Abstract Characterization of Potential Anti-Infective

ABSTRACT

CHARACTERIZATION OF POTENTIAL ANTI-INFECTIVE AGENTS OF Burkholderia pseudomallei TARGETING IspF

Joy Michelle Blain, Ph.D. Department of Chemistry and Biochemistry Northern Illinois University, 2016 James R. Horn, Co-Director Timothy J. Hagen, Co-Director

There is an urgent need for new anti-infective agents. Pathogenic organisms continue to develop resistance to modern-day antibiotics, while there has been a significant decline in the development of antibiotics over the last 30 years. One particular pathway of interest in the development of new antibiotics is the biosynthesis of isoprenoids. These compounds are essential chemical building blocks in pathogenic organisms such as Plasmodium falciparum, Mycobacterium tuberculosis, and

Burkholderia pseudomallei. Notably, there are two distinct pathways in the biosynthesis of isoprenoids called the Mevalonate (MVA) pathway and the 2C-methyl-D-erythritol phosphate (MEP) pathway. The MEP pathway enzymes are attractive targets in developing novel anti-infective agents due to the fact that humans only use the MVA pathway to produce isoprenoids. Burkholderia pseudomallei is a Gram-negative bacterium found in untreated water and soil in Southeast Asia and northern Australia. The bacterium has been classified as a category B bioterrorism agent/disease by the category B bioterrorism agent/disease by the Centers for Disease Control and Prevention.

Antibiotic resistance to traditional anti-microbial agents in such pathogenic organisms has increased due to the primary mechanisms of resistance, which are enzyme inactivation, drug efflux from the cell, and target deletion. Consequently, new anti- infective agents targeting B. pseudomallei are of great interest.

The focus of this research project was to examine the inhibition of the fifth enzyme in the MEP pathway, 2C-methyl-D-erythritol 2,4-cyclodiphosphate synthase

(IspF) using small molecules that were developed based on fragment hits and structure activity relationships (SAR). In this study, enzyme engineering was utilized for the development of a high throughput thermal-shift assay. The assay was used to assess the relative potency of compounds targeting IspF. The inhibitor binding affinities and thermodynamics were determined using Surface Plasmon Resonance (SPR) and

Isothermal Titration Calorimetry (ITC). Enzyme inhibition values for the most potent inhibitors were then evaluated using a High Performance Liquid Chromatography

(HPLC) activity assay. Finally, structural analysis for two compounds in complex with

IspF was completed to assess the mode of binding in the active site, which will aid in future inhibitor design strategies. NORTHERN ILLINOIS UNIVERSITY DE KALB, ILLINOIS

AUGUST 2016

CHARACTERIZATION OF POTENTIAL ANTI-INFECTIVE AGENTS OF Burkholderia pseudomallei TARGETING IspF

A DISSERTATION SUBMITTED TO THE GRADUATE SCHOOL

IN PARTIAL FULFILLMENT OF THE REQUIREMENTS

FOR THE DEGREE OF

DOCTOR OF PHILOSOPHY

DEPARTMENT OF CHEMISTRY AND BIOCHEMISTRY

Doctoral Director:

James R. Horn, Co-Director

Timothy J. Hagen, Co-Director

ACKNOWLEDGEMENTS

I would like to first thank my graduate co-advisor, Professor James Horn, for his continual support, guidance, and patience. He has constantly challenged me throughout the last five years and I have become a better scientist because of his encouragement. I would also like to thank my co-advisor, Professor Timothy Hagen, for his guidance and enthusiasm. It has been an honor to work with both Professors Horn and Hagen.

It is with tremendous gratitude that I acknowledge Rick Walter and Gina Raneiri for their invaluable help at the Advanced Photon Source. They have not only helped me immeasurably with my research but have become more than just colleagues. I have enjoyed all of our conversations and appreciate all your encouragement over the last two years. I would also like to thank Matt Pokross for all the knowledge he has passed on to me to in determining the crystal structures.

I am also thankful to my committee members Professor Gary Baker and Professor Oliver

Hofstetter. They have continually supported and aided in my work through numerous valuable discussions. Many thanks to Dr. Heike Hofstetter for her help with the NMR assay and all our conversations.

My family and friends have been an incredible support system throughout my graduate school career. Their support when I needed it the most has been amazing and this would not have been possible without you. To my parents, Leland and Soon Ik Blain thank you for continual love and support. You have always encouraged me in my academic interests. I would not be

iii where I am now without your inspiration and willingness to help me. To my sisters,

Jennifer and Jill, my best friends and always being there for me and supporting me. To my nephews and niece, Leonidas, Leandra, and Liev, you are the light in my life. Thanks to my paternal grandparents, Phillip and Nora Blain for your endless support. My maternal grandparents, Jin Darl Son and Do Hee Kim, have passed away, but the work ethic of my grandfather and kindness of my grandmother have guided me throughout my life.

I wish to thank my aunt and uncle Eun Ik and Duk Yoon Song for their love and kindness and my cousins Eun Sun Song and Eun Hee Lee and her family, Young, Yu Jin, Yoori, and Yoo

Hyun Lee for their love and support. I would also like to thank my aunt Connie Blain, and my uncle and aunt, Robert and Coleen Blain and their family, for their encouragement. I am also grateful to my maternal uncles and aunts living in Korea: Dea Ik Son, Dong Ik Son, Ik Soon Son,

Kyong Ik Son, their spouses and all of my cousins. Although we live far apart and are not able to see each other you are always close in my thoughts.

Several thanks go out to all members of the Horn and Hagen labs, past and present. You have provided valuable feedback throughout our conversations. The group members have become not only colleagues but also friends.

I would like to acknowledge Northern Illinois University and the Department of

Chemistry and Biochemistry for funding support. Additionally, I would like to thank the NIU

Foundation, National Institute of Health (1R15AI113653-01), and National Science Foundation

(MCB-0953323) for providing the funding and financial support I have received.

DEDICATION

To Leonidas “Leo” Lucien Blain Christen:

March 23, 2010 – April 8, 2010

Always on my mind Forever in my heart

TABLE OF CONTENTS

Page

LIST OF TABLES ...... x

LIST OF FIGURES ...... xi

LIST OF APPENDICES ...... xv

LIST OF ABBREVIATIONS ...... xvi

LIST OF SYMBOLS ...... xx

CHAPTER

1 INTRODUCTION

Cellular Targets for Antibiotic Development...... 5

Inhibition of Cell Wall Biosynthesis...... 7

Inhibition on DNA/RNA Synthesis ...... 9

Inhibition of Protein Biosynthesis ...... 11

Mechanisms of Antibiotic Resistance ...... 13

Investigation of Novel Targets and Novel Classes of Antibiotics ...... 15

Known Targets of Antibiotics ...... 15

Bacterial Genes which are Essential, Conserved ...... 16

Biosynthesis of Isoprenoids as a Potential Anti-Infective Target ...... 17

Two Distinct Isoprenoid Biosynthetic Pathways Mevalonate and Non- mevalonate ...... 19

vi Mevalonate Isoprenoid Biosynthetic Pathway ...... 19

Discovery of the Non-mevalonate Isoprenoid Biosynthetic Pathway ... 21

Methylerythritol Phosphate (MEP) Pathway an Attractive Target for Developing Novel Antibiotics ...... 23

MEP pathway Regulation ...... 25

Known MEP Pathway Inhibitors ...... 25

Targeting Cyclodiphosphate Synthase, IspF ...... 29

Active Site of Cyclodiphosphate synthase (IspF) ...... 31

Cyclodiphosphate synthase (IspF) Hydrophobic Cavity ...... 34

Targeting Burkholderia pseudomallei IspF ...... 34

2 DEVELOPMENT OF A HIGH-THROUGHPUT THERMAL-SHIFT ASSAY

Introduction ...... 36

Methods ...... 39

Expression and Purification of Burkholderia pseudomallei IspF ...... 39

Circular Dichroism...... 40

Solvent Accessible Surface Area Calculations ...... 41

Computational Stability Predictions ...... 41

Site-Directed Mutagenesis of Burkholderia pseudomallei IspF ...... 42

Thermal-Shift Assay ...... 44

Preparation of IspF Substrate ...... 44

IspF Enzymatic Rate Determination ...... 45

Crystallization of Burkholderia pseudomallei IspF Mutant ...... 46

X-Ray Diffraction ...... 46

vii X-Ray Structure Solution ...... 47

Results and Discussion ...... 48

Purification of Burkholderia pseudomallei IspF...... 48

Determination of Burkholderia pseudomallei IspF Melting Temperature ...... 51

Analysis of Burkholderia pseudomallei IspF Mutations ...... 54

Analysis of Burkholderia pseudomallei IspF Mutants Stability ...... 59

Enzymatic Rate of Burkholderia pseudomallei IspF ...... 62

Crystal Structure of Burkholderia pseudomallei IspF Mutant Q151E .. 64

Analysis of Burkholderia pseudomallei IspF Inhibitors in Thermal- Shift Assay ...... 68

Pyrimidine Series ...... 68

Tryptophan and Tryptophan hydroxamate Compounds ...... 77

Imidazothiazole Series ...... 79

Imidazole Series ...... 89

Conclusions ...... 96

3 CHARACTERIZATION OF BINDING AFFINITY OF BURKHOLDERIA PSEUDOMALLEI ISPF INHIBITORS

Introduction ...... 98

Methods ...... 99

Determination of Dissociation Constant by SPR ...... 99

Determination of Dissociation Constant by ITC ...... 101

Determination of Binding Affinity using Fluorescence Polarization .... 101

Results and Discussion ...... 104

viii Fusion Series Binding Affinity by SPR ...... 104

Binding Affinity of Imidazole Series by SPR...... 107

Imidazothiazole Series Binding Affinity by SPR ...... 117

Thermodynamic Analysis of Binding by ITC ...... 122

Development of a Fluorescence Polarization Assay ...... 127

Conclusions ...... 132

4 DEVELOPMENT OF ISPF ENZYME INHIBITION ASSAY USING HPLC

Introduction ...... 134

Methods ...... 135

Determination of Burkholderia pseudomallei IspF Inhibition ...... 135

Results and Discussion ...... 136

Formation of CDP-ME2P ...... 136

Burkholderia pseudomallei IspF Enzymatic Rate ...... 140

Inhibition of Burkholderia pseudomallei IspF ...... 144

Conclusions ...... 159

5 STRUCTURAL ANALYSIS OF INHIBITORS USING CRYSTALLOGRAPHY

Introduction ...... 161

Methods ...... 162

Co-Crystallization of Burkholderia pseudomallei IspF Inhibitors ...... 162

Burkholderia pseudomallei IspF Inhibitor Soaks ...... 163

Results and Discussion ...... 163

BpIspF Complex with L-tryptophan hydroxamate ...... 163

ix BpIspF Complex with HGN-0289 ...... 169

Conclusions ...... 174

6 SUMMARY, CONCLUSIONS, AND FUTURE DIRECTIONS 175

REFERENCES ...... 180

APPENDICES ...... 194

LIST OF TABLES

Page

Table 2-1 Sequence of Burkholderia pseudomallei IspF ...... 40

Table 2-2 Primers for mutagenesis of Burkholderia pseudomallei IspF ...... 43

Table 2-3 Surface burial of individual residues of BpIspF ...... 56

Table 2-4 Determination of apparent binding affinity of pyrimidine series ...... 72

Table 2-5 Analysis of tryptophan analogs apparent binding affinity ...... 78

Table 2-6 Imidazothiazole series thermal-shift analysis ...... 81

Table 2-7 Imidazole series thermal-shift assay analysis ...... 90

Table 3-1 Binding affinity and stoichiometry of fusion series by SPR ...... 105

Table 3-2 Binding affinity of imidazole series by SPR ...... 112

Table 3-3 Binding affinity of imidazothiazole series by SPR ...... 119

Table 3-4 Thermodynamic analysis of fusion compounds binding by ITC ...... 123

Table 3-5 Tryptophan and tryptophan hydroxamate binding affinity determined by ITC ...... 125

Table 4-1 Compounds screened in HPLC enzyme inhibition assay ...... 145

Table 4-2 Summary of BpIspF Enzyme Inhibition by Compounds ...... 157

LIST OF FIGURES

Page

Figure 1-1 FDA approved antibacterial drugs from 1980-2014 ...... 2

Figure 1-2 Timeline of new class of antibiotics...... 3

Figure 1-3 Chemical structures of representative antibiotics ...... 4

Figure 1-4 Antibiotic targets within the cell...... 6

Figure 1-5 Drug-target interaction of cell wall biosynthesis by -lactams ...... 8

Figure 1-6 Drug-target interaction of topoisomerase inhibition by Quinolones ...... 10

Figure 1-7 Drug-target interaction of aminoglycoside inhibition of protein biosynthesis ...... 12

Figure 1-8 Mechanisms of antibiotic resistance by cellular targets ...... 14

Figure 1-9 Representation of the biosynthesis of isoprenoids from the isoprene building blocks ...... 18

Figure 1-10 Biosynthesis of DMAPP and IPP in the mevalonate (MVA) pathway ...... 20

Figure 1-11 The elucidation of the MEP pathway ...... 22

Figure 1-12 The biosynthesis of DMAPP and IPP by the MEP pathway...... 24

Figure 1-13 Substrates of MEP pathway enzymes and their known inhibitors ...... 28

Figure 1-14 Proposed IspF mechanism...... 29

Figure 1-15 Escherichia coli IspF displaying three active sites at interface ...... 30

Figure 1-16 Active site of Escherichia coli IspF ...... 32

Figure 1-17 Cytosine ring hydrogen bonds to backbone of EcIspF ...... 33

Figure 2-1 Coupled plate-based IspF assay ...... 37

xii Figure 2-2 HPLC method to monitor IspF activity ...... 37

Figure 2-3 Plate-based assay measuring resorufin...... 38

Figure 2-4 Burkholderia pseudomallei IspF purification by His-trap ...... 49

Figure 2-5 Size exclusion chromatography of BpIspF ...... 50

Figure 2-6 Wavelength scan of Burkholderia pseudomallei IspF ...... 52

Figure 2-7 Thermal unfolding of Burkholderia pseudomallei IspF...... 53

Figure 2-8 Burkholderia pseudomallei IspF homotrimer residues for mutagenesis...... 57

Figure 2-9 FoldX stability predictions for single point mutations ...... 58

Figure 2-10 Thermal unfolding of Burkholderia pseudomallei IspF mutants ...... 61

Figure 2-11 Enzymatic rate of Burkholderia pseudomallei IspF wild-type and mutant Q151E ...... 63

Figure 2-12 Overlay of Burkholderia pseudomallei IspF wild-type and Q151E Mutant ...... 65

Figure 2-13 Side chain comparison of Q151E mutant shows small differences structurally ...... 66

Figure 2-14 Active site overlay of BpIspF wild-type and mutant ...... 67

Figure 2-15 Observed binding model of various MMP inhibitors ...... 69

Figure 2-16 Zinc chelate binding modes designed to mimic MMP inhibitors ...... 70

Figure 2-17 Active site of EcIspF hydrogen bonding of cytosine ring...... 70

Figure 2-18 Example of BpIspF Unfolding for the thermal shift assay ...... 71

Figure 2-19 Structure of Imidazothiazole for substituents ...... 79

Figure 2-20 Crystal structure of BpIspF in complex with FOL717 ...... 80

Figure 3-1 EDC/NHS coupling chemistry for SPR ...... 100

Figure 3-2 Instrumental setup of fluorescence polarization assay ...... 103

xiii Figure 3-3 Docking HGN-0006/HGN-0007 display structural differences in binding . 106

Figure 3-4 Fragment hits identified through an STD-NMR screen ...... 110

Figure 3-5 SPR chromatogram of L-tryptophan hydroxamate ...... 111

Figure 3-6 Crystal structure of FOL955 fragment hit ...... 116

Figure 3-7 Crystal structure of FOL535 fragment hit ...... 121

Figure 3-8 Titration of HGN-0006 with BpIspF to determine Kd,app ...... 123

Figure 3-9 Binding affinity of HGN-0003 determined by ITC at pH 7.4...... 126

Figure 3-10 Illustration of FP principle of enzyme binding fluorophore ...... 128

Figure 3-11 Wavelength scans of fluorescent ligand ...... 129

Figure 3-12 Fluorescence polarization of BpIspF binding HGN-0201 ...... 131

Figure 4-1 IspF enzymatic reaction monitored by HPLC ...... 135

Figure 4-2 IspF substrate, CDP-ME2P, formation ...... 137

Figure 4-3 Chromatogram of cytosine/adenosine molecules in CDP-ME2P formation 137

Figure 4-4 IspD reaction monitored by HPLC ...... 138

Figure 4-5 IspE reaction monitored by HPLC ...... 139

Figure 4-6 Full enzymatic profile of BpIspF ...... 141

Figure 4-7 Overlay of IspF reaction from 2 to 45 minutes by HPLC ...... 142

Figure 4-8 Analysis of BpIspF rate of enzymatic reaction ...... 143

Figure 4-9 Enzyme inhibition by HGN-0003 ...... 148

Figure 4-10 Enzymatic rate in presence of HGN-0003 ...... 149

Figure 4-11 Determination of Inhibition by HGN-0107...... 150

Figure 4-12 Enzymatic rate in presence of HGN-0107 ...... 151

Figure 4-13 Inhibition by HGN-0201 ...... 152

xiv

Figure 4-14 Rate in presence of HGN-0201 ...... 153

Figure 4-15 Inhibition of BpIspF by HGN-0289 ...... 154

Figure 4-16 Enzymatic rate in presence of HGN-0289 ...... 155

Figure 5-1 Structure of L-tryptophan hydroxamate and N-isopropyl-3-methylimidazo [2,1-b]thiazole-2-carboxamide ...... 162

Figure 5-2 Crystal structure of L-tryptophan hydroxamate complex ...... 165

Figure 5-3 Known mode of hydroxamic acids binding to metal ions ...... 166

Figure 5-4 Electron density of L-tryptophan hydroxamate ...... 166

Figure 5-5 BpIspF and EcIspF overlaid with L-tryptophan hydroxamate ...... 168

Figure 5-6 Co-crystal of HGN-0289 with BpIspF ...... 171

Figure 5-7 Electron density HGN-0289 ...... 172

Figure 5-8 Overlay of FOL535 with HGN-0289 ...... 173

Figure 5-9 Chemical structure of HGN-0321 ...... 174

LIST OF APPENDICES

APPENDIX A: Burkholderia pseudomallei IspF VECTOR ...... 194

APPENDIX B: SPR SENSOGRAMS OF IMIDAZOLE SERIES ...... 199

APPENDIX C: SPR SENSOGRAMS OF IMIDAZOTHIAZOLE SERIES ...... 208

APPENDIX D: IDENTIFICATION OF NEW HIT COMPOUNDS FOR BpIspF ...... 212

APPENDIX E: ANALYSIS OF INHIBITOR BINDING BY STD-NMR ...... 217

APPENDIX F: X-RAY CRYSTAL STRUCTURE STATISTICS ...... 224

LIST OF ABBREVIATIONS

(His)6 hexahistidine affinity tag

Ac acetyl

ADP adenosine diphosphate

AMP ampicillin

ATP adenosine triphosphate

Bp Burkholderia pseudomallei

BSA bovine serum albumin

CD circular dichroism

CDP cytidine diphosphate

CDP-ME 4-diphosphocytidyl-2-C-methylerythritol

CDP-ME2P 4-diphosphopcytidyl-2C-methyl-D-erythritol 4-phosphate

CMP cytidine monophosphate

CMPK cytidine monophosphate kinase

CoA acetyl Coenzyme A

CTP cytidine triphosphate

D-G3P D-glyceraldehyde 3-phosphate

DMAPP dimethylallyl pyrophosphate

DMSO dimethyl sulfoxide

DNA deoxyribonucleic acid

DSF differential scanning fluorimetry

xvii DTT dithiothreitol

DXP 1-deoxy-D-xylulose 5-phosphate

Dxs deoxyxylulose 5-phosphate synthase

Ec Escherichia coli

EDC N-Ethyl-N′-(3-dimethylaminopropyl)carbodiimide hydrochloride

FDA US Food and Drug Administration

FOL fragment of life

FP fluorescence polarization

Gnd-HCl guanidine hydrochloride

GPP geranyl pyrophosphate

HEPES 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid

HGN compound code

HMBDP (E)-4-hydroxy-3-methylbut-2-enyl diphosphate

HMG 3-hydroxy-3-methylglutaryl

HMGR 3-hydroxy-3-methylglutaryl Coenzyme A reductase

HPLC high performance liquid chromatography

IC50 inhibitory concentration at 50%

IPP isopentyl pyrophosphate

IPTG isopropyl -D-thiogalactoside

IspC/Dxr 1-deoxy-D-xylulose 5-phosphate reductoisomerase

IspD 2-C-methyl-D-erythritol 4-phosphate cytidylyltransferase

IspE 4-diphosphocytidyl-2-C-methyl-D-erythritol kinase

IspF 2-C-methyl-D-erythritol 2,4-cylcodiphosphate synthase

xviii IspG 4-hydroxy-3-methylbut-2-en-1-yl diphosphate synthase

IspH 4-hydroxy-3-methylbut-2-enyl diphosphate reductase

ITC isothermal titration calorimetry

LB luria-bertani

Le Lycopersicon esculentum

MEcDP 2C-methyl-D-erythritol 2,4-cyclodiphosphate

MEP 2C-methyl-D-erythritol 4-phosphate

MEP 2C-methyl-D-erythritol 4-phosphate pathway mRNA messenger ribonucleic acid

MS mass spectrometry

MVA mevalonate pathway

MWCO molecular weight cut-off

NaCl sodium chloride

NAD+ nicotinamide adenine dinucleotide oxidized

NADH nicotinamide adenine dinucleotide reduced

NADP+ nicotinamide adenine dinucleotide phosphate oxidized

NADPH nicotinamide adenine dinucleotide phosphate reduced

NHS N-hydroxysuccinimide

NMR nuclear magnetic resonance

OD600 optical density at 600 nm

PBP penicillin-binding protein

PBS phosphate buffer saline

PCR polymerase chain reaction

xix PDB protein data bank

PEEK polyether ether ketone

PEG polyethylene glycol

PEP phosphoenol pyruvate

PES polyethersulfone

PK pyruvate kinase

RNA ribonucleic acid

ROS reactive oxygen species rpm rotations per minute

SAR structure activity relationship

SDS sodium dodecylsulfate

SE size exclusion

SPR surface plasmon resonance

STD-NMR saturation transfer difference nuclear magnetic resonance

TBLASTN translated basic local alignment search tool of nucleotide

TCEP tris(2-chloroethyl) phosphate

TEV tobacco etch virus

μRIU refractive index units

LIST OF SYMBOLS

Å Angstrom

Ab anisotropy of the bound probe

Af anisotropy of the free probe

Aobs observed anisotropy

C2H2MgO4∙2H2O Magnesium formate

Da Dalton

DPN-1 restriction enzyme which digests methylated DNA eV electron volts

γ gamma

Kd,app dissociation constant kDa kilo-Dalton

LT total probe concentration

Mg2+ magnesium ion

MgCl2 magnesium chloride

Ni2+ nickel ion nm nanometer

Ri rate in the presence of inhibitor

Ro rate of reaction in the absence of inhibitor

RT concentration of enzyme

Tm melting temperature

xxi Zn2+ zinc ion

ZnCl2 zinc chloride

alpha

beta

Δ delta

ΔASAap changes in apolar surface area

ΔASApol changes in polar surface area

ΔCp,B change in heat capacity of binding under constant pressure

ΔCp,u change in heat capacity of unfolding under constant pressure

ΔG° gibbs free energy of binding standard state

ΔG°mut gibbs free energy of equilibrium unfolding of mutant protein standard state

ΔG°WT gibbs free energy of equilibrium unfolding of wild-type standard state

ΔH° enthalpy of binding standard state

ΔS° entropy of binding standard state

ΔS°B change in the entropy of binding standard state

ΔTm change in melting temperature standard state

λ lambda

λmax wavelength at maximum signal

chi

CHAPTER 1

INTRODUCTION

The use of antibiotics has increased the average human lifespan,1 as well as improved livestock production,2 due to their ability to minimize the adverse outcomes of bacterial infections. Despite their importance in human health, the number of FDA approved antibacterial drugs has significantly decreased since the 1980’s (Figure 1-1).3 Furthermore, since 1985, the development of antibiotics has declined by greater than 80%, due to the high cost of bringing these drugs to market. On average, pharmaceutical companies spend $5 billion on research and development to bring a new drug to market, where approximately 80% of those drugs fail during testing.4 There were 20 novel classes of antibiotics that were introduced that treat bacterial infections from 1930 – 1962, but since 1962, there have been only two new classes of antibiotics that reached the market, with the remainder being structural analogs of the existing 20 classes

(Figure 1-2 and Figure 1-3).5 In the last 5 years, only six antibiotic drugs have been approved by the FDA.6,7

Figure 1-1. FDA approved antibacterial drugs from 1980-2014. Adapted from Ventola 2

Figure 1-2. Introduction of new classes of antibiotics timeline. Adapted from Walsh.8

Figure 1-3. Chemical Structure of antibiotics based on timeline shown in Figure 1-2.

5 Cellular Targets for Antibiotic Development

Antibiotics are generally categorized by their cellular target. Examples of antibiotic targets/categories include deoxyribonucleic acid (DNA) replication, transcription, protein biosynthesis, folate biosynthesis, and cell wall biosynthesis (Figure 1-4).7,9,10,11 Inhibiting these cellular targets would lead to a series of metabolic miscues, such as induction of reactive oxygen species (ROS), metabolic stress, alteration of cell wall integrity and permeability, mistranslation of proteins, as well as other effects, all which eventually lead to bacterial cell death.8

Figure 1-4. Selected antibiotic targets within the cell. Adapted from Wright.10

7 Inhibition of Cell Wall Biosynthesis- Bacterial cell walls are composed of peptidoglycans, which are covalently linked peptide-sugar repeats. The peptidoglycans maintain the cell’s shape and promote survival under harsh environments. The repeating peptide-sugar units are assembled and maintained by a complex series of enzymatic steps, which include transpeptidases and transglycosylases.8,9,12 Disruption of the cell wall synthetic machinery compromises the formation of the cell coat, as well as cell function, often leading to bacterial cell death.

Penicillins and cephalosporins, antibiotics that are classified as -lactams, are able to covalently modify and inhibit transpeptidases that cross-link the peptidoglycan units (Figure 1-5).

Analogously, vancomycin binds to D-alanyl-D-alanine dipeptide of NAM/NAG-peptides, N- acetylglucosamine and N-acetylmuramic acid. This action blocks transglycosylase and transpeptidase activity thereby inhibiting peptidoglycan biosynthesis.9

Figure 1-5. Drug-target interaction of -lactams inhibition of cell wall biosynthesis. Adapted from Kohanski8 (yellow circle: autolysin, green circle: penicillin-binding protein, pink triangle: -lactam)

Inhibition of DNA/RNA Synthesis- Levofloxacin and Ciprofloxacin are examples of antibiotics that inhibit DNA replication by targeting topoisomerase II and IV, enzymes that are involved in the breaking and rejoining of DNA, to increase or reduce DNA supercoiling, as it proceeds through replication. Topoisomerases from eukaryotes and bacteria are structurally and functionally distinct, which allows selective targeting of pathogenic topoisomerases. The aminocoumarins and quinolone class of antibiotics target DNA gyrase, a bacterial topoisomerase. These antibiotics stall DNA replication by trapping the topoisomerase as it breaks and re-ligates the DNA strand (Figure 1-6). The antibiotic subclass of rifamycin, from the family of ansamycine, targets the synthesis of ribonucleic acid (RNA) from DNA. Rifamycin is a natural product that acts allosterically, blocking nascent RNA transcript initialization, as it binds the exit site of a previously initiated transcription reaction in RNA polymerase.7,8 Disruption of these replication/transcription steps would ultimately initiate a DNA stress response mechanism that leads to bacterial cell death.8

Figure 1-6. Drug-target interaction of topoisomerase inhibition by quinolones. Adapted from Kohanski8 (pink circle: topoisomerase and purple triangle: quinolone)

11 Inhibition of Protein Biosynthesis-The translation of RNA to protein involves three steps, which are initiation, elongation, and termination. The translation machinery are anchored on the ribosome as it progresses along the RNA, resulting in the final peptide product. After biosynthesis the peptide can undergo processes such as post-translation modification and protein folding. The bacterial ribosome is made up of two subunits 50S and 30S. The larger of the two subunits, 50S, is targeted by the lincosamide, clindamycin, amphenicols, oxazolidinones, and the macrolide erythromycin. The smaller subunit, 30S, is targeted by tetracyclines and aminocyclitols. The mechanism by which these inhibitors work is by blocking initiation or translation of peptides. Changing the peptide synthesis fragments by alteration of factors associated with initiation and elongation or through obstruction of access to the amino acid building blocks can result in blocking peptide initiation and translation that results in cell death

(Figure 1-7).7,8

Figure 1-7. Drug-target interaction of aminoglycoside inhibition of protein biosynthesis. Adapted from Kohanski.8 (beige triangle: aminoglycoside, blue rectangle: membrane protein)

13 Mechanisms of Antibiotic Resistance

The use of antibiotics has resulted in a dramatic increase in human life expectancy, as otherwise deadly bacterial infections can typically be treated.13 Despite the important role antibiotics have played in improving human health, their overuse, as well as misuse, has led to increased bacterial resistance. Bacteria, through often sophisticated mechanisms, develop methods to evade the antibiotic’s mechanism of action. In a review by Soares, et al., the mechanisms of bacterial resistance against antibiotics were described as falling into three categories: efflux of the antibiotic from the cell, modification of the antibiotic target, and inactivation of the antibiotic (Figure 1-8).10,14

Figure 1-8. Mechanisms of antibiotic resistance by cellular targets. Adapted from Wright.10

15 Two examples of antibiotic resistance are observed with quinolone-based antibiotics.

These include alteration of the target and over expression of efflux pump systems and/or impermeability of the membrane. On the other hand, antibiotic resistance of -lactam antibiotics is due to deactivation by -lactamase enzymes, which hydrolyze the -lactam core.

Alternatively, resistance to vancomycin is caused by bacteria utilizing a low-affinity precursor,

D-alanine-D-lactate or D-alanine-D-serine, which replaces the high-affinity precursor D-alanine-

D-alanine. In turn this causes modification of the vancomycin-target binding. Finally, in the vast majority of observed antibiotic resistance cases, there is a decrease in membrane permeability.9,11,12,,15

Investigation of Novel Classes of Antibiotics and Novel Targets

The problem with antibiotic resistance is magnified due to the lack of new antibiotics, with only two new classes introduced since 1962. The majority of anti-bacterials that have been developed are derived from known scaffolds. To effectively combat antibiotic resistance requires the identification of novel classes of antibiotics with distinguishing scaffolds. In addition, antibiotic disruption of pathways not been previously targeted will help delay the emergence of antibiotic resistance.

Known targets of antibiotics- One strategy of developing novel classes of antibiotics is exploitation of known targets, such as protein biosynthesis and cell wall biosynthetic pathways.

Antibiotic classes such as macrolides and � –lactams, including erythromycin and penicillin, target these two essential bacterial cellular processes and are highly effective. Targeting these two pathways is a particularly interesting approach due to the many components involved in these processes. Many of these components are still being activity studied. While transpeptidases are well studied and amply targeted by numerous �-lactam antibiotics, the transglycosylases and

16 the machinery involved in the assembly of the target polysaccharide portion of the cell wall still remain a viable target for exploration in developing new classes of antibiotics. Furthermore, the size and diverse components of the translation machinery leave potential for development of novel classes of antibiotics that target the ribosome and ribosomal subunits.5

Targeting Bacterial Genes that are Essential- An alternative method of introducing new classes of antibiotics entails delving into the genomes of known pathogens to identify conserved and essential pathways or genes. Antibiotics that target proteins that are not present in mammals are often ideal, as they possess reduced potential for cross-reactivity. There were >350 genes identified by Payne et. al that participate in bacterial pathways. The identified genes encoded proteins participating in cell division, lipid metabolism, nucleotide and amino acid metabolism, partitioning of chromosomes, protein secretion, and signal transduction, as well as other bacterial pathways.16 Especially attractive examples identified by the study are the genes involved in bacterial communication and bacterial infectivity, which were found to be essential and constitutively expressed. Virulence factors, which are molecules responsible for the pathogenicity of bacteria,17 along with the proteins involved in biofilm formation have recently been investigated as potential anti-infective targets. The ability to inhibit bacterial infectivity to the host (virulence) as well as disrupt communication through biofilms would improve clearance of the weakened pathogen from the host.18,19 Researchers have also explored bacteriolytically therapeutics, such as bacteriophage-derived lysins, which can damage cell wall peptidoglycans selectively without disturbing normal microbiota. For example, in extracellular infections, lysins that are able to influence the pathogen within the host could eradicate the infection prior to inactivation of the lysin by the host immune response.,19,20

17 Biosynthesis of Isoprenoids as a Potential Anti-Infective Target- The biosynthesis of isoprenoids is a vital process in all-living organisms, including pathogenic organisms.

Isoprenoids are a large class of natural products and metabolites in nature, also known as terpenoids, with over 40,000 compounds characterized which include drugs, lipids, and hormones.21,22 The structural diversity of isoprenoids is immense and includes carotenoids, dolichols, quinones, steroids, and sesquiterpenes (Figure 1-9). Isoprenoids are taxonomically widespread and involved in multiple biological functions such as hormone signaling, electron transport, post-translational regulation of lipid biosynthesis, protein cleavage and degradation, membrane structure, and transcription regulation. The structural diversity of this large class of natural products are wholely derived from differences in only two, five-carbon subunits, dimethylallyl pyrophosphate (DMAPP) and isopentyl pyrophosphate (IPP). These two subunits,

DMAPP and IPP, are products of two independent pathways in nature called the mevalonate

(MVA) pathway in eukaryotes and the 2C-methyl-D-erythritol 4-phosphate pathway (MEP) in higher order eukaryotes and many bacteria.23,24,25,26

Figure 1-9. Representation of the biosynthesis of isoprenoid from the isoprene building blocks dimethylallyl pyrophosphate (DMAPP) and isopentyl pyrophosphate (IPP). Adapted from Tachibana.27

Two Distinct Isoprenoid Biosynthetic Pathways for Mammalians and non-Mammalians

Mammalian Isoprenoid Biosynthetic Pathway The mevalonate pathway (MVA) was first recognized in the late 1950s.28 Consisting of six enzymatic steps,29 the MVA pathway converts two acetyl Coenzyme A (Ac-CoA or acetyl-CoA) molecules to DMAPP and IPP (Figure 1-10).

In the first reaction, two molecules of acetyl-CoA are condensed by the enzyme acetoacetyl-CoA thiolase to give acetoacetyl-CoA. The product is then converted by HMG-CoA synthase to 3- hydroxy-3-methylglutaryl-CoA (HMG-CoA) by addition of another acetyl-CoA molecule.

Reduction of HMG-CoA is facilitated by HMG-CoA reductase (HMGR) to mevalonate, which then undergoes two phosphorylation steps by mevalonate kinase and phosphomevalonate kinase to produce mevalonate-5-diphopshate. In the final step, the intermediate mevalonate-5- diphosphate is decarboxylated to IPP, which can be converted to DMAPP by IPP isomerase.30

Besides mammals, the mevalonate pathway has also been identified in archaea, some eubacteria, fungi, and higher order plants.26

Figure 1-10. Biosynthesis of DMAPP and IPP in the mevalonate (MVA) pathway.

21 Discovery of the Non-mammalian Isoprenoid Biosynthetic Pathway -Until the early

1990’s, the MVA pathway was thought to be the single source of DMAPP and IPP biosynthesis until the discovery of the methylerythritol phosphate (MEP) pathway. 21,31,32 When studying triterpenoid biosynthesis in Rhodopseudomonas palustris bacteria, Rohmer, et al. found that the carbon isotope labeling distribution was not consistent with the expected labeling pattern for the

MVA pathway (Figure 1-11b), suggesting the existence of an alternate non-MVA isoprenoid pathway. 28,33,34 Similarly, Brechbuhler-Bader, et al., observed inconsistencies with 13C labeling patterns in plants.35 In 1991, Zhou and White found that supplementing bacterial cells with [13C]- glucose, which is metabolized by the glycolytic pathway (Figure 1-11a), incorporated the label into the isoprene chain of ubiquinone, a quinone-based analog (Figure 1-9). On the other hand, when bacteria were supplemented with unlabeled glucose and [13C]-acetate there was no labeling of the ubiquinone chain. When bacteria were supplemented with doubly labeled [4,5-13C]- glucose, conclusive evidence was found that isoprenoids in bacteria resulted in the condensation of C2 from pyruvate and C3 from G3P units, thereby explaining the observed labeling pattern

(Figure 1-11C).28,31,36Arioni, et al., found that plants also utilize the MEP pathway due to incorporation of [13C]-labeling in DMAPP and IPP using [13C]-deoxyxylulose.37

Over time, algae, bacteria, plants, and many other pathogenic organisms were found to exploit the MEP pathway. Two important examples are the highly infectious agents

Mycobacterium tuberculosis and Plasmodium falciparum, among other Apicomplexan parasites.

Due to mammalians exclusive use of the MVA pathway as a source of isoprenoid biosynthesis, the MEP pathway enzymes have been identified as promising targets in the development of novel anti-infective agents as well as herbicides.21,38,39

22 a) Glycolysis

b) MVA Pathway

c) MEP Pathway

Figure 1-11. The elucidation of the MEP pathway by feeding [13C]-labeled carbon sources to bacterial. a) [13C]-labeled glucose metabolism by glycolysis. b) Predicted integration of the [13C]-label by the MVA pathway. c) Observed integration of the [13C]-label by the MEP pathway.

Methylerythritol Phosphate (MEP) Pathway an Attractive Target for Developing Novel Antibiotics

The enzymes of the methylerythritol phosphate (MEP) pathway (also referred to as the non-mammalian isoprenoid biosynthetic pathway) (Figure 1-12), became the target of novel anti- infective agents in the early 1990’s.7 The MEP pathway presents a new class of targets as well as the potential for developing a new class of antibiotics. The MEP pathway utilizes seven enzymes in the biosynthesis of DMAPP and IPP that starts with pyruvate and D-glyceraldehyde 3- phosphate (D-G3P) to form DXP, which is catalyzed by 1-deoxy-D-xylulose 5-phosphate (DXP) synthase (Figure 1-12). DXP is then converted to 2C-methyl-D-erythritol 4-phosphate (MEP) by reductoisomerase IspC, which is the first committed step of the MEP pathway. MEP undergoes cytidylation by IspD with a cofactor of cytidine triphosphate (CTP) followed by phosphorylation by IspE with adenosine triphosphate to form 4-diphosphocytidyl-2C-methyl-D-erythritol 2- phosphate (CDP-ME2P). The fifth enzymatic step, which is the focus of this dissertation, catalyzes the cyclization of CDP-ME2P to 2C-methyl-D-erythritol 2,4-cyclodiphosphate

(MEcDP) with the release of a cytidine monophosphate (CMP). The enzyme IspG then performs a reductive ring opening of MEcDP to form (E)-4-hydroxy-3-methylbut-2-enyl diphosphate

(HMBDP) and the final enzymatic step is performed by a reductase, IspH, to generate the isoprene building blocks DMAPP and IPP.

Figure 1-12. The biosynthesis of DMAPP and IPP by the methylerythritol phosphate (MEP) pathway.

25 MEP pathway Regulation- The regulation of the MEP pathway in pathogens is not well known.22,40 Brown, et al., determined that overexpression of the gene dxs, which encodes DXP synthase, in Mycobacterium tuberculosis leads to accumulation of the downstream product

HMBDP of the MEP pathway via IspC (Dxr) and IspG (GcpE) upregulation. These results suggested DXP synthase is involved in transcriptional control, consequently it would control flux at the beginning of the pathway. Furthermore, the authors proposed IspG (GcpE) was the rate- limiting step of the pathway.37 IspD and IspF are expressed as a bifunctional IspDF enzyme in certain organisms such as Campylobactor jejuni, Mesorhizobium loti, and Agrobacterium tumefaciens.22,41,42,43 Multiple studies have focused on the bifunctional IspDF enzyme as a prospective regulatory mechanism in the biosynthesis of isoprenoids. It is known to catalyze the conversion of MEP to MEcDP in a number of bacterial species. In early reports, Rodriguez-

Concepcion proposed the IspDF organization of enzymes within the pathway could increase flux by substrate channeling through a complex of IspDF/IspE.39 Research by Lherbbert, et al., did not observe an increased metabolic flux through the IspDF complex, which suggests that the bifunctional enzyme serves in some other regulatory role.44

Known MEP Pathway Inhibitors- The MEP pathway has been validated as a target for the development of inhibitors as potential antibiotics. The first inhibitor discovered targeting DXP synthase was Fluoropyruvate (Figure 1-13, 1), a pyruvate mimetic, that after elimination of the fluoride ion during decarboxylation will generate a deactivated acetylated cofactor.45 The utilization of library screens identified arylpyrimidinone (2) as a DXP synthase inhibitor with an

46 IC50 value in the low micromolar range. The molecule ketoclomazone (3), which is a known herbicide, was also found to target DXP synthase.47 In 2012, Smith, et al., showed acylphosphonates could selectively inhibit Escherichia coli DXP synthase when they contain a

26 methyl-, ethyl-, or butyl- group at the phosphorus center with a inhibition constant, Ki, in the low micromolar range.48

The most potent inhibitor of the MEP pathway, fosmidomycin (Figure 1-10, compound

4), targets the second enzyme, IspC. Fosmidomycin and analog FR-900098 (compound 6), both phosphonate-containing compounds, are substrate mimetics that inhibited two malarial parasites,

Plasmodium falciparum and Plasmodium vinckei.49,50 In combination with clindamycin, a protein biosynthesis inhibitor, fosmidomycin was evaluated in Phase III clinical trials for treating malaria. The results from the clinical trial showed that the dual-compound treatment, while showing efficacy, was not a cure for malaria in all patients.51 Currently, Phase II clinical trials are being performed using fosmidomycin in combination with piperaquine as a non-artemisinin- based therapy for acute noncomplicated Plasmodium falciparum.52 Subsequent experiments have found that the phosphonate, hydroxamate, and the propyl linker were important for fosmidomycin activity.53 Therefore, this reduced potential active derivatives to compounds such as compound 5.42 Results for compounds containing an aryl ring at the -position, compound 8, had varied results due to the size.54

Developing inhibitors against the remainder of the enzymes within the MEP pathway has been difficult due to the relatively polar nature of their active sites, which possess pockets that electrostatically complement the substrate’s diphosphate groups. There have been two reported inhibitors of IspD with one being a substrate mimetic, desmethylerythritol (compound 8) that binds at the active site55 and another that was found to bind to an allosteric site of the enzyme

(compound 9), which is the most potent IspD inhibitor to date.56 IspE was also found to be sensitive to substrate analogs that contain a cytidine (compound 10).57 The most potent inhibitor of IspE, which is a kinase, was an alkyne-sulfonamide (compound 11) that contains a

27 tetrahydrothiophene substituent as a mimetic of the cytidyl ribose ring.45 Initial IspF inhibitors mimicked the CDP-ME2P substrate preserving the cytidyl moiety and replacing the methylerythritol portion with a fluorescent group (compound 12).58 Baumgartner, et al., found that modifying these inhibitors increased the IC50 into the upper micromolar range (compound

59 13). In 2010, the most potent IspF inhibitor was reported with an IC50 value in the low micromolar range (compound 14), which contained a thiazolopyrimidine core.60 Inhibitors of

IspG and IspH have also included the diphosphate functional group (compounds 15 & 16) as well as the bisphosphonate group (compound 17). The propylargyl diphosphate (compounds 15

& 16) actually showed activity against both IspG and IspH while the pyridyl analog (compound

17) was only active against IspH.45

Figure 1-13. Substrates of MEP pathway enzymes and their known inhibitors (Adapted from Hale (et al.).45

29 Targeting Cyclodiphosphate Synthase (IspF) IspF is of great interest as a target to disrupt the biosynthesis of isoprenoids since it has been shown to be essential in multiple pathogenic organisms such as Mycobacterium tuberculosis,61 Escherichia coli, Bacillus subtilis.62 Beutow, et. al,61 attempted an IspF knockout mutant in M. tuberculosis but was only successful in isolating the mutant when a functional copy of the gene was co-introduced. In another study, ispF and yacN (IspF homolog) genes were deleted in E. coli and B. subtilis, which resulted in altered growth rates and morphology.62 The results of these studies determined that the ispF gene is essential in M. tuberculosis, E. coli, and

B. subtilis.

Figure 1-14. Proposed IspF mechanism displaying the substrate position and charge stabilization by Mg2+ and Zn2+.65

Figure 1-15. Ribbon representation of Escherichia coli IspF crystal structure. Highlighted is one of the three active sites located at the monomer-monomer interface. IspF co-crystallized with the products MEcDP, CMP, and the metal co-factors (PDB: 1H48 and 1U3L).63,64 IspF has also been shown to bind downstream isoprenoid intermediates in the hydrophobic cavity such as geranyl pyrophosphate (GPP). 63

31 Active Site of Cyclodiphosphate synthase (IspF)- IspF has been proposed to be physiologically active as a trimer with three active sites located between the three symmetrical monomer-monomer interfaces (Figure 1-15).65 It also requires two metal co-factors, Zn2+ and

Mg2+, which stabilize the charge of its pentavalent transition state (Figure 1-14).26,64,66,67,68,69

The crystal structure of E. coli IspF reveals three active site sub-binding pockets, named Pocket I,

II, and III, which bind the ribosyl group, methylerythritol group, and the cytidyl nucleotide base, respectively (Figure 1-16).58,64 Pocket I encloses the riboxyl-5’-diphosphate moiety and has a conserved Asp46 at the bottom of the active site, it also has no direct interaction with the substrate. Pocket II was shown to contain a flexible loop (residues Pro62 to Ala71) that becomes well-ordered upon binding the substrate MEcDP. Furthermore, the pocket contains hydrophobic residues Ile57 and Leu76, as well as hydrophilic residues Ser35, Ser73, and Asp63 that orients the C-methyl-D-erythritol phosphate moiety for attack by the -phosphoryl group.

Pocket III was observed to bind the cytidine group. The pocket is formed by Ala131, Thr133,

Lys104, and Leu106. Additionally, Pocket III binds the nucleobase with four hydrogen bonds provided by residues Ala100 to Leu106 (Figure 1-17).64 As shown previously, the Mg2+ and Zn2+ ions are important for catalyzing the IspF reaction. In particular, the Zn2+ ion coordinates to

Asp8, His10, and His42 and attacks the -phosphoryl group, which coordinates the ion tetrahedrally. The Mg2+ ion holds the - and -phosphates of CDP-ME2P to prepare for the 2- phosphoryl attack on the -phosphoryl center, which thereby stabilizes the negative charge that develops on the CMP.64

Figure 1-16. Sub-binding pockets within the active site of Escherichia coli IspF (PDB: 1H48)63

Figure 1-17. Active site diagram of Escherichia coli IspF highlighting cytosine-enzyme interactions. Cytosine ring has four hydrogen bonds with the backbone of Ala102 to Leu106, Pocket III. (PDB: 1H48)63

34 Cyclodiphosphate synthase (IspF) Hydrophobic Cavity-Structural studies have shown he presence of a hydrophobic cavity in the channel of the IspF homotrimer (Figure 1-14),63,65,66,67 which has been observed to accommodate downstream isoprenoids. Richard, et al. found that the polar diphosphate of geranyl diphosphate (GDP) interacts with an arginine “cap” within the hydrophobic cavity that was first reported in 2004.63,65 The arginine “cap”, which consists of three arginines, was shown to coordinate to a sulfate ion when downstream isoprenoids were not present.65 Further structural studies by Kemp, et al.67 and Ni, et al66 showed that farnesyl diphosphate (FDP) could also be trapped in the hydrophobic cavity. 31P NMR and mass spectrometry studies showed that GDP predominately binds in the cavity.67 Kemp, et al. also found that the downstream isoprenoids IDP, GDP, and FDP bound to EcIspF with a ratio of

1:4:2,but this finding was not confirmed with structural studies.67 Additionally, crystallographic studies on a double mutant of the enzyme, R142M and E144L, disrupted six hydrogen bonds to the diphosphate moiety but maintained binding of the isoprenoid diphosphate group. This indicates that the hydrophobic interactions with the ligand dominate binding of the downstream isoprenoids.61 Overall, this suggests the potential for feedback regulation of the MEP pathway by inhibiting IspF with downstream isoprenoids.70

Targeting Burkholderia pseudomallei IspF- In 2004, Eberl, et al. used a TBLASTN search of E. coli homologues which identified that Burkholderia pseudomallei utilizes the MEP pathway.71 Studies have shown that the IspF product, 2C-methyl-D-erythritol 2,4- cyclodiphosphate (MEcDP), accumulates to high levels in a variety of organisms as a response to oxidative stress.72,73,74,75,76,77,78 This implicates MEcDP could act as an anti-stressor through metal-dependent enzyme regulation by chelation of cations or antioxidant regulation.

Goncharenko, et al. found that adding MEcDP to a “nonculturable” form of Mycobacterium

35 smegmatis would reactivate growth in vitro, which suggests it may play a role in the transition to and from latency79 involving chromatin condensation-decondensation regulation. This result coupled with the fact that IspF binds downstream isoprenoids suggest cyclodiphosphate synthase

IspF could regulate the MEP pathway.

The goal of this research project was to develop and implement assays to study small molecule inhibitors on the fifth enzyme in the MEP pathway, IspF, from Burkholderia pseudomallei in an effort to identify lead molecules as potential anti-infective agents. This dissertation will describe the development of a high-throughput thermal shift assay to assess relative binding potencies of potential IspF inhibitors (Chapter 2). The most potent compounds were analyzed using Surface Plasmon Resonance (SPR), Isothermal Titration Calorimetry (ITC), and Fluorescence Polarization (FP) to determine dissociation constants and the binding energetics (Chapter 3). The most promising compounds were tested for their ability to inhibit

Burkholderia pseudomallei IspF using a HPLC-based enzymatic assay (Chapter 4). Finally, X- ray crystallography was used to study the mode of inhibitor binding and future inhibitor design

(Chapter 5).

CHAPTER 2

DEVELOPMENT OF A HIGH-THROUGHPUT THERMAL SHIFT ASSAY TO ASSESS

THE RELATIVE AFFINITY OF BURKHOLDERIA PSEUDOMALLEI ISPF

INHIBITORS

In drug discovery, it is imperative to have a method to assess the ability of compounds to bind or inhibit the target. This step is crucial to help guide the next step in molecular design for increased affinity, selectivity, and specificity. A high-throughput assay is preferred when screening a large library of compounds to rapidly identify compounds that are the most potent.

Chemical libraries typically consist of thousands to hundreds of thousands of compounds with diverse chemical structures that can easily be modified for synthesis of analogs.

The current methods reported in the literature to assess IspF activity do not have a direct readout of the enzymatic product due to the absence of a chromophore in 2-C-methyl-D- erythritol 2,4-cyclodiphosphate. Consequently, established IspF enzyme inhibition assays are either highly complex and/or labor-intensive. In 2006, Rohdich reported the first IspF assay which consisted of a plate-based assay using three additional enzymes that allows monitoring the consumption of nicotinamide adenine dinucleotide (NADH) (Figure 2-1).80 An alternative IspF assay was reported by Freel-Meyers in 2012, which utilizes an HPLC method to monitors the formation of the by-product, cytidine monophosphate (Figure 2-2).81 The HPLC method is a time-intensive

37 assay that requires 36 hours to acquire an 8-point dose-response curve. In a third approach, reported in 2011, the product CMP is used to generate ADP in the presence of ATP and nucleotide monophosphate kinase. Through the use of three enzymes, ADP can then be converted to resorufin, which fluoresces at 590 nm (Figure 2-3).82 Consequently, the development of a simple and rapid method to directly assess the relative affinity of compounds in a high throughput manner would be advantageous in screening potential IspF inhibitors.

Figure 2-1. Plate-based IspF assay by coupling cytidine monophosphate to monitor NADH consumption.81

CDP-ME2P MEcPP CMP

Figure 2-2. HPLC method to monitor IspF activity using the cytidine chromophore of the substrate and cytidine monophosphate that is released.82

Figure 2-3. Plate-based assay converting CMP to ADP, which is then converted to resorufin for detection.83

Thermal-shift assays have been used successfully in drug discovery efforts.83,84,85,86 These assays monitor the unfolding of the target enzyme in the absence and presence of inhibitors to identify compounds that result in a significant shift in the enzyme’s melting temperature (e.g., ≥4

º C), which indicates the compounds are stabilizing the native state of the enzyme. Analyzing all compounds at the same concentration can provide a rank order of potency, as the degree of shift is proportional to the affinities of the compounds for the enzyme. As B. pseudomallei IspF was found to be extremely thermostable, with a melting temperature >100°C at physiological pH, a thermal shift analysis was not possible. In this study, enzyme engineering was utilized to develop

39 a high-throughput IspF thermal shift assay by destabilizing the enzyme to lower the temperature at which it unfolds. Furthermore, similar to naturally observed destabilized enzymes, 87,88 the identified IspF variant Q151E possessed decreased stability without loss of enzyme activity.

Methods

Expression and Purification of Burkholderia pseudomallei IspF

An IspF expression vector, IspF-pET14b (Novagen), which possesses an N-terminal hexa-His tag before the IspF gene, was obtained from the SSGCID. Starting from a single colony, a 5-mL (LB- 100 µg/mL ampicillin) culture of E. coli BL21(DE3) containing the pET14b-His-3C-IspF89 was grown overnight at 37°C. The overnight culture, 500 µL, was used to start a 50mL subculture that was used to inoculate a 1 L (LB-100 µg/mL ampicillin) culture once

° mid-log phase was reached. Cells were grown at 37 C until the OD600 reached 0.5-0.8. The culture was induced with isopropyl �--D-thiogalactoside (IPTG, 1 mM) and allowed to shake overnight (20 C, 235 rpm). The cells were harvested by centrifugation (8000 rpm, 20 min) and stored at -20C.

The frozen pellet was resuspended in 10 mM TRIS pH 8.0 and sonicated for six cycles of

20 sec on/20 sec off with a Branson 450 digital sonicator followed by centrifugation at 4°C (20 min, 15,000 rpm) to pellet insoluble protein and cell debris. The soluble fraction was loaded onto an immobilized Ni2+ affinity column (Histrap HP, GE Healthcare). The His-tagged IspF was eluted using a 70-mL linear gradient from 20 mM to 500 mM imidazole (in 50 mM sodium phosphate, 500 mM NaCl, and 1 mM DTT). The single peak corresponding to IspF was further 40 purified on a HiLoad 26/60 Superdex 75 prep grade FPLC size exclusion column (20 mM TRIS and 150 mM NaCl).90

Table 2-1 DNA and Protein Sequence of Burkholderia pseudomallei IspF

DNA Sequence - ATGGCTCATCACCATCACCATCATATGGGTACCCTGGAAGCTCAGACCCAGGGTCCTGGTTCGATGGATT TCAGAATCGGACAAGGCTACGACGTGCACCAGCTCGTTCCCGGACGCCCGCTCATCATCGGCGGCGTGA CGATCCCTTACGAGCGGGGGCTGCTCGGCCACTCGGACGCCGACGTGCTGCTGCACGCGATCACCGAC GCGCTCTTCGGCGCGGCGGCGCTCGGCGACATCGGCCGCCACTTCTCCGACACCGATCCGCGCTTCAAG GGCGCCGACAGCCGCGCGCTGTTGCGCGAATGCGCATCGCGCGTCGCGCAGGCGGGTTTTGCGATCCG CAACGTCGACAGCACGATCATCGCGCAGGCGCCGAAGCTCGCGCCGCATATCGACGCGATGCGCGCGA ACATCGCGGCCGATCTCGACTTGCCGCTCGATCGCGTGAACGTGAAGGCGAAGACGAACGAGAAGCTC GGCTATCTCGGGCGCGGTGAGGGGATCGAGGCGCAGGCGGCCGCGCTCGTCGTGCGCGAAGCGGCCG CGTGAAAACAGCACGAACAAGTTCTGCA Protein Sequence - MAHHHHHHMGTLEAQTQGPGSMDFRIGQGYDVHQLVPGRPLIIGGVTIPYERGLLGHSDADVLLHAITDA LFGAAALGDIGRHFSDTDPRFKGADSRALLRECASRVAQAGFAIRNVDSTIIAQAPKLAPHIDAMRANIAADL DLPLDRVNVKAKTNEKLGYLGRGEGIEAQAAALVVREAAA

Circular Dichroism

All experiments were performed on an Aviv Instruments circular dichroism spectrometer model 215. Wavelength scans were carried out at 25oC from 250 to 200 nm in phosphate- buffered saline (PBS) (10 mM sodium phosphate, 150 mM sodium chloride, pH 7.4) or a chemical denaturant phosphate buffer (10 mM sodium phosphate, 150 mM sodium chloride, 4M guanidine hydrochloride). A quartz cuvette was used with a 1-mm path length. IspF temperature denaturation experiments were monitored at 222 nm over a temperature range of 25 to 100°C at

o 1 C/min, 2 nm bandwidth, and 2 second averaging time. The melting temperatures, Tm values, were determined using a Boltzmann fit with Origin Labs.

Solvent Accessible Surface Area

The high resolution crystal structure of Burkholderia pseudomallei IspF (PDB: 3QHD)90 was used to calculate solvent accessible surface areas with the NACCESS program91 using a water sphere with a 1.4 radius and a slice width of 0.25 . An in-house Fortran program92 was used to determine the change in solvent accessible surface upon folding (ASA). The ASA value was partitioned into the change in polar (ΔASApol) and apolar (ΔASAap) surface area upon folding.

Computational Stability Measurements

The program FoldX allows the assessment of the effect single and multiple mutations may have on protein stability, folding, and dynamics.93 Here, stability calculations, performed using FoldX, were used to identify residue substitutions that may destabilize IspF. Specifically, two approaches were used: (1) isosteric introduction of charged residue (either Asp or Glu) at the interface of two monomers and (2) substitution of bulky hydrophobic residues with alanine. Both strategies are aimed at destabilizing the favorable packing interactions that lead to IspF’s high thermal stability. Twelve single point mutations were identified based on the two criteria: I5A,

Q7E, Q13E, Q88E, N95D, N117D, N130D, N136D, L139A, R144A, Q151E, and L155A.

Before analysis using FoldX in Yasara, RepairPDB was run on a representative crystal structure of IspF (PDB: 3MBM)90 to optimize residues that have high energy torsion angles or clashes.

42 FoldX was used to calculate the predicted change in stability (ΔΔG°) for the twelve mutants based on the following equation:

; Eq. 2-1 where ΔΔG° values less than zero are predicted to be stabilizing mutants and those values greater than zero are predicted to be destabilizing mutants.93

Site-Directed Mutagenesis of Burkholderia pseudomallei IspF

Starting with a modified pET14b vector containing Burkholderia pseudomallei IspF, site- directed mutagenesis was carried out to create single point mutations. A total of 50 ng of template dsDNA was mixed with 10X reaction buffer, forward oligonucleotide (125 ng/µL), reverse oligonucleotide (125 ng/µL), 10 mM dNTP, 2% v/v DMSO, and PfuTurbo DNA polymerase (2.5 U/µL). Site-directed mutagenesis was performed using a Bio-rad iQ5 Real-Time

PCR with one cycle at 95oC for 30 seconds, which was followed by a second cycle at 95oC for

30 seconds, 55 oC for 60 sec, and 68 oC for 6 minutes, repeated sixteen times. DPN-1 (20 U/µL),

1 µL, was added to the reaction tube and incubated at 37oC for 1 hour. The DPN-1 digested PCR product was then transformed into XL-1 Blue competent cells using standard methods. Single colonies were picked and grown overnight in 5 mL LB/Amp at 37 oC in a tabletop incubator shaker. DNA was isolated using the Promega Wizard Plus SV Miniprep DNA Purification

System and sent for sequence verification at the University of Chicago Comprehensive Cancer

Center DNA Sequencing and Genotyping Facility.

43 Table 2-2. Primers used for site-directed mutagenesis of Burkholderia pseudomallei IspF

Primer Sequence (5’→3’) I5A Forward GGATTTCAGAGCGGGACAAGGCTACGACGTGCACCAGCTCGTTCC I5A Reverse GGAACGAGCTGGTGCACGTCGTAGCCTTGTCCCGCTCTGAAATCC Q7E Forward GATTTCAGAATCGGAGAAGGCTACGACGTGC Q7E Reverse GCACGTCGTAGCCTTCTCCGATTCTGAAATC Q13E Forward GCTACGACGTGCACGAACTCGTTCCCGGACG Q13E Reverse CGTCCGGGAACGAGTTCGTGCACGTCGTAGC Q88E Forward ATCGCGCGTCGCGGAAGCGGGTTTTGCGA Q88E Reverse TCGCAAAACCCGCTTCCGCGACGCGCGAT N95D Forward GTTTTGCGATCCGCGACGTCGACAGCACG N95D Reverse CGTGCTGTCGACGTCGCGGATCGCAAAAC N117D Forward CGATGCGCGCGGACATCGCGGCC N117D Reverse GGCCGCGATGTCCGCGCGCATCG N130D Forward CGCTCGATCGCGTGGACGTGAAGGCGAAG N130D Reverse CTTCGCCTTCACGTCCACGCGATCGAGCG N136D Forward AAGGCGAAGACGGACGAGAAGCTCGGCTAT N136D Reverse ATAGCCGAGCTTCTCGTCCGTCTTCGCCTT L139A Forward CGAACGAGAAGGCGGGCTATCTCGGGCGCGG L139A Reverse CCGCGCCCGAGATAGCCCGCCTTCTCGTTCG R144A Forward GCTCGGCTATCTCGGGGCAGGTGAGGGGATCG R144A Reverse CGATCCCCTCACCTGCCCCGAGATAGCCGAGC Q151E Forward GGGATCGAGGCGGAAGCGGCCGCGC Q151E Reverse GCGCGGCCGCTTCCGCCTCGATCCC L155A Forward GGCGGCCGCGGCGGTCGTGCGCGAAGC L155A Reverse GCTTCGCGCACGACCGCCGCGGCCGCC

Thermal Shift Assay

Small molecule screening by fluorescence based thermal unfolding was performed using a Bio-Rad iQ5 Real-Time PCR equipped with a 96-well PCR plate. Purified Burkholderia pseudomallei IspF mutants were diluted to the appropriate concentration with a buffer consisting of 100 mM Glycine-NaOH, pH 9.0, 100 mM NaCl, and 100 mM NaI. All assay experiments contained 5 µg enzyme per well and 4X SYPRO orange dye with a final volume of 25 µL, resulting in an enzyme concentration of 4 µM. Compounds were dissolved at 10 mM in DMSO and serial diluted into the assay buffer to a final concentration of 500 µM with 25% DMSO, with a resultant final concentration of 100 µM per well. The PCR plates were sealed with optical seal, shaken, and centrifuged after enzyme and compound were added. Thermal scanning (25-95oC at

1.0oC/min) was performed and fluorescence intensity (ex. 530 nm em. 570 nm) measured every

10 seconds in 0.5oC increments. The melting temperature was determined using the first derivative of the raw fluorescence intensity to identify the peak minimum.

Preparation of 4-diphosphocytidyL-2C-methyl-D-erythritol 2-phosphate (CDP-ME2P)

CDP-ME2P was prepared following a modified procedure of Bitok et al.81 A mixture of

MEP (5 mM), CTP (5 mM), E. coli IspD (2 µM), inorganic pyrophosphates (7 units), BSA (1 mg/mL) in 50 mM HEPES (pH 7.4) was incubated at room temperature for 128 minutes to facilitate the conversion of CTP (2.38 min) to CDP-ME (1.56 min). The reaction was monitored at 272 nm using a Varian Prostar reverse phase HPLC with ion-pairing reagent tetrabutyl

45 ammonium. The column, Altima C18 3 µ, 53 mm x 7 mm Rocket column, was run using a gradient of 0-33.5% Buffer B at a flow rate of 1.5 mL/min to elute the analytes of interest

(Buffer A = 100 mM phosphate buffer, 5 mM tetrabutyl ammonium bisulfate, pH 6.0 and Buffer

B = 100 mM phosphate buffer, 5 mM tetrabutyl ammonium bisulfate in 30% acetonitrile). Once the IspD reaction was complete, phosphonenol pyruvate (7 mM), pyruvate kinase (3.5 units), and

LeIspE (2 µM) was added to the reaction mixture with continued monitoring by HPLC for CDP-

ME2P formation (CDP-ME2P = 2.47 min) after 182 minutes.81

HPLC-based IspF Activity

IspF reaction mixtures contained 50 mM phosphate buffer, pH 7.4, 5 mM MgCl2, 100

µM CDP-ME2P, 2 nM BpIspF and 50 µg/mL BSA in a total volume of 250 µL. All reactions were performed at 25 oC. The reaction, initiated by the addition of IspF, was terminated at 2, 4, 6, and 10 min by addition of 40 µL of reaction mixture to 80 µL of cold 0.1% SDS. The quenched reaction mixture was vortexed for 5 seconds. Before HPLC analysis proteins were removed using a 3 kDa MWCO VWR Polyethersulfone Membrane centrifugal filter. Using a Varian

Prostar HPLC with PEEK tubing, samples (100 µL) were injected into a reversed-phase ion-pair

HPLC with an Altima C18 3 µ, 53 mm x 7 mm Rocket column. The column used a gradient of

0-33.5% Buffer B at a flow rate of 1.5 mL/min to elute the analytes of interest (Buffer A = 100 mM phosphate buffer, 5 mM tetrabutyl ammonium bisulfate, pH 6.0 and Buffer B = 100 mM phosphate buffer, 5 mM tetrabutyl ammonium bisulfate in 30% acetonitrile). The peak areas of

46 CMP (1.24 min) and CDP-ME2P (2.47 min) were determined and the concentration of CMP was calculated as a fraction of the total peak area.81

Crystallization of Burkholderia pseudomallei IspF Q151E

Purified BpIspF (13 mg/mL) was run on a Superdex 75 10/30GL size exclusion column

(GE Healthcare) using a 20 mM HEPES pH 7.0, 150 mM NaCl, 2 mM DTT running buffer.

Fractions corresponding to a single peak representing trimeric IspF were pooled and concentrated to 32 mg/mL using an Amicon ultrafiltration device with a 3 kDa cut-off membrane. Protein crystals were grown using the hanging drop vapor diffusion method. The trays were incubated at 19oC and BpIspF crystals were observed in three days. Initial crystallization drops were composed of 1.0 µL protein solution (32 mg/mL of BpIspF in SE buffer) and 1.0 µL crystallization buffer (100 mM Tris, pH 7.2, 200 mM NaCl, 20% PEG 4000,

5 µM ZnCl2 or 200 mM C2H2MgO4∙2H2O, pH 5.9, 20% PEG 3350). The crystallization tray wells contained 1000 µL of crystallization buffer.90

X-Ray Diffraction

Crystals were loop mounted and flash frozen in a glycerol containing cryoprotectant solution and loop mounted using standard methods.94 Briefly, for each crystal growth condition,

2X-concentrated cryoprotectant stocks were prepared by mixing 100% glycerol with crystal tray well solution to a final glycerol concentration of 40% (v/v). Prior to crystal mounting, 0.5 µL of the 2X-concentrated cryoprotectant stock was diluted two-fold by adding an equal volume of crystal growth drop solution. Crystals were then immediately transferred from the growth drop

47 using an appropriately sized LithoLoop mount (Molecular Dimensions) and briefly washed in the cryoprotectant solution. Crystals were then re-looped, flash frozen by immersion in a liquid nitrogen bath, and seated in the appropriate wells of a Unipucks (MiTeGen), which had been chilled in a liquid nitrogen bath. Pucks containing the samples were transferred to the beamline and samples were mounted for data collection using the robotic automounter provided at the beamline.

Data were collected at Southeast Regional Collaborative Access Team (SER-CAT) 22-

BM beamline95 at the Advanced Photon Source, Argonne National Laboratory.96 The beamline energy was tuned to 12.398eV (1.000Å). A total of 180-degrees of data were collected from each sample, using a 1-second per 1-degree oscillation range. Data were collected at a distance of 140 mm on a MAR 225 CCD detector (Rayonix LLC, Evanston IL, USA). Data were processed using HKL2000.97

X-Ray Structure Solution

All IspF structures in this study were determined using molecular replacement. Molecular replacement was performed using Phaser98 with Burkholderia pseudomallei IspF (PDB ID:

3QHD)90 as the input model. One or two IspF molecules were observed in the asymmetric unit, dependent upon the crystal form (form 1 or 2: C2 with a = 157.7 or 116.9, b = 69.4 or 68.1, c =

116.4 or 60.4). The Phenix program suite was used for all refinements.99 The ligand, along with residues that were not modeled in the input structure, were built into observed electron density

48 using iterative rounds of refinement and manual model building with Phenix, Refmac, and

Coot.18,100,101

Results and Discussion

Purification of Burkholderia pseudomallei IspF

The wild-type and mutant IspF enzymes were purified using a Ni2+ affinity column with a gradient elution from 50 mM to 250 mM imidazole (Figure 2-4). Fractions containing protein were collected, from 25-35 minutes, and concentrated to a volume of ten milliliters using a 3000

MWCO spin concentrator. Further purification was performed using a Sephadex-75 prep grade size exclusion chromatography (SEC) column. Fractions containing protein were collected from

90-100 minutes (Figure 2-5). The purified protein was stored at -80°C and aliquots removed as needed for circular dichroism analysis, the thermal-shift assay, and the HPLC activity assay.

Figure 2-4. Chromatogram of Ni2+ affinity purification of Burkholderia pseudomallei IspF.

Figure 2-5. Size exclusion chromatogram of Burkholderia pseudomallei IspF purification.

51 Determination of Burkholderia pseudomallei IspF Melting Temperature

The thermal stability of the enzyme was evaluated using circular dichroism at an IspF concentration of 10 µM in an effort to develop a thermal-shift assay to assess the relative affinities of inhibitors. Initially, a far-UV scan was performed from 250-200 nm. The analysis confirmed the expected alpha helical and beta sheet secondary structure of the enzyme (Figure 2-

6).102 Surprisingly, a thermal melt did not show any loss of secondary structure up to 100°C

(Figure 2-7). Subsequently, to promote unfolding, the enzyme was incubated with a chemical denaturant, guanidine hydrochloride. In the presence of 4 M guanidine hydrochloride the melting temperature was determined to be approximately 92°C (Figure 2-7). The exact melting temperature could not be determined due to an incomplete unfolding transition. Despite the addition of a high concentration of chemical denaturant, the enzyme maintained a high thermal stability. This result indicated that such an approach would not be viable for a useful thermal- shift assay to assess IspF inhibitors. For this technique to work, at least two conditions must be met. First, the range of temperature over which thermal denaturation is observed must be sufficiently broad to allow clear identification of the temperature of the unfolding event. Second, the melting temperature cannot be too close to the maximum temperature for the technique

(100°C). The native IspF, even in the presence of the chemical denaturant guanidine hydrochloride, violates both requirements.

Figure 2-6. Wavelength scan of trimeric Burkholderia pseudomallei IspF displaying alpha helical and beta sheet characteristics.

Figure 2-7 Thermal unfolding of wild-type enzyme in the presence (open squares) and absence (filled triangles) of 4 M guanidine HCl.

54 Identification of Burkholderia pseudomallei IspF Destabilizing Mutants

Due to the extreme thermostability of BpIspF, an alternative approach was devised to develop a high throughput assay for screening inhibitors. Identification of amino acid residues that could be mutated to affect thermal stability of the enzyme was analyzed. In order to generate useful thermal destabilizing mutants of the enzyme, amino acid residues that do not affect enzyme activity had to be identified, which likely involved residues distant from the active site.

Two approaches were used to engineer a destabilized IspF with the disruption of monomer- monomer packing interactions and the introduction of charged residues. The first approach analyzed each IspF residue for both apolar and polar surface burial (Table 2-3). Mutagenesis of a residue that is found to have a large apolar surface area to a much smaller amino acid, such as alanine, reduces the hydrophobic interactions that drive packing efficiency and will likely destabilize the native state of the enzyme. Using this protocol, three hydrophobic residues within the subunit interfaces were calculated to have a significant loss in apolar surface area that upon mutation to alanine that could lead to destabilization: I5A, L139A, and L155A (Figure 2-8).

Additionally, the arginine “cap” is in close proximity to a sulfate ion located within the channel of the homotrimer (Figure 5-2). Due to its close proximity to the sulfate ion, it was hypothesized that mutating R144 would eliminate binding and destabilize the enzyme. The second approach involves introducing buried charged residues to create an energetic penalty. Aspartate and glutamate mutations were substituted for isosteric asparagine and glutamine residues, respectively. The specific mutations made were at positions: Q7E, Q13E, Q88E, N95D, N117D,

N130D, and Q151E. Based on the crystal structure of IspF (PDB: 3QHD)90 the positions for

55 mutagenesis were found to be 12-25 Å from the active site (Figure 2-8). Next, FoldX was used to calculate ΔΔG values for destabilization (Figure 2-9). It is important to note, these values represent one-third of the predicted stabilization or destabilization effect due to the trimeric nature of the IspF structure. A total of eleven residues were predicted to have a destabilizing effect (ΔΔGfold>0) based upon their mutation to alanine (approach 1) or to an isosteric, charged residue (approach 2). Mutagenesis of Arg144 to an alanine was the only change predicted to have an overall stabilizing effect (Figure 2-9).

56 Table 2-3. Analysis of surface burial of individual residues of the Burkholderia pseudomallei IspF enzyme. (PDB: 3QHD)90 res# res ap tot pol tot res# res ap tot pol tot 1 MET 4.23 55.09 99 THR 3.51 13.27 2 ASP 12.96 12.41 101 ILE 16.09 0 3 PHE 66.32 20.11 108 ALA 8.1 7.41 4 ARG 36.16 44.91 111 ILE 22.8 0 5 ILE 110.87 12.87 112 ASP 4.41 19.65 6 GLY 8.61 2.51 115 ARG 0 18.69 7 GLN 44.25 67.74 127 ASP 1.69 39.12 8 GLY 0.75 1.41 128 ARG 15.16 40.69 9 TYR 81.61 17.02 129 VAL 0.88 0.69 10 ASP 5.64 15.19 130 ASN 17.64 36.87 11 VAL 41.75 12.84 131 VAL 0 12.46 12 HIS 20.2 7.36 132 LYS 83.5 1.78 13 GLN 28.52 32.91 133 ALA 1.82 17.3 34 LEU 5.62 4.86 134 LYS 57.7 38.37 35 GLY 4.98 0 135 THR 0 4.05 44 HIS 0.08 0 136 ASN 4.67 37.3 48 ASP 10.37 29.19 137 GLU 47.87 24.48 52 GLY 18.26 15.73 138 LYS 12.63 0 53 ALA 0.62 3.95 139 LEU 108.25 0 54 ALA 0 0.95 140 GLY 16.89 3.06 55 ALA 52.94 7.14 141 TYR 62.44 4.15 57 GLY 20.99 10.94 144 ARG 0 2.11 58 ASP 20.43 33.35 146 GLU 4.61 7.29 59 ILE 7.74 0 151 GLN 7.13 41.91 61 ARG 0 9.63 152 ALA 3.35 0.96 94 ARG 1.6 25.91 153 ALA 42.84 0.69 95 ASN 16.44 39.52 154 ALA 1.99 1.83 96 VAL 0 0.18 155 LEU 46.4 0 97 ASP 17.12 28 158 ARG 0 13.74 98 SER 3.52 0 - - - -

Figure 2-8. Ribbon representation of Burkholderia pseudomallei IspF homotrimer highlights interdomain mutation designed to decrease protein stability (magenta). The homotrimer contains three subunits that are designated by the following color scheme A (cyan), subunit B (pale cyan), and subunit C (green cyan). All three monomers contain a single zinc ion (blue spheres). The B subunit zinc ion is not visible in the picture rotation.

Figure 2-9. FoldX stability predictions for single point mutations.

59 Melting Temperature Determination of Burkholderia pseudomallei IspF Mutants

The twelve single point IspF mutants were generated by site-directed mutagenesis and then expressed and purified according to the same procedure that was used for the wild-type

IspF. The mutants were then analyzed to determine the melting temperature under three separate conditions (Figure 2-10). The first condition studied was physiological pH (7.4) where a total of three mutants were found to have a destabilizing effects. All three destabilization mutants involved the incorporation of a charged residue that decreased the melting temperature by 13-

31°C: N95D, N130D, and Q151E. The next condition examined was pH 9.0, which was predicted to introduce additional folding penalties for the acid residue substitution by favoring generation of a formal charge on the mutant residues: Q7E, Q13E, Q88E, N95D, N117D,

N130D, N136D, and Q151E. The three mutants that were destabilized at physiological pH were destabilized even further (by 6-22°C). One additional mutant, Q7E, was also destabilized at this pH (by 20°C). In an attempt to lower the melting temperature further, several alternative chemical denaturants and salts to guanidine hydrochloride were analyzed in the presence of all

BpIspF single point mutations. Of note, sodium iodide displayed a destabilization effect on all variants and wild-type enzyme. The iodine molecule is larger than a chlorine and therefore can disrupt interactions between hydrophobic residues.103 The disruption of those interactions can further destabilize the enzyme to a temperature that can aid in the rank ordering of inhibitor binding using this type of assay.5 Overall, the addition of the chemical denaturant NaI showed a significant destabilization effect on all single point mutants, by 12-34°C. In addition, NaI acted as a better chemical denaturant on the wild-type enzyme in comparison to guanidine

60 hydrochloride. The N130D and Q151E showed the most significant destabilization of the enzyme at >53°C and >36°C, respectively. To ensure the mutants were binding an inhibitor similarly to the wild-type IspF, the two mutants were initially screened with a control inhibitor,

L-tryptophan hydroxamate. The N130D mutant was determined not to bind the compound and was therefore not used in the thermal-shift assay. The Q151E mutant, however, displayed the same binding affinity for the compound as the wild-type IspF and could therefore be used in the thermal-shift assay to analyze potential IspF inhibitors.

Figure 2-10. Melting temperatures analysis of Burkholderia pseudomallei IspF mutants under various conditions. (* indicates the melting temperature could not be detected and is >100°C)

62 Enzymatic Activity of Burkholderia pseudomallei IspF

For this study, enzyme activity was a mandatory criterion in identifying usable destabilizing mutations. Although the twelve identified mutations involved changes that were distant from the active site, this did not guarantee that the mutants would maintain native IspF activity, so this had to be tested. Consequently, the Q151E variant, which was destabilized by

>36°C and bound the control compound with similar affinity as wild-type enzyme, was examined for activity using the HPLC assay. Enzymatic activity was analyzed to determine the rates of both the wild-type and Q151E variant. The rate of the wild-type was measured as 4.3 ±

0.3 µmol/min and the mutant was measured to be 6.0 ± 1.0 µmol/min (Figure 2-11). Overall, the rates of both the wild-type and mutant enzyme were found to be similar indicating that the active site of the enzyme was not affected by mutating glutamine to glutamate at position 151.

Figure 2-11. Enzymatic reaction rate of Burkholderia pseudomallei IspF Q151E mutant. The Q151E mutant was determined to have a rate of 6.0 ± 1.0 µmol/min (blue squares) in comparison to native IspF (green squares).

64 Structure of Burkholderia pseudomallei IspF Mutant Q151E

Structural analysis using X-ray crystallography confirmed the Q151E variant did not significantly affect the structure of the IspF enzyme. An overlay of the structures for the wild- type and Q151E mutant of IspF showed that the quaternary structure of the enzyme was not affected by incorporation of a charged residue within each monomer-monomer interface (Figure

2-12). In addition, Q151E displayed only minimal differences in side chain conformation relative to the wild-type (Figure 2-13). Finally, examination of the wild-type versus mutant active sites verifies that the close proximity of the charged residue to the active site did not modify the binding pocket (Figure 2-14).

Figure 2-12. Overlay of Burkholderia pseudomallei IspF wild-type and Q151E mutant. The wild- type structure (PDB: 3F0D)90 is shown in cyan and the crystal structure determined for the Q151E mutant in this study is displayed in magenta. (RMSD: 0.152 Å)

Figure 2-13. Side chain comparison of the Q151E mutant (magenta) shows small differences structurally due to incorporating a charged residue. (RMSD: 0.3 Å)

Figure 2-14. Active site overlay of Burkholderia pseudomallei IspF wild-type and Q151E mutant confirms active site was unaffected by mutagenesis. The IspF structure (PDB: 3F0D)90 is shown in cyan and the crystal structure solved for the Q151E mutant is displayed in magenta. (RMSD: 0.152 Å)

68 Analysis of BpIspF Q151E Melting Thermal-shifts in Presence of Inhibitors

The mutant enzyme Q151E was successfully destabilized, resulting in a Tm of 64°C. This represents at least a 36°C decrease in Tm relative to wild-type IspF. Because this variant also maintained a wild-type level of activity, its use was justified to examine inhibitor binding using the thermal-shift assay. This high-throughput assay can be used to efficiently screen compounds for their apparent affinity. Since the magnitude of the thermal-shift is dependent upon both the concentration and binding affinity of the inhibitor, a relative ranking of inhibitor binding affinities can be determined by keeping the concentration of inhibitor constant. The compounds that display the highest apparent affinity, defined as those displaying a shift in melting temperature of ≥4°C, were then tested in an alternative assay to determine dissociation constants.

A total of four chemical series were studied using the temperature assay to determine a rank order of compound affinities.

Previous research on IspF inhibitors found coordination to the active site zinc was important for binding. This mechanism of action has been observed for other classes of enzymes.

For example, matrix metalloprotease (MMP) inhibitors are known to bind to the active site zinc via coordination by three histidine residues, similar to IspF. Furthermore, it has been shown that these inhibitors typically contain a peptidomimetic backbone that chelates with the zinc (Figure

2-15)104,105 In collaboration with Dr. Michael Clare at Emerald Biostructures, a similar strategy was used to design a new chemical series that mimics MMP inhibitors, which contained 5 or 6- member ring systems (Figure 2-16). The first chemical series (synthesized by Zheng Zhang) consisted of thirty-one compounds based around a common pyrimidine core (Table 2-4). The

69 thermal-shift assay indicated that the pyrimidine substituents which contain hydrogen bond donor/acceptor groups have an overall higher apparent affinity (Figure 2-18). These compounds employ a similar mechanism of binding to that of the IspF substrate, CDP-ME2P, which also contains hydrogen bond donor/acceptor groups in the cytosine ring that hydrogen bonds to the enzyme (Figure 2-17). Therefore, these analogs could be utilizing the same interactions as the cytosine ring and increasing the affinity of the ligand for IspF. In addition, the compounds that were converted to disodium salts were in general more active than their ethyl ester counterparts likely due to increased solubility. The pyrimidine core is expected to coordinate to the active site zinc and the substituents are proposed to bind in the cytidine pocket, where the hydrogen bond donor/acceptor can interact with the backbone of the enzyme.

Figure 2-15 Observed binding model of various MMP inhibitors that mimic peptidomimetic backbone inhbitors.104,105

Figure 2-16 The 6 and 5-member zinc binding molecules designed to mimic MMP inhibitors. Where X = Y = O or N.

Figure 2-17 Active site of E. coli IspF displaying hydrogen bonding of cytosine ring to backbone. (PDB: 1H48)63

Figure 2-18 Example of normalized thermal-shift assay in the absence and presence of compound. The black squares represent the unfolding of BpIspF in the presence of 5% DMSO as a control. The open triangles display the shift in melting temperature in the presence of 100 M of compound 2 containing 5% DMSO.

Table 2-4 continued 72

Table 2-4. Table of potential pyrimidine IspF inhibitors synthesized for this study. The Tm determined in the thermal-shift assay assesses the relative ranking of pyrimidines binding to IspF. ΔTm @ 100 Compound Structure µM

1 6.0

2 5.5

3 5.5

4 5.0

5 4.7

6 4.5

Continued on following page Table 2-4 continued 73

7 4.0

8 3.2

9 3.2

10 3.0

11 3.0

12 2.8

13 2.8

Continued on following page Table 2-4 continued 74

14 2.7

15 2.7

16 2.5

17 2.5

18 2.5

19 2.5 Continued on following page Table 2-4 continued 75

20 2.5

21 2.3

22 2.2

23 2.2

24 2.0

25 1.5

26 1.5 Continued on following page Table 2-4 continued 76

27 1.0

28 0.8

29 0.7

30 0.3

31 0.2

Continued on following page 77 Surface plasmon resonance has also been used to study binding of potential Escherichia coli IspF inhibitors.106 The study showed that L-tryptophan hydroxamate bound to E. coli IspF with a Kd,app of 2 µM. Four tryptophan analogs, the two enantiomers of the amino acid as well as their hydroxamate derivatives, synthesized by Gashaw Goshu, were evaluated using the thermal- shift assay for their ability to bind Burkholderia pseudomallei IspF (Table 2-5), which possesses

30% sequence identity to E. coli IspF. The results demonstrated that the hydroxamate derivatives possessed a higher relative affinity compared to their respective amino acids (Table 2-5). The hydroxamate functional group in MMP inhibitors is known to coordinate to the active site zinc through a tetrahedral or trigonal bipyramidal geometry (Figure 2-15).107 Overall these interactions are predicted to increase the apparent affinity of the ligand by providing more interactions to the enzyme. In addition, L-tryptophan hydroxamate exhibited a slightly larger shift in melting temperature than D-tryptophan hydroxamate and is expected to bind with higher affinity to the IspF. This result is probably not surprising since most enzyme active sites are stereospecific and designed to operate only on the L-enantiomorphs of chiral compounds, since that form is more prevalent in nature.

78 Table 2-5. Tryptophan analogs synthesized to study IspF binding.

Compound Structure ΔTm @ 500 µM

1 5.5

2 5.2

3 3.3

4 3.3

79 The last two chemical series studied in the thermal-shift assay are structurally similar, both containing an imidazole ring. The first is an imidazothiazole series (synthesized by Gashaw

Goshu). The series includes two thiazolopyrimidine compounds, which are structurally similar to the imidazothiazole core, which can be considered as controls based on literature reports demonstrating they are inhibitors of IspF from plant, parasitic, and bacterial organisms with IC50 values between 10 µM and 200 µM108 (Table 2-6). The two control compounds 1 and 2 generated thermal shift values >17°C against Burkholderia pseudomallei IspF. The imidazothiazole analogs have three positions available for substitution, two located on the imidazole ring (R1 and R2) and one on the thiazole ring (R0) (Figure 2-19). A general trend observed in the imidazothiazole series was substitution at R1 and R2, increased the apparent affinity (Figure 2-20). Substitutions at these positions are predicted to primarily extend into the cytidine pocket for potential for interactions. These ligands can potentially hydrogen bond with the backbone of the enzyme resulting in the increased apparent affinity (Figure 2-17).

Substituents at R0, which points towards the lipophilic pocket, were found to bind with a lower apparent affinity when the substituent was bulky. It is hypothesized that the lipophilic pocket cannot accommodate the bulky substituents, with the notable exceptions of 3 and 5.

Figure 2-19. Imidazothiazole compounds contained substituents at C2 (R0), C5 (R1), and C6 (R2).

Figure 2-20 Crystal structure of BpIspF in complex with FOL717 displaying interactions of imidazole ring substituents with the cytidine pocket. BpIspF is shown in cyan with the ligand FOL717 (magenta) N4 binding to zinc (1.9Å) and the hydroxyl group hydrogen bonding to K134 (2.9Å). (PDB: 3MBM) 90

Table 2-6 continued 81

Table 2-6. Imidazothiazole series thermal-shift analysis for apparent binding affinity.

Compound Structure ΔTm @ 500 µM

1 18.8

2 17.1

3 8.3

4 5.8

Continued on following page Table 2-6 continued 82

5 5.8

6 4.3

7 4.3

8 4.3

Continued on following page Table 2-6 continued 83

9 3.6

10 3.5

11 2.7

12 2.5

13 1.0

14 0.8

Continued on following page Table 2-6 continued 84

15 0.7

16 0.7

17 0.6

18 0.6

19 0.5

20 0.5

21 0.5

Continued on following page Table 2-6 continued 85

22 0.5

23 0.5

24 0.5

25 0.3

26 0.3

27 0.3

Continued on following page Table 2-6 continued 86

28 0.3

29 0.2

30 0.2

31 0.2

32 0.2

33 0.2

Continued on following page Table 2-6 continued 87

34 0.0

35 0.0

36 0.0

37 0.0

38 0.0

39 0.0

Continued on following page Table 2-6 continued 88

40 0.0

41 -0.2

42 -0.2

43 -2.2

Continued on following page 89 The final series examined in the thermal-shift assay were the imidazole compounds synthesized by Zheng Zhang and Gashaw Goshu (Table 2-7). In general, the apparent affinities of the imidazole compounds were shown to be weaker than the imidazothiazole series (lower

ΔTm values). The phenyl substituents that are di-substituted (3, 4, 6, and 7) with an electron- withdrawing group were more potent than those that are mono-substituted (5, and 10-17). There was one compound observed to have a significant destabilization effect in the imidazole series, compound 30, which was quite interesting, suggesting it may favor binding to the unfolded state relative to the native state. Alternatively, it may bind strongly to zinc ions such that it chelates metal ions, thus decreasing IspF stability. Although the inhibitor destabilized the enzyme it could still be a good inhibitor potentially due its strong binding affinity for zinc if it bound IspF in a sequence/structure-dependent manner. Therefore, compounds that are found to have a significant negative thermal shift should not be eliminated from screening in the determination of dissociation constants.

Table 2-7 continued 90

Table 2-7. Imidazole series thermal-shift analysis for apparent binding affinity. Compound Structure ΔTm @ 500 µM

1 4.5

2 3.7

3 3.7

4 3.5

5 3.5

Continued on following page Table 2-7 continued 91

6 3.5

7 3.3

8 3.3

9 3.2

10 3.2

Continued on following page Table 2-7 continued 92

11 3.2

12 3.0

13 2.8

14 2.7

15 2.5

Continued on following page Table 2-7 continued 93

16 2.5

17 2.5

18 2.3

19 1.8

20 1.8

Continued on following page Table 2-7 continued 94

21 1.8

22 1.7

23 1.3

24 1.0

25 1.0

Continued on following page Table 2-7 continued 95

26 0.8

27 0.5

28 0.3

29 -7.2

Continued on following page 96 Conclusions

Due to BpIspF’s high thermostability, a protein engineering approach was used to destabilize the enzyme allowing the development of a thermal-shift assay to evaluate potential small molecule inhibitors. Analysis of twelve single point mutants using the program FoldX yielded estimates of destabilization ΔΔG° values for each variant of the enzyme. Eleven variant forms of the enzyme were predicted to destabilize BpIspF. These mutants were produced and analyzed, with three variants displaying destabilization at physiological pH. As a result of these mutants, the thermostable enzyme Burkholderia pseudomallei IspF was destabilized sufficiently to produce a Tm value at which small molecule inhibitors could be reliably analyzed for apparent binding affinity by monitoring unfolding temperatures. Additionally, of these three mutants, the

Q151E variant was found to not affect enzyme activity and was therefore used as the basis for a thermal-shift assay. The use of a high-throughput thermal-shift assay provided a rank order of apparent affinities within four chemical series. This allowed the library of compounds to be narrowed down to a reasonable number of compounds for subsequent experimental analysis to determine dissociation constants. These methods for incorporating isosteric charged residues into any target protein of interest could be further refined and optimized as a broadly applicable approach to developing high throughput thermal-shift assays for enzymes that possess extreme thermostability.

Thermal-shift assays have been used across a wide variety of applications, such as the identification of Bacillus anthracis PurE inhibitors85 Toxoplasma gondii Enoyl-Acyl Carrier

Protein Reductase inhibitors,109 the recognition of caffeine and analogs of caffeine by an anti- caffeine camelid antibody,110 as well as the stabilization of CC-chemokine receptor 5 by small

97 molecules.111 Here, the engineering BpIspF variant opens new opportunities for the thermal-shift assay to be used to evaluate compounds that bind the target, BpIspF.

CHAPTER 3

CHARACTERIZATION OF BINDING AFFINITY OF BURKHOLDERIA

PSEUDOMALLEI ISPF INHIBITORS

The determination of a ligand’s affinity for an enzyme can be measured by a variety of methods such as mass spectrometry (MS),112,113,114,115 fluorescence polarization (FP),116,117,118 isothermal titration calorimetry (ITC),119,120,121 nuclear magnetic resonance (NMR),122,123,124 and surface plasmon resonance (SPR).125,126,127 Each technique has its own advantages and disadvantages in the determination of the ligand’s affinity for the enzyme. For example, NMR and MS are well- suited for efficiently screening large compound libraries to quickly identify a subset of molecules that bind to the active site of an enzyme target. Such reduced compound subsets can then be evaluated using titration experiments to determine binding affinities.114,123,115 FP, first reported in 1920,128 was used for high throughput screening in the 1990s, often in competitive binding assays.116 However, there are inherent limitations with FP assays, such as inconsistent probe fluorescence at high concentrations (>100 µM), non-specific interactions, and compound aggregates, which can make determining ligand affinity difficult.5 SPR has been used since the

1990s to analyze protein binding thermodynamics and kinetics.129,130 Similarly in the 1990s, ITC became widely employed in drug discovery to examine the thermodynamic profile of ligand- enzyme interactions.131 ITC has the advantage of being able to study enzyme-ligand interactions at lower concentrations due its ability to detect small differences in heats of reaction.

99 One of the major limitations for any of these techniques in obtaining weak affinity dissociation constants for small organic ligands that bind their protein targets weakly (Kd values mM to μM) is solubility. This chapter will discuss the analysis of BpIspF inhibitor dissociation constants determined using ITC, SPR, and FP.

Methods

Surface Plasmon Resonance (SPR)

SPR binding experiments were performed on a Reichert SR7000DC (with a SR7500DC upgrade) using a high density 500,000 Da carboxymethyl dextran SR7000 gold sensor slide. IspF was coupled to the sensor slide using amine coupling chemistry132 (Figure 3-1). Initially, the surface was activated by injecting a mixture of 80 mg/ml EDC (N-Ethyl-N′-(3- dimethylaminopropyl)carbodiimide hydrochloride) and 20 mg/ml NHS (N-hydroxysuccinimide) for 7 min using a 50 μl/min flow rate. IspF, diluted to a final concentration of 120 μg/ml in 10 mM 4.5 pH sodium acetate buffer, was injected over the surface for 7 min at 30 μl/min. Excess succinamide ester groups were quenched with an injection of 1M ethanolamine for 3 min with a

50 μl/min flow rate. A total of 4000-8000 μRIU (Refractive index units) of IspF was coated on the slide. Binding experiments were conducted at 25°C in running buffer (50 mM sodium phosphate pH 7.4/150 mM NaCl/4.76% DMSO) at 50 μl/ min. Compounds were dissolved in

DMSO to initial concentrations between 4-20 mM, dependent upon solubility. These were further diluted (20-fold) into running buffer to produce a final concentration of 200-1000 μM compound with 4.76% DMSO, thereby minimizing buffer mismatch. Compounds were serially diluted two-fold using running buffer and injected over the IspF surface. Blank (running buffer

100 only) injections were included between each compound injection, allowing for double subtraction (reference cell and buffer) during data processing. Data processing and analysis were performed using Scrubber (BioLogic Software).

Figure 3-1. Overview of BpIspF coupling chemistry. A 500,000 Da carboxymethyl dextran SR7000 gold sensor slide (sphere) was used to coat BpIspF on the surface via EDC/NHS coupling chemistry. The surface was first activated with EDC to form an o-Acylisourea leaving group. N-hydroxysuccinimide was then used to convert the activated carboxylate to an NHS ester intermediate, which is a good leaving group. The amine groups in the enzyme can then form amide bonds with the activated surface.

101 Isothermal Titration Calorimetry

ITC experiments were performed using a Microcal VP-ITC titration calorimeter

(Microcal, LLC, Northampton, MA). Buffers were matched for all ITC experiments by dialyzing the protein overnight against 4L of dialysis buffer (20 mM sodium phosphate monobasic, pH 7.4,

150 mM NaCl). The ligand was dissolved in the dialysis buffer and centrifuged for 5 min at

14,000 rpm. The extinction coefficient of BpIspF (11,560 M-1 cm-1 at 280 nm) was determined with 6M guanidine hydrochloride based using the Edelhoch method.133 Protein and ligand concentrations were measured in triplicate and the average value was used to calculate the final

IspF concentration. The concentrations of the tryptophan analogs were calculated by measuring absorbance at 280 nm using the published extinction coefficient of tryptophan (5540 M-1 cm-

1).134 The concentrations of fusion compounds were based on dry weight. The ITC experiments were performed with BpIspF (cell) and ligand (syringe) diluted to appropriate concentrations in dialysis buffer. All experiments included a total of 28 injections with each injection being separated by 240 seconds at 25°C with a stirring speed of 307 rpm. The initial injection was 2

μL, which was omitted in data processing, to account for dilution across the syringe. The remaining injections were 10 μL. Dilution heats of the ligand were measured in a separate experiment by injecting the ligand (syringe) into dialysis buffer (cell). Data were analyzed using

Origin 7.0 with an ITC add-on supplied by the manufacturer (Microcal, Northampton, MA).

Fluorescence Polarization

The dissociation constant (Kd,app) of the fluorescent probe, HGN-0201, for BpIspF was evaluated using a Biotek Synergy 2 fluorescent plate reader. Purified IspF was spin concentrated to 40 mg/ml using a 3,000 MWCO spin concentrator. The stock IspF was serially diluted in FP

102 buffer (20mM sodium phosphate monobasic, pH 7.0, 300 mM NaCl, 2 mM DTT) to produce final concentrations ranging from 0-500 μM with 100 μM HGN-0201 at 5% DMSO in a 96-well black plate (Corning No. 3915). The plate was incubated at 30oC for 30 min followed by a 30 s

shake then read (λex = 330, λem = 528). The Kd,app was determined by nonlinear regression analysis (Equation 3-1) in Origin 7.0 using the following relationship.118

; Eq. 3-1

where Aobs is the observed anisotropy, Af and Ab are the anisotropy of the free and bound probe, respectively, LT is the total probe (HGN-0201) concentration, RT is the concentration of enzyme, and KD is the apparent dissociation constant of the probe. A schematic representation of the instrumental setup is shown in Figure 3-2.

103

Figure 3-2. Instrumental setup of fluorescence polarization assay using a Fluoromax 4 Spectrofluorimeter. Fluorophore is excited with a short pulse of vertically polarized light (z) that causes photoselection. The polarizer (x) can be rotated from the vertical to horizontal position. The decay of vertically polarized fluorescence (Ivv) is measured then the polarizer is moved to 135 the horizontal orientation (Ivh) to be detected.

104 Results and Discussion

Determination of Dissociation Constants by Surface Plasmon Resonance (SPR)

Due to the labor-intensive nature of the enzyme inhibition assays, it was determined that

Surface Plasmon Resonance (SPR) would be utilized as the primary method for measuring the dissociation constant of inhibitors targeting BpIspF. The dissociation constants, Kd,app, were determined using steady-state equilibrium experiments. The first ligands that were screened, the fusion series,136 were designed to link the zinc and cytidine active site sub-pockets thereby providing mimetics of the IspF substrate (Table 3-1). These were hypothesized to enhance affinity compared to the control cytidine diphosphate (CDP) (Figure 1-16). Cytidine diphosphate was found to bind with a stoichiometry of one with a Kd,app of 75 µM (Table 3-1). This was examined via the synthesis of four, diverse zinc binding groups synthesized by Zheng Zhang: imidazothiazole, thiazole, pyridine, and azabenzimidazole. To investigate these interactions with the enzyme the compounds were analyzed by in silico docking using SYBYL X-2.0. Docking studies provided predicted KD values (as pKD values). As shown in Table 3-1 these values were comparable to experimental Kd,app values determined using SPR, which ranged from 70-200 µM.

The zinc binding moieties were predicted to direct the major conformational differences observed in the docking poses (Figure 3-3). For example, compound HGN-0006 has a water- mediated hydrogen bond to the active site zinc ion, while HGN-0007 has no interaction with the zinc ion in comparison to HGN-0006. The interaction with the zinc ion in HGN-0006 was hypothesized to increase the affinity of the ligand due to the additional point of interaction with the enzyme.

105 Table 3-1. Binding affinity and stoichiometry of fusion series by SPR utilizing cytidine diphosphate as a control compound. Docking studies were done using SYBYL X-2.0 where pKD, predicted log KD, were determined that correlate to SPR Kd,app values. *SPR ran by Sriram Jakkaraju.

Compound SPR Kd,app SPR Estimated Docking Score Stoichiometry (pKD) (μM)

CDP 75 1:1 N.A.

HGN-0005 180 3:1 8.11

HGN-0009 148 3:1 8.34

HGN-0006 70 3:1 8.91

HGN-0008 200 3:1 7.13

HGN-0007 90 3:1 8.45

106

Figure 3-3. Docking of HGN-0006 and HGN-0007 in the active site of BpIspF (cyan) displays structural differences in modes of binding. HGN-0006 (magenta) coordinates to zinc (blue sphere) via a water-bridge (red sphere) while HGN-0007 (yellow) has no interaction with zinc.

107 To identify new fragment hits that bind to MEP pathway enzymes, Saturation Transfer

Difference (STD) NMR experiments were performed using Mycobacterium paratuberculosis

IspD, Mycobacterium abscessus IspE, and Burkholderia pseudomallei IspF by Dr. Darren

Begley (SSGCID). As shown in Figure 3-4, the NMR experiments were able to detect molecules that bound to IspD, IspE, and IspF. Analogs of FOL955 (synthesized, by Zheng Zhang and

Gashaw Goshu) where N1 in the imidazole ring was moved to C5 for ease of synthesis and derivatization at N5 (Figure 3-6). A series of twenty analogs were synthesized and screened using SPR to determine dissociation constants and initial structure-activity relationships (SAR).

The compound L-tryptophan hydroxamate was used as a control ligand for the imidazole series.

106 It was reported to bind to Escherichia coli IspF with a Kd,app of 2 µM using SPR. The ligand was found to bind to BpIspF with a Kd,app of 60  3 µM (Figure 3-5).

SPR analysis of the imidazole compounds was completed starting with the benzyl derivatives including fluoro- substituted analogs, e.g. HGN-0095, which in general were better than the chloro- substituted analogs, e.g. HGN-0051. The compound HGN-0021, a benzyl- substituted imidazole, was shown to be essentially inactive with a Kd,app >2,000 µM. Analogs containing chlorine or fluorine were hypothesized to increase favorable interactions with the enzyme in the cytidine pocket (Figure 1-16). Due to the higher polarizability of chlorine these analogs could have an increased affinity for the enzyme in comparison to the fluoro-substituted analogs.137 Relative to HGN-0021, the fluorine substituted benzyl analogs at the ortho- (HGN-

0106) and para- (HGN-0093) positions were determined to be effectively equal in potency with dissociation constants of 96 µM and 83 µM, respectively. In comparison, a fluorine substituent on the meta- position, HGN-0109, increased the affinity two-fold to 45 µM suggesting that at the meta- position fluorine may introduce more favorable interactions with the enzyme via electron

108 donating amino acids. The chlorine substituted benzyl analogs, HGN-0090 and HGN-0094, were found to enhance binding affinity when substituted at the meta- (Kd,app = 22 µM) and para- (Kd,app

= 21 µM) positions but inactive in the ortho-position (HGN-0091). Since chlorine has a larger electron density, they may be interacting more strongly with electron donating groups in the active site of the enzyme but absent when in the ortho- position. Analysis of the disubstituted chlorobenzyl analog HGN-0092 was found to be inactive, which could be due to deactivation of the ring with the addition of the second chlorine molecule.

The phenethyl derivatives corresponding to the above compounds were also evaluated.

Compound HGN-0022 was found to be inactive (Kd,app > 2000 µM) suggesting that unsubstituted analogs may not be able to interact with the enzyme to the same extent as those containing a halogen. The fluorine analogs were determined to be more active than the chlorine substituted analogs. The compound HGN-0097 contained a m-fluoro substituent and was found to be inactive while HGN-0109, a m-fluoro benzyl analog, bound to the enzyme with a Kd,app of 45

µM. The decrease in affinity is most likely due to the loss of the fluorine interacting with the electron donating amino acid, such as E137, because it was moved further into the cytidine pocket when the linker was extended. The ortho- (HGN-0107) had an SPR Kd,app of 51 µM showing an enhanced affinity as did the para- (HGN-0095) with a Kd,app of 36 µM. The increase in the affinity from the benzyl analogs are opposite from the observation with HGN-0109. The linker extension for the ortho- and para-fluoro (HGN-0095 and HGN-0107) phenethyl substituted compounds appears to introduce additional favorable interactions with the enzyme enhancing ligand affinity for the enzyme in comparison to the benzyl analogs (HGN-0093 and

HGN-0106). Overall, the mono- and disubstituted chlorine analogs were not active against

BpIspF with an SPR Kd,app >2,000 µM. The linker length in the monosubstituted chlorine

109 compounds HGN-0096, HGN-0108, and HGN-0051 are bringing the substituent closer to the cytosine pocket. Since the cytosine pocket does not contain electron-donating groups for the chlorine it is binding similar to HGN-0022. Furthermore, the disubstituted analogs (HGN-0098,

HGN-0099, and HGN-0100) deactivate the ring which decreases its affinity for the enzyme.

Finally, a fluorescent ligand, HGN-0201, was developed for use in a potential fluorescence polarization assay. The dansyl chloride group has previously been used in literature to modify amines, carboxylic acids, and hydroxyl functional groups to fluorescently label small

138 molecules. The compound, HGN-0201, was determined to have an SPR Kd,app of 78 µM.

Although, the fluorescent compound was hypothesized to bind through the imidazole nitrogen, as indicated by the crystal structure of FOL955 (Figure 3-6), there are alternative modes by which the ligand may interact with the active site zinc ion. It is possible that HGN-0201 could also bind through the sulfonamide or the phenol. The latter group also lends itself to the ready synthesis of new IspF inhibitors.

Figure 3-4. STD-NMR screen identified seventy-two molecules that target BpIspF with three displayed (red). Additionally select MpIspD fragment hits (cyan) and MaIspE fragment hits (black) identified in the screen are shown above.

110

111 A B

Figure 3-5. The determination of Kd,app of control, HGN-0003, screened from 7.81-1000 µM. (A) SPR sensogram displaying the response in µRIU for each concentration tested and (B) semi- logarithmic plot of the data from (A) demonstrating that the Kd,app as determined by SPR was found to be 60  3 µM.

Table 3-2 continued 112

Table 3-2. Binding affinity of imidazole series for Burkholderia pseudomallei IspF determined by SPR utilizing L-tryptophan hydroxamate as a control.

Compound Structure Kd,app (µM)

FOL955 >2000 µM

HGN-0021 >2000 µM

HGN-0106 96 ± 2 µM

HGN-0109 45 ± 1 µM

Continued on following page Table 3-2 continued 113

HGN-0093 83 ± 2 µM

HGN-0091 >2000 µM

HGN-0090 22 ± 6 µM

HGN-0094 21 ± 2 µM

HGN-0092 >2000 µM

Continued on following page Table 3-2 continued 114

HGN-0022 >2000 µM

HGN-0107 51 ± 1 µM

HGN-0097 >2000 µM

HGN-0095 36 ± 1 µM

HGN-0096 >2000 µM

HGN-0108 >2000 µM

Continued on following page Table 3-2 continued 115

HGN-0051 >2000 µM

HGN-0098 >2000 µM

HGN-0099 Non-specific

HGN-0100 >2000 µM

HGN-0201 78 ± 8 µM

Continued on following page 116

Figure 3-6. Structure of Burkholderia pseudomallei IspF in complex with FOL955. FOL955 (magenta) binds to zinc (blue sphere) of BpIspF (cyan) through the imidazole N with a bond distance of 2.1Å.The crystal structure indicates substitution at the 5-position of the imidazole ring will point towards the cytidine pocket. (PDB: 3QHD)90

117 Analogs of FOL535 were synthesized by Gashaw Goshu with substituents on the 2, 5, and 6- position of the imidazothiazole ring and screened using SPR (Table 3-3). The parent compound (HGN-0245) was determined to bind to the enzyme with an affinity of 407 µM.

FOL535 (i.e., HGN-0245) was co-crystallized with Burkholderia pseudomallei IspF and found to bind to the zinc through N7 of the imidazo[2,1-b]thiazole ring.90 Furthermore, FOL535 (HGN-

0245) provided a better initial dissociation constant as a starting scaffold to FOL955 which was shown to bind with an affinity >2000 µM (Table 3-2). The analogs based on FOL535 were synthesized with an amide rather than an ester linkage. There were two reasons for this decision: ease of synthesis and lability of the bond. The ester can be more readily hydrolyzed to the acid whereas the amide is stable. Substituents at C2, which point toward the lipophilic pocket, indicated that the addition of large groups decreased the affinity of the ligand. This can be observed by comparing HGN-0289, which bound with a dissociation constant of 210 µM, and

HGN-0282, which bound with a dissociation constant of 1200 µM. Therefore, the amide bond improved binding affinity with small, hydrophobic groups in comparison to the fragment hit

FOL535 (HGN-0245). On the other hand, bulky hydrophobic substituents are unfavorable because the pocket may not accommodate the size of the hydrophobic group, as observed in the thermal-shift assay. Although HGN-0321 contains a bulky substituent at the 5-position as well as the 2-position, it was found to have a Kd,app of 25 µM, which did not follow the observed SAR with HGN-0282 and HGN-0289. The ligand, HGN-0321, could potentially chelate to the zinc through the carbonyl which increases the chance of the pyridine ring to hydrogen bond with the backbone of the enzyme in the cytidine pocket and accommodate the benzoyl substituent.

Alternatively, the compound could coordinate to the zinc through the pyridyl nitrogen similar to the observed binding of FOL8395 (PDB: 3JVH)90.

118 In addition, analogs were made with the hit molecule FOL717 (HGN-0247) identified in the STD-NMR screen that bound with the same affinity as HGN-0289 with a KD of 210 µM, due to a hydrogen bond between the primary alcohol and K134 (Figure 3-7). The substituent on the imidazole ring points towards the cytidine pocket, which can accommodate larger substituents and have more interactions with the backbone of the enzyme, Ala102 to Ala108. Therefore,

HGN-0381 was designed to add a hydrogen bond donor groups that extends into the cytidine pocket. However, the compound was determined to have a decreased affinity (Kd,app 620 µM).

This decrease in affinity could be due to N7 of the imidazole ring not being at a bonding distance with the zinc ion and therefore increasing the linker length could restore affinity. Additionally, the imidazothiazole core was substituted with a different zinc binding group, a pyridine ring, to determine the effect of changing the core. Two compounds were synthesized, HGN-0316 and

HGN-0317, which altered the position of the nitrogen in the ring. SPR results revealed that the nitrogen in the 4-position had a higher affinity for the enzyme with a Kd,app of 855 µM for HGN-

0316 whereas with the nitrogen in the 3-position as for HGN-0317 it was only 1360 µM. The two compounds containing the pyridine ring displayed a decrease in binding affinity due to the increased flexibility within the molecule. Overall, maintaining the rigidity of the ligands with the imidazothiazole core while including hydrophobic interactions with the lipophilic pocket and hydrogen bonding with the cytidine pocket seem to result in increased binding affinity.

Table 3-3 continued 119

Table 3-3. Binding affinity of imidazothiazole compounds to BpIspF using SPR with L- tryptophan hydroxamate as control at 25°C.

Compound Structure Kd,app (µM)

HGN-0245 407 ± 8

HGN-0289 210 ± 60

HGN-0282 1200 ± 3

HGN-0321 25 ± 1

HGN-0247 210 ± 10

Continued on following page Table 3-3 continued 120

HGN-0318 620 ± 9

HGN-0316 855 ± 8

HGN-0317 1360 ± 3

Continued on following page Table 3-3 continued 121

Figure 3-7. Structure of the BpIspF (cyan) active site in complex with FOL535 and FOL717. Substituents at the 2-position of FOL535 (yellow), are observed to bind in the lipophilic pocket while substituents at the 6-position of FOL717 (magenta) bind in the ribose pocket of BpIspF. (PDB: 3K14 and 3MBM)90

Continued on following page 122

Thermodynamic Analysis of Binding by Isothermal Titration Calorimetry (ITC)

ITC was used as an alternative method to study dissociation constants as well as the underlying thermodynamics of inhibitor binding to trimeric IspF. Two structurally similar fusion compounds, HGN-0006 and HGN-0009, were investigated to analyze the thermodynamic contributions of binding HGN-0006 (Figure 3-8), HGN-0009, and the control cytidine diphosphate. ITC revealed a stoichiometry of three with the fusion compounds in comparison to

CDP, which bound one molecule to the homotrimer. Additionally, the control compound CDP had a much larger enthalpic contribution to binding the target, BpIspF. The ITC binding affinity was in close agreement to the Kd,app values of the two fusion compounds determined by SPR.

Using the assumption that ligand-protein interactions dominates the enthalpic contribution, optimizing this parameter can be used to guide the synthesis of more potent compounds.139 Since both HGN-0006 and HGN-0009 displayed a significantly less favorable enthalpic contributions than CDP, which had a ∆H of -24.7 kcal/mol, modification of CDP would likely lead to compounds that bind with a higher affinity. Although, CDP has a significant unfavorable entropic contribution of binding to BpIspF, -T∆S of 19.4 kcal/mol, and balancing enthalpic and entropic contributions of binding need to be considered to increase affinity (Table 3-4).

123 Table 3-4. Thermodynamic analysis of imidazothiazole fusion compounds binding to BpIspF by ITC.

Compound N ∆G° ∆H° -T∆S° pH Kd,app (μM) (kcal/mol) (kcal/mol) (kcal/mol)

CDP 7.4 134 ± 5 0.84 ± 0.01 5.28 ± 0.05 -24.7 ± 0.5 19.4 ± 0.5

HGN-0009 7.4 45.87 ± 3 ± 0 5.916 ± 0.002 -5.95 ± 0.07 0.03 ± 0.07 0.08 (fixed)

HGN-0006 7.4 3 ± 0 22 ± 1 6.35 ± 0.05 -8.61 ± 0.02 2.25 ± 0.05 (fixed)

Figure 3-8. ITC titration of HGN-0006 with Burkholderia pseudomallei IspF to determine Kd,app at physiological pH and 25°C.

124 The SPR control compound L-tryptophan hydroxamate, HGN-0003, was also studied using ITC, along with the D-enantiomer and their L- and D- amino acids. This allowed determination of the thermodynamic contributions of the hydroxamate moiety and enantiomeric specificity. The L- and D-hydroxamates were both shown to bind more tightly than L- and D- tryptophan and BpIspF was more specific for the L-enantiomer (Table 3-5). Overall, the thermodynamic analysis at pH 7.4 indicates the binding of the L- and D-tryptophan hydroxamate analogs to the enzyme were enthalpically-driven with small entropic contributions. Based on the structure of IspF, the stoichiometry was expected to be either 1.0 or 3.0. The binding stoichiometry of L-tryptophan hydroxamate was 1.4, which suggests just over one ligand binding to the IspF trimer (Figure 3-9). Thelemann, et al, observed a similar binding stoichiometry of 1.4 in the bis-sulfonamides with Arabidopsis thaliana IspF. 115 In comparison, the D-tryptophan hydroxamate was found to have approximately less than one ligand (N=0.7) binding to the IspF trimer with a two-fold decrease in affinity. The observed differences in stoichiometry may likely reflect issues in determining an accurate concentration of the ligand, with a true stoichiometry of

1:1. This may be due to the hydroxamate moiety changing the extinction coefficient of tryptophan, which was used for concentration determination. Due to weak binding affinities, a sigmoidal binding titration was not observed in initial ITC experiments of D- and L-tryptophan using a 10:1 ratio. The experiment was then run at 100:1 ratio, due to the low c-value, and the stoichiometry was correspondingly fixed during model fitting.140 Additionally, L-tryptophan hydroxamate was studied at a slightly lower pH (6.0) to explore the affinity under crystallization pH conditions. The binding of the ligand under the slightly acidic pH condition was ~20-fold weaker and was entropically driven. The pH dependent binding event could be explained by a change in protonation of the enzyme or ligand between the two pH values, 7.4 to 6.0.

125 Table 3-5. Tryptophan and tryptophan hydroxamate binding affinities determined differences in binding affinity for BpIspF by ITC at 25°C. pH dependent binding was observed in HGN-0003 which can be used to optimize crystallization conditions due to ligands binding more favorably under physiological conditions.

Kd,app Compound pH N ∆G° (kcal/mol) ∆H° (kcal/mol) -T∆S° (kcal/mol) (μM)

30 ± 2 6.2 ± 0.1 HGN-0003/ 7.4 1.4 -6.649 ± 0.002 2.1 ± 0.1 L- tryptophan hydroxamate 4.19 ± 0.01 HGN-0003/ 6.0 848 ± 2 1 1.1 ± 0.2 -5.2 ± 0.2 L- tryptophan hydroxamate -6.650 ± 0.001 HGN-0778/ 7.4 64 ± 7 0.7 5.7 ± 0.1 0.9 ± 0.1 D- tryptophan hydroxamate

L- 7.4 >1,000 1 - - - tryptophan

D- 7.4 >1,000 1 - - - tryptophan

126

Figure 3-9. ITC titration of HGN-0003 with Burkholderia pseudomallei IspF to determine Kd,app at physiological pH and 25°C.

127 Development of a Fluorescence Polarization Assay (FP)

To explore an alternative approach to assess dissociation constants, Kd,app, in a high- throughput manner, a fluorescence polarization assay was developed. Specifically, the goal was to develop a competitive fluorescence polarization assay in which a library of inhibitors could be screened in a plate-based assay to identify compounds which competed for the same active site.

The fluorescent molecule HGN-0201 was synthesized with the imidazole moiety utilizing a dansyl group as the fluorophore. The methodology behind fluorescence polarization is that the fluorophore’s degree of polarization is inversely proportional to its molecular rotation. Once the fluorophore is bound to the macromolecule of interest, in this case Burkholderia pseudomallei

IspF, it will slow down the molecule’s rotation increasing its fluorescence polarization signal

(Figure 3-10). The fluorescent properties of HGN-0201 were analyzed, including the excitation and emission spectra. The absorbance of the molecule was measured from 300-550 nm with a

λmax of 330 nm (Figure 3-11a). Analogously, the emission profile was collected from 350-750 nm, which possessed a λmax of 548 nm (Figure 3-11b).

128

Fluorophore

Fast Rotation Low Polarization Slow Rotation High Polarization

Figure 3-10. Illustration of fluorescence polarization principle of enzyme binding fluorophore Fluorescence polarization of the small molecule will be low initially due to it rotating quickly in solution. When macromolecule is added into the solution and as the ligand binds it will decrease the rate at which the ligand rotates and give a high polarization. The polarization value increases as more ligand binds to the macromolecule.

129

Figure 3-11. Wavelength scans of HGN-0201 for determination of λmax values. (A) Excitation scan of HGN-0201 identified a λmax at 330 nm. (B) Emission scan of HGN-0201 identified a λmax at 548 nm.

130 The dissociation constant of HGN-0201 for Burkholderia pseudomallei IspF was determined using fluorescence polarization/anisotropy. The triplicate data set determined the

Kd,app to be 102.7 ± 9.8 µM (Figure 3-12). A competitive assay with 0-700 µM L-tryptophan hydroxamate was then attempted with 150 µM BpIspF, the concentration at which approximately

50% binding was observed. Overall the experimental data were inconsistent in the competitive assay and the Kd,app value of the ligand was not determined based upon this methodology. The anisotropy value determined at 50% binding was lower than the value measure in the binding assay of 5-(dimethylamino)-N-(2-(5-(4-hydroxyphenyl)-1H-imidazol-1-yl)ethyl)naphthalene-1- sulfonamide, HGN-0201. Scans of the competitive ligand displayed overlapping of the fluorescent signal, which interfered with the measured anisotropy values (data not shown).

Therefore, a competitive assay was not viable with the current fluorescent molecule.

131

Figure 3-12. Fluorescence polarization titration of BpIspF with HGN-0201 determined the dissociation constant to be 102.7 ± 9.8 µM.

132 Conclusion

The binding affinity of several groups of potential IspF inhibitors including the fusion series, imidazole series, imidazothiazole series, and tryptophan series were assessed using ITC and SPR. The fusion compounds, which were designed to link the cytidine and zinc binding pockets, were found to have the same affinity for the target in comparison to the control CDP. In silico studies were performed for the fusion compounds, which were able to predict the rank order of affinity that was observed by SPR. In addition, the mode of ligand binding was predicted by docking where HGN-0006 coordinates to zinc through a water-mediated hydrogen bond. Furthermore, the docking pose suggested the zinc-binding moiety of HGN-0009 does not coordinate to the metal ion, which is in agreement with its weaker affinity for BpIspF. ITC binding studies of two structurally similar fusion compounds, only differing by a methylene group, determined the binding was enthalpically driven. The addition of a methylene group decreased the enthalpic contribution by 2.66 kcal/mol likely due to moving the zinc binding moiety away from the active site zinc.

L-tryptophan hydroxamate had been previously shown to bind to E. coli IspF and was studied as an IspF control compound using both ITC and SPR. The compound was observed to bind with an affinity of 30 ± 2 µM by ITC and 60  3 µM by SPR. The D-tryptophan hydroxamate and the tryptophan enantiomers were also studied by ITC to determine the enantiomeric selectivity, as well as the importance of the hydroxamate functional group. The tryptophan enantiomers were found to bind with weak affinity >1000 µM indicating the hydroxamic acid group is important for increasing the binding affinity for BpIspF. Additionally, the L- tryptophan hydroxamate was found to bind more tightly to the enzyme with a KD of 30

133 µM relative to the D-tryptophan hydroxamate with a KD of 64 µM. The binding affinities of the tryptophan compounds were found to correlate to the thermal-shift analysis (Table 2-5).

To identify novel molecules that target the BpIspF active site, STD-NMR experiments were used, revealing BpIspF bound several compounds from the Fragments of Life (FOL) library including FOL955, FOL535, and FOL717. A series of analogs based on FOL955 were studied for binding by SPR that were established to have Kd,app ranging from 20 µM to greater than 2000

µM. Compounds containing a fluorine substituent in the ortho and para position were observed to bind more tightly than the chlorine substituted analogs in the phenethyl series. The meta and para chloro benzyl substituents were shown to bind more tightly than the flouro substituents in the imidazole series. In particular, a fluorescently labeled imidazole compound, HGN-0201, which possessed a dansyl group had a Kd,app of 78 µM, opening its possible use in fluorescence polarization assays. Analogs of FOL535 and FOL717 were studied by SPR and were found to have Kd,app ranging from 210-1360 µM. Analogs of HGN-0321, which potentially interacts with

IspF’s active site zinc ion through the 5-postion carbonyl, where the electron acceptor oxygen is not present, would give insight into the mode of binding. Alternatively, HGN-0321 could bind through the pyridyl nitrogen observed with FOL8395 (PDB: 3JVH)90 which can be examined using an analog which does not contain the carbonyl at the 5-position. Furthermore, analogs of

HGN-0321 could lead to more potent inhibitors of BpIspF. The thermal-shift assay (Chapter 2) was able to provide insight about relative affinities of the compounds, but unfortunately, correlation to the experimentally determined dissociation constants did not agree. This was particularly evident in the imidazole series. The FP assay was able to successfully measure the affinity at 102 µM, but the development of a competitive assay was not effective due to interference of the fluorophore signal by the competitive ligands.

CHAPTER 4

DEVELOPMENT OF AN ISPF ENZYME INHIBITION ASSAY USING HIGH-

PERFORMANCE LIQUID CHROMATOGRAPHY

Although ligands studied in Chapter 3 bind the IspF enzyme, whether the compounds inhibit IspF activity is not known. Therefore, the development of an assay to study the effect of the compounds on enzyme activity, as well as SAR was essential. Bitok, et. al, previously reported a high performance liquid chromatography (HPLC) assay to analyze enzymatic activity of IspF (Figure 4-1).81 This assay monitors the cytosine chromophore that is present in both the substrate, 4-diphosphocytidyl-2-C-methyl-D-erythritol 2-phosphate (CDP-ME2P) and by- product, cytidine monophosphate (CMP). Due to their identical chromophores, it is necessary to separate the analytes of interest using an ion-pairing reagent. This chapter describes the synthesis of the IspF substrate, CDP-ME2P, using the starting substrate MEP and two MEP pathway enzymes IspD and IspE (Figure 4-2). IspF activity was analyzed in the absence and presence of compounds to determine percent inhibition. Compounds were evaluated at a single concentration

(two-fold higher than the compound’s SPR-derived Kd,app). The data presented below demonstrates that not all ligands that bind IspF inhibit the enzyme. Four of eleven compounds analyzed in the HPLC assay were determined to inhibit IspF by >50%.

135

Figure 4-1. IspF enzymatic reaction to convert CDP-ME to CDP-ME2P. The chromophore of the substrate, CDP-ME2P (1), and the by-product, CMP (3), are monitored to determine the rate of the IspF reaction at 272 nm.

Methods

General methods for preparation of 4-diphosphocytidyl-2C-methyl-D-erythritol 2-phosphate

(CDP-ME2P) and reaction rate determination of Burkholderia pseudomallei IspF were followed as stated in Chapter 2.

HPLC-based IspF Activity with Inhibitors

The reaction rate determination of Burkholderia pseudomallei IspF were followed as stated in

Chapter 2 in the presence of IspF inhibitors. The final reaction mixtures contained 5% dimethyl sulfoxide, which was used to facilitate small molecule solubility. Inhibitors were tested at a concentration that was twice the measured Kd,app value to ensure the compound was in excess of the protein. Percent inhibition was calculated using the Equation 4-1, where Ro is the rate of

81 reaction in the absence of inhibitor and Ri is the rate in the presence of inhibitor. Samples ran in

Caren Freel Meyers lab at John’s Hopkins School of Medicine.

136

% Inhibition = x 100 Eq. 4-1

Results and Discussion

Formation of 4-diphosphocytidyL-2C-methyl-D-erythritol 2-phosphate (CDP-ME2P)

The enzymatic synthesis of the IspF CDP-ME2P substrate was performed to allow analysis of IspF enzymatic activity and its inhibition by inhibitors. The substrate, CDP-ME2P, was formed via two enzymatic reactions from the starting material, 2C-methyl-D-erythritol 4- phosphate (MEP), which was organically synthesized in a 7-step reaction sequence by the Freel-

Meyers Lab (Figure 4-2).141 The reaction was monitored using HPLC following absorbance at

272 nm. Both cytosine and adenosine nucleobases are observed in the chromatogram (Figure 4-

3). The IspD reaction was monitored for the conversion of cytidine triphosphate, CTP (Sigma), which transfers the cytidine monophosphate group to MEP, to form 4-diphosphocytidyl-2C- methylerythritol, CDP-ME. The IspD reaction took 128 min for complete conversion to CDP-

ME (Figure 4-4). The IspE reaction was performed in flux by addition of pyruvate kinase and its substrate phosphoenol pyruvate (PEP) to regenerate ATP within the reaction. The reaction was monitored for an additional 182 minutes during which time CDP-ME was converted to CDP-

ME2P (Figure 4-5).

137

Figure 4-2. The IspF substrate, CDP-ME2P, was generated using two enzymatic reactions in the MEP pathway starting with synthesized 2C-methyl-D-erythritol 4-phosphate (MEP). The reactions were completed in flux to ensure the IspD product was formed before addition of IspE.

Figure 4-3. Chromatogram of cytidine and adenosine nucleotides elution profile. Elution was performed on an Altima C18 3µ, 53 mm x 7 mm Rocket column at 1.5 mL/min monitored at 272 nm. A linear gradient from 0-33.5% buffer B was used to elute the analytes of interest.

138

Det 168-272nm Det 168-272nm Det 168-272nm Det 168-272nm SS-02-53 5 mM CDPME2P Synthesis IspD 128m SS-02-53 5 mM CDPME2P Synthesis IspD 51m SS-02-53 5 mM CDPME2P Synthesis IspD 74m Area 1200

1000 51 min

800 74 min

600 mAU 105 min 400

200

128 min

8611 96 7817 0

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 Minutes Figure 4-4. The IspD reaction was monitored from 51 to 128 minutes. Conversion of CTP (tr = 3.10 min) to CDP-ME (tr = 2.15 min) was monitored. Slight degradation of CDP-ME to CDP (tr = 2.00 min) can be observed in the chromatogram.

139

4000 182m IspDE

3500 165m IspDE

3000 138m IspDE

118m IspDE 2500 96m IspDE 2000 mAU 75m IspDE

1500 44m IspDE

1000 15m IspDE

128m IspD 500

15m IspD

49187

16839

10124 203 0

1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6 3.8 4.0 4.2 4.4 4.6 4.8 5.0 Minutes Figure 4-5. Formation of CDP-ME2P monitored over time by following IspD and IspE reaction progress at 272 nm. Listed times describe time post enzyme addition first for IspD (IspD), then for IspE (IspDE). Conversion of CDP-ME (tr = 2.10 min) to CDP-ME2P (tr = 3.10 min) can be observed in the chromatogram with ADP (tr = 3.80 min) and ATP (tr = 4.20 min).

140 Determination of Burkholderia pseudomallei IspF Enzymatic Activity

After synthesis of the IspF substrate (CDP-ME2P) was completed and confirmed, the rate of enzymatic activity for the Burkholderia pseudomallei IspF could be determined. The reaction was initiated using 500 nM, 50 nM, and 2 nM IspF, which revealed 2 nM enzyme with 100 M of substrate, where optimum conditions. The enzymatic reaction was found to be >90% complete after 0.5 min using 50 nM and 500 nM IspF to initiate the reaction. Under these conditions the rate of the enzymatic reaction could be measured from 0 to 45 minutes. Aliquots of the reaction were quenched at 0, 0.5, 1, 2, 4, 6, 10, and 45 minutes to determine the concentration of CMP at each time point (Figure 4-6, Figure 4-7). The reaction was 81% complete after 45 minutes with an observed rate of 4.3  0.3 mol/min between 2 and 10 minutes (Figure 4-8). The rate of the enzyme in absence of inhibitor was then used to calculate the percent inhibition.

141

Figure 4-6. Enzymatic activity of BpIspF at 2 nM concentration. The reaction was monitored from 0 to 45 minutes and determined to be >95% complete.

142

1800 45 min

1600

1400 10 min

1200

1000 6 min

800 mAU

600 4 min

400

200 2 min

267328

40918

84352 20087 0

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6 3.8 4.0 4.2 4.4 4.6 Minutes Figure 4-7. Time course of CMP formation by BpIspF. The progress of the conversion of CDP- ME2P (tr = 3.0 min) to CMP (tr = 1.1 min) by BpIspF is demonstrated in this overlay of chromatograms from 2 to 45 minutes. Reaction progress was monitored at 272 nm.

143

Figure 4-8. Analysis of BpIspF rate through linear fit. A rate of 4.3  0.3 M/min was calculated.

144 Inhibition of Burkholderia pseudomallei IspF

Single point screens of inhibitors were carried out to discover those that bound with Kd,app value ≤210 M (Chapter 3, Table 4-1). Compounds were screened at concentrations that were two times the Kd,app value, unless solubility was a limiting factor. Reaction time points were taken at 2, 4, and 6 minutes to determine the rate of the inhibited reaction, which was fit using a linear regression. To assess the extent of the reaction a sample was also taken at 45 minutes.

There were a total of eleven compounds screened in the HPLC assay to determine percent inhibition (Table 4-1).

145 Table 4-1. List of compounds screened in the HPLC enzyme inhibition assay displaying those with a Kd,app values of ≤210 μM.

Compound Structure Solubility SPR Kd,app Enzyme Inhibition Screen HGN-0003 N >200 μM 60 μM 120 μM

NH2

N O OH HGN-0247 N 1000 μM 210 μM 400 μM

HO N S CH HGN-0289 3 50 μM 210 μM 50 μM O N CH3 N S N CH 3 H C HGN-0321 3 O 100 μM 25 μM 50 μM CH3 O N N S N N HGN-0090 OH 200 μM 22 μM 50 μM

N Cl N HGN-0093 OH 200 μM 83 μM 200 μM

N N F HGN-0094 OH 200 μM 21 μM 50 μM

N N Cl HGN-0106 OH 200 μM 96 μM 200 μM

F N N

HGN-0109 OH 200 μM 45 μM 100 μM

N F N

HGN-0107 OH 200 μM 51 μM 100 μM

N N HGN-0201 OH 200 μM 78 μM 160 μM

N N N S O O

N CH3 H C 3

146 A total of four of the eleven compounds that were screened had greater than 50% inhibition of Burkholderia pseudomallei IspF. The compound HGN-0003 (L-tryptophan hydroxamate) which was previously identified to bind E. coli IspF106 and used as a control for the binding assays, was run at a concentration of 120  to determine percent inhibition (Figure

4-9). The compound exhibited 55% inhibition with a rate of 1.74  0.4 mol/min (Figure 4-10).

The observed inhibition of BpIspF by L-tryptophan hydroxamate supports the notion that compounds may bind across numerous IspF containing species.115,60,142 The three remaining compounds that showed >50% enzyme inhibition were designed to bind to the active site zinc, with two analogs based on the imidazole series and one on the imidazothiazoles. The compound

HGN-0107 was run at 100 for its single point inhibition screen (Figure 4-11). The compound displayed 68% inhibition with a rate of 1.2 mol/min (Figure 4-12). A related analog HGN-0106 displayed 49% inhibition, suggesting the additional methylene group in compound HGN-0107 allows the ligand to interact more strongly with the electron donating group in the active site of

BpIspF to a greater extent than HGN-0106. In the imidazole analogs, placing the fluoro- substituent to the meta-position (HGN-0109) decreased the inhibitory effect while having it at the para-position (HGN-0093) enhanced the activity of the enzyme. By comparison, having a chloro- substituent on either the meta- or para- position (HGN-0090 and HGN-0094) did not show significant inhibition of BpIspF. The fluorescent compound HGN-0201 was analyzed for inhibition at 160 (Figure 4-13) and was determined to be the best inhibitor of IspF discovered in this study, displaying 75% inhibition with a rate of 0.97 mol/min (Figure 4-14).

Although the rationale for synthesis of the imidazole series was based upon a hypothesis of binding through the imidazole nitrogen, the fluorescent compound HGN-0201 also has the

147 potential to bind through the sulfonamide. The results need further examination to determine the exact mode of binding.

A total of three compounds in the imidazothiazole series: HGN-0245, HGN-0289, and

HGN-0321 were analyzed in the HPLC assay with HGN-0289 displaying an inhibitory effect

>50% on BpIspF. HGN-0289 was tested at 50 due to its lower solubility of the compound

(Figure 4-15). The compound displayed 57% inhibition with a rate of 1.6 mol/min (Figure 4-

16). The original hit molecule from the STD-NMR screen HGN-0247 also displayed a weak inhibitory effect at 400  with 32% inhibition, indicating a slightly bulkier functional group on the thiazole ring increases activity. Finally, the rate observed for HGN-0321 was the same as the enzyme in the presence 5% DMSO, thereby demonstrating that it is not a BpIspF inhibitor. The substituents on the imidazothiazole ring are much bulkier than in HGN-0289 and these could be causing steric clashes with active site side chains thereby preventing binding and leading to its observed inactivity in the assays.

Figure 4-9. Determination of BpIspF inhibition by L-tryptophan hydroxamate. Aliquots of reaction were measure from 2 to 45 minutes (bottom to top). Conversion of CDP-ME2P (tr = 3.0 min) to CMP (tr = 1.1 min) was monitored at 272 nm.

148

149

Figure 4-10. Comparison of BpIspF rate in absence and presence of HGN-0003 (L-tryptophan hydroxamate). The rate was determined to be 1.74  0.4 µmol/min at 120 

Figure 4-11. Determination of BpIspF inhibition by 4-(1-(2-fluorophenethyl)-1H-imidazol-5-yl)phenol (HGN-0107). Aliquots of reaction were measure from 2 to 45 minutes (bottom to top). Conversion of CDP-ME2P (tr = 3.0 min) to CMP (tr = 1.1 min) was monitored at 272 nm.

150

151

Figure 4-12. Comparison of BpIspF rate in absence and presence of HGN-0107, 4-(1-(2- fluorophenethyl)-1H-imidazol-5-yl)phenol. The rate was determined to be 1.22 µmol/min at 100 M.

Figure 4-13. HGN-0201 Determination of BpIspF inhibition by 5-(dimethylamino)-N-(2-(5-(4-hydroxyphenyl)-1H-imidazol-1- yl)ethyl)naphthalene-1-sulfonamide, HGN-0201. Aliquots of reaction were measure from 2 to 45 minutes (bottom to top). Conversion of CDP-ME2P (tr = 3.0 min) to CMP (tr = 1.1 min) was monitored at 272 nm.

152

153

Figure 4-14. Comparison of BpIspF rate in absence and presence of HGN-0201, 5- (dimethylamino)-N-(2-(5-(4-hydroxyphenyl)-1H-imidazol-1-yl)ethyl)naphthalene-1- sulfonamide. The rate was determined to be 0.97 µmol/min at 160 

Figure 4-15. Determination of BpIspF inhibition by N-isopropyl-3-methylimidazo[2,1-b]thiazole-2-carboxamide, HGN-0289. Aliquots of reaction were measure from 2 to 45 minutes (bottom to top). Conversion of CDP-ME2P (tr = 3.0 min) to CMP (tr = 1.1 min) was monitored at 272 nm.

154

155

Figure 4-16. Comparison of BpIspF rate in absence and presence of HGN-0289, N-isopropyl-3- methylimidazo[2,1-b]thiazole-2-carboxamide. The rate was determined to be 1.64 µmol/min at 50 M.

156 Bitok, et al. has previously shown that the substrate, MEP, is able to enhance E. coli IspF activity.70 Therefore, BpIspF was also screened in the presence of 500  2C-methyl-D- erythritol 4-phosphate (MEP) to examine potential enhancement of activity due to the presence of the IspD substrate. The rate of conversion of CDP-ME2P to CMP by BpIspF was found to be reduced to 3.3 mol/min in the presence of MEP (13% reduction). This value is within experimental error of native BpIspF activity. Therefore, in this study, enhanced activity was not observed in the presence of MEP.

The enzyme activity of BpIspF was also examined in the presence of DMSO, alone.

Since all reaction mixtures contained 5% DMSO to facilitate small molecule dissolution, it was important to demonstrate that this component of the reaction buffers was not influencing activity measurements. The activity of BpIspF in the presence of 5% DMSO was found to be 3.8

mol/min. This value is within experimental error of native activity and, therefore, DMSO was demonstrated to have had no effect upon the experimental results.

Overall, out of the eleven compounds screened in the HPLC assay, four compounds displayed >50% inhibition of Burkholderia pseudomallei IspF at concentrations two-fold above the Kd,app value (Table 4-2). The result lends promise as these protocols and methods are further refined.

Table 4-2 continued 157

Table 4-2. Summary of BpIspF enzymatic rates in the presence of compounds.

Compound Structure Rate (µmol/min) % Inhibition

DMSO 3.82  0.3 -

NH HGN-0003 2 1.74  0.4 54.5  0.2

N O OH

N HGN-0247 2.68 29.8 HO N S

CH3 O N HGN-0289 1.64 CH3 57.1 N S N

CH3

H3C O

CH3 O N HGN-0321 3.71 - N S N

HGN-0090 2.59 32.2 N Cl N

HGN-0094 3.15 - N N

Continued on following page Table 4-2 continued 158

HGN-0106 F 1.94 49.2 N N

HGN-0109 2.91 - N F N

HGN-0093 5.38 -

N N

HGN-0107 F 1.22 68.1

N N

HGN-0201 0.97 74.6

N N N S O O

N CH3

H3C

Continued on following page 159 Conclusions

An HPLC-based enzyme inhibition assay was successfully designed and implemented for determination of single point, percent inhibition values. Using this assay, eleven potential inhibitors were tested and four were found to generate ≥50% inhibition. The data also provided insight into an initial SAR of the imidazole and imidazothiazole series. In the imidazole series, ortho fluoro- substituent was found to be more potent than the chloro- substituent series.

Furthermore, extending the substituent with a two-carbon linker unit (HGN-0107) was found to be more potent than a one carbon linker (HGN-0106). The inhibition studies were shown to correlate to the FTS assay where the ortho fluoro- substituent displayed at shift of 3.2°C, for both the benzyl and phenethyl analogs, and the ortho chloro- substituent had a shift of 2.5°C, benzyl analog and 1.0°C, phenethyl analog. The fluorescent compound, HGN-0201, was determined to have the best inhibitory effect observed for BpIspF with 75% inhibition, suggesting that the bulkier functional group may have increased interactions with the enzyme. The molecule could also be offering an alternative mode of binding through the sulfonamide rather than the imidazole nitrogen and this moiety may even increase binding affinity to the active site zinc.

HPLC assay data for the imidazothiazole series are in agreement with SAR and SPR data

(Chapter 3), confirming that bulky functional groups on the thiazole ring (e.g., HGN-0321) are likely not accommodated in the binding pocket. The enzyme inhibition data were based upon single point runs, the lack of inhibition by HGN-0321, which had a KD of 25 µM, could be within the error of the HPLC assay. Substitution of the ester linkage found in the original fragment hit (HGN-0247 – 2.68 µmol/min) to and amide (HGN-0289 – 1.64 µmol/min) generated greater than a two-fold increase in inhibition activity (57% inhibition at 50 µM) The inhibition data showed good correlation to SPR KD values when comparing compounds that

160 were tested at the same concentration, with the exception of HGN-0093 which displayed enhance

IspF activity. The enhanced IspF activity with HGN-0093 would need to be studied further to determine the error associated with the assay to conclude if there is actual enchantment of activity. Based upon the findings of this study, the imidazole series appear to be the best inhibitors of BpIspF tested to date.

CHAPTER 5

STRUCTURAL ANALYSIS OF INHIBITORS USING CRYSTALLOGRAPHY

The crystal structure of Burkholderia pseudomallei IspF is well documented in the protein data bank (PDB), with over ten published structures.102 These IspF structures are either in complex with low molecular weight ligands or are ligand-free. High resolution X-ray crystal structures of inhibitor-enzyme complexes reveal atomic-level insight into the binding interactions. These structures help illuminate existing interactions that drive the binding event, as well as provide a path for potential design efforts to enhance ligand binding. Inhibitors that undergo structure- guided modifications can be tested for increased or decreased potency in any number of assays or biophysical techniques including additional crystal structures. It is this iterative process of structure determination, insight, synthesis, assay, structure determination that constitutes the basis of so-called rational drug design. To gain structural insight in the design of novel, high- potency BpIspF inhibitors, a structural biology component was explored. As a result, two high resolution inhibitor-enzyme complex structures were determined for BpIspF bound to HGN-003 and HGN-0289 (Figure 5-1).

162

Figure 5-1. Structures of L-tryptophan hydroxamate, HGN-0003, and N-isopropyl-3- methylimidazo[2,1-b]thiazole-2-carboxamide, HGN-0289.

Methods

General methods for Crystallization of Burkholderia pseudomallei IspF, X-ray Diffraction, and

X-Ray Structure Solution were followed as stated in Chapter 2.

Co-Crystallization of Burkholderia pseudomallei IspF Inhibitors

Attempts to co-crystallize BpIspF with various inhibitors were undertaken. Briefly, inhibitors were dissolved at 100 mM in DMSO. A total of 2.0 µL of DMSO solubilized inhibitor was incubated with 8.0 µL of BpIspF at 32 mg/mL at 4oC for 2 hours. Samples were then centrifuged at 14,000 rpm for 20 min at 4oC to remove potential precipitation that occurred during incubation. Crystallization trials were performed manually using standard hanging drop vapor diffusion methods in 24-well trays. Crystallization drops were composed of 1.0 µL protein-inhibitor solution (~20 mM inhibitor, 25.6 mg/ml BpIspF in SE buffer) mixed with 1.0

µL crystallization buffer (100 mM Tris, pH 7.2, 200 mM NaCl, 20% PEG 4000, 5 0 mM

163 inhibitor, 25.6 mg/ml B formate, pH 5.9, 20% PEG 3350). Crystallization tray well reservoirs contained 1000 µL of crystallization buffer.90

Inhibitor Soaks with Burkholderia pseudomallei IspF

Attempts were also made to soak inhibitors into pre-formed, apo crystals of BpIspF.

Inhibitors were dissolved in methanol at 25 mM for these crystal soaks. For any given compound, a 1.0 µL drop of inhibitor was added to a coverslip and allowed to evaporate to dryness. The compound was then resuspended in 2.0 µL of crystallization buffer (100 mM Tris, pH 7.2, 200 mM NaCl, 20% PEG 4000, 5 µM ZnCl2 or 200 mM C2H2MgO4∙2H2O, pH 5.9, 20%

PEG 3350) and two to three apo crystals were transferred to the drop. The coverslip was inverted over a well on a crystal tray and sealed using vacuum grease. Well reservoirs contained 1000 µL of crystallization buffer. Crystal trays containing these soaking experiments were incubated for at least three weeks to maximize the possibility that compound would bind BpIspF.90

Results and Discussion

Burkholderia pseudomallei IspF Complex with L-tryptophan hydroxamate

L-tryptophan hydroxamate has been shown to be one of the highest affinity inhibitor of

BpIspF. The observed dissociation constant of the ligand was 30 µM (ITC-Chapter 3) and it showed an inhibition of enzyme activity of 55% at 120 µM (HPLC-Chapter 4). While this

164 106 compound was previously reported to be an EcIspF inhibitor (Kd,app of 2 µM), the mode of target binding was unknown.

The crystal structure of BpIspF with HGN-0003, determined at 1.47 Å, revealed a non- traditional mode of binding. HGN-0003 interacts with the active site zinc ion through the primary amine as well as the amide nitrogen (Figure 5-2). The traditional mode of hydroxamic acid binding to zinc has been shown to be mediated through either the carbonyl or hydroxyl group with tetrahedral coordination. There are also some structures in which the hydroxamic acid moiety of a compound bind the metal through both the carbonyl and hydroxyl group via trigonal bipyramidal geometry (Figure 5-3).143,144,145,146 The ligand HGN-0003 was refined to an occupancy of 1.0 with a B-factor of 25 Å2 (Figure 5-4). The bonding distance of the primary amine to the zinc ion is ~1.9 Å, similar to the zinc bonding distances with Asp10, His12, and

His44. The ~2.9 Å bonding distance of the amide nitrogen to the zinc is slightly long for ideal trigonal bipyramidal coordination but it is within measurement error. Finally, a water-mediated hydrogen bond was observed between the backbone amide nitrogen of His36 and the hydroxyl group of HGN-0003.

165

Figure 5-2. The Burkholderia pseudomallei IspF crystal structure with HGN-0003 displays an unusual mode of binding for the hydroxamic acid functional group to zinc. The ligand, HGN- 0003 (magenta) binds the zinc (blue) by filling out a trigonal pyramidal geometry coordination sphere. The three BpIspF monomers in the trimeric enzyme are colored cyan (with bound ligand), green, and light cyan.

166

Figure 5-3. Known modes of hydroxamic acids binding to metal ions. Hydroxamic acids are observed to chelate to metals through a) carbonyl and hydroxyl group (PDB: 4R7M, cyan),146 as well as b) the carbonyl (PDB:5TLN, green).143

Figure 5-4. Electron density from the active site of BpIspF in complex with HGN-003. Clear binding of the compound (magenta) to the zinc (blue sphere) is evident in this view. Electron density is contoured at 1.5σ

167 To help reveal potential structural differences between BpIspF and EcIspF that may lead to the observed differences in binding affinities, the crystal structure of native E. coli IspF

(1JY8) was overlaid with the current structure of the L-tryptophan hydroxamate-B. pseudomallei

IspF complex (Figure 5-5). Based on the comparison of the two structures, there are no obvious differences between the crystal structures to explain the difference in observed binding to HGN-

0003 (RMSDC = 0.256 Å).

168

Figure 5-5. Active site overlap of the structures of Burkholderia pseudomallei IspF in complex with HGN-003 and Escherichia coli IspF. The EcIspF (green) crystal structure was aligned to BpIspF (cyan) in complex with HGN-0003 (magenta) to illustrate potential structural differences between the two enzymes as they relate to ligand binding residues.

169 Burkholderia pseudomallei IspF Complex with HGN-0289

To guide the synthesis of ligands with increased affinity for BpIspF, the ligand, HGN-

0289, was co-crystallized with the enzyme (Figure 5-6a). HGN-0289 has an N-isopropyl-3- methylimidazo[2,1-b]thiazole-2-carboxamide core with an isopropyl amide group attached to the thiazole ring. The ligand was designed to bind to the zinc through the imidazole nitrogen. HGN-

0289 binds IspF with a dissociation constant of 210 µM (SPR- Chapter 3) and displays 57% inhibition at 50 µM (HPLC assay-Chapter 4). Structural data, determined at 1.51 , confirmed binding of the imidazole nitrogen to the zinc with the isopropyl amide functional group pointing into the lipophilic pocket (Figure 5-6b). The crystal structure indicates that the addition of a slightly bulkier functional group off the thiazole ring, in future synthetic efforts, could increase interactions with the lipophilic pocket, something to be considered in future inhibitor design.

Furthermore, the addition of functional groups on C2 or C3 of the imidazole ring may lead to higher affinity inhibitors due to increasing interactions in the lipophilic pocket. The ligand HGN-

0289 refined to an occupancy of 0.81 with a B-factor of 30 Å 2 (Figure 5-7). The ligand coordinates with the zinc tetrahedrally through the imidazole nitrogen with a bond length of 1.9

Å. The ligand HGN-0289 was synthesized as an analog of a fragment identified from a NMR

STD screen, FOL535 (PDB: 3K14).

As shown in Figure 5-8, an overlay of the crystal structures of BpIspF with HGN-0289 and the fragment hit from the STD-NMR screen, FOL535, shows both ligands bind to the zinc ion with a tetrahedral coordination (RMSD = 0.071 Å). For both structures, the lipophilic substituent points into the lipophilic pocket, as would be expected. This suggests that extending

170 into this pocket with a slightly bulkier lipophilic group may increase the affinity of the ligand by providing more and better interactions with the enzyme via the sub-pocket.

171

Figure 5-6. Ribbon representations of the co-crystal structure of BpIspF in complex with HGN- 0289. (A) Overview of the BpIspF trimer. HGN-0289 is shown bound to the cyan monomer. (B) Close up of the active site providing details of HGN-0289 binding. HGN-0289 is pictured in magenta and the zinc as a blue sphere. The tight binding interaction (1.9 ) between the imidazole nitrogen of HGN-0289 and the zinc is evident in this view. Some BpIspF residues involved in ligand binding are also shown (and labeled). Worthy of specific mention is the orientation of the HGN-0289 propyl amide functional group with respect to the lipophilic pocket (residues Phe63, Ile59, and Phe70 of the pocket are identified).

172

Figure 5-7. Electron density for the active site of BpIspF in complex with HGN-0289. HGN-0289 (magenta) binds to zinc (gray) of BpIspF (yellow) via N7 of the imidazothiazole ring. The electron density map displays both positive (green) and negative (red) in addition to density which is occupied (blue) by BpIspF and HGN-0289.

173

Figure 5-8. Overlay of active site atoms for BpIspF in complex with HGN-0289 and the ligand from a published BpIspF co-crystal structure (PDB: 3K14). 90 The published ligand structure was chosen for this comparison because it is the same molecule that was discovered as a BpIspF fragment hit by STD-NMR screening (FOL535).

174 Conclusion

The crystal structures of BpIspF in complex with HGN-0003 and HGN-0289 provide insight to help drive future design strategies for both the tryptophan and N-isopropyl-3- methylimidazo[2,1-b]thiazole-2-carboxamide scaffolds. The information could help lead to the development of inhibitors with an increased affinity to the target Burkholderia pseudomallei

IspF. The structure of HGN-0003 indicates that extending functional groups off the indole ring nitrogen or C2 into the ribose pocket may allow new interactions with the enzyme. In addition, lipophilic substituents at C7 of the indole ring of tryptophan are hypothesized to extend into the lipophilic pocket and increase the affinity of the ligand for BpIspF. Functional groups which have electron donating and accepting groups that extend into the ribose and cytidine pocket are hypothesized to lead to an increase in affinity for BpIspF, which was observed in HGN-0321

(Table 3-3). The development of methods to screen and produce BpIspF/inhibitor co-crystal structures is a promising start for the use of X-ray crystallography in guiding the synthesis of novel, high affinity compounds that can inhibit this important target.

Figure 5-8 Structure of HGN-0321, 3-methyl-5-(4-methylbenzoyl)-N-(pyridin-3- ylmethyl)imidazo[2,1-b]thiazole-2-carboxamide, which showed a binding affinity of 25 M by SPR (Table 3-3).

CHAPTER 6

OVERALL CONCLUSIONS

According to the World Health Organization in 2013, there were 1.5 million deaths due to tuberculosis alone.147 In addition, the number of multi-drug resistant tuberculosis cases has significantly increase from 31% in 1993 to 88% in 2014.148 Therefore, the demand for the development of novel anti-infective agents has significantly increased in conjunction with the decreasing number of FDA-approved anti-bacterials.6,7 As a result, there is a need to determine new pathways to target bacteria and/or to develop novel classes of anti-infective agents. The goal of this project was to identify potential small molecule inhibitors of Burkholderia pseudomallei through targeting the fifth enzyme in the MEP pathway, IspF. Due to a 30% sequence identity of

IspF across eukaryotic and prokaryotic bacterial species,90 inhibitors of Burkholderia pseudomallei could have the potential to target several infectious disease causing organisms.

The current methods to analyze inhibition of the IspF enzyme are based upon a labor-intensive

HPLC assay or a highly coupled plate-based assays.

To open new opportunities for the discovery of novel IspF inhibitors, the initial study developed a new assay to assess the potency of potential small molecule inhibitors. Biophysical characterization of Burkholderia pseudomallei IspF revealed the enzyme was extremely thermostable with a melting temperature >100°C. Therefore, a protein engineering approach was used to develop a thermal shift assay. Single-site IspF variants were generated with the aim of destabilizing the interface of IspF’s three subunits. Several variants destabilized IspF

176 significantly (by at least 30ºC), which allowed thermal melt analysis of small molecule inhibitors to assess apparent binding affinity. Of equal importance, enzymatic activity of variant

Q151E was not affected by the mutation. The development of such an assay provided an efficient, convenient method for rank ordering the apparent affinities. The thermal shift assay was used to evaluate four chemical series of compounds in order to narrow down the number of compounds for subsequent determination of the dissociation constants.

The binding affinities of several groups of potential IspF inhibitors including the fusion series, imidazole series, imidazothiazole series, and tryptophan analogs were assessed using SPR and ITC. The fusion compounds, which were designed to link the cytidine and zinc binding pockets of the IspF active site, were found to have the same affinity for the target as the control,

CDP. The fusion compounds were further examined by in silico studies which were able to predict the rank order of affinities that were observed by SPR. By further utilizing in silico docking methods, the mode of ligand binding was predicted for both HGN-0006 and HGN-0009.

ITC binding studies were conducted on two structurally similar fusion compounds, HGN-0006 and HGN-0009. These compounds differ by only a single methylene group. The results show that binding of both compounds was enthalpically-driven. The addition of a methylene group decreased the enthalpic contribution by 2.66 kcal/mol. It is plausible that this decrease is coming from a slight change in the orientation of the zinc binding moiety relative to the zinc along with the entire compound shifting to accommodate the presence of the extra methylene group in the active site.

Tryptophan analogs were also a focus of this study because of their historical precedent as metal binders. L-tryptophan hydroxamate was previously shown to bind E. coli IspF. Here, both L- and D-tryptophan hydroxamate enantiomers, as well as both tryptophan enantiomers,

177 were studied by ITC and SPR. The L-tryptophan hydroxamate was observed to bind with dissociation constants of 30 ± 2 M by ITC and 60  3 M by SPR. Whereas the D-enantiomer was found to bind to BpIspF slightly weaker (64 7 M by ITC). The tryptophan enantiomers were found to bind with weak affinities, Kd >1000 M, indicating the hydroxamic acid group is important for binding to BpIspF. The thermal shift assay also showed good correlation to SPR and ITC binding affinity values for all compounds in the tryptophan series (Table 2-5).

STD-NMR experiments were used to identify new fragment hits that bind BpIspF. These new hits served as a starting point for more detailed binding characterization using the primary techniques and protocols developed in this study. A series of analogs were studied for binding by

SPR based on FOL955 that were established to have Kd,app ranging from 20 M to greater than

2000 M. Compounds containing a fluorine substituent were observed to bind more tightly than the chlorine-substituted analogs. In particular, a fluorescently labeled derivative, HGN-0201, had a Kd,app of 78 M. In addition to being a potentially exciting lead for further development, the compound could potentially be used in fluorescence polarization assays. Analogs of fragment hits FOL535 and FOL717 were determined to have Kd,app values ranging from 25 to 1200 M.

One compound of particular interest that was developed from FOL535, HGN-0321, had one of the better Kd,app values observed in this study: 25 M. The synthesis of more analogs of HGN-

0321 could lead to more exciting and potent inhibitors of BpIspF. The thermal-shift assay

(Chapter 2) was able to provide insight about relative affinities of the compounds, but it unfortunately did not have a high degree of correlation to dissociation constants determined by other, more reliable methods used in this study. In addition, an FP was used to accurately measure the affinity of HGN-0201 for BpIspF: 102 M. However, the development of a

178 competitive assay was not effective due to interference of the fluorophore signal by the ligands that compete for active site binding in BpIspF. Finally, those compounds that were found to have dissociation constants ≤200 M by SPR were also analyzed in the enzyme inhibition assay.

An enzyme inhibition assay was developed and used to screen eleven compounds for single-point, percent inhibition determination. A total of four of the eleven compounds that were tested using the assay displayed ≥50% inhibition. The fluorescent compound, HGN-0201, was determined to have the best inhibitory effect on BpIspF with 75% inhibition. This suggests that the bulkier functional group may increase interactions with the enzyme, thereby providing another starting point for future inhibitor design. Additionally, the molecule has the potential to have an alternative mode of binding through the sulfonamide, rather than via the expected imidazole nitrogen mechanism, which could result in an increased affinity for the zinc.

Crystal structures for both HGN-0003 and HGN-0289 BpIspF complexes provided potential insight into future design strategies of both the tryptophan and N-isopropyl-3- methylimidazo[2,1-b]thiazole-2-carboxamide scaffolds. The information should help lead to the development of inhibitors with increased affinities for the target Burkholderia pseudomallei

IspF. The structure of HGN-0003 indicates that extending functional groups off the indole ring into the ribose pocket may create new interactions with the enzyme. In addition, electron withdrawing or donating groups on the benzene ring of the tryptophan are hypothesized to affect the affinity of the ligand. Functional groups that extend into the ribose and cytidine pocket could lead to an increase in affinity for BpIspF, as this work has already shown with HGN-0321 (Table

3-3). This study’s structural insight into BpIspF inhibition will help guide the synthesis of novel derivatives of these two chemical classes as well as others.

179 This work started with compounds that had weak mM affinities and little potential for further development and culminated with a scaffold having an encouraging 25 M affinity that can be used to continue the development of novel anti-infective agents. A high-throughput assay was developed to assess relative affinities of IspF inhibitors. This assay can be used to efficiently identify those compounds to be moved forward for binding affinity analysis using SPR and ITC.

This study also delivered a method for reliable Kd,app determination so that only those compounds with Kd,app values ≤200 µM could be identified for assessment with more labor- intensive BpIspF inhibition evaluation. In closing, this work has succeeded in new method development, as well as delivering sufficiently high-affinity, BpIspF-binding compounds. Both aspects will aid the design of next-generation novel compounds that target BpIspF with drug-like binding affinities, which may ultimately be used as new antibiotic treatments.

REFERENCES

(1) Hughes, D. (2003) Exploiting genomics, genetics and chemistry to combat antibiotic resistance. Nat. Rev. Genet. 4, 432–441.

(2) Buchanan, N. P., Hott, J. M., Cutlip, S. E., Rack, A. L., Asamer, A., and Moritz, J. S. (2008) The Effects of a Natural Antibiotic Alternative and a Natural Growth Promoter Feed Additive on Broiler Performance and Carcass Quality. J. Appl. Poult. Res. 17, 202–210.

(3) Ventola, C. L. (2015) The antibiotic resistance crisis: part 1: causes and threats. P T Peer-Rev. J. Formul. Manag. 40, 277–283.

(4) Krans, B. The Decline of Antiobiotic Development. Healthline.

(5) Coates, A. R., Halls, G., and Hu, Y. (2011) Novel classes of antibiotics or more of the same?: New antibiotic classes are urgently needed. Br. J. Pharmacol. 163, 184–194.

(6) Healthline News. Antibiotic Resistance: Why Our Best Medical Weapon is Losing Its Edge.

(7) Spellberg, B., Powers, J. H., Brass, E. P., Miller, L. G., and Edwards, J. E. (2004) Trends in antimicrobial drug development: implications for the future. Clin. Infect. Dis. Off. Publ. Infect. Dis. Soc. Am. 38, 1279–1286.

(8) Walsh, C. (2003) Where will new antibiotics come from? Nat. Rev. Microbiol. 1, 65–70.

(9) Kohanski, M. A., Dwyer, D. J., and Collins, J. J. (2010) How antibiotics kill bacteria: from targets to networks. Nat. Rev. Microbiol. 8, 423–435.

(10) Falconer, S. B., Czarny, T. L., and Brown, E. D. (2011) Antibiotics as probes of biological complexity. Nat. Chem. Biol. 415–423.

(11) Wright, G. D. (2010) Q&A: Antibiotic resistance: where does it come from and what can we do about it? BMC Biol. 8, 123.

(12) Jovetic, S., Zhu, Y., Marcone, G. L., Marinelli, F., and Tramper, J. (2010) β-Lactam and glycopeptide antibiotics: first and last line of defense? Trends Biotechnol. 28, 596–604.

(13) Oeppen, J. (2002) DEMOGRAPHY: Enhanced: Broken Limits to Life Expectancy. Science 296, 1029–1031.

(14) Soares, G., Figueiredo, L., Faveri, M., Cortelli, S., Duarte, P., and Feres, M. (2012) Mechanisms of action of systemic antibiotics used in periodontal treatment and mechanisms of bacterial resistance to these drugs. J. Appl. Oral Sci. 20, 295–309.

181

(15) Ruiz, J. (2003) Mechanisms of resistance to quinolones: target alterations, decreased accumulation and DNA gyrase protection. J. Antimicrob. Chemother. 51, 1109–1117.

(16) Payne, D. J., Gwynn, M. N., Holmes, D. J., and Pompliano, D. L. (2007) Drugs for bad bugs: confronting the challenges of antibacterial discovery. Nat. Rev. Drug Discov. 6, 29–40.

(17) Cross, A. S. (2008) What is a virulence factor? Crit. Care 12, 197.

(18) Nathan, C. (2012) Fresh Approaches to Anti-Infective Therapies. Sci. Transl. Med. 4, 140sr2–140sr2.

(19) Clatworthy, A. E., Pierson, E., and Hung, D. T. (2007) Targeting virulence: a new paradigm for antimicrobial therapy. Nat. Chem. Biol. 3, 541–548.

(20) Fischetti, V. A. (2008) Bacteriophage lysins as effective antibacterials. Curr. Opin. Microbiol. 11, 393–400.

(21) Sacchettini, J. C., and Poulter, C. D. (1997) Creating isoprenoid diversity. Science 277, 1788– 1789.

(22) Penuelas, J., and Munnebosch, S. (2005) Isoprenoids: an evolutionary pool for photoprotection. Trends Plant Sci. 10, 166–169.

(23) Dubey, V. S., Bhalla, R., and Luthra, R. (2003) An overview of the non-mevalonate pathway for terpenoid biosynthesis in plants. J. Biosci. 28, 637–646.

(24) Dubey, V. S. (2002) Mevalonate-independent pathway of isoprenoids synthesis: A potential target in some human pathogens. Curr Sci 83, 685–688.

(25) Gabrielsen, M., Rohdich, F., Eisenreich, W., Gräwert, T., Hecht, S., Bacher, A., and Hunter, W. N. (2004) Biosynthesis of isoprenoids: A bifunctional IspDF enzyme from Campylobacter jejuni. Eur. J. Biochem. 271, 3028–3035.

(26) Rohdich, F., Eisenreich, W., Wungsintaweekul, J., Hect, S., Schuhr, C., and Racher, A. (2001) Biosynthesis of terpenoids. Eur J Biochem 268, 3190–3197.

(27) Tachibana, A., Yano, Y., Otani, S., Nomura, N., Sako, Y., and Taniguchi, M. (2000) Novel prenyltransferase gene encoding farnesylgeranyl diphosphate synthase from a hyperthermophilic archaeon, Aeropyrum pernix: Molecular evolution with alteration in product specificity. Eur. J. Biochem. 267, 321–328.

182 (28) Chaykin, S., Law, J., Phillips, A. H., Tchen, T. T., and Bloch, K. (1958) PHOSPHORYLATED INTERMEDIATES IN THE SYNTHESIS OF SQUALENE. Proc. Natl. Acad. Sci. U. S. A. 44, 998–1004. (29) Miziorko, H. M. (2011) Enzymes of the mevalonate pathway of isoprenoid biosynthesis. Arch. Biochem. Biophys. 505, 131–43.

(30) Zheng, W., Sun, F., Bartlam, M., Li, X., Li, R., and Rao, Z. (2007) The crystal structure of human isopentenyl diphosphate isomerase at 1.7 A resolution reveals its catalytic mechanism in isoprenoid biosynthesis. J. Mol. Biol. 366, 1447–58.

(31) Rohmer, M. (2003) Mevalonate-independent methylerythritol phosphate pathway for isoprenoid biosynthesis elucidation and distribution. Pure Appl Chem 75, 375–388.

(32) Hunter, W. N., Bond, C. S., Gabrielsen, M., and Kemp, L. E. (2003) Structure and reactivity in the non-mevalonate pathway of isoprenoid biosynthesis. Biochem. Soc. Trans. 31, 537–42.

(33) Flesch, G., and Rohmer, M. (1988) Prokaryotic hopanoids: The biosynthesis of the bacteriohopane skeleton. Eur J Biochem 175, 405–411.

(34) Rohmer, M., Knani, M., Simonin, P., Sutter, B., and Sahm, H. (1993) Isoprenoid biosynthesis in bacteria: a novel pathway for the early steps leading to isopentenyl diphosphate. Biochem. J. 295 ( Pt 2), 517–24.

(35) Brrechbuhler-Bader, S., Coscia, C., Loew, P., von Szczepanski, C., and Arigoni, D. (1968) The chemistry and biosynthesis of laganin. Chem Commun Lond. 136–137.

(36) Zhou, D., and White, R. H. (1991) Early steps of isoprenoid biosynthesis in Escherichia coli. Biochem. J. 273 ( Pt 3), 627–634.

(37) Arigoni, D., Sagner, S., Latzel, C., Eisenreich, W., Bacher, A., and Zenk, M. H. (1997) Terpenoid biosynthesis from 1-deoxy-D-xylulose in higher plants by intramolecular skeletal rearrangement. Proc. Natl. Acad. Sci. U. S. A. 94, 10600–5.

(38) Moreno, S. N. J., and Li, Z.-H. (2008) Anti-infectives targeting the isoprenoid pathway of Toxoplasma gondii. Expert Opin. Ther. Targets 12, 253–63.

(39) Rodriguez-Concepcion, M. (2004) The MEP Pathway: A New Target for the Development of Herbicides, Antibiotics and Antimalarial Drugs. Curr. Pharm. Des. 10, 2391–2400.

(40) Brown, A., Eberl, M., Crick, D., Jomaa, H., and Parish, T. (2010) The nonmevalonate pathway of isoprenoid biosynthesis in mycobacterium tuberculosis is essential and transcriptionally regulated by dxs. J Bacteriol 193, 2424–2433.

183 (41) Gabrielsen, M., Bond, C., Hallyburton, I., Hect, S., Bacher, A., Eisenreich, W., Rohdich, F., and Hunter, W. (2004) Hexameric assembly of the bifunctional methylerythritol 4,- cyclodiphosphate synthase and protein-protein associations in the deoxy-xylulose-dependent pathway of isoprenoid precurson biosynthesis. J Biol Chem 279 2, 52753–52761.

(42) Testa, C. A., Lherbet, C., Pojer, F., Noel, J. P., and Poulter, C. D. (2006) Cloning and expression of IspDF from Mesorhizobium loti. Characterization of a bifunctional protein that catalyzes non-consecutive steps in the methylerythritol phosphate pathway. Biochim. Biophys. Acta 1764, 85–96.

(43) Perez, G., Bergua, M., Boronat, A., and Imperial, S. (2010) Cloning and functional characterization of an enzyme from helicobacter pylori that catalyzes two steps of the methylerythritol phosphate pathway for isoprenoid biosynthesis. Biochem Biophys Acta Gen Subj 1800, 919–928.

(44) Lherbet, C., Pojer, F., Richard, S. B., Noel, J. P., and Poulter, C. D. (2006) Absence of substrate channeling between active sites in the Agrobacterium tumefaciens IspDF and IspE enzymes of the methyl erythritol phosphate pathway. Biochemistry (Mosc.) 45, 3548–53.

(45) Hale, I., Neill, P., Berry, N., Odom, A., Sharma, R., and The, M. (2012) The MEP Pathway and the development of inhibitors as potential anti-infective agents. Med Chem Commun 3, 418– 433.

(46) Mao, J., Eoh, H., He, R., Wang, Y., Wan, B., Franzblau, S. G., Crick, D. C., and Kozikowski, A. P. (2008) Structure-activity relationships of compounds targeting mycobacterium tuberculosis 1-deoxy-D-xylulose 5-phosphate synthase. Bioorg. Med. Chem. Lett. 18, 5320–3.

(47) Matsue, Y., Mizuno, H., Tomita, T., Asami, T., Nishiyama, M., and Kuzuyama, T. (2010) The herbicide ketoclomazone inhibits 1-deoxy-D-xylulose 5-phosphate synthase in the 2-C-methyl- D-erythritol 4-phosphate pathway and shows antibacterial activity against Haemophilus influenzae. J. Antibiot. (Tokyo) 63, 583–8.

(48) Smith, J. M., Vierling, R. J., and Meyers, C. F. (2012) Selective inhibition of E. coli 1-deoxy-D- xylulose-5-phosphate synthase by acetylphosphonates(). MedChemComm 3, 65–67.

(49) Jomaa, H., Wiesner, J., Sanderbrand, S., Altincicek, B., Weidemeyer, C., Hintz, M., Türbachova, I., Eberl, M., Zeidler, J., Lichtenthaler, H. K., Soldati, D., and Beck, E. (1999) Inhibitors of the nonmevalonate pathway of isoprenoid biosynthesis as antimalarial drugs. Science 285, 1573–6.

(50) Yajima, S., Hara, K., Iino, D., Sasaki, Y., Kuzuyama, T., Ohsawa, K., and Seto, H. (2007) Structure of 1-deoxy-D-xylulose 5-phosphate reductoisomerase in a quaternary complex with a

184 magnesium ion, NADPH and the antimalarial drug fosmidomycin. Acta Crystallograph. Sect. F Struct. Biol. Cryst. Commun. 63, 466–70. (51) Saadou, I. (2009) A Comparative Assessment of the Efficacy of Fosmidomycin-Clindamycin Versus Sulfadoxine-Pyrimethamine for the Treatment of Children with Uncomplicated Plasmodium Falciparum Malaria. Cliincal Trial, Albert Schweitzer Hospital.

(52) (2015) A Phase IIa Proof of Concept Study to Explore the Efficacy, Tolerability and Safety of Fosmidomycin Sodium when Administered with Piperaquine Tetraphosphate to Adults and Older Children with Acute Uncomplicated Plasmodium Falciparum Malaria. Clinical Trial, Jomaa Pharma GmbH.

(53) Andaloussi, M., Lindh, M., Björkelid, C., Suresh, S., Wieckowska, A., Iyer, H., Karlén, A., and Larhed, M. (2011) Substitution of the phosphonic acid and hydroxamic acid functionalities of the DXR inhibitor FR900098: an attempt to improve the activity against Mycobacterium tuberculosis. Bioorg. Med. Chem. Lett. 21, 5403–7.

(54) Behrendt, C. T., Kunfermann, A., Illarionova, V., Matheeussen, A., Pein, M. K., Gräwert, T., Kaiser, J., Bacher, A., Eisenreich, W., Illarionov, B., Fischer, M., Maes, L., Groll, M., and Kurz, T. (2011) Reverse fosmidomycin derivatives against the antimalarial drug target IspC (Dxr). J. Med. Chem. 54, 6796–802.

(55) Lillo, A. M., Tetzlaff, C. N., Sangari, F. J., and Cane, D. E. (2003) Functional expression and characterization of EryA, the erythritol kinase of Brucella abortus, and enzymatic synthesis of L- erythritol-4-phosphate. Bioorg. Med. Chem. Lett. 13, 737–9.

(56) Witschel, M. C., Höffken, H. W., Seet, M., Parra, L., Mietzner, T., Thater, F., Niggeweg, R., Röhl, F., Illarionov, B., Rohdich, F., Kaiser, J., Fischer, M., Bacher, A., and Diederich, F. (2011) Inhibitors of the herbicidal target IspD: allosteric site binding. Angew. Chem. Int. Ed Engl. 50, 7931–5.

(57) Crane, C. M., Hirsch, A. K. H., Alphey, M. S., Sgraja, T., Lauw, S., Illarionova, V., Rohdich, F., Eisenreich, W., Hunter, W. N., Bacher, A., and Diederich, F. (2008) Synthesis and characterization of cytidine derivatives that inhibit the kinase IspE of the non-mevalonate pathway for isoprenoid biosynthesis. ChemMedChem 3, 91–101.

(58) Crane, C., Kaiser, J., Ramsden, N., Law, S., Rohdich, F., Eisenreich, W., Hunter, W., Bacher, A., and Diederich, F. (2006) Fluorescent inhibitors for IspF, and enzyme in the non-mevalonate pathway for isoprenoid biosynthesis and a potential target for antimalarial therapy.

(59) Baumgartner, C., Eberle, C., Diederich, F., Lauw, S., Rohdich, F., Eisenreich, W., and Bacher, A. (2007) Structure-based design and synthesis of the first weak non-phosphate inhibitors for IspF, and enzyme in the non-mevalonate pathway of isoprenoid biosynthesis. Helv Chim Acta 90, 1043–1068.

185

(60) Geist, J. G., Lauw, S., Illarionova, V., Illarionov, B., Fischer, M., Gräwert, T., Rohdich, F., Eisenreich, W., Kaiser, J., Groll, M., Scheurer, C., Wittlin, S., Alonso-Gómez, J. L., Schweizer, W. B., Bacher, A., and Diederich, F. (2010) Thiazolopyrimidine Inhibitors of 2-Methylerythritol 2,4- Cyclodiphosphate Synthase (IspF) from Mycobacterium tuberculosis and Plasmodium falciparum. ChemMedChem 5, 1092–1101.

(61) Beutow, L., Brown, A., Parish, T., Hunter, W., and BMC. (2007) The structure of mycobacteria 2C-methyl-D-erythritol-4-cyclodiphosphate synthase, and essential enzyme, provides a platform for drug discovery. Biol 7 68 2.

(62) Campbell, T., and Brown, E. (2002) L and Characterization of the depletion of 2-C-methyl-D- erythritol-4-cyclodiphosphate synthase in Escherichia coli and bacillus subtilis. J Bacteriol 184 2, 5609–5618.

(63) Kemp, L., Alphey, M., Bond, C., Ferguson, M., Hecht, S., Bacher, A., Eisenreich, W., Rohdich, F., and Hunter, W. (2005) The identification of isoprenoids that bind in the intersubunit cavity of escherichia coli 2C-methyl-D-erythritol-4-cyclodiphosphate synthase by complementary biophysical methods. Acta Crystallogr Sect Biol Crystallogr 61 2, 45–52.

(64) Steinbacher, S., Kaiser, J., Wungsintaweekul, J., Hect, S., Eisenreich, W., Gerhardt, S., Bacher, A., and Rohdich, F. (2002) Structure of 2C-methyl-D-erythritol-4-cyclodiphosphate synthase involved in mevalonate-independent biosynthesis of isoprenoids. J Mol Biol 316 2, 79– 88.

(65) Sgraja, T., Kemp, L. E., Ramsden, N., and Hunter, W. N. (2005) A double mutation of Escherichia coli 2C-methyl-D-erythritol-2,4-cyclodiphosphate synthase disrupts six hydrogen bonds with, yet fails to prevent binding of, an isoprenoid diphosphate. Acta Crystallograph. Sect. F Struct. Biol. Cryst. Commun. 61, 625–9.

(66) Richard, S., Ferrer, J., Bowman, M., Lillo, A., Tetzlaff, C., Cane, D., and Noel, J. (2002) Structure and mechanism of 2-C-methyl-D-erythritol 4-cyclodiphosphate synthase. J Biol CHem 277 2, 8667–8672.

(67) Herz, S., Wungsintaweekul, J., Schuhr, C., Hect, S., Sagner, S., Fellermeier, M., Eisenreich, W., Zenk, M., Bacher, A., and Rohdich, F. (2000) Lüttgen, H., and Biosynthesis of terpenoids: YgbB protein converts 4-diphosphocytidyl-2C-methyl-D-erythritol 2-phosphate to 2C-methyl-d- erythritol 4-cyclodiphosphate. Proc Natl Acad Sci USA 97 2, 2486–2490.

(68) Kemp, L. E., Bond, C. S., and Hunter, W. N. (2002) Structure of 2C-methyl-D-erythritol 2,4- cyclodiphosphate synthase: an essential enzyme for isoprenoid biosynthesis and target for antimicrobial drug development. Proc. Natl. Acad. Sci. U. S. A. 99, 6591–6596.

186 (69) Ni, S., Robinson, H., Marsing, G., Bussiere, D., and Kennedy, M. (2004) Structure of 2C- methyl-D-erythritol-4-cyclodiphosphate synthase from shewanella oneidensis at 1.6Å: Identification of farnesyl pyrophosphate trapped in a hydrophobic cavity. Acta Crystallogr Sect Biol Crystallogr 60 2, 1949–1957.

(70) Bitok, J. K. (2012, October 5) Targeting Non-Mammalian Isoprenoid Biosynthesis: Regulation and Modulation of Cyclodiphosphate Synthase, IspF. Johns Hopkins University, Baltimore, MD.

(71) Eberl, M., Hintz, M., Reichenberg, A., Kollas, A.-K., Wiesner, J., and Jomaa, H. (2003) Microbial isoprenoid biosynthesis and human γδ T cell activation. FEBS Lett. 544, 4–10.

(72) Ogrel, O. D., Fegeding, K. V., Kharatian, E. F., Sudarikov, A. B., and Ostrovsky, D. N. (1996) The ability of a recombinant Escherichia coli strain to synthesize 2-C-methyl-D-erythritol-2,4- cyclopyrophosphate correlates with its tolerance to in vitro induced oxidative stress and to the bactericidal action of murine peritoneal macrophages. Curr. Microbiol. 32, 225–8.

(73) Ostrovsky, D., Shashkov, A., and Sviridov, A. (1993) Bacterial oxidative-stress substance is 2-C-methyl-D-erythritol 4-cyclopyrophosphate. Biochem J 295 2, 901–902.

(74) Ostrovsky, D., Shipanova, I., Sibeldina, L., Shashkov, A., Kharatian, E., Malyarova, I., and Tantsyrev, G. (1992) A new cyclopyrophosphate as a bacterial antistressor? FEBS Lett. 298, 159– 61.

(75) Ostrovsky, D., Diomina, G., Lysak, E., Matveeva, E., Ogrel, O., and Trutko, S. (1998) Effect of oxidative stress on the biosynthesis of 2-C-methyl-d-erythritol-4-cyclopyrophosphate and isoprenoids by several bacterial strains. Arch Microbiol 171 2, 69–72.

(76) Ershov, I. V. (2007) [2-C-Methylerythritol phosphate pathway of isoprenoid biosynthesis as a target in identifying of new antibiotics, herbicides, and immunomodulators (Review) ]. Prikl. Biokhim. Mikrobiol. 43, 133–57.

(77) Rivasseau, C., Seemann, M., Boisson, A.-M., Streb, P., Gout, E., Douce, R., Rohmer, M., and Bligny, R. (2009) Accumulation of 2-C-methyl-D-erythritol 2,4-cyclodiphosphate in illuminated plant leaves at supraoptimal temperatures reveals a bottleneck of the prokaryotic methylerythritol 4-phosphate pathway of isoprenoid biosynthesis. Plant Cell Environ. 32, 82–92.

(78) MongéLard, G., Seemann, M., Boisson, A.-M., Rohmer, M., Bligny, R., and Rivasseau, C. (2011) Measurement of carbon flux through the MEP pathway for isoprenoid synthesis by 31P- NMR spectroscopy after specific inhibition of 2-C-methyl-d-erythritol 2,4-cyclodiphosphate reductase. Effect of light and temperature: Carbon flux through the MEP pathway. Plant Cell Environ. 34, 1241–1247.

187 (79) Goncharenko, A., Ershov, Y., Salina, E., Wiesner, J., Vostroknutova, G., Sandanov, A., Kaprelyants, A., and Ostrovsky, D. (2007) The role of 2-C-methylerythritol-4- cyclopyrophosphate in the resuscitation of the “nonculturable” forms of mycobacterium smegmatis. Microbiol. 76 2, 147–152.

(80) Illarionova, V., Kaiser, J., Ostrozhenkova, E., Bacher, A., Fischer, M., Eisenreich, W., and Rohdich, F. (2006) Nonmevalonate Terpene Biosynthesis Enzymes as Antiinfective Drug Targets: Substrate Synthesis and High-Throughput Screening Methods. J. Org. Chem. 71, 8824–8834.

(81) Bitok, J. K., and Meyers, C. F. (2012) 2 C -Methyl- D -erythritol 4-Phosphate Enhances and Sustains Cyclodiphosphate Synthase IspF Activity. ACS Chem. Biol. 7, 1702–1710.

(82) Narayanasamy, P., Eoh, H., Brennan, P. J., and Crick, D. C. (2010) Synthesis of 4- Diphosphocytidyl-2-C-Methyl-D-Erythritol 2-Phosphate and Kinetic Studies of Mycobacterium tuberculosis IspF. Chem. Biol. 17, 117–122.

(83) Waldron, T. T., and Murphy, K. P. (2003) Stabilization of proteins by ligand binding: application to drug screening and determination of unfolding energetics. Biochemistry (Mosc.) 42, 5058–5064.

(84) Samala, G., Nallangi, R., Devi, P. B., Saxena, S., Yadav, R., Sridevi, J. P., Yogeeswari, P., and Sriram, D. (2014) Identification and development of 2-methylimidazo[1,2-a]pyridine-3- carboxamides as Mycobacterium tuberculosis pantothenate synthetase inhibitors. Bioorg. Med. Chem. 22, 4223–4232.

(85) Kim, A., Wolf, N. M., Zhu, T., Johnson, M. E., Deng, J., Cook, J. L., and Fung, L. W.-M. (2015) Identification of Bacillus anthracis PurE inhibitors with antimicrobial activity. Bioorg. Med. Chem. 23, 1492–1499.

(86) Mohedas, A. H., Wang, Y., Sanvitale, C. E., Canning, P., Choi, S., Xing, X., Bullock, A. N., Cuny, G. D., and Yu, P. B. (2014) Structure–Activity Relationship of 3,5-Diaryl-2-aminopyridine ALK2 Inhibitors Reveals Unaltered Binding Affinity for Fibrodysplasia Ossificans Progressiva Causing Mutants. J. Med. Chem. 57, 7900–7915.

(87) Pey, A. L., Mesa-Torres, N., Chiarelli, L. R., and Valentini, G. (2013) Structural and Energetic Basis of Protein Kinetic Destabilization in Human Phosphoglycerate Kinase 1 Deficiency. Biochemistry (Mosc.) 52, 1160–1170.

(88) Guo, J., Luk, L. Y. P., Loveridge, E. J., and Allemann, R. K. (2014) Thermal Adaptation of Dihydrofolate Reductase from the Moderate Thermophile Geobacillus stearothermophilus. Biochemistry (Mosc.) 53, 2855–2863.

188 (89) Choi, R., Kelley, A., Leibly, D., Nakazawa Hewitt, S., Napuli, A., and Van Voorhis, W. (2011) Immobilized metal-affinity chromatography protein-recovery screening is predictive of crystallographic structure success. Acta Crystallograph. Sect. F Struct. Biol. Cryst. Commun. 67, 998–1005.

(90) Begley, D. W., Hartley, R. C., Davies, D. R., Edwards, T. E., Leonard, J. T., Abendroth, J., Burris, C. A., Bhandari, J., Myler, P. J., Staker, B. L., and Stewart, L. J. (2011) Leveraging structure determination with fragment screening for infectious disease drug targets: MECP synthase from Burkholderia pseudomallei. J. Struct. Funct. Genomics 12, 63–76.

(91) Hubbard, S. J., and Thornton, J. M. (1993) NACCESS. University College London, Department of Biochemistry and Molecular Biology.

(92) Murphy, K. P., Bhakuni, V., Xie, D., and Freire, E. (1992) Molecular basis of co-operativity in protein folding. III. Structural identification of cooperative folding units and folding intermediates. J. Mol. Biol. 227, 293–306.

(93) Schymkowitz, J., Borg, J., Stricher, F., Nys, R., Rousseau, F., and Serrano, L. (2005) The FoldX web server: an online force field. Nucleic Acids Res. 33, W382–388.

(94) Teng, T.-Y. (1990) Mounting of crystals for macromolecular crystallography in a free- standing thin-film. J. Appl. Crystallogr. 23, 387–391.

(95) Supporting institutions for SER-CAT may be found at www.ser-cat.org/members.html.

(96) This research used resources of the Advance Photon Source, a U.S. Department of Energy (DOE) Office of Science User Facility operated for the DOE Office of Science by Argonne National Laboratory under Contract No. DE-AC02-06CH11357.

(97) Otwinowski, Z., and Minor, W. (1997) Processing of X-ray Diffraction Data Collected in Oscillation Mode. Macromol. Crystallograpy 276, 307–326.

(98) McCoy, A. J., Grosse-Kunstleve, R. W., Adams, P. D., Winn, M. D., Storoni, L. C., and Read, R. J. (2007) Phaser crystallographic software. J. Appl. Crystallogr. 40, 658–674.

(99) Adams, P. D., Afonine, P. V., Bunkoczi, V. B., Chen, I. W., Davis, N., Echols, J. J., Headd, L. W., Hung, G. J., Kapral, R. W., Grosse-Kunstleve, A. J., McCoy, N. W., Moriarty, R., Oeffner, R. J., Read, D. C., Richardson, J. S., Terwilliger, T. C., and Zwart, P. H. (2010) PHENIX: a comprehensive Python-based system for macromolecular structure solution.

(100) Murshudov, G. N., Vagin, A. A., and Dodson, E. J. (1997) Refinement of Macromolecular Structures by the Maximum-Likelihood Method. Acta Crystallogr. D Biol. Crystallogr. 53, 240– 255.

189 (101) Emsley, P., Lohkamp, B., Scott, W. G., and Cowtan, K. (2010) Features and development of Coot. Acta Crystallogr. D Biol. Crystallogr. 66, 486–501.

(102) Berman, H. M., Westbrook, J., Feng, Z., Gilliland, G., Bhat, T. N., Weissig, H., Shindyalov, I. N., and Bourne, P. E. (2000) The Protein Data Bank. Nucleic Acids Res. 28, 235–242.

(103) Baldwin, R. L. (1996) How Hofmeister ion interactions affect protein stability. Biophys. J. 71, 2056–2063.

(104) Martin, D. P., Blachly, P. G., Marts, A. R., Woodruff, T. M., de Oliveira, C. A. F., McCammon, J. A., Tierney, D. L., and Cohen, S. M. (2014) “Unconventional” Coordination Chemistry by Metal Chelating Fragments in a Metalloprotein Active Site. J. Am. Chem. Soc. 136, 5400–5406.

(105) Whittaker, M., Floyd, C. D., Brown, P., and Gearing, A. J. H. (1999) Design and Therapeutic Application of Matrix Metalloproteinase Inhibitors. Chem. Rev. 99, 2735–2776.

(106) Ramsden, N. L., Buetow, L., Dawson, A., Kemp, L. A., Ulaganathan, V., Brenk, R., Klebe, G., and Hunter, W. N. (2009) A structure-based approach to ligand discovery for 2C-methyl-D- erythritol-2,4-cyclodiphosphate synthase: a target for antimicrobial therapy. J. Med. Chem. 52, 2531–2542.

(107) Jacobsen, J. A., Major Jourden, J. L., Miller, M. T., and Cohen, S. M. (2010) To bind zinc or not to bind zinc: An examination of innovative approaches to improved metalloproteinase inhibition. Biochim. Biophys. Acta BBA - Mol. Cell Res. 1803, 72–94.

(108) Geist, J., Lauw, S., Illarionova, V., Illarionova, B., Fischer, M., Rohdich, F., Eisenreich, W., Kaiser, J., Groll, M., Scheurer, C., Wittlin, S., Schweizer, W., Bacher, A., and Diederich, F. (2010) Gräwert, T., Alonso-Gómez, J.L., and Thiazolopyrimidine inhibitors of 2-methylerythritol 4,- cyclodiphosphate synthase (IspF) from mycobacterium tuberculosis and plasmodium falciparum. ChemMedChem 5 2, 1092–1101.

(109) Afanador, G. A., Muench, S. P., McPhillie, M., Fomovska, A., Schön, A., Zhou, Y., Cheng, G., Stec, J., Freundlich, J. S., Shieh, H.-M., Anderson, J. W., Jacobus, D. P., Fidock, D. A., Kozikowski, A. P., Fishwick, C. W., Rice, D. W., Freire, E., McLeod, R., and Prigge, S. T. (2013) Discrimination of Potent Inhibitors of Toxoplasma gondii Enoyl-Acyl Carrier Protein Reductase by a Thermal Shift Assay. Biochemistry (Mosc.) 52, 9155–9166.

(110) Franco, E. J., Sonneson, G. J., DeLegge, T. J., Hofstetter, H., Horn, J. R., and Hofstetter, O. (2010) Production and characterization of a genetically engineered anti-caffeine camelid antibody and its use in immunoaffinity chromatography. J. Chromatogr. B 878, 177–186.

190 (111) Knepp, A. M., Grunbeck, A., Banerjee, S., Sakmar, T. P., and Huber, T. (2011) Direct Measurement of Thermal Stability of Expressed CCR5 and Stabilization by Small Molecule Ligands. Biochemistry (Mosc.) 50, 502–511.

(112) Bovet, C., Wortmann, A., Eiler, S., Granger, F., Ruff, M., Gerrits, B., Moras, D., and Zenobi, R. (2007) Estrogen receptor-ligand complexes measured by chip-based nanoelectrospray mass spectrometry: An approach for the screening of endocrine disruptors. Protein Sci. 16, 938–946.

(113) Jonker, N., Kool, J., Irth, H., and Niessen, W. M. A. (2011) Recent developments in protein–ligand affinity mass spectrometry. Anal. Bioanal. Chem. 399, 2669–2681.

(114) Annis, D. A., Nickbarg, E., Yang, X., Ziebell, M. R., and Whitehurst, C. E. (2007) Affinity selection-mass spectrometry screening techniques for small molecule drug discovery. Curr. Opin. Chem. Biol. 11, 518–526.

(115) Thelemann, J., Illarionov, B., Barylyuk, K., Geist, J., Kirchmair, J., Schneider, P., Anthore, L., Root, K., Trapp, N., Bacher, A., Witschel, M., Zenobi, R., Fischer, M., Schneider, G., and Diederich, F. (2015) Aryl Bis-Sulfonamide Inhibitors of IspF from Arabidopsis thaliana and Plasmodium falciparum. ChemMedChem n/a–n/a.

(116) Lea, W. A., and Simeonov, A. (2011) Fluorescence polarization assays in small molecule screening. Expert Opin. Drug Discov. 6, 17–32.

(117) Lei, Y., Hu, T., Wu, X., Wu, Y., Bao, Q., Zhang, L., Xia, H., Sun, H., You, Q., and Zhang, X. (2015) Affinity-Based Fluorescence Polarization Assay for High-Throughput Screening of Prolyl Hydroxylase 2 Inhibitors. ACS Med. Chem. Lett. 6, 1236–1240.

(118) (2006) Technical Resource Guide Fourth Edition Fluorescence Polarization. Invitrogen.

(119) pochetti, giorgio, Pochetti, G., and Montanari, R. (2012) Isothermal titration calorimetry to determine the association constants for a ligand bound simultaneously to two specific protein binding sites with different affinities. Protoc. Exch.

(120) Wöhri, A. B., Hillertz, P., Eriksson, P.-O., Meuller, J., Dekker, N., and Snijder, A. (2013) Thermodynamic studies of ligand binding to the human homopentameric glycine receptor using isothermal titration calorimetry. Mol. Membr. Biol. 30, 169–183.

(121) Yi, F., and Regan, L. (2008) A Novel Class of Small Molecule Inhibitors of Hsp90. ACS Chem. Biol. 3, 645–654.

(122) Meyer, B., and Peters, T. (2003) NMR spectroscopy techniques for screening and identifying ligand binding to protein receptors. Angew. Chem. Int. Ed Engl. 42, 864–890.

191 (123) Venkitakrishnan, R. P., Benard, O., Max, M., Markley, J. L., and Assadi-Porter, F. M. (2012) Use of NMR Saturation Transfer Difference Spectroscopy to Study Ligand Binding to Membrane Proteins, in Membrane Protein Structure and Dynamics (Vaidehi, N., and Klein-Seetharaman, J., Eds.), pp 47–63. Humana Press, Totowa, NJ.

(124) Davis, B. (2013) Screening Protein–Small Molecule Interactions by NMR, in Protein-Ligand Interactions (Williams, M. A., and Daviter, T., Eds.), pp 389–413. Humana Press, Totowa, NJ.

(125) Kuroki, K., and Maenaka, K. (2011) Analysis of Receptor–Ligand Interactions by Surface Plasmon Resonance, in Immune Receptors (Rast, J. P., and Booth, J. W. D., Eds.), pp 83–106. Humana Press, Totowa, NJ.

(126) Patching, S. G. (2014) Surface plasmon resonance spectroscopy for characterisation of membrane protein–ligand interactions and its potential for drug discovery. Biochim. Biophys. Acta BBA - Biomembr. 1838, 43–55.

(127) Shepherd, C. A., Hopkins, A. L., and Navratilova, I. (2014) Fragment screening by SPR and advanced application to GPCRs. Prog. Biophys. Mol. Biol. 116, 113–123.

(128) Jameson, D. M., and Croney, J. C. (2003) Fluorescence polarization: past, present and future. Comb. Chem. High Throughput Screen. 6, 167–173.

(129) Roos, H., Karlsson, R., Nilshans, H., and Persson, A. (1998) Thermodynamic analysis of protein interactions with biosensor technology. J. Mol. Recognit. JMR 11, 204–210.

(130) Lin, S., Shih-Yuan Lee, A., Lin, C.-C., and Lee, C.-K. (2006) Determination of Binding Constant and Stoichiometry for Antibody-Antigen Interaction with Surface Plasmon Resonance. Curr. Proteomics 3, 271–282.

(131) Núñez, S., Venhorst, J., and Kruse, C. G. (2012) Target–drug interactions: first principles and their application to drug discovery. Drug Discov. Today 17, 10–22.

(132) Hermanson, G. T. (2013) Bioconjugate techniques Third edition. Elsevier/AP, London ; Waltham, MA.

(133) Pace, C. N., Vajdos, F., Fee, L., Grimsley, G., and Gray, T. (1995) How to measure and predict the molar absorption coefficient of a protein. Protein Sci. Publ. Protein Soc. 4, 2411– 2423.

(134) Mach, H., Middaugh, C. R., and Lewis, R. V. (1992) Statistical determination of the average values of the extinction coefficients of tryptophan and tyrosine in native proteins. Anal. Biochem. 200, 74–80.

192 (135) Visser, A. J. W. G., and Rolinski, O. J. Bsic Photophysics.

(136) Zhang, Z., Jakkaraju, S., Blain, J., Gogol, K., Zhao, L., Hartley, R. C., Karlsson, C. A., Staker, B. L., Edwards, T. E., Stewart, L. J., Myler, P. J., Clare, M., Begley, D. W., Horn, J. R., and Hagen, T. J. (2013) Cytidine derivatives as IspF inhibitors of Burkolderia pseudomallei. Bioorg. Med. Chem. Lett. 23, 6860–6863.

(137) Carey, F. A. Structure and physical properties. Organohalogen Compd. Encyclopedia Britannica.

(138) Bartzatt, R. (2001) Dansylation of hydroxyl and carboxylic acid functional groups. J. Biochem. Biophys. Methods 47, 189–195.

(139) Geschwindner, S., Ulander, J., and Johansson, P. (2015) Ligand Binding Thermodynamics in Drug Discovery: Still a Hot Tip? J. Med. Chem. 58, 6321–6335.

(140) Turnbull, W. B., and Daranas, A. H. (2003) On the Value of c: Can Low Affinity Systems Be Studied by Isothermal Titration Calorimetry? J. Am. Chem. Soc. 125, 14859–14866.

(141) Lagisetti, C., Urbansky, M., and Coates, R. M. (2007) The Dioxanone Approach to (2 S ,3 R )-2- C -Methylerythritol 4-Phosphate and 2,4-Cyclodiphosphate, and Various MEP Analogues. J. Org. Chem. 72, 9886–9895.

(142) Geist, J. (2011) Development of Inhibitors for the Enzyme IspF using High-Throughput Screening and Structure-Based Design. ETH, Zurich.

(143) Holmes, M. A., and Matthews, B. W. (1981) Binding of hydroxamic acid inhibitors to crystalline thermolysin suggests a pentacoordinate zinc intermediate in catalysis. Biochemistry (Mosc.) 20, 6912–6920.

(144) Huguet, F., Melet, A., Alves de Sousa, R., Lieutaud, A., Chevalier, J., Maigre, L., Deschamps, P., Tomas, A., Leulliot, N., Pages, J.-M., and Artaud, I. (2012) Hydroxamic Acids as Potent Inhibitors of Fe II and Mn II E. coli Methionine Aminopeptidase: Biological Activities and X-ray Structures of Oxazole Hydroxamate- Ec MetAP-Mn Complexes. ChemMedChem 7, 1020–1030.

(145) Drinkwater, N., Vinh, N. B., Mistry, S. N., Bamert, R. S., Ruggeri, C., Holleran, J. P., Loganathan, S., Paiardini, A., Charman, S. A., Powell, A. K., Avery, V. M., McGowan, S., and Scammells, P. J. (2016) Potent dual inhibitors of Plasmodium falciparum M1 and M17 aminopeptidases through optimization of S1 pocket interactions. Eur. J. Med. Chem. 110, 43– 64.

(146) Mistry, S. N., Drinkwater, N., Ruggeri, C., Sivaraman, K. K., Loganathan, S., Fletcher, S., Drag, M., Paiardini, A., Avery, V. M., Scammells, P. J., and McGowan, S. (2014) Two-Pronged

193 Attack: Dual Inhibition of Plasmodium falciparum M1 and M17 Metalloaminopeptidases by a Novel Series of Hydroxamic Acid-Based Inhibitors. J. Med. Chem. 57, 9168–9183.

(147) (2016) Tuberculosis. World Health Organization.

(148) (2014, October 9) Tuberculosis (TB) Fact Sheet. Centers for Disease Control and Prevention.

(149) Sachs, G., Shin, J. M., and Howden, C. W. (2006) Review article: the clinical pharmacology of proton pump inhibitors. Aliment. Pharmacol. Ther. 23, 2–8.

(150) Madigan, M. T., and Brock, T. D. (Eds.). (2012) Brock biology of microorganisms 13. ed., global ed. Pearson, Boston, Mass.

(151) Kinoshita, M. (2014) Anticonvulsant sulfonamides in epilepsy and other neurological disorders.

(152) Werner, L. H., Habicht, E., and Zergenyi, J. (1978) Sulfonamide Diuretics, in Diuretic Agents (Cragoe, E. J., Ed.), pp 38–55. AMERICAN CHEMICAL SOCIETY, WASHINGTON, D. C.

(153) Davis, T. W., Kerr, R. B., and Bogoch, A. (1959) Experience with glipasol (R. P. 2259)-an antidiabetic sulfonamide drug. Can. Med. Assoc. J. 81, 101–107.

(154) Silverman, R. B., and Holladay, M. W. (2014) The organic chemistry of drug design and drug action Third edition. Elsevier/AP, Academic Press, is an imprint of Elsevier, Amsterdam ; Boston.

(155) Lepre, C. A., Moore, J. M., and Peng, J. W. (2004) Theory and Applications of NMR-Based Screening in Pharmaceutical Research. Chem. Rev. 104, 3641–3676.

(156) Begley, D. W. Applying ligand-observe NMR spectroscopy to enable fragment-based lead discovery.

(157) Viegas, A., Manso, J., Nobrega, F. L., and Cabrita, E. J. (2011) Saturation-Transfer Difference (STD) NMR: A Simple and Fast Method for Ligand Screening and Characterization of Protein Binding. J. Chem. Educ. 88, 990–994.

(158) Saturation Transfer Difference (STD) NMR. University of Wisconsin Madison Department of Chemistry.

APPENDIX A – EXPRESSION VECTOR OF Burkholderia pseudomallei IspF

Expression Vector AVA0421

Antibiotic marker: Ampicillin and Chloramphenicol

Sequencing primers: Novagen T7 promoter primer (5’→3’) and T7 terminator primer (3’→5’)

Notes:  This is a T7 promoter-based expression vector constructed from pET14b (Novagen). The cloned gene is expressed as an N-terminal fusion with a His6-3C cleavage site.  Use BL21(DE3) or similar host cell for expression and standard expression and purification condition or T7 and His-tag.

195 Translation of DNAMAN1{RF+1}(1-550) Universal code Total amino acid number: 182, MW=19347 Max ORF starts at AA pos 1(may be DNA pos 1) for 182 AA (546 bases), MW=19347

His6-tag 10 20 30 40 50 60 1 ATGGCTCATCACCATCACCATCATATGGGTACCCTGGAAGCTCAGACCCAGGGTCCTGGT 1 M A H H H H H H M G T L E A Q T Q G P G

70 80 90 100 110 120 61 TCGATGGATTTCAGAATCGGACAAGGCTACGACGTGCACCAGCTCGTTCCCGGACGCCCG 21 S M D F R I G Q G Y D V H Q L V P G R P

130 140 150 160 170 180 121 CTCATCATCGGCGGCGTGACGATCCCTTACGAGCGGGGGCTGCTCGGCCACTCGGACGCC 41 L I I G G V T I P Y E R G L L G H S D A

190 200 210 220 230 240 181 GACGTGCTGCTGCACGCGATCACCGACGCGCTCTTCGGCGCGGCGGCGCTCGGCGACATC 61 D V L L H A I T D A L F G A A A L G D I

250 260 270 280 290 300 241 GGCCGCCACTTCTCCGACACCGATCCGCGCTTCAAGGGCGCCGACAGCCGCGCGCTGTTG 81 G R H F S D T D P R F K G A D S R A L L

310 320 330 340 350 360 301 CGCGAATGCGCATCGCGCGTCGCGCAGGCGGGTTTTGCGATCCGCAACGTCGACAGCACG 101 R E C A S R V A Q A G F A I R N V D S T

370 380 390 400 410 420 361 ATCATCGCGCAGGCGCCGAAGCTCGCGCCGCATATCGACGCGATGCGCGCGAACATCGCG 121 I I A Q A P K L A P H I D A M R A N I A

430 440 450 460 470 480 421 GCCGATCTCGACTTGCCGCTCGATCGCGTGAACGTGAAGGCGAAGACGAACGAGAAGCTC 141 A D L D L P L D R V N V K A K T N E K L

490 500 510 520 530 540 481 GGCTATCTCGGGCGCGGTGAGGGGATCGAGGCGCAGGCGGCCGCGCTCGTCGTGCGCGAA 161 G Y L G R G E G I E A Q A A A L V V R E

550 560 570 541 GCGGCCGCGTGAAAACAGCACGAACAAGTTCTGCA 181 A A A stop

196 Full DNA sequence

TTCTCATGTTTGACAGCTTATCATCGATAAGCTTTAATGCGGTAGTTTATCACAGTTA AATTGCTAACGCAGTCAGGCACCGTGTATGAAATCTAACAATGCGCTCATCGTCATC CTCGGCACCGTCACCCTGGATGCTGTAGGCATAGGCTTGGTTATGCCGGTACTGCCG GGCCTCTTGCGGGATATCGTCCATTCCGACAGCATCGCCAGTCACTATGGCGTGCTG CTAGCGCTATATGCGTTGATGCAATTTCTATGCGCACCCGTTCTCGGAGCACTGTCC GACCGCTTTGGCCGCCGCCCAGTCCTGCTCGCTTCGCTACTTGGAGCCACTATCGAC TACGCGATCATGGCGACCACACCCGTCCTGTGGATATCCGGATATAGTTCCTCCTTT CAGCAAAAAACCCCTCAAGACCCGTTTAGAGGCCCCAAGGGGTTATGCTAGTTATTG CTCAGCGGTGGCAGCAGCCAACTCAGCTTCCTTTCGGGCTTTGTTAGCAGCCGGATC CTCGAGCATATGGCTGCCGCGCGGCACCAGGCCGCTGCTGTGATGATGATGATGATG GCTGCTGCCCATGGTATATCTCCTTCTTAAAGTTAAACAAAATTATTTCTAGAGGGA AACCGTTGTGGTCTCCCTATAGTGAGTCGTATTAATTTCGCGGGATCGAGATCTCGA TCCTCTACGCCGGACGCATCGTGGCCGGCATCACCGGCGCCACAGGTGCGGTTGCTG GCGCCTATATCGCCGACATCACCGATGGGGAAGATCGGGCTCGCCACTTCGGGCTC ATGAGCGCTTGTTTCGGCGTGGGTATGGTGGCAGGCCCCGTGGCCGGGGGACTGTTG GGCGCCATCTCCTTGCATGCACCATTCCTTGCGGCGGCGGTGCTCAACGGCCTCAAC CTACTACTGGGCTGCTTCCTAATGCAGGAGTCGCATAAGGGAGAGCGTCGACCGAT GCCCTTGAGAGCCTTCAACCCAGTCAGCTCCTTCCGGTGGGCGCGGGGCATGACTAT CGTCGCCGCACTTATGACTGTCTTCTTTATCATGCAACTCGTAGGACAGGTGCCGGC AGCGCTCTGGGTCATTTTCGGCGAGGACCGCTTTCGCTGGAGCGCGACGATGATCGG CCTGTCGCTTGCGGTATTCGGAATCTTGCACGCCCTCGCTCAAGCCTTCGTCACTGGT CCCGCCACCAAACGTTTCGGCGAGAAGCAGGCCATTATCGCCGGCATGGCGGCCGA CGCGCTGGGCTACGTCTTGCTGGCGTTCGCGACGCGAGGCTGGATGGCCTTCCCCAT TATGATTCTTCTCGCTTCCGGCGGCATCGGGATGCCCGCGTTGCAGGCCATGCTGTC CAGGCAGGTAGATGACGACCATCAGGGACAGCTTCAAGGATCGCTCGCGGCTCTTA CCAGCCTAACTTCGATCACTGGACCGCTGATCGTCACGGCGATTTATGCCGCCTCGG CGAGCACATGGAACGGGTTGGCATGGATTGTAGGCGCCGCCCTATACCTTGTCTGCC TCCCCGCGTTGCGTCGCGGTGCATGGAGCCGGGCCACCTCGACCTGAATGGAAGCC GGCGGCACCTCGCTAACGGATTCACCACTCCAAGAATTGGAGCCAATCAATTCTTGA AGACGAAAGGGCCTCGTGATACGCCTATTTTTATAGGTTAATGTCATGATAATAATG GTTTCTTAGACGTCAGGTGGCACTTTTCGGGGAAATGTGCGCGGAACCCCTATTTGT TTATTTTTCTAAATACATTCAAATATGTATCCGCTCATGAGACAATAACCCTGATAA ATGCTTCAATAATATTGAAAAAGGAAGAGTATGAGTATTCAACATTTCCGTGTCGCC CTTATTCCCTTTTTTGCGGCATTTTGCCTTCCTGTTTTTGCTCACCCAGAAACGCTGGT GAAAGTAAAAGATGCTGAAGATCAGTTGGGTGCACGAGTGGGTTACATCGAACTGG ATCTCAACAGCGGTAAGATCCTTGAGAGTTTTCGCCCCGAAGAACGTTTTCCAATGA TGAGCACTTTTAAAGTTCTGCTATGTGGCGCGGTATTATCCCGTGTTGACGCCGGGC AAGAGCAACTCGGTCGCCGCATACACTATTCTCAGAATGACTTGGTTGAGTACTCAC CAGTCACAGAAAAGCATCTTACGGATGGCATGACAGTAAGAGAATTATGCAGTGCT GCCATAACCATGAGTGATAACACTGCGGCCAACTTACTTCTGACAACGATCGGAGG ACCGAAGGAGCTAACCGCTTTTTTGCACAACATGGGGGATCATGTAACTCGCCTTGA TCGTTGGGAACCGGAGCTGAATGAAGCCATACCAAACGACGAGCGTGACACCACGA TGCCTGCAGCAATGGCAACAACGTTGCGCAAACTATTAACTGGCGAACTACTTACTC TAGCTTCCCGGCAACAATTAATAGACTGGATGGAGGCGGATAAAGTTGCAGGACCA

197 CTTCTGCGCTCGGCCCTTCCGGCTGGCTGGTTTATTGCTGATAAATCTGGAGCCGGTG AGCGTGGGTCTCGCGGTATCATTGCAGCACTGGGGCCAGATGGTAAGCCCTCCCGTA TCGTAGTTATCTACACGACGGGGAGTCAGGCAACTATGGATGAACGAAATAGACAG ATCGCTGAGATAGGTGCCTCACTGATTAAGCATTGGTAACTGTCAGACCAAGTTTAC TCATATATACTTTAGATTGATTTAAAACTTCATTTTTAATTTAAAAGGATCTAGGTGA AGATCCTTTTTGATAATCTCATGACCAAAATCCCTTAACGTGAGTTTTCGTTCCACTG AGCGTCAGACCCCGTAGAAAAGATCAAAGGATCTTCTTGAGATCCTTTTTTTCTGCG CGTAATCTGCTGCTTGCAAACAAAAAAACCACCGCTACCAGCGGTGGTTTGTTTGCC GGATCAAGAGCTACCAACTCTTTTTCCGAAGGTAACTGGCTTCAGCAGAGCGCAGAT ACCAAATACTGTCCTTCTAGTGTAGCCGTAGTTAGGCCACCACTTCAAGAACTCTGT AGCACCGCCTACATACCTCGCTCTGCTAATCCTGTTACCAGTGGCTGCTGCCAGTGG CGATAAGTCGTGTCTTACCGGGTTGGACTCAAGACGATAGTTACCGGATAAGGCGC AGCGGTCGGGCTGAACGGGGGGTTCGTGCACACAGCCCAGCTTGGAGCGAACGACC TACACCGAACTGAGATACCTACAGCGTGAGCTATGAGAAAGCGCCACGCTTCCCGA AGGGAGAAAGGCGGACAGGTATCCGGTAAGCGGCAGGGTCGGAACAGGAGAGCGC ACGAGGGAGCTTCCAGGGGGAAACGCCTGGTATCTTTATAGTCCTGTCGGGTTTCGC CACCTCTGACTTGAGCGTCGATTTTTGTGATGCTCGTCAGGGGGGCGGAGCCTATGG AAAAACGCCAGCAACGCGGCCTTTTTACGGTTCCTGGCCTTTTGCTGGCCTTTTGCTC ACATGTTCTTTCCTGCGTTATCCCCTGATTCTGTGGATAACCGTATTACCGCCTTTGA GTGAGCTGATACCGCTCGCCGCAGCCGAACGACCGAGCGCAGCGAGTCAGTGAGCG AGGAAGCGGAAGAGCGCCTGATGCGGTATTTTCTCCTTACGCATCTGTGCGGTATTT CACACCGCATATATGGTGCACTCTCAGTACAATCTGCTCTGATGCCGCATAGTTAAG CCAGTATACACTCCGCTATCGCTACGTGACTGGGTCATGGCTGCGCCCCGACACCCG CCAACACCCGCTGACGCGCCCTGACGGGCTTGTCTGCTCCCGGCATCCGCTTACAGA CAAGCTGTGACCGTCTCCGGGAGCTGCATGTGTCAGAGGTTTTCACCGTCATCACCG AAACGCGCGAGGCAGCTGCGGTAAAGCTCATCAGCGTGGTCGTGAAGCGATTCACA GATGTCTGCCTGTTCATCCGCGTCCAGCTCGTTGAGTTTCTCCAGAAGCGTTAATGTC TGGCTTCTGATAAAGCGGGCCATGTTAAGGGCGGTTTTTTCCTGTTTGGTCACTGATG CCTCCGTGTAAGGGGGATTTCTGTTCATGGGGGTAATGATACCGATGAAACGAGAG AGGATGCTCACGATACGGGTTACTGATGATGAACATGCCCGGTTACTGGAACGTTGT GAGGGTAAACAACTGGCGGTATGGATGCGGCGGGACCAGAGAAAAATCACTCAGG GTCAATGCCAGCGCTTCGTTAATACAGATGTAGGTGTTCCACAGGGTAGCCAGCAGC ATCCTGCGATGCAGATCCGGAACATAATGGTGCAGGGCGCTGACTTCCGCGTTTCCA GACTTTACGAAACACGGAAACCGAAGACCATTCATGTTGTTGCTCAGGTCGCAGAC GTTTTGCAGCAGCAGTCGCTTCACGTTCGCTCGCGTATCGGTGATTCATTCTGCTAAC CAGTAAGGCAACCCCGCCAGCCTAGCCGGGTCCTCAACGACAGGAGCACGATCATG CGCACCCGTGGCCAGGACCCAACGCTGCCCGAGATGCGCCGCGTGCGGCTGCTGGA GATGGCGGACGCGATGGATATGTTCTGCCAAGGGTTGGTTTGCGCATTCACAGTTCT CCGCAAGAATTGATTGGCTCCAATTCTTGGAGTGGTGAATCCGTTAGCGAGGTGCCG CCGGCTTCCATTCAGGTCGAGGTGGCCCGGCTCCATGCACCGCGACGCAACGCGGG GAGGCAGACAAGGTATAGGGCGGCGCCTACAATCCATGCCAACCCGTTCCATGTGC TCGCCGAGGCGGCATAAATCGCCGTGACGATCAGCGGTCCAGTGATCGAAGTTAGG CTGGTAAGAGCCGCGAGCGATCCTTGAAGCTGTCCCTGATGGTCGTCATCTACCTGC CTGGACAGCATGGCCTGCAACGCGGGCATCCCGATGCCGCCGGAAGCGAGAAGAAT CATAATGGGGAAGGCCATCCAGCCTCGCGTCGTGAACGCCAGCAAGACGTAGCCCA

198 GCGCGTCGGCCGTAACAACACCATTTAAATGGAGTGGTTACAAATGGAGTGGTTAA TTAACAACACCATTTGTCGACGCTCTCCCTTATGCGACTCCTGCATTAGGAAGCAGC CCAGTAGTAGGTTGAGGCCGTTGAGCACCGCCGCCGCAAGGAATGGTGCATGCAAG GAGATGGCGCCCAACAGTCCCCCGGCCACGGGGCCTGCCACCATACCCACGCCGAA ACAAGCGCTCATGAGCCCGAAGTGGCGAGCCCGATCTTCCCCATCGGTGATGTCGG CGATATAGGCGCCAGCAACCGCACCTGTGGCGCCGGTGATGCCGGCCACGATGCGT CCGGCGTAGAGGATCGAGATCTCGATCCCGCGAAATTAATACGACTCACTATAGGG AGACCACAACGGTTTCCCTCTAGAAATAATTTTGTTTAACTTTAAGAAGGAGATATA CCATGGCTCATCACCATCACCATCATATGGGTACCCTGGAAGCTCAGACCCAGGGTC CTGGTTCGATGGATTTCAGAATCGGACAAGGCTACGACGTGCACCAGCTCGTTCCCG GACGCCCGCTCATCATCGGCGGCGTGACGATCCCTTACGAGCGGGGGCTGCTCGGC CACTCGGACGCCGACGTGCTGCTGCACGCGATCACCGACGCGCTCTTCGGCGCGGC GGCGCTCGGCGACATCGGCCGCCACTTCTCCGACACCGATCCGCGCTTCAAGGGCGC CGACAGCCGCGCGCTGTTGCGCGAATGCGCATCGCGCGTCGCGCAGGCGGGTTTTG CGATCCGCAACGTCGACAGCACGATCATCGCGCAGGCGCCGAAGCTCGCGCCGCAT ATCGACGCGATGCGCGCGAACATCGCGGCCGATCTCGACTTGCCGCTCGATCGCGTG AACGTGAAGGCGAAGACGAACGAGAAGCTCGGCTATCTCGGGCGCGGTGAGGGGA TCGAGGCGCAAGCGGCCGCGCTCGTCGTGCGCGAAGCGGCCGCGTGAAAACAGCAC GAACAAGTTCTGCAGCCAAGCTTCTCGAGGATCCGGCTGCTAACAAAGCCCGAAAG GAAGCTGAGTTGGCTGCTGCCACCGCTGAGCAATAACTAGCATAACCCCTTGGGGC CTCTAAACGGGTCTTGAGGGGTTTTTTGCTGAAAGGAGGAACTATATCCGGATATCC ACAGGACGGGTGTGGTCGCCATGATCGCGTAGTCGATAGTGGCTCCAAGTAGCGAA GCGAGCAGGACTGGGCGGCGGCCAAAGCGGTCGGACAGTGCTCCGAGAACGGGTG CGCATAGAAATTGCATCAACGCATATAGCGCTAGCAGCACGCCATAGTGACTGGCG ATGCTGTCGGAATGGACGATATCCCGCAAGAGGCCCGGCAGTACCGGCATAACCAA GCCTATGCCTACAGCATCCAGGGTGACGGTGCCGAGGATGACGATGAGCGCATTGT TAGATTTCATACACGGTGCCTGACTGCGTTAGCAATTTAACTGTGATAAACTACCGC ATTAAAGCTTATCGATGATAAGCTGTCAAACATGAGAA

APPENDIX B – IMIDAZOLE SPR SENSOGRAMS

FOL955*

HGN-0021*

A B

Figure B1. The dissociation constant of HGN-0106 displaying A) Sensogram displaying the response in µRIU for each concentration tested and B) semi-logarithmic plot to determined

Kd,app.

A B

Figure B2. The dissociation constant of HGN-0109 displaying A) Sensogram displaying the response in µRIU for each concentration tested and B) semi-logarithmic plot to determined Kd,app

200 A B

Figure B3. The dissociation constant of HGN-0093 displaying A) Sensogram displaying the response in µRIU for each concentration tested and B) semi-logarithmic plot to determined

Kd,app.

A B

Figure B4. The dissociation constant of HGN-0091 displaying A) Sensogram displaying the response in µRIU for each concentration tested and B) semi-logarithmic plot to determined

Kd,app.

201 A B

Figure B5. The dissociation constant of HGN-0090 displaying A) Sensogram displaying the response in µRIU for each concentration tested and B) semi-logarithmic plot to determined Kd,app.

A B

Figure B6. The dissociation constant of HGN-0094 displaying A) Sensogram displaying the response in µRIU for each concentration tested and B) semi-logarithmic plot to determined Kd,app.

202

A B

Figure B7. The dissociation constant of HGN-0092 displaying A) Sensogram displaying the response in µRIU for each concentration tested and B) semi-logarithmic plot to determined Kd,app.

A B

Figure B8. The dissociation constant of HGN-0022 displaying A) Sensogram displaying the response in µRIU for each concentration tested and B) semi-logarithmic plot to determined Kd,app.

203 A B

Figure B9. The dissociation constant of HGN-0107 displaying A) Sensogram displaying the response in µRIU for each concentration tested and B) semi-logarithmic plot to determined Kd,app.

A B

Figure B10. The dissociation constant of HGN-0097 displaying A) Sensogram displaying the response in µRIU for each concentration tested and B) semi-logarithmic plot to determined Kd,app.

204 A B

Figure B11. The dissociation constant of HGN-0095 displaying A) Sensogram displaying the response in µRIU for each concentration tested and B) semi-logarithmic plot to determined Kd,app.

A B

Figure B12. The dissociation constant of HGN-0096 displaying A) Sensogram displaying the response in µRIU for each concentration tested and B) semi-logarithmic plot to determined Kd,app.

205 A B

Figure B13. The dissociation constant of HGN-0108 displaying A) Sensogram displaying the response in µRIU for each concentration tested and B) semi-logarithmic plot to determined Kd,app.

A B

Figure B14. The dissociation constant of HGN-0051 displaying A) Sensogram displaying the response in µRIU for each concentration tested and B) semi-logarithmic plot to determined Kd,app.

206 A B

Figure B15. The dissociation constant of HGN-0098 displaying A) Sensogram displaying the response in µRIU for each concentration tested and B) semi-logarithmic plot to determined Kd,app.

A B

Figure B16. The dissociation constant of HGN-0099 displaying A) Sensogram displaying the response in µRIU for each concentration tested and B) semi-logarithmic plot to determined Kd,app.

207 A B

Figure B17. The dissociation constant of HGN-0100 displaying A) Sensogram displaying the response in µRIU for each concentration tested and B) semi-logarithmic plot to determined Kd,app.

A B

Figure B18. The dissociation constant of HGN-0201 displaying A) Sensogram displaying the response in µRIU for each concentration tested and B) semi-logarithmic plot to determined Kd,app.

* Compounds were run by Sriram Jakkaraju and sensograms are not included.

APPENDIX C – IMIDAZOTHIAZOLE SPR SENSOGRAMS

A B

Figure C1. The dissociation constant of HGN-0245 displaying A) Sensogram displaying the response in µRIU for each concentration tested and B) semi-logarithmic plot to determined Kd,app. A B

Figure C2. The dissociation constant of HGN-0289 displaying A) Sensogram displaying the response in µRIU for each concentration tested and B) semi-logarithmic plot to determined Kd,app.

209 A B

Figure C3. The dissociation constant of HGN-0282 displaying A) Sensogram displaying the response in µRIU for each concentration tested and B) semi-logarithmic plot to determined Kd,app.

A B

Figure C4. The dissociation constant of HGN-0321 displaying A) Sensogram displaying the response in µRIU for each concentration tested and B) semi-logarithmic plot to determined Kd,app.

210 A B

Figure C5. The dissociation constant of HGN-0247 displaying A) Sensogram displaying the response in µRIU for each concentration tested and B) semi-logarithmic plot to determined Kd,app.

A B

Figure C6. The dissociation constant of HGN-0318 displaying A) Sensogram displaying the response in µRIU for each concentration tested and B) semi-logarithmic plot to determined Kd,app.

211 A B

Figure C7. The dissociation constant of HGN-0316 displaying A) Sensogram displaying the response in µRIU for each concentration tested and B) semi-logarithmic plot to determined Kd,app.

A B

Figure C8. The dissociation constant of HGN-0317 displaying A) Sensogram displaying the response in µRIU for each concentration tested and B) semi-logarithmic plot to determined Kd,app.

APPENDIX D – IDENTIFICATION OF NEW HIT COMPOUNDS FOR BpIspF

The use of a thermal-shift assay has been widely implemented to generate a rank ordering of inhibitors for protein target binding.84,85,86,87,109 The shift in melting temperature in the presence of inhibitors can be measured rapidly, >2000 compounds in an hour, which makes the assay a powerful technique. Therefore, the development of an internal high throughput thermal- shift assay for BpIspF was a critical alternative to STD-NMR, which had to be done in collaboration with an exterior partner. The development of this assay for BpIspF was made more challenging by the high thermal stability of the enzyme. However, a successful effort in rational, point directed mutagenesis allowed the development of a form of BpIspF with sufficiently decreased Tm, but retaining native activity, to allow for small molecules to be screened. Using this assay, BpIspF could be screened against a library consisting of drug components from the

US, Europe, and Japan, natural products with known biological activity, and bioactive compounds that include non-drug enzyme inhibitors, receptor blockers, membrane active compounds, and cellular toxins. The library is called the Spectrum Collection (MicroSource

Discovery Systems, Inc.) and contains a total of 2400 compounds. The BpIspF Q151E variant was screened against the entire library to identify new fragment hits.

Method:

General procedure for temperature-shift assay was followed as reported in Chapter 2.

Results:

213 Screening for new lead compounds in the spectrum collection identified three chemical classes of compounds that were found to change the melting temperature of BpIspF Q151E by >±4°C.

The first chemical class of compounds determined to have an apparent affinity for the enzyme were -lactams, known to inhibit cell wall biosynthesis in bacteria.9,12 This class constituted a total of eight compounds out of the fifty-six in the library (D1). Since only eight of the fifty-six compounds in this class found in the library were identified by the assay, the apparent affinities were not simply specific to the -lactam core.

The second class of compounds identified was sulfoxides, which are proton pump inhibitors that reduce gastric acid production (D2). Sulfoxides have been found to be the most potent inhibitors available for gastrointestinal acid secretion149 A total of only three out of thirty- two sulfoxides found in the library were identified by the thermal-shift assay as hits against

BpIspF, once again showing that the assay was working selectively and that the enzyme is not simply promiscuously binding sulfoxides.

The last class of compounds identified to have an apparent affinity for BpIspF was sulfonamides, which inhibit folate synthesis (D3).150 Although sulfonamides were originally identified as antibacterials, they have been found to also function as anticonvulsants, diuretics, and anti-diabetic agents.151,152,153 Therefore, it is not surprising that this broad spectrum class might also have potential activity against BpIspF. However, this might also prove problematic for the ultimate development of a BpIspF drug from the class since there is clear precedent for cross-reactivity to other systems. As for the other two classes identified by the assay, only three of the eighty-four sulfonamides present in the library were identified as hits for BpIspF binding, so this once again does not suggest simple promiscuity against an entire class.

Table D1 contintued 214

Blind screening of chemical libraries to identify binders for a target molecule can lead to false positives. The identification of false positive hits is particularly exacerbated when libraries are populated with molecular classes that are known to be promiscuous binders, like the catechols.154 Though, well designed libraries will minimize or totally eliminate such compound classes since they serve little purpose for high specificity inhibitor design.154 To test the applicability of the Spectrum Collection to the goals of this study, the library was also screened against another enzyme in the MEP pathway, EcIspD. The fragment hits identified for BpIspF were not the same compounds that showed apparent affinity for EcIspD, confirming that these compounds were indeed BpIspF-specific, and not promiscuous, binders and that the library was applicable to the goals of this study.

Table D1. List of -lactams identified in the thermal-shift assay at Northwestern University.

-lactams Structure ΔTm

Cefamandole 33.57

Cefonicid 28.48

Continued on following page Table D1 contintued 215

Cefoperazone 18.74

Cefazolin 17.95

Cefotaxime 10.93

Cefdinir 6.66

Cephradine 5.15

Azlocillin 4.55

Continued on following page 216 Table D2. List of Sulfoxides identified in the thermal-shift assay at Northwestern University.

Sulfoxides Structure ΔTm

Lansoprazole 19.25

Esomeprazole 5.12

Rabeprazole 4.28

Table D3. List of Sulfonamides identified in the thermal-shift assay at Northwestern University.

Sulfonamides Structure ΔTm

Ethoxzolamide 10.31

Sulfapyridine -5.62

Sulfamonoethoxine -9.63

APPENDIX E – ANALYSIS OF INHIBITOR BINDING BY STD-NMR

The use of NMR to screen drug-like molecules was first reported in the 1990’s and has been advancing rapidly over the years. NMR has become a powerful tool to study nucleic acid and protein interactions with small molecules155 using Saturation-transfer difference NMR, or

STD-NMR. STD-NMR is a powerful technique that allows epitope mapping of small molecule- protein interactions. This helps to determine which portion(s) of small molecule ligands are within 5Å of the protein.122,156 The STD-NMR experiment depends on the exchange rate of bound and unbound ligand (Figure E1) and nuclear Overhauser effect (NOE) to identify protons which are interacting with the protein (Figure E2). Therefore, this approach can be utilized for those compounds with a dissociation constant between 10-8 and 10-3 M due to the time scale which NMR is capable of measuring (µs-s).157 Yet another limitation is that the concentration of available ligand must be in great excess of the protein, typically by a factor of 50-100. Such high concentrations can lead to problems with less soluble compounds and may result in non-specific interactions.158 To avoid this limitation it is best to run compounds well below their solubility limits.

218

Figure E1. Scheme of a Saturation Transfer Difference – NMR Assay (STD-NMR). Adapted from Viegas.157

Figure E2. Scheme of Nuclear Overhauser Effect to determine STD. Adapted from Begley.156 219 Methods:

HGN-0201: NMR samples were prepared by diluting dialyzed protein (36 mg/mL) to 70

µM in D2O. Compound was assayed at a ligand concentration of 210 µM with 25 µL deuterated dimethyl sulfoxide (d6-DMSO) in a 500 µL sample volume. All experiments were conducted on a Varian Inova 500 MHz NMR spectrometer equipped with a 5 mm TBO BB-1H/13C/D Z-GRD

Z5723/0007 probe at a temperature of 25 °C. Screening was carried out using ligand-observed proton-based one-dimensional STD-NMR. Briefly, 64 scans were acquired over a 16 ppm sweep width at 32k scans, with a total recycle delay of 2.0 s. A low-power 15 ms spin-lock pulse was added to filter out low-level protein peaks to suppress the bulk water signal. STD-NMR pre- saturation was completed using a 2.4 s long train of Gaussian-shaped pulses with a spectral width of 500 Hz focused at -3 ppm, with reference irradiation set to 30 ppm. NMR spectra were processed and analyzed using MestReNova (Mestrelab).

HGN-0003: NMR samples were prepared by diluting dialyzed protein (36 mg/mL) to 70

µM in D2O. Compound was assayed at ligand concentrations of 3500 µM with D2O present in a

500 µL sample volume. The general procedure was followed as reported for HGN-0201.

Results:

Two compounds were studied using one-dimensional STD-NMR to determine the protons interacting with BpIspF. The compound, HGN-0201, synthesized for the development of a fluorescence polarization assay, SPR derived KD of 78 µM, was studied at a ratio of 30:1 by

STD-NMR. Proton mapping was determined using difference NMR. Difference NMR indicated the biphenyl ring, imidazole ring, phenol ring, and the N-dimethyl groups were in contact with 220 the enzyme (Figure E3). Furthermore, knowledge of the protons that are not interacting with the enzyme could be used to guide the synthesis of compounds that would potentially have increased affinity for BpIspF.

In order to map specific interactions of the control compound, HGN-0003, to BpIspF, it was also studied by STD-NMR. HGN-0003 was found to have a KD of 60 µM by SPR and a KD of 30 µM by ITC. HGN-0003 was run at a 50:1 ratio in STD-NMR. A limitation of this experiment was due to the exchangeable protons within HGN-0003, the amide NH proton and amine protons. Since these protons rapidly exchange with deuterium they were not observed in the NMR and could not be analyzed for interactions with the enzyme. The mapping of protons observed in the NMR, which interact with the enzyme, was determined to be the indole ring as well as the methylene group of HGN-0003 (Figure E4). Therefore, interactions of exchangeable protons, found in the hydroxamate function group, indicated to be important for binding in the crystal structure were not observed in the difference NMR. 221 IspF. Bp Figure E3. Determination of HGN-0201 protons that are observe saturation E3. Determination observe protonsfrom of HGN-0201 that are Figure transfer 222 IspF. Bp Figure E4. Determination saturation protons from transfer of HGN-0003 that observe Figure 223

1 H NMR (300 MHz, CD3OD): δ 8.55 (d, J = 8.4 Hz, 1H), 8.21 (d, J = 8.7 Hz, 1H), 8.02 (dd, J =

7.5 Hz, J = 1.2 Hz, 1H), 7.58 – 7.50 (m, 3H), 7.26 (d, J = 7.5 Hz, 1H), 7.05 – 7.02 (m, 2H), 6.83

– 6.80 (m, 2H), 6.76 (s, 1H), 4.03 (t, J = 6.6 Hz, 3H), 2.92 (t, J = 6.3 Hz, 3H), 2.88 (s, 6H).

(500 MHz, CD3OD): δ 7.62 (d, J = 7.5 Hz, 1H), 7.35 (d, J = 8.5 Hz, 1H), 7.13 – 7.09 (m, 2H),

7.05 – 7.02 (m, 1H), 3.53 (t, J = 7.0 Hz, 1H), 3.18 (dd, J = 14 Hz, J = 7.0 Hz, 1H), 3.00 (dd, J =

14 Hz,J = 7.0 Hz, 1H).

APPENDIX F – X-RAY CRYSTAL STRUCTURE STATISTICS

Table F1. Data processing and refinement statistics for BpIspF Q151E variant

Space Group C121 abc 116.8, 68.1, 60.4

ɣ 90.0, 96.6, 90.0

Resolution 1.72

Completeness 1.0

# Unique Reflections 50230

Rwork/Rfree 0.1659/0.2036

Burkholderia pseudomallei IspF 463

Zn2+ 3

H2O 458

RMSD

Bond length (Å) 0.011

Bond angle (o) 1.26

Ramachandran

Preferred (%) 0.98

Allowed (%) 0.02

Outliers (%) 0.0

225 Table F2. Data processing and refinement statistics for BpIspF in complex with HGN-0003

Space Group C121 abc 116.5, 68.1, 60.7

ɣ 90.0, 96.9, 90.0

Resolution 1.47

Completeness 0.99

# Reflections

Unique 79224

Rwork/Rfree 0.1800/0.1756

Res

Burkholderia pseudomallei IspF 474

Zn2+ 3

H2O 435

RMSD

Bond length (Å) 0.011

Bond angle (o) 0.92

Ramachandran

Preferred (%) 0.99

Allowed (%) 0.01

Outliers (%) 0.0

226 Table F3. Data processing and refinement statistics for BpIspF in complex with HGN-0289

Space Group C121 abc 116.7, 67.9, 60.5

ɣ 90.0, 96.5, 90.0

Resolution 1.51

Completeness 1.00

# Reflections

Unique 73745

Rwork/Rfree 0.1548/0.1740

Res

Burkholderia pseudomallei IspF 458

Zn2+ 3

H2O 459

RMSD

Bond length (Å) 0.014

Bond angle (o) 1.28

Ramachandran

Preferred (%) 0.99

Allowed (%) 0.01

Outliers (%) 0.0