Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Genetic and phenotypic characterization of novel Bordetella pertussis lipid A modifications Shah, Nita Reva 2014

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

Item Metadata

Download

Media
24-ubc_2014_september_shah_nita. [ 6.38MB ]
Metadata
JSON: 24-1.0165993.json
JSON-LD: 24-1.0165993-ld.json
RDF/XML (Pretty): 24-1.0165993-rdf.xml
RDF/JSON: 24-1.0165993-rdf.json
Turtle: 24-1.0165993-turtle.txt
N-Triples: 24-1.0165993-rdf-ntriples.txt
Original Record: 24-1.0165993-source.json
Full Text
24-1.0165993-fulltext.txt
Citation
24-1.0165993.ris

Full Text

GENETIC AND PHENOTYPIC CHARACTERIZATION OF NOVEL BORDETELLA PERTUSSIS LIPID A MODIFICATIONS  by Nita Reva Shah  B.Sc., The University of British Columbia, 2008  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Microbiology and Immunology)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  August 2014  © Nita Reva Shah, 2014 ii  Abstract Lipopolysaccharide (LPS) is a component of the outer membrane in most Gram-negative bacteria. The lipid A region of LPS anchors the molecule to the outer membrane and forms the first barrier between Gram-negative bacteria and the extracellular environment. Lipid A is also important for bacterial interaction and activation of host immune cells through binding to Toll-like receptor 4 (TLR4), activation of which results in a downstream inflammation response. The work presented in this thesis explores the genetic basis for different lipid A structures and the effects of this structural variability on activation of TLR4 and resistance to cationic antimicrobial peptides (CAMPs). Penta-acyl lipid A from B. pertussis strain BP338 is modified with glucosamine (GlcN), and mutational analyses revealed that LgmA, LgmB, and LgmC are required for this modification. Bioinformatic analysis suggests the following hypothetical model: LgmA transfers N-acetylglucosamine (GlcNAc) from UDP-GlcNAc to the carrier lipid C55P, LgmC removes the acetyl group, and LgmB transfers GlcN from C55P to the lipid A phosphate. Glycosyltransferase assays with LgmA-expressing E. coli membranes show that LgmA transfers GlcNAc onto a lipid, therefore supporting the first step in this model. Site-directed mutagenesis has also identified a putative active site in LgmA and LgmC. Killing assays show the GlcN modification in B. pertussis increases resistance to numerous CAMPs and to outer membrane perturbation by EDTA. Lipid A from B. pertussis strain 18-323 exhibits low levels of TLR4 activation. Lipid A of strain 18-323 has no GlcN modification (due to an incomplete lgm locus) and a shorter C3’ acyl chain compared to strain BP338 (due to a difference in LpxA). Complementation of 18-323 with BP338 lipid A-modifying genes lpxA and/or the lgm locus increase TLR4 activation, though the GlcN modification had a dominant effect. In hexa-acyl E. coli lipid A, shortening the C3 and C3’ acyl chains had been shown to decrease TLR4 activation and resistance to polymyxin B, but increase activation of the limulus amebocyte lysate assay. Thus, varying the structure of lipid A, in both B. pertussis and E. coli, can affect TLR4 activation and resistance of the bacteria to CAMPs.  iii  Preface I designed, performed, and analyzed the data from all experiments detailed in this thesis, with the following exceptions:  Mass spectrometry analysis and structural determination of all lipid A structures and isolation of LPS from the strains 18-323, 18-323 + pPtacLgmABCD, 18-323 + pPtacLpxA338, and 18-323 + pPtacLgmABCDLpxA338 was performed in the laboratory of Dr. Martine Caroff (Université de Paris-Sud, Orsay, France) by Sami Albitar-Nehme and Alexey Novikov.  Whole genome sequencing of Bordetella hinzii strain ATCC 51730 and Bordetella trematum strain CCUG 13902 and assembly of the draft genomes was performed in the laboratory of Dr. Martin Hirst (University of British Columbia, Vancouver, Canada) by Michelle Moksa.  I supervised Emma Kim in the cloning of the vectors: pET30LgmA and pPtacLgmAB, and in the analysis of the B. pertussis 18-323 sequence raw reads for the presence of the lgm locus genes.  I supervised Andrew Low in the cloning of vectors: pBBR2LgmA D76G D77G, pBBR2LgmA D127G, pBBR2LgmA D129G, pBBR2LgmA D127G D129G, pBBR2LgmA D159N, pBBR2LgmC D80G D81G, pBBR2LgmC H130G, pBBR2LgmC D187G D189G, pBBR2LgmC E313G; and in the experiments shown in Figure 22 and Figure 25.  A version of Sections 3.2.1, 4.3, and 4.4 have been published. Shah, N. R., S. Albitar-Nehme, E. Kim, N. Marr, A. Novikov, M. Caroff, and R. C. Fernandez. 2013. Minor modifications to the phosphate groups and the C3' acyl chain length of lipid A in two Bordetella pertussis strains, BP338 and 18-323, independently affect Toll-like receptor 4 protein activation. The Journal of Biological Chemistry 288:11751-11760.  A version of Section 3.6 has been published. Novikov, A., N. R. Shah, S. Albitar-Nehme, S. M. Basheer, I. Trento, A. Tirsoaga, M. Moksa, M. Hirst, M. B. Perry, A. E. Hamidi, R. C. Fernandez, and M. Caroff. iv  2013. Complete Bordetella avium, Bordetella hinzii and Bordetella trematum lipid A structures and genomic sequence analyses of the loci involved in their modifications. Innate Immunity: 10.1177/1753425913506950. I completed the genomic analysis to support the lipid A structures and wrote these sections of the paper.  The original draft genome sequences in Section 3.6 have been published. Shah, N. R., M. Moksa, A. Novikov, M. B. Perry, M. Hirst, M. Caroff, and R. C. Fernandez. 2013. Draft genome sequences of Bordetella hinzii and Bordetella trematum. Genome announcements 1:e00838-13.   A version of Section 4.2 has been published. Shah, N. R., R. E. W. Hancock, R. C. Fernandez. Bordetella pertussis lipid A glucosamine modification confers resistance to cationic antimicrobial peptides and increases resistance to outer membrane perturbation. Antimicrobial Agents and Chemotherapy 58(8):4931-4  Biohazard Approval Certificate numbers under which this research was conducted are: H06-0225, H06-0226, B06-0225, and B06-0226 issued by the Biosafety Committee, Office of Research Services at the University of British Columbia.  Radioisotope Approval Certificate number under which this research was conducted is MICB-3399-18 issued by the University of British Columbia Committee on Radioisotopes and Radiation Hazards. v  Table of Contents  Abstract ........................................................................................................................................................ ii Preface ......................................................................................................................................................... iii Table of Contents ........................................................................................................................................ v List of Tables ............................................................................................................................................... x List of Figures ............................................................................................................................................. xi List of Abbreviations ............................................................................................................................... xiv Acknowledgements ................................................................................................................................ xviii Dedication .................................................................................................................................................. xx Chapter 1: Introduction ............................................................................................................................. 1 1.1 Bordetella pertussis and whooping cough ................................................................................... 1 1.1.1 Clinical manifestation ............................................................................................................. 1 1.1.2 Molecular pathogenesis ........................................................................................................... 2 1.1.2.1 BvgAS regulation system ............................................................................................... 2 1.1.2.2 Adhesins ......................................................................................................................... 3 1.1.2.3 Toxins ............................................................................................................................. 3 1.1.2.4 Other virulence factors ................................................................................................... 4 1.1.2.5 Immune evasion and modulation .................................................................................... 4 1.1.3 Other Bordetella species ......................................................................................................... 5 1.2 Lipopolysaccharide ...................................................................................................................... 6 1.2.1 LPS structure ........................................................................................................................... 8 1.2.1.1 Lipid A ......................................................................................................................... 10 1.2.1.2 Core sugars ................................................................................................................... 10 1.2.1.3 O-antigen ...................................................................................................................... 10 vi  1.2.2 LPS Biogenesis ..................................................................................................................... 11 1.2.2.1 Raetz lipid A biosynthesis pathway.............................................................................. 11 1.2.2.2 Synthesis of the core sugars.......................................................................................... 12 1.2.2.3 O-antigen synthesis and ligation................................................................................... 12 1.2.2.4 Transport of LPS to the outer membrane ..................................................................... 13 1.2.3 Biological role of LPS ........................................................................................................... 13 1.2.3.1 LPS and membrane integrity ........................................................................................ 14 1.2.3.2 Host-pathogen interaction ............................................................................................ 14 1.2.4 LPS modification systems ..................................................................................................... 16 1.2.4.1 Acyl chain modifications .............................................................................................. 17 1.2.4.2 Phosphate modifications ............................................................................................... 17 1.2.4.3 Core sugars and O-antigen modifications .................................................................... 19 1.3 B. pertussis LPS ......................................................................................................................... 19 1.3.1 B. pertussis LPS structure ..................................................................................................... 20 1.3.2 Lipid A structure of strains BP338 and 18-323 ..................................................................... 20 1.3.3 GlcN modification and the lgm locus .................................................................................... 23 1.3.4 Bioinformatic analysis of the lgm locus and a hypothetical model ....................................... 25 1.4 Hypotheses ................................................................................................................................ 28 1.5 Thesis goals ............................................................................................................................... 29 Chapter 2: Materials and methods .......................................................................................................... 31 2.1 Bacterial growth conditions ....................................................................................................... 31 2.2 Strains, plasmids, and primers ................................................................................................... 31 2.2.1 Bacterial strains ..................................................................................................................... 31 2.2.2 Plasmids ................................................................................................................................ 33 2.2.3 Primers .................................................................................................................................. 35 vii  2.3 Cloning of vectors and deletion strains ..................................................................................... 37 2.3.1 General cloning techniques ................................................................................................... 37 2.3.2 Generating markerless deletion mutants ............................................................................... 39 2.3.3 Site-directed mutagenesis ...................................................................................................... 42 2.3.4 Vectors to complement B. pertussis strains........................................................................... 44 2.3.5 Vectors to complement E. coli strain R0138 ......................................................................... 46 2.3.6 Vector for LgmA expression in E. coli strain BL-21 ............................................................ 47 2.4 Preparation of bacterial cells for mass spectrometry analysis and TLR4-activation assays ...... 47 2.5 Isolation of LPS and lipid A ...................................................................................................... 48 2.5.1 Isolation of B. pertussis LPS for TLR4 activation assays ..................................................... 48 2.5.2 Isolation of E. coli LPS for TLR4 and LAL activation assays .............................................. 48 2.5.3 Isolation of lipid A for mass spectrometry analysis .............................................................. 49 2.6 Mass spectrometry analysis ....................................................................................................... 50 2.7 HEK-Blue hTLR4 activation assay ........................................................................................... 50 2.7.1 Maintenance of HEK-Blue cells ........................................................................................... 50 2.7.2 hTLR4 activation assay ......................................................................................................... 51 2.8 E. coli growth curves ................................................................................................................. 51 2.9 Bacterial survival assays ............................................................................................................ 52 2.9.1 Polymyxin B growth curve assay .......................................................................................... 52 2.9.2 Killing assays ........................................................................................................................ 52 2.10 LAL activation assay ................................................................................................................. 54 2.11 Glycosyltransferase assay .......................................................................................................... 54 2.12 Western blot analysis ................................................................................................................. 56 2.13 Whole genome sequencing ........................................................................................................ 56 2.14 Reverse transcriptase PCR......................................................................................................... 57 viii  2.15 Bioinformatic analysis tools ...................................................................................................... 58 2.16 Statistical analysis ..................................................................................................................... 58 Chapter 3: Identification and characterization of the lgm locus in Bordetella .................................... 59 3.1 Introduction ............................................................................................................................... 59 3.2 The lgm locus is required for lipid A GlcN modification in B. pertussis .................................. 59 3.2.1 lgmA, lgmB, and lgmC are required for lipid A GlcN modification...................................... 60 3.2.2 Potential flippase replacements are not required for GlcN modification .............................. 66 3.3 LgmA functions as a GlcNAc transferase ................................................................................. 75 3.4 Identification of the putative active site of LgmA ..................................................................... 79 3.5 Identification of the putative active site of LgmC ..................................................................... 84 3.6 The lgm locus in other Bordetella species ................................................................................. 90 3.7 Discussion .................................................................................................................................. 98 Chapter 4: The biological effects of lipid A modifications in B. pertussis .......................................... 102 4.1 Introduction ............................................................................................................................. 102 4.2 Effect of GlcN modification on resistance to CAMPs and membrane stability ...................... 102 4.2.1 GlcN modification increases resistance to CAMPs ............................................................ 103 4.2.2 GlcN modification increases resistance to OM perturbation .............................................. 106 4.3 lgm locus and lpxA are involved in the differences in lipid A structure between strains BP338 and 18-323 ............................................................................................................................................ 108 4.3.1 Difference in GlcN modification is due an incomplete lgm locus ...................................... 108 4.3.2 Difference in C3’ acyl chain length is due to LpxA ............................................................ 112 4.4 C3’ acyl chain length and GlcN modification individually affect hTLR4 activation .............. 114 4.5 Discussion ................................................................................................................................ 116 Chapter 5: The biological effects of varying the C3 and C3’ acyl chain lengths in E. coli hexa-acyl lipid A ....................................................................................................................................................... 121 ix  5.1 Introduction ............................................................................................................................. 121 5.2 Generation of R0138 E. coli strains with different C3 and C3’ acyl chain lengths ................. 122 5.3 C3 and C3’ acyl chain lengths affect bacterial growth ............................................................ 128 5.4 C3 and C3’ acyl chain lengths affect resistance to polymyxin B ............................................ 130 5.5 C3 and C3’ acyl chain lengths affect hTLR4 activation .......................................................... 133 5.6 C3 and C3’ acyl chain lengths affect LAL activation ............................................................. 135 5.7 Discussion ................................................................................................................................ 137 Chapter 6: Discussion ............................................................................................................................. 141 6.1 Discussion ................................................................................................................................ 141 6.2 Conclusions ............................................................................................................................. 149 6.3 Future directions ...................................................................................................................... 150 References ................................................................................................................................................ 153 Appendices ............................................................................................................................................... 161 Appendix A Mass spectra ..................................................................................................................... 161 Appendix B Sequence alignments ........................................................................................................ 165    x  List of Tables  Table 1. List of bacterial strains .................................................................................................................. 33 Table 2. List of plasmids ............................................................................................................................. 34 Table 3. List of primers ............................................................................................................................... 35 Table 4. Summary of markerless deletion mutant cloning ......................................................................... 42 Table 5. Summary of site-directed mutant primer sets ............................................................................... 44 Table 6. List of CAMPs and antibiotics used in killing assays ................................................................... 54 Table 7. GlcN modification of BP338 lgm locus mutants .......................................................................... 64  xi  List of Figures  Figure 1. Gram-negative bacterial membrane schematic. ............................................................................. 7 Figure 2. General structure of LPS ............................................................................................................... 9 Figure 3. Schematic for the TLR4-MD-2-lipid A interaction. .................................................................... 15 Figure 4. Lipid A structure of B. pertussis strains BP338 and 18-323 ....................................................... 21 Figure 5. hTLR4 activation by BP338 and 18-323 B. pertussis strains ...................................................... 22 Figure 6. Initially identified B. pertussis lgm locus .................................................................................... 24 Figure 7. Predicted topologies of the Lgm proteins .................................................................................... 26 Figure 8. Hypothetical model for lipid A GlcN modification in B. pertussis ............................................. 27 Figure 9. Schematic for the generation of markerless deletion mutants. .................................................... 41 Figure 10. Transcription of lgmABCD in BP338 and 18-323 + pPtacLgmABCD ..................................... 62 Figure 11. Mass spectrometry analysis of BP338 and BP338lgmABCDKO lipid A .................................. 63 Figure 12. hTLR4 activation by BP338 lgm locus mutants ........................................................................ 65 Figure 13. Predicted topologies of BP1945 and LgmE............................................................................... 68 Figure 14. Current schematic of the B. pertussis lgm locus and transcription of lgmE in BP338 .............. 69 Figure 15. Schematic of the lgmE mutant ................................................................................................... 70 Figure 16. Transcription of lgm locus genes in BP338 lgmEKO strains .................................................... 71 Figure 17. hTLR4 activation by BP338 lgmEKO strains ........................................................................... 72 Figure 18. hTLR4 activation by BP338 BP1945KO strains ....................................................................... 74 Figure 19. Expression of LgmA-His in E. coli ........................................................................................... 77 Figure 20. LgmA glycosyltransferase assay ............................................................................................... 78 Figure 21. Identifying conserved residues in LgmA ................................................................................... 80 Figure 22. hTLR4 activation by LgmA mutants ......................................................................................... 81 Figure 23. Phyre predicted model of the structure of LgmA ...................................................................... 83 xii  Figure 24. Identifying conserved residues in LgmC ................................................................................... 86 Figure 25. hTLR4 activation by LgmC mutants ......................................................................................... 87 Figure 26. Phyre predicted model of the structure of LgmC ...................................................................... 89 Figure 27. lgm locus of the sequenced Bordetella species .......................................................................... 93 Figure 28. Comparison of the start of LgmC between Bordetella species ................................................. 94 Figure 29. Neighbour-joining trees of the lgm locus genes in Bordetella species ...................................... 97 Figure 30. Influence of the B. pertussis lipid A GlcN modification on CAMP susceptibility.................. 104 Figure 31. Complementation of BP338LgmABCDKO rescues resistance to polymyxin B .................... 105 Figure 32. Influence of the B. pertussis lipid A GlcN modification on OM stabilization ........................ 107 Figure 33. Genetic analysis of the lgm locus of B. pertussis strains BP338 and 18-323 .......................... 110 Figure 34. Lipid A structures of B. pertussis 18-323 strains complemented with BP338 lipid A-modifying genes ......................................................................................................................................................... 111 Figure 35. Alignment of LpxA from various Gram-negative species....................................................... 113 Figure 36. hTLR4 activation by LPS from B. pertussis 18-323 strains with BP338 lipid A-modification genes ......................................................................................................................................................... 115 Figure 37. Model for how the B. pertussis GlcN modification facilitates CAMP resistance and OM stabilization ............................................................................................................................................... 120 Figure 38. Viability of E. coli R0138 strains complemented with lpxA at 30°C and 42°C ...................... 123 Figure 39. Lipid A structures of E. coli R0138 strains complemented with exogenous lpxA .................. 127 Figure 40. Growth of E. coli R0138 strains with varying acyl chain lengths ........................................... 129 Figure 41. Polymyxin B resistance of E. coli R0138 strains with varying acyl chain lengths, assessed by growth curve assay .................................................................................................................................... 131 Figure 42. Polymyxin B resistance of E. coli R0138 strains with varying acyl chain lengths, assessed by percent survival assay ............................................................................................................................... 132 Figure 43. Activation of hTLR4 by LPS from E. coli R0138 strains with varying acyl chain lengths .... 134 xiii  Figure 44. Activation of the LAL assay by LPS from E. coli R0138 strains with varying acyl chain lengths ....................................................................................................................................................... 136  xiv  List of Abbreviations  ABC: adenosine triphosphate (ATP)-binding cassette ADP: adenosine diphosphate Amp: ampicillin AmpR: ampicillin resistant Ara4FN: N-formylated 4-amino-4-deoxy-L-arabinose Ara4N: 4-amino-4-deoxy-L-arabinose ATP: adenosine triphosphate BG: Bordet-Gengou bp: base pair Bvg: Bordetella virulence gene BvgA-P: activated BvgA C10-OH: hydroxydecanoic acid C12-OH: hydroxydodecanoic chain C14-OH: hydroxymyristoyl chain C16-OH: hydroxypalmitoyl chain C55P: undecaprenyl-phosphate C55PP: undecaprenyl-pyrophosphate CAMP: cationic antimicrobial peptide cAMP: cyclic adenosine monophosphate cfu: colony forming units CP: cytoplasm CyaA: adenylate cyclase dH2O: deionized water xv  DMEM: Dulbecco’s modified Eagle’s medium DNA: deoxyribonucleic acid dpm: degradations per minute EDTA: ethylenediaminetetraacetic acid FHA: filamentous hemagglutinin GalNAc: galactosamine GlcN: glucosamine GlcN[14C]: carbon-14-labeled glucosamine GlcNAc: N-acetylglucosamine G-protein: guanine nucleotide-binding protein h: hour(s) HEPES: 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid His: histidine HRP: horseradish peroxidase hTLR4: human Toll-like receptor 4 IFN-γ: interferon γ IL-1β: interleukin-1β IL-6: interleukin 6 IM: inner membrane IPTG: isopropyl β-D-1-thiogalactopyranoside IRF3: interferon regulatory transcription factor 3 KanR: kanamycin resistant Kdo: 3-deoxy-D-manno-octo-2-ulosonic acid L,D-Hep: L-glycero-D-mannoheptose LAL: limulus amebocyte lysate xvi  LB: Luria-Bertani (lysogeny broth) LPS: lipopolysaccharide MALDI-MS: matrix-assisted laser desorption/ionization mass spectrometry MD-2*: dimerizing MD-2 min: minute(s) mRNA: messenger ribonucleic acid MyD88: myeloid differentiation primary response gene 88 nalR: nalidixic acid resistant NFκB: nuclear factor kappa-B nt: nucleotide OD: optical density OM: outer membrane ORF: open reading frame pBD1: porcine β-defensin 1 PBS: phosphate-buffered saline PCR: polymerase chain reaction Phyre: Protein Homology/analogy Recognition Engine Pi: inorganic phosphate PP: periplasm PRN: pertactin PT: pertussis toxin RLU: relative light units RNA: ribonucleic acid rpm: rotations per minute RT-PCR: reverse transcriptase polymerase chain reaction xvii  s: second(s) SDS-PAGE: sodium dodecyl sulfate polyacrylamide gel electrophoresis SP: surfactant protein sRNA: small ribonucleic acid SS: Stainer-Scholte TCT: tracheal cytotoxin TetR: tetracycline resistant TLR2: Toll-like receptor 2 TLR4*: dimerizing Toll-like receptor 4 (TLR4) TLR4: Toll-like receptor 4 TNFα: tumor necrosis factor-α UDP: uridine diphosphate UMP: uridine monophosphate vag: vir-activated genes vrg: vir-repressed genes w/v: weight per volume xviii  Acknowledgements  First, and foremost, I would like to thank Dr. Rachel C. Fernandez for all her support and advice over the years. Her mentorship has helped transform me from a timid undergraduate student to a confident young researcher – I am truly indebted to her.  I am very grateful to my committee: Dr. Robert E.W. Hancock, Dr. Zakaria Hmama, and Dr. Natalie Strynadka, for their insightful comments and guidance.  Dr. Martine Caroff has been an invaluable collaborator, whose advice and discussions have been greatly appreciated.  I would especially like to express my gratitude to my fellow lab members over the years – their support and sharing in both laughs and commiserations will not be forgotten. Thank you to Dr. Nico Marr, for all your patience when guiding me in my first forays into research. Thank you to Dr. Nina Maeshima, for always taking the time to turn away from your own work to discuss my research. Thank you to the students I supervised: Emma Kim, Rose Lee, and Andrew Low, for your contributions to my thesis work and publications, and to Dr. Jackie Felberg, for setting up the HEK hTLR4 assay in our lab, and Leigh Hobbs, for testing the different LAL assay kits.  I have greatly benefited from the support of my floor and wing mates, with their willingness to share equipment, bacterial strains, and knowledge.  I would also like to thank the Natural Sciences and Engineering Research Council of Canada, the Canadian Institute for Health Research, and the University of British Columbia for their funding. xix   Finally, my heartfelt thanks go to my family and friends, for all their support and understanding.       xx  Dedication   For my parents, L.P.S. and P.C.S., who have always been there for me.  For my brother, N.P.S., who shares my love of science. 1  Chapter 1: Introduction 1.1 Bordetella pertussis and whooping cough Bordetella pertussis is a Gram-negative, aerobic, coccobacillus that was first isolated in 1906 from a human patient, and was identified as the causative agent of whooping cough. This bacterium infects the human respiratory tract and is an obligate strictly human pathogen with no known animal or environmental reservoir (Fedele, Bianco, et al. 2013, Mattoo and Cherry 2005).  Whooping cough is an acute respiratory illness that is a significant cause of infant death worldwide (Mattoo and Cherry 2005). Despite extensive childhood vaccination programs, there are approximately 40 million cases of pertussis every year, 200 000 to 400 000 of which result in death, with most of these deaths occurring in infants (Crowcroft, Stein, et al. 2003, Leung, Robson, et al. 2007). Immunity gained via vaccination or infection is short-lived, thus adults and adolescents are susceptible to infection by B. pertussis and act as a reservoir for this human-restricted pathogen. These individuals are the primary source of transmission to partially vaccinated or unvaccinated infants (Gustafsson, Hessel, et al. 2006, Hallander, Gustafsson, et al. 2005). The ability of B. pertussis to evade and modulate the immune system likely contributes to the persistence of this pathogen in humans (Mattoo and Cherry 2005).  1.1.1 Clinical manifestation The most severe symptoms of whooping cough are often observed in infants, with neonates having the highest risk of death, at 3%. As summarized by Mattoo and Cherry, individuals infected with B. pertussis can present with classical pertussis, mild respiratory symptoms, or be asymptomatic. Classic pertussis illness, known as whooping cough, lasts 6 to 12 weeks, and presents in three stages. After a 7 to 10 day incubation period, the first stage of the disease, the catarrhal stage, lasts 7 to 14 days and initially resembles rhinovirus infection with the patient developing a mild cough. Throughout this stage, the cough worsens until it develops into a paroxysmal cough, which is the defining characteristic of the second stage of disease: the paroxysmal stage. During this period, violent coughing fits are interrupted by sharp intakes 2  of breath, resulting in a whooping sound, for which the disease has been named. The force of these coughing fits often leads to vomiting and consequently weight loss. During this 2 to 6 week stage, there is a risk of complications, such as pneumonia, otitis media, seizures, and encephalopathy. The last stage of the disease (if not leading to fatality), the convalescent stage, presents as a gradual decrease in the severity of the cough, and may last for about 1 to 2 weeks (Mattoo and Cherry 2005). 1.1.2 Molecular pathogenesis B. pertussis causes disease by means of a variety of virulence factors, which are under control of the two component Bordetella virulence gene (Bvg)AS regulatory system (Cummings, Bootsma, et al. 2006). Adhesins are surface proteins that adhere to the airway epithelial cells, thereby enabling the bacterium to remain in the respiratory tract, despite the expulsatory movement of airway cilia (de Gouw, Diavatopoulos, et al. 2011). Toxins are secreted proteins that are required for colonization and persistence in the host. In addition to this, many of these bacterial factors also have immune evasive and immunomodulatory properties (Mattoo and Cherry 2005). 1.1.2.1 BvgAS regulation system The BvgAS two-component signal transduction system consists of BvgS, the sensor kinase, and BvgA, the response regulator. Activation of BvgS by unknown environmental signals triggers autophosphorylation of BvgS, which leads to phosphorylation and activation of BvgA. Activated BvgA (BvgA-P) can then bind to BvgA-specific promoters upstream of virulence genes to activate transcription of these vir-activated genes (vag) (Decker, James, et al. 2012). BvgA-P also activates BvgR, a repressor protein that suppresses expression of a subset of genes (vir-repressed genes, vrg) that are not expressed during the virulence-associated Bvg+ phase (Merkel, Boucher, et al. 2003). In contrast, during the Bvg- phase, BvgAS is not activated; consequently there is no up-regulation of the vag genes, and the vrg genes are expressed. During the Bvgi intermediate phase, some Bvg+ phase genes are expressed, in addition to a subset of exclusively intermediate phase genes, such as bipA. Modulating in vitro growth conditions, by growing B. pertussis at 25°C or in the presence of MgSO4 or nicotinic acid, suppresses the Bvg system, 3  resulting in Bvgi or Bvg- phase growth. However, the Bvg- phase in vivo is associated with survival in the environment by Bordetella bronchiseptica, and since B. pertussis is a human-restricted pathogen, the role of the Bvg- phase in B. pertussis has long thought to be a relic from the ancestral B. bronchiseptica (Cummings, Bootsma, et al. 2006).  1.1.2.2 Adhesins The two major B. pertussis adhesins that facilitate direct binding to human epithelial cells are filamentous hemagglutinin (FHA) and fimbriae (de Gouw, Diavatopoulos, et al. 2011). FHA is a secreted and surface-associated protein that is critical for early colonization. It has four binding domains that facilitate binding to glycolipids on the surface of ciliated and non-ciliated cells of the respiratory epithelium (Alonso, Reveneau, et al. 2002, Mattoo and Cherry 2005). Fimbriae are long, filamentous surface protein structures that consist of a major subunit (Fim2 or Fim3) that makes up the extended structure capped by the tip protein, FimD. The major subunit of fimbriae is known to bind sugars found throughout the mammalian respiratory tract, such as heparin sulfate, and FimD facilitates binding to monocytes (Geuijen, Willems, et al. 1996, Hazenbos, Geuijen, et al. 1995).  1.1.2.3 Toxins B. pertussis possesses numerous toxins, including pertussis toxin (PT), adenylate cyclase (CyaA), and tracheal cytotoxin (TCT) (Mattoo and Cherry 2005).  PT is a secreted, hexameric AB5 toxin, in which the catalytic subunit (PtxA) is an adenosine diphosphate (ADP)-ribosylating toxin, and the B subunits bind the surface of eukaryotic cells. Entry of PT into the host cell results in activation of PtxA, leading to inactivation of guanine nucleotide-binding proteins (G-proteins) and downstream histamine sensitization and suppression of the immune response. CyaA functions as an adenylate cyclase and hemolysin. The N-terminal region of CyaA facilitates entry of the C-terminal region into the host cell, upon which the C-terminal region is activated by the host enzyme calmodulin. Activated CyaA proceeds to convert adenosine triphosphate (ATP) into cyclic adenosine monophosphate (cAMP), leading to an excessive level of the important secondary messenger cAMP in the host cell, resulting in an overall anti-4  inflammatory and antiphagocytic effect (as reviewed in (Carbonetti 2010)). Unlike the protein toxins PT and CyaA, TCT is a disaccharide-tetrapeptide breakdown product of peptidoglycan (Rosenthal, Nogami, et al. 1987). In most Gram-negative bacteria, this product is recycled by the cell via transport back into the cytoplasm by AmpG. However, B. pertussis lacks a functional AmpG, and therefore constitutively releases TCT from the cell (Mielcarek, Debrie, et al. 2006). TCT has toxic effects on host cells, and can result in mitochondrial bloating, disruption of tight junctions, and damage to ciliated cells (Fedele, Bianco, et al. 2013, Mattoo and Cherry 2005). 1.1.2.4 Other virulence factors An important class of B. pertussis virulence factors is the autotransporter proteins, which consists of a C-terminal beta-barrel domain that facilitates the transport of the N-terminal passenger domain across the outer membrane (OM) of the bacterium. The passenger domain is the functional region of the protein, and varies between different autotransporters. Pertactin (PRN), the first identified and characterized autotransporter in Bordetella, is suggested to be involved in adhesion to eukaryotic cells. Two other autotransporter proteins, BrkA and Vag8, are required for resistance against killing by complement (Fernandez and Weiss 1994, Marr, Shah, et al. 2011). Lipopolysaccharide (LPS) also plays a crucial role in B. pertussis infection, and this is discussed in section 1.3. 1.1.2.5 Immune evasion and modulation Many of the virulence factors of B. pertussis have immunomodulatory properties, and can work in synergy with one another. For example, FHA, which functions as the primary adhesin of B. pertussis, also induces apoptosis of human phagocitic and epithelial cells. PT and CyaA assist in this effort to remove phagocytes by intoxicating alveolar macrophages via inactivation of G-proteins and inhibiting monocyte and neutrophil activities by increasing intracellular cAMP, respectively. Additionally, PT inhibits immune cell recruitment and suppresses antibody responses and CyaA promotes apoptosis of phagocytes (as reviewed in (de Gouw, Diavatopoulos, et al. 2011)). The autotransporter Vag8 also plays a role in immune evasion through binding of C1-inhibitor to the surface of the bacterium, thus protecting the 5  pathogen against killing by complement (Marr, Shah, et al. 2011). BrkA, another autotransporter, protects B. pertussis by inhibiting the classical pathway of complement, specifically by preventing deposition of the complement protein C4b on the surface of the bacterium, though the mechanism of this protection is still unknown (Barnes and Weiss 2001).  1.1.3 Other Bordetella species Genus Bordetella is a member of Class Betaproteobacteria and is made up of many host-associated species. To date, there are eight sequenced Bordetella species: B. pertussis, Bordetella parapertussis, B. bronchispetica, Bordetella avium, Bordetella hinzii, Bordetella homesii, Bordetella trematum, and Bordetella petrii (Shah, Moksa, et al. 2013). B. pertussis, B. bronchispetica, and B. parapertussis have been reclassified as “subspecies”, due to the similarity of their genomes (Gerlach, von Wintzingerode, et al. 2001). B. parapertussis is split into two lineages of strains: the human-adapted strains, which only infect humans, and the ovine-adapted strains, which only infect sheep. Human-adapted B. parapertussis is similar to B. pertussis, in that it is a strict human pathogen and causes the whooping cough disease, though B. parapertussis cases of whooping cough tend to be less severe. This may be due to the lack of PT, since B. pertussis is the only species that possesses this toxin. B. bronchiseptica causes kennel cough in small four-legged mammals, and is also able to survive in the environment. B. avium and B. hinzii both infect birds; B. avium is pathogenic whereas B. hinzii is a commensal in the respiratory tracts of fowls, but it can infect immunocompromised humans (Spears, Temple, et al. 2003, Vandamme, Hommez, et al. 1995). B. homesii is associated with septicemia in humans and B. trematum has been isolated from ear infections and skin wounds in humans (Mattoo and Cherry 2005, Weyant, Hollis, et al. 1995). Finally, B. petrii is the only identified environmental species of this genus, and is proposed to be the progenitor of the host-associated Bordetella species (von Wintzingerode, Schattke, et al. 2001). Bordetella ansorpii is the most recent addition to genus Bordetella; it was isolated in 2005 from an immunocompromised patient, though its genome has yet to be sequenced (Ko, Peck, et al. 2005).  6  1.2 Lipopolysaccharide Gram-negative bacterial cells have two membranes that enclose the cytoplasm: the inner membrane (IM) and the outer membrane (OM), both of which contain integral or membrane-associated proteins (Figure 1). The IM surrounds the cytoplasm and consists of a symmetrical lipid bilayer of phospholipids. Peptidoglycan, a sugar-peptide mesh-like molecule that surrounds the cell, is found in the periplasm, the region between the IM and OM. The OM is an asymmetric lipid bilayer, the inner layer of which is composed of phospholipids, whereas the outer layer is made of LPS. Also known as endotoxin, LPS plays an important role in maintaining the integrity of the OM and in the host-pathogen interactions (Needham and Trent 2013, Raetz and Whitfield 2002).   7     Figure 1. Gram-negative bacterial membrane schematic. Gram-negative bacteria have two membranes: the inner membrane (IM) and the outer membrane (OM). Between these two membranes is the periplasm, which contains a thin layer of peptidoglycan. The IM is a symmetric bilayer, consisting primarily of phospholipids. The OM is an asymmetric membrane consisting of phospholipids in the inner leaflet and lipopolysaccharide (LPS) in the outer leaflet    8  1.2.1 LPS structure LPS consists of three domains: lipid A, core sugars, and in some cases, the O-antigen (Figure 2 A). The lipid A region is amphipathic; the acyl chains are hydrophobic and anchor LPS to the OM, whereas the head group is hydrophilic, and interacts with the extracellular milieu. The lipid A domain is linked to the core sugars that extrude from the surface of the cell. Rough bacterial strains are characterized by the lack of O-antigen, and therefore the LPS of these strains consist of only the lipid A and core sugar regions. Smooth bacterial strains, however, have repeating O-antigen subunits linked to the core sugars, though the O-antigen is not added to every LPS molecule in the cell and the length of the O-antigen in most strains varies between different LPS molecules. This results in a heterogeneous mixture of LPS structures in a single bacterium (as reviewed by (Caroff and Karibian 2003)). 9   Figure 2. General structure of LPS A) LPS consists of three main domains: lipid A, the core sugars (separated into the inner core and the outer core), and the O-antigen repeating polysaccharide. The core sugars and O-antigen make up the polysaccharide region. The ‘n’ denotes a number of repeats of the O-antigen subunit. B) Structure of E. coli lipid A, consisting of a di-glucosamine (GlcN) backbone decorated with phosphate groups and 6 acyl chains. Acyl chain length is denoted by the number directly below the chain. Acyl chains are either primary acyl chains, which are attached directly to the di-GlcN backbone, or secondary acyl chains, which are attached to a primary acyl chain. The numbering system of the di-GlcN backbone carbons is in red.    10  1.2.1.1 Lipid A The structure of lipid A varies between different Gram-negative bacterial species. Generally, lipid A consists of a di-glucosamine (GlcN) backbone with four to seven primary and secondary acyl chains linked to the C2, C3, C2’, or C3’ carbons of the di-GlcN backbone (Figure 2B). Primary acyl chains are linked directly to the di-GlcN backbone and secondary acyl chains are connected to a primary acyl chain via an acyl-oxy-acyl linkage. The primary acyl chains at the C2 or C2’ positions are connected to the di-GlcN backbone by amid linkages whereas the C3 or C3’ position primary acyl chains are linked via ester bonds. The basic lipid A structure also has phosphate groups at the C1 and C4’ positions, though these groups can be modified or removed by further modifications to lipid A. At the C6’ position, one or two 3-deoxy-D-manno-octo-2-ulosonic acid (Kdo) sugars link the lipid A domain to the core sugar domain of LPS (as reviewed by (Caroff and Karibian 2003, Raetz and Whitfield 2002)). Some Gram-negative bacteria, such as Aquifex pyrophilus, deviate from this basic structure by using a di-galacturonic acid molecule backbone instead of the more common di-GlcN backbone (Plotz, Lindner, et al. 2000). 1.2.1.2 Core sugars The core sugars of LPS can vary greatly between different bacteria, though the inner core region (the sugars proximal to the lipid A) tends to be conserved between members of a certain genus or family. The inner core often contains Kdo and L-glycero-D-mannoheptose (L,D-Hep) sugars, whereas the sugars of outer core (distal to lipid A) are quite variable, likely due to selective pressures from the environment. The outer core region acts as a linker between the core polysaccharide and the O-antigen in smooth Gram-negative species (Raetz and Whitfield 2002). 1.2.1.3 O-antigen In addition core sugars, smooth bacterial strains also have an O-antigen polysaccharide linked to the core polysaccharide of a fraction of the LPS molecules in the OM. The O-antigen is the most diverse region of LPS, and often varies between strains of a single species. It consists of a linear or branched repeating 11  pattern of sugars that varies in length between different LPS molecules in a single bacterium (Raetz and Whitfield 2002). 1.2.2 LPS Biogenesis LPS is synthesized at the IM, and then transported to the outer leaflet of the OM. The first step is biosynthesis of lipid A-Kdo via the Raetz lipid A biosynthesis pathway, followed by a step-wise process of transfer and modification of the core sugars by individual glycosyltransferases directly onto lipid A-Kdo to produce lipid A-core. Synthesis of the lipid A-core molecule occurs at the cytoplasmic side of the IM, after which it is flipped to the periplasmic face of the IM. In smooth bacterial strains, the O-antigen is also built at the cytoplasmic side of the IM, then translocated to the periplasmic side of the IM, before ligation to produce the completed smooth LPS molecule, before transport to the surface (Raetz and Whitfield 2002). 1.2.2.1 Raetz lipid A biosynthesis pathway The first stage of LPS biosynthesis is the Raetz lipid A biosynthesis pathway, which results in the production of lipid A-Kdo, take place in the cytoplasm and the inner leaflet of the IM, as reviewed by Raetz et al. (Raetz, Guan, et al. 2009). The enzymes of this pathway (Lpx proteins) are highly conserved across most Gram-negative bacterial species, though some species have variations in these Lpx proteins or additional modification enzymes that allow for atypical lipid A structures. In the general Raetz pathway, which was characterized in Escherichia coli, the first step is transfer of β-hydroxymyristoyl chain (C14-OH) from an acyl carrier protein to the C3 position of uridine diphosphate (UDP)-N-acetylglucosamine (GlcNAc) by LpxA (Anderson and Raetz 1987). LpxA is a homotrimer with an active site located between two monomeric subunits (Robins, Williams, et al. 2009). Orthologs of LpxA in other bacterial species can add acyl chains of different lengths to UDP-GlcNAc, based on the structure of the hydrocarbon ruler region of the LpxA active site (Shah, Albitar-Nehme, et al. 2013). LpxC irreversibly deacetylates the product of LpxA, thereby committing UDP-3-O-(acyl)-GlcN to the Raetz pathway, and this is followed by transfer of another C14-OH to the C2 position by LpxD, resulting in UDP-2,3-diacyl-12  GlcN (Kelly, Stachula, et al. 1993, Raetz, Guan, et al. 2009). LpxH then cleaves UDP, to produce lipid X (2,3-diacyl-GlcN-1-phosphate), though some bacterial species lack LpxH, and instead the enzyme LpxI generates lipid X from UDP-2,3-diacyl-GlcN (Metzger and Raetz 2010). Then one molecule of UDP-2,3-diacyl-GlcN and one molecule of lipid X are joined by LpxB to form a β,1’-6 linkage (Raetz, Guan, et al. 2009). LpxK subsequently phosphorylates the 4’ position of this molecule, to generate lipid IVA, a tetra-acyl precursor of lipid A (Ray and Raetz 1987). In E. coli, WaaA (formerly KdtA) adds two Kdo sugars to the 6’ position of lipid IVA, followed by addition of the secondary acyl chains by LpxM and LpxL to produce hexa-acyl Kdo2-lipid A. LpxM adds the C12-OH secondary acyl chain at the C2’ position and LpxL adds the C14-OH secondary acyl chain at the C3’ position. Bacteria lacking either LpxM or LpxL often generate penta-acyl lipid A, if they lack additional acyl-chain modifying enzymes (Raetz, Guan, et al. 2009). 1.2.2.2 Synthesis of the core sugars Core sugars are added to Kdo2-lipid A by a set of sequential glycosyltransferases at the cytosolic face of the IM. For example, in E. coli strains with an R1 LPS core structure, L,D-Hep is transferred to the Kdo proximal to the lipid A by the glycosyltransferase WaaC, followed by transfer of a second L,D-Hep by another glycosyltransferase WaaF onto the first L,D-Hep (Clementz and Raetz 1991, Gronow, Brabetz, et al. 2000). Successive glycosyltransferases and modifying enzymes build the remaining sugars in the R1 core, resulting in a molecule consisting of lipid A and the core sugars (Raetz and Whitfield 2002). Lipid A-core is then flipped to the periplasmic face of the IM by the ATP-binding cassette (ABC) transporter MsbA, upon which either the lipid A-core is directly transported to the OM via the Lpt system or the O-antigen is ligated to the outer core polysaccharide and then transported to the OM (Sperandeo, Deho, et al. 2009).  1.2.2.3 O-antigen synthesis and ligation The repeating sugar subunits of the O-antigen polysaccharide are synthesized on the IM carrier lipid undecaprenyl-phosphate (C55P) before the fully synthesized O-antigen is ligated to the distal end of 13  rough LPS via WaaL on the periplasmic face of the IM. Most bacteria have one of three mechanisms for synthesizing the O-antigen: the Wzy-dependent pathway, the ABC-transporter-dependent pathway, or the synthase-dependent pathway. In all three pathways O-antigen synthesis is initiated by transfer of a sugar-1-phosphate to C55P to generate undecaprenyl-pyrophosphate (C55PP)-sugar by a transferase enzyme, such as WecA in Salmonella enterica serovar Typhimurium. This is followed by synthesis of the long, repeating sugar chain on C55PP, and flipping of C55PP-O-antigen to the periplasmic face of the IM, allowing WaaL to transfer the O-antigen from C55PP to the rough LPS to generate a smooth LPS molecule (as reviewed by (Raetz and Whitfield 2002)). 1.2.2.4 Transport of LPS to the outer membrane Completion of LPS biosynthesis and MsbA activity results in LPS embedded in the periplasmic side of the IM by the hydrophobic acyl chains of lipid A. The Lpt system functions to transport this amphipathic molecule from the outer leaflet of the IM, across the periplasm and OM, to the outer leaflet of the OM. The current model for transport of LPS to the bacterial surface by the Lpt system is as follows: first, the ABC transporter complex, consisting of LptBCFG, extracts LPS from the periplasmic face of the IM and transfer it to LptA, which binds LPS (Chng, Gronenberg, et al. 2010, Ruiz, Gronenberg, et al. 2008). LptA is hypothesized to shield the hydrophobic acyl chains of LPS within a hydrophobic pocket during transport across the hydrophilic periplasmic environment, followed by transfer to the periplasmic OM lipoprotein LptE (Tran, Trent, et al. 2008). LptE is proposed to plug the pore of the β-barrel, integral OM protein LptD, thereby allowing the LptE-LptD complex to insert LPS directly into the outer leaflet of the OM through an opening in the LptD β-barrel wall (Freinkman, Chng, et al. 2011). 1.2.3 Biological role of LPS LPS makes up the outer most layer of the bacterial membrane, and therefore acts as the first barrier between the bacterium and the extracellular environment (Needham and Trent 2013). Furthermore, LPS plays an important role in the interaction between the bacterium and the host: the polysaccharide region 14  extruding from the bacterial surface interacts with host factors and the lipid A region affects the immune system via activation of Toll-like receptor 4 (TLR4) (Caroff and Karibian 2003). 1.2.3.1 LPS and membrane integrity The barrier created by LPS consists of tightly packed acyl chains inside the OM and di-GlcN backbone with negatively-charged phosphate groups making up the outer surface of the membrane. These negatively-charged phosphate groups are bridged by divalent cations, which also partially neutralize the overall negative charge of the OM surface, thereby increasing the integrity of the membrane. The barrier function of LPS also protects the bacterium from cationic antimicrobial peptides (CAMPs), which are found in both the environment, when produced by environmental bacteria, and in the host (Hancock 1997). CAMPs gain entry into bacterial cells by displacing the bridging cations and interacting with the negatively-charged phosphates of LPS, leading to self-promoted uptake of the CAMPs into the cell. Some bacteria are able to increase resistance to CAMPs by modifying the structure of lipid A (Hancock 1997, Raetz, Reynolds, et al. 2007).  1.2.3.2 Host-pathogen interaction Both major domains of LPS, the polysaccharide region and the lipid A region, are involved in the host-pathogen interplay. The polysaccharide region, especially the O-antigen that extrudes from the cell in smooth bacterial strains, is highly antigenic, though these long chains also protect bacteria from numerous factors, such as complement and antibiotics (Caroff and Karibian 2003). Furthermore the O-antigen of some bacterial species aid infection by adhering to mammalian tissues, as found in the LPS of Actinobacillus pleuropneumoniae (Boekema, Stockhofe-Zurwieden, et al. 2003).    15     Figure 3. Schematic for the TLR4-MD-2-lipid A interaction. Interaction with LPS (lipid A is shown in red for simplicity) promotes formation of the TLR4-MD-2 heterodimer (TLR4 and MD-2 shown in blue). Interaction between one TLR4-MD-2 heterodimer (blue) and another TLR4*-MD-2* heterodimer (orange) is mediated by interactions between lipid A bound to MD-2 (blue) and TLR4* (orange). This TLR4 dimerization leads to activation and downstream production of proinflammatory cytokines.     16  The lipid A region of LPS is recognized by the TLR4-MD-2 heterodimer that is found on many cell types, including macrophages and dendritic cells (Trent, Stead, et al. 2006). The acyl chains of lipid A sit in the hydrophobic pocket of MD-2 and the di-GlcN backbone sits at the top of the MD-2 pocket where the phosphate groups of lipid A interact with both TLR4 and TLR4* (the incoming, dimerizing TLR4 unit). This interaction results in dimerization and activation of TLR4 (Figure 3), which leads to the initiation of signaling cascades (Park, Song, et al. 2009). If TLR4 dimerizes at the surface of the host cell, signaling via the myeloid differentiation primary response gene 88 (MyD88)-dependent pathway results in early activation of the transcription factor nuclear factor kappa-B (NFκB) and production of pro-inflammatory cytokines, such as tumor necrosis factor-α (TNFα) and interleukin-1β (IL-1β). Internalization of TLR4 into an endosome results in signaling via the MyD88-independent pathway, which results in activation of interferon regulatory transcription factor 3 (IRF3) and the production of type I interferons, along with “late” activation of NFκB (Kagan, Su, et al. 2008, Maeshima and Fernandez 2013). Activation of TLR4 also leads to downstream production of costimulatory molecules that are required for the generation of an adaptive immune response (Raetz and Whitfield 2002). 1.2.4 LPS modification systems Since the structure of LPS can affect its numerous biological roles, many Gram-negative bacteria have developed mechanisms to modify LPS structure to adapt to environmental stresses. Enzymes that facilitate LPS modification generally function at the periplasmic side of the IM, thereby modifying LPS before transport across the periplasm, or at the OM, thus modifying LPS that is already present at the surface (Needham and Trent 2013). Modification of the lipid A region, including the acyl chains and the phosphate groups, can affect CAMP resistance and TLR4 activation (Raetz, Reynolds, et al. 2007). Altering the structure of the polysaccharide region can also affect interactions with the immune system, such as increasing resistance to complement-mediated bacterial killing (Needham and Trent 2013). LPS modification systems do not modify every molecule of LPS within a bacterium; therefore, they generate heterogeneity in the LPS of a single bacterium (Caroff and Karibian 2003). 17  1.2.4.1 Acyl chain modifications Though unmodified E. coli LPS is hexa-acylated (Figure 2), the lack of secondary acyltransferases (LpxL and LpxM) and the presence of acyl chain modifying enzymes in E. coli and other bacterial species can result in hepta-, penta-, or tetra-acyl LPS molecules (Trent, Stead, et al. 2006). Importantly, the number and position of the acyl chains in LPS affect the ability of LPS to activate TLR4, likely due to the position of these differently shaped LPS molecules in the MD-2-TLR4 binding pocket (Figure 3) (Needham, Carroll, et al. 2013). In Salmonella, PagP transfers C16-OH secondary acyl chain to the C2 position, resulting in hepta-acyl LPS molecules. This modification increases resistance to CAMPs, perhaps by increasing OM integrity due to more tightly packed LPS in the membrane (Guo, Lim, et al. 1998, Trent, Stead, et al. 2006). PagP-modification of Salmonella LPS also decreases activation of TLR4 (Kawasaki, Ernst, et al. 2004). However, PagP plays a slightly different role in B. bronchiseptica, since it incorporates the secondary C16-OH chain at the C3’ position, resulting in hexa-acyl lipid A. This modification in B. bronchiseptica does not affect CAMP resistance, though it is required for persistent infection of the mouse respiratory tract (Preston, Maxim, et al. 2003). In contrast, PagL removes the C3 position primary acyl chain, and in Salmonella, this modification does not affect resistance to CAMPs, but does attenuate activation of TLR4 (Trent, Pabich, et al. 2001, Trent, Stead, et al. 2006). LpxO also modifies acyl chains, but instead of adding or removing them, LpxO hydroxylates the acyl chain and is hypothesized to be involved in coordination of the stress response (Needham and Trent 2013) . 1.2.4.2 Phosphate modifications The phosphate groups of lipid A (Figure 2) are important for both susceptibility to CAMPs, since they lend to the overall negative charge of the OM, and for activation of TLR4, as they interact with key positive residues in TLR4 and TLR4* to promote dimerization. Thus, bacterial modification of these phosphate groups can act as a mechanism to evade the host immune system by decreasing susceptibility to CAMPs and modulating activation of TLR4 (Maeshima and Fernandez 2013, Needham and Trent 2013). For example, Francisella tularensis LpxE and LpxF remove the phosphate groups at the 1 and 4’ 18  positions, respectively, although LpxF functions only in the absence of the C3’ secondary acyl chain (Wang, Karbarz, et al. 2004, Wang, McGrath, et al. 2006). Removal of the 4’-phosphate group in F. tularensis results in increased resistance to CAMPs, but no difference in the ability of lipid A to activate TLR4. Interestingly, removal of the 1-phosphate in Salmonella that expresses LpxE results in a monophosphoryl lipid A species that has attenuated TLR4 activation (Raetz, Reynolds, et al. 2007).  In Salmonella and E. coli, enzymes encoded by the Arn locus function to modify the 4’-phosphate of lipid A with a 4-amino-4-deoxy-L-arabinose (Ara4N) moiety, resulting in greater resistance to the CAMP polymyxin B, though no reported difference in TLR4 activation (Gunn, Lim, et al. 1998, Trent, Stead, et al. 2006). The increase in CAMP resistance is thought to be due to a decrease in the overall negative charge of the OM by addition of the positively charged Ara4N group. This, hypothetically, results in less interaction of the LPS with the positively charged CAMPs, thereby increasing CAMP resistance (Needham and Trent 2013). The Arn pathway has been studied in E. coli and Salmonella to determine the mechanism by which the lipid A is modified with Ara4N (as reviewed by (Raetz, Reynolds, et al. 2007)). ArnA and ArnB initiate the pathway by synthesizing N-formylated UDP-Ara4N (UDP-Ara4FN) in the cytoplasm (Breazeale, Ribeiro, et al. 2005). This is followed by transfer of Ara4FN to C55P, the IM carrier lipid, by the GT2 family glycosyltransferase ArnC and then removal of the formate group by ArnD, to generate C55P-Ara4N (Breazeale, Ribeiro, et al. 2005). Next, ArnE and ArnF, two short proteins that each consist of 4 transmembrane helices, function as a flippase to transport C55P-Ara4N from the cytoplasmic face to the periplasmic face of the IM (Yan, Guan, et al. 2007). Finally ArnT, a member of the GT83 family of glycosyltransferases, adds Ara4N to the 4’-phosphate group of LPS (Trent, Ribeiro, et al. 2001). An ArnT homolog in Pseudomonas aeruginosa also modifies LPS with Ara4N, though in this bacterium both the 1- and 4’-phophates are generally modified and this modification increases both CAMP resistance and the inflammatory response, likely by increasing TLR4 activation (Gellatly, Needham, et al. 2012, Moskowitz, Ernst, et al. 2004). Thus, as seen with the PagP 19  acyl chain modification, homologous enzymes can modify lipid A in slightly different ways and result in different biological effects. 1.2.4.3 Core sugars and O-antigen modifications The polysaccharide region of LPS can also be modified in a variety of manners, such as addition of sugar, phosphate, and phosphoethanolamine groups, by the bacterium to adapt to different environmental conditions (Needham and Trent 2013). One such example is the addition of sialic acid and glucose to the core sugars of Neisseria gonorrhoeae LPS. This increases the affinity of the LPS for the complement regulatory proteins Factor H and complement component 4b (C4b)-binding protein, thus leading to increased resistance to complement-mediated killing (Needham and Trent 2013, Ram, Ngampasutadol, et al. 2007).  1.3 B. pertussis LPS LPS plays an important role in B. pertussis infection. As mentioned in Section 1.2.3, LPS is crucial to the structural integrity of the OM and interacts with the host immune system via activation of TLR4, which leads to downstream release of proinflammatory cytokines. Activation of human TLR4 (hTLR4) by penta-acyl B. pertussis LPS from strain Tohama I is not as robust as activation by hexa-acyl E. coli LPS, but still results in the production of proinflammatory cytokines such as TNFα, interleukin 6 (IL-6), and interferon γ (IFN-γ) (Marr, Novikov, et al. 2010). In addition, B. pertussis LPS acts as an antigen during natural infection, and protects the bacterium from surfactant proteins (SPs) A and D (Schaeffer, McCormack, et al. 2004, Trollfors, Lagergard, et al. 2001). SPA and SPD are lipid-binding lectins that are expressed in the human lower respiratory tract and damage bacterial membranes by binding to LPS (Chaby, Garcia-Verdugo, et al. 2005). The trisaccharide domain of B. pertussis LPS (see Section 1.3.1) prevents binding of SPA and SPD, likely through steric hindrance, therefore protecting the bacterium from this host clearance mechanism (Fedele, Bianco, et al. 2013, Schaeffer, McCormack, et al. 2004). 20  1.3.1 B. pertussis LPS structure B. pertussis LPS consists of three regions: lipid A, the core polysaccharide, and a trisaccharide, which is attached to the distal end of the core polysaccharide in a fraction of the LPS molecules. Therefore, B. pertussis LPS separates into two bands by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) analysis: a faster-migrating band B, which consists of only the lipid A and core sugars, and band A, which is band B plus the trisaccharide. The branched core polysaccharide contains 10 sugar subunits, including only one Kdo sugar that links the core sugars to the lipid A domain (Caroff, Brisson, et al. 2000). Though the lipid A structure can vary between different strains of B. pertussis, in general, it has a di-GlcN backbone decorated with a phosphate group at the 1 and 4’ positions, and 5 acyl chains. A characteristic feature of B. pertussis lipid A is a short C10-OH primary acyl chain at the C3 position (Caroff, Brisson, et al. 2000, Marr, Novikov, et al. 2010). 1.3.2 Lipid A structure of strains BP338 and 18-323 The lipid A of B. pertussis wild-type strain BP338, a nalidixic acid-resistant derivative of strain Tohama I, consists of a di-GlcN backbone and generally 5 acyl chains: three C14-OH primary acyl chains at positions C2, C2’, and C3’, one C10-OH primary acyl chain at position C3, and a C14 secondary acyl chain at the C2’ position (Figure 4, left). Furthermore, both phosphate groups may be modified with GlcN moieties, such that the bacterium contains LPS species with no GlcN modification, one GlcN modification at either the 1- or 4’-phosphate, or GlcN modification at both phosphate groups (Marr, Tirsoaga, et al. 2008). However, the lipid A structure of B. pertussis wild-type strain 18-323 differs from that of strain BP338 in two ways: 1) 18-323 does not have GlcN-modified lipid A, and 2)18-323 lipid A has C10-OH or C12-OH acyl chains at the C3’ position, which are shorter than the C14-OH acyl chain at this position in BP338 (Figure 4) (Marr, Novikov, et al. 2010). These differences in lipid A structure are likely responsible for the greater ability of BP338 LPS to activate hTLR4 compared to LPS from the 18-323 strain (Figure 5) (Marr, Novikov, et al. 2010).   21   Figure 4. Lipid A structure of B. pertussis strains BP338 and 18-323 Both lipid A structures consists of a di-GlcN backbone decorated with 1- and 4’-phosphates and five acyl chains. BP338 lipid A (left) has GlcN modifications at the phosphate groups, circled in red, and a 14-carbon C3’ acyl chain (green rectangle). 18-323 lipid A (right) has a 10- or 12-carbon C3’ acyl chain (green rectangle). Acyl chain length is denoted by the number directly below or next to the chain. (Marr, Novikov, et al. 2010)    22    Figure 5. hTLR4 activation by BP338 and 18-323 B. pertussis strains From Marr et al. 2010 (Marr, Novikov, et al. 2010), used with permission. Relative light units (RLUs; firefly luciferase/Renilla luciferase) as a readout of nuclear factor-kB activation by transfected HEK-293 cells after 4 h of stimulation with heat-killed B. pertussis BP338, its isogenic mutant BP338GlcN−, mouse challenge strain 18-323 (multiplicity of infection, 75), or the synthetic triacylated lipopeptide and Toll-like receptor 2 (TLR2) agonist Pam3CSK4 (1 mg/mL). n.s., not significant. ***, P < 0.001.     23  1.3.3 GlcN modification and the lgm locus A transposon mutant of B. pertussis strain BP338 that lacks the lipid A GlcN modification has decreased levels of hTLR4 activation compared to the wild type BP338, though this GlcN-negative strain activates hTLR4 at a higher level than that of 18-323 (Figure 5) (Marr, Hajjar, et al. 2010). Interestingly, no difference is observed in the ability to activate murine TLR4 between B. pertussis BP338 strains that do or do not exhibit GlcN-modified lipid A (Marr, Hajjar, et al. 2010). The transposon in the BP338 mutant that lacks the GlcN modification disrupts the BP0398 locus tag (lgmB).  Upstream of BP0398 is BP0399 (lgmA), and these two genes have overlapping start and stop codons, suggesting they may function together (Figure 6) (Marr, Tirsoaga, et al. 2008). Directly downstream of this gene pair is another set of two genes with overlapping start and stop codons: BP0397 (lgmC) and BP0396 (lgmD) (Marr, Tirsoaga, et al. 2008, Shah, Albitar-Nehme, et al. 2013). Thus, the lgm locus was originally predicted to encode the four genes: lgmA, lgmB, lgmC, and lgmD.   24   Figure 6. Initially identified B. pertussis lgm locus The originally described B. pertussis lgm locus consists of lgmA (1059 bp), lgmB (1620 bp), lgmC (1104 bp), and lgmD (456 bp), formerly locus tags BP0399 to BP0396, respectively. The intergenic region between lgmB and lgmC is 188 bp.    25   1.3.4 Bioinformatic analysis of the lgm locus and a hypothetical model The lipid A GlcN modification in B. pertussis strain BP338 is abrogated in the lgmB transposon mutant, suggesting LgmB is required for this GlcN modification. LgmB is a predicted homolog of Salmonella and E. coli glycosyltransferase ArnT, which modifies lipid A phosphate groups with Ara4N. The start codon of lgmB overlaps with stop codon of lgmA, a predicted homolog of ArnC, another glycosyltransferase in the lipid A Ara4N-modification pathway of Salmonella and E. coli (Marr, Tirsoaga, et al. 2008). The downstream gene lgmC encodes a member of the YdjC-like pfam protein family, and members of this family are involved in cleavage of cellobiose-phosphate. TTHB029, a member of the YdjC-like protein family, shows structural similarities with SpPgdA, a peptidoglycan GlcNAc deacetylase from Streptococcus pneumonia (Imagawa, Iino, et al. 2008).  Lastly, lgmD is a small gene predicted to encode a protein composed primarily of four transmembrane helices (Figure 7D), similar to the structures of ArnE and ArnF, which form the C55P-Ara4N flippase in the Salmonella lipid A modification pathway (Shah, Albitar-Nehme, et al. 2013, Yan, Guan, et al. 2007).   Based on these observations, the lipid A GlcN modification (lgm) locus, consisting of lgmA, lgmB, lgmC, and lgmD (formerly locus tags BP0399 to BP0396, respectively), is proposed to be required for GlcN modification of lipid A in B. pertussis strain BP338 (Marr, Tirsoaga, et al. 2008). Furthermore, these Lgm proteins are predicted to function in the pathway summarized in Figure 8: LgmA transfers GlcNAc from UDP-GlcNAc to the IM carrier lipid C55P, then LgmC removes the acetyl group, followed by flipping of C55P-GlcN from the cytoplasmic to the periplasmic face of the IM by LgmD, and finally LgmB transfers GlcN from C55P to the phosphate of lipid A (Marr, Tirsoaga, et al. 2008). However, though previous data suggest lgmB is required for lipid A GlcN modification in B. pertussis, the involvement of the other lgm genes was still unknown at the beginning of this project, as was the function of the Lgm proteins.  26   Figure 7. Predicted topologies of the Lgm proteins A) LgmA, B) LgmB, C) LgmC, D) LgmD. Membrane topologies were predicted with TMHMM (Sonnhammer, von Heijne, et al. 1998). The amino acid position is along the x-axis, the probability of a transmembrane helix is along the y-axis (red vertical bars). For the horizontal lines: thick red line is predicted transmembrane region, blue is predicted cytoplasmic region, pink is predicted periplasmic region. The diagrams on the right depict a visual representation of the TMHMM prediction. The upper horizontal black line represents the periplasmic leaflet of the IM, the lower black line represents the cytoplasmic leaflet of the IM, and the coloured line represents the protein. For LgmC (C), since no transmembrane helices are predicted, the prediction by TMHMM of LgmC being placed in the periplasm or cytoplasm is not accurate. 27    Figure 8. Hypothetical model for lipid A GlcN modification in B. pertussis In this proposed model for the Lgm pathway in B. pertussis, LgmA transfers GlcNAc from UDP-GlcNAc to the IM carrier lipid C55P, resulting in the byproducts UMP and Pi. Then LgmC deacetylates C55P-GlcNAc, to generate C55P-GlcN, which is flipped to the periplasmic side of the IM by LgmD, followed by transfer of GlcN from C55P to lipid A by LgmB. PP, periplasm; IM, inner membrane; CP, cytoplasm; C55P, undecaprenyl phosphate; GlcNAc, N-acetyl-glucosamine; UMP, uridine monophosphate; Pi, inorganic phosphate; GlcN, glucosamine.    28  1.4 Hypotheses Based on the bioinformatics analysis of the lgm locus, I hypothesize that all four genes of the lgm locus, lgmA, lgmB, lgmC, and lgmD, are required for GlcN modification of B. pertussis lipid A. Furthermore, since LgmA is a homolog of the GT2 glycosyltransferase ArnC, which transfers Ara4FN onto C55P, I predict that LgmA functions to transfer GlcNAc from UDP-GlcNAc to C55P (Figure 8) (Breazeale, Ribeiro, et al. 2005).  Modification of lipid A structures, especially with charged groups, often affect resistance to CAMPs (Needham and Trent 2013). Therefore, I hypothesize that the addition of the positively-charged GlcN moiety to the negatively-charged phosphate groups of B. pertussis lipid A will affect resistance to a variety of CAMPs. Since the negative charge of the phosphate groups is important for stabilizing the OM via coordination of cations (Needham and Trent 2013), I also predict these GlcN modification will affect OM stabilization in B. pertussis.  The lipid A structures of two B. pertussis strains, BP338 and 18-323, differ in two ways: the presence of the GlcN modification on the phosphate groups, and the length of the C3’ acyl chain (Figure 4). Furthermore, the ability of these strains to activate hTLR4 greatly varies: BP338 induces moderate levels of hTLR4 activation whereas 18-323 activates hTLR4 at very low levels. A lack of GlcN modification in BP338 LPS results in a reduction in hTLR4 activation, however, this level of activation is still higher compared to 18-323 LPS (Marr, Novikov, et al. 2010). I therefore hypothesize that each difference in lipid A structure between B. pertussis BP338 and 18-323 strains individually affects hTLR4 activation.   Many Gram-negative bacteria, such as Salmonella species and E. coli, have hexa-acyl LPS. Removal and addition of acyl chains from hexa-acyl LPS can decrease TLR4 activation by these molecules and affect resistance to CAMPs (Trent, Stead, et al. 2006). However the effect of more subtle changes, such as 29  minor changes to the lengths of the acyl chains, have not been studied in species with hexa-acyl LPS. I hypothesize that gradually decreasing the length of the C3 and C3’ C14-OH acyl chains in E. coli hexa-acyl LPS will affect many biological characteristics of the bacterium, including resistance to CAMPs and hTLR4 activation.  1.5 Thesis goals A primary goal of this thesis is to characterize the lgm locus genes, which are the genes required for the modification of B. pertussis lipid A with GlcN. lgmA, lgmB, lgmC, and lgmD (locus tags BP0399 to BP0396, respectively) have been proposed to be required for this modification based on transposon mutagenesis and bioinformatics analysis (Marr, Tirsoaga, et al. 2008). Furthermore, these analyses suggest a hypothetical pathway for GlcN modification of lipid A (summarized in Figure 8) in which LgmA is proposed to transfer GlcNAc from UDP-GlcNAc to the carrier lipid C55P (Marr, Tirsoaga, et al. 2008). Therefore, demonstrating the function of LgmA could support or contradict this hypothetical model. To further characterize the first two enzymes of this model, mutational analysis of LgmA and LgmC, the proposed C55P-GlcNAc deacetylase, could identify functionally important amino acids in these two enzymes, and perhaps describe a putative active site region.  Analysis of the lgm locus in other Bordetella species may also highlight features of this locus, since many Bordetella species modify lipid A with GlcN. B. hinzii and B. trematum, however, have unmodified lipid A (Novikov, Shah, et al. 2013). Determining the genetic basis for the lack of GlcN modification in these two species may reveal clues as to the requirements for a functional Lgm pathway.  Another significant goal of this thesis is to characterize the biological effects of lipid A modifications in B. pertussis, such as the effect on CAMP resistance and membrane stability. Additionally, an important effect of lipid A modifications is the ability to influence activation of hTLR4 (Maeshima and Fernandez 30  2013, Needham and Trent 2013). The differences between the lipid A structures of the two B. pertussis strains BP338 and 18-323 provides a suitable model to dissect the relative effect of two different lipid A structural modifications on hTLR4 activation: the length of the C3’ acyl chain and the GlcN modification.   The final goal of this thesis is to characterize the biological effects of gradually shortening the C3 and C3’ acyl chain in hexa-acyl E. coli. Some attributes to explore are bacterial growth, resistance to polymyxin B, and activation of hTLR4 and the Limulus amebocyte lysate (LAL) assay by LPS that contains these different lipid A structures.     31  Chapter 2: Materials and methods 2.1 Bacterial growth conditions B. pertussis strains were grown on BG agar supplemented with 15% defibrinated sheep’s blood (Dalynn) at 37 °C or in Stainer-Scholte (SS) broth with 0.06% bovine serum albumin (Sigma-Aldrich) shaking at 180 rpm at 37°C (Stainer and Scholte 1970). All BP338 strain derivatives were grown in the presence of 30 µg/ml nalidixic acid and wild type strain 18-323 was grown in the absence of antibiotics, unless otherwise stated.  All E. coli strains were grown in LB broth shaking at 200 rpm or on LB agar with appropriate antibiotics. E. coli strain R0138 was grown on LB agar at 30°C or in LB broth at 30°C or 42°C, as specified, and E. coli strains DH5α, S17-1, and BL-21 were grown at 37°C. R0138 strains were grown in media supplemented with 12 µg/ml tetracycline and R0138 strains with pBlueScript II KS (-) derivatives were grown with 1 mM Isopropyl β-D-1-thiogalactopyranoside (IPTG) when grown at 42°C.   B. pertussis and E. coli strains containing derivatives of pMMB67HE or pBlueScript II KS (-) were grown with 100 µg/ml ampicillin (Amp), strains containing derivatives of pBBR1MCS were grown with 34 µg/ml chloramphenicol, and strains containing derivatives of pSS4245, pET30b, or pBBR1MCS2 were grown with 50 µg/ml kanamycin.  2.2  Strains, plasmids, and primers 2.2.1 Bacterial strains B. pertussis wild-type strains BP338 and 18-323, or mutants of these strains, were used in all experiments. B. pertussis strain BP338 is a nalidixic acid-resistant derivative of wild type strain Tohama I.  32  E. coli strain DH5α (Invitrogen) was used for cloning, plasmid maintenance, and transformation. E. coli strain DH5α has the following genotypic features: endA mutation eliminates non-specific endonuclease I activity, therefore allowing improved plasmid maintenance and preparations; hsd mutation which inactivates the methylation and restriction system that allows E. coli to recognize foreign deoxyribonucleic acid (DNA), therefore allowing transformation with polymerase chain reaction (PCR)-generated DNA. E. coli strain S17-1, which contains a chromosomally-integrated transfer (tra) locus, was used for conjugation. E. coli strain BL-21 was used for protein expression from pET30 vectors, since it contains a deletion of lon and a mutation in ompT, both of which result in decreased proteolysis, and contains the DE3 prophage, which encodes T7 ribonucleic acid (RNA) polymerase that is required for expression from T7 promoters, as found in pET30 vectors. The E. coli strain R0138 (Galloway and Raetz 1990) has lpxA under the control of a temperature sensitive promoter (lpxA2), such that the endogenous LpxA protein is expressed at low temperatures (i.e. 30°C), but is not expressed at higher temperatures (i.e. 42°C). The recA mutation in R1038 protects against reversion of lpxA2 to wild type by preventing DNA recombination.   33  Table 1. List of bacterial strains Strain Description Source      BP338 nalR derivative of B. pertussis Tohama I wild-type strain A. Weiss      BP338LgmABCDKO ΔlgmABCD This thesis      BP338LgmAKO ΔlgmA This thesis      BP338LgmBKO ΔlgmB This thesis      BP338LgmCKO ΔlgmC This thesis      BP338LgmDKO ΔlgmD This thesis      BP338BP1945KO ΔBP1945 This thesis      BP338LgmDKOBP1945KO ΔlgmD ΔBP1945 This thesis      18-323 B. pertussis wild type strain A. Weiss      B. avium strain ATCC 5086 Wild type strain M. Caroff      B. hinzii strain ATCC 51730 Wild type strain M. Caroff      B. trematum strain CCUG 13902 Wild type strain M. Caroff      E. coli R0138 TetR, lpxA2 recA C. Raetz      E. coli BL-21 DE3 Δlon ompT  Novagen      E. coli DH5α endA hsd Invitrogen      E. coli S17-1  (Simon, Priefer, et al. 1983)  2.2.2 Plasmids The pSS4245 vector was a generous gift from Scott Stibitz. pSS4245 contains the following features that allow for the generation of markerless deletion mutants in B. pertussis strains: plasmid-encoded I-SceI restriction endonuclease under the control of a Bvg-regulated promoter, such that I-SceI is only expressed under Bvg+ conditions (i.e. no supplementation of the media with MgSO4), an I-SceI enzyme cut site, kanamycin resistance cassette, and a streptomycin resistance cassette that provides resistance in B. pertussis, but not in E. coli strain S17-1.   34  Table 2. List of plasmids  Cloning details are in section 2.3. Plasmids Description Source pSS4245 Suicide vector (replicates in E. coli, but not B. pertussis), KanR (Inatsuka, Xu, et al. 2010) pSS4245LgmAKO pSS4245 containing the BP338 lgmA upstream and downstream regions; for generation of BP338LgmAKO This thesis pSS4245LgmBKO pSS4245 containing the BP338 lgmB upstream and downstream regions; for generation of BP338LgmBKO This thesis pSS4245LgmCKO pSS4245 containing the BP338 lgmC upstream and downstream regions; for generation of BP338LgmCKO This thesis pSS4245LgmDKO pSS4245 containing the BP338 lgmD upstream and downstream regions; for generation of BP338LgmDKO This thesis pSS4245LgmABCDKO pSS4245 containing the BP338 lgmA upstream and the lgmD downstream regions; for generation of BP338LgmABCDKO This thesis pSS4245BP1945KO pSS4245 containing the BP338 BP1945 locus tag upstream and downstream regions; for generation of BP338BP1945KO and BP338LgmDKOBP1945KO This thesis pBBR1MCS2 Broad-range vector, KanR (Kovach, Elzer, et al. 1995) pBBR2Pcpn pBBR1MCS2 containing the Pcpn heat shock promoter (constitutive expression) This thesis pBBR2LgmA pBBR2Pcpn containing lgmA of BP338 This thesis pBBR2LgmA D76G D77G D76G and D77G mutations in pBBR2LgmA This thesis pBBR2LgmA D127G D127G mutation in pBBR2LgmA This thesis pBBR2LgmA D129G D129G mutation in pBBR2LgmA This thesis pBBR2LgmA D127G D129G D127G and D129G mutations in pBBR2LgmA This thesis pBBR2LgmA D159N D159N mutation in pBBR2LgmA This thesis pBBR2LgmA W163R W163R mutation in pBBR2LgmA This thesis pBBR2LgmA D159N W163R D159N and W163R mutations in pBBR2LgmA This thesis pBBR2LgmC pBBR2Pcpn containing lgmC of BP338 This thesis pBBR2LgmC D80G D81G D80G and D81G mutations in pBBR2LgmC This thesis pBBR2LgmC H130G H130G mutation in pBBR2LgmC This thesis pBBR2LgmC D187G H189G D187G and H189G mutations in pBBR2LgmC This thesis pBBR2LgmC E313G E313G mutation in pBBR2LgmC This thesis pBBR2LgmC-E G15STOP The lgmE G15STOP mutation in pBBR2LgmC This thesis pBBRLpxA338 pBBR1MCS2 containing lpxA of BP338 This thesis pNMLgmAB pBBR1MCS2 containing lgmA and lgmB (Shah, Albitar-Nehme, et al. 2013) pMMB67HE Broad-range vector, AmpR, Ptac promoter (IPTG-inducible) (Furste, Pansegrau, et al. 1986) pPtacLgmABCD pMMB67HE containing lgmA, lgmB, lgmC, and lgmD of BP338 (Shah, Albitar-Nehme, et al. 2013) pPtacLgmABCD-E G15STOP pPtacLgmABCD with a G15STOP mutation in lgmE (glycine 15 mutated to a stop codon) This thesis pPtacLgmABC pMMB67HE containing lgmA, lgmB, and lgmC of BP338 This thesis pPtacLgmABC-E G15STOP pPtacLgmABC with a G15STOP mutation in lgmE (glycine 15 mutated to a stop codon) This thesis pPtacLgmAB pMMB67HE containing lgmA and lgmB of BP338 This thesis pPtacLpxA338 pMMB67HE containing lpxA of BP338 This thesis 35  Plasmids Description Source pPtacLgmABCDLpxA338 pMMB67HE containing lgmA, lgmB, lgmC, lgmD, and lpxA of BP338 This thesis pET30b Expression vector, KanR Novagen pET30LgmA pET30b containing lgmA with a C-terminal 3xGly linker connected to a 6xHis tag This thesis pBlueScript II KS (-) AmpR, lac promoter (IPTG-inducible) Stratagene pBSlpxAEC pBlueScript II with lpxA from E. coli strain DH5α, forward direction with respect to lac promoter This thesis pBSlpxA338 pBlueScript II with lpxA from B. pertussis strain BP338, forward direction with respect to lac promoter This thesis pBSlpxA338rev pBlueScript II with lpxA from B. pertussis strain BP338, reverse direction with respect to lac promoter This thesis pBSlpxA18323 pBlueScript II with lpxA from B. pertussis strain 18-323, forward direction with respect to lac promoter This thesis pBSlpxA18323rev pBlueScript II with lpxA from B. pertussis strain 18-323, reverse direction with respect to lac promoter This thesis   2.2.3 Primers  Table 3. List of primers Sequences are listed 5’ to 3’. Underlined sequences represent restriction enzyme cut sites. Primer Sequence Restriction enzyme sites lgmAKO1fw GGAATTCTGGTAGCCGTGCCGCAGCC EcoRI lgmAKO1rev CCTTAATTAAACCACGGCGAAATTGACGGG PacI lgmAKO2fw CCTTAATTAATGACTCTCGCTACCCGATCC PacI lgmAKO2rev CCCCTAGGCAGCGTGGCCAGCACCAG AvrII lgmBKO1fw CGAATTCGCGACTTCCGCCTGATGGAC EcoRI lgmBKO1rev CCTTAATTAATCATTGGGAACGCGCCTTGGC PacI lgmBKO2fw GCTTAATTAACGACGAATGTCAGGAAGGCCA PacI lgmBKO2rev CCCCTAGGGACCGCGCTCAGGCGTC AvrII lgmCKO1fw GGAATTCGCCGATGTGGGTGGTCGAC EcoRI lgmCKO1rev CCTTAATTAATCAGACTCACTTTCCGCCAC PacI lgmCKO2fw CCTTAATTAAAATGAGTTCTTCCTCGAGACAAAC PacI lgmCKO2rev CCCCTAGGTCCGTGGCCGACTTCGTACG AvrII lgmDKO1fw GGAATTCAGCGCATCTGGCTGCGG EcoRI lgmDKO1rev CCTTAATTAATCATAAACGGCTTGCCAGGC PacI lgmDKO2fw CCTTAATTAAGGCCGCTCAGGTACCGGGC PacI lgmDKO2rev CCCCTAGGACTGCCCTCGGAGCAAAGCG AvrII BP1945KO1fw CGAATTCACGAACGCCAGCCCCGCCACC EcoRI BP1945KO1rev CCTTAATTAATCAGTGGATTTTTTTACAGCATAC PacI BP1945KO2fw CCTTAATTAAAAAGCCGGCTGGACAGTTG PacI BP1945KO2rev CCCCTAGGTGGTCGCGCAACGCGCACA AvrII lgmAD76GD77Gfw CGTCGTCGGCGGCGGCAGCACCGACGATAC  lgmAD76GD77Grev TGCTGCCGCCGCCGACGACGATGATTTCCC  lgmAD127Gfw GTCTGGGCGCCGACATGCAGCATCCGCCCGAACT  lgmAD127Grev GATGCTGCATGTCGGCGCCCAGACAGATGACGGC  36  Primer Sequence Restriction enzyme sites lgmAD129Gfw GTCTGGACGCCGGCATGCAGCATCCGCCCGAACT  lgmAD129Grev GATGCTGCATGCCGGCGTCCAGACAGATGACGGC  lgmAD127GD129Gfw CTCTGGGCGCCGGCATGCAGCATCCGCCCGAACT  lgmAD127GD129Grev GATGCTGCATGCCGGCGCCCAGACAGATGACGGC  lgmAD159Nfw GCGCAACGACGAGCCGTGGTTCAAGCGTGT  lgmAD159Nrev ACCACGGCTCGTCGTTGCGCTGGCGCCGCA  lgmAW163Rfw GCGCGACGACGAGCCGCGGTTCAAGCGTGT  lgmAW163Rrev ACCGCGGCTCGTCGTCGCGCTGGCGCCGCA  lgmAD159NW163Rfw GCGCAACGACGAGCCGCGGTTCAAGCGTGT  lgmAD159NW163Rrev ACCGCGGCTCGTCGTTGCGCTGGCGCCGCA  lgmCD80GD81Gfw GTGCGGCGGTGGTTTTGGCATGAACGAGGC  lgmCD80GD81Grev TGCCAAAACCACCGCCGCACACCGCGATGC  lgmCH130Gfw GACCTGGGCGTCGGTGTCGATTTCACCGAA  lgmCH130Grev TCGACACCGACGCCCAGGTCGACGTCCAGG  lgmCD187GH189Gfw ACGTCGGCGGCGGCCAGCACGTGCATCAGC  lgmCD187GH189Grev GTGCTGGCCGCCGCCGACGTAGTCGGGCGC  lgmCE313Gfw CGCCGCCGGATACGAGGTGCTGGCGCACCC  lgmCE313Grev GCACCTCGTATCCGGCGGCGCGCTGGGAGG  BPlgmDlikeG15STOPfw TTCATTGCCGTCTGATGCGCCGCGGCCGCC  BPlgmDlikeG15STOPrev GCGCATCAGACGGCAATGAACCAGGCAATC  BP0399fw1 ATCTGTCTGGACGCCGACATGC  BP0399rev1 CAGGTAACCGCCGTAGGACAGC  BP0398fw1 TTCTTCGTCCACCAGCATTTCG  BP0398rev1 AGCTGCAGGTCGAACGGATAGG  BP0397-RTfw CGGAACATTCCGGACCTTACCC  BP0397-RTrev CCACCAGCACGTGCATCAGC  BP0396-RTfw GCGCGCAGCTTGTCATAGGC  BP0396-RTrev CAAACCCCCAAAACCATCAAGG  BPlgmDlike-RTfw CCGTCGCGGTCGCCTGCGTG  BPlgmDlike-RTrev TGACGGAACGCCCACAGGCG  vag8fw1 CCCCAAGCTTCGTCCGAGCACGGTATCAACG HindIII vag8rev1 CGCTCTAGACACATAGATCCCGGCGACTTCC XbaI lgmAfw4 CCATCGATTTCGCCGTGGTGTGTTCATG ClaI lgmArev3 GCTCTAGATCATTGGGAACGCGCCTTG XbaI lgmAfw3-NdeI GGGAGTCATATGTGTTCATGTATACCGAATTCCG NdeI lgmArev3-XhoI CCGCTCGAGACCACCACCTTGGGAACGCGCCTTGGC XhoI lgmCfw3 GGAATTCCATGACCAGTGAACGATACGA EcoRI lgmCrev3 GCTCTAGATCATAAACGGCTTGCCAG XbaI BPlpxAfw2 CCCCAAGCTTCTGCCGCATCGCTACCCGA HindIII BPlpxArev1 GCTCTAGACCGGCGACCATGCCTATG XbaI lpxAECfw1 GGAATTCGGCCTGATACGTGATTGATAAAT EcoRI lpxAECrev1 GCTCTAGACGCTGTTCAGTCATTAACGA XbaI BPlpxAfw3 GCTCTAGACGATCCGTAGCCTGGAAGA XbaI BPlpxArev3 GGAATTCCGGCGACCATGCCTATG EcoRI lpxAECfw1 GGAATTCGGCCTGATACGTGATTGATAAAT EcoRI lpxAECrev1 GCTCTAGACGCTGTTCAGTCATTAACGA XbaI BPlpxAfw3 GCTCTAGACGATCCGTAGCCTGGAAGA XbaI BPlpxArev3 GGAATTCCGGCGACCATGCCTATG EcoRI   37  2.3 Cloning of vectors and deletion strains 2.3.1 General cloning techniques All restriction endonuclease enzymes and other cloning-related enzymes were from New England Biolabs, unless otherwise indicated. The cloned genes in all vectors were confirmed by sequencing and PCR amplification. Plasmids were introduced into E. coli and B. pertussis strains using the following techniques.  A. CaCl2-heat shock transformation To prepare transformation-competent E. coli DH5α, S17-1, BL-21, or R0138 samples, 3 ml LB broth with no antibiotics was inoculated with E. coli and grown overnight while shaking at 200 rotations per minute (rpm) at 37°C (30°C for R0138). 1 ml of this overnight culture was used to inoculate 100 ml LB broth and grown to an OD600 of 0.4 to 0.5 at 37°C (30°C for R0138) and shaking at 200 rpm. These cells were harvested by centrifugation at 2100 g at 4°C for 10 min and immediately resuspended in 10 ml ice cold, sterile 50 mM CaCl2 and incubated on ice for 10 min. The cells were harvested again and resuspended in 2 ml ice cold, sterile 50 mM CaCl2 + 15% glycerol (w/v). Aliquotes of 50 µl were dispensed per Epindorf tubes and then stored at -80°C for storage until use for transformation.  To introduce a vector into E. coli DH5α, S17-1, BL-21, or R0138 cells: a 50 µl aliquot of transformation-competent cells was thawed on ice, then 1 µl of vector or a 20 µl ligation reaction was added to these cells, mixed and incubated on ice for 30 min. The mixture was then heat-shocked at 42°C for 45 s followed by a recovery period of 2 min on ice. 1 ml of LB broth was added to these cells and they were incubated at 37°C (30°C for R0138 strains) and shaken at 200 rpm for 1 h. Cells were then plated onto LB agar containing the required selection antibiotics and incubated at 37°C (30°C for R0138 strains) for approximately 16 h (24-30 h for R0138 strains) or until colonies were visible.  38  B. Electroporation This protocol was used to introduce vectors into B. pertussis strains. To prepare electrocompetent bacterial samples: B. pertussis was first grown on BG agar containing the appropriate antibiotics for three days at 37°C. These bacteria were used to inoculate 300 ml SS broth supplemented with the appropriate antibiotics at an initial OD600 of 0.01 and grown at 37°C shaking at 180 rpm until an OD600 of 0.8 (approximately 48 hours). Bacteria were harvested by centrifugation at 11 350 g for 10 min at 4°C followed by two washes with 150 ml sterile deionized water (dH2O) and one wash with 150 ml sterile 272 mM sucrose + 15% glycerol (w/v). The bacteria were then resuspended in 6 ml sterile 272 mM sucrose + 15% glycerol (w/v) and 400 µl aliquots were stored at -80°C for storage until use.  Introduction of a vector into an electrocompetent bacterial sample was accomplished as follows: 400 µl electrocompetent cells were thawed on ice and added to a 0.2 cm electrode gap electroporation cuvette (VWR). One µg of vector was added to the cuvette and the mixture was pulsed at 2.5 kV with a GenePulser Xcell (BioRad). The vector-bacteria mixture was immediately transferred into 1 ml SS broth and then incubated at 37°C shaking at 180 rpm for 1 h. The bacterial cells were then pelleted by centrifugation at 5 160 g for 5 min, then resuspended in 100 µl SS broth and plated onto BG agar containing the appropriate selection antibiotics. The BG agar plates were incubated at 37°C for 3-5 days, until colonies were visible.  C. Conjugation A diparental mating protocol was used to introduce vectors into B. pertussis strains, as previously described (Marr, Hajjar, et al. 2010). First, the vector was introduced into E. coli S17-1 donor strain via CaCl2-heatshock transformation. The B. pertussis acceptor strain was grown on BG agar containing the appropriate antibiotics for 3 days at 37°C and the E. coli S17-1 donor strain containing the vector was grown on LB agar supplemented with the appropriate antibiotics overnight at 37°C. B. pertussis cells 39  were then resuspended directly from the BG agar plate into SS broth and the E. coli cells were similarly resuspended in LB broth. The equivalent of 1 ml of B. pertussis at an OD600 of 1.0 and the equivalent of 100 µl of E. coli at an OD600 of 1.0 were mixed together and poured onto a mating plate (Marr, Hajjar, et al. 2010) and incubated at 37°C for 5 to 7 h. Bacterial cells were then removed from the mating plate with a sterile swab and streaked onto BG agar containing the appropriate selection antibiotics for both the parental B. pertussis strain and the vector. BG plates were incubated at 37°C for 3 to 5 days, until colonies were visible. 2.3.2 Generating markerless deletion mutants The following general protocol was used to generate markerless deletion mutants in B. pertussis strains, as summarized in Figure 9. I will use the generation of BP338LgmAKO as an example, but a similar method was used for all other markerless deletion mutants. Approximately 300 to 500 base pairs of the up and downstream regions of the gene targeted for deletion, lgmA, were cloned into the vector pSS4245 to generate pSS4245LgmAKO, such that the upstream nucleotide (nt) region was directly joined to the downstream nt region by the cut site PacI. This vector was then introduced into E. coli strain S17-1 via CaCl2-heat shock transformation (Section 2.3.1). pSS4245LgmAKO was then introduced into the B. pertussis strain BP338 by conjugation (Section 2.3.1) and plated onto Bordet-Gengou (BG) agar supplemented with the appropriate selection antibiotics (i.e. naladixic acid, to select for BP338, and streptomycin, to select for pSS4245LgmAKO) and with 50 mM MgSO4. The presence of 50 mM MgSO4 in the growth medium suppresses the Bvg system, therefore the restriction endonuclease I-SceI is not expressed from pSS4245LmgAKO. Since pSS4245 cannot replicate in Bordetella species, this vector must integrate into the chromosome via homologous recombination. It will likely integrate at either the upstream or downstream region of lgmA, thus introducing the entire vector into this region. The BG agar plates were incubated at 37°C for 3 to 5 days, until colonies were visible, and these colonies were then streaked onto a BG agar plate supplemented with only nalidixic acid. Since MgSO4 is not present in this growth medium, the Bvg system of B. pertussis would be active, and I-SceI, which is under the control of 40  a Bvg promoter, would now be expressed. I-SceI endonuclease would cleave the I-SceI restriction enzyme cut site present in pSS4245, therefore resulting in a double-stranded break in the chromosomal DNA at this site. The bacteria would attempt to repair this DNA break via homologous recombination, which would result in either removal of the originally inserted vector, pSS4245LgmAKO, or removal of the entire vector plus lgmA, resulting in a markerless deletion mutant of lgmA. Since the BG agar no longer contains streptomycin, there is no selection pressure to maintain the pSS4245 vector in the chromosome, and the resulting colonies were screened for kanamycin susceptibility, to indicate a loss of the pSS4245 vector, and by PCR for the generation of a markerless lgmA mutant. The lack of the targeted gene at the specific site in the chromosome was confirmed in all markerless deletion mutants by PCR amplification of the targeted chromosomal DNA region and sequencing of this product.   To generate the markerless deletion mutant, BP338LgmAKO, the lgmA intergenic upstream and downstream regions were PCR amplified from genomic DNA of B. pertussis strain BP338 using the primer sets (lgmAKO1fw and lgmAKO1rev) and (lgmAKO2fw and lgmAKO2rev), respectively. These regions were cloned into the suicide vector pSS4245 using EcoRI, PacI, and AvrII, such that the lgmA intergenic upstream region is directly joined to the lgmA intergenic downstream region via a PacI restriction enzyme cut site sequence. This resulted in the generation of pSS4245LgmAKO. As described above, pSS4245LgmAKO was used to generate the markerless deletion mutant BP338LgmAKO such that the lgmA gene was replaced by a PacI restriction enzyme cut site sequence. A similar method was used to generate all markerless deletion mutants in this thesis, and the specific primer sets and parental strains are summarized in Table 4.   41     Figure 9. Schematic for the generation of markerless deletion mutants. In this schematic, the generation of BP338LgmAKO, the lgmA clean deletion mutant of B. pertussis strain BP338 is used as an example. Vector and strain names are underlined.  The pSS4245 vector contains an I-SceI enzyme cleavage site and the I-SceI gene (pink arrow) under the control of the pertussis toxin promoter, which induces expression under Bvg+ conditions.  The pSS4245LgmAKO vector (red lines) contains the region upstream of lgmA (green triangle) and the region downstream of lgmA (blue triangle).  B. pertussis strain BP338 chromosomal DNA is depicted with a black line.  In the first step, pSS4245LgmAKO was introduced into BP338 under Bvg- conditions and, via homologous recombination, the vector inserted into the BP338 chromosome (in this case, homologous recombination via the region upstream of lgmA is depicted).  This strain was then grown under Bvg+ conditions, therefore inducing the expression of the I-SceI restriction enzyme.  I-SceI would cleaved the chromosomal DNA at the I-SceI cleavage site encoded by pSS4245LgmAKO, therefore promoting homologous recombination in this region in an effort to repair the double stranded break in the DNA.  Two homologous combination events are probable:  1) homologous recombination via the upstream region, which leads to the removal of the vector, leaving the original BP338 lgmA gene intact, or 2) homologous recombination via the downstream region, which leads to the removal of the vector and the lgmA gene, resulting in the generation of a markerless lgmA mutant of BP338.     42  Table 4. Summary of markerless deletion mutant cloning Deletion mutant Parental strain Suicide plasmid Upstream primer set Downstream primer set BP338LgmAKO BP338 pSS4245LgmAKO lgmAKO1fw lgmAKO1rev lgmAKO2fw lgmAKO2rev BP338LgmBKO BP338 pSS4245LgmBKO lgmBKO1fw lgmBKO1rev lgmBKO2fw lgmBKO2rev BP338LgmCKO BP338 pSS4245LgmCKO lgmCKO1fw lgmCKO1rev lgmCKO2fw lgmCKO2rev BP338LgmDKO BP338 pSS4245LgmDKO lgmDKO1fw lgmDKO1rev lgmDKO2fw lgmDKO2rev BP338LgmABCDKO BP338 pSS4245LgmABCDKO lgmAKO1fw lgmAKO1rev lgmDKO2fw lgmDKO2rev BP338BP1945KO  BP338 pSS4245BP1945KO BP1945KO1fw BP1945KO1rev BP1945KO2fw BP1945KO2rev BP338LgmDKOBP1945KO BP338LgmDKO pSS4245BP1945KO BP1945KO1fw BP1945KO1rev BP1945KO2fw BP1945KO2rev  2.3.3 Site-directed mutagenesis The following general protocol was used to generate site-directed mutants. I will use pBBR2LgmA D127G as an example, though all site-directed mutants were generated using a similar procedure. The non-mutated original vector pBBR2LgmA was used as the template in a PCR reaction with the primers lgmAD127Gfw and lgmAD127Grev. The 60 µl PCR mixture contained: 12 µl 5X Q5® reaction buffer (New England Biolabs), 12 µl 5X GC enhancer  (New England Biolabs), 1.2 µl 10 mM dNTP (BioBasic), 3 µl template, 3 µl of each primer at 10 uM, 0.6 µl Q5® High-Fidelity DNA Polymerase, dH2O for the remaining reaction volume. The PCR reaction was as follows: 5 min at 95°C, 18 cycles of (50 s at 95°C, 50 s at 60 to 67°C, 8 min at 72°C), 10 min at 72°C. The reaction was treated with DpnI (addition of 1 unit/µl DpnI and incubation at 37°C for 1 h) to remove the original methylated pBBR2LgmA vector followed by inactivation of DpnI by incubation at 80°C for 20 min. This reaction was introduced into E. coli DH5α cells by CaCl2-heatshock transformation (Section 2.3.1) and plated onto LB agar containing kanamycin to select for the plasmid. Transformant colonies were screened for the presence of the site-directed mutation by sequencing. Table 5 summarizes the primer sets used to generate site-directed mutants with the previously described method.  43  To clone pPtacLgmABCD-E G15STOP, pPtacLgmABCD and pBBR2LgmC-E G15STOP vectors were digested with NotI and AatII, and the fragment containing the G15STOP mutation from pBBR2LgmC-E G15STOP was ligated to the pPtacLgmABCD vector backbone, to transfer the lgmE G15STOP mutation to this vector. Similarly, the lgmE G15STOP mutation was cut out of pPtacLgmABCD-E G15STOP and ligated into pPtacLgmABC with enzymes DraIII and DrdI, to generate pPtacLgmABC-E G15STOP.   44   Table 5. Summary of site-directed mutant primer sets Site-directed mutant vector Template vector Primer set pBBR2LgmA D76G D77G pBBR2LgmA lgmAD76GD77Gfw, lgmAD76GD77Grev pBBR2LgmA D127G pBBR2LgmA lgmAD127Gfw, lgmAD127Grev pBBR2LgmA D129G pBBR2LgmA lgmAD129Gfw, lgmAD129Grev pBBR2LgmA D127G D129G pBBR2LgmA lgmAD127GD129Gfw, lgmAD127GD129Grev pBBR2LgmA D159N pBBR2LgmA lgmAD159Nfw, lgmAD159Nrev pBBR2LgmA W163R pBBR2LgmA lgmAW163Rfw, lgmAW163Rrev pBBR2LgmA D159N W163R pBBR2LgmA lgmAD159NW163Rfw, lgmAD159NW163Rrev pBBR2LgmC D80G D81G pBBR2LgmC lgmCD80GD81Gfw, lgmCD80GD81Grev pBBR2LgmC H130G pBBR2LgmC lgmCH130Gfw, lgmCH130Grev pBBR2LgmC D187G H189G pBBR2LgmC lgmCD187GD189Gfw, lgmCD187GD189Grev pBBR2LgmC E313G pBBR2LgmC lgmCE313Gfw, lgmCE313Grev pBBR2LgmC-E G15STOP pBBR2LgmC BPlgmDlikeG15STOPfw, BPlgmDlikeG15STOPrev  2.3.4 Vectors to complement B. pertussis strains I complemented the lgmA, lgmB, and lgmC knockout mutants in B. pertussis strain BP338 (BP338LgmAKO, BP338LgmBKO, and BP338LgmCKO, respectively) with vectors expressing the respective wild-type gene.   To generate the lgmA and lgmC complementing vectors, I first generated pBBR2Pcpn, a vector that encoded a constitutively-expressing promoter, Pcpn. The Pcpn heat shock promoter was cut out of pBBRPcpnBrkA (Marr, Shah, et al. 2011) using the KpnI sites, and cloned into the KpnI site of pBBR1MCS2 in the same direction as the kanamycin resistance cassette. Then I cloned lgmA or lgmC into this vector such that the expression of these genes was driven by the Pcpn promoter. The B. pertussis BP338 lgmA gene was PCR amplified with primers lgmAfw4 and lgmArev3, and the product was cloned into the ClaI and XbaI sites of pBBR2Pcpn to generate the vector pBBR2LgmA. To construct pBBR2LgmC, I PCR amplified lgmC from BP338 with primers lgmCfw3 and lgmCrev3, and then cloned this PCR product into the EcoRI and XbaI sites of pBBR2Pcpn.  45  To complement BP338LgmBKO, the region containing lgmA and lgmB was cloned from pNMLgmAB into pMMB67HE using HindIII and XbaI to generate the vector pPtacLgmAB, in which lgmA and lgmB are under the control of the IPTG-inducible promoter, Ptac.   The vector pPtacLgmABC was used to complement the full lgm locus mutant BP338LgmABCDKO. The region containing lgmA, lgmB, and lgmC was cut out of pPtacLgmABCD with XhoI and cloned into pMMB67HE cut with SalI, since XhoI and SalI have identical overhanging regions.  I also complemented B. pertussis wild type strain 18-323 with vectors encoding lipid A-modifying genes from strain BP338: lpxA, the lgm locus, or both. First I cloned the vector pBBRLpxA338. lpxA and the intergenic upstream region of lpxA, which presumably contains the endogenous lpxA promoter, was PCR amplified from B. pertussis BP338 chromosomal DNA with primers BP338lpxAfw2-HindIII and BP338lpxArev1-XbaI. The PCR product was cloned into the XbaI and HindIII sites of pBBR1MCS2. Then, lpxA from B. pertussis BP338 and the intergenic upstream region of lpxA were cut out of pBBRLpxA338 with HindIII and XbaI, and the ends were blunted with Klenow. pPtacLgmABCD was cut with EcoRI to cut out the LgmABCD locus, and the ends of the vector backbone were blunted with Klenow, into which the DNA fragment containing lpxA from BP338 and the intergenic upstream region of lpxA were cloned. A clone was selected in which the lpxA gene is in the forward direction in comparison to the pPtac promoter, resulting in the vector pPtacLpxA338.   To complement 18-323 with both lpxA and the lgm locus from B. pertussis strain BP338, I generated the vector pPtacLgmABCDLpxA338. lpxA from BP338 and the intergenic upstream region of lpxA were cut out of pBBRLpxA338 with HindIII and XbaI, and the ends were blunted with Klenow. pPtacLgmABCD was cut with SapI, resulting in a 13.04 kb linear vector. The ends of this vector were blunted with Kelnow, and pPtacLgmABCD was ligated to the blunt-ended fragment of lpxA from BP338 and the 46  intergenic upstream region. The resulting vector contains the lgmABCD locus under the control of the Ptac promoter, and the lpxA gene from BP338 under the control of its endogenous promoter. 2.3.5 Vectors to complement E. coli strain R0138 E. coli strain R0138 encodes a temperature-sensitive lpxA gene (lpxA2), and was complemented with vectors encoding lpxA from E. coli strain DH5α, B. pertussis strain BP338, or B. pertussis strain 18-323. I constructed the vector pBSLpxAEC, by PCR amplifying lpxA from E. coli DH5α chromosomal DNA with primers lpxAECfw1 and lpxAECrev1 and cloning this PCR product into pBlueScript II KS (-) using XbaI and EcoRI. This resulted in lpxA from E. coli oriented in the forward direction in reference to the lac promoter.  I constructed two vectors with lpxA from B. pertussis strain BP338: pBSLpxA338 and pBSLpxA338rev. To clone pBSLpxA338, the lpxA gene from B. pertussis strain BP338 was cut out of pPtacLpxA338 and cloned into pBlueScript II KS (-) using EcoRI and XbaI, resulting in lpxA from BP338 in the forward orientation in reference to the lac promoter. In pBSLpxA338rev, however, the BP338 lpxA gene was cloned in the reverse orientation in reference to the lac promoter by cutting lpxA out of pPtacLpxA338 and cloning it into pBlueScript II KS (-) using EcoRI and SalI.   I also cloned the two vectors pBSLpxA18323rev and pBSLpxA18323, which both encode the lpxA gene from B. pertussis strain 18-323. First, I PCR amplified 18-323 lpxA from chromosomal DNA with primers BPlpxAfw3 and BPlpxArev3 and cloned this PCR product into pBlueScript II KS (-) with EcoRI and XbaI, such that lpxA from 18-323 is in the reverse orientation in reference to the lac promoter. This generated the vector pBSLpxA18323rev. Next, I cloned pBSLpxA18323 by excising an internal region of lpxA of 18-323 (that contains the single base pair difference when compared to lpxA of BP338) from pBSLpxA18323rev with AscI and SphI and cloning it into pBSLpxA338, which was also cut by AscI and 47  SphI remove the BP338 lpxA internal region. This resulted in lpxA of 18-323 in the forward orientation in reference to the lac promoter. 2.3.6 Vector for LgmA expression in E. coli strain BL-21 pET30LgmA was used to express LgmA in E. coli strain BL-21(DE3). First, lgmA was PCR amplified with primers lgmAfw3-NdeI and lgmArev3-XhoI to generate a product of lgmA with a C-terminal 3x glycine linker. Then, both this PCR product and the pET30b vector were digested with NdeI and XhoI and ligated together to generate a construct with lgmA connected to a C-terminal 6x histidine (His) tag by a 3x glycine linker.  2.4 Preparation of bacterial cells for mass spectrometry analysis and TLR4-activation assays B. pertussis strains were grown on BG agar at 37 °C for 3 to 4 days, and these cells were used to inoculate SS broth, with the appropriate antibiotics, at an OD600 of 0.01. Cultures were grown at 180 rpm at 37 °C until an OD600 of 0.6 to 0.9, and samples from each culture were grown on BG agar to confirm a hemolytic phenotype. Bacterial cells were harvested into phosphate-buffered saline (PBS) pH = 7.4 to an OD600 of 5 and heat-inactivated by incubation at 56 °C for 1 h. To confirm heat-killing, 50 µl of each cell suspension was spotted onto BG agar supplemented with no antibiotics and incubated at 37 °C for 5 days. Heat-killed cells were stored at -20 °C. These cells were used for stimulation in a TLR4 activation assay, generation of highly purified LPS for use as the stimulus in the TLR4 activation assays, and direct lipid A isolation for use in matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS) analysis.  For E. coli R0138 strains, a similar protocol was used to prepare the bacterial cells, with the following exceptions: E. coli R0138 strains were grown on LB agar at 30°C for 48 h, after which a single colony was used to inoculate LB broth supplemented with 1 mM IPTG. This culture was grown for 16-20 h, shaking at 200 rpm at 30°C for R0138 wild type strain and at 42°C for the complemented strains (R0138 + pBSlpxAEC, R0138 + pBSlpxA338, and R0138 + pBSlpxA18323). E. coli cells were heat-killed at 48  80°C for 1 h, and to confirm the bacteria were killed, 50 µl of each cell suspension was spotted onto LB agar supplemented with no antibiotics and incubated at 30 °C for 3 days. These E. coli R0138 cells were used for direct isolation of lipid A for use in MALDI-MS analysis.  2.5 Isolation of LPS and lipid A 2.5.1 Isolation of B. pertussis LPS for TLR4 activation assays B. pertussis LPS preparations were extracted by an ammonium hydroxide-isobutyric acid method by members of the Caroff lab (M. Caroff Patent WO 2004/062690 A1, March 20, 2012). Primary extracts were subjected to a standard enzyme treatment (DNase, RNase, and proteinase K) and finally repurified with the acidified chloroform-methanol-water procedure as described (Tirsoaga, Novikov, et al. 2007). To be sure that no specific LPS molecular species were discriminated against during the process, all intermediate and final products were analyzed by MALDI-MS. High purity of the resulting LPS preparations was evidenced by three different methods. The absence of contaminating (non-LPS) peaks was attested by positive-ion MALDI mass spectra analysis; the absence of detectable protein contaminants was demonstrated by Tricine-SDS-PAGE (Schagger and von Jagow 1987) and silver staining (Tsai and Frasch 1982) loading up to 250 ng of LPS from each preparation; and the absence of lipoprotein content was further demonstrated based on the lack of detectable levels of cysteine by analysis with an amino acid analyzer (Hitachi L-8800, equipped with a 2620MSC-PS column, ScienceTec, Les Ulis, France). 2.5.2 Isolation of E. coli LPS for TLR4 and LAL activation assays The protocol was adapted from Rezania et al. (Rezania, Amirmozaffari, et al. 2011). E. coli R0138 strains were grown on LB agar at 30°C for 48 hours. 50 ml LB broth supplemented with 1 mM IPTG was inoculated with freshly grown colonies and incubated at 42°C at 200 rpm for 16-20 h. Cells were harvested by centrifugation at 10 000 g for 5 min and washed twice with 0.15 M PBS (pH = 7.2) solution. Cells were then resuspened in 10 ml 0.15 M PBS (pH = 7.2), sonicated on ice for 10 min, and treated with 49  proteinase K (100 µg/ml proteinase K incubated at 65°C for 1 h) and then with RNase and DNase (40 µg/ml RNase A, 20 µg/ml DNase I, 0.02% MgSO4, and 0.004% chloroform incubated at 37°C for 16-20 h). LPS was extracted using a hot phenol extraction method. Briefly, 10 ml 90% phenol at 65-70°C was added to 10 ml of the cell mixture and incubated for 15 min at 65-70°C, then mixed in an ice bath for 15 min. The phases were separated by centrifugation at 8 500 g for 15 min and the aqueous layer was re-extracted using this hot phenol method. Sodium acetate was added to the final aqueous phase to a final concentration of 0.5 M and 10 volumes of ethanol were added to this mixture and stored at -20°C for 16-20 h to precipitate the LPS. This mixture was centrifuged at 2 000 g at 4°C for 10 min, and the remaining LPS pellet was resuspended in water and dialyzed against water at 4°C for 16-20 h to remove any residual phenol. The LPS sample was then lyophilized and resuspended in DMSO to a final concentration of 1 mg/ml.  This purified E. coli LPS was used for HEK-Blue hTLR4 activation assays and LAL chromogenic assays. 2.5.3 Isolation of lipid A for mass spectrometry analysis For MALDI-MS analysis, lipid A was isolated directly by hydrolysis of bacterial cells as described previously by members of the Caroff lab (El Hamidi, Tirsoaga, et al. 2005, Tirsoaga, El Hamidi, et al. 2007). Briefly, lyophilized bacterial cells (10 mg) were suspended in 200 µl of a isobutyric acid:1 M ammonium hydroxide (5:3, v/v) mixture and were kept for 2 h at 100°C in a screw-cap test tube under magnetic stirring. The suspension was cooled in ice water and centrifuged at 2000 g for 10 min. The recovered supernatant was diluted with 2 volumes of water and lyophilized. The sample was then washed once with 200 µl of methanol by centrifugation at 2000 g for 10 min. Finally, lipid A was extracted from the pellet in 100 µl of a chloroform:methanol:water (3:1.5:0.25, v/v) mixture. In some of the spectra, peaks corresponding to small contaminants were identified and marked with an “X” (peaks at m/z 1349 and 1377 in Figure 34).  50  2.6 Mass spectrometry analysis Mass spectrometry was performed in the Caroff lab. LPS samples were dispersed in water at 1 µg/µl. Lipid A extracts in chloroform-methanol-water were used directly. In both cases, a few microliters of sample solution were desalted with a few grains of ion-exchange resin Dowex 50W-X8 (H+). 0.5 to 1 µl aliquots of the solution were deposited on the target, covered with matrix solution, and allowed to dry. Dihydroxybenzoic acid (Sigma-Aldrich) was used as matrix. It was dissolved at 10 mg/ml in 0.1 M citric acid solution in the same solvents as those used for the analytes (Therisod, Labas, et al. 2001). Different analyte/matrix ratios (1:2, 1:1, 2:1, v/v) were tested to obtain the best spectra. Analyses were performed on a PerSeptive Voyager-DE STR time-of-flight mass spectrometer (Applied Biosystems) in linear mode, with delayed extraction. Negative- and positive-ion mass spectra were recorded. The ion-accelerating voltage was set at -20 kV, and the extraction delay time was adjusted to obtain the best resolution and signal-to-noise ratio.  2.7 HEK-Blue hTLR4 activation assay 2.7.1 Maintenance of HEK-Blue cells HEK-Blue (InvivoGen) cell lines hTLR4 and Null2 cells were maintained in complete Dulbecco’s modified Eagle’s medium (DMEM) containing 10% heat-inactivated fetal calf serum, 100 units/ml penicillin, 100 µg/ml streptomycin (Gibco, Life Technologies), 2 mM GlutaMAX, 1 mM pyruvate (Life Technologies), and 100 µg/ml Normocin (InvivoGen). Both the Null2 and the hTLR4 cell lines express secreted alkaline phosphatase under the control of an NFκB promoter. In addition, the hTLR4 cell line is stably transfected with human TLR4, MD-2, and CD14 co-receptor genes, such that hTLR4 activation leads to transcription from the NFκB promoter and expression of secreted alkaline phosphatase. Null2 cells were grown in the presence of 100 µg/ml Zeocin (InvivoGen), and hTLR4 cells were grown in the presence of 100 µg/ml Zeocin, 200 µg/ml Hygrogold, and 30 µg/ml Blasticidin (InvivoGen). Cells were incubated at 37°C in humid air with 5% CO2. 51  2.7.2 hTLR4 activation assay The HEK-Blue (Invivo-Gen) manufacturer’s guidelines were followed to assay hTLR4 activation by purified LPS or heat-killed bacterial samples. Briefly, HEK-Blue hTLR4 and Null2 were seeded in 96-well plates at 25 000 or 50 000 cells/well, 100 µl/well. All media used in this assay were pyrogen-free and consisted of complete DMEM, with the absence of the following antibiotics: Normocin, Zeocin, Hygrogold, and Blasticidin. HEK-Blue cells were incubated at 37°C (as described in Section 2.7.1) and at 24 h, 100 µl of fresh medium was added to each well, and the cells were incubated for another 24 h. HEK-Blue cells were then washed with 100 µl of medium and stimulated with 100 ng/ml purified B. pertussis LPS, 0.1 ng/ml purified E. coli LPS, or heat-killed B. pertussis cells (prepared as described in Section 2.4) at a 1/10 dilution. After 24 h of incubation, the supernatants were assayed for secreted alkaline phosphatase activity by mixing 20 µl of supernatant with 180 µl of QUANTI-Blue (InvivoGen) reagent in a 96-well plate and incubating this mixture at 37°C in the dark until the color of the mixture started turning blue. Absorbance at 650 nm was used as a read out of alkaline phosphatase activity, which indicates NFκB activation via hTLR4. In each assay, the Null2 cell line and unstimulated hTLR4 cells were used as negative controls.  2.8 E. coli growth curves E. coli R0138 strains were grown on LB agar with appropriate antibiotics for 48 h at 30° C, and growth from these agar plates was used to inoculate LB broth with the appropriate antibiotics and grown shaking at 200 rpm at 30°C for 16 to 20 h. These cultures were used to inoculate fresh LB broth supplemented with the appropriate antibiotics at an initial OD600 of 0.1, and all complemented strains were also supplemented with 1 mM IPTG (R0138 + pBSlpxAEC, R0138 + pBSlpxA338, and R0138 + pBSlpxA18323). Cultures were grown at 30°C or 42°C shaking at 200 rpm, and growth was monitored by OD600 measurements over time.  52  2.9 Bacterial survival assays 2.9.1 Polymyxin B growth curve assay E. coli R0138 strains were grown on LB agar with the appropriate antibiotics for 48 h at 30°C and these freshly grown bacterial were used to inoculate LB broth supplemented with the appropriate antibiotics and IPTG and grown at 42°C and shaking at 200 rpm for 16 to 20 h. Bacterial cultures were then diluted to an OD600 of 0.1 and incubated at 42°C and shaking at 200 rpm in the presence of polymyxin B concentrations: 0, 0.5, or 1.0 µg/ml. Growth was monitored via OD600 measurements over time. 2.9.2 Killing assays Protocol was adapted from Wiegand et al (Wiegand, Hilpert, et al. 2008). E. coli R0138 strains were grown on LB agar supplemented with appropriate antibiotics for 48 h at 30°C. LB broth with the appropriate antibiotics and IPTG was inoculated with freshly grown R0138 colonies and incubated at 42°C at 200 rpm for 16 to 20 h. 1 ml of this culture was then used to inoculate 4 ml of LB broth supplemented with the appropriate antibiotics and IPTG, and this was grown at 42°C at 200 rpm for 1 h. Cells were then diluted in LB broth to an OD600 of 0.1, and mixed with an equal volume of polymyxin B diluted in LB broth for final concentrations of: 0, 0.5, 1.0, or 2.0 µg/ml polymyxin B in polypropylene 96-well plates. After incubation in polymyxin B for 5 min, 100 µl of 1/100 and 1/1000 diluted samples were plated on LB agar plates with no antibiotics and grown at 30°C for 48 h, after which the colony forming units (cfu) were determined to calculate percent survival. n = 3 for each condition (i.e. each strain at a specific polymyxin B concentration).  For B. pertussis assays, B. pertussis strains were grown on BG agar with no antibiotics at 37°C for 3 days, after which their hemolytic phenotype was confirmed. These cells were used to inoculate 20 ml SS broth supplemented with nalidixic acid at an initial OD600 of 0.01. Cultures were grown at 37°C while shaking at 180 rpm for 24 to 30 h. These cultures were then diluted into SS salts (10.72 g glutamic acid, 0.24 g proline, 2.50 g NaCl, 0.50 g KH2PO4, 0.20 g KCl, 0.10 g MgCl2-6H2O, 0.02 g CaCl2, 3.175 g Tris-HCl, 53  0.59 g Tris base dissolved in 1L distilled deionized water, pH 7.6) to an OD600 of 0.002. The killing agent (peptide, antibiotic, lysozyme, and/or ethylenediaminetetraacetic acid (EDTA)) was diluted from the stock solutions into SS salts. 50 µl of the diluted bacterial culture was mixed with 50 µl of the diluted killing agent in polypropylene 96-well plates and incubated at 37°C for 2 h. A list of CAMP and antibiotics used as killing agents in this assay is in Table 6.  For the B. pertussis spot assays, after the 2 h incubation, the bacteria and killing agent mixture was serially diluted in SS salts and 2 µl of each 10 fold dilution was spotted onto BG agar plates with no antibiotics and incubated at 37°C for 72 h to visualize growth.  For the B. pertussis percent survival assays, following the 2 h incubation, 100 µl of a 1/100 and 1/1000 dilution of the bacteria and killing agent mixture was plated onto BG agar plates with no antibiotics. After incubation at 37°C for 72 h, the percent survival was determine from the number of cfu for the samples with the killing agent compared to bacterial samples incubated with no killing agents. n = 3 for each condition (i.e. killing agent at a specific concentration).   54  Table 6. List of CAMPs and antibiotics used in killing assays The estimated net charges are at pH 7.0 and expressed to the nearest integer value.  CAMP or Antibiotic Sequence  Net Charge Structure Class Polymyxin B (Sigma-Aldrich)  +5 cyclic, lipidated polymyxin Polymyxin E (Sigma-Aldrich)  +5 cyclic, lipidated polymyxin LL-37 LLGDFFRKSKEKIGKEFKRIVQRIKDFLRNLVPRTES +6  helical cathelicidin Indolicidin ILPWKWPWWPWRR-NH2 +3 extended cathelicidin HHC-10 KRWWKWIRW-NH2 +3  helical synthetic CP28 KWKLFKKIGIGAVLKVLTTGLPALKLTK-NH2 +7  helical Insect hybrid Gentamicin (Sigma-Aldrich)  +5 trisaccharide aminoglycoside  2.10 LAL activation assay The LAL assay detects the presence of endotoxin, also known as LPS, in a sample through the interaction of LPS with Factor C from the Horse Shoe crab, Limulus polyphemus. The Pyrochrome LAL assay (Associates of Cape Cod Inc.) manufacturer’s guidelines were followed to assay LAL activation by 10 µg/ml purified LPS from E. coli strains. Briefly, LPS samples were diluted to 10 µg/ml in Glucashield Reconstitution Buffer using pyrogen-free instruments and a Pyroplate 96-well microplate (Associates of Cape Cod Inc.). An equal volume of freshly reconstituted Pyrochrome reagent was added to each sample and mixed by pipetting the samples up and down two times, carefully, to ensure no bubbles were formed. The presence of LPS was detected by monitoring absorbance at A405 over time.  2.11 Glycosyltransferase assay The protocol was adapted from Song et al. (Song, Guan, et al. 2009) and the butanol extraction protocol was adapted from Ravishankar et al. (Ravishankar, Kumar, et al. 2005). To demonstrate glycosyltransferase activity of LgmA, E. coli membranes were used as the source of both the assayed protein, LgmA, and the acceptor lipid and UDP-GlcN[14C]Ac was the donor molecule. An initial 10 ml 55  culture was inoculated with E. coli BL-21 (DE3) strains and grown for 16 to 20 h. This 10 ml culture was then added to 90 ml LB and grown for an additional 5 to 7 h, at which point cells were harvested from 500 µl of culture for Western blot analysis. The bacteria were induced with a final concentration of 1 mM IPTG and grown for another 16 h, upon which 500 µl of induced bacterial cells were harvested for Western blot analysis. The full 100 ml culture was then harvested by centrifugation at 6 000 g for 15 min at 4°C. From this point onwards, the cells were placed on ice, or centrifuged or stored at 4°C, to prevent degradation of the cell components. Cells were washed in 50 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) buffer and then resuspended in 10 ml 50 mM HEPES buffer. Cells were then disrupted via 5 passages through a cell homogenizer, EmulsiFlex-05 (Avestin), at an average of 9 000 psi, followed by removal of unbroken cells by centrifugation at 10 000 g for 20 min. The membranes were collected from the resulting supernatants via centrifugation at 100 000 g for 1 h. The membrane pellet was then resuspended in 50 mM HEPES buffer and the total protein concentration was determined by the Bradford assay (Bradford 1976) using Protein Assay reagent (BioRad). Membranes were then washed in 50 mM HEPES buffer and resuspended in 50 mM HEPES buffer at a total protein concentration of 5 mg/ml. The crude membrane preparation was stored at 4°C and used within a week as the enzyme source in the glycosyltransferase assay.  To assay transfer of GlcNAc from UDP-GlcNAc onto a lipid, the following reaction was mixed: 4 µl crude membrane preparation at 5 mg/ml, 2 µl 0.4 mM UDP-GlcN[14C]Ac (specific activity of 250 mCi/mmol) (American Radiolabeled Chemicals), 2 µl 1% Triton X-100, 2 µl 500 mM HEPES buffer (pH 7.5), 2 µl 5 mM MgCl2. The remaining reaction volume was filled with dH2O to a total volume of 20 µl. The mixed reaction was incubated at 30°C for 1 h, followed by extraction of the lipids into butanol (Ravishankar, Kumar, et al. 2005) by the addition of 100 µl of butanol and 100 µl of water. This mixture was then agitated for 5 s and centrifuged at 2 000 g for 1 min to separate the aqueous and organic phases. Fifty µl of the top organic phase, which contains the extracted lipids, was added to 3 ml scintillation fluid 56  in a scintillation vial, and the degradations per minute (dpm) of each sample was determined by a LS 6000IC Scintillation counter (Beckman Coulter) with the 14C settings and a 5 min count setting. The presence of 14C in the organic phase measures transfer of GlcN[14C]Ac onto a lipid.  2.12 Western blot analysis E. coli cells were harvested from liquid cultures by centrifugation and the resultant cell pellet was resuspended in sample buffer and boiled for 5 min. Samples were run on SDS-PAGE at 180 V for 55 min to separate the proteins followed by transfer of the proteins onto Immobilon-P membrane (Millipore) at 100 V for 60 min on ice (Oliver and Fernandez 2001). The remaining SDS-PAGE gel, after transfer, was stained with Coomassie Brilliant Blue to confirm equal levels of protein were present in comparable samples (Oliver and Fernandez 2001). The Immobilon-P membrane was then probed with rabbit anti-HisTag antiserum, His-Probe (G-18) (Santa Cruz Biotechnology), at 1:10 000 dilution, followed by probing with secondary antiserum horseradish peroxidase (HRP)-conjugated goat anti-rabbit (ICN Biomedicals) at 1:20 000 dilution (Oliver and Fernandez 2001). The HRP-conjugated antibody was detected with Western Lightning® ECL (Perkin Elmer) followed by exposure to X-OMAT LS film (Kodak) for 30 s, as per the manufacturers’ guidelines. PageRuler Prestained protein ladder (Fermentas Life Sciences) was used to estimate molecular mass.  2.13 Whole genome sequencing Library preparation and genome sequencing was performed in the Hirst laboratory. The sequencing data for B. hinzii ATCC 51730 and B. trematum CCUG 13902 were obtained from PCR-free random fragment libraries sequenced on the MiSeq (Illumina, Hayward, CA) platform using indexed paired-end 250-nucleotide (nt) v2 chemistry and resulted in ~700-fold coverage for each genome. The nonindexed read length was 250 nt, with 84.4% of the postfilter paired-end reads having Q30 or greater. The sequence reads were subsampled (~2.2 M reads) and assembled into contigs using Velvet (Zerbino and Birney 57  2008) with a k-mer of 151. A total of 1,850,984/2,212,976 reads were assembled for B. hinzii and 1,878,624/2,212,458 reads were assembled for B. trematum, resulting in 98 contigs for B. hinzii and 83 contigs for B. trematum. These whole-genome shotgun projects have been deposited in GenBank under accession no. AWNM00000000 (B. hinzii) and no. AWNL00000000 (B. trematum).  2.14 Reverse transcriptase PCR RNA was extracted from freshly grown bacterial cells using the Quick-RNA MiniPrep kit (Zymo Research), following the manufacturer’s directions. Any contaminating DNA was removed using the DNA-free kit (Ambion), as per the manufacturer’s instructions. The DNA-free RNA sample was then tested for DNA contamination by PCR using 0.1 µg of the RNA sample as a template and chromosome-specific primers. If the RNA sample was confirmed to be DNA-free by PCR, 1 µg RNA was converted to cDNA using SuperScript II RNase H- Reverse Transcriptase (Invitrogen), in accordance with the manufacturer’s instructions for first-strand cDNA synthesis, with Random Primers (Invitrogen) and the RNaseOUT reagent (Invitrogen). PCR of cDNA was performed with the following specifications: cDNA from the equivalent of 0.1 µg RNA was the template for each reaction, 2 µl 10X ThermoPol® buffer (New England Biolabs), 4 µl 5X GC enhancer (New England Biolabs), 0.4 µl 10 mM dNTP (BioBasic), 1 µl each primer at 10 µM, 0.2 µl Taq DNA Polymerase. The PCR reaction was as follows: 5 min at 94°C, 30 cycles of (45 s at 94°C, 45 s at 63°C, 1 min at 72°C), 5 min at 72°C. For each reverse transcriptase (RT)-PCR, the successful conversion into cDNA was tested with positive control primers (vag8 primers).  The following primer pairs were used to detect transcription of genes in RT-PCR experiments (Table 3): lgmA, BP0399fw1 and BP0399rev1; lgmB, BP0398fw1 and BP0398rev1; lgmC, BP0397-RTfw1 and BP0397-RTrev1; lgmD, BP0396-RTfw1 and BP0396rev1; lgmE, BPlgmDlike-RTfw1 and BPlgmDlike-RTrev1; vag8, vag8fw1 and vag8rev1.  58  2.15 Bioinformatic analysis tools Bioinformatic tools were used to analyze individual sequences, or to compare multiple sequences. For all these web-based bioinformatics tools, I used the default settings. To predict the protein family of a single peptide sequence based on protein sequence similarity, I used Pfam 27.0 (http://pfam.sanger.ac.uk/) (Punta, Coggill, et al. 2012). To analyze a protein sequence for predicted membrane helices, and analyze the predicted membrane topologies of proteins, I used the TMHMM server v 2.0 (http://www.cbs.dtu.dk/services/TMHMM/) (Sonnhammer, von Heijne, et al. 1998). To compare multiple protein sequences and generate a multiple sequence alignment, I used ClustalW2 (http://www.ebi.ac.uk/Tools/msa/clustalw2/) (Thompson, Higgins, et al. 1994). Then, I generated neighbour-joining trees based on the ClustalW2 protein sequence alignments using The Methodes et Algorithmes pour la bio-informatique LIRMM website (http://www.phylogeny.fr/version2_cgi/one_task.cgi?task_type=bionj) (Gascuel 1997). Finally, to compare two nucleotide or protein sequences, I used NCBI Blast2 algorithms (http://blast.ncbi.nlm.nih.gov/Blast.cgi?PAGE_TYPE=BlastSearch&PROG_DEF=blastn&BLAST_PROG_DEF=megaBlast&BLAST_SPEC=blast2seq)  2.16 Statistical analysis Data were analyzed using one-way ANOVA with a Bonferroni post-test to compare groups, unless otherwise specified. GraphPad Prism 5 or GraphPad Prism 6 software was used for all analyses.  59  Chapter 3: Identification and characterization of the lgm locus in Bordetella 3.1 Introduction The lipid A region of B. pertussis LPS can be modified with GlcN moieties at the phosphate groups in the wild-type strain BP338 (Figure 4), which is a nalidixic acid resistant derivative of strain Tohama I. Abrogation of this GlcN modification was mapped to the gene locus tag BP0398 (lgmB) in a GlcN-negative transposon mutant, suggesting lgmB is required for this modification (Marr, Tirsoaga, et al. 2008). Furthermore, LgmB is a homolog of the glycosyltransferase ArnT, which transfers Ara4N onto lipid A in Salmonella (Trent, Ribeiro, et al. 2001). I hypothesized that the genes surrounding lgmB may also be involved in GlcN modification of lipid A in B. pertussis (Figure 6). BP0399 (lgmA), which is found directly upstream of lgmB, is a homolog of another enzyme in the Ara4N modification pathway: ArnC, which transfers Ara4FN onto the lipid carrier C55P (Breazeale, Ribeiro, et al. 2005). Additionally, bioinformatics analysis of BP0397 (lgmC) suggests deacetylase activity and BP0396 (lgmD) is a short protein that primarily consists of four predicted transmembrane helices, a similar structure to ArnE and ArnF, which act as C55P-Ara4N flippases (Imagawa, Iino, et al. 2008, Yan, Guan, et al. 2007). Based on these analyses, I hypothesized that LgmA, LgmB, LgmC, and LgmD are involved in a pathway that modifies the phosphate groups of LPS with GlcN, as summarized in Figure 8. Briefly, I predict LgmA transfer GlcNAc onto the carrier lipid C55P, followed by deacetylation by LgmC to generate C55P-GlcN. In this hypothetical model, C55P-GlcN is then flipped from the cytoplasmic to the periplasmic face of the IM by LgmD, and finally LgmB transfers GlcN onto the phosphate of lipid A.   3.2 The lgm locus is required for lipid A GlcN modification in B. pertussis Locus tags BP0399 to BP0396 had been annotated in the B. pertussis Tohama I genome as putative open reading frames (ORFs) with unknown function (Parkhill, Sebaihia, et al. 2003). Based on transposon mutagenesis of BP0398 (lgmB), our lab originally postulated lgmA, lgmB, lgmC, and lgmD (BP0399 to BP0396, respectively) were part of the lipid A glucosamine modification locus (Marr, Tirsoaga, et al. 60  2008, Shah, Albitar-Nehme, et al. 2013). The first step in characterizing the lgm locus would be to determine whether all four lgm genes are required for GlcN modification of lipid A in B. pertussis. 3.2.1 lgmA, lgmB, and lgmC are required for lipid A GlcN modification The ORFs for lgmA, lgmB, lgmC, and lgmD were all annotated as hypothetical proteins (Parkhill, Sebaihia, et al. 2003), therefore I first set out to confirm all four of these lgm genes are actually expressed in the BP338 strain. Previously, BP0399 and BP0398 (lgmA and lgmB, repectively) were shown to be transcribed in B. pertussis strain BP338 (Marr, Tirsoaga, et al. 2008), and I used RT-PCR to confirm these results, and to show lgmC and lgmD are also transcribed in BP338 (Figure 10).   Since lgmB was disrupted by the insertion of a transposon in B. pertussis BP338 to generate a strain lacking GlcN-modified lipid A, it is possible that insertion of the transposon affected expression of a downstream gene. Therefore, to confirm lgmB is required for GlcN modification of lipid A, and to determine if lgmA, lgmC, and lgmD are also needed for this modification, I generated markerless deletion mutants of each of the individual lgm locus genes and of the full lgm locus. This generated the following mutant strains: BP338LgmAKO, BP338LgmBKO, BP338LgmCKO, BP338LgmDKO, and BP338LgmABCDKO (full lgm locus deletion mutant). The lipid A from these strains was analyzed via negative-ion MALDI mass spectrometry. The presence of the GlcN modification is represented by peaks observed at m/z 1720 (one GlcN at either 1-phosphate or 4′-phosphate) and 1881 (GlcN at both 1-phosphate and 4′-phosphate), i.e. 161 and 322 mass units higher than the major unmodified penta-acyl lipid A species observed at m/z 1559 (Figure 11). Wild type BP338 and the lgmD mutant have GlcN-modified lipid A, whereas the lgmA, lgmB, lgmC, and the full lgmABCD locus mutants lack this modification (Figure 11, Table 7, Appendix A  ). I demonstrated these results were not due to downstream polar effects by complementing each mutant (BP338LgmAKO, BP338LgmBKO, and BP338LgmCKO) with a vector containing the deleted gene (pBBR2LgmA, pPtacLgmAB, and pBBR2LgmC, respectively), and analyzed the lipids A of these strains via negative-ion MALDI mass spectrometry. I found, in each 61  case, that complementation of the missing gene resulted in restoration of the lipid A GlcN modification phenotype (Table 7, Appendix A  ). This suggests that only lgmA, lgmB, and lgmC are required for GlcN modification of lipid A in B. pertussis, but lgmD is not. To further assess the dispensability of lgmD, I complemented the full lgm locus mutant either with lgmABCD or with only lgmABC (vectors pPtacLgmABCD and pPtacLgmABC, respectively) and tested these strains for lipid A GlcN modification in a similar manner. I found GlcN modification was restored in both strains, showing lgmD is not required for modification of lipid A with GlcN.   Previously, a transposon mutant of B. pertussis strain BP338 lacking the lipid A GlcN modification was shown to have decreased hTLR4 activation when compared to wild type BP338 (Marr, Hajjar, et al. 2010). Since I had generated several BP338 strains that lack GlcN-modified lipid A, I set out to confirm the link between GlcN modification and hTLR4 activation. I used the HEK-Blue hTLR4 activation assay to determine the ability of these strains to activate hTLR4 by stimulating HEK cells expressing hTLR4 with heat-killed bacteria (Figure 12). My results showed that strains with the GlcN-modified lipid A (BP338, BP338LgmDKO, and BP338LgmABCDKO complemented with either pPtacLgmABCD or pPtacLgmABC) have approximately 1.5 to 3.5 times higher levels of hTLR4 activity in this assay when compared with strains without the GlcN modification (BP338LgmAKO, BP338LgmBKO, BP338LgmCKO, and BP338LgmABCDKO). The lgmB transposon mutant, which lacks the GlcN modification, had previously been shown to also have decreased levels of hTLR4 activation compared to BP338.   62    Figure 10. Transcription of lgmABCD in BP338 and 18-323 + pPtacLgmABCD Reverse transcriptase PCR (RT-PCR) of B. pertussis BP338 and 18-323 + pPtacLgmABCD (18-323 + pABCD) with vag8 (+), lgmA, lgmB, lgmC, or lgmD primers. Chromosomal DNA from BP338 was used as a positive control for each PCR reaction to indicate the size of the expected product. vag8 is a Bvg+ phase gene. Left-most lane is a DNA ladder in kilobases (kb).    63   Figure 11. Mass spectrometry analysis of BP338 and BP338lgmABCDKO lipid A Structural analysis of lipid A with negative-ion MALDI mass spectrum analysis. Mass spectra of BP338 (A) and BP338LgmABCDKO (B), the full lgm locus deletion mutant. Peaks at m/z 1559 represent penta-acyl lipid A that lack GlcN modification, peaks at m/z 1720 represent penta-acyl lipid A with one GlcN modification at either phosphate group, and peaks at m/z 1881 represent penta-acyl lipid A with a GlcN modification at both phosphate groups. The peaks at m/z 1333 and 1494 represent tetra-acyl species. arb. u., arbitrary units. C and D, the lipid A structures present in BP338 (C) and BP338LgmABCDKO (D) as determined by mass spectral analysis. The numbers at the bottom of the structures indicate the length of the acyl chains. From Shah et al. 2013 (Shah, Albitar-Nehme, et al. 2013), used with permission.  64  Table 7. GlcN modification of BP338 lgm locus mutants Summary of lipid A glycosylation, as determined by negative-ion MALDI mass spectrum analysis. See Appendix A   for mass spectra. pLgmA, pBBR2LgmA; pLgmAB, pPtacLgmAB; pLgmC, pBBR2LgmC; pLgmABCD, pPtacLgmABCD; pLgmABC, pPtacLgmABC. Modified from Shah et al. 2013 (Shah, Albitar-Nehme, et al. 2013), used with permission.  Strain GlcN-modified lipid A BP338 Yes BP338LgmAKO No BP338LgmAKO + pLgmA Yes BP338LgmBKO No BP338LgmBKO + pLgmAB Yes BP338LgmCKO No BP338LgmCKO + pLgmC Yes BP338LgmDKO Yes BP338LgmABCDKO No BP338LgmABCDKO + pLgmABCD Yes BP338LgmABCDKO + pLgmABC Yes     65   Figure 12. hTLR4 activation by BP338 lgm locus mutants hTLR4 activation measured with HEK-Blue NFκB hTLR4 activity assay. Null2 all, stimulation of HEK-Blue Null2 cell line that lacks TLR4 expression with all LPS variants; Blank, medium only with no HEK-Blue cells; Unstim, HEK-Blue hTLR4 cells stimulated with medium only; BP338LgmABCDKO, the full lgm locus deletion mutant; BP338LgmABCDKO + pLgmABCD, full lgm locus deletion mutant complemented with pPtacLgmABCD; BP338LgmABCDKO + pLgmABC, full lgm locus deletion mutant complemented with pPtacLgmABC. Green bars represent strains with GlcN-modified lipid A and red bars represent strains with no GlcN modification (from Table 7). Absorbance at 650 nm (A650). Graph shows the results of one representative experiment of three, n = 6 replicates per experiment. One-way ANOVA with a Bonferroni post-test used for statistical analysis. p values: < 0.001 (***). Modified from Shah et al. 2013 (Shah, Albitar-Nehme, et al. 2013), used with permission.     66  3.2.2 Potential flippase replacements are not required for GlcN modification In my initial model (Figure 8), I hypothesized the need for a flippase to translocate C55P-GlcN from the cytoplasmic to the periplasmic face of the IM, since LgmA and LgmC are predicted to have globular domains in the cytoplasm, and LgmB, which primarily consists of predicted transmembrane helices, is predicted to have a short periplasmic C-terminal region (Figure 7). I had originally hypothesized that LgmD functions as this C55P-GlcN flippase, due to the similarities in predicted membrane topology between LgmD and ArnE and ArnF, which function to flip C55P-Ara4N from the cytoplasmic to the periplasmic face of the IM (Yan, Guan, et al. 2007). However, the lgmD mutant strain (BP338LgmDKO) still has GlcN-modified lipid A, suggesting LgmD is not required for the Lgm pathway. Alternatively, if LgmD does act as a flippase, another protein may compensate for the role of LgmD in the lgmDKO strain, therefore resulting in GlcN-modified lipid A, even in the absence of lgmD. ArnE and ArnF are small proteins that are made up of four transmembrane helices, just as lgmD is (Figure 7). Therefore, I searched for similar proteins in the B. pertussis Tohama I genome that may function as a C55P-GlcN flippase in the Lgm pathway.  Upon closer inspection of the lgm locus in B. pertussis, I identified a previously unannotated ORF, which I have named lgmE. lgmE is found within the original lgm locus, but in the opposite direction compared to lgmABCD, and it overlaps with lgmC and lgmB (Figure 14A). LgmE is predicted to contain four transmembrane helices (Figure 13A) and is a member of the GtrA-like Pfam family, members of which are predicted to function as flippases (Guan, Bastin, et al. 1999). Therefore, I hypothesized that LgmE may function either as a possible C55P-GlcN flippase replacement of LgmD in the Lgm pathway, or simply fulfills this role in the Lgm pathway, regardless of LgmD. Since lgmE had not previously been described, I first demonstrated that lgmE is transcribed in B. pertussis strain BP338 using RT-PCR with lgmE-specific primers (Figure 14B). Next, I tested whether LgmE is required for GlcN modification in B. pertussis. Since lgmE overlaps with both lgmB and lgmC (Figure 14A), deleting the lgmE gene would 67  also disrupt lgmB and lgmC. Therefore, instead of knocking out lgmE, I generated a single base pair mutant of lgmE on the plasmids pPtacLgmABCD and pPtacLgmABC such that the residue Gly15 of lgmE was mutated to a stop codon and this mutation in the overlapping lgmC was a silent mutation (Figure 15). This generated the plasmids pPtacLgmABCD-EG15STOP, which encodes for lgmABCD with a mutated lgmE, and pPtacLgmABC-EG15STOP, which contains lgmABC and a mutated lgmE. I then complemented the full lgm locus mutant, BP338LgmABCDKO, with these plasmids to generate a strain that either lacks only lgmE expression (BP338lgmABCDKO + pPtacLgmABCD-EG15STOP) or lacks both lgmD and lgmE expression (BP338lgmABCDKO + pPtacLgmABC-EG15STOP). The results showed these two strains transcribe the expected lgm genes by RT-PCR with primers specific to each lgm gene: BP338lgmABCDKO + pPtacLgmABCD-EG15STOP transcribes all five lgm genes whereas BP338lgmABCDKO + pPtacLgmABC-EG15STOP does not transcribe lgmD (Figure 16). Since the Gly15STOP mutation in lgmE stops production of LgmE at the translational level, lgmE is still expected to be transcribed. I tested these strains for the ability to activate hTLR4 using the HEK-Blue hTLR4 activation assay, and found both strains had wild-type levels of hTLR4 activation (Figure 17). Thus, both lgmE mutant strains have GlcN-modified lipid A, so LgmE is not required for the Lgm pathway.    68   Figure 13. Predicted topologies of BP1945 and LgmE A) LgmE, B) BP1945. Membrane topologies were predicted with TMHMM (Sonnhammer, von Heijne, et al. 1998). The amino acid position is along the x-axis, the probability of a tramsmembrane helix is along the y-axis (red vertical bars). For the horizontal lines: red is transmembrane region, blue is cytoplasmic region, pink is periplasmic region. The diagrams on the right depict a visual representation of the TMHMM prediction. The upper horizontal black line represents the periplasmic leaflet of the IM, the lower black line represents the cytoplasmic leaflet of the IM, and the thick grey line represents the protein. 69    Figure 14. Current schematic of the B. pertussis lgm locus and transcription of lgmE in BP338 A) B. pertussis lgm locus consists of lgmA, lgmB, lgmC, lgmD, and lgmE (previously referred to as lgmD-like (Novikov, Shah, et al. 2013)). B) Reverse transcriptase PCR (RT-PCR) of B. pertussis BP338 with lgmE primers. Control to confirm no DNA contamination of total RNA preparation from BP338: lane 1 – PCR of BP338 chromosomal DNA with lgmA primers, lane 2 – PCR of RNA purification from BP338 with lgmA primers. Control to show successful conversion of RNA to cDNA: lane 3 – PCR of BP338 cDNA with lgmA primers. Test for lgmE transcription: lane 4 – PCR of PtaclgmABCD DNA with lgmE primers, lane 5 – PCR of BP338 cDNA with lgmE primers. Left-most lanes are DNA ladders in kilobases (kb). Band in lane 5 indicates B. pertussis strain BP338 transcribes lgmE.     70   Figure 15. Schematic of the lgmE mutant Gly15STOP mutation in lgmE results from changing ‘g’ to ‘t’. This same mutation in the overlapping lgmC results in a ‘c’ to ‘a’ mutation, which is silent (no change in the serine residue). Mutations are highlighted in red.   71   Figure 16. Transcription of lgm locus genes in BP338 lgmEKO strains Reverse transcriptase PCR (RT-PCR) of B. pertussis BP338LgmABCDKO + pPtacLmgABCD-EG15STOP (lgmABCDKO strain complemented with lgmABCD and a mutated lgmE) and BP338LgmABCDKO + pPtacLgmABC-EG15STOP (lgmABCDKO strain complemented with lgmABC and a mutated lgmE) with lgmA, lgmB, lgmC, lgmD, and lgmE -specific primers (lanes labeled A, B, C, D, and E, respectively). Left-most markings (1.0 and 0.5) are size markers in kilobases (kb). BP338ABCDKO is BP338LgmABCDKO.    72   Figure 17. hTLR4 activation by BP338 lgmEKO strains hTLR4 activation measured with HEK-Blue NFκB hTLR4 activity assay. Null2 all, stimulation of HEK-Blue Null2 cell line that lacks TLR4 expression with all LPS variants; Blank, medium only with no HEK-Blue cells; Unstim, HEK-Blue hTLR4 cells stimulated with medium only; BP338LgmABCDKO, the full lgm locus deletion mutant; BP338LgmDKO, full lgmD deletion mutant; + pLgmABCD, complemented with pPtacLgmABCD vector; + pLgmABC, complemented with pPtacLgmABC vector; + pLgmABCD-EG15STOP, complemented with pPtacLgmABCD-EG15STOP vector that has the lgmE G15STOP mutation; + pLgmABC-EG15STOP, complemented with pPtacLgmABC-EG15STOP vector that has the lgmE G15STOP mutation. Absorbance at 650 nm (A650). Graph shows the results of one representative experiment of three, n = 4 replicates per experiment. One-way ANOVA with a Bonferroni post-test used for statistical analysis. p values: < 0.001 (***), no significant difference (ns).     73  A second possible candidate for a replacement flippase for lgmD is locus tag BP1945. BP1945 consists primarily of two predicted transmembrane helices (Figure 13A) and is found adjacent to the Bps locus, which is involved in poly-GlcNAc synthesis. To determine if BP1945 is involved in GlcN-modification of lipid A in B. pertussis strain BP338, I generated clean, markerless deletion mutants of BP1945 in both wild type BP338 and in the lgmD mutant to generate the strains BP338BP1945KO and BP338LgmDKOBP1945KO, respectively. I then tested the ability of these strains to activate hTLR4 using the HEK-Blue hTLR4 activation assay to determine if these strains had GlcN-modified lipid A, since strains lacking the modification have lower levels of hTLR4 activation. Both of these BP1945KO strains activated hTLR4 to the same level as wild type BP338 (Figure 18), demonstrating the BP1945 mutant strains have GlcN-modified lipid A. As such, BP1945 does not replace the activity of LgmD in the lgmD mutant and is not required for GlcN modification in B. pertussis.   74   Figure 18. hTLR4 activation by BP338 BP1945KO strains  hTLR4 activation measured with HEK-Blue NFκB hTLR4 activity assay. Null2 all, stimulation of HEK-Blue Null2 cell line that lacks TLR4 expression with all LPS variants; Blank, medium only with no HEK-Blue cells; Unstim, HEK-Blue hTLR4 cells stimulated with medium only; BP338LgmABCDKO, the full lgm locus deletion mutant; BP338LgmDKO, full lgmD deletion mutant; BP338BP1945KO, BP1945 locus tag deletion mutant; BP338LgmDKOBP1945KO, lgmD and BP1945 locus tag double deletion mutant. Absorbance at 650 nm (A650). Graph shows the results of one representative experiment of three, n = 4 replicates per experiment. One-way ANOVA with a Bonferroni post-test used for statistical analysis. p values: < 0.001 (***), no significant difference (ns).     75   3.3 LgmA functions as a GlcNAc transferase Thus far, I had established LgmA, LgmB, and LgmC are required for GlcN modification of B. pertussis lipid A. LgmA is a homolog of the glycosyltransferase ArnC and is a predicted inner membrane protein, due to two predicted transmembrane helices (Figure 7). Since ArnC functions to transfer Ara4FN from UDP-Ara4FN to the inner membrane carrier lipid C55P in the Salmonella and E. coli Ara4N lipid A modification pathway (Raetz, Reynolds, et al. 2007), I predicted LgmA functions in a similar manner to transfer GlcNAc from UDP-GlcNAc to C55P in B. pertussis (Figure 8). To test the function of LgmA, I performed a glycosyltransferase assay using the radioactively-labelled substrate, UDP-GlcN[14C]Ac, and crude membrane preparations from E. coli strains as the source of both LgmA and the acceptor lipid. The E. coli strains either expressed histidine-tagged LgmA (BL21 + pLgmA) or contained an empty vector control (BL21 + pET30). Immunoblot detection with an anti-HisTag antibody showed histidine-tagged LgmA (LgmA-His) was expressed in BL21 + pLgmA after induction with 1 mM IPTG (band at ~40 kDa) but not expressed in the empty-vector control (Figure 19). Furthermore, LgmA-His localized to the membrane preparation of the sample, which was used as the source of the enzyme in the glycosyltransferase assay (Figure 19). The lipids from each glycosyltransferase reaction were extracted using butanol, and this lipid fraction was measured for radioactivity, which would indicate the transfer of the radioactive substrate, UDP-GlcN[14C]Ac, to a lipid. Negative control reactions, with either no UDP-GlcN[14C]Ac or no membrane preparation, and therefore no enzyme source, had very low levels of radioactivity in the lipid fraction, therefore indicating no transfer of GlcN[14C]Ac to a lipid acceptor in these reactions. The reaction with the membrane preparation from BL21 + pET30 as the enzyme source had higher than background levels of radioactivity in the lipid fraction, indicating background levels of GlcN[14C]Ac transfer to lipids, likely from glycosyltransferases that naturally reside in E. coli membranes. However, reactions with LgmA-His-containing membrane preparations (BL21 + pLgmA) had a 40% to 60% increase in radioactivity of the lipid fraction when compared to BL21 + pET30 (Figure 76  20). This indicates the presence of LgmA increases transfer of GlcN[14C]Ac to lipids in E. coli, and therefore, LgmA likely acts as a GlcNAc glycosyltransferase.    77   Figure 19. Expression of LgmA-His in E. coli  Western transfer and immunoblot detection with anti-HisTag antibody of E. coli BL-21 (DE3) cells harboring an empty vector plasmid (BL21 + pET30) or an LgmA-His expressing vector (BL21 + pLgmA). Induced cells were induced with 1 mM IPTG for 24 hours. 1/10 membr prep is a 1/10 dilution of the crude membrane preparation that is at 5 mg/ml total protein concentration. Orange arrow indicates LgmA-His protein band. Immunoblot shows one representative experiment of three, with similar results.    78       Figure 20. LgmA glycosyltransferase assay Transfer of GlcN[14C]Ac to the lipid fraction as measured by degradations per minute (dpm) of butanol-extracted lipids from each reaction. Blank, no reaction added; no UDP-GlcN[14C]Ac, control reaction with no radioactive substrate (UDP-GlcN[14C]Ac) but contains BL21 + pLgmA membrane preparation; no membrane, control reaction with no membrane preparation; BL21 + pET30, reaction with empty vector membrane preparation; BL21 + pLgmA, reaction with LgmA-His-containing membrane preparation. Graph shows the results of one representative experiment of three, n = 3 replicates for BL21 + pET30 and BL21 + pLgmA experiments. Student’s t-test used for statistical analysis to compare BL21 + pET30 and BL21 + pLgmA reactions. p value < 0.001 (***).    79  3.4 Identification of the putative active site of LgmA LgmA is predicted to be a member of the GT2 family of glycosyltransferases. Members of this family are predicted to have a conserved DXD motif, which may be involved in the coordination of divalent cations in the active site (Lairson, Henrissat, et al. 2008). To identify a putative active site of LgmA, I aligned LgmA with two GT2 family homologs: ArnC and Ftn_0545. ArnC functions in the Salmonella Ara4N lipid A modification pathway to transfer Ara4FN onto the carrier lipid C55P (Breazeale, Ribeiro, et al. 2005) and, similarly, Ftn_0545 functions to transfer N-acetyl-galactosamine (GalNAc) to C55P in the Francisella novicida lipid A-modification pathway (Song, Guan, et al. 2009). I identified the following conserved residues that I hypothesized to be involved in LgmA function: D76, D77, D127, and D129; the last two residues appear to be a possible conserved DXD motif (Figure 21A). These residues were mutated by Andrew Low to glycine in the construct pBBR2LgmA to generate the following mutants: D76G D77G, D127G, D129G, and D127G D129G. To test if these mutations affect the function of LgmA, the lgmA mutant, BP338LgmAKO, was complemented with either wild type lgmA (pBBR2LgmA) or mutated lgmA and RT-PCR was used to confirm these complemented strains all transcribed lgmA (Figure 22 insert). These complemented strains were shown to have GlcN modified-lipid A by stimulation of HEK-Blue hTLR4 cells with heat-killed bacteria to assess hTLR4 activation (Figure 22). The lgmAKO strain was found to have 28 to 51% lower levels of hTLR4 activation compared to wild type BP338, but this phenotype was partially rescued by complementation with wild type lgmA (+ plgmA), which had 36 to 73% higher levels of hTLR4 activation compared to lgmAKO. Complementation of the lgmAKO strain with the lgmA mutants D76G D77G, D127G, D129G, D127G D129G resulted in no recovery of hTLR4 activation, therefore these strains have unmodified lipid A, suggesting these amino acids may be required for LgmA activity.   80   Figure 21. Identifying conserved residues in LgmA A) ClustalW2 (Thompson, Higgins, et al. 1994) alignment of B. pertussis Tohama I LgmA, F. novicida U112 Ftn-0545, and E. coli K-12 ArnC. Residues in red text and below a red arrow were targeted for mutational analysis. ‘*’: Identical amino acid residues; ‘:’: conserved amino acid residues; ‘.’: semi-conserved amino acid residues.  B) ClustalW2 (Thompson, Higgins, et al. 1994) alignment of LgmA from Bordetella species: Bpe, B. pertussis Tohama I; Bbr, B. bronchiseptica RB50; Bpa, B. parapertussis 12822; Bav, B. avium 197N; Bhi, B. hinzii ATCC 51730; and Btr, B. trematum CCUG 13902. Bolded sequences (B. hinzii and B. trematum) have no lipid A GlcN-modification whereas all other species listed have the lipid A GlcN-modification. Residues in green text and below a green arrow are conserved in all GlcN-modification strains, but variable in non-GlcN-modifying strains and these residues were targeted for mutational analysis.    81   Figure 22. hTLR4 activation by LgmA mutants hTLR4 activation measured with HEK-Blue NFκB hTLR4 activity assay. Null2 all, stimulation of HEK-Blue Null2 cell line that lacks TLR4 expression with all LPS variants; Blank, medium only with no HEK-Blue cells; Unstim, HEK-Blue hTLR4 cells stimulated with medium only; BP338LgmAKO, lgmA deletion mutant; + plgmA, complemented with pBBR2LgmA. Absorbance at 650 nm (A650). Graph shows the results of one representative experiment of three, n = 4 replicates per experiment. One-way ANOVA with a Bonferroni post-test used for statistical analysis. p values: < 0.0001 (****), < 0.001 (***), no significant difference (ns). Bars with p value indicator directly above are in comparison with BP338LgmAKO + plgmA. Inset shows transcription of lgmA in each strain using RT-PCR and lgmA-specific primers. DNA ladder sizes (left-most lane), from top to bottom: 1.0, 0.75, 0.5, 0.25 kilobases (kb). Inset shows one representative experiment of three.  82  Since the structure of LgmA is unknown, I generated a hypothetical structure of LgmA using the Protein Homology/analogy Recognition Engine (Phyre) protein fold recognition server (Kelley and Sternberg 2009) to assess the position of these residues on a predicted 3-dimenstional model (accessed on December 2, 2013). The highest ranked Phyre-predicted model for LgmA covered amino acids 24 to 345 of the 352 residue long peptide, with an estimated precision of 100% and an E value of 1.3 x 10-21. The model was based upon the structure of the Bacteroides fragilis strain NCTC 9343 protein BF2801 (PDB ID 3BCV), which is a predicted glycosyltransferase. I located the amino acid residues that, when mutated, result in a lack of LgmA activity on this structural model (Figure 23) and found these residues are clustered in one region of the predicted LgmA structure. I hypothesize the region of the predicted LgmA structure where D76, D77, D127, and D129 cluster may function as the active site.    83   Figure 23. Phyre predicted model of the structure of LgmA Model of LgmA structure by the Protein Homology/analogy Recognition Engine (Phyre) (Kelley and Sternberg 2009). The predicted transmembrane helix region (Figure 7) is green, the remainder of the protein is cyan. Residues that, when mutated, resulted in a lack of GlcN modification of lipid A are red, other mutated residues, that did not affect lipid A GlcN modification, are yellow. PyMOL was used to visualize this structure.     84  3.5 Identification of the putative active site of LgmC In the hypothetical model (Figure 8), after LgmA transfers GlcNAc onto C55P, I predicted LgmC functioned to deacetylate C55P-GlcNAc to generate C55P-GlcN. Recently, LgmC was shown to function as I had originally predicted, to remove the acetyl group from C55P-GlcNAc in B. bronchiseptica (Llewellyn, Zhao, et al. 2012). LgmC is predicted to be a member of the YdjC-like protein family by Pfam. To identify putative active site residues, I aligned LgmC with the following homologs to find conserved residues: YdjC, Ftn_0544, and TTHB209, all three of which are also members of the YdjC-like family. YdjC is hypothesized to be involved in the cleavage of cellobiose-phosphate, Ftn_0544 functions to deacetylate C55P-GalNAc in the F. novicida lipid A modification system, and TTHB209 is structurally similar to SpPgdA, a GlcNAc deacetylase in Streptococcus pneumoniae (Imagawa, Iino, et al. 2008, Llewellyn, Zhao, et al. 2012, Song, Guan, et al. 2009). Based on the ClustalW2 alignment of these proteins (Figure 24), I hypothesized the conserved residues D80, D81, H130, D187, H189, and E313 are required for LgmC activity, since these residues are conserved amongst the aforementioned YdjC-like family proteins. These residues were mutated in the plasmid pBBR2LgmC to generate the following mutants: D80G D81G, H130G, D187G H189G, and E313G. Then the lgmC mutant strain, BP338LgmCKO, was complemented with wild type lgmC (+ plgmC) or the lgmC mutant plasmids and these strains were shown to transcribed lgmC by RT-PCR with lgmC-specific primers (Figure 25 insert). To determine if these mutations in LgmC affect function, the ability of the complemented lgmCKO strains to activate hTLR4 was tested, thereby assessing the presence of the lipid A GlcN modification in these different strains. The lgmCKO strain had 27 to 54% lower levels of hTLR4 activation compared to the BP338 wild type (Figure 25), indicating a lack of GlcN-modified lipid A, as previously confirmed by mass spectrometry analysis (Section 3.2.1). Complementation with wild-type lgmC (+ lgmC) resulted in a partial rescue of hTLR4 activation, such that it was 10 to 40% greater than lgmCKO hTLR4 activation, but is was still lower than wild type levels of hTLR4 activation, suggesting lower-than-wild type-levels of GlcN modification (Figure 25). However, complementation with the lgmC mutant plasmids did not rescue 85  hTLR4 activation (Figure 25), showing these strains did not have GlcN-modified lipid A. This suggests the mutated amino acid residues are required for LgmC function, since LgmC is required for GlcN modification of lipid A.  86                                           Figure 24. Identifying conserved residues in LgmC ClustalW2 (Thompson, Higgins, et al. 1994) alignment of B. pertussis Tohama I LgmC, Enterobacter cloacae EcWSU1 YdjC, F. novicida U112 Ftn_0544 , and Thermus thermophilus HB8 TTHB029. Residues in red text and below a red arrow were targeted for mutational analysis. ‘*’: Identical amino acid residues; ‘:’: conserved amino acid residues; ‘.’: semi-conserved amino acid residues.    87   Figure 25. hTLR4 activation by LgmC mutants hTLR4 activation measured with HEK-Blue NFκB hTLR4 activity assay. Null2 all, stimulation of HEK-Blue Null2 cell line that lacks TLR4 expression with all LPS variants; Blank, medium only with no HEK-Blue cells; Unstim, HEK-Blue hTLR4 cells stimulated with medium only; BP338LgmCKO, lgmC deletion mutant; + plgmC, complemented with pBBR2LgmC. Absorbance at 650 nm (A650). Graph shows the results of one representative experiment of three, n = 4 replicates per experiment. One-way ANOVA with a Bonferroni post-test used for statistical analysis. p values: < 0.0001 (****), < 0.001 (***). Bars with p value indicator directly above are in comparison with BP338LgmCKO + plgmC. Inset shows transcription of lgmC in each strain using RT-PCR and lgmC-specific primers. DNA ladder sizes (left-most lane), from top to bottom: 1.0, 0.75, 0.5, 0.25 kilobases (kb). Inset shows one representative experiment of three.   88  Similarly to LgmA, the structure of LgmC is not yet solved, though LgmC is predicted to be a cytoplasmic protein (Figure 7). To analyze the possible positions of these residues, especially in relation to one another, I generated a hypothetical structure for LgmC via the Phyre protein fold recognition server. The structure with the highest estimated precision and lowest E value (100% and 9.1 x 10-26, respectively), was based on the previously solved structure of Eterococcus faecalis strain V583 protein EF_3048, which is a putative enzyme involved in cellobiose-phosphate cleavage (PDB ID 2I5I). This hypothetical model covers 100 % of the LgmC peptide, and predicts a primarily alpha helical secondary structure along with some beta sheet structure (Figure 26) (Kelley and Sternberg 2009). Mutation of the conserved amino acids (D80G D81G, H130G, and E313G) resulted in a lack of LgmC function, and these residues are located in the vicinity of one another and around the same cavity on the predicted structure of LgmC. I hypothesize this region may function as the active site of LgmC.    89   Figure 26. Phyre predicted model of the structure of LgmC Model of LgmC structure by the Protein Homology/analogy Recognition Engine (Phyre) (Kelley and Sternberg 2009). LmgC protein is green. Residues that resulted in a lack of GlcN modification of lipid A when mutated are in red. PyMOL was used to visualize this structure.    90  3.6 The lgm locus in other Bordetella species Thus far, I have analyzed the lgm locus of B. pertussis strain BP338, which is a derivative of the Tohama I strain, and demonstrated lgmA, lgmB, and lgmC are required for GlcN modification of lipid A in this strain. However, analysis of the lipid A structures of additional Bordetella species has revealed that other Bordetella also have GlcN-modified lipid A. Specifically, B. bronchispetica, B. parapertussis, and B. avium have been shown to have the lipid A GlcN modification, whereas B. hinzii and B. trematum do not (Novikov, Shah, et al. 2013). I hypothesized the lack of GlcN modification in B. hinzii and B. trematum was because of a clear genetic reason, such as the lack of the lgm locus in these species. However, the genome sequences of B. hinzii and B. trematum were not publically available at the time. I attempted to sequence the lgm locus in these species with primers specific to the B. pertussis or B. avium lgm loci. However, I was unsuccessful in my efforts, and it later became clear that this is likely due to the low degree of nucleotide sequence similarity of the Bordetella lgm loci, despite conserved amino acid residues in some regions of the Lgm proteins. Therefore we sequenced these genomes (strains B. hinzii ATCC 51730 and B. trematum CCUG 13902) on a MiSeq (Illumina) platform. Sequencing resulted in ~700-fold coverage for each genome and assembly with Velvet algorithms resulted in 98 contigs for B. hinzii and 83 contigs for B. trematum (Shah, Moksa, et al. 2013). I then used the tblastn algorithm, which searches a translated nucleotide sequence with a protein sequence query, to search the B. hinzii and B. trematum genomes, and the other Bordetella species, for the LgmABCDE proteins. Figure 27 summarizes the results of this analysis. I note that lgmA, lgmB, lgmC, lgmD, and lgmE were present in the other subspecies, B. bronchiseptica and B. parapertussis.   B. avium, alternatively, contains the three required lgm genes: lgmABC, along with lgmE (locus tag BAV2926), but no lgmD. B. avium has GlcN-modified lipid A, suggesting a fully functional set of lgmABC genes (Novikov, Shah, et al. 2013). Further analysis of the B. avium lgm locus revealed the LgmC homolog, BAV2925, is annotated to have a truncated N-terminus compared to the other Bordetella 91  LgmC proteins (Figure 28). This truncated LgmC would no longer encode conserved amino acids in the N-terminal region that are likely required for LgmC function, specifically the residues equivalent to B. pertussis LgmC D80 and D81 (see Section 3.5). However, there are methionine residues present upstream of the predicted start site of B. avium lgmC that may act as a start codon for this gene (Figure 28). If the B. avium lgmC gene begins translation at one of these upstream methionine residues, the aforementioned conserved D80 and D81 residues would be expressed at the N-terminus of the protein, which would agree with the presence of a functional LgmC, as required for the GlcN modification.  Analyzing the B. hinzii and B. trematum genomes revealed the presence of homologs of lgmABC in both species, which are the only genes that have thus-far been shown to be required for GlcN modification of lipid A. Therefore, this reflected the hypothesis that the unmodified lipid A produced by these species is not due to the lack of an lgm locus.. Furthermore, the residues I had previously identified as putative active sites for LgmA and LgmC were also conserved amongst the Bordetella species genomes I analyzed. B. hinzii has lgmE in a similar position as in the other Bordetella species, though lgmD is found elsewhere in the genome. B trematum does not possess lgmE, but like in B. hinzii, lgmD is found at a different location on the chromosome. Earlier in this chapter I have shown a B. pertussis strain BP338 that lacks LgmD and LgmE still had the GlcN modification, so the lack of lgmE in B. trematum is unlikely to be the cause of unmodified lipid A. Next, I aligned the LgmA proteins of the different Bordetella species to identify amino acids residues that are conserved amongst the species that have GlcN-modified lipid A (B. pertussis, B. bronchiseptica, B. parapertussis, and B. avium) but are more variable in B. hinzii and B. trematum, which do not have this modification. I hypothesized differences in these residues may be the reason why B. hinzii and B. trematum lack the GlcN modification. I identified the B. pertussis LgmA residues D159 and W163 (Figure 21B) and mutated these residues to the corresponding residues found in B. hinzii to generate the LgmA mutants D159N, W163R, and D159N W163R on a vector. I then complemented BP338LgmAKO with these mutated lgmA genes and tested the 92  ability of these complemented strains to activate hTLR4 to assess the presence or absence of the lipid A GlcN modification. The D159N, W163R, and D159N W163R mutant lgmA successfully complemented lgmAKO, as evident by the rescue of hTLR4 activation levels, showing these amino acids are not essential for LgmA activity (Figure 22). These results suggest the differences between B. hinzii and B. trematum LgmA at these residues are not the reason for the lack of GlcN-modified lipid A in these species. I also mapped the positions of these residues onto the Phyre-predicted model of LgmA, and found they are predicted to be located elsewhere on the structure compared to the previously identified putative active site residues (Figure 23). Therefore, I have not found a clear genetic reason for the lack of GlcN-modified lipid A in B. hinzii and B. trematum.     93    Figure 27. lgm locus of the sequenced Bordetella species Schematic representation of the lgm locus in Bordetella species: B. pertussis Tohama I, B. bronchiseptica RB50, B. parapertussis 12822, B. avium 197N and ATCC 35086, B. hinzii ATCC 51730, and B. trematum CCUG 13902. lgmA is orange, lgmB is yellow, lgmC is green, lgmD is blue and lgmE is diagonally-striped in light blue. Double diagonal black line represents the following region is in a different location of the chromosome. The directionality of lgmD in B. trematum is unknown in reference to lgmABC. Locus tags are given for B. bronchiseptica RB50, B. parapertussis 12822 and B. avium 197N; the open reading frame for lgmE has not been assigned a locus tag and therefore remains unlabeled. Modified from Novikov et al. 2013(Novikov, Shah, et al. 2013), used with permission.   94   Figure 28. Comparison of the start of LgmC between Bordetella species ClustalW2 (Thompson, Higgins, et al. 1994) alignment of LgmC protein sequence of Bordetella species: Bpe, B. pertussis Tohama I; Bbr, B. bronchiseptica RB50; Bpa, B. parapertussis 12822; Bav, B. avium 197N and ATCC 35086; Bhi, B. hinzii ATCC 51730; and Btr, B. trematum CCUG 13902. B. avium LgmC: annotated sequence is black, the upstream region (translated in frame) is red, X highlighted in red is a stop codon, red ‘M’s highlighted in blue are upstream methionine codons that could act as start codons. ‘DD’ highlighted in green are conserved aspartic acid residues. ‘*’: Identical amino acid residues; ‘:’: conserved amino acid residues; ‘.’: semi-conserved amino acid residues.  95  To explore the relationship between LgmABCDE proteins from the Bordetella species I generated neighbour-joining trees based on ClustalW2 alignments (Figure 29). For the LgmC tree, I used the longest possible B. avium LgmC protein, as depicted in Figure 28, though this is still truncated compared to the other Bordetella LgmC proteins. The three subspecies, B. pertussis, B. bronchiseptica, and B. parapertussis, have very closely related Lgm proteins compared to the other Bordetella species. B. avium, B. hinzii, and B. trematum cluster in a branch separate from the three Bordetella subspecies. However, regarding LgmA, LgmB, and LgmC, the three Lgm proteins shared by B. avium, B. hinzii, and B trematum, the two bird-associated species, B. hinzii and B. avium, cluster separately compared to B. trematum. Therefore, even though B. hinzii Lgm proteins are more similar to B. avium compared to B. trematum, B. avium has GlcN-modified lipid A, and therefore a functional Lgm pathway, whereas B. hinzii and B. trematum do not (Novikov, Shah, et al. 2013). LgmD and LgmE are both four-transmembrane helix proteins (Figure 7, Figure 13), and have very low sequence identity between the different Bordetella species (Appendix B  ). The LgmD and LgmE proteins cluster with themselves (Figure 29F), supporting my annotation of the ORFs in the three subspecies as the lgm genes and in the sequenced species B. hinzii and B. trematum, even though B. trematum LgmD appears to be distantly related to the other LgmD proteins (Figure 29D).     96  A. LgmA                                                                                               +-------------------------------------------------------------------------------------------------------------ArnC  |  |  |                                                                                                    +--------------Bav  |                                                                                          +---------+  |                                                                                          |         +----------Bhi  |                                                                                 +--------+  |                                                                                 |        +----------------------Btr  |                                                                                 |  +---------------------------------------------------------------------------------+                                                                                    |                       +Bpa                                                                                    |                      ++                                                                                    |                      ||                                                                                    +----------------------++Bpe                                                                                                           |                                                                                                           +Bbr   B. LgmB                                                                                                   +----------------------------------------------------------------------------------------------------------------ArnT  |  |  |                                                                                                  +-----------Bav  |                                                                                         +--------+  |                                                                                         |        +----------Bhi  |                                                                                   +-----+  |                                                                                   |     +-------------------Btr  |                                                                                   |  +-----------------------------------------------------------------------------------+                                                                                      |                              +Bbr                                                                                      |                              |                                                                                      |                              |                                                                                      +------------------------------+Bpe                                                                                                                     |                                                                                                                     +Bp   C. LgmC  +---------------------------------------------------------------------------------------------------------Ftn  |  |  |                                                                                                                 +Bpa  |                                                        +--------------------------------------------------------+  |                                                        |                                                        |Bbr  |                                                        |                                                        |  |                                                        |                                                        +-Bpe  |                                                        |  +--------------------------------------------------------+                                                           |                   +----------------------Bav                                                           |    +--------------+                                                           |    |              |                                                           +----+              +------------------Bhi                                                                |                                                                +---------------------------------Btr     97  D. LgmD  +-----------------------------------------------------------------------LgmEBav  |  |  |                                                         +---------Bhi  |                                                      +--+  |                                                      |  +---------------------------------------------------------Btr  |                                                      |  +------------------------------------------------------+                                                         |  +Bpe                                                         |  |                                                         +--+Bbr                                                            |                                                            |                                                            +Bpa  E. LgmE  +----------------------------------------------------------------------------------------------------------------GtrA  |  |  |                                                                                                   +---------------Bav  |                                                                                              +----+  |                                                                                              |    +------Bhi  |                                                                                              |  +----------------------------------------------------------------------------------------------+                                                                                                 |                +Bbr                                                                                                 |                |                                                                                                 +----------------+Bpa                                                                                                                  |                                                                                                                  |                                                                                                                  +Bpe  F. LgmD and LgmE                                                  +D-Bbr                                                  |                                                  |-D-Bpa                                            +-----+                                            |     |                                            |     +D-Bpe  +-----------------------------------------+  |                                         |      +--D-Bhi  |                                         +------+  |                                                +----------------------------------------------------------------D-Btr  |  |  |                                                           +---E-Bav  |                                                       +---+  |                                                       |   +----E-Bhi  +-------------------------------------------------------+                                                          |   +E-Bbr                                                          |   |                                                          +---+                                                              |E-Bpa                                                              |                                                              +E-Bpe   Figure 29. Neighbour-joining trees of the lgm locus genes in Bordetella species Neighbour-joining trees of: A) LgmA, B) LgmB, C) LgmC, D) LgmD, E) LgmE, F) LgmD and LgmE. Bpe, B. pertussis Tohama I; Bbr, B. bronchiseptica RB50; Bpa, B. parapertussis 12822; Bav, B. avium 197N; Bhi, B. hinzii ATCC 51730; and Btr, B. trematum CCUG 13902. A) ArnC from E. coli K-12 rooted the LgmA tree. B) ArnT from E. coli K-12 rooted the LgmB tree. C) Ftn, Ftn_0544 from F. novicida U112 rooted the LgmC tree. D) LgmEBav, B. avium LgmE rooted the LgmD tree. E) GtrA from E. coli K-12 rooted the the LgmE tree. F) prefix ‘D-’ indicates LgmD, prefix ‘E-’ indicates LgmE. Neighbour-joining trees were built based on ClustalW2 (Thompson, Higgins, et al. 1994) alignments of the protein sequences (see Appendix B   for ClustalW2 alignments) using the Methodes et Algorithmes pour la bio-informatique LIRMM website (http://www.phylogeny.fr/version2_cgi/one_task.cgi?task_type=bionj) (Gascuel 1997). 98  3.7 Discussion I found that lgmA, lgmB, and lgmC are required for the modification of B. pertussis lipid A with GlcN, while lgmD, lgmE, and locus tag BP1945 are not. Mutational analysis of LgmA and LgmC identified residues D76, D77, D127, and D129 in LgmA and D80, D81, H130, D187, H189, and E313 in LgmC that may be part of a putative active site in these enzymes. Furthermore, I did not uncover any clear genetic reason for the lack of lipid A GlcN modification in B. hinzii and B. trematum.  I also demonstrated LgmA transfers GlcNAc onto a molecule that localizes to the lipid fraction. LgmC has recently been shown to function as a deacetylase by removing the acetyl group from C55P-GlcNAc to produce C55P-GlcN (Llewellyn, Zhao, et al. 2012), as I had originally hypothesized (Figure 7). Both LgmA and LgmC are required to modify the lipid A of B. pertussis with GlcN, and we predict these two enzymes function in the same pathway. Therefore, since the substrate for LgmC is C55P-GlcNAc, and LgmA transfers GlcNAc onto a lipid, the acceptor molecule for transfer by LgmA is probably the inner membrane carrier lipid C55P. ArnC, a homolog of LgmA in E. coli and Salmonella, also uses C55P as the acceptor when it transfers Ara4FN onto the carrier lipid in the Ara4N-lipid A modification pathway. This further supports the deduction that LgmA functions in the B. pertussis lipid A GlcN modification pathway by transferring GlcNAc onto C55P.  I had originally predicted that a flippase is required for the Lgm pathway, as ArnE and ArnF play this role in the Ara4N lipid A-modification pathway. Yet, my candidates, lgmD, lgmE, and BP1945, are not required for GlcN modification of lipid A, either individually or in combination with lgmD. It is possible that the presence of just one of these three genes may be sufficient to flip C55P-GlcN, since I did not concurrently knockout all three genes. If all three proteins, LgmD, LgmE, and BP1945, functioned in the same role as a C55P-GlcN flippase, I would predict knocking out two of these three proteins would decrease levels of lipid A GlcN modification in the bacterium. However, I did not observe any significant 99  decrease in hTLR4 activation by these double knock-out strains compared to wild type BP338, indicating these strains do not have a decrease in GlcN-modification of lipid A. Therefore, it is more likely that another protein functions as a flippase in this system, or no flippase is required, and spontaneous flipping of C55P-GlcN from the cytoplasmic to the periplasmic face of the IM is sufficient for the Lgm pathway.   Though translation of lgmE to produce the protein LgmE is not required for GlcN-modification of lipid A, generation of the lgmE transcript is not prevented by the mutations in the lgmEKO strains, in which a STOP codon was introduced at residue Gly15. I showed lgmE is transcribed in B. pertussis strain BP338, and since this transcript would be exactly complementary to, at the very least, the start of lgmC mRNA and to the end of lgmB mRNA, lgmE mRNA may be involved in the Lgm system, perhaps at a regulatory level.  The lack of GlcN-modification in B. hinzii and B. trematum lipid A remains unclear, since both strains have a complete copy of lgmA, lgmB, and lgmC, the required genes for the Lgm pathway. One possibility is that an as yet unidentified gene that is required for GlcN modification of lipid A is present in B. pertussis, and the other Bordetella species that possess this modification, but is absent from B. hinzii and B. trematum. Another possibility to consider is the presence of an inactivating mutation in lgmA, lgmB, or lgmC in these two species, though differences in LgmA at the residues equivalent to B. pertussis D159 and W163 are not responsible for this lack of modification. In B. pertussis, the lgm locus is part of the Bordetella virulence gene (Bvg) regulon (Marr, Tirsoaga, et al. 2008). Genome analysis indicates that the bvgAS two-component regulatory system genes are present in B. hinzii and B. trematum, suggesting that the lack of GlcN modification of lipid A is not due to the absence of the bvgAS genes. Culture conditions can also affect the expression of genes in the lgm locus (Marr, Tirsoaga, et al. 2008); however, all strains were grown under similar conditions when analyzed for lipid A structure (Novikov, Shah, et al. 2013). 100  Therefore, the lack of GlcN modification in B. hinzii and B. trematum may be due to the aforementioned differences in the lgm locus or to other variations in gene regulation and expression.   Mutational analysis of LgmA and LgmC identified key residues that may be involved in the function of these proteins. LgmA is a GT2 family glycosyltransferase with a GT-A fold, which is characterized by two adjacent β/α/β Rossmann domains (Lairson, Henrissat, et al. 2008). Glycosyltransferases with a GT-A fold often contain a conserved DXD motif that functions in the coordination of a divalent cation and/or ribose in the active site, and this has previously been considered a ‘signature’ of GT-A fold glycosyltransferases. However, a DXD motif is not found in all GT-A fold glycosyltransferases, and many non-glycosyltransferase enzymes possess conserved DXD motifs (Lairson, Henrissat, et al. 2008). In LgmA, I identified a conserved DXD motif that, when either one or both aspartates were mutated to glycine, resulted in no GlcN modification of B. pertussis lipid A. This suggests the aspartic acid residues of the conserved DXD motif in LgmA (D127 and D129) are required for activity. Whether these residues are involved in coordination of a cation remains to be seen.  LgmC is a member of the YdjC-like family of proteins. By aligning LgmC with other members of this protein family, including TTHB029 from Thermus thermophiles HB8, I identified several conserved residues: D80, D81, H130, D187, D189, and E313. Mutation of these residues resulted in a lack of GlcN-modified lipid A in B. pertussis, suggesting they are required for LgmC function. The structure of TTHB029 has been determined by crystallography (Imagawa, Iino, et al. 2008), and five hypothetical functionally important motifs have been identified. Motif 1 is DDXG and aligns with D80 and D81 of LgmC, which are part of a DDXG motif in LgmC. Motif 2 is GXH and the histidine of this motif in TTHB029 aligns closely with H130 in LgmC, which is also part of a GXH amino acid pattern. Motif 3 is a THXDXH motif in TTHB029, though only the DXH part of this motif is conserved in LgmC (D187 and H189). Motifs 1, 2, and 3 have been shown to bind Mg2+ in the TTHB029 crystal structure, suggesting 101  these residues may also be involved in coordinating divalent cations in LgmC. The last two motifs, 4 and 5, are presumed to be involved in the active site. Motif 4 contains H215 in TTHB029, which is equivalent to H295 in LgmC, and motif 5 contains R232 in TTHB029, which aligns with R310 in LgmC. I identified E313 of LgmC as a semi-conserved residue in the YdjC protein family, which is found close to the conserved residue R310, and mutational analysis suggests E313 may be involved in LgmC function. Motif 4 and 5 may be expanded to include additional conserved residues, following further mutational experiments.  In these mutational experiments of LgmA and LgmC, large amino acids, such as aspartic acid and histidine, were changed into glycine, which has a hydrogen atom as a side chain. This is a very drastic change in amino acid shape, and could result in deformation of the secondary structure of the protein in this region, which could result in abrogation of protein function, even if these specific residues are not directly involved in enzyme activity. Therefore, to confirm the requirement of the identified amino acids in protein function, further mutational analysis of these proteins is required. For example, mutating aspartic acid to leucine or asparagine, both of which have a similar shape to aspartic acid but lack the carboxylic acid group, may clarify the importance of this anionic carboxylic acid group.  lgmA and lgmB are regulated by the BvgAS virulence regulatory system in B. pertussis, and by an additional, unknown system, since lgmA and lgmB are optimally transcribed during growth in SS broth compared to growth on BG agar, which not observed for vag8, a known Bvg-regulated protein (Marr, Tirsoaga, et al. 2008). Despite the unknown component of lgmA and lgmB regulation, Bvg-regulation of these genes suggests involvement of the lgm locus in the virulence of B. pertussis. However, further analysis is required to elucidate the role of this GlcN modification, and other lipid A modifications, in pathogenesis and interaction with the host. 102  Chapter 4: The biological effects of lipid A modifications in B. pertussis 4.1 Introduction Changes to lipid A structure can affect many factors, such as membrane stability, resistance to CAMPs, and activation of TLR4 (Caroff, Karibian, et al. 2002, Needham and Trent 2013). In the previous chapter I described how the lgm locus, specifically lgmA, lgmB, and lgmC, are required for modification of B. pertussis lipid A with GlcN at the phosphate groups. In this chapter, I will explore the relationship between B. pertussis lipid A structure and these previously mentioned factors. First I will assess the effect of the lipid A GlcN modification on the susceptibility of B. pertussis to a variety of CAMPs and on the stability of the bacterial membrane. Then, I will determine the individual and combined contribution of two structural features of B. pertussis lipid A to activation of hTLR4: the GlcN modification of the phosphate groups and the length of the C3’ acyl chain.   4.2 Effect of GlcN modification on resistance to CAMPs and membrane stability B. pertussis colonizes the human respiratory tract, and therefore must contend with bacterial-clearance mechanisms present in the nasal cavity, trachea, and lung, including several CAMPs (de Gouw, Diavatopoulos, et al. 2011, Laube, Yim, et al. 2006). One mechanism for resistance to CAMPs in some bacteria is modification of lipid A, such as the addition of Ara4N by ArnT or phosphoethanolamine by EptA to the phosphate groups in Salmonella lipid A (Needham and Trent 2013). Modification of the lipid A phosphates by GlcN in B. pertussis may also affect resistance to CAMPs, along with stabilization of the OM. We had previously suggested this modification did not affect resistance to the CAMP polymyxin B (Marr, Hajjar, et al. 2010). However, in these earlier polymyxin B susceptibility experiments, bacteria were grown on BG agar with a range of polymyxin B concentrations. Since lgmA and lgmB (BP0399 and BP0398, respectively) have reduced levels of transcription during growth on BG agar compared to SS broth, the lipids A from these bacteria were likely not optimally modified with GlcN (Marr, Tirsoaga, et 103  al. 2008). Thus, I hypothesized the GlcN modification may increase resistance to CAMPs, if the bacteria are grown in SS broth, and not on BG agar.  4.2.1 GlcN modification increases resistance to CAMPs I tested the effect of B. pertussis lipid A GlcN modification on CAMP resistance by incubating wild-type BP338 bacteria and bacteria lacking the GlcN modification (BP338LgmABCDKO) with a range of CAMP concentrations. The bacteria were all initially grown in SS broth, to ensure optimal lipid A GlcN modification, before incubation with CAMPs or antibiotics for 2 h. I used this 2 h incubation method instead of growing the bacteria over 48 h in the presence of CAMPs to determine the minimum inhibitory concentration because recent studies have shown growth in the presence of sub-inhibitory concentrations of cationic peptides can upregulate LPS modifications involved in resistance to these very peptides (Fernandez, Jenssen, et al. 2012). I tested different structural classes of CAMPs, including: cyclic, lipidated CAMPs (polymyxin B and polymyxin E); alpha helical CAMPs (LL-37, HHC-10, and CP28); and an extended conformation CAMP (indolicidin) (Table 6). As shown in Figure 30, a greater percent of BP338 survived exposure to the different CAMPs when compared to its mutant, BP338LgmABCDKO, that lacks GlcN modification on lipid A. The greatest difference in survival between these two strains was observed when the bacteria were incubated with polymyxin B and polymyxin E. Complementation of the GlcN mutant with pPtacLgmABCD, which restored GlcN modification of LPS (Table 7), also restored wild-type levels of susceptibility to polymyxin B (Figure 31). Therefore, the lipid A GlcN modification in B. pertussis increases resistance to a variety of CAMPs. I observed no difference in susceptibility to another positively-charged antibiotic, the aminoglycoside gentamicin, between these two B. pertussis strains (Figure 30). This highlights the specificity of the GlcN modification resistance mechanism to CAMPs.   104   Figure 30. Influence of the B. pertussis lipid A GlcN modification on CAMP susceptibility B. pertussis lipid A GlcN modification increased resistance to CAMPs. B. pertussis BP338 and  BP338LgmABCDKO (GlcN mutant) strains were incubated with a range of killing agents including bacterial polymyxin B and polymyxin E (colistin), human LL-37, bovine indolicidin, insect CP28, and synthetic HHC-10. Gentamicin was used as a control. The concentrations of the killing agent are along the x-axis and percent survival (% survival) is along the y-axis. Graphs show the results of one representative experiment of three, n = 3 replicates per experiment. Statistical significance was determined by ANOVA, with a Bonferroni post-test to compare groups. **** P < 0.0001; ** P < 0.01; ns = no significant difference.    105      Figure 31. Complementation of BP338LgmABCDKO rescues resistance to polymyxin B B. pertussis BP338, BP338LmgABCDKO, and BP338LgmABCDKO + pPtacLgmABCD (complemented) strains were incubated with a range of polymyxin B (PmB) concentrations for 2 hours, then diluted in a 1/10 dilution series before 2 µl of each sample at each dilution was spotted onto a BG agar plate. These are results of one representative experiment of three.    106  4.2.2 GlcN modification increases resistance to OM perturbation The OM provides a protective barrier against many antimicrobial factors in the airway, such as lysozyme, which needs to permeate the OM to gain access to its substrate peptidoglycan (Laube, Yim, et al. 2006). CAMPs interact with the OM at sites where adjacent LPS molecules are bridged by divalent cations, causing perturbation of the OM and consequent self-promoted uptake of the CAMP (Hancock 1997). Any changes to the integrity of the OM, e.g. to engender increased resistance to CAMPs, have the potential to affect the ability of pathogens to survive in the respiratory tract. I tested the effect of lipid A GlcN modification on OM stabilization by incubating BP338 and the GlcN mutant, BP338LgmABCDKO, with EDTA, lysozyme, or both. EDTA perturbs the OM by chelating cations that bridge the phosphate groups of LPS molecules. I found that BP338 was more resistant to killing by EDTA alone compared to the GlcN mutant (Figure 32), showing that the GlcN modification stabilizes the OM, likely by decreasing the need for divalent cation stabilization. There was no significant difference in survival when the bacteria were incubated with lysozyme alone. However, membrane perturbation by 2 mg/ml EDTA rendered both bacterial strains more susceptible to killing by lysozyme, although the GlcN mutant had a larger decrease in survival when compared to BP338. This supports the conclusion that the GlcN modification on B. pertussis lipid A increases resistance to perturbation of the OM.   107   Figure 32. Influence of the B. pertussis lipid A GlcN modification on OM stabilization B. pertussis lipid A GlcN modification increased resistance to OM perturbation. B. pertussis BP338 and  BP338LgmABCDKO (GlcN mutant) strains were incubated with a range of killing agents including EDTA, lysozyme, and EDTA+lysozyme. The concentrations of the killing agent are along the x-axis and percent survival (% survival) is along the y-axis. Graph shows the results of one representative experiment of three, n = 3 replicates per experiment. Statistical significance was determined by ANOVA, with a Bonferroni post-test to compare groups. **** P < 0.0001; *** P < 0.001; ns = no significant difference.     108  4.3 lgm locus and lpxA are involved in the differences in lipid A structure between strains BP338 and 18-323 Previously, we have shown that the lack of GlcN-modified lipid A results in a decrease in activation of hTLR4 by B. pertussis BP338 LPS. However, the LPS of another B. pertussis wild-type strain, 18-323, exhibited even lower levels of hTLR4 activation (Figure 5) (Marr, Novikov, et al. 2010). Analysis of the structure of 18-323 lipid A revealed two differences compared to BP338 lipid A: 1) 18-323 lacks the GlcN modification, and 2) the C3’ acyl chain of 18-323 lipid A is only 10 or 12 carbons long, compared to BP338 lipid A, which has a C14-OH acyl chain at this position (Figure 4) (Marr, Novikov, et al. 2010). Consequently, I hypothesized each structural difference between BP338 and 18-323 lipid A individually affect activation of hTLR4 by LPS. To test this hypothesis, I planned to gradually modify the lipid A of 18-323 to resemble that of BP338. In order to do this, I first needed to determine the genetic basis for the lack of GlcN and the production of shorter C3’ acyl chain length in 18-323 lipid A.  4.3.1 Difference in GlcN modification is due an incomplete lgm locus In B. pertussis strain BP338, the lgm locus, specifically lgmA, lgmB, and lgmC, are required for the lipid A GlcN modification. Prior to the completion of the B. pertussis 18-323 genome (Park, Zhang, et al. 2012), the individual raw sequence reads for this strain had been released, and these were analyzed for the presence of lgmA, lgmB, lgmC, and lgmD by blastn algorithms. This analysis revealed that 18-323 possesses a complete lgmA, but lgmB has a mutation at bp 981 where bases TT are deleted, resulting in a frameshift, and an early stop codon (Figure 33AB). No matches to lgmC or lgmD were found in the 18-323 genome sequence. The absence of lgmC and lgmD in the 18-323 genome was confirmed via PCR using primers to amplify internal fragments of each of the lgm locus genes (Figure 33C). In addition, the existence of the dinucleotide deletion in lgmB was verified by sequencing using primers spanning the deletion site. The absence of a complete lgm locus in 18-323 thus explains the lack of GlcN-modified lipid A in this strain. 109   To corroborate these results, I complemented 18-323 with pPtacLgmABCD, thereby introducing the lgm locus of BP338 into 18-323. Negative-ion MALDI-MS analysis of the lipid A from this strain shows an extra set of peaks at m/z 1664 and 1692, i.e. 161 mass units (the mass of GlcN) higher relatively to the corresponding peaks of major unmodified penta-acylated lipid A molecular species observed in the wild type strain spectrum at m/z 1503 and 1531 (Figure 34AB). These additional peaks indicate the presence of C10-OH or C12-OH C3′ acyl chains and the addition of a GlcN modification in 18-323 lipid A.    110   Figure 33. Genetic analysis of the lgm locus of B. pertussis strains BP338 and 18-323 A) Lgm loci of BP338 and 18-323, as determined by sequence analysis. These data were provided by the pathogen genomics group at the Wellcome Trust Sanger Institute and can be obtained from the Sanger Institute website. The red ‘X’ represents a TT deletion mutation at bp 981 of lgmB in 18-323; the forward arrows and the reverse arrows illustrate the annealing sites of the primers used for the PCR shown in (C). B) Comparison of lgmB sequence between BP338 (top nucleotide sequence) and 18-323 (bottom nucleotide sequence) in the region of the TT deletion in 18-323 lgmB. LgmB amino acid code is above (BP338) or below (18-323) the nucleotide sequence, such that the amino acid letter aligns with the third nucleotide in the corresponding codon. The mutation in 18-323 lgmB, and the downstream sequence is highlighted to point out the frameshift mutation. C) PCR of the lgm locus genes in BP338 and 18-323 using gene-specific primers. Expected positive bands: 0.48 kb (lgmA primers, A), 0.51 kb (lgmB primers, B), 0.50 kb (lgmC primers, C), and 0.40 kb (lgmD primers, D). Modified from Shah et al. 2013 (Shah, Albitar-Nehme, et al. 2013), used with permission.    111   Figure 34. Lipid A structures of B. pertussis 18-323 strains complemented with BP338 lipid A-modifying genes (A) 18-323 wild type strain, (B) 18-323 + pPtacLgmABCD (complemented with the lgm locus of BP338), (C) 18-323 + pPtacLpxA338 (complemented with lpxA of BP338), (D) and 18-323 + pPtacLgmABCDLpxA338 (complemented with the lgm locus of BP338 and lpxA of BP338). Arrows labeled with C10OH, C12OH, or C14OH indicate the 10-, 12-, or 14-carbon acyl chains absent in the tetra-acyl and present in the respective penta-acyl lipid A species. Arrows labeled with GlcN indicate the addition of GlcN at a phosphate group. The lipid A structures are summarized to the right of the mass spectra. Numbers at the bottom of the structures indicate the length of the acyl chains. Structures with GlcN modifications (B and D) have one GlcN added to either of the phosphate groups (peaks at m/z 1664, 1692, or 1720). arb. u., arbitrary units. From Shah et al. 2013 (Shah, Albitar-Nehme, et al. 2013), used with permission.  112  4.3.2 Difference in C3’ acyl chain length is due to LpxA Lipid A of BP338 has C14-OH C3′ acyl chains, whereas that of 18-323 has C10-OH and C12-OH C3′ acyl chains. LpxA is an essential enzyme that catalyzes the reaction that substitutes an acyl chain onto the C3 of UDP-GlcNAc, and this C3 acyl chain can then become the C3′ acyl chain, when two GlcN backbone subunits are joined by LpxB, to form the di-GlcN moiety of lipid A (Raetz, Guan, et al. 2009). I compared BP338 and 18-323 LpxA sequences and found a single amino acid difference: amino acid 173 is a serine in BP338 LpxA and a leucine in 18-323 LpxA. Residue 173 in B. pertussis LpxA  is the equivalent of residue G176 in E. coli LpxA G176, which is positioned at the tip of the active site in close proximity to the C14 carbon of the 14-carbon long acyl chain that would be transfered onto GlcNAc by the activity of E. coli LpxA (Figure 35) (Williams and Raetz 2007). I hypothesized that the larger L173 in 18-323 LpxA occludes the tip of the active site, therefore allowing only C10-OH and C12-OH acyl chains into the active site, whereas the smaller S173 in BP338 LpxA allows C14-OH acyl chains.   To demonstrate that the difference at amino acid 173 of B. pertussis LpxA is the reason for the difference in C3′ acyl chain lengths between BP338 and 18-323, I introduced BP338 lpxA into 18-323 via the vector pPtacLpxA338. Analysis of the lipid A structural modifications of this strain via negative-ion MALDI-MS shows that introduction of BP338 lpxA into 18-323 generates a new major peak at m/z 1559. This peak corresponds to the addition of C3′ acyl chains with 14 carbons, as opposed to wild type 18-323, which only has peaks at m/z 1503 and 1531 (C10-OH and C12-OH C3′ acyl chains, respectively) (Figure 34AC). This shows that the LpxA from BP338 (with the single amino acid difference) alone is sufficient to introduce C14-OH C3′ acyl chains onto the lipid A of 18-323.    113   Figure 35. Alignment of LpxA from various Gram-negative species Species (top to bottom) are: E. coli K-12 (EC), P. aeruginosa (PA), B. bronchiseptica (Bbronch), Bordetella parapertussis (Bpara), B. pertussis Tohama I BP338 (Tohama), B. pertussis 18-323 (18323), R. sphaeroides (RS), and P. gingivalis (Pging). The alignment was constructed with ClustalW2 (Thompson, Higgins, et al. 1994). The down arrow indicates the single amino acid difference between LpxA of BP338 and 18-323 (amino acid 173), which corresponds to G176 in E. coli LpxA. The bottom right inset shows the active site region between chains A (gray) and B (brown) of the E. coli LpxA homo-trimer and the amino acids surrounding the 14-carbon acyl chain (pink). E. coli LpxA amino acids: H160 (green), G173 (blue), G176 (red), and H191 (orange). E. coli LpxA structure PDB ID 2QIA (Williams and Raetz 2007) was used to generate this figure in Swiss-Pdb Viewer (ExPASy) (Guex and Peitsch 1997). From Shah et al. 2013 (Shah, Albitar-Nehme, et al. 2013), used with permission.  114  4.4 C3’ acyl chain length and GlcN modification individually affect hTLR4 activation I set out to explore the effect of these minute structural differences in B. pertussis lipid A (the presence of C14-OH acyl chains at the C3′ position and the GlcN modification) on hTLR4 activation. As such, I first generated a recombinant 18-323 strain that was complemented with both lpxA from BP338 and the full lgm locus (18-323 + pPtacLgmABCDLpxA338). The mass spectral analysis shows that lipid A from this strain displays a mixture of lipid A species, including species that have longer C14-OH C3′ acyl chains (peaks m/z 1559 and 1720) and species that contain the GlcN modification (peaks m/z 1664, 1692, and 1720) (Figure 34D). We tested the ability of purified LPS from these 18-323-derived strains to activate hTLR4 with the HEK-Blue hTLR4 activation assay (Figure 36). Our results show that 18-323 with BP338lpxA (18-323 + pPtacLpxA, which now has longer acyl chains at the C3′ position) was more effective in activating hTLR4 than wild type 18-323. When 18-323 lipid A was modified with GlcN at the phosphate groups (18-323 + pPtacLgmABCD), there was also a significant increase in hTLR4 activation over the wild type levels. By incorporating both these modifications into 18-323 to produce a longer C3′ acyl chain and the GlcN modification (18-323 + pPtacLpxA338LgmABCD), a significant increase in hTLR4 activationwas observed when compared to activation by LPS from the wild type (18-323) and from 18-323 with the longer acyl chain (18-323 + pPtacLpxA338). However, there was no difference in hTLR4 activation level when comparing 18-323 + pPtacLpxA338LgmABCD to 18-323 LPS that only has the GlcN modification (18-323 + pPtacLgmABCD). These results suggested that each modification of B. pertussis LPS alone was sufficient to cause an increase in hTLR4 activation. These results also suggest that when the GlcN modification is present, increasing the C3′ acyl chain length does not further increase hTLR4 activation. However, because there are varying levels of these modifications in each of the 18-323 strains, the relative contribution of each modification is difficult to determine.    115    Figure 36. hTLR4 activation by LPS from B. pertussis 18-323 strains with BP338 lipid A-modification genes hTLR4 activation measured with the HEK-Blue NFκB hTLR4 activation assay. Null2 all, stimulation of HEK-Blue Null2 cell line that lacks TLR4 expression with all LPS variants; Blank, medium only with no HEK-Blue cells; Unstim, HEK-Blue hTLR4 cells stimulated with medium only; 18-323 + pPtacLpxA338, 18-323 complemented with lpxA from BP338; 18-323 + pPtacLgmABCD, 18-323 complemented with lgm locus from BP338; 18-323 + pPtacLgmABCDLpxA338, 18-323 complemented with both lpxA and lgm locus from BP338. Absorbance at 650 nm (A650). Graph shows the results of one representative experiment of three, n = 6 replicates per experiment. p values: <0.01 (**), < 0.001 (***). From Shah et al. 2013 (Shah, Albitar-Nehme, et al. 2013), used with permission.     116  4.5 Discussion I obtained data to show that the B. pertussis lipid A GlcN modification increased resistance of the bacteria to CAMPs, but this modification did not appear to affect resistance to the positive aminoglycoside, gentamicin. Furthermore, the GlcN modification increased resistance to OM perturbation. I also demonstrated that the genetic basis for the 18-323 lipid A structure was due to the absence of a complete lgm locus, which explains the lack of the GlcN modification, and a single amino acid difference in LpxA, which explains the shorter C3’ acyl chain when compared to BP338. This allowed me to test the effect of each individual structural difference between BP338 and 18-323 lipid A on LPS activation of hTLR4 by complementing 18-323 with BP338 lipid A-modifying genes. I found each modification (increasing the C3’ acyl chain length or adding GlcN to the phosphates) alone increased the level of hTLR4 activation, but the addition of both modifications resulted in the same level of hTLR4 activation as when only the GlcN is added to lipid A.  Similar modifications of lipid A in different species do not always have the same effects. For example, Ara4N modification of Salmonella LPS does not affect activation of TLR4, whereas the same modification in Pseudomonas aeruginosa increases activation of TLR4 (Gellatly, Needham, et al. 2012). In this context, the recent findings that the lgmB gene in B. bronchiseptica, the causative agent of kennel cough in small mammals, is important for resistance to polymyxin B and porcine β-defensin 1 (pBD1) (Rolin, Muse, et al. 2014) does not necessarily mean the same is true for B. pertussis, especially considering the differences in LPS between these two species. B. pertussis LPS has 5 acyl chains and no O-antigen, whereas B. bronchiseptica LPS is hexa-acylated and has a long O-antigen, which has been shown to confer resistance to CAMPs (Banemann, Deppisch, et al. 1998, Marr, Tirsoaga, et al. 2008, Rolin, Muse, et al. 2014). Despite these considerable variations in LPS structure, GlcN modification still confers resistance to polymyxin B in both bacterial species. Furthermore, when comparing the effect on TLR4 activation in the natural hosts, the LgmB-mediated modification in B. bronchiseptica does not 117  affect activation of mouse TLR4, whereas the GlcN modification in B. pertussis increases activation of hTLR4 (Marr, Hajjar, et al. 2010, Rolin, Muse, et al. 2014). This adds a layer of complexity in the B. pertussis and human system that is not present in B. bronchiseptica and mice. The GlcN-modification in B. bronchispetica has also been linked to successful transmission between mice and colonization at lower infectious doses (Rolin, Muse, et al. 2014).   Previously, Ara4N modification to lipid A was proposed to protect Salmonella against CAMPs by decreasing the overall negative charge of the OM, thereby decreasing the affinity for positively charged CAMP molecules (Gunn, Lim, et al. 1998). I propose that in addition to this charge-masking mechanism of CAMP resistance, the modification of lipid A with a positively-charged sugar (GlcN in B. pertussis) may also function to stabilize the OM, thereby further excluding CAMPs from gaining access to the bacterium (Needham and Trent 2013). Bacteria with GlcN-modified lipid A were more resistant to lysozyme in the presence of the cation-chelator EDTA (Figure 32), which perturbs the OM by chelating the stabilizing cations that bridge the negatively charged phosphate groups of lipid A (Needham and Trent 2013). I propose that in wild-type B. pertussis, which modifies the phosphates of a proportion of its lipid A with GlcN, the positively-charged GlcN groups are able to coordinate negatively-charged phosphate groups on other unmodified lipid A molecules, thereby stabilizing the OM, even in the absence of cations, as described in Figure 37.   Stabilization of the OM is also critical for protection against numerous agents present in the airway that contribute to the clearance of infections, such as lysozyme, which is found at 1 mg/ml in sputum (Laube, Yim, et al. 2006, Needham and Trent 2013). To kill Gram-negative cells, lysozyme must traverse the OM to gain access to the peptidoglycan, which it degrades to promote cell death. Agents in the respiratory tract, such as lactoferrin, destabilize the OM, thus allowing lysozyme greater access to the peptidoglycan (Laube, Yim, et al. 2006). This type of synergy has been found between many antimicrobial factors in the 118  airway (Laube, Yim, et al. 2006). Therefore, the increase in resistance to OM perturbation afforded by the lipid A GlcN modification may also protect B. pertussis against lysozyme and other infection-clearing mechanisms.  Differences between enzymes in the Raetz lipid A biosynthesis pathway (Raetz, Guan, et al. 2009) can also be responsible for the variations observed in lipid A structures between different strains and species. Williams and Raetz suggest that the length of the acyl chain at the C3 and C3′ positions of lipid A is controlled by the hydrocarbon ruler region of LpxA, that is, the amino acids located near the active site in proximity to the acyl chain of the substrate, such as G173 and G176 in E. coli LpxA (Williams and Raetz 2007). Previous work, where mutating G173 of E. coli LpxA changed the acyl chain length specificity of the enzyme, supports this theory (Wyckoff, Lin, et al. 1998). My data also support this theory because amino acid 173 in B. pertussis LpxA is equivalent to G176 in E. coli, and if B. pertussis LpxA has a serine in this position (as seen in BP338), a C14-OH acyl chain can be added at the C3′ position, whereas if a leucine is found in position 173 (as seen in 18-323), only C10-OH and C12-OH acyl chains are found. Therefore, I propose that the larger leucine is occluding the tip of the active site of LpxA, thus only allowing 10- or 12-carbon acyl chains to be added. An alignment of LpxA sequences from several Gram-negative species suggests a similar correlation between strains with shorter acyl chains at the C3 and/or C3′ positions (e.g. Pseudomonas aeruginosa and Rhodobacter sphaeroides) and the presence of a larger amino acid at positions equivalent to E. coli LpxA G173, G176, or both (Figure 35). However, for P. gingivalis LpxA, this correlation, which is mostly based on sequence alignment, does not hold true and may be a reflection of the decreased sequence similarity of the P. gingivalis LpxA. Thus, although this analysis generally supports the role of both G173 and G176 of E. coli LpxA as a hydrocarbon ruler for the length of the acyl chain added to lipid A at the C3 and C3′ positions, there are likely more complexities involved in determining acyl chain length, especially in more distantly related LpxA species.  119  These results help to shed light on how the GlcN modification and C3’ acyl chain length of penta-acyl LPS can affect various aspects of the bacterium, from CAMP resistance to activation of hTLR4. Yet, many bacteria have hexa-acyl LPS, such as E. coli and Salmonella. Studies looking at the removal and addition of entire acyl chains in E. coli hexa-acyl LPS demonstrate how such drastic changes to the acylation pattern can sometimes affect CAMP resistance and activation of TLR4. However, the effects of slight changes to the acyl chain length of hexa-acyl E. coli LPS have, thus far, not been studied.      120    Figure 37. Model for how the B. pertussis GlcN modification facilitates CAMP resistance and OM stabilization Top row depicts BP338 lipid A with the GlcN modification (blue) on some phosphate groups (red). Bottom row depicts BP338 lacking the GlcN modification (BP338GlcN-). CAMP (green helix), cation (orange), coordination between oppositely charged elements (purple-dashed line). ‘+’ and ‘-’ indicate the charge of the correspondingly coloured factor. 1. In BP338, positively charged GlcN modification repels positively charged CAMPs; in BP338GlcN-, the negatively charged, exposed phosphates groups attract the positively charged CAMPs. 2. In BP338, the positively charged GlcN that modifies some lipid A phosphate groups coordinates other negatively charged un-modified phosphates, thereby stabilizing the OM; in BP338GlcN-, the negatively charged phosphate groups rely on coordinating positively charged divalent cations to stabilize the OM.  121  Chapter 5: The biological effects of varying the C3 and C3’ acyl chain lengths in E. coli hexa-acyl lipid A 5.1 Introduction Major changes to the acylation pattern of lipid A can have a wide range of effects on bacteria (Raetz, Reynolds, et al. 2007). For example, addition of an acyl chain at the C2 secondary position by PagP to generate hepta-acyl lipid A in Salmonella promotes resistance to CAMPs, and also attenuates TLR4 activation (Miller, Ernst, et al. 2005, Raetz, Reynolds, et al. 2007, Trent, Stead, et al. 2006). Thus, a single acyl chain modification can have multiple biological effects.  In regards to TLR4 activation, not only the number of acyl chains, but also their position on the lipid A di-GlcN backbone affects the degree of activation (Needham, Carroll, et al. 2013). Coats et al. (Coats, Berezow, et al. 2011) suggest that in penta-acyl lipid A species, longer fatty acid chains correlate with greater TLR4 activity. This hypothesis is based on the observation that penta-acyl E. coli LPS with acyl chain lengths of 12 and 14 carbons is a weaker TLR4 agonist in comparison with Porphyromonas gingivalis penta-acyl LPS, which contains C15, C16, and C17 acyl chains. Both these species have penta-acyl lipid A, and I have shown in B. pertussis penta-acyl lipid A, increasing the C3’ acyl chain in isogenic strains increases hTLR4 activation (Section 4.4). This situation may be different in hexa-acyl LPS species, since hexa-acyl LPS tends to be a robust activator of TLR4, suggesting very strong interactions with TLR4 and MD-2 (Maeshima and Fernandez 2013). As such, changing the acyl chain length may or may not affect hTLR4 activation by hexa-acyl LPS.   Varying the length of the acyl chains in LPS may also affect other aspects, such as growth, resistance to CAMPs, and activation of the LAL assay, which is used in the biotechnology industry to detect endotoxin contamination.  122  5.2 Generation of R0138 E. coli strains with different C3 and C3’ acyl chain lengths To study the effects of lipid A acyl chain length variation in a hexa-acyl strain of E. coli, I generated isogenic strains with varying C3 and C3’ acyl chains. The E. coli strain R0138 (Galloway and Raetz 1990) has lpxA under the control of a temperature sensitive promoter, such that endogenous LpxA is expressed at low temperatures (i.e. 30°C), but is not expressed at higher temperatures (i.e. 42°C). LpxA is the first enzyme in the Raetz LPS biosynthesis pathway that adds the acyl chain onto C3 of UDP-GlcNAc. Once the di-glucosamine backbone of lipid A is formed, this added acyl chain is substituted at either the C3 or C3’ position in the final LPS molecule (Raetz, Reynolds, et al. 2007). LpxA from B. pertussis strain BP338 is able to add at least C10-OH and C14-OH acyl chains and B. pertussis strain 18-323 LpxA can add C10-OH and C12-OH acyl chains (Section 4.3.2). Therefore, to generate isogenic R0138 E. coli strains that differ only in the lengths of the C3 and C3’ primary acyl chains, I transformed R0138 with plasmids containing lpxA from E. coli DH5α (R0138 + pBSlpxAEC), B. pertussis strain BP338 (R0138 + pBSlpxA338), or B. pertussis strain 18-323 (R0138 + pBSlpxA18323). In addition, I also complemented R0138 with lpxA from B. pertussis strains BP338 and 18-323 in the reverse orientation in the pBS vector (R0138 + pBSlpxA338rev and R0138 + pBSlpxA18323rev, respectively) as negative controls.   Since lpxA is an essential gene, R0138 is unable to grow at 42°C, as the endogenous LpxA is not expressed. However, if R0138 is complemented with exogenous LpxA it should grow at 42°C. I found all the R0138 strains can grow at 30°C, which is expected since the endogenous LpxA is expressed under these conditions (Figure 38). However, only the strains that are complemented with exogenous lpxA in the forward direction in the vector (R0138 + pBSlpxAEC, R0138 + pBSlpxA338, and R0138 + pBSlpxA18323) were able to grow at 42°C, suggesting successful complementation with functional LpxA in these strains (Figure 38).  123   Figure 38. Viability of E. coli R0138 strains complemented with lpxA at 30°C and 42°C LB broth with the appropriate antibiotics was inoculated with freshly grown R0138 colonies and incubated for 24 hours shaking at 200 rpm at 30°C or 42°C. Images were taken of 2 ml of culture with Olympus Camedia C-5060 camera. Images show the results of one representative experiment of three.    124  If the exogenous LpxA is fully functional in the Raetz lipid A biosynthesis pathway in these R0138 E. coli strains, I expect these strains to contain lipid A with different C3 and C3’ acyl chain lengths. Analysis of the lipid A from these strains grown at 42°C (expression of the exogenous LpxA only) with mass spectrometry showed R0138 complemented with E. coli lpxA (R0138 + pBSlpxAEC) (Figure 39B) had primarily hexa-acyl lipid A species with C14-OH C3 and C3’acyl chains (1797 m/z peak), as seen in wild-type R0138 grown at 30°C (Figure 39A). However, both strains also had a minor peak at 1769 or 1770 m/z, indicating the presence of lipid A species with acyl chains shortened by two carbons, perhaps at the C3 or C3’ position. Complementation of R0138 with B. pertussis strain BP338 lpxA (R0138 + pBSlpxA338) resulted in hexa-acyl lipid A with C3 and C3’ acyl chains of C10-OH, C12-OH, and C14-OH (Figure 39C), as indicated by the following peaks: 1685 m/z represents lipid A with 10-carbon long C3 and C3’ acyl chains; 1713 m/z represents lipid A with a 10-carbon C3 acyl chain and a 12-carbon C3’ acyl chain or a C12-OH C3 acyl chain and a C10-OH C3’ acyl chain; 1741 m/z represents lipid A with a C10-OH C3 acyl chain and a C14-OH C3’ acyl chain, a C12-OH acyl chain at both the C3 and C3’ positions, or a C14-OH C3 acyl chain and C10-OH C3’ acyl chain; 1770 m/z represents lipid A with a C12-OH C3 acyl chain and a C14-OH C3’ acyl chain or a C14-OH C3 acyl chain and a C12-OH C3’ acyl chain; and finally 1779 m/z represents lipid A with a 14-carbon acyl chain at both the C3  and C3’ positions. Therefore, compared to R0138 + pBSlpxAEC, the R0138 + pBSlpxA338 strain has lipid A with overall shorter C3 and C3’ acyl chains. However, complementing R0138 with B. pertussis strain 18-323 lpxA (R0138 + pBSlpxA18323) results in a strain with the shortest acyl chains, with only C10-OH and C12-OH acyl chains at the C3 and C3’ positions (Figure 39D). When grown at 42°C, all three complemented strains have slightly higher peaks representing lipid A with 7 fatty acyl chains, which could suggest a greater level of hepta-acyl lipid A in comparison to R0138 grown at 30°C. This could be due to the addition of an acyl chain by PagP to the hexa-acyl lipid A species in response to the temperature stress. Therefore, I confirmed the generation of isogenic E. coli strains with hexa-acyl lipid A 125  with gradually decreasing C3 and C3’ acyl chain lengths (R0138 + pBSlpxAEC to R0138 + pBSlpxA338 to R0138 + pBSlpxA18323) when grown at 42°C.      126     127   Figure 39. Lipid A structures of E. coli R0138 strains complemented with exogenous lpxA A) R0138 (30°C), B) R0138 + pBSlpxAEC (42°C), C) R0138 + pBSlpxA338 (42°C), D) and R0138 + pBSlpxA18323 (42°C). Peaks at m/z 1685 represent hexa-acyl lipid A with acyl chains lengths (C3, C3’) of (10, 10) carbons, m/z 1713 represent hexa-acyl lipid A with acyl chains lengths (C3, C3’) of (10,12) and (12, 10) carbons, m/z 1741 represent hexa-acyl lipid A with acyl chains lengths (C3, C3’) of (10,14), (12, 12) and (14, 10) carbons, m/z 1770 represent hexa-acyl lipid A with acyl chains lengths (C3, C3’) of (12,14) and (14, 12) carbons, and m/z 1797 represent hexa-acyl lipid A with acyl chains lengths (C3, C3’) of (14,14) carbons. 4FA, 4 fatty acid chains; 5FA, 5 fatty acid chains; 6FA, 6 fatty acid chains; 7FA, 7 fatty acid chains. The hexa-acyl lipid A structures are summarized to the right of the mass spectra. The numbers at the bottom of the structures indicate the length of the acyl chains     128  5.3 C3 and C3’ acyl chain lengths affect bacterial growth Though all three lpxA-complemented R0138 strains were able to grow at the non-permissive temperature, the growth rate could be affected by the difference in C3 and C3’ primary acyl chain lengths of the lipid A. To test this assumption, I grew wild type R0138 and the three lpxA-complemented strains at 30°C and 42°C. I found all four strains grew at similar rates at 30°C, conditions under which the endogenous LpxA is expressed (Figure 40A). At 42°C, when only the complemented LpxA is functioning in the bacterium, wild-type R0138 is unable to grow (Figure 40B). R0138 + pBSlpxAEC and R0138 + pBSlpxA338, which have the longer C3 and C3’ acyl chains, grow at similar rates at 42°C whereas the strain with the shortest acyl chains, R0138 + pBSlpxA18323, grows at a slower rate (Figure 40B). The association between shorter C3 and C3’ acyl chains and a reduced growth rate could be the result of the enzymes in the lipid A biosynthesis pathway being less efficient with substrates with shorter acyl chains, or a decrease in fitness of the bacteria with shorter acyl chains due to factors such as membrane stability, or a combination of both these issues.    129      Figure 40. Growth of E. coli R0138 strains with varying acyl chain lengths Growth curves of R0138 strains at 30°C (A) and 42°C (B). Graphs show the results of one representative experiment of three, n = 2 repeats per experiment.    130  5.4 C3 and C3’ acyl chain lengths affect resistance to polymyxin B Polymyxin B is a cationic peptide that is often used as a last resort for the treatment of multidrug-resistant Gram-negative pathogens (Zavascki, Goldani, et al. 2007). Polymyxin B, and other cationic peptides, are taken up by the bacterium via interaction with LPS and cause cell death through destabilization of the OM (Zavascki, Goldani, et al. 2007). In hexa-acyl E. coli lipid A, changes in the C3 and C3’ acyl chain lengths may affect susceptibility to polymyxin B, since these changes could influence membrane stability or the interaction between polymyxin B and LPS. I used two methods to compare polymyxin B resistance between the lpxA-complemented strains with different acyl chain lengths. First, I grew the bacteria in the presence of different polymyxin B concentrations and monitored growth over time. Using this method, I found a marked difference in growth at 0.5 µg/ml polymyxin B (half the minimum inhibitory concentration of polymyxin B for E. coli): as C3 and C3’ acyl chains decrease in length (R0138 + pBSlpxAEC to R0138 + pBSlpxA338 to R0138 + pBSlpxA18323), I observe a longer delay in growth when compared to growth in no polymyxin B (Figure 41). To confirm that these results were not due to other LPS modifications that might be induced by polymyxin B (Fernandez, Jenssen, et al. 2012), as I discussed in the previous chapter, I used a second method in which the bacteria were incubated with polymyxin B for 5 min followed by cfu determination to monitor killing of the bacteria. In this case, I also see an effect of C3 and C3’ acyl chain length on polymyxin B resistance: at 0.5 µg/ml polymyxin B the strain with the shortest acyl chains (R0138 + pBSlpxA18323) has decreased survival compared to strains with longer acyl chains (R0138 + pBSlpxAEC and R0138 + pBSlpxA338) (Figure 42).    131     Figure 41. Polymyxin B resistance of E. coli R0138 strains with varying acyl chain lengths, assessed by growth curve assay R0138 strains were grown in polymyxin B (PmB) (0, 0.5, and 1.0 µg/ml) at 42°C and growth was followed over time by OD600 readings of cultures. A) R0138 + pBSlpxAEC (R0138 + lpxAEC), B) R0138 + pBSlpxA338 (R0138 + lpxA338), C) R0138 + pBSlpxA18323 (R0138 + lpxA18323)   Graphs show the results of one representative experiment of three, n = 2 replicates per experiment.  132    Figure 42. Polymyxin B resistance of E. coli R0138 strains with varying acyl chain lengths, assessed by percent survival assay R0138 strains were incubated in polymyxin B (PmB) for 5 minutes and then plated on LB agar for cfu counting to monitor bacterial killing and determine percent survival (% survival). R0138 + pBSlpxAEC (R0138 + lpxAEC), R0138 + pBSlpxA338 (R0138 + lpxA338), R0138 + pBSlpxA18323 (R0138 + lpxA18323). Graphs show the results of one representative experiment of three, n = 3 replicates per experiment. p-values: <0.05 (*), <0.01 (**).     133  5.5 C3 and C3’ acyl chain lengths affect hTLR4 activation The C3’ acyl chain length affected hTLR4 activation by penta-acyl B. pertussis LPS (Section 4.4). However hexa-acyl E. coli LPS has a higher level of hTLR4 activation than B. pertussis strain BP338 LPS (Marr, Novikov, et al. 2010), therefore minor changes in the length of acyl chains may not affect hTLR4 activation as they do in B. pertussis LPS. To determine if variation in C3 and C3’ acyl chain lengths of E. coli lipid A affect hTLR4 activation, I stimulated HEK-Blue hTLR4 cells with purified LPS from the lpxA-complemented R0138 strains. I found a significant difference in hTLR4 activation between R0138 + pBSlpxAEC and R0138 + pBSlpxA18323 LPS (Figure 43), showing that as the C3 and C3’ acyl chain lengths decrease, hTLR4 activation by LPS also decreases. This trend also holds true for R0138 + pBSlpxA338, which has intermediate C3 and C3’ acyl chain lengths. LPS from this E. coli strain consistently has intermediate levels of hTLR4 activation compared to R0138 + pBSlpxAEC and R0138 + pBSlpxA18323, across all repeated experiments, though this difference is not always statistically significant.    134     Figure 43. Activation of hTLR4 by LPS from E. coli R0138 strains with varying acyl chain lengths hTLR4 activation measured with the HEK-Blue NFκB hTLR4 activation assay. Null2 all, stimulation of HEK-Blue Null2 cell line that lacks TLR4 expression with all LPS variants; Blank, medium only with no HEK-Blue cells; Unstim, HEK-Blue hTLR4 cells stimulated with medium only; R0138 + lpxAEC, R0138 complemented with lpxA from E. coli; R0138 + lpxA338, R0138 complemented with lpxA from B. pertussis strain BP338; R0138 + lpxA18323, R0138 complemented with lpxA from B. pertussis strain 18-323. Absorbance at 650 nm (A650). Graph shows the results of one representative experiment of three, n = 5 replicates per experiment. p value: <0.01 (**).     135  5.6 C3 and C3’ acyl chain lengths affect LAL activation The LAL assay detects LPS through interaction between LPS and the Factor C protein of the Horseshoe crab, which leads to a downstream enzymatic cascade that is linked to an experimental read out, such as a change in colour (Muta, Miyata, et al. 1991). We used the Pyrochrome LAL assay to assess the activation of the LAL cascade by LPS purified from the three lpxA-complemented E. coli R0138 strains. We found R0138 + pBSlpxAEC, the strain with the longest lipid A C3 and C3’ acyl chains, had the lowest levels of LAL activation when compared to R0138 + pBSlpxA338 and R0138 + pBSlpxA18323, which have gradually shorter C3 and C3’ acyl chains in comparison (Figure 44). Therefore, E. coli hexa-acyl LPS with longer C3 and C3’ acyl chains (14 carbons long) activates the LAL assay to a lesser degree compared with LPS with shorter C3 and C3’ acyl chains.    136   Figure 44. Activation of the LAL assay by LPS from E. coli R0138 strains with varying acyl chain lengths Pyrochrome LAL assay used to assay purified LPS at 10 ug/ml. – control, endotoxin-free water. Absorbance at 405 nm (A405). Graph shows the average of 4 individual experiments, n = 3 replicates per experiment. p-values: < 0.001 (***).     137  5.7 Discussion Variation in the acyl chain lengths of LPS can affect many aspects of bacterial physiology, from membrane stability to interaction with the host immune system.   Complementing E. coli strain R0138 with lpxA from B. pertussis strains BP338 and 18-323 could provide further information regarding the lipid A biosynthesis pathway in B. pertussis, and also raises some queries regarding selectivity in this pathway. LPS from B. pertussis strain BP338 has C14-OH acyl chains at the C3’ position and C10-OH acyl chains at the C3 position, suggesting LpxA from this strain is able to transfer both C10-OH and C14-OH acyl chains (Shah, Albitar-Nehme, et al. 2013). However, when R0138 is complemented with lpxA from BP338 and grown at the non-permissive temperature, C10-OH, C12-OH, and C14-OH acyl chains are added at the C3 and C3’ positions. This indicates BP338 LpxA can also transfer C12-OH acyl chains, even though C12-OH chains are not observed in the final LPS structure in B. pertussis BP338 (Caroff, Brisson, et al. 2000, Marr, Tirsoaga, et al. 2008). Therefore, there is likely another level of control over the lengths of the C3 and C3’ acyl chains in B. pertussis strain BP338, other than LpxA. Though LpxA controls the acyl chain length attached at the C3 position of UDP-GlcNAc during the initial step in LPS biosynthesis, another enzyme in this pathway likely plays a role to ensure the C3’ position acyl chain is 10 carbons long and the C3’ position acyl chain is 14 carbons long in the final lipid A structure. For example, LpxH could specifically cleave uridine-monophosphate from substrates with a C10-OH acyl chain at the C3 position to generate lipid X molecules with only a 10-carbon long chain at this position, therefore resulting in LPS molecules with only C10-OH acyl chains at the C3’ position. Alternatively, LpxB in B. pertussis BP338 could specifically catalyze the joining of a lipid X molecule with a C10-OH acyl chain at the C3 position and a UDP-2,3-diacyl-GlcN molecule with C14-OH acyl chain at the C3 position. Since lipid A of B. pertussis strain 18-323 also consistently has a C10 acyl chain at the C3 position, the selective mechanism found in BP338 is probably also present in 18-323. However, the C3’ position acyl chain in 18-323 LPS can be 10 or 12 carbons long, even when 138  complemented with BP338 lpxA which allows this strain to have 10-, 12-, or 14-carbon long acyl chains at this position. Therefore, the selectivity for the C3’ acyl chain length in BP338, which ensures only 14-carbon acyl chains at this position, is not present in 18-323.  The presence of alternate C3 and C3’ acyl chain lengths in the complemented E. coli R0138 strains demonstrated that the LPS biosynthesis pathway of E. coli has a certain level of substrate flexibility in the downstream enzymes that allows them to accommodate lipid A with different acyl chain lengths. However, we did observe a slight growth defect in the strain with the shortest acyl chains (R0138 + pBSlpxA18323), which could result from a decreased efficiency of the Raetz pathway with substrates with shorter acyl chains. This could also be a result of decreased membrane stability due to decreased acyl chain lengths, which is predicted to decrease the overall hydrophobic interactions of the LPS acyl chains in the OM (Guo, Lim, et al. 1998).  A decrease in membrane stability is sometimes associated with an increase in susceptibility to CAMPs (Guo, Lim, et al. 1998). As discussed in the previous chapter, CAMPs are an important part of the innate immune system of many organisms, and are secreted by epithelial and immune cells in humans (Miller, Ernst, et al. 2005). CAMPs are also found in the environment, such as polymyxin B, which is a CAMP produced by Gram-positive environmental bacteria. Polymyxin B interacts with surface-exposed regions of LPS at the bacterial OM and displaces the cations Ca2+ and Mg2+, therefore destabilizing the OM and allowing polymyxin B to interact with the fatty acyl chains. This results in the insertion of polymyxin B into the membrane, which leads to OM leakiness (Zavascki, Goldani, et al. 2007). Salmonella that lacks the ability to generate hepta-acyl LPS (pagP mutant) has decreased resistance to CAMPS, suggesting a relationship between the level of acylation of LPS and CAMP resistance (Guo, Lim, et al. 1998). My findings also support this hypothesis, since decreasing the C3 and C3’ acyl chain lengths of hexa-acyl E. coli LPS also decreases resistance to polymyxin B. 139   The acylation pattern of LPS can also be associated with different levels of TLR4 activation, though this is not always the case. In section 4.4, I have shown in B. pertussis penta-acyl LPS, increasing the length of the C3’ acyl chain also increases hTLR4 activation. However, modifying the acyl chains of LPS does not always result in a difference in TLR4 activation. Removal of the primary C3 acyl chain in hexa-acyl E. coli LPS did not affect TLR4 activation whereas removal of the secondary C3’ acyl chain decreased TLR4 activation by LPS (Needham, Carroll, et al. 2013). This suggests the location of the acyl chains is an important factor in the ability of LPS to activate TLR4. Since I modified the lengths of both the C3 and C3’ acyl chains in E. coli, I am unable to determine if the decrease in hTLR4 activation as acyl chain lengths decrease is the effect of altering the C3 acyl chain, the C3’ acyl chain, or both. In either case, the longer acyl chains in both penta-acyl and hexa-acyl LPS, in B. pertussis and E. coli, respectively, are associated with higher levels of hTLR4 activation.   Excessive release of proinflammatory cytokines in response to TLR4 activation can lead to in endotoxic shock, where these mediators damage small blood vessels, which can lead to septic shock and multiple organ failure (Raetz and Whitfield 2002). Therefore, when proteins are purified from bacteria for use in humans, it is important to ensure they are endotoxin-free. The LAL assay is used in the biotechnology industry to determine the endotoxicity of a sample, and this assay is based on the interaction between the Factor C enzyme of the Horseshoe crab and the di-GlcN backbone of lipid A (Chen and Mozier 2013, Muta, Miyata, et al. 1991). However, differences in lipid A structure can have different effects on LAL and TLR4 activation, since these two different methods of detecting lipid A depend on interaction with different proteins. For example, when comparing tetra-acyl and hexa-acyl lipid A: tetra acyl lipid A has higher levels of LAL activation whereas hexa-acyl lipid A has higher levels of TLR4 activation (Gutsmann, Howe, et al. 2010). My results also indicate a discrepancy between LAL activation and hTLR4 activation, as decreasing the C3 and C3’ acyl chain lengths decreases hTLR4 activation but 140  increases LAL activation. This brings into question the effectiveness of the LAL assay for detecting differently shaped LPS molecules and casts further doubt on whether the LAL assay is a good reflection of TLR4 activation. In conclusion, I found very small decreases in the C3 and C3’ acyl chain lengths of E. coli hexa-acyl lipid A affect bacterial growth, resistance to polymyxin B, activation of hTLR4, and activation of the LAL cascade.    141  Chapter 6: Discussion 6.1 Discussion The vast variety of lipid A structures amongst different Gram-negative bacterial species and strains can be attributed to both differences in Raetz lipid A biosynthetic pathway enzymes and also the presence of additional modification enzymes (Needham and Trent 2013, Raetz, Guan, et al. 2009). It is becoming increasingly evident that modifications to the structure of lipid A allow bacteria to adapt to different environments, as evident in the interaction between various human-adapted bacteria and the immune system (Needham and Trent 2013). Activation of hTRL4 by LPS can be significantly affected by alterations in lipid A structure, especially modification of the phosphate groups, which play a key role in the dimerization and subsequent activation of TLR4 (Maeshima and Fernandez 2013, Raetz, Reynolds, et al. 2007).  In B. pertussis strain BP338, the phosphate groups of lipid A are modified with GlcN moieties via LgmA, LgmB, and LgmC, and this modification resulted in an increase in CAMP resistance, greater OM stabilization, and an increase in hTLR4 activation by LPS. I have used 14C-labelled substrate in an assay and shown LgmA functions to transfer GlcNAc onto a lipid, which is likely the inner membrane carrier lipid C55P. This is the proposed first step in lipid A GlcN-modification pathway of B. pertussis, and since the globular domain of LgmA was predicted to be cytoplasmic (Figure 7) and UDP-GlcNAc is found in the cytoplasm, I theorize that LgmA functions at the cytoplasmic side of the IM. I hypothesize the next step in this pathway is performed by the predicted cytoplasmic enzyme LgmC, which was proven to functions in the removal of the acetyl group from C55P-GlcNAc to produce C55P-GlcN (Llewellyn, Zhao, et al. 2012). Then, I predict LgmB functions to transfer GlcN from C55P onto lipid A in B. pertussis because the LgmB homolog ArnT functions in a similar manner to transfer Ara4N from C55P to lipid A in E. coli and Salmonella (Trent, Ribeiro, et al. 2001).  I had originally postulated that a flippase enzyme translocates C55P-GlcN from the cytoplasmic side of the IM to the periplasmic side of the IM, 142  followed by transfer of GlcN onto lipid A by LgmB in the periplasm. However, the three flippase candidates, LgmD, BP1945, and LgmE, were shown not to be required for the GlcN modification of lipid A. An alternative enzyme may function as this flippase, such as the putative Gram-negative lipid II flippase of peptidoglycan synthesis, MurJ, which has 14 transmembrane helices (Butler, Davis, et al. 2013, Ruiz 2008). It is also possible that a flippase is not required for this pathway, either because spontaneous flipping of C55P-GlcN is sufficient, or because the transfer of GlcN to lipid A might occur in the cytoplasm, therefore negating the need for a C55P-GlcN flippase. However, the intrinsic rate of phospholipids flipping from one leaflet to another across a cytoplasmic membrane is relatively slow (half times in the order of hours to days) (Pomorski and Menon 2006), and this is likely insufficient for the GlcN-modification pathway in B. pertussis. Furthermore, though it is possible that GlcN-modification of lipid A occurs in the cytoplasm, LgmB is predicted to have very short cytoplasmic loops but longer periplasmic loops and a predicted periplasmic C-terminal region. In addition, a homolog of LgmB, ArnT, functions at the periplasmic side of the IM in the Ara4N lipid A-modification pathway (Casella and Mitchell 2008) and numerous residues along the putative periplasmic loop 5 of ArnT are required for function (Impellitteri, Merten, et al. 2010). This suggests the functional domains of LgmB are on the periplasmic face of the IM, and a flippase is probably required to transfer C55P-GlcN from the cytoplasmic to the periplasmic face of the IM, though this flippase has yet to be identified.  An interesting aspect of the lgm locus in B. pertussis is the presence of the previously unannotated ORF lgmE, a cis-encoded antisense RNA. Though LgmE was not required for the GlcN-modification of lipid A in B. pertussis, the RNA of lgmE may play a role in the regulation of the lgm locus. The full length of this RNA is not known, however, at the very least it is complementary to part of the lgmB and lgmC messenger RNAs (mRNA), and may act as a regulatory small RNA (sRNA). If lgmABCD is transcribed as a single transcript, lgmE RNA could anneal to the middle of this long transcript. Alternatively, if lgmB and lgmC are transcribed as separate mRNA, the lgmE RNA could bind to both mRNA molecules. Most 143  cis-encoded antisense sRNAs have been studied in plasmid and phage systems, where the sRNA is often involved in maintaining copy numbers. The role of cis-ecoded antisense sRNA in chromosomal systems is less clear. Some chromosomally-encoded sRNAs promote degradation of the mRNA they anneal to whereas others can repress expression. The lgmE transcript most resembles a group of cis-encoded antisense sRNAs that are encoded by an operon and the sRNA is complementary to the region in between ORFs on the opposite strand. In this case, the sRNA could function to promote cleavage of the full length operon mRNA between the ORFs to produce two mRNAs. In the case of the lgm locus, lgmE sRNA could anneal to the region between lgmB and lgmC, therefore promoting cleavage of the mRNA into an mRNA encoding lgmA and lgmB and a separate mRNA with lgmC and lgmD. Alternatively, this operon-sRNA architecture could lead to transcription termination. In this case, if lgmABCD was transcribed as a single mRNA, lgmE sRNA would anneal to the end of lgmB and the intergenic region between lgmB and lgmC to generate a quasi-hairpin structure, resulting in intrinsic (Rho-independent) transcription termination at this site, and therefore, lower levels of lgmC translation (Waters and Storz 2009). Thus, regardless of whether LgmE is expressed, lgmE RNA could play a role in the regulation of the lgm pathway, though the extent and function of such regulation remains to be seen. lgmE could be related to the Bvg-independent regulation of the lgm locus that is linked to growth on BG agar compared to growth in SS broth (Marr, Tirsoaga, et al. 2008).  B. pertussis shows a range of susceptibility to numerous CAMPs (Fernandez and Weiss 1996), and pBD1 is able to protect against B. pertussis infection in a newborn piglet model, demonstrating the protective capability of CAMPs against this pathogen (Elahi, Buchanan, et al. 2006). Recently, Taneja et al. have shown an LPS-independent mechanism in B. pertussis that increased CAMP resistance, namely the addition of a D-alanine group to an outer membrane component by the dra gene locus (Taneja, Ganguly, et al. 2013). B. pertussis has therefore evolved two separate mechanisms for CAMP resistance (modification of an outer membrane component with D-alanine and modification of lipid A with a 144  glucosamine group), suggesting this strategy of immune evasion plays an important role in the survival of this human-restricted pathogen.  The GlcN modification plays a dual role in B. pertussis, since it both increases hTLR4 activation (Marr, Hajjar, et al. 2010), leading to greater inflammation, which would tend to lead to increased LL-37 expression or release by degranulation from attracted neutrophils (De Smet and Contreras 2005), and it also increases resistance against CAMPs. I propose this lipid A modification may have originally arisen in Bordetella as an adaptation to the environment, to increase OM stabilization and protect against CAMPs and other environmental factors. In the human host, GlcN modification of B. pertussis lipid A continues to protect against antimicrobial factors, such as CAMPs, but this change in lipid A structure also affects hTLR4 activation. My results suggest GlcN-modified lipid A provides greater resistance to some CAMPs compared to others (Figure 30), though no pattern is observed based on the charge of the CAMPs at physiological pH (Table 6). We did, however, observe greater resistance to the polymyxins, which are cyclic, lipidated peptide antibiotics produced by environmental bacterial species, and this is likely due to different interactions between lipid A and CAMPs of different sequences and structures.  Other Gram-negative bacteria also modify the negatively-charged phosphates of lipid A with positively charged amino sugars. Salmonella and E. coli add Ara4N to lipid A phosphates, as does P. aeruginosa, and F. turlarensis can modify the lipid A phosphates with galactosamine (Needham and Trent 2013). In Salmonella, E. coli, and P. aeruginosa, the addition of Ara4N to lipid A increases resistance to polymyxins B, though this has yet to be reported in F. tularensis (Needham and Trent 2013). Modification of the phosphate groups of Salmonella and E. coli LPS with another positively-charged group, phosphoethanolamine, also increases resistance to CAMPs, such as polymyxins B, as does the removal of the negatively-charged phosphate groups themselves in F. tularensis LPS (Needham and Trent 2013). These observations suggest that a more positively-charged OM is related to greater resistance to 145  positively-charged CAMPs, which supports the hypothesis that the decrease in negative charge of the OM results in a decreased affinity for the positively charged CAMPs, thus endowing greater resistance to CAMPs. I propose another contributing factor to the CAMP resistance afforded by the addition of positively-charged groups to the phosphates of lipid A may be OM stabilization by the bridging of these positively-charged groups with unmodified negatively charged phosphates (Figure 37). OM stabilization would also protect these Gram-negative bacteria from membrane-perturbation agents, such as those encountered in the environment and in hosts. Therefore, I speculate modification of lipid A with positively charged groups provide an underappreciated protection against such membrane-destabilizing agents, especially in the context of host-pathogen interactions.  However, the role of lipid A acylation patterns and modification of these acyl chains in CAMP resistance and TLR4 activation is not as clear as the role of phosphate modification. Differences in lipid A acylation patterns can affect CAMP resistance, though this is not always the case (Trent, Stead, et al. 2006). Salmonella with mutations in pagP, the enzyme that adds a secondary palmitate acyl chain to the primary C2 carbon of Salmonella lipid A to produce hepta-acyl lipid A molecules, have decreased CAMP resistance (Guo, Lim, et al. 1998). This mutation also decreases OM permeability in Salmonella, and it is therefore hypothesized that lipid A with 7 acyl chains, as found in wild-type Salmonella, has more tightly packed LPS, thus increasing the barrier activity of the OM compared to the pagP mutant, which has only hexa-acyl lipid A (Guo, Lim, et al. 1998, Trent, Stead, et al. 2006). The activity of PagP in B. bronchiseptica also generates hepta-acyl lipid A species, however this does not increase resistance to CAMPs, though in B. bronchiseptica PagP adds a secondary palmitate chain at the C3’ position (Preston, Maxim, et al. 2003). The removal of the C3 acyl chain by PagL in Salmonella would presumably result in a less-tightly packed LPS in the OM; however, this modification also does not affect CAMP resistance (Trent, Stead, et al. 2006). My results show decreasing the length of the C3 and C3’ position acyl chains in E. coli results in decreased in resistance to the CAMP polymyxins B. Decreasing the acyl chain lengths 146  in E. coli may result in LPS in the OM that is not as tightly packed, and therefore may increase OM permeability, leading to increased susceptibility to CAMPs. Overall, whereas some data support the hypothesis that greater levels of LPS acylation results in greater CAMP resistance, there are likely other factors involved, since this does not hold true in the case of PagP in Bordetella or PagL in Salmonella. Lipid A acylation patterns also play a complex role in activation of TLR4. Needham et al found hTLR4 activation is affected not just by the number of acyl chains on an LPS molecule, but also the position of the acyl chains. For example, addition of a secondary palmitate group at the C2 position of hexa-acyl E. coli LPS, to generate hepta-acyl LPS, does not significantly affect hTLR4 activation, though this same modification in penta-acyl E. coli LPS decreases hTLR4 activation (Needham, Carroll, et al. 2013). My findings show the length of these acyl chains also affect hTLR4 activation by LPS. Increasing C3’ acyl chain length in non-GlcN-modified penta-acyl B. pertussis lipid A increases activation of hTLR4, as does an increase in the C3 and C3’ acyl chain lengths in hexa-acyl E. coli. However, in the case of E. coli, it is unclear whether the effect on hTLR4 activation is due to modification of the C3 acyl chain, the C3’acyl chain, or both. Furthermore, it is unknown whether changing the length of any other acyl chains in these lipid A structures would affect hTLR4 activation. In the case of B. pertussis LPS, I observed that increasing the C3’ acyl chain in GlcN-modified LPS did not affect hTLR4 activation, therefore suggesting that the effect of some acyl chain modifications may depend on the presence or absence of other lipid A modifications.  Differential activation of hTLR4 by varying lipid A structures is due to the interaction of lipid A with the human TLR4-MD-2 receptor. The solved structures of hexa-acyl E. coli lipid A, with human TLR4-MD-2 and tetra-acyl lipid IVA bound to human MD-2 has revealed some interactions that are important for hTLR4 activation. E. coli lipid A is a good agonist of hTLR4 whereas lipid IVA is a poor activator of hTLR4. In general, these structures indicate the hydrophobic acyl chains of lipid A and lipid IVA sit inside the pocket formed by MD-2, and the phosphate groups sit above the MD-2 pocket to interact with amino 147  acid residues in both TLR4 and TLR4*, thereby promoting dimerization and activation (Figure 3) (Ohto, Fukase, et al. 2007, Park, Song, et al. 2009). When comparing the solved structures for E. coli lipid A and lipid IVA within MD-2, the two molecules are sitting in these co-crystals in opposite orientations, such that the 1-phosphate of E. coli lipid A is in the equivalent position as the 4’-phosphate of lipid IVA (Park, Song, et al. 2009). The orientation of the lipid A or lipid IVA inside MD-2 can have important implications for which phosphate group (1- or 4’-) and which acyl chain interacts with the dimerizing TLR4* (Figure 3) during activation of TLR4. It is possible that the specific orientation of the di-GlcN backbone in each of these structures is due to a crystallization artifact, and each molecule is able to interact with MD-2-TLR4 in either orientation in vivo. Furthermore, since there are fewer acyl chains to fit inside the MD-2 pocket in lipid IVA, the entire molecule sits deeper inside MD-2 and the phosphate groups of lipid IVA are positioned lower and at an angle compared to the phosphate groups of E. coli lipid A. This difference in phosphate group positioning is likely a main cause of the decreased hTLR4 activation by lipid IVA compared to E. coli lipid A, since these negatively-charged phosphate groups interact with positively-charged amino acid residues in TLR4 and TLR4*. Another important difference in these structures is the position of the acyl chains: all four acyl chains of lipid IVA sit inside the MD-2 pocket, however only five of the six acyl-chains of E. coli lipid A sit inside the MD-2 pocket, the sixth C2 acyl chain is positioned between the MD-2 molecule and TLR4*. The interaction of the C2 acyl chain with TLR4* is hypothesized to be important for TLR4 activation, though if hexa-acyl lipid A molecules can interact with TLR4-MD-2 in the flipped orientation, other acyl chains, such as the C3’ acyl chain, may also be involved in this manner (Maeshima and Fernandez 2013, Park, Song, et al. 2009). I have shown that decreasing C3 and C3’ acyl chain lengths in hexa-acyl E. coli lipid A results in a decrease in hTLR4 activation. Based on the structure of E. coli lipid A and TLR4-MD-2, I hypothesize that the decrease in the acyl chain lengths likely lowers the position of the lipid A molecule in the MD-2 pocket, therefore positioning the phosphate groups at a suboptimal level. This would result in decreased interaction between the phosphate groups with positively charged amino acids in TLR4 and TLR4*, 148  therefore resulting in a reduction of hTLR4 activation. If E. coli LPS can also interact with TLR4-MD-2 in the opposite orientation, such that the 1-phosphate interacts with TLR4*, it is also possible the C3’ acyl chain may sit outside of the MD-2 pocket and interact with TLR4*, in which case a shorter C3’ acyl chain would also decrease interaction with TLR4*.   Unlike E. coli, many Gram-negative species have penta-acyl lipid A, as found in B. pertussis. Though the structure of TLR4-MD-2 bound to penta-acyl lipid A has yet to be determined, the solved structures of tetra-acyl lipid IVA and hexa-acyl lipid A within MD-2 can also shed light on how B. pertussis lipid A may interact with TLR4-MD-2 (Ohto, Fukase, et al. 2007, Park, Song, et al. 2009). However, it still remains unclear which orientation a penta-acyl lipid A molecule would have inside the MD-2 pocket; would it sit with the 4’-phosphate interacting with TLR4*, as lipid IVA does, or would it sit with the 1-phosphate in this position, as seen in hexa-acyl lipid A? Furthermore, based on these structures alone, it is unclear whether all five acyl chains of B. pertussis lipid A would sit inside the MD-2 pocket or whether a single acyl chain would extrude from this pocket to interact with TLR4*. Recent work in our laboratory suggests that all 5 acyl chains of B. pertussis lipid A lie within the MD-2 pocket in vivo and the dimerization of TLR4 is instead mediated by a MD-2-TLR4* interaction (Maeshima N, Fernandez RC, unpublished data). Despite these uncertainties, the aforementioned solved structures still provide a basis for interaction between TLR4-MD-2, lipid A, and the incoming dimerizing TLR4*-MD-2*-lipid A (Figure 3). In the case of B. pertussis lipid A, which has only five acyl chains, I propose, in the absence of the GlcN modification on the phosphate groups and with shorter C3′ acyl chains, as seen in 18-323, the phosphate groups of lipid A are not able to interact with the key positively charged residues in TLR4 and/or TLR4* to cause dimerization and activation. However, when the C3′ acyl chain length is increased, this may position the lipid A molecule slightly more out of the MD-2 pocket and slightly closer to the positively charged residues in TLR4 and TLR4*, thus significantly increasing TLR4 dimerization and activation. Alternatively, when the phosphates are modified with positively charged GlcN, there is 149  greater hTLR4 activation when compared with unmodified lipid A. Based on ongoing work in our laboratory, the GlcN moieties on this penta-acyl lipid A likely interacts with negatively charged residues in TLR4 and TLR4* to promote dimerization (Maeshima N, Evans-Atkinson T, Hajjar AM, Fernandez RC, in revision). Adding both GlcN and longer acyl chain modifications to 18-323 lipid A also increases hTLR4 activation when compared with both wild type 18-323 and 18-323 with only the longer acyl chain modification, but this level is statistically equal to that of 18-323 with only the GlcN modification. This suggests that the slight increase in the C3′ acyl chain length in the GlcN-modified lipid A does not change the position within the TLR4-MD-2 complex enough to significantly affect the interaction between the negatively charged residues in TLR4 and TLR4* and the GlcNs.  Since activation of TLR4 by LPS leads to downstream NFκB and IRF3 activation and the generation of both an inflammatory response and a type I interferon response, respectively, LPS is a possible candidate for pharmacological uses, exemplified by the use of monophosphoryl lipid A as an adjuvant (Maeshima and Fernandez 2013). What makes LPS an even more attractive contender is the possibility of modulating the activation of hTLR4 by adjusting the structure of the lipid A region of LPS. Therefore, building a repertoire of tools to modify the structure of lipid A could aid in the design of superior adjuvants – even slight changes to the lipid A, such as changing the acyl chain lengths, may allow fine-tuning the hTLR4 response.  6.2 Conclusions In conclusion, I have shown lgmA, lgmB, and lgmC are required for glycosylation of lipid A in B. pertussis, whereas lgmD, lgmE, and BP1945 are not. This leaves the identity or requirement for a C55P-GlcN flippase in the GlcN modification pathway unclear. However, I have demonstrated LgmA transfers GlcNAc from UDP-GlcNAc onto a lipid, which is likely the C55P, therefore supporting the first proposed step in this pathway. LgmC has been shown by others to function in the removal of the acetyl group from 150  C55P-GlcNAc to generate C55P-GlcN (Llewellyn, Zhao, et al. 2012), which also supports the proposed GlcN modification pathway. Furthermore I characterized a putative active site for LgmA (D76, D77, D127, and D129) and LgmC (D80, D81, H130, D187, H189, and E313).   I also elucidated the genetic basis for the structural differences in lipid A between B. pertussis strains BP338 and 18-323. The lack of GlcN modification in 18-323 is due to the absence of a complete lgm locus and the lack of 14-carbon long acyl chains at the C3’ position of the lipid A in 18-323 is due to a single amino acid difference in the hydrocarbon ruler region of the enzyme LpxA between the two B. pertussis strains.  Furthermore, I demonstrated lipid A modifications in B. pertussis affect different biological interactions. The GlcN modification increases B. pertussis resistance to a variety of CAMPs, though not to the aminoglycoside gentamicin, and also increases resistance to OM perturbation. Additionally, the GlcN modification and longer C3’ acyl chain lengths in B. pertussis lipid A individually increase activation of hTLR4. However, when both modifications are present, the same level of hTLR4 activation is observed as with only the GlcN modification, suggesting the GlcN modification may have a dominant effect on increasing activation.  In regards to E. coli lipid A structural differences, I demonstrated shortening the C3 and C3’ acyl chains results in a minor growth defect, decreases resistance to polymyxin B, attenuates activation of hTLR4 by LPS, and increases LAL activation by LPS.  6.3 Future directions There are many avenues for future research related to this project. Firstly, there are still questions regarding the regulation of the lgm locus, and consequently the GlcN modification of LPS, in B. pertussis. 151  lgmA and lgmB have a higher level of transcription when the bacteria are grown in SS broth compared to on BG agar (Marr, Tirsoaga, et al. 2008). The signal facilitating this difference in transcription is still unknown. Furthermore, there may be other Bvg-independent environmental signals, such as growth phase or the presence of CAMPs, which may further modulate regulation of the lgm locus.  Linked to these inquiries is the role of lgmE. I found lgmE mRNA overlaps, and would therefore anneal to, at least lgmB and lgmC mRNA. However, depending on the actual length of the lgmE transcript, which is still unknown, it may also overlap with lgmA. Additionally, the manner in which lgmABCD is transcribed, e.g. as one single transcript or as four individual transcripts, may affect how lgmE interacts and regulates expression of the lgm locus genes. If the lgmE transcript does anneal to lgmABCD mRNAs, then determining whether this association promotes the degradation of lgmABCD mRNAs or protects the lgmABCD mRNAs from degradation would shed light on how this locus is regulated. Also, lgmE may play a role in the non-Bvg-associated regulation of lgmA and lgmB.  Another avenue of research is explaining the lack of GlcN-modified lipid A in B. hinzii and B. trematum, since both species have a complete copy of lgmA, lgmB, and lgmC; thus far, the only genes that are required for this modification in B. pertussis. It is possible that another, as yet unidentified gene is required for the GlcN modification in B. pertussis, and this additional gene is not present in B. hinzii or B. trematum. Alternatively, an inactivating mutation in lgmA, lgmB, or lgmC could be responsible for the lack of modification in B. hinzii and B. trematum, or the lgm genes may not be expressed in these two species.  Though I have shown evidence that LgmA transfers GlcNAc onto a lipid, which is likely C55P, we have yet to prove the lipid acceptor of this reaction is the inner membrane carrier lipid C55P. Furthermore, the function of LgmB, the glycosyltransferase predicted to transfer GlcN from C55P to the phosphate group 152  of lipid A, has also not been proven. To complete the model for the Lgm pathway, determining whether a flippase is required, and if so, identifying this elusive flippase would also be important. Determining which face of the IM each Lgm enzyme functions at may reveal clues as to the requirement of a flippase in this system, because if all the Lgm enzymes function at the cytoplasmic face of the IM, there would be no need for a C55P-sugar flippase.  Further analysis of the LgmA and LgmC active sites is another topic to research. The amino acids I identified as possibly being part of the putative active sites of these enzymes need to be confirmed by further mutational analysis, perhaps by mutating the residues to amino acids that take up the same amount of space, but lack the functional group. Additional residues can also be targeted for mutational analysis, to broaden the characterization of the active sites of not only LgmA and LgmC, but also LgmB. GT83 glycosyltransferases are not well studied, so further study of LgmB, a member of this protein family, may lead to insights applicable to other GT83 protein family members.   153  References 1. Alonso, S., N. Reveneau, K. Pethe, and C. Locht. 2002. Eighty-kilodalton N-terminal moiety of Bordetella pertussis filamentous hemagglutinin: adherence, immunogenicity, and protective role. Infection and immunity 70:4142-4147. 2. Anderson, M. S., and C. R. Raetz. 1987. Biosynthesis of lipid A precursors in Escherichia coli. A cytoplasmic acyltransferase that converts UDP-N-acetylglucosamine to UDP-3-O-(R-3-hydroxymyristoyl)-N-acetylglucosamine. The Journal of biological chemistry 262:5159-5169. 3. Banemann, A., H. Deppisch, and R. Gross. 1998. The lipopolysaccharide of Bordetella bronchiseptica acts as a protective shield against antimicrobial peptides. Infection and immunity 66:5607-5612. 4. Barnes, M. G., and A. A. Weiss. 2001. BrkA protein of Bordetella pertussis inhibits the classical pathway of complement after C1 deposition. Infection and immunity 69:3067-3072. 5. Boekema, B. K., N. Stockhofe-Zurwieden, H. E. Smith, E. M. Kamp, J. P. van Putten, and J. H. Verheijden. 2003. Adherence of Actinobacillus pleuropneumoniae to primary cultures of porcine lung epithelial cells. Veterinary microbiology 93:133-144. 6. Bradford, M. M. 1976. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Analytical biochemistry 72:248-254. 7. Breazeale, S. D., A. A. Ribeiro, A. L. McClerren, and C. R. Raetz. 2005. A formyltransferase required for polymyxin resistance in Escherichia coli and the modification of lipid A with 4-Amino-4-deoxy-L-arabinose. Identification and function oF UDP-4-deoxy-4-formamido-L-arabinose. The Journal of biological chemistry 280:14154-14167. 8. Butler, E. K., R. M. Davis, V. Bari, P. A. Nicholson, and N. Ruiz. 2013. Structure-function analysis of MurJ reveals a solvent-exposed cavity containing residues essential for peptidoglycan biogenesis in Escherichia coli. Journal of bacteriology 195:4639-4649. 9. Carbonetti, N. H. 2010. Pertussis toxin and adenylate cyclase toxin: key virulence factors of Bordetella pertussis and cell biology tools. Future microbiology 5:455-469. 10. Caroff, M., J. Brisson, A. Martin, and D. Karibian. 2000. Structure of the Bordetella pertussis 1414 endotoxin. FEBS letters 477:8-14. 11. Caroff, M., and D. Karibian. 2003. Structure of bacterial lipopolysaccharides. Carbohydrate research 338:2431-2447. 12. Caroff, M., D. Karibian, J. M. Cavaillon, and N. Haeffner-Cavaillon. 2002. Structural and functional analyses of bacterial lipopolysaccharides. Microbes and infection / Institut Pasteur 4:915-926. 13. Casella, C. R., and T. C. Mitchell. 2008. Putting endotoxin to work for us: monophosphoryl lipid A as a safe and effective vaccine adjuvant. Cellular and molecular life sciences : CMLS 65:3231-3240. 14. Chaby, R., I. Garcia-Verdugo, Q. Espinassous, and L. A. Augusto. 2005. Interactions between LPS and lung surfactant proteins. Journal of endotoxin research 11:181-185. 15. Chen, L., and N. Mozier. 2013. Comparison of Limulus amebocyte lysate test methods for endotoxin measurement in protein solutions. Journal of pharmaceutical and biomedical analysis 80:180-185. 16. Chng, S. S., L. S. Gronenberg, and D. Kahne. 2010. Proteins required for lipopolysaccharide assembly in Escherichia coli form a transenvelope complex. Biochemistry 49:4565-4567. 17. Clementz, T., and C. R. Raetz. 1991. A gene coding for 3-deoxy-D-manno-octulosonic-acid transferase in Escherichia coli. Identification, mapping, cloning, and sequencing. The Journal of biological chemistry 266:9687-9696. 154  18. Coats, S. R., A. B. Berezow, T. T. To, S. Jain, B. W. Bainbridge, K. P. Banani, and R. P. Darveau. 2011. The lipid A phosphate position determines differential host Toll-like receptor 4 responses to phylogenetically related symbiotic and pathogenic bacteria. Infection and immunity 79:203-210. 19. Crowcroft, N. S., C. Stein, P. Duclos, and M. Birmingham. 2003. How best to estimate the global burden of pertussis? Lancet Infect Dis 3:413-418. 20. Cummings, C. A., H. J. Bootsma, D. A. Relman, and J. F. Miller. 2006. Species- and strain-specific control of a complex, flexible regulon by Bordetella BvgAS. Journal of bacteriology 188:1775-1785. 21. de Gouw, D., D. A. Diavatopoulos, H. J. Bootsma, P. W. Hermans, and F. R. Mooi. 2011. Pertussis: a matter of immune modulation. FEMS microbiology reviews 35:441-474. 22. De Smet, K., and R. Contreras. 2005. Human antimicrobial peptides: defensins, cathelicidins and histatins. Biotechnology letters 27:1337-1347. 23. Decker, K. B., T. D. James, S. Stibitz, and D. M. Hinton. 2012. The Bordetella pertussis model of exquisite gene control by the global transcription factor BvgA. Microbiology 158:1665-1676. 24. El Hamidi, A., A. Tirsoaga, A. Novikov, A. Hussein, and M. Caroff. 2005. Microextraction of bacterial lipid A: easy and rapid method for mass spectrometric characterization. Journal of lipid research 46:1773-1778. 25. Elahi, S., R. M. Buchanan, S. Attah-Poku, H. G. Townsend, L. A. Babiuk, and V. Gerdts. 2006. The host defense peptide beta-defensin 1 confers protection against Bordetella pertussis in newborn piglets. Infection and immunity 74:2338-2352. 26. Fedele, G., M. Bianco, and C. M. Ausiello. 2013. The Virulence Factors of Bordetella pertussis: Talented Modulators of Host Immune Response. Archivum immunologiae et therapiae experimentalis 61:445-457. 27. Fernandez, L., H. Jenssen, M. Bains, I. Wiegand, W. J. Gooderham, and R. E. Hancock. 2012. The two-component system CprRS senses cationic peptides and triggers adaptive resistance in Pseudomonas aeruginosa independently of ParRS. Antimicrobial agents and chemotherapy 56:6212-6222. 28. Fernandez, R. C., and A. A. Weiss. 1994. Cloning and sequencing of a Bordetella pertussis serum resistance locus. Infection and immunity 62:4727-4738. 29. Fernandez, R. C., and A. A. Weiss. 1996. Susceptibilities of Bordetella pertussis strains to antimicrobial peptides. Antimicrobial agents and chemotherapy 40:1041-1043. 30. Freinkman, E., S. S. Chng, and D. Kahne. 2011. The complex that inserts lipopolysaccharide into the bacterial outer membrane forms a two-protein plug-and-barrel. Proceedings of the National Academy of Sciences of the United States of America 108:2486-2491. 31. Furste, J. P., W. Pansegrau, R. Frank, H. Blocker, P. Scholz, M. Bagdasarian, and E. Lanka. 1986. Molecular cloning of the plasmid RP4 primase region in a multi-host-range tacP expression vector. Gene 48:119-131. 32. Galloway, S. M., and C. R. Raetz. 1990. A mutant of Escherichia coli defective in the first step of endotoxin biosynthesis. The Journal of biological chemistry 265:6394-6402. 33. Gascuel, O. 1997. BIONJ: an improved version of the NJ algorithm based on a simple model of sequence data. Molecular biology and evolution 14:685-695. 34. Gellatly, S. L., B. Needham, L. Madera, M. S. Trent, and R. E. Hancock. 2012. The Pseudomonas aeruginosa PhoP-PhoQ two-component regulatory system is induced upon interaction with epithelial cells and controls cytotoxicity and inflammation. Infection and immunity 80:3122-3131. 35. Gerlach, G., F. von Wintzingerode, B. Middendorf, and R. Gross. 2001. Evolutionary trends in the genus Bordetella. Microbes and Infection 3:61-72. 36. Geuijen, C. A., R. J. Willems, and F. R. Mooi. 1996. The major fimbrial subunit of Bordetella pertussis binds to sulfated sugars. Infection and immunity 64:2657-2665. 155  37. Gronow, S., W. Brabetz, and H. Brade. 2000. Comparative functional characterization in vitro of heptosyltransferase I (WaaC) and II (WaaF) from Escherichia coli. European journal of biochemistry / FEBS 267:6602-6611. 38. Guan, S., D. A. Bastin, and N. K. Verma. 1999. Functional analysis of the O antigen glucosylation gene cluster of Shigella flexneri bacteriophage SfX. Microbiology 145 ( Pt 5):1263-1273. 39. Guex, N., and M. C. Peitsch. 1997. SWISS-MODEL and the Swiss-PdbViewer: an environment for comparative protein modeling. Electrophoresis 18:2714-2723. 40. Gunn, J. S., K. B. Lim, J. Krueger, K. Kim, L. Guo, M. Hackett, and S. I. Miller. 1998. PmrA-PmrB-regulated genes necessary for 4-aminoarabinose lipid A modification and polymyxin resistance. Molecular microbiology 27:1171-1182. 41. Guo, L., K. B. Lim, C. M. Poduje, M. Daniel, J. S. Gunn, M. Hackett, and S. I. Miller. 1998. Lipid A acylation and bacterial resistance against vertebrate antimicrobial peptides. Cell 95:189-198. 42. Gustafsson, L., L. Hessel, J. Storsaeter, and P. Olin. 2006. Long-term follow-up of Swedish children vaccinated with acellular pertussis vaccines at 3, 5, and 12 months of age indicates the need for a booster dose at 5 to 7 years of age. Pediatrics 118:978-984. 43. Gutsmann, T., J. Howe, U. Zahringer, P. Garidel, A. B. Schromm, M. H. Koch, Y. Fujimoto, K. Fukase, I. Moriyon, G. Martinez-de-Tejada, and K. Brandenburg. 2010. Structural prerequisites for endotoxic activity in the Limulus test as compared to cytokine production in mononuclear cells. Innate immunity 16:39-47. 44. Hallander, H. O., L. Gustafsson, M. Ljungman, and J. Storsaeter. 2005. Pertussis antitoxin decay after vaccination with DTPa Response to a first booster dose 31/2-61/2 years after the third vaccine dose. Vaccine 23:5359-5364. 45. Hancock, R. E. 1997. The bacterial outer membrane as a drug barrier. Trends in microbiology 5:37-42. 46. Hazenbos, W. L., C. A. Geuijen, B. M. van den Berg, F. R. Mooi, and R. van Furth. 1995. Bordetella pertussis fimbriae bind to human monocytes via the minor fimbrial subunit FimD. The Journal of infectious diseases 171:924-929. 47. Imagawa, T., H. Iino, M. Kanagawa, A. Ebihara, S. Kuramitsu, and H. Tsuge. 2008. Crystal structure of the YdjC-family protein TTHB029 from Thermus thermophilus HB8: structural relationship with peptidoglycan N-acetylglucosamine deacetylase. Biochemical and biophysical research communications 367:535-541. 48. Impellitteri, N. A., J. A. Merten, L. E. Bretscher, and C. S. Klug. 2010. Identification of a functionally important loop in Salmonella typhimurium ArnT. Biochemistry 49:29-35. 49. Inatsuka, C. S., Q. Xu, I. Vujkovic-Cvijin, S. Wong, S. Stibitz, J. F. Miller, and P. A. Cotter. 2010. Pertactin is required for Bordetella species to resist neutrophil-mediated clearance. Infection and immunity 78:2901-2909. 50. Kagan, J. C., T. Su, T. Horng, A. Chow, S. Akira, and R. Medzhitov. 2008. TRAM couples endocytosis of Toll-like receptor 4 to the induction of interferon-beta. Nature immunology 9:361-368. 51. Kawasaki, K., R. K. Ernst, and S. I. Miller. 2004. Deacylation and palmitoylation of lipid A by Salmonellae outer membrane enzymes modulate host signaling through Toll-like receptor 4. Journal of endotoxin research 10:439-444. 52. Kelley, L. A., and M. J. Sternberg. 2009. Protein structure prediction on the Web: a case study using the Phyre server. Nature protocols 4:363-371. 53. Kelly, T. M., S. A. Stachula, C. R. Raetz, and M. S. Anderson. 1993. The firA gene of Escherichia coli encodes UDP-3-O-(R-3-hydroxymyristoyl)-glucosamine N-acyltransferase. The third step of endotoxin biosynthesis. The Journal of biological chemistry 268:19866-19874. 156  54. Ko, K. S., K. R. Peck, W. S. Oh, N. Y. Lee, J. H. Lee, and J. H. Song. 2005. New species of Bordetella, Bordetella ansorpii sp. nov., isolated from the purulent exudate of an epidermal cyst. Journal of clinical microbiology 43:2516-2519. 55. Kovach, M. E., P. H. Elzer, D. S. Hill, G. T. Robertson, M. A. Farris, R. M. Roop, 2nd, and K. M. Peterson. 1995. Four new derivatives of the broad-host-range cloning vector pBBR1MCS, carrying different antibiotic-resistance cassettes. Gene 166:175-176. 56. Lairson, L. L., B. Henrissat, G. J. Davies, and S. G. Withers. 2008. Glycosyltransferases: structures, functions, and mechanisms. Annual review of biochemistry 77:521-555. 57. Laube, D. M., S. Yim, L. K. Ryan, K. O. Kisich, and G. Diamond. 2006. Antimicrobial peptides in the airway. Current topics in microbiology and immunology 306:153-182. 58. Leung, A. K. C., W. L. M. Robson, and H. D. Davies. 2007. Pertussis in adolescents. Adv Ther 24:353-361. 59. Llewellyn, A. C., J. Zhao, F. Song, J. Parvathareddy, Q. Xu, B. A. Napier, H. Laroui, D. Merlin, J. E. Bina, P. A. Cotter, M. A. Miller, C. R. Raetz, and D. S. Weiss. 2012. NaxD is a deacetylase required for lipid A modification and Francisella pathogenesis. Molecular microbiology 86:611-627. 60. Maeshima, N., and R. C. Fernandez. 2013. Recognition of lipid A variants by the TLR4-MD-2 receptor complex. Frontiers in cellular and infection microbiology 3:3. 61. Marr, N., A. M. Hajjar, N. R. Shah, A. Novikov, C. S. Yam, M. Caroff, and R. C. Fernandez. 2010. Substitution of the Bordetella pertussis lipid A phosphate groups with glucosamine is required for robust NF-kappaB activation and release of proinflammatory cytokines in cells expressing human but not murine Toll-like receptor 4-MD-2-CD14. Infection and immunity 78:2060-2069. 62. Marr, N., A. Novikov, A. M. Hajjar, M. Caroff, and R. C. Fernandez. 2010. Variability in the lipooligosaccharide structure and endotoxicity among Bordetella pertussis strains. The Journal of infectious diseases 202:1897-1906. 63. Marr, N., N. R. Shah, R. Lee, E. J. Kim, and R. C. Fernandez. 2011. Bordetella pertussis autotransporter Vag8 binds human C1 esterase inhibitor and confers serum resistance. PloS one 6:e20585. 64. Marr, N., A. Tirsoaga, D. Blanot, R. Fernandez, and M. Caroff. 2008. Glucosamine found as a substituent of both phosphate groups in Bordetella lipid A backbones: role of a BvgAS-activated ArnT ortholog. Journal of bacteriology 190:4281-4290. 65. Mattoo, S., and J. D. Cherry. 2005. Molecular pathogenesis, epidemiology, and clinical manifestations of respiratory infections due to Bordetella pertussis and other Bordetella subspecies. Clinical microbiology reviews 18:326-382. 66. Merkel, T. J., P. E. Boucher, S. Stibitz, and V. K. Grippe. 2003. Analysis of bvgR expression in Bordetella pertussis. Journal of bacteriology 185:6902-6912. 67. Metzger, L. E. t., and C. R. Raetz. 2010. An alternative route for UDP-diacylglucosamine hydrolysis in bacterial lipid A biosynthesis. Biochemistry 49:6715-6726. 68. Mielcarek, N., A. S. Debrie, D. Raze, J. Quatannens, J. Engle, W. E. Goldman, and C. Locht. 2006. Attenuated Bordetella pertussis: new live vaccines for intranasal immunisation. Vaccine 24:S54-S55. 69. Miller, S. I., R. K. Ernst, and M. W. Bader. 2005. LPS, TLR4 and infectious disease diversity. Nature reviews. Microbiology 3:36-46. 70. Moskowitz, S. M., R. K. Ernst, and S. I. Miller. 2004. PmrAB, a two-component regulatory system of Pseudomonas aeruginosa that modulates resistance to cationic antimicrobial peptides and addition of aminoarabinose to lipid A. Journal of bacteriology 186:575-579. 71. Muta, T., T. Miyata, Y. Misumi, F. Tokunaga, T. Nakamura, Y. Toh, Y. Ikehara, and S. Iwanaga. 1991. Limulus factor C. An endotoxin-sensitive serine protease zymogen with a mosaic 157  structure of complement-like, epidermal growth factor-like, and lectin-like domains. The Journal of biological chemistry 266:6554-6561. 72. Needham, B. D., S. M. Carroll, D. K. Giles, G. Georgiou, M. Whiteley, and M. S. Trent. 2013. Modulating the innate immune response by combinatorial engineering of endotoxin. Proceedings of the National Academy of Sciences of the United States of America 110:1464-1469. 73. Needham, B. D., and M. S. Trent. 2013. Fortifying the barrier: the impact of lipid A remodelling on bacterial pathogenesis. Nature reviews. Microbiology 11:467-481. 74. Novikov, A., N. R. Shah, S. Albitar-Nehme, S. M. Basheer, I. Trento, A. Tirsoaga, M. Moksa, M. Hirst, M. B. Perry, A. E. Hamidi, R. C. Fernandez, and M. Caroff. 2013. Complete Bordetella avium, Bordetella hinzii and Bordetella trematum lipid A structures and genomic sequence analyses of the loci involved in their modifications. Innate immunity. 75. Ohto, U., K. Fukase, K. Miyake, and Y. Satow. 2007. Crystal structures of human MD-2 and its complex with antiendotoxic lipid IVa. Science 316:1632-1634. 76. Oliver, D. C., and R. C. Fernandez. 2001. Antibodies to BrkA augment killing of Bordetella pertussis. Vaccine 20:235-241. 77. Park, B. S., D. H. Song, H. M. Kim, B. S. Choi, H. Lee, and J. O. Lee. 2009. The structural basis of lipopolysaccharide recognition by the TLR4-MD-2 complex. Nature 458:1191-1195. 78. Park, J., Y. Zhang, A. M. Buboltz, X. Zhang, S. C. Schuster, U. Ahuja, M. Liu, J. F. Miller, M. Sebaihia, S. D. Bentley, J. Parkhill, and E. T. Harvill. 2012. Comparative genomics of the classical Bordetella subspecies: the evolution and exchange of virulence-associated diversity amongst closely related pathogens. BMC genomics 13:545. 79. Parkhill, J., M. Sebaihia, A. Preston, L. D. Murphy, N. Thomson, D. E. Harris, M. T. Holden, C. M. Churcher, S. D. Bentley, K. L. Mungall, A. M. Cerdeno-Tarraga, L. Temple, K. James, B. Harris, M. A. Quail, M. Achtman, R. Atkin, S. Baker, D. Basham, N. Bason, I. Cherevach, T. Chillingworth, M. Collins, A. Cronin, P. Davis, J. Doggett, T. Feltwell, A. Goble, N. Hamlin, H. Hauser, S. Holroyd, K. Jagels, S. Leather, S. Moule, H. Norberczak, S. O'Neil, D. Ormond, C. Price, E. Rabbinowitsch, S. Rutter, M. Sanders, D. Saunders, K. Seeger, S. Sharp, M. Simmonds, J. Skelton, R. Squares, S. Squares, K. Stevens, L. Unwin, S. Whitehead, B. G. Barrell, and D. J. Maskell. 2003. Comparative analysis of the genome sequences of Bordetella pertussis, Bordetella parapertussis and Bordetella bronchiseptica. Nature genetics 35:32-40. 80. Plotz, B. M., B. Lindner, K. O. Stetter, and O. Holst. 2000. Characterization of a novel lipid A containing D-galacturonic acid that replaces phosphate residues - The structure of the lipid A of the lipopolysaccharide from the hyperthermophilic bacterium Aquifex pyrophilus. Journal of Biological Chemistry 275:11222-11228. 81. Pomorski, T., and A. K. Menon. 2006. Lipid flippases and their biological functions. Cellular and molecular life sciences : CMLS 63:2908-2921. 82. Preston, A., E. Maxim, E. Toland, E. J. Pishko, E. T. Harvill, M. Caroff, and D. J. Maskell. 2003. Bordetella bronchiseptica PagP is a Bvg-regulated lipid A palmitoyl transferase that is required for persistent colonization of the mouse respiratory tract. Molecular microbiology 48:725-736. 83. Punta, M., P. C. Coggill, R. Y. Eberhardt, J. Mistry, J. Tate, C. Boursnell, N. Pang, K. Forslund, G. Ceric, J. Clements, A. Heger, L. Holm, E. L. Sonnhammer, S. R. Eddy, A. Bateman, and R. D. Finn. 2012. The Pfam protein families database. Nucleic acids research 40:D290-301. 84. Raetz, C. R., Z. Guan, B. O. Ingram, D. A. Six, F. Song, X. Wang, and J. Zhao. 2009. Discovery of new biosynthetic pathways: the lipid A story. Journal of lipid research 50 Suppl:S103-108. 158  85. Raetz, C. R., C. M. Reynolds, M. S. Trent, and R. E. Bishop. 2007. Lipid A modification systems in gram-negative bacteria. Annual review of biochemistry 76:295-329. 86. Raetz, C. R., and C. Whitfield. 2002. Lipopolysaccharide endotoxins. Annual review of biochemistry 71:635-700. 87. Ram, S., J. Ngampasutadol, A. D. Cox, A. M. Blom, L. A. Lewis, F. St Michael, J. Stupak, S. Gulati, and P. A. Rice. 2007. Heptose I glycan substitutions on Neisseria gonorrhoeae lipooligosaccharide influence C4b-binding protein binding and serum resistance. Infection and immunity 75:4071-4081. 88. Ravishankar, S., V. P. Kumar, B. Chandrakala, R. K. Jha, S. M. Solapure, and S. M. de Sousa. 2005. Scintillation proximity assay for inhibitors of Escherichia coli MurG and, optionally, MraY. Antimicrobial agents and chemotherapy 49:1410-1418. 89. Ray, B. L., and C. R. Raetz. 1987. The biosynthesis of gram-negative endotoxin. A novel kinase in Escherichia coli membranes that incorporates the 4'-phosphate of lipid A. The Journal of biological chemistry 262:1122-1128. 90. Rezania, S., N. Amirmozaffari, B. Tabarraei, M. Jeddi-Tehrani, O. Zarei, R. Alizadeh, F. Masjedian, and A. H. Zarnani. 2011. Extraction, Purification and Characterization of Lipopolysaccharide from Escherichia coli and Salmonella typhi. Avicenna journal of medical biotechnology 3:3-9. 91. Robins, L. I., A. H. Williams, and C. R. Raetz. 2009. Structural basis for the sugar nucleotide and acyl-chain selectivity of Leptospira interrogans LpxA. Biochemistry 48:6191-6201. 92. Rolin, O., S. J. Muse, C. Safi, S. Elahi, V. Gerdts, L. E. Hittle, R. K. Ernst, E. T. Harvill, and A. Preston. 2014. Enzymatic Modification of Lipid A by ArnT Protects Bordetella bronchiseptica against Cationic Peptides and Is Required for Transmission. Infection and immunity 82:491-499. 93. Rosenthal, R. S., W. Nogami, B. T. Cookson, W. E. Goldman, and W. J. Folkening. 1987. Major fragment of soluble peptidoglycan released from growing Bordetella pertussis is tracheal cytotoxin. Infection and immunity 55:2117-2120. 94. Ruiz, N. 2008. Bioinformatics identification of MurJ (MviN) as the peptidoglycan lipid II flippase in Escherichia coli. Proceedings of the National Academy of Sciences of the United States of America 105:15553-15557. 95. Ruiz, N., L. S. Gronenberg, D. Kahne, and T. J. Silhavy. 2008. Identification of two inner-membrane proteins required for the transport of lipopolysaccharide to the outer membrane of Escherichia coli. Proceedings of the National Academy of Sciences of the United States of America 105:5537-5542. 96. Schaeffer, L. M., F. X. McCormack, H. Wu, and A. A. Weiss. 2004. Bordetella pertussis lipopolysaccharide resists the bactericidal effects of pulmonary surfactant protein A. J Immunol 173:1959-1965. 97. Schagger, H., and G. von Jagow. 1987. Tricine-sodium dodecyl sulfate-polyacrylamide gel electrophoresis for the separation of proteins in the range from 1 to 100 kDa. Analytical biochemistry 166:368-379. 98. Shah, N. R., S. Albitar-Nehme, E. Kim, N. Marr, A. Novikov, M. Caroff, and R. C. Fernandez. 2013. Minor modifications to the phosphate groups and the C3' acyl chain length of lipid A in two Bordetella pertussis strains, BP338 and 18-323, independently affect Toll-like receptor 4 protein activation. The Journal of biological chemistry 288:11751-11760. 99. Shah, N. R., M. Moksa, A. Novikov, M. B. Perry, M. Hirst, M. Caroff, and R. C. Fernandez. 2013. Draft Genome Sequences of Bordetella hinzii and Bordetella trematum. Genome announcements 1. 100. Simon, R., U. Priefer, and A. Puhler. 1983. A broad host range mobilization system for invivo genetic-engineering - transposon mutagenesis in Gram-negative bacteria. Bio-Technol 1:784-791. 159  101. Song, F., Z. Guan, and C. R. Raetz. 2009. Biosynthesis of undecaprenyl phosphate-galactosamine and undecaprenyl phosphate-glucose in Francisella novicida. Biochemistry 48:1173-1182. 102. Sonnhammer, E. L., G. von Heijne, and A. Krogh. 1998. A hidden Markov model for predicting transmembrane helices in protein sequences. Proceedings / ... International Conference on Intelligent Systems for Molecular Biology ; ISMB. International Conference on Intelligent Systems for Molecular Biology 6:175-182. 103. Spears, P. A., L. M. Temple, D. M. Miyamoto, D. J. Maskell, and P. E. Orndorff. 2003. Unexpected similarities between Bordetella avium and other pathogenic bordetellae. Infection and immunity 71:2591-2597. 104. Sperandeo, P., G. Deho, and A. Polissi. 2009. The lipopolysaccharide transport system of Gram-negative bacteria. Biochimica et biophysica acta 1791:594-602. 105. Stainer, D. W., and M. J. Scholte. 1970. A simple chemically defined medium for the production of phase I Bordetella pertussis. Journal of general microbiology 63:211-220. 106. Taneja, N. K., T. Ganguly, L. O. Bakaletz, K. J. Nelson, P. Dubey, L. B. Poole, and R. Deora. 2013. D-alanine modification of a protease-susceptible outer membrane component by the Bordetella pertussis dra locus promotes resistance to antimicrobial peptides and polymorphonuclear leukocyte-mediated killing. Journal of bacteriology 195:5102-5111. 107. Therisod, H., V. Labas, and M. Caroff. 2001. Direct microextraction and analysis of rough-type lipopolysaccharides by combined thin-layer chromatography and MALDI mass spectrometry. Analytical chemistry 73:3804-3807. 108. Thompson, J. D., D. G. Higgins, and T. J. Gibson. 1994. CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic acids research 22:4673-4680. 109. Tirsoaga, A., A. El Hamidi, M. B. Perry, M. Caroff, and A. Novikov. 2007. A rapid, small-scale procedure for the structural characterization of lipid A applied to Citrobacter and Bordetella strains: discovery of a new structural element. Journal of lipid research 48:2419-2427. 110. Tirsoaga, A., A. Novikov, M. Adib-Conquy, C. Werts, C. Fitting, J. M. Cavaillon, and M. Caroff. 2007. Simple method for repurification of endotoxins for biological use. Applied and environmental microbiology 73:1803-1808. 111. Tran, A. X., M. S. Trent, and C. Whitfield. 2008. The LptA protein of Escherichia coli is a periplasmic lipid A-binding protein involved in the lipopolysaccharide export pathway. The Journal of biological chemistry 283:20342-20349. 112. Trent, M. S., W. Pabich, C. R. Raetz, and S. I. Miller. 2001. A PhoP/PhoQ-induced Lipase (PagL) that catalyzes 3-O-deacylation of lipid A precursors in membranes of Salmonella typhimurium. The Journal of biological chemistry 276:9083-9092. 113. Trent, M. S., A. A. Ribeiro, S. Lin, R. J. Cotter, and C. R. Raetz. 2001. An inner membrane enzyme in Salmonella and Escherichia coli that transfers 4-amino-4-deoxy-L-arabinose to lipid A: induction on polymyxin-resistant mutants and role of a novel lipid-linked donor. The Journal of biological chemistry 276:43122-43131. 114. Trent, M. S., C. M. Stead, A. X. Tran, and J. V. Hankins. 2006. Diversity of endotoxin and its impact on pathogenesis. Journal of endotoxin research 12:205-223. 115. Trollfors, B., T. Lagergard, J. Taranger, E. Bergfors, R. Schneerson, and J. B. Robbins. 2001. Serum immunoglobulin G antibody responses to Bordetella pertussis lipooligosaccharide and B. parapertussis lipopolysaccharide in children with pertussis and parapertussis. Clinical and diagnostic laboratory immunology 8:1015-1017. 116. Tsai, C. M., and C. E. Frasch. 1982. A sensitive silver stain for detecting lipopolysaccharides in polyacrylamide gels. Analytical biochemistry 119:115-119. 160  117. Vandamme, P., J. Hommez, M. Vancanneyt, M. Monsieurs, B. Hoste, B. Cookson, C. H. W. Vonkonig, K. Kersters, and P. J. Blackall. 1995. Bordetella Hinzii Sp-Nov, Isolated from Poultry and Humans. Int J Syst Bacteriol 45:37-45. 118. von Wintzingerode, F., A. Schattke, R. A. Siddiqui, U. Rosick, U. B. Gobel, and R. Gross. 2001. Bordetella petrii sp nov., isolated from an anaerobic bioreactor, and amended description of the genus Bordetella. Int J Syst Evol Micr 51:1257-1265. 119. Wang, X., M. J. Karbarz, S. C. McGrath, R. J. Cotter, and C. R. Raetz. 2004. MsbA transporter-dependent lipid A 1-dephosphorylation on the periplasmic surface of the inner membrane: topography of francisella novicida LpxE expressed in Escherichia coli. The Journal of biological chemistry 279:49470-49478. 120. Wang, X., S. C. McGrath, R. J. Cotter, and C. R. Raetz. 2006. Expression cloning and periplasmic orientation of the Francisella novicida lipid A 4'-phosphatase LpxF. The Journal of biological chemistry 281:9321-9330. 121. Waters, L. S., and G. Storz. 2009. Regulatory RNAs in bacteria. Cell 136:615-628. 122. Weyant, R. S., D. G. Hollis, R. E. Weaver, M. F. Amin, A. G. Steigerwalt, S. P. O'Connor, A. M. Whitney, M. I. Daneshvar, C. W. Moss, and D. J. Brenner. 1995. Bordetella holmesii sp. nov., a new gram-negative species associated with septicemia. Journal of clinical microbiology 33:1-7. 123. Wiegand, I., K. Hilpert, and R. E. Hancock. 2008. Agar and broth dilution methods to determine the minimal inhibitory concentration (MIC) of antimicrobial substances. Nature protocols 3:163-175. 124. Williams, A. H., and C. R. Raetz. 2007. Structural basis for the acyl chain selectivity and mechanism of UDP-N-acetylglucosamine acyltransferase. Proceedings of the National Academy of Sciences of the United States of America 104:13543-13550. 125. Wyckoff, T. J., S. Lin, R. J. Cotter, G. D. Dotson, and C. R. Raetz. 1998. Hydrocarbon rulers in UDP-N-acetylglucosamine acyltransferases. The Journal of biological chemistry 273:32369-32372. 126. Yan, A., Z. Guan, and C. R. Raetz. 2007. An undecaprenyl phosphate-aminoarabinose flippase required for polymyxin resistance in Escherichia coli. The Journal of biological chemistry 282:36077-36089. 127. Zavascki, A. P., L. Z. Goldani, J. Li, and R. L. Nation. 2007. Polymyxin B for the treatment of multidrug-resistant pathogens: a critical review. The Journal of antimicrobial chemotherapy 60:1206-1215. 128. Zerbino, D. R., and E. Birney. 2008. Velvet: algorithms for de novo short read assembly using de Bruijn graphs. Genome research 18:821-829.    161  Appendices Appendix A  Mass spectra   Structural analysis of lipid A with negative-ion MALDI mass spectra of the following B. pertussis strains: A) BP338LgmAKO, B) BP338LgmBKO, C) BP338LgmCKO, D) BP338LmgDKO, E)BP338LgmAKO + pBBR2LgmA (complementation of BP338 lgmA mutant), F) BP338LgmBKO + pPtacLgmAB (complementation of BP338 lgmB mutant, G) BP338LgmCKO + pBBR2LgmC (complementation of BP338 lgmC mutant), H) BP338LgmABCDKO + pPtacLgmABCD (complementation of BP338 full lgm locus mutant in trans with full lgm locus, lgmABCD), I) BP338LgmABCDKO + pPtacLgmABC (complementation of BP338 full lgm locus mutant in trans with lgmABC), J) 18-323 + pMMB67HE (18-323 wild type complemented with an empty vector, a control for the pPtac vectors, as seen in Figure 34). For (A-I), peaks at m/z 1559 represent penta-acyl lipid A that lack GlcN modification, peaks at m/z 1720 represent penta-acyl lipid A with one GlcN modification at either phosphate group, and peaks at m/z 1881 represent penta-acyl lipid A with two GlcN modifications, one at each phosphate group. For (J), peaks at m/z 1520 or 1530 represent penta-acyl lipid A with C10-OH or C12-OH at the C3’ position, respectively.  162    163   164        165  Appendix B  Sequence alignments ClustalW2 (Thompson, Higgins, et al. 1994) alignments of Bordetella LgmABCDE that are the basis for the trees in Figure 29. A) LgmA, B) LgmB, C) LgmC, D) LgmD, E) LgmE, F) LgmD and LgmE. Bpe, B. pertussis Tohama I; Bbr, B. bronchiseptica RB50; Bpa, B. parapertussis 12822; Bav, B. avium 197N; Bhi, B. hinzii ATCC 51730; and Btr, B. trematum CCUG 13902. A) ArnC from E. coli K-12. B) ArnT from E. coli K-12. C) Ftn, Ftn_0544 from F. novicida U112. D) LgmEBav, B. avium LgmE. E) GtrA from E. coli K-12. F) prefix ‘D-’ indicates LgmD, prefix ‘E-’ indicates LgmE. ‘*’: Identical amino acid residues; ‘:’: conserved amino acid residues; ‘.’: semi-conserved amino acid residues.  A) LgmA Bav             --MQAEFRSRALPAPLPAPLPTSEAAFEVAAETR-SFYVEAQISCIVPCLNEADNLCVLL 57 Bhi             --MQVEFRSQAFPASLPP----SEAPLEMAGEVR-SFYVEAQVSCIVPCLNEADNLRVLL 53 Btr             --MRYEFRSSSLPAALAG----TPASADIEAETRQAFFVPALVSCVVPCLNEADNLRLLL 54 Bbr             MFMYTEFRSQLLAGAGTS---AAGQSARMAVLAG-DGATGVQVSCVVPGLNEAANLRVLV 56 Bpa             MFMYTEFRSQLLAGAGTS---AAGQSARMAVLAG-DGATGVQVSCVVPGLNEAANLRVLV 56 Bpe             VFMYTEFRSQLLAGAGTS---AAGQSARMAVLAG-DGATGVQVSCVVPGLNEAANLRVLV 56 ArnC            -------------------------MFEIHPVKK--------VSVVIPVYNEQESLPELI 27                                             :             :* ::*  **  .*  *:  Bav             PALRNRLESLCSAWEIIVIDDGSTDNTAELMANWSGLDG--FRYIQLARNFGKEAAISAG 115 Bhi             PALRSRLDSMCSAWEIIVIDDGSSDATPELMAEWTALDG--FRYIQLARNFGKEAAISAG 111 Btr             PALRNRLEAMCDRWEIIVVDDGSTDDTADLMAQWTLVDG--FRYVQLARNFGKEAALSAG 112 Bbr             PALRACLEQWCASWEIIVVDDGSTDDTAELMAQWSAVEG--IRYVQLSRNFGKEAALTAG 114 Bpa             PALRACLEQWCASWEIIVVDDGSTDDTAELMAQWSAVEG--IRYVQLSRNFGKEAALTAG 114 Bpe             PALRACLEQWCASWEIIVVDDGSTDDTAELMAQWSAVEG--IRYVQLSRNFGKEAALTAG 114 ArnC            RRTTTACESLGKEYEILLIDDGSSDNSAHMLVEASQAENSHIVSILLNRNYGQHSAIMAG 87                        :     :**:::****:* :..::.: :  :.  :  : * **:*:.:*: **  Bav             LEAADGDVVICLDADMQHPPALIEEMLRRWQAGSEMVYAVRRNRDDEGYFKRLGSDWFYR 175 Bhi             LEAADGDVVICLDADMQHPPALIEEMLRRWRAGAEMVYALRRDRNDESRFKQLGSHWFYR 171 Btr             LEAADGEAVICMDADLQHPPALIEEMLRRWRAGAEMVYAVRRDRQDEGRFKRFGTACFYR 172 Bbr             LEAADGDAVICLDADMQHPPELIGDMLAAWRNGAEMVYAVRRQRDDEPWFKRVGARAFYR 174 Bpa             LEAADGDAVICLDADMQHPPELIGDMLAAWRNGADMVYAVRRQRDDEPWFKRVGARAFYR 174 Bpe             LEAADGDAVICLDADMQHPPELIGDMLAAWRNGADMVYAVRRQRDDEPWFKRVGARAFYR 174 ArnC            FSHVTGDLIITLDADLQNPPEEIPRLVAKADEGYDVVGTVRQNRQD-SWFRKTASKMINR 146                 :. . *: :* :***:*:**  *  ::     * ::* ::*::*:*   *:: .:  : *   166  Bav             LLSGSR-VDVPAGAGDFRLMDRRVVNALVALPERTRFMKGLYAWVGFKCEPLPYTPDSRM 234 Bhi             LLSGSR-VDVPAGAGDFRLMDRRVVEALVALPERTRFMKGLYAWVGFKSEALPYTPDSRM 230 Btr             LMSGAR-VQVPPGAGDFRLLDRRVVQALVSLPERTRFMKGLYAWVGFKSEALTYVPQARV 231 Bbr             LLSTARGVEVPPHAGDFRLMDRRVVEALVALPERTRFMKGLYAWVGFKSQAVPYTPQARR 234 Bpa             LLSTARGVEVPPHAGDFRLMDRRVVEALVALPERTRFMKGLYAWVGFKSQAVPYTPQARR 234 Bpe             LLSTARGVEVPPHAGDFRLMDRRVVEALVALPERTRFMKGLYAWVGFKSQAVPYTPQARR 234 ArnC            LIQRTTGKAMGDYGCMLRAYRRHIVDAMLHCHERSTFIPILANIFARRAIEIPVHHAERE 206                 *:. :    :   .  :*   *::*:*::   **: *:  *   .. :.  :.     *   Bav             HGQSHFRPMKLIGLAMDGLTAFTTWPLRVVSMLGLVFALMSMVYGAYLVGAYFLDG--NP 292 Bhi             HGRSHFRPMRLFGLALDGLTAFTTWPLRAVSLLGVLFALLALVYGAYLVGAYLLVG--NA 288 Btr             HGESHFRPWRLIRLGLDGLTAFTTWPLRMVSLLGLALAVLAMIYGGYLVAAYFAQG--NA 289 Bbr             HGASHFSAWKLFRLACDGLTAFTTWPLRLVSLIGVLFALLSLSYGGYLVADYLISG--NA 292 Bpa             HGASHFSAWKLFRLACDGLTAFTTWPLRLVSLIGVLFALLSLSYGGYLVADYLISG--NA 292 Bpe             HGASHFSAWKLFRLACDGLTAFTTWPLRLVSLIGVLFALLSLSYGGYLVADYLISG--NA 292 ArnC            FGESKYSFMRLINLMYDLVTCLTTTPLRMLSLLGSIIAIGGFSIAVLLVILRLTFGPQWA 266                 .* *::   :*: *  * :*.:** *** :*::*  :*: .:  .  **   :  *   .  Bav             VSGWTTIVTALLFFAGINLLSLGVVGEYVARIFDEVKGRPLYITRQRRGRARAARKANKK 352 Bhi             VSGWTTIMTALLFFAGVNLISLGVVGEYVARIFDEVKGRPLYIARQRRGRARAARKARAQ 348 Btr             VSGWTTIVTALLFFAGINLISLGVVGEYVARVFDEVKGRPLYIARQRRGRARGARQGEAR 349 Bbr             VSGWTTIVTALLFFAGINLISLGVVGEYVARIFDEVKGRPLFIARQRRGRAKRAAKARSQ 352 Bpa             VSGWTTIVTALLFFAGINLISLGVVGEYVARIFDEVKGRPLFIARQRRGRAKRAAKARSQ 352 Bpe             VSGWTTIVTALLFFAGINLISLGVVGEYVARIFDEVKGRPLFIARQRRGRAKRAAKARSQ 352 ArnC            AEGVFMLFAVLFTFIGAQFIGMGLLGEYIGRIYTDVRARPRYFVQQVIRPSSKENE---- 322                 ..*   :.:.*: * * :::.:*::***:.*:: :*:.** ::.:*    :    :    B) LgmB Bbr             MTLATRSMQPHAVSPGQARSWPLPAAGWLLLA-VGVWLAFLSWMRPLALPDEGRYAGVAW 59 Bpe             MTLATRSMQPHAVSPGQARSWPLPAAGWLLLA-VGVWLAFLSWMRPLALPDEGRYAGVAW 59 Bpa             MTLATRSMQPHAVSPGQARSWPLPAAGWLLLA-VGVWLAFLSWMRPLALPDEGRYAGVAW 59 Bav             -------MNTAVLSAR-SATVRWPA--WLVLAGVAVWLAFLAGIRPLTLPDEGRYGGVAW 50 Bhi             -------MSSVALPVSGVARHRIPA--WLVLAAVALWLMGLAWSRPLTLPDEGRYAGVAW 51 Btr             -------MKAVALSRAPADAPRLPA--WVVLAGIAVWLACLAWARPLTLPDEGRYAGVAW 51 ArnT            -------MKSVRYLIG-------------LFAFIACYYLLPISTRLLWQPDETRYAEISR 40                        *..                   ::* :. :       * *  *** **. ::   Bbr             DMLRNGSFAVPLIDGMPYFHKPPLYYWLAELSFRLFGVNEWAARLPSALAAWASAVALYL 119 Bpe             DMLRNGSFAVPLIDGMPYFHKPPLYYWLAELSFRLFGVNEWAARLPSALAAWASAVALYL 119 Bpa             DMLRNGSFAVPLIDGMPYFHKPPLYYWLAELSFRLFGVNEWAVRLPSALAAWASAVALYL 119 Bav             EMLRSHSYLVPLMDGMPYFHKPPLYYWLAQASFAVFGLSEWSARLPSLLIAWMSIAGVYA 110 Bhi             EMLRSHSYLVPLMDGMPYFHKPPLYYWLAQLSFSVFGLSEWAARLPSLLIAWASIAGVYG 111 Btr             EMLRSDSPMVPLMNGMPYFHKPPLYYWLAQMSFAVFGLNEWAARLPSLLIAWASVAALYA 111 ArnT            EMLASGDWIVPHLLGLRYFEKPIAGYWINSIGQWLFGANNFGVRAGVIFATLLTAALVTW 100                 :** . .  ** : *: **.**   **: . .  :** .::..*    : :  : . :    Bbr             FVRR-HRDAASATLCVLVLATLPLFFGGAQYANMDMLVAGMITLCVLAGADTALRVRGGQ 178 Bpe             FVRR-HRDAASATLCVLVLATLPLFFGGAQYANMDMLVAGMITLCVLAGADTALRVRGGQ 178 Bpa             FVRR-HRDAASATLCVLVLATLPLFFGGAQYANMDMLVAGMITLCVLAGADTALRVRGGQ 178 Bav             FSRR-YRGEAFALCAVLVLSTMPFFYGGAQFANMDMSVAGLITLCVLAGADTIMRVSQGQ 169 Bhi             FARR-YRGERFALCAVLVLSSMPFFYGGAQFANMDMPVAGMITLCVLAGADTIMRVAAGL 170 Btr             FARR-YRGERFALTAAAVLSTMPFFYGGAQFANTDMSVAGLIALCVLAGVHTALCAAAGQ 170 ArnT            FTLRLWRDKRLALLATVIYLSLFIVYAIGTYAVLDP----FIAFWLVAGMCSFWLAMQAQ 156                 *  *  *.   *  .. :  :: :.:. . :*  *     :*:: ::**  :   .  .    167  Bbr             AWR---AMALATGVCAALAMLAKGLIGLVLPGAILLAWLAWRRDWRGLRALLWPPAILAF 235 Bpe             AWR---AMALATGVCAALAMLAKGLIGLVLPGAILLAWLAWRRDWRGLRALLWPPAILAF 235 Bpa             AWR---AMALATGVCAALAMLAKGLIGLVLPGAILLAWLAWRRDWRGLRALLWPPAILAF 235 Bav             PWR---YMSLATALAAALAVLAKGLIGVVLPAAILFFWLLMRRDWRGFKALIWPPAILLF 226 Bhi             PWR---RMSLATALLAALAVLAKGLIGIVLPGAILFFWLLMRRDFHGFKALVWPPAIALF 227 Btr             PWR---RWALTTAAAAGLAVLAKGLIGLVLPGAIVLGWLLLRRDWRGIKALLWPPAIGMF 227 ArnT            TWKGKSAGFLLLGITCGMGVMTKGFLALAVPVLSVLPWVATQKRWKDLFIYGWL-AVISC 215                 .*:      *  .  ..:.:::**::.:.:*   :: *:  :: ::.:    *  *:     Bbr             AVVAVPWFWLMQVRYPGFFQYFFVHQHFERFAQTGFNNVQPFWFYLPVIAGLALPWSLWA 295 Bpe             AVVAVPWFWLMQVRYPGFFQYFFVHQHFERFAQTGFNNVQPFWFYLPVIAGLALPWSLWA 295 Bpa             AVVAVPWFWLMQVRYPGFFQYFFVHQHFERFAQTGFNNVQPFWFYLPVIAGLALPWSLWA 295 Bav             LLVAVPWFVEMQLRYPSFFHYFFVYQHFERFALSGFNNVQPFWFYPPVLAGLALPWSLWL 286 Bhi             LLVALPWFVDMQLRYPGFFHYFFVYQHFERFALSGFNNVQPFWFYPPVLAGLTLPWSLWM 287 Btr             LLVAVPWFAYLQWRYPGFFHYFFIYQQFQRFTLTGFNNVQPFWFYPPVLIGLTLPWSLWL 287 ArnT            VLTVLPWGLAIAQREPNFWHYFFWVEHIQRFALDDAQHRAPFWYYVPVIIAGSLPWLGLL 275                  :..:**   :  * *.*::***  ::::**:  . ::  ***:* **: . :***      Bbr             GGLLRKQFWAADADPDGLRRLALVWLAVIVAFFSMPQSKLVGYIMPVLPPLAFLLAEVVM 355 Bpe             GGLLRKQFWAADADPDGLRRLALVWLAVIVAFFSMPQSKLVGYIMPVLPPLAFLLAEVVM 355 Bpa             GGLLRKQFWAADADPDGLRRLALVWLAVIVAFFSMPQSKLVGYIMPVLPPLAFLLAEVVM 355 Bav             GGALRRGFWAAE-DTDGLRRLMLIWLLVILVFFSLPSSKLIGYILPAVPALAFLVAELVM 345 Bhi             GGTLRRGFWGAE-DVDGLRRLMLIWLVVVVGFFSLPSSKLIGYILPAVPALAFLIAELVL 346 Btr             GGALRRSFWRHE-DQDGLRGLMLLWLAVIVGFFSIPSSKLIGYVLPALPPLAFLVADLVL 346 ArnT            PGALYTGWKNRK---HSATVYLLSWTIMPLLFFSVAKGKLPTYILSCFASLAMLMAHYAL 332                  * *   :   .   ..     * *  : : ***:...**  *::. ...**:*:*. .:  Bbr             GALRDPAVARATRRMARVSALVAVAICVTAVFVASFNARGSSRELALSLRGELRPDDTLV 415 Bpe             GALRDPAVARATRRMARVSALVAVAICVTAVFVASFNARGSSRELALSLRGELRPDDTLV 415 Bpa             GALRDPAVARATRRMARVSALVAVAICVTAVFVASFNARGSSRELALSLRGELRPDDTLV 415 Bav             GAWN-----QGMRRRALLSLGLGAVLCLLGIIIATLNPRGGSGPLGEQVRAEAQTGDTMV 400 Bhi             PAWE-----RGQRRRAQVCLGVAMALCVTGILVATFNPRGGSGPLGEQVRGEAGPYDTMV 401 Btr             PAWE-----QGRRARVWVSAGVAAALCVTGIAVATLKPRGGNGPLARQVITLMQPGDTTV 401 ArnT            LAAKNN--PLALRINGWINIAFGVTGIIATFVVSPWGPMNTPVWQTFESYKVFCAWSIFS 390                  * .      . *    :   .. .  :  . ::.  . .       .      . .     Bbr             ALHTYPFDLQLYAHAARP--------MWVVDDWSNPEIPKRDNWRRELYDAVQFEPALGE 467 Bpe             ALHTYPFDLQLYAHAARP--------MWVVDDWSNPEIPKRDNWRRELYDAVQFEPALGE 467 Bpa             ALHTYPFDLQLYAHAARP--------MWVVDDWSNPEIPKRDNWRRELYDAVQFEPALGE 467 Bav             ALHHFPFDLGIYTASTEP--------IWVVDDWSNPEIPTRDNWRKELYDAAIFEPEVGR 452 Bhi             ALHHFPFDLGIYTASTEP--------LWVVDDWSNPEIPTRDNWRKELYDAAIFEPEVGK 453 Btr             ALHQFPFDLGVYGNLREP--------VWVVDDWRNPEIPTRDNWRKELYDAAQFEPEVGQ 453 ArnT            LWAFFGWYTLTNVEKTWPFAALCPLGLALLVGFSIPDRVMEGKHPQFFVEMTQESLQPSR 450                     : :          *        : :: .:  *:   ..:  : : : .  .   ..  Bbr             RLLVSADTFQQRLCQAPEGSRYWVWGTAADEE---AYAPLRG------QAARFADARRS- 517 Bpe             RLLVSADTFQQRLCQAPEGSRYWVWGTAADEE---AYAPLRG------QAARFADARRS- 517 Bpa             RLLVSADTFQQRLCQAPEGSRYWVWGTAADEE---AYAPLRG------QAARFADARRS- 517 Bav             RVLVSNEVFNARLCAAPTGSRYWVWGQPSDND---AYVSIKG------EAPYFSDGRRL- 502 Bhi             RVLVSNAVFNERLCAAPTGSRYWVWGQTSDGD---AYAAIRG------EAPRFVDGRRQ- 503 Btr             GVLVSNEDFNARLCRADTGARFWIWGQPSDQD---AYPALRG------ESALVGDSRRR- 503 ArnT            YILTDSVGVAAGLAWSLQRDDIIMYRQTGELKYGLNYPDAKGRFVSGDEFANWLNQHRQE 510                  :*..   .   *. :       ::  ..: .    *   :*      : .   : :*     168  Bbr             -----LWLVLVDEAFKGRVCDGTPTGG------------- 539 Bpe             -----LWLVLVDEAFKGRVCDGTPTGG------------- 539 Bpa             -----LWLVLVDEAFKGRVCDGTPTGG------------- 539 Bav             -----VWRVVVNDALRERVCAGKPIGGLPQK--------- 528 Bhi             -----VWRIDVNDAVRQRVCGGKPIGGSPQK--------- 529 Btr             -----VWRIEVNDAVRQRVCGQTPTGGSPRKSAPPEQAE- 537 ArnT            GIITLVLSVDRDEDINSLAIPPADAIDRQERLVLIQYRPK 550                      :  :  :: ..  .       .                C) LgmC  Bav             ------------------------------------------------------------ Bhi             MTKSQSVVTTEGHHEGYQPAGHVRPGWRQAGEFDAQHGDQPVRCGGGAADKDEPGNLTKQ 60 Btr             MPECQQIVTAERHHEGDQPADHIGPCGRQAGQLYAEHDHQPVGGGRSGTDADEPGNLAKH 60 Bbr             -------MTSERYDEGQQPAHHVGPQRGQAGPFHAGDRDGPMGGGRGASDGNEPGNLSQQ 53 Bpa             -------MTSERYDEGQQPAHHVGPQRGQAGPFHAGDRDGPMGGGRGASDGNEPGNLSQQ 53 Bpe             -------MTSERYDEGQQPAHHVGPQRGQAGPFHAGDRDGPMGGGRGASDGNEPGNLSQQ 53 Ftn             ------------------------------------------------------------                                                                               Bav             ---MRMTVMGTTMKASAYSRCIAICGDDFGMDASIDRAIFQLLDAGRMSAVSCMSTGASF 57 Bhi             GGHAEVKVMDTTMKASAHGRRIAVCGDDFGMDASIDHAIFQLVDAGRLSAVSCMSTGASF 120 Btr             GGQAGKSVRDAGMDTLRQARRVAICGDDFGMDAGIDHAILRLIQARRLSAASCMSSAPGF 120 Bbr             AAHVEEVSPLRNQAGDVRCRRIAVCGDDFGMNEAIDGALIELAGAGRLSAVSCMPLAPAF 113 Bpa             AAHVEEVSPLRNQAGDVRCRRIAVCGDDFGMNEAIDGALIELAGAGRLSAVSCMPLAPAF 113 Bpe             AAHVEEVSPLRNQAGDVRCRRIAVCGDDFGMNEAIDGALIELAGAGRLSAVSCMPLAPAF 113 Ftn             -----------------MVKKIIICADDFGMSDNINSAIINLLEKKIINATSCMPNMPAF 43                                    : : :*.*****.  *: *::.*     :.*.***.  ..*  Bav             AAHAQDLKTRA---VDIGLHLNFTEPLSPADG-------GMLPLRALLLRAYTGRLNANW 107 Bhi             ARHAADLKRRA---ADTGLHLNLTQALSPADA-------GMLPLKTLLLRAYAGRLDRTQ 170 Btr             ATRGAELLNSG---ADIGLHLNLTERLGLADT-------PRPALRRLLWRAYSRQLDLQW 170 Bbr             AADAPALARLD---VDLGVHVDFTEAFAGAAP-------AAPGLAALLWRAYAGQLDPDW 163 Bpa             AADAPALARLD---VDLGVHVDFTEAFAGAAP-------AAPGLAALLWRAYAGQLDPDW 163 Bpe             AADAPALARLD---VDLGVHVDFTEAFAGAAP-------AAPGLAALLWRAYAGQLDPDW 163 Ftn             KLGIAQLKKIYNDFSHVGIHLNLTEGNAFTNPKSITRNGKFLSLSKLLVKSKLRAINYDD 103                       *        . *:*:::*:  . :             *  ** ::    ::     Bav             LRQEIARQLDAFEDRIGHAPHYVDGHQHVHQLPGVRQALLAELRQR-YAGQRPWLRLTPA 166 Bhi             VRQEIDRQLDAFEDQMGQAPHYVDGHQHVHQLPGVRGPLLEALRQR-YPRQRPWLRLTPA 229 Btr             VNQEICRQLDRFEDVLGLAPDYIDGHQHVHQLPGVRQLLLAELQRR-YAGRRPWLRLTSV 229 Bbr             IDARLASQFDAFERAFGRAPDYVDGHQHVHQLPGILPRLRALLKRR-YAGQRIWLRHTAP 222 Bpa             IDARLASQFDAFERAFGRAPDYVDGHQHVHQLPGILPRLRALLKRR-YAGQRIWLRHTAP 222 Bpe             IDARLASQFDAFERAFGRAPDYVDGHQHVHQLPGILPRLRALLKRR-YAGQRIWLRHTAP 222 Ftn             VYNELKAQINNFIEDWGALPDFIDGHQHVHHFPIIRKAVINLYKDFNMYTKQTYIRSTYK 163                 :  .:  *:: *    *  *.::*******::* :   :    :      :: ::* *    Bav             GTLQGMPWMPTLKAQLIAALGGHALAAQVRREAWPSNGRFFGVYGFQGGERVYAAFLHHW 226 Bhi             GAMQGLPLAAVFKAHAIAGLGGHALAAQARQDGWPGNRRFFGAYGFGGGRRAYASLLHHW 289 Btr             AALDGLPWAAHIKAHAIASLGGHALAARARALSWPSNRGFLGVYGFEGGRRGYARMLHHW 289 Bbr             GLQFGLPLAEAAKARLIGALGAGALARAAGQEGWQTNRRMLGVYGFTGGPRRYAGLLHHW 282 Bpa             GLQFGLPLAEAAKARLIGALGAGALARAAGQEGWQTNRRMLGVYGFTGGPRRYAGLLHHW 282 Bpe             GLQFGLPLAEAAKARLIGALGAGALARAAGQEGWQTNRRMLGVYGFTGGPRRYAGLLHHW 282 Ftn             MDKS------DFKSLIIYRSGAKKFYNMLIKNNIKHNSSFAGVYSLESDNQDFRKVILEA 217                             *:  *   *.  :           *  : *.*.: .. : :  .: .    169  Bav             LSNALDGDLIMCHPALPG--PVEHAEQRVAELAVLSSAELGEWLVANGLLVQRLSLRRPA 284 Bhi             LFNALDGDLIMCHPALPG--PIEHAAQRVAEFEVLSSPELGEWLVANGLSVARLSQMVPA 347 Btr             LVHARDGDLLMCHPALAG--EIEHAAQRKAEYEVLADPDLGGWLAVNGLSVQRMSVILAS 347 Bbr             LMNARDGDLLMCHPGWPQVHGAAHASQRAAEYEVLAHPELGTWLARNGLRIVRLSQVRGR 342 Bpa             LMNARDGDLLMCHPGWPQVHGAAHASQRAAEYEVLAHPELGTWLARNGLRIVRLSQVRGR 342 Bpe             LMNARDGDLLMCHPGWPQVHGAAHASQRAAEYEVLAHPELGTWLARNGLRIVSLSQVRGR 342 Ftn             YTEIKDGGIIMCHPAADIDIKDPISQSRIKEFAYFNSKQALQDQKDHNIVL--------- 268                   .  **.::****.         : .*  *   :   :       :.: :           Bav             LASETRQGSPLMR------TLQPLSGP- 305 Bhi             GSREASIGASLKRQGWGADANPPLISR- 374 Btr             AAWAAP-GAVGAR-------WRPATGR- 366 Bbr             QASQESGKVRNVPHSG---SFRRLASRL 367 Bpa             QASQESGKVRNVPHSG---SFRRLASRL 367 Bpe             QASQESGKVRNVPHFG---SFRRLASRL 367 Ftn             ----------------------------    D) LgmD Bbr             MSSSSKQTPKTIKERFLHAFFFEIIAIGLSAPVAAWAMDQPLFDMGVLTAVIAWIALLWN 60 Bpa             MSSSSKQTPKTIKERFLHAFFFEIIAIGLSAPVAAWAMDQPLFDMGVLTAVIAWIALLWN 60 Bpe             MSSSSRQTPKTIKERFLHAFFFEIIAIGLSAPVAAWAMDQPLFDMGVLTAVIAWIALLWN 60 Bhi             MT----QAKKTLKERFFHAFLFEILAIGLCAPVAAWAMGKSLFEMGVLTAVIAWIALLWN 56 Btr             -MEVTLITLSPLKRRIVYVSLFELFAILLSTLILMALSDGSAQNSLPVAVIVSATAVLWN 59 LgmEBav         -------------MPALFRQIALFILVGCAAAATHWLAAVLCVEFGGMAPAWANVVG-WL 46                                 ..  :  :: :  .:            :   ::   :  .  *   Bbr             MVYNAGFERLERR-FGVVRTMPVRVAHAVGFELGLVLIIVPLAAWWLAISFWEAFMLDIG 119 Bpa             MVYNAGFERLERR-FGVVRTMPVRVAHAVGFELGLVLIIVLLAAWWLAISFWEAFMLDIG 119 Bpe             MVYNAGFERLERR-FGVVRTMPVRVAHAVGFELGLVLIIVPLAAWWLAISFWEAFMLDIG 119 Bhi             MVYNAGFDRLENH-MGWTRTLRLRVVHALGFETGLILIVIPLAAWWLDISLWQAFVLDIA 115 Btr             YLYNLGFEAWERRNHVMQRTLRVRCIHAVGFEGGLLLFCLPVYMLWYGVGPLVALGMELT 119 LgmEBav         LAFAVSFSGHYRLTFRHLALSWIVAARRFFLVSAAGFAVNELAFVWLLHTTRLPYELLLG 106                   :  .*.   .          :   : . :  .  :    :   *       .  : :   Bbr             LLMFYLPYAFFYNLAYDKLR---ARWWG-RIEPAGA 151 Bpa             LLMFYLPYAFFYNLAYDKLR---ARWWG-RIEPAGA 151 Bpe             LLMFYLPYAFFYNLAYDKLR---ARWWG-RIEPAGA 151 Bhi             LVLFYLPYAFFYNLGYDKARGPVLRWLARRARYAGA 151 Btr             LMVFFLFYTFVFTLVFDKIFTLPQHYAKLTPAEQG- 154 LgmEBav         LILIVLACLTFVASRLWAFR-----HKPARATRH-- 135                 *::: *    .      E) LgmE Bav             MPALFRQIALFILVGCAAAATHWLAAVLCVEFGGMAPAWANVVGWLLAFAVSFSGHYRLT 60 Bhi             MPALFRQIAWFVFVGCAAAATHWLVAVLCVEFAGLAPAWANVAGWLVAFVVSFSGHYRLT 60 Bpa             MRGLLRQIAWFIAVGCAAAATHWAVAVACVEWAGLPPLGANVVGWLLAFVVSFTGHFRLT 60 Bpe             MRGLLRQIAWFIAVGCAAAATHWAVAVACVEWAGLPPLGANVVGWLLAFVVSFTGHFRLT 60 Bbr             MRGLLRQIAWFIAVGCAAAATHWAVAVACVEWAGLPPLGANVVGWLLAFVVSFTGHFRLT 60 GtrA            ---MLKLFAKYTSIGVLNTLIHWVVFGVCIYVAHTNQALANFAGFVVAVSFSFFANAKFT 57                    ::: :* :  :*   :  ** .   *:  .      **..*:::*. .** .: ::*   170  Bav             FRHLALSWIVAARRFFLVSAAGFAVNELAFVWLLHTTRLPYELLLGLILIVLACLTFVAS 120 Bhi             FRHLTLSWIIAARRFFLVSAAGFALNEAAYVWLLHATRLPYDLLLALILVGLAFLTFVAS 120 Bpa             FRHLAASWTIAARRFFLVSALGFAINELSYAWLLHATSLPYDVLLALVLIGLAFLTFVAS 120 Bpe             FRHLAASWTIAARRFFLVSALGFAINELSYAWLLHATSLPYDVLLALVLIGLAFLTFVAS 120 Bbr             FRHLAASWTIAARRFFLVSALGFAINELSYAWLLHATSLPYDVLLALVLIGLAFLTFVAS 120 GtrA            FKASTT----TMRYMLYVGFMGTLS--ATVGWAADRCALPPMITLVTFSAISLVCGFVYS 111                 *:  :     : * :: *.  *      :  *  .   **  : *  .        ** *  Bav             RLWAFRHKPARATRH 135 Bhi             RLWAFRHKPAAGPHR 135 Bpa             RLWAFRHRHAP---- 131 Bpe             RLWAFRHRHAP---- 131 Bbr             RLWAFRHRHAP---- 131 GtrA            KFIVFRDAK------ 120                 :: .**.            F) LgmD and LgmE E-Bpa           -------------MRGLLRQIAWFIAVGCAAAATHWAVAVACVEWAGLPPLGANVVGWLL 47 E-Bpe           -------------MRGLLRQIAWFIAVGCAAAATHWAVAVACVEWAGLPPLGANVVGWLL 47 E-Bbr           -------------MRGLLRQIAWFIAVGCAAAATHWAVAVACVEWAGLPPLGANVVGWLL 47 E-Bav           -------------MPALFRQIALFILVGCAAAATHWLAAVLCVEFGGMAPAWANVVGWLL 47 E-Bhi           -------------MPALFRQIAWFVFVGCAAAATHWLVAVLCVEFAGLAPAWANVAGWLV 47 D-Bbr           MSSSSKQTPKTIKERFLHAFFFEIIAIGLSAPVAAWAMDQPLFDMGVLTAVIAWIA---L 57 D-Bpa           MSSSSKQTPKTIKERFLHAFFFEIIAIGLSAPVAAWAMDQPLFDMGVLTAVIAWIA---L 57 D-Bpe           MSSSSRQTPKTIKERFLHAFFFEIIAIGLSAPVAAWAMDQPLFDMGVLTAVIAWIA---L 57 D-Bhi           MT----QAKKTLKERFFHAFLFEILAIGLCAPVAAWAMGKSLFEMGVLTAVIAWIA---L 53 D-Btr           -MEVTLITLSPLKRRIVYVSLFELFAILLSTLILMALSDGSAQNSLPVAVIVSATA---V 56                                 .   :  :. :  .:            :   :.   :  .   :  E-Bpa           AFVVSFTGHFRLTFRH-LAASWTIAARRFFLVSALG---FAINELSYAWLLHATSLPYDV 103 E-Bpe           AFVVSFTGHFRLTFRH-LAASWTIAARRFFLVSALG---FAINELSYAWLLHATSLPYDV 103 E-Bbr           AFVVSFTGHFRLTFRH-LAASWTIAARRFFLVSALG---FAINELSYAWLLHATSLPYDV 103 E-Bav           AFAVSFSGHYRLTFRH-LALSWIVAARRFFLVSAAG---FAVNELAFVWLLHTTRLPYEL 103 E-Bhi           AFVVSFSGHYRLTFRH-LTLSWIIAARRFFLVSAAG---FALNEAAYVWLLHATRLPYDL 103 D-Bbr           LWNMVYNAGFERLERR-FGVVRTMPVRVAHAVGFELGLVLIIVPLAAWWLAISFWEAFML 116 D-Bpa           LWNMVYNAGFERLERR-FGVVRTMPVRVAHAVGFELGLVLIIVLLAAWWLAISFWEAFML 116 D-Bpe           LWNMVYNAGFERLERR-FGVVRTMPVRVAHAVGFELGLVLIIVPLAAWWLAISFWEAFML 116 D-Bhi           LWNMVYNAGFDRLENH-MGWTRTLRLRVVHALGFETGLILIVIPLAAWWLDISLWQAFVL 112 D-Btr           LWNYLYNLGFEAWERRNHVMQRTLRVRCIHAVGFEGGLLLFCLPVYMLWYGVGPLVALGM 116                  :   :.  :    .:       :  *  . :.      :        *       .  :  E-Bpa           LLALVLIGLAFLTFVAS----------RLWAFRHRHAP---- 131 E-Bpe           LLALVLIGLAFLTFVAS----------RLWAFRHRHAP---- 131 E-Bbr           LLALVLIGLAFLTFVAS----------RLWAFRHRHAP---- 131 E-Bav           LLGLILIVLACLTFVAS----------RLWAFRHKPARATRH 135 E-Bhi           LLALILVGLAFLTFVAS----------RLWAFRHKPAAGPHR 135 D-Bbr           DIGLLMFYLPYAFFYNLAYDKLR---ARWWG-RIEPAGA--- 151 D-Bpa           DIGLLMFYLPYAFFYNLAYDKLR---ARWWG-RIEPAGA--- 151 D-Bpe           DIGLLMFYLPYAFFYNLAYDKLR---ARWWG-RIEPAGA--- 151 D-Bhi           DIALVLFYLPYAFFYNLGYDKARGPVLRWLARRARYAGA--- 151 D-Btr           ELTLMVFFLFYTFVFTLVFDKIFTLPQHYAKLTPAEQG---- 154                  : *::. *    .             :                  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

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"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            data-media="{[{embed.selectedMedia}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0165993/manifest

Comment

Related Items