Molecular Classification of a Vector and Roles of Bacterial Surface Proteins in Pathogenesis

Neorickettsia spp.: Molecular Classification of a Vector and Roles of Bacterial Surface Proteins in Pathogenesis

Dissertation

Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the Graduate School of The Ohio State University

Kathryn Elizabeth Gibson, D.V.M.

Graduate Program in Veterinary Biosciences

The Ohio State University

2011

Dissertation Committee

Yasuko Rikihisa, Advisor

Prosper Boyaka

Steven Krakowka

Michael Oglesbee

Kathryn Gibson

2011

Abstract

Neorickettsia spp. are Gram-negative, obligate intracellular bacteria of the family

Anaplasmataceae. Given their prevalence throughout the world, their propensity to cause human disease and deadly animal diseases, and the continuing discovery of new

Neorickettsia spp., they are significant pathogens requiring better comprehension. The overall objective of this dissertation is to determine the host-bacterium relationships of

Neorickettsia. Chapter 1 details background on Neorickettsia spp., with emphasis on

Neorickettsia risticii and Neorickettsia sennetsu. The objective of Chapter 2 was to demonstrate the lineage of all N. risticii-infected trematode life stages. The study established the molecular identification of the N. risticii adult trematode host and its immature life stages, demonstrating as hypothesized that all life stages harboring N. risticii belong to the same clade. The objective of Chapter 3 was to determine the major surface proteins of N. sennetsu involved in host-pathogen interaction and to determine the roles of the major surface proteins. Four proteins: the 51-kDa antigen (P51),

Neorickettsia surface proteins 2 (Nsp2) and 3 (Nsp3), and heat-shock protein 60 (GroEL), were found to have the highest surface expression. It was hypothesized that the two major

β-barrel proteins, P51 and Nsp3 function as porins. The outer membrane fraction of N. sennetsu, as well as native P51 and Nsp3 were incorporated into proteoliposomes and tested by porin-swelling assays, and it was confirmed that P51 is a large porin. The ii

objective of Chapter 4 was to determine levels of variation within predicted surface- exposed proteins of N. risticii, with the hypothesis being that geographic and temporal variation would occur among strains of N. risticii. Variation in P51 demonstrated geographic separation of N. risticii strains. Nsp2, Nsp3, and strain-specific antigen 3

(Ssa3) demonstrated temporal variation. Variety within the β-barrel proteins P51, Nsp2, and Nsp3 occurred mainly within regions predicted as external loops. Ssa3 variation mainly occurred in an N-terminal localized repeat region and consisted of changes in the number of 52-aa repeats. The objective of Chapter 5 was to determine causes of cytokine and chemokine induction in N. sennetsu infection, thus potential reasons for disease symptoms. The studies in this chapter first validated the mouse model of Sennetsu neorickettsiosis in immunocompetent BALB/c mice and then demonstrated cytokine and chemokine production by quantitative reverse-transcriptase polymerase chain reaction within spleens of infected mice during disease. Cytokine and chemokine production in the whole mouse was replicated in vitro within splenocyte and bone marrow-derived macrophage (BMDM) cultures, with significant induction of IL-1β, CXCL2, and IL-12A

(p35) mRNAs occurring within whole bacteria and the Sarkosyl-purified bacterial outer membrane fraction of N. sennetsu. It was hypothesized that P51 is a major cause of cytokine induction, is recognized by Toll-like receptor 2, and that cytokine/chemokine induction is MyD88 dependent. Preliminary studies showed that cytokine/chemokine levels were reduced in MyD88-knockout mouse BMDMs and TLR2-knockout mouse splenocytes incubated with whole bacteria and in TLR2-knockout mouse BMDMs incubated with whole bacteria and bacterial outer membrane fraction. In conclusion,

iii

these studies demonstrated important host-pathogen relationships during Neorickettsial infection and disease useful for further understanding of these bacteria.

Dedication

Dedicated to my family, both human and animal

Acknowledgement

I would foremost like to thank my advisor, Dr. Yasuko Rikihisa for her guidance, assistance, wisdom, patience, and resources. I would like to thank the members of my dissertation committee, Dr. Proper Boyaka, Dr. Steven Krakowka, and Dr. Michael

Oglesbee for their valuable input and advice. I would like to thank all the members past and present of the Rikihisa laboratory, including Dr. Jason Mott, Dr. Mingqun Lin, Dr.

Yumi Kumagai, Dr. Junji Matsuo, Dr. Rahman Akhlakur, Dr. Hua Niu, Dr. Tzung-Huei

Lai, and Dr. Yan Ge, for their technical advice and assistance. I would like to thank future veterinarians Susanne Moesta and Gabrielle Pastenkos for their hard work and dedication during their summer research programs in the Rikihisa Laboratory. I would like to thank the members of the Mass Spectrometry and Proteomics Facility, including

Dr. Karen Green-Church for their assistance. I appreciate Dr. Hiroshi Nikaido at the

University of California, Berkeley for his valuable advice. I also would like to acknowledge The Ohio State University College of Veterinary Medicine and Department of Veterinary Biosciences. These works were funded by grant R01AI30010 and T32

RR0070703 from the National Institutes of Health.

Vita

June 1996 ……………………………… Granville High School

May 2000 ……………………………… B.A. Biology and Music, Minor in

Chemistry, Salem College

2001-2003 ……………………………… Graduate research associate, Department of

Veterinary Biosciences, The Ohio State

University

August 2003 ……………………………… M.S. Veterinary Biosciences, The Ohio

State University

2003-2007 ……………………………… D.V.M., The Ohio State University

2007-2008 ……………………………… Graduate research associate, Department of

Veterinary Biosciences, The Ohio State

University

2008-2010 ……………………………… Post-doctoral fellow, Department of

Veterinary Biosciences, The Ohio State

University

2010-Present ……………………………… Graduate research associate, Department of

Veterinary Biosciences, The Ohio State

University vii

Publications

Gibson, K., Y. Rikihisa, C. Zhang, and C. Martin (2005). “Neorickettsia risticii is vertically transmitted in the trematode Acanthatrium oregonense and horizontally transmitted to bats.” Environ Microbiol 7: 203-212.

Gibson, K. and Y. Rikihisa (2008). “Molecular link of different stages of the trematode host of Neorickettsia risticii to Acanthatrium oregonense.” Environ Microbiol 10: 2064- 2073.

Lin, M., C. Zhang, K. Gibson, and Y. Rikihisa (2009). “Analysis of complete genome sequence of Neorickettsia risticii: Causative agent of Potomac horse fever.” Nucleic Acids Res 37(18): 6076-6091.

Gibson, K., Y. Kumagai, and Y. Rikihisa (2010). “Proteomic analysis of Neorickettsia sennetsu surface-exposed proteins and porin activity of the major surface protein P51.” J Bacteriol 192(22): 5898-5905.

Fields of Study

Major Field: Veterinary Biosciences

viii

Table of Contents

Title Page ..……………………………………………………………………….………. i

Abstract ……………………………….………………………………………….……… ii

Dedication ………………………….……………………………………………………. v

Acknowledgement .………………..………….………………………………………… vi

Vita …………………………………………..………………………………………… vii

List of Tables …………………………………………………………………….….…... x

List of Figures ……………………………………………………………………….…. xii

List of Symbols and Abbreviations …………………………………………………….. xv

Chapter 1: Introduction ………………………………………………………………….. 1

Chapter 2: Molecular Identification of the Trematode Host of Neorickettsia risticii by

18S rRNA ……………………………………………………………………………… 14

Chapter 3: Identification of Neorickettsia sennetsu Surface-Exposed Proteins and

Characterization of P51 as a Porin ……………………………………….…………….. 41

Chapter 4: Identification of Geographical and Temporal Variation within N. risticii

Surface-Exposed Proteins and Their Antigenicity within Naturally-Infected Horses …. 80

Chapter 5: Cytokine Induction by N. sennetsu Outer Membrane Proteins ……….…... 110

References …………………………………………………………………………….. 157

Appendix A …………………………………………………………………………… 179 ix

List of Tables

Table 1. Trematode 18S rRNA sequence sample information ………………………… 24

Table 2. 18S rRNA Region 3 (767 bp) percent sequence identity for all trematode samples …………………………………………………………………………………. 25

Table 3. 18S rRNA Region 1 percent sequence identity for N. risticii-positive trematodes ……………………………………………………………………………… 26

Table 4. 18S rRNA Region 2 percent sequence identity for N. risticii-positive trematodes ……………………………………………………………………………… 26

Table 5. 18S rRNA Region 4 percent sequence identity for N. risticii-positive trematodes ……………………………………………………………………………… 26

Table 6. NCBI BLAST search results for trematode 18S rRNA sequences …………… 33

Table 7. Streptavidin affinity-purified and proteomics-identified proteins for N. sennetsu

Miyayama ……………………………………………………………………………… 54

Table 8. Amino acid differences among predicted P51 transmembrane domains …….. 65

Table 9. Primers utilized for Neorickettsia PCR amplification ………………………... 83

Table 10. Sequences amplified for Neorickettsia ……………………………………… 85

Table 11. Proteomics-identified proteins for two N. risticii strains ……………………. 89

Table 12. PHF-positive sera from naturally-infected horses and negative sera ……….. 92

Table 13. Primer pairs for RT-PCR, qPCR, and qRT-PCR …………………………... 122

Table 14. Expression analysis of N. sennetsu surface proteins in four and eight-week-old

BALB/c mouse spleens ……………………………………………………………….. 132

Table 15. Means for qRT-PCR results for cytokine induction studies in spleens in the

BALB/c mouse model of disease ……………………………………………………... 141

Table 16. Means for qRT-PCR results for mouse splenocyte cytokine induction studies ………………………………………………………………………………… 180

Table 17. Means for qRT-PCR results for mouse BMDM cytokine induction studies ………………………………………………………………………………… 183

List of Figures

Figure 1. Life cycle of the digenetic trematode hosts of Neorickettsia spp. …………….. 2

Figure 2. Life cycle of Acanthatrium oregonense, the trematode host of N. risticii ……. 6

Figure 3. Known range of Neorickettsia in the world …………………………………. 13

Figure 4. Primer locations on the 18S rRNA gene of the N. risticii-positive trematode ……………………………………………………………………………….. 19

Figure 5. Dendrogram of Region 3 (767 bp total, including all insertions) of the 18S rRNA gene sequences obtained from all N. risticii-positive trematode samples, P. longiforme, and L. linstowi ….…………………………………………………….…… 27

Figure 6. Dendrogram of the 1,588 bp (including all insertions) aligned 18S rRNA gene sequences (a 1,512 bp region for both PAFlukeA and PASnailA) obtained from the N. risticii-positive trematode samples FlukeA and SnailA and ten other species of digenetic trematodes ……………………………………………………………………………… 29

Figure 8. Streptavidin affinity-purified N. sennetsu Miyayama surface proteins ……… 53

Figure 9. Western blotting analysis of N. sennetsu Miyayama surface proteins ………. 58

Figure 10. Recombinant proteins and antisera to Nsp2 and Nsp3 ………………….….. 59

xii

Figure 11. Surface localization and predicted secondary structure of P51 …………….. 63

Figure 12. Surface localization of Nsp3 ………………………………………….…….. 67

Figure 13. Surface localization of Nsp2 ……………………………………………….. 68

Figure 14. Surface localization of GroEL ……………………………………………… 69

Figure 15. Porin activity of the N. sennetsu outer membrane fraction incorporated into proteoliposomes ………………………………………………………………………... 70

Figure 16. Porin activity of HPLC-separated P51 and Nsp3 fractions incorporated into proteoliposomes ………………………………………………………………………... 72

Figure 17. N. risticii SS-biotinylation gels …………………………………………….. 88

Figure 18. Western blotting against rP51, rGroEL, rNsp2, and rNsp3 using PHF-positive sera from naturally-infected horses …………………………………………………….. 91

Figure 19. P51 amino acid sequence variations among Neorickettsia sequences ……... 94

Figure 20. Dendrograms of P51 amino acid sequence variations among Neorickettsia sequences ………………………………………………………………………………. 95

Figure 21. Nsp2 amino acid sequence variations ………………………………………. 97

Figure 22. Nsp3 amino acid sequence variations ………………………………………. 99

Figure 23. Ssa3 amino acid sequence variations ……………………………………... 102

Figure 24. Ssa1 amino acid sequence variations ……………………………………... 104

Figure 25. ICR mouse body and spleen weights ……………………………………… 124

Figure 26. Body weights, splenic weights, and clinical signs for four-week-old CF-1 and

BALB/c mice …………………………………………………………………………. 126

xiii

Figure 27. Body weights, splenic weights, and clinical signs for eight-week-old BALB/c

4 6 mice inoculated with 1 × 10 or 1 × 10 N. sennetsu infected or control P388D1 cells …………………………………………………………………………………… 128

Figure 28. Pathologic changes in N. sennetsu-infected BALB/c mice ……………….. 130

Figure 29. RT-PCR results for four-week-old N. sennetsu infected and control BALB/c mice …………………………………………………………………………………… 133

Figure 30. Quantitative RT-PCR results for N. sennetsu-infected P388D1 cells with low infectivity (5%, Low inf.) and high infectivity (>80%, High inf.) cells and four-week-old

BALB/c mouse livers and spleens (n = 3) ………………………………………….… 134

Figure 31. Reactivity of N. sennetsu-infected mouse sera against rNsp2 and rNsp3 … 135

Figure 32. Recombinant Nsp mouse immunization experiment N. sennetsu challenge ……………………………………………………………………………… 137

Figure 33. In vitro neutralization with α-rP51, α-rNsp2, and α-rNsp3 ……………..… 139

Figure 34. In vivo cytokine mRNA expression in BALB/c mouse spleens during Sennetsu neorickettsiosis ………………………………………………………………………... 140

Figure 35. Cytokine production within splenocytes incubated fresh and previously-frozen

N. sennetsu and N. sennetsu outer membrane fraction ……………………………….. 143

Figure 36. MyD88-dependent cytokine production by N. sennetsu in BMDMs ……... 145

Figure 37. TLR2-dependent cytokine production by N. sennetsu in vitro and link to outer membrane fraction ……………………………………………………………………. 147

Figure 38. Recombinant protein incubation of splenocytes and BMDMs …………..... 149

xiv

List of Symbols and Abbreviations

°C: Degrees Celsius

=: Equals

>: Greater than

<: Less than

−: Minus

×: Multiplied by

#: Number of

%: Percent

+: Plus

±: Plus or minus

2: Squared

16S: Subunit of the small (30S) subunit of the 70S prokaryotic ribosome

18S: Subunit of the small (40S) subunit of the 80S eukaryotic ribosome

α: Alpha or anti

β: Beta

γ: Gamma

Δ: Change in

µg: Microgram

µl: Microliter

µm: Micrometer aa: Amino acid

ANOVA: Analysis of variance

AtpA: ATP synthase F1, alpha subunit

BALB/c: “Bagg albino” inbred albino mouse strain

BCA: Bicinchoninic acid

BL21(DE3): Escherichia coli competent cells used for recombinant protein

expression

BMDM: Bone marrow-derived macrophage bp: Base pair

BLAST: Basic Local Alignment Search Tool blastn: Nucleotide BLAST algorithm blastp: Protein BLAST algorithm

C: Control

C57/BL6J: C57 black 6 inbred mouse strain cd or CD: Conserved domain cDNA: Complementary DNA

CF-1: Carworth farms CF-1 colony albino outbred mouse strain cm: Centimeter

CO2: Carbon dioxide

C-terminal region: Carboxy-terminal region of a protein

Ctl: Control xvi

Cy3: Cyanine 3 fluorescent dye

DGG: Digestive gland-gonad complex of the snail dH2O: Autoclaved distilled, deiozined, and filtered water

DH5α: Escherichia coli competent cells used for amplification of

recombinant protein-containing plasmids

Diffs: Differences

DNA: Deoxyribonucleic acid

DnaK: Heat-shock protein 70

DOB: Date of birth

E: External loop of a β-barrel protein

ECL: Enhanced chemiluminescence

EcoRI: Restriction endonuclease which recognizes DNA site

5′ GAATTC 3′

EDTA: Ethylenediaminetetraacetic acid

E value: Expect value for Basic Local Alignment Search Tool result

FabF: 3-oxoacyl-(acyl-carrier-protein) synthase II

FBS: Fetal bovine serum

FtsH: ATP-dependent metalloprotease

FusA: Translation elongation factor G g: Gravity (for centrifugations) or grams

Ga: Gauge

Gap: Glyceraldehyde-3-phosphate dehydrogenase, type I gapdh: Gene encoding glyceraldehyde 3-phosphate dehydrogenase xvii

GlnA: Glutamine synthetase, type I

GltA: Citrate synthase

GlyA: Serine hydroxymethyltransferase gp: Glycoprotein

GroEL: Heat-shock protein 60 groESL: Genes encoding GroEL and GroES (heat-shock protein 10) h: Hour

Hib: Haemophilus influenzae type b

His: Histidine

HMM: Hidden Markov model

HPLC: High-pressure liquid chromatography

HRP: Horseradish peroxidase

Hsp: Heat-shock protein

HtpG: Heat-shock protein 90

IACUC: Institutional Animal Care and Use Committee

I: Infected

ICR: Institute for Cancer Research outbred albino mouse strain of Swiss

mouse origin, also called CD-1(ICR)

ID: Identification

IFA: Immunofluorescence assay

IFN: Interferon

IgG: Immunoglobulin G

IL: Interleukin xviii

Inf: Infectivity iNOS: Nitric oxide synthase

IOE: Ixodes ovatus Ehrlichia

IP: Intraperitoneal

IPTG: Isopropyl β-D-1-thiogalactopyranoside kDa: Kilodalton kg: Kilogram

KHCO3: Potassium bicarbonate

KO: Knock out

L-: Amino acid spatial configuration form L

L-929: Mouse fibroblast-like cell line from a C3H/An strain mouse

LamB: Lambda phage receptor protein

LPS: Lipopolysaccharide

M: Molar

MgCl2: Magnesium chloride ml: Milliliter min: Minute mM: Millimolar

Mol mass: Molecular mass of a protein

MPP: Mitochondrial processing peptidase mRNA: Messenger RNA murE: Gene encoding UDP-N-acetylmuramoylalanyl-D-glutamate-2,6-

diamiaopimelate ligase xix

MyD88: Myeloid differentiation primary response gene (88) n: Number of samples in a given group

N: Nucleus or negative

NaCl: Sodium chloride

Nano-LC/MS/MS: Capillary-liquid chromatography-nanospray tandem mass

spectrometry

NaPi: Sodium phosphate buffer

NBF: Neutral-buffered formalin

NCBI: National Center for Biotechnology Information

Neg: Negative ng: Nanogram

NH4Cl: Ammonium chloride no.: Number

NotI: Restriction endonuclease which recognizes DNA site

5′ GCGGCCGC 3′

NP-40: Nonyl phenoxypolyethoxylethanol

NR: Neorickettsia risticii

NS: Neorickettsia sennetsu

Nsp: Neorickettsia surface protein nt: Nucleotide

N-terminal region: Amino-terminal region of a protein

OD: Optical density

OGC: Octyl-β-glucoside xx

OM: Outer membrane

Omp or OMP: Outer membrane protein

ORF: Open-reading frame

P: Probability value

P388D1: Mouse macrophage-like cell line from dilute brown agouti mice

P51: 51-kDa antigen

PAGE: Polyacrylamide gel electrophoresis

PAMP: Pathogen-associated molecular pattern

PBS: Phosphate-buffered saline, pH 7.4

PBST: Phosphate-buffered saline containing Tween-20

PCR: Polymerase chain reaction pET 33b(+): Plasmid vector used for recombinant protein transformation

PGS: Phosphate-buffered saline containing gelatin and saponin

Pfam: Protein family database

PGE: Prostaglandin

PHF: Potomac horse fever p.i.: Post inoculation pI: Theoretical isoelectric point

Pnp: Polyribonucleotide nucleotidyltransferase

PRR: Pattern-recognition receptor qRT-PCR: Quantitative reverse-transcription polymerase chain reaction

RNA: Ribonucleic acid rNsp2: Recombinant Nsp2 xxi

rNsp3: Recombinant Nsp3

RPMI: Roswell Park Memorial Institute culture medium

RpsB: Ribosomal protein S2 rRNA: Ribosomal RNA rP51: Recombinant P51 rpoB: Gene encoding the β subunit of RNA polymerase from bacteria

RT-PCR: Reverse-transcription polymerase chain reaction s: Second

SalI: Restriction endonuclease which recognizes DNA site

5′ GTCGAC 3′

SCID: Severe combined immunodeficiency

ScrY: Plasmid-encoded sucrose channel

SdhA: Succinate dehydrogenase, flavoprotein subunit

SDS: Sodium dodecyl sulfate

Seqs: Sequences

SF agent: Stellantchasmus falcatus agent sp.: Species (singular)

SP: Signal peptide

SPD: Salmon-poisoning disease spp.: Species (plural)

Ssa: Strain-specific antigen

Std dev: Standard deviation

SucA: 2-oxoglutarate dehydrogenase xxii

SucC: Succinyl-CoA synthetase, β-subunit

T: Type strain

TBS: Tris-buffered saline, pH 7.4

TBST: Tris-buffered saline containing Tween-20

TLR: Toll-like receptor

TM: Transmembrane domain of a β-barrel protein

TolC: Outer membrane efflux protein

TpiA: Triosephosphate isomerase

Trp: Tandem-repeat protein

Tsf: Translation elongation factor Ts

VirB9-1: Type IV secretion system protein subunit

VLPT: Variable-length PCR target protein

Wsp: Wolbachia surface protein

WT: Wild type wt/vol: Weight per volume

Zn: Zinc

xxiii

Chapter 1:

Introduction

Neorickettsia spp. are Gram negative obligate intracellular α-proteobacteria of the family

Anaplasmataceae in the order Rickettsiales (Dumler, Barbet et al. 2001). They are related to members of the genera Ehrlichia, Anaplasma, and Wolbachia. Neorickettsia is unique among these other genera in that it is comprised of the only bacteria known to infect trematodes, maintained by transovarial passage through generations of these parasitic worms (Cordy and Gorham 1950; Bennington and Pratt 1960; Gibson, Rikihisa et al.

2005). Neorickettsia is horizontally transmitted to mammals, causing human and animal diseases (Figure 1) (Cordy and Gorham 1950; Philip, Hadlow et al. 1953; Fukuda, Kitao et al. 1954; Misao and Kobayashi 1954; Holland, Weiss et al. 1985).

Figure 1. Life cycle of the digenetic trematode hosts of Neorickettsia spp. Vertical transmission occurs within the trematode (green arrows). Horizontal transmission occurs between the adult trematode and mammalian host (purple arrow).

Neorickettsia risticii:

Neorickettsia (formerly Ehrlichia) risticii, an obligate intracellular bacterium that infects monocytes, macrophages, intestinal epithelial cells, and mast cells is the causative agent of Potomac horse fever (PHF) (Rikihisa and Perry 1984; Rikihisa, Perry et al. 1984;

Holland, Ristic et al. 1985; Chae, Kim et al. 2002). PHF was first identified in 1979 along the Potomac River in Maryland. It was characterized as a potentially fatal disease

(30% mortality) of horses with clinical signs including pyrexia, anorexia, and periodic

cases of diarrhea; blood work on infected horses also demonstrated leukopenia (Knowles,

Anderson et al. 1984). Upon analysis of macrophage-like cells within colonic and intestinal specimens of a PHF-infected pony, Rickettsial-like organisms were discovered

(Rikihisa, Perry et al. 1984). These organisms were propagated in human histiocytes and were identified as similar to other bacteria of the genus Ehrlichia (Rikihisa and Perry

1984). The organism was also isolated from a PHF naturally-infected pony in a culture of canine monocytes and then re-isolated from a pony that was experimentally infected with the culture of canine monocytes and had demonstrated clinical signs of PHF (Holland,

Ristic et al. 1985). The experimentally-supported bacterial causative agent of PHF (also called equine monocytic ehrlichiosis) was originally named Ehrlichia risticii (Holland,

Weiss et al. 1985).

It was originally hypothesized that E. risticii was transmitted to the horse by a tick vector, as is demonstrated in other Ehrlichial and Rickettsial organisms (Hechemy, Oteo et al.

2006). Attempts were made to isolate E. risticii from ticks collected from PHF-endemic regions (Carroll and Schmidtmann 1986) and to experimentally infect ticks of the species

Dermacentor variabilis (Schmidtmann, Robl et al. 1986). However, no tick species were demonstrated to harbor and/or transmit E. risticii.

Through an increased study of genetic information, it was soon discovered that E. risticii had strikingly similar genomic, as well as morphologic characteristics to Neorickettsia helminthoeca (Rikihisa, Stills et al. 1991; Pretzman, Ralph et al. 1995). N. helminthoeca is known to be perpetuated through nature in a digenetic trematode host (Philip, Hadlow et al. 1953; Philip, Hadlow et al. 1954). Furthermore, another Ehrlichia-like agent,

known as the SF agent and isolated from the digenetic trematode Stellantchasmus falcatus was determined to be genetically similar to E. risticii (Wen, Rikihisa et al. 1996).

Thus, it was hypothesized that E. risticii may have a similar natural pattern and infect digenetic trematodes instead of ticks or other arthropods. The bacterium was eventually reclassified to the genus Neorickettsia, and the infection of digenetic trematodes by N. risticii was confirmed (Barlough, Reubel et al. 1998; Kanter, Mott et al. 2000; Dumler,

Barbet et al. 2001; Mott, Muramatsu et al. 2002).

In the case of the digenetic trematode harboring N. risticii, the trematode begins in a river or stream as an egg. The egg is devoured by a snail and hatches into a miracidium, or the miracidium first hatches in the water then penetrates the snail. This snail (Elimia livenscens, Elimia virginica, or Juga spp.) serves as the trematode’s first intermediate host. The trematode develops into a sporocyst (Barlough, Reubel et al. 1998; Reubel,

Barlough et al. 1998; Kanter, Mott et al. 2000; Mott, Muramatsu et al. 2002), which can asexually reproduce more sporocysts or produce numerous motile cercariae. The cercarial stage then leaves the snail and bores into the larval stage of a mayfly, caddisfly, or stonefly. In this second intermediate host, the trematode develops into an encysted metacercaria (Burns 1961; Chae, Pusterla et al. 2000; Mott, Muramatsu et al. 2002).

When the adult insect is then ingested by a bat (Eptesicus fuscus or Myotis yumanensis), the trematode breaks out of its cyst, attaches to the lumen of the small intestine, and matures into an adult (Pusterla, Johnson et al. 2003; Gibson, Rikihisa et al. 2005). Eggs from gravid trematode adults are shed in the feces. If the feces lands in water, then the life cycle begins anew (Figure 2) (Macy 1939; Knight and Pratt 1955; Schell 1985;

Pusterla, Johnson et al. 2003; Gibson, Rikihisa et al. 2005). Horses, being outside the natural life cycle of the digenetic trematode are considered accidental hosts and are infected by ingestion of insects containing N. risticii-infected metacercaria (Mott,

Muramatsu et al. 2002; Rikihisa 2004). N. risticii DNA is found in most developmental stages of the digenetic trematode, including egg, cercarial and sporocyst, metacercarial, and adult (Barlough, Reubel et al. 1998; Reubel, Barlough et al. 1998; Kanter, Mott et al.

2000; Madigan, Pusterla et al. 2000; Pusterla, Johnson et al. 2000; Mott, Muramatsu et al. 2002; Park, Kim et al. 2003; Pusterla, Johnson et al. 2003; Gibson, Rikihisa et al.

2005). The miracidium stage is difficult to visualize, but is assumed to be infected, given the evidence of N. risticii infection of larvae within trematode eggs. Furthermore, vertical transmission of N. risticii from adult to egg has been recognized (Gibson, Rikihisa et al.

2005).

Figure 2. Life cycle of Acanthatrium oregonense, the trematode host of N. risticii.

Photomicrographs are modified from (Gibson, Rikihisa et al. 2005; Gibson and Rikihisa

2008).

PHF is a rapidly progressing and potentially life-threatening disease of horses observed mainly in North America, South America, and Europe (Holland, Ristic et al. 1985; Ristic,

Holland et al. 1986; Vidor, Bissuel et al. 1988; van der Kolk, Bernadina et al. 1991;

Dutra, Schuch et al. 2001). The clinical signs of PHF include depression, anorexia, fever, dehydration, laminitis, abortion, and watery diarrhea. Although complications, such as laminitis can be fatal or necessitate humane euthanasia, clinical signs of PHF vary from

severe to mild or nonexistent (Knowles, Anderson et al. 1984; Rikihisa 1998; Rikihisa

2004). PHF can thus be highly detrimental to the equine industry. Depending on the virulence of the disease and financial or emotive situations, the economic burden of PHF could range from the cost of veterinary care to the loss of a highly valuable animal and his/her potential progeny. If a pleasure horse contracts PHF and is hospitalized, the total bill might extend over five thousand dollars, especially since the horse will likely be kept in isolation during their hospitalization. If a racing horse dies from PHF, the loss could be millions of dollars (not including future progeny, training, and other expenses) and devastating, both emotionally and financially to the owner, trainer and other involved parties. PHF is mainly diagnosed by immunofluorescence assay (IFA) or by polymerase chain reaction (PCR) testing based on the 16S ribosomal RNA (rRNA) gene (Holland,

Ristic et al. 1985; Wen, Rikihisa et al. 1995; Mott, Rikihisa et al. 1997). The only effective treatment is the administration of tetracycline, a broad-spectrum bacteriostatic antibiotic, in the early stages of the disease; although there are bactrin vaccines based on

N. risticii Illinois, the efficacy of these vaccines is highly questionable (Knowles,

Anderson et al. 1984; Rikihisa and Jiang 1988; Palmer 1989; Rikihisa and Jiang 1989;

Chaichanasiriwithaya, Rikihisa et al. 1994; Dutta, Vemulapalli et al. 1998).

Previous studies have explored the intracellular life of N. risticii and bacterial invasion of host cells. Electron microscopy revealed that the bacterium is not free in the cytoplasm; rather it resides in a membrane-bound compartment (Rikihisa, Perry et al. 1984). This compartment containing N. risticii does not fuse with lysosomes, suggesting that the bacterium is able to arrest the maturation of the compartment, creating a niche in which

to survive and reproduce (Wells and Rikihisa 1988). It has further been demonstrated that, even though the bacterium parasitizes professional phagocytes, N. risticii invades its host cells by receptor-mediated endocytosis, rather than by phagocytosis (Messick and

Rikihisa 1993).

Recently, whole genome sequencing of N. risticii IllinoisT was performed (Lin, Zhang et al. 2009). At 879,977 nucleotides (nt) in size, with 86.9% of the genome encoding potential RNAs and proteins, and with 839 open-reading frames (ORFs), the N. risticii genome is highly reduced. There are 344 ORFs with hypothetical or unknown function, no capacity to produce lipopolysaccharide (LPS) or peptidoglycan, an incomplete glycolysis pathway, the inablilty to synthesize most amino acids, and a faulty DNA repair system. N. risticii is thus dependent on its hosts for survival, possibly having a symbiotic relationship with its trematode host (Lin, Zhang et al. 2009).

Neorickettsia sennetsu: Neorickettsia (formerly Rickettsia, Ehrlichia) sennetsu is a human pathogen with an affinity for monocytes and macrophages (Hoilien, Ristic et al. 1982) and causes the disease Sennetsu neorickettsiosis (also known as Sennetsu ehrlichiosis, Sennetsu rickettsiosis, human Sennetsu rickettsiosis, Sennetsu disease, infectious or rickettsial mononucleosis, Hyuga fever, and glandular fever) (Fukuda 1958; Tachibana and Kobayashi 1975; Hastriter, Kelly et al. 1987; Brouqui and Raoult 1990; Hotopp, Lin et al. 2006). N. sennetsu is the first human pathogen to have been isolated and cultured in the family Anaplasmataceae (Fukuda, Kitao et al. 1954; Misao and Kobayashi 1954). Symptoms of Sennetsu neorickettsiosis include lymphadenopathy, pyrexia, inappetence, lethargy, sleeplessness, and overall malaise (Misao and Kobayashi 1954; Tachibana, 8

Shishime et al. 1987; Rikihisa 1991). Treatment involves tetracycline therapy and is normally highly successful in resolving the symptoms (Fukuda 1958). Geographically, N. sennetsu infections mainly occur in western Japan, although antibodies to N. sennetsu have also been found in humans in Malaysia, and one strain of N. sennetsu has been isolated from Malaysia (Fukuda 1958; Holland, Ristic et al. 1985; Rikihisa, Zhang et al. 2004). Recently, N. sennetsu has been discovered in Laos (Newton, Rolain et al. 2009).

N. sennetsu was first identified as the causative agent of Sennetsu neorickettsiosis in 1953, and the resulting type strain (Miyayama) was propagated from a case on the island of Kyushu, Japan (Misao and Kobayashi 1954). It was noted that the disease occurred mostly between the months of August and November (Misao and Kobayashi 1954). Although originally not isolated from fish, a similar bacterium, known as the SF agent was identified from encysted trematodes (S. falcatus) within grey mullet fish. Furthermore, it was indicated that most patients suffering from Sennetsu neorickettsiosis had eaten raw grey mullet fish and that volunteers fed raw grey mullet fish developed symptoms similar to those of Sennetsu neorickettsiosis (Fukuda, Kitao et al. 1954).

Given the morphologic similarities of N. sennetsu to Rickettsia spp., the bacterium was originally named Rickettsia sennetsu, and it was suspected that the bacterium may use an arthropod vector, such as a trombiculid mite, to gain entry into a human host.

Experiments were performed on the species Leptotrombidium (Leptotrombidium) fletcheri, although infection was not attained in the mites fed on R. sennetsu-infected mice, nor was R. sennetsu passed from infected mouse to uninfected mouse through a mite vector (Hastriter, Kelly et al. 1987). Later findings determined that R. sennetsu was not serologically related to other Rickettsia spp. but was related to Ehrlichia canis; the inclusion bodies (morulae) formed by R. sennetsu also did not resemble the freeform cytoplasmic organisms characteristic of Rickettsia spp. (Anderson, Hopps et al. 1965; 9

Shirahama 1967; Ristic, Huxsoll et al. 1981; Hoilien, Ristic et al. 1982). Thus, R. sennetsu was reclassified to the Ehrlichia genus and eventually again reclassified to

Neorickettsia, given the striking physical and genetic similarities to N. helminthoeca, as noted with N. risticii (Fukuda and Yamamoto 1981; Pretzman, Ralph et al. 1995; Dumler, Barbet et al. 2001).

Although hypothesized to be a digenetic trematode, the vector of N. sennetsu is still unknown (Hotopp, Lin et al. 2006). Current research on N. sennetsu infections in Laos have demonstrated that Anabas testudineus, or the climbing perch is an intermediate host for the unknown digenetic trematode harboring N. sennetsu (Newton, Rolain et al. 2009).

Currently, P388D1 (a murine macrophage-like cell line derived from the dilute brown agouti mouse) cells (Koren, Handwerger et al. 1975) are the standard for culturing N. sennetsu in the Rikihisa laboratory (Cole, Ristic et al. 1985; Hotopp, Lin et al. 2006). N. sennetsu has also been propagated in vivo within human volunteers, a pigtailed macaque (Macaca mulatta), mice, and even horses; although horses show no clinical signs of disease (Misao and Kobayashi 1954; Kitao, Farrell et al. 1973; Rikihisa, Pretzman et al. 1988).

As done for N. risticii, whole genome sequencing has been performed on N. sennetsu MiyayamaT (Hotopp, Lin et al. 2006), and there is noted to be a strong similarity between

N. sennetsu and N. risticii, with their 16S rRNA sequences being over 99% identical to each other (Lin, Zhang et al. 2009). The determined N. sennetsu genome size is 859,006 nt with 87.5% of the genome coding potential RNAs and proteins. Out of the determined 935 N. sennetsu ORFs, approximately 43% (403 out of 935 ORFs) are considered coding for hypothetical (288 ORFs), conserved hypothetical (51 ORFs) or unknown function (64 ORFs) proteins. As seen with N. risticii, there is no LPS, no peptidoglycan synthesis,

limited amino acid synthesis, an incomplete glycolysis pathway, and a faulty DNA repair system (Hotopp, Lin et al. 2006).

Other Neorickettsia spp.: Other members of the genus Neorickettsia include the abovementioned N. helminthoeca, the causative agent of salmon-poisoning disease (SPD) in wild and domestic canids. SPD is an acute disease with mortality rates in untreated dogs around 90%. Rapidly-emerging clinical signs involve fever, depression, diarrhea and resulting dehydration, pronounced lymphadenopathy, and death. The geographic range of SPD was originally designated from northern coastal regions of California to Washington (Cordy and Gorham 1950; Philip, Hadlow et al. 1954). In the Rikihisa laboratory, N. helminthoeca DNA was isolated from two diagnostic canine blood samples submitted from southern California in 2001 (Rikihisa, Zhang et al. 2004), and multiple cases have been recently diagnosed by PCR in the Rikihisa laboratory from 2007 to 2010 (data not published).

N. helminthoeca was the first infectious agent identified as transmitted by a digenetic trematode. Naturally infecting the digenetic trematode Nanophyetus salmincola, N. helminthoeca is transmitted throughout the trematode’s lifecycle: through rediae and cercariae identified in the freshwater snail Oxytrema silicula (formerly Goniobasis pliciera and the trematode’s first intermediate host), metacercariae within salmonid fishes (second intermediate host), and adult gravid trematodes in canids (definitive host) (Philip 1953; Philip, Hadlow et al. 1953; Knapp and Millemann 1981). Therefore, domestic dogs develop SPD through eating raw or undercooked fish containing metacercariae of N. salmincola infected with N. helminthoeca. Given the potential for high morbidity and mortality, SPD is a definite threat to the domestic canine community. Furthermore, dogs and wild canids such as coyotes, foxes, and wolves, and many other wild mammals such

as bears (although the latter don’t develop SPD [they may develop mild clinical signs]), are natural parts of the digenetic trematode’s lifecycle and shed trematode eggs in their feces, thus perpetuating the spread of N. helminthoeca-infected trematodes (Millemann and Knapp 1970).

Although the geographic range of the bacterium originally seemed to be contained to the northwestern coastal United States, occurrences of SPD appear to be expanding in range. The first indication was in 1984, when SPD was reported in domestic dogs on Vancouver Island, thus extending the geographic range into Canada (Booth, Stogdale et al. 1984). From 2000 to 2005, N. helminthoeca DNA was amplified from dogs displaying clinical signs of SPD in Brazil. These dogs were residents of Brazil throughout their lives, and fragments from multiple genes (including the 16S rRNA gene, rpoB, and groESL) were PCR-amplified and compared to determine the appropriate species (Headley, Vidotto et al. 2004; Headley, Scorpio et al. 2006). These data have demonstrated the expanding range of N. helminthoeca. O. silicula has only been reported on the western coast of the

United States, but it seems likely that other related snail and fish species are also able to be infected by N. salmincola (Simms, Donham et al. 1931; Bennington and Pratt 1960; Gebhardt, Millemann et al. 1966; Baldwin, Millemann et al. 1967; Millemann and Knapp 1970).

Another previously-mentioned bacterium found in the Neorickettsia genus is the SF agent. This bacterium has been isolated from Japan and Oregon and can cause mild clinical signs in the dog, although it has not been thoroughly characterized (Fukuda and Yamamoto 1981; Rikihisa, Zhang et al. 2004). The causative agent of Elokomin fluke fever has also been characterized as a Neorickettsial organism, showing similarity to N. helminthoeca and infecting and producing mild clinical signs in the black bear (Ursus americanus) (Farrell, Leader et al. 1973). There have also been other uncharacterized 12

Neorickettsia-like organisms in Oregon (Pusterla, Johnson et al. 2000) and in Laos (Newton, Rolain et al. 2009), suggesting the presence of multiple unknown species of unknown virulence and host range. In all, Neorickettsia spp. are found in multiple countries throughout North and South America, Western Europe, Southeast Asia, and Australia (Figure 3) (Philip, Hadlow et al. 1953; Fukuda, Kitao et al. 1954; Misao and Kobayashi 1954; Park, Kim et al. 2003; Headley, Vidotto et al. 2004; Rikihisa, Zhang et al. 2004; Rikihisa 2005; Headley, Scorpio et al. 2009; Newton, Rolain et al. 2009).

Figure 3. Known range of Neorickettsia in the world. Countries reporting Neorickettsia are shown in red.

Chapter 2:

Molecular Identification of the Trematode Host of Neorickettsia risticii by 18S rRNA

Introduction:

Trematodes have been historically classified and identified with morphologic keys.

Because the identification of a trematode requires careful study of morphologic characteristics, only the adult stage can be used. This adult must be alive when preserved and specially stained to obtain a definitive genus and species. If the trematode is immature, dead, torn, too gravid, or otherwise damaged, then identification is hindered

(Pritchard and Kruse 1982). Previously, the identification of adult gravid trematodes harboring N. risticii was made based on morphological keys: Acanthatrium oregonense of the family Lecithodendriidae (Gibson, Rikihisa et al. 2005), yet immature stages were unidentifiable to genus and species (Reubel, Barlough et al. 1998; Kanter, Mott et al.

2000).

The objective of Chapter 2 is to demonstrate the lineage of all N. risticii-infected trematode life stages, using the hypothesis that all life stages of the trematode harboring

N. risticii are from A. oregonense. No molecular information had previously been available for Acanthatrium spp. Therefore, to establish the life cycle of the trematodes in the present study, the 18S rRNA gene of several life stages of N. risticii-positive trematode stages obtained from a PHF-endemic region was sequenced. The 18S rRNA 14

gene is considered a molecular clock, since sequence variation closely follows the evolutional tree and is less influenced by environmental selection pressure (Woese 1987;

Olsen and Woese 1993; Littlewood and Olson 2001). There are multiple trematode 18S rRNA whole and partial sequences available in GenBank, including two from members of the family Lecithodendriidae: Prosthodendrium longiforme and Lecithodendrium linstowi (Littlewood, Rohde et al. 1999; Littlewood and Olson 2001; Olson, Cribb et al.

2003). These sequences allowed us to determine the relative genetic relationship among

N. risticii-infected trematodes and among them and other sequenced trematode species.

Experimental Procedures:

N. risticii-positive adult trematodes. DNA was extracted from four adult trematodes and four trematode eggs obtained from the same E. fuscus bat collected in 2002 from

Chambersburg, PA (labeled TW2) (Gibson et al., 2005). DNA samples from two of the adult gravid trematodes that were both N. risticii-positive by nested PCR performed using outer primers ER5-3 (5′ ATT TGA GAG TTT GAT CCT GG 3′) and ER3-2 (5′ GTT

TTA AAT GCA GTT CTT GG 3′) and nested primers E.ris1 (5′ GGA ATC AGG GCT

GCT TGC AGC CT 3′) and E.ris2 (5′ TGT GGG TAC CGT CAT TAT CTT CCC CAC

T 3′) (Kanter, Mott et al. 2000; Gibson, Rikihisa et al. 2005) were used in trematode PCR analyses.

N. risticii-positive sporocysts and cercariae from snails. Collections of sporocysts and cercariae were obtained from digestive gland-gonad complexes (DGG) of E. virginica snails acquired 8/14/2005 and 7/7/2003 from the Susquehanna River in a PHF-endemic

region in Northumberland, PA as previously described (Kanter, Mott et al. 2000; Mott,

Muramatsu et al. 2002). Northumberland is approximately 169 km Northeast of

Chambersburg, PA. In brief, the snail shells were cracked just above the first whorl, and the snail was carefully extracted. The upper regions of the DGG, staying caudal to the stomach, were collected and teased apart with 20-Ga needles in a Petri dish containing phosphate-buffered saline, pH 7.4 (PBS). Cercariae and sporocysts would normally produce grayish-white clouds within the PBS, which were confirmed as the organisms by microscopy. The sporocysts/cercariae were then collected by Pasteur pipette and stored in a 1.5 ml microcentrifuge tube in PBS at −20°C until used. DNA was extracted using the tissue protocol in the DNA Mini Kit (Qiagen, Valencia, CA) with modifications to use all possible pooled cercariae/sporocysts per snail sample and decrease amounts of Buffer AE

(Elution Buffer; Qiagen). DNA was stored at −20ºC, until tested for N. risticii DNA by

PCR using primers amplifying the 16S rRNA gene described above. N. risticii-positive

DNA extracted from sporocysts and cercariae collections from three different snails were used for trematode PCR analyses.

N. risticii-positive metacercariae. DNA extracted from a pool of four caddisflies of similar color and morphology collected 06/17/2004 (obtained at or near a gas station close to the Susquehanna River in Northumberland, PA) that was N. risticii PCR positive was used for trematode PCR analyses. Metacercariae were isolated under light microscopy from ten mayflies of similar color and morphology that were obtained at or near the gas station (see caddisfly description above) on 07/05/2005. In brief, mayflies were teased apart in PBS, using 20-Ga needles. When a metacercaria was identified, it

was isolated as cleanly as possible from surrounding insect tissue, collected by Pasteur pipette, added to a 1.5 microcentrifuge tube, and stored in PBS at −20°C until used. To extract DNA, the tissue protocol in the DNA Mini Kit (Qiagen) was modified including an overnight incubation and increased Proteinase K (Invitrogen Life Technologies,

Carlsbad, CA) concentration for improved tissue digestion, an extra washing step with

AW1 (Wash Buffer I; Qiagen), and decreased Buffer AE (Elution Buffer; Qiagen). DNA that was extracted from a collection of 15 metacercariae and was N. risticii PCR-positive was used for trematode PCR analyses.

Trematode 18S rRNA PCR and sequencing. Primers were selected or modified based on the 18S rRNA sequences of P. longiforme (GenBank Accession number AY222148) and

L. linstowi (AY222147), and one was created de novo using the P. longiforme and L. linstowi sequences (Figure 4). First-round PCR on a GeneAmp PCR System 9700 (Bio-

Rad Laboratories, Hercules, CA) used the universal primer pair 18S-E (5′ CCG AAT

TCG TCG ACA ACC TGG TTG ATC CTG CCA GT 3′; base pair [bp] positions −34 to

0, in relation to the 1,922 bp 18S rRNA fragment) and 18S-F (5′ CCA GCT TGA TCC

TTC TGC AGG TTC ACC TAC 3′; bp 1,952-1,923) (Littlewood, Rohde et al. 1999;

Littlewood and Olson 2001). PCR conditions were as follows: preincubation of 5 min at

94ºC, followed by 40 cycles of 1 min denaturation at 94ºC, 1 min annealing at 50ºC, and

1 min blocking at 72ºC, then 7 min extension at 72ºC, followed by a 4ºC hold for infinity.

Nested and hemi-nested PCR were performed to amplify Regions 1, 3, and 4. Region 1 was amplified with 18S-E and SHRrev (5′ GCA CTC AAA TTT GTT CAA AG 3′; bp

844-825); SHRrev was revised from primer SHR (Park, Kim et al. 2003). Hemi-nested

PCR conditions were the same as the first-round PCR, except step-wise increases in the annealing temperature (50°C, 52°C, and 58°C) and hot starts were employed. Region 3 was amplified with newly-designed primer 715F (5′ CTT CCA GAT GCT CTT AAC 3′; bp 779-796) and 18S-3rev (5′ GCA TCA CAG ACC TGT TAT 3′; bp 1,581-1,564); 18S-

3rev was revised from primer 18S-3 (Littlewood and Olson 2001). Nested PCR conditions were the same as the first-round PCR, except annealing was at 52ºC. Region 4 was amplified with 1270F (5′ ACT TAA AGG AAT TGA CGG 3′; bp 1,211-1,228)

(Littlewood and Olson 2001) and 18S-F. Hemi-nested PCR conditions were the same as first-round PCR, except annealing was performed at 50ºC. SHFrev (5′ GGG TGC ATT

TAT TAG AAC 3′; bp 178-195), revised from primer SHF (Park, Kim et al. 2003) and

SHRrev were used for single step PCR for Region 2. PCR conditions were the same as for Region 4 hemi-nested PCR.

18S-E 18S-F 715F 18S-3rev

SHFrev SHRrev 1270F

18S rRNA gene

1 Region 1 824

797 1,563 Region 3

196 824 1,229 1,922 Region 2 Region 4

1 1,922 18S rRNA gene fragment

Figure 4. Primer locations on the 18S rRNA gene of the N. risticii-positive trematode.

Green arrows indicate forward primers. Red arrows indicate reverse primers. Primers in bold are universal trematode 18S rRNA primers. Primers ending in “rev” were revised from previously-designed primers to better fit the family Lecithodendriidae. The shaded primer was created de novo. Regions 1-4 and the 1,922 bp gene fragment amplified by primers are shown below the 18S rRNA gene depiction. Figure is derived from (Gibson and Rikihisa 2008).

PCR products obtained for Regions 1, 2, 3, and 4 were cloned into pCRII vectors using the TA Cloning Kit Dual Promoter (Invitrogen), according to manufacturer’s instructions.

Plasmids were either sequenced with the ABI BigDye Terminator v 3.0 Cycle

Sequencing Kit (Applied Biosystems, Foster City, CA) and an ABI Prism 310 Genetic

Analyzer (Applied Biosystems) or were submitted for sequencing at the Ohio State

University Plant-Microbe Genomics Facility.

Sequence alignment and analysis. Sequences were aligned using the CLUSTAL W

(slow/accurate) method in the MegAlign program of DNAStar (DNAStar, Madison, WI).

Multiple sequences for an individual DNA sample were combined into a longer sequence

(Table 1). If multiple bases appeared for a given bp position within multiple plasmids for an individual DNA sample, then the most common base for that position was selected. If no base majority was available for a given bp position, then a letter designating both base letters was employed, using the standard International Union of

Biochemistry/International Union of Pure and Applied Chemistry nucleic acid code format. Each separately-sequenced region and total combined sequence for an individual sample was compared to the nucleotide collection database for any organism with the nucleotide megablast algorithm of the Basic Local Alignment Search Tool (BLAST) from the National Center for Biotechnology Information (NCBI) website.

Dendrograms and bootstrapping. Combined sequences for PAFlukeA and PASnailA

(1,922 bp long) were aligned by the CLUSTAL W method of MegAlign (DNAStar) with

18S rRNA sequences from ten other digenetic trematodes using a common region, whose combined sequence size was 1,588 bp (the region was a 1,512 bp fragment for both

PAFlukeA and PASnailA). To confirm the validity of internal nodes, Seqboot of PHYLIP

(v3.66) (Felsenstein 1989) with 1,000 replicates was employed. Bootstrap values were acquired using the programs DNADist, Neighbor, and Consense of PHYLIP. A dendrogram was created and plotted using the DNADist, Neighbor, and Drawgram

programs of PHYLIP. Sections containing a majority of sequences with three or more deletions within the common 1,588 bp region were then removed from the aligned sequences, and node validity, bootstrap values, and a dendrogram were again obtained on the resulting 1,538 bp region. PCR results for all ten Region 3 sequences (767 bp) and the corresponding regions of P. longiforme (763 bp) and L. linstowi (767 bp) were aligned with MegAlign (DNAStar). Node validity, bootstrap values, and a dendrogram were obtained using PHYLIP as described earlier.

Results:

18S rRNA gene sequence of N. risticii-positive adult trematodes. In a previous N. risticii study (Gibson, Rikihisa et al. 2005), numerous adult gravid trematodes of homogeneous morphology were recovered from the intestines of an E. fuscus bat from PA. Samples of this trematode batch were identified as A. oregonense, and adult trematodes, as well as individual eggs from the adult trematodes were N. risticii-positive by PCR. In the present study, DNA samples from two adult gravid trematodes (FlukeA and FlukeB) from this batch, were confirmed by PCR to be N. risticii-positive and used. The 18S rRNA gene was amplified from FlukeA within Regions 1, 3 and 4. Sequences of four plasmids were identical or showed clear majority for each region. The resulting sequences were combined into a 1,922 bp sequence (Accession no. EU019964) representing an almost complete 18S rRNA gene. This sequence was labeled PAFlukeA. The trematode 18S rRNA gene was amplified from FlukeB within Region 3. Sequences (767 bp) from two plasmids were identical, and the sequence was labeled PAFlukeB (Accession no.

EU019965, Table 1). 21

18S rRNA gene sequence of N. risticii-positive sporocysts and cercariae from snails. N. risticii-positive DNA samples from pooled cercariae and sporocysts of three snails collected in PA were used in the present study. The DNA samples were labeled SnailA,

SnailB, and SnailC. Regions 1, 3, and 4 were amplified from SnailA. Sequencing was performed on two plasmids for each region and deposited under six separate sequences

(PASnailA1-6), given the base differences (Accession nos. EU019974-EU019978). The six sequences were combined into a 1,922 bp sequence (PASnailA) representing an almost complete 18S rRNA gene. Regions 1 and 3 were amplified from SnailB.

Sequencing was performed on two plasmids for each region and deposited under four separate sequences (PASnailB1-4, Accession nos. EU019970-EU019973). The four sequences were combined into a 1,563 bp fragment (PASnailB) in this paper. Regions 3 and 4 were amplified from SnailC. Sequences of two plasmids for each Region were identical. The resulting sequences were combined into a 1,126 bp fragment labeled

PASnailC (Accession no. EU019963, Table 1).

18S rRNA gene sequence of N. risticii-positive metacercaria from insects. DNA from four whole caddisflies collected in PA that were pooled and tested N. risticii positive

(Caddis) was used. The 18S rRNA gene was amplified from Caddis within Regions 2 and

3. The sequences from four plasmids for Region 2 were identical (PACaddis-1,

Accession no. EU019969). Sequences of two plasmids for Region 3 were different

(PACaddis-2 and 3, Accession nos. EU019967 and EU019968). The resulting sequences were combined into a 1,563 bp fragment, labeled PACaddis in this paper (Table 1).

Region 4 could not be amplified. Instead, the 702-bp sequences (Accession nos.

EU019961 and EU019962) when analyzed by the NCBI BLAST megalign algorithm were most closely related to the 18S rRNA gene of Oecetis avara (Accession no.

AF286300) with 689-690 bp out of 703 bp identical (98%). O. avara is a member of the order Trichoptera and, thus a caddisfly (Whiting 2002). Of note, the N. risticii-positive caddisflies in a previous study (Mott, Muramatsu et al. 2002) were morphologically identified as Oecetis inconspicua, suggesting that the caddifly host DNA was unintentionally amplified instead of the trematode DNA for Region 4.

To prevent erroneous amplification of insect DNA, metacercariae were isolated from ten mayflies collected in PA. DNA (labeled May) was extracted from 15 pooled N. risticii- positive metacercariae. Trematode DNA was amplified from May within Regions 3 and

4. Two plasmids were sequenced for both regions. The resulting sequences within each region were identical, and Regions 3 and 4 were combined into a 1,126 bp fragment of the 18S rRNA gene, labeled PAMay (Accession no. EU019966, Table 1).

Trematode Trematode Trematode Position Fragment size Accession Plasmid # sequence life stage host (bp) (bp) nos. Fluke A/ Adult Eptesicus fuscus 12 1-1,922 1,922 EU019964 PAFlukeA FlukeB/ Adult E. fuscus 2 797-1,563 767 EU019965 PAFlukeB EU019974 EU019975 SnailA/ Sporocyst EU019976 Elimia virginica 6 1-1,922 1,922 PASnailAa and cercarial EU019977 EU019978 EU019979 EU019970 SnailB/ Sporocyst EU019971 E. virginica 4 1-1,563 or 1,564 1,563 or 1,564 PASnailBb and cercarial EU019972 EU019973 SnailC/ Sporocyst E. virginica 3 797-1,922 1,126 EU019963 PASnailC and cercarial

EU019969 Caddis/ Metacercarial Caddisfly 6 196-1,563 1,368 EU019967 PACaddisc EU019968

May/PAMay Metacercarial Mayfly 4 797-1,922 1,126 EU019966 Table 1. Trematode 18S rRNA sequence sample information. Table is modified from

(Gibson and Rikihisa 2008). a PASnailA is a combination of the following deposited GenBank sequences: PASnailA-1

(EU019974), PASnailA-2 (EU019975), PASnailA-3 (EU019976), PASnailA-4

(EU019977), PASnailA-5 (EU019978), and PASnail-6 (EU019979). b PASnailB is a combination of the following deposited GenBank sequences: PASnailB-1

(EU019970), PASnailB-2 (EU019971), PASnailB-3 (EU019972), and PASnailB-4

(EU019973). c PACaddis is a combination of the following deposited GenBank sequences: PACaddis-1

(EU019969), PACaddis-2 (EU019967), and PACaddis-3 (EU019968).

All trematode life stages are genetically most similar to each other. Region 3 was amplified the most successfully of the four Regions. Therefore, all ten Region 3 sequences and the corresponding regions of P. longiforme and L. linstowi were utilized for alignment comparison. There was 98.7-100.0% sequence identity among the ten

Region 3 sequences from N. risticii-positive trematode samples (Table 2). All ten Region

3 sequences clustered in a group distinct from P. longiforme and L. linstowi with a bootstrap value of 100% (Figure 5). Regions 1, 2 and 4 also demonstrated alignment similarities among the N. risticii-positive trematode sequences with 95.3-99.5% sequence identity in Region 1, 94.1-99.8% identity in Region 2, and 99.1-99.9% identity in Region

4 (Tables 3, 4, and 5).

PA PA PA PA PA PA PA PA PAMay FlukeB SnailA-3 SnailA-4 SnailB-3 SnailB-4 SnailC Caddis-2 Caddis-3 PAFlukeA 99.5 99.7 99.5 99.5 99.5 99.5 99.2 100 99.5

PAFlukeB 99.7 99.5 99.5 99.5 99.5 98.7 99.5 99.5 PASnail A-3 99.7 99.7 99.7 99.7 99.0 99.7 99.7 PASnail A-4 99.5 99.5 99.5 98.7 99.5 99.5 PASnail B-3 99.5 99.5 98.7 99.5 99.5 PASnail B-4 99.5 98.7 99.5 99.5 PASnailC 98.7 99.5 99.5 PACaddis -2 99.2 99.7 PACaddis -3 99.5 Table 2. 18S rRNA Region 3 (767 bp) sequence percent identity for all trematode samples. Table is derived from (Gibson and Rikihisa 2008).

PASnailA-1 PASnailA-2 PASnailB-1 PASnailB-2 PAFlukeA 99.5 99.4 98.7 95.3 PASnailA-1 99.4 98.9 95.3 PASnailA-2 98.3 95.4 PASnailB-1 95.4 Table 3. 18S rRNA Region 1 sequence percent identity for N. risticii-positive trematodes.

Table is derived from (Gibson and Rikihisa 2008). Region 1 (824 bp) sequences were compared among PAFlukeA, PASnailA-1, PASnailA-2, PASnailB-1, and PASnailB-2

(825 bp with a 1 bp insertion).

PASnailA-1 PASnailA-2 PASnailB-1 PASnailB-2 PACaddis-1 PAFlukeA 99.5 99.5 98.9 94.3 99.8 PASnailA-1 99.4 99.0 94.1 99.7 PASnailA-2 98.4 94.4 99.7 PASnailB-1 94.1 98.7 PASnailB-2 94.4 Table 4. 18S rRNA Region 2 sequence percent identity for N. risticii-positive trematodes.

Table is derived from (Gibson and Rikihisa 2008). Region 2 (629 bp) sequences were compared among PAFlukeA, PASnailA-1, PASnailA-2, PASnailB-1, PASnailB-2, and

PACaddis-1.

PASnailA-5 PASnailA-6 PASnailC PAMay PAFlukeA 99.3 99.6 99.7 99.3 PASnailA-5 99.1 99.3 99.4 PASnailA-6 99.9 99.1 PASnailC 99.3 Table 5. 18S rRNA Region 4 sequence % identity for N. risticii-positive trematodes.

Table is derived from (Gibson and Rikihisa 2008). Region 4 (694 bp) sequences were compared among PAFlukeA, PASnailA-5, PASnailA-6, PASnailC, and PAMay.

Figure 5. Dendrogram of Region 3 (767 bp total, including all insertions) of the 18S rRNA gene sequences obtained from all N. risticii-positive trematode samples, P. longiforme, and L. linstowi. Figure is derived from (Gibson and Rikihisa 2008). Bootstrap values greater than 80% are given. *, bootstrap value is 100%. Bar, 1% divergence in sequences.

N. risticii-positive trematodes are most related to family Lecithodendriidae members.

Based on NCBI BLAST searches, the two most similar sequences for PAFlukeA and

PASnailA were the partial 18S rRNA gene of P. longiforme with 1,781 bp out of 1,854 bp (96%) and 1,775 bp out of 1,854 bp (95%) identical, respectively and the partial 18S rRNA gene of L. linstowi with 1,776 bp out of 1,859 bp (95%) and 1,770 bp out of 1,859 27

bp (95%) identical, respectively. It was discovered that the N. risticii-positive trematode sequence started 65 bp upstream of the L. linstowi and P. longiforme available sequences and contained discrepancies of four or seven bp on the 3′ end of L. linstowi and P. longiforme, respectively (refer to Table 6). Therefore, the 5′ and 3′ ends of the 18S rRNA genes compared were cut to a region of similarity (bp 134-1,645, in relation to FlukeA and SnailA), and the insertions and deletions within each sequence were left intact.

PAFlukeA and PASnailA 18S rRNA gene sequences were compared to ten different species of digenetic trematodes, including P. longiforme, L. linstowi, and other trematodes pathogenic and/or parasitic to humans or domestic animals. Prosthogonimus ovatus (Accession no. AY222149), a parasite and pathogen of chickens belongs to the superfamily Plagiorchioidea, which also contains the family Lecithodendriidae (Kigston

1978; Leok, Inoue et al. 2002). Paragonimus iloktsuenensis (Accession no. AY222141, a parasite of dog, cat and rat lungs), and Paragonimus westermani (Accession nos.

AY222140 [1] and AJ287556 [2], parasitizes human lungs) are from the order

Playiorchiida, which is same order as the family Lecithodendriidae (Lee, Koo et al. 1989;

Dekumyoy, Waikagul et al. 1998). N. salmincola (Accession no. AY222138) may directly cause clinical disease in humans and is the reservoir of N. helminthoeca (Nyberg,

Knapp et al. 1967; Eastburn, Fritsche et al. 1987). Zalophotrema hepaticum (Accession no. AJ224884, causes meningoencephalitis in sea lions) and Fasciola gigantica

(Accession no. AJ004804, a zoonotic pathogen), though of different orders from each other and the family Lecithodendriidae (Fauquier, Gulland et al. 2004; McManus and

Dalton 2006) have similarities with the N. risticii-positive trematode sequences upon

NCBI BLAST searches. Given the significance of Schistosoma spp. in human disease

(Cheever and Andrade 1967; Pearce 2005; McManus and Dalton 2006), two species:

Schistosoma japonicum (Accession no. M62652) and Schistosoma mansoni (Accession no. Z11590) were also included. Sequence divergence and bootstrap values demonstrated the clustering of PAFlukeA and PASnailA within the family Lecithodendriidae with a

100% bootstrap value (Figure 6).

trematodes. Figure is derived from (Gibson and Rikihisa 2008). Bootstrap values greater than 80% are given. *, bootstrap value is 100%. Bar, 10% divergence in sequences.

A second dendrogram was created using the same 18S rRNA region as in Figure 6, but deletions of three or more bp within the majority of aligned sequences were removed, resulting in a 1,538 bp region. Removing deletions or insertions affecting the majority of aligned rRNA sequences has been used to represent the uneven distribution of evolutional mutations, although this practice has been criticized in that it does not demonstrate the full ribosome molecule accurately (Littlewood and Olson 2001; Olson, Cribb et al. 2003).

Sequence divergence and bootstrap values in both dendrograms demonstrated the clustering of PAFlukeA and PASnailA within the family Lecithodendriidae with a 100% bootstrap value (Figure 7).

Figure 7. Dendrogram of the 1,538 bp aligned region (including all insertions) of the 18S rRNA gene sequences, with internal gaps of 3 bp or greater in the majority of the aligned sequences removed (a 1,512 bp region for both PAFlukeA and PASnailA). Figure is derived from (Gibson and Rikihisa 2008). The sequences were from the N. risticii- positive trematode samples Fluke A and Snail A and ten other species of digenetic trematode. Accession numbers are shown in parentheses. Bootstrap values greater than

80% are given. *, bootstrap value is 100%. Bar, 10% divergence in sequences.

As the 18S rRNA gene does not experience evolutionary mutations uniformly along its entire sequence, due to secondary structure interference (Olson, Cribb et al. 2003; Xia,

Xie et al. 2003), separate comparisons of Regions 1, 2, 3, and 4 and derived combinations of the regions were made for all N. risticii-infected trematode samples by NCBI BLAST searches. All 18S rRNA gene fragments of all tested N. risticii-infected trematode life stages were the most similar to the 18S rRNA gene fragments of P. longiforme and L. linstowi with 95-96% identity to P. longiforme and 94-96% identity to L. linstowi (Table

6).

Trematode strain ID Region(s)a Most related species Alignment (bp) Comparison size (bp) % identity P. longiformeb 729 762 95% 1 L. linstowic 724 762 95% P. longiforme 744 770 96% 3 L. linstowi 737 771 95% PAFlukeA L. linstowi 663 692 95% 4 P. longiforme 653 686 95% P. longiforme 1781 1854 96% 1, 3, and 4 L. linstowi 1776 1859 95% P. longiforme 743 769 96% PAFlukeB 3 L. linstowi 735 771 95% P. longiforme 726 763 95% 1 L. linstowi 722 762 94% P. longiforme 743 769 96% 3 L. linstowi 735 771 95% PASnailA L. linstowi 660 692 95% 4 P. longiforme 650 686 94% P. longiforme 1775 1854 95% 1, 3, and 4 L. linstowi 1770 1859 95% P. longiforme 737, 728 762, 763 96, 95% 1 L. linstowi 746, 727 763, 764 97, 95% P. longiforme 743 769 96% PASnailBd 3 L. linstowi 735 771 95% P. longiforme 1453, 1444 1502, 1503 96% 1 and 3 L. linstowi 1454, 1435 1505, 1506 96, 95% P. longiforme 745 770 96% 3 L. linstowi 738 771 95% L. linstowi 664 693 95% PASnailC 4 P. longiforme 654 687 95% P. longiforme 1080 1122 96% 3 and 4 L. linstowi 1080 1127 95% P. longiforme 604 634 95% 2 L. linstowi 598 634 94% P. longiforme 738 770 95% PACaddis 3 L. linstowi 731 771 94% P. longiforme 1315 1375 95% 2 and 3 L. linstowi 1302 1376 94% P. longiforme 743 769 96% 3 L. linstowi 735 771 95% L. linstowi 662 692 95% PAMay 4 P. longiforme 652 686 95% P. longiforme 1077 1120 96% 3 and 4 L. linstowi 1076 1126 95% Table 6. NCBI BLAST search results for trematode 18S rRNA sequences. Table is modified from (Gibson and Rikihisa 2008). a Refer to Figure 2 for explanations of the sequenced regions. b All sequences labeled P. longiforme are from Accession no. AY222148. c All sequences labeled L. linstowi are from Accession no. AY222147. d Due to the multiple sequences in the four total plasmids of PASnailB, two different versions of Region 1 and two different versions of Region 2 were created, submitted to

GenBank and NCBI BLAST searched. The presence of a deletion in Region 1 resulted in 33

two different identities for Region 1 and the Regions 1 and 3 combination. Both varieties are shown, separated by commas.

Discussion:

Identification of digenetic trematodes has been shifting from morphologic to genetic.

With this transformation comes the prospect of identifying immature life stages and comparing different trematode life stages to determine if the stages come from the same trematode species. Chapter 2 discusses first constructing 18S rRNA primers for trematodes of the family Lecithodendriidae, based on GenBank-deposited sequences and previously-designed trematode primers. Next, the results of PCR and subsequent sequencing of trematode DNA obtained from N. risticii-positive trematode life stages with these primers are demonstrated; results include an almost complete 18S rRNA gene, comparison of the sequences with known members of the family Lecithodendriidae, and comparison of 18S rRNA fragments within and among the life stages. According to the results, the trematode harboring N. risticii in each tested life stage is the same species, and this trematode species is most related to two species from the family

Lecithodendriidae. This information will not only provide a basis for determining if digenetic trematodes harboring N. risticii differ in strain or species in different regions, but it will also provide future insight into trematode identification in different regions of the United States and in different countries for epidemiological studies of N. risticii prevalence.

Wolbachia sp. strains are members of the family Anaplasmataceae and thus are related to

Neorickettsia spp. They are intracellular bacteria that infect other parasites, including arthropods (such as ticks and the tsetse fly) and filarial nematodes such as Dirofilaria immitis (heartworm). Similar to N. risticii, Wolbachia is vertically and transovarially transmitted (Werren 1997; Fenn and Blaxter 2004; Gibson, Rikihisa et al. 2005).

According to genetic research, strains of Wolbachia have coevolved with their parasitic hosts. Wolbachia strains infecting nematodes have even developed a mutual relationship.

The nematode provides cloistered protection and essentials, such as phosphorous and amino acids, for bacterial survival (Wu, Sun et al. 2004; Foster, Ganatra et al. 2005). In return, Wolbachia may provide basic substrates for the nematode, including riboflavin, flavin adenine dinucleotide, heme, nucleotides or nucleotide-precursors, and glutathione

(Foster, Ganatra et al. 2005; Fenn and Blaxter 2006; Heider, Blaxter et al. 2006).

Wolbachia may even assist its nematode host in parasitizing a mammalian host. The cell membrane of Wolbachia contains a peptidoglycan similar to the peptidoglycan-derived cytotoxin found in Neisseria gonorrhoeae and Bordetella pertussis. Like peptidoglycan- derived cytotoxin, it is hypothesized that the Wolbachia peptidoglycan destroys mucosal layers, initiates sloughing of ciliated cells, and induces inflammation through IL-1 and

IL-6 activation, which would aid the nematode in pathogenesis (Melly, McGee et al.

1984; Martin, Rosenthal et al. 1987; Taylor, Cross et al. 2000; Taylor, Cross et al. 2001;

Cloud and Dillard 2002; Saint Andre, Blackwell et al. 2002; Foster, Ganatra et al. 2005;

Thomson, Crossman et al. 2006). The communal bond between bacterium and worm is so strong that killing the Wolbachia has detrimental effects on the filarial nematode,

including reduced larval growth, delay of larval molting, sterilization of the female, and death (Hoerauf, Nissen-Pahle et al. 1999; Hoerauf, Volkmann et al. 2000; Fenn and

Blaxter 2004; Heider, Blaxter et al. 2006).

Is it possible that N. risticii and its trematode host have evolved a similar relationship?

Could N. risticii be providing the trematode with advantages over its non-infected associates through improved nutrition and/or pathogenesis? Or, is the relationship more like the parasitic bond between arthropods and Wolbachia? In this case, both vertical and rare horizontal transmission of the bacterium among arthropods can occur, thus exposing

Wolbachia to new hosts, new benefits, and new dangers, such as the bacteriophage WO

(Wright, Sjostrand et al. 1978; Masui, Kuroiwa et al. 2001; Fenn, Conlon et al. 2006;

Gavotte, Henri et al. 2007). These Wolbachia strains have more adverse effects on their arthropod hosts, including feminization and killing of males, but they also boast advantages, such as making females more reproductively fit. Furthermore, unlike filarial nematodes, infected arthropods are not adversely affected if the bacteria are killed.

By studying the genetic sequences, researchers may better understand the relationship between N. risticii and its trematode host and, indeed, among all Neorickettsial organisms and their trematode hosts. It could be seen with 16S rRNA, 18S rRNA, and other genes when on the evolutionary time scale N. risticii diverged from other Neorickettsia spp. and how similar N. risticii strains and their trematode hosts are to each other. It could determined if certain trematode genes are tailored to strengthening a mutualistic link, similar to how Ls-ppe-1 in the filarial nematode Litomosoides sigmodontis is hypothesized to provide Wolbachia with phosphorous, so that the bacterium can produce

nucleotides for use by the trematode (Heider, Blaxter et al. 2006). It could be discerned if bacterial genetic material from N. risticii has at some time been passed to its trematode host, as is seen in filarial nematodes containing Wolbachia genetic material (Fenn,

Conlon et al. 2006). Many questions regarding evolution and ecology of N. risticii can be derived, explored, and answered.

Genetic information about the trematode host of N. risticii will also benefit in comparison against trematode hosts of other Neorickettsia spp. As indicated in Figures 6 and 7, N. salmincola (the trematode host of N. helminthoeca) does not cluster with FlukeA or

SnailA. The percentage difference in the 18S rRNA sequenced regions was about 10%.

There is also another important aspect to ribosomal RNA: the secondary structure. These sequencing data, along with previous data from Olson et al. do not include potential secondary structure similarities among compared 18S rRNAs. These similarities could alter the phylogeny, given that the 18S rRNA gene does not experience evolutionary mutations uniformly along its entire sequence (Olson, Cribb et al. 2003; Xia, Xie et al.

2003).

To compensate for deletions and insertions affecting the majority of aligned 18S rRNA sequences, removing these regions has been used in certain cases (Littlewood and Olson

2001; Olson, Cribb et al. 2003). However, this practice has acquired criticism in that it does not demonstrate the full ribosome molecule and the resulting secondary structures.

Attempts were also made to remove regions in which most of the alignments demonstrated a deletion at a certain point. Regions of 3 bp or more were isolated and removed, and phylogenetic analysis with bootstrapping was again performed. These

results, displayed in Figure 8 were very similar in the grouping of the N. risticii-positive trematode with other members of the family Lecithodendriidae as before. It is possible that this narrow field of view is not as affected by secondary structure as it would be for larger more encompassing studies performed previously (Littlewood, Rohde et al. 1999;

Littlewood and Olson 2001; Olson, Cribb et al. 2003). Since all the 18S rRNA N. risticii- positive trematode partial sequences are extremely similar, even without secondary structure analysis it is obvious that these sequences are from the same species. Given the multiple levels with which 18S rRNA genes must be compared, it will be useful to compare sequences of protein-creating genes as well (Olsen and Woese 1993). Future experiments could explore and compare protein sequences, including those for S. falcatus and unknown trematode hosts of other Neorickettsia spp.

To date, no 18S rRNA sequencing data are available for S. falcatus: the trematode host of the SF agent, which is the Neorickettsial organism most closely related to N. risticii by

16S rRNA comparison and is highly related to N. risticii by p51, groEL, and murE sequence comparison (Wen, Rikihisa et al. 1996; Rikihisa, Zhang et al. 2004). It is possible the trematode hosts of N. risticii and the SF agent are genetically related. Given that both N. salmincola and S. falcatus parasitize fish in the metacercarial life stage, it is possible that their genetic identities are more similar to each other and the still unknown trematode host of N. sennetsu which also parasitizes fish (Fukuda, Kitao et al. 1954;

Fukuda and Yamamoto 1981; Wen, Rikihisa et al. 1996; Newton, Rolain et al. 2009). If information is gathered about the sequences of the trematode hosts and comparisons are made among different trematode hosts and within the same trematode hosts in different

geographical regions, the pattern of Neorickettsia evolution and coevolution with the trematode host could be elicited. These comparisons could shed light on incidences on the origins of new Neorickettsial species, differing characteristics (formation of morulae versus individual organisms, and increased virulence within a species; such as the 081 strain of N. risticii (Chaichanasiriwithaya, Rikihisa et al. 1994).

Unlike strains of Wolbachia, Neorickettsia spp. can be horizontally transmitted from trematodes to mammalian cells, thereby causing PHF in horses, Sennetsu neorickettsiosis in humans, and SPD in canids (Rikihisa, Chaichanasiriwithaya et al. 1994). While no 18S rRNA information is available for S. falcatus, based on these data it is obvious the N. risticii-infected trematode is not related to N. salmincola. Perhaps Neorickettsia spp. have been evolving with their trematode hosts for long periods of time, allowing for this separation among the species of trematode hosts. While Wolbachia spp. are found within almost all known filarial nematode species (Bandi, Trees et al. 2001), Neorickettsia spp. have only been reproducibly identified in A. oregonense and family Lecithodendriidae trematodes, N. salmincola, and S. falcatus (Nyberg, Knapp et al. 1967; Wen, Rikihisa et al. 1996; Pusterla, Johnson et al. 2003; Gibson, Rikihisa et al. 2005; Gibson and Rikihisa

2008). This suggests that the extent of Neorickettsial infection among trematode species is not as broad as Wolbachia infection of nematodes and insects. However, as previously reported, it is also possible more species of trematodes are infected with N. risticii or other Neorickettsial organisms (Pusterla, Johnson et al. 2000; Park, Kim et al. 2003;

Pusterla, Johnson et al. 2003; Newton, Rolain et al. 2009).

In conclusion, Chapter 2 demonstrates the relatedness of life stages of the digenetic trematode infected with N. risticii. With the morphologic identification of the trematode as A. oregonense (Gibson, Rikihisa et al. 2005), the clustering of immature life stage 18S rRNA sequences with the adult sequences, and the clustering of all sequences with other members of the trematode family Lecithodendriidae, there is strong support that all life stages are A. oregonense.

Chapter 3:

Identification of Neorickettsia sennetsu Surface-Exposed Proteins and

Characterization of P51 as a Porin

Introduction:

Little information has been obtained on proteins of Neorickettsia spp., in particular surface proteins. The number and sizes of proteins potentially involved in stimulating adaptive immunity in N. sennetsu were identified through recognition of bacterial antigens by N. sennetsu-infected equine sera (Rikihisa, Pretzman et al. 1988). Numbers and sizes of potential surface proteins in N. risticii has also been hypothesized using I125 radiolabeling and Western blotting with anti-N. risticii mouse serum (Kaylor, Crawford et al. 1991). However, due to technical limitations, these protein bands could not be associated with any specific proteins. Although Neorickettsia spp. are genetically most closely related to tick-borne Anaplasma and Ehrlichia spp. (Pretzman, Ralph et al. 1995;

Dumler, Barbet et al. 2001), P44, Msp2, Omp-1/P28/P30, Asp62, Asp55 and Esp73 outer membrane proteins, which are likely significant in nutrient transport, immunoevasion and virulence in Anaplasma and Ehrlichia spp. are not detected in the N. sennetsu genome

(Hotopp, Lin et al. 2006; Ge and Rikihisa 2007; Ge and Rikihisa 2007; Huang, Wang et al. 2007; Kumagai, Huang et al. 2008).

Gram-negative bacteria generally have porins spanning their outer membranes. These proteins enable the transport of hydrophilic molecules, such as amino acids, sugars, and other nutrients (Nikaido 2003). As seen in other members of Anaplasmataceae, N. sennetsu is limited in its ability to synthesize necessary compounds, including amino acids and enzymes for intermediary metabolism and glycolysis (Hotopp, Lin et al. 2006).

Therefore, porins are an absolute necessity for the survival of these bacteria. Before this study, the only porins defined in the order Rickettsiales were major outer membrane proteins of Anaplasma phagocytophilum named P44s (Huang, Wang et al. 2007) and

Omp-1F and P28 in Ehrlichia chaffeensis (Kumagai, Huang et al. 2008). These porins contain 16 (P44s) or 12 (Omp-1F and P28) transmembrane passes, and some are large enough to allow the slow diffusion of tetrasaccharides. P44/Msp2 and Omp-1/P28/P30 proteins belong to pfam01617, and the N. sennetsu genome was reported to encode only one hypothetical protein from this family (NSE_0875) (Hotopp, Lin et al. 2006), later named Neorickettsia surface protein 3 (Nsp3) (Lin, Zhang et al. 2009).

Chapter 3 demonstrates that the Neorickettsia surface is composed of unique proteins as well as proteins conserved among the family Anaplasmataceae. The overall objective is to determine the identity and roles of the major surface proteins of N. sennetsu involved in host-pathogen interaction. Surface-exposed proteins of N. sennetsu were isolated and identified by proteomics. Further confirmed was the surface-exposure of the four highest amino acid coverage proteins: Neorickettsia surface protein 2 (Nsp2), Nsp3, the 51-kDa antigen (P51), and heat-shock protein 60 (GroEL), by immunofluorescence labeling with antibodies to respective recombinant proteins.

The hypothesis was thus formulated that the two major β-barrel proteins, Nsp3 and P51 are porins. To test this hypothesis, the outer membrane fraction was isolated from host cell-free N. sennetsu and examined for porin activity using an in vitro proteoliposome swelling assay. Next, antibodies against the dominant protein P51 were used to examine neutralization of the porin activity. Third, purified native P51 and Nsp3 from isolated outer membrane fraction were acquired by high-pressure liquid chromatography (HPLC) and were tested to see whether these proteins have porin activity. Identification of surface-exposed proteins and the protein with major porin activity will help in better understanding N. sennetsu and the disease Sennetsu neorickettsiosis.

Experimental Procedures:

Culturing and isolation of N. sennetsu. N. sennetsu MiyayamaT (Misao and Kobayashi

2 1954) was cultured in P388D1 cells in 75-cm flasks containing Cellgro RPMI 1640 1X with L-glutamine (Mediatech, Inc., Herdon, VA) supplemented with 10% fetal bovine serum (FBS) (U.S. Biotechnologies, Inc., Pottstown, PA) and 2-3% L-glutamine, 200 mM solution (Invitrogen) at 37°C under 5% CO2. N. sennetsu was isolated as previously described with some modifications (Ge and Rikihisa 2007). As host cells became confluent, the cells were knocked loose of the flasks by banging the flasks against the soft back of a chair and split into multiple new flasks containing fresh media. Decreases in

FBS percentage and additions of uninfected P388D1 cells were made to compensate for too low or too high infectivity levels, respectively. Flasks with high N. sennetsu infectivity (>80%) and large numbers of bacteria per host cell were scraped down using

24-cm cell scrapers (TPP, Switzerland), swirled to suspend the cells in the medium, 43

poured into 50-ml centrifuge tubes (Corning Costar Laboratory Science, Park Ridge, IL), three flask contents per tube, and centrifuged at 448 × g for 5 min to pellet the cells. The media was poured out, and all of the pellets were resuspended in 30 ml of plain RPMI media (final volume was about 33 ml). Half of the suspension volume was then homogenized on ice 50-100 times in a 40 ml Dounce homogenizer (Kimble/Kontes,

Vineland, NJ) and added to a fresh 50 ml centrifuge tube. This was repeated for the remaining volume, and the Dounce homogenizer and pestle were washed with 5 ml of plain RPMI to obtain any leftover bacteria. Host cell debris and unbroken cells were pelleted by centrifugation at 448 × g for 5 min, and the supernatant containing freed bacteria was filtered through a GD/X 5-µm nylon microfiber syringe filter (Whatman,

Inc., Florham Park, NJ), followed by a GD/X 2.7-µm glass microfiber syringe filter

(Whatman).

Biotinylation of N. sennetsu surface proteins, using Sulfo-NHS-SS-Biotin. N. sennetsu isolated from a total of forty-four 75-cm2 flasks was surface biotinylated at 4°C for 30 min as previously described (Ge and Rikihisa 2007). In brief, the bacteria were washed three times with PBS containing 1 mM MgCl2, pH 7.4 and centrifugation at 8,000 × g for

3 min at 4°C. Biotinylation was then performed, using 1 mg/ml of EZ Link Sulfo-NHS-

SS-Biotin (Pierce Biotechnology, Rockford, IL) suspended in PBS containing 1 mM

MgCl2, pH 7.4 for 30 min at 4°C with 15 rotations per minute using a Heto Rotamix, RK

10 VS (Heto-Holten A/S, Allerod, Denmark). The reaction was quenched through washing the bacteria three times with PBS containing 1 mM MgCl2 and 500 mM glycine, pH 7.4 and centrifugation at 8,000 × g for 3 min at 4°C. Biotinylated bacteria were lysed

by sonication for 2 min in 1 ml of radioimmunoprecipitation buffer (25 mM Tris-HCl, pH

7.6, 150 mM NaCl, 1% NP-40, 1% sodium deoxycholate, and 0.1% sodium dodecyl sulfate [SDS]) containing a 1:100 dilution of protease inhibitor cocktail set II

(Calbiochem, San Diego, CA), followed by 30 min incubation on ice with vortexing every 5 min. The supernatant containing biotinylated N. sennetsu proteins were collected by centrifugation at 16,000 × g for 10 min at 4°C, and stored at −80°C in 10% glycerol

(final concentration) until future usage.

Biotinylated N. sennetsu proteins were purified using streptavidin-agarose gel (Pierce) and separated by 12% SDS-polyacrylamide gel electrophoresis (PAGE). The gel was fixed, stained with GelCode blue (Pierce), and submitted to the Mass Spectrometry &

Proteomics Facility (Campus Chemical Instrument Center, The Ohio State University) for protein identification by capillary liquid chromatography-nanospray tandem mass spectrometry (Nano-LC/MS/MS) as previously described (Ge and Rikihisa 2007). Nano-

LC/MS/MS-identified N. sennetsu proteins were analyzed for similarity to other proteins using the NCBI BLAST algorithms blastp and protein-protein blastp (Altschul, Madden et al. 1997). Identified N. sennetsu proteins were also examined for motifs using the

NCBI BLAST against deposited conserved domains search with a default E value threshold of 0.01 (Marchler-Bauer, Anderson et al. 2007).

Western blotting of SS-biotinylated proteins with anti-N. sennetsu, anti-rP51, and anti- rGroEL. Sulfo-NHS-SS-biotinylated and streptavidin-purified N. sennetsu proteins were separated by SDS-PAGE using a 10% running gel and a 4% stacking gel at 4°C. After transfer to Trans-Blot transfer medium pure nitrocellulose membranes (0.45 µm, Bio-

Rad), the membranes were blocked in Tris-buffered saline, pH 7.4 (TBS) containing 5%

Difco skim milk (Becton, Dickinson and Company, Franklin Lakes, NJ) and 0.1%

Tween-20 (Sigma-Aldrich, St. Louis, MO) (Blocking buffer), incubated with a 1:250 dilution of N. sennetsu horse antiserum from Pony 41 collected 07/26/1986 (α-N. sennetsu-1) (Rikihisa, Pretzman et al. 1988), 1:250 and 1:100 dilutions of N. sennetsu horse antiserum from Pony 36 collected 03/31/1987 (α-N. sennetsu-2) (Rikihisa,

Pretzman et al. 1988), 1:500 dilution of anti-N. risticii recombinant P51 rabbit sera (α- rP51), which was previously shown to react with N. sennetsu P51 (Rikihisa, Zhang et al.

2004), or 1:500 dilution of anti-N. sennetsu recombinant GroEL rabbit sera (α-rGroEL)

(Zhang, Ohashi et al. 1997) in Blocking buffer, and washed four times in TBS containing

0.1% Tween-20 (Sigma) (TBST). The membranes were then incubated in a 1:1,000 dilution of horseradish peroxidase (HRP)-conjugated anti-rabbit (for α-rP51 and α- rGroEL) (Cell Signaling Technology, Inc., Danvers, MA) or 1:1,000 dilution of HRP- conjugated anti-horse (for α-N. sennetsu-1 and α-N. sennetsu-2) (Kirkegaard & Perry

Laboratories, Inc., Gaithersburg, MD) in Blocking buffer and then washed four times in

TBST. Enhanced chemiluminescence (ECL) Western blotting substrate (Thermo Fisher

Scientific, Inc., Waltham, MA) and a LAS3000 Intelligent Dark Box (FUJIFILM Life

Science, Stamford, CT) were used to visualize the protein bands. Restore Western blot stripping buffer (Pierce) was employed, according to the manufacturer’s protocol to completely remove primary and secondary antibodies and allow re-probing the membrane. Antibody removal was confirmed through incubation with ECL Western blotting substrate (Thermo Fisher Scientific) and 10 min standard exposure of the membrane with the LAS300 Intelligent Dark Box (FUJIFILM). 46

Anti-Nsp2 and anti-Nsp3 peptide sera and recombinant Nsp2 and Nsp3. Based on

Protean analysis (DNAStar), 15-mer peptides were chosen from N. sennetsu Miyayama

Nsp2 (aa136-149 with a C-terminal cysteine added: PSGWATSKENNKKLC) and Nsp3

(aa107-120 with a C-terminal cysteine added: KKDTKLRTKVPASNC), and rabbit antisera were produced at EZBiolabs (Carmel, IN). Sera were ammonium sulfate precipitated by

EZBiolabs to enrich immunoglobulins. Antisera for Nsp2 (α-Nsp2136-149) and Nsp3 (α-

Nsp3107-120) were reconstituted with dH2O.

Recombinant Nsp2 and Nsp3 and anti-rNsp2 and anti-rNsp3 sera. Full length nsp2 and nsp3, minus their signal sequences were PCR-amplified from N. sennetsu genomic DNA and cloned into pET 33b(+) vectors (EMD, Madison, WI) using NotI and EcoRI sites and

NotI and SalI sites, respectively. The recombinant plasmids were amplified by transformation into DH5α cells (Invitrogen), and inserts were confirmed by sequencing.

The recombinant plasmids were then transformed into BL21(DE3) cells (EMD). One mM isopropyl β-D-1-thiogalactopyranoside (IPTG) was used to induce protein expression, and cells were sonicated for 30 s on ice to separate proteins into supernatant and pellet fractions. Production of recombinant Nsp2 (rNsp2) and recombinant Nsp3 (rNsp3) was confirmed through staining 10% SDS-PAGE gels with GelCode blue and through

Western blotting with HRP-conjugated anti-His (Sigma).

Mouse antisera to rNsp2 and rNsp3 were produced by inoculating six-week-old female

BALB/c mice with 7-10 mg/kg of His-tag-purified rNsp2 or rNsp3 treated with 50-100 column volume washes of 0.1% Triton X-114 (Sigma) to remove endotoxin (Reichelt,

Schwarz et al. 2006) (<5 endotoxin units/kg by the ToxinSensor Gel Clot Endotoxin

Assay Kit [GenScript, Piscataway, NJ]). Recombinant protein was suspended in PBS (the recombinant suspension was tested overnight for sterility in Lysogeny broth) with 10 µg

Quil A (Accurate Chemicals, Westbury, NY) saponin/mouse and injected subcutaneously. Mice were inoculated three times at two week intervals and sacrificed two weeks after the last inoculation for harvesting immune sera. All mice were housed and treated according to the Institutional Animal Care and Use Committee (IACUC) protocol 2008A0066 and IACUC rules and regulations.

P51 secondary structure prediction. The secondary structure of N. sennetsu P51 was predicted using a combination of the programming algorithm in the PRED-TMBB web server (Bagos, Liakopoulos et al. 2004), the hydrophobicity and hydrophobic movement profile (Jeanteur, Lakey et al. 1991), and MegAlign (DNAStar) alignment and analysis of the P51 sequences of N. sennetsu Miyayama (Hotopp, Lin et al. 2006), N. risticii Illinois

(Lin, Zhang et al. 2009), and the SF agent Hirose strain (Rikihisa, Zhang et al. 2004).

Double immunofluorescence labeling of isolated N. sennetsu. N. sennetsu organisms

2 were isolated from four to six 75-cm flasks of >80% N. sennetsu-infected P388D1 cells, affixed to glass slides using a Shandon Cytospin 4 cytocentrifuge (Thermo Fisher

Scientific), and fixed with 4% paraformaldehyde in PBS for 15-30 min at room temperature. The fixed bacteria were washed with TBS and incubated with rabbit antibodies against outer membrane proteins for 1 h in PBS, including pre-absorbed α-

Nsp2136-149 or α-Nsp3107-120, α-rNsp3, α-rP51, or α-rGroEL. After washing in 2× PBS containing 0.05% Tween 20 (Sigma) (PBST) the bacteria were incubated with pre- absorbed α-N. sennetsu-1 for 1 h. After washing in 2× PBST the bacteria were incubated

in a combination of Alexa Fluor 488-conjugated goat anti-rabbit IgG (Invitrogen) and

Cy3-conjugated goat anti-horse IgG (Jackson ImmunoResearch Laboratories, West

Grove, PA) in PBS for 30 min. As negative controls, paraformaldehyde-fixed N. sennetsu were incubated with normal horse sera (N. sennetsu and N. risticii negative) and normal rabbit serum.

Double immunofluorescence labeling of N. sennetsu within P388D1 host cells. N. sennetsu-infected P388D1 cells (approximately 60-90% infected) were affixed to glass slides using a Shandon Cytospin 4 cytocentrifuge (Thermo Fisher Scientific), and fixed with 4% paraformaldehyde in PBS for 15 min at room temperature. The fixed cells were washed with TBS and incubated with antibodies to P51 and Nsp3 for 1 h in PBS containing 0.1% gelatin and 0.3% saponin (PGS), including rabbit anti-N. risticii recombinant P51 (α-rP51, which was previously shown to react with N. sennetsu P51)

(Rikihisa, Zhang et al. 2004), α-Nsp3107-120 pre-absorbed with P388D1 cells, and pre- absorbed α-rNsp3. After washing in 2× PBST, the bacteria were incubated with pre- absorbed horse α-N. sennetsu-1 in PGS for 1 h. After washing in 2× PBST, the bacteria were incubated in a combination of Alexa Fluor 488-conjugated goat anti-rabbit IgG

(Invitrogen) or Alexa Fluor 488-conjugated goat anti-mouse IgG (Invitrogen) and Cy3- conjugated goat anti-horse IgG (Jackson ImmunoResearch Laboratories) in PGS for 30 min.

Outer membrane fraction and isolation of native P51 and Nsp3. N. sennetsu were purified from P388D1 cells by sonication at setting 2 for 32 s and 5 µm filtration using a W-380 ultrasonic processor (Heat Systems, Farmington, NY). A 0.1% (wt/vol) Sarkosyl-

insoluble outer membrane fraction was prepared, and outer membrane proteins were solubilized with 2% (wt/vol) octyl-β-glucoside (OGC, Pierce) (Huang, Wang et al. 2007;

Kumagai, Huang et al. 2008). For antibody neutralization, 25 µg of pelleted outer membrane was treated for 1 h with 20 µl of α-rP51 or control rabbit sera and washed with

10 mM Tris-HCl (pH 8.0) before solubilization with 2% OGC.

HPLC was performed by Dr. Yumi Kumagai as previously described (Kumagai, Huang et al. 2008), and the fractions were tested by GelCode blue staining and Western blotting with respective antibodies for P51 and Nsp3. The combined P51 and Nsp3 fractions were concentrated by evaporation and dialyzed in 50 mM Tris-HCl, pH 8.0 with 1% OGC.

Protein amounts were determined by bicinchoninic acid (BCA) protein assay (Pierce).

Western blotting for rNsp2, rNsp3, and HPLC protein products. Western blotting was performed against rNsp2, rNsp3, OGC-solubilized N. sennetsu outer membrane, and

HPLC-separated P51 and Nsp3 fractions. Antisera used were α-rP51, α-Nsp2136-149, α- rNsp2, α-Nsp3107-120, and α-rNsp3. When Western blotting was performed against rNsp2 with α-Nsp2136-149 and against rNsp3 with α-Nsp3107-120, the primary antibodies were first pre-absorbed against non-transformed E. coli BL21(DE3) proteins separated by SDS-

PAGE and transferred to a nitrocellulose membrane. The secondary antibody used was

HRP-conjugated goat anti-rabbit or HRP-conjugated goat anti-mouse (Cell Signaling

Technology). Protein bands were visualized with ECL by incubating the membrane with

LumiGLO chemiluminescent reagent (Pierce). Images were captured using a LAS3000 image documentation system (FUJIFILM).

Proteoliposome-swelling assay. Twenty-five µg of outer membrane protein preincubated with α-rP51 or normal rabbit sera was incorporated into proteoliposomes for testing the porin activity of the outer membrane fraction of N. sennetsu as previously performed

(Huang, Wang et al. 2007; Kumagai, Huang et al. 2008). In brief, lipid films were created utilizing egg phosphatidylcholine (Avanti Polar Lipids, Alabaster, AL) and dicetyl phosphate (Sigma), impurities were removed from the lipid films by washes in organic solvents (benzene and ether), membrane proteins were added to the lipid films, the protein-lipid mixtures were sonicated to incorporate the proteins into the proteoliposomes, and the proteoliposomes were reconstituted with 15% Dextran T-40 (40 kDa). The same molar concentration of protein, 10 µg and 4.8 µg of P51 and Nsp3, respectively or 5 µg and 2.4 µg of P51 and Nsp3, respectively were also reconstituted into proteoliposomes. Porin activity of the outer membrane fraction, P51, and Nsp3 was determined through proteoliposome swelling utilizing different solutes (Nikaido and

Rosenberg 1983; Nikaido, Nikaido et al. 1991; Huang, Wang et al. 2007; Kumagai,

Huang et al. 2008).

Statistical analysis. The unpaired Student’s t-test was applied to determine the differences among swelling levels. A P value of less than 0.05 was considered significant.

Results:

Nano-LC/MS/MS of streptavidin-affinity purified surface proteins and in silico analysis of N. sennetsu. To experimentally identify bacterial surface-exposed proteins, N. sennetsu were biotinylated with Sulfo-NHS-SS-Biotin, and biotinylated proteins were affinity

purified. GelCode blue protein staining of the SDS-PAGE gel detected multiple protein bands (Figure 8). Forty-two of the total 936 (4.5%) N. sennetsu ORFs were identified by

Nano-LC/MS/MS in eight bands. The majority of proteins identified were demonstrated as expressed proteins for the first time. GroEL (NSE_0642), P51 (NSE_0242), a hypothetical protein (NSE_0456), a hypothetical protein (NSE_0873) and a putative outer surface protein (NSE_0875) were distributed in multiple protein bands. Table 7 summarizes the number of peptides identified for each protein, the percentage of amino acids covered by the identified peptides for each protein, and protein properties including predicted molecular mass, isoelectric point, and presence of a signal peptide. Through analysis of the entire N. sennetsu genome, seven proteins were predicted to be outer membrane proteins by PSORTb (v.2.0) (Gardy, Laird et al. 2005), and Nano-LC/MS/MS demonstrated the surface expression of five of these proteins: P51, an Omp85 family protein (NSE_0718), an outer membrane efflux protein (NSE_0798), NSE_0456,

NSE_0498. Nine lipoproteins were predicted by LipoP (v.1.0) (Juncker, Willenbrock et al. 2003), and a single lipoprotein: NSE_0591 was found by Nano-LC/MS/MS to be surface exposed.

Figure 8. Streptavidin affinity-purified N. sennetsu Miyayama surface proteins. Sulfo-

NHS-SS-Biotin-labeled N. sennetsu surface proteins were affinity purified, separated by

12% SDS-PAGE, and stained with GelCode blue. Bands 1-8 were subjected to Nano-

LC/MS/MS analysis. Lane M (marker), Precision Plus prestained protein standards (Bio-

Rad).

% coverage Mol mass Signal Locus ID Protein name pIa Band(s)d (query)b (kDa) peptidec NSE_0875 Neorickettsia surface protein 3 (Nsp3) 63.0 (95) 26.3 5.75 Yes (24-25) 7 (1, 4, 5, 8) NSE_0873 Neorickettsia surface protein 2 (Nsp2) 61.8 (35) 33.3 6.19 Yes (24-25) 6 (7) NSE_0642 Heat-shock protein 60 (GroEL) 61.7 (88) 58.2 5.24 No 3 (3, 4) NSE_0242 51-kDa antigen (P51) 60.7 (138) 54.6 7.12 Yes (20-21) 3 (1, 2, 4, 7) NSE_0019 Heat-shock protein 70 (DnaK) 32.9 (39) 68.5 5.18 No 2 NSE_0456 Hypothetical protein 31.5 (33) 77.2 8.44 Yes (29-30) 4 (5) NSE_0718 Outer membrane protein, Omp85 family 30.0 (20) 83.6 8.57 Yes (19-20) 1 NSE_0908 Strain-specific antigen 3 (Ssa3) 26.2 (10) 36.5 5.37 No 5 NSE_0005 Iron-binding protein 25.4 (13) 37.6 6.79 No 5 Yes (24-25 NSE_0914 Peptidase, M16 family 22.9 (17) 49.0 6.18 4 or 20-21)e NSE_0664 OmpH 20.3 (14) 22.6 6.23 Yes (21-22) 8 NSE_0732 Hypothetical protein 19.9 (7) 35.2 8.94 Yes (20-21) 6 NSE_0591 Hypothetical protein 19.2 (12) 53.9 4.44 Yes (21-22) 1 NSE_0913 Peptidase, M16 family 19.2 (11) 49.8 8.41 Yes (19-20) 4 NSE_0947 Translation elongation factor Ts (Tsf) 17.6 (5) 31.6 6.05 No 6 NSE_0210 Type IV secretion system protein VirB9-1 15.4 (5) 29.9 8.36 Yes (23-24) 7 NSE_0495 Heat-shock protein 90 (HtpG) 15.0 (11) 71.3 5.54 No 2 NSE_0218 Serine hydroxymethyltransferase (GlyA) 15.0 (7) 45.8 7.17 No 4 NSE_0645 Rotamase family protein 14.6 (13) 68.4 6.59 Yes (35-36) 2 NSE_0063 Putative phosphate ABC transporter 13.6 (4) 37.8 6.16 Yes (17-18) 5 Periplasmic serine protease, DO/DeqQ NSE_0166 12.3 (7) 50.4 8.57 Yes (24-25) 4 family NSE_0725 Pentapeptide repeat domain protein 12.0 (4) 62.7 5.50 HMMf (39-40) 3 NSE_0923 Putative competence protein ComL 11.0 (2) 24.4 6.01 No 8 NSE_0423 ATP-dependent metalloprotease (FtsH) 10.8 (8) 70.2 5.89 HMM (23-24) 2 Succinyl-CoA synthetase, β-subunit NSE_0251 10.2 (6) 42.4 5.98 No 4 (SucC) Polyribonucleotide nucleotidyltransferase NSE_0057 9.78 (9) 82.0 5.99 No 1 (Pnp) NSE_0621 Cytochrome c oxidase, subunit II 9.41 (2) 28.9 5.58 HMM (18-19) 7 NSE_0350 Glutamine synthetase, type I (GlnA) 9.19 (4) 52.6 6.23 No 3 NSE_0472 Putative hydrolase 8.60 (2) 31.4 5.92 No 7 NSE_0254 Triosephosphate isomerase (TpiA) 7.88 (1) 25.9 5.45 No 7 NSE_0798 Outer membrane efflux protein (TolC) 7.11 (3) 49.0 7.77 Yes (17-18) 4 NSE_0900 Hypothetical protein 6.96 (1) 25.7 8.63 Yes (23-24) 8 NSE_0420 Malate dehydrogenase 6.40 (4) 81.3 4.94 No 1 NSE_0687 Translation elongation factor G (FusA) 5.06 (3) 76.1 5.21 No 2 NSE_0948 Ribosomal protein S2 (RpsB) 5.03 (3) 33.3 9.33 No 5 NSE_0578 2-oxoglutarate dehydrogenase (SucA) 4.97 (4) 102 6.01 No 1 3-oxoacyl-(acyl-carrier-protein) synthase NSE_0453 4.10 (1) 43.7 5.70 HMM (25-26) 4 II (FabF) Glyceraldehyde-3-phosphate NSE_0434 3.85 (1) 36.4 5.99 No 5 dehydrogenase, type I (Gap) Succinate dehydrogenase, NSE_0376 3.52 (2) 66.1 5.92 No 2 flavoprotein subunit (SdhA) NSE_0498 Hypothetical protein 2.81 (3) 92.4 6.54 No 1 NSE_0730 Citrate synthase (GltA) 2.76 (2) 44.4 6.62 No 4 NSE_0119 Heat shock protein ClpB protein 2.46 (2) 94.9 6.02 No 1 Table 7. Streptavidin affinity-purified and proteomics-identified proteins for N. sennetsu

Miyayama. Shaded proteins are the four major surface proteins identified. Table is modified from (Gibson, Kumagai et al. 2010). a Theoretical isoelectric point of the given protein as predicted by ExPASy Compute pI/MW tool (Gasteiger, Hoogland et al. 2005).

b Indicates percentage coverage of proteins by identified peptides. Numbers in parentheses are the total number of peptides detected. c Signal peptide presence as determined by the Center for Biological Sequence Analysis

SignalP v.3.0 (Bendtsen, Nielsen et al. 2004). Parentheses indicate amino acids between which cleavage is predicted to occur in the given protein. d Indicates band location(s) (as shown in Figure 8) of the given protein. Bands listed in parentheses are locations in which the protein is found in minority (fewer identified peptides than the major band). e For this protein, the neural networks predicted cleavage between amino acids 24 and 25, and HMM predicted cleavage between amino acids 20 and 21. f Indicates that only the hidden Markov model (HMM) predicts a signal peptide in the given protein.

P51 and Nsp3, found so far only in Neorickettsia spp., had the largest numbers of identified peptides (138 and 95, respectively), and Nsp3 and Nsp2 (another unique protein of Neorickettsia spp.) had the largest percentages of protein coverage (63.0% and

61.8%, respectively). P51 is a major antigen in N. sennetsu, N. risticii, and the SF agent and is predicted to be an outer membrane protein (Dutta, Shankarappa et al. 1991;

Rikihisa, Zhang et al. 2004). Nsp1 (NSE_0872), Nsp2 (NSE_0873), and Nsp3

(NSE_0875) are translated by tandem genes (Lin, Zhang et al. 2009). When their amino acid sequences were compared by protein-protein blastp, Nsp1 and Nsp2 were 23% identical, Nsp2 and Nsp3 were 22% identical, and Nsp1 and Nsp3 were 25% identical. 55

Although three nsp paralogs are present in the N. sennetsu genome (Lin, Zhang et al.

2009), Nsp1 was not identified in the current analysis.

Proteomics analysis further characterized unique and hypothetical Neorickettsia proteins.

This is the first time NSE_0908, known as strain-specific antigen 3 (Ssa3) (Lin, Zhang et al. 2009) was found to be expressed. Strain-specific antigens (Ssas) of N. risticii were previously proposed to be a potential cause of vaccine failure (Biswas, Vemulapalli et al.

1998). One of three N. sennetsu Ssa paralogs, Ssa3 (Lin, Zhang et al. 2009) was found to be surface exposed, although no signal peptide for secretion was predicted for Ssa3 or any other Ssas (SignalP, v.3.0). A strongly-identified protein, NSE_0456, is conserved in

Neorickettsia (Hotopp, Lin et al. 2006; Lin, Zhang et al. 2009), has no homology to other bacterial proteins, and has unknown function. Other hypothetical proteins identified by proteomics, conserved in Neorickettsia, and not found in other bacteria include

NSE_0687, NSE_0591 (no blastp results other than N. risticii), and NSE_0732.

Several proteins identified as surface exposed by proteomics have homologs in Ehrlichia and Anaplasma. Proteins identified with E. chaffeensis and/or A. phagocytophilum also identified in N. sennetsu were GroEL (NSE_0642), DnaK (NSE_0019), the Omp85 family protein, NSE_0664 (renamed OmpH-like outer membrane protein), a rotamase family protein (NSE_0645), a serine protease of the DO/DeqQ family (NSE_0166),

VirB9-1 (NSE_0210), a pentapeptide repeat domain protein (NSE_0725), and translation elongation factor G (FusA, NSE_687) (Ge and Rikihisa 2007; Ge and Rikihisa 2007;

Huang, Lin et al. 2008). Two M16 peptidases (NSE_0913 and NSE_0914) were detected in N. sennetsu. Although homologs are present in Ehrlichia and Anaplasma spp. (E

values ≤ 6 × 10−37), these proteins have not been previously identified as surface exposed

(Ge and Rikihisa 2007; Ge and Rikihisa 2007; Huang, Lin et al. 2008). NSE_0498 and

NSE_0900 also have homologs in Ehrlichia and Anaplasma, yet this is the first time surface protein expression was identified.

Recognition of SS-biotinylated proteins by N. sennetsu, rP51, and rGroEL antisera. N. sennetsu horse antisera strongly recognized protein bands corresponding to P51 (α-N. sennetsu-1) and GroEL (α-N. sennetsu-2). There are also two faint protein bands detected by α-N. sennetsu-1 at 35 kDa and 26 kDa. These two bands are at similar molecular masses as bands 5 and 7 in the SDS-PAGE gel submitted for proteomics and may be recognizing Nsp2 (predicted mass: 33 kDa), Nsp3 (26 kDa), and/or Ssa3 (37 kDa). There is also a faint protein band detected by α-N. sennetsu-2 corresponding to P51 (Figure 9).

Figure 9. Western blotting analysis of N. sennetsu Miyayama surface proteins. Sulfo-

NHS-SS-biotinylated and streptavidin-affinity purified surface proteins were separated by

10% SDS-PAGE, transferred to nitrocellulose membrane and incubated with α-N. sennetsu-1, α-N. sennetsu-2, α-rP51, or α-rGroEL. Lane 1, α-N. sennetsu-1. Asterisks indicate two faint bands. Lane 2, α-N. sennetsu-2. Asterisk indicates one faint band. Lane

3, α-rP51. Lane 4, α-rGroEL. The bands were detected by ECL.

Nsp recombinant protein and antisera production. Western blotting using HRP- conjugated anti-His demonstrated expression of rNsp2 and rNsp3 was primarily in the insoluble fraction. Using these recombinant proteins Nsp2136-149, α-Nsp3107-120, α-rNsp2, and α-rNsp3 were demonstrated to be specific. There was no cross-reaction between the

Nsp2 and Nsp3 antibodies (Figure 10). 58

Figure 10. Recombinant proteins and antisera to Nsp2 and Nsp3. Figure is modified from

(Gibson, Kumagai et al. 2010). (A) Two-dimensional structure prediction of N. sennetsu

Nsp2 (left) and Nsp3 (right) with respect to the outer membrane lipid bilayer by the posterior decoding method with the dynamic programming algorithm in PRED-TMBB.

The N-terminal signal peptides (24 aa from both proteins) were removed from the structures. Peptide sequences used for α-Nsp2 and α-Nsp3 production are indicated by flanking yellow squares. (B) Recombinant Nsp2 and rNsp3 produced in BL21(DE3) insoluble fractions, separated by 10% SDS-PAGE, and stained with GelCode blue. For rNsp2 (expected molecular mass = 35 kDa): Lane 1, GelCode blue-stained rNsp2; lane 2,

Western blotting with anti-His; lane 3, Western blotting with α-Nsp2136-149; lane 4,

Western blotting with α-Nsp3107-120. For rNsp3 (28 kDa): Lane 1, GelCode blue-stained rNsp3; lane 2, Western blotting with anti-His; lane 3, Western blotting with α-Nsp3107-120; lane 4, Western blotting with α-Nsp2136-149. Recombinant Nsp2 (C) and rNsp3 (D) were separated by 10% SDS-PAGE and transferred to nitrocellulose membranes for Western blotting analyses. Anti-His was used to probed non-induced supernatant fraction (lane 1),

IPTG-induced supernatant fraction (lane 2), non-induced pellet fraction (lane 3), and

IPTG-induced pellet fraction (lane 4). Lane 5, IPTG-induced pellet fraction probed with

α-Nsp2136-149; lane 6, IPTG-induced pellet fraction probed with α-Nsp3107-120; lane 7, His- tag-purified recombinant protein probed with α-rNsp2; lane 8, His-tag-purified protein probed with α-rNsp3. All Western blotting bands were detected by ECL.

P51 surface distribution and predicted secondary structure. Procedures used to isolate bacteria from infected host cells may strip off bacterial surface proteins and/or alter their distribution patterns. To examine P51 expression and P51 localization among individual 60

bacterium within infected cells with minimal manipulation, N. sennetsu-infected P388D1 cells were prefixed with paraformaldehyde and permeabilized with saponin. Although it was difficult to discern individual bacteria within infected cells as they were densely packed and overlapped, the majority of intracellular bacteria were labeled with α-rP51 in a rosary-like pattern (Figure 11A).

To confirm bacterial surface exposure and distribution of P51, host cell-free N. sennetsu was prefixed with paraformaldehyde, which does not allow antibody penetration across biological membranes (Wang, Kikuchi et al. 2006) and incubated with α-rP51.

Approximately 50% of isolated bacteria were strongly labeled with α-rP51. This labeling was noted in bacteria ranging in size from 0.6-1.2 µm in diameter; the same size range was noted for non-labeled bacteria. Individual bacteria were often labeled in a distinct and uniform punctate ring-like surface-staining pattern, suggesting P51 antigen exists as multiple clusters along the entire circumference of a bacterium (Figure 11B).

Representative negative controls demonstrated the specificity of the antibodies and lack of bleed-through of another antibody labeling in double immunofluorescence labeling

(Figure 11C and D).

The two-dimensional structure of P51 was then examined, using PRED-TMBB (Bagos,

Liakopoulos et al. 2004). The discrimination value of the P51 amino acid sequence was

2.908, which is below the threshold value of 2.965, making P51 more likely to be a β- barrel protein localized to the outer membrane. Next was examined whether P51 meets the established criteria of porin structure including a length between seven and 16 amino acids with the periplasmic turns having fewer than eight amino acids, an even number of

strands, N- and C-termini of the protein facing the periplasm, lipophilic amino acids on the ends of each β-strand, and amphipathic β-strands being in conserved regions

(Jeanteur, Lakey et al. 1991; Schulz 2002; Nikaido 2003). P51 was predicted to have 18 amphipathic and antiparallel transmembrane β-strands (Figure 11E and F). These predicted transmembrane domains of P51 are highly conserved among the type strains of

N. sennetsu and N. risticii and the first isolate of the SF agent (Table 8).

Figure 11. Surface localization and predicted secondary structure of P51. Figure is from

(Gibson, Kumagai et al. 2010). (A) Labeling pattern of P51 on N. sennetsu within a

P388D1 host cell by double immunofluorescence assay. Infected P388D1 cells were fixed in paraformaldehyde, permeabilized with PGS, incubated with α-rP51 and α-N. sennetsu-

1 (α-NS), stained with Alexa Fluor 488 (green) goat anti-rabbit IgG and Cy3 (red) goat anti-horse IgG, and visualized by fluorescence microscopy. Note that the majority of bacteria are stained with α-rP51. Dashed lines indicate the outlines of the cell and nucleus

(labeled N). Scale bar, 5 µm. (B) Surface localization of P51 on N. sennetsu by double immunofluorescence assay. Host cell-free N. sennetsu were fixed in paraformaldehyde, incubated with α-rP51 and α-N. sennetsu-1 (α-NS), stained with Alexa Fluor 488 goat anti-rabbit IgG and Cy3 goat anti-horse IgG, and visualized by fluorescence microscopy.

Note the regular approximately 0.1 µm diameter dotted pattern of P51. Scale bar, 1 µm.

Negative controls are incubated with α-rP51 and control horse serum (N. risticii and N. sennetsu negative serum, labeled Ctl Horse) (C), or control rabbit serum (labeled Ctl

Rabbit) and α-NS (D). Scale bars, 5 µm. (E) Hydrophobicity and hydrophobic moment profiles (Jeanteur, Lakey et al. 1991) for the P51 sequence. The x axis demonstrates the amino acid number. The y axis depicts the probability of the presence of a transmembrane domain. The black line denotes the presence of normal β-strands. The broken red line indicates twisted β-strands. The N-terminal signal peptide (20 aa) was removed from the structure. The blue numbers demonstrate the locations of the predicted

β-strands. (F) Secondary structure of the N. sennetsu P51 mature protein with 18 transmembrane domains, based on results from (E), PRED-TMBB (Bagos, Liakopoulos et al. 2004), and alignment of the P51 sequences of N. sennetsu Miyayama (Accession no. YP_506136), N. risticii Illinois (YP_003081464), and the SF agent Hirose strain

(AAL12490) using MegAlign (DNAStar) and analysis. Listed amino acids span the outer membrane, with lipophilic residues labeled in red and charged residues labeled in blue.

Source Amino acid at transmembrane domain positiona: T2 T3 T5 T6 T7 T8 T10 T11 T13 T14 T15 T18 2 3 6 3 6 2 5 6 3 6 4 4 7 2 6 4 NSb K K L A I N A N V G A T T M G A NRc R  F V  S  D M   S A   T SFd  E F  V  T D M A G S A A N T Table 8. Amino acid differences among predicted P51 transmembrane domains. Table is from (Gibson, Kumagai et al. 2010). a Bullets represent positions conserved relative to the P51 predicted transmembrane (T) domains T1-T18 of the N. sennetsu Miyayama type strain. T1, T4, T9, T12, T16, and T17 are 100% identical among all strains. Numbers listed below transmembrane domains are the amino acid positions (N-terminus to C-terminus) demonstrating differences. b Miyayama type strain of N. sennetsu P51 protein (Accession no. YP_506136). c Illinois type strain of N. risticii P51 protein (Accession no. YP_003081464). d SF agent Hirose strain P51 protein (Accession no. AAL12490).

Nsp2 and Nsp3 surface distribution and predicted secondary structure. To determine the pattern of labeling within infected cells, N. sennetsu-infected P388D1 cells were prefixed with paraformaldehyde and permeabilized with saponin. Approximately 5-25% of bacteria were labeled with α-Nsp3107-120 and α-rNsp3 within infected cells (Figure 12A).

Nsp2 levels were undetectable by either α-rNsp2 or α-Nsp2136-149 in N. sennetsu-infected

P388D1 cells.

In the absence of saponin permeabilization, approximately 15 and 20% of isolated bacteria were strongly labeled with α-Nsp3107-120 and α-rNsp3, respectively, confirming the bacterial surface exposure of Nsp3. The pattern of Nsp3 surface labeling appeared mainly as polar clusters, and labeling on smaller bacteria (0.6-0.9 µm diameter) seemed more condensed and stronger than on larger bacteria (>0.9 µm diameter, Figure 12 B).

Antisera to Nsp2136-149 recognized about 1% of bacteria with a punctate pattern which was detected on smaller bacteria (0.6-0.9 µm diameter, Figure 13). Antisera to rNsp2 poorly labeled individual bacteria.

Figure 12. Surface localization of Nsp3. Figure is derived from (Gibson, Kumagai et al.

2010). (A) Labeling pattern of Nsp3 on N. sennetsu within a P388D1 host cell by double immunofluorescence assay. Infected P388D1 cells were fixed in paraformaldehyde, permeabilized with PGS, incubated with α-Nsp3107-120 or α-rNsp3 and α-NS, stained with

Alexa Fluor 488 goat anti-mouse IgG and Cy3 goat anti-horse IgG, and visualized by fluorescence microscopy. Dashed lines indicate the outlines of the cells and nuclei

(labeled N). Arrows indicate α-rNsp3 staining. Scale bars, 5 µm. (B) Double immunofluorescence labeling of host cell-free N. sennetsu fixed in paraformaldehyde, incubated with α-Nsp3107-120 or α-rNsp3 and α-NS, stained with Alexa Fluor 488 goat anti-rabbit IgG and Cy3 goat anti-horse IgG, and visualized by fluorescence microscopy.

Note the polar labeling of Nsp3. Scale bars, 1 µm. (C) Negative control is incubated with control mouse serum (Ctl Mouse) and α-NS. Scale bars, 5 µm.

Figure 13. Surface localization of Nsp2. Double immunofluorescence labeling of host cell-free N. sennetsu fixed in paraformaldehyde, incubated with α-Nsp2136-149 and α-NS, stained with Alexa Fluor 488 goat anti-rabbit IgG and Cy3 goat anti-horse IgG, and visualized by fluorescence microscopy. The upper and lower panels are from two different oil-immersion fields. Note punctate labeling pattern on small bacteria. Bar, 1

µm.

GroEL surface distribution. To verify the localization of GroEL on the bacterial surface, host cell-free N. sennetsu was also prefixed with paraformaldehyde, incubated with α- rGroEL and α-N. sennetsu-1, and examined by immunofluorescence microscopy.

Labeling patterns of GroEL were heterogeneous; surface ring-like structures, punctate structures, rings within the bacterial perimeter, and extracellular released GroEL patterns were all noticed in these photomicrographs (Figure 14). 68

Figure 14. Surface localization of GroEL. Host cell-free N. sennetsu were fixed in paraformaldehyde, incubated with α-rGroEL and α-N. sennetsu-1, stained with Alexa

Fluor 488 (green) goat anti-rabbit IgG and Cy3 (red) goat anti-horse IgG. Note the ring- like, punctate, and potential inner surface staining of GroEL. Bar, 1 µm.

Porin activity of the isolated N. sennetsu outer membrane fraction. Since a large amount of P51 was detected on the N. sennetsu bacterial surface, it was first examined whether the isolated N. sennetsu outer membrane fraction has porin activity and whether this activity can be reduced by α-rP51. Porin activity was measured by the proteoliposome swelling assay (Nikaido and Rosenberg 1983; Nikaido, Nikaido et al. 1991) using outer membrane fraction derived from N. sennetsu preincubated with either control serum or α- rP51. When control serum-incubated proteoliposome was mixed with 33 mM (isosmotic solute concentration) of L-glutamine, the monosaccharide arabinose or glucose, or with the tetrasaccharide stachyose, swelling was observed, indicating that the N. sennetsu outer membrane fraction has strong porin activity. When α-rP51-incubated N. sennetsu

outer membrane fraction was utilized in the proteoliposome swelling assay (Figure 15A and C), swelling was significantly reduced with arabinose, glucose, and stachyose compared to the control serum-incubated outer membrane fraction (Figure 15B and C), suggesting P51 is the major porin of N. sennetsu. The antibody was more effective in blocking larger solutes (such as stachyose) than smaller solutes (such as L-glutamine), suggesting steric hindrance of solute diffusion (Figure 15C).

Figure 15. Porin activity of the N. sennetsu outer membrane fraction incorporated into proteoliposomes. Figure is from (Gibson, Kumagai et al. 2010). Optical density changes for the first 20 s in α-rP51-treated (A) and control rabbit-treated (using pre-immune rabbit serum) (B) N. sennetsu outer membrane fraction, using 33 mM L-glutamine (open

squares), arabinose (open diamonds), glucose (filled squares), and stachyose (filled diamonds) for a representative reading of three independent experiments. (C) Initial swelling over 60 s for α-rP51-treated and control rabbit-treated (Control) N. sennetsu outer membrane fraction for three independent experiments. Average blank values (lipid film without protein mixed with 2% OGC in 10 mM Tris-HCl) for three independent experiments of each solute were subtracted from each point. *, P < 0.05; **, P < 0.01 by unpaired Student’s t-test.

Porin activity of isolated native P51 and Nsp3. The isolated N. sennetsu outer membrane fraction contained two major proteins: P51 and Nsp3 (Figure 16A). To experimentally determine if P51 and Nsp3 had porin activity, native P51 and Nsp3 were isolated from the outer membrane fraction by HPLC (Figure 16A). Proteoliposome swelling assays were performed using the same molar concentration of P51 and Nsp3 proteins with L- glutamine, glucose, sucrose, and stachyose as solutes. P51 showed statistically-greater swelling than Nsp3 for glucose, sucrose, and stachyose (Figure 16B). Nsp3 did not demonstrate significant swelling compared with the blank 60 s after the addition of any tested solutes. The results showed that P51 has prominent porin activity, whereas Nsp3 did not have significant porin activity.

Figure 16. Porin activity of HPLC-separated P51 and Nsp3 fractions incorporated into proteoliposomes. Figure is from (Gibson, Kumagai et al. 2010). (A) Lane 1, GelCode blue staining of the OGC-solubilized Sarkosyl-insoluble N. sennetsu outer membrane fraction. Lane 2, HPLC P51 fraction, GelCode blue stain. Lane 3, Western blotting of

P51 fraction with α-rP51. Lane 4, HPLC Nsp3 fraction, GelCode blue stain. Lane 5,

Western blotting of Nsp3 fraction with α-Nsp3107-120. (B) Initial swelling over 60 s for

P51 and Nsp3 fractions for three independent experiments. Average blank values (lipid film without protein 1% OGC in 50 mM Tris-HCl) for three independent experiments of each solute were subtracted from each point. Glucose and sucrose results were obtained

using 10 µg P51 and 4.8 µg Nsp3. L-glutamine and stachyose results utilized 5 µg P51 and 2.4 µg Nsp3. *, P < 0.05; **, P < 0.01 by unpaired Student’s t-test.

Discussion:

Chapter 3 details the first-reported identification of surface-expressed proteins in

Neorickettsia. Expressions of most of these proteins by Neorickettsia spp. were established for the first time. Similar to Anaplasma and Ehrlichia spp., heat-shock proteins, a type IV secretion apparatus protein, and serine protease (HtrA) were surface exposed (Ge and Rikihisa 2007; Ge and Rikihisa 2007; Kumagai, Matsuo et al. 2010).

Two major β-barrel proteins were confirmed by immunofluorescence labeling: the 51- kDa antigen (P51, NSE_0242), and Neorickettsia surface protein 3 (Nsp3, NSE_0875).

P51, a unique protein that has so far only been found in Neorickettsia spp. (Rikihisa,

Zhang et al. 2004), has long been assumed a potential surface-exposed protein. The presence of P51 on the surface of Neorickettsia spp. has been demonstrated for the first time. Protein secondary structure analysis revealed P51 and Nsp3 are predicted to be multi-span β-barrel outer membrane proteins.

While N. sennetsu lacks most genes required for glycolysis (Hotopp, Lin et al. 2006), isolated N. sennetsu can metabolize exogenous L-glutamine and generate ATP and CO2

(Weiss, Dasch et al. 1988; Weiss, Williams et al. 1989). The present study demonstrated

N. sennetsu indeed has a porin: P51 is the major surface-exposed protein and a major porin of N. sennetsu, which can allow the diffusion of L-glutamine and sugars. Compared

with A. phagocytophilum P44s (Huang, Wang et al. 2007) and E. chaffeensis Omp-1F

(Kumagai, Huang et al. 2008), P51 is larger and predicted to have more transmembrane domains (18 β-strands). The pore size of P51 also appears to be larger than that of P44 or

Omp-1 as it allowed diffusion of tetrasaccharide stachyose at a similar rate as glucose or sucrose. Although bacterial porins generally cannot transport sugars larger than disaccharides (approximately 600 kDa pore size) (Koebnik, Locher et al. 2000), among

α-proteobacteria, Omp2a of Brucella abortus can transport the tetrasaccharide maltotetraose when expressed in Escherichia coli (Marquis and Ficht 1993). In most cases, porin proteins have been demonstrated or predicted to contain 16 β-strands

(Nikaido 2003). There are several instances of 18-β-strand porins, including a major outer membrane protein of Campylobacter jejuni (Zhang, Meitzler et al. 2000; Labesse,

Garnotel et al. 2001), a lambda phage receptor protein of E. coli called LamB (Schirmer,

Keller et al. 1995), and a plasmid-encoded sucrose channel called ScrY found in E. coli and Salmonella (Hardesty, Ferran et al. 1991). P51 is, therefore, at the upper end of the sizes of known bacterial porins. The size and abundance of porins are expected to be advantageous for Neorickettsia to acquire a variety of nutrients in the host cytoplasms, as they may have less concern for preventing toxic molecule intake within their protected intracellular environments, similar to mitochondria (Nikaido 2003).

P51 was initially identified in N. risticii and was predicted to be a surface antigen, due to strong antigenic response to the 51-kDa N. risticii antigen by N. risticii and N. sennetsu immune sera (Dutta, Penney et al. 1988; Dutta, Shankarappa et al. 1991; Shankarappa,

Dutta et al. 1992), and a P51 ortholog was later sequenced in N. sennetsu (Rikihisa,

Zhang et al. 2004). This study showed antibodies against rP51 neutralize porin activity in

N. sennetsu. These results corroborate observations from a previous study which showed immune serum against N. risticii inhibits L-glutamine metabolism of isolated N. risticii

(Messick and Rikihisa 1994). There is a small population of bacteria not labeled with α- rP51 within host cells, possibly indicating differential expression of P51 among N. sennetsu populations. Future studies will be required to confirm this preliminary observation.

Chapter 3 has also characterized the N. sennetsu major surface protein Nsp3. Using NCBI conserved domains searches, Nsp1 and Nsp3 were found to belong to pfam01617, and

Nsp2 showed a weak similarity to members of pfam01617 (E value = 0.18). Yet, unlike

P44s and Omp-1/P28 members which belong to pfam01617 and have porin activity,

Nsp3 has no significant porin activity. This is in agreement with the Nsp3 structure prediction which showed only eight β-strands. β-barrel proteins with eight-β-strands and similar structures to porins, such as β8 proteins, are suggested to serve roles other than forming pores, such as in the invasion of host cells (Miller, Bliska et al. 1990; Cirillo,

Heffernan et al. 1996; Baldermann, Lupas et al. 1998; Baldermann and Engelhardt 2000).

With the smaller percentage (15-20%) of smaller diameter (<0.9 µm) bacteria labeled by

α-Nsp3107-120 and α-rNsp3, it is possible that Nsp3 might have such a role. The roles of

Nsp3 and Nsps in Neorickettsia infection remain to be further characterized.

Several heat-shock proteins (hsps) were additionally identified as potential surface- exposed proteins, including GroEL (NSE_0642) and DnaK (NSE_0019). There are numerous reports of both hsps being discovered on the surfaces of different species of

bacteria and interacting with host cell components, including the aforementioned E. chaffeensis and A. phagocytophilum (Ge and Rikihisa 2007; Ge and Rikihisa 2007).

Legionella pneumophila GroEL was discovered to label as an outer surface protein by electron microscopy (EM) and anti-GroEL immunostaining and assist in the invasion of

HeLa cells (Garduno, Faulkner et al. 1998; Garduno, Garduno et al. 1998). The

Salmonella enterica Typhimurium serovar uses a 66-kDa GroEL homolog to interact with host colonic mucus (Ensgraber and Loos 1992). GroEL from Haemophilus ducreyi enables the bacterium to adhere to HEp-2 (epidermoid carcinoma) cells, and EM reveals surface labeling of GroEL with anti-GroEL immunostaining (Frisk, Ison et al. 1998).

Purified GroEL from H. ducreyi is able to bind to various host cell carbohydrate receptors which are also bound by whole bacteria (Pantzar, Teneberg et al. 2006).

Haemophilus influenzae DnaK (hsp-70) is surface-localized and binds to sulfogalactoglycolipids, which are found on gastric mucosa (Hartmann, Lingwood et al.

2001; Mamelak, Mylvaganam et al. 2001). While the irregular surface distribution pattern of GroEL suggests binding of GroEL to the bacterial surface, which may have occurred inside infected cells or during isolation of bacteria prior to biotin labeling, these reports suggest that hsps may contribute to N. sennetsu adhesion to epithelial intestinal cells during invasion or adhesion to mononuclear leukocytes.

An Ssa protein (NSE_0908) was detected in the present study. Although Ssas were considered surface antigens, this was not previously demonstrated (Vemulapalli, Biswas et al. 1995). Ssas of N. risticii have previously been shown to convey strain-specific protection when used as vaccines in mice (Vemulapalli, Biswas et al. 1998). In N.

sennetsu, α-N. sennetsu-1 weakly recognized a protein band consistent with the size of band 5, which is the location of Ssa3, making it possible that α-N. sennetsu-1 is recognizing Ssa3, although further tests are warranted. It was previously determined that two genes, NSE_0904 (molecular mass = 29.5 kDa) and NSE_0908 (molecular mass =

36.5 kDa) and one degenerate gene (NSE_0905) are related to the deposited 85-kDa and

50-kDa Ssas of N. risticii (Hotopp, Lin et al. 2006). Further analysis revealed that two other proteins within this region, NSE_0906 (hypothetical protein, molecular mass = 51.3 kDa) and NSE_0907 (hypothetical protein, molecular mass = 33.8 kDa) and a non-coding region between NSE_0904 and NSE_0905 (bp 805,548-805,581) also demonstrate mild similarities to the 85-kDa and the 50-kDa Ssas. It is possible that the Ssa region within N. sennetsu was duplicated into multiple genes, some of which became degenerate. While only NSE_0908 was detected in the present study, other Ssas may be reserved for surface expression under different conditions.

The two M16 family peptidases discovered by proteomics as surface-exposed

(NSE_0913 and NSE_0914) have a 2 bp intergenic region, suggesting these proteins likely form an operon and are distinct from each other. Motifs for the M16 superfamily active (pfam00675) and inactive (pfam05193) domains are present in both proteins (E values = 6 × 10−7 or smaller) (Marchler-Bauer, Anderson et al. 2007), yet the aforementioned HXXEH inverted Zn metalloendopeptidase motif is only present in

NSE_0914 (Barrett, Rawlings et al. 2004). These findings, along with the approximately

50-kDa size of each protein emulates mitochondrial processing peptidases (MPPs), which are of the M16B family, contain dissimilar α and β subunits of approximately 50-kDa

molecular masses, and in which only the β subunits are enzymatically active (Ito 1999;

Taylor, Smith et al. 2001). NSE_0913 and NSE_0914 also contain orthologs throughout the α-proteobacteria, including Anaplasma and Ehrlichia spp. Similar proteins to MPP have been previously discovered in Rickettsia prowasekii, Mycobacterium spp., and

Bacillus subtilis (Bolhuis, Koetje et al. 2000). Furthermore, A. phagocytophilum and E. chaffeensis employ an operon style for two of their M16 family proteins (APH1158 and

APH 1159 for A. phagocytophilum and ECH1057 and Ech1058 for E. chaffeensis). Of these proteins, Aph1159 and Ech1057 contain the HXXEH motif, whereas Aph1158 and

Ech1058 do not. It is possible that the α-proteobacterial forms of M16 peptidases are similar in function to the MPPs and may be involved in protein processing, nutrient acquisition, gene suppression (Bacillus subtilis) or some other related role (Bolhuis,

Koetje et al. 2000).

There are some proteins that are most likely not surface exposed. This has been seen in similar experiments (Ge and Rikihisa 2007; Ge and Rikihisa 2007), and reasons previously suggested, such as naturally lysed bacteria or damaged outer membranes could also apply to this case. Normally, these proteins have low protein coverage. NSE_0005, an iron-binding protein and considered periplasmic in location (as predicted by PSORTb v.2.0) is an exception, and identified peptides produces 25.4% total and specific protein coverage (Table 7). It is possible that this protein is highly expressed and thus picked up in larger quantities than the other minor proteins that are not surface-exposed.

In conclusion, Chapter 3 defined the N. sennetsu surface proteome. This functional study demonstrated that a hypothesized surface-exposed protein P51 has prominent porin

activity, which can be blocked by specific antibody. This study additionally provided new critical data as a basis for future studies, which will contribute toward better understanding this unique trematode-borne human pathogen and its environmental persistence.

Chapter 4:

Identification of Geographical and Temporal Variation within N. risticii

Surface-Exposed Proteins and Their Antigenicity within Naturally-Infected Horses

Introduction:

Little is known about N. risticii surface-exposed proteins, and this missing information is crucial in the understanding of bacterium-host cell interactions. Antigenic and potential surface proteins ranging between 28 and 110 kDa in mass were previously detected by

Western blotting, but these proteins were not identified (Dutta, Mattingly et al. 1989).

Immunoprecipitation of N. risticii labeled with I125 and N. risticii immune mouse sera revealed potential surface proteins ranging from 25 to 62 kDa in mass, although these proteins were not identified (Kaylor, Crawford et al. 1991). Antigenic proteins of 70, 55,

51, and 44-kDa masses have been demonstrated utilizing recombinant proteins; again the proteins were not identified (Dutta, Shankarappa et al. 1991). Two highly- immunodominant proteins in two N. risticii strains were identified as GroEL and P51

(Vemulapalli, Biswas et al. 1998), but it was not shown whether these proteins were surface exposed. Ssa was suggested as a surface immunogenic protein with potential use in vaccine production, although it was not determined to be bacterial surface exposed

(Dutta, Vemulapalli et al. 1998; Vemulapalli, Biswas et al. 1998).

Chapter 4 has the objective of determining levels of variation within predicted surface- exposed proteins of N. risticii. The Chapter 4 hypothesis was that geographic and temporal variation would occur among strains of N. risticii within these surface-exposed proteins. As demonstrated in Chapter 3, the identification of Neorickettsia proteins is now achievable with the availability of whole genome sequencing data on both N. sennetsu

Miyayama (Hotopp, Lin et al. 2006) and N. risticii Illinois (Lin, Zhang et al. 2009), and

N. risticii surface-expressed proteins are able to be identified in vitro through Sulfo-NHS-

SS-biotinylation and proteomics (Gibson, Kumagai et al. 2010). It was thus endeavored to determine 1) major surface proteins by proteomics analysis on N. risticii, 2) horse immune recognition of N. risticii surface proteins, and 3) strain variations in aligned sequences of these major surface proteins with respect to their predicted secondary structures.

Experimental Procedures:

Culturing and isolation of N. risticii strains. N. risticii IllinoisT (Holland, Ristic et al.

1985) and a Pennsylvania strain (PA-1) (Mott, Muramatsu et al. 2002) were cultured in

2 P388D1 cells in 75-cm flasks containing RPMI 1640 (Mediatech) supplemented with 5-

10% FBS (U.S. Biotechnologies) and 4-6 mM L-glutamine (Invitrogen) at 37°C under

5% CO2. N. risticii was isolated from highly-infected P388D1 cells as previously described for N. sennetsu Miyayama (Gibson, Kumagai et al. 2010).

Biotinylation and streptavidin-affinity purification of N. risticii surface proteins.

Biotinylation of purified N. risticii Illinois and PA-1 from twenty-five 75-cm2 flasks using EZ Link Sulfo-NHS-SS-Biotin (Pierce) and subsequent bacterial lysis and 81

collection of solubilized bacterial proteins were performed as previously described in

Chapter 3 (Gibson, Kumagai et al. 2010). Streptavidin purification of SS-biotinylated N. risticii proteins was then performed, followed by SDS-PAGE and fixation and GelCode blue (Pierce) staining of the gel (Gibson, Kumagai et al. 2010). Proteins from seven bands from N. risticii Illinois and proteins from four bands from PA-1 were identified by

Nano-LC/MS/MS as previously described (Ge and Rikihisa 2007).

Western blotting using recombinant proteins. Recombinant P51 (rP51, 57 kDa), cloned from N. risticii Illinois (NRI_0235), and rNsp2 (35 kDa) and rNsp3 (28 kDa), cloned from N. sennetsu Miyayama (NSE_0873 and NSE_0875, respectively), were expressed by transformed BL21(DE3) cells using isopropyl-β-D-thiogalactopyranoside induction and His-tag purified as described previously (Rikihisa, Zhang et al. 2004; Gibson,

Kumagai et al. 2010). Recombinant GroEL (55 kDa), derived from N. sennetsu

Miyayama (NSE_0642), was acquired from stored aliquots (Zhang, Ohashi et al. 1997).

Fifty µg of each recombinant protein were separated by SDS-PAGE, transferred to nitrocellulose membranes, and cut into strips. Western blotting was then performed on these strips using 1:500 dilutions of known positive horse sera samples as determined by

IFA (Pretzman, Rikihisa et al. 1987; Mott, Rikihisa et al. 1997). The membrane was subsequently incubated with a 1:1000 dilution of horseradish peroxidase-conjugated goat anti-horse (Kirkegaard & Perry Laboratories, Inc., Gaithersburg, MD) as secondary antibody. Enhanced chemiluminescence (ECL) LumiGLO chemiluminescent reagent

(Pierce) and a LAS3000 image documentation system (FUJIFILM Medical Systems

USA, Stamford, CT) were used to visualize the protein bands with 300 s exposure. Bands

were aligned using Precision Plus prestained protein standards (Bio-Rad Laboratories,

Hercules, CA).

Polymerase chain reaction, sequencing, and sequence alignment. DNA was purified from buffy coat of PHF-positive horses or culture of N. risticii in P388D1 cells using the

DNeasy Blood and Tissue Kit (Qiagen), according to manufacturer’s instructions. PCR amplification was then performed using either Phusion or Taq DNA polymerase (New

England BioLabs, Ipswich, MA) and primers designed for conserved regions through alignment of multiple Neorickettsia spp. and/or N. risticii strains (Table 9).

Primer Direction Sequence (5′-3′) Primer pair(s) Gene(s) amplified GP38 Forward GGTTAGTTCTTTGCATTTTACTG GP40 nsp2 GP40 Reverse CAAAGTAGCGTGCAGTATATC GP38 nsp2 GP42 Forward TACCACGACTTCAGTGCTG GP43 nsp2, NRI_0840, nsp3 GP43 Reverse CTTCTTCAGCGAAACCTTC GP42 nsp2, NRI_0840, nsp3 GP41 Forward CTTGACGATGGACTTCTTG GP39 nsp3 GP39 Reverse CACAATTAGGACCGCAAC GP41 nsp3 Nsp1-743F Forward CACACAATATTGAAGCTGGTATAG Nsp2-371R nsp1, nsp2 Nsp2-371R Reverse CTTTCAGCGAGCTTACCTG Nsp1-743F nsp2 Nsp2-225F Forward CTATCTTAATGGTACTGTGATAAG Nsp2/3-NCR-R nsp2, NRI_0840 Nsp2/3-NCR-R Reverse GTTCACCTCTTTGAAGTTTCATAG Nsp2-225F nsp2, NRI_0840 Nsp2-231F-SF Forward CAATGGTACTGTGATAAGAGAATTC Nsp2/3-NCR-R nsp2, NRI_0840 PER51-7 Forward TGTATAAACTTAGCAAGATATTAC TM6 p51 TM6 Reverse CAGCGATGGAAGATACATC PER51-7 p51 PER51-14 Reverse ACACTTGGTGTTAATGTAAGG PER51-7 p51 51K-F7 Forward GTCTTCCAAAGATCGATGTCC 51K-F7 p51 51K-R5 Reverse TTCCGTAACCGGTTTCAAAG PER51-7 p51 KM324 Forward CCGGCTGTTGAAAAAACGACATCA KM325 p51 KM325 Reverse AGCTCATACGTGCTTCCAGTGATG KM324 p51 KM017 Forward GTAACATTCGGAGAGAAGGGTTC KM018 p51, NRI_0234 KM018 Reverse GAGAACAAGATTATAGGGATCCAAGT KM017 p51, NRI_0234 838-1a Forward GGTAAGGATGAAGCAAAAGCAGTAC 838-4, 840-2 ssa1, ssa2, ssa3 838-4 Reverse CTGGTGCATAGTGCACTTCC 838-1a ssa1 840-1 Forward CTAGTGCATCAAAAGGCGTGAG 840-2 ssa3 840-2 Reverse CATTACCTGGACTTTCGAACAGC 838-1a, 840-1 ssa1, ssa2, ssa3 NCR839/840-1 Forward CATAACTTAGGGCTACTATCCC 840-3, NCR840/841-1 ssa3 840-3 Reverse GTGAGAACATTGCCTACTTTATC NCR839/840-1 ssa3 840-3F Forward GATAAAGTAGGCAATGTTCTCAC NCR840/841-1 ssa3 NCR840/841-1 Reverse CTTGTTATGGTAACCTGCTTG NCR839/840-1, 840-3F ssa3 Table 9. Primers utilized for Neorickettsia PCR amplification.

Sequencing was performed by the Ohio State University Plant-Microbe Genomics

Facility. Sequences containing whole genes or gene fragments were translated and aligned mainly through the CLUSTAL W (slow/accurate) method in the MegAlign program of DNAStar. P51 was first aligned by CLUSTAL V (PAM250) method, and

Ssa3 was aligned both by CLUSTAL W and manually. External loops were also aligned separately by CLUSTAL W for both P51 and Nsp3. Amino acid variations in N. risticii strains and other Neorickettsia spp. for all proteins were determined in relation to N. risticii Illinois. Protein alignments of the same size (including deletions as dashes) were analyzed by PHYLIP to obtain bootstrap values for 1000 replicates (using the programs

SeqBoot, Protdist, Neighbor, and Consense) and to create dendrograms (using the programs Protdist, Neighbor, and Drawgram) (Felsenstein 1989). Protein properties, including antigenicity profiles and β-sheet predictions were determined using the Protean program (DNAStar). Gene and protein sequence homologies were also demonstrated using NCBI BLAST algorithms, including blastn, protein-protein blastp, and blastp

(Altschul, Madden et al. 1997; Zhang, Schwartz et al. 2000).

Prediction of secondary structures. Predictions for Nsp2 and Nsp3 were based on a combination of the programming algorithm in the PRED-TMBB web server (Bagos,

Liakopoulos et al. 2004), hydrophobicity and hydrophobic movement profiles (Jeanteur,

Lakey et al. 1991), and MegAlign (DNAStar) alignments and analyses of all available N. risticii strain and Neorickettsia spp. sequences.

GenBank Accession numbers. GenBank Accession numbers of all sequences determined in this study are shown in Table 10.

Sample IDa Location/Year Fragment size (bp)b Gene(s) amplifiedc Accession no. 2091 nsp2, nsp3 HQ857586 PA-1 Pennsylvania/2000 765 ssa1 (p) HQ857584 1812 ssa3 HQ857585 673 p51 (p) HQ857589 Herodia Pennsylvania/1999 2133 nsp2, nsp3 HQ857588 1460 ssa3 HQ857587 2420 nsp2, nsp3 HQ857591 081 Ohio/1991 717 ssa3 (p) HQ857590 676 p51 (p) HQ857594 MN Minnesota/2002 2156 nsp2, nsp3 HQ857593 1029 ssa3 (p) HQ857592 2103 nsp2, nsp3 HQ857596 OV Kentucky/1993 863 ssa3 HQ857595 IA03-1 Iowa/2003 1550 nsp2 (p), nsp3 HQ875741 623 nsp2 (p) HQ875742 IL01-1 Illinois/2001 489 nsp3 (p) HQ875743 IN01-1 Indiana/2001 1879 nsp2 (p), nsp3 HQ875744 IN02-1 Indiana/2002 2052 nsp2 (p), nsp3 HQ875745 542 p51 (p) HQ875747 IN02-2 Indiana/2002 733 nsp3 (p) HQ875746 542 p51 (p) HQ906674 IN03-1 Indiana/2003 2110 nsp2, nsp3 HQ906673 IN03-2 Indiana/2003 1361 nsp2, nsp3 (p) HQ906675 673 p51 (p) HQ906678 594 p51 (p) HQ906679 KY03-1 Kentucky/2003 306 p51 (p) HQ906680 2095 nsp2, nsp3 HQ906677 1129 ssa3 (p) HQ906676 KY03-2 Kentucky/2003 1398 nsp2, nsp3 (p) HQ906681 KY03-3 Kentucky/2003 1042 nsp2 (p), nsp3 (p) HQ906682 259 p51 (p) HQ906685 OH07-1 Ohio/2007 721 ssa1 (p) HQ906683 1739 ssa3 HQ906684 OH07-2 Ohio/2007 259 p51 (p) HQ906686 1558 nsp2 (p), nsp3 (p) HQ906688 OH07-3 Ohio/2007 995 ssa3 (p) HQ906687 654 p51 (p) HQ906691 OH07-4 Ohio/2007 1118 nsp2 (p), nsp3 (p) HQ906690 1029 ssa3 (p) HQ906689 OH10-1 Ohio/2010 768 ssa3 (p) HQ906692 OH10-2 Ohio/2010 660 p51 (p) HQ906693 676 p51 (p) HQ906695 TN02-1 Tennessee/2002 622 p51 (p) HQ906696 1893 nsp2 (p), nsp3 HQ906694 1171 nsp2 HQ906697 SF Oregon Oregon/2004 842 nsp3 HQ906698 370 ssa3 (p) HQ906699 Table 10. Sequences amplified for Neorickettsia. Most sequences were amplified by

Gabrielle Pastenkos. a All samples, except for PA-2 and SF Oregon are from naturally-infected horses. PA-2 is the horse 2 isolate from an experimental infection utilizing N. risticii-infected insects from Pennsylvania (Mott, Muramatsu et al. 2002). Both 081 and OV are strains of N. risticii previously described and with unique morphologies and sequences

(Chaichanasiriwithaya, Rikihisa et al. 1994; Wen, Rikihisa et al. 1995; Kanter, Mott et al. 2000). SF Oregon is a strain of the SF agent (Rikihisa, Zhang et al. 2004). b The largest fragment size acquired containing the given gene(s) is shown. Multiple fragments may be present for a sample. c p, partial sequence for the given gene was obtained.

P51 sequences used in this study are N. risticii Illinois (YP_003081464), PA-1

(AAM18377), PA-2 (AAM18376), Eclipse (AAC01597), SqCaddis (AAM18381),

SqMouse (AAM18380), S21 (AAG03352), TW2-1 (AAR22503), TW2-2 (AAR22504),

25-D (AAB46983), 90-12 (AAB46982), CM1-1 (AAR22501), 081 (AAG03354), OV

(AAG03353), Doc (AAC01595), Oregon (AAC01600), N. sennetsu Miyayama

(NSE_0242, YP_506136), Kawano (AAR23991), Nakazaki (AAR23990), 11908

(AAL79561), SF Hirose (AAL12490), SF Oregon (AAR23988), Dr. Pepper

(AAC01596), Ms. Annie (AAC01599), SHSN-1 (AAB95417), SHSN-2 (AAB95418),

SRC (AAB95419), SCID/CB17 (AAG09962), Snail 2121 (AAF20073), CF-1-snail 2121

(AAF20072), Shasta-horse (AAF43112), Caddis-1 (AAF26718), Caddis-2 (AAF26748),

Siskiyou horse-1 (AAF20069), Siskiyou horse-2 (AAF20070), Siskiyou horse-3

(AAF20071), Juga-1 (AAC01598), and Stonefly-1 (AAF26749). Nsp2 sequences include

N. risticii Illinois (NRI_0839, YP_003082043) and N. sennetsu Miyayama (YP_746740).

Previously-deposited Nsp3 sequences include N. risticii Illinois (NRI_0841,

YP_003082045) and N. sennetsu Miyayama (YP_506742). Ssa3 sequences include N. risticii Illinois (NRI_0872, YP_003082075) and N. sennetsu Miyayama (NSE_0908, 86

YP_506773). The Ssa1 sequence is from N. risticii Illinois (NRI_0870, YP_003082073), and other Ssas are from 25-D (AAC31427) and 90-12 (AAC31428).

Results:

Nano-LC/MS/MS of streptavidin-affinity purified surface proteins. Given that only N. risticii Illinois genome (NC_013009) has been sequenced (Lin, Zhang et al. 2009), these data were used for proteomic analyses. Four N. risticii proteins in N. risticii Illinois (1984 isolate) and five N. risticii proteins (with conserved peptide sequences in relation to N. risticii Illinois) in PA-1 (2000 isolate) contained two or more peptide queries identified by Nano-LC/MS/MS (Table 2). Proteins identified for N. risticii Illinois were P51,

GroEL (NRI_0614), a conserved hypothetical protein (NRI_0567), and Nsp3. The largest protein coverage and the largest number of peptides identified were both from P51.

Proteins identified in PA-1 also included P51 and GroEL; the largest number of peptides was from P51. Minor proteins identified in PA-1 strain were DnaK (NRI_0017), ATP synthase F1, alpha subunit (AtpA, NRI_0132), and Ssa3 (NRI_0872, Figure 17 and Table

11).

Figure 17. N. risticii SS-biotinylation gels. Streptavidin affinity-purified N. risticii protein bands 1-9 of Illinois (A) and 1-4 of PA-1 (B) were submitted to the Mass Spectrometry and Proteomics Facility (Campus Chemical Instrument Center, The Ohio State

University).

Mol Mass % (query) peptide Signal Locus ID Protein name pIa (kDa) coverageb peptidec N. risticii Illinois NRI_0235 51-kDa antigen (P51) 54.9 8.44 52.9 (139) Yes (20-21) NRI_0614 Heat-shock protein 60 (GroEL) 58.1 5.23 46.9 (36) No NRI_0841 Neorickettsia surface protein 3 (Nsp3) 25.7 5.96 12.4 (2) Yes (24-25) PA-1 NRI_0235 P51 54.9 8.44 38.1 (41) Yes (20-21) NRI_0614 GroEL 58.1 5.23 49.8 (36) No NRI_0017 Heat-shock protein 70 (DnaK) 68.4 5.18 2.83 (6) No NRI_0132 ATP synthase F1, alpha subunit (AtpA) 55.9 5.29 3.54 (3) No NRI_0872 Strain-specific antigen 3 (Ssa3) 41.9 6.01 2.62 or 5.24d (2) No Table 11. Proteomics-identified proteins for two N. risticii strains. a Theoretical isoelectric point of the given protein as predicted by ExPASy Compute pI/MW tool (Gasteiger, Hoogland et al. 2005). b Indicates percentage coverage of proteins by all peptides. Numbers in parentheses are the number of peptide queries for each protein identified in the given band. c Signal peptide presence as determined by the Center for Biological Sequence Analysis

SignalP v.3.0 (Bendtsen, Nielsen et al. 2004). Parentheses indicate amino acids between which cleavage is predicted to occur in the given protein. d The peptide detected twice was within the repeated region of Ssa3, therefore the percentage coverage could be two different percentages.

Immune recognition of major surface antigens by PHF-positive horse sera. Bacterial surface-exposed proteins are generally major antigens (Brogden 2007). To determine whether N. risticii surface antigens are recognized by naturally-infected horses, sera from

15 naturally-infected horses with IFA titers greater than or equal to 1:80 using N. risticii- infected P388D1 cells as antigen (Pretzman, Rikihisa et al. 1987; Mott, Rikihisa et al.

1997) and three PHF-negative horse sera were utilized to probe rP51, rGroEL, rNSP2,

and rNSP3 (Figure 1, Table 3). Though only Nsp3 was detected on the surface of N. risticii Illinois in this study, Nsp2 was included because both Nsp3 and Nsp2 from N. sennetsu Miyayama are significant surface proteins (Gibson, Kumagai et al. 2010). All

15 PHF-positive samples demonstrated recognition of rP51, with 11 out of 15 sera having strong recognition. N. sennetsu Miyayama GroEL is 98% identical to N. risticii Illinois

GroEL, and antisera to rGroEL of N. sennetsu cross-reacts with GroEL from multiple species of Rickettsiales, including N. risticii (Zhang, Ohashi et al. 1997). Six out of 15

PHF-positive serum samples demonstrated strong reactivity to rGroEL, with the rest having weak to no reactivity. Nsp2 and Nsp3 from N. sennetsu Miyayama are 83% and

84% identical to Nsp2 and Nsp3 from N. risticii Illinois, respectively, using protein- protein blastp. Only one serum sample reacted strongly to rNsp2, with the rest having weak to no reactivity. Three sera reacted strongly to rNsp3, with the rest having weak to no reactivity. All negative controls did not recognize any of the recombinant proteins.

Figure 18. Western blotting against rP51, rGroEL, rNsp2, and rNsp3 using PHF-positive sera from naturally-infected horses. Recombinant P51, rGroEL, rNsp2, and rNsp3 were separated by SDS-PAGE and probed with 1:500 dilution of PHF-positive horse sera

(PHF sera, 1-10) and negative sera (Neg sera, N1-N3). All bands were aligned using

Precision Plus prestained protein standards (Bio-Rad). The arrow in panel A indicates the expected location of P51. Arrows in panels B-E indicate the locations of the recombinant proteins. Information regarding the sera samples is given in Table 3. Western blotting was performed by Susanne Moesta.

Horse IDa Clinical signsb Location Year IFA titer 1 (OH10-1) A, F, De, Deh, C Johnstown, OH 2010 >1:10,240 2 (OH10-2) A, F, De, C, L, Et, EUTH Grove City, OH 2010 >1:10,240 3 A, F, De, Deh, L, Et, EUTH Richwood, OH 2010 >1:10,240 4 A, De, F Galloway, OH 2010 >1:10,240 5 A, De, Deh, F, C, L Dayton, OH 2010 >1:10,240 6 A, F, C, L, EUTH Loveland, OH 2010 >1:10,240 7 U Indiana 2010 1:5120 8 A, Di, De, Deh, F, L Troy, OH 2008 1:1280 9 U Kentucky 2008 1:1280 10 U Indiana 2008 1:1280 11 A, F, Di, De, Deh Columbus, OH 2008 1:1280 12 A, F, Di Cattaraugus, NY 2010 1:640 13 U Indiana 2008 1:640 14 A, F, C Oak Hill, OH 2008 1:80 15 A, F Utica, OH 2008 1:80 N1 U New Jersey 2010 <1:20 N2 U Ohio 2010 <1:20 N3 U New Jersey 2010 <1:20 Table 12. PHF-positive sera from naturally-infected horses and negative sera. Clinical sign records were obtained by Susanne Moesta. a Sera 1 and 2 are from the same horses as buffy coats OH10-1 and OH10-2, respectively, as identified in Table 10. b A, anorexia; F, fever; De, depression; Deh, dehydration; C, colic; L, laminitis; Et, endotoxemia; EUTH, euthanized; UN, Unknown; Di, diarrhea.

Sequence variation in P51. P51 sequences are known to be strain variable (Kanter, Mott et al. 2000; Rikihisa, Zhang et al. 2004). Since P51 was found to be the major target of horse immune recognition, the parts of the P51 in which sequence variations occur were examined. N. sennetsu P51 was predicted to have 18 transmembrane β-barrel proteins with nine external loops (Gibson, Kumagai et al. 2010). N. sennetsu and SF agent, which are closely related to N. risticii (Fukuda and Yamamoto 1981; Rikihisa 1991; Rikihisa,

Zhang et al. 2004) were included for comparison. P51 alignments of a total of 52 sequences and sequence fragments from N. risticii during a 26-year period throughout the 92

United States revealed high variability within regions corresponding to external loops 2 and 4 (Figure 19). Forty-three P51 sequence fragments (aa 136-176) containing most of external loop 2 (aa 120-176), and 36 P51 sequence fragments (aa 259-286) containing the entire external loop 4 were analyzed using PHYLIP (Figure 20A and B). Both loops 2 and 4 created patterns of clustering for sequences from states in the Eastern and

Midwestern United States (East/Midwest US) and sequences from Japan, Malaysia, and

US states bordering the Pacific Ocean (Pacific coast). The California strain Doc and the

Ohio strain 081 did not follow this pattern, both being in East/Midwest US in external loop 2 and Pacific coast in external loop 4. In external loop 2, N. risticii Illinois was only loosely associated with the other East/Midwest US sequences; in external loop 4, N. risticii Illinois tightly clustered with several East/Midwest US sequences. External loop 4 of 081 clustered with SF agent strains rather than with other N. risticii strains.

Figure 19. P51 amino acid sequence variations among Neorickettsia sequences. Amino acids different from N. risticii Illinois, including insertions and deletions are divided by the number of sequences plotted for each amino acid position (# aa diffs). The horizontal axis displays P51 amino acid positions (aa position) including the signal peptide and all detected amino acid insertions (515 aa total). SP, signal peptide. E, external loop; and

TM, transmembrane domain are based on the predicted secondary structure (Gibson,

Kumagai et al. 2010). The number of sequences available at each amino acid position on

P51 (# seqs) is shown below.

Figure 20. Dendrograms of P51 amino acid sequence variations among Neorickettsia sequences. Dendrograms of P51 from (A) a 41-aa fragment (counting all insertions) including the majority of predicted external loop 2 with 43 sequences and (B) a 31-aa fragment (counting all insertions) including the entire predicted external loop 4 with 36 sequences are shown with bootstrap values greater than 50.0% for 1000 replicates. *, bootstrap value of 90.0% or greater. East/Midwest US, sequences from states in the

Eastern and Midwestern US. Pacific coast, sequences from Japan, Malaysia, and US states bordering the Pacific Ocean.

Sequence variation in Nsp2. Nsp2 sequences of N. risticii, other than the sequence from

N. risticii Illinois, have not been determined. Nsp2 was predicted to have eight transmembrane β-barrel domains with four external loops. A total of 20 Nsp2 proteins and protein fragments were aligned, with amino acid variations determined in relation to

N. risticii Illinois, and variations mainly occurred in external loops, with the most variation occurring within external loop 4 (Figure 21A). Full-length Nsp2 (including the signal peptide), with 11 sequences total, as well as the external loop 4 region (aa 244-

297) with 19 sequences total were analyzed by PHYLIP (Figure 21B and C). For full- length Nsp2 and external loop 4, most N. risticii strains obtained after the year 2000

(post-2000 strains, Table 10) were 100% identical, whereas other strains were more diverse (Figure 21B and C). Nsp2 for both N. risticii Illinois and Herodia (which were

100% identical) were unique to all other N. risticii strains. The 081 strain clustered with

SF Oregon, rather than with other N. risticii strains.

Figure 21. Nsp2 amino acid sequence variations. (A) Amino acids different from N. risticii Illinois, including insertions and deletions are divided by the number of sequences plotted for each amino acid position (# aa diffs). The horizontal axis displays Nsp2 amino acid positions (aa position) including the signal peptide and all detected amino acid insertions (309 aa total). SP, signal peptide. E, external loop; and TM, transmembrane domain are based on the predicted secondary structure. The number of sequences available at each amino acid position on Nsp2 (# seqs) is shown below. (B) Dendrograms of Nsp2 from the full-length protein, including the signal peptide (12 sequences total) and

(C) the predicted external loop 4 (55 aa, including all insertions; 19 sequences total) are shown with bootstrap values greater than 50.0% for 1000 replicates. *, bootstrap value of

90.0% or greater. Post-2000 sequences are shown in the shaded areas.

Sequence variation in Nsp3. Nsp3 sequences of N. risticii, except for the sequence from

N. risticii Illinois have also not been determined. Nsp3 was predicted to have eight transmembrane β-barrel proteins with four external loops. Alignment of a total of 21

Nsp3 proteins and protein fragments demonstrated the highest variation within predicted external loop 2, yet there was less variation in the C-terminal region comprising external loops 3 and 4 (Figure 22A). Fourteen full-length Nsp3 sequences (including signal peptides) and 17 external loop 2 regions (aa 102-136) were analyzed by PHLYIP (Figure

22B and C). As seen in Nsp2, N. risticii Illinois had marked differences to other sequences, in particular to most post-2000 strains (Table 10). TN02-1 and IL01-1 had the highest similarity to N. risticii Illinois in external loop 2.

Figure 22. Nsp3 amino acid sequence variations. (A) Amino acids different from N. risticii Illinois, including insertions and deletions are divided by the number of sequences plotted for each amino acid position (# aa diffs). The horizontal axis displays Nsp3 amino acid positions (aa position) including the signal peptide and all detected amino acid insertions (264 aa total). SP, signal peptide. E, external loop; and TM, transmembrane domain are based on the predicted secondary structure. The number of sequences

available at each amino acid position on Nsp3 (# seqs) is shown below. (B) Dendrograms of Nsp3 from the full-length protein, including the signal peptide (14 sequences total) and

(C) the predicted external loop 2 (57 aa, including all insertions; 17 sequences total) are shown with bootstrap values greater than 50.0% for 1000 replicates. *, bootstrap value of

90.0% or greater. Post-2000 sequences are shown in the shaded areas.

Sequence variation in Ssa3. Ssa3 sequences of N. risticii, other than that of N. risticii

Illinois have not been ascertained. Ssa3 was included in the analysis, since unknown Ssas were previously reported as major N. risticii surface antigens in a 1984 Maryland strain

25-D and a 1990 Maryland strain 90-12 (Biswas, Vemulapalli et al. 1998) and a small amount Ssa3 was detected in both N. risticii PA-1 in this study, and N. sennetsu

Miyayama (Gibson, Kumagai et al. 2010). There was no signal peptide predicted for Ssa3

(Lin, Zhang et al. 2009), and Ssa3 was not predicted to have a β-barrel structure. It was originally shown that ssas contain a wide variety of mainly small repeats of 10-55 bp in size (Biswas, Vemulapalli et al. 1998). Tandem repeats ranging in size from 63-156 bp are present in ssa1, ssa2, and ssa3 of N. risticii Illinois (Lin, Zhang et al. 2009). The N terminus of Ssa3 contains 2.2 copies of a 52-aa (156 bp) tandem repeat in N. risticii

Illinois (aa 53-196) (Lin, Zhang et al. 2009). Thirteen Ssa3 proteins and protein fragments were aligned and compared. Within this N-terminal repeated region,

Neorickettsia spp. consisted of anywhere from zero to four repeated 52-aa peptides arranged in tandem followed by a terminal 40-aa peptide similar to the 52-aa repeats (for

N. risticii Illinois: 50% identical, E value = 6 × 10−8, using protein-protein blastp). It 100

appears that the number of 52-aa repeats increases over time; six post-2000 strains (Table

10) have four repeats. There is further variety in the form of point mutations within the

52-aa repeats and 40-aa peptide. In addition, the 40-aa peptide in SF Oregon was truncated by 9 aa (31 aa in length, with the downstream sequence aligning with the other

Neorickettsia sequences downstream of their 40-aa peptides) (Figure 23). Of note, there are β-sheets predicted to encompass most of the repeated region (aa 40-67; 76-119; 128-

167) and scattered within the C-terminal region (aa 235-433).

101

Figure 23. Ssa3 amino acid sequence variations. Ssa3 changes in amino acids, including insertions and deletions were divided by the number of sequences for each amino acid position (# aa diffs). The length of Ssa3 (horizontal axis) displays Ssa3 amino acid positions (aa position) and includes all amino acid insertions (537 aa total), and the number of sequences available at each amino acid position on Ssa3 (# seqs) is given. The amplified Ssa3 repeat region location (aa 53-196, in relation to N. risticii Illinois) and variety are demonstrated below. Dark gray boxes indicate the 52-aa repeats found in the

N. risticii sequences. Light gray boxes indicate the terminal 40-aa peptide found in all

102

Neorickettsia sequences. Black lines indicate amino acid variations in relation to N. risticii Illinois. The box containing diagonal lines in SF indicates a 9-aa truncation in the

40-aa peptide in relation to the other Neorickettsia spp.

Sequence variation in ssa1. Ssa1 sequences of N. risticii, other than that of N. risticii

Illinois have not been determined. Given the closest similarities between ssa1 of N. risticii Illinois and the unknown ssas from N. risticii strains 25-D (isolated in 1984) and

90-12 (isolated in 1990) (Lin, Zhang et al. 2009), two ssa1 fragments were amplified, sequenced, and translated from PA-1 and OH07-1. When sequence variation comparison was attempted among N. risticii Illinois, PA-1, OH07-1, and N. sennetsu Miyayama, the variation present was extensive, demonstrating the extent of overlaps and repeats (Figure

24A). Protein alignments among PA-1, OH07-1, and N. risticii Illinois Ssa1 fragments and whole protein demonstrated unique repeats and overlapping amino acid patterns with greatest homology between PA-1 and OH07-1 (Figure 24B). PA-1 (aa 11-249) and

OH07-1 (aa 1-239) Ssa1 fragments were aligned with corresponding regions from N. risticii Illinois Ssa1 (aa 246-469) and the Ssas from 25-D (aa 287-507) and 90-12 (aa

579-817). Interestingly, Ssa1 fragments from PA-1 and OH07-1 clustered with the 90-12

Ssa, rather than with N. risticii Illinois Ssa1 and the 25-D Ssa, suggesting a chronological trend (Figure 24C).

103

Figure 24. Ssa1 amino acid sequence variations. (A) Ssa1 changes in amino acids, including insertions and deletions and Ssa1 sequences from N. risticii Illinois, PA-1,

OH07-1, and N. sennetsu Miyayama Ssa1 were divided by the number of sequences for each amino acid position. The length of the Ssa1 fragment (horizontal axis) includes all amino acid insertions (274 aa total), and the number of sequences available at each amino acid position on Ssa3 (# sequences) is given. (B) Ssa1 alignment results for PA-1 and

OH07-1 protein fragments and N. risticii Illinois whole protein using the protein-protein blastp (Altschul, Madden et al. 1997). Axes labeled N. risticii Illinois Ssa1 represent the size of the entire N. risticii Illinois Ssa1 protein (501 aa). Axes labeled PA-1 Ssa1 represent the size of the PA-1 Ssa1 protein fragment (255 aa). Axes labeled OH07-1 Ssa1 represent the size of the OH07-1 Ssa1 protein fragment (239 aa). Lines indicate regions of homology between the two compared sequences. (C) The dendrogram of a 241-aa 104

fragment of Ssa1, including all insertions (five sequences total) is shown. *, bootstrap value of 90.0% or greater.

Discussion:

The genes p51, nsp2, nsp3, and ssa3 are uniquely evolved in Neorickettsia spp. The gene p51 is a single copy gene and demonstrates only loose associations with other proteins of the family Anaplasmataceae (Hotopp, Lin et al. 2006; Lin, Zhang et al. 2009). The nsps and ssas are both potential operons, consisting of three genes tandemly arranged (Lin,

Zhang et al. 2009). The nsps belong to pfam01617, and similar to Ehrlichia chaffeensis omp-1 (p28) genes (also from pfam01617) (Miura and Rikihisa 2007), the proteins encoded by nsp2 and nsp3 were strain variable. As seen in the ssas, other members of the family Anaplasmataceae have genes encoding proteins containing strain-variable tandem repeats (involving amino acid variation and changes in the numbers of tandem repeats), including Trp120 (formerly gp120), Trp47 (formerly gp47), and VLPT (variable-length

PCR target) from E. chaffeensis and Trp140 (formerly gp140), Trp36 (formerly gp36), and gp19 from Ehrlichia canis (Sumner, Childs et al. 1999; Doyle, Cardenas et al. 2005;

Zhang, Luo et al. 2008). Of note, the proteins encoded by the ssas are not homologous to any proteins of the family Anaplasmataceae by blastp. Among p51, the nsps, and the ssas, there have been no signs of intragenomic recombination events, which are seen in the Anaplasma p44/msp2 expression locus (Barbet, Meeus et al. 2003; Lin, Rikihisa et al.

2003).

105

Proteomics results performed on two strains of N. risticii established that P51 is a dominant surface-expressed protein. The recognition of recombinant P51 by PHF horse sera, even by 1:80 IFA titer sera suggests P51 is expressed and highly recognized within the present day naturally-infected horses. Despite P51 amino acid sequence variation among N. risticii strains, this strong universal recognition by horse immune sera suggests rP51 may serve as a defined serodiagnostic antigen. Furthermore, the study suggests that there are immunodominant conserved peptide sequences within P51 which might serve as even more specific PHF diagnostic antigens.

Sequence comparison of these surface-exposed proteins of N. risticii strains, with respect to the predicted protein secondary structure, the majority of which are clinical isolates, indicates there are hot spots within the genes with greater strain divergence. These include external loops 2 and 4 in P51, external loop 4 in Nsp2, external loop 2 in Nsp3, and the repeated region of Ssa3. P51 showed strong geographical association; and Nsp2,

Nsp3, and Ssa3 showed temporal association. Importantly, N. risticii Illinois (upon which vaccines for PHF are produced) is distinct from most East/Midwest US strains (P51) and most post-2000 strains (Nsp2, Nsp3, and Ssa3), which may be a contributing factor in

PHF vaccine failure (Vemulapalli, Biswas et al. 1995; Dutta, Vemulapalli et al. 1998).

There are outlier strains which do not fit the geographical and temporal patterns. These include 081 (Chaichanasiriwithaya, Rikihisa et al. 1994; Wen, Rikihisa et al. 1995), the

Kentucky strain OV (Wen, Rikihisa et al. 1995), and the Kentucky strain Herodia.

Unique sequences in other N. risticii strains, such as TN02-1 (P51, Nsp2, and Nps3),

KY03-3 (Nsp2), IL01-1 (Nsp3), and OH10-1 (Ssa3), suggest that variation contrary to

106

the popular geographical and temporal influences may be more widespread. When additional contemporary sequences and sequences from more varied geographic regions become available, these analyses are expected to improve.

Possible explanations for extensive DNA sequence variation within Neorickettsia include the defective DNA repair systems in both N. risticii and N. sennetsu (Hotopp, Lin et al.

2006; Lin, Zhang et al. 2009). This would result in higher mutation rates for

Neorickettsia (LeClerc, Li et al. 1996), which would agree with the temporal changes and the production of outlier strains of N. risticii. P51 variation showed substantial geographical association, suggesting these variations were selected under local environmental pressures. It is possible that geographical association of N. risticii sequence variation is due to N. risticii strains being selected within essential reservoir trematode populations. In addition, diverse N. risticii strains may have emerged due to selective pressures inflicted on the infected trematodes and/or on the trematodes’ hosts

(Barlough, Reubel et al. 1998; Chae, Pusterla et al. 2000; Kanter, Mott et al. 2000;

Pusterla, Madigan et al. 2000; Mott, Muramatsu et al. 2002; Park, Kim et al. 2003;

Pusterla, Johnson et al. 2003; Gibson, Rikihisa et al. 2005; Gibson and Rikihisa 2008).

However, since Neorickettsia spp. are known (N. risticii and N. helminthoeca) and suspected to be vertically transmitted within their trematode hosts (Nyberg, Knapp et al.

1967; Gibson, Rikihisa et al. 2005; Rikihisa, Dumler et al. 2005), mammalian infection is not expected to be required for maintaining Neorickettsia in the natural environment.

Humoral immunity would, thus not play any direct role in creating genetic diversity within N. risticii populations. Regardless the cause, this genetic variation would result in

107

increased N. risticii survival as a species, and subsequent surface antigen diversity would have links to variable PHF virulence and clinical signs, which may result in vaccine failure.

It is important to note the external loops demonstrating more conserved sequences among

N. risticii strains and Neorickettsia spp. These include external loops 3 and 4 from Nsp3 and external loops 3, 5, and 8 from P51. Even though the mammalian immune system is not expected to play a role in N. risticii evolution, the conservation in these external regions suggests that they are vital for the bacterium. It is possible that peptides from these external loops may be used singly or in conjunction to create more effective vaccines for N. risticii.

Genes encoding the two original Ssas, called P85 (90-12) and P50 (25-D) are most related to ssa1 from N. risticii Illinois (Vemulapalli, Biswas et al. 1995; Biswas,

Vemulapalli et al. 1998; Dutta, Vemulapalli et al. 1998; Lin, Zhang et al. 2009), but they also show similarities to ssa2 and the non-coding region between ssa1 and ssa2 using blastn. Although both are Maryland isolates, the 25-D strain is six years older than the

90-12 strain (Biswas, Vemulapalli et al. 1998), suggesting both temporal variation and the potential development of chimeras of multiple Ssas and non-coding regions in P50,

P85, and post-2000 Ssa1 (due to the similarities of PA-1 and OH07-1 Ssa1 fragments to

P85). It is possible that the high variability of Ssa1 may have prevented PA-1 Ssa1 from being identified by proteomics. However, there is the obvious lack of large numbers of peptides identified by proteomics for Ssas in N. risticii Illinois using the isogenic Illinois strain sequence data and in N. sennetsu using Miyayama isogenic strain data (Gibson,

108

Kumagai et al. 2010). It is likely that Ssas are not a dominant surface protein in mammalian cells.

In conclusion, these data meet the objective of Chapter 4 by demonstrating the variety present within major surface proteins of N. risticii, as well as suggesting conservation among geographical regions and time periods. In addition, P51 is implicated as the major surface antigen of N. risticii. The data will help developing better diagnostic methods.

109

Chapter 5:

Cytokine Induction by N. sennetsu Outer Membrane proteins

Introduction:

Understanding bacterial pathogenesis is key to developing effective treatments and prevention of severe disease. Cytokines and chemokines are known to modulate disease processes and immunity (Tizard 2008). The overall objective of Chapter 5 is to determine causes of cytokine induction in N. sennetsu infection, thus potential reasons for disease symptoms. Mice were originally used to propagate N. sennetsu (Fukuda, Kitao et al.

1954; Misao and Kobayashi 1954; Misao and Kobayashi 1955) and to perform experiments on increasing bacterial growth and purifying complement-fixing bacterial antigens (Tachibana and Kobayashi 1975; Tachibana, Kusaba et al. 1976). It has been previously shown that mice demonstrated progressing clinical signs with N. sennetsu infection (Misao and Kobayashi 1955), although no previous studies on pathogenesis were performed, nor were any cytokine studies reported. Therefore, underlying objectives of Chapter 5 are 1) to develop a mouse model of Sennetsu neorickettsiosis to determine cytokine responses, 2) to identify critical cytokines, and 3) to determine the mechanism(s) of cytokine induction.

110

In vitro and in vivo pathogenesis studies have been performed on the highly-related bacterium N. risticii. Mouse pathogenesis studies have demonstrated splenomegaly, thymic atrophy, and reduction lymphocyte proliferation and IL-2 production by spleen cells from infected CF-1 mice (Rikihisa, Johnson et al. 1987). Mouse models have been used to study antibiotic effectiveness (Rikihisa and Jiang 1989) and potential N. risticii vaccines (Rikihisa 1991), identify N. risticii antigens (Kaylor, Crawford et al. 1991;

Biswas, Vemulapalli et al. 1998) demonstrate transmissibility of N. risticii from trematodes to mammal (Kanter, Mott et al. 2000), and demonstrate differences in disease according to inbred and outbred mouse strains (Williams and Timoney 1994). In vitro cytokine expression has been studied using thioglycolate-induced mouse peritoneal macrophages, demonstrating IL-1α induction, which was predicted to play a role in disease, immunosuppression detected as a reduction of IL-6, and PGE2 by E. coli LPS when the macrophages were first incubated with N. risticii (van Heeckeren, Rikihisa et al. 1993), and the killing of N. risticii by IFNγ in a nitric oxide-dependent manner (Park and Rikihisa 1991; Park and Rikihisa 1992).

Cytokine induction by pathogenic bacteria is due to recognition of pathogen-associated molecular patterns (PAMPs), which are by majority not found in the eukaryote. PAMPs include LPS, peptidoglycan, lipoproteins, flagellin, and CpG bacterial DNA (Tizard

2008). In addition, heat-shock proteins, found on both prokaryotes and eukaryotes may serve as PAMPs and/or may bind and work in conjunction with other PAMPs (Bulut,

Faure et al. 2002; Bulut, Shimada et al. 2009; Tsan and Gao 2009). These PAMPs are then recognized by pattern recognition receptors (PRRs), including Toll-like receptors

111

(TLRs) (Tizard 2008). The adaptor molecule, MyD88 then plays an important role for most TLRs in the translation of PAMP recognition into a cytokine response (McGettrick and O'Neill). Both N. sennetsu and N. risticii lack LPS and flagellin, and cannot synthesize peptidoglycan (Hotopp, Lin et al. 2006; Lin, Zhang et al. 2009), suggesting that other proteins are involved in host cytokine/chemokine production, resulting in disease symptoms and clinical signs.

Bacterial research has also demonstrated roles of porins in MyD88-dependent cytokine/chemokine production, including Haemophilus influenzae type b (Hib)

(Galdiero, D'Amico et al. 2001), Salmonella (Galdiero, de L'ero G et al. 1993), and

Shigella dysenteriae (Biswas, Banerjee et al. 2007; Biswas, Banerjee et al. 2009).

Cytokine/chemokine activation by porins also appears to occur mainly through recognition of the porin by TLR2 and a co-receptor, such as CD14 in Hib (Galdiero,

Finamore et al. 2004), CD11a/CD18 in Salmonella (Galdiero, Vitiello et al. 2002), and

TLR6 in S. dysenteriae (Biswas, Banerjee et al. 2007). Given that P51 (a major surface protein of N. sennetsu) is a porin (Gibson, Kumagai et al. 2010), there is a strong possibility that P51 is a PAMP and plays a role in cytokine/chemokine production during

N. sennetsu infection and disease.

In accordance with the objectives for Chapter 5, the BALB/c disease model for Sennetsu neorickettsiosis was developed and utilized to determine cytokine/chemokine expression in vivo. It was hypothesized, given the prominence of P51 in the outer membrane fraction and its identification as a porin, that P51 is a major cause of cytokine/chemokine production and, given the porin qualities of P51 (Gibson, Kumagai et al. 2010), P51 is

112

recognized by TLR2 and that cytokine/chemokine production is MyD88 dependent. In vitro studies using splenocytes and bone marrow-derived macrophages (BMDMs) from wild-type and MyD88 and TLR2 knockout mice were incubated with whole bacteria, the outer membrane fraction of N. sennetsu, and recombinant proteins for P51, Nsp2 and

Nsp3. In addition, a recombinant Nsp immunization study is described with the hypothesis that antibodies to rNsp2 and/or rNsp3 reduce bacterial load and/or clinical signs of disease in the mouse. The work described here determined N. sennetsu mouse pathogenesis and roles of cytokines and will be valuable in understanding and treating symptoms of Sennetsu neorickettsiosis.

Experimental procedures:

In vitro culture of N. sennetsu in P388D1 cells. N. sennetsu Miyayama (Misao and

Kobayashi 1954) was grown in P388D1 cells (non-synchronized culture) in RPMI

(Mediatech) supplemented with 5-10% FBS (U.S. Biotechnologies) and 4-6 mM of L- glutamine (Invitrogen) at 37°C under 5% CO2 – 95% air atmosphere. Infectivity was monitored by DiffQuik staining (Dade Bahring, Deerfield, IL), and cells were suspended in RNAlater (Invitrogen) during low (5%) and high (>80%) infectivity until RNA extraction was performed.

Inoculation of ICR mice with N. sennetsu. All mice were housed and treated according to the IACUC protocol 2008A0066 and IACUC rules and regulations. The first mouse strain utilized in developing a mouse model of disease was the ICR strain (Charles River

Laboratories). Four-week old female mice of the ICR strain were inoculated via

113

intraperitoneal (IP) route with 1 × 102, 1 × 104, or 1 × 106 N. sennetsu Miyayama highly- infected (> 95%, non-synchronized culture) P388D1 cells suspended in 0.1 ml of RPMI

6 (n = 5). Negative control mice were IP inoculated with 1 × 10 uninfected P388D1 cells

(n = 5). Clinical signs were monitored twice daily, and body weights were checked once daily, and blood was taken from the submandibular sinus every 5 days. With no clinical signs observed and no significant weight loss documented, all mice were euthanized 15 days post infection by CO2 inhalation, followed by cardiac exsanguination. Blood was collected, and spleens were weighed and preserved in 10% neutral-buffered formalin

(NBF).

Inoculation of CF-1 outbred mice with N. sennetsu. It was then planned to test multiple strains of outbred and inbred mice, according to previous results obtained for N. risticii infection and disease (Williams and Timoney 1994). Pilot study candidates included CF-

1 (outbred) and BALB/c (inbred) mice. Four-week-old female mice of the CF-1 strain

(Charles River Laboratories, Spencerville, OH) were IP inoculated with 1 × 106 N. sennetsu Miyayama highly-infected (> 95%, non-synchronized culture) P388D1 cells suspended in 0.2 ml of RPMI (n = 3). Negative control mice were IP inoculated with 1 ×

6 10 uninfected P388D1 cells (n = 3). Clinical signs were monitored twice daily, and body weights were checked once daily, and blood was taken from the submandibular sinus every 5 days. With only mild clinical signs observed and no significant weight loss documented, all mice were euthanized 17 days post inoculation by CO2 inhalation, followed by cardiac exsanguination. Blood was collected, spleens were weighed, samples

114

of spleen and liver were saved in RNAlater, and samples of spleen and liver were preserved in 10% NBF.

Inoculation of BALB/c mice with N. sennetsu. Four-week-old female mice of the

BALB/c strain (Charles River Laboratories, Spencerville, OH) were IP inoculated with 1

6 × 10 N. sennetsu Miyayama highly-infected (> 95%, non-synchronized culture) P388D1 cells suspended in 0.2 ml of RPMI (n = 3). Negative control mice were IP inoculated

6 with 1 × 10 uninfected P388D1 cells (n = 3). Clinical signs were monitored twice daily, and body weights were checked once daily, and blood was taken from the submandibular sinus every 5 days. Due to progressive weight loss and clinical signs approaching early removal criteria, all mice were euthanized nine days post infection by CO2 inhalation, followed by cardiac exsanguination. Blood was collected; spleens were weighed; samples of spleen and liver were saved in RNAlater; and samples of spleen and liver were preserved in 10% NBF.

A dose-dependent pilot study was also performed, using eight-week-old female mice of the BALB/c strain (Charles River Laboratories). The mice were IP inoculated with 1 ×

104 or 1 × 106 N. sennetsu Miyayama highly-infected (> 95%, non-synchronized culture)

P388D1 cells suspended in 0.2 ml of RPMI (n = 3). Negative control mice were IP

4 6 inoculated with 1 × 10 or 1 × 10 uninfected P388D1 cells (n = 3). Clinical signs were monitored twice daily, and body weights were checked once daily, and blood was taken from the submandibular sinus every 5 days. Due to progressive weight loss and clinical signs approaching early removal criteria in 1 × 106 N. sennetsu-infected mice, these mice and their respective controls were euthanized 9 days post infection by CO2 inhalation,

115

followed by cardiac exsanguination. Clinical signs in 1 × 104 infected mice waxed and waned, and did not meet early removal criteria, and the mice and their respective controls were euthanized 23 days post infection. Blood was collected; spleens and livers were weighed; samples of spleen and liver were saved in RNAlater; and samples of spleen, liver, thymus, kidney, and the gastrointestinal tract were preserved in 10% NBF.

Immunization study using rNsp2 and rNsp3 inoculation of BALB/c mice and subsequent

N. sennetsu challenge. Six-week-old mice of the BALB/c strain were inoculated with rNsp2, rNsp3, or PBS, as described in Chapter 3. After three immunizations (day 49), mice were IP inoculated with 1 × 106 N. sennetsu Miyayama highly-infected (> 95%, non-synchronized culture) P388D1 cells suspended in 0.2 ml of RPMI (n = 5). Clinical signs were monitored twice daily, body weights were checked once daily, and blood was taken from the submandibular sinus every 5 days. Due to progressive weight loss and clinical signs approaching early removal criteria, all mice were euthanized nine days post infection by CO2 inhalation, followed by cardiac exsanguination. Blood was collected; spleens and livers were weighed; samples of spleen and liver were saved in RNAlater; and samples of spleen, liver, thymus, kidney, and the gastrointestinal tract were preserved in 10% NBF.

Host cell-free N. sennetsu. N. sennetsu were released from P388D1 host cells by sonication for 30 s at setting 2 using a W-380 ultrasonic processor (Heat Systems). Cell fragments were pelleted by centrifugation at 448 × g for 5 min, the supernatant was filtered through a 5 µm-pore-size nylon microfiber syringe filter (Whatman), and the

116

bacteria were pelleted by centrifugation at 10,000 × g for 10 min and either used immediately or stored at −80°C.

In vitro neutralization with α-rP51, α-rNsp2, and α-rNsp3. Rabbit rP51 antisera and mouse rNsp2 and rNsp3 antisera were used in preliminary in vitro infection and binding

5 studies. For infection studies: 4.5 × 10 P388D1 cells/well were incubated overnight in a

24-well plate in RPMI containing 5% FBS and 4-6 mM L-glutamine. For rP51 experiments, N. sennetsu filtered bacteria were incubated in preimmune rabbit sera or α-

6 rP51 (bacteria from 2 × 10 infected P388D1 cells in 28 µl centrifuged sera) at room temperature for 1 h with 20 rotations per minute. For rNsp experiments, N. sennetsu filtered bacteria were incubated in preimmune mouse sera, α-rNsp2, or α-rNsp3 (bacteria

6 from 3 × 10 infected P388D1 cells in 40 µl centrifuged sera) at room temperature for 1 h with 20 rotations per minute. The bacteria and sera were added to the uninfected P388D1 cells at a concentration of 5.5 × 106 cells worth of N. sennetsu per well. The media in each well was changed 1 day post inoculation (p.i.), and the P388D1 cells were incubated for a total of nine days. DNA was extracted from the P388D1 cells, using a DNeasy minikit (Qiagen) and subjected to qPCR, using gapdh and 16S rRNA primers (Table 13).

Quantitative polymerase chain reaction (qPCR). DNA was extracted from buffy coats spleens, livers, and duodenums of mice using a DNeasy minikit (Qiagen), according to manufacturer’s instructions. Quantitative PCR used a Stratagene Mx3005P real-time thermal cycler (Foster City, CA), Maxima SYBR Green/ROX qPCR Master Mix

(Fermentas) and primers designed to amplify N. sennetsu 16S rRNA and mouse gapdh

117

(Table 13). Expression of N. sennetsu 16S rDNA was normalized against gapdh DNA levels.

N. sennetsu outer membrane isolation. N. sennetsu bacteria were resuspended at 200 µg per 25 µl of 10 mM sodium phosphate buffer, pH 7.4 (NaPi). One hundred µg/ml DNase

(Sigma), 5 mM MgCl2, and 2% Sarkosyl (Sigma) in NaPi equal to the volume of NaPi used to resuspend the bacteria were then added and mixed gently by pipetting and inversion, and the mixture was incubated at 37°C for 30 min. Five mM of EDTA were added, the tube was then centrifuged for 1 h at 10,000 × g at 15°C, and the resulting pellet was washed at least twice with 1 ml NaPi and twice with 1 ml RPMI at 10,000 × g at 15°C for 10 min. The outer membrane fraction was used immediately afterwards or was stored for up to 3 h at −80°C.

Preparation of splenocytes and bone marrow-derived macrophages from wild-type

(C57/BL6) and MyD88 and TLR2 knockout mice. Whole spleens from euthanized

MyD88-/-, TLR2-/-, and corresponding C57/BL6 wild-type mice (all original stock mice were generously provided by Dr. Xin Li from the Massachusetts General Hospital) were removed under a laminar hood, placed in sterile RPMI, and teased apart with 20-Ga needles. After centrifugation at 448 × g for 5 min, the pellet was vortexed for 15 s, and erythrocytes were lysed by resuspending the pellet in RBC lysis buffer (0.83% NH4Cl,

0.1% KHCO3, and 0.004% Na2EDTA; pH 7.3) for 2 min. Three ml of FBS were added slowly, and the splenocytes were centrifuged and vortexed as above. The splenocytes were then resuspended in 5 ml RPMI containing 5% FBS and 4-6 mM of L-glutamine, the cells were assessed for viability by Trypan blue staining (greater than 99% live

118

required), the volume was adjusted to 107 cells/ml, and 100 µl (1 × 106 cells) were added to a 96-well plate. BMDMs were aspirated, using sterile RPMI and 30-Ga needles from femurs and tibias and centrifuged at 448 × g for 5 min. Lysis of erythrocytes was performed using 3 ml RBC lysis buffer for 2 min, followed by the addition of 47 ml

RPMI and centrifugation. The cells were then suspended at 107 cells in 17 ml RPMI and

2 ml of colony-stimulating factor-1 derived from L-929 cells (Lin and Stewart 1973),

2 added to a 75-cm flask, and incubated for 7-10 days at 37°C under 5% CO2 – 95% air atmosphere.

Inoculation of splenocytes and BMDMs with N. sennetsu whole bacteria, bacterial outer membrane fraction, and recombinant proteins. For splenocyte inoculation: N. sennetsu isolated bacteria, outer membrane fraction, and recombinant proteins (His-tag-purified rNsp2 and rNsp3 treated with at least 25× column volume washes with 0.1% Triton X-

114 to reduce endotoxin levels and dialyzed against and washed with PBS) were suspended in RPMI containing 5% FBS and 4-6 mM of L-glutamine, and 100 µl were added to each well (200 µl/well total). Splenocytes were disrupted and homogenized by rotary mortar and pellet pestle (Kimble Chase, Vineland, NJ) in Buffer RLT (Qiagen) containing β-mercaptoethanol (Sigma) 2 and/or 8 h p.i. and stored at −80°C until RNA extraction was performed.

For BMDM inoculation: BMDMs were gently released from the 75-cm2 flask using

Cellstripper buffer (Mediatech) and cell scrapers and plated in 24 or 48-well plated overnight in RPMI containing 10% FBS and 4-6 mM of L-glutamine. The media was removed and replace with media containing appropriate dilutions of N. sennetsu isolated

119

bacteria, outer membrane fraction, and recombinant proteins (His-tag-purified rP51 and rNsp2 treated with at least 25× column volume washes with 0.1% Triton X-114).

BMDMs were disrupted and homogenized by rotary mortar and pellet pestle (Kimble

Chase, Vineland, NJ) in Buffer RLT (Qiagen) containing β-mercaptoethanol (Sigma) 2 h p.i. and stored at −80°C until RNA extraction was performed. Negative controls for splenocytes and BMDMs included plain RPMI, PBS treated with the Sarkosyl outer membrane purification protocol, and filtered (5 µm) uninfected P388D1 lysate. LPS from

Escherichia coli 0111:B4 (Sigma) was used for positive controls.

Reverse transcription-polymerase chain reaction and quantitative reverse transcription- polymerase chain reaction (RT-PCR and qRT-PCR). RNA was extracted from spleens, splenocytes, and BMDMs using an RNeasy minikit (Qiagen), and first-strand complementary DNAs (cDNAs) were synthesized using oligo-dT primers and reverse transcriptase (Fermentas, Glen Burnie, MD), according to manufacturer’s instructions.

For N. sennetsu protein mRNA qRT-PCR, RNA was first treated with DNase

(Invitrogen), and cDNAs were synthesized using random primers and SuperScript III

(Invitrogen), according to manufacturer’s instructions. RT-PCR was performed on mouse spleens using primers to amplify N. sennetsu p51, nsp1, nsp2, nsp3, and groEL and mouse gapdh (Table 13), using 40 cycles and a 60°C annealing temperature, except for nsp1 and nsp2 amplifications, which employed a ten-cycle touchdown (Hecker and Roux

1996) from 65-56°C before using 40 cycles and a 55°C annealing temperature.

Quantitative PCR on cDNA samples used a Stratagene Mx3005P real-time thermal cycler

(Foster City, CA), Maxima SYBR Green/ROX qPCR Master Mix (Fermentas),

120

cytokine/chemokine primers designed (according to exon and intron locations) to prevent

DNA contamination (Miura and Rikihisa 2009), and primers designed to amplify N. sennetsu p51, nsp1, nsp2, nsp3, and groEL (Table 13).

121

Gene Primer name 5′→3′ sequence ER5-3 ATTTGAGAGTTTGATCCTGG ER3-2 GTTTTAAATGCAGTTCTTGG 16S E.ris1 GGAATCAGGGCTGCTTGCAGCCT rRNAa E.ris2 TGTGGGTACCGTCATTATCTTCCCCACT NS16S-q1F CGTAACGCGTGGGAACTTG NS16S-q1R AGCTAATCTAACGCGGGCCTATCTGA NSnsp1-q1F AGAATGCCTATATGGACGAACTG nsp1 NSnsp1-q1R TAGGTATACATATCCCCCATTAG NSnsp2-q1F TGACTCCCATACAACCAGACCAG nsp2 NSnsp2-q1R TGTACCGTTTATCAGGGCTTCAA NSnsp3-q1F GAGTATACGGGCGCAAAAGCAGACG NSnsp3-q1R CATTCATAGGCACCATCACACCAACC nsp3a NSE_0875 CR1F TGACTCCCATACAACCAGACCAG NSE_0875 CR4R TGTACCGTTTATCAGGGCTTCAA NSp51-q1F TGGTACGCATTATCGCATCACTGGAAGC p51 NSp51-q1R TCACACCGGAAGCAAGAACAAAAC NSgroEL-q1F TATGGTGGCCCGGAAGTAAC groEL NSgroEL-q1R ACCAAGATAGTAGCAGTAGTAGTCCCATC gapdh-q1F GAATACGGCTACAGCAACAG gapdh-q1R TGGAAATTGTGAGGGAGATG gapdhb KM0179 GGCATTGCTCTCAATGACAA KM0180 TGTGAGGGAGATGCTCAGTG KM0179 GGCATTGCTCTCAATGACAA gapdh KM0180 TGTGAGGGAGATGCTCAGTG KM0195 GGGCCTCAAAGGAAAGAATC IL-1β KM0196 TACCAGTTGGGGAACTCTGC KM0207 CTCTCAAGGGCGGTCAAAAAGTT CXCL2 KM0208 TCAGACAGCGAGGCACATCAGGTA KM0185 CATCTTCTCAAAATTCGAGTGACAA TNFα KM0186 TGGGAGTAGACAAGGTACAACCC KM0191 GCAACGGGAAGATTCTGAAG IL-1α KM0192 TGACAAACTTCTGCCTGACG KM0205 CGCTGCTGCTGCTGGCCACCA CXCL1 KM0206 CTATGACTTCGGTTTGGGTGCAG KM0288 CATCGATGAGCTGATGCAGT IL-12A KM0289 CAGATAGCCCATCACCCTGT KM0181 GCGTCATTGAATCACACCTG IFNγ KM0182 TGAGCTCATTGAATGCTTGG KM0256 ATTGCTCCCTTCCGAAGTTT iNOS KM0257 TGCAGGATGTCCTGAACGTA KM0201 CATGGGTCTTGGGAAGAGAA IL-10 KM0202 AACTGGCCACAGTTTTCAGG Table 13. Primer pairs for RT-PCR, qPCR, and qRT-PCR. The first half of the table

(above the double line) utilized DNA and cDNAs synthesized with random primers for N. sennetsu gene DNA and mRNA amplification. The second half (below the double line) uses cDNAs synthesized with oligo-dT primers for cytokine/chemokine mRNA amplification. Primer pair KM0179 and KM0180 and all primers in the second half were created by Dr. Koshiro Miura, and most are listed previously (Miura and Rikihisa 2009).

122

a The 16S rRNA primers ER5-3, ER3-2 (for first-step PCR), E.ris1, and E.ris2 (for nested

PCR) and nsp3 primers NSE_0875 CR1 F and NSE_0875 CR4R were used for RT-PCR only. b The gapdh primers gapdh-q1F and gapdh-q1R were used for both qPCR and qRT-PCR.

The primers KM0179 and KM0180 were used for RT-PCR on random primer-amplified cDNAs.

Statistical analysis. Analysis of variance (ANOVA) and the unpaired Student’s t-test were applied to determine the differences among weights and qRT-PCR results. A P value of less than 0.05 was considered significant.

Results:

Sennetsu neorickettsiosis in the mouse model. ICR mice did not demonstrate overt clinical signs of Sennetsu neorickettsiosis, and there was no weight loss indicative of disease (Figure 25A). However, all N. sennetsu-inoculated ICR mice demonstrated a statistically-significant increase in splenic weight in relation to the control mice (Figure

25B), and all mice had N. sennetsu PCR-positive spleens. No adverse clinical signs were noted in the control mice. It is possible that the ICR mouse strain could be used in a future model of subclinical infection.

123

Figure 25. ICR mouse body and spleen weights. ICR mouse body weights from 0-15 days post-infection (A) and spleen weights, given as a percentage of body weight 15 days post infection (B). Horizontal bar, average splenic weight (n = 5). **, P < 0.01 by ANOVA between N. sennetsu-infected samples and control samples. There is no statistical difference among N. sennetsu-infected splenic weights.

124

CF-1 mice demonstrated mild clinical signs, including ruffled fur. No adverse clinical signs were noted in the control mice. This mouse strain could be a future model of subclinical infection, potentially being more representative of the mild disease state seen

6 in humans. All BALB/c mice inoculated with 1 × 10 N. sennetsu-infected P388D1 cells began to demonstrate clinical signs on day 6 post infection (ruffled fur and mild weight loss). Until euthanasia nine days post infection, these clinical signs progressed to include squinting, hunched appearance, weakness, depression, lethargy, dehydration, inappetence, and significant weight loss (Figure 26). BALB/c mice inoculated with 1 ×

4 10 N. sennetsu-infected P388D1 cells demonstrated only mild waxing and waning clinical signs (ruffled fur, Figure 27). This lower dosage could be used to create a future model of subclinical infection, potentially being more representative of the mild disease state seen in humans.

125

Figure 26. Body weights, splenic weights, and clinical signs for four-week-old CF-1 and

BALB/c mice. (A) Mouse body weights from 0-17 (CF-1) and 0-9 (BALB/c) days post- infection. (B) Spleen weights, given as a percentage of body weight 17 (CF-1) and nine

(BALB/c) days post infection. Horizontal bar, average splenic weight (n = 3). *, P <

126

0.05; **, P < 0.01 by Student’s t-test. (C) Clinical signs for CF-1 and BALB/c mice, according to appearance (1, normal; 2, ruffled fur; 3, ruffled fur + squinting + hunched; 4, ruffled fur + squinting + hunched + weak) and activity(1, normal; 2, less active; 3, depressed and lethargic; 4, moribund) (Williams and Timoney 1994).

127

Figure 27. Body weights, splenic weights, and clinical signs for eight-week-old BALB/c

4 6 mice inoculated with 1 × 10 or 1 × 10 N. sennetsu infected or control P388D1 cells. (A)

Mouse body weights from 0-23 (1 × 104) and 0-9 (1 × 106) days post-infection. (B)

Spleen weights, given as a percentage of body weight 23 (1 × 104) and nine (1 × 106) 128

days post infection. Horizontal bar, average splenic weight (n = 3). (C) Liver weights, given as a percentage of body weight 23 (1 × 104) and nine (1 × 106) days post infection.

Horizontal bar, average hepatic weight (n = 3). *, P < 0.05; **, P < 0.01 by Student’s t- test. (D) Clinical signs for CF-1 and BALB/c mice, according to appearance (1, normal;

2, ruffled fur; 3, ruffled fur + squinting + hunched; 4, ruffled fur + squinting + hunched + weak) and activity(1, normal; 2, less active; 3, depressed and lethargic; 4, moribund)

(Williams and Timoney 1994).

On necropsy of N. sennetsu-infected BALB/c mice, significant findings included splenomegaly (Figure 26 and 27), hepatomegaly (Figure 27), pallor in the kidneys and periphery of liver lobes, reduced abdominal fat, dialated intestinal loops, and fecal staining around the anus suggestive of diarrhea. Microscopy of hematoxylin and eosin- stained tissue samples revealed increased red pulp and reduced organization in the spleen, perivascular cuffing in the liver and kidney, severe depletion of lymphocytes in the thymus, and potential Neorickettsial organisms in mononuclear cells on blood films. No adverse clinical signs were noted in the control mice (Figure 28).

129

Figure 28. Pathologic changes in N. sennetsu-infected BALB/c mice. Hematoxylin- and

6 eosin-stained liver (A), spleen (B), thymus (C), and kidney (D) of 1 × 10 P388D1 group 130

N. sennetsu-infected and control BALB/c mice nine days post infection. Liver and spleen are from four-week-old mice. Thymus and kidney are from eight-week-old mice. Note perivascular infiltration in the liver and kidney (arrows), increase in red pulp and lack of organization in the spleen, and severe thymic lymphoid depletion in the infected mice.

Bars, 100 μm for panels (A) through (D). (E) Potential N. sennetsu organisms (red arrow) in a mononuclear cell in a blood smear stained by DiffQuik from a four-week-old N.

6 sennetsu-infected mouse (1 × 10 N. sennetsu-infected P388D1 cell inoculum group) nine days post infection. Bar, 10 µm.

Expression of N. sennetsu surface proteins in P388D1 cells and a Sennetsu disease mouse model. To examine expression of these surface-exposed proteins in vitro and in vivo, the murine leukemia cell line P388D1 and BALB/c mice were experimentally infected with

N. sennetsu. Samples of N. sennetsu-infected P388D1 cells were collected during low infectivity (approximately 5% of the cells were infected) and during heavy infection

(>80% of the cells were infected, and the cells were starting to rupture).

RT-PCR preliminary results in the BALB/c mouse spleen demonstrated the mRNA expression of all four major surface protein of N. sennetsu: Nsp2, Nsp3, P51, and GroEL.

The nsp1 expression was not detected (Table 14 and Figure 29).

131

Genea Primer name Target size (bp) # Expressionb ER5-3c 16S ER3-2 6/6 (I) 382 rRNA E.ris1 0/6 (C) E.ris2 NSnsp1-q1F 0/6 (I) nsp1 121 NSnsp1-q1R 0/6 (C) NSnsp2-q1F 6/6 (I) nsp2 157 NSnsp2-q1R 0/6 (C) NSE_0875 CR1F 6/6 (I) nsp3 486 NSE_0875 CR4R 0/6 (C) NSp51-q1F 6/6 (I) p51 134 NSp51-q1R 0/6 (C) NSgroEL-q1F 6/6 (I) groEL 158 NSgroEL-q1R 0/6 (C) KM0179d 6/6 (I) gapdh 202 KM0180 6/6 (C) Table 14. Expression analysis of N. sennetsu surface proteins in four and eight-week-old

BALB/c mouse spleens. a All genes, except gapdh are N. sennetsu genes; gapdh is a Mus musculus gene. The genes 16S rRNA and gapdh are used as controls for the presence/absence of N. sennetsu and mouse RNA, respectively. b Demonstrates the number of mice RT-PCR positive out of the total mice for the given gene fragment from N. sennetsu-infected (I) and control (C) groups. c Primers for the amplification of the 16S rRNA gene fragment are used in first-step

(ER5-3 and ER3-2) and nested (E.ris1 and E.ris2) PCR as previously described (Kanter,

Mott et al. 2000). d Primers for the amplification of the mouse gapdh gene fragment were designed previously (Miura and Rikihisa 2009).

132

Figure 29. RT-PCR results for four-week-old N. sennetsu infected and control BALB/c

6 mice. Refer to Table 14 above. Infected, 1 × 10 N. sennetsu-infected P388D1 cell

6 inoculum group; n = 3. Control, 1× 10 uninfected P388D1 cell inoculums group; n = 3. dH2O, autoclaved distilled, deionized, and filtered water negative control; NS DNA, N. sennetsu DNA positive control; RT−, without reverse transcriptase; RT+, with reverse transcriptase. Results for eight-week-old BALB/c mice were similar.

Quantitative RT-PCR results demonstrated levels of N. sennetsu protein expression in vitro and in vivo, using N. sennetsu DNA as standards and normalized against 16S rRNA levels. Relative mRNA levels among in vitro samples demonstrated that nsp3 had the highest expression, followed by groEL and p51, and lastly nsp2 with the lowest levels.

Relative mRNA levels among mouse liver and spleen samples were similar to each other yet different from in vitro results with p51 expression being higher in vivo than observed in vitro. The levels of nsp2 in the mice were similar to the in vitro results in that they

133

were noticeably lower than the other genes. Expression of nsp1 was again not detected

(Figure 30).

Figure 30. Quantitative RT-PCR results for N. sennetsu-infected P388D1 cells with low infectivity (5%, Low inf.) and high infectivity (>80%, High inf.) cells and four-week-old

BALB/c mouse livers and spleens (n = 3). Standards for qRT-PCR were dilutions of

DNA from N. sennetsu-infected P388D1 cells, and samples were normalized against 16S rRNA levels.

Immunogenicity of rNsp2 and rNsp3. Western blotting against rNsp2 and rNsp3 showed that sera from all three mice inoculated with 1  104 N. sennetsu-infected cells 23 days post infection selectively recognized rNsp2, but not rNsp3 (Figure 31). Similarity of

134

protein loading amounts of rNsp2 and rNsp3 was confirmed by Western blot analysis using anti-His staining (see Figure 10).

Figure 31. Reactivity of N. sennetsu-infected mouse sera against rNsp2 and rNsp3.

Western blotting was performed using sera collected 23 days post infection from 1 × 104

N. sennetsu infected and control group eight-week-old BALB/c mice to probe rNsp2 and rNsp3 (n = 3).

Recombinant Nsps immunization study. With the high levels of Nsp3 mRNA during

BALB/c infection and the strong recognition of rNsp2 by BALB/c N. sennetsu antisera, an immunization study was performed to determine the effectiveness of rNsp2 and rNsp3 against Sennetsu neorickettsiosis. Mice were immunized three times with recombinant proteins, and antisera from rNsp2 and rNsp3 groups demonstrated strong recognition of their respective recombinant proteins with no crossreaction (see Chapter 3, Figure 10C and D). All mice began to demonstrate clinical signs six days post N. sennetsu infection.

Clinical signs appeared to be somewhat less severe in rNsp2 and rNsp3-inoculated mice, although body weights, liver weights, and bacterial loads did not support these observations. There was a statistically-significant decrease in splenic weight in the 135

rNsp2-inoculated mice in relation to the control mice (Figure 32). Of note, N. sennetsu

DNA levels were also checked in buffy coat, spleen, and intestine of control mice. N. sennetsu levels were highest in the buffy coat, followed by spleen, and minute quantities were found in the duodenal section of the intestine (Figure 32F). Although it is possible that N. sennetsu is infecting intestinal epithelium, as is seen in N. risticii (Rikihisa, Perry et al. 1984; Chae, Kim et al. 2002), it is also possible that N. sennetsu from blood within the duodenal segment was amplified. Future experiments, such as in situ hybridization, are required to determine the range of N. sennetsu host cells.

136

Figure 32. Recombinant Nsp mouse immunization experiment N. sennetsu challenge. (A)

Mouse body weights from 0-9 days post-infection. (B) Spleen weights, given as a percentage of body weight nine days post infection. *, P < 0.05 by Student’s t-test.

Horizontal bar, average splenic weight. (C) Liver weights, given as a percentage of body weight nine days post infection. (D) Bacterial loads in the buffy coats of control and

137

rNsp2 and rNsp3-immunized mice. (E) Bacterial loads in the spleens of control and rNsp3-immunized mice. (F) Bacterial loads in the buffy coat, spleens and livers of control mice nine days post infection. (D), (E), and (F) are 16S rDNA levels normalized against gapdh DNA levels and values shown are means ± standard deviations for five mice.

In vitro neutralization with rP51, rNsp2, and rNsp3 antisera. Mouse antisera from rP51- inoculated rabbits and rNsp2 and rNsp3-inoculated mice were used to incubate isolated

N. sennetsu before infecting P388D1 cultures. Nine days post infection, there was no significant change in bacterial number in P388D1 infected with rP51-incubated bacteria

(2.7% average neutralization) a statically-significant decrease in the quantity of N. sennetsu DNA in P388D1 infected with rNsp3-incubated bacteria (76% average neutralization) and a statistically-significant increase in P388D1 infected with rNsp2- incubated bacteria (−71% average neutralization). These findings do not reflect the antibody titers in the antisera, and future studies will be required to determine the effectiveness of neutralization in vitro (Figure 33).

138

Figure 33. In vitro neutralization with α-rP51, α-rNsp2, and α-rNsp3. The 16S rDNA levels are relative to 16S rDNA levels from the respective preimmune sera (rabbit sera

[Ctl Rabbit] for α-rP51 and mouse sera [Ctl Mouse] for α-rNsp2 and α-rNsp3) nine days post infection. Values shown are means ± standard deviations for three wells. *, P <

0.05; **, P < 0.01 by Student’s t-test between antisera and respective preimmune levels.

Cytokine expression in vivo in BALB/c model of Sennetsu neorickettsiosis. Cytokine expression was tested in the spleens of four-week-old 1 × 106 N. sennetsu-infected

P388D1 inoculated and control BALB/c mice described above. It was found that mRNA levels of IL-1β, CXCL2, TNFα, IFNγ, and iNOS were significantly higher in N. sennetsu-infected mice than in control mice (Figure 34 and Table 15). These data indicate that these proinflammatory cytokines and chemokines, as well as iNOS play important roles in Sennetsu neorickettsiosis.

139

Figure 34. In vivo cytokine mRNA expression in BALB/c mouse spleens during Sennetsu neorickettsiosis. All cytokine transcript levels are relative to respective cytokine

6 transcript levels from control mice inoculated with 1 × 10 uninfected P388D1. Values shown are means ± standard deviations for three mice. C, mice inoculated with 1 × 106

6 uninfected P388D1. NS, mice inoculated with 1 × 10 N. sennetsu-infected P388D1. *, P

< 0.05; **, P < 0.01 by Student’s t-test.

140

Mouse sample Inoculation Gene Means Std dev IL-1β 1.66 ±6.03 × 10−1 CXCL2 3.30 × 102 ±183 1 × 106 heavily N. sennetsu- TNFα 1.01 ±2.30 × 10−1 infected P388D cells/mouse IFNγ 1.85 ±4.98 × 10−1 Female BALB/c 1 iNOS 8.79 × 10−1 ±2.93 × 10−1 Four-weeks-old IL-10 1.31 ±2.13 × 10−1 Spleens IL-1β 2.29 × 10−1 ±1.97 × 10−1 (n = 3) CXCL2 4.03 × 10−1 ±3.53 × 10−1 Performed 09/2008 1 × 106 uninfected TNFα 5.28 × 10−1 ±1.18 × 10−1 −2 −2 P388D1 cells/mouse IFNγ 3.22 × 10 ±2.72 × 10 iNOS 9.79 × 10−2 ±8.13 × 10−2 IL-10 1.12 ±7.76 × 10−2 Table 15. Means for qRT-PCR results for cytokine induction studies in spleens in the

BALB/c mouse model of disease. Std dev, standard deviation. All qRT-PCR results are nine days post infection. Quantitative RT-PCR results are normalized by gapdh transcript levels.

In vitro cytokine expression in splenocytes. In order to create an in vitro model of cytokine expression demonstrated in the mouse model, splenocytes were isolated from

C57/BL6 mice (in anticipation of utilizing knockout mice). When N. sennetsu isolated bacteria were freshly isolated from P388D1 host cells and incubated for eight hours with splenocytes, there were significant increases in IL-1β (4× average increase), CXCL2

(101×), IFNγ (144×), and iNOS (18×) mRNA levels in relation to P388D1 lysate control mRNA levels. When frozen bacteria and Sarkosyl-derived bacterial outer membrane fraction were used to incubate splenocytes for 8 h, IL-1β and CXCL2 levels were both markedly increased. However, levels of both IFNγ and iNOS were drastically reduced; in particular the IFNγ levels in whole frozen bacteria were significantly less than the control, the IFNγ levels in the outer membrane fraction (2× the average PBS control

141

levels) were reduced in relation to the freshly-isolated bacteria levels (144× the average

P388D1 lysate control levels), and the iNOS levels were reduced in outer membrane fraction (5× less than N. sennetsu whole frozen bacteria). This suggests that IFNγ and iNOS expression is induced by some component of the bacterium other than the outer membrane fraction (potentially a soluble protein), and possibly requires living bacteria

(Figure 35, Table 16).

142

Figure 35. Cytokine production within splenocytes incubated fresh and previously-frozen

N. sennetsu and N. sennetsu outer membrane fraction. All splenocyte cytokine transcript

143

levels are at 8 h p.i. and use 1 × 106 splenocytes/well. (A) Bacteria were freshly isolated from infected P388D1 cells. All cytokine transcript levels are relative to respective cytokine transcript levels from splenocytes incubated with P388D1 lysate. C, splenocytes

6 incubated with filtered lysate from 7 × 10 P388D1 cells per well. NS, splenocytes

6 incubated with N. sennetsu isolated from 7 × 10 N. sennetsu heavily-infected P388D1 cells. LPS, splenocytes incubated with 50 µg/ml LPS from E. coli 0111:B4. (B) Bacteria were previously frozen at −80°C. All cytokine transcript levels are relative to respective cytokine transcript levels from splenocytes incubated with PBS treated with the Sarkosyl fractionation protocol. C, splenocytes incubated with PBS treated with the Sarkosyl fractionation protocol. NS, 750 µg/ml N. sennetsu whole bacteria. OM, 250 µg/ml N. sennetsu outer membrane fraction (derived from 750 µg/ml N. sennetsu whole bacteria).

LPS, splenocytes incubated with 500 ng/ml LPS from E. coli 0111:B4. Values shown are means ± standard deviations for three wells. NS, 1 × 106 splenocytes inoculated with 50

µg/ml N. sennetsu bacteria. *, P < 0.05; **, P < 0.01 by the Student’s t-test between control and experimental groups.

Cytokine expression is MyD88 dependent. When BMDMs from MyD88 knockout mice were inoculated with N. sennetsu-isolated bacteria, there was a decrease in mRNA levels for IL-1β (5.4× reduction, in relation to wild-type levels), CXCL2 (9.5×), TNFα (4.4×), and IL-12A (p35) (15×, Figure 36, Table 17).

144

Figure 36. MyD88-dependent cytokine production by N. sennetsu in BMDMs. All cytokine transcript levels are 2 h p.i. and are relative to respective wild-type or knockout

6 cytokine transcript levels from BMDMs inoculated with 7 × 10 uninfected P388D1 filtered lysate. Values shown are means ± standard deviations for two wells. C, 4.5 × 105

6 BMDMs inoculated with 7 × 10 uninfected P388D1 filtered lysate per well. NS, 4.5 ×

5 6 10 BMDMs inoculated with bacteria from 7 × 10 N. sennetsu heavily-infected P388D1 per well. WT, C57J/BL6 wild-type mice. KO, MyD88 knockout mice.

Cytokine expression is TLR2 dependent, and TLR2-dependent cytokine expression is triggered by the outer membrane fraction of N. sennetsu. When splenocytes and BMDMs from TLR2 knockout mice were incubated with N. sennetsu isolated bacteria and/or 145

Sarkosyl-purified N. sennetsu outer membrane fraction, there were statistically- significant decreases in mRNA levels of IL-1β, CXCL2, and IL-12A in relation to wild- type mRNA levels. IL-1β and CXCL2 levels were reduced by 11× and 242×, respectively for whole bacteria inoculation of TLR2 knockout splenocytes. In BMDMs, IL-1β average levels were reduced by 25× for whole bacteria and reduced by 219× for the outer membrane fraction. For CXCL2, average BMDM levels were reduced by 17× for whole bacteria and 177× for the outer membrane fraction. For IL-12A, levels for knockout mice were on average 3× less for whole bacteria and 41× less for outer membrane fraction. Of interest, the outer membrane fraction of N. sennetsu had mRNA levels comparable to the

BMDMs incubated with PBS treated with the Sarkosyl fractionation protocol, whereas N. sennetsu whole bacteria still had mild activity, indicating that the outer membrane fraction is involved with TLR2-dependent cytokine/chemokine induction (Figure 37,

Table 17).

146

Figure 37. TLR2-dependent cytokine production by N. sennetsu in vitro and link to outer membrane fraction. (A) All splenocyte cytokine transcript levels are at 8 h p.i. and are relative to respective cytokine transcript levels from splenocytes incubated with RPMI.

Values shown are means ± standard deviations for three wells. C, 1 × 106 splenocytes incubated with RPMI. NS, 1 × 106 splenocytes inoculated with 50 µg/ml N. sennetsu bacteria. (B) All BMDM cytokine transcript levels are at 2 h p.i. and are relative to respective cytokine transcript levels from BMDMs inoculated with PBS treated with the

Sarkosyl fractionation protocol. Values shown are means ± standard deviations for three

147

wells. C, 2 × 105 BMDMs inoculated with PBS treated with the Sarkosyl fractionation protocol. NS, 4.5 × 105 BMDMs inoculated with bacteria from 7 × 106 N. sennetsu

6 heavily-infected P388D1 per well. OM, 1 × 10 splenocytes inoculated with 10 µg/ml outer membrane fraction, derived from 50 µg/ml N. sennetsu bacteria. WT, C57J/BL6 wild-type mice. KO, TLR2 knockout mice. **, P < 0.01 by the Student’s t-test between C and NS levels or C and OM levels.

In vitro cytokine response using recombinant proteins. Recombinant Nsp2, rNsp3, and rP51 were used as inoculate for splenocytes and/or BMDMs. In the splenocytes studies, there is a noticeable increase in CXCL2 levels with rNsp2 incubation. There is, however induction in the TLR knockout mice, resulting in only a 3.7× reduction in average

CXCL2 expression between wild-type and TLR knockout splenocytes and suggesting potential LPS contamination. Since Nsp2 levels in vitro and in vivo are much lower than the two dominant proteins: P51 and Nsp3, the rNsp2 quantity was reduced to one-tenth that of rP51 for BMDM inoculations. This time, there was no evident cytokine induction comparable to N. sennetsu or outer membrane fraction induction levels. There also was no strong cytokine induction with rNsp3 and rP51 incubations (Figure 38, Tables 16 and

17). Recombinant P51 recovery after Triton X-114 wash was very poor, suggesting that, even though there was protein detected by BCA assay, the rP51 was washed off the His- tag column. It is likely, from these preliminary experiments that native proteins will be needed for future experiments.

148

Figure 38. Recombinant protein incubation of splenocytes and BMDMs. (A) All splenocyte cytokine transcript levels are relative to respective cytokine transcript levels from splenocytes inoculated with incubated with PBS treated with the Sarkosyl fractionation protocol at 8 h p.i. Values shown are means ± standard deviations for three wells. C, 1 × 106 splenocytes inoculated with PBS treated with the Sarkosyl fractionation protocol at 8 h p.i. NS, 1 × 106 splenocytes inoculated with 50 µg/ml N. sennetsu bacteria at 8 h p.i. rNsp2, 1 × 106 splenocytes inoculated with 10 µg/ml rNsp2 at 8 h p.i. rNsp3, 1

149

× 106 splenocytes inoculated with 10 µg/ml rNsp3 at 8 h p.i. LPS, 1 × 106 splenocytes inoculated with 100 ng/ml LPS from E. coli 0111:B4 at 8 h p.i. WT, C57J/BL6 wild-type mice. KO, TLR2 knockout mice. (B) All BMDM C57J/BL6 wild-type mouse cytokine transcript levels are relative to respective cytokine transcript levels from BMDMs inoculated with PBS treated with the Sarkosyl fractionation protocol at 2 h p.i. Values shown are means ± standard deviations for three wells. C, 2 × 105 splenocytes inoculated with PBS treated with the Sarkosyl fractionation protocol at 2 h p.i. NS, 2 × 105 splenocytes inoculated with 50 µg/ml N. sennetsu bacteria at 2 h p.i. OM, 2 × 105 splenocytes inoculated with outer membrane fraction derived from 50 µg/ml N. sennetsu bacteria at 2 h p.i. rP51, 2 × 105 splenocytes inoculated with 7.5 µg/ml rP51 at 2 h p.i. rNsp2, 2 × 105 splenocytes inoculated with 0.75 µg/ml rNsp2 at 2 h p.i. LPS, 2 × 105

BMDMs inoculated with 100 ng/ml LPS from E. coli 0111:B4 at 2 h p.i. *, P < 0.05; **,

P < 0.01 by Student’s t-test between respective wild-type and knockout groups (A) or respective control and experimental groups (B).

150

Discussion:

N. sennetsu is the only human pathogen of the family Anaplasmataceae to infect and cause progressive disease in an immunocompetent mouse model. Normally, an immunosuppressed mouse model, such as the SCID mouse or a related strain able to infect mice (IOE and Ehrlichia muris, for examples) have to be used to study Ehrlichia and Anaplasma human pathogens (Wen, Rikihisa et al. 1995; Sotomayor, Popov et al.

2001; Borjesson and Barthold 2002; Miura and Rikihisa 2009). This work determined the usefulness of N. sennetsu could serve as a model to identify virulence factors, cytokine responses, and protective antigens. Regarding the variety in severity of clinical signs: from non-existent in the ICR mice to potentially fatal in the BALB/c, these different mouse strains may be useful in the future for creating a more realistic model of human

Sennetsu neorickettsiosis and determining characteristics of the mouse strains which result in resistance and susceptibility to disease.

Despite more abundant expression of Nsp3 over Nsp2, infected animals appear to recognize Nsp2 preferentially. This poor immune recognition of Nsp3 may facilitate survival and persistence of Nsp3-expressing N. sennetsu in mice. Perhaps the highly- immunogenic Nsp2 thwarts the immune system from recognizing the more highly- expressed Nsp3. However, in the absence of antibodies, Nsp3 is still more highly expressed than Nsp2 in cell culture, suggesting this may be an intrinsic property of N. sennetsu in mammalian cells. Also, in the case of N. risticii naturally-infected horses in

Chapter 4, rNsp2 and rNsp3 recognition by N. risticii antisera is not common, and rNsp3 recognition appears stronger than rNsp2 recognition. Interestingly, Nsp1 was

151

undetectable by proteomics in cell culture, as well as by qRT-PCR in mice. It would be beneficial to see whether these trends in Nsp expression are seen in human patients or in potential trematode vectors. Although more studies are necessary, the present work suggests rNsp2 may serve as a specific diagnostic antigen for Sennetsu neorickettsiosis. It is also possible, given the in vitro infection data using rNsp3 antisera and the conserved external loops noted in Chapter 4, that Nsp3 may serve as a vaccine candidate.

Compared to using whole mice, in vitro cytokine/chemokine expression assays have several advantages. First, it enables the testing of wild-type and knock out conditions in a quick, efficient, and statistically-significant manner. Splenocyte cultures allow for the study of a more heterogeneous condition, better mimicking the in vivo conditions.

BMDMs permit the examination of just macrophages, which are primary host cells for N. sennetsu, and determine their responses to N. sennetsu and bacterial components.

Disadvantages include the inability to study the whole animal impact, which would involve not just one cell type or cells from one organ, but multiple cells in multiple tissue and organs with multiple interacting factors which may play a role in disease manifestation.

There were differences noted between in vivo and in vitro cytokine/chemokine mRNA expression and among in vitro results. High levels of IL-12A were noted in BMDMs, but this was not seen in splenocytes. This is likely due to the used of mononuclear cells as opposed to the mixed populations seen in splenocyte cultures and mouse spleens. There was also a reduction in IFNγ mRNA levels induced by frozen bacteria versus freshly- isolated bacteria and a reduction in iNOS mRNA levels induced by bacterial outer

152

membrane fraction versus the corresponding quantity of N. sennetsu frozen bacteria. It was previously demonstrated in N. risticii that when IFNγ was induced, reactive nitrogen intermediates including iNOS were produced, which could kill the bacteria (Park and

Rikihisa 1992). Given the reduction in IFNγ in whole frozen bacteria and the bacterial outer membrane fraction, there may be a soluble protein reduced in frozen bacteria, a lipoprotein removed by Sarkosyl treatment (Huang, Lin et al. 2008), or perhaps the induction is due to an active process, such as the type IV secretion system (Lin, Zhang et al. 2009; Rikihisa and Lin). Furthermore, there was a more evident IL-1β, CXCL2, and

IL-12A mRNA expression reduction in TLR2 knockout mice for outer membrane- incubated BMDMs versus whole frozen bacteria-incubated BMDMs, indicating that outer membrane fraction component(s) are recognized by TLR2.

The findings of MyD88 and TLR2 dependence, the lack of LPS and peptidoglycan in N. sennetsu, and the identification of the bacterial outer membrane fraction inducing IL-1β,

CXCL2, and IL-12A in vitro strongly implicate two proteins: P51 and Nsp3 as PAMPs.

IL-1β is a secreted pro-inflammatory cytokine with varied functions, including increasing production of acute-phase proteins, leading to liver damage and causing disease symptoms such as fever, general malaise, lethargy, and anorexia (Tizard 2008). Both N. sennetsu and N. risticii (van Heeckeren, Rikihisa et al. 1993) cytokine expression data show increases in IL-1 indicating its importance. IL-12 has been demonstrated to be involved in IFNγ induction, which is involved in bacterial clearance (Park and Rikihisa

1991; Park and Rikihisa 1992), and IL-12 is known to be induced by TLR signaling, although this does not appear to be the case in A. phagocytophilum (Pedra, Tao et al.

153

2007). Given its porin activity, P51 has the potential to be a major cause of cytokine/chemokine induction. The upregulation of IL-1β, CXCL2, and TNFα correlate with pro-inflammatory findings in other bacterial porins, including those in Hib,

Salmonella, and S. dysenteriae (Galdiero, Vitiello et al. 2002; Galdiero, Finamore et al.

2004; Biswas, Banerjee et al. 2007). Nsp3 is also a possibility. The Nsps are related to

Wolbachia surface proteins (Wsps) (Lin, Zhang et al. 2009), and recombinant Wsp, has been demonstrated to upregulate mRNAs for IL-1β, IL-6, and TNF in vitro (Porksakorn,

Nuchprayoon et al. 2007).

The role of CXCL2 in Sennetsu neorickettsiosis is currently unknown. To date, this chemokine has not been studied in N. risticii. The major role of the chemokine CXCL2

(equivalent to human IL-8) is to recruit neutrophils (Kobayashi 2008). As stated, N. sennetsu invades monocytes and macrophages, not neutrophils (Rikihisa, Dumler et al.

2005); should bacteria attempt to invade, neutrophils will in fact destroy the N. risticii

(Messick and Rikihisa 1993). Also, neutrophilia is not a hallmark finding in Neorickettsia infection (Rikihisa, Dumler et al. 2005). There are other listed roles for CXCL2/IL-8, including reducing bacterial invasion in enteric bacterial pathogens, such as Shigella flexerni (Sansonetti, Arondel et al. 1999); bone marrow suppression (Broxmeyer, Sherry et al. 1993); and angiogenesis (Koch, Polverini et al. 1992). There is evidence of blood abnormalities, including leukopenia during infections with Neorickettsia (Rikihisa,

Dumler et al. 2005). It is also possible that the goal of CXCL2 induction in recruiting neutrophils to further recruit monocytes and macrophages. It is still unclear exactly how

CXCL2 induction is of benefit to N. sennetsu. Future in vivo studies with knockout mice

154

(MyD88, TLR2, and CXCL2 receptor) will assist in determining if CXCL2 induction, and TLR2 recognition and MyD88 dependent signaling are advantageous to the bacterium (Pedra, Tao et al. 2007). Further experiments using human cell lines and/or blood-derived macrophages will be needed to determine if human IL-8 follows a similar pattern to that of murine CXCL2.

Studies on recombinant protein inoculation is limited, and future studies using HPLC fractionated N. sennetsu outer membrane fraction (as was performed in Chapter 3) to acquire native P51 and Nsp3 and determine if one or both proteins in involved in cytokine/chemokine production. There may be conformational changes in rP51 and rNsp3 reducing their effectiveness at inducing cytokine/chemokines. It is also possible that Nsp2 has a role in cytokine/chemokine production with the levels of CXCL2 induced in splenocytes. However, this is offset by the low protein levels found in N. sennetsu

(Figure 30).

Chapter 5 has demonstrated the uses of in vivo and in vitro models of N. sennetsu infection and disease. Although it remains to be seen if P51 is indeed the major cause of cytokine/chemokine induction, strong evidence has been presented to show that the outer membrane fraction of N. sennetsu is prominent in MyD88-dependent induction.

155

Conclusions:

Neorickettsia are the only bacteria known to infect digenetic trematodes, have wide geographic and mammalian host ranges, and are capable of causing disease in animals and humans. Through these studies, the host-pathogen relationships Neorickettsia and its trematode and mammalian hosts share have been explored. Shared lineage of the N. risticii-infected trematodes of both adult and immature life stages in Pennsylvania has been confirmed and linked to A. oregonense. Surface-expressed proteins have been identified for N. sennetsu, and one major β-barrel protein, P51 has been demonstrated to have porin activity. Major surface-expressed proteins of N. risticii have been identified, and geographic and temporal variations within these proteins among N. risticii strains have been demonstrated. Finally, cytokine and chemokine expression during N. sennetsu infection and disease has been studied, demonstrating a major role for the bacterial outer membrane fraction. These studies, presented in this dissertation, have thus analyzed the host-bacterium relationships of Neorickettsia in the pursuit of better understanding and combating these unique bacteria.

156

References

Altschul, S. F., T. L. Madden, et al. (1997). "Gapped BLAST and PSI-BLAST: a new generation of protein database search programs." Nucleic Acids Res 25(17): 3389-3402.

Anderson, D. R., H. E. Hopps, et al. (1965). "Comparison of the ultrastructure of several rickettsiae, ornithosis virus, and Mycoplasma in tissue culture." J Bacteriol 90(5): 1387-1404.

Bagos, P. G., T. D. Liakopoulos, et al. (2004). "PRED-TMBB: a web server for predicting the topology of beta-barrel outer membrane proteins." Nucleic Acids Res 32(Web Server issue): W400-404.

Baldermann, C. and H. Engelhardt (2000). "Expression, two-dimensional crystallization, and three-dimensional reconstruction of the beta8 outer membrane protein Omp21 from Comamonas acidovorans." J Struct Biol 131(2): 96-107.

Baldermann, C., A. Lupas, et al. (1998). "The regulated outer membrane protein Omp21 from Comamonas acidovorans is identified as a member of a new family of eight- stranded beta-sheet proteins by its sequence and properties." J Bacteriol 180(15): 3741-3749.

Baldwin, N. L., R. E. Millemann, et al. (1967). ""Salmon poisoning" disease. 3. Effect of experimental Nanophyetus salmincola infection on the fish host." J Parasitol 53(3): 556-564.

Bandi, C., A. J. Trees, et al. (2001). "Wolbachia in filarial nematodes: evolutionary aspects and implications for the pathogenesis and treatment of filarial diseases." Vet Parasitol 98(1-3): 215-238.

157

Barbet, A. F., P. F. Meeus, et al. (2003). "Expression of multiple outer membrane protein sequence variants from a single genomic locus of Anaplasma phagocytophilum." Infect Immun 71(4): 1706-1718.

Barlough, J. E., G. H. Reubel, et al. (1998). "Detection of Ehrlichia risticii, the agent of Potomac horse fever, in freshwater stream snails (Pleuroceridae: Juga spp.) from northern California." Appl Environ Microbiol 64(8): 2888-2893.

Barrett, A. J., N. D. Rawlings, et al., Eds. (2004). Handbook of proteolytic enzymes. Amsterdam ; Boston, Elsvier Academic Press.

Bendtsen, J. D., H. Nielsen, et al. (2004). "Improved prediction of signal peptides: SignalP 3.0." J Mol Biol 340(4): 783-795.

Bennington, E. and I. Pratt (1960). "The life history of the salmon poisoning fluke, Nanophyetus salmincola (Chapin)." J Parasitol 46: 91-100.

Biswas, A., P. Banerjee, et al. (2009). "Porin of Shigella dysenteriae directly promotes toll-like receptor 2-mediated CD4+ T cell survival and effector function." Mol Immunol 46(15): 3076-3085.

Biswas, A., P. Banerjee, et al. (2007). "Porin of Shigella dysenteriae activates mouse peritoneal macrophage through Toll-like receptors 2 and 6 to induce polarized type I response." Mol Immunol 44(5): 812-820.

Biswas, B., R. Vemulapalli, et al. (1998). "Molecular basis for antigenic variation of a protective strain-specific antigen of Ehrlichia risticii." Infect Immun 66(8): 3682- 3688.

Bolhuis, A., E. Koetje, et al. (2000). "Did the mitochondrial processing peptidase evolve from a eubacterial regulator of gene expression?" Mol Biol Evol 17(1): 198-201.

Booth, A. J., L. Stogdale, et al. (1984). "Salmon Poisoning Disease in Dogs on Southern Vancouver Island." Can Vet J 25(1): 2-6.

Borjesson, D. L. and S. W. Barthold (2002). "The mouse as a model for investigation of human granulocytic ehrlichiosis: current knowledge and future directions." Comp Med 52(5): 403-413. 158

Brogden, K. A. (2007). Virulence mechanisms of bacterial pathogens. Washington, D.C., ASM Press.

Brouqui, P. and D. Raoult (1990). "In vitro susceptibility of Ehrlichia sennetsu to antibiotics." Antimicrob Agents Chemother 34(8): 1593-1596.

Broxmeyer, H. E., B. Sherry, et al. (1993). "Comparative analysis of the human macrophage inflammatory protein family of cytokines (chemokines) on proliferation of human myeloid progenitor cells. Interacting effects involving suppression, synergistic suppression, and blocking of suppression." J Immunol 150(8 Pt 1): 3448-3458.

Bulut, Y., E. Faure, et al. (2002). "Chlamydial heat shock protein 60 activates macrophages and endothelial cells through Toll-like receptor 4 and MD2 in a MyD88-dependent pathway." J Immunol 168(3): 1435-1440.

Bulut, Y., K. Shimada, et al. (2009). "Chlamydial heat shock protein 60 induces acute pulmonary inflammation in mice via the Toll-like receptor 4- and MyD88- dependent pathway." Infect Immun 77(7): 2683-2690.

Burns, W. C. (1961). "Penetration and development of Allassogonoporus vespertilionis and Acanthatrium oregonense (Trematoda: Lecithodendriidae) cercariae in caddis fly larvae." J Parasitol 47: 927-932.

Carroll, J. F. and E. T. Schmidtmann (1986). "American Dog Tick (Acari, Ixodidae), Summer Activity on Equine Premises Enzootic for Potomac Horse Fever in South-Central Maryland." Journal of Economic Entomology 79(1): 62-66.

Chae, J. S., M. S. Kim, et al. (2002). "Detection of Neorickettsia (Ehrlichia) risticii in tissues of mice experimentally infected with cercariae of trematodes by in situ hybridization." Vet Microbiol 88(3): 233-243.

Chae, J. S., N. Pusterla, et al. (2000). "Infection of aquatic insects with trematode metacercariae carrying Ehrlichia risticii, the cause of Potomac horse fever." J Med Entomol 37(4): 619-625.

159

Chaichanasiriwithaya, W., Y. Rikihisa, et al. (1994). "Antigenic, morphologic, and molecular characterization of new Ehrlichia risticii isolates." J Clin Microbiol 32(12): 3026-3033.

Cheever, A. W. and Z. A. Andrade (1967). "Pathological lesions associated with Schistosoma mansoni infection in man." Trans R Soc Trop Med Hyg 61(5): 626- 639.

Cirillo, D. M., E. J. Heffernan, et al. (1996). "Identification of a domain in Rck, a product of the Salmonella typhimurium virulence plasmid, required for both serum resistance and cell invasion." Infect Immun 64(6): 2019-2023.

Cloud, K. A. and J. P. Dillard (2002). "A lytic transglycosylase of Neisseria gonorrhoeae is involved in peptidoglycan-derived cytotoxin production." Infect Immun 70(6): 2752-2757.

Cole, A. I., M. Ristic, et al. (1985). "Continuous propagation of Ehrlichia sennetsu in murine macrophage cell cultures." Am J Trop Med Hyg 34(4): 774-780.

Cordy, D. R. and J. R. Gorham (1950). "The pathology and etiology of salmon disease." Am J Pathol 26(4): 617-637.

Dekumyoy, P., J. Waikagul, et al. (1998). "Human lung fluke Paragonimus heterotremus: differential diagnosis between Paragonimus heterotremus and Paragonimus westermani infections by EITB." Trop Med Int Health 3(1): 52-56.

Doyle, C. K., A. M. Cardenas, et al. (2005). "Molecular characterization of E. canis gp36 and E. chaffeensis gp47 tandem repeats among isolates from different geographic locations." Ann N Y Acad Sci 1063: 433-435.

Dumler, J. S., A. F. Barbet, et al. (2001). "Reorganization of genera in the families Rickettsiaceae and Anaplasmataceae in the order Rickettsiales: unification of some species of Ehrlichia with Anaplasma, Cowdria with Ehrlichia and Ehrlichia with Neorickettsia, descriptions of six new species combinations and designation of Ehrlichia equi and 'HGE agent' as subjective synonyms of Ehrlichia phagocytophila." Int J Syst Evol Microbiol 51(Pt 6): 2145-2165.

160

Dutra, F., L. F. Schuch, et al. (2001). "Equine monocytic Ehrlichiosis (Potomac horse fever) in horses in Uruguay and southern Brazil." J Vet Diagn Invest 13(5): 433- 437.

Dutta, S. K., B. L. Mattingly, et al. (1989). "Antibody response to Ehrlichia risticii and antibody reactivity to the component antigens in horses with induced Potomac horse fever." Infect Immun 57(10): 2959-2962.

Dutta, S. K., B. E. Penney, et al. (1988). "Disease features in horses with induced equine monocytic ehrlichiosis (Potomac horse fever)." Am J Vet Res 49(10): 1747-1751.

Dutta, S. K., B. Shankarappa, et al. (1991). "Molecular cloning and analysis of recombinant major antigens of Ehrlichia risticii." Infect Immun 59(3): 1162- 1169.

Dutta, S. K., R. Vemulapalli, et al. (1998). "Association of deficiency in antibody response to vaccine and heterogeneity of Ehrlichia risticii strains with Potomac horse fever vaccine failure in horses." J Clin Microbiol 36(2): 506-512.

Eastburn, R. L., T. R. Fritsche, et al. (1987). "Human intestinal infection with Nanophyetus salmincola from salmonid fishes." Am J Trop Med Hyg 36(3): 586- 591.

Ensgraber, M. and M. Loos (1992). "A 66-kilodalton heat shock protein of Salmonella typhimurium is responsible for binding of the bacterium to intestinal mucus." Infect Immun 60(8): 3072-3078.

Farrell, R. K., R. W. Leader, et al. (1973). "Differentiation of salmon poisoning disease and Elokomin fluke fever: studies with the black bear (Ursus americanus)." Am J Vet Res 34(7): 919-922.

Fauquier, D., F. Gulland, et al. (2004). "Meningoencephalitis in two stranded California sea lions (Zalophus californianus) caused by aberrant trematode migration." J Wildl Dis 40(4): 816-819.

Felsenstein, J. (1989). "PHYLIP -- Phylogeny Inference Package (Version 3.2)." Cladistics 5: 164-166.

161

Fenn, K. and M. Blaxter (2004). "Are filarial nematode Wolbachia obligate mutualist symbionts?" Trends Ecol Evol 19(4): 163-166.

Fenn, K. and M. Blaxter (2006). "Wolbachia genomes: revealing the biology of parasitism and mutualism." Trends Parasitol 22(2): 60-65.

Fenn, K., C. Conlon, et al. (2006). "Phylogenetic relationships of the Wolbachia of nematodes and arthropods." PLoS Pathog 2(10): e94.

Foster, J., M. Ganatra, et al. (2005). "The Wolbachia genome of Brugia malayi: endosymbiont evolution within a human pathogenic nematode." PLoS Biol 3(4): e121.

Frisk, A., C. A. Ison, et al. (1998). "GroEL heat shock protein of Haemophilus ducreyi: association with cell surface and capacity to bind to eukaryotic cells." Infect Immun 66(3): 1252-1257.

Fukuda, T. (1958). "Rickettsial mononucleosis (synonyms: hyuganetsu, kagaminetsu)." Kitasato Arch. Exp. Med. 31: 51-56.

Fukuda, T., T. Kitao, et al. (1954). "Studies on the causative agent of “Hyuganetsu” disease. I. Isolation of the agent and its inoculation trial in human beings." Med Biol 32(4): 200-209.

Fukuda, T. and S. Yamamoto (1981). "Neorickettsia-like organism isolated from metacercaria of a fluke, Stellantchasmus falcatus." Jpn J Med Sci Biol 34(2): 103- 107.

Galdiero, F., C. de L'ero G, et al. (1993). "Release of cytokines induced by Salmonella typhimurium porins." Infect Immun 61(1): 155-161.

Galdiero, M., M. D'Amico, et al. (2001). "Haemophilus influenzae porin contributes to signaling of the inflammatory cascade in rat brain." Infect Immun 69(1): 221-227.

Galdiero, M., E. Finamore, et al. (2004). "Haemophilus influenzae porin induces Toll- like receptor 2-mediated cytokine production in human monocytes and mouse macrophages." Infect Immun 72(2): 1204-1209.

162

Galdiero, M., M. Vitiello, et al. (2002). "Porins from Salmonella enterica serovar Typhimurium activate the transcription factors activating protein 1 and NF- kappaB through the Raf-1-mitogen-activated protein kinase cascade." Infect Immun 70(2): 558-568.

Garduno, R. A., G. Faulkner, et al. (1998). "Immunolocalization of Hsp60 in Legionella pneumophila." J Bacteriol 180(3): 505-513.

Garduno, R. A., E. Garduno, et al. (1998). "Surface-associated hsp60 chaperonin of Legionella pneumophila mediates invasion in a HeLa cell model." Infect Immun 66(10): 4602-4610.

Gardy, J. L., M. R. Laird, et al. (2005). "PSORTb v.2.0: expanded prediction of bacterial protein subcellular localization and insights gained from comparative proteome analysis." Bioinformatics 21(5): 617-623.

Gasteiger, E., C. Hoogland, et al. (2005). Protein identification and analysis tools on the ExPASy server. The proteomics protocols handbook. J. M. Walker. Totowa, N.J., Humana Press: 571-607.

Gavotte, L., H. Henri, et al. (2007). "A survey of the bacteriophage WO in the endosymbiotic bacteria Wolbachia." Mol Biol Evol 24(2): 427-435.

Ge, Y. and Y. Rikihisa (2007). "Identification of novel surface proteins of Anaplasma phagocytophilum by affinity purification and proteomics." J Bacteriol 189(21): 7819-7828.

Ge, Y. and Y. Rikihisa (2007). "Surface-exposed proteins of Ehrlichia chaffeensis." Infect Immun 75(8): 3833-3841.

Gebhardt, G. A., R. E. Millemann, et al. (1966). ""Salmon poisoning" disease. II. Second intermediate host susceptibility studies." J Parasitol 52(1): 54-59.

Gibson, K., Y. Kumagai, et al. (2010). "Proteomic analysis of Neorickettsia sennetsu surface-exposed proteins and porin activity of the major surface protein P51." J Bacteriol 192(22): 5898-5905.

163

Gibson, K. E. and Y. Rikihisa (2008). "Molecular link of different stages of the trematode host of Neorickettsia risticii to Acanthatrium oregonense." Environ Microbiol 10(8): 2064-2073.

Gibson, K. E., Y. Rikihisa, et al. (2005). "Neorickettsia risticii is vertically transmitted in the trematode Acanthatrium oregonense and horizontally transmitted to bats." Environ Microbiol 7(2): 203-212.

Hardesty, C., C. Ferran, et al. (1991). "Plasmid-mediated sucrose metabolism in Escherichia coli: characterization of scrY, the structural gene for a phosphoenolpyruvate-dependent sucrose phosphotransferase system outer membrane porin." J Bacteriol 173(2): 449-456.

Hartmann, E., C. A. Lingwood, et al. (2001). "Heat-inducible surface stress protein (Hsp70) mediates sulfatide recognition of the respiratory pathogen Haemophilus influenzae." Infect Immun 69(5): 3438-3441.

Hastriter, M. W., D. J. Kelly, et al. (1987). "Evaluation of Leptotrombidium (Leptotrombidium) fletcheri (Acari: Trombiculidae) as a potential vector of Ehrlichia sennetsu." J Med Entomol 24(5): 542-546.

Headley, S. A., D. G. Scorpio, et al. (2006). "Neorickettsia helminthoeca in dog, Brazil." Emerg Infect Dis 12(8): 1303-1304.

Headley, S. A., D. G. Scorpio, et al. (2009). "Neorickettsia helminthoeca and salmon poisoning disease: A review." Vet J 187(2): 165-173.

Headley, S. A., O. Vidotto, et al. (2004). "Suspected cases of Neorickettsia-like organisms in Brazilian dogs." Ann N Y Acad Sci 1026: 79-83.

Hechemy, K. E., J. A. Oteo, et al. (2006). "A century of rickettsiology: emerging, reemerging rickettsioses, clinical, epidemiologic, and molecular diagnostic aspects and emerging veterinary rickettsioses: an overview." Ann N Y Acad Sci 1078: 1-14.

Hecker, K. H. and K. H. Roux (1996). "High and low annealing temperatures increase both specificity and yield in touchdown and stepdown PCR." Biotechniques 20(3): 478-485.

164

Heider, U., M. Blaxter, et al. (2006). "Differential display of genes expressed in the filarial nematode Litomosoides sigmodontis reveals a putative phosphate permease up-regulated after depletion of Wolbachia endobacteria." Int J Med Microbiol 296(4-5): 287-299.

Hoerauf, A., K. Nissen-Pahle, et al. (1999). "Tetracycline therapy targets intracellular bacteria in the filarial nematode Litomosoides sigmodontis and results in filarial infertility." J Clin Invest 103(1): 11-18.

Hoerauf, A., L. Volkmann, et al. (2000). "Targeting of Wolbachia endobacteria in Litomosoides sigmodontis: comparison of tetracyclines with chloramphenicol, macrolides and ciprofloxacin." Trop Med Int Health 5(4): 275-279.

Hoilien, C. A., M. Ristic, et al. (1982). "Rickettsia sennetsu in human blood monocyte cultures: similarities to the growth cycle of Ehrlichia canis." Infect Immun 35(1): 314-319.

Holland, C. J., M. Ristic, et al. (1985). "Isolation, experimental transmission, and characterization of causative agent of Potomac horse fever." Science 227(4686): 522-524.

Holland, C. J., M. Ristic, et al. (1985). "Adaptation of Ehrlichia sennetsu to canine blood monocytes: preliminary structural and serological studies with cell culture-derived Ehrlichia sennetsu." Infect Immun 48(2): 366-371.

Holland, C. J., E. Weiss, et al. (1985). "Ehrlichia-Risticii Sp-Nov - Etiological Agent of Equine Monocytic Ehrlichiosis (Synonym, Potomac Horse Fever)." International Journal of Systematic Bacteriology 35(4): 524-526.

Hotopp, J. C., M. Lin, et al. (2006). "Comparative genomics of emerging human ehrlichiosis agents." PLoS Genet 2(2): e21.

Huang, H., M. Lin, et al. (2008). "Proteomic analysis of and immune responses to Ehrlichia chaffeensis lipoproteins." Infect Immun 76(8): 3405-3414.

Huang, H., X. Wang, et al. (2007). "Porin activity of Anaplasma phagocytophilum outer membrane fraction and purified P44." J Bacteriol 189(5): 1998-2006.

165

Ito, A. (1999). "Mitochondrial processing peptidase: multiple-site recognition of precursor proteins." Biochem Biophys Res Commun 265(3): 611-616.

Jeanteur, D., J. H. Lakey, et al. (1991). "The bacterial porin superfamily: sequence alignment and structure prediction." Mol Microbiol 5(9): 2153-2164.

Juncker, A. S., H. Willenbrock, et al. (2003). "Prediction of lipoprotein signal peptides in Gram-negative bacteria." Protein Sci 12(8): 1652-1662.

Kanter, M., J. Mott, et al. (2000). "Analysis of 16S rRNA and 51-kilodalton antigen gene and transmission in mice of Ehrlichia risticii in virgulate trematodes from Elimia livescens snails in Ohio." J Clin Microbiol 38(9): 3349-3358.

Kaylor, P. S., T. B. Crawford, et al. (1991). "Passive transfer of antibody to Ehrlichia risticii protects mice from ehrlichiosis." Infect Immun 59(6): 2058-2062.

Kigston, N. (1978). Trematodes. Diseases of poultry. M. S. Hofstad. Ames, Iowa State University Press: 777-781.

Kitao, T., R. K. Farrell, et al. (1973). "Differentiation of salmon poisoning disease and Elokomin fluke fever: fluorescent antibody studies with Rickettsia sennetsu." Am J Vet Res 34(7): 927-928.

Knapp, S. E. and R. E. Millemann (1981). Salmon poisoning disease. Infectious diseases of wild mammals. J. W. Davis, L. H. Karstad and D. O. Trainer. Ames, Iowa, Iowa State University Press: 376–387.

Knight, R. A. and I. Pratt (1955). "The life-histories of Allassogonoporus vespertilionis Macy and Acanthatrium oregonense Macy (Trematoda: Lecithodendriidae)." J Parasitol 41(3): 248-255.

Knowles, R. C., C. Anderson, et al. (1984). Acute Equine Diarrhea Syndrome (AEDS): A Preliminary Report. Proceedings of the 29th annual convention of the American Association of Equine Practitioners. F. J. Milne. Las Vegas, Nevada, American Association of Equine Practitioners: 353-357.

Kobayashi, Y. (2008). "The role of chemokines in neutrophil biology." Front Biosci 13: 2400-2407. 166

Koch, A. E., P. J. Polverini, et al. (1992). "Interleukin-8 as a macrophage-derived mediator of angiogenesis." Science 258(5089): 1798-1801.

Koebnik, R., K. P. Locher, et al. (2000). "Structure and function of bacterial outer membrane proteins: barrels in a nutshell." Mol Microbiol 37(2): 239-253.

Koren, H. S., B. S. Handwerger, et al. (1975). "Identification of macrophage-like characteristics in a cultured murine tumor line." J Immunol 114(2 pt 2): 894-897.

Kumagai, Y., H. Huang, et al. (2008). "Expression and porin activity of P28 and OMP-1F during intracellular Ehrlichia chaffeensis development." J Bacteriol 190(10): 3597-3605.

Kumagai, Y., J. Matsuo, et al. (2010). "Cyclic di-GMP signaling regulates invasion by Ehrlichia chaffeensis of human monocytes." J Bacteriol 192(16): 4122-4133.

Labesse, G., E. Garnotel, et al. (2001). "MOMP, a divergent porin from Campylobacter: cloning and primary structural characterization." Biochem Biophys Res Commun 280(1): 380-387.

LeClerc, J. E., B. Li, et al. (1996). "High mutation frequencies among Escherichia coli and Salmonella pathogens." Science 274(5290): 1208-1211.

Lee, S. H., K. H. Koo, et al. (1989). "[Experimental infection of Paragonimus iloktsuenensis to albino rats, dogs and cats]." Kisaengchunghak Chapchi 27(3): 197-202.

Leok, C. S., I. Inoue, et al. (2002). "Morphology of the oviduct fluke, Prosthogonimus ovatus, isolated from Indonesian native chickens and histopathological observation of the infected chickens." J Vet Med Sci 64(12): 1129-1131.

Lin, H. and C. C. Stewart (1973). "Colony formation by mouse peritoneal exudate cells in vitro." Nat New Biol 243(127): 176-177.

Lin, M., C. Zhang, et al. (2009). "Analysis of complete genome sequence of Neorickettsia risticii: causative agent of Potomac horse fever." Nucleic Acids Res 37(18): 6076-6091.

167

Lin, Q., Y. Rikihisa, et al. (2003). "Mechanisms of variable p44 expression by Anaplasma phagocytophilum." Infect Immun 71(10): 5650-5661.

Littlewood, D. T. and P. D. Olson (2001). Small Subunit rDNA and the Platyhelminthes: Signal, Noise, Conflict, and Compromise. Interrelationships of the Platyhelminthes. D. T. J. Littlewood and R. A. Bray. New York, Taylor & Francis: 262-278.

Littlewood, D. T., K. Rohde, et al. (1999). "The interrelationships of all major groups of Platyhelminthes: phylogenetic evidence from morphology and molecules." Biological Journal of the Linnean Society 66(1): 75-114.

Macy, R. W. (1939). "A new trematode, Acanthatrium oregonense, (Lecithodendriidae), from bats of the genus Myotis." Am Midl Nat 22: 640-641.

Madigan, J. E., N. Pusterla, et al. (2000). "Transmission of Ehrlichia risticii, the agent of Potomac horse fever, using naturally infected aquatic insects and helminth vectors: preliminary report." Equine Vet J 32(4): 275-279.

Mamelak, D., M. Mylvaganam, et al. (2001). "Hsp70s contain a specific sulfogalactolipid binding site. Differential aglycone influence on sulfogalactosyl ceramide binding by recombinant prokaryotic and eukaryotic hsp70 family members." Biochemistry 40(12): 3572-3582.

Marchler-Bauer, A., J. B. Anderson, et al. (2007). "CDD: a conserved domain database for interactive domain family analysis." Nucleic Acids Res 35(Database issue): D237-240.

Marquis, H. and T. A. Ficht (1993). "The omp2 gene locus of Brucella abortus encodes two homologous outer membrane proteins with properties characteristic of bacterial porins." Infect Immun 61(9): 3785-3790.

Martin, S. A., R. S. Rosenthal, et al. (1987). "Fast atom bombardment mass spectrometry and tandem mass spectrometry of biologically active peptidoglycan monomers from Neisseria gonorrhoeae." J Biol Chem 262(16): 7514-7522.

168

Masui, S., H. Kuroiwa, et al. (2001). "Bacteriophage WO and virus-like particles in Wolbachia, an endosymbiont of arthropods." Biochem Biophys Res Commun 283(5): 1099-1104.

McGettrick, A. F. and L. A. O'Neill (2010). "Localisation and trafficking of Toll-like receptors: an important mode of regulation." Curr Opin Immunol 22(1): 20-27.

McManus, D. P. and J. P. Dalton (2006). "Vaccines against the zoonotic trematodes Schistosoma japonicum, Fasciola hepatica and Fasciola gigantica." Parasitology 133 Suppl: S43-61.

Melly, M. A., Z. A. McGee, et al. (1984). "Ability of monomeric peptidoglycan fragments from Neisseria gonorrhoeae to damage human fallopian-tube mucosa." J Infect Dis 149(3): 378-386.

Messick, J. B. and Y. Rikihisa (1993). "Characterization of Ehrlichia risticii binding, internalization, and proliferation in host cells by flow cytometry." Infect Immun 61(9): 3803-3810.

Messick, J. B. and Y. Rikihisa (1994). "Inhibition of binding, entry, or intracellular proliferation of Ehrlichia risticii in P388D1 cells by anti-E. risticii serum, immunoglobulin G, or Fab fragment." Infect Immun 62(8): 3156-3161.

Millemann, R. E. and S. E. Knapp (1970). "Biology of Nanophyetus salmincola and "salmon poisoning" disease." Adv Parasitol 8: 1-41.

Miller, V. L., J. B. Bliska, et al. (1990). "Nucleotide sequence of the Yersinia enterocolitica ail gene and characterization of the Ail protein product." J Bacteriol 172(2): 1062-1069.

Misao, T. and Y. Kobayashi (1954). "Studies on infectious mononucleosis. I. Isolation of etiologic agent from blood, bone marrow and lymph node of a patient with infectious mononucleosis by using mice." Tokyo Iji Shinshi 71: 683-686.

Misao, T. and Y. Kobayashi (1955). "Studies on infectious mononucleosis (glandular fever). 1. Isolation of the etiological agent from blood, bone marrow and lymph nodes of a patient infected with infectious mononucleosis by using mice." Kyushu J. Med. Sci. 6: 145-152.

169

Miura, K. and Y. Rikihisa (2007). "Virulence potential of Ehrlichia chaffeensis strains of distinct genome sequences." Infect Immun 75(7): 3604-3613.

Miura, K. and Y. Rikihisa (2009). "Liver transcriptome profiles associated with strain- specific Ehrlichia chaffeensis-induced hepatitis in SCID mice." Infect Immun 77(1): 245-254.

Mott, J., Y. Muramatsu, et al. (2002). "Molecular analysis of Neorickettsia risticii in adult aquatic insects in Pennsylvania, in horses infected by ingestion of insects, and isolated in cell culture." J Clin Microbiol 40(2): 690-693.

Mott, J., Y. Rikihisa, et al. (1997). "Comparison of PCR and culture to the indirect fluorescent-antibody test for diagnosis of Potomac horse fever." J Clin Microbiol 35(9): 2215-2219.

Newton, P. N., J. M. Rolain, et al. (2009). "Sennetsu neorickettsiosis: a probable fish- borne cause of fever rediscovered in Laos." Am J Trop Med Hyg 81(2): 190-194.

Nikaido, H. (2003). "Molecular basis of bacterial outer membrane permeability revisited." Microbiol Mol Biol Rev 67(4): 593-656.

Nikaido, H., K. Nikaido, et al. (1991). "Identification and characterization of porins in Pseudomonas aeruginosa." J Biol Chem 266(2): 770-779.

Nikaido, H. and E. Y. Rosenberg (1983). "Porin channels in Escherichia coli: studies with liposomes reconstituted from purified proteins." J Bacteriol 153(1): 241-252.

Nyberg, P. A., S. E. Knapp, et al. (1967). ""Salmon poisoning" disease. IV. Transmission of the disease to dogs by Nanophyetus salmincola eggs." J Parasitol 53(4): 694- 699.

Olsen, G. J. and C. R. Woese (1993). "Ribosomal RNA: a key to phylogeny." Faseb J 7(1): 113-123.

Olson, P. D., T. H. Cribb, et al. (2003). "Phylogeny and classification of the Digenea (Platyhelminthes: Trematoda)." Int J Parasitol 33(7): 733-755.

170

Palmer, J. E. (1989). "Prevention of Potomac horse fever." Cornell Vet 79(3): 201-205.

Pantzar, M., S. Teneberg, et al. (2006). "Binding of Haemophilus ducreyi to carbohydrate receptors is mediated by the 58.5-kDa GroEL heat shock protein." Microbes Infect 8(9-10): 2452-2458.

Park, B. K., M. J. Kim, et al. (2003). "Identification of trematode cercariae carrying Neorickettsia risticii in freshwater stream snails." Ann N Y Acad Sci 990: 239- 247.

Park, J. and Y. Rikihisa (1991). "Inhibition of Ehrlichia risticii infection in murine peritoneal macrophages by gamma interferon, a calcium ionophore, and concanavalin A." Infect Immun 59(10): 3418-3423.

Park, J. and Y. Rikihisa (1992). "L-arginine-dependent killing of intracellular Ehrlichia risticii by macrophages treated with gamma interferon." Infect Immun 60(9): 3504-3508.

Pearce, E. J. (2005). "Priming of the immune response by schistosome eggs." Parasite Immunol 27(7-8): 265-270.

Pedra, J. H., J. Tao, et al. (2007). "IL-12/23p40-dependent clearance of Anaplasma phagocytophilum in the murine model of human anaplasmosis." FEMS Immunol Med Microbiol 50(3): 401-410.

Philip, C. B. (1953). "Nomenclature of the rickettsiaceae pathogenic to vertebrates." Ann N Y Acad Sci 56(3): 484-494.

Philip, C. B., W. J. Hadlow, et al. (1953). Neorickettsia helminthoeca, a new rickettsia- like disease agent of dogs in western United States transmitted by a helminth. International Congress on Microbiology report: proceedings of the 6th congress. Rome, International Association of Microbiological Societies. 2: 256-257.

Philip, C. B., W. J. Hadlow, et al. (1954). "Studies on salmon poisoning disease of canines. I. The rickettsial relationships and pathogenicity of Neorickettsia helmintheca." Exp Parasitol 3(4): 336-350.

171

Porksakorn, C., S. Nuchprayoon, et al. (2007). "Proinflammatory cytokine gene expression by murine macrophages in response to Brugia malayi Wolbachia surface protein." Mediators Inflamm 2007: 84318.

Pretzman, C., D. Ralph, et al. (1995). "16S rRNA gene sequence of Neorickettsia helminthoeca and its phylogenetic alignment with members of the genus Ehrlichia." Int J Syst Bacteriol 45(2): 207-211.

Pretzman, C. I., Y. Rikihisa, et al. (1987). "Enzyme-linked immunosorbent assay for Potomac horse fever disease." J Clin Microbiol 25(1): 31-36.

Pritchard, M. H. and G. O. W. Kruse (1982). The collection and preservation of animal parasites. Lincoln, University of Nebraska Press.

Pusterla, N., E. Johnson, et al. (2000). "Molecular detection of an Ehrlichia-like agent in rainbow trout (Oncorhynchus mykiss) from Northern California." Vet Parasitol 92(3): 199-207.

Pusterla, N., E. M. Johnson, et al. (2003). "Digenetic trematodes, Acanthatrium sp. and Lecithodendrium sp., as vectors of Neorickettsia risticii, the agent of Potomac horse fever." J Helminthol 77(4): 335-339.

Pusterla, N., J. E. Madigan, et al. (2000). "Helminthic transmission and isolation of Ehrlichia risticii, the causative agent of Potomac horse fever, by using trematode stages from freshwater stream snails." J Clin Microbiol 38(3): 1293-1297.

Reichelt, P., C. Schwarz, et al. (2006). "Single step protocol to purify recombinant proteins with low endotoxin contents." Protein Expr Purif 46(2): 483-488.

Reubel, G. H., J. E. Barlough, et al. (1998). "Production and characterization of Ehrlichia risticii, the agent of Potomac horse fever, from snails (Pleuroceridae: Juga spp.) in aquarium culture and genetic comparison to equine strains." J Clin Microbiol 36(6): 1501-1511.

Rikihisa, Y. (1991). "Cross-reacting antigens between Neorickettsia helminthoeca and Ehrlichia species, shown by immunofluorescence and Western immunoblotting." J Clin Microbiol 29(9): 2024-2029.

172

Rikihisa, Y. (1991). "Protection against murine potomac horse fever by an inactivated Ehrlichia risticii vaccine." Vet Microbiol 27(3-4): 339-350.

Rikihisa, Y. (1991). "The tribe Ehrlichieae and ehrlichial diseases." Clin Microbiol Rev 4(3): 286-308.

Rikihisa, Y. (1998). Rickettsial diseases in horses. Equine internal medicine. S. M. Reed and W. M. Bayly. Philadelphia, Saunders: 112-123.

Rikihisa, Y. (2004). Rickettsial Diseases. Equine Internal Medicine, 2nd ed. S. M. Reed, W. M. Bayly and D. C. Sellon. Philadelphia, PA, W. B. Saunders: 96-109.

Rikihisa, Y., W. Chaichanasiriwithaya, et al. (1994). Distinct antigenic differences of new Ehrlichia risticii isolates. Proc. 7th International Conferences on Equine Infectious Diseases, Tokyo, Japan, R&W Publishing, Newmarket, U.K.

Rikihisa, Y., J. S. Dumler, et al. (2005). Neorickettsia. The Proteobacteria. G. M. Garrity, D. J. Brenner, N. R. Krieg and J. T. Staley. New York, Springer. 2: 132- 137.

Rikihisa, Y. and B. M. Jiang (1988). "In vitro susceptibilities of Ehrlichia risticii to eight antibiotics." Antimicrob Agents Chemother 32(7): 986-991.

Rikihisa, Y. and B. M. Jiang (1989). "Effect of antibiotics on clinical, pathologic and immunologic responses in murine Potomac horse fever: protective effects of doxycycline." Vet Microbiol 19(3): 253-262.

Rikihisa, Y., G. C. Johnson, et al. (1987). "Reduced immune responsiveness and lymphoid depletion in mice infected with Ehrlichia risticii." Infect Immun 55(9): 2215-2222.

Rikihisa, Y. and M. Lin (2010). "Anaplasma phagocytophilum and Ehrlichia chaffeensis type IV secretion and Ank proteins." Curr Opin Microbiol 13(1): 59-66.

Rikihisa, Y. and B. Perry (1984). "Causative agent of Potomac horse fever." Vet Rec 115(21): 554-555.

173

Rikihisa, Y., B. Perry, et al. (1984). "Rickettsial link with acute equine diarrhoea." Vet Rec 115(15): 390.

Rikihisa, Y., C. I. Pretzman, et al. (1988). "Clinical, histopathological, and immunological responses of ponies to Ehrlichia sennetsu and subsequent Ehrlichia risticii challenge." Infect Immun 56(11): 2960-2966.

Rikihisa, Y., H. Stills, et al. (1991). "Isolation and continuous culture of Neorickettsia helminthoeca in a macrophage cell line." J Clin Microbiol 29(9): 1928-1933.

Rikihisa, Y., C. Zhang, et al. (2004). "Analysis of p51, groESL, and the major antigen P51 in various species of Neorickettsia, an obligatory intracellular bacterium that infects trematodes and mammals." J Clin Microbiol 42(8): 3823-3826.

Rikihisa, Y. J. S. D. a. G. A. D. (2005). Neorickettsia. The Proteobacteria, Part C, Bergey's Manual of Systematic Bacteriology. K. Brenner, Staley and Garrity New York, NY Springer 2: 132-137.

Ristic, M., C. J. Holland, et al. (1986). "Diagnosis of equine monocytic ehrlichiosis (Potomac horse fever) by indirect immunofluorescence." J Am Vet Med Assoc 189(1): 39-46.

Ristic, M., D. L. Huxsoll, et al. (1981). "Evidence of a serologic relationship between Ehrlichia canis and Rickettsia sennetsu." Am J Trop Med Hyg 30(6): 1324-1328.

Saint Andre, A., N. M. Blackwell, et al. (2002). "The role of endosymbiotic Wolbachia bacteria in the pathogenesis of river blindness." Science 295(5561): 1892-1895.

Sansonetti, P. J., J. Arondel, et al. (1999). "Interleukin-8 controls bacterial transepithelial translocation at the cost of epithelial destruction in experimental shigellosis." Infect Immun 67(3): 1471-1480.

Schell, S. C. (1985). Handbook of trematodes of North America north of Mexico. Moscow, Idaho, University Press of Idaho.

Schirmer, T., T. A. Keller, et al. (1995). "Structural basis for sugar translocation through maltoporin channels at 3.1 A resolution." Science 267(5197): 512-514.

174

Schmidtmann, E. T., M. G. Robl, et al. (1986). "Attempted transmission of Ehrlichia risticii by field-captured Dermacentor variabilis (Acari: Ixodidae)." Am J Vet Res 47(11): 2393-2395.

Schulz, G. E. (2002). "The structure of bacterial outer membrane proteins." Biochim Biophys Acta 1565(2): 308-317.

Shankarappa, B., S. K. Dutta, et al. (1992). "Antigenic and genomic relatedness among Ehrlichia risticii, Ehrlichia sennetsu, and Ehrlichia canis." Int J Syst Bacteriol 42(1): 127-132.

Shirahama, S. (1967). "[Electron microscopic studies of glandular fever pathogen]." Nippon Densenbyo Gakkai Zasshi 41(2): 82-98.

Simms, B. T., C. R. Donham, et al. (1931). "Salmon Poisoning." Journal of the American Veterinary Medical Association 78: 181-195.

Sotomayor, E. A., V. L. Popov, et al. (2001). "Animal model of fatal human monocytotropic ehrlichiosis." Am J Pathol 158(2): 757-769.

Sumner, J. W., J. E. Childs, et al. (1999). "Molecular cloning and characterization of the Ehrlichia chaffeensis variable-length PCR target: an antigen-expressing gene that exhibits interstrain variation." J Clin Microbiol 37(5): 1447-1453.

Tachibana, N. and V. Kobayashi (1975). "Effect of cyclophosphamide on the growth of Rickettsia sennetsu in experimentally infected mice." Infect Immun 12(3): 625- 629.

Tachibana, N., T. Kusaba, et al. (1976). "Purification of complement-fixing antigens of Rickettsia sennetsu by ether treatment." Infect Immun 13(4): 1030-1036.

Tachibana, N., E. Shishime, et al. (1987). "[Two cases of spotted fever rickettsiosis in Kyushu]." Kansenshogaku Zasshi 61(10): 1166-1172.

Taylor, A. B., B. S. Smith, et al. (2001). "Crystal structures of mitochondrial processing peptidase reveal the mode for specific cleavage of import signal sequences." Structure 9(7): 615-625.

175

Taylor, M. J., H. F. Cross, et al. (2000). "Inflammatory responses induced by the filarial nematode Brugia malayi are mediated by lipopolysaccharide-like activity from endosymbiotic Wolbachia bacteria." J Exp Med 191(8): 1429-1436.

Taylor, M. J., H. F. Cross, et al. (2001). "Wolbachia bacteria in filarial immunity and disease." Parasite Immunol 23(7): 401-409.

Thomson, N., L. Crossman, et al. (2006). "Bacterial home from home." Nat Rev Microbiol 4(3): 168-170.

Tizard, I. R. (2008). Veterinary immunology : an introduction. St. Louis, MO, Saunders, an imprint of Elsevier Inc.

Tsan, M. F. and B. Gao (2009). "Heat shock proteins and immune system." J Leukoc Biol 85(6): 905-910. van der Kolk, J. H., W. E. Bernadina, et al. (1991). "[A horse seropositive for Ehrlichia risticii]." Tijdschr Diergeneeskd 116(2): 69-72. van Heeckeren, A. M., Y. Rikihisa, et al. (1993). "Tumor necrosis factor alpha, interleukin-1 alpha, interleukin-6, and prostaglandin E2 production in murine peritoneal macrophages infected with Ehrlichia risticii." Infect Immun 61(10): 4333-4337.

Vemulapalli, R., B. Biswas, et al. (1995). "Pathogenic, immunologic, and molecular differences between two Ehrlichia risticii strains." J Clin Microbiol 33(11): 2987- 2993.

Vemulapalli, R., B. Biswas, et al. (1998). "Cloning and molecular analysis of genes encoding two immunodominant antigens of Ehrlichia risticii." Microb Pathog 24(6): 361-372.

Vemulapalli, R., B. Biswas, et al. (1998). "Studies with recombinant proteins of Ehrlichia risticii: identification of strain-specific antigen as a protective antigen." Vet Parasitol 76(3): 189-202.

Vidor, E., G. Bissuel, et al. (1988). "Serologie positive a Ehrlichia risticii chez une jument presentant un tableau dehrlichiose equine." Pratique Vet. Equine 20: 5–10. 176

Wang, X., T. Kikuchi, et al. (2006). "Two monoclonal antibodies with defined epitopes of P44 major surface proteins neutralize Anaplasma phagocytophilum by distinct mechanisms." Infect Immun 74(3): 1873-1882.

Weiss, E., G. A. Dasch, et al. (1988). "Substrate utilization by Ehrlichia sennetsu and Ehrlichia risticii separated from host constituents by renografin gradient centrifugation." J Bacteriol 170(11): 5012-5017.

Weiss, E., J. C. Williams, et al. (1989). "Energy metabolism of monocytic Ehrlichia." Proc Natl Acad Sci U S A 86(5): 1674-1678.

Wells, M. Y. and Y. Rikihisa (1988). "Lack of lysosomal fusion with phagosomes containing Ehrlichia risticii in P388D1 cells: abrogation of inhibition with oxytetracycline." Infect Immun 56(12): 3209-3215.

Wen, B., Y. Rikihisa, et al. (1995). "Diversity of 16S rRNA genes of new Ehrlichia strains isolated from horses with clinical signs of Potomac horse fever." Int J Syst Bacteriol 45(2): 315-318.

Wen, B., Y. Rikihisa, et al. (1995). "Ehrlichia muris sp. nov., identified on the basis of 16S rRNA base sequences and serological, morphological, and biological characteristics." Int J Syst Bacteriol 45(2): 250-254.

Wen, B., Y. Rikihisa, et al. (1996). "Characterization of the SF agent, an Ehrlichia sp. isolated from the fluke Stellantchasmus falcatus, by 16S rRNA base sequence, serological, and morphological analyses." Int J Syst Bacteriol 46(1): 149-154.

Werren, J. H. (1997). "Biology of Wolbachia." Annu Rev Entomol 42: 587-609.

Whiting, M. F. (2002). "Phylogeny of the holometabolous insect orders: molecular evidence." Zoologica Scripta 31(1): 3–15.

Williams, N. M. and P. J. Timoney (1994). "Variation in susceptibility of ten mouse strains to infection with a strain of Ehrlichia risticii." J Comp Pathol 110(2): 137- 143.

Woese, C. R. (1987). "Bacterial evolution." Microbiol Rev 51(2): 221-271.

177

Wright, J. D., F. S. Sjostrand, et al. (1978). "The ultrastructure of the rickettsia-like microorganism Wolbachia pipientis and associated virus-like bodies in the mosquito Culex pipiens." J Ultrastruct Res 63(1): 79-85.

Wu, M., L. V. Sun, et al. (2004). "Phylogenomics of the reproductive parasite Wolbachia pipientis wMel: a streamlined genome overrun by mobile genetic elements." PLoS Biol 2(3): E69.

Xia, X., Z. Xie, et al. (2003). "18S ribosomal RNA and tetrapod phylogeny." Syst Biol 52(3): 283-295.

Zhang, Q., J. C. Meitzler, et al. (2000). "Sequence polymorphism, predicted secondary structures, and surface-exposed conformational epitopes of Campylobacter major outer membrane protein." Infect Immun 68(10): 5679-5689.

Zhang, X., T. Luo, et al. (2008). "Genetic and antigenic diversities of major immunoreactive proteins in globally distributed Ehrlichia canis strains." Clin Vaccine Immunol 15(7): 1080-1088.

Zhang, Y., N. Ohashi, et al. (1997). "Ehrlichia sennetsu groE operon and antigenic properties of the GroEL homolog." FEMS Immunol Med Microbiol 18(1): 39-46.

Zhang, Z., S. Schwartz, et al. (2000). "A greedy algorithm for aligning DNA sequences." J Comput Biol 7(1-2): 203-214.

178

Appendix A: Additional Tables

179

Mouse sample Inoculation Gene Means Std dev 2 h p.i.: 6.17 × 10−1 ±1.30 × 10−1 IL-1β 8 h p.i.: 1.09 ±7.83 × 10−2 2 h p.i.: 1.74 ±2.26 × 10−1 CXCL2 8 h p.i.: 1.27 × 101 ±1.42 N. sennetsu bacteria freshly TNFα 2 h p.i.: 4.83 × 10−1 ±7.20 × 10−2 isolated from 2 h p.i.: 1.05 ±7.28 × 10−2 6 CXCL1 −1 2.18 × 10 P388D1 8 h p.i.: 1.11 ±1.26 × 10 cells/well IFNγ 8 h p.i.: 2.33 × 10−1 ±4.87 × 10−2 iNOS 8 h p.i.: 5.65 × 10−1 ±7.23 × 10−2 2 h p.i.: 7.78 × 10−1 ±5.56 × 10−2 IL-12A 8 h p.i.: 1.17 ±7.15 × 10−2 2 h p.i.: 7.86 × 10−1 ±5.28 × 10−2 IL-10 8 h p.i.: 6.25 × 10−1 ±1.07 × 10−1 2 h p.i.: 4.10 × 10−1 ±2.65 × 10−2 IL-1β 8 h p.i.: 2.65 × 10−1 ±3.81 × 10−2 2 h p.i.: 7.68 × 10−2 ±8.90 × 10−3 CXCL2 8 h p.i.: 1.26 × 10−1 ±2.09 × 10−2 Male C57/BL6J TNFα 2 h p.i.: 2.98 × 10−2 ±2.77 × 10−2 DOB: 09/26/2010 2 h p.i.: 9.86 × 10−1 ±4.95 × 10−2 Filtered lysate from 2.18 × CXCL1 Splenocytes 8 h p.i.: 1.06 ±1.12 × 10−1 106 P388D cells/well 1 × 106/well (n = 3) 1 IFNγ 8 h p.i.: 1.62 × 10−3 ±1.89 × 10−4 Performed 11/2010 iNOS 8 h p.i.: 3.14 × 10−2 ±1.86 × 10−2 2 h p.i.: 5.83 × 10−1 ±5.77 × 10−2 IL-12A 8 h p.i.: 8.06 × 10−1 ±1.21 × 10−1 2 h p.i.: 6.90 × 10−1 ±5.98 × 10−2 IL-10 8 h p.i.: 5.52 × 10−1 ±3.95 × 10−2 2 h p.i.: 9.13 × 10−1 ±6.14 × 10−2 IL-1β 8 h p.i.: 8.67 × 10−1 ±1.60 × 10−1 2 h p.i.: 1.36 ±4.27 × 10−1 CXCL2 8 h p.i.: 8.95 × 10−1 ±1.58 × 10−1 TNFα 2 h p.i.: 7.96 × 10−1 ±1.50 × 10−1 2 h p.i.: 1.08 ±8.70 × 10−2 100 ng/ml LPS from E. coli CXCL1 8 h p.i.: 1.05 ±1.14 × 10−1 0111:B4/well IFNγ 8 h p.i.: 8.12 × 10−1 ±2.31 × 10−1 iNOS 8 h p.i.: 1.47 ±4.16 × 10−1 2 h p.i.: 9.99 × 10−1 ±7.46 × 10−2 IL-12A 8 h p.i.: 1.44 ±1.38 × 10−1 2 h p.i.: 7.97 × 10−1 ±6.37 × 10−2 IL-10 8 h p.i.: 5.98 × 10−1 ±8.08 × 10−2 Continued

Table 16. Means for qRT-PCR results for mouse splenocyte cytokine induction studies.

All qRT-PCR results are normalized by gapdh transcript levels. DOB, date of birth. KO, knockout. Std dev, standard deviation. If not marked, transcript levels were 8 h p.i.

180

Table 16 continued

Mouse sample Inoculation Gene Means Std dev IL-1β 1.46 ±4.60 × 10−1 750 µg/ml N. sennetsu CXCL2 9.00 ±2.32 frozen bacteria/well IFNγ 1.49 × 10−3 ±1.80 × 10−4 iNOS 2.35 ±6.92 × 10−1 IL-1β 2.71 ±5.57 × 10−1 250 µg/ml N. sennetsu outer CXCL2 2.71 × 101 ±4.77 Male C57/BL6J −3 −4 membrane/well IFNγ 5.48 × 10 ±7.27 × 10 DOB: 09/23/2010 −1 −1 iNOS 4.97 × 10 ±1.84 × 10 Splenocytes IL-1β 1.46 × 10−1 ±3.64 × 10−2 1 × 106/well (n = 3) PBS treated with the CXCL2 6.30 × 10−2 ±1.44 × 10−2 Performed 12/2010 b Sarkosyl protocol/well IFNγ 2.52 × 10−3 ±4.30 × 10−4 iNOS 1.42 × 10−2 ±1.39 × 10−2 IL-1β 1.24 ±1.65 × 10−1 500 ng/ml LPS from E. coli CXCL2 1.66 ±2.71 × 10−1 0111:B4/well IFNγ 9.83 × 10−1 ±1.32 × 10−1 iNOS 2.27 ±8.26 × 10−1 6 µg/ml N. sennetsu frozen IL-1β 1.14 bacteria/well CXCL2 9.08 × 10−1 30 µg/ml N. sennetsu IL-1β 1.58 bacteria/well CXCL2 1.41 Male C57/BL6J 2 µg/ml N. sennetsu outer IL-1β 4.72 × 10−1 DOB: 09/23/2010 −1 membrane/well CXCL2 2.73 × 10 Splenocytes 10 µg/ml N. sennetsu outer IL-1β 5.35 × 10−1 1 × 106/well (n = 1) membrane/well CXCL2 4.71 × 10−1 Performed 12/2010 b PBS treated with the IL-1β 2.17 × 10−1 Sarkosyl protocol/well CXCL2 6.79 × 10−2 1 µg/ml LPS from E. coli IL-1β 8.27 × 10−1 0111:B4/well CXCL2 6.05 × 10−1 Continued

181

Table 16 continued

Mouse sample Inoculation Gene Means Std dev 50 µg/ml N. sennetsu frozen IL-1β 1.47 ±3.58 × 10−1 bacteria/well CXCL2 1.56 ±1.50 × 10−1 IL-1β 6.56 × 10−1 ±1.67 × 10−1 Male C57/BL6J 10 µg/ml rNsp2/well CXCL2 4.91 ±1.72 DOB: 09/23/2010 IL-1β 3.79 × 10−1 ±7.35 × 10−2 Splenocytes 10 µg/ml rNsp3/well CXCL2 4.63 × 10−1 ±1.08 × 10−1 1 × 106/well (n = 3) IL-1β 1.37 × 10−1 ±1.74 × 10−2 Performed 02/2011 RPMI/well CXCL2 6.46 × 10−3 ±1.13 × 10−3 100 ng/ml LPS from E. coli IL-1β 9.95 × 10−1 ±1.45 × 10−1 0111:B4/well CXCL2 1.29 ±8.16 × 10−2 50 µg/ml N. sennetsu frozen IL-1β 1.19 ±6.40 × 10−2 bacteria well CXCL2 3.77 × 10−1 ±5.50 × 10−2 IL-1β 9.83 × 10−1 ±8.23 × 10−2 Male TLR KO 10 µg/ml rNsp2/well CXCL2 5.65 ±7.22 × 10−1 DOB: 07/14/2010 IL-1β 8.38 × 10−1 ±3.55 × 10−2 Splenocytes 10 µg/ml rNsp3/well CXCL2 5.36 × 10−1 ±2.30 × 10−2 1 × 106/well (n = 3) IL-1β 3.46 × 10−1 ±4.70 × 10−2 Performed 02/2011 RPMI/well CXCL2 2.73 × 10−2 ±4.14 × 10−3 100 ng/ml LPS from E. coli IL-1β 1.42 ±2.08 × 10−1 0111:B4/well CXCL2 3.19 ±3.13 × 10−1

182

Mouse sample Inoculation Gene Means Std dev IL-1β 6.08 × 10−3 ±3.75 × 10−4 CXCL2 1.80 × 10−1 ±1.14 × 10−2 N. sennetsu bacteria freshly TNFα 9.90 × 10−2 ±6.68 × 10−3 isolated from IL-1α 1.31 × 10−2 ±6.62 × 10−4 6 −1 −3 7 × 10 P388D1 cells/well CXCL1 1.59 × 10 ±8.14 × 10 iNOS 4.25 × 10−3 ±2.16 × 10−3 IL-12A 1.61 × 10−3 ±2.89 × 10−4 IL-10 1.43 × 10−1 ±1.93 × 10−3 IL-1β 1.12 × 10−3 ±8.98 × 10−5 Female C57/BL6J CXCL2 4.81 × 10−3 ±6.82 × 10−4 DOB: 04/13/2010 TNFα 1.26 × 10−2 ±1.74 × 10−3 BMDMs Filtered lysate from 7 × 106 IL-1α 2.96 × 10−3 ±5.40 × 10−5 5 −2 −4 4.5 × 10 /well P388D1 cells/well CXCL1 4.48 × 10 ±6.19 × 10 (n = 2) iNOS 9.88 × 10−4 ±8.01 × 10−4 Performed 08/2010 IL-12A 4.01 × 10−5 ±4.80 × 10−5 IL-10 1.18 × 10−1 ±1.00 × 10−2 IL-1β 9.88 × 10−1 ±2.77 × 10−2 CXCL2 1.27 ±1.43 × 10−2 TNFα 1.15 ±2.42 × 10−2 50 µg/ml LPS from E. coli IL-1α 1.05 ±6.31 × 10−3 0111:B4/well CXCL1 8.90 × 10−1 ±1.57 × 10−1 iNOS 6.03 × 10−1 ±8.92 × 10−2 IL-12A 2.90 ±7.66 × 10−1 IL-10 7.94 × 10−1 ±8.02 × 10−2 Continued

Table 17. Means for qRT-PCR results for mouse BMDM cytokine induction studies.

DOB, date of birth. KO, knockout. Std dev, standard deviation. NA, not available (only one result). All qRT-PCR results are 2 h p.i. and are normalized by gapdh transcript levels.

183

Table 17 continued

Mouse sample Inoculation Gene Means Std dev IL-1β 2.51 × 10−3 ±1.20 × 10−3 CXCL2 5.87 × 10−2 ±3.16 × 10−2 N. sennetsu bacteria freshly TNFα 2.58 × 10−2 ±2.51 × 10−3 isolated from IL-1α 1.10 × 10−2 ±1.01 × 10−3 6 −2 −2 7 × 10 P388D1 cells/well CXCL1 9.56 × 10 ±1.08 × 10 iNOS 5.82 × 10−4 ±9.38 × 10−5 IL-12A 1.28 × 10−3 ±1.55 × 10−4 IL-10 2.36 × 10−1 ±5.74 × 10−2 IL-1β 2.49 × 10−3 ±2.64 × 10−4 Male MyD88 KO CXCL2 1.52 × 10−2 ±4.68 × 10−4 DOB: 09/26/2010 TNFα 1.47 × 10−2 ±4.11 × 10−4 BMDMs Filtered lysate from 7 × 106 IL-1α 2.29 × 10−3 ±3.77 × 10−4 5 −2 −3 4.5 × 10 /well P388D1 cells/well CXCL1 6.73 × 10 ±3.75 × 10 (n = 2) iNOS 1.57 × 10−4 NA Performed 08/2010 IL-12A 4.66 × 10−4 ±1.73 × 10−4 IL-10 1.59 × 10−1 ±7.93 × 10−3 IL-1β 1.08 × 10−1 ±8.45 × 10−3 CXCL2 7.07 × 10−2 ±2.07 × 10−2 TNFα 2.62 × 10−1 ±2.05 × 10−2 50 µg/ml LPS from E. coli IL-1α 2.10 × 10−1 ±1.67 × 10−2 0111:B4/well CXCL1 6.58 × 10−2 ±1.60 × 10−3 iNOS 9.87 × 10−2 ±5.80 × 10−3 IL-12A 6.36 × 10−2 ±1.29 × 10−2 IL-10 1.11 × 10−1 ±7.74 × 10−3 Continued

184

Table 17 continued

Mouse sample Inoculation Gene Means Std dev IL-1β 9.93 × 10−2 ±8.24 × 10−3 50 µg/ml N. sennetsu CXCL2 1.26 ±2.21 × 10−2 frozen bacteria/well IL-12A 2.46 × 10−1 ±7.42 × 10−3 IL-1β 1.43 × 10−1 ±1.78 × 10−3 10 µg/ml N. sennetsu outer CXCL2 1.36 ±1.55 × 10−1 membrane/well IL-12A 3.36 × 10−1 ±9.25 × 10−2 IL-1β 2.47 × 10−3 ±1.76 × 10−4 Male C57/BL6J 7.5 µg/ml rP51/well CXCL2 5.16 × 10−2 ±7.24 × 10−3 DOB: 09/23/2010 IL-12A 1.18 × 10−2 ±9.18 × 10−3 BMDMs IL-1β 5.39 × 10−3 ±5.37 × 10−4 2 × 105/well (n = 3) 0.75 µg/ml rNsp2/well CXCL2 5.43 × 10−2 ±2.61 × 10−3 Performed 02/2011 IL-12A 4.22 × 10−3 ±4.00 × 10−3 IL-1β 5.94 × 10−4 ±7.25 × 10−5 PBS treated with the CXCL2 4.51 × 10−3 ±1.91 × 10−3 Sarkosyl protocol/well IL-12A 5.77 × 10−3 ±5.47 × 10−3 IL-1β 1.02 ±1.18 × 10−1 100 ng/ml LPS from E. CXCL2 1.32 ±3.16 × 10−1 coli 0111:B4/well IL-12A 1.40 ±2.19 × 10−1 IL-1β 1.05 × 10−1 ±7.63 × 10−3 50 µg/ml N. sennetsu CXCL2 8.45 × 10−1 ±7.02 × 10−2 frozen bacteria/well Male TLR KO IL-12A 7.83 × 10−3 ±6.39 × 10−3 DOB: 07/14/2010 IL-1β 1.73 × 10−2 ±2.15 × 10−3 10 µg/ml N. sennetsu outer BMDMs CXCL2 9.12 × 10−2 ±1.75 × 10−2 membrane/well 2 × 105/well (n = 3) IL-12A 7.84 × 10−4 ±1.01 × 10−3 Performed 02/2011 IL-1β 1.59 × 10−2 ±3.98 × 10−3 PBS treated with the CXCL2 5.29 × 10−2 ±3.80 × 10−3 Sarkosyl protocol/well IL-12A 5.55 × 10−4 ±7.69 × 10−4

185