD. Class 1B enzymes are only found in Gram-positive bacteria, and many bacteria have both a Class 1A and Class 1B DHOD.

9.2.2.2 Mechanisms of the pyrimidine half-reactions The pyrimidine half-reaction of DHODs is the reductive half-reaction. This has been studied by stopped-flow methods and detailed interpretations have been possible for Class 1A and Class 2 enzymes [1,2]. The reaction pathways of these two classes are qualitatively similar but have important quantitative differences. In less than one millisecond after mixing a Class 1A or Class 2 DHOD with DHO (in the absence of O2, at 4°; 򐂰Fig. 9.3), the flavin absorbance spectrum shifts significantly to longer wavelengths, indicating the formation of the DHO complex. This complex rapidly reacts to form a reduced enzyme-orotate charge-transfer complex characterized by absorbance that is significantly higher around 550 nm than that of the free reduced enzyme; this charge-transfer band disappears upon orotate dissociation. The release of orotate from the reduced Class 1A enzyme is rapid, so that the charge-transfer complex is short-lived [2]. In contrast, reduced Class 2 enzymes bind orotate very tightly and release it very slowly [1]; somehow, the reduction of the flavin

208

9 Flavoenzymes in pyrimidine metabolism 0.2 DHODox DHODox-DHO DHODred-orotate DHODred-DHO

A

0.1

0 300

500

700

l (nm) Fig. 9.3: Absorbance spectra observed during the reaction of E. coli DHOD with DHO. The reaction was conducted in a stopped-flow spectrophotometer in the absence of oxygen at 4°. The spectrum of the free oxidized enzyme is shown in black. Within a millisecond after mixing with saturating DHO, the spectrum undergoes a large red shift, indicating the formation of the Michaelis complex. Flavin reduction occurs tens of milliseconds afterwards, giving the reduced enzyme-orotate charge-transfer complex shown in blue. Orotate dissociates very slowly, and ultimately the high excess of DHO traps the reduced enzyme (green spectrum).

markedly decreases the rate of orotate dissociation, perhaps by changing the dynamics of the loop covering the active site. These vastly different rates of orotate release between the Class 1A and Class 2 enzymes control the chemical possibilities of the non-pyrimidine reactions (below) – the active sites of the Class 1A enzymes are vacated at a catalytically competent rate, opening the binding site for the next substrate, while the analogous sites remains blocked in the Class 2 enzymes so that a different binding site is needed for the next substrate. When DHO is oxidized, two C-H bonds of DHO are broken: the hydrogen at the 6-position is transferred as a hydride to FMN, while the somewhat acidic hydrogen at the 5-position (pKa estimated to be ~20 [16]) is removed as a proton by the active site base. These two bond cleavages could occur simultaneously in a concerted reaction, where there is only one transition state separating reactants and products. Alternatively, the two C-H bond cleavages could occur in distinct steps, each with its own transition state, separated by an intermediate (򐂰Scheme 9.3). These extremes of mechanism were probed using deuterium kinetic isotope effects. Stopped-flow studies on the reductive half-reactions of the Class 1A enzyme from Lactococcus lactic showed that the oxidation of DHO was concerted [2]. In contrast, similar studies on the Class 2 enzymes from Escherichia coli and Homo sapiens demonstrated a stepwise mechanism [17]. The mechanisms of DHO oxidation/FMN reduction had previously been examined by measuring deuterium isotope effects on steady-state kinetic parameters for Class 1A [18], Class 2 [19], and Class 1B [20] DHODs. Transient kinetics contradict some of these steady-state studies, but more broadly, the important conclusion emerges that Class 1A DHODs react with a different mechanism than Class 2 DHODs. The rate constant for reduction of DHOD measured by stopped-flow methods increases with increasing pH until it levels off above a pKa value, which is expected as the active site base becomes more deprotonated. A pKa of 8.3 was determined for the

9.2 Pyrimidine/dihydropyrimidine interconversions O H H

HN O

N H

H

O

H FMN CO2앥

O

O

B

N H

H FMN CO2앥

O

O

:B

H FMN CO2앥

H

HN

앥

H H N H

H

HN

O HN

쎵 쎵

O :B

N H

HN

앥 CO2앥 FMNH

H

HN N H

HB

O HB

O

209

H FMN CO2앥

O

N H

HB

앥 CO2앥 FMNH

Scheme 9.3: Concerted and stepwise mechanisms for the oxidation of DHO.

Class 1A enzyme from L. lactis, consistent with the thiolate of Cys being the active site base [2]. The behavior of the Class 2 enzyme from E. coli is more subtle [1,4,17]. The rate constant for flavin reduction increased until it plateaued above an apparent pKa value of 9.5 [1]; a similar result was obtained for the H. sapiens enzyme [17]. This pKa value might be assigned to the ionization of the active site base – a Ser. However, this would be an extreme perturbation of the pKa from its value in aqueous solution of about 14. The structures of these enzymes do not reveal interactions (such as a nearby positive charge) that could provide more than 6 kcal mol−1 of stabilization. Indeed, the Ser is in a somewhat hydrophobic environment, so that the thermodynamic pKa might be expected to increase rather than decrease. Instead of being the thermodynamic pKa of the active site Ser, the pKa represents a change in rate-determining step in the twostep reaction. At “low” pH values, the observed rate constant for flavin reduction is determined partially by the individual rate constants for each C-H bond cleavage. As the pH increases, however, the small amount of the alkoxide form of the active site Ser increases, increasing the rate constant for the deprotonation of DHO; eventually, this becomes too fast to contribute to the net rate constant for reduction and it no longer increases with pH [4]. This concept was corroborated by the loss of a deuterium isotope effect for DHO labeled at the 5-position, but not the 6-position [17]. Thus the pKa is a kinetic pKa, not a thermodynamic value. A proton captured from DHO by the rare alkoxide form of Ser is transferred to bulk solvent through a short network of hydrogen bonds (򐂰Scheme 9.4) connecting the active site Ser, an internal water partly held in place by a Phe, and a Thr which communicates with the external solvent [21]. Mutating the residues that create this hydrogen-bonding network decreases the rate constant for reduction by varying degrees [4]. The pyrimidine binding sites of each class of DHOD are essentially superimposable. They all have Asn residues, a Lys, and a Thr or Ser that make hydrogen bonds to the orotate in crystal structures, and the pyrimidine lies over the flavin in the same orientation. It is very surprising, then, that the nearly identical catalytic apparati of the Class 1A and Class 2 enzymes do not tolerate the interchange of the catalytic bases [2]. Thus mutating the Cys of the Class 1A enzyme from L. lactis to Ser produces an enzyme that is reduced by DHO 4–5 orders of magnitude slower than the wild-type enzyme [2,4], despite the ability of Ser to support a very rapid reaction in the very similar-looking Class 2 enzyme. Class 1A enzymes lack the Thr and Phe that comprise the proton-transfer network

210

9 Flavoenzymes in pyrimidine metabolism Ser

DHO O O H H H HN O

N H

H CO2앥

Thr O O

H

H

Bulk solvent H O H

H pKa 앑14

FMN Ser DHO O O앥 H H O HN C5 H H H O N CO2앥 H FMN

Ser

Thr

H

O

Bulk solvent H O

OA O HN

H O

HH

Thr

O

Bulk solvent H O

H

H

O앥

C5

N CO2앥 H FMNH앥

H

H

O

Scheme 9.4: Proton transfers in the reductive half-reaction of Class 2 DHODs.

supporting the reactivity of the Ser base in Class 2 enzymes. Interestingly, the triple mutant Class 1A enzyme that had all the components of the Class 2 proton-transfer system – Ser, Thr, and Phe – reacted much more slowly than the single Cys → Ser mutant enzyme; adding a new proton-transfer route was decidedly ineffective [22].

9.2.2.3 Mechanisms of the non-pyrimidine half-reactions In contrast to the subtle differences in the pyrimidine half-reactions of the phylogenetic classes of DHODs, there are very large differences between the non-pyrimidine half reactions, most obviously reflected in the different oxidizing substrates. The mechanisms of these reactions have been studied in much less detail than the mechanisms of the pyrimidine half-reactions, but key features are clear.

Class 2 – Ubiquinone After DHO reduces Class 2 DHODs, orotate is held tightly and dissociates only slowly [1]. Ubiquinone, the oxidizing substrate, cannot gain access to N5 of reduced FMN through the occupied pyrimidine binding site. Thus the site of hydride transfer is inaccessible, so that a single-electron transfer mechanism is required for flavin oxidation. The putative ubiquinone binding site is a tunnel (򐂰Fig. 9.4), lined mostly with hydrophobic residues, formed by helices at the N-terminus of the enzyme [23]. No structure has been reported yet of ubiquinone bound to a DHOD, but there are many structures of drug molecules bound to this site. The closest approach possible for a quinone to FMN, partly blocked by a conserved Tyr, is at the 8-methyl edge. From this pocket, hydride transfer from FMN to ubiquinone is impossible; the reaction must proceed through sequential single-electron transfers. Nonetheless, semiquinones were not observed in stopped-flow experiments when menadione [1] or Q1 (a ubiquinone analog) [24] were used to oxidize reduced enzyme. This indicates that the first electrontransfer step determines the rate of the pathway.

9.2 Pyrimidine/dihydropyrimidine interconversions

211

A

O

B

O

O

OH

O

FMNH앥

O

FMN R

O O

FMN R

O 앥O

R

O OH

Fig. 9.4: Reaction of ubiquinone. (A) Class 2 DHODs, such as that from E. coli shown here, have accessory helices (orange) which are the surface that contacts the membrane and form a mostly hydrophobic tunnel thought to be the site of ubiquinone binding. No structures with ubiquinone bound have been reported; it has been docked manually (pink) in a plausible conformation, bringing the quinone moiety close to the 8-methyl edge of FMN (yellow, of course). Tyr 318 partly prevents direct contact of ubiquinone and FMN. Ubiquinone in this model approaches Arg 102, which might be important in stabilizing negative charge as the quinone is reduced. Note that orotate can be seen edge-on and covers N5 of FMN, the site of hydride transfers in flavins. (B) The geometry of quinone binding requires sequential single-electron transfers to oxidize the flavin, forming an intermediary flavin semiquinone-ubiquinone semiquinone pair; this intermediate has not been detected. Negative charge on the ubiquinone semiquinone might be stabilized by Arg 102 shown in (A), and also by His 19 (not shown in (A)). One or both of these residues might protonate ubiquinol.

Class 1A – Fumarate Orotate vacates the active site of the reduced enzyme rapidly, opening the site for reaction with fumarate, a hydride acceptor. Structures show how fumarate can occupy the site by making many of the interactions with Asn residues and Lys that were made by DHO and orotate [25]. Importantly, this binding mode positions one carbon of the carbon-carbon double bond of fumarate over N5 of FMN so that it can receive a hydride, while the other carbon is positioned next to the active site Cys, which would function as an acid in this half-reaction (򐂰Fig. 9.5). Consistent with this concept, the reaction of reduced enzyme with fumarate is faster at lower pH (Nelson, Rider, & Palfey, unpublished).

212

9 Flavoenzymes in pyrimidine metabolism A

앥

B

FlH

Fl ox O

앥

O

앥

O

O O

Cys

앥

Fl

FlH

O

SH Cys

쎵 쎵

O 앥

O

O

앥

O

H S d앥 Cys

앥

S

Cys

d앥 H

앥

O O

O O

SH

Cys

O

앥

O

앥

SH

앥

O

앥

O

앥

O

Flox

앥

O

Fl ox O

앥

앥

O

O O

앥

S

Cys

Fig. 9.5: Oxidation of Class 1A DHODs by fumarate. (A) Class 1A DHODs use the same site for both half-reactions. One of the carboxylates of fumarate occupies the same position that the carboxylate of orotate occupies, forming hydrogen bonds with a lysine and other residues, while the other carboxylate forms a hydrogen bond with Asn residues. This positions one of the olefinic carbons of fumarate over N5 of FMN and the other beneath Cys. (B) Two reaction mechanisms are possible – stepwise and concerted. The stepwise reaction (top row) forms an enediolate intermediate, which have precedent in enzymology. The concerted mechanism would go through a single transition state.

Class 1B – NAD The 8-methyl edge of FMN in the Class 1B enzyme is near the Fe2S2 cluster of the pyrK subunit, and this is adjacent to an FAD, which is in turn adjacent to an NAD binding site [9]. This internal electron transport chain moves the reducing equivalents, initially donated as a hydride by DHO, to NAD in the form of a hydride. Because the Fe2S2 cluster transfers electrons singly, flavin semiquinones – both FMN and FAD – are obligate intermediates. Semiquinones have indeed been observed in titrations and kinetically [26]. However, the preferred catalytic pathway of the Class 1B enzymes is still not known in detail. This is not surprising considering the inherent complexity. The prosthetic groups can hold up to five electrons; there are many ways to transiently distribute two electrons

9.2 Pyrimidine/dihydropyrimidine interconversions

FMNsq Fe2S2,ox FADox FMNox Fe2S2,ox FADox

FMNox Fe2S2,red FADox

FMNox Fe2S2,ox FADsq

0 e앥

1 e앥

FMNred Fe2S2,ox FADox

FMNred Fe2S2,red FADox

FMNsq Fe2S2,red FADox

FMNred Fe2S2,ox FADsq

FMNsq Fe2S2,ox FADsq

FMNsq Fe2S2,red FADsq

FMNox Fe2S2,red FADsq

FMNsq Fe2S2,ox FADred

FMNox Fe2S2,ox FADred

FMNox Fe2S2,red FADred

2 e앥

3 e앥

213

FMNred Fe2S2,ox FADred FMNsq Fe2S2,red FADred

FMNred Fe2S2,red FADred

FMNred Fe2S2,red FADsq

4 e앥

5 e앥

Scheme 9.5: Complexities of electron-distribution in Class 1B DHODs, and possible catalytic cycles.

donated by DHO among the two flavins and the iron-sulfur cluster (򐂰Scheme 9.5). Furthermore, in the steady-state, the enzyme could cycle between the fully oxidized and 2-electron reduced states, or between the 2-electron reduced and 4-electron reduced states, and, if fortuitous single-electron redox events occur, cycles between states containing odd numbers of electrons are imaginable. Significant progress remains to be made in understanding the Class 1B DHODs.

9.2.3 Dihydrouridine synthases 9.2.3.1 General Newly transcribed tRNA molecules are modified extensively. One of the most common modifications is the reduction of specific uracil moieties to dihydrouracil by DUSs (򐂰Scheme 9.6) [27]. The biological function of dihydrouracil is uncertain, but it has been suggested to increase the flexibility of the tRNA molecule. This suggestion is corroborated by the higher amounts of dihydrouracil found in tRNA isolated from psychrophiles than mesophiles, and the lower amounts found in the tRNA isolated from thermophiles. Presumably, at the cold temperatures of psychrophilic growth, more dihydrouracil is needed to disrupt the rigid nucleic acid packing, while at high temperatures where tRNA ought to be inherently flexible, less dihydrouracil is needed. DUSs have TIM-barrel domains with FMN and a Cys at the active site [11,28]. They also have C-terminal helical domains that seem to play a role in binding tRNA. A structure has been determined of a DUS with tRNA covalently bound by a Michael adduct to

214

9 Flavoenzymes in pyrimidine metabolism O

O

HN O

HN O

N

NADPH 쎵

DUS

N

NADP 쎵

Scheme 9.6: Reaction catalyzed by DUSs.

the active site Cys [11], which formed spontaneously upon co-expression of the enzyme and tRNA. The complex shows that tRNA makes extensive electrostatic interactions with the protein.

9.2.3.2 Pyrimidine half-reactions DUSs reduce tRNA in their oxidative half-reactions. Transient kinetic studies on the reaction of yeast DUS2 with in vitro-transcribed tRNALeu showed that the reaction was far too slow to be physiologically relevant [6]. However, when the tRNA was isolated from a yeast strain whose DUS2 was knocked-out (thus assuring that the tRNA would not already be reduced), rapid reactions were obtained. tRNA transcribed in vivo would be exposed to other modifying enzymes. These observations strongly suggested that in order for tRNALeu to be a good substrate, at least one other modification must be present; this putative modification has not been identified. Fast reactions also required the active site Cys, consistent with its assignment as the catalytic acid; its mutation to Ala slowed the reaction of in vivo transcribed tRNALeu to the level of in vitro transcribed tRNALeu with the wild-type enzyme.

9.2.3.3 Non-pyrimidine half-reaction NADPH reduces the FMN of yeast DUS2 in its reductive half-reaction [6]. The use of deuterium-labeled substrate showed that the proR hydride of NADPH is transferred. Stopped-flow experiments showed that there was a kinetic isotope effect of 3.5 on reduction, indicating that hydride transfer was at least partly rate determining. Mutating the active site Cys to Ala did not slow reduction, indicating that it had no role in this half-reaction.

9.2.4 Dihydropyrimidine dehydrogenases 9.2.4.1 General DHPDs use NADPH to reduce uracil or thymine in the first step of the catabolism of these compounds [7]. Mammalian DHPDs are very large enzymes – dimers of ~110 kDa

9.3 Methylations

FADA

N-terminal [4Fe-4S]A

FMNA C-terminal [4Fe-4S]B

NADPHA

C-terminal [4Fe-4S]A

NADPHB

FADB

N-terminal [4Fe-4S]B

215

UracilA

UracilB

FMNB

Fig. 9.6: Electron-transfer pathways of DHPDHs. Reducing equivalents donated by NADPH to FAD are transferred one at a time through the iron-sulfur clusters to FMN, where they are used to reduce uracil or thymine. Each redox center is labeled with A or B to indicate on which subunit of the dimer it resides. Note that the complete electron-transport pathway uses components from both subunits.

subunits. Each subunit is comprised of five domains. The pyrimidine substrates react with FMN at a TIM barrel domain (domain IV) that is highly similar to Class 1A DHODs, while NADPH (bound to domain III) reacts with FAD bound to domain II. Other domains contain four Fe4S4 clusters. As with the Class 1B DHODs, electrons move between flavins by the iron-sulfur clusters, only more intervene in DHPDHs. The dimers are intertwined so that each subunit contributes clusters of the internal electron-transport chain (򐂰Fig. 9.6).

9.2.4.2 Pyrimidine half-reaction When operating in the physiological direction, DHPDs reduce uracil or thymine in the oxidative half-reaction [7]. Electrons, donated singly from an iron-sulfur cluster near the 8-methyl edge of FMN, reduce it to the hydroquinone. The active site has three Asn residues for binding pyrimidines, analogous to the active site of DHODs. Covering the active site is a loop with a Cys that functions as an acid. Mutating this Cys impairs turnover significantly. The mechanism of reduction of uracil was found to be stepwise based on the measurement of substrate tritium kinetic isotope effects and solvent isotope effects.

9.3 Methylations 9.3.1 Overview Pyrimidine rings are frequently methylated in biochemistry. Many of these are not redox reactions; these use S-adenosyl methionine to donate a methyl group. However, some methylating enzymes transfer a carbon at the methylene oxidation level,

216

9 Flavoenzymes in pyrimidine metabolism

donated from methylenetetrahydrofolate (CH2THF), and subsequently reduce it to the methyl-level of oxidation. This is the case for flavin-dependent thymidylate synthases (FDTSs), and for folate/FAD-dependent methyl transferase (TrmFO), which convert uracil moieties to thymine moieties. These enzymes must perform many chemical tasks in order to form a methylated uracil moiety. The carbon-transfer reaction is generally thought to be the reaction of a nucleophilic pyrimidine with the electrophilic methylene. The methylene group carried by CH2THF is very electrophilic upon imine formation. Pyrimidines, however, are not extremely nucleophilic at the 5-position – an activation mechanism is required. These issues have been studied for many decades in the classic thymidylate synthases, which do not use flavins; concepts from the flavin-independent thymidylate synthases have guided research on the flavindependent methylating enzymes.

9.3.2 Flavin-dependent thymidylate synthase The conversion of dUMP to dTMP is critical for DNA synthesis. It has been estimated that as many as one-third of bacteria perform this reaction with FDTS. FDTS is found predominately in bacteria that spend at least some of their lives in anaerobic environments. Many pathogenic bacteria use FDTS. The substrates of FDTSs are dUMP, CH2THF, and NADPH. Many structures of FDTSs have been solved [29–31]. The enzyme is a homotetramer of oblong subunits with a unique fold (򐂰Fig. 9.7). The FAD prosthetic groups lie extended within the crevices between subunits, with the isoalloxazines at the ends of the enzyme. There are four active sites per tetramer, each formed by residues from three of the four subunits. Several structures show deoxynucleotides bound next to the si-face of FAD. The binding site of CH2THF is on the re-face of FAD [32]. Interestingly, the methylene carbon of CH2THF is much too far from its ultimate destination, the 5-position of dUMP. It seems likely that the crystallography has captured a complex with the pteridine moiety of CH2THF located on the wrong side of the isoalloxazine. It is worth noting that the structure of the oxidized enzyme was determined but the reduced form would be catalytically relevant; this might have important conformational consequences. The mechanism of FDTS is far from settled, but key facts are emerging. It is widely believed that NADPH reduces the enzyme in the reductive half-reaction and then NADP dissociates. Curiously, the enzyme is not completely stereospecific for the hydride transferred from NADPH [33] – a rarity in enzymology. This might be a hint that the reaction of NADPH is fortuitous and the physiological substrate has yet to be identified. The mechanisms of transfer of the methylene carbon from CH2THF to dUMP and the reduction of the methylene have been investigated. The flavin-independent thymidylate synthases activate the uracil moiety of dUMP by the Michael addition of an active site Cys to the enone of dUMP. No such Cys is present in FDTS, nor has a reasonable alternative nucleophile been found. The proximity of N5 of FAD to the 6-position of dUMP suggested the possibility that hydride donated by FAD could be the activating nucleophile, but this concept was disproved by experiments using 5-fluoro-dUMP [34]. The fluoro analog of the substrate cannot make product. At some point in the conversion of dUMP to dTMP, the 5-position is deprotonated; the analogous reaction of the

9.3 Methylations

217

A

B

dUMP

FAD

Fig. 9.7: Structure of the FDTS-dUMP complex. (A) The enzyme is a tetramer of identical polypeptides. FAD molecules can be seen in the crevices facing upwards and downwards. The four active sites of the tetramer are formed by three subunits at the ends of the enzyme, where dUMP is bound. (B) A close-up is shown of the active site from the left side, top, of the tetramer in (A). The uracil ring of dUMP is in contact with the isoalloxazine of FAD, with N5 in the proper position to reduce the pyrimidine moiety.

C-F bond would be the formation of F+, an impossibility under biochemical conditions. However, the enone moiety of 5-fluoro-dUMP can be reduced. Therefore, if the mechanism starts with reduction of uracil, mixing reduced enzyme, 5-fluoro-dUMP, and CH2THF with the enzyme should oxidize the enzyme – the blockage by the C-F bond would occur afterwards. This was not the case; 5-flouro-dUMP and CH2THF do not oxidize the enzyme. NMR studies showed that solvent deuterium was incorporated into product dTMP by the enzyme from Thermotoga maritima [35]. When the reaction was performed at 37° – well below the physiological temperature of the enzyme – solvent-derived hydrogen was distributed between the 6-position of dTMP and the methyl group. This is strong evidence that label washes into N5 of the reduced flavin which transfers a hydride to the 6-position of the pyrimidine, followed by a 1,3-hydride shift to give dTMP. Notably, stopped-flow and quenching experiments have detected intermediates in the oxidative half-reaction. These observations lead to the mechanism

218

9 Flavoenzymes in pyrimidine metabolism 앥

O

O

H HN O

HN O

N R

N쎵 쎵

HN O

쎵

N R

N H

THF

R

HA THF

HN 쎵

N R

R

앥

N

N

O NH

O

O

O

N

CH2THF R

O

쎵

O

N

:B

HN

NH

N R

dUMP

앥

O

R

O

O NH

N O

HN O

N

dTMP N

H H

R exo- dTMP

Fig. 9.8: Working model for the synthesis of dTMP by FDTS. Novel features are the activation of uracil as a nucleophile by electrostatic polarization, highlighted in maize, and a 1,3-hydride shift, highlighted in blue.

in 򐂰Fig. 9.8 as a current working model. A number of unpublished lines of evidence suggest that the uracil moiety of dUMP is activated by electrostatic polarization upon binding.

9.3.3 Folate/FAD-dependent methyl transferase (TrmFO) TrmFO superficially performs the same reaction as FDTS – the reductive methylation of a uracil moiety [36]. Its substrates are tRNA, CH2THF, and NAD(P)H, with an apparent preference for NADPH (򐂰Scheme 9.7) [37]. However, the mechanism of TrmFO appears to be very different FDTS. The crystal structure of TrmFO from Thermus thermophilus (򐂰Fig. 9.9) reveals a Cys near FAD at the putative active site [38]. Additionally, there is a conserved Cys that is quite distant from the flavin in a pocket that binds glutathione. Both Cys residues are vital for activity. In the enzyme from Bacillus subtilis, the distant Cys makes a covalent adduct with an artificial RNA helix containing 5-fluorouracil, an analog incapable of making product [39]. Furthermore, a significant amount of Bacillus TrmFO expressed in E. coli is purified with an unusual natural flavin derivative, identified as an N5-adduct between reduced FAD, a methylene group, and the Cys near FAD [40]. It was proposed that the adduct is on the catalytic pathway – that it is formed by the transfer of the methylene group from CH2THF to the nucleophilic N5 of reduced flavin and sulfur of Cys (򐂰Fig. 9.10). An equilibrium would exist between this adduct to the protein and the highly electrophilic iminium formed by elimination of Cys. The iminium would react with the uracil of tRNA activated by the distant Cys in

9.3 Methylations

219

O HN H2N

N

NADPH 쎵 HN O CH2THF

H N

O

N TrmFO

쎵

N N R

O HN H2N TrmFO

N

NADP 쎵 HN O THF

H N N H

O H N

R

N

쎵

Scheme 9.7: Reaction catalyzed by TrmFO.

Cys 223

THF

FAD

Cys 51

Fig. 9.9: Structure of TrmFO from T. thermophilus. The cartoon representation is colored according to residue number, with the N-terminus being dark blue. Note that despite the net resemblance of the TrmFO and FDTS reactions, the structure of TrmFO has no resemblance to that of FDTS (򐂰Fig. 9.7). The two catalytically important Cys residues are shown in space-filling representation. The pteridine moiety of tetrahydrofolate (THF) was found next to the flavin, but the rest of the molecule was too disordered to be visualized.

O

HN

N CH2THF

tRNA

N

O

N H

N

N

O

NH

HN O tRNA

N

O

앥

N H

O

N

R2

NH

HN

O

H2N

NH

O

S Cys226

N

R1

N H

N

O

N

쎵

N

O

N

R1

H2N

HN

O

N H

N

앥

N

N

NH

O

HS Cys226

N

O R2

N

앥

R1

O

HN

N

O

tRNA

N

O

N H

N

N

N

R1

R1

O

앥

N

N H

O

O

tRNA

N

쎵

N

N

R1

NH

R2

O HA :B

앥

N NH

S Cys53

S Cys226

N

N

앥

N

N H

O

O

O HS Cys53

THF R1

tRNA

N

O

N H

H N

앥

O

HN

N

HN O

HN H2N

NH

R2

NH

O

S Cys226

N H

N

O

N

앥

O

O

HN

tRNA

N

O

쎵

N

NH O S Cys53

앥

N

S Cys53

앥

R1 N

Fig. 9.10: Proposed mechanism for methylation of tRNA by TrmFO. CH2THF donates its methylene group to reduced FAD by first forming the iminium, forming the reactive reduced flavin-iminium; the stable Cys adduct (box on the right) is isolated when the B. subtilis enzyme is expressed in E. coli. The flavin-iminium is attacked by the enolate formed by the Michael addition of Cys to uracil on tRNA, leading to carbon-transfer, reduction of the enone moiety on the nascent product, and finally enzyme-tRNA adduct cleavage.

H2N

HN

O

N H

N

R1

O

220 9 Flavoenzymes in pyrimidine metabolism

9.4 References

221

a Michael addition. It is interesting to note that the distance from the Cys that forms the adduct with RNA to N5 of FAD is 19 Å; either TrmFO undergoes a massive conformational change to bring reactive centers close, or it dimerizes. This role proposed for the reduced flavin in the TrmFO reaction – as an electrophilic carbon transfer-agent – is highly unusual and suggests that new flavin mechanisms still remain to be uncovered.

Acknowledgements I gratefully acknowledge the invigorating collaborations with the students from my laboratory which helped develop the perspectives presented here. I especially want to thank Professors David P. Ballou (University of Michigan), Kaj Frank Jensen (University of Copenhagen), and the late Vincent Massey (University of Michigan) for critical support. The author’s lab has been supported by NIH R01GM61087 and NSF CHE 1213620.

9.4 References [1] Palfey BA, Björnberg O, Jensen KF. Insight into the chemistry of flavin reduction and oxidation in Escherichia coli dihydroorotate dehydrogenase obtained by rapid reaction studies. Biochemistry 2001;40:4381–90. [2] Fagan RL, Jensen KF, Björnberg O, Palfey BA. Mechanism of flavin reduction in the class 1A dihydroorotate dehydrogenase from Lactococcus lactis. Biochemistry 2007;46:4028–36. [3] Bjornberg O, Rowland P, Larsen S, Jensen KF. Active site of dihydroorotate dehydrogenase A from Lactococcus lactis investigated by chemical modification and mutagenesis. Biochemistry 1997;36:16197–205. [4] Kow RL, Whicher JR, McDonald CA, Palfey BA, Fagan RL. Disruption of the proton relay network in the class 2 dihydroorotate dehydrogenase from Escherichia coli. Biochemistry 2009;48:9801–9. [5] Bjornberg O, Grüner AC, Roepstorff P, Jensen KF. The activity of Escherichia coli dihydroorotate dehydrogenase is dependent on a conserved loop identified by sequence homology, mutagenesis, and limited proteolysis. Biochemistry 1999;38:2899–908. [6] Rider LW, Ottosen MB, Gattis SG, Palfey BA. Mechanism of dihydrouridine synthase 2 from yeast and the importance of modifications for efficient tRNA reduction. J Biol Chem 2009;284:10324–33. [7] Schnackerz KD, Dobritzsch D, Lindqvist Y, Cook PF. Dihydropyrimidine dehydrogenase: a flavoprotein with four iron–sulfur cluster. Biochimica et Biophysica Acta (BBA) - Proteins & Proteomics 2004;1701:61–74. [8] Nielsen FS, Jensen KF, Rowland P, Larsen S. Purification and characterization of dihydroorotate dehydrogenase a from lactococcus lactis, crystallization and preliminary X-Ray diffraction studies of the enzyme. Protein Sci 1996;5:852–6. [9] Rowland P, Nørager S, Jensen KF, Larsen S. Structure of dihydroorotate dehydrogenase B: electron transfer between two flavin groups bridged by an iron-sulphur cluster. Structure 2000. [10] Rowland P, Nielsen FS, Jensen KF, Larsen S. The crystal structure of the flavin containing enzyme dihydroorotate dehydrogenase A from Lactococcus lactis. Structure 1997;8:1227–38. [11] Yu F, Tanaka Y, Yamashita K, Suzuki T, Nakamura A, Hirano N, Suzuki T, Yao M, Tanaka I. Molecular basis of dihydrouridine formation on tRNA. Proc Natl Acad Sci 2011;108:19593–8. [12] Inaoka DK, Sakamoto K, Shimizu H, Shiba T, Kurisu G, Nara T, Aoki T, Kita K, Harada S. Structures of Trypanosoma cruzi dihydroorotate dehydrogenase complexed with substrates

222

[13]

[14]

[15]

[16]

[17] [18] [19] [20]

[21] [22] [23] [24]

[25]

[26]

[27]

[28]

[29]

[30]

9 Flavoenzymes in pyrimidine metabolism and products: atomic resolution insights into mechanisms of dihydroorotate oxidation and fumarate reduction. Biochemistry 2008;47:10881–91. Schneider G, Schnackerz KD, Lindqvist Y. Crystal structure of dihydropyrimidine dehydrogenase, a major determinant of the pharmacokinetics of the anti-cancer drug 5-fluorouracil. EMBO J 2001;20:650–60. Schneider G, Schnackerz KD, Lindqvist Y. Crystal structure of the productive ternary complex of dihydropyrimidine dehydrogenase with NADPH and 5-iodouracil implications for mechanism of inhibition and electron transfer. J Biol Chem 2002;277:13155–166. Nielsen FS, Andersen PS, Jensen KF. The B form of dihydroorotate dehydrogenase from Lactococcus lactis consists of two different subunits, encoded by the pyrDb and pyrK genes, and contains FMN, FAD, and [FeS] redox centers. J Biol Chem 1996;271:29359–65. Argyrou A, Washabaugh MW. Proton transfer from the C5-proR/proS positions of L-dihydroorotate: general-base catalysis, isotope effects, and internal return. J Am Chem Soc 1999;121:12054–62. Palfey BA, Fagan RL. Analysis of the kinetic isotope effects on initial rates in transient kinetics. Biochemistry 2006;45:13631–40. Pascal RA, Walsh CT. Mechanistic studies with deuterated dihydroorotates on the dihydroorotate oxidase from Crithidia fasciculata. Biochemistry 1984;23:2745–52. Hines V, Johnston M. Analysis of the kinetic mechanism of the bovine liver mitochondrial dihydroorotate dehydrogenase. Biochemistry 1989;28:1222–6. Argyrou A, Washabaugh MW, Pickart CM. Dihydroorotate dehydrogenase from Clostridium oroticum is a class 1B enzyme and utilizes a concerted mechanism of catalysis. Biochemistry 2000;39:10373–84. Small YA, Guallar V, Soudackov AV, Hammes-Schiffer S. Hydrogen bonding pathways in human dihydroorotate dehydrogenase. J Phys Chem B 2006;110:19704–10. McDonald CA, Palfey BA. Substrate binding and reactivity are not linked: grafting a protontransfer network into a class 1A dihydroorotate dehydrogenase. Biochemistry 2011;50:2714–6. Liu S, Neidhardt EA, Grossman TH, Ocain T, Clardy J. Structures of human dihydroorotate dehydrogenase in complex with antiproliferative agents. Structure 2000;8:25–33. Malmquist NA, Gujjar R, Rathod PK, Phillips MA. Analysis of flavin oxidation and electrontransfer inhibition in Plasmodium falciparum dihydroorotate dehydrogenase. Biochemistry 2008;47:2466–75. Inaoka DK, Sakamoto K, Shimizu H, Shiba T, Kurisu G, Nara T, Aoki T, Kita K, Harada S. Structures of Trypanosoma cruzi dihydroorotate dehydrogenase complexed with substrates and products: atomic resolution insights into mechanisms of dihydroorotate oxidation and fumarate reduction. Biochemistry 2008;47:10881–91. Combe JP, Basran J, Hothi P, Leys D, Rigby SEJ, Munro AW, Scrutton NS. Lys-D48 is required for charge stabilization, rapid flavin reduction, and internal electron transfer in the catalytic cycle of dihydroorotate dehydrogenase B of Lactococcus lactis. J Biol Chem 2006;281:17977–88. Whipple JM, Lane EA, Chernyakov I, D’Silva S, Phizicky EM. The yeast rapid tRNA decay pathway primarily monitors the structural integrity of the acceptor and T-stems of mature tRNA. Genes & Development 2011;25:1173–84. Park F, Gajiwala K, Noland B, Wu L, He D, Molinari J, Loomis K, Pagarigan B, Kearins P, Christopher J, Peat T, Badger J, Hendle J, Lin J, Buchanan S. The 1.59 A resolution crystal structure of TM0096, a flavin mononucleotide binding protein from Thermotoga maritima. Proteins 2004;55:772–4. Mathews II, Deacon AM, Canaves JM, McMullan D. Functional analysis of substrate and cofactor complex structures of a thymidylate synthase-complementing protein. Structure 2003;11:677–90. Sampathkumar P, Turley S, Ulmer JE, Rhie HG, Sibley CH, Hol WGJ. Structure of the Mycobacterium tuberculosis flavin dependent thymidylate synthase (MtbThyX) at 2.0Å resolution. J Mol Biol 2005;352:1091–104.

9.4 References

223

[31] Graziani S, Bernauer J, Skouloubris S, Graille M. Catalytic mechanism and structure of viral flavin-dependent thymidylate synthase ThyX. J Biol Chem 2006;281:24048–57. [32] Koehn EM, Perissinotti LL, Moghram S, Prabhakar A, Lesley SA, Mathews II, Kohen A. Folate binding site of flavin-dependent thymidylate synthase. Proc Natl Acad Sci USA 2012;109:15722–7. [33] Agrawal N, Lesley SA, Kuhn P, Kohen A. Mechanistic studies of a flavin-dependent thymidylate synthase. Biochemistry 2004;43:10295–301. [34] Gattis SG, Palfey BA. Direct observation of the participation of flavin in product formation by thyX-encoded thymidylate synthase. J Am Chem Soc 2005;127:832–3. [35] Koehn EM, Fleischmann T, Conrad JA, Palfey BA, Lesley SA, Mathews II, Kohen A. An unusual mechanism of thymidylate biosynthesis in organisms containing the thyX gene. Nature 2009;458:919–23. [36] Delk AS, Nagle DP, Rabinowitz JC, Straub KM. The methylenetetrahydrofolate-mediated biosynthesis of ribothymidine in the transfer-RNA of Streptococcus faecalis: incorporation of hydrogen from solvent into the methyl moiety. Biochem Biophys Res Com 1979;86:244–51. [37] Yamagami R, Yamashita K, Nishimasu H, Tomikawa C, Ochi A, Iwashita C, Hirata A, Ishitani R, Nureki O, Hori H. The tRNA recognition mechanism of folate/FAD-dependent tRNA methyltransferase (TrmFO). J Biol Chem 2012;287:42480–94. [38] Nishimasu H, Ishitani R, Yamashita K, Iwashita C, Hirata A, Hori H, Nureki O. Atomic structure of a folate/FAD-dependent tRNA T54 methyltransferase. Proc Natl Acad Sci USA 2009;106:8180–5. [39] Hamdane D, Argentini M, Cornu D, Myllykallio H, Skouloubris S, Hui-Bon-Hoa G, GolinelliPimpaneau B. Insights into folate/FAD-dependent tRNA methyltransferase mechanism: role of two highly conserved cysteines in catalysis. J Biol Chem 2011;286:36268–80. [40] Hamdane D, Argentini M, Cornu D, Golinelli-Pimpaneau B, Fontecave M. FAD/folatedependent tRNA methyltransferase: flavin as a new methyl-transfer agent. J Am Chem Soc 2012;134:19739–45.

10 Excited state electronic structure of flavins and flavoproteins from theory and experiment Goutham Kodali and Robert J. Stanley

Abstract Flavins are a premiere photobiological cofactor. In the past 15 years it has become apparent that Nature utilizes this function in all oxidation and protonation states, as well as various spin states. An elucidation of the detailed electronic structure of these transient photoexcited states has benefited from advances in experimental and theoretical tools, including Stark spectroscopy, ultrafast laser spectroscopy, and ab-initio and semi-empirical quantum mechanical methods. Progress in these areas over the past eleven years for simple flavins and flavoproteins are highlighted in this review. In addition, several studies about the application of flavin photochemistry to artificial DNA repair and photosynthesis are illustrated.

10.1 Introduction Chemical reactivity takes many forms. Covalent bond formation, isomerization, electron and proton transfer, and changes in multiplicity all modify the chemical identity and properties of molecules. Several other transformations become possible upon resonant absorption of a quantum of light. Photochemical transformations include all of the above but also open up reactions involving energy transfer. The reaction rate will depend strongly on the driving force of the reaction and, in all but unimolecular reactions, the intermolecular distance and orientation of the reacting partners. At close distances, orbital overlap can play a dominant role, while a larger distances it is primarily the electric fields of the reactants, both static and time-varying, that mediate the reaction. Biological molecules are perfect examples of how these reaction parameters have been optimized to achieve a single outcome. The placement of cofactors and substrate molecules in a protein, the driving force of the reaction, and the local environment set up by the protein are key to quenching of subsequent or side reactions. This is the point of view taken by Losi in his review of flavin photoreceptors [1], and one that the authors’ own research program has supported for some time. In this review, we explore the photochemistry of flavins in the context of how photoexcited state charge redistribution and orientation of reacting partners are employed synergistically towards the final product outcome. Thirty years ago, though ubiquitous, flavins were almost exclusively viewed as biological cofactors whose functional diversity was derived from their unique ability to act as one or two electron transfer agents in their ground electronic state (see 򐂰Scheme 10.1 for structure and numbering of the three neutral oxidation states). However, work begun

226

10 Excited state electronic structure of flavins and flavoproteins R H3C

N

9 7

H3C

N

9a 10 10a 1

8

4a

5a

N

6

4

2

O

NH 3

5

O

FIox R H3C

N

9 7

H3C

N

9a 10 10a 1

8

4a

5a

N H

6

4

R

9a 10 10a 1

8 7

H3C

H N

N

9

4a

5a 6

N H FIH2

O

NH 3

O

FIH •

H3C

2

4

2

O

NH 3

O

Scheme 10.1: Neutral flavin oxidation states and numbering.

in the 1960s by Claud Rupert [2] suggested that these highly colored molecules might be involved in the blue light mediated repair of UV-damaged DNA, as described earlier by Dulbecco [3]. The repair enzyme was dubbed DNA photolyase by Werbin and Mintao, who supplied the first evidence that photolyase was a flavoprotein [4]. The field was advanced dramatically when the Sancar group cloned and overexpressed Escherichia coli photolyase in milligram quantities [5]. Since the discovery of DNA photolyase and the recognition that flavins could serve as biological light sensors, there has been an ever-accelerating number of papers highlighting new roles for flavins as a central photobiological cofactor. In parallel, there has been increasing interest in utilizing the excited state properties of flavins to study basic photochemical processes to making photonic materials that can act as photosensors, artificial photolyases, or for solar energy conversion. Many aspects of this research up through 1991 can be found in the excellent reviews by Heelis [6] and Tollin [7]. For a review summarizing flavin spectroscopy from 1991–2001 see the review by Stanley [8]. The aim of the present review is to provide an update of the many new advances that have been made since about 2001. These advances rest on several factors, primarily improved time resolution and signal to noise ratios in both optical and electron paramagnetic resonance spectroscopies, the employment of Stark spectroscopy, and advances in computational methods with concomitant hardware improvements. The increasing availability of genomic sequencing and facile approaches to mutagenesis have complemented the advances in physical methods. The review is structured as follows. A brief introduction is given on how charge distributions are described. We then take a “Stark-centric” view of the excited state electronic structure of flavin as determined by our Stark spectroscopy experiments. These experiments form a basis for comparison with theoretical calculations, which include the

10.2 Flavin photophysics and the electronic structure of its excited states

227

singlet, doublet, and charge transfer excited states that are relevant to flavin photobiology. From here, experiments and calculations about photolyases serve to show how theory and experiment have competed to form a working picture of this primal photoflavoprotein. Moving to ultrafast spectroscopy, we highlight the evidence for vibronic coherence in flavin and the importance of dark electronic states. New data on the vibrational modes of ground and excited state flavins are discussed, as discovered through state of the art infrared and Raman approaches. These are complimented by some new results using low temperature spectroscopy. The emphasis turns to flavin photochemistry, including new and interesting reactions, as well as updates on the complexity of standard photoredox reactions that many of us use on a routine basis. We close the review with a recounting of potential applications of flavin excited states. These synthetic and genetically engineered systems represent photosensors and other artificial protein motifs, as well as flavin-based materials for solar energy conversion. Because of the large number of papers in this broad subject, many meritorious reports are necessarily omitted. Many of these are already widely acknowledged. Here, we hope to bring lesser known but equally meritorious work to light.

10.2 Flavin photophysics and the electronic structure of its excited states In 1950, Gregorio Weber began his studies on the fluorescence of various flavins [9]. The field was advanced with the application of semi-empirical electronic structure methods to experiment as exemplified by P.S. Song [10–13]. Song also made seminal experimental contributions in assigning spectroscopic transitions of flavins in a variety of solvents and oxidation states [12,14]. The role of flavin charge transfer, acting as either a donor or acceptor, was explored by several groups. Much of the information about flavin electronic structure comes from studies in flavoproteins from ca. 1960–1990, particularly by Massey [15–17], Ghisla [18–20], Edmonson and Tollin [21], Visser [22–24], and Mataga [25–27]. A central goal in all of these studies was to determine the excited state charge distribution of flavins in its many redox, spin, and protonation states.

10.2.1 Moments of the charge distribution A description of the charge density of excited electronic states (e.g. S1, S2, T1, etc.) is the necessary starting point for their subsequent evolution and photochemical reactivity. The charge density, which can be expanded in monopole (ion), dipole, quadrupole (polarizability) and higher-order terms, determines the potential of the molecule and thus how it interacts with other charge distributions and electric fields. If we consider only neutral molecules then the monopole term is zero and the dipole moment and polarizability provide familiar descriptions how charge is distributed in the electronic states of flavins and other chromophores. Dipole moments describe the partitioning of charge into electron-rich and electron-deficient parts of the molecule. This becomes the major determinant in how solvent interacts with the chromophore. The polarizability represents how easily electron density can be redistributed under the influence of an external electric field, such as when that field originates in the solvent (reaction field). However, the quantity of first consequence in our discussion is the is the transition dipole moment, as it determines to what degree the molecule can absorb a photon of light.

228

10 Excited state electronic structure of flavins and flavoproteins

____›

The transition dipole moment, m fi (where i is the index of the initial state and f is the index of the final state), is an intrinsic property of the chromophore ____›that describes the magnitude and preferred direction for the absorption of a photon. | m fi |2 is proportional to the more familiar molar decadic extinction coefficient, ε. As will be shown, vectorial descriptions of charge redistribution resulting require a ____› from an optical transition ____› knowledge of the magnitude and direction of m fi. The computation of m fi is done routinely, although the accuracy of excited ____› state calculations rarely serves as more than a semi-quantitative guide to the true m fi, particularly when solvent is involved. Of great importance is the experimental determination of the transition dipole moment, as these measurements often reveal hidden electronic____ states and serve to test computational › . methods. The primary method for measuring m fi is linear dichroism ___› The permanent ground and excited state dipole moments, µ k, k = 0, 1, 2, etc. are vectors whose directions and magnitudes are a signature property of a molecule. Symmetric molecules have zero ground state dipole moment (though this is not necessarily the case for the excited state) while asymmetric molecules like flavins can have rather large ground state dipole moments. For solvated molecules, a large permanent ground state molecular dipole will orient the solvation shell through dipolar and inductive interactions with the solvent molecules. This can lead to changes in the dielectric properties of the solution, providing a way to measure the magnitude, if not the direction of the molecular dipole moment.

10.2.2 Experimental techniques for the determination of excited state electronic structure The experimental ___› determination of excited state dipoles for chromophores in the condensed phase, µ k, k > 0, is a difficult task. To our knowledge, there is no technique that can determine ___› the excited state dipole moment of large solvated molecules without reference to µ 0, except for gas phase approaches where resolution of rotational lines is obtained [28]. For condensed phase studies, several techniques have been employed: solvatochromism [29], time-resolved molecular relaxation [30], time-resolved microwave conductivity [31], and Stark (or electroabsorption) spectroscopy [29,32]. Solvatochromism [29], the and easiest to implement, affords a measure ___› most ___› familiar ___› of the difference dipole, Δµ k0 = µ k − µ 0, through its interaction with the solvent reaction field. This produces energy shifts that change the absorption and emission maxima of chromophores when the excited state charge distribution is no longer at minimum energy with respect to the ground state solvation shell. Several theories have been employed [33], but almost all suffer from the assumption that the ground and excited state dipoles are parallel, which is often not the case. The theory of Bakhshiev [34] does not make this assumption, but the angle between the two permanent dipoles is often underdetermined from the available data. Time-resolved molecular relaxation [31] monitors the time-resolved solvatochromic shift. It makes use of Bakhshiev’s theory [34] but requires that the molecule be emissive, which is not always the case. Time-resolved microwave conductivity [31] is an interesting approach that requires the use of a microwave cavity filled with nonpolar organic solvents, wherein the excited state, generated by pulsed optical excitation, leads to a transient change in the dielectric properties of the solution. It is highly equipment-intensive and somewhat limited by the requirements of dielectric spectroscopy in apolar media.

10.3 Linear dichroism measurements

229

Stark spectroscopy has been widely employed to study biological chromophores [35–50]. The technique relies on the application of an external electric field to shift both ground and excited state dipoles (electrochromism) to obtain both the magnitude and ___› direction of Δµ k0. To prevent poling of dipolar molecules in the applied field, most studies are done on frozen or immobilized samples.

10.3 Linear dichroism measurements of reduced anionic flavin transition dipole moments and complimentary calculations ____›

The m k0, k = 1–2, for oxidized flavins was determined by linear dichroism (LD) of flavin in stretched polyvinylalcohol (PVA) films [51], or in single crystals of flavodoxin [52]. ____› The use of reduced flavodoxin single crystals also afforded a measurement of m k0 for the neutral semiquinone state. These results have been addressed in an earlier review [8]. In spite of their importance as photobiological chromophores, the transition dipole moments of reduced flavins have not been explored until recently. Both neutral and anionic forms, FlH2 and FlH−, are found in proteins and the cofactor spectra show significant differences that are solvent dependent [53]. For example, FMNH2 in aqueous buffer at room temperature shows a broad band starting at 500 nm with maxima at 395 and 295 nm, whereas in FMNH−, the long wavelength band, beginning at 500 nm, is subsumed into a much broader feature with λmax ~342 nm. The high energy band shifts by –10 nm to 285 nm. In ethanolic solutions of tetraacetylriboflavin (TARF) at 298 K the neutral FlH2 spectrum starts at about 500 nm with a soft shoulder at ~400 nm and peaks at 340 nm and 297 nm. The FlH− spectrum has the same behavior above 400 nm but comes to a slightly sharper peak at ~360 nm with the 296 nm peak being almost twice as intense as the neutral. Ghisla et al. obtained spectra of neutral and anionic TARF in ethanol at 77 K [17]. At low temperature, the 404 nm band sharpens considerably. The anionic form also shows considerable resolution. They surmised that there are three π → π* transitions, at ca. 410, 350, and 290 nm (S0 → Sn n = 1–3). Almost no other experimental data were available to confirm this assignment, with the exception of fluorescence polarization work by Visser’s group in 1979 [53]. Over time, however, this interpretation was apparently discounted or ignored in the literature, and the near-UV/vis spectrum was interpreted as a single π → π* transition with a broad vibronic progression peaking around 400 nm. We measured the transition dipole moment of anionic flavin mononucleotide, FMNH− using UV/vis and IR linear dichroism of FMNH− in stretched PVA films [54]. This technique had been used previously by Johansson [55] and Norden [51] for oxidized flavin. IR dichroism was used to determine the degree of orientation of the FMNH− molecules in the stretched PVA films. The isotropic absorption and linear dichroism spectra as shown in 򐂰Fig. 10.1. The peak in the reduced LD at 365 nm and the plateau at 420 nm suggested two difference electronic transitions, not one. These data provided unequivocal evidence that the near-UV spectrum, previously considered to be a broad S0 → S1 band, is composed of two separate transitions. Ironically, most current laser studies of reduced flavins use Ti:sapphire lasers, whose doubled output straddles these S0 → S1 and S0 → S2 absorptions. Even though this study was published in 2007 (and the Ghisla paper appeared in 1974! [17]) many transient absorption experiments on reduced flavins have

230

10 Excited state electronic structure of flavins and flavoproteins

9

1.2

9a

8

R N

10a

10

1.0

5 7

5a 6

N

LDr & Abs (OD)

0.8

4a

N

1 2 3 4 NH

O m290 m440 m350

O

0.6 0.4

Abs LDr

0.2 0.0 300

350

400

450

500

550

600

Wavelength (nm)

Fig. 10.1: LD of FMNH− with normalized absorption spectrum shown for comparison. The inset shows the directions of the lowest bright π → π* transitions.

ignored this fundamental aspect of excited state electronic structure and their analyses must be thrown into question as a result.

10.4 Experimental studies of excited state electronic structure of flavins and complementary calculations 10.4.1 Oxidized flavin Earlier determinations of the electronic structure of flavins in the oxidized state have been reviewed by Stanley [8]. They consist of solvatochromic, time-resolved molecule relaxation, and___time-resolved microwave conductivity experiments [31,56]. These tech› niques give | Δµ 10 | ~ 1 D. In the late 1990s we constructed a Stark spectrometer with the goal of elucidating the magnitude and direction of excited state dipole moments of flavins in simple solvents as well as those serving as protein cofactors. Stark spectroscopy [32] uses an external electric ____›light to___›obtain ___› ___› field and polarized both the magnitude and direction of Δµ k0 so that µ k may be evaluated. m 0k and µ 0 must be obtained by other methods. Using ground state quantum calculations from other ___› groups (as well as our own), ____› reasonable values for the direction and magnitude of µ ___›0 µk can be derived. Using the m k0 from other studies it ____›is possible ____›determine ___› to uniquely ___› in the visible and near-UV. The angle between m and Δ µ , ζ ≡ ∠ m , Δ µ , prok0 k0 k0 k0 k0 ___› vides the direction of µ k in the molecular frame. An analysis of the data also allows an __ estimation of the average change in polarizability between the states, TrΔa k0, which can be thought ___of as the change in polarizability along the axis of the transition dipole __ › moment. The Δµ k0 and TrΔa k0 are scaled by f, the local field correction factor [57]. This correction factor depends on the dielectric properties of the solvent and the shape of the solute molecule. f would be 1.0 for vacuum and is about 1.7 for butanol (BuOH). Determination of the value for f in a protein is complicated.

10.4 Oxidized flavin

___›

231

Our initial report on Δµ k0 for 3 mM solutions of FMN and FAD in glycerol/water glasses and for N(3)-methyl, ___N(10)-isobutyl-7-8-dimethylisoalloxazine (“N3-MeFl”) in ___› › µ ~ 1.5 D with ζ ≅ 70° and Δ µ ~ 2.7 D with ζ 20 ≅ 50° butanol gave (for N3-MeFl) Δ 10 __ 10 __ 20 __ __ 3 3 [58]. Here TrΔa 10 ≈ 5 Å ___and TrΔa 20 ≈ 55 Å , for f = 1.7. The data agreed with previ› ous ___determinations for Δµ 10 from other methods (as cited___above). ___› No determinations › › seemed__ at least of Δµ 20 had been previously available, and the ratio of Δµ 20/Δµ 10 ~ 2 __ __ __ in qualitative agreement with solvatochromism data. The ratio of TrΔa 20/TrΔa 10 ≈ 10 showed that the S2 state is much more polarizable than the S1 state. However, in the end these experiments were plagued by a systematic ___error in the apparatus. › In 2001 we reported revised and accurate Δµ k0 (k = 1–2) for N3-MeFl in ethanol (EtOH), butanol, and 3-methyltetrahydrofuran (3-MTHF) frozen glasses at 77 K by Stark spectroscopy [59]. Typical data from that study are shown in 򐂰Fig. 10.2. The top panel in 򐂰Fig. 10.2 shows the 77 K absorption spectrum of N3-MeFl in BuOH, along with the Stark spectra taken at two different probe polarizations in the bottom panel. What should be immediately apparent is that the electric field effect on the 450 nm transitions (S0 → S1) is much smaller than the field effect on the 370 nm band (S0 → S2), even though the extinction of the 450 nm band is about twice that of the 370 nm transition. This clearly indicates how differently charge is redistributed in S1 and S2 at the moment of photoexcitation, a critical piece of information for understanding the subsequent evolution of the excited state. The fact that the Stark spectra look like the 2nd derivative of the absorption spectrum is an indication that charge redistribution is dominated by a change in dipole moment. A large change in polarizability would appear as a 1st derivative feature, and some of that is apparent in the S0 → S2 transition. Because of improvements in the spectrometer we were able to make measurements at lower flavin concentration, between 800–900 μM, ruling out issues of aggregation. ___› However, the most important difference was that the Δµ 20___ was larger than our earlier › work suggested. For example, N3-MeFl in BuOH gave Δµ 10~ 1.5 D ___with___ζ10 ≅ 63°, in ___› › › /Δµ ~ 3, 50% line with our earlier values, but Δµ 20 ~ 4.3 D with ζ 20 ≅ 15° leads to Δµ 20__ __ 10 3 higher than our original estimate. The polarizability changes were TrΔa 10 ≈ 6 Å and ___› __ __ __ 3 TrΔa 20 ≈ 52 Å with TrΔa 20/TrΔa 10 ≈ 9. Importantly, Δµ k0 measured as a function of

0.6

500

450

400

N(3)-Flavin (0.8 mM) in 1-Butanol

0.4 De/n (M1 ) at 106 V/cm Abs. (norm.)

350 nm

0.2 0.0 2x104

c  58 deg. c  90 deg.

0

2x104 20000 22000 24000 26000 28000 30000 32000 Wavenumber (cm1)

Fig. 10.2: Low temperature (top) and Stark (bottom) spectra of N3-MeFl in BuOH glass.

232

10 Excited state electronic structure of flavins and flavoproteins

solvent polarity were relatively invariant, suggesting that solvent polarization did not produce a measurable (or differential) effect on charge redistribution at the dipolar __ level. This is in contrast to TrΔa k0, which showed an increase with ___decreasing ___› dielectric › constant. It is worth noting that assignments about the direction of µ 1 and µ 2 were made on the basis of semi-empirical calculations on gas phase lumiflavin. The state of the art in computing ground state dipole moments for flavins in physically reasonable solvents has progressed significantly since 2001 (see below) and our original assignments are in need of further refinement.

10.4.2 Excited state structure of OYE and OYE charge transfer complex Having explored the Stark spectroscopy of flavins in simple solvents we turn to oxidized flavoproteins. A central question in this area is how the protein tunes the electronic properties of the flavin to effect catalysis, a topic explored in model systems as discussed below. There are literally hundreds of studies exploring this concept for ground state electron transfer, but for photobiological systems it will be the excited state electronic structure of the flavin, and its interaction with the protein, that modulates its function. We have explored this idea initially with a non-photobiological protein, old yellow enzyme (OYE). OYE uses FMN as a cofactor for electron transfer [60]. The oxidized cofactor has a higher structure absorption spectrum, making Stark measurements more accurate. Further, the FMN cofactor forms charge transfer (CT) complexes with aromatic ligands like p-chlorophenol and p-hydroxybenzaldehyde (pictured in 򐂰Fig. 10.3 with amino acid residues within 4 Å of the N3-H, from PDB 1OYB). In the case of p-Cl-phenol, CT occurs from the ligand to the oxidized flavin with a pronounced and characteristic new red band appearing in the absorption spectrum, making it possible to monitor the influence of the protein matrix on the OYE-only Stark spectrum, and determine the characteristics of the CT complex in the OYE-p-Cl-phenolate system [41]. The results are shown in 򐂰Fig. 10.4. The differences in the absorption spectra can be seen in the top panels for OYE (left) and OYE+ligand (right) at both 298 K (RT) and 77 K (LT). The CT transition peaks at 15,000 cm−1, or about 670 nm. Stark spectra at two different probe polarizations are shown in the bottom panels and the ___› difference__between unliganded and liganded OYE is dramatic. An analysis of the Δµ k0 and TrΔa k0 for FMN in OYE were identical to those obtained for N3-MeFl in

rDA = 3.6 Å

Fig. 10.3: OYE with p-HBA CT ligand (1OYB).

10.4 Oxidized flavin A

900 12500

750

600

233

450 nm

OYE

10000 7500

77 K 298 K

5000

e (M1 cm1)

Band II (S0  S2)

Band I (S0  S1)

2500 0 12500 OYE: p-Cl-phenol

10000 7500

CT Band

5000 2500 0 12000 14000 16000 18000 20000 22000 24000 26000 28000 30000 Energy (cm1) 900

B

750

600

450 nm

0.004

Band I (S0  S1)

0.002

Band II (S0  S2)

e/n (M1 ) at 106 V/cm

0.000 0.002 0.004 0.00033

c  50 ° c  90°

OYE

c  51 ° c  90°

0.00000

0.00033

OYE: p-Cl-phenol 12000 14000 16000 18000 20000 22000 24000 26000 28000 30000 Energy (cm1)

Fig. 10.4: OYE without and with p-Cl-phenolate ligand absorption spectra at 298 K and 77 K (top), and Stark spectra (bottom).

nonpolar solvents (this was also the case for FADOX in E. coli photolyase [35]). A great uncertainty___ here is how __ to calculate f, the local field correction, for the protein, but the › ratios of Δµ k0 and TrΔa k0 make for useful comparisons with the simple solvent results. Thus, the Stark spectrum may contain evidence of cofactor polarization by the nearby amino acids and the cofactor binding site electric field, but this is not evident at our level of analysis.

234

10 Excited state electronic structure of flavins and flavoproteins

___›

The Stark spectra of the CT complex is illuminating. The CT band has Δµ CT ≈ 12 D and ζ A ≈ 0°, giving a quantitative measure of the degree of CT. The angle of 0° is particularly gratifying as it matches the expectation for Mulliken CT where the CT dipole lies along the CT transition dipole moment, typically pointing between donor and acceptor mole3 == cules [61]. The TrΔa CT ≈ 370 Å is also significantly larger than for the flavin cofactor alone, again suggesting that excited state electron density is significantly more polarizable (and thus easier to transfer) than charge distributed in localized MOs on the flavin. ___›

10.4.3 DNA photolyase and Δµ k0 The relevance of these measurements is brought into focus in elucidating the photoinduced electron transfer (PET) mediated repair of cyclobutylpyrimidine dimers by DNA photolyase. To place this into the proper context we provide here a short summary of what has been learned about this protein, for which many excellent reviews are available [62–65]. The ca. 50 kD monomeric protein binds FAD non-covalently as its catalytic cofactor. The oxidation state of the flavin shuttles between FADH− and FADH• during the repair process. A second chromophore, a folate or deazaflavin, acts as a photo antenna to transfer light energy to FADH• to reduce it to its catalytically active FADH− state. Several groups have examined this energy transfer process [66–74] but some believe it is not physiologically relevant [75]. The structure of the active site [76] is shown in 򐂰Fig. 10.5. Once the substrate is bound, blue light absorption by FADH− leads to ultrafast PET. After earlier attempts [77–81] had failed to resolve intermediates in the repair mechanism, MacFarlane and Stanley [82] performed subpicosecond blue/UV pump/probe measurements on the Anacystis nidulans protein and dT2- and dT5-based cyclobutane pyrimidine dimer (CPD) substrates. Overexpressed A. nidulans DNA photolyase (PL) does not contain the deazaflavin antenna because the host cell, E. coli, does not synthesize it. The PET lifetime was clocked at 32 ps with immediate repair of one of the two cyclobutane bonds following PET to the CPD. The second CPD bond scission was not resolved directly but estimated to occur between 100–600 ps after light absorption. This model was confirmed in a highly detailed blue/UV pump/probe study of dimeric substrates

CPD Ado

Flavin Trp390

Fig. 10.5: Active site of A. nidulans PL with CPD and Trp390 (from PDB 1TEZ).

10.4 Oxidized flavin

235

with A. nidulans PL by the Zhong group in 2011 [83]. One interesting addition to the repair mechanism gleaned from this work is that the PET lifetime is substrate dependent, with τUT < τUU < τTU λ). However, when the donor-acceptor separation becomes less than 10 Å, the ET dynamics can occur in the range of femtoseconds to hundreds of picoseconds, which are on the similar time scales of local environment relaxation of solvent and the protein [12,68,69]. Under these circumstances, the ET dynamics are in nonequilibrium and the processes follow non-exponential dynamical behaviors [12,69,70]. Usually, the extended Sumi-Marcus two-dimensional model [71] with quantum correction [72] can be used to simulate experimental results, providing molecular insight into short-range protein ET. Specific vibrational modes have been observed to greatly enhance ET dynamics in chemical systems [72,73], effectively reducing the ET barrier, but such a quantum effect has not been directly observed in protein ET. Furthermore, when the donor and acceptor are in close proximity, e.g. within van der Waals contact, the ET dynamics could be even faster than the local relaxation, leading to a “frozen” environment with a heterogeneous distribution of electrostatics. Flavoproteins are ideal model systems for studies of intraprotein ET at such short distances [26,74–81]. In flavoproteins, oxidized flavin, FAD or FMN, can be photoreduced to anionic radical flavin, FAD•− or FMN•−, through intraprotein ET from neighboring aromatic residues (tryptophan or tyrosine) in hundreds of femtoseconds to a few picoseconds [26,74,75]. Such ET has been observed in many flavoproteins, including flavodoxin [73], riboflavin-binding protein [74], DNA photolyase [78,80–82] and cryptochrome [26,83]. In this section, we report our recent study on the ET dynamics in a model system of flavodoxin [84]. By extensive characterization of these ET dynamics for twelve mutant flavodoxins, we have observed ultrafast nonequilibrium ET dynamics for both forward (FET) and backward (BET), modulated by different reduction potentials. Significantly, we are able to directly observe the vibrationally enhanced BET process and subsequent vibrational cooling dynamics at the active site of the protein.

17.3.1 Experiment design, reaction scheme and probing strategy Flavodoxin functions as an electron shuttle through its cofactor of FMN [85]. Structurally, flavodoxin consists of five-parallel β-strands surrounded by several α-helices as shown in 򐂰Fig. 17.5 (left). The prosthetic FMN group is noncovalently but tightly bound by a series of interactions with protein residues [86]. One of the key interactions is the π-π stacking of the isoalloxazine ring with Y98 and W60, both of which are within van der Waals contact of the isoalloxazine ring (򐂰Fig.17.5, right), forming a sandwich configuration. Such interactions exclude water from the binding pocket and only the o-xylene ring of the flavin is exposed to solvent. We have recently characterized the

402

17 Ultrafast dynamics of flavins and flavoproteins A

B

Y98 D95 3.34 Å

G61

3.46 Å

W60

Fig. 17.5: (A) X-ray crystallographic structure of oxidized D. vulgaris flavodoxin (PDB code 2FX2). (B) A close-up view of the local configuration at the FMN-binding site. The FMN cofactor (in yellow) is sandwiched between two neighboring aromatic residues W60 (in green) and Y98 (in orange) at the van der Waals contact. Also shown are the acidic residue D95 and the backbone carbonyl group of G61 in close proximity of FMN.

dynamics of the active-site solvation [87] and observe local relaxation on wide time scales from a few, to tens and to hundreds of picoseconds. With well-characterized active-site relaxation dynamics, we can quantitatively examine the short-range nonequilibrium ET dynamics in flavodoxin. Here, with perturbation of the flavin environment by site-directed mutagenesis in conjunction with femtosecond-resolution data acquisition, we have systematically studied eleven mutants to evaluate the short-rang ET dynamics with different redox energies, including two mutations of neighboring residues of G61 and D95 which have been identified as influencing the reduction potentials of FMN [88,89]. These mutants and the wild-type protein can be classified into four cases (򐂰Fig. 17.5, right): (i) Y98 as the only ET donor with mutations of W60A and W60F; (ii) W60 as the only ET donor with Y98F, Y98A, Y98H and Y98R; (iii) two identical ET donors of Y60/Y98 and W60/W98 with mutations of W60Y and Y98W, respectively; and (iv) both W60 and Y98 as the ET donors with the wild type and G61V, G61A, D95N mutants. For twelve proteins with four different sets of ET reactions, we can summarize the kinetics as follows: k1 FET W k2

Y

Y

FMN* 

FMN 

k3 BET W k4

Y FMN†  W

k5 cooling

Y

(a)

W

(b)

FMN 

For cases (i) and (ii), the proteins only contain one ET donor and thus the ET dynamics correspond to reaction (a) and (b), respectively. For cases (iii) and (iv), both reactions (a) and (b) occur, but for the former the two donors are either the same Y or W in the reactions. In 򐂰Fig. 17.6, we show all the visible absorption spectra for various species involved in these reactions. 򐂰Fig. 17.6B shows two typical absorption spectra of wild-type flavodoxin and Y98W mutant with the peaks of S1←S0 (450 nm) and S2←S0 (380 nm) transitions of FMN in the binding pocket. For Y98W, the absorption extends to more than

403

800

740

630

410

lpr (nm)

700

Normalized DA

A

480 500 510 515 520 530 540 580

17.3 Electron transfer in model flavodoxin

Flavin* Flavin• Indole Phenol

1

5 lpu

WT Y98W

500

lpr (nm)

550

630

580

Probing FMN†

510 515 520 530 540

e (x103M1cm1)

Flavodoxin 10

500

0 B

600

0 400

500

600

700

800

Wavelength (nm)

Fig. 17.6: (A) Normalized (transient) absorption spectra of protein-bound FMN* (red) and anionic flavin semiquinone (dark yellow), and free cationic radicals of indole (blue) and phenol (dark purple) in organic solvents. The various probe wavelengths for the transient absorption experiments are marked on the top. (B) Steady-state absorption spectra of wild-type (blue) and mutant Y98W (red) flavodoxin. The pump wavelength at 400 nm and multiple probe wavelengths (inset) for probing the vibrationally hot FMN molecules are marked by a series of arrows.

650 nm, suggesting charge-transfer character due to electronic delocalization of FMN through stacking with the two aromatic tryptophan residues of W60 and W98 [90]. If high-frequency vibrational modes are involved in the BET reactions and thus enhance the ET rates, the hot FMN† could be detected at the red edge of the ground-state absorption (򐂰Fig. 17.6B, insert). To directly observe such a quantum effect, i.e., the formation of FMN†, the ET reaction rates, k1 and k3 in (a) or k2 and k4 in (b), must be faster than the cooling relaxation (k5). Otherwise, the FMN† could not be accumulated due to the rate-determining slow ET dynamics. Thus, it is necessary to resolve all elementary steps by monitoring the dynamics of each intermediate involved in the reactions. 򐂰Fig. 17.6A shows the normalized absorption spectra of three intermediate radicals and the excited FMN [91–93]. We can tune different wavelengths to follow the whole dynamics of reactants (FMN*), intermediates (FMN•−, W+ and Y+) and products (FMN† and FMN) and thus map out the entire ET cycle and hot vibrational cooling, if any, in the protein.

17.3.2 Femtosecond charge separation, frozen active-site configuration and critical free energies We observe the fluorescence transients gated at the emission peak (538 nm) for the wild type and mutants. Except for Y98W, for all other mutants the fluorescence transients show dynamics in the 135–340 fs range [84]. We do not observe any fluorescence dynamics for Y98W due to the nature of charge-transfer excitation. Except for W60A, fluorescence transients for the other mutants show single-exponential behavior.

404

17 Ultrafast dynamics of flavins and flavoproteins

Specifically for case (i), we obtain 302 fs (k1−1) for W60F; for case (ii), we get 258, 247, 204, and 193 fs (k2−1) for Y98F, Y98A, Y98H and Y98R, respectively; for case (iii), we have 258 fs for W60Y; and for case (iv), we find 157, 154, 149 and 135 fs for the wildtype, G61V, G61A, and D95N, respectively. For W60A, we can fit a double-exponential decay with 340 fs with the amplitude of 96% and 1.82 ps (4%) or a stretched exponential decay with 345 fs and a stretched parameter of 0.9 [12,69,70,94]. We note that the W60A mutant is less stable than W60F because the aromatic residue W60 behaves as a critical gate for entering into the binding pocket [95]. Thus, the dominant 340-fs component represents the major ET dynamics and the minor 1.82-ps component probably reflects a loose, unstable configuration. The observed dynamics of charge separation are ultrafast, in the 135–340 fs range, and our reported active-site relaxation takes longer than 1 ps. Thus, upon initial excitation the ET is much faster than relaxation of the local environment. The ET reaction thus occurs in a nonequilibrium configuration with the active site nearly frozen on the time scale of ET. The concept of reorganization energy (λ) may not be applicable here and the forward ET reaction mainly depends on the reduction potentials (mainly enthalpy difference) of the donor and acceptor, the driving force of the reaction, similar to the ET reaction in gas phase determined by the donor ionization energy and acceptor electron affinity. Overall, the observed time scales follow the trend of driving-force changes of the FET reactions, but for the fixed donor(s) the FET dynamics have minor variations with the mutants. Thus, the change of the FET dynamics is mainly from the difference (320 meV) of the reduction potentials of W and Y. The reduction of FMN* to FMN•− depends on the total available aromatic residues (W and/or Y) in proximity.

17.3.3 Ultrafast charge recombination, vibrational quantum effect and hot ground-state cooling

򐂰Fig. 17.7 shows the typical absorption transients of four ET cases with a wide range of

detection wavelengths from 800 to 400 nm to monitor the dynamics from the reactants (FMN*), to charge-separated intermediates (FMN•−, Y•+) and final products (FMN† and FMN) of Y98 (W60F mutant) in case (i). 򐂰Fig. 17.7A shows the entire evolution of the transients. Similar transients are observed at 800 and 740 nm and we only detect the excited reactant FMN*. The dynamics follow a single exponential decay in 0.49 ps, slower than that from the fluorescence detection (0.3 ps), which seems a general phenomenon [96], i.e., the absorption may sense more donor-acceptor configurations than the fluorescence detection. When the probe wavelength is tuned to the blue side, the dynamics clearly become slower (򐂰Fig. 17.7A). In particular, when the dynamics were probed from 540 to 500 nm on the red side of the ground-state absorption, the transients become slower. Clearly, if only the excited FMN* and intermediates FMN•−/Y•+ are detected, the dynamics should not change dramatically even though the percentage of each species could change. Thus, the only possibility is that we observe the hot ground state FMN† after charge recombination (򐂰Fig. 17.7B); such vibrationally excited states have absorption to the red of the original ground-state absorption. Therefore, we see a series of hot ground-state relaxations and the dynamics become slower with vibrational cooling from higher- to lower-energy states. At 480 nm, we observe recovery of the ground-state and the dynamics become faster again due to cancellation of slow-cooling contributions by ground-state formation signals (򐂰Fig. 17.7C). At

17.3 Electron transfer in model flavodoxin

lpr

405

Y98 (mutant W60F) 1

lpr A 800 nm 740 nm 540 nm 530 nm 520 nm 515 nm

800 nm

740 nm 0 540 nm

510 nm 500 nm 480 nm 410 nm

X 0.5 530 nm

Normalized DA

520 nm

0

5

10

B FMN*  FMN•  Y FMN†

1

515 nm

lpr  510 nm 0 0

510 nm

5

10

C FMN* FMN•  Y FMN† FMN

1 500 nm 0 480 nm 1

410 nm 0

5

lpr  480 nm

X 0.5 0

5

10

15

10

20

Delay time (ps)

Fig. 17.7: Normalized femtosecond-resolved absorption transients of mutant W60F flavodoxin probed from 800 to 410 nm for the ET reaction with only donor Y98. Inset (A) shows the entire evolution of the dynamics with the probe wavelengths. Insets (B) and (C) show the deconvolution of the transients into various species probed at 510 and 480 nm, respectively. Note that at 480 nm the transient becomes faster again as a result of cancellation of ground-state FMN recovery (light-orange line) by the positive cooling signal of FMN† (green line).

406

17 Ultrafast dynamics of flavins and flavoproteins

410 nm, we observe a positive signal again from contributions of the excited FMN† and the intermediates of FMN•− and Y•+. Thus, the global fitting of all these transients gives 0.95 ps (k3−1) for charge recombination and 3.7–4.0 ps (k5−1) for vibrational relaxation of hot ground states. The observation of ground-state vibrational excitation after BET in flavodoxin is significant. Vibrationally coupled ET through high-frequency modes is seen in BET of several chemical systems [72,73,97]. For flavodoxin, the BET dynamics are even faster than the active-site relaxation and thus the active-site motions are only partially coupled with the BET dynamics. With a van der Waals distance of 3.34 Å between the donor tyrosine and acceptor isoalloxazine ring [86], the FMN•−…Y•+ ion pair formed here would proceed to vibrational motions as observed in gas-phase benzene/iodine charge-transfer complex [98]. The ultrafast BET, faster than active-site relaxation, must efficiently channel ionic energy into high-frequency vibrational energy of the ground-state FMN and/or Y, leading to the formation of hot FMN†, a molecular picture similar to hot ground-state benzene formation in BET of benzene/iodine charge-transfer complex [98]. Such ultrafast ET reactions in flavodoxin bear some similarity with gas-phase bimolecular chargetransfer reactions [98,99]. If we consider the absorption shift from 500 to 540 nm due to hot ground-state formation, a total vibration energy of ~1,500 cm−1 is obtained for hot FMN† with a few vibrational quantum numbers (ν = 1–3), given some high-frequency modes of FMN in 500–1,500 cm−1 [100]. The cooling dynamics occur in a few picoseconds, slower than all ET dynamics, and thus enable us to observe formation of the hot FMN† and follow its subsequent cooling dynamics. The time scale in 3–4 picoseconds is similar to that of the FMN† cooling in polar solvents [49,52], implying that vibrational energy may flow into the water molecules around the entrance of the active site. Such cooling dynamics is also mixed with the active-site relaxation induced by fast charge recombination. In ET with W donor and dual donor of W and Y, we observe the similar pattern of the striking vibrational excitation and subsequent cooling dynamics.

17.3.4 Photoinduced redox cycle, reaction time scales, and vibrational coupling generality The photoinduced ET cycle at the active site of flavodoxin is given in 򐂰Fig. 17.8A. Four fundamental processes are involved in the redox cycle: Forward ET (τFET), subsequent back ET (τBET), following vibrational excitation and cooling (τC), continuous active-site solvation (τS). The active-site relaxation evolves with the entire dynamic process from the initial excitation, to charge separation, to charge recombination and finally to vibrational cooling. The complete solvation at the active site occurs on multiple time scales in 1.0 ps (53%), 25 ps (26%) and 670 ps (21%), involving various motions at the active site including neighboring hydration water molecules [87]. The 1.0-ps dynamics mainly come from the initial local relaxation of hydration water networks around the active site at the protein surface [87,101]. Thus, the four time scales involved are critical to our observation and the understanding of the molecular mechanism of ET. For flavodoxin, τFET is ultrafast and the solvation appears frozen, a gas-phase type of bimolecular charge-transfer reactions. τBET is also very fast on the similar time scale of initial solvation. Thus, the BET is a nonequilibrium ET dynamics and theoretically could be treated by the extended Sumi-Marcus two-dimensional model (solvent and nuclear coordinates), leading to a nonexponential behavior typically in a stretched exponential

17.3 Electron transfer in model flavodoxin A

B

For wa rd el 13 ect 53

hn

407 Excited state

Ionic state

fer ns tra n ro fs 40

Ground state

Y98

ps tr a ns fe

r

.3 -4 a 2.5 vibr te

sta n d-

ps tio na l

co

olin g

B

1.6 n 50 . 9 c tr o ele ack

Intra mole cular disto rtion (q)

(X ) reo rga niz ati on

hn

So lve nt

G ro u

FMN W60

Fig. 17.8: (A) The complete photoinduced redox cycle in flavodoxin with all resolved elementary dynamics and their time scales. The active-site relaxation (not shown) is involved in all these processes on the picosecond time scales. (B) Schematic presentation of the three potential surfaces involved in the photoinduced redox cycle along two coordinates, intramolecular distortion and solvent reorganization. The three dots represent the equilibrium configurations on the three surfaces. The high-frequency ground-state vibrational excitation is not shown during the transition from the ionic state to the ground state.

decay. However, such BET dynamics partially overlap with the following cooling dynamics and thus we simply treat the BET as an exponential decay without any stretched behavior. Due to the stacking nature of donor and acceptor and the hydrating water molecules around the entrance of the active site and not intercalating between the donor and acceptor, the charge-separated complex (FMN•−…Y•+/W•+) could involve significant intermolecular vibrations and thus lead to high-frequency vibrational excitation of the products FMN†…Y†/W†, resulting in very fast BET dynamics. However, if τFET and τBET are much longer than τS, we would observe a single-exponential decay for both equilibrium ET dynamics with a relaxed actives-site configure. If τFET or τBET is longer than τC, we would not observe the vibrational-cooling process because the rate-determining step is the longer ET dynamics and the hot products cannot be accumulated. Thus, flavodoxin is an ideal model system for direct observation of quantum vibrational effects on acceleration of BET dynamics with all four elementary processes resolved in an ET cycle in the active-site environment of the protein. With the eleven mutants and wild-type flavodoxin, we have resolved all three elementary dynamics of FET, BET and vibrational cooling. With the reported reduction potentials of FMN in these mutants, the ET dynamics mainly follow the reduction potential trend of the donor W or Y but slightly change with that of FMN upon mutation. For the stacked, compact FMN…Y/W pair, the ET reaction occurs mainly between the donor and acceptor. The environment has a principally static effect but does have a minor influence on the dynamics of the system. 򐂰Fig. 17.8B shows the schematic potential energy surfaces of three states along two coordinates of intramolecular distortion (q) and solvent reorganization (X). The initial FET mainly evolves along the intramolecular nuclear coordinate. After charge separation, the ionic complex continuously

408

17 Ultrafast dynamics of flavins and flavoproteins

moves along the nuclear coordinate and meanwhile also evolves along the solvation coordinate to reach the minimum crossing seam to the ground-state surface with highfrequency vibrational excitation (not shown in 򐂰Fig. 17.8B) to conserve the total energy. The molecular picture revealed here should be applicable to many other flavoproteins due to the common structural configurations and interactions between aromatic residues and cofactor flavin. Thus, the vibrational excitation after charge recombination and subsequent cooling dynamics should be general to flavoproteins even though the cooling dynamics could be often buried in the rate-determining slow ET process.

17.4 Enzymatic reactions and repair photocycles in DNA photolyases DNA photolyases are photoenzymes that repairs ultraviolet (UV)-damaged DNA under blue-light irradiation. There are two major types of UV-damage to DNA, cyclobutane pyrimidine dimer (CPD) and (6-4) photoproduct (64PP), which are repaired by CPD photolyase and (6-4) photolyase, respectively. Our ultrafast spectroscopic studies with molecular biology [12,69,70,94] have elucidated the complete molecular mechanisms of the DNA repair by mapping out the entire functional dynamics as described below.

17.4.1 Dynamics and mechanism of cyclobutane pyrimidine dimer repair by CPD photolyase CPD photolyase contains a fully reduced flavin adenine dinucleotide (FADH−) as the catalytic cofactor and electron donor [11]. Based on previous studies [11,12,102–106], a sequential repair mechanism of thymine dimer splitting has been proposed, as shown in 򐂰Fig. 17.9. Our initial study has shown that the forward ET from FADH−* to thymine dimer (TT) occurs in 250 ps (1/kFET) and the total decay of intermediate FADH• in 700 ps (1/ktotal) [12,103]. These dynamics usually follow a stretched-exponential decay behavior, reflecting heterogeneous ET dynamics controlled by the active-site solvation [12,68,102]. However, in our earlier study, we focused on detection of the cofactor flavin dynamics in the visible region and thus no detailed information about the dimer splitting was obtained. To reveal how the thymine dimer splits, we have recently extended our detection wavelengths from visible to deep UV light to catch thymine-related intermediates. To discover how the electron tunnels in the repair, we have used different dimer substrates to lay out electron tunneling pathways.

17.4.1.1 Sequential splitting dynamics of the cyclobutane ring We observe a striking pattern of the transient absorption signals of the complex of E. coli CPD photolyase with substrate TT, probed at fifteen wavelengths [70,94]. At 430 nm, the signal is the summation of all three flavin species (FADH−*, FADH• and FADH−) and decays to zero upon completion of repair. From 335 nm in the UV region, we have captured the formation and decay of thymine-related intermediates and from 300 nm we clearly observe the long-component formation of final repaired thymines. By knowing the dynamics of FADH−* and the absorption coefficients of FADH• and FADH−, only with the sequential model shown in 򐂰Fig. 17.9 (and not any other synchronously or asynchronously concerted schemes of thymine splitting and electron return

17.4 Enzymatic reactions and repair photocycles in DNA photolyases

409

Fig. 17.9: Enzyme-substrate complex structure and one sequential repair mechanism with all elementary reactions. X-ray complex structure of A. nidulans CPD photolyase with DNA containing a repaired photoproduct of thymine dimer. E. coli CPD photolyase has a similar structure. Two critical conserved residues in the active site, E283 (E274 in E. coli) near the substrate and N386 (N378 in E. coli) near the cofactor, are also shown. The thymine dimer is flipped out of DNA and inserted into the active site. A close-up view shows the relative positions of the catalytic cofactor FADH−, the conserved residues E283 and N386, and the repaired substrate with the electron tunneling pathways in repair. Shown in the sequential repair scheme (bottom panel) are forward electron transfer (FET, reaction rate kFET) from FADH−* to thymine dimer upon light excitation, followed by back electron transfer (BET, reaction rate kBET) without repair, and the repair channel including splitting of two bonds of C5-C5’ (reaction rate ksp1) and C6-C6’ (reaction rate ksp2) in thymine dimer with subsequent electron return (ER, reaction rate kER) after complete ring splitting. ktotal is the overall decay rate of intermediate state FMNH• after the initial charge separation.

[104,105,107]) can we systematically and globally fit all the absorption transients and thus obtain the entire dynamics of thymine dimer splitting. Our data indicate that, in contrast to the reaction schemes proposed in previous computational studies [105,107], the thymine dimer is split by a sequential pathway. We note, however, that we are not able to detect TT−, giving an upper limit of less than 10 ps for cleavage of the first C5-C5’ bond (1/ksp1), consistent with the theoretical prediction of a nearly barrierless process [105,107,108]. The slow formation (kFET) and ultrafast decay (ksp1) result in negligible accumulation of TT− population. However, we do observe the formation and decay of T-T− intermediate after the initial bond cleavage (򐂰Fig. 17.9). The decay dynamics with τ = 87 ps mainly represents cleavage of the second C6-C6’ bond. Given that the total repair quantum yield is 0.82 and the forward

410

17 Ultrafast dynamics of flavins and flavoproteins nsfer tron tra elec d r  a O rw 250 ps O R R Fo

N

5’

3’

O FADH

O

FADH O

R

R

O

HN

tr o

R

O

N

N

5’

3’

•

NH N

N

5’

3’

r Back electron transfe 2.4 ns O

R

NH

nr

R

O

R

O

HN O

O

90 p ttin s g

N

El ec

O HN

NH

O

O

NH N

N

5’

3’

O

p li

3’

O

R

HN O

N

5’

C6

R

N

6–

hn

O

O

ing litt sp 5’ ps – C 10 

FADH *

C5

NH

’s

HN

C

70 0 ps etu rn

Fig. 17.10: Complete photocycle of CPD repair by photolyase. All resolved elementary steps of CPD (thymine dimer) repair with reaction times, showing the complete repair photocycle on the ultrafast time scales and the elucidated molecular mechanism.

ET yield is 0.85 [11,12], the effective splitting yield is 0.96. This results in cleavage of the second bond with τ = 90 ps (1/ksp2), much slower than predicted by theoretical calculations [105,107,108], and the back ET without the second-bond splitting in 2.4 ns (1/kBET). After cleavage of the second bond, we observe the spectral signature of T− at around 290 nm, which decays with τ = 700 ps (1/kER), reflecting decoupling of electron return from T− to FADH• from the second bond-breaking. Thus, the two repaired thymines are formed in two sequential steps with τ = 90 and 700 ps, respectively, upon initial electron injection. The complete photocycle of CPD repair by photolyase is summarized in 򐂰Fig. 17.10.

17.4.1.2 Electron tunneling pathways and functional role of adenine moiety The FADH− cofactor in CPD photolyase has an unusual bent stacked conformation with the isoalloxazine and adenine rings in close proximity (򐂰 Fig. 17.9). The crystal structure of the A. nidulans photolyase in complex with CPD complex shows the adenine moiety of FADH− in van der Waals contact with both base moieties of CPD, 3.1 Å to the 5’ side and 3.2 Å to 3’, and at 3.6 Å to the C-10’ methyl of the isoalloxazine ring (dashed red lines in 򐂰Fig. 17.9) [109]. Intramolecular electron hopping from the isoalloxazine ring to the adenine moiety is unfavorable due to their relative reduction potentials (ΔG ~ +0.1 eV) [110] and we in any case do not observe any fast

17.4 Enzymatic reactions and repair photocycles in DNA photolyases

411

quenching of FADH−* fluorescence in the absence of substrate [12]. It is thought that the repair reaction by CPD photolyase involves electron tunneling [15–17], although the cyclic electron tunneling pathways, forward (kFET) and backward (kBET) or return (kER), are a matter of some debate. One view is that electron tunneling is mediated by the intervening adenine over a total distance of about 8 Å [111]. An alternative model suggests that tunneling occurs directly from the o-xylene ring of FADH− to the 3’ side of CPD over a distance of 4.3 Å [112,113]. To elucidate the specific electron tunneling pathway, we have used a series of substrates, UU, UT, TU and TT (chemical structures in 򐂰Fig. 17.11), as electron acceptors to follow electron tunneling.

lpr 710 nm lpr 620 nm t 250 ps t 700 ps

B 1.0

TT O

O

CH3 CH3

HN O

NH N

N

5’

3’

O

Normalized DA

A

TU

O

t 185 ps

N

N

5’

3’

O

UU O

HN O

t 210 ps

NH N

N

5’

3’

O

0.5 0.0

CH3

O

HN O

t 63 ps

500

1000

1500

NH N

N

5’

3’

Delay time (ps)

2000

TU

0.0 0

O

2500

1000

300

600

T

900

lpr 300 nm

Flavins

300 600 900 1500

1000

TT UT

2000

2500

lpr  335 nm

0.8 0.5

2500

U/T

0.0

0 500

2000

1500

1.0

0.5

Flavins UU U

0.4 UU

0.0 TU

0.0 0

T/U Flavins

1.0 0

UT

500

TT UT UU TU

D 1.0 O

t 1220 ps

0 C 1.0

Normalized DA

t 73 ps

TT UT UU TU

0.0 NH

O

0.5

0.5

O

HN

t 85 ps

Normalized DA (a. u.)

CH3

Normalized DA

O

lpr 270 nm

1.0

0

500

1000

0

300 1500

600 2000

900 2500

Delay time (ps)

Fig. 17.11: Femtosecond-resolved transient absorption dynamics of DNA repairs with different combination of bases. (A) Transient absorption signals of repair with TT, TU, UU and UT probed at 710 and 620 nm. The dynamics of FADH−* (blue line) was probed at 710 nm. The signal at 620 nm is the combination of FADH−* and intermediate FMNH• (red line) contributions. The chemical structures of various CPD substrates are also shown with highlight at uracil sides. The blue shading of U indicates the forward electron tunneling to the 5’ side of DNA and the red shading for electron return starting at the 3’ side after the complete two-bond splitting. (B–D) Repair dynamics of TT (orange), UT (blue), UU (green) and TU (dark red) probed at 270 nm (B), 300 nm (C) and 335 nm (D). Insets show the deconvolution of total flavin-related species (dashed red), CPD intermediate anions T-U−/U-U− (dashed cyan) and T− (dashed dark red), and the products of T/U (dashed dark yellow) of repair with TU, UT and UU in (B), (C) and (D), respectively.

412

17 Ultrafast dynamics of flavins and flavoproteins

The forward ET dynamics, detected at 710 nm (blue lines in 򐂰Fig. 17.11A), occur with τ = 63, 73, 85 and 250 ps for UT, UU, TU and TT. Generally, uracil has a higher reduction potential by 0.1 V than thymine [110], which provides a larger driving force and results in the faster forward ET dynamics. Significantly, the ET rates increase with U at the 5’ position, indicating that the electron tunneling ends at the 5’ side of CPD (򐂰 Fig. 17.9). This observation proves that the adenine moiety mediates forward ET toward the 5’ side of CPD and enhances the ET rate through a superexchange mechanism, ruling out direct electron transfer from FADH− to CPD. This conclusion is further supported by a comparison between the repair of CPD and 64PP. The (6-4) photolyase, which specifically repairs the 64PP, exhibits a similar stacked FADH− configuration and forward ET dynamics (280 ps) as CPD photolyase [69], but the shortest direct distance from the cofactor to 64PP is 6.3 Å, some 2 Å longer than seen with the CPD photolyase [114]. These observations strongly suggest a common mechanism whereby the electron from FADH−* tunnels through the adenine moiety to substrate. We have also done a series of studies by UV detection for these substrates to gain further information on the splitting of the cyclobutane ring and the subsequent electron return process. 򐂰Fig. 17.11B–D shows three typical signals probed at 270, 300 and 335 nm. With systematic analyses, we observe splitting of the second bond with τ = 35 ps with both UU and UT and 75 ps for TU, similar to that of TT. Thus, after the initial electron tunneling event to the 5’ side and the subsequent prompt splitting of the C5-C5’ bond, the resulting radicals are much more stable in TT and TU than UU and UT due to stabilization by the methyl group at the C5 position, thus resulting in a decrease in the rate of the second-bond C6-C6’ breakage by a factor of two. Finally, electron return from the repaired substrates take 185 and 210 ps for (T+U)− and (U+U)−, respectively, and 1220 ps for (U+T)−, with (T+T)− having an intermediate value of 700 ps (򐂰Fig. 17.11A–D). Given that electron return must occur in the inverted Marcus region, the electron from U− returns faster to the FADH• than that from T− back to restore the active state FADH− and complete the repair photocycle. Clearly, after repair the electron moves to the 3’ side, so that T+U− and U+U− have faster back electron tunneling rates than U+T−. Thus, from the forward ET dynamics as well as the dynamics of cleavage of the second bond with these substrates, the forward electron tunnels from the isoalloxazine ring to the first carbon atom linked to the ring through a covalent bond (1.5 Å), then to the adenine moiety, and finally to the 5’ side of CPD at a total distance of 8.2 Å, rather than taking the shortest though-space distance of 4.3 Å. After the complete cleavage of the two C-C bonds of substrate, the electron remains at the 3’ side and tunnels back along the original adenine-mediated pathway (򐂰Fig. 17.9). The electron tunneling, both forward and return, thus has unique directionality and the adenine moiety plays a critical functional role. The reaction times of all elementary steps are shown in 򐂰Fig. 17.12A and the reaction energy profiles along the reaction coordinate are given in 򐂰Fig. 17.12B. According to the transition-state theory, we estimate an activation energy of ~0.174 eV (4.02 kcal/mol) for T−-T bond breaking, ~0.170 eV (3.93 kcal/mol) for T−-U bond splitting, and ~0.152 eV (3.51 kcal/mol) for U−-T and U−-U bond cleavage (򐂰Fig. 17.12B). The radical stabilization in T−-T by the methyl group(s) at the C5 (and C5’) position thus leads to ~0.022 eV (0.51 kcal/mol) greater activation of the C6-C6’ bond splitting than in U−-U. On the other hand, before the C6-C6’ splitting, the futile back ET

17.4 Enzymatic reactions and repair photocycles in DNA photolyases A

413



T–T Reaction time (ps)



U+ T

10 3



T+T



T –U TT 10 2





U– T  U–U

U+U T +U 

TU UU UT

T–T  T –U 

U– T  U–U

101 FET B

10 ps

Anionic surface  – U– U T T U– T – T U – UU – UT (normal DG ET region) DG (inverted ET region)

SP2 90 ps

ER 700 ps

DG 

3.5 kcal/mol 4.0 kcal/mol

T – U T–T

Rin gr ec los ur e

Free energy

Dynamics

BET



U+ T T+T



U+U T +U

DG (inverted ET region)

TT Neutral surface

T+T

Reaction coordinate

Fig. 17.12: Reaction times and free energy diagrams of the elementary steps in splitting of CPD substrates. (A) Reaction times of each elementary step observed in repair of various substrates (TT, TU, UU and UT), including the forward ET, the C6-C6’ bond splitting, futile back ET and final electron return to complete the photocycle. Note the order in reaction times of three ET processes for four different substrates and also the faster bond splitting than the back ET and electron return. (B) The free energy profiles along the reaction coordinate after the forward ET in repair of CPD substrates with the time scales of the dynamics shown at the top. On the anionic surface, the solid curve represents the splitting of TT− while the dashed curve for UU−. On the neutral surface, the bond-breaking activation barrier (dashed curve) is very high according to theoretical calculations. Note the different regions, normal or inverted, of three ET processes and the ring reclosure after the futile back ET.

results in the neutral intermediates to reclose the ring by formation of the C5-C5’ bond again and return to the original ground state (򐂰Fig. 17.12B). The splitting of C6-C6’ bond in the neutral ground state after back ET is probably insignificant. The splitting of the CPD lesion on the neutral surface (򐂰Fig. 17.12B) is unlikely and the recent theoretical studies have shown that a free energy greater than 1.7 eV is required to

414

17 Ultrafast dynamics of flavins and flavoproteins

lengthen and break the C5-C5’ bond [105,115]. Thus, the free energy for back ET after the C5-C5’ breaking in T−-T must be around -0.34 eV (–1.96 eV + 1.7 eV – 0.08 eV) (򐂰Fig. 17.12B). Assuming the reorganization energy to be similar as in the forward ET, 1.21 eV, the back ET thus switches to the normal Marcus region (򐂰Fig. 17.12B). Although thymine has a reduction potential more negative than uracil by 0.11 V, which would lead to a larger driving force and a faster back ET, it seems that the methyl group at C5/C5’ position significantly stabilizes the anion intermediate, resulting in a lower driving force and strategic slowing of the back ET. With T at the 5’ side, the stabilization energy is about 0.15 eV by comparison of T−-U vs. U−-U, while T at the 3’ side the stabilization energy is about 0.02 eV of U−-T vs. U−-U, indicating that the electron is mainly localized at the 5’ side (with only a partial distribution on the 3’ side) so that the methyl group at the C5’ position has some stabilization effect.

17.4.2 Dynamics and mechanism of repair of UV-induced (6-4) photoproduct by (6-4) photolyase The (6-4) photolyase also contains a fully reduced flavin adenine dinucleotide (FADH−) as the catalytic cofactor (򐂰Fig. 17.13 and 򐂰Fig. 17.14A) [11,114,116]. Upon excitation the FADH−* donates an electron to the 64PP to generate a charge-separated radical pair (FADH• + 64PP•−) [12,18,114]. However, because the formation of 64PP involves C6C4 bond formation and –OH group transfer from the 3’-pyrimidine to the 5’-pyrimidine,

Fig. 17.13: Enzyme-substrate complex structure and a possible repair scheme. Shown is the X-ray structure of D. melanogaster (6-4) photolyase bound to DNA containing a (6-4) photoproduct. The photolyase from A. thaliana has a similar structure with the conserved histidine residue in the active site (H364 in A. thaliana and H365 in D. melanogaster). The 64PP is flipped out of DNA and inserted into the active site. A close-up view shows the relative positions of the catalytic cofactor FADH−, the conserved H364 (H365) residue and the 64PP substrate with a proposed scheme for electron and proton transfers in the repair reaction. Shown in the repair scheme are forward electron transfer (FET) after light excitation, back electron transfer (BET) without repair, and the repair channel including initial proton transfer (PT) and late proton and electron return (PR and ER) after repair, with corresponding reaction rates (k1–k4) for the indicated steps.

17.4 Enzymatic reactions and repair photocycles in DNA photolyases

415

the direct C6-C4 bond splitting in 64PP•− unaccompanied by –OH transfer would simply produce two damaged bases (򐂰Fig. 17.13). Remarkably, (6-4) photolyase catalyzes this complex repair with no measurable side reactions. Various hypothetical repair models [114,116–121] have been proposed to rationalize the bond breakage/formation and group arrangements necessary for repair. Most models invoke PT [116,117,122] from a neighboring histidine residue, leading to the plausible scheme shown in 򐂰Fig. 17.13.

*

FADH

8 6

FADH •

450 500 550 600 650 Wavelength (nm)

4 2

FADH

64PP

E64PL

400 500 600 Wavelength (nm)

E64PL64PP 1.0

l fl 550 nm lpr 800 nm

0.5 t 225 ps b 0.8

0.5

lpr 640 nm

0.0 0

50

100

150

0

700

0.5

250

0.0

D 1.0

500

1000 1500 2000 2500 Delay time (ps)

500 E64PL64PP

0.5

With 64PP l fl 550 nm lpr 800 nm

1000 1500 2000 2500 Delay time (ps) H364N/M/Y/A/D/K WT(D2O) WT(H2O)

1.5 1.0 0.5 0.0 0

1 2 Visible light (hr)

3

WT(H2O) DA0.26

lpr 640 nm

WT(D2O) DA0.14 H364N DA0

0.0 0

Without 64PP t 3.0 ns b 1.0

0.0

Normalized DA

0

C 1.0



B 1.0

Inter.

300

Normalized DA

*

Repaired 64PP (a.u.)

10

FADH

Normal signal (a.u.)

E64PL

Fluo. (a.u.)

Absorption (mM1 cm1)

A 12

0

200 400 600 800 1000 1200 Delay time (ps)

Fig. 17.14: Femtosecond-resolved dynamics of flavin species involved in the damaged DNA repair. (A) Absorption spectra and coefficients of purified protein with FMNH•, converted active form FADH−, damaged 64PP, and FADH−* determined by this study. Also determined is a 64PPrelated reaction intermediate (Inter.) around 325 nm. The fluorescence emission of FADH−* is shown in inset with an arrow indicating the gated wavelength. (B) Normalized signals detected by both fluorescence (gated around the emission peak of 550 nm) and absorption (probed at 800 nm) methods with and without the substrate in the active site show the same lifetime and forward ET decays. The ET dynamics is best represented by a stretched-single-exponential decay. (C) Transient absorption signal probed at 640 nm with both FADH−* (blue curve) and FMNH• detection (green). The total FMNH• signal is from the two contributions of the initially formed one (dashed purple; k1 formation and k2+k3 decay in 򐂰Fig. 17.13) and the branched one in the repair channel (dashed cyan; k3 formation and k4 decay). Note the flat signal in tens of picosecond shown in inset, reflecting a fast apparent rise signal. (D) Transient absorption signals probed at 640 nm of the mutant H364N and the wild-type enzyme in D2O compared with the wild type in H2O, all showing the similar initial flat signals. But, for H364N the signal decays to zero and in D2O the signal drops to about a half of the original plateau in H2O. The corresponding relative steady-state quantum yield measurements are shown in inset.

416

17 Ultrafast dynamics of flavins and flavoproteins

17.4.2.1 Ultrafast electron and proton transfer dynamics To reveal the repair photocycle, we have performed ultrafast spectroscopy to identify the reaction intermediates. We have first characterized the initial ET by analysis of the excited FADH− dynamics using femtosecond fluorescence spectroscopy [12]. By monitoring the weak FADH−* emission at 550 nm (򐂰Fig. 17.14A) we find that the fluorescence of FADH−* in the absence of substrate is 3 ns but becomes dramatically shorter in the enzyme-substrate complex (򐂰Fig. 17.14B). The transient in the complex is best represented by a stretched-single-exponential decay, Ae−(t/τ)β, with τ = 225 ps and β = 0.8, reflecting a heterogeneous ET dynamics [12,103] which is continuously modulated by solvation of the active site as has been recently reported for (6-4) photolyase [102] and also observed in photosynthesis [68]. These results have been confirmed by transient absorption spectroscopy of the excited-state flavin dynamics (λpr = 800 nm) as shown in 򐂰Fig. 17.14B. Importantly, with mutagenesis of the potential proton donor, His364 [118] in the active site, to charged (K and D), polar (N and Y) and hydrophobic (A and M) residues, we observe similar ET dynamics that follow a stretched-single-exponential decay with τ = 147–281 ps and β = 0.8–0.9. This finding, along with the observation that 64PP appears to be in its canonical form in the crystal structure of the enzymecomplex [114], excludes an earlier repair model [117,118] that requires an oxetane precursor before photochemical excitation. We conclude that after injection of one electron, 64PP repair takes place entirely in the anionic ground state of 64PP•−. After the photoinduced charge separation (FADH•+64PP•−), the reaction can evolve along two pathways, back ET (k2) or 64PP repair (k3) (򐂰Fig. 17.13). By knowing the forward ET dynamics of FADH−*, we can map out the temporal evolution of FADH• by probing at wavelengths from 500 to 700 nm (򐂰Fig. 17.14A) to follow the 64PP repair. Clearly, in 򐂰Fig. 17.14C, the transient probed at 640 nm (red curve) shows a dramatically different behavior than when probed at 800 nm (blue curve) due to the capture of the radical FADH• (green curve). Strikingly, we observe an apparent rise signal of FADH• with τ = 45 ps (initial flat part in inset of 򐂰Fig. 17.14C) followed by a remarkably longtime plateau, indicating that complete 64PP repair takes longer than several nanoseconds. Because the forward ET occurs with τ = 225 ps, the 45-ps process must represent the overall decay of the initially formed FADH• ((k2+k3)−1 in 򐂰Fig. 17.13 and dashed purple curve in 򐂰Fig. 17.14C) but appears an apparent absorption increase. Slower formation and faster decay results in less FADH• accumulation. The dynamics of the branched FADH• in the repair channel (򐂰Fig. 17.13 and dashed cyan curve in 򐂰Fig. 17.14C) exhibits a complex formation (largely determined by the k1 process) and a slow decay (k4) but its amplitude is mainly determined by the k3 (proton transfer) rate. Upon deconvolution, we obtain back ET with τ = 50 ps (k2−1) and repair with τ = 425 ps (k3−1) to form a 64PP•− anionic intermediate. From these kinetics we obtained a value for repair branching of 0.097, which is in excellent agreement with the reported steadystate repair quantum yield of ~0.1 [123]. This suggests that after the k3 step, all subsequent reaction steps proceed to the final 64PP repair without any back ET that would result in a futile cycle. The same dynamics for FADH•− are observed throughout the 640–500 nm region. These findings reveal that the underlying molecular mechanism for the low repair quantum yield of (6-4) photolyase (~0.1) compared to CPD photolyase (~0.9) [11] is the faster rate of back ET (k2−1 = 50 ps) from 64PP•− to FADH• relative to the rate of proton transfer to 64PP•− (k3−1 = 425 ps) which is an essential step in catalysis.

17.4 Enzymatic reactions and repair photocycles in DNA photolyases

417

Surprisingly, for the series of mutants of H364N/M/Y/A/D/K used to test the proposed reaction mechanism we find that all transients probed in the 500–700 nm region show similar back ET dynamics in the range of 70–260 ps but decay to zero without any long plateaus as shown in 򐂰Fig. 17.14D for the H364N mutant. This observation is critical to the proposed reaction scheme as it indicates that in these mutants even though ET from FADH−* to 64PP is essentially normal the repair channel is completely shut off and all FADH• formed by initial charge separation follows a futile ET cycle back to FADH− by charge recombination (򐂰Fig. 17.13). These results are also consistent with our steadystate quantum yield measurements which reveal a total lack of repair with any of the mutants (򐂰Fig. 17.14D, inset). Collectively, our data, in agreement with an earlier report [118], indicate that H364 in the active site is a functional residue which is essential for repair, and proton transfer from H364 to 64PP•− is conceivably the rate-limiting step (k3) in the reaction scheme (򐂰Fig. 17.13). To test for proton transfer from H364 during the repair reaction, the reaction has been carried out in D2O. As shown in 򐂰Fig. 17.14D for the wild-type (6-4) photolyase, we observe a different transient with an obviously lower plateau, reflecting a slower repair process with τ = 1100 ps but with the similar forward (212 ps) and back (60 ps) ET dynamics as seen in H2O. Thus, with deuterated H364, repair through D+ transfer slows by a factor of more than two. The crystal structure shows a H-bond distance of 2.7 Å between the –NH of H364 and the –OH at the C5 position of the 5’ base (򐂰Fig. 17.13) and hence proton transfer from H364 to the –OH is quite feasible. The lower plateau in the transient in D2O is approximately half of that seen in H2O (򐂰Fig. 17.14D) and thus corresponds to a 50% decrease in the repair branching, in good agreement with the steady-state measured quantum yield ratio of 1:2 for D2O to H2O (򐂰Fig. 17.14D, inset). We have also studied both the repair dynamics and the steady-state enzyme activity over a pH range of 7 to 9 and do not see any changes, consistent with the observation that H364 remains protonated over a wide pH range [122]. All these results are consistent with proton transfer from H364 to 64PP•− to generate a protonated neutral radical 64PPH•, as a key step in the repair pathway (򐂰Fig. 17.13). This critical proton transfer, facilitated by the initial photoinduced electron transfer, completely blocks the futile back ET from 64PP•− and allows the reaction to proceed to repair with 100% efficiency after this step. To further confirm the proposed model, we have examined the repair processes by detection of the 64PP-related species and the recovery of FADH− in the UV region. We have obtained the transient dynamics of the 64PP intermediate formed in about 425 ps with protonated and in about 1100 ps with deuterated H364 and determined its absorption coefficient with a peak around 325 nm (򐂰Fig. 17.14A). Thus, in this system we are able to detect the repair reaction intermediates of both the cofactor (FADH•) and of the substrate (64PPH•). After protonation, the substrate radical intermediate decays in nanoseconds, corresponding to a series of atom arrangements with bond breaking and formation to complete the 64PP repair on a time scale of tens of nanoseconds.

17.4.2.2 Catalytic repair photocycle Based on these findings and previous data, including the crystal structure of the enzyme-substrate complex [114], we have proposed the catalytic photocycle for repair

418

17 Ultrafast dynamics of flavins and flavoproteins ard electron transfe r Forw 225 ps

FADH • FADH * H

hn

O

O

HN 6 O

N

O

O

N

N II

H N

N

4

5’

O

HN

I

H

H

O

H364

N

N

NH

NH

5’

O N

3’

FADH 50 ps Back electron transfer

NH O

N

V

O H N

5’

HN

O

O O N

n to



o Pr

O

N

N

rn

3’

N H

5’

etu

O

NH

HN

on r

N

III

H N

5’

3’ O

10 an ns d el ec tr

NH N

O H

H

HN O

N

Proton transfer 425 ps

3’

IV

O H N

O N

3’

B on d

ran rear

m ge

en

t

Fig. 17.15: Repair photocycle of (6-4) thymine photoproduct by (6-4) photolyase. The resolved elementary steps include a forward electron transfer in 225 ps upon excitation, a back electron transfer in 50 ps without any repair, and a parallel, catalytic proton transfer between the enzyme (H364) and the substrate, induced by the initial electron transfer, in 425 ps. This proton transfer is the determinant in the repair and determines the overall repair quantum yield. The subsequent repair reactions involve a series of atom arrangements with bond breaking and making and final proton and electron returns (to H364 residue and flavin cofactor) to convert the 64PP to two thymine bases on the time scales of longer than ten nanoseconds.

of thymine 64PP shown in 򐂰Fig. 17.15. In this scheme, the primary reactions are the initial electron transfer (I to II) and the subsequent proton transfers (II to III). The ETinduced proton from a His residue in (6-4) photolyase to the 64PP is a key step in the repair photocycle, like the “dividing line” in the transition state and making the subsequent reactions “downhill” without the possibility of back reaction. This critical step competes with the back ET resulting in an overall repair quantum yield of about 0.1, which is probably the maximum value that can be achieved for such a structurally and chemically challenging reaction, through slowing down back ET and speeding up the proton transfer processes. Also, the scheme that emerges from our data in which transient oxetane formation facilitates the oxygen-atom transfer from the 5’ to 3’ base followed by C6-C4 bond split (IV) would be less prone to mutagenic side reaction. In this scheme, following oxygen atom transfer and C-C bond cleavage the proton returns to H364 residue and the electron returns to FADH• to restore the enzyme to its active form and the 64PP to two repaired thymine bases (V).

17.5 Signal transduction in blue-light photoreceptors

419

17.5 Signal transduction in blue-light photoreceptors Significant studies have been recently carried out to understand the molecular mechanism of photo-sensing proteins in which a flavin cofactor is chromophore. Three classes of flavoprotein photoreceptors have been recognized; phototropin with LOV (Light, Oxygen and Voltage) domain [124], blue-light sensory protein with Blue Light sensing Using Flavin adenine dinucleotide (BLUF ) domain [125], and cryptochrome with a Photolyase Homology Region (PHR) domain [126]. Understanding the dynamics of the flavin is essential to unraveling the mechanisms of photoreceptors in signal transduction. The signal-transduction mechanisms of FMN in LOV domain and of FAD in BLUF domain have been well-characterized to date, but the photochemistry of FAD for signaling formation in the PHR domain of cryptochrome remains elusive.

17.5.1 Photoaddition of cysteine to flavin in phototropin Phototropins, which regulate phototropism in plants, contain two ~12 kDa FMNbinding LOV domains at the N-terminus and a serine/threonine kinase domain at the C-terminus [127]. The oxidized FMN-bound LOV domain absorbs around 450 nm in the dark. Blue-light illumination triggers a photocycle involving intersystem crossing from the electronically excited singlet state to the triplet state of flavin in nanoseconds, reversible formation of a blue-shifted signaling state with absorbance around 390 nm in microseconds, and a slow restoration to the original dark state in tens to hundreds of seconds [128]. The signaling state is formed through the light-induced formation of an adduct between a nearby cysteine residue and C(4a) of the FMN [129]. Although the photochemistry of flavin in the LOV domain has been well characterized, some questions remain unanswered. First, the mechanism of thio-adduct formation has not been identified. An ionic mechanism has been proposed because a protonated triplet state of flavin was observed by ultrafast absorption spectroscopy [41]. However, ultrafast infrared spectroscopic studies show the marker bands for an unprotonated triplet state of flavin [130,131]. Recent transient absorption spectroscopic studies support a mechanism involving a radical of FMNH• [132]. Second, it is not well understood why individual LOV domains differ so markedly in the kinetics and quantum yield of the photocycle [133]. The basis for these differences may have to do with protein flexibility or solvent accessibility, and the specific environment of the flavin chromophore.

17.5.2 Switching of flavin hydrogen bond in BLUF protein BLUF domain is present in various proteins that are involved in the photophobic response and in light-controlled gene expression in bacteria [134]. This domain exhibits typical features of oxidized FAD in the dark. Upon blue-light illumination, the FAD forms an intermediate signaling state in tens to hundreds of picoseconds having ~10 nm red-shifted absorption. This red-shifted signaling state returns to the original dark state on a seconds time scale. It has been postulated that the signaling state is formed after a light-induced switching of a hydrogen bond between an active site glutamine and the FAD [135]. However, two different mechanisms of the signaling-state formation have been proposed. The first one involves ET from a nearby tyrosine upon blue-light excitation of the FAD [135], while the second one involves both ET and proton transfer from the tyrosine [136]. Recent studies by ultrafast visible and infrared transient absorption

420

17 Ultrafast dynamics of flavins and flavoproteins

methods have provided evidence for ET from tyrosine to FAD* to form FAD•− with concomitant bleaching of a vibrational mode of tyrosine. The hydrogen bond between glutamine and FAD is reversed through a radical pair mechanism [137,138]. Thus, the mechanism of photoactivity in BLUF protein has not been firmly established and the intermediates between the dark and signaling states need to be further identified [54].

17.5.3 Ultrafast flavin dynamics in cryptochrome Cryptochrome serves as a blue-light receptor that regulates plant development and synchronizes animal circadian rhythm [139,140]. Recently, it has been noted that cryptochrome works like rhodopsin as an important component in rapid light perception [141]. A light-induced conformational change involving the C-terminal domain of the protein has been proposed to mediate the signal transduction [13,139,142,143], although the active state of the cofactor flavin is still under debate and the photocycle has not been established. We have systematically examined the excited-state dynamics of all oxidation states of the FAD cofactor with femtosecond resolution in both insect cryptochrome [26] and plant cryptochrome. We observe ultrafast photoreduction of oxidized FAD in subpicosecond and of neutral radical semiquinoid FADH• in tens of picoseconds. Also, we find that the excited anionic semiquinoid FAD•− and hydroquinoid FADH− could have long lifetimes, enabling the excited states to persist longer for functional pathways. In photolyases, the excited FADH− has a long (nanosecond) lifetime, long enough for the ET reaction with damaged DNA for efficient DNA-repair function. In insect cryptochromes the excited FAD•−* has complex decay dynamics on a time scale ranging from a few to hundreds of picoseconds. Decay is believed to occur through conical intersection(s) with a flexible bending motion to modulate the functional channel. These unique properties of anionic flavins suggest a universal mechanism of ET for the initial functional steps of the blue-light photolyase/cryptochrome flavoprotein family. Other transientabsorption studies of plant cryptochrome have reported proton transfer in 1.7 µs upon excitation, leading to the formation of a flavin neutral radical [144]. Proton transfer has not been observed in animal cryptochrome. Although the dynamics of ET and proton transfer between FAD and neighboring protein environments have been studied [26,144–146], further investigations are required to ascertain whether and how ET and proton transfer processes participate in signal transduction.

17.6 Conclusions We have summarized here recent progress in understanding of molecular mechanisms of several important flavoproteins. Using femtosecond spectroscopy in conjunction with site-directed mutagenesis, we have been able to follow the entire evolution of functional dynamics from the initial state, through functional intermediates to the final biological products and determine all actual time scales of biological changes, identify key functional residues at the active sites and thus reveal the complete molecular mechanisms at the most fundamental level. Flavins are ideal for ultrafast studies with their spectral signatures in the visible-light region. With wavelength tunability we have followed the oxidation state changes of flavin during catalytic reactions in real time, thus mapping out the entire photocycle. For flavodoxin, we have observed the fastest electron transfer (~100 fs) in proteins, the quantum vibrational effect on ET dynamics,

17.7 References

421

and the localized ET reactions confined in active-site nanospaces. For photolyase, we have resolved the two repair photocycles for damaged DNA of cyclobutane pyrimidine dimer and (6-4) photoproduct with a series of new findings: electron-transfer directionality, adenine-moiety mediation, sequential breaking of the cyclobutane ring, enzymeassisted proton transfer, and electrostatic and dynamic control of the repair efficiency. For photoreceptors, we have discussed several blue-light sensory flavoproteins for signal transduction. The dynamics and mechanisms have been gradually revealed, including formation of a cysteine adduct of the FAD in phototropin and hydrogen-bond switching in BLUF protein, to induce conformational changes. These advances are essential to understanding the functional processes and elucidating the molecular mechanisms. With the powerful integration of femtosecond temporal resolution and single-residue spatial resolution, we expect more exciting discoveries on flavoproteins forthcoming.

Acknowledgements We like to thank our group members at Ohio State University, past and present, who made significant contributions to the flavin story told here, especially to Chaitanya Saxena, Ya-Ting Kao, Ting-Fang He, Chuang Tan and Zheyun Liu. The work is supported in part by the National Institute of Health (Grant GM074813), the Packard Fellowship and the Camille Dreyfus Teacher-Scholar (to D.Z.), and the Ohio State University Pelotonia fellowship (to J.L.).

17.7 References [1] Walsh C. Flavin coenzymes - at the crossroads of biological redox chemistry. Accounts Chem Res 1980;13:148–55. [2] Massey V, Hemmerich P. Active-site probes of flavoproteins. Biochem Soc Trans 1980;8:246–57. [3] Muller F. The flavin redox-system and its biological function. Top Curr Chem 1983;108: 71–107. [4] Fraaije MW, Mattevi A. Flavoenzymes: diverse catalysts with recurrent features. Trends Biochem Sci 2000;25:126–32. [5] Massey V. The chemical and biological versatility of riboflavin. Biochem Soc Trans 2000;28:283–96. [6] Bornemann S. Flavoenzymes that catalyse reactions with no net redox change. Nat Prod Rep 2002;19:761–72. [7] Joosten V, van Berkel WJ. Flavoenzymes. Curr Opin Chem Biol 2007;11:195–202. [8] Senda T, Senda M, Kimura S, Ishida T. Redox control of protein conformation in flavoproteins. Antioxid Redox Signal 2009;11:1741–66. [9] Miura R. Versatility and specificity in flavoenzymes: control mechanisms of flavin reactivity. Chem Rec 2001;1:183–94. [10] Müller F, Eds. Chemistry and biochemistry of flavoenzymes. CRC Press, Boca Raton, FL, 1990. [11] Sancar A. Structure and function of DNA photolyase and cryptochrome blue-light photoreceptors. Chem Rev 2003;103:2203–37. [12] Kao YT, Saxena C, Wang L, Sancar A, Zhong D. Direct observation of thymine dimer repair in DNA by photolyase. Proc Natl Acad Sci USA 2005;102:16128–32. [13] Partch CL, Sancar A. Photochemistry and photobiology of cryptochrome blue-light photopigments: the search for a photocycle. Photochem Photobiol 2005;81:1291–304.

422

17 Ultrafast dynamics of flavins and flavoproteins

[14] Lin C, Shalitin D. Cryptochrome structure and signal transduction. Annu Rev Plant Biol 2003;54:469–96. [15] Lin C, Todo T. The cryptochromes. Genome Biol 2005;6:220. [16] Kennis JT, Groot ML. Ultrafast spectroscopy of biological photoreceptors. Curr Opin Struct Biol 2007;17:623–30. [17] Christie JM. Phototropin blue-light receptors. Annu Rev Plant Biol 2007;58:21–45. [18] Zhong D. Ultrafast catalytic processes in enzymes. Curr Opin Chem Biol 2007;11:174–81. [19] Kao YT, Saxena C, He TF, et al. Ultrafast dynamics of flavins in five redox states. J Am Chem Soc 2008;130:13132–9. [20] Heelis PF. The photophysical and photochemical properties of flavins (isoalloxazines). Chem Soc Rev 1982;11:15–39. [21] Ghisla S, Massey V, Lhoste J-M, Mayhew SG. Fluorescence and optical characteristics of reduced flavines and flavoproteins. Biochemistry 1974;13:589–97. [22] Land EJ, Swallow AJ. One-electron reactions in biochemical systems as studied by pulse radiolysis 2. riboflavin. Biochemistry 1969;8:2117–25. [23] Sakai M, Takahashi H. One-electron photoreduction of flavin mononucleotide: time-resolved resonance Raman and absorption study. J Mol Struct 1996;379:9–18. [24] Massey V, Palmer G. On existence of spectrally distinct classes of flavoprotein semiquinones. A new method for quantitative production of flavoprotein semiquinones. Biochemistry 1966;5:3181–9. [25] Massey V. The chemical and biological versatility of riboflavin. Biochem Soc Trans 2000;28:283–96. [26] Kao YT, Tan C, Song SH, et al. Ultrafast dynamics and anionic active states of the flavin cofactor in cryptochrome and photolyase. J Am Chem Soc 2008;130:7695–701. [27] Stanley RJ, MacFarlane AW, IV. Ultrafast excited state dynamics of oxidized flavins: direct observations of quenching by purines. J Phys Chem A 2000;104:6899–906. [28] Malicka J, Groth M, Karolczak J, Czaplewski C, Liwo A, Wiczk W. Influence of solvents and leucine configuration at position 5 on tryptophan fluorescence in cyclic enkephalin analogues. Biopolymers 2001;58:447–57. [29] Reichardt C. Solvents and solvent effects in organic chemistry 3rd ed.; Wiley-VCH: Weinheim, 2003. [30] Barrio JR, Tolman GL, Leonard NJ, Spencer RD, Weber G. Flavin 1, N-6 ethenoadenine dinucleotide - dynamic and static quenching of fluorescence. Proc Natl Acad Sci USA 1973;70:941–3. [31] Weber G. Fluorescence of riboflavin and flavin-adenine dinucleotide Biochem J 1950;47: 114–21. [32] van den Berg PAW, Feenstra KA, Mark AE, Berendsen HJC, Visser AJWG. Dynamic conformations of flavin adenine dinucleotide: simulated molecular dynamics of the flavin cofactor related to the time-resolved fluorescence characteristics. J Phys Chem B 2002;106:8858–69. [33] Kotowycz G, Teng N, Klein MP, Calvin M. 220-Mhz Nuclear magnetic resonance study of a solvent-induced conformational change in flavin adenine dinucleotide. J Biol Chem 1969;244:5656–62. [34] Kainosho M, Kyogoku Y. High-resolution proton and phosphorus nuclear magnetic-resonance spectra of flavin-adenine dinucleotide and its conformation in aqueous-solution. Biochemistry 1972;11:741–52. [35] Copeland RA, Spiro TG. Ultraviolet Resonance raman-spectroscopy of flavin mononucleotide and flavin adenine-dinucleotide. J Phys Chem 1986;90:6648–54. [36] Weber G, Tanaka F, Okamoto BY, Drickamer HG. Effect of pressure on molecular complex of isoalloxazine and adenine. Proc Natl Acad Sci USA 1974;71:1264–6. [37] Visser AJWG. Kinetics of stacking interactions in flavin adenine-dinucleotide from timeresolved flavin fluorescence. Photochem Photobiol 1984;40:703–6.

17.7 References

423

[38] Hahn J, Michel-Beyerle M-E, Rösch N. Conformation of the flavin adenine dinucleotide cofactor FAD in DNA-photolyase: a molecular dynamics study. J Mol Model 1998;4:73–82. [39] Islam SDM, Susdorf T, Penzkofer A, Hegemann P. Fluorescence quenching of flavin adenine dinucleotide in aqueous solution by pH dependent isomerisation and photo-induced electron transfer. Chem Phys 2003;295:137–49. [40] Gauden M, Yeremenko S, Laan W, et al. Photocycle of the flavin-binding photoreceptor AppA, a bacterial transcriptional antirepressor of photosynthesis genes. Biochemistry 2005;44: 3653–62. [41] Kennis JTM, Crosson S, Gauden M, van Stokkum IHM, Moffat K, van Grondelle R. Primary reactions of the LOV2 domain of phototropin, a plant blue-light photoreceptor. Biochemistry 2003;42:3385–92. [42] Holzer W, Penzkofer A, Fuhrmann M, Hegemann P. Spectroscopic characterization of flavin mononucleotide bound to the LOV1 domain of Phot1 from Chlamydomonas reinhardtii. Photochem Photobiol 2002;75:479–87. [43] Chosrowjan H, Taniguchi S, Mataga N, Tanaka F, Visser AJWG. The stacked flavin adenine dinucleotide conformation in water is fluorescent on picosecond timescale. Chem Phys Lett 2003;378:354–8. [44] Moser CC, Keske JM, Warncke K, Farid RS, Dutton PL. Nature of biological electron-transfer. Nature 1992;355:796–802. [45] Qiu W, Kao Y-T, Zhang L, et al. Protein surface hydration mapped by site-specific mutations. Proc Natl Acad Sci USA 2006;103:13979–84. [46] Zhang L, Wang L, Kao Y-T, et al. Mapping hydration dynamics around a protein surface. Proc Natl Acad Sci USA 2007;104:18461–6. [47] Jovanovic SV, Simic MG. One-electron redox potentials of purines and pyrimidines. J Phys Chem 1986;90:974–8. [48] Candeias LP, Steenken S. Electron-transfer in di(deoxy)nucleoside phosphates in aqueoussolution - rapid migration of oxidative damage (via adenine) to guanine. J Am Chem Soc 1993;115:2437–40. [49] Wolf MM, Schumann C, Gross R, Domratcheva T, Diller R. Ultrafast infrared spectroscopy of riboflavin: dynamics, electronic structure, and vibrational mode analysis. J Phys Chem B 2008;112:13424–32. [50] Weigel A, Dobryakov AL, Veiga M, Perez Lustres JL. Photoinduced processes in riboflavin: superposition of pi pi*-n pi* states by vibronic coupling, transfer of vibrational coherence, and population dynamics under solvent control. J Phys Chem A 2008;112:12054–65. [51] Kondo M, Nappa J, Ronayne KL, Stelling AL, Tonge PJ, Meech SR. Ultrafast vibrational spectroscopy of the flavin chromophore. J Phys Chem B 2006;110:20107–10. [52] Gil M, Wang YL, Douhal A. Ultrafast dynamics of lumichrome in solution and in chemical and biological caging media. J Photoch Photobio A 2012;234:146–55. [53] Partch CL, Clarkson MW, Özgür S, Lee AL, Sancar A. Role of structural plasticity in signal transduction by the cryptochrome blue-light photoreceptor. Biochemistry 2005;44: 3795–805. [54] Lukacs A, Zhao RK, Haigney A, et al. Excited state structure and dynamics of the neutral and anionic flavin radical revealed by ultrafast transient mid-IR to visible spectroscopy. J Phys Chem B 2012;116:5810–8. [55] Lennon BW, Williams CH, Jr., Ludwig ML. Crystal structure of reduced thioredoxin reductase from Escherichia coli: structural flexibility in the isoalloxazine ring of the flavin adenine dinucleotide cofactor. Protein Sci 1999;8:2366–79. [56] Visser AJWG, Ghisla S, Lee J. Picosecond fluorescence dyanmcis of reduced flavins. Flavins and Falvopoteins 1990;1990:49–54. [57] Leenders R, Kooijman M, van Hoek A, Veeger C, Visser AJWG. Flavin dynamics in reduced flavodoxins - a time-resolved polarized fluorescence study. Eur J Biochem 1993;211:37–45.

424

17 Ultrafast dynamics of flavins and flavoproteins

[58] Bruggeman YE, Schoenmakers RG, Schots A, et al. Monoclonal-antibodies against 2 electron reduced riboflavin and a quantification of affinity constants for this oxygen-sensitive molecule. Eur J Biochem 1995;234:245–50. [59] Zhao RK, Lukacs A, Haigney A, et al. Ultrafast transient mid IR to visible spectroscopy of fully reduced flavins. Phys Chem Chem Phys 2011;13:17642–8. [60] Xiong P, Nocek JM, Vura-Weis J, Lockard JV, Wasielewski MR, Hoffman BM. Faster interprotein electron transfer in a [myoglobin, b(5)] complex with a redesigned interface. Science 2010;330:1075–8. [61] Gray HB, Winkler JR. Long-range electron transfer. Proc Natl Acad Sci USA 2005;102:3534–9. [62] Stubbe J, Nocera DG, Yee CS, Chang MC. Radical initiation in the class I ribonucleotide reductase: long-range proton-coupled electron transfer? Chem Rev 2003;103:2167–201. [63] Page CC, Moser CC, Chen XX, Dutton PL. Natural engineering principles of electron tunnelling in biological oxidation-reduction. Nature 1999;402:47–52. [64] Lin JP, Balabin IA, Beratan DN. The nature of aqueous tunneling pathways between electrontransfer proteins. Science 2005;310:1311–3. [65] Berlin YA, Burin AL, Siebbeles LDA, Ratner MA. Conformationally gated rate processes in biological macromolecules. J Phys Chem A 2001;105:5666–78. [66] Daizadeh I, Medvedev ES, Stuchebrukhov AA. Effect of protein dynamics on biological electron transfer. Proc Natl Acad Sci USA 1997;94:3703–8. [67] Marcus RA, Sutin N. Electron transfers in chemistry and biology. Biochim Biophys Acta 1985;811:265–322. [68] Wang HY, Lin S, Allen JP, et al. Protein dynamics control the kinetics of initial electron transfer in photosynthesis. Science 2007;316:747–50. [69] Li J, Liu ZY, Tan C, et al. Dynamics and mechanism of repair of ultraviolet-induced (6-4) photoproduct by photolyase. Nature 2010;466:887–U124. [70] Liu Z, Guo X, Tan C, et al. Electron tunneling pathways and role of adenine in repair of cyclobutane pyrimidine dimer by DNA photolyase. J Am Chem Soc 2012;134:8104–14. [71] Sumi H, Marcus RA. Dynamic effects in electron-transfer reactions. J Chem Phys 1986;84:4894–914. [72] Walker GC, Akesson E, Johnson AE, Levinger NE, Barbara PF. Interplay of solvent motion and vibrational-excitation in electron-transfer kinetics - experiment and theory. J Phys Chem 1992;96:3728–36. [73] Nagasawa Y, Yartsev AP, Tominaga K, Bisht PB, Johnson AE, Yoshihara K. Dynamical aspects of ultrafast intermolecular electron-transfer faster than solvation process - substituent effects and energy-gap dependence. J Phys Chem 1995;99:653–62. [74] Zhong D, Zewail AH. Femtosecond dynamics of flavoproteins: charge separation and recombination in riboflavine (vitamin B-2)-binding protein and in glucose oxidase enzyme. Proc Natl Acad Sci USA 2001;98:11867–72. [75] Mataga N, Chosrowjan H, Taniguchi S, Tanaka F, Kido N, Kitamura M. Femtosecond fluorescence dynamics of flavoproteins: comparative studies on flavodoxin, its site-directed mutants, and riboflavin binding protein regarding ultrafast electron transfer in protein nanospaces. J Phys Chem B 2002;106:8917–20. [76] Park H-W, Kim S-T, Sancar A, Deisenhofer J. Crystal-structure of DNA photolyase from Escherichia coli. Science 1995;268:1866–72. [77] Li YF, Heelis PF, Sancar A. Active-site of DNA photolyase - tryptophan-306 is the intrinsic hydrogen-atom donor essential for flavin radical photoreduction and DNA-repair in vitro. Biochemistry 1991;30:6322–9. [78] Aubert C, Vos MH, Mathis P, Eker APM, Brettel K. Intraprotein radical transfer during photoactivation of DNA photolyase. Nature 2000;405:586–90. [79] Byrdin M, Eker APM, Vos MH, Brettel K. Dissection of the triple tryptophan electron transfer chain in Escherichial coli DNA photolyase: Trp382 is the primary donor in photoactivation. Proc Natl Acad Sci USA 2003;100:8676–81.

17.7 References

425

[80] Saxena C, Sancar A, Zhong D. Femtosecond dynamics of DNA photolyase: energy transfer of antenna initiation and electron transfer of cofactor reduction. J Phys Chem B 2004;108:18026–33. [81] Wang H, Saxena C, Quan D, Sancar A, Zhong D. Femtosecond dynamics of flavin cofactor in DNA photolyase: radical reduction, local solvation, and charge recombination. J Phys Chem B 2005;109:1329–33. [82] Lukacs A, Eker APM, Byrdin M, Brettel K, Vos MH. Electron hopping through the 15 angstrom triple tryptophan molecular wire in DNA photolyase occurs within 30 ps. J Am Chem Soc 2008;130:14394. [83] Brazard J, Usman A, Lacombat F, et al. Spectro-temporal characterization of the photoactivation mechanism of two new oxidized cryptochrome/photolyase photoreceptors. J Am Chem Soc 2010;132:4935–45. [84] He T-F, Guo L, Guo X, Chang C-W, Wang L, Zhong D. Direct observation of vibrationally coupled electron transfer in protein flavodoxin. Submitted 2012. [85] Sancho J. Flavodoxins: sequence, folding, binding, function and beyond. Cell Mol Life Sci 2006;63:855–64. [86] Watt W, Tulinsky A, Swenson RP, Watenpaugh KD. Comparison of the crystal structures of a flavodoxin in its three oxidation states at cryogenic temperatures. J Mol Biol 1991;218:195–208. [87] Chang CW, He TF, Guo L, et al. Mapping solvation dynamics at the function site of flavodoxin in three redox states. J Am Chem Soc 2010;132:12741–7. [88] O’Farrell PA, Walsh MA, McCarthy AA, Higgins TM, Voordouw G, Mayhew SG. Modulation of the redox potentials of FMN in desulfovibrio vulgaris flavodoxin: thermodynamic properties and crystal structures of glycine-61 mutants. Biochemistry 1998;37:8405–16. [89] Zhou ZM, Swenson RP. Electrostatic effects of surface acidic amino-acid-residues on the oxidation-reduction potentials of the flavodoxin from desulfovibrio-vulgaris (Hildenborough). Biochemistry 1995;34:3183–92. [90] Swenson RP, Krey GD. Site-directed mutagenesis of tyrosine-98 in the flavodoxin from Desulfovibrio-Vulgaris (Hildenborough) - Regulation of oxidation-reduction properties of the bound FMN cofactor by aromatic, solvent, and electrostatic interactions. Biochemistry 1994;33:8505–14. [91] Karen A, Sawada MT, Tanaka F, Mataga N. Dynamics of excited flavoproteins - picosecond laser photolysis studies. Photochem Photobiol 1987;45:49–53. [92] Edmondson DE, Tollin G. Semiquinone formation in flavo- and metalloflavoproteins. Top Curr Chem 1983;108:109–38. [93] Ganapathi MR, Hermann R, Naumov S, Brede O. Free electron transfer from several phenols to radical cations of non-polar solvents. Phys Chem Chem Phys 2000;2:4947–55. [94] Liu Z, Tan C, Guo X, et al. Dynamics and mechanism of cyclobutane pyrimidine dimer repair by DNA photolyase. Proc Natl Acad Sci USA 2011;108:14831–6. [95] Genzor CG, Perales-Alcon A, Sancho J, Romero A. Closure of a tyrosine/tryptophan aromatic gate leads to a compact fold in apo flavodoxin. Nat Struct Biol 1996;3:329–32. [96] Wan CZ, Xia TB, Becker HC, Zewail AH. Ultrafast unequilibrated charge transfer: a new channel in the quenching of fluorescent biological probes. Chem Phys Lett 2005;412: 158–63. [97] Gladkikh V, Burshtein AI, Feskov SV, Ivanov AI, Vauthey E. Hot recombination of photogenerated ion pairs. J Chem Phys 2005;123:244510. [98] Cheng PY, Zhong D, Zewail AH. Femtosecond real-time probing of reactions .21. Direct observation of transition-state dynamics and structure in charge-transfer reactions. J Chem Phys 1996;105:6216–48. [99] Zhong D, Bernhardt TM, Zewail AH. Femtosecond real-time probing of reactions. 24. Time, velocity, and orientation mapping of the dynamics of dative bonding in bimolecular electron transfer reactions. J Phys Chem A 1999;103:10093–117.

426

17 Ultrafast dynamics of flavins and flavoproteins

[100] Eisenberg AS, Schelvis JPM. Contributions of the 8-methyl group to the vibrational normal modes of flavin mononucleotide and its 5-methyl semiquinone radical. J Phys Chem A 2008;112:6179–89. [101] Zhong D. Hydration dynamics and coupled water-protein fluctuations probed by intrinsic tryptophan. Adv Chem Phys 2009;143:83–149. [102] Chang CW, Guo L, Kao YT, et al. Ultrafast solvation dynamics at binding and active sites of photolyases. Proc Natl Acad Sci USA 2010;107:2914–9. [103] Kao YT, Saxena C, Wang L, Sancar A, Zhong D. Femtochemistry in enzyme catalysis: DNA photolyase. Cell Biochem Biophys 2007;48:32–44. [104] MacFarlane AWT, Stanley RJ. Cis-syn thymidine dimer repair by DNA photolyase in real time. Biochemistry 2003;42:8558–68. [105] Hassanali AA, Zhong D, Singer SJ. An AIMD study of the CPD repair mechanism in water: reaction free energy surface and mechanistic implications. J Phys Chem B 2011;115:3848–59. [106] Kao YT, Song QH, Saxena C, Wang L, Zhong D. Dynamics and mechanism of DNA repair in a biomimetic system: flavin-thymine dimer adduct. J Am Chem Soc 2012;134:1501–3. [107] Masson F, Laino T, Rothlisberger U, Hutter J. A QM/MM Investigation of thymine dimer radical anion splitting catalyzed by DNA photolyase. Chemphyschem 2009;10:400–10. [108] Harrison CB, O’Neil LL, Wiest O. Computational studies of DNA photolyase. J Phys Chem A 2005;109:7001–12. [109] Mees A, Klar T, Gnau P, et al. Crystal structure of a photolyase bound to a CPD-like DNA lesion after in situ repair. Science 2004;306:1789–93. [110] Seidel CAM, Schulz A, Sauer MHM. Nucleobase-specific quenching of fluorescent dyes.1. Nucleobase one-electron redox potentials and their correlation with static and dynamic quenching efficiencies. J Phys Chem 1996;100:5541–53. [111] Antony J, Medvedev DM, Stuchebrukhov AA. Theoretical study of electron transfer between the photolyase catalytic cofactor FADH(-) and DNA thymine dimer. J Am Chem Soc 2000;122:1057–65. [112] Prytkova TR, Beratan DN, Skourtis SS. Photoselected electron transfer pathways in DNA photolyase. P Natl Acad Sci USA 2007;104:802–7. [113] Acocella A, Jones GA, Zerbetto F. What is adenine doing in photolyase? J Phys Chem B 2010;114:4101–6. [114] Maul MJ, Barends TR, Glas AF, et al. Crystal structure and mechanism of a DNA (6-4) photolyase. Angew Chem Int Ed Engl 2008;47:10076–80. [115] Boggio-Pasqua M, Groenhof G, Schafer LV, Grubmuller H, Robb MA. Ultrafast deactivation channel for thymine dimerization. J Am Chem Soc 2007;129:10996–7. [116] Hitomi K, DiTacchio L, Arvai AS, et al. Functional motifs in the (6-4) photolyase crystal structure make a comparative framework for DNA repair photolyases and clock cryptochromes. Proc Natl Acad Sci USA 2009;106:6962–7. [117] Zhao XD, Liu JQ, Hsu DS, Zhao SY, Taylor JS, Sancar A. Reaction mechanism of (6-4) photolyase. J Biol Chem 1997;272:32580–90. [118] Hitomi K, Nakamura H, Kim ST, et al. Role of two histidines in the (6-4) photolyase reaction. J Biol Chem 2001;276:10103–9. [119] Joseph A, Prakash G, Falvey DE. Model studies of the (6-4) photoproduct photolyase enzyme: laser flash photolysis experiments confirm radical ion intermediates in the sensitized repair of thymine oxetane adducts. J Am Chem Soc 2000;122:11219–25. [120] Borg OA, Eriksson LA, Durbeej B. Electron-transfer induced repair of 6-4 photoproducts in DNA: a computational study. J Phys Chem A 2007;111:2351–61. [121] Domratcheva T, Schlichting I. Electronic structure of (6-4) DNA photoproduct repair involving a non-oxetane pathway. J Am Chem Soc 2009;131:17793–9. [122] Schleicher E, Hitomi K, Kay CW, Getzoff ED, Todo T, Weber S. Electron nuclear double resonance differentiates complementary roles for active site histidines in (6-4) photolyase. J Biol Chem 2007;282:4738–47.

17.7 References

427

[123] Hitomi K, Kim ST, Iwai S, et al. Binding and catalytic properties of Xenopus (6-4) photolyase. J Biol Chem 1997;272:32591–8. [124] Huala E, Oeller PW, Liscum E, Han IS, Larsen E, Briggs WR. Arabidopsis NPH1: a protein kinase with a putative redox-sensing domain. Science 1997;278:2120–3. [125] Iseki M, Matsunaga S, Murakami A, et al. A blue-light-activated adenylyl cyclase mediates photoavoidance in Euglena gracilis. Nature 2002;415:1047–51. [126] Lin C, Robertson DE, Ahmad M, et al. Association of flavin adenine dinucleotide with the Arabidopsis blue light receptor CRY1. Science 1995;269:968–70. [127] Briggs WR, Christie JM. Phototropins 1 and 2: versatile plant blue-light receptors. Trends Plant Sci 2002;7:204–10. [128] Kottke T, Hegemann P, Dick B, Heberle J. The photochemistry of the light-, oxygen-, and voltage-sensitive domains in the algal blue light receptor phot. Biopolymers 2006;82:373–8. [129] Salomon M, Christie JM, Knieb E, Lempert U, Briggs WR. Photochemical and mutational analysis of the FMN-binding domains of the plant blue light receptor, phototropin. Biochemistry 2000;39:9401–10. [130] Alexandre MTA, Domratcheva T, Bonetti C, et al. Primary reactions of the LOV2 domain of phototropin studied with ultrafast mid-infrared spectroscopy and quantum chemistry. Biophys J 2009;97:227–37. [131] Pfeifer A, Majerus T, Zikihara K, et al. Time-resolved fourier transform infrared study on photoadduct formation and secondary structural changes within the phototropin LOV domain. Biophys J 2009;96:1462–70. [132] Bauer C, Rabl CR, Heberle J, Kottke T. Indication for a radical intermediate preceding the signaling state in the LOV domain photocycle. Photochem Photobiol 2011;87:548–53. [133] Moglich A, Yang X, Ayers RA, Moffat K. Structure and function of plant photoreceptors. Annu Rev Plant Biol 2010;61:21–47. [134] Losi A, Gartner W. Old Chromophores, New photoactivation paradigms, trendy applications: flavins in blue light-sensing photoreceptors. Photochem Photobiol 2011;87:491–510. [135] Anderson S, Dragnea V, Masuda S, Ybe J, Moffat K, Bauer C. Structure of a novel photoreceptor, the BLUF domain of AppA from Rhodobacter sphaeroides. Biochemistry 2005;44:7998– 8005. [136] Laan W, van der Horst MA, van Stokkum IH, Hellingwerf KJ. Initial characterization of the primary photochemistry of AppA, a blue-light-using flavin adenine dinucleotide-domain containing transcriptional antirepressor protein from Rhodobacter sphaeroides: a key role for reversible intramolecular proton transfer from the flavin adenine dinucleotide chromophore to a conserved tyrosine? Photochem Photobiol 2003;78:290–7. [137] Bonetti C, Mathes T, van Stokkum IH, et al. Hydrogen bond switching among flavin and amino acid side chains in the BLUF photoreceptor observed by ultrafast infrared spectroscopy. Biophys J 2008;95:4790–802. [138] Gauden M, van Stokkum IHM, Key JM, et al. Hydrogen-bond switching through a radical pair mechanism in a flavin-binding photoreceptor. Proc Natl Acad Sci USA 2006;103:10895–900. [139] Liu H, Liu B, Zhao C, Pepper M, Lin C. The action mechanisms of plant cryptochromes. Trends Plant Sci 2011;16:684–91. [140] Ozturk N, Song SH, Ozgur S, et al. Structure and function of animal cryptochromes. Cold Spring Harb Symp Quant Biol 2007;72:119–31. [141] Fogle KJ, Parson KG, Dahm NA, Holmes TC. Cryptochrome is a blue-light sensor that regulates neuronal firing rate. Science 2011;331:1409–13. [142] Liu B, Liu HT, Zhong D, Lin CT. Searching for a photocycle of the cryptochrome photoreceptors. Curr Opin Plant Biol 2010;13:578–86. [143] Ozturk N, Selby CP, Annayev Y, Zhong D, Sancar A. Reaction mechanism of Drosophila cryptochrome. Proc Natl Acad Sci USA 2011;108:516–21. [144] Langenbacher T, Immeln D, Dick B, Kottke T. Microsecond light-induced proton transfer to flavin in the blue light sensor plant cryptochrome. J Am Chem Soc 2009;131:14274–80.

428

17 Ultrafast dynamics of flavins and flavoproteins

[145] Brazard J, Usman A, Lacombat F, et al. Spectro-temporal characterization of the photoactivation mechanism of two new oxidized cryptochrome/photolyase photoreceptors. J Am Chem Soc 2010;132:4935–45. [146] Immeln D, Weigel A, Kottke T, Perez Lustres JL. Primary events in the blue light sensor plant cryptochrome: intraprotein electron and proton transfer revealed by femtosecond spectroscopy. J Am Chem Soc 2012;134:12536–46.

Index

14-3-3 protein 134, 137, 139, 140 2,6-dichlorindophenol 146, 147 2-heptyl-4-hydroxy quinoline-N-oxide 154 2-ketopropyl-Coenzyme M carboxylase/ oxidoreductase 167 4-(methylnitrosamino)-1-(3-pyridyl)-1butanone 91 4a-peroxyflavin 290 60s loop 326, 327, 328, 332 absorbance spectra 5, 10, 11, 13, 14, 19, 20, 362, 366 acid-labile sulfide 142 adenylate cyclase 363 AFM. See atomic force microscopy agarose beads 286 AhpD 175 alcaligin 33, 47, 49 aldehyde oxidase 103, 104, 105, 111, 121 alkanesulfonate monooxygenase 255 alkanesulfonates to sulfites and aldehydes 256 Arg297 and conformational changes 262 Baeyer-Villiger mechanism 265 alkanesulfonate monooxygenase system SsuD/SsuE 255 amine 1, 2, 18, 20 aminobenzoate 9, 13, 14 anthrachelin 30 Antigen-antibody interaction 287 Antley-Bixler syndrome 91 AppA 237, 251, 252 aromatic hydroxylase 1, 2, 13, 16 aromatic substrate 2, 4 Atomic Force Microscopy 277, 279, 280, 284, 288 ATP synthase 142 aureochromes 364 autophosphorylation 361, 368, 369, 370, 375, 385, 386, 388 azurin 287, 297 bacterial luciferase 11, 25 Baeyer-Villiger monooxygenase 16, 17, 18, 19, 21, 22, 23

Baeyer-Villiger oxidation 58, 67 Baeyer-Villiger reaction 18, 19 bFMO monooxygenase 40 biological motors 281 biomolecular modeling 336 blue-light photoreceptor 393, 394, 421, 423 blue-light receptor 420 blue-light using flavoproteins 237 BLUF 393, 419, 421, 427 BLUF domains 362, 363, 380, 381, 392 Boltzmann ensemble 342 boronate 1, 2 bovine serum albumin 286 Brownian dynamics 342 BVMO single component 58 C4a flavin-thiol adduct 170 C4a-(hydro)peroxyflavin 255 C4a-hydroperoxyflavin 29, 35, 36, 37, 42, 43, 44, 46 C4a-hydroxyflavin 44 C4a-peroxyflavin nucleophilic attack on sulfonyl group 265 calcium-dependent protein kinase 127, 130, 134 carbon-sulfur bond cleavage 256 charge 336, 337 charge distribution 227, 228, 236 charge recombination 397, 404, 406, 408, 417, 425 charge redistribution 225, 228, 231, 232, 236 charge-transfer 171, 173, 188, 232, 234 charge-transfer absorbance 9 charge-transfer band 207 charge-transfer complex 5, 132, 133, 207 charge-transfer interaction 5, 9 CHMO. See cyclohexanone monooxygenase cholesterol biosynthesis 91 cholesterol oxidase 288 circadian clock 361, 364, 371, 378, 379, 382, 387, 389, 391 CoA disulfide reductases 169 CoADR 192, 200

430

Index

CoADR-RHD 192 coelichelin B 34 coenzyme Q. See ubiquinone complex ii. See succinate dehydrogenase Compound I 86 conformational change 5, 6, 7, 9, 16, 19, 25, 183, 185, 199, 369, 373, 374, 375, 377, 378, 391 continuum electrostatics 336 CPD photolyase 408 Criegee intermediate 17, 19, 21, 22, 23, 26 cryptochrome 361, 362, 363, 371, 372, 376, 380, 381, 386, 387, 388, 389, 390, 391, 393, 420, 422, 427. See blue-light receptor cyclobutane pyrimidine dimer 234, 235, 236, 246, 250 cyclohexanone 17, 19, 25, 26, 27 cyclohexanone monooxygenase 17, 18, 22, 23, 37, 40, 49, 50 CYP26 91 CYP51 79, 91 Cys-S-S-CoA 188, 191 cysteine-based redox center 167 cysteine sulfenic acid 167, 188, 189 cytochrome b5 74, 79, 80, 81, 86, 87, 93, 126, 127, 128, 129, 132, 138, 139 cytochrome bc1 135 cytochrome c 128, 132, 134, 135, 139, 141, 142 cytochrome c551 287, 297 cytochrome P450 19, 20, 73, 74, 78, 79, 80, 81, 86, 87, 88, 91, 92, 93 cytochrome P450 oxidoreductase deficiency 91, 93 cytochrome P450 reductase 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93 deazaflavin 234, 250 density functional theory 238, 242, 252 DepH 169, 188 desulfonation of alkanesulfonates 255 DFS. See dynamic force spectroscopy DHODH. See Dihidroorotate dehydrogenase dielectric constant 336 dienone 12, 14 diethylene-triamine-pentaacetate 346 differential equation 353 diflavin oxidoreductase 73, 74, 78, 90 Dihidroorotate dehydrogenase 289

dihydrolipoamide dehydrogenase 167, 168, 170, 174, 175, 197 E3 E3BD 174 dihydroorotate 289, 297 dihydroorotate dehydrogenase 204, 205, 206, 207, 208, 210, 213, 215, 289 dihydropyrimidine dehydrogenase 204, 205, 215 dihydropyrimidines 204, 205 dihydrouracil 204, 205, 213 dihydrouridine synthase 204, 205, 213, 214 dihydroxybenzoate 2, 4-dihydroxybenzoate 6, 7, 14, 24 3, 4-dihydroxybenzoate 4, 6, 12, 13, 14, 15 dimer interface 4 dipole moment 227, 228, 231, 239, 240, 242, 248 disulfide 18, 20 disulfide oxidoreductases 169, 197 DNA photolyase 226, 234, 235, 236, 237, 246, 248, 249, 250, 251, 252, 253, 393, 394, 395, 396, 399, 400, 401, 408, 409, 410, 411, 412, 414, 415, 416, 417, 418, 420, 421, 422, 423, 424, 425, 426, 427, 428 DNA repair 225, 244, 249, 250, 251 DNA replication 278 docking 341 DTPA 346 E3-ubiquitin ligase 378 effector 6 EH2 169, 170, 171, 172, 173, 176, 180, 185, 192 EH2•NADP+ 171, 176, 192 EH2•NADPH 172, 173, 176 electroabsorption 228, 249 electronic coupling 353 electronic structure 225, 226, 227, 228, 230, 232, 235, 236, 238, 239, 240, 241, 243, 244, 253 electron paramagnetic resonance spectroscopy 105, 110, 113, 122 electron transfer 129, 130, 131, 132, 133, 134, 135, 137, 138, 139, 140, 393, 394, 396, 397, 398, 400, 401, 402, 403, 404, 405, 406, 407, 408, 409, 410, 412, 413, 414, 415, 416, 417, 418, 419, 420, 421, 423, 424, 425, 426 electron transfer proteins 341

Index electrophilic aromatic substitution 1, 2, 14, 16, 56 electrostatic energy 339 electrostatic interactions 322, 323, 333 electrostatic potential 336, 339 electrostatics 9, 13, 15, 16, 26, 373, 375 elimination reaction 1, 9, 13, 14, 19, 20, 25, 26 enterobactin 30, 47 excited states 227, 239, 240, 241, 242, 243, 245, 248, 250, 252 F0. See flavin, 7,8-didemethyl-8-hydroxy5-deaza FAD 1-carba-1-deaza- 15 FAD-binding domain 166, 167, 169, 183, 194 FDR 165, 166, 167, 169, 170, 172, 173, 176, 177, 181, 183, 188, 189, 192, 194, 196 Group 1 CXXXXC sequence motif 167 Group 1-fold 167 Group 3 SFXXC motif 167 femtosecond spectroscopy 393, 420, 428 Fenton reaction 30 ferredoxin 285, 290, 291, 292, 293, 294, 341, 342 NADP+ reductase 129, 131, 132 Ferredoxin-NADH-Reductase catalytic cycle 350 ferredoxin-NADP+ reductase 73, 74, 75, 78, 82, 83, 93 ferredoxin-NADP-reductase 341, 349 ferredoxin reductase 285, 290, 291, 292, 293, 294 ferricrocin 32, 33, 48 ferricyanide 146 ferritin 30 fish odor syndrome 59 flavin 7,8-didemethyl-8-hydroxy-5-deaza- 373 flavin adenine dinucleotide 321, 322, 329, 330, 331, 333 flavin C4a-alkoxide 15 flavin C4a-cysteinyl adduct 366, 368, 369 flavin C4a-(hydro)peroxide 51, 52 flavin C4a-hydroperoxide 1, 2, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 23 flavin C4a-hydroxide 1, 9, 10, 12, 13, 14, 19, 20, 22

431

flavin C4a-peroxide 1, 11, 13, 16, 19, 21, 22, 23 flavin-containing monooxygenase 16, 19, 20, 21, 22, 27 flavin-containing reductase 351 flavin-dependent monooxygenase 49, 256 flavin-dependent thymidylate synthase 204, 216, 218 flavin-dependent two-component systems flavin transfer mechanisms channeling or diffusion 269 flavin mononucleotide 321, 333 flavin reductase 51, 52, 64, 67 flavin as substrate 256 SsuE 255 flavin-specific monooxygenase 51, 52 flavocytochrome P450BM3 73, 74, 76, 79, 85, 93 flavodoxin 73, 74, 75, 76, 78, 79, 93, 229, 241, 285, 290, 291, 292, 293, 294, 322, 323, 324, 325, 326, 327, 328, 329, 330, 331, 332, 333, 342, 393, 394, 395, 396, 397, 398, 399, 400, 401, 402, 403, 405, 406, 407, 420, 424, 425 Flavoprotein Disulfide Reductases 165 flavoprotein monooxygenase 1, 2 Flavoprotein thiol/disulfide linked oxidoreductases 165 flavosemiquinone 236 fluorescence 14 fluorescence spectra 5 FMN 8-isopropyl 368 FMOs 59 FNR-Fd complex 285, 291, 293 FO conformation 183, 185 folate 234, 250, 373 FR conformation 183, 185 free radicals 9 FRET. See fluorescence resonant energy transfer fumarate 205, 206, 211 fumarate reductase 143, 144, 145, 146, 147, 148, 151, 152, 153, 154 FYX-051 111, 117, 118, 119, 124 GliT 169, 188 glucose oxidase 44, 50 glutathione amide reductase 176 glutathione reductase 171, 176, 179, 182, 192, 193, 196, 197, 199, 201 gout 104, 116, 118, 119, 124

432

Index

GR. See glutathione reductase Group 1-fold 168, 170, 171, 173, 192 halogenases 62, 63 bipyrrole homocoupling 64 sequential dichlorination 64 W box motif 55 Hammett relationship 15 helix-dipole 167 heme oxygenase 74, 87, 93 heteroatom hydroxylation reactions 58 HlmI 188, 199 HO. See heme oxygenase Hodgkin index 342 homodimer 166, 169, 174, 183, 185, 188, 194 HQ. See hydroquinone human tumour repressor 288 hydride transfer 6, 9 hydrogen peroxide 4, 6, 10, 11, 13, 20, 21 hydrogen-bond network 8, 9, 13, 14, 15, 25, 375, 377 (hydro)peroxyflavin 64 hydrophobicity 321, 323 hydroquinone 321, 325, 332 hydroxybenzoate 4-hydroxylase 3 hyperuricemia 104, 124 in conformation 6, 8, 9, 13, 14, 15 indole dioxygenase 74 interchange thiol 173, 183, 185, 188, 192 interface domain 166, 167, 169, 174, 183, 188, 193, 194, 195, 196 intersystem crossing 244 iron 29, 30, 32, 33, 34, 47, 48, 49 iron-sulfur 145, 146, 151 iron-sulfur cluster 205, 207, 212, 215 isofunctional proteins 342 IucD 33, 34, 35, 40, 42 Kelch repeat F-box 364, 371, 382 ketoadipate 2, 24 ketone 17 KIE. See kinetic isotope effect kinase 361, 363, 364, 368, 369, 370, 371, 381, 382, 383, 385 kinesin 283, 296 kinetic isotope effect 35, 214 Krebs cycle. See tricarboxylic acid cycle lactate dehydrogenase 288 lactoferrin 30

L-aspartate

oxidase 148 lignin 2 linear dichroism 228, 229, 230, 241, 242 lipoamide dehydrogenase 169, 173, 197, 198, 201 L-lysine monooxygenase 34, 42, 44 L-ornithine 29, 33, 34, 35, 37, 38, 41, 42, 43, 44, 45, 48 L-ornithine monooxygenase 33, 34 LOV domain 361, 362, 363, 364, 365, 366, 367, 368, 369, 370, 371, 375, 380, 381, 382, 383, 384, 385, 392, 419, 427 lysine hydroxylase 45 magnetoreception 371, 377, 387, 390 Marcus theory 353 master equation 352 MbsG 34, 40, 42, 44, 45, 46, See L-lysine monooxygenase membrane binding domain 76, 78, 79 menaquinone 151, 152, 154 mercaptans. See thiol mercuric ion reductase 167, 168, 170, 171, 173, 178, 180, 196, 197, 198 NmerA 180 methenyltetrahydrofolate 373 methionine synthase 73, 74, 93 methylmercaptan 366 Metropolis Monte Carlo 341 MICAL 2 Michael addition 216, 221 microscopic pKa value 346 microstate model 345 mitochondrial electron transport 143 molecular dynamics 396 molecular motor 280, 283, 284 molybdenum 125, 126, 127, 128, 129, 131, 133, 134, 135, 137, 138 molybdenum enzyme 125, 138 monooxygenase 1, 4, 15, 16, 17, 21, 22, 23, 24, 26, 27, 28, 59 Baeyer-Villiger BVMOs 58 Baeyer-Villiger oxidation 60 classification 53 CoA-activated substrates 58 enantioselective oxygenations 51 fingerprints 54 flavin-containing FMOs 58 from Rhodococcus jostii RHA1 genome 55 G box fingerprint 54

Index GD fingerprint 54 long-chain alkane hydroxylation 60 mechanistic aspects. See Chapters by Ballou; Sobrado; and Ellis N-hydroxylating NMOs 58 ornithine hydroxylases 58 oxidation and desulfurization of sulfonates 60 oxidation of aldehydes coupled with generation of bioluminescence. See Tu protein folds 53 reactions 52 regioselective oxygenations 51 Rossmann-fold domains 60 sequence fingerprints 54 siderophore NMOs 58 single component 56 subclass A and B 52 single-component enzymes 51 SsuD 255 subclass A 51 interaction with NAD(P)H 65 subclass A phylogenetic overview 57 subclass B 51 subclass B phylogenetic overview 59 subclass C 51 TIM-barrel fold 60 subclass D 51 acyl–CoA dehydrogenase fold 61 subclass E 51 styrene epoxidation 62 subclass F 51 halogenases 62 subclass F - halogenases 55 subclass F phylogenetic overview 63 two-component cofactor-independent 62 two-component enzymes 51 monooxygenase reactions Baeyer-Villiger oxidations 52 epoxidations 52 halogenations 52 hydroxylations 52 sulfoxidations 52 Mycobacterium tuberculosis 32, 48, 50 Mycobacterium tuberculosis LipDH 175 mycobactin 32, 33, 44, 45, 48, 50 mycobactin T 33 mycothione reductase 167, 177

433

NAD 2, 3, 4, 6, 20 NADH oxidases 169, 189 NADP 1, 4, 5, 9, 14, 16, 18, 19, 20, 22, 23, 27, 28, 349 NADPH 1, 4, 5, 6, 8, 13, 15, 16, 18, 21, 24 neutral semiquinone 366 neutravidin 286 N-hydroxylating monooxygenase 16, 20, 21, 22, 29, 33, 49 N-hydroxylation 29, 34 nitrate reductase 125, 126, 127, 128, 129, 130, 131, 132, 133, 134, 135, 136, 137, 138, 139, 140 nitric oxide synthase 73, 74, 76, 79, 93 NmerA 178, 183, 198 N-methyltryptophan oxidase 44, 50 NMO. See N-hydroxylating monooxygenase non-heme iron 142 non-ribosomal antibiotic dithiol oxidoreductases 169 novel reductase 1 73, 74 obligate homodimers 169 Old Yellow Enzyme 299–321 one-electron transfer 374 open conformation 6, 7, 14, 15 optical trap 283 Optical Tweezers 277, 279, 280, 282, 283, 284, 285, 296 ornithine monooxygenase 21 orotate 204, 207, 209, 210, 211, 289 out conformation 6, 7, 9, 14 oxidase 9, 11, 13, 15, 17, 19, 20, 21, 25, 27 oxidation and desulfurization of sulfonates. See Ellis oxygen 2, 4, 9, 10, 13, 16, 18, 20, 21, 23, 27, 28 oxygen atom transfer 126, 131, 138 OYE alkenes conjugated with esters as substrates 312 asymmetric alkene reductions 299 asymmetric organic synthesis 299 catalysts for industrial applications 299 charge transfer complexes 300 crystal structure 301 electron-deficient alkenes 301 enone and enal substrates 306 gelbe Ferment 300 His191 and Asn194 hydrogen bonds 302 nitro alkene substrates 315

434

Index

Saccharomyces pastorianus 299 stereoselective reductions 299 stereoselectivity 299 substrate scope 299 substrate specificity 305 Thr37 302 Trp116 affects stereoselectivity 304 Tyr196 proton donor 302 W116I and W116F mutants altered stereochemical outcome 304 P450 catalysis 85, 86, 87 p53. See human tumor repressor PAMO. See phenylacetone monooxygenase para-hydroxybenzoate hydroxylase 56 partition function 345 PAS domains 361, 364, 365, 369, 382, 383 peptide synthetase 33 peroxidase 166, 169, 175, 176, 189, 190, 192, 197, 198, 201 peroxiredoxin reductase 169, 187 phenazine ethosulfate 146 phenazine methosulfate 146 phenol 18 phenolate 8, 13, 14 phenol hydroxylase 3, 4, 9, 10, 14, 15, 24, 25, 26 phenylacetone monooxygenase 18, 22, 23, 40, 50 from Thermobifida fusca 55 phenyl boronic acids 18 phosphodiesterase 363 phosphorylation 369, 373, 375, 385, 386, 388 photochemistry 225, 227, 236, 244, 245, 246, 251 photocycle 366, 368, 369, 381, 383, 384, 385, 390 photoexcited states 225 photoinduced electron transfer 234, 235, 236, 244, 245 photolyase homology region 419 photolyases 226, 227, 249 photoreceptor 361, 362, 363, 364, 366, 368, 370, 371, 375, 377, 378, 379, 380, 381, 382, 383, 384, 385, 386, 387, 388, 389, 390, 391, 392 photosynthesis 225 phototropin 361, 363, 364, 365, 366, 368, 369, 370, 383, 384, 386, 393, 419, 421, 423, 427 phototropism 363, 364, 370, 382, 386

p-hydroxybenzoate 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 23, 24, 25, 26 p-hydroxybenzoate hydroxylase 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 20, 23, 24, 25, 26, 35, 49, 289, 290, 297 pH titration 344 PhzS 2, 24 pKa value 346 Poisson-Boltzmann equation 322, 337 Poisson-Boltzmann equation, linearized 338 polarizability 227, 230, 231, 249, 336 polarization 336 prostacyclin synthase 90 proteasome 378, 379 protein unfolding 278, 284 protonation 349 proton-coupled electron transfer 321 proton transfer 393, 414, 416, 417, 418, 420, 421, 427, 428 Pseudomonas aeruginosa 30, 32, 33, 48, 49, 50 PvdA 34, 35, 37, 38, 40, 41, 42, 43, 44, 45, 49 pyochelin 30 pyocyanin 2, 24 pyoverdin 30, 34, 48 pyridine nucleotide 165, 167, 169, 172, 173, 183, 185, 186, 187, 188, 191, 199 pyridine nucleotide-dependent 165, 169 Pyrimidines 203, 216 quantum yield 362, 366, 373, 380 quinone 321, 322, 325 radical 361, 366, 374, 375, 376, 377, 384, 386, 387, 388, 389, 390, 391 radical pair 366, 374, 375, 377 rate constant 352 reaction field 336, 337, 339 reactive oxygen species 30, 45, 104 redox active cysteine disulfide 165, 167 redox titration 344 reductase 1, 3, 6, 10 reduction potential 6 regio 51 regioselective 56 regioselective bromination 62 regioselective chlorination 62 regioselective hydroxylation 56 reorganization energy 353 rhodanese domain 190, 192

Index riboflavin 5-deaza-riboflavin 373 Rieske protein 135 RNA 213 Rossmann-fold 167 ruler mechanism 29, 42, 46 rupture force 286, 287, 288 selenide 1, 2 selenocysteine 165, 167, 179, 181 semiquinone 10, 210, 212, 229, 236, 239, 240, 241, 244, 253, 321, 332, 333, 351, 361, 373, 374, 375, 377, 389, 390 SidA 34, 35, 37, 38, 40, 41, 42, 43, 44, 45, 49 siderophore 20, 21, 29, 30, 32, 33, 34, 45, 46, 47, 48, 49, 50 siderophore A 34, 49 similiarity index 342 single-electron transfer 10, 11, 19, 205, 210 single-molecule spectroscopy 15 Single-molecule techniques 277, 279, 289 singlet-state 374, 377 small angle X-ray scattering study 181 SMFS. See single-molecule force spectroscopy SNF1-related kinase 134, 140 solvatochromism 228, 231, 243, 248 solvent kinetic isotope effect 368 solvent screening 336 sqr 166, 194, 195, 196, See succinate dehydrogenase SQRs 169, 194, 195 SQ. See semiquinone squalene monooxygenase 74, 87, 91 SsuD modeling of octanesulfonate substrate 264 pH dependence for kcat and kcat/Km 267 SsuD enzyme TIM-barrel structure 259 SsuD flavin binding site superimposition with LadA 264 SsuD loop deletion variants 263 SsuD structure insertion regions 261 SsuE alteration in the kinetic mechanism 270 equilibrium ordered mechanism 259 SsuE and SsuD complex pull-down and cross-linking experiments 270

435

SsuE enzyme ordered sequential mechanism 258 ssu operon SsuE and SsuD 256 Stark spectroscopy 225, 226, 229, 230, 231, 232, 236, 238, 243, 246, 248 stereoselective 56 STM. See scanning tunnelling microscopy streptavidin 286 styrene monooxygenases SMO 56 subclass E 56 substrate inhibition 14 substrate-induced electrochromism 235, 236 succinate dehydrogenase 141, 142, 143, 144, 147, 148, 151, 155 succinate oxidase 147 sulfenic acid 18, 20 sulfhydryl oxidases 188 sulfide 19 quinone oxidoreductases 169 sulfite oxidase 125, 126, 127, 128, 131, 135, 136, 138, 139, 140 sulfite reductase 73 sulfonate uptake 256 sulfoxidation 58 synaptobrevin 288 syntaxin 288 tau operon 256 taurine dioxygenase (TauD) 256 tautomerization 14 thioether 1, 2 thiol 1, 2, 18, 20, 366 thiol/disulfide interchange 172 thiol-disulfide reductase 22, 23 thioredoxin 165, 166, 167, 168, 169, 170, 177, 179, 181, 182, 183, 184, 187, 189, 196, 197, 198, 199, 201 thioredoxin-like fold 3 thioredoxin reductase 166, 167, 168, 169, 170, 177, 179, 181, 182, 183, 184, 187, 196, 198, 199 thromboxane A2 synthase 90 thymine 203, 204, 205, 214, 215, 216 TIM-barrel 205, 206, 213 time-dependent density functional theory 236, 238, 239, 240, 241, 242 titration curve 344 transferrin 30, 288

436

Index

transition dipole moment 227, 228, 229, 230, 234, 240, 241, 249 triacetylfusarinine C 32, 33 tricarboxylic acid cycle 143, 147, 157 trimethylcyclopentenylacetyl-coenzyme A monooxygenase 2-oxo-Ô-4,5,5- 22, 23 triplet state 361, 366, 367, 377, 383 TrmFO 204, 216, 218, 223 tRNA 204, 205, 213, 214, 218, 221, 222, 223 TrxR 167, 177, 181, 182, 183, 184, 185, 186, 187, 188 trypanothione reductase 167, 168, 177, 198, 201 tryptophanyl cation radical 374 two-component enzyme system 255 two-component monooxygenases 2, 17 ubiquinone 142, 151, 152, 154, 205, 206, 210 ultrafast laser spectroscopy 225

ultrafast spectroscopy 227 uracil 203, 204, 205, 213, 214, 215, 216, 218 vicibactin synthase O 34 xanthine dehydrogenase 103, 104, 105, 110, 111, 114, 115, 116, 326, 327, 328, 329, 330, 332 xanthine oxidase 103, 104, 111, 112, 114, 115, 116, 121, 123, 326, 327, 328, 329, 330, 332 xanthine oxidoreductase 103, 104, 105, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119, 120 yersiniabactin 33 YUCCA 16, 21, 22 zeitlupe protein 364, 371