Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Iron-containing monooxygenases in Mycobacterium tuberculosis cholesterol degradation : biochemical and… Capyk, Jenna 2012

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


24-ubc_2012_spring_capyk_jenna.pdf [ 31.64MB ]
JSON: 24-1.0072626.json
JSON-LD: 24-1.0072626-ld.json
RDF/XML (Pretty): 24-1.0072626-rdf.xml
RDF/JSON: 24-1.0072626-rdf.json
Turtle: 24-1.0072626-turtle.txt
N-Triples: 24-1.0072626-rdf-ntriples.txt
Original Record: 24-1.0072626-source.json
Full Text

Full Text

Iron-containing monooxygenases in Mycobacterium tuberculosis cholesterol degradation: biochemical and phylogenetic perspectives by Jenna Capyk B.Sc., The University of British Columbia, 2007 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES (Biochemistry and Molecular Biology) THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  March 2012 © Jenna Capyk, 2012  Abstract Mycobacterium tuberculosis (Mtb) is the human pathogen that causes tuberculosis. A gene cluster encoding a cholesterol degradation pathway plays a role in Mtb virulence. Two iron-containing monooxygenases in this pathway were characterized with respect to their roles in bacterial cholesterol catabolism: the Rieske oxygenase (RO) KshAB, and the cytochrome P450 (P450) Cyp125. These enzymes are predicted to catalyze the first ringopening step and the first transformation of the steroid side chain, respectively. Cyp125A1 (Mtb) and Cyp125A14P (Rhodococcus jostii RHA1) were expressed in R. jostii RHA1 and characterized in vitro using the Mtb reductase KshB. Both enzymes were purified with the heme iron in a predominantly high spin state and exhibiting thiolate ligation of the heme iron. Both P450s bound cholesterol and 4-cholesten-3-one with apparent submicromolar affinity. Cyp125A1 was demonstrated to catalyze C26-monohydroxylation of both steroids. KshA (a terminal oxygenase) and KshB (an oxygenase reductase) of Mtb were produced in Escherichia coli and characterized in vitro. KshAB had over twenty times the apparent substrate specificity for steroid substrates with isopropionyl-CoA side chains than for the corresponding 17-keto steroids. The apparent KMO2 with a CoA thioester-bearing steroid was 90 ± 10 µM whereas that for the corresponding 17-keto steroid was in excess of 1.2 mM. These results suggest that the physiological substrate(s) for KshAB is likely a CoA thioester intermediate of cholesterol side chain degradation. A comprehensive phylogenetic analysis was undertaken to consolidate the available RO literature. Six hundred fifty enzymes that are fully representative of the RO terminal oxygenase (RO-O) sequences in the NCBI database were collected and aligned to a structurebased sequence template. The structure-based alignment was also used to objectively define the structurally conserved positions that were included in phylogenetic reconstruction. The resulting analysis revealed a level of RO-O diversity that has been unrecognized in previous literature and that necessitates a different approach to RO-O classification. A classification scheme based on the system currently in use for P450s was proposed. This work provides significant insight into the cholesterol degradation pathway of Mtb and the RO-O protein family and contributes to potential commercial applications in bioremediation, biocatalysis, and Mtb therapeutics. ii  Preface Parts of this thesis have been published in peer-reviewed journals. Characterization of Cyp125A14P appeared in Molecular Microbiology (Rosloniec, K. Z., Wilbrink, M. H., Capyk, J. K., Mohn, W. W., Ostendorf, M., van der Geize, R., Dijkhuizen, L., and Eltis, L.D. “Cytochrome P450 125 (CYP125) catalyses C26-hydroxylation to initiate sterol side-chain degradation in Rhodococcus jostii RHA1” (2009) Mol Micro 74 (5), 1031-1043). In this study, I was responsible for purifying Cyp125A14P from R. jostii RHA1, performing protein analytical procedures including steroid-binding assays, and preparing documentation on the methods and results of these experiments. This published work is located primarily in section 3.1. Characterization of Cyp125A1 appeared in the Journal of Biological Chemistry (Capyk, J.K., Kalscheuer, R., Stewart, G. R., Liu, J., Kwon, H., Zhao, R., Okamoto, S., Jacobs, W. R. Jr., Eltis, L. D., and Mohn, W. W. “Mycobacterial Cytochrome P450 125 (Cyp125) Catalyzes the Terminal Hydroxylation of C17 Steroids” (2009) J Biol Chem 284 (51) 35534-35542). In this study, I was responsible for producing and purifying enzymes, performing protein analytical procedures including steroid-binding and activity assays, performing small molecule analytical procedures, and writing the manuscript. This published work is located primarily in section 3.1. Characterization of KshAB appeared in the Journal of Biological Chemistry (Capyk, J.K., D’Angelo, I., Strynadka, C.N., and Eltis, L. D. “Characterization of 3-Ketosteroid 9!Hydroxylase, a Rieske Oxygenase in the Cholesterol Degradation Pathway of Mycobacterium tuberculosis” (2009) J Biol Chem 284 (15) 9937-9946; and Capyk, J.K., Casabon, I., Gruninger, R., Strynadka, N.C., and Eltis, L.D. “Activity of 3-Ketosteroid 9!Hydroxylase (KshAB) Indicates Cholesterol Side Chain and Ring Degradation Occur Simultaneously in Mycobacterium tuberculosis” (2011) J Biol Chem 286 (47) 40717-40724) and the Handbook of Metalloproteins (Capyk, J.K., Strynadka, N. C., and Eltis, L.D., “3Ketosteroid 9!-Hydroxylase” (2010) Handbook of Metalloproteins, John Wiley and Sons, Ltd.). In these studies I was responsible for producing and purifying enzymes, performing iii  protein analytical procedures including activity assays, performing small molecule analytical procedures, contributing to structural interpretation and comparative structural analysis, and writing the manuscripts. This published work is located primarily in section 3.2. The global Rieske oxygenase phylogenetic analysis appeared in the Journal of Biological Inorganic Chemistry (Capyk, J.K. and Eltis, L. D. “Phylogenetic analysis reveals the surprising diversity of an oxygenase class” (2012) J Biol Inorg Chem 17 (3) 425-436). I was responsible for all aspects of this work. This published work is located primarily in section 3.3.  iv  Table of Contents Abstract ..................................................................................................................................... ii! Preface...................................................................................................................................... iii! Table of Contents...................................................................................................................... v! List of Tables ............................................................................................................................ x! List of Figures .......................................................................................................................... xi! List of Abbreviations ............................................................................................................. xiii! Acknowledgements................................................................................................................. xv! 1 Introduction............................................................................................................................ 1! 1.1! Steroids ..................................................................................................................................... 1! 1.1.1! Steroid metabolism............................................................................................................ 2!! Eukaryotic steroid biosynthesis ................................................................................. 2!! Bacterial steroid degradation ..................................................................................... 5! 1.1.2! Cholesterol......................................................................................................................... 7! 1.2! Bacterial cholesterol metabolism.............................................................................................. 8! 1.2.1! Cholesterol catabolic pathway........................................................................................... 9! 1.2.2! Cholesterol catabolism regulation and import................................................................. 14! 1.2.3! Side chain degradation .................................................................................................... 16! 1.2.4! Ring degradation ............................................................................................................. 17! 1.3! Mycobacterium tuberculosis................................................................................................... 19! 1.3.1! Related organisms ........................................................................................................... 20!! Mycobacteria............................................................................................................ 20!! Mycobacterial cell envelope ............................................................................. 20!! Mycobacterium tuberculosis complex ..................................................................... 23! 1.3.2! Infection cycle ................................................................................................................. 23!! Human immune response......................................................................................... 29! 1.3.3! Determinants of Mtb pathogenicity ................................................................................. 31!! Modulation of intracellular killing mechanisms ...................................................... 32!! Cell envelope............................................................................................................ 34!  v! Latent infection and dormancy ................................................................................ 35!! Lipid-biased metabolism.......................................................................................... 36!! Other proposed factors in pathogenicity .................................................................. 38!! Pathogenicity vs. host immune response ................................................................. 39! 1.3.4! Role of cholesterol in Mtb virulence ............................................................................... 40!! Host cholesterol and establishment of infection ...................................................... 41!! Mtb cholesterol catabolism ...................................................................................... 41!! The role of cholesterol catabolism in virulence ....................................................... 42! 1.4! Oxygenases ............................................................................................................................. 43! 1.4.1! Mononuclear iron-containing oxygenases....................................................................... 44!! Iron-containing monooxygenases in the cholesterol degradation pathway ............. 46! 1.4.2! Cytochromes P450 .......................................................................................................... 49!! Reactions and applications....................................................................................... 50!! Mechanism ............................................................................................................... 51!! Classification............................................................................................................ 52! 1.4.3! Rieske oxygenases........................................................................................................... 52!! Reactions and applications....................................................................................... 54!! Structure ................................................................................................................... 56!! Mechansim ............................................................................................................... 60!! Eukaryotic Rieske oxygenases................................................................................. 61!! Classification............................................................................................................ 62! 1.5! Aim of this study..................................................................................................................... 64!  2 Materials and Methods......................................................................................................... 65! 2.1! Chemicals and reagents .......................................................................................................... 65! 2.2! Bacterial strains and growth ................................................................................................... 66! 2.3! DNA manipulation and plasmid construction ........................................................................ 66! 2.3.1! Oligonucleotides.............................................................................................................. 67! 2.3.2! Plasmid construction ....................................................................................................... 68! 2.4! Protein production and purification ........................................................................................ 69! 2.4.1! Production of Cyp125A14P and Cyp125A1 ................................................................... 69! 2.4.2! Production of proteins in E. coli BL21(DE3).................................................................. 70! 2.4.3! Ni2+-NTA purification of His-tagged proteins ................................................................ 72! 2.4.4! FPLC protein purification ............................................................................................... 72! 2.5! Small molecule analysis.......................................................................................................... 73! vi  2.6! Protein analytics...................................................................................................................... 74! 2.6.1! Protein quantification ...................................................................................................... 74! 2.6.2! Evaluation of cofactor content......................................................................................... 74!! Identification of flavin cofactors.............................................................................. 74!! Determination of iron content .................................................................................. 75!! Determination of acid-labile sulfur content ............................................................. 75! 2.6.3! Spectroscopic analysis..................................................................................................... 76!! Determination of P450 iron spin state equilibrium .................................................. 76!! P450 CO-bound spectra ........................................................................................... 76!! Reduced and oxidized spectra of iron-sulfur cluster proteins.................................. 76! 2.7! Enzyme activity assays ........................................................................................................... 77! 2.7.1! P450 substrate binding .................................................................................................... 77! 2.7.2! Cyp125A1 activity assay................................................................................................. 77! 2.7.3! KshAB activity assay ...................................................................................................... 78! 2.7.4! KshAB coupling assay .................................................................................................... 79! 2.7.5! KstD activity assay .......................................................................................................... 79! 2.8! Phylogenetic analysis of RO-Os ............................................................................................. 80! 2.8.1! Phylogenetic analysis of full-length RO-O ! subunits ................................................... 80! 2.8.2! Structure-based template ................................................................................................. 80! 2.8.3! Sequence collection ......................................................................................................... 81! 2.8.4! Alignment construction ................................................................................................... 81! 2.8.5! Phylogenetic reconstruction ............................................................................................ 82! 2.8.6! Analysis of Rieske and catalytic domains ....................................................................... 83!  3 Results.................................................................................................................................. 84! 3.1! Characterization of Cyp125 .................................................................................................... 84! 3.1.1! Protein production and purification................................................................................. 84!! Cyp125A14P from R. jostii RHA1 .......................................................................... 84!! Cyp125A1 of Mtb H37Rv........................................................................................ 85!! Electronic absorption spectroscopy ......................................................................... 85! 3.1.2! Steroid binding assays ..................................................................................................... 87!! Cyp125A14P of R. jostii RHA1............................................................................... 87!! Cyp125A1 of Mtb H37Rv........................................................................................ 89! 3.1.3! Cyp125A1 activity assays ............................................................................................... 90! 3.2! Characterization of KshAB..................................................................................................... 94! vii  3.2.1! Protein production and purification................................................................................. 94! 3.2.2! Reconstitution of activity ................................................................................................ 95! 3.2.3! Reaction products and coupling ...................................................................................... 96! 3.2.4! Steady state kinetic analysis ............................................................................................ 98! 3.2.5! Reaction of KstD with 4-AD and 4-BNC-CoA............................................................. 102! 3.2.6! KshAH37Rv crystal structure............................................................................................ 103!! Overall structure..................................................................................................... 103!! Structure of the KshAH37Rv active site .................................................................... 106!! Substrate docking simulations ............................................................................... 107!! Comparison with other RO-O alpha subunits ........................................................ 110!! Orientation of the active site channel ............................................................. 110!! Minimal catalytic domain ............................................................................... 111!! Phylogenetic analysis of RO-O structures ...................................................... 112! 3.3! RO-O global phylogenetic analysis ...................................................................................... 113! 3.3.1! Alignment and tree construction ................................................................................... 114! 3.3.2! Overall tree topology..................................................................................................... 116! 3.3.3! RO-O sequence diversity............................................................................................... 118! 3.3.4! Two distinct RO-O groups ............................................................................................ 119! 3.3.5! Origins of the RO-O catalytic domain........................................................................... 123! 3.3.6! RO-O classification scheme .......................................................................................... 124!  4 Discussion .......................................................................................................................... 126! 4.1! Initiation of cholesterol side chain degradation by Cyp125 ................................................. 126! 4.1.1! Multiple P450 C26-hydroxylases .................................................................................. 127! 4.1.2! Formation of the C26 carboxylic acid ........................................................................... 129! 4.1.3! Physiological reductase ................................................................................................. 130! 4.1.4! Structure of Cyp125A1 and Cyp124A1 ........................................................................ 132! 4.2! Order of reactions in cholesterol degradation....................................................................... 133! 4.2.1! Pathway elucidation strategies ...................................................................................... 133! 4.2.2! Insights from KshAB substrate specificity.................................................................... 135! 4.2.3! Implications of KshAB substrate specificity................................................................. 137! 4.2.4! Differences between species and strains ....................................................................... 138! 4.3! Role of cholesterol degradation in Mtb physiology and pathogenicity ................................ 138! 4.3.1! Stage in life cycle .......................................................................................................... 141! 4.3.2! Side chain degradation .................................................................................................. 143! viii  4.3.3! Ring degradation ........................................................................................................... 144!! Discrepancies in kshA knockout phenotypes ......................................................... 146! 4.3.4! Toxicity of cholesterol and its metabolites.................................................................... 148! 4.3.5! Prospects for therapeutic development.......................................................................... 150! 4.4! Implications of global RO-O phylogenetic analysis............................................................. 151! 4.4.1! RO-O sequence diversity............................................................................................... 152!! Limitations of common bioinformatic tools .......................................................... 153! 4.4.2! Two domain fusion events possible in RO-O evolution ............................................... 154! 4.5! Future directions ................................................................................................................... 155  Bibliography......................................................................................................................................159 Appendices............................................................................................................................ 180! Appendix A Bacterial strains in the thesis ..................................................................................... 180! Appendix B X-Ray crystallography data collection and stats ....................................................... 181! B.1! Crystal growth and data collection .................................................................................. 181! B.2! KshA crystallography statistics ....................................................................................... 182! B.3! KshAB (2ZYL) metal-ligand distances ........................................................................... 183! B.4! Docking of substrate molecular models in KshA crystal structure ................................. 184! Appendix C Proposed RO-O nomenclature system....................................................................... 185!  ix  List of Tables Table 1.1  Upregulation of cholesterol-degradation gene cluster ......................................... 11!  Table 1.2  Published Rieske terminal oxygenase structures ................................................. 59!  Table 2.1  Oligonucleotides .................................................................................................. 67!  Table 2.2  Expression plasmids ............................................................................................ 68!  Table 2.3  Incubation temperatures for protein production in E. coli BL21(DE3)............... 71!  Table 2.4  Purification buffer A............................................................................................ 71!  Table 3.1  Apparent dissociation constants of Cyp125 for steroids ..................................... 89!  Table 3.2  Apparent steady state kinetic parameters of KshAB using various steroids ....... 99!  Table 4.1  Apparent binding affinities and steady-state kinetic and binding affinities of  Cyp124A1, Cyp125A1, and Cyp142A1 for cholesterol and 4-cholesten-3-one .................. 129! Table 4.2 Phenotypes of targeted Mtb gene knockouts ..................................................... 140! Table 4.3  Histopathology of guinea pigs infected with Mtb strains .................................. 147!  x  List of Figures Figure 1.1 Chemical structure of cholesterol......................................................................... 1! Figure 1.2 Major steps in eukaryotic steroid biosynthetic pathway ...................................... 4! Figure 1.3 Cholesterol in phospholipid bilayer....................................................................... 8! Figure 1.4 Schematic of cholesterol degradation gene cluster in Mtb................................. 10! Figure 1.5 Summary of cholesterol degradation in Mtb ...................................................... 13! Figure 1.6 Schematic of Mtb cellular envelope ................................................................... 21! Figure 1.7 Mtb infection cycle ............................................................................................. 25! Figure 1.8 Activities of KshAB and KstD ........................................................................... 48! Figure 1.9 Cytochrome P450 reaction cycle........................................................................ 51! Figure 1.10 Schematic of Rieske monooxygenase systems................................................. 53! Figure 1.11 Reactions catalyzed by Rieske oxygenases...................................................... 55! Figure 1.12 Rieske terminal oxygenase structure ................................................................. 57! Figure 1.13 Rieske oxygenase catalytic cycle ..................................................................... 61! Figure 3.1 SDS-PAGE gels of Cyp125 expression and purification ................................... 84! Figure 3.2 UV-vis spectra of Cyp125 .................................................................................. 86! Figure 3.3 Binding of steroids to purified Cyp125A14P..................................................... 88! Figure 3.4 Binding of 4-cholesten-3-one to Cyp125A1 ...................................................... 90! Figure 3.5 Turnover of 4-cholesten-3-one by Cyp125A1 ................................................... 91! Figure 3.6 Spectra of purified putative ferredoxins ............................................................. 92! Figure 3.7 Dependence of the Cyp125A1 reaction on oxygen concentration ........................ 93! Figure 3.8 Spectra of reduced and oxidized KshA and KshB ................................................ 95! Figure 3.9 KshAB substrates under investigation................................................................... 96! Figure 3.10 Fragmentation pattern of KshA-catalyzed transformation of ADD .................... 97! Figure 3.11 Steady state kinetic analyses of KshAB .............................................................. 99! Figure 3.12 Dependence of KshAB reaction rate on O2 concentration ................................ 101! Figure 3.13 Turnover of 4-AD and 4-BNC-CoA by KstD-lysate ........................................ 102! Figure 3.14 Crystal structure of KshA from Mtb.................................................................. 104! Figure 3.15 The secondary structure of the KshAH37Rv catalytic domain............................. 106! Figure 3.16 Two conformations of ADD docked in the active site of KshAH37Rv................ 108!  xi  Figure 3.17 Docking of 1,4-BNC-CoA in active site of KshAH37Rv ..................................... 109! Figure 3.18 Orientation of the active site channel in RO-Os................................................ 111! Figure 3.19 Radial phylogram of structurally characterized ROs ........................................ 113! Figure 3.20 Multiple sequence alignment of RO-Os structural models ............................... 115! Figure 3.21 Topology diagram of the secondary structural elements conserved in all RO-Os of known structure ................................................................................................................ 116! Figure 3.22 Consensus tree of 98 RO-O representative sequences from a bootstrap analysis of 100 datasets........................................................................................................................... 117! Figure 3.23 Consensus trees of the Rieske (A) and catalytic (B) domains of 98 representative RO-O sequences.................................................................................................................... 121! Figure 3.24 Topology diagram of RO-O Rieske domains.................................................... 122! Figure 4.1 4-Cholesten-3-one bound in the active site channel of Cyp125A1..................... 132! Figure 4.2 Virulence of Mtb strains in SCID mice ............................................................... 146!  xii  List of Abbreviations 4-AD  4-androstene-3,17-dione  ADD  1,4-androstadiene-3,17-dione  BCD  2-hydroxypropyl-"-cyclodextrin  BCG  Bacillus Calmette-Guerin  BLAST  basic local alignment search tool  1,4-BNC  3-oxo-23,24-bisnorchol-1,4-dien-22-oic acid  4-BNC  3-oxo-23,24-bisnorchol-4-en-22-oic acid  1,4-BNC-CoA  3-oxo-23,24-bisnorchol-1,4-dien-22-oyl coenzyme A thioester  4-BNC-CoA  3-oxo-23,24-bisnorchol-4-en-22- oyl coenzyme A thioester  CDART  conserved domain architecture retrieval tool  CDD  conserved domain database  CoA  coenzyme A  DOHNAA  9,17-dioxo-1,2,3,4,10,19-hexanorandrostan-5-oic acid  EDTA  ethylenediaminetetraacetic acid  ETC  electron transport chain  FAD  flavin adenine dinucleotide  FAS  ferrous ammonium sulfide  FMN  flavin mononucleotide  FPLC  fast protein liquid chromatography  GC-MS  gas chromatography-coupled mass spectrometry  3-HSA  3-hydroxy 9,10-secandrost-1,3,5(10)-triene-9,17-dione  3-HSBNC  3-hydroxy-9-oxo-9,10-seco-23,24-bisnorchola-1,3,5(10)-trien-22-oic acid  3-HSBNC-CoA  3-hydroxy-9-oxo-9,10-seco-23,24-bisnorchola-1,3,5(10)-trien-22-oyl coenzyme A thioester  HPLC  high performance liquid chromatography  IFN-#  interferon-#  IPTG  isopropyl "-D-1 thiogalactopyranoside  KstD  3-ketosteroid-$1-dehydrogenase xiii  KshAB  3-ketosteroid 9!-hydroxylase  LB  lysogeny broth  Mtb  Mycobacterium tuberculosis  NADH  nicotinamide adenine dinucleotide (reduced)  NAD+  nicotinamide adenine dinucleotide (oxidized)  NADPH  nicotinamide adenine dinucleotide phosphate (reduced)  NAPS  Nucleic Acid Protein Service Unit  NCBI  National Center for Biotechnology Information  Ni2+-NTA  nickel-nitrilotriacetic acid  NMR  nuclear magnetic resonance  OD600  optical density at 600 nm  9-OHAD  9-hydroxy 4-androstene-3,17-dione  P450  cytochrome P450  PCR  polymerase chain reaction  PDIM  phthiocerol dimycocerates  PDB  Protein Data Bank  RD1  Region of Difference 1  RO  Rieske oxygenase  RO-O  Rieske oxygenase (terminal oxygenase component)  r.m.s.d.  root mean square deviation  SDS-PAGE  sodium dodecyl sulfate polyacrylamide gel electrophoresis  TB  tuberculosis  TNF-!  tumour necrosis factor !  TraSH  transposon site hybridizaton  UV-vis  ultraviolet-visible  xiv  Acknowledgements I would like to thank my supervisor, Prof. Lindsay Eltis, for his guidance, support, and instruction in various aspects of research and academia. I am also grateful for the many opportunities he has given me to build my skill sets both in the lab and outside of it. I would like to thank my committee members, Prof. Natalie Strynadka and Prof. Grant Mauk, for their support, collaboration, and valuable advice. There are several current and past members of the Eltis lab who are owed thanks for their help with this project. Jie Liu deserves many thanks for sharing her technical expertise, especially her exceptional cloning ability. Christine Florizone provided much support by establishing initial protocols, helping me become oriented in the lab, and maintaining many components of the lab infrastructure. I would also like to thank other past and present members of the Eltis lab, including Dr. Carly Huitema, Dr. Rahul Singh, Dr. Tim Machonkin, Dr. Hao-Ping Chen, Dr. Nicolas Seghezzi, Jennifer Lian, Antonio Ruzzini, Joseph Roberts, Hyukin Kwon, Rafael Zhao, Jonathan Penfield, James Round, and Adam Crowe for their support, instruction, discussion, assistance, and collaboration. I owe much gratitude to the talented collaborators I have worked with and without whom this work would not have been possible. These individuals include Dr. Isräel Casabon (Eltis lab, University of British Columbia); Dr. Igor D’Angelo and Dr. Robert Gruninger (Strynadka lab, University of British Columbia); Gordon Stewart and Professor William Mohn (Mohn Lab, University of British Columbia); Kamila Rosloniec, Maarten Wilbrink, and Dr. Robert van der Geize (Lubbert Dijkhuizen lab, University of Groningen); Dr. Rainer Kalscheuer and Professor William Jacobs Jr. (Jacobs lab, Howard Hughs Medical Institute); Dr. Lan Ly (Department of Microbial and Molecular Pathogenesis at Texas A & M University); and Dr. Paul J. Converse (Center for Tuberculosis Research in the Department of Medicine at Johns Hopkins University).  xv  I would like to acknowledge the National Science and Engineering Research Council and the Michael Smith Foundation for Health Research for providing financial support in the form of studentships held over the course of my degree. These years would have been a very different experience indeed without the wonderful relationships I have been lucky enough to build with members of the lab. In this regard, I would like to thank Dr. Sachi Okamoto for her instruction, patience, and valuable friendship. Thank you to both Dr. Carola Dresen and Dr. Elitza Tocheva for providing much needed perspective and fun. Very special thanks to Dr. Katherine Yam for years of friendship, support, and academic release. I would also like to thank my parents, Susan and Richard Capyk for their enduring support throughout my degree, and for giving me the freedom to pursue whatever goals I choose. Thank you to my wee brother, Nicholas Capyk, and my adopted sister, Chantel Au, for their encouragement. Finally, I would like to thank Lauren Moccia, my partner in crime (and life). His constant support and friendship has made a very real contribution to any accomplishment I have made since we joined forces. Ours is truly the strongest and most fruitful collaboration I have the privilege to be a part of.  xvi  1  Introduction  1.1  Steroids Steroids are important organic molecules found in both eukaryotic and prokaryotic  organisms. These molecules belong to the isoprenoid family that are derived from 5-carbon isoprene units and are ubiquitous in all life forms (1,2). Steroids constitute a distinct class of isoprenoid with a core structure derived from the gonane chemical skeleton. Comprised of 17 carbons as four fused cycloalkanes, gonane is the simplest steroid (3). Rings A, B, and C are cyclohexanes while ring D is a cyclopentane. In animals, steroids play several vital biological roles. For example, many human sex hormones are steroids, and steroid bile acids are important for solubilization of fats and fat-soluble vitamins in the digestive tract (4). Cholesterol (Figure 1.1) is required for controlling cellular membrane fluidity in animals, while ergosterol plays a similar role in fungi (5). Plants also synthesize a wide variety of steroids collectively referred to as phytosterols (5). Steroids can act as a carbon source or membrane component in bacteria, and pentacyclic hopanoids in the membranes of many bacterial species play much the same role as steroids in eukaryotic membranes (5).  Figure 1.1  Chemical structure of cholesterol. Rings and carbons are labeled according to steroid conventions.  1  1.1.1  Steroid metabolism The anabolic and catabolic components of steroid metabolism are generally  segregated between eukaryotic and prokaryotic forms of life. Most steroids are synthesized by eukaryotes and degraded by prokaryotes (5). There are exceptions to this generalization, including reports of de novo steroid synthesis in the prokaryotes Methylococcus capsulatus (6), Gemmata obscuriglobus (7), and some myxobacteria (8,9). Several cyanobacteria and Mycobacterium smegmatis (10) have also been suggested to synthesize steroids, but these studies have been met with skepticism (5,9). Legitimate endogenous steroids occur sparsely in bacteria and represent a small range of compounds compared with those in eukaryotic systems (5). Neither steroids nor hopanoids are synthesized by any known archaea (9). Some animals, such as many insects and nematodes, are also unable to synthesize steroids and need to consume them as part of their diet (11). Steroid catabolism is undertaken almost completely by prokaryotic organisms.  Eukaryotic steroid biosynthesis  Steroid biosynthesis starts with acetyl-Coenzyme A (CoA) units in animals and fungi (12). These units are the basic building blocks used to form the key intermediates isopentyl5-pyrophosphate and dimethylallylpyrophosphate via the mevalonate pathway. An analogous pathway for isoprenoid precursor biosynthesis is found in the cytoplasm of higher plants and macroalgae (9). The non-mevalonate (or 2-C-methyl-D-erythritol 4-phosphate-1-deoxy-Dxylulose-5-phosphate) pathway is used to produce these intermediates in many microalgae and vascular plant plastids (13). The latter pathway uses pyruvate and glyceraldehyde-3phosphate as the initial compounds for activated isoprenoid synthesis (13). In eukaryotes, steroid biosynthesis from isopentyl-5-pyrophosphate and dimethylallylpyrophosphate follows a common route to the formation of squalene epoxide (5) (Figure 1.2). These two basic units are combined to form geranyl pyrophosphate, three of which are needed to synthesize the 29-carbon branched multi-desaturated alkane, squalene.  2  Squalene oxygenase (also known as squalene epoxidase) then incorporates one atom of oxygen from O2 to form (3S)-2,3-oxidosqualene (squalene epoxide). Formation of this intermediate marks the divergence of steroid and hopanoid biosynthesis. Given the requirement for dioxygen (O2) in this reaction, correlations have been made between eukaryotic evolution and the appearance of appreciable amounts of oxygen in the earth’s atmosphere (5,14). Squalene epoxide is cyclized to form the first steroid in the pathway: lanosterol in animals and fungi, and cycloartenol in plants. There are examples of both protosterols in steroid-synthesizing bacteria (8). In animals, cholesterol is formed by transformation of lanosterol, including incorporation of oxygen from 10 additional molecules of O2 by four separate oxygenases.  3  Figure 1.2  Major steps in eukaryotic steroid biosynthetic pathway. Intermediate metabolites are labeled.  Multiple arrows indicate more than one enzymatic step.  The paradigm of steroid synthesis and metabolic processing in animals involves modification of central metabolic compounds rather than specific de novo biosynthesis of many molecules. Cholesterol is an obligatory steroid intermediate in the production of all other human steroids (4). Some important steroid classes are the 21-carbon progestogens (eg. pregnenolone), 19-carbon androgens (eg. testosterone), 18-carbon estrogens (eg. estradiol), 21-carbon glucocorticoids (eg. cortisol), and 21-carbon mineralocorticoids (eg. aldosterol) (4). These compounds are required for multiple physiological processes including metabolic regulation, sexual maturation, and stress responses (4). Complete mineralization of steroids  4  does not take place in humans. These compounds are hydroxylated at various positions, often by cytochromes P450 (P450s), to form bile acids and excreted into the gastrointestinal tract (4). Although a significant proportion of excreted bile acids is transformed by gut microbes and resorbed for recirculation, these metabolites are ultimately excreted (4).  Bacterial steroid degradation  Typical bacteria do not synthesize steroids or require exogenous steroids for growth. Many bacteria, however, are able to acquire and degrade steroids, thus participating in the global carbon cycle. Notable characterizations of bacterial steroid degradation pathways have taken place in Nocardia restricticus (15), Commamonas testosteroni TA441 (16), various Rhodococcal species (17), and Mycobacterium tuberculosis (Mtb) (18); aspects of microbial steroid degradation have also been studied in many other strains (19). Steroid metabolism in bacteria is regulated at the transcriptional level. For example, the genome of Rhodococcus jostii RHA1 includes four putative steroid degradation gene clusters (20). One of these gene clusters is specifically upregulated in response to cholesterol (21) (Mohn et al. unpublished results) while another is upregulated in response to cholic acid (Mohn et al. unpublished results). Changes in the transcript levels of some steroid degradation genes, especially in pathogens like Mtb, are often not obvious from the results of whole genome transcriptional screens, such as microarray studies. Although these pathways exhibit a degree of specific regulation, high levels of constitutive expression for specific genes in the pathway result in low ratios of upregulation in comparative transcriptional studies (22). This effect was observed for the cholesterol degradation cluster of R. jostii RHA1 (Mohn et al. unpublished results). For this reason, failure to identify a specific gene as significantly upregulated in the presence of a steroid substrate does not exclude involvement of that gene in steroid catabolism. Bacterial steroid degradation involves import of these molecules using specific uptake systems followed by chemical degradation via the aerobic or anaerobic bacterial steroid  5  catabolic pathways. Systems enabling steroid import have been identified in many bacterial species, including human intestinal tract commensal organisms (23) and several actinomycetes (24). Complete steroid mineralization has been reported during both aerobic and anaerobic growth, and the pathways responsible for steroid metabolism under both conditions have been studied (23). The oxidative breakdown pathway contains several oxygenases that incorporate oxygen from O2 to activate the ring structure for degradation (17). This pathway has been identified in several bacteria, most of which are actinomycetes. Indeed, it has been proposed that the aerobic steroid degradation pathway could be nearly ubiquitous in these organisms (24). The aerobic steroid degradation pathway will be discussed in further detail in the context of the actinomycete cholesterol degradation pathway (Section 1.2.1). Bacteria without access to O2 use the anaerobic steroid degradation pathway. Complete transformation of specific steroids to CO2 using nitrate as an electron acceptor has been documented in !-Proteobacteria strain 72Chol (25) and Sterolibacterium denitrificans (26). Investigation of the anoxic degradation of testosterone by the latter organism has demonstrated that the initial reactions in this pathway mirror those of the oxic pathway. Thus, anoxic testosterone degradation is initiated by oxidation of the C17 hydroxyl group and C1C2 bond to form 1,4-androstadiene-3,17-dione (ADD) (27). Following the formation of this intermediate the anoxic pathway diverges, proceeding via reduction of the "4 double bond, #hydration of C1 with concomitant reduction of the "1 double bond, and reduction of the C3 ketone (28). Further study is needed to elucidate the mechanism of ring cleavage in anaerobic steroid metabolism. The metabolic significance of bacterial steroid degradation varies according to the context of growth. For some bacteria, steroids can act as a major electron donor and source of carbon and energy (29,30). In commensal bacteria in the human gut, bacterial transformations of bile acids are proposed to act as an electron sink during fermentative growth (23). The products of these pathways are also resorbed by the host system for reexcretion, contributing to the relationship between host and bacterium (23). It has been proposed that steroid metabolism by pathogenic bacteria can influence host organisms through cellular signaling mechanisms (31). This latter point will be discussed in further detail in the context of Mtb (Section 1.3.4). Elements of bacterial steroid catabolism have  6  also been exploited for the industrial production of steroid drugs from inexpensive raw materials such as cholesterol and sitosterol (32,33). The growing number of complete bacterial genomes available allows for identification of entire steroid degradation pathways that can be studied for applications in industrial processes. 1.1.2  Cholesterol Cholesterol (Figure 1.1) is a 27-carbon steroid constituting an essential component of  animal cell membranes. This molecule is a precursor for the synthesis of almost all other animal steroids (5). In addition to its prevalence animals, small amounts of cholesterol are also synthesized in plants and fungi. Modifications to the gonane cycloalkane ring skeleton characteristic of cholesterol include a C3 !-hydroxyl group, a double bond between C5 and C6, !-methyl groups on C10 and C13, and an 8-carbon 1,5-dimethyl hexane side chain at C17. Cholesterol in animals is both synthesized and absorbed from nutritional sources (34). A critical component of animal cell membranes, cholesterol helps mediate membrane fluidity over a range of physiologically relevant temperatures (35), a function achieved in part by regulation of membrane lipid desaturation levels in prokaryotes (36). Membrane cholesterol increases membrane fluidity at lower temperatures by interfering with close packing of phospholipid tails and decreases membrane fluidity as higher temperatures through interactions limiting the mobility of the same structures (35). Specific interactions between different parts of the cholesterol molecule with other cell membrane components mediate the effect of cholesterol on membrane properties. The C3-hydroxyl group interacts with polar head groups of phospholipids and sphingolipids while the steroid core and aliphatic sidechain interact with the fatty acid moities of these lipids (Figure 1.3) (37). Membrane cholesterol also reduces cation permeability and has specific receptor functions as a component of lipid rafts (34). Cholesterol is required for formation of cavolae and clathrincoated pits, making this lipid essential for many forms of intracellular trafficking, cell signaling, and nerve conduction (34).  7  Figure 1.3 Cholesterol in phospholipid bilayer. Figure from The Cell: A Molecular Approach (37), © Geoffrey M. Cooper, 2000, by permission.  Unlike animals, bacteria do not use cholesterol as a major structural lipid although some employ hopanoid compounds that serve the same function in vitro (38). Cholesterol produced by animals, however, plays a role in the biology of many bacteria. Some bacteria use cholesterol as a source of carbon and energy while others exploit the presence of cholesterol in animal membranes to interact with animal cells. The cholesterol-dependent cytolysins, for example, are pore-forming toxins used by many Gram-negative bacterial pathogens that rely on cholesterol in the host membrane as toxin receptors (39). Similarly, animal membrane cholesterol-rich microdomains, or membrane rafts, are proposed requirements for the entry of several intracellular pathogenic bacteria into mammalian cells (40). 1.2  Bacterial cholesterol metabolism Over the past several decades, several genera of Actinomycetales bacteria have been  shown to be able to degrade exogenous cholesterol. Cholesterol mineralization has been  8  characterized in many strains (eg. Gordonia (41), Rhodococcus (42) (21), and Mycobacterium (30) (43) (21)) (19). The ecological niches of cholesterol-degrading bacteria vary from intracellular animal pathogens to non-pathogenic organisms. Bacterial cholesterol degradation has been studied not only in the context of bacterial physiology, but for the potential application of the relevant enzymes to chemical and industrial processes. For example, the bacterial enzyme cholesterol oxidase is commonly used to assay the cholesterol content of human serum (44). Other applications of cholesterol-degrading enzymes include transformation of steroids into industrially valuable intermediates (19,32,45), biosynthesis of antifungal antibiotics, and use as biocatalytic insecticides (44). To this end, steroid transformations of many steroids have long been studied in both medical and industrial contexts (46), with actinomycetes receiving much of the attention for their relative efficiency in specific steroid transformations (47). 1.2.1  Cholesterol catabolic pathway A recent upsurge in biochemical characterization of the actinomycete cholesterol  degradation pathway has been enabled by the discovery of a gene cluster that is upregulated during growth on cholesterol in the non-pathogenic soil actinomycete R. jostii RHA1 (21) (Figure 1.3). Many of the genes in this ~80 gene cluster, along with much of their operonic structure, are conserved in the actinomycete pathogen Mtb (Rv3494c-3574) (21). The involvement of these genes in cholesterol degradation was recently confirmed by a transposon mutagenesis study (48). Several microarray studies have demonstrated upregulation of genes in this cluster in response to growth on cholesterol and within macrophages (Table 1.1) (21,49-51). These discoveries prompted the proposal of a cholesterol degradation pathway in Mtb and research into the associated enzymes (21). This pathway is discussed in Sections 1.2.1-1.2.4 with a focus on the elucidation of cholesterol degradation in Mtb.  9  Figure 1.4  Schematic of cholesterol degradation gene cluster in Mtb. Genes are colour coded with respect to their known or predicted function: uptake, red;  side chain degradation, magenta; A and B ring degradation, green; 9,17-dioxo-1,2,3,4,10,19-hexanorandrostan-5-oic acid degradation, blue; gene regulation, yellow. Genes of unknown function are shown in black. Clusters of genes regulated by KstR1 and KstR2 are labeled; genes not regulated by either transcriptional regulator are shown in grey. Genes identified by transposon mutagenesis studies as important for Mtb growth on cholesterol (#) (48), survival in a macrophage infection model (*) (52), and survival in a mouse model (+) (22) are indicated.  10  Table 1.1  Upregulation of cholesterol-degradation gene cluster  Gene  Name  3494c 3495c 3496c 3497c 3498c 3499c 3500c 3501c 3502c 3503c 3504 3505 3506 3513c 3514 3515c 3516 3517 3518c 3519 3520c 3521 3522 3523 3524 3525c 3526 3527 3528c 3529c 3530c 3531c 3532 3533c 3534c 3535c 3536c 3537 3538 3539 3540c 3541c 3542c 3543c  mce4F mce4E mce4D mce4C mce4B mce4A supB supA hsd4A fdxD fadE26 fadE27 fadD17 fadD18 fadD19 echA19  cholesterolgrown R. jostii RHA11  macrophages (24 hr)2  ! ! ! ! ! ! ! ! ! ! !  ! !  macrophages (24 hr, 17 strains)4  macrophages (2 hr)5  ! ! !  ! ! !  ! !  !  ! !  ! ! !  cyp142  ltp4 ltp3  macrophages (24 hr)3  ! ! ! ! !  ! ! !  kshA  ! !  ! ! !  ! ! !  hsaF hsaG hsaE kstD hsd4B  ! ! ! ! !  ! ! ! ! !  ltp2  ! !  ! ! ! !  fadE29  !  !  ! ! !  !  !  !  !  ! ! ! !  !  ! ! !  !  11  Gene  Name  cholesterolgrown R. jostii RHA11 ! ! ! ! ! ! ! ! ! !  macrophages (24 hr)2  macrophages (24 hr)3  macrophages (24 hr, 17 strains)4 ! ! !  macrophages (2 hr)5  3544c fadE28 ! ! ! 3545c cyp125 ! ! ! 3546 fadA5 ! ! 3547 3548c 3549c ! 3550 echA20 ! 3551 ipdA 3552 ipdB ! ! 3553 ! 3554 ! 3555c ! 3556c fadA6 ! ! 3557c kstR2 ! ! 3558 3559c ! 3560c fadE30 ! 3561 fadD3 ! ! ! 3562 fadE31 ! ! ! 3563 fadE32 ! ! 3564 fadE33 ! 3565 3566c 3567c hsaB ! ! 3568c hsaC ! ! 3569c hsaD ! ! 3570c hsaA ! ! 3571 kshB ! ! ! ! 3572 ! 3573c fadE34 ! ! 3574 kstR ! ! 1 Genes upregulated in R. jostii RHA1 during in vitro growth on cholesterol vs. pyruvate (21) 2 Genes upregulated in Mtb clinical isolate 1254 isolated from either resting or activated macrophages 24 hr post-infection vs. cells used for macrophage infection (7H9 growth) (51) 3 Genes upregulated in Mtb CDC1551 isolated from both resting and activated macrophages 24 hr post-infection vs. cells used for macrophage infection (7H9-OADC growth) (49) 4 Genes upregulated in Mtb isolated from either resting or activated macrophages 24 hr post infection vs. cells used for macrophage infection (7H9-OADC growth). A checkmark indicates that 15/17 strains of Mtb tested exhibited this transcriptional regulation pattern (50) 5 Genes upregulated in Mtb CDC1551 isolated from both resting and activated macrophages 2 hr post-infection vs. cells used for macrophage infection (7H9-OADC growth) (49)  The cholesterol catabolic pathway is an example of the aerobic steroid degradation paradigm. Incorporation of oxygen from O2 is used to activate various components of the cholesterol structure for degradation. At least four oxygenase enzymes act during cholesterol 12  degradation (17). A summary of our current understanding of the cholesterol degradation pathway in Mtb is represented in Figure 1.5. Many steps in this pathway have direct analogs in degradation pathways for other steroids in other bacteria (eg. testosterone degradation in C. testosteroni (53) and cholic acid degradation in Pseudomonas sp. Chol1 (29)). The intermediate 4-androstene-3,17-dione (4-AD) (Figure 1.5, 1 R = keto group) is common to all proposed aerobic bacterial steroid degradation pathways (17), although its physiological relevance in all systems is unproven.  Figure 1.5  Summary of cholesterol degradation in Mtb. Schematic of the proposed cholesterol degradation  pathway of Mtb. The processes of cholesterol side chain and partial ring degradation are shown as parallel pathways. Enzymes known to catalyze specific reactions in Mtb are labeled in bold type. Reactions that have not been demonstrated in vitro are shown as dashed arrows.  Complete breakdown of cholesterol involves two conceptually distinct processes: degradation of the aliphatic side chain and degradation of the steroid ring structure (17). Recent descriptions of the pathway in the literature show complete side chain degradation to the 17-ketosteroid before cleavage of the ring structures (17,48). These schemes also place the first oxidation of ring A before the first side chain transformation step. These publications notwithstanding, the physiological order of reactions in cholesterol degradation has not been  13  demonstrated in the context of a wild-type bacterium. Studies involving several bacterial strains have revealed ring-degradation intermediates bearing partially degraded side chains, indicating that the two processes can occur concurrently to some extent (29,41,54). Cumulative studies indicate that the cholesterol-degrading enzymes exhibit a degree of substrate promiscuity. Indeed, Sih et al. observed in 1968 that side chain degradation and ring degradation do not have a compulsory order but proceed simultaneously and independently such that blockages in one pathway result in accumulation of end products of the other (55). Similarly, selective degradation of the side chain, which is advantageous for industrial pharmaceutical intermediate production, has been achieved in several actinomycetes using various methods, without which simultaneous degradation of the ring structure and side chain is observed (47). Both the physiological relevance of this functional enzymatic flexibility and the preferred order of reactions remain to be determined. Regardless of their order, it has been proposed that the metabolic fate of carbons originating from the ring structure is different from the fate of side chain-derived carbons. Specifically, while carbon originating in ring A is mineralized to CO2 (30), carbon derived from the cholesterol side chain is proposed to be incorporated into cellular lipids (30,56). 1.2.2  Cholesterol catabolism regulation and import Transcriptional regulation of cholesterol catabolic genes has been best characterized  in Mtb. In this pathogen, these genes are under the control of the transcriptional regulators KstR1 (Rv3574) and KstR2 (Rv3557c) (57) (Figure 1.4). These proteins are both TetR-type transcriptional repressors that bind to specific DNA motifs in the absence of an effector molecule. Both transcriptional regulators are highly conserved among actinomycetes and both bind regulatory sequences in their own promoter regions (57,58). This mechanism allows for negative feedback in the absence of high effector molecule concentrations and limits inappropriate upregulation of the genes in their respective regulons. KstR1 of Mtb was found to bind a 14-nucleotide inverted repeat to directly control the transcription of 74 genes, 51 of which are in the cholesterol degradation gene cluster (58). Six putative operons, encompassing genes Rv3548c to Rv3565, contain several genes that are upregulated in the presence of cholesterol (21) but are not predicted to be part of the KstR1 regulon (58). The  14  three intergenic regions between divergently transcribed genes in this region contain a distinct 14-nucleotide motif that is recognized by KstR2, the gene for which is also encoded in this block (57). Together, these two regulators control transcription of the vast majority of the cholesterol degradation gene cluster encompassing Rv3494c to Rv3574. KstR1 and KstR2 act independently, however, and although both regulators respond when cholesterol is a major carbon source, there is no apparent overlap in their regulons (21,57). The implications of having two separate regulators for Mtb temporal gene expression have not been fully explored. Likewise, the activating ligands for KstR1 and KstR2 have yet to be identified. Further research may demonstrate that differences between these regulators reflect the suggested involvement of cholesterol catabolic genes in the degradation of multiple lipid species (57,58). Specific transport systems are required for efficient transport of steroid molecules across the bacterial cellular envelope. This function has been attributed to proteins encoded by the mce4 locus of actinomycetes (24). Originally named for their supposed role in mammalian cell entry (59), this family of proteins is conserved among actinomycetes. Indeed, most of these bacteria have more than one paralogue encoded by physically separated loci (24). These genes are proposed to encode members of a novel class of complex ATPbinding cassette transporter, with putative exported virulence-associated proteins also encoded in some loci (60). The exact role of the putative proteins encoded by many of these genes has yet to be determined. The 11-gene mce4 operon of R. jostii RHA1 was shown to constitute a steroid uptake system and this function is proposed to extend to the mce4 loci of other actinomycetes (24). In accordance with this prediction, the mce4 locus of Mtb (Rv3494c-Rv3503c) was found to have a role in cholesterol import (30). In addition to the ATPase and two permeases typical of classical ATP-binding cassette transporters, this operon encodes for eight proteins of unknown function (24). These genes bear considerable similarity to genes in some plants and Gram-negative bacteria, an observation that has prompted the suggestion that the encoded proteins are involved in transport across complicated cell walls and membranes (24). This system has been found to exhibit specificity for steroids with long hydrophobic side chains and uptake of other steroids is presumably managed by separate transport machinery (17).  15  1.2.3  Side chain degradation The aliphatic cholesterol side chain is broken down by a process similar to  mammalian !-oxidation (19). In mammals, a large multifunctional enzyme catalyzes several enzymatic steps in this pathway (61). In contrast, discrete enzymes are assumed to carry out these reactions in bacteria. Side chain degradation is initiated by hydroxylation of one of the two terminal methyl groups and subsequent oxidation to the carboxylic acid (62). A soluble P450 enzyme system is proposed to carry out this reaction (62). Early investigations in Nocardia species indicated the C27 steroid is transformed to a C22 steroid via standard fatty acid degradation with concomitant release of one molecule each of acetate and propionate (55). Complete side chain degradation was later recognized to proceed through formation of the C26-CoA thioester by an acyl-CoA synthetase (29). This transformation is followed by successive rounds of catalysis with an acyl-CoA dehydrogenase, hydratase, dehydrogenase, and thiolase to liberate one acetyl-CoA and one propionyl-CoA molecule (63). Liberation of a final propionyl-CoA molecule (via a retroaldol mechanism) forms the 17-ketosteroid (63). As it is currently predicted, the pathway does not involve intermediates with shortened, carboxylic acid-bearing side chains. Instead, side chains bearing the thioester, oxidized thioester, hydrated thioester, and ketonized thioester groups characteristic of !-oxidation are predicted to occur in cholesterol side chain degradation (63). Several side chain-degrading enzymes have recently been proposed to be encoded by genes in the Mtb cholesterol-degradation gene cluster (21). In Mtb, two P450s encoded by genes in the KstR1 regulon are candidates to catalyze the initial side chain transformation that forms the 26-carboxylic acid: Cyp125A1 (Rv3545c) and Cyp142A1 (Rv3518c) (Figure 1.4). The former protein is encoded by a gene in the putative igr operon alongside predicted !-oxidation genes (Rv3540c-3545c) (21,56,64). Two other igr genes, fadE28 and fadE29, have been demonstrated to catalyze 1,4-desaturation of 4-BNC-CoA (64). Coupled with the predicted functions of Rv3542c, Rv3541c, and ltp2 (Rv3540c), this result suggests that the genes following cyp125 in igr the operon encode enzymes that catalyze the transformation of 17-propionate-CoA steroids to their respective 17-keto derivatives (64). A putative !ketoacyl CoA thiolase (Rv3546) in the KstR regulon has been partially characterized (65). 16  Additionally, the fadD19 gene of Rhodococcus rhodocrous DSM43269 has been characterized as an ATP-dependent acyl-CoA synthetase required for degradation of 14branched sterol side chains (66). Characterization of FadD19 (Rv3515) of Mtb has demonstrated that this enzyme catalyzes formation of the initial C26-CoA thioester in Mtb cholesterol degradation (Casabon et al. unpublished data). 1.2.4  Ring degradation The individual steps leading to degradation of the cholesterol steroid nucleus in Mtb  and related actinomycetes are better characterized than side chain degradation. Beginning in 1962, Sih and coworkers published a series of papers outlining various aspects of steroid degradation. The described processes include the reactions leading to ring B cleavage (15,67,68), formation and cleavage of the ring A catechol intermediate (69,70), and processing of downstream metabolites (71). The enzymes responsible for individual steps in the steroid nucleus-degradation pathway, however, were not identified until much more recently. Genes encoding a bacterial steroid ring degradation pathway were described in C. testosteroni TA441 by Horinouchi et al. in 2003 (16) and 2004 (53). Identification of the R. jostii RHA1 and Mtb cholesterol-degradation gene clusters in 2007 enabled more complete characterizations of the enzymes participating in cholesterol ring degradation in these organisms (21). Through the combined efforts of several laboratories, the enzymatic reactions resulting in degradation of rings A and B are fairly well understood. Many of these enzymes are homologous to those that degrade other poly-cyclic organic compounds (eg. biphenyl) (16,17,21,72). The initial cleavage of the steroid core structure is achieved through three sequential reactions that result in opening of ring B (47) (Figure 1.5). The steroid nucleus is first transformed by oxidation of the C3 hydroxyl group to a ketone and concomitant isomerization of the !5 double bond to !4 (15,67,68). This has been proposed as the first reaction in bacterial cholesterol degradation (73,74). Depending on the organism, C3 oxidation is catalyzed by either the O2-dependent cholesterol oxidase (75-77) or the NAD(P)+-dependent 3"-hydroxysteroid dehydrogenase (3"-HSD) (Rv1106c) (73,78). 17  Although the genomes of several mycobacteria encode for proteins similar to both enzymes, the balance of evidence suggests that 3!-HSD catalyzes this reaction during cholesterol metabolism in Mtb (18,48,74,78). The next reactions in the pathway are desaturation of the C1-C2 bond and hydroxylation at the " position of C9. The former reaction is catalyzed by the flavin-containing 3-ketosteroid-#1-dehydrogenase (KstD) (Rv3537) (79-84). A twocomponent Rieske monooxygenase, 3-ketosteroid-9"-hydroxylase (KshAB) (Rv3526 and Rv3571) is proposed to catalyze the 9"-hydroxylation (85). Although the order of the two reactions is not clear, the cumulative actions of these two enzyme systems result in nonenzymatic cleavage of ring B and aromatization of ring A. In several organisms it appears that the two reactions can take place in either order (82,84) and the physiological substrates for both enzymes is a matter of debate (85-88). The next stage of cholesterol ring catabolism uses a strategy conserved in bacterial aromatic degradation: incorporation of O2-derived hydroxyl groups to form dihydroxylated metabolites cleaved by further dioxygenases (17) (Figure 1.5). Following ring B opening, ring A is hydroxylated on carbon 4 by the two-component flavin-dependent monooxygenase (HsaAB) (Rv3570c and Rv3567c) using reduced nicotinamide adenine dinucleotide (NADH) and O2 as cosubstrates and forming a catecholic intermediate (54). Ring A is then cleaved between C4 and C5 by the Fe2+-containing extradiol dioxygenase HsaC (Rv3568c) (72); thus both rings A and B are linearized while rings C and D remain intact. The meta-cleavage product-hydrolase HsaD (Rv3569c) catalyzes the next reaction to separate the 6-carbon fragment 2-hydroxyhexadienoate from rings C and D (89). The fate of the remaining molecule, including the intact C and D rings, has not been well studied. Degradation of the three carbon side chain on C8 has been proposed to proceed by another !-oxidation-like mechanism (90) involving a second acyl-CoA synthetase (91), dehydrogenase, and hydratase, and resulting in the release of acetyl-CoA (71,90). The enzymes responsible for these reactions, however, have yet to be identified. Similarly, the metabolic fate of cholesterol rings C and D is uncertain.  18  As mentioned above, cholesterol degradation takes place in both environmental and pathogenic organisms. The contributions of specific cholesterol-degradation enzymes to Mtb pathogenicity will be discussed below. 1.3  Mycobacterium tuberculosis The human pathogen Mycobacterium tuberculosis causes tuberculosis (TB), a disease  which kills 1.7 million people every year (92). About one third of all people are infected with Mtb, although only about one tenth of infected individuals will go on to develop active disease while the balance exhibit a latent and symptomless infection (92). Although the first antibiotic therapy for TB, streptomycin, was developed over 60 years ago (93), TB control on a global scale is still a significant challenge. One of the contributing factors to TB disease control is resistance to antibiotic therapies (92). The currently recommended drug course is at least six months of multi-drug therapy. The significant negative side effects of current drugs, long treatment times, and uncertain drug supply in developing nations are all factors leading to low compliance for antibiotic therapies and resultant drug resistance (92). Drug resistance in Mtb is rampant, and multidrug-resistant, extensively drug resistant, and extremely drug resistant strains have been reported in all sections of the globe (92). Multidrug-resistant strains are resistant to isoniazid and rifampicin and recommended treatment of these strains involves a two year long antibiotic drug course (92). Extensively drug resistant strains are resistant to both of the abovementioned first line drugs, plus any fluroquinone, streptomycin and any injectable therapy, while extremely resistant strains are resistant to all known therapies (92). The global burden of TB is also enhanced by the disease’s cooperative effect with the human immunodeficiency virus (HIV). Immunosuppression as a result of HIV infection can cause activation of latent Mtb infections; TB is the leading cause of death for AIDS patients (92). Despite a concentrated research effort in recent years, no new drugs have been approved for the treatment of TB in over 40 years. There are, however, promising results for new drugs and drug combinations now in clinical trials (94,95).  19  1.3.1  Related organisms Mtb is taxonomically classified with the monoderm actinomycete bacterial phylum  characterized by a high genomic G-C content. These bacteria range from soil-dwelling organisms like the non-pathogenic R. jostii RHA1 to intracellular pathogens like Mtb and Mycobacterium leprae. The former category of bacteria are abundant in many ecological environments and contribute to the global carbon cycle through their prolific catalytic abilities (17,45,96). Bacteria in this phylum are also well known for their production of secondary metabolites, including many antibiotics (96). Due to similarities between their metabolic pathways, several actinomycetes (eg. Rhodococcus and Nocardia species) have been used as model organisms to probe the metabolic pathways of Mtb. A table with the names of the strains discussed in this thesis and their taxonomic relationships to Mtb is provided as Appendix A.  Mycobacteria  Members of the genus Mycobacterium exhibit a thick, mycolic acid-containing cell wall (97). Within this genus, Mtb is part of a group of mycobacterial intracellular pathogens that also includes Mycobacterium avium, Mycobacterium marinum, and M. leprae. The latter organism is an obligate intracellular pathogen with a highly reduced genome that is often used to study gene essentiality in the context of intracellular survival. Many pathogenic mycobacteria are used as model organisms to investigate disease in specific host models (98). For example, granuloma formation has been studied in M. marinum infection in a live zebrafish model (99). Although not usually considered a pathogen, M. smegmatis has also been used as a genetic model organism for Mtb (98).  Mycobacterial cell envelope  Although taxonomically classified as Gram-positive bacteria based on sequence identity criteria, Mtb and other mycobacteria do not exhibit the cellular envelope typical of 20  Gram-positive organisms. Instead, they have a more complicated structure incorporating a pseudo periplasmic space formed by intercalation of mycolic acids on the exterior of the cell wall with lipids belonging to an outer capsule (97) (Figure 1.6). Like Gram-positive organisms, mycobacteria exhibit a cellular membrane comprised of a lipid bilayer, integral membrane proteins, and other associated molecules. A thick layer of peptidoglycan, similar to that found in Gram-positive bacteria, surrounds the cellular membrane (97). Unlike typical Gram-positive organisms, however, Mtb exhibits further components of the cell wall, and an outer layer of lipids and proteins separating the cell wall from the extracellular space (100).  Figure 1.6  Schematic of Mtb cellular envelope. From Jackson et al. 2007 (101). ! Elsevier, 2006, by  permission.  Like almost all characterized bacteria, mycobacteria exhibit a peptidoglycancontaining cell wall immediately outside of the plasma membrane. The cell wall core structure of mycobacteria, however, is characteristic and exhibits three distinct layers (listed in order from most proximal to most distal from the plasma membrane): peptidoglycan, arabinogalactan, and mycolic acids (97). The peptidoglycan of mycobacteria is essentially similar to that of other bacteria and incorporates "(1-4)-linked N-acetylglucosamine and Nacetylmuramic acid residues that form glycan strands cross-linked together with short 21  peptides. The mycobacterial peptidoglycan is distinct due to partial glycolylation of the glycan strand muramic acid residues and atypical peptide cross-links (97). There is also evidence that mycobacterial peptidoglycan is synthesized at the tips of the bacillus rather than through lateral wall extension like other bacteria (102). Mycobacteria have a layer of the polysaccharide arabinogalactan and a layer of mycolic acids covalently linked to the exterior surface of the peptidoglycan. The arabinogalactan is important for cell envelope integrity and serves to attach the mycolic acids to the cell wall core structure (97). The galactan chain is comprised of approximately 30 residues linked by alternating 1-5 and 1-6 glycosidic bonds with arabininan chains attached at three branch points. The arabinian chains are also about 30 residues long and are capped by ester linkages to mycolic acid residues. Mycolic acids make up the outer layer of the cell wall core structure and are the major determinant of cell wall permeability (97). These molecules comprise long chains (60-90 carbons) that form a waxy barrier around the cell. Another distinguishing feature of the mycobacterial cellular envelope not observed in Gram-positive strains is an outer layer of polysaccharides, lipids and proteins (97,101). This outer layer is associated with the cell wall core structure by intercalation with the mycolic acid residues protruding outward from the cell and arguably forms the outer barrier to a space functioning similar to that of the periplasmic space in Gram-negative bacteria (97,101). The lipids of this layer include a variety of multiple-methyl branched fatty acids including phthiocerol dimycocerates (PDIM), phenolic glycolipids, diacyltrehaloses, triacyltrehaloses, polyacyltrehaloses, and sulfatides. While the hydrophobic lipid tails interact with the hydrophobic mycolic acids, the sugars of the trehalose ester family of lipids are oriented away from the cell and interact with a hydrophilic polysaccharide capsule, the outmost layer of the mycobacterial cell envelope (101). Fast growing species like M. smegmatis express a porin-like protein that appears to be important for rapid nutrient uptake through the outer envelope layer, but Mtb lacks such a protein (97). Restriction of nutrient uptake due to the relative impermeability of the mycobacterial cell envelope has been suggested to be a limiting factor in bacterial growth rate (97). The thickened hydrophobic nature of the mycobacterial cell wall has also been proposed to act as a barrier to the potentially toxic surfactant effects of some steroid molecules and allow them to be metabolized safely.  22  Mycobacterium tuberculosis complex  Mtb shares significant genetic similarity with a group of organisms called the Mycobacterium tuberculosis complex, a group that includes Mycobacterium africanum, Mycobacterium bovis, Mycobacterium microti, Mycobacterium canetti, Mycobacterium pinnepetii, and Mycobacterium caprae (50). All members of this Mtb complex are believed to have derived from a common Mtb ancestor through multiple gene deletion events over the past 15000 years (103). Members of the Mtb complex have identical 16S rRNA and their proteins share greater than 99.9 % amino acid sequence identity. Mtb complex members all cause tuberculosis-like disease in various animal hosts. The exception is the avirulent Mycobacterium bovis Bacillus Calmette-Guerin (BCG), which was differentiated from a virulent M. bovis strain through 230 passages in liquid medium and which is used as a live attenuated vaccine against mycobacterial infection. Many strains of these species are studied for their virulence in humans or animals and are used as model organisms for studying Mtb infection. Obviously there are limitations to such studies as, notwithstanding the high level of DNA-homology, the range of hosts infected by these organisms indicates significant phenotypic differences. While Mtb represents only one species in the Mtb complex, it encompasses many strains that also exhibit phenotypically relevant genetic variation (50). Significant differences in the physiology of different strains have been detected, making studies investigating a single strain difficult to interpret in a wider context (104). Indeed, even maintenance of the reference strain H37Rv by different laboratories has been demonstrated to result in genotypic alterations with significant phenotypic consequences (105). This genomic heterogeneity has contributed to the phenomenon of studies performed in different laboratories giving contradictory results (50,104). 1.3.2  Infection cycle Mtb is a human pathogen and no significant part of its growth cycle takes place  outside of the human body (106). A diagram of the Mtb infection cycle is provided as Figure 1.7 (106). While the information in this and the following section reflect current theories of Mtb pathogenesis and the human immune response, they reflect both directly observed 23  phenomenon and conjectures based on in vitro experiments designed to probe the Mtb infection process. The bacterium is transmitted from host to host via the aerosol droplets created when individuals with active tuberculosis cough or sneeze. Host phagocytes engulf the bacterium and a granuloma or “tubercle” forms around them. Formation of this structure effectively contains the infection within discreet structures in the lung tissue and this stage of the infection that can persist for decades (106). In an estimated 5-10% of chronically infected individuals, this structure breaks down, releasing viable bacteria and renewing the infection cycle (106). Acute disease caused by Mtb is rare in adults and the vast majority of clinical cases are a result of a change in immune status leading to reactivation of chronic infection (107).  24  Figure 1.7  Mtb infection cycle. From Russell et al. 2010 (106). ! AAAS, 2010. Reprinted with permission  from AAAS.  The infection of a new host commences with phagocytosis of the bacterium by host alveolar macrophages (106). Upon phagocytosis of many other foreign particles phagosomelysozome fusion takes place leading to acidification of the lumen with concomitant release of various degradative enzymes. In the case of Mtb phagocytosis, however, the phagosome fails to fuse with lysozomes and instead undergoes only mild acidifiction (pH 6.4) (108). The 25  Mtb-containing phagosome therefore remains a relatively unhostile environment in which the bacilli are able to survive and replicate (109). This membrane-bound compartment is not, however, completely isolated from the rest of the cell. The phagosome interacts with the early endosomal pathway and in this way exchanges molecules with the host cell (107). Intracellular Mtb induces a proinflammatory response in its host macrophage via the toll-like receptor-agonists on the bacterial cell surface (107) and this leads to the cytokine production that is important for cell recruitment and granuloma formation. The human tubercular granuloma is a stratified structure that forms around Mtbcontaining macrophages and is characteristic of chronic infection (106,107). This structure comprises multiple macrophage and lymphocyte cell types, as well as a fibrous cuff of extracellular matrix material. Directly following phagocytosis, the host macrophage migrates from the airway to the host epithelium (107). The granuloma is formed through remodeling of the local lung tissue and recruitment of various immune cells. This process is the result of a robust immune response and is mediated by a wide variety of chemokines and cytokines resulting from a signaling cascade initiated by the Mtb-infected macrophages (107). Macrophages with intracellular bacilli produce tumour necrosis factor (TNF)-! and inflammatory chemokines, which leads to the recruitment of successive waves of neutrophils, macrophages, mature dendritic cells, T cells, NK cells, and B cells (107). These arriving immune cells contribute to the growing granuloma structure and produce additional inflammatory cytokines. The order of immune cell recruitment also helps form the stratified structure of the granuloma. A fibrous extracellular matrix forms around the periphery of the structure, separating macrophages, foamy macrophages, giant cells, and granulocytes on the interior from the peripheral lymphocytes, and contributing to the stratified appearance (107). Early stages of granuloma formation are characterized by significant neovascularization, while maturation of “active” granulomas, including thickening of the fibrous sheath, results in less vascular penetrance to the interior of the structure (106,107). Mtb bacilli are generally understood to exist within the phagocytes at the interior of the granuloma but have also been observed in macrophages outside the fibrous capsule at the site of lymphocyte proliferation (107). Although granuloma development is thought to follow a specific progression, discrete lesions in the same patient can exhibit vastly different stages of granuloma development 26  (110). Lung granulomas are often the only such structures considered in the literature, however these bodies can also form in other tissues (most notably the lung draining lymph nodes) (111). Most Mtb infections do not result in acute clinical symptoms and instead enter a latency phase often associated with Mtb dormancy after the initial immune response and granuloma formation (112). The term “latent infection” describes a patient who harbors a population of viable Mtb bacilli but who does not exhibit clinical symptoms and is noninfectious (112). “Dormancy” describes a physiological state of the bacterium usually understood to include non-replication and a significantly depressed or quiescent metabolic rate (112). This state of the disease has been difficult to study, partially due to a lack of appropriate models. Many animal models of latency and in vitro models of bacterial dormancy have been developed (113). Viable but non-replicating bacteria have been identified in infections of transiently antibiotic-treated mice (114) and infection of rabbits appears to elicit human-like granuloma formation and latency (115); however, no animal model has been shown to faithfully reproduce the conditions of the human granuloma (97,112). A non-human primate model using macaque monkeys shows promise as a more faithful infection model, but the spontaneous reactivation rate is much higher than in humans and the metabolic state of bacteria during latent infection remains to be determined (112,116). Moreover, the macaque model is limited by high monetary costs and limited reagents. In vitro models of dormancy have also been developed. The most thoroughly used of these is probably the Wayne model in which progressive oxygen depletion by bacilli growing in a sealed vessel induces a non-replicative state (117). Other models inducing dormancy-like states through hypoxia (through a protocol distinct from the Wayne model) (118), nitric oxide, nutrient depletion (119), depressed pH, or a combination of these stresses have also been developed (112). Latent infections are often described in terms of quiescence of both the bacilli and the host immune response, but there is evidence to suggest that an ongoing balance between bacterial growth and immune response occurs throughout infection. Perhaps the simplest demonstration of this is the predictable reactivation of infection upon immune system  27  compromization. Such a change in immune status results in reduction of CD4+ T cell numbers and associated cytokine production and shifts the balance in favour of bacterial survival and growth (112). There is also evidence from an unstable plasmid-based bacterial “replication clock” that bacterial growth and death occurs at a steady and balanced pace throughout latent infection in a mouse model (120). Very low replication rates are also indicated by DNA mutation rates during latent infection in humans (118,121). Other evidence supports the hypothesis that the bacterium enters a dormant state during latent infection as a response to stress induced by the intraphagasomal environment, or via a stochastic mechanism (112). The latter hypothesis can be incorporated into the theory that latency is maintained by a population of dormant “persister” cells (112). Such variation in physiological states within an Mtb population has been observed within macrophages (122). Additionally, various specialized mechanisms have been demonstrated to allow for maintained viability through non-replicative and non-sporulative states induced by various in vitro models (117-119). The occurrence of such mechanisms in Mtb suggests that a dormant bacterial state plays a role in human infection. Some of these dormancy survival-associated genes have also been demonstrated to be upregulated in various infection models including human infection (118). In either case, the mechanisms enabling Mtb to establish, survive, and recover from latent infections are an area of concentrated investigation for anti-mycobacterial therapeutic development (112). While the stratified granuloma appears to be a significant factor for the maintenance of latent infection, transmission of Mtb to a new host requires the breakdown of this structure to release viable bacilli into the airways (107). This process is initiated by development of a necrotic core comprised of host and bacteria-derived cellular debris and free viable bacilli within the granuloma (107). Consistent with the finding that the environment within the granuloma is hypoxic (123), reduction of vascularization within this structure precipitates necrosis. Structural deterioration in the necrotic centre eventually leads to caseation of the granuloma, a process involving decay of the internal and external structure and eventual rupture (107). This process results in the release of viable bacilli into the airways and pulmonary veins and leads to transmission to new hosts and dissemination within the current host.  28  Mtb is known to cause disease in almost any organ of the body. These conditions are referred to as extrapulmonary TB. Accumulation of bacilli in the lymph nodes of infected patients is especially common, although wider spread through the circulatory system is also observed. Such dissemination is proposed to be driven by bacterial mechanisms (99). Although extrapulmonary TB is a significant source of mortality (especially in children and HIV-infected individuals) (92), biochemical and genetic investigations into Mtb physiology have focused almost exclusively on studies of pulmonary TB.  Human immune response  Upon infection of a naïve host with Mtb, a response involving both innate and adaptive immunity is initiated. Throughout Mtb infection cytokines play a key role in mediating both adaptive and innate immunity, with interferon-! (IFN-!) playing a central role (111). Neutrophils are typically the first cells recruited by the infected alveolar macrophages, likely through signaling mediated by the chemokine interleukin 8. Although neutrophils typically mount an antibacterial immune response though phagocytosis and intracellular killing mechanisms, in Mtb infection these leukocytes likely support the adaptive immune system and amplify cellular recruitment through cytokine signaling (111). Chemokine signals also recruit blood monocytes that differentiate into phagocytic macrophages upon reaching the site of infection (111). Macrophages are “activated” by IFN-!, and these activated phagocytes are thought to be the primary cells responsible for bacterial killing during infection (111). In fact, strong activation of macrophages by IFN-! in vitro is able to overcome the survival mechanisms of intracellular Mtb bacilli to effect bacterial killing. Natural killer cells are recruited to the developing granuloma and recognize infected macrophage via lipid antigen presentation on CD1 (124). Although natural killer cells are competent to lyse Mtb-infected macrophages in vitro, their in vivo role in the immune response to Mtb is thought to involve production of IFN-! and stimulation of further cytokine production in the macrophages (125).  29  Antigen presentation via the MHC Class I, MHC Class II and CD1 systems precipitate the activation of adaptive immune cells in the lung draining lymph nodes (111). This process includes recruitment of CD4+ and CD8+ T cells to the tubercular lesion and initiation of immunoglobulin G production by B cells. While both CD4+ and CD8+ T cells are evident in tuberculosis infection, CD4+ cells seem to play a dominant role. CD4+ T cells recognize peptide antigens presented on MHC Class II molecules that derive their cargo from endocytic vesicles, including Mtb-containing phagosomes (111). Once an infected macrophage is recognized, CD4+ T cells release cytokines, with IFN-! dominant among them. These signals activate bactericidal mechanisms within the macrophage, promoting bacterial killing (111). The importance of CD4+ lymphocytes to the control of Mtb infection is underlined by the increased susceptibility to infection seen in CD4+- and MHC Class IIdeficient mice (126) and in HIV+ patients with depleted CD4+ T cells (60). There are two sub-categories of CD4+ T cells: TH1 and TH2 cells. The former is involved in the abovementioned macrophage activation via IFN-!, while the latter inhibits excessive macrophage activation and promotes B cell-mediated humoural immunity (127). The contribution of CD8+ T cells to antimycobacterial immunity is underlined by the high levels of Mtb-specific CD8+ T cells exhibited by patients with latent infections (128). The role of these cells in the infection process includes cytokine-mediated and cytotoxic effects that have been demonstrated in a variety of model organisms (111). These cells recognize both peptide antigens on MHC Class I molecules and lipid antigens presented by CD1 molecules. Both of these presentation systems derive their cargo from cytoplasmic sources, making the exact mechanism by which CD8+ cells obtain Mtb antigens unclear (111). Although data on phagosome escape is somewhat contradictory (104), some studies have suggested that Mtb secretes specific proteins enabling entry into the host cytoplasm (129). In considering the signaling cytokines released by cells in both the adaptive and innate immune response it is important to remember that while some effects are attributable to specific “signal and response” interactions between defined cell types, all of the signals also reinforce the overall immune response by providing supporting roles in activation or recruitment of other immunity components.  30  Cytokine signaling, dominated by IFN-!, ultimately results in activation of the phagocytic macrophages to trigger bactericidal mechanisms within the phagosome (111). Macrophages kill phagocytotically-aquired intracellular bacteria through acidification of the phagosome, secretion of degradative hydrolases and oxidative species into the lumen, and subsequent fusion with pre-formed lysozomes (130). The latter results in exposure of the bacterium to low pH, hydrolases, proteases, super oxide dismutase, and lysozymes, and ultimately leads to bacterial death (130). A major bactericidal mechanism in the maturing phagosome is production of nitric oxide (NO) and other reactive nitrogen intermediates (131). Specifically, NO and superoxide (O2-) are both introduced into the phagosomal lumen where they react to form the powerfully oxidizing peroxynitrite anion (ONOO-). This compound interacts with bacterial protein methionine residues to form methionine sulfoxide and causes other oxidative damage (131). These stresses are reflected in studies examining transcription patterns of intracellular Mtb that suggest an intraphagosomal environment limited in nutrients and presenting oxidative and nitrosative stresses (49-51,104). Other stresses reflected in these studies include depleted levels of carbohydrate carbon sources and iron. The importance of B-cell mediated humoural immunity in protection against Mtb has been historically debated. Although a live vaccine of M. bovis BCG is protective for extrapulmonary childhood tuberculosis in some populations, geographic inconsistency and lack of protection in adult or pulmonary tuberculosis has lead to questions about the universal importance of antibodies in Mtb infection (106). These observations notwithstanding, monoclonal antibodies to several Mtb antigens have been shown to decrease the severity of disease in mice (111). 1.3.3  Determinants of Mtb pathogenicity The determinants of Mtb pathogenicity have been studied genetically since  differentiation of strain H37 by repeated cultures of single colonies on defined media in the 1930’s (132). Repeated culture of differentiated strains to apparent stability resulted in formation of the virulent H37Rv and avirulent H37Ra strains (133). Similarly, the M. bovis  31  strain BCG is an avirulent strain differentiated from a parent virulent strain through 230 successive subcultures in glycerinated beef bile over 13 years (134). Careful genetic testing identified that a region of the M. bovis BCG genome, termed the Region of Difference 1 (RD1) and conserved in virulent Mtb, had been lost in BCG with respect to other M. bovis strains (135). It was therefore concluded that this region contained some major determinants of mycobacterial virulence. Today, higher resolution genetic and transcriptomic techniques are used to investigate the genes and pathways important for the infection process. Microarray transcriptomic studies have been carried out to probe the physiological state of Mtb during growth in in vivo infection models (49-51,136) and in vitro models of dormancy (119,136,137). Transposon Site Hybridization (TraSH) studies identifying mutants with growth defects in macrophages or animal models have also been employed to probe the importance of specific genes (22,52,138,139). Interestingly, many of the genes required for survival in infection models have been found to be constitutively expressed and are therefore are not identified in comparative transcriptomic studies (22). This finding underlines the bacterium’s evolutionary adaptation to an intracellular pathogenic lifestyle. Identification of “essential” genes for infection has been further challenged by inconsistency in the results of such genome-wide screens between strains, models, and experimental approach (50). Nonetheless, it is clear that Mtb uses several strategies of immune evasion that act on different timescales within the context of infection. While some so-called virulence factors are important in the acute stage of infection, separate factors may be involved in latent infection and reactivation. Indeed, it has been proposed that the regulatory mechanisms controlling the temporal expression of various genes can themselves be considered important virulence factors (140).  Modulation of intracellular killing mechanisms  A major category of Mtb virulence factor contributes to the mechanisms the bacterium uses to interact with human immune cells to avoid sterilization. These mechanisms include the arrest of phagosome maturation to avoid the intracellular killing mechanisms of the macrophage and prevent the bactericidal consequences of lysozome fusion (111,138).  32  The processes of phagosome maturation and lysozomal fusion require acquisition and interaction of several macrophage factors at the phagosome membrane and interference with any of these components can result in phagosome maturation arrest (141). Unsurprisingly an Mtb transposon mutant screen probing this process pointed to the involvement of several Mtb physiological properties including replication rate, components of the cellular envelope, and specific export systems (138). Several mechanisms involving mycobacterial glycolipids have been proposed to facilitate blocking of phagosome maturation by intracellular Mtb (142), with the strongest evidence being for the involvement of the mycobacterial cell wall component mannosylated lipoarabinomannan (143). This glycolipid appears to be shed from the outer bacterial envelope and incorporated into lipid rafts of the macrophage membrane where it has several putative roles in disrupting the phagosome maturation pathway (111,143). Other mechanisms such as those involving secretion of the bacterial protein kinase PknG (144), activation of host mitogen-activated protein kinase, and phagosome coronin-1 retention (145) have also been proposed to inhibit phagosome maturation (111). Mtb is able to resist intraphagosomal sterilization by employing remediatory mechanisms for nitrogen-based oxidants and through interfering with MHC II-mediated antigen presentation (131). Mtb manages oxidative stresses like ONOO- with a range of enzymes including an NADH-dependent peroxidase and a peroxynitrite reductase comprised of the proteins dihydrolipoamide dehydrogenase (Lpd), dihydrolipoamide succinyltransferase (SucB), peroxiredoxin alkyl hydroperoxide reductase (AhpC), and the thioredoxin-like protein AhpD (146). There is also evidence that the mycobacterial proteosome plays a role in remediating NO-related damage (144). Mtb also has be ability to mediate the CD4+ response by inhibiting IFN-!-induced upregulation of MHC II (131). With this antigen-presentation system inhibited, CD4+ cells are unable to recognize infected macrophages and are limited in their capacity for macrophage activation through cytokine production. Mtb is only able to manipulate the MHC II system, however, and is not able to effect similar control of CD8+ cells and natural killer cells, which use the MHC I and CD1 antigen presentation systems (111). In addition to inhibiting the CD4+ response by disrupting antigen presentation, Mtb is able to shift the major immune response from the cell-mediated TH1 response to the TH2 response through suppression of CD4+ TH1 differentiation (111). The TH1 response is typified 33  by proinflammatory cytokine signaling with IFN-!, TNF-" and interleukin 12, while the TH2 response promotes humoural immunity and has a suppressive effect on inflammation by producing the cytokines interleukin 4, interleukin 5, and interleukin 13 (147). The TH2 response is much less effective in suppressing intracellular infection.  Cell envelope  Many of the unique aspects of the mycobacterial cellular envelope have been linked to pathogenicity or innate immunity to antibiotics and host defenses (97,101). For example, the glycolylation of muramic acid residues in mycobacterial peptidoglycan has been shown to impart increased resistance to both lysozyme and #-lactam antibiotics (148). Similarly, the atypical peptidoglycan peptide crosslinks may confer resistance to specific hydrolases (97). The decreased permeability for many compounds attributable to the cell wall-associated mycolic acids has also been suggested to confer a degree of innate insensitivity to many small effector molecules (97). Many of the multimethyl-branched fatty acids in the outer lipid layer of the mycobacterial envelope are found exclusively in pathogenic mycobacteria and are absent in attenuated strains (101). Specific associations with this class of lipid and virulence-associated functions include links between PDIM and distinct membrane permeability properties (149), protection against reactive nitrogen intermediates (150), and decreased inflammatory cytokine production (149); sulfatides and inhibition of phagosomelysozome fusion, innate immunity proinflammatory response, and intraphagosomal secretion of oxidative species (101); and acyltrehaloses and inhibition of T cell proliferation (151). Mtb also possess specialized virulence-associated secretion systems to aid in the translocation of molecules across the complicated cellular envelope. One of these is encoded in the RD1 locus and is suggested to have evolved from a DNA conjugation system (144,152). The genes encoding proteins in this system are required during infection in a mouse model (52).  34  Latent infection and dormancy  A major impediment to global tuberculosis control is the Mtb’s ability to enter an apparently altered metabolic state characteristic of a latent infection (112). Bacterial dormancy has been proposed to contribute to the large proportion of the human population exhibiting latent Mtb infections, thus maintaining large pools of potential disease (112). Dormancy has also been proposed to account for the innate drug resistance characteristic of Mtb and the requisite long courses of drug therapy (112). In support of this theory, nonreplicating bacteria induced in vitro through nutrient starvation or oxygen depletion have been found to remain viable and exhibit significant resistance to antibiotics (117,153). The importance of dormancy to mycobacterial survival inside the phagosome is supported by identification of dormancy-associated genes in screens of upregulated or required genes within this environment (51). Current models suggest that Mtb needs to transition between non-respiring and respiring states to enter and subsequently recover from a state of dormancy (118). These transitions are enabled, at least in part, by proteins regulated by the DosR regulon (118,154). Proteins in this regulon facilitate a transition into the dormant metabolic state in response to hypoxia, nitric oxide, and carbon monoxide (49,104,154), and have been suggested to contribute to slowing the rate of aerobic respiration (118). The three-gene dosRST operon (the DosR operon) (118) encodes the transcription factor response regulator DosR, and the two DosR-regulating sensor kinases DosS and DosT. While both kinases bind heme as a prosthetic group aiding in environmental sensing, DosT is a hypoxia sensor inhibited by O2 bound to the heme iron and DosS is a redox sensor with a much greater autophosphorylation activity in the presence of a ferrous heme iron. Both DosS and DosT are can bind CO and NO with their heme iron ions and binding of both compounds increases autophosphorylation activity in the kinases (154). Functional DosR gene regulation has been shown to be required for transition into and out of dormancy and mutants of the necessary components exhibit loss of viability during these transitions (118).  35  In addition to the DosR regulon, other specific factors required for dormancy have been identified. For example, specific enzymes referred to as resuscitation-promoting factors are required as a supplement in many models to retain viability during transition from dormancy. These proteins are proposed to facilitate remodeling of the bacterial wall to reverse the thickening of this structure observed during dormancy and enable growth (97,155). Unlike most intracellular pathogens, the Mtb genome also includes a very high number of toxin-antitoxin loci that have been proposed to contribute to dormancy. These systems are upregulated in in vitro dormancy models and could contribute to Mtb dormancy through production of a bacteristatic toxin and repression of antitoxin production under specific environmental conditions (112). The toxin-antitoxin loci of Mtb have also been proposed to play a role in regulating replication and translation under nutrient stress (97).  Lipid-biased metabolism  One of the emerging trends in Mtb virulence research is a growing understanding that lipid metabolism plays a crucial role in enabling survival in vivo. Early indications that lipid metabolism is important during infection included survival defects in strains deficient in lipolysis-related genes, and alterations in intracellular Mtb cellular lipids compared to in vitro grown cells (104). The biosynthesis of specific lipids has also long been recognized to be important for host-pathogen interactions, with compounds such as PDIM often referred to as “virulence lipids” (101). More recently, transcriptomic studies in both in vitro and in vivo models have suggested that Mtb exhibits increases in triacylglycerol biosynthesis accompanied by reduced production of energy and biosynthetic precursors via the tricarboxylic acid cycle and other central metabolic pathways during infection (136,139). This diversion of carbon from energy production to lipid anabolism has been associated with decreased growth rate and increased innate antibiotic resistance of intracellular mycobacteria (139). Studies to date have revealed that intracellular Mtb viability depends on both catabolic and anabolic strategies with a lipid bias.  36  Studies probing the characteristics of the intraphagosomal environment have suggested that this milieu is poor in carbohydrate carbon sources (49,51,52). Indeed, the 1956 investigation by Bloch and Segal (156) demonstrated that bacterial cells isolated from mouse lung were much more metabolically responsive to fatty acids than to carbohydrates. These authors proposed that this shift was an adaptation to growth within the host (156). Such findings support the hypothesis that lipids, which are presumably available to the bacterium from the host phagosomal membrane and other macrophage-derived sources, constitute an important nutrient for intracellular Mtb (48,157). In support of this theory, microarray analysis of human lung granulomas showed significant upregulation of genes involved in lipid sequestration and metabolism in the host tissue (157). Transposon site hybridization screens in a mouse model identified several lipid catabolism genes as important for bacterial survival (52). The lipid-biased catabolism of intracellular Mtb is not surprising, given our understanding of the nutrients available to the bacterium in the macrophage or granuloma caseum. Material derived from the caseum of human granulomas was found to include significant amounts of cholesterol, cholesteryl esters, triacylglycerols, and lactosylceramide (157). Foam cell formation induced by Mtb infection results in alterations in macrophage physiology leading to accumulation of intracellular lipid droplets rich in cholesterol and triacylglycerol (122,157,158). This material is presumed to form the bulk of caseous debris given the observation that the lipid profile of human caseum correlates well with these cells (130,157). These findings suggest that Mtb manipulates the host macrophage metabolism, possibly via Mtb-derived oxygenated mycolic acids (158), and causes changes in lipidprocessing and accumulation of these carbon sources in the granuloma environment (157). Furthermore, intracellular accumulation of host-derived triacylglycerol droplets (122) and incorporation of host-derived cholesterol in the bacterial cell wall (159) have both been demonstrated for Mtb. It is notable that Mtb lipid-biased catabolism is selectively observed during intracellular growth, indicating the existence of regulatory mechanisms controlling the expression of the necessary genes (51,156). Mtb has also been demonstrated to co-catabolize multiple carbon sources for maximum growth in vitro (160) and genes for carbohydrate import are required for macrophage and mouse survival (22).  37  There are significant central metabolic consequences associated with a primarily lipid-based catabolism, including production of large pools of propionyl-CoA and the accumulation of reduced cofactors (48,161-163). The intracellular concentrations of such metabolites are characteristic of the specialized metabolic state of Mtb during infection and require specific mechanisms to maintain biological homeostasis. For example, enzymes of the methyl citrate cycle, including isocitrate lyase, are required for Mtb survival during infection (162) and are upregulated in the macrophage environment (122). This cycle is involved in processing propionyl-CoA and its activity results in the production of pyruvate and succinate. Enzymes in this pathway in Mtb have been specifically linked to metabolizing propionyl-CoA units derived from metabolism of cholesterol (48), methyl-branched fatty acids, and odd-chain-length fatty acids (104). Such processing is absolutely necessary as accumulation of propionyl-CoA is toxic to Mtb in vivo (163). It has even been suggested that transcription of isocitrate lyase may be rate-limiting for growth on cholesterol (48). Another propionyl-CoA-processing mechanism used by the bacterium that also serves to maintain redox balance (161), is direct incorporation of propionate units from this molecule into cellular lipids (101). Modulation of lipid anabolism by WhiB3 in response to redox imbalance has been shown to be important for regulation of cytokine production in the innate immune response (161).  Other proposed factors in pathogenicity  In addition to those listed above, there have been many other factors proposed to play a role in Mtb pathogenicity. Given the length of human-Mtb co-evolution, the variety of Mtb virulence factors is unsurprising and one could postulate that the vast majority of Mtb genes contribute in some way to its pathogenicity in humans. Nonetheless, some specific factors proposed to be important for Mtb pathogenicity are involved in regulation and maintenance of general bacterial physiology. The ability to alter gene expression in response to many environmental signals, with pH chief among them (49), indicates that the sensory and signaling systems governing these responses are key to mycobacterial survival inside the host. Mtb incorporates thirteen sigma factors that contribute to pathogenicity by enabling  38  transcriptional control of pathogenicity-related genes under a variety of environmental conditions (97). Mtb also expresses several members the WhiB family of proteins, found only in actinomycetes (97). WhiB proteins with putative pathogenic roles include: WhiB4, a putative redox sensor (164); WhiB2, a putative regulator of cell division; WhiB3, which is proposed to interact with the virulence-associated transcription factor RpoV (165); and WhiB7, which is induced in the presence of eukaryotic lipids and subinhibitory antibiotic concentrations (166). WhiB7 is proposed to induce expression of a regulon promoting resistance to a wide range of antibiotics. The relatively slow growth rate of Mtb has also been proposed to contribute to bacterial virulence (167,168), as most pathogenic mycobacteria are slow-growing (97). Although exhibiting a constitutively lower growth rate than nonpathogenic species like M. smegmatis, Mtb can nonetheless modulate growth rate through a variety of mechanisms. The mce1 locus was found to be a key component enabling the shift between slow and fast growth rates (167) and intraphagosomal survival (22). In fact, mutation of the putative transporters encoded in the mce1, mce2, and mce3 loci all resulted in attenuated virulence in a mouse model (169). Genes of the BER system of DNA repair have also been identified as important survival factors in vivo (52), presumably by maintaining integrity of the bacterial chromosome in the face of intraphagosomal stresses.  Pathogenicity vs. host immune response  Many of the determinants of Mtb pathogenicity represent an interplay between bacterial- and human-centric mechanisms. Although most of the available literature attempts to deconvolute this situation by attributing processes either to “bacterial virulence” or “human immune system,” co-evolution of these two species has resulted in much more subtlety. Formation of the granuloma, for example, is proposed to benefit the host through infection containment, but also benefits the bacteria by providing a level of protection from inflammatory cytokines and providing a mechanism for infection of naïve hosts (170). In fact, the compromised immune systems of HIV-positive people can actually lower Mtb transmission rates as human immunity plays an important role in granuloma formation and the establishment of infection (106). In a zebrafish model of M. marinum, the bacterium was  39  found to contribute to granuloma formation through RD1-dependent macrophage recruitment and direction (99). The same authors also posited that macrophages are a vehicle for pathogen hematogenous dissemination through phagocytosis and subsequent migration (99). Indeed, it has been proposed that the evolutionary success of Mtb can be in great part credited to its propensity to establish non-symptomatic latent infections that enable maintenance of a significant pool of viable bacteria in its host species (104). The “containment” of the bacterial infection within the granuloma has also been touted as an effective immune response, sparing the host TB symptoms. It is not surprising, therefore, that disease progression is mediated by both host and bacterial factors. It is therefore likely that the bacterium is adapted to modulate the host immune response rather than eliminate or circumvent it altogether.  1.3.4  Role of cholesterol in Mtb virulence The 2007 discovery of the Mtb cholesterol degradation cluster (21) initiated a series  of investigations into the contribution of cholesterol to mycobacterial pathogenicity. The putative role of cholesterol in the ability of Mtb to survive infection and cause disease is supported by the fact that several genes in the cholesterol degradation pathway have been identified in screens for Mtb mutants deficient for growth or survival under various conditions simulating human infection (22,48,52) (Figure 1.3) and in transcriptomic studies of genes upregulated during intracellular growth (49-51). The abundance of host-derived cholesterol in Mtb-containing macrophages (158) and within the granuloma (157) is suggestive, given the low availability of non-lipid carbon sources generally believed to characterize these environments. That Mtb has maintained the capacity for cholesterol metabolism over the course of evolution as a human pathogen is also suggestive of a role for this pathway in Mtb pathogenicity.  40  Host cholesterol and establishment of infection  Cholesterol has been shown to be important in the infection cycle of Mtb as a component in several host-mediated processes. Some broad observations have also been made regarding the status of the host with respect to cholesterol. The findings that high dietary cholesterol results in an increase in the number of pulmonary bacilli (171) and that elevated blood cholesterol is responsible for a defect in the antimycobacterial immune response (172) are suggestive of a role for host cholesterol in Mtb infection and immunity, but do not point to any specific mechanisms. Specific studies probing the intersection of host cholesterol homeostatis and Mtb infections have shed some light on the ways in which this lipid contributes to the infection process. The presence of cholesterol in the host membrane has been found to be necessary for mycobacterial phagocytosis (173,174). Furthermore, uptake of Mtb by macrophages specifically requires accumulation of cholesterol in lipid rafts at the site of uptake, and this requirement appears to be specific for mycobacteria (144). The mce4A gene in the cholesterol degradation gene cluster also appears to play a role in uptake as exogenous production of its product in non-pathogenic Escherichia coli facilitates uptake by non-phagocytic HeLa cells (175). Cholesterol is essential for association of coronin 1 on the cytoplasmic side of the phagosomal membrane, a condition that appears to inhibit phagosome-lysozome fusion in mice and is also specific to mycobacteria-containing phagosomes (144). Indeed, cholesterol depletion of macrophages with established intracellular mycobacteria results in phagosome maturation and bacterial killing (176).  Mtb cholesterol catabolism  While cholesterol is required for host membrane-mediated processes, cholesterol catabolism for the production of energy and biosynthetic precursors has also been proposed to play a significant role in Mtb virulence (18). Targeted knockouts of several of the genes required for this process have been used to probe its role in Mtb pathogenicity. Defects in cholesterol import resulting from disruption of an ATP-binding cassette transporter component (Rv3501c) in the mce4 locus led to severe attenuation of growth in mouse lungs  41  during later stages of infection (30). The 6-gene igr operon (named for a defect in intracellular growth) within the KstR regulon encompasses genes Rv3540c-Rv3545c that have been predicted to have a role in degradation of the cholesterol side chain (56). Deletion of this operon resulted in abrogation of in vitro growth on cholesterol (56) and attenuation of growth in macrophages and in a mouse infection model (177). The thiolase encoded by the fadA5 gene (Rv3546) is similarly predicted to have a role in cholesterol side chain degradation and is also required for full virulence in a mouse model (65). Additionally, enzymes involved in the degradation of the ring structure of Mtb have been demonstrated to have a profound impact on bacterial pathogenicity. In particular, deletion of two genes involved in B-ring cleavage (Rv3526 and Rv3571) resulted in severe growth defects in both mouse and macrophage infection models (31) and deletion of the ring-cleaving dioxygenase HsaC (Rv3568) caused attenuation of virulence in a mouse model and decreased pathology in a guinea pig infection model (72). Generally, mutation of enzymes involved in early steps in cholesterol degradation has a greater deleterious effect on Mtb growth on cholesterol than those with a role in downstream processes (48). These findings notwithstanding, mutation of the gene encoding the first ring-transforming enzyme 3!-HSD (Rv1106c) does not affect bacterial growth in macrophages nor pathology in a guinea pig infection model (178). In the related pathogen, Rhodococcus equi, enzymes required for 9,17-dioxo-1,2,3,4,10,19hexanorandrostan-5-oic acid (DOHNAA, Figure 1.5 7 R = keto group) degradation have a profound effect on pathogenicity in a horse model, so much so that strains deficient in this capability have been proposed as valuable vaccine prospects (90). Interestingly, the fate of cholesterol-derived carbon in Mtb has been found to be dependent on the origin of the carbon atom within the cholesterol structure; carbon from the terminus of the side chain is incorporated into cellular lipids while carbon from ring A is completely mineralized to CO2 (30).  The role of cholesterol catabolism in virulence  While there is a growing body of knowledge indicating that Mtb cholesterol catabolism plays an important role in pathogenicity, determination of which functions of  42  cholesterol breakdown are important for virulence vs. which constitute necessary consequences of other processes has proven difficult. For example, incorporation of cholesterol-derived carbon into lipid species could potentially be important for a pathogenically-associated envelope composition or could contribute to alleviation of redox stress that is proposed to result from fatty acid catabolism (101,161). In addition, there have been several non-nutritional functions proposed for processes or intermediates involved in cholesterol breakdown. For example, it has been suggested that cholesterol uptake by Mtb may be an important signal indicating macrophage entry and resulting in induction of specific virulence-associated genes (24). It has also been suggested that binding of host cholesterol by proteins in the Mce4 system could help the bacterium manipulate its surroundings by specific localization within the host or communication with the host cell via membrane alteration (24). Similarly, the HsaC enzyme is proposed to have a role in dissemination of Mtb within the host, suggesting that cholesterol degradation may be involved with macrophage manipulation (72). Cholesterol with an oxygenated side chain modulates cholesterol homeostasis in mammalian cells (179) and a role for cholesterol degradation intermediates in communication with the host cell is possible. Mtb cholesterol accumulation in the free-lipid zone of its outer cell wall has also been shown to alter the membrane properties, including making it less permeable to some small molecules (159). The significance of cholesterol degradation to pathogenicity is uncertain, as is the role of the process in the overall physiology of Mtb. One way to explore the cholesterol degradation pathway’s significance in both contexts is to investigate individual enzymes that act in this pathway and use their physiologic properties as a probe to look at the greater process. One class of enzyme that is represented more than once in the actinomycete cholesterol degradation pathway is the iron-dependent monooxygenases. 1.4  Oxygenases Oxygenation of organic compounds is a powerful chemical process involved in both  catabolic and anabolic processes. Incorporation of dioxygen-derived oxygen atoms is an effective method for activating various organic compounds and can lower energy barriers for  43  further reactions. This strategy consequently plays a major role in the global carbon cycle through facilitating reclamation of many otherwise recalcitrant organic compounds (180). Unsurprisingly, enzymes with many different catalytic mechanisms have evolved to perform these nearly ubiquitous reactions. The reaction between triplet O2 (S=1) and a singlet organic cosubstrate (S=0) is thermodynamically favourable (181) but spin forbidden and therefore kinetically slow (182). While presenting a chemical challenge for biological systems using this reaction, the spin-forbidden reaction also allows for control of oxidation and oxygenation of organic compounds by O2 by limiting their spontaneous reactivity (180). Such control increases the usefulness of these reactions in a biological context. Enzymatic catalysis speeds oxygenase reactions by orders of magnitude through activation of the organic substrate to lower the energy barrier for attack by O2 or through activation of O2 using redox chemistry. The latter is accomplished by reductive O2 activation and avoids the highly endothermic alternative: inversion of an oxygen electron to give the singlet species (180). The oxygenase redox chemistry is often accomplished with the aid of organic or inorganic cofactors (182) and flavins, porphyrins, quinones, various vitamins, and metal ions are all examples of cosubstrates or prosthetic groups used by oxygenase systems (183). Enzymatic activation of O2 necessarily produces oxygen species with the potential to cause oxidative damage to other cellular components. For this reason, oxygenases are both spatially and temporally regulated so that reactive oxygen intermediates are only generated in the presence of productive cosubstrates (180). 1.4.1  Mononuclear iron-containing oxygenases A component of many oxygenase prosthetic groups is the iron ion. This metal ion’s  relative environmental abundance, biologically accessible redox potentials, multiple oxidation states, and variable coordination geometry make it a common component of many redox systems. Incorporation of iron into iron-sulfur cluster redox centres is integral to the catalytic and sensing functionalities of many proteins. Iron ions are also used in enzyme active sites to effect some of the most energetically challenging chemistry known to biology. For example, the C-H bond of methane has a dissociation energy of 104 kcal mol-1 (184) and through activation of O2 by the diferrous active site of methane monooxygenase this 44  molecule is hydroxylated to form methanol (184,185). Both Fe2+ and Fe3+ ions occur in the active sites of various enzymes, the catalytic cycles of which may or may not involve a change in the oxidation state of the iron (182). This thesis will focus on investigations of monooxygenases that employ mononuclear iron in their catalytic prosthetic groups. Activation of dioxygen by Fe2+ sites in enzymes involves redox chemistry in which one or more electrons from the metal are donated to a coordinated oxygen molecule (182,183). Unlike unactivated organic substrates, the transition metal can react directly with O2, which results in alteration of the electronic structure of O2 to enable a kinetically favourable reaction with a cosubstrate (181). Enzymes using a variety of mononuclear iron coordination environments have been observed. Some of the best-studied are those incorporating a porphyrin ring, such as the P450s. These enzymes contain an iron protoporphyrin IX (heme-b) prosthetic group in which the iron is coordinated by the central porphyrin nitrogen atoms in conjunction with an axial cysteine ligand situated on one face of the porphyrin ring (181). This prosthetic group is also found in non-P450 oxygenases, including secondary amine monooxygenase, prostagladin H synthase, indoleamine 2,3-dioxygenase, and tryptophan 2,3-dioxygenase, all of which possess an axial histidine ligand to the heme iron rather than cysteine (181). Another common mononuclear iron-binding motif is the 2-His-1-carboxylate facial triad (180,186). This amino acid grouping, which incorporates two histidyl residues and one aspartyl or glutamyl residue, has been observed in several protein scaffolds of disparate evolutionary origin including the extradiol dioxygenases, Rieske oxygenases (ROs), !ketoglutarate-utilizing oxygenases, tetrahydropterin-containing oxygenases, and several oxidase families (180). This arrangement allows for flexibility in catalytic capacity as it provides a variety of coordination geometries through spatial arrangement of the amino acid residues and monodentate or bidentate coordination by the carboxylate residue. The threeamino acid motif is also beneficial from a catalytic point of view as concentration of the three ligands on one face of the metal ion allows for up to three open coordination sites for substrate coordination on the distal face of the iron (180). This arrangement enables direct interaction of both O2 and its cosubstrate with the Fe2+ catalytic centre. The range of 45  mechanisms proposed for oxygenases with 2-His-1-carboxylate facial triad Fe2+ centres includes four iron oxidation states and all forms of reduced O2 (180). This variety reflects the catalytic flexibility of these sites. Other non-heme mononuclear Fe2+-binding motifs include the 3-His-1-Glu coordination in the cupin protein superfamily, a 3-His motif found in some oxygenases with a cupin fold, and motifs consisting of two or four histidine residues in proteins of distinct tertiary structure (187).  Iron-containing monooxygenases in the cholesterol degradation  pathway  There are three putative iron-containing monooxygenases encoded in the cholesterol degradation gene cluster of Mtb. The first two are the P450s Cyp142 (Rv3518c), which exhibits a destabilizing mutation in the clinical isolate strain (188) CDC1551 (62), and Cyp125A1 (Rv3545c). Compared to most bacteria, actinomycete genomes encode a large number of P450s; the Mtb genome includes genes for 20 of these enzymes (189). Cyp125A1 was identified in transposon mutant screens as being necessary for bacillary survival during infection in a mouse model (52) and during in vitro growth on cholesterol (48). In accordance with this data, cyp125 was found to be upregulated during intramacrophage growth of Mtb (49-51) and is one of the most highly upregulated genes in R. jostii RHA1 during growth on cholesterol (21). Cyp125A1 is proposed to catalyze the hydroxylation and subsequent oxidation of an aliphatic carbon chain as the first step in a cycle of !-oxidation (56). In Mtb, the cyp125 gene is encoded in the igr operon, alongside genes with a high similarity to genes in the testosterone degradation cluster of C. testosteroni (21). Although testosterone does not exhibit the aliphatic side chain at C17 characteristic of cholesterol, several igr genes have been proposed to orchestrate a round of !-oxidation resulting in the removal of the last three side chain carbons from the steroid core structure (64). Regardless of its exact role in cholesterol metabolism, Cyp125A1 has been demonstrated to be resistant to inhibition by physiologically relevant levels of NO (190).  46  The third putative iron-containing monooxygenase is the two-component Rieske monooxygenase, KshAB (Rv3526 and Rv3571). KshA is proposed to be the terminal oxygenase of this two-component enzyme system while KshB is proposed to provide electrons to enable oxygenation by KshA, presumably without forming a stable proteinprotein complex (85). In accordance with these predictions, sequence analyses indicate that KshB binds a flavin prosthetic group with its N-terminal region and a plant-type [2Fe2S] cluster at its C-terminus. This architecture classifies KshB as a type II (FNRC-type) reductase by a recently proposed RO classification system (191). KshA is predicted to exhibit the canonical RO terminal oxygenase structure with an N-terminal Rieske domain and Cterminal catalytic domain. Like cyp125, both kshA and kshB were found to be upregulated in cholesterol-grown R. jostii RHA1 (21) and in Mtb isolated from macrophages (49,51). Mtb !kshA and !kshB mutants also both exhibit attenuated growth on cholesterol (48). Gene disruption studies in Rhodococcus erythropolis SQ1 (85) and M. smegmatis (33) have established that KshAB of these species catalyzes the 9"-hydroxylation of 4-AD and 1,4androstadiene-3,17-dione (ADD) to 9"-hydroxy-4-androstene-3,17-dione (9-OHAD) and 3hydroxy-9,10-secandrost-1,3,5(10)-triene-9,17-dione (3-HSA), respectively (Figure 1.7). The substrate specificity of KshAB and its order of action with respect to the ketosteroid dehydrogenase have yet to be determined. Likewise, the role of KshAB in Mtb cholesterol degradation has not been substantiated.  47  Figure 1.8  Activities of KshAB and KstD. Compounds represent: (R = ketone) (1) 4-androstene-3,17-dione  (4-AD), (2) 1,4-androstadiene-3,17-dione (ADD), (3) 9!-hydroxy-4-androstene-3,17-dione (9-OHAD), and (4) 3-hydroxy-9,10-secandrost-1,3,5(10)-triene-9,17-dione (3-HSA); (R = isopropionate) (1) 3-oxo-23,24bisnorchol-4-en-22-oic acid (4-BNC), (2) 3-oxo-23,24-bisnorchola-1,4-dien-22-oic acid (1,4- BNC), (3) 9hydroxy-3-oxo-23,24-bisnorchol-4-en-22-oic acid (9-OH-4-BNC), and (4) 3-hydroxy-9-oxo-9,10-seco-23,24bisnorchola-1,3,5(10)-trien-22-oic acid (3-HSBNC); (R = isopropionyl-CoA) (1) 3-oxo-23,24-bisnorchol-4-en22-oyl-Coenzyme A thioester (4-BNC-CoA), (2) 3-oxo-23,24-bisnorchola-1,4-dien-22-oyl-Coenzyme A thioester (1,4-BNC-CoA), (3) 9-hydroxy-3-oxo-23,24-bisnorchol-4-en-22-oyl-Coenzyme A thioester (9-OH-4BNC-CoA) and (4) 3-hydroxy-9-oxo-9,10-seco-23,24-bisnorchola-1,3,5(10)-trien-22-oyl-CoA thioester (3HSBNC-CoA).  Although the coordination environments of the iron ions at the active sites of P450s and ROs are different, these enzyme classes exhibit significant similarities. For example, in both families a wide variety of substrates can be accommodated through variation of the features of the active site pocket. The nature of substrate binding within the pocket also dictates the reaction specificity for both classes of enzyme. The possibility of uncoupled reactions resulting in production of reduced oxygen species without substrate turnover is also a factor for both types of oxygenase and often occurs in the presence of poor substrates (192194). Such uncoupling can result in the release of reactive oxygen species like hydrogen peroxide or of two molecules of water. From a structural point of view, both ROs and P450s  48  exhibit low amino acid sequence identity but conserved core tertiary structures. Both classes are thus independent examples of enzymatic diversity evolved from a central catalytic scaffold. Further similarities with respect to range of reactions, enzyme components, and electron transport chains (ETCs) will be discussed below in the context of each family. 1.4.2  Cytochromes P450 P450s contain a non-covalently bound heme group that is required for activity. The  name of this enzyme family is atypical in that it derives from the characteristic spectroscopic absorbance at 450 nm of the reduced CO-bound heme group that is referred to as the “Soret band” (195). Although most abundant in eukaryotes, in which many of these systems are membrane-bound, soluble P450s are common in prokaryotic systems and play a significant role in many cellular processes. The Fe-protoporphyrin IX prosthetic group, which bears ethylene groups at C3 and C8 and a methyl group at C18, is usually bound to the protein via non-covalent interactions although some P450s have been found to have covalently-bound heme (183). In all P450s, enzymatic activity is dependent upon heme iron ligation by a conserved cysteinyl residue. This thiolate-ferrous iron ligation gives rise to the characteristic absorbance of the CO-bound complex at 450 nm (181,194). P450s are external monooxygenases, meaning that they derive the electrons needed to perform oxygen activation from a source distinct from the oxygenated substrate. P450s derive these electrons from NADH or nicotinamide adenine dinucleotide phosphate (NADPH) via an ETC. The P450 ETC comprises flavin-containing reductase, iron-sulfur cluster-containing ferredoxin, and heme-containing active site components. P450 systems exhibit considerable variation in the nature of prosthetic groups in the former two components. Both flavin adenine dinucleotide (FAD) and flavin mononucleotide (FMN) are used in the reductase components which can derive electrons from either NADH or NADPH. Similarly, the nature and number of iron sulfur clusters in the ETC varies from system to system (196). The organization of these components is also variable such that in threecomponent systems the different components are located on three separate polypeptides and in two-component systems a single reductase-ferredoxin protein complements the heme-  49  bearing oxygenase. Several systems have also been discovered in which all three components are located on a single polypeptide that exhibits considerably greater reaction rates than multi-protein systems (183). A ten-category classification scheme for P450s has been developed based on the nature of the ETC (196).  Reactions and applications  P450s catalyze a wide variety of oxidation reactions including mono-hydroxylation, epoxidation, heteroatom-dealkylation and oxidation, oxidative deamination, dehalogenation, dehydrogenation, dehydration, isomerization, C-C bond cleavage, and reduction (181,183). Arguably the most common and undoubtedly the most studied of these reactions are monohydroxylation reactions. These reactions are carried out with a great degree of regioand stereo-specificity, often in spite of significant energetic disadvantages. For example, hydroxylation of a tertiary aliphatic carbon is energetically favourable to that of a primary carbon; however, P450s are able to catalyze the hydroxylation of terminal methyl groups with over 99% specificity (18). This characteristic is due in great part to the specific binding orientation of the substrate within the enzyme active site. Aside from performing many essential chemical transformations required for both prokaryotic and eukaryotic life, P450s have been explored for industrial applications. To this end, several P450 systems have been exogenously expressed in both bacterial (Escherichia coli) and yeast (Saccharomyces cerevisiae) model organisms (183). Similarly, members of this family have been evolved in a laboratory setting (197) for improved utility in industrial processes (198,199). Although recombinant P450s are not currently being used on an industrial scale (183), examples of P450 systems designed for use in industrial protocols include a S. cerevisiae strain expressing four mammalian P450s for production of hydrocortisone (200), expression of human P450s in E. coli for synthesis of drug metabolites (201), expression of a human P450 in fission yeast Schizosaccharomyces pombe for production of drug metabolites (183,202), and expression of P450(BM3) from Bacillus megaterium in E. coli for production of artemisinin (203).  50  Mechanism  The reaction cycle of P450s involves the input of exogenous electrons through the heme iron to activate O2 for attack on the organic substrate (181) (Figure 1.8). In the resting state of the enzyme (1), the heme iron is typically in a low spin ferric state with a square bipyramidal coordination geometry in which the heme contributes the equatorial ligands and the axial ligands are a conserved cysteine residue and a water molecule (194). Substrate binding in the active site pocket displaces the water, thus inducing a high-spin ferric iron with an open coordination site (2). This shift in spin state is concomitant with a shift in redox potential that enables transfer of an electron from the ETC to generate a five-coordinate highspin ferrous heme iron (3). O2 binds to this centre and is reduced through transfer of one electron from the iron ion to generate a dioxy-ferric heme active site species (4). The next step is rate-limiting and involves transfer of a second electron from the ETC to generate a ferric-peroxide species (5a). Two sequential protonations of the distal oxygen atom then lead to heterolytic O-O bond cleavage and formation of the ferryl-oxo active oxidant, compound I (6) (194,204). Addition of the oxygen atom to an organic substrate (7) and subsequent product release results in regeneration of the resting enzyme.  Figure 1.9  Cytochrome P450 reaction cycle. Reprinted with permission from Denisov et al. 2005 (194).  Copyright 2005 American Chemical Society.  51  Classification  As mentioned above, P450s have been classified with respect to their ETC components (196), but a more informative classification scheme reflecting the taxonomic inter-relationships in this enzyme class has also been developed (205). This latter system classifies P450s based on amino acid sequence identity of the heme-bearing oxygenase component and is more closely indicative of both evolutionary origin and enzyme function than the system based on the somewhat promiscuous ETC. In the sequence-based classification scheme that was developed by Nebert et al. (206) P450s are separated into families and subfamilies. Members of different families typically do not share greater than 40% amino acid sequence identity while those in different subfamilies are less than 55% identical (205). This system effectively classifies P450s into hundreds of functionally relevant classes rather than a handful of broader categories. It is advantageous for this class of enzyme given the great functional diversity and low sequence identity characteristic of P450s. While the oxygenase components of P450s exhibit a highly conserved heme-binding core and tertiary structure, the overall sequence identity of these proteins can be quite low (< 20 %) (183,196) and this makes classification into a small number of categories both impractical and uninformative. 1.4.3  Rieske oxygenases ROs catalyze the oxidation of relatively stable organic substrates such as  unsubstituted aromatic compounds. This catalytic ability has led to their recruitment for many bacterial catabolic pathways and for biosynthesis of critical compounds in eukaryotes. A subset of the so-called Rieske-type proteins (207), ROs have been given many names in the literature including bacterial ring-hydroxylating oxygenases (RHOs), and aromatic ring hydroxylases (ARHs). These names are not truly representative as ROs do not occur exclusively in bacteria and many perform reactions other than ring hydroxylation (208-213). In 2000, Gibson and Parales suggested the family name be changed to “Rieske non-heme iron oxygenase” (214) to reflect the presence of the mononuclear iron and Rieske cluster in the terminal oxygenase. The Rieske cluster was first named in the literature in 1975 (215) for 52  Dr. John S. Rieske who characterized this cluster while studying the cytochrome complexes of the mitochondrial electron transport chain. It has since been established that Rieske clusters are coordinated in protein domains of conserved basic fold that can exist independently as small ferredoxins or be incorporated into larger proteins or complexes. There are also examples of hypothetical proteins incorporating Rieske domains coupled with putative oxygenase domains different from the canonical ROs. An unambiguously descriptive name for ROs would therefore be unwieldy: Rieske Bet v1-like mononuclear HisHis-carboxylate-ligated iron oxygenases. Accordingly, the term Rieske oxygenase (RO) will be used for these multi-component enzymes, with RO-O referring to the terminal oxygenase component. Like the P450s, ROs function as two- or three-component systems (Figure 1.9) with an ETC acting to transfer NADPH-derived electrons to the terminal oxygenase (RO-O) active site. The RO-O is the catalytic component of the RO system (216). The active site lies in the RO-O’s C-terminal catalytic domain and contains the mononuclear catalytic Fe2+. The metal ion is coordinated by a 2-His-1-carboxylate facial triad motif, allowing for coordination of up to three additional ligands on the iron’s distal face (180). The N-terminal Rieske domain coordinates a Rieske iron-sulfur cluster that ensures the efficient and precise delivery of electrons to the catalytic centre. Like all Rieske clusters, the two irons in the [2Fe2S] cluster are coordinated by two cysteinyl and two histidyl residues, respectively.  Figure 1.10  Schematic of Rieske monooxygenase systems. Components of the electron transport chain and  prosthetic groups are labeled. Variations for each component are shown vertically. Flow of electrons, substrates, and products are indicated with arrows.  53  The ETC of ROs comprises one or two proteins in addition to the Rieske cluster of the RO-O (216). Electrons are first transferred from NAD(P)H to the non-covalently bound flavin cofactor of the reductase component. This protein sometimes also contains an additional iron sulfur cluster that transfers electrons from the reductase directly to the RO-O Rieske cluster in two-component systems (216). In three-component systems, a small ferredoxin shuttles electrons from the reductase to the oxygenase. The number of electron transfer components and the nature of their prosthetic groups vary significantly and have been used as a basis for classification (191,217). Furthermore, like the corresponding components of P450s, RO ETC components can be functionally and physiologically promiscuous (191).  Reactions and applications  Although the ROs that are best represented in the literature catalyze mono- and dioxygenations of aromatic compounds (214), other types of RO-catalyzed oxidations have been demonstrated in aromatic ring-hydroxylating ROs and ROs with other physiological roles (Figure 1.11). The former is exemplified by naphthalene dioxygenase from Pseudomonas sp. strain NCIB 9816-4 (NDO9816-4). In addition to naphthalene oxygenation, this enzyme is capable of transforming dozens of substrates by varied oxidative chemistry including sulfoxidation, dealkylation, and desaturation (213). There also several examples of ROs with non-hydroxylating physiologically relevant activities such as N-demethylation (218), O-demethylation (219), and desaturation (208). Among these ROs are RedGA3(2) and McpGA3(2) of Streptomyces coelicolor A3(2) which catalyze oxidative cyclization (209). The latter enzymes are involved in antibiotic biosynthesis, and as such represent a significant departure from the paradigm of ROs in bacterial catabolism.  54  Figure 1.11  Reactions catalyzed by Rieske oxygenases. Proposed reactions of the Rieske oxygenases A)  naphthalene dioxygenase from Psuedomonas putida NCIB 9816-4 (213); B) 2-oxoquinoline 8-monooxygenase from Psuedomonas putida 86 (220); C) vanillate demethylase from Pseudomonas sp. strain HR199 (219); D) Ndemethylase from Pseudomonas putida CBB5 (218); E) choline monooxygenase from Arabidopsis thaliana (Roo26B1) (221); F) Neverland from Bombix mori (208); and G) the RedG protein of Streptomyces coelicolor A3(2) (209).  55  Due to their ability to catalyze a range of stereo-specific chemistries on a broad range of substrates, bacterial ring-hydroxylating ROs have been studied with the aim of exploiting these activities for applications in bioremediation and industrial biocatalysis. For example, biphenyl dioxygenase (BPDO) of the biphenyl degradation pathway has been studied for bioremediation of polychlorinated biphenyl (222,223). These efforts have included directed evolution to augment the enzyme’s PCB-transforming capabilities (224) and the heterologous production of BPDO in plants for phytoremediation (225). More recently, protein engineering efforts have shown promise for enzyme design strategies for increasing the substrate range of BPDO (226). ROs have also been studied for applications in the synthesis of commercially valuable products such as indigo, a commercial dye, and indinavir, an HIV protease inhibitor (227,228). The RO KshAB has similarly been investigated for production of valuable steroid intermediates (33,45).  Structure  The first crystal structure of an RO-O, published in 1998, was that of the !3"3 NDO9816-4 (229) (PDB: 1NDO). The structural model consists of a mushroom-shaped assembly with three !" protomers arranged around a central three-fold rotational axis such that the three ! subunits form a ring and the three " subunits are situated on one face of that ring (Figure 1.12). More recent crystal structures have been published with both !3 and !3"3 configurations and an RO-O with an !6 configuration of two stacked !3 units has also been proposed (230). The " subunit in some ROs is generally thought to play a structural role in trimer stabilization, but this chain has also been demonstrated to contribute to substrate specificity (231) and reaction rate (232) in some enzymes.  56  Figure 1.12 Rieske terminal oxygenase structure. The structure of Pseudomonas sp. NCIB 9816-4 (PDB ID 1NDO) (229) is shown as a ribbon diagram. (A) Structure of the !-subunit monomer. The Rieske and catalytic domains are shown in green and blue, respectively. Iron ions and acid-labile sulfur are represented as red and orange spheres. (B) and (C) are views of the !3"3 structure such that the plane of the page is parallel and perpendicular to the threefold axis of symmetry, respectively. Within each !" protomer, the ! subunit is a dark shade and " subunit is a light shade.  The basic features of the ! subunit are conserved among all RO-Os and consist of an N-terminal (~150 aa) Rieske domain and a larger (~250 aa) C-terminal catalytic domain (Figure 1.11) (216). Within one monomer, the two metal centres are separated by ~43 Å, a distance prohibitive to efficient electron transfer. The Rieske cluster of the adjacent subunit in the trimer, however, is within ~12 ! of the catalytic iron and this distance is suitable for electron transfer (216). The RO-O trimeric structure is therefore functionally essential for electron transfer from the Rieske cluster to the active site iron. The Rieske cluster iron ions are coordinated by the conserved Cys-X-His-X16-18-Cys-X2-His motif in the Rieske domain (216). The three mononuclear iron ligands are a pair of histidines in a conserved Asp-X2-HisX4-His motif situated on a bent-helix near the N-terminus of the catalytic domain (NDO9816-4 57  His208 and His213), and an aspartate over 100 residues downstream and situated on an adjacent helix (NDO9816-4 Asp362). Comparative sequence analysis failed to identify this third ligand due to its considerable distance downstream from the histidine ligands and the high sequence diversity of the RO-O C-terminus. The NDO9816-4 structural model also suggested an important role for the conserved aspartate in the above-mentioned Asp-X2-His-X4-His motif; it forms hydrogen bonds with N!1 of the first histidine in this motif and N"2 of the second histidine in the Rieske binding motif of the adjacent subunit to link the two metallocentres through a hydrogen bond network. This residue has been proposed to play an essential role in catalysis (233,234) and may impart various catalytic properties depending on its orientation. The structures of 17 different RO-Os have now been published (Table 1.2), all of which share the RO-O canonical fold but differ from one another by inclusion or omission of significant secondary structural features. The basic fold of the Rieske domain (216) consists of three antiparallel #-sheets (207,229). The four Rieske-cluster ligands are situated in two loops: the first occurring between sheets two and three, and the second between two strands in sheet three. These loops form a pincher-like structure between which the cluster is ligated (207,229). The various C-terminal catalytic domains also exhibit conserved features, but also include many more significant insertions than the Rieske domains. These adaptations illustrate the evolutionary flexibility of the RO-O structural scaffold to accommodate diverse substrates through modification of the basic enzyme structure. This strategy is again analogous to what is observed in the P450s. The RO-O conserved fold is defined by the SCOP database (235) as belonging to the Bet v1-like superfamily of proteins with a TATAbinding protein-like fold and comprises a central anti-parallel #-sheet flanked by !-helices on the face distal to the Rieske domain (216,229).  58  Table 1.2  Published Rieske terminal oxygenase structures  Structural modela  PDB IDb  Enzyme Name  Organism  4o structure  TDOF1  3EN1 (236)  toluene 2,3-dioxygenase  Pseudomonas putida F1  !3 "3  BPDOLB400  2XR8  Biphenyl dioxygenase  Burkholderia xenovorans LB400  !3 "3  BPDOB-356  3GZY  biphenyl dioxygenase  Comomonas testosteroni B-356  !3 "3  BPDORHA1  1ULI (237)  biphenyl dioxygenase  Rhodococcus jostii RHA1  !3 "3  CUMDOIP01  1WQL (238)  cumene dioxygenase  Pseudomonas fluorescens IP01  !3 "3  NDO9816-4  1O7W (239)  naphthalene dioxygenase  Pseudomonas sp. NCIB 9816-4  !3 "3  NBDOJS765  2BMO (240)  nitrobenzene dioxygenase  Comomonas sp. JS765  !3 "3  RHDOCHY-1  2CKF (241)  ring-hydroxylating dioxygenase  Sphingomonas sp. CHY-1  !3 "3  BPDOB1  2GBW (242)  biphenyl dioxygenase  Sphingomonas yanoikuyae B1  !3 "3  OMO86  1Z02 (220)  2-oxoquinoline-8monooxygenase  Pseudomonas putida 86  !3  CARDOJ3  1WW9 (243)  carbazole 1,9!-dioxygenase  Janthinobacterium sp. J3  !3  CARDOIC177  3GCF (244)  carbazole 1,9!-dioxygenase  Nocardioides aromaticivorans IC177  !3  CARDOKA1  3GKQ  carbazole 1,9!-dioxygenase  Novosphingobium sp. KA1  !3  NDO12038  2B1X (245)  naphthalene dioxygenase  Rhodococcus sp. NCIMB 12038  !3 "3  KshAH37Rv  2ZYL (246)  3-ketosteroid-9!hydroxyxlase  Mycobacterium tuberculosis H37Rv  !3  DMO  3GKE (210)  dicamba monooxygenase  Stenotrophomonas maltophilia  !3  putativeTM1040  3N0Q  (putative)  Silicibacter sp. TM1040  !3  a  Names of RO-O structural models will be used throughout the thesis. Only one representative structure is listed for each enzyme. When more than one structure was available, the highest resolution structure without bound substrate or inhibitor and with both iron centres in a reduced state was selected. b  The major limitation of the current RO structural data is the narrow taxonomic and functional range of the enzymes studied to date. The seventeen enzymes for which x-ray crystallographic data are available represent only 6 orders of bacteria: three from the genus 59  Pseudomonas, four from the family Sphingomonadaceae, four from the order Actinomycetales, and three from the order Burkholderiales. Only the putative RO-O from Silicibacter sp. TM1040 and DMO represent enzymes from independent orders. The functional redundancy is even more problematic; three structural models are of RO-Os that dihydroxylate carbazole and four are of biphenyl-transforming ROs. This overall lack of diversity has lead to erroneous inferences about ROs in general, and has contributed to nonrepresentative RO classification schemes.  Mechansim  The catalytic mechanism of ROs, like that of the P450s, involves initial binding of substrate followed by binding and activation of oxygen, substrate transformation, and finally product release to regenerate the resting enzyme (Figure 1.13) (180,216,247,248). The mononuclear iron of the resting enzyme is in the ferrous state (216,229) with either one or two solvent ligands in addition to the three protein-derived ligands (216). Substrate binds in a pocket adjacent to the iron such that the atom to be modified is orientated towards the iron (220,239,249). The position of the iron relative to the substrate seems to be mediated by the oxidation state of the Rieske cluster via a conserved carboxyl residue, usually aspartate, that bridges the inter-subunit boundary (233,234). Thus, when the cluster is diferric, there is insufficient space between the substrate and the iron for O2 to bind to the latter (220,250). Upon reduction of the cluster, an ~1 Å shift of the mononuclear iron away from the substrate allows O2 to bind side-on to the catalytic iron adjacent to the carbon(s) to be hydroxylated (238,239). The exact nature of the activated oxygen species is not yet known; however, current studies indicate that an iron(III)-hydroperoxy intermediate is formed which may directly attack the substrate through concerted transfer of a second electron from the Rieske cluster (216), or form an iron(V)-oxo intermediate before substrate hydroxylation via a radical mechanism (248,251). Recent data obtained using a mononuclear non-heme iron(III)peroxo complex model compound suggests the former species is the active oxidant (252). In either case, reduction of the mononuclear iron to a ferrous state appears to complete the reaction cycle and aid in product release (253).  60  Figure 1.13  Rieske oxygenase catalytic cycle. The Rieske cluster is shown as a white rectangle; the  mononuclear iron is shown as a circle; the active site is shown as a shaded oval. Parallel paths represent different models of the reaction mechanism.  Eukaryotic Rieske oxygenases  Although members of the RO family have been described in taxonomically diverse eukaryotic systems, these homologues have been largely overlooked in the RO literature. Indeed, ROs are often referred to as “bacterial” enzymes. Sequence information is available for eukaryotic ROs from fungi, plants, animals, and brown algae, although most genetic and biochemical studies have focused on enzymes found in green plants, insects, and round worms. The eukaryotic ROs play a different metabolic role than their prokaryotic counterparts, carrying out biosynthetic or detoxification reactions rather than catabolic processes (221). The best studied animal RO is the product of the neverland gene (208). This has been studied in Drosophila melanogaster (fruit fly) (208), Apis mellifera (honey bee) (254), Bombix mori (silk moth) (208), Danio rerio (zebra fish) (208), Xenopus laevis (frog) (208), and Caenorhabditis elegans, where it has also been called Daf36 (255). While no mammalian 61  homologue of Neverland has been identified (208), this protein is conserved in species as disparate as the single-celled Casaspora owczarzaki, Nematostella vectensis (sea anemonae), and Gallus gallus (chicken) (256). Moreover, it is the only identified RO detectable in the genomes in which it is found (208). Neverlandmelanogaster shares 31% amino acid sequence identity with Daf36 and 23% identity with Mtb KshA (256). Consistent with this identity, Neverland is proposed to transform a steroid substrate(s) in steroid hormone production (255), possibly acting as a cholesterol dehydrogenase (208). Among plants, ROs are exemplified by the five that occur in Arabidopsis thaliana and that are conserved in several other plant species (221): PaO, CaO, TICC55, PTC52 and CMO. The first four of these enzymes have been implicated in chlorophyll b metabolism (221). CAO has been characterized in vitro and catalyzes two sequential monohydroxylations of the same methyl carbon in the chlorophyll b biosynthetic pathway of chloroplasts (212). Contrary to what is observed in many RO substrates, the hydroxylated carbon is not part of a ring structure. The fifth, CMO, appears to be more distantly related than the other four and has been proposed to be the result of a horizontal gene transfer event (221). CMO catalyzes the monooxygenation of choline to form betaine aldehyde. The orientation of the exons of each gene provides evidence that all five enzymes existed before the divergence of monocots and dicots in flowering plants and are not the product of a recent gene duplication event (221). The relationship between the A. thaliana ROs and a limited number of bacterial ROs has been recently investigated by Gray et al. (221).  Classification  The first RO classification scheme was proposed by Batie et al in 1991. It did not incorporate sequence information, as this was very limited at the time. Instead, the Batie classification system used the components of the electron transport chain to group ROs that had been studied either genetically or biochemically (217). Although this system describes important RO components and is easily expandable to accommodate the discovery of new  62  ETCs, it carries no information about function or evolutionary relationships. In fact, the promiscuity of the ETC components can result in their “swapping” during RO evolution. More recent RO classification schemes have been based on phylogenetic analyses. In 1996 Werlen et al. grouped fifteen RO-Os into four phylogenetically-determined families designated by their respective substrate classes (257). A study published by Nam et al. in 2001 (258) compared RO-Os based on their amino acid sequences. Using this analysis, the authors proposed four RO-O group designations based on the degree of intra-group sequence identity. A more recent study published by Kweon et al. (191) used phylogenetic analyses of each of the ETC components to define “classification keys”. These keys were integrated into a phylogenetic analysis of the RO-Os to construct a scheme that incorporates determinants used in both the Nam and Batie classification schemes. This key-based classification scheme identified five RO types further divided into subtypes based on the presence or absence of a beta subunit (191). This analysis was subsequently used as the basis for a computer program that was designed to automatically classify new ROs (259). Although comparative sequence analysis is a powerful tool for constructing classification systems, the sequence diversity of RO-Os presents significant challenges for robust phylogenetic analysis and interpretation. These challenges include lack of diversity in the sequences examined, validity of the multiple sequence alignment, classification based on nested phylogenetic clades, and limited feasibility of functional inference. The high variability of RO-O sequences limits the power of the basic local alignment search tool (BLAST) searches (260) commonly used to retrieve sequences for analysis. Use of this tool thus limits the diversity of the sequence pool obtained using this approach and results in an experiment that explores the sequence space most closely surrounding biochemically characterized enzymes rather than representing the total available RO-O spectrum. It has also been challenging to construct robust multiple sequence alignments for RO-Os because of their high sequence diversity in the C-terminus of the catalytic domain. Indeed, most published alignments are highly ambiguous or erroneous in this section. Subsequent phylogenetic reconstruction is impacted by such misalignment, which inflates the apparent phylogenetic distance between similar sequences that align properly and less similar  63  sequences that may be erroneously aligned (261). The phylogenetically-based RO classification schemes to date have also relied to varying extents on classification through nested genetic clades and have thus effectively classified proteins based on similarity to the most closely interrelated group. This approach restricts the possibility of functional inference to members of the least diverse clans. Finally, as demonstrated by the P450 enzyme family, the predictive power of a specific classification scheme is limited by the catalytic and sequence diversity of the enzyme family. The wide range of physiologically relevant catalytic functions exhibited by ROs precludes grouping these enzymes into a small number of functionally relevant categories. At the outset of work toward this thesis, no inclusive scheme addressing these issues for RO classification had yet been published. 1.5  Aim of this study The aim of this study is to biochemically characterize Cyp125 and KshAB in the  context of cholesterol degradation in actinomycetes generally and in Mtb specifically. In the case of KshAB, an additional objective was to study KshA in the context of the RO enzyme family. To this end, the proteins under investigation were overexpressed in E. coli and R. jostii RHA1 and purified in their active forms. Substrate binding and steady-state kinetic studies using commercially available substrates or those obtained through collaboration with Dr. Israël Casabon were performed to investigate the role of these proteins in cholesterol degradation. These studies were also used to investigate the cholesterol degradation pathway via substrate specificities of these enzymes. Structural studies of KshA were performed through collaboration with Prof. Natalie Strynadka to further interpret kinetic data and compare KshA to other ROs. Further characterization of KshA within this enzyme family was accomplished through a global structure-guided phylogenetic analysis of RO sequences and a new classification system for ROs was proposed. Collaborations with Prof. William Mohn, Dr. Robert van der Geize, and Prof. William Jacobs facilitated the investigation of the role of these genes in bacterial growth in vitro on cholesterol and in animal models of infection (presented in discussion). The work included in this thesis has contributed to the understanding of bacterial cholesterol degradation. Studies described herein have also contributed significantly to the understanding of the structure and evolution of ROs.  64  2 2.1  Materials and Methods Chemicals and reagents ADD and 3-oxo-23,24-bisnorchol-4-en-22-oic acid (4-BNC) were purchased from  Steraloids, Inc. (Newport, RI). 4-AD, cholesterol, 4-cholesten-3-one, and 2-hydroxypropyl!-cyclodextrin (BCD) were purchased from Sigma-Aldrich (St. Louis, MO). Jeffamine M600 was purchased from Hampton Research (La Jolla, CA, USA). Restriction enzymes and Expand High Fidelity polymerase chain reaction (PCR) System were purchased from New England Biolabs (Ipswich, MA) and Roche (Laval, Quebec), respectively. All other reagents were of high performance liquid chromatography (HPLC) or analytical grade. Water for buffers was purified with a Barnstead Nanopure DiamondTM system (Dubuque, IA) to a resistivity of at least 18 M". 26-Cholestenoic acid was a gift from Dr. Robert van der Geize (Groningen Biomolecular Sciences and Biotechnology Institute (GBB), University of Groningen). An ethylenediamine tetraacetic acid (EDTA)-bridged !-cylodextrin dimer was a gift from Dr. José Vázquez Tato (Universidad de Santiago de Compostela). 3-Oxo-23,24-bisnorchol-4en-22-oyl-coenzyme A thioester (4-BNC-CoA), 3-oxo-23,24-bisnorchol-1,4-dien-22-oic acid (1,4-BNC) and 3-oxo-23,24-bisnorchol-1,4-dien-22-oyl-coenzyme A thioester (1,4BNC CoA) were prepared by Dr. Israël Casabon (University of British Columbia). Solutions of 4-AD, ADD and 1,4-BNC up to 25 mM for use in kinetic assays were prepared fresh daily in ethanol. Solutions of 4-BNC up to 50 mM were prepared fresh in 94 % ethanol containing 60 mM NaOH. Aqueous solutions of up to 6 mM 4-BNC-CoA or 1,4BNC-CoA were used for storage and activity assays.  65  2.2  Bacterial strains and growth Escherichia coli BL21(DE3) and E. coli DH5! were routinely used for protein  production and DNA propagation, respectively. Unless otherwise noted, these strains were grown in lysogeny broth (LB) at 37 °C, 200 r.p.m. R. jostii RHA1 was also used for protein production and was grown in LB at 30 °C, 200 r.p.m. Single colonies of bacteria were obtained by incubating cells on LB agar plates at the abovementioned temperatures. Electro-competent E. coli was prepared by growing cells under standard conditions to an optical density at 600 nm (OD600) of 0.5 followed by cooling on ice for 10 min and harvesting by centrifugation. Cells were then washed three times at 4 °C with 20 mM potassium phosphate buffer containing 10% glycerol. Following the final wash, excess liquid was removed by inverting the centrifuge vessel and cells were stored at -80 °C in 30 µL aliquots. Electro-competent R. jostii RHA1 cells were prepared using the same method but using sterile water to wash the cells and aliquots of 100 µL for storage. Cells were transformed via electropiration using a MicroPulser aparatus from BioRad (Hercules, CA) with Bio-Rad 0.1 cm GenePulser Cuvettes. For E. coli cells, 1-2 µL concentrated plasmid (20-175 ng mL-1) was mixed in a cold electropiration cuvette with 20 µL cell suspension. Cells were then subjected to an electric pulse at 2.0 kV and the contents of the cuvette were transferred into 0.5 mL LB. This suspension was incubated for 30 min at 37 °C without shaking and subsequently transferred to LB agar plates. R. jostii RHA1 cells were transformed using a similar method but with 80 µL electro-competent cell suspension, and incubation in 200 µL LB at 30 °C for four hours prior to plating. 2.3  DNA manipulation and plasmid construction DNA was propagated, digested, and ligated using standard protocols (262), and DNA  plasmids were purified as described previously (263).  66  2.3.1  Oligonucleotides Oligonucleotides were purchased from Integrated DNA Technologies (San Diego,  CA) through the Nucleic Acid Protein Service Unit (NAPS) at the University of British Columbia. A list of oligonucleotides used to construct plasmids used in this thesis is provided in Table 2.1. Table 2.1  Oligonucleotides  Gene  Expression plasmid  Ro04679  pTip-QC1cyp125RHA1  fwd rev  Rv3545c  pTipCP125  fwd rev  Rv3527  pETKC1  fwd rev  Rv3503c  pETFD1  fwd rev  Rv3526  pETKA1  fwd rev  Rv3571c  pETKB1  fwd rev  Rv3571c  pETKB3  fwd rev  Rv3537  pETKD1  fwd rev  iscRSUAhscBAf dxorf3 cluster  pPAISC-1  fwd rev  Oligonucleotide sequence 5'-CATATGGCGCAGC CCAATCTTCCAGAGGG-3' 5'-GGATCCCAGTGTCTG ACCGGGCAACCG-3' 5’-GCAATAGCATATGCC CAGCCCCAATCTGCCG-3’ 5’-GCCAAGCTTCTGTTCC GCAGTGGGATCGAAATC-3’ 5'-GTGTATCCATATGCCTG ACGATCAGCCGG-3' 5'-GTCAAGCTTTTATCGGTT TGGGTCGTGGTGC-3' 5'-GTGTATCCATATGCGGG TGATCGTGGACCGAG-3' 5'-GCCAAGCTTCTACTCA CCACGCGACAACGCC-3' 5’-GCAATAGCATATGAGTA CCGACACGAGTGGGGTCG-3’ 5’-TCTAAGCTTTTGCTCGGC GGGCACGTCGT-3’ 5'-GCAATAGCATATGACCG AGGCAATTGGAGACGAGC-3' 5'-TCCAAGCTTCTCGTCGT AGGTCACTTCCACCG-3' 5’-CGGAAGGCATATGACC GAGGCAATT-3’ 5’-GACAAGCTTCTACTCGT CGTAGGTCACT-3’ 5’-TACGCTAGCACTGTGCA GGAGTTCGACGTCG-3’ 5’-CAGAATTCTCAGCGCT TTCCCGCCTG-3’ 5'-GCTCTAGATGGCTAGT CCCGTTCCTC-3' 5'-CACTCTAGACCTAA TTACTCGGATAAGAGC-3'  Restriction site NdeI BamHI NdeI HindIII NdeI HindIII NdeI HindIII NdeI HindIII NdeI HindIII NdeI HindIII NheI EcoRI XbaI XbaI  67  2.3.2  Plasmid construction Construction of plasmids for protein production was performed by ligation of PCR-  amplified DNA fragments containing the genes of interest into the multiple cloning sites of restriction enzyme-digested vectors. A list of the plasmids used for this work and their properties is provided in Table 2.2. The plasmids used in this thesis were constructed by members of Dr. Robert van der Geize’s lab (pTipQC1cyp125RHA1 (264)), Jie Liu in the lab of Prof. Lindsay Eltis (pETKD1 (265), pTipCP125 (266), pETKA1 (246), pETKB1 (246), pETKC1, pETFD1), Yong Ge in the Eltis lab (pPAISC-1 (267)), and myself (pETKB3). Table 2.2  Expression plasmids  Expression plasmid pTip-  Protein product  Origin of cloned gene  Parent vector  Tag  Cyp125A14P  R. jostii RHA1  pTipQC1  C-(His x 6)  pTipCP125  Cyp125A1  Mtb H37Rv  pTipQC2  none  pETKC1  KshC  Mtb H37Rv  pET28b+  N-(His x 6) thrombin  pETFD1  FdxD  Mtb H37Rv  pET28b+  N-(His x 6) thrombin  pETKA1  KshA  Mtb H37Rv  pETKB1  KshB-his  Mtb H37Rv  pETKB3  KshB  Mtb H37Rv  pET41b  none  pETKD1  KstDMtb  Mtb H37Rv  pET28a+  N-(His x 6) thrombin  QC1cyp125RHA1  pET41b-cthrombin pET41b-cthrombin  C-(His x 8) thrombin  C-(His x 8) thrombin  The pETKB3 plasmid was constructed for production of KshB from Mtb in E. coli BL21(DE3). The kshB gene was amplified from genomic DNA of Mtb H37Rv using polymerase chain reactions containing 0.2 µg template DNA, 0.9 units of the Expand High Fidelity PCR System polymerase, 50 !M of each dNTP, and 75 pmol of each forward and reverse oligonucleotide (Table 2.1) in a volume of 25 !L. Reactions were subject to 25 temperature cycles using a Stratagene Robocycler Gradient 96 instrument (La Jolla, CA) as follows: 95 °C for 45 s, 45 °C for 45 s, and 72 °C for 90 s. The kshB amplicon and pET41b  68  vector were each digested with NdeI and HindIII, purified, and ligated together to yield pETKB3. Amplification of the plasmid insert using T7 oligonucleotides followed by sequencing by NAPS confirmed the presence of the kshB gene. 2.4  Protein production and purification Fast protein liquid chromatography (FPLC) was performed using an ÄKTA Explorer  (Amersham Pharmacia Biotech, Quebec) equipped with a 1 x 10 cm column of SourceTM15Q (GE Healthcare) resin. Nickel-nitrilotriacetic (Ni2+-NTA) acid affinity chromatography was performed using a 1.8 x 20 cm gravity-driven column with 5 mL Ni2+-NTA agarose resin (Qiagen, Hilden). KshA, KshB-his, KshC, and FdxD were assumed to be O2-labile and therefore purified anaerobically. This included purging the headspace of ultracentrifugation tubes and septa-equipped glass bottles containing lysed cells with argon or nitrogen, performing gravity-driven chromatography and FPLC inside or interfaced to a Labmaster Model 100 glove box (M. Braun, Inc. Peabody, MA) operated at <5 ppm O2, and using buffers equilibrated with the atmosphere inside the glove box following cell lysis. Stirred cell protein concentration systems were used for anaerobic purifications while centrifugal protein concentration was used for aerobic purifications. 2.4.1  Production of Cyp125A14P and Cyp125A1 Cyp125A14P and Cyp125A1 were produced in R. jostii RHA1 using the expression  plasmids pTip-QC1cyp125RHA1 and pTipCP125, respectively. R. jostii RHA1 cells were transformed with the appropriate plasmid by electroporation and grown on LB-agar plates containing 25 !g/mL chloramphenicol for two days, after which a single colony was used to inoculate 50 mL of LB which was incubated at 30 °C (200 rpm). At OD600 1.0 (~2-3 days) 1 L of medium was inoculated with 10 mL of this pre-culture and incubated at 30 °C. At an OD600 of 0.5, thiostrepton was added to a final concentration of 50 !g/mL and the culture was incubated for a further 20 hrs. The cells were then harvested by centrifugation (4600 g, 4 °C, 10 min) and washed with 0.1 M potassium phosphate buffer, pH 8.0. Cell pellets were flash frozen in liquid nitrogen and stored at -80 °C until use.  69  Cells of R. jostii RHA1 containing Cyp125 were lysed by bead beating. Pellets from 2 L Cyp125A14P and 3 L Cyp125A1 were suspended in 20 mM potassium phosphate, pH 7.4 and 25 mM HEPES, pH 7.5, respectively. Resuspension buffers were supplemented with DNase I (Roche diagnostics, IN). The suspended cells were subjected to 5 rounds of bead beating for 3 minutes, alternating with 10 min cooling on ice. The lysed cells were centrifuged for 7 min at 4,600 g, and the supernatant was further clarified by centrifugation at 10,000 g for 45 min, followed by syringe-driven filtration through a 0.45 µm membrane to give the cell-free extract. 2.4.2  Production of proteins in E. coli BL21(DE3) KshA (expression vector pETKA1), KshB (pETKB3), KshB-his (pETKB1), KshC  (pETKC1), FdxD (pETFD1), and KstD (pETKD1) were heterologously produced in E. coli BL21(DE3). Production of all of the above but KstD was improved by producing them in cells that also contained pPAISC-1 (267), a vector containing the isc genes of P. aeruginosa PA01 involved in FeS cluster assembly. Cells were grown in LB supplemented with 25 !g/mL kanamycin, 7.5 !g/mL tetracycline and an HCl-solubilized solution of minerals (268) for production of KshA, KshB, KshB-his, KshC, and FdxD. KstDMtb was produced using LB containing 25 µg/mL kanamycin. One litre of medium in a 2 L flask inoculated with 10 mL of an overnight culture was incubated at 200 r.p.m. at the temperatures indicated in Table 2.3. For FdxD production, 500 mL of medium in a 2 L flask was inoculated with 5 mL of the preculture. At OD600 0.5 for all cultures isopropyl !-D-1 thiogalactopyranoside (IPTG) was added to a final concentration of 0.5 mM and cultures were incubated for a further 18 to 20 hrs before harvesting by centrifugation. Pellets were washed twice with 20 mM sodium phosphate, pH 8.0 containing 10% glycerol and frozen at -80 °C until use.  70  Table 2.3  Incubation temperatures for protein production in E. coli BL21(DE3)  Plasmid pETKC1 pETFD1 pETKA1 pETKB1 pETKB3 pETKD1  Protein product KshC FdxD KshA KshB-his KshB KstD  Incubation temperature (°C) 30 37 25 37 37 20  E. coli pellets containing the above mentioned recombinant proteins were resuspended in their initial purification buffer (Buffer A, Table 2.4) supplemented with DNase I. All cell-free extracts except that of KstD were prepared by passing cell suspensions five times through an Emulsiflex-05 homogenizer (Avestin, Ottawa, Ont.) operated at 10,000 psi at 4 °C. In the case of KshA, ferrous ammonium sulfate (FAS) was added to a final concentration of 0.25 mM after the first pass. Cell debris was removed by ultracentrifugation at 10,000 g for 45 min at 4 °C followed by filtration of the supernatant through a 0.45 µm filter. KstD cell-free extract was prepared by subjecting cell suspensions to five rounds of bead beating using an MP Biomedicals FastPrep-24 bead beater (Solon, OH) set to 5.0 for 20 s with 4 min on ice between rounds. Cell debris was pelleted by centrifugation at 16,100 g for 10 min at 4 °C. The supernatant, referred to hereafter as KstD-lysate, was separated from the pellet and kept on ice until use. Table 2.4  Purification buffer A  Protein Cyp125A14P Cyp125A1 KshC FdxD KshA KshB-his KshB KstD  Buffer 20 mM potassium phosphate 25 mM HEPES 20 mM sodium phosphate 20 mM sodium phosphate 20 mM sodium phosphate 20 mM sodium phosphate 25 mM HEPES 50 mM Tris-HCl  pH 7.4 7.5 8.0 8.0 8.0 8.0 7.5 7.4  Additional Component none none 10% glycerol 10% glycerol 10% glycerol 10% glycerol 5% glycerol 0.175 mg/mL phenylmethane-sulfonylfluoride  71  2.4.3  Ni2+-NTA purification of His-tagged proteins Cyp125A14P, KshA, KshB-his, KshC, and FdxD were purified by Ni2+-NTA affinity  chromatography. For purification of all but Cyp125A14P, cell-free extract was loaded onto a column equilibrated with 20 mM sodium phosphate, pH 8.0 containing 10% glycerol. Protein was eluted using an imidazole step gradient according to the instructions of the manufacturer. Coloured fractions eluted in 20 mM sodium phosphate containing 150 mM imidazole and 10% glycerol, pH 8.0 and were exchanged into 25 mM HEPES, pH 7.5 containing 5% glycerol (KshA, KshB-his, FdxD), or 25 mM HEPES, pH 8.3 (KshC). Cellfree extract containing Cyp125A14P was loaded on a column equilibrated with 0.1 M potassium phosphate, pH 7.4, washed with the same buffer containing 0.5 M NaCl, eluted with this buffer further supplemented with 50 mM L-histidine, and exchanged into 0.1 M potassium phosphate, pH 7.4. Proteins were concentrated to 15-20 mg/mL using YM5 (FdxD), YM10 (KshC), or YM30 (others) membranes (Amicon, Oakville, Ontario). Cleavage of the His-tags of KshA, KshC, and FdxD was performed in an anaerobic glovebox. Protein solutions of ~ 5 mg mL-1 were brought to 50 mM NaCl (KshA) or 20 mM NaCl (KshC and FdxD) and human ! thrombin (HTI, Essex Junction, VT) was added to a thrombin:protein molar ratio of 1:410 (KshA), 1:500 (KshC), or 1:740 (FdxD). Reactions were allowed to proceed for 16 hours at room temperature. KshA was further purified by FPLC. 2.4.4  FPLC protein purification Cyp125A1, KshA, and KshB were purified using anion exchange FPLC. Protein  samples were loaded onto a SourceTM15Q column equilibrated with 25 mM HEPES, pH 7.5. For KshA and KshB purifications this buffer also contained 5% glycerol, and this buffer was further supplemented with 1 mM dithiothreitol and 0.25 mM FAS for KshA purifications. The column was washed with three column volumes of the loading buffer, and 5 column volumes of 0.1 M NaCl in the same buffer at a flow rate of 3 mL min-1. Cyp125A1 was  72  eluted with a linear gradient from 0.10 – 0.16 M NaCl over 6 column volumes with Cyp125A1 (red) eluting at 0.14 M NaCl. KshA was eluted using a step gradient of NaCl, with KshA (brown) eluting at 0.15 M NaCl. KshB was eluted with a linear gradient of 0.22 to 0.36 M NaCl over 12 column volumes with KshB (orange) eluting at 0.34 M NaCl. Eluted proteins were exchanged into the column equilibration buffer, concentrated to 15-25 mgmL-1, flash frozen as beads in liquid nitrogen, and stored at -80 °C until use. 2.5  Small molecule analysis HPLC analysis was carried out using a Waters 2695 Separations HPLC module  equipped with a Waters 2996 photodiode array detector and a 250 x 4.60mm C18 Prodigy 10u ODS-Prep or 50 ! 4.6 mm LUNA 3 µm PFP(2) (KstD activity only) column (Phenominex, Torrance, CA). Gas chromatography-coupled mass spectrometry (GC-MS) was performed using an HP 6890 series GC system fitted with an HP-5MS 30 m x 250 "m column and an HP 5973 mass-selective detector (Hewlett-Packard, Palo Alto, CA). Nuclear magnetic resonance (NMR) spectra were recorded using a Varian Unity 500 MHz and Inova 600 MHz (Varian Inc., Mississauga, ON, Canada). Chemical shifts (#) are reported in ppm relative to CHCl3 (1H: d = 7.24) and CDCl3 (13C: d = 77.0) as internal standards. Ultraviolet-visible (UV-vis) absorption spectra were recorded using a Cary 5000 spectrophotometer equipped with a thermojacketed cuvette holder (Varian, Walnut Creek, CA). Steroid concentrations were measured spectrophotometrically using the following extinction coefficients for aqueous solution at 22 °C: $248 = 8.7 mM-1cm-1 for 4-BNC; $248 = 8.5 mM-1cm-1 for 1,4-BNC; $248 = 17.2 M-1cm-1 for 4-BNC-CoA; and $248 = 16.9 M-1cm-1 for 1,4-BNC-CoA. The extinction coefficient for 4-BNC was measured using a solution containing a gravimetrically determined amount of the compound. The extinction coefficient for 1,4-BNC was calculated from the ratios of values for 4-AD and ADD, and for 4-AD and 4-BNC, assuming a linear relationship for the contributions of the %1 bond and isopropionate group. The values for 4-BNC-CoA and 1,4-BNC-CoA were calculated from linear sums of the values for the steroid and CoA components.  73  2.6  Protein analytics Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) was  performed using a Bio-Rad MiniPROTEAN III apparatus with a 12% resolving gel. Gels were stained with Coomassie Blue according to standard protocols. 2.6.1  Protein quantification Protein concentrations were determined using the Micro BCATM protein assay kit  (Pierce, Rockford, IL) using bovine serum albumen as a standard. KshA concentrations were routinely determined using !280 = 142 mM-1cm-1 and !324 = 23.2 mM-1cm-1. KshB concentrations were routinely determined using !349 = 12.0 mM-1cm-1. P450 protein concentrations were calculated from the reduced CO-bound difference spectrum using the extinction coefficient !!450-490 = 91 mM-1cm-1 (269). 2.6.2  Evaluation of cofactor content  Identification of flavin cofactors  Flavins were identified using the method of Faeder & Siegel (270). Briefly, 5 "L of 400 "M KshB was diluted to 500 "L using 0.1 M potassium phosphate, pH 7.7 containing 0.1 mM EDTA in a light-shielded tube. The mixture was heated at 100 °C for 3 min, rapidly cooled on ice, then centrifuged at 10,000 g for 30 min at 4 °C. The supernatant was passed through a 0.2 "m filter, and 100 "L was injected onto the HPLC column equilibrated with aqueous 0.5% phosphoric acid and operated at a flow rate of 1 mLmin-1. Flavins were eluted using a gradient of 0 to 100% methanol in 60 mL. Solutions of FMN and FAD were used as standards.  74  Determination of iron content  Iron content was determined using the Ferene S assay (271) adapted for 96-well plate format. Briefly, 80 !L standards containing 5-75 !M FeCl2 and protein samples containing 1 to 2 nmol of iron were incubated for 10 min with 10 !L of 12 N HCl. Ten !L of 80% trichloroacetic acid was then added and protein precipitate was removed by centrifugation. Supernatants were added to 20 !L 45% sodium acetate in a 96-well plate, to which was then added 100 !L of Ferene S reagent (0.75 mM Ferene S, 10 mM L-ascorbic acid, 45% sodium acetate). Absorbances were read at 562 nm using a VMax Kinetic Microplate Reader (Molecular Devices, Sunnyvale, CA).  Determination of acid-labile sulfur content  Acid-labile sulfur content of samples was determined colorimetrically using the N,Ndiethyl-p-phenylene diamine assay (272). Briefly, 25 µL standards containing 5-25 µM Na2S were made from a stock solution of 1 mM Na2S in 0.1 M NaOH. These, and 25 µL samples of protein containing 5-15 nmol acid-labile sulfur were prepared in 3 mL glass vials equipped with septa, and the headspace was exchanged with N2 gas. To each vial, 250 µL of 1% zinc acetate was added, followed by agitation, addition of 12.5 µL 12% NaOH, and thorough mixing. After 30 min, 50 µL of 0.1% N,N-diethyl-p-phenylene diamine in 5.5 M HCl (made fresh and kept anaerobic and dark) was added followed by addition of 25 µL 40 mM FeCl3 in 1.2 M HCl below the surface of the liquid. After a further 30 min, the septa were removed from the vials, reactions were brought to 1.5 mL with water, protein samples were centrifuged at 16,100 g for 10 min to remove precipitated protein and absorbances were read at 670 nm.  75  2.6.3  Spectroscopic analysis  Determination of P450 iron spin state equilibrium  The proportion of purified P450 containing high-spin ferric heme iron was estimated by comparing the spectra of CYP125A1 and Cyp125A14P to linear combinations of the spectra of P450s in high and low spin states (273,274). High spin samples were generated by adding 0.5% Triton X-100 (Cyp125A14P) or 12 !M 4-cholesten-3-one in 10% BCD (Cyp125A1) to 3.3 µM enzyme. The substrate-free cytochrome P450cam from Pseudomonas putida (275) was used as a low spin standard. The same values were obtained when using low spin Cyp125A14P generated through addition of 40% methanol was used as a low spin standard.  P450 CO-bound spectra  The CO-bound form of P450s was generated by first incubating samples with ~8 mM sodium dithionite for 10 min and then slowly bubbling them with CO for 30 s. Spectra between 250 and 600 nm were recorded immediately following CO addition.  Reduced and oxidized spectra of iron-sulfur cluster proteins  Reduced KshA was prepared by buffer exchange into 0.1 M potassium phosphate, pH 7.0 by gel filtration chromatography under anaerobic conditions while oxidized KshA samples were prepared by addition of excess potassium ferricyanide under aerobic conditions followed by similar buffer exchange. Oxidized KshB samples were likewise generated through aerobic buffer exchange while reduced KshB was prepared via addition of excess NADH and subsequent buffer exchange under anaerobic conditions. Spectra were recorded in the same buffer used for gel filtration chromatography and reduced spectra were recorded anaerobically. 76  2.7  Enzyme activity assays  2.7.1  P450 substrate binding Substrate-induced spectral responses of Cyp125A14P and Cyp125A1 were recorded  in 0.1 mM potassium phosphate, pH 7.0 by titrating solutions of P450 with 1.0 mM stock solutions of cholesterol, 5!-cholestane-3"-ol, and 4-cholesten-3-one in 10% BCD or of an EDTA-linked "-cyclodextrin dimer (cholesterol titrations with Cyp125A1) (276). Equilibrium dissociation constants were calculated using a quadratic equation (Equation 1).  Equation 1  In this equation, #A is the change in absorbance observed in the sample, [S]T is the total ligand concentration, [E]T is the total enzyme concentration, KD is the equilibrium dissociation constant, and #AMax is the change in absorbance at infinite ligand concentration. A non-linear least-squares fit of the equation to the data was obtained using the program R ( 2.7.2  Cyp125A1 activity assay Cyp125A1 activity was measured by monitoring substrate and product concentrations  by HPLC. Solvents used for HPLC analysis were 0.5% aqueous phosphoric acid (A) and methanol (B). Compounds were detected at 242 nm and eluted at a flow rate of 1 mLmin-1 with a gradient from 80-90% B over 5 min, followed by isocratic elution at 90% B for 10 min, a gradient from 90-100% B for 5 min, and further isocratic elution for 15 min. Under these conditions, retention times for 4-cholesten-3-one and 4-cholesten-3-one-26-ol were 38.5 and 22.0 min, respectively. Concentrations of 4-cholesten-3-one and 4-cholesten-3-one26-ol were calculated from their respective peak areas using a standard curve of 4-cholesten3-one and assuming that the extinction coefficients for the two compounds are similar; the R2 77  value for the standard curve was > 0.99. The standard assay was performed in air-saturated 0.1 M potassium phosphate, pH 7.0 containing 900 !M NADH, 50 !M 4-cholesten-3-one, 1.5 !M KshB, and 1.6 !M Cyp125A1. Stock solutions of 1 mM 4-cholesten-3-one were made in 10% BCD and stored at 4 °C; stock solutions of 180 mM NADH were made fresh daily in water and stored on ice. Assays were conducted in multiple tubes containing 200 !L of the standard assay mixture. At each time point the reaction was quenched by the addition of 200 !L methanol and vigorous mixing. To investigate the dependence of the Cyp125A1-catalyzed reaction on O2 concentration, 200 !L of the standard assay mixture without substrate was equilibrated with a head space continuously flushed with an N2/O2 gas accompanied by vigorous stirring for 10 min. Reactions were initiated by addition of 50 !M 4-cholesten-3-one and were quenched after 15 minutes using 200 !L methanol. Steroid concentrations were determined by HPLC analysis. 2.7.3  KshAB activity assay KshAB activity was measured by following O2 consumption using an Oxygraph  Clarke-type electrode interfaced to a computer (Hansatech Instruments, Norfolk) according to the manufacturers’ instructions. Standard activity assays were performed at 22 °C in a total volume of 1 mL air-saturated 0.1 M potassium phosphate, pH 7.0 containing 430 !M NADH, and either 0.8 µM KshB and 0.4 µM KshA (4-AD and ADD); or 1.1 !M KshB and 0.2 !M KshA (other substrates). Reactions were initiated by adding substrate (4-BNC-CoA, 1,4BNC-CoA) or KshB (AD, ADD, 4-BNC, 1,4-BNC) after equilibration of all other components for 3 min. Stock solutions were prepared fresh daily. KshA was thawed, exchanged into 0.1 M potassium phosphate, pH 7.0 anaerobically using gel filtration chromatography, and stored in a sealed vial on ice. Aliquots were withdrawn as required using gas-tight syringes. Reaction rates were corrected for O2 consumption observed prior to reaction initiation. Apparent steady-state kinetic parameters for all substrates were determined by measuring initial rates of O2 consumption in the presence of various concentrations of substrate. The apparent steady-state kinetic parameters for O2 were 78  measured in the presence of 380 !M ADD or 57 !M 1,4-BNC-CoA. In these experiments, the reaction buffer was equilibrated for 20 min by vigorous bubbling with mixtures of N2 and O2 prior to the reaction, and the reaction cuvette was continually flushed with the same gas mixture during the reaction. Kinetic parameters were evaluated by fitting the MichaelisMenten equation to the data using the least-squares fitting and dynamic weighting options of LEONORA (277). Inhibition of the reaction by CoASH was evaluated under standard conditions at the respective KM values of ADD, 1,4-BNC, and 1,4-BNC-CoA. 2.7.4  KshAB coupling assay Determination of KshAB coupling was performed by measuring the consumption of  O2 and steroid substrate in the same reaction. Coupling experiments with ADD were conducted in air-saturated 0.1 M potassium phosphate, pH 7.0 with 1 mM phenylalanine in the presence of 1.3 !M KshA, and concentrations of all other components corresponding to those of the standard assay. Coupling experiments using 4-AD as a substrate were performed as above with buffer equilibrated with 80% O2. One mM phenylalanine was not found to change the rate of reaction with either substrate (results not shown). After 5 min, reactions were quenched by diluting 200 !L of the reaction mixture in 200 !L methanol. One hundred !L of this mixture was injected onto the column equilibrated with 30% methanol in 0.5% aqueous phosphoric acid. The column was operated at a flow-rate of 1 mLmin-1 and the sample was eluted with a step gradient of methanol, with phenylalanine eluting with 30% methanol and ADD and 4-AD eluting at 80% methanol. A standard curve of the peak area ratios of substrate and phenylalanine was used to quantify the substrate. O2 consumption was monitored using the oxygen electrode. 2.7.5  KstD activity assay Reactions investigating KstD activity were performed at 22 °C in a 1 mL volume  containing 200 !M AD or 4-BNC-CoA, 400 !M NAD+, and 0.04 mg (total protein) KstDlysate in 50 mM Tris/HCl, pH 7.4. For reactions containing 4-BNC-CoA the reaction pH was adjusted to 7.4 with NaOH before addition of the KstD-lysate. At various time points, 200 !L  79  of each reaction was quenched by rapid dilution into 200 !L methanol, evaporated to remove methanol, filtered, and analyzed by HPLC by Dr. Israël Casabon using the described conditions (265). 2.8  Phylogenetic analysis of RO-Os  2.8.1  Phylogenetic analysis of full-length RO-O ! subunits Super-positions of crystallographic coordinates were calculated using the SSM  algorithm in COOT (278). Z scores were calculated using DALI (279). Multiple sequence alignments were generated using STRAP (280) and ClustalX (281). Manual inspection of molecular models and construction of figures was done using Pymol (282). Phylogenetic analyses were performed using PHYLIP (283). 2.8.2  Structure-based template All published structures of independent RO-Os in the Protein Data Bank (PDB),  excluding variant structures were collected. Where more than one structure of the same enzyme was published, the highest resolution structure with no ligand and both metal centres in the reduced state was selected. The number and diversity of these structures was insufficient to make an effective template for the alignment of a broad range of RO-Os. Two approaches were therefore combined to construct the structure-based template: an alignment of the published RO-O structures (Table 1.2), and the alignment from the National Centre for Biotechnology Information (NCBI) manually curated domain cd00680 (RO-O catalytic domain) as represented in the program Cn3D (284). The structural alignment of the 17 ROOs in the PDB (Table 1.2) was produced with the STAMP algorithm (285) implemented in Multiseq (286). Two passes were performed with the following parameters: similarity set to three, comparison residues set to ten, and slow scan option. The alignment for NCBI domain family cd00680 contained 98 sequences at the time of writing and represents a multiple sequence alignment of structurally conserved residues in the RO-O catalytic domain with unaligned segments between the aligned blocks. ClustalX (281) was used with default 80  parameters to align these unaligned segments and the Rieske domain of these 98 sequences. A chimera structure-based template alignment was then constructed by performing profile-toprofile alignments using Clustal as implemented in SeaView (287). The STAMP alignment was used as the template profile for the Rieske domain and the cd00680-derived alignment was used as the template profile for the catalytic domain. The resultant alignment was named Template A. 2.8.3  Sequence collection To compile a list of sequences representative of known RO-O diversity, we used the  conserved domain architecture retrieval tool (CDART) with the conserved domain database (CDD) at NCBI to identify all proteins with the canonical domain architecture: N-terminal Rieske domain and C-terminal RHO_alpha domain. This search returned over 3400 sequences. Many additional sequences, such as KshA from Mtb, did not appear in the list retrieved by CDART as their catalytic domains were not recognized by NCBI CDD. The latter were manually selected from the NCBI database by collecting all sequences over 300 amino acids in length from the ~ 4000 sequences annotated as containing only a Rieske domain. This yielded approximately 400 sequences, from which were eliminated those with a C-terminal Rieske domain or no mononuclear iron binding motif, resulting in about 200 sequences. These ~200 sequences together with the ~3400 possessing the canonical domain structure were aligned within the taxonomic groupings provided in the CDART output with ClustalX using the Gonnet series matrix with default parameters. Sequences were eliminated from the analysis such that within each taxonomic group, sequences varied by a pair-wise distance (calculated using prodist in the PHYLIP package (283)) of !1.0. This data set contained 633 sequences. 2.8.4  Alignment construction A multiple sequence alignment of all of the 633 representative sequences was  constructed in SeaView (287) using Template A as a reference. In constructing the multiple sequence alignment, efforts were taken to minimize bias through process standardization.  81  Accurate alignment of the catalytic domain, however, necessitated the use of multiple strategies due to their high sequence diversity. Specifically, either MUSCLE (288) or Clustal algorithms were used, and sequences were aligned to Template A individually or in small, pre-aligned clusters. In this workflow, sequences were not added to a growing alignment; each sequence or small group was aligned to Template A and subsequently superimposed upon previously aligned sequences. In the final alignment used for phylogenetic reconstruction, we included only structurally conserved positions as determined by the STAMP alignment of structurally characterized RO-Os. Structurally homologous residues were identified using the Qres score calculated in STAMP. This is a metric for how similar the position of a particular residue is to the corresponding residues of the other proteins in the alignment. I took the average Qres score for each position in the alignment, performed a histogram analysis, and selected a Qres cutoff of ! 0.6 for positions considered structurally homologous. This constraint drastically reduced erroneously aligned sequence fragments that are prevalent between blocks of structurally conserved residues. The alignment used for the final phylogenetic analysis does not include sequences in the cd00680 template; it comprises the 17 sequences for which there are available structures and the 633 sequences selected from the NCBI database. The entire alignment was edited to remove all positions that had an average Qres value of less than 0.6 or were in small groups (four or less residues) of structurally conserved positions that did not align well over all sequences. This resulted in a final alignment of 159 residues for 650 proteins. 2.8.5  Phylogenetic reconstruction The final alignment was used to calculate a maximum likelihood tree using proml  program in the PHYLIP suite (283). This tree was subsequently trimmed to a smaller number of representative sequences based on pair-wise distances calculated using prodist; a cutoff of 1.3 resulted in 98 representative sequences for which the amino acid sequence identity of the truncated sequences does not exceed 40%. This trimmed alignment was used in bootstrap  82  analyses to calculate a consensus tree using 100 datasets generated using the PHYLIP program seqboot and calculating the corresponding maximum likelihood trees with proml using 12 jumbles per dataset. The consensus tree was calculated using the consense program of PHYLIP. The results of this analysis were similar using neighbour-joining and maximumlikelihood methods. Trees were visualized using Dendroscope (289) and Geneious (290). 2.8.6  Analysis of Rieske and catalytic domains Separate phylogenetic analyses were performed on the Rieske and catalytic domains.  The individual domains of the 17 RO-O structures were used to perform new STAMP alignments. Due to removed restraints from the second domain, this resulted in more positions meeting the structurally conserved criteria for each domain (96 and 87 positions for the catalytic and Rieske domains respectively). Maximum-likelihood consensus trees for each of the two domains were calculated using 100 datasets and 5 jumbles per dataset.  83  3  Results  3.1  Characterization of Cyp125  3.1.1  Protein production and purification  Cyp125A14P from R. jostii RHA1  Recombinant R. jostii RHA1 Cyp125A14P bearing a 6-His tag was homologously produced in R. jostii RHA1 using the expression plasmid pTipQC1-cyp125RHA1. The recombinant protein made up a significant proportion (~ 15%) of total protein in the R. jostii cell-free extract as evaluated by SDS-PAGE chromatography (Figure 3.1 A). Addition of !aminolevulinic acid and other additives that are usually necessary to promote expression of properly folded and soluble P450 proteins in E. coli was not needed for homologous production of Cyp125A14P in R. jostii RHA1. The enzyme was purified using Ni2+-NTA affinity chromatography, yielding ~ 5 mg purified material per litre of culture. SDS-PAGE analysis indicated that the Cyp125A14P constituted greater than 95% of total protein in the preparation (Figure 3.1 A).  Figure 3.1  SDS-PAGE gels of Cyp125 expression and purification. Purification of Cyp125A14P (A) and  Cyp125A1 (B) produced in R. jostii RHA1. Lane 1: broad range protein standards; lane 2: cell-free extract; lane 3: purified protein. The molecular weights (kDa) are indicated at the left of each protein standard lane.  84  Cyp125A1 of Mtb H37Rv  Cyp125A1 from Mtb H37Rv was produced heterologously in R. jostii RHA1 using the expression plasmid pTipCP125. Purification of Cyp125A1 by anion exchange chromatography yielded ~4 mg purified protein per litre of harvested culture. The purified protein had a mass of 46.4 kDa as measured by MALDI-TOF mass spectrometry analysis. SDS-PAGE analysis indicated that Cyp125A1 constituted > 99% of the protein in the purified sample (Figure 3.1 B).  Electronic absorption spectroscopy  The spectroscopic properties of isolated Cyp125A14P and Cyp125A1 were atypical of bacterial P450s in that, as purified, both proteins contained heme iron that was primarily in the high spin state. The UV-vis spectra of the two proteins were similar, exhibiting peaks at 280 nm and 392 nm, with a shoulders centred at 422 nm (Figure 3.2). This finding is consistent with the spectroscopic properties of Cyp125A1 found by Ouellet et al. (190). As isolated, Cyp125A1 had an absorbance ratio A392/A280 of 1.11 and was estimated to be 70% in the high spin state. Likewise, Cyp125A14P was estimated to contain ~93% high spin iron. Exposure of dithionite-reduced Cyp125A14P and Cyp125A1 to CO revealed reduced COdifference spectra with maxima at 451 nm and 449 nm, respectively (Figure 3.2 insets). These peaks indicate that the heme iron thiolate ligation remained intact during protein purification.  85  Figure 3.2  UV-vis spectra of Cyp125. Spectra of 2.9 µM Cyp125A14P (A) and 3.4 µM Cyp125A1 (B) as  isolated with oxidized heme iron (solid line), incubated with 10 µM cholesterol (A) or 4-cholesten-3-one (B) (dashed line), and upon reduction with excess sodium dithionite (dotted line) are shown. The reduced COdifference spectra are shown in the inset graphs. Spectra were collected at 25 °C in 0.1 mM potassium phosphate, pH 7.0.  86  3.1.2  Steroid binding assays  Cyp125A14P of R. jostii RHA1  The capacity of Cyp124A14P to bind sterols was investigated using spectroscopic assays. Sterol-binding was investigated by examining the relative absorbance at 392 nm (high spin heme iron) and 422 nm (low spin heme iron); ligand binding adjacent to the heme iron causes a shift in the iron from a hexa-coordinated low spin state to a penta-coordinated high spin state in P450s. Following the addition of cholesterol, 5!-cholestanol, or 4-cholesten-3one in a solution of 10% 2-hydroxypropyl-"-cyclodextrin, Cyp125A14P exhibited a spectral change with a pronounced trough at 422 nm and a peak at 392 nm, which is consistent with the decrease in the low-spin character of the heme iron associated with substrate binding (Figure 3.3). No such spectral shift was observed upon addition of only BCD, which was used to solubilize the steroid ligands. The difference spectrum also exhibited a perturbation at 395 nm in comparison to the typical type I binding spectrum. A perturbation at the same wavelength was observed upon addition of 5-cholestene-26-oic acid-3"-ol in 10% BCD, although the acid elicited no underlying type I spectral change at concentrations up to 20 #M (Figure 3.3 A). Cholesterol also induced a type I binding spectrum when added in the presence of other solubilizing agents such as Triton WR1339 and dimethylsulphoxide. The spectral shifts, however, were much weaker than in the presence of 2-hydroxypropyl-"cyclodextrin.  87  Figure 3.3  Binding of steroids to purified Cyp125A14P. (A) Spectral responses of 3.7 µM purified  Cyp125A14P induced by 10 µM cholesterol (solid line) and 10 µM 5-cholestene-26-oic acid-3!-ol (dashed line). The dependence of the absorbance change of Cyp125A14P at 422 nm on cholesterol (B), 5"-cholestanol (C), and 4-cholestene-3-one (D) concentration. The best fit of Equation 1 (Materials and Methods) is represented as a grey line. Spectra were collected at 25 °C in 0.1 mM potassium phosphate, pH 7.0. Steroids were prepared as 1.0 mM stock solutions in 10% BCD.  Using Equation 1 (Materials and Methods), apparent KD values for cholesterol, 5!cholestane-3"-ol, and 4-cholesten-3-one were evaluated (Table 3.1). The concentrations of enzyme calculated using this equation (4.0, 4.3, and 3.6 #M, respectively) were within 15% of the enzyme concentration calculated using the extinction coefficient for the reduced COdifference spectrum of $450-490 = 91 mM-1cm-1 (3.7 #M), although this extinction coefficient has not been independently verified for this isozyme. The high quality fit of the equation to the binding data (Figure 3.3 B, C, D) supports a binding stoichiometry of 1:1 and suggests that Cyp125A14P isolated with the protocol described here does not harbor a ligand despite the proportion of high-spin iron. Finally, Cyp125A14P exhibited maxima at 451 nm in COdifference spectra taken after each binding experiment, which indicates that the heme-thiolate ligation remained intact.  88  Table 3.1  Apparent dissociation constants of Cyp125 for steroidsa KD (!M)  Cyp125A14P cholesterol 5!-cholestanol 4-cholesten-3-one Cyp125A1  0.20 (0.08)b 0.15 (0.03) 0.20 (0.03)  cholesterol 4-cholesten-3-one  0.20 (0.02) 0.27 (0.05)  a  Cholesterol was prepared as an aqueous stock solution in 25 mM of an EDTA-bridged BCD dimer. Other steroids were prepared as aqueous stock solutions in 10% BCD b Values for standard error for triplicate data are shown in parentheses.  Cyp125A1 of Mtb H37Rv  Like Cyp125A14P, incubation of Cyp125A1 with 4-cholesten-3-one and cholesterol resulted in a type I spectral shift that was not elicited by the addition of BCD alone. Specifically, when Cyp125A1 was exposed to 4-cholesten-3-one the 422 nm shoulder disappeared and the A392/A280 ratio increased to 1.20 (Figure 3.4). Binding of 5-cholestene26-oic acid 3"-ol, a proposed product of the Cyp125A1 reaction (264), was not observed. The apparent KD of Cyp125A1 for 4-cholesten-3-one was 0.27 ± 0.05 #M (Table 3.1). Attempts to determine a KD for cholesterol under these conditions were unsuccessful as a partial reversal of the spectral shift was observed at concentrations of BCD exceeding 0.02%. Such concentrations of BCD were attained well before the enzyme was saturated with steroid. To solve this problem, 5 mM cholesterol was solubilized in 25 mM of an EDTAbridged "-cyclodextrin dimer (276). Using this reagent, we were able to obtain a binding curve for cholesterol with an apparent KD of 0.20 ± 0.02 #M (Table 3.1). The binding reaction with cholesterol under these conditions took much longer to equilibrate (~15 min) than for 4-cholesten-3-one with BCD. With the dimer, 4-cholesten-3-one was not as soluble as cholesterol, and a similar problem of partial reversal of the spectral shift was observed using 4-cholesten-3-one solubilized in this manner. Finally, using either steroid complexed with the dimer, a shoulder at 422 nm remained, even at saturation, which was not the case for binding studies performed using the 4-cholesten-3-one with BCD. Several additional solvent  89  systems were tried for both steroids without success. Thus, a direct comparison of the binding affinity of Cyp125A1 for cholesterol versus 4-cholesten-3-one was not possible.  Figure 3.4  Binding of 4-cholesten-3-one to Cyp125A1. Difference spectra of 3.3 µM Cyp125A1 upon the  addition of 0.33 to 10 µM 4-cholesten-3-one are overlaid. Spectra were collected at 25 °C in 0.1 mM potassium phosphate, pH 7.0. The resultant binding curve is inset with error bars representing the S.E. calculated from triplicate data. The solid line represents the best fit of Equation 1 to the data with the fitted parameters: KD = 0.27 ± 0.05 µM, [Cyp125A1] = 3.3 ± 0.2 µM, !A422Max = 0.055 ± 0.001.  3.1.3  Cyp125A1 activity assays Based on the binding results, we investigated the ability of Cyp125A1 to transform 4-  cholesten-3-one and cholesterol. 4-Cholesten-3-one solubilized in 10% BCD was used for quantitative Cyp125A1 activity analysis because of its ability to fully convert the P450 to a high spin state, short equilibration time for binding, and UV-vis spectrum allowing for quantification by HPLC analysis. Bioinformatic analyses were performed to identify potential redox partners for Cyp125A1. These analyses indicated that the rhodococcal and mycobacterial cholesterol catabolic clusters contain more oxygenases than NAD(P)Hutilizing reductases. Moreover, the four steroid catabolic clusters identified in R. jostii RHA1 90  do not appear to each encode a full complement of reductases, suggesting that the reductase encoded by one cluster might be used by the oxygenases encoded by another cluster (21). Additionally, a kshB deletion mutant of R. erythropolis SQ1 was unable to degrade the sterol side chain, prompting the authors to propose that KshB may function as part of the sterol 26hydroxylase enzyme system (85). Accordingly, we attempted to reconstitute Cyp125A1 activity in vitro using KshB (Rv3571) as the reductase. In assays containing Cyp125A1, KshB and NADH, 4-cholesten-3-one was consumed and a product was formed (Figure 3.5). Assuming that this product has a similar extinction coefficient to the substrate, it was formed in a 1:1 ratio with depleted 4-cholesten-3-one, and no other product was detected after incubation up to 24 hr. Substrate turnover was dependent on each of NADH, KshB, and native Cyp125A1 and omitting any of these components or replacing Cyp125A1 with enzyme which had been heat inactivated by incubation at 65 °C for 30 min abolished substrate consumption and product formation. The reconstituted Cyp125A1 also transformed cholesterol. The rate of the observed reaction using either steroid was very low, with a 4-cholesten-3-one specific activity of 0.01 U/mg (Figure 3.5). If Cyp125A1 was saturated with steroid under these conditions, this rate would correspond to an apparent kcat value of 31 h-1.  Figure 3.5  Turnover of 4-cholesten-3-one by Cyp125A1. Depletion of 4-cholesten-3-one and accumulation of  26-hydroxy-4-cholesten-3-one (inset) are represented by squares and circles, respectively. Assays contained 900 µM NADH, 50 µM 4-cholesten-3-one, 1.5 µM KshB and 1.6 µM Cyp125A1 in 0.1M potassium phosphate, pH 7.0. 91  In attempts to increase the rate of reaction, each of two potential ferredoxin components were added to the reaction mixture. Both FdxD (Rv3503c, 6.7 KDa) and KshC (Rv3527, 16.7 KDa), encoded in the cholesterol degradation gene cluster of Mtb, are predicted to encode small ferredoxins. The physiological role of neither protein has been determined. Each protein was heterologously produced in E. coli and purified anaerobically with Ni2+-NTA chromatography. The absorption spectra of both purified proteins had broad absorption bands consistent with FeS clusters (Figure 3.6). Inclusion of either purified protein, with or without the histidine tag, however, failed to increase the rate of Cyp125A1catalyzed steroid hydroxylation. Substitution of KshB for purified toluate 1,2-dioxygenase reductase component (XylZ) of Psuedomonas putida mt-2 (231) or biphenyl dioxygenase reductase and ferredoxin (BphFG) from Commomonas testosteroni B-356 (291) both resulted in a decrease in Cyp125A1 reaction rate.  Figure 3.6  Spectra of purified putative ferredoxins. The UV-vis spectra of 9 µM KshC (A) and 55 µM FdxD  (B) purified by Ni2+-NTA chromatography are shown. Spectra were collected at 25 °C in 0.1 mM potassium phosphate, pH 7.0.  The rate of the Cyp125A1 reaction was also investigated with respect to oxygen concentration. There was a linear correlation between Cyp125A1-dependent steroid consumption and oxygen concentration in buffer containing from 116 to 1160 !M O2. This finding indicates that the apparent KMO2 of Cyp125A1 was > 1.2 mM. The estimated apparent kcat/KMO2 value of Cyp125A1 was 11 M-1s-1.  92  Figure 3.7 Dependence of the Cyp125A1 reaction on oxygen concentration. The rate of 26-hydroxy-4cholesten-3-one production is shown as circles. The linear line of best fit with slope = 0.0011 min-1 (R2 value of fit 0.95) is shown as a solid line. Assays contained 900 µM NADH, 50 µM 4-cholesten-3-one, 1.5 µM KshB and 1.6 µM Cyp125A1 in 0.1M potassium phosphate, pH 7.0.  The reaction products were identified by their mass fragmentation patterns as modified 4-cholesten-3-one or cholesterol bearing a hydroxyl group on the aliphatic side chain. For example, the ion representing the total mass of each molecule (472 and 546 m/z for reactions with 4-cholesten-3-one and cholesterol, respectively) differed from the substrate molecules by the mass of one trimethylsianol-derivatized hydroxyl group (88 a.m.u.). Additionally, fragments representing the steroid nucleus (271 m/z for 4-cholesten-3-one missing C20-C27 and 255 m/z for cholesterol missing C20-C27 and the oxygen at C3) remain unaltered in both substrate-product pairs. Using a combination of 1H, HSQC, HMBC, and COSY NMR, the following chemical shifts for the hydroxylated 4-cholesten-3-one product were determined: 1H NMR (500 MHz, CDCl3): ! = 5.72 (s, 1H, 4), 3.50 (dd, 3J = 6.0, 10.0 Hz, 1H, 26x), 3.43 (dd, 3J = 6.0, 10.0 Hz, 1H, 26y), 2.40 (m, 4H, 2xy, 6xy), 2.01 (m, 2H, 1x, 12x), 1.83 (m, 2H, 7x, 16x), 1.69 (m, 1H, 1y), 1.61 (m, 2H, 15x, 25), 1.51 (m, 2H, 11x, 8), 1.44 (m, 1H, 11y), 1.40 (m, 4H, 20, 22x, 23x, 24x), 1.27 (m, 1H, 16y), 1.18 (s, 3H, 19xyz), 1.12 (m, 4H, 12y, 15y, 17, 23y), 1.02 (m, 4H, 7y, 14, 22y, 24y), 0.92 (m, 7H, 9, 21xyz, 27xyz), 0.71 (s, 3H, 18xyz); 13C NMR 93  (125.8 MHz, CDCl3): ! = 199.7 (3), 171.7 (5), 68.2 (26), 55.9 (17), 55.7 (14), 53.6 (9), 42.3 (13), 39.4 (12), 38.5 (10), 36.0 (22), 35.8 (4), 35.6 (25), 35.5 (1, 20), 35.4 (8), 33.8 (2), 33.5 (24), 32.8 (6), 31.9 (7), 28.0 (16), 24.0 (15), 23.3 (23), 20.8 (11), 18.5 (21), 17.2 (19), 16.5 (27), 11.8 (18). Notably, five methyl groups were distinguishable in the spectrum for 4cholesten-3-one, with the C26 and C27 protons represented by overlapping doublets centered at ! = 0.87 (data not shown). The spectrum of the reaction product did not contain this overlapping doublet, instead containing a signal at ! = 0.92 that is consistent with a downshifted methyl group and two protons on a hydroxylated carbon (! = 3.50, 3.43). These results enabled the identification of the product as 26-hydroxy-4-cholesten-3-one. The stereochemistry at C25 was not determinable from this experiment. 3.2  Characterization of KshAB  3.2.1  Protein production and purification Anaerobic purification of KshA of Mtb H37Rv from E. coli using the expression  plasmid pETKA1 yielded 5 mg KshA per litre of culture. Over 95% of the protein in the preparation was KshA, as determined by SDS-PAGE analysis. The preparation contained 8.2 ± 0.7 mol iron and 2.4 ± 0.4 mol sulfur per mol of KshA protomer. Aerobic incubation of KshA for 10 min with 2 mM EDTA followed by buffer exchange by gel filtration chromatography lowered the iron content to 3.9 ± 0.3 mol per mol KshA without affecting the specific activity. The spectrum of oxidized KshA had maxima at 324 and 455 nm (Figure 3.8) characteristic of Rieske-type [2Fe2S] clusters. The preparation had an R-value (A280/A324) of 5.8 and a specific activity of 0.76 ± 0.07 units/mg in the standard assay containing 430 µM ADD.  94  Figure 3.8 Spectra of reduced and oxidized KshA and KshB. UV-vis spectra of KshA (A) and KshB (B) in reduced and oxidized forms are shown as heavy and light lines, respectively. All spectra were recorded in 0.1M potassium phosphate, pH 7.0.  Purification of Mtb KshB from E. coli containing pETKB3 by anion exchange chromatography yielded 25 mg KshB from one litre of culture. This material contained 1.7 ± 0.1 mol iron and 2.0 ± 0.5 mol sulfur per KshB monomer. The flavin of KshB eluted from a C18 HPLC column with a retention time and spectrum consistent with FAD. The spectrum of oxidized KshB had maxima at 274, 349 and 457 nm (Figure 3.8), which is consistent with its flavin and [FeS] cluster content. The preparation had an R-value (A274/A349) of 4.6. Production of Mtb KstD in E. coli resulted in very little soluble protein in the cell-free extract, with most protein being incorporated into inclusion bodies, and significant enrichment of KstD by Ni2+-NTA chromatography was not observed. Various expression strains, temperatures, and conditions were tried and did not improve solubility. Nonetheless the cell-free extract (KstD-lysate) of E. coli cells containing KstD was active toward steroid substrates (see below). 3.2.2  Reconstitution of activity Under the conditions of the standard oxygraph assay, the rate of O2 consumption was  directly proportional to KshA concentrations from 0.1 to 4.0 !M. Using the 17-keto steroid substrates 4-AD and ADD (Figure 3.9), KshA was not saturated with KshB, even at a tenfold excess of KshB over the KshA protomer. Optimal signal-to-noise ratios were obtained  95  using 0.4 !M KshA and 0.8 !M KshB and these concentrations were used in assays containing these substrates. KshA saturation with KshB was observed for steroid substrates with isopropionic acid or isopropionyl CoA thioester groups at C17 (Figure 3.9) and 0.2 µM KshA and 1.1 µM KshB were used in these reactions. A variant of KshB fused to a Cterminal polyhistidine-tag was also tested, but resulted in 50% lower specific activity despite containing comparable levels of iron and sulfur as wild-type KshB. Finally, assays conducted over a pH range from 6.5 to 8.0 revealed that while KshAB activity was greater at lower pH, the signal-to-noise ratio was also significantly lower. Accordingly, assays were performed at pH 7.0. O  A) O  O  B) O  4-AD  ADD  O  C)  OH  O  D)  OH  O  4-BNC  E)  O  1,4-BNC O S  O N H  O  O O N OPOPO N N O- OH OH O  O  1,4-BNC-CoA  -O  NH2 N N  O OH PO OH  Figure 3.9 KshAB substrates under investigation. A) 4-androstene-3,17-dione (4-AD); B) 1,4-androstadiene3,17-dione (ADD); C) 3-oxo-23,24-bisnorchol-4-en-22-oic acid (4-BNC); D) 3-oxo-23,24-bisnorchola-1,4dien-22-oic acid (1,4-BNC); E) 3-oxo-23,24-bisnorchola-1,4-dien-22-oyl-Coenzyme A thioester (1,4-BNCCoA). Not shown is 3-oxo-23,24-bisnorchol-4-en-22-oyl-Coenzyme A thioester (4-BNC-CoA) which differs from E) by a 1-2 saturated bond on ring A.  3.2.3  Reaction products and coupling Under the standard assay conditions, KshAB transformed ADD to a single product  whose GC retention time and mass spectrum were identical to those of 3-HSA (Figure 3.10). This is the expected reaction product following 9!-hydroxylation of the ADD steroid nucleus. Moreover, HPLC and oxygraph analyses performed on a methanol-quenched 96  reaction revealed that 1.02 ± 0.07 moles of O2 were consumed per mole of ADD consumed. As the transformation of 4-AD proceeded much more slowly (see below), coupling studies of this reaction were conducted in buffer equilibrated with 80% O2, which increased the reaction rate ~3-fold. In this reaction, 1.1 ± 0.1 mol O2 were consumed per mole of 4-AD consumed. When 650 U catalase was added to the reaction with either substrate at a time point corresponding to the methanol quench, no O2 production was detected. This finding indicates that no H2O2 was produced. Taken together, these results indicate that the KshAB-catalyzed transformations of both 4-AD and ADD are well coupled to O2 consumption.  Figure 3.10 Fragmentation pattern of KshA-catalyzed transformation of ADD. GC-MS Rt = 15.32 min; EI-MS (70 eV, EI); m/z: 300 ([M+], 15%), 134 (100%), 121 (15%), 77 (8%), 55 (6%). This retention time and fragmentation pattern correspond to those of authentic 3-hydroxy-9,10-seconandrost-1,3,5(10)-triene-9,17-dione (3-HSA).  Consistent with KshAB catalyzing the 9!-hydroxylation of 4-AD and ADD, each of 4-BNC, 1,4-BNC, 4-BNC-CoA, and 1,4-BNC-CoA (Figure 3.9) stimulated the depletion of oxygen in the presence of KshA, KshB, and NADH in the standard oxygen electrode activity assay. For each steroid substrate, the amount of oxygen consumed corresponded with the calculated amount of substrate added. These data are consistent with full coupling between oxygen and steroid substrate consumption and indicate that the extinction coefficients calculated for each of 1,4-BNC, 4-BNC-CoA, and 1,4-BNC-CoA are accurate. To confirm that KshAB catalyzes the 9!-hydroxylation of the thioester substrates, a reaction mixture of KshAB with 1,4-BNC-CoA was examined by HPLC. The product peak exhibited the 97  retention time and spectrum identical to those of 3-HSBNC-CoA (Figure 1.8). Moreover, alkaline hydrolysis yielded 3-hydroxy-9-oxo-9,10-seco-23,24-bisnorchola-1,3,5(10)-trien-22oyl-coenzyme A thioester (3-HSBNC) and CoASH, as determined by GC-MS and HPLC analyses, confirming the product’s identity. 3.2.4  Steady state kinetic analysis In air-saturated buffer the initial rates of O2 utilization by KshAB in the presence of a  variety of steroid substrates (Figure 3.9) displayed Michaelis-Menten kinetics (Figure 3.11 A) with the estimated apparent parameters indicated in Table 3.2. For some substrate concentrations, however, initial rates of O2 consumption were too low to afford reliable estimates of KM and kcat/KM (Table 3.2). Under the standard reaction conditions the apparent specificity constants were 1,4-BNC-CoA > 4-BNC-CoA > 4-BNC > ADD > 1,4-BNC > 4AD. Significantly, apparent kcat/KM values for all compounds except 1,4-BNC were greater than for either previously proposed 17-keto substrates, 4-AD and ADD (85). In fact, the value for 1,4-BNC-CoA (160,000 ± 20,000 M-1s-1) is over twenty times greater than that for the previously proposed physiological substrate ADD (246), suggesting that the physiological substrate(s) for KshAB is a CoA-thioester intermediate of cholesterol side chain degradation.  98  Figure 3.11 Steady state kinetic analyses of KshAB. Shown is the dependence of the initial velocity of O2 consumption on ADD (A), 4-AD (B), 1,4-BNC-CoA (C), and 4-BNC-CoA (D) concentration in air saturated buffer. Assays were performed in 0.1M potassium phosphate, pH 7.0 containing 430 µM NADH, and either 0.8 µM KshB and 0.4 µM KshA (4-AD and ADD) or 1.1 µM KshB and 0.2 µM KshA (4-BNC-CoA and 1,4-BNCCoA).  Table 3.2  a  Apparent steady state kinetic parameters of KshAB using various steroids  Substrate  KM (µM)  kcat (s-1)  kcat/KM (M-1 s-1)  4-AD  24 (16)a  0.07 (0.01)  3,000 (2,000)  ADD  110 (20)  0.80 (0.05)  7,600 (700)  4-BNC  3 (2)  0.08 (0.01)  30,000 (10,000)  1,4-BNC  70 (10)  0.25 (0.02)  3,500 (400)  4-BNC-CoA  6.8 (0.1)  0.61 (0.03)  90,000 (10,000)  1,4-BNC-CoA  17 (3)  2.7 (0.2)  160,000 (20,000)  Values reported in parentheses represent standard error. 99  For substrates with the same group at C17, the turnover numbers (kcat) were consistently greater for substrates with !1,4 desaturated A rings vs. those with !4 desaturated structure. Significantly, the turnover numbers for the CoA thioester substrates were approximately nine and three times greater than for the corresponding 17-keto steroids 4-AD and ADD, respectively. The KM values for 4-BNC-CoA and 1,4-BNC-CoA were also 3.5 and 6.5 times less than those for their respective 17-keto derivatives. In the presence of 380 "M ADD, the initial rates of O2 utilization by KshAB depended linearly on the concentration of O2 up to ~1.2 mM, the maximal attainable in this assay (Figure 3.12). This finding suggests that the enzyme’s apparent KMO2 exceeds this value. Fitting the Michaelis-Menten equation to the data yielded a kcat/KMO2 value of 2450 ± 80 M-1s-1. The estimates of KMO2 (46 ± 120 mM) and kcat (110 ± 290 s-1), however, were poor. The reactivity of KshAB with O2 was also investigated in the presence of 1,4-BNC-CoA. In the presence of 57 "M 1,4-BNC-CoA the apparent steady-state kinetic parameters for O2 were KMO2 = 90 ± 10 "M, kcat = 2.5 ± 0.1 s-1, and kcat/KMO2 = 29,000 ± 2,000 M-1s-1. The latter value is an order of magnitude greater than that measured in the presence of ADD. Perhaps more significantly, the KMO2 in the presence of 1,4-BNC-CoA is similar to that of the other oxygenases in the cholesterol degradation pathway HsaC (72) and HsaAB (54), as well as the O2 concentration in the lung tissues of healthy rabbits (123).  100  Figure 3.12 Dependence of KshAB reaction rate on O2 concentration. Shown is the dependence of the initial velocity of O2 consumption of KshAB on O2 concentration in the presence of ADD (A) and 1,4-BNC-CoA (B). Assays were performed in 0.1M potassium phosphate, pH 7.0 containing 430 µM NADH, and either 0.8 µM KshB, 0.4 µM KshA, and 380 µM ADD, or 1.1 µM KshB, 0.2 µM KshA 57 µM 1,4-BNC-CoA.  The apparent specificity constants for steroid metabolites were measured in airequilibrated buffer, so the high KMO2 for the keto-substituted steroid means O2 was well below saturating levels. In contrast, the lower KMO2 in the presence of a CoA thioester substrate means that in the presence of this steroid KshAB is closer to being saturated with O2. Thus, the increased reactivity of KshAB for O2 in the presence of the CoA thioester substrates could account for the greater apparent substrate specificity compared to the 17-keto substrates. The ability of CoASH to inhibit enzymatic turnover of either ADD or 1,4-BNC-CoA was tested to further probe the role of the CoA moiety in KshAB reactivity. These assays were carried out under the conditions of the standard assay with the steroid substrates at concentrations corresponding to their respective KM values. No inhibition was observed at CoASH concentrations up to 10-fold greater than substrate concentration.  101  3.2.5  Reaction of KstD with 4-AD and 4-BNC-CoA To further explore the possibility that CoA thioesters are the relevant substrates for  cholesterol ring catabolic enzymes in Mtb, the reaction rates of KstD with each of 4-AD and 4-BNC-CoA were investigated (Figure 3.12). Each compound was incubated with 40 !g (total protein) KstD-lysate and the reaction progress was followed by HPLC analysis in a discontinuous assay. After two hours, 87% of 4-BNC-CoA had been converted to 1,4-BNCCoA while only 30% of 4-AD had been turned over. Complete conversion is not observed over extended periods without the addition of fresh enzyme, suggesting that there is a timedependent decrease in enzyme activity. Under these conditions, the specific activity of 1,4BNC-CoA was estimated at 195 !M mg-1 min-1 while that for 4-AD was 25 !M mg-1 min-1. Similar results were obtained using each of three different concentrations of KstD-lysate.  Figure 3.13 Turnover of 4-AD and 4-BNC-CoA by KstD-lysate. The concentration of 4-AD (squares) and 4BNC-CoA (circles) as measured by HPLC analysis is shown. Assays were performed in 50 mM Tris/HCl, pH 7.4 containing 400 µM NAD+, 200 µM steroid, and 0.04 mg (total protein) KstD-lysate.  102  3.2.6  KshAH37Rv crystal structure  Overall structure  The structure of Mtb KshA was been determined to 2.3 Å resolution (KshAH37Rv structural model, PDB entry 2ZYL) by Dr. Igor D’Angelo through collaboration with Prof. Natalie Strynadka (246). Although the protein was crystallized aerobically, the crystallographic data suggests an intact mononuclear iron active site and Rieske cluster. The refined KshAH37Rv structural model includes residues 14-374 (excluding residues 283-284 for which electron density was not visible), three iron ions, two acid-labile sulfur atoms, and 176 water molecules. The overall structure of KshAH37Rv exhibits many features common to all known RO-O structures. Specifically, the KshAH37Rv protomer comprises an N-terminal Rieske domain encompassing residues 24 to 153 that harbors the [2Fe2S] Rieske cluster and a larger C-terminal catalytic domain that contains the catalytic iron centre and includes residues 154 to 374 (Figure 3.14). Three KshAH37Rv protomers are arranged in what has been described as a “head-to-tail assembly” around the three-fold symmetry axis in the crystal to make up the trimeric quaternary structure of the functional enzyme. In this assembly, the Rieske cluster of one protomer is within approximately 12 Å of the active site iron of the adjacent protomer, allowing for electron transfer through a network of hydrogen bonds over this distance. The KshAH37Rv trimer measures 90 Å at its greatest width and features a conical central pore 20-30 Å in diameter.  103  Figure 3.14 Crystal structure of KshA from Mtb. (A) Ribbon representation of the KshAH37Rv trimer viewed along the three-fold symmetry axis. Monomers are colored in different shades of green. Iron ions and acidlabile sulfur atoms are shown as orange and yellow spheres, respectively. (B) Ribbon representation of the KshAH37Rv monomer. The Rieske domain is labeled and represented in light brown; the catalytic domain is labeled and shown in green. The location and position of the beta strand !19 and terminal helix "9 are indicated. (C) Mononuclear iron coordination sphere. The ribbon representation of KshAH37Rv is shown as semi-transparent background. The residues ligating the iron are shown as stick representations. The mononuclear iron and solvent molecule are shown as orange and red spheres, respectively. 2Fo-Fc electron density (blue mesh) is shown at 1.7 #. Residual Fo-Fc density green mesh is shown at 3.5 #. (D) Active site and adjacent Rieske center of KshAH37Rv. The distance between atoms involved in proposed electron transfer is indicated. Residues from the adjacent molecule are indicated with a (*). Metal centers are shown as spheres, side chains are labeled and shown as stick diagrams.  The Rieske domain of KshAH37Rv is highly similar to that of other !3 RO-Os with root mean square deviation (r.m.s.d.) values of 1.9 Å over 119 C! atoms, 1.6 Å over 117 C! atoms, and 2.0 Å over 121 C! atoms for carbazole 1,9!-dioxygenase (CARDOJ3) of  104  Janthinobacterium sp. strain J3 (PDB entry 1WW9) (243), Dicamba monooxygenase (DMO) of Stenotrophomonas maltophilia (PDB entry 3GTE) (211), and 2-oxyquinoline 8monoxygenase (OMO86) of Pseudomonas putida strain 86 (PDB entry 1Z02) (220), respectively. As in these enzymes, the Rieske domain of KshAH37Rv is made up of three antiparallel !-sheets, the third of which harbors the Rieske cluster ligands. Loops between strands !4 and !5 and strands !6 and !7 contain residues Cys67 and Cys86 that coordinate one iron of the Rieske cluster, and His69 and His86 that coordinate the other. The catalytic domain of KshAH37Rv exhibits the TATA-binding protein (TBP)-like fold typical of RO-Os. This domain is composed of an eight-stranded anti-parallel !-sheet (strands !12-!19) (Figure 3.15) flanked by a series of "-helices on the face of the sheet distal to the Rieske domain (Figure 3.14). The active site iron is positioned at the center of this domain and accessible to solvent through a channel lined with primarily hydrophobic residues. The catalytic domain features a C-terminal helix ("9, residues 357-373) situated on the exterior of the KshA trimer where it forms contacts with residues 103-106 of the Rieske domain of the adjacent subunit (Figure 3.14). In the tertiary structure, this helix is located in the same place as a !-hairpin in CARDOJ3 and OMO86 previously termed the “trimerization domain,” and like this hairpin appears to stabilize the trimer. No structure is present at this position in the DMO structure, however, suggesting that it is not crucial for the integrity of !3 RO-O trimers. KshA also bears an additional ! strand (!19) (Figure 3.14 B, Figure 3.15) at the C-terminal end of the central ! sheet that is not observed in other "3 RO-O structures. This feature is shared with the "3!3 ROs. In the latter proteins, however, this strand is Nterminal to a structurally conserved ! helix !7, rather than C-terminal to this structure, as in KshAH37Rv.  105  Figure 3.15 The secondary structure of the KshAH37Rv catalytic domain. !-Helices and "-strands are represented to scale by cylinders and arrows, respectively. ! Strands and ! helices are numbered. Elements conserved in all structurally characterized ROs are shown in white. The positions of mononuclear iron ligands and Asp178 are indicated with asterisks.  Structure of the KshAH37Rv active site  The ferrous active site iron is coordinated in a distorted five-coordinate tetragonal geometry (Figure 3.13) by residues His181, His186, Asp304 (bidentate), and an exogenous ligand. The former two residues are situated on helix !4 while the latter residue is on helix !5. These form the His-His-carboxylate facial triad conserved among all characterized ROOs (216) and typical of many Fe2+-dependent oxygenases (180). The exogenous ligand in the KshAH37Rv structural model was refined as a single solvent molecule, although residual density was present adjacent to this species (Figure 3.13 C). After refinement experiments with a single molecule of O2 and two fully or partially occupied solvent molecules, it was concluded that the observed density likely represents a mixture of species at partial occupancies, as has been postulated for other ROs (237,238,242). As in the other RO-Os, the Rieske cluster and mononuclear iron center of adjacent subunits of KshAH37Rv interact via a network of hydrogen bonds over the subunit interface. 106  This network includes the conserved residue Asp178, the carboxylate of which bridges the metallocenters by forming hydrogen bonds with each of the N!2 atom of His89 (2.8 Å), and the N"1 atom of His181 (3.1 Å). In other RO-Os this residue has been implicated in catalysis, although its precise role is unclear (233,234). The conformation of this aspartate is similar in all previous RO structures, but in KshAH37Rv the #1 angle differs by ~150° such that the bond between the $ and % carbons points toward the catalytic domain rather than the Rieske domain of the adjacent subunit (Figure 3.14). It is unclear, however, whether the occurrence of this different rotomer in KshAH37Rv has any mechanistic significance.  Substrate docking simulations  Attempts to produce high quality KshA crystals with bound substrate by either cocrystallization or soaking experiments were not successful. In order to investigate the likely spatial implications of KshA CoA thioester substrates, docking simulations were performed using the ligand-free KshAH37Rv structural model (246) and structural models of various steroid substrates. Docking experiments with the substrate ADD were performed by Dr. Igor D’Angelo using the program Autodock (292). The two most highly ranked (as well as biologically compelling) solutions of these experiments show ADD fitting tightly into a pocket adjacent to the active site iron and extending past it into the protein interior (Figure 3.16). The two orientations differ by an approximate 180° rotation about the short axis parallel to the plane of the rings of the steroid molecule such that either ring A or D is oriented toward the interior of the protein. The latter of these two orientations appears to be physiologically relevant as C9 of the steroid structure occupies a position 3.8 Å from the catalytic iron in an orientation conducive to $-hydroxylation. Indeed, its position relative to the catalytic iron is very similar to that of the hydroxylated atom in 2-oxoquinoline, C8, in the OMO86:2-oxoquinoline complex (220). The proposed substrate-binding pocket extends further past the catalytic iron than is observed in structures of other RO-Os, allowing for hydroxylation near the centre of the bulky fused four-ring steroid structure. The residues lining the identified binding cavity  107  (Val176, Gln204, Tyr232, Met238, Asn240, Asn257, Phe301, and Trp308) (Figure 3.16) are conserved in known and predicted steroid-transforming RO-Os, but not in other RO-Os of known structure. The conservation of first shell residues, the tight fit of ADD into the pocket, and the position of the carbon to be hydroxylated with respect to the mononuclear iron strongly suggest that the docking simulation correctly identified the substrate-binding pocket.  Figure 3.16 Two conformations of ADD docked in the active site of KshAH37Rv. ADD and conserved residues of the steroid binding pocket are shown as stick representations. Carbon atoms of the two conformations of ADD are shown in yellow and gray and those of KshA are shown in green. The mononuclear iron is represented as a red sphere. The Connolly surface (293) of the substrate binding pocket is represented as a semitransparent surface.  Docking simulations with the substrates 1,4-BNC and 1,4-BNC-CoA were performed by Dr. Robert Gruninger using the program Autodock Vina (294). Initial docking simulations of 1,4-BNC revealed a positioning of the substrate that is inconsistent with the known outcome of the reaction: C5 and C6 of ring A, rather than C9, were positioned close to the catalytic iron. This orientation had a calculated binding energy of -8.6 kcal mol-1. By contrast, simulations using 1,4-BNC-CoA resulted in a single, well-defined conformation of the steroid moiety with a docking energy of -11.1 kcal mol-1 (Figure 3.17). This conformation 108  was conducive to 9!-hydroxylation. The CoA moiety adopted more than one conformation in different simulation experiments, all of which had similar binding energies. These conformations show significantly different potential binding positions for CoA at the mouth of the KshAH37Rv active site channel. The conformation with the lowest binding energy is shown (Figure 3.17).  Figure 3.17 Docking of 1,4-BNC-CoA in active site of KshAH37Rv. 1,4-BNC-CoA is shown as a ball and stick representation. KshAH37Rv (PDB ID 2ZYL) is shown as a green, semitransparent surface. KshA amino acids within 4 Å of the CoA group of 1,4-BNC-CoA are shown as sticks and labeled. Carbon atoms of amino acid residues and the substrate are colored green and yellow, respectively. Oxygen, phosphorus, and nitrogen atoms are shown in red, orange, and blue, respectively. The catalytic iron is shown as a rust-coloured sphere.  The docked model shows the CoA group bound in a pocket at the mouth of the active site channel. The model also shows several predicted polar contacts including between: the adenine moiety, Ser251 and the carboxylate of Tyr249; the ribose hydroxyl group and Asn244; the ribose phosphate, Ser253, and Asn244; and the phosphate proximal to the steroid group, Asp216, and Tyr246. Most of the above-mentioned residues are not conserved  109  in KshAs from R. erythropolis SQ1, R. rhodochrous DSM43269, and R. jostii RHA1, all of which are predicted to be involved in steroid catabolism (17,295,296). The substrate specificities of the rhodococcal enzymes, however, have not been defined and it is likely that many do not act on CoA thioester substrates.  Comparison with other RO-O alpha subunits Orientation of the active site channel  Although the active site iron occupies a position within the catalytic domain of KshAH37Rv conserved in all known RO-O structures, there are striking differences in the size and orientation of both the substrate binding pocket and the active site channel in the KshAH37Rv structure. In KshAH37Rv, the mouth to the active site channel is found between the C-terminus of helix !5 and the loop connecting strands "15 and "16. This is in stark contrast to the channel of the other !3 ROs, which open to solvent between the second and third strands of the central " sheet (211,220,243,244) ("14 and "15 in KshA), or the channels of the !3"3 ROs which exhibit openings in a similar position near the N- and C-termini of strands "13 and "14, respectively (236-242,245) (Figure 3.18). Concomitant with the different position of the channel mouth, the KshAH37Rv active site channel descends to the active site iron at an angle of almost 90° compared to that of other RO-Os. The KshAH37Rv channel is also significantly longer (~28 Å) than in OMO86 and CARDOJ3 (~18 Å) as measured from the mononuclear iron to the surface of the protein. Two factors contribute to this distinct channel. First, 5-7 additional residues in each of "13 and "14 in KshA effectively occlude the channel mouth observed in other structures. Second, the absence of secondary structural features between "14 and "15 allow for the distinctive opening of KshAH37Rv. Despite these differences, the width of the channel mouth in KshAH37Rv (~9 Å measured at various points along the channel length: eg. C# of Phe182 and C$2 of Leu255) is similar to that of OMO and CARDO (~10 Å). The channel observed in KshAH37Rv allows for access to the steroid-binding pocket along a linear channel, whereas conservation of the  110  channel present in other RO-Os would necessitate significant rotation of the steroid substrate to allow for binding in the active site.  Figure 3.18 Orientation of the active site channel in RO-Os. The catalytic domains of KshAH37Rv (green) and CARDOJ3 (243) (orange) are shown in the same orientation as (A) ribbon diagrams with the semi-transparent Connolly surface displayed, and (B) cross sections of the Connolly surface. The orientation of the active site channel is indicated with a black arrow, the catalytic iron is shown as a dark red sphere. Minimal catalytic domain  Analysis of the secondary structural features of all RO-O structures available in June 2008 revealed that KshAH37Rv exhibits a catalytic domain devoid of specific secondary structural features that are present in the other RO-O structures. Compared to OMO86 and CARDOJ3, KshAH37Rv is missing a !-strand between helix "4 and strand !13 and a ! hairpin 111  and ! strand between !14 and !15 (Figure 3.15); these elements make up the trimerization domain in these structures. DMO exhibits "-helical motifs at these positions. Compared to the structures of "3!3 RO-Os, the catalytic domain of KshAH37Rv is missing helical structures between !14 and !15 and a ! strand which pairs with the N-terminal portion of the protein between "6 and "7. Some "3!3 ROs also exhibit further insertions of secondary structural features. These comparisons suggest that KshAH37Rv exhibits a minimal and possibly archetypical core catalytic domain including elements from strand !12 to helix "7, within which region KshAH37Rv exhibits only features conserved in all known RO-O structures. The structural features absent in KshAH37Rv occur as insertions into this core structure and are specific to structural models belonging to the "3 or "3!3 subfamilies both in their position within the domain and in the identity of the inserted motif. Phylogenetic analysis of RO-O structures  To investigate the distinct features of the KshAH37Rv structural model in the context of RO-O phylogeny, a structure-based sequence alignment was performed using the structures and sequences of KshAH37Rv and nine additional structurally characterized RO-Os. Phylogenetic reconstruction guided by this alignment (Figure 3.19) indicates that KshAH37Rv forms a unique structural subfamily within RO-Os of known structure. This is consistent with the observation that many of the structural features that differ between KshAH37Rv and the other RO-O structural models are conserved within the !3 and !3"3 subfamilies of structural characterized RO-Os.  112  Figure 3.19 Radial phylogram of structurally characterized ROs. The !3 and !3"3 enzymes are clustered at the bottom and top of the tree, respectively. Structure-based sequence alignments were performed using: KshAH37Rv from Mtb H37Rv (PDB accession code 2ZYL); OMO86, 2-oxoquinoline 8-monooxygenase from P. putida 86 (1Z02), CARDOJ3, carbazole 1,9!-dioxygenase from Janthinobacerium sp. strain J3 (1WW9); BPDOB1, biphenyl dioxygenase from Sphingomonas yanoikuyae B1 (2GBW); BPDORHA1 from R. jostii RHA1 (1ULI); NDO9816-4, naphthalene dioxygenase from Pseudomonas sp. strain NCIB 9816-4 (1O7W); NDO12038 from Rhodococcus sp. strain NCIMB 12038 (2B1X); CUMDOIP01, cumene dioxygenase from Pseudomonas fluorescens IP01 (1WQL); NBDOJS765 from Comamonas sp. strain JS765 (2BMO); and RHDOCHY-1, ringhydroxylating dioxygenase from Sphingomonas sp. strain CHY-1 (2CKF).  3.3  RO-O global phylogenetic analysis The phylogenetic analysis of RO-O ! subunit structural models was expanded to  investigate the phylogenetic relationships and diversity of RO-Os of known sequence. A  113  global RO-O phylogenetic analysis was conducted using a structure-guided sequence alignment template and a range of sequences truly representative of all RO-Os available in current databases. 3.3.1  Alignment and tree construction A structure-based template was used to aid in accurate alignment of RO-O sequences  and to provide criteria to ensure that only structurally conserved positions were included during phylogenetic analyses. This is crucial for RO-Os as their low sequence similarity otherwise precludes accurate alignment of the C-terminal portion of these diverse enzymes. Indeed, without a robust structural template, the conserved acidic residue (Asp or Glu) that coordinates the catalytic iron does not align properly for a wide range of RO-Os using several available algorithms. Using the template enabled the identification and proper alignment of such key conserved residues. RO-O sequence diversity likewise results in a high degree of alignment ambiguity outside of the structurally conserved regions. The quality of the multiple sequence alignment has a profound effect on subsequent phylogenetic analysis (261), and removing blocks of ambiguously aligned sequence from the multiple sequence alignment increases the legitimacy of the phylogenetic reconstruction, although resulting in lower bootstrap values (297). The structure-based sequence alignment comprised 159 positions meeting the structurally conserved criteria (Figure 3.20). These are well-distributed over the length of the protein and span a variety of topological features (Figure 3.21). Although the alignment is accurate for structurally conserved sequence segments, the alignments of sequences between such blocks of residues were of extremely low quality. In an attempt to improve this quality, several algorithms were tried. The dissimilarity of amino acid sequence in these regions, however, precluded accurate alignment. These sequence segments were therefore eliminated from the final analysis to include only data representing accurate phylogenetic information.  114  Figure 3.20 Multiple sequence alignment of RO-Os structural models. The multiple sequence alignment has been edited to include structurally conserved positions. The secondary structural components for each sequence segment are shown below the alignment. Metal-coordinating residues are shaded in dark grey. The “bridging” aspartate is shown in a black box. The Asn-Trp-Lys motif conserved in Group I enzymes is indicated with a horizontal bar.  115  Figure 3.21 Topology diagram of the secondary structural elements conserved in all RO-Os of known structure. Elements are labeled sequentially and do not reflect the numbering of any specific protein. Metal-coordinating residues and the conserved bridging aspartate are labeled with asterisks. The positions of structural insertions are marked. Conserved structural features included in the phylogenetic reconstruction are shown in red.  3.3.2  Overall tree topology The consensus tree (Figure 3.22) has a pronounced bilobal structure: Group I contains  41 representative sequences, including all of the best-studied !3"3 RO-Os; and Group II contains 57 representative sequences. Consistent with what was found by Nam et al. (258) and Gray et al. (221), these can be most easily distinguished by an Asn-Trp-Lys motif conserved in group I enzymes (Figure 3.20) and located 11-12 amino acids upstream of the mononuclear iron-binding motif in the N-terminus of the catalytic domain. Both groups contain taxonomically diverse sequences, with plant and animal sequences represented by sequences in both lobes on the tree.  116  Figure 3.22 Consensus tree of 98 RO-O representative sequences from a bootstrap analysis of 100 datasets. Groups I and II are labeled. Proteins with associated structural data are shaded purple. Other proteins that have been functionally characterized in some respect are shaded green. Structurally characterized enzymes are labeled with their respective PDB accession numbers and molecular model (Table 1.2); biochemically characterized enzymes are labeled with their respective enzyme names and bacterial strain (See Appendix C); protein sequences are labeled by GI# where no functional data is available. Sequences associated with recognizable ! subunits are indicated with an asterisk. Bootstrap values are indicated for the most robust groupings of " 4 sequences.  The topology of the tree with respect to groups I and II is reproducible between neighbour-joining and maximum-likelihood methods of calculation and the consensus tree  117  shows a well-defined central node for each group (Figure 3.22). Such consistency notwithstanding, the bootstrap value for the node separating the two groups was only 61. This value was influenced by three outlier representative sequences (1Z02 OMO86, 117628169, and 33865952). When these sequences were removed from the bootstrap analysis, the value for this node increased to 80. 1Z02 OMO86 represents the three CARDO structures in addition to that for OMO86. 3.3.3  RO-O sequence diversity Inspection of the RO-O global phylogenetic tree reveals a degree of RO-O diversity  that has been unappreciated to this point. This oversight is illustrated by the fact that the beststudied RO-Os are concentrated in very small sections of the tree. For example, biphenyl dioxygenase of R. jostii RHA1 (BPDORHA1; 1ULI) represents all ten !3"3 RO-O structural models on the final tree (Figure 3.22). This finding implies that the sum of research into these systems has only explored a tiny portion of the RO-O functional landscape. Moreover, all of the sequences in three out of four groups of the Nam classification scheme (258) are also represented in the same three-sequence branch containing BPDORHA1. Sequences associated with structures of !3 proteins and sequences from Nam group 1 are more widely spread over the tree. This distribution is consistent with group 1 having the lowest intra-group sequence identity scores in the Nam classification system (258). The narrow view of ROs that has been represented previous to this work is further illustrated by the distribution of associated " subunits across the tree topology. Whereas all three previous classification schemes are dominated by RO groups with " subunits, our analysis reveals only five of 98 representative sequences are encoded by genes operonically associated (either upstream or downstream) with genes encoding a recognizable " subunit; four of these are in the small group represented by BPDORHA1. This is not an exhaustive list, however, as gene information was not available for about one quarter of the sequences. Also, given the high sequence variability of the RO-Os, " subunits may not be recognizable based on primary structure alone. Similarly, the RO-O diversity revealed in this work indicates that  118  additional multimeric arrangements are possible, such as the !6 structure suggested for phthalate dioxygenase (230). Distribution of the structurally characterized proteins across the consensus tree indicates that we likely have a good understanding of the core RO-O structural elements. Given this distribution, the conservation of core structural features between all RO-O structural models likely indicates that the core components defined in Figure 3.21 are common to all RO-Os. Conservation of the core fold with specific structural insertions or modifications allows for accommodation of a wide range of substrates using a common catalytic scaffold and mechanism. The different orientation of the KshAH37Rv active site channel with respect to the other available structural models demonstrates the plasticity of RO-O protein architecture. The RO-O tree exhibits notably low bootstrap values for clans within the two main groups. These values are also indicative of high diversity between the representative sequences. Although small robust clusters of 2-4 representative sequences are apparent throughout the tree, no further relationships can be inferred between members of the same group. The exception is a more robust clan in Group II that includes the bacterial KshAH37Rv and the eukaryotic Neverlandmelanogaster. The wide RO-O sequence diversity we uncovered indicates why previous studies relying on standard BLAST searches and previously studied enzymes were limited to the very narrow phylogenetic space surrounding these starting points. A standard BLAST search identifies a limited number of sequences related by any range of residues within the sequence, while the domain-centric approach used for sequence collection in this work uses specific motifs indicative of a particular protein domain to return all protein sequences predicted to incorporate that domain. 3.3.4  Two distinct RO-O groups To investigate evolutionary determinants of the dichotomous RO-O tree topology we  conducted individual phylogenetic analyses of the Rieske and catalytic domains (Figure 3.23). We separated the two domains for each sequence between !2 and "9 (Figure 3.21) and  119  performed a new structure-based sequence alignment on each domain independently. In the absence of the restraints of the second domain, 87 and 96 residues aligned in structurally conserved positions for the Rieske and catalytic domains, respectively. As more residues were included in each individual domain analysis, there was increased information available for these two phylogenetic reconstructions.  120  Figure 3.23 Consensus trees of the Rieske (A) and catalytic (B) domains of 98 representative RO-O sequences. Proteins with associated structural data are shaded purple. Other proteins that have been functionally characterized in some respect are shaded green. Protein sequences are labeled by GI# where no functional data is available. Sequences associated with recognizable ! subunits are indicated with an asterisk.  121  The Rieske domain tree topology is similar to that for the full-length RO-O alignment while the catalytic domain consensus tree is significantly different and exhibits a single central node. The tree resulting from each single-domain analysis exhibits very similar clans as the analysis for the full-length protein, albeit with significantly lower bootstrap values. The two groups apparent in the Rieske domain tree are consistent with structural features at Insertion 1 of the RO-O structural models (Figure 3.21); the insertions at this point in the polypeptide differ between the structures of Group I and II enzymes (Figure 3.24). It should be noted that the sequence segments associated with these insertions are not included during the calculation of the tree as they are not structurally homologous between the two groups.  Figure 3.24 Topology diagram of RO-O Rieske domains. Topology diagrams of Group I and Group II Rieske domains are shown with ! sheets and " helixes represented by arrows and cylinders, respectively. Metal coordinating residues are labeled with asterisks. Secondary structural features are not shown to scale. The major structural insertion present in group I enzymes is highlighted in orange.  122  3.3.5  Origins of the RO-O catalytic domain A search of similar protein structures using the Dali database (279) revealed that RO-  O catalytic domains most closely resemble the Bet_v1-like family of proteins. These are proteins characterized by a large hydrophobic binding pocket capable of binding bulky ligands (298). In keeping with what we observe for RO-O catalytic domains, the members of this family exhibit low sequence similarity but a common polypeptide fold (298). The catalytic domains of most RO-O structural models are more similar to several other RO-Os (often all the other RO-Os) than to other members of the Bet_v1-like family. Interestingly, this is not true for the catalytic domain of KshAH37Rv. Excluding DMO, this domain is most similar to several other Bet_v1-like proteins than to other RO-Os. Although not conclusive, this suggests that the domain in KshAH37Rv may more closely resemble the RO-O ancestral catalytic domain than do the other structurally characterized RO-Os. It is clear from the available sequence data that the vast majority of proteins with ROO catalytic domains exhibit the canonical domain architecture featuring an N-terminal Rieske domain. I did find one group of genes, however, that is conserved in several species of fungi from subkingdom Dikaryia and that encode an RO-O-like catalytic domain but no recognizable Rieske domain. Interestingly, many of these genes are operonically associated with genes encoding RO-O-like Rieske domains, suggesting that these may be RO-Os exhibiting the two canonical domains on different polypeptides. These hypothetical proteins share a distinctive Glu-Cys-X-His-Cys-X-X-(X)-His sequence as part of the mononuclear iron-binding motif. This motif is also shared among five Group I representative sequences in the global analysis, including the structurally characterized putative RO-O from Silicibacter sp. TM1040 (PDB 3N0Q). This group of proteins, which lacks a fused Rieske domain, could be just one example of several such atypical RO-Os. Such proteins are likely hard to detect in the sequence database due to the imperfect ability of the current algorithms to reliably recognize genuine RO-O catalytic domains.  123  3.3.6  RO-O classification scheme The results of the RO-O global phylogenetic analysis clearly indicate that previously  proposed classification systems do not sufficiently describe the diversity of RO-O sequence, function, and taxonomic distribution. Based on this analysis, it is unlikely to be feasible to draw functional inferences from broad RO-O categorizations. Indeed, with the exception of the observed two main groups and one other large robust clan, broad categories of RO-Os do not seem to exist. Instead RO-Os could be categorized in smaller, catalytically relevant groups by a system analogous to the scheme developed for P450 classification (205). This proposal is logical given the similarities of these two enzyme families outlined in Section 1.4 of this thesis. The proposed classification scheme would group RO-Os into families and subfamilies by amino acid sequence identities and assign names for each enzyme based on these designations (Appendix C). Analogous to the P450 naming system, each enzyme would be named by the abbreviation “Roo” followed by a number denoting the family (40% amino acid sequence identity), a letter denoting the subfamily (55% sequence identity), and a second number identifying the specific enzyme. At this level of amino acid identity, ambiguous alignment of the catalytic domain drops off significantly, eliminating the need for comprehensive structure-based alignment and allowing simple pair-wise alignments of fulllength sequences. The sequences of 121 RO-Os specifically identified and named in the literature were collected for classification. Inconsistencies in RO terminology, however, make it possible that some previously studied ROs were not identified in this search. These sequences were grouped into families and subfamilies as described above and named according to the first publication date for each enzyme (Appendix C). This analysis resulted in 40 families of ROOs thus far studied. As in the P450s, most families contain enzymes with the same catalytic function, based on the available data. Some families or subfamilies, however, contain enzymes of closely related function. For example, family Roo1A incorporates both BPDOLB400 (Roo1A5) and TDOF1 of P. putida F1 (Roo1A2). 124  Not all families are shown on the representative consensus trees in Figures 3.22 and 3.23 as many are represented by a sequence belonging to a related family. The number of families represented by each branch varies depending on the segments of sequence that are not structurally conserved and consequently not included in the phylogenetic reconstruction. In general, however, the full-length sequences represented in these trees are <30% identical, a number which is well below the 40% cutoff for designation as part of the same family. Each sequence on the tree that is represented by a GI # therefore represents one or more asyet unexplored RO-O families.  125  4 4.1  Discussion Initiation of cholesterol side chain degradation by Cyp125 The biochemical characterization of Cyp125A1 presented in this thesis establishes  that this enzyme catalyzes the hydroxylation of the side chain terminal methyl groups of both cholesterol and 4-cholesten-3-one. These direct observations of substrate consumption and product formation are supplemented with data reflecting the in vivo role of this enzyme in bacterial cholesterol degradation. While wild-type R. jostii RHA1 shows robust growth on cholesterol (21), a !cyp125 strain exhibited no growth on cholesterol and an inability to transform this steroid (264). From this data it was concluded that cholesterol degradation is initiated by transformation of the side chain by Cyp125A14P in this organism. A !cyp125 knockout of M. bovis BCG was likewise unable to grow on cholesterol (266). This strain was nonetheless found to consume cholesterol and quantitatively convert it to 4-cholesten-3-one. This finding suggests that while steroid side chain degradation was blocked, this process is not required for the initial oxidation of the steroid nucleus (266). A !cyp125 mutant of Mtb CDC1551 incubated with cholesterol also accumulated 4-cholesten-3-one (299). Additionally, this strain exhibited less PDIM during growth on cholesterol, an observation attributed to the inability to incorporate carbon atoms derived from the cholesterol side chain into these lipids (299). All of these findings point to the conclusion that Cyp125A1 has a significant physiological role in initiating cholesterol side chain degradation through oxygenation of the terminal methyl group in mycobacteria. The cumulative findings also indicate that degradation of the steroid ring structure, beyond the action of 3"-HSD, is not possible without some degree of side chain degradation. The human P450 Cyp27A1 (300) and DAF-9 of C. elegans (301) both also catalyze C26 hydroxylation of cholesterol but exhibit low sequence similarity to Cyp125A1. This reaction is therefore a likely example of convergent evolution between bacterial and animal P450s (62).  126  4.1.1  Multiple P450 C26-hydroxylases Cyp125A1 has a physiological role in initiating cholesterol side chain degradation.  Compensatory activity, however, has been observed in !cyp125 knockouts of Mtb H37Rv. The Mtb H37Rv knockout strain lacking the entire cyp125-encoding igr operon was found to retain the ability to transform cholesterol and to incorporate C14 from the cholesterol side chain into its cellular lipids (56). Likewise, a more specific !cyp125 knockout of Mtb H37Rv showed no growth defect on cholesterol compared to wild type bacteria (266). These observations suggest that another enzyme capable of steroid 26-hydroxylation was present in the cells lacking Cyp125A1. Contrasting these results, the !cyp125 knockout of M. bovis BCG was unable to grow on cholesterol and accumulated 4-cholesten-3-one in cholesterolcontaining cultures. This finding suggests a difference in monooxygenase activity between the two mycobacterial species. Bioinformatic analyses revealed eight genes predicted to encode oxidoreductases in Mtb H37Rv that do not occur in M. bovis BCG. Three of these genes are predicted to encode monooxygenases: Rv1265c (cyp130), Rv3121 (cyp141), and Rv3618 (a predicted flavin-dependent monooxygenase) (266). It was hypothesized that one of these proteins could exhibit a facultative ability to compensate for the loss of Cyp125A1 activity, therefore explaining the differences in phenotype between the two strains. When tested, the former two enzymes failed to exhibit detectable steroid 26-hydroxylase activity (299). Differences in the ability to compensate for !cyp125 mutation were also found between strains of Mtb. A study published by Ouellet and colleagues demonstrated that while Mtb H37Rv was able to use cholesterol in the absence of Cyp125A1, a !cyp125 mutant of Mtb CDC1551 was wholly deficient in this capacity (299). In these experiments, the !cyp125 strain was also unable to incorporate cholesterol side-chain-derived carbon into PDIM and, like the M. bovis BCG !cyp125 knockout strain, was found to accumulate 4-cholesten-3-one in cultures containing cholesterol. These results suggested that the genes responsible for additional steroid 26-hydroxylase activities were not those that differed between Mtb and M. bovis, but rather reflected strain-specific genetic differences between different members of the Mtb complex. Inspection of the two Mtb genome sequences by the same authors revealed 127  a 639-bp deletion in CDC1551 upstream of the KstR-regulated cyp142 gene encompassing the promoter and N-terminal amino acids of the predicted protein product. This mutation was predicted to result in an absence of Cyp142A1 activity in Mtb CDC1551 (62). Likewise, the cyp142 gene of M. bovis BCG was determined to be defective (302), explaining the phenotype of the !cyp125 mutant of this strain. Additionally, the P450 Cyp124A1 exhibits significant structural homology and sequence identity to Cyp125A1 and was therefore also tested for its ability to compensate for !cyp125 mutation. Both Cyp142A1 and Cyp124A1 are capable of steroid 26-hydroxylation (62). Only the former enzyme, however, can support bacterial growth on cholesterol in the absence of Cyp125A1 activity (62). Biochemical studies indicate that Cyp125A1 and Cyp142A1 catalyze C26 hydroxylation of cholesterol and 4-cholesten-3-one with similar catalytic efficiencies, while Cyp124A1 exhibits significantly lower catalytic activity for both compounds (Table 4.1) (62). Interestingly, while the steady-state kinetic parameters for Cyp125A1 and Cyp142A1 are similar, the former enzyme produces the steroid (25S)-26hydroxy steroid derivatives while the other two P450s yields products of 25R stereochemistry (62). Genetic and protein expression studies presented in the same work indicate that Cyp124A1 is not expressed at detectable levels in cholesterol-grown cultures of either Mtb strain, while Cyp142A1 was detectable only in Mtb H37Rv cells. These data suggest that Cyp142A1 compensates for the loss of Cyp125A1 activity in gene knockouts of this strain and that other cholesterol catabolic genes are capable of metabolizing compounds with C25 carbons of both R- and S- configurations.  128  Table 4.1  Apparent binding affinities and steady-state kinetic and binding affinities of Cyp124A1,  Cyp125A1, and Cyp142A1 for cholesterol and 4-cholesten-3-onea Enzyme  KD (nM)  KM (µM)  kcat (min-1)  kcat/ KM (µM-1min-1)  12 (5)b  1.5 (0.2)  0.129  Cholesterol Cyp124A1 Cyp125A1  110 (60)  11 (4)  28 (3)  2.6  Cyp142A1  18 (5)  8 (2)  17 (1)  2.2  Cyp124A1  1060 (30)  21 (2)  12 (1)  0.56  Cyp125A1  1200 (100)  21 (2)  175 (8)  8.4  Cyp142A1  110 (20)  12 (2)  84 (5)  7.1  4-Cholesten-3-one  a  Table adapted from Johnston et al. 2010 (62), ! the American Society for Biochemistry and Molecular Biology, Inc., adapted by permission. b Values in parentheses represent standard error  4.1.2  Formation of the C26 carboxylic acid In work published by Ouellet et al. (299) and Johnston et al. (62) Cyp125A1,  Cyp142A1, and Cyp124A1 were all demonstrated to catalyze the transformation of cholesterol and 4-cholesten-3-one to 26-carboxylic acids. In each case, the corresponding alcohol was formed first and two further oxidation reactions subsequently transformed these compounds to their respective aldehyde and carboxylic acid derivatives. Catalysis of sequential steroid oxidations by the same P450 enzyme has several precedents in the literature and can proceed with or without dissociation of intermediate compounds (303). In the experiments presented in this thesis, Cyp125A1 catalyzed the transformation of cholesterol and 4-cholesten-3-one to their respective C26-hydroxy compounds in a 1:1 ratio with no evidence of further oxidation products. Likewise, Cyp125A1 activity reconstituted in vitro using an E. coli-derived ETC produced 26-hydroxycholesterol as the only detectable product (304). One explanation for these discrepancies is differences in the ETC used in the in vitro experiments. The experiments on Cyp125A1 described in this thesis were performed using a two-component enzyme system incorporating the Mtb reductase KshB. Experiments with Cyp125A1, Cyp142A1, and Cyp124A1 in which the carboxylic acid products were observed were performed with spinach ferredoxin and ferredoxin reductase (62,299). The ETC can have a significant impact on catalytic efficiency, as demonstrated by the ~50%  129  reduction in KshAB activity observed upon addition of a poly-histidine tag to the reductase component. Similarly, lower Cyp125A1 reaction rates were also observed using ETCs from other bacterial ROs. The rate of formation of 26-hydroxy-4-cholesten-3-one from 4cholesten-3-one by Cyp125A1 also differs between the work presented in this thesis and that presented by Johnston et al. (62). While the latter experiments using the spinach-derived ETC revealed a turnover number of 175 ± 8 min-1 (62), turnover numbers around 0.5 min-1 were measured for reactions using KshB as a reductase (266). Together, these findings support the theory that the Cyp125A1 reaction rate, and potentially the ability or efficiency of the enzyme to catalyze further oxidation, is influenced by the ETC. Another possible explanation for differences in steroid side chain oxidation products under various experimental conditions is corresponding differences in the steroid solubilization agent present. In experiments with Cyp125A1, Ouellet et al. only observed further oxidation of 26-hydroxy-4-cholesten-3-one when the substrate was prepared with Tween-20 but not when BCD was used (299). The results presented in this thesis were obtained using BCD as a solubilization agent. In contrast to these results, Johnston et al. observed formation of the C26-alcohol, -aldehyde, and -carboxylic acid derivatives of both cholesterol and 4-cholesten-3-one in the presence of BCD but performed steady state kinetic experiments under conditions “optimized for the formation of the respective 26-hydroxy derivatives” that eliminated further oxidation products. Thus, cumulative evidence suggests that subtle changes in reaction conditions affect the enzyme’s ability to form the aldehyde and carboxylic acid products. 4.1.3  Physiological reductase Several findings are consistent with the hypothesis that KshB is the physiological  reductase for Cyp125A1. While the Mtb genome encodes 20 P450 enzymes, many fewer ETC components have been identified in this bacterium. Such a discrepancy in numbers necessitates that the same ETC components be used by more than one terminal oxygenase. In this work, the ability of KshB to provide reducing equivalents for the Cyp125A1-catalyzed reaction was demonstrated in vitro. This finding is consistent with genetic evidence that  130  KshB is the cognate reductase for Cyp125 in R. erythropolis (85). Van der Geize and coworkers demonstrated that sterol side chain degradation was abolished in the !kshB knockout strain R. erythropolis SQ4, strongly suggesting that KshB is involved as a reductase in this process. Additionally, a !kshB mutant of Mtb exhibited phenotypes distinct from those exhibited by the !kshA mutant, including a thickened cell wall and a deficiency in penta-acylated trehalose biosynthesis (31). These results are also consistent with KshB acting as a reductase for more than one oxygenase. A physiological role for KshB in reduction of both KshA and Cyp125A1 would potentially be the first described example of an oxidoreductase exhibiting activity toward two classes of oxygenase in vivo. The biological capacity of KshB to support Cyp125A1 hydroxylation activity does not eliminate the possibility that another redox partner performs this function in vivo. A review published in 2010 by Ouellet and coworkers (305) identified nine demonstrated or predicted P450 ETC components and postulated that combinations of these proteins would support the activity of the full complement of Mtb P450s. Of these proteins, FdxB (Rv3554) and FdxD (Rv3503c) are encoded in the cholesterol degradation gene cluster (Figure 1.4). In this publication, the former protein is described as a predicted fused ferredoxin-ferredoxin reductase, while the latter protein is a predicted ferredoxin. In work presented here, FdxD did not increase Cyp125A1 function when KshB was used as a reductase. While it is possible that FdxB could comprise all or part of the Cyp125A1 ETC, the N-terminal domain is more closely related to the di-iron oxygen-dependent fatty acid desaturases and "-carotene hydroxylases and exhibits the eight conserved histidines presumed to be the metallo-ligands in these enzymes. This observation suggests that this protein may represent a fused oxygenase-reductase enzyme and would thus be unlikely to be involved in Cyp125A1 activity. It is also possible that the physiological ETC for Cyp125A1 is encoded outside of the cholesterol degradation gene cluster, as P450s are not always encoded at physically proximal loci and several genes in the KstR regulon are located elsewhere in the genome.  131  4.1.4  Structure of Cyp125A1 and Cyp124A1 Crystal structures of Cyp125A1 in ligand-free (304), 4-AD-bound (304), and 4-  cholesten-3-one-bound (299) states have been determined to resolutions of 1.4 Å, 2.0 Å, and 1.58 Å, respectively. These molecular models reveal a roughly rectangular active site channel mouth leading to a deep cavity lined with hydrophobic residues (304). The rectangular shape of the outer channel allows for tight binding to the roughly planar steroid ring structure (299). The narrowing of this cavity toward the catalytic heme group presumably contributes to the specificity of the enzyme in side chain oxidation as it effectively prevents the steroid nucleus of 4-AD from approaching the heme iron (304). The tight fit of the steroid substrate within the active site channel has also been proposed to be responsible for orienting the terminal methyl group in a favourable position and the tertiary C25 in a disfavourable position for hydroxylation relative to the catalytic iron (299). These steric constraints are proposed to be instrumental in achieving selective hydroxylation of the primary carbon, a reaction that is energetically less favourable (299,306).  Figure 4.1 4-Cholesten-3-one bound in the active site channel of Cyp125A1. View of the cholests-4-en-3-onebound Cyp125A1 clipped by a plane through the substrate binding tunnel orthogonal (A) or parallel (B) to the plane of the tetracyclic steroid nucleus. The heme prosthetic group is shown in orange. The residues preventing the ligand from sliding towards the heme are labeled. Figure taken from Ouellet et al. (2010) (299), ! Blackwell Publishing Ltd., 2010, with permission.  132  4.2  Order of reactions in cholesterol degradation Our knowledge of the Mtb cholesterol degradation pathway comprises a list of  catabolic “limits” defining the biochemical capacities of specific enzymes, rather than their physiological roles. Both the in vitro nature of experiments exploring this pathway and a level of facultative promiscuity for some enzymes contribute to the limitations inherent in our understanding of the pathway. Thus, it is established that 3!-HSD, but not KstD, can transform cholesterol before Cyp125A1 catalyzes the first side chain reaction and that Cyp125A1 is capable of transforming cholesterol or 4-cholesten-3-one (78,266,299). Fad28 and Fad29, which perform a dehydrogenation contributing to cleavage of the final three side chain-derived carbons to form the 17-keto group, can also use substrates with both an intact steroid ring structure and with cleaved rings A and B (64). Likewise, side chain degradation can proceed to its conclusion before the action of either KstD or KshAB (64,65), but both of these enzymes can transform intermediates of side chain degradation. Similarly, the ringdegrading enzymes HsaAB, HsaC, and HsaD can all transform 17-keto steroid metabolites (54,72,89), but can also catalyze their respective reactions using substrates with partially degraded side chains (64). 4-Cholesten-3-one accumulates in cultures incapable of side chain transformation, but 3!-HSD, the enzyme responsible for forming this compound from cholesterol, showed 3-fold lower activity towards cholesterol than 5-pregnene-3!-ol-20-one, a sterol with a shortened side chain (78). These findings all demonstrate the catalytic flexibility of enzymes involved in both side chain and ring degradation that complicates the elucidation of their true physiological substrates. 4.2.1  Pathway elucidation strategies The current view of the metabolites occurring in the cholesterol degradation pathway  derives primarily from in vitro or gene knockout studies. While these experiments provide valuable information about the biochemical capacities of the relevant enzymes, neither approach faithfully recapitulates the physiological conditions for Mtb in the human host. One of the major limitations of studies using gene knockouts is that genetic manipulation of the bacterium imposes a non-physiological condition in any experimental setup. In a simple case,  133  a mutationally-blocked metabolic pathway may result in the detection of metabolites that would normally never occur in the cell. This phenomenon occurs when compounds are effectively forced as far down a pathway as biochemically possible due to accumulation of up-stream metabolites at the point of metabolic disruption. Such results are especially common in processes, like cholesterol degradation, that involve reactions at more than one position in the metabolite. More complicated situations can also arise when the knockout results in significant metabolic stress due to considerations such as metabolite toxicity or redox imbalance. In these cases it can be even more difficult to tease out the actual biochemical reaction and metabolic consequences directly related to the mutated gene. It is also important to consider the intrinsically non-physiological nature of experiments involving wild type Mtb grown in the lab or in vitro studies of purified components. The metabolic state of Mtb that is grown under laboratory conditions cannot be expected to be representative of a naturally occurring physiological state. The discrepancy in growth conditions may explain why 4-AD and ADD were observed to accumulate in Mtb culture supernatants (64,65) while results presented in section 3.2.4 of this thesis suggest that the KstD and KshAB reactions cleaving ring B likely take place before formation of the C17ketone. Likewise, the experimental conditions used for in vitro assays can have a profound effect on enzyme activities. Such an effect is exemplified by differences between results presented here and those measured by others for Cyp125A1 reaction rates (62), and KstD substrate specificities (87). In the latter case, no reaction was discernable for KstD using 4AD as a substrate,while a reaction rate was measurable using methods described here. Indeed, while substrate specificities of purified enzymes are suggestive of physiologically relevant reactions, the reaction conditions used for these assays are far removed from those in the bacterial cytoplasm. Full understanding of the cholesterol catabolites in Mtb therefore requires us to probe these compounds in Mtb growing under conditions representative of human infection.  134  4.2.2  Insights from KshAB substrate specificity The results presented in section 3.2.4 of this thesis suggest that steroid CoA thioesters  are the likely physiological substrates for Mtb KshAB. The steady-state kinetic analyses demonstrate that both KstD and KshAB transform the CoA thioesters more efficiently than 17-keto steroids, the previously proposed physiological substrates. This finding is especially dramatic in KshAB’s reactivity with oxygen. The kcat/KMO2 in the presence of 1,4-BNC-CoA is an order of magnitude greater than that measured in the presence of ADD. Perhaps more significantly, the KMO2 of 90 ± 10 !M is similar to that of other oxygenases in the cholesterol degradation pathway (54,72) as well as to the O2 concentration in the lung tissues of healthy rabbits (123). This KMO2 can be compared to the value in excess of 1.2 mM measured in the presence of ADD. The apparent specificity constants for steroid metabolites were measured in air-equilibrated buffer, so the high KMO2 for the keto-substituted steroid means O2 was well below saturating levels. In contrast, the lower KMO2 in the presence of a CoA thioester substrate means that in the presence of the latter steroid, KshAB is closer to being saturated with O2. Thus, the increased reactivity of KshAB for O2 in the presence of the CoA thioester substrates could account for the greater apparent substrate specificity compared to the 17keto substrates. The contribution of the CoA moiety to the lower KMO2 was not directly explored in the experiments described here, but reasonable hypotheses can be made based on current understanding of the RO mechanism that is outlined in section The catalytic mechanism, as we understand it, involves binding of the substrate followed by reduction of the Rieske cluster and resultant shifting in the position of the catalytic iron to enable O2 binding (216,220,250). It is therefore possible that the CoA moiety aids in substrate binding in a mode conducive to communication between the Rieske centre and the enzyme active site. If the 17-keto substrates are able to bind the enzyme in a mode that does not enable efficient communication between these two metallocentres, it could result in lower reactivity with O2 even at seemingly saturating steroid concentrations. The greater activity of KshAB with CoA-bearing thioesters could also be the result of increased reactivity of these substrates with the active oxidant in the enzyme active site. This hypothesis, however, does not explain  135  the significantly higher KMO2 for the 17-keto substrates. Indeed, reaction rates with ADD measured at ~ 1.2 mM oxygen are similar to the kcat estimated for reactions with 1,4-BNCCoA (2.5-2.7 sec-1). This observation indicates that, once formed, the active oxidant can react with either class of substrate at comparable rates, or that another step is rate-limiting in these reactions. The steady-state kinetic parameters reveal that the turnover numbers and KM values for 4-BNC-CoA and 1,4-BNC-CoA are significantly different, but that the specificity constants for the two compounds differ by less than two fold. The kinetic differences between the !4 and !1,4 thioesters mirror those for 4-AD and ADD. This trend of greater kcat and KM for !1,4 compounds than for their !4 counterparts is also evident in the comparison of 4-BNC and 1,4-BNC, notwithstanding the relatively low specificity constant for the latter steroid. The mechanistic determinants for KshAB’s low turnover number for 1,4-BNC compared to other !1,4 substrates are not clear from the available data. The relatively small difference in KshAB substrate preference between candidate physiologically relevant !4 and !1,4 compounds has implications for the order of the KshAB- and KstD-catalyzed reactions (Figure 1.8). Given that KshAB does not exhibit a strong specificity for one class of compound over the other, the order of these two reactions may be determined by the substrate specificity of KstD for compounds with and without a !-hydroxyl group at C9. This study suggests 4-BNC-CoA and 1,4-BNC-CoA as candidate physiological substrates for KshAB, but does not address the possibility of other CoA thioester substrates. The molecular docking experiment of 1,4-BNC-CoA in the KshAH37Rv (2ZYL) active site demonstrates that there is room for this large molecule in the enzyme active site channel, and that favorable electrostatic interactions with the CoA moiety may exist and could contribute to KshA substrate specificity. Nonetheless, KshAB inhibition by CoASH was not observed, suggesting that the significantly greater apparent specificity constants for the CoA thioester substrates is not solely attributable to KshA affinity for the CoA moiety. The side chain degradation pathway proposed for Mtb (64,65) includes CoA thioester steroids with 8-carbon and 5-carbon side chains in addition to the 3-carbon isopropionyl-CoA thioester steroids tested here. Additionally, compounds with oxidized and hydrated side chains are produced 136  during each round of !-oxidation. Further studies are needed to determine whether KshAB preferentially hydroxylates a particular thioester. 4.2.3  Implications of KshAB substrate specificity A major implication of CoA thioesters as physiological substrates for KshAB is  concurrent cholesterol side chain and ring degradation in Mtb. All previous biochemical experiments with purified Mtb cholesterol ring-degrading enzymes have used compounds representing end products of side chain degradation (54,72,89,246). The CoA thioesters examined in this study represent hypothetical intermediates in cholesterol side chain degradation. The kintic results therefore suggest that the true physiological substrates for some ring-degrading enzymes may also be side chain degradation intermediates. This notion counters the paradigm that cholesterol side chain degradation in Mtb is completed before breakdown of the ring structure. The order of Mtb cholesterol catabolism has implications for the probable steroid molecules produced during cholesterol degradation and for the design of small molecule inhibitors for the cholesterol ring-degrading enzymes. KshAB has been implicated in the management of potentially immunomodulatory steroids that have been suggested to play a role in early infection (31). The results presented in this thesis broaden the range of potential compounds in this category. Additionally, these results expand the potential substrate range and the industrial potential of steroid ring-degrading enzymes from other species. To this end, protein engineering efforts in KshA enzymes from R. rhodocrous involving substitution of significant segments of protein have resulted in active enzymes that exhibit altered substrate specificities (307). It remains unclear what degree of side chain degradation is completed before the activity of each ring-degrading enzyme in vivo, or whether physiologically relevant parallel pathways exist. This study, in combination with other investigations into KshAB reactivity (31), strongly suggests the existence of more than one functional pathway in Mtb.  137  4.2.4  Differences between species and strains For many steps of the Mtb cholesterol degradation pathway the interplay between ring  and side chain degradation seems to differ between bacterial strains. For example, C3 oxidation and ring A isomerization to the !4 form is able to precede side chain transformation in Mycobacterium smegmatis mc2155 (74), Mtb CDC1551 (299) and M. bovis BCG (266). Likewise a R. rhodocrous !cyp125 knockout strain can perform ring oxidation (264). In contrast R. jostii RHA1 requires oxygenation of the side chain terminal methyl group for initial ring oxidation to take place (264) but is able to grow on 4-cholesten-3-one in the absence of Cyp125A14P. This finding indicates either that R. jostii RHA1 exhibits distinct substrate specificities of the ring degrading enzymes, compared to mycobacterial species, that enable bacterial growth due to steroid core degradation in the absence of side chain degradation or that this strain incorporates multiple pathways for steroid degradation including a C26-hydoxylase capable of transforming 4-cholesten-3-one (264). The relative substrate specificities of KshAB for !4 and !1,4 compounds also appear to differ between species. For example, KshAB from M. smegmatis exhibits a much lower capacity to transform !4 steroids than Mtb KshAB, based on metabolite accumulation studies in !kstD knockout cultures (84,159). Similarly, four purified KshABs from R. rhodocrous DSM43269 exhibit different relative substrate specificities for a variety of steroids, including those of !4 and !1,4 configurations (296,308). The potential for differences in steroid metabolism in both knockout and wild type strains of different species or genera need to be considered when interpreting the results of experiments performed in one strain in the context of the physiology of a distinct strain. 4.3  Role of cholesterol degradation in Mtb physiology and pathogenicity Several studies involving screens of genetic knockouts (22,48,52), transcriptomic  investigations (49-51), and targeted genetic mutations (30,31,56,65,72,90,177) have implicated genes in the cholesterol degradation gene cluster as being important for Mtb virulence or survival under conditions designed to model human infection (18). While this body of evidence continues to build, interpreting the exact role and importance of cholesterol 138  catabolism in Mtb physiology is far from straight forward. Firstly, while targeted knockouts of several genes in the KstR regulon have indicated that their products play a significant role in Mtb pathogenesis or intracellular survival (30,31,65,72), removal of some enzymes in the pathway was found to have no deleterious effect (178)(P. Converse, R. Kalscheuer, L. Ly, and W. Jacobs, personal communication). Secondly, there are three general roles proposed for Mtb cholesterol catabolism and it remains unclear which are most important to Mtb physiology and pathogenesis. These three roles are: catabolism of the steroid nucleus as a source of carbon and energy, use of breakdown products as biosynthetic precursors for physiologically important cell components, and modification of the cholesterol molecule for use in manipulation of the host physiology. In addition to these catabolically-associated functions, the ability of Mtb to sense cholesterol or cholesterol metabolites has also been proposed as a mechanism for detecting the intracellular environment of the host macrophage (24).  139  Table 4.2  Phenotypes of targeted Mtb gene knockoutsa,b Gene product name  Role  1106c  3!-HSD  ring degradation  5301c  Yrb4A  import  3518c  Cyp142A1  side chain degradation  3526  KshA  3540c-3545c  (igr operon)  Rv number  3545c  Cyp125A1  3546  FadA5  3568  HsaC  3571  KshB  c  Attenuation in macrophage infection model(s)  Attenuation in animal infection model(s)d  Attentuation of in vitro growth on cholesterol  Comments  ! (178)  Attenuation in macrophage and guinea pig infection models tested and not observed  ! (177)  ! (177)  ! (56) ! (62)  ring degradation  ! (31)  ! (31)  ! (31)  side chain degradation  ! (30)  ! (30)  ! (30)  side chain degradation side chain degradation ring degradation ring degradation  ! (299)  ! (31)  ! (65) ! (72) ! (31)  Contradictory results in different laboratories, see Section  Phenotype observed for knockout in Mtb CDC1551 but not H37Rv (266). Part of igr operon.  ! (65) ! (72) ! (31)  a  Unless otherwise stated, blank columns indicate experiment was not performed Primary references most directly addressing observed phenotype are listed c The predicted role with the most cumulative evidence is listed d Encompasses various mouse and guinea pig infection models b  Given the variety of potential roles for cholesterol catabolism, it is possible that only specific segments of the cholesterol degradation pathway contribute significantly to Mtb intracellular survival and/or virulence. This theory is supported by the observation that knockouts of ring-degrading genes earlier in the pathway tend to elicit stronger phenotypes than downstream enzymes (48). It follows that while Mtb can grow on cholesterol as a carbon source (30), partial cholesterol degradation may be sufficient to support this nutritional requirement. Similarly, the generation of degradation intermediates with intact steroid core structures may be crucial for a signaling role while downstream catabolic enzymes are used to capitalize on the available cholesterol carbon that may not be a critical carbon source. A similar argument can be made that the generation of biosynthetic precursors from side chain degradation products is potentially more important to physiology than cholesterol catabolism for nutrition. In this model, complete degradation of cholesterol would  140  be expected to confer a selective advantage, thus being consistent with the conservation of the entire gene cluster, but could include some processes that are not critical to the bacterium. This theory could also help explain how availability and efficient import of cholesterol would increase bacterial survival in some infection models (30) while decreasing efficiency of uptake could have a remediative effect on growth of a mutant with a different and potentially toxic mutation (56). In considering the potential roles of cholesterol in intracellular survival it is potentially telling that this pathway is not conserved in the related mycobacterial intracellular pathogen M. leprae. This observation suggests that while cholesterol degradation in Mtb may have important roles in the specific infection cycle of this pathogen, it is unlikely that this process plays a broadly critical role in intracellular survival. 4.3.1  Stage in life cycle As described in section 1.3.2 of this thesis, the environmental conditions and  physiological state of Mtb are not constant throughout the infection cycle. It logically follows that different biochemical processes are likely important during establishment of infection in a naïve host, survival during the latent stage of infection, and emergence from dormancy with concomitant increase in proliferation. In addition to these broad stages of the infection cycle processes such as formation and stratification of the granuloma, caseation of this structure, and dissemination to form new lesions in the current host all presumably require various adaptations in the bacterial physiological state to enable survival during the dynamic infection process. Such adaptations are hinted at in the temporally-dependent transcriptional adaptation of Mtb during macrophage infection (49). Given the diversity in biologically relevant conditions throughout infection, it is possible that all or part of the cholesterol degradation pathway is important for one or more specific stage of infection or pathological process. Several studies have resulted in hypotheses that cholesterol degradation is important to the bacterium during a specific phase in Mtb infection. Pandey and colleagues concluded from experiments inhibiting cholesterol import that degradation of this steroid is important during the chronic stage of infection (30). Yam and coworkers came to a similar conclusion  141  considering results of infection model studies with a !hsaC mutant, although these authors also noted that Mtb appears to use cholesterol earlier in the infection process (72). Nesbitt and coworkers found cholesterol degradation to be important in chronic infections based on experiments with a !fadA5 mutant in a mouse infection model (65). These authors further concluded that cholesterol catabolism as a source of carbon and energy is not crucial for acute infection. Conversely, Chang and colleagues found no difference in persistence of their !igr mutant, but did observe a lag in the increase of colony-forming units in the lung and spleen in immuno-competent mice (56). This finding potentially indicates a role for cholesterol degradation in early infection (56). While on some level it is perplexing to consider why different components of the cholesterol degradation process appear to be important for seemingly disparate Mtb physiological functions, these contradictory findings may reflect multiple roles for cholesterol degradation during infection. For example, if a signaling role were important for a specific time period in Mtb infection, this does not preclude metabolism of the cholesterol ring structure as a prominent source of nutrition during a separate phase of infection. Another consideration in examining the disparate results of infection model studies of mutant Mtb is the interpretability of the results of a specific model in the context of human disease. As explored in section 1.3.2, while Mtb can cause disease and mortality in common model systems like mice and guinea pigs, the progression of the infection and accompanying environmental conditions inside the host do not necessarily recapitulate those of human infection (112). It is therefore speculative to interpret infection model studies in the context of determining the importance of cholesterol degradation during different stages in human disease. One of the major factors determining the infection stage during which cholesterol degradation might have particular pathological significance is the time scale that the necessary components are available to the bacterium. As mentioned in section 1.2.1, Mtb cholesterol catabolism requires the action of at least four oxygenases. Additionally, oxidation of 26-hydroxy-steroids by Cyp125A1 or Cyp142A1 to form the carboxylic acid derivatives (62) is expected to consume a further two molecules of O2 (303). The hypoxic environment of the granuloma (123) therefore calls into question the significance of cholesterol degradation in bacteria within this environment. This consideration is especially pertinent 142  given the KMO2 measured for KshAB (265), HsaAB (54), and HsaC (72). All of these values are about 50 times greater than the O2 concentration associated with the granuloma (123). While O2 concentration is limited in this environment, the granuloma caseum is a rich source of cholesterol (157). The ubiquitous occurrence of cholesterol in all host cell membranes, however, may negate its relative abundance as a factor in determining the temporal importance of cholesterol degradation during infection. 4.3.2  Side chain degradation The role of side chain degradation in Mtb has primarily been interrogated through  knockouts of fadA5 (65) and the neighbouring igr gene cluster (56). The igr cluster includes the cyp125 gene. The exact contribution of Cyp125A1 to Mtb pathogenesis has been complicated by both the presence of compensatory activities in Mtb H37Rv and by the concurrent knockout of the other igr genes in some studies. The !cyp125 mutation in Mtb CDC1551 completely abrogates mycobacterial side chain transformation through an inability to initiate this process, but this strain has not been used for infection model studies. The products of two other igr genes, FadE28 and FadE29, have been demonstrated to catalyze a reaction in cholesterol side chain degradation and the remaining three genes are proposed to work in concert with FadE28 and FadE29 to remove the final side chain-derived propionylCoA molecule and form a 17-keto product (64). A knockout of the igr operon resulted in a growth defect in the early infection stages of a mouse model but eventual bacterial growth comparable to wild type levels (56). Further, this growth defect was eliminated by reducing cholesterol uptake via inactivation of the Mce4 uptake system, suggesting that the in vivo growth defect is a result of metabolite toxicity. These results suggest that side chain degradation is not essential for growth in a mouse model. A !fadA5 knockout strain of Mtb exhibited an inability to grow on cholesterol in vitro and a decrease in 4-AD and ADD production compared to wild type (65). The authors of this study interpret the latter result as a consequence of incomplete side chain degradation in the mutant strain. This strain also exhibited a growth attenuation phenotype after four weeks in a mouse infection model (65), a finding that contrasts the early growth defect seen in the !igr strain (56). While the relative importance of side chain degradation in various infection stages remains obscure, the 143  presence of more than one P450 that is able to initiate cholesterol side chain degradation may indicate an important role for this specific reaction, or for the entire side chain degradation process, in Mtb physiology. The occurrence of a 26-hydroxylated steroid in the mycobacterial catabolism of cholesterol catabolism is potentially significant in pathogenesis as hydroxysteroids are important signaling molecules in mammalian systems. Specifically, (25R)-26hydroxycholesterol (also called 27-hydroxycholesterol) plays a role in at least three processes: regulating cholesterol uptake and metabolism as an LXR ligand (309,310), preventing accumulation of cholesterol in macrophages and regulating macrophage differentiation (311), and modulating the activity of human estrogen receptors (312). The signaling roles of steroid metabolites with hydroxylated side chains are particularly intriguing considering that cyp125 is not essential for Mtb H37Rv growth on cholesterol but may contribute to the pathogen’s survival in animal models (56). 4.3.3  Ring degradation Targeted genetic knockouts of several genes involved in cholesterol ring degradation  have been used to investigate the role of this process during infection. Results of these studies have revealed somewhat contradictory phenotypes for knockouts of different pathway genes or knockouts of the same gene subjected to different experimental conditions. Mutation of the gene encoding 3!-HSD (Rv1106c), the first ring-transforming enzyme, resulted in growth comparable to wild type Mtb in both macrophage and guinea pig infection models while completely abrogating in vitro growth on cholesterol (178). It has been proposed that this reaction can also be performed by ChoD (Rv3409c) and that elimination of this enzyme causes growth defects in macrophage infection models (313). The finding that ChoD is not required for cholesterol degradation in the presence of 3!-HSD strongly suggests that the phenotype observed upon its deletion is not due to inhibition of cholesterol degradation (178). While a homologue of ChoD from M. smegmatis showed no in vitro cholesterol oxidase activity when expressed in E. coli (74), it remains possible that ChoD provides some level of compensatory activity for the mutation of Rv1106c in vivo. Perhaps 144  the strongest phenotype related to cholesterol degradation has been observed in knockouts of kshA and kshB that were evaluated in macrophage and mouse infection models (31). Deletion of either gene resulted in complete attenuation of virulence in a SCID mouse infection model that was reversible upon complementation with an intact copy of the appropriate gene (Figure 4.2). Similarly, colony-forming units isolated from the lung tissue of BALBc mice and from resting and IFN-!-activated macrophages were significantly reduced upon kshA or kshB mutation (31). The authors of this study found no evidence for toxicity of cholesterol or a cholesterol metabolite in in vitro growth studies of the mutants and concluded that the decreased virulence could be attributed to disruption of cholesterol metabolism (31). In contrast to this finding, mutation of hsaC resulted in the accumulation of catechol intermediates and an accompanying toxic effect. The observed toxicity obscures the mechanism for the observed attenuation of virulence in mouse and guinea pig infection models, although it provides some insight as to the temporal profile of cholesterol catabolism in Mtb infection (72). Finally, mutation of genes that are proposed to be involved in degradation of the downstream ring-degradation intermediate 9,17-dioxo-1,2,3,4,10,19hexanorandrostan-5-oic acid (Figure 1.5, compound 7 (R = ketone)) in the related pathogen R. equi resulted in attenuation in a macrophage infection model that correlated with growth defects on steroid substrates (90). Taken as a body of work, these studies do not seem to present a cohesive argument for the importance of cholesterol degradation in Mtb pathogenesis, but rather indicate that further investigations into the ring-degrading enzymes are necessary to evaluate their relative contributions.  145  Figure 4.2 Virulence of Mtb strains in SCID mice. Shown are the results of SCID mouse infection studies with wild type (WT), !kshA mutant (YH!kshA) and complemented stain (YHkshAComp) (Panel A), or !kshB (YH!kshB) and complemented strain (YHkshBComp) (Panel B). Figure from Hu et al., 2010 (31), © 2009 The Authors, Journal compilation © 2009 Blackwell Publishing Ltd., with permission.  Discrepancies in kshA knockout phenotypes  The phenotypes observed in genetic mutation studies for KshAB carried out by Hu et al. are striking in their degree of attenuation and convincing in their reversal upon complementation (31). Despite such a clear result in these experiments, equally clear and nearly opposite results were obtained for a !kshA knockout strain of Mtb through collaboration with Dr. Rainer Kalscheuer of the laboratory of William R. Jacobs Jr., Dr. Lan H. Ly in the Department of Microbial and Molecular Pathogenesis at Texas A & M  146  University, and Dr. Paul J. Converse from the Center for Tuberculosis Research in the Department of Medicine at Johns Hopkins University. In direct contradiction to the results obtained by Hu and coworkers, a !kshA deletion mutant of Mtb H37Rv exhibited no attenuation in a SCID mouse model (Dr. Rainer Kalscheuer, personal communication). Additionally, the same mutant exhibited colony-forming unit counts similar to wild type bacterium in the lungs and spleens of infected guinea pigs (Dr. Paul J. Converse, personal communication). Although these experiments demonstrated that the number of bacteria in infected guinea pig lungs were similar for mutant and wild type bacteria, the degree of pathology in this organ actually seemed to increase upon kshA deletion. Histopathological examination of the lungs of infected animals four and eight weeks after infection (Table 4.2) suggested a hypervirulent phenotype for bacteria lacking kshA and slight attenuation when kshA is provided under a constitutive promotor. Table 4.3 Week 4  Week 8  Histopathology of guinea pigs infected with Mtb strainsa strain Wild type !kshA !kshA-compc Wild type !kshA !kshA-comp  granulomas / field 8 10 4 12 7.8 6  Subjective score (1-4)b 1.7 2.3 1.6 1.8 2.1 1.4  a  Animals were infected via aerosol as described in Yam et al., 2009 (72). Subjective score from 1-4 was assessed for each field to describe size of granulomatous lesion and degree of necrosis. c kshA was provided in trans under control of the constitutive hsp60 promotor. b  There does not seem to be an obvious explanation either for the hyper-virulence observed for one !kshA mutant strain or for the obvious discrepancies between the results of the two sets of experiments. Hu et al. suggested that the robust phenotype of the !kshA mutant in their experiments could be due to accumulation of steroids with intact ring structures being shunted back into the host metabolism where these compounds could have an immunomodulatory effect that aids in clearing the infection (31). Paradoxically, a similar effect could be used to rationalize hyper-virulence of the !kshA mutant in guinea pig lung pathology, given the contributions of both the host and pathogen in granuloma development. Under this hypothesis, the different phenotypes observed for the two experiments could be  147  attributed to significant differences in the infection processes between the two infection models (116). Comparison of the two SCID mouse infection results, however, reveals a more direct contradiction with fewer differences in experimental conditions. Hu and coworkers used a complemented strain with kshA expression under control of the native promotor while the other mutants were complemented with constitutively expressed kshA. This difference, however, would be irrelevant in explaining the non-complemented phenotypes. A more likely explanation is genetic differences between the parent strains used to make the deletion mutants. Although both groups used a strain of Mtb H37Rv, recent data suggests that these strains maintained in different laboratories exhibit genetic differences with significant phenotypic consequences (105). Genetic variation between the parent strains used by the two research groups could affect bacterial metabolism in such a way that cholesterol utilization is altered in one strain with respect to the other and different phenotypes are elicited upon disruption of this pathway. It is similarly possible that genetic differences between the SCID mice used in the two sets of experiments had a phenotypically relevant effect on the experiment. 4.3.4  Toxicity of cholesterol and its metabolites Toxicity in cultures grown in the presence of cholesterol has been observed in  several studies investigating mutation of genes involved in the cholesterol degradation pathway. In this context, toxicity is defined as depressed growth in the presence of cholesterol, or its degradation metabolites, compared to identical growth conditions in the absence of this steroid. As some basal growth of Mtb and M. bovis BCG is supported in growth medium used to grow these strains in the absence of a carbon source supplement, depression of bacterial growth can be in comparison to basal growth or growth on other carbon sources. Toxicity has been observed for cholesterol degradation metabolites with both partially degraded and intact steroid ring structures. In !hsaC mutants of Mtb and M. bovis BCG, catecholic substrates for this enzyme were found to accumulate and elicit significant toxicity (72). The mechanisms of catechol toxicity have been established and involve oxidation to the  148  quinone derivatives and subsequent deleterious interaction with cellular components by a variety of mechanisms (314). 4-Cholesten-3-one accumulation was also found to be toxic in Mtb CDC1551 cells incapable of metabolizing this compound further (299). Moreover, 3cholesten-4-one toxicity was observed in the presence of glycerol for the CDC1551 !cyp125 knockout strain at 0.01 mM of the steroid and for the wild type bacteria at 0.1 mM 4cholesten-3-one. No toxicity was observed in !cyp125 knockouts of M. bovis BCG growing on a combination of cholesterol and acetate, but lower growth was observed for this strain in 4-cholesten-3-one-containing media than the basal growth observed in the absence of any added carbon source. This finding calls into question the mechanism of 4-cholesten-3-one toxicity observed in Mtb CDC1551 (299). The data seem to indicate that, in contrast to the catechol toxicity observed with HsaC mutants, toxicity of 4-cholesten-3-one is dependent upon the overall metabolic state of the bacterium, which is influenced by the nature of the primary carbon source. Toxicity was also observed in cultures containing cholesterol and glycerol for an Mtb H37Rv strain in which the 6-gene igr operon was mutated (30). This effect was reduced in the absence of a functional Mce4 cholesterol uptake system (30). This strain was, however, able to incorporate labeled carbon from C26 of cholesterol into its cellular lipid fraction. This ability, coupled with the multiple steroid 26-hydroxylase activities in the Mtb H37Rv strain, suggests that the observed toxicity can be attributed to loss of an igr gene other than cyp125 (266,299). The metabolite responsible for the observed toxic effect would therefore likely correspond to a side chain degradation intermediate representing a possible substrate for one of the other igr genes. A toxic effect of cholesterol or a downstream metabolite was not observed in Mtb lacking 3"-HSD (178) or FadA3 (65). Likewise, deletion of kshA or kshB resulted in growth similar to wild type on media containing both glucose and cholesterol (31). These results, coupled with the lack of a steroid-dependent toxic effect for !cyp125 M. bovis BCG in the presence of another carbon source (266), suggest that the toxicity observed in the abovementioned experiments is not a necessary consequence of exposure to or accumulation of molecules with an intact steroid ring structure. Instead, the observation of such a phenotype in some mycobacterial cholesterol-degradation gene knockout strains appears to  149  be dependent upon the specific nature of the accumulating compound(s) and to the overall physiological state of the bacterium. 4.3.5  Prospects for therapeutic development The balance of evidence collected so far indicates that the entire process of  cholesterol degradation is not critical to Mtb pathogenesis and survival in the host. This conclusion, however, does not preclude the possibility that specific components of the cholesterol catabolic pathway represent feasible targets for antimycobacterial therapeutics. While some reactions, such as 26-oxygenation, may have an important role in pathogen-host communication, other enzymes exhibit reaction-specific characteristics that make their inhibition with small molecules attractive from a therapeutic point of view. Consideration of such possibilities requires one to put aside the axiom that good therapeutic targets must be essential for bacterial survival or essential for a pathogenic process. Although the pharmacokinetic properties and delivery mechanisms of potential inhibitory compounds are an over-riding consideration with respect to therapeutic development, this discussion focuses on the biochemical potential for decreasing virulence through enzyme inhibition. One example of a non-essential enzyme that represents a potential therapeutic target is KshAB. As human cells do not contain any ROs, compounds designed to inhibit this enzyme are less likely to elicit toxic effects in the host. While deletion of kshA has produced variable results in different infection models (Section, inhibition of its activity by a small molecule that induces enzyme uncoupling represents an as-yet untested mechanism in Mtb killing. In other ROs, poor substrates and some inhibitors have been shown to elicit uncoupling of NADH and O2 consumption from substrate transformation, resulting in production of reduced oxygen species such as hydrogen peroxide (192,193). This approach could therefore have the doubly deleterious effect of disrupting the cholesterol degradation pathway at a point that potentially results in significant reduction in virulence (31) and introducing a novel source of oxidative stress to the cell. Similar uncoupling and hydrogen peroxide production could be achieved by targeting one of the P450s in the pathway (194), although the abundance of P450s in the human body could result in a higher potential for  150  cross-reactivity with host proteins. Neither Cyp125A1 nor Cyp142A1 are “essential” enzymes due to their functional redundancy in the cholesterol degradation pathway. Although the effect on virulence of a !cyp125!cyp142 double mutation has not been assessed, the similar ability of each of the encoded enzymes to transform the same substrates suggests it may be possible to inhibit both enzymes with the same compound and eliminate side chain transformation. More research is needed to determine the importance of individual genes in the pathway and the consequences of their inactivation on various stages of infection. The temporal profile of cholesterol utilization by Mtb during the infection process is especially pertinent to discussions about possible therapeutic application. If cholesterol is metabolized for energy during the latent infection stage, toxicity caused by inhibition of such enzymes as HsaC (72) could have a significant impact on cells that are relatively resistant to many other pharmaceutical interventions (112). Targeting cholesterol degradation must also be considered as complementary to other therapeutic approaches. This mindset is especially important given that active tuberculosis is treated with multi-drug therapies to increase efficacy and reduce the development of resistance in the bacterial population (92). Taking this approach into consideration, the stress of preventing cholesterol degradation through enzyme inhibition could be efficacious for bacterial killing in the presence of established pharmaceuticals. 4.4  Implications of global RO-O phylogenetic analysis The global phylogenetic analysis described in section 3.3 of this thesis challenged  previous ideas in both the methodologies used for enzyme phylogenetic analysis and in the understanding of ROs as an enzyme class. This work does not contradict findings in either of these areas, but effectively broadens the range of considerations that should be used to construct complete phylogenetic representations for enzyme families. Employing such consideration through use of inclusive search tools and methodology for phylogenetic analysis that is mindful of the particularities of ROs has enabled the generation of a more comprehensive view of this enzyme family than has been previously considered.  151  4.4.1  RO-O sequence diversity The overall tree topology confirms and extends previous biochemical and  phylogenetic studies. For example, a study of the five ROs conserved in several species of plant (221) found that four were closely related in both their primary sequences and biochemical functionality in the metabolism of chlorophyll, whereas CMO was more distantly related. The global RO-O phylogenetic tree also shows CMOthaliana as a member of Group I whereas the other four cluster closely together as Group II enzymes. This tree is also consistent with the phylogenetic relationships found in the Nam (258) and Kweon (191) analyses, although it presents a very different perspective on RO-O diversity. This study represents a major advancement over previous work with respect to recognizing and analyzing the great diversity characteristic of RO-Os. The methodology that was used involved collecting all available sequences and therefore the study presents an accurate assessment of the range of RO-Os currently known. Furthermore, using only structurally conserved positions for the phylogenetic reconstruction minimized the introduction of artifacts that can result from the difficulty inherent in aligning highly diverse sequences. When such artifacts are introduced, tree topology can be greatly affected by improperly aligned sequences, which have the effect of making properly aligned sequences appear to be more closely related than they really are (297). The result of the improved sequence sampling and alignment techniques described here is a much broader and more realistic view of RO-O phylogeny. For example, this perspective allows recognition of the actual prevalence of RO-O features such as the ! subunit. In light of this analysis, ROs must now be considered an enzyme class with far broader relevance than has been recognized previously. As the vast majority of the RO-O global phylogenetic tree remains unexplored biochemically, the diversity recognized in this study represents enormous opportunity for application of novel RO activities in areas such as biocatalysis, bioremediation, and medicine. Indeed, the range and biological importance of ROs could well rival that of the P450s. Consistent with this notion, the insight into RO-O  152  diversity gained from this analysis strongly suggests that, like P450s, a small number of categories is not sufficient for RO functional classification.  Limitations of common bioinformatic tools  The results of the global RO-O phylogenetic analysis serve not only to expand understanding of this enzyme family, but also to explain the discrepancy between this work and previous analyses and to illustrate some limitations of commonly-used methods. The most striking difference between this analysis and the phylogenetic reconstructions conducted by Nam et al. (258) and Kweon et al. (191) is the breadth of RO-O sequences included. These differences are readily attributable to the methods used to collect the sequences that were subsequently used for alignment and phylogenetic analysis. Both previous studies used BLAST (260) searches of characterized RO-Os to establish a pool of test sequences. In the context of the global RO-O phylogenetic tree, sequences collected in this way represent closely clustered groups rather than representative samples. Indeed, RO-O sequence diversity is so high that even recognition based on domain-characteristic motifs is imperfect and the catalytic domain of several RO-Os, including KshA, are missed by this more inclusive algorithm. BLAST searches are extremely useful in identifying proteins that are closely related to one-another in sequence and are therefore invaluable in initial investigations into the potential functionality of an uncharacterized sequence. This same attribute, however, limits their utility in the collection of highly diverse sequences. Another major difference between the analysis described in this thesis and previous work with RO-O sequences is the use of a structurally based template to improve sequence alignment accuracy and provide criteria for sequence segments to be used in phylogenetic reconstruction. The techniques applied to RO-O phylogenetic analysis described in this thesis were designed specifically to address the challenges for this protein family, but the principals of achieving representative sequence pools and meaningful alignments are applicable to all such analyses. As genome-sequencing technology becomes increasingly accessible, the amount of data being generated begins to limit the capacity of commonly used bioinformatic tools to  153  meaningfully process this data in a fully automated way. While there is enormous potential for discovery in the vastness of current sequence databases, examination of this data without appropriate experimental approaches can, as in the case of RO-Os, result in an incomplete picture that is not representative of our current body of knowledge. 4.4.2  Two domain fusion events possible in RO-O evolution Two lines of evidence suggest that modern RO-Os arose from two distinct fusion  events between ancestral Rieske and catalytic domains. First, the distinct two-group topology of the Rieske domain consensus tree is consistent with at least two independent ancestors. As this tree was calculated using sequence information from protein segments that are conserved between the two groups, the different secondary structural elements present between the second and third !-sheets of the Rieske domain were not included in tree calculation. The tree topology therefore derives from the distinct sequence trends for the two groups, which suggests that the insertion(s) were likely present in the two ancestral Rieske domains rather than representing insertion events after domain fusion. There are several subclasses of Rieske domain with a variety of such indels (261), and further analysis is needed to investigate the possible origins of those in modern RO-Os. The second line of evidence consistent with at least two domain fusion events derives from sequences in the catalytic domain. A motif in this domain, Asn-Trp-Lys, is highly conserved in only Group I enzymes, such that RO-Os bearing this motif have Rieske domains of a particular origin. This motif occurs on a helix separated from the Rieske domain by the central !-sheet. Significant physical separation of this sequence and the Rieske domain suggests that conservation of the motif in the catalytic domain is not a direct result of association with a specific Rieske domain. Two domain fusion events during RO-O evolution would elegantly explain these observations, but it is nonetheless possible that these conserved features arose following a single fusion event via a distinct mechanism. Such a mechanism could be illegitimate recombination with laterally transferred genes or multiple insertion and deletion events that resulted in strong selective advantages for the surviving two configurations.  154  The single node observed for the catalytic domain tree indicates that association with a specific Rieske domain does not apply strong evolutionary constraints on the sequence of the catalytic domain. This single-node topology could be a result of the higher sequence diversity observed in the catalytic domain with respect to the Rieske domain. This topology could also indicate that the proposed ancestral catalytic domains had a more recent common ancestor than their Rieske domain counterparts. The implication that the biolobal structure of the RO-O consensus tree derives primarily from the Rieske domain does not necessarily indicate that the sequence of this domain is the major determinant in RO-O evolution, but rather that its greater overall sequence conservation allows for such patterns to be discernable in the analysis. 4.5  Future directions Work presented in this thesis establishes that Cyp125A1 initiates cholesterol side  chain degradation in Mtb through oxidation of a terminal methyl group. Subsequent work by Johnston and coworkers (62) demonstrated that in some strains of Mtb, Cyp142A1 is able to perform the same reaction, with opposite stereochemistry at C25, and that the presence of either enzyme can support growth on cholesterol. The latter study also demonstrated that while both P450s catalyze oxidation of the C26-hydroxy to the C26-carboxylic acid steroid, there is significant accumulation of the hydroxylated compound in in vitro assay mixtures. The capacity of 26-hydroxycholesterol and 26-hydroxy-4-cholesten-3-one of either C25 stereochemistry to facilitate manipulation of the host by the pathogen is an intriguing area for investigation. The potential influence of these steroids over host cells could be investigated by monitoring changes in cytokine production, the transcriptome, and the gross physiology of macrophages in response to application of these compounds. Such experiments could be performed using macrophages with and without intracellular wild type or !cyp125!cyp142 Mtb, or by direct administration of the pertinent steroids to macrophage. Such experiments have the potential to detect whether or not the host has an immunologically relevant response to these steroid signals, or if these compounds are capable of regulating some other component of host physiology, such as lipid homeostasis. Similar experiments using other cell types associated with the granuloma may also be illuminating.  155  The investigations of KshAB activity described in this thesis strongly suggest that the physiological substrate(s) for this enzyme is an intermediate of cholesterol side chain degradation bearing a CoA-thioester. This work does not, however, identify the exact compound KshAB hydroxylates in vivo. Two experimental avenues can be explored with respect to this question: investigation of enzyme activity using in vitro assays and structural determination and detection of the cholesterol metabolome in a physiologically relevant context. In the first case, in vitro enzyme activity assays using both KshAB and KstD could be conducted using a wider range of cholesterol side chain degradation intermediates to establish each enzyme’s preferred substrate under the assay conditions. These assays could help to identify the likely order of action of the two enzymes, as well as suggest the degree of side chain degradation likely to take place before transformation by either enzyme. In addition to enzyme activity assays, co-crystallization experiments with KshA and potential physiological substrates could be revealing with respect to the specific substrate components that contribute to higher enzyme activity or affinity. Such knowledge could be useful in the design of inhibitory compounds. The second approach to exploring the physiological substrates for all of the cholesterol-degrading enzymes is detection of the relevant metabolites in Mtb that is growing in conditions modeling natural infection. This experiment is not trivial for several reasons. Firstly, although growth on cholesterol in vitro could enrich for cholesterol-derived metabolites, these conditions cannot be considered physiologically relevant, as evidenced by accumulation of ADD and 4-AD in cultures of cholesterol-grown Mtb (64,65). Secondly, while growth in macrophages or a similar condition provides an environment that more closely approximates that of infection, cholesterol degradation may occur at a very low level under many conditions, which potentially makes detection of pathway intermediates exceedingly difficult. This problem may be partially ameliorated by infecting cells containing radiolabeled cholesterol. Nonetheless, conducting such studies in the most informative way requires greater knowledge of the temporal profile of cholesterol utilization by Mtb and the specific conditions governing it than is currently available.  156  The principal question in the field of mycobacterial cholesterol degradation is the importance of this process to bacterial physiology and pathogenicity. Attempts to address this question through elimination of specific enzyme activities in the cholesterol degradation pathway have met with mixed results. For example, genetic mutation of kshA performed by different research groups resulted in completely different phenotypes in a SCID mouse infection model. To address the possibility of variations in bacterial strain, as well as to demonstrate the importance of this reaction in the context of the range of clinically relevant Mtb strains, !kshA knockouts could be constructed using a variety of lab and clinical Mtb stains. These genetic mutants could then be tested in both macrophage and animal infection models and the results could be interpreted in the context of KshA importance during Mtb infection. Obviously such studies would be highly resource-intensive. An alternative approach for investigating the importance of KshA in the context of steroid ring opening could be to test mutants of kstD in infection models. As both the KshAB and KstD reactions are required for ring B opening, elimination of KstD could shed light on the specific importance of breaking the steroid core structure. While further studies focused on KshA and KstD could provide insight as to the importance of this component of cholesterol degradation to Mtb growth and survival in the host, true appreciation of the role of the entire cholesterol degradation process will likely only derive from greater understanding of mycobacterial physiology in general in the context of infection. Whether or not KshAB is essential for Mtb infection, use of inhibitors for this enzyme that uncouple NADH and O2 consumption from steroid hydroxylation could be an effective therapeutic strategy, especially in combination with other anti-mycobacterial compounds. The feasibility of this approach could be tested by expressing enzymes that catalyze formation of hydrogen peroxide or other reactive oxygen species in Mtb under the control of the KshA promotor. A strain containing such a construct could be tested in infection models in isolation or in combination with currently employed tuberculosis therapies to assess the potential effect of KshAB uncoupling. Uncoupling inhibitor lead compounds could be evaluated in vitro through detection of produced hydrogen peroxide by measuring generation of O2 in an oxygen electrode assay upon addition of catalase to the reaction cuvette.  157  The RO-O global phylogenetic analysis presented in section 3.3 of this thesis significantly increases the scope of RO-O diversity previously appreciated in the literature. It also suggests that the evolutionary history of modern RO-Os may have involved multiple fusion events between ancestral Rieske and catalytic domains. To test this hypothesis, a similar analysis could be carried out for isolated Rieske domains. These domains exist as individual polypeptides, as well as fused to a great variety of enzymatic and redox-associated domains. Such an analysis would thus encompass all proteins including a Rieske domain, and would likely be highly informative regarding the evolutionary relationships between the wide variety of proteins incorporating this domain.  158  Bibliography 1.  Eschenmoser, A., Ruzicka, L., Jeger, O., and Arigoni, D. (1955) Helvetica Chimica Acta 38, 1890  2. 3.  Eschenmoser, A., and Arigoni, D. (2005) Helvetica Chimica Acta 88, 3011-3050 Edgren, R. A., and Stanczyk, F. Z. (1999) Contraception 60, 313  4.  Nelson, D. L., and Cox, M. M. (2005) Lehninger Principals of Biochemistry. 4 Ed., W. H. Freeman and Company, New York. pp 816-829  5.  Summons, R. E., Bradley, A. S., Jahnke, L. L., and Waldbauer, J. R. (2006) Philos Trans R Soc Lond B Biol Sci 361, 951-968  6. 7.  Bird, C. W., Lynch, J. M., Pirt, F. J., and Reid, W. W. (1971) Nature 230, 473-474 Pearson, A., Budin, M., and Brocks, J. J. (2003) Proc Natl Acad Sci U S A 100, 15352-15357  8.  Bode, H. B., Zeggel, B., Silakowski, B., Wenzel, S. C., Reichenbach, H., and Muller, R. (2003) Mol Microbiol 47, 471-481  9.  Volkman, J. K. (2003) Appl Microbiol Biotechnol 60, 495-506  10.  Lamb, D. C., Kelly, D. E., Manning, N. J., and Kelly, S. L. (1998) FEBS Lett 437, 142-144  11.  Gilbert, L. I., Rybczynski, R., and Warren, J. T. (2002) Annu Rev Entomol 47, 883916  12.  Disch, A., and Rohmer, M. (1998) FEMS Microbiol Lett 168, 201-208  13.  Schwender, J., Seemann, M., Lichtenthaler, H. K., and Rohmer, M. (1996) Biochem J 316 ( Pt 1), 73-80  14.  Chen, L. L., Wang, G. Z., and Zhang, H. Y. (2007) Biochem Biophys Res Commun 363, 885-888  15.  Sih, C. J., and Bennett, R. E. (1962) Biochim Biophys Acta 56, 584-592  16.  Horinouchi, M., Hayashi, T., Yamamoto, T., and Kudo, T. (2003) Appl Environ Microbiol 69, 4421-4430  17.  Yam, K., Van der Geize, R., and Eltis, L. (2010) Catabolism of aromatic compounds and steroids by Rhodococcus, In Biology of Rhodococcus. (Ed. Alvarez). in Biology of Rhodococcus (Alvarez ed.), Springer. pp 133-170 159  18.  Ouellet, H., Johnston, J. B., and Montellano, P. R. (2011) Trends Microbiol 19, 530539  19.  Martin, C. K. (1977) Adv Appl Microbiol 22, 29-58  20.  McLeod, M. P., Warren, R. L., Hsiao, W. W., Araki, N., Myhre, M., Fernandes, C., Miyazawa, D., Wong, W., Lillquist, A. L., Wang, D., Dosanjh, M., Hara, H., Petrescu, A., Morin, R. D., Yang, G., Stott, J. M., Schein, J. E., Shin, H., Smailus, D., Siddiqui, A. S., Marra, M. A., Jones, S. J., Holt, R., Brinkman, F. S., Miyauchi, K., Fukuda, M., Davies, J. E., Mohn, W. W., and Eltis, L. D. (2006) Proc Natl Acad Sci U S A 103, 15582-15587  21.  van der Geize, R., Yam, K., Heuser, T., Wilbrink, M. H., Hara, H., Anderton, M. C., Sim, E., Dijkhuizen, L., Davies, J. E., Mohn, W. W., and Eltis, L. D. (2007) Proc Natl Acad Sci U S A 104, 1947-1952  22.  Rengarajan, J., Bloom, B. R., and Rubin, E. J. (2005) Proc Natl Acad Sci U S A 102, 8327-8332  23.  Hylemon, P. B., and Harder, J. (1998) FEMS Microbiol Rev 22, 475-488  24.  Mohn, W. W., van der Geize, R., Stewart, G. R., Okamoto, S., Liu, J., Dijkhuizen, L., and Eltis, L. D. (2008) J Biol Chem 283, 35368-35374  25.  Harder, J., and Probian, C. (1997) Arch Microbiol 167, 269-274  26. 27.  Tarlera, S., and Denner, E. B. (2003) Int J Syst Evol Microbiol 53, 1085-1091 Chiang, Y. R., Fang, J. Y., Ismail, W., and Wang, P. H. (2010) Microbiology 156, 2253-2259  28.  Leu, Y. L., Wang, P. H., Shiao, M. S., Ismail, W., and Chiang, Y. R. (2011) J Bacteriol 193, 4447-4455  29.  Birkenmaier, A., Holert, J., Erdbrink, H., Moeller, H. M., Friemel, A., Schoenenberger, R., Suter, M. J., Klebensberger, J., and Philipp, B. (2007) J Bacteriol 189, 7165-7173  30. 31.  Pandey, A. K., and Sassetti, C. M. (2008) Proc Natl Acad Sci U S A 105, 4376-4380 Hu, Y., van der Geize, R., Besra, G. S., Gurcha, S. S., Liu, A., Rohde, M., Singh, M., and Coates, A. (2010) Mol Microbiol 75, 107-121  32.  Kieslich, K. (1985) J Basic Microbiol 25, 461-474  160  33.  Andor, A., Jekkel, A., Hopwood, D. A., Jeanplong, F., Ilkoy, E., Konya, A., Kurucz, I., and Ambrus, G. (2006) Appl Environ Microbiol 72, 6554-6559  34.  Ikonen, E. (2008) Nat Rev Mol Cell Biol 9, 125-138  35.  Quinn, P. J., and Chapman, D. (1980) CRC Crit Rev Biochem 8, 1-117  36.  Aguilar, P. S., and de Mendoza, D. (2006) Mol Microbiol 62, 1507-1514  37.  Cooper, G. M. (2000) The Cell: A Molecular Approach, 2 ed., Sinauer Associates, Sunderland (MA)  38.  Ourisson, G., Rohmer, M., and Poralla, K. (1987) Annu Rev Microbiol 41, 301-333  39.  Hotze, E. M., and Tweten, R. K. (2012) Biochim Biophys Acta 1818, 1028-1038  40.  Hartlova, A., Cerveny, L., Hubalek, M., Krocova, Z., and Stulik, J. (2010) Microbiol Immunol 54, 237-245  41.  Drzyzga, O., Navarro Llorens, J. M., Fernandez de Las Heras, L., Garcia Fernandez, E., and Perera, J. (2009) Int J Syst Evol Microbiol 59, 1011-1015  42.  Watanabe, K., Shimizu, H., Hidetaka, A., Nakamura, R., Suzuki, K., and Komagata, K. (1986) J. Gen. Appl. Microbiol. 32, 137-147  43.  Av-Gay, Y., and Sobouti, R. (2000) Can J Microbiol 46, 826-831  44.  Doukyu, N. (2009) Appl Microbiol Biotechnol 83, 825-837  45.  van der Geize, R., and Dijkhuizen, L. (2004) Curr Opin Microbiol 7, 255-261  46.  Hogg, J. A. (1992) Steroids 57, 593-616  47.  Donova, M. V. (2007) Prikl Biokhim Mikrobiol 43, 5-18  48.  Griffin, J. E., Gawronski, J. D., Dejesus, M. A., Ioerger, T. R., Akerley, B. J., and Sassetti, C. M. (2011) PLoS Pathog 7, e1002251  49.  Rohde, K. H., Abramovitch, R. B., and Russell, D. G. (2007) Cell Host Microbe 2, 352-364  50.  Homolka, S., Niemann, S., Russell, D. G., and Rohde, K. H. (2010) PLoS Pathog 6, e1000988  161  51.  Schnappinger, D., Ehrt, S., Voskuil, M. I., Liu, Y., Mangan, J. A., Monahan, I. M., Dolganov, G., Efron, B., Butcher, P. D., Nathan, C., and Schoolnik, G. K. (2003) J Exp Med 198, 693-704  52.  Sassetti, C. M., and Rubin, E. J. (2003) Proc Natl Acad Sci U S A 100, 12989-12994  53.  Horinouchi, M., Kurita, T., Yamamoto, T., Hatori, E., Hayashi, T., and Kudo, T. (2004) Biochem Biophys Res Commun 324, 597-604  54.  Dresen, C., Lin, L. Y., D'Angelo, I., Tocheva, E. I., Strynadka, N., and Eltis, L. D. (2010) J Biol Chem 285, 22264-22275  55.  Sih, C. J., Tai, H. H., Tsong, Y. Y., Lee, S. S., and Coombe, R. G. (1968) Biochemistry 7, 808-818  56.  Chang, J. C., Miner, M. D., Pandey, A. K., Gill, W. P., Harik, N. S., Sassetti, C. M., and Sherman, D. R. (2009) J Bacteriol 191, 5232-5239  57.  Kendall, S. L., Burgess, P., Balhana, R., Withers, M., Ten Bokum, A., Lott, J. S., Gao, C., Uhia-Castro, I., and Stoker, N. G. (2010) Microbiology 156, 1362-1371  58.  Kendall, S. L., Withers, M., Soffair, C. N., Moreland, N. J., Gurcha, S., Sidders, B., Frita, R., Ten Bokum, A., Besra, G. S., Lott, J. S., and Stoker, N. G. (2007) Mol Microbiol 65, 684-699  59.  Flesselles, B., Anand, N. N., Remani, J., Loosmore, S. M., and Klein, M. H. (1999) FEMS Microbiol Lett 177, 237-242  60.  Zhang, F., and Xie, J. P. (2011) Mol Cell Biochem 352, 1-10  61.  Wakil, S. J. (1989) Biochemistry 28, 4523-4530  62.  Johnston, J. B., Ouellet, H., and Ortiz de Montellano, P. R. (2010) J Biol Chem 285, 36352-36360  63.  Szentirmai, A. (1990) Journal of Industrial Microbiology 6, 101-116  64.  Thomas, S. T., van der Ven, B. C., Sherman, D. R., Russell, D. G., and Sampson, N. S. (2011) J Biol Chem 286, 43668-43678  65.  Nesbitt, N. M., Yang, X., Fontan, P., Kolesnikova, I., Smith, I., Sampson, N. S., and Dubnau, E. (2010) Infect Immun 78, 275-282  66.  Wilbrink, M. H., Petrusma, M., Dijkhuizen, L., and van der Geize, R. (2011) Appl Environ Microbiol 77, 4455-4464  162  67.  Chang, F. N., and Sih, C. J. (1964) Biochemistry 3, 1551-1557  68.  Sih, C. J., and Rahim, A. M. (1963) J Pharm Sci 52, 1075-1080  69.  Gibson, D. T., Wang, K. C., Sih, C. J., and Whitlock, H., Jr. (1966) J Biol Chem 241, 551-559  70.  Sih, C. J., Lee, S. S., Tsong, Y. Y., and Wang, K. C. (1966) J Biol Chem 241, 540550  71.  Lee, S. S., and Sih, C. J. (1967) Biochemistry 6, 1395-1403  72.  Yam, K. C., D'Angelo, I., Kalscheuer, R., Zhu, H., Wang, J. X., Snieckus, V., Ly, L. H., Converse, P. J., Jacobs, W. R., Jr., Strynadka, N., and Eltis, L. D. (2009) PLoS Pathog 5, e1000344  73.  Chiang, Y. R., Ismail, W., Heintz, D., Schaeffer, C., Van Dorsselaer, A., and Fuchs, G. (2008) J Bacteriol 190, 905-914  74.  Uhia, I., Galan, B., Morales, V., and Garcia, J. L. (2011) Environ Microbiol 13, 943959  75.  Kreit, J., and Sampson, N. S. (2009) FEBS J 276, 6844-6856  76.  Machang'u, R. S., and Prescott, J. F. (1991) Can J Vet Res 55, 332-340  77.  Gadda, G., Wels, G., Pollegioni, L., Zucchelli, S., Ambrosius, D., Pilone, M. S., and Ghisla, S. (1997) Eur J Biochem 250, 369-376  78.  Yang, X., Dubnau, E., Smith, I., and Sampson, N. S. (2007) Biochemistry 46, 90589067  79.  Plesiat, P., Grandguillot, M., Harayama, S., Vragar, S., and Michel-Briand, Y. (1991) J Bacteriol 173, 7219-7227  80.  Morii, S., Fujii, C., Miyoshi, T., Iwami, M., and Itagaki, E. (1998) J Biochem 124, 1026-1032  81.  Molnar, I., Choi, K. P., Yamashita, M., and Murooka, Y. (1995) Mol Microbiol 15, 895-905  82.  van der Geize, R., Hessels, G. I., van Gerwen, R., van der Meijden, P., and Dijkhuizen, L. (2001) FEMS Microbiol Lett 205, 197-202  163  83.  van der Geize, R., Hessels, G. I., van Gerwen, R., Vrijbloed, J. W., van Der Meijden, P., and Dijkhuizen, L. (2000) Appl Environ Microbiol 66, 2029-2036  84.  Brzostek, A., Sliwinski, T., Rumijowska-Galewicz, A., Korycka-Machala, M., and Dziadek, J. (2005) Microbiology 151, 2393-2402  85.  van der Geize, R., Hessels, G. I., van Gerwen, R., van der Meijden, P., and Dijkhuizen, L. (2002) Mol Microbiol 45, 1007-1018  86.  van der Geize, R., Hessels, G. I., and Dijkhuizen, L. (2002) Microbiology 148, 32853292  87.  Knol, J., Bodewits, K., Hessels, G. I., Dijkhuizen, L., and van der Geize, R. (2008) Biochem J 410, 339-346  88.  Thomas, S. T., Yang, X., and Sampson, N. S. (2011) Bioorg Med Chem Lett 21, 2216-2219  89.  Lack, N. A., Yam, K. C., Lowe, E. D., Horsman, G. P., Owen, R. L., Sim, E., and Eltis, L. D. (2010) J Biol Chem 285, 434-443  90.  van der Geize, R., Grommen, A. W., Hessels, G. I., Jacobs, A. A., and Dijkhuizen, L. (2011) PLoS Pathog 7, e1002181  91.  Miclo, A., and Germain, P. (1990) Appl Microbiol Biotechnol 32, 594-599  92.  WHO. (2010) Global Tuberculosis Control: WHO report 2010, Geneva, Switzerland  93.  Werner, E. (1950) Kinderarztl Prax 18, 170-174  94.  Andries, K., Verhasselt, P., Guillemont, J., Gohlmann, H. W., Neefs, J. M., Winkler, H., Van Gestel, J., Timmerman, P., Zhu, M., Lee, E., Williams, P., de Chaffoy, D., Huitric, E., Hoffner, S., Cambau, E., Truffot-Pernot, C., Lounis, N., and Jarlier, V. (2005) Science 307, 223-227  95.  Stover, C. K., Warrener, P., VanDevanter, D. R., Sherman, D. R., Arain, T. M., Langhorne, M. H., Anderson, S. W., Towell, J. A., Yuan, Y., McMurray, D. N., Kreiswirth, B. N., Barry, C. E., and Baker, W. R. (2000) Nature 405, 962-966  96.  Ventura, M., Canchaya, C., Tauch, A., Chandra, G., Fitzgerald, G. F., Chater, K. F., and van Sinderen, D. (2007) Microbiol Mol Biol Rev 71, 495-548  97.  Hett, E. C., and Rubin, E. J. (2008) Microbiol Mol Biol Rev 72, 126-156, table of contents  164  98.  Shiloh, M. U., and DiGiuseppe Champion, P. A. (2010) Curr Opin Microbiol 13, 8692  99.  Davis, J. M., and Ramakrishnan, L. (2009) Cell 136, 37-49  100.  Hoffmann, C., Leis, A., Niederweis, M., Plitzko, J. M., and Engelhardt, H. (2008) Proc Natl Acad Sci U S A 105, 3963-3967  101.  Jackson, M., Stadthagen, G., and Gicquel, B. (2007) Tuberculosis (Edinb) 87, 78-86  102.  Thanky, N. R., Young, D. B., and Robertson, B. D. (2007) Tuberculosis (Edinb) 87, 231-236  103.  Brosch, R., Gordon, S. V., Marmiesse, M., Brodin, P., Buchrieser, C., Eiglmeier, K., Garnier, T., Gutierrez, C., Hewinson, G., Kremer, K., Parsons, L. M., Pym, A. S., Samper, S., van Soolingen, D., and Cole, S. T. (2002) Proc Natl Acad Sci U S A 99, 3684-3689  104.  Russell, D. G., VanderVen, B. C., Lee, W., Abramovitch, R. B., Kim, M. J., Homolka, S., Niemann, S., and Rohde, K. H. (2010) Cell Host Microbe 8, 68-76  105.  Ioerger, T. R., Feng, Y., Ganesula, K., Chen, X., Dobos, K. M., Fortune, S., Jacobs, W. R., Jr., Mizrahi, V., Parish, T., Rubin, E., Sassetti, C., and Sacchettini, J. C. (2010) J Bacteriol 192, 3645-3653  106.  Russell, D. G., Barry, C. E., 3rd, and Flynn, J. L. (2010) Science 328, 852-856  107.  Russell, D. G. (2007) Nat Rev Microbiol 5, 39-47  108.  Sturgill-Koszycki, S., Schlesinger, P. H., Chakraborty, P., Haddix, P. L., Collins, H. L., Fok, A. K., Allen, R. D., Gluck, S. L., Heuser, J., and Russell, D. G. (1994) Science 263, 678-681  109.  Mwandumba, H. C., Russell, D. G., Nyirenda, M. H., Anderson, J., White, S. A., Molyneux, M. E., and Squire, S. B. (2004) J Immunol 172, 4592-4598  110.  Lin, P. L., Pawar, S., Myers, A., Pegu, A., Fuhrman, C., Reinhart, T. A., Capuano, S. V., Klein, E., and Flynn, J. L. (2006) Infect Immun 74, 3790-3803  111.  Gupta, A., Kaul, A., Tsolaki, A. G., Kishore, U., and Bhakta, S. (2012) Immunobiology 217, 363-374  112.  Chao, M. C., and Rubin, E. J. (2010) Annu Rev Microbiol 64, 293-311  113.  Flynn, J. L., and Chan, J. (2001) Infect Immun 69, 4195-4201  165  114.  McCune, R. M., Feldmann, F. M., Lambert, H. P., and McDermott, W. (1966) J Exp Med 123, 445-468  115.  Manabe, Y. C., Kesavan, A. K., Lopez-Molina, J., Hatem, C. L., Brooks, M., Fujiwara, R., Hochstein, K., Pitt, M. L., Tufariello, J., Chan, J., McMurray, D. N., Bishai, W. R., Dannenberg, A. M., Jr., and Mendez, S. (2008) Tuberculosis (Edinb) 88, 187-196  116.  Dharmadhikari, A. S., and Nardell, E. A. (2008) Am J Respir Cell Mol Biol 39, 503508  117.  Wayne, L. G., and Hayes, L. G. (1996) Infect Immun 64, 2062-2069  118.  Leistikow, R. L., Morton, R. A., Bartek, I. L., Frimpong, I., Wagner, K., and Voskuil, M. I. (2010) J Bacteriol 192, 1662-1670  119.  Betts, J. C., Lukey, P. T., Robb, L. C., McAdam, R. A., and Duncan, K. (2002) Mol Microbiol 43, 717-731  120.  Gill, W. P., Harik, N. S., Whiddon, M. R., Liao, R. P., Mittler, J. E., and Sherman, D. R. (2009) Nat Med 15, 211-214  121.  Lillebaek, T., Dirksen, A., Vynnycky, E., Baess, I., Thomsen, V. O., and Andersen, A. B. (2003) J Infect Dis 188, 1032-1039  122.  Daniel, J., Maamar, H., Deb, C., Sirakova, T. D., and Kolattukudy, P. E. (2011) PLoS Pathog 7, e1002093  123.  Via, L. E., Lin, P. L., Ray, S. M., Carrillo, J., Allen, S. S., Eum, S. Y., Taylor, K., Klein, E., Manjunatha, U., Gonzales, J., Lee, E. G., Park, S. K., Raleigh, J. A., Cho, S. N., McMurray, D. N., Flynn, J. L., and Barry, C. E., 3rd. (2008) Infect Immun 76, 2333-2340  124.  Mendelson, M., Walters, S., Smith, I., and Kaplan, G. (2005) Tuberculosis (Edinb) 85, 407-413  125.  Vankayalapati, R., and Barnes, P. F. (2009) Tuberculosis (Edinb) 89 Suppl 1, S77-80  126.  Caruso, A. M., Serbina, N., Klein, E., Triebold, K., Bloom, B. R., and Flynn, J. L. (1999) J Immunol 162, 5407-5416  127.  Jager, A., and Kuchroo, V. K. (2010) Scand J Immunol 72, 173-184  128.  Tufariello, J. M., Chan, J., and Flynn, J. L. (2003) Lancet Infect Dis 3, 578-590  166  129.  van der Wel, N., Hava, D., Houben, D., Fluitsma, D., van Zon, M., Pierson, J., Brenner, M., and Peters, P. J. (2007) Cell 129, 1287-1298  130.  Russell, D. G., Vanderven, B. C., Glennie, S., Mwandumba, H., and Heyderman, R. S. (2009) Nat Rev Immunol 9, 594-600  131.  Flynn, J. L., and Chan, J. (2003) Curr Opin Immunol 15, 450-455  132.  Steenken, W., Oatway, W. H., and Petroff, S. A. (1934) J Exp Med 60, 515-540  133.  Steenken, W., Jr., and Gardner, L. U. (1946) Am Rev Tuberc 54, 62-66  134.  Branch, A. (1927) Can Med Assoc J 17, 720-721  135.  Mahairas, G. G., Sabo, P. J., Hickey, M. J., Singh, D. C., and Stover, C. K. (1996) J Bacteriol 178, 1274-1282  136.  Shi, L., Sohaskey, C. D., Pfeiffer, C., Datta, P., Parks, M., McFadden, J., North, R. J., and Gennaro, M. L. (2010) Mol Microbiol 78, 1199-1215  137.  Muttucumaru, D. G., Roberts, G., Hinds, J., Stabler, R. A., and Parish, T. (2004) Tuberculosis (Edinb) 84, 239-246  138.  Pethe, K., Swenson, D. L., Alonso, S., Anderson, J., Wang, C., and Russell, D. G. (2004) Proc Natl Acad Sci U S A 101, 13642-13647  139.  Baek, S. H., Li, A. H., and Sassetti, C. M. (2011) PLoS Biol 9, e1001065  140.  Rohde, K., Yates, R. M., Purdy, G. E., and Russell, D. G. (2007) Immunol Rev 219, 37-54  141.  Vieira, O. V., Botelho, R. J., and Grinstein, S. (2002) Biochem J 366, 689-704  142.  Vergne, I., Fratti, R. A., Hill, P. J., Chua, J., Belisle, J., and Deretic, V. (2004) Mol Biol Cell 15, 751-760  143.  Welin, A., Winberg, M. E., Abdalla, H., Sarndahl, E., Rasmusson, B., Stendahl, O., and Lerm, M. (2008) Infect Immun 76, 2882-2887  144.  Nguyen, L., and Pieters, J. (2005) Trends Cell Biol 15, 269-276  145.  Ferrari, G., Langen, H., Naito, M., and Pieters, J. (1999) Cell 97, 435-447  146.  Bryk, R., Lima, C. D., Erdjument-Bromage, H., Tempst, P., and Nathan, C. (2002) Science 295, 1073-1077  167  147.  Korbel, D. S., Schneider, B. E., and Schaible, U. E. (2008) Microbes Infect 10, 9951004  148.  Raymond, J. B., Mahapatra, S., Crick, D. C., and Pavelka, M. S., Jr. (2005) J Biol Chem 280, 326-333  149.  Camacho, L. R., Constant, P., Raynaud, C., Laneelle, M. A., Triccas, J. A., Gicquel, B., Daffe, M., and Guilhot, C. (2001) J Biol Chem 276, 19845-19854  150.  Rousseau, C., Winter, N., Pivert, E., Bordat, Y., Neyrolles, O., Ave, P., Huerre, M., Gicquel, B., and Jackson, M. (2004) Cell Microbiol 6, 277-287  151.  Saavedra, R., Segura, E., Leyva, R., Esparza, L. A., and Lopez-Marin, L. M. (2001) Clin Diagn Lab Immunol 8, 1081-1088  152.  Flint, J. L., Kowalski, J. C., Karnati, P. K., and Derbyshire, K. M. (2004) Proc Natl Acad Sci U S A 101, 12598-12603  153.  Gengenbacher, M., Rao, S. P., Pethe, K., and Dick, T. (2010) Microbiology 156, 8187  154.  Kumar, A., Toledo, J. C., Patel, R. P., Lancaster, J. R., Jr., and Steyn, A. J. (2007) Proc Natl Acad Sci U S A 104, 11568-11573  155.  Russell-Goldman, E., Xu, J., Wang, X., Chan, J., and Tufariello, J. M. (2008) Infect Immun 76, 4269-4281  156.  Bloch, H., and Segal, W. (1956) J Bacteriol 72, 132-141  157.  Kim, M. J., Wainwright, H. C., Locketz, M., Bekker, L. G., Walther, G. B., Dittrich, C., Visser, A., Wang, W., Hsu, F. F., Wiehart, U., Tsenova, L., Kaplan, G., and Russell, D. G. (2010) EMBO Mol Med 2, 258-274  158.  Peyron, P., Vaubourgeix, J., Poquet, Y., Levillain, F., Botanch, C., Bardou, F., Daffe, M., Emile, J. F., Marchou, B., Cardona, P. J., de Chastellier, C., and Altare, F. (2008) PLoS Pathog 4, e1000204  159.  Brzostek, A., Pawelczyk, J., Rumijowska-Galewicz, A., Dziadek, B., and Dziadek, J. (2009) J Bacteriol 191, 6584-6591  160.  de Carvalho, L. P., Fischer, S. M., Marrero, J., Nathan, C., Ehrt, S., and Rhee, K. Y. (2010) Chem Biol 17, 1122-1131  161.  Singh, A., Crossman, D. K., Mai, D., Guidry, L., Voskuil, M. I., Renfrow, M. B., and Steyn, A. J. (2009) PLoS Pathog 5, e1000545  168  162.  McKinney, J. D., Honer zu Bentrup, K., Munoz-Elias, E. J., Miczak, A., Chen, B., Chan, W. T., Swenson, D., Sacchettini, J. C., Jacobs, W. R., Jr., and Russell, D. G. (2000) Nature 406, 735-738  163.  Savvi, S., Warner, D. F., Kana, B. D., McKinney, J. D., Mizrahi, V., and Dawes, S. S. (2008) J Bacteriol 190, 3886-3895  164.  Alam, M. S., Garg, S. K., and Agrawal, P. (2007) Mol Microbiol 63, 1414-1431  165.  Steyn, A. J., Collins, D. M., Hondalus, M. K., Jacobs, W. R., Jr., Kawakami, R. P., and Bloom, B. R. (2002) Proc Natl Acad Sci U S A 99, 3147-3152  166.  Morris, R. P., Nguyen, L., Gatfield, J., Visconti, K., Nguyen, K., Schnappinger, D., Ehrt, S., Liu, Y., Heifets, L., Pieters, J., Schoolnik, G., and Thompson, C. J. (2005) Proc Natl Acad Sci U S A 102, 12200-12205  167.  Beste, D. J., Espasa, M., Bonde, B., Kierzek, A. M., Stewart, G. R., and McFadden, J. (2009) PLoS One 4, e5349  168.  Davenport, M. P., Belz, G. T., and Ribeiro, R. M. (2009) Trends Immunol 30, 61-66  169.  Gioffre, A., Infante, E., Aguilar, D., Santangelo, M. P., Klepp, L., Amadio, A., Meikle, V., Etchechoury, I., Romano, M. I., Cataldi, A., Hernandez, R. P., and Bigi, F. (2005) Microbes Infect 7, 325-334  170.  Flynn, J. L., and Chan, J. (2005) Trends Microbiol 13, 98-102  171.  Schafer, G., Guler, R., Murray, G., Brombacher, F., and Brown, G. D. (2009) PLoS One 4, e8448  172.  Martens, G. W., Arikan, M. C., Lee, J., Ren, F., Vallerskog, T., and Kornfeld, H. (2008) Infect Immun 76, 3464-3472  173.  Gatfield, J., and Pieters, J. (2000) Science 288, 1647-1650  174.  Peyron, P., Bordier, C., N'Diaye, E. N., and Maridonneau-Parini, I. (2000) J Immunol 165, 5186-5191  175.  Saini, N. K., Sharma, M., Chandolia, A., Pasricha, R., Brahmachari, V., and Bose, M. (2008) BMC Microbiol 8, 200  176.  de Chastellier, C., and Thilo, L. (2006) Cell Microbiol 8, 242-256  177.  Chang, J. C., Harik, N. S., Liao, R. P., and Sherman, D. R. (2007) J Infect Dis 196, 788-795  169  178.  Yang, X., Gao, J., Smith, I., Dubnau, E., and Sampson, N. S. (2011) J Bacteriol 193, 1473-1476  179.  Gale, S. E., Westover, E. J., Dudley, N., Krishnan, K., Merlin, S., Scherrer, D. E., Han, X., Zhai, X., Brockman, H. L., Brown, R. E., Covey, D. F., Schaffer, J. E., Schlesinger, P., and Ory, D. S. (2009) J Biol Chem 284, 1755-1764  180.  Kovaleva, E. G., and Lipscomb, J. D. (2008) Nat Chem Biol 4, 186-193  181.  Sono, M., Roach, M. P., Coulter, E. D., and Dawson, J. H. (1996) Chem Rev 96, 2841-2888  182.  Pau, M. Y., Lipscomb, J. D., and Solomon, E. I. (2007) Proc Natl Acad Sci U S A 104, 18355-18362  183.  Torres Pazmino, D. E., Winkler, M., Glieder, A., and Fraaije, M. W. (2010) J Biotechnol 146, 9-24  184.  Lipscomb, J. D. (1994) Annu Rev Microbiol 48, 371-399  185.  Woodland, M. P., and Dalton, H. (1984) J Biol Chem 259, 53-59  186.  Hegg, E. L., and Que, L., Jr. (1997) Eur J Biochem 250, 625-629  187.  Straganz, G. D., and Nidetzky, B. (2006) Chembiochem 7, 1536-1548  188.  Valway, S. E., Sanchez, M. P., Shinnick, T. F., Orme, I., Agerton, T., Hoy, D., Jones, J. S., Westmoreland, H., and Onorato, I. M. (1998) N Engl J Med 338, 633-639  189.  McLean, K. J., Clift, D., Lewis, D. G., Sabri, M., Balding, P. R., Sutcliffe, M. J., Leys, D., and Munro, A. W. (2006) Trends Microbiol 14, 220-228  190.  Ouellet, H., Lang, J., Couture, M., and Ortiz de Montellano, P. R. (2009) Biochemistry 48, 863-872  191.  Kweon, O., Kim, S. J., Baek, S., Chae, J. C., Adjei, M. D., Baek, D. H., Kim, Y. C., and Cerniglia, C. E. (2008) BMC Biochem 9, 11  192.  Lee, K. (1999) J Bacteriol 181, 2719-2725  193.  Bernhardt, F. H., and Kuthan, H. (1981) Eur J Biochem 120, 547-555  194.  Denisov, I. G., Makris, T. M., Sligar, S. G., and Schlichting, I. (2005) Chem Rev 105, 2253-2277  195.  Luthra, A., Denisov, I. G., and Sligar, S. G. (2011) Arch Biochem Biophys 507, 26-35  170  196.  Hannemann, F., Bichet, A., Ewen, K. M., and Bernhardt, R. (2007) Biochim Biophys Acta 1770, 330-344  197.  Bloom, J. D., Labthavikul, S. T., Otey, C. R., and Arnold, F. H. (2006) Proc Natl Acad Sci U S A 103, 5869-5874  198.  Glieder, A., Farinas, E. T., and Arnold, F. H. (2002) Nat Biotechnol 20, 1135-1139  199.  Nazor, J., Dannenmann, S., Adjei, R. O., Fordjour, Y. B., Ghampson, I. T., Blanusa, M., Roccatano, D., and Schwaneberg, U. (2008) Protein Eng Des Sel 21, 29-35  200.  Szczebara, F. M., Chandelier, C., Villeret, C., Masurel, A., Bourot, S., Duport, C., Blanchard, S., Groisillier, A., Testet, E., Costaglioli, P., Cauet, G., Degryse, E., Balbuena, D., Winter, J., Achstetter, T., Spagnoli, R., Pompon, D., and Dumas, B. (2003) Nat Biotechnol 21, 143-149  201.  Vail, R. B., Homann, M. J., Hanna, I., and Zaks, A. (2005) J Ind Microbiol Biotechnol 32, 67-74  202.  Peters, F. T., Dragan, C. A., Kauffels, A., Schwaninger, A. E., Zapp, J., Bureik, M., and Maurer, H. H. (2009) J Anal Toxicol 33, 190-197  203.  Dietrich, J. A., Yoshikuni, Y., Fisher, K. J., Woolard, F. X., Ockey, D., McPhee, D. J., Renninger, N. S., Chang, M. C., Baker, D., and Keasling, J. D. (2009) ACS Chem Biol 4, 261-267  204.  Rittle, J., Younker, J. M., and Green, M. T. (2010) Inorg Chem 49, 3610-3617  205.  Nelson, D. R., Koymans, L., Kamataki, T., Stegeman, J. J., Feyereisen, R., Waxman, D. J., Waterman, M. R., Gotoh, O., Coon, M. J., Estabrook, R. W., Gunsalus, I. C., and Nebert, D. W. (1996) Pharmacogenetics 6, 1-42  206.  Nebert, D. W., Nelson, D. R., Coon, M. J., Estabrook, R. W., Feyereisen, R., FujiiKuriyama, Y., Gonzalez, F. J., Guengerich, F. P., Gunsalus, I. C., Johnson, E. F., and et al. (1991) DNA Cell Biol 10, 1-14  207.  Link, T. A. (1999) Advances in Inorganic Chemistry 47, 83-157  208.  Yoshiyama-Yanagawa, T., Enya, S., Shimada-Niwa, Y., Yaguchi, S., Haramoto, Y., Matsuya, T., Shiomi, K., Sasakura, Y., Takahashi, S., Asashima, M., Kataoka, H., and Niwa, R. (2011) J Biol Chem 286, 25756-25762  209.  Sydor, P. K., Barry, S. M., Odulate, O. M., Barona-Gomez, F., Haynes, S. W., Corre, C., Song, L., and Challis, G. L. (2011) Nat Chem 3, 388-392  171  210.  Dumitru, R., Jiang, W. Z., Weeks, D. P., and Wilson, M. A. (2009) J Mol Biol 392, 498-510  211.  D'Ordine, R. L., Rydel, T. J., Storek, M. J., Sturman, E. J., Moshiri, F., Bartlett, R. K., Brown, G. R., Eilers, R. J., Dart, C., Qi, Y., Flasinski, S., and Franklin, S. J. (2009) J Mol Biol 392, 481-497  212.  Oster, U., Tanaka, R., Tanaka, A., and Rudiger, W. (2000) Plant J 21, 305-310  213.  Resnick, S. M., Lee, K., and Gibson, D. T. (1996) Journal of Industrial Microbiology 17, 438-457  214.  Gibson, D. T., and Parales, R. E. (2000) Curr Opin Biotechnol 11, 236-243  215.  Prince, R. C., Lindsay, J. G., and Dutton, P. L. (1975) FEBS Lett 51, 108-111  216.  Ferraro, D. J., Gakhar, L., and Ramaswamy, S. (2005) Biochem Biophys Res Commun 338, 175-190  217.  Batie, C. J., Ballou, D. P., and Correll, C. C. (1991) Phthalate dioxygenase reductase and related flavin-iron-sulfur containing electron transferases. in Chemistry and Biochemistry of Flavoenzymes (Muller, F. ed.), CRC Press, Boca Raton, London. pp 543-556  218.  Summers, R. M., Louie, T. M., Yu, C. L., and Subramanian, M. (2011) Microbiology 157, 583-592  219.  Priefert, H., Rabenhorst, J., and Steinbuchel, A. (1997) J Bacteriol 179, 2595-2607  220.  Martins, B. M., Svetlitchnaia, T., and Dobbek, H. (2005) Structure 13, 817-824  221.  Gray, J., Wardzala, E., Yang, M., Reinbothe, S., Haller, S., and Pauli, F. (2004) Plant Mol Biol 54, 39-54  222.  Furukawa, K., Suenaga, H., and Goto, M. (2004) J Bacteriol 186, 5189-5196  223.  Pieper, D. H. (2005) Appl Microbiol Biotechnol 67, 170-191  224.  Kumamaru, T., Suenaga, H., Mitsuoka, M., Watanabe, T., and Furukawa, K. (1998) Nat Biotechnol 16, 663-666  225.  Sylvestre, M., Macek, T., and Mackova, M. (2009) Curr Opin Biotechnol 20, 242247  226.  Mohammadi, M., Viger, J. F., Kumar, P., Barriault, D., Bolin, J. T., and Sylvestre, M. (2011) J Biol Chem 286, 27612-27621  172  227.  Boyd, D. R., and Bugg, T. D. (2006) Org Biomol Chem 4, 181-192  228.  Wackett, L. (2002) Enzyme microb. tech. 31, 577-587  229.  Kauppi, B., Lee, K., Carredano, E., Parales, R. E., Gibson, D. T., Eklund, H., and Ramaswamy, S. (1998) Structure 6, 571-586  230.  Tarasev, M., Kaddis, C. S., Yin, S., Loo, J. A., Burgner, J., and Ballou, D. P. (2007) Arch Biochem Biophys 466, 31-39  231.  Ge, Y., Vaillancourt, F. H., Agar, N. Y., and Eltis, L. D. (2002) J Bacteriol 184, 4096-4103  232.  Parales, R. E., Emig, M. D., Lynch, N. A., and Gibson, D. T. (1998) J Bacteriol 180, 2337-2344  233.  Tarasev, M., Pinto, A., Kim, D., Elliott, S. J., and Ballou, D. P. (2006) Biochemistry 45, 10208-10216  234.  Pinto, A., Tarasev, M., and Ballou, D. P. (2006) Biochemistry 45, 9032-9041  235.  Murzin, A. G., Brenner, S. E., Hubbard, T., and Chothia, C. (1995) J Mol Biol 247, 536-540  236.  Friemann, R., Lee, K., Brown, E. N., Gibson, D. T., Eklund, H., and Ramaswamy, S. (2009) Acta Crystallogr D Biol Crystallogr 65, 24-33  237.  Furusawa, Y., Nagarajan, V., Tanokura, M., Masai, E., Fukuda, M., and Senda, T. (2004) J Mol Biol 342, 1041-1052  238.  Dong, X., Fushinobu, S., Fukuda, E., Terada, T., Nakamura, S., Shimizu, K., Nojiri, H., Omori, T., Shoun, H., and Wakagi, T. (2005) J Bacteriol 187, 2483-2490  239.  Karlsson, A., Parales, J. V., Parales, R. E., Gibson, D. T., Eklund, H., and Ramaswamy, S. (2003) Science 299, 1039-1042  240.  Friemann, R., Ivkovic-Jensen, M. M., Lessner, D. J., Yu, C. L., Gibson, D. T., Parales, R. E., Eklund, H., and Ramaswamy, S. (2005) J Mol Biol 348, 1139-1151  241.  Jakoncic, J., Jouanneau, Y., Meyer, C., and Stojanoff, V. (2007) FEBS J 274, 24702481  242.  Ferraro, D. J., Brown, E. N., Yu, C. L., Parales, R. E., Gibson, D. T., and Ramaswamy, S. (2007) BMC Struct Biol 7, 10  173  243.  Nojiri, H., Ashikawa, Y., Noguchi, H., Nam, J. W., Urata, M., Fujimoto, Z., Uchimura, H., Terada, T., Nakamura, S., Shimizu, K., Yoshida, T., Habe, H., and Omori, T. (2005) J Mol Biol 351, 355-370  244.  Inoue, K., Ashikawa, Y., Umeda, T., Abo, M., Katsuki, J., Usami, Y., Noguchi, H., Fujimoto, Z., Terada, T., Yamane, H., and Nojiri, H. (2009) J Mol Biol 392, 436-451  245.  Gakhar, L., Malik, Z. A., Allen, C. C., Lipscomb, D. A., Larkin, M. J., and Ramaswamy, S. (2005) J Bacteriol 187, 7222-7231  246.  Capyk, J. K., D'Angelo, I., Strynadka, N. C., and Eltis, L. D. (2009) J Biol Chem 284, 9937-9946  247.  Bugg, T. D., and Ramaswamy, S. (2008) Curr Opin Chem Biol 12, 134-140  248.  Chakrabarty, S., Austin, R. N., Deng, D., Groves, J. T., and Lipscomb, J. D. (2007) J Am Chem Soc 129, 3514-3515  249.  Ferraro, D. J., Okerlund, A. L., Mowers, J. C., and Ramaswamy, S. (2006) J Bacteriol 188, 6986-6994  250.  Yang, T. C., Wolfe, M. D., Neibergall, M. B., Mekmouche, Y., Lipscomb, J. D., and Hoffman, B. M. (2003) J Am Chem Soc 125, 2034-2035  251.  Ballou, D., and Batie, C. (1988) Prog Clin Biol Res 274, 211-226  252.  Cho, J., Jeon, S., Wilson, S. A., Liu, L. V., Kang, E. A., Braymer, J. J., Lim, M. H., Hedman, B., Hodgson, K. O., Valentine, J. S., Solomon, E. I., and Nam, W. (2011) Nature 478, 502-505  253.  Wolfe, M. D., Altier, D. J., Stubna, A., Popescu, C. V., Munck, E., and Lipscomb, J. D. (2002) Biochemistry 41, 9611-9626  254.  Yamazaki, Y., Kiuchi, M., Takeuchi, H., and Kubo, T. (2011) Insect Biochem Mol Biol 41, 283-293  255.  Rottiers, V., Motola, D. L., Gerisch, B., Cummins, C. L., Nishiwaki, K., Mangelsdorf, D. J., and Antebi, A. (2006) Dev Cell 10, 473-482  256.  Yoshiyama, T., Namiki, T., Mita, K., Kataoka, H., and Niwa, R. (2006) Development 133, 2565-2574  257.  Werlen, C., Kohler, H. P., and van der Meer, J. R. (1996) J Biol Chem 271, 40094016  174  258.  Nam, J. W., Nojiri, H., Yoshida, T., Habe, H., Yamane, H., and Omori, T. (2001) Biosci Biotechnol Biochem 65, 254-263  259.  Baek, S., Kweon, O., Kim, S. J., Baek, D. H., Chen, J. J., and Cerniglia, C. E. (2009) J Microbiol Methods 76, 307-309  260.  Altschul, S. F., Gish, W., Miller, W., Myers, E. W., and Lipman, D. J. (1990) J Mol Biol 215, 403-410  261.  Lebrun, E., Santini, J. M., Brugna, M., Ducluzeau, A. L., Ouchane, S., SchoeppCothenet, B., Baymann, F., and Nitschke, W. (2006) Mol Biol Evol 23, 1180-1191  262.  Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989) Molecular cloning : a laboratory manual, 2nd ed., Cold Spring Harbor Laboratory, Cold Spring Harbor, NY  263.  Pulleyblank, D., Michalak, M., Daisley, S. L., and Glick, R. (1983) Mol Biol Rep 9, 191-195  264.  Rosloniec, K. Z., Wilbrink, M. H., Capyk, J. K., Mohn, W. W., Ostendorf, M., van der Geize, R., Dijkhuizen, L., and Eltis, L. D. (2009) Mol Microbiol 74, 1031-1043  265.  Capyk, J. K., Casabon, I., Gruninger, R., Strynadka, N. C., and Eltis, L. D. (2011) J Biol Chem 286, 40717-40724  266.  Capyk, J. K., Kalscheuer, R., Stewart, G. R., Liu, J., Kwon, H., Zhao, R., Okamoto, S., Jacobs, W. R., Jr., Eltis, L. D., and Mohn, W. W. (2009) J Biol Chem 284, 3553435542  267.  Gomez-Gil, L., Kumar, P., Barriault, D., Bolin, J. T., Sylvestre, M., and Eltis, L. D. (2007) J Bacteriol 189, 5705-5715  268.  Vaillancourt, F. H., Han, S., Fortin, P. D., Bolin, J. T., and Eltis, L. D. (1998) J Biol Chem 273, 34887-34895  269.  Omura, T., and Sato, R. (1964) J Biol Chem 239, 2379-2385  270.  Faeder, E. J., and Siegel, L. M. (1973) Anal Biochem 53, 332-336  271.  Zabinski, R., Munck, E., Champion, P. M., and Wood, J. M. (1972) Biochemistry 11, 3212-3219  272.  Chen, J. S., and Mortenson, L. E. (1977) Anal Biochem 79, 157-165  273.  Jefcoate, C. R. (1978) Methods Enzymol 52, 258-279  274.  Jung, C., Ristau, O., and Rein, H. (1991) Biochim Biophys Acta 1076, 130-136  175  275.  Eltis, L. D., Karlson, U., and Timmis, K. N. (1993) Eur J Biochem 213, 211-216  276.  Alcalde, M. A., Antelo, A., Jover, A., Meijide, F., and Tato, J. V. (2009) J Incl Phenom Macrocycl Chem 63, 309-317  277.  Cornish-Bowden, A. (1994) Analysis of enzyme kinetic data, Oxford University Press, Oxford ; New York  278.  Emsley, P., and Cowtan, K. (2004) Acta Crystallogr D Biol Crystallogr 60, 21262132  279.  Holm, L., Kaariainen, S., Rosenstrom, P., and Schenkel, A. (2008) Bioinformatics 24, 2780-2781  280.  Gille, C., and Frommel, C. (2001) Bioinformatics 17, 377-378  281.  Larkin, M. A., Blackshields, G., Brown, N. P., Chenna, R., McGettigan, P. A., McWilliam, H., Valentin, F., Wallace, I. M., Wilm, A., Lopez, R., Thompson, J. D., Gibson, T. J., and Higgins, D. G. (2007) Bioinformatics 23, 2947-2948  282.  DeLano, W. L. (2002) The PyMOL Molecular Graphics System.  283.  Felsenstein, J. (1989) Cladistics 5, 164-166  284.  Wang, Y., Geer, L. Y., Chappey, C., Kans, J. A., and Bryant, S. H. (2000) Trends Biochem Sci 25, 300-302  285.  Russell, R. B., and Barton, G. J. (1992) Proteins 14, 309-323  286.  Roberts, E., Eargle, J., Wright, D., and Luthey-Schulten, Z. (2006) BMC Bioinformatics 7, 382  287. 288.  Gouy, M., Guindon, S., and Gascuel, O. (2010) Mol Biol Evol 27, 221-224 Edgar, R. C. (2004) Nucleic Acids Res 32, 1792-1797  289.  Huson, D. H., Richter, D. C., Rausch, C., Dezulian, T., Franz, M., and Rupp, R. (2007) BMC Bioinformatics 8, 460  290.  AJ, D., B, A., S, B., M, C., A, C., C, D., M, F., J, H., M, K., S, M., R, M., S, S.-H., S, S., T, T., and A, W. (2011) Available from  291.  Imbeault, N. Y., Powlowski, J. B., Colbert, C. L., Bolin, J. T., and Eltis, L. D. (2000) J Biol Chem 275, 12430-12437  176  292.  Morris, G. M., Goodsell, D. S., Halliday, R. S., Huey, R., Hart, W. E., Belew, R. K., and Olson, A. J. (1998) J. Comp. Chem. 19, 1639-1662  293.  Connolly, M. L. (1983) Science 221, 709-713  294.  Trott, O., and Olson, A. J. (2010) J Comput Chem 31, 455-461  295.  van der Geize, R., Hessels, G. I., Nienhuis-Kuiper, M., and Dijkhuizen, L. (2008) Appl Environ Microbiol 74, 7197-7203  296.  Petrusma, M., Hessels, G., Dijkhuizen, L., and van der Geize, R. (2011) J Bacteriol 193, 3931-3940  297.  Talavera, G., and Castresana, J. (2007) Syst Biol 56, 564-577  298.  Radauer, C., Lackner, P., and Breiteneder, H. (2008) BMC Evol Biol 8, 286  299.  Ouellet, H., Guan, S., Johnston, J. B., Chow, E. D., Kells, P. M., Burlingame, A. L., Cox, J. S., Podust, L. M., and de Montellano, P. R. (2010) Mol Microbiol 77, 730-742  300.  Norlin, M., von Bahr, S., Bjorkhem, I., and Wikvall, K. (2003) J Lipid Res 44, 15151522  301.  Motola, D. L., Cummins, C. L., Rottiers, V., Sharma, K. K., Li, T., Li, Y., SuinoPowell, K., Xu, H. E., Auchus, R. J., Antebi, A., and Mangelsdorf, D. J. (2006) Cell 124, 1209-1223  302.  Driscoll, M. D., McLean, K. J., Levy, C., Mast, N., Pikuleva, I. A., Lafite, P., Rigby, S. E., Leys, D., and Munro, A. W. (2010) J Biol Chem 285, 38270-38282  303.  Guengerich, F. P., Sohl, C. D., and Chowdhury, G. (2011) Arch Biochem Biophys 507, 126-134  304.  McLean, K. J., Lafite, P., Levy, C., Cheesman, M. R., Mast, N., Pikuleva, I. A., Leys, D., and Munro, A. W. (2009) J Biol Chem 284, 35524-35533  305.  Ouellet, H., Johnston, J. B., and Ortiz de Montellano, P. R. (2010) Arch Biochem Biophys 493, 82-95  306.  Johnston, J. B., Ouellet, H., Podust, L. M., and Ortiz de Montellano, P. R. (2011) Arch Biochem Biophys 507, 86-94  307.  Petrusma, M., Dijkhuizen, L., and van der Geize, R. (2012) J Bacteriol 194, 115-121  308.  Petrusma, M., Dijkhuizen, L., and van der Geize, R. (2009) Appl Environ Microbiol 75, 5300-5307  177  309.  Javitt, N. B. (2002) J Lipid Res 43, 665-670  310.  Chen, W., Chen, G., Head, D. L., Mangelsdorf, D. J., and Russell, D. W. (2007) Cell Metab 5, 73-79  311.  Hansson, M., Ellis, E., Hunt, M. C., Schmitz, G., and Babiker, A. (2003) Biochim Biophys Acta 1593, 283-289  312.  DuSell, C. D., Umetani, M., Shaul, P. W., Mangelsdorf, D. J., and McDonnell, D. P. (2008) Mol Endocrinol 22, 65-77  313.  Brzostek, A., Dziadek, B., Rumijowska-Galewicz, A., Pawelczyk, J., and Dziadek, J. (2007) FEMS Microbiol Lett 275, 106-112  314.  Monks, T. J., and Jones, D. C. (2002) Curr Drug Metab 3, 425-438  315.  Kabsch, W. (1993) J. Appl. Cryst. 26, 795-800  316.  Adams, P. D., Grosse-Kunstleve, R. W., Hung, L. W., Ioerger, T. R., McCoy, A. J., Moriarty, N. W., Read, R. J., Sacchettini, J. C., Sauter, N. K., and Terwilliger, T. C. (2002) Acta Crystallogr D Biol Crystallogr 58, 1948-1954  317.  McCoy, A. J. (2007) Acta Crystallogr D Biol Crystallogr 63, 32-41  318.  Terwilliger, T. C. (2000) Acta Crystallogr D Biol Crystallogr 56, 965-972  319.  Murshudov, G. N., Vagin, A. A., and Dodson, E. J. (1997) Acta Crystallogr D Biol Crystallogr 53, 240-255  320.  Brunger, A. T., Adams, P. D., Clore, G. M., DeLano, W. L., Gros, P., GrosseKunstleve, R. W., Jiang, J. S., Kuszewski, J., Nilges, M., Pannu, N. S., Read, R. J., Rice, L. M., Simonson, T., and Warren, G. L. (1998) Acta Crystallogr D Biol Crystallogr 54, 905-921  321.  Collaborative Computational Project, n. (1994) Acta Crystallogr D Biol Crystallogr 50, 760-763  322.  Schuttelkopf, A. W., and van Aalten, D. M. (2004) Acta Crystallogr D Biol Crystallogr 60, 1355-1363  323.  Mayo, S. L., Olafson, B. D., and Goddard, W. A. (1990) J. Phys. Chem. 94, 88978909  324.  Stewart, J. J. (2007) J Mol Model 13, 1173-1213  178  325.  Schneidman-Duhovny, D., Inbar, Y., Nussinov, R., and Wolfson, H. J. (2005) Nucleic Acids Res 33, W363-367  179  Appendices Appendix A Bacterial strains in the thesis  Genus Mycobacterium Mycobacterium Mycobacterium Mycobacterium Mycobacterium Mycobacterium Mycobacterium Mycobacterium Mycobacterium Mycobacterium Mycobacterium Mycobacterium Mycobacterium Mycobacterium Commamonas Rhodococcus Rhodococcus Rhodococcus  Species tuberculosis tuberculosis tuberculosis smegmatis avium leprae marinum africanum bovis bovis microi canetti pinnepetii caprae testosteroni erythropolis jostii rhodocrous  Strain H37Rv CD1551 Erdman  BCG  TA441 SQ1 RHA1 DSM43269  Closest relationship to Mtb H37Rv Mtb strain Mtb strain Mtb strain Mycobacterium Mycobacterium Mycobacterium Mycobacterium Mtb complex Mtb complex Mtb complex Mtb complex Mtb complex Mtb complex Mtb complex Actinomyete Actinomyete Actinomyete Actinomyete  180  Appendix B X-Ray crystallography data collection and stats B.1  Crystal growth and data collection  Crystals of KshA were grown aerobically at room temperature (18 °C) using the sitting drop method. Drops contained a 1:1 ratio of 290-400 !M KshA in 25 mM HEPES, pH 7.0 with 1 mM DTT and 0.25 mM FAS; and crystallization solution containing 1.1 M sodium malonate, 0.1 M HEPES, pH 7.0, and 0.5 % v/v Jeffamine M600. Single dark-red crystals appeared in 3-5 weeks and grew to their full size (200 !m x 300 x 50 !m) in approximately 3 months. Prior to data collection, KshA crystals were serially transferred between solutions of mother liquor supplemented with increasing amounts of ethylene glycol (5-20 %) and flash frozen in liquid N2. X-ray data collections were performed under cryogenic conditions using an in-house rotating anode X-ray generator (CuKa radiation " = 1.542 Å) and at the Canadian Light Source (CLS, Beamline CMCF1, " = 1.000 Å). Data were processed using XDS (315). Substructure solution and initial phasing were performed using the PHENIX program suite (316). The wavelength used (" = 1.542 Å) and redundant nature of the data allowed for the location of three iron and six sulfur sites using the HYSS (HYbrid Sub-structure Search) heavy atom search routine, accounting for one molecule in the asymmetric unit. Initial SAD phasing was performed using PHASER (317) (as implemented in PHENIX), enabling the calculation of a readily interpretable electron density map. The initial electron density map was then subjected to various cycles of density modification coupled to gradual phase extension and initial backbone tracing using the program Resolve (318). The resulting partial model was iteratively re-built using COOT (278) and the structure refined using REFMAC (319) as well as, for simulated annealing, CNS (320). Electron density maps were calculated using the FFT function of the CCP4 suite (321). The completed model was then refined using the higher resolution data acquired at the CLS. The coordinates and structure factors for M. tuberculosis KshA were deposited in the Brookhaven Protein Databank with accession code 2ZYL.  181  B.2  KshA crystallography statistics KshA (FeS SAD/SAS)a  KshA  X-ray source  Cu-K!  CLS CMCF1  Wavelength (Å)  1.542  1  Spacegroup  P321  P321  Unit cell (Å)  Diffraction Data  a=b=116.1, c=80.8  a=b=116.2, c=81.0  Resolution range (Å)  61-2.5  50-2.3  Highest shell (Å)  2.5-2.6  2.3-2.36  Total observations  86,930 (1,820)  204,950 (14,713)  Unique reflections  19,620 (827)  53,315 (3,841)  4.4 (2.2)  14.6 (3.1)  7.0 (39.1)  8.1 (43.5)  98.6 (97.4)  98.0 (95.1)  I / !I Rsym(%)  b  Completeness (%) Phasing Resolution range (Å) No. of used sites  20-2.6 3 Fe – 6 S  Figure of Merit (FOM)  0.26  FOM after density modification Refined Model  0.62  Resolution range (Å)  20-2.3  No. Reflections  26,909  Rfactor/Rfree (%)c  19/23  Total No. atoms:  3,100  Protein (No. atoms):  2,916  Solvent (No. atoms):  176  2  Mean B values (Å ) Protein  42.2  Non-Heme Fe  36.2  [2Fe2S]  32.1  rmsd Bond lengths (Å)  0.021  Bond angles (deg)  2.42  Iron-sulfur single wavelength anomalous diffraction. b Rsym = !h!iI(hkl) - "I(hkl)#/!h!iI(hkl) c Rwork = !$$Fo$ - $Fc$$/!$Fo$. Rfree is the Rwork value for 5% of the reflections excluded from the refinement. Data for the highest resolution shell are given in parentheses.  182  B.3  KshAB (2ZYL) metal-ligand distances  Metal-ligand distances (Å) Defining atom Fe-N!2, Asp181 Fe-N!2, His186 Fe-O"1, Asp304 Fe-O"2, Asp304 Fe-S1  Value 2.2 2.1 2.2 2.5 2.1  Metal-ligand angles (degrees) 181-Fe-86 181-Fe-O"1, Asp304 181-Fe-O"2, Asp304 186-Fe-O"1, Asp304 186-Fe-O"2, Asp304 O"1, Asp304 -Fe-O"2, Asp304 S1-Fe-81 S1-Fe-86 S1-Fe-O"1, Asp304 S1-Fe-O"2, Asp304  102 143 92 92 87 54 101 123 98 141  183  B.4  Docking of substrate molecular models in KshA crystal structure  ADD was designed using PRODRG (322), energy was minimized using the Dreiding force field (323), and PM6 semi-empirical charges were calculated using MOPAC2007 (324). After removing water molecules from the KshA model, docking simulations were performed using PATCHDOCK (325) and Autodock version 4.0 (292) with active site residues held rigid and the initial torsion angles of ADD randomly set. Structural figures and graphics were rendered using PYMOL (282). Models of 1,4-BNC and 1,4-BNC-CoA were generated using PRODRG (322) and were docked to KshA (PDB 2ZYL) with AutoDock Vina (294) using default parameters. A grid of 20 ! 24 ! 20 Å centered about the active site was defined as the searchable area for placement of the ligand. The top ten docked ligand conformations were output and the one with the lowest binding energy was chosen as the best structure.  184  Appendix C Proposed RO-O nomenclature system GI number 2822265 225698039  Enzyme name isopropyl benzene dioxygenase  557070  toluene dioxygenase chlorobenzene dioxygenase  151091  biphenyl dioxygenase  312597117  Organism Pseudomonas putida RE204 Pseudomonas putida F1  Previous Abbreviation  Classification code  IpbAaRE204  Roo1A1  TDOF1  Roo1A2  TcbAaP51  Roo1A3  BphAKF707  Roo1A4  biphenyl dioxygenase  Pseudomonas sp. P51 Pseudomonas pseudoalcaligenes KF707 Burkholderia xenovorans LB400  BPDOLB400  Roo1A5  1168640  benzene dioxygenase  Pseudomonas putida ML2  BedC1ML2  Roo1A6  3023401  biphenyl dioxygenase  Pseudomonas sp. KKS102  BphA1KKS102  Roo1A7  607172  biphenyl dioxygenase  Rhodococcus globerulus P6  BDOP6  Roo1A8  927232  biphenyl dioxygenase  BpdM5  Roo1A9  1354284  N/A  XylC1RB1  Roo1A10  295789305  BPDOB-356  Roo1A11  IBDOJR1  Roo1A12  3184044  biphenyl dioxygenase isopropyl benzene dioxygenase chlorobenzene dioxygenase ethylbenzene dioxygenase chlorobenzene dioxygenase  Rhodococcus sp. M5 Cycloclasticus oligotrophus RB1 Comamonas testosteroni B356  55670318  biphenyl dioxygenase  62738279 37651310  cumene dioxygenase dibenzothiophene dioxygenase  3059204  1685013 2264417  Pseudomonas sp. JR1 Burkholderia sp. PS12 Pseudomonas fluorescens CA4  TecA1PS12  Roo1A13  EBDOCA-4  Roo1A14  Ralstonia sp. JS705  McbAaJS705  Roo1A15  BPDORHA1  Roo1A16  CUMDOIP01  Roo1A17  DbdCa  Roo1A18  biphenyl dioxygenase  Rhodococcus jostii RHA1 Pseudomonas fluorescens IP01 Xanthobacter polyaromaticivorans Rhodococcus erythropolis TA421  BphATa421  Roo1A19  3426122  dioxin dioxygenase  Sphingomonas sp. RW1  DxnA1RW1  Roo1B1  137442  vanillate demthylase  Pseudomonas sp. ATCC19151  VanA19151  Roo2A1  3915234  Vanillate demethylase naphthalene dioxygenase  Pseudomonas sp. HR199 Pseudomonas putida NCIB 9816-4  VanAHr199  Roo2A2  NDO9816-4  Roo3A1  PAH dioxygenase 2,4-Dinitrotoluene dioxygenase 2-nitrotoluene 2,3dioxygenase gentisate 1,2dioxygenase nitrobenzene dioxygenase polyaromatic hydrocarbon oxygenase  Pseudomonas putida OUS82  PahAcOUS82  Roo3A2  Burkholderia sp. DNT  DntAcDNT  Roo3A3  Pseudomonas sp. JS42  NtdAcJS42  Roo3A4  NagIU2  Roo3A5  Comamonas sp. JS765  NBDOJS765  Roo3A6  Burkholderia sp. RP007  PhnAcRP007  Roo3B1  4105709  28948321 391844 1477923 1773277 2828018 67464649  3820519  Ralstonia sp. U2  185  GI number 33456990  Enzyme name  Organism  Previous Abbreviation  Classification code  PhnA1A5  Roo3C1  RHDOCHY-1  Roo3C2  BPDOB1  Roo3C3  Cycloclasticus sp. A5  126030179  PAH dioxygenase ring hydroxylating dioxygenase  145579344  biphenyl dioxygenase  Sphingobium yanoikuyae B1  158346890  N/A  Sphingomonas sp. LH128  PhnA1fLH128  Roo3C4  94481140  PAH-dioxygenase  Sphingomonas sp. A4  ArhA1A4  Roo3D1  196166905  N/A  Acidovorax sp. NA3  PhnAcNA3  Roo3E1  21431747  benzoate dioxygenase 2-halobenzoate 1,2dioxygenase anthranilate dioxygenase Phthalate dioxygenase  Acinetobacter sp. ADP1  BenAADP1  Roo4A1  Burkholderia cepacia 2CBS  CbdA2CBS  Roo4A2  Acinetobacter sp. ADP1 Pseudomonas putida  AntAADP1 Pht3  Roo4B1 Roo5A1  Burkholderia cepacia DB01  PDODB01  Roo5A2  Arabidopsis thaliana  Paothaliana  Roo6A1  Oryza sativa Japonica Group  Paosativa  Roo6A2  TftA1AC1100  Roo7A1  PobAPOB310  Roo8A1  2073550  Phthalate dioxygenase Pheophorbide a oxygenase Pheophorbide a oxygenase 2,4,5trichlorophenoxyacetic acid oxygenase phenoxybenzoate dioxygenase 3-chlorobenzoate 3,4dioxygenase  CbaABR60  Roo9A1  1263180  cumate dioxygenase  Pseudomonas putida F1  CmtAbF1  Roo10A1  1841362  aniline oxygenase  Pseudomonas putida UCC22  TdnA1UCC22  Roo11A1  1395141  analine dioxygenase toluenesulfonate methylmonooxygenase oxyquinoline monooxygenase Carbazole 1,9adioxygenase Carbazole 1,9adioxygenase Carbazole 1,9adioxygenase diphenylamine dioxygenase Carbazole 1,9adioxygenase Carbazole 1,9adioxygenase Carbazole 1,9adioxygenase Phenanthrene dioxygenase  Acinetobacter sp. YAA  AtdAYAA  Roo11B1  Comamonas testosteroni T-2  TsaM2T-2  Roo12A1  Pseudomonas putida 86  OMO86  Roo13A1  Pseudomonas sp. CA10  CARDOCA10  Roo14A1  Pseudomonas stutzeri OM1  CARDOOM1  Roo14A2  Janthinobacterium sp. J3  CARDOJ3  Roo14A3  Burkholderia sp. JS667 Nocardioides aromaticivorans IC177  DpaAaJS667  Roo14B1  CARDOIC177  Roo14C1  Sphingomonas sp. KA1  CARDOKA1  Roo14D1  Sphingomonas sp. CB3  CARDOCB3  Roo15A1  Nocardioides sp. KP7  PhdAKP7  Roo16A1  758210 3511232 3914350 4128221 41688605 108706168  508289 473250  13661652 67463926 2317678 3293058 75765412 226935179 258588294 292659576 3243167 5931577  Sphingomonas sp. CHY-1  Burkholderia cepacia AC1100 Pseudomonas pseudoalcaligenes POB310 Comamonas testosteroni BR60  186  GI number  27657409 68053509 116805452  77543327 120401546 4585359  Enzyme name PAH ringhydroxylating dioxygenase polyaromatic ring hydroxylase ring-hydroxylating dioxygenase PAH ringhydroxylating dioxygenase Ring hydroxylating dioxygenase  78101541  N/A naphthalene dioxygenase  38524454  N/A  11038552  pyrene dioxygenase PAH ringhydroxylating dioxygenase ring-hydroxylating dioxygenase ring-hydroxylating dioxygenase  27657415 33333858 33333865 32562912 4406507 4455072  76781905 13242054 51338947 27531093 49072886 3643998 21309823 187472112 224510546 324106141  biphenyl dioxygenase ortho-halobenzoate 1,2-dioxygenase diterpenoid dioxygenase 2aminobenzenesulfonate dioxygenase Phthalate 3, 4 dioxyenase Phthalate 3, 4 dioxyenase Phthalate 3, 4 dioxyenase Phthalate 3, 4 dioxyenase 2-hydroxybenzoate 5hydroxylase 3-ketosteroid-9-alphahydroxylase 3-ketosteroid-9-alphahydroxylase 3-ketosteroid-9-alphahydroxylase 3-ketosteroid-9-alphahydroxylase  Organism  Previous Abbreviation  Classification code  Mycobacterium sp. 6PY1 Mycobacterium vanbaalenii PYR-1  Pdo26PY1  Roo16A2  PAHPYR-1  Roo16A3  Mycobacterium sp. SNP11  PdhASNP11  Roo16A4  Terrabacter sp. FLO Mycobacterium vanbaalenii PYR-1  ARHFLO  Roo16A5  RHDOPYR-1  Roo16A6  NidAI24  Roo16B1  NDO12038  Roo16B2  NarAaSAO101  Roo16B3  NidAPYR-1  Roo16C1  Mycobacterium sp. 6PY1  Pdo16PY1  Roo16C2  Mycobacterium sp. S65  NidAS65  Roo16C3  Rhodococcus sp. I24 Rhodococcus sp. NCIMB12038 Rhodococcus opacus strain SAO101 Mycobacterium vanbaalenii PYR-1  Mycobacterium sp. S65  PdoAS66  Roo16C4  Bacillus sp. JF8  BphAJF8  Roo16D1  Pseudomonas aeruginosa 142 Pseudomonas abietaniphila BKME-9  OhbB142  Roo17A1  DitABKME-9  Roo18A1  Alcaligenes sp. O-1  AbsAaO-1  Roo19A1  Arthrobacter keyseri 12B  PDO12B  Roo20A1  Rhodococcus jostii RHA1  PDORHA1  Roo10A2  Terrabacter sp. DBF63 Mycobacterium vanbaalenii PYR-1  PDODBF63  Roo20A3  PhtAaPYR-1  Roo20A4  HybBJB2  Roo21A1  KshA1SQ1  Roo22A1  KshA2SQ1  Roo22A2  KshAH37Rv  Roo22A3  KshA143269  Roo22A4  Pseudomonas aeruginosa JB2 Rhodococcus erythropolis SQ1 Rhodococcus erythropolis SQ1 Mycobacterium tuberculosis H37Rv Rhodococcus rhodochrous DSM 43269  187  GI number  Organism Rhodococcus rhodochrous DSM 43269 Rhodococcus rhodochrous DSM 43269 Rhodococcus rhodochrous DSM 43269 Rhodococcus rhodochrous DSM 43269  15219408  Enzyme name 3-ketosteroid-9-alphahydroxylase 3-ketosteroid-9-alphahydroxylase 3-ketosteroid-9-alphahydroxylase 3-ketosteroid-9-alphahydroxylase dibenzofuran dioxygnease terephthalate 1,2dioxygenase oxygenase salicylate 1hydroxylase alpha subunit anthranilate dioxygenase terephthalate 1,2dioxygenase oxygenase ring hydroxylating dioxygenase ethylbenzene dioxygenase choline monooxygenase choline monooxygenase chloriphillide a oxygenase chlorophillide a oxygenase chlorophillide a oxygenase chlorophillide a oxygenase  41469406  N/A  30686966  N/A  126722226  biphenyl dioxygenase ring-hydroxylating dioxygenase ring-hydroxylating dioxygenase Dicamba odemethylase dicamba monooxygenase aminopyrrolnitrin oxidase PrnD Abnormal Dauer formation 36  324106148 324106151 324106157 324106163 22036072  22779305  52207945 29568950  78210747 23330204 35764412 297725227 30688478 15224625 9944967 110289597  33333862 33333869 55584974 254574850 68345267 71983709  Terrabacter sp. YK3  Delftia tsuruhatensis T7  Sphingomonas sp. CHY-1 Burkholderia cepacia DB01  Previous Abbreviation  Classification code  KshA243269  Roo22A5  KshA343269  Roo22A6  KshA443269  Roo22A7  KshA543269  Roo22A8  DfdAYK3  Roo23A1  TerZT7  Roo24A1  PhnA1bCHY-1  Roo24B1  AndAcDB01  Roo24C1  Comamonas sp. strain E6 Sphingopyxis macrogoltabida TFA  TphA2E6  Roo24C2  ThnA1TFA  Roo25A1  Rhodococcus jostii RHA1  EtbA1RHA1  Roo25A2  Oryza sativa Japonica Group  CMOsativa  Roo26A1  Arabidopsis thaliana  CMOthaliana  Roo26B1  Arabidopsis thaliana  TIC55thaliana  Roo27A1  Oryza sativa Japonica Group  LLS1sativa  Roo28A1  Oryza sativa Japonica Group  Caosativa  Roo28A2  Arabidopsis thaliana  Caothaliana  Roo28A3  Oryza sativa Japonica Group  PTC52sativa  Roo29A1  Arabidopsis thaliana Rhodococcus rhodochrous K37  PTC52thaliana  Roo29B1  BphAK37  Roo30A1  Mycobacterium sp. S65  NidXS65  Roo31A1  Mycobacterium sp. S65 Stenotrophomonas maltophilia DI-6  PdoXS65  Roo31A2  DdmCDi-6  Roo32A1  Stenotrophomonas maltophilia  DMO  Roo32A2  Pseudomonas fluorescens Pf-5  PrnDPf-5  Roo33A1  Daf-36elegans  Roo34A1  Caenorhabditis elegans  188  GI number  Enzyme name  Organism  95007733  Neverland  Bombyx mori  317176037  Neverland  Apis mellifera  Previous Abbreviation  Classification code  Neverlandmori Neverlandmellife  Roo35A1  ra  Roo35B1  Neverlandpulcher 336088229  Neverland  Hemicentrotus pulcherrimus  rimus  Roo35C1  Neverlandintestin 336088225  Neverland  Ciona intestinalis  alis-1  Roo35D1  Neverlandintestin 336088227  Neverland  Ciona intestinalis  336088223  Neverland  Danio rerio  336088221  Neverland  Xenopus laevis  158028612  Drosophila melanogaster Methylosulfonomonas methylovora M2  gaster  Roo36A1  MsmAM2  Roo37A1  195973390  Neverland methanesulfonate monooxygenase (hypo)xanthine hydroxylase  Klebsiella oxytoca M5al  HpxDM5al  Roo38A1  297787799  N/A  Ruegeria sp. TM1040  putativeTM1040  Roo39A1  2815311  N/A  Streptomyces coelicolor A3(2)  RedGA3(2)  Roo40A1  4633080  alis-2  Roo35E1  Neverlandrerio  Roo35F1  Neverlandlaevis Neverlandmelano  Roo35F2  189  


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items