AN ABSTRACT OF THE DISSERTATION OF

Jamie L. Everman for the degree of Doctor of Philosophy in Microbiology presented on December 9, 2014.

Title: Disease Models and Infectious Phenotypes of Mycobacterium avium subspecies paratuberculosis

Abstract approved: ______Luiz E. Bermudez

Mycobacterium avium subspecies paratuberculosis (MAP) is the causative agent of Johne’s disease, a chronic inflammatory bowel disease that affects ruminant populations worldwide. The characteristic stages of the disease make diagnosis difficult, resulting in silent transmission among animals in a herd for years before proper detection of the infection. The extensive prevalence of Johne’s disease has driven a continuous effort to more readily understand the pathogenesis of the bacterium and to develop more effective preventative measures to curb the spread of the disease within herds. In this dissertation, we aim to create a more effective model for studying MAP infection within the intestinal mucosa, to utilize the milk-induced virulence phenotype to study how opsonization affects cell infection, and to study the metabolic interaction between MAP and the host phagocyte during infection. We describe a novel in vitro cell culture passage model which indicates that MAP changes during passage between bovine epithelial cells and macrophages, developing a more pro-inflammatory phenotype. We show that the inflammatory MAP phenotype not only increases gene expression of lipid metabolism- and modification-related genes, but that it is also composed of a set of lipids that are unique to the phenotype. Ultimately, we were able to identify these inflammatory- related transcripts in naturally MAP-infected bovine tissues, thus validating our

model and indicating that the changing MAP phenotype may be a contributing factor in driving the development of inflammation within MAP infected animals. By using a different infectious phenotype that develops after MAP exposure to milk, a reservoir and transmission source of the bacterium, we demonstrate that opsonization of MAP results in more efficient translocation across an epithelial monolayer. Upon infection, we determine that macrophages more readily kill opsonized MAP in a rapid and specific manner. Furthermore, we begin to characterize one of the highest upregulated genes in this milk-induced phenotype, MAP1203, and its interaction with the intestinal epithelium. We establish that the putative cell wall associated protein is involved in both binding to and invasion of bovine MDBK epithelial cells when over-expressed in MAP during infection. Together, these data indicate the importance of the infectious phenotype developed after milk exposure and its role in the pathogenesis and transmission of Johne’s disease. Finally, we utilize Acanthamoeba castellanii (amoeba) as a phagocytic host and describe the influence that MAP infection has on the metabolic activity of the cell. We detail how MAP stimulates the metabolism of the amoeba and how that stimulation directly mirrors the pattern of survival and the intracellular burden of MAP over the course of infection. We identify bacterial mutants that result in excessive or deficient stimulation of the metabolic activity within host cell and by utilizing phenotype arrays, we illustrate that amoeba change the use of specific carbon sources based on the MAP strain used for infection. These data aid in beginning to understand the bacterial mechanisms that drive the metabolic interactions between MAP and the phagocytic host. Our results describe novel model systems for studying Johne’s disease and further our understanding about the host-pathogen interactions that occur within the intestinal mucosa of infected cattle. We show that the shifting phenotypes of MAP may be important contributing factors for detection, diagnosis, and in driving the progression of Johne’s disease. These findings offer new methods and the

identification of new bacterial phenotype targets that could be used as the basis of developing more efficacious strategies in detecting and preventing Johne’s disease.

©Copyright by Jamie L. Everman December 9, 2014 All Rights Reserved

Disease Models and Infectious Phenotypes of Mycobacterium avium subspecies paratuberculosis

by Jamie L. Everman

A DISSERTATION

submitted to

Oregon State University

in partial fulfillment of the requirements for the degree of

Doctor of Philosophy

Presented December 9, 2014 Commencement June 2015

Doctor of Philosophy dissertation of Jamie L. Everman presented on December 9, 2014

APPROVED:

Major Professor, representing Microbiology

Chair of the Department of Microbiology

Dean of the Graduate School

I understand that my dissertation will become part of the permanent collection of Oregon State University libraries. My signature below authorizes release of my dissertation to any reader upon request.

Jamie L. Everman, Author

ACKNOWLEDGEMENTS

I would first and foremost like to thank Dr. Luiz Bermudez for allowing me to conduct my doctoral research in his lab, for continuous discussion and exchange of ideas, and for the constant support he has offered throughout my time in the lab. His lessons, mentoring, contributions, and guidance have been invaluable in shaping me into becoming the research scientist I am today. I would also like to thank my committee members Dr. Dan Rockey, Dr. Martin Schuster, Dr. Manoj Pastey, Dr. Lia Danelishvili and Dr. Ricardo Letelier for their support and input to my projects. I would like to extend my sincerest thanks to all of the Bermudez lab members past and present. To the graduate students, technicians, and research staff, Lia Danelishvili, Lmar Babrak, Sasha Rose, Brendan Jeffrey, Rashmi Gupta, Jessica Chinison, Michael McNamara, and Laura Hauck, your suggestions, brainstorming, help, and laughs made working in the lab a wonderful experience. To all of the undergraduates and professional veterinary students I have worked with, Erin Flannery, Navid Ziaie, Eugenio Mannucci, Sadie Rice, and Lucy Garcia Flores, your contributions to all our projects have been immense and are greatly appreciated. To all of the students, faculty, and administrative staff in Dryden hall, thanks you for being my sounding board and for all the friendship throughout the years. A big thanks to the Department of Microbiology, especially Mary Fulton, and the Department of Biomedical Sciences, specifically Beth Chamblin, Denny Weber, and Jayne Theurer, for providing assistance with all the paperwork, travel, and official business, and for listening to and helping me with all my questions and concerns. Many thanks go to the Department of Veterinary Medicine for technical advice and guidance on bovine sample preparation, blood collection, milk collection, and general veterinary techniques. The following personnel were wonderfully forgiving with all my questions and early morning interruptions, and offered a friendly face and suggestions during various stages of my projects: Robert Murray (LARC), Dr. Helen Diggs and Dr. Jennifer Grossman (Laboratory Research

Veterinarians), Teresa Sawyer (TEM), Darlene Joyner (LARC), Hayden Bush and Jeff Behm (OSU Dairy Center), and the Histology laboratory. To all the members of every softball, soccer, dodgeball, ultimate frisbee team, and gym class I have been a part of while in Corvallis, for all the fun memories and for the maintenance of my sanity, I give my sincerest thanks. Last but certainly not least, I would like to extend my love and immense gratitude towards my family. Mom and Dad, thanks for lending words of confidence during the tough days and for pretending to understand all my research when I got excited talking about it when it worked. Finally, to my wonderful boyfriend Brian, thank you so much for standing by me for all these years all the way from Colorado, it has been a long road but your encouragement and support has been invaluable.

CONTRIBUTION OF AUTHORS

Chapter 2: Jamie L. Everman conducted the experiments, contributed to experimental design, preparation of the manuscript, and data analysis. Dr. Luiz E. Bermudez contributed to the experimental design, data analysis, manuscript preparation, and funding of the project.

Chapter 3: Jamie L. Everman conducted the experiments, contributed to experimental design, preparation of the manuscript, and data analysis. Dr. Torsten M. Eckstein performed and analyzed lipidomics and HPLC-ES/MS data. Dr. John Bannantine designed and produced the MAP K10 microarray slides and contributed to array protocols and experimental design. Jonathan Roussey and Dr. Paul Coussens contributed bovine tissue samples from their ongoing study at Michigan State University. Dr. Luiz E. Bermudez was involved in the experimental design, data analysis, manuscript composition, and funding of the project.

Chapter 4: Jamie L. Everman conducted the experiments, contributed to experimental design, preparation of the manuscript, and data analysis. Erin F. Flannery conducted the screening of the MAP libraries and optimization of metabolic activity assays. Eugenio U. Mannucci contributed to the experimental design and optimization of phenotype arrays with amoeba. Dr. Luiz E. Bermudez contributed to the experimental design, data analysis, manuscript preparation, and funding of the project.

Chapter 5: Jamie L. Everman conducted the experiments, contributed to experimental design, preparation of the manuscript, and data analysis. Lucero Garcia Flores conducted induction and purification assays of MAP1203 protein. Dr. Luiz E. Bermudez contributed to the experimental design, data analysis, manuscript preparation, and funding of the project.

Appendix 1: Jamie L. Everman contributed to experimental design, preparation of the manuscript, conducted experiments, and data analysis. Navid Ziaie and Jessica Bechler conducted experiments and contributed to the analysis of data and preparation of the manuscript. Dr. Luiz E. Bermudez contributed to the experimental design, data analysis, manuscript preparation, and funding of the project.

TABLE OF CONTENTS Page Chapter 1: Introduction and Background ...... 1 Classification of Mycobacterium avium subspecies paratuberculosis ...... 1 MAP Growth Characteristics ...... 2 Johne’s Disease Prevalence ...... 3 Zoonotic Potential of MAP ...... 4 Stages and Progression of Johne’s Disease ...... 6 Johne’s Diagnostics ...... 9 Johne’s Treatment and Prevention Efforts ...... 11 Models Used to Study Johne’s Disease and MAP Pathogenesis ...... 14 Intracellular Survival Mechanisms of MAP ...... 17 Shifting Bacterial Phenotypes During Infection ...... 19 Scope of Doctoral Dissertation ...... 21

Chapter 2: Antibodies against invasive phenotype-specific antigens increase Mycobacterium avium subspecies paratuberculosis translocation across a polarized epithelial cell model and enhance killing by bovine macrophages ...... 23

Abstract ...... 24 Introduction ...... 25 Methods and Materials ...... 28 Results ...... 33 Discussion ...... 39 Acknowledgements ...... 45 References ...... 55

Chapter 3: Characterization of the inflammatory phenotype of Mycobacterium avium subspecies paratuberculosis using a novel cell culture passage model .... 59

Abstract ...... 60 Introduction ...... 61 Materials and Methods ...... 64

TABLE OF CONTENTS (Continued) Page Results ...... 70 Discussion ...... 76 Acknowledgements ...... 84 References ...... 95

Chapter 4: Mycobacterium avium subspecies paratuberculosis induces change in the metabolism of phagocytic host Acanthamoeba castellanii during infection ...... 101

Abstract ...... 102 Introduction ...... 103 Methods and Materials ...... 105 Results ...... 110 Discussion ...... 117 Acknowledgements ...... 121 References ...... 133

Chapter 5: MAP1203 involved in binding and invasion of epithelial cells by Mycobacterium avium subspecies paratuberculosis ...... 136

Abstract ...... 137 Introduction ...... 138 Methods and Materials ...... 140 Results ...... 143 Discussion ...... 147 Acknowledgements ...... 148 References ...... 158

Chapter 6: Discussion and Conclusions ...... 159 Bibliography ...... 169 Appendices ...... 181

TABLE OF CONTENTS (Continued) Page Appendix 1: Additional Manuscripts Establishing Caenorhabditis elegans as a model for Mycobacterium avium subspecies hominissuis infection and intestinal colonization...... 182

Appendix 2: Abstracts of Additional Publications Appendix 2.1 Danelishvili L, JL Everman, MJ McNamara, and LE Bermudez (2009) Inhibition of the plasma-membrane­ associated serine protease cathepsin G by Mycobacterium tuberculosis Rv3364c suppresses caspase-1 and pyroptosis in macrophages. Front. Micro. 1-14 vol 2 ...... 207

Appendix 2.2 Bannantine JP, Everman JL, Rose SJ, Babrak L, Katani R, Barletta RG, Talaat AM, Gröhn YT, Chang Y-F, Kapur V and Bermudez LE (2014) Evaluation of eight live attenuated vaccine candidates for protection against challenge with virulent Mycobacterium avium subspecies paratuberculosis in mice. Front. Cell. Infect. Microbiol. 4:88. doi:10.3389/fcimb.2014.00088 ...... 208

Appendix 2.3 Danelishvili L, Babrak L, Rose SJ, Everman J, Bermudez LE. 2014. Mycobacterium tuberculosis alters the metalloprotease activity of the COP9 signalosome. mBio 5(4):e01278-14. doi:10.1128/mBio.01278-14 ...... 210

LIST OF FIGURES

Figure Page

1.1 Evolution of MAP from closely related Mycobacterium avium Complex members ...... 1

1.2 Inflammatory States of Johne’s Disease ...... 8

2.1 Effect of opsonization of milk-exposed MAP on invasion of MDBK epithelial cells ...... 46

2.2 Translocation of opsonized milk-exposed MAP across an MDBK epithelial transwell monolayer ...... 47

2.3 Immune serum opsonization of milk-exposed MAP initiates phenotype which decreases uptake of bacteria by BOMAC via cell-mediated killing ...... 49

2.4 Inhibition of cellular killing mechanisms and effect on uptake of opsonized MAP by BOMAC cells...... 50

2.5 Role of nitric oxide in the killing of immune serum opsonized MAP during BOMAC infection ...... 51

2.6 Role of macrophage extracellular traps (METs) in the killing of immune serum opsonized MAP during BOMAC infection ...... 52

2.7 Effect of opsonization on MAP survival in a murine model ...... 53

3.1 In vitro cell culture passage model ...... 86

3.2 Inflammatory response to MDBK epithelial cells during in vitro cell culture passage model ...... 87

3.3 Lipid profiled of MAP phenotypes ...... 89

3.4 Abundance of cell wall lipid Para-LP-01 in MAP phenotypes ...... 91

3.5 Bovine intestinal tissue samples ...... 92

3.6 MAP transcripts from infected bovine intestinal tissue ...... 93

4.1 Amoeba readily phagocytize MAP during infection ...... 122

4.2 Temporal pattern of MAP viability during amoeba infection ...... 123

4.3 Stimulation of amoeba metabolism by MAP infection ...... 124

4.4 Impact of MAP infection on metabolism of amoeba during long-term infection...... 125

4.5 Identification of MAP mutant strains exhibiting altered metabolic stimulation of amoeba ...... 126

4.6 Verification of MAP mutants using bovine macrophage infection ...... 127

4.7 MAP mutant strain growth dynamics within Acanthamoeba castellanii ...... 128

4.8 Carbon source metabolic utilization phenotypes of MAP and MAP mutant infected amoeba ...... 130

5.1 Construction of dominant-negative mutants of MAP1203 ...... 149

5.2 Induction of MAP1203 protein expression in M. smegmatis ...... 150

5.3 Role of MAP1203 in M. smegmatis uptake and survival within macrophages ...... 151

5.4 Role of MAP1203 in epithelial cell invasion by M. smegmatis ...... 152

5.5 Induction of MAP1203 protein expression in MAP cultures ...... 153

5.6 Role of MAP1203 in epithelial cell invasion by MAP ...... 154

5.7 Colony sizes of MAP containing pJAM::MAP1203 constructs ...... 155

5.8 Expression and purification of N-terminal and C-terminal 6xHN-tagged MAP1203 proteins ...... 156

6.1 Experimental findings described in the chapters of this dissertation ...... 167

LIST OF TABLES

Table Page

3.1 List of primer sets for bovine and bacterial transcript analysis ...... 85

3.2 Microarray analysis of MAP phenotypes ...... 88

3.3 Supplemental – Diagnostic scores of cattle with Johne’s disease ...... 94

4.1 Supplemental – Biolog carbon source plate layout of PM-M1 plate ...... 129

5.1 Primers used for creation of MAP1203 expression vectors ...... 148

LIST OF APPENDIX FIGURES

Figure Page

Appendix 1.1 C. elegans feed on MAH ...... 196

Appendix 1.2 MAH-td104 does not affect median or total lifespan of C. elegans ...... 197

Appendix 1.3 MAH colonization persists after pulse-chase analysis ...... 198

Appendix 1.4 MAH colonize lumen of C. elegans intestinal tract ...... 199

Appendix 1.5 Transmission electron microscopy of MAH-colonized C. elegans ...... 200

Appendix 1.6 Binding of HEp-2 cells and colonization of C. elegans by MAH ΔGPL/4B2 mutant...... 202

1

Chapter 1: Introduction

Classification of Mycobacterium avium subspecies paratuberculosis Mycobacterium avium subspecies paratuberculosis (MAP) is a member of the Mycobacterium genus classified within the phylum Actinobacteria. The Mycobacterium genus is composed of a variety of pathogenic and non-pathogenic species which are characterized by their lipid-rich bacterial cell wall that allows for acid-fast positive staining used readily for diagnostic purposes. MAP is grouped within the Mycobacterium avium Complex (MAC) which includes the slow-growing members Mycobacterium avium subspecies hominissuis (MAH), Mycobacterium avium subspecies avium (MAA), and Mycobacterium avium subspecies sylvaticum (MAS). Recent multilocus sequencing analysis has determined that contrary to previous phylogenetic trees that identify each member of MAC as having one shared common ancestor prior to evolving. MAH is, in actuality, the common node of MAC from which the ovine and bovine strains of MAP and the human and avian strains of MAA and MAS independently evolved (Figure 1.1) (Turenne, Collins et al. 2008).

Figure 1.1: Evolution of MAP from closely related Mycobacterium avium Complex members (Turenne, Collins et al. 2008)

Historically, mycobacterial species were classified based on their phenotypic traits, thus M. avium (now MAA/MAH) and M. paratuberculosis (now MAP) were considered two distinct species based on their vastly different growth rates and disease hosts. In 1988, it was suggested that due to the high DNA homology shown using hybridization assays, and that the two be classified as one species and as separate subspecies; thus, the nomenclature M. avium subspecies avium and M. 2 avium subspecies paratuberculosis began (Saxegaard and Baess 1988). The complete sequencing of the 4.8 megabase MAP genome has allowed for the further characterization of the similarities and differences between MAC members (Li, Bannantine et al. 2005, Bannantine, Li et al. 2014). Using published genomes, it has been calculated that MAH, MAA, and MAP are >97% genetically identical at the nucleotide level, and have 100% identical 16s ribosomal RNA nucleotide sequences (Bannantine, Zhang et al. 2003). This level of genetic similarity is astounding given the dramatic differences in growth rate, host range, and diseases caused by each of the 3 subspecies, with early research questioning if the uncommon genetic factors between the subspecies may be responsible for causing the dramatically different disease states between members of MAC.

MAP Growth Characteristics On a phylogenetic tree the genus Mycobacterium organizes into 2 distinct groupings: fast-growing species which generally tend to be lesser or non-pathogenic and take less than 7 days to culture, and slow-growing species of which members often have higher pathogenicity and require more than 7 days to grow in culture. MAP is a member of the latter group and is one of the slowest growing species of mycobacteria. Compared to a standard laboratory culture of Escherichia coli, which has a doubling time of ~20 minutes, laboratory strains of MAP grown in broth culture are demonstrated to have a single doubling time of ~24 hours in vitro and >36 hours in vivo (Elguezabal, Bastida et al. 2011). This extremely slow growth requires MAP laboratory strains to be grown for 4-6 weeks on solid medium, while environmentally acquired strains, isolates from animals, and diagnostic tests for MAP in fecal samples can require up to 4 months for growth and confident diagnostic results for MAP isolates. Enhanced growth is encouraged by the addition of multiple supplements to Middlebrook 7H10 agar used to classically culture mycobacteria in the lab. The addition of glycerol and casein add carbon and nitrogen to the growth media, creating a richer environment to facilitate growth. The supplement OADC is added and 3 contains oleic acid, a long-chain fatty acid for enhanced biosynthesis of the waxy, lipid-rich cells wall, albumin for increased protein concentration and protection from toxic compounds, dextrose for increased carbon metabolism, and catalase which contributes to the detoxification of peroxides and oxidants in the media (Rawat, Johnson et al. 2007, Vilcheze, Av-Gay et al. 2008, Zimbro 2009). All mycobacterial isolates are grown on the aforementioned cocktail of media; however, MAP requires an additional supplement for culture in the laboratory setting. Mycobactin J, a siderophore used for scavenging iron from the environment for use by the bacterium, is a required supplement for MAP growth. Genomic analysis indicates that this requirement is in part due to the absence of the gene entA, which serves as the first gene in the cluster of genes required for the synthesis of mycobactin during growth (Li, Bannantine et al. 2005). In part due to its inability to synthesize mycobactin for iron acquisition, this trait renders MAP inefficient at replicating in the environment.

Johne’s Disease Prevalence MAP is the etiologic agent responsible for Johne’s disease among cattle, goats, sheep, deer, camelids, and other ruminants in both captivity and the wild. Also referred to as paratuberculosis, Johne’s disease is a wide spread and highly prevalent infection around the world. In the United States alone it is estimated that the disease has a $250 million - $1.5 billion economic impact on the cattle industry due to loss of milk production, removal of animals from production, and the culling of infected animals (Stabel 1998, Ott, Wells et al. 1999). Johne’s disease presents a global threat as the identification of the disease has been reported in nearly every country. The economic trade and distribution of animals, as well as the natural movement of wildlife populations, greatly contribute to the spread and transmission of the disease. The true prevalence of Johne’s disease in the United States is difficult to accurately calculate due to many factors. Similar to other bovine diseases such as bovine tuberculosis, caused by Mycobacterium bovis, and bovine spongiform encephalopathy, caused by misfolded prion proteins, the detection of MAP is required to be reported to the United States Department of Agriculture’s National Animal 4

Health Reporting System. However, testing for the presence of MAP within older, more established herds, and newborn calves is completed at the discretion of farm owners. Thus, many cattle ranchers choose to not know their Johne’s disease status rather than absorb the cost of testing, potential early culling of cattle, or suffer the public stigma from consumers after being identified as a Johne’s positive farm. Nevertheless, education programs have been relatively successful in encouraging and increasing the testing prevalence for MAP within the United States over the past few decades. Using currently reported rates and disease models, it is estimated that over 68% of dairy herds within the United States are infected with Johne’s disease, and that ~10% of beef herds carry the disease (2014).

Zoonotic Potential of MAP While Johne’s disease has been attributed to ruminants since 1895 upon its discovery by the German veterinarians Heinrich Johne and L. Frothengham (Johne 1895), it has also been suggested to play a role in a variety of human diseases. Though controversial, MAP infection has been suggested as a contributor to Crohn’s disease, a similar inflammatory enteric disease in humans. MAP was first isolated from a human patient exhibiting Crohn’s disease in 1985 (Chiodini, Van Kruiningen et al. 1984). It has been noted that MAP can been isolated from patients with Crohn’s or Irritable Bowel Syndrome, though inconsistency drives the skepticism of its exact role as recovery rates from patients range from 0% to 100% depending on the study (Sartor 2005, Frank, St Amand et al. 2007). Anecdotal evidence and proponents of this hypothesis claim that anti-mycobacterial therapies can provide relief to those suffering from the symptoms of Crohn’s disease. One study supports anti­ mycobacterial treatment as a cure for Crohn’s disease, though the study did not include appropriate controls to come to such definite conclusions (Shafran, Kugler et al. 2002). On the other hand, Selby et al demonstrated that though treatment with an anti-mycobacterial cocktail of clarithromycin, rifabutin, and clofazimine provided transient relief which was attributed to their wide-spectrum antibacterial properties, the treatment did not provide long-term relief against the symptoms of the intestinal 5 inflammatory diseases compared to treatment with a placebo (Selby, Pavli et al. 2007). Followed by even more controversy, additional studies have offered the hypothesis that infection with MAP could trigger the neurologic disease multiple sclerosis (MS). The proteins MAP_2694 and FprB have homology to the human NMDA receptor and myelin, respectively, found along neuronal synapses and axons suggesting the cross-reactivity of antibodies from infection could lead to the autoimmune reaction seen in patients with MS (Cossu, Cocco et al. 2011, Cossu, Masala et al. 2012). Alternatively, studies suggest that patients suffering from Crohn’s disease, having a deteriorated intestinal lining and a disrupted microbiome (Manichanh, Rigottier-Gois et al. 2006, Dicksved, Halfvarson et al. 2008) may simply be more susceptible to MAP infection, as immunocompromised patients are more susceptible to other mycobacterial diseases such as MAH and tuberculosis. To support the hypothesis that a disrupted microbiome is in part to blame for Crohn’s symptoms, trials utilizing the treatment of patients with probiotics, prebiotics, and antibiotic therapy to re-establish a healthy balance of intestinal flora have been proven to be successful therapies for combating Crohn’s symptoms (Sartor 2004). While MAP may be associated with the disease, much more evidence exists in favor of other more well understood and previously described enteric pathogens being the culprit. The overgrowth of enteric bacteria such as invasive Escherichia coli within the ileum (Neut, Bulois et al. 2002), as well as the shift in abundance of Bacteroides, Enterococcus, and Klebsiella species has been widely studied as a factor in other inflammatory bowel diseases (Sartor 2004). While studies are divided on how MAP may contribute to the development, triggering, or sustainment of Crohn’s disease, more research with appropriate research designs must be completed to truly understand the role of MAP in human Crohn’s disease. While the direct zoonotic potential of MAP remains unclear, it is mandated by federal and state bylaws of the United States Department of Agriculture (USDA) that animals that either test positive or have been treated for MAP infection be removed from production lines of meat and dairy products meant for human consumption. The 6 guidelines are set, in part, due to the ability to isolate MAP from both muscle tissue and milk samples from infected cattle (Taylor, Wilks et al. 1981, Chiodini and Hermon-Taylor 1993, Alonso-Hearn, Molina et al. 2009). It was proven that MAP is not only found within the milk of symptomatic animals (Taylor, Wilks et al. 1981), but asymptomatic dairy cattle as well (Sweeney, Whitlock et al. 1992), and conflicting reports detail the effect of pasteurization on the inactivation of MAP in dairy samples. A handful of studies indicate that small-batch pasteurization treatments used on farms for feeding of young animals are capable of eliminating any viable MAP organisms (Stabel 2001), and that MAP can only survive the pasteurization process used for cheese production (Stabel and Lambertz 2004). Other studies support that MAP survive all laboratory-simulated pasteurization processes (Chiodini and Hermon-Taylor 1993) and can also be detected by both PCR and plate- grown culture methods in naturally-infected commercially-processed milk samples (Grant, Hitchings et al. 2002). Further research is required to fully understand the role, if any, as a causative or contributing factor of human Crohn’s disease. A more thorough understanding of the resiliency of MAP to decontamination efforts is also required to more accurately and safely understand the survival of MAP during food processing and to determine whether it poses a problem in the manufacturing of products meant for human consumption. It is important to note, however, that currently, manufacturers and food producers are confident in their ability to provide healthy and safe products free of MAP to both beef and dairy consumers in the United States.

Stages and Progression of Johne’s Disease Young animals are more susceptible to becoming infected with Johne’s disease, and can acquire the infection by 1 of 3 main transmission routes: in utero, via ingestion of colostrum and milk from an infected dam, or from fecal contamination of water and feed. Upon infection, young animals do not outwardly appear sick or ill. This silent stage of the disease can last from months to years, with the animal serving as host to the bacterium but not appearing to be infected. Little to no inflammation 7 within the intestine of the animal is observed during necropsy and those in the early stages of infection have been reported to have high levels of both tissue-localized, and circulating anti-inflammatory markers including the potent immunosuppressive signal IL-10 (Weiss, Evanson et al. 2005, Weiss, Evanson et al. 2006). During this initial disease stage, animals usually do not test positive by current serum measurement testing standards, nor do they shed any detectable amount of bacteria in their feces. At a certain undetermined time, infected animals shift to the next disease stage, termed subclinical, where the animal still does not exhibit a high degree of inflammation but is able to shed MAP in its feces and is able to transmit the bacterium to other animals. The identification of such animals at this stage of infection via diagnostics involving fecal culture, PCR identification, or antibody ELISA are often inconsistent as the accuracy of the test depends on how far the animal is into the subclinical stage, if they are actively shedding MAP in the feces, or if the animal has yet to mount an antibody response in their serum. The majority of animals infected with Johne’s disease is in this dangerous subclinical state of infection and can remain there for many years. Unable to be accurately diagnosed, these animals appear externally healthy though are actively shedding bacteria and serving as a dangerous carrier of the disease among otherwise healthy herds of animals for years before transitioning to the next stage of the disease. The next stage, after an unknown trigger or specific timepoint after infection, is when animals transition from the subclinical into the clinical stage of Johne’s disease. Clinical and advanced clinical animals are those most readily identifiable as being infected with Johne’s disease. A shift in the inflammatory state of the animal is triggered, leading to a surge in pro-inflammatory signals within the animal, both systemically as well as within the intestinal mucosa. Infiltrating tissue macrophages meant to engulf and rid the animal of MAP infection instead form granulomatous lesions within the mucosal tissue. There lesions spread throughout the tissue, leading to the destruction of the intestinal barrier and the fusion of the villi of the epithelium. Together, these changes make up the severe inflammation within the intestinal tissue seen within clinically infected animals (Figure 1.2). The severe intestinal 8 inflammation causes the inability of the infected animal to take up nutrients from its diet and results in profuse diarrhea, dramatic weight loss, and visible wasting of the animal.

Figure 1.2: Inflammatory States of Johne’s Disease Early stages of Johne’s disease (silent and subclinical) are characterized by a lack of inflammation and visible signs of disease (left panel). Later stages (clinical and severe) are well described to have severe inflammation, presence of MAP-containing granulomas, fused or destroyed villi, and influx of immune cells (right panel). (Chacon, Bermudez et al. 2004) 9

To complicate subclinical and clinical infections, a phenomenon can occur leading to the presence of ‘super-shedders’ within a herd. The animals deemed super-shedders are difficult to detect as current diagnostics enumerate fecal bacteria counts on agar slants on a scale of 0, <20, <50, or >50. Super shedders are often missed as their samples fall on the spectrum of >50 colony forming units during diagnosis. In comparison to a clinically infected animal sample that can result in 50 – 100 colonies on an agar slant, super-shedders can result in samples that produce thousands of colonies on an agar slant. While in the subclinical stage of disease and appearing to be healthy to an observer, these animals are capable of shedding upwards of 50 billion bacteria per day; making the act of preventing fecal cross-contamination of water and feed nearly impossible. These confounding factors are why it is important to develop more accurate diagnostics with enhanced detection limits on the upper and lower spectrum of the disease in order to identify animals in the silent and subclinical stages of Johne’s, determine the severity of their disease state, and to prevent the passage of MAP to young uninfected animals.

Johne’s Diagnostics Currently, there are a variety of methods that can be used to detect and diagnose Johne’s disease and MAP infection in cattle, each with its own set of advantages and disadvantages. Due to the staged progression of Johne’s disease and the silent nature of early infection, young animals are the most difficult to diagnose. It is suggested, and often required to wait to test animals for MAP infection until 18 months of age as testing earlier produces inconclusive test results. The first types of diagnostic tests involve identification of the bacterium directly via culture or identification of bacterial DNA in suspected samples. Based on National Veterinary Services Laboratory protocols, suspected MAP-containing samples, including soil, tissue, and fecal samples, are processed by using a multi-step process which includes treatment with hexadecylpyridinium chloride, and the antibiotics amphotericin B, nalidixic acid, and vancomycin prior to culture on Middlebrook 7H10 agar supplemented with mycobactin J or on Herrold’s Egg Yolk Agar supplemented with mycobactin J and the cocktail of amphotericin B, nalidixic 10

acid, and vancomycin (PDYKEMA 2011). The disadvantages of bacterial culture are that the organism must be viable for identification, samples must be incubated and stored for up to 4 months to confidently indicate the presence or absence of MAP, and long incubation times can lead to contamination with other organisms from fecal or environmental samples. One benefit of fecal sampling is that it can be completed using pooled samples, which dramatically decreases the cost of the diagnostic testing and enables the ability to test larger herds at a single time (Wells, Godden et al. 2003). Upon growth of colonies, the bacterial species is confirmed using polymerase chain reaction (PCR), which can verify the colonies are MAP and eliminate the possibility of other closely related mycobacterial species. Currently, PCR identification tests utilize the MAP insertion sequence IS900 which is specific to MAP and only a handful of other relatively rare Mycobacterium avium subspecies avium strains (Cousins, Whittington et al. 1999). Bacterial gene-based detection can also be done in the absence of culture on DNA samples extracted directly from fecal or tissue samples. These tests are more sensitive as the bacteria do not have to be viable in a sample, only present for DNA to be detectable, although PCR detection can also be inconsistent between samples. Diagnostics based on the detection of MAP in samples are limited by the stage at which the animal is at during the infection. Animals must be shedding MAP in fecal samples in order for detection to be possible and silent and subclinical-infected animals often do not shed MAP in the feces or do so at a level that is below the level of detection using these methods. Tissue samples offer much more reliable results, but are only available upon necropsy following the death or culling of an animal suspected of being infected with Johne’s disease. The second method most often used for diagnosis of MAP infection is the detection of the presence of antibodies against MAP in the serum and blood of animals. Enzyme-linked Immunosorbant Assays (ELISA) is the most sensitive test for detection of infection within a herd (Sockett, Conrad et al. 1992, Whitlock, Wells et al. 2000). The assays utilize MAP antigens and can quantify the antibody titer and level of protection an animal has mounted to the bacterium. This method of detection is one of the most simple, as materials are relatively inexpensive and can be 11

completed with multiple samples in a short amount of time, and only a small amount of blood from cattle to be tested is required. ELISA detection is also one of the better tests to use as an animal can have circulating antibodies against MAP much sooner than when they begin to shed bacteria in the feces, allowing for earlier detection of animals in the silent or subclinical stages of infection. However, antibody detection also has an array of disadvantages. Repeated testing of the same animal sample has shown significant variability in the results which make definitive diagnostics difficult for herd management decisions (Adaska, Munoz-Zanzi et al. 2002). The tests can provide inconsistent results depending on the antibody titer, and the time elapsed since last stimulation of humoral protection, and many animals exhibit a delayed or absent humoral response to MAP infection, even during the clinical stages of the disease (Chiodini, Van Kruiningen et al. 1984). Prior immunization for Johne’s disease, the administration of the PPD skin test for bovine tuberculosis (Bastida and Juste 2011), or the presence of anti-Mycobacterium bovis antibodies from either natural infection or immunization, also interfere with the serum antibody detection ELISA tests for Johne’s disease (Marassi 2005). Thus, it is important to know the health status of the animal being tested prior to choosing the appropriate diagnostic test to enhance the sensitivity and specificity of the test and its results.

Johne’s Treatment and Prevention Efforts Johne’s disease does not currently have a cure and while antibiotic therapy can decrease the appearance of disease characteristics within infected animals, it has a long history of ineffectiveness. Historically, treatments of mycobacterial diseases were characterized by the use of multi-antibiotic cocktails which came with high costs and lengthy required treatment times where patient compliance or frequent and repetitive dosing schedules can often be problematic. Since MAP treatment is plagued by these difficulties, and as there is no approved treatment or drug for Johne’s disease within the United States, it is generally advised to abide by herd management guidelines and cull infected animals in order to remove them from the healthy herd to prevent further spread of the infection. In the case of economically valuable animals such as those with genetic value for reproductive purposes, 12

treatment is often based on the suggestion of the veterinarian. Interestingly, drug treatments for MAP are more readily based on anecdotal clinical evidence as in vitro drug efficacy against MAP is often not indicative of efficacy in vivo (St. Jean 1996, Krishnan, Manning et al. 2009, Collins 2014). The suggested treatment for MAP infection is most often a drug cocktail of isoniazid and rifampin which must be administered daily for the remaining lifespan of the animal (St. Jean 1996). This treatment proves to be time intensive and costly and serves as a deterrent for treatment of animals with the disease. While it is suggested to use additional drugs such as amikacin or clofazimine in tandem with the cocktail to ensure efficacy of at least 2 antibiotics during the course of treatment of an infected animal, these drugs are extremely costly for long-term treatment options (St. Jean 1996). More recent studies on the antibiotic monensin have shown that treatment with the drug can lessen the seropositivity of animals (Hendrick, Duffield et al. 2006), and can lessen the tissue colonization of MAP in calves fed milk containing the antibiotic (Fecteau and Whitlock 2011). Interestingly, monensin, sold under the retail name Rumensin® is an additive already included in cattle feed to prevent a variety of diseases and improve weight gain (Elanco; Greenfield, IN). The role of monensin in feed in its current levels and its role in MAP pathogenicity has not been established and while current studies indicate it has a protective effect, research has yet to be completed on the effect the drug has on shedding and transmission of MAP from infected animals. Rather than antibiotic treatment, prophylactic measures are more commonly encouraged to prevent the incidence of MAP infection in a herd. Animals within a herd may be vaccinated in order to prevent the spread of MAP between animals. Cattle owners can opt to vaccine their herds with Mycopar, the only approved vaccine in the United States, which is an inactivated whole-cell vaccine. While vaccination of animals within a herd is an option, there are multiple requirements that must be met, and many disadvantages to take into consideration. Owners must get permission from a state accredited veterinarian to obtain vaccine dosages, and they must be able to prove that MAP infection is already present in the herd meant to be vaccinated. The only animals allowed to be vaccinated are replacement heifers and bulls and the animals must be between 1 and 35 days old in order to be considered for preventative 13

treatment. As for the mode of action of the US-approved vaccine, Mycopar does not protect against acquisition of the infection, nor does it prevent the shedding of bacteria in the feces. Rather, the vaccine prevents the clinical disease symptoms such as severe inflammation and body wasting from progressing to the advanced stage (2014, Collins 2014). This in turn limits the amount of bacteria shed in the stool but by no means stops shedding from occurring. Lastly, the vaccination of animals with whole-cell inactivated MAP produces antibodies which are semi-protective in nature. However, this immune response results in the inability to use ELISA detection as a method of diagnosing MAP within the herd as the antibodies in the serum create a false-positive reaction (Bastida and Juste 2011). Thus, vaccinated herds must utilize fecal culture or other methods of detection of live bacteria in place of serum detection methods and diagnostics. These limitations and the search for a more complete and effective vaccine are what drive current research efforts to develop a more optimal vaccine for protection against Johne’s disease. As outlined above, antibiotic therapy and vaccination can be expensive and ineffective at preventing the spread of Johne’s disease both within a farm and between animal facilities. Therefore, it is strongly encouraged by veterinarians and animal practitioners to follow strict herd management guidelines which are provided by the USDA and Johne’s risk management teams (Godden 2008, 2014, Collins 2014). Using these guidelines alone, it is believed that a farm can prevent transmission of disease to young animals much more effectively than the sole use of vaccination and antibiotic treatment options, and are much easier and more cost and time effective to implement within a dairy operation. The USDA guidelines are extensive, though the basic strategy entails removing young animals from whole herds upon birth and ensuring that these young susceptible animals are housed in clean pens free of any chance of interacting with MAP-contaminated feed and water. Furthermore, as feeding of colostrum to young animals within the first days of life is an important and strictly followed practice, it is of the utmost importance that young animals must be weaned and fed colostrum, and later milk, from dams that are free of MAP in order to prevent transmission (Godden 2008). 14

In addition to herd management guidelines, it is highly encouraged to test for MAP before acquiring and introducing new animals to a new herd. Furthermore, if the Johne’s-positive status of an animal is known, it is illegal to move said animal across state lines in order to halt the spread of the disease (Collins 2014).

Models Used to Study Johne’s Disease and MAP Pathogenesis

As with the majority of pathogens that infect both humans and animals, the infectious processes of MAP that occur within the ruminant host are complicated, multi-factorial, and occur in a variety of stages. These factors make it difficult to unravel molecular mechanisms occurring at each particular stage of Johne’s disease progression in the ruminant host environment. While cattle, goats, and other large ruminant models can and are often used to study Johne’s disease as they are the natural host, they have a distinct set of disadvantages. Large animal models such as cattle, goats, and sheep require large amounts of space for the proper enclosures and pens of the animals and that space must be suitable for the biosafety containment standards of the pathogen of interest. Daily costs for the care of the animals are a significant factor in experimental design as the number of animals and the length of the incubation of infection can increase quickly for a long term, multi-variable study. This factor is unavoidable for studying Johne’s disease, as the immunological and pathological mechanisms that occur during the clinical stage of disease are the most studied and take years to develop. A recent published review of all ruminant studies involving MAP infection over the past century indicated that while large animal models are common, the lack of consistency among the models varied dramatically among factors such as animal model and breed used, age of the animal upon infection and sacrifice, MAP strain used, infection dose and site of administration, length of infection prior to necropsy, and tissues sampled for downstream analysis (Begg and Whittington 2008). While current efforts have been made to establish consistency among large animal models and MAP research (Hines, Stabel et al. 2007), historical inconsistencies result in questioning of the validity of certain study conclusions, and make it difficult to form a general consensus about disease mechanisms as experimental differences are extensive between studies. 15

In order to understand the distinct mechanisms of MAP infection, initiation of disease, and its manipulation of the host, scientific models must be employed to study and characterize Johne’s disease and test novel vaccination strategies. Murine models are readily used as they offer a system that is cost-effective, and test groupings are almost genetically identical which provides less confounding data due to age and breeding backgrounds. Mouse studies have been widely used to study the initiation of the disease and validate vaccine candidates for protection against infection. Using the murine intestine, it was demonstrated that MAP utilize both M-cells within Peyer’s patches for uptake into the mucosa and also invade intestinal enterocytes to access the tissue, though each route results in different disease dissemination outcomes (Bermudez, Petrofsky et al. 2010). Mouse models have been utilized as the first in vivo stage to evaluate the efficacy of vaccine candidates identified using in vitro cell cultures, and to identify the most successful vaccine strains prior to moving to larger animal models such as goats and cattle (Scandurra, de Lisle et al. 2010, Bannantine, Everman et al. 2014). Murine systems have also been used to determine the effect of probiotic treatment on the prevention and treatment of intestinal MAP colonization (Cooney, Steele et al. 2014) and to establish the role of T-cells and immune reactivity to MAP infection in the intestinal tissue (Koc, Bargen et al. 2014). However, as the immune system and physiology of mice differs from that of ruminants previous vaccine efficacy studies have shown that findings from murine models may not accurately translate to ruminant models (Bannantine, Everman et al. 2014). Additionally, while the utilization of a murine model allows insight into early infection processes, it is harnessed by the limitation that mice do not develop the same long-term, clinical, inflammatory infection to MAP as mounted by and expressed in ruminants. For more simplified models, mammalian cell lines are extremely useful in the characterization of molecular interactions that occur over the course of infection. Cancer cells lines are the most commonly used as they are readily obtainable, grow quickly, and provide a realistic reproduction of the cells that the bacteria interact with within the host. Unfortunately, the immortalization of many cell lines can result in effects that may potentially alter the true nature of the cell. The effects that must be 16

kept in mind include but are not limited to excess/suppressed immune reaction, increased mutation rate due to repetitive culturing, increased metabolic activity, and alteration in phenotype due to 2-dimentioanl culture environments. Both murine (RAW 264.7 and J774) and bovine macrophage (BOMAC) lines are readily used to study interactions between MAP and the phagocytic host. These studies have described how MAP stimulates immune signals, intracellular survival by manipulation of phagosome-lysosme fusion and nitric oxide synthase acquisition (Kuehnel, Goethe et al. 2001, Miller, Fratti et al. 2004), and the role of opsonization on uptake (Stabel and Stabel 1995, Hostetter, Kagan et al. 2005) (Everman and Bermudez, In review). Likewise, the bovine epithelial Madin-Darby Bovine Kidney epithelial cell line (MDBK), serves as a model for the intestinal epithelial lining which MAP first encounters and crosses to begin infection. MDBK cells have aided in the identification of invasion-related proteins such as Cdc42 (Alonso-Hearn, Patel et al. 2008), characterization of novel phenotypes initiated after infection within milk and the intracellular environment (Patel, Danelishvili et al. 2006), and investigation of specific genes and proteins upregulated in novel phenotypes such as the hypothetical transcriptional regulator LuxR (Alonso-Hearn, Eckstein et al. 2010). In addition to cancer cell lines, primary cell lines can be obtained and cultured directly from the host and provide the most accurate depiction of host-microbe interactions in a simplified environment of a culture dish. Most utilized in MAP investigations are bovine monocyte-derived macrophages that are isolated directly from the peripheral blood mononuclear cells of bovine blood samples. These cell lines have aided in the study of MAP-induced immune responses during infection (Weiss, Evanson et al. 2005), the differences between MAP and its close relative MAH in survival with host cells (Woo, Sotos et al. 2006), and the screening of libraries for attenuated MAP strains that may be candidates for vaccination studies (Lamont, Talaat et al. 2014) In addition to mammalian cell lines, other models have been utilized to study the infectious mechanisms of MAP. Early studies demonstrated the ability of Mycobacterium avium subspecies avium (MAA) to invade and replicate with the amoeba Dictyostelium discoideum (Skriwan, Fajardo et al. 2002) and showed that invasion of the amoeba Acanthamoeba castellanii by Mycobacterium avium strain 17

101 produced a more virulent phenotype of the bacterium (Cirillo, Falkow et al. 1997). As MAP spends much time as an environmental resident after being shed in the feces of infected hosts, in was hypothesized that MAP may interact with a variety of species of amoeba in the soil. Not only has MAP been shown to interact with and be isolated from free-living amoeba in the environment (White, Birtles et al. 2010), but Tenant and Bermudez identified that the genes upregulated by MAP during intracellular amoeba infection were similar to those upregulated during macrophage infection, lending evidence toward the notion that amoeba could serve as a useful model for intracellular MAP infection (Tenant and Bermudez 2006). Ultimately, the current models used to study MAP infection are either too broad (large animal/ruminant models) to answer molecular mechanistic question or are too simplified (single cell culture models) to fully understand the complicated interplay between the array of cell types and the immune responses that are present in the intestinal mucosa. To address this issue, under-utilized models, including Acanthamoeba castellanii, and the development of novel models are needed in order to further understand the bacterial mechanisms and host cell response during the complicated multi-stage infection. New models are needed to simplify the intestinal mucosa so that studies are able to more adequately understand the pathogenic methods used by MAP during ruminant infection, without the utilization of large animal models.

Intracellular Survival Mechanisms of MAP The biology of the innate immune response and host phagocytes utilize an array of biochemical mechanisms useful in recognizing and eliminating pathogens encountered during disease and infection. While these mechanisms are useful against a multitude of microbes, others are able to withstand or manipulate these actions for their advantage or prolonged survival over the course of infection. Intracellular organisms including bacteria, parasites, and viruses have evolved their own set of mechanisms which allow for the microbes to not only survive but to replicate, alter the immune response, and encourage successful infection within the host cell. 18

Mycobacteria are well studied and thoroughly described as one genus of intracellular pathogens that manipulate the host cell to prevent death and to thrive in the intracellular compartment of host phagocytes. MAP is capable of surviving within the phagosome of host cells by utilizing a number of bacterial-induced strategies during infection. During the progression of phagocytosis, endocytic/phagocytic vesicles acquire markers which transition the vesicle from early endosomes to late endosomes and then signal for the fusion of endosomes/phagosomes and lysosomes, which contain a variety of lytic enzymes. This fusion results in the development of the bactericidal phagolysosomal compartment which is responsible for the elimination of invading pathogens (Kinchen and Ravichandran 2008). Mycobacterial phagosomes are capable of restricting the progression of maturation at the early endosome stage by manipulating the decoration pattern of rab family GTPase molecules on the phagosomal surface. Mycobacterium­ containing phagosomes acquire and concentrate rab5, an early endosomal marker, on the membrane of the vesicle over the course of infection, while inhibiting the acquisition of rab7 which inhibits transition to the late endosomal stage of vesicle maturation (Via, Deretic et al. 1997). The gradual acidification of the endosome from a neutral pH 7 to as low as pH 4.5 by membrane bound ATPase transporter complexes during maturation is another key mechanism that protects the host from subsequent infection (Cain, Sipe et al. 1989). While mycobacteria are capable of inhibiting the complete maturation of the vesicles in which it resides, the bacterium can also protect itself from the toxic effects of acidification. Mycobacterium avium-containing vesicles are capable of preventing the acidification of its phagocytic compartment by inhibiting the recruitment of vesicular ATPase to the phagosomal membrane (Sturgill-Koszycki, Schlesinger et al. 1994). More specifically, it has recently been demonstrated that the Mycobacterium tuberculosis protein PtpA, a secreted protein tyrosine phosphatase, is the protein responsible for the inhibition of acidification as it binds to subunit H of the ATPase, preventing the proton pump from functioning correctly, and contributes to the mechanism that results in the ATPase exclusion from the membrane of the endocytic vesicle (Wong, Bach et al. 2011). The ability to prevent complete endosomal 19

maturation and the maintenance of a neutral pH enables the bacteria within the vesicle to evade the toxic environment of the maturing phagosome and the lytic components found within the lysosome meant to protect the host from microbial infection. Mammalian cells also mount a nitrosative attack on invading and foreign microbes in attempt to protect the host from infection. Within bovine macrophage cells, MAP infection is capable of stimulating the inducible nitric oxide (iNOS) response and the release of nitric oxide within the phagocytic compartments. However, while artificially produced levels of nitric oxide are lethal to MAP, bovine macrophages do not produce near the level required, even in the presence of the immune response activator IFNγ, to result in more than a minimal mycobactericidal effect on the viability of the intracellular MAP (Zhao, Collins et al. 1997). While host cells may not mount a nitric oxide response high enough to kill intracellular MAP, the bacteria may also protect itself by preventing iNOS function in the phagosome. Studies with the tuberculosis vaccine strain Mycobacterium bovis strain BCG (M. bovis BCG) demonstrated that iNOS colocalizes with phagosomes containing inert latex beads and E. coli bacteria in an actin-dependent manner, but do not colocalize with those containing M. bovis BCG (Miller, Fratti et al. 2004). The study concluded that M. bovis BCG, as well as other mycobacteria, may facilitate the prevention or reduction the level of recruitment of iNOS to the membrane of the mycobacterium-containing phagosome, thus contributing to its intracellular survival.

Shifting Bacterial Phenotypes During Infection Mycobacterial species have a wide range of hosts they are capable of infecting and environments they are able to survive within. MAP is capable of infecting and surviving within ruminants (cattle, sheep, goats, bison, deer, etc), non-ruminants (camelids), and humans, while it is also able to cause infection in mice, non-human primates, horses, and rabbits (Collins 2013). In addition to mammalian infections, MAP can survive within phagocytic environmental amoeba and is capable of surviving for prolonged amounts of time in soil (White, Birtles et al. 2010). The wide array of environments MAP can survive within requires it to have a large repertoire of 20

genetic elements that allow it to adapt to each unique environment. The expression of such elements in an environment- or host-specific manner is the unique phenotype of the bacterium that is required for its survival and replication. MAP is capable of expressing a variety of phenotypes depending on both the environment is it in and the stage of infection it is at. Within cattle, the mammary epithelium and milk have been shown to serve as a reservoir to MAP prior to transmission. Passage through bovine mammary epithelial (MAC-T) cells elicits a phenotype which is more readily able to invade subsequent MDBK epithelial cells for enhanced initiation of infection (Patel, Danelishvili et al. 2006). Within the mammary gland, the exposure to milk can result in a dramatic shift in the phenotype of MAP, resulting in the upregulation of a variety of invasion- and virulence-related genes (MAP1203, MAP4088), and an increase in a hypothetical transcriptional regulator (MAP0482 – luxR) which can potentially direct the expansion of additional virulence components prior to infection (Patel, Danelishvili et al. 2006). Upon initial interaction with host cells, MAP changes its phenotype and alters the expression of genes that are important for invasion of the host epithelial mucosa. MAP upregulates the gene MAP3464, an activator of the host rho family GTPase protein Cdc42, nearly 6-fold upon epithelial cell interaction resulting in cytoskeletal rearrangement and engulfment of the pathogen (Alonso-Hearn, Patel et al. 2008). Together, these phenotypic changes lead to the rapid changing of bacterial proteins which result in the increased virulence and enhanced infection upon interaction with the intestinal mucosa of the host during the early stages of infection. During the course of infection MAP can alter its phenotype for the purpose of evading the immune system, prolonging its intracellular survival, or initiating the progression of Johne’s disease. Upon intracellular uptake the mechanisms listed previously, including inhibition of endosome maturation, acidification, and recruitment of iNOS, are enabled for the intracellular survival of the bacterium. Particular genes are capable of being upregulated prior to introduction to the intracellular environment which aid not only in uptake, but in the persistence within the phagosome. MAP1203, a hypothetical invasion and intracellular persistence (iipA) gene, is upregulated during milk exposure (Patel, Danelishvili et al. 2006, 21

Alonso-Hearn, Eckstein et al. 2010). The homologous iipA protein in Mycobacterium marinum is responsible for not only invasion of, but survival within, macrophages (Gao, Pak et al. 2006). Interestingly, the intracellular phenotype of MAP and other mycobacterial species is not limited to mammalian host cells. The genes that are upregulated during macrophage infection with MAH are quite similar to the gene expression which is upregulated during intracellular infection of amoeba (Tenant and Bermudez 2006). As MAP are readily found within free-living amoeba (White, Birtles et al. 2010), and can infect Acanthamoeba castellanii and Acanthamoeba polyphaga, which induces increased resistance to chlorine inactivation of the bacterium (Whan, Grant et al. 2006), it is a strong possibility that MAP also alters its phenotype during intracellular infection of amoeba. One a larger scale, MAP has been shown to alter its phenotypic components during the various stages of Johne’s disease. Bacteria isolated from MAP-infected bovine intestinal tissue exhibit a different proteome than bacteria grown in laboratory culture or isolated from fecal samples (Weigoldt, Meens et al. 2011), providing additional evidence that a more virulent MAP phenotype may be responsible for driving the progression of Johne’s disease. When taken together, it is evident that the phenotypic expression of MAP, and the ability of the bacterium to alter such expression in a variety of environmental and host conditions, indicates its important role in virulence of the bacterial organism. Further studies on such phenotypes, their specific components, and the role they play in host manipulation and enhanced bacterial survival are needed to fully elucidate the function of the shifting phenotypes during the course of MAP infection and Johne’s disease.

Scope of Doctoral Dissertation The aims of the following dissertation are to investigate the mechanisms used by MAP during infection of the bovine host by utilizing both existing and newly developed in vitro and in vivo models, molecular tools, and the phenotypes derived by the bacterium over the course of infection. In light of in vitro models that inadequately mimic the intestinal environment and interactions encountered by MAP 22

during infection, we set out to develop a novel model to more accurately understand the molecular changes and adaptations that the MAP bacterium make in order to cause Johne’s disease. Using the virulent phenotype that MAP develops in milk, we aimed to identify whether opsonization would serve as a useful and successful prevention strategy, ultimately in the form of passive vaccination. We aimed to characterize the function of one of the most upregulated genes, MAP1203, during milk exposure and understand its role in bovine epithelial interaction and initiation of Johne’s disease. Lastly, we utilized the environmental phagocyte Acanthamoeba castellanii as a model organism to understand the metabolic impact that MAP infection has on phagocytes during infection, and how those changes may hurt, or help, the ultimate survival of the intracellular pathogen during infection. Ultimately, the work described here, enhances the field of MAP research, and the tools available to study the disease. Our studies expand upon the phenotypic changes acquired by MAP, the role of those changes upon infection of host epithelial cells, the utilization of those changes for new prevention strategies, and the further understanding of the influence MAP infection has on the host phagocytes during infection. The following data add novel observations and findings to the field of MAP and Johne’s disease research, while providing new models and systems to utilize in order to continue to ask and to answer questions about the bacterial virulence mechanisms, and to enhance management options of MAP infection in ruminant populations worldwide.

23

Chapter 2

Antibodies against invasive phenotype-specific antigens increase Mycobacterium avium subspecies paratuberculosis translocation across a polarized epithelial cell model and enhance killing by bovine macrophages

Jamie L. Everman1, 2 , and Luiz E. Bermudez 1,2

1 Department of Microbiology, College of Science, Oregon State University 2 Department of Biomedical Sciences, College of Veterinary Medicine, Oregon State University

Frontiers in Microbiology – Manuscript In Preparation

24

Abstract Johne’s disease, caused by Mycobacterium avium subspecies paratuberculosis (MAP), is a severe chronic enteritis which affects large populations of ruminants globally. Prevention strategies to combat the spread of Johne’s disease among cattle herds involve adhering to strict calving practices to ensure young susceptible animals do not come in contact with MAP-contaminated colostrum, milk, or fecal material. Unfortunately, the current vaccination options available are associated with high cost and suboptimal efficacy. To more successfully combat the spread of Johne’s disease to young calves, an efficient method of protection is needed. In this study, we examined passive immunization as a mode of introducing protective antibodies against MAP to prevent the passage of the bacterium to young animals via colostrum and milk. Utilizing the infectious MAP phenotype developed after bacterial exposure to milk, we demonstrate that in vitro opsonization with serum from Johne’s-positive cattle results in enhanced translocation across a bovine MDBK polarized epithelial cell monolayer. Furthermore, immune serum opsonization of MAP results in a rapid host cell-mediated killing by bovine macrophages in an oxidative-, nitrosative-, and extracellular DNA trap-independent manner. This study illustrates that antibody opsonization of MAP expressing an infectious phenotype leads to the killing of the bacterium during the initial stage of macrophage infection.

25

Introduction Johne’s disease is a chronic enteritis caused by the pathogen Mycobacterium avium subspecies paratuberculosis (MAP). The global burden of the disease is widespread and outdated studies estimate that it results in $250 million to $1.5 billion per year in culled herds and loss of milk production among the US dairy industry alone (Stabel 1998, Ott, Wells et al. 1999). The most successful of current prevention strategies involves managing the spread of disease by implementing carefully planned calving practices to ensure that young animals receive colostrum and milk from Johne’s-free dams, preventing the exposure of young susceptible animals to contaminated feces, and culling and removing animals that test positive for the bacterium. Multiple vaccine formulations exist, though only one is commercially available in the United States. Overall, vaccination rates are generally low and herd- management is the most common and economically feasible form of Johne’s prevention worldwide. Published studies, and the product information for the commercially available vaccine Mycopar (Boehringer Ingelheim Vetmedica, Inc.) explain that while vaccination limits the progression of cases to the clinical stage of the disease, it does not prevent shedding of MAP in the feces, nor does it prevent vaccinated animals from becoming infected (Wentink, Bongers et al. 1994). Due to these factors and its associated cost, strict timeline of administration, and suboptimal efficacy, there is a continuous push to develop more efficacious vaccines to combat MAP infection. Unfortunately, the results obtained from the pipeline of determining host toxicity and vaccine efficacy from in vitro cultures and mouse models, did not translate in a successful vaccine trial in ruminant hosts due to unappreciated differences in immunity and pathogenesis of the infection between animal species (Hines, Turnquist et al. 2014). Furthermore, the phenotypic changes that occur within MAP during infection (Everman et al, In Review) or during exposure to different environmental or host reservoirs (Cirillo, Falkow et al. 1997, Patel, Danelishvili et al. 2006, Alonso-Hearn, Eckstein et al. 2010) may result in ineffective vaccine efficacy. There is a possibility that due to the incorrect set of antigens being the focus of vaccine development, chosen vaccine candidates are not representative of the most 26

relevant antigens during the stages of Johne’s disease in the animal. This is certainly a limitation of the current vaccine target approach, with consequent inefficient protection over the full course of the disease. Compared to vaccine-induced (active) immunity, which requires the host immune system to mount a response to introduced antigens, passive immunity provides immediate protection in the form of pre-formed antibodies. Neonatal calves have a narrow repertoire of gammaglobulins due to their immature immune systems and early protection of the animal is provided by uptake of maternal immunoglobulins concentrated in the colostrum during the first feedings in the early hours of life. These colostrum-delivered antibodies provide immediate immunity against naturally occurring enteric and respiratory pathogens which can lead to deadly diarrheal and pneumonic diseases in animals that do not receive proper feedings of colostrum (Godden 2008). Experimental vaccination of pregnant cows has shown to provide protection against pathogens such as Escherichia coli (Reiter and Brock 1975, Nagy 1980), Cryptosporidium parvum (Perryman, Kapil et al. 1999), and rotavirus (Saif, Redman et al. 1983), by the resulting mounted antibody titers which are passed to the neonate during initial feedings of colostrum. This passive transfer of opsonizing antibodies enables host phagocytes to eliminate potentially harmful pathogens by phagocytosis and intracellular killing, or by triggering antibody- dependent cell-mediated cytotoxicity (ADCC) for the elimination of the pathogen in the mucosal tissues of young animals. Previous studies have shown that MAP can reside within and acquire an infectious phenotype in the presence of milk, and in the mammary gland, with significant alteration in the gene expression of the pathogen (Koenig, Hoffsis et al. 1993, Patel, Danelishvili et al. 2006, Antognoli, Garry et al. 2008, Alonso-Hearn, Eckstein et al. 2010). This infectious phenotype may provide a novel and unstudied array of surface antigen biomarkers that may be used for the development of more effective preventative strategies. Considering the inherent susceptibility of young animals to infection, the altered and infectious MAP phenotypes from milk exposure, the protective mechanisms of passively obtained immunoglobulins, and the knowledge that young calves are at the highest risk of acquiring infection from uptake 27

of contaminated milk, we hypothesize that maternal passive immunity may serve as a protective mechanism against the passage of MAP to young animals. In the current study, we investigate the role of serum opsonization and its effect on the bacterial interaction with bovine epithelial cells and macrophages during infection. We reveal that serum opsonization of the infectious phenotype of MAP enhances the translocation of the bacteria across an epithelial barrier. Upon interaction with bovine macrophages, the immune serum opsonized pathogen expressing the milk-exposed infectious phenotype is rapidly killed. Based on our findings, we begin to understand how the method of passive immunity may serve as a successful approach for protection against acquiring MAP in the milk during the feeding of young animals.

28

Methods and Materials

Bacterial culture. Mycobacterium avium subspecies paratuberculosis strain K10 (ATCC BAA-968) was cultured at 37°C on 7H10 agar (BD; Franklin Lakes, NJ) supplemented with casein hydrolysate (1 g/L; BD), 10% (vol/vol) oleic acid, albumin, dextrose, and catalase (OADC; Hardy Diagnostics; Santa Maria, CA), and ferric mycobactin J (2 mg/L; Allied Monitor, Fayette, MO) for 3-4 weeks. Prior to experiments, a bacterial suspension was made in HBSS (Corning; Corning, NY), passed through a 22-gauge needle 5 times to disperse clumps, and allowed to settle for 10 minutes. The top half of the inoculum was used as a single-cell suspension for experiments (Patel, Danelishvili et al. 2006).

Mammalian cell culture. Madin-Darby Bovine Kidney (MDBK) epithelial cells (CCL-22) and murine RAW 264.7 macrophage cultures (TIB-71) were obtained from the American Type Culture Collection (ATCC; Manassas, VA). Both cells lines were cultivated in Dulbecco’s Modified Eagle’s Medium (DMEM) supplemented with 10% heat-inactivated fetal bovine serum (FBS; Gemini Bio-Products; West

Sacramento, CA), at 37°C in 5% CO2. The SV40 mutagenized bovine macrophage cell line (BOMAC) was a gift from the USDA and cultivated in RPMI-1640 (Corning) supplemented with 10% heat-inactivated fetal bovine serum at 37°C in 5%

CO2 (Stabel and Stabel 1995).

Serum samples. Bovine serum. Serum samples that were positive and negative for antibodies against MAP antigens, as determined by USDA diagnostic ELISA and Western blot against whole cell MAP lysates, were obtained from the USDA Sample Repository (Ames, IA). Serum samples from 10 positive and 10 negative animals were acquired, positive and negative samples were pooled, and aliquots stored at ­ 80°C prior to use. Mouse serum. Six-week old female C57BL/6 mice were purchased from Jackson Laboratory (Bar Harbor, ME) and held under observation for 1 week prior to use. MAP was exposed to milk as described below and whole MAP lysate was isolated by bead-beating bacteria with 0.1 mm glass beads in HBSS 5 29

times for 30 seconds in a Mini-BeadBeater (Biospec Products; Bartlesville, OK) at a speed of 4,800 oscillations/minute. Protein lysate was cleared of intact cells and cell wall debris by centrifugation at 9500 × g for 30 minutes at 4°C and plated to confirm no viable bacteria existed in the antigen preparation. Ten mice were pre-bled for control non-immune serum prior to immunization on day 0. Animals were injected subcutaneously with 0.1 mg of MAP antigen in incomplete Freund’s adjuvant (Sigma; St. Louis, MO) in 3 dorsal administration sites on day 1 and with 50 μg of MAP antigen in incomplete Freund’s adjuvant in 2 dorsal sites every 2 weeks. Test- bleeds were conducted to confirm antibody titer by Western blot analysis using a 1:5000 dilution of serum as the primary antibody and detected using a goat anti- mouse IRdye800 secondary detection antibody as per manufacturer’s instructions (Licor; Lincoln, NE). Blood was collected via cardiac puncture and serum was isolated by centrifugation at 5500 × g for 15 minutes at 4°C. Samples were pooled and aliquots stored at -80°C. All animal procedures were completed in strict accordance with guidelines set by the institutional animal care and use committee (Oregon State University Animal Care and Use Protocol #4490).

Bacterial exposure to raw milk and serum opsonization. Freshly collected whole milk from Holstein-Friesian cows was collected from a bulk tank at the Oregon State University Dairy Center. Milk was separated into 3 fractions by centrifugation for 20 minutes at 13,000 × g; milk fat (top layer) and milk particulate (pellet) were discarded and raw milk (middle fraction) was treated with polymyxin B (5 µg/ml), amphotericin (22 µg/ml), carbenicillin (25 µg/ml), and trimethoprim (2.5 µg/ml) overnight at 4°C. Samples were centrifuged again for 20 minutes at 13,000 × g to pellet any precipitated material, supernatant was collected, and aliquots frozen at -20°C. Prepared raw milk samples were inoculated with MAP for 24 hours at 37°C with shaking at 200 rpm. To isolate milk-exposed MAP, suspension was centrifuged at 2000 × g for 15 minutes at 4°C, and pellet was washed in HBSS twice for 10 minutes at 3500 × g at 4°C. For in vitro opsonization experiments, 1.5 x 107 milk- exposed MAP were mixed with a 1:50 dilution of serum in 1 ml total in HBSS and 30

incubated for 1 hour in a rotating hybridization oven at 37°C. Mock controls were incubated in an identical manner in the absence of serum.

Invasion assays. MDBK, BOMAC, or RAW 264.7 cells were seeded into 48-well plates and grown to 80% confluence prior to experiments. Cells were infected at an MOI of 10:1 with opsonized MAP samples and controls in complete DMEM or RPMI and infections were synchronized at 220 × g for 5 minutes prior to incubation at 37°C with 5% CO2. For invasion assays, infections were incubated for 15, 30, 45 minutes or 1 hour depending on experimental design, cells were washed 3 times with HBSS and depending on experiment, fresh DMEM supplemented with amikacin (200μg/ml) was added to each well for 2 hours to kill extracellular bacteria and washed 3 times with HBSS prior to lysis (Bermudez and Young 1994). To quantify bacterial uptake cells were lysed with 0.1% triton X-100 in deionized water, samples collected, serially diluted, and plated for CFU determination.

Transwell monolayer translocation assays. A transwell insert (Costar; Tewksbury, MA) with 3.0 mm pores was inserted into a 24-well tissue culture plate (Corning). 104 MDBK cells/well were added to the apical chamber of each well and DMEM was supplemented into the basal chamber upon seeding. Both apical and basal chamber media was replenished with fresh media every other day. The integrity of the monolayers was measured every 2 days by two different methods: Trypan Blue permeability and transmembrane resistance. Every 2 days, 0.4% trypan blue was added to the apical chamber of test wells and basal chamber samples were collected at 1, 5, 10, and 30 minutes. Samples were measured for trypan blue absorbance at 580 nm using a spectrophotometer (Boiadjieva, Hallberg et al. 1984, Mangum, Everitt et al. 1990). Transwell cultures were considered intact monolayers and used for assays once the trypan blue permeability read-out was less than OD580 0.010 at 30 minutes post-exposure and transmembrane resistance was greater than 400 Ω/cm3 (Bermudez, Sangari et al. 2002). For translocation assay, MAP was exposed to milk and opsonized as described above. Transwell inserts were infected with 106 MAP/well in the apical chamber in DMEM and fresh media was added to the basal chamber and 31

incubated at 37°C in 5% CO2. Samples were obtained at 6- and 24-hours post- infection by collecting 500 µl from the basal chamber, at which time media was replaced with equal volume media which was collected at 24 hours post-infection. Data from 6-hour samples were obtained from direct quantification of colonies and 24-hour samples were obtained by the addition of samples quantified from both the 6­ and 24-hour timepoints.

Mammalian cell response inhibitor assays. To block cellular response mechanisms, cells were treated with 10 U/ml catalase (Sigma), 300 U/ml superoxide dismutase (Sigma), a combination of 10 U/ml catalase and 300 U/ml superoxide dismutase, 50 µM diphenyleneiodonium chloride (DPI; Sigma), 250 µM NG- monomethyl-L-arginine (L-NMMA monoacetate; ENZO Life Sciences; Farmingdale, NY), DNase I 100 U/ml (Roche; Basel, Switzerland) and their respective buffers of 0.1 M potassium phosphate (catalase and superoxide dismutase separately), 0.2 M potassium phosphate (mix of catalase/superoxide dismutase), dimethylsulfoxide (DMSO; Sigma; used for DPI buffer), deionized water (L-NMMA), and DNase Buffer alone (DNase) as previously described (De Assis, Da Costa et al. 2000, Brinkmann, Reichard et al. 2004, Aulik, Hellenbrand et al. 2012). Host cells were incubated for 30 minutes with each compound or buffer in RPMI and washed 2 times with HBSS prior to bacterial infection. Cells were then infected as described above with opsonization samples suspended in RPMI supplemented with each inhibitor or buffer for 15 minutes, washed, lysed with 0.1% triton X-100 in deionized water, serially diluted, and plated for CFU. Milk-exposed opsonized MAP samples and host cells were incubated in cellular inhibitors individually for the same amount of time to determine the effect of each inhibitor on bacterial viability and host cell viability and quantified using CFU counts or trypan blue staining (0.4% w/v), respectively.

Statistical analysis and data interpretation. Results are reported as the mean of at least 2 replicate experiments each performed in triplicate ± standard error. Statistical comparisons between experimental groups and control groups were determined using the Student’s t test with p<0.05 denoting statistical significance. GraphPad Prism 32

version 6.0 software was used for the construction of graphs, data interpretation, and statistical analysis.

33

Results

Whole serum opsonization influences the interaction of MAP with MDBK epithelial cells Opsonization aids in the uptake of pathogens by host phagocytes and has been demonstrated to enhance killing of intracellular pathogens during infection (Weber, Ducry et al. 2014). Previous investigations have shown that opsonization of plate- grown MAP increases the efficiency at which the bacterium is able to be ingested and increases survival within the intracellular compartment (Hostetter, Kagan et al. 2005). However, the effect opsonization has on the uptake of MAP by the intestinal mucosa is unknown. To determine the effect of opsonization on the interaction between MAP cultured in milk (hereafter referred to as the infectious phenotype) and the bovine intestinal epithelium, we evaluated how opsonization influenced the invasion of and the translocation across polarized bovine MDBK epithelial cells. The infectious phenotype of MAP was recovered and opsonized with PBS (mock), bovine serum from Johne’s negative animals (non-immune serum), or bovine serum from Johne’s positive animals (immune serum). Invasion assays indicated that neither non-immune nor immune serum opsonization had any significant effect on the ability of MAP to invade MDBK epithelial cells compared to mock opsonized samples (Figure 2.1). While the epithelial cell invasion capability of MAP isn’t significantly affected by opsonization, the mechanism of uptake may provide a difference in the survival or efficiency of translocation across the epithelium. To investigate the fate of opsonized MAP after uptake by MDBK epithelial cells, a transwell culture assay was employed. After 4 days in culture, monolayers were determined to be impermeable and intact as demonstrated by both transmembrane resistance measurements (Figure 2.2a) and a trypan blue permeability assay performed 4 days after cell seeding (Figure 2.2b). Opsonization with immune serum resulted in enhanced translocation of MAP across an intact polarized MDBK epithelial cell monolayer with 3.1-fold more bacteria being recovered from the basolateral chamber after 6 hours, and almost 3.7-fold increase in recovered MAP after 24 hours post-infection compared to non-immune serum opsonized bacteria 34

(Figure 2.2c). These data indicate that opsonization with sera from Johne’s infected, immune positive cattle has little effect on uptake of bacteria by MDBK epithelial cells, but results in significantly enhanced translocation of infectious MAP across the polarized epithelial cell monolayer during infection.

Bovine macrophage uptake of serum opsonized MAP Once established that opsonization of MAP increased the level of translocation across a polarized MDBK epithelial cell monolayer, we wanted to address whether enhanced translocation contributes to the virulence of MAP by allowing greater numbers to be taken up by and survive within tissue macrophages, or whether greater translocation results in enhanced uptake and subsequent killing of MAP by macrophages. To assess the MAP-macrophage interaction in this system, we analyzed what effect opsonization had on the uptake of MAP by an immortalized bovine macrophage cell line (BOMAC). Preliminary assays were carried out to determine if our experimental model was able to replicate the effect of macrophage uptake of MAP grown under standard conditions as previously described (Hostetter, Kagan et al. 2005). Consistent with previous studies, opsonization of broth cultured MAP with serum samples results in a 2-fold increase in uptake of MAP by bovine macrophages after a 15 minute infection (Figure 2.3a). It was also shown that neither mock (PBS), non-immune serum, nor immune serum opsonization treatments of milk-cultured or 7H9-cultured MAP had any detrimental effect on the viability of each pool of bacteria prior to infection (data not shown), suggesting that any observable change is due to the response of the host cell and not due to a difference of inoculum levels or impact on viability resulting from serum components, culture, or opsonization treatments occurring before infection. Prior studies demonstrated the increased ability of the infectious phenotype of MAP to invade MDBK epithelial cells (Patel, Danelishvili et al. 2006), and our data indicates that the infectious phenotype also has a significantly greater rate of uptake by BOMAC cells after a 15 minutes infection as 8.55 ± 1.15% of MAP expressing the infectious phenotype are ingested by BOMAC cells (Figure 2.3b; black bar) compared to 3.72 ± 0.32% of 7H9-exposed MAP (Figure 2.3a; black bar) after 35

infection (p = 0.0054). To determine the effect of the infectious phenotype on the interaction between opsonized MAP and BOMAC cells, infection assays were carried out to assess the amount bacteria that were able to be taken up by macrophages. Intracellular bacterial quantification demonstrated that the infectious phenotype resulting from milk exposure results in a 75% decrease in BOMAC uptake of immune serum opsonized samples compared to mock or non-immune serum opsonized MAP populations (Figure 2.3b). This decrease was observed over a variety of infection lengths and sample treatments. Initial 3 hour assays, which included a 1 hour infection followed by 2 hours of antibiotic treatment, and short-term infections of 15 minutes, 30 minutes and 1 hour in the absence of antibiotic treatment were carried out and result in nearly identical data (15 minute assay shown in Figure 2.3b; other data not shown). Collectively, these data indicate that immune serum opsonization of infectious MAP results in a significant decrease in the amount of bacteria taken up by BOMAC cell cultures during infection over both long- and short-term infections. Immunoglobulins in the serum, particularly IgG and IgA, are responsible for opsonizing pathogens and toxins to enhance uptake and subsequent killing, or for neutralization and protection from disease, respectively (Janeway 2001). To ascertain which function the serum antibodies have in our model and the mechanism responsible for the decrease in uptake of immune serum opsonized MAP, the supernatant from each experimental group during BOMAC cell infection was collected and the number of viable bacteria quantified (Figure 2.3c). After 15 minutes of infection approximately equal numbers of MAP were recovered from the supernatant of mock, non-immune, and immune serum opsonized infections. These findings indicate that the decrease in uptake of MAP opsonized with immune serum was not due to a neutralizing effect of the antibodies blocking uptake by macrophages. Rather, the infectious phenotype opsonized with immune serum results in the killing of MAP as there is no difference in the viable amount of bacteria within the supernatants of each treatment. The mechanism used to kill immune serum opsonized bacteria is rapid, as significant changes in viability occur after only a short 15 minute incubation period. In total, these data illustrate that this infectious 36

phenotype of MAP serves as a key mechanism for recognition by immune serum and for the specific response initiated by BOMAC cells during infection.

Determining macrophage mechanisms for killing of immune serum opsonized MAP Innate cellular mechanisms are vital for allowing host cells to react in a rapid and non-specific manner to pathogen associated molecular patterns (PAMPs) (Medzhitov 2001), as well as antibody coated particles (Greenberg and Grinstein 2002). The phagocytic vacuole is hard-wired to fuse with lysosomes and deliver numerous oxidative stress molecules that eliminate the pathogen within the vacuole (Janeway 2001). To identify which host cellular mechanism(s) are being employed in the early killing of immune serum opsonized MAP, we inhibited a variety of cellular mechanisms known to be utilized during the process of phagocytosis. Though mycobacteria are well described to inhibit this fusion process (Kuehnel, Goethe et al. 2001, Vergne, Chua et al. 2004), antibody-mediated phagocytosis may utilize different vacuole trafficking signals and fusion mechanisms which lead to a different outcome of mycobacterial survival. Chemical inhibitors of classical oxidative mechanisms were added to infections to determine if there was an effect on the viability of opsonized and non-opsonized milk-exposed MAP (Figure 2.4) over the course of the 15 minute infection. The inhibition of hydrogen peroxide production by the addition of catalase, and the blockage of superoxide anion radical efficacy by superoxide dismutase (SOD) results in no change in the recovery of immune serum opsonized populations of MAP. As the conversion of superoxide by superoxide dismutase results in the release of oxygen and hydrogen peroxide, to combat both oxidative mechanisms we treated cells with a combination of catalase and superoxide dismutase in tandem, which results in no change from the control groups (Figure 2.4). The membrane bound NADPH oxidase enzyme complex is responsible for the production of a variety of reactive oxygen species (ROS) including but not limited to peroxides, oxygen radicals, and oxidative stress components that can kill bacteria within the cell (Janeway 2001). To antagonize the collective production of these toxic products, cells were treated with the NADPH oxidase inhibitor 37

diphenyleneiodonium chloride (DPI). While the overall levels of uptake were lower in DPI treated samples, the overall pattern demonstrating a lower uptake of immune serum opsonized MAP remains the same as with the addition of other inhibitors. It was noted that DPI-treated MAP controls result in a 35% decrease in the viability of the bacteria, while there was a 10% decrease in host cell viability in DPI-treated controls compared to DMSO vehicle treated samples, accounting for the decrease in overall uptake in the DPI-treated assays. All other controls demonstrate negligible changes in viability of either MAP or host cell controls during treatment with inhibitors or their respective vehicle controls (data not shown). In addition to the stimulation of an oxidative burst upon phagocytosis, host cells can mount a nitrosative burst, specifically by the production of nitric oxide, to aid in the destruction of ingested pathogens. To analyze the role of nitric oxide during phagocytosis of opsonized milk-exposed MAP, treatment was completed with G the nitric oxide inhibitor N -monomethyl-L-arginine (L-NMMA). Treatment with L­ NMMA results in no change in uptake and intracellular bacterial load of immune serum opsonized MAP compared to control treated samples (Figure 2.5). There was no effect on viability of the MAP or cell populations upon addition of L-NMMA in controls (data not shown). To identify if extracellular killing mechanisms were responsible for the rapid killing of immune serum opsonized MAP, we investigated the role of extracellular traps during our infection model. Neutrophil extracellular traps (NETs) have been described as a toxic tool employed upon infection with both gram-negative and gram- positive organisms (Brinkmann, Reichard et al. 2004, Aulik, Hellenbrand et al. 2012). Upon exposure to stimuli, cells release a sticky DNA net into the extracellular environment which is characterized by the presence of histones and elastase enzymes. This phenomenon has recently been reported to be used by macrophages as well, resulting in macrophage extracellular traps (METs) in response to TNF-α (Mohanan, Horibata et al. 2013) and a variety of pathogens and toxins (Aulik, Hellenbrand et al. 2012, Liu, Wu et al. 2014). To determine if METs were stimulated in response to infection with our bacterial populations, BOMAC cultures were treated with DNase prior to and during each uptake assay. DNase treatment identifies that the dissolution 38

of DNA complexes outside of the cell does not alter the pattern of selective immune- serum opsonized MAP killing (Figure 2.6a), nor does it alleviate the lack of extracellular immune serum opsonized MAP during the assays (Figure 2.6b).

Opsonization of infectious MAP phenotype in a murine model system The observation that bovine macrophages were able to more effectively eliminate the infectious phenotype of MAP after opsonization with immune serum in such a rapid manner has not previously been described prior to this study. To understand if the phenomenon was specific to the physiology of bovine macrophages or could be observed in other systems, we completed the opsonization of the infectious phenotype of MAP and infection in a murine model. We first produced antibodies against the infectious phenotype of MAP in C57BL/6 mice. Western blot analysis of infectious MAP lysates and milk protein alone confirm that collected mouse serum contained antibodies which recognize only MAP bacterial lysate proteins and not bovine milk proteins (Figure 2.7a). To identify the effect the collected mouse serum had on the interaction of opsonized infectious MAP with murine RAW 267.4 macrophages, we conducted similar 15 minute macrophage infection assays as done with the bovine model (Figure 2.8b and 2.8c). Murine RAW 264.7 macrophages exhibit a significantly lower level of immune serum opsonized bacteria compared to non-immune serum and mock opsonized samples (Figure 2.8b). Additionally, the decrease in bacterial uptake by murine macrophages is not an effect of neutralization, as the lack of intracellular MAP in the uptake assays are not recovered from the supernatants of each culture (Figure 2.8c). These data validate that the pattern of rapid host cell mediated killing is consistent between both the bovine and murine in vitro model systems, and provides further support that the infectious phenotype of MAP serves as an important and relevant bacterial phenotype for studies involving opsonization, and the resulting rapid phagocyte killing, as a means of preventing infection in multiple species models.

39

Discussion At present, the most effective preventative measures against Johne’s disease are strict calf management practices so as to avoid the exposure of neonates and young animals to colostrum, milk, feed, and soil contaminated with MAP. The currently approved vaccine protocol involves administration of the bacterin within a strict window between 0 – 30 days of age. Unfortunately, the vaccine can be ineffective at stopping the spread of the disease through a herd and results in variable protection against disease depending on the immune status, age, infection status, and other unknown factors about the animal. Treatment options are difficult and not generally encouraged as they include antibiotic regimens which are expensive, are required for the duration of the animal’s life, and are often unable to fully eliminate the infection. Most often, infected animals are culled in order to prevent the transmission of MAP to the rest of the herd. Novel vaccination strategies would allow for more successful methods of preventing infection among young animals, thus slowing the cycle of transmission and incidence of Johne’s disease within herds. The delivery of antibodies in colostrum and milk is an important mechanism which passively transfers protective IgA and IgG to neonatal and suckling animals that are born with undeveloped and immature immune systems. The intestinal physiology of neonatal calves is quite unique within the first 24 hours of life as the animal needs to take up as much protective immunoglobulin as it can to establish a protective immune system. Neonatal animals express increased levels of the neonatal Fc receptor (FcRn) in various tissues including the lung (Mayer, Kis et al. 2004), and intestine (Kuo, Baker et al. 2010). These receptors allow for the maximum level of absorption of maternal IgG molecules into circulation for the establishment of immunity and to provide local protection against pathogen exposure early in life. Meanwhile, IgA molecules are utilized as neutralizing protection against pathogens the young animal may encounter prior to the establishment of a complete immune system. The passive transfer of antibodies via milk from experimentally vaccinated dams can result in protection against commonly encountered farm pathogens (Reiter and Brock 1975, Saif, Redman et al. 1983, Perryman, Kapil et al. 1999). In this study we aimed to identify whether the shift to an infectious bacterial phenotype in milk 40

(Patel, Danelishvili et al. 2006), its direct transmission to young animals (Taylor, Wilks et al. 1981, Sweeney, Whitlock et al. 1992), and the natural passage of maternally-derived antibodies would be useful as a protective mechanism against an intracellular pathogen such as MAP. Our initial results suggest that antibodies do not play a neutralizing role during initial epithelial cell infection as there was no significant change in the invasion ability between opsonized and mock treated MAP after 1 hour of infection. Alternatively, rather than protecting the epithelial barrier from infection, opsonization of the infectious MAP phenotype resulted in a significantly increased rate of trafficking of MAP across an intact polarized MDBK epithelial cell transwell culture. Epithelial cells, limited in their protective abilities, may be transporting opsonized MAP to a location where they are more likely to be appropriately dealt with by professional immune cells in the mucosal tissue for a more successful and positive outcome for the host. As MAP is an intracellular pathogen with an arsenal of mechanisms used to avoid host cell-mediated killing mechanisms, we asked whether the enhanced translocation across the epithelial cell layer was advantageous to the uptake by and survival within host phagocytes, or whether it resulted in a protective effect by triggering enhanced killing of the pathogen. Astoundingly, upon opsonization of MAP expressing the milk-induced infectious phenotype, the effect of opsonization changed from one of enhanced uptake (Figure 2.3a) to one that resulted in the immediately killing of the immune serum opsonized bacteria (Figure 2.3b). This effect was solely a result of the phenotype and antibody recognition, as preliminary experiments ruled out reagent and protocol influence. The decrease in uptake of immune serum opsonized MAP by macrophages was not an effect of neutralizing IgA molecules, as we would have expected to see a higher amount of unbound and non-phagocytized but viable MAP in the supernatants to make up for the decrease in intracellular MAP in each respective sample (Figure 2.3b and 2.3c). We predict that the dramatic change in the outcome of opsonized samples and the impact on MAP viability are an effect of the specific milk-induced infectious bacterial phenotype and the antibody recognition of serum from clinically infected cattle. The more effective match between surface exposed biomarkers and the 41

deposition of antibodies during opsonization may be a key factor in the development of a more specific, and here shown as toxic, response by macrophages upon infection with immune serum opsonized, milk delivered MAP. As we observed a rapid bacterial killing effect triggered by immune serum opsonization of MAP, we focused on identifying the specific innate immune mechanism responsible for the bacterial killing in our model. Uptake of opsonized pathogens results in the fusion of the bacteria-containing phagosome with intracellular lysosomes containing bactericidal components. The inhibition of the host cell-initiated mechanisms including catalase, superoxide anions, and the functional NADPH oxidase complex, resulted in no change in the viability of immune serum opsonized MAP during the infection of bovine macrophages, indicating that the oxidative burst triggered during uptake or phagosome-lysosome fusion is not responsible for the rapid killing of the bacteria (Figure 2.4). The nitrosative burst within host phagocytes can also be used to eliminate intracellular pathogens, although mycobacteria are described to inhibit the recruitment of nitric oxide synthase to the phagosomal membrane during macrophage infection (Miller, Fratti et al. 2004), and the level of nitric oxide produced during infection only results in minimal mycobactericidal activity against MAP (Zhao, Collins et al. 1997). Regardless of the role of nitric oxide during intracellular MAP infection, inducible nitric oxide has been described to play a role in the response to opsonized particles, as demonstrated by luminescence assays in human macrophages (Gross, Dugas et al. 1998); though it has also been described that IgG-opsonized particles induce apoptosis in macrophages in an oxidative-dependent and nitric oxide-independent manner (Kim, Kwon et al. 2003). We illustrate that the inhibition of nitric oxide with L-NMMA resulted in no change in the intracellular viability of opsonized MAP in our model, indicating that the nitrosative burst is not responsible for the killing of the immune serum opsonized MAP upon macrophage infection. From these studies, we pinpoint a variety of cellular mechanisms that are not responsible for the killing of opsonized MAP. Further investigation into additional mechanisms of cellular protection including Cu/Zn toxicity and the activation of antibody-dependent cell-mediated cytotoxicity 42

and their role in killing of opsonized MAP will elucidate the potential novel mechanism used to provide a toxic outcome for MAP infection within our model. As an extracellular mechanism for eliminating pathogens, extracellular traps offer an efficient method of protection against infection in a tissue. The milieu that composes extracellular traps, including histones, enzymes such as elastase and a variety of cellular proteases, catch trap disease causing organisms, resulting in their destruction and degradation. Our investigation shows that the addition of DNase, which has the ability to disrupt the DNA matrix formed in the traps, thus rendering them unable to trap bacteria, has no effect on the killing on immune serum opsonized MAP in our model system (Figure 2.6). Furthermore, the formation of NETs is dependent on the reactive oxygen species produced by NADPH oxidase activity during infection (Fuchs, Abed et al. 2007). Though it is unknown if the same is true for macrophage extracellular traps, the resulting inhibition of NADPH oxidase with DPI in our study, which shows no change in the host-mediated killing of immune serum opsonized MAP, offers further evidence towards macrophage extracellular traps not being the responsible cellular mechanism for the specific host-mediated killing in our model. It is important to address the current inability for protection to be conferred to a neonate from a dam that is currently infected with MAP and should, in theory, be providing protective antibodies to her suckling calf. There are a variety of physiological, molecular, and immunological reasons that may explain why calves aren’t currently protected by the ingestion of milk from an infected cow. It is quite possible that MAP-contaminated milk does transfer both specific and non-specific antibodies to the young calf which confers a certain level of protection against MAP infection. Transferred antibodies may provide a level of protection which increases the infectious dose required for infection to be established, or it may simply allow for a slower progression of the disease as fewer bacteria are able to invade the epithelial lining in the beginning of infection. It is also clear that MAP and the closely related Mycobacterium avium subspecies hominissuis readily change their phenotype based on the environment they are in or the host they are infecting, including milk (Patel, Danelishvili et al. 2006), amoeba (Cirillo, Falkow et al. 1997, Tenant and Bermudez 43

2006), bovine macrophages (Bermudez, Petrofsky et al. 2004, Patel, Danelishvili et al. 2006), and intracellular passage between epithelial cells and macrophages in the intestinal tissue (Everman et al, In Review). The phenotype that a young animal is initially infected with, that which is expressed by the bacterium as it begins to replicate and spread throughout the host intestinal tissues, and the phenotype that occurs upon the transition to the clinical, inflammatory stage of the disease are potentially each characterized by their own particular set of antigens that are different from one another. Depending on the phenotype of the bacterium to which the pregnant cow produced antibodies which would be naturally passed in the milk, it is possible that complete or even partial protection may not be provided due to the mismatch between the dominant antigen epitopes of the bacterium and the maternal antibody recognition sites from her mounted humoral response. Overall, little is known about the protection that maternal antibodies delivered in colostrum and milk provide against MAP infection, if it exists at all; though it is clear that full protective passive immunity is not transferred as calves ultimately become infected with MAP from early life feedings. As seen with other pathogens (Reiter and Brock 1975, Saif, Redman et al. 1983), it is a possibility that the most protective antibody titers against MAP would have to be induced by a vaccination with the relevant set of infectious MAP antigens during pregnancy in order for the production and adequate passage of protective antibodies to her calf. To further understand the effect of passively transferred antibodies and the protective effect against MAP infection, neonatal calves would serve as the ideal model. The uptake of colostrum and milk immunoglobins is dependent on the unique and specific physiology of the neonatal calf. Antibodies are able to more likely survive degradation due to the lack of digestive enzymes such as pepsin which are secreted in low amounts in the first hours of life (Thivend 1980), and only in the intestinal epithelium of newborn calves are antibodies highly transported across the epithelial barrier which begins to mature, divide, and set the appropriate intestinal acidic pH as early as 12 hours after birth (Stott, Marx et al. 1979, Kuo, Baker et al. 2010). Using a neonate model would appropriately demonstrate the protective ability 44

of passive immunity and its effect in protecting against Johne’s disease derived from MAP contaminated milk and colostrum sources. In total, we demonstrate that passive antibody transfer of MAP-opsonizing antibodies serves a protective role against the initiation of infection of MAP in bovine and murine in vitro cell models. Opsonization of a relevant infectious phenotype that is present at the immediate time of passage from dam to calf results in greater translocation of bacteria across polarized epithelial cells. The bacteria that reach the sub-epithelial tissues are rapidly killed by macrophage-mediated bactericidal mechanisms which specifically eliminate antibody-opsonized MAP. These data offer evidence toward a novel preventative strategy that takes advantage of the infectious phenotype that MAP expresses at the time of transfer from cow to calf in the milk. The passive transfer of anti-MAP antibodies and enhanced immunity in the early hours of the life of a calf may be able to prevent the initiation of disease in young animals, and thus stop infection and the development and transmission of clinical Johne’s disease among herds.

45

Acknowledgments We would like to thank Judith Stabel at the National Animal Disease Research Unit, USDA-ARS in Ames, IA for the generous gift of the BOMAC cell line. Serum samples were obtained from the tissue repository at the National Animal Disease Research Unit, USDA-ARS Ames, IA.

46

2 .5

2 .0

1 .5

1 .0 % In v a sio n 0 .5

0 .0

u m c k r ru m M o S e S e e e u n

m m u n Im m o n -I N

Sample/Treatment

Figure 2.1: Effect of opsonization of milk-exposed MAP on invasion of MDBK epithelial cells MAP incubated in raw milk for 24 hours was isolated and opsonized in PBS (mock), or 2% serum from uninfected (non-immune) or infected (immune) cattle for 1 hour at 37°C. MDBK monolayers were infected for 1 hour at 37°C before collection of cell lysate. Percent uptake was calculated by (amount of intracellular bacteria/amount of bacteria added to the well) × 100. Data represents mean ± SEM of 2 independent experiments each performed in triplicate. Differences not significant (p>0.05) as determined by a Student’s t-test.

47

A B 5 0 0 Media Control ) 2 .0 3 M D B K C e lls 4 5 0 1 .5 4 0 0 1 .0 3 5 0 0 .5 3 0 0 Resistance (ohms/cm

2 5 0 Absorbance (580nm)0 .0 2 4 4 8 7 2 9 6 1 5 1 0 3 0

Time (hours) Time (minutes)

C *** 1 .5 1 0 5 ** 6 h o u r s * * 2 4 h o u r s 1 .0 1 0 5

5 .0 1 0 4 CFU Recovered

0

k k m u m ru m ru m r ru o c e o c e M S S e M S e S e e n e n e u n m u m m Im m u Im m u n Im n -I N o n - N o Sample/Treatment

Figure 2.2: Translocation of opsonized milk-exposed MAP across an MDBK epithelial transwell monolayer Apical chambers were seeded with 104 MDBK cells and monolayer integrity was measured. Trypan blue exclusion assay was measured every other day (a), and transmembrane resistance of each well was measured every day until monolayer resistance reached 400 Ω/cm3 (b). MAP was incubated in milk for 24 hours and then opsonized with PBS (mock), serum from Johne’s negative cattle (non-immune serum), or serum from Johne’s positive cattle (immune serum). Apical chambers of a transwells containing intact bovine epithelial monolayers were infected with 106 MAP and samples from the basal chamber were collected and quantified for translocated bacteria at 6 and 24 hours post-infection. Trypan blue permeability data 48

represent mean ± SEM of 3 independent experiments each performed in duplicate; transmembrane resistance data represent mean ± SEM of 8 wells and is representative of 3 independent experiments; translocation data represent the mean ± SEM of 2 independent experiments each performed in triplicate (* p<0.05; ** p<0.01; *** p<0.001 as determined by a Student’s t-test). 49

Figure 2.3: Immune serum A opsonization of milk-exposed MAP ***

1 0 ** initiates phenotype which decreases n s 8 uptake of bacteria by BOMAC via

6 cell-mediated killing

4 MAP were incubated in 7H9 broth (a) or % U p ta k e 2 milk (b and c) for 24 hours, then 0 opsonized with non-immune serum, m

M o c k S e ru m e S e ru immune serum, or PBS (mock) and used

Im m u n Im m u n e o n - N to infect BOMAC cells for 15 minutes. B Sample Treatment Samples were immediately collected and **** 1 2 *** cell lysates (a and b) and supernatants 1 0 (c) from each well were quantified for 8 bacterial uptake and supernatant 6

% U p ta k e 4 bacterial viability, respectively. Data 2 shown represents the mean ± SEM of 4 0 independent experiments each u m u m o c k M S e r S e r n e n e performed in triplicate (**p<0.01, m m u m m u -I I n N o ***p<0.001; **** p<0.0001 as C Sample Treatment determined by a Student’s t-test). 1 0 0

8 0

6 0

4 0 % Recovered 2 0

0

u m u m o c k M S e r S e r n e n e

m m u m m u -I I o n N

Sample Treatment 50

1 5 M o c k Non-immune serum Immune serum

1 0

**** **** **** ****

% U p ta k e + + + + + + **** + + + *** + + + + + + + + 5 + + + + **** + + + + *** + +

0 0.2 M Potassium-Phosphate - + + + - - - - 0.4 M Potassium-Phosphate - - - - + + - - DMSO ------+ + C a ta la s e - - + - - + - - Superoxide Dismutase - - - + - + - - DPI ------+ Sample/Treatment

Figure 2.4: Inhibition of cellular killing mechanisms and effect on uptake of opsonized MAP by BOMAC cells BOMAC cultures were pre-incubated in RPMI with the cellular mechanism inhibitors or respective vehicle controls for 30 minutes prior to infection as follows: catalase (10 U/ml)/0.2 M potassium phosphate, superoxide dismutase (SOD; 300 U/ml)/0.2 M potassium phosphate, diphenyleneiodonium chloride (DPI; 50 μM)/DMSO, a combination of catalase/SOD (10 U/ml and 300 U/ml, respectively)/0.4 M potassium phosphate buffer. Each well was infected with 107 mock, or serum opsonized milk- exposed MAP samples in RPMI in the presence of each respective inhibitor/buffer combination or buffer alone. After 15 minute incubation, samples were washed, lysed, and quantified. Data shown represent mean ± SEM of 2 independent experiments each performed in triplicate (Mock treated samples compared to immune serum treated samples ****p<0.0001, ***p<0.001; Non-immune serum sample treatment compared to immune serum treatment samples ++++p<0.0001, +++p<0.001, ++p<0.01 as determined by Student’s t test). 51

8 ** * * **** 6

4 % U p ta k e 2

0 M o c k No treatment L-NMMA Non-immune serum Sample/Treatment Immune serum

Figure 2.5: Role of nitric oxide in the killing of immune serum opsonized MAP during BOMAC infection BOMAC cells were pre-incubated with the cellular mechanism inhibitor L-NMMA (250 μM) or water vehicle control in RPMI medium for 30 minutes prior to infection. Cells were infected with 107 mock or serum opsonized milk-exposed MAP in RPMI in the presence of inhibitor and buffer alone. After 15 minute infection, samples were washed, lysed, and quantified. Data shown represent mean ± SEM of 2 independent experiments each performed in triplicate (*p<0.05, **p<0.01, ****p<0.0001 as determined by Student’s t test).

52

A 1 0 M o c k 8 Non-immune serum Immune serum 6

4 % U p ta k e 2

0

N o D N a s e D N a se I / tr e a tm e n t B u ffe r DNase Buffer

B 1 0 0

9 0

8 0

7 0

6 0 % Recovered 5 0 1 0 0

N o D N a s e D N a se I / tr e a tm e n t B u ffe r DNase Buffer

Sample/Treatment

Figure 2.6: Role of macrophage extracellular traps (METs) in the killing of immune serum opsonized MAP during BOMAC infection BOMAC cells were pre-incubated with DNase I/DNase buffer (100 U/ml) or DNase buffer alone in RPMI medium for 30 minutes prior to infection. Cells were infected with 107 mock or serum opsonized milk-exposed MAP in RPMI in the presence of inhibitor and buffer alone. After 15 minute infection, samples were washed, lysed, and quantified. Data shown represent mean ± SEM of 2 independent experiments each performed in triplicate.

53

A

B ** C 8 0 *** 4 0

6 0 3 0

4 0 2 0 % U p ta k e 2 0 % Recovered 1 0

0 0 k u m k ru m ru m u m M o c S e r S e M o c S e S e r n e n e u n e u n e m m u Im m I Im m u n - -Im m o n N o N Sample/Treatment Sample/Treatment

Figure 2.7: Effect of opsonization on MAP survival in a murine model Serum was collected from mice after subcutaneous vaccination with protein preparations of the infectious phenotype of MAP over a 3 month timeframe. Equal amounts of proteins from either the infectious MAP phenotype or milk proteins only separated on an SDS-PAGE protein gel and Western blot was conducted using terminal serum as the primary probing antibody (1:5000), and probed with a goat α­ mouse IR800 secondary antibody for visualization (a). The infectious phenotype of MAP was opsonized with pre-bleed non-immune serum, terminal bleed immune serum, or PBS (mock) and used to infect RAW 264.7 macrophage cells for 15 minutes. Cell lysate (b) and supernatant (c) samples from each well were quantified 54

for bacterial uptake and supernatant bacterial viability, respectively. Data shown represents the mean ± SEM of 2 experiments each performed in triplicate (**p<0.01, ***p<0.001; **** p<0.0001 as determined by a Student’s t-test).

55

References Alonso-Hearn, M., T. M. Eckstein, S. Sommer and L. E. Bermudez (2010). "A Mycobacterium avium subsp. paratuberculosis LuxR regulates cell envelope and virulence." Innate Immun 16(4): 235-247.

Antognoli, M. C., F. B. Garry, H. L. Hirst, J. E. Lombard, M. M. Dennis, D. H. Gould and M. D. Salman (2008). "Characterization of Mycobacterium avium subspecies paratuberculosis disseminated infection in dairy cattle and its association with antemortem test results." Vet Microbiol 127(3-4): 300-308.

Aulik, N. A., K. M. Hellenbrand and C. J. Czuprynski (2012). "Mannheimia haemolytica and its leukotoxin cause macrophage extracellular trap formation by bovine macrophages." Infect Immun 80(5): 1923-1933.

Bermudez, L. E., M. Petrofsky and F. Sangari (2004). "Intracellular phenotype of Mycobacterium avium enters macrophages primarily by a macropinocytosis-like mechanism and survives in a compartment that differs from that with extracellular phenotype." Cell Biol Int 28(5): 411-419.

Bermudez, L. E., F. J. Sangari, P. Kolonoski, M. Petrofsky and J. Goodman (2002). "The efficiency of the translocation of Mycobacterium tuberculosis across a bilayer of epithelial and endothelial cells as a model of the alveolar wall is a consequence of transport within mononuclear phagocytes and invasion of alveolar epithelial cells." Infect Immun 70(1): 140-146.

Bermudez, L. E. and L. S. Young (1994). "Factors affecting invasion of HT-29 and HEp-2 epithelial cells by organisms of the Mycobacterium avium complex." Infect Immun 62(5): 2021-2026.

Boiadjieva, S., C. Hallberg, M. Hogstrom and C. Busch (1984). "Methods in laboratory investigation. Exclusion of trypan blue from microcarriers by endothelial cells: an in vitro barrier function test." Lab Invest 50(2): 239-246.

Brinkmann, V., U. Reichard, C. Goosmann, B. Fauler, Y. Uhlemann, D. S. Weiss, Y. Weinrauch and A. Zychlinsky (2004). "Neutrophil extracellular traps kill bacteria." Science 303(5663): 1532-1535.

Cirillo, J. D., S. Falkow, L. S. Tompkins and L. E. Bermudez (1997). "Interaction of Mycobacterium avium with environmental amoebae enhances virulence." Infect Immun 65(9): 3759-3767.

De Assis, M. C., A. O. Da Costa, T. C. Barja-Fidalgo and M. C. Plotkowski (2000). "Human endothelial cells are activated by interferon-gamma plus tumour necrosis factor-alpha to kill intracellular Pseudomonas aeruginosa." Immunology 101(2): 271­ 278. 56

Fuchs, T. A., U. Abed, C. Goosmann, R. Hurwitz, I. Schulze, V. Wahn, Y. Weinrauch, V. Brinkmann and A. Zychlinsky (2007). "Novel cell death program leads to neutrophil extracellular traps." J Cell Biol 176(2): 231-241.

Godden, S. (2008). "Colostrum management for dairy calves." Vet Clin North Am Food Anim Pract 24(1): 19-39.

Greenberg, S. and S. Grinstein (2002). "Phagocytosis and innate immunity." Curr Opin Immunol 14(1): 136-145.

Gross, A., N. Dugas, S. Spiesser, I. Vouldoukis, C. Damais, J. P. Kolb, B. Dugas and J. Dornand (1998). "Nitric oxide production in human macrophagic cells phagocytizing opsonized zymosan: direct characterization by measurement of the luminol dependent chemiluminescence." Free Radic Res 28(2): 179-191.

Hines, M. E., 2nd, S. E. Turnquist, M. R. Ilha, S. Rajeev, A. L. Jones, L. Whittington, J. P. Bannantine, R. G. Barletta, Y. T. Grohn, R. Katani, A. M. Talaat, L. Li and V. Kapur (2014). "Evaluation of novel oral vaccine candidates and validation of a caprine model of Johne's disease." Front Cell Infect Microbiol 4: 26.

Hostetter, J., R. Kagan and E. Steadham (2005). "Opsonization effects on Mycobacterium avium subsp. paratuberculosis--macrophage interactions." Clin Diagn Lab Immunol 12(6): 793-796.

Janeway, C. J., C.A.; Travers, P.; Walport, M.; Shlomchik, A. (2001). Immunobiology, Garland Science.

Kim, J. S., H. Y. Kwon, W. H. Choi, C. Y. Jeon, J. I. Kim, J. Kim, J. Y. Lee, Y. S. Kim and J. B. Park (2003). "Phagocytosis of serum- and IgG-opsonized zymosan particles induces apoptosis through superoxide but not nitric oxide in macrophage J774A.1." Exp Mol Med 35(3): 211-221.

Koenig, G. J., G. F. Hoffsis, W. P. Shulaw, S. Bech-Nielsen, D. M. Rings and G. St- Jean (1993). "Isolation of Mycobacterium paratuberculosis from mononuclear cells in tissues, blood, and mammary glands of cows with advanced paratuberculosis." Am J Vet Res 54(9): 1441-1445.

Kuehnel, M. P., R. Goethe, A. Habermann, E. Mueller, M. Rohde, G. Griffiths and P. Valentin-Weigand (2001). "Characterization of the intracellular survival of Mycobacterium avium ssp. paratuberculosis: phagosomal pH and fusogenicity in J774 macrophages compared with other mycobacteria." Cell Microbiol 3(8): 551-566.

Kuo, T. T., K. Baker, M. Yoshida, S. W. Qiao, V. G. Aveson, W. I. Lencer and R. S. Blumberg (2010). "Neonatal Fc receptor: from immunity to therapeutics." J Clin Immunol 30(6): 777-789.

Liu, P., X. Wu, C. Liao, X. Liu, J. Du, H. Shi, X. Wang, X. Bai, P. Peng, L. Yu, F. Wang, Y. Zhao and M. Liu (2014). "Escherichia coli and Candida albicans induced 57

macrophage extracellular trap-like structures with limited microbicidal activity." PLoS One 9(2): e90042.

Mangum, J. B., J. I. Everitt, J. C. Bonner, L. R. Moore and A. R. Brody (1990). "Co­ culture of primary pulmonary cells to model alveolar injury and translocation of proteins." In Vitro Cell Dev Biol 26(12): 1135-1143.

Mayer, B., Z. Kis, G. Kajan, L. V. Frenyo, L. Hammarstrom and I. Kacskovics (2004). "The neonatal Fc receptor (FcRn) is expressed in the bovine lung." Vet Immunol Immunopathol 98(1-2): 85-89.

Medzhitov, R. (2001). "Toll-like receptors and innate immunity." Nat Rev Immunol 1(2): 135-145.

Miller, B. H., R. A. Fratti, J. F. Poschet, G. S. Timmins, S. S. Master, M. Burgos, M. A. Marletta and V. Deretic (2004). "Mycobacteria inhibit nitric oxide synthase recruitment to phagosomes during macrophage infection." Infect Immun 72(5): 2872­ 2878.

Mohanan, S., S. Horibata, J. L. McElwee, A. J. Dannenberg and S. A. Coonrod (2013). "Identification of macrophage extracellular trap-like structures in mammary gland adipose tissue: a preliminary study." Front Immunol 4: 67.

Nagy, B. (1980). "Vaccination of cows with a K99 extract to protect newborn calves against experimental enterotoxic colibacillosis." Infect Immun 27(1): 21-24.

Ott, S. L., S. J. Wells and B. A. Wagner (1999). "Herd-level economic losses associated with Johne's disease on US dairy operations." Prev Vet Med 40(3-4): 179­ 192.

Patel, D., L. Danelishvili, Y. Yamazaki, M. Alonso, M. L. Paustian, J. P. Bannantine, L. Meunier-Goddik and L. E. Bermudez (2006). "The ability of Mycobacterium avium subsp. paratuberculosis to enter bovine epithelial cells is influenced by preexposure to a hyperosmolar environment and intracellular passage in bovine mammary epithelial cells." Infect Immun 74(5): 2849-2855.

Perryman, L. E., S. J. Kapil, M. L. Jones and E. L. Hunt (1999). "Protection of calves against cryptosporidiosis with immune bovine colostrum induced by a Cryptosporidium parvum recombinant protein." Vaccine 17(17): 2142-2149.

Reiter, B. and J. H. Brock (1975). "Inhibition of Escherichia coli by bovine colostrum and post-colostral milk. I. Complement-mediated bactericidal activity of antibodies to a serum susceptible strain of E. coli of the serotype O 111." Immunology 28(1): 71­ 82.

Saif, L. J., D. R. Redman, K. L. Smith and K. W. Theil (1983). "Passive immunity to bovine rotavirus in newborn calves fed colostrum supplements from immunized or nonimmunized cows." Infect Immun 41(3): 1118-1131. 58

Stabel, J. R. (1998). "Johne's disease: a hidden threat." J Dairy Sci 81(1): 283-288.

Stabel, J. R. and T. J. Stabel (1995). "Immortalization and characterization of bovine peritoneal macrophages transfected with SV40 plasmid DNA." Vet Immunol Immunopathol 45(3-4): 211-220.

Stott, G. H., D. B. Marx, B. E. Menefee and G. T. Nightengale (1979). "Colostral immunoglobulin transfer in calves II. The rate of absorption." J Dairy Sci 62(11): 1766-1773.

Sweeney, R. W., R. H. Whitlock and A. E. Rosenberger (1992). "Mycobacterium paratuberculosis cultured from milk and supramammary lymph nodes of infected asymptomatic cows." J Clin Microbiol 30(1): 166-171.

Taylor, T. K., C. R. Wilks and D. S. McQueen (1981). "Isolation of Mycobacterium paratuberculosis from the milk of a cow with Johne's disease." Vet Rec 109(24): 532­ 533.

Tenant, R. and L. E. Bermudez (2006). "Mycobacterium avium genes upregulated upon infection of Acanthamoeba castellanii demonstrate a common response to the intracellular environment." Curr Microbiol 52(2): 128-133.

Thivend, P. T., R.; Guilloteau, P. (1980). Digestive adaptation in the preruminant. Digestive physiology and metabolism in ruminants. Y. T. Ruckenbusch, P. Lancaster, M.T.P. Press Ltd.: 561-586.

Vergne, I., J. Chua, S. B. Singh and V. Deretic (2004). "Cell biology of mycobacterium tuberculosis phagosome." Annu Rev Cell Dev Biol 20: 367-394.

Weber, S. S., J. Ducry and A. Oxenius (2014). "Dissecting the Contribution of IgG Subclasses in Restricting Airway Infection with Legionella pneumophila." J Immunol 193(8): 4053-4059.

Wentink, G. H., J. H. Bongers, A. A. Zeeuwen and F. H. Jaartsveld (1994). "Incidence of paratuberculosis after vaccination against M. paratuberculosis in two infected dairy herds." Zentralbl Veterinarmed B 41(7-8): 517-522.

Zhao, B., M. T. Collins and C. J. Czuprynski (1997). "Effects of gamma interferon and nitric oxide on the interaction of Mycobacterium avium subsp. paratuberculosis with bovine monocytes." Infect Immun 65(5): 1761-1766.

59

Chapter 3

Characterization of the inflammatory phenotype of Mycobacterium avium subspecies paratuberculosis using a novel cell culture passage model

Jamie L. Everman,1,2 Torsten M. Eckstein,3 Jonathan Roussey,4 Paul Coussens,4,5 John P. Bannantine,6 and Luiz E. Bermudez1,2

Department of Microbiology, College of Science, Oregon State University, Corvallis, Oregon1; Department of Biomedical Sciences, College of Veterinary Medicine, Oregon State University Corvallis, Oregon2; Department of Microbiology, Immunology and Pathology, College of Veterinary Medicine and Biomedical Sciences, Colorado State University, Fort Collins, Colorado3; Comparative Medicine and Integrative Biology Program, Michigan State University, East Lansing, Michigan4; Department of Microbiology and Molecular Genetics, Michigan State University, East Lansing, Michigan5; National Animal Disease Center, USDA Agricultural Research Service, Ames, Iowa6

Microbiology – In Review

60

Abstract Understanding the pathogenic mechanisms of Mycobacterium avium subspecies paratuberculosis (MAP) and the host responses to Johne’s disease is complicated by the multifaceted disease progression, late-onset host reaction, and the lack of available ex vivo infection models. Here, we describe a novel cell culture passage model that mimics the course of infection in vivo. The developed model simulates the interaction of MAP with the intestinal epithelial cells, followed by infection of macrophages, and return to the intestinal epithelium. Using this model we determined that during the initial stage of infection of MDBK epithelial cells, MAP infection triggers no inflammatory response. After passage through macrophages, bacterial re-infection of MDBK epithelial cells was associated with increased levels of the pro-inflammatory signals such as IL-6, CCL5, IL-8, and IL-18, paired with decreased levels of TGF-β. Transcriptome analysis of MAP from each stage of epithelial cell infection identified increased expression of lipid biosynthesis and lipopeptide modification genes in the inflammatory phenotype of MAP. Total lipid analysis by HPLC-ES/MS indicates different lipidomic profiles between the two phenotypes and a unique set of lipids in the inflammatory MAP phenotype. We provide supporting evidence to the validity of our model by studying the presence of selected upregulated lipid-modification gene transcripts in samples of ileal tissue from cows diagnosed with Johne’s disease. By using the cell culture passage model, we provide evidence that MAP alters its lipid composition during intracellular infection and acquires a pro-inflammatory phenotype, which likely is associated with the inflammatory phase of Johne’s disease. 61

Introduction Mycobacterium avium subspecies paratuberculosis (MAP) is the causative agent of Johne’s disease, a chronic intestinal inflammatory disease that affects ruminants worldwide. The disease is particularly devastating to the dairy industry, with an economic loss estimate as high as $1.5 billion annually in the United States alone (Stabel 1998, Ott, Wells et al. 1999). The development of new approaches for control of the disease is needed and depends on increased understanding of the pathogenesis of MAP and its interaction with the ruminant host. However, investigation of the molecular mechanisms that drive the disease is made difficult by factors including the long incubation period between the silent subclinical infection and the severe clinical stage of the disease, and the facilities and costs associated with housing experimentally infected animals during the development of disease. Initiation of infection occurs early in the life of a calf; however, clinical signs of disease do not appear until years after initial exposure. Infection can remain undetected for a number of years and antigen-specific and immune signal-based diagnostic tests usually fail to identify animals in the pre-clinical stages of the disease. When clinical signs emerge it is too late for therapeutic intervention, and the animal producer resorts to the culling and removal of diseased animals from the herd. The phenotype of a bacterial population plays an important role in its ability to cause and maintain a long term infection in the host, as it has been shown with pathogens such as Neisseria gonorrhoeae (Hagblom, Segal et al. 1985), Salmonella Typhimurium (Diard, Garcia et al. 2013), Pseudomonas aeruginosa (Penketh, Pitt et al. 1983), and Mycobacterium tuberculosis (Ryan, Hoff et al. 2010). Thus, environmental and host factors may potentially influence the phenotype of MAP prior to and during infection. Early studies on this phenomenon have identified a variety of phenotypes between geographically distinct MAP isolates (Whittington, Marsh et al. 2011), illustrated proteomic differences between laboratory culture and mucosal derived isolates (Weigoldt, Meens et al. 2011). MAP develops more invasive phenotypes in response to both the hyperosmolar environment of raw milk and growth within the mammary epithelium (Patel, Danelishvili et al. 2006). As phenotypic changes are quite pervasive among pathogens and utilized to drive 62

disease, further investigation is needed to determine if differing MAP phenotypes develop while within the host over the course of Johne’s disease. There is a limited number of well-defined models used to study the interaction between MAP and the mammalian host. At present, simple in vitro culture models of single cell monolayers are used to explore the interaction of MAP with the epithelial intestinal cell or with phagocytic macrophage cells. Mouse models are typically employed to investigate initiation and prevention of infection during the early stages of Johne’s disease (Bermudez, Petrofsky et al. 2010, Scandurra, de Lisle et al. 2010, Bannantine, Everman et al. 2014), while goat and calf models are more commonly used to study vaccine efficacy and the later stages of the disease (Khare, Lawhon et al. 2012, Facciuolo, Kelton et al. 2013, David, Barkema et al. 2014, Hines, Turnquist et al. 2014). While large animals provide useful models since they are the natural ruminant hosts of MAP and progress through the entire spectra of disease states, they come with a variety of disadvantages. Pitfalls include the necessity of large spaces, extended time courses, and high costs, as animal infection must be allowed to progress for many years prior to development of the clinical signs of disease. In light of the drawbacks of current models, a more elaborate in vitro system is needed to mimic and study the different environments encountered by MAP during infection. Natural infection within ruminant hosts begins with the uptake of MAP by the intestinal epithelium (Bermudez and Young 1994, Bermudez, Petrofsky et al. 2010), followed by uptake by tissue macrophages (Buergelt, Hall et al. 1978, Sigurethardottir, Valheim et al. 2004), and tissue dissemination likely via macrophages and the lymphatic system to sites including but not limited to the lymph nodes, mammary tissue, liver, and epithelial tissue (Antognoli, Garry et al. 2008). In this study, we describe the development of a novel in vitro cell culture passage model which mimics the passage of bacteria from its uptake by the intestinal epithelium, spread to the tissue phagocytes, and ultimate return to the intestinal epithelium during the later stages of infection. We demonstrate that a predominant pro-inflammatory immune response is mounted upon sequential passage of bacteria and infection of epithelial cells. Our findings led us to characterize intracellular MAP populations into non-inflammatory and inflammatory phenotypes. At the genetic level, lipid 63

biosynthesis genes are more highly expressed in the inflammatory MAP phenotype, while lipidomic analysis shows that the inflammatory MAP phenotype has a distinct lipid profile compared to that of the non-inflammatory MAP phenotype. Using the cell culture passage model, we examined the inflammatory phenotype of MAP and tested our hypothesis that the constantly changing phenotypes of MAP may trigger the transition between the multiple disease stages during the course of infection. These bacterial changes could play a significant role in variable diagnostic results and low vaccination efficacy in populations of ruminants infected with or at risk of contracting MAP. Further examination of this phenomenon may provide a better platform for understanding and developing future diagnostic and treatment options.

64

Methods and Materials

Bacterial preparation. Mycobacterium avium subspecies paratuberculosis strain K10 (ATCC BAA-968) was cultured at 37°C on 7H10 agar (BD; Franklin Lakes, NJ) supplemented with casein hydrolysate (1g/L; BD), 10% (v/v) oleic acid, albumin, dextrose, and catalase (OADC; Hardy Diagnostics; Santa Maria, CA), and ferric mycobactin J (2 mg/L; Allied Monitor, Fayette, MO) for 3-4 weeks. Prior to experiments, a bacterial suspension was made in HBSS (Corning, Corning, NY), passed through a 22-gauge needle to disperse clumps, and allowed to settle for 10 minutes. The top half of the inoculum was used as a single-cell suspension for experiments as described previously (Patel, Danelishvili et al. 2006).

Mammalian cell culture. Madin-Darby bovine kidney (MDBK) epithelial cells (CCL-22) and RAW 264.7 macrophage cultures (TIB-71) were obtained from the American Type Culture Collection (ATCC; Manassas, VA). Both cells lines were cultivated in Dulbecco’s Modified Eagle’s Medium (DMEM) supplemented with 10% heat-inactivated fetal bovine serum (FBS; Gemini Bio-Products; West

Sacramento, CA), at 37°C in 5% CO2.

In vitro cell culture passage model. Cell cultivation and infection schedule was completed as described (Figure 3.1). MDBK cells were seeded in T-75 tissue culture flasks and grown to 80% confluence. MAP was prepared as described above and used to infect monolayers at an MOI of 50:1 for 4 hours at which time media was removed, cells were washed 2 times with HBSS, and replaced with fresh media. MAP infection proceeded for 24 hours at which time modified differential centrifugation was used to isolate bacteria from host cells as described (McNamara, Tzeng et al. 2012). Cells were lysed with 0.1% triton X-100 for 15 minutes and lysate was centrifuged for 15 minutes at 2,000 × g at 4°C. Pellet was suspended in 0.1% triton X-100 and centrifuged for 3 minutes at 60 × g at 4°C to remove intact cells and cell debris. The MAP-containing supernatant was collected, bacteria pelleted at 2,000 × g at 4°C for 10 minutes. Bacterial pellet was suspended in HBSS, 65

centrifuged again at 60 × g at 4°C to pellet any residual cell debris, and MAP- containing supernatant was collected and centrifuged at 5,600 × g at 4°C to pellet intracellular bacteria. The MAP pellet was suspended in HBSS (primary MAP population) and used to infect RAW 264.7 macrophages. Infection was allowed to proceed for 72 hours. Macrophages were lysed and processed using differential centrifugation as described above to isolate intracellular bacteria. Suspended pellets were used to infect MDBK cells for 24 hours, at which time differential centrifugation was used to isolate the intracellular bacteria (secondary MAP population). Mock infection passages were completed using HBSS in the absence of bacteria to account for presence of host material after differential centrifugation. MAP and mock primary and secondary populations were frozen at -80°C for downstream analysis.

Mammalian RNA Extraction and Analysis. After progression through the passage model described above (Figure 3.1) RNA from MDBK epithelial cells was extracted. DNase treatment was conducted using the RNeasy Mini Kit as per manufacturer’s instructions (QIAGEN, Valencia, CA). RNA samples were verified for absence of DNA, quality verified and quantified. cDNA was synthesized using the iScript reverse transcription supermix for Real-Time qPCR (RT-PCR) (BioRad; Hercules, CA) and transcripts were analyzed using the CFX Connect Real-Time PCR Detection System (BioRad) using primers listed in Table 3.1. PCR amplifications were run as follows: 95°C for 3 minutes, 35 cycles of 95°C for 30 seconds, 53°C or 60°C for 30 seconds (see Table 3.1 for anneal temperatures), and 72°C for 20 seconds, followed by a final extension at 72°C for 5 minutes, and a melt curve from 50°C to 95°C to confirm amplicon size. Relative change in gene expression was quantified using the

ΔΔCt method (Livak and Schmittgen 2001). β-actin was used as an internal normalization control for the amount of mammalian cDNA added to each reaction, and mock infection samples served as a baseline control for the immune signal profile of host cells that undergo the passage model in the absence of bacteria.

66

Intracellular bacterial RNA extraction. The cell culture passage model was conducted (Figure 3.1) and total intracellular bacterial RNA was isolated as described (Wren and Dorrell 2002). Briefly, infected monolayers were lysed with guanidine thiocyanate (GTC) buffer (4M guanidine thiocyanate, 0.5% N-Lauryl sarcosine, 30mM sodium citrate, pH 7.0) with 0.1M β-mercaptoethanol, and lysates were pelleted at 2,700 × g at 4°C. Pellet was washed twice with GTC buffer and pelleted at 2,700 × g at 4°C. Intact bacterial pellet was lysed in Trizol (Life Technologies, Carlsbad, CA) with 0.5 ml of 0.1 mm glass beads by bead beating in a reciprocal bead beater for 20 seconds each, incubated on ice for 1 minute, and repeated 3 times. Chloroform:isoamyl alcohol (24:1; Ambion; Carlsbad, CA) was added and mixture was centrifuged at 9,500 × g at 4°C. Aqueous layer was collected, equal volume chloroform added, mixed, and centrifuged at 9,500 × g at 4°C. Aqueous layer was collected and nucleic acids were precipitated with 0.7 volume isopropanol and 15 µg of linear polyacrylamide overnight at -20°C. Nucleic acids were pelleted at 9,500 × g at 4 °C for 30 minutes, washed with 70% ethanol, and suspended in nuclease-free water. DNase treatment and RNA clean-up were performed using the RNeasy Mini Kit (Qiagen; Valencia, CA) as per manufacturer’s instructions. RNA preparation was verified for absence of DNA contamination and quantified. For each sample, 100 ng of RNA was amplified using the Message AmpII-Bacteria RNA Amplification Kit (Ambion, Grand Island, NY) as per manufacturer’s instructions. Amplified RNA (aRNA) was synthesized using biotin-labeled UTP nucleotides (Ambion) and analyzed for quality on an Agilent Bioanalyzer 2100 at the Center for Genome Research and Biocomputing at Oregon State University. Samples were aliquoted and stored at -80°C prior to microarray analysis.

Bovine CCL5 ELISA. The cell culture passage model was conducted as described above. Supernatant from the primary and secondary sets of MAP and mock infected MDBK epithelial cells were collected after 36 hours of infection. For control samples MDBK cells were infected with DMEM alone or lipopolysaccharide from Escherichia coli 0111:B4 (100 ng/ml; Sigma-Aldrich) and supernatants collected. Samples were filtered through a 0.2 µm filter and used to measure the secreted 67

chemokine CCL5 during infection using a Bovine RANTES ELISA Kit (NeoBioLab; Cambridge, MA) as per manufacturer’s instructions.

DNA microarray. The MAP K10 specific DNA oligo spotted arrays were designed and printed at the National Animal Disease Center (USDA-ARS, Ames, IA). Arrays were post-processed, blocked, and hybridized prior to use as per standard protocol (http://derisilab.ucsf.edu/microarray/protocols.html). For hybridization total RNA was fragmented to 200 base pair fragments by incubation for 30 minutes at 94°C in fragmentation buffer (100 mM potassium acetate, 30 mM magnesium acetate, and 40 mM tris-acetate). Array chips were hybridized with 6 µg of biotinylated RNA for 16 hours using an InSlide Hybridization Oven (Thomas Scientific, Swedesboro, NJ) in a humidified chamber at 45°C. Chips were washed for 5 minutes in 5x SSC, 0.01% SDS buffer, 3 minutes in 0.5x SSC buffer, and 30 seconds in 0.1x SSC buffer, stained with streptavidin-Cy3 (Life Technologies, Grand Island, NY) in 10% bovine serum albumin in HBSS for 1 hour, and washed again in 5x SSC/0.01% SDS buffer for 5 minutes, 0.5x SSC for 1 minute, 0.01x SSC for 10 seconds, and slides were centrifuged to dry. Array chip fluorescence was measured using the Axon4000 slide reader at the Center for Genome Resources and Bioinformatics at Oregon State University. Hybridization fluorescence was analyzed using Axon4000 software and spots containing bubbles, scratches, or inconsistency in fluorescence across the circle were censored from the data. Triplicate spots on each array were averaged and each slide was normalized to control spots containing hybridization buffer only. Triplicate biological replicates were averaged and the ratio of the fluorescence values of secondary MAP/primary MAP was determined. Fluorescence values were normalized to hybridization buffer spots within each individual array chip. To confirm microarray data, cDNA was synthesized from bacterial RNA using the iScript reverse transcription supermix for RT-qPCR (BioRad) and each sample was run in triplicate for each timepoint analyzed. Primers were designed for 100-200 base pair amplicons with one primer located within the 70 base pair oligomer spotted onto the microarray chip (Table 3.1). PCR amplifications were run as follows: 1 µl cDNA, 0.2 µM forward primer, 0.2 µM reverse primer, 12.5ul SsoAdvance SYBR Green 68

Supermix (BioRad), and water to 25 µl. PCR conditions were 95°C for 3 minutes, 35 cycles of 95°C for 30 seconds, 60°C for 30 seconds, 72°C for 20 seconds, followed by a final extension at 72°C for 5 minutes, and a melting curve from 50°C to 105°C to confirm amplicon size. The data discussed in this publication have been deposited in NCBI’s Gene Expression Omnibus (Edgar, Domrachev et al. 2002) and are accessible through GEO Series accession number GSE62566 (http://www.ncbi.nlm.nih/gov/geo/query/acc.cgi?acc=GSE62566).

Lipid analysis of intracellular bacterial phenotypes. Intracellular bacterial samples isolated from the in vitro cell culture model were collected and stored at ­ 80°C. Samples were lyophilized and whole cell lipids were extracted by chloroform:methanol (2:1), dried under nitrogen, and purified using Folsch wash (chloroform:methanol:water – 4:2:1) (Eckstein, Chandrasekaran et al. 2006). Isolated lipids were separated by high-performance liquid chromatography (HPLC) fractionation and analyzed in both positive and negative mode by electrospray mass spectrometry (ES/MS) as previously described (Sartain, Dick et al. 2011) at the Central Instrument Facility at Colorado State University (Fort Collins, CO). Total spectra of lipid ions of 200 to 3000 m/z are reported and phenotype specific peaks were determined by the subtraction of the ion abundance of each phenotype from the other.

MAP-infected bovine ileal tissue. MAP-infected intestinal tissue was obtained from cattle involved in a previously approved study at Michigan State University. From dissected ileal tissue, 1 cm3 biopsy samples were collected, washed twice in PBS, snap frozen in liquid nitrogen, and stored at -80°C. Specimens were thawed in RNAlater-ICE tissue transition solution (Life Technologies; Carlsbad, CA) as per manufacturer’s instructions. 200 mg of tissue was dissected from each sample and homogenized in GTC buffer with 3.2 mm stainless steel beads for 2 minutes at setting 8 in a Bullet Blender (Next Advance; Averill Park, NY). Homogenate was transferred to a new tube and intact bacteria and debris pelleted at 5,500 × g for 5 69

minutes at 4°C. Pellet containing bovine tissue debris and MAP was collected in GTC buffer and intracellular bacterial RNA was purified as described above. For complete removal of genomic DNA, samples first underwent DNase treatment and RNA clean-up (Qiagen), then processed with a second DNase treatment with 20 U RNase-free DNase (Roche; Basel, Switzerland), and 20 U RNase-OUT (Life Technologies) for 20 minutes, followed by RNA clean-up (Qiagen). Samples were tested for presence or absence of DNA by PCR for bovine β-actin using primers listed in Table 3.1 and were amplified as follows: 95°C for 30 seconds, 60°C for 30 seconds, 68°C for 30 seconds and repeated for 40 cycles. Bacterial RNA was enriched using the MICROBEnrich Kit (Ambion) as per manufacturer’s instructions. cDNA was prepared using the iScript cDNA Synthesis kit (Bio-Rad) and used as DNA template for PCR detection of bacterial transcripts. PCR amplification reactions were conducted using the primer sets in Table 3.1 and amplification proceeded as described above. Amplicons were resolved on a 1% agarose gel and band intensities analyzed using Image J software (Schneider, Rasband et al. 2012).

Statistical analysis and data interpretation. Results are reported as the mean of three repeated experiments ± standard deviation. Statistical comparisons between experimental groups and control groups were determined using the Student’s t test with p<0.05 denoting statistical significance. GraphPad Prism version 6.0 software was used for the construction of graphs, data interpretation, and statistical analysis. 70

Results

Development of a novel in vitro cell culture passage model To be able to readily understand the progression of MAP pathogenesis, we developed an in vitro cell culture passage model that mimics the interactions between the bacterium and the host intestine over the course of infection. The goal of this study was to investigate the bacterial and cellular mechanisms and their changes over the course of infection, specifically focusing on the changes that occur within the epithelial cells between the early and late stages of infection. Our illustrated model (Figure 3.1) consists of consecutive passages of MAP through host cells starting with MDBK epithelial cells, continuing with dissemination into macrophages, and the final return to the intestinal epithelium. The cellular progression design of our model replicates the stages that MAP goes through during infection of the host. Upon oral uptake of MAP from either the environment or fecal-contaminated food and water, MAP first interacts with and invades the intestinal epithelium (Bermudez, Petrofsky et al. 2010). After translocation through the epithelial cells, bacteria are ingested by local and infiltrating macrophages in response to infection where MAP is able to reside and replicate. Following macrophage infection, we hypothesize that during the late stage of infection bacteria are released by macrophages, either during granuloma formation, cell mediated death, or bacterial-mediated escape to encourage spread of the disease. This release would result in the subsequent return to and infection of the intestinal mucosa during the late stage of the disease and is supported by in vivo evidence that MAP can be identified within the intestinal epithelium during the late stage of Johne’s disease in ruminants (Antognoli, Garry et al. 2008, Khan, Chaudhry et al. 2010). Time points were determined based on the relative progression of the disease in the host. Invasion and translocation through the intestinal epithelium occurs within a short amount of time following infection and thus we determined 24 hour infection of each epithelial infection stage would be effective in modeling the infection. Furthermore, intracellular infection with macrophages can persist for long amounts of time and require the intracellular environment for efficient growth, replication, and evasion of the immune response. Due to this we chose 72 hour 71

timepoint for macrophage infection as an appropriate representation of the survival and time within macrophages in our model. As our focus was on the hypothesized changes occurring between each stage of epithelial cell infection, we collected nucleic acid, protein, and lipids from both the mammalian cells and the isolated bacterial samples from both the primary and secondary populations of the infection model for comparison and further analysis. The development of the in vitro passage model described here represents a novel technique in which the intricate host-microbe interactions during various stages of infection may be analyzed in a simplified manner compared to current animal infection models for Johne’s disease.

Serial passage of MAP initiates inflammatory response in bovine epithelial cells To investigate if there is a change in the host immune response to infection with cell-culture passaged compared with non-passaged bacteria, we employed the model described in Figure 3.1. Complete passages were conducted with MAP bacteria, as well as passages with HBSS to serve as a control (mock) infection to account for any carryover of mammalian components during isolation of intracellular MAP. MDBK epithelial cell RNA was collected 24 hours after each MAP or mock infection, cDNA was analyzed for bovine cytokine and chemokine signals, and the expression ratios of MAP infected MDBK epithelial cells to mock infected MDBK cells were determined (Figure 3.2a). Gene expression analysis revealed that the secondary epithelial cell population produced a different gene expression profile of immune signals than the primary infected MDBK cells in the model. An increase in transcripts of the pro- inflammatory signals IL-6 and IL-8 was seen only in the second infected MDBK cell population. This second MDBK cell population also showed a significant decrease in the expression of the anti-inflammatory signal TGF-β compared to primary infected MDBK cells (p<0.005). While secondary epithelial cells infected with MAP slightly increased expression of the pro-inflammatory signals IL-18 and CCL5 when compared to mock infected cells, primary cells infected with MAP appeared to suppress the expression of CCL5 and IL-18. Together, the secondary infected MDBK cells had a significant increase in the transcript expression of CCL5 (p<0.05) and IL-18 (p<0.005) compared to MAP infected primary population of MDBK cells 72

in our model. To confirm our findings, ELISA was performed to detect the secreted CCL5 protein in the supernatants of MAP-infected MDBK cells from the passage model (Figure 3.2b). Supernatant samples collected from the secondary MAP- infected epithelial cells demonstrated a higher level of CCL5 protein production and secretion than the measured expression level of CCL5 from the primary set of infected MDBK epithelial cells. Collectively, our results indicate that early infection with unpassaged MAP in the primary MDBK epithelial population induces a decreased inflammatory state within epithelial cells. Conversely, the secondary MDBK epithelial cell population infected with cell culture passaged MAP exhibits a more inflammatory profile by means of increased pro-inflammatory (IL-6, IL-8, IL­ 18, and CCL5) cytokine and chemokine expression and by decreased anti- inflammatory (TGF-β) signals detected. These data suggest that during passage through host cells, MAP likely develops a novel phenotype responsible for the induction of a more pro-inflammatory immune response during infection with intracellular-conditioned MAP.

Bacterial gene expression changes during in vitro passage model To determine the bacterial phenotype change during the cell culture passage model, we profiled that MAP gene expression at each stage of MDBK epithelial cell infection. Intracellular MAP was collected from both primary and secondary MDBK cell populations after 24 hours of infection and RNA from intracellular MAP was extracted and hybridized to MAP K10 DNA arrays for transcriptome analysis. Genome-wide expression analysis was completed by comparing the phenotype of MAP isolated from the secondary epithelial host cells to MAP isolated from the primary set of MDBK cells (Table 3.2). Our analysis identified 52 genes upregulated more than 2-fold in the inflammatory phenotype of MAP when compared to the non­ inflammatory primary phenotype. The upregulated genes catagorized into 3 general categories: lipid biosynthesis and metabolism (MAP2974c, MAP3121, and MAP3763c), cell wall remodeling (MAP1584c, MAP0385, MAP2604c and MAP1137c), and cellular metabolism (MAP2660, MAP0808, and MAP1485c). A regulator of the transcription factor sig8 (MAP3111c) was upregulated as well as a 73

variety of hypothetical proteins that have yet to be fully described (MAP3433 and MAP3516). Our results confirm the original hypothesis that intracellular passage of MAP through host cells elicits a change in gene expression, particularly in genes involved in lipid and cell wall components of the bacteria.

Lipidomic changes in MAP during passage model infection Gene expression studies uncover substantial information on the transcriptional state of an organism. The production of macromolecules including lipids can require a variety of genetic components, non-ribosomal peptide synthetase enzymes, and post-translational modifications prior to production of a functional moiety. This can result a skewed ratio of gene expression transcripts to translated functional lipid or protein components. As lipids play a pivotal role in mycobacterial pathogenesis, we analyzed the total lipid composition of both MAP phenotypes before and after serial passage through host cells in order to ascertain the changes in the bacterial lipidome. Intracellular MAP were collected after 24 hours of infection from both the primary and secondary MDBK cell populations and whole cell lipids were extracted and analyzed using HPLC-ES/MS (Figure 3.3). The complete lipidome spectra between the primary passaged MAP population (Figure 3.3a) and secondary passaged inflammatory MAP population (Figure 3.3b) have similarities in the composition of common small ion fragments from 200 to 300 m/z, and in mycolic acid ion patterns found from 1000 to 1500 m/z. Conversely, there are dramatic differences in the lipid profiles detected from 300 to 900 m/z between the MAP phenotype isolated from the primary epithelial cell infection (Figure 3.3c) and the inflammatory MAP phenotype isolated from the secondary epithelial cell infection (Figure 3.3d). Focusing on the differences within that range, whole cell lipidome profiles were analyzed for lipids specific to each population. The inflammatory phenotype of MAP is composed of a much more diverse set of lipids, including a wider array of specific sizes and types of lipids (Figure 3.3f), compared to those specific to the primary MAP population (Figure 3.3e). The mycobacterial lipid Para-LP-01 has been described as an important cell wall component of MAP (Eckstein, Chandrasekaran et al. 2006). Para-LP-01 is known to 74

decrease in abundance following intracellular infection (Alonso-Hearn, Eckstein et al. 2010). We used this molecule as an internal validation marker for our model (Figure 3.4). When comparing the two phenotypes, Para-LP-01 was found in high abundance in the lipidomic spectra of the primary MAP phenotype at an intensity of 7.7x104 (Figure 3.4a), while the lipid composition of the cell culture passaged inflammatory phenotype contained a significantly lower abundance of Para-LP-01 at an intensity of 3.0x103 (Figure 3.4b). The loss of Para-LP-01 and differences in other lipids in the late infection phenotype suggest that the change in lipid composition may contribute to the inflammatory status within the intestinal epithelium during MAP infection of the ruminant host.

Identification of upregulated MAP transcripts from infected tissues In order to validate the in vitro cell culture passage model, we selected upregulated MAP genes from the secondary inflammatory phenotype and investigated the presence of these transcripts in the tissue samples of MAP infected animals. Two cattle at a Michigan dairy were identified as MAP-positive by standard diagnostic techniques including fecal PCR and ELISA detection of MAP antibodies (Supplemental Table 3.3). The level of bacterial burden measured by qPCR and observations of peripheral blood mononuclear cell counts indicated that one animal was in the advanced stages of Johne’s disease (animal #6211; Figure 3.5b), while another was relatively healthy and in good body condition (animal #1688; Figure 3.5a). Tissue samples were collected from various locations within the terminal ileum during necropsy, and total RNA was isolated and analyzed for the presence of bacterial transcripts (Figure 3.6a). Within each tissue sample the MAP insertion element IS900 was detected in abundance indicating MAP infection in each particular tissue section. The presence of MAP3121, MAP3433, MAP2974, and MAP1584 transcripts were detected within at least one ileal tissue sample from each of the infected animals. It was observed that detection levels within each tissue biopsy appeared to be correlated with areas that were observed to have more significant lesions during necropsy (Lesion 2 in animal #1688 and lesion 1 in animal #6211). In addition to identification of these transcripts, quantification of band intensity of the 75

detected transcripts determined that expression levels were similar to those reported from the microarray experiment (Figure 3.6b). Furthermore, band intensity determined that levels of the inflammatory phenotype-related MAP transcripts in animal #6211 were much higher than those found in animal #1688. Interestingly, animal #6211 was in a more advanced clinical stage of Johne’s disease as observations indicated the animal had stopped eating and was almost unable to stand. Necropsy and diagnostics indicated a greater degree of inflammation within the ileal tissue and almost 100 times more MAP IS900 DNA detected in fecal PCR tests. Alternatively, animal #1688 was visually healthy, active, and mobile and had a lower grade of inflammation observed upon gross examination of the ileal tissue (Supplemental Table 3.3). The identification of inflammatory MAP phenotype transcript markers within naturally occurring MAP-infected cattle, and the increase of these markers in tissue from an animal in the much more progressive stages of Johne’s disease, validates the findings reported using our novel cell culture passage model.

76

Discussion The shift in the host immune response is a key characteristic to the onset of inflammation within the intestinal tissue of ruminants and results in the distinct characteristics of Johne’s disease. While cell culture and animal models have uncovered intracellular survival mechanisms, little is known about the bacterial drivers behind the transition in the immune response and progression to the clinical stage of the disease. In this paper, we describe a novel in vitro cell culture passage model which mimics the progression of infection within the tissues of the host. Using this model, we have identified two different bacterial phenotypes that trigger opposite levels of inflammatory signals produced by infected epithelial cells, similar to the pattern of the infection in cows. These data indicate that MAP alters its phenotype during intracellular infection, the lipid composition of the bacteria change, and those changes potentially contribute to a shift in the inflammatory response by the infected host epithelial cells during Johne’s disease. Current disease models used to study Johne’s disease include in vitro cell culture systems provide important information on bacterial virulence mechanisms, yet the scope of infection is limited to the cellular level of a single particular cell line. Mouse models are used to investigate the ability of MAP to invade the intestinal mucosa and disseminate to peripheral tissues (Bermudez, Petrofsky et al. 2010), to determine preventative probiotics (Cooney, Steele et al. 2014), and to identify and validate vaccine candidates (Scandurra, de Lisle et al. 2010, Bannantine, Everman et al. 2014). Unfortunately, mouse models offer limited insight on the complete disease progression as mice do not exhibit intestinal inflammation in the same manner that ruminants do advanced stages of Johne’s disease. MAP infection of goat and calf models are used to study a variety of disease interactions including host gene expression (Khare, Lawhon et al. 2012) vaccine efficacy (Hines, Turnquist et al. 2014), both bacterial (Facciuolo, Kelton et al. 2013) and host (David, Barkema et al. 2014) biomarker discovery for enhanced diagnostics, and are used to track and develop mathematical shedding and transmission models (Mitchell, Whitlock et al. 2008); however, the excessive time and space costs of large animals make ruminant models challenging to use. Though recent efforts have been made to standardize 77

animal models used in Johne’s disease research (Hines, Stabel et al. 2007), historical literature lacks consistency. A review of the past century of Johne’s disease research sheds light on the highly variable study design in regard to age, number, and type of animal, bacterial infection strain, dosage, delivery method, and incubation time allowed prior to sample analysis (Begg and Whittington 2008). These inconsistencies allow past Johne’s animal studies to be open to a wide array of interpretation as to how those varying factors and the particular animal host used, accurately and consistently model the full stages of Johne’s disease infection. With this knowledge, we aimed to develop an in vitro passage model system composed of cell lines in order to simulate the more complex interactions occurring within the intestinal tissue of the host (Figure 3.1). By utilizing this model, we will be able to more readily understand both the host and bacterial mechanisms used during the invasion and infection process of the ruminant intestine, and discover unidentified changes or interactions that occur during the disease. During natural infection, ruminants are infected with MAP via intestinal epithelial cell uptake and the bacteria are ingested by resident and infiltrating macrophages in the epithelial tissue. MAP-containing macrophages can then traffic throughout the host, leading to system dissemination of the bacteria as demonstrated by the detection of MAP from a variety of organs including mesenteric lymph nodes, liver, muscle tissue, mammary lymph nodes, epithelium, and intestinal mucosa (Sweeney, Whitlock et al. 1992, Antognoli, Garry et al. 2008, Alonso-Hearn, Molina et al. 2009). Infected macrophages located in the epithelial tissue can then lead to the formation of granulomatous lesions within the mucosa. Once a high enough intracellular burden has been reached, MAP can trigger necrosis within infected bovine macrophages (Periasamy, Tripathi et al. 2013), suggesting that necrotic macrophages in the early stages of granuloma formation may be a contributor of returning MAP to the host epithelium during the progression of Johne’s disease. Our model described here, and the inflammatory signals measured during the differing stages of infection, are supported by the scientific literature which has reported the subclinical and clinical stages of Johne’s disease and analyzed the immunological responses specific to the mucosal intestinal tissue during each stage (Lee, Stabel et al. 78

2001, Buza, Mori et al. 2003, Weiss, Evanson et al. 2006, Stabel and Robbe- Austerman 2011, Khare, Lawhon et al. 2012). Here, we describe a more anti- inflammatory epithelial response to infection characterized by decreased expression of CCL5 and IL-18 paired with higher expression of the anti-inflammatory signal TGF-β in the primary set of MAP-infected epithelial cells. Experimental observations support the role of induced immune tolerance in mucosal tissue during early stages of Johne’s disease. TGF-β is capable of inducing ‘inflammatory anergy’ in human macrophage cells, as they maintain phagocytic capabilities, but are unable to mount a proper pro-inflammatory cytokine response (Smythies, Sellers et al. 2005). Weiss et al have described that although the total T-cell population increases during MAP infection in cattle, the majority of those cells are memory (CD2+CD62L- ) or regulatory (CD4+CD25+) cells (Weiss, Evanson et al. 2006), although it is unclear as to whether the cells deemed as regulatory T-cells in the study were under the control of FoxP3 activation. Additionally, host gene expression analysis has identified a variety of T-cell, cytokine, and additional immune cell signaling factors that are either suppressed or not expressed at all during the early stages of MAP infection within ligated ileal loops of MAP-infected neonatal calves (Khare, Lawhon et al. 2012). The suppression of CCL5 in our model is mirrored by the findings that MAP infection suppresses the expression of CCL5 (RANTES), as well as monocyte chemotractant protein-1 (MCP-1), in peripheral blood monocytes of cattle (Buza, Mori et al. 2003). The early phenotype of MAP may play a role in inducing a transient anti- inflammatory environment to gain a foothold for increased infection outcomes. We show that IL-18 is downregulated within epithelial cells during early MAP phenotype expression. IL-18 is important in immunomodulation during mycobacterial infection as IL-18-deficient mice are unable to stimulate interferon-gamma (IFNγ) and develop an excess of granulomatous lesions in response to Mycobacterium tuberculosis and Mycobacterium bovis BCG challenge (Sugawara, Yamada et al. 1999). IFNγ is a well-known immune signal for combating mycobacterial disease, serves as a marker used to aid in diagnosing MAP infection, and can be seen as early as 3 months after oral infection in cattle (Stabel and Robbe-Austerman 2011). Perhaps, by creating an early anti-inflammatory environment within the epithelium through the induction of 79

tolerance, preventing the expression of inflammatory mediators such as IL-6 and chemokines such as CCL5 and IL-8, and by suppressing IL-18, thus limiting the production of IFNγ, early MAP phenotypes create a microenvironment within the mucosal tissue which supports enhanced uptake, intracellular survival, and spread of the bacterium. This transient immunosuppressed environment could potentially allow MAP to gain an advantage and more successfully establish an infection within the host during the early stages of the disease. The clinical, late-stage of Johne’s disease is characterized by severe inflammation of the mucosal tissue within the intestine of the infected animal. Our model demonstrates that upon changing phenotypes, MAP elicits an increase in IL-6, similar to the RNA levels detected in ileal tissue collected from clinically MAP-infected cattle (Lee, Stabel et al. 2001). The abrogation of immune signal suppression, and subsequent increase in the production of the chemotactic agents IL-8 and CCL5 provide the components necessary for monocyte and macrophage recruitment which contribute to the development of MAP-containing granulomas within the intestinal tissue. Paired with increased chemokine responses, this shift allows for the influx and activation of macrophages, thus encouraging granuloma formation, increase in inflammatory signals, and production of high levels of IFNγ via increased IL-18 in the intestinal tissue. However, at this stage of infection, the pro-inflammatory response may not be sufficient to rid the host of infection. Instead, as granuloma formation and infiltrating immune cells attempt to combat the intestinal infection, the response results in uncontrollable inflammation and spread of granulomatous lesions throughout the tissue in the initiation of the severe clinical stage of Johne’s disease. It is well described that the alteration of the phenotype of a bacterium can be imperative to the response to and survival within a variety of environmental and host conditions. Several pathogens utilize particular phenotypes for enhanced infection: Neisseria meningitidis in its LPS and glycopeptides (Hagblom, Segal et al. 1985), Staphylococcus aureus reverts to a small colony variant to save metabolic energy and establish chronic infections (Tuchscherr, Medina et al. 2011), and Salmonella enterica serovar Typhimurium utilizes a biphasic growth cycle of virulent and avirulent phenotypes in order to maximize evolutionary stability and virulence 80

(Diard, Garcia et al. 2013). A variety of mycobacterial species are reported to readily alter their phenotypes in response to external stimuli. Mycobacterium tuberculosis is described to alter its phenotype, expression of virulence factors, and survival mechanisms in response to anaerobic environments (Gillespie, Barton et al. 1986), and granuloma formation (Karakousis, Yoshimatsu et al. 2004), and the close relative of MAP, Mycobacterium avium subspecies hominissuis, alters its phenotype upon the formation of microaggregates and biofilms (Babrak, et al In review) (McNabe, Tennant et al. 2011), in response to the metal matrix found within host phagosomes (Early and Bermudez 2011), and upon intracellular infection of macrophages and amoeba (Bermudez, Petrofsky et al. 2004, Harriff and Bermudez 2009). Furthermore, previous studies with MAP indicate that changing phenotypes in response to environmental and host conditions contributes to virulence and increased pathogenesis within the host (Patel, Danelishvili et al. 2006, Alonso-Hearn, Patel et al. 2008). Our data suggest that MAP changes its phenotype during the stages of intracellular infection and that these changes involve the alteration of whole cell and cell wall-associated lipids. These changes are pertinent as MAP has an abundant reliance on its lipid-rich cell wall for virulence and survival (Ehrt and Schnappinger 2007, Bansal-Mutalik and Nikaido 2014, Cambier, Takaki et al. 2014). The upregulated genes we identified in this study are involved in a number of lipid-related processes. The genes MAP2974c, MAP3121, MAP3763c, and MAP1584c are all involved in larger lipid synthesis gene clusters and open reading frames responsible for the production of multi-domain lipid synthetase enzymes required for proper lipid production. Other changes include an increase in the unique mycobacterial cell wall component mycocerosic acid (Rainwater and Kolattukudy 1985), MAP1456 (which is located in a larger cluster of genes containing ABC-transporters), and a hypothetical transcriptional regulator encoded by MAP3111c. Interestingly, rather than being responsible for the synthesis of one specific lipid, many of the genes identified in our study seem to be specific components of the larger multi-factorial pathways required for the breakdown and biosynthesis of lipids within the bacterium. The higher expression of these genes suggests that in tandem they have the ability to more globally impact the production of lipid components. Lipid analysis conducted in this 81

study identified ion fragment patterns of the lipidome and suggest that the change in gene expression results in an impact on total lipid composition. As seen from our data, the inflammatory phenotype of MAP is uniquely composed of a wide variety of types and sizes of lipids compared to the non-inflammatory MAP phenotype. Changes in lipid composition, cell wall associated molecules, or the abundance of such components have the ability to dramatically impact the virulence of the bacteria prior to and during infection. One major cell wall component known to change during intracellular infection is the high molecular weight lipopeptide Para-LP-01 which is found in abundance in the cell wall of MAP and is recognized by sera from infected cattle (Eckstein, Chandrasekaran et al. 2006). The decrease in Para-LP-01 abundance after intracellular infection (Alonso-Hearn, Eckstein et al. 2010) is also seen in the total lipid composition from intracellular MAP in our passage model with significantly higher levels of Para-LP-01 present in the bacterial lipids from the bacterial population isolated from the primary MDBK cell infection. These findings provide evidence that supports the validity of the model system described in this manuscript, and that the changing lipidome may play a role in the evolving stages of Johne’s disease. The shift in mycobacterial lipid composition in response to a changing environment could result in a variety of outcomes. The immunogenic nature of mycobacterial lipids is classically demonstrated by their addition to Freund’s adjuvant and its ability to elicit a robust immune response (Freund 1956). Mycobacterial lipid components are capable of stimulating inflammatory cytokines IL-6 and TNFα when coated onto latex beads and delivered to host immune cells (Geisel, Sakamoto et al. 2005). Alternatively, this change in bacterial components could be a useful strategy in suppressing particular inflammatory responses at certain times, producing virulence factors at specific times to trigger a particular stage of the disease, and produce an environment it needs to survive. The expression levels or variation of lipoarabinomannan (LAM) and mannose-capped LAM molecules results in a change in virulence or bacterial survival as LAM is, in part, responsible for the inhibition of maturation of mycobacterium-containing phagosomes and thus contributes to intracellular survival (Fratti, Chua et al. 2003). Not only does mannosylated LAM 82

increase intracellular survival, but MAP derived molecules are capable of suppressing macrophage inflammatory signals by prolonged stimulation of IL-10 during infection (Souza, Davis et al. 2013). Keeping each of these situations in mind, the alteration of bacterial lipids described in this study may provide a method of camouflage for the bacteria early during infection in order to avoid, or even actively suppress, the host immune response. An explanation for the changing lipid phenotype during infection can also be due to the energetic demands on the bacterial organism. The production of the lipids, specifically mycolic acids, which make up the majority of the cell envelope of MAP is metabolically costly. Mycocerosic acid, whose synthase (MAP2604c) is upregulated 2.68-fold in the inflammatory phenotype, is a dense lipid moiety based on a long 28- to 32-carbon backbone (Minnikin 1982). The production of the pentapeptide component of Para-LP-01 requires a large input of energy and resources as the gene that encodes for its production is ~20kb. The seemingly unnecessary production of such large molecules could channel required nutrients or lipid components away from pathways needed to produce functionally relevant cell wall molecules needed for the current infection stage at hand. This specifically rings true for intracellular populations which are already limited in the ions, fatty acid precursors, cholesterol, and metabolic building blocks necessary for intracellular virulence and survival (Tuchscherr, Medina et al. 2011, Subramoni, Agnoli et al. 2013). Using differing phenotypes for the initiation, survival, and transmission ensures that the pathogen is more successful, and energetically efficient, during each particular stage of disease. Lastly, we determine that upregulated transcripts identified in the inflammatory MAP phenotype are expressed in MAP-infected ileal tissue samples from cattle infected with Johne’s disease. Amplification of such transcripts from tissue samples indicate that animals testing positive for Johne’s, either at the most advanced stage of the disease or at a more subclinical stage of disease, express the genes identified in this study that are associated with the inflammatory phenotype of MAP. Our study design utilized cDNA synthesized from bovine MDBK epithelial cells to control for the potential cross reactivity to sequences in the bovine transcriptome. For a 83

molecular study such as this, healthy and uninfected ileal tissue would normally be chosen as the best control sample. Unfortunately, with MAP as ubiquitous as it is, and with detection and diagnosis of animals in the earliest stages of the disease being quite unreliable, we did not want to sacrifice and collect samples from animals which could be giving potential false-negative test results for use as our controls; therefore, we determined that the bovine transcriptome of cultured cells would serve as a reliable control for our particular research question. Not only were we able to validate that these inflammatory phenotype MAP transcripts were expressed during MAP infection in cattle, we were able to show that they are expressed at different levels dependent on the stage progression of Johne’s disease within the intestinal tissue of the animal, with higher levels identified during more advanced stages of disease. The development of the novel cell culture passage model described here enhances the ability to more readily study the intricate mechanisms occurring between the bovine host and MAP in the absence of a live animal. This model demonstrates that MAP develops an inflammatory phenotype over the course of infection, and initiates a more pro-inflammatory cytokine and chemokine response in the later stages of infection within the epithelial mucosa. The inflammatory phenotype is characterized by a distinct set of upregulated genes and contains a unique lipid composition from the non-inflammatory phenotypes of MAP. Identification of these lipid synthesis transcripts within MAP-infected bovine tissues provides evidence which supports the validity of our model and offers evidence towards the idea that the changes in MAP, its gene expression, and its lipid composition could be a driver behind the multiple stages of Johne’s disease. These changes could be a notable pathogenic mechanism of interest to study, as the characterization of the function of these lipids may provide novel compounds for enhanced vaccine targets and therapeutic strategies for animals in the advanced stages of Johne’s disease.

84

Acknowledgements We would like to thank Caprice Rosato for technical assistance with the Axon4000 Slide Reader and hybridization analysis, and Sadie Rice for technical assistance with CCL5 ELISA assays. This work was funded by a foundation grant from the Department of Microbiology, College of Science, Oregon State University. . 85

BOVINE PRIMERS Name Sequence (5’ to 3’) Anneal Source β –actin_B_Fwd CGCACCACTGGCATTGTCAT 60°C (Konnai, Usui et al. 2003) β –actin_B_Rev TCCAAGGCGACGTAGCAGAG 60°C (Konnai, Usui et al. 2003) IL-6_B_Fwd TCCAGAACGAGTATGAGG 53°C (Konnai, Usui et al. 2003) IL-6_B_Rev CATCCGAATAGCTCTCAG 53°C (Konnai, Usui et al. 2003) IL-8_B_Fwd TGCCTCATGTACTGTGTGGG 60°C (Weiss, Evanson et al. 2002) IL-8_B_Rev GGGATAAAGAAACCAAGGCG 60°C (Weiss, Evanson et al. 2002) TGF-β_B_Fwd TTCTTACCCTCGGAAAATGCCATCC 60°C This paper TGF-β_B_Rev CCATCAATACCTGCAAAGCGTG 60°C This paper CCL5_B_Fwd CATGGCAGCAGTTGTCTTTATCA 53°C This paper CCL5_B_Rev CTCTCGCACCCACTTCTTCTCT 53°C This paper IL-18_B_Fwd TCTTTGAGGATATGCCTGATTCTG 60°C This paper IL-18_B_Rev CAGACCTCTAGTGAGGCTGTCCTT 60°C This paper MAP PRIMERS MAP_16s_Fwd CGAACGGGTGAGTAACACG 60°C This paper MAP_16s_Rev TGCACACAGGCCACAAGGGA 60°C This paper IS900_Fwd_C GATGGCCGAAGGAGATTG 60°C This paper IS900_Rev_C CACAACCACCTCCGTAACC 60°C This paper MAP2131_Fwd CGTCGATGGTCAGGGCCGA 60°C This paper MAP3121_Rev AGCAGATTCGGATGTCGGCGG 60°C This paper MAP3433_Fwd CCCGAAGTCGACGAGGCGTT 60°C This paper MAP3433_Rev TGCCGACCTGCAGCCAGAAG 60°C This paper MAP2974c_Fwd GGCGTGGGAACCACCAAAAGTG 60°C This paper MAP2974c_Rev GATCTGCGGGTGCACGACCTG 60°C This paper MAP1584c_Fwd GGCGATCTTGTCGCTGACCTCG 60°C This paper MAP1584c_Rev TCAACCGGCTGACCGACCCG 60°C This paper

Table 3.1: List of primer sets for bovine and bacterial transcript analysis Primers used for real time-PCR analysis of bovine RNA transcripts during in vitro cell culture passage model infection, confirmation of microarray expression results, and detection of MAP transcripts from infected bovine ileal tissue samples.

86

Figure 3.1: In vitro cell culture passage model. A schematic of the cell culture passages used to mimic the path MAP takes during infection on ruminants. The timepoints analyzed are indicated above, and samples collected for analysis of host- microbe interactions during cell culture passage are shown below.

87

Figure 3.2: Inflammatory response of MBK epithelial cells during in vitro cell culture passage model. In vitro cell culture passage model was completed with MAP-infection and mock- infection and RNA (A) or supernatants (B) were collected from the primary and secondary MDBK cell populations. (A) Relative expression of IL-6, IL-8, TGF-β, CCL5, and IL-18 was analyzed by qPCR. Samples were normalized to β-actin expression in each sample and the relative fold changes of MAP-infected to mock- infected transcripts are reported. (B) Supernatants were collected after 36 hours of infection of each epithelial cell population and concentration of CCL5 was measured using ELISA. Data represent the mean ± SD of 2 independent experiments analyzed by qPCR or ELISA in duplicate (*p-value <0.05, **p-value <0.005 by Student’s t- test).

88

Gene Fold Gene Description Biological Process Change MAP2974c 3.20 cyclopropane-fatty-acyl-phospholipid synthase 1 Lipid Biosynthesis MAP3121 3.14 enoyl-CoA hydratase Lipid Metabolism MAP3433 2.99 hypothetical protein Unknown MAP3763c 2.98 conserved polyketide synthase associated protein 3 Lipid Metabolism MAP1584c 2.92 ATP-dependent Lon protease Stress Response Proteolysis MAP2660 2.85 NAD-dependent epimerase dehydratase Nucleic Acid Metabolism MAP0385 2.82 restriction endonuclease family protein DNA Binding MAP0808 2.81 molybdenum cofactor biosynthesis protein Metabolism MAP3111c 2.81 regulator of sig8 Metabolism MAP3516 2.77 hypothetical protein Unknown MAP1456 2.76 hypothetical esterase lipase Metabolism MAP1485c 2.72 acyl-CoA synthetase Metabolism MAA2452 2.72 hypothetical protein Unknown MAP1137c 2.71 aminoglycoside tetracycline-transport membrane protein Membrane Transport MAP0350 2.70 short chain dehydrogenase Oxidoreductase MAP2604c 2.68 mycocerosic acid synthase Oxidoreductase MAP2239 2.67 mmpl4 protein Unknown MAP2751 2.67 21 kDa protein Unknown

Table 3.2: Microarray analysis of MAP phenotypes Intracellular MAP was isolated after 24 hours from each MDBK cell population during infection using the in vitro cell culture passage model. Extracted bacterial RNA was amplified to biotinylated aRNA, and hybridization to a MAP K-10 microarray slide was measured using a streptavidin-Cy3 probe on an Axon4000 slide reader. Data represent the mean value of 3 independent experiments with gene oligomers spotted in triplicate on each slide and are reported as the ratio of gene expression of MAP from secondary MDBK population to gene expression of MAP from the primary MDBK population.

89

Figure 3.3: Lipid profiles of MAP phenotypes Intracellular MAP populations were isolated after 24 hours of infection using the in vitro cell culture passage model and total lipid extraction was performed and analyzed using HPLC-ES/MS. The complete spectra of 200 to 3000 m/z is shown (A and B) as well as a more detailed representation of the spectra from 200 to 1100 m/z (C and D). The lipidome of the primary population phenotype (A and C) and the secondary 90

inflammatory phenotype (B and D) have distinct differences in both the presence and abundance of detected lipid ion fragments. Lipid ions unique to the primary population phenotype (E) and the secondary inflammatory phenotype (F) are shown. Unique lipid profiles were determined by subtracting abundance values of each lipidome from one another. The ion of 922.0098 m/z represents the internal standard which was added to each sample prior to analysis. Spectra in figure A and B are shown on the same x-axis; spectra of figures C-F are shown on the same x-axis range.

91

Figure 3.4: Abundance of cell wall lipid Para-LP-01 in MAP phenotypes Intracellular MAP populations were isolated after 24 hours of epithelial cell infection using the in vitro cell culture passage model and equal amounts of total lipids were analyzed using HPLC-ES/MS. The ion size of 918.66 m/z, known as para-LP-01 (arrow), was identified as running in identical fractions of HPLC eluate (data not shown) with intensity of the fragment in the primary MAP population at 7.7×104 (A), and intensity in the inflammatory MAP phenotype at 3.0×103 (B). The ion of 922.0098 m/z represents the internal standard which was added to each sample prior to analysis. Figures A and B are shown on the same x-axis range.

92

Figure 3.5: Bovine intestinal tissue samples Tissue samples from animals testing positive for Johne’s disease were collected from the terminal ileum of each animal. Animal #1688 (A) was in a much healthier state, while animal #6211 (B) was extremely weak, in poor body and immune condition, and in a much later stage of infection. Biopsy punches were taken from locations indicated (1-5) after photograph (A), and prior to photograph (B), and immediately snap-frozen for downstream sample processing.

93

Figure 3.6: MAP transcripts from infected bovine intestinal tissue Ileal samples were collected from cattle that tested positive for MAP infection, and tissue sections were processed for RNA extraction. Bacterial-mammalian RNA sample mixtures were enriched for bacterial RNA using the MICROBEnrich Kit and cDNA was used for PCR identification of inflammatory MAP phenotype transcripts (A). Gel band intensities were quantified using Image J and graphed (B) to show expression differences between the late-stage, severely infected animal (#6211; black bars) and the early stage infected animal (#1688; white bars). 94

Summary of Diagnostic Scores Animal ELISA Score Fecal PCR Score Body Condition/Observations 6211 2.884 16.8 - Extremely poor/emaciated - PBMCs abnormal 1688 2.909 23.6 - Good - PBMCs normal

Table 3.3: Supplemental - Diagnostic scores of cattle with Johne’s disease Animals were tested for MAP infection prior to necropsy and tissue collection. Standard detection of MAP antibodies was measured by ELISA and score reported (values >0.700 are considered positive for MAP infection). qPCR for MAP in the feces was completed and Ct values reported. For qPCR, any detection is indicative of a positive MAP identification and lower Ct value indicates a higher bacterial burden. Body condition and PBMC counts were conducted and observations are reported.

95

References

Alonso-Hearn, M., T. M. Eckstein, S. Sommer and L. E. Bermudez (2010). "A Mycobacterium avium subsp. paratuberculosis LuxR regulates cell envelope and virulence." Innate Immun 16(4): 235-247.

Alonso-Hearn, M., E. Molina, M. Geijo, P. Vazquez, I. Sevilla, J. M. Garrido and R. A. Juste (2009). "Isolation of Mycobacterium avium subsp. paratuberculosis from muscle tissue of naturally infected cattle." Foodborne Pathog Dis 6(4): 513-518.

Alonso-Hearn, M., D. Patel, L. Danelishvili, L. Meunier-Goddik and L. E. Bermudez (2008). "The Mycobacterium avium subsp. paratuberculosis MAP3464 gene encodes an oxidoreductase involved in invasion of bovine epithelial cells through the activation of host cell Cdc42." Infect Immun 76(1): 170-178.

Antognoli, M. C., F. B. Garry, H. L. Hirst, J. E. Lombard, M. M. Dennis, D. H. Gould and M. D. Salman (2008). "Characterization of Mycobacterium avium subspecies paratuberculosis disseminated infection in dairy cattle and its association with antemortem test results." Vet Microbiol 127(3-4): 300-308.

Bannantine, J. P., J. L. Everman, S. J. Rose, L. Babrak, R. Katani, R. G. Barletta, A. M. Talaat, Y. T. Grohn, Y. F. Chang, V. Kapur and L. E. Bermudez (2014). "Evaluation of eight live attenuated vaccine candidates for protection against challenge with virulent Mycobacterium avium subspecies paratuberculosis in mice." Front Cell Infect Microbiol 4: 88.

Bansal-Mutalik, R. and H. Nikaido (2014). "Mycobacterial outer membrane is a lipid bilayer and the inner membrane is unusually rich in diacyl phosphatidylinositol dimannosides." Proc Natl Acad Sci U S A 111(13): 4958-4963.

Begg, D. J. and R. J. Whittington (2008). "Experimental animal infection models for Johne's disease, an infectious enteropathy caused by Mycobacterium avium subsp. paratuberculosis." Vet J 176(2): 129-145.

Bermudez, L. E., M. Petrofsky and F. Sangari (2004). "Intracellular phenotype of Mycobacterium avium enters macrophages primarily by a macropinocytosis-like mechanism and survives in a compartment that differs from that with extracellular phenotype." Cell Biol Int 28(5): 411-419.

Bermudez, L. E., M. Petrofsky, S. Sommer and R. G. Barletta (2010). "Peyer's patch- deficient mice demonstrate that Mycobacterium avium subsp. paratuberculosis translocates across the mucosal barrier via both M cells and enterocytes but has inefficient dissemination." Infect Immun 78(8): 3570-3577.

96

Bermudez, L. E. and L. S. Young (1994). "Factors affecting invasion of HT-29 and HEp-2 epithelial cells by organisms of the Mycobacterium avium complex." Infect Immun 62(5): 2021-2026.

Buergelt, C. D., C. Hall, K. McEntee and J. R. Duncan (1978). "Pathological evaluation of paratuberculosis in naturally infected cattle." Vet Pathol 15(2): 196-207.

Buza, J. J., Y. Mori, A. M. Bari, H. Hikono, g. Aodon, S. Hirayama, Y. Shu and E. Momotani (2003). "Mycobacterium avium subsp. paratuberculosis infection causes suppression of RANTES, monocyte chemoattractant protein 1, and tumor necrosis factor alpha expression in peripheral blood of experimentally infected cattle." Infect Immun 71(12): 7223-7227.

Cambier, C. J., K. K. Takaki, R. P. Larson, R. E. Hernandez, D. M. Tobin, K. B. Urdahl, C. L. Cosma and L. Ramakrishnan (2014). "Mycobacteria manipulate macrophage recruitment through coordinated use of membrane lipids." Nature 505(7482): 218-222.

Cooney, M. A., J. L. Steele, H. Steinberg and A. M. Talaat (2014). "A murine oral model for Mycobacterium avium subsp. paratuberculosis infection and immunomodulation with Lactobacillus casei ATCC 334." Front Cell Infect Microbiol 4: 11.

David, J., H. W. Barkema, R. Mortier, S. Ghosh, L. Guan le and J. De Buck (2014). "Gene expression profiling and putative biomarkers of calves 3 months after infection with Mycobacterium avium subspecies paratuberculosis." Vet Immunol Immunopathol 160(1-2): 107-117.

Diard, M., V. Garcia, L. Maier, M. N. Remus-Emsermann, R. R. Regoes, M. Ackermann and W. D. Hardt (2013). "Stabilization of cooperative virulence by the expression of an avirulent phenotype." Nature 494(7437): 353-356.

Early, J. and L. E. Bermudez (2011). "Mimicry of the pathogenic mycobacterium vacuole in vitro elicits the bacterial intracellular phenotype, including early-onset macrophage death." Infect Immun 79(6): 2412-2422.

Eckstein, T. M., S. Chandrasekaran, S. Mahapatra, M. R. McNeil, D. Chatterjee, C. D. Rithner, P. W. Ryan, J. T. Belisle and J. M. Inamine (2006). "A major cell wall lipopeptide of Mycobacterium avium subspecies paratuberculosis." J Biol Chem 281(8): 5209-5215.

Edgar, R., M. Domrachev and A. E. Lash (2002). "Gene Expression Omnibus: NCBI gene expression and hybridization array data repository." Nucleic Acids Res 30(1): 207-210.

97

Ehrt, S. and D. Schnappinger (2007). "Mycobacterium tuberculosis virulence: lipids inside and out." Nat Med 13(3): 284-285.

Facciuolo, A., D. F. Kelton and L. M. Mutharia (2013). "Novel secreted antigens of Mycobacterium paratuberculosis as serodiagnostic biomarkers for Johne's disease in cattle." Clin Vaccine Immunol 20(12): 1783-1791.

Fratti, R. A., J. Chua, I. Vergne and V. Deretic (2003). "Mycobacterium tuberculosis glycosylated phosphatidylinositol causes phagosome maturation arrest." Proc Natl Acad Sci U S A 100(9): 5437-5442.

Freund, J. (1956). "The mode of action of immunologic adjuvants." Bibl Tuberc(10): 130-148.

Geisel, R. E., K. Sakamoto, D. G. Russell and E. R. Rhoades (2005). "In vivo activity of released cell wall lipids of Mycobacterium bovis bacillus Calmette-Guerin is due principally to trehalose mycolates." J Immunol 174(8): 5007-5015.

Gillespie, J., L. L. Barton and E. W. Rypka (1986). "Phenotypic changes in mycobacteria grown in oxygen-limited conditions." J Med Microbiol 21(3): 251-255.

Hagblom, P., E. Segal, E. Billyard and M. So (1985). "Intragenic recombination leads to pilus antigenic variation in Neisseria gonorrhoeae." Nature 315(6015): 156-158.

Harriff, M. and L. E. Bermudez (2009). "Environmental amoebae and mycobacterial pathogenesis." Methods Mol Biol 465: 433-442.

Hines, M. E., 2nd, J. R. Stabel, R. W. Sweeney, F. Griffin, A. M. Talaat, D. Bakker, G. Benedictus, W. C. Davis, G. W. de Lisle, I. A. Gardner, R. A. Juste, V. Kapur, A. Koets, J. McNair, G. Pruitt and R. H. Whitlock (2007). "Experimental challenge models for Johne's disease: a review and proposed international guidelines." Vet Microbiol 122(3-4): 197-222.

Hines, M. E., 2nd, S. E. Turnquist, M. R. Ilha, S. Rajeev, A. L. Jones, L. Whittington, J. P. Bannantine, R. G. Barletta, Y. T. Grohn, R. Katani, A. M. Talaat, L. Li and V. Kapur (2014). "Evaluation of novel oral vaccine candidates and validation of a caprine model of Johne's disease." Front Cell Infect Microbiol 4: 26.

Karakousis, P. C., T. Yoshimatsu, G. Lamichhane, S. C. Woolwine, E. L. Nuermberger, J. Grosset and W. R. Bishai (2004). "Dormancy phenotype displayed by extracellular Mycobacterium tuberculosis within artificial granulomas in mice." J Exp Med 200(5): 647-657.

Khan, F. A., Z. I. Chaudhry, M. I. Ali, S. Khan, N. Mumtaz and I. Ahmad (2010). "Detection of Mycobacterium avium subsp. paratuberculosis in tissue samples of cattle and buffaloes." Trop Anim Health Prod 42(4): 633-638. 98

Khare, S., S. D. Lawhon, K. L. Drake, J. E. Nunes, J. F. Figueiredo, C. A. Rossetti, T. Gull, R. E. Everts, H. A. Lewin, C. L. Galindo, H. R. Garner and L. G. Adams (2012). "Systems biology analysis of gene expression during in vivo Mycobacterium avium paratuberculosis enteric colonization reveals role for immune tolerance." PLoS One 7(8): e42127.

Konnai, S., T. Usui, K. Ohashi and M. Onuma (2003). "The rapid quantitative analysis of bovine cytokine genes by real-time RT-PCR." Vet Microbiol 94(4): 283­ 294.

Lee, H., J. R. Stabel and M. E. Kehrli, Jr. (2001). "Cytokine gene expression in ileal tissues of cattle infected with Mycobacterium paratuberculosis." Vet Immunol Immunopathol 82(1-2): 73-85.

Livak, K. J. and T. D. Schmittgen (2001). "Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method." Methods 25(4): 402-408.

McNabe, M., R. Tennant, L. Danelishvili, L. Young and L. E. Bermudez (2011). "Mycobacterium avium ssp. hominissuis biofilm is composed of distinct phenotypes and influenced by the presence of antimicrobials." Clin Microbiol Infect 17(5): 697­ 703.

McNamara, M., S. C. Tzeng, C. Maier, L. Zhang and L. E. Bermudez (2012). "Surface proteome of "Mycobacterium avium subsp. hominissuis" during the early stages of macrophage infection." Infect Immun 80(5): 1868-1880.

Minnikin, D. (1982). Lipids: complex lipids their chemistry, biosythesis and roles. The biology of the mycobacteria. S. J. Ratledge C. London, Academic Press. 1: 95­ 185.

Mitchell, R. M., R. H. Whitlock, S. M. Stehman, A. Benedictus, P. P. Chapagain, Y. T. Grohn and Y. H. Schukken (2008). "Simulation modeling to evaluate the persistence of Mycobacterium avium subsp. paratuberculosis (MAP) on commercial dairy farms in the United States." Prev Vet Med 83(3-4): 360-380.

Ott, S. L., S. J. Wells and B. A. Wagner (1999). "Herd-level economic losses associated with Johne's disease on US dairy operations." Prev Vet Med 40(3-4): 179­ 192.

Patel, D., L. Danelishvili, Y. Yamazaki, M. Alonso, M. L. Paustian, J. P. Bannantine, L. Meunier-Goddik and L. E. Bermudez (2006). "The ability of Mycobacterium avium subsp. paratuberculosis to enter bovine epithelial cells is influenced by preexposure to a hyperosmolar environment and intracellular passage in bovine mammary epithelial cells." Infect Immun 74(5): 2849-2855. 99

Penketh, A., T. Pitt, D. Roberts, M. E. Hodson and J. C. Batten (1983). "The relationship of phenotype changes in Pseudomonas aeruginosa to the clinical condition of patients with cystic fibrosis." Am Rev Respir Dis 127(5): 605-608.

Periasamy, S., B. N. Tripathi and N. Singh (2013). "Mechanisms of Mycobacterium avium subsp. paratuberculosis induced apoptosis and necrosis in bovine macrophages." Vet Microbiol 165(3-4): 392-401.

Rainwater, D. L. and P. E. Kolattukudy (1985). "Fatty acid biosynthesis in Mycobacterium tuberculosis var. bovis Bacillus Calmette-Guerin. Purification and characterization of a novel fatty acid synthase, mycocerosic acid synthase, which elongates n-fatty acyl-CoA with methylmalonyl-CoA." J Biol Chem 260(1): 616-623.

Ryan, G. J., D. R. Hoff, E. R. Driver, M. I. Voskuil, M. Gonzalez-Juarrero, R. J. Basaraba, D. C. Crick, J. S. Spencer and A. J. Lenaerts (2010). "Multiple M. tuberculosis phenotypes in mouse and guinea pig lung tissue revealed by a dual- staining approach." PLoS One 5(6): e11108.

Sartain, M. J., D. L. Dick, C. D. Rithner, D. C. Crick and J. T. Belisle (2011). "Lipidomic analyses of Mycobacterium tuberculosis based on accurate mass measurements and the novel "Mtb LipidDB"." J Lipid Res 52(5): 861-872.

Scandurra, G. M., G. W. de Lisle, S. M. Cavaignac, M. Young, R. P. Kawakami and D. M. Collins (2010). "Assessment of live candidate vaccines for paratuberculosis in animal models and macrophages." Infect Immun 78(3): 1383-1389.

Schneider, C. A., W. S. Rasband and K. W. Eliceiri (2012). "NIH Image to ImageJ: 25 years of image analysis." Nat Methods 9(7): 671-675.

Sigurethardottir, O. G., M. Valheim and C. M. Press (2004). "Establishment of Mycobacterium avium subsp. paratuberculosis infection in the intestine of ruminants." Adv Drug Deliv Rev 56(6): 819-834.

Smythies, L. E., M. Sellers, R. H. Clements, M. Mosteller-Barnum, G. Meng, W. H. Benjamin, J. M. Orenstein and P. D. Smith (2005). "Human intestinal macrophages display profound inflammatory anergy despite avid phagocytic and bacteriocidal activity." J Clin Invest 115(1): 66-75.

Souza, C., W. C. Davis, T. M. Eckstein, S. Sreevatsan and D. J. Weiss (2013). "Mannosylated lipoarabinomannans from Mycobacterium avium subsp. paratuberculosis alters the inflammatory response by bovine macrophages and suppresses killing of Mycobacterium avium subsp. avium organisms." PLoS One 8(9): e75924.

Stabel, J. R. (1998). "Johne's disease: a hidden threat." J Dairy Sci 81(1): 283-288. 100

Stabel, J. R. and S. Robbe-Austerman (2011). "Early immune markers associated with Mycobacterium avium subsp. paratuberculosis infection in a neonatal calf model." Clin Vaccine Immunol 18(3): 393-405.

Subramoni, S., K. Agnoli, L. Eberl, S. Lewenza and P. A. Sokol (2013). "Role of Burkholderia cenocepacia afcE and afcF genes in determining lipid-metabolism­ associated phenotypes." Microbiology 159(Pt 3): 603-614.

Sugawara, I., H. Yamada, H. Kaneko, S. Mizuno, K. Takeda and S. Akira (1999). "Role of interleukin-18 (IL-18) in mycobacterial infection in IL-18-gene-disrupted mice." Infect Immun 67(5): 2585-2589.

Sweeney, R. W., R. H. Whitlock and A. E. Rosenberger (1992). "Mycobacterium paratuberculosis cultured from milk and supramammary lymph nodes of infected asymptomatic cows." J Clin Microbiol 30(1): 166-171.

Tuchscherr, L., E. Medina, M. Hussain, W. Volker, V. Heitmann, S. Niemann, D. Holzinger, J. Roth, R. A. Proctor, K. Becker, G. Peters and B. Loffler (2011). "Staphylococcus aureus phenotype switching: an effective bacterial strategy to escape host immune response and establish a chronic infection." EMBO Mol Med 3(3): 129­ 141.

Weigoldt, M., J. Meens, K. Doll, I. Fritsch, P. Mobius, R. Goethe and G. F. Gerlach (2011). "Differential proteome analysis of Mycobacterium avium subsp. paratuberculosis grown in vitro and isolated from cases of clinical Johne's disease." Microbiology 157(Pt 2): 557-565.

Weiss, D. J., O. A. Evanson, A. Moritz, M. Q. Deng and M. S. Abrahamsen (2002). "Differential responses of bovine macrophages to Mycobacterium avium subsp. paratuberculosis and Mycobacterium avium subsp. avium." Infect Immun 70(10): 5556-5561.

Weiss, D. J., O. A. Evanson and C. D. Souza (2006). "Mucosal immune response in cattle with subclinical Johne's disease." Vet Pathol 43(2): 127-135.

Whittington, R. J., I. B. Marsh, V. Saunders, I. R. Grant, R. Juste, I. A. Sevilla, E. J. Manning and R. H. Whitlock (2011). "Culture phenotypes of genomically and geographically diverse Mycobacterium avium subsp. paratuberculosis isolates from different hosts." J Clin Microbiol 49(5): 1822-1830.

Wren, B. and N. Dorrell (2002). Functional microbial genomics. Amsterdam ; Boston, Academic Press.

101

Chapter 4

Mycobacterium avium subspecies paratuberculosis induces change in the metabolism of phagocytic host Acanthamoeba castellanii during infection

Jamie L. Everman1,2 , Erin F. Flannery2 , Eugenio U. Mannucci3 , and Luiz E. Bermudez1, 2, 3

1 Department of Microbiology, College of Science, Oregon State University, Corvallis, OR 2 Department of Biomedical Sciences, College of Veterinary Medicine, Oregon State University, Corvallis, OR 3 College of Veterinary Medicine, Oregon State University, Corvallis, OR

Infection and Immunity – Manuscript in Preparation

102

Abstract

Mycobacterium avium subspecies paratuberculosis (MAP) is the causative agent of Johne’s disease, an infection characterized by chronic and often fatal enteritis that affects large populations of ruminants worldwide. As an intracellular pathogen, MAP thrives within both mammalian and environmental phagocytes, and global host studies have begun to unravel particular facets of the MAP-host interaction during infection. This study aims to analyze the metabolic response of host phagocytes during MAP infection and how that response affects the intracellular survival of MAP within the host. We employ the environmental protozoan host Acanthamoeba castellanii (amoeba) as a model phagocytic host and show that there is a distinct pattern of intracellular MAP burden on the host over the course of infection. By utilizing the redox dye alamarBlue, we are able to measure the effect that MAP has on the metabolism of the amoeba during the course of infection. This pattern of intracellular bacterial survival is temporally matched to the level of metabolic activity within the phagocyte, with greater activation leading to a decrease in MAP and slower metabolism resulting in the growth and expansion of the intracellular bacterial population. A constructed MAP library was screened to identify mutants which cause a shift in the metabolism of the amoeba that deviates from that of the wild-type MAP pattern. Phenotype array plates were used to identify the profiles of carbon source usage by amoeba infected with wild-type MAP and selected MAP transposon mutants. Phenotype array analysis shows that each strain results in a specific pattern of carbon utilization by infected amoeba. While a similar core set of carbon sources are used globally, there is an individual specific sub-set of carbon sources that are particular to wild-type or mutant MAP infection including carbon substrates involved in glycolysis, nucleotide precursors, and TCA cycle intermediates and inhibitors. By analyzing the phenotype array data we begin to understand the bacterial mechanisms used to drive the differential use of carbon and the stimulation of phagocyte metabolism during infection.

103

Introduction Johne’s disease is a chronic intestinal inflammatory disease that affects ruminants worldwide. The prevalence of the disease, estimated to affect approximately 70% of herds in the dairy industry, has a devastating economic effect upwards of $1.5 billion annually (Ott, Wells et al. 1999). The disease is caused by the bacterial pathogen Mycobacterium avium subspecies paratuberculosis (MAP) which can be transmitted to young animals in utero, or via milk and fecal- contamination of food and water. Similar to other mycobacterial species, MAP causes infection by preventing phagosome-lysosme fusion after uptake by macrophages and replicating within the phagocytic vacuole (Sturgill-Koszycki, Schlesinger et al. 1994, Via, Deretic et al. 1997, Stabel 1998, Chacon, Bermudez et al. 2004). By preventing host phagocyte tactics normally employed to kill intracellular pathogens, MAP thrives within the vacuole, leads to the formation of granulomatous lesions within the intestinal tissue, and disseminated infection ultimately results in severe wasting and death of the animal. Mycobacteria are well described to hijack or inhibit intracellular host mechanisms for their survival and replication with the intracellular phagocytic vacuole (Sturgill-Koszycki, Schlesinger et al. 1994, Via, Deretic et al. 1997, Fratti, Chua et al. 2003, Vergne, Chua et al. 2004, Wong, Bach et al. 2011). Intracellular bacteria residing within the phagocytic vacuole rely on the components found within the host for nutrient acquisition over the course of infection. As this supply can be limited within the intracellular vacuole, pathogens have evolved mechanisms to obtain the required nutrients during infection. Our study uses the amoeba Acanthamoeba castellanii as a model phagocytic host to explore the interactions between MAP and the host during infection, specifically focusing on the impact the bacterial infection has on the metabolic activity of the host over the course of infection. We demonstrate here that MAP infection induces a dramatic stimulation of the metabolic activity of amoeba during infection. This increased activity is MAP-dependent and is indirectly correlated with the intracellular bacterial burden within the amoeba vacuoles over a 15 day infection. Using carbon source phenotype array analysis, we identify that though global 104

metabolism is increased, wild-type MAP and MAP mutants drive the host to utilize a different pattern of carbon sources during the first 24 hours of infection.

105

Methods and Materials

Amoeba culture. Acanthamoeba castellanii Neff (amoeba; ATCC 30234) were obtained from the American Type Culture Collection (ATCC; Manassas, VA, USA). Axenic cultures were cultured in tissue-culture treated flasks in the dark at 25°C in complete 712 Peptone-Yeast-Glucose (PYG) media (10 g proteose peptone, 0.5 g yeast extract, 4 mM magnesium sulfate heptahydrate, 400 µM calcium chloride, 3.4 mM calcium chloride, 50 µM ferric ammonium sulfate hexahydrate, 2.5 mM disodium phosphate heptahydrate, 2.5 mM potassium phosphate, 0.1 M glucose, pH 6.5). For experiments, amoeba were washed off surface of flask in fresh media and seeded to appropriate density as per experimental protocol.

Mammalian cell culture. The SV40 mutagenized bovine macrophage cell line (BOMAC) was obtained from Dr. Judith Stabel at the National Animal Disease Center (USDA-ARS). Cells were cultivated in RPMI-1640 (Corning) supplemented with 10% heat-inactivated fetal bovine serum at 37°C in 5% CO2 (Stabel and Stabel 1995).

Bacterial culture. Mycobacterium avium subspecies paratuberculosis strain K10 (ATCC BAA-968) was cultured at 37°C on 7H10 agar (BD; Franklin Lakes, NJ) supplemented with casein hydrolysate (1 g/L; BD), 10% (vol/vol) oleic acid, albumin, dextrose, and catalase (OADC; Hardy Diagnostics; Santa Maria, CA), and ferric mycobactin J (2 mg/L; Allied Monitor, Fayette, MO) for 3-4 weeks prior to experiments. For experimental suspensions, bacteria were suspended in HBSS, passaged through a 23-gauge needle and clumps allowed to settle for 10 minutes. The top 50% of the suspension was used for infection (Patel, Danelishvili et al. 2006). For heat-killed samples, MAP suspensions were incubated as indicated for either 70°C or 97°C for 30 minutes prior to use in appropriate assays.

Construction of MAP transposon library. Electrocompetent MAP was prepared by washing cultures 4 times in buffer of 10% glycerol and 0.1% tween-80, followed by 2 106

washes in 10% glycerol. Cells were electroporated with 6 μg of the plasmid pTNGJC containing the heat-inducible transposome (McAdam, Weisbrod et al. 1995), and recovered for 24 hours in stationary culture of 5ml Middlebrook 7H9 broth, OADC (10% v/v), and mycobactin J (2 mg/ml) at 30°C. Cultures were pelleted at 2000 × g and cultured in 7H9 broth supplemented with OADC (10% v/v), mycobactin J (2 mg/ml) and kanamycin sulfate (400 μg/ml) for 8 weeks at 30°C. An aliquot from each culture was screened for the presence of the kanamycin gene using PCR cycle of 95°C for 5 min., 95°C for 30 sec., 60°C for 30 sec., 68°C for 1 min, for 35 cycles using the following primers: 16_pTNGJC_Kan_Fwd 5’TAATGTCGGGCAATCAGGTG 3’ and 17_pTNGJC_Kan_Rev 5’TGTTCAACAGGCCAGCCA 3’. Positive cultures were transferred to 39°C for 1 week to initiate the heat-inducible transposase activity for transfer of transposon element from the plasmid to genome. Bacterial cultures were then cultured on 7H10 plates supplemented with kanamycin sulfate as described above at 37°C for 6 weeks. Individual colonies were selected and independently cultured in 96-well plates to create the library of mutants (Alonso-Hearn, Patel et al. 2008).

Bacterial Infection of Amoeba. Amoeba seeded onto glass chamber slides were infected with MAP at an MOI of 10:1 in 712 PYG amoeba media and incubated for 1 hour at room 25°C. Slides were washed 3 times with PBS to remove extracellular bacteria, heat fixed, and stained using Kinyoun acid fast protocol to identify intracellular mycobacteria (Kinyoun 1915). Slides were visualized on a DM4000B Leica microscope under bright field and fluorescent conditions. Images were captured and analyzed using QCapture Pro7 software (QImaging; Surrey, BC, Canada).

Amoeba metabolic activity assay. 96-well tissue-culture plates were seeded with amoeba overnight in 712 PYG. Cells were infected with the following samples in 712 PYG: media only, MAP (MOI 100:1, 50:1, 10:1, 1:1), heat-killed MAP (incubated for 1 hour at 70°C), or 0.2 μm latex beads (Life Technologies; Carlsbad, CA). Infections were synchronized by centrifugation at 225 × g at 25°C for 5 107

minutes and incubated for 2 hours at 25°C in the dark, at which time wells were washed 2 times with Page’s Amoeba Saline. Fresh 712 PYG media supplemented with 10% (v/v) AlamarBlue (Life Technologies) was added to each well. Fluorescent readings of each plate were measured at 530nm/590nm (excitation/emission) every hour for 24 hours using a Tecan F200 plate reader (Tecan Group Ltd, Männedorf, Switzerland). For long-term experiments amoeba were infected as described above, synchronized, washed with Page’s amoeba saline, and fresh 712 PYG media was added to each well. At time 0, and every 12 hours until 180 hours post-infection, alamarBlue (10% v/v) was added to a new set of infected wells and fluorescent readings were taken at 0, 2, 12, and 24 hours after the addition of dye. For controls and normalization between plates when necessary, 712 PYG media, live MAP, heat- killed MAP, and latex beads were inoculated into wells in the absence of amoeba and treated in identical manner dependant on the experiments described above.

Intracellular MAP survival during amoeba infection. Amoeba were seeded and cultured in 24-well plate overnight at 25°C or 37°C. Wells were infected with MAP at an MOI of 10:1, infection synchronized at 225 × g for 5 minutes, and incubated at 25°C. At 0, 1, 2, 3, 5, 7, 10, and 15 days post-infection supernatant from wells was removed and gently washed once to remove any extracellular bacteria. Supernatant samples were centrifuged at 225 × g for 5 minutes to pellet amoeba, and both pellet and resulted supernatant were collected for colony counts. Both plate-adheared amoeba and non-adheared amoeba from supernatant were lysed by addition of 0.5% sodium dodecyl sulfate in water, passed through a 28-gauge insulin needle, and pelleted at 2000 × g for 10 minutes. Pellets were suspended and plated to quantify intracellular bacteria recovered from amoeba samples. Separate wells were infected as described above, suspended in Page’s amoeba saline, stained 1:10 with Trypan Blue (0.4% solution), and quantified for cell viability over the same time course.

MAP transposon library screen. 96-well plates containing individual MAP mutants in each well were prepared as described above and OD600 was measured to normalize for infection level. 96-well plates were seeded with amoeba and infected with MAP 108

transposon library clones at an MOI of 10:1. Plates were synchronized at 225 × g for 5 minutes and incubated for 2 hours at 25°C. Wells were gently washed twice with Page’s amoeba saline and fresh 712 PYG media containing AlamarBlue (10% v/v; Life Technologies) was added to each well. Fluorescent readings were taken every hour for 24 hours on a Tecan F200 plate reader. Each plate contained internal controls of media only and wild-type MAP infection both in the presence and absence of amoeba to normalize all library plates screened. In total ~3000 individual clones were analyzed and the clones with the greatest change of metabolic stimulation from the wild-type infection were selected for further analysis.

BOMAC uptake and survival assays. Each strain of MAP selected from the transposon library screening assays was cultured on 7H11 agar as described above. BOMAC cells were seeded overnight and infected with either wild-type MAP or each selected mutant at an MOI of 10:1. Infections were incubated for 2 hours, washed with PBS, and treated with amikacin (200 µg/ml) in RPMI for 2 hours at 37°C. For uptake assays, wells were washed twice with PBS and cells were lysed with 0.1% triton X-100 in water for 15 minutes. For survival assays, wells were washed twice with PBS and fresh media was added to each well and incubated for 24 or 96 hours prior to lysis. Lysates were collected and plated for quantification of viable intracellular bacteria.

Metabolic phenotype array. Amoeba were seeded into T-25 flasks overnight in modified 712 PYG media (mPYG; 2.5 g proteose peptone, 0.5 g yeast extract, 4 mM magnesium sulfate heptahydrate, 400 µM calcium chloride, 50 µM ferric ammonium sulfate hexahydrate, 2.5 mM disodium phosphate heptahydrate, 2.5 mM potassium phosphate, pH 6.5) to eliminate an excess of carbon sources in the media and infected with wild-type MAP or MAP transposon mutant strains at an MOI of 10:1 in mPYG. Infections were incubated for 2 hours at 25°C and amoeba were collected by low speed centrifugation at 50 × g for 5 minutes. Collected amoeba were quantified and each well of a PM-M1 MicroPlate – Carbon and Energy Source array panel (Biolog; Hayward, CA) was seeded with 5 × 104 infected amoeba in mPYG with a final 109

concentration of 10% AlamarBlue (Life Technologies). Of the 3 empty control wells in each plate used to control and normalized conditions between plates, 1 well was seeded with only mPYG/10% AlamarBlue, 1 well was seeded with 105 uninfected amoeba in mPYG/10% AlamarBlue, and for mutant infections 1 well was seeded with 105 wild-type MAP-infected amoeba in mPYG/10% AlamarBlue. Plates were wrapped in parafilm to prevent evaporation, and fluorescence values were measured at excitation/emission 530nm/590nm over a 48 hours infection using a Tecan F200 plate reader (Tecan Group Ltd, Männedorf, Switzerland). For long-term phenotype assays, amoeba were infected with MAP or PBS alone as described above for 2 hours, media removed, and fresh mPYG media added prior to incubation in the dark at 25°C. Amoeba were collected at 3 days and 6 days post-infection and phenotype assay was seeded and conducted as described above.

110

Results

MAP is readily taken up by and replicate within Acanthamoeba castellanii In order to utilize Acanthamoeba castellanii (amoeba) as a proper phagocytic host to model the metabolic interactions that occur during MAP infection, parameters of infection were first set to understand the dynamics during MAP infection of amoeba. Using Kinyoun staining, intracellular MAP are identified as acid-fast positive pink bacilli (data not shown), or as bright red bacilli under fluorescence (Figure 4.1). It is evident that amoeba are able to readily take up bacteria after only 1 hour of infection, as amoeba can be seen with an abundance of MAP within the intracellular vacuoles or in the process of uptake via phagocytosis (Figure 4.1) (Cirillo, Falkow et al. 1997, Harriff and Bermudez 2009). To understand the long-term outcome of MAP infection on amoeba and to determine the viability of intracellular MAP over the course of infection, cultures were infected and both extra- and intracellular bacterial viability was monitored over a 15 day course of infection (Figure 4.2a). In accordance with our observations with microscopy in Figure 1, the amount of MAP quantified after antibiotic treatment to remove extracellular bacteria is nearly 100% of the initial infection inoculum, indicating amoeba readily take up the majority of MAP upon infection. Within 24 hours of infection the amoeba eliminate almost 50% of the initially ingested bacteria and are able to maintain that observed reduction in intracellular bacterial burden until 3 days post-infection. After 3 days of infection the intracellular population begins to replicate and increases in number, by day 5 almost reaching the original amount of intracellular bacteria that was ingested upon initial infection. At 7, 10, and 15 days post-infection, the number of amoeba begin to exceed the space constraints of the plate and detatch into the media where they remain viable and continue to be infected with bacteria. In addition to detachment, extracellular MAP is observed in the infection media beginning at 15 days post-infection (Figure 4.2; gray bar). While the amoeba cultures are able to combat intracellular survival from 1 to 3 days post- infection, and then appear to serve as a host for increased replication from 5 to 10 days post-infection, the bacterial infection does not appear to have a detrimental 111

effect on the general growth or viability of the amoeba over the 15 day infection (Figure 4.2b). Over the complete time course of 15 days, the MAP-infected amoeba show no significant difference in the growth rate or in the amount of cells present in each sample compared to the growth rate or viability of uninfected control cultures measured over the same time course. These data indicate that Acanthamoeba castellanii have the ability to readily take up the majority of bacteria during experimental MAP infection, and we establish the temporal pattern of the intracellular and extracellular MAP viability, and viability of the amoeba hosts, over the course of a 15 day infection.

MAP stimulates the metabolism of amoeba during infection The most dramatic shift in the phagocyte-pathogen interaction between the amoeba and MAP during infection occurs during the first 24 hours of infection as amoeba are able to eliminate almost half of the original intracellular bacteria ingested during initial infection (Figure 4.2a). To determine the impact that MAP infection had on the overall metabolism of the amoeba, a metabolic activity assay was conducted using the dye AlamarBlue as an indicator of oxidative respiration and metabolic activity within the amoeba. The measured fluorescence values indicate that MAP infection has a significant and dramatic stimulation of the metabolic activity of the amoeba over the first 24 hours of infection (Figure 4.3). The stimulation observed was dose-dependent as an MOI of 10:1 results in a much more dramatic metabolic stimulation compared to the stimulation that results from infection at an MOI of 1:1 (data not shown), though both stimulation levels are significantly higher than uninfected amoeba. Since rapidly growing bacteria can stimulate a shift in AlamarBlue fluorescence (Pesch 1929), it was investigated whether the growth and metabolic activity of the bacteria is a contributing factor in the dramatic shift observed during amoeba infection. In the absence of amoeba, MAP alone results in minimal fluorescence over the time of the assay (Figure 4.3; red dotted line). To determine whether the bacterial-induced stimulation of amoeba metabolism was an active process requiring live MAP, or simply stimulated by the cellular components of MAP alone, amoeba were infected with heat-killed bacteria for an equivalent 112

duration of time. Neither the heat-killed bacteria alone, nor the heat-killed MAP- infected amoeba result in any stimulation of phagocyte metabolic activity (Figure 4.3; black lines). Finally, Acanthamoeba castellanii has been described to release an oxidative burst upon phagocytosis of inert beads (Davies 1991). To confirm that the metabolic stimulation observed in this assay was due specifically to the ingestion of the pathogen, and not a bystander effect which occurs during the process of phagocytosis itself, assays were conducted with the addition of 0.2 µm latex beads. Upon uptake of latex beads, there is negligible stimulation of metabolic activity of the amoeba recorded over the course of the assay (Figure 4.3; blue lines). In total, these data indicate that the metabolism of amoeba is highly stimulated by the ingestion of MAP during infection in a dose-, viability-, and particle-dependent manner. The metabolism of MAP infected amoeba is dramatically stimulated in response to MAP infection with in the first 24 hours of infection (Figure 4.3), coinciding with the decrease in the intracellular burden of bacteria over the same time frame (Figure 4.2a). However, as infection progresses, MAP begin to replicate and grow within the intracellular environment leading to the eventual escape into the extracellular environment. To determine the state of amoeba metabolism during the later stages of MAP infection, alamarBlue metabolism assays were conducted at several timepoints over an 8 day infection, following and recording the level of metabolism within treated amoeba (Figure 4.4). Consistent with the data shown in Figure 3, within the first 24 hours the metabolic activity of MAP-infected amoeba is dramatically increased (Figure 4.4a). Between 1 and 3 days post-infection, the metabolism of infected amoeba remains higher than uninfected amoeba cultures, though with a much lower change in MAP-induced activity, compared to the initial infection (Figure 4.4b – 4.4e). This data paired with no observed increase of intracellular MAP over the same time (Figure 4.2a) suggest that the higher metabolic levels of the host are able to keep the bacteria from replicating, or that the bacteria haven’t yet adjusted to their intracellular environment in the early stages of infection. As intracellular MAP begin to replicate (Figure 4.2a), the metabolic activity of the amoeba once again increases to a more activated level compared to previous time points leading up to the 5 day time point post-infection (Figure 4.4g). However, as 113

the bacteria begin replicating and establish a successful intracellular niche, the bacterial population returns to the original level of infection and leads to escape from the amoeba between 7 and 10 days post-infection. At this point the stimulation of metabolic activity no longer is increased compared to uninfected amoeba (Figure 4.4h – 4.4j). These data indicate that the shift in the metabolism of infected amoeba is correlated with the intracellular bacterial burden of MAP over the course of infection within the amoeba. Stimulation of metabolism appears to occur paired with a decrease in bacterial viability, while a slower, less dramatic metabolic stimulation occurs at the same time that the bacteria are replicating and increasing in number within the phagocyte.

Identification of metabolic activity mutants and validation of amoeba model As infection with MAP is capable of inducing a dramatic shift in the metabolism of the infected host phagocyte, we hypothesized that one or more virulence components of the bacterium are responsible for the metabolic stimulation. To determine which bacterial components are responsible, a transposon library of MAP was constructed and the library was screened using the alamarBlue metabolic assay to identify mutants that result in the highest change in metabolic stimulation of amoeba during 24 hours infection (Figure 4.5). Each plate contained wild-type MAP infected amoeba which served as an internal control and comparison for amoeba metabolic stimulation in order to account for any uncontrollable variables in the protocol including changes in room temperature, alteration of timepoints collected over the 24 hour infection, and any variability between array plates. As we identified that the multiplicity of infection of MAP during infection can affect the stimulation levels of amoeba (data not shown) the bacterial density of each culture was measured and infection level adjusted to ensure the multiplicity of infection was not responsible for increased or decreased metabolic activity within the amoeba. Of the ~50 plates that were screened, the highest and the lowest metabolic curves were identified and the MAP mutant strain used for infection of those respective wells was re-cultured. Each selected clone was re-analyzed twice for the metabolic activity of amoeba upon infection and the mutant pool of interest was narrowed to 30 mutants by eliminating 114

those strains that did not produce consistent results over the 3 independently completed experiments (Figure 4.5 and verification data not shown). To determine whether the identification of mutants using the metabolic stimulation assay within amoeba was a valid identification technique of mutant bacterial strains, we analyzed the virulence of each strain using mammalian host cells. Using bovine macrophages (BOMAC), each selected mutant was assayed for its level of uptake by bovine macrophages cells (BOMACs) and its ability to survive after 24­ and 96-hours post-infection (Figure 4.6). With the exception of only a few mutants, the uptake of each strain compared to the wild-type MAP is relatively consistent and not significantly different from one another or from the wild-type MAP uptake (Figure 4.6a). To determine if the strains had a different outcome during intracellular survival, survival assays were conducted and intracellular MAP were quantified and normalized to their respective uptake values to account for the differences in initial bacteria present within the cells. Notably, at 24-hours post infection the majority of the strains that more actively stimulate the metabolic activity of the amoeba exhibit equivalent or increased intracellular survival compared to the wild-type MAP infection (Figure 4.6b – blue). Alternatively, each of the mutants that are deficient in their ability to stimulate amoeba metabolism were found to have a diminished survival rate within the BOMAC cultures after 24 hours of infection (Figure 4.6b – green). When analyzed for intracellular survival within BOMACs after 96 hours of infection, it appears that in total the strains are able to resume growth, replicate within the cell, and maintain a steady rate of infection over the long-term infection. Here, we show that the mutants identified within the metabolic assay model using amoeba as the host phagocyte are also mutants in bovine macrophage cell culture after 24­ hours, thus verifying the validity of our findings using the amoeba model.

MAP mutant growth dynamics within Acanthamoeba castellanii Once we validated that the mutants identified using the metabolic stimulation assays in amoeba are also mutants in BOMAC infection, we narrowed the pool to the strains which resulted in the largest change, either stimulatory or inhibitory, to amoeba metabolism for further characterization. To compare the intracellular growth 115

dynamics of each mutant to the wild-type strain of MAP, amoeba were infected with each strain and intracellular viability was assessed at time 0, and 1 and 5 days post- infection (Figure 4.7). After incubation of each bacterial strain with amoeba, there is no significant difference in the level of bacteria that are taken up by the phagocytic hosts over a 2 hour time period (Figure 4.7a). To assess the intracellular survival of each strain after 1 day and 5 days of infection, infected cultures were lysed and intracellular bacteria quantified (Figure 4.7b). Independent of their effect on amoeba metabolic activity, 4 of the mutants do not exhibit a decrease in the number of intracellular bacteria after 1 day of infection as seen with the wild-type strain, while the other 4 mutants do show a decrease in intracellular viability after 1 day of infection (Figure 4.7b). Interestingly, while wild-type infection results in intracellular growth and an increase in quantified bacteria after 5 days of infection, each mutant strain is not only unable to maintain the original level of ingested bacteria, but each strain is also eliminated much more effectively from the host amoeba after a 5 day infection. These data indicate that each mutant strain of MAP has its own pattern of growth and unique interaction with amoeba during the course of infection and that these patterns are not directly correlated to the ability to stimulate or inhibit the metabolism of the amoeba host.

Carbon source metabolism during MAP infection Pathogens have been described to both manipulate and detoxify host products for the acquisition of nutrients including amino acids and carbon compounds. This usage of acquired nutrients could be a method in which MAP acquires the nutrients it requires to successfully replicate within the intracellular vacuole. To more accurately understand the metabolic shift that is occurring during MAP infection of amoeba, and to characterize the changes in those patterns derived from MAP-mutant infected amoeba, a phenotype array was conducted using a carbon source plate containing a different carbon source in each well of a 96-well plate, to analyze the utilization and growth of amoeba on individual carbon sources during infection. Prior to experimental assays, we optimized the 712 PYG amoeba medium for the phenotype assay as the standard, carbon-rich medium would not require use of the compound 116

found in each well as a carbon source during infection. We adjusted the components of the modified amoeba growth medium (mPYG) so that a minimal amount of carbon was present in the broth to maintain health of the amoeba and sustain basal levels of growth, but so that upon addition of an external carbon source that was readily metabolized, growth could rapidly occur and be measured. Using the alamarBlue dye as a read-out, we were able to measure the growth of wild-type MAP infected and uninfected amoeba from time 0 – 24 hours, 72 – 96 hours, and 144 – 168 hours after initial infection (Data Analysis in Progress). Utilizing the same technique to understand how the bacterial infection is shaping and altering the metabolic use of carbon components during infection, we assayed uninfected amoeba, wild-type infected amoeba, and amoeba infected with each of the 8 mutant strains selected by the experiments conducted shown in Figure 4.5 – Figure 4.8. The overall carbon usage patterns of amoeba infected with each strain of MAP are similar and by analyzing the differences we can begin to characterize the mutant strains and their interaction with the amoeba host (Figure 4.8; Supplemental Table 4.1). After 24 hours wild-type MAP infected amoeba are readily able to metabolize and grow in the presence of simple sugars and compounds such as glucose, fructose, glycerol, and mono-methyl succinate. Meanwhile, substrates such as adenosine, previously reported to be toxic to cultured cells at 50 µM (Snyder, Hershfield et al. 1978), and acetoacetic acid are toxic to amoeba as neither uninfected no any of the MAP-infected cultures are able to grow in the presence of those compounds.

Data Analysis and Experiments in Progress:  Identify specific carbon sources that are differentially used during wild-type MAP infection compared to mutant infection  Analyze sequenced genomes for mutations and determine which gene(s) are interrupted in the MAP mutants to more readily understand the cause of the mutant phenotype and its effect on the amoeba metabolism during infection

117

Discussion MAP is a successful pathogen capable of surviving the intracellular compartment upon ingestion by phagocytic cells and manipulating the host and its immune system to evade detection and killing. While the intracellular environment is the limiting factor to which nutrients the bacterium has access to during infection, the pathogen may have some influence on what those nutrients are and how they are utilized by the host cell. One mechanism of cellular manipulation includes the effect that the bacterial infection has on the metabolic state of the host phagocyte. While infection can trigger an innate response meant to rid the host of infection, that bacterial driven response may also be responsible for enhancing the survival of the pathogen by increasing its access to compounds or nutritional building blocks that aid it a survival advantage within the phagocytic vacuole. The studies of such mechanisms are difficult to conduct as unraveling the host-pathogen metabolic interaction is difficult, and the interpretation of such data can be complex to sort through. The use of the correct host cell, variables being measured and chosen timepoints for analysis are imperative in understanding the metabolic dynamics between the bacterium and the host during infection. Multiple cell types can serve as a host to MAP and mycobacterial infection ranging from mammalian macrophages to environmental amoeba. Interestingly, the bacterial response to intracellular infection within these different phagocyte hosts are similar as infection of Acanthamoeba castellanii with Mycobacterium avium results in a similar pattern of upregulated genes as that seen during macrophage infection (Tenant and Bermudez 2006). The closely related Mycobacterium avium subspecies hominissuis (MAH) can infect environmental amoeba and possess a more virulent phenotype upon isolation from Acanthamoeba castellanii (Cirillo, Falkow et al. 1997) as it also does from intracellular infection within mammalian macrophages (Bermudez, Petrofsky et al. 2004, Early and Bermudez 2011). Environmental amoeba have been shown to host MAP and serve as a potential reservoir for the long­ term survival of the bacterium in the soil of pastures at infected farms (White, Birtles et al. 2010), while conferring enhanced resistance to common disinfectant treatments involving chlorine and antibiotics (Miltner and Bermudez 2000, Whan, Grant et al. 118

2006). The similarities between amoeba and macrophages, including mycobacterial interactions and comparable characteristics and traits obtained by MAP from the infection of both types of phagocytes suggest that amoeba may serve as a significant model to further investigate host-pathogen interactions and the metabolic impact of MAP infection on the host. Bacterial metabolic adaptations are common and allow for the use of host products that would otherwise be toxic. Salmonella Typhimurium infection of the gut elicits a strong inflammatory response characterized by oxidative stress mechanisms that lead to the production of the sulfur compound tetrathionate. Toxic or unusable for most pathogens, Salmonella contains the gene TtrABC which enables the bacterium to utilize the compound as an electron acceptor, resulting in enhanced growth and dissemination of infection (Hensel, Hinsley et al. 1999). Shigella is capable of using host cell waste products, particularly pyruvate, as an energy source effectively causing no metabolic activation within the host (Kentner, Martano et al. 2014). Likewise, Mycobacterium tuberculosis has the ability to break down host lipids and cholesterol, detoxify the breakdown components, and utilize the resulting molecules for the acquisition of carbon during intracellular infection (Lee, VanderVen et al. 2013). Alternatively, bacteria can use mechanisms that allow for the manipulation of host machinery and result in the increase of nutrients that are then available to the intracellular pathogen. These mechanisms, termed ‘nutritional virulence’ by Abu Kwaik et al, are required for the successful proliferation within the intracellular environment which has a restricted level of nutrients (Abu Kwaik and Bumann 2013). Multiple pathogens utilize nutritional virulence as a survival strategy. Chlamydial inclusions acquire nutrients from rerouted and fragmented intracellular vesicles (Christian, Heymann et al. 2011, Elwell, Jiang et al. 2011), while Legionella pnemophila can trigger proteosome degradation of host proteins in both mammalian cells and environmental amoeba for an increased pool of amino acids required for nutrients and virulence (Price, Al-Quadan et al. 2011, Al-Quadan, Price et al. 2012). While MAP exhibits adaptation to infection within the host as shown by its altered proteome during in vivo infection (Weigoldt, Meens et al. 2011) and its shifting transcriptome within the intestine (Lamont, Xu et al. 2013), it is 119

unknown how MAP infection affects the metabolism of the host phagocyte during infection and if the effect is beneficial towards the outcome of the infection. Here, we identify that the ability of MAP infection to stimulate the metabolism of the host and the subsequently survive within the phagocytic vacuole. This survival, however, is adversely affected during points of high metabolic activity within the host, while a slowing of the cellular metabolic state results in the outgrowth and enhanced survival of the bacteria. To determine if MAP was influencing the metabolism for its benefit of intracellular nutrients during infection, we conducted phenotype array assays for particular carbon sources and identified metabolic effect mutants to further understand the role of the metabolic stimulation seen within wild-type MAP infection within amoeba. As our original alamarBlue assay screen was for mutants within the first 24 hours of infection, it is important to note that each of the mutants analyzed are also the 24 hour infection of BOMAC cells as well. In total, these assays provide validation to the findings that the attenuated mutants identified in our amoeba screen are, in fact, mutants within bovine macrophage cells as indicated by their increased killing after 24 hours of infection, compared to the hyper-stimulatory mutants which were more able to survive within the phagosomal compartments of the macrophages during early infection. Early analysis of our phenotype arrays indicate that while the global pattern of carbon usage is similar between amoeba infected with wild-type and mutant MAP strains, there are distinct differences among the use of particular sources. Simple sugars such as glucose, fructose and mono-methyl succinate are the most successfully used compounds resulting in the highest growth and metabolic activity of the amoeba during infection. The metabolism of infected amoeba are significantly higher than that of uninfected amoeba cultures over the same period of infection. Interestingly, similar phenotype arrays conducted using the bacterial carbon source plates to study the metabolic use of compounds by Mycobacterium tuberculosis (Mtb) indicated that Mtb was able to replicate and grow more successfully on the simple compounds of glucose and mono-methyl succinate than most other nutrient sources found on the bacterial nutrient source plate (Khatri, Fielder et al. 2013). 120

In total, these data indicate that MAP has a dramatic impact and effect on the metabolic activity of the host phagocyte during infection. The bacterial-driven effect is temporal and changes during the course of infection, seemingly affecting the growth and intracellular survival of MAP during infection. The phenotype array analysis begins to uncover the nuances used by MAP to acquire nutrients while in the intracellular environment and determine if MAP manipulates the host cell to provide nutrients that would not otherwise be present during different stages of intracellular infection. Ultimately, we aim to utilize this data to determine what effect bacterial infection has on the metabolic interaction between host and pathogen and to determine the bacterial drivers of those mechanisms using analysis of the mutant strain of MAP. It would be interesting to determine if the nutritional depletion or supplementation of the intestinal environment during MAP infection would have an effect on the viability and intracellular survival of MAP or would result in an effect on the development of disease within the animal over time.

121

Acknowledgements We would like to thank Brendan Jeffrey for technical assistance with bioinformatic graph development and analysis of MAP mutant genomes. We would also like to thank Dr. Barry Bochner at Biolog for technical assistance and discussion on experimental design,

122

Figure 4.1: Amoeba readily phagocytize MAP during infection Amoeba were infected with MAP at an MOI of 10:1 and incubated for 1 hour at 25°C. Amoeba were washed to remove any extracellular MAP, heat-fixed, and stained using Kinyoun acid-fast stain method (Kinyoun 1915). Carbol-fuchsin stained MAP was visualized as red bacilli under fluorescent microscopy. Images taken at 100x magnification and representative of 100 amoeba visualized within each experimental and control sample.

123

A Intracellular - Adhered 1 .5  1 0 6 Intracellular - Detached

Extracellular 1 .0  1 0 6

5 .0  1 0 5 CFU Recovered

0 0 1 2 3 5 7 1 0 1 5

Time (Day)

B

2 .5  1 0 6 1 0 5 abii (%) ) s t o d ( ) % ( y ilit b ia V 2 .0  1 0 6 1 0 0

1 .5  1 0 6 9 5

1 .0  1 0 6 9 0

5 .0  1 0 5 8 5 Am oeba (cells/m l) (bars) 0 8 0 0 1 2 3 5 7 1 0 1 5

Times (days)

Amoeba Viability (uninfected) Amoeba Count (uninfected)

Amoeba Viability (infected) Amoeba Count (infected)

Figure 4.2: Temporal pattern of MAP viability during amoeba infection Amoeba were infected with MAP at an MOI of 10:1, incubated for 1 hour, followed by a 2 hour treatment with amikacin (200 µg/ml) to kill extracellular MAP and replaced with 712 PYG medium. At indicated times the supernatant was collected to quantify extracellular MAP. Both detached and attached amoeba in each well were lysed and intracellular MAP was quantified (a). The viability of MAP-infected and uninfected control wells was determined using trypan blue staining and the amount of amoeba in each well was quantified at each time point over the course of the experiment (b). Data represent the mean ± SD of wells assayed in triplicate and is representative of 2 independent experiments.

124

20000

Amoeba + Media 15000 **** Amoeba + MAP Amoeba + Heat-Killed MAP Amoeba + 0.2 m Beads 10000 Media Only MAP Only Fluorescence 5000 Heat-Killed MAP Only 0.2 m Beads Only

0

(Excitation/Emission 530 nm/590nm) 0 2 4 6 8 10 12 14 16 18 20 22 24 Time (hours)

Figure 4.3: Stimulation of amoeba metabolism by MAP infection Wells seeded with amoeba and wells without amoeba were infected with live MAP (MOI 10:1), heat-killed MAP, 0.2 µm latex beads, or PYG medium alone. Plates were incubated for 1 hour at 25°C, washed to remove extracellular bacteria, debris, and beads, fresh PYG medium containing 10% alamarBlue was added to each well, and fluorescence was measured at 530 nm/590 nm (excitation/emission) every hour for 24 hours. Data represent the mean ± SD of at least 2 independent experiments each conducted in triplicate (****p-value <0.0001 as determined by Student’s t-test).

125

A B C D E 2 5 0 0 0

2 0 0 0 0

1 5 0 0 0 RFU 1 0 0 0 0

5 0 0 0

0 0 5 1 0 1 5 2 0 2 5 10 15 20 25 30 35 40 25 30 35 40 45 50 35 40 45 50 55 60 50 55 60 65 70 75

F G H I J 2 5 0 0 0

2 0 0 0 0 Relative Fluorescent Units 1 5 0 0 0 RFU 1 0 0 0 0

5 0 0 0

0 70 75 80 85 90 95 100 95 100 105 110 115 120 120 125 130 135 140 145 145 150 155 160 165 170 170 175 180 185

Time (hours)

Amoeba + PYG Medium Amoeba + MAP Infection Amoeba + Heat-Killed MAP

PYG Medium Only M A P O n ly Heat-Killed M AP

Figure 4.4: Impact of MAP infection on metabolism of amoeba during long­ term infection Amoeba were infected with live MAP, heat-killed MAP, or PYG medium alone for 2 hours, washed, and fresh PYG medium added to each well. At 0 hours (a), 12 hours (b), 24 hours (c), 36 hours (d), 48 hours (e), 72 hours (f), 96 hours (g), 120 hours (h), 144 hours (i), and 168 hours (j), alamarBlue was added to each set of wells and fluorescent readings were measured at 530nm/590nm (excitation/emission) and recorded at 0, 2, 12, and 24 hours after the addition of dye. Wells infected with live MAP, heat-killed MAP, or latex beads in the absence of amoeba were included to control for the amoeba-independent reduction of the dye over the course of the study. Data represent the mean ± SEM of 3 independent experiments each performed in triplicate.

126

4 0 0 0 0 W ild-type M AP M AP mutant - high 3 0 0 0 0 M AP mutant - low

2 0 0 0 0 RFU

1 0 0 0 0

0 0 5 1 0 1 5 2 0 2 5 Time (hours)

Figure 4.5: Identification of MAP mutant strains exhibiting altered metabolic stimulation of amoeba A MAP-transposon library was screened for metabolic activity stimulation of amoeba over a 24 hour infection. Each plate contained amoeba infected with mutants or wild- type MAP and wells containing uninfected amoeba. Each mutant-induced metabolic curve was compared to the wild-type MAP infected amoeba curve within each plate. The data here represent the MAP mutants which reproducibly demonstrated the highest (blue) and lowest (green) metabolic curves compared to the wild-type MAP infected amoeba. The wild-type MAP infected amoeba data (shown in red) represent the average of all the wild-type MAP infected amoeba curves from all ~50 plates that were screened, and are included for illustrative purposes to show the average wild- type induced metabolic curve compared to the selected MAP mutants.

127

A 1 2 BOMAC Uptake

1 0

8

6

% U p ta k e 4

2

0

6 4 3 2 6 2 9 5 6 3 4 9 G 2 G 6 1 0 1 1 G 8 D 6 D 6 1 0 E 1 1 F 1 1 4 G 5 D 7 6 D 9 T y p e 4 E 1 8 3 4 3 4 F 84 9 F 5 2 F 4 1 C 1 1 G 2 9 G 3 8 C 5 5 C 41 5 5 1 5 1 6 G 3 6 F 85 0 G 4 1 5 0 G E 1 01 0 G 8 F 1 02 5 F 25 d 1 1 G 1 2 2 2 F 4 2 G 4 2 il W B 2 0 0 BOM AC Survival (24 hours post infection)

1 5 0

1 0 0 % S u r v iv a l 5 0

0

6 2 8 4 3 2 6 9 4 8 5 8 3 0 2 G 6 1 1 1 1 D 6 G 6 1 0 G 4 1 0 G 9 y p e G 1 0 4 G 5 D 7 6 D 9 4 E 1 8 G 3 4 3 4 F 4 9 F 5 2 F 4 1 C 1 1 G E 1 1 2 9 G 23 8 C 5 5 C 1 5 G 5 F 1 5 1 6 3 6 F 5 0 G 4 1 D 6 5 0 1 0 8 F 1 2 5 F 5 d T 1 1 1 2 2 2 F 4 2 G 4 2 E il W BOM AC Survival (96 hours post infection) C 3 0 0

2 0 0

1 0 0 % S u r v iv a l

0

6 4 2 6 2 9 6 5 7 6 2 9 G 2 G 6 F 8 F 3 C 4 G 8 1 1 G 6 1 0 G 4 G 9 1 0 y p e G 1 0 F 1 1 4 G 5 D E 1 0 5 F 6 D 4 E 1 8 3 4 3 4 4 9 F 5 2 4 1 C 1 1 G E 1 1 2 9 G 3 8 C 5 5 1 5 5 F 1 5 D 1 6 3 6 F 85 0 G 34 1 D 5 0 1 0 8 F 2 5 d T 1 1 1 2 2 2 4 2 G 4 2 il W Bacterial Strain/M utant

Figure 4.6: Verification of MAP mutants using bovine macrophage infection To confirm the validity of the amoeba metabolism screen and verify MAP clones are mutants in bovine macrophages (BOMAC), macrophage uptake (a), 24 hour survival (b), and 96 hour survival of wild-type MAP and each mutant clone was assessed by a 2 hour infection followed by amikacin treatment to eliminate extracellular bacteria. Cells were lysed and intracellular bacteria were quantified at time 0- (a), 24- (b) or 96-hours (c) post-infection. Survival assay values (b and c) are normalized to the original amount of bacteria ingested in the 2 hour uptake assay for each mutant strain. Data represent the mean ± SD of 3 independent experiments each completed in duplicate. 128

D ay 0 A B 1 5 0 5 0 D ay 1 D ay 5 4 0 1 0 0

3 0

* * 2 0 * % S u r v iv a l

% U p ta k e 5 0 *

1 0

0 0

6 8 4 e 6 G 2 D 6 G 6 G 6 F 8 D 6 G 3 4 G 5 4 G 5 1 1 G 2 9 3 4 F 4 9 F 1 5 1 6 5 0 G 3 1 1 2 9 G 2 3 4 4 9 F 4 1 5 1 6 G 5 0 -T y p e ild -T y p W W ild Bacterial Strain/M utant Bacterial Strain/M utant

Figure 4.7: MAP mutant strain growth dynamics within Acanthamoeba castellanii To compare the growth dynamics of each selected mutant strain with the wild-type MAP dynamics over a 5 day course of infection in amoeba, intracellular viability assays were conducted and quantified to determine the intracellular level of bacteria at initial uptake and infection (Time 0), and after 1 day and 5 days of infection. Black bars – wild-type MAP; Blue bars – MAP mutant strains with increased amoeba metabolic stimulation; Green bars – MAP mutant strains with decreased amoeba metabolic stimulation. Figure B values were normalized to time 0 uptake values from figure A to control for the initial amount of bacteria present in each mutant-infected amoeba population. Data represent mean ± SD of 2 independent experiments each performed in triplicate (* p<0.05 as determined by a Student’s t test).

129

Well Carbon Source Well Carbon Source A1 Empty Control E1 Melibionic Acid A2 Empty Control E2 D-Melibiose A3 Empty Control E3 D-Galactose A4 α-Cyclodextrin E4 α-Methyl-D- Galactoside A5 Dextrin E5 β-Methyl-D- Galactoside A6 Glycogen E6 N-Acetyl- Neuraminic Acid A7 Maltitol E7 Pectin A8 Maltotriose E8 Sedoheptulosan A9 D-Maltose E9 Thymidine A10 D-Trehalose E10 Uridine A11 D-Cellobiose E11 Adenosine A12 β-Gentiobiose E12 Inosine B1 D-Glucose-6- Phosphate F1 Adonitol B2 α-D-Glucose-1- Phosphate F2 L- Arabinose B3 L-Glucose F3 D-Arabinose B4 α-D-Glucose F4 β-Methyl-D- Xylopyranoside B5 α-D-Glucose F5 Xylitol B6 α-D-Glucose F6 Myo-Inositol B7 3-O-Methyl-D- Glucose F7 Meso-Erythritol B8 α-Methyl-D- Glucoside F8 Propylene glycol B9 β-Methyl-D- Glucoside F9 Ethanolamine B10 D-Salicin F10 D,L- α-Glycerol- Phosphate B11 D-Sorbitol F11 Glycerol B12 N-Acetyl-D- Glucosamine F12 Citric Acid C1 D-Glucosaminic Acid G1 Tricarballylic Acid C2 D-Glucuronic Acid G2 D,L-Lactic Acid C3 Chondroitin-6- Sulfate G3 Methyl D-lactate C4 Mannan G4 Methyl pyruvate C5 D-Mannose G5 Pyruvic Acid C6 α-Methyl-D- Mannoside G6 α-Keto-Glutaric Acid C7 D-Mannitol G7 Succinamic Acid C8 N-Acetyl-β-D- Mannosamine G8 Succinic Acid C9 D-Melezitose G9 Mono-Methyl Succinate C10 Sucrose G10 L-Malic Acid C11 Palatinose G11 D-Malic Acid C12 D-Turanose G12 Meso-Tartaric Acid D1 D-Tagatose H1 Acetoacetic Acid (a) D2 L-Sorbose H2 γ-Amino-N- Butyric Acid D3 L-Rhamnose H3 α-Keto-Buytric Acid D4 L-Fucose H4 α-Hydroxy- Butyric Acid D5 D-Fucose H5 D,L-β-Hydroxy- Butyric Acid D6 D-Fructose-6- Phosphate H6 γ-Hydroxy- Butyric Acid D7 D-Fructose H7 Butyric Acid D8 Stachyose H8 2,3-Butanediol D9 D-Raffinose H9 3-Hydroxy-2- Butanone D10 D-Lactitol H10 Propionic Acid D11 Lactulose H11 Acetic Acid D12 α-D-Lactose H12 Hexanoic Acid

Supplemental Table 4.1: Biolog carbon source plate layout of PM-M1 plate 130

WT 1 0 0 0 0 U n in fe c te d

5 0 0 0

0

-5 0 0 0

-1 0 0 0 0 1 0 0 0 0 WT 4 G 5

5 0 0 0

0

-5 0 0 0

-1 0 0 0 0 1 0 0 0 0 WT 5 0 G 3 5 0 0 0

0

-5 0 0 0

-1 0 0 0 0 WT 1 0 0 0 0 1 5 D 6

5 0 0 0

0

-5 0 0 0

-1 0 0 0 0 WT 1 0 0 0 0 1 6 G 6

5 0 0 0

0

-5 0 0 0

-1 0 0 0 0 WT 1 0 0 0 0 4 9 G 4

5 0 0 0 Normalized Relative Fluorescent Units (RFU)

0

-5 0 0 0

-1 0 0 0 0 1 0 0 0 0 WT

2 9 G 2 5 0 0 0

0

-5 0 0 0

-1 0 0 0 0

1 0 0 0 0 WT 1 1 G 6 5 0 0 0

0

-5 0 0 0

-1 0 0 0 0 1 0 0 0 0 WT

3 4 F 8 5 0 0 0

0

-5 0 0 0

-1 0 0 0 0

Phenotype Array Carbon Sources (A1 - G12) 131

Figure 4.8: Carbon source metabolic utilization phenotypes of MAP and MAP mutant infected amoeba Amoeba were seeded into culture flasks and incubated overnight in mPYG prior to infection. Cells were infected at MOI 10:1 with wild-type MAP or each MAP mutant in mPYG as indicated. Infections were incubated for 2 hours at 25°C, amoeba were collected and washed once in PAS and 1x104 amoeba were seeded into each well of a PM-M1 phenotype array plate in mPYG with 10% alamarBlue. Control wells were seeded with uninfected amoeba and wild-type infected amoeba to control for plate variation between experiments. Fluorescent readings were measured at 530nm/590nm (excitation/emission) and 24 hour data is shown. Data represent the mean ± SEM of 2 independent experiments.

132

References

Abu Kwaik, Y. and D. Bumann (2013). "Microbial quest for food in vivo: 'nutritional virulence' as an emerging paradigm." Cell Microbiol 15(6): 882-890.

Al-Quadan, T., C. T. Price and Y. Abu Kwaik (2012). "Exploitation of evolutionarily conserved amoeba and mammalian processes by Legionella." Trends Microbiol 20(6): 299-306.

Alonso-Hearn, M., D. Patel, L. Danelishvili, L. Meunier-Goddik and L. E. Bermudez (2008). "The Mycobacterium avium subsp. paratuberculosis MAP3464 gene encodes an oxidoreductase involved in invasion of bovine epithelial cells through the activation of host cell Cdc42." Infect Immun 76(1): 170-178.

Bermudez, L. E., M. Petrofsky and F. Sangari (2004). "Intracellular phenotype of Mycobacterium avium enters macrophages primarily by a macropinocytosis-like mechanism and survives in a compartment that differs from that with extracellular phenotype." Cell Biol Int 28(5): 411-419.

Chacon, O., L. E. Bermudez and R. G. Barletta (2004). "Johne's disease, inflammatory bowel disease, and Mycobacterium paratuberculosis." Annu Rev Microbiol 58: 329-363.

Christian, J. G., J. Heymann, S. A. Paschen, J. Vier, L. Schauenburg, J. Rupp, T. F. Meyer, G. Hacker and D. Heuer (2011). "Targeting of a chlamydial protease impedes intracellular bacterial growth." PLoS Pathog 7(9): e1002283.

Cirillo, J. D., S. Falkow, L. S. Tompkins and L. E. Bermudez (1997). "Interaction of Mycobacterium avium with environmental amoebae enhances virulence." Infect Immun 65(9): 3759-3767.

Davies, B. C., L.S.; Edwards, S.W. (1991). "Superoxide generation during phagocytosis by Acanthamoeba castellanii: similarities to the respiratory burst of immune phagocytes." J Gen Microbiol 137: 705-710.

Early, J. and L. E. Bermudez (2011). "Mimicry of the pathogenic mycobacterium vacuole in vitro elicits the bacterial intracellular phenotype, including early-onset macrophage death." Infect Immun 79(6): 2412-2422.

Elwell, C. A., S. Jiang, J. H. Kim, A. Lee, T. Wittmann, K. Hanada, P. Melancon and J. N. Engel (2011). "Chlamydia trachomatis co-opts GBF1 and CERT to acquire host sphingomyelin for distinct roles during intracellular development." PLoS Pathog 7(9): e1002198.

Fratti, R. A., J. Chua, I. Vergne and V. Deretic (2003). "Mycobacterium tuberculosis glycosylated phosphatidylinositol causes phagosome maturation arrest." Proc Natl Acad Sci U S A 100(9): 5437-5442. 133

Harriff, M. and L. E. Bermudez (2009). "Environmental amoebae and mycobacterial pathogenesis." Methods Mol Biol 465: 433-442.

Kentner, D., G. Martano, M. Callon, P. Chiquet, M. Brodmann, O. Burton, A. Wahlander, P. Nanni, N. Delmotte, J. Grossmann, J. Limenitakis, R. Schlapbach, P. Kiefer, J. A. Vorholt, S. Hiller and D. Bumann (2014). "Shigella reroutes host cell central metabolism to obtain high-flux nutrient supply for vigorous intracellular growth." Proc Natl Acad Sci U S A 111(27): 9929-9934.

Kinyoun, J. J. (1915). "A Note on Uhlenhuths Method for Sputum Examination, for Tubercle Bacilli." Am J Public Health (N Y) 5(9): 867-870.

Lamont, E. A., W. W. Xu and S. Sreevatsan (2013). "Host-Mycobacterium avium subsp. paratuberculosis interactome reveals a novel iron assimilation mechanism linked to nitric oxide stress during early infection." BMC Genomics 14: 694.

Lee, W., B. C. VanderVen, R. J. Fahey and D. G. Russell (2013). "Intracellular Mycobacterium tuberculosis exploits host-derived fatty acids to limit metabolic stress." J Biol Chem 288(10): 6788-6800.

McAdam, R. A., T. R. Weisbrod, J. Martin, J. D. Scuderi, A. M. Brown, J. D. Cirillo, B. R. Bloom and W. R. Jacobs, Jr. (1995). "In vivo growth characteristics of leucine and methionine auxotrophic mutants of Mycobacterium bovis BCG generated by transposon mutagenesis." Infect Immun 63(3): 1004-1012.

Miltner, E. C. and L. E. Bermudez (2000). "Mycobacterium avium grown in Acanthamoeba castellanii is protected from the effects of antimicrobials." Antimicrob Agents Chemother 44(7): 1990-1994.

Ott, S. L., S. J. Wells and B. A. Wagner (1999). "Herd-level economic losses associated with Johne's disease on US dairy operations." Prev Vet Med 40(3-4): 179­ 192.

Patel, D., L. Danelishvili, Y. Yamazaki, M. Alonso, M. L. Paustian, J. P. Bannantine, L. Meunier-Goddik and L. E. Bermudez (2006). "The ability of Mycobacterium avium subsp. paratuberculosis to enter bovine epithelial cells is influenced by preexposure to a hyperosmolar environment and intracellular passage in bovine mammary epithelial cells." Infect Immun 74(5): 2849-2855.

Pesch, K. L. S., U. (1929). Combined assays for lactose and galactose by enzymatic reactions. Dietary Sugars: Chemistry, Analysis, Function and effects. V. R. Preedy, Royal Society of Chemistry: 395-404.

Price, C. T., T. Al-Quadan, M. Santic, I. Rosenshine and Y. Abu Kwaik (2011). "Host proteasomal degradation generates amino acids essential for intracellular bacterial growth." Science 334(6062): 1553-1557. 134

Snyder, F. F., M. S. Hershfield and J. E. Seegmiller (1978). "Cytotoxic and metabolic effects of adenosine and adenine on human lymphoblasts." Cancer Res 38(8): 2357­ 2362.

Stabel, J. R. (1998). "Johne's disease: a hidden threat." J Dairy Sci 81(1): 283-288.

Stabel, J. R. and T. J. Stabel (1995). "Immortalization and characterization of bovine peritoneal macrophages transfected with SV40 plasmid DNA." Vet Immunol Immunopathol 45(3-4): 211-220.

Sturgill-Koszycki, S., P. H. Schlesinger, P. Chakraborty, P. L. Haddix, H. L. Collins, A. K. Fok, R. D. Allen, S. L. Gluck, J. Heuser and D. G. Russell (1994). "Lack of acidification in Mycobacterium phagosomes produced by exclusion of the vesicular proton-ATPase." Science 263(5147): 678-681.

Tenant, R. and L. E. Bermudez (2006). "Mycobacterium avium genes upregulated upon infection of Acanthamoeba castellanii demonstrate a common response to the intracellular environment." Curr Microbiol 52(2): 128-133.

Vergne, I., J. Chua, S. B. Singh and V. Deretic (2004). "Cell biology of mycobacterium tuberculosis phagosome." Annu Rev Cell Dev Biol 20: 367-394.

Via, L. E., D. Deretic, R. J. Ulmer, N. S. Hibler, L. A. Huber and V. Deretic (1997). "Arrest of mycobacterial phagosome maturation is caused by a block in vesicle fusion between stages controlled by rab5 and rab7." J Biol Chem 272(20): 13326-13331.

Weigoldt, M., J. Meens, K. Doll, I. Fritsch, P. Mobius, R. Goethe and G. F. Gerlach (2011). "Differential proteome analysis of Mycobacterium avium subsp. paratuberculosis grown in vitro and isolated from cases of clinical Johne's disease." Microbiology 157(Pt 2): 557-565.

Whan, L., I. R. Grant and M. T. Rowe (2006). "Interaction between Mycobacterium avium subsp. paratuberculosis and environmental protozoa." BMC Microbiol 6: 63.

White, C. I., R. J. Birtles, P. Wigley and P. H. Jones (2010). "Mycobacterium avium subspecies paratuberculosis in free-living amoebae isolated from fields not used for grazing." Vet Rec 166(13): 401-402.

Wong, D., H. Bach, J. Sun, Z. Hmama and Y. Av-Gay (2011). "Mycobacterium tuberculosis protein tyrosine phosphatase (PtpA) excludes host vacuolar-H+-ATPase to inhibit phagosome acidification." Proc Natl Acad Sci U S A 108(48): 19371­ 19376. 135

Chapter 5

MAP1203 involved in binding and invasion of epithelial cells by Mycobacterium avium subspecies paratuberculosis

Jamie L. Everman1,2 , Lucero Garcia Flores1 , and Luiz E. Bermudez 1,2

1 Department of Microbiology, College of Science, Oregon State University, Corvallis, OR 2 Department of Biomedical Sciences, College of Veterinary Medicine, Oregon State University, Corvallis, OR

Infection and Immunity – Manuscript in Preparation

136

Abstract Mycobacterium avium subspecies paratuberculosis (MAP) is the causative agent of Johne’s disease, a chronic and ultimately fatal enteritis that affect ruminant populations worldwide. One of the modes of transmission of the bacterium is via milk as it is ingested by young animals. The exposure to milk has a dramatic impact on MAP, resulting in a more virulent and invasive phenotype. The gene MAP1203 was identified as being upregulated 25-fold after milk exposure and here was identify the role the MAP1203 plays during the initial interaction with and infection of epithelial cells within the bovine host. By using over-expression of the wild-type MAP1203, and dominant-negative mutants including MAP1203 containing a mutated integrin-binding RGD domain, and MAP1203 with a 15 amino acid deletion in the putative signal sequence, we show that MAP1203 plays a role in the binding to and the invasion of bovine MDBK epithelial cells during a 1 hour infection. Furthermore, we provide preliminary evidence indicating that MAP1203 is a surface-exposed protein and that the putative signal sequence is required for the processing and expression of the function protein on the surface of the bacterium. Current studies aim to identify the bovine epithelial host protein(s) that MAP1203 binds to in early stages of infection, and to characterize how the pathogen-host recognition facilitates invasion and subsequent infection.

137

Introduction Mycobacterium avium subspecies paratuberculosis (MAP) is a slow-growing pathogen which causes Johne’s disease in cattle and other ruminants. The bacterium is well described to readily change its phenotype during various stages of disease and upon interaction with a variety of environments. One method of MAP transmission to young animals is via contaminated colostrum and milk early in life. The mammary gland, its epithelial lining, and milk have all been shown to serve as a reservoir for MAP across the stages of the disease (Taylor, Wilks et al. 1981, Sweeney, Whitlock et al. 1992, Stabel and Lambertz 2004). Notably, MAP dramatically alters its phenotype when exposed to the hyperosmolar milk environment, resulting in a more virulent and infectious form of the bacterium (Patel, Danelishvili et al. 2006, Alonso- Hearn, Eckstein et al. 2010). Of interest in this phenotype is the gene MAP1203 which is upregulated nearly 25-fold upon incubation within milk (Patel, Danelishvili et al. 2006). The function of this protein may play an important role in the pathogenesis of MAP and its invasive capabilities within the intestinal mucosa. The mycobacterial operon containing the invasion and intracellular persistence (iipA) gene plays a direct role in the invasion and survival of mycobacteria within the host. Using Mycobacterium marinum which encodes for 2 iipA genes, and complementing gene knockouts with the Mycobacterium tuberculosis (Mtb) homolog Rv1477(RipA) confers increased binding, invasion, and survival of Mycobacterium marinum within macrophage cell lines (Gao, Pak et al. 2006). Other studies have characterized the Mtb protein RipA, containing a conserved NLP60-peptidoglycan hydrolase domain, as important in cell wall remodeling during mycobacterial growth (Chao, Kieser et al. 2013). As MAP1203 contains the same NLP60 peptidase domain, with similar RGD binding and signal sequence domains as the iipA gene in Mycobacterium marinum (Gao, Pak et al. 2006), and is 79% homologous to the amino acid sequence of Mtb Rv1477, we aim to understand the function of MAP1203 during invasion and infection. Utilizing dominant-negative constructs of the RGD integrin binding domain and a deletion mutant of the putative signal sequence of the protein, we 138

describe in this study the role of MAP1203 in the pathogenesis of MAP during the initiation of infection and invasion of bovine MDBK epithelial cells.

139

Methods and Materials

Bacterial culture. Mycobacterium avium subspecies paratuberculosis strain K10 (ATCC BAA-968) was cultured at 37°C on 7H10 agar (BD; Franklin Lakes, NJ) supplemented with casein hydrolysate (1 g/L; BD), 10% (vol/vol) oleic acid, albumin, dextrose, and catalase (OADC; Hardy Diagnostics; Santa Maria, CA), and ferric mycobactin J (2 mg/L; Allied Monitor, Fayette, MO) for 3-4 weeks prior to experiments. Mycobacterium smegmatis (MS) strain mc2155 (ATCC) was cultured on 7H10 agar supplemented with 10% OADC at 37°C. For experimental suspensions, bacteria were suspended in HBSS, passaged through a 23-gauge needle 5 times and clumps allowed to settle for 10 minutes. The top 50% of the suspension was used for infection (Patel, Danelishvili et al. 2006).

Mammalian cell culture. Madin-Darby bovine kidney (MDBK) epithelial cells (CCL-22) and RAW 264.7 macrophage cultures (TIB-71) were obtained from the American Type Culture Collection (ATCC; Manassas, VA). Both cells lines were cultivated in Dulbecco’s Modified Eagle’s Medium (DMEM) supplemented with 10% heat-inactivated fetal bovine serum (FBS; Gemini Bio-Products; West

Sacramento, CA), at 37°C in 5% CO2.

Dominant-Negative MAP1203 Constructs. The 1400bp MAP1203 gene was PCR amplified from genomic DNA from MAP using the primers 18_1203XbaIFwd and 19_1203XbaIRev (Table 5.1) as follows: 95°C for 5 minutes, 35 cycles of 95°C for 1 minute, 60°C for 30 seconds, 68°C for 2 minutes, followed by 68°C for 10 minutes. Amplified PCR product of MAP1203 was used as the DNA template to create the gene mutants as follows using SOE-PCR. The ΔRGD mutant was created by mutating the RGD amino acid sequence beginning at residue 417 in the amino acid protein sequence of MAP1203 to an RCE sequence, and the ΔSignalSequence (ΔSS) mutant was created by eliminating the amino acids residues 16-39 which fall on the N-terminal side of the predicted signal sequence cleavage site (Petersen, Brunak et al. 2011). ΔRGD fragments were PCR amplified using the primer pairs 1203XbaIFwd / 140

MAP1203_RGDmu_B and 1203XbaIRev / MAP1203_RGDmu_C, while the ΔSS fragments were PCR amplified using the primer pairs 1203XbaIFwd / MAP1203_SSmuB and 1203XbaIRev / MAP1203_SSmuC using the program described above. Equimolar amounts of fragments for each mutant were incubated together for 10 minutes at 50°C before PCR amplification as follows: 95°C for 5 minutes, 40 cycles of 95°C for 1 minute, 60°C for 1 minute, 68°C for 2 minutes, followed by 68°C for 10 minutes. PCR fragments were cloned into the XbaI site of the acetamide inducible plasmid pJAM2 (Triccas, Parish et al. 1998). To prepare electropcompetent cells, Mycobacterium smegmatis grown to

OD600 0.500 and plate grown MAP were washed 4 times with 10% glycerol and 0.1% tween-80 at 2000 × g for 15 minutes each and final pellet suspended in 1 ml 10% glycerol. Aliquots of 200 µl of each culture were electroporated in a 0.2 cm cuvette at 2.5 kV, 1000 ohms, and 25 µF. M.smegmatis cultures were recovered in 7H9 broth supplemented with 10% OADC for 3 hours at 37°C with shaking at 200 rpm and plated onto 7H10/OADC supplemented with kanamycin (50 µg/ml). MAP cultures were recovered in 7H9 broth supplemented with 10% OADC and mycobactin J (2 mg/L) for 24 hours at 37°C prior to plating cultures on 7H10/casein hydrolysate/OADC/mycobactin J/ kanamycin (400 µg/ml). Plate-grown MAP colonies were photographed under 40x magnification and the area of each colony was measured using ImageJ software (Schneider, Rasband et al. 2012). To induce the expression of MAP1203, M. smegmatis cultures were grown to an OD600 0.5 and induced by the addition of 0.2% acetamide and 0.2% succinic acid for 6 hours at 37°C at 200 rpm. MAP cultures were grown to an OD600 0.3 and induced with 0.2% acetamide and 0.2% succinic acid for 24 hours at 37°C in stationary culture. Proteins were extracted and analyzed using a rabbit polyclonal α­ MAP1203 antibody (1:300) in 4% nonfat dry milk as primary probe, and detected using a goat anti-rabbit IR800 secondary antibody (1:10,000) (Licor).

Invasion and Survival Assays. MDBK or RAW 264.7 cells were seeded into 48­ well plates and grown to 80% confluence prior to experiments. Cells were infected at an MOI of 10:1 with bacterial cultures complete DMEM and infections were 141

synchronized at 220 × g for 5 minutes prior to incubation for 2 hours at 37°C with 5%

CO2. Each well was washed 3 times with HBSS and treated with DMEM supplemented with amikacin (200μg/ml) for 2 hours to kill extracellular bacteria and washed 3 times with HBSS and lysed (Bermudez and Young 1994). For survival assays, complete DMEM was added after antibiotic treatment and plates were incubated for 14 hours prior to lysis. To quantify bacterial uptake, cells were lysed with 0.1% triton X-100 in deionized water, samples collected, serially diluted, and plated for CFU determination.

Purification of MAP1203. MAP1203 was PCR amplified from pJAM MAP1203 containing mutant constructs described above using the primers MAP1203_pETHindIII_Fwd and MAP1203_pETEcoRI_Rev (Table 5.1) using the amplification program described above. Fragments were cloned into the HindIII/EcoI restriction sites of the pET-6x HN-N-terminal expression vector. E. coli BL21 DE3 (LifeTechnologies; Carlsbad, CA) were transformed with the appropriate constructs and immediately cultured to OD600 0.3 in 250 ml LB broth supplemented with ampicillin (100 µg/ml) at 37°C at 200 rpm. Cultures were induced by the addition of 1 mM IPTG (Sigma-Aldrich; St. Louis, MO) for 3 hours prior to collection of the bacterial pellet for protein analysis. Bacteria were lysed using the xTractor Buffer (Clontech Laboratories, Inc; Mountain View, CA) as per manufacturer’s instructions, separated on an SDS-PAGE gel, probed for 6xHN-tagged MAP1203 expression using a goat anti-6xHN primary antibody (1:2000; Clontech Laboratories, Inc), and visualized with a rabbit anti-goat IR800 secondary antibody (1:10,000) (Licor). MAP1203 protein was purified by incubating protein extracts at 4°C with magnetic anti-6xHN beads, captured, and washed as per manufacturer’s instructions (Clontech Laboratories, Inc.).

142

Results

Analysis of MAP1203 domains and construction of dominant-negative constructs To determine the role of MAP1203 during MAP infection and its interaction with bovine epithelial cells, we first analyzed if any domains of interest were found within the hypothetical MAP1203 gene sequence. Previous findings by Gao et al described the effects of an invasion and intracellular persistence gene (iipA) in Mycobacterium marinum and its interaction with macrophage cells. To characterize the function of MAP_1203 in its entirety, as well as understand the function of such gene domains, we created dominant-negative vectors for over-expression assays in vitro (Figure 5.1). The RGD domain is a well described integrin binding domain used by viruses and bacteria to bind to and initiate the invasion process of host cells. We mutated the RGD domain found at residues 417­ 419 in the amino acid sequence of MAP_1203 to an RCE sequence, shown to abrogate the binding effect to host cell integrin molecules. Analysis of putative signal sequences or cleavage sites using SignalP 4.0 (Petersen, Brunak et al. 2011) indicated that a predicted signal sequence was present at the N-terminal portion of the protein with a cleavage site between residue 39/40. To interrupt the function of the signal sequence we deleted amino acid residues 16-39, resulting in a truncated and interrupted signal sequence and cleavage recognition site in the protein. The gene sequence of the wild-type, ΔRGD, and ΔSS gene fragments were verified and cloned into the acetamide-inducible plasmid pJAM2 for further analysis of function. Protein expression was confirmed for all assays done in M. smegmatis (Figure 5.2) and MAP (Figure 5.5) prior to analysis of bacterial quantification from each assay.

Function of MAP_1203 in macrophage and epithelial cell infection To understand the function of MAP1203 during mycobacterial infection we utilized MDBK epithelial cells and RAW 264.7 macrophages to determine if the overexpression of the wild-type protein or of each of the mutants had an effect on the binding or invasion of bacteria during infection. We used Mycobacterium smegmatis 143

(M. smegmatis) as a surrogate for the slow-growing MAP in our initial infections as it is an avirulent, fast-growing mycobacterial species which is highly amenable to transformation and manipulation. The species does not contain any gene or protein sequences homologous to MAP1203, though provides a similar avirulent mycobacterial background to provide insight into the function of proteins of interest. M. smegmatis containing the empty pJAM2 vector, or with the vector containing any of the MAP1203 constructs did not alter the level of bacterial uptake by macrophages after 1 hour of infection (Figure 5.3a). However, after 14 hours of incubation, the intracellular survival rate of M. smegmatis expressing the wild-type MAP1203 and MAP1203 ΔRGD protein was 2-fold higher than that of bacterial populations expressing the MAP1203 ΔSS construct or the empty pJAM2 vector (Figure 5.3b). These data suggest that MAP1203 plays a unique role in encouraging intracellular survival within macrophages. The RGD domain does not serve as a ligand for binding to the integrin receptors on the macrophage membrane, and data indicates that the signal sequence is required for the production or complete function of MAP1203 during macrophage infection. To assess the role of MAP1203 over-expressed in M. smegmatis we grew and induced cultures as described above and infected bovine MDBK cells for 2 hours prior to quantification of intracellular bacteria. Over-expression of MAP1203 significantly increases the invasion capability of the naturally non-invasive M. smegmatis (Figure 5.4). The RGD domain does not affect the invasion ability of the bacterium, as integrins are not present on the surface of MDBK epithelial cells. Lastly, the signal sequence deletion mutant appears to be unable to confer the invasion advantage to M. smegmatis indicating that it plays an important role in the function of MAP1203.

Role of MAP1203 during MAP infection of epithelial cells Using the surrogate M. smegmatis can offer relatively fast information into the function of a protein, but does not provide the complete genetic and physiologic make-up as that of the MAP bacterium from which it was identified. Thus, we transformed each of our constructs into MAP for the characterization of its function 144

during over-expression in its natural bacterial host. Wild-type MAP1203 was successfully expressed after 24 hours of induction within MAP cultures (Figure 5.5). To analyze the function of MAP1203 during epithelial cell infection, MAP cultures were incubated with MDBK epithelial cells for 2 hours to at 4°C to assess binding capabilities, and at 37°C to investigate the invasion rate of each culture (Figure 5.6). MAP cultures over-expressing the complete MAP1203 protein are able to bind to the surface of epithelial cells nearly 3 times more that control cultures with the empty pJAM vector (Figure 5.6a). Additionally, cultures expressing the pJAM::MAP1203 construct are able to invade MDBK epithelial cells at a significantly higher level than the empty vector containing cultures (Figure 5.6b). These data clearly indicate that MAP1203 plays a significant role in the initiation of infection between MAP and bovine epithelial cells as it is able to enhance both the binding and invasion ability of MAP during early infection.

Observational and experimental evidence that MAP1203 may be a surface associated protein The iipA gene has been described as an important invasion protein within mycobacterial species. Likewise, the Mtb Rv1477 protein is a predicted membrane surface associated protein. A protein used for virulence and binding to host cells would almost always need to be on the cell surface in order to come in contact with the host cell receptors or surface proteins to initiate and facilitate invasion and uptake during the early stages of infection. MAP1203 has not been definitively proven to be a surface associated protein via proteomic analysis of the MAP surface proteome (McNamara, unpublished data). However, as it is highly upregulated upon exposure to the milk environment, it is possible that previous proteomic analysis was not conducted on a relevant phenotype that would express MAP1203 for capture and identification. Through a series of experimental observations, we suggest that MAP1203 is a surface associated protein. From experimental observations, we note that the over-expression of MAP1203 is toxic to a variety of cells and it is described that cell membrane proteins, when over-expressed, often result in bacteria toxicity or lysis upon induction by the 145

resulting alteration of the protective bacterial membrane. Upon induction of MAP1203 expression in M. smegmatis at time 0 at the time of inoculation, rather than at OD600 0.5, we conclude that the over-expression of MAP1203 is lethal to M. smegmatis, resulting in no growth of the bacteria and no increase in the optical density of the cultures over a 5 day time period. From cloning in the pET vector for protein purification, we observe that induction and expression of wild-type and ΔRGD MAP1203 in E. coli is toxic to cells, while the signal sequence mutant is not. Even in the E. coli BL21 pLysS, meant to silence background expression of toxic proteins prior to induction, the bacteria do not grow well and optical density decreases by half after a 3 hours induction of protein, indicating protein toxicity resulting in bacterial cell lysis. Lastly, we have observed that the growth of MAP cultures containing either the pJAM vector alone, or the pJAM vector containing the inserts of MAP1203, MAP1203ΔRGD, or MAP1203ΔSS results in different colony size over the same period of incubation, suggesting that MAP1203 may have an effect on the growth rate or ability of MAP to grow in culture (Figure 5.7).

146

Discussion

The phenotype that MAP develops upon exposure to milk results in increased virulence and infectivity of MAP. The gene MAP1203, upregulated almost 28-fold in this phenotype, serves as a potentially important virulence and invasion protein for MAP and its ability to invade the intestinal epithelium upon infection of the ruminant host. Our study here aimed to characterize the function of MAP1203 and it functional domains, and show that it is a putative surface associated protein that is important for both the binding and invasion of bovine epithelial cells. The RGD binding domain is important for the binding of a variety of pathogens to integrin receptors on the surface of host cells. While, important for macrophages and other host cell types, integrin receptors are not readily found on the membrane of intestinal epithelial cells, and this we hypothesized the RGD motif of MAP1203 would not play an important role in the functionality of MAP1203. Alternatively, cell surface associated proteins require a signal sequence to be successfully processed and exported to the surface membrane. We hypothesized that the disruption of the predicted signal sequence and cleavage site on the MAP1203 protein would render the protein non-functional in its ability to cause an effect on the invasion of host cells. Our findings indicate that the RGD does not have an altered effect on the invasion of Mycobacterium smegmatis overexpressing MAP1203 containing a mutation in the RGD amino acid sequence compared to that of the overexpression of the intact wild-type protein. However, the deletion of the signal sequence of MAP1203 abrogated the effect of the protein and resulted in the same level of invasion as the empty vector control. Overexpressed in the natural host MAP, MAP1203 is involved in both binding to and invasion of epithelial cells. Current and future work will aim to identify the binding receptor of MAP1203 within the host epithelial cell. Characterization of this interaction and its effect on the bacterium and the host cell will allow us to further understand the role of MAP1203, its dramatic upregulation upon milk exposure, and its role in MAP virulence during infection. 147

Acknowledgements

We would like to thank Dr. John Bannantine at the Agricultural Research Service with the USDA for the gift of anti-MAP1203 rabbit serum for antibody detection of the MAP1203 protein in our assays.

148

Primer Name Sequence (5’ to 3’) 1203XbaIFwd gtg tct aga atgagacgca cacgctgg 1203XbaIRev gtg tctaga ccactcgatgtaacggacc MAP1203_RGDmu_B gat gac ttc gca gcg gcg cat MAP1203_RGDmu_C atg cgc cgc tgc gaa gtc atc MAP1203_SSmuB ggcgacgggatccctcgcggccggccgcgcggaag MAP1203_SSmuC ccg gcc gcg agg gat ccc gtc gcc gac aac ctg g MAP1203_pETHindIII_Fwd gtg aagctt a aga atgagacgca cacgctgg MAP1203_pETEcoRI_Rev gtg gaattc ccactcgatgtaacggacc

Table 5.1: Primers used for the creation of MAP1203 expression vectors

149

Figure 5.1: Construction of dominant-negative mutants of MAP1203 An RGD mutant and a Signal Sequence deletion was constructed to identify the role of the RGD domain and also the putative signal sequence of MAP1203 during the interaction with bovine epithelial cells. ClustalW alignment of wild-type (WT) MAP1203, RGD to RCE mutation (RGD), and the signal sequence deletion (SS) are shown. 150

6 6

4 5

3 1

0.2% acetamide + - + - + - +

p J A M p J A M :: p J A M :: p J A M :: M A P 1 2 0 3 M A P 1 2 0 3 M A P 1 2 0 3  RGD  SS

Figure 5.2: Induction of MAP1203 protein expression in M. smegmatis M. smegmatis cultures containing pJAM2, pJAM::MAP1203, pJAM::MAP1203

ΔRGD, or pJAM::MAP1203 ΔSS were grown to OD600 0.6 and protein expression induced by the addition of 0.2% acetamide and 0.2% succinic acid for 6 hours. Uninduced cultures were grown in a similar manner in the presence of 0.2% succinic acid only. 50 µg of protein extract was separated on an SDS-PAGE gel and identification of 52 kDa MAP1203 protein was completed by Western blot analysis using a rabbit anti-MAP1203 polyclonal antibody (1:300) and visualized using a goat anti-rabbit IR800 (1:10,000) secondary antibody. Blot is representative of the expression of constructs in all M. smegmatis cellular assays completed.

151

A B

2 0 0 8

** 1 5 0 6 *

1 0 0 4 % U p ta k e % S u r v iv a l 5 0 2

0 0 + pJAM Inducible Vector + pJAM Inducible Vector

S 3 S S MS GD 3  p ty S m p ty R MS GD  E P 1 2 0  R 3 E m P 1 2 0  1 2 0 3 MA P 1 2 0 3 MA P 1 2 0 AP MA P 1 2 0 3 M MA MA Sam ple Strain/Vector Sam ple Strain/V ector

Figure 5.3: Role of MAP1203 in M. smegmatis uptake and survival within macrophages M. smegmatis containing either the empty pJAM2 vector or the plasmid over- expressing the wild-type MAP1203, MAP1203 ΔRGD, or MAP1203 ΔSS were used to investigate the ability of the bacteria to invade RAW 264.7 macrophages after 2 hour infection (a) or survive within the host cell after 14 hours of intracellular infection (b). Data represent the mean ± SEM of 3 independent experiments each conducted in triplicate (*p <0.05, **p<0.01 as determined by Student’s t-test).

152

0 .2 5 *

0 .2 0 *

0 .1 5

0 .1 0 Invasion (% ) 0 .0 5

0 .0 0 + pJAM Inducible Vector

S S MS AM 2 0 3 GD R p J

M A P 1 + M

p J A Sample Strain/Vector

Figure 5.4: Role of MAP1203 in epithelial cell invasion by M. smegmatis M. smegmatis containing either the empty pJAM2 vector or the plasmid over- expressing the wild-type MAP1203, MAP1203 ΔRGD, or MAP1203 ΔSS were used to investigate the ability of the bacteria to invade bovine MDBK epithelial cells after a 2 hour infection. Data represent the mean ± SEM of 3 independent experiments each conducted in triplicate (*p<0.05 as determined by Student’s t-test).

153

k D a 6 6

4 5

0.2% Acetamide - + - +

p J A M p J A M :: M A P 1 2 0 3

Figure 5.5: Induction of mAP1203 protein expression in MAP cultures MAP cultures testing positive for the pJAM plasmid containing the MAP1203 gene were grown to OD600 0.3 and induced using 0.2% acetamide for 24 hours. Extracted proteins were run on an SDS-PAGE gel and analyzed using Western blot. Rabbit anti-MAP1203 (1:300) was used as the primary probe and visualized using a secondary goat anti-rabbit IR800 (1:10,000). MAP1203 induction is seen at 52 kDa (arrow). Image is representative of each experiment conducted in the MAP expression studies.

154

A B 2 0 ** 4 **

1 5 3

1 0 2 % B in d in g % I n v a s io n 5 1

0 0

AM 1 2 0 3 p J P 1 2 0 3 p J A M

AM-MA AM-MAP p J p J

C o n s t r u c t C o n s t r u c t

Figure 5.6: Role of MAP1203 in epithelial cell invasion by MAP MAP containing either the empty pJAM2 vector or the pJAM::MAP1203 plasmid was used to infect bovine MDBK epithelial cell monolayers to assess the role of MAP1203 in the binding (a) or invasion (b) of bovine MDBK epithelial cells after a 2 hour infection. Data represent the mean ± SEM of 4 independent experiments each conducted in triplicate (**p<0.01 as determined by Student’s t-test).

155

A B 2 .5 ****

2 .0 ) 2 **** 1 .5

1 .0 Area (units 0 .5

0 .0

0 3 GD SS JAM p P 1 2 0 3 1 2 0 3 R M + M A P 1 2

p J A J A M + M A A M + M A P p p J C o n s t r u c t

Figure 5.7: Colony sizes of MAP containing pJAM::MAP1203 constructs MAP was electroporated with pJAM, pJAM::MAP1203, pJAM::MAP1203ΔRGD, or pJAM::MAP1203ΔSS and grown on 7H10/casein/OADC/MJ plates supplemented with kanamycin (400 µg/ml) for 4 weeks. Colonies were photographed under light microscopy and images were captured under 40x magnification (a) and the area of each colony was measured using the software ImageJ. Images and data represent the mean ± SD of 20 individual colonies photographed and measured (****p<0.0001 as determined by Student’s t test).

156

N-terminal 6x-HN tag C-terminal 6x-HN tag

k D a FT W 1 W 2 Elution FT W 1 W 2 Elution 97 66

45

31

21

14

Figure 5.8: Expression and purification of N-terminal and C-terminal 6xHN­ tagged MAP1203 MAP1203 was cloned into the pET expression vector fused to either an N-terminal 6x HN tag or a C-terminal 6x HN tag for expression of a 52 kDa protein. Extracted proteins were purified using magnetic anti-6xHN beads and extracts, washes, and elution fractions were run on an SDS-PAGE gel. Gel was probed using a 1:2000 goat anti-6xHN primary antibody and visualized using a rabbit anti-goat secondary antibody (1:10,000). 157

References

Bermudez, L. E. and L. S. Young (1994). "Factors affecting invasion of HT-29 and HEp-2 epithelial cells by organisms of the Mycobacterium avium complex." Infect Immun 62(5): 2021-2026.

Patel, D., L. Danelishvili, Y. Yamazaki, M. Alonso, M. L. Paustian, J. P. Bannantine, L. Meunier-Goddik and L. E. Bermudez (2006). "The ability of Mycobacterium avium subsp. paratuberculosis to enter bovine epithelial cells is influenced by preexposure to a hyperosmolar environment and intracellular passage in bovine mammary epithelial cells." Infect Immun 74(5): 2849-2855.

Petersen, T. N., S. Brunak, G. von Heijne and H. Nielsen (2011). "SignalP 4.0: discriminating signal peptides from transmembrane regions." Nat Methods 8(10): 785-786.

Schneider, C. A., W. S. Rasband and K. W. Eliceiri (2012). "NIH Image to ImageJ: 25 years of image analysis." Nat Methods 9(7): 671-675.

Triccas, J. A., T. Parish, W. J. Britton and B. Gicquel (1998). "An inducible expression system permitting the efficient purification of a recombinant antigen from Mycobacterium smegmatis." FEMS Microbiol Lett 167(2): 151-156.

158

Chapter 6

Discussion and Conclusions

159

MAP is a widespread bacterial infection which causes Johne’s disease and the development of severe intestinal inflammation, resulting in wasting and the ultimate death of ruminants including cattle, goats, sheep, and deer worldwide. The astounding economic impact and loss of animals caused by Johne’s disease, estimated to be upwards of $1.5 billion annually in the United States dairy industry alone, drives current research efforts to further understand the disease and identify potential therapeutic strategies. The ability to study the pathogenesis of the bacterium is complicated by its multi-stage disease progression and the ability of subclinical infection to persist for many years within an animal prior to diagnosis. More efficient and effective models are required to fully understand the pathogenic mechanisms utilized by MAP throughout the silent, subclinical, and clinical stages of the disease. Developing and using both new and established systems is key to studying the changing phenotypes of MAP over the course of disease in order to pinpoint specific mechanisms, bacterial targets, and immunological triggers that could be produced at, and responsible for, each stage of Johne’s disease. The shift in disease stages during MAP infection within the host, and the characteristics of those stages have been studied but the mechanism driving the shift has yet to be fully determined. As MAP has been described to change phenotypes under a variety of environmental and host conditions, we hypothesized that the bacterium alters its phenotype over the long and complicated incubation within the bovine intestinal mucosa. However, the ability to successfully model this intestinal interaction over the complete spectrum of Johne’s disease is difficult. While ruminants are the natural host of MAP infection, large animal models are expensive, require immense amounts of time, and can produce inconsistent results due to animal variance factors out of the researchers’ control. Meanwhile, cell culture models are useful, but can oversimplify the complicated range of host-pathogen interactions. In Chapter 3, we describe the development of a novel in vitro model which was created to take advantage of the simplicity of cell culture while still representing the multicellular interaction that MAP has with the intestinal mucosa during infection. Using this model, we demonstrate that one of the initiating factors in the shift of the inflammatory state of the intestinal epithelial cells is the changing phenotype of the 160

bacteria in the tissue. The development of an inflammatory phenotype, as determined by transcriptome analysis and further confirmed by HPLC-ES/MS of the lipidome, was a result of a dramatic shift in the lipid composition of the bacterium after passage through our model. As MAP rely heavily on their lipid composition for invasion of host cells, survival within the intracellular compartment, and the regulation of immune responses, it makes sense that a shift in the lipidome of the bacterium would serve as an important virulence determinant during the course of infection. As our model is a simplified one, in the absence of the intact mucosal environment and immune system, we cannot conclude that the changing phenotype is solely responsible for the shift in disease stages. Rather, the shift in the immune response and onset of inflammation seen within cattle is most likely one mediated in part by the resulting inflammatory phenotype of MAP, in tandem with effects caused by increasing numbers of tissue-localized MAP, development of granulomas, and the presence and function of the lymphocytes, dendritic cells, and macrophages within the infected tissue. Nonetheless, the identification of a novel phenotype of MAP during the interaction with bovine epithelial cells and macrophages, and the identification of the inflammatory properties of that phenotype, results in a novel virulence trait and interesting mechanism utilized by MAP to drive the progression of infection in Johne’s diseased animals. The limitation of in vitro infection models is that while they may be able to derive information about mechanisms that may be at play, they are not representative of the process that is occurring in the natural intact host environment. By identifying the inflammatory phenotype transcripts in naturally MAP-infected bovine intestinal tissue biopsies, and illustrating that those transcripts are found at higher levels in animals exhibiting a more severe disease state, we successful validate our model and its ability to identify not only the inflammatory phenotype described in this study but its ability to be used as a model system for future investigation of Johne’s disease mechanisms. The findings from this novel cell culture model could successfully contribute to the tools used to study and understand the disease. Further analysis of the exact composition of these phenotype-specific components could greatly enhance the field of Johne’s research and prevention by the development of more sensitive and specific 161

assays used for diagnostic purposes. The knowledge that MAP expresses a different and more inflammatory phenotype during the later stages of infection, and the identification of the markers of those phenotypes, provides a toolbox of new antigens and biomarkers exposed on the surface of the bacterium to be used for the screening for and selection of candidates in the development of more effective and preventative vaccines. The desire for new prevention tools against Johne’s disease is abundant as, aside from calf management, there is currently no successful biologically-induced protection or preventative options. The production of antibodies in response to infection is the mechanism used by the humoral arm of the immune system to provide a memory response against infections and to produce protective antibodies upon subsequent infection with the same pathogen later in life. Unfortunately, cattle that are infected with Johne’s disease do not mount a humoral response to the bacterium until the very late stages of the disease, and the response mounted is often too little too later to provide any protective effect against the established infection. The basis of Chapter 2 was to characterize the feasibility of passive immunization and opsonization as a protective mechanism against MAP infection at the cellular level within the host. The notion that MAP can be readily transmitted to young animals from infected dams during early feedings of colostrum and milk, and the previously characterized infectious phenotype developed by MAP upon exposure to milk, provided a novel route and method of immunization and potential protection that had not previously been studied. Our initial hypothesis was that opsonization of MAP would confer a protective outcome against infection in bovine cell culture models. Initial experiments required a reconsideration of our hypothesis, as rather than providing a protective, neutralizing effect during infection of polarized MDBK epithelial cells, opsonized MAP were translocated across the polarized layer at a significantly higher level than MAP that were treated with non-specific serum or not opsonized at all. As the epithelial cell layer provides a barrier to the underlying mucosal tissue, but hosts a limited array of bactericidal mechanisms, we wondered whether the increased translocation would increase the macrophage killing of MAP, or whether the intracellular bacteria would be ingested by macrophages and establish 162

a niche within the phagocytic vacuole for survival. Our data indicated that upon interaction with bovine macrophages, opsonized MAP are quickly killed by innate responses initiated by the host immune cells. Our utilization of inhibitors ruled out the independent role of commonly used cellular mechanisms, including catalase, superoxide anions, NAPDH oxidase mediated compounds, nitric oxide, and macrophage extracellular traps in the killing of opsonized MAP during macrophage infection. These studies led us to determine that a novel killing mechanism was being used by our bovine macrophage (BOMAC) cell line to eliminate the antibody-coated bacteria upon immediate recognition. Current studies aim to characterize the specific killing mechanism employed and is focused on the host utilization of copper- and zinc-derived products based on their toxic effects to other pathogenic microbes and the role of antibody-dependent cell mediated cytotoxicity (ADCC) and the cellular components it uses for the elimination of antibody-opsonized pathogens. In Chapter 2, we discuss the current state of passive immunity and why it may not confer protection against MAP infection to young animals. While little is known about the level of immunoglobulin protection, if any, naturally passed to neonates in the colostrum/milk feedings, this study demonstrates that the passive transfer of anti­ mycobacterial antibodies may provide protection against infection by opsonization of the bacteria and resulting in the rapid killing of MAP upon interaction with host macrophages. Future work on this protective mechanism is required to confirm that our findings using bovine macrophages in vitro translate to protection in vivo. Although murine models can be a useful tool in confirming the efficacy of early protective strategies against MAP infection, the optimal method of determining the ability of passively transferred antibodies to opsonize MAP and protect from infection would be the use of a neonatal calf model. The unique digestive and intestinal physiology is imperative in accurately understanding the role of opsonization during MAP infection and determining whether it can protect the neonatal animal from subsequent intestinal colonization and infection. Ultimately, the results described here indicate an alternative method of immunization which may be able to provide protection against a disease for which there is currently no successful and widely utilized vaccine. 163

The interplay between MAP and the host macrophage is a complex interaction built between a host cell with the primary goal to eliminate intracellular pathogens, and the bacterium that has evolved a repertoire of functions to evade such host- mediated mechanisms and survive within the intracellular vacuole. While a number of the bacterial mediated techniques have been elucidated, including the inhibition of maturation of mycobacterium-containing endosome, the inhibition of phagosome acidification, and the genetic elements that allow for detoxification of innate host- produced products, there are a variety of interactions that have yet to be understood between the host phagocyte and MAP during infection. By using Acanthamoeba castellanii as a model phagocyte in Chapter 4, we aimed to analyze the interaction between MAP and the host phagocyte over the short- and long-term course of infection. For our studies, we chose to focus on the interaction between the intracellular survival of the bacterium and the metabolic activity of the host phagocyte, with the goal of determining how each one affects the other during disease. We demonstrate that MAP, in a viability-dependent manner, actively stimulates the metabolic activity of the amoeba host over the course of infection. This stimulation, significantly higher during early infection, results in the intracellular killing of MAP, while the growth of the bacteria and increase in the intracellular burden is matched with a lower state of amoeba metabolic activity. Using phenotype array assays, we begin to understand how amoeba react to MAP infection, or alternately how MAP manipulates the host to utilize nutrients that are more beneficial to the intracellular pathogen than they are to the host itself. Current data analysis of phenotype array patterns and the specific carbon source utilization initiated by wild- type MAP-infected amoeba from day 0-1, 3-4, and 6-7 during infection will allow us to understand the temporal pattern of MAP infection and its exploitation of the intracellular phagocyte metabolism. To determine the bacterial mechanisms involved in this phenomenon we identified MAP mutants which demonstrate a hyper- or deficient stimulation of amoeba metabolism and exhibit altered growth patterns within amoeba. As a limitation to our study, amoeba may not provide a completely identical or relevant 164

host for identifying metabolic MAP mutants. Thus, we confirmed that each of our selected strains demonstrated a mutated ability to infect and/or survive within bovine macrophages during infection to validate our identification of mutants using the amoeba model system. The investigation of the mutants and their genetic mutation(s) will allow for insight into their phenotype array patterns and altered usage of carbon sources by amoeba infected with each strain, and ultimately may shed light on the bacterial-derived mechanisms responsible for the shift in host cell metabolism. Perhaps, the knowledge of the metabolic interactions at play may aid in the treatment or prevention of Johne’s disease if the attenuation or elimination of MAP virulence or growth can be conferred upon addition, or depletion, of common carbon or nutrient sources within animal feed. As described previously in Chapter 2, the infectious phenotype derived after milk-exposure is one that results in greater virulence and invasion of host cells. By studying the highly upregulated gene MAP1203, in Chapter 5 we begin to describe the function of the protein, and its domains, in the initiation of intestinal infection. The RGD binding domain, pertinent for integrin binding and cellular uptake, is not readily used in the initial stages of infection as intestinal epithelial enterocytes do not contain integrins on their surface. It is possible that as MAP utilizes M-cells for uptake, and the surface of those cells are decorated in concentrated patterns of the integrin fibronectin, the bacterium utilizes the RGD binding domain to facilitate the initial binding and subsequent uptake by the M-cells during early stages of infection; though this hypothesis has yet to be studied. Rather, it seems that the complete protein is required for the ability of binding and invasion of the host epithelial cell layer. The putative cell membrane associated protein requires the intact signal sequence and putative cleavage site for protein functionality, as deletion mutants abrogate the effect of binding and invasion of host epithelial cells and survival within macrophages. With current studies underway, future work aims to identify the bovine epithelial protein(s) that serve as a receptor or binding partner to MAP1203 during the invasion process. The elucidation of this protein, by using protein pull-down assays, analysis using mass spectrometry, and confirmation by a yeast 2-hybrid 165

system, will allow for the determination of a potentially novel host cellular mechanism responsible for the ingestion of MAP at the beginning of infection. Overall, this study offers insight into the protein MAP1203 and its role as an important mycobacterial virulence factor. Specifically, early studies reveal it may be an important virulence protein involved in binding and invasion of the bovine intestinal epithelium upon delivery of the bacterium in milk. The identification and characterization of the hypothesized cell surface associated MAP1203 protein provides a novel and potentially interesting target for the development of new and therapeutic strategies for the protection of young animals from infection with MAP and the development of Johne’s disease. In total, the research involved in this dissertation describes the development and utilization of new models useful in the study of mycobacterial infection (Figure 6.1). The in vitro cell culture passage model offers simplified insight into the interactions that occur between the host and infectious pathogens during mucosal infection. Meanwhile, the use of amoeba serves as a useful cell model and offers insight to MAP pathogenesis and its metabolic interaction with the phagocytic host. Additionally, described in Appendix 1, for the first time we describe and characterize the nematode model C. elegans as a simple model system and tool to study intestinal colonization by mycobacteria, specifically Mycobacterium avium subspecies hominissuis, and the ability to use the nematode to identify bacterial factors important in binding and colonization of the intestinal epithelium. We demonstrate how the changing phenotype of MAP within milk, used to increase virulence of the bacterium, can be used against the pathogen. We illustrate how opsonization of the infectious MAP phenotype results in rapid killing by host immune cells and provides a potentially novel method of protection that needs to be further studied to determine its complete protective efficacy. Lastly, we begin to elucidate the role of MAP1203 as an important protein in the infectious phenotype of milk-exposed MAP and start to understand its function, and its role as a potential vaccine target, during the initiation of MAP infection within the intestinal epithelium. Our findings, development of models, investigation of phenotypes, and increased understanding of the bacterial virulence strategies of MAP hold the potential to lead to new and enhanced 166

diagnostics, functional therapeutics, and relevant vaccine candidates which may contribute to decreasing both the incidence and prevalence of Johne’s disease in cattle herds, and in other susceptible species, across the globe.

167

Figure 6.1 Experimental findings described in the chapters of this dissertation

168

Bibliography

(2014). "Johne's Information Central." Retrieved August 10, 2014, 2014.

Abu Kwaik, Y. and D. Bumann (2013). "Microbial quest for food in vivo: 'nutritional virulence' as an emerging paradigm." Cell Microbiol 15(6): 882-890.

Adaska, J. M., C. A. Munoz-Zanzi and S. K. Hietala (2002). "Evaluation of result variability with a commercial Johne's disease enzyme-linked immunosorbent assay kit and repeat testing of samples." J Vet Diagn Invest 14(5): 423-426.

Al-Quadan, T., C. T. Price and Y. Abu Kwaik (2012). "Exploitation of evolutionarily conserved amoeba and mammalian processes by Legionella." Trends Microbiol 20(6): 299-306.

Alonso-Hearn, M., T. M. Eckstein, S. Sommer and L. E. Bermudez (2010). "A Mycobacterium avium subsp. paratuberculosis LuxR regulates cell envelope and virulence." Innate Immun 16(4): 235-247.

Alonso-Hearn, M., E. Molina, M. Geijo, P. Vazquez, I. Sevilla, J. M. Garrido and R. A. Juste (2009). "Isolation of Mycobacterium avium subsp. paratuberculosis from muscle tissue of naturally infected cattle." Foodborne Pathog Dis 6(4): 513-518.

Alonso-Hearn, M., D. Patel, L. Danelishvili, L. Meunier-Goddik and L. E. Bermudez (2008). "The Mycobacterium avium subsp. paratuberculosis MAP3464 gene encodes an oxidoreductase involved in invasion of bovine epithelial cells through the activation of host cell Cdc42." Infect Immun 76(1): 170­ 178.

Antognoli, M. C., F. B. Garry, H. L. Hirst, J. E. Lombard, M. M. Dennis, D. H. Gould and M. D. Salman (2008). "Characterization of Mycobacterium avium subspecies paratuberculosis disseminated infection in dairy cattle and its association with antemortem test results." Vet Microbiol 127(3-4): 300­ 308.

Aulik, N. A., K. M. Hellenbrand and C. J. Czuprynski (2012). "Mannheimia haemolytica and its leukotoxin cause macrophage extracellular trap formation by bovine macrophages." Infect Immun 80(5): 1923-1933.

Bannantine, J. P., J. L. Everman, S. J. Rose, L. Babrak, R. Katani, R. G. Barletta, A. M. Talaat, Y. T. Grohn, Y. F. Chang, V. Kapur and L. E. Bermudez (2014). "Evaluation of eight live attenuated vaccine candidates for protection against challenge with virulent Mycobacterium avium subspecies paratuberculosis in mice." Front Cell Infect Microbiol 4: 88.

Bannantine, J. P., L. Li, M. Mwangi, R. Cote, J. A. Raygoza Garay and V. Kapur (2014). "Complete Genome Sequence of Mycobacterium avium subsp. paratuberculosis, Isolated from Human Breast Milk." Genome Announc 2(1).

Bannantine, J. P., Q. Zhang, L. L. Li and V. Kapur (2003). "Genomic homogeneity between Mycobacterium avium subsp. avium and Mycobacterium avium subsp. paratuberculosis belies their divergent growth rates." BMC Microbiol 3: 10.

Bansal-Mutalik, R. and H. Nikaido (2014). "Mycobacterial outer membrane is a lipid bilayer and the inner membrane is unusually rich in diacyl phosphatidylinositol dimannosides." Proc Natl Acad Sci U S A 111(13): 4958-4963.

Bastida, F. and R. A. Juste (2011). "Paratuberculosis control: a review with a focus on vaccination." J Immune Based Ther Vaccines 9: 8. 169

Begg, D. J. and R. J. Whittington (2008). "Experimental animal infection models for Johne's disease, an infectious enteropathy caused by Mycobacterium avium subsp. paratuberculosis." Vet J 176(2): 129-145.

Bermudez, L. E., M. Petrofsky and F. Sangari (2004). "Intracellular phenotype of Mycobacterium avium enters macrophages primarily by a macropinocytosis-like mechanism and survives in a compartment that differs from that with extracellular phenotype." Cell Biol Int 28(5): 411-419.

Bermudez, L. E., M. Petrofsky, S. Sommer and R. G. Barletta (2010). "Peyer's patch-deficient mice demonstrate that Mycobacterium avium subsp. paratuberculosis translocates across the mucosal barrier via both M cells and enterocytes but has inefficient dissemination." Infect Immun 78(8): 3570-3577.

Bermudez, L. E., F. J. Sangari, P. Kolonoski, M. Petrofsky and J. Goodman (2002). "The efficiency of the translocation of Mycobacterium tuberculosis across a bilayer of epithelial and endothelial cells as a model of the alveolar wall is a consequence of transport within mononuclear phagocytes and invasion of alveolar epithelial cells." Infect Immun 70(1): 140-146.

Bermudez, L. E. and L. S. Young (1994). "Factors affecting invasion of HT-29 and HEp-2 epithelial cells by organisms of the Mycobacterium avium complex." Infect Immun 62(5): 2021-2026.

Boiadjieva, S., C. Hallberg, M. Hogstrom and C. Busch (1984). "Methods in laboratory investigation. Exclusion of trypan blue from microcarriers by endothelial cells: an in vitro barrier function test." Lab Invest 50(2): 239-246.

Brinkmann, V., U. Reichard, C. Goosmann, B. Fauler, Y. Uhlemann, D. S. Weiss, Y. Weinrauch and A. Zychlinsky (2004). "Neutrophil extracellular traps kill bacteria." Science 303(5663): 1532-1535.

Buergelt, C. D., C. Hall, K. McEntee and J. R. Duncan (1978). "Pathological evaluation of paratuberculosis in naturally infected cattle." Vet Pathol 15(2): 196-207.

Buza, J. J., Y. Mori, A. M. Bari, H. Hikono, g. Aodon, S. Hirayama, Y. Shu and E. Momotani (2003). "Mycobacterium avium subsp. paratuberculosis infection causes suppression of RANTES, monocyte chemoattractant protein 1, and tumor necrosis factor alpha expression in peripheral blood of experimentally infected cattle." Infect Immun 71(12): 7223-7227.

Cain, C. C., D. M. Sipe and R. F. Murphy (1989). "Regulation of endocytic pH by the Na+,K+-ATPase in living cells." Proc Natl Acad Sci U S A 86(2): 544-548.

Cambier, C. J., K. K. Takaki, R. P. Larson, R. E. Hernandez, D. M. Tobin, K. B. Urdahl, C. L. Cosma and L. Ramakrishnan (2014). "Mycobacteria manipulate macrophage recruitment through coordinated use of membrane lipids." Nature 505(7482): 218-222.

Chacon, O., L. E. Bermudez and R. G. Barletta (2004). "Johne's disease, inflammatory bowel disease, and Mycobacterium paratuberculosis." Annu Rev Microbiol 58: 329-363.

Chao, M. C., K. J. Kieser, S. Minami, D. Mavrici, B. B. Aldridge, S. M. Fortune, T. Alber and E. J. Rubin (2013). "Protein complexes and proteolytic activation of the cell wall hydrolase RipA regulate septal resolution in mycobacteria." PLoS Pathog 9(2): e1003197.

Chiodini, R. J. and J. Hermon-Taylor (1993). "The thermal resistance of Mycobacterium paratuberculosis in raw milk under conditions simulating pasteurization." J Vet Diagn Invest 5(4): 629­ 631.

Chiodini, R. J., H. J. Van Kruiningen and R. S. Merkal (1984). "Ruminant paratuberculosis (Johne's disease): the current status and future prospects." Cornell Vet 74(3): 218-262. 170

Chiodini, R. J., H. J. Van Kruiningen, R. S. Merkal, W. R. Thayer, Jr. and J. A. Coutu (1984). "Characteristics of an unclassified Mycobacterium species isolated from patients with Crohn's disease." J Clin Microbiol 20(5): 966-971.

Christian, J. G., J. Heymann, S. A. Paschen, J. Vier, L. Schauenburg, J. Rupp, T. F. Meyer, G. Hacker and D. Heuer (2011). "Targeting of a chlamydial protease impedes intracellular bacterial growth." PLoS Pathog 7(9): e1002283.

Cirillo, J. D., S. Falkow, L. S. Tompkins and L. E. Bermudez (1997). "Interaction of Mycobacterium avium with environmental amoebae enhances virulence." Infect Immun 65(9): 3759-3767.

Collins, M. (2013, November 2013). "The Merck Veterinary Manual: Overview of Paratuberculosis." Retrieved October 2014, 2014, from http://www.merckmanuals.com/vet/generalized_conditions/paratuberculosis/overview_of_paratubercul osis.html.

Collins, M. M., E. (2014, 2014). "Johne's Information Center." 2014, from http://johnes.org/.

Cooney, M. A., J. L. Steele, H. Steinberg and A. M. Talaat (2014). "A murine oral model for Mycobacterium avium subsp. paratuberculosis infection and immunomodulation with Lactobacillus casei ATCC 334." Front Cell Infect Microbiol 4: 11.

Cossu, D., E. Cocco, D. Paccagnini, S. Masala, N. Ahmed, J. Frau, M. G. Marrosu and L. A. Sechi (2011). "Association of Mycobacterium avium subsp. paratuberculosis with multiple sclerosis in Sardinian patients." PLoS One 6(4): e18482.

Cossu, D., S. Masala, E. Cocco, D. Paccagnini, J. Frau, M. G. Marrosu and L. A. Sechi (2012). "Are Mycobacterium avium subsp. paratuberculosis and Epstein-Barr virus triggers of multiple sclerosis in Sardinia?" Mult Scler 18(8): 1181-1184.

Cousins, D. V., R. Whittington, I. Marsh, A. Masters, R. J. Evans and P. Kluver (1999). "Mycobacteria distenct from Mycobacterium avium subsp. paratuberculosis isolated from the faeces of ruminants possess IS900-like sequences detectable IS900 polymerase chain reaction: implications for diagnosis." Mol Cell Probes 13(6): 431-442.

David, J., H. W. Barkema, R. Mortier, S. Ghosh, L. Guan le and J. De Buck (2014). "Gene expression profiling and putative biomarkers of calves 3 months after infection with Mycobacterium avium subspecies paratuberculosis." Vet Immunol Immunopathol 160(1-2): 107-117.

Davies, B. C., L.S.; Edwards, S.W. (1991). "Superoxide generation during phagocytosis by Acanthamoeba castellanii: similarities to the respiratory burst of immune phagocytes." J Gen Microbiol 137: 705-710.

De Assis, M. C., A. O. Da Costa, T. C. Barja-Fidalgo and M. C. Plotkowski (2000). "Human endothelial cells are activated by interferon-gamma plus tumour necrosis factor-alpha to kill intracellular Pseudomonas aeruginosa." Immunology 101(2): 271-278.

Diard, M., V. Garcia, L. Maier, M. N. Remus-Emsermann, R. R. Regoes, M. Ackermann and W. D. Hardt (2013). "Stabilization of cooperative virulence by the expression of an avirulent phenotype." Nature 494(7437): 353-356.

Dicksved, J., J. Halfvarson, M. Rosenquist, G. Jarnerot, C. Tysk, J. Apajalahti, L. Engstrand and J. K. Jansson (2008). "Molecular analysis of the gut microbiota of identical twins with Crohn's disease." ISME J 2(7): 716-727. 171

Early, J. and L. E. Bermudez (2011). "Mimicry of the pathogenic mycobacterium vacuole in vitro elicits the bacterial intracellular phenotype, including early-onset macrophage death." Infect Immun 79(6): 2412-2422.

Eckstein, T. M., S. Chandrasekaran, S. Mahapatra, M. R. McNeil, D. Chatterjee, C. D. Rithner, P. W. Ryan, J. T. Belisle and J. M. Inamine (2006). "A major cell wall lipopeptide of Mycobacterium avium subspecies paratuberculosis." J Biol Chem 281(8): 5209-5215.

Edgar, R., M. Domrachev and A. E. Lash (2002). "Gene Expression Omnibus: NCBI gene expression and hybridization array data repository." Nucleic Acids Res 30(1): 207-210.

Ehrt, S. and D. Schnappinger (2007). "Mycobacterium tuberculosis virulence: lipids inside and out." Nat Med 13(3): 284-285.

Elguezabal, N., F. Bastida, I. A. Sevilla, N. Gonzalez, E. Molina, J. M. Garrido and R. A. Juste (2011). "Estimation of Mycobacterium avium subsp. paratuberculosis growth parameters: strain characterization and comparison of methods." Appl Environ Microbiol 77(24): 8615-8624.

Elwell, C. A., S. Jiang, J. H. Kim, A. Lee, T. Wittmann, K. Hanada, P. Melancon and J. N. Engel (2011). "Chlamydia trachomatis co-opts GBF1 and CERT to acquire host sphingomyelin for distinct roles during intracellular development." PLoS Pathog 7(9): e1002198.

Facciuolo, A., D. F. Kelton and L. M. Mutharia (2013). "Novel secreted antigens of Mycobacterium paratuberculosis as serodiagnostic biomarkers for Johne's disease in cattle." Clin Vaccine Immunol 20(12): 1783-1791.

Fecteau, M. E. and R. H. Whitlock (2011). "Treatment and chemoprophylaxis for paratuberculosis." Vet Clin North Am Food Anim Pract 27(3): 547-557, v.

Frank, D. N., A. L. St Amand, R. A. Feldman, E. C. Boedeker, N. Harpaz and N. R. Pace (2007). "Molecular-phylogenetic characterization of microbial community imbalances in human inflammatory bowel diseases." Proc Natl Acad Sci U S A 104(34): 13780-13785.

Fratti, R. A., J. Chua, I. Vergne and V. Deretic (2003). "Mycobacterium tuberculosis glycosylated phosphatidylinositol causes phagosome maturation arrest." Proc Natl Acad Sci U S A 100(9): 5437­ 5442.

Freund, J. (1956). "The mode of action of immunologic adjuvants." Bibl Tuberc(10): 130-148.

Fuchs, T. A., U. Abed, C. Goosmann, R. Hurwitz, I. Schulze, V. Wahn, Y. Weinrauch, V. Brinkmann and A. Zychlinsky (2007). "Novel cell death program leads to neutrophil extracellular traps." J Cell Biol 176(2): 231-241.

Gao, L. Y., M. Pak, R. Kish, K. Kajihara and E. J. Brown (2006). "A mycobacterial operon essential for virulence in vivo and invasion and intracellular persistence in macrophages." Infect Immun 74(3): 1757-1767.

Geisel, R. E., K. Sakamoto, D. G. Russell and E. R. Rhoades (2005). "In vivo activity of released cell wall lipids of Mycobacterium bovis bacillus Calmette-Guerin is due principally to trehalose mycolates." J Immunol 174(8): 5007-5015.

Gillespie, J., L. L. Barton and E. W. Rypka (1986). "Phenotypic changes in mycobacteria grown in oxygen-limited conditions." J Med Microbiol 21(3): 251-255. 172

Godden, S. (2008). "Colostrum management for dairy calves." Vet Clin North Am Food Anim Pract 24(1): 19-39.

Grant, I. R., E. I. Hitchings, A. McCartney, F. Ferguson and M. T. Rowe (2002). "Effect of commercial-scale high-temperature, short-time pasteurization on the viability of Mycobacterium paratuberculosis in naturally infected cows' milk." Appl Environ Microbiol 68(2): 602-607.

Greenberg, S. and S. Grinstein (2002). "Phagocytosis and innate immunity." Curr Opin Immunol 14(1): 136-145.

Gross, A., N. Dugas, S. Spiesser, I. Vouldoukis, C. Damais, J. P. Kolb, B. Dugas and J. Dornand (1998). "Nitric oxide production in human macrophagic cells phagocytizing opsonized zymosan: direct characterization by measurement of the luminol dependent chemiluminescence." Free Radic Res 28(2): 179-191.

Hagblom, P., E. Segal, E. Billyard and M. So (1985). "Intragenic recombination leads to pilus antigenic variation in Neisseria gonorrhoeae." Nature 315(6015): 156-158.

Harriff, M. and L. E. Bermudez (2009). "Environmental amoebae and mycobacterial pathogenesis." Methods Mol Biol 465: 433-442.

Hendrick, S. H., T. F. Duffield, K. E. Leslie, K. D. Lissemore, M. Archambault, R. Bagg, P. Dick and D. F. Kelton (2006). "Monensin might protect Ontario, Canada dairy cows from paratuberculosis milk- ELISA positivity." Prev Vet Med 76(3-4): 237-248.

Hensel, M., A. P. Hinsley, T. Nikolaus, G. Sawers and B. C. Berks (1999). "The genetic basis of tetrathionate respiration in Salmonella typhimurium." Mol Microbiol 32(2): 275-287.

Hines, M. E., 2nd, J. R. Stabel, R. W. Sweeney, F. Griffin, A. M. Talaat, D. Bakker, G. Benedictus, W. C. Davis, G. W. de Lisle, I. A. Gardner, R. A. Juste, V. Kapur, A. Koets, J. McNair, G. Pruitt and R. H. Whitlock (2007). "Experimental challenge models for Johne's disease: a review and proposed international guidelines." Vet Microbiol 122(3-4): 197-222.

Hines, M. E., 2nd, S. E. Turnquist, M. R. Ilha, S. Rajeev, A. L. Jones, L. Whittington, J. P. Bannantine, R. G. Barletta, Y. T. Grohn, R. Katani, A. M. Talaat, L. Li and V. Kapur (2014). "Evaluation of novel oral vaccine candidates and validation of a caprine model of Johne's disease." Front Cell Infect Microbiol 4: 26.

Hostetter, J., R. Kagan and E. Steadham (2005). "Opsonization effects on Mycobacterium avium subsp. paratuberculosis--macrophage interactions." Clin Diagn Lab Immunol 12(6): 793-796.

Janeway, C. J., C.A.; Travers, P.; Walport, M.; Shlomchik, A. (2001). Immunobiology, Garland Science.

Johne, H. A. L. F. (1895). "Ein eigenthumlicher fall von tuberculose beim rind [A peculiar case of tuberculosis in a cow]." Deutsche. Zeitschr. Tierm. Path. 21: 438 - 454.

Karakousis, P. C., T. Yoshimatsu, G. Lamichhane, S. C. Woolwine, E. L. Nuermberger, J. Grosset and W. R. Bishai (2004). "Dormancy phenotype displayed by extracellular Mycobacterium tuberculosis within artificial granulomas in mice." J Exp Med 200(5): 647-657.

Kentner, D., G. Martano, M. Callon, P. Chiquet, M. Brodmann, O. Burton, A. Wahlander, P. Nanni, N. Delmotte, J. Grossmann, J. Limenitakis, R. Schlapbach, P. Kiefer, J. A. Vorholt, S. Hiller and D. Bumann (2014). "Shigella reroutes host cell central metabolism to obtain high-flux nutrient supply for vigorous intracellular growth." Proc Natl Acad Sci U S A 111(27): 9929-9934. 173

Khan, F. A., Z. I. Chaudhry, M. I. Ali, S. Khan, N. Mumtaz and I. Ahmad (2010). "Detection of Mycobacterium avium subsp. paratuberculosis in tissue samples of cattle and buffaloes." Trop Anim Health Prod 42(4): 633-638.

Khare, S., S. D. Lawhon, K. L. Drake, J. E. Nunes, J. F. Figueiredo, C. A. Rossetti, T. Gull, R. E. Everts, H. A. Lewin, C. L. Galindo, H. R. Garner and L. G. Adams (2012). "Systems biology analysis of gene expression during in vivo Mycobacterium avium paratuberculosis enteric colonization reveals role for immune tolerance." PLoS One 7(8): e42127.

Khatri, B., M. Fielder, G. Jones, W. Newell, M. Abu-Oun and P. R. Wheeler (2013). "High throughput phenotypic analysis of Mycobacterium tuberculosis and Mycobacterium bovis strains' metabolism using biolog phenotype microarrays." PLoS One 8(1): e52673.

Kim, J. S., H. Y. Kwon, W. H. Choi, C. Y. Jeon, J. I. Kim, J. Kim, J. Y. Lee, Y. S. Kim and J. B. Park (2003). "Phagocytosis of serum- and IgG-opsonized zymosan particles induces apoptosis through superoxide but not nitric oxide in macrophage J774A.1." Exp Mol Med 35(3): 211-221.

Kinchen, J. M. and K. S. Ravichandran (2008). "Phagosome maturation: going through the acid test." Nat Rev Mol Cell Biol 9(10): 781-795.

Kinyoun, J. J. (1915). "A Note on Uhlenhuths Method for Sputum Examination, for Tubercle Bacilli." Am J Public Health (N Y) 5(9): 867-870.

Koc, A., I. Bargen, A. Suwandi, M. Roderfeld, A. Tschuschner, T. Rath, G. F. Gerlach, M. Hornef, R. Goethe, S. Weiss and E. Roeb (2014). "Systemic and mucosal immune reactivity upon Mycobacterium avium ssp. paratuberculosis infection in mice." PLoS One 9(4): e94624.

Koenig, G. J., G. F. Hoffsis, W. P. Shulaw, S. Bech-Nielsen, D. M. Rings and G. St-Jean (1993). "Isolation of Mycobacterium paratuberculosis from mononuclear cells in tissues, blood, and mammary glands of cows with advanced paratuberculosis." Am J Vet Res 54(9): 1441-1445.

Konnai, S., T. Usui, K. Ohashi and M. Onuma (2003). "The rapid quantitative analysis of bovine cytokine genes by real-time RT-PCR." Vet Microbiol 94(4): 283-294.

Krishnan, M. Y., E. J. Manning and M. T. Collins (2009). "Comparison of three methods for susceptibility testing of Mycobacterium avium subsp. paratuberculosis to 11 antimicrobial drugs." J Antimicrob Chemother 64(2): 310-316.

Kuehnel, M. P., R. Goethe, A. Habermann, E. Mueller, M. Rohde, G. Griffiths and P. Valentin- Weigand (2001). "Characterization of the intracellular survival of Mycobacterium avium ssp. paratuberculosis: phagosomal pH and fusogenicity in J774 macrophages compared with other mycobacteria." Cell Microbiol 3(8): 551-566.

Kuo, T. T., K. Baker, M. Yoshida, S. W. Qiao, V. G. Aveson, W. I. Lencer and R. S. Blumberg (2010). "Neonatal Fc receptor: from immunity to therapeutics." J Clin Immunol 30(6): 777-789.

Lamont, E. A., A. M. Talaat, P. M. Coussens, J. P. Bannantine, Y. T. Grohn, R. Katani, L. L. Li, V. Kapur and S. Sreevatsan (2014). "Screening of Mycobacterium avium subsp. paratuberculosis mutants for attenuation in a bovine monocyte-derived macrophage model." Front Cell Infect Microbiol 4: 87.

Lamont, E. A., W. W. Xu and S. Sreevatsan (2013). "Host-Mycobacterium avium subsp. paratuberculosis interactome reveals a novel iron assimilation mechanism linked to nitric oxide stress during early infection." BMC Genomics 14: 694. 174

Lee, H., J. R. Stabel and M. E. Kehrli, Jr. (2001). "Cytokine gene expression in ileal tissues of cattle infected with Mycobacterium paratuberculosis." Vet Immunol Immunopathol 82(1-2): 73-85.

Lee, W., B. C. VanderVen, R. J. Fahey and D. G. Russell (2013). "Intracellular Mycobacterium tuberculosis exploits host-derived fatty acids to limit metabolic stress." J Biol Chem 288(10): 6788­ 6800.

Li, L., J. P. Bannantine, Q. Zhang, A. Amonsin, B. J. May, D. Alt, N. Banerji, S. Kanjilal and V. Kapur (2005). "The complete genome sequence of Mycobacterium avium subspecies paratuberculosis." Proc Natl Acad Sci U S A 102(35): 12344-12349.

Liu, P., X. Wu, C. Liao, X. Liu, J. Du, H. Shi, X. Wang, X. Bai, P. Peng, L. Yu, F. Wang, Y. Zhao and M. Liu (2014). "Escherichia coli and Candida albicans induced macrophage extracellular trap-like structures with limited microbicidal activity." PLoS One 9(2): e90042.

Livak, K. J. and T. D. Schmittgen (2001). "Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method." Methods 25(4): 402-408.

Mangum, J. B., J. I. Everitt, J. C. Bonner, L. R. Moore and A. R. Brody (1990). "Co-culture of primary pulmonary cells to model alveolar injury and translocation of proteins." In Vitro Cell Dev Biol 26(12): 1135-1143.

Manichanh, C., L. Rigottier-Gois, E. Bonnaud, K. Gloux, E. Pelletier, L. Frangeul, R. Nalin, C. Jarrin, P. Chardon, P. Marteau, J. Roca and J. Dore (2006). "Reduced diversity of faecal microbiota in Crohn's disease revealed by a metagenomic approach." Gut 55(2): 205-211.

Marassi, C. D. F., S.; Gonzaga, J.S.; Ristow, P.; Ferreira, R.; Oelemann, W.M.R.; Fonseca, L.S.; Lilenbaum, W. (2005). Interference of anti-M. bovie antibodies in serological tests for paratuberculosis. International Colloquium on Paratuberculosis.

Mayer, B., Z. Kis, G. Kajan, L. V. Frenyo, L. Hammarstrom and I. Kacskovics (2004). "The neonatal Fc receptor (FcRn) is expressed in the bovine lung." Vet Immunol Immunopathol 98(1-2): 85-89.

McAdam, R. A., T. R. Weisbrod, J. Martin, J. D. Scuderi, A. M. Brown, J. D. Cirillo, B. R. Bloom and W. R. Jacobs, Jr. (1995). "In vivo growth characteristics of leucine and methionine auxotrophic mutants of Mycobacterium bovis BCG generated by transposon mutagenesis." Infect Immun 63(3): 1004-1012.

McNabe, M., R. Tennant, L. Danelishvili, L. Young and L. E. Bermudez (2011). "Mycobacterium avium ssp. hominissuis biofilm is composed of distinct phenotypes and influenced by the presence of antimicrobials." Clin Microbiol Infect 17(5): 697-703.

McNamara, M., S. C. Tzeng, C. Maier, L. Zhang and L. E. Bermudez (2012). "Surface proteome of "Mycobacterium avium subsp. hominissuis" during the early stages of macrophage infection." Infect Immun 80(5): 1868-1880.

Medzhitov, R. (2001). "Toll-like receptors and innate immunity." Nat Rev Immunol 1(2): 135-145.

Miller, B. H., R. A. Fratti, J. F. Poschet, G. S. Timmins, S. S. Master, M. Burgos, M. A. Marletta and V. Deretic (2004). "Mycobacteria inhibit nitric oxide synthase recruitment to phagosomes during macrophage infection." Infect Immun 72(5): 2872-2878.

Miltner, E. C. and L. E. Bermudez (2000). "Mycobacterium avium grown in Acanthamoeba castellanii is protected from the effects of antimicrobials." Antimicrob Agents Chemother 44(7): 1990-1994. 175

Minnikin, D. (1982). Lipids: complex lipids their chemistry, biosythesis and roles. The biology of the mycobacteria. S. J. Ratledge C. London, Academic Press. 1: 95-185.

Mitchell, R. M., R. H. Whitlock, S. M. Stehman, A. Benedictus, P. P. Chapagain, Y. T. Grohn and Y. H. Schukken (2008). "Simulation modeling to evaluate the persistence of Mycobacterium avium subsp. paratuberculosis (MAP) on commercial dairy farms in the United States." Prev Vet Med 83(3-4): 360­ 380.

Mohanan, S., S. Horibata, J. L. McElwee, A. J. Dannenberg and S. A. Coonrod (2013). "Identification of macrophage extracellular trap-like structures in mammary gland adipose tissue: a preliminary study." Front Immunol 4: 67.

Nagy, B. (1980). "Vaccination of cows with a K99 extract to protect newborn calves against experimental enterotoxic colibacillosis." Infect Immun 27(1): 21-24.

Neut, C., P. Bulois, P. Desreumaux, J. M. Membre, E. Lederman, L. Gambiez, A. Cortot, P. Quandalle, H. van Kruiningen and J. F. Colombel (2002). "Changes in the bacterial flora of the neoterminal ileum after ileocolonic resection for Crohn's disease." Am J Gastroenterol 97(4): 939-946.

Ott, S. L., S. J. Wells and B. A. Wagner (1999). "Herd-level economic losses associated with Johne's disease on US dairy operations." Prev Vet Med 40(3-4): 179-192.

Patel, D., L. Danelishvili, Y. Yamazaki, M. Alonso, M. L. Paustian, J. P. Bannantine, L. Meunier- Goddik and L. E. Bermudez (2006). "The ability of Mycobacterium avium subsp. paratuberculosis to enter bovine epithelial cells is influenced by preexposure to a hyperosmolar environment and intracellular passage in bovine mammary epithelial cells." Infect Immun 74(5): 2849-2855.

PDYKEMA, N.-. (2011). HPC Decontamination of Fecal Swabs and Feces Submitted for Mycobacteria Culture. N. V. S. Laboratory. Ames, Iowa.

Penketh, A., T. Pitt, D. Roberts, M. E. Hodson and J. C. Batten (1983). "The relationship of phenotype changes in Pseudomonas aeruginosa to the clinical condition of patients with cystic fibrosis." Am Rev Respir Dis 127(5): 605-608.

Periasamy, S., B. N. Tripathi and N. Singh (2013). "Mechanisms of Mycobacterium avium subsp. paratuberculosis induced apoptosis and necrosis in bovine macrophages." Vet Microbiol 165(3-4): 392-401.

Perryman, L. E., S. J. Kapil, M. L. Jones and E. L. Hunt (1999). "Protection of calves against cryptosporidiosis with immune bovine colostrum induced by a Cryptosporidium parvum recombinant protein." Vaccine 17(17): 2142-2149.

Pesch, K. L. S., U. (1929). Combined assays for lactose and galactose by enzymatic reactions. Dietary Sugars: Chemistry, Analysis, Function and effects. V. R. Preedy, Royal Society of Chemistry: 395­ 404.

Petersen, T. N., S. Brunak, G. von Heijne and H. Nielsen (2011). "SignalP 4.0: discriminating signal peptides from transmembrane regions." Nat Methods 8(10): 785-786.

Price, C. T., T. Al-Quadan, M. Santic, I. Rosenshine and Y. Abu Kwaik (2011). "Host proteasomal degradation generates amino acids essential for intracellular bacterial growth." Science 334(6062): 1553-1557.

Rainwater, D. L. and P. E. Kolattukudy (1985). "Fatty acid biosynthesis in Mycobacterium tuberculosis var. bovis Bacillus Calmette-Guerin. Purification and characterization of a novel fatty acid 176

synthase, mycocerosic acid synthase, which elongates n-fatty acyl-CoA with methylmalonyl-CoA." J Biol Chem 260(1): 616-623.

Rawat, M., C. Johnson, V. Cadiz and Y. Av-Gay (2007). "Comparative analysis of mutants in the mycothiol biosynthesis pathway in Mycobacterium smegmatis." Biochem Biophys Res Commun 363(1): 71-76.

Reiter, B. and J. H. Brock (1975). "Inhibition of Escherichia coli by bovine colostrum and post­ colostral milk. I. Complement-mediated bactericidal activity of antibodies to a serum susceptible strain of E. coli of the serotype O 111." Immunology 28(1): 71-82.

Ryan, G. J., D. R. Hoff, E. R. Driver, M. I. Voskuil, M. Gonzalez-Juarrero, R. J. Basaraba, D. C. Crick, J. S. Spencer and A. J. Lenaerts (2010). "Multiple M. tuberculosis phenotypes in mouse and guinea pig lung tissue revealed by a dual-staining approach." PLoS One 5(6): e11108.

Saif, L. J., D. R. Redman, K. L. Smith and K. W. Theil (1983). "Passive immunity to bovine rotavirus in newborn calves fed colostrum supplements from immunized or nonimmunized cows." Infect Immun 41(3): 1118-1131.

Sartain, M. J., D. L. Dick, C. D. Rithner, D. C. Crick and J. T. Belisle (2011). "Lipidomic analyses of Mycobacterium tuberculosis based on accurate mass measurements and the novel "Mtb LipidDB"." J Lipid Res 52(5): 861-872.

Sartor, R. B. (2004). Microbial influences in inflammatory bowel disease: role in pathogenesis and clinical implications. Kirsner's Inflammatory Bowel Diseases. R. B. S. Sartor, W.J. Philadelphia, Elsevier: 138 - 162.

Sartor, R. B. (2004). "Therapeutic manipulation of the enteric microflora in inflammatory bowel diseases: antibiotics, probiotics, and prebiotics." Gastroenterology 126(6): 1620-1633.

Sartor, R. B. (2005). "Does Mycobacterium avium subspecies paratuberculosis cause Crohn's disease?" Gut 54(7): 896-898.

Saxegaard, F. and I. Baess (1988). "Relationship between Mycobacterium avium, Mycobacterium paratuberculosis and "wood pigeon mycobacteria". Determinations by DNA-DNA hybridization." APMIS 96(1): 37-42.

Scandurra, G. M., G. W. de Lisle, S. M. Cavaignac, M. Young, R. P. Kawakami and D. M. Collins (2010). "Assessment of live candidate vaccines for paratuberculosis in animal models and macrophages." Infect Immun 78(3): 1383-1389.

Schneider, C. A., W. S. Rasband and K. W. Eliceiri (2012). "NIH Image to ImageJ: 25 years of image analysis." Nat Methods 9(7): 671-675.

Selby, W., P. Pavli, B. Crotty, T. Florin, G. Radford-Smith, P. Gibson, B. Mitchell, W. Connell, R. Read, M. Merrett, H. Ee, D. Hetzel and G. Antibiotics in Crohn's Disease Study (2007). "Two-year combination antibiotic therapy with clarithromycin, rifabutin, and clofazimine for Crohn's disease." Gastroenterology 132(7): 2313-2319.

Shafran, I., L. Kugler, F. A. El-Zaatari, S. A. Naser and J. Sandoval (2002). "Open clinical trial of rifabutin and clarithromycin therapy in Crohn's disease." Dig Liver Dis 34(1): 22-28.

Sigurethardottir, O. G., M. Valheim and C. M. Press (2004). "Establishment of Mycobacterium avium subsp. paratuberculosis infection in the intestine of ruminants." Adv Drug Deliv Rev 56(6): 819-834. 177

Skriwan, C., M. Fajardo, S. Hagele, M. Horn, M. Wagner, R. Michel, G. Krohne, M. Schleicher, J. Hacker and M. Steinert (2002). "Various bacterial pathogens and symbionts infect the amoeba Dictyostelium discoideum." Int J Med Microbiol 291(8): 615-624.

Smythies, L. E., M. Sellers, R. H. Clements, M. Mosteller-Barnum, G. Meng, W. H. Benjamin, J. M. Orenstein and P. D. Smith (2005). "Human intestinal macrophages display profound inflammatory anergy despite avid phagocytic and bacteriocidal activity." J Clin Invest 115(1): 66-75.

Snyder, F. F., M. S. Hershfield and J. E. Seegmiller (1978). "Cytotoxic and metabolic effects of adenosine and adenine on human lymphoblasts." Cancer Res 38(8): 2357-2362.

Sockett, D. C., T. A. Conrad, C. B. Thomas and M. T. Collins (1992). "Evaluation of four serological tests for bovine paratuberculosis." J Clin Microbiol 30(5): 1134-1139.

Souza, C., W. C. Davis, T. M. Eckstein, S. Sreevatsan and D. J. Weiss (2013). "Mannosylated lipoarabinomannans from Mycobacterium avium subsp. paratuberculosis alters the inflammatory response by bovine macrophages and suppresses killing of Mycobacterium avium subsp. avium organisms." PLoS One 8(9): e75924.

St. Jean, G. (1996). Treatment of clinical paratuberculosis in cattle. The Veterinary Clinics of North America: Food Animal Practice - Paratuberculosis (Johne's Disease). R. W. Sweeney. Philadelphia, PA, W.B. Saunders Company. 12: 417-430.

Stabel, J. R. (1998). "Johne's disease: a hidden threat." J Dairy Sci 81(1): 283-288.

Stabel, J. R. (2001). "On-farm batch pasteurization destroys Mycobacterium paratuberculosis in waste milk." J Dairy Sci 84(2): 524-527.

Stabel, J. R. and A. Lambertz (2004). "Efficacy of pasteurization conditions for the inactivation of Mycobacterium avium subsp. paratuberculosis in milk." J Food Prot 67(12): 2719-2726.

Stabel, J. R. and S. Robbe-Austerman (2011). "Early immune markers associated with Mycobacterium avium subsp. paratuberculosis infection in a neonatal calf model." Clin Vaccine Immunol 18(3): 393­ 405.

Stabel, J. R. and T. J. Stabel (1995). "Immortalization and characterization of bovine peritoneal macrophages transfected with SV40 plasmid DNA." Vet Immunol Immunopathol 45(3-4): 211-220.

Stott, G. H., D. B. Marx, B. E. Menefee and G. T. Nightengale (1979). "Colostral immunoglobulin transfer in calves II. The rate of absorption." J Dairy Sci 62(11): 1766-1773.

Sturgill-Koszycki, S., P. H. Schlesinger, P. Chakraborty, P. L. Haddix, H. L. Collins, A. K. Fok, R. D. Allen, S. L. Gluck, J. Heuser and D. G. Russell (1994). "Lack of acidification in Mycobacterium phagosomes produced by exclusion of the vesicular proton-ATPase." Science 263(5147): 678-681.

Subramoni, S., K. Agnoli, L. Eberl, S. Lewenza and P. A. Sokol (2013). "Role of Burkholderia cenocepacia afcE and afcF genes in determining lipid-metabolism-associated phenotypes." Microbiology 159(Pt 3): 603-614.

Sugawara, I., H. Yamada, H. Kaneko, S. Mizuno, K. Takeda and S. Akira (1999). "Role of interleukin­ 18 (IL-18) in mycobacterial infection in IL-18-gene-disrupted mice." Infect Immun 67(5): 2585-2589.

Sweeney, R. W., R. H. Whitlock and A. E. Rosenberger (1992). "Mycobacterium paratuberculosis cultured from milk and supramammary lymph nodes of infected asymptomatic cows." J Clin Microbiol 30(1): 166-171. 178

Taylor, T. K., C. R. Wilks and D. S. McQueen (1981). "Isolation of Mycobacterium paratuberculosis from the milk of a cow with Johne's disease." Vet Rec 109(24): 532-533.

Tenant, R. and L. E. Bermudez (2006). "Mycobacterium avium genes upregulated upon infection of Acanthamoeba castellanii demonstrate a common response to the intracellular environment." Curr Microbiol 52(2): 128-133.

Thivend, P. T., R.; Guilloteau, P. (1980). Digestive adaptation in the preruminant. Digestive physiology and metabolism in ruminants. Y. T. Ruckenbusch, P. Lancaster, M.T.P. Press Ltd.: 561­ 586.

Triccas, J. A., T. Parish, W. J. Britton and B. Gicquel (1998). "An inducible expression system permitting the efficient purification of a recombinant antigen from Mycobacterium smegmatis." FEMS Microbiol Lett 167(2): 151-156.

Tuchscherr, L., E. Medina, M. Hussain, W. Volker, V. Heitmann, S. Niemann, D. Holzinger, J. Roth, R. A. Proctor, K. Becker, G. Peters and B. Loffler (2011). "Staphylococcus aureus phenotype switching: an effective bacterial strategy to escape host immune response and establish a chronic infection." EMBO Mol Med 3(3): 129-141.

Turenne, C. Y., D. M. Collins, D. C. Alexander and M. A. Behr (2008). "Mycobacterium avium subsp. paratuberculosis and M. avium subsp. avium are independently evolved pathogenic clones of a much broader group of M. avium organisms." J Bacteriol 190(7): 2479-2487.

Vergne, I., J. Chua, S. B. Singh and V. Deretic (2004). "Cell biology of mycobacterium tuberculosis phagosome." Annu Rev Cell Dev Biol 20: 367-394.

Via, L. E., D. Deretic, R. J. Ulmer, N. S. Hibler, L. A. Huber and V. Deretic (1997). "Arrest of mycobacterial phagosome maturation is caused by a block in vesicle fusion between stages controlled by rab5 and rab7." J Biol Chem 272(20): 13326-13331.

Vilcheze, C., Y. Av-Gay, R. Attarian, Z. Liu, M. H. Hazbon, R. Colangeli, B. Chen, W. Liu, D. Alland, J. C. Sacchettini and W. R. Jacobs, Jr. (2008). "Mycothiol biosynthesis is essential for ethionamide susceptibility in Mycobacterium tuberculosis." Mol Microbiol 69(5): 1316-1329.

Weber, S. S., J. Ducry and A. Oxenius (2014). "Dissecting the Contribution of IgG Subclasses in Restricting Airway Infection with Legionella pneumophila." J Immunol 193(8): 4053-4059.

Weigoldt, M., J. Meens, K. Doll, I. Fritsch, P. Mobius, R. Goethe and G. F. Gerlach (2011). "Differential proteome analysis of Mycobacterium avium subsp. paratuberculosis grown in vitro and isolated from cases of clinical Johne's disease." Microbiology 157(Pt 2): 557-565.

Weiss, D. J., O. A. Evanson, C. de Souza and M. S. Abrahamsen (2005). "A critical role of interleukin­ 10 in the response of bovine macrophages to infection by Mycobacterium avium subsp paratuberculosis." Am J Vet Res 66(4): 721-726.

Weiss, D. J., O. A. Evanson, A. Moritz, M. Q. Deng and M. S. Abrahamsen (2002). "Differential responses of bovine macrophages to Mycobacterium avium subsp. paratuberculosis and Mycobacterium avium subsp. avium." Infect Immun 70(10): 5556-5561.

Weiss, D. J., O. A. Evanson and C. D. Souza (2006). "Mucosal immune response in cattle with subclinical Johne's disease." Vet Pathol 43(2): 127-135. 179

Wells, S. J., S. M. Godden, C. J. Lindeman and J. E. Collins (2003). "Evaluation of bacteriologic culture of individual and pooled fecal samples for detection of Mycobacterium paratuberculosis in dairy cattle herds." J Am Vet Med Assoc 223(7): 1022-1025.

Wentink, G. H., J. H. Bongers, A. A. Zeeuwen and F. H. Jaartsveld (1994). "Incidence of paratuberculosis after vaccination against M. paratuberculosis in two infected dairy herds." Zentralbl Veterinarmed B 41(7-8): 517-522.

Whan, L., I. R. Grant and M. T. Rowe (2006). "Interaction between Mycobacterium avium subsp. paratuberculosis and environmental protozoa." BMC Microbiol 6: 63.

White, C. I., R. J. Birtles, P. Wigley and P. H. Jones (2010). "Mycobacterium avium subspecies paratuberculosis in free-living amoebae isolated from fields not used for grazing." Vet Rec 166(13): 401-402.

Whitlock, R. H., S. J. Wells, R. W. Sweeney and J. Van Tiem (2000). "ELISA and fecal culture for paratuberculosis (Johne's disease): sensitivity and specificity of each method." Vet Microbiol 77(3-4): 387-398.

Whittington, R. J., I. B. Marsh, V. Saunders, I. R. Grant, R. Juste, I. A. Sevilla, E. J. Manning and R. H. Whitlock (2011). "Culture phenotypes of genomically and geographically diverse Mycobacterium avium subsp. paratuberculosis isolates from different hosts." J Clin Microbiol 49(5): 1822-1830.

Wong, D., H. Bach, J. Sun, Z. Hmama and Y. Av-Gay (2011). "Mycobacterium tuberculosis protein tyrosine phosphatase (PtpA) excludes host vacuolar-H+-ATPase to inhibit phagosome acidification." Proc Natl Acad Sci U S A 108(48): 19371-19376.

Woo, S. R., J. Sotos, A. P. Hart, R. G. Barletta and C. J. Czuprynski (2006). "Bovine monocytes and a macrophage cell line differ in their ability to phagocytose and support the intracellular survival of Mycobacterium avium subsp. paratuberculosis." Vet Immunol Immunopathol 110(1-2): 109-120.

Wren, B. and N. Dorrell (2002). Functional microbial genomics. Amsterdam ; Boston, Academic Press.

Zhao, B., M. T. Collins and C. J. Czuprynski (1997). "Effects of gamma interferon and nitric oxide on the interaction of Mycobacterium avium subsp. paratuberculosis with bovine monocytes." Infect Immun 65(5): 1761-1766.

Zimbro, M. J. P., D.A.; Miller, S.M.; Wilson, G.E.; Johnson, J.A. (2009). Difco & BBL Manual: Manual of Microbiological Culture Media. D. Becton, and Company: 356-358.

180

APPENDICES

181

Appendix 1

Establishing Caenorhabditis elegans as a model for Mycobacterium avium subspecies hominissuis infection and intestinal colonization

Jamie L. Everman1,2 , Navid R. Ziaie1 , Jessica Bechler1 , and Luiz E. Bermudez1,2

1 Department of Microbiology, College of Science, Oregon State University, Corvallis, OR 2 Department of Biomedical Sciences, College of Veterinary Medicine, Oregon State University, Corvallis, OR

Infection and Immunity – In Review

182

Abstract The nematode Caenorhabditis elegans (C. elegans) has become a model system for studying the disease interaction between pathogens and the host. To determine whether the transparent nematode could serve as a useful model for Mycobacterium avium subspecies hominissuis (MAH) infection of the intestinal tract, worms were fed MAH and assayed for the effects of the bacterial infection on the worm. It was observed during feeding that viable MAH increases in the intestinal lumen in a time dependent manner. Ingestion of MAH was deemed non-toxic to worms as MAH-fed populations have similar survival curves to those fed E. coli strain OP50. Pulse-chase analysis using E. coli strain OP50 revealed that MAH colonize the intestinal tract, as viable MAH remain within the intestine after the assay. Visualization of intestinal MAH using histology and transmission electron microscopy demonstrate that MAH localizes to the intestinal lumen, as well as establishes direct contact with intestinal epithelium. Bacterial colonization appears to have a detrimental effect on the microvilli of the intestinal epithelial cells. The MAH ΔGPL strain with a mutation in glycopeptidolipid production is deficient in binding to human epithelial cells (HEp-2), as well as deficient in its ability to bind to and colonize the intestinal tract of C. elegans as efficiently as wild-type MAH. These data indicate the C. elegans may serve as a useful model system for MAH pathogenesis and in determining the mechanisms used by MAH during infection and colonization of the intestinal epithelium.

183

Introduction Caenorhabditis elegans (C. elegans) is a ubiquitous nematode which lives in soil and feeds on bacteria. Due to its transparent nature, simple and streamlined body structure, and its defined genome, C. elegans has become a widely used model organism for studying genetics, immunology, and host-pathogen interactions. The intestinal epithelium of C. elegans and human intestinal epithelial cells are quite similar, sharing comparable morphology, structure, and function, which includes acting as a first defense against invading pathogenic bacteria. The nematode has been characterized extensively as a model for bacterial pathogenesis of medically relevant organisms, including Pseudomonas aeruginosa, Staphylococcus aureus, and Enterococcus faecalis (Garsin, Sifri et al. 2001, Irazoqui, Troemel et al. 2010). It has been demonstrated that C. elegans can feed on fast-growing strains of mycobacteria including Mycobacterium fortuitum and Mycobacterium marinum (Couillault and Ewbank 2002); however, the growth and feeding of C. elegans on slow-growing mycobacterial species has not been described. The ease of use of the nematode and the array of tools available make it a desirable candidate as a model system of other human pathogens of interest. Mycobacterium avium subsp. hominissuis (MAH) is a member of the Mycobacterium avium Complex (MAC) and is an environmental bacterium known to cause opportunistic infections in humans with immunodeficiency including those with cystic fibrosis, HIV/AIDS, and pre-existing respiratory pathology (Hawkins, Gold et al. 1986, Prince, Peterson et al. 1989). The bacterium is capable of establishing infection in both the intestinal and the respiratory epithelia where it ultimately invades and infects sub-mucosal macrophages. There, the bacterium establishes an intracellular niche where it can disseminate via lymph nodes. Current efforts are being taken to understand the mechanisms used by MAH for the transmission and colonization of the epithelial mucosa. Identification of proteins that interact with host cells can be seen as candidates for the development of novel approaches to prevent the disease. C. elegans and MAH are both abundant in the environment and there is a natural possibility of interaction between the two organisms. In this study, the 184

nematode C. elegans is characterized for the first time as a model organism to investigate MAH infection and virulence. We describe here how C. elegans feeds on MAH without significant consequences on its health or lifespan. Our results also demonstrate that MAH is able to colonize the intestinal tract of the worm in a stable non-transient manner, and that colonization of the gut results from a close association between the pathogen and the apical membrane and microvilli of the intestinal epithelial. Establishing C. elegans as a model system for MAH infection allows for the further characterization of the pathogenic mechanisms employed by MAH, and for increased progress towards therapeutic and prevention strategies.

185

Methods and Materials

Nematode propagation. Caenorhabditis elegans (C. elegans) strain N2 were obtained as a gift from the laboratory of Dr. Dee Denver at Oregon State University. Nematodes were maintained in monoxenic cultures with the addition of Escherichia coli strain OP50 and propagated on nematode growth medium (NGM) agar plates at 25°C as previously described (Brenner 1974).

Bacterial culture. E. coli strain OP50 was grown in Luria-Bertani (LB) broth overnight prior to inoculation of NGM plates. Mycobacterium avium subspecies hominissuis (MAH) strain 104 and strain A5 was grown on Middlebrook 7H10 agar supplemented with 10% w/v oleic acid-albumin-dextrose-catalase (OADC; Hardy Diagnostics; Santa Maria, CA) for 10 days at 37°C. MAH A5 ΔGPL mutant was grown on 7H10 medium described above supplemented with kanamycin sulfate (400 µg/ml). Bacterial suspensions were processed through a 23-gauge syringe, clumps allowed to settle for 10 minutes, and top 2 ml of suspension collected and used for assays.

Mammalian cell culture. Human epithelial (HEp-2; CCL-23) cells were obtained from the American Type Culture Collection (ATCC; Manassas, VA). Cells were grown in RPMI medium supplemented with 10% heat-innactivated fetal bovine serum (FBS; Gemini Bio-Products; West Sacramento, CA) in 37°C with 5% CO2.

Nematode MAH feeding assays. NGM plates were prepared as described above, or supplemented with 400 µM 5-Fluoro-2’-deoxyuridine, a DNA synthesis inhibitor which allows for the synchronization of worm cultures (FUdR; Sigma-Aldrich; St. Louis, MO) (Mitchell, Stiles et al. 1979). Agar plates with and without FUdR were seeded with 108 MAH-td104, or appropriate MAH strain as per experimental design, and inoculated with equal volume synchronized nematode cultures at the L4 stage of growth. Plates were incubated for 1, 3, or 5 days at 25°C prior to sample collection. At each timepoint, worms were collected in M9 salt solution, pelleted, and 186

anesthetized using 70% ethanol. Samples were spotted onto glass slides, and visualized on a DM4000B Leica microscope. Images were captured and analyzed using QCapture Pro7 software.

Pulse-chase feeding assay. Pulse-chase analysis was modified for mycobacterial isolation and conducted as previously described (Chou, Chiu et al. 2013). Synchronized worms at the L4 growth stage were collected in 1x M9 saline solution and washed twice. Individual NGM plates supplemented with FUdR (400 µM) were seeded with 108 MAH as per experimental design. Each plate was seeded with equivalent number of worms and incubated at 25°C. After 5 days, worms were collected in M9 saline, washed twice, and seeded onto new NGM-FUdR plates containing a lawn of E. coli OP50. After 24 hours of incubation, worms were collected in M9 saline supplemented with 25 mM levamisole hydrochloride (Sigma- Aldrich) for paralysis and prevention of expulsion or uptake of bacteria during washes. Worms were collected at 225 × g for 2 minutes, washed twice in M9 saline solution, treated with amikacin sulfate (200 µg/ml) for 2 hours at room temperature to kill all extracellular bacteria surrounding the worms, and washed twice with HBSS. For visualization, suspensions were analyzed under fluorescent microscopy. To quantify pulse-chase assays, suspensions were homogenized using a handheld motorized pestle (VWR; Radnor, PA) for 1 minute in 0.1% triton X-100 and deionized water, and samples were serially diluted and quantified.

Nematode survival curve. Synchronized nematode cultures were prepared and 30 worms were picked using a platinum wire pick and transferred to NGM plates supplemented with FUdR (400 µM) which were seeded with either 108 E. coli strain OP50 or 108 MAH strain 104. Cultures were monitored and scored for nematode survival. Worms that did not move or respond to gentle agitation by platinum wire pick were deemed dead, and worms found on the plastic sides and lid of petri plate as well as worms that burst were censored from the data counts. Survival of populations was assessed for 30 days and analyzed using a Kaplan-Meier survival analysis.

187

Histology and transmission electron microscopy. Specimen Preparation. Synchronized nematodes at the L4 growth stage were picked and seeded onto NGM agar plates supplemented with FUdR (400 µM) for 24 hours in the absence of any seeded bacteria to remove any extracellular or residual E. coli from the worm cultures. Worms were then picked and seeded onto plates containing either 108 MAH-td104 for experimentally fed samples or containing no bacteria for starved control samples and allowed to incubate for 5 days at 25°C. Histology. Worms were collected in M9 saline solution and washed twice at 50 × g for 2 minutes to remove any extracellular bacteria in suspension. Nematodes were fixed in 10% buffered formalin for 5 minutes at room temperature, washed in M9 solution, suspended in 1% low-melt agarose. Agarose-encased worms were embedded in resin, and sections mounted onto glass slides by the Veterinary Diagnostic Lab at Oregon State University. Specimens were acid-fast stained and visualized. Transmission Electron Microscopy. Worms were collected in M9 solution, washed twice, and pellet was suspended in fixative buffer of 2.5% glutaraldehyde, 1% paraformaldehyde, and 0.1 M sodium cacodylate. Worms were cut in half and incubated in fixative buffer overnight at 4°C. Specimen sections were stained, dehydrated, and infiltrated for TEM visualization by the Electron Microscopy Facility at Oregon State University as previously described (Hall 1995).

HEp-2 binding assay. Bacterial suspensions were prepared in Hank’s Balanced Salt Solution (HBSS; Corning; Corning, NY) and HEp-2 cells were infected at an MOI of 10:1. Infections were synchronized at 225 × g for 5 minutes at 4°C. Plates were incubated for 1 hour at 4°C, washed 3 times with PBS, and monolayers lysed for 15 minutes in 0.1% triton X-100 in water. Lysates were serially diluted and quantified using CFU counts.

Statistical analysis and data interpretation. Results are reported as the mean of at least 2 independent experiments ± standard error. For binding and pulse-chase assays statistical comparisons between experimental groups and control groups were determined using the Student’s t test with p<0.05 denoting statistical significance. 188

Survival curve data was analyzed using Kaplan Meier Survival Analysis. GraphPad Prism version 6.0 software was used for the construction of graphs, data interpretation, and all statistical analyses.

189

Results

C. elegans are able to feed on MAH It is well documented that C. elegans can be used as a classical infection model for a variety of pathogens (Couillault and Ewbank 2002, Balla and Troemel 2013). Previous work has determined that nematodes can be cultivated on fast growing strains of Mycobacterium fortuitum and Mycobacterium marinum (Couillault and Ewbank 2002). Our investigation first examined whether nematodes would feed on MAH. NGM plates were seeded with 108 MAH strain 104 containing the plasmid pJDC60-tdTomato which contains a tomato red fluorescent protein under a constitutive mycobacterial L5 promotor for identification of MAH within the intestinal tract of the nematode (MAH-td104). Synchronized worms in the L4 growth stage were seeded onto MAH-td104 containing plates and allowed to feed for 1, 3, and 5 days. Visualization of worms at each timepoint demonstrated that C. elegans fed on MAH-td104 in a time-dependent manner with pharynx-, grinder-, and intestinal-localized MAH-td104 increasing in intensity over time (data not shown). During the 5 day incubation, worms were able to readily feed, mate, and produce apparently healthy progeny (observation; data not shown). In order to determine if the variety of life stages played a role in uptake and feeding on MAH-td104, 5­ Fluoro-2’-deoxyuridine (FUdR) was added to MAH seeded NGM plates. The addition of FUdR inhibits DNA synthesis, thus preventing progeny to be produced and allows for the maintenance of a synchronous population during experiments and prevents the over population of a plate during longer timepoints. Synchronized cultures at the L4 stage of growth seeded onto NGM-FUdR plates demonstrated similar feeding trends on MAH-td104 during an identical feeding time course (Appendix Figure 1.1). Previous work demonstrated that fluorescence does not indicate viability, as GFP can still be detected within the intestine after bacteria are killed and digested for nutrients (Hsiao, Chen et al. 2013). To identify whether intestinal-localized MAH­ td104 was viable, worms fed for 5 days were collected, treated with levamisole to maintain the state of bacterial uptake by preventing pharyngeal uptake or expulsion 190

by the worms, and were treated with amikacin to remove residual extracellular MAH from the feeding assays. Quantification of worm lysates indicated that the intracellular MAH visualized after a 5-day feeding was viable as greater than 105 MAH were isolated from each culture of MAH-fed nematodes. From these data, we determined that C. elegans will feed on MAH if it is the sole source of bacteria during growth, and that the bacteria reside within the digestive tract after ingestion.

C. elegans survival upon MAH-td104 feeding While C. elegans can feed on a variety of bacteria in experimental studies, such pathogens, notably Salmonella and Pseudomonas are rapidly lethal (Irazoqui, Troemel et al. 2010). It is unknown whether the high fatty acid, lipid-rich composition of mycobacteria has an effect on the longevity of worms in culture. Feeding on MAH can be visualized for 5 days, as indicated by fluorescent microscopy (Appendix Figure 1.1); however, that timeframe may not be adequate to establish whether MAH is toxic or lethal to the nematodes. To determine lethality of MAH on C. elegans, a survival curve was conducted in order to compare the lifespan of C. elegans fed on standard E. coli strain OP50 cultures to the lifespan of those fed on MAH-td104. Equal numbers of synchronized worms were picked and fed on E. coli or MAH in the presence of FUdR for 35 days. Kaplan-Meier survival analysis indicated that feeding on MAH-td104 had no observable effect on worm lifespan as median survival of both the MAH-fed animals and OP50-fed animals is 18 days after the start of feeding (Appendix Figure 1.2). Furthermore, there is no significant difference between the total lifespan of OP50-fed worms (31 days) and that of MAH- fed worms (35 days).

MAH colonization of the C. elegans intestine C. elegans is capable of ingesting a wide variety of environmental bacteria for nutrients. As the nematodes can feed on MAH, and the colonization appears to remain within the intestinal tract at 5 days post-infection, we analyzed whether the presence of MAH within the intestine was transient, or if it was a longer-lived, more permanent state of colonization. To answer this question, we used pulse-chase 191

analysis during feeding. Nematodes were fed MAH-td104 for 5 days at which point they were selected and placed onto a clean NGM plate for 2 hours to remove any MAH that was on the outside of the worm bodies. Worms were then placed onto a plate containing E. coli strain OP50 and fluorescent microscopy indicated that after 24 hours of pulse-chase feeding on E. coli, red MAH-td104 was still readily visible within the intestinal lumen of the worms (Appendix Figure 1.3).

Visualization of the MAH-colonized C. elegans intestine Fluorescent microscopy analysis of MAH-fed worms indicates that the bacteria are capable of colonizing the intestine of the worms during infection. As MAH is an intracellular pathogen known for its ability to invade and survive within host cells we wanted to determine whether the red fluorescent MAH seen in our photos were in the lumen of the intestine or if the bacteria were able to invade the intestinal epithelium and cause an intracellular infection within the intestinal tract. To analyze the cross-sections of MAH-fed animals, worms were seeded onto plates and fed bacteria as described above. Worms were picked at 5-days post-infection and fixed for histology. As a control, worms were seeded onto plates in the absence of any bacteria for the same incubation time and processed as described. Acid-fast stained cross-sections indicate that MAH was capable of colonizing the lumen of the intestine at high levels during feeding and infection (Appendix Figure 1.3c – 1.3f), while starved worms indicate an absence of pink acid-fast bacilli, or any bacilli within the lumen of the intestine (Appendix Figure 1.3a and 1.3b) In order to visualize if MAH interacts with the intestinal epithelium of C. elegans following feeding, MAH-fed and starved worms were prepared and visualized using transmission electron microscopy. As a healthy and uninfected control, starved worms demonstrate an absence of bacteria within the lumen which appears compact in size (Appendix Figure 1.4a and 1.4b). An undisturbed intestinal tract can be clearly seen in the control worms as intact and tightly clustered microvilli line the apical membrane of the intestinal epithelial cells. Alternatively, MAH-fed worms present with a grossly distended lumen that is filled with colonizing MAH bacteria (Appendix Figure 1.5c – 1.5f). MAH can be seen both in the luminal space, 192

as well as in direct contact with the microvilli located on the apical membrane of the intestinal epithelium. Both at the site of contact, as well as in the vicinity of these sites, it can be appreciated that the neatly layered, tightly clustered microvilli seen in the control sections are shortened, appear damaged, and in a much looser association in the presence of MAH colonization (Appendix Figure 1.5c – 1.5f).

Binding and colonization of MAH glycopeptidolipid mutant ΔGPL/4B2 We next aimed to determine if the C. elegans model could identify MAH mutants and lead to the understanding of bacterial components responsible for the colonization of the intestinal epithelium within the nematode. The previously described MAH ΔGPL/4B2 mutant (Yamazaki, Danelishvili et al. 2006) was analyzed for binding to HEp-2 epithelial cells during a 1 hour infection (Appendix Figure 1.6a). Compared to the parental MAH A5 strain the MAH ΔGPL/4B2 mutant is deficient in its ability to bind to human HEp-2 cells during infection. Worms were allowed to feed on the wild-type and mutant MAH strains for 5 days and intestinal binding was quantified to assess the ability of worms to feed on and retain each bacterial strain (Appendix Figure 1.6b). The MAH ΔGPL/4B2 mutant is found at significantly lower levels than the wild-type MAH strain within the intestinal tract of the worms. After 5 days of feeding on each bacterial strain, nematodes underwent pulse-chase analysis with a 24-hour feeding on E. coli strain OP50 prior to intestinal MAH quantification (Appendix Figure 1.6c). The ability of the MAH ΔGPL/4B2 mutant to colonize the C. elegans intestinal epithelium is lower compared to the wild- type MAH infection as indicated by a significant decrease in the amount of bacteria localized in the intestine after pulse-chase analysis.

Discussion C. elegans has been illustrated as an extremely useful model system used to study a variety of research questions from genetics and cell biology, to pathogenesis and host-microbe interactions. The transparent nature of the nematode and simple body structure provides a simplified in vivo model to mimic physiological and pathogenic mechanisms that occur during infection. The manipulation and use of C. 193

elegans is simple, inexpensive, and there are a variety of tools available for studying the processes occurring within the host. C. elegans provides an intestinal physiology that is relevant for the study of MAH infection. The bacterium is known to invade the enterocytes of the intestinal epithelium, rather than the Peyer’s Patches and M- cells that are used by other mycobacteria (Sangari, Goodman et al. 2001). Likewise, the relatively similar physiology and epithelial make-up of the intestinal epithelium of the enterocytes provides a high degree of similarity and hypothesized interaction between bacterial pathogens such as mycobacterium and the host during infection. Here, we show that C. elegans can serve as a useful tool and model system to study the pathogenesis of MAH, specifically its interaction and infection of the intestinal epithelium. Nematodes naturally feed on environmental bacteria for nutrients and we show that the ability of populations to feed on MAH is no exception. Unlike the feeding on other pathogens such as Pseudomonas and Salmonella, the ingestion of MAH is not rapidly toxic. These observations suggest that MAH does not exude rapidly toxic components within the intestinal tract of the worm during infection resulting in a longer-lived infection within the C. elegans host. It is described that M. avium first colonizes the host intestinal epithelium prior to the observation of systemic bacterial infection (Roth, Owen et al. 1985). Our histology images and pulse-chase data demonstrate that MAH are found within the lumen of the intestinal tract over the 5-day infection, quickly increasing in number and colonizing in a stable, non-transient manner. These data are supported by TEM images of MAH-infected worms which indicate that MAH has a direct interaction during colonization with the intestinal epithelium. This interaction results in damage to the villi on the epithelial cells. As our infection only progressed until 5-days post- feeding, it would be interesting to analyze whether longer infection of C. elegans with MAH would allow for progression of the colonization to an invasive infection where MAH may be able to invade the mucosal tissue within the worm gut. These observations suggest that MAH is able to colonize the intestinal epithelium to potentially establish a more long-term and chronic infection within the host. Ultimately, we aim to understand the bacterial mechanisms and factors that are responsible for MAH infection within the host. The identification of such factors 194

used to initiate the disease could be useful in developing therapeutics to prevent the infection from establishing or to treat infections that are in the early stages of colonization. By establishing C. elegans as a useful model system for MAH colonization, we have identified a new model system that may provide a more realistic host scenario for the colonization capabilities of MAH and MAH mutants of interest for further study. Furthermore, our data indicate that the MAH ΔGPL/4B2 mutant, which demonstrates a decreased ability to form biofilm (Yamazaki, Danelishvili et al. 2006), and a decreased ability to bind to human epithelial HEp-2 cells, also shows a decreased ability in binding to and colonizing the intestinal tract of C. elegans.

195

Acknowledgements We would like to thank the lab of Dr. Dee Denver at Oregon State University for supplying C. elegans, E. coli strain OP50 stocks, and technical advice on worm techniques. We would like to thank Teresa Sawyer for technical assistance with preparation and imaging of TEM samples and the Oregon State University Veterinary Medicine Histology lab for technical assistance. We acknowledge that this TEM material is based upon work supported by the National Science Foundation via the Major Research Instrumentation (MRI) Program under Grant No. 1040588. We gratefully acknowledge financial support for acquisition of the TEM instrument from the Murdock Charitable Trust and the Oregon Nanoscience and Microtechnologies Institute (ONAMI).

196

Appendix Figure 1.1: C. elegans feed on MAH C. elegans were seeded onto NGM plates supplemented with FUdR (400 μM) and seeded with MAH expressing a fluorescent tomato red marker. Worms were allowed to feed for 1, 3, or 5 days at which time worms were collected, washed, and mounted on glass slides for microscopic observation. Images are representative of 20 worms visualized per experiment and independently repeated 5 times. All images are shown at 400x magnification.

197

1 0 0 E . c o li O P 5 0 M A H - td 1 0 4 7 5

5 0

2 5 Percent survival

0 0 1 0 2 0 3 0 4 0 Time (days)

Appendix Figure 1.2: MAH-td104 does not affect median or total lifespan of C. elegans Worms were picked and incubated on NGM agar to remove external bacteria for 3 hours. Thirty worms were individually placed onto NGM plates supplemented with FUdR (400 µM) and seeded with either 108 E. coli strain OP50 or MAH-td104. Worms were scored every 2 days for survival and worms that ruptured or crawled up the sides of the plate were censored and removed from the study. Kaplan-Meier statistics were used to construct and analyze growth characteristics. Data represents survival from one experiment and is representative of 2 independently completed experiments.

198

Appendix Figure 1.3: MAH colonization persists after pulse-chase with E. coli strain OP50 C. elegans were placed onto NGM-FUdR (400 μg/ml) plates and seeded with 108 MAH-td104 and allowed to feed for 5 days. Worms were collected and moved to a new plate to remove extracellular bacteria. Worms were then transferred to a plate seeded with E. coli strain OP50, and allowed to feed for 24 hours. Nematodes were collected and mounted onto glass slides for microscopic observation. Images are representative of 20 worms visualized per experiment and independently repeated 3 times. All images are shown at 400x magnification.

199

Appendix Figure 1.3: MAH colonize lumen of C. elegans intestinal tract Worms were seeded onto NGM plates supplemented with FUdR (400 µM) for 5 days and samples were fixed, set into agarose blocks and paraffin, sectioned, and acid-fast stained. Starved worms were seeded onto plates in the absence of bacteria (A and B) onto plates containing MAH (D – F). Acid-fast positive bacilli within the intestinal space are indicated by black/white arrows. Non-specific staining of fat deposits by carbol-fuchsin is indicated by white asterisks (*). Images are representative of 10 worms sectioned and analyzed per treatment. All images are shown at 630x magnification.

200

Appendix Figure 1.5: Transmission electron microscopy of MAH-colonized C. elegans Worms were seeded onto NGM plates supplemented with FUdR (400 µM) for 5 days and samples were fixed, processed, and visualized by transmission electron microscopy on an FEI Titan 80-200 microscope. Starved worms were seeded onto plates in the absence of bacteria (A and B) or onto plates seeded with 108 MAH (C – F). Panel F illustrates a magnified view taken from panel E (dotted line) and shows disruption on the microvilli (bracket). Key: white arrowheads - MAH, lumen – 201

luminal space, mv – microvilli, am – apical epithelial membrane, bm – basal epithelial membrane. Panels A – E scale bar is 2 µm, Panel F scale bar is 500 nm.

202

A 2 5 **

2 0 HEp-2 Binding

1 5

1 0 % B in d in g

5

0 WT MAH 4B2/GPL

S t r a in B 1 .0 1 0 7 **

1 .0 1 0 6 Feeding Assay

5 ( C F U /m1 l) .0 1 0 MAH Recovered

1 .0 1 0 4 M A H A 5 4 B 2 / GPL

S t r a in

C 1 .0 1 0 7 ****

1 .0 1 0 6 Pulse C hase

1 .0 1 0 5 ( C F U /m l) 1 .0 1 0 4 MAH Recovered

1 .0 1 0 3 MAH A5 4B2/GPL

S t r a in

Appendix Figure 1.6: Binding of HEp-2 cells and colonization of C. elegans by MAH ΔGPL/4B2 mutant The MAH ΔGPL/4B2 mutant and the parental strain MAH A5 were used for HEp-2 binding assays (A). HEp-2 cells were infected at an MOI of 10:1 with each strain and binding was allowed to progress for 1 hour at 4°C. Wells were lysed and quantified for percent of bound bacteria to the surface of epithelial cells during assay. Equivalent numbers of C. elegans were seeded onto NGM-FUdR (400 µM) plates containing 108 of each mutant strain of MAH, and allowed to feed at 25°C for 5 days. Worms were collected, washed with levamisole (25 mM), treated with amikacin (200 203

µg/ml), and lysed for quantification of intracellular bacteria. Worms were homogenized immediately after MAH feeding to determine binding ability after feeding (B). Analysis for colonization using pulse-chase analysis was conducted by transferring worms to NGM-FUdR plates seeded with E. coli strain OP50 for 24 hours prior to homogenization and quantification (C). Data represent the mean ± SEM of 2 independent experiments each performed in triplicate (**p-value<0.01, ****p-value<0.0001 as determined by Student’s t-test).

204

References

Balla, K. M. and E. R. Troemel (2013). "Caenorhabditis elegans as a model for intracellular pathogen infection." Cell Microbiol 15(8): 1313-1322.

Brenner, S. (1974). "The genetics of Caenorhabditis elegans." Genetics 77(1): 71-94.

Chou, T. C., H. C. Chiu, C. J. Kuo, C. M. Wu, W. J. Syu, W. T. Chiu and C. S. Chen (2013). "Enterohaemorrhagic Escherichia coli O157:H7 Shiga-like toxin 1 is required for full pathogenicity and activation of the p38 mitogen-activated protein kinase pathway in Caenorhabditis elegans." Cell Microbiol 15(1): 82-97.

Couillault, C. and J. J. Ewbank (2002). "Diverse bacteria are pathogens of Caenorhabditis elegans." Infect Immun 70(8): 4705-4707.

Garsin, D. A., C. D. Sifri, E. Mylonakis, X. Qin, K. V. Singh, B. E. Murray, S. B. Calderwood and F. M. Ausubel (2001). "A simple model host for identifying Gram- positive virulence factors." Proc Natl Acad Sci U S A 98(19): 10892-10897.

Hall, D. H. (1995). "Electron microscopy and three-dimensional image reconstruction." Methods Cell Biol 48: 395-436.

Hawkins, C. C., J. W. Gold, E. Whimbey, T. E. Kiehn, P. Brannon, R. Cammarata, A. E. Brown and D. Armstrong (1986). "Mycobacterium avium complex infections in patients with the acquired immunodeficiency syndrome." Ann Intern Med 105(2): 184-188.

Hsiao, J. Y., C. Y. Chen, M. J. Yang and H. C. Ho (2013). "Live and dead GFP- tagged bacteria showed indistinguishable fluorescence in Caenorhabditis elegans gut." J Microbiol 51(3): 367-372.

Irazoqui, J. E., E. R. Troemel, R. L. Feinbaum, L. G. Luhachack, B. O. Cezairliyan and F. M. Ausubel (2010). "Distinct pathogenesis and host responses during infection of C. elegans by P. aeruginosa and S. aureus." PLoS Pathog 6: e1000982.

Mitchell, D. H., J. W. Stiles, J. Santelli and D. R. Sanadi (1979). "Synchronous growth and aging of Caenorhabditis elegans in the presence of fluorodeoxyuridine." J Gerontol 34(1): 28-36.

Prince, D. S., D. D. Peterson, R. M. Steiner, J. E. Gottlieb, R. Scott, H. L. Israel, W. G. Figueroa and J. E. Fish (1989). "Infection with Mycobacterium avium complex in patients without predisposing conditions." N Engl J Med 321(13): 863-868.

Roth, R. I., R. L. Owen, D. F. Keren and P. A. Volberding (1985). "Intestinal infection with Mycobacterium avium in acquired immune deficiency syndrome (AIDS). Histological and clinical comparison with Whipple's disease." Dig Dis Sci 30(5): 497-504. 205

Sangari, F. J., J. Goodman, M. Petrofsky, P. Kolonoski and L. E. Bermudez (2001). "Mycobacterium avium invades the intestinal mucosa primarily by interacting with enterocytes." Infect Immun 69(3): 1515-1520.

Yamazaki, Y., L. Danelishvili, M. Wu, E. Hidaka, T. Katsuyama, B. Stang, M. Petrofsky, R. Bildfell and L. E. Bermudez (2006). "The ability to form biofilm influences Mycobacterium avium invasion and translocation of bronchial epithelial cells." Cell Microbiol 8(5): 806-814.

Yamazaki, Y., L. Danelishvili, M. Wu, M. Macnab and L. E. Bermudez (2006). "Mycobacterium avium genes associated with the ability to form a biofilm." Appl Environ Microbiol 72(1): 819-825. 206

Appendix 2: Abstracts of Additional Publications

Appendix 2.1 Danelishvili L, JL Everman, MJ McNamara, and LE Bermudez (2001) Inhibition of the plasma-membrane-associated serine protease cathepsin G by Mycobacterium tuberculosis Rv3364c suppresses caspase-1 and pyroptosis in macrophages. Front. Micro. 1-14 vol 2

Title: Inhibition of the plasma-membrane-associated serine protease cathepsin G by Mycobacterium tuberculosis Rv3364c suppresses caspase-1 and pyroptosis in Macrophages

Authors: Lia Danelishvili, Jamie L. Everman, Michael J. McNamara, and Luiz E. Bermudez

Journal: Frontiers in Microbiology; 2011 pg 1-14 Vol 2

Abstract:

Tuberculosis is a disease associated with the infection of a great part of the world's population and is responsible for the death of two to three million people annually. Mycobacterium tuberculosis infects macrophages and subverts its mechanisms of killing. The pathogen suppresses macrophage apoptosis by many different mechanisms. We describe that, upon uptake by macrophages, M. tuberculosis overexpresses an operon Rv3361c-Rv3365c and secretes Rv3364c. The Rv3365c knockout strain is deficient in apoptosis inhibition. The Rv3364c protein binds to the serine protease cathepsin G on the membrane, inhibiting its enzymatic activity and the downstream activation of caspase-1-dependent apoptosis. In summary, M. tuberculosis prevents macrophage pyroptosis by a novel mechanism involving cytoplasmic surveillance proteins.

207

Appendix 2.2 Bannantine JP, Everman JL, Rose SJ, Babrak L, Katani R, Barletta RG, Talaat AM, Gröhn YT, Chang Y-F, Kapur V and Bermudez LE (2014) Evaluation of eight live attenuated vaccine candidates for protection against challenge with virulent Mycobacterium avium subspecies paratuberculosis in mice. Front. Cell. Infect. Microbiol. 4:88. doi: 10.3389/fcimb.2014.00088

Title: Evaluation of eight live attenuated vaccine candidates for protection against challenge with virulent Mycobacterium avium subspecies paratuberculosis in mice

Authors: John P. Bannantine, Jamie L. Everman, Sasha J. Rose, Lmar Babrak, Robab Katani, Raúl G. Barletta, Adel M. Talaat, Yrjö T. Gröhn, Yung-Fu Chang, Vivek Kapur, and Luiz E. Bermudez

Journal: Frontiers Cellular and Infection Microbiology; 2014 doi: 10.3389/fcimb.2014.00088

Abstract: Johne's disease is caused by Mycobacterium avium subsp. paratuberculosis (MAP), which results in serious economic losses worldwide in farmed livestock such as cattle, sheep, and goats. To control this disease, an effective vaccine with minimal adverse effects is needed. In order to identify a live vaccine for Johne's disease, we evaluated eight attenuated mutant strains of MAP using a C57BL/6 mouse model. The persistence of the vaccine candidates was measured at 6, 12, and 18 weeks post vaccination. Only strains 320, 321, and 329 colonized both the liver and spleens up until the 12-week time point. The remaining five mutants showed no survival in those tissues, indicating their complete attenuation in the mouse model. The candidate vaccine strains demonstrated different levels of protection based on colonization of the challenge strain in liver and spleen tissues at 12 and 18 weeks post vaccination. Based on total MAP burden in both tissues at both time points, strain 315 (MAP1566::Tn5370) was the most protective whereas strain 318 (intergenic Tn5367 insertion between MAP0282c and MAP0283c) had the most colonization. Mice vaccinated with an undiluted commercial vaccine preparation displayed the highest bacterial burden as well as enlarged spleens indicative of a strong infection. Selected vaccine strains that showed promise in the mouse model were moved forward into a 208

goat challenge model. The results suggest that the mouse trial, as conducted, may have a relatively poor predictive value for protection in a ruminant host such as goats.

209

Appendix 2.3 Danelishvili L, Babrak L, Rose SJ, Everman J, Bermudez LE. 2014.Mycobacterium tuberculosis alters the metalloprotease activity of the COP9 signalosome. mBio 5(4):e01278-14. doi:10.1128/mBio.01278-14.

Title: Mycobacterium tuberculosis alters the metalloproteae activity of the COP9 Signalosome.

Authors: Lia Danelishvili, Lmar Babrak, Sasha J. Rose, Jamie Everman, and Luiz E. Bermudez

Journal: mBio; 2014 doi:10.1128/mBio.01278-14

Abstract:

Inhibition of apoptotic death of macrophages by Mycobacterium tuberculosis represents an important mechanism of virulence that results in pathogen survival both in vitro and in vivo. To identify M. tuberculosis virulence determinants involved in the modulation of apoptosis, we previously screened a transposon bank of mutants in human macrophages, and an M. tuberculosis clone with a nonfunctional Rv3354 gene was identified as incompetent to suppress apoptosis. Here, we show that the Rv3354 gene encodes a protein kinase that is secreted within mononuclear phagocytic cells and is required for M. tuberculosis virulence. The Rv3354 effector targets the metalloprotease (JAMM) domain within subunit 5 of the COP9 signalosome (CSN5), resulting in suppression of apoptosis and in the destabilization of CSN function and regulatory cullin-RING ubiquitin E3 enzymatic activity. Our observation suggests that alteration of the metalloprotease activity of CSN by Rv3354 possibly prevents the ubiquitin-dependent proteolysis of M. tuberculosis-secreted proteins.