Structure and Function of Iron-Sulfur Cluster Biosynthesis

STRUCTURE AND FUNCTION OF IRON-SULFUR CLUSTER BIOSYNTHESIS

PROTEINS AND THE INFLUENCE OF OXYGEN LIGATION

DISSERTATION

Presented in Partial Fulfillment of the Requirements for

the Degree Doctor of Philosophy in the Graduate

School of The Ohio State University

Sheref S. Mansy, B.S.

The Ohio State University 2003

Dissertation Committee:

Dr. James A. Cowan, Adviser

Dr. Ming-Daw Tsai

Dr. Venkat Gopalan Adviser

Dr. Mark P. Foster Ohio State Biochemistry Program

ABSTRACT

Members of the IscU family of proteins are among the most conserved of all

protein groups, extending across all three kingdoms of life. IscU is believed to be

involved in iron-sulfur cluster delivery to apo iron-sulfur proteins. However, most of the

evidence supporting the function of IscU stems from genetic and cellular biological

studies. Therefore, we set out to biochemically characterize human, yeast, and

prokaryotic IscU proteins. A variety of spectroscopic techniques were used to evaluate

IscU including, UV-visible absorption, Mössbauer, near- and far- UV circular dichroism,

mass spectrometry, atomic absorption, and nuclear magnetic resonance. Herein we demonstrate that IscU proteins coordinate reductively labile [2Fe-2S]2+ centers and are

capable of mediating delivery of intact cluster to apo protein targets. Furthermore,

extensive structural and dynamic data of a hyperthermophilic homologue, Thermotoga

maritima IscU, revealed that IscU adopts a mobile molten globule-like state that is vastly

different from the previously identified ferredoxin-like fold that has thus far been characterized for other metallochaperones. Such a dynamic molecule may allow for the flexibility that is necessary for the multiple roles of Fe-S cluster assembly, and

ii recognition and delivery of that cluster to a target protein. Additionally, we utilized X- ray crystallography to elucidate a high resolution structure of an oxygen ligated [4Fe-4S] high potential iron protein.

iii

To my mother and father, Wafeya and Samir

ACKNOWLEDGMENTS

I thank my adviser, J. A. Cowan, for expert intellectual guidance and for providing a highly stimulating and professional research environment. I am also grateful to those that have served my on dissertation committee, Ming-Daw Tsai, Venkat

Gopalan, Donald H. Dean, and Mark P. Foster.

One of the enjoyable aspects of research is the opportunity to collaborate with other laboratories and researchers. I thank Russ Hille and Craig Hemann for EPR and resonance Raman experiments, Kari Green-Church for mass spectrometry, Jon-David

Sears and Don Ordaz for fermentations, In-Ja L. Byeon for NMR training, M.

Sundaralingam and Yong Xiong for X-ray crystallography, Kristene K. Surerus for

Mössbauer spectroscopy (University of Wisconsin, Milwaukee), and Marco Sola for electrochemistry (University of Modena and Reggio Emilia, Italy).

My time in Italy would have been far less productive and enjoyable had it not been for Cristina Del Bianco, Rainer Küemmerle, Mariapina D'Onofrio, Karel Kubicek, and Fiorenza Cramaro. Also, I value the opportunities to have worked with people in the

Cowan laboratory, including Anjali Patwardhan, Gong Wu, Shu-pao Wu, Taejin Yoon,

Matthew W. Foster, Chun-An Chen, Manunya Nuth, and Sreedhara Alavattam.

v I greatly appreciate the funding that I received from the Chemistry-Biology Interface

Program, the Ohio State Biochemistry Program, and the Ohio State University.

VITA

February 17, 1975 Born, Eugene, Oregon, USA

1997 B.S. Microbiology, Ohio State University

1997 American Heart Association Undergraduate Student Summer Fellowship, Ohio State University

1997-1998 Graduate Research Associate, Ohio State University

1998-2001 NIH Chemistry - Biology Interface Training Grant, Ohio State University

2001-2002 Graduate Research Associate, Ohio State University

2003-present Presidential Fellowship, Ohio State University

PUBLICATIONS

1. Sheref S. Mansy, John S. Olson, Gonzalo Gonzalez, and Marie A. Gilles- Gonzalez (1998) Imidazole is a Sensitive Probe of Steric Hindrance in the Distal Pockets of Oxygen-Binding Heme Proteins. Biochemistry 37, 12452-12457.

2. Weimin Gong, Bing Hao, Sheref S. Mansy, Gonzalo Gonzalez, Marie A. Gilles- Gonzalez, and Michael K. Chan. (1998) Structure of a Biological Oxygen Sensor: A New Mechanism for Heme-Driven Signal Transduction. Proc. Natl. Acad. Sci. USA 95, 15177-15182.

vii

3. Matthew W. Foster, Sheref S. Mansy, Jungwon Hwang, James E. Penner-Hahn, Kristene K. Surerus, and J. A. Cowan. (2000) A Mutant Human IscU Protein Contains a Stable [2Fe-2S]2+ Center of Possible Functional Significance. J. Am. Chem. Soc. 122, 6805-6806.

4. Sheref S. Mansy, Yong Xiong, Craig Hemann, Russ Hille, M. Sundaralingam, and J. A. Cowan. (2002) Crystal Structure and Stability Studies of C77S HiPIP: A Serine Ligated [4Fe-4S] Cluster. Biochemistry 41, 1195-1201.

5. Gong Wu, Sheref S. Mansy, Shu-pao Wu, Kristene K. Surerus, Matthew W. Foster and J. A. Cowan. (2002) Characterization of an Iron-sulfur Cluster Assembly Protein (ISU1) from Schizosaccaromyces pombe. Biochemistry 41, 5024-5032.

6. Gong Wu, Sheref S. Mansy, Craig Hemann, Russ Hille, Kristene K. Surerus, and J. A. Cowan. (2002) Iron-Sulfur Cluster Biosynthesis: Characterization of Schizosaccaromyces pombe Isa1 J. Biol. Inorg. Chem. 7, 526-532.

7. Sheref S. Mansy, Gong Wu, Kristene K. Surerus, and J. A. Cowan. (2002) Iron- Sulfur Cluster Biosynthesis: Thermotoga maritima IscU is a Structured Iron-Sulfur Cluster Assembly Protein. J. Biol. Chem. 277, 21397-21404.

8. Ivano Bertini, J. A. Cowan, Cristina Del Bianco, Claudio Luchinat, and Sheref S. Mansy. (2003) Thermotoga maritima IscU. Structural Characterization and Dynamics of a New Class of Metallochaperone. J. Mol. Biol. 331, 907-924.

FIELD OF STUDY

Major Field: Biochemistry

viii

TABLE OF CONTENTS

Abstract...... ii

Dedication...... iv

Acknowledgments...... v

Vita...... vii

List of Tables...... xiv

List of Figures...... xv

List of Abbreviations...... xix

Chapters:

1 Introduction...... 1 1.1 Iron-Sulfur Cluster Proteins...... 1 1.1.1 Types of Protein Bound Iron-Sulfur Clusters...... 1 1.1.2 Importance of Iron-Sulfur Cluster Proteins...... 3 1.1.3 Redox Function...... 4 1.1.4 Enzymatic...... 4 1.1.5 Sensing...... 5 1.2 Metallochaperones...... 6 1.2.1 Metallochaperone Function...... 6 1.2.2 Copper Metallochaperone Structure...... 7 1.2.3 Non-Copper Metallochaperones...... 9 1.2.4 Necessity of Iron-Sulfur Cluster Metallochaperones...... 9 1.3 Iron-Sulfur Cluster Assembly...... 10 1.3.1 Nitrogen Fixation...... 10 1.3.2 Identification of Bacterial ISC Proteins...... 11

ix 2 Characterization of Human and Yeast ISC Proteins...... 13 2.1 Introduction...... 13 2.2 Materials and Methods...... 16 2.2.1 General Chemicals...... 16 2.2.2 Protein Expression...... 16 2.2.2 Human ISU Purification...... 17 2.2.3 S. pombe ISU1 and ISA1 Purification and Reconstitution...... 18 2.2.4 Wild Type S. pombe Ferredoxin Purification...... 19 2.2.5 Cys to Ser Substituted S. pombe Ferredoxin Purification...... 20 2.2.6 Human Ferredoxin Purification...... 20 2.2.7 Mutagenesis...... 21 2.2.8 UV-Visible Absorption Spectroscopy...... 24 2.2.9 Circular Dichroism...... 24 2.2.10 Native Polyacrylamide Gel Electrophoresis...... 24 2.2.11 EDC Cross-Linking...... 25 2.3 Results and Discussion...... 25 2.3.1 Human ISU Expression...... 25 2.3.2 Holo D37A Hs ISU Iron-Sulfur Characterization...... 26 2.3.3 Cys to Ala Mutagenesis...... 29 2.3.4 ISU - Ferredoxin Cross-Linking...... 31 2.3.5 ISU - Mutant Human Ferredoxin Cross-Linking...... 36 2.3.6 Native-Page of Mutant Ferredoxins...... 40 2.3.7 S. pombe ISA1 Cys Mutagenesis...... 42 2.3.8 Sp ISA1 Interaction with Human Ferredoxin...... 43 2.3.9 C - Terminal Deletion of D37A Hs ISU...... 47 2.3.10 Cys to Ser S. pombe Ferredoxins...... 47

3 Cloning and Characterization of Thermotoga maritima IscU...... 49 3.1 Introduction...... 49 3.2 Materials and Methods...... 51 3.2.1 General Chemicals...... 51 3.2.2 Cloning of T. maritima IscU...... 51 3.2.3 Mutagenesis...... 53 3.2.4 T. maritima IscU Over-Expression...... 54 3.2.5 Protein Purification...... 54 3.2.6 Mass Spectrometry...... 55 3.2.7 T. maritima IscU Cluster Reconstitution...... 58 3.2.8 Size Exclusion Chromatography...... 58 3.2.9 Dynamic Light Scattering...... 59 3.2.10 DSS Cross-Linking...... 59 3.2.11 UV-Visible Spectroscopy and Evaluation of Extinction Coefficients...... 59 3.2.12 Temperature Dependence of [2Fe-2S] Cluster Stability...... 60 3.2.13 Mössbauer Spectroscopy...... 61

x 3.2.14 EPR Spectroscopy...... 61 3.2.15 Nuclear Magnetic Resonance...... 61 3.2.16 Iron Quantitation...... 62 3.2.17 Circular Dichroism...... 62 3.2.18 D40A Tm IscUHis - Human Fd EDC Cross-Linking...... 63 3.2.19 D40A Tm IscU Cluster Transfer to Apo Hs Fd...... 63 3.3 Results...... 63 3.3.1 Cloning...... 63 3.3.1 Protein Expression...... 64 3.3.2 Mass Spectroscopic Analysis...... 67 3.3.3 Tm IscU Aggregation State...... 67 3.3.4 Cluster Coordination...... 69 3.3.5 Mössbauer and EPR Spectroscopy...... 72 3.3.6 UV-Visible Spectroscopy...... 74 3.3.7 Thermal Stability of the [2Fe-2S] Cluster...... 75 3.3.8 Circular Dichroism...... 77 3.3.9 NMR...... 78 3.3.9 Cluster Transfer to Human Ferredoxin...... 82 3.3.10 Sequence Comparison...... 82 3.3.11 D55A Tm IscU...... 85 3.4 Discussion...... 86 3.4.1 Protein Expression...... 86 3.4.2 Cluster Coordination...... 87 3.4.3 Cluster Stoichiometry...... 89 3.4.4 Thermal Stability...... 89 3.4.5 T. maritima IscU Reconstitution of Apo Human Ferredoxin..90 3.4.6 Structural Considerations...... 91

4 Nuclear Magnetic Resonance of Thermotoga maritima IscU...... 93 4.1. Introduction...... 93 4.2 Materials and Methods...... 95 4.2.1 General Chemicals...... 95 4.2.2 Protein Expression and Purification...... 95 4.2.3 NMR Spectroscopy...... 96 4.2.4 RDC...... 97 4.2.5 Secondary Structure Determination...... 99 4.2.6 15N Backbone Relaxation Measurements...... 99 4.2.7 Hydrogen Exchange...... 100 4.2.8 Data Processing and Structure Calculations...... 100 4.3 Results and Discussion...... 101 4.3.1 Solution Conditions...... 101 4.3.2 Chemical Shift Dispersion and Resonance Assignments...... 102 4.3.3 Secondary Structure...... 122 4.3.4 NOE Analysis...... 128

xi 4.3.5 Hydrogen Exchange...... 131 4.3.6 Holo D40A Tm IscU Spectra...... 134 4.3.7 Apo and Holo D40A Tm IscU Dynamics...... 139 4.3.8 Structural Features of IscU...... 146 4.3.9 Fluxionality of Apo Tm IscU...... 151 4.3.10 Comparison of the Structure and Dynamics of IscU with Other metallochaperones...... 155

5 Characterization and Relevance of the Molten Globule State of Thermotoga maritima IscU...... 158 5.1. Introduction...... 158 5.2 Materials and Methods...... 160 5.2.1 General Chemicals...... 160 5.2.2 ANS Binding...... 160 5.2.3 Near-UV CD...... 161 5.2.4 Dynamic Light Scattering...... 161 5.2.5 Free Energy of Unfolding...... 162 5.2.6 Limited Tryptic Digestion...... 162 5.2.7 Cloning of T. maritima dnaK...... 163 5.3 Results...... 165 5.3.1 ANS Binding...... 165 5.3.2 Near-UV CD...... 169 5.3.3 Dynamic Light Scattering...... 169 5.3.4 Free Energy of Unfolding...... 171 5.3.5 Tryptic Digest...... 175 5.3.6 T. maritima dnaK Cloning...... 178 5.3.7 DnaK Expression and Preliminary Purification...... 180 5.4 Discussion...... 182

6 Structural and Stability Effects of Serine Oxygen Ligation of a High Potential [4Fe-4S] Protein...... 190 6.1 Introduction...... 190 6.2 Materials and Methods...... 195 6.2.1 Protein Purification...... 195 6.2.2 WT and C77S HiPIP Crystallization...... 196 6.2.3 C77S HiPIP Crystal Data Collection...... 197 6.2.4 C77S HiPIP Structure Calculation...... 197 6.2.5 Resonance Raman Spectroscopy...... 198 6.2.6 Cluster Stability...... 199 6.2.7 Near-UV-Visible Circular Dichroism...... 199 6.3 Results...... 200 6.3.1 Cystallization of WT and C77S HiPIP...... 200 6.3.2 Protein Conformation and [4Fe-4S] Cluster Coordination...... 203 6.3.3 The Aromatic Core Surrounding the [4Fe-4S] Cluster...... 208

xii 6.3.4 Backbone Amide Hydrogen Bonds and Water Content...... 211 6.3.5 Optical Spectroscopy...... 212 6.3.6 Resonance Raman Spectroscopy...... 214 6.3.7 C77S HiPIP Cluster Acid Sensitivity...... 218 6.4 Discussion...... 219 6.4.1 Fe-S Cluster Associated Bond Lengths...... 219 6.4.2 Role of Phe48 in Solvent Protection...... 221 6.4.3 Electronic Changes in C77S HiPIP...... 223 6.4.4 Acid-Catalyzed Cluster Degradation...... 223 6.4.5 Conclusions...... 225

References...... 226

xiii

LIST OF TABLES

2.1. Primers Used for Quikchange Mutagenesis...... 23

4.1. Acquisition Parameters for NMR Data Sets Collected on Tm IscU...... 98

4.2. Nitrogen and Proton Chemical Shift Assignments of Apo D40A Tm IscU...... 107

4.3. Carbon Chemical Shift Assignments of Apo D40A Tm IscU...... 115

4.4. 3J(HN,Hα) Coupling Constants of Apo D40A Tm IscU...... 126

4.5. 1H-15N Residual Dipolar Couplings of Apo D40A Tm IscU...... 127

4.6. Hydrogen Exchange...... 132

5.1. Comparison of Denaturation Data for Molten Globule and Non-Molten

Globule Proteins...... 173

6.1. Crystal Data and Refinement Statistics of C77S HiPIP...... 202

6.2. Comparison Between C77S and WT HiPIP Mean Fe-S Cluster Bond

Lengths and Dihedral Angles...... 206

6.3. Resonance Raman Frequencies for WT and C77S HIPIPs ...... 218

xiv

LIST OF FIGURES

1.1 Common Protein Bound Fe-S Clusters...... 3

1.2 Ribbons Diagram of the Solution Structure of Enterococcus hirae CopZ...... 8

2.1 Sequence Alignment of ISU/NifU Proteins...... 28

2.2 Absorption Spectra of Hs ISU...... 29

2.3 Cys Substituted D37A Hs ISU Electrophoresis...... 32

2.4 Yeast and Human ISU EDC Cross-Linking with Hs Fd...... 34

2.5 Helix C of Bovine Ferredoxin with Carboxylate Side-Chains Labeled...... 37

2.6 Spectra of WT and Mutant Ferredoxins and their Interaction with Hs ISU...... 38

2.7 Native-PAGE of D37A Hs ISU and Hs Fd...... 41

2.8 UV-visible Spectrum of WT Sp ISA1...... 43

2.9 Sp ISA1 EDC Cross-Linking with Target Proteins...... 45

2.10 UV-visible spectrum of D37A ∆CT Hs ISU after Chemical Reconstitution...... 48

3.1 Cloning of T. maritima iscU...... 65

3.2 WT and D40A Tm IscU/His Expression...... 66

3.3 Aggregation State of Tm IscU...... 70

3.4 UV-visible Spectra of Apo and Holo D40A Tm IscU...... 72

3.5 D40A Tm IscU Mössbauer Spectra at 4.2, 100, and 200 K...... 74

xv 3.6 Cluster Thermal Stability of D40A Tm IscU and Human Ferredoxin...... 76

3.7 Circular Dichroism Spectra of D40A Tm IscU and Hs Fd...... 79

3.8 1H-15N-HSQC of 4 mM WT and D40A Apo Tm IscU in 50 mM Sodium

o Phopsphate, pH 7.0, 10% D2O at 20 C...... 80

3.9 D40A Tm IscU - Hs Fd EDC Cross-Linking and Cluster Transfer...... 83

3.10 IscU/NifU Sequence Alignment...... 84

3.11 UV-visible Spectrum of D55A Tm IscUHis...... 86

4.1 1H-15N HSQC of Apo D40A Tm IscU in 450 mM NaCl, pH 5.4 at 45 oC

and 700 MHz...... 103

4.2 2D Planes from 3D CBCANH and CBCA(CO)NH Spectra of

Apo D40A Tm IscU...... 106

4.3 Sequence Alignment of IscU Proteins Showing the Determined Secondary

Structure of Apo D40A Tm IscU...... 124

4.4 2D NOESY of Apo D40A Tm IscU in 450 mM NaCl, pH 5.4, 45 oC

at 700 MHz...... 129

4.5 Schematic Representation of NOE Connectivities...... 130

4.6 1H-15N HSQC of holo D40A Tm IscU in 200 mM NaCl, pH 7.4 at 20 oC

and 500 MHz...... 135

4.7 2D NOESY of Apo and Holo D40A Tm IscU in 200 mM NaCl, pH 7.0,

20 oC at 700 MHz...... 137

4.8 R1, R2, and Heteronuclear NOEs for Apo D40A Tm IscU in 450 mM NaCl,

pH 5.4, 45 oC at 500 MHz...... 140

xvi 4.9 Theoretical R1 and R2 Values on the Basis of the Stokes-Einstein Equation

at 20 °C and 45 °C for Monomeric and Dimeric apo D40A Tm IscU...... 141

4.10 R1 and R2 for Apo and Holo D40A Tm IscU in 200 mM NaCl, pH 7.4

at 20 °C and 500 MHz...... 144

4.11 Orientation of Tm IscU Secondary Structural Elements...... 150

5.1 ANS Fluorescence...... 167

5.2 Near-UV CD of WT and D40A Tm IscU in 100 mM Tris-HCl,

200 mM NaCl, pH 7.4 with and without 7.4 M Gdn-HCl...... 170

5.3 Gdn-HCl Denaturation of WT and D40A Tm IscU...... 172

5.4 Tryptic Digestion of WT and D40A Apo Tm IscU...... 176

5.5 Cloning of dnaK...... 179

5.6 Expression and Purification of DnaK/His...... 181

6.1 C. vinosum HiPIP with the Aromatic Side-Chains Surrounding the

[4Fe-4S] Cluster...... 193

6.2 WT and C77S HiPIP Crystals...... 201

6.3 The Association of the Four Independent Molecules in the Asymmetric Unit....203

6.4 Superposition of the Cα Backbones of C77S and WT HiPIP...... 205

6.5 C77S and WT HiPIP Cluster and Cluster Ligands...... 208

6.6 Comparison of C77S and WT HiPIP Cluster, Cluster ligands, and

Surrounding Hydrophobic Residues...... 210

6.7 Space Filling Model of WT C. vinosum HiPIP...... 211

6.8 UV-Visible Spectra of WT and C77S Reduced and Oxidized HiPIP...... 213

xvii 6.9 Near-UV-Visible CD of WT and C77S HiPIP...... 214

6.10 Resonance Raman Spectra of Reduced WT and C77S HiPIP with

458 nm and 514 nm Excitation...... 216

6.11 Resonance Raman Spectra of WT and C77S HiPIPs in the Reduced

and Oxidized states...... 217

6.12 Acid Catalyzed [4Fe-4S] Degradation...... 219

xviii

LIST OF ABBREVIATIONS

ANS 1-anilino-8-naphthalenesulfonic acid

ATP Adenosine 5'-triphosphate bp Base Pair

BSA Bovine Serum Albumin

CAPS (3-[cyclohexylamino]-1-propanesulfonic acid)

CD Circular Dichroism dH2O Distilled Water

DNA Deoxyribonucleic Acid dNTP Deoxynucleotide 5'-triphosphate

DTT Dithiothreitol

EDTA Ethylenediaminetetraacetic Acid

EDC 1-ethyl-3-(dimethylaminopropyl)carbodiimide hydrochloride

EPR Electron Paramagnetic Resonance

ESI Electrospray Ionization

ExPASy Expert Protein Analysis System

Fd Ferredoxin

FPLC Fast Protein Liquid Chromatography

xix Gdn-HCl Guanidinium Hydrochloride

Gdn-SCN Guanidinium Thiocyanate

HEPES (N-[2-hydroxyethyl]piperazine-N'-[2-ethanesulfonic acid])

HiPIP High Potential Iron Protein

HSQC Heteronuclear Single Quantum Coherence

IPTG Isopropyl Thiogalactoside

ISC Iron-Sulfur Cluster

LB Luria-Bertani

MALDI Matrix Assisted Laser Desorption/Ionization

MPD 2-methyl-2, 4-pentanediol

MW Molecular Weight

Nif Nitrogen Fixation

NMR Nuclear Magnetic Resonance

NOE Nuclear Overhauser Effect

NOESY Nuclear Overhauser Effect Spectroscopy

NTA Nitrilotriacetic Acid

ORF Open Reading Frame

PAGE Polyacrylamide Gel Electrophoresis

PCR Polymerase Chain Reaction

PMSF Phenylmethylsulfonyl Fluoride

RNA Ribonucleic Acid

RC Reaction Center

xx RDC Residual Dipolar Coupling

SDS Sodium Dodecyl Sulfate

TBE 1,1,2,2-tetrabromoethane

TFE 2,2,2-Trifluoroethanol

TOF Time of Flight

Tris Trihydroxyethylene Amine tRNA Transfer RNA

UV Ultraviolet

WT Wild Type

xxi

CHAPTER 1

INTRODUCTION

1.1 IRON-SULFUR CLUSTER PROTEINS

1.1.1 TYPES OF PROTEIN BOUND IRON-SULFUR CLUSTERS

There are three common types of iron-sulfur cluster (Fe-S) proteins, including mononuclear or rubredoxin-like, [2Fe-2S], and [4Fe-4S] containing proteins (Figure 1.1).

Typically, Fe-S clusters are ligated by four cysteine residues with additional coordination of each iron provided by inorganic sulfides. The exception is rubredoxin, with its single iron atom completely coordinated by the Sγ of Cys side-chains. Although protein ligation is typically provided by Cys residues, there are examples of non-Cys ligation to

Fe-S clusters. For instance, the [2Fe-2S] cluster of the Rieske protein is coordinated by two Cys and two His side-chains (1, 2). Additionally, [4Fe-4S] clusters can be coordinated by three Cys and one non-Cys ligand, as seen for aconitase. Aconitase catalyzes the stereospecific interconversion of citrate and isocitrate in the tricarboxylic

1 acid cycle (3, 4). Unlike the tetrahedral geometry of iron atoms in typical Fe-S clusters, the non-Cys ligated iron of aconitase is hexacoordinate, bonded to three inorganic sulfides from the cluster, one oxygen from water, and two oxygens (hydroxyl and carboxyl oxygen atoms) from the substrate (5). In the substrate unbound form, the coordination sphere of the non-Cys ligated iron is completed by solvent molecules. Due to the unique coordination properties of oxygen ligated [4Fe-4S] clusters, they possess unusual characteristics, such as the ability to easily lose the oxygen coordinated Fe under mildly oxidizing conditions resulting in a [3Fe-4S] cluster (Figure 1.1D). Also, [3Fe-4S] clusters have been observed to undergo rearrangement at high pH to a linear [3Fe-4S] cluster (6) with some evidence supporting physiological relevance (7), although this has yet to be confirmed. In addition to solvent or substrate ligation, i.e. non-protein coordination, [4Fe-4S] clusters can be coordinated by an Asp side-chain (8). However, even though oxygen ligation to [4Fe-4S] clusters does occur naturally, no conclusive

evidence exists for analogous oxygen coordination to [2Fe-2S] clusters. This is

surprising since [2Fe-2S] clusters can be coordinated by three Cys and one Ser, as has

been shown by mutagenesis techniques (9).

Figure 1.1. Common protein bound Fe-S clusters. A) Mononuclear or rubredoxin-like

Fe-S cluster. B) [2Fe-2S] cluster. C) [4Fe-4S] cluster. D) [3Fe-4S] cluser. Iron is shown in green, sulfur in yellow, and Cys side-chains in blue. Figures were generated with MOLMOL using the PDB coordinates 1FHH (A), 1CJE (B), and 1HIP (C and D).

The [3Fe-4S] was created by removing one of the iron atoms from the [4Fe-4S] cluster.

1.1.2 IMPORTANCE OF IRON-SULFUR CLUSTER PROTEINS

Iron is one of the most abundant elements in the earth's crust. Therefore, from an evolutionary perspective, iron is an attractive candidate for being one of nature's first

3 catalysts, particularly since many Fe-S proteins provide several physiologically essential

roles in both lowly and highly evolved organisms (5). For example, Fe-S proteins are

responsible for the conversion of N2 to NH3, fixation of carbon monoxide, reduction of

nitrite and sulfite, oxygenic phosphorylation, dehydration reactions, and mitochondrial

hydroxylation (10-12). In addition to enzymatic catalysis, Fe-S clusters are important for structural stabilization of tertiary folds, electron transfer reactions, and physiological sensing - signal transduction mechanisms (13).

1.1.3 REDOX FUNCTION

The mitochondrial electron transfer chain (ETC) contains 13 - 14 Fe-S cluster containing proteins (5), used to provide aerobic organisms with energy in the form of

ATP. Fe-S proteins are ideally suited for such electron transfer pathways, since they cover a wide range of redox potentials from less than 600 mV for [7Fe-8S] proteins to greater than 400 mV for high potential iron proteins (5). This large redox potential range is greater than that of any other simple cofactor involved in electron transfer reactions

(12).

1.1.4 ENZYMATIC

Fe-S clusters can directly participate in catalysis, as seen for the [4Fe-4S] cluster of aconitase, which catalyzes the interconversion of citrate and isocitrate in the tricarboxylic acid cycle (3, 4). More specifically, the non-Cys ligated iron of the cluster acts as a Lewis acid catalyzing the removal of a proton and a hydroxyl from the substrate

4 (5). Other enzymes utilize the [4Fe-4S] cluster as a primary electron donor to initiate free

radical reaction mechanisms (5), such as those of anaerobic ribonucleotide reductase

(14), pyruvate formate lyase (15, 16), lysine amino mutase (17), and biotin synthase (18).

1.1.5 SENSING

Several Fe-S proteins serve cellular sensory roles. Sensing can be achieved by a

change of redox state of an intact protein-bound Fe-S cluster, or through more elaborate

mechanisms of cluster degradation. The SoxR/SoxS two-component system of

Escherichia coli is a classical example of redox sensing (19). SoxR is a [2Fe-2S]

containing transcriptional activator that regulates gene expression in response to

superoxide and nitric oxide (19, 20). Under normal physiological redox conditions SoxR

is inactive with an overall Fe-S cluster charge of 1+. However, the cluster becomes oxidized in the presence of superoxide or nitric oxide, resulting in the activation of SoxR.

Activated SoxR binds to the soxS promoter, which in turn triggers a series of physiological changes resulting in the synthesis of proteins involved in oxidative stress protection, such as superoxide dismutase (19).

The aconitase/iron regulatory protein (IRP) system is an example of Fe-S coordination dependent sensing. Holo cytoplasmic aconitase contains a [4Fe-4S] cluster that performs a physiologically essential role (see 1.1.4). However, under low cellular iron concentrations holo aconitase loses its Fe-S cluster and gains the ability to bind specific hairpin RNA structures, named iron responsive elements (IREs), within the untranslated regions of particular messenger RNAs. Apo aconitase is referred to as IRP,

5 for iron regulatory protein (21, 22). IRE-containing messenger RNAs include those of ferritin and transferrin receptor. Under limited iron conditions, IRP binding results in, among other effects, decreased ferritin levels (an iron storage protein) and increased transferrin concentrations (a protein involved in iron absorption) (21).

A fascinating example of physiologically relevant [4Fe-4S] to [2Fe-2S] cluster conversion exists for E. coli FNR (fumerate and nitrate reduction), a protein that functions as a metabolic switch between anaerobic and aerobic physiological processes

(23). The oxygen sensitive [4Fe-4S]2+ cluster of FNR (24) rapidly converts to a

2+ [2Fe-2S] cluster in the presence of O2 followed by slower degradation steps to apo protein (24, 25). The [4Fe-4S]2+ state of the protein is an active dimer that binds DNA with high affinity, thus regulating gene expression. The [2Fe-2S]2+ form is an inactive monomer (26, 27).

1.2 METALLOCHAPERONES

1.2.1 METALLOCHAPERONE FUNCTION

Although many metalloproteins can be reconstituted from their apo states by simple incubation with the appropriate metal under anaerobic conditions and in the presence of reductant, it is not believed that such mechanisms occur in vivo. Due to the toxicity of free metals, particularly of transition metals, metals do not exist in a free state inside the cell. Their toxicity, in part, stems from the ability of free metals to interact with oxygen and solvent molecules generating free radicals via Fenton-like reactions

6 (28). Free radicals can degrade proteins, nucleic acids, and lipids. Additionally, the

physiological concentrations of some metals are low, thus necessitating mechanisms to

ensure their sequestration and coordination to appropriate proteins. This is important

because metals can adventitiously bind to proteins and potentially inhibit coordination to

their intended protein targets. Such problems are overcome by a class of proteins dubbed

metallochaperones, which coordinate and deliver their metal cargo to their intended apo

protein target (29). The first identified metallochaperone was yeast Atx1 involved in

copper homeostasis (the human homologue is HAH1) (30). Subsequently, prokaryotic

operons (for example, the cop operon of Enterococcus hirae) encoding copper

metallochaperones were identified (31). Sequence comparisons, mutagenesis, and

structural characterization of several metallochaperones from different sources revealed a

conserved metal binding motif of M(T/H)CXXC (32).

1.2.2 COPPER METALLOCHAPERONE STRUCTURE

The first reported structures of metallochaperone proteins were of the copper

delivery proteins Atx1 (33) and CopZ (34). Both proteins possessed a βαββαβ

ferredoxin-like fold (Figure 1.2). Many proteins with widely different functions and

active sites have this tertiary structural motif, including several RNA binding proteins

(35).

Figure 1.2. Ribbons diagram of the solution structure of Enterococcus hirae CopZ. The

coordinates are from the PDB accession code 1CPZ.

An interesting variation of copper metallochaperones was observed with CCS. In

eukaryotes, CCS delivers copper to dimeric SOD, a copper-zinc containing protein (36-

39). Each subunit of CCS contains two domains. One domain is homologous to other

characterized copper chaperons, i.e. contains the ferredoxin-like motif, and the other domain is homologous to SOD, but lacks the necessary residues for metal coordination and enzymatic catalysis. Based on size exclusion chromatography, dynamic light scattering, cross-linking, sedimentation, and even crystallographic structural data, a mechanism of "domain swapping" has been proposed. This mechanism involves the formation of heterodimers between CCS and SOD through their respective SOD domains

8 (40-43). Copper is then delivered to apo SOD from the βαββαβ motif of CCS, followed by CCS - SOD dissociation, and formation of holo homodimeric copper-zinc SOD.

1.2.3 NON-COPPER METALLOCHAPERONES

Since the discovery of copper metallochaperones, several proteins involved in delivery of other metals to apo protein targets have been identified. These include nickel and zinc metallochaperones (44-47). Although not technically a metallochaperone, a protein (MerP) involved in mercury detoxification (35) also has been characterized and shares several functional and structural characteristics with metallochaperones.

Interestingly, all of these proteins share the presence of a M(T/H)CXXC metal binding

motif within a ferredoxin-like βαββαβ tertiary fold.

1.2.4 NECESSITY OF IRON-SULFUR CLUSTER METALLOCHAPERONES

The mechanisms involved in in vivo iron-sulfur cluster assembly and delivery are

only now beginning to be explored. The need for Fe-S chaperone-like proteins are the

same as for other metalloproteins, that is, free iron and sulfide are cellularly toxic.

However, the differences lie in the complexity of the cofactor. Those metallochaperones

that are extensively characterized are involved in delivering single metal ions to apo

proteins, whereas Fe-S clusters are polynuclear. Therefore, an Fe-S cluster must both be

synthesized and delivered to apo protein targets.

9 1.3 IRON-SULFUR CLUSTER ASSEMBLY

1.3.1 NITROGEN FIXATION

Nitrogen fixation is the process by which atmospheric nitrogen is reduced to

ammonia. This chemically difficult process is catalyzed by nitrogenase, a large (~300

kDa) oligomeric protein complex composed of dinitrogen reductase (the Fe protein) and

dinitrogenase (the MoFe protein) (48). The Fe protein is a [4Fe-4S] containing

homodimer, while the MoFe protein is an α2-β2 tetramer coordinating eight Fe ions and

two FeMo cofactors (49). The genes encoding components required for in vivo

nitrogenase activity are called nif (for nitrogen fixation). Much of the physiological data pertaining to nitrogen fixation has been obtained from studies on Azotobacter vinelandii.

A. vinelandii is a non-symbiotic, aerobic, nitrogen fixing bacterium that expresses large quantities of nitrogenase observable by the red cell pellets of diazotrophically grown cells. Bacteriological studies by Dean and colleagues found that unlike other nif mutant

A. vinelandii phenotypes, inactivation of either nifS or nifU resulted in light brown cells

lacking nitrogenase activity, i.e. nitrogenase deficient in both Fe and MoFe cofactors

were produced (50). The result clearly suggested that nifS and nifU are needed for in vivo

Fe-S cluster coordination to nitrogenase (51, 52). Subsequent work revealed that NifS is

a pyridoxal 5-phosphate (PLP) containing enzyme that converts L-cysteine to L-alanine and sulfur via an enzyme-bound cysteine persulfide intermediate (53).

Since NifS appeared to provide the sulfur necessary for Fe-S cluster formation within nitrogenase, it was hypothesized that NifU may be the iron or iron-sulfur cluster

10 donor (50). NifU is a modular protein with amino-terminal, central, and carboxy-

terminal domains (52, 54, 55). The central domain coordinates one stable [2Fe-2S] cluster per subunit of dimeric NifU. Although the role of this Fe-S cluster is not known,

it is required for in vivo and in vitro formation of holo nitrogenase (56). Initially not

much was known regarding the remaining domains, except that the amino-terminus

contained three additional conserved cysteines, and that the carboxy-terminal region

possessed two conserved Cys.

1.3.2 IDENTIFICATION OF BACTERIAL ISC PROTEINS

Prior to the identification of NifU and NifS, proteins were not known to be

involved in in vivo formation of Fe-S clusters. Interestingly, several lines of evidence

suggested that similar systems exist for Fe-S coordination to proteins not involved in

nitrogen fixation. First, deletion of either nifU or nifS in A. vinelandii was not lethal

indicating that another cellular system may functionally substitute for these proteins (51).

Second, NifS-like and NifU-like proteins are encoded by genomes of several non-

nitrogen fixing organisms spanning all three kingdoms of life including human (55).

Third, the NifS-like protein of E. coli possesses the same enzymatic activity as that of A.

vinelandii NifS (57).

Such considerations prompted a successful search for alternative operons

involved in Fe-S cluster formation in A. vinelandii. The newly identified operon was

named isc (for iron-sulfur cluster) and contains homologues to nifS (called iscS), nifU

(iscU), nifA (iscA), and genes encoding chaperones homologous to dnaJ (hsc20 or hscB)

11 and dnaK (hsc66 or hscA). Additionally, the operon codes for a [2Fe-2S] ferredoxin (58).

Purified IscS showed nearly identical biochemical characteristics as NifS (58). The NifU homologue (IscU) is only homologous to the amino-terminal domain of NifU and therefore only contains three conserved Cys residues. The discovery of the isc operon and the presence of isc homologues throughout nature are consistent with a common mechanism for cellular iron-sulfur cluster biosynthesis.

CHAPTER 2

CHARACTERIZATION OF HUMAN AND YEAST ISC PROTEINS

2.1 INTRODUCTION

Genes encoding ISC proteins were identified in human (55) and yeast genomes

(59). Indeed, yeast was found to contain genes coding for proteins homologous to IscU

(called ISU), IscA (ISA), IscS (NFS1), Hsc20 (JAC1), Hsc66 (SSQ1), and the isc specific

[2Fe-2S] ferredoxin (FDX1 or YAH1 (60)) (59). Interestingly, two genes encode IscU

(ISU1 and ISU2) and IscA (ISA1 and ISA2) homologues. Also, a gene homologous to

the carboxy-terminal domain of NifU (NFU1) is present in yeast (59) and human

genomes. These genes were further implicated in Fe-S cluster assembly by inactivation

experiments of NFS1, JAC1, and SSQ1 that resulted in decreased activities of Fe-S

cluster containing enzymes (aconitase and succinate dehydrogenase) (59).

Although Fe-S proteins are present in mitochondria, the cytosol, and the nucleus

(61), the ISC machinery has thus far only been identified in mitochondria (62).

Interestingly, NFS1 was found to be important in mammalian and yeast mitochondrial

13 and cytosolic Fe-S cluster assembly (63-65), suggesting that mitochondrial proteins play a fundamental role in Fe-S cluster biogenesis. Further studies in Saccharomyces cerevisiae localized ISU1, ISU2, and NFU1 to the mitochondrial matrix (66). Deletions of either ISU1 or ISU2 alone had no observable phenotypic effect; however, a simultaneous deletion of both genes was lethal (66, 67). An involvement in Fe-S cluster maturation was suggested by the inability of cells carrying ISU1, ISU2, or NFU1 null

mutations to maintain normal mitochondrial aconitase and succinate dehydrogenase

activity (66, 67). Interestingly, Cys to Ala mutations of the first two Cys of ISU resulted

in complete loss of ISU activity, whereas substitution of the third conserved Cys still

retained some activity, albeit to a much lesser extent (67).

Significantly less is known about ISA proteins, since no homologues have been

biochemically characterized. Nevertheless, separate deletions of S. cerevisiae ISA1 and

ISA2 showed some defects in respiratory processes, with double deletions remaining

viable and displaying the same phenotype as that of the single deletions. Therefore, ISA1

and ISA2 do not appear to provide redundant functions since they are non-additive (68).

All three deletion constructs also resulted in decreased aconitase and succinate

dehydrogenase activity, thus implicating ISA proteins in Fe-S cluster maturation. Cys to

Ser substitutions of the three conserved Cys residues yielded inactive protein. Unlike the

ISU proteins, a difference was observed between ISA1 and ISA2. ISA1 resides in the

mitochondrial matrix, whereas ISA2 performs its function in the mitochondrial

14 intermembrane space (68). Additional experiments indicated that ISA1 may be involved

in some way with cytosolic Fe-S cluster maturation in addition to mitochondrial cluster

assembly, even though S. cerevisiae ISA1 is localized to the mitochondrion (69).

There are other identified proteins implicated in Fe-S cluster biogenesis, including ferredoxin, Atm1p, and frataxin, although their exact roles have not been determined.

Yeast ferredoxin is an essential protein (60) that shows sequence homology with both human and bacterial ISC ferredoxins (70). Ferredoxin is important for both mitochondrial and cytosolic Fe-S cluster assembly (70) and is localized to the mitochondrion (60). Atm1p is an ATP-binding cassette transporter found in the mitochondrial inner membrane and is believed to be involved in the export of necessary components for cytosolic Fe-S cluster assembly (65, 71). Also, there has been some evidence implicating frataxin in Fe-S cluster biogenesis (72).

To better understand ISC proteins in general, and the eukaryotic homologues in particular, we biochemically characterized human and yeast (Schizosaccharomyces

pombe) ISU, yeast ISA1, and human and yeast ferredoxin. Our data clearly show that

ISU and ISA are capable of assembling Fe-S clusters and delivering them to a target apo

Fe-S protein.

15 2.2 MATERIALS AND METHODS

2.2.1 GENERAL CHEMICALS

All restriction enzymes and buffers were from Life Technologies (Rockville,

MD). Pfu polymerase was obtained from Stratagene (La Jolla, CA), Ni-NTA resin was

from QIAGEN (Valencia, CA). DEAE, G-25, and G-75 resins were from Pharmacia, and

DE-52, CM-32, and P-11 were from Whatman. Expression vectors, pET21 and pET28,

were from Novagen. All other chemicals were from Sigma-Aldrich.

2.2.2 PROTEIN EXPRESSION

Typically, E. coli BL21(DE3) pLysS cells (Novagen) were used for protein

expression except for the expression of Schizosaccharomyces pombe ISU1, which used

E. coli BL21(DE3)CodonPlus-RIL (Novagen). The expression vectors were pHsISUhis, pET-28b(+)-(ISU1) (73), Isa1/pET28 (73), pSpFdx, and HuFd/pET3a, for human ISU

(Hs ISU), S. pombe ISU1 (Sp ISU1), S. pombe ISA1 (Sp ISA1), S. pombe ferredoxin (Sp

Fd), and human ferredoxin (Hs Fd), respectively. The plasmid containing human ferredoxin was a gift from J. L. Markley (University of Wisconsin, Madison). Trial expressions were also performed with E. coli BL21(DE3) pLysE (Novagen) and E. coli

BL21(DE3) DNA-Y. The DNA-Y plasmid was a gift from M. P. Foster (Ohio State

University, Columbus) and D. P. Hornby (University of Sheffield, U.K.). For preparative protein production overnight starter cultures were grown at 37 oC in LB supplemented

with the appropriate antibiotic. Chloramphenicol (35 µg/ml) was needed for maintenance

of pLysS, pLysE, RIL, and DNA-Y. Ampicillin (100 µg/ml) was used when expressing

16 human ferredoxin, and 50 µg/ml kanamycin was added to cultures for the remaining

constructs. Inoculation was achieved with 10 mL/L overnight starter cultures, and the

cells grown at 37 oC. At an A600 ~ 0.6 the cultures were induced with 1 mM IPTG and

then harvested 5 h later. Cultures were either grown in 2 L Erlenmeyer/shaker flasks or

in a 10 L fermenter at the Ohio State Fermentation Facility. Cell pellets were stored at -

80 oC until further use.

2.2.2 HUMAN ISU PURIFICATION

Our construct of human ISU contains an amino-terminal His-tag. Even though

we have a non-His-tagged construct, it was not routinely used. Additionally, our construct is slightly truncated with the first residue (not including the His-tag) being

Arg22. Although there are two isoforms of Hs ISU, Hs ISU1 and Hs ISU2, the differences between them are solely confined to the region amino-terminal to Arg22.

Therefore, we simply refer to our construct as Hs ISU. Frozen cell pellets were thawed in

3 volumes of 50 mM Tris-HCl, pH 7.4, 1 mM PMSF and sonicated at 90% power.

Insoluble material was removed by ultracentrifugation at 45000 rpm, 4 oC for 30 min.

The supernatant was loaded onto a Ni-NTA column equilibrated with binding buffer (20 mM Tris-HCl, pH 7.9, 0.5 M NaCl, 5 mM imidazole), and washed with 5 column volumes of binding buffer. For apo Hs ISU purification, the column was washed with 5 column volumes of binding buffer + 75 mM imidazole and then eluted with binding buffer + 200 mM imidazole. Samples were then desalted by repeated ultrafiltration

(Amicon). For holo Hs ISU isolation, the Ni-NTA column with bound Hs ISU was

17 washed with binding buffer + 20 mM imidazole and then eluted with binding buffer +

200 mM imidazole. The protein was then diluted 3 fold with 50 mM Tris-HCl, pH 7.4 and loaded onto a cation exchange column (P-11) equilibrated with 50 mM sodium phosphate, pH 7.4, 50 mM NaCl. The column was washed with 3 column volumes of the same buffer, 3 column volumes of the same buffer plus an additional 50 mM, and finally

eluted with 50 mM sodium phosphate, pH 7.4, 0.5 M NaCl. Finally, holo Hs ISU was

further purified by a FPLC Superose-12 gel filtration column (HR 16/50, Pharmacia) run at 0.5 ml/min using 50 mM HEPES, pH 7.4, 50 mM NaCl as the running buffer. The extinction coefficient of holo Hs ISU at 460 nm is ~ 8500 M-1 cm-1 (74). Protein concentrations of apo samples were estimated from ε278 = 7254 M-1 cm-1 (ExPASy).

2.2.3 S. POMBE ISU1 AND ISA1 PURIFICATION AND RECONSTITUTION

The purification procedures for Sp ISU1 and Sp ISA1 were similar to that

previously reported by Wu et al. (75, 76). Cells pellets were thawed in 3 volumes of 50

mM Tris-HCl, pH 7.4, 50 mM NaCl, 1 mM PMSF and sonicated at 90% power. After

sonication, solid urea was added to the crude lysate to 6 M and stirred at 4 oC for 1 h.

Insoluble material was removed by centrifugation at 15000 rpm, 4 oC for 1 h. The

supernatant was then loaded onto a Ni-NTA column equilibrated with binding buffer + 6

M urea. The column was washed with 5 column volumes of binding buffer + 6 M urea,

binding buffer + 6 M urea + 20 mM imidazole, and then eluted with binding buffer + 6 M

urea + 250 mM imidazole. The eluted protein solution was diluted 2 fold with 50 mM

Tris-HCl, pH 7.4, 50 mM NaCl, and then subsequently diluted another 4 fold with 50

18 mM Tris-HCl, pH 7.4, 50 mM NaCl, 3 M urea. The protein was concentrated to

approximately 0.2 mM via ultrafiltration (Amicon). Fresh DTT was added to 50 mM and the solution argon purged in a round bottom flask for 0.5 h. Then ferric chloride was added slowly to 1.0 mM and sodium sulfide to 1.0 mM. The mixture was incubated for an additional 0.5 h. Desalting was achieved with a G-25 column. Separation of apo and holo protein was accomplished by cation exchange (CM-32) chromatography equilibrated with 50 mM sodium phosphate, pH 7.4. Bound protein was washed with 3

column volumes of sodium phosphate, pH 7.4, 50 mM NaCl, and holo protein eluted with

50 sodium phosphate, pH 7.4, 150 mM NaCl. If different solution conditions were

needed, buffer exchange was achieved by sephadex G-25. Holo Sp ISU1 has extinction

coefficients of 26770 M-1 cm-1, 27040 M-1 cm-1, 15990 M-1 cm-1 at 278 nm, 324 nm, and

442 nm, respectively, per dimeric protein (75). Holo Sp ISA1 concentrations were based

on extinction coefficients at 328 nm and 442 nm of 44200 M-1 cm-1 and 29800 M-1 cm-1, respectively, per tetrameric protein (76).

2.2.4 WILD TYPE S. POMBE FERREDOXIN PURIFICATION

Frozen cell pellets containing over-expressed holo WT Sp Fd were thawed in 5 volumes of 50 mM Tris-HCl, pH 7.4, 20 µg/ml DNase S1 and sonicated on ice at 90% power. Cellular debris was removed by centrifugation at 15000 rpm and 4 oC for 1 h, and

the supernatant was loaded onto a Ni-NTA column equilibrated with binding buffer. The

column was then washed with 5 column volumes of binding buffer, and the protein eluted

with binding buffer + 100 mM imidazole. The eluant was diluted 4 fold with 50 mM

19 Tris-HCl, pH 7.4 and loaded onto an anion exchange column (DEAE) equilibrated with

the same buffer. The column was then washed with 400 ml of 50 mM Tris-HCl, pH 7.4,

100 mM NaCl, and the protein eluted with 50 mM Tris-HCl, pH 7.4, 1 M NaCl. The

solution was then concentrated by ultrafiltration (Amicon) and loaded onto a gel filtration

column (G-75). Pure WT Sp Fd was stored at -80 oC. Holo WT Sp Fd concentrations were estimated by use of the Fe-S cluster extinction coefficient of WT Hs Fd, i.e. ε414 =

11 mM-1 cm-1 (77).

2.2.5 CYS TO SER SUBSTITUTED S. POMBE FERREDOXIN PURIFICATION

Frozen cell pellets were thawed in 5 volumes of 50 mM Tris-HCl, pH 7.4, 50 mM

NaCl, and 5 mM β-mercaptoethanol and sonicated at 90% power. During sonication

PMSF was added to 1 mM. Insoluble material was removed by centrifugation at 15000 rpm at 4 oC for 0.5 h, and the cleared lysate loaded onto a Ni-NTA column equilibrated with binding buffer. The column was then washed with 5 column volumes of binding

buffer + 5 mM imidazole, and the protein eluted with binding buffer + 100 mM imidazole. The samples were then desalted by use of a G-25 column equilibrated with 50 mM Tris-HCl, pH 7.4. Protein concentrations were based on extinction coefficients

calculated for apo protein from ExPASy (ε278 = 1450 M-1 cm-1).

2.2.6 HUMAN FERREDOXIN PURIFICATION

Cell pellets were thawed in 3 volumes of 50 mM Tris-HCl, pH 7.4, 1 mM EDTA,

40 µg/ml DNase S1, and 6.9 µg/ml RNase and sonicated at 90% power. After sonication

20 the crude lysate was ultracentrifuged at 45000 rpm, 4 oC for 0.5 h. The supernatant was

then diluted 2 fold with 50 mM Tris-HCl, pH 7.4 and loaded onto a DE-52 anion

exchange column equilibrated with the same buffer. The column was washed with 400

ml of 50 mM Tris-HCl, pH 7.4 resulting in the elution of a bright yellow band. The Hs

Fd containing fraction was eluted with 50 mM Tris-HCl, pH 7.4, 0.5 M NaCl. Solid

ammonium sulfate was added to the eluant to 50% (w/v) followed by stirring at 4 oC for

10 min and centrifugation at 15000 rpm at 4 oC for 10 min. The large light-brown pellet

was discarded, and the ammonium sulfate concentration of the red supernatant was increased to 95% (w/v) and stirred for an additional 10 min at 4 oC. After centrifugation

(15000 rpm, 4 oC, 10 min), the supernatant was discarded and the red pellet retained.

Often the pellet was stored at -20 oC at this stage until further purification. The pellet

was then resuspended in 50 mM Tris-HCl, pH 7.4 and loaded onto a G-75 gel filtration

column equilibrated with the same buffer. Fractions were judged for purity by their

absorbance ratio at 414 and 276 nm. Highly pure samples had a 414/276 between 0.76

and 0.79. Protein concentrations were based on ε414 = 11 mM-1 cm-1 (77).

2.2.7 MUTAGENESIS

The Quikchange technique (Stratagene) was employed for mutagenesis.

Reactions contained 50 ng template, 2.5 units of cloned Pfu DNA polymerase

(Stratagene), 1X cloned Pfu buffer, 0.5 mM DTT, and 125 ng of each primer. The

primers used are listed in Table 2.1. The thermocycle was identical to that described in

the Quikchange manual (Stratagene). An aliquot (27 %) of the post-thermocycle sample

21 o was incubated with 7.5 units of Dpn I at 37 C for 2 h. Subsequently, CaCl2 competent

DH5α was transformed via heat shock with the mutant constructs (78). Cloning and mutagenesis results were confirmed by nucleotide sequencing at the Ohio State

University Plant-Microbe Genomics Facility. Plasmids containing WT sequences were used as templates for mutagenesis resulting in single point mutations. However, for all

Hs ISU constructs, except for D37A Hs ISU, plasmid encoding D37A Hs ISU was used as the template so that double mutants would be generated. The K113Stop primers

(Table 2.1) were used to mutate the codon for K113 (AAG) to a TAG stop codon thereby creating a carboxy-terminal deleted D37A Hs ISU (D37A ∆CT Hs ISU) construct.

Additionally, all Hs ISU mutations were introduced in His-tagged constructs (pET28 parent plasmid), except for D37A Hs ISU which was made both in His-tagged and non-

His-tagged constructs (pET21 parent plasmid). The gene of S. pombe ferredoxin was cloned into pET28b(+) (pSpFdx) by G. Wu and is expressed with an amino-terminal His- tag. Even though vertebrate-type ferredoxins are highly homologous, unlike most ferredoxins Sp Fd is an integral membrane protein. Our construct only contains the soluble domain of the protein. Full length Sp Fd is 631 amino acids long, and the construct used by us only contains the carboxy-terminal 117 amino acids. This region is homologous to full length mature human ferredoxin (after removal of its mitochondrial targeting sequence). The numbering of amino acids was kept as it is for full length protein. For example, the first cysteine is Cys556 even though it is the forty second amino acid in our construct.

Mutation Oligonucleotide Sequence Restriction Site Resulting Construct D37A GTGGGGGCCCCAGCATGTGGTGCCGTAATGAAATTACAG gain Apa I D37A Hs ISU

D37A GTGGGGGCCCCAGCATGTGGTGCCGTAATGAAATTACAG gain Apa I D37A T72A Hs ISU* C35A GGACTGGTGGGGGCTCCAGCAGCTGGTGCCGTAATGAAATTACAG lose Apa I C35A D37A Hs ISU C61A GCTAGGTTTAAAACATTTGGCGCAGGTTCTGCAATTGCCTCCAGC gain AlwN I C61A D37A Hs ISU C96A CGCCAAGGAGCTGGCCCTTCCTCCCGTGAAACTGCACTGCTCC lose Sac I C96A D37A Hs ISU C104A GCCTTCCTCCCGTGAAACTGCACGCGTCCATGCTGGCTGAAGATGC gain Mlu I C104A D37A Hs ISU K113Stop GGCTGAAGATGCAATCTAGGCCGCCCTGGC D37A ∆CT Hs ISU C72A AAGCAAAAAGGAGCCGCTGGTCAGGCA C72A Sp ISA1 C136A GTGAAATCGACTGCCGGTTGCGGAGAA C136A Sp ISA1 C138A TCGACTTGCGGTGCCGGAGAATCGTTT C138A Sp ISA1 # C556S CGATTTAGAAGGCGCTAGTGAAGGGTCTGTTGC E553G C556S Sp Fd $ E553 GCTAACAATATCGATTTAGAAGGCGCTAGTGAAGGGTCTG C556S Sp Fd C562S GAAGGGTCTGTTGCAAGTTCGACTTGCCATGTTATCG C562S Sp Fd C565S CTGTTGCATGTTCGACTAGCCATGTTATCGTAGATCC C565S Sp Fd C602S GGAAACAAGCAGATTGGGAAGCCAAGTACTCTTAAG C602S Sp Fd D72A GATGCAATCACTGCTGAGGAGAATGACATGCTCGATC D72A Hs Fd E73A GATGCAATCACTGATGCGGAGAATGACATGCTCGATC E73A Hs Fd D76A CTGATGAGGAGAATGCCATGCTCGATCTGGCATATGG D76A Hs Fd D79A GAGAATGACATGCTCGCTCTGGCATATGGACTAACAGAC D79A Hs Fd

Table 2.1. Primers used for Quikchange mutagenesis. Bold positions are mutation sites, and underlined portions denote regions where either a restriction site was introduced

(gain) or removed (lose) by silent mutagenesis. Only one of the two primers for each reaction is listed. The other primer used was the exact complement of what is listed.

* Unintentional introduction of T72A. #Unintentional E553G mutation. $primer designed to revert amino acid position 553 to its original E residue.

23 2.2.8 UV-VISIBLE ABSORPTION SPECTROSCOPY

UV-visible spectra were recorded on Hewlett-Packard 8425A diode array spectrophotometer using the On-Line Instrument Systems (OLIS) 4300S Operating

System software. A 1.0 cm path-length cuvette was used for all measurements. Spectra were recorded at room temperature.

2.2.9 CIRCULAR DICHROISM

Circular dichroism spectra were measured on an AVIV model 202 circular dichroism spectrometer using 3 mm path length cuvettes at 25 oC. Far-UV CD (190 -

250 nm) solution conditions were 1.5 µM Hs Fd in 5 mM sodium phosphate, pH 7.4.

Absorbance was collected every 1 nm and averaged over 10 scans. Near-UV-visible CD spectra were recorded on 50 µM protein in 5 mM sodium phosphate, pH 7.4 for Hs Fd and 50 mM Tris-HCl, pH 7.4, 50 mM NaCl for 85 µM holo D37A Hs ISU. Only one scan was collected.

2.2.10 NATIVE POLYACRYLAMIDE GEL ELECTROPHORESIS

Typically, 15% polyacrylamide gels were used, unless otherwise indicated. The buffer system was 25 mM Tris-HCl, 19.2 mM Glycine, pH 8.5, and the gels run at 25 mA. Bands were visualized by Coomassie blue staining.

24 2.2.11 EDC CROSS-LINKING

A sample (40 µM) of each protein was mixed and incubated in either 25 mM

Hepes, pH 7.4, 50 mM NaCl or 50 mM sodium phosphate, pH 7.4, 50 mM NaCl, and 5 mM EDC at room temperature for 2 h. An aliquot of the reaction mixture was then diluted with 4X SDS-PhastGel loading buffer (Pharmacia), boiled for 2 min, and loaded onto a Homogenous-20 pre-cast polyacrylamide gel (Pharmacia).

2.3 RESULTS AND DISCUSSION

2.3.1 HUMAN ISU EXPRESSION

We routinely used BL21(DE3) pLysS for protein expression, a λDE3 lysogen carrying the T7 RNA polymerase and the plasmid pLysS. This plasmid encodes T7 lysozyme, which is a natural inhibitor of T7 RNA polymerase. Low levels of T7 lysozyme, as provided by pLysS, ensures that desired recombinant protein expression only occurs upon induction, i.e. the presence of pLysS prevents "leaky" expression. Such a system is useful when the cloned gene product is toxic to the cell. The plasmid pLysE produces much higher levels of T7 lysozyme and can therefore potentially enhance the yield of highly toxic proteins. For Hs ISU, the expression was essentially identical when co-transformed with pLysS, pLysE or without either plasmid.

Another method for improving recombinant protein expression is to supplement the intracellular tRNA pool with rarely used E. coli tRNAs, since preferred codon usage differs widely between species. One example is the Arg codon AGA. This codon is

25 rarely used in E. coli, but common in eukaryotic genes. Lack of the appropriate tRNA

can result in low protein expression and misincorporation of, in this case, Lys in place of

Arg (79). The first codon in our Hs ISU construct is AGA, and so it seemed plausible that expression was diminished due to the presence of this uncommon E. coli Arg codon.

Arg However, co-transformation with DNA-Y, a plasmid providing tRNAUCU , failed to

increase Hs ISU expression.

2.3.2 HOLO D37A Hs ISU IRON-SULFUR CHARACTERIZATION

Initially it was not known whether ISU/IscU proteins have the ability to bind an

Fe-S cluster. The existence of only three highly conserved and uniquely spaced Cys

residues made predictions of IscU's possible metal content difficult (Figure 2.1).

Expression and purification of WT Hs ISU consistently yielded apo protein. However, it

was soon discovered that substitution of a highly conserved Asp with an Ala of

A. vinelandii NifU resulted in stabilization of protein bound iron (80). Therefore, we

made the analogous mutation within Hs ISU (D37A Hs ISU), and found that purified

D37A Hs ISU was red with UV-visible spectra similar to [2Fe-2S] containing proteins

(Figure 2.2A). While it is surprising that substitution of a non-Fe-S cluster ligand

resulted in increased cluster stability, such a situation is not without precedent. A Leu to

His substitution near one of the cluster ligands of E. coli FNR greatly increases the

stability of its [4Fe-4S] cluster (81). For ISU, it is probable that increased Fe-S cluster

stability is a result of decreased solvent access to the cluster, thus decreasing the

likelihood of hydrolytic attack (82). Nevertheless, being able to isolate holo protein

26 allowed for additional experiments that would otherwise not have been possible. These experiments performed either in our laboratory or through collaborations further defined

the nature of the Fe-S cluster. Mössbauer, EXAFS, EPR, iron quantitation, and size

exclusion chromatography revealed a monomeric protein with a single reductively labile

[2Fe-2S]2+ cluster (74). Thus far, Hs ISU is the only identified monomeric ISU protein.

An intriguing spectroscopic result is the near-UV-visible CD spectrum of D37A

Hs ISU (Figure 2.2B). This spectrum was similar to that reported for E. coli IscU (83),

consistent with equivalent cluster ligation for eukaryotic and prokaryotic ISU proteins.

This is significant, because most ISU proteins only contain three Cys residues.

Therefore, the [2Fe-2S] cluster is either ligated by 3 Cys and one non-Cys side-chain or

the cluster is coordinated at the interface of an oligomeric quaternary structure. Hs ISU

has four Cys and E. coli IscU possesses three. Therefore, the number of naturally

occurring Cys residues within ISU does not alter the mode of cluster coordination. Also,

even though the CD spectrum has features in common with typical [2Fe-2S] proteins,

such as human and Anabaena ferredoxin (Figure 2.2B and (84, 85)), further confirming

[2Fe-2S] coordination, there are significant spectral differences. For example, human

ferredoxin has a prominent peak at 436 nm in addition to several other bands. D37A Hs

ISU also possesses this prominent peak, which for D37A Hs ISU is at 438 nm; however,

D37A Hs ISU also contains a sharp band centered near 497 nm that is not present in the

spectrum of human or Anabaena ferredoxin. Interestingly, this peak is observed in C46S

Anabaena ferredoxin, in which one of the ligating Cys was substituted with a Ser (85).

27 No naturally occurring [2Fe-2S] proteins to date have been shown to have an oxygen ligand, although primary amino acid sequence analysis of some [2Fe-2S] proteins suggests that it may exist (86).

* HsISU1 MVLIDMS-VDLSTQVVDHYENPRNVGSLDKTSKNVGTGLVGAPACGDVMKLQIQ 53 SpISU1 MVTANVSRRMYHKNVLDHYNNPRNVGTLPKGDPDVGIGLVGAPACGDVMRLAIR 54 EcIscU ------MAYSEKVIDHYENPRNVGSFDNNDENVGSGMVGAPACGDVMKLQIK 46 AvIscU ------MAYSDKVIDHYENPRNVGKLDAQDPDVGTGMVGAPACGDVMRLQIK 46 AvNifU ------MWDYSEKVKEHFYNPKNAGAVEGAN---AIGDVGSLSCGDALRLTLK 44

* @ HsISU1 VD-EKGKIVDARFKTFGCGSAIASSSLATEWVKGKTVEEALTIKNTDIAKELC- 105 SpISU1 VN-KDGVIEDVKFKTFGCGSAIASSSYVTTMVKGMTLEEASKIKNTQIAKELC- 106 EcIscU VN-DEGIIEDARFKTYGCGSAIASSSLVTEWVKGKSLDEAQAIKNTDIAEELE- 98 AvIscU VN-EQGIIEDAKFKTYGCGSAIASSSLATEWMKGRTLEEAETIKNTQIAEELA- 98 AvNifU VDPETDVILDAGFQTFGCGSAIASSSALTEMVKGLTLDEALKISNQDIADYLDG 98

* HsISU1 LPPVKLHCSMLAEDAIKAALADYKLKQEPKKGEAEKK------142 SpISU1 LPPVKLHCSMLAEDAIKSAVKHYRSKQLTPVGTTAGAIESATA------192 EcIscU LPPVKIHCSILAEDAIKAAIADYKSKREAK------128 AvIscU LPPVKIHCSVLAEDAIKAAVRDYKHKKGLV------128 AvNifU LPPEKMHCSVMGREALQAAVANYRGETIEDDHEEGALICKCFA------147

Figure 2.1. Sequence alignment of ISU/NifU proteins. Identities are shown as inverted

text. The * denotes the 3 conserved Cys and the @ indicates the non-conserved Cys position. Only the IscU homologous region of NifU is shown. Organisms are as follows:

Hs, human; Sp, Schizosaccharomyces pombe; Ec, E. coli; Av, A. vinelandii.

Figure 2.2. Absorption spectra of Hs ISU. A) UV-visible spectra of apo WT (bottom line) and holo D37A Hs ISU (top line). B) Near-UV-visible CD spectra of holo Hs Fd

(bottom line) and holo D37A Hs ISU (top line).

2.3.3 CYS TO ALA MUTAGENESIS

Each Cys residue of D37A Hs ISU was substituted with Ala, and the protein expressed and purified (Figure 2.3). We maintained the D37A substitution to more easily determine whether each Cys residue was necessary for cluster coordination. Silent mutagenesis was used to facilitate identification of successful mutant constructs (Table

2.1). For example, to engineer the C35A D37A and C104A D37A Hs ISU constructs, an

29 Apa I site was removed and an Mlu I site was introduced, respectively, without further altering the amino acid composition of the protein. Although, Hs ISU does contain an

Apa I site, it was removed during the generation of D37A Hs ISU. Hs ISU does not contain a Mlu I restriction site. However, there is an Apa I and a Mlu I site in the lacI gene of pET28. Therefore, digestion of D37A pHsISUhis with Apa I results in two fragments of approximately 1400 bp and 4300 bp. Digestion of C35A D37A pHsISUhis only yields a single fragment of ~5700 bp. Similarly, Mlu I digestion of D37A pHsISUhis only results in one large fragment, whereas digestion of C104A D37A pHsISUhis with Mlu I gives two nucleic acid fragments. Mutagenesis was confirmed by nucleotide sequencing.

All double mutants expressed as apo protein, suggesting that each Cys does, indeed, act as a ligand to the [2Fe-2S] cluster. This is surprising since most ISU proteins only contain three Cys, and near-UV-visible CD spectra are the same for both three and four Cys containing ISUs. Although highly unlikely, this could be interpreted as Fe-S cluster coordination by all four Cys in ISUs with four Cys per monomer with three Cys containing ISUs achieving complete cysteinyl ligation by interfacial coordination of the

[2Fe-2S] cluster between subunits of an oligomeric structure. Support for this theory stems from the fact that most ISUs are dimers, and Hs ISU is a monomer. However, if

this were true then it would be expected that there is a correlation between ISU Cys

content and aggregation state, which is not the case. S. pombe ISU1 is dimeric and

possesses four cysteine residues with one [2Fe-2S]2+ cluster bound per monomeric subunit (75). Also, A. vinelandii IscU is a dimer with only three Cys per monomer and

30 has been isolated with both one and two [2Fe-2S] clusters per dimer (54). Additionally,

non-Cys ligation is supported by near-UV-visible CD (see 2.3.3) and resonance Raman

data (87). Therefore, the Fe-S cluster of ISU is more than likely coordinated by the three

conserved Cys residues and one unidentified non-Cys ligand. Possibilities include

solvent or an oxygen/nitrogen amino acid side-chain. The non-conserved Cys (Cys96)

may serve a role in in vivo Fe-S cluster assembly, without ultimately ligating the final

cluster once it is formed within Hs ISU.

2.3.4 ISU - FERREDOXIN CROSS-LINKING

NifU is a modular protein with three domains. The amino-terminal domain has

three conserved Cys and binds a labile, less stable [2Fe-2S] cluster, the central region

contains four conserved Cys and coordinates a stable [2Fe-2S] cluster, and the carboxy-

terminal domain has an unknown function with two conserved Cys (52, 55, 58). ISU is only homologous to the amino-terminus of NifU and analogously coordinates a reductively labile [2Fe-2S] cluster (58, 88, 89). Although ISU proteins do not contain this central ferredoxin-like domain, a ferredoxin has been identified in bacterial isc operons (58). For eukaryotic systems, in which ISC proteins are localized to the mitochondria (90), homologous mitochondrial ferredoxins have been identified (91, 92).

Since ISU coordinates a reductively labile [2Fe-2S] cluster, the process of intact Fe-S cluster transfer to an apo target protein may be initiated by ferredoxin mediated reduction of ISU's [2Fe-2S].

Figure 2.3. Cys substituted D37A Hs ISU electrophoresis. A) Examples of restriction digest screening of Cys mutant D37A iscU. Lane 1, 1 kb DNA ladder (labeled in bp); lane 2, Apa I digest of C35A D37A pHsISUhis; lane 3, Apa I digest of D37A pHsISUhis; lane 4, Mlu I digest of C104A D37A pHsISUhis; lane 5, Mlu I digest of D37A pHsISUhis. B) SDS-PAGE of Cys substituted D37A Hs ISU. Lane 1, low MW marker

(labeled in kDa); lane 2, C31A D37A Hs ISU; lane 3, C61A D37A Hs ISU; lane 4, C96A

D37A Hs ISU; lane 5, C104A D37A Hs ISU.

To test this hypothesis we sought to probe the possibility of ISU - ferredoxin complex formation. These experiments utilized human and yeast ISUs and Hs Fd.

Complex formation was identified by EDC cross-linking. EDC is a specific, zero-length

32 cross-linker that forms amide bonds between Lys and Asp/Glu side chains (93). The appearance of a cross-linked product is good initial evidence for a well-defined interaction between two proteins that is substantially brought about by a complementarity of acidic and basic residues. We found that incubation of holo ISU (both human and yeast) with holo Hs Fd in the presence of EDC resulted in the appearance of a new high molecular weight species that ran slightly faster than the 29 kDa marker (Figure 2.4, panels A and B), in agreement with the predicted MW of the ISU - Fd cross-linked species. Formation of the cross-linked product is unlikely to arise from non-specific interactions since cross-linking is observed also at significantly lower concentrations.

However, incubation of oxidized holo ISU with reduced holo Hs Fd did not decrease the cluster stability of ISU, suggesting that the role of Fd is not to induce cluster release from

ISU (75). Perhaps the physiological relevance of the holo ISU - holo Fd complex is to remove electrons from the ISU cluster during its biosynthesis, thereby stabilizing cluster coordination (75).

Aside from redox chemistry, the interaction of ISU with Fd raises the possibility of a physiological complex involved in Fe-S cluster transfer from holo ISU to apo Fd. In agreement with this, we find holo ISU to cross-link with apo Fd, while apo ISU does not cross-link with either apo or holo Fd (Figures 2.4C and 2.4D). Indeed, subsequent studies showed direct transfer of intact Fe-S cluster from holo ISU to apo Fd (82). It should be emphasized that Fd is not the only physiological target of ISU. In order to carry out its function, ISU must interact with a variety of Fe-S proteins.

Figure 2.4. Yeast and Human ISU EDC cross-linking with Hs Fd. A) Holo D37A Sp

ISU1 cross-linking with holo Hs Fd. Lanes 2 - 4 do not contain EDC whereas lanes 5 -7 represent reactions with EDC present. Lane 1, Low MW marker; lanes 2 and 5, D37A Sp

ISU1; lanes 3 and 6, Hs Fd; lanes 4 and 7, D37A Sp ISU + Hs Fd. B) Holo Hs ISU cross- linking with holo Hs Fd. Lanes 1 - 2 and 3 - 4 are from different gels. Lanes 1 and 4, low

MW marker; lane 2, D37A Hs ISU + Hs Fd; lane 3, D37A Hs ISU + Hs Fd + EDC. C)

Apo and holo Sp ISU1 cross-linking with apo and holo Hs Fd. All lanes represent reactions with EDC. Lane 1, low MW marker; lane 2, apo D37A Sp ISU1; lane 3, apo Hs

Fd; lane 4, apo D37A Sp ISU1 + holo Hs Fd; lane 5, apo D37A Sp ISU1 + apo Hs Fd; holo D37A Sp ISU1 + holo Hs Fd; holo D37A Sp ISU1 + apo Hs Fd. D) Apo and holo Hs

ISU cross-linking with apo and holo Hs Fd. All lanes are aliquots from reactions containing EDC. Lane 1, low MW marker, lane 2; holo D37A Hs ISU + holo Hs Fd; lane

3, apo D37A Hs ISU + holo Hs Fd; lane 4, holo D37A Hs ISU + apo Hs Fd; lane 5, apo

D37A Hs ISU + apo Hs Fd; lane 6, holo Hs Fd; lane 7, holo D37A Hs ISU; lane 8, high

MW marker. Apo Sp ISU1 and apo D37A Hs ISU precipitate in the presence of EDC resulting in low intensity bands as seen in panels C and D.

35 2.3.5 ISU - MUTANT HUMAN FERREDOXIN CROSS-LINKING

ISU proteins have a Lys rich domain at their carboxy-terminus that could potentially function as a protein interaction domain with complementary acidic residues

on the target protein. The interaction of Fd with adrenodoxin reductase and CytP450scc has been extensively characterized revealing that the acidic side-chains of D72, E73,

D76, and D79 of Hs Fd (Figure 2.5) mediate complex formation, with D76 and D79 being critical (94-96). Therefore, we generated Asp/Glu to Ala Hs Fd mutants to test their effect on complex formation with ISU. All Fd constructs were isolated to high purity with 414/276 absorbance ratios of 0.73, 0.79, 0.70, 0.75, and 0.76 for WT, D72A,

E73A, D76A, and D79A Hs Fd, respectively (Figure 2.6A). Additionally, the mutations did not result in an alteration of the folding properties of the protein as evidenced by far-

UV and near-UV-visible CD spectroscopy (Figures 2.6B and 2.6C). Incubation of any of these holo Hs Fd constructs with holo D37A Hs ISU in the presence of EDC resulted in the appearance of an ISU - Fd cross-linked species (Figure 2.6D). Although the result suggests that these carboxylate side-chains are not involved in complex formation, it may be that the affinity of the complex was decreased but not sufficiently as to inhibit EDC cross-linking. A more physiologically relevant experiment, carried out by S. Wu, utilized cluster transfer from holo D37A Hs ISU to apo Hs Fd as a means of evaluating the effect of these Fd mutations. S. Wu found that holo D37A Hs ISU was not able to transfer cluster to apo D76A Hs Fd and apo D79A Hs Fd, consistent with these residues mediating complex formation. Holo D37A Hs ISU was capable of transferring Fe-S cluster equivalents to apo D72A Hs Fd and apo E73A Hs Fd.

Figure 2.5. Helix C of bovine ferredoxin with carboxylate side-chains labeled. The PDB coordinates 1CJE were used to generate this figure.

Figure 2.6. Spectra of WT and mutant ferredoxins and their interaction with Hs ISU. For

panels A, B, and C spectra from bottom to top are WT, D72A, E73A, D76A, and D79A, respectively. A) UV-visible absorption spectra. B) Far-UV CD. C) Near-UV-visible

CD. D) EDC Cross-linking between holo D37A Hs ISU and holo Hs Fd. Lane 1, Low

MW marker; lane 2, WT Hs Fd + D37A Hs ISU; lane 3, D72A Hs Fd + D37A Hs ISU;

lane 4, E73A Hs Fd + D37A Hs ISU; lane 5, D76A Hs Fd + D37A Hs ISU; lane 6, D79A

Hs Fd + D37A Hs ISU; lane 7, WT Hs Fd; lane 8, D37A Hs ISU.

39 2.3.6 NATIVE-PAGE OF MUTANT FERREDOXINS

To test whether complex formation could be detected by electrophoresis, samples

of Hs ISU and Hs Fd were incubated together and analyzed by native-PAGE (Figure 2.7).

Unfortunately, even though cluster transfer between ISU and Fd is known to occur (82),

we were unable to observe a stable complex by native-PAGE. Cluster transfer did not

occur under the conditions used for this experiment because of the lack of reducing

equivalents in the reaction mixture. Typically, cluster transfer reactions utilize DTT in

order to ensure that the Cys of the apo protein target are reduced, thus capable of

accepting cluster. Also, it is interesting to note that WT, D72A and D76A, and D76A

and D79A Hs Fd run differently on the gel. Native-PAGE is sensitive to the size, shape, and charge of the molecule.

Figure 2.7. Native-PAGE of D37A Hs ISU and Hs Fd. Lanes 8 - 15 contain holo

protein. Lane 1, holo WT Hs Fd; lane 2, apo WT Hs Fd; lane 3, holo D37A Hs ISU; lane

4, holo WT Hs Fd + holo D37A Hs ISU; lane 5, holo WT Hs Fd + apo D37A Hs ISU;

lane 6, apo WT Hs Fd + holo D37A Hs ISU; lane 7, apo WT Hs Fd + apo D37A Hs ISU;

lane 8, D72A Hs Fd; lane 9, D72A Hs Fd + D37A Hs ISU; lane 10, E73A Hs Fd; lane 11,

E73A Hs Fd + D37A Hs Fd; lane 12, D76A Hs Fd; lane 13, D76A Hs Fd + D37A Hs

ISU; lane 14, D79A Hs Fd; lane 16, D79A Hs Fd + D37A Hs ISU.

41 2.3.7 S. POMBE ISA1 CYS MUTAGENESIS

ISA is a protein encoded by the isc operon in bacterial species with homologues

in eukaryotes. The function of ISA is unknown, but is thought to be involved in Fe-S

cluster biogenesis. G. Wu cloned and expressed ISA1 from S. pombe and found it to be a

tetrameric protein with one [2Fe-2S]2+ per monomeric subunit (76). The UV-visible

spectrum of WT Sp ISA1 is shown in Figure 2.8. Like ISU, ISA proteins only have three

conserved Cys. Therefore, we mutated each Cys to Ala to probe the ligands to the Fe-S

cluster. Since Sp ISA1 is recombinantly expressed in E. coli as inclusion bodies, a

reconstitution step was necessary to test whether an Fe-S cluster could bind to these

proteins. It was found that each mutant (C72A, C136A, and C138A Sp ISA1) could

coordinate a significantly less stable [2Fe-2S]2+ cluster (76), under the reconstitution conditions employed for WT protein. While these proteins still are able to coordinate cluster, the high degree of instability indirectly suggests that all the Cys residues of ISA are involved in cluster ligation (76).

Figure 2.8. UV-visible spectrum of WT Sp ISA1 in 50 mM Tris-HCl, pH 7.4.

2.3.8 Sp ISA1 INTERACTION WITH HUMAN FERREDOXIN

ISA proteins contain a carboxy-terminal Lys rich domain similar to that of ISU.

Therefore, it seemed reasonable that ISA would interact with Hs Fd in a similar fashion

as ISU. EDC cross-linking was used as an initial test for complex formation and did

show formation of a Sp ISA1 - Hs Fd complex (Figure 2.9A). Analogous to ISU - Hs Fd cross-linking, holo Sp ISA1 cross-linked to apo and holo Hs Fd, whereas apo Sp ISA1 did not cross-link with either (Figure 2.9B). This is consistent with a complex involved in transferring cluster from holo ISA to apo Fd. Indeed, cluster transfer between these two

43 proteins has been observed (97). The Asp/Glu to Ala mutants of Hs Fd failed to inhibit

EDC cross-linking with Sp ISA1 (Figure 2.9C). However, S. Wu found that cluster

transfer from holo Sp ISA1 to Hs Fd did not occur in D76A and D79A Hs Fd constructs, just as in the case of Hs ISU. Similar to Hs ISU, holo Sp ISA1 could transfer cluster to apo D72A Hs Fd and apo E73A Hs Fd. Further support for the interaction domain of Hs

Fd being the same for ISU and ISA is found by the lack of formation of a termolecular

complex by EDC cross-linking when all three protein were incubated together (Figure

2.9D). Additionally, no cross-linking was observed between Hs ISU and Sp ISA1.

Figure 2.9. Sp ISA1 EDC cross-linking with target proteins. A 40 µM sample of each protein in 25 mM Hepes, pH 7.4, and 50 mM NaCl was incubated at room temperature for 2 h with or without 5 mM EDC. A) Holo Sp ISA1 cross-linking with holo WT Hs Fd.

Lane 1, low MW BRL marker; lane 2, Sp ISA1; lane 3, Hs Fd; lane 4, Sp ISA1 + Hs Fd; lane 5, Sp ISA1 + EDC; lane 6, Hs Fd + EDC; lane 7, Sp ISA1 + Hs Fd + EDC; lane 8, high MW BRL marker. B) Apo and holo Sp ISA1 - Hs Fd cross-linking. Lane 1, Low

MW BRL marker; lane 2, holo Hs Fd + holo Sp ISA1; lane 3, holo Hs Fd + apo Sp ISA1; lane 4, apo Hs Fd + Holo Sp ISA1; lane 5 apo Hs Fd + apo Sp ISA1; lane 6, holo Hs Fd; lane 7, holo Sp ISA1; lane 8, apo Sp ISA1. C) Holo Sp ISA1 cross-linking with holo mutant Hs Fd. Lane 1, low MW BRL marker; lane 2, WT Hs Fd + Sp ISA1; lane 3,

D72A Hs Fd + Sp Isa1; lane 4, E73A Hs Fd + Sp ISA1; lane 5, D76A Hs Fd + Sp ISA1; lane 6, D79A Hs Fd + Sp ISA1; lane 7, WT Hs Fd; lane 8, Sp ISA1. D) Sp ISA1 - D37A

Hs ISU EDC cross-linking. Lane 1, low MW BRL marker; lane 2, holo Sp ISA1 + holo

D37A Hs ISU; lane 3, holo Sp ISA 1 + apo D37A Hs ISU; lane 4, apo Sp ISA1 + holo

D37A Hs ISU; lane 5, apo Sp ISA1 + apo D37A Hs ISU; lane 6, apo Sp ISA1 + holo

D37A Hs ISU + holo Hs Fd; lane 7, holo Sp ISA1+ holo D37A Hs ISU + holo Hs Fd; lane 8, apo Sp ISA1. Mass labels are in kDa.

46 2.3.9 C-TERMINAL DELETION OF D37A HS ISU

Since it is believed that ISU and ISA proteins interact with at least some of their target proteins through their carboxy-terminal Lys rich domains, a carboxy-terminal deletion mutant of D37A Hs ISU (D37A ∆CT Hs ISU) was constructed. D37A ∆CT Hs

ISU could bind an Fe-S cluster after chemical reconstitution (Figure 2.10). However, the yield and stability of the cluster was low, thus precluding meaningful experiments.

2.3.10 CYS TO SER S. POMBE FERREDOXINS

Since a convenient assay was developed to monitor and quantitate cluster transfer kinetics from holo ISU to apo ferredoxin (82), we sought to further characterize the mechanistic details of this process through mutagenesis. Cys to Ser substitutions of each

Cys of S. pombe ferredoxin were made in the hopes of defining the order of target protein

Cys nucleophilic attack on the cluster. Although, natural Ser coordination to an Fe-S cluster has never been observed (98), Fe-S clusters can be coordinated by three Cys and one Ser ligand, albeit resulting in a significantly less stable cluster with variable effects dependent on which Cys is lost (9). Ser coordination of several mutant [2Fe-2S] ferredoxins has been previously reported (9, 77, 85, 99). We had hoped that the kinetics of cluster transfer from holo ISU to a series of Cys to Ala mutants of Sp Fd would suggest the order of Sp Fd Cys binding. However, even though each Cys to Ser mutant of Sp Fd constructed could be chemically reconstituted with an Fe-S cluster, none of them were able to accept an Fe-S cluster from ISU.

Figure 2.10. UV-visible spectrum of D37A ∆CT Hs ISU after chemical reconstitution

with iron and sulfide. Reconstitution was as described for Sp ISU1 in materials and

methods.

CHAPTER 3

CLONING AND CHARACTERIZATION OF THERMOTOGA MARITIMA IscU

3.1 INTRODUCTION

The eukaryotic ISU proteins that we have studied show NMR behavior consistent with an unfolded state (100, 101), i.e. their spectra (1H and 1H-15N HSQC) lack chemical

shift dispersion. To identify an IscU with a more stable fold, we cloned, isolated, and

biochemically characterized an IscU from the hyperthermophilic bacterium Thermatoga

maritima. T. maritima represents one of the deepest and most slowly evolving

eubacterial lineages (102), with an optimal growth temperature of 80 oC (103). The

genomic sequence shows genes encoding both IscU (Tm IscU) and IscS located next to

each other, although no other isc genes are found clustered within this region. Located

elsewhere in the T. maritima genome are heat shock proteins that are homologous to

those identified in other organisms (104) and several ferredoxins. Additionally,

T. maritima expresses a second NifS-like protein that has been crystallographically

characterized (105). Possibly the functions of the missing ISC proteins are provided by

49 other proteins involved in iron homeostasis, as has been identified in E. coli (106, 107).

Consistent with this is the fact that the T. maritima operon encoding iscU and iscS

contains genes implicated in other Fe-S cluster assembly paths identified in higher

organisms.

The non-ISC encoding genes near T. maritima iscU and iscS are homologous to

those typically found in the suf operon, initially identified in E. coli (108). The suf operon encodes SufA (an IscA homologue (109)), SufB, SufC, SufD, SufS (an IscS homologue (107)), and SufE (108). SufC has ATPase activity and interacts with SufB, both of which are membrane proteins (110). SufB and SufD are paralogues (111), and

SufD is believed to be important for the stability of the [2Fe-2S] protein FhuF (107). The suf operon is controlled by the Fur repressor and induced under iron limiting conditions

(111). Interestingly, neither the complete suf nor the complete isc operon is present in

T. maritima. Instead, the T. maritima operon appears to be a combination of both, consisting of sufC, sufB, an ORF encoding a protein with unknown function (TM1370), iscS, and iscU, in that order. It may be that suf and isc operons originated from one common operon as found in T. maritima.

Herein we report the characterization of an evolutionarily ancient IscU. Similar to other bacterial and eukaryotic organisms (74, 83, 87), Tm IscU binds a [2Fe-2S] cluster. The holo form of the native protein proved to be relatively labile, but following prior precedent (74, 89) was stabilized by substitution of a conserved aspartate by alanine

(D40A). Analysis of CD spectra over a wide temperature range demonstrate Tm IscU to possess significant secondary structure that is not dependent upon the coordination of an

50 Fe-S cluster, while the dispersion of resonances in 1H-15N HSQC spectra indicate

significant tertiary structure. Furthermore, Tm IscU proves to be competent for Fe-S cluster transfer to human ferredoxin (Hs Fd), consistent with a conserved role in Fe-S cluster biogenesis. This result provides strong support for a common conserved recognition mechanism for both prokaryotic and eukaryotic IscU-type proteins.

3.2 MATERIALS AND METHODS

3.2.1 GENERAL CHEMICALS

57Fe was from Pennwood Chemicals. All restriction enzymes and the copper staining kit were from Life Technologies (Rockville, MD). Pfu DNA polymerase and

BL21CodonPlus(DE3)-RIL were from Stratagene (La Jolla, CA); PCR Purification Kit and Ni-NTA resin were purchased from QIAGEN (Valencia, CA). Protein expression vectors pET21, pET28, and BL21(DE3) pLysS were from Novagen (Madison, WI).

Oligonucleotides were from Integrated DNA Technologies, Inc (Coralville, IA). CM-32 and DE-52 were from Whatman (Aston, PA). Homogenous-20 precast polyacrylamide gels, G-75, and Superose-12 resin were from Pharmacia.

3.2.2 CLONING OF T. MARITIMA iscU

T. maritima genomic DNA was obtained from the American type culture collection (ATCC # 43589D). Chromosomal DNA (25 ng, 50 ng, and 100 ng), 0.3 µM of each primer, 2.5 units of cloned Pfu DNA polymerase, 1X cloned Pfu buffer, and 0.2 mM each dNTP were used to amplify iscU (TIGR locus: TM1372) by PCR. The total volume

51 was 50 µL. Primers were: 5’-GGGCCCGGCATATGGTTTTCAAGATGATG-3’ and 5’-

CCGGCCGGATCCTTAAGGCCGTGAAATCTTTTTG-3’ where underlined regions denote Nde I and BamH I sites respectively, and the bold position indicates a G to A

substitution resulting in an amino-terminal Met rather than a Val for improved

recombinant expression in E. coli. The thermocycle used was as described in the Pfu

DNA polymerase manual (Stratagene): 95 oC for 45 s followed by 25 cycles of 95 oC for

45 s, 55 oC for 45 s, and 72 oC for 1 min with an additional one cycle of 72 oC for 10 min.

Pfu DNA polymerase was added to the reaction mixture during the first annealing step

(55 oC) of the thermocycle. The PCR product was purified with the Qiagen PCR

purification kit and then lyophilized via speed-vac (Savant). To the lyophilized PCR

product the following components were added: 10 mM DTT, 0.1 mg/ml BSA, 40 units of

Nde I, 20 units of BamH I, sterile water, and 1X React 3 buffer (BRL). Digestion

samples were incubated at 37 oC for 16 - 18 h. Approximately 7 µg of vector (pET28 or

pET21) was similarly digested. After digestion the samples were purified with the

Qiagen gel extraction kit by directly loading the mixtures onto the column as described in

the Qiagen manual, i.e. the digested DNA was not purified from an agarose gel. The

purified DNA was lyophilized and resuspended in 10 µL of sterile H2O. From these

samples, 2.0 µL of vector was mixed with 3.0 µL of insert, 1X T4 ligase buffer, and 1 unit of T4 ligase (Invitrogen). The volume was brought to 10 µL with sterile H2O and the mixture incubated at 14 oC for 16 h. Subsequently, 4 µL of the ligation mixture was

electroporated into DH5α and plated on LB supplemented with 100 µg/ml ampicillin for

pET21 constructs or LB supplemented with 50 µg/ml kanamycin for pET28 constructs.

52 Colonies were screened by plasmid purification (Qiagen mini-prep kit) and digestion (1

µg sample of the DNA) with 10 units EcoR I, 1X React 3 (BRL) at 37 oC for 1 h (total

volume was 10 µL). The entire mixture plus 1 µL of 10X DNA loading buffer (50%

(v/v) glycerol, 0.1 M EDTA, 1% (w/v) SDS, 0.1% w/v bromphenol blue, 0.1% (w/v) xylene cyanole) was run on a 1 % (w/v) agarose gel with TBE as the running buffer (10.8 g Tris-base, 5.5 g boric acid, 0.74 g EDTA per L dH2O). Positive clones were confirmed

by nucleotide sequencing at the Ohio State University Plant-Microbe Genomics Facility.

Cloning into pET21 and pET28 yielded constructs without a tag (pTmIscU) and with an

amino-terminal His-tag (pTmIscUHis), respectively.

3.2.3 MUTAGENESIS

The Quikchange technique (Stratagene) was employed for the D40A point

mutation. Reactions contained 50 ng template (either pTmIscU or pTmIscUHis), 2.5 units of cloned Pfu DNA polymerase, 1X cloned Pfu buffer, 0.5 mM DTT, and 125 ng of each primer. Primers were 5’-

GGGAAAGAACATCTCTTGTGGCGCCGAAATCACACTCTAC-3’ and 5’-

GTAGAGTGTGATTTCGGCGCCACAAGAGATGTTCTTTCCC-3’ where the bold positions indicate the mutation. The thermocycle was identical to that described in the

Quikchange manual (Stratagene). An aliquot (27 %) of the post-thermocycle sample was

o incubated with 7.5 units of Dpn I at 37 C for 2 h. Subsequently, CaCl2-competent DH5α

was transformed via heat shock with the mutant constructs (78). Cloning and

mutagenesis results were confirmed by nucleotide sequencing at the Ohio State

53 University Plant-Microbe Genomics Facility. The generation of D55A pTmIscUHis followed an identical procedure as for D40A pTmIscU/His except that the following primers were used: 5'-GGAAGACGGTGTTGTGAAAGCTGCTAAGTTCGAGGG-3' and 5'-CCCTCGAAC TTAGCAGCTTTCACAACACCGTCTTCC-3', where the bold positions indicates the mutation site.

3.2.4 T. MARITIMA IscU OVER-EXPRESSION

BL21CodonPlus(DE3)-RIL was used for protein expression. A 100 ml LB culture (supplemented with either 100 µg/ml ampicillin and 35 µg/ml chloramphenicol for pTmIscU constructs or 50 µg/ml kanamycin and 35 µg/ml chloramphenicol for pTmIscUHis constructs) was grown overnight as a starter culture. The entire starter culture was used as an inoculum for a 10 L fermentation at the OSU fermentation facility and grown to an A600 ~ 0.6 prior to induction with 1 mM IPTG. Cells were pelleted 5 h post-induction and stored at –80o C for future use.

3.2.5 PROTEIN PURIFICATION

Cell pellets were resuspended in 5 volumes of 50 mM Tris-HCl, pH 7.4, 2 mM β-

mercaptoethanol, 20 µg/ml DNase, and 5 µg/ml RNase and lysed by sonication.

Insoluble material was removed by centrifugation at 15000 rpm, 4o C for 1 h. For His-

tagged constructs the cleared lysate was applied to a Ni-NTA column equilibrated with

binding buffer (20 mM Tris-HCl, pH 7.9, 5 mM imidazole, 500 mM NaCl). The column

was then washed with 5 column volumes of binding buffer, 5 volumes of binding buffer

54 + 15 mM imidazole, and the protein was subsequently eluted with binding buffer + 100 mM imidazole. His-tagged protein was exchanged with 50 mM sodium phosphate, pH

7.4 via repeated ultrafiltration (Amicon).

The cleared lysate containing non His-tagged protein was loaded onto a cation

exchange column (CM32) equilibrated with 50 mM sodium phosphate, pH 7.4 and

washed with one cleared lysate volume of phosphate buffer. The flow through and wash

fractions were combined. NaCl was added to 50 mM and β-mercaptoethanol was added

to 5 mM and the solution was incubated at 85o C for 0.5 h. The sample was then

centrifuged at 15000 rpm, 4o C for 10 min and the supernatant loaded onto an anion

exchange column (DE-52). The column was washed with 3 column volumes of 50 mM

Tris-HCl, pH 7.4. The flow through and wash fractions were combined and concentrated

via ultrafiltration. Subsequently, protein solution was loaded onto a G-75 gel filtration

column equilibrated with 50 mM sodium phosphate, pH 7.4. The fractions with a λmax at

278 nm were pooled and confirmed to be pure IscU via SDS-PAGE. All apo protein samples were stored at either 4 oC or –80 oC.

HuFd/pET3a (encoding human ferredoxin) was a gift from J. L. Markley.

BL21(DE3) pLysS HuFd/pET3a was essentially expressed and purified as previously

described (77). Apo Hs Fd was prepared as described by Nishio and Nakai (112).

55 3.2.6 MASS SPECTROMETRY

All mass spectra were acquired at the campus chemical instrument center at Ohio

State University and all solutions were made in Barnstead purified water. MW

determination for extinction coefficient calculation was determined by electro-spray ionization (ESI) using a Micromass Q-TOF(tm) II (Micromass, Wythenshawe, UK) mass spectrometer equipped with an orthogonal electrospray source (Z-spray) operated in

positive ion mode. Sodium iodide was used for mass calibration for a calibration range

of m/z 100 - 2500. Salt buffers from the protein samples were cleaned using manual

syringe protein traps from Michrom BioResources (Auburn, CA). Proteins were

prepared in a solution containing 50% acetonitrile/50% water, 0.1% formic acid at a

concentration of 50 pmol/µl and infused into the electrospray source at a rate of 5 - 10 µL

min-1. Optimal ESI conditions were: capillary voltage 3000 V, source temperature 110o C

and a cone voltage of 60 V. The ESI gas was nitrogen. Q1 was set to optimally pass ions

from m/z 100 - 2000 and all ions transmitted into the pusher region of the TOF analyzer

were scanned over m/z 100 - 3000 with a 1 s integration time. Data was acquired in

continuum mode until acceptable averaged data was obtained (10 -15 min). ESI data

were deconvoluted using MaxEnt I provided by Micromass.

In-gel tryptic digests were performed on copper stained protein bands (5, 10, and

20 µg WT Tm IscU) separated on a 15% SDS-PAGE gel. Incised bands were washed

with 50% methanol (HPLC grade)/5% acetic acid (JT Baker UltrexII Ultrapure) and dried

in HPLC grade acetonitrile. Samples were then reduced with DTT (5 mg/ml in 100 mM

ammonium bicarbonate) and alkylated with iodoacetamide (15 mg/ml in 100 mM

56 ammonium bicarbonate). The protein band was dehydrated with acetonitrile, rehydrated

with 100 mM ammonium bicarbonate, and dehydrated again with acetonitrile. Promega

modified trypsin (20 ng/µl in 100 mM ammonium bicarbonate, total addition = 50 µl)

was added to each gel piece and rehydrated on ice for 10 min. Samples were centrifuged

and excess trypsin solution removed. Ammonium bicarbonate (50 mM) was added to 20

µl, vortexed and centrifuged briefly, and incubated at room temperature overnight. A solution of 50% acetonitrile/5% formic acid (EM Science ACS 88%) was added to 30 µl, vortexed for 10 min, and centrifuged. The supernatant was isolated and an additional 30

µl of 50% acetonitrile/5% formic acid was added, vortexed for 10 min, and centrifuged.

The supernatant was analyzed by matrix assisted laser desorption/ionization time-of- flight (MALDI-TOF) performed on a Bruker Reflex III (Bruker, Breman, Germany) mass spectrometer operated in linear, positive ion mode with a N2 laser. Laser power was used at the threshold level required to generate signal. Accelerating voltage was set to 28 kV.

The instrument was calibrated with protein standards whose masses encompased the molecular weights of the protein samples (typically mixtures of bradykinin fragment 1-5

and ACTH fragment 18-39 as appropriate). Salt buffers from the protein samples were

cleaned using ZipTips (Millipore, Bedford, MA) according to manufacturer directions.

α-Cyano-4-hydroxycinnamic acid was used as the matrix and prepared as a saturated

solution in 50% acetonitrile/ 0.1% trifluoroacetic acid (in water). Aliquots of 1 µL of

matrix and 1 µL of sample were thoroughly mixed together; 0.5 µL of this was spotted on

the target plate and allowed to dry. Protein identification based on peptide masses was

determined by the program ProFound .

57 3.2.7 T. MARITIMA IscU CLUSTER RECONSTITUTION

Protein (0.5 mM) in 50 mM sodium phosphate, pH 7.4, 50 mM NaCl, 50 mM

DTT was repeatedly degassed and argon purged for 1 h. Fresh FeCl3 and Na2S were then

slowly added to 1 mM. The mixture was allowed to incubate anaerobically at room

temperature for 0.5 h. Insoluble material was removed by centrifugation at 15000 rpm,

4o C for 5 min. The supernatant was desalted by a G-25 column equilibrated with 50 mM

Tris-HCl, pH 7.4 or 50 mM sodium phosphate, pH 7.4, and the colored protein fraction was collected. His-tagged holo protein was concentrated and stored at –80 oC until further use. Non His-tagged holo protein was then loaded onto a DE-52 column equilibrated with 50 mM Tris-HCl, pH 7.4 and washed with 5 column volumes of the same buffer. The holo-protein was eluted with 50 mM Tris-HCl, pH 7.4, 100 mM NaCl.

3.2.8 SIZE EXCLUSION CHROMATOGRAPHY

FPLC gel filtration using a Superose-12 column (HR 16/50) at a flow rate of 0.5 ml/min was used for aggregation state determination. The running buffer was 50 mM

HEPES, pH 7.4, 50 mM NaCl. A Gel Filtration Calibration Kit (Pharmacia) was used to calibrate the column. The standards were RNase A (13,700 Da), chymotripsinogen A

(25,000 Da), ovalbumin (43,000 Da), and albumin (67,000 Da). Blue Dextrin was used to determine the dead volume. MWs were determined by plotting log MW of standards vs. Kav where Kav = (Ve - V0)/(Vt - V0) and Ve = elution volume; V0 = dead volume; Vt =

total column volume.

58 3.2.9 DYNAMIC LIGHT SCATTERING

Light scattering was recorded on a DynaPro-801 (Protein Solutions,

Charlottesville, VA) with a temperature-controlled microsampler. The laser wavelength

and scattering angle was 8294 Å and 90o, respectively. The instrument software

(Dynamics, version 3.27) was used to calculate the MW of sample from its measured

translational diffusion coefficient. Solution conditions were 50 µM protein in 100 mM

Tris-HCl, pH 7.4 at 22.7 oC with varying NaCl concentrations of 0, 50, 100, 200, 300,

and 450 mM NaCl. Samples were filtered through 0.02 µm filters (Anatop-10 from

Whatman) immediately prior to measurements.

3.2.10 DSS CROSS-LINKING

D40A Tm IscU (50 µM, 25 µM, 10 µM, and 5 µM) was incubated with 2.5 mM disuccinimidyl suberate (DSS) in 50 mM sodium phosphate, pH 7.4, 50 mM NaCl and incubated at room temperature for 1 h. The DSS stock solution was made in dimethylsulfoxide. SDS-PAGE loading buffer (4X: 100 mM Tris-HCl, pH 6.8, 1.0%

(w/v) SDS, 0.5% (v/v) β-mercaptoethanol, 0.05% (w/v) bromophenol blue) was then added, the samples boiled for 2 min, and then run on a 15% polyacrylamide gel at 20 mA.

3.2.11 UV-VISIBLE SPECTROSCOPY AND EVALUATION OF EXTINCTION

COEFFICIENTS

UV-visible spectra were recorded on a Hewlett-Packard 8425A diode array spectrophotometer using the On-Line Instrument Systems (OLIS) 4300S Operating

59 System software. A 1.0 cm path-length cuvette was used for all measurements. All

solutions used for extinction coefficient determination were prepared in Barnstead

purified deionized water. Apo D40A Tm IscU was initially dialyzed extensively against a

volatile buffer (100 mM ammonium bicarbonate, pH 7.0) then against unbuffered water.

Finally, the protein was passed through a G-25 column equilibrated with water. After the

absorption spectrum was collected, the sample was lyophilized and the mass of the

protein sample determined. Using Beer’s law and the MW determined by ESI, the

extinction coefficient of apo D40A Tm IscU was calculated. The concentration of protein

used for holo D40A Tm IscU extinction coefficient calculation was determined by incubating holo D40A Tm IscU in 50 mM Tris-HCl, pH 7.4, 50 mM NaCl, 1 mM EDTA at 60o C for 0.5 h followed by desalting via a G-25 column. Total protein concentration

was then determined from the 278 nm absorption of apo D40A Tm IscU using the

previously determined extinction coefficient.

3.2.12 TEMPERATURE DEPENDENCE OF [2Fe-2S] CLUSTER STABILITY

Holo protein (30 µM) in 100 mM Tris-HCl, pH 7.4, 50 mM NaCl was repeatedly degassed and argon purged in a stoppered cuvette while kept on ice. Absorption data at

412 nm was acquired on a Hewlett-Packard diode array spectrometer (HP 8453) with

HP8453 Win system software. Temperature control was achieved with a Peltier

temperature controller (HP 8909A). Absorption data at temperatures ranging from 20 oC

to 86 oC were collected at 2 oC increments with a 0.5 min equilibration period at each

temperature prior to measurement.

60 3.2.13 MÖSSBAUER SPECTROSCOPY

Mössbauer spectra of 57Fe D40A Tm IscUHis was recorded on a constant

acceleration spectrometer, model MS-1200D from Ranger Scientific, using a Janis

SuperVaritemp cryostat (model 8DT), a Lakeshore temperature controller (model 340),

and a 57Co source from Isotope Products Laboratory. The 57Fe reconstituted protein was prepared as described above (see Tm IscU cluster reconstitution) except that 57Fe3+ was used instead of regular FeCl3. Kristene K. Surerus (University of Wisconsin, Milwaukee)

acquired and analyzed the Mössbauer spectra.

3.2.14 EPR SPECTROSCOPY

EPR signals were recorded with an X-band Bruker ESP 300 spectrometer

equipped with an Oxford liquid helium cryostat at 15 K by C. Hemann in the R. Hille

laboratory. Holo D40A Tm IscUHis concentrations were 0.9 mM for untreated samples,

and 0.4 mM for ascorbate reduced samples. Holo D40A Tm IscUHis was reduced with

0.2 mM and 2 mM ascorbate and immediately frozen.

3.2.15 NUCLEAR MAGNETIC RESONANCE

15N- labeled apo D40A Tm IscU was expressed in minimal media supplemented

15 15 with NH4Cl (113). N-HSQC spectra were recorded on a Bruker 600 Avance DMX

spectrometer operating at 600 MHz. The pulse sequence was as previously described

(114).

61 3.2.16 IRON QUANTITATION

Iron content of holo D40A Tm IscU was determined by atomic absorption using a

Perkin Elmer Zeeman 5000 graphite furnace atomic absorption spectrometer. An iron

standard solution (GFS Chemicals, Inc.) was used to construct a standard curve with a r2

> 0.999. Sample loading was automated (AS40) and all data were run in duplicate and averaged. Absorption was measured at 305.9 nm and integrated for 7 s. All solutions were in Barnstead purified deionized water and 2% nitric acid. Background iron concentrations of solutions were measured and found to be negligible.

3.2.17 CIRCULAR DICHROISM

Circular dichroism spectra were measured on an AVIV model 202 circular dichroism spectrometer. Far-UV CD spectra was acquired with a 0.1 mm path length cuvette. Experimental conditions were essentially as recommended by Johnson (115).

Protein concentrations were 0.08 mM WT and D40A Tm IscU (based on monomeric protein) and 0.15 mM Hs Fd in 10 mM sodium phosphate pH 7.0, 10 mM NaCl, except for holo D40A Tm IscU (10 mM Tris-HCl, pH 7.0, 50 mM NaCl). Spectra acquired at

20 oC were determined per 0.2 nm in triplicate and averaged. At elevated temperatures

spectra were only recorded once. Secondary structure quantitation was determined via

the self-consistent method (116) with the Dicroprot V2.5 version 5.0 package (117).

Near-UV-Vis CD spectra were recorded in a 3 mm path length cuvette per 1 nm. Protein

concentrations were 40 µM in 50 mM Tris-HCl, pH 7.0. Buffer spectra were always

subtracted.

62 3.2.18 D40A Tm IscUHis - HUMAN Fd EDC CROSS-LINKING

Proteins (40 µM) were incubated with 5 mM EDC in 50 mM sodium phosphate, pH 7.4 at room temperature for 2 h. The reactions were stopped by the addition of SDS-

PAGE loading buffer, boiled for 2 min, and then loaded onto a homogenous-20 pre-cast polyacrylamide gel.

3.2.19 D40A Tm IscU CLUSTER TRANSFER TO APO Hs Fd

Reactions were initiated by the addition of 0.1 mM holo D40A Tm IscUHis in 50 mM sodium phosphate, pH 7.4 to a solution containing 0.1 mM apo human ferredoxin and 5 mM DTT in the same buffer. Reactions were incubated on ice, stopped by the addition of loading buffer, and immediately frozen at –80o C. Reaction products were separated on a 7% native-PAGE and visualized by Coomassie Blue staining. This experiment was performed by G. Wu.

3.3 RESULTS

3.3.1 CLONING

T. maritima iscU (426 nucleotides long) was successfully amplified by PCR

(Figure 3.1), and inserted into pET28 and pET21 generating His-tag (Tm IscUHis) and non His-tag (Tm IscU) constructs, respectively. Potential clones were initially screened by EcoR I restriction digestion. Both pET plasmids contain one EcoR I site in their polyclonal region that is retained after Nde I - BamH I ligation with PCR amplified

63 T. maritima iscU. Additionally, T. maritima iscU has an EcoR I restriction site near its 5' end. Therefore, a positive clone yields two bands after EcoR I digestion with lengths of

~400 bp and ~5800 bp. Background plasmids, i.e. not containing insert, only show one large band after EcoR I digestion of ~5400 bp. Clones were confirmed by nucleotide sequencing.

3.3.1 PROTEIN EXPRESSION

Tm IscU expression was high for all constructs in BL21CodonPlus(DE3)-RIL, greater than that obtained from BL21(DE3), with typical yields of greater than 700 mg of pure protein from a 10 L fermentation. BL21CodonPlus(DE3)-RIL provides tRNAs for the rarely used E. coli codons of Arg, Ile, and Leu. These codons are more common in

T. maritima and are found within the iscU gene. Surprisingly, expressed WT Tm IscU and His-tagged WT Tm IscU (Tm IscUHis) were observed to run as two closely spaced bands during SDS-PAGE. These bands were found in a 1:1 ratio and persisted from the time of lysis through all purification steps (Figure 3.2) and give rise to similar mass spectrometric patterns (see 3.3.2). PMSF had no effect on the appearance or ratio of the two bands and so they do not seem to be a result of endoproteinase activity. D40A Tm

IscU and D40A Tm IscUHis only expressed as one band migrating at the same position as the lighter MW band of the corresponding WT protein.

Figure 3.1. Cloning of T. maritima iscU. The DNAs were analyzed using 1% (w/v) agarose gels stained with ethidium bromide. A) PCR products. Lane 1 is a 1 kb DNA marker. Lanes 2, 3, and 4 are aliquots of different PCR reactions using 25 ng, 50 ng, and

100 ng chromosomal DNA, respectively. B) EcoR I digest screening of transformants potentially carrying pTmIscUHis (top gel) or pTmIscU (bottom gel). Lane 1 is a kb

DNA ladder. Lanes 2 - 9 are potential clones. Lane 10 is EcoR I digested pET28, and lane 11 is undigested pET28. Positive clones are in lanes 2, 3, and 9 for pTmIscUHis, and lanes 2 and 8 for pTmIscU.

Figure 3.2. WT and D40A Tm IscU/His expression. Small 10 mL cultures were grown and induced as described in materials and methods and used for lanes 2 - 6. Lanes 3 – 6 are cleared lysates after incubation at 85o C for 0.5 h followed by centrifugation to remove insoluble material. Lane 1: Low MW markers, lane 2: cleared lysate containing

WT Tm IscU, i.e. not subjected to a heat step, lane 3: WT Tm IscU, lane 4: WT Tm

IscUHis, lane 5: D40A Tm IscU, lane 6: D40A Tm IscUHis, lane 7: WT Tm IscU after all the purification steps described in material and methods. The MW markers are labeled in kDa. Lane 7 is from a different lane of the same gel.

66 3.3.2 MASS SPECTROSCOPIC ANALYSIS

ESI spectra of each of the two bands identified by SDS-PAGE for WT Tm IscU

showed the same two major peaks corresponding to full-length protein and to protein

lacking the amino-terminal Met in approximately a 1:1 ratio (16070 Da and 15940 Da,

respectively) whereas D40A Tm IscU showed complete loss of the amino-terminal Met with a MW of 15893 Da. These values are in close agreement with their predicted MWs

of 16071 Da, 15940 Da, and 15894 Da, respectively. In-gel tryptic digests of the two

bands of WT Tm IscU yielded essentially identical MALDI-TOF spectra for all three

protein concentrations tested; that is, peptide fragments with identical MWs were

observed for both SDS-PAGE bands without the appearance or disappearance of any

signal. The search engine ProFound identified both bands as consisting of Tm IscU,

further confirming that neither of the bands was due to an impurity.

3.3.3 Tm IscU AGGREGATION STATE

Apo WT and D40A Tm IscU migrated as a monomer on a gel filtration column,

whereas holo D40A Tm IscU eluted as a dimer with apparent MWs of 18 kDa and 35

kDa, respectively (Figure 3.3A). No high MW aggregates were observed to elute. The

holo protein sample used was not entirely free of apo D40A Tm IscU, and so a red peak

eluted at an elution time consistent with a dimer and an additional colorless peak eluted at

a time corresponding to a monomer. To ensure that the differences in aggregation state

observed were not due to protein concentration, apo D40A Tm IscU was run twice, once

at a relatively lower concentration (0.3 mM) and once at high concentration (1.0 mM).

67 The elution profile of apo D40A Tm IscU at both concentrations were identical suggesting that the differences between apo and holo profiles are not due to protein concentration.

However, other techniques were consistent with a dimeric state for apo Tm IscU.

For example, dynamic light scattering of apo WT and apo D40A Tm IscU clearly indicated a single dimeric species in solution. Holo D40A Tm IscU dynamic light scattering data was non-monodispersive, indicating the existence of several aggregation states in solution that could not be accurately fit. The dimeric nature of apo D40A Tm

IscU was further confirmed by cross-linking. (Figure 3.3B). DSS is a homobifunctional

N-hydroxysuccinimide ester that forms covalent amide bonds between primary amines

(mostly ε-amine of lysine side-chain) 11 Å apart. Incubation of D40A Tm IscU with DSS resulted in the appearance of two strong bands corresponding to monomeric and dimeric states on a SDS-PAGE gel. Higher MW aggregates were observed at higher concentrations, but were significantly diminished upon dilution. Under conditions of high dilution, cross-linking occurs predominantly within an oligomer rather than between oligomers. Therefore the number of main bands observed suggests the number of subunits in a given protein (118).

Although the inconsistency of the gel filtration profiles with other data is unusual, a dimeric state of apo and holo D40A Tm IscU is consistent with subsequent NMR studies (see 4.3.7). Also, there exists a precedent for a lack of correlation in measurements of the aggregation state of multimeric proteins from gel filtration. Radaev et al. (119) have described studies of 3-deoxy-D-manno-octulosonate 8-phosphate

68 synthase that was initially characterized as a trimer by gel filtration analysis and later

demonstrated to be tetrameric. Additionally, a diversity of ISC aggregation states has

been observed. For example, human ISU is a monomer, E. coli and A. vinelandii IscU

are dimeric, A. vinelandii NifU is dimeric, and Helicobacter pylori NifU is a tetramer

(52, 74, 87, 120, 121) . Interestingly, similar problems in characterizing the aggregation

state of ISA have been encountered (76, 122). The difficulty in determining the

oligomeric state of ISC proteins is probably a reflection of their dynamic nature (see

4.3.7) and their involvement in protein - protein interactions.

3.3.4 CLUSTER COORDINATION

The bacterial pellets of induced cells were dark brown in color for all constructs.

Following centrifugation to remove solids, His-tagged WT and D40A Tm IscU lysates

yielded a red band that was bound by a Ni-NTA column. The non-His-tagged constructs

had reddish-brown lysates, but there was not color after the 85 oC purification step.

These results suggest proper Fe-S cluster assembly in vivo in E. coli, and that the His-tag does not interfere with Fe-S cluster coordination. Although a fraction of Tm IscU expressed as holo protein, the overall holo concentration appeared to be quite low as judged by UV-visible spectroscopy. Therefore, attempts were made to reconstitute Tm

IscU. Simple anaerobic incubation of apo D40A Tm IscU/His with DTT and a two-fold molar excess of iron and sulfide without the presence of denaturant resulted in a deep red protein with UV-visible spectra similar to [2Fe-2S] containing proteins (Figure 3.4).

Figure 3.3. Aggregation state of Tm IscU. A) FPLC gel filtration elution profiles.

Samples are: apo WT, 0.3 mM apo WT Tm IscU; apo D40A low, 0.3 mM apo D40A Tm

IscU; apo D40A (high), 1.0 mM apo D40A Tm IscU; holo D40A, 0.2 mM holo D40A Tm

IscU. The UV absorbance setting of the instrument was not equivalent for each sample.

Therefore, the peak intensities are not correlated to protein concentration between

different samples. The inset shows a plot of log MW vs. Kav. B) DSS cross-linking of apo D40A Tm IscU. Lanes 1 and 2 are low MW markers. Lane 2 - 5 represent cross-

linking of differing concentrations of protein. Lanes 2, 3, 4, and 5 represent 50, 25, 10, and 5 µM protein reactions, respectively.

70 Increasing the iron and sulfide concentration only resulted in the appearance of

adventitiously bound iron, as judged by Mössbauer analysis. Holo D40A Tm IscU was further purified by anion exchange chromatography. Apo D40A Tm IscU did not bind to

DE-52, whereas holo D40A Tm IscU was observed to bind. However, holo D40A Tm

IscUHis did not stick to DE-52, presumably due to the increased pI of the protein

resulting from the His-tag. Iron content was measured by atomic absorption

spectroscopy (monitoring the absorption at 305.9 nm) following control experiments to establish a standard calibration curve. Freshly purified holo D40A Tm IscU was found to contain 1.2 ± 0.1 Fe per monomeric subunit. Although we were able to reconstitute WT

Tm IscU, the yield of holo protein was low making biochemical characterization difficult.

This is in agreement with previous results indicating increased cluster stability upon an

Asp to Ala substitution at amino acid position 40 (Tm IscU numbering) (74, 80).

Although it is not currently understood why a mutation at this position increases cluster

stability, such a situation is not without precedent. A Leu to His substitution near one of

the cluster ligands of E. coli FNR greatly increases the stability of its [4Fe-4S] cluster

(81). As a result of the low yield of purified holo WT Tm IscU, and the relative instability of its cluster, all holo protein studies were carried out with reconstituted D40A

Tm IscU/His. In vivo cluster formation in WT Tm IscU is most likely facilitated by chaperones and IscS (84, 87, 120).

Figure 3.4. UV-visible spectra of 60.6 µM apo (dashed line) and 24.5 µM holo D40A Tm

IscU (solid line) in 50 mM Tris-HCl, pH 7.4.

3.3.5 MÖSSBAUER AND EPR SPECTROSCOPY

Mössbauer is a powerful technique to characterize iron containing proteins (5,

28). Mössbauer is a form of γ-ray resonance spectroscopy, that utilizes radiation emitted from the excited state of the 57Fe isotope (for Fe studies) to probe high-energy nuclear

72 transitions. 57Fe is quadrupolar with a nuclear spin of 3/2. The effect is that the quadruple moment splits the observed Mössbauer bands into doublets (quadrupole splitting). Their separation and location (isomer shift) are sensitive indicators of the oxidation and ligation states of the iron (5, 28).

Mössbauer spectra of 57Fe reconstituted D40A Tm IscUHis show a single iron- containing species with isomer shifts and quadrupole splitting of δ = 0.29 ± 0.03 mm/s and ∆EQ = 0.58 ± 0.03 mm/s, respectively at 100 K and δ = 0.27 ± 0.03 mm/s and ∆EQ =

0.56 ± 0.03 mm/s at 200 K (Figure 3.5). A similar spectrum was obtained at 4.2 K (δ =

0.28 ± 0.03 mm/s, ∆EQ = 0.56 ± 0.03 mm/s), with no hyperfine splitting observed, consistent with a diamagnetic diferric [2Fe-2S]2+ cluster, but cannot exclude the possibility of one or two non-cysteinyl cluster ligands (13). The data does, however, exclude formation of a [4Fe-4S] cluster. Holo D40A Tm IscUHis was EPR silent thus precluding the possibility of a rubredoxin-type center or a [3Fe-4S]+ cluster. Attempts to reduce and trap an EPR active species of holo D40A Tm IscUHis were unsuccessful as a result of the reductive lability of the cluster (54, 74, 87, 89). While reduction is likely to be rapid, we were unable to trap the subsequent reduced species prior to degradation by manual freeze quenching.

Figure 3.5. D40A Tm IscU Mössbauer spectra at 4.2 (A) , 100 (B), and 200 K (C).

74 3.3.6 UV-VISIBLE SPECTROSCOPY

Apo D40A Tm IscU had a λmax at 278 nm with an extinction coefficient of 15105

M-1 cm-1, which was determined by weighing a completely desalted and lyophilized

protein where the absorbance spectrum had previously been measured. Holo D40A Tm

IscU had λmax at 278 and 412 nm with shoulders around 320, 450, and 580 nm (Figure

3.4). The known protein concentration, determined by converting holo to apo protein,

yielded extinction coefficients of 41097, 22188, and 13854 M-1 cm-1 at 278 nm, 320 nm,

and 412 nm, respectively, for dimeric holo D40A Tm IscU. The absorption pattern and

relative intensities are similar to those of previously studied IscU proteins and human

ferredoxin (a [2Fe-2S] containing protein). However the absolute extinction coefficients

for the latter two cluster-centered transitions are lower than expected for a one cluster per

monomer ratio (52) and are consistent with one cluster per dimeric D40A Tm IscU.

3.3.7 THERMAL STABILITY OF THE [2Fe-2S] CLUSTER

Cluster loss from D40A Tm IscU and Hs Fd was monitored by visible absorption

(Figure 3.6). Hs Fd exhibited a typical profile with retention of cluster up to a critical temperature followed by rapid cluster loss, presumably due to loss of necessary structural elements. Hs Fd began to lose cluster at 54o C and was completely apo by 64o C. At higher temperatures Hs Fd began to precipitate even under anaerobic conditions as evidence by an increase in absorption due to light scattering. D40A Tm IscU showed more unusual thermal behavior. Almost immediately upon increasing the temperature the cluster of D40A Tm IscU began to degrade and was essentially absent by 55o C. It is

75 important to note that the cluster is thermodynamically unstable even at ambient temperatures, but takes longer to degrade. Figure 3.6 highlights the instability of the

IscU-bound cluster relative to the protein (compare below), and how the thermal stability of a ferredoxin cluster compares with that from IscU. At higher temperatures degradation is accelerated, and so a Tm calculation for such a thermal profile is meaningless since significant degradation occurs over a wide temperature range. Therefore, this experiment serves to illustrate the difference in thermal stability of the [2Fe-2S] cluster between these two proteins, rather than to assign a specific value.

Figure 3.6. Cluster thermal stability of D40A Tm IscU (solid line) and human ferredoxin

(dashed line). Conditions are as described in materials and methods.

76 3.3.8 CIRCULAR DICHROISM

The secondary structural content of apo D40A and apo WT Tm IscU was

determined by CD. Both were found to be have identical CD spectra indicating 36.7 %

α-helix, 13.1 % antiparallel β-sheet, 11.3 % parallel β-sheet, 20.2 % β-turn, and 19.1 %

other at 20 oC with negligible spectral changes observed at 70 oC (Figure 3.7). The

experimental error of 8.6 % in these measurements is defined as the difference between

the calculated and the experimentally determined CD spectrum. Interestingly, holo

D40A Tm IscU also had an identical CD spectrum consistent with no major structural

rearrangement upon [2Fe-2S] cluster coordination. The far-UV CD spectra of Holo

D40A Tm IscU were only acquired to 195 nm, as opposed to 180 nm for the other

samples, due to the presence of Tris-HCl rather than phosphate buffer. Since iron

coordinates to Tris-HCl to a significantly lesser extent than to phosphate, Tris-HCl was

used as a buffer for the holo D40A Tm IscU sample even though it is less transparent than

phosphate. In contrast to D40A Tm IscU, holo Hs Fd showed obvious structural changes

upon increased temperature from 20 oC to 60 oC with the negative peak shifting from 204

nm to 200 nm and the positive peak moving from 195 nm to 187 nm. A temperature of

60 oC was chosen to avoid protein precipitation as had been previously observed at higher temperatures (see cluster thermal stability). After allowing the Hs Fd sample to cool

slowly to 20 oC the spectrum looked intermediate to the 20 oC and 60 oC spectra with

negative and positive peaks at 203 nm and 188 nm, respectively, although the error in the

positive peak is large.

77 The near-UV-visible portion of the CD spectrum for holo D40A Tm IscUHis was

similar to that reported for E. coli IscU (83). The spectrum also has features in common

with human and Anabaena ferredoxin consistent with [2Fe-2S] coordination (84, 85). In

addition to the peak near 440 nm there is an additional peak near 480 nm for D40A Tm

IscUHis that is not present in WT ferredoxin. However, this peak is observed in C46S

Anabaena ferredoxin (85).

3.3.9 NMR

High resolution 1H-15N HSQC spectra for apo WT and D40A Tm IscU were

similar and showed significant dispersion of signals indicating the presence of substantial tertiary structure (Figure 3.8). Previous measurements in our laboratory of human and yeast homologues have shown 1H-15N HSQC spectra that lack dispersion of cross peaks, indicating the absence of significant tertiary structural elements (100, 101).

Figure 3.7. Circular dichroism spectra of D40A Tm IscU and Hs Fd. Panels A, B, and C are far-UV CD spectra. Panel D shows near-UV-visible CD spectra. Spectra were acquired at 20 oC (solid lines) and 70 oC (for Tm IscU) or 60 oC (for Hs Fd) (dashed lines). Panel A: apo WT Tm IscU, panel B: D40A Tm IscU (holo protein is shown displaced for ease of comparison with apo), panel C: Hs Fd (the middle spectral line is of protein cooled to 20 oC after incubation at 60 oC), panel D: upper spectrum is of holo

D40A Tm IscU and lower is of holo Hs Fd.

Figure 3.8. 1H-15N-HSQC of 4 mM apo WT (A) and D40A (B) Tm IscU in 50 mM

o sodium phosphate, pH 7.0, 10% D2O at 20 C.

81 3.3.9 CLUSTER TRANSFER TO HUMAN FERREDOXIN

Although no EDC cross-linking occurred between holo D40A Tm IscU and holo

Hs Fd, cluster transfer between equimolar concentrations of holo D40A Tm IscUHis and apo Hs Fd was easily achieved on ice (Figure 3.9). The cluster transfer reaction was

monitored by native PAGE, exploiting the difference in migration of apo and holo human

ferredoxin. Holo Hs Fd formation was quite rapid with completion of the reaction in less

than 10 min. At this temperature and time scale, no significant D40A Tm IscUHis cluster

degradation occurs. The yield of D40A Tm IscUHis mediated reconstitution of holo Hs

Fd was determined to be 80%. Cluster insertion into the apo Fd was most likely

precluded by disulfide bond formation. Furthermore, formation of holo Hs Fd is not

observed under conditions where standard solution reconstitution methods are used where free iron, sulfide and DTT are incubated with apo protein.

3.3.11 SEQUENCE COMPARISON

Figure 3.10 shows that Tm IscU is homologous to other IscU proteins from higher organisms (26% identity and 47% similarity to human ISU) and contains the three conserved Cys and the conserved Asp at position 40 (T. maritima numbering). However,

Tm IscU contains a ~18 amino acid insertion in the middle of the protein not found in most other prokaryotic or eukaryotic IscU proteins. This insertion is found within the

Bacillus subtilis IscU homologue named YurV (35% identity, 62% similarity to Tm

IscU). These insertions are clearly homologous with 21% identity and theoretical pI’s

4.1 and 4.2 for Tm IscU and B. subtilis IscU, respectively.

Figure 3.9. D40A Tm IscU - Hs Fd EDC cross-linking and cluster transfer. A) EDC cross-linking between holo D40A Tm IscU and holo Hs Fd. Lanes 2 - 3 are aliquots of reactions without EDC and lanes 5 - 7 represent reactions with EDC present. Lane 1, low

MW marker; lane 2, D40A Tm IscU; lane 3, Hs Fd; lane 4, D40A Tm IscU + Hs Fd; lane

5, D40A Tm IscU; lane 6, Hs Fd; lane 7, D40A Tm IscU + Hs Fd. B) Cluster transfer from holo D40A Tm IscUHis to apo Hs Fd. Lane 1, 40 µM apo Hs Fd; lane 2: 0.1 mM holo Hs Fd; lane 3: 0.1 mM holo D40A Tm IscUHis. Lanes 4 and 5 are cluster transfer reactions (50 µM holo D40A Tm IscUHis and 50 µM apo Hs Fd) terminated after 10 min and 1 h, respectively.

* TmIscU ----MVFKMMYSEAILDYANSKKFRGKLDDAT---VIEEGKNISCGDEITLYLKVE--DG 51 BsIscU MSFNANLDTLYRQVIMDHYKNPRNKGVLNDSI---VVDMN-NPTCGDRIRLTMKLD--GD 54 EcIscU ------MAYSEKVIDHYENPRNVGSFDNNDENVGSGMVGAPACGDVMKLQIKVN-DEG 51 AvIscU ------MAYSDKVIDHYENPRNVGKLDAQDPDVGTGMVGAPACGDVMRLQIKVN-EQG 51 HsISU1 MVLIDMS-VDLSTQVVDHYENPRNVGSLDKTSKNVGTGLVGAPACGDVMKLQIQVD-EKG 58 SpISU1 MVTANVSRRMYHKNVLDHYNNPRNVGTLPKGDPDVGIGLVGAPACGDVMRLAIRVN-KDG 59 AvNifU ------MWDYSEKVKEHFYNPKNAGAVEGAN---AIGDVGSLSCGDALRLTLKVDPETD 50

* TmIscU VVKDAKFEGMGCVISQASASLMLERIIGERVEEIFSLIEEAEKMSRGENFDEG-KLKNVT 110 BsIscU IVEDAKFEGEGCSISMASASMMTQAIKGKDIETALSMSKIFSDMMQGKEYDDSIDLGDIE 114 EcIscU IIEDARFKTYGCGSAIASSSLVTEWVKGKSLDEAQ------AIKNTD 92 AvIscU IIEDAKFKTYGCGSAIASSSLATEWMKGRTLEEAE------TIKNTQ 92 HsISU1 KIVDARFKTFGCGSAIASSSLATEWVKGKTVEEAL------TIKNTD 99 SpISU1 VIEDVKFKTFGCGSAIASSSYVTTMVKGMTLEEAS------KIKNTQ 100 AvNifU VILDAGFQTFGCGSAIASSSALTEMVKGLTLDEAL------KISNQD 91

* TmIscU LMSDIK-NYPARVKCFILAWKTLKEALKKISRP------142 BsIscU ALQGVS-KFPARIKCATLSWKALEKGVAKEEGGN------147 EcIscU IAEELE-LPPVKIHCSILAEDAIKAAIADYKSKREAK------128 AvIscU IAEELA-LPPVKIHCSVLAEDAIKAAVRDYKHKKGLV------128 HsISU1 IAKELC-LPPVKLHCSMLAEDAIKAALADYKLKQEPKKGEAEKK------142 SpISU1 IAKELC-LPPVKLHCSMLAEDAIKSAVKHYRSKQLTPVGTTAGAIESATA------192 AvNifU IADYLDGLPPEKMHCSVMGREALQAAVANYRGETIEDDHEEGALICKCFA------147

Figure 3.10. IscU/NifU sequence alignment. Inverted text are amino acid identities and bold positions are highly conserved residues. The conserved Cys are identified with a *.

Only the IscU homologous region of NifU is shown. Organisms are as follows: Tm,

T. maritima; Bs, B. subtilis; Ec, E. coli; Av, A. vinelandii; Hs, human; Sp,

Schizosaccharomyces pombe.

84 3.3.10 D55A Tm IscU

The remaining ligand to the [2Fe-2S] cluster of IscU proteins is not known. To try to identify the nature of this remaining ligand we were guided by the following observations. One, although His coordination to [2Fe-2S] clusters has been previously observed, Tm IscU does not contain a single His residue. Second, some spectroscopic and stability data suggest that oxygen may be a ligand to the cluster. Third, the only naturally occurring protein bound oxygen Fe-S cluster ligand identified is provided by an

Asp side-chain, even though these cases are for [4Fe-4s] clusters. Fourth, Asp55 is a highly conserved IscU residue (Figure 3.10). Therefore, we generated a His-tagged

D55A Tm IscU construct (D55A Tm IscUHis) to probe whether Asp55 is the remaining ligand. A fraction of D55A Tm IscUHis recombinantly expressed in the holo form, thus indicating that Asp55 is not likely to be an Fe-S cluster ligand (Figure 3.11).

Interestingly, D55A Tm IscU His runs as a doublet on SDS-PAGE similarly to WT protein.

Figure 3.11. UV-visible spectrum of D55A Tm IscUHis in binding buffer + 100 mM

imidazole. The inset is a SDS-PAGE of the protein.

3.4 DISCUSSION

3.4.1 PROTEIN EXPRESSION

The expression and purification of D40A Tm IscU/His yielded large amounts of

high purity protein with no complications. However, the expression of WT Tm IscU/His

was more problematic. Although, large amounts of highly purified WT Tm IscU/His

were easily obtained, they were consistently present as two forms as judged by SDS-

86 PAGE. The observance of two bands suggests some sort of post-translational

modification, but we were only able to identify differences in amino-terminal Met

processing. Small highly charged proteins have often been observed to behave oddly on

SDS-PAGE (123, 124); however, WT Tm IscU and WT Tm IscUHis have predicted pI’s of 5.7 and 7.1, respectively. The γ-subunit of T. maritima hydrogenase also

recombinantly expresses as two forms with a MW difference of greater than 7 kDa (125).

The difference was shown to be due to protein cleavage as judged by mass spectroscopy

and amino-terminal sequencing and was not dependent upon a heating step during

purification (125). This cleavage site is not found within Tm IscU. Mass-spectrometric

analysis of an A. vinelandii ferredoxin cloned from its isc operon revealed that

recombinant expression in E. coli yielded protein 52 Da larger than native protein,

although the recombinant protein ran faster on a SDS-PAGE (126). The difference in

MW was not due to amino-terminal cleavage as judged by amino acid sequencing and

was not definitively characterized (126). The most reasonable explanation for the double

band observed for WT Tm IscU, given the fact that each band is essentially the same

species by mass spectrometric analysis, is that the protein binds SDS in two

stoichiometries and therefore runs as a doublet, as previously observed and documented

for the E. coli outer cell membrane protein OmpA (127, 128). It is interesting to note that

a single point mutation known to stabilize cluster coordination yielded Tm IscU in a

single form, perhaps by stabilizing a specific “conformer”.

87 3.4.2 CLUSTER COORDINATION

Reconstitution of D40A Tm IscU/His with iron and sulfide yielded holo protein with Mössbauer, UV-visible absorption, and CD spectra characteristic of [2Fe-2S] coordination. The UV-visible spectra are similar to other reported IscU proteins in that their λmax are roughly in the same positions. However, holo D40A Tm IscU had sharper, better defined peaks between 400 and 600 nm. Although the near-UV-visible CD spectrum had features indicative of a [2Fe-2S] cluster and was similar to WT ferredoxin

(both Anabaena and human), the wavelengths and relative intensities of CD bands of peaks and troughs in the spectrum of holo D40A Tm IscUHis looked much like those observed for C46S Anabaena ferredoxin in which one of the Cys cluster ligands is substituted with a Ser (85). No naturally occurring [2Fe-2S] proteins to date have been

shown to have an oxygen ligand, although primary amino acid sequence analysis of some

[2Fe-2S] proteins suggests that it may exist (86). IscU proteins have three conserved Cys

that upon mutation result in loss of Fe-S cluster coordination (74, 80). The nature of the

remaining ligand has yet to be identified. Possibilities include solvent or an

oxygen/nitrogen amino acid side-chain, however, the absence of His in Tm IscU makes

nitrogen ligation unlikely. Resonance Raman results from A. vinelandii IscU are

consistent with either complete Cys ligation or single serinate ligation (87). Other than

Asp40, the conserved oxygen side-chains derive from Asp13, Asp55, Ser69, Ser71, and

Glu83. Asp55 is unlikely to be a ligand since D55A Tm IscUHis expresses as holo

protein (see 3.3.10). Future structural and mutagenesis studies may help to better define

the coordination environment of IscU-type clusters.

3.4.3 CLUSTER STOICHIOMETRY

The experimentally determined extinction coefficients suggest one cluster per dimer, consistent with direct estimation of the iron content of D40A Tm IscU to be 1.2 ±

0.1 Fe per monomer. Iron concentrations were accurately determined by atomic absorption and protein concentrations determined from the experimental extinction values. The occurrence of low concentrations of adventitiously bound iron, evidenced by

Mössbauer data, and small losses in protein during desalting steps account for the slightly higher Fe : protein ratio identified. Correction for this results in close agreement with a single [2Fe-2S] cluster per dimer of protein. Agar et al. have isolated forms of A. vinelandii IscU with both one and two clusters per dimer and have shown that the one cluster form is more stable to iron chelators (54). Thus far, the only monomeric IscU identified is the human ISU, which binds a single cluster per monomer (74); however, other dimeric IscU-type proteins all appear to coordinate one cluster per protein subunit.

Accordingly, the cluster stoichiometry for the holo form of Tm IscU may need to await further structural characterization.

3.4.4 THERMAL STABILITY

The thermal stability of the [2Fe-2S] cluster in D40A Tm IscU is distinct from that of other ferredoxin-type clusters in Fe-S proteins from both thermophilic as well as mesophilic organisms. The [4Fe-4S] cluster of T. maritima ferredoxin is stable past the boiling point of water with a calculated transition temperature of 125 oC (129). Even the

89 [2Fe-2S] cluster of human ferredoxin shows greater thermal stability than D40A Tm

IscU. The thermal profile of human ferredoxin is typical of Fe-S cluster proteins in

which no cluster is lost up until a specific temperature is reached, after which the cluster

is rapidly released following loss of necessary structural elements. Indeed, the thermal

cluster profile of human ferredoxin corresponds to changes in structure observed by CD

and is consistent with what is currently known regarding the influences of metal cofactor

coordination on proper protein folding (130, 131). Interestingly, D40A Tm IscU did not

exhibit a thermal profile similar to any Fe-S protein characterized thus far. Cluster loss

was evident over a wide temperature range proceeding without an obvious thermal

threshold. Such behavior is particularly surprising considering that T. maritima can

survive temperatures up to 90 oC (103) and suggests a stabilizing role by other proteins

(such as chaperones) in the intracellular environment. Furthermore, no structural change

was observed by CD over the tested temperature range or following cluster coordination.

Accordingly, unlike many other Fe-S cluster binding proteins, cluster loss does not coincide with loss of secondary and tertiary structural elements. The data is also consistent with the ability to reconstitute D40A Tm IscU in the absence of low levels of denaturant that are occasionally required to allow the protein to properly fold around the cluster (77).

90 3.4.5 T. MARITIMA IscU RECONSTITUTION OF APO HUMAN FERREDOXIN

Cluster transfer between two proteins from extremely distant organisms such as

T. maritima and Homo sapiens demonstrates the high degree of conservation of the Fe-S

cluster assembly apparatus and is consistent with the notion of Fe-S proteins being evolutionarily ancient (11, 132). The cluster transfer reaction from D40A Tm IscUHis to

apo human ferredoxin was rapid even at low temperatures where background IscU cluster

loss is negligible. Apo human ferredoxin retains much of its secondary structure (Figure

3.7) (84) that presumably is recognized by IscU. This recognition motif must be

conserved for such divergent sources of IscU to be competent in assisted Fe-S cluster

maturation of ferredoxin. This is the first reported example of cluster transfer from an

IscU-type domain to a target ferredoxin. It has been previously shown, in an example

from a distinct protein family, that E. coli IscA can transfer an Fe-S cluster to apo E. coli

ferredoxin (122), and that a protein homologous to the carboxy-terminal domain of NifU

can similarly transfer a cluster to ferredoxin (112). The differences in the physiological

roles between these different protein families are not currently known.

3.4.6 STRUCTURAL CONSIDERATIONS

The circular dichroism (Figure 3.7) and NMR (Figure 3.8) data reveal for the first time structural information regarding this family of proteins. Other IscU family members that have been studied in our laboratory (namely, human and yeast homologues) lack CD and NMR features that support secondary and tertiary structural elements. Most likely this reflects conformational flexibility that is of inherent functional relevance for these

91 proteins. Our expectation that such a protein from a thermophilic organism might

possess a degree of structural stability that is often associated with proteins from such organisms has been borne out. The ability of the holo Tm IscU to functionally transfer

cluster to a human apo Fd speaks to the functional equivalence of this family of proteins.

Undoubtedly other family members adopt key structural and functional states, but

typically these are not detected by standard structural methods. The thermophile

apparently diminishes the conformational flexibility sufficiently to allow these states to

be identified.

CHAPTER 4

NUCLEAR MAGNETIC RESONANCE OF THERMOTOGA MARITIMA IscU

4.1. INTRODUCTION

The proteins involved in the maturation of copper, nickel, zinc, heme, and iron- sulfur cluster containing proteins have received considerable interest (30, 44, 45, 54, 58,

74, 133). Of these systems only the structures of copper, nickel, and zinc metallochaperones have been solved to date (34, 42, 45-47, 134-137). Additionally, the structures of proteins involved in mercury (35) and heme transport have been elucidated

(138). All of the metallochaperone structures solved thus far share the presence of a ferredoxin-like βαββαβ motif (45, 139, 140).

In contrast to the examples just cited, the polynuclearity of Fe-S clusters requires a biosynthetic apparatus that is considerably more complex for assembly and delivery of such centers. Not only does the cell have to acquire potentially toxic inorganic components, they must be used to synthesize and deliver an Fe-S cluster to target partner proteins. The proteins involved in Fe-S cluster assembly belong to the Isc (for iron-sulfur

cluster) family and include IscU, IscS, and IscA (58). Additionally, a transcriptional

93 repressor (141), a [2Fe-2S] ferredoxin (70, 75), and molecular chaperones (59, 120) have

been implicated in Fe-S cluster biogenesis. IscS provides sulfur equivalents to IscU via

catalytic cysteine desulfurization (51, 53, 87, 89). IscU is believed to function as a

platform for the assembly of a nascent [2Fe-2S]2+ cluster that is subsequently delivered to

target apo Fe-S proteins (54, 73). IscU coordinates a reductively labile [2Fe-2S]2+ cluster

that is stabilized by substitution of a highly conserved Asp with Ala at amino acid

position 40 (Thermotoga maritima numbering) (58, 73, 74, 89). IscU-like proteins are extremely evolutionarily conserved, spanning all three kingdoms of life (55). In eukaryotes, Isc proteins are localized within the mitochondrion and prove to be essential for the maturation of both intra- and extra-mitochondrial proteins (67, 90). The role of

IscA is less clear and appears to be non-essential, possibly providing a redundant function to IscU (68, 69, 76, 122, 142, 143).

To date the structural information regarding IscU is quite limited, a likely reflection of its intrinsic folding properties. With the hope of identifying a model IscU system for detailed structural examination, we had earlier cloned and characterized an

IscU from the hyperthermophile T. maritima (Tm IscU) (see Chapter 3 and (73)). Tm

IscU showed CD and 1H-15N HSQC NMR spectra consistent with a well-folded protein

that was not dependent on Fe-S coordination. Furthermore, Tm IscU could functionally

substitute for human IscU in vitro, as demonstrated by Fe-S cluster transfer from Tm IscU

to human ferredoxin (73). Herein we report further NMR characterization of apo and

holo D40A Tm IscU and demonstrate unique structural and motional characteristics that

distinguish it from other members of this class of proteins. In particular, IscU adopts a

94 mobile molten globule-like state that is vastly different from the previously identified

ferredoxin-like fold that has thus far been characterized for other metallochaperones.

Such a dynamic molecule may allow for the flexibility that is necessary for the multiple roles of Fe-S cluster assembly, and recognition and delivery of that cluster to a target protein.

4.2 MATERIALS AND METHODS

4.2.1 GENERAL CHEMICALS

15 13 D2O, NH4Cl, and C-glucose were from Isotec. 2,2,2-trifluoroethanol (TFE)

was from Fluka.

4.2.2 PROTEIN EXPRESSION AND PURIFICATION

BL21(DE3)CodonPlus-RIL (Stratagene) D40A pTmIscU was used for the

expression of unlabeled samples as previously described (73). Preparation of labeled apo

and holo D40A Tm IscU was as previously reported for unlabeled samples, except that

immediately prior to induction the cells were gently pelleted and resuspended in M9

minimal media (78) supplemented with 20 µg/ml ampicillin, 25 µg/ml chloramphenicol,

15 and the appropriate isotopically labeled nitrogen (0.1 g NH4Cl/L) and carbon sources

(0.4 g 13C-glucose/L). Bacterial cultures were induced with 1 mM IPTG and harvested

12 h post induction. While the concentrations of labeled nitrogen and carbon sources

were significantly lower than typically employed for isotopic enrichment, trial runs using

both isotope sources established concentration regimes that were satisfactory for

production of high levels of labeled protein. Tm IscU was purified as previously described (73).

4.2.3 NMR SPECTROSCOPY

NMR spectra were acquired on Bruker Avance spectrometers at 700 MHz with

1H, 13C, 15N triple-resonance probe heads (TXI) with self-shielded triple axis gradients

and at 500 MHz with a TXI cryoprobe with a self-shielded z-axis gradient. Quadrature

detection in the indirect dimensions was used and water suppression was achieved via

WATERGATE (144). Solution conditions typically were 1.8 mM apo D40A Tm IscU,

450 mM NaCl, pH 5.4, 10% D2O at 318 K, unless otherwise indicated. Further experimental parameters are listed in Table 4.1.

Triple resonance experiments were used to assign the backbone. These included

CBCANH (145), CBCA(CO)NH (145) and HNCA (146) at 700 MHz and a TROSY-

HNCO (147) at 500 MHz. Side-chain carbon and proton assignments were made by a

(H)CCH-TOCSY (148) at 700 MHz.

HNHA (149), HNHB (150), and 1H-15N-NOESY-HSQC (151) spectra at 700

MHz allowed for torsion angle calculations. Using the HNHA spectrum the backbone

3 dihedral ϕ angles were derived from JHNHα coupling constants determined from the

3 appropriate Karplus relationship. More specifically, JHNHα values larger than 7 Hz were

constrained to ϕ angles between –155o and –85o and for those lower than 4.5 Hz the ϕ

96 angles were constrained within –70o and –30o (151). The ψ angles for residues (i-1) were

1 15 determined from NOE peak intensity ratios of dαN(i-1,i) and dNα(i,i) in a H- N-NOESY-

HSQC. Ratio values of residue i-1 greater than one are indicative of β-sheets with ψ

values ranging between 60o and 180o, whereas values smaller than one indicate a right

handed α-helix with ψ values between –60o and –20o (152).

NOEs were measured from 2D NOESY, 3D 1H-13C-HSQC-NOESY (with a

mixing time of 80 ms) (151), and 3D 1H-15N-HSQC-NOESY (151) (with a mixing time

of 95 ms) spectra collected at 700 MHz. Additionally for screening purposes, 2D

NOESY spectra of apo D40A Tm IscU were recorded under the same solution conditions

listed above except for the following separate changes: 298 K (at 500 MHz), 343 K (at

600 MHz), pH 7.4 (at 700 MHz), and 5% TFE (at 700 MHz). Also, a 2D NOESY of 1.8

mM (based on monomer) of apo and holo Asp40Ala Tm IscU in 50 mM sodium

phosphate, pH 7.4, 200 mM NaCl, 10% D2O at 500 MHz and 293 K was acquired.

Triple-resonance TROSY-HNCO experiments were acquired in both three- and two-

dimensional (1H-13C) modes (153-155) to directly detect hydrogen bonds (156).

4.2.4 RDC

Solution conditions were 1.8 mM apo D40A Tm IscU, 450 mM NaCl, 5% penta-

ethyleneglycol dodecyl ether, 1% n-hexanol, 10% D20, pH 5.4 (157). Spectra were acquired at 700 MHz and 300 oC. The pulse sequence was as previously reported (158).

Experiment Dimensions of acquired data (nucleus) Spectral width (Hz)

t1 t2 t3 F1 F2 F3

1H-15N-HSQC 128(15N) 1024(1H) 2941 10000

[1H-1H]-NOESY 1024(1H) 2048(1H) 8389 8389

2D TROSY-HNCO 200(13C) 2048(1H) 2012 6010

long-range TROSY-HNCO 100(13C) 2048(1H) 2012 6010

2D HNHB 128(15N) 2048(1H) 2941 7575

CBCANH 140(13C) 40(15N) 1024(1H) 14084 2838 8389

CBCA(CO)NH 140(13C) 40(15N) 1024(1H) 14084 2838 8389

HNCA 140(13C) 48(15N) 1024(1H) 7143 2838 8389

TROSY-HNCO 30(13C) 32(15N) 2048(1H) 2012 2027 8013

HNHA 256(1H) 48(15N) 2048(1H) 8389 2841 8389

3D HNHB 140(1H) 30(15N) 2048(1H) 7576 3125 7576

(H)CCH-TOCSY 220(13C) 48(13C) 2048(1H) 12500 12500 8389

1H-15N-HSQC-NOESY 256(1H) 40(15N) 2048(1H) 7936 2909 7936

1H-13C-HSQC-NOESY 256(1H) 72(13C) 2048(1H) 7932 14286 7936

15 1 R1 128( N) 1024( H) 2027 6010

15 1 R2 128( N) 1024( H) 2027 6010

RDC 512(15N) 1400(1H) 2909 7716

15N-NOE 256(15N) 1024(1H) 2027 8012

15 1 D2O EXCHANGE 128( N) 2048( H) 2027 7003

Table 4.1. Acquisition parameters for NMR data sets collected on Tm IscU.

98 4.2.5 SECONDARY STRUCTURE DETERMINATION

Regions of secondary structure were determined by a combination of factors. The

method of Wishart and Sykes (159) was used for chemical shift analysis. Subsequently, chemical shift data, calculated backbone dihedral angles of ϕ and ψ, RDC data, NOE patterns, and directly detected hydrogen bonds were used to define regions of secondary structure.

4.2.6 15N BACKBONE RELAXATION MEASUREMENTS

A series of 1H-15N HSQC spectra were collected with flip-back (160) and phase

15 sensitive (161-163) pulse sequences at 500 MHz. N longitudinal relaxation rates, R1, were measured with delays of 2.0, 25, 50, 100, 200, 400, 600, 900, 1200, and 1500 ms.

15 A total of 8 scans was collected for each experiment. N transverse relaxation rates, R2, were measured with delays of 18.4, 36.8, 55.2, 73.6, 92.0, 110.4, 128.8, 147.2, 165.6,

184.0, 220.8, and 257.6 ms using τm of 450 µs. A total of four scans were collected for each delay time. Heteronuclear 1H-15N NOEs were measured by taking the ratio of peak

1 volumes acquired with and without H saturation (160). Relaxation rates R1 and R2 were

calculated by fitting cross-peak intensities as a function of the delay time to a single

exponential decay using the Levenburg-Marquardt algorithm (164, 165). Uncertainties

were determined by using a Monte Carlo approach (166).

99 4.2.7 HYDROGEN EXCHANGE

Concentrated apo D40A Tm IscU was rapidly diluted with D2O to a final concentration of 93% (v/v). A 1H-15N HSQC (167) spectrum was acquired after 1 h, 18 h, and one week. It has been our experience with IscU proteins from several sources that they do not survive lyophilization. However, Tm IscU is highly soluble thus making the experiment possible by rapid dilution of concentrated protein with D2O.

4.2.8 DATA PROCESSING AND STRUCTURE CALCULATIONS

NMR data were processed with XWINNMR (Bruker) and analyzed with XEASY

(ETH, Zurich) (168). The 15N and 13C-NOESY-HSQC and 2D NOESY cross-peaks were integrated and converted into inter-proton upper distance limits using the program

CALIBA (169). The calibration curves obtained were iteratively adjusted as the structure calculations proceeded.

Structure calculations were performed by simulated annealing calculations within torsion angle dynamics simulations with DYANA (170) and CYANA (171) programs.

One hundred random conformers were annealed in 10000 steps using NOE, dihedral angle constraints (obtained as described above) and hydrogen bonds. H-bond constraints were incorporated into the structure calculations as upper and lower distance limits. The upper limit between the NH proton and the hydrogen bond acceptor was set to 2.4 Å, while distances of 2.6 Å (lower) and 3.3 Å (upper) were used between the N and the acceptor atoms. In addition to the directly detected H-bonds (see above), any H-bond

100 donor-acceptor pair that was found to have the right distance and orientation in all initial output structures was subsequently constrained to within hydrogen bond distances in later

runs. The program MOLMOL was used for structure display (172).

4.3 RESULTS AND DISCUSSION

4.3.1 SOLUTION CONDITIONS

The D40A derivative of Tm IscU was used for NMR experiments as a result of

the increased stability of the [2Fe-2S] cluster (73), which allowed for direct comparison between apo and holo protein. Amino acid substitutions of non-ligating residues has occasionally been found to stabilize Fe-S clusters (73, 74, 81, 89). Wild type and D40A

Tm IscU are likely structurally equivalent as reflected by nearly identical 1H-15N HSQC

spectra (73)), and thus cluster stability is probably a reflection of solvent accessibility

(82). Initial screening of solution conditions was monitored by 1H-15N HSQC and 2D

NOESY spectra. Although apo Tm IscU is soluble to greater than 6 mM, the 1H-15N

HSQC spectral line widths significantly narrowed at decreased protein (1.8 mM) and

increased NaCl (0.45 M) concentrations. Indeed, 3.3 mM apo D40A Tm IscU at 25 oC

and 300 mM NaCl displays only two thirds of the 1H-15N HSQC backbone cross-peaks

seen under optimal solution conditions. Lower concentrations of protein were not

routinely used, because trials using lower concentrations yielded lower quality spectra

and fewer NOEs. The spectral quality was further increased by elevating the temperature

to 45 oC (a 1H-15N HSQC spectrum at high protein concentration and low temperature is

101 shown in Figure 3.8). For example, few 2D NOESY cross-peaks were observed at room

temperature. However, at 45 oC, the number of intra- and inter-residue NOEs was

significantly increased. Temperatures higher than 45 oC did not further increase the

number of NOEs. A pH of 5.4 was chosen to facilitate labile amide proton detection.

Indeed, NOESY spectral quality was decreased at pH 7.4. Low concentrations of TFE also did not increase the number of detectable NOEs. The high salt and high temperature conditions are consistent with the conditions known to be necessary for optimal

T. maritima growth (103, 173).

4.3.2 CHEMICAL SHIFT DISPERSION AND RESONANCE ASSIGNMENTS

The proton spectrum showed good chemical shift dispersion with reasonably defined peaks from -0.05 ppm to 9.55 ppm, including the aromatic side-chain and backbone amide regions. The 1H-15N HSQC spectrum also contained chemical shift dispersion indicative of a well-folded protein with amide backbone 1H resonances between 9.55 (Thr43) and 5.19 (Leu24) (Figure 4.1). Excluding side chain resonances, approximately 133 of the 139 possible backbone amide peaks (accounting for the processed amino-terminal Met and two Pro) were observed in the 1H-15N HSQC spectrum.

102

Figure 4.1. 1H-15N HSQC of apo D40A Tm IscU in 450 mM NaCl, pH 5.4 at

45 oC and 700 MHz (A) and an expansion of the spectrum from 5.0 to 10.0 ppm in the proton dimension and 100.0 to 132.0 ppm in the nitrogen dimension (B). Assigned peaks are labeled with their corresponding residue number.

103

104 Sequential backbone connectivities relied on 3D spectra, including: CBCANH,

CBCA(CO)NH, and HNCA (Figure 4.2). These spectra further allowed for the assignments of Cα, Cβ, N, and NH resonances. HNHA and HNHB spectra were used to assign Hα and Hβ, respectively. Side chain resonances were assigned via (H)CCH-

TOCSY and the carbonyl C’ was assigned via TROSY-HNCO. Excluding the post- translationally processed amino-terminal Met (73) nine amino acids were not assigned in any of the NMR spectra (V2, L12, D13, K18, Y45, F86, K107, N108, and T131).

Furthermore, none of the side-chain resonances of L24, L44, V64, I65, L74, L111, N117, and W129 were detected. The backbone resonances of these residues were identified by the more sensitive HNCA spectrum. It is interesting to note that a disproportionate number of hydrophobic residues were represented within this category. Specific regions of the protein contained double resonances (separated by ~20 Hz) for each detected atom of the amino acid indicating an exchange between two conformations on a time scale longer than ms. These amino acids were 3, 23-25, 33, 42-43, 67-71, 74, 78, 93-94, 109,

115-117, 129, 132, and 135. Several of these residues also lacked side-chain resonances.

At lower temperature (25 oC) 44 of the previously assigned backbone amide resonances

were no longer present in the CBCANH. Furthermore, several residues that exhibited

double resonances at 45 oC were not detected at 25 oC (24, 42-43, 69, 71, 74, 116, and

129). The residues with double resonances at 25 oC were 8, 11, 33, 35, 37, 39, 40, 50,

54-55, 59-60, 63, 68, 87, 93, 115, 128, 134, and 135. For Lys19 and Phe86, only their

CO resonances were assigned. A summary of 1H, 15N, and 13C resonance assignments is

listed in Tables 4.2 and 4.3.

105

Figure 4.2. 2D planes from 3D CBCANH and CBCA(CO)NH spectra of apo D40A Tm

IscU. The Cα and Cβ of i and i-1 residues are present in CBCANH spectra, and the Cα and Cβ of the i-1 residue is shown in the adjacent CBCA(CO)NH spectrum. Connections between the Cα of adjacent residues are shown by a line.

106

Table 4.2. Nitrogen and proton chemical shift assignments (ppm) of apo D40A Tm IscU.

Chemical shifts for the second conformation (where available) are in parenthesis.

107

N NH HA HB HG HD HE

1 MET

2 VAL

3 PHE 122.09 8.08 4.14 1.43 (8.12) 4 LYS 117.72 8.66 3.81 2.11 1.27,1.60

5 MET 117.72 7.65 4.28 2.43 2.36,2.27

6 MET 117.72 8.13 4.20 2.33,2.20 2.44,2.84

7 TYR

8 SER 116.78 7.99 4.53 3.12

9 GLU 123.40 8.17 4.85

10 ALA 122.72 8.33 3.50 0.16

11 ILE 118.08 8.38 3.84 2.06 2.46(QG1) 2.62(QG2) 12 LEU

13 ASP

14 TYR 4.38

15 ALA 121.78 8.93 4.04 1.86

16 ASN 112.72 7.97 4.95 3.12,2.95

17 SER 114.59 7.47 4.44 4.31,4.14

18 LYS

19 LYS

Continued

108 Table 4.2 continued.

20 PHE 124.28 8.17 4.22 2.11,1.94

21 ARG 117.72 7.03 5.14 1.68,1.85 1.65 3.09

22 GLY 112.72 8.48 4.39,3.74

23 LYS 112.41 7.44 3.34 1.22 1.06 1.65 3.09 (7.4) (3.37) (1.12) (1.1) (1.62) 24 LEU 117.41 5.20 4.72 1.51,1.62 1.12 1.11 (5.16) 25 ASP 124.28 8.86 4.59 2.68,2.77 (8.92) (4.61) (2.7,2.79) 26 ASP 117.72 8.51 4.79 2.70

27 ALA 118.97 7.40 3.99 1.68

28 THR 119.28 9.34 4.39 4.22

29 VAL 115.84 7.94 4.62 2.11 1.14(QG1) 1.04(QG2) 30 ILE 126.78 8.10 5.07 1.52 0.99(QG1) 1.13 0.48(QG2) 31 GLU 128.03 8.72 4.87 1.85,2.04

32 GLU 126.78 8.80 5.49 2.19,2.02 2.34

33 GLY 111.78 9.46 4.77,3.80 (9.34) (4.74,3.85) 34 LYS 120.84 8.74 5.20 2.02,1.86 1.50 1.80 3.07

35 ASN 117.72 8.61 5.12 2.89

36 ILE

37 SER 119.59 8.47 5.01

38 CYS 117.41 7.99 4.50 3.88,4.14

39 GLY 109.59 8.18 4.06,3.97

40 ALA 117.09 7.72 5.70 1.52

Continued

109 Table 4.2 continued.

41 GLU 120.53 8.82 5.14 2.28

42 ILE 120.53 9.25 5.40 1.87 1.68(QG1) 0.72 (9.16) (5.46) (1.89) 0.95(QG2) (0.76) (1.71, 0.92) 43 THR 123.34 9.55 4.96 4.19 0.86(QG2) (9.48) (4.99) (4.23) (0.92) 44 LEU 125.53 9.32 5.02 1.48

45 TYR

46 LEU 4.76 1.99 0.77

47 LYS 124.91 8.81 5.05 1.69 1.49 1.88,1.85 3.03

48 VAL 128.03 9.04 5.14 2.09 0.91(QG1) 0.90(QG2) 49 GLU 126.16 8.95 4.79 2.11,1.94 2.32

50 ASP 125.84 9.50 5.00 3.19, 2.82

51 GLY 101.73 8.67 3.72,4.37

52 VAL 120.53 8.03 4.74 2.14 0.84

53 VAL 123.97 8.24 4.35 2.44 0.94,0.82

54 LYS 131.47 8.75 4.37 2.02,1.68 1.65 1.80 3.17

55 ASP 112.72 8.09 4.99 2.87,2.48

56 ALA 122.41 8.70 5.42 1.43

57 LYS 118.97 9.01 5.33 1.85 1.23,1.33 1.24,1.62 3.15,3.03

58 PHE 116.78 9.22 6.45 3.21,2.95

59 GLU 118.66 9.11 4.82 2.27 2.58,2.35

60 GLY 108.66 8.41 4.39,3.73

61 MET 117.09 8.66 5.02 2.11 2.70

Continued

110 Table 4.2 continued.

62 GLY 106.16 8.27 4.51,4.27

63 CYS 117.09 8.06 4.02 3.21

64 VAL 118.34 8.41 3.81 2.13 0.98,0.73

65 ILE 119.28 8.85 4.14 1.94

66 SER 117.41 9.18 4.69 3.63

67 GLN 119.91 7.98 4.21 2.18 2.54,2.64 (7.94) (4.19) (2.21) 68 ALA 122.09 9.00 4.48 1.77 (9.06) (4.52) (1.72) 69 SER 108.03 8.33 4.23 4.04 (8.38) (4.2) 70 ALA 118.34 7.17 3.34 0.50 (7.23) (3.37) (0.44) 71 SER 112.72 7.54 4.25 4.48,4.31 (7.6) (4.23) 72 LEU 120.96 8.47 3.78 0.77 1.33 0.33

73 MET 124.59 8.19 4.49 2.49,2.44

74 LEU 120.53 8.31 3.83 2.10 (8.25) 75 GLU 116.16 7.60 4.36 2.19,2.02

76 ARG 115.53 7.50 4.67 2.19

77 ILE 110.22 8.22 4.25 2.11 1.54(HG12) 1.02(QG2) 1.33(HG13) 78 ILE 121.47 7.53 3.62 1.98 2.03(HG12) 1.13(QG2) 1.28(QD1) (7.49) (3.65) (2.02) 0.98(HG13) (1.1) (1.26) (2.06, 0.95) 79 GLY 116.47 9.34 4.56,3.71

80 GLU 118.97 8.17 4.84 2.28,2.11 2.32

81 ARG 119.91 9.20 5.02 2.36,2.26 1.98,2.11 3.57,3.49

82 VAL 123.66 7.84 3.69 2.37 1.30,1.13

Continued

111 Table 4.2 continued.

83 GLU 115.53 9.36 4.44 2.28,2.11 2.62,2.57

84 GLU 118.34 7.16 4.65 2.27,2.45 2.64,2.48

85 ILE 121.47 8.47 3.78 1.94 1.40 (QG1) 0.98 1.30 (QG2) 86 PHE

87 SER 117.41 7.99 4.67 3.38,3.12

88 LEU 123.34 8.17 4.84 3.42 2.09 1.91

89 ILE 117.41 8.67 4.10 1.94 1.87

90 GLU

91 GLU 115.22 9.36 4.44 2.36,2.11 2.01

92 ALA 121.16 8.50 3.78 1.01

93 GLU 118.66 6.58 4.30 3.37,3.12 (6.63) (4.26) (3.33,3.0) 94 LYS 120.22 8.92 4.20 2.11 1.738,1.68 1.91,1.86 3.18,3.22 (8.97) (4.19) (2.06) (1.72,1.7) (1.92,1.91) (3.14,3.17) 95 MET 118.66 8.33 3.53 2.02,1.86 0.84,-0.14

96 SER 114.91 8.06 3.76 3.96,3.86

97 ARG 116.78 7.26 4.63 2.19,1.94 1.73 3.56

98 GLY 106.78 7.95 3.99,4.27

99 GLU 118.97 7.85 4.58 2.36,1.95 2.48,2.35

100 ASN 118.03 8.46 2.88,2.70

101 PHE 118.03 7.63 5.21 3.12 7.31 7.63

102 ASP 121.16 9.48 4.72 2.56,2.97

103 GLU 127.41 8.67 4.16 2.35,2.25 2.63,2.45

104 GLY 108.66 8.62 4.04

Continued

112 Table 4.2 continued.

105 LYS 118.97 7.71 4.42 2.02,1.77 1.67,1.54 1.79,1.73 3.12

106 LEU 116.80 7.67 4.18 1.86,1.43 1.71 0.76

107 LYS

108 ASN

109 VAL 122.80 8.01 3.78 2.10 1.35(QG1) 0.85(QG2) 110 THR 107.72 7.92 4.09 4.81

111 LEU 125.53 8.19 4.32 1.86,1.52 2.00 1.09

112 MET 112.72 7.75 4.70 2.45

113 SER 114.59 8.00 3.97 4.22,4.05

114 ASP 116.16 7.60 4.36 2.19,2.03

115 ILE 113.03 7.34 4.51 2.23 1.73(HG12) 1.10 (QG2) 1.01(QD1) (7.28) (4.48) (2.27) 1.49(HG13) (1.09) (0.98) (1.76,1.42) 116 LYS 120.22 7.89 3.94 2.08 1.69,1.64 (7.93) 117 ASN 117.09 7.65 5.03 0.93 (7.6) 118 TYR 118.34 8.60 5.19 1.77,1.38

119 PRO 4.44 2.61,2.21 2.29,2.20 3.98,3.63

120 ALA 118.66 8.81 4.14 1.44

121 ARG 113.97 8.41 4.67 2.28 3.41 3.42

122 VAL 116.16 7.61 3.79 2.44 1.22 1.17

123 LYS 116.78 7.19 4.25 2.53,2.36

124 CYS 117.72 7.89 4.65 2.70,2.53

125 PHE 118.97 8.15 4.59 3.20,3.04

Continued

113 Table 4.2 continued.

126 ILE 117.72 8.36 4.37

127 LEU 122.09 8.94 4.20 1.60

128 ALA 121.16 8.25 4.44 1.86

129 TRP 113.03 8.33 4.11 1.77,1.27 (8.39) 130 LYS 119.28 8.32 4.35 2.02,2.19 2.07 1.89,1.87 3.09

131 THR

132 LEU 117.41 7.46 4.65 2.36,2.19 (7.4) 133 LYS 120.84 8.32 3.90 1.62,1.27 1.07,0.65 1.62 3.01

134 GLU 118.66 8.41 3.65 0.88,1.33 2.85

135 ALA 121.47 8.17 4.13 1.51 (8.25) (4.2) (1.46) 136 LEU 119.28 8.30 4.35

137 LYS 115.22 8.09 4.02 1.93 1.60 1.93 3.24

138 LYS 122.09 9.15 4.67 3.04,2.87 2.05 2.32 3.18

139 ILE 119.28 8.18 3.78 2.62

140 SER 114.59 7.56 4.67 4.26

141 ARG 122.09 8.13 2.36

142 PRO 4.40 2.34,2.07 2.15,2.07 3.88,3.79

114

Table 4.3. Carbon chemical shift assignments (ppm) of apo D40A Tm IscU.

115

CA CB C CG CD CE

1 MET

2 VAL

3 PHE 51.23 29.36 169.47

4 LYS 63.43 35.92 176.42

5 MET 57.18 27.16 176.87

6 MET 56.85 26.54

7 TYR 59.20 36.00

8 SER 56.24 61.89 171.12

9 GLU 51.55 28.11 175.44

10 ALA 52.95 12.79

11 ILE

12 LEU

13 ASP

14 TYR 62.07 36.20 176.52

15 ALA 53.10 17.17 175.97

16 ASN 50.92 37.49 173.02

17 SER 56.86 61.23 174.77

18 LYS

19 LYS 171.11

20 PHE 54.06 36.87 171.17

21 ARG 52.17 29.67 174.26 26.93 41.20

22 GLY 41.86 166.17

Continued

116 Table 4.3 continued.

23 LYS 51.86 33.80 21.71 27.31 39.88

24 LEU 50.83

25 ASP 53.73 38.60 173.38

26 ASP 49.99 36.85 172.75

27 ALA 50.61 17.19 174.53

28 THR 62.49 67.79 172.19 21.16

29 VAL 58.43 33.10 169.17 19.21(CG1) 19.31(CG2) 30 ILE 58.11 38.41 172.80 25.80(CG1) 12.28 14.91(CG2) 31 GLU 52.20 32.50 170.88

32 GLU 52.49 30.29 173.48 34.21

33 GLY 42.48 169.58

34 LYS 53.11 34.35 172.76 22.54 27.42 39.87

35 ASN 52.18 39.05

36 ILE 61.77 35.24 172.70

37 SER 59.68 60.30 173.58

38 CYS 55.00 40.30 171.52

39 GLY 44.36 172.01

40 ALA 48.43 21.85 172.64

41 GLU 53.42 31.23 171.48 33.93

42 ILE 57.20 40.60 171.22 25.31(CG1) 11.17 15.00(CG2) 43 THR 60.30 67.82 170.71 20.05

44 LEU 51.11

45 TYR

Continued

117 Table 4.3 continued.

46 LEU 51.39 41.40 173.26 29.69 24.76(CD1) 24.28(CD2) 47 LYS 53.12 32.17 171.51 23.21 26.84 40.11

48 VAL 58.42 32.16 172.50 19.08(CG1) 18.80(CG2) 49 GLU 53.09 31.55 34.14

50 ASP 53.12 37.63 173.41

51 GLY 43.73 171.16

52 VAL 58.42 33.10 173.23 19.49(CG1) 18.65(CG2) 53 VAL 60.62 27.48 171.67

54 LYS 54.36 31.23 173.84 22.74 26.22 39.73

55 ASP 50.93 41.55 170.71

56 ALA 49.67 21.39 171.95

57 LYS 52.33 38.11 172.11 25.07 27.05 40.57

58 PHE 53.43 41.24 171.11

59 GLU 52.18 31.86 173.59 34.88

60 GLY 43.60 169.47

61 MET 51.86 33.11 171.61 33.92

62 GLY 43.11 168.93

63 CYS 56.87 26.55 176.32

64 VAL 64.06 29.68 174.95

65 ILE 60.64 31.55 176.44

66 SER 59.36 60.78 175.20

67 GLN 55.62 26.86 175.55 33.97

68 ALA 53.13 17.48 176.60

Continued

118 Table 4.3 continued.

69 SER 57.78 61.23 176.75

70 ALA 53.12 13.73 175.49

71 SER 61.23 62.31

72 LEU 62.70 47.79 176.81 32.08 22.75

73 MET 54.05 30.92 175.30

74 LEU 55.61 176.05

75 GLU 56.25 27.80 175.44

76 ARG 54.05 26.56 175.69 22.94 30.27

77 ILE 60.94 36.85 173.47 24.29(CG1) 14.09 17.00(CG2) 78 ILE 63.43 34.98 174.67 28.31(CG1) 12.32 15.54(CG2) 79 GLY 42.48 172.02

80 GLU 52.16 28.40 173.55 33.64

81 ARG 54.06 28.73 175.92 26.27 41.12

82 VAL 65.60 29.04 174.33 22.26(CG1) 19.34(CG2) 83 GLU 58.12 26.56 174.14 34.24

84 GLU 55.30 27.48 176.84 33.21

85 ILE 63.43 34.37 21.25(CG1) 18.14 19.32(CG2) 86 PHE 177.64

87 SER 56.23 61.85 171.11

88 LEU 52.47 41.54 174.06 28.53 24.78(CD1) 24.47(CD2) 89 ILE 56.86 37.49 27.21(CG1) 10.67 16.97(CG2) 90 GLU 58.10 26.56 174.34

91 GLU 53.45 21.86 174.32

Continued

119 Table 4.3 continued.

92 ALA 54.05 14.38 174.17

93 GLU 56.87 26.55 175.73

94 LYS 56.88 30.29 176.79 22.96 27.10 40.10

95 MET 56.86 30.31 173.89 26.84

96 SER 59.06 60.61 169.94

97 ARG 53.86 29.35 174.53

98 GLY 44.04 172.25

99 GLU 52.48 29.05 173.15 34.76

100 ASN 52.49 35.93 171.50

101 PHE 52.49 39.68 170.70

102 ASP 50.61 38.43 173.58

103 GLU 57.15 27.18 176.47 34.33

104 GLY 44.36 173.70

105 LYS 55.92 30.92 23.22 27.17 39.46

106 LEU

107 LYS

108 ASN

109 VAL 63.34 29.26 174.42

110 THR 62.78 66.54 173.64 21.16

111 LEU 55.79 40.30 175.70

112 MET 52.78 29.98 173.45 18.38

113 SER 59.98 60.60 174.45

114 ASP 56.25 30.29 175.41

Continued

120 Table 4.3 continued.

115 ILE 59.06 35.93 174.59 24.89(CG1) 12.11 16.02(CG2) 116 LYS 55.77 30.62 174.77 22.13 27.19 40.02

117 ASN 54.54 38.63 172.39

118 TYR 51.86 43.88

119 PRO 63.59 29.90 172.61 25.17 48.34

120 ALA 52.17 16.54 175.65

121 ARG 52.80 29.01 177.45

122 VAL 60.90 29.36 169.93 18.60(CG1) 17.38(CG2) 123 LYS 54.98 29.04 175.48

124 CYS 51.54 35.91 171.27

125 PHE 55.31 38.42 177.53

126 ILE 61.87 35.92 172.82

127 LEU 58.11 34.72 175.94 26.66(CG1) 25.99(CG2) 128 ALA 52.49 16.86 176.79

129 TRP 54.36 176.80

130 LYS 57.18 30.29 171.50 23.72 27.47 39.81

131 THR

132 LEU 57.16 27.17 178.05

133 LYS 58.76 30.31 177.56 22.29 26.99 39.96

134 GLU 56.55 30.30 176.84 33.79

135 ALA 52.50 16.23 175.80

136 LEU 58.42 30.51 177.83

137 LYS 56.23 27.96 175.64 23.51 27.31 40.22

Continued

121 Table 4.3 continued.

138 LYS 53.42 34.98 173.81

139 ILE 63.11 34.67 176.59

140 SER 59.06 60.31 173.94

141 ARG 56.09 30.30

142 PRO 62.76 29.90 25.04 48.20

4.3.3 SECONDARY STRUCTURE

The chemical shift index (159) was used as an initial indication of the secondary structural content of apo D40A Tm IscU and was consistent with the φ and ψ dihedral angles (Figure 4.3). The φ dihedral angles were calculated from J-coupling constants

(Table 4.4), and the ψ dihedral angles were calculated from NOE peak intensity ratios of

1 15 dαN(i-1,i) and dNα(i,i) in a H- N-NOESY-HSQC. The secondary structural elements were further confirmed by NOE assignments. For instance, within helical regions a series of dNN(i, i+2) and dαN(i, i+2) in addition to side-chain NOEs were detected.

Additionally, one dαN(i, i+3) NOE was assigned in helix 1 (between F3 and M6). Long- range NOEs indicated the presence of an anti-parallel β-sheet composed of β-strands 1, 2, and 3. The portion of the β-sheet composed of β-strands 2 and 3 appears to be more tightly formed than the other regions of the β-sheet, as evidenced by the number of

NOEs. The formation of this β-sheet was confirmed by the direct detection of backbone

122 NH to CO hydrogen bonds between T43 and E59 and between K47 and D55. The RDC

data (Table 4.5) were also consistent with the determined secondary structure. Apo

D40A Tm IscU was aligned with the magnetic field by use of a non-ionic liquid crystalline solution (157). Generally, each residue within each secondary structural element was oriented in the same direction with respect to the magnetic field. Taken together, all of the data indicate that Tm IscU is composed of six α-helices and one three- stranded anti-parallel β-sheet.

123

Figure 4.3. Sequence alignment of IscU proteins showing the determined secondary structure of apo D40A Tm IscU. An “a” indicates data consist with an α-helix, whereas a

“b” represents data consistent with a β-sheet. CSI, average chemical shift index results

from Cα, Cβ, C’, and Hα. PHI and PSI angles were calculated as described in materials

and methods. The RDC data are summarized as (+) for positive residual dipolar coupling

values and (-) for negative residual dipolar coupling values. Residues labeled as (+)

1 15 within the D20 category indicate that their amide resonances were observed in a H- N

HSQC after 18 h in 93% D2O.

124

125 residue 3J(HN,Hα) (Hz) residue 3J(HN,Hα) (Hz) residue 3J(HN,Hα) (Hz) 3 9.69 49 8.39 93 7.93 4 3.9 50 8.89 94 6.51 5 4.56 51 4.69 95 3.79 6 5.1 52 7.98 96 4.51 8 1.85 53 8.64 97 7.07 9 6.41 55 6.36 98 9.45 10 4.46 56 5.02 99 8.83 11 3.4 57 7.3 101 8.12 16 8.31 58 7.67 102 9.94 17 2.99 59 8.84 103 3.43 19 5.22 61 6.16 104 7.09 21 4.52 62 3.64 105 7.35 22 8.37 63 4.13 106 5.96 23 8.66 64 1.0 109 3.13 24 8.17 67 3.98 110 5.53 26 7.82 69 3.1 111 3.74 27 4.07 70 1.11 112 8.69 28 6.77 71 3.69 113 2.9 29 7.03 74 1.66 114 4.7 30 9.83 76 5.26 115 8.14 31 8.88 77 6.77 118 8.79 32 8.02 78 3.5 120 4.63 33 7.98 79 3.27 121 6.96 34 8.13 80 7.45 124 7.7 35 6.29 81 7.48 126 3.28 38 4.89 82 2.24 129 4.24 39 6.98 83 2.82 130 2.03 40 8.26 84 9.22 132 4.06 41 6.71 85 3.0 135 2.9 42 7.21 87 5.67 136 3.56 43 9.36 88 6.91 137 2.88 44 8.85 89 2.18 138 4.29 47 9.39 91 2.43 139 3.1 48 10.37 92 3.52 140 5.57

Table 4.4. 3J(HN,Hα) coupling constants of apo D40A Tm IscU.

126

residue RDC (Hz) residue RDC (Hz) residue RDC (Hz) 3 -11.56 52 -11.99 99 -10.22 4 -10.07 53 -10.07 100 2.91 6 -0.78 54 -2.41 101 -9.79 8 78.32 55 5.18 102 -7.17 10 -8.94 56 2.84 103 -6.24 11 -4.61 57 1.56 104 -0.14 15 -16.25 58 0.14 105 1.84 16 -7.17 59 1.92 106 -6.38 17 -4.47 60 2.62 109 -11.07 19 -12.27 61 0.43 110 16.32 20 6.31 62 1.35 111 -14.54 21 4.75 63 0.28 112 -9.51 22 8.66 66 8.09 113 -4.82 23 4.19 67 -2.7 114 -7.59 25 11.14 68 1.7 115 -16.18 26 -1.7 69 0.43 116 -10.93 27 3.62 70 -1.77 117 -7.87 28 -7.17 71 -0.07 118 3.33 29 1.21 72 4.33 121 -6.1 30 1.35 76 12.2 122 2.84 31 -0.71 77 7.66 125 8.3 32 -1.06 78 -5.82 126 -3.48 33 2.13 79 -3.12 127 -10.36 34 0.71 81 -2.98 128 3.05 35 2.06 82 -1.84 129 -11.92 38 2.62 83 -2.55 130 -12.13 39 9.86 84 7.38 132 -2.41 40 4.47 85 -12.49 133 -1.28 41 -2.27 87 0.85 134 -3.97 42 1.99 88 -6.53 135 -15.32 43 1.14 89 -1.7 136 -2.84 44 9.15 91 -3.05 137 -11.14 47 -8.3 93 -9.36 138 -6.31 48 -3.26 94 -4.11 139 0.85 49 -10.36 96 -11.07 140 3.9 50 -5.96 97 0.78 141 -16.46 51 0.92 98 2.41

Table 4.5. 1H-15N Residual Dipolar Couplings of apo D40A Tm IscU.

127 4.3.4 NOE ANALYSIS

The 2D NOESY spectrum of apo D40Ala Tm IscU at 45 oC displayed only 847

NOE cross-peaks, which is significantly less than would be expected for a protein of Tm

IscU’s size (Figure 4.4). More specifically, the aromatic region of the spectrum was not

greatly populated, and there were few NOEs arising from side-chain aliphatic methyls

that typically exist in a protein’s hydrophobic core. Nevertheless, sequential i and i+1

NOEs were found throughout the protein, thus confirming the sequence specific resonance assignments. NOEs indicative of helical structure were observed within regions previously identified as helical from non-NOE data. However, these were mostly limited to within i to i+2 NOEs. The lack of i to i+3 NOEs, while retaining good agreement with the chemical shift index, suggests flexibility of the helices rather than an equilibrium between fully formed and partially folded states. The largest number of long-range NOEs was between the β-strands of the anti-parallel β-sheet (Figure 4.5).

Apo D40A Tm IscU had only 66 long-range NOEs, of which 30 were within this β-sheet.

The remaining regions of Tm IscU had very few long-range NOEs. The region with the least number of NOEs was between α4 and α5.

128

Figure 4.4. 2D NOESY of apo D40A Tm IscU in 450 mM NaCl, pH 5.4, 45 oC at 700

MHz.

129

Figure 4.5. Schematic representation of NOE connectivities. The upper left portion of the diagram shows only backbone NOEs, and the lower right region shows all of the H-H

NOEs for apo D40A Tm IscU. The intensity of the points is directly proportional to the overall intensity of the NOE.

130 4.3.5 HYDROGEN EXCHANGE

Concentrated apo Tm IscU in H2O was rapidly diluted in D2O to a final concentration of 93%. After 1 h 37% of the backbone resonances were observed in a 1H-

15N HSQC spectrum, whereas after 18 h only 25% of the peaks remained (Table 4.6). All of the backbone amides of apo D40A Tm IscU exchanged after one week in D2O.

Generally, amino acids that showed protection from hydrogen exchange resided in regions of secondary structure, whereas rapidly exchangeable residues were in regions lacking defined secondary structure. Cys38 and Cys63 were in regions protected from hydrogen exchange, whereas Cys124 was not. Thus, it appears that the hydrophobic pocket that anchors the Fe-S cluster contains Cys38 and Cys63

131

Table 4.6. Hydrogen Exchange. A "x" indicates that the residue was detected in a 1H-

15 N HSQC after addition of D2O to 93% after the indicated time.

132 residue 1 h 18 h residue 1 h 18 h residue 1 h 18 h 1 47 x x 93 x x 2 48 94 3 x 49 95 x x 4 x x 50 96 5 51 97 6 x x 52 x 98 7 53 x x 99 8 54 x 100 9 55 x x 101 x 10 x x 56 x 102 11 57 x x 103 12 58 x x 104 13 59 x x 105 14 60 106 15 61 107 16 62 108 17 63 x x 109 x x 18 64 110 19 x 65 111 20 x x 66 112 21 67 x 113 22 68 114 x x 23 69 x 115 24 70 x x 116 25 x 71 x x 117 26 72 118 x x 27 73 119 28 74 120 29 x x 75 x x 121 30 76 122 31 x x 77 x x 123 32 78 x x 124 33 79 x x 125 34 80 x x 126 35 81 x x 127 36 82 128 x 37 83 129 x x 38 x 84 130 39 85 x 131 40 x x 86 132 x 41 87 133 x x 42 x x 88 134 43 x 89 x 135 x x 44 x x 90 136 x x 45 91 137 46 92 x x 138

133 .4.3.6 Holo D40A Tm IscU SPECTRA

As a result of the instability of the Fe-S cluster of IscU proteins, NMR spectra

were acquired on D40A Tm IscU under solution conditions more amenable to holo

protein studies (50 mM sodium phosphate, pH 7.4, 200 mM NaCl, 20 oC).

Unfortunately, the solution conditions necessary for improved cluster stability yielded lower quality NMR spectra than under the optimized solution conditions. For example,

29 backbone amide cross-peaks assigned under optimized solution conditions for apo

D40A Tm IscU were missing in the 1H-15N HSQC spectrum acquired at 20 oC (Figure

4.6). Under these solution conditions the pattern of cross-peaks in the 1H-15N HSQC

spectra of apo and holo D40A Tm IscU were similar. However, 15 of the missing

backbone resonances of apo protein at 20 oC (that were seen for apo at 45 oC) reappeared

in the 1H-15N HSQC spectrum of holo D40A Tm IscU at 20 oC. This suggests that the

Fe-S cluster stabilized the protein fold, which is corroborated by relaxation measurements

(see 4.3.7). Unfortunately, such stabilizing influences were not readily apparent from 2D

NOESY spectral comparisons. The 2D NOESY of holo protein was not improved over that of apo under these solution conditions (Figure 4.7). Indeed, the observed NOESY cross-peaks were significantly broadened for holo D40A Tm IscU. Such line broadening

is likely due to the paramagnetic effects of the Fe-S cluster, particularly since both apo

and holo protein under these solution conditions exist in the same aggregation state (see

4.3.7). Thus, the highest quality NOESY spectra were acquired for apo D40A Tm IscU at

45 oC.

134

Figure 4.6. 1H-15N HSQC of holo D40A Tm IscU in 200 mM NaCl, pH 7.4 at 20 oC and

500 MHz (A) and an expansion of the spectrum (B).

135

136

Figure 4.7. 2D NOESY of apo (A) and holo (B) D40A Tm IscU in 200 mM NaCl, pH

7.0, 20 oC at 700 MHz.

137

138 4.3.7 APO AND HOLO D40A Tm IscU DYNAMICS

15 The N R1 and R2 (Figure 4.8) for apo D40A Tm IscU in 450 mM NaCl, pH 4.5

at 45 oC were consistent with a stable protein fold and were near the values expected for

dimeric Tm IscU (174) (Figure 4.9) on the basis of the Stokes-Einstein equation at this

-1 temperature for a globular protein . The mean R1 and R2 values were 2.32 ± 0.24 s and

-1 9.27 ± 1.01 s , respectively. Only Leu88 and Arg121 had R2 values indicative of an

exchange process. The average heteronuclear 1H-15N NOE was 0.73 ± 0.13 s-1 at 500

MHz. Although this average value is indicative of a stable protein fold at 45 oC, the

variability between residues is greater than that typically observed for a rigid protein,

suggesting that some conformational flexibility must exist. However, no negative

heteronuclear NOEs were observed, as they often are for highly dynamic proteins (175,

176).

139

Figure 4.8. R1, R2 and heteronuclear NOEs for apo D40A Tm IscU in 450 mM NaCl, pH

5.4, 45 oC at 500 MHz.

140

Figure 4.9. Theoretical R1 and R2 values on the basis of the Stokes-Einstein equation at

20 °C and 45 °C for monomeric (dashed curves) and dimeric (solid curves) apo D40A Tm

IscU. Experimental points for R1 (circles) and R2 (squares) at both temperatures are

shown in the plot.

141 Relaxation rates were additionally measured for apo D40A Tm IscU in 50 mM

o 15 sodium phosphate, pH 7.4, 200 mM NaCl, at 20 C. The average N R1 and R2 values were 1.48 ± 0.18 and 15.17 ± 3.23, respectively. These values were again near that expected for dimeric Tm IscU at this temperature (Figure 4.10A and 4.10C). Transverse

o relaxation rates revealed a dynamic protein at 20 C, showing highly variable R2 values throughout the protein sequence. G33, K54, E59, G98, T110, W129, and R141 appeared to be disordered, while several other residues were experiencing an exchange process

(M6, I89, F101, Y118, K137, and K138, and I139) as evidenced by large R2 values.

Holo D40A Tm IscU relaxation measurements were recorded under the same

o 15 conditions as the 20 C apo data set (Figure 4.10B and 4.10D) and yielded mean N R1 and R2 values of 1.71 ± 0.17 and 19.03 ± 1.80, respectively, again indicating an essentially dimeric structure. Although these mean rates are similar to those determined for apo protein, their distribution throughout the protein sequence was significantly

different in that the uniformity of the relaxation rates on a per residue basis was much

greater for holo than for apo. That is, the number of residues undergoing an exchange

process or exhibiting a high degree of disorder was significantly less for holo D40A Tm

IscU than for apo protein at 20 oC. Again, it is important to note that even though holo

protein appeared more rigid than apo D40A Tm IscU at 20 oC, both samples were more

dynamic than apo D40A Tm IscU at 45 oC. Although such a thermoprofile is intriguing,

it is plausible that the higher temperature allows Tm IscU to achieve a thermodynamic

mixture of states among low-lying conformers in such a way that the fraction occupying

the lower level is higher. At lower temperature the number of kinetically accessible

142 conformational states may be significantly lower and may not include those more favored

thermodynamically. Fe-S coordination would then favor one of these conformational

states over the others, possibly stabilizing the same fold as that selected for at higher

temperatures. Although, the data indicates that apo D40A Tm IscU at 45 oC has a more static three-dimensional fold than holo at 20 oC, it is clear that the Fe-S cluster

contributes to some extent to the stabilization of the tertiary fold. At the very least, the

15 cluster may stabilize a pre-existing tertiary fold. A comparison of N R1 and R2 for apo and holo D40A Tm IscU under the same conditions is shown in Figure 4.10.

143

Figure 4.10. R1 and R2 for apo (panels A and C) and holo (panels B and D) D40A Tm

IscU in 200 mM NaCl, pH 7.4 at 20 °C and 500 MHz.

144

145 4.3.8 STRUCTURAL FEATURES OF IscU

The secondary structural content of apo D40A Tm IscU is consistent with

previous circular dichroism measurements on apo WT and apo and holo D40A Tm IscU

(73). The identified secondary structural elements are six α-helices and three β-strands, the latter forming an anti-parallel β-sheet. After a first short helix encompassing residues

4 to 10, there are three β-strands (28-35, 40-49, 53-62), followed by helices 2 (63-77), 3

(82-85), 3’ (91-96), 4 (110-115) and 5 (129-139). Helix 3’ resides in a ~ 18 amino acid insertion region of the protein which is only present in IscUs of lowly evolved organisms

(73). Although there is sufficient evidence supporting the formation of all of these helices, they clearly are not rigid, as evidenced by the relative lack of i to i+3 NOEs.

This suggests that these helices experience some distortions while participating in slow dynamic processes.

Most of the 66 long-range NOEs that could be detected involve the secondary structural elements. In particular, 30 of them occur within the β-sheet. This number is sufficient to attempt a structural model of the latter. Therefore, a structure calculation

(using Dyana) was performed on a stretch of D40A Tm IscU spanning residues 24 to 65, i.e. a region of Tm IscU encompassing the β-sheet. The results are shown in Figure 4.11

(panels A and B). Although the RMSD within the family of structures is high, the structural features are indeed typical of an antiparallel β-sheet. Particularly, strands 2 and

3 are relatively well ordered and show an interesting distribution of hydrophobic amino acids on one side of the β-sheet (even-numbered residues) and hydrophilic residues on the other (odd-numbered residues). Another important piece of structural information

146 extracted from these calculations is the placement of two of the three cysteine residues

(Cys38 and Cys63), which are implicated in binding the Fe-S cluster in holo Tm IscU

(74, 80). Cys38 is located in a loop connecting strands β1 and β2, while Cys63 immediately follows the third strand of the β-sheet. This structural arrangement results in

Cys38 and Cys63 being spatially proximal (sulfur atoms as close as 6 to 7 Å in some structures), which is consistent with their binding to a [2Fe-2S] cluster in the holo state.

It is interesting to note that an amino acid position (Asp40) identified by site-directed mutagenesis to be important for cluster stability (58, 73, 74), occurs within a structured region of Tm IscU at the beginning of strand β2. Furthermore, the derivative D40A is found to be hydrogen-bonded to Lys34 at the end of strand β1, thus stabilizing the β-

sheet. Such stabilization at the ends of the β-strands likely stabilizes the loop region

connecting them, which for Tm IscU contains two of the Fe-S cluster ligands (Cys38 and

Cys63). Possibly this region within D40A Tm IscU is less dynamic than that of WT

protein, thus stabilizing Fe-S cluster coordination. However, a more likely reason for the

stability of the D40A derivative stems from the change in solvent accessibility to the

cluster binding pocket. Residue 40 is positioned on the hydrophilic side of the β-sheet

that holds cluster binding Cys residues. We have previously reported studies of cluster

stability and cluster transfer kinetics that address the importance of solvent access to the

cluster binding pocket in mediating both of these properties (75, 82). Localization of

residue 40 to the close proximity of the cluster provides a rational explanation for the

aforementioned influence of the D40A substitution on the stability and kinetic data.

147 The β-sheet is followed by one long helix (α2) and four short helices (α3, α3’,

α4, and α5). Among the remaining long-range NOEs, some connect helices 2 and 3 with helix 3’ while a few others connect helices 2 and 5. Again, Dyana calculations involving helices 2, 3, and 3’ and 2 and 5, respectively, were performed. In the latter case, the long stretch of residues between helices 2 and 5 was substituted by a flexible linker made of pseudoatoms in such a way as to decrease the degrees of freedom and to avoid steric hindrance from meaningless orientations of the intervening residues. In both calculations the helical structures were imposed through backbone dihedral angles. The results are shown in Figure 4.11 (panels C, D, E, F). Despite the scarcity of long-range NOEs, some super-secondary structural features appear among helices 2, 3, 3’, and 5, such as the reciprocal orientation of helices 2 and 5. Helix 5 contains a Lys rich region that could potentially function as a protein-protein interaction domain. Such a motif has previously been observed in partner protein interactions (adrenodoxin reductase and CytP450scc) with ferredoxin (94-96). In fact, these Lys lie on the same face of helix 5 and are oriented near Cys63 suggesting that this putative protein interaction domain places the partner protein near the Fe-S cluster binding site. Furthermore, there are a few NOEs connecting helix 5 with the residues immediately following the β-sheet and containing

Cys63, thus supporting the idea that the Lys rich domain is near the Fe-S cluster binding site. The remaining Cys (Cys124) is located between α4 and α5. Although it is tempting to look for close contacts between these two regions, Dyana calculations do not show any recognizable structural arrangements. One surprising feature of the three cysteine residues in Tm IscU (Cys38, Cys63, and Cys124) is the chemical shifts of their Cβ

148 resonances, which suggests an oxidized, or cystine, state for Cys38 (40.3 ppm) and possibly also for Cys124 (35.9 ppm), and to a reduced state for Cys63 (26.5 ppm).

Although all three Cys are believed to be involved in Fe-S cluster coordination (74, 80),

there were no detectable NOEs between either Cys38 or Cys63 with Cys124. This is a

likely result of Cys124 being in an unstructured region of Tm IscU.

149

Figure 4.11. Orientation of Tm IscU secondary structural elements. Panels A, C, and E show families of structures from Dyana calculations. Panels B, D, and F show only one molecule of the family. A and B) Structure of the anti-parallel β-sheet with cysteines 38 and 63 labeled. C and D) Structure of helices 2, 3, and 3’. E and F) Structure of helices

2 and 5 with Cys63, Lys130, Lys133, and Lys137 labeled.

150 4.3.9 FLUXIONALITY OF APO Tm IscU

Even under the best experimental conditions (pH 5.4, 45°C, 0.45 M NaCl, < 2

mM protein) apo Tm IscU is not a well behaved protein by NMR standards. As many as

17 out of the 143 residues could not be detected fully, nine of them lacking even the

backbone signals and another eight lacking the side chains completely. On statistical

grounds, this is expected to reduce up to 15-20% the number of observable NOEs.

However, the total number of observed NOEs (847) is by far less than expected for a

protein of this size (2-3 x 103). Furthermore, only 66 of the 847 NOEs are long-range.

This observation alone clearly explains why no 3D structures of apo Tm IscU can be

deduced from the present data.

At the same time, a number of experimental facts point to a state of the protein in

solution which is far from what would be classified as “unfolded”. First, there is clear

evidence for the presence of secondary structural elements, six α-helices and three β- strands, involving 77 residues and thus accounting for approximately 54% of the whole structure. This evidence relies on φ and ψ dihedral angle information, chemical shift index, or both. In addition, another 20 residues have dihedral angles and/or chemical shift index values inconsistent with a random coil structure, increasing the total to 97 residues (approximately 68% of the protein) that are not in a random coil structure.

Second, the proton spectrum of apo Tm IscU has well dispersed peaks, including regions indicative of tertiary structure, such as the aromatic region and upfield of 0 ppm.

Proton resonances in the 0 to –0.3 ppm region arising from ring current shifted aliphatic

151 side-chains are not usually detected in unfolded or molten globule proteins (177, 178).

Moreover, the 1H-15N-HSQC NMR spectrum shows good chemical shift dispersion.

Third, R1, R2 and NOE data on backbone amide nitrogens at two temperatures

point to an essentially normal behavior for the large majority of residues in terms of order

parameter. This observation rules out large sub-nanosecond mobility as the primary

culprit for the lack of NOEs. Indeed, the number of short-range NOEs, that are expected

to be the most sensitive to mobility on this time scale, is not far from the number

expected for a protein of this size, while the number of long-range NOEs is dramatically

small. And lastly, nitrogen R1 and R2 values point to a protein of double molecular weight, suggesting a dimeric quaternary structure.

All of these observations are consistent with extensive long-range order in this protein. However, this has to be reconciled with the lack of long-range NOEs. Mobility on longer than nanosecond time scales may account for strong reduction of NOEs.

Extensive conformational changes may bring nuclei that are far in primary sequence into close contact for a small fraction of time, and far apart for a larger fraction of time.

Under these circumstances the NOE between these two nuclei is scaled down by the

fraction of time that the two nuclei are in close contact. Each of the two nuclei may be in

contact with other nuclei for the remaining time. As a result, a nucleus may actually give

rise to a number of small NOEs rather than to a strong NOE. However, as first suggested

by Jardetzky (179), this situation would increase, rather than decrease the number of

long-range NOEs. Subsequently, these could be used for tertiary structure calculations

even if the resulting structure represents a meaningless average of different 3D structures.

152 This situation is expected if conformational equilibria occur over time scales that are longer than the rotational time of the molecule (a few nanoseconds) but shorter than the reciprocal frequency separation among the signals from the same nucleus in different environments (of the order of milliseconds) (180). At time scales much longer than this reciprocal frequency separation, the spectrum would consist of a superposition of the spectra from each different conformer, each with its own pattern of long-range NOEs, reduced in intensity but still visible.

Apparently, none of these two extreme situations can fully explain the Tm IscU

NMR data. However, it is possible that multiple conformational exchange processes occur on a time scale that is of the order of the reciprocal frequency separation. In this case, dramatic exchange broadening may occur (180-186), which can easily render the

NOE unobservable, especially when coalescence is approached. The resulting picture would be that of a protein that has reasonably well defined secondary structural elements, but with a tertiary structure that is fluxional among widely different conformational arrangements. The occurrence of dimers would be easily accommodated into this picture, and could involve quaternary contacts that may be transiently formed at the expenses of some intramolecular tertiary contacts. The time scale of these processes, at the fields employed in the present work (500-800 MHz) would roughly cover two orders of magnitude, from about 0.1 to 10 ms. It is also interesting to note that seven of the eight residues for which the side chain could not be identified are hydrophobic residues, suggesting that the fluxional behavior stems from a failure to form a stable hydrophobic core.

153 Strong support for such an interpretation comes from the comparison of NMR data taken at 20 °C and 45 °C. At 45 °C, 17 residues escaped detection, while 23 showed double resonances. At 25 °C as many as 61 residues were missing, while of the remaining ones 31 showed double resonances. This behavior suggests that many different conformational equilibria exist with different time scales. By lowering the temperature some of these equilibria pass from fast exchange to coalescence (and several more signals disappear) while others pass from coalescence to slow exchange (and some signals reappear as doubled). Unfortunately, further increases of temperature above

45 °C did not seem to cause reappearance of more signals.

The NMR data indicative of fluxionality is also consistent with what is usually attributed to a molten globule-like state. Molten globules are generally defined as proteins which possess native-like secondary structure but little to no static tertiary structure (177, 187-189). Indeed, the main NMR characteristic of molten globules is a strong reduction of long-range NOEs (176, 177, 190, 191). Although many of these conformational states were initially characterized in non-native protein folds under partially denaturing conditions, there are now several examples of dynamic protein folds whose physiological function requires a high degree of conformational flexibility (192-

195). The extent of the flexibility attributed to molten globule proteins is wide ranging.

If apo Tm IscU is regarded as a molten globule, it shows evidence for a degree of folding

that is not always observed for molten globules. For instance, apo Tm IscU’s chemical

shifts are more dispersed than for many studied molten globules (177, 196), and the

dynamic data of apo Tm IscU under optimal solution conditions is quite normal for a well

154 folded protein. Indeed, unlike the molten globule state of apo myoglobin (175), apo

D40A Tm IscU has no negative heteronuclear NOEs. Further support for a tertiary fold stems from the hydrogen exchange rates of apo D40A Tm IscU, which even at 45 oC are

significantly slower than for both unstructured (197) and several molten globule proteins

(198). Nevertheless, the dramatic lack of long-range NOEs for Tm IscU involves the

majority of residues and is not restricted to those with undetectable side-chains. Lack of

long-range NOEs for side-chains that are clearly observed and not broadened must have a

different origin, such as large amplitude fast fluctuations. This feature makes IscU a

peculiar molten globule in that it is more compact but yet more fluid than typical molten

globules.

4.3.10 COMPARISON OF THE STRUCTURE AND DYNAMICS OF IscU WITH

OTHER METALLOCHAPERONES

It is evident from the determined secondary structure and from their spatial

organization (Figure 4.11) that IscU proteins represent a new protein fold involved in

metalloprotein maturation. All previously structurally characterized metallochaperone

proteins have revealed a common ferredoxin-like fold (βαββαβ), of which IscU is now

determined not to be a member. IscU flexibility appears to be an inherent property

necessary for physiological function. A rigid protein fold presumably cannot accomplish

Fe-S cluster synthesis and the recognition of various differing target proteins.

Furthermore, it is reasonable to expect conformational changes of IscU upon target

protein interactions that are necessary for Fe-S cluster release from IscU and subsequent

155 coordination to the target protein. Thus, the observation of multiple conformational states by NMR is consistent with the roles that IscU must fulfill. Interaction with a particular target protein may stabilize a specific IscU conformational state.

Analogous conformational flexibility does not exist in other structurally characterized metallochaperones. Generally, for these systems any observed flexibility is localized within the metal binding loop (34, 35, 135). However, it is important to note that these metallochaperones possess a ferredoxin-like fold and function to transfer only a single metal ion. Furthermore, these relatively rigid chaperones transfer metal to other conformationally restricted proteins that have similar folds (42, 139), that is a chaperone with a ferredoxin-like fold transfers metal to another protein with a ferredoxin-like fold.

Or the metal-transfer domain of the metallochaperone is tethered to another domain that is structurally homologous to the target protein (41, 42, 199). IscU differs greatly from this situation and thus would be expected to possess distinct structural and dynamic characteristics. For instance, IscU directly transfers [2Fe-2S] units (82) to proteins that are structurally distinct, and can transfer cluster to proteins that lack tertiary structure in the apo state (73). In contrast, copper chaperones and their targets show similar structural folds (34, 35, 135).

Studies of metallochaperones have revealed several common themes for metallochaperone/transport systems including: little to no change in secondary or tertiary structure upon conversion from apo to holo states (34, 35, 45, 47, 73, 134-138, 200), metal dependent recognition of partner proteins (i.e. the holo form of the metallochaperone has an increased affinity for target proteins over its corresponding apo

156 state) (30, 75, 76, 201, 202), a solvent exposed metal binding site (82, 139), and lastly, the metallochaperone structures solved thus far share the presence of a ferredoxin-like

βαββαβ motif (139, 140). IscU shares all of these characteristics except for the ferredoxin-like fold. It is intriguing that an entirely new metallochaperone fold with vastly different dynamic properties still shares many of these characteristics and suggests that these shared properties are necessary for metal transfer, whereas the differences reflect the specific function of the particular metallochaperone.

157

CHAPTER 5

CHARACTERIZATION AND RELEVANCE OF THE MOLTEN GLOBULE

STATE OF THERMOTOGA MARITIMA IscU

5.1. INTRODUCTION

Elucidation of the structure of IscU would provide critical insight to our

understanding of IscU chemistry; however, the structural characterization of IscU has proved to be challenging. IscU proteins appear to lack a rigid tertiary fold and so structural characterization has thus far been exclusively focused on a homologue from the hyperthermophile Thermotoga maritima. Initial structural characterization, by circular dichroism, correctly predicted a high degree of secondary structure. Other spectroscopic and functional evidence seemed consistent with a high degree of secondary structure, and the presence of a stable tertiary fold. The strongest evidence for such was the high degree of chemical shift dispersion observed by 1H-15N HSQC NMR experiments.

However, subsequent efforts to further define the three dimensional structure of

158 T. maritima IscU (Tm IscU) by NMR failed to identify a unique tertiary fold (203).

Surprisingly, the vast majority of the NMR data was consistent with a stable and rigid protein fold. Nevertheless, the number of detectable long-range NOEs was significantly below that which is normally observed for a protein of Tm IscU’s size, and provided too few distance constraints for satisfactory structural calculations of the tertiary structure.

Much of the NMR data was, however, similar to that previously characterized for molten globule states of proteins. The term molten globule is generally used to describe a state in which a protein possesses a high degree of secondary structure without a stable tertiary fold (204-206). Molten globules often show large degrees of chemical shift dispersion but lack long-range NOEs (176, 177). The main difference between Tm IscU’s NMR results and that of previously characterized molten globule proteins is the dynamic data.

Molten globules usually show motion on a ns-ps time scale detectable by 15N relaxation

measurements (175). Tm IscU did not appear to have significant ns-ps motion. In

addition to the dynamic data several lines of evidence, such as resonance splitting and

hydrogen exchange, suggested that flexibility existed on a ms time scale. Thus, it is

presently unclear what tertiary state Tm IscU is in, and whether that state reflects what is

typically described as a molten globule. Therefore, we sought to characterize apo Tm

IscU by methods typically employed to identify molten globules. These methods include near-UV CD, ANS binding, free energy of unfolding, hydrodynamic radius measurements, and limited tryptic digestion. Apo and holo protein behave similarly by

NMR criteria. However, due to complications arising from the instability of the bound

Fe-S cluster of holo Tm IscU, the majority of the structural data was acquired on apo Tm

159 IscU. The data suggest unusual dynamic behavior that is not fully consistent with typical

protein states, such as fully folded, fully unfolded, or molten globule. For instance, the

existence of a stable tertiary fold is supported both by near-UV CD spectra and

measurements of the hydrodynamic radius, while other data are less clearly interpretable

and may be viewed as consistent with either a molten globule or a fully folded state.

However, all of the data is consistent with our previous hypothesis of a protein sampling

multiple discrete tertiary conformations in which these structural transitions occur on a

‘slow’ time scale. The significance of such dynamics and how they may be affected by

partner protein interactions are discussed.

5.2 MATERIALS AND METHODS

5.2.1 GENERAL CHEMICALS

ANS was obtained from molecular probes (Eugene, OR), the low molecular

weight protein ladder was from BRL, and all other chemicals were from Sigma-Aldrich

(St. Louis, MO). Protein samples were expressed and purified as previously described

(73). All protein concentrations were calculated based on monomer equivalents.

5.2.2 ANS BINDING

ANS concentration was determined by absorbance using the previously determined extinction coefficient at 350 nm (207). Excitation was at 371 nm with a slit width of 3 nm, and emission was monitored at 482 nm with a slit width of 10 nm at

160 25 oC. The buffer was 100 mM Tris-HCl, pH 7.4. NaCl concentrations were varied between 0 and 200 mM. Data were corrected for the inner-filter effect when necessary

(208).

5.2.3 NEAR-UV CD

Circular dichroism spectra were measured on an AVIV model 202 circular dichroism spectrometer using a 1 cm path-length cuvette. Protein concentrations were

0.08 mM in 100 mM Tris-HCl, pH 7.4 at 25 oC. Spectra were recorded in 0, 50, and 200

mM NaCl. Buffer spectra were subtracted. Spectra are averages of three measurements except for those in Gdn-HCl.

5.2.4 DYNAMIC LIGHT SCATTERING

Light scattering was recorded on a DynaPro-801 (Protein Solutions,

Charlottesville, VA) with a temperature-controlled microsampler. The laser wavelength

and scattering angle was 8294 Å and 90o, respectively. The instrument software

(Dynamics, version 3.27) was used to calculate the hydrodynamic radius using the

measured translational diffusion coefficient and the Stokes-Einstein equation: DT = (kT)/f

= (kT)/(6πηR), which can be rewritten as R = (kT)/(6πηDT). DT is the translational

diffusion coefficient, k is the Boltzmann constant, T is the absolute temperature, f is the

frictional coefficient, η is the viscosity of the solution, and R is the hydrodynamic radius.

Solution conditions were 50 µM protein in 100 mM Tris-HCl, pH 7.4 at 22.7 oC with

161

varying NaCl concentrations of 0, 50, 100, 200, 300, and 450 mM NaCl. Samples were filtered through 0.02 µm filters (Anatop-10 from Whatman) immediately prior to measurements.

5.2.5 FREE ENERGY OF UNFOLDING

Gdn-HCl induced unfolding was monitored by CD. Measurements were on an

AVIV model 202 circular dichroism spectrometer using a 0.1 mm path-length cuvette, repeated three times, and averaged. Protein (0.08 mM) was incubated in 100 mM Tris-

HCl, pH 7.4 and 0, 0.5, 1, 1.5, 2, 2.5, 3.0, 3.2, 3.5, 3.6, 3.7, 3.8, 3.9, 4.0, 4.1, 4.2, 4.3, 4.4,

4.5, 4.7, 5.0, 5.5, 6.0, 7.0, and 7.3 M Gdn-HCl overnight at room temperature. The Gdn-

HCl concentrations were prepared by mixing stocks of 100 mM Tris-HCl, 200 mM NaCl, pH 7.4 and an identical solution with 8.0 M Gdn-HCl in which the pH was adjusted after dissolving all of the components. Unfolding was followed by monitoring the ellipticity at

222 nm as a function of Gdn-HCl concentration and analyzed by the linear extrapolation method (209), which assumes a two state unfolding mechanism.

5.2.6 LIMITED TRYPTIC DIGESTION

Reactions were carried out at the Campus Chemical Instrument Center at Ohio

State University. WT and D40A Tm IscU were reduced with DTT (5 mg/ml in 100 mM ammonium bicarbonate), alkylated with iodoacetamide (15 mg/ml in 100 mM ammonium bicarbonate), and then mixed with 1:25 (w/w) Promega sequence grade

162 modified trypsin and incubated at 37 oC. After 5, 15, 30, 60, and 120 min aliquots were removed and immediately desalted by use of ZipTips (Millipore, Bedford, MA)

according to the manufacturer’s directions, and analyzed by electro-spray ionization

(ESI) using a Micromass Q-TOF(tm) II (Micromass, Wythenshawe, UK) mass

spectrometer equipped with an orthogonal electrospray source (Z-spray) operated in

positive ion mode. Sodium iodide was used for mass calibration for a calibration range

of m/z 100 - 2500. Proteins were prepared in a solution containing 50% acetonitrile/50%

water, 0.1% formic acid at a concentration of 50 pmol/µl and infused into the

electrospray source at a rate of 5 - 10 µL min-1. Optimal ESI conditions were: capillary

voltage 3000 V, source temperature 110o C and a cone voltage of 60 V. The ESI gas was

nitrogen. Q1 was set to optimally pass ions from m/z 100 - 2000 and all ions transmitted

into the pusher region of the TOF analyzer were scanned over m/z 100 - 3000 with a 1 s

integration time. Data was acquired in continuum mode until acceptable averaged data

was obtained (10 -15 min). ESI data was deconvoluted using MaxEnt I, provided by

Micromass. The experiment was repeated identically except that at each time point an

aliquot was removed and mixed with SDS-PAGE loading buffer, boiled for 3 min, and

subsequently loaded onto a homogenous-20 precast polyacrylamide gel from Pharmacia

(Uppsala, Sweden).

5.2.7 CLONING OF T. MARITIMA dnaK

T. maritima genomic DNA was obtained from the American type culture

collection (ATCC # 43589D). Chromosomal DNA (10 ng, 50 ng, and 100 ng), 0.3 µM of

163 each primer, 2.5 units of cloned Pfu DNA polymerase, 1X cloned Pfu buffer, and 0.2 mM

each dNTP were used to amplify dnaK (TIGR locus: TM0373) via PCR. The total

reaction volume was 50 µL. The primers were 5'-

GGGCCCGGCATATGGCAGAAAAGAAAGAATTCG-3’ and

5’-CCGGCCGGATCCTTACTGATTTGATGTTTCTCC-3’. The underlined regions

denote Nde I and BamH I sites, respectively. The thermocycle was similar to that used

for T. maritima iscU cloning (see 3.2.2), except that the extension time was increased

since dnaK (1788 nucleotides) is much larger than iscU (426 nucleotides). The thermocycle was one step of 95 oC for 45 s followed by 25 cycles of 95 oC for 45 s, 55 oC

for 45 s, and 72 oC for 3 min and 40 seconds, and one final cycle of 72 oC for 10 min.

Pfu DNA polymerase was added to the reaction mixture during the first annealing step

(55 oC) of the thermocycle. The entire PCR product was purified with the Qiagen PCR kit and subsequently lyophilized with a speed-vac (Savant). Vectors (pET21 and pET28) were purified from 10 mL cultures of DH5α carrying the appropriate plasmid using a

Qiagen mini-prep kit and lyophilized. To these lyophilized DNA samples (both insert and vector) the following was added: 10 mM DTT, 0.1 mg/ml BSA, 15 units Nde I, 15 units BamH I, 1X React 3 buffer (BRL), and sufficient dH2O to bring the total volume to

45 µL. Digests were incubated at 37 oC for 16 h. Then, one unit of shrimp alkaline phosphatase (USB) was added to the vector digests only and incubated for an additional

2 h at 37 oC. All samples were purified with the Qiagen gel extraction kit without

actually performing an agarose gel extraction. The samples were then lyophilized again.

Insert was resuspended in 10 µL of sterile dH2O, and vector fragments were resuspended

164 in 20 µL. The ligation reactions utilized two different insert/vector ratios. The first

condition used 1 µL of vector and 6 µL of insert. The second condition utilized 2 µL of

vector and 3 µL insert. For both conditions 1X ligase buffer and 1 unit T4 ligase were

used. Sterile dH2O was added to bring the total reaction volume to 10 µL. Subsequently,

the samples were incubated at 16 oC for 16 h. All of the ligation mixture was used to transform CaCl2 competent DH5α (78). Transformants were initially screened by Hind

III digestion of purified plasmid (Qiagen mini prep kit). Hind III digestion reactions

contained ~1 µg plasmid, 5 units Hind III, and 1X React 2 buffer (total reaction volume =

~10 µL) at 37 oC for 1 h. Then, 2 mL of DNA loading buffer was added, and 10 µL of

the resulting mixture was loaded onto a 1% (w/v) agarose gel with TBE as the running buffer. Positive samples were further confirmed by nucleotide sequencing at the Ohio

State University Plant-Microbe Genomics Facility. Cloning into pET21 yielded a

construct without a tag or additional residues (pTmDnaK). Cloning into pET28 resulted

in the addition of an amino-terminal His-tag (pTmDnaKHis).

5.3 RESULTS

5.3.1 ANS BINDING

ANS is a structural probe whose fluorescence is greatly dependent on its polar

environment. For example, ANS fluorescence is negligible in aqueous solution, whereas

in an apolar environment its fluorescence is greatly increased and blue shifted. ANS addition to native or fully unfolded protein solutions usually does not result in significant

165 fluorescence. However, in the presence of partially folded or molten globule proteins

ANS exhibits a large degree of fluorescence, presumably by penetrating the protein’s hydrophobic core (210). Under the current solution conditions and in the absence of protein, ANS showed weak fluorescence with a maximum emission at 516 nm. Upon Tm

IscU addition (either WT or D40A) the fluorescence increased approximately 16 fold and blue shifted to 478 nm (Figure 5.1). Increasing sodium chloride concentrations resulted in decreased ANS fluorescence in the presence of protein. For example, in 200 mM

NaCl the ANS fluorescence decreased 2 fold. As a control, apo human ferredoxin was examined for ANS binding. Human ferredoxin is regarded as a typical low molecular weight [2Fe-2S] protein that is not considered a molten globule in the apo or holo state.

Under identical solution conditions used for the Tm IscU experiments, ANS fluorescence increased less than 2 fold in the presence of apo human ferredoxin, and slightly increased with increasing NaCl concentrations. As is commonly observed (211, 212), Scatchard plots of ANS – Tm IscU titrations were not linear indicating multiple ANS binding sites with the affinity of the first binding site being approximately 1 µM. Interestingly, no significant ANS fluorescence was observed in the presence of holo D40A Tm IscU. This lack of fluorescence was completely unaffected by the ionic strength of the solution. As a control, human ferredoxin was again used. Holo Hs Fd also lacked ANS fluorescence and was unaffected by NaCl concentrations. Whether the Fe-S cluster stabilizes the protein fold of Tm IscU, or inhibits binding by occupying the ANS binding site, or simply quenches the fluorescence of bound ANS is not known.

166

Figure 5.1. ANS fluorescence. A) ANS fluorescence without protein. B) Apo WT Tm

IscU ANS fluorescence. C) Apo D40A Tm IscU ANS fluorescence. D) Holo D40A Tm

IscU ANS fluorescence. E) Apo Hs Fd ANS fluorescence. F) Holo Hs Fd ANS fluorescence. The concentrations of all the samples were identical (50 µM protein and 10

µM ANS). The inset is an expansion of the y-axis. Thin black line, protein without

ANS; thick black line, ANS; green line, ANS + 50 mM NaCl; blue line, ANS + 200 mM

NaCl.

167

168 5.3.2 NEAR-UV CD

Near-UV CD spectra are sensitive to the protein’s tertiary structure surrounding

its aromatic residues. Only fully folded proteins show significant signals within this

region, whereas fully unfolded and molten globules do not (213). Both WT and D40A

Tm IscU showed near-UV CD spectra indicative of a stable tertiary fold with minima at

297 and 299 nm and a maximum near 292 nm. These signals were lost in the presence of

high concentrations of Gdn-HCl (Figure 5.2). Sodium chloride did not significantly

affect the near-UV CD spectra.

5.3.3 DYNAMIC LIGHT SCATTERING

-7 2 The measured translational diffusion coefficient (DT) was 8.83 x 10 cm /s and

8.54 x 10 –7 cm2/s for apo WT and apo D40A Tm IscU, respectively. Using the Stokes-

Einstein equation the hydrodynamic radius was calculated to be 2.5 ± 0.1 for WT Tm

IscU and 2.6 ± 0.1 nm for D40A Tm IscU, which is normal for a dimeric protein of Tm

IscU’s size. Sodium chloride did not have a significant effect on the measured DT and consequently did not influence the hydrodynamic radius. The translational diffusion coefficient and the hydrodynamic radius are related to the frictional coefficient of the molecule, which is indicative of the overall shape of the protein in solution. In particular, the ratio of the measured frictional coefficient (f) over that calculated for the smallest rigid sphere capable of accommodating the protein (fsphere) is typically between 1.2 and

1.3 for globular proteins. Values larger than this are observed for non-globular

169

Figure 5.2. Near-UV CD of WT (solid line) and D40A Tm IscU (dashed line) in 100 mM

Tris-HCl, 200 mM NaCl, pH 7.4 with and without 7.4 M Gdn-HCl (dotted lines).

proteins with elongated tertiary folds (214, 215). The f/fsphere for both WT and D40A apo

Tm IscU was calculated to be 1.2. DLS data on holo D40A Tm IscU were not

monodispersive, thus precluding reasonable fits of the data. However, the result it

consistent with an increased ability of holo protein to self associate, as has been observed

with other metallochaperones (34).

170 5.3.4 FREE ENERGY OF UNFOLDING

The unfolding of Tm IscU was achieved by incubation with increasing concentrations of Gdn-HCl and monitored by CD (Figure 5.3). Assuming a two state mechanism for unfolding, the calculated free energy of unfolding (∆GD(H2O)) was 3.9 kcal/mol and 5.8 kcal/mol for WT and D40A Tm IscU, respectively. The dependence of the free energy on Gdn-HCl concentration is often represented by the m value, and reflects the amount of polypeptide that is exposed to solvent upon unfolding (216, 217).

The m values for WT and D40A Tm IscU were 0.9 and 1.4, respectively. A comparison of unfolding data between rigid protein folds and molten globules from both mesophilic and thermophilic organisms is reported in Table 5.1.

171

Figure 5.3. Gdn-HCl denaturation of WT (open circles) and D40A Tm IscU (closed circles). Solution conditions were 100 mM Tris-HCl, 200 mM NaCl, pH 7.4 with varying concentrations of Gdn-HCl.

172

Table 5.1. Comparison of denaturation data for molten globule and non-molten globule proteins. * The m value is in kcal/mol/M. Entries in bold are those of molten globule

proteins. For proteins that follow a three state denaturation profile N, I, D represent

natured (fully folded), intermediate (molten globule), and denatured states, respectively.

The last 11 entries are thermophilic proteins, whereas the remaining are mesophilic. Ox

and red indicate oxidized and reduced, respectively.

173 ∆G(H2O) protein organism T (oC) pH denaturant (kcal/mol) m* reference ferredoxin bovine 25 8.5 urea 5.02 (218) ferredoxin Anabaena 7120 25 7.5 Gdn-HCl 6.30 2.3 (219) adenylate kinase S. cerevisiae 15.2 7.5 Gdn-HCl 5.48 7.3 (220) maltose binding protein E. coli 25 7.5 Gdn-HCl 9.5 10 (221) ribonuclease A bovine 25 7.0 Gdn-HCl 8.74 2.94 (222) flavodoxin Anabaena 7120 25 6.0 urea 4.77 (223) apo flavodoxin Anabaena 7120 25 7.0 urea 4.09 (224) myoglobin horse 25 7.6 urea 15.02 (225) apo myoglobin horse 8.0 Gdn-HCl 3.16 (226) P1A-P2B complex S. cerevisiae 7.0 urea 2.37 0.62 (227) serum albumin human 25 7.0 urea 3.49 (228) serum albumin (N-I) bovine 6.0 Gdn-HCl 3.0 (229) serum albumin (I-U) bovine 6.0 Gdn-HCl 3.0 1.2 (229) α lactalbumin human 25 7.3 Gdn-SCN 6.8 (230) α-lactalbumin (N-D) bovine 25 8.0 Gdn-HCl 2.3 (231) apo α-lactalbumin (N-I) bovine 25 8.7 urea 0.8 (232) α-lactalbumin (I-D) bovine 25 8.0 Gdn-HCl 1.6 (231) IPMDH (N-I) E. coli 27 7.0 urea 14.7 5.6 (233) IPMDH (I-U) E. coli 27 7.0 urea 17.7 2.2 (233) IPMDH (N-I) T. thermophilus 27 7.0 urea 4.51 1.2 (233) IPMDH (I-U) T. thermophilus 27 7.0 urea 16.3 1.4 (233) cytochrome c-ox T. thermophilus 20 7.0 Gdn-HCl 17.9 3 (234) cytochrome c-ox T. thermophilus 20 7.0 Gdn-HCl 6.0 1.7 (234) cytochrome c-red T. thermophilus 20 7.0 Gdn-HCl 25.1 3.8 (234) cytochrome c-red T. thermophilus 20 7.0 Gdn-HCl 8.4 1.5 (234) Dihydrofolate Reductase T. maritima 25 7.8 Gdn-HCl 30.12 7.15 (235) Pyrrolidone carboxyl peptidase P. furiosus 40 7.0 Gdn-HCl 1.82 4.67 (236) AFEST A. fulgidus 20 7.5 Gdn-HCl 8.6 2.1 (237) EST2 A. acidocaldarius 20 7.5 Gdn-HCl 6.4 2.1 (237) [4Fe-4S] ferredoxin-red A. ambivalens RT 10.0 Gdn-HCl 16.7 (238)

174 5.3.5 TRYPTIC DIGEST

Tryptic digests under non-denaturing conditions of WT and D40A Tm IscU were monitored by mass spectrometry and SDS-PAGE. Both proteins were relatively resistant to protease digestion, with nearly full-length protein fragments persisting for over 19 h.

After 5 min, the predominant species present was full-length protein. However, after 15 min, a slightly truncated species appeared that predominated for the duration of the measurements. This fragment spanned Met5 to Pro142, i.e. full-length protein minus the first four residues. The limited proteolytic digestion of WT and D40A Tm IscU were not identical. For instance, after 30 min several additional large peptides between 6 and 14 kDa were observed for D40A Tm IscU, but not for WT. Additionally, less than 30% of

WT Tm IscU remained undigested after 19 h, whereas for D40A Tm IscU over 40% of the protein survived after 19 h of trypsin treatment. The trypsin sites that resulted in large fragments of D40A Tm IscU after 30 min of proteolytic digestion were Lys4,

Lys18, Arg21, Lys34, Lys47, Arg121, Lys123, Lys130, Lys137, and Lys138 (Figure

5.4).

175

Figure 5.4. Tryptic digestion of WT and D40A apo Tm IscU. ESI spectra of WT (A) and

D40A (B) apo Tm IscU after 30 min of tryptic digestion. Insets are SDS-PAGE of aliquots after 0 min, 5 min, 15 min, 30 min, 1 h, and 19 h tryptic digestion. C) Sequence of Tm IscU showing regions of secondary structure and solvent accessible trypsin sites indicated by inverted text. Note that the sequence shown is for WT protein, whereas the experimentally determined solvent exposed trypsin sites were determined for D40A Tm

IscU. D and E) Structural models of Tm IscU showing solvent-accessible trypsin sites in red and positions of Cys residues in black.

176

177 Both proteins were treated with Cys alkylating agents under non-denaturing

conditions prior to tryptic digestion to prevent by-product cystine formation.

Interestingly, full length protein with zero, one, and two alkylated Cys were observed, indicating that two of these Cys, at least to some extent, were solvent accessible. This is not surprising since previous biochemical evidence suggested a solvent accessible Fe-S cluster binding domain (82). Based on peptide frequency and peak intensity ratios, it

appeared that Cys124 was preferentially alkylated, consistent with our previous NMR

results indicating that Cys124 resides within an unstructured and solvent accessible

region of Tm IscU (203). However, since the alkylating agent was present in solution

throughout the course of the experiment it was not possible to unambiguously identify the

modified Cys residues. Additionally, significant quantities of Tm IscU without any

modified Cys were observed indicating that some degree of solvent protection exists for

the Fe-S cluster ligands.

5.3.6 T. MARITIMA dnaK CLONING

Cloning of dnaK (1788 nucleotides) was achieved by PCR and ligation into

pET21 and pET28 (~5.4 kb) (Figure 5.5). Colonies were initially screened by restriction

digestion of purified plasmids. The gene encoding dnaK has a Hind III site near its 3' end

(nucleotide position 1311), and pET vectors contain a Hind III site within their polyclonal

region that is retained after ligation using Nde I and BamH I sites. If no insert is ligated

into the vector, then the digestion product is full-length linearized plasmid. If dnaK

cloning is successful, then a small fragment of 502 nucleotides is produced after Hind III

178 digestion. Nucleotide sequencing from the 5' end of the gene confirmed that dnaK was cloned, and that there were no mutations within the first 807 nucleotides of pTmDnaKHis or the first 571 nucleotides of pTmDnaK.

Figure 5.5. Cloning of dnaK. All three 1.0% (w/v) agarose gels were run in TBE. Lanes

1, 5, 6, 11, 12, and 23 are 1 kb DNA ladders. Lanes 2, 3, and 4 are PCR products containing aliquots from reactions using 10 ng, 50 ng, and 100 ng chromosomal DNA, respectively. Lanes 7 - 10 are potential dnaK clones in pET21. Lanes 8 - 10 are positive.

Lanes 13 - 22 are potential dnaK clones in pET28. Lanes 13 - 15 and 17 - 22 are positive.

179 5.3.7 DnaK EXPRESSION AND PRELIMNARY PURIFICATION

DnaK is a 596 amino acids protein with a theoretical MW and pI of 66051 Da and

5.27, respectively. BL21(DE3)CodonPlus-RIL was used for protein expression and yielded high quantities of soluble protein. DnaK contains nucleotide and peptide binding domains (239). Peptide release is driven by ATP hydrolysis. Thus, the ADP bound state has higher affinity for target protein binding than the ATP bound form. Additionally, the binding of peptide is mostly through hydrophobic interactions. Although small scale trial expressions (Figure 5.6A) yielded nearly pure protein after incubation at high temperature, larger scale preparations did not produce such high purity protein.

Therefore, we tried purifying DnaK/His under a variety of conditions to determine which conditions result in the highest purity DnaK/His (Figure 5.6A). Glycerol addition to

DnaKHis protein solution was used in order to interfere with hydrophobic mediated binding of contaminant proteins to DnaKHis. Similarly separate additions of ATP, ADP, and NaCl were attempted. However, no significant differences were observed. The His- tag construct of DnaK did not bind to the Ni-NTA column with high affinity, resulting in

DnaKHis elution between 20 mM and 60 mM imidazole. Similar chromatographic problems were encountered with DnaK. DnaK eluted from a FPLC gel filtration column in several different fractions corresponding to different aggregation states. Such behavior has been previously observed for T. maritima DnaK, and may reflect different states of the protein, such as nucleotide and peptide bound and unbound forms (104). Similarly,

DnaK eluted from an anion exchange column at several different salt concentrations

(Figure 5.6B). DnaK/His did not bind cation exchange resins.

180

Figure 5.6. Expression and purification of DnaK/His. A) SDS-PAGE of crude lysate supernatants containing DnaKHis from 10 mL cultures after 5 h IPTG induction and heating for 0.5 h under differing conditions. Lane 1, MW marker; lane 2, 75 oC; lane 3;

80 oC; lane 4, 85 oC; lane 5, 85 oC + 2 mM ATP; lane 6, 85 oC + 2 mM ADP; lane 7,

85 oC + 200 mM NaCl; lane 8; 85 oC + 5% glycerol. Lane 9 is from a different gel and is

the insoluble fraction after sonication and centrifugation. B) SDS-PAGE of DnaK anion

exchange (DE-52) fractions. The protein was in 50 mM sodium phosphate, pH 7.4 prior

to loading and the column was equilibrated with 50 mM Tris-HCl, pH 7.4. All washes

were with Tris-HCl, pH 7.4 with varying concentrations of NaCl. Lane 1 is a MW

marker and lane 2 contains the flow-thru fraction. Lanes 3 - 6 contain NaCl washes with

50 mM, 100 mM, 200 mM, and 500 mM NaCl, respectively.

181 5.4 DISCUSSION

Molten globule proteins have been shown to posses a variety of biochemical

characteristics that are now often used to identify new members of this protein fold

family. These characteristics include: ANS binding (210, 240), far-UV CD reflective of a

significant degree of secondary structure without near-UV CD signals (198), an expanded

hydrodynamic radius that is nevertheless compact in comparison with fully unfolded

proteins (241), decreased stability as reflected by ∆GD(H2O) measurements (232), and

increased susceptibility to protease digestion (242, 243). However, comparison of these

data between different molten globules is complicated by the fact that what is labeled in

the literature as a molten globule is wide ranging, covering the full range of possibilities between that of fully folded and fully unfolded, i.e. those that are more similar to a rigid protein fold and those that lie nearer to an unfolded state. Additional complications are

related to the method used to induce the molten globule state, since most characterized

molten globules do not exist under physiological conditions.

The data described here for Tm IscU are not fully consistent with a rigid protein fold, a molten globule state, or a completely unstructured fold. Indeed, none of the data suggests an unstructured conformation, and so Tm IscU’s structure must either be a unique molten globule or posses a more typical fold with atypical dynamics. The present data that most clearly indicates a molten globule state for Tm IscU is ANS binding.

Much of our previous NMR data was acquired under high salt conditions (between 200

and 450 mM NaCl). Thus, much of these current experiments were done under similar

conditions so as to allow for facile comparison with previous NMR results. Since ANS

182 binds to hydrophobic regions of proteins (244) it seemed plausible that high salt would promote hydrophobic interactions resulting in forced ANS binding. However, ANS binding to Tm IscU was found to decrease with increasing salt, thus eliminating forced binding as a possibility. Increased structural rigidity of the protein under high salt conditions is an unlikely cause for the decreased ANS fluorescence since the hydrodynamic radius and near-UV CD of Tm IscU are unaffected by NaCl. A more reasonable explanation is interference with electrostatic interactions between the negatively charged sulfonyl group of ANS (ionized between pH 1.5 – 12) (245) and Tm

IscU. Although ANS is routinely used to probe for hydrophobic patches, it is known that both the non-polar anilinonaphthalene and the charged sulfonyl group can significantly contribute to protein binding (246).

Much of the remaining data is indicative of a rigid protein fold. In particular, Tm

IscU exhibits near-UV CD signals that are not observed for unfolded proteins. These bands reflect the asymmetric environment around the protein’s aromatic residues, and suggest the formation of a stable hydrophobic core. Additionally, the hydrodynamic radius of Tm IscU is normal for a globular protein fold. The measured translational diffusion coefficient can be used to calculate the protein’s frictional coefficient, which for

Tm IscU is also typical for a globular fold. Even if this method of frictional coefficient analysis introduces some degree of error, it is clear that the hydrodynamic radius of

Tm IscU is far from that of an unfolded protein. Using the method of Tanford (247, 248), the hydrodynamic radius of fully unfolded Tm IscU is expected to be approximately 26 nm for a monomer and 40 nm for a dimer.

183 The free energy of denaturation and the tryptic digest data are less clearly

interpretable. Both techniques are heavily protein dependent, thus making comparisons

between different proteins difficult. For example, values of the free energy of unfolding

vary widely among a variety of rigid proteins relative to protein states described as molten globule. Therefore, even though molten globules show decreased stability in comparison with their associated native state, such differences are only apparent for induced molten globules. For proteins that only exist in a molten globule state under physiological conditions there is no associated rigid state for comparison. Nevertheless, it is apparent that the ∆GD(H2O) for Tm IscU is normal. Indeed, nearly 6 M Gdn-HCl is

required to fully unfold both WT and D40A Tm IscU. The results do tend to suggest that

D40A Tm IscU is more stable than WT Tm IscU with ∆∆GD(H2O) of 1.9 kcal/mol.

Interestingly, this value is similar to the difference in activation energies for Fe-S cluster

transfer from D40A and WT Tm IscU to apo ferredoxin, with ∆G* of 15.0 kcal/mol and

13.7 kcal/mol, respectively, and ∆∆G* of 1.3 kcal/mol (82). Wu et al. have previously

reported that the stability of the cluster and the cluster transfer rate constant is strongly influenced by the solvation state of the cluster pocket (82), and so it is not unreasonable to expect that the difference in activation energies and cluster stability would dominate

any difference in the denaturation free energy of WT versus D40A derivative protein.

Alternatively, although less likely, the ∆∆GD may reflect distinct unfolding pathways.

However, the relative magnitude of the m values (0.9 and 1.4 for WT and D40A Tm

IscU, respectively) provide further support of the relevance of solvation in defining the

relative stabilities of these proteins. The magnitude of m reflects the amount of

184 polypeptide that is exposed to solvent upon unfolding (216, 217), and so the difference in

solvent accessibility for folded and unfolded states should be greater in the case of the

D40A derivative, as is indeed observed. Accordingly, while we are unable to compare

the denaturation of an induced “molten globule” state versus its native form, we can nevertheless compare the parameters for WT and the mutant forms. These data are completely consistent with the solvation model that we have previously outlined for the

cluster binding pocket and do lend some measure of credence to the data and the

interpretation that we present here. Certainly the stabilization of a protein fold by

substitution of a hydrophilic with a hydrophobic residue is not without precedent.

Additionally, the packing of the hydrophobic core of these proteins may not be identical since their near-UV CD are not completely superimposable. It is, however, apparent that neither protein is easily unfolded by denaturant, and therefore exhibits a large degree of stability.

Analysis of limited proteolytic digestion data suffers from the same complications of that of the free energy of unfolding, that is both techniques are more easily understood

when comparing native and molten globule states of the same protein. Nevertheless, it is evident that Tm IscU is significantly more compact than an unfolded protein with nearly

full-length protein being the predominant species present for over 2 h and persisting for

over 19 h. This ability of Tm IscU to withstand large amounts of trypsin for long periods

of time indicates that many of Tm IscU’s trypsin sites are buried for a significant period

of time, and thus are not solvent or protease accessible during these periods. Similar

behavior for a dynamic protein experiencing motion on a µs-ms time scale has been

185 observed (249). Of the twenty two possible trypsin sites (not including Arg141, which is followed by a Pro) nine were cleaved within 30 min, suggesting that the regions surrounding these sites are the most solvent accessible or least structured segments of

Tm IscU.

Comparison of the identified trypsin cleavage sites with previous structural data

(203) is revealing. Between α1 and β2 four of the possible seven trypsin sites were cleaved (Figure 4), and from the carboxy-terminus of α4 to the end of the molecule five of the possible seven trypsin sites were cleaved. In fact, the least structured part of Tm

IscU, as determined by NMR data, lies between α4 and α5 encompassing Cys124. Of the two possible trypsin sites within this region both were cleaved, further confirming that this region of the protein is unstructured. It is also interesting to note that a putative protein-protein interaction domain within α5 of IscU contains several Lys residues believed to mediate binding with target proteins. Of the four Lys within α5 of Tm IscU, three were cleaved by trypsin consistent with these Lys being solvent accessible and thus available for protein-protein interactions. By limited trypsin digest criteria the most stable part of Tm IscU appears to lie in the region from β3 to α4, with none of the seven possible trypsin sites cleaved. However, it should be noted that the large peptides produced by limited tryptic digestion at these sites were only observed for D40A Tm IscU and not for WT. Also, these large peptides were subsequently rapidly degraded resulting in large populations of nearly full-length protein and fully digested peptide fragments without significant populations of intermediate species. Such behavior suggests that Tm

IscU does not consist of independently folded units, and thus requires the whole molecule

186 to fold properly. Therefore, the large fragments resulting from initial cleavage are not structurally stable and thus are rapidly degraded. It appears that D40A Tm IscU is slightly more stable than WT.

While the various studies that we have performed to investigate the solution structure of Tm IscU appear to indicate contradictory conformational states, it can be reconciled by a model in which Tm IscU alternates between different conformations on a millisecond time-scale, as previously proposed (203). Those techniques that are not influenced by conformational flexibility, such as near-UV CD and dynamic light scattering, show characteristics indicative of a well-folded protein. Such a result is expected if each conformational state that the molecule experiences is well defined relative to the time scales of the measurements. However, those techniques that are influenced by dynamic processes may result in data that resembles less folded structures.

For instance, ANS may only penetrate the hydrophobic core of the protein during transitions between distinct conformational states. Similarly, a protein with multiple conformations in equilibrium may possess buried trypsin cleavage sites in each of these conformations that are nonetheless exposed to solvent during conformational transitions.

To our knowledge Tm IscU is the first molten globule-like protein exhibiting slow motion dynamics. Typically molten globule proteins display motion on a ns-ps time scale. Since IscU possesses slower motion, the methods commonly used to identify molten globules yield unusual and surprising results. In particular, ANS binding to Tm

IscU is consistent with a molten globule state, whereas the results from other non-NMR experiments support a non-molten globule state. Future characterization of other

187 members of this distinct family of molten globule-like proteins should therefore focus on comparing the results of ANS binding experiments with data from near-UV CD and dynamic light scattering measurements. If trends are found that resemble the results

described herein, then the possibility of slow motion conformational exchange should be

further confirmed by NMR criteria. Therefore, Tm IscU represents a new class of protein

that exist in two or more discrete conformational states that are interchangeable on a

slow, millisecond, timescale.

It is presumed that structural flexibility is necessary for IscU’s in vivo function.

This function requires holo IscU to dock with an apo protein target and then transfer an

intact [2Fe-2S] cluster via a sequence of bond cleavage and bond formation events where

the ligand set from IscU is replaced with that from the target protein. Such a complex

reaction presumably demands flexibility on the part of IscU as it prepares the cluster for

transfer and then proceeds through a series of intermediate and transition states before cluster transfer can be achieved. Such a reaction presumably requires an “opening up” of

the cluster binding domain prior to transfer. Given the fact that IscU can transfer cluster

to a number of distinct protein targets, the need for plasticity in protein structure is

readily understood. Other than recognition of diverse apo targets for cluster transfer,

IscU needs to be capable of interacting with a variety of proteins with different folds to

execute its function. Such proteins include sulfur donors, iron donors, holo ferredoxins,

and chaperones. It is, therefore, plausible that upon interaction with one of these partner

proteins a specific IscU conformation may be selected. One identified partner protein is

the chaperone Hsc66 (120), although the role of this partner protein is not clear, and IscU

188 appears to be able to carry out its function in vitro without the presence of Hsc66 (73,

82). Nonetheless, an Hsc66 binding motif within Escherichia coli IscU has been identified and consists of a hydrophobic-Pro-Pro-hydrophobic-hydrophilic sequence that is usually followed by a basic residue. In IscU proteins from more highly evolved organisms, including E. coli, the sequence is strictly LPPVK (250). However, more lowly evolved organisms, such as Gram-positive bacteria, often lack this exact sequence.

For example, in Tm IscU the analogous region spans residues 117-121 and is NYPAR.

There are two interesting aspects of organisms with this altered motif. One, these IscU proteins contain an ~18 amino acid insertion with unknown function (73) that is structured and contains an α-helix (203). Second, the genomes of organisms with such

IscU proteins do not appear to encode Hsc66. However, they do code for a homologous chaperone, DnaK, which has also been found to bind to this same region of E. coli IscU in addition to other parts of the protein (250). Furthermore, this chaperone-binding motif resides in a region of Tm IscU with the least number of NOE contacts, exhibits rapid hydrogen exchange properties, contains solvent exposed trypsin sites, and for which chemical shift analysis and dihedral angle calculations were unable to identify any secondary structural element (203). That is, this particular region of Tm IscU seems to be the least structured part of the protein. Whether Hsc66 or DnaK binding to this region will induce local structural changes or rigidify the entire protein fold possibly by selecting for one of the conformational states has yet to be determined.

189

CHAPTER 6

STRUCTURAL AND STABILITY EFFECTS OF SERINE OXYGEN LIGATION

OF A HIGH POTENTIAL [4Fe-4S] PROTEIN

6.1. INTRODUCTION

High potential iron proteins (HiPIPs) are small [4Fe-4S]3+/2+ containing proteins believed to be involved in electron transfer reactions in photosynthetic and non- photosynthetic bacteria (251). For example, in Chromatium vinosum, a purple sulfur

photosynthetic bacterium, HiPIP is thought to reduce the photo-oxidized hemes of the

reaction center under particular physiological conditions (252, 253). Generally, in purple

photosynthetic bacteria, the cyclic electron flow between the reaction center (RC) and the

cytochrome bc1 complex is in part mediated by quinones that carry electrons from the RC

to the cytochrome bc1 complex. The RC, cytochrome bc1 complex, and the quinones are

all embedded in the membrane. The return of electrons to the RC from the cytochrome

bc1 complex is mediated by soluble periplasmic proteins (254). Initially it was thought

that this soluble protein was exclusively cytochrome c2 (254). However, it was later

190 discovered that cytochrome c8 (255-257), HiPIP (253, 256, 258-262), and the membrane associated cytochrome cy (254, 263, 264) can all perform this function. Subsequent studies suggested that the particular protein used to shuttle electrons back to the RC is regulated by the physiological conditions of the cell. For instance, ratios of HiPIP, cytochrome c8, and the cytochrome bc1 complex are not static with respect to the RC, and do correlate with changes in growth conditions (251). When C. vinosum is grown autotrophically, in the presence of Na2S and Na2S2O3, HiPIP is preferentially used as the

direct electron donor to the RC. However, growth on organic compounds, without Na2S

and Na2S2O3, results in C. vinosum alternatively using soluble c-type cytochromes instead of HiPIP (252).

HiPIPs display a unique range of positive reduction potentials of +50 to +450 mV

(265), whereas those from low potential [4Fe-4S]2+/1+ proteins are in the range of -100 to

-650 mV (266). Interestingly, the [4Fe-4S]1+ state of HiPIP can be achieved under

extreme conditions (dithionite reduction under anaerobic conditions in 70-80%

dimethylsulfoxide), and gives rise to EPR signals similar to that of reduced low potential

ferredoxins (267). Reduced HiPIP has an overall charge of 2+ with the iron atoms being

antiferromagnetically coupled resulting in a diamagnetic state, and are thus EPR silent.

Oxidized HiPIP has an overall 3+ charge with a net spin of 1/2, and therefore is strongly

paramagnetic (268). The oxidized state of HiPIP formally has three ferric and one

ferrous ion, whereas the reduced state of the cluster is usually represented by two ferric

and two ferrous ions. However, in solution all the iron ions are in the 2.5+ oxidation

state when reduced (269, 270), while oxidized HiPIP has two ferric ions and a mixed-

191 valence pair, i.e. two Fe2.5+ ions (270-272). The ligands to the Fe-S cluster of

Ectothiorhodospira halophila HiPIP are Cys33, Cys36, Cys50, and Cys66. Through the

hyperfine shifts of α and β carbon and proton Cys resonances, Bertini et al. determined

that in the oxidized state Cys33 and Cys66 are bound to ferric ions and Cys36 and Cys50

are bonded to Fe2.5+ ions for the majority of the protein present in solution (~80%). The

remaining oxidized species contain Cys50 and Cys66 bonded to Fe2.5+ ions with the rest

of the Cys ligands bonded to ferric ions (273). A similar situation exists for oxidized C.

vinosum HiPIP, in which Cys43 and Cys46 are covalently attached to Fe3+ ions and the

mixed valence pair is ligated by Cys63 and Cys77 for 60% of the protein. The remaining oxidized C. vinosum protein utilizes Cys43 and Cys77 to ligate the ferric ions with the mixed valence pair coordinated by Cys46 and Cys63 (274).

Native C. vinosum HiPIP is a small (9258 Da) acidic protein with a pI of 3.88 and a redox potential of 0.35 V (275). The structure of HiPIP was first solved by Carter in

1974 (276). The X-ray crystal structure revealed a cubane [4Fe-4S] cluster ligated by

Cys43, Cys46, Cys63, and Cys77. Surprisingly, the protein did not show much regular secondary structure, but did posses some interesting structural characteristics, such as the surrounding of the Fe-S cluster with the hydrophobic side-chains of Tyr19, Phe48,

Phe66, Trp76, and Trp80 (Figure 6.1). This led many to attempt to alter the reduction potential of HiPIPs by the introduction of point mutations at sites defining the hydrophobic core. However, rather than perturbing the reduction potential significantly, these mutations had the effect of decreasing cluster stability, and so the aromatic side- chains surrounding the cluster appear to protect the [4Fe-4S] cluster from hydrolytic

192 attack rather than to modulate reduction potential (266, 277, 278). Such considerations have also led to investigations of cluster assembly and disassembly pathways (130, 279,

280), which are relevant in the context of iron sensing Fe-S proteins and mechanisms of cellular iron homeostasis (281). It now appears that the redox potential is tuned by the backbone amide dipoles around the Fe-S cluster (282).

Figure 6.1. C. vinosum HiPIP with the aromatic side-chains surrounding the [4Fe-4S] cluster shown in blue. The iron and sulfur atoms of the Fe-S cluster are represented by green and yellow spheres, respectively. The figure was generated by MOLMOL with the coordinates of 1HIP from the protein data bank.

193 The majority of known [4Fe-4S] proteins contain clusters coordinated by four

cysteines. Aconitase provided the first example of non-cysteinyl coordination, in which

one of the ligands is a solvent oxygen (283). Other examples include Ni-Fe hydrogenase

from Desulfovibrio gigas with a [4Fe-4S] cluster ligated by histidine (284), and

Pyrococcus furiosus ferredoxin with a [4Fe-4S] cluster ligated by an aspartate (8). As in

the case of aconitase, the oxygen ligated iron in P. furiosus ferredoxin can be lost

thereby generating a [3Fe-4S] cluster (235, 285).

Site-directed mutagenesis is now commonly applied to substitutions of cluster-

bound cysteines in an attempt to identify or alter the ligands to a Fe-S cluster. In some

instances it has been found that the isosteric serine can substitute for cysteine as an

unnatural ligand to the cluster. However, due to the low success rate, decreased stability

of the cluster, and complete lack of natural examples, it appears that cysteine is

significantly favored over serine. Indeed, for Azotobacter vinelandii ferredoxin I, protein

rearrangement resulting in remote cysteinyl ligation is preferred over coordination to a

serine substituted in place of a ligating cysteine (286). Although there are no examples of

natural serine coordination to canonical Fe-S clusters, serine has been identified as a

ligand to the molybdenum center of dimethylsulfoxide reductase (287), and a serine of

the oxidized nitrogenase molybdenum-iron protein P-cluster may serve an auxiliary role

by providing additional coordination to one of the iron atoms (288).

The X-ray crystal structures of an introduced Ser ligand to a Fe(Cys)4 rubredoxin center (289) and a [2Fe-2S] cluster have been solved (85, 99). However, no such examples exist for a [4Fe-4S] cluster. Relative to WT protein, it has been previously

194 shown that C77S HiPIP coordinates a less stable [4Fe-4S] cluster that is ligated by the Oγ of Ser77 and is accompanied by no gross structural perturbations (290, 291). Although the NMR solution structure of C77S HiPIP has been solved, a detailed comparison between WT and C77S HiPIP clusters and the details of cluster coordination was precluded as a result of the paramagnetism of the cluster. X-ray crystallography can provide additional and more accurate structural details of the cluster environment.

Therefore, to gain further insight into the structural and stability effects of Ser oxygen ligation of a [4Fe-4S] cluster we crystallized, solved the structure, and probed the stability of C. vinosum C77S HiPIP.

6.2 MATERIALS AND METHODS

6.2.1 PROTEIN PURIFICATION

Preparation and initial purification procedures for WT and C77S HiPIPs were essentially as previously described (113, 290). BL21(DE3) pLysS, transformed with either pHiPIP or C77S pHiPIP, was fermented on a 10 L scale at the Ohio State

Fermentation Facility in LB supplemented with 100 µg/ml ampicillin and 35 µg/ml chloramphenicol. At an optical density of 0.7 at 600 nm, the cultures were induced with

1 mM IPTG, and 5 h later pelleted and stored at -80 oC. Cell pellets were later thawed in

5 volumes of 10 mM Tris-HCl, 10 mM NaCl, pH 8.0 (H buffer) and sonicated at 90%

power. During sonication PMSF was added to 1 mM. Then, one volume of -20 oC

acetone was added drop-wise to the crude lysate, followed by centrifugation at 15000

195 rpm at 4 oC for 0.5 h. The cleared lysate was loaded onto an anion exchange column

(DE-52) equilibrated with H buffer and washed with greater than 5 column volumes of the same buffer. Bound protein was eluted with 10 mM Tris-HCl, 150 mM NaCl, pH

8.0. Further purification was achieved by gel filtration (G-75, Pharmacia) with H buffer as the running buffer. To insure that there was no apo contaminant of the C77S HiPIP sample, the protein was again loaded onto an anion exchange column (DEAE) at 4 oC equilibrated with H buffer. The column was washed with two column volumes of the same buffer and then four column volumes of H buffer + 50 mM NaCl. The protein was eluted with 10 mM Tris-HCl, 200 mM NaCl, pH 8.0. A G-25 column equilibrated with

10 mM Tris-HCl, pH 8.0 4 oC was used to desalt the C77S HiPIP sample. Purified WT

HiPIP had an A388/A282 absorbance ratio of 0.41, and C77S had an A388/A282 of 0.32.

HiPIPs were concentrated via ultrafiltration (Amicon) to 1.4 mM and 1.1 mM for WT and C77S HiPIP, respectively. The molecular weight of our WT HiPIP construct is 9447

Da. Protein concentrations were determined by the molar extinction coefficients at 282 nm and 388 nm of 41300 cm-1M-1 and 16100 cm-1M-1, respectively (275).

6.2.2 WT AND C77S HiPIP CRYSTALLIZATION

Hollow rectangular rod shaped crystals with typical dimensions of 0.05 x 0.05 x

0.5 mm grew within four days in the dark at room temperature via the hanging drop technique. C77S HiPIP (2 µl) was mixed with an equal volume of reservoir solution

(45.5% (v/v) saturated ammonium sulfate, and 49.8 mM KCl), and 2 µl of 50% 2-methyl-

2, 4-pentanediol (MPD) and allowed to equilibrate with 1 ml reservoir solution via vapor

196 diffusion. WT HiPIP crystallization was essentially the same as for C77S protein except

that 2 µl protein was mixed with 2 µl reservoir without MPD addition, and the reservoir solution was 49.8 mM KCl and 43.2% (v/v) saturated ammonium sulfate.

6.2.3 C77S HiPIP CRYSTAL DATA COLLECTION

Crystal data was collected at 100 K at beam line 14BM-C of the Advanced

Photon Source (APS) at Argonne National Labs. The C77S HiPIP crystals were transferred to the reservoir solution supplemented with 6% glycerol and immediately frozen inside a loop prior to data collection. Extensive attempts of freezing the crystal as

the whole hollow rod were unsuccessful due to its fragility. Therefore, one wall of the

hollow rod was cut into a 0.01 x 0.02 x 0.04 mm piece and subsequently used for data

collection. Crystal data were processed with DENZO and SCALEPACK (292).

6.2.4 C77S HiPIP STRUCTURE CALCULATION

The structure was solved by Y. Xiong and M. Sundaralingam using the molecular replacement method as provided by the program AmoRe (293). H42Q HiPIP (294) was

used as the search model. The model was refined to 1.9 Å using the program CNS (295).

The crystallized C77S HiPIP contains five additional amino acid residues at the amino-

terminus not found in native HiPIP or in the model (113). These additional residues were

fitted to the electron density according to its known sequence (113). A total of 188 water

molecules were added for the four independent molecules. The final R/R-free is

197 19.3%/23.4% in the resolution range of 15-1.9 Å. The crystal data, and structure

refinement statistic are shown in Table 6.1. The coordinates have been deposited in the

Protein Data Bank (PDB) with accession code 1JS2.

6.2.5 RESONANCE RAMAN SPECTROSCOPY

Protein concentrations were 3.7 mM WT HiPIP and 3.3 mM C77S HiPIP in 10 mM Tris-HCl, pH 8.0. Oxidized samples were prepared by the addition of a 10 fold

molar excess of K3Fe(CN)6, incubation on ice for 5 min., desalting via a short G-25

(Pharmacia) column, and then immediately frozen on dry ice until further use.

Resonance Raman spectra were obtained with a Chromex 500IS single-stage 0.5

meter imaging spectrograph equipped with an 1800 groove/mm grating. The detector

was a Princeton Instruments, Inc. LN/CCD-1024TKB liquid nitrogen cooled, 1024 X

1024 pixel array, back-thinned, charge-coupled device (CCD). All excitation lines were

generated by a Coherent INNOVA 307 Ar+ ion laser. Scattered photons were collected at

180o from a frozen sample (approximately 35 µl) placed in the sample well of a nickel

plated sample holder affixed to the coldfinger of an APD Cryogenics Inc. Displex DE-

204SL two stage, closed cycle helium refrigerator held at 30 K by a Lakeshore 330

Autotuning Temperature Controller. Rayleigh scattering was rejected by using a

holographic notch filter (Kaiser Optical Systems, Inc.) that was angle tuned for maximum

spectral coverage. The detector control software used was the CCD Spectrometric

Multichannel Analysis (CSMA) software version 2.2a provided by Princeton

Instruments, Inc. Spectra were calibrated using a software package based on the ASYST

198 scientific software package (Asyst Software Technologies) and further developed in the laboratory of Dr. Terry Gustafson at The Ohio State University. Indene, frozen in one of the three sample wells, was used for calibration. A quadratic fit to the indene Raman bands served as the calibration function for converting from pixel number on the CCD to wavenumber shift. Raman band center frequencies were determined using the PEAKFIT program (Jandel Scientific) running on a PC/DOS platform. The resonance Raman spectra were collected by C. Hemann and R. Hille.

6.2.6 CLUSTER STABILITY

A 100 mM buffered protein solution in a 1.0 cm path-length cuvette was repeatedly degassed under vacuum and argon purged. Buffers were pH 10.0 CAPS, pH

7.0 sodium phosphate, and sodium citrate at pH 4.0 and 3.0. Reactions were initiated at room temperature by the addition of argon-purged HiPIP to a concentration of 35 µM and monitored by the decrease of cluster absorbance at 386 nm on a Hewlett-Packard 8425A diode array spectrophotometer using the On-Line Instrument Systems (OLIS) 4300S

Operating System software.

6.2.7 NEAR UV-VISIBLE CIRCULAR DICHROISM

CD spectra were recorded with an Aviv model 202 circular dichroism spectrometer. Near-UV-visible spectra were acquired with 3 mm path-length cuvettes.

Absorbance was measured every 1 nm for 100 µM protein in 50 mM Tris-HCl, pH 7.4.

Only one scan was collected for each spectrum.

199 6.3 RESULTS

6.3.1. CYSTALLIZATION OF WT AND C77S HiPIP

The C. vinosum HiPIP construct used had three additional residues (Met-Glu-Phe)

at its amino-terminus (113). To confirm that the crystallization conditions were similar

for our construct as to that of native protein we performed trial crystallization screens

with WT before attempting C77S HiPIP. Two laboratories have previously crystallized

C. vinosum HiPIP (276, 294), and in both cases ammonium sulfate was the primary

precipitant. More specifically, Carter et al. used protein adjusted to pH 7.9 with NaOH

and 67% (v/v) saturated ammonium sulfate as the precipitant. The crystals formed were

of the P212121 space group (276). More recently, Parisini et al. solved the structure of

HiPIP to higher resolution (1.2 Å). In that case, C. vinosum HiPIP in 40 mM Tris-HCl,

180 mM KCl was crystallized with 2 M ammonium sulfate. These crystals were also indexed to the P212121 space group (294). We were able to crystallize our WT HiPIP

construct in a similar fashion by the hanging drop method with 49.8 mM KCl and 43.2%

(v/v) saturated ammonium sulfate at room temperature, in the dark, within one week.

(Figure 6.2A). However, high levels of Tris-HCl were found to interfere with crystal formation. Since there are already two high resolution crystal structures reported for

C. vinosum WT HiPIP, no X-ray diffraction data was collected on our WT construct.

Using the conditions determined for WT protein as a guide, conditions were screened for

C77S HiPIP crystallization. Interestingly, unlike WT HiPIP, improved crystal quality was achieved by the addition of MPD (Figure 6.2B). For C77S HiPIP, the reservoir

200 solution was 45.5% (v/v) saturated ammonium sulfate and 49.8 mM KCl. Additionally,

the droplet of protein solution that was equilibrated with the reservoir solution contained

16.7% (v/v) MPD. C77S HiPIP crystals diffracted to 1.9 Å resolution and were indexed

to the monoclinic space group P21 (Table 6.1). The cell volume indicated that there were

four independent protein molecules in the asymmetric unit (Figure 6.3).

Figure 6.2. WT (A) and C77S (B) HiPIP Crystals.

201 Space Group: P21 Cell parameters: a = 67.06Å, α = 90° b = 30.18Å, β = 111.27° c = 68.74Å, γ = 90° No. of protein molecules in a.u.:a 4 Wavelength: 1.653Å Resolution: 1.9Å Unique Reflections:b 35,617 Completeness (%): 90.4 Rsym (%): 9.4 Refinement (15Å-1.9Å) R-work/R-free: 23.4/19.3 Protein atoms: 2672 Water molecules: 188 RMSD bonds: 0.010 RMSD angles: 1.3

Table 6.1. Crystal data and refinement statistics of C77S HiPIP. a a.u.: asymmetric unit; b Friedel pairs unmerged.

202

Figure 6.3. The association of the four independent molecules in the asymmetric unit.

6.3.2 PROTEIN CONFORMATION AND [4Fe-4S] CLUSTER COORDINATION

In agreement with previous NMR results, the overall tertiary structure of C77S

HiPIP was essentially unaltered when compared with WT (291) with a rmsd of 0.89 Å

(Figure 6.4). Moreover, the [4Fe-4S] cluster itself was remarkably similar to the WT cluster (Table 6.2, Figure 6.5). Although the differences between the mean cluster interatomic distances for WT and C77S HiPIP are within experimental error, a trend of shorter Fe – inorganic sulfide (S*) bond lengths associated with the Ser ligated Fe (Fe4) is apparent. The [4Fe-4S] cluster protein ligands of C77S HiPIP were not greatly

203 perturbed. However, there was a noticeable lengthening of the preserved Cys – Fe bonds, and a shortening of the Ser77 – Fe bond (Table 6.2), indicating a slight migration of the

[4Fe-4S] cluster towards the introduced oxygen ligand. Three of the four C77S HiPIP molecules within the asymmetric unit had Ser77 – Fe4 distances between 2.10 and 2.14 Å with the fourth molecule having a distance of 2.27 Å. Nevertheless, the mean Ser77 –

Fe4 distance was 0.11 Å shorter than the Cys77 – Fe4 distance found for WT HiPIP. It is also worth noting that no attempts were made to maintain the reduced state of the crystal during data acquisition. However, the conditions used were identical to that for the

HiPIP structures reported by Parisini et al. (294) and are unlikely to result in oxidation at

100 K. Therefore, good comparisons between WT and C77S HiPIP structures can be made and are supported by resonance Raman results (see 6.3.6 and 6.4.1).

204

Figure 6.4. Superposition of the Cα backbones of C77S (yellow) and WT (white) HiPIP.

205

Table 6.2. Comparison between C77S and WT HiPIP mean Fe-S cluster bond

lengths and dihedral angles. a Taken from Parisini et al., (294). b Present work. c Positive and negative numbers indicate an increase and decrease, respectively, for C77S relative to WT HiPIP. Standard deviations are less than 0.01 for Fe-S* bond distances, less than

0.05 for Fe-protein ligand bond distances (except for Ser77-Fe4, 0.08), less than 4 for dihedral angles, and less than 0.11 for hydrogen bond distances (except for N81-Sγ46,

0.16).

206

WTa C77Sb Differencec Fe - S* distances, Å Fe1-S*2 2.31 2.29 -0.02 Fe2-S*1 2.30 2.30 0.00 Fe3-S*3 2.32 2.31 -0.01 Fe4-S*4 2.28 2.18 -0.10 Mean 2.30 2.27 -0.03

Fe1-S*1 2.33 2.32 -0.01 Fe2-S*2 2.31 2.31 0.00 Fe3-S*4 2.32 2.30 -0.02 Fe4-S*3 2.30 2.19 -0.11 Mean 2.32 2.28 -0.04

Fe1-S*3 2.24 2.26 0.02 Fe2-S*4 2.22 2.27 0.05 Fe3-S*2 2.25 2.18 -0.07 Fe4-S*1 2.28 2.20 -0.08 Mean 2.25 2.23 -0.02

Fe - protein ligand (Sγ/Oγ) bond distances, Å Cys43-Fe1 2.26 2.33 +0.07 Cys46-Fe2 2.30 2.36 +0.06 Cys63-Fe3 2.27 2.36 +0.09 Cys/Ser77-Fe4 2.27 2.16 -0.11

Fe-(Sγ/Oγ)-Cβ-Cα dihedral angles, deg Cys43 -67 -62 +5 Cys46 -173 -171 +3 Cys63 127 125 -2 Cys/Ser77 -79 -99 -20

Hydrogen bond distances, Å

N48-Sγ46 3.45 3.45 0.00

N81-Sγ46 3.67 3.51 -0.16

N65-Sγ63 3.30 3.26 -0.04

N79-(Sγ/Oγ)77 3.37 3.17 -0.20 N49-S*2 3.82 3.78 -0.04 N77-S*1 3.41 3.43 +0.02

207

Figure 6.5. C77S and WT HiPIP cluster and cluster ligands. The superposition of C77S

(colored) and WT (white) HiPIP was centered on the [4Fe-4S] cluster.

6.3.3 THE AROMATIC CORE SURROUNDING THE [4Fe-4S] CLUSTER

The residues defining the aromatic hydrophobic core are Tyr19, Phe48, Phe66,

Trp76, and Trp80. In particular, Tyr19 has been implicated in maintaining cluster stability due to its close proximity to the cluster (296). We have found only slight

208 differences in Tyr19 positioning with respect to the cluster (Figure 6.6). The Tyr19 Cδ1-

S*3 and Cδ1-(S/Oγ)77 distances for C77S HiPIP are 3.71 Å and 5.29 Å, respectively, and

are 3.69 Å and 5.41 Å, respectively, for WT HiPIP. Furthermore, Trp76 and Trp80

remain relatively unchanged. Banci et al. noted an altered Phe66 orientation between

crystal and solution structures of WT HiPIP (297). We also observe a slightly altered

Phe66 orientation (Phe66 tilts 29.6o in the C77S HiPIP structure). However, the most

dramatic difference observed in the C77S HiPIP structure is the rearrangement of Phe48, which is tilted 34.5o towards S*46 (Figure 6.6). The importance of Phe48 in excluding

solvent from the Fe-S cluster binding pocket of the protein can be visualized by a space

filling model as shown in Figure 6.7.

209

Figure 6.6. Comparison of C77S and WT HiPIP cluster, cluster ligands, and surrounding hydrophobic residues after superimposing their Cα backbones. A) Crystal structures of

C77S HiPIP (yellow), WT HiPIP (white, (294)), H42Q HiPIP (magenta, (294)). B) Same

as panel A except viewed from a different angle. C) Solution structures of WT HiPIP

(297). D) Solution structures of WT HiPIP showing all of the backbone in yellow, the

Fe-S cluster in red, and Phe48 in magenta.

210

Figure 6.7. Space filling model of WT C. vinosum HiPIP. The protein is shown in yellow, Phe48 in blue, and the Fe-S cluster in red. WT HiPIP coordinates were from

Carter et al. (276).

6.3.4 BACKBONE AMIDE HYDROGEN BONDS AND WATER CONTENT

None of the backbone amide to cysteinyl-S or S* hydrogen bonds were disrupted in the C77S HiPIP structure (Table 6.2). These hydrogen bonds have been suggested to stabilize the structure as well as to modulate the reduction potential of HiPIP (298). All of the [4Fe-4S] cluster associated hydrogen bond distances were essentially unaltered

211 except for N81-Sγ46 and N79-(Sγ/Oγ)77. A similar deviation for the N81-Sγ46 hydrogen bond was observed for H42Q HiPIP and therefore does not seem to be a manifestation of an altered Fe-S ligand. The shortening of the N79-(Sγ/Oγ)77 hydrogen bond is expected due to the greater electronegativity of an oxygen atom.

As previously noted, the major consideration with regard to cluster stability is solvent accessibility to the cluster. Of the three conserved water positions noted by

Parisini et al. (294), none were found to be significantly altered in the C77S HiPIP structure. Therefore, neither increased solvent accessibility to the cluster nor disruption of stabilizing hydrogen bonds can be used as an explanation for the decreased stability of

C77S HiPIP.

6.3.5. OPTICAL SPECTROSCOPY

Even though the X-ray crystal structures showed nearly identical structures, the electronic environment of the Fe-S cluster was clearly altered as can be seen by comparison of WT and C77S HiPIP UV-visible spectra. For example, the extinction coefficient of the Fe-S cluster for both reduced and oxidized C77S HiPIP cluster was less than that of WT (Figure 6.8). A more dramatic difference was observed by near-UV- visible CD (Figure 6.9). CD bands within this region reflect both the electronic and the geometric arrangement of the Fe-S cluster. WT HiPIP had CD peaks at 556 and 453 nm, while C77S HiPIP had CD absorption maxima at 556 and 451 nm. Additionally, there were differences below 400 nm. Although both proteins had CD bands at 556 nm and near 450 nm, the ratio of these peaks was dramatically different. For WT protein, the 556

212 peak was less intense than that near 450 nm, with a 556/453 absorbance ratio of 0.7. This

ratio was different for the oxygen ligated [4Fe-4S] protein. C77S HiPIP had a 556/453

absorbance ratio of 1.1, i.e. the 556 nm peak became more intense than the band near 450

nm. This increase in intensity of the longer wavelength band seems to be a characteristic

of oxygen ligated Fe-S clusters when compared to their complete Cys ligated forms.

Figure 6.8. UV-visible spectra of WT (solid line) and C77S (dashed line) reduced (A) and oxidized (B) HiPIP. The spectra were scaled so that the absorbance at the λmax were equal for WT and C77S HiPIP.

213

Figure 6.9. Near-UV-visible CD of WT (bottom line) and C77S HiPIP (top line).

6.3.6 RESONANCE RAMAN SPECTROSCOPY

Raman spectroscopy is a sensitive technique to characterize metal coordination to

biological molecules (299, 300), and was therefore used to complement the

crystallographic and spectroscopic data. Resonance Raman spectra with 458 nm, 488 nm, and 514 nm excitation were collected on reduced WT and C77S HiPIP (Figures 6.10 and 6.11). Spectra were also acquired on oxidized WT and C77S HiPIP (Figure 6.11).

Important spectral features are compared in Table 6.3. Consistent with previous studies

(301, 302), there was an upshift of the totally symmetric vibration of the [4Fe-4S] core

214 b (A1 mode where b refers to a bridging mode (302)) upon oxidation for both WT and

C77S HiPIP. This upshift was 4 cm-1 for both forms of HiPIP (from 339 cm-1 to 343 cm-1 for C77S and from 337 cm-1 to 341 cm-1 for WT HiPIP), which compares favorably with

the 3 - 5 cm-1 upshifts reported previously for WT (301, 302). At both oxidation states

the cluster’s totally symmetric vibration upshifted by 2 cm-1 for C77S relative to WT

HiPIP. Above 340 cm-1, the resonance Raman spectrum of reduced WT HiPIP was

dominated by the asymmetric Fe-S(Cys) mode at 361 cm-1 and a broad, less intense band

centered at 393 cm-1 in the proper energy range for a symmetric Fe-S(Cys) mode. Minor

alterations within this region were observed for C77S HiPIP and are further discussed

below (see discussion).

215

Figure 6.10. Resonance Raman spectra of reduced WT and C77S HiPIP with 458 nm

(panel A) and 514 nm (panel B) excitation. The samples were frozen at 30 K. The spectrograph was set with an 1800 groove/mm holograph grating and a 50 µm input slit width for each spectrum.

216

Figure 6.11. Resonance Raman spectra of WT and C77S HiPIPs in (A) the reduced and

(B) oxidized states. All spectra were recorded on frozen samples at 30 K using an excitation wavelength of 488 nm and a laser power of approximately 120 mW. The spectrograph was set with an 1800 groove/mm holograph grating and a 50 µm input slit width for each spectrum. The (*) denotes bands due to frozen solvent.

217 WT, reduced C77S, reduced WT, oxidized C77S, oxidized

Mainly terminal Fe-S stretching modes

393 389 406 406

361 363 376 379

Mainly bridging Fe-S stretching modes

390 389

337 339 341 343

274 273

249 248

Table 6.3. Resonance Raman Frequenciesa for WT and C77S HIPIPs. The excitation

wavelength was 488 nm. a All frequencies are in cm-1.

6.3.7 C77S HIPIP CLUSTER ACID SENSITIVITY

Both WT and C77S HiPIPs were quite stable at basic pH showing no signs of cluster degradation after 1 h at room temperature (Figure 6.12). WT HiPIP continued to

remain stable at all pHs tested. However, under more acidic conditions the C77S HiPIP

cluster began to degrade. At pH 3.0 the cluster of C77S HiPIP rapidly degraded with a

t1/2 of 4 min.

218

Figure 6.12. Acid catalyzed [4Fe-4S] degradation. A) Cluster degradation of C77S

(solid line) and WT (dashed line) HiPIP at pH 10.0 (top lines), pH 7.0 (middle lines), and pH 4.0 (bottom lines). Each tick mark on the y-axis represents a change in absorbance at

386 nm of 0.5. B) Cluster degradation at pH 3.0. Data were fit to a first order exponential decay.

6.4 DISCUSSION

6.4.1 Fe-S CLUSTER ASSOCIATED BOND LENGTHS

The shortening of the Ser77 - Fe4 bond by 0.11 Å is significantly less than that observed for Anabaena C49S ferredoxin (0.3 Å) (85) and C. pasteurianum C42S

219 rubredoxin (0.4 Å) (289). The reasons for this are unclear. C42S rubredoxin loses a

Cys42(Sγ)-N44 hydrogen bond upon Ser substitution, which may facilitate the lengthening of the Ser42 - Fe bond. However, C49S ferredoxin retains all of its cluster- associated hydrogen bonds. A more likely explanation is the rigidity of the polypeptide backbone. EXAFS results of single Cys to Ser mutants of rubredoxin suggests that interior ligand mutants have Ser - Fe bond lengths more similar to WT than exterior ligand mutants (C42 is an exterior ligand, i.e. more solvent exposed) (289). Such findings are consistent with greater backbone rigidity maintaining a particular ligand - Fe bond length. NMR studies have shown that HiPIP retains its structural integrity in the presence of 3 M guanidinium hydrochloride and only partially unfolds in 3.3 to 4.4 M guanidinium hydrochloride (303). Furthermore, the [4Fe-4S] cluster of HiPIP is deeply buried within its hydrophobic core (276). Thus, the conservation of protein ligand – Fe bond lengths for C77S HiPIP is likely a consequence of its great structural rigidity and its ability to surround the [4Fe-4S] cluster with hydrophobic residues.

Resonance Raman spectra of C77S and WT HiPIP revealed slight differences in

b cluster coordination. In particular, the increase in energy of the A1 mode upon oxidation

has been attributed to an overall shortening of the Fe-S* bonds of the cluster (301).

Interestingly, this same mode was shifted 2 cm-1 higher in the C77S protein form of

HiPIP relative to WT in both oxidation states. This is consistent with the crystallographic

data that revealed a minor shortening of the Fe-S* bond lengths in the C77S protein

b relative to WT reduced HiPIP. This upshift in the A1 mode was also reported in a study

of four serine variants of P. furiosus ferredoxin (235). Relative to the complete

220 cysteinyl-ligated mutant form of P. furiosus ferredoxin (D14C), the upshifts of this mode

ranged from 2 - 5 cm-1 for single Cys to Ser variants. Resonance Raman spectra could

not be acquired for all of the individual Cys to Ser cluster ligands of C. vinosum HiPIP

due to cluster instability (290).

In the study of the serine variants of P. furiosus ferredoxin (235), the asymmetric

Fe-S(Cys) modes proved most sensitive to changes in cluster ligation with the general

trend being downshifts of varying degree for both the symmetric and asymmetric Fe-St (t denotes terminal protein-ligand modes) modes relative to the system with complete cysteinyl ligation. Our data compares very closely with that of the D14S P. furiosus ferredoxin mutant in which the symmetric Fe-St mode shifted down by 3 cm-1 from 395

cm-1 in the D14C mutant to 392 cm-1, the asymmetric Fe-St mode shifts down 1 cm-1

-1 -1 b -1 from 363 cm to 362 cm , and the A1 mode mentioned above shifts up by only 2 cm .

The other serine mutants in this study showed much larger shifts in one or more of these modes. In the present study, relative to reduced WT HiPIP, these same modes in the reduced C77S HiPIP resonance Raman spectrum shifted by -4cm-1, +2cm-1 and +2cm-1.

These minor shifts reflect the subtle changes seen in the crystallographic data for reduced

C77S HiPIP, in which the Fe-Sγ bonds elongated by less than 0.1 Å and the Fe-Sγ-Cβ-Cα dihedral angles changed by less than 5 degrees relative to WT (Table 6.2).

6.4.2 ROLE OF PHE48 IN SOLVENT PROTECTION

Comparisons of all of the C. vinosum HiPIP crystal structures solved to date are nearly identical with the largest significant discrepancy being the positioning of Phe48.

221 Even within our single crystal the Phe48 orientation differs among the four molecules of

the asymmetric unit. An analysis of the accepted NMR structures of WT and C77S

HiPIP reveals even greater heterogeneity for Phe48 positioning (291, 297). Interestingly,

none of the other residues showed such a large degree of motion (Figure 6.6, panel D).

Thus, a unique situation exists in which the flexibility of a side-chain can be corroborated

both by crystal and solution structures of a protein. Such flexibility at this location within

HiPIP is particularly significant since Phe48 protects the cluster from its most solvent exposed region, which is consistent with NMR studies that identify Phe48 as being

crucial in maintaining solvent exclusion from the cluster (304). Therefore, the normal

breathing of the molecule may ultimately result in Phe48 controlling solvent access to the

cluster by alternating between “open” and “closed” conformations similar to that of the

distal histidine of hemoglobin which controls ligand entry into the heme pocket via a

swinging gate mechanism (305). Phe66 orientation is also altered between the different

HiPIP structures, although to a much lesser extent than for Phe48. The flexibility of

Phe66 may be indicative of a less frequently available solvent path to the cluster, or a manifestation of its ability to accommodate structural changes such as those encountered upon the “opening” of Phe48 and subsequent solvent entry. Surprisingly, a comparison of solved HiPIP structures from different bacteria reveals that neither Phe48 nor Phe66 are absolutely conserved. An analysis of their structures indicates that those HiPIPs with analogous Phe48 and Phe66 residues, such as Ectothiorhodospira halophila HiPIP (306), more likely go through a cluster degradation pathway similar to that described for

222 C. vinosum HiPIP. However, those with a non-conserved Phe48, as in

Ectothiorhodospira vacuolata HiPIP (278, 307, 308), or non-conserved Phe48 and Phe66

positions, as in Rhodocyclus tenuis HiPIP (309), appear to have different faces of the

Fe-S cluster that are maximally exposed to solvent, and therefore may have different

routes for solvent entry.

6.4.3. ELECTRONIC CHANGES IN C77S HiPIP

The UV-visible absorption and the near-UV-visible CD spectra are consistent

with previous NMR characterization of the altered electronic properties of the Fe-S

cluster of C77S HiPIP (290). For oxidized WT C. vinosum HiPIP, two of the iron ions

are in the ferric oxidation state and two exist as Fe2.5+. However, in solution there is an

equilibrium between two species, where one species ligates the mixed valence pair by

Cys63 and Cys77 and the other species coordinates the Fe2.5+ ions by Cys46 and Cys63.

The former species predominates for WT protein and accounts for 60% of the form

present in solution (310). However, for C77S HiPIP the equilibrium shifts resulting in the latter species accounting for 60% of the protein in solution (310). Even though there are changes in the electronic properties of the cluster, the redox potential only decreases

by ~30 mV (290).

6.4.3 ACID-CATALYZED CLUSTER DEGRADATION

The reasons for the decreased cluster stability of C77S HiPIP are not readily

apparent from the crystal structure. Neither cluster geometry, protein backbone or side-

223 chain orientations, water positioning, nor hydrogen bonding to the cluster are significantly perturbed. Therefore, there must be other non-structural factors associated with the intrinsic properties of Fe-S coordination by a serine ligand that make such interactions less favorable. All Cys to Ser mutations of Fe-S cluster ligating residues reported in the literature thus far have resulted in decreased cluster stability. Our observation of increased cluster susceptibility to acid-catalyzed degradation is consistent with the findings of Wedd and colleagues (289). Xiao et al. noted complete loss of Fe coordination to all single Cys to Ser mutants of C. pasteurianum rubredoxin within 5 min under acidic conditions, whereas native rubredoxin required at least an hour under the same conditions (289). Such differences in stability can be attributed to the much greater difficulty in maintaining a deprotonated serine rather than a cysteine. For a simple Fe-S cluster as in rubredoxin, which contains no inorganic sulfides, the mechanism of acid-catalyzed cluster disassembly is not significantly perturbed inasmuch as both mutant and native cluster degradation pathways most likely initiate with the protonation of protein ligand side-chains, although the order of ligand protonation may be altered. However, for a more complex Fe-S cluster, as found in HiPIP, the mechanism by which the cluster degrades is likely to be more significantly affected. The initial steps of the HiPIP [4Fe-4S] acid catalyzed cluster degradation pathway are thought to initiate by protonation of the bridging inorganic sulfides, not the protein ligands (88). Thus, the introduction of an easily protonated protein ligand allows for an altered and more facile route to the loss of Fe-S coordination.

224 Often Cys to Ser mutant Fe-S proteins express in the apo-form (9). Indeed, Fe-S coordination is a necessary step in the proper folding of HiPIP (130). Therefore, interference with the formation of the appropriate Fe-S(Cys) bonds may result in an inability of the protein to attain its functional conformation (130, 281). Mutations of cysteines that initiate cluster formation would be expected to result in a greater decrease in in vivo and in vitro formation of holo-protein than mutations of residues that coordinate to the cluster in the final stages of assembly. This again can be readily explained by a comparison of the pKas of cysteine and serine. The energetic penalty for deprotonating a serine (pKa ~ 16) is much greater than that for deprotonating a cysteine

(pKa ~ 8.35) (28, 77, 289). Such an effect would be particularly dominating in the early stages of cluster assembly since the protein has not yet had the opportunity to create an interior cluster-coordinating cavity, which could potentially alter the pKa of the subsequent ligating residues, thus making bond formation more favorable.

6.4.4 CONCLUSIONS

We have determined the crystal structure and report the results for the first Ser ligated [4Fe-4S] protein, thus confirming the presence of the Ser(Oγ)-Fe bond. Ser ligation effectively preserves the overall geometry and bond lengths of the HiPIP protein and Fe-S cluster. The major effect of Ser ligation appears to be sensitivity to proton mediated cluster disassembly, as a result of the increased pKa for Ser relative to Cys, rather than from structural factors. Furthermore, Phe48 has been identified as a likely candidate for the control of solvent access to the Fe-S cluster.

225

REFERENCES

1. Mason, J. R., and Cammack, R. (1992) Annu. Rev. Microbiol. 46, 277-305.

2. Iwata, S., Saynovits, M., Link, T. A., and Michel, H. (1996) Structure 15, 567- 579.

3. Kennedy, M. C., and Stout, C. D. (1992) Adv. Inorg. Chem. 38, 323-339.

4. Beinert, H., Kennedy, M. C., and Stout, C. D. (1996) Chem. Rev. 96, 2335-2373.

5. Beinert, H. (2000) J. Biol. Inorg. Chem. 5, 2-15.

6. Kennedy, M. C., Kent, T. A., Emptage, M., Merkle, H., Beinert, H., and Munck, E. (1984) J. Biol. Chem. 259, 14463-14471.

7. Gailer, J., George, G. N., Pickering, I. J., Prince, R. C., Kohlhepp, P., Zhang, D., Walker, F. A., and Winzerling, J. J. (2001) J. Am. Chem. Soc. 123, 10121-10122.

8. Calzolai, L., Gorst, C. M., Zhao, Z.-H., Teng, Q., Adams, M. W. W., and La Mar, G. N. (1995) Biochemistry 34, 11373-11384.

9. Moulis, J.-M., Davasse, V., Golinelli, M.-P., Meyer, J., and Quinkal, I. (1996) J. Biol. Inorg. Chem. 1, 2-14.

10. Sticht, H., and Rosch, P. (1998) Progress in Biophys. Mol. Biol. 70, 95-136.

226 11. Wachtershauser, G. (1992) Prog. Biophys. Mol. Biol. 58, 85-2002.

12. Rees, D. C., and Howard, J. B. (2003) Science 300, 929-931.

13. Beinert, H., Holm, R. H., and Munck, E. (1997) Science 277, 653-659.

14. Ollagnier, S., Meier, C., Mulliez, E., Gaillard, J., Schuenemann, V., Trautwein, A., Mattioli, T., Lutz, M., and Fontecave, M. (1999) J. Am. Chem. Soc. 121, 6344-6350.

15. Broderick, J. B., Duderstadt, R. E., Fernandez-Recio, J., Wojtuszewski, K., Henshaw, T. F., and Johnson, M. K. (1997) J. Am. Chem. Soc. 119, 7396-7397.

16. Kulzer, R., Pils, T., Kappl, R., Huttermann, J., and Knappe, J. (1998) J. Biol. Chem. 273, 4897-4903.

17. Lieder, K. W., Booker, S., Ruzicka, F. J., Beinert, H., Reed, G. H., and Frey, M. (1998) Biochemistry 37, 2578-2585.

18. Duin, E. C., Lafferty, M. E., Crouse, B. R., Allen, R. M., Sanyal, I., Flint, D. H., and Johnson, M. K. (1998) Biochemistry 36, 11811-11820.

19. Hidalgo, E., Ding, H., and Demple, B. (1997) Trends Biochem. Sci. 22, 207-210.

20. Gaudu, P., and Weiss, B. (1996) Proc. Natl. Acad. Sci. USA 93, 10094-10098.

21. Hentze, M., and Kuhn, L. C. (1996) Proc. Natl. Acad. Sci. USA 93, 8175-8182.

22. Haile, D. J., Rouault, T. A., Harford, J. B., Kennedy, M. C., Blondin, G. A., Beinert, H., and Klausner, R. D. (1992) Proc. Natl. Acad. Sci. USA 89, 11735- 11739.

23. Beinert, H., Kiley, P. J. (1999) Curr. Opin. Chem. Biol. 3, 152-157.

227 24. Khoroshilova, N., Beinert, H., and Kiley, P. J. (1995) Proc. Natl. Acad. Sci. USA 92, 2499-2503.

25. Khoroshilova, N., Popescu, C. V., Munck, E., Beinert, H., and Kiley, P. J. (1997) Proc. Natl. Acad. Sci. USA 94, 6087-6092.

26. Lazazzera, B. A., Beinert, H., Khoroshilova, N., Kennedy, M. C., and Kiley, P. J. (1996) J. Biol. Chem. 271, 2762-2768.

27. Lazazzera, B. A., Bates, D. M., and Kiley, P. J. (1993) Genes Dev. 7, 1993-2005.

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

29. O'Halloran, T. V., and Culotta, V. C. (2000) J. Biol. Chem. 275, 25057-25060.

30. Pufahl, R., Singer, C. P., Peariso, K. L., Lin, S.-J., Schmidt, P. J., Fahrni, C. J., Culotta, V. C., Penner-Hahn, J. E., and O'Halloran, T. V. (1997) Science 278, 853-856.

31. Odermatt, A., Suter, H., Krapf, R., and Solioz, M. (1993) J. Biol. Chem. 268, 12775-12779.

32. Rosenzweig, A. C. (2001) Acc. Chem. Res. 34, 119-128.

33. Rosenzweig, A. C., Huffman, D. L., Hou, M. Y., Wernimont, A. K., Pufahl, R., and O'Halloran, T. V. (1999) Structure 7, 605-617.

34. Wimmer, R., Herrmann, T., Solioz, M., and Wuthrich, K. (1999) J. Biol. Chem. 274, 22597-22603.

35. Steele, R., A., and Opella, S. J. (1997) Biochemistry 36, 6885-6895.

36. Culotta, V. C., Klomp, L. W., Strain, J., Casareno, R. I., Krems, B., and Gitlin, J. D. (1997) J. Biol. Chem. 272, 23469-23472.

228 37. Rae, T., Schmidt, P. J., Pufahl, R. A., Culotta, V. C., and O'Halloran, T. V. (1999) Science 284, 805-808.

38. Rae, T. D., Torres, A. S., Pufahl, R. A., and O'Halloran, T. V. (2001) J. Biol. Chem. 276, 5166-5176.

39. Schmidt, P. J., Rae, T. D., Pufahl, R. A., Hamma, T., Strain, J., O'Halloran, T. V., and Culotta, V. C. (1999) J. Biol. Chem. 274, 23719-23725.

40. Lamb, A. L., Torres, A. S., O'Halloran, T. V., and Rosenzweig, A. C. (2000) Biochemistry 39, 14720-14727.

41. Lamb, A. L., Torres, A. S., O'Halloran, T. V., and Rosenzweig, A. C. (2001) Nature Struct. Biol. 8, 751-755.

42. Lamb, A. L., Wernimont, A. K., Pufahl, R. A., Culotta, V. C., O'Halloran, T. V., and Rosenzweig, A. C. (1999) Nature Struct. Biol. 6, 724-729.

43. Hall, L. T., Sanchez, R. J., Holloway, S. P., Zhu, H., Stine, J. E., Lyons, T. J., Demeler, B., Schirf, B., Hansen, J. C., Nersissian, A. M., Valentine, J. S., and Hart, P. J. (2000) Biochemistry 39, 3611-3623.

44. Soriano, A., and Hausinger, R. P. (1999) Proc. Natl. Acad. Sci. USA 96, 11140- 11144.

45. Banci, L., Bertini, I., Ciofi-Baffoni, S., Finney, L. A., Outten, C. E., and O'Halloran, T. V. (2003) J. Mol. Biol. 323, 883-897.

46. Song, H. K., Mulrooney, S. B., Huber, R., and Hausinger, R. P. (2001) J. Biol. Chem. 276, 49359-49364.

47. Remaut, H., Safarov, N., Ciurli, S., and Beeumen, J. V. (2001) J. Biol. Chem. 276, 49365-49370.

48. Burgess, B. K., and Lowe, D. J. (1996) Chem. Rev. 96, 2983-3011.

229 49. Smith, B. E., and Eady, R. R. (1992) Eur. J. Biochem. 205, 1-15.

50. Dean, D., Bolin, J. T., and Zheng, L. (1993) J. Bacteriol. 175, 6737-6744.

51. Zheng, L., White, R. H., Cash, V. L., Jack, R. F., and Dean, D. R. (1993) Proc. Natl. Acad. Sci. USA 90, 2754-2758.

52. Fu, W., Jack, R. F., Morgan, T. V., Dean, D. R., and Johnson, M. K. (1994) Biochemistry 33, 13455-13463.

53. Zheng, L., White, R. H., and Dean, D. R. (1994) Biochemistry 33, 4714-4720.

54. Agar, J. N., Krebs, C., Frazzon, J., Huynh, B. H., Dean, D. R., and Johnson, M. K. (2000) Biochemistry 39, 7856-7862.

55. Hwang, D. M., Dempsey, A., Tan, K.-T., and Liew, C.-C. (1996) J. Mol. Evol. 43, 536-540.

56. Frazzon, J., Fick, J. R., and Dean, D. R. (2002) Biochem. Soc. Trans. 30, 680-685.

57. Flint, D. H. (1996) J. Biol. Chem. 271, 16068-16074.

58. Zheng, L., Cash, V. L., Flint, D. H., and Dean, D. R. (1998) J. Biol. Chem. 273, 13264-13272.

59. Strain, J., Lorenz, C. R., Bode, J., Smolen, G. A., Garland, S. A., Vickery, L. E., and Culotta, V. C. (1998) J. Biol. Chem. 273, 31138-31144.

60. Barros, M. H., and Nobrega, F. G. (1999) Gene 233, 197-203.

61. Lill, R., Dickert, K., Kaut, A., Lange, H., Pelzer, W., Prohl, C., and Kispal, G. (1999) Biol. Chem. 380, 1157-1166.

62. Muhlenhoff, U., and Lill, R. (2000) Biochim. Biophys. Acta 1459, 370-382.

230 63. Land, T., and Rouault, T. A. (1998) Mol. Cell. 2, 807-815.

64. Nakai, M., Yoshihara, Y., Hayashi, H., and Kagamiyama, H. (1998) FEBS Lett. 433, 143-148.

65. Kispal, G., Csere, P., Prohl, C., and Lill, R. (1999) EMBO J. 18, 3981-3989.

66. Schilke, B., Voisine, C., Beinert, H., and Craig, E. (1999) Proc. Natl. Acad. Sci. USA 96, 10206-10211.

67. Garland, S. A., Hoff, K., Vickery, L. E., and Culotta, V. C. (1999) J. Mol. Biol. 294, 897-907.

68. Jensen, L. T., and Culotta, V. C. (2000) Mol. Cell. Biol. 20, 3918-3927.

69. Kaut, A., Lange, H., Diekert, K., Kispal, G., and Lill, R. (2000) J. Biol. Chem. 275, 15955-15961.

70. Lange, H., Kaut, A., Kispal, G., and Lill, R. (2000) Proc. Natl. Acad. Sci. USA 97, 1050-1055.

71. Leighton, J., and Schatz, G. (1995) EMBO J. 14, 188-195.

72. Muhlenhoff, U., Richhardt, N., Gerber, J., and Lill, R. (2002) J. Biol. Chem. 277, 29810-29816.

73. Mansy, S. S., Wu, G., Surerus, K. K., and Cowan, J. A. (2002) J. Biol. Chem. 277, 21397-21404.

74. Foster, M. W., Mansy, S. S., Hwang, J., Penner-Hahn, J. E., Surerus, K. K., and Cowan, J. A. (2000) J. Am. Chem. Soc. 122, 6805-6806.

75. Wu, G., Mansy, S. S., Wu, S., Surerus, K. K., Foster, M. W., and Cowan, J. A. (2002) Biochemistry 41, 5024-5032.

231 76. Wu, G., Mansy, S. S., Hemann, C., Hille, R., Surerus, K. K., and Cowan, J. A. (2002) J. Biol. Inorg. Chem. 7, 526-532.

77. Xia, B., Cheng, H., Bandarian, V., Reed, G. H., and Markley, J. L. (1996) Biochemistry 35, 9488-9495.

78. Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989) Molecular cloning: a laboratory manual, 2 ed., Cold Spring Harbor Laboratory Press, Plainview, NY.

79. You, J., Cohen, R. E., and Pickart, C. M. (1999) BioTechniques 27, 950-954.

80. Agar, J. N., Yuvaniyama, P., Jack, R. F., Cash, V. L., Smith, A. D., Dean, D. R., and Johnson, M. K. (2000) J. Biol. Inorg. Chem. 5, 167-177.

81. Bates, D. M., Popescu, C. V., Khoroshilova, N., Vogt, K., Beinert, H., Munck, E., and Kiley, P. J. (2000) J. Biol. Chem. 275, 6234-6240.

82. Wu, S., Wu, G., Surerus, K. K., and Cowan, J. A. (2002) Biochemistry 41, 8876- 8885.

83. Urbina, H. D., Silberg, J. J., Hoff, K. E., and Vickery, L. E. (2001) J. Biol. Chem. 276, 44521-44526.

84. Iametti, S., Bera, A. K., Vecchio, G., Grinberg, A., Bernhardt, R., and Bonomi, F. (2001) Eur. J. Biochem. 268, 2421-2429.

85. Hurley, J. K., Weber-Main, A. M., Hodges, A. E., Stankovich, M. T., Benning, M. M., Holden, H. M., Cheng, H., Xia, B., Markley, J. L., Genzor, C., Gomez- Moreno, C., Hafezi, R., and Tollin, G. (1997) Biochemistry 36, 15109-15117.

86. Hagen, W. R., Silva, P. J., Amorim, M. A., Hagedoorn, P.-L., Wassink, H., Haaker, H., and Robb, F. T. (2000) J. Biol. Inorg. Chem. 5, 527-534.

87. Agar, J. N., Zheng, L., Cash, V. L., Dean, D. R., and Johnson, M. K. (2000) J. Am. Chem. Soc. 122, 2136-2137.

232 88. Foster, M. W., Bian, S., Surerus, K. K., and Cowan, J. A. (2001) J. Biol. Inorg. Chem. 6, 266-274.

89. Yuvaniyama, P., Agar, J. N., Cash, V. L., Johnson, M. K., and Dean, D. R. (2000) Proc. Natl. Acad. Sci. USA 97, 599-604.

90. Lill, R., and Kispal, G. (2000) Trends Biochem. Sci. 25, 352-356.

91. Hang, H., Kispal, G., Kaut, A., and Lill, R. (2000) Proc. Natl. Acad. Sci. USA 97, 1050-1055.

92. Lutz, T., Westermann, B., Neupert, W., and Herrmann, J. M. (2001) J. Mol. Biol. 307, 815-825.

93. Gergely, J., and Grabarek, Z. (1990) Anal. Biochem. 185, 131-135.

94. Grinberg, A. V., Hannemann, F., Schiffler, B., Muller, J., Heinemann, U., and Bernhardt, R. (2000) Proteins 40, 590-612.

95. Coghlan, V., and Vickery, L. E. (1991) J. Biol. Chem. 266, 18606-18612.

96. Brandt, M. E., and Vickery, L. E. (1993) J. Biol. Chem. 268, 17126-17130.

97. Wu, S., and Cowan, J. A. (2003) Biochemistry 42, 5784-5791.

98. Mansy, S. S., Xiong, Y., Hemann, C., Hille, R., Sundaralingam, M., and Cowan, J. A. (2002) Biochemistry 41, 1195-1201.

99. Yeh, A. P., Ambroggio, X. I., Andrade, S. L. A., Einsle, O., Chatelet, C., Meyer, J., and Rees, D. C. (2002) J. Biol. Chem. 277, 34499-34507.

100. Foster, M. W. (2000) in Chemistry, Ohio State University, Columbus.

101. Wu, G., and Cowan, J. A. (2000) Unpublished data.

233 102. Bocchetta, M., Gribaldo, S., Sanangelantoni, A., and Cammarano, P. (2000) J. Mol. Evol. 50, 366-380.

103. Stetter, K. O. (1996) FEMS Microbiol. Rev. 18, 149-158.

104. Michelini, E. T., and Flynn, G. C. (1999) J. Bacteriol. 181, 4237-4244.

105. Kaiser, J. T., Clausen, T., Bourenkow, G. P., Bartunik, H.-D., Steinbacher, S., and Huber, R. (2000) J. Mol. Biol. 297, 451-464.

106. Muller, K., Matzanke, B. F., Schunemann, V., Trautwein, A. X., and Hantke, K. (1998) Eur. J. Biochem. 258, 1001-1008.

107. Patzer, S. I., and Hantke, K. (1999) J. Bacteriol. 181, 3307-3309.

108. Tokumoto, U., and Takahashi, Y. (2002) J. Biol. Chem. 277, 28380-28383.

109. Ollagnier-De-Choudens, S., Nachin, L., Sanakis, Y., Loiseau, L., Barras, F., and Fontecave, M. (2003) J. Biol. Chem. 278, 17993-8001.

110. Mihara, H., and Esaki, N. (2002) Appl. Microbiol. Biotechnol. 60, 12-23.

111. Nachin, L., El Hassouni, M., Loiseau, L., Expert, D., and Barras, F. (2001) Mol. Microbiol. 39, 960-972.

112. Nishio, K., and Nakai, M. (2000) J. Biol. Chem. 275, 22615-22618.

113. Agarwal, A., Tan, J., Eren, M., Tevelev, A., Lui, S. M., and Cowan, J. A. (1993) Biochem. Biophys. Res. Commun. 197, 1357.

114. Bodenhausen, G., and Ruben, D. J. (1980) Chem. Phys. Lett. 69, 185.

115. Johnson, W. C., Jr. (1990) Proteins 7, 205-214.

234 116. Sreerama, N., and Woody, R. W. (1994) J. Mol. Biol. 242, 497-501.

117. Deleage, G., and Geourjon, C. (1993) Comp. Appl. Biosc. 9, 197-199.

118. Davies, G. E., and Stark, G. R. (1970) Proc. Natl. Acad. Sci. USA 66, 651-656.

119. Radaev, S., Dastidar, P., Patel, M., Woodard, R. W., and Gatti, D. (2000) J. Biol. Chem. 275, 9476-9484.

120. Hoff, K. E., Silberg, J. J., and Vickery, L. E. (2000) Proc. Natl. Acad. Sci. USA 97, 7790-7795.

121. Olson, J. W., Agar, J. N., Johnson, M. K., and Maier, R. J. (2000) Biochemistry 39, 16213-16219.

122. Ollagnier-De-Choudens, S., Mattioli, T., Takahashi, Y., and Fontecave, M. (2001) J. Biol. Chem. 276, 22604-22607.

123. Jung, Y.-S., Roberts, V. A., Stout, C. D., and Burgess, B. K. (1999) J. Biol. Chem. 274, 2978-2987.

124. Vidakovic, M. S., Fraczkiewicz, G., and Germanas, J. P. (1996) J. Biol. Chem. 271, 14734-14739.

125. Verhagen, M. F. J. M., O'Rourke, T. W., Menon, A. L., and Adams, M. W. W. (2001) Biochim. Biophys. Acta 1505, 209-219.

126. Jung, Y.-S., Sheridan-Gao, H. S., Christiansen, J., Dean, D. R., and Burgess, B. K. (1999) J. Biol. Chem. 274, 32402-32410.

127. Schnitman, C. A. (1974) J. Bacteriol. 118, 442-453.

128. Hindennach, I., and Henning, U. (1975) Eur. J. Biochem. 59, 207-213.

235 129. Pfeil, W., Gesierich, U., Kleemann, G. R., and Sterner, R. (1997) J. Mol. Biol. 272, 591-596.

130. Natarajan, K., and Cowan, J. A. (1997) J. Am. Chem. Soc. 119, 4082-4083.

131. Pozdnyakova, I., and Wittung-Stafshede, P. (2001) J. Am. Chem. Soc. 123, 10135-10136.

132. Hall, D. O., Cammack, R., and Rao, K. K. (1971) Nature 233, 136-138.

133. Schulz, H., Hennecke, H., and Thony-Meyer, L. (1998) Science 281, 1197-1200.

134. Wernimont, A. K., Huffman, D. L., Lamb, A. L., O'Halloran, T. V., and Rosenzweig, A. C. (2000) Nature Struct. Biol. 7, 766-771.

135. Arnesano, F., Banci, L., Bertini, I., Huffman, D. L., and O'Halloran, T. V. (2001) Biochemistry 40, 1528-1539.

136. Banci, L., Bertini, I., Del Conte, R., Markey, J., and Ruiz-Duenas, J. (2001) Biochemistry 40, 15660-15668.

137. Banci, L., Bertini, I., Ciofi-Baffoni, S., D'Onofrio, M., Gonnelli, L., Marhuenda- Egea, F. C., and Ruiz-Duenas, F. J. (2002) J. Mol. Biol. 317, 415-429.

138. Arnoux, P., Haser, R., Izadi, N., Lecroisey, A., Delepierre, M., Wandersman, C., and Czjzek, M. (1999) Nature Struct. Biol. 6, 516-520.

139. Rosenzweig, A. C., and O'Halloran, T. V. (2000) Curr. Opin. Chem. Biol. 4, 140- 147.

140. Ciurli, S., Safarov, N., Miletti, S., Dikiy, A., Christensen, S. K., Kornetzky, K., Bryant, D. A., Vandenberghe, I., Devreese, B., Samyn, B., Remaut, H., and Beeumen, J. V. (2002) J. Biol. Inorg. Chem. 7, 623-631.

236 141. Schwartz, C. J., Giel, J. L., Patschkowski, T., Luther, C., Ruzicka, F. J., Beinert, H., and Kiley, P. J. (2001) Proc. Natl. Acad. Sci. USA 98, 14895-14900.

142. Krebs, C., Agar, J. N., Smith, A. D., Frazzon, J., Dean, D. R., Huynh, B. H., and Johnson, M. K. (2001) Biochemistry 40, 14069-14080.

143. Takahashi, Y., and Nakamura, M. (1999) J. Biochem. 126, 917-926.

144. Piotto, M., Saudek, V., and Sklender, V. (1992) J. Biomol. NMR 2, 661-666.

145. Kay, L. E., Ikura, M., Tschudin, R., and Bax, A. (1990) J. Magn. Reson. 89, 496- 514.

146. Stonehouse, J., Clowes, R. T., Shaw, G. L., Keeler, J., and Laue, E. E. (1995) J. Biomol. NMR 5, 226-232.

147. Salzmann, M., Pervushin, K., Wider, G., Senn, H., and Wuthrich, K. (1998) Proc. Natl. Acad. Sci. USA 95, 13585-13590.

148. Kay, L. E., Xu, G. Y., Singer, A. U., Muhandiram, D. R., and Forman-Kay, J. D. (1993) J. Magn. Reson. Ser. B 101, 333-337.

149. Vuister, G. W., and Bax, A. (1993) J. Am. Chem. Soc. 115, 7772-7777.

150. Archer, S. J., Ikura, M., Torchia, D. A., and Bax, A. (1991) J. Magn. Reson. 95, 636-641.

151. Wider, G., Neri, D., Otting, G., and Wuthrich, K. (1989) J. Magn. Reson. 85, 426- 431.

152. Gagne, R. R., Tsuda, S., Li, M. X., Chandra, M., Smillie, L. B., and Sykes, B. D. (1994) Protein Sci. 3, 1961-1974.

153. Grzesiek, S., and Bax, A. (1992) J. Magn. Reson. 96, 432-440.

237 154. Schleucher, J., Sattler, M., and Griesinger, C. (1993) Angew. Chem. Int. Ed. 32, 1489-1491.

155. Kay, L. E., Xu, G. Y., and Yamazaki, T. (1994) J. Magn. Reson. A109, 129-133.

156. Banci, L., Felli, I. C., and Kummerle, R. (2002) Biochemistry 41, 2913-2929.

157. Ruckert, M., and Otting, G. (2000) J. Am. Chem. Soc. 122, 7793-7797.

158. Ottiger, M., Delaglio, F., and Bax, A. (1998) J. Magn. Reson. 131, 373-378.

159. Wishart, D. S., and Sykes, B. D. (1994) J. Biomol. NMR 4, 171-180.

160. Grzesiek, S., and Bax, A. (1993) J. Am. Chem. Soc. 115, 12593-12594.

161. Marion, D., and Wuthrich, K. (1983) Biochem. Biophys. Res. Commun. 113, 967- 974.

162. Kay, L. E., Nicholson, L. K., Delaglio, F., Bax, A., and Torchia, D. A. (1992) J. Magn. Reson. 97, 359-375.

163. Kay, L. E., Torchia, D. A., and Bax, A. (1989) Biochemistry 28, 8972-8979.

164. Marquardt, D. W. (1963) J. Soc. Ind. Appl. Math. 11, 431-441.

165. Press, W. H., Flannery, B. P., Teukolsky, S. A., and Vetterling, W. T. (1998) Numerical recipes in C- the art of scientific computing, Cambridge University Press, New York.

166. Palmer, A. G., Rance, M., and Wright, P. E. (1991) J. Am. Chem. Soc. 113, 4371- 4380.

238 167. Schleucher, J., Schwendinger, M., Sattler, M., Schmidt, P. J., Schedletzky, O., Glaser, S. J., Sorensen, O. W., and Griesinger, C. (1994) J. Biomol. NMR 4, 301- 306.

168. Eccles, C., Guntert, P., Billeter, M., and Wuthrich, K. (1991) J. Biomol. NMR 1, 111-130.

169. Guntert, P., Braun, W., and Wuthrich, K. (1991) J. Mol. Biol. 217, 517-530.

170. Guntert, P., Mumenthaler, C., and Wuthrich, K. (1997) J. Mol. Biol. 273, 283- 298.

171. Herrmann, T., Guntert, P., and Wuthrich, K. (2002) J. Mol. Biol. 319, 209-227.

172. Koradi, R., Billeter, M., and Wuthrich, K. (1996) J. Mol. Graphics 14, 51-55.

173. Martins, L. O., Carreto, L. S., Da Costa, M. S., and Santos, H. (1996) J. Bacteriol. 178, 5644-5651.

174. Farrow, N. A., Zhang, O., Forman-Kay, J. D., and Kay, L. E. (1997) Biochemistry 36, 2390-2402.

175. Cavagnero, S., Nishimura, C., Schwarzinger, S., Dyson, H. J., and Wright, P. E. (2001) Biochemistry 40, 14459-14467.

176. Yao, J., Chung, J., Wright, P. E., and Dyson, H. J. (2001) Biochemistry 40, 3561- 3571.

177. Matthews, J. M., Norton, R. S., Hammacher, A., and Simpson, R. J. (2000) Biochemistry 39, 1942-1950.

178. Koshiba, T., Yao, M., Kobashigawa, M., Demura, M., Nakagawa, M., Tanaka, I., Kuwajima, K., and Nitta, K. (2000) Biochemistry 39, 3248-3257.

179. Jardetzky, O. (1980) Biochim. Biophys. Acta. 621, 227-232.

239 180. Burgi, R., Pitera, J., and Van Gunsteren, W. F. (2001) J. Biomol. NMR 19, 305- 320.

181. Fejzo, J., Krezel, A. M., Westler, W. M., Macura, S., and Markey, J. (1991) Biochemistry 16, 3807-3811.

182. Alexandrescu, A. T., Evans, P. A., Pitkeathly, M., Baum, J., and Dobson, C. M. (1993) Biochemistry 32, 1707-1718.

183. Eliezer, D., and Wright, P. E. (1996) J. Mol. Biol. 263, 531-538.

184. Schulman, B. A., Kim, P. S., Dobson, C. M., and Redfield, C. (1997) Nat. Struct. Biol. 4, 630-634.

185. Eliezer, D., Jennings, P. A., Dyson, H. J., and Wright, P. E. (1997) FEBS Lett. 417, 92-96.

186. Smith, L. J., Dobson, C. M., and Van Gunsteren, W. F. (1999) J. Mol. Biol. 286, 1567-1580.

187. Darby, N. J., and Creighton, T. E. (1993) Protein Structure, Oxford university press, Inc., New York.

188. Guijarro, J. I., Jackson, M., Chaffotte, A. F., Delepierre, M., Mantsch, H. H., and Goldberg, M. E. (1995) Biochemistry 34, 2998-3008.

189. Troullier, A., Reinstadler, D., Dupont, Y., Naumann, D., and Forge, V. (2000) Nature Struct. Biol. 7, 78-86.

190. Xiandong, P., Jonas, J., and Silva, J. L. (1993) Proc. Natl. Acad. Sci. USA 90, 1776-1780.

191. Yuan, C., Byeon, I.-J., Poi, M.-J., and Tsai, M.-D. (1999) Biochemistry 38, 2919- 2929.

240 192. Wright, P. E., and Dyson, H. J. (1999) J. Mol. Biol. 293, 321-331.

193. Dyson, H. J., and Wright, P. E. (2002) Curr. Opin. Struct. Biol. 12, 54-60.

194. Twigg, P. D., Parthasarathy, G., Guerrero, L., Logan, T. M., and Casper, D. L. D. (2001) Proc. Natl. Acad. Sci. USA 98, 11259-11264.

195. Kim, A. S., Kakalis, L. T., Abdul-Manan, N., Liu, G. A., and Rosen, M. K. (2000) Nature 404, 151-158.

196. Samuel, D., Kumar, T. K. S., Srimathi, T., Hsieh, H.-C., and Yu, C. (2000) J. Biol. Chem. 275, 34968-34975.

197. Kobayashi, T., Ikeguchi, M., and Sugai, S. (2000) J. Mol. Biol. 299, 757-770.

198. Baum, J., Dobson, C. M., Evans, P. A., and Hanley, C. (1989) Biochemistry 28, 7- 13.

199. Torres, A. S., Petri, V., Rae, T. D., and O'Halloran, T. V. (2001) J. Biol. Chem. 276, 38410-38416.

200. Banci, L., Bertini, I., Ciofi-Baffoni, S., Huffman, D. L., and O'Halloran, T. V. (2001) J. Biol. Chem. 276, 8415-8426.

201. Hamza, I., Schaefer, M., Klomp, L. W., and Gitlin, J. D. (1999) Proc. Natl. Acad. Sci. USA 96, 13363-13368.

202. Larin, D., Mekios, C., Das, K., Ross, B., Yang, A.-S., and Gilliam, T. C. (1999) J. Biol. Chem. 1999, 28497-28504.

203. Bertini, I., Cowan, J. A., Del Bianco, C., Luchinat, C., and Mansy, S. S. (2003) J. Mol. Biol. In press.

204. Ptitsyn, O. B. (1987) J. Protein Chem. 6, 273-293.

241 205. Ohgushi, M., and Wada, A. (1983) FEBS Lett. 164, 21-24.

206. Kuwajima, K. (1977) J. Mol. Biol. 114, 241-258.

207. Weber, G., and Young, L. R. (1964) J. Biol. Chem. 239, 1415-1421.

208. Lakowicz, J. R. (1983) Principles of Fluorescence Spectroscopy, Plenum Press, New York.

209. Felitsky, D. J., and Record, M. T. (2003) Biochemistry 42, 2202-2217.

210. Semisotnov, G. V., Rodionova, N. A., Razgulyaev, O. I., Uversky, V. N., Gripas, A. F., and Gilmanshin, R. I. (1991) Biopolymers 31, 119-128.

211. Reddy, G. B., Das, K. P., Petrash, J. M., and Surewicz, W. K. (2000) J. Biol. Chem. 275, 4565-4570.

212. Kusmierczyk, A. R., and Martin, J. (2000) J. Biol. Chem. 275, 33504-33511.

213. Chaffotte, A., Guillou, Y., Delepierre, M., Hinz, H., and Goldberg, M. E. (1991) Biochemistry 30, 8067-8074.

214. Eisenberg, D., and Crothers, D. (1979) Physical Chemistry with Applications to the life Sciences, Benjamin/Cummings Publishing Company, Inc., Menlo Park.

215. Schurmann, G., Haspel, M., Grumet, M., and Erickson, H. P. (2001) Mol. Cell. Biol. 12, 1765-1773.

216. Tanford, C. (1964) J. Am. Chem. Soc. 86, 2050-1090.

217. Schellman, J. A. (1987) Biopolymers 26, 549-559.

218. Burova, T. V., Bernhardt, R., and Pfeil, W. (1995) Protein Sci. 4, 909-916.

242 219. Hurley, J. K., Caffrey, M. S., Markley, J. L., Cheng, H., Xia, B., Chae, Y., Holden, H. M., and Tollin, G. (1995) Protein Sci. 4, 58-64.

220. Spuergin, P., Abele, U., and Schulz, G. E. (1995) Eur. J. Biochem. 231, 405-413.

221. Betton, J., and Hofnung, M. (1996) J. Biol. Chem. 271, 8046-8052.

222. Ahmad, F., Yadav, S., and Taneja, S. (1992) Biochem. J. 287, 481-485.

223. Fernandez-Recio, J., Romero, A., and Sancho, J. (1999) J. Mol. Biol. 290, 319- 330.

224. Genzor, C. G., Beldarrain, A., Gomez-Moreno, C., Lopez-Lacomba, J. L., Cortijo, M., and Sancho, J. (1996) Protein Sci. 5, 1376-1388.

225. Schechter, A. N., and Epstein, C. J. (1968) J. Mol. Biol. 35, 567-589.

226. Sirangelo, I., Bismuto, E., and Irace, G. (1994) FEBS Lett. 338, 11-15.

227. Tchorzewski, M., Krokowski, D., Boguszewska, A., Liljas, A., and Grankowski, N. (2003) Biochemistry 42, 3399-3408.

228. Muzammil, S., Kumar, Y., and Tayyab, S. (2000) Proteins 40, 29-38.

229. Wange, C., Lascu, I., and Giartosio, A. (1998) Biochemistry 37, 8457-8464.

230. Takase, K., Nitta, K., and Sugai, S. (1974) Biochim. Biophys. Acta 371, 352-359.

231. Xie, D., Bhakuni, V., and Freire, E. (1991) Biochemistry 30, 10673-10678.

232. Creighton, T. E., and Ewbank, J. J. (1993) Biochemistry 32, 3694-3707.

233. Montono, C., Yamagishi, A., and Oshima, T. (1999) Biochemistry 38, 1332-1337.

243 234. Wittung-Stafshede, P. (1998) Biochim. Biophys. Acta 1382, 324-332.

235. Brereton, P. S., Duderstadt, R. E., Staples, C. R., Johnson, M. K., and Adams, M. W. W. (1999) Biochemistry 38, 10594-10605.

236. Ogasahara, K., Nakamura, M., Nakura, S., Tsunasawa, S., Kato, I., Yoshimoto, T., and Ytani, K. (1998) Biochemistry 37, 17537-17544.

237. Del Vecchio, P., Graziano, G., Granata, V., Barone, G., and Mandrich, L. (2002) Biochemistry 41, 1364-1371.

238. Wittung-Stafshede, P., Gomes, C. M., and Teixeira, M. (2000) J. Inorg. Biochemistry 78, 35-41.

239. Schlee, S., and Reinstein, J. (2002) Cell. Mol. Life Sci. 59, 1598-1606.

240. Ptitsyn, O. B., Pain, R. H., Semisotnov, G. V., Zerovnik, E., and Razgulyaev, O. I. (1990) FEBS Lett. 262, 20-24.

241. Fink, A. L., Calciano, L. J., Goto, Y., Kurotsu, T., and Palleros, D. R. (1994) Biochemistry 33, 12504-12511.

242. de Laureto, P. P., De Filippis, V., Di Bello, M., Zambonin, M., and Fontana, A. (1995) Biochemistry 34, 12596-12604.

243. de Laureto, P. P., Frare, E., Gottardo, R., Van Dael, H., and Fontana, A. (2002) Protein Sci. 11, 2932-2946.

244. Stryer, L. (1965) J. Mol. Biol. 13, 482-495.

245. Turner, D. C., and Brand, L. (1968) Biochemistry 7, 3381-3390.

246. Matulis, D., and Lovrien, R. E. (1998) Biophys. J. 74, 422-429.

244 247. Tanford, C., Kawahara, K., and Lapanje, S. (1967) J. Am. Chem. Soc. 89, 729- 736.

248. Reynolds, J. A., and Tanford, C. (1970) J. Biol. Chem. 245, 5161-5165.

249. McElroy, C., Manfredo, A., Wendt, A., Gollnick, P., and Foster, M. P. (2003) J. Mol. Biol. 323, 463-473.

250. Hoff, K. G., Ta, D. T., Tapley, T. L., Silberg, J. J., and Vickery, L. E. (2002) J. Biol. Chem. 277, 27353-27359.

251. Lieutaud, C., Nitschke, W., Vermeglio, A., Parot, P., and Schoepp-Cothenet, B. (2003) Biochem. Biophys. Acta 1557, 83-90.

252. Vermeglio, A., Li, J., Schoepp-Cothenet, B., Pratt, N., and Knaff, D. B. (2002) Biochemistry 41, 8868-8875.

253. Venturoli, G., Mamedov, M., Mansy, S. S., Musiani, F., Strocchi, M., Francia, F., Semenov, A. Y., Cowan, J. A., and Ciurli, S. (2003) Submitted.

254. Zannoni, D., and Daldal, F. (1993) Arch. Microbiol. 160, 413-413.

255. Samyn, F., De Smet, L., Van Driessche, G., Meyer, T. e., Bartsch, R. G., Gusanovich, M. A., and Van Veeumen, J. J. (1996) Eur. J. Biochem. 236, 689- 696.

256. Kerfeld, C. A., Chan, C., Hirasawa, M., Kleis-SanFrancisco, S., Yeates, T. O., and Knaff, D. B. (1996) Biochemistry 35, 7812-7818.

257. Bersch, B., Blackledge, M. J., Meyer, T. E., and Marion, D. (1996) J. Mol. Biol. 264, 567-584.

258. Menin, L., Schoepp, B., Parot, P., and Vermeglio, A. (1997) Biochemistry 1997, 12183-12188.

245 259. Schoepp, B., Parot, P., Menin, L., Gaillard, J., Richaud, P., and Vermeglio, A. (1995) Biochemistry 34, 11736-11742.

260. Hochkoeppler, A., Ciurli, S., Venturoli, G., and Zannoni, D. (1995) FEBS Lett. 357, 70-74.

261. Hochkoeppler, A., Zannoni, D., Ciurli, S., Meyer, J., Cusanovich, M. A., and Tollin, G. (1996) Proc. Natl. Acad. Sci. USA 93, 6998-7002.

262. Nagashima, K. V. P., Matsuura, K., Shimada, K., and Vermeglio, A. (2003) Biochemistry 41, 14028-14032.

263. Jenny, F. E., Jr., and Daldal, F. (1993) EMBO J. 12, 1283-1292.

264. Jenny, F. E., Jr., Prince, R. C., and Daldal, F. (1994) Biochemistry 33, 2469-2502.

265. Bartsch, R. G. (1978) Methods Enzymol. 53, 329-340.

266. Cowan, J. A., and Lui, S. M. (1998) Adv. Inorg. Chem. 45, 313-350.

267. Cammack, R. (1973) Biochem. Biophys. Res. Commun. 54, 548-554.

268. Moss, T. H., Petering, D., and Palmer, G. (1969) J. Biol. Chem. 244, 2275-2277.

269. Dickson, D. P. E., Johnson, C. E., Cammack, R., Evans, M. C. W., Hall, D. O., and Rao, K. K. (1974) Biochem. J. 139, 105.

270. Middleton, P., Dickson, D. P. E., Johnson, C. E., and Rush, J. D. (1980) Eur. J. Biochem. 104, 289-296.

271. Bertini, I., Campos, A. P., Luchinat, C., and Teixeira, M. (1993) J. Inorg. Biochem. 52, 227-234.

272. Bertini, I., Ciurli, S., and Luchinat, C. (1995) Struct. Bonding 83, 1-54.

246 273. Bertini, I., Donaire, A., Felli, I. C., Luchinat, C., and Rosato, A. (1997) Inorg. Chem. 36, 4798-4803.

274. Banci, L., Bertini, I., Savellini, G. G., and Luchinat, C. (1996) Inorg. Chem. 35, 4248-4253.

275. Dus, K., DeKlerk, H., Sletten, K., and Bartsch, R. G. (1967) Biochim. Biophys. Acta 140, 291-311.

276. Carter, C. W., Kraut, J., Freer, S., Xuong, N.-H., Alden, R. A., and Bartsch, R. G. (1974) J. Biol. Chem. 249, 4212-4225.

277. Iwagami, S. G., Creagh, A. L., Haynes, C. A., Borsari, M., Felli, I. C., Piccioli, M., and Eltis, L. D. (1995) Protein Sci. 4, 2562-72.

278. Bertini, I., Couture, M. J., Donaire, A., Eltis, L. D., Felli, I. C., Luchinat, C., Picciolo, M., and Rosato, A. (1996) Eur. J. Biochem. 241, 440-452.

279. Bian, S., and Cowan, J. A. (1999) Coord. Chem. Rev. 190-192, 1049-1066.

280. Bian, S., and Cowan, J. A. (1998) J. Am. Chem. Soc. 120, 3532-3533.

281. Bentrop, D., Bertini, I., Iacoviello, R., Luchinat, C., Niikura, Y., Piccioli, M., Presenti, C., and Rosato, A. (1999) Biochemistry 38, 4669-4680.

282. Jensen, G. M., Warshel, A., and Stephen, P. J. (1994) Biochemistry 33.

283. Robbins, A. H., and Stout, C. D. (1989) Proc. Natl. Acad. Sci. USA 86, 3639- 3643.

284. Volbeda, A., Charon, M.-H., Piras, C., Hatchikian, E. C., Frey, M., and Fontecilla-Camps, J. C. (1995) Nature 373, 580-587.

285. Beinert, H., and Thomson, A. J. (1983) Arch. Biochem. Biophys. 222, 333-361.

247 286. Shen, B., Jollie, D. R., Diller, T. C., Stout, C. D., Stephens, P. J., and Burgess, B. K. (1995) Proc. Natl. Acad. Sci. USA 92, 10064-10068.

287. Schindelin, H., Kisker, C., Hilton, K., Rajagopalan, K. V., and Rees, D. C. (1996) Science 272, 1615-1621.

288. Kim, J., and Rees, D. C. (1992) Science 257, 1677-1682.

289. Xiao, Z., Lavery, M. J., Ayhan, M., Scrofani, S. D. B., Wilce, M. C. J., Guss, J. M., Tregloan, P. A., George, G. N., and Wedd, A. G. (1998) J. Am. Chem. Soc. 120, 4135-4150.

290. Agarwal, A., Li, D. L., and Cowan, J. A. (1996) J. Am. Chem. Soc. 118, 927-928.

291. Bentrop, D., Bertini, I., Capozzi, F., Dikiy, A., Eltis, L., and Luchinat, C. (1996) Biochemistry 35, 5928-5936.

292. Otwinowski, Z., and Minor, W. (1997) Methods Enzymol. 276, 307-326.

293. Navaza, J. (1994) Acta Cryst. A50, 157-163.

294. Parisini, E., Capozzi, F., Lubini, P., Lamzin, V., Luchinat, C., and Sheldrick, G. M. (1999) Acta Cryst. D55, 1773-1784.

295. Brunger, A. T., Adams, P. D., Clore, G. M., DeLano, W. L., Gros, P., Grosse- Kunstleve, R. W., Jinag, J., Kuszewski, J., Nilges, M., Pannu, N. S., Read, R. J., Rice, L. M., Simonson, T., and Warren, G. L. (1998) Acta Cryst. D54, 905-921.

296. Agarwal, A., Li, D. L., and Cowan, J. A. (1995) Proc. Natl. Acad. Sci. USA 92, 9440-9444.

297. Banci, L., Bertini, I., Dikiy, A., Kastrau, D. H. W, Luchinat, C., and Sompornpisut, P. (1995) Biochemistry 34, 206-219.

248 298. Carter, C. W. (1977) in Iron-Sulfur Proteins (Lovenberg, W., Ed.) pp 157-204, Academic Press, New York.

299. Long, T. V., and Loehr, T. M. (1970) J. Am. Chem. Soc. 92, 6384-6386.

300. Mansy, S., Wood, T. E., Sprowles, J. C., and Tobias, R. S. (1974) J. Am. Chem. Soc. 96, 1762-1770.

301. Backes, G., Mino, Y., Loehr, T. M., Meyer, T. E., Cusanovich, M. A., Sweeney, W. V., Adman, E. T., and Sanders-Loehr, J. (1991) J. Am. Chem. Soc. 113, 2055- 2064.

302. Czernuszewicz, R. S., Macor, K. A., Johnson, M. K., Gewirth, A., and Spiro, T. G. (1987) J. Am. Chem. Soc. 109, 7178-7187.

303. Bertini, I., Cowan, J. A., Luchinat, C., Natarajan, K., and Piccioli, M. (1997) Biochemistry 36, 9332-9339.

304. Soriano, A., and Cowan, J. A. (1996) Inorg. Chim. Acta 251, 285-290.

305. Perutz, M. F., and Mathews, F. S. (1966) J. Mol. Biol. 21, 199-202.

306. Breiter, D. R., Meyer, T. E., Rayment, I., and Holden, H. M. (1991) J. Biol. Chem. 266, 18660-18667.

307. Benning, M. M., Meyer, T. E., Rayment, I., and Holden, H. M. (1994) Biochemistry 33, 2476-2483.

308. Banci, L., Bertini, I., Eltis, L. D., Felli, I. C., Kastrau, D. H. W., Luchinat, C., Piccioli, M., Pierattelli, R., and Smith, M. (1994) Eur. J. Biochem. 225, 715-725.

309. Rayment, I., Wesenber, G., Meyer, T. E., Cusanovich, M. A., and Holden, H. M. (1992) J. Mol. Biol. 228, 672-686.

249 310. Babini, E., Bertini, I., Borsari, M., Capozzi, F., Dikiy, A., Eltis, L., and Luchinat, C. (1996) J. Am. Chem. Soc. 118, 75-80.

250