A Dissertation

entitled

Structural, Enzymatic, and Inhibitory Studies of Two Mycobacterium tuberculosis-

Mycomembrane Lipid Esterases

by

Christopher M. Goins

Submitted to the Graduate Faculty as partial fulfillment of the requirements for the

Doctor of Philosophy Degree in Chemistry

______Dr. Donald R. Ronning, Committee Chair

______Dr. Peter R. Andreana, Committee Member

______Dr. John J. Bellizzi, Committee Member

______Dr. Scott M. Leisner, Committee Member

______Dr. Amanda Bryant-Friedrich, Dean College of Graduate Studies

The University of Toledo

May 2018

Copyright 2018, Christopher M. Goins

This document is copyrighted material. Under copyright law, no parts of this document may be reproduced without the expressed permission of the author. An Abstract of

Structural, Enzymatic, and Inhibitory Studies of Two Mycobacterium tuberculosis Mycomembrane Lipid Esterases

Christopher M. Goins

Submitted to the Graduate Faculty as partial fulfillment of the requirements for the Doctor of Philosophy Degree in Chemistry

The University of Toledo May 2018

Mycobacterium tuberculosis (Mtb), the causative agent of tuberculosis (TB) is notoriously difficult to treat due to its impervious cell wall. However, the biosynthetic pathways responsible for its protective barrier have proven to be an Achilles heel, as many TB drugs target the enzymes within these unique pathways. Based on this proven approach, we too have decided to target enzymes responsible for the construction and maintenance of the outermost lipid membrane, the mycomembrane, for drug development. The focus of this dissertation is on the structural, enzymatic, and inhibitory study of two Mtb mycomembrane lipid esterases. Both enzymes are secreted, essential, have similar protein folds and utilize classical catalytic triads, yet perform different types of chemistry.

The Antigen 85 complex (Ag85) of Mtb catalyzes the acyltransfer of the mycolic acids

(MA) to produce trehalose-dimycolate (TDM) or the mycolylarabinogalactan (mAG), two hallmark lipids of the unique mycomembrane. Three homologous mycolytransferases comprise the Ag85 complex (Ag85A, B, and C) and despite the term complex, act independently through a ping-pong catalytic mechanism. Based on early structural studies, an interfacial mechanism model was proposed that detailed substrate arrangement within Ag85s; however, numerous problems exist with this model. To address those concerns, we sought a crystal structure of the acyl-enzyme intermediate form of Ag85C; however MAs are highly insoluble. Therefore, Ag85C was successfully co-crystalized with tetrahydrolipstatin (THL), an FDA approved lipase inhibitor, iii which mimics core structural attributes of MAs. The Ag85-THL structure served as a basis for the modeling of the acyl-enzyme intermediate. Based on structural similarities to previously solved

Ag85 structures, mutagenic studies, and computational simulations, Ag85s were shown to undergo structural changes upon acylation that limit substrate hydrolysis and promote substrate transfer. Based on these findings, a new model of substrate binding was proposed that satisfies issues with the previous interfacial model and results in the chemical requirement for the β- hydroxyl of MAs for Ag85 catalysis. Outside the TB field, the Ag85C-THL structure is the second protein-THL structure to ever be solved and provides a new molecular perspective on lipase inhibition by THL.

Ebselen and thiopene compounds have previously been reported to inhibit Ag85C covalently and non-covalently, respectively. A library of ebselen and thiophene derivatives has been synthesized and tested for in vitro and in vivo properties. As a result of no clear structure activity relationship, two chemically dissimilar ebselen compounds were selected for further characterization. Based on the crystal structures of Ag85C in covalent complex with azido- ebselen and adamantyl-ebselen, key protein-drug interactions were assessed that influence in vitro inhibition properties. Furthermore, using differential scanning fluorimetry, ebselen modification was shown to significantly influence protein stability. Unfortunately, all thiophene derivatives displayed no inhibition towards Ag85C due to the loss of specificity to the sugar binding site.

Finally, screening Ag85C against a small set of lactone containing compounds and two drug libraries comprising over 1500 TB active compounds resulted in the identification of two scaffolds that can be utilized for further development targeting Ag85s.

The second lipid esterase of interest is encoded by the rv3802c gene, which resides in gene clusters responsible for cell wall biosynthesis. Rv3802c encodes an enzyme with thioesterase/phospholipase A activity, is retained in the cell wall, and has been shown to modulate lipid content of the mycomembrane as a result of cellular stress and overexpression. To further

iv characterize Rv3802, we set out to obtain the first crystal structure of the enzyme. The X-ray crystal structure of Rv3802 was solved with two molecules of polyethylene glycol (PEG) bound within the enzyme active site. On the basis of PEG binding, the lipid binding site of Rv3802 was structurally identified and characterized. Comparison of the Mtb Rv3802-PEG structure with an apo Mycobacterium smegmatis (Msmeg) ortholog structure resulted in the identification of dynamic regions required for lipid binding. The Rv3802-PEG structure provides a molecular basis for the binding of phosphatidylinositol-based substrates, further suggesting the enzyme responsible for the decomposition of glycerophospholipids of the mycomembrane.

In efforts to develop inhibitors of Rv3802, two fluorescence based assays were developed and compared. Using these assays, two drug libraries were screened against Rv3802, which resulted in the identification of multiple compounds with low micromolar affinity. Unfortunately, identified compounds failed to produce morphological differences in Msmeg cells and displayed modest in vivo activity against two strains of Mtb.

THL has been shown capable of inhibiting numerous human lipases and covalently modifying 261 lipid esterases in mycobacteria, two of which are Ag85C and Rv3802. To better understand how THL inhibits both enzymes, a library of stereoderivatives was screened against both enzymes. We found that the stereochemistry of the β-lactone ring is important for cross enzyme reactivity, while the stereochemistry of the peptidyl side arm influences enzyme specificity and stability of the covalent THL-enzyme complex. Observed in vitro inhibition data was rationalized using the Ag85C-THL structure and molecular modeling of Rv3802 and THL.

Findings within this dissertation advance the structural and enzymatic understanding of two essential Mtb lipid esterases, while providing a basis for future development of novel inhibitors specific to Ag85C and Rv3802.

v

I dedicate this dissertation to my late grandfather who bestowed upon me the gift of scientific curiosity at an early age. He inspired and developed my interest in understanding how things work, from bridges to radios. While my interest may have shifted from manmade machines to the micro machinery of biology, my desire to never stop learning is owed to him.

Acknowledgements

The scientific findings contained within this dissertation would not have been possible without the help, guidance, and collaborative efforts of multiple individuals. I would like to thank my advisor, Dr. Donald Ronning, for the guidance and dissemination of biochemical and structural biology knowledge over the course of my graduate studies.

Previous lab members Dr. Jared Lindenberger and Dr. Lorenza Favrot were critical in the early stages of my research training, for that I am grateful. I would also like to thank all my fellow lab members, past and present, particularly Michael Banco for the intellectual conversations and making graduate school enjoyable. Additionally, Celine Schreidah, an undergraduate researcher in our group, was instrumental in the crystallization of Rv3802.

I would like to thank the following collaborators for their respective contributions: Steven Dajnowicz and Dr. Jerry Parks for in silico energetic calculations and MD simulations, Dr. Steven Sucheck and Dr. Sandeep Thanna for ebselen and thiophene derivatives, Dr. Mary Jackson’s lab for testing compounds against Mtb, Dr.

Todd Lowary’s lab for acyl-arabinose compounds, and Dr. George A. O’Doherty’s lab for THL stereoderivatives. I would also like to thank Dr. Panne Burckel and Dr. Lief

Hanson for their assistance with SEM and MALDI-MS, respectively. Finally, my path to a PhD would not have been possible without all the teachers and professors that have educated, trained, and inspired me along the way.

vi

Table of Contents

Abstract ...... iii

Acknowledgements ...... vi

Table of Contents ...... vii

List of Tables ...... xv

List of Figures ...... xvi

List of Abbreviations ...... xx

List of Symbols ...... xxii

1 Introduction ...... 1

1.1 Tuberculosis ...... 1

1.1.1 Background and disease state ...... 1

1.1.2 Current treatment and rise of drug resistance ...... 2

1.2 Mtb cell envelope ...... 4

1.2.1 General architecture and composition ...... 4

1.2.2 Mtb lipids ...... 8

1.2.2.1 Mycolic acids ...... 8

1.2.2.2 Glycerophospholipids ...... 11

1.3 α/β-hydrolase fold: lipid esterases ...... 13

1.4 Project overview ...... 16

1.4.1 Antigen 85C ...... 16 vii

1.4.2 Rv3802 ...... 16

1.4.3 Tetrahydrolipstatin Inhibition ...... 17

2 Antigen 85 Catalysis ...... 19

2.1 Background ...... 19

2.1.1 Antigen 85s ...... 19

2.1.2 Ag85 structure and substrate binding sites ...... 20

2.1.3 Mechanism of catalysis ...... 23

2.1.4 Substrate coordination ...... 25

2.2 Methods ...... 27

2.2.1 Molecular cloning of Ag85C and mutants ...... 27

2.2.2 Expression and purification of Ag85C and mutants ...... 27

2.2.3 Ag85C Activity assay ...... 28

2.2.3.1 Mutant activity ...... 29

2.2.3.2 kinact/KI determination for tetrahydrolipstatin ...... 29

2.2.4 Crystallization and structure determination ...... 30

2.2.5 Structural and sequence alignments ...... 31

2.3 Results ...... 31

2.3.1 Mutant and wild type cloning, expression, and purification ...... 31

2.3.2 Tetrahydrolipstatin inhibition ...... 33

2.3.3 Ag85C tetrahydrolipstatin structure ...... 34

2.3.4 Structural comparisons and conservation of residues of interest ...... 37

2.3.5 Mutant activity ...... 40

2.4 Discussion ...... 41

viii

2.4.1 Ag85C-THL structure and inhibition...... 41

2.4.2 Promoting substrate transfer over hydrolysis ...... 42

2.4.3 Substrate coordination and acceptor molecule activation ...... 47

2.5 Conclusions ...... 52

3 Inhibiting Antigen 85C ...... 54

3.1 Background ...... 54

3.1.1 Antigen 85s as a drug target ...... 54

3.1.2 Previously developed inhibitors ...... 55

3.1.3 Inhibition by ebselen ...... 57

3.1.4 Covalent warheads of serine esterases ...... 59

3.2 Methods ...... 60

3.2.1 Protein expression and purification ...... 60

3.2.2 Evaluation and characterization of ebselen derivatives ...... 60

3.2.2.1 In vitro inhibition (% activity and IC50 determination) .....60

3.2.2.2 Crystallization and structure determination ...... 61

3.2.2.3 kinact/Ki determination of selected derivatives ...... 63

3.2.2.4 Differential scanning fluorimetry ...... 64

3.2.3 In vitro screening of thiophene derivatives ...... 64

3.2.4 In vitro screening of β and γ-lactones ...... 65

3.2.5 Drug library screening ...... 65

3.2.5.1 Screening parameters ...... 65

3.2.5.2 Hit validation ...... 66

3.3 Results ...... 66

ix

3.3.1 Evaluation and characterization of ebselen derivatives ...... 66

3.3.1.1 in vitro inhibition ...... 66

3.3.1.2 Crystal structures of Ag85C and ebselen derivatives ...... 70

3.3.1.3 Influence of ebselen on protein stability ...... 74

3.3.2 Evaluation of thiophene derivatives...... 75

3.3.3 Evaluation of β and γ-lactones ...... 76

3.3.4 Drug library screening ...... 77

3.4 Discussion ...... 79

3.4.1 Ebselen derivatives ...... 79

3.4.2 Thiophene derivatives ...... 85

3.4.3 β and γ-lactones...... 86

3.4.4 Identified lead molecules ...... 87

3.5 Conclusion ...... 88

4 Structural, enzymatic, and biological characterization of Rv3802 ...... 90

4.1 Background ...... 90

4.1.1 rv3802c gene and encoded enzymatic activity ...... 90

4.1.2 Msmeg ortholog structure ...... 91

4.1.3 Rv3802 biological function...... 91

4.2 Methods ...... 92

4.2.1 Wild type and mutant cloning ...... 92

4.2.2 Protein expression and purification ...... 94

4.2.3 Activity assay development ...... 95

4.2.3.1 Resorufin butyrate assay ...... 95

x

4.2.3.2 4-methylumbelliferyl heptanoate assay ...... 96

4.2.4 Crystallization and structure determination ...... 97

4.2.5 Structural analysis and sequence conservation ...... 98

4.2.6 Cysteine quantification ...... 98

4.2.7 WT and mutant activity ...... 99

4.2.8 Phospholipase A assay ...... 99

4.2.9 Evaluation of acyl-arabinose substrates ...... 100

4.3 Results ...... 100

4.3.1 Molecular cloning, expression, and purification...... 100

4.3.2 Activity assays ...... 101

4.3.2.1 Resorufin butyrate assay ...... 101

4.3.2.2 4-methylumbelliferyl heptanoate assay ...... 102

4.3.3 Rv3802 structure and lipid binding site ...... 103

4.3.4 Structural movement upon lipid binding ...... 107

4.3.5 Substrate organization and mechanism of catalysis...... 109

4.3.6 Catalytic mutant activity ...... 111

4.3.7 Protein dynamics required for catalysis ...... 112

4.3.8 Cysteine quantification ...... 114

4.3.9 Evaluation of acyl-arabinose substrates ...... 115

4.4 Discussion ...... 115

4.4.1 1st and 2nd generation Rv3802 assay comparison ...... 115

4.4.2 Rv3802 Structure and PLA classification ...... 116

4.4.3 Mechanism of catalysis ...... 118

xi

4.4.4 Dynamics of lipid binding ...... 120

4.4.5 Structure-based insights on biological function ...... 122

4.5 Conclusion ...... 125

5 Inhibiting Rv3802 ...... 126

5.1 Background ...... 126

5.1.1 Rv3802 as a drug target ...... 126

5.1.2 Inhibition by tetrahydrolipstatin ...... 126

5.2 Methods ...... 128

5.2.1 Protein expression and purification ...... 128

5.2.2 Drug library screening ...... 128

5.2.2.1 Screening parameters ...... 128

5.2.2.2 “Hit” validation ...... 129

5.2.2.3 Ki determination ...... 129

5.2.3 In vivo effects on Msmeg by Mtb active compounds ...... 129

5.2.3.1 Growth inhibition of Msmeg ...... 129

5.2.3.2 Scanning electron microscopy of Msmeg treated cells ....130

5.3 Results ...... 131

5.3.1 Identification of lead molecules ...... 131

5.3.1.1 TB alliance library ...... 131

5.3.1.2 GSK library ...... 133

5.3.2 Biological activity of lead molecules ...... 135

5.3.2.1 Effects on Msmeg growth ...... 135

5.3.2.2 Influence on Msmeg morphology...... 137

xii

5.4 Discussion ...... 138

5.4.1 Structure based rationale for Rv3802 inhibitors ...... 138

5.4.2 Biological evaluation of lead molecules ...... 141

5.5 Conclusions ...... 143

6 Influence of THL stereochemistry on Ag85C & Rv3802 inhibition ...... 145

6.1 Background ...... 145

6.1.1 Reversible covalent inhibition by tetrahydrolipstatin ...... 145

6.1.2 Non-specific inhibition of Mtb lipid esterases ...... 147

6.1.3 THL inhibition of Ag85C and Rv3802 ...... 148

6.2 Methods ...... 149

6.2.1 Protein expression and purification ...... 149

6.2.2 Initial screening of tetrahydrolipstatin stereoderivatives ...... 149

6.2.3 Enzymatic characterization ...... 150

6.2.3.1 Ag85C kinact/Ki determination ...... 150

6.2.3.2 Analysis of covalent inhibition of Rv3802 ...... 151

6.2.3.3 Rv3802 IC50 determination ...... 151

6.2.4 Molecular modeling of Rv3802 tetrahydrolipstatin complex ...... 152

6.3 Results ...... 153

6.3.1 Initial inhibition screening ...... 153

6.3.2 Ag85C inhibition ...... 154

6.3.3 Rv3802 inhibition ...... 156

6.4 Discussion ...... 159

6.4.1 Ag85C inhibition ...... 159

xiii

6.4.2 Rv3802 inhibition ...... 161

6.4.3 Influence of stereochemistry on cross reactivity & specificity ...... 164

6.5 Conclusions ...... 164

References ...... 166

xiv

List of Tables

1.1 First and second line treatments for TB ...... 4

2.1 DNA primers used for Ag85C S148A and S148T mutants ...... 27

2.2 Ag85C-THL structure X-ray and refinement data ...... 36

3.1 Ag85C Ebs derivative structures X-ray and refinement data ...... 71

3.2 Ebs derivative inhibition summary ...... 80

4.1 DNA primers used for Rv3802 WT and mutant cloning ...... 94

4.2 Rv3802-PEG structure X-ray and refinement data ...... 104

4.3 Comparison of 1st and 2nd generation Rv3802 assays ...... 116

5.1 Properties of Rv3802 inhibitors ...... 143

6.1 IC50 values for THL stereoderivatives ...... 158

xv

List of Figures

1-1 Mtb cell envelope ...... 6

1-2 Mycolic acid species ...... 10

1-3 Glycerophospholipids of Mtb ...... 13

1-4 Typical α/β-hydrolase fold ...... 14

1-5 General esterase mechanism ...... 15

2-1 Ag85 structures ...... 21

2-2 Substrate binding sites of Ag85s...... 23

2-3 Catalytic mechanism of TDM biosynthesis by Ag85C ...... 24

2-4 Previously proposed substrate coordination ...... 26

2-5 Ag85C fluorescence based assay ...... 29

2-6 Sequence confirmation of Ag85C mutants ...... 32

2-7 SDS-PAGE of purified WT Ag85C and mutants ...... 33

2-8 THL inhibition ...... 34

2-9 Inhibition progress curves ...... 34

2-10 Ag85C THL structure ...... 37

2-11 Ag85C hydrophobic pocket and global changes upon THL modification ...... 37

2-12 Active and Inactive forms of Ag85C ...... 39

2-13 Conservation of amino acids involved in His260 sequestration ...... 40

2-14 Enzymatic activity of S148 mutants ...... 41 xvi

2-15 MD simulations of Ag85C acyl-enzyme forms ...... 44

2-16 REMD simulation of Ag85C-MA-His260cat ...... 46

2-17 MD simulation of mycolated Ag85C with acceptor molecule ...... 48

2-18 Direct and Indirect acceptor molecule activation ...... 49

2-19 Visualization of the Ag85 catalytic cycle ...... 51

3-1 Previous Ag85 inhibitors ...... 56

3-2 Mechanism of ebselen modification ...... 57

3-3 Ebselen modified Ag85C ...... 58

3-4 Ebselen derivative library ...... 59

3-5 Lactone scaffold ...... 60

3-6 Screening of ebselen derivatives ...... 67

3-7 Ebselen derivative IC50 Curves ...... 68

3-8 Selected ebselen derivatives for further characterization ...... 69

3-9 kinact/Ki Determination for selected derivatives ...... 70

3-10 Fo-Fc omit map for ebselen derivatives ...... 72

3-11 Ag85C-ebselen derivative interactions ...... 73

3-12 Ag85C structural changes upon ebselen derivative modification...... 74

3-13 Raw DSF data ...... 75

3-14 Averaged DSF melt peak ...... 75

3-15 Second generation thiophene library screening ...... 76

3-16 Lactone screening results ...... 76

3-17 TB alliance drug screening results ...... 77

3-18 Dose dependence of TB alliance hits ...... 78

xvii

3-19 GSK drug screening results ...... 79

3-20 Dose dependence of GSK hits ...... 79

3-21 Energetics of Ag85C-ebselen interactions ...... 82

3-22 Thiophene derivatives ...... 86

3-23 Tested lactones ...... 87

3-24 Structures of identified drug screening hits ...... 87

4-1 Rv3802 Msmeg ortholog structure ...... 91

4-2 SDS-PAGE of WT and Rv3802 mutants ...... 101

4-3 RfB assay and Michaelis-Menten kinestics ...... 102

4-4 4MH Z’ determination ...... 103

4-5 Mtb Rv3802 structure ...... 105

4-6 Rv3802 fatty acid binding site ...... 107

4-7 Identification of dynamic regions ...... 109

4-8 Substrate binding sites of Rv3802 ...... 110

4-9 Arrangement of PEG and water in Rv3802 active site ...... 111

4-10 Catalytic mutant activity ...... 112

4-11 Influence of helical movement on Rv3802 catalysis ...... 113

4-12 Cysteine quantification ...... 114

4-13 Evaluation of acyl-arabinose hydrolysis ...... 115

4-14 Rv3802 secondary structure and transmembrane predications ...... 118

4-15 MD simulation of apo Rv3802 ...... 121

4-16 Model of Rv3802 with PI bound ...... 124

5-1 Rv3802 THL derivatives...... 127

xviii

5-2 Rv3802 TB Alliance drug screen results ...... 132

5-3 Dose dependence of TB Alliance “hits” ...... 133

5-4 Ki determination for best inhibitors ...... 133

5-5 Rv3802 GSK library screening ...... 134

5-6 Dose dependence of GSK “hits” ...... 134

5-7 Treated and untreated Msmeg cultures ...... 135

5-8 Influence of inhibitors on Msmeg cultures under heat stress ...... 136

5-9 Colonies on agar plates doped with inhibitors ...... 137

5-10 SEM images of treated and untreated Msmeg cells ...... 138

5-11 Rv3802 lipid binding site for drug targeting...... 139

5-12 Structures of lead molecules ...... 140

5-13 Best inhibitors modeled into Rv3802 lipid binding site ...... 141

6-1 Mode of THL inhibition...... 146

6-2 Human FAS-THL structure ...... 147

6-3 Screening of THL stereoderivatives against Ag85C and Rv3802 ...... 153

6-4 THL and 2’-epi-THL kinact/KI Determination for Ag85C ...... 155

6-5 THL and 2’-epi-THL reaction progress curves ...... 156

6-6 Reversible covalent inhibition of Rv3802 by THL...... 157

6-7 Rv3802 IC50 curves for THL stereoderivatives ...... 158

6-8 Ag85C-THL binding site ...... 160

6-9 Model of THL binding to Rv3802 ...... 163

xix

List of Abbreviations

AG ...... Arabinogalactan Ag85 (A,B,C) ...... Antigen 85 A, B, or C

CDP-DAG ...... Cytidine diphosphate-diacylglycerol CL ...... Cardiolipin CM ...... Cytoplasmic membrane

DSF ...... Differential scanning fluorimetry DFT ...... Density functional theory

Ebs...... Ebselen Em ...... Emission ES ...... Extracellular space Ex ...... Excitation

FA ...... Fatty acid FAS ...... Fatty acid synthase

GSK...... GlaxoSmithKline GuHCl ...... Guanidinium hydrochloride

HEPES ...... 4-(2-hydroxymethyl)-1-piperazineethanesulfonic acid HIV ...... Human Immunodeficiency Virus

IC50 ...... Inhibitory concentration for 50 % inhibition

LM...... Lippomannan

MA ...... Mycolic acid mAG ...... Mycolylarabinogalactan Mavium ...... Mycobacterium avium Mbovis ...... Mycobacterium bovis MCB ...... Monochlorobimane MD ...... Molecular dynamics MDR-TB ...... Multiple-drug resistant Tuberculosis MIC ...... Minimal inhibitory concentration xx

MPD ...... (+/-)-2-Methyl-2,4-pentanediol Msmeg ...... Mycobacterium smegmatis Mtb ...... Mycobacterium tuberculosis

NIH ...... National Institute of Health

PAT ...... Polyacylated Trehalose PE ...... Phosphatidylethanolamine PG ...... Peptidoglycan PI ...... Phosphatidylinositol PIM ...... Phosphatidylinositol Mannoside Pks13 ...... Polyketide synthase 13 PLA ...... Phospholipase A PS ...... Phosphatidylserine

REMD ...... Replica exchange molecular dynamics RfB ...... Resorufin butyrate RFU ...... Relative fluorescence units RMS ...... Root mean square RMSD ...... Root mean square distance RMSF ...... Root mean square fluctuation

SAR ...... Structure activity relationship SEM ...... Scanning electron microscopy STD ...... Standard deviation

TB ...... Tuberculosis TDR-TB ...... Total-drug resistant Tuberculosis THL ...... Tetrahydrolipstatin TRIS ...... tris(hydroxymethyl0aminomethane TS ...... Transition state

WT ...... Wild type

XDR-TB ...... Extensively-drug resistant Tuberculosis

xxi

List of Symbols

˚ ...... Degree ± ...... Plus or minus Å ...... Angstrom

μM ...... micromolar μL ...... microliter σ ...... sigma (signal) λ ...... Wavelength

M ...... Molar mM ...... millimolar nM ...... nanomolar mL ...... milliliter C ...... Celsius Ki ...... Inhibitor binding constant Km ...... Michaelis-Menten constant Vmax ...... Maximal reaction velocity X ...... Time V’o ...... uninhibited reaction rate V’ ...... Inhibited reaction rate kinact ...... Rate of inactivation kobs ...... Observed reaction rate Y ...... Product concentration Yo ...... Product concentration time zero Vi ...... Initial rate

xxii

Chapter 1

Introduction

1.1 Tuberculosis

1.1.1 Background and disease state

Tuberculosis (TB) is a centuries old disease that still plagues modern society1. It is believed that one third of the world’s population is infected with the Mycobacterium tuberculosis (Mtb) bacilli, the causative agent of TB1. However, those infected with Mtb will not always develop an active infection in their lifetime2. Only 5 to 15 % of infected individuals will develop an active form of TB, this translated to approximately 10.4 million active cases of TB in 20151, 2. Most cases of TB are considered curable; nevertheless, approximately 1.4 million people succumbed to the disease in 20151. The number of mortalities increases to ~1.8 million deaths when considering patients co- infected with the Human Immunodeficiency Virus (HIV)1. Increasing rates of patients co- infected with HIV and TB is of great concern as treatment is further complicated1, 3. This is reflected in the 2015 TB statistics as patients co-infected with HIV account for only 11

% of total active TB cases, yet account for 22 % of total deaths1. Due to prevalence, mortality rate, and length of treatment, TB is considered one of the top 10 infectious diseases our world faces1. 1

In 1882, Robert Koch discovered Mtb to be the causative agent of TB4. Mtb is a highly aerobic gram positive bacterium, which slowly replicates every 12 to 24 hours5.

Mtb primarily afflicts the pulmonary system and therefore is readily spread through coughing or speaking2. Following inhalation of aerosolized bacilli, Mtb is phagocytosed by alveolar macrophages in the lungs2, 6. However, Mtb has evolved to evade cellular destruction by macrophages and can even exist indefinitely within macrophage cells in a latent state2, 7. The infection progresses upon the rupture of the macrophage cells by replicating bacilli, which then spread to the alveolar epithelium and pulmonary lymph nodes, resulting in the formation of granulomas by responding immune cells2. Mtb has developed ways to evade T-cell recognition, which limits initial immune response2, 8.

Again, Mtb can enter a latent phase and reside within the granulomas indefinitely or the bacilli can continue to divide, rupturing the granulomas and begin to spread throughout the lungs and to other organs2, 9. At this stage, the infection would be considered clinically “active,” the patient is considered contagious, and would begin to show signs of chest pain, coughing, fever, and loss of weight2.

1.1.2 Current treatment and rise of drug resistance

The current medical treatment of TB is far from modern, with little change in antibiotic regiments within the past 5 decades10, 11. The first antibiotic used to treat TB was streptomycin in the 1940s; however, drug resistance quickly arose and streptomycin is now reserved as a second line drug10, 12. Following streptomycin, TB was successfully treated with just isoniazid; however, treatment required 18 months to prevent the disease from returning10. Treatment could be shortened to 9 months and antibiotic resistance reduced when isoniazid is given in combination with rifampin10. Currently, drug sensitive 2

TB is treated for 2 months with a combination of isoniazid, rifampin, and pyrazinamide followed by 6 more months of just isoniazid and rifampin10. Long, multi-drug approaches to treat TB are required due to the ability of Mtb to enter into a latent or dormant state, in doing so the bacteria become much more difficult to kill than when actively dividing10, 11,

13, 14. Latency or non-replicating bacteria is problematic for two reasons: 1) drugs that target biosynthetic pathways that are only required for cell replication are not as efficient and 2) drug accessibility to the bacteria encapsulated by granulomas10, 11. These two problems are best highlighted in the significant decrease in the efficacy of isoniazid

(inhibits cell wall biosynthesis) against latent Mtb when compared to the stellar antibacterial properties against active Mtb11, 15.

Most cases of reported TB occur within underdeveloped or developing nations in

Africa and Asian countries1. The geographical location of TB cases is problematic for a variety of reasons: general access to health care, disease diagnosis, access to drugs, cost of treatment, and poorly informed populations are major issues with these regions1. The stated issues influence the treatment of any disease; however, given the required 6 to 9 month treatment of multiple pills a day, these geographical concerns are extremely problematic for the treatment of TB1, 10. As a result of failed adherence to prescribed antibiotic treatments and poor diagnosis, drug resistant strains of Mtb have rapidly arisen in modern times and are of major concern1. Drug resistant forms of TB can be characterized into multiple-drug resistant (MDR-TB), extensively-drug resistant (XDR-

TB), or total-drug resistant (TDR-TB). MDR-TB is resistant to one of the four first line drugs, XDR-TB is resistant to all first line drugs, and TDR-TB is resistant to all known drugs used to treat TB1, 11, 16, 17. A list of first and second line drugs used to treat TB is 3

given in table 1-1. In 2015, there were ~500,000 new cases of MDR-TB (~ 5 % of all TB cases) and an additional 100,000 of MDR-TB cases that were also resistant to rifampicin1. Treatment of MDR-TB and XDR-TB require the heavy use of second and third line drugs (not mentioned in table 1-1) for durations up to and exceeding 20 months; these rigorous treatment plans can be successful, but result in severe side effects10, 18-20.

Due to the rise of drug resistant TB and the lack of a new first-line drugs since the

1960’s, the need for novel antibiotics is critical1.

Table 1.1: first and second line drugs used to treat TB, respective date discovered, target, and mode of action. Table adapted from Zumla et al. Nat. Rev. Drug Discov., 201311 Drug (Discovery Date) Target Mode of Action First line drugs Isoniazid (1952) Enoyl-ACP (acyl carrier protein) Mycolic acid synthesis reductase Rifampicin (1963) RNA polymerase, β-subunit DNA transcription Pyrazinamide (1954) S1 component of 30S ribosomal Protein synthesis subunit Ethambutol (1961) Arabinosyl transferase Arabinogalactan synthesis Second line drugs Para-amino salicylic acid (1944) Dihydropteroate synthase Folate synthesis Streptomycin (1944) S12 and 16SrRNDA components Protein Synthesis of 30S ribosomal subunit Ethionamide (1961) Enoyl-ACP (acyl carrier protein) Mycolic acid synthesis reductase Ofloxacin (1980) DNA gyrase and topoisomerase DNA supercoiling Capreomycin (1963) Interbridge B2a between 30S and Protein synthesis 50S ribosomal subunits Kanamycin (1957) 30S ribosomal subunit Protein synthesis Amikacin (1972) 30S ribosomal subunit Protein Synthesis Cycloserine (1955) D-alanine racemase and ligase Peptidoglycan synthesis

1.2 Mtb cell envelope

1.2.1 General architecture and composition

Mtb is a rod shaped bacillus that is considered a Gram positive bacterium21.

However, due to the unique waxy composition of the Mtb cell envelope, identification of

4

Mtb is conducted using acid fast (Ziehl-Neelsen) staining as mycobacteria do not retain crystal violet stain used for gram staining21, 22. The cell envelope of Mtb can be separated into four unique components: the cytoplasmic membrane (CM), the periplasmic region composed of the peptidoglycan (PG) - arabinogalactan (AG), the mycomembrane (MM), and the extracellular space (ES) or often referred to as the outer capsule23. A cartoon representation of the Mtb cell envelope is given in Figure 1-1. In Mycobacterium smegmatis (Msmeg), a non-pathogenic mycobacteria species commonly used to study

Mtb, the thickness of the CM and MM were determined to be ~7 and ~8 nM, respectively24. The periplasmic region between the CM and MM consisting of the PG and

AG is approximately ~20 nM in width. The cell envelope of M.smeg is therefore ~ 35 nM in width24. Interestingly, it was found that MDR and XDR Mtb have overall thicker PG layers and therefore thicker cell walls than drug susceptible Mtb25.

5

Figure 1-1: Mtb cell envelope. Cartoon representation of cell envelope not drawn to scale. Abbreviations are as follows: cytoplasmic membrane (CM), peptidoglycan (PG), arabinogalactan (AG), mycomembrane (MM), and extracellular space (ES). Figure adapted from Jackson Cold Spring Harbor perspectives in medicine, 201423.

The innermost CM is composed mostly of glycerophospholipids23.

Glycerophospholipids found within the CM that are not unique to Mtb include phosphatidylglycerol, phosphatidylinositol (PI), phosphatidylserine (PS), phosphatidylethanolamine (PE), and cardiolipin (CL)23. Unique to mycobacteria are the di, tetra, and hexa mannosylated forms of PI, those being the phosphatidylinositol mannosides (PIM)23, 26, 27. PIM species are now believed to account for a major portion of the CM28. Additionally, PIM species can be further mannosylated to form lipomannan

(LM) and glycosylated with arabinose to form ManLAM, these lipoglycans extend into the periplasmic region and are also found in the outer leaflet of my mycomembrane23, 29.

6

Countering the CM is the free-floating PG composed of repeating units of N- acetylglucosamine and acetyluramic acid23, 30. The terminal galactose of AG is covalently linked to the PG through a phosphoryl-N-acetylglucosaminosylrhamnosyl linkage30. The

AG consists of repeating units of β(1-5) or β(1-6) O-linked galactose that are then modified at the C5 position by repeat units of either α(1-5), α(1-3), or β(1-2) linked arabinose monomers30. The AG is capped by a branched hexaarbinofuranosyl moiety, which in tern is esterified on the 5-hydroxyl of the terminal arabinose by mycolic acid

(MA)30.

The unique mycomembrane is a highly hydrophobic lipid bilayer and is composed mostly of MA containing species (60-90 carbons in length)23, 28. The inner leaflet of the mycomembrane is covalently linked to the AG, consisting primarily of mycolated AG or referred to as the mycolylarabinogalactan (mAG)23. PIM and polyacylated trehalose

(PAT) species can also be found within the inner leaflet of the mycomembrane23.

Countering the mAG is the outer leaflet of the mycomembrane. This free-floating outer layer is mostly comprised of trehalose 6’-monomycolate (TMM), trehalose 6,6’- dimycolate (TDM) or known as cord factor, sulfolipids, PIM, LM, and ManLAM; however, the exact composition remains uncertain23, 31, 32. The lipid rich mycomembrane of Mtb provides a protective barrier for the organism and represents another hurdle that must be crossed in antibiotic developement33.

Extending into the ES from the outer layer of the mycomembrane are the saccharide groups of the lipoglycans PIM, LM and ManLAM23. The ES or outer capsule is composed mostly of large polysaccharides (D-glucan, D-arabino-D-mannan, and D- mannan) and is considered loosely attached to the mycomembrane34, 35. In addition to 7

polysaccharides, the ES is composed of proteins, specifically, 205 were found in cell culture filtrates23, 36. Despite the waxy nature of Mtb, only 1-6 % of the outer capsule consists of lipids35.

Aside from providing an exceptional environmental barrier for the organism, the lipids, lipoglycans, polysaccharides, and proteins that make up the Mtb cell envelope are directly involved in pathogenesis and virulence23. During initial macrophage infection,

ManLAM and PIM bind to mannose receptors of macrophages, allowing for uptake37-39.

These same lipids in addition to TDM and sulfolipids have been associated with the prevention of phagosome maturation23. Additionally, the lipids and glycolipids of Mtb have been associated with the immune response. For example, TDM, TMM, MA, sulfolipids, ManLAM, and PIM species are all considered CD1 antigens23, 40, 41. Aside from lipids of the cell envelope, secreted proteins found within the periplasmic and ES regions can elicit strong immune responses, a well-studied example being the secreted

Antigen 85 (Ag85A, Ag85B, and Ag85C) mycolyltransferases42-44. Much effort has gone into understanding the function, biosynthesis, and inhibition of lipids, glycolipids, and lipid associated proteins of Mtb due to their role in general cell viability and pathogenesis.

1.2.2 Mtb lipids

1.2.2.1 Mycolic Acids

The hallmark lipids of mycobacteria are the large, hydrophobic MAs found within the mycomembrane and ES23. The quantity, size and structure of MAs can differ between mycobacteria species28, 45, 46. In general, MAs have two branches, the shorter α-chain and 8

the longer meromycolate chain (Figure 1-2A)45. Indeed MAs are highly hydrophobic, the

α-chain in Mtb is 20 to 26 carbons in length and the meromycolate chain is between 42 to

62 carbons in length47. The meromycolate chain differs with regard to the presence of 1 or 2 double bonds or cyclopropanes, methoxy, methyl, or ketone adducts45. These chemical modifications are consistent in location along the meromycolate chain, an example of a MA with a keto cis-cyclopropane configuration is depicted in Figure 1-

2A45. The unique structures of MAs can influence cell wall fluidity, pathogenesis, and virulence48, 49.

The biosynthesis of MAs occur within the cytoplasm of the cell, is energetically costly, and is believed to require over 30 enzymes45. The two alkyl chains of MAs are generated independent of each other in respective fatty acid synthase (FAS) cycles45. The

α-chain is generated through the elongation of acetyl-CoA by the multi-domain Rv2524, which comprises FAS I45, 50. Following carboxylation of the elongated acyl-CoA fatty acid by AccD4, biosynthesis of the α-chain component (carboxyacyl-CoA) of MA is complete45, 51. Biosynthesis of the larger and diverse meromycolate chain occurs through multiple iterations of the multi-protein FAS II cycle45. It should be noted that inhibition of the NADH dependent enoyl reductase InhA of FAS II by isoniazid and ethionamide prohibits MA biosynthesis and thereby hindering Mtb growth15, 45. Biosynthesis of the meromycolate chain is complete upon activation of the elongated fatty acid by FadD32, generating meromycoloyl-AMP45, 51. The two precursor fatty acids of MA, carboxylacyl-

CoA and meromycoloyl-AMP, are joined together through a condensation reaction facilitated by the multi-domain, polyketide synthase 13 (Pks13) to produce a mycolic β- ketoester52. The mycolic β-ketoester is transferred from the thioesterase domain of Pks13 9

to the 6-hydroxyl of trehalose to produce trehalose mycolic β-keteoester53. The final step of MA biosynthesis is the reduction of the β-ketone of trehalose mycolic β-ketoester to a

β-hydroxyl by the CmrA reductase to produce the final molecule of TMM54. TMM is exported from the cytoplasm by the MmpL3 transporter to the periplasmic space where it can be utilized for the production of the mycomembrane44, 45, 55. The Ag85 mycolytransferases facilitate the transfer of MA from the 6-hydroxyl of TMM to the 6’- hydroxyl of another molecule of TMM to produce TDM (Figure 1-2B) or to the 5- hydroxyl of the terminal arabinose of the AG to form the mAG (Figure1-2C)30, 44, 56.

Figure 1-2: Mtb MA and MA lipids of the mycomembrane. A) Depicted is a keto cis- cyclopropane MA with respective alkyl-chains and β-hydroxyl labeled. B) TDM is a further mycolated form of TMM on the 6’-hydroxyl. C) Depicted is a mono mycolated penta arabinose on the 5-hydroxyl. Figure adapted from Goins et al. Journal of Biological Chemistry, 201857.

10

1.2.2.2 Glycerophospholipids

While MA substituents may be the hallmark lipids of mycobacteria, another class of mycobacterial lipids that are of equal if not greater importance than MAs are the glycerophospholipids. As stated early, glycerophospholipids mainly comprise the CM and play critical roles in mycomembrane fluidity and Mtb pathogenesis23. The shared scaffold of phosphatidylglycerol, PE, PS, PI, and PIM species is 1,2-diacyl-sn-glycerol 3- phosphate or phosphatidic acid (Figure 1-3)58.

Phosphatidic acid is synthesized in two steps with the respective acyl chains embedded in the CM during synthesis58. First, glycerol phosphate acyltransferase catalyzes an acyl transfer from a molecule of acyl-CoA to the sn2 position of Glycero-3- phosphate, forming 2-acyl-sn-glycerol-3-phosphate58. A second molecule of acyl-CoA is then used by acylglycerol phosphate acyltransferase to transfer the second acyl chain to the sn1 position of 2-acyl-sn-glycerol-3-phosphate, producing phosphatidic acid58. The respective fatty acids (FA) used for phosphatidic acid synthesis are synthesized by FAS I, the same enzyme responsible for the synthesis of the α-chain of MAs59. Common FAs used in phosphatidic acid synthesis are palmitic, oleic, and tuberculostearic acid (Figure

1-3B)60. Palmitic and oleic acids are commonly found at the sn2 position while tuberculostearic acid is found at the sn1 position of phosphatidic acid61. Before synthesis of the respective glycerophospholipids can occur, phosphatidic acid is activated by a CTP synthetase through the addition of a phosphoester linked molecule of CMP to produce cytidine diphosphate-diacylglycerol (CDP-DAG)58, 62.

CDP-DAG can be utilized in three distinct pathways to generate the respective glycerophospholipids: pathway 1 produces phosphatidylglycerol and CL, pathway 2 11

yields PS and PE, and pathway 3 yields the PI and PIM species58. Pathway 1 proceeds through the removal of CMP and addition of glycerol-3-phosphate by a phosphatidylglycerol phosphate synthetase followed by a dephosphorylation event to yield phosphatidylglycerol58. Phosphatidylglycerol can be converted into CL by a CL synthase which utilizes a molecule of CDP-DAG, yielding CL and CMP58, 63. Pathway 2 proceeds again through the removal of CMP from CDP-DAG and is replaced with a molecule of serine by a PS synthetase to yield PS63. Following a decarboxylation event,

58 PS is converted into PE through he liberation of a molecule of CO2 . Pathway 3 is responsible for the biosynthesis of PI and subsequent mannosylated PIM forms. The biosynthesis of PI occurs in one step with the removal of CMP and the addition of myo- inositol by PI synthetase64. PI is mannosylated first at the 2 position (PIM1) and then at the 6 position (PIM2) by the mannoysltransferases PimA and PimB, respectively65. PIM2 is then subject to the selective acylation of palmitic acid at the 5 hydroxyl of the first mannose sugar by PatA, producing Ac1-PIM2 (This position is denoted as R3 in Figure

1-3C)66. Ac1-PIM2 can be subject to further acylation at the 3 hydroxyl of the inositol ring to generate Ac2-PIM2, the acyltransferase responsible has yet to be determined (This

66 location is denoted as R4 in Figure 1-3C) . Both Ac1-PIM2 and Ac2-PIM2 can be further mannosylated by PimC to form Acyl-PIM3 species67. Upon a fourth mannosylation step by an unknown mannosyltransferase, Ac1-PIM4 and Ac2-PIM4 are believed to be translocated across the CM to the periplasmic side67. Ac1-PIM4 can only further be mannosylated to PIM6, while Ac2-PIM4 serves as the lipid anchor for the larger glycolipids of LM and manLAM66, 67.

12

Figure 1-3: Glycerophospholipids of Mtb. A) Common glycerophospholipids of Mtb share a phosphatidic acid core. R1 and R2 = any fatty acid from panel B (R1 = sn1, R2 = sn2 positions). B) Three most common FAs of glycerophospholipids60. C) Structure of PIM2, 4, and 6. Panel C was adapted from Albesa-Jove et al. Nature Communications, 201566.

1.3 α/β-hydrolase fold: lipid esterases

The α/β-hydrolase superfamily is one of the largest protein superfamilies shown to catalyze a diverse set of hydrolytic reactions68, 69. The unique protein fold was identified in 1992 upon comparison of the structures of hydrolytic enzymes

(acetylcholinesterase, carboxypeptidase II, dienelactone hydrolase, haloalkane dehalogenase, and a lipase)70. The defining feature of this fold is its central β-sheet that is comprised of 8, mostly parallel β-strands that are connected with α-helices (Figure1-4)70.

Furthermore, it was found these enzymes all possess a catalytic triad consisting of a nucleophilic residue - histidine - and an acidic residue. Importantly, the nucleophilic residue resides on a tight turn connecting α-helix to β-strand, this structural motif is often referred to as the nucleophilic elbow (Figure 1-4)70. Serine, cysteine, or aspartate can be

13

the nucleophilic residue of the catalytic triad, while the acidic residue is either aspartate or glutamate; however, the base is always histidine69.

Figure 1-4: Structure of haloalkane dehalogenase used to identify the α/β-hydrolase superfamily. α-helices, β-strands, and connecting loops are uniquely colored. Catalytic residues are in yellow (Nucleophile: Asp124, Base: His298, Acid: Asp160)(PDB: 2HAD)71

Divergent evolution has resulted in the conservation of the catalytic triad within

α/β-hydrolase superfamily; however, enzymes of this superfamily have vastly different protein sequences and substrate binding sites70. As a result, α/β-hydrolases can facilitate a variety of hydrolytic reactions. A major portion of α/β-hydrolases are considered to be esterases68. Esterases catalyze the hydrolysis of esters through nucleophilic attack on the ester carbonyl by the catalytic nucleophile68. Nucleophilic attack results in a covalent intermediate with the enzyme, referred to as the acyl-enzyme intermediate68. The acyl- enzyme intermediate is subject to hydrolysis through nucleophilic attack by an activated water molecule68. Therefore, esterases transform ester-containing substrates into two distinct products: an alcohol and a carboxylate (Figure 1-5)68. However, esterases can catalyze acyl transfer, if an acceptor molecule containing a free hydroxyl is subject to

14

nucleophilic activation over water. In this case, the products of this reaction would be the newly acylated acceptor molecule and the deacylated donor molecule (Figure 1-5)72.

Figure 1-5: Cartoon representation of a general esterase mechanism. Substrate 2 determines if acyl transfer or hydrolysis occurs.

Enzymes that have esterase activity are often divided into numerous categories based on performed chemistry and utilized substrates. For the sake of simplicity, esterases that perform chemistry on lipids will be referred to as lipid esterases, unless otherwise noted. Due to the lipid rich cell wall of Mtb, the Mtb genome encodes a high number of lipid esterases that are believed to be involved in lipid metabolism73. Two essential Mtb lipid esterases are the focus of this dissertation work. One lipid esterase is considered an acyltransferase and the other a Phospholipase A. As stated above, an acyltransferase catalyzes the transfer of an acyl group from a donor molecule to an acceptor molecule. Phospholipase A enzymes catalyze the hydrolysis of a fatty acid from the sn1 or sn2 position of glycerophospholipids.

15

1.4 Project Overviews

1.4.1 Antigen 85C

Mtb encodes three highly secreted, homologous mycolyltransferases referred to as the Ag85 complex (Ag85A, Ag85B, and Ag85C). Ag85s are α/β-hydrolases, use a classical catalytic triad (Ser-His-Glu), and catalyze the acyl transfer of MA from TMM to produce either TDM or mAG, two major components of the Mtb mycomembrane. Mtb is significantly impaired upon loss of Ag85C and is not viable in the absence of another homolog. Due to their essential nature and recognition by the human immune system,

Ag85s represent an appealing target for antibiotic and vaccine development.

Chapter two of this dissertation describes an updated model of substrate coordination based on the X-ray crystal structure of Ag85C in covalent complex with tetrahydrolipstatin (THL), a common, covalent lipase inhibitor that is also a moderate inhibitor of Mtb. Additionally, insights on how mycolyltransferases limit substrate hydrolysis through structural rearrangements are presented. Chapter three looks at the evaluation and characterization of synthesized derivatives of known covalent (ebselen) and non-covalent (thiophene) inhibitors of Ag85C. Finally, screening Ag85C against a variety of β and γ lactones in addition two TB active drug libraries resulted in the identification of novel lead molecules for further development.

1.4.2 Rv3802

Rv3802 is encoded by the rv3802c gene, which resides in a gene cluster responsible for MA biosynthesis. The enzyme is essential for Mtb viability, secreted to the periplasmic space, and possesses thioesterase and phospholipase A (PLA) activity.

The enzyme uses a classical catalytic triad (Ser-His-Asp) to perform chemistry and was 16

predicted to have an α/β-hydrolase fold. The exact biological role of Rv3802 remains uncertain; however, the enzyme has been shown to modulate the lipid content of the mycomembrane upon cell stress (decrease glycerophospholipid content, increase mycolated content).

Chapter four of this dissertation looks at the structural characterization of Mtb

Rv3802 and dynamics required for substrate binding. The first solved crystal structure of

Rv3802 is presented and characterized in addition to the identification of the lipid- binding site that is shown to readily accommodate glycerophospholipid substrates.

Protein dynamics required for lipid binding are investigated through mutagenesis and use of a PLA assay. Additionally, two fluorescence-based assays developed to monitor enzymatic activity in real time, both being amenable to high throughput drug screening are presented and compared. Finally, insights on biological activity are discussed based on structural and in vitro studies with mycomembrane substrates. Chapter five details efforts to inhibit Rv3802. Lead molecules were identified upon screening Rv3802 against two drug libraries. Compounds were tested against Msmeg and Mtb to evaluate biological activity.

1.4.3 Tetrahydrolipstatin Inhibition

THL is a potent inhibitor of human lipases and is FDA approved for the treatment of obesity. Inhibition by THL is covalent, yet often considered reversible due to hydrolysis of the ester linked drug-enzyme complex. THL is a promiscuous lipid esterase inhibitor and has been shown capable of covalently modifying over 200 lipid esterases in

Mtb; subsequently, THL is proficient at inhibiting Mtb growth. Both Ag85C and Rv3802 were identified as potential targets of THL within Mtb. Chapter 6 of this research looks at 17

the influences of THL stereochemistry on the inhibition of Rv3802 and Ag85C by THL.

The unique stereochemistry of THL was found to influence cross enzyme reactivity as well as the stability of the acyl-enzyme inhibited complex. Finally, the stability of covalent inhibition by THL is discussed as the Ag85C-THL and the only other protein-

THL structure (human FAS) are compared.

18

Chapter 2

Antigen 85 Catalysis

2.1 Background

2.1.1 Antigen 85s

Within the Mtb genome resides the fbpA, fbpB, and fbpC genes that encode for

Ag85A, B, and C, respectively74. These three enzymes are often referred to as the Ag85 complex; however, they are considered biological monomers and do not form a larger complex44. Each respective protein contains an N-terminal signal sequence that results in secretion from the cytoplasm42. Comprising over 15 % of the total protein content found in cell culture fluids, Ag85s are highly secreted proteins of Mtb with Ag85B being the most prominent42, 75. Ag85s are aptly named as they elicit an immune response in humans42. Early studies found Ag85s to exclusively bind human fibronectin; therefore these enzymes were initially referred to as the fibronectin binding proteins76.

In 1997, Belisle and colleagues discovered the biological role of Ag85s as mycolyltransferases, showing all three enzymes capable of catalyzing the acyl transfer of

MA from TMM to a second molecule of TMM to produce TDM44. Later studies confirmed Ag85s also catalyze acyltransfer to fragments of the AG, showing Ag85s responsible for mAG biosynthesis56, 77. These in vitro characterization are supported as 19

Mtb strains with disrupted Ag85 genes resulted in mycomembranes deficient in MA content and disrupted cell walls74, 78. Specifically, Msmeg strains deficient in the corresponding Ag85A ortholog had lower levels of TDM present, while Mtb strains deficient in Ag85C had lower levels of mycolated cell wall content (mAG)78, 79. This would suggest that Ag85A is responsible for TDM biosynthesis while Ag85C is responsible for mAG biosynthesis. Therefore, despite having sequence similarities greater than 75 % and being able to perform the same type of chemistry, it is believed that the three Ag85 homologs have unique roles in mycobacteria77.

2.1.2 Ag85 Structure and Substrate Binding sites

The X-ray crystal structure of Mtb Ag85C was the first of the Ag85s to be solved80. Ronning and colleagues discovered Ag85C to have an α/β-hydrolase fold with 9

α-helices and a central β-sheet comprised of 8 β–strands. Shortly thereafter, the structures of Ag85A and Ag85B were also solved81, 82. All three Ag85s share a similar α/β- hydrolase fold with nearly identical structures and identical positioning of catalytic residues (Figure 2-1)80-82. Between the three homologs all secondary structures are conserved with differences observed in unstructured loops and the N-terminus.

20

Figure 2-1: X-ray crystal structures of Ag85s reveal the enzymes to have near identical structures and active sites (Ag85A PDB: 1SFR, Ag85B PDB: 1F0N, and Ag85C PDB: 1DQZ)80-82.

As previously stated, Ag85s catalyze the transfer of MA from TMM to either the

5-hydroxyl of a terminal arabinose in the AG or to the 6’-hydroxyl of another molecule of TMM44, 56, 77. Based on the substrates involved, Ag85s should therefore have a sugar binding site and a site to accommodate the alkyl chains of MA. Ag85s were crystallographically determined to have two sugar binding sites based on the crystal structure of Ag85B in complex with trehalose molecules81. The first trehalose binding site is adjacent to the active site, with the 6’-hydroxyl of trehalose oriented towards the nucleophilic Ser124 (Figure 2-2A)81. The secondary trehalose binding site is distal to the active site and resides at the base of α-helix 5 and the middle of α-helix 9 (Figure 2-

2A)81. These sugar binding sites were also crystallographically confirmed in Ag85C through complexation with octylthioglucoside molecules (Figure 2-2B)82. The sugar of the octylthioglucoside molecule in the active site resides in an identical position to that of trehalose; additionally, the non-hydrolysable thiol linkage sits above the nucleophilic

Ser124 with the octyl chain occupying a hydrophobic cleft opposite that of the sugar 21

binding site (Figure 2-2B)82. The Ag85C-octylthioglucoside structure therefore highlights a potential mode of TMM binding to Ag85C, splitting the active site into two distinct regions, the sugar binding site and the MA binding site. The bifurcation of the active site into two distinct regions is further supported by the Ag85C-diethylphosphate structure, which mimics the tetrahedral intermediate of the enzymatic reaction80. One ethyl arm of diethylphosphate is pointed towards the sugar binding site while the other ethyl arm is positioned in the hydrophobic cleft (Figure 2-2C)80. Aside from substrate binding, the

Ag85C-diethylphosphate structure confirmed the backbone amides of Met124 and Leu40 comprise the oxyanion hole required for stabilization of the tetrahedral intermediate formed during nucleophilic attack on substrate80. Aside from mapping key sites of the active site, a noticeable structural change is observed upon modification of Ag85C with diethylphosphate. The α9-helix is found relaxed relative to the native form; as a result, the connecting loop that is elongated over the hydrophobic cleft of the active site (Figure

2-2D). Furthermore, His260 is displaced away from Ser124. The implications of these structural changes with respect to catalysis will be examined in this chapter.

22

Figure 2-2: Substrate binding sites of Ag85s: A) Ag85B in complexation with trehalose reveals two sugar binding sites (PDB: 1F0P)81. B) Ag85C in complexation with octylthioglucoside confirms sugar binding sites and resembles a potential binding mode for TMM (PDB: 1VA5)82. C) The nucleophilic Ser124 of Ag85C covalently modified by diethylphosphate, resembling the tetrahedral intermediate (PDB: 1DQY)80. D) Structural changes upon modification of Ser124 with diethylphosphate (Green: Ag85C- diethylphosphate, PDB: 1DQY; White: Apo Ag85C, PDB: 1DQZ)80.

2.1.3 Mechanism of catalysis

Ag85s catalyze the transfer of an acyl chain from a donor molecule to an acceptor molecule through a Ping-Pong reaction mechanism using acid base catalysis80. In general, the reaction mechanism can be broken down into two half reactions. The simplified mechanism of TDM production from two molecules of TMM by Ag85C is illustrated in

Figure 2-3. The reaction begins upon binding of the acyl donor molecule, in this case,

23

80 TMM1 . Ser124 is deprotonated by His260, which acts as a general base and is in turn stabilized by the deprotonated Glu228 of the catalytic triad80. The “activated” Ser124 is now free to undergo nucleophilic attack on the carbonyl carbon of the ester linked MA of

TMM80. Nucleophilic attack results in a tetrahedral intermediate between enzyme and

TMM1, with the resulting oxyanion stabilized by the amide backbones of Met124 and

Leu4080. Upon collapse of the tetrahedral intermediate, a free molecule of trehalose is liberated with the enzyme acylated through an ester linkage, this state is referred to as the acyl-enzyme intermediate80. The second half reaction therefore proceeds upon binding of

80 a second molecule of TMM . Binding of TMM2 positions the 6’-hydroxyl to be positioned for deprotonation by His26080. The now deprotonated or “activated” 6’- hydroxyl of the acceptor molecule undergoes nucleophilic attack on the carbonyl carbon of the acyl-enzyme intermediate ester80. Again, a tetrahedral intermediate is formed and stabilized by the oxyanion hole80. Collapse of the tetrahedral intermediate results in the release of product, TDM, and turnover of the enzyme back to the original catalytic state80.

Figure 2-3: Mechanism of TDM biosynthesis by Ag85C. The reaction requires two molecules of TMM and produces a molecule of trehalose and TDM, R1 and R2 are MAs. Figure adapted from Goins et al. the Journal of Biological Chemistry, 201857, 80.

24

2.1.4 Substrate coordination (Portions of text reproduced from Goins et al. the

Journal of Biological Chemistry, 2018)57

As previously stated, Ag85s have two sugar binding sites: the secondary site, which is located near the middle of the α9-helix and the base of the α5-helix, and the active site, with the 6-hydroxy pointing toward the nucleophilic serine residue (Figure 2-

4A)81. On the basis of these observed binding sites, Anderson and coworkers proposed an interfacial mechanism model in which TMM initially binds at the secondary site outside the active site, stimulating a conformational change of the side chain of F232 in Ag85A and B (Leu230 in Ag85C) that allows TMM to then enter the active site and undergo nucleophilic attack as the initiating step of the first half-reaction (substrate flow depicted with arrows in Figure 2-4A)81. Following this step, the liberated trehalose molecule transiently resides in the active site until it is released as the product of the first half- reaction81. This scheme suggests that, in the acyl-enzyme intermediate form, the α-chain of MA is buried in a hydrophobic hole with the meromycolate chain flipped out away from the enzyme and residing within the mycomembrane (Figure 2-4A)81. The second half-reaction would then proceed upon binding of a second molecule of TMM to the secondary site followed by translocation to the active site81. TDM is thereby produced following nucleophilic attack on the MA-enzyme intermediate by the second molecule of

TMM in the active site81.

Major problems exist with the proposed interfacial mechanism. First, the model does not account for the production of mAG. Second, the hydrophobic cleft used to translocate the second molecule of TMM2 to the active site would be too sterically hindered to allow for the second molecule of TMM to enter the active site; this is due to 25

the acyl chains of the Ag85-MA acyl-enzyme intermediate occupying this region and the enzyme undergoing structural changes upon acylation that further restrict this cleft

(Figure 2-4C). Furthermore, if an acceptor molecule did approach from this orientation, the attacking hydroxyl could not be properly oriented to satisfy the Burgi-Dunitz angle required for nucleophilic attack on a carbonyl carbon83. Due to these reasons, an updated model of substrate coordination with respect to enzyme catalysis is required.

Figure 2-4: Proposed substrate coordination during Ag85 catalysis. A) The interfacial mechanism proposed by Anderson and colleagues relies on the translocation of TMM from the secondary binding site to the active site (PDB: 1F0P)81. B) TMM is sizable, hydrophobic molecule45. C) The Ag85C-diethylphosphate structure highlights how sterically hindered the proposed translocation site becomes upon enzyme acylation (PDB: 1DQY)80.

This chapter addresses concerns with substrate coordination based on the crystal structure of Ag85C in covalent complex with tetrahydrolipstatin (THL). THL is a ubiquitous inhibitor of Mtb lipid esterases and human lipases73, 84. The resulting structure is the second protein-THL structure to be ever solved, providing insights on general THL inhibition. Fortuitously, THL allowed for the trapping of an acyl-enzyme complex, which resembles the core attributes of a mycolated Ag85C. On this structural basis, we were able to provide an updated model of substrate coordination to satisfy issues raised with 26

the interfacial mechanism. Additionally, this structure provided insights into how the enzyme undergoes structural changes to facilitate substrate transfer over hydrolysis.

2.2 Methods (Portions of text reproduced from Goins et al. the Journal of Biological

Chemistry, 2018)57

2.2.1 Molecular cloning of Ag85C and mutants

The Mtb wild type (WT) fbpC gene was previously cloned into a pET-29 plasmid by Dr. Lorenza Favrot85. The recombinant Ag85C expression construct included a non- cleavable C-terminal poly-Histidine tag with the N-terminal signal sequence removed.

S148A and S148T mutants were generated using the Agilent QuickChange® Lightning kit. The plasmid harboring the WT fbpC gene was used as the template for both mutagenesis reactions. Primers used for mutagenesis are given in table 2.1. PCR components and parameters were the same as recommended by the Agilent

QuikChange® Lightning kit, an annealing temperature of 68.0 °C was used for both mutants. Resulting plasmids were sent for DNA sequencing to confirm mutagenesis.

Table 2.1: DNA primers used for mutagenesis. Primer Name DNA Sequence S148A Forward 5’-ggttgaggaagcccgccaacgacgcggcgta-3’ S148A Compliment 5’-tacgccgcgtcgttggcgggcttcctcaacc-3’ S148T Forward 5’-ggttgaggaagcccgtcaacgacgcggcgta-3’ S148T Compliment 5’-tacgccgcgtcgttgacgggcttcctcaacc-3’

2.2.2 Expression and purification of Ag85C and mutants

Recombinant Mtb WT Ag85C was expressed and purified as previously published with S148A and S148T following identical experimental procedures85. In short, 27

chemically competent T7 Express Escherichia coli cells were transformed with the desired pET-29 C-terminal poly-Histidine construct. Inoculated cultures were grown at

37 °C in Luria-Bertani broth to a density of 0.6 OD600nm. Incubation temperature of the cultures were then dropped to 16 °C and induced with 1 mM isopropyl β-D-1- thiogalactopyranoside. After 24-36 hours, induced cultures were harvested and resuspended in 20 mM TRIS pH 8.0 buffer containing 5 mM β-mercaptoethanol and placed at – 80 °C for storage.

Induced cells were thawed, lysozyme and DNase I added, and incubated on ice, cell lysis was complete following sonification. The crude cellular lysate was clarified by centrifugation and loaded onto a metal (cobalt) affinity chromatography column equilibrated with lysis buffer. Following washing with 15 column volumes of lysis buffer, protein was eluted with a 15 column volume gradient of imidazole, 0 to 150 mM.

Eluted protein was pooled and loaded on to a 5 mL anion exchange column equilibrated with washing buffer (20 mM TRIS pH 8.0 buffer containing 1 mM EDTA and 0.3 mM

TCEP). Following washing with 10 column volumes, protein was eluted with a 0 to 1 M

NaCl gradient of 15 column volumes. Eluted protein was pooled and subjected to ammonium sulfate precipitation (2.8 M). Precipitated protein was pelleted via centrifugation and resuspended in 10 mM TRIS pH 7.5, 2 mM EDTA for crystallization or 50 mM sodium phosphate pH 7.5 for enzymatic assays. Resuspended protein was dialyzed overnight against respective buffer to remove residual ammonium sulfate.

2.2.3 Ag85C Activity assay

A previously described fluorescence based assay was used to monitor Ag85C activity85. In brief, the assay monitors the acyltransfer of butyrate from resorufin butyrate 28

(RfB) to trehalose, producing a fluorescent molecule of resorufin and 6-butyl-trehalose

(Figure 2-5). In general, reactions consisted of 500 nM enzyme, 4 mM trehalose, and 100

μM RfB. Reactions were ran in 50 mM at 37 °C in a 50 mM sodium phosphate buffer pH

7.5, using λex = 500 nm and λemit = 590 nm. Fluorescent reads were obtained using

Synergy H4 plate reader. All reactions were repeated in triplicate, with a negative control

(no enzyme) used to subtract background hydrolysis of RfB.

Figure 2-5: Fluorescence based assay used to monitor Ag85C activity.

2.2.3.1 Mutant activity

Enzymatic activity for mutants was determined with an identical assay as described above; conditions for transesterase activity are as follows: 500 nM respective enzyme, 4 mM trehalose (500 mM buffer stock), 100 μM Resorufin butyrate (10 mM DMSO stock). Kinetic reads were initiated immediately following the titration of RfB. For hydrolase activity, reactions were identical sans trehalose. Data were converted to product concentration using a resorufin standard curve, background water hydrolysis of resorufin butyrate was subtracted from all triplicate data, and rates determined using PRISM 7 with a linear fit. Reported transesterase activity has the enzymatic rate of RfB hydrolysis subtracted.

2.2.3.2 kinact/Ki determination for tetrahydrolipstatin

THL was serial diluted from a 30 mM stock in DMSO resulting in a range of final reaction concentrations of 300 to 12.5 μM. Kinetic reads were initiated following the titration of RfB immediately after the titration of THL or an equal volume of DMSO.

Kinetic reads were conducted at 37 °C in a 50 mM sodium phosphate buffer pH 7.5.

29

Relative fluorescent units were converted to product concentration using a resorufin standard curve. Background water hydrolysis of RfB was subtracted from the triplicate data and the rates determined using PRISM 7 with a one-phase association equation, Y =

Y0 + (Plateau – Y0)(1-exp(-kx)); where x = time, Y = [Product], (Y0 + (Plateau – Y0) =

Vi/kobs, and k = kobs. kinact/KI was determined by plotting the kobs versus inhibitor

86 concentration and fitting the data with the equation kobs = ( kinact/(1+(Ki/[I]))) .

2.2.4 Crystallization and structure determination

Recombinant WT Ag85C was concentrated to 150 μM (~ 5 mg/mL). Ag85C-THL crystals were obtained through co-crystallization via incubation of THL (20 mM stock in

DMSO) at a 1:1.2 molar ratio of protein to compound for 90 minutes on ice prior to drop set up. Using the hanging drop method, crystals formed after a week at 16 °C in a 1:1 protein to well solution (0.1 M Calcium chloride dehydrate, 0.05 M BIS-TRIS pH 6.5,

22.5 % v/v (+/-)-2-Methyl-2,4-pentanediol (MPD), Hampton Research Index HR2-144).

Crystals were cryoprotected through the addition of 0.2 μL glycerol (10 % v/v final drop concentration) immediately prior to looping and flash-cooling in liquid nitrogen. X-ray diffraction data were collected at 100 K using synchrotron radiation (λ = 0.97872 Å) on beam line F of LS-CAT, Advanced Photon Source at Argonne National Lab.

Diffraction data were indexed, integrated, and scaled using HKL200087. The resulting diffraction data set was indexed and scaled as P21212. The phase solution came from molecular replacement using the native Ag85C structure (PDB: 1DQZ) with one molecule determined to be in the asymmetric unit. Residues not fitting 2Fo-Fc electron density were deleted and the resulting model subjected to a rigid body refinement, followed by simulated annealing (PhenixRefine)88. Deleted residues and the THL 30

modified Ser124, were built manually using COOT89. Restraints for the THL modified

Ser124 were generated with eLBOW90. Glycerol and MPD molecules were added using

Ligand Fit91. The progressing model was subjected to rounds of XYZ coordinate, real- space, occupancy, and B-factor refinements in between manual builds and ligand additions (PhenixRefine)88. Model building and refinement ended when an Rwork and

Rfree values of 0.163/0.167 were reached with 97 % residues being Ramachandran favored and 1 % outliers. The two outlier residues, Gly29 and Ser86, can be found within loop regions and are positioned based on the difference density maps giving correlation coefficients of 0.97 and 0.86 for those two residues, respectively. Model statistics were validated with MolProbity92. These residues have been flagged as outliers in prior Ag85C structures93.

2.2.5 Structural and sequence alignments

The sequences of 464 known mycolyltransferases were aligned using Clustal

Omega94. The resulting alignment was used to generate the frequency of a given amino acid at a determined position, indicating sequence conservation. Figures were generated using WebLogo 395. Structural alignments were generated using PyMOL96.

2.3 Results (Portions of text reproduced from Goins et al. the Journal of Biological

Chemistry, 2018)57

2.3.1 Mutant and wild type cloning, expression, and purification

Cloning of the Ag85C S148A and S148T mutants was straightforward and required only one attempt using the Agilent QuikChange® Lightning kit. The Ag85

S148A gene was completely sequenced; however, sequencing reads were of low quality 31

for the C-terminus of the S148T plasmid. Nonetheless, the desired mutants were confirmed via DNA sequencing. ClustalW sequence alignments between sequenced plasmids and known Mtb Ag85C sequence are given in Figure 2-6. The N-terminal signal sequence that was previously removed is also indicated by (---) marks with regard to the full WT sequence.

Figure 2-6: Sequence alignment from mutant plasmids to known Mtb Ag85C sequence. Respective mutants were confirmed via DNA sequencing.

Levels of recombinant protein expression of Ag85C mutants were equal to that of

WT (~25 mg of protein per 4L culture). Ag85C mutants appeared to have similar levels of stability compared to WT based on a lack of precipitation when being concentrated.

Additionally, both mutants were purified in an identical fashion to that of WT (SDS-

PAGE of purified WT and mutant Ag85s given in Figure2-7). Levels of protein purity appear to be identical to that of WT.

32

Figure 2-7: SDS-PAGE of purified WT, S148A, and S148T mutants, anticipated mass of ~33 kDa. Lane 1: ladder, 2: WT Ag85C, 3: ladder, 4: S148A Ag85C, 5: S148T Ag85C.

2.3.2 Tetrahydrolipstatin inhibition

THL is a well-known lipid esterase inhibitor that suppresses Mtb growth73.

Covalent inhibition results from nucleophilic attack by the serine nucleophile on the carbonyl center in the β-lactone ring of THL (Figure 2-8A)97. Covalent inhibition is both

86 time and concentration dependent . Therefore, kinact/KI was determined for THL

86 inhibition of Ag85C, as kinact/KI takes into account concentration and time dependence .

-3 -1 The kinact/KI was measured to be 7.9 ± 1.0 x 10 μM-1 min by monitoring the rate of transesterification by Ag85C in the presence of varying concentrations of THL (Figure 2-

8B). The KI of THL is therefore 8.8 ± 0.7 μM. The raw inhibition progress curves used to determine kobs are provided in Figure 2-9. The observed progress curves are indicative of covalent inhibition as they plateau as a function of time, indicating a decrease in total enzyme active sites as a function of covalent inhibition86.

33

Figure 2-8: Covalent inhibition of Ag85C by THL. A) Chemical structure of THL and resulting structure upon covalent attack by Ag85C. B) kinact/KI plot of THL inhibition. Figure reproduced from Goins et al. the Journal of Biological Chemistry, 201857.

Figure 2-9: Progress curves of Ag85C with varying amount of THL added. The curves displayed are the average of triplicate reactions and are fitted with the kobs equation. Figure reproduced from Goins et al. the Journal of Biological Chemistry, 201857.

2.3.3 Ag85C tetrahydrolipstatin structure

Mtb Ag85C was co-crystallized with THL, resulting in a structure of 1.45 Å resolution (crystallographic and refinement statistics in Table 2.2). Continuous difference density was present for all atoms of THL. The Fo-Fc omit map for the THL-modified

Ser124 and 2Fo-Fc map for surrounding residues are shown in Figure 2-10A. Covalent modification of Ser124 by THL results in an ester linkage between the drug and enzyme and yields the β-hydroxyl of THL as a result of ring opening. The carbonyl of that ester

34

linkage points directly toward the identified oxyanion hole of the backbone amides of

Met125 and Leu40 (Figure 2-10B)80. The alkyl chains of THL lie within a hydrophobic cleft extending back toward the secondary trehalose binding site, which resides below the terminal methyl carbon of the palmitic core chain. The hexanoyl tail approaches the hydrophobic hole; however, it does not extend down into this channel, which is occupied by seven water molecules (Figure 2-11A). The peptidyl side arm of THL is extended toward the α9-helix, displacing the catalytic His260. Additionally, the α9-helix adopts a relaxed conformation relative to the kinked apo structure, displacing the helix away from the protein core (Figure 2-11B). Unfortunately, interpretable difference density for residues 216 to 221, which account for a dynamic loop connecting the α9-helix to the preceding β7-strand, was not present and therefore was not modeled. Two glycerol molecules are present in the trehalose active site location, with one hydroxyl pointing toward the acyl enzyme intermediate and the β-hydroxyl of THL (Fo-Fc omit map figure

3A). Aside from these noted observations, the overall protein fold is identical to that of apo Ag85C with the two structural models exhibiting an RMSD of = 0.23 Å (Figure 2-

11B).

35

Table 2.2: X-ray data collection and refinement statistics (molecular replacement). One crystal was used for this structure. Values in parentheses, unless otherwise indicated, represent data in the highest-resolution shell. Table reproduced from Goins et al. the Journal of Biological Chemistry, 201857. Ag85C-THL (PDB: 5VNS) Data collection Space group P 21 21 2

a, b, c (Å) 68.38, 122.78, 40.26 α, β, γ () 90.00, 90.00, 90.00 Resolution (Å) 35.12 – 1.45 (1.5-1.45) Total reflections 773689 (61070) (Unique)

Rmerge 0.06 (0.53)

Rmeas 0.07 (0.57) CC1/2 0.97 (0.91) I /σI 18.4 (3.8) Completeness (%) 99.91 (99.18) Redundancy 7.3 (7.3) Refinement Resolution (Å) 35.12 – 1.45

Rwork / Rfree 0.161/0.173 No. atoms Protein 2190 Ligand/ion 28 Water 281 B-factors Protein 15.0 Ligand/ion 28.9 Solvent 28.70 R.m.s. deviations Bond lengths (Å) 0.007 Bond angles () 1.209 Ramachandran Favored 97 % Outliers 1 %

36

Figure 2-10: Ag85C-THL structure. A) Fo-Fc likelihood-weighted omit map (blue) contoured to 3.0 σ for S124 modified by THL and glycerol molecules in the active site, and 2Fo-Fc map (red) contoured to 1.5 σ for surrounding active site residues. B) Surface rendering of Ag85C with S124 modified with THL and two glycerol molecules occupying the trehalose binding portion of the active site. Figure reproduced from Goins et al. the Journal of Biological Chemistry, 201857.

Figure 2-11: Hydrophobic pocket and global changes upon THL modification. A) Hydrophobic pocket occupied by 7 water molecules. B) Alignment of THL modified Ag85C with apo enzyme. Regions of difference are highlighted in green and wheat, the dark gray and wheat molecule is Ag85C-THL, the light gray and green molecule is Ag85C; PDB: 1DQZ80. Figure reproduced from Goins et al. the Journal of Biological Chemistry, 201857.

2.3.4 Structural comparisons and conservation of residues of interest

In the catalytically active enzyme, the nucleophilic hydroxyl of Ser124 is hydrogen bonded to the ε-nitrogen of the His260 imidazole ring, and the δ-nitrogen of

His260 is hydrogen bonded to Glu228 (Figure 2-12A)80. In this form, the α9-helix is

37

kinked toward the active site with a connecting dynamic loop positioned away from the active site, allowing for substrate binding (Figure 2-12A). In structures exhibiting a disrupted catalytic triad, His260 rotates about both χ1 and χ2 to form an interaction with the side chain of Ser148 (Figure 2-12A)80, 85, 93, 98. Disruption of hydrogen bonds with residues of the α9-helix therefore relaxes the helix, allowing the dynamic loop to be positioned over the active site, resulting in the side chain of Leu217 being oriented between His260 and the nucleophilic Ser124 (Figure 2-12A and B)93. Residue conservation at position 217 is split between leucine and isoleucine in all mycolyltransferases (Figure 2-13). Positioning of the aliphatic side arm of THL is similar to that of Leu217 (Figure 2-12B). In both the disrupted catalytic and THL-modified forms, the aliphatic substituents help orient the His260 side chain such that it maintains a hydrogen bond with Ser148 and abrogates potential solvent interactions with His260, decreasing the potential for water activation (Figure 2-12B).

This repositioning of His260 by disruption of the catalytic triad results in the formation of a hydrogen bond between the δ-nitrogen of the imidazole ring and the neighboring hydroxyl of Ser148. Ser148 and the neighboring Trp265, which forms a hydrogen bond with Ser148 through its indole nitrogen, are both highly conserved among all known mycolyltransferases (Figure 2-13). In every Ag85C structure with disrupted catalytic triads, a similar displacement of H260 is present that mimics this acylated form

(Figure 2-12B)80, 85, 93, 98.

38

Figure 2-12: Observed structural changes between the active and disrupted catalytic forms. A) Catalytic triad disruption results in relaxation of the α9-helix with His260 adopting a sequestered position (white: Ag85C-S124A; PDB: 4QEK, cyan: trehalose- bound Ag85B; PDB: 1F0P)81, 93. B) Consistent displacement of His260 results in hydrogen bond formation with Ser148. The aliphatic side arm of THL mimics the hydrophobic interactions of the side chain of Leu217 in the diethyl phosphate-modified Ag85C and S124A mutant structures (Wheat: Ag85C-THL, PDB: 5VNS; green: Ag85C- diethyl phosphate, PDB: 1DQY; white: Ag85C-S124A, PDB: 4QEK)80, 93. Figure reproduced from Goins et al. the Journal of Biological Chemistry, 201857.

39

Figure 2-13: Web logo showing sequence conservation for (A) Leu217, (B) Ser148, and (C) Trp265 across 464 known mycolyltransferases including Ag85A, Ag85B, and Ag85C sequences. Figure reproduced from Goins et al. the Journal of Biological Chemistry, 201857.

2.3.5 Mutant activity

Based on the structural alignments and high sequence conservation, we sought to determine if Ser148 plays a direct role in sequestering the catalytic His260 to limit hydrolysis of the acyl-enzyme intermediate. To investigate potential effects of Ser148 on enzymatic activity, we mutated Ser148 to alanine and threonine. Both variants lacked observable transesterase activity, but exhibited hydrolase activity at a lower level than that of WT (43.46 ± 3.1 and 25.06 ± 1.4 % of observed WT hydrolysis, S148A and

S148T, respectively)(Figure 2-14).

40

Figure 2-14: Enzymatic activity of WT, S148A, and S148T mutants plotted as a function of reaction rates, average and SEM given. Reaction rates are a function of resorufin production per minute (μM/min). Transesterase activity is lacking in both mutants. Figure reproduced from Goins et al. the Journal of Biological Chemistry, 201857.

2.4 Discussion (Portions of text reproduced from Goins et al. the Journal of

Biological Chemistry, 2018)57

2.4.1 Ag85C-THL structure and inhibition

THL, known commercially as Orlistat, is a synthetically stable derivative of lipstatin, a natural product synthesized by Streptomyces toxytricini that inhibits the human pancreatic lipase99, 100. THL was shown to inhibit the thioesterase domain of the human FAS101. Functioning as a versatile lipid esterase inhibitor, THL exhibits growth inhibition on Mtb via covalent modification of various endogenous lipid esterases,

Ag85C being reported as one of fourteen validated targets73. Here we have shown that covalent inhibition occurs within minutes and initial binding affinities are in the low μM.

These values fall within the range of previously reported ebselen derived covalent

41

inhibitors of Ag85C that are known to efficiently inhibit Mtb growth (Ebselen derivatives discussed in Chapter 3)98, 102.

A significant difference in THL inhibition of Ag85C compared to human FAS is the lack of observed hydrolysis of the covalent enzyme adduct. The structure of FAS in complex with THL contains two molecules in the asymmetric unit: one with an intact ester linkage and the other in a hydrolyzed form84. A later study found that movement of the hexanoyl tail resulted in the repositioning of the β-hydroxyl of THL103. Repositioning of the β-hydroxyl disrupts hydrogen bonding to the stationary catalytic histidine, allowing for water activation and subsequent hydrolysis of the acyl-enzyme intermediate103. However, in the Ag85C-THL structure we find that covalent modification results in structural changes, specifically the catalytic histidine, His260, is physically displaced by the peptidyl side arm and is instead within hydrogen bonding distance to neighboring Ser148 (2-12B). The Ag85C-THl complex in addition to the sequestered positioning of His260 to Ser148 was shown to be stable via MD simulations

(Data not shown, MD simulation conducted by Steven Dajnowicz and Jerry M. Parks), thereby limiting water activation and subsequent hydrolysis of Ag85C-THL. This stable interaction is experimentally supported as Ag85C was successfully crystallized with THL in only a slight molar excess of THL to enzyme, 1.2:1, respectively. Differences between inhibition of Ag85C and human FAS by THL are discussed in further detail in chapter 6.

2.4.2 Promoting substrate transfer over hydrolysis

A similar displacement of His260 toward Ser148 is observed in both the tetrahedral intermediate (represented by the Ag85C-diethyl phosphate complex) and the nucleophilic S124A mutant structures (Figure 2-12A)80, 93. Although both of these 42

structures lack a bulky substituent to force the dislocation of His260, the side chain of

His260 occupies a sequestered conformation due to structural rearrangement, as a result of hydrogen bond disruption to the nucleophilic Ser124. Additionally, the side chain of

Leu217 falls between His260 and the nucleophilic Ser124 has a function of loop restructuring. In a sequestered conformation, His260 would not be able to activate a water molecule for nucleophilic attack on the acyl-enzyme ester linkage. Based on consistent structural observations we therefore hypothesized that upon acylation, Ag85s adopt conformational changes that limit hydrolysis of the acyl-enzyme intermediate through sequestering of His260.

Using the Ag85C-THL structure as a molecular basis, we were able to model a truncated form of the mycolated Ag85 acyl-enzyme intermediate and perform MD simulations to further validate our structure-based hypothesis. All MD simulations were conducted by Steven Dajnowicz and Jerry M. Parks, models were generated by myself and Steven, methods for model generation and MD simulations can be found in Goins et al. the Journal of Biological Chemistry, 201857. We modeled the mycolated Ag85C as two forms: one with His260 in the sequestered form with α9-helix in a relaxed position

(Ag85C-MA-His260seq), as observed in the Ag85-THL structure, and the second form being an active form with His260 in the catalytic position and α9-helix kinked towards the active site (Ag85C-MA-His260cat). Models of the two acyl-enzyme intermediate forms are depicted in Figure 2-15A and B.

43

Figure 2-15: MD models of the mycolated Ag85C intermediate. A) Ag85C-MA- His260seq. B) Ag85C-MA-His260cat. C) Hydrogen bonding of His260 with Ser148 in the sequestered position is stable during MD simulations. When His260 is placed in the catalytic position without acceptor molecule present the side chain freely samples conformations making limited hydrogen bonding interactions with the β-hydroxyl of MA. Figure reproduced from Goins et al. the Journal of Biological Chemistry, 201857.

The acyl-enzyme model, Ag85C-MA-His260seq, with His260-hydrogen bonded to

Ser148 was found to be highly stable (Figure 2-15C). During the course of a 100 ns MD simulation, this interaction was maintained almost entirely, suggesting this “sequestered” form of the enzyme to be a stable energetic minimum. When His260 and the α9-helix were placed in a catalytic state, the hydrogen bond interaction between His260 and the proximal β-hydroxyl of MA was rarely sampled (Figure 2-15C). Over the course of the simulation the α9-helix began to relax in conformation and His260 began sampling towards the sequestered confirmation; however, His260 did not form a hydrogen bond with Ser148. This would suggest that indeed upon enzyme acylation, the enzyme prefers a structural state where His260 is displaced away from the acyl-enzyme ester. To remove starting model bias to see if the model would fully transition from the unstable catalytic

44

form to the stable sequestered form, a replica exchange molecular dynamic (REMD) simulation was conducted on the Ag85C-MA-His260cat.

REMD is an enhanced sampling method that allows the system to explore phase space more extensively and efficiently than conventional MD simulations104-106. Steven

Dajnowicz, Micholas D. Smith, and Jerry M. Parks performed the REMD MD simulation

(Methods can be found in Goins et al. the Journal of Biological Chemistry, 201857. The distance between His260 and the β-hydroxyl of MA, and the distance between the nucleophilic Ser124 and the α9-helix are plotted as a 2D heat map (Figure 2-16). When both of these distances are small, the enzyme is in the catalytic state. Alternatively, when the distance between His260 and β-hydroxyl of MA is large (> 5 Å), two states are observed. In state 1, the α9 helix was kinked toward the active site, similar to what was observed in the catalytic state, but the distance between His260 and β-hydroxyl of MA was between 5 to 7 Å. In the REMD simulations the enzyme transitioned toward the relaxed form of the enzyme (state 2). In state 2, the α9 helix was further away from the active site, while His260 remained dynamic and sampled many conformations.

Importantly, in many states His260 moved closer to Ser148 but never formed a hydrogen bond with the hydroxyl of Ser148. The conformations of His260 observed in the initial conventional MD simulation of Ag85C-MA-H260cat are consistent with the corresponding REMD simulation.

45

Figure 2-16: Heat map of Ag85C-MA-His260cat conformations sampled during the REMD simulation. The starting catalytic position was not frequently sampled when the α9-helix remained kinked towards the active site in the catalytic position; instead His260 randomly sampled phase space, similar to what was observed in the initial MD simulation (state 1). In state 2, the α9-helix relaxes similar to what is observed in the sequestered form. While His260 is positioned towards Ser148, a stable hydrogen bond to Ser148 is never observed to form. Figure reproduced from Goins et al. the Journal of Biological Chemistry, 201857.

Our hypothesis that Ag85s undergo structural changes to limit substrate hydrolysis upon acylation is further supported by the REMD and MD simulations.

However, a quick transition from the catalytic to sequestered form of the enzyme was not observed, negating a simplified two state system that is observed crystallographically.

Alternatively, the simulations suggest that a number of conformations are plausible along the pathway from catalytic to hypothesized sequestered form that is observed crystallographically. In most cases the structural rearrangement upon acylation is indeed sufficient to limit the activation of water by His260, thereby limiting hydrolysis of the acyl-enzyme intermediate. However, the enzyme may not always adopt a partially sequestered or fully sequestered form, resulting in water activation and the hydrolysis of

46

the acyl-enzyme intermediate. This hypothesis is supported as low levels of hydrolytic activity are observed with WT Ag85C (Figure 2-14). The low levels of hydrolytic activity remained when the highly conserved Ser148 was mutated to alanine or threonine to investigate a potential role in sequestering His260. Similar to the wild-type enzyme, this residual hydrolytic activity to can be attributed to His260 sampling active conformations.

Interestingly, the Ser148 mutants did lose all detectable acyltransferase activity, indicating that Ser148 plays a role in the molecular positioning of the active site in

Ag85C for acyltransfer to occur.

2.4.3 Substrate coordination and acceptor molecule activation

The previous understanding of the mycolated form of Ag85s was that the α-chain of MA was buried in the hydrophobic hole while the meromycolate chain was flipped outward into the mycomembrane81. As stated earlier, such an arrangement would prohibit the translocation of acceptor molecule from the secondary sugar-binding site to the active site, as suggested by the interfacial mechanism. Based on the Ag85C-THL structure, we propose that in the acylated state, both the α-alkyl chain of the MA and the portion of the meromycolate chain proximal to the ester linkage most likely lie within the active site hydrophobic cleft. The region of the meromycolate chain distal to the ester linkage may remain embedded in the mycomembrane. This positioning properly orients the carbonyl of the acyl-enzyme intermediate toward the oxyanion hole. Furthermore, it allows for nucleophilic attack from the respective 6’ or 5 hydroxyl of either trehalose of TMM or arabinose of the AG upon binding to the identified sugar-binding site of the active site.

This mode of acceptor binding would therefore circumvent the issues of steric restraint 47

raised with the interfacial mechanism model and allow for the production of both TDM and mAG.

To investigate this potential mode of acceptor molecule binding to the acyl- enzyme intermediate, we simply modeled trehalose into the Ag85C-MA-His260cat model to yield the Ag85C-MA-Trehalose model for MD production (Figure 2-17A, Steven

Dajnowicz and Jerry M. Parks performed MD simulation). The Ag85C-MA-His260cat model was chosen as the acyl-enzyme form because this configuration is required to activate the incoming acceptor molecule nucleophile. The modeled in trehalose molecule was placed nearly identical to the location observed in the Ag85B trehalose structure81. In stark contrast to the MD simulation of Ag85C-MA-His260cat, the enzyme remained in a stable catalytic state in the presence of an acceptor molecule present (Figure 2-17B). The stark difference between the MD simulations of Ag85C-MA-His260cat and Ag85C-MA-

Trehalose suggests that binding of acceptor molecule drives a conformational change from the partial or full-sequestered state back to a stable catalytic form of the acyl- enzyme intermediate required for acyltransfer.

Figure 2-17: MD model of the mycolated Ag85C intermediate with acceptor molecule of trehalose present. A) Ag85C-MA-Trehalose MD model. B) In the catalytic potion when accept molecule is present, His260 exclusively samples both the 6’ hydroxyl of trehalose and the β-hydroxyl of MA. Figure reproduced from Goins et al. the Journal of Biological Chemistry, 201857.

48

Throughout the Ag85C-MA-Trehalose MD simulation, His260 was found to exclusively sample either the β-hydroxyl of MA or the 6-hydroxyl of trehalose. Two potential pathways for the activation of the incoming nucleophile are therefore chemically viable based on this binding mode. Central to both pathways is the β-hydroxyl of MA as it is positioned to allow for either direct or indirect activation of the incoming nucleophile based on the stable catalytic conformation of His260 in the presence of trehalose (Figure 2-18).

Figure 2-18: Nucleophilic activation pathways for the second half reaction (6*OH denotes the mycolated hydroxyl of TMM). A) Direct activation of the 6’ hydroxyl of TMM for the second half-reaction by His260, stabilized by the β-hydroxyl of MA. B) Indirect activation scheme proceeding through a proton transfer from the β-hydroxyl of MA to His260 followed by the proton transfer from the 6’ hydroxyl to deprotonated β- hydroxyl of MA. Figure reproduced from Goins et al. the Journal of Biological Chemistry, 201857.

Indirect activation would proceed through a concerted proton transfer from the β- hydroxyl of MA to His260, while the proton is being transferred from the incoming

49

nucleophilic hydroxyl of the acceptor molecule to the β-hydroxyl of MA (Figure 2-18B).

The now deprotonated hydroxyl is free to undergo nucleophilic attack on the acyl- enzyme ester. The indirect pathway is chemically feasible and would be similar to a proton shuttle mechanism107-109. The more conventional direct activation pathway would consist of a simple proton transfer from the incoming hydroxyl nucleophile by His260

(Figure 2-18A). Again, the deprotonated hydroxyl is now free to undergo nucleophilic attack on the acyl-enzyme ester. Regardless of the pathway chosen, the β-hydroxyl of

MA stabilizes the deprotonated nucleophile prior to nucleophilic attack. Therefore, the final reduction of the β-ketone of MA to β-hydroxyl during biosynthesis of MA is crucial for Ag85 catalysis45. Interestingly, when the reductase responsible for the reduction of the

MA β-ketone to the β-hydroxyl is knocked out, a significant decrease in mycolated content is observed54.

Our proposed organization of substrates and intermediates as the reaction proceeds is depicted in Figure 2-19 and outlined in the corresponding figure legend. The proposed alternative model satisfies the issues raised with the previous model and highlights the importance of the dynamic nature of the α9-helix in the enzymatic activities of Ag85A, B, and C77. Specifically, this arrangement of substrates in the enzyme explains how both TMM and AG can act as acceptor molecules and why both are selectively mycolated on the 6’ or 5 hydroxyls, respectively. Therefore, the secondary trehalose binding site is not necessary for this catalytic reaction scheme, but the affinity for trehalose and sugar-based detergents cannot be ignored80, 81. As a consequence of the conformational change between native and acyl-enzyme forms, the secondary binding site changes shape and thereby likely changes the affinity of this site for carbohydrates. 50

Therefore, it seems most likely that the secondary binding site of Ag85 has affinity to trehalose to maintain Ag85s at the surface of the mycomembrane. This association is most important in the native form of the enzyme where it interacts with the mycomembrane surface through non-covalent interactions, but is less important when

Ag85 is in a covalent complex with a MA that is embedded in the mycomembrane.

Figure 2-19: Proposed structure-based catalytic cycle of Ag85s. The light blue surface corresponds to the α9-helix and dynamic loop, yellow to the MA binding site, and red to the trehalose, TMM, or arabinose-binding site. A) Apo enzyme exhibits a large TMM1 binding site (PDB: 1DQZ)80. B) Ag85C octylthioglucoside structure mimics the initial 82 TMM1 binding event (PDB: 1VA5) . C) Tetrahedral transition state (TS) of the first half-reaction modeled by the Ag85C-diethyl phosphate structure, His260 sequestered and the α9-helix relaxed (PDB: 1DQY)80. A free trehalose molecule leaves, completing the first half-reaction. D) Model of the Ag85C-MA intermediate based on the Ag85C-THL structure, α9-helix is relaxed and His260 sequestered. E) The second half-reaction proceeds through the binding of TMM2 or AG to the sugar-binding portion of the active site. The α9-helix and His260 are restored to the catalytic position as a result of acceptor binding. The model of Ag85C-MA-trehalose is shown. F) TS for the second half- reaction, again modeled by Ag85C-diethyl phosphate, leading to the formation of TDM or mAG and subsequent product release (PDB: 1DQY)80. Figure reproduced from Goins et al. the Journal of Biological Chemistry, 201857.

51

2.5 Conclusions

Mtb Ag85C was successfully crystalized with the versatile lipid esterase inhibitor

THL, which resulted in the second ever protein-THL structure73, 84. Displacement of the catalytic His260 resulted in stable inhibition and preservation of the acyl-enzyme complex, which is in stark contrast to what was observed with THL inhibition of human

FAS thioesterase domain. Therefore, the Ag85C-THL structure alone provides a secondary molecular basis for THL inhibition and can be utilized to further enhance the potency and selectivity of THL towards Ag85s and other lipid esterases.

The structural changes that afford stability of the Ag85C-THL complex were also found to exist in previous Ag85C structures with disrupted catalytic triads. Therefore, we proposed that upon enzyme acylation, structural changes occur to limit substrate hydrolysis that ultimately promotes substrate transfer. This hypothesis is supported via mutagenesis and MD simulations, which suggest a variety of sequestered conformations of His260 can exist that ultimately limit water activation. However, an active conformation of the enzyme was shown to be stable in the presence of acceptor molecule, suggesting the enzyme undergoes a conformational change between a “sequestered” form to an active form required to facilitate acyltransfer.

A new mode of substrate coordination was proposed that satisfies issues raised with the previous model. Our simplified model requires only one sugar-binding site for catalysis to occur and results in the chemical requirement of the β-hydroxyl of MA.

Based on our proposed substrate coordination, incoming acceptor hydroxyls of the second half reaction may undergo nucleophilic activation through either a direct or indirect mechanism. Ultimately, the Ag85C-THL structure provided a basis for the 52

modeling of the elusive mycolated acyl-enzyme form and allowed for the complete visualization of the catalytic cycle of mycolyltransferases.

53

Chapter 3

Inhibiting Antigen 85C

3.1 Background

3.1.1 Ag85s as drug targets

The genes encoding Ag85A, B, and C when considered independent of each other are deemed non-essential for the growth of Mtb110. However, as stated in chapter two, a knock out of either Ag85A or Ag85C severely impairs cell wall biosynthesis74, 78. As a consequence of an impaired cell wall structure via mycolyltransferase suppression or elimination, Mtb under these conditions is more susceptible to small molecules and known TB drugs such as isoniazid, rifampin, vancomycin, and erythromycin78, 79, 111.

Therefore, even moderate inhibition of Ag85s may significantly increase the efficacies of known TB drugs111.

While not considered essential as independent genes, in combination the fbp genes encoding Ag85s are considered synthetically lethal112. Consequently, inhibition of at least two of the three Ag85s would be sufficient enough to inhibit Mtb growth without the addition of a second drug. This idea becomes even more appealing given that the general structures and specifically the active sites of Ag85A, B, and C are nearly identical80-82. Therefore, if designed appropriately, a strong Ag85C inhibitor would also 54

inhibit Ag85A and Ag85B or vice versa. An inhibitor of all three enzymes would reduce the potential for drug resistance, as the same mutation would have to occur simultaneously in two independent genes. Finally, the cellular location of Ag85s makes them an appealing target for drug development. Ag85s are secreted from the cytoplasm and can be found in cell wall fractions and in the ES42. Therefore, Ag85s are far more accessible targets than essential enzymes found in the cytoplasm, which resides behind the impervious Mtb cell wall. Due to the listed reasons above, Ag85s are intriguing targets for drug development.

3.1.2 Previously developed inhibitors

Previously developed inhibitors of Ag85s can be broken down into two categories: covalent and non-covalent inhibitors. Two predominant scaffolds have been used to non-covalently target Ag85s, thiophenes and alkyl-thio-arabinofuranosides113-115.

A small library of alkyl-thio-arabinofuranosides derivatives were previously synthesized by the Sucheck group to target the TMM binding site of Ag85s by mimicking an acylated arabinose molecule113. Direct evidence of in vitro inhibition against Ag85s was never determined; however, the best molecule had an MIC value of 256 μg/mL against

Msmeg113. The heterocyclic thiophene scaffold was found by Warrier et al. to weakly inhibit Mtb growth in vivo (MIC 100-200 μM)115. Lipid extracts indicated an overall decrease in TDM, suggesting Ag85s as a potential target115. Their group showed weak binding to Ag85C through solution based NMR115. Thiophene selectivity was increased towards Ag85C by the Sucheck group through the conjugation of the thiophene core to thiol linked arabinose114. The addition of the arabinose moiety afforded affinity to the sugar binding site of the active site; the best derivative having a Ki value of ~18 μM 55

against Ag85C114. To date, the thiophenyl-thioarabinofuranosides are the best, non- covalent inhibitors of Ag85C reported in literature. The in vitro evaluation of a second- generation library of thiophene inhibitors synthesized by the Sucheck group will be discussed in this chapter116.

Figure 3-1: Ag85C inhibitors.

The first covalent inhibitors of Ag85s used a phosphonate group as the covalent warhead. The phosphonate group afforded covalent modification of the serine nucleophile and results in irreversible inhibition. Gobec and colleagues made a small library of phosphonate derivatives with alkyl chains of various lengths and trehalose mimics. The best compound had a reported IC50 value of 2 μM against Ag85C and an

MIC of 200 μg/mL against Mycobacterium avium (Mavium)117. As with the non-covalent inhibitors, the phosphonate derivatives were designed to target both MA and sugar binding sites of the active site117.

The second covalent inhibitor of Ag85s was discovered by the Ronning lab via in vitro screening of Ag85C against the National Institute of Health (NIH) clinical library85.

56

The selenium containing molecule, ebselen (Ebs), was discovered to be a potent inhibitor of Ag85s in vitro and Mtb in vivo85.

3.1.3 Inhibition by ebselen (Portions of text reproduced from Goins et al. ACS

Infectious Diseases, 2017)98

Inhibition of Ag85C by Ebs proceeds through a unique mode of covalent allosteric modification. Ebs is low molecular weight, organoselnium molecule that readily modifies free thiols118. Ebs was shown to covalently modify the only cysteine,

Cys209, in Ag85C, which is conserved, non-catalytic, and solvent-accessible through the reversible oxidation of the γ-sulfur of Cys209 by the selenium of Ebs (Figure 3-2).

Figure 3-2: Mechanism of Ebs modification. Inhibition is reversible in the presence of a reducing agent.

The resulting modification of Cys209 leads to the relaxation of the native, kinked

α9 helix, displacing the catalytically relevant His260 and Glu228 residues from the active site, and thereby rendering the enzyme inactive through covalent allosteric inhibition

(Figure 3-3)85. This form of the inhibited enzyme is similar to the proposed sequestered form of Ag85s that occurs upon acylation, discussed in chapter 257. The crystal structure of Mtb Ag85C covalently modified by Ebs was solved at 1.4 Å93. However, due to dynamic movement of Ebs, complete electron density was lacking for the secondary phenyl ring of Ebs93. As a result, a full assessment of protein-inhibitor interactions was

57

not possible, resulting in a cursory assignment of interactions between Phe254 and the primary phenyl ring of Ebs, and a hypothesized interaction between the guanidinium moiety of Arg239 and the carbonyl of the amide of Ebs93.

Figure 3-3: Ebs modified Ag85C. Inhibition of Ag85C proceeds through covalent, allosteric inhibition85. Reversible modification of Cys209 by Ebs results in active site disruption through structural changes (Ag85C-Ebs, PDB: 4QDU)85, 93.

Ebs has an MIC against H37Rv Mtb of 12.5 μg/ml, reduced TDM biosynthesis, and was shown to potently inhibit Ag85C in vitro85, 102. In efforts to optimize the original

Ebs scaffold, the Sucheck group has synthesized a library of Ebs derivatives (Figure 3-

4)102. The in vitro inhibitory properties against Ag85C are provided in this chapter.

Furthermore, two crystal structures of Ag85C in complex with Ebs derivatives are presented that allow for a complete assessment of protein-drug interactions. These structures allowed for the calculation of protein-drug interaction energetics that support a structure based rationale for inhibitory differences between Ebs derivatives. Finally, the influence of Ebs modification on Ag85C stability is investigated. All together, this chapter provides a complete characterization of Ag85C inhibition by Ebs and its derivatives.

58

Figure 3-4: Ebs derivative library. Compounds are numbered as to how they will be referred.

3.1.4 Covalent warheads of serine esterases

Covalent inhibitors of enzymes are often considered problematic due to the potential for cross reactivity with unintended targets119. However, if the covalent inhibitor is designed with specificity in mind, the benefits can outweigh the risk. Covalent inhibitors can have prolonged resonance times in enzyme active sites if designed to be reversible119-121. While on the other hand, covalent inhibitors can act as suicide inhibitors if the covalent warhead chosen is nonreversible119. Given that Ag85s are secreted from

Mtb, prolonged protein-drug residence times would be ideal. Furthermore, serine proteases, lipases, and esterases have been frequent targets for covalent inhibitor development; therefore, Ag85s are ideal candidates for covalent inhibitors119.

In this chapter we investigate two types of covalent warheads with the ambition of developing selective covalent inhibitors towards Ag85s. Based on the covalent inhibition of Ag85s by THL, a small group of β and γ-lactones were screened against Ag85C to determine general viability of scaffolds containing these covalent warheads. In addition to the development of known covalent inhibitors, Ag85C was screened against two TB

59

active drug libraries, one from the TB alliance and the other from GlaxoSmithKline

(GSK) in hopes of discovering further scaffolds for future development.

Figure 3-5: Lactones investigated for Ag85C inhibitors.

3.2 Methods

3.2.1 Protein expression and purification

WT Ag85C was expressed and purified in identical fashion as described in chapter 2, section 2.2.2 for inhibition and crystallographic studies. Again following ammonium sulfate precipitation, purified enzyme was dialyzed into 50 mM sodium phosphate pH 7.5 for assays and 10 mM TRIS pH7.5 and 2 mM EDTA for crystallization.

3.2.2 Evaluation and characterization of ebselen derivatives

3.2.2.1 In vitro inhibition (% activity and IC50 determination), (Portions of text were reproduced from Thanna et al. the Journal of Organic Chemistry, 2017)102

To determine % activity levels for Ebs derivatives at a set concentration and time point, Ag85C was reacted with 5 μM of respective Ebs derivative (10 mM DMSO stock) for 40 min at room temperature. The enzymatic activity of Ag85C, covalently modified with these analogues, was evaluated using the RfB assay described in chapter 2, section

2.2.3. Again, reactions were initiated upon titration of RfB with reads conducted in triplicate. Data were processed using PRISM 5.

60

An apparent IC50 value for each Ebs derivative was obtained for Mtb Ag85C by varying the concentration of inhibitor, ranging from 40.0 μM to 312.0 nM (10 mM

DMSO stocks). Compounds were incubated with Ag85C for 15 min at room temperature, prior to kinetic read initiation. Again, enzymatic activity was assessed in triplicate using the RfB assay, with linear rates assessed in PRISM 5. The apparent IC50 was calculated using the following equation: IC50 = [(50 − A)(B −A)] × (D − C) + C, where points were expressed in percent inhibition: A = the point on the curve that is less than 50%, B = the point on the curve that is greater than or equal to 50%, C = the concentration of inhibitor that gives the A % inhibition, and D= the concentration of inhibitor that gives the B % inhibition122.

3.2.2.2 Crystallization and structure determination (Portions of text were reproduced from Goins et al. ACS Infectious Diseases, 2017)98

Based on observed in vitro inhibition, two Ebs derivatives were selected for crystallization studies. Crystals for the Ag85C azido-Ebs complex were achieved through co-crystallization using the hanging drop vapor diffusion method. Purified Ag85C at 4.2 mg/mL with azido-Ebs (10 mM DMSO stock) added in a 1:1.1 molar ratio, respectively, and allowed to incubate on ice for 15 minutes prior to drop set up. The Ag85C azido-Ebs complex crystallized in 1:1 ratio of protein to well solution after a week of incubation at

16 °C. The well solution was composed of 0.1 M ammonium acetate, 0.05 M HEPES pH

7.5, and 12.5 % w/v polyethylene glycol 3,350. Crystals were cryoprotected through the addition of 0.25 μL of glycerol to the drop immediately prior to looping and flash cooling in liquid nitrogen. X-ray diffraction data were collected at the Advanced Photon Source

(APS) at Argonne National Labs, on LS-CAT beam line F. 61

Crystals for the Ag85C adamantyl-Ebs complex were also obtained via co- crystallization using the hanging drop vapor diffusion method. Adamantyl-Ebs (40 mM

DMSO stock) was added to purified Ag85C at 4.0 mg/mL in a 1:1.1 protein to inhibitor molar ratio and allowed to incubate on ice for 15 minutes prior to drop setup. Crystals formed in a 1:1 ratio of protein to well solution, comprising 0.1 BIS-TRIS pH 5.5, and

25% w/v polyethylene glycol 3,350 after four days of incubation at 16 °C. Crystals were cryoprotected through the addition of 0.25 μL glycerol to the drops immediately prior to looping and flash cooling in liquid nitrogen. X-ray diffraction data were collected at the

APS LS-CAT beam line G.

Both data sets were indexed, integrated, and scaled using HKL200087. The X-ray crystal structures of Mtb Ag85C covalently modified at residue Cys209 by azido-Ebs and adamantyl-Ebs were solved at 2.01 Å and 1.3 Å resolution, respectively. The Ag85C- azido-Ebs crystals had a C2221 space group with the phase solution coming from molecular replacement using Phaser MR with the previous Ag85C-Ebs structure with both Ebs and the Ebs modified Cys209 omitted from the search model (PDB: 4QDU)88.

Two molecules were determined to be in the asymmetric unit and confirmed by the molecular replacement solution. The Ag85C-adamantyl-Ebs crystals had a P43212 space group, the phase solution was also solved by molecular replacement with the same search model as described above (PDB: 4QDU) and supported one molecule per asymmetric unit. Each structure was subjected to a rigid body and simulated annealing refinement to remove model bias followed by rounds of XYZ coordinate, real-space, occupancy, and

B-factor refinements (Phenix Refine) between manual modelling with COOT88, 89.

Following rounds of model refinement and manual building, final Rwork/Rfree values of 62

0.1623/0.1968 and 0.1660/0.1763 were obtained for the azido and adamantyl structures, respectively. Restraints for both azido and adamantyl ligands covalently linked to C209 were generated using eLBOW (electronic Ligand Building and Optimization

Workbench)90. Ligand geometries were further optimized using REEL90.

3.2.2.3 kinact/KI determination of selected derivatives (Portions of text were reproduced from Goins et al. ACS Infectious Diseases, 2017)98

kinact/KI determination of selected derivatives was determined in a similar manner to that of THL described in chapter 2, section 2.2.3.2 using the RfB assay. Ag85C in the presence of trehalose was quickly titrated with Ebs or the corresponding Ebs derivative

(10 mM DMSO stocks, inhibitor was serial diluted for desired concentrations) and an equivalent v/v % DMSO for the uninhibited reaction immediately prior to RfB addition and the subsequent fluorometric measurement. Reactions were performed in triplicate.

After subtracting the rate of background fluorescence production and converting from relative fluorescent units to product concentration (resorufin standard curve given in figure S2), kobs was calculated by fitting the triplicate kinetic data in PRISM 7 with a one- phase association equation, Y = Y0 + (Plateau – Y0)(1-exp(-kx)). This is equivalent to the commonly used [Product] = Vi/kobs[1-exp(-kobs*t), where Y = [Product], (Y0 + (Plateau –

86 Y0) = Vi/kobs, k = kobs, and x = t or time . To obtain the best fit, time points 0 to 5 minutes were used for 3 and 2 μM Ebs, while 0 to 10 minutes were used for 1.5, and 1

μM Ebs reactions. All data for azido Ebs was fitted using 0 to 5 minutes, while 0 to 40 minutes were used for adamantyl Ebs. kinact/KI was determined by plotting kobs vs inhibitor concentration and fitting the data to a line through the origin.

63

3.2.2.4 Differential scanning fluorimetry (Portions of text were reproduced from

Goins et al. ACS Infectious Diseases, 2017)98

Relative protein stability was determined by differential scanning fluorimetry

(DSF) using a Bio-Rad CFX96 Real-Time PCR Detection System. Purified Ag85C in the crystallization buffer was concentrated to 36 μM (1.2 mg/ml) and incubated with 100 μM of either Ebs, Ebs derivatives, or an equal volume of DMSO (1 % v/v) for 15 minutes.

Ebs and both derivative stocks were at 10 mM in DMSO. SYPRO® Orange was added to a final concentration of 2.5x (5,000x DMSO stock, Life Technologies, diluted to 200x stock with buffer), after 25 μL of the respective master stock reaction solution was aliquoted into the 96 well PCR plate and sealed (Bio-Rad). To determine what conditions produced the best melt curves, concentrations of protein and SYPRO Orange were varied from 0.5 to 2 mg/ml and 1 to 5x, respectively,. The plate was centrifuged at 700 rpm (651 x g) for 2 minutes to remove any air bubbles. The melt curve protocol is as follows: reactions were initially incubated at 25 °C for 3 minutes followed by a 0.5 degree increase, being held for 3 seconds, followed by a fluorescent read using the FRET setting, until a final temperature of 95 °C was achieved. Data were analyzed using the Bio-Rad

CFX Manager 3.1 software. Triplicate data sets were averaged in Excel and plotted in

Prism 7 to yield the final melt peaks.

3.2.3 In vitro screening of thiophene derivatives

Compounds were dissolved in DMSO to a stock concentration of 10 mM.

Thiophene derivatives were screened at 100 uM with 0 and 1 hr incubation periods, using the RfB assay with single reactions. Reaction rates were determined with a linear fit and plotted in excel. The average and standard deviation of the DMSO controls were 64

determined and denoted on the screening plot. Compounds displaying rates that were lower than 3 times the standard deviation of the DMSO controls were considered statistical hits.

3.2.4 In vitro screening of β and γ-lactones

N-octanoyl-L-Homoserine and N-(β-ketocaproyl)-L-Homoserine and Lactones were purchased from Cayman chemical. Gibberellic acid, (4R)-4-Benzyl-3-(5- hexenoyl)dihydrop-2(3H)-furanone, clasto-lactacystin β-lactone, and andrographolide were purchased from Sigma. Respective lactones were dissolved in DMSO to a set concentration of 10 mM. Compounds were screened using the previously described RfB assay, in triplicate at reaction concentrations of 100 μM. Following titration of drug and

RfB, kinetic reads were initiated. Progress curves with background subtracted were assessed for signs of covalent inhibitions. For over night incubation, enough reaction master mix was made to allow for two, triplicate readings at time point “0” and overnight. Reactions were kept at 4 ˚C for overnight incubation.

3.2.5 Drug library screening

3.2.5.1 Screening parameters

The Ag85C RfB assay was previously determine to have a Z’ value of 0.82, indicating the assay suitable for high through put drug screening123. A Z’ value is a statistical evaluation of assay reliability based on signal to noise and calculated error of control reactions123. A Z’ between 0.5 and 1 is considered statically sound, with 1 being the theoretical best123. The TB alliance and GSK libraries combined contain over 1500 compounds indicated to be active against various strains of Mtb. Both libraries were screened by hand, in singlet reactions (40 drugs per kinetic read), using a multichannel 65

pipette over the course of multiple days. Compounds came in 10 mM DMSO stocks and were therefore screened at 100 μM with 15 minute incubation prior to kinetic read. To account for variation between plate reads, reactions were plotted as a function of V’/V’o, where V’ was inhibited rate and V’o the uninhibited DMSO control for that specific read set. The average and standard deviation of the positive controls was calculated in excel.

Compounds with inhibition values 3 times lower than the standard deviation of DMSO controls were pursued as statistical hits.

3.2.5.2 Hit validation

Seven lead molecules from the TB alliance library were selectively chosen from initial hits based on compound structure, price, and availability (ChemBridge). Two lead molecules were selected for dose dependence from the GSK library (Taken from library wells). Dose dependence was used for hit validation by serial diluting compounds dissolved at 10 mM in DMSO. Compounds were tested in triplicate at 100, 50, and 25

μM. Again, compounds were allowed to incubate with enzyme prior to RfB titration and kinetic read. Data was processed in PRSIM 7.

3.3 Results

3.3.1 Evaluation and characterization of ebselen derivatives (Portions of text reproduced from Thanna et al. Journal of Organic Chemistry, 2017 and Goins et al.

ACS Infectious Diseases, 2017)98, 102

3.3.1.1 In vitro inhibition

Ebs derivatives were initially screened at various concentrations and time points to evaluate both time and concentration effects (some of this preliminary data can be 66

found in Dr. Lorenza Favrots dissertation). Based on those observations, 40 minutes and

5 μM were chosen as a median time point to standardize all Ebs derivatives to (Figure 3-

6). Compound 3a, Ebs, resulted in 17 ± 3 % enzymatic activity, all derivatives were benchmarked to this value. Compounds 3b, 3d, 3e, 3f, 3i, 3j, all showed comparable levels of inhibition at this given concentration and time point.

Figure 3-6: Initial screening of Ebs derivatives. (Figure adapted from Thanna et al. Journal of Organic Chemistry, 2017)102

To directly compare concentration dependence of Ebs derivatives, IC50 values were determined for the entire library (Figure 3-7). Respective inhibitors were allowed to incubate with enzyme 15 minutes prior to kinetic read. Ebs was found to have an apparent IC50 of 5.1 μM. Compounds 3b and 3h had IC50 almost 10 fold lower in value, being 540 nM and 720 nM, respectively. However, remaining inhibitors all had similar

IC50 values, aside from compounds 3c and 3l, being roughly 5 times worse than ebs.

Compound 4 had an IC50 value over 100 μM.

67

Figure 3-7: IC50 curves of Ebs derivatives (Figure adapted from Thanna et al. Journal of Organic Chemistry, 2017)102

Based on the initial characterization of derivatives, no clear trend was present.

Despite a variety of chemical modifications, most inhibitors behaved in similar fashion to the original Ebs. To further investigate the lack of an apparent trend or structure activity relationship (SAR), two bulkier, derivatives were chosen for further characterization. One

68

being comparable to Ebs, azido-Ebs (compound 3j), and a worse inhibitor, adamantyl-

Ebs (compound 3l)(Figure 3-8).

Figure 3-8: Ebs and its derivatives chosen for further analysis; (a) Ebs, compound 3a (b) p-azido Ebs, compound 3j (c) adamantyl Ebs, compound 3l (Reproduced from Goins et al. ACS Infectious Diseases, 2017)98

Per the recommendation of an editor at ACS Infectious Diseases, the kinact/KI values were determined for selected derivatives. As stated early, kinact/KI evaluation is the best way to characterize covalent inhibitors as time and concentration are considered86. In agreement with initial evaluation, EBS was found to have the best kinact/KI value being

0.31 ± 0.01 uM-1Min-1 (Figure 3-9). Azido-Ebs was found to have a slightly lower value than Ebs, being 0.18 ± 0.01 uM-1Min-1, while Adamantyl-Ebs was nearly 50 times lower than Ebs (Figure 3-9). Do to the quick rate at which Ebs inhibits Ag85C, higher

86 concentrations were not attainable; therefore, a kinact/KI was calculated using a linear fit .

Unfortunately, the rate of inactivation and binding coefficients cannot be deconvoluted from each other using this method. However, based on raw progress curves at the highest concentrations tested, Ebs and Azi-Ebs inhibit Ag85C to near completeness within 5 minutes, while Adamantyl-Ebs required nearly half an hour. Additionally, it is apparent that slightly more Azi-Ebs is required and takes longer to fully inhibit Ag85C when compared to Ebs. These observations are consistent with the previously determined IC50 values.

69

Figure 3-9: kinact/KI Determination for Ebs, Azido-Ebs, and Adamantyl-Ebs. (a-c) Progress curves of the enzymatic reactions in the presence of specific inhibitor at specific concentrations as a function of increasing reaction product (resorufin, μM) over time (minutes). The average of the triplicate reads is plotted and fitted with a one-phase association equation. (d-f) Resulting kinact/KI plots for Ebs, adamantyl-Ebs, and azido- Ebs. (Reproduced from Goins et al. ACS Infectious Diseases, 2017)98

3.3.1.2 Crystal structures of Ag85C and ebselen derivatives

The Ag85C-azido-Ebs and Ag85C-adamantyl-Ebs structures were solved at 2.01 and 1.30 Å, respectively. Complete data collection and refinement statistics are presented in table 3.1. The generated 2Fo-Fc map for the azido-Ebs structure displayed electron density up to the second phenyl ring of the modifier, where partial density was observed when contouring the map at 1.5σ. Difference density was not observed for the linear azide moiety. As a result, further refinement of this portion of the inhibitor involved modeling it as a terminal amine.

70

Table 3.1: Crystallographic and Refinement Statistics for Ag85C-Ebs derivative structures. Parentheses indicate the value for the highest resolution shell. (Table reproduced from Goins et al. ACS Infectious Diseases, 2017)98 Ag85C – Azido Ebselen Ag85C – Adamantyl Ebselen PDB ID 5KWJ 5KWI Data Collection Diffraction source APS LS-CAT F APS LS-CAT G Wavelength (Å) 0.97872 0.97857 Space group C2221 P43212 a, b, c (Å) 88.474, 88.479, 161.950 63.405, 63.405, 160.210 α, β, γ (°) 90, 90, 90 90, 90, 90 Mosaicity (°) 0.605 0.318 Resolution range (Å) 49.51-2.01 40.85-1.30 No. of unique reflections 42348 (3917) 81206 (7968) Completeness (%) 99.32 (93.13) 99.96 (99.67) Redundancy 14.9 (15.0) 13.0 (6.9) 〈 I/σ(I)〉 20.91 (5.50) 20.00 (2.69) R meas 0.104 (0.563) 0.071 (0.578) Overall B factor from 24.59 14.90 Wilson plot (Å2) Refinement Resolution range (Å) 49.51 – 2.01 40.85 - 1.3 Completeness (%) 99.32 (93.13) 99.94 (99.67) Rwork/Rfree (%) 16.23 / 19.68 16.60 / 17.63 Total No. of Atoms 4781 2533 Protein 4452 2237 Solvent 328 296 RM.S. deviations Bonds (Å) 0.008 0.006 Angles (°) 1.090 1.550 Average B factors (Å2) 24.59 14.88 Protein 25.80 15.80 Solvent 34.00 25.20 Ramachandran plot Most favoured (%) 96.81 97.53 Allowed (%) 3.19 2.47

The calculated, likelihood-weighted Fo-Fc omit maps for both structures were generated by omitting the modeled covalent modifier as well as the -sulfur and -carbon

71

atoms of Cys209 (figure 3-10). Strong difference density for the covalent linkage between the Cys209 sulfur atom of Ag85C and the selenium of the Ebs derivatives is present in both structures. With regards to azido-Ebs, difference density for the primary and secondary phenyl rings is observed, yet not the azide moiety itself, further supporting the truncated p-amino model.

Figure 3-10: Calculated Fo-Fc omit maps for the modified Cys209 of Ag85C and respective Ebs derivatives. Both maps are contoured at 3σ in blue mesh. Atom color: carbon is beige in (a), orange in (b), nitrogen blue, oxygen red, sulfur yellow, selenium light orange. (a) Ag85C covalently modified by azido-Ebs (b) Ag85C covalently modified by adamantyl-Ebs. (Reproduced from Goins et al. ACS Infectious Diseases, 2017)98

In both structures, the derivatized end (p-amino and adamantyl) of the covalent modifier is pointing toward solvent, with the carbonyl of the amide linkage oriented toward Cβ of Phe252. In agreement with the previous Ag85C-Ebs structure, the primary phenyl ring of the Ebs derivative forms a centered, yet tilted π - π stacking interaction with the aromatic side chain of Phe254 in both the azido and adamantyl structures (Figure

3-11)93. Notably, a favorable cation - π interaction may be split between the parallel orientation of the guanidinium moiety of Arg239 to the sp2-hybridized nitrogen of the

72

amide linkage and to the secondary sp2-hybridized phenyl ring of the modifier (Figure 3-

11a). Conversely, in the adamantyl-Ebs structure this secondary phenyl ring of Ebs is omitted and replaced with a bulky hydrophobic, sp3-hybridized adamantyl group. As a result, the observable cation - π interaction is no longer split and solely resides between

Arg239 and the nitrogen of the amide linkage with the guanidinium moiety once again parallel to the inhibitor (Figure 3-11b).

Figure: 3-11: Crystallographic interactions of Ebs derivatives with Ag85C. Interactions observed in the crystal structure of Ag85C–azido-Ebs, are similar to those observed in the Ag85C–adamantyl-Ebs structure (b). (Reproduced from Goins et al. ACS Infectious Diseases, 2017)98

Consistent with the previously published Ag85C-Ebs structure, a relaxation of the

α9 helix and restructuring of a connective loop (Gly210 - Pro223) is observed as a result of covalent modification of Cys209 (Figure 3-12)85, 93. However, there is a noticeable difference in the observed movement of the α9 helix when the azido and adamantyl Ebs structures are superimposed with the previous Ag85C-Ebs structure (RMSD = 0.220 and

0.201 Å2, respectively, compared with PDB entry 4QDU)93. The movement of the α9 helix shifts residue Phe226 by 1.8 and 1.6 Å in the azido and adamantyl structures, respectively. Helical movement from native to modified with respect to the azido-Ebs is

9.5 Å when measured from the amide backbone nitrogen of Lys225. In the new

73

structures, helical movement directly results in a nearly identical conformational change to Trp157, rotating χ2 by +110 ° with respect to the previous Ag85C-Ebs structure. This side chain rotation reorients Trp157 perpendicular to the α9 helix compared to the previous parallel orientation observed in 4QDU (Figure 3-12a and b)93. An additional consequence of the larger shift in the α9 helix is the elongation and restructuring of the

14-residue loop connecting the β7 strand to the α9 helix (Gly210 - Pro223).

Figure 3-12: Further observed structural shift of α9 helix away from the active site, restructuring the connecting loop, resulting in the loss of thermal stability. (a) Alignment of the Ag85C-Ebs structure (Gray, PDB:4QDU) with the Ag85C-azido-Ebs structure (Beige)93. (b) Alignment of the Ag85C-Ebs structure (Gray, PDB:4QDU) with the Ag85C-adamantyl-Ebs structure (Orange)93. (Reproduced from Goins et al. ACS Infectious Diseases, 2017)98

3.3.1.3 Influence of ebselen on protein stability

Native Ag85C with no drug present displayed a melt temperature of 55.0 ± 0 °C, whereas Ag85C modified with Ebs and azido-Ebs displayed melt temperatures of 40.7 ±

0.3 °C and 41.0 ± 0 °C, respectively. Ag85C modified with adamantyl-Ebs resulted in a bi-phasic melt curve, yielding two melt temperatures of 41.0 ± 0 °C and 54.5 ± 0 °C (raw melt curves given in figure 3-13, helical movement highlighted in figure 3-14 with averaged melt peaks).

74

Figure 3-13: Raw triplicate data for the obtained melt curves. (b) Resulting melt peaks from DSF. DMSO corresponds to sample with No Drug Added. (Reproduced from Goins et al. ACS Infectious Diseases, 2017)98

Figure 3-14: Influence of structural changes on protein stability. a) Overlay of α-helix 9 between native enzyme (Green, PDB:1DQZ), Ebs modified enzyme (Gray, PDB:4QDU), azido-Ebs modified enzyme (Beige), and adamantyl-Ebs modified enzyme93. (b) Ag85C modified with Ebs or either of the derivatives displays a 14 °C thermal shift from native enzyme. (Reproduced from Goins et al. ACS Infectious Diseases, 2017)98

3.3.2 Evaluation of Thiophene derivatives

Screening of the second-generation thiophene library resulted in the identification of no “hits” with regard to Ag85C (Figure 3-15)116. Control reactions with DMSO are plotted in red, no statistical hits were observed. Reactions 50 and 51 are with Ebs and

THL, serving as controls.

75

Figure 3-15: Screening results of thiophene library. No compounds were identified as strong inhibitors. Reactions 50 and 51 are controls.

3.3.3 Evaluation of β and γ-lactones

Lactones screened displayed no significant inhibition towards Ag85C (Figure 3-

16). Lactones based on naturally occurring compounds were allowed to incubate overnight with Ag85C to evaluate time dependencies. No significant decrease in enzymatic activity was observed (Figure 3-16A). It should be noted that Clasto- lactacystin-β-lactone was the only β-lactone screened. Interestingly, this molecule also contains a γ-lactam. All other lactones tested, including the synthetic series in Figure 3-

16B were γ-lactones. All reaction progress curves were inspected for signs of covalent inhibition; however, none was observed as all were linear in nature.

Figure 3-16: Lactones screened against Ag85C activity displayed little to no inhibition. A) Lactones derived from natural sources. B) Synthetic, hydrophobic lactones. 76

3.3.4 Drug library screening

The TB alliance drug library contains over 1400 compounds believed to be active against TB; however lack known targets. Screening Ag85C against this library resulted in the identification of 25 potential “hits” (Figure 3-17). “Hits” were considered to decrease enzymatic activity by over 38 % based on the 3σ cut off.

Figure 3-17: TB alliance drug screen results. 3σ cut off for hits is denoted by red line above V’/V’o of 0.6. Reactions near zero were negative control reactions with no enzyme present.

Selected hits per guidelines outlined in methods section 3.3.4 were serial diluted to evaluate dose dependence. All selected inhibitors indeed displayed dose dependence; however, compounds 141, 1410, and 1516 displayed inhibition levels not consistent with values observed in the preliminary screen (Figure 3-18). However, compounds 308, 77

3344, 403, and 801 displayed similar levels of inhibition as determined through library screening and are considered valid “hits.

Figure 3-18: Dose dependence of selected “hits” from preliminary screen.

The GSK library contains 177 drug molecules known to inhibit TB, yet lack identified targets124. Screening against Ag85C resulted in the identification of two compounds that were barely below the 3σ cut off (Figure 3-19). Identified hits displayed dose dependence and moderate inhibition levels similar to screening results (Figure 3-

17). Therefore, these two hits were considered validated.

78

Figure 3-19: Screening results from GSK library. 3σ cut off denoted by green line.

Figure 3-20: Dose dependence of “hits” identified in intial GSK library screening.

3.4 Discussion

3.4.1 Ebselen derivatives (Portions of text reproduced from Goins et al. ACS

Infectious Diseases, 2017)98

Derivatization of Ebs at the site of the secondary benzyl ring resulted very little enhancement or reduced inhibitory properties, in vitro or in vivo. Summary of inhibitory

79

data provided in table 3.2. In general, the better in vitro inhibitors did correlate to better in vivo activity against H37Rv strain of Mtb; however, no clear rationale was present for observed inhibition values with regard to a classical SAR study102. On this basis, two derivatives with differing chemical and inhibitory attributes, in addition to Ebs, were selected for further characterization.

Table 3.2 Summary of in vivo and in vitro inhibition values of Ebs derivatives. (Table adapted from Thanna et al. Journal of Organic Chemistry, 2017)102 Derivative MIC (μg/mL) Initial Activity (%) App. IC50 (μM) 3a 12.5 17 ± 3 5.1 3b 25 24 ± 17 0.5 3c 12.5 36 ± 4 28.6 3d 25 20 ± 4 1.2 3e 25 20 ± 10 6.5 3f 12.5 15 ± 2 8.8 3g 50 30 ± 16 5.3 3h 25 19 ± 2 0.7 3i 25 21 ± 7 1.0 3j 25 20 ± 4 1.1 3k 25 31 ± 12 1.5 3l 50 59 ± 14 25 3m 25 40 ± 17 4.1 3n 25 44 ± 15 3.7 4 100 80 ± 7 >100 6 N/A 61 ± 11 2.02

Determination of kinact/KI revealed that inhibition of Ag85C by Ebs proceeds slightly faster than that of azido-Ebs, and almost 50-fold faster than adamantyl-Ebs.

Covalent inhibition can proceed through three schemes: nonspecific, quiescent, and mechanism-based86. Non-specific is a result of multiple accessible covalent sites, that readily undergo modification, yet each site may not lead to inhibition; therefore, a saturation/full inhibition point may never be obtained with respect to inhibitor concentration. Quiescent and mechanism-based modification rely on the non-covalent

80

binding affinity of an inhibitor as well as the rate of covalent modification86. While inhibition of Ag85C by Ebs and its derivatives is most likely represented by quiescent inhibition, as full inhibition is certainly obtainable and there must be an initial short lived non-covalent binding event, we analyzed the data as non-specific for two major reasons.

Complete inhibition by Ebs and azido-Ebs at higher concentrations (2 and 3 μM) is obtained within two to four minutes, drastically decreasing in time as concentration increases. This becomes problematic as the transition from initial rate to terminal rate is simply lost since the time-resolution of our experimental setup is not sufficiently high, therefore accurately determining kobs is challenging. The second issue is the high concentration of enzyme required, 500 nM, for the assay, which limits our ability to use lower inhibitor concentrations as we are already near a 1 to 1 stoichiometric equivalent of inhibitor to enzyme. For these reasons, kinact/KI was determined using non-specific inhibition analysis as a reasonable compromise.

Similar to the previously determined structure of Ag85C by Ebs, modification of

Cys209 by adamantyl-Ebs and azido-Ebs resulted in structural rearrangements that disrupt the catalytic triad85, 93. However, the two novel structures allowed for the resolution of both Ebs derivatives, allowing for a more accurate assessment of protein- drug interactions. The primary ring of Ebs was found within π-π stacking distances to

Phe254 and the second major interaction observed being the positively charged side arm of Arg239 with the amide linkage to the modified portion of Ebs. To investigate the influences of these observed chemical interactions, energetics of protein-drug interactions using density functional theory (DFT) calculations were performed (Figure 3-21). The azido-Ebs structure was used as a basis for calculations of Ebs-protein interaction 81

energetics, the primary amine was simply removed from the model. Computational calculations were performed by Steven Dajnowicz and Jerry M. Parks, methods and results can be found in Goins et al. ACS infectious diseases, 201798.

Figure 3-21: Crystallographic interactions of Ebs derivatives with Ag85C and calculated energetics of interacting monomers after QM geometry optimization. (a) Interactions observed in the crystal structure of Ag85C–azido-Ebs, are similar to those observed in the Ag85C–adamantyl-Ebs structure (d). (b, c, e, f) Interaction energetic curves for respective residue derivative interaction. (Reproduced from Goins et al. ACS Infectious Diseases, 2017)98

At the equilibrium geometry (intermonomer distance of 3.90 Å) the noncovalent interaction energy between Phe254 and adamantyl-Ebs was computed to be -8.0 kcal/mol in the condensed phase (i.e., ε = 10). The computed interaction energy minimum for the

Phe254 and azido-Ebs pair is -10.2 kcal/mol at an intermonomer separation of 3.86 Å, and -8.9 kcal/mol at 3.55 Å for Phe254 and Ebs. Thus, all Ebs derivatives are capable of forming favorable stacking interactions with Phe254, which is expected to contribute to initial drug recognition and binding.

82

Of greater intrigue to us are the interaction energetics between the cationic guanidinium side chain of Arg239 and respective Ebs molecules, since the observed noncovalent interaction differs between the two derivatives. The calculated minimum interaction energy for Arg239 and adamantyl-Ebs is -4.1 kcal/mol for the condensed phase model at 5.04 Å. Conversely, the interaction energy for Arg239 and azido-Ebs is -

7.0 kcal/mol for the condensed phase model at 3.78 Å, approximately double that computed for the adamantyl-Ebs system. The Arg239 interaction with Ebs at an equilibrium distance of 3.76 Å is -7.4 kcal/mol. The comparable values for the Ebs and azido-Ebs – Arg239 interactions are consistent with a constructive split cation – π interaction, whereas the higher interaction energetics for the Arg – adamantyl-Ebs support much weaker cation – sp3 dispersive interactions. This interaction increases the distance between the guanidinium moiety and the amide nitrogen by 1.3 Å when compared to Ebs and azido-Ebs. This increase in distance may be the cause for the reduced interaction energies calculated for the Arg239 - adamantyl-Ebs model.

Comparing both sets of complete interaction energy curves for Ebs and the two derivatives, we see that each molecule shares similar interaction energies with Phe254.

However, the interactions with Arg239 and the site of derivatization display obvious differences. These observations suggest that interactions between the inhibitor and

Arg239 may facilitate initial noncovalent drug binding to Ag85C and may partially explain the differences in the observed in vitro kinetic inhibition properties. Indeed, derivatives with chemical moieties that favorably interact with Arg239, have enhanced inhibitory properties towards Ag85C. In addition, the shape of the secondary benzyl ring substituents may influence inhibition. This is best observed when comparing methoxy 83

and fluorinated derivatives. Derivatives with substituents at the para position display overall better inhibition levels than ortho substituted derivatives (3h: para-fluoro > 3i: ortho-fluoro, 1b: para-methoxy > 1c: ortho-methoxy). Based on the structures of Ebs derivatives with Ag85C, derivatives with ortho-substituted substituents may have steric clash with the enzyme. Therefore, inhibition of Ag85C by Ebs is governed by interactions with Phe254, Arg239, and overall shape complementarity to the shallow, solvent exposed Ebs binding pocket.

We were curious to determine whether or not modification of Ag85C by Ebs had any significant impact on the overall thermal stability of the enzyme. To answer this question, native and covalently modified Ag85C were subjected to thermal shift assays using DSF. DSF is becoming a routine approach to investigate thermal stability of proteins for the optimization of buffers and ligands for crystallization experiments as well as for high through-put drug screening of drug and fragment libraries125-128. Upon modification by Ebs and its derivatives, the thermal stability of Ag85C decreases by 14

°C. The significant loss in thermal stability is not surprising because the covalent modification of Cys209 results in the movement of a major secondary structure and connecting loop. Previous studies examining protein stability of thermophiles concluded proteins with relatively dynamic secondary structures tend to be less thermal stable129. As a result, various hydrogen bonds are disrupted, as highlighted in our previous paper investigating various covalent modifiers of Cys20993. Additionally, factors associated with higher protein thermal stability include buried hydrophobic patches, the presence of numerous salt bridges, and a lack of solvent channels through the protein core structure130, 131. Each of these stabilizing characteristics is lost in Ag85C upon covalent 84

modification by Ebs and its derivatives (Figure 3-14). The decrease in thermal stability lowers the protein melt temperature to near physiological temperature (37 °C), suggesting that the modified enzyme will not only be inhibited enzymatically, but will begin to denature in the host environment during active and dormant stages of bacterial growth2,

132.

3.4.2 Thiophene derivatives

None of the second generation thiophene inhibitors displayed significant inhibition against Ag85C at 100 μM. The original thiophene core displayed weak affinity towards Ag85C; however, this was significantly enhanced by the addition of an arabinose moiety113. The arabinose moiety was crystallographically shown to bind to the trehalose- binding site, with the thiophene substituent residing in the MA binding cleft113.

Unfortunately, the arabinose moiety was removed in the second-generation thiophene library and replaced with other “drug like” substituents116. Library of synthesized thiophene derivatives can be found in Thanna et al. Organic and Biomecular Chemistry,

2016116. Higher concentrations of thiophene derivatives were not tested against Ag85C, due to the fact that thiophenes were found to potently inhibit Pks13 in Mtb115. While none of the compounds in the second-generation thiophene library inhibit Ag85C, a hand full of compounds display stellar inhibitory properties against H37Rv strains of Mtb and were shown to inhibit MA biosynthesis through Pks13 inhibition116.

85

Figure 3-22: Thiophene molecules. Loss of arabinose in second-generation derivatives resulted in loss of Ag85C inhibition levels.

3.4.3 β and γ-lactones

Ag85C was found to be highly selective with regard to lactone inhibition. The majority of lactones screened were 5 member cyclic γ-lactones (Figure 3-23). While, a case for lack of shape complementarity could be argued for andrographolide, and gibberellic acid, the less sterically hindered (4R)-4-Benzyl-3-(5-hexenoyl)dihydrop-

2(3H)-furanone, N-octanoyl-L-Homoserine and N-(β-ketocaproyl)-L-Homoserine should have been able to access the active site of Ag85C. Primarily based on the lack of inhibition observed with N-octanoyl-L-Homoserine and N-(β-ketocaproyl)-L-

Homoserine, γ-lactones most likely are not suitable for Ag85C inhibition. Therefore, β- lactones serve as a more promising covalent warhead towards Ag85s, as displayed by

THL57. However, as observed with clasto-lactacystin β-lactone, the presence of a β- lactone within a molecule is not sufficient enough to inhibit the enzyme. The influence of stereochemistry with respect to substituents of the β-lactone rings is further investigated in chapter 6 through the use of THL stereoderivatives.

86

Figure 3-23: Tested lactones

3.4.5 Identified lead molecules

Screening of Ag85C against two TB active libraries resulted in the identification of potential lead molecules for future development. However, given the levels of observed activity of these molecules towards Ag85C, Ag85s are most likely not the biological target of these TB active compounds. Nonetheless, information was gained on potential scaffolds for future development (Figure 3-24).

Figure 3-24: Structures of identified lead molecules based on drug library screening.

87

Fortuitously, validated compounds of the TB alliance provide a small SAR comparison between compounds 403 and 308, as well as 334 and 801. Based on these molecules, a potential inhibitor of Ag85C is linear and can be separated into two fragments, an aromatic substituent connected with a hydrazide linkage to a thiourea linked hydrophobic or heterocyclic substituent. This proposed arrangement goes hand in hand with the bifurcated active site of Ag85s, being the MA binding site and the trehalose-binding site57. Therefore, these molecules can be modified to specifically target respective portions of the active site, if indeed the thiourea linkage resides near the catalytic Ser124. If this assumption holds true, these molecules may also be modified to be covalent in nature. Previously, thioureas and urea scaffolds have been selectively tuned to covalently inhibit serine esterases133, 134. Indeed, this hypothesis has been proven true through the independent development of triazole containing urea scaffolds by

Abhishek Vartak of the Sucheck group (manuscript in preparation). I have enzymatically shown that these triazole containing compounds covalently inhibit Ag85C; however, I have elected to omit these compounds from this dissertation for the sake of length. A manuscript detailing these compounds is currently being prepared for publication.

3.5 Conclusion

A library of Ebs derivatives was evaluated and enzymatically characterized for

Ag85C inhibition. The lack of an apparent SAR was rationalized through structural and computational studies. Ebs binding was found to be influenced by protein-drug interactions governed by Phe254 and Arg239; additionally, the shallow, surface exposed

Ebs binding site induces steric limitations on the secondary benzyl ring of Ebs. Finally, 88

Ebs modification of Ag85C was biophysically evaluated through the development of a

DSF assay which indicated covalent modification of Cys209 by Ebs significantly reduced protein stability.

A second-generation thiophene library was tested for Ag85C inhibition; however, loss of an arabinose moiety abolished all inhibitory properties with regard to Ag85C.

Excellent in vivo inhibition of Mtb by compounds of the second-generation thiophene library were subsequently attributed to the inhibition of Pks13, not Ag85C. A small set of lactones were screened against Ag85c and suggest that only β-lactones may serve as covalent warheads against the enzyme. Finally, screening Ag85C against two drug libraries resulted in the identification of lead molecules with modest inhibition. A common scaffold was identified among hits and may serve as a potential scaffold for further development.

89

Chapter 4

Structural, enzymatic, and biological characterization of Rv3802

4.1 Background (Portions of text reproduced from Goins et al. the Journal of

Biological Chemistry, 2017)135

4.1.1 rv3802c gene and encoded enzymatic activity

Residing within the gene cluster that encodes proteins responsible for MA and arabinogalactan biosynthesis is the rv3802c gene136. The biological function of rv3802c remains in question; however the gene has been determined to be essential for Mtb viability110, 137, 138. In addition to being required for general viability, the Mavium ortholog of Mtb rv3802c has been shown to be associated with intestinal epithelium invasion139. The rv3802c gene encodes an N-terminal translocation signal sequence that includes a predicted transmembrane region140. Rv3802 was found only in cell wall extracts upon cellular lysis and maintaining a molecular mass indicative of an uncleaved protein, suggesting the enzyme remains anchored in the cytoplasmic membrane140. Due to sequence similarities with cutinase-like enzymes, the rv3802c gene product was named

CULP6, cutinase-like protein 6, despite lacking cutinase activity140. However, Rv3802 has been shown to possess both thioesterase and Phospholipase A (PLA) activity utilizing residues S175-H299-D268 as the catalytic triad136, 140. Substrates shown susceptible to 90

acyl-hydrolysis by Rv3802 include: polyoxyethylene (20) sorbitan monolaurate, p- nitrophenyl butyrate, laurate, myristate, palmitate, and sterate (estase activity); phosphatidylserine, phosphatidylcholine, and phosphatidylethanolamine (PLA activity)136, 140.

4.1.2 Msmeg ortholog structure

The structure of the Msmeg ortholog, MSMEG_6394, has been solved, revealing that the enzyme possesses an α/β-hydrolase fold (Figure 4-1)138. The apo structure presented two intact disulfides and a small cavity leading to the identified active site residues138. However, no sizable substrate binding site was observed138.

Figure 4-1: Structure of the Rv3802 ortholog MSMEG-6394. A) MSMEG-6394 has an α/β-hydrolase fold, catalytic residues highlighted in yellow. B) small pocket leading to catalytic residues (PDB: 3AJA)138.

4.1.3 Rv3802 biological function

Mtb Rv3802 shares a high level of sequence identity with a variety of

Mycobacterium homologs (68.8% sequence identity between encoded Mtb Rv3802 and

MSMEG_6394) in addition to the Corynebacterium glutamicum ortholog NCgl2775137.

91

Upon heat-induced cell stress, NCgl2775 was shown to play an essential role in the modulation of the mycomembrane composition137. As a result, the total mycolate content increased while glycerophospholipid content decreased137. This elevated mycolate:glycerophospholipid ratio was also observed when both MSMEG_6394 and

NCgl2775 were overexpressed under physiological conditions137. Meniche et al. therefore proposed naming this class of enzymes the “envelope lipids regulation factor” or the ElrF family137. Additionally, a conditional knock out of MSMEG_6394, resulted in Msmeg cells having an the elongated, jagged cell morphology138. Based on these findings Crellin and colleagues proposed that Rv3802 play an essential roll in mycomembrane maintenance and potentially cellular division through the decomposition of glycerophospholipids138. Based on these previous studies, we found Rv3802 to be an intriguing target for biochemical characterization and inhibitor development. Therefore, we pursued the structure of Rv3802 to gain insights into biological function and as a basis for structure guided drug development towards this unexplored, essential enzyme.

4.2 Methods (Portions of text reproduced from Goins et al. the Journal of Biological

Chmiestry, 2017)135

4.2.1 Wild type and mutant cloning

An E. coli codon-optimized, synthetic gene encoding the periplasmic domain of

Mtb Rv3802 was purchased from Integrated DNA Technologies. The bases of 5’ –

CGCTGTTCCAGGGACCT - 3’ and 5’ – GCGTCCGGATCCGAA – 3’ were added to the 5’ and 3’ ends of the gene, respectively. The resulting gene was PCR amplified using primers of identical sequences (5’ and 3’ amplify primers, Table 4.1) to those added to 92

the gene and inserted into a pET32-derived plasmid linearized with PshA1 using a

Gibson Assembly™ Master Mix. The protein expression construct produces protein with the following sequence: Met - His6 – Ser2 – Gly - Cut Site - Rv3802c. The

N132C/N288C mutant was cloned in pieces overlapping the site of mutation using the

WT construct as a template. Using the 5’ amplify primer and the N132C reverse primer generated the first fragment. The N132 Forwards and N288C Reverse primers were used to generate the middle fragment. The N288C Forwards and the 3’ amplify primers were used to generate the end fragment (Primer sequences in Table 4.1). The full gene was constructed by amplifying the front and middle fragments together, followed by the middle and end fragments. The two resulting fragments were then amplified together to produce a full-length mutant gene and inserted into an expression plasmid in an identical fashion to WT. Mutants of the catalytic residues (S175A, H299A, and D268N) were again generated in fragments using the 5’ and 3’ amplify primers in combination to primers designed to incorporate the mutation of interest. The 5’ amplify primer was used with respective “Reverse” primer and 3’ amplify primer was used with respective

“Forward” primer to generate two gene fragments for each mutant (Primers listed in table

4.1). The halves were amplified together using the amplified fragments with the 5’ and 3’ amplify primers and inserted into the same pET32 plasmid as WT. WT and mutant plasmids were sequenced to confirm the desired DNA sequence.

93

Table 4.1: Primers used for WT and mutant cloning Primer Name DNA Sequence 5’ Amplify 5’ – CGCTGTTCCAGGGACCT - 3’ 3’ Amplify 5’ – GCGTCCGGATCCGAA – 3’ N132C Reverse 5’ - TCAGAGGGCAGTGGAACTG - 3’ N132C Forward 5’ - CAGTTCCACTGCCCTCTGA - 3’ N288C Reverse 5’ - GCAAGGGTGCAAAGAGTCGTA - 3’ N288C Forward 5’ - TACGACTCTTTGCACCCTTGC - 3’ S175A Reverse 5’ - CTC CTT GCG CGA ACC CGA – 3’ S175A Forward 5’ - TCG GGT TCG CGC AAG GAG – 3’ H299A Reverse 5’ - ACA TTG CCG CAA CCG GTT – 3’ H299R Forward 5’ - AAC CGG TTG CGG CAA TGT – 3’ D268N Reverse 5’ - CCG CGC AAA TTA AGT TTC CTT GTG – 3’ D268N Forward 5’ - CAC AAG GAA ACT TAA TTT GCG CGG – 3’

4.2.2 Protein expression and purification

Protein expression and purification were adapted and modified from a previously published methodology140. WT and mutant proteins were expressed and purified in an identical fashion. Chemically competent T7 Express E. coli cells (New England Biolabs) were transformed with plasmid encoding WT or mutant protein. Subsequent cultures were grown at 37 °C in Lysogeny Broth media containing carbenicillin. Protein expression was induced for 3 hours at 37 °C, following the addition of 1 mM IPTG once culture density reached an OD600nm of 0.6. Induced cells were pelleted and resuspended in a 50 mM TRIS pH 7.5, 300 mM NaCl buffer and placed at -80 °C for storage.

Induced cells were thawed in a warm bath, followed by the addition of DNase 1 and Lysozyme, and allowed to incubate on ice for half an hour. Resuspended cells were then sonicated and the crude lysate pelleted at 18,515 x g for 40 minutes at 4 °C.

Supernatant was discarded and the pellet was resuspended in 50 mM TRIS pH 7.5, 3 M

NaCl buffer. Resuspended cells were again pelleted at 18,515 x g for 40 minutes at 4 °C,

94

with the supernatant again discarded. The resulting pelleted inclusion bodies containing

Rv3802 were solubilized in a 50 mM TRIS pH 7.5 and 4 M Guanidine HCl (GuHCl) and pelleted at 18,515 x g for 40 minutes at 4 °C. The supernatant containing soluble

Rv3802c was loaded onto an equilibrated 5 mL HiTrap Talon Crude cobalt column (GE

Healthcare) and was washed for 15 column volumes with 50 mM TRIS pH 7.5, 4 M

GuHCl. Protein was eluted over a 20 column-volume gradient with increasing imidazole concentration. Rv3802 was refolded through the removal of GuHCl as a result of extensive dialysis using a 50 mM TRIS pH 7.5, 300 mM NaCl buffer. Refolded protein was centrifuged at 18,515 x g for 1 hr at 4 °C to remove any residual insoluble protein.

Soluble, pure protein was dialyzed into either 100 mM HEPES pH 7.5 for crystallization or 50 mM NaPO4 pH 7.5 for cysteine quantification, 4MH and PLA assays.

4.2.3 Activity assay development

4.2.3.1 Resorufin butyrate assay

Crellin and colleagues developed an absorbance based assay using p-nitrophenyl butyrate to monitor the hydrolytic activity of Rv3802138. However, Rv3802 has low affinity to p-nitrophenyl butyrate, Km of 5.2 mM; additionally, absorbance based assays are inherently less sensitive than fluorescence based assays138. Therefore, we sought to develop a fluorescence-based assay using a more specific substrate to the enzyme. The first generation assay utilized the same flourogenic reporter as the Ag85C assay, resorufin butyrate (RfB). RfB was dissolved in DMSO to make a 15 mM stock. The assay consisted simply of 1 μM Rv3802 in 50 mM sodium phosphate pH 7.5 buffer with reactions initiated upon titration of RfB. All fluorescent reads were conducted at 37 °C 95

using λex = 500 nm and λem = 590 nm on a Synergy H4 Plate Read (Biotek). For

Michaelis-Meneten kinetic determination, RfB was serial diluted from 15 mM to 0.5 mM and titrated in triplicate into reactions with and without enzyme present. Relative fluorescent units (RFU) was converted to product concentration using a resorufin standard curve. Background was subtracted from reactions with enzyme, Michaelis-

Menten parameters were determined using PRISM 7. Z’ value for RfB assay was determined using 24 reactions with and without enzyme present, reactions consisted of 1

μM enzyme and 100 μM RfB in 50 mM sodium phosphate pH 7.5 buffer. Z’ was calculated using: Z' = 1-(3*stdev(a)+3*stdev(b))/(| AvgRate(a)-AvgRate(b) |) where a = reactions with enzyme and b = reactions with no enzyme123.

4.2.3.2 4-methylumbelliferyl heptanoate assay

The second generation Rv3802 fluorescence-based assay uses 4MH (Sigma

Aldrich) as the fluorogenic reporter. In the presence of enzyme, the heptyl chain is hydrolyzed from 4MH, producing the now fluorescent 4-methylumbelliferone molecule and heptanoic acid. All assays were performed in triplicate in a 50 mM sodium phosphate pH 7.5 buffer. All fluorescent reads were conducted at 37 °C using λex = 360 nm and λem

= 450 nm on a Synergy H4 Plate Read (Biotek). The methods for Michaelis-Menten kinetics for 4MH are detailed in section 4.2.4.7. To determine if the second-generation assay was amenable to high throughput drug screening, the Z’ value was determined. 24 reactions containing enzyme and 24 reactions not containing enzyme were used for Z’ calculation. 50 nM of WT enzyme was used with 75 μM 4MH (Based on Michaelis-

Menten values). Z’ was calculated using: Z' = 1-(3*stdev(a)+3*stdev(b))/(| AvgRate(a)-

AvgRate(b) |) where a = reactions with enzyme and b = reactions with no enzyme123. 96

4.2.4 Crystallization and structure determination

Purified Rv3802 was concentrated to 4.5 mg/mL for crystallization. Screening of

Rv3802 against the Index Screen (Hampton Research) using the hanging drop method resulted in the formation of a single orthorhombic crystal. The crystal formed in a 1 to 1 ratio of well solution (0.1 M BIS-TRIS pH 6.5, 25 % w/v PEG 3350) to protein, requiring

1 month to grow with incubation at 16 °C. The resulting crystal was looped and flash- cooled in liquid nitrogen. To date, this crystal has not been able to be reproduced.

Extensive efforts to crystalize WT Rv3802 and mutants have been conducted to no success. Attempts to crystalize the protein included varying protein concentration, complex with inhibitors, removal of the His tag, varying protein buffers, and screening against diluted well solutions. X-ray diffraction data on the lone crystal were collected using synchrotron radiation at the Advanced Photon Source, LS-CAT beamline F, at

Argonne National Laboratory. A single crystal was used for this structure.

Diffraction data were indexed and scaled in a P212121 space group using

HKL200087. The phase solution with 2 molecules in the asymmetric unit came from molecular replacement (PHASR-MR) using a model generated with SWISS-MODEL that was based on the previously solved MSMEG_6394 structure (PDB: 3AJA)88, 141.

Atoms lacking 2Fo-Fc density were deleted and the resulting model was subjected to rigid body and simulated annealing refinements (PHENIX Refine)88. Deleted atoms were built back into corresponding Fo-Fc and 2Fo-Fc difference density maps using COOT89.

The progressing model was subjected to rounds of XYZ coordinates, real-space, occupancies, and individual B-factors refinements (PHENIX-Refine) with manual modeling in COOT88, 89. The four molecules of PEG were added using LigandFit and 97

subjected to model refinement142. Model refinement was complete once a final

Rwork/Rfree of 0.1582/0.1830 was achieved. The 0.4 % Ramachandran outlier is due to

V237 and is warranted given the strong electron density for all atoms of the residue in both molecules within the asymmetric unit. MolProbity was utilized to validate the structure92.

4.2.5 Structural analysis and sequence conservation

X-ray crystal structure alignments, measurements, and subsequent analysis were performed using PyMOL96. Sequence alignments were performed with ClustalW and figures were generated through the use of T-COFFEE and BOXSHADE94, 143. Jpred4 was used for secondary structure prediction144. Transmembrane regions were predicted using the TMHMM Server145.

4.2.6 Cysteine quantification

To ensure proper disulfide formation of the engineered N132C/N288C mutant, thiols present in WT and mutant were quantified and compared using fluorescent labeling with a non-reversible thiol reactive probe. Respective enzymes were compared in non- reduced and reduced states. 5 μM of respective enzyme in 50 mM sodium phosphate buffer pH 7.5 was incubated with 0.3 mM TCEP (reduced) or equivalent volume buffer

(non-reduced) for 30 minutes at 37 °C (200 μL reaction volumes). Samples were titrated with 240 μM monochlorobimane (100 mM stock, DMSO), mixed, and incubated at 37 °C for 30 minutes. A standard curve to quantify free thiols was generated by serial diluting

L-cysteine in the presence of 0.3 mM TCEP, reduced for a half hour at 37 °C, and modified with 240 μM monochlorobimane for 30 minutes at 37 °C. Fluorescent reads

98

were conducted in triplicates using λex = 390 nm and λem = 490 nm on a Synergy H4 Plate

Read (Biotek).

4.2.7 WT and Mutant activity

To assess activity levels of WT and N132C/N288C mutant Rv3802, Michaelis-

Menten kinetics were determined for 4MH. Briefly, 50 nM of WT Rv3802c and 75 nM

N132C/N288C mutant were used with varying concentrations of serial diluted 4MH (20 mM DMSO stock). All fluorescent reads were conducted at 37 °C using λex = 360 nm and

λem = 450 nm on a Synergy H4 Plate Read (Biotek). Background fluorescence as a consequence of 4MH hydrolysis was subtracted and relative fluorescence units were converted to concentration of standard curve. Reaction rates and Michaelis-Menten kinetic parameters were determined using PRISM 7. To determine percent activity for catalytic mutants, 4MH was titrated into 50 nM WT and respective mutants (final concentration of 4MH was 75 μM). Linear rates were assessed in PRISM with background subtracted. % Activity was determined by dividing observed rates of mutant by WT enzyme and multiplying by 100.

4.2.8 Phospholipase A assay

PLA activity was assessed using a modified version of a previous established

Rv3802-PIM hydrolysis experiment136. Briefly, soybean PI (Avanti® Polar Lipids Inc.) was resuspended in 50 mM sodium phosphate pH 7.5 forming a 2 mg/ml suspension of

PI. Reactions consisted of 300 μg of PI suspension, 50 μg of respective enzyme or equivalent volume of buffer, and were brought to a final volume of 400 μL with 50 mM sodium phosphate pH 7.5 buffer. Reactions were incubated for 1 hour at 37 °C, being mixed every 10 minutes, and quenched with 200 μL of 50:50:0.3 CHCl3:CH3OH:HCl. 99

The lower organic phase was extracted, spotted onto a silica TLC plate, and resolved using a mobile phase of 80:20:2 CHCl3:CH3OH:NH4OH. Inositol-containing species were visualized by treating the TLC plate with 5 % H2SO4 in methanol and charring.

4.2.9 Evaluation of acyl-arabinogalactan substrates

Acylated arabinose compounds were obtained from the Todd Lowary group at the

University of Alberta. Solid compounds were suspended in a 50 mM sodium phosphate pH 7.5 buffer to making a 1 mg/ml suspension. For hydrolysis experiments, 100 μg of respective acyl arabinose was added to 50 μg of Rv3802 in 50 mM sodium phosphate.

The reaction mixture was incubated at 37 °C for 2 hours, being mixed every 15 minutes and quenched upon the addition of a 1:1 solution of CHCl3:CH3:OH. The lower organic phase was extracted and spotted on a silica TLC plate and resolved using a mobile phase of 90:10 CHCl3:CH3OH. Spots were developed using 5 % H2SO4 in methanol and charring.

4.3 Results (Portions of text reproduced from Goins et al. the Journal of Biological

Chemistry, 2017)135

4.3.1 Molecular cloning, expression, and purification

DNA sequencing of constructs of WT and Rv3802 mutants confirmed the presence of either the correct WT gene or desired mutation within the WT gene. All mutants were successfully purified in an identical fashion to WT with similar yield; however, expression levels of Rv3802 H299A were noticeably higher than WT. It should be noted that all WT and mutant forms were refolded through dialysis and were sensitive to concentration within the dialysis bag. Therefore, protein must be at a low 100

concentration (~ 0.25 mg/ml) within the dialysis bag for refolding or significant precipitation is observed. The developed expression and purification system for Rv3802 and mutants resulted in workable yields of highly pure proteins (2-3 ml of protein at 5 mg/ml for a 2 L culture). SDS-PAGE of WT Rv3802 and mutants are given in Figure 4-

2. While the D269N mutant was successfully cloned, the protein was never expressed or purified for studies.

Figure 4-2: SDS-PAGE of WT and Rv3802 mutants, expected molecular weight of ~29.9 kDa. A) Non-reducing SDS-PAGE of WT (Lane 1), N132C/N288C mutant (Lane 2), ladder lane 3. B) Reducing SDS-PAGE of S175A (Lane 1), H299A (Lane 2), and ladder lane 3. Panel A was adapted from Goins et al. the Journal of Biological Chemistry, 2017135.

4.3.2 Activity assays

4.3.2.1 Resorufin butyrate assay

Rv3802 was shown to successfully catalyze the hydrolysis of RfB; however, the best signal to noise using 100 μM RfB required 1 μM of enzyme (Figure 4-3A). In general the observed signal to noise is approximately 3:1 at these concentrations. The Km

101

-1 of RfB was determined to be 24.32 ± 3.91 μM and Vmax = 0.074 ± 0.004 μM*Min

(Figure 4-3B). The Z’ value was determined to be 0.52 based on the (+) Avg = 1520

RFU/Min, the (-) Avg = 519 RFU/Min, (+) STD = 115 RFU/Min, (-) STD = 48 RFU/min

(RFU = relative fluorescence units).

Figure 4-3: Rv3802 RfB assay. A) Typical reaction progress curves observed with 1 μM Rv3802 and 100 μM RfB. RfB has significant levels of hydrolysis in water. B) Michaelis-Menten curve for RfB.

4.3.2.2 4-methylumbelliferyl heptanoate assay

Rv3802 was found to readily hydrolyze 4MH, requiring substantially less enzyme than the RfB assay. 50 nM of WT enzyme was used with 75 μM 4MH to determine a Z’ value for the assay (Figure 4-4). Obtained values for Z’ determination are as follows: (+) Avg = 9560

RFU/Min, the (-) Avg = 139.1 RFU/Min, (+) STD = 482 RFU/Min, (-) STD = 26.6

RFU/min. Z’ for the 4MH assay is therefore 0.84. Using these concentrations of enzyme and substrate, the reaction is essentially over within 8 minutes, as the plate reader used to obtain fluorescent reads max out at 100,000 RFU. The signal to noise for these conditions is ~ 68. Michaelis-Menten values for 4MH are given in section 4.3.7.

102

Figure 4-4: Z’ determination for 4MH assay (Figure reproduced from Goins et al. the Journal of Biological Chemistry, 2017)135.

4.3.3 Rv3802 structure and lipid binding site

Mtb Rv3802 crystals diffracted to 1.7 Å, having a space group of P212121 with two molecules present in the asymmetric unit having a Cα RMS difference of 0.14 Å (X- ray diffraction and model refinement statistics are given in Table 4.2). Molecule A was used for structural analyses if not otherwise noted.

103

Table 4.2: X-ray data collection and refinement statistics, (*) denotes values in the outer shell. Rv3802-PEG (PDB: 5W95) Data collection Space group P 21 21 21 Cell dimens a, b, c (Å) 57.19, 91.30, 101.39 α, β, γ ) 90, 90, 90 Resolution (Å) 44.32 - 1.72 (1.78 - 1.72) * Rmerge 0.09 (0.51) I / σI 23.23 (6.08) Completeness (%) 99.63 (96.24) Redundancy 14.8 (14.8) Refinement Resolution (Å) 44.32 - 1.72 No. reflections 56631 Rwork / Rfree 0.1582/0.1830 No. atoms Protein 3921 Ligand/ion 64 Water 648 B-factors (Å2) Protein 17.70 Ligand/ion 33.10 Water 28.00 R.m.s. deviations Bond lengths (Å) 0.007 Bond angles () 1.062 Ramachandran (%) Favored 98.0 Outliers 0.4

As predicted by homology to the Msmeg ortholog, Mtb Rv3802 was found to have

135 an α/β hydrolase fold (Figure 4-5) . Rv3802 has 10 helices (9 α-helices and one 310 helix) and a single, central β-sheet comprised of 5 parallel β-strands (Figure 4-5). The nucleophilic S175 is located on the aptly named nucleophilic elbow connecting β3-strand to α3-helix. The remaining catalytic triad residues, H299 and D268, are located on α9- helix and a connecting loop between β5-strand and α8-helix, respectively (Figure 4-5). 104

Two intact disulfides are present in the Rv3802 structure: C72 - C164 and C264 - C271

(Figure 4-5). These conserved disulfides are also present in the MSMEG_6394 structure138. During the early stages of model refinements, two regions of linear difference density converging at the active site were observed. This non-protein electron density is interpreted as two molecules of PEG found bound to each of the two protein molecules in the asymmetric unit. Difference density maps for PEG molecules in protein molecule A can be found in Figure 4-6. PEG 3350 (average molecular weight is 3350 kDa or ~75 repeating monomer units) was present in the crystallization solution; however, the bound PEG polymers consisted of only 5 ethylene glycol monomer units.

Figure 4-5: The X-ray crystal structure of Mtb Rv3802 with PEG bound. The enzyme has an α/β-hydrolase fold with the catalytic triad residues of S176, H299, and D2269 shown in addition to the two conserved disulfides, C72-C164 and C264-C271. The two molecules of PEG found bound to Rv3802 are shown in spheres. Atom colors are CPK sans carbon, which is white for protein and wheat for the two PEG molecules. Helices and β-strands are numbered and colored in increasing respective color intensity from N- term to C-term. Figure reproduced from Goins et al. the Journal of Biological Chemistry, 2017135.

Due to the presence of PEG molecules within Rv3802 and PEG resembling the fatty acid chains of lipids, the lipid-binding site of Rv3802 and presumptive orthologs

105

have been structurally identified (Figure 4-6). The fatty acid portion of the lipid-binding site begins at the active site, with the terminal oxygen of PEG1 positioned within hydrogen bonding distance to both S175 and H299, 3.2 and 3.8 Å respectively. The terminal oxygen of the PEG2 resides 6.1 Å away from the terminal oxygen of PEG1 and is positioned between the side chains of Y142 and I270. Both PEG molecules extend away from the active site in a linear fashion, flanked on both sides by mostly aliphatic amino acid side chains that are positioned on helices α6 and α8 as well as various loop regions (Figure 4-6A). Without ligand present, the aliphatic side chains involved in PEG binding form a hydrophobic core of van der Waals interactions as seen in the previously- solved apo MSMEG_6394 structure (Figure 4-6B)138. While Rv3802 and MSMEG_6394 have a relatively high level of sequence identity for all encoded residues, 68.8%, the residues that make up the identified fatty acid binding region possess a higher level of conservation, 76.2% between Mtb and Msmeg specifically, as well as a higher level of conservation with other mycobacterial orthologs (Figure 4-6C).

106

Figure 4-6: The identified lipid tail (fatty acid) binding site is highly conserved and hydrophobic. (A) Fo-Fc likelihood weighted omit map contoured to 3 σ for salmon colored PEG molecules. Amino acid side chains within 4 Å of the PEG molecules are shown in cyan, active site residues are orange with other non-carbon atoms in CPK color. These residues comprise the lipid tail binding site. (B) Without ligand present, the amino acid side chains of the lipid tail binding site form a wall of hydrophobic interactions as seen in the apo MSMEG_6394 structure (PDB: 3AJA)138. (C) Sequence alignment of the fatty acid binding region from Mycobacteria. Residues of differing sequence are shaded, starred residues are shown in panels A and B, while a red underline corresponds to secondary structure location. Figure reproduced from Goins et al. the Journal of Biological Chemistry, 2017135.

4.3.4 Structural movement upon lipid binding

As stated above, Rv3802 and the previously solved apo MSMEG_6394 ortholog share a common core scaffold with two disulfides present. However, a noticeable difference between the Mtb and Msmeg structures exists between residues 279-298

(Figure 4-7A). In order to accommodate the two bound PEG molecules, α8 helix and flanking residues are shifted away from the adjacent loop in the Rv3802 structure.

107

The N-terminal region of α8 helix is extended outward 8.8 Å upon PEG binding, when compared to the apo Msmeg ortholog (Figure 4-7B). The opening of α8-helix alters a connecting loop to the neighboring 310-7 helix; however, this does not affect 310-7 helix positioning. The loop connecting α8 to the upstream catalytic H299 is unstructured. Due to poor difference density, this initially modeled region was ultimately removed in the deposited structure. Moreover, it is also unresolved in the Msmeg ortholog structure. This structurally dynamic region is the hinge point for α8 helix movement, being comprised of mostly alanine and glycine residues in both enzymes.

Due to a high sequence similarity in the active site between Rv3802 and its ortholog, a direct active site comparison between the two conformations is feasible.

Throughout the remainder of this chapter, analogous residue numbers will be given with the Rv3802 number first and the MSMEG_6394 number second (e.g. T83/84). The top of

α8 helix abuts the active site; despite significant movement between the open and closed states of α8 helix upon PEG binding, little movement is observed in the residues of the active site (Figure 4-7C). However, the nucleophilic S176 of MSMEG_6394 is positioned toward the presumed oxyanion hole consisting of the backbone amides of

T83/84 and N176/177, while the nucleophilic S175 of Rv3802 is hydrogen bonded to the catalytically relevant H299. Uncertainty remains as to whether this difference is a function of the modest 2.9 Å resolution of the MSMEG_6394 structure or represents a true structural difference between the open, substrate-bound form and the closed, substrate-free form.

108

Figure 4-7: Superposition of Mtb Rv3802 (gray) and ortholog MSMEG_6394 (purple, PDB: 3AJA), residue numbering is given as Rv3802/MSMEG_6394138. (A) The α8-helix and surrounding residues are found to be structurally dynamic upon substrate binding. (B) The N-terminus of α8-helix moves ~9 Å between the open and closed states, the region shown spans from residues 265/266 to 302/303. (C) Despite dynamic movement near the active site, catalytic and surrounding residues are positioned similarly between the open and closed states. The side chain of N176 in Rv3802 was modeled to give the highest likely conformation for hydrogen bond formation, resulting in the discrepancy with the previously solved MSMEG_6394 N177138. Figure reproduced from Goins et al. the Journal of Biological Chemistry, 2017135.

4.3.5 Substrate organization and mechanism of catalysis

When α8 helix is in the open conformation, a sizable substrate binding site is apparent (Figure 4-8A). However, in the closed state visualized in the Msmeg ortholog, only a small cavity leading to the nucleophilic serine is observed138. Therefore, for a large substrate possessing hydrophobic, aliphatic character to bind, α8 helix must move to expose the substrate binding site. It is expected that initial substrate binding positions the scissile ester/thioester bond in proximity to S175, affording bifurcation of the active site into two distinct areas that accommodate either the hydrophobic tail or the hydrophilic

109

head group. With the reasonable expectation that the PEG molecules bind within the lipid tail-binding site, the lipid head-binding site becomes apparent (Figure 4-8B). The head group-binding site is unchanged from that observed in the closed state of Msmeg ortholog, being the observed small cavity described previously138. This site is comprised of: the hydrophobic side chains of W84, L102, F174, A292, A300, the negatively charged residues or carbonyl backbones of T83, E85, A300, and the positively charged side chain of K100. These residues are fully conserved between Rv3802 and the Msmeg ortholog aside from K100 and A300, these two residues are L and N in Msmeg, respectively.

Figure 4-8: Structure-based insights on substrate binding sites and arrangement for catalysis. (A) Identification of the Rv3802 lipid tail (fatty acid) binding site results in the identification of the lipid head binding site that is linear to that of the lipid tail. Additionally, a noticeable solvent channel leads to the active site directly intersecting the lipid tail and lipid head sites in a perpendicular manner. PEG molecules are wheat and water molecules are red spheres, cavity surfaces shown (B) Residues comprising the proposed lipid head binding site adjacent to the catalytic triad are depicted, catalytic residues are in orange. Figure reproduced from Goins et al. the Journal of Biological Chemistry, 2017135.

Perpendicular to and intersecting the substrate binding site is a solvent channel leading to the active site (Figure 4-8A). In both protein molecules within the asymmetric unit, ordered water molecules are found alongside H299, opposite of the lipid head group-binding site. In molecule A, the terminal PEG1 oxygen is 3.2 Å from the S175

110

hydroxyl, while the S175 hydroxyl is 3.0 Å from the ε nitrogen of H299 (Figure 4-9A).

Additionally, a water molecule is found 4.2 Å away from H299. In molecule B, the terminal PEG1 oxygen is pointing towards the plausible oxyanion hole, 3.4 Å equidistant to both backbone amides of T83 and N176. A water molecule is now found within hydrogen bonding distance, 2.9 Å to the ε nitrogen of H299 at (Figure 4-9B). It should be noted that while not depicted in the figures, solvent molecules are positioned in a similar fashion in the lipid head binding sites of both protein molecules A and B.

Figure 4-9: Positioning of PEG and water in the active sites of both protein molecules in the asymmetric unit (Fo-Fc likelihood weighted omit map contoured to 4 σ). (A) The terminal oxygen of PEG, denoted by *, is within hydrogen bonding distance of the catalytic S175 in protein molecule A. Additionally, a water molecule is 4.2 Å away from H299, which is positioned towards S175. This configuration resembles initial substrate binding and the activation of S175 for nucleophilic attack on substrate. (B). In protein molecule B, the terminal oxygen of PEG is pointed towards the presumptive oxyanion hole, while a water molecule is within hydrogen bonding distance to H299. This configuration resembles water activation by the enzyme to hydrolyze the acyl-enzyme intermediate. Figure reproduced from Goins et al. the Journal of Biological Chemistry, 2017135.

4.3.6 Catalytic mutant activity

We were interested in generating a chemically inert form of Rv3802 for crystallization purposes. Interestingly, mutation of the nucleophilic serine to an alanine resulted only in a 48.4 ± 0.5 % decrease in enzymatic activity compared to WT.

However, mutation of the catalytic histidine resulted in complete loss of enzymatic activity (Figure 4-10). The D269N mutant was never expressed and purified for enzymatic studies.

111

Figure 4-10: enzymatic activity of catalytic mutants. No detectable activity was observed for H299A mutant.

4.3.7 Protein dynamics required for catalysis

To investigate the effects α8 movement has on catalysis and to validate the structural observations, Mtb Rv3802 was mutated to limit α8 movement. In the closed state, N133 and E289 of the Msmeg ortholog are ~5 Å apart; therefore, we mutated the corresponding residues in Mtb Rv3802, N132 and N288, to cysteine residues to form an engineered disulfide (Figure 4-11A). The resulting N132C/N288C mutant would therefore have limited α8 helical movement due to covalent tethering.

The 4MH assay was used to compare the enzymatic activities of WT and covalently tethered forms of Rv3802. Michaelis-Menten kinetics were determined for both WT and N132C/N288C mutant using 4MH, Figure 4-11B (WT: Km = 19.88 ± 1.18

-1 -1 μM, Vmax = 0.51 ± 0.01 μMmin , kcat = 10.05 ± 0.17 Min ; N132C/N288C: Km =

-1 -1 37.21 ± 3.11 μM, Vmax = 0.38 ± 0.01 μMmin , kcat = 5.04 ± 0.15 Min ). The affinity of the N132C/N288C mutant for 4MH decreases by 53 % and results in a twofold decrease in the kcat when compared to WT enzyme.

112

The residual hydrolase activity of the N132C/N288C mutant towards 4MH indicated that N132C/N288C mutant retained sufficient dynamics or space near the active site for 4MH to bind, despite the hindered helical movement. While 4MH has a 7-carbon alkyl chain, the substrate is significantly shorter than phosphatidyl-based substrates that could occupy the entire lipid tail binding site. Therefore, PLA activity of WT and

N132C/N288C mutant was tested using PI as a substrate. Following an hour of incubation with PI, hydrolyzed product indicative of PLA activity was observed for WT

(Figure 4-11C, lane 1). However, no PLA activity was observed for the N132C/N288C mutant (Figure 4-11C, lane 2).

Figure 4-11: Influence of α8-helix movement on enzymatic hydrolysis (A) In the closed state, α8-helix blocks the lipid tail (fatty acid) portion of the lipid-binding site. Based on the proximity of N133 and E289 in the MSMEG_6394 closed state, the corresponding amino acids in Rv3802, N132 and N288, were both mutated to cysteine to create an engineered disulfide and limit α8 helix movement. Surface rendering and gray cartoon is Rv3802 with MSMEG_6394 depicted in purple (MSMEG_6394 PDB: 3AJA)138. (B) Michaelis-Menten kinetics of WT compared to N132C/N288C mutant using 4MH as the substrate. Both the affinity for 4MH and the kcat decreases by half as a result of hindered helical movement. (C) TLC plate of reaction products when the much larger PI is used as the substrate. A shift indicative of PLA1 or PLA2 activity is observed with WT Rv3802, reaction 1 (Rf value of 0.19). However, no acyl hydrolysis is observed with the disulfide mutant, reaction 2, or when enzyme is lacking, reaction 3. Figure reproduced from Goins et al. the Journal of Biological Chemistry, 2017135.

113

4.3.8 Cysteine quantification

To ensure the loss in enzymatic activity of the N132C/N288C mutant was due to the engineered disulfide, multiple controls were performed to investigate proper disulfide bond formation. Both WT and N132C/N288C Rv3802 run as single monomeric bands on a non-reducing SDS-PAGE gel, indicating the lack of intermolecular disulfide bonds

(Figure 4-11A). Additionally, the amount of free thiols present in both WT and

N132C/N288C Rv3802 was quantified under non-reducing and reducing conditions using monochlorobimane. Non-reduced samples had levels of fluorescence equal to background, indicating that no free thiols are present in both WT and N132C/N288C mutant (Figure 4-12B). The resulting ratio of quantified cysteines in reduced WT to

N132C/N288C Rv3802 was 0.68 ± 0.04, consistent with the 4:6 ratio of cysteines present in WT and N132C/N288C mutant, suggesting that all cysteine residues were forming disulfides and were indeed intramolecular (Figure 4-11C).

Figure 4-12: Non-reducing SDS-PAGE gel of WT and Cys mutant and free thiol quantification. (A) Non-reducing SDS-PAGE of WT and Mutant Rv3802 (MW ~29 kDa). Lane 1: WT, lane 2: Cys Mutant, Lane 3: Ladder. (B) Non-reduced WT and Cys mutant have background levels of fluorescence (RFU = relative fluorescence unit) upon monochlorobimane medication (MCB), mean and SEM shown with triplicate data. (C) Free thiol quantification of reduced WT and Cys mutant were determined to be 21.3 ± 1.6 μM and 31.2 ± 1.3 μM, respectively. The observed 2:3 ratio of free thiol is consistent with the 4 cysteines in WT and 6 in the Cys mutant when both enzymes were set at 5 μM.

114

Mean and SEM shown with triplicate data (D) standard curve of reduced L-cysteine and MCB, triplicate data shown. Figure reproduced from Goins et al. the Journal of Biological Chemistry, 2017135.

4.3.9 Evaluation of acyl-arabinose substrates

To determine if Rv3802 could hydrolyze MA from the mAG, smaller acylated arabinose compounds were tested in place of the highly insoluble mAG. Following two hours of incubation with enzyme, no evidence of esterase activity towards acyl-arabinose compounds were observed (Figure 4-13).

Figure 4-13: Evaluation of Rv3802 hydrolytic activity towards acylated arabinose compounds. A) Compounds tested, acyl chains were of 15-carbon length. B) Lane 1: tri- arabinose, no enzyme; lane 2: tri-arabinose, enzyme; lane 3: penta-arabinose, no enzyme; lane 4: penta-arabinose, enzyme.

4.4 Discussion (Portions of text reproduced from Goins et al. the Journal of

Biological Chemistry, 2017)135

4.4.1 1st and 2nd generation Rv3802 assay comparison

The first generation fluorescence based assay-using RfB as substrate required a high concentration of enzyme and a low signal to noise due to the rapid hydrolysis of RfB

115

in the presence of water. Therefore, comparatively speaking, this assay did not substantially enhance the way Rv3802 activity was detected from the previous absorbance based assay using p-nitrophenyl butyrate138. While, the affinity towards RfB is significantly better than p-nitrophenyl butyrate resulting in less required substrate, the poor signal to noise is a severe limitation. Despite this problem, the RfB assay is amenable to high throughput drug screening with a Z’ value of 0.52; however, a higher number of false positives may be observed due to the poor signal to noise123.

The second generation Rv3802 assay improved upon the downfalls of the RfB assay. 4MH proved to be a far superior fluorogenic reporter for Rv3802 esterase activity when compared to RfB. Briefly, 20 times less enzyme is required, the signal to noise is

~22 fold higher resulting in a higher Z’ value (Table 4.3). Therefore, the 4MH assay met our goal of developing a superior assay for Rv3802 activity. The only concern with the

4MH assay is the wavelengths used for fluorophore excitation and emission, as these lower wavelengths occasionally overlap with the conjugated ring systems of drug molecules.

Table 4.3: Comparison of 1st and 2nd generation Rv3802 fluorescence-based assays Property RfB 4MH Enzyme concentration 1 μM 50 nM Substrate affinity (Km) 24.32 ± 3.91 μM 19.88 ± 1.18 μM Signal to Noise ~ 3 ~ 68 Z’ value 0.52 0.84 Ex & Em Wavelengths λex: 500 nm, λem: 590 nm λex: 360 nm, λem: 450 nm

4.4.2 Rv3802 Structure and PLA classification

Given the structural and enzymatic attributes of Rv3802, the Rv3802 family of enzymes most likely falls within the PLA2 enzyme class146. Rv3802 was previously 116

determined to have PLA activity towards phosphatidyl species (phosphatidylserine, phosphatidylcholine, and phosphatidylethanolamine)136. In contrast to thioesters, the ester moieties of phospholipids are generally stable in aqueous solutions, so the enzymatic activity observed in vitro is strongly suggestive of biological activity. The two general classes of PLAs differ with regards to specificity for either the sn-1 acyl chain (PLA1) or sn-2 acyl chain (PLA2) of the glycerol phosphate moiety of phosphatidyl species147. For unknown reasons, PLA2 enzymes possess numerous disulfides, whereas the PLA1 enzymes typically lack disulfides146. Mtb Rv3802 was found to possess two disulfides

(Figure 4-5). These two disulfides are present in the Msmeg ortholog structure and were found to be conserved across all Rv3802 orthologs138.

The first conserved disulfide, C72 - C164, covalently links the N-terminal peptides to the base of α2 helix (Figure 4-5). This is of potential structural importance as the 41 residues preceding V75, C72 included, are predicted to be unstructured until the predicted N-terminal transmembrane region (I12-I34)(Figure 4-14). Therefore, this disulfide may help retain proper protein fold of the structured catalytic hydrolase domain while being immediately preceded by a large, unstructured domain. The second conserved disulfide, C264 - C271, induces a kink in a loop to appropriately position the catalytically important D268 residue. This second disulfide therefore ensures the proper positioning of D268 relative to H299, maintaining the catalytic triad despite movement of the nearby α7-helix upon lipid binding.

117

Figure 4-14: (A) Secondary structure and (B) trans membrane region predictions. Figure reproduced from Goins et al. the Journal of Biological Chemistry, 2017135.

4.4.3 Mechanism of catalysis

PLA2 enzymes catalyze fatty acid hydrolysis through two general mechanisms: a calcium dependent His/Asp mechanism or through calcium independent mechanisms146.

Given the identified catalytic triad and the lack of calcium-dependent enzyme activity,

Rv3802 is a calcium independent PLA2 enzyme. Flanked by aromatic side chains (W84,

F174) and displaying little movement of catalytic residues between open and closed enzymatic states, the active site of Rv3802 resembles that of a typical hydrolase enzyme

(Figure 4-7C)69. These features observed in the Rv3802 active site are known to help promote substrate hydrolysis over transfer69. These structural attributes help explain why

Rv3802 has no reported transferase activity in the presence of acyl acceptors136.

Based on the enzymatic activity, structure, and catalytic triad of Rv3802, catalysis most likely proceeds through a Ping-Pong reaction mechanism. The reaction can be 118

broken down into two half-reactions. The first half-reaction proceeds through nucleophilic attack on the carbon of the ester/thioester carbonyl by S175, which is deprotonated by H299. The terminal oxygen of PEG1 would therefore mimic the carbon of the carbonyl moiety of substrate prior to nucleophilic attack, depicted in Figure 4-9A.

Following attack, the plausible oxyanion hole consisting of backbone amide nitrogen atoms of T83 and N176 stabilizes the tetrahedral intermediate. Tetrahedral intermediate collapses yields the acyl-enzyme intermediate and the release of the now deacylated lipid at the end of the first half-reaction. The acyl-enzyme intermediate form of the enzyme is approximated by the terminal oxygen of PEG1 from protein molecule B (Figure 4-9B).

The second half-reaction would therefore proceed through the activation of a water molecule by H299 and subsequent nucleophilic attack on the acyl-enzyme intermediate.

The nucleophilic water molecule is observed within hydrogen bonding distance to H299 in protein molecule B and is depicted in Figure 4-9B. Tetrahedral collapse following this second nucleophilic attack results in the release of a free fatty acid and a regenerated enzyme.

While this mechanism most likely describes the esterase activity of Rv3802, it is interesting to note that catalysis can occur without the nucleophilic serine (Figure 4-10).

Based on these observations, Rv3802 can appropriately bind substrate and orient a water molecule in place of the serine hydroxyl for catalysis to occur. A similar phenomenon has previously been observed in serine proteases148. Therefore, to completely abolish Rv3802 esterase activity, the catalytic histidine must be removed to eliminate water or hydroxyl activation.

119

4.4.4 Dynamics of lipid binding

Based on the structural similarity of PEG to the short-chain fatty acids of various lipids, it is therefore logical that a similar repositioning of α8 helix must occur for binding of natural substrate and subsequent catalysis to occur. The Rv3802

N132C/N288C mutant, with limited α8 helix movement, displayed a lower ability to bind

4MH, resulting in a significant loss in activity compared to WT enzyme (Figure 4-11B).

However, when the larger PI substrate was used, no PLA activity was observed for the

N132C/N288C mutant (Figure 4-11C). This loss in activity as a result of hindered helical movement highlights the importance of protein dynamics with respect to substrate binding. Given the large variety of possible structural forms of α8-helix and adjacent loops, it is difficult to fully assess how this dynamic interplay affects substrate recognition.

Initial substrate recognition may proceed through binding of the lipid head, as this site does not change between open and closed states of the enzyme. A MD simulation of apo Rv3802 suggests α8-helix quickly adapts the closed conformation and remains in that state without substrate present when in the aqueous phase (Figure 4-15, Simulation conducted by Steven Dajnowicz, methods can be found in Goins et al. the Journal of

Biological Chemistry, 2017). Therefore, substrate binding may drive helical opening in a solvated environment. However, given the uncertainty of the true biological function of

Rv3802, it is difficult to fully assess the mechanism of lipid binding based on the

Rv3802-PEG structural observations and mutant kinetics. PLA2 enzymes have been found to bind only the surface-exposed lipid head group, reside partially within the

120

membrane itself and bind the entire lipid or extract the lipid from the membrane for hydrolysis146, 149.

Figure 4-15: MD simulation of apo Rv3802. (A) Superposition of Mtb Rv3802-PEG, the Mtb Rv3802 MD model at 1 ns, and the Msmeg 6394. (B) Root mean square distance (RMSD) of Cα atoms for residues in α8-helix compared to Cα atoms of the entire protein. The α8-helix quickly adopts the closed conformation and remaining closed for the 100 ns simulation. (C) Representation of dynamic regions as a function of calculated root mean square fluctuations. Regions of high fluctuation have greater tube thickness and are of red hue. The equilibrated apo Rv3802 model prior to MD simulations is depicted. Figure reproduced from Goins et al. the Journal of Biological Chemistry, 2017135.

121

When α8-helix is in the open state, a vast hydrophobic lipid-binding site is apparent (Figure 4-6). In agreement with typical lipid binding sites, the identified Rv3802 lipid binding site is lined with aliphatic amino acid side chains (Figure 4-6)150. The

Rv3802 hydrophobic binding channel is relatively linear, extending away from the active site, therefore resulting in a site capable of binding two alkyl chains in a parallel linear arrangement similar to the arrangement of the two observed PEG molecules. Therefore, the binding of PEG molecules observed within Rv3802 provides a structural basis for the binding of the two acyl chains of the phosphatidyl substrate.

4.4.5 Structure-based insights on biological function

Meniche and colleagues determined that Rv3802 orthologs NCgl2775 and

MSMEG_6394 play an essential role in modulating lipid composition137. Specifically, they noted the increase of the mycolic acid to glycerophospholipid ratio, suggesting this family of enzymes play a direct role in decreasing glycerophospholipid content within the mycomembrane137. Therefore, it is of little surprise that Rv3802 did not hydrolyze acyl- chains from the tested acyl-arabinose compounds that mimic the mAG. It should be noted that NCgl2775 is not essential under physiological growth, whereas the Mtb orthologs are essential137. This differing essentiality of Rv3802 orthologs may be attributed to the differing compositions of the outer membranes between mycobacteria and corynebacteria46. For example, the inner layer of the mycomembrane of corynebacteria has a much higher composition of cardiolipin relative to other glycerophospholipids, when compared to mycobacteria46.

PI species comprise a major portion of the glycerophospholipids found within the plasma membrane and mycomembrane layers in both Mtb and Msmeg24 Previously, 122

Rv3802 was shown in vitro to hydrolyze a fatty acid from PIM2 upon treatment of

Msmeg cell wall extracts with the enzyme136. PIMs play an important role in the permeability of the Mtb cell wall, proper cell division, and influencing pathogenesis26, 41,

151, 152 Interestingly, when a conditional knockout of MSMEG_6394 was cultured, cell morphology was jagged and elongated compared to WT138. PIM2 generally has four acyl substituents, two on the phosphoglycerol, one on the inositol, and the fourth on the mannose moiety. The observed lipid-binding site of Rv3802 is significantly larger than the palmitoyl binding pocket of other hydrolases known to hydrolyze esters of PIM66.

Specifically, the enzyme PatA resides within the PIM biosynthetic pathway and is responsible for the transfer of a palmitoyl group to the 6 position on the mannose sugar66.

Based on the larger Rv3802 lipid binding site compared to PatA, a sensible biological function for Rv3802 is the hydrolysis of one or both of the fatty acids from the phosphogylcerol component of PIM or other PI species. PIM has been shown to be an antigen of CD1d-restricted T cells; however, PIM binding to CD1d is abolished upon the hydrolysis of the diacyl phosphoglycerol moiety following treatment with a PLA2 enzyme41. While Rv3802 has the enzymatic ability to facilitate this chemistry, further in- depth studies are required to truly validate the biological context of such action.

Using the Rv3802-PEG structure as a basis, PI was modeled into the open form of the enzyme (Figure 4-16). This model provides a reasonable structural basis for binding of phosphatidyl-based substrates to Rv3802. Structurally, PI can be easily accommodated within the lipid-binding site of Rv3802 in a manner that promotes catalysis. As modeled, the sn-2 acyl chain would be subject to hydrolysis. While little positive charge is present at this site to counter the negatively charged phosphodiester moiety, stabilization of 123

interactions with the phosphate moiety may be afforded through solvent or by backbone amide nitrogen atoms of the highly dynamic loop upon restructuring as a function of substrate binding (A292-E296). Due to the proximity of this loop to the lipid head- binding site, this region may also be relevant to sugar binding of larger PI species, such as PIM.

Figure 4-16: PI modeled into the lipid binding site of Rv3802. The fatty acid chains are easily accommodated within the lipid binding site allowing the sn-2 ester to be positioned by the nucleophilic S175, placing the inositol moiety within the proposed sugar-binding site. FA1: tuberculostearic acid, sn-1 position; FA2: palmitic acid, sn-2 position of glycerol. Figure reproduced from Goins et al. the Journal of Biological Chemistry, 2017135.

The essential nature of Rv3802 to Mtb viability makes it an alluring new target for future drug development. The identification of the lipid-binding site of Rv3802 highlights an extremely intriguing and potentially promising site for drug development.

The presented structure and subsequent analysis can be utilized for designing inhibitors specific to Rv3802.

124

4.5 Conclusions

The first ever X-ray crystal structure of Rv3802 was solved. Due to the presence of PEG within Rv3802, the lipid-binding site of Rv3802 was structurally identified.

Comparison of the Rv3802-PEG structure with the apo ortholog revealed dynamic regions required for substrate binding. Structural observations were confirmed via covalent tethering through the introduction of an engineered disulfide bond. Two fluorescent assays in addition to a PLA assay were developed and evaluated to monitor

Rv3802 activity in vitro. The structurally identified lipid binding site of Rv3802 was shown to accommodate PI substrates, further suggesting a biological role of glycerophospholipid degradation in mycobacteria.

125

Chapter 5

Inhibiting Rv3802

5.1 Background

5.1.1 Rv3802 as a drug target

The biological function of Rv3802 may remain uncertain; however, the rv3802c gene was determined to be putatively essential for Mtb viability110. The Rv3802 ortholog in the non-pathogenic Msmeg, MSMEG_6394, was also found to be essential for general viability137, 138. As stated in chapter 4, a conditional knockout of MSEG_6394 in Msmeg cells resulted in a jagged, elongated cellular morphology138. Based on these findings,

Rv3802 is therefore a potential target for drug development. However, Rv3802 was only found in cell wall fractions and is believed to reside in the periplasmic space due to a helical transmembrane anchor140. As a consequence, in vitro inhibitors of Rv3802 have to cross the mycomembrane to affect Rv3802 in vivo. On this basis, the development of drugs specific to Rv3802 which also have biological activity against Mtb may prove challenging.

5.1.2 Inhibition by tetrahydrolipstatin

As a consequence of an uncertain biological function and until recently a structure to guide drug design efforts, only one effort has been made to design inhibitors specific to 126

Rv3802. As stated in chapter 4, Rv3802 is a serine esterase with Phospholipase A activity136. Given these attributes, Rv3802 is potently inhibited by THL138, 153. THL has a reported Ki of 0.87 μM for Mtb Rv3802 with IC50 values ranging between 2 and 5 μM depending on the assay utilized138, 153. Using the THL scaffold as a basis, West and colleagues synthesized a small library of THL compounds with derivatized peptidyl arms

(Figure 5-1)153. A handful of derivatives display enhanced levels of in vitro and

153 correlated in vivo inhibition . The best derivative has an IC50 19-fold lower than THL

153 against Rv3802 and 12 fold lower IC50 against H37Rv Mtb (1.3 μM) . While the enhanced in vitro inhibition is undeniable, the enhanced levels of in vivo inhibition may or may not be a result of Rv3802 inhibition due to the promiscuity of THL with regard to

Mtb lipid esterases73. Unfortunately, until a definitive biological function of Rv3802 is determined, it may prove difficult to fully assess if Rv3802 inhibitors are indeed targeting the enzyme in vivo. While not as conclusive as a lipidomics study of treated cells, a plausible way to assess if Rv3802 or Msmeg ortholog is being inhibited in vivo is comparison of cell morphology to that of the Msmeg cell line with a conditional

MSMEG_6394 knockout.

Figure 5-1: THL and peptidyl arm derivatives153.

Chapter 5 details our efforts in discovering new inhibitory scaffolds that can serve as a basis for future drug design specific to Rv3802. To do so, the fluorescence-based 127

assays described in chapter 4 were utilized to screen two drug libraries containing over

1,500 compounds believed to be active against TB. Identified hits were validated and further enzymatically characterized. The best molecules were tested against Msmeg to test for changes in cell morphology as well as two strains of Mtb to evaluate biological activity.

5.2 Methods

5.2.1 Protein expression and purification

WT Rv3802 was expressed and purified in an identical manner as described in section 4.2.1 and 4.2.2.Rv3802 was extensively dialyzed into a 50 mM sodium phosphate pH 7.5 buffer for in vitro assay work.

5.2.2 Drug library screening

5.2.2.1 Screening parameters

The two drug libraries screened were the TB Alliance and GSK TB active drug libraries described in Chapter 3. All compounds were again screened at 100 μM with 15 minute preincubation before kinetic reads were initiated. The larger TB alliance library was screened with the first generation RfB assay described in section 4.2.3.1. Briefly, reactions consisted of 1 μM enzyme, 100 μM RfB, and 100 μM drug or DMSO. The

GSK library was screened using the second generation 4MH assay described in section

4.2.3.2. Briefly, reactions consisted of 50 nM enzyme, 75 μM 4MH, and 100 μM drug or

DMSO. To account for multiple days required to screen the TB alliance library, data was plotted as a function of inhibited reaction rates divided by uninhibited reaction rates for that specific read (40 compounds per read). GSK data was plotted simply as a function of 128

reaction rate. Again, “hits” were determined based on inhibition levels below 3σ or 3 times the standard deviation of all DMSO control reactions.

5.2.2.2 “Hit” Validation

Identified “hits” were validated through dose dependence. Compounds from the

TB alliance library were purchased from ChemBridge and dissolved in DMSO to a working concentration of 10 mM. “Hits” identified from the GSK library were simply serial diluted from the screen. In either case, compounds were tested at 100, 50 and 25

μM. The RfB assay was used for the TB alliance “hits” and the 4MH assay used for the

GSK “hits.”

5.2.2.3 Ki determination

Three compounds identified from the TB Alliance library displayed promising levels of in vitro inhibition of Rv3802 upon initial dose dependence studies. Compounds were set at 30 mM in DMSO and serial diluted with DMSO. Reaction concentrations of compounds ranged from 300 to 1.56 μM for Ki determination. The RfB assay was used for Ki determination, following background subtraction; triplicate rates were plotted in

PRISM 7 and data fit with the Morrison equation. The Morrison equation is used for tight

86 binding inhibitors to determine Ki apparent values .

5.2.3 In vivo effects on Msmeg by Mtb active compounds

5.2.3.1 Growth inhibition of Msmeg

Two approaches were utilized to investigate the influence of identified compounds on Msmeg growth: liquid cultures and solid media for cell imaging studies.

100 mL flasks of sterile LB media and 0.1 % v/v Tyloxapol were inoculated with Msmeg cells and allowed to grow at 37 °C until an OD600nm of 0.7 was reached. Compounds 129

dissolved in DMSO were added to a final concentration of 15 μM, the OD600nm was checked every hour for 4 hours and then 12 hours later to evaluate cell death. To investigate influence of compounds on early Msmeg growth, 15 μM of compounds were added to sterile LB media with 0.1 % v/v Tyloxapol prior to inoculation. OD600nm was checked following 24 hours of incubation at 37 °C. Finally, to evaluate the affects of cellular stress, Msmeg was grown in LB media with 0.1 % v/v Tyloxapol to an OD600nm of 0.2 at 37 °C, a final concentration of 15 μM of compound 59 was added and the temperature increased to 42 °C. Culture density was checked every hour for 5 hours and then 20 hours later.

For solid media studies, liquid cultures of Msmeg were grown in LB media with

0.1 % v/v Tyloxapol to an OD600nm of 0.7. 50 μL of cells were platted on LB agar plates with no antibiotic added. Initially, Kirby Bauer disks were utilized to establish an MIC value for Msmeg; however, no detectable cell death was observed. LB agar plates were therefore made with compounds or equal volume of DMSO (50 uL) added to the media at

200, 100, and 50 μg. Again, 50 uL of Msmeg cells grown to a density of 0.7 OD600nm were spread on agar plates and allowed to grow for 24 hours. Regions of limited growth were used for microscopy studies.

5.2.3.2 Scanning electron microscopy of Msmeg treated cells.

Msmeg cells treated with compounds of interest were imaged using scanning electron microscopy (SEM) to evaluate change in cell morphology as a consequence of

MSMEG_6394 (Rv3802 Msmeg ortholog) inhibition. Treated and untreated cells were fixed and imaged using a method adapted from Crellin and colleagues to study the cell morphology of Msmeg with a MSMEG_6394 conditional knockout138. 50 uL of 2.5 % 130

v/v glutaraldehyde (Sigma Aldrich) in a phosphate buffered saline was titrated onto cells.

Following 2 minutes of incubation, cells were scrapped into a micro centrifuge tube and allowed to incubate for another half hour at room temperature. Cells were gently pelleted at 5,000 rpm for five minutes and washed 3 times with phosphate buffered saline and then 3 times with dH20. Following washing, cells were slowly dehydrated through washes of 10, 30, 50, 70, and 90 % ethanol, cells were incubated with respective concentration for 15 minutes each. For SEM imaging, cells were resuspended in hexamethyldisilizane and spotted on plastic film fixed to SEM spuds. Samples were gold coated and imaged using a JOEL JSM-7500F at 2 kV. Dr. Panne Burckel coated samples in gold and greatly assisted with cell imagineg.

5.3 Results

5.3.1 Identification of lead molecules

5.3.1.1 TB alliance library

Screening Rv3802 against the TB alliance library of compounds resulted in the identification of 51 compounds that were considered a statistical “hit.” Due to the low signal to noise of the RfB assay, the threshold for statistical “hits” was lowered to 4σ, which represents 50 % inhibition. Based off this criterion, the “hit” rate is lower than 4

%.

131

Figure 5-2: TB alliance library screening results. Red line signifies 3σ cut off, green line signifies 4σ cut off. Blue reaction markers near 0 V’/V’o are mostly control reactions without enzyme present.

On the basis of % inhibition, structure, chemical attribute, price, and availability,

8 molecules were chosen for further evaluation. The 8 molecules that were chosen all displayed dose dependence inhibition, validating these screening “hits.” The best molecules with regard to Rv3802 in vitro inhibition were compounds “59-F3,” “60-D5,” and “74-A4.” Molecule naming was based off drug plate and subsequent well ID. From here on out these molecules will simply be referred to as compound 59, 60, and 74. Based on observed inhibition, all three compounds have IC50 values below 25 μM and were chosen for further characterization.

132

Figure 5-3: Dose dependence of TB alliance drug screening “hits”

Compounds 59, 60, and 74 were originally screened against Rv3802 between 300 and 1.56 μM for IC50 determination. However, the observed level of drug concentration required for 50 % inhibition was close to enzyme concentration; therefore, a Morrison Ki equation was used to fit the data. The resulting Ki apparent values are 2.78 ± 0.28, 3.23 ±

0.49, and 4.01 ± 0.78 μM for compounds 59, 60, and 74, respectively (Figure 5-4).

Figure 5-4: Ki determination for best lead molecules.

5.3.1.2 GSK library

Screening of the smaller, 177 molecule TB active library from GSK resulted in the identification of 3 potential “hits.” The statistical cut off of 3σ represented 56 %

133

inhibition level. Due to relatively inconsistent rates of positive reactions, this value is higher than previous drug screens within this dissertation.

Figure 5-5: GSK library screening.

The three best “hit” molecules were further tested for dose dependence.

Compounds 45 and 61 failed to show dose dependence, indicating false hits. Compound

208 displayed dose dependence; however, levels of inhibition were far below initial screening results. As such, none of these molecules were pursued.

Dose Dependence: 1.2 Rv3802 GSK hits No Inhibitor uM 1.0 25 uM 0.8 50 uM

o

’

V 100 uM

/ 0.6

’

V 0.4

0.2

0.0 5 1 8 4 6 0 2 Figure 5-6: Dose dependence of identified “hits” from GSK library.

134

5.3.2 Biological activity of lead molecules

5.3.2.1 Effects on Msmeg growth

Compounds 59, 60, and 74 were selected for in vivo studies. Compounds were sent to Dr. Mary Jackson’s lab at Colorado State University Mycobacteria Research

Laboratories for MIC determination against Mtb. During this time however, we sought to further characterize potential in vivo effects of identified inhibitors on non-pathogenic

Msmeg.

Initially, smaller 5 mL cultures of Msmeg were pursued for these studies; however, cells continually fell out of solution due to aggregation. On this basis, larger

100 mL cultures were used which limited the concentration of compound we could test.

Titration of 15 μM of respective drug into cultures of Msmeg that were in the growth phase at 37 °C failed to effect cell growth upon inspection of OD600nm over various time points. Treated and untreated Msmeg cultures are shown after 12 hours of titration in

Figure 5-7.

Figure 5-7: Msmeg cultures treated with DMSO or compounds of interest. None of the identified compounds reduced culture growth at 15 μM.

To assess if these molecules hindered growth from an early onset, 15 μM of respective drug compound was titrated into LB media with 0.1 % tyloxapol prior to inoculation with Msmeg cells. However, upon monitoring the OD600nm after 24 hours of 135

growth, no difference was observed between treated and untreated cells, suggesting a higher concentration or continual titration of drug is required.

Based on the identification that the Msmeg ortholog of Rv3802 was identified to influence mycomembrane lipid composition under heat induced cell stress, we tested to see if identified compounds were more potent under cellular stress137. Only compound 59 was chosen for this experiment. No difference was observed between DMSO and drug treated cultures following an hour after titration. However, a noticeable trend between treated and untreated cells was observed between hours 1 and 4 (Figure 5-8).

Unfortunately, cell growth continued after hour 4, and trended similarly to that of untreated cells.

Figure 5-8: Culture density of treated and untreated Msmeg cells under heat stress conditions.

Upon plating Msmeg cells on LB agar plates and placing Kirby Bauer disks with varying concentrations of compounds 59, 60, and 74, no observed growth inhibition zone was observed. Therefore, LB agar plates were doped with 200 μg of respective compound or equal % v/v DMSO prior to Msmeg plating. Upon incubation at 37 °C for 24 hours, no significant difference was observed between treated and untreated plates (Figure 5-9).

136

Cell colonies were selected on the edge between growth and non-growth areas for cell imaging studies.

Figure 5-9: LB agar plates doped with DMSO or respective drug compound. Plates were incubated at 37 °C for 24 hours. Plates are labeled with DMSO or respective drug number.

5.3.2.2 Influence on Msmeg morphology

Msmeg cells with a conditional knockout of the Rv3802 ortholog produced cells with elongated and jagged morphology138. We were therefore interested in seeing if

Msmeg cells treated with identified Rv3802 inhibitors produced a similar cellular morphology. However, no significant change in morphology was observed between cells grown on DMSO or drug doped LB agar plates (Figure 5-10). This experiment was repeated a handful of times and yielded similar results.

137

Figure 5-10: SEM images of treated and untreated Msmeg cells. No signs of change in morphology were observed that was similar to the jagged morphology reported in the Msmeg cells with MSMEG_6394 conditional knockout138. A) DMSO control. B) compound 59. C) compound 60. D) compound 74.

5.4 Discussion

5.4.1 Structure based rationale for Rv3802 inhibitors

The lipid-binding site of Rv3802 has characteristics of an ideal site for drug targeting (Figure 5-11). The large hydrophobic FA portion of the active site can afford entropic drug binding through the sequestering of apolar chemical attributes given this region generally excludes solvent154, 155. While entropic contributions facilitate drug binding, ideally enthalpic contributions to drug binding are considered ideal with regard to drug specificity and affinity154, 155. While the lipid binding site is largely hydrophobic

138

this region can still afford specific chemical interactions through: aromatic interactions with Phe130, Trp84, or Phe174; electrostatic interactions with Glu229, or general hydrogen bond interactions through any of the numerous backbone carbonyls or amides found within the FA binding portion of the lipid binding site. Opposite of the FA binding site is the inositolphosphate binding site which includes the polar residues of Glu85 and

Lys100135. However, this portion of the lipid-binding site is solvent exposed. Aside from the expansive nature of the Rv3802 lipid-binding site, the shape is generally linear.

Therefore, an ideal inhibitor of Rv3802 is bifurcated in a linear, chemical fashion as is the lipid-binding site with regard to the FA and inositolphosphate binding sites.

Figure 5-11: Rv3802 lipid binding site with distinct regions identified. Below is a space filling model of residues that line the lipid binding site. A noticeable linear channel is observed through the lipid binding site. Sphere color: white – carbon, red – oxygen, blue –nitrogen. Nucleophilic serine is denoted in yellow (PDB: 5W95)135.

139

Indeed, all the molecules identified from the TB alliance drug library screening were linear in nature (Figure 5-12). In general, most compounds had aromatic or heterocyclic substituents on either end and were adjoined through an amide linkage.

Based on these structural similarities, these compounds most likely bind in a similar manner. Interestingly compounds 60-D5, 63-F4, and 68-B3 share an identical scaffold and allow a small SAR analysis. Based on the observed dose dependence between these compounds, the halogenated substitution pattern of the amide linked benzene influences inhibition. In this case, the meta position is preferred to the ortho substitution. However, the bulkier trifluoromethyl group at the ortho position significantly decreases inhibition levels as observed by compound 68-B3.

Figure 5-12: Structures of identified lead molecules from the TB alliance drug library. Best molecules are bolded.

Significant effort was spent on obtaining a crystal structure of Rv3802 in complex with compounds 59, 60, and 74. Preliminary crystals were obtained of Rv3802 and respective compounds; unfortunately, obtained crystals either diffracted poorly, were of insufficient quality for indexing, or were simply salt (crystallization and diffraction studies not included). Regardless, the Rv3802-PEG structure allows for the loose modeling of these compounds within the lipid binding sites. Potential binding

140

orientations of compounds 60 and 74 are depicted in Figure 5-13. The fused tricycle compound 59 most likely resides mostly within the hydrophobic FA portion of the lipid binding site. Based off these binding modes, significant room still remains for further chemical derivatization.

Figure 5-13: Identified lead molecules 60 and 74 modeled into the lipid-binding site of Rv3802. Compound 60 is magenta, compound 74 is green (PDB:5W95)135.

5.4.2 Biological evaluation of lead molecules

The best in vitro inhibitors of Rv3802 failed to produce noticeable levels of in vivo inhibition in Msmeg when tested at modest concentrations (~5 μM). Given the structural and enzymatic similarities between Mtb Rv3802 and Msmeg MSMEG_6394, identified inhibitors of Rv3802 should also be capable of inhibiting the Msmeg ortholog135, 138. Compounds failed to inhibit growth of Msmeg cultures in both lag and exponential phases of growth. Additionally, attempts to evaluate changes in cellular morphology consistent with Rv3802 ortholog suppression failed to identify any

141

observable changes between treated and untreated cells. However, a small change was observed in Msmeg cultures under heat induced cellular stress when treated with the best

Rv3802 inhibitor, 59. Based on the observed growth curves, a higher concentration or multiple titrations of compound 59 are required to further decrease culture growth under stress conditions. Regardless, this preliminary data may suggest that compound 59 was targeting the Msmeg ortholog of Rv3802 in vivo, given that the essential nature of

Rv3802 appears to be amplified under cellular stress; therefore, requiring lower concentration of inhibitor137.

From the onset of these studies, we assumed the identified compounds would have noticeable activity against mycobacteria given these compounds were discovered in a “TB active library.” However, we quickly learned the identified compounds were not as active as initially assumed. While we weren’t interested in obtaining an MIC value for

Msmeg as the compounds were being tested by our collaborators in Mtb, more insights on appropriate concentrations to use for cell morphology studies could be gained by determining an MIC value for Msmeg. MIC values can be determined using a resazurin based assay that monitors cell growth and only required microliters of cell culture156.

Shortly following the SEM imaging studies, Dr. Mary Jackson’s lab determined the MIC values against Mtb H37Rv to be 40 μM for compound 59 and above 80 μM for compounds 60 and 74 (Table 5.1). Based on these values, the identified in vitro inhibitors of Rv3802 were poor growth inhibitors of Mtb, despite having similar levels of in vitro inhibition as THL138. Two plausible explanations for the stark difference in Mtb inhibition levels are: 1) THL inhibits more enzymes than just Rv3802 in vivo and 2)

Difference in accessibility to Rv3802. THL is calculated to be roughly twice as 142

hydrophobic as compound 59, based on ClogP values (Table 5.1). A logP value of -2 to 4 is suggested to be ideal for cellular uptake across a normal CM (ClogP = computed partitioning coefficient, logP = experimentally determined partitioning coefficient)157.

However, one would assume that a more hydrophobic compound would remain within the mycomembrane and not access the periplasmic region where Rv3802 is believed to reside140. On this rationale, the first scenario is more likely; however, the accessibility of compound 59 to Rv3802 may still be a concern. Compound 59 has an MIC value ~220 fold higher than that of the first line TB drug, isoniazid, against the H37Rv strain of Mtb; therefore, significant improvements must be made to this scaffold to make it a viable Mtb inhibitor158.

Table 5.1: Chemical and inhibitory properties of identified in vitro inhibitors of Rv3802 and in vivo inhibition. (*) Indicates previously published values138, 153. ClogP values determined in ChemDraw. Compound ClogP Ki app (μM) Mtb H37Rv MIC (μM) Mtb H37Rv IC50 (μM) 59 4.71 2.78 ± 0.28 40 -- 60 5.13 3.23 ± 0.49 > 80 -- 74 3.88 4.01 ± 0.78 > 80 -- THL* 8.61 0.87 ± 0.3 -- 15

5.5 Conclusions

Over 1500 “TB active” compounds were tested against Rv3802 to identify potential lead molecules for future development. Three promising scaffolds were identified in vitro and determined to have low μM affinity towards Rv3802. Identified molecules had structures amenable for shape complementarity towards the lipid-binding site of Rv3802. Efforts to crystallize Rv3802 with compounds failed to produce usable protein-drug complex structures; however, rudimentary modeling suggests significant room remains for further derivatization of compounds.

143

Despite promising levels of in vitro inhibition, modest concentrations of compounds failed to inhibit Msmeg growth and produce morphological changes indicative of MSMEG_6294 inhibition under physiological conditions. However, treatment of Msmeg cells under stress conditions may increase the potency of compound

59 which would correlate with an increase in Rv3802 essentiality under cell stress.

Determined MIC values towards Mtb indicated compounds were far less “TB active” than previously thought. Compound 59 displayed the best in vivo inhibition; however, levels were 220 fold worse then the first line drug isoniazid. On this basis, morphology studies will require higher concentrations of compounds than what was tested in this study to identify if indeed the identified compounds target Rv3802 in vivo.

144

Chapter 6

Influence of tetrahydrolipstatin stereochemistry on Ag85C and Rv3802 inhibition

6.1 Background (Portions of text reproduced from Goins et al. ACS Biochemistry,

2018)159

6.1.1 Reversible covalent inhibition by tetrahydrolipstatin

THL is a stable, synthetic derivative of the naturally occurring human pancreatic lipase inhibitor, lipstatin, produced by Streptomycesan toxytricini97. Initially, THL was found to inhibit the pancreatic lipase, gastric lipases, and the carboxyl ester lipase

(cholesterol esterase). Human lipase inhibition results in the reduction of fat absorption, leading THL (Orlistat) to be used as an FDA approved treatment for obesity160. Inhibition of the cholesterol esterase by THL is covalent, yet reversible in aqueous solutions97.

Covalent inhibition proceeds through an acyl-enzyme type inhibitor-enzyme complex as a result of nucleophilic attack on THL by the enzyme97. Nucleophilic attack occurs on the carbonyl center of the β-lactone ring by the serine nucleophile, resulting in ring opening

(Figure 6-1A B)97. However, this acyl-enzyme complex is reversible due to water activation by the enzyme103. Subsequent hydrolysis of the acyl-enzyme complex leads to the recovery of enzymatic activity and inactivation of the inhibitor (Figure 6-1B and C)97.

This inactivation of the β-lactone moiety is analogous to β-lactam inactivation by β- 145

lactamases and carbapenumases and may be the reason for the modest THL MIC against

Mtb and non-tuberculosis mycobacteria161, 162.

Figure 6-1: THL structure and inhibition. A) THL has 4 stereocenters at the 2, 3, 5, and 2’ carbons, carbon 1 of the carbonyl is subject to nucleophilic attack.1 B) Nucleophilic attack by the catalytic serine of serine esterases results in β-ring opening and an ester linked enzyme-THL covalent complex97. C) Covalent inhibition is considered reversible as a result of water activation by the catalytic histidine yielding an inactive, hydrolyzed THL and an active enzyme97, 103. Figure adapted from Goins et al. ACS Biochemistry, 2018159.

An additional in vivo target of THL was later found to be the thioesterase domain of the human fatty acid synthase (FAS)101. Due to the overexpression of FAS in a variety of tumor cells, THL has been proven to be a potential cancer therapeutic101, 163, 164. In

2007, the X-ray crystal structure of THL in complex with the thioesterase domain of FAS was solved having two protein molecules in the asymmetric unit84. The THL acyl- enzyme covalent complex was observed in one protein molecule, while a bound, hydrolyzed form of THL was observed in the second protein molecule84. In the intact

THL acyl-enzyme protein molecule, the catalytic histidine is hydrogen bonded to the β- hydroxyl of THL that results from ring opening84. However, when this interaction is disrupted, the catalytic histidine is free to activate a water molecule for nucleophilic attack on the THL acyl-enzyme ester linkage103. The THL-FAS structure and subsequent

146

computational modeling provides a molecular basis for the observed reversible covalent inhibition of lipases by THL.

Figure 6-2: Human FAS-THL structure. Hydrogen bonding of the catalytic histidine to the β-hydroxyl results in reduced water activation. However, when disrupted the catalytic histidine can activate water for hydrolysis (PDB: 2PX6)84, 103.

6.1.2 Non-specific inhibition of Mtb lipid esterases

The ability of THL to inhibit a variety of lipases is not exclusive to humans. THL is known to target a myriad of lipid esterases (serine esterases) in mycobacteria73.

Mycobacterium bovis (Mbovis) cell lysate treated with biotin or fluorescently-labeled

THL derivatives resulted in the identification of 261 target proteins using mass spectrometry analysis of the enriched sample (biotin-THL pull-down), 14 of the 261 target proteins were cross validated using the fluorescently-labeled THL and non- enriched sample73. However, this non-specific targeting of lipid esterases only results in a

153 modest IC50 against Mtb strain H37Rv of ~15 μM . Given the high number of potential targets of THL in Mtb, the question stands, what attributes of THL make the molecule such a promiscuous inhibitor of lipid esterases in mycobacteria? One plausible rationale is that the two-alkyl chains of THL mimic the fatty acid chains of mycobacterial lipids,

147

giving the drug affinity to the lipid binding sites of lipid esterases23. More so, there may be an aspect of substrate mimicry as the core scaffold of THL is structurally similar to that of mycolic acids, a major lipid essential for Mtb growth, viability, and pathogenesis45. An important structural attribute of THL is the presence of four stereocenters within the molecule. These centers reside on carbons 2 and 3 of the β- lactone ring, carbon 5 on the palmitic core, and carbon 2’ on the peptidyl side chain

(Figure 6-1A). To date, no one has truly investigated how or why THL can inhibit numerous lipid esterases.

6.1.3 THL Inhibition of Ag85C and Rv3802

The Mbovis homolog of Ag85C was identified as one of the 14 validated targets of THL73. Chapter 2 discusses the in vitro inhibition of Ag85C by THL and presents the crystal structure of the Ag85C-THL complex57. Briefly, THL has low μM affinity and covalently modified Ag85C within minutes57. Covalent inhibition by THL stimulates a conformational change that results in the displacement of the catalytic histidine of

Ag85C, yielding an extremely stable, acyl-enzyme inhibited complex57. The observed inhibitory mechanism and specific non-covalent interactions between Ag85C and THL is in stark contrast to that of human FAS and THL57, 84.

The Mbovis homolog of Rv3802 was identified as one of the 261 potential target proteins of THL, but was not one of the 14 cross validated targets73. THL has been reported to have high nM affinity towards Rv3802138. While THL inhibition was believed to be covalent, no studies conclusively confirmed this nor show how stable covalent inhibition was. As stated in chapter 5, a series of THL derivatives were developed and tested against Rv3802 in vitro showing nM IC50 values with correlated improvements of 148

inhibition in vivo against Mtb153. However, due to a lack of structure for Rv3802 at the time, no molecular basis for inhibition was feasible.

Chapter 6 takes a closer look at inhibition of the two essential lipid esterases of

Ag85C and Rv3802 by THL. A library of THL stereoderivatives was utilized to test how stereochemistry affects cross reactivity, selectivity, and stability of covalent inhibition by

THL. Two stereocenters of THL were found to influence cross reactivity, while the other two centers influence selectivity and stability of the acyl-enzyme covalent complex.

6.2 Methods (Portions of text reproduced from Goins et al. ACS Biochemistry,

2018)159

6.2.1 Protein expression and purification

Ag85C was expressed and purified as described in section 2.2.2. Again, protein was dialyzed into 50 mM sodium pohosphate pH 7.5 for assay work. Rv3802 was expressed and purified as described in section 4.2.2. As with Ag85C, Rv3802 was extensively dialyzed into a 50 mM sodium phosphate pH 7.5 buffer for following assays.

6.2.2 Initial screening of tetrahydrolipstatin stereoderivatives

Ag85C enzymatic activity was assessed using the RfB trehalose assay detailed in section 2.2.3. Rv3802 activity was assessed using the 4MH assay detailed in section

4.2.3.2; however only 20 nM enzyme was used to slow reaction rates. THL stereoderivatives were provided from the O’Doherty lab at Northeastern University.

For initial screening THL and stereoderivatives were set at a working stock concentration of 30 mM in DMSO, for initial screening at 100 μM inhibitors were diluted to 10 mM in DMSO. Reactions components are given above for respective enzyme.

149

Upon titration of inhibitor to a reaction master mix sufficient for 8 reactions, triplicate reactions were quickly aliquoted into a black 384-well plate, respective fluorophore (RfB for Ag85C and 4MH for Rv3802) was titrated and the kinetic read acquired. This was considered time point “0 minutes.” Following an hour of incubation at room temperature, triplicate reactions were again aliquoted and titrated with RfB or 4MH and followed by kinetic data acquisition. This was considered time point “60 minutes.” Linear rates were assessed from time point 2 to 6 minutes. Following background rate subtraction % enzymatic activity was calculated by (V’/V’o)*100 where V’ = inhibited rate and V’o = uninhibited rate of DMSO control. Triplicate data were plotted and analyzed using

PRISM 7.

6.2.3 Enzymatic characterization

6.2.3.1 Ag85C kinact/Ki determination

To determine kinact/Ki values for THL and 2’-epi-THL inhibitors were serial diluted with DMSO from 30 mM to 1.25 mM (final reaction concentration of 300 to 12.5

μM). Reaction components are given above for Ag85C. A master mix solution of enzyme, trehalose, and buffer was aliquoted into a black 384-well plate followed by the addition of the respective concentration of inhibitor in triplicate. RfB was immediately titrated into reactions followed by kinetic read acquisition. Kinetic reads were taken every 30 seconds for 40 minutes. Following background subtraction and conversion of relative florescence units to concentration of product (resorufin) using a resorufin standard curve (0.1 to 6.0 μM), triplicate data were plotted using PRISM 7. The kobs value was determined by fitting the progress curves from time points 0 to 40 minutes with a one phase association equation Y = Y0 + (Plateau – Y0)(1-exp(-kx)); where x = 150

86 time, Y = [Product], (Y0 + (Plateau – Y0) = Vi/kobs, and k = kobs . Determined kobs were then plotted as a function of inhibitor concentration and fitted with kobs =

(kinact/(1+(Ki/[I]))) to calculate kinact/Ki values. Data was plotted and fit using PRISM 7.

6.2.3.2 Analysis of covalent inhibition of Rv3802

Covalent inhibition of Rv3802 by THL was determined by evaluating reaction progress curves of enzyme titrated with THL for time dependent inhibition. THL was diluted from 30 mM to 10 μM in DMSO and further serial diluted to 0.625 μM (final reaction concentration of 100 to 6.25 nM). A master mix of 50 nM enzyme (increase from given value above) and buffer was aliquoted into black 384-well plate followed by titration of serial diluted THL in triplicate. Immediately after, 4MH was titrated and kinetic read initiated. Following background subtraction, triplicate data were plotted in

PRISM 7. A similar approach was implemented with the remaining “epi” stereoderivatives. While progress curves for 5-epi-THL and 2’,5-epi-THL began to plateau as a function of time, indicative of covalent inhibition, they quickly transitioned to a phase exhibiting an increase in reaction rate over time. To directly compare the stability of covalent inhibition by THL and the “epi” stereoderivatives, compounds were diluted to 10 μM in DMSO and titrated in triplicate into a master mix of enzyme and buffer resulting in a final concentration of 100 nM inhibitor (same reaction conditions listed for the activity assay). Kinetic reads were initiated upon titration of 4MH.

Following background subtraction, data were plotted in PRISM 7.

6.2.3.3 Rv3802 IC50 determination

IC50 values were determined through the titration of serial diluted THL and respective derivative into aliquoted master mix of enzyme and buffer (same reaction 151

conditions listed activity assay). THL and “epi” stereoderivatives were diluted to 60 μM with DMSO and serial diluted to 0.625 μM (final reaction concentration range of 600 to

6.25 nM). For “ent” stereoderivatives, the 30 mM stocks were serial diluted to 3.125 mM

(final reaction concentration range of 300 to 3.125 μM). Following titration of respective derivative concentration, reads were immediately initiated by titration of 4MH and kinetic reads obtained. A linear rate was fit to data between time points 2 and 6 minutes.

Following background subtraction, % enzymatic activity was determined by

(V’/V’o)*100 where V’ = inhibited rate and V’o = uninhibited rate of DMSO control.

The % enzymatic activity was plotted as a function of inhibitor concentration and IC50 values determined by fitting the data with the “[Inhibitor] vs normalized response –

Variable slope” equation where y = 100/(1+(xHillslope)/(IC50Hillslope)) in PRISM 7.

6.2.4 Molecular modeling Rv3802 tetrahydrolipstatin complex

The atomic coordinates for Rv3802 (PDB: 5W95) and serine-THL from the

Ag85C-THL structure (PDB: 5VNS) were obtained from the PDB57, 135. Solvent and ligand molecules were removed from the Rv3802 structure and the serine THL aligned to the serine nucleophile and lipid-binding site of Rv3802 using PyMOL96. Resulting coordinates were inserted into the Rv3802 PDB file in place of the catalytic serine.

Rotatable bonds of the peptidyl arm were manually adjusted in COOT to fit within the solvent channel of Rv3802 and the resulting model subjected to simple model perturbation using Phenix Dynamics88, 89.

152

6.3 Results (Portions of text reproduced from Goins et al. ACS Biochemistry, 2018)159

6.3.1 Initial inhibition screening

The THL stereoderivatives tested are as follows: 2’-epi, 5-epi, 2’5,-epi being epimers of the respective centers of THL and their subsequent enantiomers (ent1 = enantiomer of THL, ent2 = enantiomer of 2’-epi, ent3 = enantiomer of 5-epi, ent4 = enantiomer of 2’5-epi). THL carbon centers are numbered in figure 6-1A. The inhibition potential of THL and its stereoderivatives was evaluated by screening each of the targeted enzymes against 100 μM of the respective inhibitor. To initially assess time- dependent covalent inhibition, inhibitors were screened with a 0 and 60 minutes preincubation period prior to kinetic reads (Figure 6-3).

Figure 6-3: Resulting percent enzymatic activity of Ag85C and Rv3802 following treatment with 100 μM of respective THL stereoderivative for 0 and 60 minutes of incubation. Figure adapted from Goins et al. ACS Biochemistry, 2018.159

Ag85C displays a high level of stereoselectivity preference towards THL. The only stereoderivative of THL that substantially inhibits the Ag85C is 2’-epi-THL, which is the most structurally similar stereoderivative to THL. Initial inhibition values are 31.8

± 5.9 and 35.8 ± 2.2 % for THL and 2’-epi-THL, respectively. However, inhibition levels 153

increase by ~56 % for both inhibitors following an hour incubation period, resulting in a near complete loss of enzymatic activity for both cases. Time-dependent inhibition is also observed for all other stereoderivatives, despite lower levels of inherent inhibition.

Specifically, 5-epi-THL increases in inhibition levels from 12.3 ± 2.6 to 57.8 ± 1.6 % following an hour of preincubation. A noticeable trend is apparent with regards to generally lower levels of inhibition for ent1-THL, ent2-THL, ent3-THL, and ent4-THL when compared to the respective epimer.

In contrast to Ag85C, Rv3802 appears to have less specificity with regards to

THL stereochemistry when screened at 100 μM. At this concentration THL, 2’-epi-THL,

5-epi-THL, and 2’,5-epi-THL all completely inhibit the enzyme with no preincubation time. However, the respective enantiomers of these derivatives show lower levels of inhibition, 85.1 ± 1.4, 83.9 ± 0.6, 83.0 ± 2.9, and 79.0 ± 1.7 % for ent1-THL, ent2-THL, ent3-THL, and ent4-THL, respectively; while minor, time-dependent inhibition is observed for the “ent” derivatives as inhibition levels increase between 4 to 6 % for these four derivatives given an hour preincubation period.

6.3.2 Ag85C inhibition

As stated above, the only stereoderivative that displayed significant and comparable inhibition to THL was 2’-epi-THL. Our previous work with Ag85C and THL indicated that the acyl-enzyme (THL-Ag85C) complex was considerably more stable than the complex of THL and FAS57. Due to the limited hydrolysis of the covalent complex, THL inhibition for Ag85C was previously characterized using a kinact/Ki analysis (Chapter 2)57. Given that covalent inhibition is both concentration and time dependent, this enzymatic analysis quantifies both components of inhibition86. Therefore, 154

an identical approach was used to assess the difference in binding affinity and rate of covalent inhibition between THL and 2’-epi-THL. The kinact/Ki value for THL and 2’-epi-

THL was experimentally determined to be 4.2 ± 0.7 and 1.1 ± 0.3 x 10-3 μM-1 min-1, respectively (Figure 6-4). The raw progress curves used to determine kobs is given in figure 6-5. Covalent inhibition occurs at a similar rate between THL and 2’-epi-THL.

However, the difference in stereochemistry of the formamide moiety at the 2’ carbon results in a ~4.5 fold lower binding affinity which translates to a ~4 fold lower kinact/Ki value for 2’-epi-THL compared to THL.

Figure: 6-4: kinact/Ki plot for THL and 2’-epi-THL. While inhibition occurs at a similar rate (kinact), Ag85C has a lower affinity (KI) towards 2’-epi-THL compared to THL, resulting in a lower kinact/Ki value. Data is plotted as a function of triplicate data, the mean value is shown with standard deviation error bars given. Figure reproduced from Goins et al. ACS Biochemistry, 2018159.

155

Figure: 6-5: Reaction progress curves of THL and 2’-epi-THL plotted as a function of product formation (Resorufin, μM) over time (Minutes). Data is shown as the average of triplicates with error bars given. Curves are fit with the one phase association equation 159 used to determine kobs. Figure reproduced from Goins et al. ACS Biochemistry, 2018 .

6.3.3 Rv3802 inhibition

Initial screening of Rv3802 at higher concentrations of THL suggests inhibition is relatively non-stereospecific; however, there was a noticeable difference when the stereocenters at the 2 and 3 carbons are changed. Given that Rv3802 is a serine esterase,

THL inhibition is believed to be covalent138. Therefore, THL inhibition of Rv3802 was initially attempted to be characterized using a kinact/KI analysis. Indeed, when THL was titrated into Rv3802 from 100 to 6.25 nM with no preincubation period, reaction progress curves that begin to plateau over time are observed, indicating covalent inhibition (Figure

6-6A). When THL was set at a 2:1 molar ratio, 100 nM THL to 50 nM enzyme, covalent inhibition appears to occur within minutes; however, the reaction progress curve never truly plateaus (Inlay of Figure 6-6A). When THL, 2’-epi-THL, 5-epi-THL, and 2’,5-epi-

THL were tested at a 5:1 molar ratio of inhibitor to enzyme, the same phenomenon is observed (Figure 6-6B). THL and 2’-epi-THL appear to plateau within 5 minutes; however, an increase in slope is observed over the following 15 minutes (Figure 6-6B).

This initial inhibition followed by an increase in signal is more evident with the 5-epi- 156

THL and 2’,5-epi-THL stereoderivatives (Figure 6-6B). Due to this phenomenon, a kinact/KI analysis was not feasible.

Figure 6-6: Reaction progress curves of Rv3802 inhibition by THL and “epi” stereoderivatives. A) Titration of Rv3802 with varying concentration of THL results in reaction progress curves that plateau as a function of time. Inlayed is a smaller scaled y- axis to show 50, 75, and 100 nM inhibition curves. Progress curves are shown as the average of triplicate reactions. B) Regeneration of Rv3802 activity is observed over time when the enzyme is reacted with 5:1 molar ratio (inhibitor to enzyme) of THL and respective “epi” stereoderivative, suggesting a hydrolysis of the acyl-enzyme covalent complex. Figure adapted from Goins et al. ACS Biochemistry, 2018159.

Therefore, to compare inhibition of Rv3802 by THL and its stereoderivatives,

IC50 values were determined using linear rates over time points 2 to 6 minutes with inhibition reactions consisting of 20 nM Rv3802. Determined IC50 values for all stereoderivatives are given in table 6.1, IC50 curves can be found in Figure 6-7. 2’-epi-

THL was found to have the best IC50 value of 9.2 ± 0.5 nM. Though, THL, 2’-epi-THL,

5-epi-THL, 2’,5-epi-THL all have comparable IC50 values in the low nM (9.2 to 28.7 nM)

157

concentration range. Interestingly, the “ent” series of derivatives possess IC50 values that are 3 orders of magnitude worse. These range from 15.9 to 62.1 μM. However, there is a correlation between the “epi” series and their respective enantiomers given that 2’,5-epi-

THL and its respective enantiomer, ent4-THL, have the highest IC50 values relative to the other “epi” and corresponding “ent” stereoderivatives.

Table 6.1: Determined IC50 values for THL and its stereoderivatives using 20 nM Rv3802. Table reproduced from Goins et al. ACS Biochemistry, 2018159. Inhibitor IC50 Value THL 16.2 ± 0.8 nM 2’-epi-THL 9.2 ± 0.5 nM 5-epi-THL 12.6 ± 1.2 nM 2’,5-epi-THL 28.7 ± 3.1 nM ent1-THL 15.9 ± 1.0 μM ent2-THL 15.9 ± 1.1 μM ent3-THL 16.6 ± 0.9 μM ent4-THL 62.1 ± 4.1 μM

Figure 6-7: IC50 curves for respective THL stereoderivative for Rv3802. Figure reproduced from Goins et al. ACS Biochemistry, 2018159.

158

6.4 Discussion (Portions of text reproduced from Goins et al. ACS Biochemistry,

2018)159

6.4.1 Ag85C inhibition

Initial screening of Ag85C against the THL stereoderivatives indicated the enzyme to be highly stereospecific with regard to THL (Figure 6-3). These findings are surprising given the apo Ag85C structure displays a large, hydrophobic cleft adjacent to the catalytic residues on the surface of the protein (Figure 6-8A). The hydrophobic cleft is thought to be the mycolic acid binding site and therefore should readily accommodate a variety of THL diastereomers given the large size of mycolic acid57, 80, 81. The alkyl chains of THL occupy this hydrophobic cleft as observed in the Ag85C-THL X-ray structure (Figure 6-8B)57. The Ag85C-THL structure highlights that THL binds in a shape-complimentary manner to the enzyme post covalent modification, which possess a hydrophobic pocket that is larger than the analogous pocket in the apo form of Ag85C57.

However, a change in stereochemistry at the C2’ position on the peptidyl arm reduces binding affinity as observed in the 2’-epi-THL stereoderivative (Figure 6-4). Yet, this small change in stereochemistry does not appear to influence the reversible nature of the covalent acyl-enzyme complex, most likely due to H260 being displaced in a similar manner to that of THL (Figure 6-8B).

159

Figure 6-8: Ag85C structural forms. A) A large hydrophobic cleft adjacent the active site is apparent in the apo Mtb Ag85C X-ray crystal structure, catalytic residues in orange (PDB: 1DQZ)80. The hydrophobic cleft is thought to be the mycolic acid-binding site of Ag85C57. B) The X-ray crystal structure of Ag85C in covalent complex with THL (yellow) allows for the molecular assessment of observed stereoselective inhibition by THL and resulting stability of the acyl-enzyme complex through displacement of the catalytic H260 (PDB: 5VNS)57. Figure reproduced from Goins et al. ACS Biochemistry, 2018159.

Given the stereoselective inhibition of Ag85C, these results suggest that initial

THL binding to Ag85C is also highly shape-specific. This is best illustrated when the stereochemistry of the C5 carbon is switched. This change in stereochemistry would likely position the peptidyl side arm into the hydrophobic cleft where the palmitic tail resides, reducing initial binding affinity (Figure 6-8B). Indeed, a significant reduction in inhibition level is observed with the 5-epi-THL stereoderivative. More so, when the stereocenters are swapped at the C2 and C3 positions, inhibition by THL and subsequent

“ent” derivatives is nearly abolished. Additionally, lower levels of enzymatic activity are observed post hour incubation, suggesting that these inhibitors retain their covalent attachment to the enzyme. Based on the crystal structure of Ag85C-THL covalent complex, there is sufficient volume in the active site to accommodate these changes57.

However, the lack of apparent inhibition by “ent” derivatives suggests that the

160

stereochemistry of the β-lactone ring with respect to the rest of the THL molecule is highly important for proper orientation of the carbonyl electrophile of THL in the active site to promote nucleophilic attack by the enzyme. This is likely the reason why the “ent” series derivatives, which have the same stereochemistry at the C2 and C3 carbons as that of mycolic acid, the natural Ag85C substrate, are poorer inhibitors than those derivatives possessing alternative stereochemistry45.

6.4.2 Rv3802 inhibition

Previous studies have reported low μM IC50 values and an apparent KI of 5 nM for THL to Rv3802136, 138. Therefore, it is of little surprise that we see almost no enzymatic activity when THL is screened at 100 μM; however, we were uncertain as to how stereochemistry would influence inhibition. Interestingly, all “epi” derivatives displayed IC50 values in the low nM range, suggesting the enzyme is capable of binding a large variety of structurally diverse THL configurations (Table 6.1). This may be attributed to the highly dynamic helix that defines the lipid-binding site of the enzyme

(Figure 6-9A)135. When this helix is in the open conformation a sizable hydrophobic channel leading to the active site is observed (Figure 6-9B)135. THL was modeled with little manipulation into the hydrophobic channel with room found for the peptidyl side arm in a solvent channel that is perpendicular to the lipid-binding site (Figure 6-9C).

Positioning of the peptidyl side arm into this cannel positions the arm near the catalytic histidine and may explain the difference in observed reversible covalent inhibition between the “epi” derivatives (Figure 6-6B). In the case of THL and 2’-epi-THL, the side arm may be partially displacing H299 similarly to that of the analogous histidine in

Ag85C. Alternatively, THL and 2’-epi-THL may be binding in a fashion that allows a 161

stable hydrogen bond to form between H299 and the β–hydroxyl that forms on C3 of

THL following covalent complex formation similar to that observed in the THL and FAS enzyme84. In either possibility, water activation by the conserved histidine is reduced.

However, when the stereochemistry of the peptidyl arm is changed, 5-epi-THL and 2’,5- epi-THL, the interaction that may otherwise reduce water activation is abolished and hydrolysis of the acyl-enzyme rapidly occurs. Such a binding mode would therefore explain the differences in observed regeneration of enzymatic activity between “epi” derivatives (Figure 6-6B). Additionally, this binding model would explain the inhibition differences observed in a prior study, which synthesized a library of THL peptidyl arm derivatives. THL derivatives were mentioned in chapter 5. Briefly, West et al. found that the smaller the side arm derivative, the better the inhibition; given that the solvent channel is more sterically hindered than the lipid binding site, this binding mode may explain their reported SAR study (THL derivative depicted in Figure 5-1)153.

162

Figure 6-9: Modeled Rv3802-THL complex. A) When the dynamic helix of Rv3802 is in the open form, a large hydrophobic channel leading to the catalytic residues (orange) is apparent. This channel is believed to be the fatty acid binding site for phosphatidylinositol substrates (PDB: 5W95)135. B) The Rv3802 surface rendering is clipped to highlight the size of the hydrophobic channel and to show a solvent channel leading to the active site135. C) Resulting Rv3802-THL covalent complex model based on stereoderivative inhibition data. THL is readily accommodated within the lipid binding site and solvent channel of Rv3802. Figure reproduced from Goins et al. ACS Biochemistry, 2018159.

Despite Rv3802 being able to tightly bind the structurally diverse “epi” THL derivatives, the “ent” derivatives had IC50 values three orders of magnitude worse.

Changes in stereochemistry at the C2 and C3 carbons of the β-lactone ring greatly reduce inhibition similar to that observed for Ag85C. This stereospecific trend is most likely not based on a form of substrate mimicry, as Rv3802 is believed to perform chemistry on glycerophospholipids not mycolic acid containing substrates57, 137, 138. This trend would therefore be supported by the previous argument that the stereochemistry of the β-lactone ring is important for proper alignment of the β-lactone ring of THL to the enzyme for nucleophilic attack to occur by the serine nucleophile.

163

6.4.3 Influence of stereochemistry on cross reactivity, specificity, and stability

Stereochemistry of THL is important with regard to the inhibition of Ag85C and

Rv3802. However, it is not the direct result of simple mycolic acid mimicry as the preferred stereochemistry of THL is opposite that of the hallmark lipid. Mostly, the stereochemistry of the β-lactone ring of THL was found to be highly important for cross- enzyme covalent inhibition and this efficacy is likely due to the proper positioning of the covalent warhead for nucleophilic attack by the enzyme. However, the hexanoyl tail and palmitic core of THL most likely afford general affinity to the lipid binding sites of lipid esterases. A major limitation of the THL scaffold and β-lactone based inhibitors is the reversible nature of covalent inhibition. Upon hydrolysis off the enzyme, THL is rendered inactive for further inhibition. Neutralization of THL upon ring opening can be compared to the neutralization of potent β-lactam compounds by β-lactamases and carbapenumases161, 162, 165, 166. Therefore, to increase the efficacy of β-lactone based inhibitors, the stability of the covalent complex needs to be addressed. In the case of

THL, the peptidyl arm may be derivatized to enhance the stability of the acyl-enzyme complex.

6.5 Conclusions

A library of THL stereoderivatives were tested against two major lipid esterases of Mtb to evaluate how the shape of THL influence cross reactivity, specificity, and stability of the resulting covalent complex. Ag85C was found to be highly stereospecific, while Rv3802 was not. Covalent inhibition of Ag85C is far more stable than that of

Rv3802, which was observed to be prone to hydrolysis. The two stereocenters of the β- 164

lactone ring of THL influence cross enzyme reactivity as a consequence of geometric requirements for nucleophilic attack. The stereocenters of the peptidyl arm were found to influence enzyme selectivity and stability of the acyl-enzyme intermediate. While the two alkyl chains of THL most likely afford general affinity towards Mtb lipid esterases, the unique stereochemistry is required for subsequent covalent inhibition.

165

References

1. World Health Organization (2016) 2016 Global Tuberculosis Report. 2. Pai, M., Behr, M. A., Dowdy, D., Dheda, K., Divangahi, M., Boehme, C. C., Ginsberg, A., Swaminathan, S., Spigelman, M., Getahun, H., Menzies, D., and Raviglione, M. (2016) Tuberculosis, Nature Reviews Disease Primers 2, 16076. 3. Havlir, D. V., Getahun, H., Sanne, I., and Nunn, P. (2008) Opportunities and challenges for HIV care in overlapping HIV and TB epidemics, JAMA 300, 423- 430. 4. Donoghue, H. D., Spigelman, M., Greenblatt, C. L., Lev-Maor, G., Bar-Gal, G. K., Matheson, C., Vernon, K., Nerlich, A. G., and Zink, A. R. (2004) Tuberculosis: from prehistory to Robert Koch, as revealed by ancient DNA, The Lancet infectious diseases 4, 584-592. 5. Sakamoto, K. (2012) The pathology of Mycobacterium tuberculosis infection, Veterinary pathology 49, 423-439. 6. Orme, I. M., Robinson, R. T., and Cooper, A. M. (2015) The balance between protective and pathogenic immune responses in the TB-infected lung, Nature immunology 16, 57-63. 7. Ferguson, J. S., Voelker, D. R., McCormack, F. X., and Schlesinger, L. S. (1999) Surfactant protein D binds to Mycobacterium tuberculosis Bacilli and Lipoarabinomannan via carbohydrate-lectin interactions resulting in reduced phagocytosis of the bacteria by macrophages1, The Journal of immunology 163, 312-321. 8. Wolf, A. J., Desvignes, L., Linas, B., Banaiee, N., Tamura, T., Takatsu, K., and Ernst, J. D. (2008) Initiation of the adaptive immune response to Mycobacterium tuberculosis depends on antigen production in the local lymph node, not the lungs, Journal of Experimental Medicine 205, 105-115. 9. Lin, P. L., Ford, C. B., Coleman, M. T., Myers, A. J., Gawande, R., Ioerger, T., Sacchettini, J., Fortune, S. M., and Flynn, J. L. (2014) Sterilization of granulomas is common in active and latent tuberculosis despite within-host variability in bacterial killing, Nature medicine 20, 75-79. 10. Horsburgh Jr, C. R., Barry III, C. E., and Lange, C. (2015) Treatment of tuberculosis, New England Journal of Medicine 373, 2149-2160. 11. Zumla, A., Nahid, P., and Cole, S. T. (2013) Advances in the development of new tuberculosis drugs and treatment regimens, Nature reviews Drug discovery 12, 388-404.

166

12. Marshall, G., Blacklock, J., Cameron, C., Capon, N., Cruickshank, R., Gaddum, J., Heaf, F., Hill, A. B., Houghton, L., and Hoyle, J. C. (1948) Streptomycin treatment of pulmonary tuberculosis: a Medical Research Council investigation, British Medical Journal 2, 769-782. 13. Johnson, J. L., Hadad, D. J., Dietze, R., Noia Maciel, E. L., Sewali, B., Gitta, P., Okwera, A., Mugerwa, R. D., Alcaneses, M. R., and Quelapio, M. I. (2009) Shortening treatment in adults with noncavitary tuberculosis and 2-month culture conversion, American journal of respiratory and critical care medicine 180, 558- 563. 14. Blumberg, H. M., and Ernst, J. D. (2016) The Challenge of Latent TB Infection, JAMA 316, 931-933. 15. Banerjee, A., Dubnau, E., Quemard, A., Balasubramanian, V., Um, K. S., Wilson, T., Collins, D., De Lisle, G., and Jacobs, W. R. (1994) inhA, a gene encoding a target for isoniazid and ethionamide in Mycobacterium tuberculosis, Science 263, 227- 229. 16. Loewenberg, S. (2012) India reports cases of totally drug-resistant tuberculosis, The Lancet 379, 205. 17. Udwadia, Z. F., Amale, R. A., Ajbani, K. K., and Rodrigues, C. (2012) Totally drug- resistant tuberculosis in India, Clinical Infectious Diseases 54, 579-581. 18. Dooley, K. E., Mitnick, C. D., Ann DeGroote, M., Obuku, E., Belitsky, V., Hamilton, C. D., Makhene, M., Shah, S., Brust, J. C., and Durakovic, N. (2012) Old drugs, new purpose: retooling existing drugs for optimized treatment of resistant tuberculosis, Clinical Infectious Diseases 55, 572-581. 19. Falzon, D., Jaramillo, E., Schünemann, H., Arentz, M., Bauer, M., Bayona, J., Blanc, L., Caminero, J., Daley, C., and Duncombe, C. (2011) WHO guidelines for the programmatic management of drug-resistant tuberculosis: 2011 update, European Respiratory Society 20. Jacobson, K. R., Tierney, D. B., Jeon, C. Y., Mitnick, C. D., and Murray, M. B. (2010) Treatment outcomes among patients with extensively drug-resistant tuberculosis: systematic review and meta-analysis, Clinical infectious diseases 51, 6-14. 21. Glickman, M. S., and Jacobs, W. R. (2001) Microbial pathogenesis of Mycobacterium tuberculosis: dawn of a discipline, Cell 104, 477-485. 22. Bishop, P., and Neumann, G. (1970) The history of the Ziehl-Neelsen stain, Tubercle 51, 196-206. 23. Jackson, M. (2014) The mycobacterial cell envelope—Lipids, Cold Spring Harbor perspectives in medicine 4, a021105. 24. Hoffmann, C., Leis, A., Niederweis, M., Plitzko, J. M., and Engelhardt, H. (2008) Disclosure of the mycobacterial outer membrane: cryo-electron tomography and vitreous sections reveal the lipid bilayer structure, Proceedings of the National Academy of Sciences 105, 3963-3967. 25. Velayati, A. A., Farnia, P., Ibrahim, T. A., Haroun, R. Z., Kuan, H. O., Ghanavi, J., Farnia, P., Kabarei, A. N., Tabarsi, P., and Omar, A. R. (2009) Differences in cell wall thickness between resistant and nonresistant strains of Mycobacterium tuberculosis: using transmission electron microscopy, Chemotherapy 55, 303-307. 167

26. Korduláková, J., Gilleron, M., Mikusõ vá, K. n., Puzo, G., Brennan, P. J., Gicquel, B., and Jackson, M. (2002) Definition of the First Mannosylation Step in Phosphatidylinositol Mannoside Synthesis PimA IS ESSENTIAL FOR GROWTH OF MYCOBACTERIA, Journal of Biological Chemistry 277, 31335- 31344. 27. Boonyarattanakalin, S., Liu, X., Michieletti, M., Lepenies, B., and Seeberger, P. H. (2008) Chemical synthesis of all phosphatidylinositol mannoside (PIM) glycans from Mycobacterium tuberculosis, Journal of the American Chemical Society 130, 16791-16799. 28. Bansal-Mutalik, R., and Nikaido, H. (2014) Mycobacterial outer membrane is a lipid bilayer and the inner membrane is unusually rich in diacyl phosphatidylinositol dimannosides, Proceedings of the National Academy of Sciences 111, 4958-4963. 29. Pitarque, S., Larrouy-Maumus, G., Payré, B., Jackson, M., Puzo, G., and Nigou, J. (2008) The immunomodulatory lipoglycans, lipoarabinomannan and lipomannan, are exposed at the mycobacterial cell surface, Tuberculosis 88, 560-565. 30. Kaur, D., Guerin, M. E., Skovierova, H., Brennan, P. J., and Jackson, M. (2009) Biogenesis of the cell wall and other glycoconjugates of Mycobacterium tuberculosis, Advances in Applied Microbiology 69, 23-78. 31. Rath, P., Saurel, O., Czaplicki, G., Tropis, M., Daffé, M., Ghazi, A., Demange, P., and Milon, A. (2013) Cord factor (trehalose 6, 6′-dimycolate) forms fully stable and non-permeable lipid bilayers required for a functional outer membrane, Biochimica et Biophysica Acta (BBA)-Biomembranes 1828, 2173-2181. 32. Chiaradia, L., Lefebvre, C., Parra, J., Marcoux, J., Burlet-Schiltz, O., Etienne, G., Tropis, M., and Daffé, M. (2017) Dissecting the mycobacterial cell envelope and defining the composition of the native mycomembrane, Scientific reports 7, 12807. 33. Draper, P., and Daffé, M. (2005) The cell envelope of Mycobacterium tuberculosis with special reference to the capsule and outer permeability barrier, In Tuberculosis and the tubercle bacillus, American Society of Microbiology, 261- 274. 34. Lemassu, A., and Daffé, M. (1994) Structural features of the exocellular polysaccharides of Mycobacterium tuberculosis, Biochemical Journal 297, 351- 357. 35. Ortalo-Magne, A., Dupont, M. A., Lemassu, A., Andersen, A. B., Gounon, P., and Mamadou, D. (1995) Molecular composition of the outermost capsular material of the tubercle bacillus, Microbiology 141, 1609-1620. 36. Sonnenberg, M. G., and Belisle, J. T. (1997) Definition of Mycobacterium tuberculosis culture filtrate proteins by two-dimensional polyacrylamide gel electrophoresis, N-terminal amino acid sequencing, and electrospray mass spectrometry, Infection and Immunity 65, 4515-4524. 37. Villeneuve, C., Gilleron, M., Maridonneau-Parini, I., Daffé, M., Astarie-Dequeker, C., and Etienne, G. (2005) Mycobacteria use their surface-exposed glycolipids to infect human macrophages through a receptor-dependent process, Journal of lipid research 46, 475-483.

168

38. Neyrolles, O., and Guilhot, C. (2011) Recent advances in deciphering the contribution of Mycobacterium tuberculosis lipids to pathogenesis, Tuberculosis 91, 187-195. 39. Venisse, A., Fournié, J. J., and Puzo, G. (1995) Mannosylated lipoarabinomannan interacts with phagocytes, The FEBS Journal 231, 440-447. 40. Ly, D., Kasmar, A. G., Cheng, T.-Y., de Jong, A., Huang, S., Roy, S., Bhatt, A., van Summeren, R. P., Altman, J. D., and Jacobs, W. R. (2013) CD1c tetramers detect ex vivo T cell responses to processed phosphomycoketide antigens, Journal of Experimental Medicine 210, 729-741 41. Fischer, K., Scotet, E., Niemeyer, M., Koebernick, H., Zerrahn, J., Maillet, S., Hurwitz, R., Kursar, M., Bonneville, M., and Kaufmann, S. H. (2004) Mycobacterial phosphatidylinositol mannoside is a natural antigen for CD1d- restricted T cells, Proceedings of the National Academy of Sciences of the United States of America 101, 10685-10690. 42. Wiker, H. G., and Harboe, M. (1992) The antigen 85 complex: a major secretion product of Mycobacterium tuberculosis, Microbiological reviews 56, 648-661. 43. Beatty, W. L., and Russell, D. G. (2000) Identification of mycobacterial surface proteins released into subcellular compartments of infected macrophages, Infection and immunity 68, 6997-7002. 44. Belisle, J. T., Vissa, V. D., Sievert, T., Takayama, K., Brennan, P. J., and Besra, G. S. (1997) Role of the Major Antigen of Mycobacterium tuberculosis in Cell Wall Biogenesis, Science 276, 1420-1422. 45. Marrakchi, H., Laneelle, M. A., and Daffe, M. (2014) Mycolic acids: structures, biosynthesis, and beyond, Chemistry & biology 21, 67-85. 46. Bansal-Mutalik, R., and Nikaido, H. (2011) Quantitative lipid composition of cell envelopes of Corynebacterium glutamicum elucidated through reverse micelle extraction, Proceedings of the National Academy of Sciences 108, 15360-15365. 47. Barry, C. E., Lee, R. E., Mdluli, K., Sampson, A. E., Schroeder, B. G., Slayden, R. A., and Yuan, Y. (1998) Mycolic acids: structure, biosynthesis and physiological functions, Progress in lipid research 37, 143-179. 48. Liu, J., Barry, C. E., Besra, G. S., and Nikaido, H. (1996) Mycolic acid structure determines the fluidity of the mycobacterial cell wall, Journal of Biological Chemistry 271, 29545-29551. 49. Glickman, M. S., Cox, J. S., and Jacobs, W. R. (2000) A novel mycolic acid cyclopropane synthetase is required for cording, persistence, and virulence of Mycobacterium tuberculosis, Molecular cell 5, 717-727. 50. Fernandes, N. D., and Kolattukudy, P. E. (1996) Cloning, sequencing and characterization of a fatty acid synthase-encoding gene from Mycobacterium tuberculosis var. bovis BCG, Gene 170, 95-99. 51. Portevin, D., de Sousa-D'Auria, C., Montrozier, H., Houssin, C., Stella, A., Lanéelle, M.-A., Bardou, F., Guilhot, C., and Daffé, M. (2005) The Acyl-AMP Ligase FadD32 and AccD4-containing Acyl-CoA Carboxylase Are Required for the Synthesis of Mycolic Acids and Essential for Mycobacterial Growth IDENTIFICATION OF THE CARBOXYLATION PRODUCT AND DETERMINATION OF THE ACYL-CoA CARBOXYLASE COMPONENTS, Journal of Biological Chemistry 280, 8862-8874. 169

52. Portevin, D., de Sousa-D'Auria, C., Houssin, C., Grimaldi, C., Chami, M., Daffé, M., and Guilhot, C. (2004) A polyketide synthase catalyzes the last condensation step of mycolic acid biosynthesis in mycobacteria and related organisms, Proceedings of the National Academy of Sciences 101, 314-319. 53. Gavalda, S., Bardou, F., Laval, F., Bon, C., Malaga, W., Chalut, C., Guilhot, C., Mourey, L., Daffé, M., and Quémard, A. (2014) The polyketide synthase Pks13 catalyzes a novel mechanism of lipid transfer in mycobacteria, Chemistry & biology 21, 1660-1669. 54. Lea-Smith, D. J., Pyke, J. S., Tull, D., McConville, M. J., Coppel, R. L., and Crellin, P. K. (2007) The reductase that catalyzes mycolic motif synthesis is required for efficient attachment of mycolic acids to arabinogalactan, Journal of Biological Chemistry 282, 11000-11008. 55. Grzegorzewicz, A. E., Pham, H., Gundi, V. A., Scherman, M. S., North, E. J., Hess, T., Jones, V., Gruppo, V., Born, S. E., Korduláková, J., Chavadi, S. S., Morisseau, C., Lenaerts, A. J., Lee, R. E., McNeil, M. R., and Jackson, M. (2012) Inhibition of mycolic acid transport across the Mycobacterium tuberculosis plasma membrane, Nature chemical biology 8, 334-341. 56. Sanki, A. K., Boucau, J., Ronning, D. R., and Sucheck, S. J. (2009) Antigen 85C- mediated acyl-transfer between synthetic acyl donors and fragments of the arabinan, Glycoconjugate journal 26, 589-596. 57. Goins, C. M., Dajnowicz, S., Smith, M. D., Parks, J. M., and Ronning, D. R. (2018) Mycolyltransferase from Mycobacterium tuberculosis in covalent complex with tetrahydrolipstatin provides insights into Antigen 85 catalysis, Journal of Biological Chemistry, jbc. RA117. 001681. 58. Crellin, P. K., Luo, C. Y., and Morita, Y. S. (2013) Metabolism of plasma membrane lipids in mycobacteria and corynebacteria, In Lipid Metabolism, InTech. 59. Gago, G., Diacovich, L., Arabolaza, A., Tsai, S. C., and Gramajo, H. (2011) Fatty acid biosynthesis in actinomycetes, FEMS microbiology reviews 35, 475-497. 60. Okuyama, H., Kameyama, Y., Fujikawa, M., Yamada, K., and Ikezawa, H. (1977) Mechanism of diacylglycerophosphate synthesis in mycobacteria, Journal of Biological Chemistry 252, 6682-6686. 61. Okuyama, H., Kankura, T., and Nojima, S. (1967) Positional distribution of fatty acids in phospholipids from Mycobacteria, The Journal of Biochemistry 61, 732- 737. 62. Nigou, J., and Besra, G. S. (2002) Cytidine diphosphate-diacylglycerol synthesis in Mycobacterium smegmatis, Biochemical Journal 367, 157-162. 63. Mathur, A., Murthy, P. S., Saharia, G., and Venkitasubramanian, T. (1976) Studies on cardiolipin biosynthesis in Mycobacterium smegmatis, Canadian journal of microbiology 22, 354-358. 64. Salman, M., Lonsdale, J. T., Besra, G. S., and Brennan, P. J. (1999) Phosphatidylinositol synthesis in mycobacteria, Biochimica et Biophysica Acta (BBA)-Molecular and Cell Biology of Lipids 1436, 437-450. 65. Guerin, M. E., Kaur, D., Somashekar, B., Gibbs, S., Gest, P., Chatterjee, D., Brennan, P. J., and Jackson, M. (2009) New Insights into the Early Steps of Phosphatidylinositol Mannoside Biosynthesis in Mycobacteria PimB′ IS AN 170

ESSENTIAL ENZYME OF MYCOBACTERIUM SMEGMATIS, Journal of Biological Chemistry 284, 25687-25696. 66. Albesa-Jové, D., Svetlíková, Z., Tersa, M., Sancho-Vaello, E., Carreras-González, A., Bonnet, P., Arrasate, P., Eguskiza, A., Angala, S. K., and Cifuente, J. O. (2016) Structural basis for selective recognition of acyl chains by the membrane- associated acyltransferase PatA, Nature communications 7, 10906. 67. Guerin, M. E., Korduláková, J., Alzari, P. M., Brennan, P. J., and Jackson, M. (2010) Molecular basis of phosphatidyl-myo-inositol mannoside biosynthesis and regulation in mycobacteria, Journal of Biological Chemistry 285, 33577-33583. 68. Bugg, T. D. (2004) Diverse catalytic activities in the αβ-hydrolase family of enzymes: activation of H 2 O, HCN, H 2 O 2, and O 2, Bioorganic chemistry 32, 367-375. 69. Jochens, H., Hesseler, M., Stiba, K., Padhi, S. K., Kazlauskas, R. J., and Bornscheuer, U. T. (2011) Protein engineering of α/β‐hydrolase fold enzymes, Chembiochem 12, 1508-1517. 70. Ollis, D. L., Cheah, E., Cygler, M., Dijkstra, B., Frolow, F., Franken, S. M., Harel, M., Remington, S. J., Silman, I., and Schrag, J. (1992) The α/β hydrolase fold, Protein Engineering, Design and Selection 5, 197-211. 71. Franken, S. M., Rozeboom, H. J., Kalk, K. H., and Dijkstra, B. W. (1991) Crystal structure of haloalkane dehalogenase: an enzyme to detoxify halogenated alkanes, The EMBO journal 10, 1297. 72. Röttig, A., and Steinbüchel, A. (2013) Acyltransferases in bacteria, Microbiology and Molecular Biology Reviews 77, 277-321. 73. Ravindran, M. S., Rao, S. P., Cheng, X., Shukla, A., Cazenave-Gassiot, A., Yao, S. Q., and Wenk, M. R. (2014) Targeting lipid esterases in mycobacteria grown under different physiological conditions using activity-based profiling with tetrahydrolipstatin (THL), Molecular & cellular proteomics 13, 435-448. 74. Armitige, L. Y., Jagannath, C., Wanger, A. R., and Norris, S. J. (2000) Disruption of the genes encoding antigen 85A and antigen 85B ofMycobacterium tuberculosis H37Rv: effect on growth in culture and in macrophages, Infection and immunity 68, 767-778. 75. Harth, G., Lee, B. Y., Wang, J., Clemens, D. L., and Horwitz, M. A. (1996) Novel insights into the genetics, biochemistry, and immunocytochemistry of the 30- kilodalton major extracellular protein of Mycobacterium tuberculosis, Infection and immunity 64, 3038-3047. 76. Abou-Zeid, C., Ratliff, T., Wiker, H., Harboe, M., Bennedsen, J., and Rook, G. (1988) Characterization of fibronectin-binding antigens released by Mycobacterium tuberculosis and Mycobacterium bovis BCG, Infection and immunity 56, 3046-3051. 77. Backus, K. M., Dolan, M. A., Barry, C. S., Joe, M., McPhie, P., Boshoff, H. I., Lowary, T. L., Davis, B. G., and Barry, C. E., 3rd. (2014) The three Mycobacterium tuberculosis antigen 85 isoforms have unique substrates and activities determined by non-active site regions, The Journal of biological chemistry 289, 25041-25053. 78. Jackson, M., Raynaud, C., Laneelle, M. A., Guilhot, C., Laurent-Winter, C., Ensergueix, D., Gicquel, B., and Daffe ́, M. (1991) Inactivation of the antigen 171

85C gene profoundly affects the mycolate content and alters the permeability of the Mycobacterium tuberculosis cell envelope, Molecular Microbiology 31, 1537 - 1587. 79. Nguyen, L., Chinnapapagari, S., and Thompson, C. J. (2005) FbpA-dependent biosynthesis of trehalose dimycolate is required for the intrinsic multidrug resistance, cell wall structure, and colonial morphology of Mycobacterium smegmatis, Journal of bacteriology 187, 6603-6611. 80. Ronning, D. R., Klabunde, T., Besra, G. S., Vissa, V. D., Belisle, J. T., and Sacchettini, J. C. (2000) Crystal structure of the secreted form of antigen 85C reveals potential targets for mycobacterial drugs and vaccines, Nature Structural Biology 7, 141 - 146. 81. Anderson, D. H., Harth, G., Horwitz, M. A., and Eisenberg, D. (2001) An interfacial mechanism and a class of inhibitors inferred from two crystal structures of the Mycobacterium tuberculosis 30 kDa major secretory protein (Antigen 85B), a mycolyl transferase, Journal of molecular biology 307, 671-681. 82. Ronning, D. R., Vissa, V. D., Besra, G. S., Belisle, J. T., and Sacchettini, J. C. (2004) Mycobacterium tuberculosis antigen 85A and 85C structures confirm binding orientation and conserved substrate specificity, The Journal of biological chemistry 279, 36771-36777. 83. Dunitz, J., Lehn, J., and Wipff, G. (1974) Stereochemistry of reaction paths at carbonyl centres, Tetrahedron 30, 1563-1572. 84. Pemble, C. W., Johnson, L. C., Kridel, S. J., and Lowther, W. T. (2007) Crystal structure of the thioesterase domain of human fatty acid synthase inhibited by Orlistat, Nature structural & molecular biology 14, 704-709. 85. Favrot, L., Grzegorzewicz, A. E., Lajiness, D. H., Marvin, R. K., Boucau, J., Isailovic, D., Jackson, M., and Ronning, D. R. (2013) Mechanism of inhibition of Mycobacterium tuberculosis antigen 85 by ebselen, Nature communications 4, 2748. 86. Copelan, R. A. (2005) Evaluation of Enzyme Inhibitors in Drug Discovery, John Wiley & Sons. 87. Otwinowski, Z., and Minor, W. (1997) Processing of X-ray Diffraction Data Collected in Oscillation Mode, Methods in Enzymology 276, 307-326. 88. Adams, P. D., Afonine, P. V., Bunkoczi, G., Chen, V. B., Davis, I. W., Echols, N., Headd, J. J., Hung, L. W., Kapral, G. J., Grosse-Kunstleve, R. W., McCoy, A. J., Moriarty, N. W., Oeffner, R., Read, R. J., Richardson, D. C., Richardson, J. S., Terwilliger, T. C., and Zwart, P. H. (2010) PHENIX: a comprehensive Python- based system for macromolecular structure solution, Acta crystallographica. Section D, Biological crystallography 66, 213-221. 89. Emsley, P., Lohkamp, B., Scott, W. G., and Cowtan, K. (2010) Features and development of Coot, Acta crystallographica. Section D, Biological crystallography 66, 486-501. 90. Moriarty, N. W., Grosse-Kunstleve, R. W., and Adams, P. D. (2009) electronic Ligand Builder and Optimization Workbench (eLBOW): a tool for ligand coordinate and restraint generation, Acta crystallographica. Section D, Biological crystallography 65, 1074-1080. 172

91. Terwilliger, T. C., Klei, H., Adams, P. D., Moriarty, N. W., and Cohn, J. D. (2006) Automated ligand fitting by core-fragment fitting and extension into density, Acta Crystallographica Section D: Biological Crystallography 62, 915-922. 92. Chen, V. B., Arendall, W. B., Headd, J. J., Keedy, D. A., Immormino, R. M., Kapral, G. J., Murray, L. W., Richardson, J. S., and Richardson, D. C. (2010) MolProbity: all-atom structure validation for macromolecular crystallography, Acta Crystallographica Section D: Biological Crystallography 66, 12-21. 93. Favrot, L., Lajiness, D. H., and Ronning, D. R. (2014) Inactivation of the Mycobacterium tuberculosis antigen 85 complex by covalent, allosteric inhibitors, The Journal of biological chemistry 289, 25031-25040. 94. Sievers, F., Wilm, A., Dineen, D., Gibson, T. J., Karplus, K., Li, W., Lopez, R., McWilliam, H., Remmert, M., and Söding, J. (2011) Fast, scalable generation of high‐quality protein multiple sequence alignments using Clustal Omega, Molecular systems biology 7, 539. 95. Crooks, G. E., Hon, G., Chandonia, J.-M., and Brenner, S. E. (2004) WebLogo: a sequence logo generator, Genome research 14, 1188-1190. 96. The PyMOL Molecular Graphics System, Version 1.6 Schrödinger, LLC. 97. Borgström, B. (1988) Mode of action of tetrahydrolipstatin: a derivative of the naturally occurring lipase inhibitor lipstatin, Biochimica et Biophysica Acta (BBA)-Lipids and Lipid Metabolism 962, 308-316. 98. Goins, C. M., Dajnowicz, S., Thanna, S., Sucheck, S. J., Parks, J. M., and Ronning, D. R. (2017) Exploring covalent allosteric inhibition of antigen 85C from Mycobacterium tuberculosis by ebselen derivatives, ACS Infectious Diseases 3, 378-387. 99. Weibel, E., Hadvary, P., Hochuli, E., Kupfer, E., and Lengsfeld, H. (1987) Lipstatin, an inhibitor of pancreatic lipase, produced by Streptomyces toxytricini, The Journal of antibiotics 40, 1081-1085. 100. Hadváry, P., Lengsfeld, H., and Wolfer, H. (1988) Inhibition of pancreatic lipase in vitro by the covalent inhibitor tetrahydrolipstatin, Biochemical Journal 256, 357- 361. 101. Kridel, S. J., Axelrod, F., Rozenkrantz, N., and Smith, J. W. (2004) Orlistat is a novel inhibitor of fatty acid synthase with antitumor activity, Cancer Research 64, 2070-2075. 102. Thanna, S., Goins, C. M., Knudson, S. E., Slayden, R. A., Ronning, D. R., and Sucheck, S. J. (2017) Thermal and Photoinduced Copper-Promoted C–Se Bond Formation: Synthesis of 2-Alkyl-1, 2-benzisoselenazol-3 (2 H)-ones and Evaluation against Mycobacterium tuberculosis, The Journal of Organic Chemistry 82, 3844-3854. 103. Fako, V. E., Zhang, J. T., and Liu, J. Y. (2014) Mechanism of orlistat hydrolysis by the thioesterase of human fatty acid synthase, ACS catalysis 4, 3444-3453. 104. Sanbonmatsu, K., and Garcia, A. (2002) Structure of Met‐enkephalin in explicit aqueous solution using replica exchange molecular dynamics, Proteins: Structure, Function, and Bioinformatics 46, 225-234.

173

105. Zhang, W., Wu, C., and Duan, Y. (2005) Convergence of replica exchange molecular dynamics, The Journal of chemical physics 123, 154105. 106. Periole, X., and Mark, A. E. (2007) Convergence and sampling efficiency in replica exchange simulations of peptide folding in explicit solvent, The Journal of chemical physics 126, 01B601. 107. Fisher, S. Z., Kovalevsky, A. Y., Domsic, J. F., Mustyakimov, M., McKenna, R., Silverman, D. N., and Langan, P. A. (2009) Neutron structure of human carbonic anhydrase II: implications for proton transfer, Biochemistry 49, 415-421. 108. Chen, L., Kong, X., Liang, Z., Ye, F., Yu, K., Dai, W., Wu, D., Luo, C., and Jiang, H. (2011) Theoretical study of the mechanism of proton transfer in the esterase EstB from Burkholderia gladioli, The Journal of Physical Chemistry B 115, 13019-13025. 109. Tomanicek, S. J., Standaert, R. F., Weiss, K. L., Ostermann, A., Schrader, T. E., Ng, J. D., and Coates, L. (2013) Neutron and X-ray crystal structures of a perdeuterated enzyme inhibitor complex reveal the catalytic proton network of the Toho-1 β-lactamase for the acylation reaction, Journal of Biological Chemistry 288, 4715-4722. 110. Sassetti, C. M., Boyd, D. H., and Rubin, E. J. (2003) Genes required for mycobacterial growth defined by high density mutagenesis, Molecular microbiology 48, 77-84. 111. Harth, G., Zamecnik, P. C., Tabatadze, D., Pierson, K., and Horwitz, M. A. (2007) Hairpin extensions enhance the efficacy of mycolyl transferase-specific antisense oligonucleotides targeting Mycobacterium tuberculosis, Proceedings of the National Academy of Sciences 104, 7199-7204. 112. Lisa Y. Armitige, C. J., Audrey R. Wanger, And Steven J. Norris. (2000) Disruption of the Genes Encoding Antigen 85A and Antigen 85B of Mycobacterium tuberculosis H37Rv: Effect on Growth in Culture and in Macrophages, Infection and Immunity 68, 767-778. 113. Sanki, A. K., Boucau, J., Srivastava, P., Adams, S. S., Ronning, D. R., and Sucheck, S. J. (2008) Synthesis of methyl 5-S-alkyl-5-thio-D-arabinofuranosides and evaluation of their antimycobacterial activity, Bioorganic & medicinal chemistry 16, 5672-5682. 114. Ibrahim, D. A., Boucau, J., Lajiness, D. H., Veleti, S. K., Trabbic, K. R., Adams, S. S., Ronning, D. R., and Sucheck, S. J. (2012) Design, synthesis, and X-ray analysis of a glycoconjugate bound to Mycobacterium tuberculosis antigen 85C, Bioconjugate chemistry 23, 2403-2416. 115. Warrier, T., Tropis, M., Werngren, J., Diehl, A., Gengenbacher, M., Schlegel, B., Schade, M., Oschkinat, H., Daffe, M., and Hoffner, S. (2012) Antigen 85C inhibition restricts Mycobacterium tuberculosis growth through disruption of cord factor biosynthesis, Antimicrobial agents and chemotherapy 56, 1735-1743. 116. Thanna, S., Knudson, S. E., Grzegorzewicz, A., Kapil, S., Goins, C. M., Ronning, D. R., Jackson, M., Slayden, R. A., and Sucheck, S. J. (2016) Synthesis and evaluation of new 2-aminothiophenes against Mycobacterium tuberculosis, Organic & biomolecular chemistry 14, 6119-6133.

174

117. Gobec, S., Plantan, I., Mravljak, J., Švajger, U., Wilson, R. A., Besra, G. S., Soares, S. L., Appelberg, R., and Kikelj, D. (2007) Design, synthesis, biochemical evaluation and antimycobacterial action of phosphonate inhibitors of antigen 85C, a crucial enzyme involved in biosynthesis of the mycobacterial cell wall, European journal of medicinal chemistry 42, 54-63. 118. Xu, K., Zhang, Y., Tang, B., Laskin, J., Roach, P. J., and Chen, H. (2010) Study of highly selective and efficient thiol derivatization using selenium reagents by mass spectrometry, Analytical chemistry 82, 6926-6932. 119. Singh, J., Petter, R. C., Baillie, T. A., and Whitty, A. (2011) The resurgence of covalent drugs, Nature reviews. Drug discovery 10, 307-317. 120. Bradshaw, J. M., McFarland, J. M., Paavilainen, V. O., Bisconte, A., Tam, D., Phan, V. T., Romanov, S., Finkle, D., Shu, J., and Patel, V. (2015) Prolonged and tunable residence time using reversible covalent kinase inhibitors, Nature chemical biology 11, 525-531. 121. Otrubova, K., Brown, M., McCormick, M. S., Han, G. W., O'Neal, S. T., Cravatt, B. F., Stevens, R. C., Lichtman, A. H., and Boger, D. L. (2013) Rational design of fatty acid amide hydrolase inhibitors that act by covalently bonding to two active site residues, Journal of the American Chemical Society 135, 6289-6299. 122. Caldwell, G. W., and Yan, Z. (2016) OPTIMIZATION IN DRUG DISCOVERY, Springer. 123. Zhang, J. H., Chung, T. D., and Oldenburg, K. R. (1999) A simple statistical parameter for use in evaluation and validation of high throughput screening assays, Journal of biomolecular screening 4, 67-73. 124. Ballell, L., Bates, R. H., Young, R. J., Alvarez‐Gomez, D., Alvarez‐Ruiz, E., Barroso, V., Blanco, D., Crespo, B., Escribano, J., and González, R. (2013) Fueling Open‐Source Drug Discovery: 177 Small‐Molecule Leads against Tuberculosis, ChemMedChem 8, 313-321. 125. Pantoliano, M. W., Petrella, E. C., Kwasnoski, J. D., Lobanov, V. S., Myslik, J., Graf, E., Carver, T., Asel, E., Springer, B. A., and Lane, P. (2001) High-density miniaturized thermal shift assays as a general strategy for drug discovery, Journal of biomolecular screening 6, 429-440. 126. Vedadi, M., Arrowsmith, C. H., Allali-Hassani, A., Senisterra, G., and Wasney, G. A. (2010) Biophysical characterization of recombinant proteins: a key to higher structural genomics success, Journal of structural biology 172, 107-119. 127. Geders, T. W., Gustafson, K., and Finzel, B. C. (2012) Use of differential scanning fluorimetry to optimize the purification and crystallization of PLP-dependent enzymes, Acta Crystallographica Section F: Structural Biology and Crystallization Communications 68, 596-600. 128. Lo, M. C., Aulabaugh, A., Jin, G., Cowling, R., Bard, J., Malamas, M., and Ellestad, G. (2004) Evaluation of fluorescence-based thermal shift assays for hit identification in drug discovery, Analytical biochemistry 332, 153-159. 129. Scandurra, R., Consalvi, V., Chiaraluce, R., Politi, L., and Engel, P. C. (1998) Protein thermostability in extremophiles, Biochimie 80, 933-941.

175

130. Szilágyi, A., and Závodszky, P. (2000) Structural differences between mesophilic, moderately thermophilic and extremely thermophilic protein subunits: results of a comprehensive survey, Structure 8, 493-504. 131. Sadeghi, M., Naderi-Manesh, H., Zarrabi, M., and Ranjbar, B. (2006) Effective factors in thermostability of thermophilic proteins, Biophysical chemistry 119, 256-270. 132. Sasindran, S. J., and Torrelles, J. B. (2011) Mycobacterium tuberculosis infection and inflammation: what is beneficial for the host and for the bacterium?, Frontiers in microbiology 2. 133. Adibekian, A., Martin, B. R., Wang, C., Hsu, K. L., Bachovchin, D. A., Niessen, S., Hoover, H., and Cravatt, B. F. (2011) Click-generated triazole ureas as ultrapotent in vivo–active serine hydrolase inhibitors, Nature chemical biology 7, 469-478. 134. Aaltonen, N., Savinainen, J. R., Ribas, C. R., Rönkkö, J., Kuusisto, A., Korhonen, J., Navia-Paldanius, D., Häyrinen, J., Takabe, P., and Käsnänen, H. (2013) Piperazine and piperidine triazole ureas as ultrapotent and highly selective inhibitors of monoacylglycerol lipase, Chemistry & biology 20, 379-390. 135. Goins, C. M., Schreidah, C. M., Dajnowicz, S., and Ronning, D. R. (2017) Structural basis for lipid binding and mechanism of the Mycobacterium tuberculosis Rv3802 phospholipase, Journal of Biological Chemistry 239, 1363- 1372 136. Parker, S. K., Barkley, R. M., Rino, J. G., and Vasil, M. L. (2009) Mycobacterium tuberculosis Rv3802c encodes a phospholipase/thioesterase and is inhibited by the antimycobacterial agent tetrahydrolipstatin, PloS one 4, e4281. 137. Meniche, X., Labarre, C., De Sousa-D'Auria, C., Huc, E., Laval, F., Tropis, M., Bayan, N., Portevin, D., Guilhot, C., and Daffé, M. (2009) Identification of a stress-induced factor of Corynebacterineae that is involved in the regulation of the outer membrane lipid composition, Journal of bacteriology 191, 7323-7332. 138. Crellin, P. K., Vivian, J. P., Scoble, J., Chow, F. M., West, N. P., Brammananth, R., Proellocks, N. I., Shahine, A., Le Nours, J., and Wilce, M. C. (2010) Tetrahydrolipstatin inhibition, functional analyses, and three-dimensional structure of a lipase essential for mycobacterial viability, Journal of Biological Chemistry 285, 30050-30060. 139. Miltner, E., Daroogheh, K., Mehta, P. K., Cirillo, S. L., Cirillo, J. D., and Bermudez, L. E. (2005) Identification of Mycobacterium avium genes that affect invasion of the intestinal epithelium, Infection and immunity 73, 4214-4221. 140. West, N. P., Chow, F. M., Randall, E. J., Wu, J., Chen, J., Ribeiro, J. M., and Britton, W. J. (2009) Cutinase-like proteins of Mycobacterium tuberculosis: characterization of their variable enzymatic functions and active site identification, The FASEB Journal 23, 1694-1704. 141. Biasini, M., Bienert, S., Waterhouse, A., Arnold, K., Studer, G., Schmidt, T., Kiefer, F., Cassarino, T. G., Bertoni, M., and Bordoli, L. (2014) SWISS-MODEL: modelling protein tertiary and quaternary structure using evolutionary information, Nucleic acids research 42, W252-W258.

176

142. Terwilliger, T. C., Klei, H., Adams, P. D., Moriarty, N. W., and Cohn, J. D. (2006) Automated ligand fitting by core-fragment fitting and extension into density, Acta Crystallographica Section D 62, 915-922. 143. Notredame, C., Higgins, D. G., and Heringa, J. (2000) T-Coffee: A novel method for fast and accurate multiple sequence alignment, Journal of molecular biology 302, 205-217. 144. Drozdetskiy, A., Cole, C., Procter, J., and Barton, G. J. (2015) JPred4: a protein secondary structure prediction server, Nucleic acids research 43, W389-W394. 145. Krogh, A., Larsson, B., Von Heijne, G., and Sonnhammer, E. L. (2001) Predicting transmembrane protein topology with a hidden Markov model: application to complete genomes, Journal of Molecular Biology 305, 567-580. 146. Dennis, E. A., Cao, J., Hsu, Y.-H., Magrioti, V., and Kokotos, G. (2011) Phospholipase A2 enzymes: physical structure, biological function, disease implication, chemical inhibition, and therapeutic intervention, Chemical reviews 111, 6130-6185. 147. Dennis, E. A. (2015) Introduction to thematic review series: phospholipases: central role in lipid signaling and disease, Journal of lipid research 56, 1245-1247. 148. Carter, P., and Wells, J. A. (1988) Dissecting the catalytic triad of a serine protease, Nature 332, 564-568. 149. Mouchlis, V. D., Bucher, D., McCammon, J. A., and Dennis, E. A. (2015) Membranes serve as allosteric activators of phospholipase A2, enabling it to extract, bind, and hydrolyze phospholipid substrates, Proceedings of the National Academy of Sciences 112, E516-E525. 150. Marsh, D. (2003) Lipid–binding proteins: Structure of the phospholipid ligands, Protein science 12, 2109-2117. 151. Parish, T., Liu, J., Nikaido, H., and Stoker, N. G. (1997) A Mycobacterium smegmatis mutant with a defective inositol monophosphate phosphatase gene homolog has altered cell envelope permeability, Journal of bacteriology 179, 7827-7833. 152. Morita, Y. S., Velasquez, R., Taig, E., Waller, R. F., Patterson, J. H., Tull, D., Williams, S. J., Billman-Jacobe, H., and McConville, M. J. (2005) Compartmentalization of lipid biosynthesis in mycobacteria, Journal of Biological Chemistry 280, 21645-21652. 153. West, N. P., Cergol, K. M., Xue, M., Randall, E. J., Britton, W. J., and Payne, R. J. (2011) Inhibitors of an essential mycobacterial cell wall lipase (Rv3802c) as tuberculosis drug leads, Chemical Communications 47, 5166-5168. 154. Freire, E. (2005) A thermodynamic guide to affinity optimization of drug candidates, Springer, 291-307 155. Bissantz, C., Kuhn, B., and Stahl, M. (2010) A medicinal chemist’s guide to molecular interactions, Journal of medicinal chemistry 53, 5061-5084. 156. Sarker, S. D., Nahar, L., and Kumarasamy, Y. (2007) Microtitre plate-based antibacterial assay incorporating resazurin as an indicator of cell growth, and its application in the in vitro antibacterial screening of phytochemicals, Methods 42, 321-324.

177

157. Vraka, C., Nics, L., Wagner, K.-H., Hacker, M., Wadsak, W., and Mitterhauser, M. (2017) LogP, a yesterday's value?, Nuclear Medicine and Biology 50, 1-10. 158. Bulatovic, V. M., Wengenack, N. L., Uhl, J. R., Hall, L., Roberts, G. D., Cockerill, F. R., and Rusnak, F. (2002) Oxidative stress increases susceptibility of Mycobacterium tuberculosis to isoniazid, Antimicrobial agents and chemotherapy 46, 2765-2771. 159. Goins, C. M., Sudasinghe, T. D., Liu, X., Wang, Y., O'Doherty, G. A., and Ronning, D. R. (2018) Characterization of Tetrahydrolipstatin and Stereo-derivatives on the Inhibition of Essential Mycobacterium tuberculosis Lipid Esterases, ACS Biochemistry, In Press, DOI: 10.1021/acs.biochem.8b00152 160. Heck, A. M., Yanovski, J. A., and Calis, K. A. (2000) Orlistat, a new lipase inhibitor for the management of obesity, Pharmacotherapy: The Journal of Human Pharmacology and Drug Therapy 20, 270-279. 161. Hugonnet, J.-E., Tremblay, L. W., Boshoff, H. I., Barry, C. E., and Blanchard, J. S. (2009) Meropenem-clavulanate is effective against extensively drug-resistant Mycobacterium tuberculosis, Science 323, 1215-1218. 162. Dhar, N., Dubée, V., Ballell, L., Cuinet, G., Hugonnet, J.-E., Signorino-Gelo, F., Barros, D., Arthur, M., and McKinney, J. D. (2015) Rapid cytolysis of Mycobacterium tuberculosis by faropenem, an orally bioavailable β-lactam antibiotic, Antimicrobial agents and chemotherapy 59, 1308-1319. 163. Menendez, J., Vellon, L. a., and Lupu, R. (2005) Antitumoral actions of the anti- obesity drug orlistat (Xenical™) in breast cancer cells: blockade of cell cycle progression, promotion of apoptotic cell death and PEA3-mediated transcriptional repression of Her2/neu (erb B-2) oncogene, Annals of Oncology 16, 1253-1267. 164. Kuhajda, F. P. (2006) Fatty acid synthase and cancer: new application of an old pathway, Cancer research 66, 5977-5980. 165. Soroka, D., de La Sierra-Gallay, I. L., Dubée, V., Triboulet, S., Van Tilbeurgh, H., Compain, F., Ballell, L., Barros, D., Mainardi, J. L., and Hugonnet, J. E. (2015) Hydrolysis of clavulanate by Mycobacterium tuberculosis β-lactamase BlaC harboring a canonical SDN motif, Antimicrobial agents and chemotherapy 59, 5714-5720. 166. Cohen, K. A., El-Hay, T., Wyres, K. L., Weissbrod, O., Munsamy, V., Yanover, C., Aharonov, R., Shaham, O., Conway, T. C., and Goldschmidt, Y. (2016) Paradoxical hypersusceptibility of drug-resistant mycobacteriumtuberculosis to β- lactam antibiotics, EBioMedicine 9, 170-179.

178