ROLES OF MICROBIALLY-PRODUCED METABOLITES IN REGULATING HOST GUT HEALTH A Dissertation Presented to the Faculty of the Graduate School of Cornell University In Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy by Samantha Ann Marie Scott December 2021 © 2021 Samantha Ann Marie Scott ROLES OF MICROBIALLY-PRODUCED METABOLITES IN REGULATING HOST GUT HEALTH Samantha Ann Marie Scott, Ph.D. Cornell University 2021 Abstract: The human gut microbiome includes trillions of bacteria colonizing the intestinal tract. These bacteria exist in close contact with the intestinal epithelial barrier, composed of a single layer of cells separating the lumen from the immune-cell rich lamina propria and the rest of the body. It has become evident that inflammation and perturbations to immune homeostasis associated with diseases such as inflammatory bowel diseases (IBDs), as well as disruptions in our gut due to infection with pathogenic microbes, can lead to a myriad of functional changes in terms of intestinal permeability and physiology. We have determined that three gut microbially-produced, dietary-derived metabolites activate host receptors and affect downstream host proteins to protect the host during inflammation associated with IBDs as well as during infection with the human pathogen enterohemorrhagic Escherichia coli (EHEC) O157:H7 and the murine pathogen Citrobacter rodentium. These metabolites can improve intestinal barrier function and mediate colonization resistance against pathogens. We envision that our findings hold great promise for potential prophylactic or therapeutic treatments for host intestinal ailments. BIOGRAPHICAL SKETCH Samantha Ann Marie Scott was raised in Falls, Pennsylvania. She attended the University of Scranton in Scranton, Pennsylvania where she earned her B.S. degree in Biochemistry, Cell and Molecular Biology in 2015. She spent time volunteering, working as a microbiology lab teaching assistant, as well as conducting research during her undergraduate years. Samantha’s undergraduate research with Dr. Michael Sulzinski focused on determining the distribution of the human, plant, and fungal pathogen Burkholderia gladioli on mushrooms and onions grown in the United States. She also completed a summer research internship in the lab of Dr. Gregory Yochum at the Pennsylvania State College of Medicine working to design and validate transcription activator like effectors (TALEs) to repress axis inhibition protein 2 (AXIN2) gene expression. In 2015, Samantha began her graduate studies at Cornell University in the Field of Microbiology. She joined the lab of Dr. Pamela Chang in 2016. Her thesis work focused on understanding the effect and mechanism of action of microbially-produced metabolites in models of inflammatory bowel diseases and enteric infections. She also participated in several teaching and outreach opportunities at Cornell and in the Ithaca area during her years in graduate school. iii To all those who have supported me on this journey iv ACKNOWLEDGMENTS First and foremost, I would like to thank my advisor and mentor Dr. Pamela Chang for her patience, guidance, and support throughout my Ph.D. research. Without her, my work would not be possible. She saw potential in me and helped to shape me into the scientist I am today. I would also like to thank my committee members Dr. David Russell and Dr. Yuxin Mao for their helpful suggestions and encouragement over the years. Many thanks to past and present members of the Chang lab, Bibudha Parasar, Gael Nicolas, Jingjing Fu, Lin Han, Amanda Conwell, Laura Frazier, Kien Malarney, and Tracy Zheng for their experimental aid, help with brainstorming, and fun times both in and out of the lab. I would also like to thank past and present members of the Baskin lab, the Marquis lab, and Dr. Rod Getchell for their experimental advice and generosity over the years. I thank the members of the Department of Microbiology and Immunology and the Field of Microbiology for their support of my Ph.D. research. I would also be remiss if I did not thank members of the CARE staff for their time, expertise, and patience in training me to work with mice and caring for our mouse colonies. Outside of the lab, I am very fortunate to have a wonderful support system, and there are too many people to thank each one. Tremendous thanks to my husband Anthony Caselli for his unwavering love and support. I cannot thank him enough for the strength he has given me. Thanks to my fur babies for their companionship and love – my angel Maggie, Roscoe Lee, and Owen West. Thanks to my parents Randy and Patricia Scott for their help and guidance and the many opportunities I have had in life that allowed me to pursue a Ph.D. Thanks to my Ithaca friends and Scranton pals for always listening to me ramble about my research and for helping me take my mind off it when needed. v TABLE OF CONTENTS Biographical sketch i i i Dedication i v Acknowledgements v Table of contents v i List of figures viii Chapter 1: Introduction 1 1.1 The human gut microbiome 1 1.1.1 Gut bacteria produce metabolites, mediate immune responses, and colonization 1 resistance 1.1.2 Further elucidation of how microbially-produced metabolites influence the host 4 immune response and colonization resistance is needed 1.2 Inflammatory bowel diseases (IBDs) 5 1.2.1 The intestinal epithelial barrier is disrupted during IBD 5 1.2.2 Microbially-derived metabolites play a key role in ameliorating IBD 7 1.2.3 Further elucidation of the roles of microbially-derived metabolites in ameliorating 12 IBD is needed 1.3 Enterohemorrhagic Escherichia coli (EHEC) O157:H7 13 1.3.1 EHEC injects bacterial effector proteins, disrupts host proteins during infection, 13 and induces inflammation 1.3.2 Murine pathogen Citrobacter rodentium serves as a physiologically relevant in 15 vivo model of EHEC infection 1.3.3 Microbially-derived metabolites can mediate resistance to EHEC and C. 16 rodentium infections 1.3.4 Further elucidation of the roles of microbially-derived metabolites during EHEC 19 and C. rodentium infection is needed 1.4 References 20 Chapter 2: Microbial tryptophan metabolites regulate gut barrier function via the 36 aryl hydrocarbon receptor 2.1 Abstract 36 2.2 Introduction 36 2.3 Materials and Methods 40 2.4 Results 53 2.4.1 Trp Feeding Ameliorates DSS Colitis in Mice and Is Dependent on the Gut 53 Microbiota. 2.4.2 Specific Trp Metabolites Improve Intestinal Barrier Function In Vitro. 56 2.4.3 Trp Metabolites Improve Disease Outcomes in DSS Colitis in Mice. 62 2.4.4 Trp Metabolites Are Ligands of the AhR. 75 2.4.5 Effect of Trp Metabolites Is Partially Dependent on AhR In Vivo. 77 2.4.6 Trp Metabolites Inhibit Activation of Myosin IIA In Vivo in an AhR-Dependent 88 Manner 2.4.7 Trp Metabolites Inhibit Activation of Ezrin In Vitro. 90 2.4.8 Trp Metabolites Inhibit Activation of Ezrin In Vivo in an AhR-Dependent 95 vi Manner. 2.5 Discussion 97 2.6 References 103 Chapter 3: Dopamine receptor D2 confers colonization resistance via gut 109 microbial metabolites 3.1 Abstract 109 3.2 Introduction 109 3.3 Materials and Methods 110 3.4 Results 125 3.4.1 Trp Feeding Ameliorates C. rodentium Infection in Mice and is Dependent on the 125 Gut Microbiota. 3.4.2 Specific Trp Metabolites Mediate Colonization Resistance Against C. rodentium 127 Infection in Mice. 3.4.3 Trp Metabolites Exert Modest Effects on C. rodentium and EHEC Growth and 136 Virulence. 3.4.4 Trp Metabolites are Ligands of DRD2. 137 3.4.5 Effect of Trp Metabolites is Dependent on DRD2 Expressed on Intestinal 141 Epithelial Cells in Vivo. 3.4.6 Trp Metabolites Inhibit C. rodentium and EHEC Pedestal Formation in a DRD2- 147 Dependent Manner. 3.5 Discussion 152 3.6 References 154 Chapter 4: Conclusion 158 vii LIST OF FIGURES Figure Page 2.1 Dietary Trp ameliorates a mouse model of colitis in a microbiome- 55 dependent manner. 2.2 Microbial tryptophan (Trp) metabolites reduce disassembly of tight (TJ) 56 and adherens junction (AJ) proteins within the intestines using a mouse model of colitis. 2.3 Tryptophan metabolites indole-3-aldehyde (I3A), indole-3-pyruvate 58 (IPyA), and indole-3-ethanol (IEt) decrease epithelial permeability. 2.4 Tryptophan metabolites indole-3-aldehyde (I3A), indole-3-pyruvate 59 (IPyA), and indole-3-ethanol (IEt) decrease epithelial permeability. 2.5 Tryptophan metabolites indole-3-aldehyde (I3A), indole-3-pyruvate 60 (IPyA), and indole-3-ethanol (IEt) do not change levels of mRNA transcripts or protein expression of TJ and AJ proteins. 2.6 Chemical structures of tryptophan (Trp) metabolites that were tested and 61 did not affect epithelial permeability in this study. 2.7 Additional tryptophan metabolites do not affect epithelial permeability. 62 2.8 LC-MS quantification of tryptophan (Trp) metabolites during Trp-rich 64 diet in mouse model of colitis. 2.9 Tryptophan metabolites I3A, IPyA, and IEt ameliorate a mouse model of 66 colitis. 2.10 Tryptophan metabolite indole-3-aldehyde (I3A) prevents disruption of TJ 67 and AJ proteins in a mouse model of colitis. 2.11 Tryptophan metabolite indole-3-pyruvate (IPyA) prevents disruption of 68 TJ and AJ proteins in a mouse model of colitis. 2.12 Tryptophan metabolite indole-3-ethanol (IEt) prevents disruption of TJ 69 and AJ proteins in a mouse model of colitis. 2.13 Tryptophan metabolites indole-3-aldehyde (I3A), indole-3-pyruvate 70 (IPyA), and indole-3-ethanol (IEt) inhibit TNFR1 signaling. 2.14 Tryptophan metabolites indole-3-aldehyde (I3A), indole-3-pyruvate 70 (IPyA), and indole-3-ethanol (IEt) induce IL-10R in DSS colitis. 2.15 LC-MS quantification of Trp metabolites after oral gavage of I3A during 71 mouse model of colitis. 2.16 LC-MS quantification of Trp metabolites after oral gavage of IPyA during 72 mouse model of colitis. 2.17 LC-MS quantification of Trp metabolites after oral gavage of IEt during 74 mouse model of colitis. 2.18 Tryptophan metabolites I3A, IPyA, and IEt are aryl hydrocarbon receptor 76 (AhR) ligands, and the activity of IPyA can be blocked with an AhR inhibitor. 2.19 Effect of a high-Trp diet in a mouse model of colitis is dependent on AhR. 78 2.20 Effect of high tryptophan (Trp) diet in decreasing TJ and AJ disassembly 79 in a mouse model of colitis is dependent on the aryl hydrocarbon receptor (AhR). viii 2.21 Effect of high tryptophan (Trp) diet in mouse model of colitis does not 81 differ in AhR wildtype versus heterozygote mice. 2.22 Effect of tryptophan metabolite IEt in mouse model of colitis is dependent 82 on AhR. 2.23 Effect of tryptophan metabolite IEt in decreasing TJ and AJ disassembly 83 in mouse model of colitis is dependent on AhR. 2.24 Effect of tryptophan metabolite IPyA in mouse model of colitis is 84 dependent on AhR. 2.25 Effect of tryptophan metabolite I3A in mouse model of colitis is 86 dependent on the aryl hydrocarbon receptor (AhR). 2.26 Trp feeding and metabolites prevent myosin IIA activation during mouse 89 model of colitis, which is AhR dependent. 2.27 Activation of the actin regulatory protein ezrin increases epithelial 91 permeability caused by TNFα, and metabolites prevent increased permeability. 2.28 Activation of the actin regulatory protein ezrin increases sinuous 92 phenotype caused by TNFα, which is prevented by Trp metabolites. 2.29 Activation of the actin regulatory protein ezrin increases mislocalization 93 of AJ proteins caused by TNFα, which is prevented by Trp metabolites. 2.30 Activation of the actin regulatory protein ezrin increases sinuous 94 phenotype and mislocalization of TJ and AJ proteins caused by TNFα, which is prevented by Trp metabolites. 2.31 Trp feeding and metabolites prevent ezrin activation in mouse model of 96 colitis, which is AhR dependent. 2.32 Proposed model of Trp-derived metabolite function in regulating gut 102 epithelial permeability. 3.1 Dietary tryptophan (Trp) protects against a mouse model of EHEC 126 infection using Citrobacter rodentium, strain DBS100. 3.2 Tryptophan (Trp) metabolites prevent weight loss in a mouse model of 127 EHEC infection using Citrobacter rodentium, strain DBS100. 3.3 LC–MS quantification of tryptophan (Trp) metabolites during Trp-rich 128 diet in mouse model of EHEC infection with C. rodentium, strain DBS100. 3.4 The Trp metabolites I3A, IEt, and IPyA protect against Citrobacter 130 rodentium infection in mice. 3.5 LC–MS quantification of tryptophan (Trp) metabolites during I3A 131 treatment in mouse model of EHEC infection with C. rodentium, strain DBS100. 3.6 LC–MS quantification of tryptophan (Trp) metabolites during IPyA 133 treatment in mouse model of EHEC infection with C. rodentium, strain DBS100. 3.7 LC–MS quantification of tryptophan (Trp) metabolites during IEt 135 treatment in mouse model of EHEC infection with C. rodentium, strain DBS100. 3.8 Tryptophan metabolites have modest effects on C. rodentium and EHEC 137 growth in vitro. ix 3.9 Effects of I3A, IPyA, and IEt depend on dopamine receptor D2 (DRD2). 138 3.10 I3A, IPyA, and IEt are ligands of dopamine receptor D2 (DRD2), which 139 signals via Gαi. 3.11 Dopamine protects against a mouse model of EHEC infection using 140 Citrobacter rodentium, strain DBS100. 3.12 Effects of the Trp diet and metabolites in protecting against Citrobacter 142 rodentium infection depend on dopamine receptor D2 (DRD2) in intestinal epithelial cells (IECs). 3.13 Effects of the tryptophan (Trp) metabolites I3A and IPyA in protecting 143 against Citrobacter rodentium infection depend on dopamine receptor D2 (DRD2) in intestinal epithelial cells (IECs). 3.14 Effects of the tryptophan (Trp) diet and metabolites in preventing weight 144 loss during Citrobacter rodentium infection depend on dopamine receptor D2 (DRD2) in intestinal epithelial cells (IECs). 3.15 Effects of the tryptophan (Trp) diet in protecting against Citrobacter 145 rodentium infection do not depend on dopamine receptor D2 (DRD2) in macrophages. 3.16 Effects of the tryptophan (Trp) diet in protecting against Citrobacter 146 rodentium infection do not depend on dopamine receptor D2 (DRD2) in dendritic cells. 3.17 Effects of the tryptophan (Trp) diet in protecting against Citrobacter 147 rodentium infection do not depend on dopamine receptor D2 (DRD2) in CD4+ T cells. 3.18 Tryptophan (Trp) metabolites decrease actin pedestal formation via DRD2 148 during Citrobacter rodentium infection. 3.19 Tryptophan (Trp) diet and metabolites decrease levels of the actin 150 regulatory proteins N-WASP and IRTKS via dopamine receptor D2 (DRD2) during Citrobacter rodentium infection. 3.20 Trp metabolites decrease EHEC actin pedestal formation via DRD2. 151 3.21 Proposed model of tryptophan (Trp)-derived metabolite conferral of 153 colonization resistance against attaching and effacing pathogens (AE, e.g., EHEC O157:H7 and Citrobacter rodentium) via the dopamine receptor D2 (DRD2). x Chapter 1: Introduction 1.1 The human gut microbiome Over 1000 diverse bacterial species, along with archaea, viruses, fungi, and eukaryotes, inhabit the human gastrointestinal tract. Collectively termed the gut microbiome, trillions of these bacteria occupy the gut and vary from person to person based on factors including genetics, environment, and diet (1,2,3). The density and composition of the microbiota varies across the length of the digestive tract based on gradients in pH and oxygen, the availability of nutrients, differences in the mucus layer lining the intestine and host secretions including antimicrobial peptides (AMPs) and bile acids. The acidic stomach typically harbors the most acidotolerant, aerotolerant, and fewest number of bacteria, with approximately 101 bacterial cells per gram of contents. This number increases throughout the small intestine from duodenum to ileum, and peaks in the colon at about 1012 bacterial cells per gram of contents (4). The gut microbiome, once referred to as a “forgotten organ,” contributes to human health in a myriad of ways, from digestion and production of important vitamins and metabolites to the development of the immune system and prevention of pathogen invasion (4). However, due to the close association between the microbiota and the host, the delicate balance of intestinal homeostasis may be disrupted, and the gut can become a site for significant inflammatory diseases including inflammatory bowel diseases (IBDs) and infection with pathogenic microorganisms (1,5). In recent years, the gut microbiome has become increasingly studied to elucidate the ways in which commensal bacteria and their by-products may contribute to the amelioration of these diseases and infections in an effort to develop potential prophylactic or therapeutic tools. We seek to contribute to this growing body of knowledge. 1.1.1 Gut bacteria produce metabolites, mediate immune responses, and colonization resistance 1 Members of the gut microbiome contribute to human nutrition through digestion of complex or otherwise undigestible substrates. While the human body can break down protein, lipids, simple sugars and starches, many bacteria are specialized to digest other sources of energy (6, 7). Bacteroides species have been found to break down xyloglucans found in lettuce and onions, and members of Firmicutes, Actinobacteria and Verrucomicrobia can efficiently break down plant cell walls (7, 8, 9). The ability to break down such materials is not only key to host nutrition but is also significant for the by-products that may be released in the process. These microbially- produced metabolites are sometimes referred to as postbiotics. Within the human body, vitamins serve as essential coenzymes for many metabolic processes. Members of the gut microbiome, largely in the colon, can produce vitamins such as K2 and many B-vitamins either de novo or from exogenous sources (10). Various bacterial species in the gut have been found to produce neurotransmitters such as 5-hydroxytryptamine (5-HT, serotonin), norepinephrine (NE), gamma-aminobutyric acid (GABA), and dopamine (DA) (11- 17). These microbially-produced neurotransmitters not only function in the brain, but also in the gut itself, participating in what is known as the gut-brain axis. Dopamine receptors, for example, are present in the gastrointestinal tract, and dopamine has been found to play roles within the intestine, including in promoting absorption (16, 18, 19). Gut microbes also produce short chain fatty acids (SCFAs) like acetate, propionate, and butyrate from complex carbohydrates. These SCFAs are widely studied postbiotics that have been found to affect the immune system, regulation of intestinal epithelial barrier function and play roles in colonization resistance against invading pathogenic bacteria (20, 21, 22). Similar functions have been attributed to secondary bile acids, which colonic gut microbes produce by converting primary bile acids that have been produced in the liver (22-25). Of increasing importance in the 2 field are amino acid derivatives. Members of the gut microbiome can break down amino acids such as tryptophan (26). Microbially-produced tryptophan (Trp) metabolites include kynurenines and indole and its derivatives, which have also been found to affect immune responses in the gut, intestinal epithelial barrier function, and colonization resistance (22, 23, 27, 28). The intestinal epithelium of the colon is composed of a single layer of primarily simple columnar epithelial cells. At the base of crypts, pluripotent intestinal epithelial stem cells allow for the proliferation and differentiation of additional cell types that play key roles in the gut (5, 29- 31). Given the close association between the microbiota and this single layer of cells that separates intestinal contents from the immune cell-rich lamina propria beneath, the immune system must develop tolerance to the commensal microbes that inhabit the healthy human gut. The mucus layer covering the intestinal epithelium serves as the first barrier against bacteria and other antigens, and largely prevents direct contact between luminal bacteria and intestinal epithelial cells (IECs) (29). Goblet cells and Paneth cells secrete mucins and AMPs, which limit bacterial populations by killing and inhibiting the growth of bacterial cells. Bacterial populations are also limited by the production of secretory immunoglobulin A (sIgA) by plasma cells in the lamina propria. sIgA blocks bacteria from binding to epithelial cells, mediates immune exclusion and has been shown to suppress the activity of bacterial type III secretion systems (T3SS) (5, 29-31). Goblet cells, along with microfold cells (M cells), allow for transport of luminal antigens across the epithelial barrier so immune cells can sample luminal contents. Macrophages and transepithelial dendritic cells (DCs) can also sample the contents of the intestinal lumen. Presentation of luminal antigens and bacterial by-products lead to the differentiation of additional immune cell types in the lamina propria, including regulatory T cells (Tregs) and Th17 cells. Innate 3 lymphoid cells (ILCs) produce pro-inflammatory cytokines and chemokines which can protect against bacterial and helminthic infections (5, 29-31). The aryl hydrocarbon receptor (AhR), activated by dietary-derived, microbially-produced ligands such as indoles, ensures maintenance of T cells and ILCs (32, 33). The combination of the physical barrier of the intestinal epithelium, mucus, sIgA, and the immune cells makes up the “mucosal firewall.” Without this, intestinal epithelial barrier damage could occur and allow bacteria and bacterial antigens to translocate to the lamina propria (5). However, if such translocation does occur, these bacteria can be eliminated by resident immune cells such as macrophages (5, 29-31). Along with the more indirect effect of training the immune system, the commensal microbiota can inhibit the growth of pathogens in more direct way, referred to as colonization resistance. The “founder hypothesis” provides an initial explanation for such resistance, as microbes acquired during infancy occupy a niche within the gut and a stronghold on resources for that niche, excluding competing microbes (34). Commensal bacteria compete with invaders for energy sources such as sugars. They can also produce bacteriocins or metabolites such as SCFAs that can inhibit the growth of pathogens, or secondary bile acids that interfere with quorum sensing (QS) systems of some pathogens (22, 35-41). 1.1.2 Further elucidation of how microbially-produced metabolites influence the host immune response and colonization resistance is needed Given the complexity of roles that the gut microbiota plays within the gastrointestinal tract and its influence throughout the body, disruption of microbial diversity or lack of adequate nutrients for bacterial production of metabolites leading to chronic disease is not surprising. Dysbiosis in the gut caused by antibiotic administration can result in persistent antibiotic-induced colitis, and a high fat, low fiber diet may also contribute to the development of disease and leave 4 the gut at increased risk of infection (42-44). There is a need to better understand how the gut microbiota, through the production of dietary-derived metabolites, can affect the intestinal epithelium, host immune response, and resistance to invading pathogens. We have determined that specific microbially-produced Trp metabolites modulate host receptors in the intestinal epithelium to regulate intestinal epithelial barrier function and mediate colonization resistance against enteric pathogens. 1.2 Inflammatory bowel diseases (IBDs) Inflammatory bowel diseases, or IBDs, include ulcerative colitis (UC) and Crohn’s disease (CD). These diseases affect about 1.6 million Americans, with 70,000 new cases being diagnosed in the United States per year and a treatment burden exceeding 6 billion dollars (44, 45). These are chronic diseases with no cure and are associated with poor quality of life, as they often result in the afflicted undergoing hospitalization and surgery to remedy symptoms, which include severe diarrhea, gastrointestinal pain, fatigue, and weight loss (42-45). IBD is among the top three conditions to put individuals at risk for developing colorectal cancer, with approximately 18% of IBD patients developing colorectal cancer after 30 years, compared to about a 4% risk in the general population (46). Host genetics can contribute to dysbiosis, as can diet, pathogen exposure and antibiotic treatment. Studies have shown decreased microbial diversity in IBD patients, with a reduction in Firmicutes, and an increase in Bacteroidetes and, in CD patients, Enterobacteriaceae (47, 48). In genetically predisposed mice where colitis is not observed in a germ-free state, colitis can be induced with introduction of microbiota (49). Such examples further illustrate the interplay between host genetics, the immune system, and the microbiota in the development of IBD. 1.2.1 The intestinal epithelial barrier is disrupted during IBD 5 IECs are connected to one another via junctional complexes, composed of tight junctions (TJs) and, basolaterally, adherens junctions (AJs). AJs provide the adhesive force that seals cells together, and TJs mediate paracellular transport between cells. These junctional complexes help to regulate paracellular permeability, preventing the passage of bacteria and antigens that can activate the immune cells in the lamina propria (50, 51). TJs are complexes composed of several protein types including pore- and barrier-forming claudins, barrier-forming occludin, and scaffolding zonula occludens (ZO) proteins. ZO proteins such as ZO-1 connect claudins and occludin to the actin cytoskeleton which in turn provides a mechanism by which TJ opening can be controlled. Similarly, AJs are composed of barrier- forming E-cadherin, which is linked to the actin cytoskeleton by catenins, including β-catenin. The presence of pro-inflammatory cytokines such as tumor necrosis factor alpha (TNFα) and interferon gamma (IFNγ) is known to affect proteins of the junctional complexes, leading to increased intestinal permeability (49-51). TNFα mediates this permeability via activation of the NF-κB pathway, which leads to an increase in phosphorylation of the myosin light chain (MLC) by myosin light chain kinase (MLCK). Phosphorylation activates the motor protein MLC, causing junctional opening as it exerts its activity on the actin cytoskeleton (52, 53). Additional cytoskeleton associated proteins are linked to junctional opening, including ezrin, deletion of which causes defects in barrier function and has been implicated in intestinal homeostasis (54). Opening of junctional complexes allows bacteria and luminal contents access to the lamina propria where these antigens can activate immune cells and lead to inflammation. During such inflammation, naïve T cells differentiate into various T cell subsets, including Th1, Th2, or Th17 cells (55). Th1 cells, important for the elimination of intracellular invading bacteria, produce TNFα and IFNγ, activating macrophages and triggering apoptosis of epithelial cells, further 6 compromising barrier function. Th2 cells, which protect against parasites and are associated with allergic reactions, produce cytokines such as IL-4, IL-5, IL-9, and IL-13, leading to increased intestinal permeability and more apoptosis (55, 56). Th17 cells, important for elimination of extracellular pathogens and fungi, produce IL-17A and IL-22 which contribute to neutrophil recruitment, additional inflammation, and apoptosis (55, 57). Tregs, which normally suppress abnormal immune responses via production of anti-inflammatory cytokines like IL-10, have been found at reduced levels in IBD patients (55, 58, 59). 1.2.2 Microbially-derived metabolites play a key role in ameliorating IBD Current clinical treatments for IBD include the use of corticosteroids, immunomodulators, cytokine-directed treatments such as infliximab, and alterations of the microbiome with antibiotics, prebiotics, and probiotics. Surgery including total or subtotal colectomy is also an option for some IBD patients (60). More recent research has elucidated microbially-produced postbiotics as a potential therapy for IBD. AhR is key to many of the protective effects of postbiotics in models of IBD. The canonical AhR ligands 2,3,7,8‐tetrachlorodibenzo‐p‐dioxin (TCDD) and 6-formylindolo(3,2-b)carbazole (FICZ) have been shown to ameliorate colitis through various mechanisms. TCDD treatment in mice has been shown to abrogate symptoms of DSS-induced colitis including weight loss and inflammation. It was found to induce prostaglandin E2 production in the mouse colon, which may contribute to its amelioration of colitis (61). Similarly, FICZ treatment was protective in several mouse models of colitis, leading to decreased inflammation, downregulation of pro-inflammatory cytokines TNFα and IFNγ, and induction of IL-22 (62). In vitro, FICZ has been found to attenuate increases in MLCK and phosphorylated MLC (pMLC) and stabilize TJs in Caco-2 monolayers treated with TNFα and IFNγ (63). 7 Looking to the gut microbiome, commensal bacteria have also been shown to activate AhR. Propionibacterium freudenreichii produces 1,4-dihydroxy-2-naphthoic acid (DHNA) which activates AhR both in vitro and in vivo, ameliorating DSS-induced colitis symptoms in an AhR- dependent manner (64). Similarly, Lactobacillus species are capable of modulating AhR and ameliorating colitis in mice due to their ability metabolize tryptophan (65-67). Lactobacillus bulgaricus, for example, activates AhR and increases expression of downstream genes and prostaglandin E2. During DSS-induced colitis, L. bulgaricus administration led to reduced mortality, weight loss and decreased inflammation (65). Interestingly, tryptophan levels are often decreased in patients with IBD. In one study, indoleamine 2,3-dioxygenase (IDO), which catalyzes the conversion of tryptophan to kynurenine, was found to be increased in the intestinal epithelium and lamina propria of patients with active CD compared to healthy individuals, along with decreased levels of tryptophan and an increased ratio of kynurenine to tryptophan (68). Decreased serum levels of tryptophan and increased activation of kynurenine pathways have been observed in other studies with IBD patients (69). Tryptophan can be broken down in vivo into a number of metabolites including kynurenine via host IDO, as well as into other metabolites including indoles and indole-containing compounds by intestinal bacteria. Despite the aforementioned negative correlation between kynurenine levels and disease activity in IBD patients, a chemically induced colitis model produces more severe symptoms in IDO deficient mice (70). Kynurenine has also been found to activate AhR and lead to reduction of symptoms during DSS-induced colitis (71, 72). However, a recent study found that Ido-/- mice have altered gut microbial communities, including a reduction in species that can produce indole-containing metabolites from tryptophan, and that a fecal microbiota transfer from 8 Ido-/- mice to wild type mice led to a decrease in severity of DSS-induced colitis in those mice (73). Increasing dietary tryptophan in both porcine and murine models of IBD leads to a reduction of clinical symptoms (74-76). The ameliorative effect of tryptophan appears to depend on several factors. Ahr-/- mice fed a diet with increased tryptophan did not experience a reduction in colitis symptoms when administered DSS compared to wild type mice (77). In Card9-/- mice, tryptophan is no longer metabolized by the gut microbiota to metabolites such as indole-3-acetic acid (IAA) known to activate AhR, resulting in increased intestinal inflammation. The microbiome is dysregulated in these mice, with decreases in several commensal species including L. reuteri. Upon introduction of three Lactobacillus species, L. murinus, L. reuteri and L. taiwanensis, Card9- /- mice are capable of metabolizing tryptophan to produce metabolites that act as AhR ligands, and intestinal inflammation is decreased (67). The combination of a different Lactobacillus species, L. plantarum, and tryptophan has also been shown to reduce symptoms of DSS-induced colitis in mice, including a reduction in pro-inflammatory cytokines such as TNFα and stabilization of TJ proteins. This effect was also associated with the production of the microbially-derived tryptophan metabolite, and AhR ligand, IAA (78). Bacteroides thetaiotaomicron has been found to produce IAA as well as indole-3-propionic acid (IPA), another AhR ligand. In a mouse model of DSS-induced colitis, production of these metabolites by B. thetaiotaomicron led to increased differentiation of naïve T cells into Th2 and Treg cells compared to pro-inflammatory Th1 and Th17 cells, resulting in overall decreased colonic inflammation (79). Despite the apparent AhR activity of IAA in vivo, recent work utilizing Caco-2 monolayers as an in vitro model of the intestinal epithelium indicated that in this cell model, IAA reduced the expression of TNFα independent of AhR (80). Another study by the same 9 group elucidated TLR4 as a potential novel IAA receptor, mediating the effects of IAA on TNFα downregulation in Caco-2 monolayers (81) Indeed, indole and many indole-containing compounds derived from tryptophan have been found to have a myriad of protective effects in models of IBD. In HCT-8 intestinal epithelial cells, treatment with indole induced transcriptional changes associated with TJ and actin cytoskeleton strengthening, including increased expression of barrier-forming claudins, and decreased expression of the pore-forming claudin 2. This indole treatment improved barrier function and reduced TNFα-induced inflammation. Similar results were seen with the indole derivative 5- hydroxyindole, though IAA, 7-hydroxyindole, and 2-hydroxyindole had no effect, and isatin treatment had the opposite effect in this study (82). Indole-3-carbinol (I3C) is an AhR ligand found to reduce inflammation in mouse colitis models and prevent azoxymethane (AOM)-induced tumor formation (83-86). Recent work indicates that I3C alters the host microbial community, largely through increasing levels of IL-22, and ultimately leads to increased production of the SCFA butyrate in the gut, which contributes to the anti-inflammatory effect (83). During DSS-induced colitis, I3C ameliorates necroptosis and inflammation in an AhR-dependent manner, suppressing NF-κB activation and decreasing expression of proinflammatory cytokines including TNFα (87). Others have found that I3C treatment during DSS induced expression of TJ proteins and modulated host immune response, increasing Treg populations and decreasing Th17 cell counts, again in an AhR-dependent manner (88). In addition to effects on TJ proteins and inflammation, Qazi et al. found that depletion of I3C in the mouse diet affects microbial diversity, leading to an outgrowth of the pathobiont Parvibacter caecicola, and additional changes in members of Firmicutes and Bacteroidetes during DSS treatment (89). 10 IAA, along with tryptamine (TrA), were found to be depleted in the liver and cecum of mice fed a high fat diet. Both are AhR ligands, and in vitro treatment with these metabolites led to a decrease in mRNA transcript levels of pro-inflammatory TNFα (90). TrA may be produced within the gut by Clostridium sporogenes, where it can affect ion secretion within the colon (91). More recently, it has been identified as inducing stabilization of TJs in Caco-2 cells treated with pro-inflammatory compounds like LPS (92). In a model of DSS-induced colitis, levels of indole, IPA, and indole-3-aldehyde (I3A) were decreased. IPA was also found to be reduced in the serum of active UC patients. In vitro, both IPA and I3A improved barrier function in T84 intestinal epithelial cells, and in vivo, IPA treatment reduced symptoms of colitis and decreased inflammation (93). In mice fed a high fat diet, IPA also decreased intestinal permeability. In this study, along with TrA, IPA again was found to reduce permeability in T84 monolayers (94). IPA may be produced in the gut by Peptostreptococcus russellii, as is indole-3-acrylic acid (IA), another metabolite that decreases inflammation (95). Like TrA, IPA is also produced by C. sporogenes. When Trp was administered in vivo along with C. sporogenes, the IPA produced led to a subsequent reduction in intestinal permeability as IPA activates pregnane X receptor (PXR) (96, 97). PXR may also be activated by IAA (96). Indole-3-lactic acid (ILA), produced by Bifidobacterium infantis, can modulate inflammation in vitro in an AhR-dependent manner (98, 99). L. reuteri also produces ILA, as well as I3A, which upon activating AhR in CD4+ T cells, leads to the induction of CD4+CD8αα+ double-positive intraepithelial lymphocytes (DP IELs), an immunoregulatory T cell type similar to Tregs (100). These DP IELs may be able to provide a protective benefit during IBD. L. reuteri growth may be enhanced in the presence of increased tryptophan, and with it, increase production of I3A. I3A produced by L. reuteri can induce IL-22 production and, in an AhR-dependent manner, 11 provide protection against infection with Candida albicans, as well as during DSS-induced colitis (66). In a mouse model of primary sclerosing cholangitis, a liver disease associated with increased intestinal permeability, I3A also acted via the AhR/IL-22 axis to improve gut barrier function (101). Indole-3-pyruvic acid (IPyA) is yet another AhR ligand capable of ameliorating colitis symptoms and inflammation via T cell regulation, increasing populations of IFNγ- IL-10+ CD4+ T cells, as well as increasing the frequency of CD103+ CD11b- DCs (102). Synthetic indole derivatives such as MA-35, an IAA derivative, and an IPA derivative colon-targeted prodrug, IPA-azo-ANA, show promise in colitis models (103, 104). However, we are still learning about the role of microbially-produced tryptophan metabolites. Studies continue to identify new metabolites within the gut, such as 3-methyl-2-oxindole, elucidate their roles as AhR ligands or otherwise, and quantify physiological concentrations (105). 1.2.3 Further elucidation of the roles of microbially-derived metabolites in ameliorating IBD is needed Microbially-produced Trp metabolites, many of them indole-containing, show great potential as potential prophylactics or therapeutics for IBD through their activity as ligands for AhR and other receptors, and their roles in decreasing inflammation and improving intestinal epithelial barrier function. While there is extensive evidence for this, much work remains to be done. Not all Trp metabolites have been well studied, including indole-3-ethanol (IEt), and quantification of Trp metabolites in vivo has only recently begun to be established. It is well established that the effects of many Trp metabolites are AhR-dependent and that some of these metabolites regulate TJ integrity, but a mechanistic connection between Trp metabolite activation of AhR and TJs and AJs was not well elucidated prior to our work. We determined that I3A, IPyA, and IEt acted in an AhR-dependent manner to ameliorate the increased TJ and AJ permeability 12 associated with our IBD models via a pathway including host proteins ezrin and MLC. We next sought to investigate how these metabolites might mediate protection against enteric pathogens and the role of host receptors and proteins during such infections. 1.3 Enterohemorrhagic Escherichia coli (EHEC) O157:H7 Various extracellular pathogens can infect the gastrointestinal tract and cause severe disease by disrupting the intestinal epithelial barrier and causing inflammation. Some of the most well-known extracellular gut pathogens include those that are attaching and effacing (A/E), such as enteropathogenic Escherichia coli (EPEC) and enterohemorrhagic Escherichia coli (EHEC). EHEC serotype O157:H7 was first characterized during the investigation of an outbreak of hemorrhagic colitis, a condition defined by severe abdominal cramps and bloody diarrhea (106). Along with hemorrhagic colitis, EHEC O157:H7 can cause potentially fatal hemolytic uremic syndrome (HUS), characterized by thrombocytopenia, hemolytic anemia, and acute kidney failure. In the United States alone, EHEC O157:H7 causes over 70,000 illnesses per year, including more than 2,000 hospitalizations and 60 deaths. Transmitted via contaminated food, water, person to person, and animal contact, and with a very low infectious dose of fewer than 100 colony forming units (CFU), it is classified as a nationally notifiable infection. Despite regulatory measures being put in place, outbreaks of EHEC O157:H7 are still relatively common in the United States and other developed nations, with the primary source of infection being ground beef or produce (107- 109). 1.3.1 EHEC injects bacterial effector proteins, disrupts host proteins during infection, and induces inflammation EHEC is an A/E pathogen, as bacteria bind intimately to intestinal epithelial cells, manipulating host cell actin to induce pedestal formation, and efface the microvilli of cells. EHEC 13 first mediates contact with intestinal epithelial cells through the type 4 hemorrhagic coli pilus (HCP) and E. coli common pilus (ECP) (110). Within the EHEC genome, the locus of enterocyte effacement (LEE) pathogenicity island codes for a T3SS, a needle-like injection system that allows for the translocation of many effector proteins, including translocated intimin receptor (Tir), into the host cell. Once Tir is injected into and expressed on the surface of host cells it, along with nucleolin, which is upregulated by Shiga toxin (Stx), acts as a receptor to the bacterial outer membrane protein intimin, furthering the intimate attachment of EHEC to host cells. Tir interacts with host protein insulin receptor tyrosine kinase substrate (IRTKS), which binds the Tir cytoskeleton-coupling protein (known as Tccp or EspFu). Tccp binds to host neural Wiskott- Aldrich syndrome protein (N-WASP) and the actin related protein (Arp) 2/3 complex, causing rearrangement of actin to contribute to pedestal formation. Because actin is linked to junctional complexes, when actin is pulled and rearranged, this contributes to junctional complex disruption and increased intestinal permeability (109-112). The effector protein EspF is also linked to the disruption of tight junction proteins, including occludin and ZO-1 during EHEC infection. Another effector protein, Map, binds to ezrin/radixin/moesin-binding phosphoprotein 50 (EBP50) which links the actin cytoskeleton to the plasma membrane via ezrin (109-112). Additionally, A/E pathogens are known to trigger phosphorylation of MLC via MLCK, which increases intestinal epithelial barrier disruption during infection through its motor function on the actin cytoskeleton (113). Shiga toxins Stx1 and Stx2, encoded by bacteriophages and released after phage-mediated lysis, are largely responsible for induction of HUS due to their action on the kidneys. They can also bind to the globotriaosylceramide-3 (Gb-3) receptor on Paneth cells in the intestine. Stx can 14 be internalized into cells and once activated, blocks protein synthesis, leading to apoptosis (109, 110). Initial immune attacks against EHEC occur as pattern recognition receptors such as Toll- like receptors (TLRs) recognize bacterial antigens, including lipopolysaccharide (LPS), flagellin, and HCP proteins. Activation of these receptors leads to stimulation of the immune system and production of pro-inflammatory cytokines such as TNFα and IFNγ by macrophages and T cells (109, 114-116). IL-8 produced in response to infection also serves as a chemoattractant for neutrophils to accumulate in the colonic mucosa (109, 115, 116). DC activation leads to production of IL-23, which in turn increases the production of IL-22 from ILCs, neutrophils, and CD4+ T cells, including Th17 cells. IL-22 stimulates protection of the gut barrier by upregulating epithelial wound repair and production of antimicrobial proteins. Production of IL-21, also by T cells, boosts plasma cell differentiation and immunoglobulin G (IgG) production (109, 117). Clearance of EHEC from infected individuals relies heavily on this immune response, particularly as clinical treatments are limited. 1.3.2 Murine pathogen Citrobacter rodentium serves as a physiologically relevant in vivo model of EHEC infection Mice are innately resistant to EHEC infection, so infection with Citrobacter rodentium provides a physiologically relevant model to study EHEC in vivo (118-120). The natural mouse pathogen C. rodentium utilizes the same T3SS-driven mechanism of infection as EHEC. Indeed, EHEC and C. rodentium genomes both feature the LEE pathogenicity island and share 41 open reading frames (ORFs). Though it does not produce Stx, infection of mice with C. rodentium can lead to diarrhea, largely due to disruption of tight junctions, along with colonic hyperplasia, a feature unique to C. rodentium infection (111, 118-120). The degree of infection with C. rodentium 15 is highly dependent on mouse strain, and the composition of the existing intestinal microbiome is also important to infection severity, further illustrating the interaction between the intestinal microbiome and A/E pathogens like EHEC and C. rodentium (118-122). 1.3.3 Microbially-derived metabolites can mediate resistance to EHEC and C. rodentium infections Treatments for EHEC infection are largely supportive, consisting of administration of fluids, dialysis, and plasma exchange. Antibiotic treatment of EHEC is contraindicated as antibiotic-mediated killing and lysis of bacteria can lead to release of Stx (109, 123-125). In fact, antibiotic use during EHEC infection has been linked to increased occurrence of HUS in patients (126). Additionally, use of anti-diarrheal medication is usually not recommended as it can delay clearance of bacteria and Stx from the intestinal tract (109). Looking to the gut microbiome for potential treatment options, researchers have found some interesting interactions between the microbiota and EHEC. Commensal B. thetaiotaomicron has been shown to increase EHEC virulence as it increases succinate levels in the gut, leading to increased expression of the transcription factor Cra, which is important for virulence gene expression. Mice infected with C. rodentium experienced increased weight loss, intestinal permeability and mortality when also colonized by B. thetaiotaomicron (127). Expression of the LEE pathogenicity island during C. rodentium infection is also increased by microbially-produced 1,2-propanediol (128). Alternatively, other commensals have been shown to provide protection during infection with EHEC. Bifidobacterium longum, via production of the SCFA acetate, improves intestinal barrier function during EHEC infection (129, 130). Similarly, the SCFA butyrate also protects against infection with C. rodentium both in vitro and in vivo, exerting antimicrobial effects via 16 inhibition of histone deacetylase 3 and subsequent increase in macrophage differentiation (131, 132). L. reuteri also produces an antimicrobial compound, hydroxypropionaldehyde (HPA), when provided with glycerol under anaerobic conditions (133). Indole is yet another interesting microbially-produced metabolite in the context of EHEC and C. rodentium infections. E. coli species, including EHEC, produce indole and it serves as a key molecule during quorum sensing. (134, 135). C. rodentium, however, does not produce indole (136). Beyond quorum sensing, the effects of indole during these infections are an increasingly studied area of research. Kumar and Sperandio have reported that indole exists in higher concentrations in the lumen compared to closer to the intestinal barrier, and by sensing indole concentrations, EHEC can regulate expression of virulence genes. Their findings indicate that the high luminal concentration of indole leads to repression of virulence genes, and that these genes are better expressed closer to the epithelium where there is a lower concentration of indole. They proposed a mechanism involving the histidine sensor kinase CpxA acting as an indole sensor. Kumar and Sperandio altered the concentration of indole in a murine model through antibiotic treatment and reconstitution with wild-type, indole-producing B. thetaiotaomicron or a TnaA mutant strain of B. thetaiotaomicron, which does not produce indole, or a C. rodentium strain capable of producing indole (137). Their use of B. thetaiotaomicron in this context is potentially problematic given the findings of Curtis et al. that this commensal increases EHEC virulence through its effects on succinate concentration in the gut (127). Additionally, in contrast to Kumar and Sperandio’s findings, Hirakawa et al. previously reported that indole can lead to increased secretion of effector proteins EspA and EspB, thereby increasing EHEC’s ability to attach to host cells, with additional confirmation of this effect in a recent paper utilizing a synthetic compound to decrease indole levels (138, 139). Hirakawa and 17 coworkers have also reported that indole upregulates the expression of exporter genes including AcrD leading to increased drug resistance in E. coli, a mechanism mediated by BaeSR and CpxAR (140). Work by Bansal et al. has shown that indole induces negative chemotaxis in EHEC and decreases EHEC biofilm formation, motility, and attachment to cells in vitro (141). Others have also reported that indole can reduce expression of genes within the LEE pathogenicity island and reduce EHEC biofilm formation and motility. These effects were also seen in C. rodentium (142). Similarly, other metabolites produced during the microbial breakdown of tryptophan in the gut exert effects during EHEC infection. During C. rodentium infection, IDO deficient mice were more resistant and had fewer symptoms upon infection with the pathogen compared to their wildtype counterparts. This may be due to decreased production of kynurenine, which was found to induce B cell apoptosis and reduce antibody production (143). Others have found beneficial roles for indole derivatives during EHEC infection. In vitro, 5-hydroxyindole and 7-hydroxyindole have been shown to reduce EHEC biofilm formation, though isatin increased biofilm production, repressing the TnA operon and indole synthesis as well as autoinducer 2 (AI-2) transporter genes, while upregulating flagellar genes associated with motility (144). Skatole, I3A, and indole-3- acetonitrile (IAN) also inhibit biofilm formation by EHEC (145, 146). Another tryptophan metabolite, 5-HT, reduces LEE virulence gene expression in EHEC and C. rodentium by acting on CpxA (147). In a model of aberrant Cyp1a1 expression in which AhR ligands were depleted, an increase in the severity of C. rodentium infection was observed. I3C, as well its acid condensation products 3,3'-diindolylmethane (DIM) and 5,11-dihydroindolo-[3,2-b]carbazole (ICZ), ameliorated the symptoms associated with C. rodentium infection by increasing Th17 cell populations and 18 production of IL-22 (148). I3C also reduced C. rodentium growth and adhesion to Caco-2 cells in an in vitro model (149). I3A has been shown to reduce LEE virulence gene expression in EHEC without affecting growth, reduce biofilm formation and motility, as well as reduce the cytotoxic effect of Stx. I3A had similar effects on C. rodentium, reducing LEE gene expression and inhibiting pedestal formation. In vivo, treatment with I3A also ameliorated symptoms of C. rodentium infection (142). Data such as these indicate that while interactions between the gut microbiota, metabolites and EHEC may be complex, they are worth exploring in search of potential therapies for infection. 1.3.4 Further elucidation of the roles of microbially-derived metabolites during EHEC and C. rodentium infection is needed Research indicates that microbially-produced Trp metabolites may play an important role in mitigating the effects of infection with A/E pathogens. While there is data to indicate that indole and indole-containing compounds like I3A exert protective effects, other Trp metabolites, including IPyA and IEt, are less well studied in the context of infection with extracellular enteric pathogens. Furthermore, the host receptors and pathways that these metabolites act on must be better understood. We identified a novel receptor, dopamine receptor D2 (DRD2) for I3A, IPyA and IEt, and determined that these metabolites act via DRD2 to mediate colonization resistance to infection with EHEC and C. rodentium via a pathway including host proteins IRTKS, N-WASP and the Arp2/3 complex. 19 1.4 References 1. Clemente, Jose C., et al. "The impact of the gut microbiota on human health: an integrative view." Cell 148.6 (2012): 1258-1270. 2. Hooper, Lora V., and Andrew J. Macpherson. "Immune adaptations that maintain homeostasis with the intestinal microbiota." Nature Reviews Immunology 10.3 (2010): 159-169. 3. Huttenhower, Curtis, et al. "Structure, function and diversity of the healthy human microbiome." nature 486.7402 (2012): 207. 4. Sommer, Felix, and Fredrik Bäckhed. "Know your neighbor: Microbiota and host epithelial cells interact locally to control intestinal function and physiology." Bioessays 38.5 (2016): 455-464. 5. Belkaid, Yasmine, and Timothy W. Hand. "Role of the microbiota in immunity and inflammation." Cell 157.1 (2014): 121-141. 6. Porter, Nathan T., and Eric C. Martens. "The critical roles of polysaccharides in gut microbial ecology and physiology." Annual review of microbiology 71 (2017): 349-369. 7. Shreiner, Andrew B., John Y. Kao, and Vincent B. Young. "The gut microbiome in health and in disease." Current opinion in gastroenterology 31.1 (2015): 69. 8. Larsbrink, Johan, et al. "A discrete genetic locus confers xyloglucan metabolism in select human gut Bacteroidetes." Nature 506.7489 (2014): 498-502. 9. Flint, Harry J., et al. "Microbial degradation of complex carbohydrates in the gut." Gut microbes 3.4 (2012): 289-306. 10. Das, Promi, Parizad Babaei, and Jens Nielsen. "Metagenomic analysis of microbe-mediated vitamin metabolism in the human gut microbiome." BMC genomics 20.1 (2019): 208. 11. Cryan, John F., and Timothy G. Dinan. "Mind-altering microorganisms: the impact of the gut microbiota on brain and behaviour." Nature reviews neuroscience 13.10 (2012): 701-712. 20 12. Holzer, Peter, and Aitak Farzi. "Neuropeptides and the microbiota-gut-brain axis." Microbial endocrinology: The microbiota-gut-brain axis in health and disease. Springer, New York, NY, 2014. 195-219. 13. Nicholson, Jeremy K., et al. "Host-gut microbiota metabolic interactions." Science 336.6086 (2012): 1262-1267. 14. Lyte, Mark. "Probiotics function mechanistically as delivery vehicles for neuroactive compounds: microbial endocrinology in the design and use of probiotics." Bioessays 33.8 (2011): 574-581. 15. Forsythe, Paul, and Wolfgang A. Kunze. "Voices from within: gut microbes and the CNS." Cellular and molecular life sciences 70.1 (2013): 55-69. 16. Asano, Yasunari, et al. "Critical role of gut microbiota in the production of biologically active, free catecholamines in the gut lumen of mice." American Journal of Physiology-Gastrointestinal and Liver Physiology 303.11 (2012): G1288-G1295. 17. Barrett, E., et al. "γ‐Aminobutyric acid production by culturable bacteria from the human intestine." Journal of applied microbiology 113.2 (2012): 411-417. 18. Vaughan, Carl J., et al. "Identification and regional distribution of the dopamine D1A receptor in the gastrointestinal tract." American Journal of Physiology-Regulatory, Integrative and Comparative Physiology 279.2 (2000): R599-R609. 19. Li, Zhi Shan, et al. "Physiological modulation of intestinal motility by enteric dopaminergic neurons and the D2 receptor: analysis of dopamine receptor expression, location, development, and function in wild-type and knock-out mice." Journal of Neuroscience 26.10 (2006): 2798-2807. 21 20. Louis, Petra, and Harry J. Flint. "Diversity, metabolism and microbial ecology of butyrate- producing bacteria from the human large intestine." FEMS microbiology letters 294.1 (2009): 1- 8. 21. Chang, Pamela V., et al. "The microbial metabolite butyrate regulates intestinal macrophage function via histone deacetylase inhibition." Proceedings of the National Academy of Sciences 111.6 (2014): 2247-2252. 22. Chang, Pamela V. "Chemical mechanisms of colonization resistance by the gut microbial metabolome." ACS Chemical Biology (2020). 23. Nicolas, Gael R., and Pamela V. Chang. "Deciphering the chemical lexicon of host–gut microbiota interactions." Trends in pharmacological sciences (2019). 24. Jones, Mitchell L., et al. "The human microbiome and bile acid metabolism: dysbiosis, dysmetabolism, disease and intervention." Expert opinion on biological therapy 14.4 (2014): 467- 482. 25. Chen, Mei Lan, Kiyoshi Takeda, and Mark S. Sundrud. "Emerging roles of bile acids in mucosal immunity and inflammation." Mucosal immunology 12.4 (2019): 851-861. 26. Steed, Ashley L., et al. "The microbial metabolite desaminotyrosine protects from influenza through type I interferon." Science 357.6350 (2017): 498-502. 27. Cervenka, Igor, Leandro Z. Agudelo, and Jorge L. Ruas. "Kynurenines: Tryptophan’s metabolites in exercise, inflammation, and mental health." Science 357.6349 (2017): eaaf9794. 28. Roager, Henrik M., and Tine R. Licht. "Microbial tryptophan catabolites in health and disease." Nature communications 9.1 (2018): 1-10. 29. Peterson, Lance W., and David Artis. "Intestinal epithelial cells: regulators of barrier function and immune homeostasis." Nature Reviews Immunology 14.3 (2014): 141-153. 22 30. Shreiner, Andrew B., John Y. Kao, and Vincent B. Young. "The gut microbiome in health and in disease." Current opinion in gastroenterology 31.1 (2015): 69. 31. Honda, Kenya, and Dan R. Littman. "The microbiome in infectious disease and inflammation." Annual review of immunology 30 (2012): 759-795. 32. Qiu, Ju, et al. "The aryl hydrocarbon receptor regulates gut immunity through modulation of innate lymphoid cells." Immunity 36.1 (2012): 92-104. 33. Stockinger, Brigitta, et al. "The aryl hydrocarbon receptor: multitasking in the immune system." Annual review of immunology 32 (2014): 403-432. 34. Litvak, Yael, and Andreas J. Bäumler. "The founder hypothesis: a basis for microbiota resistance, diversity in taxa carriage, and colonization resistance against pathogens." PLoS pathogens 15.2 (2019). 35. Ubeda, Carles, Ana Djukovic, and Sandrine Isaac. "Roles of the intestinal microbiota in pathogen protection." Clinical & translational immunology 6.2 (2017): e128. 36. Sassone-Corsi, Martina, and Manuela Raffatellu. "No vacancy: how beneficial microbes cooperate with immunity to provide colonization resistance to pathogens." The Journal of Immunology 194.9 (2015): 4081-4087. 37. Zmora, Niv, et al. "Personalized gut mucosal colonization resistance to empiric probiotics is associated with unique host and microbiome features." Cell 174.6 (2018): 1388-1405. 38. Buffie, Charlie G., and Eric G. Pamer. "Microbiota-mediated colonization resistance against intestinal pathogens." Nature Reviews Immunology 13.11 (2013): 790-801. 39. Ducarmon, Q. R., et al. "Gut microbiota and colonization resistance against bacterial enteric infection." Microbiology and Molecular Biology Reviews 83.3 (2019): e00007-19. 23 40. Leshem, Avner, Timur Liwinski, and Eran Elinav. "Immune-Microbiota Interplay and Colonization Resistance in Infection." Molecular Cell (2020). 41. Pickard, Joseph M., and Gabriel Núñez. "Pathogen Colonization Resistance in the Gut and Its Manipulation for Improved Health." The American journal of pathology (2019). 42. Durack, Juliana, and Susan V. Lynch. "The gut microbiome: Relationships with disease and opportunities for therapy." Journal of Experimental Medicine 216.1 (2019): 20-40. 43. Podolsky, Daniel K. "Inflammatory bowel disease." New England Journal of Medicine 325.13 (1991): 928-937. 44. Baumgart, Daniel C., and Simon R. Carding. "Inflammatory bowel disease: cause and immunobiology." The Lancet 369.9573 (2007): 1627-1640. 45. Dahlhamer, James M., et al. "Prevalence of inflammatory bowel disease among adults aged≥ 18 years—United States, 2015." Morbidity and Mortality Weekly Report 65.42 (2016): 1166- 1169. 46. Andersen, Nynne Nyboe, and Tine Jess. "Has the risk of colorectal cancer in inflammatory bowel disease decreased?" World journal of gastroenterology: WJG 19.43 (2013): 7561. 47. Round, June L., and Sarkis K. Mazmanian. "The gut microbiota shapes intestinal immune responses during health and disease." Nature reviews immunology 9.5 (2009): 313-323. 48. Walker, Alan W., et al. "High-throughput clone library analysis of the mucosa-associated microbiota reveals dysbiosis and differences between inflamed and non-inflamed regions of the intestine in inflammatory bowel disease." BMC microbiology 11.1 (2011): 7. 49. Xavier, R. J., and D. K. Podolsky. "Unravelling the pathogenesis of inflammatory bowel disease." Nature 448.7152 (2007): 427-434. 24 50. Turner, Jerrold R. "Intestinal mucosal barrier function in health and disease." Nature reviews immunology 9.11 (2009): 799-809. 51. Zihni, Ceniz, et al. "Tight junctions: from simple barriers to multifunctional molecular gates." Nature reviews Molecular cell biology 17.9 (2016): 564. 52. Al-Sadi, Rana, Michel Boivin, and Thomas Ma. "Mechanism of cytokine modulation of epithelial tight junction barrier." Frontiers in bioscience: a journal and virtual library 14 (2009): 2765. 53. Ye, Dongmei, Iris Ma, and Thomas Y. Ma. "Molecular mechanism of tumor necrosis factor-α modulation of intestinal epithelial tight junction barrier." American Journal of Physiology- Gastrointestinal and Liver Physiology 290.3 (2006): G496-G504. 54. Casaletto, Jessica B., et al. "Ezrin-mediated apical integrity is required for intestinal homeostasis." Proceedings of the National Academy of Sciences 108.29 (2011): 11924-11929. 55. Geremia, Alessandra, et al. "Innate and adaptive immunity in inflammatory bowel disease." Autoimmunity reviews 13.1 (2014): 3-10. 56. Romagnani, Sergio. "Lymphokine production by human T cells in disease states." Annual review of immunology 12.1 (1994): 227-257. 57. Zhou, Liang, et al. "IL-6 programs TH-17 cell differentiation by promoting sequential engagement of the IL-21 and IL-23 pathways." Nature immunology 8.9 (2007): 967-974. 58. Valencia, Xavier, et al. "TNF downmodulates the function of human CD4+ CD25hi T- regulatory cells." Blood 108.1 (2006): 253-261. 59. Boden, Elisa K., and Scott B. Snapper. "Regulatory T cells in inflammatory bowel disease." Current opinion in gastroenterology 24.6 (2008): 733-741. 25 60. Baumgart, Daniel C., and William J. Sandborn. "Inflammatory bowel disease: clinical aspects and established and evolving therapies." The Lancet 369.9573 (2007): 1641-1657. 61. Takamura, Takeyuki, et al. "Activation of the aryl hydrocarbon receptor pathway may ameliorate dextran sodium sulfate‐induced colitis in mice." Immunology and cell biology 88.6 (2010): 685-689. 62. Monteleone, Ivan, et al. "Aryl hydrocarbon receptor-induced signals up-regulate IL-22 production and inhibit inflammation in the gastrointestinal tract." Gastroenterology 141.1 (2011): 237-248. 63. Yu, Min, et al. "Aryl hydrocarbon receptor activation modulates intestinal epithelial barrier function by maintaining tight junction integrity." International journal of biological sciences 14.1 (2018): 69. 64. Fukumoto, Suguru, et al. "Identification of a probiotic bacteria‐derived activator of the aryl hydrocarbon receptor that inhibits colitis." Immunology and cell biology 92.5 (2014): 460-465. 65. Takamura, Takeyuki, et al. "Lactobacillus bulgaricus OLL1181 activates the aryl hydrocarbon receptor pathway and inhibits colitis." Immunology and cell biology 89.7 (2011): 817-822. 66. Zelante, Teresa, et al. "Tryptophan catabolites from microbiota engage aryl hydrocarbon receptor and balance mucosal reactivity via interleukin-22." Immunity 39.2 (2013): 372-385. 67. Lamas, Bruno, et al. "CARD9 impacts colitis by altering gut microbiota metabolism of tryptophan into aryl hydrocarbon receptor ligands." Nature medicine 22.6 (2016): 598. 68. Gupta, Nitin K., et al. "Serum analysis of tryptophan catabolism pathway: correlation with Crohn's disease activity." Inflammatory bowel diseases 18.7 (2012): 1214-1220. 69. Nikolaus, Susanna, et al. "Increased tryptophan metabolism is associated with activity of inflammatory bowel diseases." Gastroenterology 153.6 (2017): 1504-1516. 26 70. Takamatsu, Manabu, et al. "IDO1 Plays an immunosuppressive role in 2, 4, 6-trinitrobenzene sulfate–induced colitis in mice." The Journal of Immunology 191.6 (2013): 3057-3064. 71. Lanis, Jordi M., et al. "Tryptophan metabolite activation of the aryl hydrocarbon receptor regulates IL-10 receptor expression on intestinal epithelia." Mucosal immunology 10.5 (2017): 1133-1144. 72. Mezrich, Joshua D., et al. "An interaction between kynurenine and the aryl hydrocarbon receptor can generate regulatory T cells." The Journal of Immunology 185.6 (2010): 3190-3198. 73. Shin, Ji-Hee, et al. "Gut microorganisms and their metabolites modulate the severity of acute colitis in a tryptophan metabolism-dependent manner." European Journal of Nutrition (2020): 1- 11. 74. Kim, Connie J., et al. "l-Tryptophan exhibits therapeutic function in a porcine model of dextran sodium sulfate (DSS)-induced colitis." The Journal of nutritional biochemistry 21.6 (2010): 468- 475. 75. Shizuma, Toru, Hidezo Mori, and Naoto Fukuyama. "Protective effect of tryptophan against dextran sulfate sodium-induced experimental colitis." Turk J Gastroenterol 24.1 (2013): 30-35. 76. Wang, Bin, et al. "Dietary L-Tryptophan Regulates Colonic Serotonin Homeostasis in Mice with Dextran Sodium Sulfate-Induced Colitis." The Journal of Nutrition (2020). 77. Islam, Jahidul, et al. "Dietary tryptophan alleviates dextran sodium sulfate-induced colitis through aryl hydrocarbon receptor in mice." The Journal of nutritional biochemistry 42 (2017): 43-50. 78. Shi, Jialu, et al. "Protective effects of tryptophan-catabolizing Lactobacillus plantarum KLDS 1.0386 against dextran sodium sulfate-induced colitis in mice." Food & Function (2020). 27 79. Li, Keying, et al. "Bacteroides thetaiotaomicron relieves colon inflammation by activating aryl hydrocarbon receptor and modulating CD4+ T cell homeostasis." International Immunopharmacology 90 (2021): 107183. 80. Chowdhury, Md Mazharul Islam, et al. "Suppression of TNFα expression induced by indole- 3-acetic acid is not mediated by AhR activation in Caco-2 cells." Bioscience, Biotechnology, and Biochemistry 85.4 (2021): 902-906. 81. Chowdhury, Md Mazharul Islam, et al. "TLR4 may be a novel indole-3-acetic acid receptor that is implicated in the regulation of CYP1A1 and TNFα expression depending on the culture stage of Caco-2 cells." Bioscience, Biotechnology, and Biochemistry (2021). 82. Bansal, Tarun, et al. "The bacterial signal indole increases epithelial-cell tight-junction resistance and attenuates indicators of inflammation." Proceedings of the national academy of sciences 107.1 (2010): 228-233. 83. Busbee, Philip B., et al. "Indole-3-carbinol prevents colitis and associated microbial dysbiosis in an IL-22–dependent manner." JCI insight 5.1 (2020). 84. Kawajiri, Kaname, et al. "Aryl hydrocarbon receptor suppresses intestinal carcinogenesis in ApcMin/+ mice with natural ligands." Proceedings of the national academy of sciences 106.32 (2009): 13481-13486. 85. Díaz-Díaz, Carol J., et al. "The aryl hydrocarbon receptor mediates the chemopreventive effect of indole-3-carbinol in an inflammation-associated colorectal tumorigenesis model." Annals of surgery 264.3 (2016): 429. 86. Metidji, Amina, et al. "The environmental sensor AHR protects from inflammatory damage by maintaining intestinal stem cell homeostasis and barrier integrity." Immunity 49.2 (2018): 353- 362. 28 87. Peng, Chunting, et al. "Indole-3-carbinol ameliorates necroptosis and inflammation of intestinal epithelial cells in mice with ulcerative colitis by activating aryl hydrocarbon receptor." Experimental Cell Research 404.2 (2021): 112638. 88. Riemschneider, Sina, et al. "Indol-3-Carbinol and Quercetin Ameliorate Chronic DSS-Induced Colitis in C57BL/6 Mice by AhR-Mediated Anti-Inflammatory Mechanisms." International journal of environmental research and public health 18.5 (2021): 2262. 89. Qazi, Aisha, et al. "Depletion of AhR Ligands in Diet Worsens Inflammatory Conditions in Mice with Chronic DSS‐Induced Colitis." The FASEB Journal 35 (2021). 90. Krishnan, Smitha, et al. "Gut microbiota-derived tryptophan metabolites modulate inflammatory response in hepatocytes and macrophages." Cell reports 23.4 (2018): 1099-1111. 91. Williams, Brianna B., et al. "Discovery and characterization of gut microbiota decarboxylases that can produce the neurotransmitter tryptamine." Cell host & microbe 16.4 (2014): 495-503. 92. Grosheva, Inna, et al. "High-throughput screen identifies host and microbiota regulators of intestinal barrier function." Gastroenterology (2020). 93. Alexeev, Erica E., et al. "Microbiota-derived indole metabolites promote human and murine intestinal homeostasis through regulation of interleukin-10 receptor." The American journal of pathology 188.5 (2018): 1183-1194. 94. Jennis, M., et al. "Microbiota‐derived tryptophan indoles increase after gastric bypass surgery and reduce intestinal permeability in vitro and in vivo." Neurogastroenterology & motility 30.2 (2018): e13178. 95. Wlodarska, Marta, et al. "Indoleacrylic acid produced by commensal peptostreptococcus species suppresses inflammation." Cell host & microbe 22.1 (2017): 25-37. 29 96. Venkatesh, Madhukumar, et al. "Symbiotic bacterial metabolites regulate gastrointestinal barrier function via the xenobiotic sensor PXR and Toll-like receptor 4." Immunity 41.2 (2014): 296-310. 97. Dodd, Dylan, et al. "A gut bacterial pathway metabolizes aromatic amino acids into nine circulating metabolites." Nature 551.7682 (2017): 648-652 98. Meng, Di, et al. "Indole-3-lactic acid, a metabolite of tryptophan, secreted by Bifidobacterium longum subspecies infantis is anti-inflammatory in the immature intestine." Pediatric Research (2020): 1-10. 99. Ehrlich, Amy M., et al. "Indole-3-lactic acid associated with Bifidobacterium-dominated microbiota significantly decreases inflammation in intestinal epithelial cells." BMC microbiology 20.1 (2020): 1-13. 100. Cervantes-Barragan, Luisa, et al. "Lactobacillus reuteri induces gut intraepithelial CD4+ CD8αα+ T cells." Science 357.6353 (2017): 806-810. 101. D’Onofrio, Fiorella, et al. "Indole-3-Carboxaldehyde Restores Gut Mucosal Integrity and Protects from Liver Fibrosis in Murine Sclerosing Cholangitis." Cells 10.7 (2021): 1622. 102. Aoki, Reiji, et al. "Indole-3-pyruvic acid, an aryl hydrocarbon receptor activator, suppresses experimental colitis in mice." The Journal of Immunology 201.12 (2018): 3683-3693. 103. Kanehara, Keigo, et al. "The indole compound MA-35 attenuates tumorigenesis in an inflammation-induced colon cancer model." Scientific reports 9.1 (2019): 1-13. 104. Lee, Hanju, et al. "Preparation and evaluation of colon-targeted prodrugs of the microbial metabolite 3-indolepropionic acid as an anticolitic agent." Molecular pharmaceutics 18.4 (2021): 1730-1741. 30 105. Dong, Fangcong, et al. "Intestinal microbiota-derived tryptophan metabolites are predictive of Ah receptor activity." Gut microbes 12.1 (2020): 1-24. 106. Riley, Lee W., et al. "Hemorrhagic colitis associated with a rare Escherichia coli serotype." New England Journal of Medicine 308.12 (1983): 681-685. 107. Rangel, Josefa M., et al. "Epidemiology of Escherichia coli O157: H7 outbreaks, united states, 1982–2002." Emerging infectious diseases 11.4 (2005): 603. 108. Kaper, James B., James P. Nataro, and Harry LT Mobley. "Pathogenic escherichia coli." Nature reviews microbiology 2.2 (2004): 123-140. 109. Ho, Nathan K., et al. "Pathogenicity, host responses and implications for management of enterohemorrhagic Escherichia coli O157: H7 infection." Canadian Journal of Gastroenterology and Hepatology 27.5 (2013): 281-285. 110. Croxen, Matthew A., and B. Brett Finlay. "Molecular mechanisms of Escherichia coli pathogenicity." Nature Reviews Microbiology 8.1 (2010): 26-38. 111. Guttman, Julian A., et al. "Attaching and effacing pathogen‐induced tight junction disruption in vivo." Cellular microbiology 8.4 (2006): 634-645. 112. Guttman, Julian A., et al. "Evidence that tight junctions are disrupted due to intimate bacterial contact and not inflammation during attaching and effacing pathogen infection in vivo." Infection and immunity 74.11 (2006): 6075-6084. 113. Yuhan, Robert O, et al. "Enteropathogenic Escherichia coli-induced myosin light chain phosphorylation alters intestinal epithelial permeability." Gastroenterology 113.6 (1997): 1873- 1882. 31 114. Miyamoto, Yukiko, et al. "Role of Shiga toxin versus H7 flagellin in enterohaemorrhagic Escherichia coli signalling of human colon epithelium in vivo." Cellular microbiology 8.5 (2006): 869-879. 115. Dahan, Stephanie, et al. "Enterohemorrhagic Escherichia coli infection induces interleukin-8 production via activation of mitogen-activated protein kinases and the transcription factors NF-κB and AP-1 in T84 cells." Infection and immunity 70.5 (2002): 2304-2310. 116. Ledesma, Maria A., et al. "The hemorrhagic coli pilus (HCP) of Escherichia coli O157: H7 is an inducer of proinflammatory cytokine secretion in intestinal epithelial cells." PLoS one 5.8 (2010). 117. Silberger, Daniel J., Carlene L. Zindl, and Casey T. Weaver. "Citrobacter rodentium: a model enteropathogen for understanding the interplay of innate and adaptive components of type 3 immunity." Mucosal immunology 10.5 (2017): 1108-1117. 118. Crepin, Valerie F., et al. "Citrobacter rodentium mouse model of bacterial infection." Nature protocols 11.10 (2016): 1851. 119. Mundy, Rosanna, et al. "Comparison of colonization dynamics and pathology of mice infected with enteropathogenic Escherichia coli, enterohaemorrhagic E. coli and Citrobacter rodentium." FEMS microbiology letters 265.1 (2006): 126-132. 120. Mundy, Rosanna, et al. "Citrobacter rodentium of mice and man." Cellular microbiology 7.12 (2005): 1697-1706. 121. Flynn, Andrew N., and Andre G. Buret. "Tight junctional disruption and apoptosis in an in vitro model of Citrobacter rodentium infection." Microbial pathogenesis 45.2 (2008): 98-104. 122. Willing, Benjamin P., et al. "Altering host resistance to infections through microbial transplantation." PloS one 6.10 (2011). 32 123. Goldwater, Paul N., and Karl A. Bettelheim. "Treatment of enterohemorrhagic Escherichia coli (EHEC) infection and hemolytic uremic syndrome (HUS)." BMC medicine 10.1 (2012): 12. 124. Hickey, Christina A., et al. "Early volume expansion during diarrhea and relative nephroprotection during subsequent hemolytic uremic syndrome." Archives of pediatrics & adolescent medicine 165.10 (2011): 884-889. 125. Obrig, Tom G. "Escherichia coli Shiga toxin mechanisms of action in renal disease." Toxins 2.12 (2010): 2769-2794. 126. Wong, Craig S., et al. "Risk factors for the hemolytic uremic syndrome in children infected with Escherichia coli O157: H7: a multivariable analysis." Clinical Infectious Diseases 55.1 (2012): 33-41. 127. Curtis, Meredith M., et al. "The gut commensal Bacteroides thetaiotaomicron exacerbates enteric infection through modification of the metabolic landscape." Cell host & microbe 16.6 (2014): 759-769. 128. Connolly, James PR, et al. "Host-associated niche metabolism controls enteric infection through fine-tuning the regulation of type 3 secretion." Nature communications 9.1 (2018): 1-14. 129. Fukuda, Shinji, et al. "Bifidobacteria can protect from enteropathogenic infection through production of acetate." Nature 469.7331 (2011): 543-547. 130. Fukuda, Shinji, et al. "Acetate-producing bifidobacteria protect the host from enteropathogenic infection via carbohydrate transporters." Gut microbes 3.5 (2012): 449-454.133. 131. Schulthess, Julie, et al. "The short chain fatty acid butyrate imprints an antimicrobial program in macrophages." Immunity 50.2 (2019): 432-445. 33 132. Osbelt, Lisa, et al. "Variations in microbiota composition of laboratory mice influence Citrobacter rodentium infection via variable short-chain fatty acid production." PLoS pathogens 16.3 (2020): e1008448. 133. Bertin, Yolande, et al. "Lactobacillus reuteri suppresses E. coli O157: H7 in bovine ruminal fluid: Toward a pre-slaughter strategy to improve food safety?." PloS one 12.11 (2017). 134. Walters, Matthew, and Vanessa Sperandio. "Quorum sensing in Escherichia coli and Salmonella." International Journal of Medical Microbiology 296.2-3 (2006): 125-131. 135. Wang, Dandan, Xuedong Ding, and Philip N. Rather. "Indole can act as an extracellular signal in Escherichia coli." Journal of Bacteriology 183.14 (2001): 4210-4216. 136. Schauer, David B., et al. "Genetic and biochemical characterization of Citrobacter rodentium sp. nov." Journal of clinical microbiology 33.8 (1995): 2064-2068. 137. Kumar, Aman, and Vanessa Sperandio. "Indole signaling at the host-microbiota-pathogen interface." mBio 10.3 (2019): e01031-19. 138. Hirakawa, Hidetada, et al. "Secreted indole serves as a signal for expression of type III secretion system translocators in enterohaemorrhagic Escherichia coli O157: H7." Microbiology 155.2 (2009): 541-550. 139. Hirakawa, Hidetada, et al. "In vitro activity of AST-120 that suppresses indole signaling in Escherichia coli, which attenuates drug tolerance and virulence." PloS one 15.4 (2020): e0232461. 140. Hirakawa, Hidetada, et al. "Indole induces the expression of multidrug exporter genes in Escherichia coli." Molecular microbiology 55.4 (2005): 1113-1126. 141. Bansal, Tarun, et al. "Differential effects of epinephrine, norepinephrine, and indole on Escherichia coli O157: H7 chemotaxis, colonization, and gene expression." Infection and immunity 75.9 (2007): 4597-4607. 34 142. Bommarius, Bettina, et al. "A family of indoles regulate virulence and Shiga toxin production in pathogenic E. coli." PLoS One 8.1 (2013). 143. Harrington, Lynne, et al. "Deficiency of indoleamine 2, 3-dioxygenase enhances commensal- induced antibody responses and protects against Citrobacter rodentium-induced colitis." Infection and immunity 76.7 (2008): 3045-3053. 144. Lee, Jintae, et al. "Enterohemorrhagic Escherichia coli biofilms are inhibited by 7- hydroxyindole and stimulated by isatin." Appl. Environ. Microbiol. 73.13 (2007): 4100-4109. 145. Choi, S. H., et al. "Inhibitory effect of skatole (3‐methylindole) on enterohemorrhagic E Escherichia coli O 157: H 7 ATCC 43894 biofilm formation mediated by elevated endogenous oxidative stress." Letters in applied microbiology 58.5 (2014): 454-461. 146. Lee, Jin‐Hyung, Moo Hwan Cho, and Jintae Lee. "3‐Indolylacetonitrile decreases Escherichia coli O157: H7 biofilm formation and Pseudomonas aeruginosa virulence." Environmental microbiology 13.1 (2011): 62-73. 147. Kumar A, Russell RM, Pifer R, et al. The Serotonin Neurotransmitter Modulates Virulence of Enteric Pathogens [published online ahead of print, 2020 Jun 4]. Cell Host Microbe. 2020 148. Schiering, Chris, et al. "Feedback control of AHR signalling regulates intestinal immunity." Nature 542.7640 (2017): 242-245. 149. Wu, Yanbei, et al. "Indole-3-Carbinol Inhibits Citrobacter rodentium Infection through Multiple Pathways Including Reduction of Bacterial Adhesion and Enhancement of Cytotoxic T Cell Activity." Nutrients 12.4 (2020): 917. 35 Chapter 2: Microbial tryptophan metabolites regulate gut barrier function via the aryl hydrocarbon receptor This work is adapted from ‘Scott, Samantha A., Jingjing Fu, and Pamela V. Chang. "Microbial tryptophan metabolites regulate gut barrier function via the aryl hydrocarbon receptor." Proceedings of the National Academy of Sciences U.S.A., 117.32 (2020): 19376-19387.’ 2.1 Abstract Inflammatory bowel diseases (IBDs), including Crohn’s disease and ulcerative colitis, are associated with dysbiosis of the gut microbiome. Emerging evidence suggests that small-molecule metabolites derived from bacterial breakdown of a variety of dietary nutrients confer a wide array of host benefits, including amelioration of inflammation in IBDs. Yet, in many cases, the molecular pathways targeted by these molecules remain unknown. Here, we describe roles for three metabolites—indole-3-ethanol, indole-3-pyruvate, and indole-3-aldehyde—which are derived from gut bacterial metabolism of the essential amino acid tryptophan, in regulating intestinal barrier function. We determined that these metabolites protect against increased gut permeability associated with a mouse model of colitis by maintaining the integrity of the apical junctional complex and its associated actin regulatory proteins, including myosin IIA and ezrin, and that these effects are dependent on the aryl hydrocarbon receptor. Our studies provide a deeper understanding of how gut microbial metabolites affect host defense mechanisms and identify candidate pathways for prophylactic and therapeutic treatments for IBDs. 2.2 Introduction Inflammatory bowel diseases (IBDs), including Crohn’s disease and ulcerative colitis, afflict millions worldwide and greatly affect patient quality of life. These diseases are due to chronic inflammation in the intestines caused by a breakdown in immune homeostasis within the 36 gut (1, 2). Although the exact etiology of this disease remains unknown (3), several factors have been identified that regulate intestinal homeostasis, including the integrity of the gut epithelium (4). This single layer of cells provides a physical and immunological barrier to the intestinal luminal contents, which include the gut microbiota and dietary nutrients. Many IBDs are characterized by compromised gut epithelial barrier function that enables these antigens access to the underlying host tissue in which the immune cells reside. This epithelial leakiness leads to activation of the host immune response and increased inflammation that drives disease pathology. The gut microbiome, or collection of microorganisms that reside in the intestines, plays a major role in regulating immune homeostasis in the intestines (5, 6). These microbes include bacteria, fungi, archaea, and viruses, which are estimated to rival the number of cells in the human body (7). The gut microbiota collectively contains ∼100 times the number of genes in the human genome and encompasses a wide variety of species with microbial densities ranging from 109 cells/mL in the small intestine and 1012 cells/g of luminal content in the large intestine. The gut microbiome modulates myriad physiological functions that greatly affect the overall health of the host, including metabolism and the gut–brain axis (8–10). For example, the gut microbiota break down indigestible dietary fiber and biosynthesize essential nutrients that provide numerous benefits to the host. In addition, these microbes regulate the host immune system by modulating the activities of various immune cell types, ultimately leading to changes in inflammation. Despite increasing evidence that the gut microbiota are critical elements in modulating intestinal immune homeostasis, the mechanisms that regulate many of these pathways remain unknown. Small-molecule metabolites that are produced by the gut microbiota are emerging as 37 ]]]important factors in maintaining intestinal immune homeostasis (11–14). These molecules are biosynthesized by gut microbes and secreted into the intestinal lumen, where they modulate the activities of many host cell types, including intestinal epithelial cells (IECs), and, for metabolites that are able to penetrate the gut epithelium, immune cells within the lamina propria. One of the most widely characterized classes of metabolites are short-chain fatty acids such as acetate, propionate, and butyrate, which derive from bacterial anaerobic fermentation of complex polysaccharides within dietary fiber (15). These metabolites exhibit pleiotropic effects on many different immune cell types, including macrophages, dendritic cells (DCs), B and T lymphocytes, and IECs, and contribute to an anti-inflammatory environment within the gut (11–14). Dietary nutrients, including essential amino acids such as tryptophan (Trp), also serve as substrates for enzymes within the gut microbiota. Bacterial catabolism of Trp produces many metabolites, several of which can modulate aspects of the host immune response, including differentiation of regulatory CD4+ T cells (Tregs), maintenance of intestinal epithelial lymphocytes, activation of natural killer T cells, and protection against mouse models of multiple sclerosis and colorectal cancer (16–20). Additional studies have examined the effects of Trp metabolites in colitis models and have focused on various cell types including DCs, Tregs, goblet cells, and innate lymphoid cells (ILCs) (21–25). Intriguingly, indole-3-pyruvate (IPyA) and indole-3-aldehyde (I3A) can ameliorate mouse models of colitis (24, 25), and these reports focus on the effects of these metabolites on modulating CD103+ DCs, T regulatory 1 (Tr1) cells, and Nkp46+ ILCs. Host-produced Trp metabolites such as kynurenine (Kyn) and the microbially produced Trp metabolites indole (IND) and indole-3-propionate (IPA) have been shown to regulate intestinal barrier function by modulating the gut epithelium via IL-10 receptor (IL10R) expression (26) and alternative pathways, including those activated by Toll-like receptor 4 (TLR4) 38 and the pregnane X receptor (PXR), respectively (27–30). Collectively, these studies have investigated only a limited number of Trp metabolites, and their physiological effects on function of the gut epithelial barrier, which is critical to host–microbiota immune homeostasis, remain largely unexplored for most Trp-derived metabolites. Thus, we were motivated to explore the effects on the gut epithelial barrier of a broader collection of Trp metabolites, whose activities toward intestinal barrier function remain largely uncharacterized. In this study, we have identified and characterized new roles for three Trp metabolites— indole-3-ethanol (IEt) (also known as tryptophol), IPyA, and I3A—in modulating gut barrier integrity via tight junctions (TJs) and adherens junctions (AJs), which together comprise the apical junctional complex (AJC) (31), a major regulator of intestinal permeability. TJ proteins (e.g., ZO1 and occludin) form a barrier against free diffusion at the apical site of adjacent epithelial cell membranes to prevent paracellular transport of molecules between cells (32), and interactions of AJ proteins (e.g., E-cadherin and β-catenin), which are located basolaterally but subjacent to TJs, form adhesive forces between adjacent epithelial cells to seal the intestinal barrier. Here, we have demonstrated that a Trp-rich diet ameliorates morbidity and inflammation using a mouse model of colitis induced by dextran sodium sulfate (DSS) gut epithelial damage. We have determined that the active Trp metabolites produced in vivo are IEt, IPyA, and I3A, while antibiotic treatment to deplete Trp-metabolizing bacteria decreases their production. We also demonstrated that treatment of mice with IEt, IPyA, and I3A individually also improves disease outcomes in DSS colitis. At a mechanistic level, a major portion of the metabolites’ effects is mediated by the aryl hydrocarbon receptor (AhR), which is known to recognize indole containing Trp metabolites (33, 34). Finally, we determined the Trp metabolites affect AJC function by modulating the levels and activity of key proteins including myosin IIA and ezrin that regulate filamentous actin (F-actin). Collectively, 39 these studies provide a mechanistic understanding of how small-molecule Trp metabolites produced by the gut microbiota modulate host processes, including gut barrier function, to benefit the host in diseases such as IBDs that negatively affect the intestinal epithelium. 2.3 Materials and Methods Tissue culture Caco-2 and HEK 293T cells were obtained from the American Tissue Culture Company and cultured according to their guidelines. DMEM, penicillin/streptomycin (P/S), 0.05% Trypsin, and DPBS were obtained from Corning, and Seradigm Premium Grade Fetal Bovine Serum (FBS) and Transfectagro were purchased from VWR. Cell culture Transwell inserts (transparent PET membrane, 12-well, 0.4 µm pore size) and Falcon 12- well companion plates were obtained from BD Falcon. Polybrene and puromycin were purchased from EMD Millipore Sigma. Lipofectamine 2000 was purchased from Thermo Fisher Scientific. Recombinant human TNFα was purchased from R&D Systems, and GNF351 was a kind gift from Avery August. Metabolites L-Tryptophan (Trp) was purchased from Chem-Impex International, Inc. Indole-3- aldehyde (I3A) and indole-3- pyruvate (IPyA) were obtained from Biosynth. Tryptophol (IEt) was obtained from Alfa Aesar, and kynurenine (Kyn) was purchased from Cayman Chemical Company. Indole (IND), indole-3-acetamide (IAM), DL-indole- 3-lactate (ILA), 5- hydroxytryptamine (5HT), and indole-3-acetic acid (IAA) were purchased from Sigma Aldrich. Tryptamine hydrochloride (TrA) was obtained from TCI America, and indole-3-propionate (IPA) was obtained from Alfa Aesar. Indole-3-acrylate (IA) was purchased from Santa Cruz Biotechnology. Histology 40 Hematoxylin was purchased from VWR, and eosin Y was obtained from Acros Organics. Canada Balsam was obtained from Ward’s Science, and xylenes was obtained from Macron Fine Chemicals. Antibiotics Ampicillin and vancomycin hydrochloride were purchased from Sigma Aldrich. Neomycin sulfate hydrate and metronidazole were obtained from Alfa Aesar. Western blotting DC Protein Assay kit and Clarity Western ECL substrate were purchased from BioRad. SuperSignal West Pico Chemiluminescent Substrate was purchased from Thermo Fisher Scientific. Bovine serum albumin (BSA) was purchased from VWR, and non-fat dry milk was purchased from Laboratory Product Sales. Protease inhibitor (cOmplete) tablets were obtained from Roche. Sodium β-glycerophosphate was obtained from Alfa Aesar, and sodium orthovanadate was purchased from MP Biomedicals. Sodium fluoride was purchased from Chem- Impex International, Inc., and sodium pyrophosphate decahydrate was purchased from Fisher Scientific. Immunofluorescence Paraformaldehyde (PFA) 32% solution, EM grade, was purchased from Electron Microscopy Sciences, and DAPI Prolong Diamond was purchased from Thermo Fisher Scientific. Fisher HealthCare Tissue Plus OCT Compound was purchased from Fisher Scientific. Transfection and plasmids PEI was obtained from Polyplus Transfection. pGudluc6.1 and pRLTK were obtained from Gary Perdew (Pennsylvania State University), and pB2X and pRenilla were obtained from Ruslan Medzhitov (Yale University). pCMV-VSV-G, wildtype ezrin, T567A ezrin, and T567E ezrin 41 mutants (C-terminal FLAG tagged) in pQCXIP expression vectors were a kind gift of Anthony Bretscher (Cornell University) LC-MS LC-MS grade methanol, water, acetonitrile, and formic acid (FA) were obtained from Fisher Scientific qPCR RNABee was obtained from Tel-Test, Inc. Diethylpyrocarbonate (DEPC) and chloroform were purchased from Sigma Aldrich. Isopropanol and ethanol were purchased from VWR. Glycogen (Roche) was purchased from Krackeler Scientific. OligodT was purchased from Integrated DNA Technologies (IDT). dNTP was purchased from BioBasic. MMLV reverse transcriptase was purchased from Clontech, and PerfeCta SYBR Green SuperMix, Low ROX, were obtained from Quanta Biosciences. Antibodies Primary antibodies for ZO1 (clone 1A12, 33-9100) and occludin (clone 3F10, 33-1500) were purchased from Thermo Fisher Scientific. Anti-E-cadherin (clone 67A4, 562869) and anti- β-catenin (clone 14, 610154) antibodies were purchased from BD Biosciences. Anti-MLCK (clone K36, M7905) and anti-α-tubulin (clone B512, T5168) antibodies were purchased from EMD Millipore Sigma. Primary antibodies for MLC (clone E-4, sc-28329) was purchased from Santa Cruz Biotechnology and p-MLC (Ser19) (3675) was purchased from Cell Signaling Technology. Anti-ezrin antibody (CPTC-ezrin-1 supernatant) was purchased from the University of Iowa Developmental Studies Hybridoma Bank, and the p-ezrin (T567) rabbit antisera was generously provided by Anthony Bretscher (Cornell University). Anti-mouse horseradish peroxidase (HRP) (170-5947) and anti- rabbit HRP (170-6515) secondary antibodies for Western blot were 42 purchased from BioRad. Donkey anti- mouse Alexa Fluor 594 (A21293), donkey anti-rabbit Alexa Fluor 594 (R37119), and Alexa Fluor 647 Phalloidin (A22287) were purchased from Thermo Fisher Scientific. qPCR primers1 Gene Target Direction Sequence (5-3) species Ahr Human Forward TGGGTCCAGTCTAATGCACG Reverse TGCTCTGTTCCTTCCTCATCT Cyp1a1 Human Forward GTGATCCCAGGCTCCAAGAG Reverse AGAAGAAACTCCGTGGCCG Cyp1a2 Human Forward GCTGAATGGCTTCTACATCCCC Reverse GCGGTGAGGAACCGCTC Cyp1b1 Human Forward AACGTACCGGCCACTATCAC Reverse GCACTCGAGTCTGCACATCA Ahrr Human Forward GCAGCGGAGATGAAAATGAGG Reverse TTCCGATTCGCACAGACTGG Zo1 Human Forward GACGTTTCCCCACTCTGAAA Reverse AGAGCACAGCAATGGAGGAA Rpl13a Human Forward TCCTCCTTTTCCAAGCGGC Reverse GGCCCAGCAGTACCTGTTT Il10r Mouse Forward AGTCTTCAGTTCTCAGGACGC Reverse GCAATGAATTCTAGGCTCAGGC Tnfr Mouse Forward GCTGTTGCCCCTGGTTATCT 43 Reverse ATGGAGTAGACTTCGGGCCT Rpl13a Mouse Forward GCTGCCGAAGATGGCGGAGG Reverse ACCACCACCTTCCGGCCCA 1Primers were purchased from IDT. Mice C57Bl/6 mice were acquired from Jackson Laboratories. Ahr+/- and Ahr-/- mice were obtained from Gary Perdew (Pennsylvania State University). All mice were subsequently bred and maintained at the animal facility of Cornell University and used at 8-12 weeks of age in accordance with the guidelines of the Institutional Animal Care and Use Committee and the Cornell Center for Animal Resources and Education (Protocol number 2015-0069). Mice were co-housed for 7 d prior to use and fed Envigo Teklad global irradiated 18% protein rodent diet meal 2918 as standard chow. Mice were anesthetized with Butler Schein Animal Health Isothesia Isoflurane, USP, 3% vol/vol. Dextran sodium sulfate (DSS) was obtained from Chem-Impex International, Inc., and fluorescein isothiocyanate (FITC)-dextran (FD4, average molecular weight 3,000 – 5,000) was obtained from Sigma Aldrich. In vivo experiments Antibiotic treatment Mice were administered an antibiotic cocktail of ampicillin (9 mg/kg), metronidazole (9 mg/kg), neomycin (9 mg/kg), and vancomycin (4.5 mg/kg) via oral gavage every 12 h. Mice were pre-treated with antibiotics for 7 dprior to receiving high tryptophan diet or metabolites and were continued to be administered the antibiotic cocktail for the remainder of the experiment. High tryptophan diet 44 Standard chow (Envigo Teklad global irradiated 18% protein rodent diet meal 2918) or high tryptophan (Trp) diet was provided to mice ad libitum in feeding jars. The high tryptophan diet was prepared by supplementing standard chow (2 g Trp/kg diet) with an additional 40 g Trp/kg diet. Mice received high tryptophan diet for 7 d prior to DSS treatment and then for the remainder of the experiment. Metabolite treatment Mice were administered I3A (1000 mg/kg), IPyA (2900 mg/kg), or IEt (600 mg/kg). All metabolites were dissolved in dimethylsulfoxide (DMSO) and administered via oral gavage every 12 h. Control mice received equivalent volumes of DMSO via gavage. Mice were pre-treated with metabolite gavage for 2 d prior to DSS treatment and then for the remainder of the experiment. DSS colitis Mice were treated with 3% DSS (wt/vol) in their drinking water (ad libitum) for 7 d. The amounts of water consumed by each treatment group were monitored and compared to control groups. Weight loss Mice were weighed daily at the same time each day. Each mouse weight was normalized to itself and control mice on day 0. 4 kDa FITC-dextran intestinal permeability assay On the last day of the experiment, mice were fasted for 4 h prior to gavage with FD4 (900 mg/kg) in PBS. Prior to euthanasia and 4 h post-gavage, 100 µl of blood was collected via retro- orbital bleed and centrifuged to obtain at least 50 µl of serum. Concentration of FD4 in the serum was determined using a SpectraMax Gemini EM Microplate Reader and a standard curve from a serial dilution of FD4 in PBS. 45 Colon length Entire colons were resected upon euthanasia and measured. Disease activity index Stool consistency (0 = normal, 1 = mildly soft, 2 = soft, 3 = very soft, 4 = diarrhea) and blood in the colon (0 = none, 1 = mildly red, 2 = red, 3 = dark red, 4 = black) were assessed and added together to obtain a disease activity index (DAI) score between 0 and 8. Histopathology A portion of the distal colon was flushed with PBS, excised and fixed in 10% neutral buffered formalin, paraffin-embedded, sectioned (5 μm), and stained with Harris hematoxylin and eosin Y. Samples were blinded, imaged using an Olympus CX41RF microscope, and given a score between 0 and 4, where 0 = normal pathology, 1 = mild, multifocal individual crypt epithelial cell attenuation and goblet cell depletion without inflammatory cell infiltrate, 2 = mild, multifocal crypt epithelial cell loss without inflammatory cell infiltrate, 3 = moderate, multifocal epithelial cell loss with mild lamina propria lymphocytic and neutrophilic infiltrate, and 4 = severe, diffuse surface epithelial cell erosion with extensive crypt epithelial cell necrosis and loss, mild lamina propria lymphocytic and rare neutrophilic infiltrate, and moderate submucosal edema. LC-MS Feces contents were collected fresh from mice and immediately flash frozen in liquid nitrogen. Frozen samples were dried on a VirTis Benchtop K Series Manifold Freeze Dryer. Dried samples were crushed and resolubilized in methanol (10x the volume of the dry weight of the samples) and rocked at room temperature for 1 h before collecting the supernatant, which was then dried down. Immediately prior to LC-MS analysis, the samples were resuspended in methanol (10x the volume of the dry weight of the samples) and filtered. LC-MS analysis was performed on 46 an Agilent 6230 electrospray ionization–time-of-flight (ESI–TOF) MS coupled to an Agilent 1260 HPLC equipped with an Agilent Poroshell 120 ECC18 reverse phase column (3 x 50 mm, 2.7 µm) using a flow rate of 0.5 ml/min. The gradient was ramped from 90% water and 0.1% FA (Solvent A) and 10% acetonitrile and 0.1% FA (Solvent B) to 50% A and 50% B for 0.5 min. The gradient was then ramped to 35% A and 65% B for an additional 0.5 min, then to 15% A and 85% B for 4.5 min, followed by 0% A and 100% B for 0.75 min. The gradient was then held constant at 0% A and 100% B for an additional minute. For detection, the MS was equipped with a dual ESI source operating in positive or negative mode, acquiring in extended dynamic range from m/z 100–3200 at one spectrum per s; gas temperature: 325 °C; drying gas 10 L/min; nebulizer: 20 psi; fragmentor: 80 V. Quantification of metabolites was determined by integrating the extracted ion count of the exact masses of the metabolites, which were determined using commercial standards. Standard curves in which known amounts of metabolite were utilized to determine the amount of each metabolite in each sample, which was normalized to the dry weight of the fecal samples Caco-2 monolayers Cell culture Caco-2 cells were seeded at 15,000 cells/insert in Transwell inserts in 12-well companion plates. Monolayers were grown in DMEM supplemented with 10% FBS and P/S at 37 °C and 5% CO2. Media was replaced every 2-3 days. Treatment On day 18, metabolites were added to the insert (1 mM unless otherwise specified) in DMSO. Equivalent volumes of DMSO were applied to control cells. On day 20, media and metabolites were replaced, and human TNFα (hTNFα, 20 ng/ml) was added to the basolateral 47 chamber. For monolayers treated with GNF351 (1 μM), this inhibitor was added initially to the media on day 17 and replenished on day 18 and 20. Transepithelial electrical resistance Transepithelial electrical resistance (TEER) was measured using a World Precision Instruments EVOM2 Epithelial Voltohmmeter. TEER was normalized to the surface area of the insert, each insert itself, and control inserts on day 0. 4 kDa FITC-dextran permeability assay Monolayers were washed, and media was replaced with Hank’s Balanced Salt Solution (HBSS) 22 h after addition of hTNFα. FD4 was added to the insert at 1 mg/ml in HBSS. After 2 h, 200 µl of HBSS was removed from each well. Concentration of FD4 in the well was determined using a SpectraMax Gemini EM Microplate Reader and a standard curve from a serial dilution of FD4 in HBSS. Ezrin transduction of Caco-2 monolayers HEK 293T cells were seeded at 105 cells/well in 6-well plates to be 90-95% confluent the next day. The next day, cells were washed and media was replaced with Transfectagro and 5 µl PEI (1 mg/ml), 0.33 µg of pCMV- VSV-G, and 1 µg of pQCXIP vector, or wildtype ezrin, T567A or T567E FLAG-tagged mutants in the pQCXIP expression vector were added. The cells were incubated at 37 °C for 4 h, after which media was replaced, and the cells were incubated for an additional 2 d. Afterwards, polybrene was added at 4 µg/ml, and HEK 293T cell supernatant was collected and syringe filtered using a 0.45 µm filter. Filtered supernatant was added drop by drop to Caco-2 cells, and the cells were incubated for 2 d at 37 °C. Transduced cells were then split 1:2 into media containing puromycin (1 µg/ml). After selection, successful transduction was 48 confirmed by anti-FLAG Western blotting, and cells were utilized to set up monolayers as previously described above. Immunofluorescence For mouse samples, a portion of the distal colon was flushed with PBS, excised, and frozen in OCT. OCT blocks were sectioned (5 μm) on a Thermo Scientific Microm HM 525 cryostat, adhered to a glass slide, washed with PBS, and fixed with 4% PFA prior to staining. Caco-2 monolayers were washed with PBS 24 h after addition of hTNFα and fixed with 4% PFA prior to staining. Fixed samples were permeabilized with 0.5% Triton X-100 in PBS at room temperature for 15 min, blocked with 5% BSA in PBS at room temperature for 1 h, and then incubated with the appropriate antibodies in 5% BSA in PBS at room temperature for 2 h. Antibodies against ZO1, occludin, E-cadherin, β-catenin, MLC, MLCK, and p-MLC were each diluted 1:1 in glycerol and used at a dilution of 1:100. The ezrin antibody and p- ezrin antisera were also diluted 1:1 in glycerol and used at a dilution of 1:200. Samples were incubated with appropriate species-specific Alexa Fluor 594 antibodies and Alexa Fluor 647 Phalloidin at a dilution of 1:500 in 5% BSA in PBS in the dark at room temperature for 1 h, then mounted with DAPI Prolong Diamond overnight. Samples were imaged with a Zeiss LSM 800 confocal laser scanning microscope equipped with 20X 0.8 NA and 40X 1.4 NA Plan Apochromat objectives, 405, 488, 561, and 640 nm solid-state lasers, and two GaAsP PMT detectors. Images shown are maximum intensity z-stack projections. Relative brightness of stained cells was quantified using FIJI ImageJ and normalized to control cells. For images of Caco-2 monolayers co-stained for ZO1 and actin, Pearson’s correlation coefficient (R2) was determined using FIJI ImageJ to quantify co- localization. In Caco-2 monolayers stained for tight junction proteins ZO1 or occludin, cells were traced from vertex to 49 vertex using either a freehand or straight line in FIJI ImageJ. The linearity index is the average ratio of the length of the freehand line to the length of the straight line for each vertex-to-vertex segment traced from 10 cells from each of the three replicate inserts per treatment. Western blot For mouse samples, a portion of the distal colon was flushed with PBS. Epithelial cells were scraped off using a pipette tip and lysed in 1X RIPA lysis buffer (150 mM sodium chloride, 1.0% Triton X-100, 0.5% sodium deoxycholate, 0.1% SDS, 25mM Tris, 1 mM EDTA) with 1X protease inhibitor (cOmplete tablets) and sodium β-glycerophosphate (17.5 mM), sodium orthovanadate (1 mM), sodium fluoride (20 mM), and sodium pyrophosphate decahydrate (5 mM) added immediately prior to use. Samples from biological replicates were pooled before quantification by the DC Protein assay kit. Caco-2 monolayers were washed with PBS 24 h after addition of hTNFα, and cells were lysed with 100 µl 1X RIPA lysis buffer with protease and phosphatase inhibitors as above. Samples from biological replicates were pooled prior to quantification as above. Lysates were sonicated using a Heat Systems Ultrasonic Processer XL sonicator. Protein concentrations were determined using the DC Protein Assay kit and BioTek PowerWave XS2 plate reader. Protein concentrations were normalized using 6X Laemmli buffer and ddH2O. Lysates were resolved on polyacrylamide gels and transferred to nitrocellulose. Membranes were blocked with 5% BSA in 25 mM Tris, 150 mM sodium chloride, and 0.1% Tween-20 solution (TBS-T) rocking at room temperature for 1 h, then probed with the appropriate antibodies in 5% milk or BSA in TBS-T with 0.05% sodium azide with rocking at 4 °C overnight. Occludin, β- catenin, and MLC primary antibodies were each diluted 1:1 with glycerol, then used at a 1:500 dilution in 5% BSA in TBS-T for Western blotting. E-cadherin, MLCK, and p-MLC primary antibodies were 50 each diluted 1:1 with glycerol, then used at a 1:500 dilution in 5% milk in TBS-T. The ezrin antibody and p-ezrin antisera were diluted 1:1 in glycerol and used at 1:2,000 and 1:1,000, respectively, in 5% BSA in TBS-T. The antibodies for α-tubulin and GAPDH were used at 1:10,000 and 1:2,500, respectively, in 5% BSA in TBS-T. Overnight incubation in primary antibody was followed by washing and incubation with appropriate species-specific HRP antibody diluted at 1:10,000 in 5% milk in TBS-T with rocking at room temperature for 1 h. Western blots were developed using chemiluminescence detection reagents on a BioRad ChemiDoc MP. Occludin, E-cadherin, β-catenin, MLC, MLCK, ezrin, p-ezrin, α-tubulin, and GAPDH Western blots were developed using BioRad Clarity Western ECL substrate, and p-MLC Western blots were developed using SuperSignal West Pico Chemiluminescent Substrate. Densitometry was performed using FIJI ImageJ and normalized to the housekeeping protein (e.g., GAPDH) and control lysate bands. RNA isolation and qPCR analysis For mouse samples, intestinal epithelial cells were collected from the distal colon, and RNA was purified using RNABee (300 µl) according to the manufacturer’s instructions and quantified using a GE Nanovue. Using a BioRad C1000 Touch Thermal Cycler, RNA was reverse transcribed using an oligo(dT) primer and MMLV reverse transcriptase. cDNA samples were analyzed using PerfeCta SYBR Green SuperMix, Low ROX, and a BioRad CFX96 Real- Time PCR Detection System. PCR amplification conditions were as follows: 95 °C (3 min) and 40 cycles of 95°C (15 s) and 60 °C (45 s). Relative expression of transcripts was normalized to the ribosomal protein L13a (Rpl13a). Data are represented as the fold induction over control samples. Caco-2 monolayers were washed with PBS 24 h after addition of hTNFα, and cells were lysed with 1 ml RNABee per insert prior to RNA purification and qPCR as above. 51 AhR and NF-kB luciferase activity assays Passive lysis buffer (5X PLB) was prepared with 125 mM Tris, pH 7.8, 10 mM 1,2- diaminocyclohexane tetraacetic acid (CDTA), 10 mM DTT, 5 mg/mL BSA, 5% (vol/vol) Triton X-100, and 50% (vol/vol) glycerol in ddH2O. An aqueous solution of 1X firefly luciferase substrate was prepared containing 75 mM HEPES, pH 8.0, 4 mM MgSO4, 20 mM DTT, 0.1 mM EDTA, 0.53 mM ATP, 0.27 mM coenzyme A, and 0.47 mM D-luciferin (firefly) in ddH2O. An aqueous solution of 1X Renilla luciferase buffer was prepared containing 7.5 mM sodium acetate, pH 5.0, 400 mM sodium sulfate, 10 mM CDTA, 15 mM sodium pyrophosphate, and 0.025 mM 2- (4-aminophenyl)-6-methylbenzothiazole. A 100X Renilla luciferase substrate was prepared by diluting coelenterazine to 0.55 mM in anhydrous methanol and added to 1X Renilla luciferase buffer immediately prior to the assay. AhR activity assay HEK 293T cells were plated at 105 cells/well in a 24-well plate. At 50-60% confluency, the cells were transfected overnight with 1 µg of pGudluc6.1 and 100 ng of pRLTK per well. The next day, media was replaced, and cells were treated for 24 h with I3A, IPyA, or IEt (1 mM) or with an equivalent volume of DMSO. After 24 h, cells were washed with PBS, lysed with 1X PLB, and 20 µl of lysate was added to a 96-well white opaque plate. Afterwards, 50 µl of 1X firefly luciferase substrate was added to each well, and luminescence was measured for 10 min using a Turner BioSystems Veritas Microplate Luminometer. Immediately after, 50 µl of the 1X Renilla substrate was added to each well, and luminescence was measured for 10 min. Luciferase activity was determined by subtracting a blank and calculating the ratio of the firefly luciferase signal to the Renilla luciferase signal. NF-kB activity assay 52 HEK 293T cells were plated at 105 cells/well in a 12-well plate. At 80-90% confluency, the cells were transfected overnight with 1 µg of pB2X and 0.1 µg of pRenilla per well. The next day, transfected cells were re-plated at 104 cells/well into a 96-well plate and incubated overnight. The next morning, cells were treated for 3 h with I3A, IPyA, or IEt (1 mM) or with an equivalent volume of DMSO. After 3 h, TNFα (10 ng/ml) was added to appropriate wells for 3 h. After 3 h, cells were washed with PBS, lysed with 1X PLB, and 20 µl of lysate was added to a 96-well white opaque plate. Afterwards, 50 µl of 1X firefly luciferase substrate was added to each well, and luminescence was measured for 10 min using a Turner BioSystems Veritas Microplate Luminometer. Immediately after, 50 µl of the 1X Renilla substrate was added to each well, and luminescence was measured for 10 min. Luciferase activity was determined by subtracting a blank and calculating the ratio of the firefly luciferase signal to the Renilla luciferase signal. Statistical analysis Experiments were completed at least three independent times. Error bars signify standard deviation from the mean. Statistical significance was determined using one-way ANOVA followed by post-hoc Tukey’s test. 2.4 Results 2.4.1 Trp Feeding Ameliorates DSS Colitis in Mice and Is Dependent on the Gut Microbiota. We began by examining the effects of a Trp-rich diet on morbidity and inflammation using a mouse model of colitis (Figure 2.1A). As measured by changes in weight loss, colon length, disease activity index, and histopathology, we found that Trp feeding exhibited protective effects (Figure 2.1 B and D–G). A major feature of DSS colitis that leads to this morbidity and inflammation is an increase in permeability of the gut epithelium. Importantly, Trp feeding during DSS colitis also 53 attenuated the increase in gut permeability, as measured by flux of the fluorescent tracer fluorescein isothiocyanate (FITC)-dextran from the intestinal lumen to the blood (Figure 2.1C). We further found that the improvement in intestinal permeability is likely due to modulation of TJ and AJ by immunofluorescence (IF) staining of key protein components of these structures within the intestinal tissues (Figure 2.1 H–M and 2.2). Aspects of these experiments confirm previously reported findings that Trp feeding improves disease outcomes in DSS colitis (35–38). In those studies, however, effects of Trp feeding were attributed to immune pathways, including production of prostaglandin E2 (PGE2), and gut permeability due to changes in AJC expression was not addressed (36). To examine whether the effects of the Trp-rich diet are due to bacterial catabolism of Trp, we also pretreated certain groups of mice with a mixture of antibiotics containing ampicillin, metronidazole, neomycin, and vancomycin (ABX) to deplete bacteria that contribute to Trp metabolite production. For all the above-described readouts of physiological function during DSS colitis, we found that ABX treatment eliminated the effects of Trp feeding (Figure 2.1). 54 Figure 2.1 Dietary Trp ameliorates a mouse model of colitis in a microbiome-dependent manner. (A) C57BL/6 mice were pretreated with antibiotics (ABX: ampicillin [9 mg/kg], metronidazole [9 mg/kg], neomycin [9 mg/kg], and vancomycin [4.5 mg/kg], intragastrically) for 7 d and then fed a Trp-rich diet (42 g Trp/kg diet) or standard chow (2 g Trp/kg diet) for 7 d, followed by administration of DSS (3%, wt/vol) or vehicle for 7 d (ad libitum) with continued antibiotic treatment and Trp feeding. (B) The mice were weighed daily. (C) Mice were orally gavaged with FITC-dextran (900 mg/kg) on day 14, and serum levels of FITC- dextran were measured 4 h later. (D) On day 14, the mice were euthanized, and colon lengths were measured, (E) disease activity index was measured, and (F and G) the distal colon was stained with hematoxylin and eosin (H&E) and blindly scored (0 = none, 1 = very mild, 2 = mild, 3 = moderate, 4 = severe) for epithelial damage, mononuclear and polymorphonuclear infiltrate, and submucosal edema. (Scale bar: 50 μm.) (H–M) Colon sections were stained for TJ and AJ proteins and imaged by confocal microscopy (see also Figure A1.1, for occludin and β-catenin). (Scale bars: 20 μm.) (J–M) Relative (rel.) brightness of images with error as SD from the mean was calculated (n = 15). (H and J) ZO1. (I and L) E-cadherin. (K) Occludin. (M) β-catenin. Data are representative of at least three 55 independent experiments; n = 5 mice per group. One-way ANOVA followed by post hoc Tukey’s test: **P < 0.01, ***P < 0.001, n.s.: not significant. Figure 2.2 Microbial tryptophan (Trp) metabolites reduce disassembly of tight (TJ) and adherens junction (AJ) proteins within the intestines using a mouse model of colitis. C57Bl/6 mice were pre-treated with antibiotics (ABX: ampicillin (9 mg/kg), metronidazole (9 mg/kg), neomycin (9 mg/kg), and vancomycin (4.5 mg/kg), i.g.) for 7 d, then fed a Trp-rich diet (42 g Trp/kg diet) or standard chow (2 g Trp/kg diet) for 7 d, followed by administration of dextran sodium sulfate (DSS, 3%, wt/vol) or vehicle for 7 d (ad libitum) with continued antibiotic treatment and Trp feeding. Colon sections were stained for (A) occludin and (B) -catenin and imaged by confocal microscopy. Data are representative of at least 3 independent experiments, n = 5 mice per group. 2.4.2 Specific Trp Metabolites Improve Intestinal Barrier Function In Vitro. To determine the identities of the Trp metabolites that mediate this protective effect, we performed an in vitro screen of Trp and 12 of its metabolites using the human Caco-2 IEC cell line, which forms a polarized monolayer that resembles the gut epithelium (39). We found that IEt, IPyA, and I3A (50 μM to 1 mM) all attenuated the increase in epithelial permeability caused by treatment with a proinflammatory cytokine, TNFα, an inflammatory factor associated with IBDs, in a dose-dependent manner as measured by changes in trans-epithelial electrical resistance (TEER) and FITC-dextran flux (Figure 2.3). Upon examination of TJ protein localization by IF, we also found that TNFα treatment causes a striking sinuous phenotype wherein the Caco-2 cells become jagged, as opposed to the typical smooth, chicken-wire–like appearance of monolayers (Figure 2.4). This sinuous phenotype, which was previously reported (40), is likely due to an 56 increased contractility of the actin cytoskeleton as indicated by increased colocalization of F-actin with TJ proteins (Figure 2.4 C and H). Importantly, IEt, IPyA, and I3A all prevent this effect, which does not depend on changes in messenger RNA transcript or protein levels of the TJ and AJ proteins (Figures 2.4 and 2.5). The remaining Trp metabolites that were tested did not have any effects on these phenotypes, so it is unlikely that they are the active Trp metabolites for this phenotype (Figures 2.6 and 2.7). 57 Figure 2.3 Tryptophan metabolites indole-3-aldehyde (I3A), indole-3-pyruvate (IPyA), and indole-3-ethanol (IEt) decrease epithelial permeability. Polarized Caco-2 monolayers were pre- treated with metabolite for 2 d at the indicated concentrations of metabolite or vehicle (0.1% DMSO) and then stimulated with TNF (20 ng/ml) at the indicated times. Epithelial permeability was measured by (A-C) transepithelial electrical resistance (TEER) or (D-F) FITC-dextran paracellular transport from the apical to basolateral compartment. Data are representative of at least 3 independent experiments. One-way ANOVA followed by post-hoc Tukey’s test: n = 3, *p<0.05, ***p<0.001, n.s. = not significant. 58 Figure 2.4 Tryptophan metabolites indole-3-aldehyde (I3A), indole-3-pyruvate (IPyA), and indole-3-ethanol (IEt) decrease epithelial permeability. (A-L) Polarized Caco-2 monolayers were pre-treated with metabolite (1 mM) for 2 d and then stimulated with TNF (20 ng/ml) for 24 h, after which the cells were stained with the indicated antibodies and imaged by confocal microscopy. (C) Magenta = ZO1, green = actin, white = merged signal. Insets: 2.5x magnification. Scale bars = 20 m. (G) Linearity index for ZO-1 was calculated (n = 30). (H) Pearson’s correlation coefficient (R2) was calculated for ZO1 and actin. (I-L) Relative (rel.) brightness of images with error as standard deviation from the mean was calculated (n = 15). Data are representative of at 59 least 3 independent experiments. One-way ANOVA followed by post-hoc Tukey’s test: n = 3, **p<0.01, n.s. = not significant. Figure 2.5 Tryptophan metabolites indole-3-aldehyde (I3A), indole-3-pyruvate (IPyA), and indole-3-ethanol (IEt) do not change levels of mRNA transcripts or protein expression of TJ and AJ proteins. Polarized Caco-2 monolayers were pre-treated with metabolite (1 mM) for 2 d and then stimulated with TNF (20 ng/ml) for 24 h. (A,F,K) mRNA was isolated and analyzed by qPCR for Zo1. (B-E, G-J, L-O) Protein lysates were generated and analyzed by Western blotting for the indicated proteins and quantified by densitometry. Data are representative of at least 3 independent experiments. One-way ANOVA followed by post-hoc Tukey’s test: n = 3, n.s. = not significant. 60 Figure 2.6 Chemical structures of tryptophan (Trp) metabolites that were tested and did not affect epithelial permeability in this study. Abbreviations: tryptophan (Trp), indole (IND), kynurenine (Kyn), indole-3-acetamide (IAM), indole-3-lactate (ILA), tryptamine (TrA), indole-3-acetic acid (IAA), indole-3-propionate (IPA), indole-3-acrylate (IA), 5-hydroxytryptamine (5HT). 61 Figure 2.7 Additional tryptophan metabolites do not affect epithelial permeability. Polarized Caco-2 monolayers were pre-treated with indicated metabolite (1 mM) for 2 d and then stimulated with TNF (20 ng/ml). Epithelial permeability was measured by (A-J) TEER at the indicated times or (K-T) FITC-dextran paracellular transport from the apical to basolateral compartment after 4 h. Abbreviations: tryptophan (Trp), indole (IND), kynurenine (Kyn), indole-3-acetamide (IAM), indole-3-lactate (ILA), tryptamine (TrA), indole-3-acetic acid (IAA), indole-3-propionate (IPA), indole-3-acrylate (IA), 5-hydroxytryptamine (5HT). Data are representative of at least 3 independent experiments. One-way ANOVA followed by post-hoc Tukey’s test: n = 3, ***p<0.001, n.s. = not significant. 2.4.3 Trp Metabolites Improve Disease Outcomes in DSS Colitis in Mice. Based on these in vitro studies, we hypothesized that IEt, IPyA, and I3A would ameliorate DSS colitis in mice. We first verified that these three metabolites are indeed produced from Trp catabolism during DSS colitis by targeted mass spectrometry (MS)-based metabolomics using liquid chromatography-MS (LC-MS) (Figure 2.8). Importantly, these studies enabled us to 62 establish therapeutically relevant doses. We then administered the same ABX treatment to deplete the gut microbiota and prevent further microbial metabolism of IEt, IPyA, and I3A, followed by giving each metabolite to mice by oral gavage prior to the induction of DSS colitis and examined changes in weight loss, colon length, disease activity index, and histopathology as above (Figure 2.9). All three of these metabolites ameliorated each of these measures of disease severity in mice. Furthermore, IEt, IPyA, and I3A also prevented increased intestinal permeability (Figure 2.9C) and disassembly of the AJC during DSS colitis, as established by IF staining of the TJ and AJ proteins (Figures 2.10–1.12). We also examined the effects of IEt, IPyA, and I3A on TNFR1 expression in IECs, which increases in DSS colitis, and found that these metabolites inhibit expression of this receptor (Figure 2.13 A–C). We further examined an important downstream target of TNFR1, NF-κB activation, using a reporter cell line, and found that these metabolites inhibit activity of this important transcription factor that is responsible for the transcription of many inflammatory mediators (Figure 2.11D). In addition, we determined the effects of IEt, IPyA, and I3A on IL-10R expression in IECs, which decreases in DSS colitis, and found that these metabolites increase expression when administered alone and in combination with DSS in vivo (Figure 2.14). We also found that oral gavage of IEt, IPyA, and I3A led to sustained levels of each metabolite within the intestinal lumen during DSS colitis by LC-MS analysis (Figures 2.15–2.17). 63 Figure 2.8 LC-MS quantification of tryptophan (Trp) metabolites during Trp-rich diet in mouse model of colitis. C57Bl/6 mice were pre-treated with antibiotics (ABX: ampicillin (9 mg/kg), metronidazole (9 mg/kg), neomycin (9 mg/kg), and vancomycin (4.5 mg/kg), i.g.) for 7 d, then fed a Trp-rich diet (42 g Trp/kg diet) or standard chow (2 g Trp/kg diet) for 7 d, followed by administration of dextran sodium sulfate (DSS, 3%, wt/vol) or vehicle for 7 d (ad libitum) with continued antibiotic treatment and Trp feeding. Metabolites from the fecal colonic contents were measured using mass spectrometry using commercial standards. Data are representative of at least 3 independent experiments, n = 5-10 mice per group. One-way ANOVA followed by post- hoc Tukey’s test: *p<0.05, **p<0.01, ***p<0.001, n.s. = not significant. 64 65 Figure 2.9 Tryptophan metabolites I3A, IPyA, and IEt ameliorate a mouse model of colitis. (A) C57BL/6 mice were pretreated with antibiotics (ABX: ampicillin [9 mg/kg], metronidazole [9 mg/kg], neomycin [9 mg/kg], and vancomycin [4.5 mg/kg], intragastrically) for 7 d, followed by I3A (1,000 mg/kg), IPyA (2,900 mg/kg), or IEt (600 mg/kg) for 2 d and then administered DSS (3%, wt/vol) for 7 d (ad libitum) with continued antibiotic and metabolite treatment. (B) The mice were weighed daily. (C) Mice were orally gavaged with FITC-dextran (900 mg/kg) on day 9, and serum levels of FITC-dextran were measured 4 h later. (D) On day 9, the mice were euthanized, and colon lengths were measured, (E) disease activity index was measured, and (F and G) the distal colon was stained with H&E and blindly scored (0 = none, 1 = very mild, 2 = 66 mild, 3 = moderate, 4 = severe) for epithelial damage, mononuclear and polymorphonuclear infiltrate, and submucosal edema. (Scale bar: 50 μm.) Data are representative of at least three independent experiments; n = 5 mice per group. One-way ANOVA followed by post hoc Tukey’s test: ***P < 0.001, n.s.: not significant. Figure 2.10 Tryptophan metabolite indole-3-aldehyde (I3A) prevents disruption of TJ and AJ proteins in a mouse model of colitis. C57Bl/6 mice were pre-treated with antibiotics (ABX: ampicillin (9 mg/kg), metronidazole (9 mg/kg), neomycin (9 mg/kg), and vancomycin (4.5 mg/kg), i.g.) for 7 d, followed by I3A (1000 mg/kg) for 2 d, and then administered DSS (3%, wt/vol) for 7 d (ad libitum) with continued antibiotic and metabolite treatment. (A-H) Colon sections were stained for TJ and AJ proteins and imaged by confocal microscopy. Scale bars = 20 m. (E-H) Relative (rel.) brightness of images with error as standard deviation from the mean was calculated (n = 15). Data are representative of at least 3 independent experiments, n = 5 mice per group. One- way ANOVA followed by post-hoc Tukey’s test: ***p<0.001, n.s. = not significant. 67 Figure 2.11 Tryptophan metabolite indole-3-pyruvate (IPyA) prevents disruption of TJ and AJ proteins in a mouse model of colitis. C57Bl/6 mice were pre-treated with antibiotics (ABX: ampicillin (9 mg/kg), metronidazole (9 mg/kg), neomycin (9 mg/kg), and vancomycin (4.5 mg/kg), i.g.) for 7 d, followed by IPyA (2900 mg/kg) for 2 d, and then administered DSS (3%, wt/vol) for 7 d (ad libitum) with continued antibiotic and metabolite treatment. (A-H) Colon sections were stained for TJ and AJ proteins and imaged by confocal microscopy. Scale bars = 20 m. (E-H) Relative (rel.) brightness of images with error as standard deviation from the mean was calculated (n = 15). Data are representative of at least 3 independent experiments, n = 5 mice per group. One- way ANOVA followed by post-hoc Tukey’s test: ***p<0.001, n.s. = not significant. 68 Figure 2.12 Tryptophan metabolite indole-3-ethanol (IEt) prevents disruption of TJ and AJ proteins in a mouse model of colitis. C57Bl/6 mice were pre-treated with antibiotics (ABX: ampicillin (9 mg/kg), metronidazole (9 mg/kg), neomycin (9 mg/kg), and vancomycin (4.5 mg/kg), i.g.) for 7 d, followed by IEt (600 mg/kg) for 2 d and then administered DSS (3%, wt/vol) for 7 d (ad libitum) with continued antibiotic and metabolite treatment. (A-H) Colon sections were stained for TJ and AJ proteins and imaged by confocal microscopy. Scale bars = 20 m. (E-H) Relative (rel.) brightness of images with error as standard deviation from the mean was calculated (n = 15). Data are representative of at least 3 independent experiments, n = 5 mice per group. One-way ANOVA followed by post-hoc Tukey’s test: ***p<0.001, n.s. = not significant. 69 Figure 2.13 Tryptophan metabolites indole-3-aldehyde (I3A), indole-3-pyruvate (IPyA), and indole-3-ethanol (IEt) inhibit TNFR1 signaling. (A-C) C57Bl/6 mice were pre-treated with antibiotics (ABX: ampicillin (9 mg/kg), metronidazole (9 mg/kg), neomycin (9 mg/kg), and vancomycin (4.5 mg/kg), i.g.) for 7 d, followed by IEt (600 mg/kg), IPyA (2900 mg/kg), or I3A (1000 mg/kg) for 2 d and then administered DSS (3%, wt/vol) for 7 d (ad libitum) with continued antibiotic and metabolite treatment. RNA from intestinal epithelial cells (IECs) was harvested and analyzed for Tnfr1 levels by qPCR. (D) HEK 293T cells expressing luciferase downstream of NF- kB binding site were pre-treated with indicated metabolite (1 mM) for 3 h and then stimulated with TNF (10 ng/ml) for an additional 3 h before luminescence was measured. Data are representative of at least 3 independent experiments, (A-C) n = 5 mice per group, (D) n = 3. One-way ANOVA followed by post-hoc Tukey’s test: *p<0.05, **p<0.01, ***p<0.001, n.s. = not significant. Figure 2.14 Tryptophan metabolites indole-3-aldehyde (I3A), indole-3-pyruvate (IPyA), and indole-3-ethanol (IEt) induce IL-10R in DSS colitis. C57Bl/6 mice were pre-treated with antibiotics (ABX: ampicillin (9 mg/kg), metronidazole (9 mg/kg), neomycin (9 mg/kg), and vancomycin (4.5 mg/kg), i.g.) for 7 d, followed by (A) I3A (1000 mg/kg), (B) IPyA (2900 mg/kg), or (C) IEt (600 mg/kg) for 2 d and then administered DSS (3%, wt/vol) for 7 d (ad libitum) with continued antibiotic and metabolite treatment. RNA from IECs was harvested and analyzed for Il10r levels by qPCR. Data are representative of at least 3 independent experiments, n = 5 mice per group. One-way ANOVA followed by post-hoc Tukey’s test: ***p<0.001, n.s. = not significant. 70 Figure 2.15 LC-MS quantification of Trp metabolites after oral gavage of I3A during mouse model of colitis. C57Bl/6 mice were pre-treated with I3A (1000 mg/kg) for 2 d and then administered DSS (3%, wt/vol) for 7 d (ad libitum) with continued metabolite treatment. Metabolites from the fecal colonic contents were measured using mass spectrometry using commercial standards. Data are representative of at least 3 independent experiments, n = 5-10 mice per group. One-way ANOVA followed by post-hoc Tukey’s test: ***p<0.001, n.s. = not significant. 71 Figure 2.16 LC-MS quantification of Trp metabolites after oral gavage of IPyA during mouse model of colitis. C57Bl/6 mice were pre-treated with IPyA (2900 mg/kg) for 2 d and then administered DSS (3%, wt/vol) for 7 d (ad libitum) with continued metabolite treatment. 72 Metabolites from the fecal colonic contents were measured using mass spectrometry using commercial standards. Data are representative of at least 3 independent experiments, n = 5-10 mice per group. One-way ANOVA followed by post-hoc Tukey’s test: *p<0.05, **p<0.01, ***p<0.001, n.s. = not significant. 73 Figure 2.17 LC-MS quantification of Trp metabolites after oral gavage of IEt during mouse model of colitis. C57Bl/6 mice were pre-treated with IEt (600 mg/kg) for 2 d and then administered DSS (3%, wt/vol) for 7 d (ad libitum) with continued metabolite treatment. Metabolites from the fecal 74 colonic contents were measured using mass spectrometry using commercial standards. Data are representative of at least 3 independent experiments, n = 5-10 mice per group. One-way ANOVA followed by post-hoc Tukey’s test: **p<0.01, ***p<0.001, n.s. = not significant. 2.4.4 Trp Metabolites Are Ligands of the AhR. We next set out to establish the molecular mechanism by which the Trp metabolites exert their protective effects on gut permeability in DSS colitis. We hypothesized that IEt, IPyA, and I3A are recognized by AhR, a host receptor that is activated in several cell types by many indole- containing ligands, including IPyA and I3A (24, 25), and that AhR may mediate the effects of these metabolites during DSS colitis. We tested this hypothesis by first assessing whether IEt, IPyA, and I3A activate AhR and found that all three metabolites activate this receptor, using a luciferase reporter assay in HEK 293T cells (Figure A1.18A). We then examined effects of these metabolites on AhR-dependent gene expression and found increased transcription of AhR and canonical downstream target genes in polarized Caco-2 monolayers (Figure A1.18 B–F). We also found that an inhibitor of AhR, GNF351, partially blocks the effects of IPyA in Caco-2 monolayers challenged with TNFα, based on an examination of TEER, FITC-dextran flux, and IF of the AJC proteins (Figure A1.18 G–P). 75 Figure 2.18 Tryptophan metabolites I3A, IPyA, and IEt are aryl hydrocarbon receptor (AhR) ligands, and the activity of IPyA can be blocked with an AhR inhibitor. (A) AhR luciferase reporter cell line was treated with each metabolite for 24 h, and luminescence was measured. RLU = relative luminescence units. (B-F) Polarized Caco-2 monolayers were treated with I3A, IPyA, and IEt (1 mM) for 8 h, and mRNA levels of downstream AhR target genes (indicated) were analyzed by qPCR. (G-H) Polarized Caco-2 monolayers were pre-treated with the AhR antagonist GNF351 76 for 24 h. Next, IPyA was added 2 d prior to stimulation with TNF (20 ng/ml). Epithelial permeability was measured by (G) TEER and (H) FITC-dextran flux. (I-P) Samples were stained for TJ and AJ proteins indicated. Scale bars = 20 m. (M-N) Linearity indices for ZO1 and occludin were calculated (n = 30). (O-P) Relative (rel.) brightness of images with error as standard deviation from the mean was calculated (n = 15). Data are representative of at least 3 independent experiments. One-way ANOVA followed by post-hoc Tukey’s test: n = 3, **p<0.01, ***p<0.001, n.s. = not significant. 2.4.5 Effect of Trp Metabolites Is Partially Dependent on AhR In Vivo. To examine whether the physiological effects of IEt, IPyA, and I3A toward gut barrier integrity are mediated by AhR, we treated Ahr−/− or Ahr+/− littermate control mice with a Trp- rich diet and then challenged the mice with DSS colitis (Figure 2.19A). We found that Trp feeding decreases morbidity and inflammation in Ahr+/− mice but not in Ahr−/− mice, as determined by changes in weight loss, colon length, disease activity index, and histopathology (Figure 2.19 B and D–G). The improvements in gut permeability (Figure 2.19 C) and disassembly of the AJC within intestinal tissues conferred by these metabolites during DSS colitis were also abrogated in the Ahr−/− mice compared to their littermate controls (Figure 2.19 H–Q and 2.20). 77 Figure 2.19 Effect of a high-Trp diet in a mouse model of colitis is dependent on AhR. (A) Ahr+/− or Ahr−/− mice were fed a Trp-rich diet (42 g Trp/kg diet) or standard chow (2 g Trp/kg diet) for 7 d, followed by administration of DSS (3%, wt/vol) or vehicle for 7 d (ad libitum) with continued Trp feeding. (B) The mice were weighed daily. (C) Mice were orally gavaged with FITC-dextran (900 mg/kg) on day 14, and serum levels of FITC-dextran were measured 4 h later. (D) On day 14, the mice were euthanized, and colon lengths were measured, (E) disease activity index was measured, and (F and G) the distal colon was stained with H&E and blindly scored (0 = none, 1 = very mild, 2 = mild, 3 = moderate, 4 = severe) for epithelial damage, mononuclear and polymorphonuclear infiltrate, and submucosal edema. (Scale bar: 50 μm.) (H–M) Colon sections were stained for TJ and AJ proteins and imaged by confocal microscopy (see also Figure A1.17, for occludin and β-catenin). (Scale bars: 20 μm.) (J–M) Relative (rel.) brightness of images with 78 error as SD from the mean was calculated (n = 15). (H and J) ZO1. (I and L) E-cadherin. (K) Occludin. (M) β-catenin. (N–Q) TJ and AJ protein levels were determined by Western blotting with the indicated antibodies and (O–Q) quantified by densitometry (n = 3). Data are representative of at least three independent experiments; n = 5 mice per group. One-way ANOVA followed by post hoc Tukey’s test: **P < 0.01, ***P < 0.001, n.s.: not significant. Figure 2.20 Effect of high tryptophan (Trp) diet in decreasing TJ and AJ disassembly in a mouse model of colitis is dependent on the aryl hydrocarbon receptor (AhR). Ahr+/- or Ahr-/- mice were fed a Trp-rich diet (42 g Trp/kg diet) or standard chow (2 g Trp/kg diet) for 7 d, followed by administration of dextran sodium sulfate (DSS, 3%, wt/vol) or vehicle for 7 d (ad libitum) with continued Trp feeding. Colon sections were stained for (A) occludin and (B) -catenin and imaged by confocal microscopy. Scale bars = 20 m. Data are representative of at least 3 independent experiments, n = 5 mice per group. Importantly, we found no difference in the effects of Trp feeding during DSS colitis in Ahr+/+ compared to Ahr+/− mice, validating the use of Ahr+/− mice as controls for the above studies (Figure 2.21). To assess the extent to which IEt, IPyA, and I3A are the active Trp metabolites produced in vivo during DSS colitis, we preadministered these metabolites to the mice and found that the effects of each metabolite are partially dependent on AhR (Figure 2.22 and 2.23, IEt; 2.24, IPyA; and 2.25, I3A). In these studies, we administered IEt at therapeutically relevant levels as determined in Figure 2.8 (Figure 2.22) to Ahr+/− and Ahr−/− mice and found that it ameliorated morbidity and inflammation in Ahr+/− mice but not in Ahr−/− mice, as determined by changes in weight loss, colon length, disease activity index, and histopathology (Figure 2.22 A, B, and D–G). We also determined that IEt improved intestinal permeability (Figure 2.22C) via 79 modulation of the AJC in intestinal tissues within these mice, and these effects were also dependent on AhR (Figure 2.22 H–Q and 2.23). In parallel experiments, we found similar effects with therapeutically relevant amounts of IPyA and I3A (as determined in Figure 2.8) in improving disease outcomes during DSS colitis and intestinal permeability in these mice due to modulation of the TJ and AJ proteins (Figures 2.24 and 2.25). Together, these studies suggest that these metabolites are the active ones produced during Trp supplementation that confer the majority of the effects on intestinal epithelial barrier function that are dependent on AhR in a mouse model of DSS colitis. 80 Figure 2.21 Effect of high tryptophan (Trp) diet in mouse model of colitis does not differ in AhR wildtype versus heterozygote mice. (A) Ahr+/+ or Ahr+/- mice were fed a Trp-rich diet (42 g Trp/kg diet) or standard chow (2 g Trp/kg diet) for 7 d, followed by administration of dextran sodium sulfate (DSS, 3%, wt/vol) or vehicle for 7 d (ad libitum) with continued Trp feeding. (B) The mice were weighed daily. (C) Mice were orally gavaged with FITC-dextran (900 mg/kg) on day 14, and serum levels of FITC-dextran were measured 4 h later. (D) On day 14, the mice were sacrificed, and colon lengths were measured, (E) disease activity index was measured, and (F-G) 81 the distal colon was stained with H&E and blindly scored (0 = none, 1 = very mild, 2 = mild, 3 = moderate, 4 = severe) for epithelial damage, mononuclear and polymorphonuclear infiltrate, and submucosal edema. Scale bar = 50 m. (H-O) Colon sections were stained for TJ and AJ proteins (indicated) and imaged by confocal microscopy. Scale bars = 20 m. (L-O) Relative (rel.) brightness of images with error as standard deviation from the mean was calculated (n = 15). (H,L) ZO1; (I,M) Occludin; (J,N) E-cadherin; (K,O) -catenin. Data are representative of at least 3 independent experiments, n = 5 mice per group. One-way ANOVA followed by post-hoc Tukey’s test: *p<0.05, ***p<0.001, n.s. = not significant. Figure 2.22 Effect of tryptophan metabolite IEt in mouse model of colitis is dependent on AhR. (A) Ahr+/− or Ahr−/− mice were pretreated with IEt (600 mg/kg) for 2 d and then administered 82 DSS (3%, wt/vol) for 7 d (ad libitum) with continued metabolite treatment. (B) The mice were weighed daily. (C) Mice were orally gavaged with FITC-dextran (900 mg/kg) on day 9, and serum levels of FITC-dextran were measured 4 h later. (D) On day 9, the mice were euthanized, and colon lengths were measured, (E) disease activity index was measured, and (F and G) the distal colon was stained with H&E and blindly scored (0 = none, 1 = very mild, 2 = mild, 3 = moderate, 4 = severe) for epithelial damage, mononuclear and polymorphonuclear infiltrate, and submucosal edema. (Scale bar: 50 μm.) (H–M) Colon sections were stained for TJ and AJ proteins and imaged by confocal microscopy (see also Figure A1.19 for occludin and β-catenin). (Scale bars = 20 μm.) (J–M) Relative (rel.) brightness of images with error as SD from the mean was calculated (n = 15). (H and J) ZO1. (I and L) E-cadherin. (K) occludin. (M) β-catenin. (N–Q) TJ and AJ protein levels were determined by Western blotting with the indicated antibodies and (O–Q) quantified by densitometry (n = 3). Data are representative of at least three independent experiments; n = 5 mice per group. One-way ANOVA followed by post hoc Tukey’s test: **P < 0.01, ***P < 0.001, n.s.: not significant. Figure 2.23 Effect of tryptophan metabolite IEt in decreasing TJ and AJ disassembly in mouse model of colitis is dependent on AhR. Ahr+/- or Ahr-/- mice were pre-treated with IEt (600 mg/kg) for 2 d and then administered DSS (3%, wt/vol) for 7 d (ad libitum) with continued metabolite treatment. Colon sections were stained for (A) occludin and (B) -catenin and imaged by confocal microscopy. Scale bars = 20 m. Data are representative of at least 3 independent experiments, n = 5 mice per group. 83 Figure 2.24 Effect of tryptophan metabolite IPyA in mouse model of colitis is dependent on AhR. (A) Ahr+/- or Ahr-/- mice were pre-treated with IPyA (2900 mg/kg) for 2 d and then administered 84 DSS (3%, wt/vol) for 7 d (ad libitum) with continued metabolite treatment. (B) The mice were weighed daily. (C) Mice were orally gavaged with FITC-dextran (900 mg/kg) on day 9, and serum levels of FITC-dextran were measured 4 h later. (D) On day 9, the mice were sacrificed, and colon lengths were measured, (E) disease activity index was measured, and (F-G) the distal colon was stained with H&E and blindly scored (0 = none, 1 = very mild, 2 = mild, 3 = moderate, 4 = severe) for epithelial damage, mononuclear and polymorphonuclear infiltrate, and submucosal edema. Scale bar = 50 m. (H-O) Colon sections were stained for TJ and AJ proteins and imaged by confocal microscopy. Scale bars = 20 m. (L-O) Relative (rel.) brightness of images with error as standard deviation from the mean was calculated (n = 15). (H,L) ZO1; (I,M) Occludin; (J,N) E- cadherin; (K,O) -catenin. (P-S) TJ and AJ protein levels were determined by Western blotting with the indicated antibodies and (Q-S) quantified by densitometry (n = 3). Data are representative of at least 3 independent experiments, n = 5 mice per group. One-way ANOVA followed by post- hoc Tukey’s test: **p<0.01, ***p<0.001, n.s. = not significant. 85 Figure 2.25 Effect of tryptophan metabolite I3A in mouse model of colitis is dependent on the aryl hydrocarbon receptor (AhR). (A) Ahr+/- or Ahr-/- mice were pre-treated with I3A (1000 86 mg/kg) for 2 d and then administered DSS (3%, wt/vol) for 7 d (ad libitum) with continued metabolite treatment. (B) The mice were weighed daily. (C) Mice were orally gavaged with FITC- dextran (900 mg/kg) on day 9, and serum levels of FITC-dextran were measured 4 h later. (D) On day 9, the mice were sacrificed, and colon lengths were measured, (E) disease activity index was measured, and (F-G) the distal colon was stained with H&E and blindly scored (0 = none, 1 = very mild, 2 = mild, 3 = moderate, 4 = severe) for epithelial damage, mononuclear and polymorphonuclear infiltrate, and submucosal edema. Scale bar = 50 m. (H-O) Colon sections were stained for TJ and AJ proteins and imaged by confocal microscopy. Scale bars = 20 m. (L- O) Relative (rel.) brightness of images with error as standard deviation from the mean was calculated (n = 15). (H,L) ZO1; (I,M) Occludin; (J,N) E-cadherin; (K,O) -catenin. (P-S) TJ and AJ protein levels were determined by Western blotting with the indicated antibodies and (Q-S) quantified by densitometry (n = 3). Data are representative of at least 3 independent experiments, n = 5 mice per group. One-way ANOVA followed by post-hoc Tukey’s test: **p<0.01, ***p<0.001, n.s. = not significant. 87 2.4.6 Trp Metabolites Inhibit Activation of Myosin IIA In Vivo in an AhR-Dependent Manner The previous studies implicate AhR as an upstream target of the Trp metabolites. We next set out to elucidate downstream mechanisms that would directly explain the metabolites’ effects on gut permeability in the context of TNFα treatment of Caco-2 monolayers as an in vitro model and DSS colitis in vivo. Because TNFα triggers increased permeability by opening up the AJC with mechanical force via contraction of the actin cytoskeleton (41, 42), we hypothesized that IEt, IPyA, and I3A might modulate the activity of actin regulatory proteins, such as nonmuscle myosin IIA (MyoIIA), a motor protein that generates force to alter the actin skeleton, to decrease intestinal permeability (43). Thus, we examined the activation state of MyoIIA in DSS-treated mice fed with Trp-rich or standard diets. MyoIIA is phosphorylated at serine 19 (p-MLC, for phospho-myosin light chain), an activation mark, by myosin light-chain kinase (MLCK). We found, by Western blotting and IF staining of intestinal tissue, that a Trp-rich diet reversed the increases in p-MLC and MLCK levels induced in vivo in DSS colitis (Figure 2.26 A–D, Q, and R). Importantly, we examined the effects in Ahr−/− mice and found that they were attenuated, indicating that the effects on MyoIIA are downstream of and dependent upon AhR activation by Trp metabolites (Figure 2.26). Furthermore, we also assessed the effects of the three relevant Trp metabolites—IEt, IPyA, and I3A—and found that treatment of mice with each of these metabolites, individually, led to similar effects for each metabolite in inhibiting MyoIIA activation that are also AhR dependent, suggesting that these metabolites contribute to the effects of Trp feeding (Figure 2.26 E–Q and S– U.) 88 Figure 2.26 Trp feeding and metabolites prevent myosin IIA activation during mouse model of colitis, which is AhR dependent. (A–D, Q, and R) Ahr+/− or Ahr−/− mice were fed a Trp-rich diet 89 (42 g Trp/kg diet) or standard chow (2 g Trp/kg diet) for 7 d, followed by administration of DSS (%, wt/vol) or vehicle for 7 d (ad libitum) with continued Trp feeding. (E–Q and S–U) Alternatively, Ahr+/− or Ahr−/− mice were pretreated with I3A (1,000 mg/kg), IPyA (2,900 mg/kg), or IEt (600 mg/kg) for 2 d and then administered DSS (3%, wt/vol) for 7 d (ad libitum) with continued metabolite treatment. Activated myosin levels within intestinal tissue were determined by (A–P) Western blotting against MLCK, p-MLC, and myosin light chain (MLC) and quantified by densitometry (n = 3) or (Q–U) immunofluorescence, followed by confocal microscopy and quantification of relative (rel.) brightness (n = 15). (Scale bar: 20 μm.) Data are representative of at least three independent experiments; n = 5 mice per group. One-way ANOVA followed by post hoc Tukey’s test: ***P < 0.001, n.s.: not significant. 2.4.7 Trp Metabolites Inhibit Activation of Ezrin In Vitro. Ezrin, a founding member of the ezrin/radixin/moesin protein family, is a major upstream regulator of MyoIIA-mediated actin assembly that physically links the actin cytoskeleton to the plasma membrane upon activation, which is marked by phosphorylation of threonine 567 (p-ezrin) and leads to mechanical opening of the AJC and increased permeability (44). We hypothesized that activation of ezrin might serve as a critical event that is inhibited by IEt, IPyA, and I3A, ultimately leading to the effects of these metabolites in preventing the increase of epithelial permeability caused by TNFα. To test this mechanistic hypothesis, we retrovirally transduced into Caco-2 monolayers either wild-type (WT) ezrin or one of two mutant forms of ezrin, one of which could not be activated (T567A, phosphodeficient) and the other of which was constitutively active (T567E, phosphomimetic) (45). We then treated these monolayers with IEt, IPyA, and I3A, subsequently challenged them with TNFα, and then evaluated epithelial permeability by TEER and FITC-dextran flux, and the sinuous phenotype by IF for AJC markers. We found that T567E- and T567Aexpressing monolayers exhibited, respectively, exaggerated and attenuated phenotypes relative to WT-expressing or control monolayers, indicating the expected functionality of these T567 mutants of ezrin (Figures 2.27–2.30). Interestingly, metabolite treatment attenuated the increased epithelial permeability in WT- and T567A-expressing monolayers (which retain 90 endogenous WT ezrin); however, the effects of the metabolites were largely abrogated in T567E- expressing monolayers (Figure 2.27). We observed similar effects on the sinuous phenotype caused by TNFα treatment of Caco-2 monolayers (Figures 2.28–2.30). These data indicate that constitutively active ezrin can override the effects of the Trp metabolites and strongly suggest that the mechanism of action of these metabolites is via the activity of the key actin regulatory protein ezrin. Figure 2.27 Activation of the actin regulatory protein ezrin increases epithelial permeability caused by TNF, and metabolites prevent increased permeability. Polarized Caco-2 monolayers expressing the indicated ezrin mutants (phosphodeficient, T567A; phosphomimetic, T567E) versus wildtype (WT) ezrin were pre-treated with I3A, IPyA, and IEt (1 mM), followed by TNF (20 ng/ml) for 24 h. (A-C) TEER and (D-F) FITC-dextran flux were measured. Data are representative of at least 3 independent experiments. One-way ANOVA followed by post-hoc Tukey’s test: **p<0.01, ***p<0.001, n.s. = not significant. 91 Figure 2.28 Activation of the actin regulatory protein ezrin increases sinuous phenotype caused by TNF, which is prevented by Trp metabolites. Polarized Caco-2 monolayers expressing the indicated ezrin mutants (phosphodeficient, T567A; phosphomimetic, T567E) versus wildtype (WT) ezrin were pre-treated with indicated metabolites (1 mM), followed by TNF (20 ng/ml) for 24 h. Cells were fixed, permeabilized, and stained for TJ proteins, followed by confocal microscopy. Insets: 2.5x magnification. Scale bars = 20 m. Data are representative of at least 3 independent experiments. 92 Figure 2.29 Activation of the actin regulatory protein ezrin increases mislocalization of AJ proteins caused by TNF, which is prevented by Trp metabolites. Polarized Caco-2 monolayers expressing the indicated ezrin mutants (phosphodeficient, T567A; phosphomimetic, T567E) versus wildtype (WT) ezrin were pre-treated with indicated metabolites (1 mM), followed by TNF (20 ng/ml) for 24 h. Cells were fixed, permeabilized, and stained for AJ proteins, followed by confocal microscopy. Insets: 2.5x magnification. Scale bars = 20 m. Data are representative of at least 3 independent experiments. 93 Figure 2.30 Activation of the actin regulatory protein ezrin increases sinuous phenotype and mislocalization of TJ and AJ proteins caused by TNF, which is prevented by Trp metabolites. Polarized Caco-2 monolayers expressing the indicated ezrin mutants (phosphodeficient, T567A; phosphomimetic, T567E) versus wildtype (WT) ezrin were pre-treated with indicated metabolites (1 mM), followed by TNF (20 ng/ml) for 24 h. Cells were fixed, permeabilized, and stained for TJ and AJ proteins, followed by confocal microscopy. (A-C, G-I, M-R) Relative (rel.) brightness of images with error as standard deviation from the mean was calculated (n = 15). (D-F, J-L) Linearity indices were calculated (n= 30). Data are representative of at least 3 independent 94 experiments. One-way ANOVA followed by post-hoc Tukey’s test: n = 3, **p<0.01, ***p<0.001, n.s. = not significant. 2.4.8 Trp Metabolites Inhibit Activation of Ezrin In Vivo in an AhR-Dependent Manner. Finally, to ascertain whether the metabolites inhibit ezrin activation in a physiologically relevant context, we fed DSS-treated mice a Trp-rich diet and found that, relative to a standard diet, Trp feeding attenuates the increase in p-ezrin in intestinal tissues caused by DSS treatment by Western blotting and IF (Figure 2.31 A–C, M, and N). We also found that treatment of mice with each of the active Trp metabolites (IEt, IPyA, or I3A) also inhibits ezrin activation, suggesting that these metabolites, which are produced in vivo from bacterial catabolism of Trp, may mediate the effects of Trp feeding during DSS colitis (Figure 2.31 D–M and O–Q). Finally, we found that the effects of both the Trp-rich diet and treatment with the individual metabolites IEt, IPyA, and I3A exhibited a substantial, but not complete, dependence upon AhR during DSS challenge because the activation of ezrin was decreased in Ahr+/− mice, but not in Ahr−/− mice (Figure 2.31). 95 Figure 2.31 Trp feeding and metabolites prevent ezrin activation in mouse model of colitis, which is AhR dependent. (A–C, M, and N) Ahr+/− or Ahr−/− mice were fed a Trp-rich diet (42 g Trp/kg diet) or standard chow (2 g Trp/kg diet) for 7 d, followed by administration of DSS (3%, wt/vol) or vehicle for 7 d (ad libitum) with continued Trp feeding. (D–M and O–Q) Alternatively, Ahr+/− or Ahr−/− mice were pretreated with I3A (1,000 mg/kg), IPyA (2,900 mg/kg), or IEt (600 mg/kg) for 2 d and then administered DSS (3%, wt/vol) for 7 d (ad libitum) 96 with continued metabolite treatment. Activated ezrin levels within intestinal tissue were determined by (A–L) Western blotting against p-ezrin and quantified by densitometry (n = 3) or (M–Q) immunofluorescence, followed by confocal microscopy and quantification of relative (rel.) brightness (n = 15). (Scale bar: 20 μm.) Data are representative of at least three independent experiments, n = 5 mice per group. One-way ANOVA followed by post hoc Tukey’s test: ***P < 0.001, n.s. = not significant. 2.5 Discussion Small-molecule metabolites produced by the gut microbiota are being increasingly appreciated to modulate numerous aspects of host physiology during health and disease (11–14). Despite this growing evidence, the identities of the majority of these molecules, their precise effects on the host, and the mechanisms by which they act remain largely uncharacterized. In this study, we focused on the effects of specific Trp metabolites that are produced by the gut microbiota in improving intestinal barrier function. In particular, we identified roles for three Trp metabolites—IEt, IPyA, and I3A—in increasing gut barrier integrity during challenge with a proinflammatory cytokine in vitro and in vivo using a mouse model of DSS colitis (Figure 2.32). In addition, we demonstrated that the effects of these metabolites were largely dependent on AhR, a receptor that provides many diverse functions for the host (33, 34) and inhibits activation of the actin regulatory proteins MyoIIA and ezrin. Supplementation of diets with Trp, which could be replicated by protein-rich diets, has been shown to be beneficial to the host by improving disease outcomes in various DSS colitis models (35–38). Critically, the identities of the metabolites and the affected host pathways were not fully characterized. We found that a Trp-rich diet provides the host with microbiota-derived metabolites, including IEt, IPyA, and I3A, which ameliorate morbidity using a mouse model of DSS colitis. These metabolites are produced by mouse and human gut bacteria, including Lactobacillus reuteri and Clostridium sporogenes, although to our knowledge, the amounts of 97 these metabolites have not been quantified within humans (16, 25, 46). Our data suggest that a mechanism by which these metabolites help protect the host is to increase gut barrier integrity through maintenance of the AJC during challenge with TNFα and DSS, which both cause increased intestinal permeability (Figure 2.32). We also showed that treatment of mice with IEt, IPyA, and I3A inhibits TNFR1 expression during DSS colitis and NF-κB activation by TNFα (Figure A1.11), which suggests that these metabolites inhibit TJ and AJ disassembly via the TNFα pathway characterized by Ma et al. and Wang et al. (47–49). We further demonstrated that these metabolites increase IL-10R expression in IECs during DSS colitis (Figure 2.14), which has been reported to protect against intestinal insults that target the AJC, including proinflammatory cytokines (50, 51). These results differ from previous studies in which Trp metabolites, including IPyA and I3A, were demonstrated to target alternative immune pathways, including induction of Tr1 and ILCs (24, 25), that also protect against colitis through their anti-inflammatory activities and barrier protection, respectively. It is likely that these small molecules have multiple physiologically relevant targets in vivo and that their overall effects are synergistic based on interactions with receptors and pathways in multiple host-cell types. Nevertheless, the gut epithelium, the apical side of which directly faces the microbiota-rich lumen, is a highly relevant cell type to investigate, given that it likely experiences the highest concentrations of microbiota- derived metabolites and has major roles in controlling permeability and tissue homeostasis, as well as instruction of several immune cell types (4). In our study, we identified several Trp metabolites in addition to IEt, IPyA, and I3A that also arise from Trp catabolism by the gut microbiota using mass spectrometry (MS)-based metabolomics (Figure 2.8). These include IND, indole-3-acetamide, tryptamine, indole-3-acetic acid, indole-3-lactate, indole-3-acrylate, and IPA. Although these metabolites are also produced in 98 vivo, our data suggest that they likely do not modulate gut epithelial permeability in this context, based on our initial screen demonstrating that they do not change TEER or FITC dextran flux in the presence of TNFα (Figures 2.6 and 2.7). Nevertheless, it is possible that these additional Trp metabolites may mediate alternative pathways in vivo that contribute to reduced morbidity in the DSS colitis model via immune-mediated pathways as described above. Notably, Mani and coworkers have shown that IPA improves intestinal barrier function in chemically induced colitis within mice; however, this metabolite targets pathways different from the ones we have identified here, including TLR4 and PXR activation (27). An interesting future direction would be to understand if these pathways synergize with metabolite-activated pathways identified in this study to ameliorate DSS colitis in mice. Trp-derived Kyn has been shown to improve DSS colitis by inducing IL-10R expression (26); however, we did not detect significant levels of this metabolite after Trp feeding (Figure 2.8). IND has also been shown to improve DSS colitis, most likely by increasing the expression of TJ and AJ proteins (28, 29); however, this metabolite did not prevent increased permeability caused by TNFα as determined by TEER and FITC dextran flux (Figure 2.7 B and L). Our results may differ from these studies because the levels of certain Trp metabolites produced in vivo can vary due to the composition of gut microbiota. Also, these reports utilized in vitro model systems and murine colitis models different from those included in this study. A major aspect of our study was the elucidation of the mechanism of action of Trp metabolites in the context of compromised gut barrier integrity during DSS-mediated epithelial damage. We demonstrated that the effects of IEt, IPyA, and I3A depend on AhR, a host receptor the traditional role of which is to recognize and detoxify foreign molecules within the host by up- regulating xenobiotic metabolizing enzymes (52). Recent studies have uncovered major roles for AhR in regulating many different aspects of the host immune response (33, 34). In the context of 99 colitis, certain xenobiotic AhR ligands that are not microbial metabolites, such as 2-(1′H-indole- 3′-carbonyl)-thiazole-4-carboxylic acid methyl ester and 2,3,7,8-tetrachlorodibenzodioxin, have been shown to decrease inflammation via immunologically mediated pathways, including increased expression of Tregs and production of PGE2 and RegIIIγ within the colon (23, 36, 53); however, these studies did not address the effects of microbially derived AhR ligands in this disease model. Here, we demonstrate that the effects of Trp feeding and its resulting metabolites IEt, IPyA, and I3A in improving morbidity and intestinal permeability in DSS colitis via maintenance of gut-barrier function through the AJC are largely AhR-dependent. Thus, these data implicate AhR as a host receptor directly targeted by the Trp metabolites in this context. It is important to note that our study does not examine direct effects of the Trp-derived metabolites IEt, IPyA, and I3A on distinct immune cell types (e.g., CD4+ T cell subsets) in DSS colitis. Nevertheless, our results suggest that Trp metabolite modulation of the AJC plays a role in strengthening gut epithelial barrier integrity in DSS colitis. Finally, to understand the molecular mechanisms by which IEt, IPyA, and I3A ultimately attenuate pathological increases in intestinal permeability, we showed that either a Trp-rich diet or individual treatment with each of the three metabolites inhibits activation of ezrin, a key actin regulatory protein that controls and maintains the AJC in endothelial and epithelial cells (Figures 2.27-2.25, 2.31 and 2.32) (54, 55). Activation of ezrin is a two-step process involving recruitment of ezrin to the plasma membrane via binding to phosphatidylinositol-4,5- bisphosphate, rendering threonine (Thr) 567 more accessible to phosphorylation (44). This modification (p-Thr567) causes a dramatic conformational change that unmasks ezrin’s binding site to F-actin, allowing p-ezrin to tether the cytoskeleton to the plasma membrane (Figure 2.32). Ezrin’s activation and association with the actin cytoskeleton leads to a mechanical opening of the AJC via the action of MyoIIA 100 motors and, hence, to increased permeability (Figure 2.32) (55, 56). Critically, we demonstrated that a Trp-rich diet or supplementation with IEt, IPyA, or I3A inhibits ezrin activation in vitro and in vivo and that a constitutively active form of ezrin could bypass the effects of the Trp metabolites (Figures 2.27–2.30 and 2.31). Furthermore, we determined that the effect of these metabolites in inhibiting ezrin is AhR-dependent in vivo (Figure 2.31), suggesting that activation of AhR by the metabolites occurs upstream of effects on ezrin, MyoIIA, and AJC integrity (Figure 2.32). Interestingly, a previous study demonstrated that ezrin protein levels are dependent on AhR in the retina (57), suggesting that the mechanism elucidated here may be generalizable to other tissues and physiological or pathological contexts. In conclusion, we have identified and extensively characterized roles for several Trp metabolites—IEt, IPyA, and I3A—that are derived from the gut microbiota and that improve intestinal epithelial barrier function during challenges with inflammatory stimuli, including proinflammatory cytokines and DSS colitis. We have demonstrated that the effects of these metabolites are largely dependent on AhR in vivo, which suggests that the host possesses receptors that are capable of recognizing these small molecule metabolites that act as signaling agents from the gut microbiota. We have also demonstrated that these Trp metabolites inhibit activation of actin regulatory proteins that control intestinal epithelial permeability by maintaining the AJC, including MyoIIA and ezrin. Critically, we found that inhibition of MyoIIA and ezrin activity by these metabolites is dependent upon AhR, supporting a model for the mechanism of action of these metabolites that takes into account a potential upstream receptor and downstream proteins that directly mediate effects on intestinal permeability. Thus, our study identifies how a class of gut microbial metabolites abundant in protein-rich diets can modulate host defense mechanisms such as the intestinal barrier by targeting specific host receptors and pathways. Interesting future 101 directions include understanding the effects of these metabolites on alternative host pathways that include different immune cell types. We envision that the metabolites and host pathways that they affect may inspire or serve as potential targets for the development of prophylactic and preventative measures that could greatly improve patient outcomes in IBDs such as ulcerative colitis. Figure 2.32 Proposed model of Trp-derived metabolite function in regulating gut epithelial permeability. (A) Trp metabolites, indole-3-ethanol, indole-3-pyruvate, and indole-3-aldehyde, produced by the gut microbiota, decrease intestinal permeability caused by proinflammatory cytokines (e.g., TNFα). (B) Trp metabolites decrease epithelial permeability through activation of AhR and modulation of TJs and AJs via actin regulatory proteins (e.g., ezrin and nonmuscle myosin II). Cartoon schematic of the model. F-actin: filamentous actin. 102 2.7 References 1. K. J. Maloy, F. Powrie, Intestinal homeostasis and its breakdown in inflammatory bowel disease. Nature 474, 298–306 (2011). 2. A. Kaser, S. Zeissig, R. S. Blumberg, Inflammatory bowel disease. Annu. Rev. Immunol. 28, 573–621 (2010). 3. B. Khor, A. Gardet, R. J. Xavier, Genetics and pathogenesis of inflammatory bowel disease. Nature 474, 307–317 (2011). 4. L. W. Peterson, D. Artis, Intestinal epithelial cells: Regulators of barrier function and immune homeostasis. Nat. Rev. Immunol. 14, 141–153 (2014). 5. E. A. Grice, J. A. Segre, The human microbiome: Our second genome. Annu. Rev. Genomics Hum. Genet. 13, 151–170 (2012). 6. J. Qin et al.; MetaHIT Consortium, A human gut microbial gene catalogue established by metagenomic sequencing. Nature 464, 59–65 (2010). 7. R. Sender, S. Fuchs, R. Milo, Are we really vastly outnumbered? Revisiting the ratio of bacterial to host cells in humans. Cell 164, 337–340 (2016). 8. J. C. Clemente, L. K. Ursell, L. W. Parfrey, R. Knight, The impact of the gut microbiota on human health: An integrative view. Cell 148, 1258–1270 (2012). 9. Y. Belkaid, T. W. Hand, Role of the microbiota in immunity and inflammation. Cell 157, 121– 141 (2014). 10. G. Sharon, T. R. Sampson, D. H. Geschwind, S. K. Mazmanian, The central nervous system and the gut microbiome. Cell 167, 915–932 (2016). 11. M. G. Rooks, W. S. Garrett, Gut microbiota, metabolites and host immunity. Nat. Rev. Immunol. 16, 341–352 (2016). 103 12. P. C. Dorrestein, S. K. Mazmanian, R. Knight, Finding the missing links among metabolites, microbes, and the host. Immunity 40, 824–832 (2014). 13. M. S. Donia, M. A. Fischbach, Small molecules from the human microbiota. Science 349, 1254766 (2015). 14. T. S. Postler, S. Ghosh, Understanding the holobiont: How microbial metabolites affect human health and shape the immune system. Cell Metab. 26, 110–130 (2017). 15. G. R. Nicolas, P. V. Chang, Deciphering the chemical lexicon of host-gut microbiota interactions. Trends Pharmacol. Sci. 40, 430–445 (2019). 16. L. Cervantes-Barragan et al., Lactobacillus reuteri induces gut intraepithelial CD4+ CD8αα+ T cells. Science 357, 806–810 (2017). 17. J. D. Mezrich et al., An interaction between kynurenine and the aryl hydrocarbon receptor can generate regulatory T cells. J. Immunol. 185, 3190–3198 (2010). 18. S. S. Iyer et al., Dietary and microbial oxazoles induce intestinal inflammation by modulating aryl hydrocarbon receptor responses. Cell 173, 1123–1134.e11 (2018). 19. V. Rothhammer et al., Type I interferons and microbial metabolites of tryptophan modulate astrocyte activity and central nervous system inflammation via the aryl hydrocarbon receptor. Nat. Med. 22, 586–597 (2016). 20. K. Kawajiri et al., Aryl hydrocarbon receptor suppresses intestinal carcinogenesis in ApcMin/+ mice with natural ligands. Proc. Natl. Acad. Sci. U.S.A. 106, 13481–13486 (2009). 21. M. Wlodarska et al., Indoleacrylic acid produced by commensal peptostreptococcus species suppresses inflammation. Cell Host Microbe 22, 25–37.e6 (2017). 22. B. Lamas et al., CARD9 impacts colitis by altering gut microbiota metabolism of tryptophan into aryl hydrocarbon receptor ligands. Nat. Med. 22, 598–605 (2016). 104 23. J. D. Abron et al., An endogenous aryl hydrocarbon receptor ligand, ITE, induces regulatory T cells and ameliorates experimental colitis. Am. J. Physiol. Gastrointest. Liver Physiol. 315, G220–G230 (2018). 24. R. Aoki, A. Aoki-Yoshida, C. Suzuki, Y. Takayama, Indole-3-pyruvic acid, an aryl hydrocarbon receptor activator, suppresses experimental colitis in mice. J. Immunol. 201, 3683– 3693 (2018). 25. T. Zelante et al., Tryptophan catabolites from microbiota engage aryl hydrocarbon receptor and balance mucosal reactivity via interleukin-22. Immunity 39, 372–385 (2013). 26. J. M. Lanis et al., Tryptophan metabolite activation of the aryl hydrocarbon receptor regulates IL-10 receptor expression on intestinal epithelia. Mucosal Immunol. 10, 1133–1144 (2017). 27. M. Venkatesh et al., Symbiotic bacterial metabolites regulate gastrointestinal barrier function via the xenobiotic sensor PXR and Toll-like receptor 4. Immunity 41, 296–310 (2014). 28. Y. Shimada et al., Commensal bacteria-dependent indole production enhances epithelial barrier function in the colon. PLoS One 8, e80604 (2013). 29. T. Bansal, R. C. Alaniz, T. K. Wood, A. Jayaraman, The bacterial signal indole increases epithelial-cell tight-junction resistance and attenuates indicators of inflammation. Proc. Natl. Acad. Sci. U.S.A. 107, 228–233 (2010). 30. M. Yu et al., Aryl hydrocarbon receptor activation modulates intestinal epithelial barrier function by maintaining tight junction integrity. Int. J. Biol. Sci. 14, 69–77 (2018). 31. J. R. Turner, Intestinal mucosal barrier function in health and disease. Nat. Rev. Immunol. 9, 799–809 (2009). 32. C. Zihni, C. Mills, K. Matter, M. S. Balda, Tight junctions: From simple barriers to multifunctional molecular gates. Nat. Rev. Mol. Cell Biol. 17, 564–580 (2016). 105 33. B. Stockinger, P. Di Meglio, M. Gialitakis, J. H. Duarte, The aryl hydrocarbon receptor: Multitasking in the immune system. Annu. Rev. Immunol. 32, 403–432 (2014). 34. V. Rothhammer, F. J. Quintana, The aryl hydrocarbon receptor: An environmental sensor integrating immune responses in health and disease. Nat. Rev. Immunol. 19, 184–197 (2019). 35. J. Islam et al., Dietary tryptophan alleviates dextran sodium sulfate-induced colitis through aryl hydrocarbon receptor in mice. J. Nutr. Biochem. 42, 43–50 (2017). 36. T. Takamura et al., Activation of the aryl hydrocarbon receptor pathway may ameliorate dextran sodium sulfate-induced colitis in mice. Immunol. Cell Biol. 88, 685–689 (2010). 37. T. Shizuma, H. Mori, N. Fukuyama, Protective effect of tryptophan against dextran sulfate sodium-induced experimental colitis. Turk. J. Gastroenterol. 24, 30–35 (2013). 38. C. J. Kim et al., l-Tryptophan exhibits therapeutic function in a porcine model of dextran sodium sulfate (DSS)-induced colitis. J. Nutr. Biochem. 21, 468–475 (2010). 39. I. Hubatsch, E. G. E. Ragnarsson, P. Artursson, Determination of drug permeability and prediction of drug absorption in Caco-2 monolayers. Nat. Protoc. 2, 2111–2119 (2007). 40. S. Luescher, C. Urmann, V. Butterweck, Effect of hops derived prenylated phenols on TNF-α induced barrier dysfunction in intestinal epithelial cells. J. Nat. Prod. 80, 925–931 (2017). 41. L. Shen, Tight junctions on the move: Molecular mechanisms for epithelial barrier regulation. Ann. N. Y. Acad. Sci. 1258, 9–18 (2012). 42. T. Suzuki, Regulation of intestinal epithelial permeability by tight junctions. Cell. Mol. Life Sci. 70, 631–659 (2013). 43. M. Vicente-Manzanares, X. Ma, R. S. Adelstein, A. R. Horwitz, Non-muscle myosin II takes centre stage in cell adhesion and migration. Nat. Rev. Mol. Cell Biol. 10, 778–790 (2009). 106 44. R. G. Fehon, A. I. McClatchey, A. Bretscher, Organizing the cell cortex: The role of ERM proteins. Nat. Rev. Mol. Cell Biol. 11, 276–287 (2010). 45. R. Viswanatha, P. Y. Ohouo, M. B. Smolka, A. Bretscher, Local phosphocycling mediated by LOK/SLK restricts ezrin function to the apical aspect of epithelial cells. J. Cell Biol. 199, 969– 984 (2012). 46. B. B. Williams et al., Discovery and characterization of gut microbiota decarboxylases that can produce the neurotransmitter tryptamine. Cell Host Microbe 16, 495–503 (2014). 47. T. Y. Ma, M. A. Boivin, D. Ye, A. Pedram, H. M. Said, Mechanism of TNF-alpha modulation of Caco-2 intestinal epithelial tight junction barrier: Role of myosin light-chain kinase protein expression. Am. J. Physiol. Gastrointest. Liver Physiol. 288, G422–G430 (2005). 48. T. Y. Ma et al., TNF-alpha-induced increase in intestinal epithelial tight junction permeability requires NF-kappa B activation. Am. J. Physiol. Gastrointest. Liver Physiol. 286, G367–G376 (2004). 49. F. Wang et al., Interferon-γ and tumor necrosis factor-α synergize to induce intestinal epithelial barrier dysfunction by up-regulating myosin light chain kinase expression. Am. J. Pathol. 166, 409–419 (2005). 50. K. L. Madsen, S. A. Lewis, M. M. Tavernini, J. Hibbard, R. N. Fedorak, Interleukin 10 prevents cytokine-induced disruption of T84 monolayer barrier integrity and limits chloride secretion. Gastroenterology 113, 151–159 (1997). 51. K. L. Madsen et al., Interleukin-10 gene-deficient mice develop a primary intestinal permeability defect in response to enteric microflora. Inflamm. Bowel Dis. 5, 262–270 (1999). 52. I. A. Murray, A. D. Patterson, G. H. Perdew, Aryl hydrocarbon receptor ligands in cancer: Friend and foe. Nat. Rev. Cancer 14, 801–814 (2014). 107 53. S. Fukumoto et al., Identification of a probiotic bacteria-derived activator of the aryl hydrocarbon receptor that inhibits colitis. Immunol. Cell Biol. 92, 460–465 (2014). 54. J. B. Casaletto, I. Saotome, M. Curto, A. I. McClatchey, Ezrin-mediated apical integrity is required for intestinal homeostasis. Proc. Natl. Acad. Sci. U.S.A. 108, 11924–11929 (2011). 55. M. Koss et al., Ezrin/radixin/moesin proteins are phosphorylated by TNF-alpha and modulate permeability increases in human pulmonary microvascular endothelial cells. J. Immunol. 176, 1218–1227 (2006). 56. D. M. Adyshev et al., Ezrin/radixin/moesin proteins differentially regulate endothelial hyperpermeability after thrombin. Am. J. Physiol. Lung Cell. Mol. Physiol. 305, L240–L255 (2013). 57. S.-Y. Kim et al., Deletion of aryl hydrocarbon receptor AHR in mice leads to subretinal accumulation of microglia and RPE atrophy. Invest. Ophthalmol. Vis. Sci. 55, 6031–6040 (2014). 108 Chapter 3: Dopamine receptor D2 confers colonization resistance via gut microbial metabolites This work is adapted from ‘Scott, Samantha A., Jingjing Fu, and Pamela V. Chang. “Dopamine receptor D2 confers colonization resistance via gut microbial metabolites.” Submitted.’ 3.1 Abstract The gut microbiota provide protection against infection by intestinal pathogens through a process known as colonization resistance. Established colonization resistance pathways either modulate host defense or directly affect the pathogen by competitive exclusion. We show that gut microbial metabolites derived from bacterial catabolism of the essential amino acid tryptophan activate a host neurotransmitter receptor, dopamine receptor D2 (DRD2), on intestinal epithelial cells to mediate colonization resistance against enterohemorrhagic Escherichia coli serotype O157:H7, a major cause of severe gastrointestinal illness and mortality, and the closely related murine pathogen Citrobacter rodentium. Further, these metabolites prevent colonization of such pathogens by decreasing epithelial expression of actin regulatory proteins hijacked by these pathogens via DRD2. Our results reveal a noncanonical host pathway by which the gut microbiota mediate colonization resistance. 3.2 Introduction The gut microbiome plays major roles in modulating host physiology. One such function is colonization resistance, or the ability of the microbial collective to protect the host against enteric pathogens (1–3). Although these microbes may outcompete some pathogens or modulate host defense provided by the gut barrier and intestinal immune cells, this phenomenon remains poorly understood. Emerging evidence suggests that small-molecule metabolites produced by the gut microbiota may mediate this process. Yet, of the estimated 500,000 metabolites in the gut (4, 5), 109 only a handful can protect the host from infection by enteric pathogens (6), and the affected host pathways remain largely uncharacterized. A comprehensive understanding of mechanisms underlying colonization resistance may inspire prophylactic and therapeutic approaches for improving gut health and treating gastrointestinal infections, which afflict millions globally. Enterohemorrhagic Escherichia coli (EHEC) serotype O157:H7 is a food-borne pathogen that causes gastroenteritis, enterocolitis, bloody diarrhea, and, in certain individuals, acute renal failure (hemolytic uremic syndrome) (7, 8). EHEC infects the host by forming attaching and effacing (AE) lesions known as actin pedestals on the gut epithelium, from which it secretes virulence factors, including Shiga toxin, into the host (9–11). Citrobacter rodentium, a murine AE pathogen, is a widely used model for EHEC infection due to a highly similar infection strategy, encoded by the locus of enterocyte effacement pathogenicity island shared with EHEC and enteropathogenic E. coli (12–14). Interestingly, the gut microbiota provide colonization resistance to C. rodentium by enhancing host defense pathways (15) and by competing with the pathogen for nutrients within the gut (16). Emerging evidence suggests that specific small-molecule metabolites produced by the gut microbiota can mediate colonization resistance against AE pathogens (17, 18). Because of their roles in modulating immunity (19–22) and improving gut barrier function during inflammation (23), we focused attention on bacterial metabolites derived from the essential amino acid tryptophan (Trp). 3.3 Materials and Methods Bacterial strains and culture EHEC EDL-933 and TUV93-0 was a kind gift of Tobi Doerr (Cornell University) and John Leong (Tufts University), respectively. C. rodentium DBS100 was obtained from Gregory 110 Sonnenberg (Weill Cornell Medicine). Difco MacConkey agar was purchased from BD Biosciences. Bacteriological agar was purchased from VWR, yeast extract and tryptone were purchased from IBI Scientific, and sodium chloride was purchased from Fisher Scientific. Tissue culture Caco-2 and HEK 293T cells were obtained from the American Tissue Culture Company and cultured according to their guidelines. HEK 293T cells stably expressing a GloSensor cAMP reporter were obtained from Thomas Gardella (Harvard Medical School). DMEM, penicillin/streptomycin (P/S), 0.05% trypsin, and DPBS were obtained from Corning, and Seradigm Premium Grade Fetal Bovine Serum (FBS) and Transfectagro were purchased from VWR. Cell culture transwell inserts (transparent PET membrane, 12-well, 0.4 μm pore size) and Falcon 12-well companion plates were obtained from BD Falcon. Lipofectamine 2000 was purchased from Thermo Fisher Scientific. Polybrene and puromycin were purchased from EMD Millipore Sigma. Metabolites and compounds L-Tryptophan (Trp) was purchased from Chem-Impex International, Inc. Indole-3- aldehyde (I3A) and indole-3-pyruvate (IPyA) were obtained from Biosynth. Tryptophol (IEt) was obtained from Alfa Aesar, and kynurenine (Kyn) was purchased from Cayman Chemical Company. Indole (IND), indole-3-acetamide (IAM), DL-indole-3-lactate (ILA), 5- hydroxytryptamine (5HT), and indole-3-acetic acid (IAA) were purchased from Sigma Aldrich. Indole-3-propionate (IPA) and dopamine hydrochloride (DA) were obtained from Alfa Aesar. Tryptamine hydrochloride (TrA) was obtained from TCI America, and indole-3-acrylate (IA) was purchased from Santa Cruz Biotechnology. Histology 111 Hematoxylin was purchased from VWR, and eosin Y was obtained from Acros Organics. Canada Balsam was obtained from Ward’s Science, and xylenes was obtained from Macron Fine Chemicals. Antibiotics Ampicillin and vancomycin hydrochloride were purchased from Sigma Aldrich. Neomycin sulfate hydrate and metronidazole were obtained from Alfa Aesar. Western blotting DC Protein Assay kit was purchased from BioRad. SuperSignal West Pico Chemiluminescent Substrate was purchased from Thermo Fisher Scientific. Bovine serum albumin (BSA) was purchased from VWR, and non-fat dry milk was purchased from Laboratory Product Sales. Protease inhibitor (cOmplete) tablets were obtained from Roche. Sodium β- glycerophosphate was obtained from Alfa Aesar, and sodium orthovanadate was purchased from MP Biomedicals. Sodium fluoride was purchased from Chem-Impex International, Inc., and sodium pyrophosphate decahydrate was purchased from Fisher Scientific. Immunofluorescence Paraformaldehyde (PFA) 32% solution, EM grade, was purchased from Electron Microscopy Sciences, and DAPI Prolong Diamond was purchased from Thermo Fisher Scientific. Fisher HealthCare Tissue Plus OCT Compound was obtained from Fisher Scientific. LC-MS LC-MS grade methanol, water, acetonitrile, and formic acid (FA) were obtained from Fisher Scientific. qPCR 112 RNABee was obtained from Tel-Test, Inc. Diethylpyrocarbonate (DEPC) and chloroform were purchased from Sigma Aldrich. Isopropanol and ethanol were purchased from VWR. Glycogen (Roche) was purchased from Krackeler Scientific. Random hexamers were purchased from Thermo Fisher Scientific. dNTP was purchased from BioBasic. MMLV reverse transcriptase was purchased from Clontech, and PerfeCta SYBR Green SuperMix, Low ROX, were obtained from Quanta Biosciences. Antibodies Anti-DRD2 (clone B-10, sc-5303), anti-DRD3 (clone 9F4, sc-136170), anti-DRD4 (clone 2B9, sc-136169), anti-Arp3 (clone A-1, sc-48344), and anti-N-WASP (clone C-1, sc-271484) were purchased from Santa Cruz Biotechnology. Anti-IRTKS (clone 2A4, WH0055971M1) was purchased from Sigma Aldrich. Anti-mouse HRP (170-5947) secondary antibody for Western blot was purchased from BioRad. Donkey anti-mouse Alexa Fluor 594 (A21293) and Alexa Fluor 647- phalloidin (A22287) were purchased from Thermo Fisher Scientific. Anti-α-tubulin (clone B512, T5168) antibody was purchased from EMD Millipore Sigma. Mice C57Bl/6, Drd2fl/fl (#020631, Drd2loxP), Villin-Cre (#021504), Cd11c-Cre (#008068), LysM-Cre (#004781), and Cd4-Cre (#022071) mice were acquired from Jackson Laboratories. All mice were subsequently bred and maintained at the animal facility of Cornell University and used at 8-12 weeks of age in accordance with the guidelines of the Institutional Animal Care and Use Committee and the Cornell Center for Animal Resources and Education (Protocol number 2015- 0069). Mice were co-housed for 7 d prior to use and fed Envigo Teklad global irradiated 18% 113 protein rodent diet meal 2918 as the conventional diet. Mice were anesthetized with Butler Schein Animal Health Isothesia Isoflurane, USP, 3% vol/vol. Fluorescein isothiocyanate (FITC)-dextran (FD4, average molecular weight 3,000 – 5,000) was obtained from Sigma Aldrich. qPCR primers1 Gene Target Direction Sequence (5-3) species tir EHEC Forward GAAGTCGGCACCTGCGAATCA Reverse GCATAGGGACCGTGCAGAATC eae (intimin) EHEC Forward TGTCGCACTAACAGTCGCTT Reverse GCAACCACGGGAAATGATGG espF EHEC Forward TTCACCGGAGTAAGACGCAC Reverse CTGCTTCTACACTAGGGCGG tccp EHEC Forward TAGCTCCATCAGCGCAACAA Reverse GCGCTGCCTCACATTAGGA rpoA EHEC Forward GCGCTCATCTTCTTCCGAAT Reverse CGCGGTCGTGGTTATGTG tir C. rodentium Forward CAGGCTAAACGTCAGCAGGA Reverse TCGGCGGATTTCGTCTATGG eae (intimin) C. rodentium Forward TCAGCATAGCGGAAGCCAAA Reverse TGCTACCGCCTTGCACATAA espF C. rodentium Forward AATGGAATTGGTCAGGCCGT Reverse ACTGAAAAGCTCGCACCTCC rpoA C. rodentium Forward GCCCTGTTGACGATCTGGAA 114 Reverse GCTCAACCTCAGTACGCTGT 1Primers were purchased from IDT. Plasmids DRD2-Tango (Plasmid #66269), pCDNA3.1(+)-CMV-bArrestin2-TEV (Plasmid #107245), and lentiCRISPRv2 puro (Plasmid #98290) were obtained from Addgene. pCDNA3.1- DRD2 was generated for this paper by subcloning Drd2 into the pCDNA3.1 backbone using EcoR1 and Xho1. VSVg and PAX2 packaging plasmids were obtained from Jeremy Baskin (Cornell University). In vivo experiments Antibiotic treatment Mice were administered an antibiotic cocktail of ampicillin (9 mg/kg), metronidazole (9 mg/kg), neomycin (9 mg/kg), and vancomycin (4.5 mg/kg) via oral gavage every 12 h. Mice were pre-treated with antibiotics for 7 d prior to receiving high tryptophan diet, metabolites, or DA. Mice were continued to be administered the antibiotic cocktail while receiving high tryptophan diet, metabolites, or DA. Neomycin treatment was discontinued 24 h prior to C. rodentium infection, and ampicillin, metronidazole, and vancomycin were continued to be administered for the remainder of the experiment. High tryptophan diet Conventional diet (Envigo Teklad global irradiated 18% protein rodent diet meal 2918) or high tryptophan (Trp) diet was provided to mice ad libitum in feeding jars. The high tryptophan diet was prepared by supplementing conventional diet (2 g Trp/kg diet) with an additional 40 g Trp/kg diet. Mice received high tryptophan diet for 7 d prior to C. rodentium infection and then for the remainder of the experiment. 115 Metabolite treatment Mice were administered I3A (1000 mg/kg), IPyA (2900 mg/kg), or IEt (600 mg/kg). I3A, IPyA, and IEt were dissolved in dimethylsulfoxide (DMSO) and administered via oral gavage every 12 h. Control mice received equivalent volumes of DMSO via gavage. Mice were pre-treated with metabolite for 2 d prior to C. rodentium infection and then for the remainder of the experiment by oral gavage. Dopamine treatment Mice were administered DA (100 mg/kg) via intraperitoneal (IP) injection every 24 h. Control mice received equivalent volumes of PBS (vehicle) via IP injection. Mice were pre-treated with DA via IP injection for 2 d prior to C. rodentium infection and then for the remainder of the experiment. C. rodentium infection C. rodentium was grown in LB broth overnight with shaking at 37 °C. The next day, bacteria were subcultured until mid-log (OD600 = 0.4 – 0.6) phase and mice were infected with 108 CFU by oral gavage. Mice were euthanized 10 d post-infection. Weight loss Mice were weighed daily at the same time each day. Each mouse weight was normalized to itself and control mice on day 0. 4 kDa FITC-dextran intestinal permeability assay On the last day of the experiment, mice were fasted for 4 h prior to gavage with FD4 (900 mg/kg) in PBS. Prior to euthanasia and 4 h post-gavage, 100 µl of blood was collected via retro- orbital bleed and centrifuged to obtain at least 50 µl of serum. Concentration of FD4 in the serum 116 was determined using a SpectraMax Gemini EM Microplate Reader and a standard curve from a serial dilution of FD4 in PBS. CFU quantification Colon contents were collected, and a portion of the distal colon was flushed with PBS. Colon contents and tissue were weighed and homogenized in PBS. Colon contents were homogenized using a Corning LSE Vortex Mixer, and colon tissue was homogenized using a Brinkmann KINEMATICA Ch-6010 KRIENS-LU Benchtop Homogenizer. Homogenates were serially diluted, plated onto MacConkey agar, and incubated for 24 h at 37 °C. C. rodentium colonies were counted to determine CFU/g of colon contents and tissue. Only plates with 30 to 300 colonies were enumerated. Histopathology and crypt height measurement A portion of the distal colon was flushed with PBS, excised and fixed in 10% neutral buffered formalin, paraffin-embedded, sectioned (5 μm), and stained with Harris hematoxylin and eosin Y. Samples were blinded, imaged using an Olympus CX41RF microscope, and given a score between 0 and 4, where 0 = normal pathology, 1 = mild, multifocal individual crypt epithelial cell attenuation and goblet cell depletion without inflammatory cell infiltrate, 2 = mild, multifocal crypt epithelial cell loss without inflammatory cell infiltrate, 3 = moderate, multifocal epithelial cell loss with mild lamina propria lymphocytic and neutrophilic infiltrate, and 4 = severe, diffuse surface epithelial cell erosion with extensive crypt epithelial cell necrosis and loss, mild lamina propria lymphocytic and rare neutrophilic infiltrate, and moderate submucosal edema. Crypt height was measured using Olympus cellSens Entry software by measuring ten well-oriented crypts per mouse. LC-MS 117 Fecal contents were collected fresh from mice and immediately flash frozen in liquid nitrogen. Frozen samples were dried on a VirTis Benchtop K Series Manifold Freeze Dryer. Dried samples were crushed and resolubilized in methanol (10x the volume of the dry weight of the samples) and rocked at room temperature for 1 h before collecting the supernatant, which was then dried down. Immediately prior to LC-MS analysis, the samples were resuspended in methanol (10x the volume of the dry weight of the samples) and filtered. LC-MS analysis was performed on an Agilent 6230 electrospray ionization–time-of-flight (ESI–TOF) MS coupled to an Agilent 1260 HPLC equipped with an Agilent Poroshell 120 ECC18 reverse phase column (3 x 50 mm, 2.7 µm) using a flow rate of 0.5 ml/min. The gradient was ramped from 90% water and 0.1% FA (Solvent A) and 10% acetonitrile and 0.1% FA (Solvent B) to 50% A and 50% B for 0.5 min. The gradient was then ramped to 35% A and 65% B for an additional 0.5 min, then to 15% A and 85% B for 4.5 min, followed by 0% A and 100% B for 0.75 min. The gradient was then held constant at 0% A and 100% B for an additional minute. For detection, the MS was equipped with a dual ESI source operating in positive or negative mode, acquiring in extended dynamic range from m/z 100–3200 at one spectrum per s; gas temperature: 325 °C; drying gas 10 L/min; nebulizer: 20 psi; fragmentor: 80 V. Quantification of metabolites was determined by integrating the extracted ion count of the exact masses of the metabolites, which were determined using commercial standards. Standard curves in which known amounts of metabolite were utilized to determine the amount of each metabolite in each sample, which was normalized to the dry weight of the fecal samples. In vitro experiments Caco-2 monolayers 118 Caco-2 cells were seeded at 15,000 cells per transwell insert in 12-well companion plates. Monolayers were grown in DMEM supplemented with 10% FBS and P/S at 37 C and 5% CO2. Media was replaced every 2-3 days. Haloperidol treatment Caco-2 monolayers were grown as above. Haloperidol in DMSO was added initially to monolayers at 10 μM on day 17 and replenished on day 18 and 20. Equivalent volumes of DMSO were applied to control cells. CRISPR/Cas9 knockout of Drd2, Drd3, and Drd4 Drd2 forward (5-ATGGGAGTTTCCCAGTGAAC-3) and reverse (5- GTTCACTGGGAAACTCCCAT-3), Drd3 forward (5- CAGGCCATTGCCGAAGACGA-3) and reverse (5- TCGTCTTCGGCAATGGCCTG-3), and Drd4 forward (5- CAACCTGTGCGCCATCAGCG-3) and reverse (5- CGCTGATGGCGCACAGGTTG-3) guide RNAs were cloned into the lentiviral CRISPR plasmid lentiCRISPRv2 puro, following digestion with BsmBI. HEK 293T cells were seeded at 105 cells/well in 6-well plates. At 90-95% confluency, the cells were washed, media was replaced, and cells were co-transfected with 1 µg of lentiCRISPR plasmid and 0.33 µg of VSVg and 0.33 µg of PAX2 packaging plasmids using 3 µl Lipofectamine 2000 per well. The cells were incubated at 37 C for 2 d. Afterwards, the HEK 293T cell supernatant was collected and syringe-filtered using a 0.45 µm filter. For lentiviral transduction, filtered supernatant was added dropwise to Caco-2 cells seeded in a 10 cm dish, and the cells were incubated for 2 d at 37 C. Polybrene was added to a final concentration of 4 μg/ml. Transduced cells were then split 1:1 into media containing puromycin (2 µg/ml). After selection, successful transduction was confirmed by Western blotting, and cells were utilized to set up monolayers. 119 4 kDa FITC-dextran permeability assay Caco-2 monolayers were grown as above. On day 18, media was replenished, and 100 μM I3A, IPyA, or IEt in DMSO were applied to appropriate inserts. Equivalent volumes of DMSO were applied to control cells. On day 20, media and metabolites were replaced, and appropriate inserts were infected with EHEC at MOI 50. Monolayers were washed, and media was replaced with Hank’s Balanced Salt Solution (HBSS) 14 h post-EHEC infection. FD4 was added to the insert at 1 mg/ml in HBSS. After 2 h, 200 µl of HBSS was removed from each well. Concentration of FD4 in the well was determined using a SpectraMax Gemini EM Microplate Reader and a standard curve from a serial dilution of FD4 in HBSS. Growth curves and viable CFU counts EHEC or C. rodentium were grown in LB broth overnight with shaking at 37 °C. The next day, bacteria were cultured in fresh media with 100 μM or 1 mM I3A, IPyA, or IEt in DMSO. An equivalent volume of DMSO was added to controls. Cultures were incubated with shaking at 37 °C. OD600 absorbance readings were measured after 2, 4, 6, 8, 12, and 24 h. At 24 h, cultures were serially diluted, plated on LB agar, and incubated overnight at 37 °C. Colonies were counted to determine viable CFUs. Only plates with 30 to 300 colonies were enumerated. Western blot Caco-2 monolayers were grown as above. On day 18, media was replenished, and 100 μM I3A, IPyA, or IEt in DMSO were applied to appropriate inserts. Equivalent volumes of DMSO were applied to control cells. On day 20, media and metabolites were replaced, and appropriate inserts were infected with EHEC at MOI 50. After 16 h, monolayers were washed with PBS, and cells were lysed with 100 µl 1X RIPA lysis buffer (150 mM sodium chloride, 1.0% Triton X-100, 0.5% sodium deoxycholate, 0.1% SDS, 25 mM Tris, 1 mM EDTA) with 1X protease inhibitor 120 (cOmplete tablets) and sodium β-glycerophosphate (17.5 mM), sodium orthovanadate (1 mM), sodium fluoride (20 mM), and sodium pyrophosphate decahydrate (5 mM) added immediately prior to use. Samples from biological replicates were pooled prior to quantification using the DC Protein Assay kit. Lysates were sonicated using a Heat Systems Ultrasonic Processer XL sonicator. Protein concentrations were determined using the DC Protein Assay kit and BioTek PowerWave XS2 plate reader. Protein concentrations were normalized using 6X Laemmli buffer and ddH2O. Lysates were resolved on SDS-polyacrylamide gels and transferred to nitrocellulose. Membranes were blocked with 5% BSA in 25 mM Tris, 150 mM sodium chloride, and 0.1% Tween-20 solution (TBS-T) rocking at room temperature for 1 h, then probed with the appropriate antibodies in 5% milk or BSA in TBS-T with 0.05% sodium azide with rocking at 4 C overnight. DRD2, DRD3, and DRD4 primary antibodies were each diluted 1:1 with glycerol, then used at a 1:500 dilution in 5% BSA in TBS-T. The α-tubulin antibody was used at 1:2,500 in 5% BSA in TBS-T. Overnight incubation in primary antibody was followed by washing (3 x TBS-T) and incubation with the appropriate species-specific HRP antibody diluted at 1:10,000 in 5% milk in TBS-T with rocking at room temperature for 1 h. Western blots were developed using SuperSignal West Pico Chemiluminescent Substrate on a BioRad ChemiDoc MP. Densitometry was performed using FIJI and normalized to the housekeeping protein and control lysate proteins. Immunofluorescence For mouse samples, A portion of the distal colon was flushed with PBS, excised, and frozen in OCT. Blocks were sectioned (5 μm) on a Thermo Scientific Microm HM 525 cryostat, adhered to a glass slide, washed with PBS, and fixed with 4% PFA prior to staining. Caco-2 monolayers were grown as above. On day 18, media was replenished, and 100 μM I3A, IPyA, or IEt in DMSO 121 were applied to appropriate inserts. Equivalent volumes of DMSO were applied to control cells. On day 20, media and metabolites were replaced, and appropriate inserts were infected with EHEC at MOI 50. After 16 h, monolayers were washed with PBS and fixed with 4% PFA prior to staining. Fixed samples were permeabilized with 0.5% Triton X-100 in PBS for 15 min, blocked with 5% BSA in PBS for 1 h, and then incubated with the appropriate antibodies in 5% BSA in PBS at room temperature for 2 h. Antibodies against N-WASP, IRTKS, and Arp3 were each diluted 1:1 in glycerol and used at a dilution of 1:100. Samples were incubated with the appropriate species-specific secondary Alexa Fluor 594 antibody at a dilution of 1:500 in 5% BSA in PBS in the dark at room temperature for 1 h, then mounted with DAPI Prolong Diamond overnight. Alternatively, the samples were incubated with Alexa Fluor 647-phalloidin at a dilution of 1:500 in 5% BSA in PBS in the dark at room temperature for 1 h, then mounted with DAPI Prolong Diamond overnight. Samples were imaged with a Zeiss LSM 800 confocal laser scanning microscope equipped with 20X 0.8 NA and 40X 1.4 NA Plan Apochromat objectives, 405, 488, 561, and 640 nm solid-state lasers, and two GaAsP PMT detectors or a Zeiss LSM 880 confocal laser scanning microscope equipped with a 40X 1.4 NA Plan Apochromat objective, 405, 458, 488, 514, 561, and 633 nm solid-state lasers, two PMT channels, and a 32 channel GaAsP detector array. Images shown are maximum intensity z-projections. Relative brightness of stained cells was quantified using FIJI and normalized to control cells. Using distal colon cryosections or Caco-2 monolayers stained with DAPI Prolong Diamond and Alexa Fluor 647-phalloidin, the number of bacteria and actin pedestals were counted. Percent pedestal formation is the result of the number of pedestals counted divided by total bacteria counted, multiplied by 100. For five replicate images per mouse, a range of 12 to 85 pedestals was 122 counted per image. For each of the three replicated Caco-2 monolayer inserts per treatment, a range of 14 to 83 pedestals were counted per image. RNA isolation and qPCR analysis EHEC or C. rodentium were grown in LB broth overnight with shaking at 37 °C with 100 μM or 1 mM I3A, IPyA, or IEt dissolved in DMSO. An equivalent volume of DMSO was added to control samples. The next day, bacteria were subcultured until mid-log phase (OD600 = 0.4 – 0.6) in 5 ml fresh LB or 5 ml preconditioned LB (supernatant from overnight culture) to activate locus of enterocyte effacement-pathogenicity island expression with 100 μM or 1 mM I3A, IPyA, or IEt dissolved in DMSO or an equivalent volume of DMSO. Cultures were centrifuged and resulting cell pellets were lysed with 1 ml RNABee prior to RNA purification according to the manufacturer’s instructions. RNA was quantified using a GE Nanovue. Using a BioRad C1000 Touch Thermal Cycler, RNA was reverse transcribed using random hexamers and MMLV reverse transcriptase. cDNA samples were analyzed using PerfeCta SYBR Green SuperMix, Low ROX, and a BioRad CFX96 Real-Time PCR Detection System. PCR amplification conditions were as follows: 95 °C (3 min) and 40 cycles of 95 °C (15 s) and 60 °C (45 s). Relative expression of mRNA transcripts was normalized to the RNA polymerase subunit alpha rpoA. Data are represented as the fold induction over control samples. Luciferase activity assays Passive lysis buffer (5X PLB) was prepared with 125 mM Tris, pH 7.8, 10 mM 1,2- diaminocyclohexane tetraacetic acid (CDTA), 10 mM DTT, 5 mg/mL BSA, 5% (vol/vol) Triton X-100, and 50% (vol/vol) glycerol in ddH2O. An aqueous solution of 1X firefly luciferase substrate was prepared containing 75 mM HEPES, pH 8.0, 4 mM MgSO4, 20 mM DTT, 0.1 mM EDTA, 0.53 mM ATP, 0.27 mM coenzyme A, and 0.47 mM D-luciferin (firefly) in ddH2O. An 123 aqueous solution of 1X Renilla luciferase buffer was prepared containing 7.5 mM sodium acetate, pH 5.0, 400 mM sodium sulfate, 10 mM CDTA, 15 mM sodium pyrophosphate, and 0.025 mM 2- (4-aminophenyl)-6-methylbenzothiazole. A 100X Renilla luciferase substrate was prepared by diluting coelenterazine to 0.55 mM in anhydrous methanol and added to 1X Renilla luciferase buffer immediately prior to the assay. DRD2-Tango assay HEK 293T cells were plated at 105 cells/well in a 24-well plate. At 50-60% confluency, the cells were transfected overnight with DRD2-Tango and pCDNA3.1(+)-CMV-bArrestin2-TEV. The next day, media was replaced, and cells were treated for 24 h with 100 μM I3A, IPyA, IET, or DA or with an equivalent volume of DMSO. After 24 h, cells were washed with PBS, lysed with 1X PLB, and 20 µl of lysate was added to a 96-well white flat-bottom plate. Afterwards, 50 µl of 1X firefly luciferase substrate was added to each well, and luminescence was measured for 10 min using a Turner BioSystems Veritas Microplate Luminometer. Immediately after, 50 µl of the 1X Renilla substrate was added to each well, and luminescence was measured for 10 min. Luciferase activity was determined by calculating the ratio of the firefly luciferase signal to the Renilla luciferase signal. GloSensor assay HEK 293T cells stably expressing a GloSensor cAMP reporter were plated at 105 cells/well in a 24-well plate. At 50-60% confluency, the cells were transfected overnight with pCDNA3.1- DRD2. The next day, media was replaced, and the cells were incubated with D-luciferin (0.47 mM) for 30 min at room temperature. Cells were then treated for 15 min with 100 μM I3A, IPyA, IEt, or DA (unless otherwise indicated for EC50 dose determination) or with an equivalent volume of DMSO. After 15 min, cells were washed with PBS, lysed with 1X PLB, and 20 µl of lysate was 124 added to a 96-well white flat-bottom plate. Afterwards, luminescence was measured using a Turner BioSystems Veritas Microplate Luminometer. Similarity ensemble approach (SEA) SEA was utilized to predict protein targets for I3A, IPyA, and IEt. SEA is provided by the Shoichet Laboratory in the Department of Pharmaceutical Chemistry at the University of California, San Francisco (UCSF) at sea.bkslab.org. Non-human protein targets were excluded. Statistical analysis Experiments were completed at least three independent times. Error bars signify standard deviation from the mean. Statistical significance was determined using one-way ANOVA followed by post-hoc Tukey’s test. 3.4 Results 3.4.1 Trp Feeding Ameliorates C. rodentium Infection in Mice and is Dependent on the Gut Microbiota. We pre-treated wildtype C57Bl/6 mice with or without a broad-spectrum antibiotic cocktail to deplete the gut microbiota, followed by a Trp-rich diet for 7 d and infection with C. rodentium (Figure 3.1A). Compared to mice fed conventional diet, Trp-fed mice exhibited lower colonic C. rodentium burden (Figure 3.1B–C), decreased intestinal permeability (Figure 3.1D), decreased colonic inflammation (Figure 3.1E–G), and reduced weight loss (Figure 3.2A). These effects were abrogated by antibiotic treatment, suggesting that the microbiota provide protection against severe disease during infection in the presence of Trp. 125 Figure 3.1 Dietary tryptophan (Trp) protects against a mouse model of EHEC infection using Citrobacter rodentium, strain DBS100. (A) C57Bl/6 mice were pre-treated with antibiotics (-/+ ABX, 9 mg/kg each of metronidazole, ampicillin, and neomycin, 4.5 mg/kg of vancomycin) for 7 d, followed by conventional (2 g Trp/kg diet, ad libitum) or Trp-rich (42 g Trp/kg diet, ad libitum) diet for 7 d. The mice were then administered C. rodentium (CR, oral gavage, 108 colony-forming units, CFU) with continued ABX (except neomycin) and Trp feeding. Feces were collected before infection and euthanasia to determine Trp metabolite levels (see Fig. S2A). (B–C) Bacterial load in colon content and tissue was measured at the peak of infection, 10 d post-infection (day 17). (D) Mice were orally gavaged with 4 kDa FITC-dextran 10 d post-infection, and serum levels were measured after 4 h. (E–G) Colon sections were stained with H&E and (E) blindly scored (0 = none, 1 = very mild, 2 = mild, 3 = moderate, 4 = severe) for epithelial damage, mononuclear and polymorphonuclear infiltrate, and submucosal edema. (F) Crypt heights were measured. (G) Scale bar: 50 m. Data are representative of at least 3 independent experiments, n=5 mice per group, bars = mean, error bars = standard deviation. One-way ANOVA followed by post-hoc Tukey’s test: ***p<0.001, n.s. = not significant. 126 Figure 3.2 Tryptophan (Trp) metabolites prevent weight loss in a mouse model of EHEC infection using Citrobacter rodentium, strain DBS100. C57Bl/6 mice were pre-treated with antibiotics (-/+ ABX, 9 mg/kg each of metronidazole, ampicillin, and neomycin, 4.5 mg/kg of vancomycin) for 7 d, followed by conventional (2 g Trp/kg diet) or (A) Trp-rich (42 g Trp/kg diet, ad libitum) diet for 7 d or (B–D) Trp metabolites, I3A (1000 mg/kg), IEt (600 mg/kg), or IPyA (2900 mg/kg), by oral gavage daily for 2 d. (A–D) The mice were then administered C. rodentium (CR, oral gavage, 108 colony-forming units, CFU) with continued ABX (except neomycin) and Trp feeding or metabolite treatment. The mice were weighed daily. Asterisks denote statistical significance of comparisons between CR and CR + Trp/metabolite. Data are representative of at least 3 independent experiments, n=5 mice per group, bars = mean, error bars = standard deviation. One- way ANOVA followed by post-hoc Tukey’s test: ***p<0.001. 3.4.2 Specific Trp Metabolites Mediate Colonization Resistance Against C. rodentium Infection in Mice. To identify specific Trp metabolites that mediate the effects of the Trp diet, we performed targeted mass spectrometry (MS)-based metabolomics before and after C. rodentium infection during Trp feeding and found that the most highly abundant metabolites were indole, indole-3- ethanol (IEt), indole-3-pyruvate (IPyA), and indole-3-aldehyde (I3A) (Figure 3.3). Though indole 127 can decrease EHEC and C. rodentium virulence, the effects and mechanisms of action of IEt, IPyA, and I3A on these pathogens remain mostly unknown (24–26). 128 Figure 3.3 LC–MS quantification of tryptophan (Trp) metabolites during Trp-rich diet in mouse model of EHEC infection with C. rodentium, strain DBS100. C57Bl/6 mice were pre-treated with antibiotics (-/+ ABX, 9 mg/kg each of metronidazole, ampicillin, and neomycin, 4.5 mg/kg of vancomycin) for 7 d, followed by conventional (2 g Trp/kg diet, ad libitum) or Trp-rich (42 g Trp/kg diet, ad libitum) diet for 7 d The mice were then administered C. rodentium (CR, oral gavage, 108 colony-forming units, CFU) with continued ABX (except neomycin) and (A) Trp feeding. Feces were collected before infection and euthanasia to determine Trp metabolite levels. Metabolites from the fecal colonic contents were measured by mass spectrometry, using commercial standards for quantification. Abbreviations: Indole-3-aldehyde (I3A), indole-3- pyruvate (IPyA), tryptophol (IEt), indole (IND), indole-3-acetamide (IAM), DL-indole-3-lactate (ILA), indole-3-acetic acid (IAA), indole-3-propionate (IPA), tryptamine hydrochloride (TrA), indole-3-acrylate (IA). Data are representative of at least 3 independent experiments, n = 5-10 mice per group. One-way ANOVA followed by post-hoc Tukey’s test: *p<0.05, **p<0.01, ***p<0.001, n.s. = not significant. To determine the roles of these Trp metabolites on C. rodentium infection, we pre-treated C57Bl/6 mice with or without antibiotics to prevent microbial metabolism of the individual Trp metabolites, followed by administration of IEt, IPyA, and I3A for 2 d prior to and during infection (Figure 3.4A). Treatment with the individual metabolites also lowered colonic C. rodentium burden (Figure 3.4B–C), decreased intestinal permeability (Figure 3.4D), decreased colonic inflammation (Figure 3.4E–G), and decreased weight loss (Figure 3.2B–D). We verified that levels of IEt, IPyA, and I3A in the colon were increased following metabolite administration and did not change with antibiotic treatment (Figures 3.5-3.7). Thus, these microbial Trp metabolites mediate colonization resistance against C. rodentium. 129 Figure 3.4 The Trp metabolites I3A, IEt, and IPyA protect against Citrobacter rodentium infection in mice. (A) C57Bl/6 mice were pre-treated with antibiotics (ABX) for 7 d, followed by Trp metabolites, I3A (1000 mg/kg), IEt (600 mg/kg), or IPyA (2900 mg/kg), by oral gavage daily for 2 d. The mice were then administered C. rodentium (CR, oral gavage, 108 CFU) with continued ABX (except neomycin) and metabolite treatment. Feces were collected before infection and euthanasia to determine Trp metabolite levels (see Fig. S2B–D). (B–C) Bacterial load in colon content and tissue was measured at the peak of infection, 10 d post-infection (day 12). (D) Mice were orally gavaged with 4 kDa FITC-dextran 10 d post-infection, and serum levels were measured after 4 h. (E–G) Colon sections were stained with H&E and (E) blindly scored as described in Fig. 1. (F) Crypt heights were measured. (G) Scale bar: 50 m. Data are representative of at least 3 independent experiments, n=5 mice per group, bars = mean, error bars = standard deviation. One- way ANOVA followed by post-hoc Tukey’s test: ***p<0.001, n.s. = not significant. 130 Figure 3.5 LC–MS quantification of tryptophan (Trp) metabolites during I3A treatment in mouse model of EHEC infection with C. rodentium, strain DBS100. C57Bl/6 mice were fed I3A (1000 mg/kg) by oral gavage daily for 2 d. The mice were then administered C. rodentium (CR, oral 131 gavage, 108 colony-forming units, CFU) with continued I3A treatment. Feces were collected before infection and euthanasia to determine Trp metabolite levels. Metabolites from the fecal colonic contents were measured by mass spectrometry, using commercial standards for quantification. Abbreviations: Indole-3-aldehyde (I3A), indole-3-pyruvate (IPyA), tryptophol (IEt), indole (IND), indole-3-acetamide (IAM), DL-indole-3-lactate (ILA), indole-3-acetic acid (IAA), indole-3-propionate (IPA), tryptamine hydrochloride (TrA), indole-3-acrylate (IA). Data are representative of at least 3 independent experiments, n = 5-10 mice per group. One-way ANOVA followed by post-hoc Tukey’s test: *p<0.05, **p<0.01, ***p<0.001, n.s. = not significant. 132 Figure 3.6 LC–MS quantification of tryptophan (Trp) metabolites during IPyA treatment in mouse model of EHEC infection with C. rodentium, strain DBS100. C57Bl/6 mice were fed IPyA (2900 mg/kg) by oral gavage daily for 2 d. The mice were then administered C. rodentium (CR, oral gavage, 108 colony-forming units, CFU) with continued IPyA treatment. Feces were collected 133 before infection and euthanasia to determine Trp metabolite levels. Metabolites from the fecal colonic contents were measured by mass spectrometry, using commercial standards for quantification. Abbreviations: Indole-3-aldehyde (I3A), indole-3-pyruvate (IPyA), tryptophol (IEt), indole (IND), indole-3-acetamide (IAM), DL-indole-3-lactate (ILA), indole-3-acetic acid (IAA), indole-3-propionate (IPA), tryptamine hydrochloride (TrA), indole-3-acrylate (IA). Data are representative of at least 3 independent experiments, n = 5-10 mice per group. One-way ANOVA followed by post-hoc Tukey’s test: *p<0.05, **p<0.01, ***p<0.001, n.s. = not significant. 134 Figure 3.7 LC–MS quantification of tryptophan (Trp) metabolites during IEt treatment in mouse model of EHEC infection with C. rodentium, strain DBS100. C57Bl/6 mice were fed IEt (600 mg/kg) by oral gavage daily for 2 d. The mice were then administered C. rodentium (CR, oral gavage, 108 colony-forming units, CFU) with continued IEt treatment. Feces were collected before 135 infection and euthanasia to determine Trp metabolite levels. Metabolites from the fecal colonic contents were measured by mass spectrometry, using commercial standards for quantification. Abbreviations: Indole-3-aldehyde (I3A), indole-3-pyruvate (IPyA), tryptophol (IEt), indole (IND), indole-3-acetamide (IAM), DL-indole-3-lactate (ILA), indole-3-acetic acid (IAA), indole-3- propionate (IPA), tryptamine hydrochloride (TrA), indole-3-acrylate (IA). Data are representative of at least 3 independent experiments, n = 5-10 mice per group. One-way ANOVA followed by post-hoc Tukey’s test: *p<0.05, **p<0.01, ***p<0.001, n.s. = not significant. 3.4.3 Trp Metabolites Exert Modest Effects on C. rodentium and EHEC Growth and Virulence. To determine if IEt, IPyA, and I3A affect the pathogens directly, we treated C. rodentium with each metabolite and found that IEt and IPyA minimally affected bacterial growth and virulence factor expression, whereas I3A had modest effects, as previously reported (Figure 3.8A– E) (26). We also treated EHEC with IEt, IPyA, and I3A and found that each metabolite had only modest effects on this pathogen as well (Figure 3.8F–K). Because of these small effects on the pathogen and the substantial effects of the metabolites in vivo, we reasoned that IEt, IPyA, and I3A may mediate colonization resistance primarily via host pathways. 136 Figure 3.8 Tryptophan metabolites have modest effects on C. rodentium and EHEC growth in vitro. C. rodentium (CR) (A–E) and EHEC O157:H7 (F–K) were cultured in the presence of I3A, IPyA, and IEt at the indicated concentrations. (A and F) Growth was monitored by measuring OD600 absorbance readings over 24 h. (B and G) Cultures were plated after 24 h, and CFUs were counted. (C–E and H–K) Bacteria were cultured in Luria Broth (LB) or preconditioned LB (supernatant from overnight culture) to activate locus of enterocyte effacement-pathogenicity island expression with I3A, IPyA, or IEt at the indicated concentrations. RNA was isolated after cultures reached mid-log phase (OD600 = 0.4–0.6), and cDNA was synthesized and analyzed by qPCR for the indicated genes. Relative expression of mRNA transcripts was normalized to the RNA polymerase subunit alpha rpoA. Data are represented as the fold induction over control samples. One-way ANOVA followed by post-hoc Tukey’s test: n=3, *p<0.05, **p<0.01, ***p<0.001, all other points not indicated are n.s. = not significant. 3.4.4 Trp Metabolites are Ligands of DRD2. To identify potential host receptors for these metabolites, we utilized Similarity Ensemble Approach (27), a computational tool for predicting protein targets of small molecules, with IEt, 137 IPyA, and I3A as ligands. All three metabolites were predicted to bind the dopamine receptors D2 (DRD2), D3 (DRD3), and D4 (DRD4), G protein-coupled receptors (GPCRs) with canonical roles as neurotransmitter receptors in the central and peripheral nervous systems (28, 29). We found that DRD2–4 inhibition using the pan-inhibitor haloperidol (30) reduced the effects of the metabolites in decreasing epithelial permeability during EHEC infection of a human intestinal epithelial cell (IEC) line, Caco-2 (Figure 3.9A). Further, we found that this effect depended exclusively on Drd2, using CRISPR/Cas9 knockout of Drd2, Drd3, and Drd4 in Caco-2 cells during EHEC infection (Figure 3.9B–E). Critically, we found that IEt, IPyA, and I3A are bona fide DRD2 ligands, with potencies similar to the endogenous ligand, dopamine, using both the GloSensor and Tango GPCR assays, which quantify cyclic AMP and arrestin recruitment, respectively (Figure 3.10) (31, 32). We also pre-treated C57Bl/6 mice with or without the antibiotic cocktail, followed by dopamine and C. rodentium infection. We found that dopamine had similar effects to the Trp diet, IEt, IPyA, and I3A in decreasing morbidity during infection (Figure 3.11A–G; I). Collectively, these data suggest that DRD2 may be a host receptor in the gut that recognizes the Trp metabolites. Figure 3.9 Effects of I3A, IPyA, and IEt depend on dopamine receptor D2 (DRD2). (A) Polarized Caco-2 monolayers were pre-treated with haloperidol (HAL, 10 μM) for 24 h, followed by metabolites (100 μM) for 2 d, and then infection with EHEC O157:H7 for 16 h. (B) Western blot analysis of Caco-2 cells to verify CRISPR/Cas9-mediated knockout (KO) of Drd2, Drd3, and 138 Drd4. -tubulin is shown as a loading control. (C–E) Caco-2 monolayers (WT vs. KO) were pre- treated with metabolites (100 μM) for 2 d and then infected with EHEC O157:H7 for 16 h. (A, C– E) Epithelial permeability was measured by FITC-dextran flux. One-way ANOVA followed by post-hoc Tukey’s test: *p<0.05, ***p<0.001, n.s. = not significant. Figure 3.10 I3A, IPyA, and IEt are ligands of dopamine receptor D2 (DRD2), which signals via Gi. (A, C–F) HEK 293T cells overexpressing either DRD2 and a split luciferase-based cAMP sensor (GloSensor) or (B) DRD2-Tango and a -arrestin-TEV fusion were incubated with dopamine (DA), I3A, IPyA, or IEt (1 mM each in A–B; concentrations indicated in C–F) for 15 min (A, C–F) or 24 h (B), after which luminescence was measured to quantify ligand-induced (A) decrease in cAMP or (B) increase in -arrestin recruitment. RLU = relative luminescence units. One-way ANOVA followed by post-hoc Tukey’s test: ***p<0.001, n.s. = not significant. 139 Figure 3.11 Dopamine protects against a mouse model of EHEC infection using Citrobacter rodentium, strain DBS100. C57Bl/6 mice were pre-treated with antibiotics (-/+ ABX, 9 mg/kg each of metronidazole, ampicillin, and neomycin, 4.5 mg/kg of vancomycin) for 7 d, followed by dopamine (DA, 100 mg/kg, intraperitoneal injection) daily for 2 d. The mice were then administered C. rodentium (CR, oral gavage, 108 colony-forming units, CFU) with continued ABX (except neomycin) and DA treatment. (B) The mice were weighed daily. Asterisks denote statistical significance of comparisons between CR and CR + DA. (C–D) Bacterial load in colon content and tissue was measured at the peak of infection, 10 d post-infection (day 12). (E) Mice were orally gavaged with 4 kDa FITC-dextran on day 12, and serum levels were measured after 4 h. (F–G and I) Colon sections were stained with H&E and (F) blindly scored (0 = none, 1 = very mild, 2 = mild, 3 = moderate, 4 = severe) for epithelial damage, mononuclear and polymorphonuclear infiltrate, and submucosal edema. (G) Crypt heights were measured. (I) Shown are representative images. Scale bar: 50 m. (H) Intestinal cryosections were stained with DAPI and Alexa Fluor 647-phalloidin. Pedestal formation = # of pedestals (n = 141–173) divided by 140 total number of bacteria. (J–O) Intestinal cryosections were stained with antibodies against the indicated actin regulatory proteins. Shown are maximum intensity z-projections. Scale bar: 20 m. (M–O) Image brightness was quantified using FIJI. Data are representative of at least 3 independent experiments, n=5 mice per group, bars = mean, error bars = standard deviation. One- way ANOVA followed by post-hoc Tukey’s test: **p<0.01, ***p<0.001, n.s. = not significant. 3.4.5 Effect of Trp Metabolites is Dependent on DRD2 Expressed on Intestinal Epithelial Cells in Vivo. We first assessed whether DRD2 expression in the gut epithelium, which expresses this receptor and is the initial target tissue of AE pathogens, mediates the effects of the Trp metabolites in promoting colonization resistance to C. rodentium infection. We generated Drd2fl/fl x Vil-Cre mice, which lack Drd2 expression in IECs, and treated these mice with a Trp-rich diet, IEt, IPyA, or I3A, followed by infection with C. rodentium (Figure 3.12A). We found that Drd2fl/fl control littermates that were treated with Trp diet or individual metabolites exhibited decreased colonic C. rodentium burden (Figures 3.12B–C and 3.13A–B), gut permeability (Figures 3.12D and 3.13C), colonic inflammation (Figures 3.12E–G and 3.13D–F), and weight loss (Figure 3.14). However, these effects were abrogated in the Drd2fl/fl x Vil-Cre mice, indicating that DRD2 expression in IECs mediates the effects of the Trp metabolites. 141 Figure 3.12 Effects of the Trp diet and metabolites in protecting against Citrobacter rodentium infection depend on dopamine receptor D2 (DRD2) in intestinal epithelial cells (IECs). (A) Drd2fl/fl x Villin (Vil)-Cre mice were fed a conventional (2 g Trp/kg diet, ad libitum) or Trp-rich (42 g Trp/kg diet, ad libitum) diet for 7 d or Trp metabolites (shown: IEt (600 mg/kg); see Fig. S7 for I3A and IPyA), by oral gavage daily for 2 d, and then infected with C. rodentium (CR, oral gavage, 108 CFU) with continued Trp feeding or metabolite treatment. (B–C) Bacterial load in colon content and tissue was measured at the peak of infection, 10 d post-infection (Trp diet: day 17; metabolite: day 12). (D) Mice were orally gavaged with 4 kDa FITC-dextran on 10 d post- infection, and serum levels were measured after 4 h. (E–G) Colon sections were stained with H&E and (E) blindly scored as described in Fig. 1. (F) Crypt heights were measured. (G) Scale bar: 50 m. Data are representative of at least 3 independent experiments, n=5 mice per group, bars = mean, error bars = standard deviation. One-way ANOVA followed by post-hoc Tukey’s test: *p<0.05, **p<0.01, ***p<0.001, n.s. = not significant. 142 Figure 3.13 Effects of the tryptophan (Trp) metabolites I3A and IPyA in protecting against Citrobacter rodentium infection depend on dopamine receptor D2 (DRD2) in intestinal epithelial cells (IECs). Drd2fl/fl x Villin (Vil)-Cre mice were fed Trp metabolites, shown I3A (1000 mg/kg) and IPyA (2900 mg/kg), by oral gavage daily for 2 d, and then infected with C. rodentium (CR, oral gavage, 108 CFU) with continued metabolite treatment. (A–B) Bacterial load in colon content and tissue was measured at the peak of infection, 10 d post-infection (day 12). (C) Mice were orally gavaged with 4 kDa FITC-dextran on 10 d post-infection, and serum levels were measured after 4 h. (D–F) Colon sections were stained with H&E and (D) blindly scored (0 = none, 1 = very mild, 2 = mild, 3 = moderate, 4 = severe) for epithelial damage, mononuclear and polymorphonuclear infiltrate, and submucosal edema. (E) Crypt heights were measured. (F) Shown are representative images. Scale bar: 50 m. Data are representative of at least 3 independent experiments, n=5 mice per group, bars = mean, error bars = standard deviation. One- way ANOVA followed by post-hoc Tukey’s test: *p<0.05, **p<0.01, ***p<0.001, n.s. = not significant. 143 Figure 3.14 Effects of the tryptophan (Trp) diet and metabolites in preventing weight loss during Citrobacter rodentium infection depend on dopamine receptor D2 (DRD2) in intestinal epithelial cells (IECs). Drd2fl/fl x Villin (Vil)-Cre mice were fed a conventional (2 g Trp/kg diet, ad libitum) or (A) Trp-rich (42 g Trp/kg diet, ad libitum) diet for 7 d or (B–D) Trp metabolites, I3A (1000 mg/kg), IEt (600 mg/kg), or IPyA (2900 mg/kg), by oral gavage daily for 2 d, and then infected with C. rodentium (CR, oral gavage, 108 CFU) with continued Trp feeding or metabolite treatment. (A–D) The mice were weighed daily. Asterisks denote statistical significance of comparisons between Drd2fl/fl CR + Trp/metabolite and Drd2fl/fl x Vil-Cre CR + Trp/metabolite. Data are representative of at least 3 independent experiments, n=5 mice per group, bars = mean, error bars = standard deviation. One-way ANOVA followed by post-hoc Tukey’s test: ***p<0.001, n.s. = not significant. We further examined the effects of the Trp diet, IEt, IPyA, and I3A in C. rodentium- infected Drd2fl/fl mice lacking Drd2 expression in immune cells that also express DRD2 and are involved in host immunity against this pathogen, including macrophages (MPs), dendritic cells (DCs), and CD4+ T cells, using Drd2fl/fl x LysM-Cre, Drd2fl/fl x CD11c-Cre, and Drd2fl/fl x CD4-Cre mice, respectively. The effects of the Trp diet were not eliminated in the mice lacking 144 Drd2 in MPs, DCs, and CD4+ T cells, suggesting that the metabolites act via DRD2 expressed in IECs (Figures 3.15–3.17). Figure 3.15 Effects of the tryptophan (Trp) diet in protecting against Citrobacter rodentium infection do not depend on dopamine receptor D2 (DRD2) in macrophages. (A) Drd2fl/fl x LysM- Cre mice were fed a conventional (2 g Trp/kg diet, ad libitum) or Trp-rich (42 g Trp/kg diet, ad libitum) diet for 7 d and then infected with C. rodentium (CR, oral gavage, 108 CFU) with continued Trp feeding. (B) The mice were weighed daily. Asterisks denote statistical significance of comparisons between Drd2fl/fl CR + Trp and Drd2fl/fl x LysM-Cre CR + Trp. (C) Mice were orally gavaged with 4 kDa FITC-dextran on day 17, and serum levels were measured after 4 h. (D–E) Bacterial load in colon content and tissue was measured at the peak of infection, 10 d post- infection (day 17). (F–H) Colon sections were stained with H&E and (F) blindly scored (0 = none, 1 = very mild, 2 = mild, 3 = moderate, 4 = severe) for epithelial damage, mononuclear and polymorphonuclear infiltrate, and submucosal edema. (G) Crypt heights were measured. (H) Shown are representative images. Scale bar: 50 m. Data are representative of at least 3 independent experiments, n=5 mice per group, bars = mean, error bars = standard deviation. One- way ANOVA followed by post-hoc Tukey’s test: ***p<0.001, n.s. = not significant. 145 Figure 3.16 Effects of the tryptophan (Trp) diet in protecting against Citrobacter rodentium infection do not depend on dopamine receptor D2 (DRD2) in dendritic cells. (A) Drd2fl/fl x CD11c-Cre mice were fed a conventional (2 g Trp/kg diet, ad libitum) or Trp-rich (42 g Trp/kg diet, ad libitum) diet for 7 d and then infected with C. rodentium (CR, oral gavage, 108 CFU) with continued Trp feeding. (B) The mice were weighed daily. Asterisks denote statistical significance of comparisons between Drd2fl/fl CR + Trp and Drd2fl/fl x CD11c-Cre CR + Trp. (C) Mice were orally gavaged with 4 kDa FITC-dextran on day 17, and serum levels were measured after 4 h. (D–E) Bacterial load in colon content and tissue was measured at the peak of infection, 10 d post- infection (day 17). (F–H) Colon sections were stained with H&E and (F) blindly scored (0 = none, 1 = very mild, 2 = mild, 3 = moderate, 4 = severe) for epithelial damage, mononuclear and polymorphonuclear infiltrate, and submucosal edema. (G) Crypt heights were measured. (H) Shown are representative images. Scale bar: 50 m. Data are representative of at least 3 independent experiments, n=5 mice per group, bars = mean, error bars = standard deviation. One- way ANOVA followed by post-hoc Tukey’s test: ***p<0.001, n.s. = not significant. 146 Figure 3.17 Effects of the tryptophan (Trp) diet in protecting against Citrobacter rodentium infection do not depend on dopamine receptor D2 (DRD2) in CD4+ T cells. (A) Drd2fl/fl x CD4- Cre mice were fed a conventional (2 g Trp/kg diet, ad libitum) or Trp-rich (42 g Trp/kg diet, ad libitum) diet for 7 d and then infected with C. rodentium (CR, oral gavage, 108 CFU) with continued Trp feeding. (B) The mice were weighed daily. Asterisks denote statistical significance of comparisons between Drd2fl/fl CR + Trp and Drd2fl/fl x CD4-Cre CR + Trp. (C) Mice were orally gavaged with 4 kDa FITC-dextran on day 17, and serum levels were measured after 4 h. (D–E) Bacterial load in colon content and tissue was measured at the peak of infection, 10 d post- infection (day 17). (F–H) Colon sections were stained with H&E and (F) blindly scored (0 = none, 1 = very mild, 2 = mild, 3 = moderate, 4 = severe) for epithelial damage, mononuclear and polymorphonuclear infiltrate, and submucosal edema. (G) Crypt heights were measured. (H) Shown are representative images. Scale bar: 50 m. Data are representative of at least 3 independent experiments, n=5 mice per group, bars = mean, error bars = standard deviation. One- way ANOVA followed by post-hoc Tukey’s test: ***p<0.001, n.s. = not significant. 3.4.6 Trp Metabolites Inhibit C. rodentium and EHEC Pedestal Formation in a DRD2- dependent Manner. We then sought to determine mechanisms by which DRD2 mediates colonization resistance in IECs. Because the actin regulatory proteins IRTKS, N-WASP, and the Arp2/3 complex constitute the initial host proteins hijacked by C. rodentium during intestinal colonization (12, 13), we examined the effects of the Trp diet and metabolites on the formation of actin pedestals by these proteins during infection. We found that the Trp diet, IEt, IPyA, and I3A all reduced pedestal formation during C. rodentium infection in Drd2fl/fl mice, but this decrease was eliminated in the Drd2fl/fl x Vil-Cre mice (Figure 3.18A). We further found that the Trp diet and metabolites decreased Arp3, N-WASP, and IRTKS expression in the gut epithelium before and during C. rodentium infection in Drd2fl/fl mice, and these effects were also abrogated in the Drd2fl/fl x Vil-Cre mice (Figures 3.18B–C and 3.19). Dopamine had similar effects on actin pedestals and these actin regulatory proteins in C. rodentium infection (Figure 3.11H, J–O). Finally, we determined that IEt, IPyA, and I3A acted via DRD2 to protect against EHEC infection of Caco-2 monolayers by a similar mechanism (Figure 3.20), suggesting that our findings may be generalizable to related AE human pathogens. 147 Figure 3.18 Tryptophan (Trp) metabolites decrease actin pedestal formation via DRD2 during Citrobacter rodentium infection. Drd2fl/fl x Villin (Vil)-Cre mice were fed a conventional (2 g Trp/kg diet, ad libitum) or Trp-rich (42 g Trp/kg diet, ad libitum) diet for 7 d or Trp metabolites, I3A (1000 mg/kg), IEt (600 mg/kg), or IPyA (2900 mg/kg), by oral gavage daily for 2 d, and then infected with C. rodentium (CR, oral gavage, 108 CFU) with continued Trp feeding or metabolite treatment. Intestinal cryosections were stained with (A) DAPI and Alexa Fluor 647-phalloidin or (B–C) antibodies against indicated proteins. Pedestal formation = # of pedestals (Trp diet, n=268– 273; I3A, n=140–187; IPyA, n=162–241; IEt, n=118–133) divided by total number of bacteria. Shown are maximum intensity z-projections (B), and brightness was quantified using FIJI (C). Data are representative of at least 3 independent experiments, n=5 mice per group, bars = mean, error bars = standard deviation. One-way ANOVA followed by post-hoc Tukey’s test: *p<0.05, **p<0.01, ***p<0.001, n.s. = not significant. 148 Figure 3.19 Tryptophan (Trp) diet and metabolites decrease levels of the actin regulatory proteins N-WASP and IRTKS via dopamine receptor D2 (DRD2) during Citrobacter rodentium infection. Drd2fl/fl x Villin (Vil)-Cre mice were fed a conventional (2 g Trp/kg diet, ad libitum) or Trp-rich 149 (42 g Trp/kg diet, ad libitum) diet for 7 d or Trp metabolites, I3A (1000 mg/kg), IEt (600 mg/kg), or IPyA (2900 mg/kg), by oral gavage daily for 2 d, and then infected with C. rodentium (CR, oral gavage, 108 CFU) with continued Trp feeding or metabolite treatment. Intestinal cryosections were stained with antibodies against indicated proteins. Shown are maximum intensity z-projections (A–B), and brightness was quantified using FIJI (C–D). Data are representative of at least 3 independent experiments, n=5 mice per group, bars = mean, error bars = standard deviation. One- way ANOVA followed by post-hoc Tukey’s test: *p<0.05, **p<0.01, ***p<0.001, n.s. = not significant. Figure 3.20 Trp metabolites decrease EHEC actin pedestal formation via DRD2. (A) Polarized Caco-2 (WT or Drd2 KO) monolayers were pre-treated with metabolites (100 M) for 2 d and then infected with EHEC O157:H7 for 16 h. Monolayers were fixed and stained with DAPI and Alexa Fluor 647-phalloidin. Pedestal formation = # of pedestals (I3A, n = 162–170; IPyA, n = 100–106; IEt, n = 144–158) divided by total number of bacteria. (B–G) Polarized Caco-2 monolayers were pre-treated with metabolites (100 M) for 2 d and then infected with EHEC O157:H7 for 16 h. The cells were then fixed, permeabilized and stained with antibodies against the indicated proteins and imaged by confocal microscopy. Shown are maximum intensity z- projections. Scale bars: 20 m. (E–G) Image brightness was quantified using FIJI, n = 3. Data are representative of at least 3 independent experiments, bars = mean, error bars = standard deviation. 150 One-way ANOVA followed by post-hoc Tukey’s test: **p<0.01, ***p<0.001, n.s. = not significant. 3.5 Discussion We demonstrate that metabolites produced by the gut microbiota mediate colonization resistance against AE pathogens, including EHEC and C. rodentium, by decreasing the expression of host actin regulatory proteins that these pathogens target during initial colonization (Figure 3.21). Crucially, these Trp metabolites activate the neurotransmitter receptor DRD2 expressed on IECs to control the expression of these host proteins. Interestingly, a recent high-throughput screen showed that DRD2 may be activated by additional microbial metabolites, suggesting that this receptor may be an important mediator of host-microbe interactions in the gut (33). Our studies reveal that, beyond its canonical roles in the nervous system, DRD2 plays an unconventional role in mediating host defense in the gut. DRD2 and its downstream pathways could serve as targets for prophylactics or therapeutics for improving gut health and treating gastrointestinal infection with certain AE pathogens. 151 Figure 3.21 Proposed model of tryptophan (Trp)-derived metabolite conferral of colonization resistance against attaching and effacing pathogens (AE, e.g., EHEC O157:H7 and Citrobacter rodentium) via the dopamine receptor D2 (DRD2). (A) The Trp-derived metabolites indole-3- ethanol (IEt), indole-3-pyruvate (IPyA), and indole-3-aldehyde (I3A), are produced by the gut microbiota. (B) The Trp metabolites IEt, IPyA, and I3A reduce bacterial pathogen burden through activation of DRD2 and decreasing actin regulatory proteins (e.g., IRTKS, N-WASP, Arp2/3 complex) that are hijacked by AE pathogens to form actin pedestals. (C) Cartoon schematic of the model. Eae, Tir, and Tccp are virulence factors involved in actin pedestal formation (Eae and Tir: EHEC and C. rodentium; Tccp: EHEC only). 152 3.6 References 1. C. G. Buffie, E. G. Pamer, Microbiota-mediated colonization resistance against intestinal pathogens. Nat. Rev. Immunol. 13, 790–801 (2013). 2. P. T. McKenney, E. G. Pamer, From Hype to Hope: The Gut Microbiota in Enteric Infectious Disease. Cell. 163, 1326–1332 (2015). 3. S. Caballero, E. G. Pamer, Microbiota-Mediated Inflammation and Antimicrobial Defense in the Intestine. Annu. Rev. Immunol. 33, 227–256 (2015). 4. P. C. Dorrestein, S. K. Mazmanian, R. Knight, Finding the Missing Links among Metabolites, Microbes, and the Host. Immunity. 40, 824–832 (2014). 5. M. S. Donia, M. A. Fischbach, Small molecules from the human microbiota. Science. 349, 1254766 (2015). 6. A. J. Bäumler, V. Sperandio, Interactions between the microbiota and pathogenic bacteria in the gut. Nature. 535, 85–93 (2016). 7. M. A. Croxen, B. B. Finlay, Molecular mechanisms of Escherichia coli pathogenicity. Nat. Rev. Microbiol. 8, 26–38 (2010). 8. J. B. Kaper, J. P. Nataro, H. L. T. Mobley, Pathogenic Escherichia coli. Nat. Rev. Microbiol. 2, 123–140 (2004). 9. J. Garmendia, G. Frankel, V. F. Crepin, Enteropathogenic and Enterohemorrhagic Escherichia coli Infections: Translocation, Translocation, Translocation. Infect. Immun. 73, 2573– 2585 (2005). 10. S. E. Battle, M. J. Brady, S. K. Vanaja, J. M. Leong, G. A. Hecht, Actin Pedestal Formation by Enterohemorrhagic Escherichia coli Enhances Bacterial Host Cell Attachment and Concomitant Type III Translocation. Infect. Immun. 82, 3713–3722 (2014). 153 11. E. M. Mallick, J. J. Garber, V. K. Vanguri, S. Balasubramanian, T. Blood, S. Clark, D. Vingadassalom, C. Louissaint, B. Mccormick, S. B. Snapper, J. M. Leong, The ability of an attaching and effacing pathogen to trigger localized actin assembly contributes to virulence by promoting mucosal attachment. Cell Microbiol. 16, 1405–1424 (2014). 12. J. W. Collins, K. M. Keeney, V. F. Crepin, V. A. K. Rathinam, K. A. Fitzgerald, B. B. Finlay, G. Frankel, Citrobacter rodentium: infection, inflammation and the microbiota. Nat. Rev. Microbiol. 12, 612–623 (2014). 13. C. Mullineaux-Sanders, J. Sanchez-Garrido, E. G. D. Hopkins, A. R. Shenoy, R. Barry, G. Frankel, Citrobacter rodentium–host–microbiota interactions: immunity, bioenergetics and metabolism. Nat. Rev. Microbiol. 17, 701–715 (2019). 14. R. J. Law, L. Gur-Arie, I. Rosenshine, B. B. Finlay, In Vitro and In Vivo Model Systems for Studying Enteropathogenic Escherichia coli Infections. Cold Spring Harb. Perspect. Med. 3, a009977–a009977 (2013). 15. B. P. Willing, A. Vacharaksa, M. Croxen, T. Thanachayanont, B. B. Finlay, Altering host resistance to infections through microbial transplantation. PLoS One. 6, 2–10 (2011). 16. N. Kamada, Y.-G. Kim, H. P. Sham, B. A. Vallance, J. L. Puente, E. C. Martens, G. Nunez, Regulated Virulence Controls the Ability of a Pathogen to Compete with the Gut Microbiota. Science. 336, 1325–1329 (2012). 17. K. J. Rangan, H. C. Hang, Biochemical Mechanisms of Pathogen Restriction by Intestinal Bacteria. Trends Biochem. Sci. 42, 887–898 (2017). 18. P. V. Chang, Chemical Mechanisms of Colonization Resistance by the Gut Microbial Metabolome. ACS Chem. Biol. 15, 1119–1126 (2020). 154 19. M. G. Rooks, W. S. Garrett, Gut microbiota, metabolites and host immunity. Nat. Rev. Immunol. 16, 341–352 (2016). 20. T. S. Postler, S. Ghosh, Understanding the Holobiont: How Microbial Metabolites Affect Human Health and Shape the Immune System. Cell Metab. 26, 110–130 (2017). 21. Y. Belkaid, T. W. Hand, Role of the Microbiota in Immunity and Inflammation. Cell. 157, 121–141 (2014). 22. J. R. Brestoff, D. Artis, Commensal bacteria at the interface of host metabolism and the immune system. Nat. Immunol. 14, 676–684 (2013). 23. S. A. Scott, J. Fu, P. V Chang, Microbial tryptophan metabolites regulate gut barrier function via the aryl hydrocarbon receptor. Proc. Natl. Acad. Sci. U.S.A. 117, 19376–19387 (2020). 24. A. Kumar, V. Sperandio, Indole signaling at the host-microbiota-pathogen interface. MBio. 10, e01031-19 (2019). 25. T. Bansal, R. C. Alaniz, T. K. Wood, A. Jayaraman, The bacterial signal indole increases epithelial-cell tight-junction resistance and attenuates indicators of inflammation. Proc. Natl. Acad. Sci. U.S.A. 107, 228–233 (2010). 26. B. Bommarius, A. Anyanful, Y. Izrayelit, S. Bhatt, E. Cartwright, W. Wang, A. I. Swimm, G. M. Benian, F. C. Schroeder, D. Kalman, A Family of Indoles Regulate Virulence and Shiga Toxin Production in Pathogenic E. coli. PLoS One. 8, e54456-16 (2013). 27. M. J. Keiser, B. L. Roth, B. N. Armbruster, P. Ernsberger, J. J. Irwin, B. K. Shoichet, Relating protein pharmacology by ligand chemistry. Nat. Biotechnol. 25, 197–206 (2007). 28. C. Missale, S. Russel Nash, S. W. Robinson, M. Jaber, M. G. Caron, Dopamine receptors: From structure to function. Physiol. Rev. 78, 189–225 (1998). 155 29. R. R. Beaulieu, J; Gainetdinov, The Physiology, Signaling, and Pharmacology of Dopamine Receptors. Pharmacol. Rev. 63, 182–217 (2011). 30. J. J. Kulagowski, H. B. Broughton, N. R. Curtis, I. M. Mawer, M. P. Ridgill, R. Baker, F. Emms, S. B. Freedman, R. Marwood, S. Patel, S. Patel, C. I. Ragan, P. D. Leeson, 3-[[4-(4- Chlorophenyl)piperazin-1-yl]-methyl]-1H-pyrrolo[2,3-b]pyridine: An antagonist with high affinity and selectivity for the human dopamine D4 receptor. J. Med. Chem. 39, 1941–1942 (1996). 31. B. F. Binkowski, B. L. Butler, P. F. Stecha, C. T. Eggers, P. Otto, K. Zimmerman, G. Vidugiris, M. G. Wood, L. P. Encell, F. Fan, K. V. Wood, A luminescent biosensor with increased dynamic range for intracellular cAMP. ACS Chem. Biol. 6, 1193–1197 (2011). 32. G. Barnea, W. Strapps, G. Herrada, Y. Berman, J. Ong, B. Kloss, R. Axel, K. J. Lee, The genetic design of signaling cascades to record receptor activation. Proc. Natl. Acad. Sci. U.S.A. 105, 64–69 (2008). 33. H. Chen, P. K. Nwe, Y. Yang, C. E. Rosen, A. A. Bielecka, M. Kuchroo, G. W. Cline, A. C. Kruse, A. M. Ring, J. M. Crawford, N. W. Palm, A Forward Chemical Genetic Screen Reveals Gut Microbiota Metabolites That Modulate Host Physiology. Cell. 177, 1217-1231.e18 (2019). 156 Chapter 4: Conclusion IBDs and infections with enteric pathogens are significant medical conditions that affect millions of individuals. Identifying and understanding the mechanism of action of gut microbially produced metabolites that play a role in ameliorating these conditions may inspire development of prophylactic or therapeutic treatments based on the elucidated pathways. Thus, the effects of these microbially derived metabolites, particularly indole containing Trp derivatives, is an area of increasing study. To broaden our understanding of the roles of such metabolites during IBD, we utilized the in vivo model of DSS-induced colitis and the in vitro model of Caco-2 monolayers treated with the pro-inflammatory cytokine TNFα to screen Trp metabolites for protective effects. We determined that, in our models, I3A, IPyA, and IEt contributed to maintenance of the intestinal epithelial barrier, reducing permeability during insult. Administration of I3A, IPyA, and IEt led to a reduction in the active forms of the actin regulatory proteins ezrin and myosin IIA, which resulted in decreased junctional complex opening. Importantly, we determined that these effects were dependent on the host receptor AhR. Our future work with Trp metabolites in IBD models may address cell types other than IECs and improve the understanding of how these metabolites exert effects on the host immune system during IBDs. We aim to explore whether metabolite treatment affects the frequency of immune cell types such as CD4+ T cells, including Tregs which are associated with protective effects during IBD, and Th17 and Th1 cells that are associated with increased inflammation during IBD. Beyond determining how metabolite treatment affects immune cell populations, we will determine if the metabolites affect cytokine levels, including TNFα, IFNγ, IL-10, and IL-22, which play a role in immune homeostasis. Additionally, we will elucidate the effects of metabolite treatment in floxed mice with AhR specifically knocked out in immune cell types known to express 157 AhR, including CD4+ T cells, macrophages, and dendritic cells. We will also explore the contribution of AhR in IECs during IBD. Given our findings that I3A, IPyA, and IEt exert a protective effect in our IBD models by targeting a host receptor and downstream host proteins within the intestinal epithelium, we next sought to determine if these metabolites would also exert protective effects during infection with an enteric pathogen that hijacks host proteins within the intestinal epithelium. To this end we utilized the A/E pathogen EHEC O157:H7 and the physiologically relevant mouse model pathogen C. rodentium for the second half of my thesis work. Again, after screening various Trp metabolites, we determined that I3A, IPyA, and IEt were indeed ameliorative during these infections. While our initial work on this project was focused on the intestinal epithelial barrier once again, we discovered significant differences in the way I3A, IPyA, and IEt were modulating upstream proteins, specifically host proteins associated with the attaching and effacing aspect of EHEC or C. rodentium infection. I3A, IPyA, and IEt reduced the levels of host pedestal associated proteins N-WASP, IRTKS, and Arp2/3 and therefore reduced attachment and pedestal formation, contributing to colonization resistance. However, the effects of the metabolites during EHEC and C. rodentium infection did not appear to be significantly dependent on AhR, unlike in our previous work. We utilized SEA and CRISPR/Cas9 mutagenesis to identify potential receptors, and ultimately found that DRD2, a host neurotransmitter receptor, is a novel receptor for I3A, IPyA, and IEt, and key to their ameliorative effects during EHEC and C. rodentium infection. Interestingly, we determined that I3A, IPyA, and IEt were DRD2 receptor-specific, and did not appear to activate the other D2-like receptors DRD3 and DRD4, making these metabolites unique. As many D2-like receptor ligands are not DRD2-specific, these metabolites may hold additional 158 promise in the context of studying DRD2-specific responses or as inspiration for the development of pharmacological DRD2 agonists, which may be important in the brain as well as the gut. Ultimately, our findings underscore the importance of host receptors and pathways in mediating the effects of microbially-produced metabolites in the gut and may indeed inspire development of potential prophylactic or therapeutic treatments. Additional work with Trp metabolites may reveal yet other novel receptors for these metabolites or help to determine which bacterial species produce them within the gut. The activity of other dietary-derived metabolites can also be explored, as well as effects of metabolites during other infection models, including those where enteric infections and IBDs co-occur. It would also be interesting to explore if and how microbial diversity is affected when mice are treated with specific metabolites, and how this may impact host health. These future directions present exciting new opportunities for research that may have significant implications for human health. The three metabolites that we determined to exert effects in both our model of IBD and during A/E pathogen infection all share the indole ring but have different side chains. If the common indole ring was the key to the metabolite’s effects, then we would expect to see all the indole metabolites we considered exerting similar effects, but this was not the case in our work, though other groups have elucidated important roles for these other metabolites. Perhaps the answer to why these particular metabolites exert effects in our models lies not with the metabolites themselves, but rather with the receptors to which we have determined they bind. AhR and DRD2 are not necessarily structurally similar, nor do they play similar roles in the established literature, therefore it was surprising, as well as exciting, to us to discover DRD2 as a novel receptor for our metabolites. Based on our SEA results, it is very possible that our metabolites bind to yet other functionally and structurally distinct receptors. This promiscuity may indicate that there is 159 something similar about these receptors that has not previously been identified, and it is also possible that these receptors may work together in some sort of synergistic relationship within the gut. However, the relative importance of AhR in our IBD model and DRD2 in our A/E pathogen infection model may provide additional or new insight into the effects of these insults within the gut and the host response. It is possible that there is an evolutionary advantage to having multiple receptors recognize multiple microbially-derived metabolites that are beneficial during disease states. Given that the gut is composed of so many different bacterial species, that the microbiome varies between individuals, and that at any given time the gut may face multiple different insults, it would be advantageous for the host to be able to utilize more than one metabolite to mediate effects across different receptors to deal with these different insults, which can potentially occur at the same time. Broadly, our work provides unique insights into this complex relationship of the gut microbiome and host response to microbial metabolites. 160