UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Iron transport in two pathogenic Gram-negative bacteria Chan, Anson Chi-Kit 2011

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

Item Metadata


24-ubc_2011_spring_chan_anson.pdf [ 4.08MB ]
JSON: 24-1.0071636.json
JSON-LD: 24-1.0071636-ld.json
RDF/XML (Pretty): 24-1.0071636-rdf.xml
RDF/JSON: 24-1.0071636-rdf.json
Turtle: 24-1.0071636-turtle.txt
N-Triples: 24-1.0071636-rdf-ntriples.txt
Original Record: 24-1.0071636-source.json
Full Text

Full Text

 IRON TRANSPORT IN TWO PATHOGENIC GRAM-NEGATIVE BACTERIA  by  Anson Chi-Kit Chan B. Sc. The University of British Columbia, 2004  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY  in  THE FACULTY OF GRADUATE STUDIES (Microbiology and Immunology)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  January 2011  ? Anson Chi-Kit Chan, 2011 ii  ABSTRACT Campylobacter jejuni and Escherichia coli strain F11 are two Gram-negative pathogens with a versatile armament of iron uptake systems to cope with the fluctuating host nutrient environment. Our current understanding of Gram-negative iron uptake systems focuses heavily on a prototypical scheme involving a TonB-dependent outer membrane receptor and an ABC transporter, with little knowledge on systems that do not fall neatly into this paradigm. The primary focus of this thesis is the characterization of three such atypical iron uptake proteins from C. jejuni (ChaN and P19) and pathogenic E. coli (FetP). C. jejuni ChaN is a 30 kDa, iron-regulated lipoprotein hypothesized to be involved in iron uptake. The crystal structure of ChaN reveals that it can bind two cofacial heme groups in a pocket formed by a ChaN dimer. Each heme iron is coordinated by a single tyrosine from one monomer and the propionate groups are hydrogen bonded by a histidine and a lysine from the other monomer. Analytical ultracentrifugation studies demonstrate heme-dependent dimerization in solution. Cell fractionation of C. jejuni shows that ChaN is localized to the outer membrane. Based on these findings, the predicted in vivo role of ChaN in iron uptake is discussed. C. jejuni cFtr1-P19 and E. coli FetMP are homologous iron-regulated systems also proposed to be iron transporters. Through growth studies in both organisms, we show that P19 and FetMP are required for optimal growth under iron-limited conditions.  Furthermore, metal binding analysis demonstrates that recombinant P19 and FetP bind both copper and iron. Dimerization of P19 is shown to be metal dependent in vitro and is detected in vivo by cross-linking. Through x-ray crystallography, we have determined the structures of P19 and FetP with various metals bound, thus revealing the locations of the highly conserved copper and iron binding sites. Additionally, the crystal structure of FetP reveals two copper positions in each iii  binding site that is likely functionally important. Through mutagenesis, residues contributing to the alternative copper positions were identified. Together, these studies provide insight into the mechanism of iron transport by the two systems and allow for the development of functional models.   iv  PREFACE  The majority of work presented in this thesis is drawn from either published literature or a submitted manuscript. This thesis would not have been possible without the collaborative input for the conducted research presented here. Below is a description of the contributions made by fellow scientists and collaborators.  Chapter 3  The structural, spectroscopic, analytical ultracentrifugation and sequence analyses were published as: Chan A.C., Lelj-Garolla B., Rosell F.I., Pedersen K.A., Mauk A.G. and Murphy M.E.P. (2006) Cofacial heme binding is linked to dimerization by a bacterial heme transport protein. J Mol Biol 362, 1108-1119.  I cloned chaNR from Campylobacter genomic DNA, purified and crystallized the protein, determined the structure of ChaN and completed the bioinformatic analysis. The protein expression vector was subcloned from my construct by K. A. Pedersen. I prepared the protein samples and provided assistance for the spectroscopic and analytical ultracentrifugation studies, which were led by Drs. F. I. Rosell and B. Lelj-Garolla, respectively. I wrote the first draft of the manuscript with contributions from Drs. F. I. Rosell and B. Lelj-Garolla for their respective sections in the methods, results and discussion. Drs. A. G. Mauk and M. E. Murphy edited the manuscript. The localization and C. jejuni growth studies also presented in Chapter 3 is part of my unpublished work. I produced the constructs used in the unpublished sections and performed the experiments.    v  Chapter 4  The majority of this chapter was published as: Chan A.C., Doukov T.I., Scofield M., Tom-Yew S.A., Ramin A.B., Mackichan J.K., Gaynor E.C. and Murphy M.E.P. (2010) J Mol Biol 401, 590-604.  Dr. T. I. Doukov and I share first authorship of this paper. The expression vector was cloned by M. Scofield. M. Scofield, A. B. Ramin and I purified the protein used in biochemical studies. The SeMet, ?as isolated? and MnCu-P19 crystals were produced by M. Scofield. The apo-P19 crystal was produced A. B. Ramin. The crystal structures of these samples were solved by Dr. T. I. Doukov. Subsequently, I prepared the Cu-P19, CuFe-P19(ox) and CuFe-P19(red) crystals and solved their structures (the latter two are not published). I completed the bioinformatic analysis and performed the iron-regulated protein expression and protein dimerization studies. A. B. Ramin determined the copper-binding capacity of apo-P19. I performed all other metal binding assays. The C. jejuni P19 deletion strain was produced by J. K. Mackichan and I produced the P19 complemented strain. I led the C. jejuni iron-limited growth studies with assistance from Dr. S. A. Tom-Yew, who developed the modified protocol used in this growth study, and Marlo Firme. I performed the dual copper-iron chelation C. jejuni growth experiment. All C. jejuni growth experiments were conducted in the laboratory of Dr. E. C. Gaynor. Introduction to working C. jejuni was provided by Dr. E. C. Gaynor and members of her lab, including Heather Candon, Sarah Svensson and Meghan McLennan. I wrote the first draft of the manuscript, which was edited by Drs. T. I. Doukov, E. C. Gaynor and M. E. Murphy.    vi  Chapter 5  The majority of this chapter is drawn from a prepared manuscript that has been submitted as: Koch D., Chan A.C., Murphy M.E.P., Lilie H., Grass G. and Nies D.H. (2010) Characterization of a novel dipartite iron-uptake system from uropathogenic Escherichia coli strain F11.  All constructs and strains were produced by D. Koch, who also performed all growth, RNA and iron uptake assays. I solved all of the crystal structures, which includes the two active site mutant structures that are not included in the manuscript. D. Koch and I wrote our respective parts for the methods and results. I wrote the discussion presented in this thesis, which has been edited by Dr. M. E. Murphy, but it draws upon the submitted manuscript discussion section written by Dr. D. Nies and edited by Dr. M. E. Murphy, Dr. G. Grass and me.   vii  TABLE OF CONTENTS  Abstract ........................................................................................................................................ ii Preface......................................................................................................................................... iv Table of Contents ....................................................................................................................... vii List of Tables .............................................................................................................................. xi List of Figures ............................................................................................................................ xii List of Symbols and Abbreviations........................................................................................... xiv Acknowledgements .................................................................................................................. xvii Dedication ............................................................................................................................... xviii CHAPTER 1. Introduction ........................................................................................................1 1.1. The evolution of life alongside metals ....................................................................... 1 1.2. Iron ............................................................................................................................. 1 1.2.1. The role of iron and its general availability ..................................................... 1 1.2.2. Regulation of uptake to maintain homeostasis ................................................ 2 1.2.3. Iron acquisition systems in Gram-negative bacteria ........................................ 3 1.2.4. Siderophore-mediated iron uptake ................................................................... 4 1.2.5. Uptake of iron from host glycoproteins ........................................................... 7 1.2.6. Uptake of iron as heme or from hemoproteins ................................................ 8 1.2.7. Ferrous iron uptake ........................................................................................ 10 1.3. Copper ...................................................................................................................... 11 1.3.1. The role of copper and its availability ........................................................... 11 1.3.2. Copper transport systems ............................................................................... 13 1.4. Campylobacter jejuni ............................................................................................... 14 1.4.1. A fastidious food-borne pathogen ................................................................. 14 1.4.2. C. jejuni iron uptake regulation ..................................................................... 17 1.4.3. C. jejuni ferrous iron uptake .......................................................................... 18 1.4.4. C. jejuni ferric siderophore uptake ................................................................ 20 1.4.5. C. jejuni and heme ......................................................................................... 21 1.4.6. The C. jejuni ChaNR system ......................................................................... 21 1.4.7. The C. jejuni P19-cFtr1 system ..................................................................... 23 1.5. Uropathogenic Escherichia coli ............................................................................... 25 1.5.1. Iron and heme uptake systems of UPEC versus K-12 ................................... 27 1.6. Objectives of this thesis ........................................................................................... 28 CHAPTER 2. Materials and methods ......................................................................................30 viii  2.1. Bacterial strains and growth conditions ................................................................... 30 2.1.1. E. coli ............................................................................................................. 30 2.1.2. C. jejuni.......................................................................................................... 30 2.2. Strain construction ................................................................................................... 32 2.2.1. C. jejuni chaNR deletion mutant .................................................................... 32 2.2.2. C. jejuni p19 deletion mutant......................................................................... 37 2.2.3. Complementation of 81176?p19 ................................................................... 37 2.2.4. E. coli ............................................................................................................. 38 2.3. Growth phenotype studies........................................................................................ 38 2.3.1. C. jejuni ChaNR............................................................................................. 38 2.3.2. C. jejuni P19 .................................................................................................. 39 2.3.3. E. coli ............................................................................................................. 39 2.4. Recombinant protein expression .............................................................................. 41 2.4.1. C. jejuni ChaN ............................................................................................... 41 2.4.2. C. jejuni P19 .................................................................................................. 42 2.4.3. E. coli FetP..................................................................................................... 44 2.4.4. E. coli His-tagged FetP mutants E46Q and M90I ......................................... 45 2.5. ChaN spectroscopic heme and pH titration ............................................................. 46 2.6. Anti-ChaN and anti-P19 polyclonal antibody production ....................................... 47 2.7. C. jejuni protein localization studies ........................................................................ 47 2.7.1. Subcellular fractionation ................................................................................ 47 2.7.2. Cell surface digestion .................................................................................... 48 2.8. Crystallization and structure determination ............................................................. 49 2.8.1. C. jejuni ChaN ............................................................................................... 49 2.8.2. C. jejuni P19 .................................................................................................. 50 2.8.3. E. coli FetP..................................................................................................... 52 2.9. Protein dimerization studies ..................................................................................... 54 2.9.1. Analytical ultracentrifugation analysis of ChaN ........................................... 54 2.9.2. P19 ................................................................................................................. 56 2.10. FetP mRNA quantification ...................................................................................... 57 2.11. Sequence analyses .................................................................................................... 58 CHAPTER 3. Cofacial heme binding is linked to dimerization by C. jejuni ChaN ................59 3.1. Introduction .............................................................................................................. 59 3.2. Results ...................................................................................................................... 60 ix  3.2.1. Overall structure ............................................................................................ 60 3.2.2. Spectroscopic analyses .................................................................................. 65 3.2.3. Analytical ultracentrifugation ........................................................................ 67 3.2.4. Sequence analyses ......................................................................................... 70 3.2.5. ChaN is localized to the outer membrane ...................................................... 72 3.2.6. A ChaNR deletion mutant exhibits delayed growth ...................................... 74 3.3. Discussion ................................................................................................................ 76 CHAPTER 4. Copper binding and iron transport by C. jejuni P19 .........................................84 4.1. Introduction .............................................................................................................. 84 4.2. Results ...................................................................................................................... 85 4.2.1. The expression of P19 protein is iron-regulated ............................................ 85 4.2.2. P19 is required for optimal C. jejuni growth under iron restriction .............. 86 4.2.3. P19 binds both iron and copper at different sites .......................................... 89 4.2.4. The overall structure of P19 .......................................................................... 91 4.2.5. The two metal binding sites are co-localized ................................................ 95 4.2.6. The primary metal binding sites are preformed ............................................. 99 4.2.7. P19 shares structural homology with other copper binding proteins ............ 99 4.2.8. Residues in P19 involved in metal binding are conserved .......................... 100 4.2.9. P19 dimer formation upon metal chelation ................................................. 101 4.2.10. Ferric and ferrous iron binding by Cu-P19 ................................................ 103 4.3. Discussion .............................................................................................................. 106 CHAPTER 5. Copper binding and iron transport by uropathogenic E. coli FetP .................113 5.1. Introduction ............................................................................................................ 113 5.2. Results .................................................................................................................... 114 5.2.1. The FetMP system is involved in iron homeostasis .................................... 114 5.2.2. FetMP contributes to growth under acidic and basic conditions ................. 116 5.2.3. FetP is a periplasmic copper-binding protein .............................................. 118 5.2.4. Overall structure of FetP .............................................................................. 119 5.2.5. The FetP copper binding site ....................................................................... 124 5.2.6. The copper binding site is preformed .......................................................... 129 5.2.7. Glu46 and Met90 contributes to the dual copper positions ......................... 129 5.3. Discussion .............................................................................................................. 131 CHAPTER 6. Overview and future directions ......................................................................137 6.1. A regulatory role for ChaN? .................................................................................. 138 x  6.2. Building a P19 functional model from work involving two pathogens ................. 139 6.3. Future directions .................................................................................................... 141 6.3.1. ChaN ............................................................................................................ 141 6.3.2. P19 and FetP ................................................................................................ 143 References .................................................................................................................................146    xi  LIST OF TABLES Table 2-1. Bacterial strains used in this study .............................................................................. 31 Table 2-2. Bacterial plasmids and clones used in this study ......................................................... 33 Table 2-3. Primers used in this study ............................................................................................ 35 Table 3-1. Data collection and refinement statistics for ChaN ..................................................... 61 Table 4-1. Data collection and refinement statistics for P19 ........................................................ 92 Table 4-2. Metal ligand geometry in the primary site of P19 and Tp34a ..................................... 96 Table 4-3. Data collection and refinement statistics for iron-soaked Cu-P19 structures ............ 104 Table 5-1. Data collection and refinement statistics for FetP ..................................................... 120 Table 5-2. Ligand bond lengths in the primary site of Cu-FetP ................................................. 127   xii  LIST OF FIGURES Figure 1-1. Schematic representation of Gram-negative iron uptake systems ............................... 5 Figure 1-2. The principal copper homeostatic mechanisms in E. coli .......................................... 13 Figure 1-3. The iron uptake systems of C. jejuni .......................................................................... 19 Figure 1-4. Schematic representation of members from the Ftr1 superfamily ............................. 23 Figure 3-1. The crystal structure of heme-bound ChaN ............................................................... 62 Figure 3-2. Stereo representation of the ChaN-heme interactions ................................................ 63 Figure 3-3. Heme and pH titration effects on ChaN spectra ......................................................... 66 Figure 3-4. Sedimentation velocity analysis of ChaN .................................................................. 68 Figure 3-5. Unrooted tree of ChaN homolog sequences............................................................... 71 Figure 3-6. ChaN is likely localized to the outer membrane ........................................................ 73 Figure 3-7. The chaNR deletion mutant exhibits delayed growth ................................................ 75 Figure 4-1. Iron-regulated P19 expression is important for growth under increasing iron limitation ........................................................................................................................... 85 Figure 4-2. The P19 deletion mutant is unable to grow under iron-limited conditions ................ 87 Figure 4-3. Growth of the P19 deletion mutant is less affected by copper chelation than wild-type C. jejuni ............................................................................................................................. 89 Figure 4-4. The overall dimeric structure of the P19 monomer with bound copper .................... 94 Figure 4-5. ?As isolated? and copper-reconstituted structures depicting the copper site ligands and the flexibility of Glu44 ............................................................................................... 95 Figure 4-6. Manganese-soaked structure ...................................................................................... 97 Figure 4-7. A multiple sequence alignment of representative P19 homologues from different genera demonstrating the conservation of copper and putative iron ligands .................. 101 xiii  Figure 4-8. Cross-linking studies of wild-type C. jejuni demonstrate detection of P19 dimer in vivo .................................................................................................................................. 103 Figure 4-9. Iron-soaked structure under reducing and oxidizing conditions .............................. 105 Figure 5-1. FetMP is an iron uptake system ............................................................................... 115 Figure 5-2. FetMP stimulates growth of ECA458-fetMP at various pH values ......................... 117 Figure 5-3. General structure of the Cu-FetP dimer with bound copper .................................... 122 Figure 5-4. Metal detection in the "as isolated" FetP crystal by fluorescence............................ 122 Figure 5-5. The FetP copper binding site reveals multiple copper positions.............................. 125 Figure 5-6. The copper ligands in each conformation ................................................................ 126 Figure 5-7. A putative third metal binding site ?CuC? exists adjacent to CuB .......................... 128 Figure 5-8. The copper binding site of FetP mutants E46Q and M90I....................................... 130 Figure 6-1. Alignment of the Cu-FetP and CuFe-P19 structure active sites .............................. 140   xiv  LIST OF SYMBOLS AND ABBREVIATIONS  ?   Angstrom (1 ? = 0.1 nm) ABC transporter ATP-binding cassette transporter AHT   Anhydrotetracycline ATP   Adenosine triphosphate BCA   Bicinchoninic acid BCS   Bathocuproine disulfonic acid Bis-Tris  Bis(2-hydroxyethyl)-amino-tris(hydroxymethyl)-methane B-factor  Crystallographic thermal factor CAPS   N-cyclohexyl-3-aminopropanesulfonic acid CAT, CmR  Chloramphenicol resistance cassette cd1NiR  Cytochrome cd1 nitrite reductase CD   Circular dichroism CDTA   Trans-1,2-Cyclohexanediaminetetraacetic Acid CFU   Colony forming units ChaNR  ?Campylobacter heme-associated? system  CHES   2-(N-cyclohexylamino) ethane sulfonic acid Ci   Curie (1 Ci = 3.7?1010 decays per second) c(s)    Sedimentation coefficient distribution  DF   Desferrioxamine DSP   Dithiobis (succinimidyl propionate) EDTA   Ethylenediamine tetra-acetic acid FetMP   Iron uptake system in Escherichia coli xv  Ferene S  3-(2-Pyridyl)-5,6-bis(5-sulfo-2-furyl)-1,2,4-triazine, disodium salt hydrate Fur   Ferric uptake regulator GmR   Gentamicin resistance cassette Hemin   Fe3+-protoporphyrin IX Heme Fe2+-protoporphyrin IX; often used as a general term for an unspecified iron oxidation state His6   Poly-histidine affinity purification tag HSA   Human serum albumin IM   Inner membrane I-PCR   Inverse PCR IPTG   Isopropyl-b-D-thiogalactopyranoside KanR    Kanamycin resistance cassette Kd    Dissociation constant LB   Luria-Bertani LF   Lactoferrin MAD   Multi-wavelength anomalous dispersion MALDI-TOF  Matrix-assisted laser desorption/ionization-time of flight mass spectrometry MCO   Multicopper oxidase MEM?   Minimal essential media-alpha formulation MES   Morpholinoethanesulfonic acid MH   Mueller-Hinton MH-TV  Mueller-Hinton media containing trimethoprim and vancomycin xvi  MOPS   3-(N-morpholino)propanesulfonic acid MWCO  Molecular weight Cut-off NB   Native buffer OD600   Optical density at 600 nm wavelength OM   Outer membrane PBP   Periplasmic binding protein PBS   Phosphate buffered saline PDB   Protein data bank PEG   Polyethylene glycol PMSF   Phenylmethylsulfonyl fluoride R.m.s.d.  Root mean square deviation Sarkosyl   Sodium lauroyl sarcosinate  SAD   Single wavelength anomalous dispersion SDS-PAGE  Sodium dodecyl sulfate polyacrylamide gel electrophoresis SeMet   Selenomethionine SSRL   Stanford Synchrotron Radiation Lightsource TAPS   N-Tris(hydroxymethyl)methyl-3-aminopropanesulfonic Acid TEA   Triethanolamine TF   Transferrin TMM   Tris-buffered mineral salts medium Tris   Tris(hydroxymethyl)aminomethane UV   Ultraviolet 2?YT    Yeast tryptone media xvii  ACKNOWLEDGEMENTS Funding of this work has been provided through a Natural Sciences and Engineering Research Council (NSERC) Canada Postgraduate Scholarship and a NSERC Undergraduate Student Research Award. This research was also funded by Canadian Institute of Health Research Grants to Dr. Michael Murphy.  I would like to thank Dr. Michael Murphy for his encouragement, guidance and supervision over the years, which made this such an invaluable opportunity. I am also extremely grateful to all of my committee members, Drs. Erin Gaynor, Bob Hancock and John Smit for their advice and support. Many growth experiments critical to my thesis have been performed in the lab of Dr. Erin Gaynor, who not only provided space and equipment, but also instruction and guidance. Members of her lab, especially Heather Candon, Sarah Svensson and Emilisa Frirdich, have been very helpful during this time. I would also like to thank Dr. John Smit, Dr. John Nomellini and Janny Lau for providing advice and assistance on top of sharing lab supplies from time to time. I am very grateful to our collaborators Dr. Dietrich Nies, Dr. Gregor Grass and Doreen Koch for fostering such an interesting and rewarding relationship for the FetMP work and Drs. Grant Mauk, Federico Rosell and Barbara Lelj-Garolla for providing their equipment and expertise that led to my first published work.  Of course, I also have to thank all of the past and present members of the Murphy lab for always providing assistance, sharing knowledge or cheering up when an experiment goes wrong and for making the lab such a wonderful place to work in. I am grateful to my friends who have provided much emotional support during the stressful times. Last, and definitely not least, I would like to thank my family for their unconditional love and support during these many years.  xviii  DEDICATION       To my parents:  Howie & Karman1  CHAPTER 1. INTRODUCTION 1.1. THE EVOLUTION OF LIFE ALONGSIDE METALS Even prior to the advent of the use of metal tools by mankind, life on earth has depended on the chemical functions made possible by incorporating metals into biological macromolecules. The most common metals used to perform critical metabolic reactions are iron, copper and zinc. However, zinc only has one physiologically available oxidation state and is therefore redox inactive; iron and copper, on the other hand, have reduction potentials that make them suitable for many biological processes (1). The usage of both metals in biological systems may have ties to the geological history of the planet as the two metals have experienced periods of differing bioavailability (2). Prior to photosynthesis and the subsequent oxygenation of the earth, ferrous iron bioavailability was dominant over cuprous copper due to higher solubility of the former. As the planet shifted from primarily anaerobic to oxygen-rich, oxidation of the surface-exposed metals increased the solubility of copper while decreasing that of iron, effectively allowing organisms to also integrate copper into chemical processes. The abundance, biochemical usefulness and historical availability of these two metals may explain why they are still heavily utilized in biological systems today.  1.2. IRON  1.2.1. THE ROLE OF IRON AND ITS GENERAL AVAILABILITY Almost all organisms require iron. The properties of iron, such as its reduction potential (spanning approximately from ?500 to +300 mV depending on the proteinaceous ligand environment) and interchangeable oxidation states (Fe2+, Fe3+ and Fe4+), makes it an essential cofactor in many cellular processes such as electron transport and oxidative stress defense (3). 2  Iron is abundant, but is not always available in free form. Although ferrous iron (Fe2+) is the more soluble form, it quickly becomes oxidized to its ferric form (Fe3+) unless in an anaerobic or acidic environment. The concentration of free ferric iron has been calculated to be ~10-18 M at pH 7 in an aerobic environment (3). Therefore, this level of free iron means that some form of active transport is required to meet the iron needs of bacteria, which are 105-106 ions per cell per generation (3).  1.2.2. REGULATION OF UPTAKE TO MAINTAIN HOMEOSTASIS In the presence of oxygen, iron can generate reactive oxygen species that can damage biological macromolecules (4). Therefore, expression of the iron uptake systems must be down-regulated when exogenous iron is abundant as a means to prevent excessive accumulation of intracellular iron. The ferric uptake regulator (Fur) protein is an iron sensor and effector system found in both Gram-negative and several Gram-positive bacteria. It was first identified in a mutant of Salmonella typhimurium that had modified expression of iron uptake systems (5). This was followed by its identification in Escherichia coli, where a fur mutant was rapidly complemented and the fur gene cloned and sequenced (6). The important contribution of Fur to establishing an infection has been demonstrated for many pathogens in various animal models, including Staphylococcus aureus (7), Helicobacter pylori (8) and Campylobacter jejuni (4). The mechanism by which Fur detects iron availability is through the requirement of iron to form a functional Fur complex. When intracellular ferrous iron is present in amounts sufficient for growth, Fur forms a dimer with Fe2+ and binds to a consensus sequence (Fur box) in the promoter region of iron-regulated proteins near the transcriptional start site. This effectively 3  hinders the binding of RNA polymerase and prevents the transcription of the iron uptake genes (9).  Identification of the Fur regulon in both Gram-negative and Gram-positive organisms has revealed many genes involved in iron homeostasis. The expression of a few proteins unrelated to iron transport but that contribute to pathogenesis have also been shown to be regulated by Fur, such as the E. coli Shiga toxins SltA and SltB (10) and hemolysin (11). In some bacteria, Fur also regulates the expression of proteins that confer resistance to oxidative stress, such as catalase, hydroperoxidases, and superoxide dismutase (SOD) (6). Fur can indirectly upregulate gene expression under iron-replete conditions through the repression of small regulatory RNA expression, such as RyhB in E. coli (12). Recently, it has been shown that Fur can also act directly as a positive regulator in certain organisms, such as H. pylori (6). In all of these cases where direct binding of Fur has been characterized, the Fur box was found to be located far upstream from the transcriptional start site. This simple change in the location of Fur binding is believed to reverse typical RNA polymerase binding inhibition and instead promote transcription. To make matters even more complicated, regulation by Fur when iron is not bound (apo-Fur) has also been demonstrated in H. pylori and has been suggested for organisms such as C. jejuni (4,6). The mechanism by which this occurs is still poorly understood.   1.2.3. IRON ACQUISITION SYSTEMS IN GRAM-NEGATIVE BACTERIA During infection, pathogenic bacteria face an environment in which the host sequesters iron in protein complexes to inhibit bacterial growth while mobilizing iron for its own uses. Host proteins such as transferrin (TF) in serum and lactoferrin (LF) in secretory fluids (and in serum at sites of infection) are generally found as a partially saturated population such that they are ready 4  to compete for any freshly introduced iron (13,14). Within cells, ferritin is a key iron storage protein complex to maintain low levels of intracellular iron (15). Together, these proteins play key roles in overall host health as an innate host defense system, as injecting iron into animal hosts has been shown to increase the virulence of invading pathogens (16). However, bacteria have acquired the means to scavenge for host iron (and cause disease) through specialized transport systems that take up a range of different iron-containing compounds.  1.2.4. SIDEROPHORE-MEDIATED IRON UPTAKE Ferric-binding complexes are too large to pass through the outer membrane (OM) porins of Gram-negative bacteria and therefore require specific transport systems. These systems typically consist of an OM receptor and an ABC transport complex consisting of a periplasmic substrate binding protein, a permease and an ATPase (Figure 1-1) (17).  The E. coli OM receptors for ferrichrome (FhuA), ferric dicitrate (FecA) and enterochelin (FepA) have had their crystal structures determined (18). Although the amino acid sequence identity between these receptors can be as low as 20%, there is a remarkable consistency in their design. They all feature a ?-barrel shape consisting of 22 45? anti-parallel strands and are filled by an N-terminal plug domain (18,19). These features appeared to be conserved amongst other TonB-dependent OM receptors as well, as they were also observed in the structures of FpvA (P. aeruginosa pyoverdine receptor), FptA (P. aeruginosa pyochelin receptor), and BtuB (E. coli vitamin B12 receptor) (18,20,21). In the folded structure of these OM receptors, the cork (or plug) domains are located inside the ?-barrel domains and act as its namesake. Due to the tight packing of the cork domain within the ?-barrel scaffold, the cork is predicted to either form a transient 5  channel or partially move out of the ?-barrel in order for substrate to pass through (19). Although the cork domain does interact with the substrate, specificity appears to be highly dependent on   Figure 1-1. Schematic representation of Gram-negative iron uptake systems The prototypical iron uptake system in Gram-negative bacteria consists of an outer membrane (OM) receptor and an ABC transporter (made up of a periplasmic binding protein and an inner membrane permease and ATPase). Following this general scheme are systems for the transport of siderophore-iron complexes (blue), heme and heme-complexes (red) and iron bound to transferrin (green). In addition, heme transport systems may include an additional protein component (hemophore) that is able to extract heme from hemoproteins prior to its delivery to the OM receptor. The transport of iron from transferrin also utilizes an additional surface-located lipoprotein component (TbpB). The energy used by the outer membrane receptor for all three transporter types is provided by the TonB complex (cyan), which consists of TonB, ExbB and ExbD.   6  the long protruding loop structures that extend up to 40 ? from the outer surface of the ?-barrel as removing or swapping the cork domains of FepA and FhuA neither reverses nor weakens substrate specificity (19,22). In the sequences of E. coli FecA and P. aeruginosa FpvA, an additional 80 residue regulatory feature is found at the N-terminus. This sequence, which is termed the N-terminal signaling domain, functions in the autoregulation of FecA and FpvA expression. When ferric-citrate is bound to FecA, for example, a conformational change in the signaling domain induces FecR, an IM anti sigma factor, to release cytoplasmic FecI, a ? factor that increases the expression of the respective uptake system. Due to the porous nature of the OM, no potential gradient is available to provide energy for transport by the OM receptor. Instead, this energy is transduced from the proton motive force across the inner membrane (IM) through the TonB-ExbB-ExbD complex (Figure 1-1). How this complex actually achieves such an intermembrane spanning feat is still unclear but current experimental evidence favors a ?shuttle? model, involving the conversion of TonB to a ?charged? state by ExbD/ExbB (23). This ?charged? conformation of TonB then traverses to the OM receptor, recognizing the target through a ?TonB Box? consensus sequence located in the plug domain of the receptor. It then discharges this stored energy to promote ligand internalization and then returns to the IM ExbD/ExbB complex for the cycle to be repeated once again. Periplasmic binding proteins (PBPs) belonging to the PBP-dependent ABC transport system are numerous due to their versatility in transporting a wide range of substrates (24). Examination of the known structures within this family reveals a highly conserved two-domain architecture even though the sequence identities can be as low as 10% between members of this family (19). One distinguishing feature between members is the composition of the hinge region 7  between the two lobe domains, which together functions to clamp onto substrates akin to a clam or a Venus-fly trap. The hinge either consists of two or three anti-parallel ?-strands (such as for the ferric iron transporter FbpA) or a long ?-helix (which includes all siderophore transport PBPs). The IM components of the ABC transport system consists of two permease domains and two ATPase domains, all of which may be fused in various ways, reducing the total number of polypeptides to form the functional unit (19). Each permease domain consists of 10 transmembrane helices, which together forms a gated channel across the IM. Rearrangement of these two domains during substrate transport is energized by ATP hydrolysis at the ATPase domains. Based on structures of BtuCD-F, the E. coli vitamin B12 transport complex, the docking of the PBP with the ABC transporter involves interactions between a highly conserved Glu residue from BtuF (PBP) and an Arg from BtuCD (ABC) (25).   1.2.5. UPTAKE OF IRON FROM HOST GLYCOPROTEINS Within the host environment, proteins used by the host to mobilize and sequester iron can be an excellent source of iron. OM receptors specific for host glycoproteins such as transferrin and lactoferrin have been characterized in organisms such as Neisseria spp. (26) and H. influenzae (27). As transferrin and lactoferrin are 80 kDa proteins, they are too large to pass through the OM intact and instead, have their bound iron ions removed for transport across the membrane. This binding and extraction process is mediated by two proteins: an OM receptor with large surface loops to interact with transferrin (TbpA) or lactoferrin (LbpA) and an OM surface located lipoprotein (TbpB/LbpB) that may act as the initial substrate binding site (Figure 1-1) (19). Although in vitro mutants of TbpB have variable deficiencies in growth when 8  transferrin is the sole iron source (28,29), TbpB can independently bind transferrin without the presence of TbpA (30) and has also been shown to preferentially bind the iron-bound form of transferrin (31). Recently, the structure of TbpB from Actinobacillus pleuropneumoniae has been determined, which features a bilobed architecture (32). Although the two lobes are structurally similar overall, the N-terminal lobe has an affinity for iron-loaded transferrin similar to that of the full length protein whereas the C-terminal lobe has no significant affinity for transferrin, likely due to the surface electrostatic differences between the two lobes. No structures are currently available for the OM receptor components.   Once the iron enters the periplasm, an ABC transport system consisting of FbpA, a periplasmic binding protein (PBP) and FbpBC (IM transporter) is responsible for shuttling Fe3+ ions into the cytoplasm. FbpA is a two ?-sheet hinge PBP that shares a similar fold with each of the two lobes of transferrin and lactoferrin (19). Upon substrate binding, the more flexible apo-form of FbpA can undergo a 20? rotation of the two domains along the ?-sheet hinge to clamp down on the Fe3+ ion (33,34).  1.2.6. UPTAKE OF IRON AS HEME OR FROM HEMOPROTEINS Heme can be an important source of iron for pathogenic organisms as heme-containing proteins are abundant inside the host. For intracellular pathogens, heme is readily available. For extracellular pathogens, host cells must be triggered to release the heme, usually through cellular damage. The released heme binds to host serum proteins such as albumin and hemopexin. Hemoglobin, released from spontaneous hemolysis in human plasma (35,36), is quickly and tightly bound by haptoglobin, a glycoprotein that is present in sufficient amounts in adult serum to prevent the accumulation of free hemoglobin (36). Heme-loaded albumin, hemopexin and 9  hemoglobin-haptoglobin is normally transported to the liver to be recycled, but can also be used by invading pathogens. For Gram-negative bacteria, these heme-containing complexes are recognized by either a TonB-dependent OM receptor or a secreted hemophore, which delivers the heme protein to a surface receptor (Figure 1-1). At the OM, the heme is generally believed to be extracted and transported intact into the periplasm. Examples of characterized hemophore-independent heme or hemoprotein receptors include PhuR from P. aeruginosa, HgpABC from H. influenzae, and HpuAB from Neisseria meningitidis (37). The best characterized hemophore dependent system is has from Serratia marcescens, which is also present in P. aeruginosa and Yersinia spp. where the hemophore HasA works in conjunction with the OM receptor HasR for heme uptake (38). HasR alone also can take up free heme and hemoglobin-bound heme, albeit with lower affinity than for HasA-heme. The only other type of hemophore characterized so far is HxuA from H. influenzae, which has been shown to bind heme-hemopexin and has been reported to be both secreted and bound the cell surface (39). However, the actual OM receptor for HxuA is not known, but may be HxuC.  Similar to siderophore receptors, an ABC transporter then ferries the heme into the cytoplasm. The two best characterized heme-specific PBPs are ShuT from Shigella dysenteriae and PhuT from P. aeruginosa, whose recent crystal structures show remarkable similarity to the siderophore PBPs, having two globular domains connected by an ?-helix (40). Alternatively, the periplasmic heme molecule may be directly incorporated as cofactors into proteins such as respiratory-chain components, but this has yet to be demonstrated definitively. As free heme is toxic, the heme that reaches the cytoplasm may be first sequestered by a ShuS/PhuS analog (41,42). This heme is then degraded by a heme oxygenase (HO) to release the iron, such as P. aeruginosa pa-HO (PigA) (43).  10   1.2.7. FERROUS IRON UPTAKE As iron in the ferrous oxidation state is highly soluble and thus does not require chelators for uptake, it can pass through the OM much more easily than ferric complexes (likely through porins).  No specific OM receptors for Fe2+ have been identified. Perhaps the best studied ferrous transporter is the anaerobically-induced FeoABC system, which was first characterized in E. coli (44). FeoB, which is found in around 50% of sequenced bacterial genomes, is in most investigated cases the major bacterial ferrous iron transporter (45). It contains a GTPase domain whose function is necessary for Fe2+ transport and an integral IM domain that likely functions as the Fe2+ permease (46). In around 80% of the sequenced genomes containing FeoB is the gene for FeoA, which is a small, cytoplasmic protein of unknown function but carries a Src Homology 3 (SH3) domain. SH3 domains are best characterized in eukaryotic proteins for their role in mediating protein-protein interactions. FeoC-like proteins are only found in ?-proteobacteria and may be transcriptional repressors. Binding protein-dependent ABC transport systems have also been shown to be ferrous iron transporters and act similarly to the Feo system without the need for specific OM receptors. However, the specificity of these systems tends to be broader, with the ability to transport other divalent metals as well (and generally with a higher affinity than iron). The SitABCD system of Salmonella typhimurium is a manganese and iron transporter that is encoded on the Salmonella pathogenicity island 1 and is required for full virulence in mice (47,48). This system was originally thought to be a major ferrous iron transporter as it was regulated by Fur and could transport iron. However, it was subsequently shown to have around 30 to 100 times higher 11  affinity for manganese and also contains a putative manganese responsive element in the promoter region (47). The constitutively expressed CorA family of IM magnesium transporters has been shown to be rather nonspecific and can transport other divalent cations, including iron (49). Bacteria also have a homolog of the eukaryotic Nramp family of proteins (MntH), which are proton-dependent, high-affinity divalent metal transporters (50). Ferrous iron and manganese are the two best characterized substrates of the mammalian Nramp systems, although they are able to transport other cations as well (47). Two other systems that were previously thought to be only found in eukaryotes were discovered and characterized recently in E. coli. These transporters include the ZIP (ZRT, IRT-like protein) family ZupT protein and the EfeUOB system. ZupT is a transmembrane permease that has been shown to have a broad divalent metal uptake spectrum (51). EfeUOB is a low pH-induced, tripartite ferrous uptake system. EfeU is a permease of the Ftr1 superfamily that has been implicated in copper-dependent iron uptake in S. cerevisiae and other fungi. EfeO contains a cupredoxin-like domain that may play a role in electron transfer (52). EfeB is a periplasmic heme-containing peroxidase that promotes the extraction of iron from heme while preserving the tetrapyrrole ring (deferrochelatase) (53).   1.3. COPPER 1.3.1. THE ROLE OF COPPER AND ITS AVAILABILITY  Copper is like iron in many ways: it is another key metal that is essential to many organisms, is an abundant element in the environment and has two easily interchangeable oxidation states that make it both an excellent cofactor for biological systems and potentially toxic by participating in Fenton chemistry (54). This redox cycling capability of copper allows it 12  to participate in the same cellular processes as iron, such as electron transport and oxidative stress defense. However, there are other characteristics of copper that set it apart from iron. The two biologically relevant oxidation states of copper are Cu+ and Cu2+. Whereas the reduction of iron from Fe3+ to Fe2+ increases solubility, the reduction of copper from Cu2+ to Cu+ changes copper to a less soluble form.  To date, no bacterial copper-dependent protein has been observed to be localized in the cytoplasm (55). The only documented cases where copper is transported through the cytosol are in Enterococcus hirae (56), Listeria monocytogenes (57) and cyanobacteria (55). Although specific copper importers have been characterized in E. hirae and L. monocytogenes, neither bacterium has a known requirement for copper in the cytosol (54). In cyanobacteria, cytosolic copper is shuttled into the thylakoid, a specialized compartment with copper-requiring photosynthetic and respiratory electron transport proteins. All other characterized copper proteins are localized either to the periplasm or the cytoplasmic membrane.  Copper availability to invading pathogens within the host environment is still poorly understood. For the typical human body, the average daily intake of copper is 1 to 3 mg (58). Much of the dietary copper is first absorbed in the duodenum, the first section of the small intestine, for its subsequent distribution to the rest of the body. To maintain copper homeostasis, however, liver hepatocytes also transport excess copper into bile through the action of the ATP7B ATPase, which re-enters the duodenum for excretion from the body (58). This constant influx and efflux of copper through the host is likely a major contributing factor to a pathogen?s copper demands.  13  1.3.2. COPPER TRANSPORT SYSTEMS As almost all bacteria have no intracellular copper requirement, copper transporters are generally aimed at avoiding copper toxicity (54). This has been best characterized in E. coli, where the detection of intracellular copper by CueR results in the expression of CopA, a P1B-type ATPase, to pump copper out of the cytoplasm (Figure 1-2) (59). All copper-transporting P-type ATPases, including the P1B subgroup, have been suggested to function in cytoplasmic copper efflux (60). In the periplasm, Cu+ is converted to the less toxic Cu2+ form by a multicopper oxidase (CueO). If the level of copper in the periplasm reaches toxic levels, CusR/S activates the    Figure 1-2. The principal copper homeostatic mechanisms in E. coli In the cytoplasm, CueR (not shown) is a copper sensor that triggers the expression of CueO and CopA. CopA transports Cu+ into the periplasm where CueO oxidizes Cu+ into less toxic Cu2+. CueO contains the twin-arginine motif and is exported as a prefolded protein via the TAT transporter. Green circles are used to represent copper ions. At excess periplasmic copper levels, the CusR/S system (not shown) activates the expression of CusCFBA, which transports copper across the OM.  14  expression of the CusCFBA system, which pumps copper into the extracellular milieu. CusCBA forms a complex that traverses both the IM and OM with CusC as a TolC-like OM protein, CusB as an IM-anchored, periplasm-spanning membrane fusion protein and CusA as an IM proton antiporter (54). CusF is a small, 10 kDa metallochaperone that binds one Cu+ ion in the periplasm and transfers it to CusB (61,62). In some strains of E. coli, an additional plasmid-borne copper detoxification system (pcoABCDRSE) has been characterized that allows survival in high copper environments in which the chromosomally-encoded copper detoxification systems (CopA-CueO and CusCFBA) would have been overwhelmed (63).  PcoC is a periplasmic copper binding protein likely acting as a copper chaperone. Similar to CueO, PcoA is a periplasmic multicopper oxidase (MCO) localized to the periplasm by the TAT pathway and can catalyze the oxidation of Cu+ bound to PcoC, which are then predicted to interact with OM PcoB for export of the copper (64). PcoD is predicted to chaperone copper for the maturation of apo-PcoA.  1.4. CAMPYLOBACTER JEJUNI 1.4.1. A FASTIDIOUS FOOD-BORNE PATHOGEN Campylobacter jejuni is a Gram-negative, microaerophilic and capnophilic, enteric pathogen of the epsilon proteobacteria class. Within this class, C. jejuni falls under the order Campylobacterales, which includes the genera Helicobacter and Wolinella, all of which feature small genomes (1.6-2.0 Mb) and the ability to have persistent, asymptomatic interactions with hosts (65). C. jejuni, for example, generally resides as a commensal in most domesticated animals, especially poultry, which is an important food source. C. jejuni is highly infectious to humans, who can become exposed during the consumption of undercooked meats, untreated milk or fecal matter-contaminated water. With a reported infectious dose as low as 500 CFU, C. jejuni 15  infection can lead to acute gastroenteritis with symptoms such as fever, abdominal pain and diarrheal illness that generally resolves in a week, but up to a fifth may have a relapse or a prolonged illness (66-68). These symptoms can vary from watery to bloody diarrhea (and stools), with a higher frequency of watery diarrhea in less developed countries and a higher frequency of bloody excretions in developed nations (68). Although infections with C. jejuni are rarely fatal for immunocompetent adults and are generally self-limiting, the extended duration of debilitating symptoms along with the high incidence rates can be a substantial burden on the economy. C. jejuni infection can also lead to more serious sequelae, as it is also the primary antecedent pathogen leading to the autoimmune neurodegenerative disease Guillain-Barr? syndrome (69). With an estimated 2.5 million cases per year in the United States alone (70), C. jejuni is recognized as a leading cause of foodborne diarrheal disease in the developed world (71) and accounts for up to 14% of cases worldwide (72). Successful evasion of the immune system, colonization of the intestinal epithelium and the induction of disease by C. jejuni is dependent on many factors. A contributing factor is the high motility of C. jejuni due to its spiral shape and polar flagella, which are important during penetration of the viscous intestinal mucus layer (73). C. jejuni also displays chemotactic behavior towards mucin, the primary component of mucus (74) and energy sources such as pyruvate (i.e. energy taxis) (75). The ability to invade human cell lines has been shown to be dependent on the capsule, proteins such as CiaB and FlaC that require the flagellum for secretion, and also adhesins such as the fibronectin-binding protein CadF, the surface-exposed lipoprotein JlpA and PBP-like Peb1 (65). The flagellum, capsule and lipooligosaccharide (LOS) surface structures of C. jejuni exhibits extensive genetic variation, which would assist in immune evasion during host colonization (76). Genetic variation is due in part to the lack of clear DNA-16  repair genes and the presence of homopolymeric tracts within regions involved in the biosynthesis and modification of these three surface carbohydrate-containing structures (65,77). C. jejuni is also naturally transformable and readily (and preferentially) takes up DNA from other C. jejuni strains in the environment (78). Taken together, a given population of C. jejuni can be quite diverse due to phase variation, gene duplications and deletions, frameshifts, point mutations, and horizontal gene transfer. Sequencing of multiple C. jejuni strains has also revealed surprising diversity in the presence or absence of additional plasmids (79) and chromosomal genes involved in areas such as energy metabolism, restriction modification, and nutrient uptake (including iron) that likely contribute to virulence (80,81). Two sequenced C. jejuni clinical isolates, NCTC 11168 and the highly invasive strain 81-176, are used in the studies presented in this thesis.  Although ex vivo experiments demonstrate that C. jejuni can adhere to and invade intestinal epithelial cells of both human and avian origin (82), the colonization outcome can be  dramatically different depending on the host species. C. jejuni has only been observed in the mucus layer with no signs of cell attachment or invasion during histological examinations of infected chick intestinal tissues (83,84). During infection of susceptible human hosts, however, it is generally believed that C. jejuni invades host cells since inflammation and bacteremia can result during infection (85) and invasion has been observed during infections of macaque monkeys (86) as well as human cell lines (87). The disparity of outcomes between the colonization of poultry as compared to humans is not fully understood, but a recent study suggests that factors specific to the mucus layer of chickens are able to inhibit epithelial cell invasion since human-derived mucus exhibited no such effect (82).  17  Whether the host is avian or human in nature, C. jejuni must be able to compete against host iron sequestration and the iron acquisition systems of other host flora for this essential metal. This is accomplished through the expression of multiple specific iron-uptake systems, which function together to exploit the many forms of iron encountered, ranging from the free ferrous form to ferric iron bound by a variety of small molecular weight chelators (siderophores), by proteins such as transferrin or lactoferrin, or by protoporphyrin IX as heme (88). Although C. jejuni can utilize siderophores produced by other organisms (xenosiderophores), genomic analyses have shown that most strains of Campylobacter do not produce any of the known bacterial siderophores themselves (89,90).   1.4.2. C. JEJUNI IRON UPTAKE REGULATION Through multiple studies, the genes that are regulated by iron have been extensively examined. Palyada et al. identified the iron-regulated genes in wild-type and fur-deficient C. jejuni NCTC 11168 utilizing DNA microarrays (4). Iron availability was shown to affect the wild-type transcription levels of 647 genes out of the predicted ~1650 open reading frames (4,77), a dramatic metabolic shift that supports the significance of this single nutrient. By comparing the transcriptome of fur mutants to wild-type C. jejuni, the Fur regulon was revealed to include at least 53 genes, many of which were not previously known to be regulated by Fur. Of the 29 genes whose expression was repressed by both iron and Fe2+-Fur, 17 encoded putative iron uptake genes. Confusingly, five genes involved in flagellum biogenesis were seen to be iron repressed but also apo-Fur activated, suggesting additional levels of regulation. Complementing this study nicely is a transcriptomic and proteomic analysis of wild-type and fur-deficient C. jejuni NCTC 11168 published at around the same period by Holmes et al, who reported similar 18  findings (91). Contrary to earlier findings (92), both of these studies do not observe Fur autoregulation. Interestingly, the C. jejuni genome also encodes a second Fur homolog. Sharing 37% sequence identity, PerR is the major regulator of the responses to oxidative stresses, regulating ~75% of the genes that were stimulated by exposure to cumene hydroperoxide, menadione or hydrogen peroxide (93). Unlike Fur, PerR appears to be upregulated directly or indirectly by iron and H2O2 but not through the action of Fur. Not surprisingly, due to the sequence similarity of Fur and PerR and the intimate ties between iron and oxidative stress, Fur was found to regulate ~30% of the oxidative stress stimulon.  1.4.3. C. JEJUNI FERROUS IRON UPTAKE The residence of C. jejuni in the oxygen-reduced gastrointestinal tract of endothermic animals means that much of the encountered iron may be in the reduced ferrous state, and is therefore likely an important source of iron. Although ferrous iron can diffuse through the OM porins of Gram-negative bacteria, a transport mechanism is required for iron to cross the cytoplasmic membrane. The Feo system in C. jejuni consists of a FeoB homolog and a FeoA-like protein (Figure 1-3). Interestingly, there are FeoB sequence differences amongst the various C. jejuni strains (94). For example, in strain 81-176, a premature stop codon has occurred, resulting in approximately half the length of a typical FeoB sequence. In other strains, frameshift mutations have occurred. Although earlier work suggests that FeoB is not involved in C. jejuni ferrous iron uptake (95), a subsequent study with strain 81-176 has demonstrated that even the truncated Campylobacter FeoB plays a major role in ferrous iron usage (96). Although ferrous iron uptake by the CorA family of magnesium transporters has been shown for organisms such   19    Figure 1-3. The iron uptake systems of C. jejuni Schematic representation of the C. jejuni iron uptake systems with known substrates indicated. A question mark is used to indicate predicted substrates or unknown pathways. CeuE is an atypical PBP with the addition of a lipidation site and is membrane anchored. CfhuA is found only in a small number of C. jejuni strains and the predicted substrate ferrichrome has not been experimentally demonstrated. Iron transport by the ChaNR system has been suggested to involve the cFbpABC system, but has yet to be demonstrated. In the work presented in this thesis, we demonstrate that ChaN is an OM heme binding protein likely involved in iron transport. We also demonstrate that periplasmic P19 is involved in iron uptake but the oxidation state of the cation is currently unknown.    20  as E. coli, S. typhimurium, and H. pylori (49,97), C. jejuni CorA is not regulated by iron (4,91) and its involvement in ferrous iron has yet to be demonstrated (95).   1.4.4. C. JEJUNI FERRIC SIDEROPHORE UPTAKE Genomic sequencing of multiple C. jejuni strains suggests that it generally does not produce any of the known siderophores (77,98). However, C. jejuni does contain genes that confer the ability to use exogenous siderophores secreted by other organisms (or xenosiderophores) (Figure 1-3). Through feeding assays, it has been shown that some Campylobacters can grow on the fungal hydroxamate siderophore ferrichrome and the catechol-type enterochelin of enteric bacteria as their sole iron sources (89).  The gene encoding a putative ferrichrome OM receptor, designated cfhuA, was identified in C. jejuni strain M129, a human clinical isolate that has not been sequenced (99). CfhuA shares 33% identity with the E. coli ferrichrome receptor FhuA and has been shown to be upregulated under iron-limited growth conditions but has yet to be shown to take up ferrichrome directly. In most fully sequenced Campylobacter strains, cfhuA is absent and therefore the contribution of this gene to host colonization is unknown.  The enterochelin receptor in C. jejuni has been recently identified as CfrA through gene deletion studies and has been shown to be important for avian intestinal colonization by strain 11168 (4). The cfrA gene has been identified in most of the sequenced strains of C. jejuni, but not in 81-176 (94). In Campylobacter coli, the ceuBCDE operon encodes a putative ABC transport system. Mutation of each of the components, including the atypical lipidated PBP CeuE, is required for efficient enterochelin transport and is believed to work in conjunction with the OM receptor CfrA (94,100). Although ceuBCDE is also found in C. jejuni and is iron 21  regulated, it might not be strictly essential for the use of enterochelin since a C. jejuni CeuE knockout has only a minor defect on enterochelin-dependent growth (4). Interestingly, homologs from related genera may have differing functions as the CeuE homolog in Helicobacter is not regulated by iron, but by nickel and is involved in nickel and cobalt uptake (101,102).  1.4.5. C. JEJUNI AND HEME Many strains of C. jejuni are invasive and can enter or cross the host intestinal epithelial layer where xenosiderophores are unavailable. Heme-containing foods are also commonly part of our diets. Therefore, heme acquisition systems are likely important during colonization of the intestinal lumen and post-invasion. C. jejuni can utilize heme, hemopexin-heme, hemoglobin and haptoglobin-hemoglobin as an iron source (9). Whether C. jejuni can utilize human serum albumin (HSA) as a heme source is not known. In C. jejuni, the ChuABCD transport system has been identified as a heme uptake system (103). Disruption of ChuA, the OM receptor, inhibited growth on heme and hemoglobin as sole iron sources (104). ChuBCD encodes the putative periplasmic and IM components of the transport system. Cj1613, a gene divergently transcribed from chuA, has been implicated in oxidatively degrading heme in the presence of ascorbic acid and is believed to be the heme oxygenase of the Chu system (104).  1.4.6. THE C. JEJUNI CHANR SYSTEM Complete sequencing of the first C. jejuni genome (strain NCTC 11168) (77) has led to the identification of two putative heme uptake proteins, Cj0177 and Cj0178, which we have named ChaN and ChaR, respectively, for Campylobacter heme association. Microarray studies have demonstrated that these two proteins are regulated by the ferric uptake regulator (4) and are 22  directly upstream of three genes encoding a putative TonB complex. The subsequent sequencing of other C. jejuni strains revealed that the genomic organization in this region is generally conserved, with an exception in C. jejuni strain 81-176 in which the genes encoding ChaNR and the TonB1 complex appear to be absent (81). Nonetheless, the cha system plays an important role in colonization of the gut as a chaR deletion severely hindered the ability of C. jejuni NCTC 11168 to colonize and proliferate within the ceca of chicks (4) and the rabbit ileal loop, a mammalian model for human gastroenteritis due to C. jejuni infections (105). Amino acid sequence alignments showed that ChaR is homologous to characterized OM heme uptake receptors such as Haemophilus influenzae Hup (31% identity), Moraxella catarrhalis MhuA (33%) and P. aeruginosa PhuR (22%) and HasR (22%). Although an earlier study suggested that C. jejuni cannot use transferrin or lactoferrin as sole iron sources (90), a recent chaR deletion study reveals a phenotype in the ability to utilize iron supplied by these two glycoproteins as iron sources (106).  The N-terminal signal sequence of ChaN contains a putative lipobox that is post-translationally lipidated after protein transport across the IM by the Sec system (107). Based upon the lipoprotein localization system of E. coli, the presence of an alanine following the cysteine in the lipobox motif suggests that ChaN is localized to the OM (107). The ChaN sequence is not similar to the transferrin or lactoferrin transport system lipoproteins TbpB and LbpB, respectively, but instead shares 30% identity with PhuW, which is associated with heme acquisition in P. aeruginosa. Deletions of either phuR or phuW exhibit significantly reduced growth on media containing heme as the sole iron source (108). Two global approaches also support a link between ChaN and iron acquisition. Holmes et al. showed that chaN transcript levels increased 25-fold for cells grown under iron-limited conditions as compared to iron-rich 23  conditions (91). Similarly, Palyada et al. demonstrated a decrease in transcription of chaN upon the addition of ferrous sulfate to C. jejuni grown in iron-limited media (4). Both of these studies show that ChaN expression is repressed by iron and are consistent with its control by the ferric uptake regulator.   1.4.7. THE C. JEJUNI P19-CFTR1 SYSTEM Examination of the C. jejuni Fur regulon reveals a member of the Ftr1 superfamily (Cjj81176_1649 in strain 81-176 and Cj1658 in strain 11168), which we have termed cFtr1. Proteins in this superfamily are integral membrane iron transporters containing 6 or 7 transmembrane helices (109). There are over 150 strains of bacteria with genes encoding Ftr1-like products. In E. coli, EfeU (YcdN), sharing 34% sequence identity with cFtr1 (E-value = 1x10-20), is induced under acidic growth conditions and has been shown to transport ferrous iron (110). cFtr1 is also distantly related to Ftr1p, a high affinity iron permease from Saccharomyces cerevisiae (24% identity, E-value = 2x10-10) (111). Sequence comparisons with E. coli EfeU and     Figure 1-4. Schematic representation of members from the Ftr1 superfamily  Four members of the Ftr1 superfamily are depicted with the peptide length listed in brackets. The signal peptide for targeting to the IM is shown in grey. A predicted periplasmic domain is found in the N-terminal portion of C. jejuni cFtr1 and E. coli FetM.  24  yeast Ftr1p, however, reveal that C. jejuni cFtr1 contains a much longer open reading frame, with the addition of an additional uncharacterized domain in the N-terminus (Figure 1-4).  Adjacent to cftr1 in C. jejuni is a gene encoding a 19 kDa protein, P19, which does not share significant sequence identity with the other components of the EfeUOB system. In an earlier study, P19 was found to be acidic (isoelectric point of 4.8) and located in the periplasmic space (112). Although p19 gene regulation by iron availability has been demonstrated by microarrays (4,91) and Fur deletion mutant studies (113), the function of P19 remained unclear, as earlier studies showed that it was not immunogenic during natural infections in humans (112) (likely due to the periplasmic localization) and a C. jejuni P19 deletion strain displayed no obvious phenotypes (9).  A study on the P19 homolog Tp34 (35% sequence identity) from Treponema pallidum, the causative agent of syphilis, has provided some possible functional insights. Unlike P19, Tp34 is a membrane bound lipoprotein located in the inner leaflet of the outer membrane and has a high (submicromolar) affinity for human lactoferrin (114). How Tp34 can be exposed to lactoferrin has been proposed to involve lactoferrin entry into the periplasmic space during membrane perturbations caused by the protein TP0453. This phenomenon has not been reported for C. jejuni and iron-binding proteins such as lactoferrin are not believed to enter the periplasm. The same study also solved the crystal structure of zinc-reconstituted Tp34, but it is unclear what metal(s) are bound in the native Tp34 structure. Studies of ChpA, a close P19 homolog (47% sequence identity) found in the marine magnetotactic vibrio strain MV-1, showed a potential role in iron uptake. This magnetotactic bacterium depends on iron acquisition for the formation of the magnetite-containing magnetosomes. These magnetosomes confer the ability to align and navigate along the Earth's 25  geomagnetic field lines. A naturally occurring non-magnetotactic mutant of MV-1 was shown to have lost the expression of ChpA and was found to have greatly reduced iron content (115). Native ChpA was purified from the periplasm of MV-1 and the copper content was measured to be ~0.5 per dimer. Reconstitution to ~1.0 copper per dimer was successful but attempts to achieve higher copper occupancy were not reported. As the loss of iron acquisition systems in pathogenic bacteria are known to have a dramatic growth defect under iron limitation, we hypothesized that the loss of P19 in C. jejuni would be detrimental to growth under iron-limited conditions, such as in human and animal hosts.   1.5. UROPATHOGENIC ESCHERICHIA COLI E. coli is an incredibly diverse Gram-negative, non-sporulating facultative anaerobe that inhabits the intestines and fecal matter of both warm-blooded animals and reptiles (116,117). Approximately 1010-1011 bacteria are found in our gut per gram of large-intestinal content, totaling to more than 500 species (117,118). Although constituting only a fraction of the total gut population, E. coli is prevalent in over 90% of the human population and is one of the first organisms to colonize the intestinal tract during infancy (117). As part of our commensal gut microflora, E. coli have been shown to participate in symbiotic relationships with the host, contributing to host colonization resistance from invading enteropathogens through the production of bacteriocins and other mechanisms (117,119,120). However, certain strains have developed the ability to stray from this mutually beneficial relationship and cause disease outcomes. These strains fall into two general categories based on general clinical symptoms arising from infection: diarrheagenic E. coli and extraintestinal pathogenic E. coli (ExPEC). ExPEC is differentiated from diarrheagenic E. coli based on the ability of ExPEC to disseminate 26  from the intestinal tract and cause disease during the colonization of other host niches, such as the urinary tract and the central nervous system (121). Interestingly, ExPEC strains are excellent colonizers of the gastrointestinal tract but do not cause gastrointestinal disease in humans (122). To colonize and infect the various extraintestinal niches, ExPEC exhibits a diverse array of virulence factors that are not found in commensal or diarrheagenic E. coli.  Under the broad ExPEC category, uropathogenic E. coli (UPEC) are a group of opportunistic intracellular pathogenic strains commonly associated with disease in humans. UPEC causes 70 to 95% of community-acquired urinary tract infections (UTIs), which in the United States alone costs $1.6 billion each year in medical expenditures (123). More specifically, UPEC generally causes cystitis (inflammation of the bladder) and less frequently pyelonephritis, where UPEC has ascended the ureters and infects the kidneys. Based on rectal and urinary tract isolate comparisons and the fact that UPEC can be asymptomatic intestinal colonizers, it is generally believed that the intestines are the primary reservoirs of UPEC (124), which can be transmitted to others through contaminated food (125) or sexual activity (126). When UPEC enters the normally sterile urinary tract, UPEC has the ability to adhere to host bladder epithelial cells, thereby allowing it to withstand the force of urine flow and invade the urothelial layer. Once inside the cells, UPEC has been shown to enter the cytosolic compartment and rapidly divide, forming a chronic reservoir of bacterial cells that cannot be easily cleared with antibiotic treatment (127). One key virulence factor that permits UPEC to disseminate and cause extraintestinal disease is its large armament of iron acquisition systems, which outnumbers the number found in both commensal isolates and lab strains such as K-12 (121). 27  1.5.1. IRON AND HEME UPTAKE SYSTEMS OF UPEC VERSUS K-12 In general, the siderophore uptake systems found in commensal isolates and the lab strain K-12 are also present in the UPEC strains. These include the hydroxamate receptors FhuA and FhuE, and the catecholate receptors FepA, Cir and Fiu. The ferric citrate receptor FecA is found in K-12 but only sporadically in commensal and uropathogenic strains. In addition to these common siderophore uptake pathways, UPEC has also been shown to carry genes for the uptake of siderophores such as salmochelin (IroN), aerobactin (IutA), yersiniabactin (FyuA) (121) as well as the general catecholate receptor Iha (128). The iron uptake system FitABCDE that is found in UPEC has been shown to be important for iron accumulation but the exact substrate has not been identified (129).  The hemin uptake receptor ChuA, which has been shown to increase the ability of E. coli to colonize the bladders and kidneys of infected mice, is also found in UPEC but not in commensal and lab E. coli strains (130). The most commonly occurring bacterial ferrous uptake system, FeoABC, is found in both pathogenic and non-pathogenic strains of E. coli, including uropathogenic strains such as the fully sequenced strain UTI89 and the partially compiled strain F11. Additionally, the SitABCD metal-ABC permease system has been identified in UPEC strains but not in their commensal cousins. E. coli SitABCD has been shown to be able to take up both ferrous iron and manganese, but with a much greater preference for manganese and is unlikely a major ferrous iron transporter (131). The Nramp-like MntH and ZIP family ZupT transporters, which transport iron as well as other divalent cations, are found in most E. coli strains (51,132). The EfeUOB ferrous uptake system is in both pathogenic and nonpathogenic strains of E. coli (53). However, a frameshift mutation has rendered the EfeUOB system in K-12 nonfunctional.  28  In some strains of E. coli, such as UPEC strains UTI89 and F11, there is a second Ftr1-like permease that shares greater similarity to C. jejuni cFtr1 than EfeU. Like cFtr1, UPEC strain F11 fetM (EcF11_1995) has a much longer open reading frame than efeU and is accompanied by fetP (EcF11_1994), a C. jejuni p19 homolog. The function of FetMP is therefore hypothesized to be similar to that of cFtr1-P19.  1.6. OBJECTIVES OF THIS THESIS The development of new treatment methods or preventative measures against pathogenic microorganisms requires a strong fundamental knowledge of their basic physiology. During both transmission and host colonization, the environment of the pathogen changes with respect to the availability and variety of iron sources. Traversing through the host alimentary and urinary tracts, a pathogen faces changes in pH, oxygen tension and moisture content among other factors, all of which are linked to iron availability. Hence, this highly competitive and challenging host environment has led to a generally larger arsenal of iron uptake systems in pathogenic organisms. The prototypical iron uptake system consists of an OM receptor that is tasked with both substrate recognition and transport, a PBP to shuttle the substrate between the two membranes, and an IM ABC transporter. Numerous examples of these components have been characterized and are described above. However, not all transport systems fall neatly into this often observed transporter scheme.  The main objective of this work has been to characterize three such atypical proteins from two pathogens that have a significant health impact. I hypothesize that ChaNR and cFtr1-P19 are two iron-regulated C. jejuni systems involved in iron uptake. To demonstrate a connection between ChaN and heme metabolism, spectroscopic metal-binding studies and x-ray 29  crystallography on recombinant ChaN protein were employed.  Based on the structural work and sequence analyses, I have shown that the heme-coordinating residues are conserved amongst ChaN homologs from diverse bacterial origins. To characterize heme-dependent dimerization by ChaN, analytical ultracentrifugation studies were conducted. The localization of ChaN was also examined using protease accessibility assays and subcellular fractionation techniques. To study the role of ChaN and ChaR in iron uptake, a C. jejuni chaNR deletion strain was constructed and grown on various iron sources. A p19 deletion strain was also used to show that the loss of P19 alone is detrimental to growth of C. jejuni on iron-restricted media. Furthermore, metal binding analysis demonstrates that recombinant P19 has distinct copper and iron binding sites. To examine the exact mode of metal coordination by P19, metal-bound crystal structures have been solved, which allowed us to propose a model of how P19 functions. The metal-dependency of dimerization and the biological relevance of the dimer in vivo were then characterized by gel filtration and cross-linking studies. FetMP is homologous to the cFtr1-P19 system and is found in uropathogenic E. coli strain F11. To examine the mode of copper binding in FetP, the P19 homolog, the crystal structures of recombinant FetP isolated from the periplasm of E. coli and a copper-reconstituted sample were solved. To isolate the role of two amino acid residues in the displacement of copper observed in the copper-reconstituted structure, the crystal structures of two active site variants were also solved. Through a collaborative study, the effect of FetMP expression in a mutant E. coli strain lacking other iron-uptake systems was examined through growth studies under iron-limited conditions.   30  CHAPTER 2. MATERIALS AND METHODS 2.1. BACTERIAL STRAINS AND GROWTH CONDITIONS  2.1.1. E. COLI Strains of E. coli used in this work are listed in Table 2-1. E. coli cultures were grown in Luria-Bertani (LB) broth, on LB-agar or in Tris-buffered mineral salts medium (pH 7, TMM) (133) containing 2 ml glycerol and 3 g casamino acids per liter with or without additional iron at 37 ?C under atmospheric conditions. When appropriate, the antibiotics kanamycin (25 ?g/ml), gentamicin (15 ?g/ml), chloramphenicol (20 ?g/ml), or ampicillin (125 ?g/ml) were included in the growth media.  2.1.2. C. JEJUNI C. jejuni strains NCTC 11168 and 81-176 were used as the wild-type reference strain for the ChaNR and P19 growth experiments, respectively (Table 2-1). C. jejuni strains were routinely cultured at 38?C on Mueller-Hinton media agar or broth (Oxoid) with vancomycin (10 ?g/ml) and trimethoprim (5 ?g/ml) (MH-TV) under standard C. jejuni growth conditions (6% O2, 12% CO2) produced by a tri-gas incubator (Heraeus) or the Campygen system (Oxoid). Minimal essential media-alpha formulation (MEM?; Gibco) or MH-TV media supplemented with deferoxamine mesylate salt (desferrioxamine) were used for iron-restricted growth conditions. When appropriate, the antibiotics kanamycin (40 ?g/ml) and chloramphenicol (15-25 ?g/ml) were included in the growth media. A supplement of 5 mM pyruvate was added to the MEM? growth medium for cultures used to perform localization studies.   31  Table 2-1. Bacterial strains used in this study Strain Description Reference    E. coli DH5? General strain for plasmid propagation. Contains the ?(lacZ)M15 partial deletion for blue-white screening. Life Technologies E. coli BL21 (DE3) Deficient in the OmpT protease. Contains an integrated ? prophage with T7 RNA polymerase gene inducible by IPTG. Used for protein overexpression. Novagen E. coli ECA458  Constructed from E. coli strain K-12 substrain W3110. Deletion in all known iron uptake systems (?entC ?fecABCDE ?feoABC ?mntH ?zupT). Reference strain used for E. coli FetMP growth studies. (134) ECA458-Gm Negative control strain in which a gentamicin resistance cassette has been introduced at the Tn7-insertion site downstream of the glmS gene Chapter 5 ECA458-fetMP E. coli ECA458 with a fetMP insertion downstream of the glmS gene Chapter 5 ECA458-fetM E. coli ECA458 with a fetM insertion downstream of the glmS gene Chapter 5 ECA458-fetP E. coli ECA458 with a fetP insertion downstream of the glmS gene Chapter 5 E. coli F11 Uropathogenic strain of E. coli. Contains fetMP. (135) C. jejuni NCTC 11168 An isolate from a diarrheic patient. The wild-type reference strain used for the ChaNR study. (77,136) 11168?chaNR C. jejuni strain NCTC 11168 with a chaNR deletion Chapter 3 C. jejuni 81-176 A raw milk outbreak isolate. The wild-type reference strain used for the P19 study. (137) 81176?p19 C. jejuni strain 81-176 with a p19 deletion Chapter 4 81176?p19C Strain 81176?p19 transcomplemented with a copy of p19 in one of three 16S ribosomal regions found in C. jejuni Chapter 4       32  2.2. STRAIN CONSTRUCTION 2.2.1. C. JEJUNI CHANR DELETION MUTANT All plasmid constructs and primers used in this study are listed in Tables 2-2 and 2-3, respectively. chaNR was amplified from genomic DNA provided by Dr. Erin Gaynor and cloned into pBluescript II SK(-) under the control of the lac promoter using the primers chaNR-for and chaNR-rev, which include an XbaI and PstI restriction site, respectively, and produces construct pBlue-chaNR. This construct was then used to produce a C. jejuni non-polar chaNR deletion strain as follows. First, primers chaNR-KO-for and chaNR-KO-rev for inverse PCR (I-PCR) were used to amplify the region that includes ~400 nucleotides in the upstream region of chaN, the pBluescript vector backbone and  ~400 nucleotides of the downstream region of chaR (these two regions of the genes are kept to recombine with the genome within C. jejuni and insert the resistance marker). This strand was then digested with XmaI and ligated to a Campylobacter KanR cassette to produce construct pBlue-chaNR-Kan, a plasmid that does not replicate within C. jejuni and thus serves as a suicide vector. pBlue-chaNR-Kan was then introduced into wild-type C. jejuni NCTC 11168 by natural transformation as previous described and KanR colonies with a disruption in chaNR were selected, forming strain 11168?chaNR (138,139). Confirmation of internal deletion of chaNR was performed by sequencing of the chaNR region and by western blotting with anti-ChaN antibodies. All sequencing was performed at the Nucleic Acid-Protein Services unit (University of British Columbia).    33  Table 2-2. Bacterial plasmids and clones used in this study Plasmid Description Reference    pBluescript II SK(-) E. coli cloning vector for blue-white screening; contains an ampicillin resistance gene Stratagene pBlue-chaNR pBluescript II SK(-) containing a copy of chaNR in the multiple cloning site Chapter 3 pBlue-chaNR-Kan pBlue-chaNR with an internal chaNR deletion that is replaced with a kanamycin resistance cassette; used to create a chaNR deletion mutant in C. jejuni Chapter 3 pET-28a(+) E. coli protein expression vector containing a T7 promoter and an N-terminal His-tag coding sequence; contains a kanamycin resistance gene. Novagen pET28a-chaN pET-28a(+) construct for the expression of recombinant ChaN Chapter 3 pET28a-p19 pET-28a(+) construct for the expression of recombinant P19 Chapter 4 pDrive Cloning vector for direct ligation of PCR products produced by non-proofreading DNA polymerases; contains ampicillin and kanamycin resistance genes. Qiagen p1658P19 pDrive ligated to cftr1-p19 Chapter 4 pCR XL-TOPO Cloning vector for direct ligation of PCR products produced by non-proofreading DNA polymerases Invitrogen pRY109 Campylobacter replicon-containing vector with a chloramphenicol resistance cassette (140) pRRK Delivery vector for the chromosomal insertion of various genes into a conserved C. jejuni rRNA gene cluster by homologous recombination; delivers a kanamycin resistance cassette for insertion selection (141) pRRK-p19 pRRK construct for the delivery of p19 Chapter 4 pGEM-T Easy Cloning vector for direct ligation of PCR products produced by non-proofreading DNA polymerases Promega 34  Plasmid Description Reference    pGEM-fetMP pGEM-T Easy construct containing a copy of fetMP in the multiple cloning site Chapter 5 pGEM-fetP pGEM-T Easy construct containing a copy of fetP in the multiple cloning site Chapter 5 pUC18R6K-mini-Tn7T-Gm Delivery vector for a Tn7-based transposon integration system; delivers a gentamicin cassette for insertion selection (142) pTNS1 Helper plasmid carrying the genes for the site-specific transposition pathway (142) pASK-IBA3 E. coli protein expression vector containing a C-terminal Strep-tag coding sequence and Tet promoter; contains a chloramphenicol resistance gene. IBA pECD1098 pASK-IBA3-based construct for the inducible expression of Strep-tagged FetM Chapter 5 pECD1099 pASK-IBA3-based construct for the inducible expression of Strep-tagged FetP Chapter 5 pECD1100 pASK-IBA3-based construct for the inducible expression of Strep-tagged FetMP Chapter 5 pET22b E. coli expression vector containing a T7 promoter, an N-terminal periplasmic leader sequence and a C-terminal His-tag coding sequence; contains an ampicillin resistance gene. Novagen pET22b-fetP pET22b-based construct for the periplasmic expression of His-tagged FetP Chapter 5 pET22b-fetP-E46Q Periplasmic His-tagged FetP E46Q mutant expression Chapter 5 pET22b-fetP-M90I Periplasmic His-tagged FetP M90I mutant expression Chapter 5      35  Table 2-3. Primers used in this study Primer Sequence (5? to 3?)   chaNR-for GCTCTAGAGCTGATTCAATATCAAAAATAAACCACTAA chaNR-rev AAACTGCAGTTGCATTTGCTTCTGCATT chaNR-KO-for CGCGCCCCCGGGTTGCTCTGTGCTTGCTAACA chaNR-KO-rev CGCGCCCCCGGGAAATAAGCTTTGGTGGTATAGC chaN-for GGAATTCCATATGGCTGTTTTGCAAAAATCATCTC chaN-rev TTGACTCGAGTCAACCTACAGCACAAAATAAACTT cftr1-p19-for TCATATCCTTTTTTTAGATTTAAG cftr1-p19-rev TTTTCCGAGTGTTTGAATATTTTTA p19comp-for CCCTCAATTTATAGTTTTAATTATGCTAATTATTGGC p19comp-rev AAAAATATTCAAACACTCGGAAAATCCGAGCG cftr1-p19-for2 TCCCCGCGGGTCAAAAATCAAGGATAATGATAATGAA cftr1-p19-rev2 CGCGGATCCTAAAAATATTCAAACACTCGGAAAA p19-for CAGCAGCGGCCTGGTGCCGCGCGGCGGCGGCGAAGTGCCGA TCGGCGATCCAA p19-rev TGCGGCCGCAAGCTTGTCGACGGAGTTATTTTGGCGTGCCTG TGTATTTGA fetMP-for TTACAGCGTCTTGCCAGCGATC fetMP-rev GTACGGCGGGTTGAATTAAGCG fetM-rev GTATCCTCTCGTCTAAAACAACGGCT fetP-for AAACCATGGAACCCATAATCGTTGTATAGCCGT fetP-rev AAACCATGGTTCCGCTGTTTTATCAAGACGTTG fetP-pASK3-for AGCGAATTCACCATGAAGAAAACCCTGATTGCC fetP-pASK3-rev TCGCTGCAGGTTCAGACCGACATATTTAAACTCGTAGCTC fetP-pET22-for AAACCATGGGCTTTAAAGAGTACCCGGCAGGC fetP-pET22-rev AAACTCGAGGCTGCCGCGCGGCACCAGGCCGCTGCTGT TCAGACCGACATATTTAAACTC E46Q-for AAAGCCGATGTTCACCTTCAGGCGGATATCCACGCTGTA   36  Primer Sequence (5? to 3?)   E46Q-rev TACAGCGTGGATATCCGCCTGAAGGTGAACATCGGCTTT M90I-for GGCACCTTCATGCCGATCGTTGCCAGCGATGGC M90I-rev GCCATCGCTGGCAACGATCGGCATGAAGGTGCC     37  2.2.2. C. JEJUNI P19 DELETION MUTANT The p19 gene with approximately 500 base pairs of flanking region on both sides was amplified by PCR from 81-176 genomic DNA.  This PCR product was cloned into the pCR XL-TOPO vector (Invitrogen) according to the manufacturer's instructions.  The resulting plasmid was used as a template for I-PCR with primers containing MfeI sites engineered into the 5' end.  The resulting product, which had ~90% of the target gene coding region removed, was digested with MfeI to create cohesive ends compatible with EcoRI. DpnI was used to selectively degrade the template plasmid methylated DNA. The digested PCR product was ligated to the chloramphenicol resistance cassette (CAT; CmR), which had been excised from pRY109 by EcoRI digest (140).  Constructs harboring p19 disrupted by CAT were selected on chloramphenicol plates.  The purified plasmid, which cannot replicate in C. jejuni, was introduced into 81-176 by natural transformation as previous described (138,139) and CmR colonies with p19 disrupted by CAT were selected. Confirmation of internal deletion of p19 with the insertion of a chloramphenicol resistance cassette was performed by amplification and sequencing of the region of cftr1-p19 using the primers cftr1-p19-for and cftr1-p19-rev and by western blotting with anti-P19 antibodies.  2.2.3. COMPLEMENTATION OF 81176?P19  Complementation of 81176?p19 was performed by inserting a wild-type copy of p19 into one of three 16S ribosomal regions found in C. jejuni following a previously described method (141). Briefly, the P19 coding region was amplified from genomic DNA using the primers p19comp-for and p19comp-rev containing XbaI and MfeI restriction sites. After digestion, the PCR product was ligated into XbaI and MfeI-digested pRRK [J. Ketley, unpublished, based on 38  pRRC integration vectors (141)] to form pRRK-p19. pRRK-p19 was introduced into 81176?p19 by natural transformation, as previously described (138,139). Briefly, 81176?p19 C. jejuni cells in log phase were streaked into 1 cm wide circles on MH-TV plates and grown for 3-4 hours at 38 ?C. ~3 ?g of purified pRRK-p19 plasmid was spotted onto the cells. The cells were left to grow for 5-7 hrs, transferred to MH-TV plates containing 40 ?g/ml kanamycin and 20 ?g/ml chloramphenicol, and grown for up to 3 days. Successful integration (to produce 81176?p19C) was assessed using primers as previously described (141) and western blotting.  2.2.4. E. COLI To analyze the FetMP-system from uropathogenic E. coli strain F11, single fetMP, fetM, or fetP chromosomal integrations were generated at the Tn7-insertion site downstream of the glmS gene of E. coli strains ECA458 (?entC ?fecABCDE ?feoABC ?mntH ?zupT) as published (142). The complete fet-operon was amplified by PCR from genomic F11 DNA using primer pairs fetMP-for/fetMP-rev (for fetMP) and fetMP-for/fetM-rev (for fetM), and cloned into the pGEM-T Easy vector (Promega). Primer pair fetP-for/fetP-rev and pGEM-fetMP were used as template to generate pGEM-fetP. Each construct was subcloned into plasmid pUC18R6K-mini-Tn7T-Gm. In the presence of the helper plasmid pTNS1, the subcloned genes were integrated into chromosomal attTn7 sites of E. coli ECA458.   2.3. GROWTH PHENOTYPE STUDIES 2.3.1. C. JEJUNI CHANR Wild-type 11168 and ChaNR deletion (11168?chaNR) strains were grown for 24 hrs with a single passage on MH-TV agar. The cells were then inoculated into MEM? at an initial 39  OD600 of 0.025, grown overnight and were harvested. Fresh unsupplemented MEM? or MEM? supplemented with FeSO4, holo-transferrin, or holo-lactoferrin was prepared in 4 ml aliquots in test tubes and inoculated with the cells at an OD600 of 0.025. The samples were then grown with agitation under microaerophilic conditions as produced by the Campygen system (Oxoid) and OD600 readings were taken at regular intervals.  2.3.2. C. JEJUNI P19 Wild-type 81-176, P19 deletion (81176?p19), and complemented P19 deletion (81176?p19C) cells were grown for 24 hrs with a single passage on MH-TV agar. Biphasic MH-TV media, consisting of 10 ml MH broth over 5 ml MH agar, was then inoculated with each strain at an initial OD600 of 0.002 and was grown overnight to an OD600 of ~0.2. 12-well plates containing 0.5 ml MH-agar overlaid with 1 ml MH-broth (biphasic media) with increasing concentrations of desferrioxamine (0 - 28 ?M) and bathocuproine disulfonic acid (0 - 150 ?M; BCS) were inoculated with each strain at a starting OD600 of 0.02. OD600 and CFU/ml were assessed after 4, 8, 12, 16 and 24 hrs of growth.  2.3.3. E. COLI To produce iron-limited cell cultures of the various E. coli strains, LB overnight cultures were diluted 1:400 in Tris-buffered mineral salts medium without iron and cultivated overnight at 37 ?C with shaking. Cultures were diluted 1:400 a second time into fresh Tris-buffered mineral salts medium without added iron. After 2 hrs of growth at 37 ?C with shaking, cultures were diluted again 1:400 into Tris-buffered mineral salts medium and used for further experiments.  40  For time-course growth experiments, 1 ?M of the iron chelator CDTA was added to iron-limited cultures of the strains ECA458-Gm, ECA458-fetM, ECA458-fetP and ECA458-fetMP. Ferrous and ferric CDTA complexes have a higher stability constant than those of other divalent metal cation chelators such as EDTA (143). ECA458-Gm is the negative control strain in which a gentamicin resistance cassette (GmR) has been introduced at the Tn7-insertion site downstream of the glmS gene. Growth was measured in Klett units over a time period of 20 hrs.  For iron uptake experiments, a filtration assay using E. coli strain ECA458 harboring plasmids pECD1098 (pASK-IBA3::fetM), pECD1099 (pASK-IBA3::fetP), pECD1100 (pASK-IBA3::fetMP) or the empty vector pASK-IBA3 as control was performed as published (144). Briefly, cells were cultivated overnight in LB medium, diluted 1:400 into TMM, cultivated overnight and diluted into fresh medium to a final turbidity of 30 Klett units. The cultures were incubated with shaking at 37 ?C to a turbidity of 60 Klett units, 200 ?g anhydrotetracycline (AHT) per liter was added to induce expression of the cloned genes, and incubation was continued for an additional hour. The cells were then harvested by centrifugation and washed twice with TMM without added casamino acids, phosphate or metals. Metal uptake was initiated by the addition of a reaction mixture leading to final concentrations of 1 ?Ci 55Fe, 5 ?M FeSO4, and 1 mM ascorbate. The 55FeCl3 (Perkin-Elmer) had a specific activity of 89.81 Ci/g. For dose-response experiments at different pH values, FeCl3, MnCl2 or ZnCl2 were added to iron-limited cultures of strains ECA458-fetMP and ECA458-Gm in double buffered MES (2-Morpholinoethanesulfonic acid)-Tris-buffered mineral salts medium (pH 5 to 9). After 24 hrs of incubation at 37 ?C with shaking, the turbidity was measured at 600 nm.     41  2.4. RECOMBINANT PROTEIN EXPRESSION  2.4.1. C. JEJUNI CHAN The portion of chaN that corresponds to the product without the N-terminal signal sequence and the predicted lipid attachment site was amplified from clone pBlue-chaNR and subcloned into pET-28a(+) utilizing the primers chaN-for and chaN-rev, producing pET28a-chaN. The resulting recombinant ChaN protein includes an N-terminal poly-His tag and residues 2-265 (residue 1 being the cysteine lipidation site). The final clone was sequenced at the Nucleic Acid-Protein Services unit (University of British Columbia). ChaN was overexpressed in Escherichia coli strain BL21(DE3), and the cells were grown in 2?YT (Difco) containing 25 ?g/ml kanamycin. The culture was incubated at 30 ?C for 18 hrs with shaking. The culture was induced with 0.2 mM isopropyl-?-D-thiogalactopyranoside (IPTG) when the cell density reached an OD600 of 1.0. The cells were lysed at 4 ?C in 300 mM NaCl, 50 mM NaH2PO4, pH 7.9 with an Emulsi Flex-C5 homogenizer (Avestin). The soluble fraction was loaded onto a ProBond nickel resin affinity column (Invitrogen) and the colorless protein (apo-ChaN) was eluted with imidazole. The buffer was exchanged by ultrafiltration with either 50 mM MES pH 6.5 (for crystallization) or 30 mM sodium phosphate pH 6.5 (for all other analyses; Buffer A). ChaN was digested with thrombin (500:1 w/w ChaN/thrombin ratio) at 4 ?C to remove the His-tag. ChaN was then applied to a Mono S HR 5/5 cation exchange column (Amersham) and eluted with NaCl. Excess salt was removed by ultrafiltration. Protein purity and digestion was analyzed by SDS-PAGE and mass spectroscopy (MSL/LMB Proteomics Core Facility, University of British Columbia). To express seleno-methionine labeled (SeMet) ChaN, E. coli strain BL21(DE3) transformed with pET28a-chaN was cultured following a previously published procedure with 42  minor modifications (145). The culture was grown to 0.6 OD600 at 37 ?C before the addition of SeMet, induced with 0.2 mM IPTG and grown overnight at 27 ?C. The cells were lysed and SeMet labeled ChaN was purified using the same protocol for unlabeled protein.  2.4.2. C. JEJUNI P19  The genes Cj1658 and p19 were first amplified from 11168 genomic DNA using the primers cftr1-p19-for2 and cftr1-p19-rev2. 3?-A overhangs were added to the PCR product and annealed to pDrive vector using the QIAGEN PCR CloningPlus Kit, forming p1658P19. The coding region of p19 corresponding to a product without the periplasmic signal sequence was amplified from p1658P19 and subcloned into pET-28a(+) utilizing the primers p19-for and p19-rev in a two-step PCR reaction as previously described (146), producing pET28a-p19. The resulting recombinant protein product includes an N-terminal poly-His tag, a thrombin cleavage site and amino acids 22-179 of native P19 (RefSeq accession number: YP_002345027). These residues have been renumbered to 2-159 in our structures with residue 1 being the first residue post-thrombin cleavage. P19 was overexpressed in E. coli strain BL21(DE3) and grown in 2?YT. Each 1 L culture was inoculated with 2 ml of overnight culture and incubated at 30 ?C with shaking. When the cell density reached an OD600 nm of 1.0, the temperature was reduced to 25 ?C and the culture was induced with 0.5 mM IPTG and then grown overnight. The cells were resuspended and lysed at 4 ?C in binding buffer (20 mM NaH2PO4, 500 mM NaCl, pH 7.8) containing 5 mM CuSO4 with an Emulsi Flex-C5 homogenizer (Avestin). The soluble fraction was loaded onto a ProBond nickel resin affinity column (Invitrogen) and the protein was eluted with increasing imidazole in binding buffer at pH 6.0. The buffer was exchanged with Native Buffer (20mM Hepes, pH 7.5; NB) by ultrafiltration. P19 was digested with thrombin (500:1 43  w/w P19/thrombin ratio) at 4 ?C to remove the His-tag. Benzamidine beads were used to remove the thrombin and purified (?as isolated?) protein was concentrated to ~20-25 mg/ml using Amicon filter devices (10,000 MWCO; Millipore). SeMet-labeled protein was produced as described for SeMet-ChaN and purified as for ?as isolated? protein.  Apo-P19 was obtained by incubating the concentrated protein with 30x molar excess ethylenediamine tetra-acetic acid (EDTA) for 10 minutes followed by desalting over an 18 ml S-200 column.  The absence (or amount) of copper and iron in all experimental samples was verified using the BCA (147) and ferene-S assays (148), respectively. Briefly, the proteinaceous component is removed with trichloroacetic acid, followed by reduction of the released Cu or Fe ions with freshly prepared ascorbic acid. The colorimetric reagents BCA and ferene S are specific for Cu+ and Fe2+, respectively. Reconstitution with copper was through incubation in NB containing 2x molar excess CuCl2 at room temperature, followed by ultrafiltration to remove excess metal. Protein concentration was determined using the Biorad Protein Assay Dye Reagent (as per the manufacturer?s protocol), in which the molar absorptivity at 280 nm was calculated to be 21,400 M?1 cm?1. The Zn2+/Cu2+ and Fe2+/Cu2+ competition assays were performed by premixing 1.5x molar excess of each ZnCl2 or FeCl3 with CuCl2, which is then added to apo-P19 in 20 mM Bis-Tris pH 6.5. After room temperature incubation for 45 min, the sample was desalted with a G-25 Sephadex column and assayed for copper content. Ferrous binding assays were performed by incubation with 200x excess dithionite or ascorbic acid for 30 min, followed by desalting in a G-25 Sephadex column equilibrated with NB supplemented with 25 mM dithionite or ascorbic acid.     44  2.4.3. E. COLI FETP To produce Strep-tagged wild-type FetP, the fetP gene was amplified by PCR from E. coli strain F11 genomic DNA using primer pair fetP-pASK3-for/fetP-pASK3-rev and cloned into vector pASK-IBA3. E. coli strain BL21(DE3) with pECD1099 (pASK-IBA3::fetP) was cultivated in LB medium with shaking at 37 ?C until an optical density of 0.8 at 600 nm. Expression of fetP was induced with 200 ?g anhydrotetracycline (AHT) per liter and incubation was continued for 3 hrs at 30 ?C. Cells were harvested by centrifugation and the periplasmic fraction was isolated as described (149). FetP protein was expressed as a Strep-tagged protein and purified using a Strep-tactin affinity chromatography column according to the manufacturer's protocol (IBA GmbH). Protein concentrations were determined with a NanoDrop ND-1000 spectrophotometer (Wilmington) using the molecular mass of the mature periplasmic protein and an extinction coefficient of 29,910 M-1 cm-1 at 280 nm, which was calculated from the primary sequence with the ProtParam tool at ExPASy (http://expasy.org/tools/protparam.html). FetP was detected by Western blot analyses after SDS-PAGE with Strep-tactin-horseradish peroxidase conjugate (IBA) and the molecular mass was determined by MALDI-TOF (UltraflexII MALDI-TOF/TOF mass spectrometry, Bruker Daltonik GmbH).  To produce His-tagged wild-type FetP, the fetP gene without the coding region for the leader sequence was amplified by PCR using primer pair fetP-pET22-for/fetP-pET22-rev and cloned into the pET22b(+) vector (Novagen) leading to plasmid pET22b-fetP. This construct was then used to express recombinant FetP with an N-terminal artificial leader sequence and a C-terminal His6-tag that includes a thrombin cleavage site for crystallographic studies. FetP was overproduced in E. coli strain BL21(DE3) grown in 2?YT. Each 1 L culture was inoculated with 45  2 ml of overnight culture and incubated at 37 ?C with shaking to an optical density of 1.0 at 600 nm. The culture temperature was reduced to 25 ?C for 30 min, gene expression induced with 0.25 mM IPTG and then grown overnight. The cells were harvested by centrifugation and the periplasmic fraction was isolated as described above. The wild-type periplasmic fraction was dialyzed into 30 mM phosphate buffer pH 7.8, 300 mM NaCl, loaded onto a HisTrap nickel resin affinity column, and FetP was eluted with increasing imidazole. The buffer was exchanged with 30 mM phosphate pH 7.8 or 20 mM Bis-Tris pH 6.5 by ultrafiltration and the protein was digested with thrombin (Haematologic Technologies). Benzamidine beads were used to remove the thrombin. The resulting periplasmic FetP protein (referred to as ?as-isolated?) was concentrated to at least 25 mg/ml using Amicon filter devices (10,000 MWCO; Millipore). To produce copper-reconstituted FetP (Cu-FetP), ?as isolated? FetP was applied to a Source Q ion exchange column and eluted with a NaCl gradient in 30 mM phosphate pH 7.8. Apo-FetP was prepared by treatment with 20x molar excess EDTA in 20 mM Bis-Tris, pH 6.5 for 3 hrs on ice, followed by repeated buffer exchange by ultrafiltration. Copper-reconstituted wild-type FetP was then prepared by direct addition of 1.5x molar excess CuCl2 to apo-FetP, incubating the mixture on ice for 20 min, and concentrated by ultrafiltration to at least 25 mg/ml protein.  2.4.4. E. COLI HIS-TAGGED FETP MUTANTS E46Q AND M90I Site directed mutagenesis by the QuikChange method (Stratagene) was used to generate the active site variants E46Q and M90I using pET22b-fetP as the template. Each primer pair is designed to overlap the mutation site and contains the mutation of interest. For FetP E46Q, the primer pair E46Q-for and E46Q-rev was used; for FetP M90I, primers M90I-for and M90I-rev 46  was used. DpnI was used to degrade template vector prior to transformation. Successful mutagenesis was confirmed by sequencing. Poly-histidine-tagged active site variants of FetP were produced in a manner similar to wild-type FetP used for crystallographic studies with the following modifications. The 1 L culture was reduced to 25 ?C at an optical density of 0.6 at 600 nm. The periplasmic fraction was dialyzed into 20 mM Tris pH 7.5, 150 mM NaCl and 10 mM imidazole, eluted with increasing imidazole and dialyzed into 20 mM Tris pH 7.5 and 150 mM NaCl. Following digestion of the mutants, thrombin was removed by gel filtration in a HiLoad 16/60 Superdex 200 column (GE Healthcare). The copper-loaded mutants (Cu-FetP E46Q and Cu-FetP M90I) were prepared in the same manner as for Cu-FetP.  2.5. CHAN SPECTROSCOPIC HEME AND PH TITRATION  The electronic absorption spectra of apo- and heme-bound forms of ChaN were recorded at 25 ?C with either a Cary 4000 or 6000i UV-visible spectrophotometer and 1 cm path length quartz cuvettes. Similarly, CD spectra of these samples were collected in the Soret and far-UV regions of the spectrum (1 and 0.1 cm path length quartz cuvettes, respectively) with a Jasco J-720 spectrometer. For these experiments, holo-ChaN was prepared by addition of a concentrated solution of hemin (Sigma-Aldrich) dissolved in dimethyl sulfoxide to a solution of thrombin-treated apo-protein in Buffer A so that the final heme concentration was 5-fold greater than the protein. Unbound heme was removed by centrifugation followed by gel filtration chromatography with a 1 x 30 cm, Sephadex G-25 Fine (GE Healthcare) column equilibrated in Buffer A. For heme titrations, ? 1 mg hemin was dissolved with 100 ?L 0.1 M NaOH and diluted to 1 ml with 30 mM sodium phosphate buffer pH 7.4 or 20 mM TEA buffer pH 7.9 to yield a 1 mM hemin solution. pH titrations were conducted by mixing equal volumes of a buffer 47  cocktail (20 mM of each of Bis-Tris, MOPS, TEA, TAPS, CAPS buffers, boric acid, and 50 mM sodium chloride) adjusted to different pH by addition of concentrated NaOH, and a protein stock prepared by dilution of ? 1.5 mM holo-ChaN with deionized water to a concentration of ? 30 mM protein. The pH values used for data analysis are those measured immediately after collection of the absorption spectra, and they typically changed by less than 0.15 pH units upon mixing of the protein and buffer solutions. For the analysis of these data, the absorption spectra were corrected for dilution, and the first derivatives were computed with a GAP function of ? 1 nm as implemented in the program Grams/AI (V. 7.0).  2.6. ANTI-CHAN AND ANTI-P19 POLYCLONAL ANTIBODY PRODUCTION Polyclonal antiserum against ChaN and P19 was generated by immunization of female New Zealand white rabbits. Recombinant protein (~0.5 mg) was homogenized with complete Freund?s adjuvant and used to immunize the rabbit by intramuscular injection. Two doses of 0.1 mg protein were administered at 3 week intervals as boosters. At week 9, the rabbit was euthanized and antiserum was harvested and purified by the addition of equivolume of saturated ammonium sulfate as per the protocol provided by Sigma.  2.7. C. JEJUNI PROTEIN LOCALIZATION STUDIES 2.7.1. SUBCELLULAR FRACTIONATION To examine the localization of ChaN, C. jejuni NCTC 11168 was grown overnight on MH-TV agar, passed once onto fresh MH-TV agar and then overnight in MEM? supplemented with 5 mM pyruvate. The cells were then pelleted and resuspended in 10 mM Tris pH 7.4 with 1 mM MgCl2. A few crystals of DNase I were then added to the cell suspension to prevent DNA-48  induced clumping, along with a protease inhibitor cocktail (Boehringer Mannheim). The supernatant was collected and concentrated by ultrafiltration to determine whether ChaN is secreted. The periplasmic fraction was isolated using an osmotic shock protocol as previously described (150) and the remaining shocked cells were lysed using a French press. The membranes were pelleted and the remaining soluble fraction was treated as the cytoplasmic fraction. The pellet was washed with 10 mM Tris pH 7.4 and then resuspended in 1% Triton X-100 or 0.5% sarkosyl (sodium lauroyl sarcosinate). Differential solubilization of the inner and outer membrane using these two detergents has been previously shown to be suitable for C. jejuni (151,152). After 35 min incubation at room temperature, the sample was centrifuged, separating the OM-concentrated fraction from the solubilized IM-concentrated fraction. The OM-concentrated pellet was then washed again prior to separation on an SDS-PAGE gel and immunoblotting using anti-ChaN and anti-CadF antibodies (a gift from Dr. Michael Konkel) as a control.  2.7.2. CELL SURFACE DIGESTION To examine whether ChaN and P19 are accessible to surface protease digestion, C. jejuni NCTC 11168 was first grown in the same manner as for the subcellular fractionation study. The cells were then spun down and washed with phosphate buffered saline (PBS) pre-warmed to 37?C. Trypsin (final concentration 100 ?g/ml; Sigma) was added to the cells, which was kept at 37?C. Samples were removed at regular intervals, mixed with PMSF to inhibit further protein degradation, pelleted and then immediately mixed with sample loading buffer and boiled. The samples were then separated by SDS-PAGE and a western blot was performed probing for ChaN, P19 and CadF. 49  2.8. CRYSTALLIZATION AND STRUCTURE DETERMINATION  All structural figures have been generated with the program PyMol (Version 1.2r3pre, Schr?dinger, LLC.)  2.8.1. C. JEJUNI CHAN ChaN and heme-reconstituted ChaN (SeMet-labeled and unlabeled) were crystallized by hanging drop vapor diffusion and microseeding with His-tagged ChaN crystals. The crystallization well contained a solution of 15% PEG 4000, 100 mM HEPES, pH 7 and a drop of this solution was mixed with an equivalent volume of heme-reconstituted ChaN with and without the His6-tag (~20 mg/ml in 50 mM MES pH 6.5). Crystals were obtained after two days of incubation at room temperature.  His-tagged SeMet-ChaN was crystallized under the same conditions as the unlabeled protein. Crystals appeared after one week of incubation at room temperature. The crystals were submersed in cryoprotectant (30% (v/v) glycerol prepared in the crystallization solution) and data was collected at 100 K. Single-wavelength anomalous dispersion (SAD) data at a wavelength of 0.9787 ? were collected under cryogenic conditions on beam-line 1-5 at the Stanford Synchrotron Radiation Lightsource (SSRL; Palo Alto, CA). The collected data was indexed, integrated and scaled with HKL2000 (153). The crystals grew in space group C2 (a = 87.92, b = 58.79, c = 82.40 and ? = 119.21). One ChaN monomer is found in the asymmetric unit. SOLVE and RESOLVE (154) were used to obtain initial phases and to produce a preliminary model through automated fitting. A complete model was built manually utilizing the program O (155) and refined to 1.9 ? using Refmac5 (156) from the CCP4 suite of programs (157). The electron density for the heme tetrapyrrole ring was clearly defined in early electron density maps. The fit of the heme was confirmed by omit 50  difference maps and was refined to an average B-factor of 52.8 ?2. The final model consists of 254 amino acid residues, one heme, and 243 water molecules. A Ramachandran plot, with 95.9% of the residues within the favored regions and 3.7% in the allowed regions, shows that the structure has excellent stereochemistry (158).   2.8.2. C. JEJUNI P19 ?As isolated? and metal-reconstituted P19 was crystallized by hanging drop vapor diffusion. P19 crystals in the P6222 space group were initially produced in a crystallization well containing 2.0 M ammonium sulfate, 0.1 M citrate, pH 5.5. All P19 crystals were flash frozen using liquid nitrogen and data was collected at 100 K. The crystal structure was determined partially from a 3-wavelength (0.9184 ?, 0.9790 ?, and 0.9791 ?) SeMet MAD experiment with one monomer per asymmetric unit by using the autoXDS script available at the Joint Center for Structural Genomics (data not shown). The Matthew?s coefficient was 2.62 assuming one monomer in the asymmetric unit, implying 53% solvent content. The highest resolution dataset was ~2.8 ?. The anomalous substructure determined by SHELXD (159) with data to 2.9 ? found 8 sites with a high correlation coefficient of 42.5. The resulting SHELXE map refined to a well connected ?protein-like? electron density map. The 8 initial sites were later confirmed to be Met88, Met63, Met15, Met102, Met27, and Met86 with the last 2 sites being false positives with low refined occupancies. Phases from the substructure were used by ArpWarp to build a partial structure with residues 11-26, 37-74, and 117-136. Even after many rounds of manual rebuilding some parts of the structure were not assigned (81-102; 151-159) mostly due to the low resolution of the dataset. The partially-built model refined to R/Rfree of 25.5/33.1% and an average B-factor value of 109 ?2.  51  Later, a new crystal form of ?as isolated? P19 with orthorhombic space group P21212 was obtained with 50% polyethylene glycol (PEG) 250 and 0.1 M 2-(N-cyclohexylamino) ethane sulfonic acid (CHES), pH 9. Diffraction data from these orthorhombic crystals was used to completely assign the amino acid sequence. Apo-P19 and Cu-P19 crystals were similarly obtained with 25-45% PEG 250 or 350 and 0.1 M CHES, pH 9-10. A drop of the well solution was mixed with an equivalent volume of P19 (~25-35 mg/ml in 20 mM Tris pH 8.0). Crystals were obtained after a few days to a few weeks of incubation at room temperature. Crystals prepared with ammonium sulfate were submersed in cryoprotectant (30% (v/v) glycerol prepared in the crystallization solution). No additional cryoprotectant was used for the crystals containing PEG.  Manganese-soaked crystals were prepared by soaking ?as isolated? P19 crystals in the well solution supplemented with 1 mM MnCl2. The crystal structure of Cu2+ and Fe2+- reconstituted P19 (CuFe-P19) under oxidizing conditions was produced by first progressively soaking Cu-P19 crystals in buffers with decreasing pH and 42% PEG 350 for 20 min under each condition. The buffers used are CHES pH 8.6, Tris pH 7.5, Bis-Tris pH 6.5 and Bis-Tris 5.5. This was followed by the direct addition of 0.5M (NH4)2Fe(SO4)2 directly to the depression plate well containing the crystals to a final concentration of 2 mM iron. The well solution progressively became yellow to indicate oxidation of the iron to Fe3+. The crystals were harvested after two hours of soaking with iron. The crystal structure of CuFe-P19 under reducing conditions was produced in a manner similar to CuFe-P19 under oxidizing conditions with the following changes: the concentration of PEG 350 was increased to 44% in each of the soaks; 10 mM dithiothreitol was included in each of the soaking solutions to ensure reduced protein; the 52  pH of the crystals was brought down to 6.5 prior to the addition of iron to the well with the crystals.  Here we report crystal structures of P19 in space group P21212 in the presence of different metals and under oxidizing and reducing conditions. All datasets were indexed, reduced with XDS or HKL2000 and scaled either with XSCALE, SCALA or HKL2000 (153,160,161). The structures were refined using Refmac5 (156) and COOT (162). The final model with the highest resolution of 1.4 ? consists of 316 amino acid residues (158 per subunit), one copper per monomer, and 353 water molecules. A Ramachandran plot, with 98.2% of the residues within the favored regions and no outliers, shows that the structure has excellent stereochemistry. Cis-peptides were observed between E28 and P29 in both subunits, both supported by well-defined electron density.  The high resolution ?as isolated? P19 crystal was soaked in well solution containing Mn2+ prior to freezing. Based on an x-ray fluorescence scan, Mn2+ comprises less than 5% of the total metal population in the crystal. To ensure that the Mn2+ did not affect the primary metal binding site in the ?as isolated? crystal structure, a second, albeit lower resolution dataset was collected on an ?as isolated? P19 crystal that was not soaked with Mn2+ (data not shown).   2.8.3. E. COLI FETP ?As-isolated? FetP, wild-type Cu-FetP and Cu-FetP mutants (E46Q and M90I) were crystallized by the sitting-drop method. A single, large crystal of ?as-isolated? FetP appeared within two weeks of incubation with a reservoir of 20% PEG 3350, 0.2 M ammonium citrate pH 7. The crystals were soaked in mother liquor supplemented with 20% ethylene glycol as a cryoprotectant, flash frozen using liquid nitrogen and data was collected at 100 K at the SSRL on 53  beam-line 9-2. An x-ray fluorescence scan was used to detect transition metals in the crystal. A 1.6 ? resolution dataset was collected using incident radiation with a wavelength of 0.9796 ?. Data were processed using iMosflm (163). MolRep (164) was used to determine initial phases of ?as-isolated? FetP using one chain from the crystal structure of P19 (PDB ID: 3NRP). Cycles of structure refinement and building were performed using Refmac5 (156) and Coot (162). The crystal structure of ?as-isolated? FetP was determined in space group P32 with four monomers per asymmetric unit and twinning with four twin domains as defined by Refmac5 (calculated distribution of 41.40, 40.74, 9.00, and 8.85%). Sufficient electron density was observed to build residues 2 to 153 of chains A and B and residues 2 to 152 of chains C and D. A clear break in the electron density was observed at residue 34 of Chain B and is therefore not modeled.  Cu-FetP was crystallized from 0.1 M Bis-Tris pH 6.5, 25% pentaerythritol ethoxylate (15/4 EO/OH), and 25 mM ammonium sulfate. Small (~0.05 mm), thin crystals of Cu-FetP in the orthorhombic space group P212121 appeared within a week and were harvested three months later to improve thickness. The crystals were flash frozen using liquid nitrogen with no additional cryoprotectant and data was collected at 100 K at the SSRL on beam-line 7-1. An x-ray fluorescence scan was used to ensure no detectable contamination by metals other than copper in the sample and to determine the optimal wavelength for maximal copper anomalous signal. A 1.7 ? resolution dataset at 0.9764 ? wavelength and a copper anomalous dataset at 1.3773 ? wavelength were collected on a single crystal. The crystal structure of Cu-FetP was determined with two monomers per asymmetric unit (PDB ID: 3NRQ). Electron density was observed to build residues 4 to 153 of chain A and 4 to 152 of chain B. MolRep was used to determine initial phases using one chain from the crystal structure of ?as-isolated? FetP. Other than the two 54  residues at the C-terminus of chain A, the B-factors of all residues in both chains are under 35 ?2, supporting correct occupancy and modeling.  Cu-FetP E46Q was crystallized in 0.1 M Bis-Tris pH 6.5, 20% PEG 550, and 40 mM CaCl2 dihydrate. Crystals in the C2 space group were harvested six weeks later, soaked in well solution containing approximately 10% higher PEG 550 as a cryoprotectant and flash frozen. A 1.4 ? resolution dataset at 1.0000 ? wavelength and a copper anomalous dataset at 1.3773 ? wavelength on a single crystal were collected in the same manner as Cu-FetP. The crystal structure contains two monomers per asymmetric unit (residues 3 to 152 of chain A and 3 to 160 of chain B). MolRep was used to determine initial phases using the crystal structure of Cu-FetP.  Cu-FetP M90I was crystallized in 0.15 M formate and 16% PEG 3350. Crystals in the C2 space group were harvested two months later and flash frozen using 30% glycerol prepared in well solution as cryoprotectant. A 1.5 ? resolution dataset at 1.0000 ? wavelength and a copper anomalous dataset at 1.3771 ? wavelength on a single crystal were collected in the same manner as Cu-FetP. The crystal structure contains two monomers per asymmetric unit (residues 1 to 152 of chain A and 3 to 160 of chain B). MolRep was used to determine initial phases using the crystal structure of Cu-FetP.   2.9. PROTEIN DIMERIZATION STUDIES 2.9.1. ANALYTICAL ULTRACENTRIFUGATION ANALYSIS OF CHAN  Sedimentation velocity experiments were conducted at 20 ?C with a Beckman Optima XL-I analytical ultracentrifuge equipped with both absorbance and interference optics. Standard aluminum double-sector centerpieces (12 mm) were filled with protein solution (400-450 ?l), and the buffer that was used in the last purification step was placed in the reference cell (Buffer 55  A). Prior to each run, the loaded cells were thermally equilibrated in the centrifuge for at least 1 hr after the instrument had reached 20 ?C under vacuum. Sedimentation velocity experiments were performed with 4-hole (AnTi60) and 8-hole (AnTi50) rotors. Sapphire and quartz windows were used with interference and absorbance optics, respectively. Radial scans were acquired with 0.003 cm radial steps in continuous mode without averaging, rotor speed was set at 50,000 rpm and for both interference and absorbance, no time interval was set between scans. Data were analyzed with a c(s) distribution of the Lamm equation solutions calculated with the program SEDFIT (165) assuming the regularization parameter p to be 0.95 (high confidence level). Sedimentation coefficient increments of 200 were used in the appropriate range for each sample. The apo-protein was run with both interference and absorbance optics at 280 nm (8, 15, 30, 50 and 80 ?M), while holo-ChaN was detected by setting the absorbance at 400 nm (17 and 36 ?M). The solution densities and partial specific volumes were calculated with the program SEDNTERP (166).  Sedimentation equilibrium experiments were run in 12mm 6-channel Epon centerpieces in the 4-hole rotor (AnTi60). The samples were run at three speeds (12,000, 20,000 and 30,000 rpm) and three different concentrations for both apo-ChaN (24, 76 and 172 ?M) and holo-ChaN (3, 14 and 21 ?M) (20 ?C, Buffer A). Data were collected with interference optics (apo) or absorption at 400 nm (holo). When the absorbance optic was used, radial increments were set at 0.001 cm; for each data set, the curve is the result of 10 averaged scans. Achievement of equilibrium conditions was confirmed with the program WinMatch (Jeffrey, L. & Yphantis, D., National Analytical Ultracentrifugation Facility, University of Connecticut). Data were analyzed with the program SEDPHAT (167) and a monomer-dimer equilibrium with mass conservation model was used to fit the data. The monomeric molecular weights of the relevant protein forms 56  was fixed (30,748 and 31,400 Da for the apo- and holo-ChaN, respectively) as was the extinction coefficients (84,558 M-1 cm-1 and 53,100 M-1 cm-1 for apo- and holo-ChaN, respectively; note the extinction coefficient for the apo-protein  is the one for the interference optics and not for absorbance). The values of logKa and protein concentration were allowed to float except that protein concentrations of the sample in the same cell at different speeds were linked and kept constant.  2.9.2. P19 The apparent molecular mass of apo- and copper-loaded P19 was determined by gel-filtration chromatography. 500 ?l samples of protein were loaded onto a Superdex 200 10/300 GL column pre-equilibrated with 50 mM Tris pH 8, 100 mM NaCl at 0.4 ml/min at room temperature. The samples and standards used were as follows: 3 mg (apo/holo) P19, 5 mg bovine serum albumin (67 kDa), 6 mg ovalbumin (45 kDa), and 3 mg equine heart myoglobin (18 kDa).  Wild-type C. jejuni strain NCTC 11168 grown under iron-restricted conditions were subjected to in vivo cross-linking studies with DSP (Thermo Fisher Scientific Inc.) to detect the P19 dimer, as per manufacturer?s instructions. After separation of the cell lysate by SDS-PAGE under non-reducing conditions, the sample was either transferred to a nitrocellulose membrane and probed for P19 or excised and soaked in 5x SDS-PAGE loading buffer containing ?-mercaptoethanol for 25 minutes to break DSP cross-links. The lane was then laid on top of another 12.5% acrylamide gel, ran in the second dimension and then transferred to a nitrocellulose membrane and probed for P19.     57  2.10. FETP MRNA QUANTIFICATION  To examine the level of FetM and FetP expression in the iron limited growth study, strains ECA458-Gm, ECA458-fetM, ECA458-fetP and ECA458-fetMP were cultivated in TMM containing 2 ml glycerol and 3 g casamino acids per liter without additional iron until the turbidity reached 100 Klett units. Total RNA was isolated as previously described (168). DNase-treatment was performed, followed by purification with phenol/chloroform and precipitation with ethanol. RNA concentration was determined photometrically and RNA quality was checked on formamide gels. To exclude experimental artifacts resulting from DNA contamination, only RNA that did not generate products in a PCR reaction with chromosomal primers was used. For the reaction, 2 ?g of total RNA and 0.1 ?g hexamer primers were incubated at 65 ?C for 5 min and cooled on ice. After addition of 0.5 mM each of dATP, dGTP, dTTP and dCTP, 20 mM DTT and 150 U of reverse transcriptase (Superscript II, Invitrogen) in reaction buffer, reverse transcription proceeded for 10 min at room temperature, followed by 1 hr at 50 ?C. The reverse transcriptase was inactivated at 70 ?C for 10 min. The resulting cDNA was amplified by PCR as published (168) and separated on an ethidium bromide-stained agarose gel. Reverse transcriptase-based amplification of rpoZ-specific RNA served as loading and process control. No fetP-specific band was visible when the RNA originated from a fetM-carrying strain or the negative control; no fetM-specific band was visible when the RNA originated from fetP-containing bacteria or the negative control. However, fetP-specific bands occurred when the RNA came from fetP- and fetMP-carrying strains and fetM-specific bands could be seen when the RNA came from fetM- or fetMP-possessing cells.     58  2.11. SEQUENCE ANALYSES  Putative homologs of ChaN and P19 were identified and their sequences were obtained from the non-redundant database at the National Center for Biotechnology Information utilizing BlastP (http://www.ncbi.nlm.nih.gov/BLAST/). The criterion used to identify closely related homologs was an E-value cutoff of 4?10-4 for ChaN and 3?10-9 for P19. For each protein alignment, a single representative sequence with the lowest E-value was chosen for each species. The sequences were aligned with ClustalX (169) and manually edited with BioEdit (170). A tree was generated using a maximum likelihood analysis with TREE-PUZZLE (171). The parameters used were 3,000,000 puzzling quartets and exact parameter estimates.   59  CHAPTER 3. COFACIAL HEME BINDING IS LINKED TO DIMERIZATION BY C. JEJUNI CHAN 3.1. INTRODUCTION ChaN is a putative lipoprotein thought to associate with the outer-membrane and interact with ChaR (9). PhuW shares 30% sequence identity with ChaN and has been shown to be important in heme acquisition in P. aeruginosa. Knockouts of either PhuR or PhuW exhibit significantly reduced growth on media containing heme as the sole iron source (108). Two global approaches also support a link between ChaN and iron acquisition. Holmes et al. showed that chaN transcript levels increased 25-fold for cells grown under iron-limited conditions as compared to iron-rich conditions (91). Similarly, Palyada et al. demonstrated a decrease in transcription of chaN upon the addition of ferrous sulfate to C. jejuni grown in iron-limited media (4). Both of these studies show that ChaN expression is repressed by iron and are consistent with its control by the ferric uptake regulator.  To explore the role of ChaN in heme acquisition, recombinant chaN was expressed and purified. The crystal structure of ChaN, the first structural representative of a diverse group of heme-binding proteins, reveals unprecedented dimerization through coordination to two cofacial heme molecules. To gain insight into some of the characteristics and consequences of this unusual heme binding motif, the spectroscopic properties and oligomeric state of ChaN and the heme-bound complex have been studied in solution. To elucidate the in vivo role of ChaN (and ChaR), the localization of ChaN was examined and iron-limited growth studies were performed on a chaNR deletion mutant.   60  3.2. RESULTS 3.2.1. OVERALL STRUCTURE To define the specific interactions of heme with ChaN, the crystal structure was determined to a resolution of 1.9 ? (PDB ID: 2G5G) from a crystal produced grown from a solution of 15% PEG 4000 and 100 mM HEPES, pH 7. Data collection and refinement statistics are listed in Table 3-1. A single ChaN monomer is found in the asymmetric unit and consists of two domains that are connected by a loop and an ?-helix (Figure 3-1a). Domain I consists of an eight-stranded ?-sheet sandwiched by six ?-helices. In the mixed ?-sheet, one strand is anti-parallel (?2) and only strands ?1 and ?2 lack an intervening helix. Interestingly, these latter two strands are composed of the N- and C-terminal segments of the polypeptide chain (Figure 3-1a). Domain II is composed mostly of three ?-helices. The termini of helices ?3 (residues 97-99) and ?5 (residues 117-119) are distorted to 310 helices. Residues 252-254 also form a small 310 helix. The electron densities at the N- and C- termini are poorly defined; thus, ChaN was modeled from residues 9 to 263.     61  Table 3-1. Data collection and refinement statistics for ChaN  SeMet-labeled ChaN   Resolution (?) 36.00-1.90 (1.95-1.90)* Rmerge 0.084 (0.333) Average I/?I 8.6 (2.7) Completeness (%) 99.0 (97.9) Redundancy 3.5 No. reflections 28,745 Rwork / Rfree 0.203 / 0.239 Number of atoms      Protein 2072     Heme 43     Water 243 B-factors (?2)      Protein 33.7     Heme 52.8     Water 42.8 r.m.s.d. bond lengths (?) 0.014 r.m.s.d. bond angles (?)    1.429   * Highest resolution shell is shown in parentheses 62   Figure 3-1. The crystal structure of heme-bound ChaN  (a) Tertiary structure of a single ChaN molecule depicting the location of heme-coordinating Tyr148 and N- and C-termini. The heme ring is colored in purple. (b) Dimeric holo-ChaN enclosing two cofacial heme molecules. Covalently lipidated N-termini of holo-ChaN may insert into a membrane situated above the plane of the figure. Blue and teal are used to indicate the two ChaN monomers. The heme and heme ligands are depicted as ball-and-stick representations and are colored in orange. (c) 2Fo - Fc electron density of the ChaN Tyr148-heme interaction, contoured at 0.8 ?. The tyrosines and neighboring Lys189 are colored the same as the monomers in (b). Water molecules are colored brown.   63  Surprisingly, the heme is not found in the inter-domain cleft but is bound at the surface of domain II (Figure 3-1). The phenol oxygen of Tyr148 is coordinated to the heme iron (2.5 ?) and is hydrogen bonded to a water molecule (2.6 ?). The NZ atom of Lys189 is 3.8 ? away from this water molecule and thus may affect the environment of the water and tyrosine ligand. The tyrosine phenol ring, located above pyrrole I and translated slightly towards pyrrole IV (Figure 3-2), forms a 105.5? angle with the ligand bond.     Figure 3-2. Stereo representation of the ChaN-heme interactions 2Fo - Fc representative electron density of one heme, the Tyr148 ligand, and other nearby residues from both ChaN monomers, contoured at 0.8 ? (blue) and 5.0 ? (cyan). The same coloring scheme as in Figure 3-1 is used. An asterisk is used to indicate the residues that originate from the symmetry related monomer.    64  The heme is located near a crystallographic two-fold axis such that a dimer is formed (Figure 3-1b). The two heme planes are separated by ~3.5 ? and the distance between the heme irons is 4.4 ?. Approximately two-thirds of each tetrapyrrole ring overlaps with the other ring. The cofacial hemes are enclosed by a pocket formed by a dimer of two ChaN monomers. This pocket is lined with the hydrophobic residues Leu178, Leu186, and Leu190 (Figure 3-2). The NZ atom of Lys197 is positioned to form a nearly ideal hydrogen bond (2.8 ?) to one of the heme propionate groups. The other propionate forms a second, longer hydrogen bond (3.3 ?) with poorer geometry to the imidazole group of His176. Both Lys197 and His176 are derived from the symmetry related monomer (Figure 3-1b). Thus, the heme dimer serves as a bridge linking the ChaN monomers. Formation of the ChaN dimer buries about 680 ?2 in solvent accessible surface area of one monomer, of which 300 ?2 is associated with the polypeptide chain and 380 ?2 with the heme (as calculated by the program AREAIMOL) (172). The solvent-exposed surface area of each heme group in the dimer is 148 ?2. The heme iron is displaced 0.2 ? from the tetrapyrrole plane towards the coordinating Tyr148 (as calculated by the program GEOMCALC) (172). The crystal structure shows that dimerization does not occlude the N-termini, which would contain the cysteine lipidation site predicted by the program PROSITE (173) for the native protein sequence. In fact, the N-termini of both monomers are located away from the dimer interface and lie approximately on the same plane at the surface of the dimer, allowing for membrane anchorage of the native protein (Figure 3-1b). A database search of known protein folds with the program DALI reveals that the structure of ChaN is most similar to proteins in the P-loop containing nucleoside triphosphate hydrolases superfamily, as classified by the SCOP database (174). The hydrolase with greatest 65  structural similarity to ChaN is the ArsA ATPase (Z-value 6.4) (175). Structural alignment of ArsA with ChaN (r.m.s.d. 4.0 ? over 149 C? atoms) reveals that the P-loop, a conserved binding motif for the phosphate moiety, is absent in ChaN. Both ChaN and ArsA are dimers; however, the dimer interfaces do not coincide. Furthermore, the nucleotide binding region of ArsA found at the dimer interface is occluded partially by a loop in ChaN. No structurally similar proteins that bind heme were found in the DALI search.  3.2.2. SPECTROSCOPIC ANALYSES The electronic absorption spectrum of apo-ChaN (Figure 3-3a) exhibits a broad band at 280 nm with a molar absorptivity of 26,200 M-1cm-1 determined according to the method of Gill and von Hippell from the absorbance of the protein when denatured in a 6 M guanidine hydrochloride solution (176). Complete denaturation of the protein was verified by loss of ellipticity between 211 and 250 nm. Extension of this measurement to shorter wavelengths was not possible because of the excessive absorbance of the solution despite the use of recrystallized guanidine hydrochloride. Addition of heme to apo-ChaN at pH 7.4 results in the development of the Soret band at 402.9 nm with a shoulder at ~370 nm and the Q1 and Q0 bands at 508.3 and 533 nm, respectively (Figure 3-3a). An additional absorbance maximum characteristic of the high-spin state of the ferric heme iron occurs at ~630 nm. At more alkaline pH, the Soret maximum shifts to 411 nm as it increases in intensity relative to the shoulder at 370 nm. Concurrently, the resolution of the Q1 and Q0 bands improves as these maxima shift to ~533 and ?561 nm, respectively. Although diminished in intensity, the band at 625 nm persists. The effect of increasing pH presumably stems from the ionization of the axial tyrosyl ligand, the water molecule to which it may form a  66   Figure 3-3. Heme and pH titration effects on ChaN spectra (a) Electronic absorption spectrum of apo-ChaN (? 12 ?M protein in 30 mM sodium phosphate buffer, pH 7.4, 25 ?C; dotted curve) after successive additions of heme (?8% saturation with each addition; solid curves and dash-dot combinations). The numbers identify peak and shoulder wavelengths. (b) pH titration of holo-ChaN in a buffer cocktail containing 20 mM of each Bis-Tris buffer, MOPS, TEA, TAPS, CAPS, boric acid, and 50 mM sodium chloride. The absorption spectra are represented as the first derivative to reduce distortions due to small but detectable light scattering that results from changing the pH. The inset illustrates the change in the first derivative measured at 420 nm as a function of pH, and the solid curve represents the result of fitting these data to a model for two deprotonations to yield pKa values of 7.77 and 9.05 (?0.06). (c and d) Comparison of the CD spectra of apo- (dotted curves) and holo-ChaN (solid curves; ? 12 ?M protein in 30 mM sodium phosphate buffer, pH 6.5, 25 ?C) collected in the visible to far-UV region (c) and in the near-UV (d) collected with a 1 and a 0.1 cm pathlength cuvettes, respectively.     67  hydrogen bond, or the nearby NZ group of Lys189. A spectrophotometric pH titration of the protein yielded a pKa of 7.73 (?0.06) for the conversion to the alkaline form, and it also revealed a second heme-linked ionization with a pKa of 9.05 (?0.06) (Figure 3-3b). Whereas neither the apo-protein nor unbound heme exhibits a significant CD transition in the Soret region, a bisignate feature is present in the spectrum of the ChaN-heme complex that is characterized by positive and negative ellipticities at 428 and 387 nm, respectively (Figure 3-3d). Further comparison of the near UV-CD spectra of apo-ChaN and the ChaN-heme complex reveals that heme binding or the ensuing dimerization of ChaN causes the ellipticity at ~260 nm to increase (Figure 3-3d). This spectroscopic change suggests that the environment of one (or more) phenylalanine residues, such as Phe169 and Phe196 that are located on Domain II, changes to reduce the conformational flexibility of this side chain(s) even though no changes are detected in the far-UV region of the spectrum (Figure 3-3c) to indicate accompanying differences in the secondary structure.   3.2.3. ANALYTICAL ULTRACENTRIFUGATION To define the dependence of ChaN dimerization on the binding of heme, analytical ultracentrifugation studies were performed. Sedimentation velocity data obtained for apo-ChaN at 20 ?C and varying protein concentrations (8 - 80 ?M) were analyzed with the program SEDFIT to obtain c(s) distribution plots (165). In Figure 3-4a, the results obtained for samples at low concentration (8 ?M) exhibit a single peak that sediments at approximately the S-value expected for a monomeric protein. As the concentration of apo-protein was increased (30, 50 and 80 ?M), a small shoulder started to appear at higher S values, indicating that apo-ChaN can dimerize in solution (Figure 3-4a). Corresponding analysis of the holo-protein included detection 68  of heme binding by monitoring sedimentation at 400 nm as shown in the c(s) plot for two concentrations of this form of the protein (18 and 36 ?M) (Figure 3-4b). In contrast to apo-ChaN, the main component of holo-ChaN sediments at approximately the value expected for a dimeric species, and a shoulder with a smaller S-value is observed. Thus, holo-ChaN is also a self-associating protein, but the dimeric species dominates at the concentrations studied here.      Figure 3-4. Sedimentation velocity analysis of ChaN (a) c(s) distribution at of apo-ChaN:  gray line, 8 ?M; black line, 80 ?M. (b) c(s) distribution of holo-ChaN: open gray circles, 18 ?M; open black circles, 36 ?M.   69  The position of the shoulders in the c(s) plot does not reflect the true values for the monomeric or dimeric sedimentation coefficients. In fact, for proteins that interact on the time-scale of sedimentation, the interconversion of species arising from dissociation and re-association contributes to boundary spread. This effect results in a shift in the position of the peaks compared to the true sedimentation values in the c(s) analysis, particularly for species that are less abundant. To obtain association constants and true sedimentation coefficients for both the monomer and the dimer, the data were analyzed with the program SEDPHAT (167). At low protein concentrations (8 and 15 ?M), these data could be fitted as a single monomeric species. At higher protein concentrations (30, 50 and 80 ?M), the data were better fit as a monomer-dimer equilibrium having a dissociation constant in the low mM range (0.25-0.4 mM) with randomly scattered deviations. The data obtained in the presence of heme were fitted globally for a monomer-dimer equilibrium, but in this case, the fitted dissociation constant was much lower, ~4 ?M. To obtain more precise association constants, sedimentation equilibrium experiments were performed at several speeds (12,000, 20,000 and 30,000 rpm) and three different protein concentrations (apo-ChaN: 24, 76 and 167 ?M and holo-ChaN: 3, 14 and 21 ?M). The results were fitted globally for a monomer-dimer equilibrium with the program SEDPHAT (167). For the apo-protein a value of log Kd of 3.24 was obtained, which corresponds to Kd of ~0.57 mM; the value obtained for the holo-protein was much lower (log Kd= 5.03, which corresponds to a Kd of ~9 ?M). Therefore, as observed in the velocity experiments, we notice a considerable decrease of the Kd of holo-ChaN.    70  3.2.4. SEQUENCE ANALYSES The sequences of 22 identified ChaN homologs were derived from Gram-negative genera and aligned to reveal key conserved residues (data not shown). The tree constructed from the alignment revealed four clusters (Figure 3-5). Cluster I (average 92% sequence identity with ChaN) contains the sequences from the campylobacters. Cluster II (~30% sequence identity with ChaN) consists of various bacteria, including pathogenic E. coli, Shigella spp. and Pseudomonas spp., the insect pathogen Photorhabdus luminescens and soil-dwelling Azotobacter vinelandii. Clusters III and IV (19-28% sequence identity with ChaN) is comprised of more distantly related homologs. Tyr148, the heme-iron coordinating residue, is conserved within Clusters I and II. A role for His176 and Lys197 in heme binding through hydrogen bond formation is supported by the conservation of these residues. His176 is absolutely conserved within Clusters I, II and III. Lys197 is conserved strictly amongst Cluster I (the sequenced Campylobacter spp.) and in Bacteroides fragilis, whereas the functionally similar arginine residue is found at the same position in all other members of Clusters II and III. Notably, the ChaN homolog in P. aeruginosa that has been implicated in heme transport, PhuW, contains the key tyrosine, histidine and arginine residues and lies within Cluster II.    71   Figure 3-5. Unrooted tree of ChaN homolog sequences The abbreviations used are (C.ups) Campylobacter upsaliensis RM3195 ZP_00371681, (C.jej) C. jejuni subsp. jejuni NCTC 11168 NP_281387, (C.col) C. coli RM2228 ZP_00367783, (P.lum) Photorhabdus luminescens subsp. laumondii TTO1 NP_929952, (S.son) Shigella sonnei Ss046 YP_311520, (E.col) Escherichia coli serotype 0164 AAK67301, (S.boy) Sh. boydii Sb227 YP_406308, (P.flu) Pseudomonas fluorescens PfO-1 YP_350525, (P.put) Ps. putida KT2440 NP_746795, (A.vin) Azotobacter vinelandii AvOP ZP_00419614, (P.aer) Ps. aeruginosa PAO1 AAC13284, (G.met) Geobacter metallireducens GS-15 YP_383317, (G.sul) G. sulfurreducens PCA NP_954180, (V.cho) Vibrio cholerae RC385 ZP_00752108, (B.fra) Bacteroides fragilis NCTC 9343 YP_213053, (S.fum) Syntrophobacter fumaroxidans MPOB ZP_00667507, (A.aeo) Aquifex aeolicus VF5 NP_214019, (P.sp.) Polaromonas sp. JS666 ZP_00505786, (R.fer.) Rhodoferax ferrireducens DSM 15236 ZP_00691601, (A.deh.) Anaeromyxobacter dehalogenans 2CP-C ZP_00402047, (R.met.) Ralstonia metallidurans CH34 ZP_00594014, (S.pom) Silicibacter pomeroyi DSS-3 YP_165670, (M.mag) Magnetospirillum magnetotacticum MS-1 ZP_00207911. The tree was generated with TREE-PUZZLE (171) and displayed with TreeView (177).    72  3.2.5. CHAN IS LOCALIZED TO THE OUTER MEMBRANE The ChaN N-terminal signal peptide sequence contains a predicted lipobox, suggesting the protein is membrane-bound. To examine the cellular localization of ChaN, wild-type C. jejuni cells were grown in iron-limited minimal media and harvested. The soluble extracellular, periplasmic and cytoplasmic fractions were then isolated from the membrane component, separated by SDS-PAGE and examined by western blot analysis. ChaN was not detected in any of the soluble fractions, but is found solely in the membrane fraction, supporting lipid anchoring of ChaN (data not shown). The membrane fraction was separated further with the detergents Triton X-100 or sarkosyl to differentially solubilize the inner from the outer membrane, as had been previous performed on C. jejuni (151,152). The OM and IM detergent-solubilized fractions were separated by SDS-PAGE and examined by western blotting using the OM protein CadF as a positive control (178). From this assay, ChaN appears to localize primarily within the OM protein-enriched sample, although a noticeable proportion of ChaN is found in the IM protein-enriched fraction (Figures 3-6a and b). The IM ChaN fraction may represent a pool of protein in transit to the OM or the result of weak OM interactions resulting in incomplete separation during detergent solubilization.  To further support OM localization of ChaN, the accessibility of ChaN to surface protease digestion was examined using CadF and P19 as OM and periplasmic controls, respectively. To pretest susceptibility of ChaN and P19 to trypsin digestion, equivalent concentrations of both recombinant proteins were exposed to trypsin and analyzed by SDS- PAGE. Based on the degradation rate, P19 appeared to be more susceptible than ChaN to trypsin (data not shown). C. jejuni cells were then either exposed to trypsin or were left untreated for 24  73    Figure 3-6. ChaN is likely localized to the outer membrane (a) C. jejuni cells grown under iron-limited conditions were lysed and the membranes were fractioned using 1% (v/v) Triton X-100. The fractions, representing enriched outer and inner membrane protein samples, were separated by SDS-PAGE and analyzed by western blot using either anti-ChaN or anti-CadF antibodies as an OM control. Due to detector saturation by ChaN, both lanes in the ChaN blot contained 40% of each fraction sample as compared to their equivalent lane in the CadF blot. Due to detergent effects, the IM samples run at a higher molecular weight. A white arrow indicates the position of the CadF bands. (b) Densitometric analysis of the bands in (a) showing the relative distribution of each protein to the two membranes. (c) Western blot of C. jejuni cells grown under iron-limited conditions and exposed to the protease trypsin for 24 hrs at 37 ?C as compared to untreated cells under the same conditions. Antibodies against P19 are used as a periplasmic control. (d) Densitometric analysis of the bands in (c) showing the proportion of protein remaining post-digestion as a ratio of the untreated control.   74  hrs at 37 ?C, harvested, separated by SDS-PAGE, and examined by western blotting (Figures  3-6c and d). Only a small proportion of soluble periplasmic P19 was degraded, indicating that in general, cellular structural integrity remained intact. More than half of the ChaN and CadF is digested upon trypsin treatment. As was previously observed in studies of CadF, this adhesin separates into two bands on a SDS-PAGE, which are both degraded to the same degree. Together, this suggests that ChaN is more surface-exposed than P19 and supports the predicted OM localization. Although it is tempting to speculate surface localization of ChaN, fidelity of membrane lipoproteins to its native leaflet may be lost during the duration of the experiment.    3.2.6. A CHANR DELETION MUTANT EXHIBITS DELAYED GROWTH To determine the role of ChaNR in the growth of C. jejuni under varying iron sources, time course liquid culture growth experiments were performed on wild-type strain NCTC 11168 and the chaNR deletion strain (11168?chaNR) (Figure 3-7). Prior to the start of each experiment, all strains were grown overnight in minimal media to reduce cellular iron content. At the start of each experiment, the growth of all strains exhibited an initial lag phase for at least 6 hrs before a detectable increase in optical density measurements was observed at 600 nm. No growth of either strain is observed under extreme iron limitation, as achieved by supplementation of the defined minimal media with 1 ?M of the iron chelator, desferrioxamine (data not shown). In Figure 3-7a, representing a growth comparison between unsupplemented minimal media and media supplemented with either FeSO4 or transferrin as the sole iron source, demonstrates that the cells are able to reach higher cell densities when iron is provided, whether it is in free form or is protein-bound. 11168?chaNR also appears to exhibit a longer lag period than the wild-type 75   Figure 3-7. The chaNR deletion mutant exhibits delayed growth Liquid cultures of C. jejuni wild-type 11168 and the deletion mutant 11168?chaNR were inoculated at OD600 nm 0.025 into MEM? media and grown for 31 hrs without supplementation or supplemented with Fe2SO4, holo-transferrin or holo-lactoferrin. Optical density at 600 nm was assessed from duplicate cultures. Panels (a) and (b) represent two separate experiments with standard deviation error bars drawn.    76  strain under all three conditions, but is eventually able to reach the same cell density as the wild-type. However, the growth difference between the wild-type and the deletion mutant is statistically significant (paired t test, P < 0.05) only for the iron-supplemented samples at the 13 hr time point and the transferrin-supplemented samples at the 25 hr time point due to fewer sample sizes (n=2), which was chosen due to technical considerations. Improvements in the statistical significance values will require repeating the experiment with an increased number of replicates. Based on the structural similarity between the iron-binding proteins transferrin and lactoferrin, growth of the two strains on these two iron sources was also compared (Figure 3-7b). Similar to growth on transferrin, a longer growth delay was observed on lactoferrin as the sole iron source. Notably, the generally observed variability in Campylobacter growth is exemplified by the growth differences of the two strains under repeated conditions, as depicted in Figures 3-7a and b, limiting direct comparisons between repeated trials.  The rate at which the deletion mutant exits lag phase in Figure 3-7b appears to be faster, resulting in reaching an equivalent OD600 nm sooner as compared to the wild-type for growth on transferrin. At the 13 hr time point, there is also a less dramatic cell density difference between the two strains under no supplementation. Again, only the transferrin-supplemented samples at the 13 hr time point show a statistical difference (P < 0.04) between the wild-type and the double deletion variant likely owing to the sample size (n=2).  3.3. DISCUSSION For a protein with transient heme binding properties, the oxidation state of the heme iron is a fundamental factor in heme binding and release. The two major determinants of the reduction potential that dictate the oxidation state of the heme iron are the axial ligands and the 77  dielectric of the heme environment (179,180). The heme irons of holo-ChaN are each coordinated by a single Tyr ligand (Figure 3-1). Although examples of heme proteins with this type of coordination environment for which electrochemical data are available are limited, a myoglobin variant with this axial ligation exhibits a lower potential than that of wild-type myoglobin (181). The effect of dielectric on the reduction potential of heme and heme proteins as first recognized by Kassner (182,183) may reflect the accessibility of the heme prosthetic group to solvent (184), but the environment provided to the heme by the protein can also vary sufficiently to influence this parameter independent of solvent accessibility (185,186). The crystal structure of holo-ChaN indicates that ~18% of the heme surface is exposed to solvent. The corresponding value of 23% is observed for heme bound to the hemophore HasA, in which Tyr and His provide axial ligands to a single heme group (187). These structural similarities suggest that the reduction potential of heme bound to ChaN should exhibit a reduction potential similar to that observed for HasA (-550 mV), consistent with our observation that holo-ChaN is not reduced by ascorbate (data not shown). A low reduction potential implies relative stabilization of the ferric state or destabilization of the ferrous state as observed for Thiosphaera pantotropha cd1NiR, in which a distal tyrosine coordinates to the heme iron in the oxidized state but becomes completely disordered on reduction of the enzyme (188,189). In light of the relatively complex behavior of holo-ChaN in solution (vide infra), electrochemical characterization of this system is likely to be challenging. The unprecedented nature of heme binding to ChaN defined by the structure determined in this work raises a number of questions regarding the spectroscopic and functional properties of such a structure that require characterization of the behavior of ChaN in solution. The spectroscopic and sedimentation velocity experiments provide initial insight into some of these 78  issues. For example, the ultracentrifugation studies provide initial quantitative information regarding the magnitude of the contribution made by heme binding to the dimerization of ChaN and are fully consistent with the observations in the previous paragraph.  The spectroscopic results demonstrate that the behavior exhibited by ChaN in solution is relatively complex. For example, the incremental addition of heme to apo-ChaN results in progressive changes in the electronic spectrum (Figure 3-3a) that are not readily transformed into unambiguous ASoret vs. [heme]/[ChaN] plots of the type classically obtained during similar titrations of apo-myoglobin. Notably, similar titrations of ShuT also result in less than ideal plots of this type (Figure 2a of Eakanunkul et al.) (190). Thus, it appears that heme proteins of this general class are not readily amenable to this type of analysis for reasons that are not yet apparent and an ultimate understanding of the mechanism by which ChaN functions in vivo may require consideration of additional factors that remain to be identified. Nevertheless, the pH-dependence of the resulting holo-ChaN complex is relatively well-behaved and exhibits two clear transitions that are readily fit to two pKa values (Figure 3-3b).  UV-vis spectroscopic studies of ShuT have shown that heme binding is insensitive to pH (190). In contrast, the absorption spectrum of holo-ChaN exhibits a strong pH dependence (Figure 3-3b). With other heme proteins, one major contribution to such behavior is the deprotonation of a coordinated water molecule to form a distal hydroxyl group ligand that results in a low-spin species or a spin-equilibrium. A related process could operate in the current case. The other titratable groups that are located in the vicinity of the coordinated Tyr residue are the water molecule with which it forms a hydrogen bond and Lys189, which is 3.8 ? from this water molecule. Small structural changes in this region of the protein could also contribute to the complex changes in solution properties observed as a function of pH. The binding of heme to 79  ChaN may also be subject to regulation by other proteins in vivo as proposed for the release of heme from HasA (187). A single axial tyrosine ligand is also found in heme transport proteins such as human serum albumin (HSA) (191) and the bacterial periplasmic binding protein ShuT (190).  Interestingly, of the heme transport proteins of known structure, the angle between the tyrosine ring and the ligand bond is comparable: holo-ChaN (105.5?), HSA (100.5?) (191) and HasA (116.4?) (187). In comparison, the crystal structures of enzymes with a tyrosyl heme ligand, such as human erythrocyte catalase (192) and cytochrome cd1 nitrite reductase (cd1NiR) (188), contain notably greater tyrosine-iron angles (135.2? and 134.5?, respectively). The tyrosine ligand in both ChaN and cd1NiR is located above pyrrole I and is translated slightly towards pyrrole IV (Figure 3-2). This orientation contrasts with human erythrocyte catalase, in which the tyrosine is situated directly above pyrrole ring IV of the heme (192) and other hemoproteins, where the tyrosine is located above the ?-meso carbon (187,191). Although these differences in the geometry of tyrosine coordination may relate to the need for reversible heme binding in transporters versus the catalytic properties of the enzymes, these differences are not discernable spectroscopically as the visible spectra of holo-ChaN (electronic absorption and CD spectra) resemble those of human erythrocyte catalase (193) and ShuT (190). Of the total solvent accessible surface buried upon dimerization of ChaN, more than 50% is contributed by the cofacial interaction of the heme groups. Heme binding stabilizes dimer formation further through hydrogen-bonding interactions of the propionate groups with residues (His176 and Lys197) of the opposing monomer (Figure 3-1b). The remaining buried surface is contributed primarily by inter-monomer contacts involving residues Val114, Arg141, Ile144, Thr145 and Asn149. As a result, the structural basis for the linkage of heme binding and dimer 80  formation observed by analytical ultracentrifugation is readily apparent. Although both apo- and holo-ChaN exhibit a monomer-dimer equilibrium, the dissociation constant for monomer formation decreases 60-fold with heme bound.  Dimeric heme has been observed previously. The classic example of such structures results upon exposure of ferrous heme to dioxygen to form ?-oxo dimers (194). Although the distance between the two heme irons (~4.4 ?) in the structure of holo-ChaN allows for the presence of the Fe-O-Fe linkage characteristic of ?-oxo dimers (Figure 3-1c), the displacement of the heme iron 0.2 ? out of the heme plane and towards Tyr148 is consistent with a five-coordinate iron group (His93Tyr myoglobin: 0.39 ?; human erythrocyte catalase: 0.14 ?) (181,192). Attempts to include an inter-heme oxygen atom resulted in an unrealistically high B-factor of the inserted atom (~82 ?2) during crystallographic refinement. Comparisons to a solution structure of a ?-oxo heme dimer analogue also reveal key differences. The Fe-Fe distance in the solution structure is ~3.5 ? (195). The greater separation observed in holo-ChaN (~4.4 ?) is due to the translation of one heme plane with respect to the other such that the hemes only overlap by approximately two-thirds. In contrast, the porphyrin rings are nearly completely overlapping in the ?-oxo heme dimer analogue. Some c-type cytochromes are described as possessing a pair of ?stacked? heme groups (196), but the arrangements of the hemes differ drastically from that observed in holo-ChaN. Firstly, the cytochrome hemes are generally arranged such that the propionate groups are oriented in the same direction whereas the ChaN hemes are arranged head-to-tail (Figure 3-1). Secondly, each of the hemes in cytochrome c is coordinated by two axial ligands, imposing a physical restraint that greatly limits the area of overlap between the two heme groups. Finally, the Fe-Fe distances within the cytochrome c family are 9 ? or greater in length (197,198) as 81  compared to 4.4 ? in holo-ChaN. Taken together, this intimate interaction of the two heme groups that is involved in the dimerization of ChaN may be functionally important in vivo.  The ability to transport heme is an important virulence factor for bacterial pathogens as heme is the most abundant source of essential iron within a mammalian host (35). Genetic mutation of the ChaN homolog PhuW in P. aeruginosa exhibited a clear growth defect using heme as the sole iron source (108). Additional homologs of ChaN are found in multiple pathogenic organisms; however, none of these has been characterized biochemically or structurally. ChaN is the first member of its lipoprotein family to be shown to bind heme specifically. We have also demonstrated that ChaN is a membrane-anchored protein likely localized to the OM (Figure 3-6). Localization of ChaN to the OM would allow interaction with ChaR, a homolog of TonB-dependent OM receptor proteins that is encoded by a gene located adjacent to chaN. A characterized heme uptake system in C. jejuni is ChuABCD (199). Disruption of ChuA, the outer-membrane receptor, inhibited growth on heme, hemoglobin, or hemoglobin-haptoglobin as a sole iron source. In contrast to PhuW knockouts, mutants of chuB, chuC, or chuD, which encode the periplasmic and inner-membrane components of the transport system, exhibited no obvious phenotype. Although both ChuA and ChaR are iron-regulated proteins (4,91), the expression of ChaR was found to be significantly upregulated during growth in the rabbit ileum, a mammalian model for human gastroenteritis due to C. jejuni infections expression (105); the expression of ChuA, on the other hand, was downregulated in the same study. This suggests that there are other regulatory factors that control expression in conjunction with iron availability and that ChaR plays a greater role during colonization of the gut.  82  Hemolytic activity has been previously observed in some strains of C. jejuni and a frequent symptom of infection by C. jejuni is bloody diarrhea (68,90,200), which would lead to the exposure of serum hemoproteins during infection. The need for multiple heme uptake systems may reflect the diversity of heme protein sources. Growth on the hemopexin-heme and haptoglobin-hemoglobin complexes as sole iron sources have been described (90). The receptors for hemopexin or haptoglobin have yet to be identified in C. jejuni and ChaR is an excellent candidate.  The destination of heme bound to ChaN is currently unknown. As all characterized heme import systems involve an ABC transporter, a periplasmic binding protein such as ChuD is a likely heme acceptor. Alternatively, ChaN may act as a chaperone that delivers externally supplied heme to periplasmic hemoproteins such as cytochromes. Similar to holo-ChaN, structures of copper and nickel metallochaperones have revealed dimeric proteins in which the metal ion binds at the subunit interface (201,202). Proposed models for metal ion delivery involve substitution of one monomer by the target protein. The heme-dependent dimerization of ChaN suggests that an analogous mechanism may operate in heme transport.  When comparing the growth of a chaNR double deletion mutant with the wild-type strain under iron-limited conditions or when an iron source is presented, an increased lag phase is observed for the mutant, which then recovers and reaches the same cell density as the wild-type strain. In contrast, growth studies using a single chaR deletion mutant by Miller et al. exhibited a noticeable growth defect as compared to the wild-type strain (106). This may be explained by technical differences between the experimental setups; however, it may also be an indicator of negative effects on transferrin/lactoferrin iron uptake by the presence of ChaN without ChaR. 83  Nonetheless, the interesting heme-binding properties of the lipoprotein ChaN requires further biochemical and in vivo studies to elucidate its exact role in C. jejuni iron metabolism.    84  CHAPTER 4. COPPER BINDING AND IRON TRANSPORT BY C. JEJUNI P19 4.1. INTRODUCTION P19 is an acidic, periplasmic protein that was originally isolated as a major band from a glycine extraction step of wild-type C. jejuni cells (112). It is a Fur-regulated protein, but was not immunogenic during natural infections in humans and the function of P19 remained unclear (112,113). The derepression of P19 expression in a Fur deletion mutant was later complemented by microarray studies that link p19 regulation and iron availability (4,91). Since its discovery, the predicted iron uptake function of P19 has been supported by studies on homologs of P19, although not by a common mechanism. ChpA, a P19 homolog from the marine magnetotactic vibrio strain MV-1, plays a key role in the formation of magnetosomes, presumably by ensuring a sufficient iron supply. (115). Native ChpA was purified from the periplasm of MV-1 and was found to have copper bound. T. pallidum Tp34 has been shown to be periplasmic but binds the large glycoprotein lactoferrin with high affinity and a zinc-reconstituted crystal structure has been solved (114).  Here, we show that P19 is an iron-regulated protein that is crucial for growth of C. jejuni under iron-limited conditions. Furthermore, copper chelation experiments suggest a role for P19 in copper utilization. We have determined that recombinant P19 is capable of binding both ferric iron and copper in a 1:1 molar ratio and that copper binding is not inhibited by zinc. Dimerization of P19 is also detected both in vitro and in C. jejuni. To explore the mechanism by which metals can bind P19, crystal structures of copper, manganese and iron-bound P19 were solved, which reveal two distinct, but proximate metal binding sites.  85  4.2. RESULTS 4.2.1. THE EXPRESSION OF P19 PROTEIN IS IRON-REGULATED  To demonstrate that the iron-regulated expression of P19 in C. jejuni strain NCTC 11168  occurs upon iron restriction for C. jejuni strain 81-176 (4,91,113), wild-type cells were grown in Mueller-Hinton (MH) media with increasing levels of the iron chelator, desferrioxamine (DF). Equivalent amounts of cell material were separated using SDS-PAGE, transferred onto a nitrocellulose membrane, and probed with anti-P19 antibodies (Figure 4-1a). In the absence of iron chelation (iron-replete conditions), minimal levels of P19 were detected. As the amount of chelator increased, there was a corresponding increase in the expression of P19.  Figure 4-1. Iron-regulated P19 expression is important for growth under increasing iron limitation  (a) Log phase C. jejuni 81-176 grown for 3 hrs in MH broth supplemented with 0, 8 or 40 ?M of the iron chelator desferrioxamine. The cells were then harvested, separated by SDS-PAGE, transferred to a nitrocellulose membrane, and probed for P19. (b) Biphasic cultures of wild-type 81-176, 81176?p19, and trans-complemented 81176?p19C were inoculated at OD600 nm 0.018 and grown for 24 hrs with increasing DF supplementation. Due to variability in maximal growth achieved, the figure is a representative triplicate culture experiment with standard deviation error bars drawn but in some cases are too small to see. The growth differences were consistently observed in experimental replicates with independent cultures. P < 0.007 (paired t test) for the wild-type compared to the mutant under all DF supplemented conditions. 86  4.2.2. P19 IS REQUIRED FOR OPTIMAL C. JEJUNI GROWTH UNDER IRON RESTRICTION  To demonstrate that iron is important for growth of C. jejuni, a titration with increasing DF was performed (Figure 4-1b). In the absence of an iron chelator, wild-type 81-176, a p19 deletion strain (81176?p19), and the complemented strain (81176?p19C), in which P19 is under the control of the constitutive KanR gene promoter in a different region of the chromosome, all achieved almost equal growth after 24 hrs. With increasing concentrations of DF, wild-type and 81176?p19C demonstrate a gradual decrease in growth over the same period whereas strain 81176?p19 grew minimally at 16 ?M or higher DF. To further show that expression of P19 facilitates growth of C. jejuni upon iron restriction, a time course growth experiment was performed with 16 ?M DF. When wild-type, 81176?p19 and 81176?p19C strains were grown in rich MH media, all three strains grew comparably, as determined by optical density measurements and CFU/ml counts (Figures 4-2a and c). When iron was restricted by the addition of DF, the wild-type and 81176?p19C strains were still able to grow, albeit at a reduced rate; OD600 nm readings for the P19 deletion mutant, however, plateaued after 8 hrs (Figure 4-2b). The CFU/ml counts revealed that growth had not merely halted, but the number of culturable cells decreased as early as 4 hrs post-iron restriction (Figure 4-2d). A western blot was performed to confirm the lack of expressed P19 protein in the mutant strain (data not shown). Trans-complementation of the deletion strain 81176?p19 demonstrates near full recovery of the wild-type phenotype (Figures 4-2b and d). To determine if copper levels in the media affect the growth of the three strains under iron replete and limited conditions, increasing concentrations of the Cu+-specific chelator bathocuproine disulfonic acid (BCS) were used in combination with DF. Under iron-replete conditions (Figure 4-3a), increasing copper chelation reduces maximal achievable growth of all  87     Figure 4-2. The P19 deletion mutant is unable to grow under iron-limited conditions  Biphasic cultures of C. jejuni wild-type 81-176, 81176?p19 and trans-complemented 81176?p19C were inoculated at OD600 nm 0.020 and grown for 24 hrs without (a and c) and with 16 ?M DF supplementation (b and d). Optical density at 600 nm (a and b) and CFU/ml counts (c and d) were assessed from triplicate cultures. The figure is a representative triplicate culture experiment with standard deviation error bars drawn but in some cases are too small to see. For (b) and (d), P < 0.027 (paired t test) for the wild-type compared to the mutant at all experimental time points after time 0.   88     89  Figure 4-3. Growth of the P19 deletion mutant is less affected by copper chelation than wild-type C. jejuni   Biphasic cultures of C. jejuni wild-type 81-176, 81176?p19 and trans-complemented 81176?p19C were inoculated at OD600 0.020 in triplicate cultures and grown for 24 hrs with increasing DF (Panel (a), 0 ?M; Panel (b), 10 ?M; Panel (c), 14 ?M) and BCS (0, 30, 75 and 150 ?M) supplementation. The figure is a representative experiment with standard deviation error bars drawn but in some cases are too small to see.   three strains. Also, in the absence of the iron chelator the P19 deletion mutant achieves greater growth than either the wild-type or complemented (81176?p19C) strains under every level of copper chelation tested. Combining copper and iron chelation results in an additive step-wise reduction in growth for the wild-type and 81176?p19C strains. The maximal growth of both strains with 75 or 150 ?M BCS and 14 ?M DF is less than with either chelator alone (Figures 4-3a and c; P < 0.03 for all comparisons between single and double chelator achievable growth values). In contrast, the growth of the mutant (81176?p19) is largely unaffected by copper chelation in the presence of DF: no statistically significant differences in growth were observed with up to 75 ?M BCS under both moderate (10 ?M DF) and severe (14 ?M DF) iron-limited conditions (Figures 4-3b and c, respectively). Only under moderate iron-limitation was a small, but statistically significant difference observed for 81176?p19 growth (P = 0.011) when comparing 0 to 150 ?M BCS.  4.2.3. P19 BINDS BOTH IRON AND COPPER AT DIFFERENT SITES  Recombinant P19 was expressed and purified from E. coli. ?As isolated? P19 was found to have a mixture of metals bound (~75% Cu and ~20% Zn based on x-ray fluorescence scans) and did not exhibit any features in the visible spectrum above 300 nm (data not shown); protein 90  samples were treated with EDTA to produce apo protein for biochemical analyses. To measure the copper binding capacity of recombinant P19, excess cupric sulfate was added to apo-P19 and incubated on ice. Unbound metal was removed by gel filtration (Sephadex G-25 Fine column) and a bicinchoninic acid (BCA) assay with and without the reducing agent yielded 1.05 ? 0.20 Cu2+ ions and no detectable Cu+ per monomer of P19. To determine if either Zn2+ or Fe3+ could compete with Cu2+ binding, equimolar FeCl3 or ZnCl2 was simultaneously added with CuCl2 to apo-P19. After a 45 min incubation at room temperature and a gel filtration step, BCA assays yielded 0.99 ? 0.11 and 1.12 ? 0.02 copper per P19 monomer for the Zn2+ and Fe3+ competition experiments, respectively. To determine the iron binding capacity of copper-loaded P19, freshly prepared ferric chloride was aerobically incubated with Cu-P19 for 60 min. After gel filtration to remove unbound iron, ferene S and BCA assays showed 1.19 ? 0.12 Fe3+ and 0.88 ? 0.03 Cu2+ ions bound per monomer, respectively. Again, no Cu+ was detected using the BCA assay in the absence of reducing agent. Ferric iron was also incubated with copper-loaded P19 in the presence of the reducing agent dithionite to assess binding of the ferrous oxidation state. After removing unbound iron by gel filtration with buffer containing dithionite and subsequent buffer exchange, less iron remained bound in these samples (0.25 ? 0.01 per monomer). Similarly, the addition of ferrous iron with ascorbic acid as a less potent reducing agent resulted in little iron bound (0.20 ? 0.04 per monomer). The ability of P19 to oxidize the phenolic compound p-phenylenediamine, typical of MCOs, was assayed but yielded no activity (data not shown). Attempts to demonstrate P19-catalyzed electron transfer between NADH and ferric iron or ferricyanide by a previously published optical method did not reveal ferrireductase activity (203).   91  4.2.4. THE OVERALL STRUCTURE OF P19  To examine the metal binding sites of recombinant P19, a protein crystal grown in 2.0 M ammonium sulfate and 0.1 M citrate pH 5.5 was used to solve a low resolution structure in space group P6222. High resolution structures in space group P21212 were then solved up to 1.4 ? resolution from crystals grown from a solution of 25-50% PEG 250 or 350 and 0.1 M CHES pH 9-10 using the low resolution solution as an initial model. These structures include ?as isolated P19?, EDTA-treated P19 (apo-P19), ?as isolated? P19 crystals soaked with Mn2+ (MnCu-P19), and Cu2+-reconstituted apo-P19 (Cu-P19). Data collection and refinement statistics are included in Table 4-1. In all the structures, two P19 molecules are found in the asymmetric unit, forming a homodimer with the same overall fold that looks much like a deflated rugby ball (Figure 4-4). Unless otherwise indicated, structural analysis was performed on the ?as-isolated? P19 structure. From tip to tip, the P19 dimer is a little over 70 ? long, ~35 ? in height and only ~20 ? deep, making P19 a relatively flat macromolecule. Analysis of the P19 dimer with the program PISA indicates that the dimer interface involves 47 residues (or 177 atoms) from each monomer, 19 possible hydrogen bonds and 2 salt bridges (204). Formation of the P19 dimer buries around 1700 ?2 (or approx. 20%) of the solvent accessible surface area of each monomer with an estimated solvation free energy gain of -27.9 kcal/M upon interface formation. ProtorP analysis of the dimer interface determined a Gap Volume Index of 1.08 ?, indicating high complementarity and close packing within the dimer (205). B-factors for these dimer interface residues are among the lowest in the structure indicating limited disorder and motion.   92  Table 4-1. Data collection and refinement statistics for P19      ?As isolated?  Cu-P19  apo-P19   Data Collectiona Resolution Range (?)  19.21-1.45  19.34-1.65  19.11-1.59      (1.53-1.45)   (1.74-1.65)  (1.68-1.59) Cell dimensions (?)  a = 54.53  55.83   54.40       b = 73.54  72.58   73.04        c = 75.06  78.82   74.82   Wavelength (?)  1.127   0.979   1.282   Unique Reflections  53633   39293   40227   Completeness (%)  98.6 (97.3)  99.5 (98.7)  98.9 (98.6)  Average I/?I   14.8 (2.3)  20.2 (2.4)  16.6 (2.2)  Redundancy   4.4   4.7   3.7    Rmerge     0.040 (0.517)  0.040 (0.661)  0.037 (0.655)     Refinement  Rwork     0.141   0.139   0.142   Rfree     0.187   0.187   0.184   No. of waters   353   291   282    Avg. B-value (?2)  23.9   25.0   23.7   r.m.s.d. bond lengths (?) 0.017   0.015   0.014   Ramachandran plot      Most-favorable (%)  98.7    98.7   98.7     Allowed (%)  1.3   1.3   1.3   PDB code   3LZL   3LZO   3LZN   a Values for the highest resolution shell are shown in parenthesis    93       MnCu-P19-1A MnCu-P19-1.9A  Data Collectiona  Resolution Range (?)  19.21-1.41  19.26-2.73     (1.49-1.41)  (2.88-2.73) Cell dimensions (?)  54.36   54.50     73.62   73.80      75.05   75.22 Wavelength (?)  0.979   1.893 Unique Reflections  54201   7756 Completeness (%)  93.1 (76.8)  92.6 (78.0) Average I/?I   22.0 (3.1)  61.9 (47.3) Redundancy   4.0   3.8 Rmerge     0.028 (0.383)  0.015 (0.019)  Refinement Rwork     0.167   0.195 Rfree     0.208   0.245 No. of waters   317   56 Avg. B-value (?2)  21.3   31.5 r.m.s.d. bond lengths (?) 0.014   0.005 Ramachandran plot      Most-favorable (%)  98.4   98.4    Allowed (%)  1.6   1.6 PDB code   3LZQ   3LZR   a Values for the highest resolution shell are shown in parenthesis    94    Figure 4-4. The overall dimeric structure of the P19 monomer with bound copper   (a) Front view ribbon diagram showing a perpendicular two-fold rotational symmetry. Copper atoms are shown in orange. Copper ligands are represented as stick models. (b) Surface of P19 after 135? rotation of the dimer. Surface exposed Met86 is colored in yellow. All other regions are colored according to the chain. (c) Front and (d) Rear view of the distribution of conserved residues mapped onto the P19 dimer surface. Regions of highest conservation are shaded blue whereas least conserved regions are in red. White arrows denote the copper and putative iron binding sites. Yellow arrows denote the location of Met86. (e) Front and (f) Rear view of the electrostatic surface potential of the P19 dimer.  95  4.2.5. THE TWO METAL BINDING SITES ARE CO-LOCALIZED The P19 monomer consists primarily of a ?-sandwich formed from 4 and 5-stranded anti-parallel ?-sheets in an immunoglobulin-like fold. On the outside surface of the 5-stranded ?-sheet of each monomer in the ?as isolated? P19 structure at pH 9 is a copper ion coordinated by His42, Met88, and His95 (Figure 4-5; see Table 4-2 for ligand distances). Completing the copper coordination sphere is His132, which lies on a large loop extending over from the opposing subunit. Typical of Cu2+ coordination by imidazole and carboxylate groups in proteins (1), the four-coordinate copper site forms a distorted tetrahedral geometry in the ?as isolated? structure and is relatively buried as it is situated almost equidistant from both faces of the dimer (Figure 4-6; Table 4-2).   Figure 4-5. ?As isolated? and copper-reconstituted structures depicting the copper site ligands and the flexibility of Glu44  The chains of each monomer in the (a) ?as isolated? and (b) copper-reconstituted structures are colored according to Figure 4-4a. The copper cations are shown in orange. Nitrogens, oxygens, and sulfurs in the depicted side chains are shown in blue, red, and yellow, respectively. Ligand-copper distances are listed in Table 4-2. The electron density shown is a 2Fo-Fc map contoured at 1.0 ?.  96  Table 4-2. Metal ligand geometry in the primary site of P19 and Tp34a    As-isolated P19, pH 9 Cu2+-P19, pH 10  Zn2+-Tp34, pH 9b  Ligand bond length (?)  His42   2.1   2.0    2.2/2.1 His95   2.1   2.0    2.1 His132  2.1   2.1    2.0 Met88   2.3/2.4   2.8    3.1/3.0 Glu44   3.6/4.0   2.4-3.1c/2.2   2.1  Ligand bond anglesd (?)  His42-Cu-Met88 112/118 His95-Cu-Met88 103/108 Glu44-Cu-His42    81/82 His42-Cu-His132 107/108  100 His132-Cu-His95 98/104   96 His95-Cu-Glu44    79/82   a Values for each of the two molecules in the asymmetric unit are given when different. b PDB ID: 2O6E c Two conformations observed d Ligand bond angles are only provided for ligands with bond lengths shorter than 2.5 ?.  97   Figure 4-6. Manganese-soaked structure  The putative iron binding site ligands are situated at the front face of the P19 dimer and the methionine-rich region is located on the opposite face of the manganese binding site. The copper cation is shown in orange while manganese is shown in purple. Each chain and ligand is colored according to Figure 4-5. The electron density shown is a manganese absorptive edge anomalous map contoured at 4.0 ?.   In the Cu-P19 crystal structure at pH 10,  nearby Glu44 (situated distal to Met88) is within ligand bonding range (2.2-2.4 ?) to the copper and the Met88-S?-Cu2+ ligand bond is lengthened by ~0.4 ? resulting in a new coordination sphere (Glu44, His42, His95, and His132) with a geometry best described as square planar (Table 4-2). An x-ray fluorescence scan of the Cu-P19 crystal ensured the absence of significant amounts of Zn or Fe species in the sample. The two copper sites in the dimer are ~30 ? apart and are unlikely to interact directly. 98  Glu44 is part of a short channel lined with other acidic residues (in particular Glu3, Asp46 and Asp92) that leads from the surface of P19 to the primary copper site. Two of these symmetry-related acidic channels open to one face of the P19 dimer and contribute to two larger regions of negative electrostatic surface potential (Figures 4-4e and f). To determine if these sites could host a second metal, ?as isolated? P19 crystals were soaked with manganese chloride (producing MnCu-P19). Initial crystallographic attempts to directly show iron coordination have been unsuccessful likely due to the rapid oxidation and instability of iron at the high pH under which the P19 crystals were produced. Based on Hard-Soft ligand theory, both Mn2+ and Fe3+ are hard metals that prefer coordination by carboxylate ligands such as those in the predicted secondary site (194). Mn2+ was used as an iron analog in the first high resolution structure of bacterioferritin, an iron storage protein that oxidizes ferrous iron to a ferric form at a ferroxidase site (206). As well, crystal structures of the homologous ferritins have also been solved using Mn2+ to map the ferroxidase sites (207).   Anomalous x-ray diffraction data at the manganese absorptive edge were collected to identify secondary metal sites in the MnCu-P19 structure. At the manganese edge, a large peak (~10 ?) in the anomalous difference map is associated with a bound Mn2+ ion. The peak is observed in the secondary, acidic residue-rich binding site of P19 (Figure 4-6).  The Mn2+ ion (modeled at 25% occupancy) is situated only 7.7 ? from the nearest copper atom. Direct coordination of Mn2+ is by Glu44 (2.5 ?) and Asp92 (2.5 ?). Glu3 is more distant (Mn-O? distances of 3.2 and 3.3 ?), allowing for electrostatic interactions. Furthermore, both Asp46 and Glu3 are hydrogen bonded (2.7 and 2.4 ?, respectively) to a water molecule that is 2.8 ? from the Mn2+ ion. Due to radiation damage to the MnCu-P19 crystal during collection of the Mn anomalous data and a native dataset, copper and zinc anomalous x-ray datasets were collected on 99  sister P19 crystals that were produced at the same time and treated as for the MnCu-P19 crystal as an additional confirmation of Mn in the secondary site. No discernable peak for the Mn2+ ion above background electron density is observed in the Cu and Zn anomalous difference maps likely owing to the lower occupancy of the Mn. However, large peaks are observed corresponding to Cu present in the primary metal sites. X-ray fluorescence scans and comparison of the zinc and copper edge anomalous maps suggest partial Zn2+ occupancy (~20%) in the primary site.   4.2.6. THE PRIMARY METAL BINDING SITES ARE PREFORMED  The apo-P19 structure was obtained from a crystal of apo-protein soaked in zinc chloride. Data was collected at the zinc anomalous edge (PDB ID: 3LZN; Table 4-1), which reveals modest zinc ion incorporation into the primary site. In the resulting refined structure, zinc was modeled at 10% occupancy and thus the structure is described as apo-P19. The overall fold of apo-P19 is unchanged as compared to the metal-bound structures (r.m.s.d. of 0.6 ? over all C? atoms, as calculated by TopMatch) (208). The metal binding residues generally are in the same well-ordered conformations (B-factors less than 20 ?2).   4.2.7. P19 SHARES STRUCTURAL HOMOLOGY WITH OTHER COPPER BINDING PROTEINS  A structural similarity search of P19 against the known protein folds at the DALI server yielded Tp34 as the top hit (Z-score = 21; r.m.s.d. 1.7 ? over 152 C? atoms; PDB ID: 2O6E) (209). Unlike P19, which is a soluble periplasmic protein, Tp34 is a cation-binding lipoprotein from the causative agent of syphilis Treponema pallidum (114) and is likely a P19 homolog (35% sequence identity, E = 2x10-20). All other hits were of moderate to lower structural 100  similarity (Z-scores < 7.2) and also had immunoglobulin-like ?-sandwich folds. Of note is another copper binding protein that appears on the list; PcoC (Z-score = 5.1; r.m.s.d. 2.5 ? over 83 C? atoms; PDB ID: 1LYQ) is a periplasmic copper protein encoded by the copper resistance operon of E. coli (210).   4.2.8. RESIDUES IN P19 INVOLVED IN METAL BINDING ARE CONSERVED  The sequences of 115 identified P19 homologs derived from both Gram-negative and Gram-positive genera were aligned to reveal key conserved residues (see Figure 4-7 for a representative alignment); sequence conservation was mapped onto the surface of the P19 crystal structure (Figures 4-4c and d). Homologs of P19 are found in diverse types of bacteria, including nonpathogenic and pathogenic species. The former includes water and soil bacteria such as Rhodopseudomonas palustris and Dechloromonas aromatica, which are used in bioremediation due to their metabolic abilities in degrading biologically harmful compounds. The latter includes human pathogens such as Bordetella, Yersinia and Treponema spp.. Interestingly, P19 homologs were only identified in the bacterial kingdom. The four residues that form the primary (copper) site (His3Met) are conserved in all but three of the 115 sequences aligned. Of those 112 sequences with conserved copper site residues, significant conservation of the residues forming the acidic secondary binding site is also observed. Glu44, Asp46 and Asp92 are absolutely conserved; Glu3 is only absent from two strains of Yersinia pestis, where a large truncation of approximately 20-30 residues has occurred at the N-terminus.      101  C.je HLEADIHALKNN-PNGFPEGFWMPYLT-IAYELKNTDTGAI---KRGTLMPMVADDGP  Y.pe HLEADIHATEGN-KNGFGAGEWIPYLT-IAYTLVNTDTGDK---QEGTFMPMVASDGP  E.co HLEADIHAVEGN-KNGFGAGEWIPYLT-ISYTLVNNDTGEK---QEGTFMPMVASDGP  V.MV HLEADIHAVVGN-ENGFAGGEWIPYLN-ISYELTKSGSDWK---KAGMFMGMVASDGP  T.pa HIEADIHANEAGKDLGYGVGDFVPYLR--VVAFLQKHGSEKVQ--KVMFAPMNAGDGP       * ^                                              ? *   ^  C.je HYGANIAMEKDKKGGFGVGNYELTFYISNPEKQ---GFGRHVDEETGVG--KWFEP  Y.pe HYGANIKMM-------GVGNYKVTYHISEPSKA---GLHRHTDSETGVG--RWWKP  E.co HYGANIKMM-------GVGNYKVTYHIEPPSKA---GMHRHTDSETGVG--RWWKP  V.MV HYGANVKLDGA-------GEYNLVFHIQPPEGH---AFMRHTDKETGVG--PWWKP  T.pa HYGANVKFEEG------LGTYKVRFEIAAPSHD---EYSLHIDEQTGVSGRFWSEP    *                                       *  Figure 4-7. A multiple sequence alignment of representative P19 homologues from different genera demonstrating the conservation of copper and putative iron ligands  The species listed here were extracted from an alignment of P19 homologues identified by BLAST with an E-value cutoff of 3e10-9. Positions shaded in black are identical while light gray denotes residue similarity. Residues identified by an asterisk are those that coordinate copper; an arrow indicates the manganese ligands; an apostrophe indicates the position of Met86. C.je: C. jejuni NCTC 11168; Y.pe: Yersinia pestis biovar Microtus str. 91001; E.co: E. coli UTI89; V.MV: Magnetite containing vibrio MV-1; T.pa: T. pallidum subsp. pallidum str. Nichols. The figure was generated using ClustalW (211) and  BioEdit (170).   A methionine-rich region (formed by Met27, Met86, and the copper ligand Met88) is found adjacent to the primary binding site of P19. Met27 and Met86 are conserved in over 90% of the 115 sequences, excluding Tp34. Generally, a Pro, Gly or Ala residue is situated in between Met86 and Met88. Notably, Met86 is the only surface accessible methionine and is positioned less than 4 ? from the copper ligand Met88.  4.2.9. P19 DIMER FORMATION UPON METAL CHELATION  As the structure of C. jejuni P19 reveals a metal coordination sphere at the dimer interface, apo- and holo-P19 were subjected to gel-filtration chromatography to determine the oligomerization state of P19 in solution. The apo-P19 sample eluted with a single major peak 102  with a calculated apparent molecular mass of ~30 kDa. This is intermediary to the size of the recombinant monomer (~18 kDa) and the dimer (~36 kDa) and the increased apparent size is likely due to the elongated shape of the monomer and the presence of  large protruding loops from the ?-sandwich core (Figure 4-4a). The holo-P19 sample elution profile consisted of a major peak at ~39 kDa, consistent with the homodimer. The observation of a dimer in the apo-P19 structure may be due to the high protein concentration or the dimer being the lowest energy packing configuration of P19 under the crystallization conditions utilized. To further examine the oligomeric state of native P19, wild-type C. jejuni grown in iron-restricted media was harvested and subjected to chemical cross-linking experiments with dithiobis (succinimidyl propionate) (DSP). After DSP exposure, the cells were resuspended in sample buffer that did not contain reducing agents, separated by SDS-PAGE, and probed by western blotting (Figure 4-8a). In the control samples (no DSP), only the expected P19 monomer is detected (calculated to ~19 kDa on the blot). In the cross-linked samples, two distinct protein bands appear, corresponding to the dimer at ~39 kDa and a fainter band at ~41 kDa of unknown identity. The cross-linkage was then broken up through ?-mercaptoethanol reduction and separated by SDS-PAGE in a second dimension. Analysis of the 2D-PAGE gel and its western blot with anti-P19 antibodies confirm the dimeric band as composed of only P19 (Figures 4-8b and c).  Together, these studies suggest that in vivo, native P19 exists as a dimer.  103   Figure 4-8. Cross-linking studies of wild-type C. jejuni demonstrate detection of P19 dimer in vivo  (a) Western blot of C. jejuni lysate after separation on an SDS-PAGE gel under non-reducing conditions. Lane 1 is the control without cross-linking showing a clear band at ~19 kDa for P19. Lane 2 is the resolution of lysates following 10 min exposure of live bacterial cells to the agent DSP. (b) Two-dimensional cross-linking analysis of C. jejuni whole cells. The first dimension is separation by SDS-PAGE under non-reducing conditions. The lane was then excised, reduced to break DSP cross-links, laid on top of another acrylamide gel, and run in the second dimension. Proteins that have not changed in size run along the diagonal; the band at ~40 kDa in (a) is comprised of P19 dimers as it is now separated into monomers (as indicated by an *). (c) Western blot of the SDS-PAGE from (b) probing for P19.   4.2.10. FERRIC AND FERROUS IRON BINDING BY CU-P19  Since the publication of the P19 work (presented in the previous subsections), I have successfully produced crystal structures of iron-bound Cu-P19 (CuFe-P19) under both oxidizing and reducing conditions. The insolubility of ferrous and ferric iron at high pH values in the presence of dioxygen, in which the P19 crystals were produced, was overcome through sequential crystal soaks in mother liquor with decreasing pH, followed by the introduction of iron. Anomalous x-ray diffraction data at the copper and iron absorptive edges were collected to identify the metal sites in the CuFe-P19 structures. Data collection and refinement statistics are    104  Table 4-3. Data collection and refinement statistics for iron-soaked Cu-P19 structures Data Collectiona  FeCu-P19(red) FeCu-P19(ox)  Resolution Range (?)  50.00-1.55  33.07-1.65     (1.61-1.55)  (1.74-1.65) Cell dimensions (?)  53.91   54.22     73.23   73.72     74.93   74.85 Wavelength (?)  1.000   0.979 Unique Reflections  43367   36846 Completeness (%)  98.9 (97.1)  100.0 (99.4) Average I/?I   29.7 (4.1)  15.0 (3.1) Redundancy   6.0   8.3 Rmerge     0.050 (0.417)  0.082 (0.639)  Refinement Rwork     0.175   0.178 Rfree     0.207   0.211 No. of waters   286   260 Avg. B-value (?2)  19.9   21.9 r.m.s.d. bond lengths (?) 0.011   0.010 Ramachandran plot      Most-favorable (%)  91.7   92.1    Allowed (%)  8.3   7.5   a Values for the highest resolution shell are shown in parenthesis     105  in Table 4-3. Examination of the reduced CuFe-P19 structure, solved to 1.55 ? resolution, reveals an iron ion (likely Fe2+, modeled at 25% occupancy) situated in the secondary, acidic residue-rich binding site of one monomer. No significant iron-associated peak is observed in the other monomer. The mode of iron coordination in reduced CuFe-P19 is almost identical to Mn2+ coordination in the MnCu-P19 structure and exhibits similarly low metal occupancies (Figure 4-9a). Direct interactions of Fe2+ is through Glu44 (2.2 ?) and Asp92 (2.9 ?), as for Mn2+ in the MnCu-P19 structure, and two water molecules (3.0 and 3.1 ?). Glu3, a nearby residue in the      Figure 4-9. Iron-soaked structure under reducing and oxidizing conditions (a) The iron binding site of iron-soaked, copper-reconstituted P19 under reducing conditions (blue) superimposed on the same site in the crystal structure of MnCu-P19 (cyan). Note that Asp92, Glu44, the two water molecules (red) and the metals of both structures are drawn but may be difficult to see due to overlap. The manganese ion is colored purple while the iron ion is colored gold. (b) The same site in the crystal structure of iron-soaked, copper-reconstituted P19 under oxidizing conditions.  106  MnCu-P19 structure with a Mn-O? distance of ~3.2 ?, has one conformation that is within ligand bonding distance to the iron (Fe-O? distance of 2.0 ?, modeled at 50% occupancy).  Inspection of the CuFe-P19 crystal structure produced under oxidizing conditions and solved to 1.65 ? resolution reveals additional interactions (Figure 4-9b). The same three Fe2+ interacting residues (Glu3, Glu44 and Asp92) are found to interact with the iron cation (predicted to be Fe3+, modeled at 75 and 100% occupancy for the two monomers), but the Fe-O bond lengths range from 2.0 to 2.2 ? and all coordinating residues adopt a single conformation. Three additional non-proteinaceous electron density maxima are found at 2.1, 2.1 and 2.3 ? from the iron and are modeled as oxygen atoms to represent water molecules or hydroxide ions. The iron sites form a slightly distorted octahedral geometry. Although there are local changes in the iron-binding site of the oxidized CuFe-P19 structure, the overall structure between the oxidized and reduced states remain much the same (r.m.s.d. of 0.1 ? over all C?).  4.3. DISCUSSION The ability of pathogens to acquire iron is essential to their survival within a host. As animal and human hosts have multiple mechanisms to sequester iron from invading organisms and mobilize it for their own use, pathogens have also co-evolved uptake systems to acquire iron from these diverse iron reservoirs. C. jejuni colonizes the mucus lining of the gut, where it is exposed to the various iron sources introduced by the host?s diet. C. jejuni can also cause extensive intestinal epithelial damage, releasing the various iron sources found in blood sera, such as transferrin (212). Any free iron is also quickly bound by host proteins such as transferrin or to low molecular weight ferric chelators called siderophores, as there is constant competition between the host and its microflora for this essential nutrient. Although the genome of C. jejuni 107  does not harbor siderophore synthesis genes, the natural flora of the gastrointestinal tract and other pathogens present produce siderophores (88). C. jejuni can hijack these iron transport systems through expression of outer membrane receptors for these iron sources, including transferrin and the siderophore enterochelin (9).  The importance of P19 to C. jejuni growth during iron limitation was demonstrated by our analyses of the P19 null mutant (81176?p19), which survived poorly during iron restriction by chelation as compared to the wild-type (Figures 4-1b and 4-2). This suggests that any redundancy of function, as provided by the other members of the Fur regulon, cannot fully compensate for this single deletion, at least in vitro. Inspection of the genomic region around p19 does not provide any clear outer membrane receptor candidates for transporting extracellular ferric iron to periplasmic P19. The absence of any dedicated receptor candidates for P19 suggests that one (or more) of the co-regulated receptors may also facilitate iron uptake by p19. A recent review describes unpublished data suggesting that cFtr1-P19 transports ferric iron derived from the fungal siderophore rhodotorulic acid (94). Once iron is transported across the outer membrane by various outer membrane receptors, such as the unidentified receptor for rhodotorulic acid, it is conceivable that the iron is then picked up by P19 in the periplasm. Furthermore, the dramatic loss of growth of the P19 deletion mutant in iron-restricted MH media indicates that P19 is able to acquire iron from sources other than rhodotorulic acid. Siderophores such as rhodotorulic acid are unlikely present in the iron restricted growth experiments as MH media is produced from casein, beef infusion and starch and C. jejuni is not known to produce siderophores (89,90). Through the copper chelation growth experiment, we have demonstrated that C. jejuni requires copper for optimal growth (Figure 4-3a). Notably, the P19 deletion mutant is 108  consistently observed to achieve greater growth as compared to both the wild-type and the complement under iron-replete conditions (Figures 4-2 and 4-3). This effect may be due to reduced competition with essential copper proteins (such as cytochrome c oxidase) for copper since P19 is not required for optimal growth under iron-replete conditions (Figures 4-2a and c). Similarly, the slight reduction in maximal growth of the deletion mutant under maximal copper chelation (150 ?M) is likely due to a lack of copper availability to other copper dependent processes, such as the function of cytochrome c oxidase. With a combination of copper and iron chelation, we have demonstrated that the P19 deletion mutant is less affected by copper restriction as compared to the wild-type strain especially under iron restriction, supporting the hypothesis that P19 affects general copper availability as it competes for copper for use as a cofactor for iron transport. Another indicator that supports P19 as a copper binding protein is the presence of a M(86)XM(88) sequence motif. This motif is a hallmark of copper homeostasis proteins such as E. coli PcoC (and its homolog in Pseudomonas syringae CopC) and CusF (62,213,214). Interestingly, P19 shares significant structural similarity with PcoC and CopC although their copper binding sites do not overlap in structural alignments between P19 and the copper-bound crystal structure of CopC. In CopC, the methionine-rich region has been shown to be selective towards the reduced copper state and a second histidine-rich site is Cu2+-specific (215,216). This mode of copper coordination is conserved in PcoC (217). In the crystal structure of P19, a surface-exposed Met-rich region is also observed, with Met88 and the neighboring three histidines as the copper ligands. In addition to the conservation of Met86 and Met88 in the MXM motif, only Pro, Gly or Ala is found in the X position. The motif makes up a loop region in the crystal structure with Met86 surface exposed and proximate to Met88, facilitating copper 109  binding (Figure 4-6). Thus, the motif may function in copper selectivity of the primary metal site. Examination of the Met88-S?-Cu ligand bond lengths and copper site ligand geometries provides clues as to the oxidation state of the bound ion. In the Cu-P19 crystal structure, the geometry is best described as square planar, a preferred geometry of Cu2+. In contrast, the ?as-isolated? copper site is closer to tetrahedral geometry, favored by both Cu+ and Cu2+ oxidation states. The stronger Met88-S?-Cu interaction (distance of 2.4 ?) in the ?as-isolated? crystal structure suggests a more reduced site, as Met preferentially binds Cu+ over Cu2+. The ?as-isolated? protein may have been partially reduced due to the cytoplasmic origin of the recombinant protein. Therefore, it is conceivable that the function of the primary metal site may involve copper cycling between the two oxidation states. Competition assays show that neither Zn2+ nor Fe3+ impedes Cu2+ binding by P19; however, the primary metal site of P19 and its homologs can bind other metals in the absence of copper in vitro. Tp34 binds Zn2+ at this site with a loss of Met and the gain of Glu in the coordination sphere (Table 4-2). The modification of the metal site could be anticipated as His3 and Glu coordination is also observed in CucA and MncA, two cupins from the cyanobacterium Synechocystis PCC 6803 that can bind copper, zinc or manganese in vitro but coordinate Cu2+ and Mn2+, respectively, in vivo (218). Furthermore, the difference in coordination with zinc bound to the Tp34 structure may be explained by Hard-Soft ligand theory (194). Methionine is a poor Zn2+ ligand and is weakly associated with the metal in the zinc-Tp34 structure (distances of 3 ? or greater, Table 4-2). To the best of our knowledge, metal coordination by methionine in biological centers has only been documented for copper and heme-iron (213,219). However, Tp34 and P19 may have differing metal preferences, as previous examples of metal exchange in 110  structurally homologous proteins have been documented (220,221). Unlike most homologs of P19, Met86 (the first methionine of the characteristic copper-binding MXM motif) is substituted by AXM in Tp34 (Figure 4-7).  We have shown that P19 has the capacity to bind both copper and iron in stoichiometric amounts in solution, most likely in neighboring sites separated by 7.7 ?. This conclusion was initially based on the manganese-soaked crystal structure and has subsequently been demonstrated in two iron-soaked crystal structures. The new structures validate the use of Mn2+ as an Fe3+ mimic since they share coordination preferences. The less than full occupancy of the Mn2+ and Fe2+ in the crystal structures and the binding assay results may be due to the lower affinity of P19 for these ions with lower positive charge.   The mode of dual metal binding, as opposed to either Fe3+ or Cu2+ occupying the same site in vivo, is supported by the in vitro preference of P19 for Cu2+ over metals such as Zn2+ and Fe3+ and a connection of P19 to both copper and iron homeostasis as demonstrated in the in vivo growth experiments (Figures 4-2 and 4-3). Due to rotational symmetry, the equivalent metal sites of each monomer are on the same face of the P19 homodimer. This arrangement facilitates accessibility to both sites at potential protein-protein interfaces with proteins such as cFtr1. Glu44 appears to play a key role between each putative copper-iron pair as this residue interacts with both metal sites depending on metal occupancy and pH. The flexibility of this residue suggests that it may be poised to receive iron as part of the acidic solvent channel and possibly mediate electron transfer during the transport process. This functional model is supported by intriguing structural similarities to the S. cerevisiae MCO Fet3p, which transports iron to the Ftr1p transporter. In Fet3p, the entrance to the type I copper site where ferrous iron binds is lined with acidic residues, in particular Asp283, Glu185, and Asp409. Electron transfer from the iron 111  to the type I copper is a prerequisite for iron uptake via the Fet3p-Ftr1p complex (222-225). In both P19 and Fet3p, a highly electronegative region likely acts as a funnel to channel in iron (Figure 4-4e). This region in P19 also coincides with a surface mapping of sequence conservation indicating that this region is functionally important (Figures 4-4c and d).  The role of P19 in C. jejuni iron metabolism is possibly through a copper-dependent iron transport mechanism that draws on similarities to the Fet3p-Ftr1p system of yeast, where the connection between iron uptake and copper is well established (109,111,226). P19 is a highly inducible periplasmic protein required for full C. jejuni growth when faced with iron limitation (Figures 4-1 and 4-2). Under these conditions, the expression of other iron transport systems are also induced, leading to an increased influx of iron into the periplasmic space. It is conceivable that P19 may then pick up the iron, either directly from an outer membrane receptor or from another periplasmic protein, and then pass the iron to cFtr1 located in the inner membrane. In C. jejuni, a MCO homolog (Cj1615) contains the characteristic type I and trinuclear copper sites and also exhibits ferroxidase activity in vitro (227). However, the null mutant had no growth defects when grown on iron limited media; instead, a major role for Cj1615 in copper homeostasis has been demonstrated (227). Thus, Cj1615 is unlikely involved in P19-mediated iron metabolism. The iron transport mechanism of P19 is likely different from that of MCOs such as yeast Fet3p, as they do not share overall structural features, including the nature of the copper sites. The function of MCOs is dependent on two distinct copper sites, the mononuclear type I site for electron transfer from ferrous iron, and the trinuclear copper cluster for the reduction of oxygen to water that is not found in P19 (228). The function of the trinuclear cluster may be replaced in P19 and its homologs by electron transfer to another electron carrier. Electron transfer is feasible 112  due to the relatively flat shape of P19 that permits electron donors or acceptors to approach within 12 ? on either side of the primary metal site. These probable mechanistic differences may explain the inability to apply standard MCO assays to demonstrate P19 ferroxidase activity. In this study, we have identified flexible residues in the copper and secondary binding sites of P19 such as Glu44 and Met88 that are likely mechanistically important. Additional mutational studies both in vitro and in C. jejuni will shed more light upon the functional role of these residues in P19 and its homologs in bacteria.   113  CHAPTER 5. COPPER BINDING AND IRON TRANSPORT BY UROPATHOGENIC E. COLI FETP 5.1. INTRODUCTION Unlike in C. jejuni, which has a single Ftr1-like gene in its genome, UPEC strains such as UTI89 and F11 have two such genes: EfeU and FetM. Similar to C. jejuni cFtr1, FetM has a much longer open reading frame (as compared to yeast Ftr1 and EfeU) and is overall much more narrowly distributed than EfeU (229). In fact, EfeU is more closely related to S. cerevisiae Ftr1 than FetM in both sequence length and identity. The genes co-regulated with EfeU and FetM also differ. EfeU belongs to a low pH-induced, tripartite ferrous uptake system with a putative cupredoxin-containing protein (EfeO) and a heme-containing peroxidase (EfeB) that is involved in extracting iron from heme (52,53,230). FetM is in the same operon as FetP, a P19 homolog, and they are believed to function together as an iron uptake system.  In this collaborative study, we show that FetM and FetP are members of an iron uptake system through expression of the individual genes or both genes in a strain of E. coli that has its known iron uptake systems removed. Experiments involving the growth of these various strains under iron-limited conditions with varying pH values and the measurement of 55Fe uptake demonstrates that FetM plays a key role in iron uptake by this system whereas FetP acts as a transport-enhancing accessory protein. To examine the coordination of copper in FetP, I have solved the x-ray crystal structure of FetP with copper bound, which revealed two copper cation positions within the active site. Two active site residues have been identified that contribute to this conformational flexibility through the determination of variant crystal structures.     114  5.2. RESULTS  5.2.1. THE FETMP SYSTEM IS INVOLVED IN IRON HOMEOSTASIS  To investigate the contribution of FetMP to iron-dependent growth, the fetMP genes were inserted as a single-copy operon into the chromosome of E. coli strain ECA458 (?entC ?fecABCDE ?feoABC ?mntH ?zupT) leading to strain ECA458-fetMP. Strain ECA458 is a mutant derivative of E. coli wild-type strain K-12 substrain W3110 devoid of all known systems required for iron-uptake from defined minimal medium. As a negative control, a gentamicin resistance-cassette was introduced at the Tn7-insertion site downstream of the glmS gene of strain ECA458 to produce strain ECA458-Gm.  ECA458-Gm was cultivated in TMM with glycerol as the carbon source and grew to ~150 Klett units (data not shown), indicating that although all known iron uptake systems have been deleted from strain ECA458, at least one low affinity iron uptake system must be present.  To circumvent this issue, the metal cation chelator CDTA was added to the growth medium, which abolished ECA458-Gm growth (Figure 5-1a). Single copies of fetM and fetP were then inserted into the chromosome of E. coli strain ECA458 to investigate the individual contributions of these two genes to iron-limited growth, producing strains ECA458-fetM and ECA458-fetP, respectively. E. coli strain ECA458-fetMP, ECA458-fetM, ECA458-fetP and ECA458-Gm were cultivated in TMM with glycerol in the presence of 1 ?M of CDTA without added iron (Figure 5-1a). After 5 hrs, ECA458-fetMP exited lag phase and reached a final turbidity of 264 Klett units. The presence of fetP alone did not rescue the iron-uptake deficient strain ECA458, whereas fetM expression in ECA458 led to a retarded onset of growth that reached a lower final turbidity (204 Klett units) compared to strain ECA458-fetMP. Expression of the genes was  115   Figure 5-1. FetMP is an iron uptake system  (a) FetMP rescues E. coli strain ECA458 in the presence of the metal chelator CDTA. The genes encoding FetMP, FetM, FetP or the GmR control were inserted into the chromosome of strain ECA458 (?entC ?fecABCDE ?feoABC ?mntH ?zupT) resulting in strains ECA458-fetMP (?), ECA458-fetM (?), ECA458-fetP (?), and ECA458-Gm (?). Metal-limited cultures of these four E. coli strains were diluted 400-fold into TMM containing 1 ?M CDTA with no added iron. The cultures were cultivated at 37 ?C with shaking and the turbidity was measured as Klett units. Shown are averages of four independent experiments. (b) The expression of FetMP increases intracellular 55Fe accumulation. Strain ECA458 containing the empty pASK-IBA3 vector (?), fetMP (?), fetM (?) or fetP (?) were incubated with 200 ?g AHT per liter for 1 hr to induce expression of the cloned genes. 55FeCl3 at a final 1 ?Ci, 1 mM ascorbate and 5 ?M FeSO4 were then added to the cells and samples were removed at the indicated time points. Averages of three independent experiments with standard deviations are shown. 116  verified by RT-PCR experiments, which indicated fetP- or fetM-mRNA levels in the respective single gene strains were between 53% and 67% of those in the fetMP-containing strain. Taken together, this indicated that fetM alone is essential for the growth of strain ECA458 under iron limitation produced by CDTA whereas the presence of fetP contributes to maximal growth of strain ECA458.   The variant E. coli strains were then examined for the uptake of radiolabeled 55Fe. However, no significant differences in iron accumulation were observed between ECA458-fetMP and the negative control strain ECA458-Gm (data not shown). Therefore, the fet genes were cloned in vector pASK-IBA3 (IBA GmbH) to increase gene expression. After transforming these vectors into strain ECA458, the ability to transport 55Fe was measured. FetP alone did not enhance iron uptake by the cells, while FetM increased this process 1.7(?0.3)-fold after 10 min (Figure 5-1b). The co-expression of fetMP increased iron accumulation further with a 2.8(?0.1)-fold difference, supporting FetM as an iron importer in E. coli that is enhanced by FetP activity.  5.2.2. FETMP CONTRIBUTES TO GROWTH UNDER ACIDIC AND BASIC CONDITIONS  To examine the effects of varying pH values and metal supplementation, the growth of strain ECA458-fetMP versus the negative control strain ECA458-Gm was compared. In unsupplemented minimal media (containing trace metals), ECA458-Gm was not able to grow below pH 6.5 or at pH 9 and displayed residual growth between these pH values (Figure 5-2a). Supplementation with either Fe3+ and to some extent Mn2+ stimulated growth of strain ECA458-Gm at neutral pH values (Figures 5-2b and c). This also indicates the presence of at least one other unknown iron/manganese uptake system that is still able to import these metal cations 117   Figure 5-2. FetMP stimulates growth of ECA458-fetMP at various pH values Strains ECA458-fetMP (?) and ECA458-Gm (?) were cultivated without added iron for 24 hrs with shaking at 37 ?C in MES-Tris-double-buffered mineral salts medium and divided into medium containing (a) no additives or (b-d) 10 ?M of various metal salts. Turbidity was determined at 600 nm after 24 hrs. Averages of three independent experiments with standard deviations shown.    118  despite the entC, fecABCDE, feoABC, mntH, and zupT deletions; however, this unknown uptake system is not sufficient for growth when CDTA is added to the medium (Figure 5-1a). Zn2+ (Figure 5-2d), Cu2+ and Ni2+ (data not shown) supplementation did not improve growth of the control strain.  When the medium is not supplemented with the chelator CDTA, the presence of fetMP made no difference under neutral pH values but stimulated growth at acidic and alkaline pH values, especially when iron was added (Figure 5-2b). The effect of Mn2+ on ECA458-fetMP growth was less pronounced than that caused by iron (Figure 5-2c). The presence of fetMP also did not improve growth when other divalent transition metal cations were added (shown for Zn2+ in Figure 5-2d; Cu2+ and Ni2+ not shown) and is comparable to growth with no supplementation (Figure 5-2a). This suggests that FetMP iron transport is not inhibited by these divalent ions and the trace levels are sufficient for FetMP function. Thus, FetMP may serve primarily as an iron import system that may also transport Mn2+.  5.2.3. FETP IS A PERIPLASMIC COPPER-BINDING PROTEIN A unique feature of the FetMP system is the presence of FetP, a predicted 175 amino acid residue (aa) protein (19,224 Da) that includes a predicted leader sequence for trafficking into the periplasm. The fetP gene was cloned into vector pASK-IBA3 (producing pECD1099), which added a C-terminal Strep-tag epitope to FetP. Following the transformation of pECD1099 into E. coli strain BL21(DE3), FetP was successfully isolated from the periplasmic fraction, indicating that FetP was indeed a periplasmic protein (data not shown).  MALDI-TOF analysis of purified FetP-Strep-tag protein yielded two major peaks corresponding to 18,676 and 18,739 Da (data not shown). The 18,676 Da signal agreed with the 119  predicted size of 18,662 Da for the periplasmic protein lacking its leader sequence within an error of 0.075%. The size difference of 63 Da is close to the atomic mass of copper (63.43 Da). Therefore, it is likely that the mature Strep-tagged FetP protein was purified from the periplasm of E. coli with and without copper bound.  5.2.4. OVERALL STRUCTURE OF FETP Because the Strep-tag may interfere with the metal-binding process, the fetP gene was recloned into pET22b(+) to produce a protein with a cleavable C-terminal His-tag and an artificial periplasmic leader sequence. Using this construct, recombinant FetP was expressed and isolated from the periplasm of E. coli and the His-tag was removed. The crystal structure of ?as-isolated? FetP was then solved to 1.6 ? resolution from a crystal grown from a solutio of 20% PEG 3350, 0.2 M ammonium citrate pH 7 (data collection and refinement statistics for all FetP crystal structures are listed in Table 5-1). The four molecules of FetP in the asymmetric unit are arranged in two dimers in which each monomer is related by non-crystallographic two-fold rotational symmetry (r.m.s.d. of 0.6 ? over all C?). Each FetP dimer is a relatively flat, oval-shaped macromolecule, with dimensions of approximately 20 x 35 x 75 ? (see Figure 5-3 for the overall FetP crystal structure with copper bound). A gap in the electron density suggests proteolytic cleavage at residue 34 on chain B, which was left unmodeled. Examination of the protein sample on an SDS-PAGE gel confirmed the presence of two additional distinct bands correlating to the FetP cleavage products along with the band for full-length FetP (data not shown).  Examination of the metal content of ?as-isolated? FetP crystals by an x-ray fluorescence scan revealed the presence of copper and nickel in approximately equal amounts (Figure 5-4).  120  Table 5-1. Data collection and refinement statistics for FetP      ?As isolated?-FetP  Cu-FetP     Data Collectiona Resolution Range (?)  50.00-1.60   38.41-1.70     (1.69-1.60)   (1.79-1.70) Space group   P32    P212121 Cell dimensions (?)  a = 134.85   a = 38.41  b = 134.85   b = 51.97   c = 45.42    c = 146.65 Wavelength (?)  0.97964   0.97641 Unique Reflections  114166   31725 Completeness (%)  93.6 (88.1)   96.7 (93.5) Average I/?I   8.3 (2.4)   11.5 (2.9) Redundancy   4.1 (3.6)   4.8 (4.2) Rmerge     0.080 (0.393)   0.030 (0.143)   Refinement  Rwork     0.171    0.189 Rfree     0.208    0.226     No. of waters   338    223  Average B-factors (?2)    All atoms   21.1    14.8    Protein   20.8    15.8    Copper   -    22.5     Waters   25.0    22.1   r.m.s.d. bond lengths (?) 0.012    0.012 Ramachandran plot      Most-favorable (%)  96.4    98.9      Allowed (%)  2.7    0.7 PDB code   3NRP    3NRQ     a Values for the highest resolution shell are shown in parenthesis    121       Cu-FetP E46Q  Cu-FetP M90I     Data Collectiona Resolution Range (?)  33.65-1.40   38.34-1.53     (1.47-1.40)   (1.61-1.53) Space group   C21    P21 Cell dimensions (?)  a = 83.16   a = 49.16 b = 36.54   b = 51.61 c = 101.00   c = 66.14      ? = 106.52   ? = 93.3 Wavelength (?)  1.000    1.000 Unique Reflections  50962    50169 Completeness (%)  87.4 (84.8)   99.5 (99.9) Average I/?I   14.9 (5.0)   12.8 (3.6) Redundancy   3.9    3.6 Rmerge     0.053 (0.235)   0.047 (0.261)   Refinement  Rwork     0.186    0.176 Rfree     0.212    0.208 No. of waters   274    348 Average B-factors (?2)    All atoms   15.4    17.1    Protein   14.5    15.6    Copper   8.8    11.9    Waters   23.9    28.0 r.m.s.d. bond lengths (?) 0.010    0.011 Ramachandran plot      Most-favorable (%)  93.1    91.9    Allowed (%)  6.5    8.1  a Values for the highest resolution shell are shown in parenthesis       122   Figure 5-3. General structure of the Cu-FetP dimer with bound copper  The two monomers are shown in blue and green, respectively. Copper atoms are shown in orange. Copper ligands are represented as stick models.   Figure 5-4. Metal detection in the "as isolated" FetP crystal by fluorescence An excitation scan of the ?as isolated? FetP crystal to detect metals with an absorption edge energy equal to or lower than that of zinc. The excitation energy used and the detected metals are indicated on the plot. The inset shows a magnified view of the fluorescing metals.  123  Fluorescence signal strength was weak, consistent with low metal occupancy in the structure. Nickel contamination of the sample is likely due to purification with Ni2+ affinity chromatography. As no other exogenous transition metals were introduced during the protein purification process and in agreement with the MALDI-TOF result of purified Strep-tagged FetP, copper was likely acquired from the media during FetP expression in E. coli. Based on these results, FetP was treated with EDTA and the crystal structure of copper-reconstituted FetP (Cu-FetP) was solved to 1.7 ? resolution from a crystal produced under 0.1 M Bis-Tris pH 6.5, 25% pentaerythritol ethoxylate (15/4 EO/OH), and 25 mM ammonium sulfate. Structural comparisons between the ?as-isolated? and copper-reconstituted FetP crystal structures reveal minimal overall fold differences (r.m.s.d. of 0.4?0.6 ? over all C? as calculated by TopMatch (208)). Unless specified, any subsequent descriptions of the structural features of FetP will be based on the copper-reconstituted structure. A single dimer was observed in the Cu-FetP crystal structure, with each monomer comprised primarily of two 4-stranded anti-parallel ?-sheets stacked upon each other in an immunoglobulin-like fold (Figure 5-3). Three single-turn ?-helices were also observed on the periphery of the FetP dimer (one from one monomer and two on the other). The ?-helix that is not found in the symmetry-related monomer is part of a loop that has the highest variability between the two monomers in the crystal structure and also the greatest crystallographic B-factor values as compared to the rest of the protein, indicating high flexibility of this region. The dimer interface was analyzed with the program PISA (204), which reveals an interface area of ~1600 ?2 (or ~20% of the total solvent accessible surface area of each monomer) and involves 45 residues from each monomer. Tight packing of the dimer interface is supported by lower than average B-factors in this region (data not shown). 124  5.2.5. THE FETP COPPER BINDING SITE Anomalous dispersion diffraction data collected at the copper anomalous edge indicate the locations of the copper ions in the structure. Two copper ions (CuA and CuB) were identified from ~30 ? symmetry-related peaks in the anomalous dispersion maps. The copper ions are buried, separated by ~29 ? and are related by the two-fold symmetry of the FetP molecule. Each copper site is located at the dimer interface and is formed by three histidines (His44, His97 and His127*), Met90 and Glu46 (Figure 5-5). The sites are located at the surface of the core ?-sandwich with two residues (His44 and Glu46) originating directly from a ?-strand and His127* that is derived from an extended loop from the opposing monomer. In each site, the copper ion and Met90 were observed in two positions, denoted ?1? and ?2? (Figure 5-6). Each conformer is refined at 50% occupancy with similar B-factors. Analysis of the two copper binding modes indicates that in both conformers the copper is four-coordinate, involving a sulfur atom of Met90 and the three imidazole ring nitrogens of His44, His97 and His127* (Table 5-2). In conformer ?2?, the copper sites adopt a tetrahedral geometry; however, in conformer ?1? of CuA, the oxygen atom of Glu46 may form a weak fifth ligand in a degenerated octahedral complex with a Cu-O? distance of 2.7 ?, drawing the Cu below the plane defined by the three His-N? atoms (Figure 5-6a). No suitable residue is present to serve as a potential sixth ligand. In contrast, the Cu-O? distance is 3.0 ? in conformer ?1? of CuB (Figure 5-6c, Table 5-2), leaving the Cu atom in the plane defined by the three His-N? atoms. Met90 and Glu46 are located on opposing sides of a plane formed by the three His residues, with Met90 in multiple conformations to maintain a Met-S?-Cu association during the Cu movement.    125   Figure 5-5. The FetP copper binding site reveals multiple copper positions A single copper atom is situated in the binding site but it exhibits two density maxima (depicted as the two orange spheres). The nearby residues are shown as a stick model and are colored according to the originating monomer as shown in Figure 5-3. The electron density in blue is a 2Fo-Fc map in blue contoured at 1 ? and the copper absorptive edge anomalous map in teal is contoured at 5 ?.   126   Figure 5-6. The copper ligands in each conformation  Copper sites A (Panels (a) and (b)) and B (Panels (c) and (d)) are shown as ball and stick representations. The two conformations of the copper cation and its ligands at each site are shown side by side to highlight the changes in ligand identity and geometry in each conformation (with the other copper position shadowed). The copper cation is shown as an orange sphere and is always bound to His44, Met90, His97 from the same monomer and His127* from the other monomer. Similar to the copper cation, Met90 is also observed in two conformations. In CuA1 (a), Cu is also weakly interacting with Glu46 (CuA1-O? distance 2.7 ?) in a degenerated octahedral arrangement with five copper ligands. The three His-N? atoms all lie on the same plane with the Cu-N? bonds forming nearly right angels (97? to 105?); the Glu46-O?-Cu-Met90-S? atoms almost form a straight line. In CuA2, the copper atom is in the center of a tetrahedral complex (b), similar to that of CuB2 (d).  127  Table 5-2. Ligand bond lengths in the primary site of Cu-FetP   Distance to copper site (?)a Residue CuA1 CuA2 CuB1 CuB2      His44 2.1 2.1 1.9 2.1 His97 1.8 2.3 1.8 2.4 His127b 2.3 2.2 2.1 2.2 Met90-1c 2.7 - 2.7 - Met90-2c - 2.3 - 2.2 Glu46 2.7 - - -      a Only ligand distances up to 3 ? are listed b His127 originates from the symmetry-related monomer  c Two conformations of Met90 are observed.      128  Adjacent to each of the Cu sites, three residues from the same monomer (Met29, Met34 and Met88) and His125* from the other monomer form another putative metal binding site, which we have termed ?CuC? (Figure 5-7). Met34 is situated on a flexible lid-like loop at the molecular surface of FetP that is observed in differing conformations in each monomer. In the more open loop conformation (as seen next to CuA), the distance between Met34-S? and His125-N? is 7.6 ? whereas in the more closed conformation, this distance is reduced to 3.5 ? (as seen next to CuB). A metal ion placed in ?CuC? within coordinating distance of these four putative ligands in the closed conformation would be approximately 7 ? from the nearest copper ion.      Figure 5-7. A putative third metal binding site ?CuC? exists adjacent to CuB  A putative Met-rich third metal binding site (?CuC?) is indicated by the red tetrahedron and is formed by Met29, Met34, and Met88 from the same monomer and His125* from the other monomer. Adjacent to CuA, the four residues represents a more open conformation.    129  5.2.6. THE COPPER BINDING SITE IS PREFORMED  Examination of the 2Fo-Fc maps of the ?as-isolated? FetP structure reveals no molecules bound in the copper binding site of chain A and either a full occupancy water or a low occupancy copper ion in chains B, C, and D (data not shown). As all of the interactions are within hydrogen bonding distance and are longer than the typical metal-ligand bonding distance (2.5 ? or greater distance except for His44 in chain B), a water molecule has been modeled into these three sites. The longer hydrogen bonding distances between the copper ligands and the modeled waters in these sites suggest that when little to no metal is bound, the loops in this region experience some flexibility but pre-form a binding site ready for incoming copper. The presence of copper, on the other hand, tightens the protein fold in the copper-binding region, as demonstrated in the Cu-FetP structure bonding lengths (Table 5-2).   5.2.7. GLU46 AND MET90 CONTRIBUTES TO THE DUAL COPPER POSITIONS To determine the role of Glu46 and Met90 in defining the position of copper in the CuA and CuB binding sites of FetP, two copper site mutations were generated (E46Q and M90I). Recombinant FetP E46Q and FetP M90I were expressed in the periplasm of E. coli, isolated and reconstituted with copper, forming Cu-FetP E46Q and Cu-FetP M90I. Crystal structures of the copper-reconstituted variants Cu-FetP E46Q and M90I were solved to 1.4 ? and 1.5 ? resolution, respectively. See Table 5-1 for structure refinement statistics. Similar to the crystal structure of wild-type Cu-FetP, two molecules of each variant are found in the asymmetric unit and are arranged in a dimer with non-crystallographic two-fold rotational symmetry. The variant structures exhibit minimal overall fold differences as compared to the wild-type Cu-FetP structure (r.m.s.d. of ~0.6 ? for both variants). In both variant crystal structures, the greatest 130  deviation in the C? backbone from wild-type is observed in the termini and the flexible loop region containing residue 34 on chain B, which is part of the ?CuC? site, and is the same location of proteolytic cleavage in the ?as isolated? structure. Anomalous dispersion diffraction data collected at the copper anomalous edge demonstrates copper binding in the variant active sites (Figure 5-8).    Figure 5-8. The copper binding site of FetP mutants E46Q and M90I Mutagenesis of FetP copper active site residues (a) Glu46 or (b) Met90 diminishes the dual copper position effect observed in the wild-type Cu-FetP crystal structure. The copper cation and its ligands are colored according to Figure 5-5. The copper absorptive edge anomalous map in teal is contoured at 5 ?.  131  In the crystal structure of Cu-FetP E46Q, mutagenesis of Glu46 has a noticeable effect on the position of the copper. In CuB, the distance between the two copper electron density maxima is reduced from 1.3 ? to 0.7 ? and only a single Met90 conformation is observed (not shown). This change results in a ~0.2 ? longer (slightly weaker) His44-Cu interaction and a ~0.2 ? reduction in Met90-Cu bond lengths, indicating a slightly stronger interaction. At the CuA site, the multiple copper positions are completely abolished and the anomalous signal at this copper site can be accounted by Met90-S? and a single copper position (Figure 5-8a). At both sites, the Gln46-Cu interaction is further weakened as the distance is increased to >3.8 ?. The crystal structure of the other active site mutant, Cu-FetP M90I, reveals a single copper electron density maximum in both copper sites (Figure 5-8b). Another significant change that is not observed in any of the other FetP crystal structures is the dual Glu46 conformations observed in the Cu-FetP M90I variant. One of the two conformations brings a Glu46-O? within bonding distance with the copper ion (~2.2 ?).   5.3. DISCUSSION  The ability of an organism to acquire iron from the host environment is crucial to its ability to colonize and persist within the host, whether as a commensal or a pathogen. E. coli strain F11 is an uropathogenic species that is able to colonize both the gastrointestinal tract and also the bladder. In the intestinal environment, colonizers such as E. coli can utilize iron from food ingested by the host. Uropathogenic E. coli can disseminate from the intestinal tract to the generally sterile urinary tract and bladder, where it would be regularly exposed to urine, a liquid with fluctuating oxygen tension, pH value, and concentration of organic matter (231). The urine of a healthy individual contains iron, although the exact species is unknown and will be affected 132  by the changing environmental conditions. At moderately acidic or alkaline conditions, ferric iron can exist in the more soluble Fe(III)(OH)2+ or Fe(III)(OH)4- complexes, respectively, instead of the insoluble Fe(III)(OH)3 (232); ferrous iron is much more stable under acidic conditions and hence is a possible form of encountered iron.  Although the forms of iron transported by the FetMP system are still unknown, this system does rescue the growth of an E. coli strain nearly devoid of other iron uptake systems under varying pH (Figures 5-1a and 5-2). Through the expression of FetM and FetP, we have demonstrated that this system increases the intracellular supply of iron, with the integral membrane protein FetM playing a central role to this task and the periplasmic protein FetP enhancing the function of FetM (Figure 5-1b).  The crystal structure of Cu-FetP reveals copper bound to residues His44, Met90, His97 and His127* from the other respective subunit. These four residues are conserved in the 100 closest homologs of FetP in a BLAST search, including the three homologs ChpA from the marine magnetotactic Vibrio strain MV-1 (115), P19 from the pathogenic Campylobacter jejuni (233) and Tp34 from Treponema pallidum (114). Recombinant Tp34, soaked in 10 mM Zn2+, displayed a Zn2+ atom in a degenerated octahedral geometry that is coordinated by three His, one Glu and one Met residue, corresponding to the copper-binding residues of FetP (114). The authors discuss that this geometry may represent a novel zinc-binding site or that zinc could have bound adventitiously to a site that would bind a different metal in vivo. Purification of FetP produced in the periplasm of E. coli revealed the presence of bound copper and no zinc, even though the binding sites of purified FetP were not saturated due to overexpression. Similarly, ChpA from the marine Vibrio contained copper after purification of the periplasmic protein fraction from its native host (115). P19 from C. jejuni, in which the equivalent active site 133  residues also bind copper in a manner similar to FetP, was shown to preferentially bind copper over zinc (Chapter 4). Although, binding of different metals to sites with very similar ligand identities and geometries has been described before (218,234), the lines of evidence herein support copper, rather than zinc, as the physiological cofactor of FetP and its orthologs.  Interestingly, in contrast to the other previously characterized FetP homologs, the copper ion in each binding site in the crystal structure of FetP exhibited two residence probability maxima (?1? and ?2?), with some additional  conformational differences between the two copper-binding sites within the FetP dimer (CuA and CuB). With the exception of CuA1, copper was found to be in a more or less tetrahedral conformation with one Met and three His residues as ligands. In CuA1, copper was in a degenerated octahedral complex with an additional Glu46 residue at a distance of 2.7 ? from the copper cation. The observed positional flexibility of the copper in the active site suggests that it may be functionally important during iron transport, potentially to facilitate electron transfer. It is possible that multiple copper positions were not observed in prior structures as the result of the crystallization pH of its homologs. The crystals of Cu-P19 (Chapter 4) and Zn-Tp34 (114) were produced at pH 9-10 whereas Cu-FetP was crystallized at pH 6.5 and may have effects on the ligand protonation state. As copper displacement is perpendicular to the plane formed by the three His ligands and is towards either Met90 or Glu46, which are located on the opposing sides of the plane, I hypothesized that these residues are required to stabilize the two copper conformers. Indeed, mutagenesis of either Met90 or Glu46 (Figure 5-8) resulted in partial or complete abolishment of the multiple copper positions observed in the wild-type Cu-FetP crystal structure (Figure 5-5). Although Cu-FetP E46Q and Cu-FetP M90I were crystallized under conditions differing from those of wild-type Cu-FetP, the pH remained unchanged (pH 6.5) and is therefore not the cause 134  of conformational differences. New crystallization conditions were found because utilization of wild-type Cu-FetP conditions for the variants produced low quality diffraction data.  An additional copper-binding site (?CuC?) is proposed to exist at the molecular surface of FetP, adjacent to the primary copper site. The putative ligands are Met29, Met34, Met88 and His125. Met29 and Met88 are conserved in the 100 closest relatives of FetP in a BLAST search, including in P19, ChpA and Tp34 whereas positions 34 and 35 are generally conserved. Approximately 80% of the sequences contain an MM, XM or MX motif (with X being any residue) and another 10% contains either an ID or VD motif and represents all of the Campylobacter spp. and Rhodospirillum rubrum. If ?CuC? is mechanistically important for iron transport by FetP, then a Met to Asp substitution may represent functional adaptation for organisms such as C. jejuni.  Met34 is situated on a flexible loop that may act as a lid to the ?CuC? binding site. The tetrahedral arrangement of the ligands in the ?CuC? site adjacent to CuB would represent a ?closed? conformation resembling a holo-CuC site; in contrast, Met34 in the equivalent site adjacent to CuA is observed in a more ?open? conformation and is positioned to possibly facilitate exogenous copper interactions. An alternative role for the ?CuC? site may be copper loading into the primary site or to allow copper shuttling between the two copper sites (e.g. ?CuC? ? CuB) during an iron transport reduction/oxidation event. A Met-rich site such as ?CuC? site would have a higher preference for Cu+ binding than CuA/B, similar to the Met-rich and His-rich copper sites of the P. aeruginosa copper chaperone CopC (215,216). This shuttling could facility the removal or addition of electrons as the ?CuC? site is located on the surface of the FetP dimer. Another model would involve binding of an additional copper for the same 135  purpose as shuttling of a single copper cation. However, the binding of more than two coppers have yet to be observed in any crystal structures.  Due to the chemical similarities between iron and manganese, it is possible that the FetMP system and by extension, the C. jejuni P19-cFtr1 system, can transport both metals in vivo. The transport of both iron and manganese has been observed in other characterized transporter families, including the ABC-type SitABCD (47) and MntH of the Nramp family (132). Additionally, a crystal structure of P19 with Mn2+ bound in the iron site has indeed been solved (Chapter 4). This site is composed of Glu3, Glu46 and Asp94 in FetP and is conserved in the closest homologs, including ChpA and Tp34.  In S. cerevisiae and other fungi, Fe3+ is reduced to Fe2+ by Fre1p and Fre2p to solubilize iron from insoluble ferric iron hydroxide complexes (109). Prior to high affinity Fe3+ uptake by Ftr1p, ferrous iron is oxidized again by the multicopper oxidase Fet3p, a glycosylated ferro-O-2-oxidoreductase of the yeast plasma membrane and a homolog of the human plasma protein ceruloplasmin (235). Fet3p contains a mononuclear copper cluster that reacts with Fe2+ and a trinuclear copper cluster that transfers the metal-derived electrons to molecular oxygen (224,236). FetP does not contain a copper cluster similar to that of Fet3p, indicating that electron transfer activity would require a different mechanism.  EfeU, the other Ftr1-like protein in E. coli, is part of a tricistronic operon under the control of the Fur repressor in response to the iron status (110) and is also repressed at high pH by CpxAR (53). When the efe operon is expressed in a strain lacking all known iron-uptake systems, the cells gained a major growth advantage (53,110). EfeB is a periplasmic hemoprotein needed to promote iron extraction from heme (53,230) while periplasmic EfeO is a novel cupredoxin-like protein hypothesized to deliver iron to EfeU (52). In contrast to Efe, however, 136  there is no evidence of heme-iron transport by the Fet system. Therefore, the two systems will likely complement each other during host colonization to transport growth sufficient iron from the wide variety of iron sources.    137  CHAPTER 6. OVERVIEW AND FUTURE DIRECTIONS In order for iron to be used in proteins within the cytoplasm of Gram-negative bacteria, three barriers must be passed: the OM, the IM, and the intervening periplasmic space, each requiring specialized classes of proteins to transport the substrate across it. The work detailed in this thesis examines the role of two types of iron-regulated transport systems from two Gram-negative pathogens.  Chapter 3 focuses on the characterization of ChaN, a membrane-associated lipoprotein in C. jejuni. Biochemical characterization of ChaN reveals heme coordination through the formation of a protein homodimer around cofacial heme molecules. This dimerization was shown to be heme-dependent and is likely conserved in the homologs of ChaN that retain the heme-interacting Tyr148, His176 and Lys197 residues. The predicted lipidation of ChaN was supported by cellular fractionation studies, which demonstrated distribution of ChaN generally to the OM. Adjacent to chaN in the C. jejuni genome is chaR, which encodes a predicted prototypical plug and ?-barrel-containing receptor in the OM. In vivo growth studies using a C. jejuni chaNR double deletion mutant revealed a delayed growth effect on free iron and transferrin as sole iron sources, but the mutant is able to achieve the same overall maximal growth as the wild-type strain.  Chapters 4 and 5 examined two homologous systems from C. jejuni (cFtr1-P19) and uropathogenic E. coli (FetMP), each containing a predicted IM transporter and a soluble periplasmic component. C. jejuni P19 was shown to be a periplasmic protein that can bind both copper and iron. Using x-ray crystallography, the copper and iron binding sites were identified. The two sites are located next to each other (with a copper-iron distance of less than 8 ?) and are formed by highly conserved residues. A deletion of p19 in C. jejuni resulted in a growth defect 138  under iron-limited conditions and also displayed an effect on copper homeostasis, especially when iron is also limited. The overall crystal structure of FetP is highly similar to P19, including the conservation of metal binding site residues. However, the copper-bound structure revealed a previously-unseen dual position of the copper cation in each binding site. Mutations in either of the copper active site residues, Met90 and Glu46, resulted in the unification of the dual copper positions in the variant crystal structures. Expression of FetM and FetP in a strain of E. coli devoid of known iron uptake systems demonstrated an iron uptake role by FetM that is assisted by the presence of FetP. This system was shown to produce an iron-dependent growth phenotype over a wide pH range that likely reflects the conditions within the host environment.   6.1. A REGULATORY ROLE FOR CHAN? Proteins of the ChaN family represent an interesting evolutionary adaptation since most TonB-dependent transport systems do not have an additional genetically-associated lipoprotein component. However, in the cases where a lipoprotein is present, it appears to contribute to the function of the OM receptor and growth under iron restriction, such as the case within the host environment. For example, a deletion in the transferrin transporter lipoprotein TbpB, which is unrelated to ChaN in both sequence and structure, has been shown to reduce transferrin-iron uptake and is important for infection (28,29). Similarly, a deletion of the P. aeruginosa heme uptake system lipoprotein PhuW, a ChaN homolog, led to a visible growth defect on heme as a sole iron source (108). Although an unequivocal role for ChaN in iron uptake is still unknown, I hypothesize that it likely interacts with and enhances or regulates the transport capabilities of ChaR.  139  In light of a subsequent study published after the ChaN biochemical work that showed ChaR involvement in transferrin and lactoferrin iron uptake (106), ChaR may be not involved in the direct transport of heme. Heme binding by ChaN could instead modulate the transport behavior of OM receptors such as ChaR, thereby affecting the iron source preference or possibly the degree of iron versus heme uptake by C. jejuni. This regulatory phenomenon has been observed in studies with Staphylococcus aureus, in which preferential iron sources are taken up in greater amounts when multiple iron sources are simultaneously presented (237). In support of ChaN involvement in transferrin-iron uptake, preliminary studies do suggest that ChaN can also bind free iron (data not shown); however, whether iron and heme binding is functionally related or biologically relevant has not yet been proven.  6.2. BUILDING A P19 FUNCTIONAL MODEL FROM WORK INVOLVING TWO PATHOGENS Through the metal binding and uptake experiments, solved crystal structures, and growth studies, a model of how P19 and FetP participates in iron uptake through their respective Ftr1-like IM permeases can be developed. Specifically, the mechanism of iron transport by P19 and FetP is predicted to involve either iron oxidation or reduction activity that is dependent on the copper situated in the primary active site identified in the crystal structures. With iron bound to the surface-located, acidic residue-rich site of the P19 or FetP, a short conduit (~7.7 ?) is provided for electron transfer between the copper and iron cations (Figure 6-1). An interaction between the two metals is also supported by the two observed copper positions, which occur on the same general axis as the one drawn through the two metals of the CuFe-P19 crystal structure. By aligning the Cu-FetP structure with the CuFe-P19 structure, one copper conformer in the 140    Figure 6-1. Alignment of the Cu-FetP and CuFe-P19 structure active sites  The metal binding ligands in the copper and iron binding sites of the crystal structures of Cu-FetP (blue) and CuFe-P19 under oxidizing conditions (yellow) are superimposed. The copper atoms are colored corresponding to the bound protein. Note that two copper conformers are observed in the Cu-FetP structure while a single conformer is observed in the CuFe-P19 structure. The iron atom of the CuFe-P19 structure is colored brown.   141  FetP structure aligns well with the copper in the P19 structure while the other copper conformer is positioned closer to the iron site when no iron is present.  This suggests that iron binding may influence the position of the copper, but the functional implications of these conformers are not currently known. Once the intermetal electron transfer occurs, the iron must be released for transport across the inner membrane, which likely involves a direct protein-protein interaction with cFtr1/FetM. However, if ferrous iron is produced by P19/FetP, in which iron binding studies suggest a weaker binding affinity, the iron may also be released into the periplasmic space and can then be recognized by a ferrous iron transporter, such as the FeoB system. This iron would likely remain reduced and soluble due to the low-oxygen content in the environments in which these two pathogens colonize.  The electron donor or acceptor is predicted to interact on the opposite side of the copper binding site, where a Met-rich secondary copper site (?CuC? in FetP) is proposed. This conserved site may be able to bind copper, which would interact with the donor/acceptor and facilitate electron delivery to or from the primary copper, depending on whether iron reduction or oxidation occurs in the biological system.   6.3. FUTURE DIRECTIONS 6.3.1. CHAN As comparisons between the chaR deletion growth studies by Miller et al. (106) and the chaNR work presented in this thesis suggests intriguing effects from ChaN expression in the absence of the receptor ChaR in C. jejuni, additional mutational studies should be performed. To that effect, single deletion strains of C. jejuni in chaN and chaR have been produced and are 142  awaiting further growth studies. Preliminary examination of growth of 11168?chaNR on heme as a sole iron source suggests similar maximal growth of both the mutant and wild-type strains (data not shown), as was the case for growth on free iron and transferrin as sole iron sources (Figure 3-7). The initial lag in growth of 11168?chaNR on free iron and transferrin may be the result of a delayed compensation by the induction of other iron uptake systems, which eventually allowed for similar maximal growth of the deletion strain as the wild-type strain. This hypothesis suggests that demonstrating a more exaggerated growth defect would require additional deletions in the other iron uptake systems in conjunction with chaNR. For example, the presence of the Chu heme uptake system in C. jejuni may make it difficult to reveal a ChaNR heme-dependent phenotype.  Further characterization of ChaN, itself, is also of much interest. To determine if ChaN plays a regulatory role, studies could be undertaken to elucidate how it modulates cellular activity. If ChaN directly interacts with ChaR and other receptors, the unidentified partners may be revealed through further interaction studies, such as co-immunoprecipitation techniques in combination with cross-linking agents. If ChaN affects the expression of other proteins, especially other transporters, it would be of interest to observe ChaN regulation through C. jejuni microarray analyses using the wild-type and chaN deletion strains. An interaction between ChaN and ChaR is predicted due to the juxtaposition of the two genes and an apparent co-regulation by the same promoter, but this has also yet to be demonstrated. Alternatively, applying a technique such as fluorescence resonance energy transfer (FRET) would allow the visualization of fluorescently-labeled ChaN and ChaR interactions within a cell (E. coli or possibly C. jejuni), as has been previously utilized in organisms such as E. coli (238) and Bacillus subtilis (239). As 143  this technique can be applied to live cells, dynamic changes in protein-protein interaction can be observed upon the addition of substrates, such as heme or iron.  The specific residues involved in the putative interaction between ChaN and ChaR can be identified through crystallography. Co-crystallization experiments have been similarly used to demonstrate such detailed interactions between the ABC transporter components of the E. coli vitamin B12transport system (25), the Serratia marcescens hemophore HasA and its receptor HasR (240), and the E. coli C-terminal periplasmic domain of TonB with the ferrichrome receptor FhuA (241).  As ChaR has been implicated in transferrin-iron uptake (106), the iron binding properties of ChaN should be further examined. Firstly, it will be interesting to determine whether heme and iron can simultaneously bind or if binding is mutually exclusive using spectroscopic techniques to measure the degree of substrate binding. Although previous attempts at producing an iron co-crystallized ChaN structure have been unsuccessful, the experience gained from the P19 iron-soaking structural studies can be applied to loading ChaN with iron. Another approach can involve the use of stable analogs that mimic iron binding to ChaN for the purpose of structural determination. As preliminary data suggests that ChaN has an iron-binding capability, the question arises as to why this occurs. Is an iron or heme moiety an important co-factor in the transport of the other or could there be substrate competition?  6.3.2. P19 AND FETP Interestingly, the genetic deletion approach taken for C. jejuni p19 revealed a severe growth defect when the mutant cells were grown under iron limitation. In contrast, the expression of FetP and FetM in an E. coli strain devoid of all known iron uptake systems 144  suggests that FetP is an accessory protein that enhances, but is not critical for FetM iron transport function. While it appears that the role of P19 in iron-limited growth cannot be compensated by the level of other C. jejuni proteins expressed under those conditions, it will be of interest to determine whether similar or additive effects also occur if cftr1 (the presumed ortholog of fetM) is deleted individually or in combination with p19. Additionally, the large periplasmic domain found in cFtr1 and FetM that is not present in EfeU or Ftr1p remains uncharacterized and is likely more amenable to experimentation than a full length construct. This domain could be examined for possible metal binding properties and interactions with P19 or FetP. The predicted role of P19 and FetP in oxidizing or reducing iron will require further investigation. Determination of the reduction potential of the copper site in these two enzymes will assist in understanding their mechanism, which may be accomplished using cyclic voltammetry. To examine the oxidation state of the metals bound to these proteins, electron paramagnetic resonance spectroscopy can be used; however, this will likely require preparation of the samples under anaerobic conditions due to the oxidative effects of atmospheric dioxygen. Although an assay for these two proteins has yet to be found, as conventional approaches do not demonstrate typical ferroxidase or ferrireductase activity, a greater understanding of their biochemical properties may lead to the discovery of such an assay in the future. From the solved crystal structures of P19 and FetP, residues involved in metal binding have been identified. Whether modifications of these residues affects metal affinity, as can be observed through x-ray crystallography or spectroscopic quantification techniques, is a potential area of future work. Mutations of interest include modifications of the metal ligands to prevent binding or to convert metal binding preferences, such as the generation of a high-affinity zinc binding site within P19 or FetP. As recent genetic tools have allowed for complementation of C. 145  jejuni deletion mutants, these vectors can instead be adapted to introduce active site mutants of p19 into the cells. Whether the expression of these mutants would elicit a metal-dependent phenotype could complement the biochemical work.   146  REFERENCES 1. Cowan, J. A. (1997) Inorganic Biochemistry: An Introduction, 2nd ed., Wiley-VCH, Inc., New York 2. Williams, R. J., and Frausto Da Silva, J. J. (2003) J Theor Biol 220, 323-343 3. Braun, V., and Killmann, H. (1999) Trends Biochem. Sci. 24, 104-109 4. Palyada, K., Threadgill, D., and Stintzi, A. (2004) J. Bacteriol. 186, 4714-4729 5. Ernst, J. F., Bennett, R. L., and Rothfield, L. I. (1978) J Bacteriol 135, 928-934 6. Carpenter, B. M., Whitmire, J. M., and Merrell, D. S. (2009) Infect Immun 77, 2590-2601 7. Horsburgh, M. J., Ingham, E., and Foster, S. J. (2001) J Bacteriol 183, 468-475 8. Bury-Mone, S., Thiberge, J. M., Contreras, M., Maitournam, A., Labigne, A., and De Reuse, H. (2004) Mol Microbiol 53, 623-638 9. van Vliet, A. H. M., Ketley, J. M., Park, S. F., and Penn, C. W. (2002) FEMS Microbiol. Rev. 26, 173-186 10. Calderwood, S. B., and Mekalanos, J. J. (1987) J Bacteriol 169, 4759-4764 11. Frechon, D., and Le Cam, E. (1994) Biochem Biophys Res Commun 201, 346-355 12. Masse, E., and Gottesman, S. (2002) Proc Natl Acad Sci U S A 99, 4620-4625 13. Abdallah, F. B., and El Hage Chahine, J. M. (2000) J Mol Biol 303, 255-266 14. Ward, P. P., Uribe-Luna, S., and Conneely, O. M. (2002) Biochem Cell Biol 80, 95-102 15. Le Brun, N. E., Crow, A., Murphy, M. E., Mauk, A. G., and Moore, G. R. Biochim Biophys Acta 1800, 732-744 16. Ratledge, C., and Dover, L. G. (2000) Annu. Rev. Microbiol. 54, 881-941 17. Andrews, S. C., Robinson, A. K., and Rodriguez-Quinones, F. (2003) FEMS Microbiol. Rev. 27, 215-237 18. Wiener, M. C. (2005) Curr. Opin. Struct. Biol. 15, 394-400 19. Krewulak, K. D., and Vogel, H. J. (2008) Biochim Biophys Acta 1778, 1781-1804 20. Cobessi, D., Celia, H., Folschweiller, N., Schalk, I. J., Abdallah, M. A., and Pattus, F. (2005) J. Mol. Biol. 347, 121-134 21. Cobessi, D., Celia, H., and Pattus, F. (2005) J. Mol. Biol. 352, 893-904 22. Scott, D. C., Cao, Z., Qi, Z., Bauler, M., Igo, J. D., Newton, S. M., and Klebba, P. E. (2001) J Biol Chem 276, 13025-13033 23. Postle, K., and Larsen, R. A. (2007) Biometals 20, 453-465 24. Holland, I. B. (2003) ABC proteins : from bacteria to man, 1st ed., Academic Press, Amsterdam ; Boston 25. Hvorup, R. N., Goetz, B. A., Niederer, M., Hollenstein, K., Perozo, E., and Locher, K. P. (2007) Science 317, 1387-1390 26. Gray-Owen, S. D., and Schryvers, A. B. (1996) Trends Microbiol 4, 185-191 27. Gray-Owen, S. D., Loosmore, S., and Schryvers, A. B. (1995) Infect Immun 63, 1201-1210 28. Anderson, J. E., Sparling, P. F., and Cornelissen, C. N. (1994) J Bacteriol 176, 3162-3170 29. Baltes, N., Hennig-Pauka, I., and Gerlach, G. F. (2002) FEMS Microbiol Lett 209, 283-287 30. Gray-Owen, S. D., and Schryvers, A. B. (1995) Infect Immun 63, 3809-3815 31. Retzer, M. D., Yu, R., Zhang, Y., Gonzalez, G. C., and Schryvers, A. B. (1998) Microb Pathog 25, 175-180 147  32. Moraes, T. F., Yu, R. H., Strynadka, N. C., and Schryvers, A. B. (2009) Mol Cell 35, 523-533 33. Bruns, C. M., Nowalk, A. J., Arvai, A. S., McTigue, M. A., Vaughan, K. G., Mietzner, T. A., and McRee, D. E. (1997) Nat Struct Biol 4, 919-924 34. Bruns, C. M., Anderson, D. S., Vaughan, K. G., Williams, P. A., Nowalk, A. J., McRee, D. E., and Mietzner, T. A. (2001) Biochemistry 40, 15631-15637 35. Stojiljkovic, I., and Perkins-Balding, D. (2002) DNA Cell Biol. 21, 281-295 36. Genco, C. A., and Dixon, D. W. (2001) Mol. Microbiol. 39, 1-11 37. Tong, Y., and Guo, M. (2009) Arch Biochem Biophys 481, 1-15 38. Ghigo, J. M., Letoffe, S., and Wandersman, C. (1997) J Bacteriol 179, 3572-3579 39. Cope, L. D., Yogev, R., Muller-Eberhard, U., and Hansen, E. J. (1995) J Bacteriol 177, 2644-2653 40. Ho, W. W., Li, H., Eakanunkul, S., Tong, Y., Wilks, A., Guo, M., and Poulos, T. L. (2007) J Biol Chem 282, 35796-35802 41. Wilks, A. (2001) Arch Biochem Biophys 387, 137-142 42. Lansky, I. B., Lukat-Rodgers, G. S., Block, D., Rodgers, K. R., Ratliff, M., and Wilks, A. (2006) J Biol Chem 281, 13652-13662 43. Ratliff, M., Zhu, W., Deshmukh, R., Wilks, A., and Stojiljkovic, I. (2001) J. Bacteriol. 183, 6394-6403 44. Hantke, K. (2003) Trends Microbiol 11, 192-195 45. Marlovits, T. C., Haase, W., Herrmann, C., Aller, S. G., and Unger, V. M. (2002) Proc Natl Acad Sci U S A 99, 16243-16248 46. Cartron, M. L., Maddocks, S., Gillingham, P., Craven, C. J., and Andrews, S. C. (2006) Biometals 19, 143-157 47. Kehres, D. G., Janakiraman, A., Slauch, J. M., and Maguire, M. E. (2002) J Bacteriol 184, 3159-3166 48. Janakiraman, A., and Slauch, J. M. (2000) Mol Microbiol 35, 1146-1155 49. Hantke, K. (1997) J Bacteriol 179, 6201-6204 50. Cellier, M. F., Bergevin, I., Boyer, E., and Richer, E. (2001) Trends Genet 17, 365-370 51. Grass, G., Franke, S., Taudte, N., Nies, D. H., Kucharski, L. M., Maguire, M. E., and Rensing, C. (2005) J Bacteriol 187, 1604-1611 52. Rajasekaran, M. B., Nilapwar, S., Andrews, S. C., and Watson, K. A. (2010) Biometals 23, 1-17 53. Cao, J., Woodhall, M. R., Alvarez, J., Cartron, M. L., and Andrews, S. C. (2007) Mol Microbiol 65, 857-875 54. Osman, D., and Cavet, J. S. (2008) Adv Appl Microbiol 65, 217-247 55. Tottey, S., Harvie, D. R., and Robinson, N. J. (2005) Acc Chem Res 38, 775-783 56. Wunderli-Ye, H., and Solioz, M. (2001) Biochem Biophys Res Commun 280, 713-719 57. Francis, M. S., and Thomas, C. J. (1997) Mol Gen Genet 253, 484-491 58. Lutsenko, S., Barnes, N. L., Bartee, M. Y., and Dmitriev, O. Y. (2007) Physiol Rev 87, 1011-1046 59. Rensing, C., and Grass, G. (2003) FEMS Microbiol Rev 27, 197-213 60. Gonzalez-Guerrero, M., Raimunda, D., Cheng, X., and Arguello, J. M. (2010) Mol Microbiol 78, 1246-1258 61. Bagai, I., Rensing, C., Blackburn, N. J., and McEvoy, M. M. (2008) Biochemistry 47, 11408-11414 148  62. Loftin, I. R., Franke, S., Blackburn, N. J., and McEvoy, M. M. (2007) Protein Sci 16, 2287-2293 63. Brown, N. L., Barrett, S. R., Camakaris, J., Lee, B. T., and Rouch, D. A. (1995) Mol Microbiol 17, 1153-1166 64. Djoko, K. Y., Xiao, Z., and Wedd, A. G. (2008) Chembiochem 9, 1579-1582 65. Young, K. T., Davis, L. M., and Dirita, V. J. (2007) Nat Rev Microbiol 5, 665-679 66. Robinson, D. A. (1981) Br Med J (Clin Res Ed) 282, 1584 67. Black, R. E., Levine, M. M., Clements, M. L., Hughes, T. P., and Blaser, M. J. (1988) J Infect Dis 157, 472-479 68. Blaser, M. J., Taylor, D. N., and Feldman, R. A. (1983) Epidemiol Rev 5, 157-176 69. Yuki, N., and Odaka, M. (2005) Curr Opin Neurol 18, 557-561 70. Friedman, C. R., Neimann, J., Wegener, H. C., and Tauxe, R. V. (2000) Epidemiology of Campylobacter jejuni infections in the United States and other industrialized nations. in Campylobacter (Nachamkin, I., and Blaser, M. J. eds.), 2nd Ed., ASM Press, Washington, D.C. pp 121-138 71. Altekruse, S. F., Stern, N. J., Fields, P. I., and Swerdlow, D. L. (1999) Emerg Infect Dis 5, 28-35 72. Sherman, P. M., Ossa, J. C., and Wine, E. (2010) Curr Opin Gastroenterol 26, 1-4 73. Szymanski, C. M., King, M., Haardt, M., and Armstrong, G. D. (1995) Infect Immun 63, 4295-4300 74. Hugdahl, M. B., Beery, J. T., and Doyle, M. P. (1988) Infect Immun 56, 1560-1566 75. Hendrixson, D. R., Akerley, B. J., and DiRita, V. J. (2001) Mol Microbiol 40, 214-224 76. Dorrell, N., Mangan, J. A., Laing, K. G., Hinds, J., Linton, D., Al-Ghusein, H., Barrell, B. G., Parkhill, J., Stoker, N. G., Karlyshev, A. V., Butcher, P. D., and Wren, B. W. (2001) Genome Res 11, 1706-1715 77. Parkhill, J., Wren, B. W., Mungall, K., Ketley, J. M., Churcher, C., Basham, D., Chillingworth, T., Davies, R. M., Feltwell, T., Holroyd, S., Jagels, K., Karlyshev, A. V., Moule, S., Pallen, M. J., Penn, C. W., Quail, M. A., Rajandream, M. A., Rutherford, K. M., van Vliet, A. H., Whitehead, S., and Barrell, B. G. (2000) Nature 403, 665-668 78. de Boer, P., Wagenaar, J. A., Achterberg, R. P., van Putten, J. P., Schouls, L. M., and Duim, B. (2002) Mol Microbiol 44, 351-359 79. Bacon, D. J., Alm, R. A., Burr, D. H., Hu, L., Kopecko, D. J., Ewing, C. P., Trust, T. J., and Guerry, P. (2000) Infect Immun 68, 4384-4390 80. Fouts, D. E., Mongodin, E. F., Mandrell, R. E., Miller, W. G., Rasko, D. A., Ravel, J., Brinkac, L. M., DeBoy, R. T., Parker, C. T., Daugherty, S. C., Dodson, R. J., Durkin, A. S., Madupu, R., Sullivan, S. A., Shetty, J. U., Ayodeji, M. A., Shvartsbeyn, A., Schatz, M. C., Badger, J. H., Fraser, C. M., and Nelson, K. E. (2005) PLoS Biol 3, e15 81. Hofreuter, D., Tsai, J., Watson, R. O., Novik, V., Altman, B., Benitez, M., Clark, C., Perbost, C., Jarvie, T., Du, L., and Galan, J. E. (2006) Infect Immun 74, 4694-4707 82. Byrne, C. M., Clyne, M., and Bourke, B. (2007) Microbiology 153, 561-569 83. Beery, J. T., Hugdahl, M. B., and Doyle, M. P. (1988) Appl Environ Microbiol 54, 2365-2370 84. Meinersmann, R. J., Rigsby, W. E., Stern, N. J., Kelley, L. C., Hill, J. E., and Doyle, M. P. (1991) Am J Vet Res 52, 1518-1522 85. Ketley, J. M. (1997) Microbiology 143 ( Pt 1), 5-21 149  86. Russell, R. G., O'Donnoghue, M., Blake, D. C., Jr., Zulty, J., and DeTolla, L. J. (1993) J Infect Dis 168, 210-215 87. Crushell, E., Harty, S., Sharif, F., and Bourke, B. (2004) Pediatr Res 55, 3-12 88. Wandersman, C., and Delepelaire, P. (2004) Annu Rev Microbiol 58, 611-647 89. Field, L. H., Headley, V. L., Payne, S. M., and Berry, L. J. (1986) Infect. Immun. 54, 126-132 90. Pickett, C. L., Auffenberg, T., Pesci, E. C., Sheen, V. L., and Jusuf, S. S. (1992) Infect. Immun. 60, 3872-3877 91. Holmes, K., Mulholland, F., Pearson, B. M., Pin, C., McNicholl-Kennedy, J., Ketley, J. M., and Wells, J. M. (2005) Microbiology 151, 243-257 92. van Vliet, A. H., Baillon, M. L., Penn, C. W., and Ketley, J. M. (1999) J Bacteriol 181, 6371-6376 93. Palyada, K., Sun, Y. Q., Flint, A., Butcher, J., Naikare, H., and Stintzi, A. (2009) BMC Genomics 10, 481 94. Stintzi, A., Van Vliet, A. H., and Ketley, J. M. (2008) Iron Metabolism, Transport and Regulation. in Campylobacter (Nachamkin, I., Szymanski, C., and Blaser, M. eds.), 3rd ed. Ed., ASM Press, Washington, DC. pp 591-610 95. Raphael, B. H., and Joens, L. A. (2003) Can. J. Microbiol. 49, 727-731 96. Naikare, H., Palyada, K., Panciera, R., Marlow, D., and Stintzi, A. (2006) Infect Immun 74, 5433-5444 97. Pfeiffer, J., Guhl, J., Waidner, B., Kist, M., and Bereswill, S. (2002) Infect Immun 70, 3930-3934 98. Chaudhuri, R. R., and Pallen, M. J. (2006) Nucleic Acids Res 34, D335-337 99. Galindo, M. A., Day, W. A., Raphael, B. H., and Joens, L. A. (2001) Curr. Microbiol. 42, 139-143 100. Richardson, P. T., and Park, S. F. (1995) Microbiology 141 ( Pt 12), 3181-3191 101. Contreras, M., Thiberge, J. M., Mandrand-Berthelot, M. A., and Labigne, A. (2003) Mol Microbiol 49, 947-963 102. Stoof, J., Kuipers, E. J., Klaver, G., and van Vliet, A. H. Infect Immun  103. Rock, J. D., van Vliet, A. H. M., and Ketley, J. M. (2001) Int. J. Med. Microbiol. 291, 125 104. Ridley, K. A., Rock, J. D., Li, Y., and Ketley, J. M. (2006) J. Bacteriol. 188, 7862-7875 105. Stintzi, A., Marlow, D., Palyada, K., Naikare, H., Panciera, R., Whitworth, L., and Clarke, C. (2005) Infect. Immun. 73, 1797-1810 106. Miller, C. E., Rock, J. D., Ridley, K. A., Williams, P. H., and Ketley, J. M. (2008) J Bacteriol 190, 1900-1911 107. Tokuda, H., and Matsuyama, S. (2004) Biochim. Biophys. Acta 1693, 5-13 108. Ochsner, U. A., Johnson, Z., and Vasil, M. L. (2000) Microbiology 146, 185-198 109. Kosman, D. J. (2003) Mol Microbiol 47, 1185-1197 110. Grosse, C., Scherer, J., Koch, D., Otto, M., Taudte, N., and Grass, G. (2006) Mol Microbiol 62, 120-131 111. Stearman, R., Yuan, D. S., Yamaguchi-Iwai, Y., Klausner, R. D., and Dancis, A. (1996) Science 271, 1552-1557 112. Janvier, B., Constantinidou, C., Aucher, P., Marshall, Z. V., Penn, C. W., and Fauchere, J. L. (1998) Res Microbiol 149, 95-107 113. van Vliet, A. H., Wooldridge, K. G., and Ketley, J. M. (1998) J Bacteriol 180, 5291-5298 150  114. Deka, R. K., Brautigam, C. A., Tomson, F. L., Lumpkins, S. B., Tomchick, D. R., Machius, M., and Norgard, M. V. (2007) J Biol Chem 282, 5944-5958 115. Dubbels, B. L., DiSpirito, A. A., Morton, J. D., Semrau, J. D., Neto, J. N., and Bazylinski, D. A. (2004) Microbiology 150, 2931-2945 116. Gordon, D. M., and Cowling, A. (2003) Microbiology 149, 3575-3586 117. Tenaillon, O., Skurnik, D., Picard, B., and Denamur, E. (2010) Nat Rev Microbiol 8, 207-217 118. Farthing, M. J. (2004) Best Pract Res Clin Gastroenterol 18, 233-239 119. Vollaard, E. J., and Clasener, H. A. (1994) Antimicrob Agents Chemother 38, 409-414 120. Hudault, S., Guignot, J., and Servin, A. L. (2001) Gut 49, 47-55 121. Wiles, T. J., Kulesus, R. R., and Mulvey, M. A. (2008) Exp Mol Pathol 85, 11-19 122. Johnson, J. R., and Russo, T. A. (2002) J Lab Clin Med 139, 155-162 123. Foxman, B. (2003) Dis Mon 49, 53-70 124. Russo, T. A., Stapleton, A., Wenderoth, S., Hooton, T. M., and Stamm, W. E. (1995) J Infect Dis 172, 440-445 125. Manges, A. R., Johnson, J. R., Foxman, B., O'Bryan, T. T., Fullerton, K. E., and Riley, L. W. (2001) N Engl J Med 345, 1007-1013 126. Foxman, B., Manning, S. D., Tallman, P., Bauer, R., Zhang, L., Koopman, J. S., Gillespie, B., Sobel, J. D., and Marrs, C. F. (2002) Am J Epidemiol 156, 1133-1140 127. Anderson, G. G., Palermo, J. J., Schilling, J. D., Roth, R., Heuser, J., and Hultgren, S. J. (2003) Science 301, 105-107 128. Leveille, S., Caza, M., Johnson, J. R., Clabots, C., Sabri, M., and Dozois, C. M. (2006) Infect Immun 74, 3427-3436 129. Ouyang, Z., and Isaacson, R. (2006) Infect Immun 74, 6949-6956 130. Torres, A. G., Redford, P., Welch, R. A., and Payne, S. M. (2001) Infect Immun 69, 6179-6185 131. Sabri, M., Leveille, S., and Dozois, C. M. (2006) Microbiology 152, 745-758 132. Makui, H., Roig, E., Cole, S. T., Helmann, J. D., Gros, P., and Cellier, M. F. (2000) Mol Microbiol 35, 1065-1078 133. Mergeay, M., Nies, D., Schlegel, H. G., Gerits, J., Charles, P., and Van Gijsegem, F. (1985) J Bacteriol 162, 328-334 134. Taudte, N., and Grass, G. (2010) Biometals 23, 643-656 135. Stapleton, A., Moseley, S., and Stamm, W. E. (1991) J Infect Dis 163, 773-779 136. Gaynor, E. C., Cawthraw, S., Manning, G., MacKichan, J. K., Falkow, S., and Newell, D. G. (2004) J Bacteriol 186, 503-517 137. Korlath, J. A., Osterholm, M. T., Judy, L. A., Forfang, J. C., and Robinson, R. A. (1985) J Infect Dis 152, 592-596 138. Gaynor, E. C., Wells, D. H., MacKichan, J. K., and Falkow, S. (2005) Mol. Microbiol. 56, 8-27 139. Davis, L., Young, K., and DiRita, V. (2008) Curr Protoc Microbiol Chapter 8, Unit 8A 2 1-8A 2 17 140. Yao, R., Alm, R. A., Trust, T. J., and Guerry, P. (1993) Gene 130, 127-130 141. Karlyshev, A. V., and Wren, B. W. (2005) Appl. Environ. Microbiol. 71, 4004-4013 142. Choi, K. H., and Schweizer, H. P. (2006) Nat Protoc 1, 153-161 143. Anderson, M. A., and Morel, F. M. M. (1982) Limnology and Oceanography 27, 789-813 151  144. Grass, G., Otto, M., Fricke, B., Haney, C. J., Rensing, C., Nies, D. H., and Munkelt, D. (2005) Arch Microbiol 183, 9-18 145. Van Duyne, G. D., Standaert, R. F., Karplus, P. A., Schreiber, S. L., and Clardy, J. (1993) J. Mol. Biol. 229, 105-124 146. Miyazaki, K., and Takenouchi, M. (2002) Biotechniques 33, 1033-1034, 1036-1038 147. Brenner, A. J., and Harris, E. D. (1995) Anal Biochem 226, 80-84 148. Haigler, B. E., and Gibson, D. T. (1990) J Bacteriol 172, 457-464 149. Bruser, T., Deutzmann, R., and Dahl, C. (1998) FEMS Microbiol Lett 164, 329-336 150. Leon-Kempis, M. R., Guccione, E., Mulholland, F., Williamson, M. P., and Kelly, D. J. (2006) Mol. Microbiol. 60, 1262-1275 151. Logan, S. M., and Trust, T. J. (1982) Infect Immun 38, 898-906 152. Page, W. J., and Taylor, D. E. (1988) J. Gen. Microbiol. 134, 2925-2932 153. Otwinowski, Z., and Minor, W. (1997) Methods Enzymol. 276, 307-326 154. Terwilliger, T. C., and Berendzen, J. (1999) Acta Crystallogr. Sect. D Biol. Crystallogr. 55, 849-861 155. Jones, T. A., Zou, J. Y., Cowan, S. W., and Kjeldgaard, M. (1991) Acta Crystallogr. Sect. A Found. Crystallogr. 47, 110-119 156. Murshudov, G. N., Vagin, A. A., and Dodson, E. J. (1997) Acta Crystallogr D Biol Crystallogr 53, 240-255 157. Collaborative Computation Project 4. (1994) Acta Crystallogr Sect D Biol Crystallogr 50, 760-763 158. Laskowski, R. A., MacArthur, M. W., Moss, D. S., and Thornton, J. M. (1993) J. Appl. Cryst. 26, 283-291 159. Sheldrick, G. M. (2008) Acta Crystallogr A 64, 112-122 160. Kabsch, W. (1993) J. Appl. Cryst. 26, 795-800 161. Evans, P. (1997) Joint CCP4 and ESF-EACMB Newsletter on Protein Crystallography 33, 22-24 162. Emsley, P., and Cowtan, K. (2004) Acta Crystallogr D Biol Crystallogr 60, 2126-2132 163. Leslie, A. G. W. (1992) Joint CCP4 + ESF-EAMCB Newsletter on Protein Crystallography 26, 27-33 164. Vagin, A. A., and Teplyakov, A. (1997) J Appl Cryst 30, 1022-1025 165. Schuck, P., Perugini, M. A., Gonzales, N. R., Howlett, G. J., and Schubert, D. (2002) Biophys J 82, 1096-1111 166. Laue, T. M., Shah, B. D., Ridgeway, T. M., and Pelletier, S. L. (1992). in Analytical Ultracentrifugation in Biochemistry and Polymer Science (Harding, S. E., Rowe, A. J., and Horton, J. C. eds.), Royal Society of Chemistry, Cambridge. pp 90-125 167. Schuck, P. (2003) Anal Biochem 320, 104-124 168. Grosse, C., Grass, G., Anton, A., Franke, S., Santos, A. N., Lawley, B., Brown, N. L., and Nies, D. H. (1999) J Bacteriol 181, 2385-2393 169. Thompson, J. D., Gibson, T. J., Plewniak, F., Jeanmougin, F., and Higgins, D. G. (1997) Nucleic Acids Res. 24, 4876-4882 170. Hall, T. A. (1999) Nucl. Acids Symp. Ser. 41, 95-98 171. Schmidt, H. A., Strimmer, K., Vingron, M., and Haeseler, A. V. (2002) Bioinformatics 18, 502-504 172. Collaborative Computation Project 4. (1994) Acta Crystallogr. Sect. D Biol. Crystallogr. 50, 760-763 152  173. Sigrist, C. J., Cerutti, L., Hulo, N., Gattiker, A., Falquet, L., Pagni, M., Bairoch, A., and Bucher, P. (2002) Brief. Bioinform. 3, 265-274 174. Holm, L., and Sander, C. (1993) J. Mol. Biol. 233, 123-138 175. Zhou, T., Radaev, S., Rosen, B. P., and Gatti, D. L. (2000) EMBO J. 19, 4838-4845 176. Gill, S. C., and von Hippel, P. H. (1989) Anal. Biochem. 182, 319-326 177. Page, R. D. W. (1996) Comput. Appl. Biosci. 12, 357-358 178. Konkel, M. E., Garvis, S. G., Tipton, S. L., Anderson, D. E., Jr., and Cieplak, W., Jr. (1997) Mol Microbiol 24, 953-963 179. Mauk, A. G., and Moore, G. R. (1997) J. Biol. Inorg. Chem. 2, 119-125 180. Moore, G. R., and Williams, R. J. (1977) FEBS Lett. 79, 229-232 181. Hildebrand, D. P., Burk, D. L., Maurus, R., Ferrer, J. C., Brayer, G. D., and Mauk, A. G. (1995) Biochemistry 34, 1997-2005 182. Kassner, R. J. (1972) Proc. Natl. Acad. Sci. U.S.A. 69, 2263-2267 183. Kassner, R. J. (1973) J. Am. Chem. Soc. 95, 2674-2677 184. Stellwagen, E. (1978) Nature 275, 73-74 185. Rafferty, S. P., Pearce, L. L., Barker, P. D., Guillemette, J. G., Kay, C. M., Smith, M., and Mauk, A. G. (1990) Biochemistry 29, 9365-9369 186. Varadarajan, R., Zewert, T. E., Gray, H. B., and Boxer, S. G. (1989) Science 243, 69-72 187. Arnoux, P., Haser, R., Izadi, N., Lecroisey, A., Delepierre, M., Wandersman, C., and Czjzek, M. (1999) Nat. Struct. Biol. 6, 516-520 188. Fulop, V., Moir, J. W., Ferguson, S. J., and Hajdu, J. (1995) Cell 81, 369-377 189. Williams, P. A., Fulop, V., Garman, E. F., Saunders, N. F., Ferguson, S. J., and Hajdu, J. (1997) Nature 389, 406-412 190. Eakanunkul, S., Lukat-Rodgers, G. S., Sumithran, S., Ghosh, A., Rodgers, K. R., Dawson, J. H., and Wilks, A. (2005) Biochemistry 44, 13179-13191 191. Wardell, M., Wang, Z., Ho, J. X., Robert, J., Ruker, F., Ruble, J., and Carter, D. C. (2002) Biochem. Biophys. Res. Commun. 291, 813-819 192. Putnam, C. D., Arvai, A. S., Bourne, Y., and Tainer, J. A. (2000) J. Mol. Biol. 296, 295-309 193. Palcic, M., and Dunford, H. B. (1979) Can. J. Biochem. 57, 321-329 194. Lippard, S. J., and Berg, J. M. (1994) Principles of Bioinorganic Chemistry, University Science Books, Mill Valley, Calif. 195. Cheng, L., Lee, J., Powell, D. R., and Richter-Addo, G. B. (2004) Acta Crystallogr. Sect. E Biol. Crystallogr. 60, m1340-m1342 196. Brig?, A., Leys, D., Meyer, T. E., Cusanovich, M. A., and Van Beeumen, J. J. (2002) Biochemistry 41, 4827-4836 197. Heitmann, D., and Einsle, O. (2005) Biochemistry 44, 12411-12419 198. Matias, P. M., Morais, J., Coelho, A. V., Meijers, R., Gonzalez, A., Thompson, A. W., Sieker, L., LeGall, J., and Carrondo, M. A. (1997) J. Biol. Inorg. Chem. 2, 507-514 199. Rock, J. D., van Vliet, A. H. M., and Ketley, J. M. (2001) Int J Med Microbiol 291, 125 200. Hossain, A., Stewart-Tull, D. E., and Freer, J. H. (1993) FEMS Immunol Med Microbiol 6, 331-339 201. Remaut, H., Safarov, N., Ciurli, S., and Van Beeumen, J. (2001) J. Biol. Chem. 276, 49365-49370 202. Wernimont, A. K., Huffman, D. L., Lamb, A. L., O'Halloran, T. V., and Rosenzweig, A. C. (2000) Nat. Struct. Biol. 7, 766-771 153  203. Dailey, H. A., Jr., and Lascelles, J. (1977) J Bacteriol 129, 815-820 204. Krissinel, E., and Henrick, K. (2007) J Mol Biol 372, 774-797 205. Reynolds, C., Damerell, D., and Jones, S. (2009) Bioinformatics 25, 413-414 206. Frolow, F., Kalb, A. J., and Yariv, J. (1994) Nat Struct Biol 1, 453-460 207. Langlois d'Estaintot, B., Santambrogio, P., Granier, T., Gallois, B., Chevalier, J. M., Precigoux, G., Levi, S., and Arosio, P. (2004) J Mol Biol 340, 277-293 208. Sippl, M. J., and Wiederstein, M. (2008) Bioinformatics 24, 426-427 209. Holm, L., Kaariainen, S., Rosenstrom, P., and Schenkel, A. (2008) Bioinformatics 24, 2780-2781 210. Wernimont, A. K., Huffman, D. L., Finney, L. A., Demeler, B., O'Halloran, T. V., and Rosenzweig, A. C. (2003) J Biol Inorg Chem 8, 185-194 211. Thompson, J. D., Gibson, T. J., and Higgins, D. G. (2002) Curr Protoc Bioinformatics Chapter 2, Unit 2 3 212. Poly, F., and Guerry, P. (2008) Curr Opin Gastroenterol 24, 27-31 213. Davis, A. V., and O'Halloran, T. V. (2008) Nat Chem Biol 4, 148-151 214. Peariso, K., Huffman, D. L., Penner-Hahn, J. E., and O'Halloran, T. V. (2003) J Am Chem Soc 125, 342-343 215. Zhang, L., Koay, M., Maher, M. J., Xiao, Z., and Wedd, A. G. (2006) J Am Chem Soc 128, 5834-5850 216. Arnesano, F., Banci, L., Bertini, I., Mangani, S., and Thompsett, A. R. (2003) Proc Natl Acad Sci U S A 100, 3814-3819 217. Djoko, K. Y., Xiao, Z., Huffman, D. L., and Wedd, A. G. (2007) Inorg Chem 46, 4560-4568 218. Tottey, S., Waldron, K. J., Firbank, S. J., Reale, B., Bessant, C., Sato, K., Cheek, T. R., Gray, J., Banfield, M. J., Dennison, C., and Robinson, N. J. (2008) Nature 455, 1138-1142 219. Castagnetto, J. M., Hennessy, S. W., Roberts, V. A., Getzoff, E. D., Tainer, J. A., and Pique, M. E. (2002) Nucleic Acids Res 30, 379-382 220. Emerson, J. P., Kovaleva, E. G., Farquhar, E. R., Lipscomb, J. D., and Que, L., Jr. (2008) Proc Natl Acad Sci U S A 105, 7347-7352 221. Stallings, W. C., Pattridge, K. A., Strong, R. K., and Ludwig, M. L. (1984) J Biol Chem 259, 10695-10699 222. Quintanar, L., Gebhard, M., Wang, T. P., Kosman, D. J., and Solomon, E. I. (2004) J Am Chem Soc 126, 6579-6589 223. Stoj, C. S., Augustine, A. J., Zeigler, L., Solomon, E. I., and Kosman, D. J. (2006) Biochemistry 45, 12741-12749 224. Taylor, A. B., Stoj, C. S., Ziegler, L., Kosman, D. J., and Hart, P. J. (2005) Proc Natl Acad Sci U S A 102, 15459-15464 225. Wang, T. P., Quintanar, L., Severance, S., Solomon, E. I., and Kosman, D. J. (2003) J Biol Inorg Chem 8, 611-620 226. Askwith, C., Eide, D., Van Ho, A., Bernard, P. S., Li, L., Davis-Kaplan, S., Sipe, D. M., and Kaplan, J. (1994) Cell 76, 403-410 227. Hall, S. J., Hitchcock, A., Butler, C. S., and Kelly, D. J. (2008) J Bacteriol 190, 8075-8085 228. Kosman, D. J. J Biol Inorg Chem 15, 15-28 154  229. Nies, D., and Grass, G. (2009) Transition metal homeostasis. in EcoSal-Escherichia coli and Salmonella: cellular and molecular biology (B?ck, A., Curtiss, R., 3rd, Kaper, J. B., Neidhardt, F. C., Nystr?m, T., Rudd, K. E., and Squires, C. L. eds.), ASM Press, Washington, D.C. pp  230. Letoffe, S., Heuck, G., Delepelaire, P., Lange, N., and Wandersman, C. (2009) Proc Natl Acad Sci U S A 106, 11719-11724 231. Brooks, T., and Keevil, C. W. (1997) Lett Appl Microbiol 24, 203-206 232. Millero, F. J. (1998) Earth and Planetary Science Letters 154, 323-329 233. Chan, A. C., Doukov, T. I., Scofield, M., Tom-Yew, S. A., Ramin, A. B., Mackichan, J. K., Gaynor, E. C., and Murphy, M. E. (2010) J Mol Biol 401, 590-604 234. Chakrabarti, P. (1989) Biochemistry 28, 6081-6085 235. de Silva, D., Davis-Kaplan, S., Fergestad, J., and Kaplan, J. (1997) J Biol Chem 272, 14208-14213 236. Sedlak, E., Ziegler, L., Kosman, D. J., and Wittung-Stafshede, P. (2008) Proc Natl Acad Sci U S A 105, 19258-19263 237. Skaar, E. P., Humayun, M., Bae, T., DeBord, K. L., and Schneewind, O. (2004) Science 305, 1626-1628 238. Vaknin, A., and Berg, H. C. (2004) Proc Natl Acad Sci U S A 101, 17072-17077 239. Hunger, K., Beckering, C. L., Wiegeshoff, F., Graumann, P. L., and Marahiel, M. A. (2006) J Bacteriol 188, 240-248 240. Krieg, S., Huche, F., Diederichs, K., Izadi-Pruneyre, N., Lecroisey, A., Wandersman, C., Delepelaire, P., and Welte, W. (2009) Proc Natl Acad Sci U S A 106, 1045-1050 241. Pawelek, P. D., Croteau, N., Ng-Thow-Hing, C., Khursigara, C. M., Moiseeva, N., Allaire, M., and Coulton, J. W. (2006) Science 312, 1399-1402   


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