ABSTRACT ZHAO, JING. Regulatory Implications of Substrate

ABSTRACT

ZHAO, JING. Regulatory Implications of Substrate, Inhibitor and Axial Ligand Binding in Multifunctional Protein Dehaloperoxidase-Hemoglobin from Amphitrite ornata. (Under the direction of Dr. Stefan Franzen).

Dehaloperoxidase-hemoglobin (DHP), first isolated from the terebellid polychaete Amphitrite ornata, is a multifunctional protein that functions as an oxygen transporter, a peroxidase, a peroxygenase and an oxidase. DHP appears to possess a mechanism to incorporate multiple functions into a small protein (15.5 kDa). Although DHP contains eight α-helices and has a 3- over-3 helical bundle folding structures and functions as a hemoglobin in A. ornata, the peroxidase function was discovered in 1996 and then both peroxygenase and oxidase functions were observed in 2014. DHP’s multiple functions are regulated by binding of substrate, inhibitor and axial ligands. For example, DHP acts as a peroxidase when bound with native substrate 2,4,6-tribromophenol and functions as a peroxygenase when bound with 2,4- dibromophenol, but 4-bromophenol is an competitive inhibitor for both functions. DHP’s peroxidase function is also regulated by hydroquinone by a proton coupled electron transfer

(PCET) mechanism. The protein backbone dynamics of DHP revealed by NMR relaxation experiments and molecular dynamics (MD) simulations suggest that DHP has a rigid protein backbone but with a few residues that experience chemical exchange and slow conformational motions. The sulfur containing residues like cysteine and methionine revealed by protein dynamics may be related to the unique autoredcution behavior observed in DHP.

Regulatory Implications of Substrate, Inhibitor and Axial Ligand Binding in Multifunctional Protein Dehaloperoxidase-Hemoglobin from Amphitrite ornata

by Jing Zhao

A dissertation submitted to the Graduate Faculty of North Carolina State University in partial fulfillment of the requirements for the degree of Doctor of Philosophy

Chemistry

Raleigh, North Carolina

2016

APPROVED BY:

______Stefan Franzen Reza A. Ghiladi Committee Chair

______Gavin Williams Tatyana Smirnova

______Denis Pelletier Graduate School Representative

DEDICATION

I dedicate this thesis to all my family, friends and mentors.

With all their love, support and encouragement, I am able to keep pursuing my ideal career.

BIOGRAPHY

On January 9th, 1989, Jing Zhao was born to Xingya Zhao and Shufang Qi. He graduated from

Wuhan Foreign Language School on 2007 and was then recommended to Chu Kochen Honors

College, Zhejiang University. Immersed in a multi-discipline environment in college, Jing was inspired by the interdisciplinary of chemistry and completed his Bachelor of Science degree in

Chemistry on 2011 under Prof. Haoran Li’s supervision. He then decided to pursue his Ph.D. at North Carolina State University with Prof. Stefan Franzen with a focus on physical chemistry. During the five years of Ph.D. study, Jing has developed interests in applying spectroscopic methods in characterizing the structural and functional properties of dehaloperoxidase-hemoglobin and using statistical learning and visualization techniques to analyze and present variety of datasets.

iii

ACKNOWLEDGMENTS

First, I would like to express my sincere gratitude to my mentor, Dr. Stefan Franzen, for enormous help, continuous encouragement and valuable guidance, both in research and in personal life. I really appreciate Stefan created a flexible and creative research environment in the lab. He provided the freedom to explore your own research idea, at same time gave valuable advice and bridged the necessary technological gap by wide collaboration all over the world.

The way he’s doing research and teaching has set an excellent model for everybody in the lab.

I would also like to thank Dr. Reza Ghiladi who has always been a very wise and helpful collaborator. Reza has given me a lot of support and many insightful advice during my Ph.D. study.

Moreover, I would like to thank past and current members of Stefan’s group: Matthew

Thompson, Dustin Lockney, Junjie Zhao, Ruqi Wang, Shu Jiang, Misun Kang, Justin Morreto,

Peter Le, Hniang Khamh, Pallavi Singh. And members from Reza’s group: David Barrios,

Nikolette McCombs, Leiah Carey. They have provided a lot of help and assistance and I feel happy that we shared a pleasant and productive working experience together.

Lastly, I want to thank my family and all my friends for their love and support over these years. Because of them, I have had a wonderful life and learning experience at North

Carolina State University during these five years.

TABLE OF CONTENTS

LIST OF TABLES ...... x LIST OF FIGURES ...... xii CHAPTER 1 An Overview of Structures and Functions of Dehaloperoxidase-Hemoglobin ...... 1 1.1 Introduction ...... 2 1.2 References ...... 11 CHAPTER 2 Measurement of Internal Substrate Binding in Dehaloperoxidase-Hemoglobin by Competition with Heme-Fluoride Binding Equilibrium ...... 1 2.1 Abstract ...... 2 2.2 Introduction ...... 3 2.3 Material and Methods ...... 4 Materials ...... 4 Fluoride Titration Assays ...... 5 Resonance Raman Spectroscopy ...... 9 Density functional theory calculations...... 10 2.4 Results ...... 12 The fluoride titration and the competitive binding between DHP-F adduct and the internal substrate binding ...... 13 pH-dependent fluoride titration ...... 19 Fe-F bonding studied by density functional theory ...... 22 2.5 Discussion ...... 25 On the choice of fluoride as the ligand for a competitive binding assay ...... 25 On the competition between non-covalent substrate and covalent fluoride binding .. 26 The role of distal histidine His55 in H-bond interactions of DHP-F adduct ...... 28 Natural selection and selective internal binding of molecules in the distal pocket .... 31 2.6 Conclusion ...... 32 2.7 References ...... 34

CHAPTER 3 Distinct Enzyme-Substrate Interactions Revealed by Two Dimensional Kinetic Comparison between Dehaloperoxidase-Hemoglobin and Horseradish Peroxidase ...... 40 3.1 Abstract ...... 41 3.2 Introduction ...... 42 3.3 Material and Methods ...... 45 Materials ...... 45 Bi-substrate Ping-Pong Mechanism for DHP’s Peroxidase Function ...... 46 Bench-top Mixing Kinetic Assays ...... 49 Fluoride Titration Assays ...... 50 Data Analysis ...... 51 3.4 Results ...... 51 3.5 Discussion ...... 59 The Significance of External and Internal Substrate Binding Sites ...... 59 Kinetic Comparison between DHP and HRP: The Tortoise and the Hare ...... 60 3.6 References ...... 63 CHAPTER 4 Kinetic Study of the Inhibition Mechanism of DHPA by para-Halogenated Phenol ...... 67 4.1 Abstract ...... 68 4.2 Introduction ...... 69 4.3 Material and Methods ...... 72 Materials ...... 72 Transient-state kinetic assays ...... 73 Bench-top mixing kinetic assays ...... 73 Data analysis ...... 74 4.4 Results ...... 76

Stopped-flow kinetics of reaction between ferric DHP and H2O2 in the presence of 4- BP ...... 76 Bench-top mixing kinetics with substrate 2,4,6-TCP and inhibitor 4-BP ...... 79

van’t Hoff analysis of the inhibition constant Ki ...... 82 Bench-top mixing kinetics with substrate 2,4,6-TCP and inhibitor 4-CP ...... 83 Bench-top mixing kinetics with substrate 2,4,6-TCP and inhibitor 4-FP ...... 84

4.5 Discussion ...... 87 4.6 References ...... 94 CHAPTER 5 Regulatory Implications of Hydroquinone for the Multifunctional Enzyme Dehaloperoxidase-Hemoglobin from Amphitrite ornata ...... 97 5.1 Abstract ...... 98 5.2 Introduction ...... 99 5.3 Material and Methods ...... 102 Materials ...... 102 Bench-top Mixing Kinetic Assays ...... 103 Stopped-flow UV-visible Kinetic Assays ...... 103 Resonance Raman Spectroscopy ...... 104 Data Analysis ...... 105 5.4 Results ...... 106 The lag phase observed in the catalytic oxidation of 2,4,6-TCP ...... 106

The reversible reaction between reduction of ferric DHP A by H2Q and oxidation of oxyferrous DHP A by 1,4-BQ ...... 107 55 Distal Histidine His acts as a proton acceptor facilitating the oxidation of H2Q by a Proton Coupled Electron Transfer (PCET) mechanism ...... 111

DHP A catalyzed oxidation of H2Q in the presence of H2O2 ...... 115 5.5 Discussion ...... 118

The oxidation of H2Q by ferric DHP A in the absence of H2O2 by a PCET mechanism ...... 118

The oxidation of H2Q by DHP and the connection to the lag phase ...... 120 5.6 References ...... 124 CHAPTER 6 Structure and Backbone Dynamics of Dehaloperoxidse- Hemoglobin A: Insights from NMR Relaxation Spectroscopy and Molecular Dynamics Simulations ...... 127 6.1 Abstract ...... 128 6.2 Introduction ...... 129 6.3 Material and Methods ...... 132 Protein Expression, labelling, and purification ...... 132

vii

NMR Sample preparation and heteronuclear relaxation measurement ...... 133 Model-free analysis using relax ...... 134 Molecular dynamic simulations and trajectory analysis ...... 135 Tunnel Calculation ...... 136 6.4 Results ...... 137 Reduced Spectral Density Mapping and Consistency Testing ...... 138

Geometric Interpretation of J(ωN) and J(0.87ωH) via Lapari-Szabo Mapping ...... 141 Model-Free Analysis using relax ...... 144 Molecular Dynamics Simulation and Computed Order Parameter ...... 148 Tunnels in the protein interior...... 149 6.5 Discussion ...... 150 DHP is Primarily a Monomer in Solution...... 150 The Role of Tunnels in Function Switching ...... 153 6.6 Conclusion ...... 157 6.7 References ...... 159 CHAPTER 7 Mechanism of Autoreduction of Hemoglobin – A Case Study of Dehaloperoxidase-Hemoglobin ...... 164 7.1 Abstract ...... 165 7.2 Introduction ...... 166 7.3 Material and Methods ...... 169 Materials ...... 169 Site-Directed Mutagenesis and Protein Purification ...... 169 Autoreduction Kinetics Measurement ...... 171 Autoreduction Kinetic Model ...... 172 7.4 Results ...... 175 7.5 Discussion ...... 179 7.6 References ...... 181

Appendix A ...... 185 A1. Derivation of the Apparent Binding Equilibrium Constant ...... 185

Appendix B ...... 190 B1. Derivation of the initial rate equations...... 190

viii

B2. Kinetics of DHP B for substrate 2,4,6-TCP and 2,4,6-TBP...... 194 B3. Kinetics and mechanism of DHP A and B catalyzed oxidation of ABTS...... 195 B4. Kinetics of HRP for substrate 2,4,6-TCP and 2,4,6-TBP...... 196

Appendix C ...... 198 C1. Singluar Value Decomposition (SVD) ...... 198 C2. Global-fitting analysis and reconstruction of the spectra ...... 199

Appendix D ...... 208

D1. Duration of lag phase is proportional to the concentration of H2Q...... 208 D2. Singular Value Decomposition (SVD) and Global-fitting Analysis of Time- Resolved Spectra ...... 208 55 D3. pKa of Distal Histidine His in Ferric and Oxy-ferrous DHP A Determined by Resonance Raman Spectroscopy ...... 213

D4. pH-dependent initial rate of DHP catalyzed oxidation reaction of H2Q...... 218

D5. Stopped-flow UV-visible spectra of reaction between DHP A and H2Q in the presence of H2O2...... 218 D6. Inhibition Kinetic Study of 1,4-BQ ...... 219

Appendix E ...... 222 E1. Spectral Density Functions of Model-Free Analysis ...... 222 E2. Cyanide Titration and Determine Its Binding Affinity in DHPA ...... 223 E3. Consistency Test of Two Datasets (500 MHz and 700 MHz) ...... 224 E4. Color Map of Squared Generalized Order Parameters ...... 225

Appendix F ...... 233

LIST OF TABLES

f Table 2.1 The fluoride dissociation constant Kd of several hemoproteins...... 14

app Table 2.2 The apparent fluoride dissociation constant Kd of DHP A in the presence of internal binding molecules ...... 16

f app Table 2.1 The fluoride dissociation constants Kd and Kd in different pH...... 19

Table 2.4 Wave number of the Fe-F stretching mode calculated using DFT methods. All structures contained heme and fluoride with various groups on the proximal and distal side as given in columns 1 and 2...... 24

Table 2.5 Solution and binding energies calculated using DFT methods including a COSMO dielectric continuum model...... 25

Table 3.2 Kinetic Parameters obtained from 2D Michaelis-Menten kinetics for DHP A ..... 54

Table 3.2 Kinetic Parameters obtained from 2D Michaelis-Menten kinetics for DHP B ...... 54

Table 3.3 Kinetic Parameters obtained from 2D Michaelis-Menten kinetics for HRP ...... 54

app Table 3.4 Apparent Fluoride Dissociation Constant Kd of DHPA and HRP with or without Substrates ...... 57

Table 4.3 Kinetic parameters obtained from global-fitting of evolutionary time-course ...... 77

Table 4.2 Michaelis-Menten parameters from bench-top mixing 4-BP inhibition kinetic assay ...... 82

Table 4.3 Michaelis-Menten parameters from bench-top mixing 4-CP inhibition kinetic assay ...... 83

Table 4.4 Michaelis-Menten parameters of substrate 4-FP ...... 85

Table 4.5 Michaelis-Menten parameters from bench-top mixing 4-FP inhibition kinetic assay ...... 87

Table 5.4 Michaelis-Menten Kinetic Parameters for Oxidation of H2Q by DHP A...... 105

Table 5.5 Michaelis-Menten Kinetic Parameters of DHP A Substrates ...... 117

1 15 Table 6.6 Average values of R1, R2 and { H}- N NOE relaxation data ...... 138

Table 6.7 Average of J(0), J(ωN) and J(0.87ωH) values ...... 140

Table 6.8 Comparison between different Hemoglobins ...... 153

Table 7.1 DNA Sequences of Primers ...... 170

Table 7.2 Apparent Autoreduction rate kobs of different proteins ...... 178

Table 7.3 Reduction Potential of Hemoglobin and Myoglobin ...... 179

Table A9 Heat of Fluoride Solution ...... 189

Table D10 Michealis-Menten parameters of inhibition study of 1,4-BQ...... 220

1 15 Table E1. DHPA Relaxation data of R1 , R2 and { H}- N NOE at 500 MHz and 700 MHz ...... 226

Table E2. DHPA Reduced spectral density J(0), J(0.87ωH) and J(ωN) at 500 MHz and 700 MHz ...... 228

Table E3. DHPA Model-free analysis dataset ...... 230

Table E4. Generalized order parameter S2 from NMR and MD simulations ...... 232

LIST OF FIGURES

Figure 1.1 Chemical structure of iron protoporphyrin IX, also known as heme b, with four edges (, , , ) labelled...... 2

Figure 1.2 Living Amphitrite ornata ...... 3

Figure 1.3 DHP catalyzed oxidative dehalogenation of trihalogenated phenol (TXP) in the presence of cosubstrate H2O2...... 4

Figure 1.4 DHP catalyzed peroxygenase reaction of 2,4-dihalogenated phenol and 5-bromoindole in the presence of cosubstrate H2O2...... 4

Figure 1.5 Overlaid structures of DHP A and SWMb...... 5

Figure 1.6 Protein sequences alignment of DHP A and B from Amphitrite ornata and intracellular hemoglobin from Alvinella pompejana...... 6

Figure 1.7 X-ray crystal structures overlay of three different substrates, 2,4,6-TBP, 2,4-DBP and 4-BP bound internally in the distal pocket of DHP, along with Xe binding site labelled in purple sphere. (a) Sideview (b) Topview...... 7

Figure 1.8 Schematic illustration of DHP’s catalytic cycle and reduction pathways involve with substrate, cosubstrates, inhibitor, regulator and axial ligand binding...... 9

Figure 2.1 Illustration of two modes of internal binding of substrate 2,4,6-TBP (a-site) and inhibitor 4-BP (b-site) in the distal pocket of DHP A with fluoride coordinated to the heme iron obtained from an overlay of the coordinate files from two X-ray crystal structures PDB entries 3LB2 and 4HF6, respectively. 7-9 ...... 6

Figure 2.2 UV-vis spectra of fluoride titration of WT DHPA and DHP A H55D mutant in the absence or in the presence of substrate. (a) Fluoride titration of WT DHPA (100 µM) in the 100 mM KPi buffer, pH 7.0. (b) Fluoride binding curve extracted from spectra data of figure 2.1a using SVD. (c) Fluoride titration of WT DHPA (50 µM) in the presence of 1 mM 2,4,6- TCP in the 100 mM KPi buffer, pH 7.0. (d) Fluoride titration of DHP A H55D mutant (50 µM) in the 100 mM KPi buffer, pH 7.0, the NaF concentration is up to 0.3 M...... 13

Figure 2.3 Resonance Raman spectra of WT ferric DHP A at pH 7.0 (red), DHP-F at pH 7.0 (blue), DHP-F at pH 5.0 (orange), DHP-F at pD 7.0 (purple), in the high frequency region. 20

Figure 2.4 Resonance Raman spectra of WT ferric DHP A at pH 7.0 (red), DHP-F at pH 7.0 (blue), DHP-F at pH 5.0 (orange), DHP-F at pD 7.0 (purple), in the low frequency region. . 20

xii

Figure 2.5 Resonance Raman spectra of DHP-F titrated with 0 mM 2,4-DBP (blue), 0.1 mM 2,4-DBP (red), 0.2 mM 2,4-DBP (purple) and 1.0 mM 2,4-DBP (black) in the low frequency region...... 22

Figure 2.6 Stick figures of one of the series of structures used for model DFT calculations. The models shown are (A) proximal-His-heme-F, (B) proximal-His-heme-F-(Ile), (C) proximal-His-heme-F-(H2O), (D) proximal-His-heme-F-(His)...... 22

Figure 3.2 Kinetics of DHPA (2.4 M) catalyzed oxidation of 2,4,6-TCP or 2,4,6-TBP in the presence of H2O2. (a) 2,4,6-TCP dimension, (b) H2O2 dimension (with 2,4,6-TCP), (c) 2,4,6- TBP dimension, (d) H2O2 dimension (with 2,4,6-TBP) in the 100 mM KPi buffer at pH 7.0.52

Figure 3.3 3D plot of kinetics of DHP (2.4 M) catalyzed oxidation of 2,4,6-TXP (X = Cl, Br) in the presence of H2O2. (a) DHP A with 2,4,6-TCP; (b) DHP A with 2,4,6-TBP; (c) DHP B with 2,4,6-TCP; (d) DHP B with 2,4,6-TBP...... 53

Figure 3.4 3D plot of kinetics of DHP (2.4 M) catalyzed oxidation of ABTS in the presence of H2O2. (a) DHP A with ABTS; (b) DHP B with ABTS...... 55

Figure 3.5 UV-vis spectra of fluoride titration of WT DHPA (a) in the absence or (b) in the presence of 1mM 2,4,6-TBP and HRP (c) in the absence or (d) in the presence of 1mM 2,4,6- TBP in the 100 mM KPi buffer at pH 7.0...... 56

Figure 3.6 The fluoride titration profile of DHP A and HRP in the presence of substrate: 2,4,6-TCP, 2,4,6-TBP and ABTS...... 58

Figure 4.1 The structure of DHP in presence of 4-BP as determined by X-ray crystallography (PDB 3LB2)...... 70

Figure 4.2 Calculated spectra from SVD analysis of the time-resolved UV-visble spectra. The measurements were conducted at [DHP]=10 µM, [H2O2]=100 µM in the 100 mM KPi buffer pH 7.0 with (a) 0 µM 4-BP; (b) 10 µM 4-BP; (c) 20 µM 4-BP; (d) 100 µM 4-BP. In the presence of 4-BP, two kinetic pathways distinguished by 4-BP bound and unbound form of DHP are observed. Thus, (b1),(c1) and (d1) represents the 4-BP unbound pathway, (b2),(c2) and (d2) represents the 4-BP bound pathway. As for (a),(b1),(c1) and (d1),the red curve is ferric DHP, blue curve is Compound ES and purple curve is Compound RH. As for (b2),(c2) and (d2), the red curve is ferric DHP bound with 4-BP, blue curve is Compound iOX and purple curve is Compound iRH...... 76

xiii

Figure 4.3 Bench-top mixing kinetic analysis of DHP catalyzed TCP oxidation reaction inhibited by 4-BP at different temperatures (a) 25 oC( b) 20 oC (c) 15 oC (d) 10 oC. Kinetic assay conditions were ferric DHP = 2.4 µm, H2O2 = 1200 µm in 100 mM KPi buffer, pH 7.0...... 81

Figure 4.4 van’t Hoff plot of of ln(Ki) vs 1/T for inhibitor 4-BP binding to ferric DHP. The Ki value was determined at pH = 7.0 in 100mM KPi buffer. 2,4,6-TCP was used as the substrate...... 82

Figure 4.5 Bench-top mixing kinetic analysis of DHP catalyzed TCP oxidation reaction o inhibited by 4-CP at 25 C. Kinetic assay conditions were ferric DHP = 2.4 µm, H2O2 = 1200 µm in 100 mM KPi buffer, pH 7.0...... 83

Figure 4.6 DHP A (2.4 µM) catalyzed oxidation of 4-FP (500 µM) in the presence of H2O2 (1200 µM) at 25 OC, forming the product 1,4-benzoquinone...... 84

Figure 4.7 (a) Temperature-dependent DHP A (2.4 µM) catalyzed oxidation of 4-FP in the presence of H2O2 (1200 µM) and (b) Arrhenius plot of kcat/Km vs 1/T...... 85

Figure 4.8 Bench-top mixing kinetic analysis of DHP catalyzed TCP oxidation reaction o inhibited by 4-FP at 25 C. Kinetic assay conditions were ferric DHP = 2.4 µm, H2O2 = 1200 µm in 100 mM KPi buffer, pH 7.0...... 86

Figure 4.9 Proposed DHP catalytic cycle and inhibition mechanism in presence of inhibitor 4-BP. The form of the heme Fe in compound iOX and iRH is not known and therefore marked with a question mark. Nonetheless, since the heme Soret band is altered in these species there is strong reason to believe that there have been changes that involve the Fe. .. 89

Figure 5.1 The lag phase of DHP A catalyzed oxidation of 2,4,6-TCP in the presence of H2Q. The blue time course represents the turnover of H2Q that forms 1,4-BQ, the red course represents the turnover of 2,4,6-TCP that yields 2,6-DCQ...... 106

Figure 5.2 Reconstructed spectra of the reaction between ferric DHP A (red) and H2Q that forms the oxyferrous DHP A (blue). The inset shows an expansion of the Q-band region. 107

Figure 5.3 Plot of the kobs vs [H2Q] for the reduction of ferric DHP A(5 µM) by H2Q in 100 mM KPi buffer, pH 7.0, 298 K...... 108

Figure 5.4 The inhibitor 4-BP inhibits reduction of ferric DHP A by H2Q. The reaction mixture consists of 5 µM Ferric DHP A reacting with 500 µM H2Q in the presence of 0, 50, 250, 500 µM 4-BP in 100 mM KPi buffer, pH 7.0 at 298K...... 109

Figure 5.5 Eyring plot for the oxidation of H2Q by ferric DHP A (5 µM) in 100 mM KPi buffer, pH 7.0...... 111

xiv

Figure 5.6 Resonance Raman spectra of metaquo DHP A as a function of pH in the high frequency region. A scheme describing the relevant equilibra for ferric DHP A is shown in the figure...... 112

Figure 5.7 Resonance Raman spectra of oxyferrous DHP A as a function of pH in the high frequency region. A photoexcitation reaction scheme for deoxyferrous DHP A is shown in Figure...... 113

Figure 5.8 Time resolved spectrum of DHP catalytic oxidation of H2Q in the presence of H2O2 (From red to purple). The subwindow shows the scale up of Q-band region. The kinetic assay condition are 5 µM DHP reacting with 500 µM [H2Q] in the presence of 1200 µM H2O2 in100 mM KPi buffer at 298K...... 116

Figure 5.9 Michaelis-Menten kinetics of catalytic oxidation of H2Q by DHP A. The kinetic assay were conducted using 2.4 µM DHP A react with varying concentrations of H2Q in the presence of 1200 µM H2O2 in 100 mM KPi buffer, pH 7.0 at 303K (red), 298K (purple), 293 K (blue), 288 K (black)...... 117

1 15 Figure 6.1 NMR relaxation data { H}- N NOE, R1, R2 and R2/ R1 that measured at 500 MHz (red) and 700 MHz (green). The α-helices are shown and labeled at the bottom (purple) and corresponding α-helical regions are shaded as column inside the plot...... 138

Figure 6.2 Application of Lipari-Szabo mapping method for analyzing the 500 MHz reduced spectral density mapping data of met-cyano DHP A. (Red) the rigid tumbling curve; (Green) the observed reduced spectral density values; (Blue) the rigid tumbling point at overall correlation time τm = 9.49 ns...... 143

Figure 6.3 (A) Comparison of generalized order parameter S2 obtained from NMR (blue) and MD (orange) for each residue in DHPA; (B) The difference histogram of generalized order parameters S2 between NMR and MD; (C) Histogram of relaxation attribute to chemical exchange Rex; (D) Histogram of effective correlation time of internal motion τe (ps) (blue, left axis) and slow effective correlation time τs (ns) (red, right axis) in two different time-scales...... 145

Figure 6.4 Color mapping of secondary structural information and NMR model-free analysis results on the met-cyano DHP A structure. (A) Color mapped eight α-helices of DHPA with calculated tunnel structures and the two xenon binding sites. (B) Color mapped generalized order parameter S2 on DHP A structure. (C) Color mapped model selection results of model- free analysis on the DHP A structure...... 147

Figure 6.5 Dimer interface of DHP observed in X-ray crystallography...... 152

Figure 7.1 DHPA monomer – dimer equilibrium with sulfuring containing residues labeled on the secondary structure...... 169

Figure 7.2 (a) Monomer autoreduction mechanism; (b) Cooperative dimerized autoreduction mechanism...... 173

Figure 7.3 The time-resolved UV-Vis spectra of CO-driven autoreduction reaction of (a) WT DHPA and (c) DHPA H55V mutant and their corresponding time course (c) WT DHPA, (d) DHPA H55V of the reaction...... 175

Figure 7.4 (a) The DHP protein concentration dependent of apparent autoreduction rate kobs. (b) The pH-dependent autoreduction rate kobs...... 177

Figure A1. Time-resolved UV-vis spectra of 10 µM DHP A catalyzed oxidation of 500 µM 2,4-DBP in the presence of 1200 µM H2O2 in at 100 mM KPi buffer at pH 7.0...... 188

Figure A2. UV-vis spectra of 50 µM DHP A binding with 5 mM 2-BP (black),3-BP (blue) and 4-BP (red) respectively in 100 mM KPi buffer at pH 7.0, respectively...... 188

Figure A3. UV-vis spectra of fluoride titration of WT DHPA (50 µM) in the presence of 1 mM 2,4-DBP in the 100 mM KPi buffer, pH 7.0...... 189

Figure B4. Kinetics of DHP B (2.4 M) catalyzed oxidation of 2,4,6-TCP or 2,4,6-TBP in the presence of H2O2. (a) 2,4,6-TCP dimension, (b) H2O2 dimension (with 2,4,6-TCP), (c) 2,4,6-TBP dimension, (d) H2O2 dimension (with 2,4,6-TBP) in the 100 mM KPi buffer at pH 7.0...... 194

Figure B5. Kinetics of DHP (2.4 M) catalyzed oxidation of ABTS in the presence of H2O2. (a) DHPA in ABTS dimension, (b) DHPA in H2O2 dimension, (c) DHPB in ABTS dimension, (d)DHPB in H2O2 dimension in the 100 mM KPi buffer at pH 7.0...... 195

Figure B3. Kinetics of HRPC (0.2 M) catalyzed oxidation of 2,4,6-TCP or 2,4,6-TBP in the presence of H2O2. (a) 2,4,6-TCP dimension, (b) H2O2 dimension (with 2,4,6-TCP), (c) 2,4,6- TBP dimension, (d) H2O2 dimension (with 2,4,6-TBP) in the 100 mM KPi buffer at pH 7.0...... 196

Figure B5. The Denaturation of HRP (0.2 µM) during the catalytic reaction in the presence of 900 µM 2,4,6-TBP and 300 µM H2O2...... 197

Figure C1. The first three orthonormal basis spectra in U matrix obtained from SVD analysis of the data matrix A(λ,t). The stopped-flow measurements were conducted at [DHP]=10 µM, [H2O2]=100 µM in the 100 mM KPi buffer pH 7.0 with (a) 0 µM 4-BP...... 201

Figure C2. The first three rows in VT matrix obtained from SVD analysis of the data matrix A(λ,t). The three vectors were global-fitted to the biexponential function. The stopped-flow measurements were conducted at [DHP]=10 µM, [H2O2]=100 µM in the 100 mM KPi buffer pH 7.0 with (a) 0 µM 4-BP...... 202

xvi

Figure C3. The b-spectra obtained from SVD analysis of the data matrix A(λ,t). The stopped- flow measurements were conducted at [DHP]=10 µM, [H2O2]=100 µM in the 100 mM KPi buffer pH 7.0 with (a) 0 µM 4-BP...... 202

Figure C5. The first three rows in VT matrix obtained from SVD analysis of the data matrix A(λ,t). The three vectors were global-fitted to the tetraexponential function. The stopped- flow measurements were conducted at [DHP]=10 µM, [H2O2]=100 µM in the 100 mM KPi buffer pH 7.0 with (b) 10 µM 4-BP, (c) 20 µM 4-BP, (d) 100 µM 4-BP...... 206

Figure C6. The b-spectra obtained from SVD analysis of the data matrix A(λ,t). The stopped- flow measurements were conducted at [DHP]=10 µM, [H2O2]=100 µM in the 100 mM KPi buffer pH 7.0 with (b) 10 µM 4-BP, (c) 20 µM 4-BP, (d) 100 µM 4-BP...... 207

Figure C7. UV-Vis spectra of preformed Compound RH titrated with 4-BP. The bench-top UV-Visible measurement was conducted at [DHP]=10 µM, [H2O2]=100 µM in the 100 mM KPi buffer pH 7.0 with 4-BP titrated in (final concentration 100 µM) after Compound RH has formed...... 207

Figure D1. Correlation of the lag phase with the concentration of H2Q. 2.4 µM DHP A catalyzed oxidation of TXP (X=Cl, Br) with 1200 µM H2O2 in the presence of H2Q in 100 mM KPi buffer, pH 7.0, 298 K...... 208

Figure D2. Time resolved UV-Visible spectrum for the reduction of ferric DHP A (5 µM) by H2Q (100 µM) in the 100 mM KPi buffer, pH 7.0 at the 60s time intervals...... 210

Figure D3. Single exponential fitting of two v vectors of VT matrix of the reaction between ferric DHP A and H2Q...... 210

Figure D4. pH dependence of second-order rate constant k1, oxidation of H2Q by ferric DHP A (5 µM) in 100 mM KPi buffer, pH 7.0 at 298 K...... 211

Figure D5. Time resolved UV-Visible spectrum for the oxidation of oxy-ferrous DHP A (5 µM) by 1,4-BQ (300 µM) in the 100 mM KPi buffer, pH 7.0 at the 60s time intervals...... 211

Figure D6. Reconstructed spectra of the reaction between Oxy-ferrous DHP A and 1,4-BQ...... 212

Figure D7. Single exponential fitting of two v vectors of VT matrix of the reaction between oxy-ferrous DHP A and 1,4-BQ...... 212

Figure D8. Plot of the kobs vs [1,4-BQ] for the oxidation of oxy-ferrous DHP A (5 µM) by 1,4-BQ in 100 mM KPi buffer, pH 7.0 at 298 K...... 212

Figure D9. pH-dependent ferric DHP A resonance Raman spectra...... 213

xvii

T Figure D10. The V2 vector of V matrix were fitted into a two pKa equilibrium model...... 215

Figure D11. pH-dependent oxy-ferrous DHP A Resonance Raman spectra...... 215

T Figure D12. The V2 vector of V matrix were fitted into equilibrium model combined with photoexciteautoxidation conversion ...... 217

Figure D13. pH dependence of initial rate Vo for the oxidation of H2Q (400 µM ) catalyzed by DHP A (2.4 µM) in the presence of H2O2 ( 1200 µM ) in 100 mM KPi buffer at 298K...... 218

Figure D14. Stopped-flow UV-visible spectra of reaction between ferric DHP A (5µM) preincubated with H2Q (55 µM) and 10-fold excess H2O2 (500 µM) in 100 mM KPi buffer, pH 7.0 at a time interval of ~83s...... 219

Figure D15. Calculated spectra of reaction species at t=0 (red); τ1/2=0.126s (Blue) and τ1/2=71.2s (Black)...... 219

Figure D16. Steady-state kinetic analysis of DHP A catalyzed TCP oxidation reaction inhibited by 1,4-BQ at 25℃...... 220

app Figure D17. Determination of Ki by linear fit of Km against the concentration of 1,4-BQ...... 221

Figure E1. (A) UV-vis spectra of DHPA cyanide titration experiments and (B) DHPA cyanide binding curve...... 223

Figure E2. J(0) consistency test of two datasets acquired at 700 MHz and 500 MHz. (a) Correlation plot of J(0); (b) Histogram distribution of the J(0) ratios...... 224

Figure E3. Computed generalized order parameters S2 vs residue number from 5 parallel MD simulations...... 225

Figure E4. Color mapped generalized order parameter S2 from NMR on DHPA structure. 225

Figure E5. Color mapped generalized order parameter S2 from MD simulations on DHPA structure...... 225

Figure F1. The time-resolved UV-Vis spectra of CO-driven autoreduction reaction of WT DHPA at pH 8.0 and corresponding time course of the reaction...... 233

Figure F2. The time-resolved UV-Vis spectra of CO-driven autoreduction reaction of DHPA C73S, M63L&M64L mutant and SWMb with their corresponding time course of the reaction...... 234

xviii

CHAPTER 1

An Overview of Structures and Functions of

Dehaloperoxidase-Hemoglobin

1.1 Introduction

Hemoglobin is an ancient protein with an ancestry that dates back 1.8 billion years.1 It is also one of the most abundant and widespread proteins in nature. It is found in organisms spanning all three kingdoms animal, plant and bacteria.1 It is well represented within the kingdoms as well. Hemoglobin, like many cytochromes and other heme proteins, contains an iron protoporphyrin IX molecule (Figure 1.1) known as heme cofactor that coordinated to the protein peptide chain.

Figure 1.1 Chemical structure of iron protoporphyrin IX, also known as heme b, with four edges (, , , ) labelled.

The primary function of hemoglobin is oxygen storage and transport.2 The heme iron is in the ferrous state (Fe2+) in order to reversibly bind to diatomic oxygen. The mammalian hemoglobin typically forms a heterotetramer with -globin and -globin subunits. But hemoglobin from more primitive species can be monomeric. For example, hemoglobin from

Glycera Dibranchiata a polychaete annelid, is monomeric.3 The major difference between multimeric hemoglobin and monomeric myoglobin is that hemoglobin has the exquisite cooperativity due to the change of its quaternary structure in response to oxygen binding.4,5

The conformational change in quaternary structure is triggered simply by the change of coordinate state of heme iron from five-coordinate (deoxy) to six-coordinate (oxyferrous) state.

In the commonly accepted point of view, a division is made for hemoproteins between reversible and irreversible binding functions. Hemoglobin or myoglobin are responsible for diatomic gas molecule (O2) storage and transport or functions as gas sensor that utilize signal molecules (CO, NO) to regulate downstream signal transduction.6 In contrast, cytochromes is mainly responsible for reversible electron-transfer function and other heme proteins, like peroxidase, peroxygenase, monooxygenase, oxidase etc conduct irreversible enzymatic oxidation.7,8 The discovery of a unique hemoglobin, known as dehaloperoxidase-hemoglobin

(DHP) challenge such opinion point of view, because the accumulated evidence strongly suggests that DHP is a multi-functional protein. It not only functions as an oxygen transporter, but also possesses multiple enzymatic functions. While the biological significance of these functions is still a topic for investigation many physical measurements and structures support the multifunctional capability of DHP.

Figure 1.2 Living Amphitrite ornata

Dehaloperoxidase-hemoglobin (DHP) isolated from the marine worm Amphitrite ornata (Figure 1.2), is the first hemoglobin identified with a biologically relevant peroxidase function.9 The oxidation of 2,4,6-tribromophenol by DHP is considered to be relevant because of the presence of this toxic compounds in benthic ecosystems. A. ornata co-inhabits shallow marine estuaries and muddy coastal regions with other polychaete worms, such as Notomatus lobatus that secret highly toxic halogenated aromatic compounds.10 One of the postulation that

A.ornata can survive in such toxic environment is because it possesses an intracellular coelomic hemoglobin, DHP, which catalytically oxidize and degrade halogenated phenols as shown in Figure 1.3.9,11

Figure 1.3 DHP catalyzed oxidative dehalogenation of trihalogenated phenol (TXP) in the presence of cosubstrate H2O2. Recently, DHP was found to possess peroxygenase and oxidase functions toward the

2,4-dihalogenated phenol and halogenated indole derivatives (Figure 1.4).12 In contrast to peroxidase reaction, peroxygenase reaction involves oxygen insertion reaction thus requires substrate internal binding in the distal pocket near the heme Fe active site.

Figure 1.4 DHP catalyzed peroxygenase reaction of 2,4-dihalogenated phenol and 5-bromo-indole in the presence of cosubstrate H2O2.

DHP has a similar folding structure compare to that of mammalian myoglobin and hemoglobin  ,  unit even though the sequence homology is only about 20%.13 The overlay structure of monomeric DHP A (PDB: 2QFK) and sperm whale myoglobin (SWMb) (PDB:

1JP6) clearly shows such resemblance in the three-dimensional structure. Both of them contain eight α-helices and share a 3-over-3 helical bundle folding structures. However, The heme of

DHP is buried more deeply than that of SWMb by circa 1.5 Å.13 Unlike mammalian hemoglobin that forms heterotetramer with a dynamic quaternary structure, DHP remains approximately 90% monomeric in solution.14 However, DHP is observed to be a dimer in the

X-ray crystallography, though its dimer interface is relatively small compared to other hemoglobins.15

DHP A SWMb

Figure 1.5 Overlaid structures of DHP A and SWMb. Two isoforms of DHP have been identified and named as DHP A and DHP B. They are distinguished by 5 amino acids: I9L, R32K, Y34N, N82S and S91G out of 137 residues in the entire sequence.16 The overall protein folding are almost the same between these two isoforms but DHP B has about 3-fold greater catalytic efficiency (kcat/Km) compared to that of

DHP A.17 The most similar protein sequence found by protein BLAST18 is an intracellular hemoglobin from Alvinella pompejana, which is also an annelid polychaete. However, unlike

A.ornata that lives in the estuary mud flats, A.pompejana is an extremophile found only at hydrothermal vents in deep sea of pacific ocean and is identified as one of the most heat- tolerant animals on earth.19 According to the alignment result, about 30% residues are the same between A. ornata and A. pompejana, more than 60% are similar. It is noteworthy that A. pompejana only contains 4 containing residues (2 cysteines and 2 methionines) despite the fact that it lives in a sulfur rich environment at hydrothermal vents.

Figure 1.6 Protein sequences alignment of DHP A and B from Amphitrite ornata and intracellular hemoglobin from Alvinella pompejana, generated by using software ESPript. Due to the multi-functional property of DHP, it has been a long term goal to search for its multiple substrate binding sites based on both structure and spectroscopy properties. Several spectroscopic methods have been applied to characterize the substrate and inhibitor binding in

DHP. Resonance Raman spectroscopy shows that when DHP binds with para-halogenated phenol, the axial water ligand is displaced which results in a 5-coordinated high spin (5cHS) heme Fe. On the contrary, DHP still favors 6cHS when titrated with 2,4,6-trihalogenated phenol.20 Fourier transform infrared spectroscopy (FTIR) has also been used to characterize interactions between substrate and bound CO. These measurements are complemented by

studies of photo-dissociation of the CO molecule by temperature derivative spectroscopy and flash photolysis to give further proof that 2,4,6-TBP, 4-BP and other molecules bind in the distal pocket and modulate the binding of diatomic ligands to the heme Fe.21 To add to the structural picture, an efficient and quantitative method to determine the relative binding affinity of different internal binding substrates was developed by using the competitive fluoride ion binding in UV-Vis titration experiments.22 By applying this strategy, we find that 2,4- dibromophenol (2,4,-DBP) has the tightest binding affinity in the distal pocket among all the known substrate or inhibitor so far. The 1H-15N heteronuclear single quantum coherence

(HSQC) NMR spectroscopy results indicated that the binding of 4-BP and 2,4,6-TCP induced chemical shift deviations at different regions. 4-BP primarily perturbs the chemical shift of

Phe60, Phe35, Lys99 and Val59, residues in the distal pocket near the inhibitor binding site.

Whereas 2,4,6-TCP induces chemical shift deviation on Arg122, His55, Ser129, Leu76, Tyr34 and

Leu100, residues that both locates at distal pocket and the dimer interface.23

2,4,6-TBP 2,4-DBP Xe 2,4,6-TBP Binding Site 2,4-DBP 4-BP His55

Xe Binding Site

4-BP

His89 His55 a b

Figure 2.7 X-ray crystal structures overlay of three different substrates, 2,4,6-TBP, 2,4-DBP and 4-BP bound internally in the distal pocket of DHP, along with Xe binding site labelled in purple sphere. (a) Sideview (b) Topview.

The various edges of the heme are in realtion to different types of substrate binding in heme proteins. -edge is the back of the distal pocket and is the most buried edge, -edge is the internal substrate binding site for benzhydroxamic acid observed in horseradish peroxidase

(HRP) and -edge is closest to the distal histidine and the exit of the distal pocket where ascorbate peroxidase bounds its substrate. Recently, the X-ray crystal structure of DHP bound with 2,4-DBP has been determined (unpublished result). So now we have structures with a series of mono-, di-, tri-bromophenol bound with DHP (Figure 1.7).24,25 Apparently, the three molecules adopt three different binding sites in the distal pocket. 2,4,6-TBP is buried deeply in the globin above the α-edge of the heme; whereas 4-BP is closer to the exit channel on the

-edge of the heme, which is near the distal histidine His55with its aromatic ring perpendicular to the heme plane; 2,4-DBP has a binding site between 2,4,6-TBP and 4-BP and a tilted aromatic plane relative to the heme plane. However, it is noteworthy that 2,4,6-TBP, 2,4-DBP and 4-BP each positions one of the bulky Br atoms into the Xe binding site above the heme α- edge.

From Figure 1.7 we can see that when 4-BP binds, the steric hindrance pushes the distal

55 histidine His to adopt the solvent exposed open conformation and the ligated H2O is displaced due to the competition. When 2,4,6-TBP binds, distal histidine His55 is still flexible and able to adopt both conformations (Figure 2.7a). 2,4-DBP binds with its phenolic hydroxyl group within hydrogen bonding distance axial ligands that bind to the heme Fe. The distal hisitidine may exert a strong influence on axial ligand binding. Recently, two 2,4,6-TCP binding sites have also been observed in the distal pocket for the Y34N, Y34N/S91G and L100F mutants of

DHP A.26 The deep internal 2,4,6-TCP binding site is very similar to the 2,4,6-TBP binding

site, but with a slight rotation of the aromatic ring. The second internal binding site of 2,4,6-

TCP was observed for the first time. 2,4,6-TCP in this second site adopts an orientation similar to 4-BP, with its aromatic ring perpendicular to the heme plane and its hydroxyl group pointing towards the exit of distal pocket. In conclusion, these observed three binding sites have reconciled previous spectroscopic and enzymatic kinetic observations for these three different molecules. 2,4,6-TBP is a peroxidase substrate that DHP catalytically oxidize; 2,4-DBP is a peroxygenase substrate which undergoes an oxygen insertion reaction; 4-BP serves as a competitive inhibitor for peroxidase function, which may be further shown to regulate other functions in DHP as well. Therefore, DHP serves as an excellent example to study protein’s structure-function relationship and the role of substrate, inhibitor, and other ligand binding in regulating multi-functional proteins.

Figure 3.8 Schematic illustration of DHP’s catalytic cycle and reduction pathways involve with substrate, cosubstrates, inhibitor, regulator and axial ligand binding.

Figure 1.8 summarizes DHP’s catalytic cycle, reduction pathways and different ligation states that involves different substrates, cosubstrates, inhibitor, regulator and axial ligand binding. The complexity of DHP is becoming clear, and yet these are just the parts of the mechanism that we have discovered and understood so far. However, we also take advantage of the inherited complexity in this system that involves substrate, cosubstrate, inhibitor and axial ligand. By studying their interactions in various systems, including catalytic cycles and regulatory pathways, we get to know more details about the systems per se.

In chapter 2, we will introduce a strategy to rapidly screen internal binding substrates or inhibitors in the distal pocket of DHP by measuring the competitive binding equilibrium between fluoride ion ligand and these large aromatic molecules. In chapter 3, we will apply a

2-dimensional kinetic model to study DHP’s peroxidase catalytic reactions that involves both substrate and cosubstrate. In chapter 4, we will study the role of competitive binding inhibitor

4-BP in DHP’s catalytic cycle. In chapter 5, we will discover a regulator molecule, 1,4-HQ that regulates DHP’s peroxidase function by a proton coupled electron transfer (PCET) mechanism. In chapter 6, we will measurebackbone dynamics of metcyano form DHP by applying NMR relaxation method and MD simulations. And in chapter 7, we will investigate the unique autoreduction pathway observed in DHP.

1.2 References

(1) Hardison, R. C. P Natl Acad Sci USA 1996, 93, 5675.

(2) Perutz, M. F. Annual Review of Biochemistry 1979, 48, 327.

(3) Volkman, B. F.; Alam, S. L.; Satterlee, J. D.; Markley, J. L. Biochemistry 1998, 37,

10906.

(4) Edelstein, S. J. Annu. Rev. Biochem. 1975, 44, 209.

(5) Baldwin, J.; Chothia, C. J. Mol. Biol. 1979, 129, 175.

(6) Kajimura, M.; Fukuda, R.; Bateman, R. M.; Yamamoto, T.; Suematsu, M. Antioxidants

& Redox Signaling 2009, 13, 157.

(7) Liu, J.; Chakraborty, S.; Hosseinzadeh, P.; Yu, Y.; Tian, S.; Petrik, I.; Bhagi, A.; Lu,

Y. Chemical Reviews 2014, 114, 4366.

(8) Poulos, T. L. Chemical Reviews 2014, 114, 3919.

(9) Chen, Y. P.; Woodin, S. A.; Lincoln, D. E.; Lovell, C. R. J Biol Chem 1996, 271, 4609.

(10) Chen, Y. P.; Lincoln, D. E.; Woodin, S. A.; Lovell, C. R. J Biol Chem 1991, 266,

23909.

(11) Belyea, J.; Gilvey, L. B.; Davis, M. F.; Godek, M.; Sit, T. L.; Lommel, S. A.; Franzen,

S. Biochemistry 2005, 44, 15637.

(12) Barrios, D. A.; D’Antonio, J.; McCombs, N. L.; Zhao, J.; Franzen, S.; Schmidt, A. C.;

Sombers, L. A.; Ghiladi, R. A. J. Am. Chem. Soc. 2014, 136, 7914.

(13) Franzen, S.; Belyea, J.; Gilvey, L. B.; Davis, M. F.; Chaudhary, C. E.; Sit, T. L.;

Lommel, S. A. Biochemistry 2006, 45, 9085.

(14) Thompson, M. K.; Franzen, S.; Davis, M. F.; Oliver, R. C.; Krueger, J. K. J. Phys.

Chem. B 2011, 115, 4266.

(15) de Serrano, V.; Chen, Z. X.; Davis, M. F.; Franzen, S. Acta Crystal. D-Biol. Cryst.

2007, 63, 1094.

(16) D’Antonio, J.; D’Antonio, E. L.; Thompson, M. K.; Bowden, E. F.; Franzen, S.;

Smirnova, T.; Ghiladi, R. A. Biochemistry 2010, 49, 6600.

(17) Zhao, J.; Lu, C.; Franzen, S. J. Phys. Chem. B 2015, 119, 12828.

(18) Gish, W.; States, D. J. Nature genetics 1993, 3, 266.

(19) Hurtado, L. A.; Lutz, R. A.; Vrijenhoek, R. C. Molecular ecology 2004, 13, 2603.

(20) Thompson, M. K.; Davis, M. F.; de Serrano, V.; Nicoletti, F. P.; Howes, B. D.;

Smulevich, G.; Franzen, S. Biophys. J. 2010, 99, 1586.

(21) Nicoletti, F. P.; Thompson, M. K.; Howes, B. D.; Franzen, S.; Smulevich, G.

Biochemistry 2010, 49, 1903.

(22) Zhao, J.; Moretto, J.; Le, P.; Franzen, S. J. Phys. Chem. B 2015, 119, 2827.

(23) Davis, M. F.; Bobay, B. G.; Franzen, S. Biochemistry 2010, 49, 1199.

(24) Zhao, J.; de Serrano, V.; Zhao, J. J.; Le, P.; Franzen, S. Biochemistry 2013, 52, 2427.

(25) De Serrano, V.; Franzen, S. Peptide Sci. 2012, 98, 27.

(26) Wang, C.; Lovelace, L. L.; Sun, S.; Dawson, J. H.; Lebioda, L. Biochemistry 2013, 52,

6203.

CHAPTER 2

Measurement of Internal Substrate Binding in Dehaloperoxidase-hemoglobin

by Competition with the Heme-Fluoride Binding Equilibrium

2.1 Abstract

The application of fluoride anion as a probe for investigating the internal substrate binding has been developed and applied to dehaloperoxidase-hemoglobin (DHP) from Amphitrite ornata.

By applying the fluoride titration strategy using the UV-vis spectroscopy, we have studied series of halogenated phenols, other substituted phenols, halogenated indoles and several natural amino acids that bind internally (and non-covalently) in the distal binding pocket of the heme. This approach has identified 2,4-dibromophenol (2,4-DBP) as the tightest binding substrate discovered thus far, with approximately 20-fold tighter binding affinity than that of

4-bromophenol (4-BP), a known internally binding inhibitor in DHP. Combined with resonance Raman spectroscopy, we have confirmed that competitive binding equilibria exist between fluoride anion and internally bound molecules. We have further investigated the hydrogen bonding network of the active site of DHP that stabilizes the exogenous fluoride ligand. These measurements demonstrate a general method for determination of differences in substrate binding affinity based on detection of a competitive fluoride binding equilibrium.

The significance of the binding that 2,4-dibromophenol binds more tightly than any other substrate is evident when the structural and mechanistic data are taken into consideration.

2.2 Introduction

Fluoride anion turns out to be an ideal ligand to probe the internal binding of substrates and inhibitors.1 The advantage of using fluoride anion as a probe the binding affinity of internal binding ligands in DHP is an improvement over H2O, which naturally binds at the axial position of the ferric heme. The native ferric form of DHP is 60% metaquo (6cHS) and 40% met (5cHS) at pH 6.0. The water binding equilibrium is the reflected by the change of the coordination state of the ferric heme between 6cHS and 5cHS, which can be observed in 2

2,3 and 3 modes in the resonance Raman spectrum. These resonance Raman data correlate well

4 with the occupancy of H2O in the room temperature X-ray crystal structure. This correlation is strengthened by the fact that the distal histidine in the X-ray crystal structure (PDB 1EW6) has two conformations, one internal that hydrogen bonds to the H2O molecule and one external that corresponds to a lack of H2O and a 5-coordinate heme. The 6cHS H2O is displaced when inhibitors bind in the distal pocket and displace H2O to form a 5cHS complex, but the signal is small since the H2O occupancy is only 60%. By replacing H2O with fluoride anion one can compare the competitive binding with a ligand that is initially bound to 100% of the heme Fe atoms. The use of fluoride also has the advantage that the assay can be conducted using UV- vis spectroscopy. Fluoride weak field ligand and therefore there is no change the spin state of the heme iron upon binding. Therefore, the Soret band shows little shift (if any) upon fluoride binding. However, fluoride anion binding can be easily monitored in UV-vis spectra because there is a charge-transfer band CT1 at 605 nm that corresponds to the formation of DHP-F adduct.5

In the present study, we have applied the fluoride titration strategy using UV-vis spectroscopy to study the interaction of DHP A with 30 molecules that bind internally (and non-covalently) in the distal binding pocket of the heme. We used resonance Raman spectroscopic methods and density function theory (DFT) calculations to investigate the hydrogen bonding (H-bonding) network of the active site of DHP A as well. By applying the fluoride titration strategy to interrogate the competitive binding interaction between fluoride anion and internally bound molecules, we were able to identify 2,4-DBP as the tightest binding substrate discovered so far. The method also confirmed that the indole derivatives, which were observed to be peroxygenase substrates, bind internally in the distal pocket despite their large size.6 The trends of the binding affinity of the 5-halogen indole series are consistent with the preliminary mechanistic studies.7,8 Moreover, the pH-dependent fluoride titration experiment has shown that only the charge neutral form of the substrate bound internally in the distal pocket as we have surmised previously based on a number of studies.9-11 The data provide further evidence that DHP is a true multi-functional enzyme.

2.3 Material and Methods

Materials

All reagents were purchased from Aldrich and ACROS and used without further purification. All chemicals were each dissolved in 100 mM, pH = 7.0 potassium phosphate

(KPi) buffer to prepare the stock solution. 2,4,6-TBP and all halogenated indoles solutions were boiled in the water bath and then cooled down to make the supersaturated solutions that meet the concentration requirement. 0.2 M and 0.6 M NaF solutions were made in the

corresponding KPi buffer. Spectra were obtained using an Agilent 8453 diode array UV-visible spectrophotometer with a Peltier-cooled sample cell at 25 ℃. Wild-type His6X (histidine- tagged) DHP A and H55D mutant were expressed in E.coli and purified as previously described.12,13 The concentration of ferric DHP A was determined by using the molar absorption coefficient, ε = 116,400 M-1 cm-1.14 The Soret band of DHP A H55D mutant is at

398 nm and the corresponding extinction coefficient is 82,500 M-1 cm-1 determined by the pyridine hemochrome method. 15

Scheme 2.1 The equilibrium scheme of competitive binding between fluoride anion and internal binding substrate or inhibitor in WT DHPA.

Fluoride Titration Assays

[E]0 in each sample was 50 µM with a total volume of 600 µL. 1 mM or 5 mM substrates were included for the fluoride competitive binding titration and the cuvette was allowed to incubate for 3 min in the thermal cell to reach thermal equilibrium before the titration. Then the fluoride stock solution were gradually titrated in and the corresponding spectra were recorded. The fluoride titration spectra were eventually normalized according to the protein concentrations in the cuvette. The binding curve were extracted from the spectra region between 300 nm to

700 nm using SVD (singular value decomposition) and then fitted into the one site reversible

app binding model to obtain the apparent fluoride binding constant Kd , which is derived in the following.

Figure 2.1 shows two non-covalent binding sites called the - (substrate) and -

(inhibitor) site, as well as the ionic binding site for fluoride with the heme Fe. The X-ray crystal structure data shown in Figure 2.1 provide one of the motivations for this study. The -site or

inhibitor binding site was observed first in 1999 (PDB 1EWA),4 but only defined as an inhibitor site in 2010 (PDB 3LB2).7 The name -site comes from the fact that this binding site is closest to the -edge of the heme. The site was discovered by two research groups using different conditions (PDB 4HF6 and 4KMW,4KN3)8,9 and is given its name because the substrate is bound nearest to the -edge of the heme. Based on more recent work we hypothesize that the -site is the substrate binding site for the peroxygenase function,6 but is apparently also a substrate inhibition binding site for the peroxidase function.8

Kinetic evidence suggests that DHP binds a number of substrates including halogenated phenols and indoles. Therefore, in order to understand the multiple functions of DHP it is imperative to characterize the interactions of large molecules in these two binding sites with ligands bound to the heme Fe. There are several possible ligands that could be considered, fluoride, azide, cyanide etc. for competitive binding studies in ferric DHP. Fluoride was chosen for this study because it has a relatively weak ionic bond that could be relatively easily displaced by non-covalent substrate and inhibitor binding. The competitive binding equilibrium is shown in the scheme 2.1, in which a square binding equilibrium scheme is used to illustrate the binding interactions between ferric DHP and both fluoride anion and substrates/inhibitors.

According to the square scheme there are four related equilibria. KL is the dissociation constant for the non-covalent interaction between internal binding molecules

f (substrates/inhibitors) and ferric DHP. Kd denotes the dissociation constant for the ionic bond

b between fluoride anion and ferric heme in the absence of any internally bound molecule. Kd is the dissociation constant between fluoride anion and ferric heme in the presence of an

app internally bound molecule. We have not included Kd in the scheme since it is not a simple

app equilibrium constant. A detailed derivation of Kd is given in the Appendix A, but the following derivation shows the crucial steps needed to obtain the final equation.

app The starting point for the derivation of Kd is to assume that the enzyme exists in four forms, which are free enzyme, E, substrate-bound enzyme, EL, fluoride bound enzyme, EF and simultaneous substrate and fluoride-bound enzyme, EFL (Scheme 2.1). The mass conservation equation is:

[퐸]0 = [퐸] + [퐸퐿] + [퐸퐹] + [퐸퐹퐿] (1) given that [E]0 is the total enzyme concentration. This equation can also be written as:

[퐸]0 [퐸] = (2) [퐿] [퐹] [퐿][퐹] 1 + + 푓 + 푏 퐾퐿 퐾 퐾 퐾푑 퐿 푑

Assuming that the fluoride ion concentration is much greater than the enzyme concentration

[F] >> [E]0, the fluoride binding fraction θ that can be monitored using the UV-Vis:

[퐸퐹] + [퐸퐹퐿] θ = (3) [퐸]0

Substituting in equation (2):

[퐸][퐹] [퐸][퐿][퐹] 푓 + 푏 퐾 퐾퐿퐾푑 θ = 푑 (4) [퐿] [퐹] [퐿][퐹] (1 + + 푓 + 푏 ) [퐸] 퐾퐿 퐾 퐾 퐾푑 퐿 푑

A rearrangement of this equation permits it to be written a standard form for a binding isotherm provided we define an apparent binding constant:

[퐹] [퐹] θ = = 푎푝푝 (5) ([퐿] + 퐾 )퐾푓퐾푏 퐾 + [퐹] 퐿 푑 푑 + [퐹] 푑 푓 푏 [퐿]퐾푑 + 퐾퐿퐾푑

app Therefore, we can define the apparent fluoride dissociation constant Kd :

푓 푏 푎푝푝 ([퐿] + 퐾퐿)퐾 퐾푑 퐾 = 푑 (6) 푑 푓 푏 [퐿]퐾푑 + 퐾퐿퐾푑

app As we can see from the Eqn. 6, Kd is dependent on both the concentration of internally bound molecules [L] and the dissociation constant KL. Due to the presence of competitive binding,

b f Kd should be much larger than Kd . Therefore, either an increase in the molecular concentration

[L] or replacement by a more tight-binding molecule L' (with smaller dissociation constant KL' value) will drive the fluoride binding equilibrium towards the dissociated form, resulting in an

app increase in Kd .

Resonance Raman Spectroscopy

DHP A samples at a final protein concentration of 100 µM were prepared in 100 mM

KPi buffer, pH 7. 0. The DHP-F complex were prepared with 0.2 M NaF in the KPi buffer with corresponding pH (pH 5.0 and pH 7.0). And the DHP-F sample at pD 7.0 were prepared and equilibrated in D2O. Samples were placed in 5 mm NMR tubes and spun with an air piston spinning sample holder (Princeton Photonics, model Raman 101). Resonance Raman spectra were obtained by excitation at the edge of the Soret band at 410 nm. The Coherent Mira 900

tunable titanium sapphire laser that generating 700 ~ 1000 nm light was pumped by a Coherent

Verdi 10 frequency-doubled diode-pumped Nd: vanadate laser that generating 10 W of 532 nm light. The near-IR output from the Ti:sapphire laser was sent through a Coherent 5-050 frequency doubler to generate the working range of 400-430 nm light for Soret band excitation.

The frequency doubled beam was collimated and cylindrically focused to a vertical line of ~5 mm and typically 90~100 mW at the sample. Raman scattered light was collected by the Spex

1877 Triplemate monochromator (2400 grooves/mm final stage grating) and was detected by a liquid N2-cooled CCD camera (ISA Spex, model CCD-3000). Spectra were measured at room temperature for 40 acquisitions with a total exposure time of 1200 seconds. The spectrograph resolution was determined to be 2.2 cm-1 using argon lamp lines. The Raman spectra were first calibrated using standard spectra of toluene, and then baseline subtracted by using a 4 point 4 polynomial extrapolation and normalized to the intensity of the ν7 band (see Figure 2.3).

Density functional theory calculations

Density functional theory (DFT) calculations were conducting using a numerical basis set implemented in the program DMol3. 16,17 The PBE functional was used. 18 The numerical basis set used is equivalent in quality to a double-zeta basis set with polarization functions on all atoms (DNP). The cutoff used from each atomic center was 20 Bohrs. The criterion for convergence was a change less than one part in 10-6 for two subsequent iterations of geometry refinement. Fluoride binding energies were carried out using a continuum dielectric for all models. The model chosen is the option COSMO in DMol3. This model is necessary because of the negative charge on the fluoride and positive charge on all unligated ferric heme models

(except those including OH-).19 A dielectric constant of 78.4 was used to model room

temperature aqueous solvation. Vibrational frequency calculations were carried out by finite difference to construct the force constant matrix. To construct the finite difference forces, each nucleus was shifted six times, along ±x, ±y and ±z. by 0.1 Bohr to obtain an estimate of the second derivative (Hessian), which is also the force constant. The calculations were carried out using the THERMAL option, with the electronic temperature set to 0.005 au (ca. 1500 K) 20.

This tool in the DFT method has no physical significance since a correction to zero Kelvin is implemented when the calculation is completed. However, it helps to ensure convergence in large systems. The systems studied here range from 97 to 109 atoms. The calculations were carried for d5 ferric Fe in the high spin state so that all of the heme systems are considered to be in the same spin state, consistent with observations by UV-vis spectroscopy.

Using the COSMO dielectric continuum model in the calculation does not describe the solvation of F at the molecular level. In order to correct for the specific solvation effect of

- H2O on the binding of the F ion, the association was carried out using the following equilibrium:

+ − 퐻푒푚푒 + 16 퐻2푂 − 퐹 → 퐻푒푚푒 − 퐹 + 16 퐻2푂 where Heme represents a model that may include proximal histidine and distal amino acid components to represent the hydrogen bonding interactions. The various combinations are given in Tables 2.4 and 2.5 and some of these structures are shown in Figure 2.6. The calculated energy is the energy required to transfer the fluoride ion from the heme Fe atom to solvent, which is modeled using 16 H2O molecules. The COSMO dielectric continuum was applied to all components of this model.

2.4 Results

In this study, we present a competitive binding assay as a general method for determining substrate binding constants needed to understand the nature of internal binding in

DHP. Since internal binding corresponds to both substrate and inhibitor binding, these measurements have a direct relationship to the peroxidase, peroxygenase and oxidase functions of DHP. In previous assays of this type strong binding ligands, such as CN, have been used to displace the substrate (or inhibitor). Instead, the assay used here started with the F-bound form of the heme, which is then challenged with various concentrations of substrate/inhibitor.

There is X-ray crystal structural evidence for both internal inhibitor and substrate binding and these structures suggest that the large molecules will compete with fluoride binding.2,3,21 The competition assay has been developed as a tool to screen for potential new substrates or inhibitors, which may further expand the repertoire of DHP functions.

Figure 2.4 UV-vis spectra of fluoride titration of WT DHPA and DHP A H55D mutant in the absence or in the presence of substrate. (a) Fluoride titration of WT DHPA (100 µM) in the 100 mM KPi buffer, pH 7.0. (b) Fluoride binding curve extracted from spectra data of figure 2.1a using SVD. (c) Fluoride titration of WT DHPA (50 µM) in the presence of 1 mM 2,4,6-TCP in the 100 mM KPi buffer, pH 7.0. (d) Fluoride titration of DHP A H55D mutant (50 µM) in the 100 mM KPi buffer, pH 7.0, the NaF concentration is up to 0.3 M.

The fluoride titration and the competitive binding between DHP-F adduct and the internal substrate binding

The fluoride titration with the WT DHP A and DHP A H55D mutant in the absence or in the presence of substrate were first studied using UV-vis spectroscopy, as shown in Figure 2.2.

The titration of fluoride to WT DHP A does not noticeably change the Soret band, which remains at 407 nm throughout the titration (Figure 2.2a). In contrast, the Q band of WT DHP

A at 506 nm was gradually diminished and the charge transfer band CT1 at 605 nm increased in intensity, which indicates the formation of the six coordinated high spin (6cHS) fluoride adduct, DHP-F. 22 The fluoride binding curve for the WT DHP A in the absence of substrate

is shown in Figure 2.2b. However, there is an additional band that increases intensity at 579 nm in the presence substrate 2,4,6-TCP (Figure 2.2c) during the fluoride titration, which may be due to the interaction between internal binding substrate 2,4,6-TCP and the axial fluoride ligand. The H55D mutants showed no binding affinity towards fluoride since no spectral changes were observed upon adding fluoride to protein solutions (Figure 2.2d). This result is not a surprise because the distal histidine, H55 is considered to be the crucial H-bond donor that stabilizes the binding of an axial fluoride ligand to the heme Fe atom. 12

The fluoride anion dissociation constants of both WT DHP A and DHP B have been measured (Table 2.1). Since these dissociation constants are measured in the absence of

f aromatic substrates and inhibitors, we denote this dissociation constant as Kd . Due to close

f similarity of DHP A and DHP B structures and sequences, Kd are nearly identical. In order

f better understand the distal environment of DHP A and B, we present a comparison of the Kd of Horse skeletal muscle myoglobin (HSMb) and HRP using the same titration method in this study. HSMb and HRP have ~ 3 fold and ~ 14 fold weaker binding affinities compared to DHP, while the Tf-trHb and CcP have ~ 17 fold and ~ 2500 fold stronger binding affinities than

DHP, respectively.

f Table 2.1 The fluoride dissociation constant Kd of several hemoproteins

-1 f Hemoprotein CT1 (nm) Fe-F (cm ) Kd (mM) WT DHP A 605 461 4.5 ± 0.1 WT DHP B 605 461 4.5 ± 0.4 Myoglobin (Horse Muscle) 607 14.2 ± 0.5 HRP 611 387 61.8 ± 4.9 TtH-NOX 23 609 1.5 ± 0.2 Tf-trHb 9 613 381 0.26 ± 0.01 CcP 11 617 (1.8 ± 0.4)  10-3

The competitive binding between large molecules (L) and fluoride binding to the heme

Fe can be quantitatively determined by monitoring the increase of the apparent dissociation

app f constant Kd given in Eqn. 6 compared to the dissociation constant Kd measured in the

app absence of substrate or inhibitors. Indeed, Kd did increase in the presence of the substrates and inhibitors (the mono-halogenated phenol series) or the classical substrates (the 2,4,6-

app trihalogenated phenol series) as shown in Table 2.2. Kd measured for L = 5 mM is

app approximately 2-fold greater than the Kd measured for L = 1 mM. For example, this can be seen for the 4-halogenated phenol series of inhibitors 4-fluoro-, 4-chloro-, or 4-bromophenol.

The inhibitor series is the best characterized, which permits us to propose an inverse correlation between the large molecule dissociation constant KL. KL follows the trend: 4-IP < 4-BP < 4-

3 app CP < 4-FP while Kd follows the opposite trend (Table 2.2).

app Table 2.2 The apparent fluoride dissociation constant Kd of DHP A in the presence of internal binding molecules

app Kd (mM) Internal Binding Molecule pKa(ph-OH) 5 mM 1 mM 4-Fluorophenol 9.89 24 16.2 ± 0.7 7.0 ± 0.3 4-Chlorophenol 9.43 24 22.4 ± 1.1 12.2 ± 0.5 4-Bromophenol 9.34 24 22.9 ± 1.1 12.3 ± 0.8 4-Iodophenol 9.20 24 N.A.(a) 16.4 ± 0.8 2-Bromophenol 8.43 25 163.4 ± 13.8 25.0 ± 0.6 3-Bromophenol 9.03 26 N.A.(b) 2,4,6-Trifluorophenol 7.20 27 46.4 ± 4.1 12.7 ± 0.7 2,4,6-Trichlorophenol 6.59 28 N.A.(a) 14.5 ± 0.8 2,4,6-Tribromophenol 6.34 28 N.A.(a) 23.8 ± 1.0 2,4-Dicholorophenol 8.05 28 N.D. 74.1 ± 4.3 2,4-Dibromophenol 7.86 28 N.D. 172.1 ± 9.0 2,6-Dibromophenol 6.89 28 21.2 ± 1.2 5.7 ± 0.5 Phenol 9.97 29 8.3 ± 0.3 N.D. p-Cresol 10.15 29 7.4 ± 0.4 N.D. 2,3,6-Trimethylphenol 10.64 30 7.7 ± 0.3 N.D. p-Methoxyphenol 10.20 24 4.8 ± 0.1 N.D. Guaiacol 9.93 31 8.0 ± 0.3 N.D. Ferulic Acid 9.21 32 8.7 ± 0.3 N.D. Benzohydroxamic Acid N.A. 10.7 ± 0.7 N.D. L-Tryptophan N.A. 5.5 ± 0.2 N.D. L-Phenylalanine N.A. 6.6 ± 0.4 N.D. L-Tyrosine 10.20 33 N.A.(c) 10.4 ± 0.6 Indole N.A. 7.9 ± 0.2 5.0 ± 0.2 5-Fluoroindole N.A. N.A.(c) 5.8 ± 0.7 5-Chloroindole N.A. N.A.(c) 8.6 ± 0.3 5-Bromoindole N.A. N.A.(c) 13.2 ± 0.5 5-Iodoindole N.A. N.A.(c) 16.8 ± 0.4 4-Bromoindole N.A. N.A.(c) 5.2 ± 0.1 6-Bromoindole N.A. N.A.(c) 6.1 ± 0.2 7-Bromoindole N.A. N.A.(c) 6.6 ± 0.2 N.A. : Not Available. N.A.(a) : Not Available because of protein denaturing. N.A.(b) : Not Available because ligand coordinates to the heme. N.A.(c) : Not Available because the limit of solubility. N.D. : Not Determined

The tighter the binding of the large molecule (i.e. the smaller the KL) the greater the

app corresponding increase in the apparent fluoride dissociation constant (the larger the Kd ).

app Using this principle, we can estimate the relative KL by measuring the Kd using a fluoride

f app titration. The Kd of DHP A is 4.51 mM, thus Kd larger than 4.51 mM indicates internal

app binding in the distal pocket. The Kd of the “native” substrate 2,4,6-TBP and inhibitor 4-BP phenol provide reference values for evaluation of other the internally bound molecules.

Based on this interpretation, the molecule 2,4-dibromophenol (2,4-DBP) has the tightest binding affinity of any molecule tested. It is therefore also the tightest binding among the three bromophenol substrates, 2,4,6-TBP, 2,4-DBP and 4-BP. Its binding affinity is approximately 14 fold of that of 4-BP. To further probe the impact of the halogen substituted

app position on the aromatic ring, we have measured the Kd for 2-BP, 3-BP and 2,6-DBP. 2-BP

app has shown larger Kd than that of 4-BP, which indicates that the ortho- position of the phenolic group may give a tighter binding affinity than the para- position. Unlike 4-BP, the binding of

2-BP to DHP gives rise a charge transfer band at 636 nm and the Soret band is less blue shifted compared to that of 4-BP. However, if both ortho- positions are occupied as in 2,6- dibromophenol (2,6-DBP) the binding affinity becomes dramatically lower. It is worth noting that 2,6-DBP has a lower binding affinity than 2,4,6-TBP. The contrast between 2,4-DBP and

2,6-DBP is striking. The former binds more tightly than 2,4,6-TBP by a factor ~7, while the latter binds ~4 times more weakly. Unlike both 2-BP and 4-BP, 3-BP directly coordinates to the heme iron with the binding affinity Kd = 184.1 ± 8.6 µM that is 25 fold smaller than that of fluoride. The UV-vis spectrum of 3-BP DHP binding complex has the Soret band at 406 nm

and a charge transfer band at 613 nm, which is quite distinct from the spectra for 2-BP and 4-

BP binding (Figure A2).

app Using the same strategy, we have measured the Kd of phenol, several substituted phenols and three aromatic amino acids. As for most internal binding molecules, their dissociation constant KL are usually much smaller than their concentration [L], especially when

[L] = 5 mM. Thus, KL can be ignored compared to [L]. Therefore, we can simplify Equation 2

app app to show that KL is inversely proportional to the Kd . Based on the Kd value, we conclude that phenol, p-cresol, guaiacol, 2,3,6-trimethylphenol, benzohydroxamic acid and ferulic acid present moderate binding affinity towards DHP, which is about half of the binding affinity of the inhibitor 4-BP. The amino acids L-phenylalanine and L-tryptophan bind poorly and are estimated to possess only one fifth binding affinity of 4-BP. In contrast, L-tyrosine shows binding affinity comparable to the 4-halogenated phenols.

Indole derivatives comprise a new class of substrate discovered along with the peroxygenase function of DHP. 6 Therefore, indole and its derivatives were also tested in

app competitive binding studies with fluoride to measure the corresponding Kd values. The 5-

app halogenated indole series (F, Cl, Br, I) clearly show a trend that the corresponding Kd increases as the radius of the halogen atom increases, which is consistent with previously reported binding affinity of 5-halogenated indoles. Moreover, the 5- position of the indole is most crucial in determining the indole derivative’s binding affinity towards DHP. Because

app Kd of 5-bromoindole is approximately 2 fold higher than that for the 4-, 6- and 7-

app bromoindoles. All of these relative binding affinity reported by Kd of fluoride titration are consistent with previous published resonance Raman spectra of indole derivatives of DHP B,

in which 2 and 3 vibrational modes referred as coordination state marker modes were monitored to tell the binding affinity of the indole derivatives.

f app Table 2.1 The fluoride dissociation constants Kd and Kd in different pH.

f app Kd (mM) Kd (mM) pH DHP A WT DHP A + 4-BP (1 mM) DHP A + 2,4,6-TCP (1 mM) pH 5.0 2.6 ± 0.1 13.4 ± 0.3 N.A. pH 6.0 3.1 ± 0.2 12.1 ± 0.2 38.7 ± 1.4 pH 7.0 4.5 ± 0.1 12.3 ± 0.8 14.5 ± 0.8 pH 8.0 17.2 ± 1.1 N.A. 66.7 ± 10.9

pH-dependent fluoride titration

In contrast to HRP, the fluoride binding affinity of DHP is relatively stable between

34 f f pH 5.0 and 7.0. The Kd of DHP A at pH 7.0 is only about 1.7 fold of Kd at pH 5.0. The large

f increase of Kd at pH 8.0 is because approximately 50% of the DHP A is ligated with hydroxide, which competes with the fluoride binding in the axial position. 35 However, some of the internal binding molecules with pKa of the phenolic group between 5.0 and 7.0 are sensitive to the pH because the population of the protonated or deprotonated species are determined by pH exponentially. It is believed that only the charge neutral from of halogenated phenols can bind internally in DHP due to the highly hydrophobic environment in the distal pocket. Indeed, for

app 2,4,6-TCP with pKa = 6.6, the increase of Kd at pH 6.0 compared to pH 7.0 implies that there are more 2,4,6-TCP bound internally due to the increase population of the charge neutral form at pH 6.0 (Table 2.3). In the contrast, 4-BP with pKa = 9.3 does not show a significant pH-

app dependent effect on the value of Kd between pH 5.0 to 7.0. To sum up, this pH-dependent

fluoride titration experiment proves that only charge neutral form of the substrate binds internally in the highly hydrophobic distal pocket.

DHP-F adduct revealed by resonance Raman spectroscopy

DHP-F pD 7.0 -1 DHP-F pH 5.0 υ 1371 cm DHP-F pH 7.0 4 WT DHPA pH 7.0 -1 υ4 1372 cm -1 υ2 1565 cm

-1 υ3 1480 cm -1 υ2 1564 cm -1 υ3 1481 cm

1000 1100 1200 1300 1400 1500 1600 -1 Wavenumber (cm ) Figure 2.3 Resonance Raman spectra of WT ferric DHP A at pH 7.0 (red), DHP-F at pH 7.0 (blue), DHP-F at pH 5.0 (orange), DHP-F at pD 7.0 (purple), in the high frequency region.

Figure 2.4 Resonance Raman spectra of WT ferric DHP A at pH 7.0 (red), DHP-F at pH 7.0 (blue), DHP-F at pH 5.0 (orange), DHP-F at pD 7.0 (purple), in the low frequency region.

Resonance Raman spectroscopy was used to characterize the DHP-F adduct in both high and low frequency region (Figures 2.3 and 2.4). Moreover, the DHP-F adduct is assigned

-1 by the strong 2 mode at 1565 cm , which indicates the presence of a 6cHS ligation, in

22 agreement with previous observations. We observe that there are two 3 modes in DHP without fluoride bound (bottom spectrum in Figure 2.3). This is in agreement with previous observations and arises from the fact that ferric DHP A exists in an equilibrium of 5cHS and

22 6cHS. Both 5cHS and 6cHS are present in the metaquo form as indicated by the 3 shift from

1564 cm-1 to 1565 cm-1 (Figure 2.3). In the low frequency region, the iron fluoride stretching band was previously assigned at 462 cm-1 for 430 nm excitation.22 It has been assigned at 461 cm-1 for 410 nm excitation in this study. The change of pH has no significant effect on the low frequency region in the resonance Raman spectra, which is reasonable because the binding affinity of fluoride is only enhanced by 1.7 fold when pH is shifted from 7.0 to 5.0. However,

-1 the Fe-F stretching band was shifted to 462 cm in the KPi buffer with D2O at pD 7.0. The increased wavenumber of Fe-F stretching frequency indicates the enhancement of the Fe-F bond strength. One possible scenario is that there is water molecule in the distal pocket that H- bonded to the fluoride anion. Replacement of H2O with D2O molecule will weaken the H-bond between water and fluoride, which in turn will enhance the Fe-F bond strength, resulting in a higher stretching frequency.

In Figure 2.5, upon the titration of the tightest binding substrate 2,4-DBP, we were able to observe the gradual diminishing of the Fe-F stretching band at 461 cm-1 as the substrate concentration is increased. This clearly shows the fluoride ligand start to dissociate upon the binding of 2,4-DBP, which indicates the strong competitive binding between fluoride anion

and internal binding molecule. If there were any simultaneous binding, we would expect the intensity of the Fe-F stretch to be preserved and the frequency to be shifted. The comparison of pH and pD shows the kind of shift that can be expected.

Figure 2.5 Resonance Raman spectra of DHP-F titrated with 0 mM 2,4-DBP (blue), 0.1 mM 2,4-DBP (red), 0.2 mM 2,4- DBP (purple) and 1.0 mM 2,4-DBP (black) in the low frequency region.

Fe-F bonding studied by density functional theory

DFT calculations were carried out on the heme models shown in Figure 2.6 and several other related models, which are described in Tables 2.4 and 2.5 in terms of the distal and proximal ligands. The vibrational frequencies and fluoride binding energies given in Tables

2.4 and 5, respectively, show that the strength of the Fe-F bond depends on both H-bonding

(cis effect) and axial ligation (trans effect). Table 2.4 shows a general trend for the wave number of the Fe-F stretching mode in cm-1 to decrease with the strength of distal H-bonding.

The trans effect produces an decrease in binding energy and wave number in direct proportion to the strength of the proximal interaction. The stronger the trans ligation (i.e. for a more basic proximal ligand), the weaker the Fe-F bond. Thus, for example, the strongest binding and highest frequency is observed for 5-coordinate heme-F, the binding energy and (Fe-F) cm-1

  decrease in the order OH (His) < H2O(His) < His < No ligand. The calculated F -binding energies shown in Table 2.5 depend on both the distal H-bonding effect and the proximal trans ligation effect. Because of the opposing distal and proximal trends we must consider the distal and proximal effects separately. In comparison with the calculated binding energies on the proximal side, the H2O(His) calculated binding energy does not follow the trend, but the other binding energies do follow the trend that increased axial (trans) ligation strength weakens the

Fe-F bond. At the present time we have no explanation for the H2O(His) anomaly. Distal ligation binding energies are complicated by the fact that there are two contributions to the binding energy. The introduction of H-bonding on the distal side weakens the Fe-F bonding interaction. However, the total binding energy of F- must also take the hydrogen binding interaction of Fe-H into account. This fact prohibits direct comparison of these binding

energies with the Fe-F frequencies, for example, since the frequencies report only on the force constant of the Fe-F bond.

Model (Fe-F) (cm-1) Proximal Distal NA NA 593 His NA 550 His His (NH) 452 His His (CH) 435 His Ile 543

His H2O 496

His (H2O) NA 470

His (H2O) H2O 489

His (H2O) His (NH) 372 His (OH-) NA 451

The DFT calculations show that the cis effect of distal histidine His or a distal water molecule significantly lowers the Fe-F stretching frequency by a H-bonding interaction, which is consistent with previously observed shifted (Fe-F) stretching frequency in the series of mutated Tf-trHb by resonance Raman spectroscopy. 36 It is relevant that DHP has an observed

(Fe-F) stretch of 461 cm-1, which is in the correct range for a model with both a proximal histidine and moderately strong H-bonding by a distal histidine. The calculated wave number was not scaled or corrected for anharmonicity. The trends on distal side agree with the general idea that hydrophobic amino acids in the distal position have weak H-bonding (e.g. Ile in Table

2.4), and therefore the (Fe-F) stretching frequency increases.

Table 2.5 Solution and binding energies calculated using DFT methods including a COSMO dielectric continuum model

Heme Species Heme – F- Heme ΔE ΔE(F- bind) ΔE(F- (aq)) Proximal / Distal E(Ha) E(Ha) (Ha)a (kcal/mol) b (kcal/mol) c NA / NA -3196.85 -3096.83 -100.021 -24.4 -13.6 His / NA -3501.47 -3401.47 -100.003 -18.8 -8.1 His / His (NH) -3806.08 -3706.07 -100.013 -21.9 -11.1 His / His (CH) -3806.08 -3706.09 -99.995 -16.5 -5.7 His / Ile -3699.02 -3599.02 -100.001 -18.4 -7.6

His / H2O -3577.86 -3477.87 -99.992 -15.5 -4.7

His(H2O) / NA -3577.86 -3477.85 -100.010 -21.1 -10.4

His(H2O)/His (NεH) -3882.48 -3782.48 -100.004 -19.2 -8.5

His(H2O) / H2O -3654.26 -3554.26 -100.008 -20.6 -9.8 His(OH-) / NA -3577.37 -3477.38 -99.986 -13.5 -2.7 His(OH-)/His (NεH) -3881.99 -3781.99 -99.998 -17.4 -6.6

a. This value is obtained by subtraction of column 2 from column 1.  b. The total energy of the F ion is -99.9429 Ha using the continuum dielectric model in H2O ( = 78.4). This value was added to E in column to obtain F binding energy in the continuum model.  c. Using this method the solvation energy of F ion in a solvent model consisting of 16 H2O molecules is -10.8 kcal/mol. The value in aqueous solution is obtained by subtracting the solvation energy of F ion in water (modeled as 16 H2O molecules) from column 4.

2.5 Discussion

On the choice of fluoride as the ligand for a competitive binding assay

Structural methods can provide information about the modes of binding of substrates and inhibitors, but they require a great deal of time and effort. The current study is not a substitute for structural studies, but a competitive binding assay that is relatively easy to implement since UV-vis spectroscopy can be used to detect the displacement of ligands bound to the heme Fe. We have already reported such an assay, in which the naturally occurring H2O

3,22 ligand is displaced by inhibitor binding. However, the H2O competition assay is of limited

utility since the shifts in the resonance Raman spectrum are quite small. The drawback of using

37 H2O ligand as a probe is that there is only about 60% of the metaquo form at pH 6.0. It is difficult to monitor the change of the coordination state in the UV-vis because there is no significant shift in the Soret, Q-band or charge-transfer band regions of the spectrum, but the structural conclusions have been corroborated by the use of electron paramagnetic resonance.38,39 CO is a diatomic probe that can also be used based on the Fe-C stretching band of the CO adduct monitored using resonance Raman spectroscopy between 400 – 600 cm-1 and the C-O vibrational mode observed by IR spectroscopy.40-42 However, CO only binds to the ferrous form of the heme, which is not the typical oxidation state for the starting point of peroxidase or peroxygenase chemistry. Moreover, CO binding is quite strong and there is no evidence for any competition with substrate binding. 35,43 Fluoride binding provides similar information to CO in that laser excitation in the Soret band region can give significant enhancement of the Fe-F stretching band at 461 cm-1 in the resonance Raman spectrum, which also provides a probe of internal large molecule binding.41,44 Fluoride has moderate binding affinity to the heme Fe so that two modes of interaction can be studied. Competitive binding will lead to loss of the Fe-F intensity. Simultaneous binding will lead to a shift in the Fe-F stretching frequency. Finally, the titration experiments can be carried out with no need for an anaerobic environment as required for CO adducts.

On the competition between non-covalent substrate and covalent fluoride binding

The competitive binding assay is based on the observation that a non-covalent internal substrate/inhibitor binding has an energy comparable to the bond ligation energy of fluoride.

f o Using the Kd values Table 2.1 we can estimate the binding free energy of fluoride. Using G

f -1 = -RTln(Ka) for the binding energy. Since Ka = 1/Kd ≈ 220 M , we can estimate the free energy as approximately -3.2 kcal/mol. Since the substrates/inhibitors can displace fluoride in same concentration range, their binding energies must be comparable to the covalent binding of fluoride itself. Of course, the competitive binding by mass action is concentration dependent. The calculated free energies from the Fe-F bond are comparable to the experimental range. The various models attempt to capture the possible range of distal and proximal effects in heme proteins with calculated binding energies ranging from -2.7 to -11.1 kcal/mol.

The computational model in Table 2.5 suggests that if the H-bond with His55 is disrupted, which occurs when the inhibitor 4-BP binds in the distal pocket, the binding energy of fluoride should increase by +3.0 kcal/mol. This value is obtained by comparing the model with a distal histidine that can H-bond (NH) and no ligand on the distal side (NA) in rows 3 and 2, respectively, of Table 2.5. The same comparison when a water molecule is on the proximal side H-bonded to the proximal histidine gives +3.4 kcal/mol in rows 5 and 2, respectively, of Table 2.5. This is the value that is relevant to the binding of substrate. The experimental binding energy of ~ -3.2 kcal/mol is consistent with the observation that the internal substrate (or inhibitor) binding can displace F. We can identify the distal H-bonding as the main cause of the destabilization of F. This conclusion is consistent with a large body of work on binding Fin heme proteins and its use as a probe of the distal environment. The

COSMO calculation gives a range values, which are comparable to the experimental value of

-3.2 kcal/mol. Since this value is at low end of the range in Table 2.5, The most appropriate model may be one in which the proximal histidine is partially polarized His(OH-), which has

been suggested based on resonance Raman data and DFT calculations.40,45,46 Calculation of the hydration energy

− − 16 퐻2푂 + 퐹 → 16 퐻2푂: 퐹 using DFT methods gives -10.8 kcal/mol, which is in the range of possible values given in

Table C1, which is taken from the CRC Handbook. Both the absolute and relative binding strengths obtained from DFT calculations are in the correct ranges and agree with experimental trends.

The role of distal histidine His55 in H-bond interactions of DHP-F adduct

The DHP-F adduct has a characteristic charge transfer band at 605 nm that can be observed in the UV-vis spectrum. This charge transfer band CT1 is due to the [a2u(π)→eg(dπ)] transition from porphyrin ring to heme iron. 47 The CT1 band of DHP is relatively blue shifted compared to the CT1 band of other hemoproteins (SWMb, hhMb, HRP and CcP) fluoride adduct, which indicates that there are fewer H-bonding interactions in the distal pocket to stabilize the fluoride anion. Similar to SWMb, the distal histidine is the only residue that responsible for stabilizing the fluoride anion in DHP. Evidence for the stabilization of fluoride by distal histidine H-bonding is found in the DHP A H55D mutant, which completely loses the ability to bind fluoride anion. Similar observations have been made for SWMb H64V and

H64L mutants. 48 The corresponding H55V mutant in DHP A is completely non-functional, consistent with the role played by the distal histidine in stabilization of ligand binding at the heme Fe. 49 However, the CT1 band of DHP-F is observed at 605 nm over the pH range 5.0 to pH 7.0, concomitant with the lack of change in the Fe-F stretching band at 461 cm-1 at both pH

5.0 and 7.0. This behavior is in contrast to HHMb, which has two pH-dependent forms of fluoride-bound complex .47 This observation is consistent with the hypothesis that the pH dependence of reactivity DHP A is due to the ionization of the substrate, rather than an effect of pH on the protein.9,11,50 The stabilization of bound fluoride by a "heme-linked" amino acid, which is most likely the distal histidine, has been observed using fluoride binding kinetics in

HRP. 34 The idea of using fluoride as a probe has been exploited in many studies of peroxidases using UV-vis and resonance Raman spectroscopy. 1,51-54 A B H N N H O His 55 N N H H F His 55 F III III Fe Fe

Scheme 2.2. Proposed H-bonding interactions with fluoride in DHP.

The distal pocket of DHP must possess flexibility to accommodate such a wide range of molecules in the distal pocket. Besides the allosteric distal histidine His55, hydrophobic residues Phe21, Phe35 and Val59 should stabilize the internal binding molecules through van de

Waals interaction. 3,37,43,55 A series of X-ray crystal structures, NMR studies and resonance

Raman studies have shown that H55 is unusually flexible in DHP, when compared to the distal histidine of other globins.3,56 The flexibility appears to be important in the competitive fluoride binding studies as well. The distal histidine, H55, is essential for stabilizing the fluoride binding in DHP. It is known that the distal histidine is the residue that fine tunes exogenous ligand binding in myoglobin. DHP has a similar scenario because the distal histidine, H55, is the only polar residue in the distal pocket. Such a role for the distal histidine has already been

established based on strong H-bond to the dioxygen ligand in the oxy-ferrous form and H-

37 bonding to H2O in the metaquo form. However, the DHP-CO turns out to be an exception as the distal histidine H55 is observed mostly adopting solvent exposed “open” conformation due to the weak H-bond interactions between H55 and heme-bound CO but rather strong H- bonds interaction between H55 and the propionates. 57 The greater flexibility of the distal histidine in DHP relative to other globins suggests that there could be stronger H-bonding in

DHP. One possible scenario is that distal histidine H55 mediates an indirect H-bonding to fluoride anion through a water molecule in the distal pocket (Scheme 2). This scenario is possible because water molecule can reside in the distal pocket as observed in the X-ray crystallographic structures PDB 1EW6 at room temperature and 2QFK at 100 K. For an intermediate H2O molecule to be present as shown in scheme 2A, His55 would need to be present in the NH tautomer. The isotope effect in this model would be explained by replacement of H2O by D2O. Another scenario is that distal histidine directly interacts with fluoride where the H-bond is N···F. In this case shown in scheme 2B the H-bonding would be weakened in D2O solution by H/D exchange, ND)···F. In the case of SWMb, the fluoride is H-bonded to the distal histidine, H64, and a solvent water molecule that significantly interacts with the same distal histidine.48 The Fe-F stretching band in DHP is 1 cm-1 higher in

D2O than that in H2O indicates that the strength of Fe-F bond is enhanced due to the weakened

H-bonding. This conclusion is also consistent with the DFT calculations given in Table 2.4.

However, we do not have sufficient information at this time to decide between the two models in scheme 2.2.

Natural selection and selective internal binding of molecules in the distal pocket

The current results show that DHP is promiscuous with regard substrate binding, which is similar to the xenobiotic metabolizing cytochrome P450 enzymes.58 However, the trend of binding affinity between brominated phenols (2,4-DBP > 2,4,6-TBP > 4-BP) still presents some selectivity towards these substrates which may be a result of evolution.59,60 Amphitrite ornata cohabitates with another polychaete marine worm Notomastus lobatus, a sub-surface filter feeder that secrets halogenated phenols to repel potential predators.61 Since A. ornata itself does not possess this “chemical weapon”, it appears that the most abundant protein in A. ornata, DHP has evolved a peroxidase function as a natural protection against the toxicity of brombinated compounds. Chemical analysis of volatile organohalogen compounds show that

N. lobatus contains abundant amount of 2,4-DBP and moderate amount of 2,4,6-TBP and 4-

BP. Thus, 2,4-DBP may be the primary substrate of DHP in nature. The structural evidence suggest that 2,4-DBP is adapted to the substrate binding site observed in the 2,4,6-TBP X-ray

8 crystal structure. First, 2,4-DBP also functions as a substrate in the presence of H2O2 (Figure

A1). Moreover, the ortho-bromo substituents make the 2,4-DBP unlikely to bind in the inhibitor binding site because of steric hinderance. While 2,6-DBP has not been identified in

N. lobatus, it has been observed in another polychaete lugworm, Arenicola brasiliensis that is also a sub-surface filter feeder.59 The ratio of the binding constants for the various brominated phenols studied in this work matches the abundance of brominated phenols in the benthic ecosystem where A.ornata lives. Therefore, it is reasonable to postulate that natural selection may play a role in shaping the internal binding site of DHP.

While bromoindoles are much less prevalent in benthic ecosystems, the binding of indoles in the distal pocket of a heme protein is similar to the binding of substrates in the mammalian enzymes, indoleamine-2,3-dioxygenase and tryptophan-dioxygenase. The determinants of substrate binding in these proteins have been probed in molecule detail using site-directed mutagenesis,62 following similar detailed studies in peroxidases.63 These observations are relevant for this study since spectroscopy has been used to quantify this binding in the dioxygenases.64 Moreover, substrate inhibition is recurrent theme in these studies of dioxygenases, as is observed in DHP as well.65,66

2.6 Conclusion

The present study permits a comparison of substrates and inhibitors that have different modes of binding. Both substrates and inhibitors bind in the distal pocket of the DHP protein in sites we call the - and -sites in Figure 2.1. From previous structural work it was known that 4- halogenated phenols (except 4-FP) act as inhibitors of DHP and bind in a well-defined binding site. One of the principal determinants of binding in the -site is presence of a Xe binding site in DHP, which is where the halogen of 4-BP resides when it is bound in the -site. 67 The hypothesized peroxygenase substrates 2,4-DCP and 2,4-DBP, are suspected to bind in a substrate binding site at the -heme edge. This alternative internal binding site is deep inside the protein. 2,21 The known peroxidase substrates, 2,4,6-TCP and 2,4,6-TBP , are believed to bind at an external site near the heme -edge at low concentration, but they are also observed to bind at the heme -edge internal site at higher concentration.2,21 The relative binding strengths of these large molecules is consistent a functional interpretation of the two major

internal modes of binding shown in Figure 2.1. The nearly perpendicular binding of the -site very near the heme Fe makes it an inhibitor since it both blocks the heme Fe so H2O2 cannot bind and it displaces the distal histidine (H55) so that it cannot act as an acid-base catalyst in the peroxidase function. The -site places the substrate in an orientation that permits H- bonding by the phenolic O-H to a bound ligand (e.g. H2O2) and it exposes the C-H in the 6- position so that it is poised for oxygen atom transfer in the peroxygenase function. This study shows that 2,4-DBP binds more tightly in the protein than any other ligand, which presumably means that it binds in the -site. The data show that although 2,4,6-TBP also binds the in - site, its binding is much weaker. While 2,4,6-TBP cannot compete with 2,4-DBP, it is capable of auto-inhibition or substrate inhibition at high 2,4,6-TBP concentration. These trends apparent in the study using a competitive fluoride binding assay provide a consistent picture of the relative strength of binding interactions needed to permit three of the protein functions of DHP to coexist; the globin, peroxidase and peroxygenase functions. It also explains the relative magnitude of inhibition for 4-BP. These provide important relative binding strengths for the bromoindoles, which will contribute to our understanding as further mechanistic studies on the peroxygenase and oxidase functions of DHP emerge.

2.7 References

(1) Neri, F.; Kok, D.; Miller, M. A.; Smulevich, G. Biochemistry 1997, 36, 8947.

(2) Zhao, J.; Zhao, J.; Franzen, S. J. Phys. Chem. B 2013, 117, 14615.

(3) Thompson, M. K.; Davis, M. F.; de Serrano, V.; Nicoletti, F. P.; Howes, B. D.;

Smulevich, G.; Franzen, S. Biophys. J. 2010, 99, 1586.

(4) Lebioda, L.; LaCount, M. W.; Zhang, E.; Chen, Y. P.; Han, K.; WHitton, M. M.;

Lincoln, D. E.; Woodin, S. A. Nature 1999, 401, 445.

(5) Smulevich, G.; Feis, A.; Howes, B. D. Acc. Chem. Res. 2005, 38, 433.

(6) Barrios, D. A.; D’Antonio, J.; McCombs, N. L.; Zhao, J.; Franzen, S.; Schmidt, A. C.;

Sombers, L. A.; Ghiladi, R. A. J. Am. Chem. Soc. 2014, 136, 7914.

(7) D'Antonio, J.; D'Antonio, E. L.; Thompson, M. K.; Bowden, E. F.; Franzen, S.;

Smirnova, T.; Ghiladi, R. A. Biochemistry 2010, 49, 6600.

(8) de Serrano, V., D'Antonio, J., Franzen, S., Ghiladi, R. A. Acta Crystallogr D 2010, 66,

529.

(9) Schkolnik, G.; Salewski, J.; Millo, D.; Zebger, I.; Franzen, S.; Hildebrandt, P. Int. J.

Mol. Sci. 2012, 13, 7466.

(10) Zhao, J. J.; Rowe, J.; Franzen, J.; He, C.; Franzen, S. Biochem. Biophys. Res. Commun.

2012, 420, 733.

(11) Schkolnik, G.; Utesch, T.; Zhao, J.; Jiang, S.; Thompson, M. K.; Mroginski, M. A.;

Hildebrandt, P.; Franzen, S. Biochem. Biophys. Res. Commun. 2013, 430, 1011.

(12) Zhao, J.; de Serrano, V.; Dumarieh, R.; Thompson, M.; Ghiladi, R. A.; Franzen, S. J.

Phys. Chem. B. 2012, 116, 12065.

(13) Ma, H.; Thompson, M. K.; Gaff, J.; Franzen, S. J. Phys. Chem. B 2010, 114, 13823.

(14) Osborne, R. L.; Taylor, L. O.; Han, K. P.; Ely, B.; Dawson, J. H. Biochem. Biophys.

Res. Commun. 2004, 324, 1194.

(15) Berry, E. A.; Trumpower, B. L. Anal. Biochem. 1987, 161, 1.

(16) Delley, B. J. Chem. Phys. 2000, 113, 7756.

(17) Delley, B. J. Chem. Phys. 1990, 92, 508.

(18) Perdew, J. P.; Burke, K.; Wang, Y. Phys. Rev. B 1996, 54, 16533.

(19) Klamt, A.; Schuurmann, G. Journal of the Chemical Society, Perkin Transactions 2

1993, 799.

(20) Delley, B. in Modern Density Functional Theory Vol. 2 Elsevier, Seminario, Politzer

Eds. 1995, 221.

(21) Wang, C.; Lovelace, L. L.; Sun, S.; Dawson, J. H.; Lebioda, L. Biochemistry 2013, 52,

6203.

(22) Nicoletti, F. P.; Thompson, M. K.; Howes, B. D.; Franzen, S.; Smulevich, G.

Biochemistry 2010, 49, 1903.

(23) Kosowicz, J. G.; Boon, E. M. Journal of inorganic biochemistry 2013, 126, 91.

(24) Breyer, E. D.; Strasters, J. K.; Khaledi, M. G. Anal. Chem. 1991, 63, 828.

(25) Le Paih, J.; Dérien, S.; Demerseman, B.; Bruneau, C.; Dixneuf, P. H.; Toupet, L.;

Dazinger, G.; Kirchner, K. Chem-Eur. J. 2005, 11, 1312.

(26) Migchels, P.; Zeegers-Huyskens, T. J. Phys. Org. Chem. 1995, 8, 77.

(27) Sigala, P. A.; Kraut, D. A.; Caaveiro, J. M. M.; Pybus, B.; Ruben, E. A.; Ringe, D.;

Petsko, G. A.; Herschlag, D. J. Am. Chem. Soc. 2008, 130, 13696.

(28) Yang, M.; Zhang, X. Environ. Sci. Technol. 2013, 47, 10868.

(29) McWhirter, C.; Lund, E. A.; Tanifum, E. A.; Feng, G.; Sheikh, Q. I.; Hengge, A. C.;

Williams, N. H. J. Am. Chem. Soc. 2008, 130, 13673.

(30) Chrystiuk, E.; Jusoh, A.; Santafianos, D.; Williams, A. Journal of the Chemical Society,

Perkin Transactions 2 1986, 163.

(31) Ragnar, M.; Lindgren, C. T.; Nilvebrant, N. O. J Wood Chem Technol 2000, 20, 277.

(32) Fazary, A. E.; Ju, Y. H. J Solution Chem 2008, 37, 1305.

(33) Nara, S. J.; Valgimigli, L.; Pedulli, G. F.; Pratt, D. A. J. Am. Chem. Soc. 2009, 132,

863.

(34) Dunford, H. B.; Alberty, R. A. Biochemistry 1967, 6, 447.

(35) Nienhaus, K.; Deng, P. C.; Belyea, J.; Franzen, S.; Nienhaus, G. U. J. Phys. Chem. B

2006, 110, 13264.

(36) Nicoletti, F. P.; Droghetti, E.; Boechi, L.; Bonamore, A.; Sciamanna, N.; Estrin, D. A.;

Feis, A.; Boffi, A.; Smulevich, G. J. Am. Chem. Soc. 2011, 133, 20970.

(37) de Serrano, V.; Chen, Z. X.; Davis, M. F.; Franzen, S. Acta Crystallogr D 2007, 63,

1094.

(38) Smirnova, T. I.; Weber, R. T.; Davis, M. F.; Franzen, S. J. Am. Chem. Soc. 2008, 130,

2128.

(39) Thompson, M. K.; Franzen, S.; Ghiladi, R. A.; Reeder, B. J.; Svistunenko, D. A. J. Am.

Chem. Soc. 2010, 132, 17501.

(40) Franzen, S.; Roach, M. P.; Chen, Y.-P.; Dyer, R. B.; Woodruff, W. H.; Dawson, J. H.

J. Am. Chem. Soc. 1998, 120, 4658.

(41) Spiro, T. G.; Soldatova, A. V.; Balakrishnan, G. Coordin Chem Rev 2013, 257, 511.

(42) Mak, P. J.; Gregory, M. C.; Sligar, S. G.; Kincaid, J. R. Biochemistry 2014, 53, 90.

(43) Nienhaus, K.; Nickel, E.; Davis, M. F.; Franzen, S.; Nienhaus, G. U. Biochemistry

2008, 47, 12985.

(44) Kitagawa, T.; Mizutani, Y. Coordin Chem Rev 1994, 135, 685.

(45) Franzen, S. J. Am. Chem. Soc. 2001, 123, 12578.

(46) Franzen, S. J. Am. Chem. Soc. 2002, 124, 13271.

(47) Droghetti, E.; Nicoletti, F. P.; Bonamore, A.; Sciamanna, N.; Boffi, A.; Feis, A.;

Smulevich, G. Journal of inorganic biochemistry 2011, 105, 1338.

(48) Aime, S.; Fasano, M.; Paoletti, S.; Cutruzzola, F.; Desideri, A.; Bolognesi, M.; Rizzi,

M.; Ascenzi, P. Biophys J 1996, 70, 482.

(49) Franzen, S.; Chaudhary, C.; Belyea, J.; Gilvey, L.; Davis, M. F.; Sit, T. L.; Lommel, S.

A. Biochemistry 2006, 45, 9085.

(50) Franzen, S.; Gilvey, L. B.; Belyea, J. L. BBA-Proteins. Proteom 2007, 1774, 121.

(51) Howes, B. D.; Veitch, N. C.; Smith, A. T.; White, C. G.; Smulevich, G. Biochem. J.

2001, 353, 181.

(52) Nissum, M.; Neri, F.; Mandelman, D.; Poulos, T. L.; Smulevich, G. Biochemistry 1998,

37, 8080.

(53) Neri, F.; Indiani, C.; Baldi, B.; Vind, J.; Welinder, K. G.; Smulevich, G. Biochemistry

1999, 38, 7819.

(54) Heering, H. A.; Jansen, M. A. K.; Thorneley, R. N. F.; Smulevich, G. Biochemistry

2001, 40, 10360.

(55) Davis, M. F.; Gracz, H.; Vendeix, F. A. P.; de Serrano, V.; Somasundaram, A.; Decatur,

S. M.; Franzen, S. Biochemistry 2009, 48, 2164.

(56) Chen, Z.; de Serrano, V.; Betts, L.; Franzen, S. Acta Crystallogr D 2009, 65, 34.

(57) Zhao, J.; de Serrano, V.; Franzen, S. Biochemistry 2014, 53, 2474.

(58) Poulos, T. L. Chem. Rev. 2014, 114, 3919.

(59) Fielman, K. T.; Woodin, S. A.; Walla, M. D.; Lincoln, D. E. Mar Ecol Prog Ser 1999,

181, 1.

(60) Lincoln, D. E., Fielman, K.T., Marinelli, R.L., Woodin, S.A. Biochem. Syst. Ecol. 2005,

33, 559.

(61) Chen, Y. P.; Woodin, S. A.; Lincoln, D. E.; Lovell, C. R. J. Biol. Chem. 1996, 271,

4609.

(62) Chauhan, N.; Basran, J.; Rafice, S. A.; Efimov, I.; Millett, E. S.; Mowat, C. G.; Moody,

P. C. E.; Handa, S.; Raven, E. L. FEBS. J. 2012, 279, 4501.

(63) Dunford, H. B. Peroxidases and Catalases: Biochemistry, Biophysics, Biotechnology and Physiology, John Wiley, Chichester 2010.

(64) Lu, C. Y.; Lin, Y.; Yeh, S. R. Biochemistry 2010, 49, 5028.

(65) Lu, C. Y.; Lin, Y.; Yeh, S. R. J. Am. Chem. Soc. 2009, 131, 12866.

(66) Efimov, I.; Basran, J.; Sun, X.; Chauhan, N.; Chapman, S. K.; Mowat, C. G.; Raven,

E. L. J. Am. Chem. Soc. 2012, 134, 3034.

(67) de Serrano, V.; Franzen, S. Biopolymers 2012, 98, 27.

CHAPTER 3

Distinct Enzyme-Substrate Interactions Revealed by Two Dimensional Kinetic Comparison between Dehaloperoxidase-Hemoglobin and Horseradish Peroxidase

3.1 Abstract

The time-resolved kinetics of substrate oxidation and cosubstrate H2O2 reduction by dehaloperoxidase-hemoglobin (DHP) on the seconds-to-minutes time scale was analyzed for peroxidase substrates 2,4,6-tribromophenol (2,4,6-TBP), 2,4,6-trichlorophenol (2,4,6-TCP) and ABTS. Substrates 2,4,6-TBP and 2,4,6-TCP show substrate inhibition at high concentration due to the internal binding at the distal pocket of DHP, whereas ABTS does not show substrate inhibition at any concentration. The data are consistent with an external binding site for the substrates with an internal substrate inhibitor binding site for 2,4,6-TBP and 2,4,6-

TCP. We have also compared the kinetic behavior of horseradish peroxidase (HRP) in terms

AH2 H2O2 of kcat, Km and Km using the same kinetic scheme. Unlike DHP, HRP does not exhibit any measurable substrate inhibition, consistent with substrate binding at the edge of heme near the protein surface at all substrate concentrations. The binding of substrates and their interactions with the heme iron were further compared between DHP and HRP using a competitive fluoride binding experiment, which provides a method for quantitative measurement of internal association constants associated with substrate inhibition. These experiments show the regulatory role of an internal substrate binding site in DHP from both a kinetic and competitive ligand binding perspective. The interaction of DHP with substrates due to internal binding actually stabilizes that protein and permits DHP to function under conditions that denature HRP. As a consequence, DHP a tortoise, a slow, but steady enzyme that wins the evolutionary race against the HRP-type of peroxidase, which is a hare, initially rapid, but flawed for this application because of the protein denaturation under the conditions of the experiment.

3.2 Introduction

Dehaloperoxidase-hemoglobin (DHP) has had a controversial history that has led to the hypothesis that it is a true multifunctional enzyme. In 1977 DHP was characterized as a hemoglobin responsible for oxygen transport in the marine worm Amphitrite ornata.1 In 1996,

DHP was discovered to possess a peroxidase function since it can catalyze the oxidative dehalogenation reaction for 2,4,6-TBP and other phenolic substrates in the presence of co-

2 substrate H2O2. Upon examination there is an apparent contradiction between globin and peroxidase functions, 3 which has been resolved by studies of function switching.4-6 Recently,

DHP was also found to be a peroxygenase for a series of halogenated indole derivatives, which further expands its functional range.7 DHP even has significant oxidase activity, which brings the total number of functions to four. As each of these functions has been discovered there has been a normal skepticism regarding the significance of the reactivity for function in the living organism. There are many known adventitious heme enzyme functions that may not be true functions in an organism. For example, the bacterial hemoglobin from Vitreoscilla shows similar peroxidase activity to that of HRP.8 And it has been known for many years that human hemoglobin (Hb) is a better peroxidase than human myoglobin (Mb), but no physiological role for this activity is known. 9,10 Sperm whale Mb has been the subject of significant protein engineering leading to significant peroxidase and peroxygenase activity in certain mutants.11-

13 Although myoglobins can be engineered to have moderately high peroxidase activity, they are not competitive with functional peroxidases. Since the moderately high peroxidase activity of human Hb is thought to have no relevance for normal function in humans, it is natural to question whether the peroxidase function of DHP is functionally relevant in A. ornata.14

2,4,6-TBP 2,4,6-TCP

(Closed)

H2O2 Heme iron binding site H55  

Fe (Open)  

H89 H2O2 Entrance

Figure 3.1 Illustration of two binding modes of two substrates, 2,4,6-TBP and 2,4,6-TCP in the distal pocket of DHP A obtained from an overlay of the coordinates files from three X-ray crystal structures of the PDB entries 4HF6, 4KMW and 4KN3. A proof of physiologically relevant function is a difficult task. However, the existence of multiple internal inhibitor binding sites in the heme distal pocket of DHP has provided evidence that DHP has evolved in ways that are unique within the globin family.15-21 These binding sites have no precedent in either the globin or the peroxidase protein family and suggest that the various toxins in the environment inhabited by A. ornata have specific differentiated interactions with DHP. 22 The internal phenol binding sites also have structural relevance since they apparently increase the resistance of DHP to denaturation under certain conditions. 23Since phenols are usually considered denaturants this property combined with the catalytic rate suggests that DHP plays a unique role as a detoxification enzyme in A. ornata.

Thus, continued exploration of the interaction of inhibitor binding and enzyme kinetics provides further evidence that the substrate interactions in DHP is not adventitious, but rather

indicative of a true function. The rationale for this behavior is that detoxification is a sufficiently high priority for the organism that the various functions that eliminate toxins may have evolved in the Hb since it is the most abundant protein in the organism. 10

The ping-pong mechanism is a name given to the two-electron oxidation of substrates via two subsequent one-electron steps in peroxidase enzymes. 24 This mechanism is considered to be a hallmark of peroxidase function.25 In this study, we use a two-dimensional Michaelis-

Menten model for H2O2 and substrate concentration to establish that DHP is a peroxidase that competes favorably with HRP at pH 7 specifically for oxidation of 2,4,6-TBP, which is considered to be the native substrate .2,21 2,4,6-TCP has also been studied as a substrate because of its higher solubility. Therefore, it will be convenient to refer to both halogated substrates using the designation 2,4,6-TXP (X=Cl, Br). Previous studies made the comparison at pH 5

26,27 since HRP is a secretory peroxidase that has optimal reactivity at low pH values found in the external environment.28 The present study of the peroxidase mechanism in DHP and HRP at pH 7 clarifies the significance of that function in a multi-functional enzyme in vivo.

This study reexamines DHP’s peroxidase kinetics in light of the relative binding constants determined by a fluoride titration technique, developed previously,29 to map the substrate binding profile in DHP A, DHP B, two isoforms of DHP that differs by five amino acid (I9L, R32K, Y34N, N81S, S91G) and HRPC. By applying a two-dimensional kinetic data strategy, we present a graphical and analytical method to better understand DHP’s peroxidase kinetic behavior, illustrate the kinetic roles of external and internal substrate binding sites and the interaction between internal binding substrate and cosubstrate H2O2. Thus, the kinetic models provide further evidence in support of a peroxidase mechanism and deepens our

understanding of the specific interactions of 2,4,6-TBP, which has been considered as the native substrate ever since its discovery in 1996.2

3.3 Material and Methods

Materials

All the reagents and chemicals were purchased from Sigma-Aldrich and ACROS and used without further purification. 2,4,6-trichlrophenol (2,4,6-TCP) and 2,4,6-tribromophenol

(2,4,6-TBP) and 2,2’-azino-bis(3-ethylbenzothiazoline-6-sulphonic acid) (ABTS) were each dissolved in 100mM, pH = 7.0 potassium phosphate (KPi) buffer to prepare the substrate stock solutions. Hydrogen peroxide solution was freshly made before each kinetic experiment. A

30% reagent grade hydrogen peroxide (H2O2) solution was added to 100 mM, pH=7.0 KPi buffer to make the stock solution. Wild-type His6X DHP A and DHP B were expressed in

30 E.coli and purified as previously described. DHP was oxidized by excess K3[Fe(CN)6] to reach the ferric state and then filtering through NAP-25 column to eliminate excess

K3[Fe(CN)6]. The concentration of ferric DHP was determined by using the Soret band molar

-1 -1 absorption coefficient, ε406nm = 116,400 M cm . Horse Radish Peroxidase (HRP, EC 1.11.1.7) isoenzyme C (Type VI) was purchased from Sigma-Aldrich without further purification. The

-1 concentration of ferric HRP was determined spectroscopically by using ε403nm = 92,000 M cm-1.

Bi-substrate Ping-Pong Mechanism for DHP’s Peroxidase Function

Scheme 3.3. Original DHP peroxidase Ping-Pong mechanism

푘1푘2푘3[퐸]0[퐻2푂2][퐴퐻2] 푉0 = (1) 푘2푘3[퐴퐻2] + 푘1(푘2+푘3)[퐻2푂2]

A peroxidase catalyzed oxidation reaction generally requires both a substrate and cosubstrate. The substrate is a reducing substrate AH2 and the co-substrate is a peroxide, which can be either H2O2 or an organic hydroperoxide. The peroxide reacts with ferric heme of the peroxidase to form the highly active oxo-ferryl species - Compound I/Compound ES for the following catalytic oxidation reaction.31,32 Subsequently, the peroxidase reaction cycle follows a ping-pong mechanism, which consists of two one-electron transfer steps; 1. a one-electron reduction by reducing substrate, which generates a substrate radical intermediate; 2. a second one-electron reduction back to the ferric heme, which generates another substrate radical intermediate. The name ping-pong has been used for a range of peroxidases to describe the radical pathway, in which both electron transfer steps lead to oxidation of the substrate. When the activation step is included the cycle consists of the three steps shown in Scheme 3.1. The initial rate can be expressed as a Michealis-Menten rate equation for both substrate AH2 and

cosubstrate H2O2 dimensions according to Scheme 3.1. However, the mechanism deviates from the assumptions of Michaelis-Menten theory in that the binding of cosubstrate to create the ES intermediate is not reversible. Moreover, the mechanism in Scheme 3.1 assumes that substrate and cosubstrate react with enzyme or its transition state independently, which implies that there is no interaction between substrate AH2 and cosubstrate H2O2. Therefore, an explicit bi-substrate ping-pong mechanism that includes the reversible binding of substrate AH2 and cosubstrate H2O2 is proposed as shown in Scheme 3.2. Although the formation of the highly active oxo-ferryl intermediate is generally believed irreversible, the formation of compound 0 does involve the reversible binding of H2O2 to the ferric heme, a step which may be impeded by steric hindrance and electrostatic interaction in the distal pocket. Compound 0 has been observed in HRP. 33 A similar phenomenon has been observed in DHP in the ferrous form,

4 where H2O2 can replace bound O2 to yield a transient H2O2-bound adduct. However, the ferric

H O AH2 compound 0 has not been observed in DHP. Based on kinetic Scheme 3.2, Km 2 2 and Km are defined explicitly by the microscopic rate constants as shown in Equations 2 ~ 5. The initial rate equation (Equations 2 and 6) can also be rearranged to the Michaelis-Menten form in both

app App substrate AH2 and cosubstrate H2O2 dimensions. However, Vm and Km are both dependent on the concentration of the complementary substrate as shown in Equation B6 and B7.

Therefore, we designed a series of kinetic assays to measure trends of initial rate V0 by varying both substrate AH2 and cosubstrate H2O2 concentrations. As a result, we can plot the initial rate as the response to two substrates concentration variables in a three-dimensional (3D) kinetic plot, and then fit the data on the kinetic “surface” to obtain the kinetic parameters.

Scheme 3.4. Proposed DHP peroxidase Bi-substrate Ping-Pong mechanism with reversible substrate binding

푘 [퐸] [퐻 푂 ][퐴퐻 ] 푉 = 푐푎푡 0 2 2 2 (2) 0 퐻2푂2 퐴퐻2 퐾푚 [퐴퐻2] + 퐾푚 [퐻2푂2] + [퐻2푂2][퐴퐻2]

In which

푘2푘4푘6 푘푐푎푡 = (3) 푘4푘6 + 푘2푘4+푘2푘6

(푘 + 푘 )푘 푘 H2O2 −1 2 4 6 퐾푚 = (4) (푘4푘6 + 푘2푘4+푘2푘6)푘1

(푘 + 푘 )푘 푘 푘 푘 + (푘 + 푘 )푘 푘 푘 푘 퐴퐻2 −3 4 1 2 5 6 −5 6 1 2 3 4 퐾푚 = (5) 푘3푘5(푘4푘6 + 푘2푘4+푘2푘6)

In Scheme 3.3, the competitive binding between substrate AH2 and cosubstrate H2O2 in the distal pocket that gives rise to substrate inhibition is included in the initial rate equation

(Equation 6). The reaction sequence leading to peroxidase chemistry requires that ferric enzyme reacts with cosubstrate H2O2 first to form the active oxo-ferryl species compound

I/Compound ES. This mechanism does not rule out the possibility that substrate AH2 can bind

to the ferric enzyme in the distal pocket before H2O2 enters. Therefore, a reversible step that

2,4,6-TBP binds to ferric enzyme is included. Substrates that bind internally in the distal pocket will impede H2O2 binding to the heme iron to form the active oxo-ferryl intermediate for the following catalytic reaction (Scheme 3.3), because DHP possesses a well-defined internal substrate binding site in the distal pockets observed in several X-ray crystal structures as shown as an overlay in Figure 3.1.

Scheme 5.3. Proposed DHP peroxidase Bi-substrate Ping-Pong mechanism includes substrate AH2 inhibition

푘푐푎푡[퐸]0[퐻2푂2][퐴퐻2] 푉0 = (6) [퐴퐻 ] 퐾퐻2푂2 (1 + 2 ) [퐴퐻 ] + 퐾퐴퐻2[퐻 푂 ] + [퐻 푂 ][퐴퐻 ] 푚 퐴퐻2 2 푚 2 2 2 2 2 퐾푆퐼

Bench-top Mixing Kinetic Assays

The kinetic assays were conducted in 100mM, pH=7.0 KPi buffer using an Agilent

8453 UV-Visible spectrophotometer equipped with Peltier temperature controller. The

catalytic reactions were carried out in a 0.4 cm path length cuvette from Starna Cells, Inc with a total volume of 1200 µL. The final enzyme concentration [E]0 in each sample was 2.4µM for

DHP and 0.2 µM for HRP. All the kinetic assays were conducted at 25 oC. The enzyme and substrate (2,4,6-TCP, 2,4,6-TBP and ABTS) were first mixed in the optical cuvette, placed in the thermal cell to allow them to reach thermal equilibrium (3 min incubation). Subsequently, an aliquot of the H2O2 stock solution was added into the cuvette to initiate the reaction. The kinetic data were obtained by monitoring the absorbance at wavelength 272 nm for 2,4,6-TBP,

273 nm for 2,4,6-TCP and 414 nm for ABTS which corresponds to the absorbance peaks of the 2,6-dibromoquinone (DBQ), 2,6-dichloroquinone (DCQ) and ABTS cation radical

•+ -1 -1 -1 (ABTS ) products with extinction coefficients ε272nm = 14,000 M cm , ε273nm = 13,200 M

-1 -1 -1 cm and , ε414nm = 31,100 M cm respectively.

Fluoride Titration Assays

The fluoride titration experiments were conducted using an Agilent 8453 UV-visible spectrophotometer operating in the standard mode. The titrations were carried out in a 0.2 cm path length cuvette obtained from Starna Cells, Inc. The enzyme concentration [E]0 in each sample was 50 µM with a total volume of 600 µL. 1 mM substrates were included for the fluoride competitive binding titration and the cuvette was allowed to incubate for 3 min in the thermal cell to reach thermal equilibrium at 25 oC before the titration. Then the fluoride stock solution were gradually titrated in and the corresponding spectra were recorded. The fluoride titration spectra were eventually normalized according to the protein concentrations in the cuvette. The binding curve were extracted from the spectra region between 300 nm to 700 nm

using SVD (singular value decomposition) and then fitted into the one site reversible binding

app model to obtain the apparent fluoride binding constant Kd .

Data Analysis

The initial rate V0 of the kinetic data was calculated based on the method of initial rate.

This consists of a linear fit of the first 5 s of kinetic data collected using the photodiode array spectrophotometer. V0 was plotted against substrate concentration in the substrate dimension and correspondingly against H2O2 concentration in the H2O2 dimension. The kinetic data were then fit to the Michaelis-Menten equation to estimate the Vmax in the H2O2 dimension using Igor

Pro 6.10, because no substrate inhibition has been observed in the H2O2 dimension in either enzyme. The kinetic data was then plotted in a 3D format, in which two horizontal axis represents two variables that are substrate AH2 and cosubstrate H2O2 concentrations, and the vertical axis shows the responses that are initial rate. The data was then fitted into equation 2

(without substrate inhibition) or equation 6 (with substrate inhibition) to generate a kinetic rate surface using the multivariate fitting in Wolfram Mathematica 10 to obtain the kinetic parameters that are listed in Tables 3.1, 3.2 and 3.3.

3.4 Results

The data will first be presented as rate profiles along both the substrate and cosubstrate concentration dimension and subsequently as 3D surfaces showing both concentration dependences in a single plot. Since the rate depends on both the substrate and cosubstrate we can refer to them as complementary “substrates”, despite the fact that H2O2 is most often referred to as a cosubstrate. Starting with the standard 2D representation, Figure 3.2 shows the

plots of initial rate of DHP A catalyzed peroxidase reaction against the substrate 2,4,6-TCP concentration (0 ~ 1600 µM) and cosubstrate H2O2 concentration (0 ~ 2400 µM).

[H O ] [2,4,6-TCP] 12 2 2 12 1600 μM 2400 μM 1200 μM 1200 μM 800 μM

) 10 10 ) 1 600 μM 500 μM

1 -

s 300 μM 200 μM s

M 100 μM 8 8 M

6 6

0 0

1 6

(

0 0

4 V 4 V 2 2

0 a 0 b 0 400 800 1200 1600 0 500 1000 1500 2000

[2,4,6-TCP] (μM) [H2O2] (μM) 2.0 2.0 [H2O2] 2400 μM d 1200 μM

) 600 μM 1.5

1.5 )

1 -

300 μM - s

100 μM s

M 6

6 -

1.0 - 1.0 0

0 [2,4,6-TBP]

1 1

x 900 μM

x (

( 600 μM 0

0 480 μM V 0.5 V 0.5 360 μM 240 μM c 120 μM 0.0 0.0 0 200 400 600 800 0 500 1000 1500 2000 [2,4,6-TBP] (μM) [H2O2] (μM)

Figure 3.2 Kinetics of DHPA (2.4 M) catalyzed oxidation of 2,4,6-TCP or 2,4,6-TBP in the presence of H2O2. (a) 2,4,6- TCP dimension, (b) H2O2 dimension (with 2,4,6-TCP), (c) 2,4,6-TBP dimension, (d) H2O2 dimension (with 2,4,6-TBP) in the 100 mM KPi buffer at pH 7.0.

It is evident that the initial rate increases proportionally to the respective concentration, either

2,4,6-TCP or H2O2, and then levels off at sufficiently high concentration. However, under certain conditions there is a reduction in rate at the highest 2,4,6-TCP substrate concentrations.

This trend is most obvious for 2,4,6-TCP when the H2O2 concentration is kept constant at 100

M. A similar kinetic pattern can be observed for substrate 2,4,6-TBP with the added observation that the rate decreases for very high 2,4,6-TBP concentration as shown in Figure

3.2 and B1. The tendency of the rate to decrease at high 2,4,6-TXP concentration is observed

in both DHP A and B. The kinetic data show that the mutual effects of complementary substrates on the initial rate can be predicted by Equation S6 and S7 based on the model described in Scheme 3. The decrease in rate at high 2,4,6-TXP concentration is predicted by the model that includes the effect of substrate inhibition. Substrate inhibition is observed most prominently at low H2O2 concentration in the 2,4,6-TCP kinetics in agreement with the model.

No cosubstrate inhibition has been observed in the H2O2 dimension even when the H2O2 concentration is up to 1000 fold greater than the enzyme concentration. Similar substrate inhibition kinetic patterns have also been observed for substrate 2,4,6-TBP. DHP B shows the same trends as DHP A.

Table 3.2 Kinetic Parameters obtained from 2D Michaelis-Menten kinetics for DHP A

DHP A 2,4,6-TCP 2,4,6-TBP ABTS -1 kcat (s ) 13.67 1.83 1.05 H2O2 Km (mM) 0.335 0.212 0.029 AH2 Km (mM) 2.07 1.02 0.448 AH2 KSI (mM) 1.22 3.62

Table 3.2 Kinetic Parameters obtained from 2D Michaelis-Menten kinetics for DHP B

DHP B 2,4,6-TCP 2,4,6-TBP ABTS -1 kcat (s ) 25.72 1.87 0.70 H2O2 Km (mM) 0.165 0.141 0.025 AH 2 Km (mM) 0.685 0.312 0.229 AH2 KSI (mM) 0.325 1.34

Table 3.3 Kinetic Parameters obtained from 2D Michaelis-Menten kinetics for HRP

HRP 2,4,6-TCP 2,4,6-TBP ABTS34 -1 kcat (s ) 571.3 223.4 52.5 H2O2 Km (mM) 0.050 0.093 0.012 AH 2 Km (mM) 5.40 4.53 5.1

The kinetic parameters listed in Table 3.1 and Table 3.2 show that both DHP A and DHP B have higher catalytic rates for oxidation of 2,4,6-TCP than of 2,4,6-TBP. But unlike the native substrate 2,4,6-TBP, 2,4,6-TCP causes more severe substrate inhibition than 2,4,6-TBP

AH2 AH2 because its KSI is ~ 50% of its Km . The substrate inhibition produced by 2,4,6-TXP

(X=Cl, Br) is presumably due to the internal binding in the distal pocket in a position that

15,17 observed in the crystal structure, in which the substrate either blocks the H2O2 heme iron axial position for H2O2 to form compound 0 or impedes the entering of H2O2 into the distal pocket.

Figure 3.4 3D plot of kinetics of DHP (2.4 M) catalyzed oxidation of ABTS in the presence of H2O2. (a) DHP A with ABTS; (b) DHP B with ABTS. A well-studied peroxidase substrate, ABTS, was also used to test DHP’s peroxidase function. 34,35 As a typical peroxidase reducing substrate, ABTS is first oxidized to form

ABTS•+ by losing one electron, further one electron oxidation produces ABTS++. Because

ABTS has much larger molecular size than that of 2,4,6-TCP and 2,4,6-TBP, it has much less likely to bind internally in the distal pocket in DHP, which implies that no substrate inhibition should be observed for ABTS. Indeed, we have not observed any decline in the initial rate even when ABTS concentration is as high as 4 mM, which is ~ 1600 fold greater than the enzyme concentration in the kinetic assay. Moreover, because ABTS does not block H2O2 entrance, its

H2O2 Km is about 1/6 ~ 1/10 of that of 2,4,6-TXP (X = Cl, Br), which indicates much higher

H2O2 binding affinity.

The peroxidase function of DHP was also compared with that of HRP using the same ping-pong mechanism kinetic scheme.34,35 In the present study, we have included the complete

2D kinetic analysis for HRP using substrate 2,4,6-TCP and 2,4,6-TBP. Apparently, HRP has much higher catalytic rate ( ~ 20 to 120 fold) than that of DHP in terms of its peroxidase

AH2 function. However, its Km is ~ 3 to 20 fold larger than that of DHP, suggesting much weaker

substrate binding on the protein surface for HRP. Moreover, no substrate inhibition has been observed for any of these three substrates in HRP. Even at very low H2O2 concentration (20

M), the initial rate does not decline when 2,4,6-TCP concentration is increased to 3600 M as shown in Figure B3 and B4. As for 2,4,6-TBP, the kinetic measurement is limited by its solubility in aqueous solution and its denaturing effect. HRP precipitates by the end of the kinetic assay for concentrations of 2,4,6-TBP greater than 900 µM. Therefore, HRP’s kinetic profile is consistent with surface substrate binding that is distant from distal pocket and heme

Fe. Stability and folding studies show that DHP is must more resistant to precipitation by 2,4,6-

TXP. 23

app Table 3.4 Apparent Fluoride Dissociation Constant Kd of DHPA and HRP with or without Substrates

app Kd (mM) DHP A HRP f Protein (Kd ) 4.5 ± 0.1 69.5 ± 4.8 + 1mM 2,4,6-TCP 14.5 ± 0.8 74.9 ± 6.9 + 1mM 2,4,6-TBP 23.8 ± 1.0 97.4 ± 15.7 + 1mM ABTS 4.6 ± 0.3 71.4 ± 13.8

The fluoride titration experiments clearly show that DHP A and HRP have different substrate binding profiles. During the titration, the fluoride anion will bind to heme Fe to form the 6 coordinated high spin (6cHS) heme-fluoride adduct, which is indicated by the charge transfer band CT1 at 605 nm for DHP and 611 nm for HRP as shown in Figure 3.6. In the

f absence of any substrate, the fluoride anion dissociation constants are denoted as Kd . In the

app presence of substrate, we denote the apparent fluoride dissociation constant as Kd . Equation

app f S8 that describe the quantitative relationship of Kd in terms of Kd and substrate concentration

[L] is given in the Appendix B. As a result, the fluoride binding affinity might become much lower because of the steric hinderance exerted by internal substrate binding. Competitive binding of this type has been demonstrated for substrates 2,4,6-TBP and 2,4,6-TCP in DHP A.

app The Kd increase ~ 5 fold and ~ 3 fold respectively indicating the weaker binding affinity of

app fluoride anion. Whereas for HRP, Kd only increase ~ 40% and ~ 8% respectively, shows that both 2,4,6-TBP and 2,4,6-TCP have much weaker interactions with the fluoride anion, which suggest a likely external binding at HRP. Substrate ABTS shows a very small effect on the fluoride binding affinity for both DHP A and HRP. Because of the much larger molecular size of ABTS compared to 2,4,6-TBP, ABTS has very little possibility to bind internally in the

distal pocket in both DHP and HRP. Therefore, the ABTS binding site must be located on the protein surface of both proteins.

4 DHP A

2 HRPC f

d 1

)/K f

d 4

-K 2 app

d 0.1 (K 4 2 2,4,6-TCP 2,4,6-TBP ABTS

Figure 3.6 The fluoride titration profile of DHP A and HRP in the presence of substrate: 2,4,6-TCP, 2,4,6-TBP and ABTS.

The fluoride titration experiments are consistent with two modes of binding for 2,4,6-

TCP and 2,4,6-TBP. The fact that competitive binding with fluoride is observed indicates that internal binding of 2,4,6-TCP and 2,4,6-TBP in DHP does occur. However, the binding is sufficiently weak that it is consistent the existence of a external substrate binding site for DHP at low concentration of 2,4,6-TCP and 2,4,6-TBP, which is needed for DHP to carry out its peroxidase function. The fluoride titration result for HRP shows the much weaker interactions between substrates and heme-bound fluoride anion. This result suggests that there is no internal binding of substrates, which is consistent with the lack of substrate inhibition in HRP. This evidence leads us to suggest that the substrates 2,4,6-TXP bind at the -edge of the heme, like other aromatic substrates of HRP.36,37

3.5 Discussion

The Significance of External and Internal Substrate Binding Sites

The substrate inhibition observed in the DHP kinetics give new insight into the internal binding site of 2,4,6-TBP and 2,4,6-TCP observed in the X-ray crystal structures. 16,20These structures provide the first -edge substrate binding site observed in any heme peroxidase.

Normally the substrate either binds at the heme -edge in BHA binding sites in HRP, or binds at the γ-edge of the heme where propionate groups would play an important role in stabilizing the substrate binding.36 Therefore, it was a surprise to find that DHP possess an deeply buried substrate binding site for 2,4,6-TCP/2,4,6-TBP on the -edge of the heme.16,20 While the data presented here suggest that this binding site is an inhibitor binding site for peroxidase function, the main function of the binding site may relate to one of the other functions that has been recently discovered. For example, the peroxygenase function involves oxygen atom transfer and requires a substrate binding site very close to the heme Fe. 7 Our working hypothesis for

DHP function suggests that the internal binding binding site is primarily a 2,4-dibromophenol

(2,4-DBP) substrate binding site for peroxygenase function, but happens to also give rise to substrate inhibition for very high 2,4,6-TXP concentration

A further hypothesis is that DHP A and B have external binding sites for the peroxidase function, which would make them similar to HRP in that respect. Flow-EPR experiments suggest that the primary functional binding of 2,4,6-TCP is external to the protein based on the detected 2,4,6-trichlorophenoxyl radical intermediate.38 The phenoxyl radical can only be observed if the substrate rapidly dissociates from the protein and then flows through the EPR

flat cell used in the experiment. Evidence for radicals can be found both in the EPR spectra and the secondary products such as 3-hydroxy-2,6-dichloroquinone.39 The isotropic character of EPR spectra also shows that the radical is not associated with a high molecular weight protein but rather is free to react in bulk solution.38 This would not be consistent with internal binding because an internally bound substrate typically undergoes a direct 2-electron peroxidase oxidation reaction without involving a substrate radical intermediate. The kinetic data obtained using ABTS as a substrate further validates the hypothesis that the peroxidase mechanism in DHP A and B involves an external substrate binding site. The large size of ABTS prohibits substrate inhibition observed with the substrates 2,4,6-TCP and 2,4,6-TBP. These results seem to suggest that the internal binding site and substrate inhibition by 2,4,6-TXP may be adventitious and the main binding is on the protein surface.

The result from previous fluoride competitive binding assay suggests that the internal

-edge site may be the substrate binding site for 2,4-DBP.29 Consistent with this interpretation, the fluoride competitive binding assay shows that 2,4,-DBP has a much tighter binding affinity than 2,4,6-TBP in the distal pocket. Therefore, we hypothesize that the -edge internal substrate binding site of DHP evolved in order for DHP to carry out to a peroxygenase function another native substrate, 2,4-DBP. But it may also become a trap for 2,4,6-TBP in the distal pocket at sufficiently high substrate concentration and thereby result in substrate inhibition.

Kinetic Comparison between DHP and HRP: The Tortoise and the Hare

Based on numerous experimental observations we hypothesize that the multifunctional properties of DHP 7 are a consequence of the role it plays in detoxification combined

1 with an essential role in O2 transport. Detoxification in benthic ecosystems is a challenge because of the large number of potentially toxic substrates. 22,40 The evidence shows that DHP has an interaction with a number of different molecules based on both structural and dynamic observations. It will be of interest to individually examine each substrate and determine the factors that give rise to catalysis. Specifically, with regard to peroxidase activity the comparative kinetic study between DHP and HRP gives insight into the structure-function relationship required for efficacy in degradation of 2,4,6-TBP. Apparently, HRP is a superior peroxidase to DHP in terms of catalytic rate kcat for most of the reducing substrates studied, including 2,4,6-TBP. However, HRP is far from an ideal enzyme for DHP’s native substrate

2,4,6-TBP since HRP is destabilized by 2,4,6-TBP. HRP starts to aggregate and precipitate during the catalytic reaction when 2,4,6-TBP concentration reaches 900 µM as shown in Figure

B5. This denaturation and inactivation of HRP could be the result of attack of phenoxy radical generated during the catalytic reaction.41 DHP is like the tortoise that wins the race in the fable by being slow and steady in its pace. From an evolutionary perspective, HRP’s peroxidase function would not be suitable for a marine worm like Amphitrite ornata to survive in an environment that contains 2,4,6-TBP. Like the hare in the fable HRP is fast, but it cannot sustain that level of activity due to its sensitivity to denaturation by the substrate. From the bioremediation perspective, HRP needs to be either immobilized42,43 or protected by additives44 so that it can used as a stable enzymatic catalyst to treat pollutants that contain halogenated phenols. Apparently DHP has evolved a strategy to solve the toxicity problems posed by these substrate by having internal binding sites to regulate its peroxidase function.

Previous studies have shown that 4-hydroquinone and 4-bromophenol can inhibit the peroxidase function of DHP by binding at the internal inhibitor binding site.29,45

The multiple binding sites of DHP are unusual, but are not unique. In fact, DHP is an example of a growing list of proteins that have multiple functions or binding sites. For example, a plant peroxidase isolated from Chamaerops excelsa palm tree (CEP) with very strong pH and thermal stability also possesses significant substrate inhibition kinetic behavior which possibly can be attributed to the existence of two substrate binding sites.46,47 Current work extends structural studies that suggest that the substrates 2,4,6-TCP and 2,4,6-TBP inhibit

DHP’s peroxidase function by substrate inhibition at high substrate concentration, which prevents the collapse of the protein stability during the catalytic reaction at high substrate concentration. In conclusion, substrate inhibition of 2,4,6-TCP and 2,4,6-TBP distinct peroxidase function of DHP from that of HRP. The internal substrate binding due to the existence of a large distal pocket above the heme in DHP plays a critical role in regulating peroxidase function of DHP by substrate inhibition.

3.6 References

(1) Weber, R. E.; Mangum, C.; Steinman, H.; Bonaventura, C.; Sullivan, B.; Bonaventura,

J. In Comp. Biochem. Phys. A 1977; Vol. 56, p 179.

(2) Chen, Y. P.; Woodin, S. A.; Lincoln, D. E.; Lovell, C. R. J. Biol. Chem. 1996, 271,

4609.

(3) Franzen, S.; Thompson, M. K.; Ghiladi, R. A. Biochim. Biophys. Acta 2012, 1824, 578.

(4) D’Antonio, J.; Ghiladi, R. A. Biochemistry 2011, 50, 5999.

(5) Davydov, R.; Osborne, R. L.; Shanmugam, M.; Du, J.; Dawson, J. H.; Hoffman, B. M.

J. Am. Chem. Soc. 2010, 132, 14995.

(6) Du, J.; Sono, M.; Dawson, J. H. Biochemistry 2010, 49, 6064.

(7) Barrios, D. A.; D’Antonio, J.; McCombs, N. L.; Zhao, J.; Franzen, S.; Schmidt, A. C.;

Sombers, L. A.; Ghiladi, R. A. J. Am. Chem. Soc. 2014, 136, 7914.

(8) Kvist, M.; Ryabova, E.; Nordlander, E.; Bülow, L. J. Biol. Inorg. Chem. 2007, 12, 324.

(9) George, P.; Irvine, D. H. Nature 1951, 168, 164.

(10) Lebioda, L. Cell. Mol. Life. Sci. 2000, 57, 1817.

(11) Wan, L.; Twitchett, M. B.; Eltis, L. D.; Mauk, A. G.; Smith, M. Proc. Natl. Acad. Sci.

1998, 95, 12825.

(12) Ozaki, S.-i.; Matsui, T.; Watanabe, Y. J. Am. Chem. Soc. 1996, 118, 9784.

(13) Lin, Y.; Wang, J.; Lu, Y. Sci. China. Chem. 2014, 57, 346.

(14) Ordway, G. A.; Garry, D. J. J. Exp. Biol. 2004, 207, 3441.

(15) Thompson, M. K.; Davis, M. F.; de Serrano, V.; Nicoletti, F. P.; Howes, B. D.;

Smulevich, G.; Franzen, S. Biophys. J. 2010, 99, 1586.

(16) Zhao, J.; de Serrano, V.; Zhao, J. J.; Le, P.; Franzen, S. Biochemistry 2013, 52, 2427.

(17) De Serrano, V.; Franzen, S. Peptide Sci. 2012, 98, 27.

(18) Davis, M. F.; Gracz, H.; Vendeix, F. A. P.; de Serrano, V.; Somasundaram, A.; Decatur,

S. M.; Franzen, S. Biochemistry 2009, 48, 2164.

(19) de Serrano, V.; Chen, Z. X.; Davis, M. F.; Franzen, S. Acta Crystal. D-Biol. Cryst.

2007, 63, 1094.

(20) Wang, C.; Lovelace, L. L.; Sun, S.; Dawson, J. H.; Lebioda, L. Biochemistry 2013, 52,

6203.

(21) Lebioda, L.; LaCount, M. W.; Zhang, E.; Chen, Y. P.; Han, K.; WHitton, M. M.;

Lincoln, D. E.; Woodin, S. A. Nature 1999, 401, 445.

(22) Fielman, K. T.; Woodin, S. A.; Walla, M. D.; Lincoln, D. E. Marine Ecology-Progress

Series 1999, 181, 1.

(23) Le, P.; Zhao, J.; Franzen, S. Biochemistry 2014, 53, 6863.

(24) Dunford, H. B. Peroxidases in Chemistry and Biology Vol. 2 (Everse, J., Everse, K. E.,

& Grisham, M. B., Eds) CRC Press, Boca Raton, FL 1991, 1.

(25) Bakovic, M.; Dunford, H. B. Biochemistry 1994, 33, 6475.

(26) Belyea, J.; Gilvey, L. B.; Davis, M. F.; Godek, M.; Sit, T. L.; Lommel, S. A.; Franzen,

S. Biochemistry 2005, 44, 15637.

(27) Osborne, R. L.; Taylor, L. O.; Han, K. P.; Ely, B.; Dawson, J. H. Biochem. Biophys.

Res. Comm. 2004, 324, 1194.

(28) Ferrari, R. P.; Laurenti, E.; Trotta, F. J. Biol. Inorg. Chem. 1999, 4, 232.

(29) Zhao, J.; Franzen, S. J. Phys. Chem. B 2013, 117, 8301.

(30) Ma, H.; Thompson, M. K.; Gaff, J.; Franzen, S. J. Phys. Chem. B 2010, 114, 13823.

(31) Dunford, H. B.; Stillman, J. S. Coord. Chem. Rev. 1976, 19, 187.

(32) Poulos, T. L.; Kraut, J. J. Biol. Chem. 1980, 255, 8199.

(33) Baek, H. K.; Van Wart, H. E. Biochemistry 1989, 28, 5714.

(34) Rodriguez-Lopez, J. N.; Smith, A. T.; Thorneley, R. N. F. J. Biol. Chem. 1996, 271,

4023.

(35) Rodríguez-López, J. N.; Gilabert, M. A.; Tudela, J.; Thorneley, R. N. F.; García-

Cánovas, F. Biochemistry 2000, 39, 13201.

(36) Gumiero, A.; Murphy, E. J.; Metcalfe, C. L.; Moody, P. C. E.; Raven, E. L. Arch.

Biochem. Biophys. 2010, 500, 13.

(37) Henriksen, A.; Smith, A. T.; Gajhede, M. J. Biol. Chem. 1999, 274, 35005.

(38) Sturgeon, B. E.; Battenburg, B. J.; Lyon, B. J.; Franzen, S. Chem. Res. Toxicol. 2011,

24, 1862.

(39) Franzen, S.; Sasan, K.; Sturgeon, B. E.; Lyon, B. J.; Battenburg, B. J.; Gracz, H.;

Dumariah, R.; Ghiladi, R. J. Phys. Chem. B 2012, 116, 1666.

(40) Lincoln, D. E.; Fielman, K. T.; Marinelli, R. L.; Woodin, S. A. Biochem. Syst. Ecol.

2005, 33, 559.

(41) Huang, Q.; Huang, Q.; Pinto, R. A.; Griebenow, K.; Schweitzer-Stenner, R.; Weber,

W. J. J. Am. Chem. Soc. 2005, 127, 1431.

(42) Alemzadeh, I.; Nejati, S. J. Hazard. Mater. 2009, 166, 1082.

(43) Huang, J.; Chang, Q.; Ding, Y.; Han, X.; Tang, H. Chem. Eng. J. 2014, 254, 434.

(44) Wu, Y.; Taylor, K. E.; Biswas, N.; Bewtra, J. K. Enzyme. Microb. Tech. 1998, 22, 315.

(45) Zhao, J.; Zhao, J.; Franzen, S. J. Phys. Chem. B 2013, 117, 14615.

(46) Cuadrado, N. H.; Arellano, J. B.; Calvete, J. J.; Sanz, L.; Zhadan, G. G.; Polikarpov, I.;

Bursakov, S.; Roig, M. G.; Shnyrov, V. L. J. Mol. Catal. B: Enzymatic 2012, 74, 103.

(47) Bernardes, A.; Textor, L. C.; Santos, J. C.; Cuadrado, N. H.; Kostetsky, E. Y.; Roig,

M. G.; Bavro, V. N.; Muniz, J. R. C.; Shnyrov, V. L.; Polikarpov, I. Biochimie 2015, 111, 58.

CHAPTER 4

Kinetic Study of the Inhibition Mechanism of DHP A

by para-Halogenated Phenol

4.1 Abstract

The mechanism of dehaloperoxidase-hemoglobin (DHP) inhibition by para-halogenated phenol were investigated using Michealis-Menten and transient-state kinetic analyses.

Transient-state kinetics using the stopped-flow technique to mix DHP and H2O2 in the presence of inhibitor concentrations less than 10-fold greater than the enzyme concentration show that

4-BP does not fully impede H2O2 entering the distal pocket to activate DHP. It is not clear whether an oxoferryl intermediate is formed under these conditions and there may be alternative pathways for H2O2 reaction in the 4-BP bound form of DHP. Two new species have been identified during the reaction of 4-BP bound form DHP in the transient-state kinetic experiment by using Singular Value Decomposition (SVD) and global-fitting analysis. Rather than forming Compound ES in the unbound form, an inhibitor bound intermediate that possesses blue-shifted Soret band and a double peaked Q band is observed. This intermediate is subsequently converted to the end-point species that is distinguished from Compound RH formed in the uninhibited enzyme. Bench-top mixing kinetics of DHP were conducted in order to determine the inhibitor binding constant and to understand the enzyme inhibition mechanism from a thermodynamic perspective. Inhibition constant, Ki of 4-bromophenol (4-BP) and has been measured over the temperature range from 283 K to 298 K, which permits determination of the enthalpy and entropy for inhibitor binding respectively, leading to the conclusion that inhibitors binding are entropically driven .

4.2 Introduction

Dehaloperoxidase-hemoglobin A (DHP A) isolated from Amphitrite ornata is the first known hemoglobin with a biologically relevant peroxidase function for 2,4,6-tribromophenol

(2,4,6-TBP) as a substrate.1 There are two globin genes in A. ornata, and the second gene product DHP B is a three-fold better peroxidase than DHP A.2 Both DHP A and DHP B

(collectively called DHP) appear to have multiple functions, which has made them the focus of intensive investigation. While it is not possible yet to determine how the function of DHP relates to the evolution of function in the peroxidase superfamily,3 inhibition by 4- bromophenol (4-BP) provides one unique feature of DHP that provides a crucial clue as to the functional origins of the enzyme.4 Due to its ability to oxidize halophenols and other halognated substrates in the presence of H2O2 in nature, DHP also has a potential application in bioremediation strategies to degrade halogenated organic pollutants produced by anthropogenic activity.5 However, it is imperative to understand the role of inhibition in order to optimize the activity of DHP both to understand its physiological role in A. ornata and for bioremediation applications.

Previous studies have also shown that trihalophenols (2,4,6-TXP), such as 2,4,6-TBP,

2,4,6-TCP and 2,4,6-TFP are significantly better substrates than the monohalophenols, 4-BP,

4-CP and 4-FP.6 In fact, 4-BP has very low turnover and can act as a potent inhibitor for other substrates.7 This point is relevant to the biological function of DHP since the native substrate

2,4,6-TBP and apparent inhibitor 4-BP are present in benthic ecosystems in a ratio of approximately 1:2.8 Recent study reveals the diameter of the halogen atom in the para position plays a crucial role in the binding of 4-halophenols in the distal pocket.9 4-halophenols with

smaller halogen atoms, chlorine or fluorine in the para position do not bind as well, and thus

4-CP and 4-FP are not as effective as inhibitors. For example, although 4-BP is an excellent inhibitor and poor substrate, 4-CP shows considerable activity as a substrate7,9. The biological significance can be approached by first understanding the thermodynamics and kinetics of inhibitor binding, which motivates the present study.

Figure 4.1 The structure of DHP in presence of 4-BP as determined by X-ray crystallography (PDB 3LB2).

The previously proposed inhibition mechanism that involves 4-BP binding in the distal pocket has consequences for peroxidase function.10 When 4-BP is bound in the internal binding site above the heme the distal histidine, His55, is displaced to a solvent-exposed conformation, in which distal His55 is too far away from the heme center to function as an acid-base catalyst that can facilitate O-O heterolysis of H2O2. Heterolytic bond cleavage is essential for the formation of Compound I or Compound ES, which is the first intermediate formed in typical peroxidases such as horseradish peroxidase (HRP)11,12 or cytochrome c peroxidase (CcP).13,14

DHP has been shown to form a Compound ES intermediate when substrate is not present,15 which has been shown to involve one or more tyrosine radicals.2,16 Based on the kinetic data it appears that the tyrosine radical observed transiently in DHP is not an electron transfer intermediate, but rather may play a protective role. The end point for the radical chemistry in the absence of substrate is an inactive (or at least severely impaired) cross-linked heme species called compound RH.17 The relationship between the internal radical pathways and compound

RH formation with normal function has not been established.

To probe the 4-BP inhibition mechanism of DHP, we have conducted transient-state kinetic study of reaction between ferric enzyme and H2O2 for a range of 4-BP concentrations using a stopped-flow UV-visible spectrophotometer. Two parallel reactions with 4-BP bound and unbound DHP were observed using singular value decomposition (SVD) and global fitting analysis of stopped-flow data. Thus, a novel inhibition mechanism was proposed based on the presence of this equilibrium. A comprehensive bench-top kinetic study of DHP catalytic process in the presence of the inhibitor 4-BP was conducted at variety of temperatures following methods developed previously. The bench-top kinetic data have been fit to a simplified inhibition kinetic model in order to determine the inhibition constant Ki and its temperature dependence, which provides the thermodynamics of the binding of 4-BP to DHP.

The kinetic measurements and analysis presented here provide further support for the novel inhibition mechanism of DHP.

4.3 Material and Methods

Materials

All the reagents and biochemicals were purchased from Aldrich and ACROS and used without further purification. 2,4,6-Trichlrophenol (TCP), 4-Bromophenol (4-BP), 4-

Chlorophenol (4-CP) and 4-Fluorophenol were each dissolved in 100mM, pH=7 potassium phosphate ( KPi ) buffer to prepare the substrate and inhibitor solution. Prepared solutions were stored at 4 oC and protected against light. The concentration of each solution was measured

-1 -1 prior to each kinetic experiment by monitor its absorbance : TCP, ε312nm = 3,752 M cm ; 4-

-1 -1 -1 -1 -1 -1 BP, ε280nm = 1,370 M cm ; 4-CP, ε280nm = 1,890 M cm and 4-FP, ε278nm = 2,230 M cm in Agilent 8453 diode array UV-Visible spectrophotometer at 25 oC. Hydrogen peroxide solution was freshly made before each kinetic experiment. A 30% reagent grade hydrogen peroxide (H2O2) solution was added to 100 mM, pH=7 KPi buffer to make a 7.2 mM stock solution. The hydrogen peroxide solution was kept on ice and protected against light during the experiment. Wild-type

16 His6X DHP A was expressed in E.coli and purified as previously described. Ferric

DHP was oxidized by excess K3[Fe(CN)6] and then filtering through Sephadex G-25 column to eliminate excess K3[Fe(CN)6], and further purified on CM-52 prior to each kinetic experiment. The concentration of ferric DHP was determined by using the Soret band molar

-1 -1 absorption coefficient, ε406nm = 116,400 M cm .

Transient-state kinetic assays

Experiments were performed on a Bio-Logic SFM-400 triple-mixing stopped-flow instrument equipped with a diode array UV-visible spectrophometer and were carried out at room temperature in 100 mM KPi buffer, pH 7. Data were collected over three time-domain regimes (2.5, 25, and 250 ms; 300 scans each) using the Bio Kinet32 software package (Bio-

Logic). Data were collected (900 scans total) over a three-time domain regime (2.5, 25, and

250 ms; 300 scans each, 83.25 s total) using Bio Kinet32 (Bio-Logic). Single–mixing experiments were performed, in which ferric DHP were preincubated with 4-BP prior to the mixing with H2O2. The final concentrations after mixing were [DHP] = 10 µM , [H2O2] = 100

µM , and [4-BP] ranging from 0, 10 ,20 ,100 µM .

Bench-top mixing kinetic assays

The kinetic assays were conducted in 100mM, pH=7 KPi buffer using an Agilent 8453

UV-Visible spectrophotometer equipped with Peltier temperature controller. The catalytic reactions were carried out in a 0.4 cm path length cuvette obtained from Starna Cells, Inc with a total volume of 1200 µL. The ferric DHP concentration [E]0 in each sample was 2.4µM.

Substrate, 2,4,6-TCP concentrations ranged from 200 to 1500 µM and inhibitor, 4-BP,4-CP or

4-FP concentrations ranged from 0 to 500 µM. The substrate and inhibitor were first mixed with ferric DHP and KPi buffer and then allowed to incubate for 3 min in the cuvette placed in the thermal cell to reach thermal equilibrium. Subsequently, 1200 µM of H2O2 solution was added into the cuvette to initiate the reaction. The kinetic data were measured by monitoring

the absorbance at wavelength 273 nm, which corresponds to the absorbance peak of the 2,6-

-1 -1 dichloroquinone (DCQ) product, with a molar absorption coefficient ε273nm = 13,200 M cm .

Data analysis

Spectra measured using a stopped-flow kinetic assay were analyzed using the SVD method, which provides a decomposition of the original absorption data matrix A(,t) in terms of basis spectra as the product of three matrices USVT.18,19 As described in detail in the

Appendix C, the first three vectors in the VT matrix (Figure C1) correspond to the time-courses of the reaction orthonormal basis spectra given in the U matrix (Figure C2) for the reaction in the absence of 4-BP. The time courses were globally-fit to a biexponential function according to the proposed two-step three-species first order reaction mechanism shown in scheme 1.

From which the rate constants k1 , kRH and the 3×3 C coefficient matrix were determined.

Scheme 4.6 Kinetic model for the global-fitting analysis of the stopped-flow data in the absence of 4-BP

In the presence of 4-BP, the VT matrix (Figure C3) was evaluated as two parallel two step reactions since there are two populations of DHP A proteins. Population 1 has a bound 4-

BP inhibitor and population 2 does not. Each of the populations has the same two processes shown in Scheme 1, but they may have different first order reactions, therefore there are four rate constants possible as shown in Scheme 2. The data were globally fit to a fourth-exponential function according to the proposed kinetic model in scheme 2, from which the rate constant k1, kRH , ki1 , kiRH and the 3×5 C coefficient matrix was determined. The basis spectra, b-spectra

(Figure C3) were calculated through B = USC and the spectra corresponding to each intermediate were reconstructed based on the analytical solution of each kinetic model. The

SVD and global fitting analysis were performed using Igor Pro 6.04.

Scheme 4.2 Kinetic model for the global-fitting analysis of the stopped-flow data in the presence of 4-BP

The initial rates of a series of substrate and inhibitor concentrations were measured as a function of temperature. The slopes of experimental progress curves were determined using the method of initial rates by linear fit of the first ten time points to provide V0. The initial rates V0 in presence of inhibitor 4-BP, 4-CP or 4-FP were then analyzed according to the

Equation 3.1, which was derived from proposed para-halogenated inhibited ping-pong mechanism based on the steady state assumption. Equation 3.1 was then simultaneously fit to all measured initial rates at different substrate and inhibitor concentrations using nonlinear

app regression. Vmax was fit globally according to the proposed inhibition mechanism. Km was

app then determined at each inhibitor concentration. Then a linear fit of Km against inhibitor concentration [4-XP] was conducted to determine the Ki. All the data analysis and fitting was conducted using Igor Pro 6.04.

4.4 Results

Stopped-flow kinetics of reaction between ferric DHP and H2O2 in the presence of 4-BP

The reactions of DHP with H2O2 in the presence or absence of 4-BP were monitored by using a single-mixing stopped-flow UV-visible spectrophotometer. It has been shown that ferric DHP reacts with H2O2 forming Compound ES, an iron(IV)-oxo species with an amino acid radical located on one or more of the three tyrosines, Y28, Y34 and Y3817 in a pH- dependent mechanism.16 The iron(IV)-oxo porphyrin π-cation radical species Compound I has not been observed in wild-type DHP for this reaction, although it has been observed in the

Y28F,Y34F,Y38F triple mutant of DHP (Data unpublished). In the absence of inhibitor 4-BP, ferric DHP reacted with 10 equivalents of H2O2 at pH 7.0 yielding Compound ES. Compound

ES formation reaches a maximum 2 seconds after mixing, and subsequently an inactivated

DHP species, known as Compound RH, was formed on a time scale of 80 seconds (Figure

4.2a). The well-defined features of Compound ES have been characterized based on the spectra extracted from the SVD analysis (see Methods) in comparison to previous work.15 The

Soret band of compound ES is observed at 420 nm and a distinctive double-peaked Q band is also evident with  and  bands at 546 and 586 nm, respectively. Compound RH has a Soret band at 411 nm and has a broad and featureless Q band in the visible region with a max at 540 nm.

Table 4.3 Kinetic parameters obtained from global-fitting of evolutionary time-course

-1 -1 -1 -1 [4-BP] (µM) k1 ( s ) kRH ( s ) ki1 ( s ) kiRH ( s ) 0 2.78 ± 0.78 0.031 ± 0.001 10 2.59 ± 0.10 0.017 ± 0.005 5.67 ± 2.43 0.276 ± 0.072 20 3.09 ± 0.19 0.041 ± 0.001 6.29 ± 0.44 0.393 ± 0.009 100 2.88 ± 1.51 0.073 ± 0.020 6.35 ± 0.90 0.536 ± 0.040

Upon preincubation of DHP with 1~10 equivalents of 4-BP and initiation of the reaction with 10 equivalents of H2O2 , two parallel pathways are observed, which are distinguished as the 4-BP bound form and the unbound form. The origin of the two parallel pathways is simply that DHP is in equilibrium with 4-BP and there are two populations at the instant that the reaction is initiated by H2O2. As expected, we observe that 4-BP unbound form has the same kinetic parameters observed previously for the reaction of DHP with H2O2. Both Compound

ES and Compound RH have been observed in the reconstructed spectrum (Figure 4.2 b1,c1,d1) with similar rate constants, k1 and kRH obtained from the fitting (Table 4.1). On the other hand, in the 4-BP bound form, a new species was observed that also possessed a double-peaked Q band at 541 and 584 nm (1 equivalent 4-BP), 543 and 585 nm (2 equivalent 4-BP), and 541 and 583 nm (10 equivalent 4-BP), while the Soret band at 408.4 nm (1 equivalent 4-BP), 417 nm (2 equivalent 4-BP) and 411 nm (10 equivalent 4-BP). The different Soret bands obtained from calculated spectra may imply the existence of equilibrium between Compound ES and

Compound iOX in the presence of 4-BP, which introduces uncertainty in the determination of the linear combinations of b-spectra to reconstruct the intermediate spectra. Since the Soret bands in the 4-BP-bound species are not the same as those for uninhibited DHP, we refer to the intermediate in this 4-BP-bound form as Compound iOX. The nomenclature is based on the fact that the intermediate are in equilibrium with the intermediate compound ES when inhibitor bound in the distal pocket, but does not necessarily mean both intermediates share the oxo-ferryl heme structure.

The spectrum of the end-point for the assay also shows that another new species with a

Soret band at 407.5 nm and distinct charge-transfer band at 622 nm has formed. This form has

been named Compound iRH (Figure 4.2-b2,c2,d2). The names compound iOX and iRH imply that the most significant difference with respect to compound ES and RH is the presence of the inhibitor in the distal pocket. Since little or no ferryl intermediate is formed when 4-BP is bound in the distal pocket, we have used the designation iOX to signify that there is oxidation by H2O2, but it may involve a different mechanism. Both ki1 and kiRH that corresponds to the formation of Compound iOX and Compound iRH slightly increase as the concentration of 4-

BP increases. Compared with k1 and kRH, ki1 is about 2-fold larger than k1, and kiRH is about 10- fold larger than kRH.

Bench-top mixing kinetics with substrate 2,4,6-TCP and inhibitor 4-BP

Scheme 4.3 Ping Pong scheme for the reaction of DHP with substrate TCP. The rate scheme incorporates non-classical competitive inhibition in presence of inhibitor 4-BP.

The kinetic model given in Equations 3.1 – 3.3 were derived from combination of previous DHP catalytic kinetic model16 and the inhibition mechanism proposed above

(Scheme 3.3). The previous DHP catalytic kinetic model is an application of the steady-state approximation to both compound ES and compound II formation in the classic peroxidase rate scheme. The steady-state approximation was invoked to derive a simplified ping pong scheme

with non-classical competitive inhibition. In the inhibition model, we believe that the inhibition does not take place at step when 4-BP binds to ferric DHP, but rather at the step where

Compound iOX/ Compound II···4-BP form. Kia and Kib correspond to the 4-BP dissociation constants of Compound iOX and Compound II···4-BP, respectively. Since Compound ES and

Compound II are two subsequent active enzyme intermediates during the two consecutive one- electron steps, we cannot distinguish between the binding and inhibitory effects of 4-BP on these two intermediates in our steady-state inhibition kinetic analysis. Therefore, we propose that the binding interaction of 4-BP will exert the same inhibitory effects on both compounds

ES and II, which leads to the assumption that Kia = Kib = Ki. Equation 1 was derived based on the steady state approximation applied to Compound iOX and Compound II···4-BP, which are the forms of compounds ES and II in the presence of inhibitor 4-BP.

푘 [퐸] [푇퐶푃] 푉 = 푐푎푡 0 (1) 0 [4– 퐵푃] 퐾푚 (1 + ) + [푇퐶푃] 퐾𝑖

푘푐푎푡 = 푘1[퐻2푂2] (2)

1 1 퐾푚 = 푘1[퐻2푂2] ( + ) (3) 푘2 푘3

Figure 4.3 Bench-top mixing kinetic analysis of DHP catalyzed TCP oxidation reaction inhibited by 4-BP at different o o o o temperatures (a) 25 C( b) 20 C (c) 15 C (d) 10 C. Kinetic assay conditions were ferric DHP = 2.4 µm, H2O2 = 1200 µm in 100 mM KPi buffer, pH 7.0.

Figure 4.3 presents the Michaelis-Menten curves of initial velocity V0 vs substrate concentration, [TCP], at 25 oC, 20 oC, 15 oC and 10 oC. The inhibition effects of 4-BP on the

DHP catalyzed substrate oxidation reaction were observed at each temperature. V0 decreases as the inhibitor concentration is increased. However, the pattern of the Michaelis-Menten curves tends to converge to the uninhibited Michaelis-Menten curve as the temperature is decreased, which indicates a decrease in the inhibitory effect due to the temperature dependence of the inhibition constants, Ki, which are reported in Table 4.2. The values in

Table 4.2 show that Ki decreases significantly as temperature increases, which means that the inhibitor is more potent at higher temperature.

Table 4.2 Michaelis-Menten parameters from bench-top mixing 4-BP inhibition kinetic assay

-1 -1 -1 T (K) kcat (s ) kcat /Km (s mM ) Ki (mM) 283 2.57 ± 0.15 1.54 ± 0.02 2.560 ± 0.036 288 5.07 ± 0.25 2.04 ± 0.02 1.735 ± 0.016 293 6.40 ± 0.28 3.95 ± 0.19 0.490 ± 0.024 298 10.42 ± 0.70 7.12 ± 1.57 0.155 ± 0.034

van’t Hoff analysis of the inhibition constant Ki

By measuring Ki at a series of temperatures, ln(Ki) vs 1/T in a van’t Hoff plot was used to

Θ Θ calculate ΔH and ΔS . Figure 4.4 shows a plot of ln(Ki) vs 1/T based on van’t Hoff equation,

Θ Θ Θ Θ Θ -ln(Ki) = ΔG /RT= ΔH /RT- ΔS /R. Thus ΔH = -135.5 ± 20.9 kJ/mol and ΔS = -526.1 ± 71.9

J/(mol•K) that corresponding to the enzyme-inhibitor complex dissociation process. The bench-top mixing kinetics shows that binding of 4-BP to DHP has an unfavorable positive enthalpy and a favorable positive entropy, which informs us that the binding process is entropy- driven. The free energy at temperatures of physiological interest are ΔGΘ =-21.28 kJ/mol at

T=298K.

-6.0

) i -7.0

ln( -8.0

-9.0 -3 3.40 3.45 3.50x10 1/T(1/K)

Figure 4.4 van’t Hoff plot of of ln(Ki) vs 1/T for inhibitor 4-BP binding to ferric DHP. The Ki value was determined at pH = 7.0 in 100mM KPi buffer. 2,4,6-TCP was used as the substrate.

Bench-top mixing kinetics with substrate 2,4,6-TCP and inhibitor 4-CP

12 [4-CP] 0 μM 10 100 μM

) 200 μM -1 8 350 μM

500 μM Ms

-6 6 (*10

0 4 V 2 o 25 C 0 0 400 800 1200 [TCP] (μM)

Figure 4.5 Bench-top mixing kinetic analysis of DHP catalyzed TCP oxidation reaction inhibited by 4-CP at 25 oC. Kinetic assay conditions were ferric DHP = 2.4 µm, H2O2 = 1200 µm in 100 mM KPi buffer, pH 7.0.

Table 4.3 Michaelis-Menten parameters from bench-top mixing 4-CP inhibition kinetic assay

-1 -1 -1 T (K) kcat (s ) kcat /Km (s mM ) Ki (mM) 298 9.02 ± 0.63 6.40 ± 0.90 0.268 ± 0.026

4-CP has shown little turnover in the presence of H2O2, thus it was also considered as a competitive inhibitor that similar to 4-BP. The same inhibition mechanism (Scheme 4.3) has been proposed for 4-CP as well. Therefore, the Michaelis-Menten curves have been globally fitted into Equation 1, giving the Michaelis-Menten parameters listed in Table 4.3. The Ki of

4-CP is 0.268 mM, which is about 70% larger than that of 4-BP, indicating that 4-CP serves as a poorer competitive substrate compared to 4-BP. This trend is also consistent with increasing Kd value of 4-CP compared to that of 4-BP.

Bench-top mixing kinetics with substrate 2,4,6-TCP and inhibitor 4-FP

0.8 1,4-Benzoquinone 246 nm 0.6

0.4

Absorbance 0.2

0.0

250 300 350 400 450 500 Wavelength(nm)

O Figure 4.6 DHP A (2.4 µM) catalyzed oxidation of 4-FP (500 µM) in the presence of H2O2 (1200 µM) at 25 C, forming the product 1,4-benzoquinone.

4-FP has shown considerable turnover catalyzed by DHP in the presence of H2O2, yielding the major product 1,4-BQ and some polymerized byproduct. Apparently, 4-FP behaves quite different compared to the 4-BP and 4-CP. Before we study the inhibition effect of 4-FP to the DHP catalyzed oxidation of TCP, we need to investigate the kinetic behavior of

4-FP as a unique monohalogenated substrate of DHP. Thus, a temperature-dependent

Michaelis-Menten kinetic assay has been conducted using 4-FP solely as the substrate. The kcat and Km of 4-FP ranging from 283 K to 298 K have been given in the Table 4.4. The activation energy for this catalytic process has been obtained to be Ea = 50.08 ± 6.33 kJ/mol by conducting the Arrhenius plot of ln(kcat/Km) vs 1/T. The low kcat indicates that 4-FP is indeed a poor substrate compared to TCP. Therefore, 4-FP falls into a different category in terms of inhibition.

Figure 4.7 (a) Temperature-dependent DHP A (2.4 µM) catalyzed oxidation of 4-FP in the presence of H2O2 (1200 µM) and (b) Arrhenius plot of kcat/Km vs 1/T.

Table 4.4 Michaelis-Menten parameters of substrate 4-FP

-1 -1 -1 T (K) kcat (s ) kcat /Km (s mM ) 283 0.409 ± 0.009 1.49 ± 0.14 288 0.455 ± 0.010 2.47 ± 0.29 293 0.498 ± 0.007 2.91 ± 0.22 298 0.596 ± 0.008 4.64 ± 0.42

4-FP functions as a competitive substrate rather than competitive inhibitor to DHP. It inhibits DHP as an alternative substrate with a tighter binding affinity than TCP in the distal pocket, but a much lower turnover rate. However, we cannot distinguish this type of inhibition from the competitive inhibition exerted by 4-BP and 4-CP by Michaelis-Menten kinetic assay.

Because the derived expression of the Michaelis-Menten equations are mathematically the

same as we can compare. The Ki in Equation 3.1 has been replaced by Km 4-FP in Equation 3.4.

Km 4-FP is no longer the inhibition constant but the Michaelis-Menten parameter for the enzyme

4-FP complex. Km 4-FP has been obtained by globally fitted the Michaelis-Menten curves at

different 4-FP concentration. Compared the Km 4-FP (0.159 mM) obtained from inhibition

kinetic analysis and the Km (0.128 mM) from which 4-FP used as substrate only, the relatively consistent values prove the validity of the kinetic scheme for 4-FP.

푉푚푎푥[푇퐶푃] 푉0 = 푎푝푝 (4) 퐾푚 + [푇퐶푃]

푉푚푎푥 = 푘1[퐻2푂2][퐸]0 (5)

푎푝푝 퐾푚 = 훼퐾푚푇퐶푃 (6)

[4 − 퐹푃] 훼 = 1 + (7) 퐾푚4−퐹푃

1 1 퐾푚4−퐹푃 = 푘1[퐻2푂2] ( + ) (8) 푘4 푘5

1 1 퐾푚푇퐶푃 = 푘1[퐻2푂2] ( + ) (9) 푘2 푘3

Scheme 4.4 Ping Pong scheme for the reaction of DHP with substrate TCP, incorporates with 4-FP that serves as competitive substrate.

12 [4-FP] 10 0 μM

) 100 μM -1 8 200 μM

350 μM Ms

-6 6 500 μM

(*10 4

0 V 2 o 25 C 0 0 400 800 1200 [TCP] (μM)

Figure 4.8 Bench-top mixing kinetic analysis of DHP catalyzed TCP oxidation reaction inhibited by 4-FP at 25 oC. Kinetic assay conditions were ferric DHP = 2.4 µm, H2O2 = 1200 µm in 100 mM KPi buffer, pH 7.0.

Table 4.5 Michaelis-Menten parameters from bench-top mixing 4-FP inhibition kinetic assay

-1 -1 -1 T (K) kcat (s ) kcat /Km (s mM ) Km 4-FP (mM) 298 9.19 ± 0.49 6.13 ± 0.64 0.159 ± 0.015

4.5 Discussion

DHP is a hemoglobin that exhibits levels of peroxidase activity comparable to typical peroxidases.20 DHP has been shown to be able to oxidize a variety of substituted phenols, such as the trihalophenols, dihalophenols, 4-chlorophenol, 4-fluorophenol, and a number of other substrates. However, the inhibition of DHP by 4-BP provokes an intriguing mechanistic question. How does DHP function in ecosystems where 4-BP is prevalent? The inhibitor 4-BP is a brominated secondary metabolite secreted by marine polychaetes that coexist with A. ornata in greater concentration than the substrate 2,4,6-TBP.8 The previously proposed inhibition mechanism placed emphasis on the internal distal pocket as the inhibitor binding site and the subsequent impact on the conformation of distal His55.7 The distal His55 has been shown to be essential for catalysis.21,22 Moreover, His55 has also been shown to be unusually flexible in DHP, when compared to other globins, meaning that it exists nearly equally in two conformers at pH 6.23 The existence of internal (closed) and external (open) conformers in globins such as Sperm Whale myoglobin has been shown to favor the internal conformation unless the pH is lowered to pH < 4.5.24 One possible cause for inhibition in DHP is that 4-BP binding forces His55 into the open conformation that is distant from the heme iron center, and thus no longer being able to function as an acid-base catalyst to facilitate heterolytic O-O bond cleavage. Therefore, significantly lower turnover of 4-BP bound ferric DHP to the active oxo-

ferryl species would be expected in the presence of 4-BP based upon this mechanism.

However, the data presented in this study reveal that not only is compound ES formed in the presence of 4-BP in the low concentration regime (<10 equivalents) in the 4-BP unbound form

DHP. In fact, a new species Compound iOX with the rate constant ki1 that is about 2-fold of k1 is formed when 4-BP binds in the distal pocket. Despite the fact that the Compound ES formation rate constant, k1, is not perturbed, the amount of compound ES decreases when inhibitor 4-BP is present, because of the formation of Compound iOX in an alternative pathway. Subsequently, Compound iOX is converted to the second new species Compound iRH with the rate constant, kiRH, that is 10-fold larger than kRH. Thus, based upon the observed transient-state kinetics, we believe that the presence 4-BP in the range from 1 - 10 equivalents relative to DHP does not impede H2O2 entering the distal pocket. Rather the effect of 4-BP appears to be a reduction of rate of subsequent reactions, which may be binding of H2O2 to the heme Fe or the activation of bound H2O2. Thus, the inhibition of the catalytic turnover is due to the decreasing yield of Compound ES and the existence of an alternative pathway in the 4-

BP bound form.

HO Br HO O Kia O IV FeIV Fe 4-BP Compound I Compound I  4-BP

H O Internal Electron 2 Transfer ket2 H O 2 2 ket1 k1 (ki1)

O Br O O ? ket3 FeIII FeIV FeIV k Ferric Compound ES et-3 Compound iOX

kRH

TCP ? kiRH k2 III k3 Fe TCPR TCPR Compound RH

TCP HO Br HO Kib ? O O kiRH FeIV IV FeIII 4-BP Fe Compound II Compound II 4-BP Compound iRH

Unlike Compound ES, Compound iOX does not clearly involve a ferryl species.

Irrespective of the mechanism for activation of H2O2, there is clearly a change in heme structure following addition of H2O2, which suggests an alternative mechanism for oxidation in the 4-BP bound form. The iOX spectrum shown in Figure 4.2-b2 appears to have a small

ferryl component, which means that a role for the heme Fe cannot be excluded. A second difference between ES and iOX is that 4-BP itself may participate in electron transfer reactions via a radical mechanism. While compound ES has been shown to be consistent of a ferryl heme with a tyrosine radical, we propose that compound iOX may involve a transient radical located on 4-BP (Figure 4.9). This hypothesis is plausible because the distance from the phenolic oxygen of 4-BP to heme iron is only 6.48 Å (Figure 4.1), which is less than the distance to either Tyr34 (11.36 Å, Tyr phenolic oxygen) or Tyr38 (9.48 Å, phenolic oxygen). Thus, if heme

Fe is involved in any way in this process, which we deem likely, 4-BP should be more readily oxidized than any of tyrosines due to its proximity. In terms of the driving force - the free energy, the bond dissociation free energy (BDFE) of phenolic hydroxyl group of 4-BP is 6.7 kJ/mol higher than that of the tyrosine.25 In other words, in terms of free energy, 4-BP is more stable than tyrosine. Thus, these two factors may counteract with each other that give arise similar rate constants for formation of Compound ES and Compound iOX.

A new species with a Soret band at 408 nm, Q band maximum at 498 nm and a charge transfer band at 622 nm was formed in presence of 4-BP. We assign this to the five-coordinated

Compound iRH. Compound iRH is a unique species that is not equilibrated with Compound

RH in the presence of 4-BP. Adding 4-BP to Compound RH does not give rise to Compound iRH (Figure C7). It is known that for ferric DHP, internal binding of 4-BP lowers the population of the six-coordinated high spin heme and increases the population of five- coordinated high spin heme16. Moreover, the charge-transfer band at 622 nm also indicates that

4-BP plays a role in forming a five-coordinate Compound iRH. In analogous fashion, it has been shown that the competitive inhibitor of horseradish peroxidase (HRP), benzohydroxamic

acid (BHA) also affects the charge transfer band of the ferric HRP.26 In the presence of BHA, the charge transfer band becomes stronger. In a manner analogous to BHA binding to HRP, the charge-transfer band at 622 nm becomes sharper and stronger as the concentration of 4-BP increases. It should be noted that, as an aromatic ligand, BHA binds at β edge of the distal pocket above the heme.27 However, the binding location of BHA in HRP is not nearly as deeply buried in the protein as 4-BP in DHP. The spectral differences beween compound iRH and RH suggest that possibilities of crosslinked heme15 and six-coordinate bis-histidine hemichrome still exist, but the Soret band is shifted due to the presence of an internally bound inhibitor.

Bench-top mixing kinetics have shown that the inhibition constant Ki has a strong

o temperature dependence. The value of Ki decreases by a factor of 3 for each increase in 5 C.

The large positive enthalpy and entropy calculated from the van’t Hoff equation suggests that the binding of 4-BP is enthalpically unfavored and entropy-driven. The crystal structure of

DHP with binding of inhibitor 4-XP (X = F, Cl, Br, I) shows that 4-XP is surrounded by the

9 hydrophobic amino acid residues V59, L100, F21, F24 and F35. Since 4-BP has pKa = 9.29, it remains protonated at pH 7.0. Thus the hydrophobic effect will give rise to a favorable entropy of binding in the distal pocket. In support of the dominance of the hydrophobic effect in the entropy, we note that 4-BP is not a highly flexible molecule. Thus, there is little conformational entropy change upon binding. Although the hydrogen bond interaction of 4-

BP with Y38 and heme propionate D and π-stacking interaction between aromatic ring of 4-

BP and F21 may contribute to the enthalpy, the binding also involves the displacement of a water molecule bound to the heme Fe atom and displacement of His55 into solvent water, which are both enthalpically disfavored, but entropically favored.28 The desolvation of 4-BP as it

binds in an internal hydrophobic pocket is most likely the major contribution to the positive entropy change of binding.

The inhibition constant Ki, is a dissociation constant of the enzyme-inhibitor complex, thus we compare the Kd value with previously reported dissociation constant for 4-halophenol to

DHP which was measured by Raman spectroscopy in the absence of H2O2 at pH 6.0. The dissociation constant follows the trend of size of the para halogen in 4-XP where X = I > Br >

Cl > F > H, with the values of 0.536 mM, 1.15 mM, 1.78 mM, 3.72 mM and 10 mM

7 respectively. The Kd of 4-BP measured using the measurement by resonance Raman spectroscopy is 1.15 mM, which is within the range we measured from 283K to 298K.

However, it also must be understood that the Kd value measured by Raman reflects the interplay between 6cHS and 5cHS by disturbing the water ligation, and thus is only indirectly related to the binding of 4-BP.

The biological role of inhibition may be explained by consideration of the relative concentrations of 4-BP, 2,4-dibromophenol and 2,4,6-TBP, which are typical secondary metabolites secreted by marine polychaetes. For example, Notomastus lobatus, excretes 4-BP,

2,4-dibromophenol, and 2,4,6-TBP to the surrounding environment with a stoichiometric ratio of 1.8:0.9:1.0.8 The reason why DHP selectively oxidizes 2,4,6-TBP, but not 4-BP, may be partially explained by the relative toxicity of these two compounds and their oxidized intermediates. 2,4,6-TBP has been shown to disrupt the function of cellular Ca2+ ion channel in neuroendocrine cell and may potentially disturb the endocrine system in invertebrates.29 It also has a negative impact on the ability of polychaetes to burrow and feed.30 On the other hand, there is no known inhibitor effect of 4-BP on respiration and the assimilation of acetate

or glucose by sediment bacteria, indicating that it has significantly lower toxicity relative 2,4,6-

TBP.31 Thus, we assume that the high priority for DHP to degrade 2,4,6-TBP is due to its high toxicity. Not only is 4-BP less toxic than 2,4,6-TBP, but oxidation of 4-BP can lead to polymerization of the radical intermediate, which may increase the toxicity relative to the starting phenol. This is distinct from the fate of 2,4,6-TXP radicals, which form quinones by disproportionation32 and hydroxyquinones by a further radical pathway.33 Thus, the differences in binding site and reactivity of 4-BP and 2,4,6-TBP may be an evolutionary consequence of the regulation of catalytic reactivity in vivo.

4.6 References

(1) Chen, Y. P.; Woodin, S. A.; Lincoln, D. E.; Lovell, C. R. J Biol Chem 1996, 271,

4609.

(2) D'Antonio, J.; D'Antonio, E. L.; Thompson, M. K.; Bowden, E. F.; Franzen, S.;

Smirnova, T.; Ghiladi, R. A. Biochemistry 2010, 49, 6600.

(3) Zamocky, M.; Gasselhuber, B.; Furtmuller, P. G.; Obinger, C. Arch. Biochem.

Biophys.

2012, 525, 131.

(4) Weber, R. E.; Mangum, C.; Steinman, H.; Bonaventura, C.; Sullivan, B.;

Bonaventura, J. Comp Biochem Physiol A Comp Physiol 1977, 56, 179.

(5) Menale, C.; Nicolucci, C.; Catapane, M.; Rossi, S.; Bencivenga, U.; Mita, D. G.;

Diano, N. J. Mol. Catal. B-Enzym. 2012, 78, 38.

(6) Osborne, R. L.; Coggins, M. K.; Raner, G. M.; Walla, M.; Dawson, J. H.

Biochemistry 2009, 48, 4231.

(7) Thompson, M. K.; Davis, M. F.; de Serrano, V.; Nicoletti, F. P.; Howes, B. D.;

Smulevich, G.; Franzen, S. Biophys J 2010, 99, 1586.

(8) Lincoln, D. E.; Fielman, K. T.; Marinelli, R. L.; Woodin, S. A. Biochem Syst Ecol

2005, 33, 559.

(9) de Serrano, V.; Franzen, S. Peptide Science 2012, 98, 27.

(10) D'Antonio, J.; Ghiladi, R. A. Biochemistry 2011, 50, 5999.

(11) Denisov, I. G.; Makris, T. M.; Sligar, S. G. J. BIol. Chem. 2002, 277, 42706.

(12) Miller, V. P.; Goodin, D. B.; Friedman, A. E.; Hartmann, C.; Demontellano, P. R. O.

J. BIol. Chem. 1995, 270, 18413.

(13) Aisen, P. Chemtracts: Biochem. Mol. Biol. 1990, 1, 441.

(14) Goodin, D. B.; Mauk, A. G.; Smith, M. Proc. Natl. Acad. Sci. U. S. A. 1986, 83,

1295.

(15) Feducia, J.; Dumarieh, R.; Gilvey, L. B. G.; Smirnova, T.; Franzen, S.; Ghiladi, R. A.

Biochemistry-Us 2009, 48, 995.

(16) Ma, H.; Thompson, M. K.; Gaff, J.; Franzen, S. J Phys Chem B 2010, 114, 13823.

(17) Franzen, S.; Thompson, M. K.; Ghiladi, R. A. Biochim. Biophys. Acta 2012, 1824,

578.

(18) Georgiadis, K. E.; Jhon, N.-I.; Einarsdottir, O. Biochemistry 1994, 33, 9245.

(19) Rittle, J.; Younker, J. M.; Green, M. T. Inorganic Chemistry 2010, 49, 3610.

(20) Belyea, J.; Gilvey, L. B.; Davis, M. F.; Godek, M.; Sit, T. L.; Lommel, S. A.;

Franzen, S. Biochemistry 2005, 44, 15637.

(21) Franzen, S.; Belyea, J.; Gilvey, L. B.; Davis, M. F.; Chaudhary, C. E.; Sit, T. L.;

Lommel, S. A. Biochemistry 2006, 45, 9085.

(22) Zhao, J. J.; de Serrano, V.; Dumarieh, R.; Thompson, M.; Ghiladi, R. A.; Franzen, S.

J. Phys. Chem. B 2012, 116, 12065.

(23) Chen, Z.; de Serrano, V.; Betts, L.; Franzen, S. Acta Crystallogr., Sect. D: Biol.

Crystallogr. 2009, D65, 34.

(24) Yang, F.; Phillips, G. N., Jr. Journal of Molecular Biology 1996, 256, 762.

(25) Warren, J. J.; Tronic, T. A.; Mayer, J. M. Chem Rev 2010, 110, 6961.

(26) Zelent, B.; Kaposi, A. D.; Nucci, N. V.; Sharp, K. A.; Dalosto, S. D.; Wright, W. W.;

Vanderkooi, J. M. J Phys Chem B 2004, 108, 10317.

(27) Gumiero, A.; Murphy, E. J.; Metcalfe, C. L.; Moody, P. C. E.; Raven, E. L. Arch.

Biochem. Biophys. 2010, 500, 13.

(28) de Serrano, V.; Chen, Z. X.; Davis, M. F.; Franzen, S. Acta Crystallogr D 2007, 63,

1094.

(29) Hassenklover, T.; Predehl, S.; Pilli, J.; Ledwolorz, J.; Assmann, M.; Bickmeyer, U.

Aquat Toxicol 2006, 76, 37.

(30) Fielman, K. T.; Woodin, S. A.; Walla, M. D.; Lincoln, D. E. Marine Ecology:

Progress Series 1999, 181, 1.

(31) Lovell, C. R.; Steward, C. C.; Phillips, T. Mar Ecol-Prog Ser 1999, 179, 241.

(32) Sturgeon, B. E.; Battenburg, B. J.; Lyon, B. J.; Franzen, S. Chem. Res. Tox. 2011, 24,

1862.

(33) Franzen, S.; Sasan, K.; Sturgeon, B. E.; Lyon, B. J.; Battenburg, B. J.; Gracz, H.;

Dumariah, R.; Ghiladi, R. The Journal of Physical Chemistry B 2011, 116, 1666.

CHAPTER 5

The Regulatory Implications of Hydroquinone for the Multifunctional Enzyme Dehaloperoxidase-Hemoglobin from Amphitrite ornata

5.1 Abstract

Hydroquinone (H2Q) has been observed to compete with the oxidation of substrates 2,4,6- tribromophenol (2,4,6-TBP) and 2,4,6-trichlorophenol (2,4,6-TCP) catalyzed by the dehaloperoxidase-hemoglobin (DHP) from Amphitrite ornata in the presence of H2O2. This competition is observed as a lag phase during which H2Q is preferentially oxidized to 1,4- benzoquinone (1,4-BQ) while totally inhibiting either 2,4,6-TBP or 2,4,6-TCP oxidation. The inhibition by H2Q is distinct from that of the native competitive inhibitor 4-bromophenol (4-

BP) since H2Q is itself oxidized and its product 1,4-BQ is not an inhibitor. Thus, once H2Q is completely consumed the inhibition is removed, and normal substrate turnover is initiated, which explains the lag phase. To probe the mechanism of lag phase, the reactions between H2Q and DHP were both studied both in the presence and in the absence of H2O2. The reversible reactions between ferric/oxyferrous DHP A and H2Q/1,4-BQ are shown to involve a proton coupled electron transfer (PCET) mechanism, where the distal histidine His55 serves as the

55 proton acceptor. The pKa of the distal histidine His has been determined by resonance Raman spectroscopy in order to corroborate its involvement in this mechanism. Consistent with the proposed mechanism, kinetic assays have shown that H2Q serves as a substrate for DHP that follows the Michaelis-Menten kinetics. Unlike H2Q, the product 1,4-BQ has a relatively large

Ki value and therefore has negligible inhibition. This study sheds light on understanding the difference between substrate and inhibitor binding sites and regulatory implication for the peroxidase and oxygen-transporter functions in DHP. It also provides information on PCET in

DHP, which is like important for resolving the switching between the ferric peroxidase catalytic function and the ferrous oxygen transport function.

5.2 Introduction

Dehaloperoxidase-hemoglobin A (DHP A) has been shown to possess distinct substrate and inhibitor binding sites in the distal pocket. 4-Br binds in the distal pocket perpendicular to the heme and serves as a non-classical competitive inhibitor to DHP A.1,2 The substrate 2,4,6-TBP binding site has been recently shown have a binding site, which is buried even more deeply in the globin.3 An additional two 2,4,6-TCP binding sites in the distal pocket have recently been discovered 4, which causes us to pose the question: how many substrates can bind and how many modes of binding are there in this relatively small protein? The internal 2,4,6-TBP substrate binding site is located above the α- edge of the heme in the crystal structure, suggesting the functional possibility of internal oxidation by a sequential two-electron mechanism.3 Consideration of the mechanism consistent with substrate binding must also include the role played by the distal histidine, which is His55 in DHP. The conserved distal histidine His55 of DHP, similar to that of myoglobin5 and horseradish peroxidase (HRP),6 has shown pH-dependent allosteric behavior which is believed to be both important for regulatory control and catalytic activation.1,7-13X-ray crystal structures have shown that the distal histidine

His55 of DHP has much greater conformational flexibility than the distal histidine His64 of

SWMb at pH 6.0.7,9 The open conformation of His64 of SWMb is only observed at pH 4.0 while it is observed at pH 6.0 in DHP.14 The distal histidine His55 has shown to be essential for

9 functioning as an acid-base catalyst in facilitating the O-O heterolysis of H2O2, resulting in formation of active oxidative species Compound ES, a ferryl species with an amino acid radical locates on the tyrosine Tyr38 or Tyr34.13,15The compound ES label used in DHP A and B is

attributed to an analogous compound ES intermediate formed in cytochrome c peroxidase

(CcP), noting that the radical in CcP is located on tryptophan instead of on tyrosine.16

Although extensive studies have focused on the mechanism whereby DHP carries out peroxidase reactions and identification of the substrate binding sites, only recently has research begun to focus on understanding the mechanism of the interplay between ferric and oxyferrous state of DHP with respect to peroxidase function.17,18 This issue is central to DHP function since DHP mostly exists in the oxyferrous form in vivo. It is not known how DHP switches to a peroxidase function from a typical hemoglobin oxygen transport function. Previous studies have shown that DHP can initiate the peroxidase function from the oxyferrous state by a direct formation of the active ferryl species, Compound II, in the presence of substrate 2,4,6-TCP

19-21 and cosubstrate H2O2, providing an exceptional example that challenges the conventional peroxidase paradigm. The normal peroxidase reaction cycle starts and ends with the ferric heme Fe. For example, for HRP the cycle consists of Ferric HRP → Compound I →

Compound II → Ferric HRP. The oxyferrous state is never produced in that cycle and it is considered an inactive dead end in most peroxidases. The alternative cycle in DHP starts with oxyferrous DHP (DHP-O2), DHP-O2 → Compound II → Ferric DHP → DHP-O2.

Surprisingly, ferric DHP can be reduced back to the oxyferrous state (Ferric DHP → DHP-O2) when exposed to the oxidation product 2,6-dichloroquinone (2,6-DCQ).17 One can surmise that the role played by 2,6-DCQ would most likely be that of a redox mediator. Quinones are easily reduced to the hydroquinone, which is in turn an excellent reducing agent. This unorthodox role for the product, 2,6-DCQ may become clearer once we understand electron transfer in

DHP. However, the product 2,6-DCQ spontaneously reacts to form 3-hydroxy-2,6-

100

dichloroquinone (3-OH-2,6-DCQ), which complicates the use of this reagent for studies of the electron transfer pathways in DHP.22 The 2,6-dibromohydroquinone is also unstable and cannot be purchased. The benzoquinone / hydroquinone system is a more stable molecule for study of the role played by hydroquinones in the DHP mechanism. During the course of such studies an unusual kinetic role for benzohydroquinone was discovered.

Herein, we report the unique kinetic behavior of benzohydroquinone (H2Q) on the DHP reaction cycle. Indeed, H2Q can reduce DHP from the ferric to the ferrous form. However, studies of enzymatic activity using 2,4,6-TCP as the substrate led to an usual observation of a

“lag phase” for the catalytic oxidation of substrate by DHP A. The lag phase differs from an inhibitory process in that there is no measurable product (2,6-DCQ) formed until H2Q is completely consumed. At the end of the lag phase turnover begins and the 2,6-DCQ product forms at a normal rate, i.e. similar to the rate observed in the presence of H2O2 but without any H2Q. In order to elucidate the mechanism of lag phase, both the interactions of H2Q with

DHP A in the absence and in the presence of H2O2 must be considered. It is highly relevant for the previous suggestion of a role for 2,6-DBQ that H2Q reversibly reduces ferric DHP A to the oxyferrous state by a proton coupled electron transfer (PCET) mechanism facilitated by the

55 55 distal histidine His , where the distal histidine His not only stabilizes the bound O2 ligand by hydrogen bonding but also serves as a proton acceptor during reduction of the heme by

H2Q. A pH-dependent distal histidine conformational behavior has been measured in both ferric state and oxyferrous states. A complete thermodynamic scheme of proton and electron transfer has been established by applying resonance Raman spectroscopy combined with previously measured electrochemistry data. The binding site of H2Q was probed by an

101

inhibition assay with the standard inhibitor 4-bromophenol (4-BP). Given the unique functional range of DHP, this study provides further insights into the difference between substrate and inhibitor binding sites, as well as the interplay between initiation of peroxidase function from the ferric state or the oxyferrous state. Such functional considerations are at the heart of any study of multifunctional proteins since they show how different functions are regulated and how they can be mutually tolerated in a single protein.

5.3 Material and Methods

Materials

All reagents were purchased from Aldrich and ACROS and used without further purification. 2,4,6-trichloropehnol (2,4,6-TCP), 2,4,6-tribromophenol (2,4,6-TBP) and 4- bromophenol (4-BP) were each dissolved in 100 mM, pH = 7.0 potassium phosphate (KPi) buffer to prepare the stock solution. H2Q and 1,4-BQ were prepared by direct weighing method

o in the 100 mM KPi buffer at desired pH. Prepared solutions were stored at 4 C and protected against light. Other concentrations were measured by monitoring their absorbance: 1,4-BQ,

-1 -1 -1 -1 -1 -1 ε246nm = 20,600 M cm ; TCP, ε312nm = 3,752 M cm ; TBP, ε316nm = 4,640 M cm ; 4-BP,

-1 -1 ε280nm = 1,370 M cm . Spectra were obtained using an Agilent 8453 diode array UV-visible spectrophotometer with a Peltier-cooled sample cell at 25 ℃. Hydrogen peroxide solution was freshly made before each kinetic experiment and is kept on ice and protected against light during the experiment. Wild-type His6X (histidine-tagged) DHP A and H55D mutant were expressed in E.coli and purified as previously described.9,23 The concentration of both ferric and oxyferrous DHP A was determined by using the molar absorption coefficient, ε = 116,400

102

M-1 cm-1.24 Oxyferrous DHP was prepared by adding excess amount of sodium dithionite

(Na2S2O4) to the purified ferric DHP. Then the solution was filtered through the PD-10 column to get rid of the remained Na2S2O4 and bubbled with oxygen for 5 min.

Bench-top Mixing Kinetic Assays

The kinetic assays were conducted using an Agilent 8453 UV-visible spectrophotometer operating in kinetic mode with a 1 second time resolution. The catalytic reactions were carried out in a 0.4 cm path length cuvette obtained from Starna Cells, Inc with a total volume of 1200

µL. The ferric DHP A concentration [E]0 in each sample was 2.4µM. For the reaction in the absence of H2O2 DHP A and KPi buffer were first mixed and allowed to incubate for 3 min in a cuvette placed in the thermal cell to reach thermal equilibrium. Subsequently, 200 µL of substrate solution was added into the cuvette to initiate the reaction. For the reaction in the presence of H2O2, the substrate H2Q were first mixed with ferric DHP A and KPi buffer and then allowed to incubate for 3 min in the cuvette placed in the thermal cell to let DHP A fully reduced to oxyferrous and to reach thermal equilibrium at the same time. Subsequently, 200

µL of 7.2 mM H2O2 solution was added into the cuvette to initiate the reaction. The kinetic data were measured by monitoring the absorbance at wavelength 246 nm, which corresponds

-1 to the absorbance peak of the 1,4-BQ, with a molar absorption coefficient ε246nm = 20,600 M cm-1.

Stopped-flow UV-visible Kinetic Assays

Experiments were performed on a Bio-Logic SFM-400 triple-mixing stopped-flow instrument equipped with a diode array UV-visible spectrophotometer and were carried out at

23 ℃ in 100 mM KPi buffer, pH 7.0. Data were collected over three time-domain regimes (2.5,

103

25, and 250 ms; 300 scans each) using the Bio Kinet32 software package (Bio-Logic). Data were collected (900 scans total) over a three-time domain regime (2.5, 25, and 250 ms; 300 scans each, 83.25 s total). Single–mixing experiments were performed, in which ferric DHP A were preincubated with H2Q prior to the mixing with H2O2. The final concentration after mixing were [DHP A] = 5 µM , [H2Q] = 55 µM , [H2O2] = 500 µM .

Resonance Raman Spectroscopy

DHP A samples at a final protein concentration of 100 µM were prepared in 100 mM

KPi buffer, pH 7. 0. Samples were placed in 5 mm NMR tubes and spun with an air piston spinning sample holder (Princeton Photonics, model Raman 101). Resonance Raman spectra were obtained by excitation at the edge of the Soret band at 410 nm using Coherent Mira 900 tunable titanium sapphire laser generating 700 ~ 1000 nm light. The Ti:sapphire laser was pumped by a Coherent Verdi 10 frequency-doubled diode-pumped Nd: vanadate laser that generating 10 W of 532 nm light. The near-IR output from the Ti:sapphire laser was sent through a Coherent 5-050 frequency doubler to generate the working range of 400 ~ 430 nm light for Soret band excitation. The frequency doubled beam was collimated and cylindrically focused to a vertical line of ~5 mm and typically 45~60 mW at the sample. Raman scattered light was collected by the Spex 1877 Triplemate monochromator (2400 grooves/mm final stage grating) and was detected by a liquid N2-cooled CCD camera (ISA Spex, model CCD-3000).

Spectra were measured at room temperature for 40 acquisitions with total exposure time of

1200 seconds. The spectra were calibrated using standard spectra of toluene and carbon tetrachloride.

104

Data Analysis

For reaction between ferric DHP A and H2Q and the reverse reaction between oxyferrous

DHP A and 1,4-BQ, the time resolved spectra measured in the bench-top kinetic assay were analyzed using the singular value decomposition (SVD) method. SVD provides a decomposition of the original absorption data matrix A(,t) in terms of basis spectra as the product of three matrices USVT. The VT matrix (Figure S2, S5), corresponding to the time- course evolution, was evaluated as a one step irreversible first-order reaction and globally fitted to a single exponential function A = c0 + c1 exp(-kobst) , from which the apparent rate constant kobs and C matrix was determined. The spectra corresponding to each reaction species were calculated based on the analytical solution of the one step irreversible first-order reaction model (Appendix D). The SVD and global fitting analysis were performed using Igor Pro 6.0.

The kinetic data for the DHP A catalyzed oxidation of H2Q in the presence of H2O2, were fitted using the short time approximation. The slope of experimental progress curve was determined by linear fit of the first ten time points to provide the initial rates V0. A series of initial rates V0 obtained as a function of the substrate concentration were then fitted to the

Michaelis-Menten equation to obtain the parameters Vmax and Km (Table 5.1).

Table 5.4 Michaelis-Menten Kinetic Parameters for Oxidation of H2Q by DHP A

-1 -3 -1 -1 T/K kcat (s ) Km (µM) kcat/Km (10 ∙s M ) 303 0.450 ± 0.008 59.7 ± 5.8 7.53 ± 0.74 298 0.280 ± 0.007 65.3 ± 8.0 4.29 ± 0.53 293 0.216 ± 0.003 87.5 ± 5.1 2.47 ± 0.15 288 0.138 ± 0.003 75.1 ± 7.8 1.84 ± 0.19

The pH-dependent Raman spectra were first baseline subtracted by using a 4 point 4 polynomial extrapolation and normalized according to the intensity of the ν4 band. Then the

105

data matrix A(휈̃, pH) was decomposed into three matrices USVT using SVD method. The second eigenvectors of VT matrix were fitted to the proposed model (Scheme S1, S2) in order to determine the corresponding pKa value.

5.4 Results

The lag phase observed in the catalytic oxidation of 2,4,6-TCP

Figure 5.1 shows the oxidation kinetics of 2,4,6-TCP in the presence of H2Q. Observation at 273 nm monitors the formation of the product 2,6-DCQ, while 246 nm monitors the oxidation of H2Q. Figure 5.1 shows that no 2,6-DCQ product is formed during the first 25 seconds of the experiment. Instead, H2Q is oxidized for a period of 25 seconds under the assay conditions of Figure 5.1. We call this delay in the formation of product, the lag phase. A lag phase is observed for both 2,4,6-TCP and the native substrate of DHP A 2,4,6-TBP. A key observation is that the duration of the lag phase depends linearly on the concentration of H2Q added to the assay mixture (Figure D1). Since H2Q reacts to form 1,4-BQ in the presence of

106

DHP A (Scheme 5.1), one can surmise that the lag phase ends when H2Q is consumed. One possible explanation for the lag phase is a reaction between H2Q and ferric DHP A that is rapid compared to turnover of either 2,4,6-TCP or 2,4,6-TBP. In this case, H2Q could completely block enzymatic turnover if it is bound inside the protein in the distal pocket in the inhibitor binding site.1 However, this inhibition mechanism is satisfactory only if the binding of 1,4-

BQ is weak compared to the binding of H2Q. To test this two-part hypothesis we have studied the reduction of ferric DHP A by H2Q and the competition between H2Q and the native inhibitor 4-BP for the inhibitor binding site in the distal pocket. We also studied the binding constant for the product of H2Q oxidation, 1,4-BQ.

Scheme 5.1. The reversible reaction between ferric DHP A with H2Q and oxyferrous DHP A with 1,4-BQ

The reversible reaction between reduction of ferric DHP A by H2Q and oxidation of oxyferrous

DHP A by 1,4-BQ

0.25 DHP A Ferric 407 nm DHP A Oxy-ferrous 417 nm 542 nm579 nm

506 nm

0.20 -3 30

x10 25 0.15 20 15 0.10 10

5 Absorbance 0.05 520 560 600

0.00 300 400 500 600 Wavelength(nm)

Figure 5.2 Reconstructed spectra of the reaction between ferric DHP A (red) and H2Q that forms the oxyferrous DHP A (blue). The inset shows an expansion of the Q-band region.

107

The time-resolved spectra (Figure D2) show that ferric DHP A has been reduced to the oxyferrous form while H2Q has been oxidized to 1,4-BQ, characterized by the increase in absorbance at 246 nm. A well-defined isosbestic point can be observed throughout the reaction time course, indicating only one product has formed in this reaction. As we can see from the

SVD analysis of the spectra (Figure 5.2), the Soret band at 407 nm and Q band at 506 nm both belong to the metaquo high spin ferric form of DHP A. A sharper red shifted Soret band at

416 nm and the α,β braches of Q band at 542 nm and 579 nm indicate the formation of 6 coordinated low-spin oxyferrous DHP A.

0.12

0.10

) -1

0.08

(s obs

k 0.06 0.04 0.02 100 200 300 400 500 [H Q](M) 2

Figure 5.3 Plot of the kobs vs [H2Q] for the reduction of ferric DHP A(5 µM) by H2Q in 100 mM KPi buffer, pH 7.0, 298 K.

The reduction of ferric DHP A by H2Q exhibits pseudo-first order kinetics in the presence of at least 10-fold excess of H2Q. The pseudo-first order rate constant kobs shows a linear dependence on [H2Q] (Figure 5.3). Thus, the second order rate constant k1 = (2.21 ± 0.04) ×

2 -1 -1 10 M s can be obtained according to the expression kobs= k1[H2Q]. Additionally, for the reverse reaction, the oxidation of oxyferrous DHP A also follows pseudo-first order kinetics in the presence of at least 10-fold excess of 1,4-BQ. The second order rate constant k-1 = (7.35

108

2 -1 -1 ± 0.25) × 10 M s (Figure D8). Thus, the equilibrium constant K1 for this reversible reaction is K1 = k1/k-1= 0.301 ± 0.012.

The pH dependence of the second-order rate constant k1 has been studied in the pH range from 5.0 to 8.0 (Figure D4). k1 is stable below pH 6.5 and rapidly increases above pH 6.5.

However, the oxidation at pH = 8.0 no longer follows first order reaction kinetics. Although the H2Q oxidation rate increases with pH, the effect is limited by the formation of hydroxy ferric DHP, which has a pKa = 8.1. Thus at pH = 8.0, approximately 50% of the ferric DHP has been ligated with hydroxide.25 The axial hydroxide ligand prevents the outer-sphere electron transfer from H2Q to the heme iron, thus inhibits the reduction of the protein.

1.0

0.8

obs

k /

obs 0.6 k'

0.4

0 100 200 300 400 500 [4-BP] (

Scheme 5.2. Inhibition model for the reduction of ferric DHP A by H2Q

109

푘′ 퐾 표푏푠 = 𝑖 (1) 푘표푏푠 퐾𝑖 + [퐼]

It is not straightforward to directly measure a putative binding of H2Q in the distal pocket since it reacts with ferric DHP A. However, we can probe whether H2Q binds in the inhibitor binding site using a competitive binding assay with the inhibitor 4-BP, which has been

1 established in inhibition assays of 2,4,6-TCP. Consistent with the hypothesis that H2Q binds in the distal pocket, the inhibitor 4-BP also acts to inhibit the reduction of the heme iron in the presence of H2Q. It is known that inhibitor 4-BP binds in the distal pocket of DHP A with binding affinity Kd =1.15 mM in the metaquo resting state measured by resonance Raman

1 spectra at room temperature. The value for under conditions of turnover is Ki = 0.155 mM as measured by an inhibition kinetic assay at 298K.26 The kinetic assays show that the rate of heme reduction by H2Q decreases in the presence of 4-BP (Figure 4). A plausible hypothesis for this inhibition behavior is that both H2Q and 4-BP compete for the same internal binding site. Since 4-BP is not reactive in the distal pocket it inhibits the oxidation of H2Q, The inhibition data have been fitted to the model illustrated below, resulting in Ki = 0.283 mM at

298K. The discrepancy between two Ki values measured in two sets of kinetic assays is due to the different kinetic assay conditions. Previously, Ki was measured in the presence of cosubstrate H2O2 under turnover conditions, whereas Ki is measured in the absence of H2O2 in this experiment.

The temperature dependence of second-order constant k1 for the oxidation of H2Q by ferric DHP A has been studied at 283 K, 288 K, 293 K and 298 K. ln (k1) vs 1/T has been plotted according to the Eyring equation (Eqn. 2) to obtain the thermodynamic parameters ΔH‡

110

= 20.0 ± 1.4 kJ/mol , ΔS‡ = -132 ± 5 J/(mol·K), which gives the Gibbs free energy for the transition state ΔG‡ = 59.4 ± 2.9 kJ/mol.

푘 푘 Δ푆‡ Δ퐻‡ 푙푛 = 푙푛 퐵 + − (2) 푇 ℎ 푅 푅푇

-0.2 -0.3

-0.4

/ ) T 1

k -0.5

ln ( -0.6 -0.7

3.40 3.45 3.50 -3 -1 1/T (10 K )

Figure 5.5 Eyring plot for the oxidation of H2Q by ferric DHP A (5 µM) in 100 mM KPi buffer, pH 7.0.

55 Distal Histidine His acts as a proton acceptor facilitating the oxidation of H2Q by a Proton

Coupled Electron Transfer (PCET) mechanism

55 The distal histidine His has been shown to be essential for the oxidation of H2Q by

DHP A. For example, the ferric DHP A mutant H55D has no detectable activity towards H2Q

55 oxidation. The pKa values for the distal histidine His have been measured in both ferric and oxyferrous forms by resonance Raman spectroscopy. The allosteric behavior of His55 is crucial in controlling the coordination state of the heme iron in a pH-dependent manner. In the ferric metaquo state, the distal histidine favors the closed conformation that stabilizes the distal water ligand at high pH, whereas it swings out of the distal pocket into the solvent and favors the open conformation at low pH. As a consequence, the coordination and spin state of heme iron switch from 6-coordinate high spin (6cHS) to 5-coordinate high spin (5cHS). This trend can

111

be clearly seen in the resonance Raman spectra as function of pH shown in Figure 5.6. The population of 5cHS gradually shifts to 6cHS as the pH rises. However, when pH is above 7.5, the rapid increase of the population of 6cHS is due to the formation of hydroxide ligated form.

25 The acid-alkaline transition in DHP A has been previously measured with a pKa = 8.1. Thus, a dual equilibrium model with different pKa values (pKa2 = 8.1) is used to fit the population curve obtained from SVD analysis. The analysis gives pKa = 5.8 ± 0.3 for the distal histidine

His55 in ferric DHP (Figure 5.6).

Figure 5.6 Resonance Raman spectra of metaquo DHP A as a function of pH in the high frequency region. A scheme describing the relevant equilibra for ferric DHP A is shown in the figure.

112

Figure 5.7 Resonance Raman spectra of oxyferrous DHP A as a function of pH in the high frequency region. A photoexcitation reaction scheme for deoxyferrous DHP A is shown in Figure.

In the oxyferrous form, the distal histidine His55 forms a hydrogen bond to the dioxygen in the closed conformation. This hydrogen bond significantly increases the oxygen binding affinity of DHP in a way that is similar to other hemoglobins.27 At low pH, this hydrogen bond is disrupted due to the open conformation of the histidine.8 Although the conformation of the distal histidine is changed upon a pH shift, this does not result in deligation of O2. Thus, the measurement of the conformation change of His55 using resonance Raman spectroscopy would have been challenging or impossible if were not for a light-induced oxidation of DHP A. It happens coincidentally that the autoxidation rate of oxyferrous DHP A is dramatically

113

accelerated by laser photoexcitation, which provides a method for estimating the population ratio between closed and open of conformations of oxyferrous DHP A. It is hypothesized that only the oxyferrous form of DHP A with distal histidine His55 in the open conformation will be excited due to the lack of hydrogen bond stabilization to the dioxygen ligand. Upon photoexcitation, oxyferrous DHP without hydrogen bond is oxidized into the ferric state. The photoexcited autoxidation has also shown a pH-dependent manner, in which the autoxidation rate is much faster at low pH, probably attributed to the formation of a protonated superoxide species (Figure 5.7). Thus, an equilibrium model combined with photoexcited autoxidation conversion was proposed and fitted to the population curve obtained from SVD analysis. Using

55 this method, the pKa of the distal histidine His in oxyferrous DHP A was determined to be

6.9 ± 0.1 (see Apendix D).

II 55 Scheme 5.3.Thermodynamic scheme of proton and electron transfer of Fe -O2~His -H The redox potential between Fe(III)/Fe(II) in the anaerobic condition at pH 7.0 has been measured to be E° = 0.221 V.28 Under aerobic conditions, the redox potential between

114

(III) (II) ° Fe /Fe -O2 is about 0.08 V higher than the value measured in the anaerobic condition, E =

0.305 V, due to the binding of O2 to the heme iron that stabilizes the ferrous state Fe and

28 changes the d electrons from high spin to low spin. The pKa values for the protein matrix of both ferric and oxyferrous DHP A that take up or expel a proton has been measured using the resonance Raman spectrophotometric method. A thermodynamic cycle shown in Scheme 4 can be established based on these data. The bond dissociation free energy (BDFE) of FeII-

55 O2~His -H can be calculated from both routes, either starting from proton transfer (PT),

II 55 BDFE (Fe -O2~His -H) = 5.73 pKa + 96.48 E° + CG = 303.8 ± 1.7 kJ/mol (CG is a solvent

29 II 55 constant, as for water, CG = 241.16 kJ/mol) or electron transfer (ET), BDFE (Fe -O2~His -

H) = 302.0 ± 0.6 kJ/mol. According to Hess’ law, the energy change should be independent of

II 55 the path, thus the consistency between these two BDFE (Fe -O2~His -H) supports the validity of the two pKa values measured by resonance Raman spectroscopy. Moreover, as for

° II 55 the reaction 1, ΔG = BDFE (H2Q) – BDFE (Fe -O2~His -H) = -RTlnK1. Since the average

29 II BDFE (H2Q) = 307.5 kJ/mol in aqueous solution, and K1 = 0.301 ± 0.012. Thus, BDFE (Fe -

55 II 55 O2~His -H) can be calculated to be 304.5 ± 0.1 (298K). The BDFE (Fe -O2~His -H) calculated from the reversible reaction equilibrium constant is within error with the value predicted from thermodynamic square scheme.

DHP A catalyzed oxidation of H2Q in the presence of H2O2

In order for H2Q to act catalytically in place of 2,4,6-TBP during the lag phase, it must be oxidized by H2O2. Yet, H2Q will reduce the heme and thus, any catalytic role for DHP must involve a ferrous, rather than ferric heme. The question of whether H2Q can be catalytically

115

oxidized by DHP A was addressed using similar methods employed for the substrates, 2,4,6-

TCP and 2,4,6-TBP (Table 5.2). The time resolved spectrum shows that DHP A starts in the oxyferrous state after incubating with H2Q for 3 min. Once H2O2 was added, the oxyferrous

DHP gradually turns to Compound II. The Soret band was shifted from 417 nm to 420 nm, α and β branches of the Q band decreased in intensity. The rising peak at 245 nm is attributed to the formation of product, 1,4-BQ. No ferric DHP intermediate was observed during the reaction time course in the much higher time resolution of the stopped-flow kinetic assay

(Figure D14, D15) suggesting that the activation of oxyferrous DHP undergoes a concerted two-electron oxidation yielding compound II, and then reduced by substrate hydroquinone in a direct two-electron reduction.

542 nm 579 nm 0.8 25

-3

0.6 15 x10 10 539 nm 578 nm 0.4 5

Absorbance 500 520 540 560 580 600 620 0.2 417 nm 245 nm 420 nm 0.0 300 400 500 600 Wavelength (nm)

The oxidation of H2Q catalyzed by DHP A in the presence of H2O2 follows Michaelis-

Menten kinetics. The kinetic parameters kcat and Km were obtained by fitting the curve of the initial rate to the Michaelis-Menten equation. The temperature dependence of kcat and Km were

116

obtained by conducting the kinetic measurement at four temperatures, 288 K, 293 K, 298 K and 303 K. The Arrhenius plot gives activation energy Ea = 69.3 ± 7.4 kJ/mol for kcat/ Km, and

Ea = 55.2 ± 4.3 kJ/mol for kcat.

303 K 1.0 298 K 293 K 288 K

) 0.8 -1

Ms 0.6 -6

(x10 0.4

o o V 0.2

0.0 0 200 400 600 800 1000 [H2Q] (M)

Table 5.5 Michaelis-Menten Kinetic Parameters of DHP A Substrates

-1 Substrate kcat(s ) KM (mM) Ea (kcat) (kJ/mol) Ea (kcat/KM) (kJ/mol)

H2Q 0.28 0.065 55.2 69.3 2,4,6-TBP (10%MeOH) 1.55 0.713 47.9 45.5 2,4,6-TCP 7.16 1.08 44.0 56.3

The pH dependence of initial rate for H2Q in this reaction is similar to that of 2,4,6-TCP.

At low pH, the initial rate is faster, however, DHP A undergoes a competing deactivation reaction to form compound RH. 13,15,30 Thus, the total amount of product that eventually produced is actually less than that of at pH 7.0. The pH dependence of initial rate can be fitted into the sigmoid curve with a midpoint value at pH = 6.13 ± 0.08.

117

5.5 Discussion

The lag phase caused by H2Q in the catalytic oxidation of 2,4,6-TCP by DHP A presents an intriguing question about the substrate specificity of DHP A. H2Q acts as an extraordinarily potent inhibitor of the oxidation of 2,4,6-TCP, while it is simultaneously oxidized to 1,4-BQ.

2,4,6-TCP starts to turn over only after H2Q has been completely consumed. The duration of the lag phase is proportional to the amount of the H2Q in the solution mixture. Thus, one can quantitatively substantiate that H2Q leads to the lag phase.

The oxidation of H2Q by ferric DHP A in the absence of H2O2 by a PCET mechanism

+ - The reaction between ferric DHP A and H2Q involves a net H and e transfer from the substrate H2Q to ferric DHP A. In principle, there are three pathways, in which this reaction can take place. The H+ and e- are either transferred by a concerted step, by a PCET mechanism, or they undergo separate proton transfer (PT) and electron transfer (ET) reactions. Since the

• II 55 thermodynamic diagrams of H2Q, HQ (Scheme S3) and Fe -O2~His -H (Scheme 4) have all been established, the thermodynamic analysis provides a basic argument to examine the possible reaction pathways.

For the reaction between ferric DHP A and H2Q, the activation energy between transition state and ground state obtained by Eyring plot gives ΔG‡ = 59.4 ± 2.9 kJ/mol.

Therefore,

a. For the initial ET and subsequent PT process:

° −1 −1 −1 ∆퐺퐸푇 = −(96.48푘퐽 ∙ 푚표푙 푉 ) ∙ (0.221 − 1.10) = 84.80 푘퐽 ∙ 푚표푙

118

The free energy is much higher than that of activation energy, thus this pathway is not

possible.

b. As for the initial PT and subsequent PT process:

° −1 −1 ∆퐺푃푇 = −(5.73푘퐽 ∙ 푚표푙 ) ∙ (5.8 − 9.85) = 23.21 푘퐽 ∙ 푚표푙

Thus the initial proton process cannot be ruled out by the thermodynamic argument.

However, forming an anion in the highly hydrophobic distal pocket requires a large

amount of solvation energy. Thus this pathway is also very unlikely.

c. PCET describes a process that ET and PT will take place simultaneously resulted in a

lower free energy for the whole process.

° −1 −1 −1 ∆퐺푃퐶퐸푇 = −(5.73 푘퐽 ∙ 푚표푙 ) ∙ (5.8 − 9.85) + (96.48 푘퐽 ∙ 푚표푙 푉 )

∙ (0.46 − 0.305) = 38.16 푘퐽 ∙ 푚표푙−1

PCET mechanism transfers the 1H+ and 1e- in a single kinetic step, circumvents the

high energy intermediates, resulted in a significantly lower activation energy.31

The reactions between H2Q and several transition metal complexes have been shown to undergo a PCET mechanism.32-34 One example that has a resemblance to DHP A is an iron- protoporphyrin-IX model compound that reacts with H2Q in a separate PCET mechanism, in which the electron acceptor and proton acceptor are far apart. In the iron-protoporphyrin-IX model compound, the heme iron accepts the electron whereas the propionate group serves as the proton acceptor.35 Although DHP A contains a similar prosthetic group heme b, compared to the iron-protoporphyrin-IX model compound, our study shows that DHP A utilizes the distal

55 histidine His as a proton acceptor. Thus, we conclude that the reaction between H2Q and

119

ferric DHP A also undergoes a separated PCET mechanism. The distance between Nε of distal histidine His55 to the heme iron is 4.8 Å in the closed conformation as measured in the wild type DHP A X-ray crystal structure.8 The distal histidine His55 has been shown to be essential for catalysis in DHP A. For example, the H55D mutant, which replaces histidine with aspartate mutant ten-fold lower reactivity 9 and the H55V mutant has no measurable catalytic activity.36

The distal histidine, His55, serves as the proton acceptor in the PCET transfer mechanism only when it maintains the closed conformation. Protonation of His55 causes it to rotate out into the solvent-exposed open conformation. Thus, there is a tight coupling of His55 with the function of DHP A. The central role of His55 in catalysis and protection has been discussed in a number of studies. 1,9,12,13 This work suggests that His55 could be important as well in the electron transfer reaction required to complete the reaction cycle.

The oxidation of H2Q by DHP and the connection to the lag phase

Scheme 5.4. Proposed mechanism of the lag phase due to the presence of H2Q

120

The catalytic reactions carried out by ferric DHP A follow a reaction cycle similar to

CcP. Normal peroxidase chemistry consists activation of the ferric state by binding of H2O2 to form the high valent active species Compound ES and Compound II, which are the initial active intermediates formed in HRP37 and CcP,16 respectively. Subsequently, substrate is oxidized in two one-electron oxidations and the heme is reduced back to the ferric state by these electrons, which enables it to carry out another catalytic cycle (Scheme 5.4). However, due to its very high reduction potential, DHP A can perform chemistry that is not possible for the other members of the peroxidase family. Specifically, DHP A and B can initiate the peroxidase reactions starting from the oxyferrous state. In fact, the oxyferrous state has been shown to be as competent as the ferric state as the starting point of peroxidase reaction

17 catalyzed by DHP. Since H2Q reduces ferric heme to the oxyferrous state, it would inactivate most peroxidases, but it does not have this effect on DHP. Furthermore, a reductant of some kind is essential in order to complete the oxyferrous peroxidase cycle shown in Scheme 5.4. It is possible that the observed behavior of H2Q is analogous to the native reductant, which may even involve transient hydroquinone formation by 2,6-DBQ. Thus, the finding of this study may have relevance for native behavior. Since DHP has a native oxygen transport function it is reasonable to suppose that DHP A initiates peroxidase chemistry from the oxyferrous state in vivo and that the reactions observed here are analogous to native reduction chemistry.

Besides acting a reductant as shown in Scheme 5.4, H2Q plays an additional role as an inhibitor. H2Q is oxidized to 1,4-BQ in the process of reducing oxyferrous DHP A to form the ferryl species compound II. The inhibition of this process by 4-BP is consistent with the hypothesis that H2Q binds in the inhibitor site in the distal pocket. The observed greater affinity

121

for 4-BP is consistent with the trend in the 4-halophenols, which have increasing dissociation constants in the order 4-FP > 4-CP > 4-BP > 4-IP. H2Q is sterically closest to 4-CP, which means it should be displaced by 4-BP as observed. If this model is correct then the inhibition of 2,4,6-TBP will persist in the presence of H2Q until it is completely consumed by catalytic oxidation in the presence of H2O2. This behavior would give rise to the observed lag phase since only after all of the H2Q has been converted to 1,4-BQ, can 2,4,6-TBP begin to be catalytically oxidized. Consistent with the proposed role for H2Q, 1,4-BQ was shown to be a very poor inhibitor, since its Ki is ~ 3.91 mM (Figure D16, D17). Thus, its inhibition effect can be ignored and DHP A returns to a normal peroxidase cycle. As expected from the lag phase kinetics, the magnitude of kcat/Km for H2Q is in slightly greater than kcat/Km for the native

-1 substrate. Based on our data, we find that kcat/Km for H2Q is ~4.3 mM s , which is intermediate

-1 between 2,4,6-TBP and 2,4,6-TCP, for which kcat/Km is ~2.0 and ~6.6 mM s , respectively.

The hypothesis that H2Q binds in the 4-BP (inhibitor) binding site in the distal pocket is consistent with steric interactions that are known in the distal pocket of DHP. Based on the inhibition of H2Q oxidation by 4-BQ one would conclude both 4-BQ and H2Q compete for the same internal binding site. Although the binding of H2Q in the inhibitor site is inhibitory for

2,4,6-TCP or 2,4,6-TBP turnover, that same site can serve as an active site for H2Q since it is activated for oxidation by electron transfer. Moreover, the hydroxyl group of H2Q bound in the inhibitor site is immediately in contact with the heme leading and near His55 to the possibility of rapid PCET. We have shown elsewhere that the inhibitor binding site is different from the 2,4,6-TBP substrate binding site3 and a recent study reveals two more binding sites for 2,4,6-TCP in the distal pocket, which are distinct from the inhibitor binding site.4 Binding

122

of substrate in any of these sites is excluded by binding of a molecule in the inhibitor site.

Thus, we find that H2Q binding in the inhibitor site provides a consistent explanation for the experimental observations, although it is still not proven conclusively.

In summary, the present study establishes that H2Q plays a unique role in the chemistry of DHP. We initiated this study in order to understand whether the reduction of the DHP that would complete the catalytic cycle and thereby resolve the functional paradox that arises from activation of an oxyferrous protein for peroxidase chemistry. In order to complete the reaction cycle there must be a reducing agent that returns DHP to the oxyferrous state. While H2Q itself is not known play that role, the present study shows that a hydroquinone has the potential to play the role of a reducing agent that completes the catalytic cycle. Moreover, the possibility of PCET suggests the possibility an alternative activation pathway in DHP that would involve reduction of bound O2 as occurs in oxygenases and oxidases. The functional complexity of dehaloperoxidase-hemoglobin continues to provide interesting examples the role that multifunctional proteins can play in marine ecosystems.

123

5.6 References

(1) Thompson, M. K.; Davis, M. F.; de Serrano, V.; Nicoletti, F. P.; Howes, B. D.;

Smulevich, G.; Franzen, S. Biophysical Journal 2010, 99, 1586.

(2) de Serrano, V.; Franzen, S. Peptide Science 2012, 98, 27.

(3) Zhao, J.; de Serrano, V.; Zhao, J. J.; Le, P.; Franzen, S. Biochemistry 2013, 52, 2427.

(4) Wang, C.; Lovelace, L. L.; Sun, S.; Dawson, J. H.; Lebioda, L. Biochemistry 2013, 52,

6203.

(5) Phillips, G. N.; Arduini, R. M.; Springer, B. A.; Sligar, S. G. Proteins-Structure

Function and Genetics 1990, 7, 358.

(6) Gajhede, M.; Schuller, D. J.; Henriksen, A.; Smith, A. T.; Poulos, T. L. Nature

Structural Biology 1997, 4, 1032.

(7) Chen, Z.; De Serrano, V.; Betts, L.; Franzen, S. Acta Crystallographica Section D-

Biological Crystallography 2009, 65, 34.

(8) de Serrano, V.; Chen, Z. X.; Davis, M. F.; Franzen, S. Acta Crystallographica Section

D-Biological Crystallography 2007, 63, 1094.

(9) Zhao, J.; de Serrano, V.; Dumarieh, R.; Thompson, M.; Ghiladi, R. A.; Franzen, S. J

Phys Chem B 2012, 116, 12065.

(10) Smirnova, T. I.; Weber, R. T.; Davis, M. F.; Franzen, S. Journal of the American

Chemical Society 2008, 130, 2128.

(11) Franzen, S.; Thompson, M. K.; Ghiladi, R. A. Biochimica Et Biophysica Acta-Proteins and Proteomics 2012, 1824, 578.

124

(12) Nicoletti, F. P.; Thompson, M. K.; Howes, B. D.; Franzen, S.; Smulevich, G.

Biochemistry 2010, 49, 1903.

(13) Thompson, M. K.; Franzen, S.; Ghiladi, R. A.; Reeder, B. J.; Svistunenko, D. A. J Am

Chem Soc 2010.

(14) Yang, F.; Phillips, G. N. Journal of Molecular Biology 1996, 256, 762.

(15) Feducia, J.; Dumarieh, R.; Gilvey, L. B. G.; Smirnova, T.; Franzen, S.; Ghiladi, R. A.

Biochemistry 2009, 48, 995.

(16) Goodin, D. B.; Mauk, A. G.; Smith, M. Proceedings of the National Academy of

Sciences of the United States of America 1986, 83, 1295.

(17) D'Antonio, J.; Ghiladi, R. A. Biochemistry 2011, 50, 5999.

(18) Du, J.; Sono, M.; Dawson, J. H. Biochemistry 2010, 49, 6064.

(19) D'Antonio, J.; D'Antonio, E. L.; Thompson, M. K.; Bowden, E. F.; Franzen, S.;

Smirnova, T.; Ghiladi, R. A. Biochemistry 2010, 49, 6600.

(20) Davydov, R.; Osborne, R. L.; Kim, S. H.; Dawson, J. H.; Hoffman, B. M. Biochemistry

2008, 47, 5147.

(21) Davydov, R.; Osborne, R. L.; Shanmugam, M.; Du, J.; Dawson, J. H.; Hoffman, B. M.

J Am Chem Soc 2010, 132, 14995.

(22) Franzen, S.; Sasan, K.; Sturgeon, B. E.; Lyon, B. J.; Battenburg, B. J.; Gracz, H.;

Dumariah, R.; Ghiladi, R. Journal of Physical Chemistry B 2012, 116, 1666.

(23) Ma, H.; Thompson, M. K.; Gaff, J.; Franzen, S. J Phys Chem B 2010, 114, 13823.

(24) Osborne, R. L.; Taylor, L. O.; Han, K. P.; Ely, B.; Dawson, J. H. Biochem Biophys Res

Commun 2004, 324, 1194.

125

(25) Nienhaus, K.; Deng, P.; Belyea, J.; Franzen, S.; Nienhaus, G. U. J Phys Chem B 2006,

110, 13264.

(26) Zhao, J.; Franzen, S. The Journal of Physical Chemistry B 2013, 117, 8301.

(27) Tsai, A.-L.; Berka, V.; Martin, E.; Olson, J. S. Biochemistry 2011, 51, 172.

(28) D'Antonio, E. L.; Chen, T. K.; Turner, A. H.; Santiago-Capeles, L.; Bowden, E. F.

Electrochemistry Communications 2013, 26, 67.

(29) Warren, J. J.; Tronic, T. A.; Mayer, J. M. Chemical Reviews 2010, 110, 6961.

(30) Franzen, S.; Gilvey, L. B.; Belyea, J. L. Biochimica Et Biophysica Acta-Proteins and

Proteomics 2007, 1774, 121.

(31) Hammes-Schiffer, S. Accounts of Chemical Research 2009, 42, 1881.

(32) Binstead, R. A.; Mcguire, M. E.; Dovletoglou, A.; Seok, W. K.; Roecker, L. E.; Meyer,

T. J. Journal of the American Chemical Society 1992, 114, 173.

(33) Song, N.; Gagliardi, C. J.; Binstead, R. A.; Zhang, M. T.; Thorp, H.; Meyer, T. J.

Journal of the American Chemical Society 2012, 134, 18538.

(34) Lam, W. W. Y.; Lee, M. F. W.; Lau, T. C. Inorganic Chemistry 2006, 45, 315.

(35) Warren, J. J.; Mayer, J. M. Journal of the American Chemical Society 2011, 133, 8544.

(36) Franzen, S.; Belyea, J.; Gilvey, L. B.; Davis, M. F.; Chaudhary, C. E.; Sit, T. L.;

Lommel, S. A. Biochemistry 2006, 45, 9085.

(37) Miller, V. P.; Goodin, D. B.; Friedman, A. E.; Hartmann, C.; Demontellano, P. R. O.

Journal of Biological Chemistry 1995, 270, 18413.

126

CHAPTER 6

Structure and Backbone Dynamics of Dehaloperoxidase-Hemoglobin A:

Insights from NMR Relaxation Spectroscopy and

Molecular Dynamics Simulations

127

6.1 Abstract

Dehaloperoxidase-hemoglobin (DHP) is the first hemoglobin identified with a biologically relevant peroxidase function. Subsequent research has shown that DHP is a multifunctional protein with peroxygenase and oxidase activities as well. Herein we report a study of the protein backbone dynamics using heteronuclear NMR relaxation methods and molecular dynamics (MD) simulations. The results show that DHP’s backbone helical regions have restricted local motions of amide N-H bonds on the picosecond – nanosecond time scale with an average order parameter of S2 = 0.87 ± 0.03. The loop regions are significantly more flexible and have consistently smaller order parameters of S2 0.76 ± 0.08. Furthermore, DHP is primarily a monomer in solution based on the overall tumbling correlation time τm (9.49 ns) calculated from model-free analysis using the program relax. However, the NMR relaxation data indicate that amide protons in the dimer interface observed in X-ray crystallography were observed to experience chemical exchange suggesting protein-protein interactions between

DHP monomers in solution. Greater conformational flexibility was also observed in certain methionine residues, which may be consistent with their role in maintaining the protein in a ferrous oxidation state. Finally, the program caver3.0 was used to identify tunnels inside DHP obtained from MD simulation snapshots that are consistent with known Xe binding sites. These tunnels provide diffusion pathways for small ligands such as O2, H2O and H2O2 to enter the distal pocket independently of the trajectory of substrates and inhibitors, both of which are aromatic molecules.

128

6.2 Introduction

Dehaloperoxidase-hemoglobin (DHP) is a ceolomic hemoglobin first isolated from the marine annelid Amphitrite ornata.1 As a member of globin superfamily, DHP shares s similar

3-over-3 -helical bundle folded structure with many myoglobins and hemoglobins.2

However, the DHP sequence is truncated relative to the majority of globins. DHP possesses

137 amino acids while typical mammalian globins have circa 153 amino acids. Moreover,

DHP’s sequence homology with other globins is less than 20%.3 DHP was first discovered in

1977 as a hemoglobin that responsible for oxygen transport. The second function of DHP was discovered in 1996 when the dehaloperoxidase function was observed.1,4 In 2014, DHP was found to also possess peroxygenase and oxidase functions.5 Therefore, DHP is a multifunctional protein with a relatively simple structure, thus it may serve as a model for hemoprotein structure-function relationships.

Since the peroxidase, peroxygenase and oxidase functions of DHP share the same catalytic intermediates: Compound 0, Compound I/Compound ES and Compound II,5,6 both ligand and substrate binding in DHP may play a role in regulating these functions. Ligand binding, which includes either O2 or H2O2 binding to the heme Fe determines whether the activity will occur by an oxidase or peroxidase/peroxygenase pathway, respectively.

Substrates, which include brominated phenols, pyrroles or indoles, may regulate DHP by an allosteric mechanism as well. There are at least three known substrate binding sites and one inhibitor binding site in DHP.7-9 Allostery has precedent in hemoglobins, in which ionic or small molecule ligand binding alters oxygen binding affinity and cooperativity between subunits.10,11 The multi-functional properties of DHP are consistent with the relatively large

129

volume of the distal pocket above the heme, which can accommodate a variety of brominated aromatic substrates mentioned above and even the amino acids phenylalanine and tyrosine.7,12-

15 Such internal substrate binding in the distal cavity is unprecedented among hemoglobins.

Binding of large molecules in the distal pocket usually arises from adventitious binding, which is frequently associated with ligation to the heme iron. DHP, on the other hand, has sufficiently high affinity for both substrates and inhibitors that these molecules can be observed in internal sites in X-ray crystal structures even when the molecules are soaked into preformed crystals.

We have advanced the hypothesis the binding of these molecules and their dynamic interactions with the protein modulate the various functions of DHP. Stated another way, the evidence suggests that the binding of a particular aromatic substrate triggers specific enzymatic activity.

Substrate and inhibitor binding have been studied using a variety of methods. The binding of various inhibitors and substrates has been characterized by X-ray crystallography, which showed native substrate 2,4,6-TBP deeply bound in the distal pocket on the -edge of the heme,7 and by 1H NMR, 19F NMR and 1H-15N HSQC spectroscopy that shows internal binding substrate causes major chemical shift deviation on residues in the distal pocket.16,17

Fourier transform infrared spectroscopy (FTIR) has also been used to characterize interactions between substrate and bound CO. These measurements are complemented by studies of photo- dissociation of the CO molecule by temperature derivative spectroscopy and flash photolysis to give further proof that 2,4,6-TBP, 4-BP and other molecules bind in the distal pocket and modulate the binding of diatomic ligands to the heme Fe.18 Resonance Raman spectroscopy has been used to provide the interactions of binding of substrate and inhibitor with the H2O

130

ligand that weakly coordinates to the ferric heme on the axial position. 19 To add to the structural picture, an efficient and quantitative method to determine the relative binding affinity of different internal binding substrates was developed by using the competitive fluoride ion binding in the titration experiments.13 All of these experimental results point to the fact that

DHP is capable of accommodating a large number of different of aromatic molecules in the distal pocket and that there are at least two modes of internal binding in addition to the external substrate binding site on the protein surface. The external or surface binding site permits the rapid release of a substrate radical intermediate in a ping-pong peroxidase mechanism,20,21 The entry of substrates and inhibitors into the distal cavity required for peroxygenase and oxidase activity, involves a relatively dynamic fluctuation in structure, which has been studied by

Steered Molecular Dynamics (SMD) simulations.22 The dynamic fluctuations involved in the interaction of aromatic molecules are distinct from the much smaller changes in structure required for diatomic O2 or H2O2 binding.

Each of the four characterized functions of DHP, specifically the oxygen storage and transport, peroxidase, peroxygenase and oxidase functions each require either O2 or H2O2 to diffuse into the distal pocket to reach the heme Fe of DHP. One method for identifying paths is to use a tunnel calculation based on the X-ray crystal structure coordinates.2,8,9,12,23-28 Since tunnels tend to have quite small radii the analysis must be combined with dynamic changes in structure to properly account for small molecule entry and exit. The dynamics are even more important for entry and exit of substrates, which involves much larger conformational changes than O2 and H2O2. Therefore, characterization of DHP’s protein dynamics and its tunnels will

131

provide crucial insight into its structure-function relationship and the mechanism whereby protein-ligand interactions result in multi-functional capability in DHP.29,30

The present work describes a heteronuclear relaxation study of dehaloperoxidase- hemoglobin (DHP) in solution by NMR spectroscopy. The NMR relaxation measurements provide insight into DHP’s backbone dynamics on the picosecond and nanosecond time scales.

We base our analysis of the relaxation dynamics on the amide proton assignments from a previous NMR study.17 Combined with molecular dynamics simulations and previous X-ray crystallography structures of DHP with two Xe binding sites highlighted,26,27 we are able to map out the tunnel networks inside DHP’s protein matrix and correlates this structure with

DHP’s backbone dynamics pattern. This will provide the important advance in our understanding of the multi-functional nature of DHP.

6.3 Material and Methods

Protein Expression, labelling, and purification

The pET-16b vector containing the 6XHis-tagged DHP A gene was transformed into

BL21 (DE3) E.coli cells. The cells were plated onto LB agar plates containing 100 μg/mL ampicilin (Amp) and allowed to grow for about 18 h at 37 °C. Subsequently, E. coli colonies were transferred to 5 ml × 4 LB broth starter growth supplemented with 100 μg/mL ampicillin.

The starter growths were incubated in a shaker at 37 °C for 8 h. The cells were pelleted from the starter growth via the centrifugation at 5000 rpm for 20 min at 4 °C. The cell pellet was

15 resuspended in 10 ml × 4 of N isotopically labeled M9 minimal medium (6.5 g Na2HPO4,

15 3.0 g NaH2PO4, 0.5g NaCl, 1.0g NH4Cl, 4.0 g D(+)-glucose, 120 mg MgSO4, 11 mg CaCl2,

132

10 mg biotin, 10 mg thiamine-HCl, 100 mg ampicillin, 10 mg hemin for each 1L) and these starter culture were incubated at 37 °C for 8h. Then this starter culture was used to inoculate

1L × 4 of M9 minimal media in a 1L flask. The cells were grown at 37 °C and agitated at 230 rpm for 16 h. The cells were then induced with 0.3 mM IPTG for another 12 h while the incubation temperature lowered to 25 °C. The purification of 15N labeled his-tagged protein followed the exact same procedure for non-labelled his-tagged protein as shown previously.

NMR Sample preparation and heteronuclear relaxation measurement

15N labeled protein samples were oxidized to ferric state and then prepared at a concentration of 0.8 mM ~ 1.0 mM in 100 mM KPi buffer, pH 7.0, containing 10% D2O, 5 mM KCN to produce the metcyano form of DHP A. This produced a six-coordinated low spin

(6cLS) ferric heme species with a spin quantum number (S = 1/2) that can be easily monitored using UV-Vis spectroscopy (Figure E1A). The binding affinity of cyanide ion to the heme was determined by a spectroscopic titration method, giving the dissociation constant Kd = 29 ± 12 nM (Figure E1B). The strong ligand field of the cyanide ligand maintains the heme in the 6cLS ferric state (S = 1/2). Because the distance dependence for dipole-dipole coupling follows a

1/r6 relation, the paramagnetic effect on backbone amide 15N relaxation is negligible when the

Fe-15N distance is greater than 7.0 Å. Based on analysis of the paramagnetic shifts we have observed hyperfine shifts of the heme methyl groups and of the side chains of Phe97 and His89.

Those studies did not provide evidence for hyperfine shifts of any of the amide protons.

Therefore, the paramagnetic effect on relaxation was ignored in our analysis.

All NMR experiments were conducted at 298K on the 500 MHz Bruker AVANCE II spectrometer and 700 MHz Bruker AVANCE III spectrometer equipped with a QCI

133

CyroProbe. Relaxation delay for T1 were: 0.1, 0.2, 0.3, 0.5, 0.6, 0.8, 1.0, 1.4 s for 500 MHz;

0.2, 0.3, 0.5, 0.7, 0.9, 1.1, 1.4, 1.7 s for 700 MHz. Relaxation delay for T2 were: 17, 34, 68, 85,

102, 136, 170, 237 ms for 500 MHz; 17, 34, 51, 68, 85, 102, 136. 170 ms for 700 MHz. For both T1 and T2 experiment, 8 transients were measured during one FID with a recycle delay time of 3.0 s. For the {1H}-15N Nuclear Overhauser Effect (NOE) experiments, spectra were recorded with and without the 1H saturation in 10 s recycle time. Proton saturation was achieved with a train of 120 pulses prior to nitrogen excitation.

All spectra were processed using Bruker software suite Topspin 3.2. The 1H dimension was zero-filled to 4096 points while the indirectly detected 15N dimension was zero-filled to

2048 points. Both dimensions were apodized with shifted qsine. The peak list of backbone amide N-H resonance of DHP A was generated based on previous assignment.17 The processed spectra were analyzed with Bruker Protein Dynamics Center. T1 and T2 were obtained by fitting the cross peak intensity over the time delay series to a single-exponential function I(t) = I(0) exp(-t/T1) and I(t) = I(0) exp(-t/T2). The standard error of the relaxation times were obtained from the uncertainty of the fits. The steady-state NOE values were determined from the ratio of peak intensities of the spectra with and without 1H saturation.

Model-free analysis using relax

The open-source program relax 4.0.0 equipped with the dual optimization approach of model-free parameters and the global diffusion tensor proposed by d’Auvergne and Gooley was used to analyze the 15N spin relaxation data of DHP A.31,32 Global diffusion and local

“model-free” model selection was performed. First, the local correlation time (τm) was estimated for each spin without the global diffusion tensor. Subsequent diffusion optimization

134

was achieved by fitting data into four specific diffusion models: sphere; oblate spheroid; prolate spheroid and ellipsoid with local model-free approach applied to each residue. Once all diffusion models converged, the best global diffusion model was used for each residue. 33 The extended model set (labeled m0 to m9) was used to describe the internal motion of amide N-H

2 2 2 bond. These models are: m0 : {τm}; m1 :{τm , S }; m2 : {τm , S , τe}; m3 : { τm , S , Rex}; m4 :

2 2 2 2 2 2 2 {τm , S , τe, Rex}; m5 : {τm , S , Sf , τs}; m6 : {τm , S , τf , Sf , τs}; m7 : {τm , S , Sf , τs, Rex}; m8 :

2 2 2 {τm , S , τf , Sf , τs, Rex}; m9 : {Rex}, where S is the squared generalized order parameter, τe is the effective internal correlation time in the original model-free formalism, τf and τs are the fast

(picosecond) and slow (nanosecond) effective internal correlation time in the extended model- free formalism. Rex is the contribution to the R2 (T2) that ascribes slow processes on the microseconds-milliseconds time scale to chemical exchange. The error propagation for all fitting parameters was calculated from Monte Carlo simulations. 31,32 The CSA (15N chemical shift anisotropy) value Δ was set to -172 ppm and the average amide N-H bond distance rN-H was set to 1.02 Å.

Molecular dynamic simulations and trajectory analysis

Molecular dynamics (MD) simulations were performed using the scalable molecular dynamics program, NAMD. 34,35 The X-ray structure of the wild type ferrous-CO DHP A

4DWU were used to construct isoelectronic metcyano form of DHP A for the MD simulations.

55 Due to the allosteric behavior of distal histidine His , both Nδ and Nε histidine tautomers were simulated in parallel. NAMD simulations were carried out with periodic boundary conditions with a model solvated and placed in a unit cell of dimensions 50.5 Å × 57.2 Å × 60.7 Å contains

9254 water molecules and with Na+ and Cl− ions added to give an ionic strength of 0.15 M.

135

The cutoff was set to 12 Å with a switching distance of 1.5 Å. The time step of 2 fs was used consistent with application of the SHAKE algorithm.36 The calculations were carried out for a total of 40 ns simulation for each trajectory. Each simulation was preceded by minimization and a 5 ns equilibration phase. The global rotational and translational motions were removed for each 40 ns trajectory by using carma, which applies the Kabsch’s algorithm to least-squares fit all frames to a reference frame. 37 40000 frames and unit vectors of amide N-H bond of 134 residues out of 137 (except the N terminus Gly1 and Pro29, Pro75) were extracted and calculated from each 40 ns simulations trajectory. The autocorrelation function of internal motion of amide N-H bond CI(t) (Equation 13) is calculated from the unit vector trajectory using a homemade Python script, of which the average of the ensemble is calculated for a 10 ps lapse increment from 0 to 20 ns. The squared generalized order parameter S2 is evaluated using a 1 ns time window as the autocorrelation function converged.

Tunnel Calculation

Caver 3.0 was used to explore the potential gas or solvent tunnels inside the met-cyano DHP

A structure.38 50 snapshots during the 4 ns molecular dynamics simulation were applied for the tunnel calculations. Three positions used as the starting points to probe the tunnels were located: 1. In the distal pocket above the heme; 2. On the α-edge of the heme and 3. On the proximal side beneath the heme. The clustered time-independent trajectories were used to construct the tunnel structures inside the protein structure and visualized in Pymol. 39

136

6.4 Results

1 The relaxation of all amino acid residues was studied based on data for R1, R2 and { H}-

15N NOE measurements at both 500 MHz and 700 MHz magnetic fields. Based on the criterion that all three experiments must be carried out at two magnetic fields, 118 out of 137 residues were available for analysis. 17 The relaxation data are shown in Figure 6.1, plotted by residue number and annotated with the corresponding secondary structure. The average values of R1,

1 15 R2 and { H}- N NOE are presented in Table 6.1, the complete dataset for each residue is given in the Table E1. We have observed {1H}-15N NOE values are consistently significantly lower in the EF and GH loop regions. It is the noteworthy that the EF loop region contains the dimer interface of DHP that has been observed by X-ray crystallography. Residues Thr71, Asp72 and

Asp126 constitute the dimer interface by forming an intermolecular hydrogen bonding network.2,12,23-26,28 Given the geometric proximity of two Cys73 in the DHP A crystallographic dimer one may suspect that these cysteines form an intermolecular disulfide bond. However, this interaction is still not proven. Although the analysis of SAXS data suggests that DHP primarily exists as monomer in solution and less than 10% forms a dimer,40 the favorable non- covalent contact between DHP monomer subunits must be sufficiently important in solution that it can give rise to the long-range contacts for residues in the EF loop regions indicated by significantly lower {1H}-15N NOE values.

137

1 15 Figure 6.5 NMR relaxation data { H}- N NOE, R1, R2 and R2/ R1 that measured at 500 MHz (red) and 700 MHz (green). The α-helices are shown and labeled at the bottom (purple) and corresponding α-helical regions are shaded as column inside the plot.

1 15 Table 6.6 Average values of R1, R2 and { H}- N NOE relaxation data

500 MHz 700 MHz {1H}-15N NOE 0.760 ± 0.065 0.821 ± 0.075 -1 R1 (s ) 1.577 ± 0.086 1.082 ± 0.062

-1 R2 (s ) 11.67 ± 0.96 13.32 ± 1.24

Reduced Spectral Density Mapping and Consistency Testing

1 15 The relaxation rate R1, R2 and { H}- N NOE can be explicitly expressed in terms of five spectral density functions: J(0), J(ωN), J(ωH + ωN), J(ωH) and J(ωH – ωN) for the amide

N-H spin pair (Equations 1-3).41 The parameters d and c represent the relaxation mechanisms

138

ascribed to dipole-dipole interaction, which are through-space coupling terms between a pair of nuclear spins and chemical shift anisotropy (CSA), respectively. The spectral density function provides probability of finding motions at given angular frequency ω, which is

15 attributed to N relaxation.

푑2 푅 = [퐽(휔 − 휔 ) + 3퐽(휔 ) + 6퐽(휔 + 휔 )] + 푐2퐽(휔 ) (1) 1 4 퐻 푁 푁 퐻 푁 푁

푑2 푐2 푅 = [4퐽(0) + 퐽(휔 − 휔 ) + 3퐽(휔 ) + 6퐽(휔 ) + 6퐽(휔 + 휔 )] + [3퐽(휔 ) + 4퐽(0)] (2) 2 8 퐻 푁 푁 퐻 퐻 푁 6 푁

푑2 훾 [6퐽(휔 + 휔 ) − 퐽(휔 − 휔 )] 푁푂퐸 = 1 + 퐻 퐻 푁 퐻 푁 (3) 4 훾푁 푅1

휇0ℎ훾푁훾퐻 1 휔푁 Where 푑 = 2 〈 3 〉 푎푛푑 푐 = ∆ , µ0 is the permeability of free space, h is Planck’s 8휋 푟푁퐻 √3

1 15 constant, γH and γN are gyromagnetic ratio of H and N nuclei, respectively, rNH is the amide bond distance that set to 1.02 Å and Δ is 15N chemical shift anisotropy (CSA) that set to -172 ppm.

It is not possible to uniquely determine five spectral density values from the three

1 15 measured relaxation data: R1, R2 and { H}- N NOE. Therefore, a standard approximation has

42,43 been made for the high Larmor frequency terms such that J(ωH) ≈ J(ωH + ωN) ≈ J(ωH – ωN).

It was further assumed that these high Larmor frequency terms are all substituted by J(0.87ωH), which reduced the number of unknown spectral densities from five to three. Thus, the explicit expression for the spectral density J(0), J(ωN) and J(0.87ωH), can be uniquely determined as shown below (Equations 4-6). This method is referred as the reduced spectral density mapping

42,43 method. The average values of J(0), J(ωN) and J(0.87ωH) are presented in Table 6.2. The complete dataset is given in the Table E2.

139

4 훾퐻 퐽(0.87휔퐻) = 2 푅1(푁푂퐸 − 1) (4) 5푑 훾푁

7푑2 푅1 − 퐽(0.87휔퐻) 퐽(휔 ) = 4 (5) 푁 3푑2 + 푐2 4

3푑2 푐2 13푑2 푅 − ( + ) 퐽(휔 ) − 퐽(0.87휔 ) 2 8 2 푁 8 퐻 퐽(0) = (6) 푑2 2푐2 + 2 3

Table 6.7 Average of J(0), J(ωN) and J(0.87ωH) values

500 MHz 700 MHz J(0) (ns/rad) 3.32 ± 0.29 3.27 ± 0.32

J(0.87ωH) (ps/rad) 5.88 ± 1.42 3.04 ± 1.33

J(ωN) (ns/rad) 0.31 ± 0.02 0.18 ± 0.01

It is a good practice to conduct consistency testing before using the two datasets for further analysis.44 Inconsistency between datasets obtained at 500 and 700 MHz may arise from variations of pH, temperature, concentration, or the water suppression pulse sequence applied during the acquisition. The consistency of the datasets can be assessed by comparing J(0), the spectral density at the zero frequency that is independent of the strength of the magnetic field.

The correlation plot of two J(0) for each residue that calculated from reduced spectral density mapping were plotted in Figure E2. The ratios between J(0) obtained at 700 MHz and 500

MHz have been plotted as the histogram in Figure E2. The distribution of the histogram was fitted to a Gaussian function (Blue). The center is at 0.972 and standard deviation is ± 0.089.

Thus, two magnetic field datasets have a reasonable mutual consistency.

140

Geometric Interpretation of J(ωN) and J(0.87ωH) via Lapari-Szabo Mapping

Although spectral density values J(0), J(ωN) and J(0.87ωH) can be obtained from the reduced spectral density mapping method, a clear physical description of the motion is still not available, especially since the reduced spectral density contains the superposition of different time-scale motional information. The model-free analysis of NMR relaxation data proposed by Lipari and Szabo solves this difficulty based on a simple idea that the correlation function of amide N-H bond vector’s global motion can be factored into the correlation functions of protein overall tumbling and amide N-H bond’s internal motions (Equation 7).45,46 As a general rule, the correlation function of internal motion of amide N-H bond can be approximated by a squared generalized order parameter S2 ∈ [0, 1] and a single exponential decay (Equation 8). Where S2 = 0 means bond vector is very flexible that isotropically sampling all orientations, whereas S2 = 1 means bond vector is very rigid without any local internal motion. Thus, in the Lipari-Szabo model-free formalism, the spectral density function

2 J(ω) can be obtained by conduct a Fourier transform of the correlation function CI(t),where S is the squared generalized order parameters describe the spatial restriction of the internal motion, τm is the overall tumbling correlation time and τe is the effective internal motion correlation time. (Equation 9)

퐶푔푙표푏푎푙(푡) = 퐶푂(푡)퐶퐼(푡) (7)

푡 − 2 2 휏 퐶퐼(푡) = 푆 + (1 − 푆 )푒 푒 (8)

2 2 2 푆 휏푚 (1 − 푆 )휏 1 1 1 퐽(휔) = [ 2 + 2] , = + (9) 5 1 + (휔휏푚) 1 + (휔휏) 휏 휏푚 휏푒

141

To analyze and visualize the reduced spectra density pairs J(ωN) and J(0.87ωH), we plotted

47 J(ωN) and J(0.87ωH) values as Cartesian coordinates and applied Lipari-Szabo mapping.

Given an expression for the spectral density function of rigid isotropic tumbling (Equation

(rigid) (rigid) 10), its Cartesian coordinates (퐽푁 (푡), 퐽퐻 (푡)) give a triangle shape-like trajectory as the function of time (0 < t < ∞) as shown in Figure 6.2.

2 푡 퐽 (rigid) (푡) = 푁 5 1 + (휔 푡)2 푁 (10) (rigid) 2 푡 퐽퐻 (푡) = 2 { 5 1 + (0.87휔퐻푡)

The expression of J(ωN) and J(0.87ωH) can be given as a linear combinations of the spectral

2 density of two rigid forms at τ and τm dictated by order parameter S as shown below (Equation

11):

2 (rigid) (rigid) (rigid) 퐽(휔푁) = 푆 (퐽푁 (휏푚) − 퐽푁 (휏)) + 퐽푁 (휏) { (11) 2 (rigid) (rigid) (rigid) 퐽(0.87휔퐻) = 푆 (퐽퐻 (휏푚) − 퐽퐻 (휏)) + 퐽퐻 (휏)

By cancelling out order parameter S2, we have:

(rigid) (rigid) 퐽(0.87휔퐻) − 퐽퐻 (휏) 퐽(휔푁) − 퐽푁 (휏) (rigid) (rigid) = (rigid) (rigid) (12) 퐽퐻 (휏푚) − 퐽퐻 (휏) 퐽푁 (휏푚) − 퐽푁 (휏)

Equation 12 describes a simple geometric relationship of three points with Cartesian

(rigid) (rigid) coordinates: A = (퐽푁 (휏푚), 퐽퐻 (휏푚)) , B = (퐽(휔푁), 퐽(0.87휔퐻)) and C

(rigid) (rigid) = (퐽푁 (휏), 퐽퐻 (휏)), which are on the same line (Figure 6.2). Thus, the squared generalized order parameter S2 is simply the ratio of distances between AB and AC .

142

Figure 6.6 Application of Lipari-Szabo mapping method for analyzing the 500 MHz reduced spectral density mapping data of met-cyano DHP A. (Red) the rigid tumbling curve; (Green) the observed reduced spectral density values; (Blue) the rigid tumbling point at overall correlation time τm = 9.49 ns. The principal diffusion tensors of DHP monomer structure have been calculated using

7 -1 7 -1 7 the program Hydronmr, which gives Dz = 1.591 × 10 s , Dy = 1.842× 10 s , Dx = 1.956 × 10

-1 7 -1 7 -1 48 s . Thus, D = (Dx + Dy)/2 = 1.899 × 10 s and D║ = Dz = 1.591 × 10 s . Therefore, the

anisotropy factor is calculated as η = D║ / D= 1.2. Given the relatively small anisotropy factor for DHP A monomer and its nature as a globin protein, the spherical diffusion model was

iso applied to estimate the overall tumbling correlation time, which gives τm = 1/(6Diso) = 9.28 ns. This estimation is very close to the average τm of 118 residues calculated from program relax which gives τm = 9.49 ± 1.65 ns. The τm = 9.49 ns is used to calculate the Cartesian coordinate of rigid tumbling point as shown in the Figure 6.2. From Figure 6.2, we can see that most of the residues are located close to the bottom leg of the “triangle” rigid tumbling curve, yet all of residues maintain a short distance from the rigid tumbling point at the overall

143

correlation time τm and stay on its left side. This tells that most of residues have a large generalized order parameters S2 that is close to 1 and have relatively slow internal motions indicated by the existence of internal motion correlation time τe. Moreover, we can already identify from the graph that residues Thr71, Val105and Phe115 have relatively smaller order parameters S2, indicates that their amide N-H bond experience more flexible internal motions.

Although Lipari-Szabo mapping method provides a simple geometric method to understand and evaluate the squared generalized order parameter S2, this method is cumbersome to conduct the model selection for the model-free analysis. Moreover, its

2 application is limited due to the fact that S can only being solved analytically for model m0:{τm

2 2 }; m1:{τm , S } and m2 : {τm , S , τe}. Therefore, a comprehensive model-free analysis is conducted using the program relax.

Model-Free Analysis using relax

The two datasets obtained at 500 and 700 MHz were used to perform the model-free analysis based on the relaxation data for the 118 observed residues. The extended models set

(m0 to m9) were used for model selection and fitting, the local motions of all residues were fitted to seven models (m1 to m5, m7 and m8) that can be divided as original model-free formalism (m1 to m4) and extended model-free formalism (m5,m7 and m8).49,50 The explicit formulas of spectral density functions of eight models (m1 to m8) are given in the Appendix

144

Figure 6.7 (A) Comparison of generalized order parameter S2 obtained from NMR (blue) and MD (orange) for each residue in DHPA; (B) The difference histogram of generalized order parameters S2 between NMR and MD; (C) Histogram of relaxation attribute to chemical exchange Rex; (D) Histogram of effective correlation time of internal motion τe (ps) (blue, left axis) and slow effective correlation time τs (ns) (red, right axis) in two different time-scales. The squared generalized order parameters S2 provide a quantitative parameter to characterize the flexibility of the amide N-H bond vector with respect to the internal motion on the picosecond-nanosecond time scale. For 118 out of 137 residues with an order parameter S2 available from the NMR measurements, the average of S2 is 0.83 ± 0.10 indicating that the protein backbone is quite rigid with only limited local motions of backbone amide N-H bond.

The overall rigidity is expected for DHP as a hemoglobin with a predominantly -helical structure. The overall distribution of the order parameter S2 clearly shows a helix-loop contrast, in which G and H helices are most rigid part in the protein structure because their average S2

145

are both 0.88, whereas EF (residues 70 – 76) and GH (residues 111 – 115) loop regions are much more flexible, their average S2 are only 0.72 and 0.69 respectively. The value of order parameter S2 has been color mapped onto the structure as shown in Figure 6.4.

The best model is selected for each residue by a statistical model selection process to conduct next step in the model-free analysis. The model selection results has the following distribution among 118 residues: m1(32), m2(19), m3(17), m4(12), m5(21), m7(14) and m8(3).

46 out 118 residues are predicted to have different levels of chemical exchange indicated by the non-zero chemical exchange term Rex that can attributed to the R2 relaxation. Chemical exchange is often coupled to slow conformational motions on the microseconds-milliseconds time scale. In practice, this refers to the breaking of hydrogen bonds, which leads to deprotection of internal amide N-H to permit hydrogen ion exchange to take place. Significant chemical exchange was observed for residues on helices B and E facing towards the distal pocket of DHP A above the heme. The other major region where significant chemical exchange was detected is in the EF loop region where the dimer interface of DHP is observed in X-ray crystallography.40 69 out 118 residues experience internal motions of the amide N-H bond indicated by the non-zero effective correlation time τe (m2, m4, m5, m7 and m8). 28 out of 69 residues show internal motions on two time scales (ps and ns) that described by extended model-free formalism (m5, m7 and m8) for their amide N-H bond. Most of these residues are located in flexible regions such as N- and C- terminal and loops. The complete results of the model-free analysis are given in Table E3.

146

G A D B H A

E C F

S2 N.D. > 0.90 0.90-0.83 0.83-0.70 < 0.70 C

N.D. m1, m2, m5 m3, m4 m7, m8

Figure 6.8 Color mapping of secondary structural information and NMR model-free analysis results on the met-cyano DHP A structure. (A) Color mapped eight α-helices of DHPA with calculated tunnel structures and the two xenon binding sites. (B) Color mapped generalized order parameter S2 on DHP A structure. (C) Color mapped model selection results of model- free analysis on the DHP A structure.

147

Molecular Dynamics Simulation and Computed Order Parameter

The NMR relaxation experiment provides a direct measurement of the spectral density function J(ω). The measurement provides a connection to actual dynamics because the spectral density function is the Fourier transform of the autocorrelation function of the N-H bond vector’s motion. Thus the order parameter S2 can also be determined from the autocorrelation function of N-H bond’s trajectory if the fluctuation of N-H bond is a stationary ergodic process after removing the global motion of the protein. Such stationary trajectories can be obtained by conducting MD simulations for comparison with experiment. The autocorrelation function of the amide N-H bond vector is described in Equation 13, where P2 is the second Legendre polynomial, µ(τ) and µ(τ + t) are time-dependent unit vectors that describe the spatial orientation of N-H bond in a fixed reference frame. The S2 is the limit of autocorrelation function when time approaches to infinity (Equation 14). In practice, S2 can be determined when the autocorrelation function converges.

퐶퐼(푡) = 〈푃2[휇(휏) ∙ 휇(휏 + 푡)]〉 (13)

2 푆 = lim 퐶퐼(푡) (14) 푡→∞

Here, we have calculated S2 from MD simulations and compared the simulated values with those obtained from NMR relaxation experiments. Because distal histidine His55 may adopt closed or open conformations, both Nδ and Nε histidine tautomers were modeled in sets of five

40 ns simulations, the order parameters were calculated as the average of all five of these simulations (Figure E4). The average order parameter S2 of 134 residues from MD simulations is 0.84 ± 0.07. The order parameters calculated from MD simulations distribute in a narrower range indicated by the smaller standard deviation as compared to those from NMR. But they

148

also presented a distinctive helix-loop contrast (Figure E5). The average of S2 in helix region is 0.87 ± 0.03, whereas in loop region, it is 0.76 ± 0.08. The complete dataset of S2 calculated from MD simulation can be found at Table E4.

In general, the order parameters S2 calculated from MD simulations are consistent with those obtained from NMR relaxation experiments (Figure 6.4A, 6.4B). One major discrepancy in S2 between NMR and MD can be observed for residues that experience chemical exchange

(Figure 6.4B, 6.4C). However, when chemical exchange occurs, experimental NMR gives much lower S2 values than those calculated by MD simulations. Chemical exchange takes place on residues that experience slow motions on the microsecond-millisecond time scale. Such slow motions could not be probed by MD simulations due to limitations in computational speed, which confined the calculations to the nanosecond time scale. Therefore, a discrepancy between the simulated and experimental order parameters, S2, is expected for any residue that undergoes chemical exchange.

Tunnels in the protein interior

We have used the program Caver 3.0 to identify tunnels that connect the internal cavities of protein to the outside solvent environment. Four major tunnels were identified that directly connect the distal cavity to the protein surface. They are the AB tunnel, starting from distal pocket and ending between helix A and B; the CD tunnel, connecting distal pocket to the CD loop region; the EF tunnel, starting from α-edge of the heme plane and exiting between helix E and F; and the GH tunnel, starting from a cavity above the α-edge of the heme and passing between helices G and H. These tunnels interconnect at a central location near the back of the distal cavity where a cavity inside DHP is known to exist. The two Xe binding sites that

149

are characterized by X-ray crystallography are located within the calculated tunnels structures as shown in Figure 6.4.27 shows that the Xe1 binding site is located at starting point of tunnel

CD that is the distal pocket right above the -edge of the heme and the Xe2 binding site is on the passageway of EF tunnel that leads to protein surface.

6.5 Discussion

DHP is Primarily a Monomer in Solution

DHP is a multi-functional hemoglobin that possesses the highly conserved 3-over-3 helical bundle folding, which is observed in many myoglobins (sperm whale, horse, human etc.) or single hemoglobin subunits. As a member of hemoglobin family, DHP A may have a tendency to form dimers or multimers in solution. Since DHP A is a crystallographic dimer in every X-ray crystal structure obtained it is possible that DHP A also forms dimers in solution.

40 Furthermore, there are only two DHP genes in A. ornata, DHP A and B.51 One or both of the DHP proteins may associate as multimers to form the giant hemoglobin (erythrocruorin), which is found in the tentacles of the annelid.1

SAXS studies provide evidence that DHP A is mostly (>80%) monomer in solution. 40

The NMR relaxation study corroborates this conclusion since the correlation time consistent with the data is that of a monomeric globin. We can compare the overall tumbling correlation time τm between 3-over-3 fold as well as 2-over-2 hemoglobins (Table 6.3). We can see that the overall tumbling correlation time τm is approximately proportional to the number of subunits in hemoglobin oligomer. Because τm is approximately inverse of the diffusion tensor and proportional to the size of the molecule for spherical model. Therefore, it is evident that

150

DHP is indeed primarily a monomer in solution as measured by NMR relaxation experiments as well. However, the both SAXS and NMR also inform us that there are weak interactions in solution that result in a population of dimers or weakly associated proteins. It is of interest to put the tendency of DHP to form dimers in context of dimeric and multimeric structures, which are common among hemoglobins.52,53 Cooperativity among subunits may regulate ligand binding affinity. Subunit interactions may also induce a conformational change that triggers signal transduction and a functional switch from a tight- to a weak-binding structural form.54-

56 Although the structure of specific hydrogen bonds and salt bridges is difficult to determine for weakly interacting proteins, the NMR relaxation data can be compared to X-ray crystal structural results to search for evidence of association in solution.

DHP A and B are both observed in a dimeric form with a small surface area interface located at EF loop region that involves residue Thr71, Asp72 and Cys73 and two residues Arg122 and Asn126 on the H helix (Figure 6.5).26 The small dimer interface area of DHP is in contrast with the extensive interface region observed in the dimeric hemoglobin from Scapharca and many other hemoglobins.57,58

151

Figure 6.5 Dimer interface of DHP observed in X-ray crystallography.

There is strong evidence for a solution phase dimeric interaction similar to that observed in the

X-ray crystal structure. For example, significantly lower {1H}-15N NOE values in the EF loop region, especially residue Thr71 indicates the existence of long range contacts that may attribute to H-bonding networks or salt bridges that are expected in the dimer interface. Moreover,

Residues Thr71, Cys73 on the EF loop and Leu127 on the H helix show significantly lower order parameters S2, the existence of rapid chemical exchange and two-time scale internal motions in the backbone amide N-H bonds. At first it may appear inconsistent that an anomalously lower value of S2 is observed for residue Leu127, which is in the middle of rigid and stable H helix. However, this residue is adjacent to residue Asp126 that is observed to be one of the critical residues corresponding to the formation of dimer in X-ray crystallography. A previous

NMR study also found that the binding of 2,4,6-trichlorophenol also can have an effect on the residues in this same region. 17 In conclusion, the existence of a monomer-dimer equilibrium for DHP A in solution is well supported by evidence from NMR relaxation experiments.

152

Despite the consistent structural conclusion that the dimer interface in solution exists at the same location observed in the X-ray crystal structure, these findings do not explain the reason for the dimer interface at this location. One would not predict this interaction a priori because the surface area of the crystallographic dimer interface is significantly smaller than other possible dimer interfaces. However, one crucial feature of this interface is that it is essentially on the opposite side of the protein from the opening to the distal pocket. Thus, it is possible that the weak dimeric interactions observed in the structural experiments are part of a regulatory switch that couples small molecule (or diatomic ligand) binding to protein-protein interactions.

Table 6.8 Comparison between different Hemoglobins

2 ° Name Oligomer Fold Ligation τm (ns) < S > T ( C) Monomer Dehaloperoxidase - Hemoglobin (DHP) 3-on-3 Ferric-CN 9.49 0.83 25 (> 90%) Glycera Dibranchiata Hemoglobin59 Monomer 3-on-3 Ferrous-CO 10.3 0.93 20 Scapharca Hemoglobin I60 Dimer 3-on-3 Ferrous-CO 17.7 0.91 25 Human Hemoglobin A61 Tetramer 3-on-3 Ferrous-CO 31.5 ~ 33 N.A. 29 Truncated Hemoglobin N62 Monomer 2-on-2 Ferric-CN 10 0.84 26.4 Hemoglobin of Truncated Lineage Monomer 2-on-2 Ferrous-CO 8.4 N.A. 25 (GlbN)63

The Role of Tunnels in Function Switching

Heteronuclear NMR relaxation experiments and MD simulations provide complementary starting points for understanding protein backbone dynamics in DHP A. Both

NMR data and MD simulations indicate that DHP A has rather high backbone rigidity consistent with the high order parameters S2 for backbone amino acids.59,60,62 Table 6.3 shows that this is expected for DHP A as a member of the hemoglobin family. For example, the

153

average order parameter S2 (0.83 ± 0.10 from NMR, 0.84 ± 0.07 from MD, 25 °C) of DHP A is comparable to that of other hemoglobins (Table 6.3). Despite significant helix-loop contrast, the high rigidity of major B, E, G and H helices on the picoseconds-nanoseconds time scale are necessary to support the tunnel structures inside the protein. These dynamical tunnels likely exist in DHP A to allow it to carry out multiple functions. These channels reveal the pathways by which O2 and H2O2 can diffuse into the active site and bind to the heme Fe. The possibility for multiple entry channels may derive from the differences in the various functions. The small molecule ligands to the heme Fe must bind independently of the large aromatic substrates, such as halogenated phenols and indole derivatives. The tunnels permits rapid entry and exit of the three main small molecules O2, H2O2 and H2O. Perhaps more important than the rate of entry/exit is the fact that these tunnels suggest possible mechanisms for control.

The entry and binding of the substrates has been studied using Steered Molecular

Dynamics (SMD).22 The entry pathway for the substrate appears to be in a region between the heme and the distal histidine (His55). It is known that for myoglobin (SWMb) and DHP, the allosteric distal histidine sidechain (His64 and His55, respectively) serve as switchable gate that controls the entry and exit of diatomic gas to the distal pocket.28,64 Thus, substrate entry pathway may conflict with one common entry channel for small molecules. Thus, the redundancy in entry channels for the small molecules may be essential to permit substrates to find and then trigger a change in function. Clearly, the distal cavity not only functions as an internal substrate binding site for the peroxidase, peroxygenase and oxidase functions, but also serves as a transient docking site for unbound diatomic gas molecules. The order of binding events is crucial as has been shown in experiments using double mixing stopped flow and

154

studies of activation of the oxy species.6 Therefore, the rigid scaffold provided by the -helical bundles folding in DHP A may also modulate the tunnel structure and control gated entry of ligands. Specific examples of this type of behavior can be seen in studies of the activation of the oxy form of the protein for peroxidase activity. If H2O2 is added to oxy-ferrous DHP, the reaction to form compound II is extremely slow. 5 However, if a substrate such as 2,4,6-TCP is present this rate is significantly more rapid. Indeed, an intermediate species has been identified as the Fe(II)-H2O2 adduct using stopped-flow spectroscopy. This means that the replacement reaction of O2 by H2O2 occurs at a measurable rate only when the structure is modulated by the binding of 2,4,6-TCP, which assumed to bind externally in this functional state of DHP. This behavior may arise from controlled entry of H2O2, but also may involve the conformation of the distal histidine, which may favor binding of H2O2 over O2 in the substrate- bound conformation. In summary, tunnels may form in a transient way to allow gas molecules or even larger aromatic substrates to diffuse into the distal pocket of DHP A in response to binding of the larger substrate molecules, which act to trigger specific functions of the protein.

The slow backbone motions around the tunnels provides long enough time to let molecules to adopt favorable configuration and pass through to reach final destination in the active site.

Chemical Exchange Behavior of Methionines Revealed by Backbone Dynamics

DHP A is a sulfur rich hemoglobin that contains 1 cysteine and 6 methionine residues in the peptide sequence. If we compare DHP A to SWMb we find an interesting contrast.

SWMb has 6 histidines and 2 methionines while DHP A has 6 methionines and 2 histidines. It is worth noting that 5 out 7 sulfur containing residues (Met49, Met69, Met108, Met136 and Cys73) in DHP A have shown significant chemical exchange on their backbone amide. Chemical

155

exchange on these residues is relevant in situations where they undergo a conformational change or interact with other molecules on the microsecond to millisecond time scale. Cys73 is part of the dimer interface and thus, we may consider it separately from the 6 methionines.

Given the high reduction potential of ferric DHP A (+221 mV)65 and the antioxidant properties of methionines, it is a reasonable hypothesis that they are a major source of reducing power inside the protein. One mechanism for maintaining the heme in the ferrous state could be electron transfer from one or more of these amino acids to the heme Fe. This would maintain the heme Fe in the ferrous state at the expense of oxidizing the methionines to sulfoxides.

The maintenance of a high redox potential is essential for globin function, which requires a ferrous Fe. It has also been shown that it is consistent with peroxidase activity albeit with a modified reaction scheme that results of ligand exchange of H2O2 with Fe-O2 to give

Fe-O2H2, which directly forms compound II. Consistent with the high reduction potential

Compound II can return to the ferrous state via ferric Fe by two one-electron reductions. In fact, the redox potential of DHP A is so high that we have observed “autoreduction”.

Autoreduction is quite unusual in heme proteins and is in stark contrast to the well-known

“autoxidation” mechanism observed in nearly every myoglobin and hemoglobin.66-68

Autoreduction is commonly observed in DHP A. For example, oxidized ferric DHP A left at 4

°C overnight in air will be more than 50% converted to the oxyferrous state. The autoreduction behavior of DHP may be an essential pathway for functional switching between peroxidase or peroxygenase that primarily using ferric state as an catalytic starting point and oxygen transporter or gas sensor that using ferrous heme.5,21,69 The relatively high flexibility of the amino acids give rise to the autoreduction phenomenon has a number of possible ramifications.

156

The conformational changes in the methionines may facilitate electron transfer. It may also permit a response to substrate binding. However, the specific functions of the methionines will require site-directed mutagenesis studies to elucidate the specific role of each residue.

6.6 Conclusion

In conclusion, we are able to quantitatively describe four dynamic features of DHP A from Amphitrite ornata by using heteronuclear NMR relaxation methods combined with MD simulations. These features are 1.) -helices are uniformly rigid, 2.) turn and loop regions are consistently less rigid and more variable, 3.) the dimer interface region shows a pattern of slow exchange indicating that the crystallographic dimeric structure is transiently formed in solution, and 4.) internal motions of methionine residues may be linked to electron transfer crucial for function switching in DHP A. Overall, the dynamical behavior of DHP A modulates the narrow tunnels that permit the exit and entry of O2, H2O2 and H2O. , These tunnels are essential passageways required for DHP to carry out its multiple functions, including oxygen transport, peroxidase, peroxygenase and oxidase activity. The dynamic picture suggests that the function switching involves subtle aspects of the whole protein structure triggered by substrate binding. The one amino acid that has been conspicuously absent in the NMR study is the distal histidine, His55, which is considered to be the most important amino acid for controlling function switching in a number of other studies. It is difficult to imagine that every aspect of function switching relies on a single amino acid. The present study complements previous work that has focused on His55 and provides the most comprehensive view yet of the

157

systematic nature of conformational changes in the DHP A that may complement His55 in providing mechanisms of function switching and regulatory control.

158

6.7 References

(1) Weber, R. E.; Mangum, C.; Steinman, H.; Bonaventura, C.; Sullivan, B.; Bonaventura,

J. Comp. Biochem. Phys. A 1977, 56, 179.

(2) LaCount, M. W.; Zhang, E.; Chen, Y. P.; Han, K.; Whitton, M. M.; Lincoln, D. E.;

Woodin, S. A.; Lebioda, L. J. Biol. Chem. 2000, 275, 18712.

(3) Bailly, X.; Chabasse, C.; Hourdez, S.; Dewilde, S.; Martial, S.; Moens, L.; Zal, F. FEBS

Journal 2007, 274, 2641.

(4) Chen, Y. P.; Woodin, S. A.; Lincoln, D. E.; Lovell, C. R. J. Biol. Chem. 1996, 271,

4609.

(5) Barrios, D. A.; D’Antonio, J.; McCombs, N. L.; Zhao, J.; Franzen, S.; Schmidt, A. C.;

Sombers, L. A.; Ghiladi, R. A. J. Am. Chem. Soc. 2014, 136, 7914.

(6) D’Antonio, J.; Ghiladi, R. A. Biochemistry 2011, 50, 5999.

(7) Zhao, J.; de Serrano, V.; Zhao, J. J.; Le, P.; Franzen, S. Biochemistry 2013, 52, 2427.

(8) Wang, C.; Lovelace, L. L.; Sun, S.; Dawson, J. H.; Lebioda, L. Biochemistry 2013, 52,

6203.

(9) Thompson, M. K.; Davis, M. F.; de Serrano, V.; Nicoletti, F. P.; Howes, B. D.;

Smulevich, G.; Franzen, S. Biophys. J. 2010, 99, 1586.

(10) Edelstein, S. J. Annu. Rev. Biochem. 1975, 44, 209.

(11) Benesch, R.; Benesch, R. E. Biochem. Biophys. Res. Commun. 1967, 26, 162.

(12) Lebioda, L.; LaCount, M. W.; Zhang, E.; Chen, Y. P.; Han, K.; WHitton, M. M.;

Lincoln, D. E.; Woodin, S. A. Nature 1999, 401, 445.

159

(13) Zhao, J.; Moretto, J.; Le, P.; Franzen, S. J. Phys. Chem. B 2015, 119, 2827.

(14) Zhao, J.; Franzen, S. J. Phys. Chem. B 2013, 117, 8301.

(15) Zhao, J.; Zhao, J.; Franzen, S. J. Phys. Chem. B 2013, 117, 14615.

(16) Davis, M. F.; Gracz, H.; Vendeix, F. A. P.; de Serrano, V.; Somasundaram, A.; Decatur,

S. M.; Franzen, S. Biochemistry 2009, 48, 2164.

(17) Davis, M. F.; Bobay, B. G.; Franzen, S. Biochemistry 2010, 49, 1199.

(18) Nienhaus, K.; Deng, P.; Belyea, J.; Franzen, S.; Nienhaus, G. U. J. Phys. Chem. B 2006,

110, 13264.

(19) Nicoletti, F. P.; Thompson, M. K.; Howes, B. D.; Franzen, S.; Smulevich, G.

Biochemistry 2010, 49, 1903.

(20) Ma, H.; Thompson, M. K.; Gaff, J.; Franzen, S. J. Phys. Chem. B 2010, 114, 13823.

(21) Zhao, J.; Lu, C.; Franzen, S. J. Phys. Chem. B 2015, 119, 12828.

(22) Zhang, Z.; Santos, A. P.; Zhou, Q.; Liang, L.; Wang, Q.; Wu, T.; Franzen, S. Biophys.

Chem. 2016, 211, 28.

(23) de Serrano, V. S.; Davis, M. F.; Gaff, J. F.; Zhang, Q.; Chen, Z.; D'Antonio, E. L.;

Bowden, E. F.; Rose, R.; Franzen, S. Acta Cryst. D-Biol. Cryst. 2010, 66, 770.

(24) de Serrano, V.; D'Antonio, J.; Franzen, S.; Ghiladi, R. A. Acta Cryst. D-Biol. Cryst.

2010, 66, 529.

(25) Chen, Z.; De Serrano, V.; Betts, L.; Franzen, S. Acta Crystal. D-Biol. Cryst. 2009, 65,

34.

(26) de Serrano, V.; Chen, Z. X.; Davis, M. F.; Franzen, S. Acta Crystal. D-Biol. Cryst.

2007, 63, 1094.

160

(27) De Serrano, V.; Franzen, S. Peptide Sci. 2012, 98, 27.

(28) Zhao, J.; de Serrano, V.; Franzen, S. Biochemistry 2014, 53, 2474.

(29) Palmer, A. G. Chem. Rev. 2004, 104, 3623.

(30) Jarymowycz, V. A.; Stone, M. J. Chem. Rev. 2006, 106, 1624.

(31) d’Auvergne, E.; Gooley, P. J. Biomol. NMR 2008, 40, 107.

(32) d’Auvergne, E.; Gooley, P. J. Biomol. NMR 2008, 40, 121.

(33) d'Auvergne, E.; Gooley, P. J. Biomol. NMR 2003, 25, 25.

(34) Nelson, M. T.; Humphrey, W.; Gursoy, A.; Dalke, A.; Kale, L. V.; Skeel, R. D.;

Schulten, K. Int. J. Supercomput. Ap. 1996, 10, 251.

(35) Phillips, J. C.; Braun, R.; Wang, W.; Gumbart, J.; Tajkhorshid, E.; Villa, E.; Chipot,

C.; Skeel, R. D.; Kale, L.; Schulten, K. J. Comp. Chem. 2005, 26, 1781.

(36) Ryckaert, J.-P.; Ciccotti, G.; Berendsen, H. J. C. J. Chem. Phys. 1977, 23, 327.

(37) Glykos, N. M. J. Comput. Chem. 2006, 27, 1765.

(38) Chovancova, E.; Pavelka, A.; Benes, P.; Strnad, O.; Brezovsky, J.; Kozlikova, B.; Gora,

A.; Sustr, V.; Klvana, M.; Medek, P.; Biedermannova, L.; Sochor, J.; Damborsky, J. PLoS

Comput. Biol. 2012, 8, e1002708.

(39) Schrodinger, LLC 2015.

(40) Thompson, M. K.; Franzen, S.; Davis, M. F.; Oliver, R. C.; Krueger, J. K. J. Phys.

Chem. B 2011, 115, 4266.

(41) Peng, J. W.; Wagner, G. J. Magn. Reson. 1992, 98, 308.

(42) Farrow, N. A.; Zhang, O.; Szabo, A.; Torchia, D. A.; Kay, L. E. J Biomol NMR 1995,

6, 153.

161

(43) Lefevre, J.-F.; Dayie, K. T.; Peng, J. W.; Wagner, G. Biochemistry 1996, 35, 2674.

(44) Morin, S.; M. Gagné, S. J. Biomol. NMR 2009, 45, 361.

(45) Lipari, G.; Szabo, A. J. Am. Chem. Soc. 1982, 104, 4546.

(46) Lipari, G.; Szabo, A. J. Am. Chem. Soc. 1982, 104, 4559.

(47) Andrec, M.; Montelione, G. T.; Levy, R. M. J Biomol NMR, 18, 83.

(48) Garcı́a de la Torre, J.; Huertas, M. L.; Carrasco, B. J. Magn. Reson. 2000, 147, 138.

(49) Chen, J.; Brooks, C. L.; Wright, P. E. J Biomol NMR, 29, 243.

(50) Clore, G. M.; Szabo, A.; Bax, A.; Kay, L. E.; Driscoll, P. C.; Gronenborn, A. M. J. Am.

Chem. Soc. 1990, 112, 4989.

(51) Han, K., Woodin, S.A., Lincoln, D.E., Fielman, K.T., Ely B. Mar. Biotechnol. 2001, 3,

287.

(52) Gelin, B. R.; Karplus, M. Proc. Natl. Acad. Sci. 1977, 74, 801.

(53) Baldwin, J.; Chothia, C. J. Mol. Biol. 1979, 129, 175.

(54) Perutz, M. F. Annu. Rev. Physiol. 1990, 52, 1.

(55) Brunori, M. FEBS Journal 2014, 281, 633.

(56) Yuan, Y.; Tam, M. F.; Simplaceanu, V.; Ho, C. Chem. Rev. 2015, 115, 1702.

(57) Royer, W. E.; Fox, R. A.; Smith, F. R.; Zhu, D.; Braswell, E. H. J. Biol. Chem. 1997,

272, 5689.

(58) Royer Jr, W. E.; Knapp, J. E.; Strand, K.; Heaslet, H. A. Trends Biochem. Sci. 2001,

26, 297.

(59) Volkman, B. F.; Alam, S. L.; Satterlee, J. D.; Markley, J. L. Biochemistry 1998, 37,

10906.

162

(60) Laine, J. M.; Amat, M.; Morgan, B. R.; Royer, W. E.; Massi, F. Biochemistry 2014, 53,

7199.

(61) Song, X.-j.; Yuan, Y.; Simplaceanu, V.; Sahu, S. C.; Ho, N. T.; Ho, C. Biochemistry

2007, 46, 6795.

(62) Savard, P.-Y.; Daigle, R.; Morin, S.; Sebilo, A.; Meindre, F.; Lagüe, P.; Guertin, M.;

Gagné, S. M. Biochemistry 2011, 50, 11121.

(63) Pond, M. P.; Majumdar, A.; Lecomte, J. T. J. Biochemistry 2012, 51, 5733.

(64) Yang, F.; Phillips Jr, G. N. J. Mol. Biol. 1996, 256, 762.

(65) D'Antonio, E. L.; Chen, T. K.; Turner, A. H.; Santiago-Capeles, L.; Bowden, E. F.

Electrochem. Commun. 2013, 26, 67.

(66) Brantley, R. E.; Smerdon, S. J.; Wilkinson, A. J.; Singleton, E. W.; Olson, J. S. J. Biol.

Chem. 1993, 268, 6995.

(67) Tsuruga, M.; Matsuoka, A.; Hachimori, A.; Sugawara, Y.; Shikama, K. J. Biol. Chem.

1998, 273, 8607.

(68) Shikama, K. Chem. Rev. 1998, 98, 1357.

(69) Tsai, A.-L.; Berka, V.; Martin, E.; Olson, J. S. Biochemistry 2012, 51, 172.

163

CHAPTER 7

The Mechanism of Autoreduction of Hemoglobin

– A Case Study of Dehaloperoxidase-Hemoglobin

164

7.1 Abstract

Hemoglobin and myoglobin are known to undergo autoxidation, which means that the oxyferrous form of the heme is oxidized to the ferric state without any external oxidant aside from O2. But dehaloperoxidase-hemoglobin (DHP) from Amphitrite ornata is observed to undergo the reverse process, which means that ferric heme is gradually reduced to oxyferrous form under aerobic conditions. The high reduction potential of DHP (+221 mV) partially explains this abnormal scenario, but the reduction source inside the protein has remained obscure. Given the fact that several amino acid residues like cysteine, methionine, tyrosine and tryptophan has the redox properties in protein and DHP has a sulfur rich protein sequence. We postulate that sulfur containing residues (cysteine and methionine) are the major reduction source that give rise to autoreduction in DHP. To investigate the role of the sulfur-containing residues, we created six mutants (C73S, C73S&M49C, S78C, M63L, M64L and

M63L&M64L) by site-directed mutagenesis and conducted a series of CO-driven autoreduction kinetic measurements. The autoreduction kinetics showed sigmoidal behavior which implies cooperativity during the reaction. DHP’s monomer-dimer equilibrium in solution may play a critical role in the autoreduction process.

165

7.2 Introduction

Dehaloperoxidase-hemoglobin (DHP) isolated from Amphitrite ornata is a multifunctional protein.1,2 DHP has very high redox potential (+ 221 mV) that may relate to the functional switch by changing the oxidation state of the heme.3,4 DHP contains eight α- helices and shares a similar 3-over-3 helical bundle folding structures compare to that of mammalian myoglobin and hemoglobin  ,  unit even though the sequence homology is only about 20%.5,6 DHP has only one cysteine residues (Cys73) in its peptide sequence, which resembles the scenario of human myoglobin (Cys110). Most mammalian myoglobin do not contain cysteine.7 Moreover, DHP is a sulfur rich hemoglobin that contains six methionine and one cysteine residues in the peptide sequence. The only cysteine Cys73 is located on the surface of the protein at the central region of dimer interface observed in every X-ray crystal structure.8,9 However, the six methionine residues are located at internal sites distributed over both helical or loop regions. (Figure 6.1). The protein NMR experiments showed that 5 out 7 of the sulfur-containing residues (Met49, Met69, Met108, Met136 and Cys73) in DHP have shown significant chemical exchange on their backbone amide N-H bonds. Chemical exchange on these residues is relevant to conformational change or interaction with other molecules on the microsecond to millisecond time scale. These structural and dynamic observations suggest that

DHP may have different redox stabilization properties than other globins. Indeed, freeze quench EPR studies show that DHP has a unique set of dynamic reduction reactions in one or more of its enzymatic functions. The transient role played by tyrosines in the reduction of the heme may also be a consideration in the global understanding of the redox properties of

DHP.10,11

166

One of the crucial aspects of a multifunctional globin is the regulation of redox state of the heme Fe atom. It is important to understand how the protein could return to the ferrous state once oxidized intermediates have been formed. In this context it is relevant that DHP has been observed to gradually reduce from the ferric state (Fe3+) to the ferrous state (Fe2+) under aerobic conditions without any external reductant. This phenomenon is quite unexpected at first because hemoglobins and myoglobins are known to undergo an autoxidation process when exposed to oxygen. This naturally occurring process is sufficiently rapid that oxyferrous hemoglobin and myoglobin will be oxidized to the ferric state within 6 hours under aerobic conditions in solution.12,13 For this reason red blood cells contain a complex known as methemoglobin reductase to reduce the heme Fe using an external source of reducing electrons.

Autooxidation is ubiquitious heme proteins.14 Because of its high reduction potential DHP has the reverse tendency under aerobic conditions; it tends to spontaneously convert from the ferric to the ferrous (or oxyferrous) form. By analogy with the well-studied autooxidation process, we call this process autoreduction. DHP has a high reduction potential of +221 mV for the heme (Fe3+ / Fe2+) couple as measured by cyclic volammetry.4 This value can be seen to be much greater than that of other hemoglobins and myoglobins (Table 6.3). Consequently, the free energy of the ferrous state of DHP is thermodynamically favored over the ferric state.

However, this aspect of the autoreduction process only explains one side of the story. We must also address the question: what is the reductant that gives rise to autoreduction? We assume that it is within the protein and therefore the electron comes from one of the redox active amino acids discussed previously. Here we see a possible connection between the relatively large number of redox active amino acids and an important aspect of function, namely the

167

maintenance of the ferrous state of DHP in the long term while permitting enzymatic activity by various oxidized forms of the protein in the short term.

Given the antioxidant properties of the sulfur-containing amino acid residues, it is a plausible to postulate that the cysteine and methionines are the major source of reducing power inside the protein. Hirota et al showed that one cysteine introduced to SWMb will accelerate the reduction in the presence of CO by forming a crosslinked dimer between two cysteine residues on SWMb mutants.15 In this mechanism, by forming disulfide bond between two cysteines, two electrons are released to reduce the heme. However, the other mechanism for autoreduction could be electron transfer from one or more of the methionines to the heme Fe.

This would maintain the heme Fe in the ferrous state at the expense of oxidizing the methionines to sulfoxides. Just as has been observed for tyrosines the multiplicity of methionines suggests that more than one of them may be oxidized during the life span of DHP.

However, unlike tyrosine oxidation, methionine oxidation is irreversible. Thus, in the end DHP would become inactive according to this mechanism.

The autoreduction behavior of DHP may be an essential pathway for functional switching between peroxidase or peroxygenase that primarily using ferric state as an catalytic starting point and oxygen transporter or gas sensor that using ferrous heme.2,16,17 To investigate the autoreduction mechanism of hemoglobin, we have designed a series of experiments to measure the kinetics of autoreduction reaction. To study the role of sulfur containing residues in DHP, several cysteine or methionine mutants have been made in DHP. We found that autoreduction reaction of DHP follows a sigmoidal kinetics which implies cooperative interaction between proteins.

168

Figure 7.1 DHPA monomer – dimer equilibrium with sulfuring containing residues labeled on the secondary structure.

7.3 Material and Methods

Materials

All Chemicals were purchase from Sigma-Aldrich and Fisher Scientific and used without further purification. The compressed CO gas tank was from Machine and Welding Supply

Company contains 10% CO balanced with Argon. The E.Z.N.A plasmid DNA mini kit was from Omega bio-tek. The QuickChange II site-directed mutagenesis kit was purchased from

Agilent Technologies Inc. The designed oligonucleotide primers were synthesized by IDT

DNA Technology Inc. Horse heart myoglobin (HHMb) was obtained from Sigma-Aldrich.

Site-Directed Mutagenesis and Protein Purification

Six mutations were generated with the Quickchange II site-directed mutagenesis kit.

Mutagenesis [melt (95 °C, 60 s), anneal (55 °C, 30 s), and extension (68 °C, 6min)] was performed for 18 cycles. The plasmid encoding wild-type(WT) DHPA (6XHis-tag) was used as a template to generate mutated plasmid using the designed primers (Table 7.1). The

M63L&M64L double mutant was made by Genewiz Inc. The mutated plasmids were extracted using a E.Z.N.A plasmid DNA mini kit after the transformation of the reaction mixture into the BL21(DE3) E. coli. The plasmids were then sent for sequencing at Genewiz Inc to confirm

169

that the desired mutation was made. WT DHPA and all mutants were expressed and purified as previously described.18 Special care was taken for S78C mutant because the protein starts to aggregate after the oxidation by potassium ferricyanide K3[Fe(CN)6] during the purification.

Except for H55V mutant, which has that its ferric state Soret band at 394 nm with ε394nm =

121,300 M-1cm-1, all other mutants and WT DHPA have Soret band at 406 nm with extinction

-1 -1 coefficient ε407nm = 116,400 M cm . HHMb has Soret band at 409 nm with extinction

-1 -1 coefficient ε409nm = 188,000 M cm .

Table 7.1 DNA Sequences of Primers

Primer Sequence C73S forward GACCGAGCCACCGATTCTGTCCCCCTTGCGTCC C73S reverse GGACGCAAGGGGGACAGAATCGGTGGCTCGGTC M49C forward CAAGAGCTCAAGTCATGCGCCAAGTTCGGTGAT M49C reverse ATCACCGAACTTGGCGCATGACTTGAGCTCTTG S78C forward GTCCCCCTTGCGTGCGACGCCAACACACTC S78C reverse GAGTGTGTTGGCGTCGCACGCAAGGGGGAC M63L forward AAAGTGTTCAACCTGCTGATGGAAGTTGCGGAC M63L reverse GTCCGCAACTTCCATCAGCAGGTTGAACACTTT M64L forward GTGTTCAACCTGATGCTGGAAGTTGCGGACCGA M64L reverse TCGGTCCGCAACTTCCAGCATCAGGTTGAACAC

Plasmid encoded recombinant sperm whale myoglobin (SWMb) was obtained from Addgene and transformed in the BL21(DE3) E. coli. The recombinant SWMb was purified according to the reported protocol using an ammonium sulfate precipitation method as the first step.19 Then

SWMb was completed oxidized by potassium ferricyanide followed by elution through a NAP-

25 size exclusion column (GE Healthcare) to remove ferricyanide and ferrocyanide. The oxidized ferric SWMb was further purified by running through the CM Sepharose Fast Flow cation exchange column (GE Healthcare). The UV-Vis spectra of purified SWMb was identical

170

to that in the literature. The protein concentration was determined by monitoring the Soret band

-1 -1 at 409 nm with extinction coefficient ε409nm = 171,000 M cm .

Autoreduction Kinetics Measurement

An aliquot of ~1 mL protein in 100 mM potassium phosphate (KPi) buffer at pH 7.0 was added with 200 L 5mM tris(2-carboxyethyl) phosphine (TCEP) solution followed by incubation for

1h to reduce potential disulfide crosslinks in DHP. It is worthy to notice that given the high redox potential of ferric DHP A (+221 mV), the ferric protein was also reduced by TCEP to the oxyferrous state in the aerobic condition. This is noteworthy since TCEP does not reduce most other heme proteins. To remove remaining TCEP in the solution, the protein aliquot was first run through the NAP-25 column, and then the protein fraction was completely oxidized to ferric state by potassium ferricyanide and eluted through a NAP-25 column to remove excess ferricyanide and ferrocynaide ion. A 1200 L 50 M protein solution aliquot was made and transferred to a conical vial sealed by septum cap. The protein solution was exchanged and filled with N2 gas for 5 min, then exchanged and filled with 10% CO gas for another 5 min.

The solution was left equilibrated with CO gas for another 30 min before transferring ~600 L into a 1mm pathlength quartz cuvette sealed with septum cap. The cuvette was fixed by a special holder and then placed in the Agilent 8453 diode array UV−visible spectrophotometer equipped with a Peltier-cooled sample cell that set a constant temperature at 37 °C. The UV-

Vis spectrometer is then set to run for 10 – 60 h with a scan every 10 min. The time course of the autoreduction reaction was extracted from the time-resolved UV- Vis spectra collected between 300 – 700 nm wavelengths using the SVD (singular value decomposition) method.

171

Autoreduction Kinetic Model

The protein solution is saturated with 10% CO gas under 1 atm at room temperature. The

Henry’s law constant for CO at 298 K is 9.71 • 10-4 M atm-1. Therefore, the CO concentration is estimated to be circa 100 M in the solution, which is in excess compared to protein concentration. Because CO diffusion and binding to the protein (k ~ 105 – 106 M-1 s-1) is much faster than the overall reduction rate the rate limiting step is the electron transfer process that reduces the ferric heme to the ferrous state. According to the proposed autoreduction mechanism shown in Figure 6.1, there are two types of autoreduction mechanism: the monomer autoreduction mechanism (km) and a further subdivision of two possible dimer autoreduction mechanisms. The dimer autoreduction mechanism is divided as ferric-ferric dimer autoreduction mechanism (kd1) and ferric-ferrous dimer autoreduction mechanism (kd2).

We hypothesized that the ferric-ferrous dimer has the faster autoreduction rate due to the cooperative reduction of the heme in the dimer, where the key assumption is that the ferric heme is reduced more rapidly when the other heme in the dimer has already been reduced to the ferrous state. Given that the apparent autoreduction rate kobs has an inverse dependence on concentration (Figure 6.2a), the concentration terms in the second order rate expression that corresponding to the dimer autoreduction mechanism are all normalized by dividing the total protein concentration [A]0 as shown in Equation 1.

172

Figure 7.2 (a) Monomer autoreduction mechanism; (b) Cooperative dimerized autoreduction mechanism.

The overall autoreduction rate equation is :

푑[Ferrous − CO] 푑푥 푥 [퐴] [퐴]2 = = 푘 [퐴] + 푘 + 푘 (1) 푚 푑1 [ ] [ ] 푑2 2 푑푡 푑푡 퐴 0 퐴 0 [퐴]0 x : Ferrous-CO protein concentration. [A] : Ferric protein concentration.

[A]0 : Total protein concentration, thus [A]0 = [A] + x. km : Rate constant of monomer autoreduction mechanism. kd1 : Rate constant of ferric-ferric dimer autoreduciton mechanism. kd2 : Rate constant of ferric-ferrous dimer autoreduction mechanism (cooperative).

The overall autoreduction rate equation is approximated by focusing on the fast, cooperative ferric-ferrous dimer autoreduction mechanism with an overall rate constant ka represented in equation 2.

173

푑푥 푥 [퐴] 푥 ([퐴]0 − 푥) ≈ 푘푎 = 푘푎 (2) 푑푡 [퐴]0 [퐴]0 [퐴]0 [퐴]0

The integrated form of this rate equation 2 is

푡 푥 2 ([퐴]0) ∫ 푘 푑푡 = ∫ 푑푥 푎 푥([퐴] − 푥) 0 푥0 0

At t = 0, x0 is supposed to be non-zero, because some reduced ferrous-CO protein has already produced by “slow” mechanisms. The integrated form of equation 2 gives equation 3 as shown below:

푡 푥 [퐴]0 [퐴]0 ∫ 푘 푑푡 = ∫ ( − ) 푑푥 푎 푥 푥 − [퐴] 0 푥0 0

푥 푥0 − [퐴]0 푘푎푡 = [퐴]0 [푙푛 − 푙푛 ] 푥0 푥 − [퐴]0

푥0[퐴]0 푥(푡) = 푘 푡 (3) − 푎 [퐴] 푥0 + ([퐴]0 − 푥0)푒 0

Therefore, the apparent autoreduction rate kobs is the reciprocal of the total protein concentration [A]0.

푘푎 푘표푏푠 = (4) [퐴]0

174

7.4 Results

1.2 a 422 nm Ferrous-CO 0.10 b 1.0 0.12 Ferric 541 nm 571 nm 0.05 e 0.10 c 407 nm 507 nm

n 0.8 0.08 a 0.00

b A r 0.06

o 0.6 Δ s 0.04

b -0.05 A 0.4 0.02 0.00 -0.10 0.2 Experimental Time Course 450 500 550 600 650 -0.15 Fitted Time Course 0.0 300 400 500 600 0 5 10 15 20 Wavelength (nm) Time (Hours)

0.20 1.2 c 424 nm Ferrous-CO Experimental Time Course 0.15 Fitted Time Course 1.0 Ferric 539 nm e 0.12 505 nm

c 394 nm 570 nm

n 0.10 0.10

a 0.8

b 0.08

A r

Δ 0.05 o 0.06

s 0.6

b 0.04

A 0.00 0.4 0.02 0.00 -0.05 0.2 450 500 550 600 650 d 0.0 -0.10 300 400 500 600 0 5 10 15 20 Time (Hours) Wavelength (nm) Figure 7.3 The time-resolved UV-Vis spectra of CO-driven autoreduction reaction of (a) WT DHPA and (c) DHPA H55V mutant and their corresponding time course (c) WT DHPA, (d) DHPA H55V of the reaction.

The time-resolved UV-Vis spectra of CO-driven autoreduction reaction of WT DHPA clearly shows the transition from the ferric state with a Soret band at 407 nm and a Q band at

507 nm to the ferrous-CO state with a soret band at 422 nm and a doublet Q band at 541 nm and 571 nm. The much sharper Soret band at 422 nm is due to the formation of 6 coordinate low spin (6cLS) ferrous-CO species in contrast to the original 5 coordinated high spin (5cHS) or water ligated 6 coordinated high spin (6cHS) ferric state. The same transition is also observed for all the other mutants, for example, the H55V mutants with a 394 nm Soret band and a 505 nm Q band at ferric state will also convert to ferrous-CO form with Soret band at

424 nm and a doublet Q band 539 nm and 570 nm. This is in contrast to the autooxidation

175

mechanism where distal histidine plays a critical role in autooxidation rate and removal of distal histidine lowers the autooxidation rate in three orders of magnitude.12

The time course of CO-driven autoreduction reaction reveal that the process is not a simple first-order reaction, but presents a sigmoidal behavior which the reaction initially accelerates and then gradually decelerates (Figure 7.3b, 7.3d). The sigmoidal kinetics implies cooperativity behavior of the protein during the autoreduction process. However, in contradistinction to the well-known cooperativity of hemoglobin, in which oxygen binding affinity increases when more oxygen are bound in the tetramer, the cooperativity observed in autoreduction kinetics suggest accelerated rate of reduction for the second ferric heme in the dimer when the first one is already reduced (Figure 7.2). Sigmoidal kinetics is also observed in amyloid fibrillation, a nucleation-dependent polymerization process.20,21 The kinetic model shown in Figure 7.2, described by Equation 3 characterizes such sigmodal behavior and showed a good fitting for WT DHPA experimental time course. The H55V mutant also showed a reasonable fit except at the beginning of the time course. We hypothesize that the removal of distal histidine His55 causes destabilization of CO binding, which disrupts the slow autoreduction mechanism but has negligible effect on the fast ferric-ferrous dimer autoreduction mechanism.

176

a 0.25 b 0.8

0.20

)

) 1

1 -

h 0.6

(

s b

b 0.15

o k 0.4 k 0.10 0.2 0.05 20 40 60 80 100 5.0 5.5 6.0 6.5 7.0 7.5 8.0 [DHP] (µM) pH

Figure 7.4 (a) The DHP protein concentration dependent of apparent autoreduction rate kobs. (b) The pH-dependent autoreduction rate kobs.

The apparent autoreduction rate kobs has the inverse dependence on protein concentration. High hemoglobin protein concentration may exert a synergeistic protective effect. This observation is in contrast to the concentration-dependent rate of SWMb reduction, where the reduction rate is increased when the concentration is increased. However, we should point out that the two cases focus on two different concentration ranges. In the case of DHP, we have studied higher concentration ranging from 15 M to 100 M, whereas in the case of

SWMb, Hirota et al studied from 2 M to 18 M.15 The autoreduction rate also showed a strong pH dependence. The optimum pH value for autoreduction is circa 7.0. Lower pH will suppress the autoreduction presumably due to the protonation of amino acid residues especially methionine and cysteine (pKa = 8.35), which makes the oxidation of these residues much more difficult. Higher pH above 7.0 also slows down the reduction rate due to stabilization of the ferric form ofthe heme Fe by the binding of a hydroxide ligand. This phenomenon can be observed by monitoring the shift of the Soret band from 414 nm to 422 nm as shown in Figure

F1.

177

Table 7.2 Apparent Autoreduction rate kobs of different proteins

-1 Protein kobs ( h ) DHPA WT 0.245 ± 0.021 C73S 0.290 ± 0.017 M49C/C73S 0.136 ± 0.029 S78C 0.349 ± 0.005 M63L 0.221 ± 0.002 M64L 0.205 ± 0.002 M63L/M64L 0.155 ± 0.011 H55V 0.223 ± 0.028 HHMb 0.086 ± 0.001 SWMb 0.098 ± 0.002

The kinetic experiment showed that the C73S mutant, which removes the only cysteine in DHP still has an autoreduction rate that is 20% higher than that of WT DHPA. When the cysteine position is swapped from 73 to 49 in the M49C/C73S double mutant, the autoreduction rate is 45% lower. However, adding a cysteine near the dimer interface will increase the autoreduction rate by 40%. Apparently, Cys73 is not the only source of reducing equivalents needed for autoreduction. This result is in contrast to the case of SWMb, where wild-type SWMb was introduced a cysteine that enhanced one order magnitude rate of autoreduction. To study the effects of other sulfur containing residues on autoreduction, we have focused on two methionine residues Met63 and Met64, which are the closest residues to the heme. The sulfur ester of the methionine sidechain is known act as a reducing agent while being oxidized itself to to a sulfoxide.22 The distances of the sulfur atoms for Met63 and Met64 to the heme vinyl group are 5.0 Å and 9.8 Å, respectively. The kinetic results suggest that removal of these methionines does reduce the rate of the autoreduction process. The autoreduction rate of the M63L/M64L double mutant is 40% slower than wild type. In

178

comparison, we have also measured the autoreduction rate of SWMb and HHMb. Both proteins have no cysteines in their wild type and have only 3 or 2 methionines in their peptide sequences. This is a control experiment to test the effect of the method used here to measure autoreduction. Using this method the myoglobins do show measurable autoreduction behavior, with about 1/3 of the rate compared to that of WT DHPA.

7.5 Discussion

Table 7.3 Reduction Potential of Hemoglobin and Myoglobin

Heme Protein # Cys # Met A.A. Length E°’ (mV) DHPA from Amphitrite Ornata* 1 6 137 2214 Glycera Dibranchiata Monomeric Hb 1 5 147 15323 Scapharca Dimeric Hb* 1 3 146 110 - 14024 Human Hb  unit 2 1 146 10225 Human Hb  unit 1 2 141 4025 Human Mb 1 2 154 597

Sperm Whale Mb 3 154 4326

Horse Heart Mb 2 153 2826 * Single chain

At first, the autoreduction phenomenon appeared to be a property that is unique to DHP.

However, when three different globins were tested (DHP, HHMb and SWMb) they all showed autoreduction behavior. Our hypothesis is that autoreduction is a common property among hemoglobins and myoglobins under reducing conditions. This hypothesis does not change the fact that DHP has significantly higher autoreduction rate and it is the only protein of the three

(and the only globin known) that undergoes autoreduction under aerobic conditions.

179

Autoreduction rate depends on two factors, the quantities of reduction sources and reduction potential of the heme. And these two factors address the kinetic and thermodynamic aspects of the autoreduction process, respectively. We propose that the sulfur containing residues account for the major reducing sources in hemoglobin and myoglobin. Table 7.3 have listed a few animal hemoglobin and myoglobin spanning from annelids, mollscus and mammalian with the number of cysteine and methionine residues in their protein sequence and the reduction potential of the heme. Given the data we collected that shown in Table 7.3, we found that the autoreduction rate has the positive correlation to the number of sulfur containing residues and the positive reduction potential value.

The autoreduction mechanism of hemoglobin and myoglobin provides a unique pathway to convert ferric state of the heme to ferrous state, so the protein returns to its function as an oxygen transporter or gas sensor without the requirement for an external reducing source, like a flavoprotein. Flavoproteins are required to provide electrons for reduction in cytochrome

P450.27,28 This reduction pathway is particularly important for DHP, because it is a critical part for DHP to conduct function switch in vivo. Unlike HHMb and SWMb CO is not necessary to drive the autoreduction in DHP. Autoreduction in DHP is also observed in anerobic (under N2) or even under O2 atmosphere, with just 1/3 rate compared to that under CO. Autoreduction switches DHP back to its classical hemoglobin’s function. Therefore, the regulation pathway that redox states switching of heme between ferric and ferrous are now complete.

180

7.6 References

(1) Chen, Y. P.; Woodin, S. A.; Lincoln, D. E.; Lovell, C. R. J. Biol. Chem. 1996, 271,

4609.

(2) Barrios, D. A.; D’Antonio, J.; McCombs, N. L.; Zhao, J.; Franzen, S.; Schmidt, A. C.;

Sombers, L. A.; Ghiladi, R. A. J. Am. Chem. Soc. 2014, 136, 7914.

(3) D’Antonio, E. L.; Bowden, E. F.; Franzen, S. Journal of Electroanalytical Chemistry

2012, 668, 37.

(4) D'Antonio, E. L.; Chen, T. K.; Turner, A. H.; Santiago-Capeles, L.; Bowden, E. F.

Electrochem. Commun. 2013, 26, 67.

(5) Franzen, S.; Belyea, J.; Gilvey, L. B.; Davis, M. F.; Chaudhary, C. E.; Sit, T. L.;

Lommel, S. A. Biochemistry 2006, 45, 9085.

(6) Bailly, X.; Chabasse, C.; Hourdez, S.; Dewilde, S.; Martial, S.; Moens, L.; Zal, F. FEBS

Journal 2007, 274, 2641.

(7) Varadarajan, R.; Zewert, T.; Gray, H.; Boxer, S. Science 1989, 243, 69.

(8) de Serrano, V.; Chen, Z. X.; Davis, M. F.; Franzen, S. Acta Crystal. D-Biol. Cryst.

2007, 63, 1094.

(9) Chen, Z.; de Serrano, V.; Betts, L.; Franzen, S. Acta Crystallographica Section D 2009,

65, 34.

(10) Thompson, M. K.; Franzen, S.; Ghiladi, R. A.; Reeder, B. J.; Svistunenko, D. A.

Journal of the American Chemical Society 2010, 132, 17501.

181

(11) Dumarieh, R.; D'Antonio, J.; Deliz-Liang, A.; Smirnova, T.; Svistunenko, D. A.;

Ghiladi, R. A. Journal of Biological Chemistry 2013, 288, 33470.

(12) Brantley, R. E.; Smerdon, S. J.; Wilkinson, A. J.; Singleton, E. W.; Olson, J. S. Journal of Biological Chemistry 1993, 268, 6995.

(13) Tsuruga, M.; Matsuoka, A.; Hachimori, A.; Sugawara, Y.; Shikama, K. Journal of

Biological Chemistry 1998, 273, 8607.

(14) Bando, S.; Takano, T.; Yubisui, T.; Shirabe, K.; Takeshita, M.; Nakagawa, A. Acta

Crystallographica Section D 2004, 60, 1929.

(15) Hirota, S.; Azuma, K.; Fukuba, M.; Kuroiwa, S.; Funasaki, N. Biochemistry 2005, 44,

10322.

(16) Zhao, J.; Lu, C.; Franzen, S. J. Phys. Chem. B 2015, 119, 12828.

(17) Tsai, A.-L.; Berka, V.; Martin, E.; Olson, J. S. Biochemistry 2012, 51, 172.

(18) Ma, H.; Thompson, M. K.; Gaff, J.; Franzen, S. J. Phys. Chem. B 2010, 114, 13823.

(19) Springer, B. A.; Sligar, S. G. Proceedings of the National Academy of Sciences 1987,

84, 8961.

(20) Tobacman, L. S.; Korn, E. D. Journal of Biological Chemistry 1983, 258, 3207.

(21) Kodaka, M. Biophysical Chemistry 2004, 107, 243.

(22) Shechter, Y. Journal of Biological Chemistry 1986, 261, 66.

(23) Addison, A. W.; Burman, S. Biochimica et Biophysica Acta (BBA) - Protein Structure and Molecular Enzymology 1985, 828, 362.

(24) Boffi, A.; Bonaventura, C.; Bonaventura, J.; Cashon, R.; Chiancone, E. Journal of

Biological Chemistry 1991, 266, 17898.

182

(25) Abraham, E. C.; Taylor, J. F. Journal of Biological Chemistry 1975, 250, 3929.

(26) Battistuzzi, G.; Bellei, M.; Casella, L.; Bortolotti, C. A.; Roncone, R.; Monzani, E.;

Sola, M. JBIC Journal of Biological Inorganic Chemistry 2007, 12, 951.

(27) Wang, M.; Roberts, D. L.; Paschke, R.; Shea, T. M.; Masters, B. S. S.; Kim, J.-J. P.

Proceedings of the National Academy of Sciences 1997, 94, 8411.

(28) Sevrioukova, I. F.; Li, H.; Zhang, H.; Peterson, J. A.; Poulos, T. L. Proceedings of the

National Academy of Sciences 1999, 96, 1863.

183

APPENDICES

184

Appendix A

Supporting Information for Chapter 2

A1. Derivation of the Apparent Binding Equilibrium Constant

Scheme A7. The equilibrium scheme of competitive binding between fluoride anion and internal binding substrate or inhibitor in WT DHPA.

Free ferric enzyme concentration is represented as [E] (upper left); Ferric enzyme bound with internal binding ligand (for example, 2,4-DBP) is [EL] (upper right); Ferric enzyme with fluoride anion coordinated is [EF] (lower left); Ferric enzyme with fluoride anion coordinated and bound internal ligand is [EFL] (lower right); Total enzyme concentration is [E]0 ; Free fluoride anion concentration [F] ; Total internal binding ligand concentration [L].

In the square competitive binding equilibrium shown in Scheme A1, three binding equilibria with defined dissociation constants describe the relative quantities of the four species, which are [E], [EL], [EF], [EFL] which have been defined above. Since the fluoride concentration

185

[F] ranges from 600 - 1500 fold greater than total enzyme concentration, [E]0, the [F] is considered as constant during the titration.

1) The binding equilibrium between ferric enzyme and internal binding ligand (for

example, 2,4-DBP) in the absence of fluoride anion, with dissociation constant KL :

[퐸][퐿] 퐾 = 퐿 [퐸퐿]

2) The fluoride anion binding equilibrium between free ferric enzyme and fluoride anion

f in the absence of internal binding ligand, with the dissociation constant Kd :

[퐸][퐹] 퐾푓 = 푑 [퐸퐹]

3) The fluoride anion binding equilibrium between ferric enzyme with bound internal

b binding ligand and fluoride anion, with the dissociation constant Kd :

[퐸퐿][퐹] 퐾푏 = 푑 [퐸퐹퐿]

According to the mass conservation of the enzyme species:

[퐸]0 = [퐸] + [퐸퐿] + [퐸퐹] + [퐸퐹퐿]

[퐸][퐿] [퐸][퐹] [퐸퐿][퐹] [퐸] = [퐸] + + + 0 퐾 푓 퐾푏 퐿 퐾푑 푑

[퐸][퐿] [퐸][퐹] [퐸][퐿][퐹] [퐸] = [퐸] + + + 0 퐾 푓 퐾 퐾푏 퐿 퐾푑 퐿 푑 Thus, [퐸] [퐸] = 0 [퐿] [퐹] [퐿][퐹] 1 + + 푓 + 푏 퐾퐿 퐾 퐾 퐾푑 퐿 푑

186

Assuming that the fluoride ion concentration is much greater than the enzyme

concentration [F] >> [E]0, the fluoride binding fraction θ that can be monitored using

the UV-Vis:

[퐸퐹] + [퐸퐹퐿] θ = [퐸]0 [퐸][퐹] [퐸][퐿][퐹] 푓 + 푏 퐾 퐾퐿퐾푑 θ = 푑 [퐿] [퐹] [퐿][퐹] (1 + + 푓 + 푏 ) [퐸] 퐾퐿 퐾 퐾 퐾푑 퐿 푑

1 [퐿] ( 푓 + 푏) [퐹] 퐾 퐾퐿퐾푑 θ = 푑 [퐿] 1 [퐿] 1 + + ( 푓 + 푏) [퐹] 퐾퐿 퐾 퐾 퐾푑 퐿 푑 [퐹] θ = [퐿] 1 + 퐾 퐿 + [퐹] 1 [퐿] + 푓 퐾 퐾푏 퐾푑 퐿 푑 [퐹] [퐹] θ = = 푎푝푝 푆푒푒 푡푒푥푡 (5) ([퐿] + 퐾 )퐾푓퐾푏 퐾 + [퐹] 퐿 푑 푑 + [퐹] 푑 푓 푏 [퐿]퐾푑 + 퐾퐿퐾푑

app Therefore, we can define the apparent fluoride dissociation constant Kd :

([퐿] + 퐾 )퐾푓퐾푏 퐾푎푝푝 = 퐿 푑 푑 푆푒푒 푡푒푥푡 (6) 푑 푓 푏 [퐿]퐾푑 + 퐾퐿퐾푑

187

Figure A1. Time-resolved UV-vis spectra of 10 µM DHP A catalyzed oxidation of 500 µM 2,4-DBP in the presence of 1200 µM H2O2 in at 100 mM KPi buffer at pH 7.0.

1.2 404 nm 4-BP 505 nm 3-BP 406 nm 2-BP 1.0 613 nm 402 nm 0.12 0.8 636 nm 0.08 0.6 0.04 0.4 Absorbance 0.00 0.2 500 550 600 650 700 0.0 300 400 500 600 700 Wavelength (nm) Figure A2. UV-vis spectra of 50 µM DHP A binding with 5 mM 2-BP (black),3-BP (blue) and 4-BP (red) respectively in 100 mM KPi buffer at pH 7.0, respectively.

188

1.2 505 nm 407 nm 605 nm 1.0 0.12 [2,4-DBP] 0.8 0.08 0.6 0.04 0.00 [Fluoride] 0.4 500 550 600 650 0.2 0.0

300 400 500 600

Figure A3. UV-vis spectra of fluoride titration of WT DHPA (50 µM) in the presence of 1 mM 2,4-DBP in the 100 mM KPi buffer, pH 7.0.

Table A9 Heat of Fluoride Solution Ho Fluoride (kcal/mol) LiF 1.13 NaF 0.218 KF -4.238 CsF -8.81 HF -14.7

Source: CRC Handbook of Tables for Applied Engineering Science By Ray E. Bolz, CRC Press

189

Appendix B

Supporting Information for Chapter 3

B1. Derivation of the initial rate equations.

Scheme B1. Proposed DHP peroxidase Bi-substrate Ping-Pong mechanism.

We can define the comprehensive rate constant for each reversible step:

′ 푘1[퐻2푂2]푘2 푘1 = 푘−1 + 푘2

′ 푘3[퐴퐻]푘4 푘3 = 푘−3 + 푘4

′ 푘5[퐴퐻]푘6 푘5 = 푘−5 + 푘6

190

Scheme B2. Reformatted Proposed DHP peroxidase Bi-substrate Ping-Pong mechanism. According to the mass conservation:

[Ferric DHP] + [Compound 0] + [Compound ES] + [Compound ES – AH2] + [Compound II] + [Compound II – AH2] = [퐸]0

According to the steady-state assumption:

[Ferric DHP] [Compound 0] [Compound ES] [Compound ES – AH ] = = = 2 1 1 1 1 ′ 푘 ′ 푘 푘1 2 푘3 4 [Compound II] [Compound II – AH ] [퐸] = = 2 = 0 1 1 1 1 1 1 1 1 + + + + + ′ 푘 ′ 푘 ′ 푘 ′ 푘 푘5 6 푘1 2 푘3 4 푘5 6

Or ′ ′ 푘1[Ferric DHP] = 푘2[Compound 0] = 푘3[Compound ES] = 푘4[Compound ES – AH2] ′ = 푘5[Compound II] = 푘6[Compound II – AH2] Thus

푑[퐴∙] = 푘 [Compound ES – AH ] + 푘 [Compound II – AH ] = 2푘′ [Ferric DHP] 푑푡 4 2 6 2 1

191

1 [퐸] ∙ 푑[푃] 1 푑[퐴∙] 0 푘′ 푣 = = = 푘′ [퐹푒푟푟𝑖푐 퐷퐻푃] = 푘′ ∙ 1 0 푑푡 2 푑푡 1 1 1 1 1 1 1 1 ′ + + ′ + + ′ + 푘1 푘2 푘3 푘4 푘5 푘6

[퐸] [퐸] = 0 = 0 1 1 1 1 1 1 푘−1 + 푘2 1 푘−3 + 푘4 1 푘−5 + 푘6 1 ′ + + ′ + + ′ + + + + + + 푘1 푘2 푘3 푘4 푘5 푘6 푘1[퐻2푂2]푘2 푘2 푘3[퐴퐻2]푘4 푘4 푘5[퐴퐻2]푘6 푘6

푘 푘 푘 푘 푘 푘 [퐸] [퐻 푂 ][퐴퐻 ] = 1 2 3 4 5 6 0 2 2 2 (푘−1 + 푘2)푘3푘4푘5푘6[퐴퐻2] + [(푘−3 + 푘4)푘1푘2푘5푘6 + (푘−5 + 푘6)푘1푘2푘3푘4][퐻2푂2] + 푘1푘3푘5(푘4푘6 + 푘2푘4+푘2푘6)

푘 푘 푘 2 4 6 [퐸] [퐻 푂 ][ ] 푘 푘 + 푘 푘 +푘 푘 0 2 2 퐴퐻2 = 4 6 2 4 2 6 (푘−1 + 푘2)푘4푘6 (푘−3 + 푘4)푘1푘2푘5푘6 + (푘−5 + 푘6)푘1푘2푘3푘4 [퐴퐻2] + [퐻2푂2] + [퐻2푂2][퐴퐻2] (푘4푘6 + 푘2푘4+푘2푘6)푘1 푘3푘5(푘4푘6 + 푘2푘4+푘2푘6)

푘푐푎푡[퐸]0[퐻2푂2][퐴퐻2] = 퐻2푂2 퐴퐻2 퐾푚 [퐴퐻2] + 퐾푚 [퐻2푂2] + [퐻2푂2][퐴퐻2]

푘푐푎푡[퐸]0[퐻2푂2][퐴퐻2] 푉 = (B1) 0 퐻2푂2 퐴퐻2 퐾푚 [퐴퐻2] + 퐾푚 [퐻2푂2] + [퐻2푂2][퐴퐻2]

In which

푘2푘4푘6 푘푐푎푡 = (B2) 푘4푘6 + 푘2푘4+푘2푘6

(푘 + 푘 )푘 푘 H2O2 −1 2 4 6 퐾푚 = (B3) (푘4푘6 + 푘2푘4+푘2푘6)푘1

(푘 + 푘 )푘 푘 푘 푘 + (푘 + 푘 )푘 푘 푘 푘 AH2 −3 4 1 2 5 6 −5 6 1 2 3 4 퐾푚 = (B4) 푘3푘5(푘4푘6 + 푘2푘4+푘2푘6)

As for substrate inhibition:

192

Scheme B8. DHP peroxidase Bi-substrate Ping-Pong mechanism includes substrate AH2 inhibition

[Ferric DHP] ∙ [AH ] 2 퐴퐻2 = 퐾푆퐼 [Ferric DHP − AH2]

[Ferric DHP − AH2] [AH2] = [Ferric DHP] 퐴퐻2 퐾푆퐼

[퐸]0 푉0 = 1 [AH2] 1 1 1 1 1 ′ ∙ (1 + 퐴퐻 ) + + ′ + + ′ + 푘 2 푘2 푘 푘4 푘 푘6 1 퐾푆퐼 3 5

푘푐푎푡[퐸]0[퐻2푂2][퐴퐻2] 푉0 = (B5) [퐴퐻 ] 퐾퐻2푂2 (1 + 2 ) [퐴퐻 ] + 퐾퐴퐻2[퐻 푂 ] + [퐻 푂 ][퐴퐻 ] 푚 퐴퐻2 2 푚 2 2 2 2 2 퐾푆퐼

193

푘푐푎푡[퐸]0[퐻2푂2] 퐾AH2 푎푝푝 1 + 푚 푉 [퐻2푂2] [퐴퐻 ] 푉 = 푚 = 2 (B6) 0 푎푝푝 [퐴퐻 ] 퐾푚 + [퐻2푂2] 1 + 2 퐴퐻2 퐻2푂2 퐾푆퐼 퐾푚 ∙ + [퐻2푂2] 퐾퐴퐻2 1 + 푚 [퐴퐻2]

푘푐푎푡[퐸]0[퐴퐻2] 퐻 푂 퐾 2 2 [퐴퐻 ] 1 + 푚 + 2 푎푝푝 [퐻 푂 ] 퐴퐻2 푉 [퐴퐻2] 2 2 퐾 [퐻2푂2] 푉 = 푚 = 푆퐼 (B7) 0 푎푝푝 퐴퐻2 퐾 + [퐴퐻2] 퐾 푚 푚 + [퐴퐻 ] 퐻2푂2 2 퐾푚 [퐴퐻2] 1 + + 퐴퐻 [퐻2푂2] 2 퐾푆퐼 [퐻2푂2]

B2. Kinetics of DHP B for substrate 2,4,6-TCP and 2,4,6-TBP.

Figure B9. Kinetics of DHP B (2.4 M) catalyzed oxidation of 2,4,6-TCP or 2,4,6-TBP in the presence of H2O2. (a) 2,4,6- TCP dimension, (b) H2O2 dimension (with 2,4,6-TCP), (c) 2,4,6-TBP dimension, (d) H2O2 dimension (with 2,4,6-TBP) in the 100 mM KPi buffer at pH 7.0.

194

B3. Kinetics and mechanism of DHP A and B catalyzed oxidation of ABTS.

Figure B10. Kinetics of DHP (2.4 M) catalyzed oxidation of ABTS in the presence of H2O2. (a) DHPA in ABTS dimension, (b) DHPA in H2O2 dimension, (c) DHPB in ABTS dimension, (d)DHPB in H2O2 dimension in the 100 mM KPi buffer at pH 7.0.

Scheme B9. Two steps oxidation mechanism for ABTS.

195

B4. Kinetics of HRP for substrate 2,4,6-TCP and 2,4,6-TBP.

Figure B3. Kinetics of HRPC (0.2 M) catalyzed oxidation of 2,4,6-TCP or 2,4,6-TBP in the presence of H2O2. (a) 2,4,6- TCP dimension, (b) H2O2 dimension (with 2,4,6-TCP), (c) 2,4,6-TBP dimension, (d) H2O2 dimension (with 2,4,6-TBP) in the 100 mM KPi buffer at pH 7.0.

Figure B4. 3D plot of kinetics of HRPC (0.2 M) catalyzed oxidation of (a) 2,4,6-TCP or (b) 2,4,6-TBP in the presence of H2O2.

196

1.6 1.4 1.2 2,4,6- TBP 1.0 0.8

0.6 Absorbance 0.4 HRP Denaturing 0.2 0.0 300 400 500 600 700 Wavelength (nm)

Figure B5. The Denaturation of HRP (0.2 µM) during the catalytic reaction in the presence of 900 µM 2,4,6-TBP and 300 µM H2O2.

197

Appendix C

Supporting Information for Chapter 4

C1. Singluar Value Decomposition (SVD)

The essence of singualr value decomposition is to lower the rank of the data covariance matrix. The time-resolved spectra A(λ,t) obtained from stopped-flow spectrophotmer is a

653×900 two-dimensional data matrix, where each column is a absorbance spectrum measured at 653 wavelengths (255.7 nm to 799.5 nm) at a specific time point. 900 spectra were collected over three-time domain regimes (2.5, 25, and 250 ms; 300 scans each, 83.25 s total). First, the data matrix A(λ,t) is applied to SVD which allows A(λ,t) to be written as a product of three matrices:

퐴(휆, 푡) = 푈푆푉푇 (C1) where U (653×653) is a matrix containing the orthonormal basis spectra (Figure C1), S

(635×900) is a diagonal matrix containing the singular values and VT (900×900) (Figure C2) is a matrix containing the evolutionary time-course of the spectra. The singular value in the S matrix is a kind of weighting factors that can be used to determine the number of species involved in the reaction. In our case, first three singualr values and corresponding orthornormal basis spectra, evolutionary time-courses were picked up for reconstruction of the calculated spectra corresponding to the reaction species involved in the reaction.

푻 푨풓(흀, 풕) = 푼풓푺풓푽풓 (C2)

198

The lower rank data matrix Ar(λ,t) can also be described by the evolution of the first-order interconversion between each species. In our case, there are only three species involved. ki corresponds to apparent rate constant, and bi(λ) is the basis spectra corresponding to each exponential decay, which is also referred as the b-spectra.

퐴푟(휆, 푡) = 퐵푇 = ∑ 푏𝑖 (휆)푒푥 푝(−푘𝑖푡) (C3) 𝑖=0

According to the equation (B.2) and (B.3), we have:

푇 퐴푟(휆, 푡) = 퐵푇 = 푈푟푆푟푉푟 = 푈푟푆푟퐶푇 (C4)

푇 where 푉푟 = 퐶푇, 퐵 = 푈푟푆푟퐶

T Therefore, instead of fitting data matrix Ar(λ,t) to BT, the Vr matrix is fitted to CT. Thus, the amount of calculation for the non-linear regression has been significantly reduced.

C2. Global-fitting analysis and reconstruction of the spectra

Scheme C10. Kinetic model of reaction between ferric DHP with H2O2 in the absence of 4-BP for the global-fitting analysis of the stopped-flow data

T Each row of Vr matrix was fitted to the biexponential function according to the proposed two-

T step three species casade first order reaction model, and these three vectors of Vr matrix are applied to global-fitting to determine the apparent rate constant k1 and kRH.

푇 −푘1푡 −푘푅퐻푡 푣𝑖 (푡) = 푐𝑖0 + 푐𝑖1푒푥푝 + 푐𝑖2푒푥푝 (C5)

199

2 푇 푇 푉 = ∑ 푣𝑖 (푡) (C6) 𝑖=0

Therefore, the 3×3 C matrix has been determined and subsequently used to calculate the b- spectra (Figure C3) according to the Equation C8:

푐00 푐01 푐02 퐶 = |푐10 푐11 푐12| (C7) 푐20 푐21 푐22

퐵 = ∑ 푏𝑖 (휆) = 푈푟푆푟퐶 (C8) 𝑖=0

Accoring to the proposed kinetic model, the analytical solution can be solved correspondingly as shown below:

−푘1푡 [퐹푒푟푟𝑖푐] = [퐸]0푒푥푝 (C9)

푘 1 −푘1푡 −푘푅퐻푡 [퐶표푚푝표푢푛푑 퐸푆] = (푒푥푝 − 푒푥푝 )[퐸]0 (C10) 푘푅퐻−푘1

−푘 푡 −푘 푡 푘1푒푥푝 푅퐻 −푘푅퐻푒푥푝 1 [퐶표푚푝표푢푛푑 푅퐻] = (1 + )[퐸]0 (C11) 푘푅퐻−푘1

For the absorbance of each species that involved in the reaction:

퐴푟(휆, 푡) = 휀퐴[퐹푒푟푟𝑖푐] + 휀퐵[퐶표푚푝표푢푛푑 퐸푆] + 휀퐶[퐶표푚푝표푢푛푑 푅퐻] (C12)

200

(푘푅퐻 − 푘1)휀퐴 + 푘1휀퐵 − 푘푅퐻휀퐶 −푘1푡 퐴푟(휆, 푡) = 휀퐶[퐸]0 + [퐸]0푒푥푝 푘푅퐻 − 푘1

푘1(−휀퐵 + 휀퐶) −푘푅퐻푡 + [퐸]0푒푥푝 푘푅퐻 − 푘1

When compared to Equation C3, we find that each pre-exponential term corresponding to the b-spectra. Therefore, we can solve εA, εB, εC for each:

휀퐴 = [푏0(휆) + 푏1(휆) + 푏2(휆)]/[퐸]0 (C13)

푘푅퐻−푘1 휀퐵 = [푏0(λ) − 푏2(휆)] /[퐸]0 (C14) 푘1

휀퐶 = 푏0(휆)/[퐸]0 (C15)

0.10 a 0.05 0.00 -0.05

Δ Absorbance -0.10 -0.15

300 400 500 600 700 800 Wavelength (nm) Figure C1. The first three orthonormal basis spectra in U matrix obtained from SVD analysis of the data matrix A(λ,t). The stopped-flow measurements were conducted at [DHP]=10 µM, [H2O2]=100 µM in the 100 mM KPi buffer pH 7.0 with (a) 0 µM 4-BP.

201

a 0.08

0.04

ΔA (a.u.) 0.00

-0.04

0.01 0.1 1 10 Time (s) Figure C2. The first three rows in VT matrix obtained from SVD analysis of the data matrix A(λ,t). The three vectors were global-fitted to the biexponential function. The stopped-flow measurements were conducted at [DHP]=10 µM, [H2O2]=100 µM in the 100 mM KPi buffer pH 7.0 with (a) 0 µM 4-BP.

a 0.8

0.4

Δ Absorbance 0.0

-0.4 300 400 500 600 700 800 Wavelength (nm) Figure C3. The b-spectra obtained from SVD analysis of the data matrix A(λ,t). The stopped-flow measurements were conducted at [DHP]=10 µM, [H2O2]=100 µM in the 100 mM KPi buffer pH 7.0 with (a) 0 µM 4-BP.

Scheme C2. Kinetic model of reaction between ferric DHP with H2O2 in the presence of 4-BP for the global-fitting analysis of the stopped-flow data In the case when 4-BP in the presence of the reaction, two parallel reactions are needed to be considered. Since Ferric DHP is preincubated with inhibitor 4-BP, there is an equilibrium between free ferric DHP and 4-BP bound ferric DHP which determined by dissociation

202

T constant Kd. Thus, each row of Vr matrix was fitted to the tetraexponential function according

T to the proposed kinetic model (Scheme B2). and these three vectors of Vr matrix are applied to global-fitting to determine the apparent rate constant k1, kRH, ki1 and kiRH.

푇 −푘1푡 −푘푅퐻푡 −푘푖1푡 −푘푖푅퐻푡 푣𝑖 (푡) = 푐𝑖0 + 푐𝑖1푒푥푝 + 푐𝑖2푒푥푝 + 푐𝑖3푒푥푝 + 푐𝑖4푒푥푝 (C16)

2 푇 푇 푉 = ∑ 푣𝑖 (푡) (C17) 𝑖=0

Therefore, the 3×5 C matrix has been determined and subsequently used to calculate the b- spectra (Figure B3) according to the equation (B.16) :

푐00 푐01 푐02 푐03 푐04 퐶 = |푐10 푐11 푐12 푐13 푐14| (C18) 푐20 푐21 푐22 푐23 푐24

2 ΄ ΄ 퐵 = ∑ 푏𝑖 (휆) = 푈푟푆푟퐶 (C19) 𝑖=0

Based on the kinetic model proposed above, we basically consider the transient inhibition kinetics as two separate reactions, both of which are two step, three species first order irreversible reactions. By comparing the b-spectrum obtained in the presence of 4-BP and in the absence of 4-BP, k1 and k2 should be able to be identified from the four rate constants obtained from the global-fitting analysis. Therefore, the other two rate constants must be k3 and k4 for the reaction in the presence of 4-BP. However, the b-spectra corresponding to long- term species Compound RH and Compound iRH are overlapped as they both corresponding to the constant term during the global-fitting. In other words, the b-spectrum obtained after

203

global-fitting analysis is actually a linear combination of the b-spectrum of Compound RH and

Compound iRH.

΄ ΄ ΄ 푏0(휆) = α푏5(휆) + (1 − α)푏6(휆) (C20)

Where b0 is the b-spectrum obtained after the global-fitting analysis, α is the coefficient for the linear combination, b5 and b6 are the b-spectra corresponding to Compound RH and

Compound iRH. Presumably, the b-spectra corresponding to the reaction of non 4-BP bound

DHP should be the same as the reaction of ferric DHP with H2O2 in the absence of 4-BP in terms of the shape, while the amplitude corresponds to the concentration.

푏΄ (휆) 푏΄ (휆) 푏΄ (휆) 훼 = 1 = 2 = 5 (C21) 푏1(휆) 푏2(휆) 푏0(휆)

Thus, the coefficient can be determined by the b΄1 over b1 or b΄2 over b2. Then b΄5 and b΄6 are obtained. The following steps for non 4-BP bound DHP are basically the same as for the reaction in the absence of 4-BP. The only difference is that b0 is replaced by b΄5, which gives:

΄ ΄ ΄ 휀퐴 = [푏5(휆) + 푏1(휆) + 푏2(휆)]/(훼[퐸]0) (C22)

΄ 푘푅퐻−푘1 ΄ 휀퐵 = [푏5(휆) − 푏2(휆)] /(훼[퐸]0) (C23) 푘1

΄ 휀퐶 = 푏5(휆)/(훼[퐸]0) (C24)

Whereas for the 4-BP bound DHP, we have

−푘푖1푡 [퐹푒푟푟𝑖푐 ··· 4 − 퐵푃] = (1 − 훼)[퐸]0푒푥푝 (C25)

204

푘 푖1 −푘푖1푡 −푘푖푅퐻푡 [퐶표푚푝표푢푛푑 𝑖퐸푆] = (푒푥푝 − 푒푥푝 )(1 − 훼)[퐸]0 (C26) 푘푖1−푘푖푅퐻

−푘 푡 −푘 푡 푘푖1푒푥푝 푖푅퐻 −푘푖푅퐻푒푥푝 푖1 [퐶표푚푝표푢푛푑 𝑖푅퐻] = (1 + )(1 − 훼)[퐸]0 (C27) 푘푖푅퐻−푘푖1

And the reconstruct spectra of 4-BP bound DHP, Compound iOX and Compound iRH

correspond to εD, εE, εF are given as:

΄ ΄ ΄ 휀퐷 = [푏6(휆) + 푏3(휆) + 푏4(휆)]/((1 − 훼)[퐸]0) (C28)

΄ 푘푖푅퐻−푘푖1 ΄ 휀퐸 = [푏6(휆) − 푏4(휆)] /((1 − 훼)[퐸]0) (C29) 푘푖1

΄ 휀퐹 = 푏6(휆)/((1 − 훼)[퐸]0) (C30)

0.10 b c 0.15 0.05 0.10 0.00 0.05 -0.05 0.00

-0.10 Δ Absorbance

Δ Absorbance -0.05 -0.10 -0.15 -0.15 -0.20 300 400 500 600 700 800 300 400 500 600 700 800 Wavelength (nm) Wavelength (nm) 0.20 0.15 d 0.10 0.05 0.00

ΔAbsorbance -0.05 -0.10 -0.15 300 400 500 600 700 800 Wavelength (nm) Figure C4. The first three orthonormal basis spectra in U matrix obtained from SVD analysis of the data matrix A(λ,t). The stopped-flow measurements were conducted at [DHP]=10 µM, [H2O2]=100 µM in the 100 mM KPi buffer pH 7.0 with (b) 10 µM 4-BP, (c) 20 µM 4-BP (d) 100 µM 4-BP.

205

40 b c 0.10 0

0.05 A (a.u)

-40 A(a.u)  Δ Δ 0.00 -3 -80x10 -0.05 0.01 0.1 1 10 0.01 0.1 1 10 Time (s) Time (s)

d 0.15

0.10

0.05 A (a.u)A  0.00

-0.05

0.01 0.1 1 10 Time (s)

Figure C5. The first three rows in VT matrix obtained from SVD analysis of the data matrix A(λ,t). The three vectors were global-fitted to the tetraexponential function. The stopped-flow measurements were conducted at [DHP]=10 µM, [H2O2]=100 µM in the 100 mM KPi buffer pH 7.0 with (b) 10 µM 4-BP, (c) 20 µM 4-BP, (d) 100 µM 4-BP.

0.6 0.6 b (No 4-BP bound) b(4-BP bound) 0.4 0.4 0.2

0.0 0.2

Absorbance

Absorbance  -0.2  0.0 -0.4 300 400 500 600 700 800 300 400 500 600 700 800 Wavenlength (nm) Wavelength (nm)

206

0.6 c (No 4-BP bound) 0.4 c (4-BP bound) 0.4 0.3 0.2

0.2 0.1

Absorbance

Absorbance Δ 0.0 Δ 0.0 -0.1 -0.2 300 400 500 600 700 800 300 400 500 600 700 800 Wavelength (nm) Wavelength (nm)

0.4 d (No 4-BP bound) 1.0 d (4-BP bound) 0.8 0.3 0.6 0.2 0.4 0.1

Absorbance 0.2 Δ Absorbance 0.0  0.0 -0.1 -0.2

300 400 500 600 700 800 300 400 500 600 700 800 Wavelength (nm) Wavelength (nm)

Figure C6. The b-spectra obtained from SVD analysis of the data matrix A(λ,t). The stopped-flow measurements were conducted at [DHP]=10 µM, [H2O2]=100 µM in the 100 mM KPi buffer pH 7.0 with (b) 10 µM 4-BP, (c) 20 µM 4-BP, (d) 100 µM 4-BP.

411 0s 120s 0.20

0.15

0.10 Absorbance 0.05

300 400 500 600 Wavelength (nm) Figure C7. UV-Vis spectra of preformed Compound RH titrated with 4-BP. The bench-top UV-Visible measurement was conducted at [DHP]=10 µM, [H2O2]=100 µM in the 100 mM KPi buffer pH 7.0 with 4-BP titrated in (final concentration 100 µM) after Compound RH has formed.

207

Appendix D

Supporting Information for Chapter 5

D1. Duration of lag phase is proportional to the concentration of H2Q.

TCP 800 μM 30 TBP 800 μM

10 Lag Phase (s)

20 40 60 80 100 [Hydroquinone] (μM)

D2. Singular Value Decomposition (SVD) and Global-fitting Analysis of

Time-Resolved Spectra

The kinetic assay was conducted using an Agilent 8453 benchtop UV-visble spectrometer. The time–resolved spectral data were first collected. Subsequently, SVD and global-fitting analysis were conducted.

Applying the SVD, we have:

푇 퐴푟(휆, 푡) = 퐵푇 = 푈푟푆푟푉푟 = 푈푟푆푟퐶푇

푇 where 푉푟 = 퐶푇, 퐵 = 푈푟푆푟퐶

208

Since at least 10-fold excess of hydroquinone (H2Q) was present in all kinetic assays, pseudo- first order approximation can be made for the reaction.

T Then each vector vi(t) of matrix Vr can be fitted to the single exponential function with a baseline:

−푘표푏푠푡 푣𝑖(푡) = 푐𝑖1 + 푐𝑖2푒푥푝

Thus the 2×2 C matrix is determined as follow:

푐 푐 퐶 = | 11 12| 푐21 푐22

퐵 = ∑ 푏𝑖 (휆) = 푈푟푆푟퐶 𝑖=0

According to the proposed kinetic model, the ferric state enzyme is going to experience an exponential decay while the oxy-ferrous form of the enzyme increases correspondingly:

−푘표푏푠푡 [퐹푒푟푟𝑖푐] = [퐸0]푒푥푝

−푘표푏푠푡 [푂푥푦 − 퐹푒푟푟표푢푠] = [퐸0](1 − 푒푥푝 )

Thus, for the total absorbance during the reaction:

퐴푟(휆, 푡) = 휀퐴[퐹푒푟푟𝑖푐] + 휀퐵[푂푥푦 − 퐹푒푟푟표푢푠]

−푘표푏푠푡 −푘표푏푠푡 −푘표푏푠푡 퐴푟(휆, 푡) = 휀퐴[퐸0]푒푥푝 + 휀퐵[퐸0](1 − 푒푥푝 ) = 휀퐵[퐸0] + (휀퐴 − 휀퐵)[퐸0]푒푥푝

Therefore, the absorption spectra for the ferric and oxy-ferrous DHP are given below:

휀퐴[퐸0] = 푏0(휆) + 푏1(휆)

휀퐵[퐸0] = 푏0(휆)

A. Reaction between ferric DHP A and H2Q

209

The reaction between H2Q and ferric DHP A (without the addition of co-substrate H2O2) is a simple reduction of the protein in a bimolecular process leading to the formation of oxyferrous DHP A. The Soret band shifted from 407 nm to 416 nm and double peaked Q bands at

542 nm and 579 nm appeared indicating the formation of oxy-ferrous DHP.

0.25 0s 6s 0.20 10s 19s 29s 0.15 39s 60s

0.10 Absorbance 0.05

0.00 300 400 500 600 Wavelength (nm)

Figure D2. Time resolved UV-Visible spectrum for the reduction of ferric DHP A (5 µM) by H2Q (100 µM) in the 100 mM KPi buffer, pH 7.0 at the 60s time intervals.

T Figure D3. Single exponential fitting of two v vectors of V matrix of the reaction between ferric DHP A and H2Q.

210

350

300 )

-1 250 M

-1 200 (s

1 150 k 100 50 5.0 5.5 6.0 6.5 7.0 7.5 pH

Figure D4. pH dependence of second-order rate constant k1, oxidation of H2Q by ferric DHP A (5 µM) in 100 mM KPi buffer, pH 7.0 at 298 K.

B. Reaction between oxy-ferrous DHP A and 1,4-BQ

The reaction between 1,4-BQ and oxyferrous DHP A leads to oxidation of the protein in a bimolecular process form ferric DHP A. This process is the reverse of the process shown in section 1A of the SI and thus shows the full reversibility of the electron transfer process and proton transfer process. The Soret band is blue shifted and the double peaked Q bands in the visible region gradually disappeared.

1s 2s 0.20 3s 4s 6s 0.15 8s 12s 20s 0.10 30s

60s Absorbance 0.05

0.00 300 350 400 450 500 550 600 Wavelength (nm) Figure D5. Time resolved UV-Visible spectrum for the oxidation of oxy-ferrous DHP A (5 µM) by 1,4-BQ (300 µM) in the 100 mM KPi buffer, pH 7.0 at the 60s time intervals.

211

0.25 542 nm 409 nm 579 nm 0.20 412 nm 25

-3

0.15 15 x10 0.10 10 5 Absorbance 0.05 520 540 560 580 600 0.00

300 400 500 600 Wavelength (nm) Figure D6. Reconstructed spectra of the reaction between Oxy-ferrous DHP A and 1,4-BQ.

Figure D7. Single exponential fitting of two v vectors of VT matrix of the reaction between oxy-ferrous DHP A and 1,4-BQ.

0.25

0.20 )

-1 0.15

(s obs

k 0.10

0.05

50 100 150 200 250 300 [1,4-BQ] (μM)

Figure D8. Plot of the kobs vs [1,4-BQ] for the oxidation of oxy-ferrous DHP A (5 µM) by 1,4-BQ in 100 mM KPi buffer, pH 7.0 at 298 K.

212

55 D3. pKa of Distal Histidine His in Ferric and Oxy-ferrous DHP A

Determined by Resonance Raman Spectroscopy

The pH-dependent full range resonance Raman spectra of high frequency region are displayed in Figure D9 and D11. The thermodynamic analysis of the resonance Raman spectral data relating to the deprotonation of the distal histidine, His55, is shown in this section.

1400 pH 5.0 pH 5.5 1200 pH 6.0 pH 6.5 1000 pH 7.0 pH 7.5 800 pH 8.0 600 400 200 0

1000 1200 1400 1600 1800 -1 Wavenumber (cm )

Figure D9. pH-dependent ferric DHP A resonance Raman spectra.

Scheme D11. pH-dependent equilibrium between 5cHS and 6cHS in ferric DHP A.

213

As for the ferric DHP, there exist two equilibria. At high pH, ferric DHP equilibrated with hydroxide ligated DHP at pKa = 8.1. At low pH, the 6 coordinated high spin (6cHS) the metaquo form equilibrated with 5 coordinated high spin (5cHS). Thus, there are three species all together, which are hydroxide ligated DHP (blue), 6cHS metaquo form (pink) and 5cHS form (red). According to the mass conservation, we have:

[DHP] + [DHP − OH] = [DHP]0

[DHP (5cHS)] + [DHP − H2O (6cHS)] = [DHP]

Then consider the Henderson-Hasselbalch relation for both equilibriums:

[DHP − OH] pH = p퐾 + log a1 [DHP]

[DHP − H2O(6cHS)] pH = p퐾 + log a2 [DHP(5cHS)]

Combined these two equations together, we can derive the partition functions for these three species:

[DHP]0 [DHP − OH] = 1 + 10pH−푝퐾푎1

pH−푝퐾푎1 pH−푝퐾푎2 10 [DHP]0 10 [DHP (5cHS)] = ∙ 1 + 10pH−푝퐾푎1 1 + 10pH−푝퐾푎2

pH−푝퐾푎1 10 [DHP]0 1 [DHP − H2O (6cHS)] = ∙ 1 + 10pH−푝퐾푎1 1 + 10pH−푝퐾푎2

214

Since hydroxide ligated DHP (blue), 6cHS metaquo form (pink) and 5cHS form (red) all have corresponding Raman spectra, we can derive the expression for the v2 vector obtained from the

SVD analysis. The pKa value is obtained by fitting the experimental data with the expression.

a 10pH−푝퐾푎1 10pH−푝퐾푎2 · b + c 푉2(pH) = + ∙ 1 + 10pH−푝퐾푎1 1 + 10pH−푝퐾푎1 1 + 10pH−푝퐾푎2

0.4 0.2 0.0 -0.2

Δ Intensity -0.4 V -0.6 1 V2

5.0 5.5 6.0 6.5 7.0 7.5 8.0 pH

T Figure D10. The V2 vector of V matrix were fitted into a two pKa equilibrium model.

1400 pH 5.0 pH 5.5 pH 6.0 1200 pH 6.5 pH 7.0 pH 7.5 1000 pH 8.0 pH 8.5 800 pH 9.0

600

400

200

1000 1200 1400 1600 -1 Wavenumber (cm )

Figure D11. pH-dependent oxy-ferrous DHP A Resonance Raman spectra.

215

Scheme S2. pH-dependent equilibrium between open and closed conformation and photoexicited population in oxy-ferrous DHP A.

Under laser irradiation, a photoinduced autoxidation of oxy-ferrous DHP was observed. We believe that only the oxy-ferrous DHP with distal histidine His55 in the open conformation will be excited due to the lack of hydrogen-bond stabilization to the dioxygen ligand. Thus, the equilibrium between oxy-ferrous DHP with hydrogen-bond (His55 in the closed conformation)

(Pink) and oxy-ferrous DHP without a hydrogen bond (His55 in the open conformation) (Red) combined with a proton associated first order photoexcited autoxidation of the open conformation oxy-ferrous DHP are considered. The photoexicited autoxidation of oxy-ferrous

DHP resulted in the five coordinated product (Black).

Since there are species, according to the mass conservation, we have:

Initial: [DHP − O2 − H] + [DHP − O2] = [DHP]0

Final: [DHP − O2 − H] + [DHP − O2]’ + [DHP] = [DHP]0

Consider the Henderson-Hasselbalch relation, we have:

216

[[DHP − O2 − H]] pH = p퐾a3 + log [DHP − O2] Thus, the partition function can be derived:

[DHP]0 [DHP − O2 − H] = 1 + 10pH−푝퐾푎3

pH−푝퐾푎3 10 [DHP]0 [DHP − O2] = 1 + 10pH−푝퐾푎3

Combined with first order photoexcited autoxidation process, we have:

+ −푘1[H ]t [DHP] = [DHP − O2](1 − e )

+ ′ −푘1[H ]t [DHP − O2] = [DHP − O2]e

Since the hydrogen bond to the dioxygen ligand has a minimal effect on the vibrational mode in high frequency region, we assume that [DHP-O2] (red) and [DHP-O2-H] (pink) have the same Raman spectrum. Then the expression of v2 vector can be derived and fitted to the experimental data to obtain the pKa value.

a · (1 − b · 10−pH) c · (b · 10 −pH + 10pH−푝퐾푎3 ) 푉2(pH) = + 1 + 10pH−푝퐾푎3 1 + 10pH−푝퐾푎3

0.2

0.0

-0.2 Intensity

Δ -0.4 V1 -0.6 V2

5 6 7 8 9 pH

T Figure D12. The V2 vector of V matrix were fitted into equilibrium model combined with photoexciteautoxidation conversion

217

• Scheme D3. Thermodynamic scheme of proton and electron transfer of H2Q and HQ .

D4. pH-dependent initial rate of DHP catalyzed oxidation reaction of H2Q.

1.2

)

-1 1.0

-6 0.8

(x10

o V 0.6

5.0 5.5 6.0 6.5 7.0 7.5 8.0 pH

Figure D13. pH dependence of initial rate Vo for the oxidation of H2Q (400 µM ) catalyzed by DHP A (2.4 µM) in the presence of H2O2 ( 1200 µM ) in 100 mM KPi buffer at 298K.

D5. Stopped-flow UV-visible spectra of reaction between DHP A and H2Q in the presence of H2O2.

The UV-visible spectral data of the oxidation of H2Q in the presence of H2O2 are shown in Figures D14 and D15. From the time-resolved spectra, a two steps kinetic model can be applied. In the first time period with a half life ~0.126 s, the Soret band and double branched

Q bands are stable at 417 nm, 542 nm and 579 nm respectively. In the second time period, the

218

Soret band gradually shifted from 417 nm to 419 nm and the double branched Q bands slightly red shifted, indicating the transformation of oxy-ferrous DHP to oxo-ferryl species Compound

II. This process is with a half life ~71.2 s.

0.6

0.4

Absorbance 0.2

0.0 300 400 500 600 700 800 Wavelength (nm)

417 542 579 0.6 419 80

t = 0 s 60

0.4 ι1/2 = 0.126 s -3 40

ι1/2 = 71.2 s x10 545 20 578 0

0.2 Absorbance

520 560 600 640 0.0 300 400 500 600 700 Wavelength (nm)

Figure D15. Calculated spectra of reaction species at t=0 (red); τ1/2=0.126s (Blue) and τ1/2=71.2s (Black).

D6. Inhibition Kinetic Study of 1,4-BQ

The Michaelis-Menten kinetic data showing inhibition of TCP oxidation by 1,4-BQ are presented in Figure S10. The Ki = 3.91 mM for the 1,4-BQ in terms of inhibition of 2,4,6-TCP

219

oxidation at 298 K, which shows that 1,4-BQ is a very weak inhibitor compared to 4-BP ,Ki =

0.155 mM.

o 10 25C

) 8 -1

Ms 6 -6

(*10 4 [1,4-BQ]

0 0 mM

V 1 mM 2 2 mM 3 mM

0 400 800 1200 [TCP] (μM) Figure D16. Steady-state kinetic analysis of DHP A catalyzed TCP oxidation reaction inhibited by 1,4-BQ at 25℃.

Table D10 Michealis-Menten parameters of inhibition study of 1,4-BQ.

-1 app [1,4-BQ] (µM) kcat (s ) Km (mM)

0 9.20 ± 0.40 1.64 ± 0.12

1 ~ 1.91 ± 0.13

2 ~ 2.21 ± 0.15

3 ~ 2.87 ± 0.18

220

2800 Ki = 3.91 mM )

2400

μM

(

app 2000

m K 1600

0 1000 2000 3000 [1,4-BQ] (μM) app Figure D17. Determination of Ki by linear fit of Km against the concentration of 1,4-BQ.

The expression of initial rate in terms of Michealis-Menten equation including inhibition is given by:

푉 [푇퐶푃] 푉 = 푚푎푥 푂 [1,4– 퐵푄] 퐾푚 (1 + ) + [푇퐶푃] 퐾𝑖

221

Appendix E

Supporting Information for Chapter 6

E1. Spectral Density Functions of Model-Free Analysis

2 2 푆 휏푚 퐽(휔) = 2 (푚1) 5 1 + (휔휏푚)

2 ( 2) ′ 2 푆 휏푚 1 − 푆 휏푒 ′ 휏푚휏푒 퐽(휔) = [ 2 + ′ 2] , 휏푒 = (푚2) 5 1 + (휔휏푚) 1 + (휔휏푒) 휏푚 + 휏푒

2 2 푆 휏푚 퐽(휔) = 2 + 푅푒푥 (푚3) 5 1 + (휔휏푚)

2 2 ′ 2 푆 휏푚 (1 − 푆 )휏푒 퐽(휔) = [ 2 + ′ 2] + 푅푒푥 (푚4) 5 1 + (휔휏푚) 1 + (휔휏푒)

2 τm is the overall tumbling correlation time, τe is the effective correlation time. S is the squared generalized order parameter. Rex is the chemical exchange that attributes to the R2 relaxation rate. As for m1 and m3, the internal motion of amide N-H bond can be considered very fast, thus the effective correlation time τe = 0.

2 (푆 2 − 푆2)휏′ 2 푆 휏푚 푓 푠 ′ 휏푚휏푠 퐽(휔) = [ 2 + ′ 2 ], 휏푠 = (푚5) 5 1 + (휔휏푚) 1 + (휔휏푠) 휏푚 + 휏푠

2 (1 − 푆 2 )휏′ 2 2 ′ 2 푆 휏푚 푓 푓 (푆푓 − 푆 )휏푠 ′ 휏푚휏푓 퐽(휔) = [ 2 + 2 + ′ 2 ], 휏푓 = (푚6) 5 1 + (휔휏푚) ′ 1 + (휔휏푠) 휏푚 + 휏푓 1 + (휔휏푓)

2 2 2 ′ 2 푆 휏푚 (푆푓 − 푆 )휏푠 퐽(휔) = [ 2 + ′ 2 ] + 푅푒푥 (푚7) 5 1 + (휔휏푚) 1 + (휔휏푠)

222

2 2 ′ 2 2 ′ 2 푆 휏푚 (1 − 푆푓 )휏푓 (푆푓 − 푆 )휏푠 퐽(휔) = [ 2 + 2 + ′ 2 ] + 푅푒푥 (푚8) 5 1 + (휔휏푚) ′ 1 + (휔휏푠) 1 + (휔휏푓)

In the extended model-free formalism, the internal motion can be further divided as ‘slow’ and

‘fast’ internal motion that characterized by τs, the slow effective correlation time and τf, the fast

2 effective correlation time. And Sf is the squared order parameter for the fast internal motion.

As for m5 and m7, the fast internal motion of amide N-H bond is assumed to be very fast, thus

τf = 0.

E2. Cyanide Titration and Determine Its Binding Affinity in DHPA

1.4 407 nm 422 nm A 0.1 B 1.2

0.0 e

c 1.0

n -0.1 a

b 0.8 A r Cyanide

o -0.2 Δ

s 0.6 b -0.3 A 0.4 -0.4 0.2 K = 0.029 μM -0.5 d 0.0

300 350 400 450 500 550 600 0 20 40 60 80 Wavelength (nm) [Cyanide] (μM) Figure E1. (A) UV-vis spectra of DHPA cyanide titration experiments and (B) DHPA cyanide binding curve.

The cyanide binding curve was fitted into the equation derived below:

[퐷퐻푃][퐿] ([퐷퐻푃] − [퐷퐻푃 − 퐿])([퐿] − [퐷퐻푃 − 퐿]) 퐾 = = 0 0 푑 [퐷퐻푃 − 퐿] [퐷퐻푃 − 퐿]

([퐷퐻푃] + [퐿] + 퐾 ) ± √([퐷퐻푃] + [퐿] + 퐾 )2 − 4[퐷퐻푃] [퐿] [퐷퐻푃 − 퐿] = 0 0 푑 0 0 푑 0 0 2

Where [DHP]0 is the total DHP concentration;

[L]0 is the total ligand concentration;

223

[DHP] is the free DHP concentration; [DHP-L] is the ligand-bound DHP concentration; [L] is the free ligand concentration.

E3. Consistency Test of Two Datasets (500 MHz and 700 MHz)

b center : 0.972 width : 0.089

0.7 0.8 0.9 1.0 1.1 1.2 1.3 J(0) (700MHz) / J(0) (500MHz)

Figure E2. J(0) consistency test of two datasets acquired at 700 MHz and 500 MHz. (a) Correlation plot of J(0); (b) Histogram distribution of the J(0) ratios.

224

E4. Color Map of Squared Generalized Order Parameters

Figure E3. Computed generalized order parameters S2 vs residue number from 5 parallel MD simulations.

90o

2 S2S N.D. > 0.90 0.90-0.83 0.83-0.70 < 0.70 Figure E4. Color mapped generalized order parameter S2 from NMR on DHPA structure.

90o

S2 N.D. > 0.90 0.90-0.83 0.83-0.70 < 0.70 Figure E5. Color mapped generalized order parameter S2 from MD simulations on DHPA structure.

225

1 15 Table E1. DHPA Relaxation data of R1 , R2 and { H}- N NOE at 500 MHz and 700 MHz

226

Table E1 (Continue)

227

Table E2. DHPA Reduced spectral density J(0), J(0.87ωH) and J(ωN) at 500 MHz and 700 MHz

228

Table E2 (Continue)

229

Table E3. DHPA Model-free analysis dataset

230

Table E3 (Continue)

231

Table E4. Generalized order parameter S2 from NMR and MD simulations

232

Appendix F

Supporting Information for Chapter 7

1.2 a 423 nm 0.15 b

1.0 0.10 e c 414 nm

n 0.8 0.05

r Δ

o 0.6 0.00

s b

A 0.4 -0.05 540 nm 570 nm 0.2 -0.10 0.0 -0.15 300 400 500 600 0 5 10 15 20 Wavelength (nm) Time (hours) Figure F1. The time-resolved UV-Vis spectra of CO-driven autoreduction reaction of WT DHPA at pH 8.0 and corresponding time course of the reaction.

233

C73S

0.20 1.4 422 nm 0.15 1.2

e 407 nm

c 1.0 0.10

a A

b 0.8 r

Δ 0.05 o

s 0.6

b 0.00 A 0.4 541 nm 571 nm -0.05 0.2 0.0 -0.10 300 400 500 600 0 5 10 15 Wavelength (nm) Time (Hours) M63L&M64L

1.0 0.15 421 nm 0.8 0.10

e 407 nm

c 0.05

n 0.6

A b

r 0.00

Δ o

s 0.4

b -0.05 A 538 nm570 nm 0.2 -0.10

0.0 -0.15 300 400 500 600 0 5 10 15 Wavelength (nm) Time (Hours) SWMb

1.4 409 nm 0.10

1.2 423 nm e

c 1.0 n

a 0.05

b r

0.8 A

Δ s

b 0.6 A 0.00 0.4 543 nm 578 nm 0.2 -0.05

300 400 500 600 0 10 20 30 40 50 Wavelength (nm) Time (Hours) Figure F2. The time-resolved UV-Vis spectra of CO-driven autoreduction reaction of DHPA C73S, M63L&M64L mutant and SWMb with their corresponding time course of the reaction.

234