ASSESMENT OF INSECT ANTMICROBIAL PEPTIDES IN MANAGEMENT OF ENTERIC REDMOUTH DISEASE IN RAINBOW TROUT (ONCORHYNCHUS MYKISS) 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 Nathaniel Alexander Smit Sibinga August 2022 © 2022 Nathaniel Alexander Smit Sibinga ASSESMENT OF INSECT ANTMICROBIAL PEPTIDES IN MANAGEMENT OF ENTERIC REDMOUTH DISEASE IN RAINBOW TROUT (ONCORHYNCHUS MYKISS) Nathaniel Alexander Smit Sibinga, Ph. D. Cornell University 2022 Insects are a rich source of bioactive compounds, and they also represent a potentially economical way to convert organic waste into high quality protein meal for use in animal feeds. However, these applications have rarely been considered simultaneously; that is, little attention has been paid to whether bioactive compounds present in insect meals from species such as black soldier fly (Hermetia illucens), yellow mealworm (Tenebrio molitor), or common housefly (Musca domestica) might affect their utility as a protein source. In this dissertation, I use enteric redmouth disease and its causative agent, Yersinia ruckeri, as a model to probe the effects of insect antimicrobial peptides on the microbiome and infection state in rainbow trout (Oncorhynchus mykiss). These studies contribute to our understanding of this topic in the following areas: 1) development of a qPCR-based assay for detection of Y. ruckeri in the intestine of fish; 2) assessment of the effects of a purified insect AMP on the survival, microbiome, and carrier state of trout infected with Y. ruckeri; 3) comparison of diets made from insects with different levels of AMP transcription on survival, microbiome, and carrier state of trout infected with Y. ruckeri. The findings of these studies lead to several conclusions in addition to opening a number of new paths of inquiry. Y. ruckeri is capable of persisting in the tissues of infected fish for at least several months after exposure, and detection rates vary significantly depending on the assay and tissue used. A diet containing the insect AMP cecropin A not only altered the composition of the gut microbiome, but also appeared to increase the load of Y. ruckeri in fish that survived infection. This suggests that at least under certain conditions, the presence of insect AMPs in the diet could lead to undesirable effects. Finally, using a genetic model of immune regulation to vary the levels of AMPs between two insect meals did not result in significant differences in infection state. BIOGRAPHICAL SKETCH Nathaniel Sibinga grew up in Cambridge, MA and Chappaqua, NY, where he graduated from Horace Greeley High School. In high school he played basketball, was a member of the Science Olympiad team, and received the Bausch + Lomb Honorary Science Award. He attended Deep Springs College in the high desert of eastern California, where he served as student body president, curriculum committee chair, and unofficial athletic director. Upon graduation from Deep Springs, he transferred to Brown University where he earned a Bachelor of Science with honors in marine biology. At Brown, Nathaniel pursued his interest in community-building by joining the Brown University Mediation Project, where he also served as treasurer. Nathaniel’s research career began in earnest in the aquaculture lab of Professor Martin Schreibman. While a masters student at Brooklyn College, he taught undergraduate courses and sold renewable energy on the streets of New York City to make ends meet. Following his time in Brooklyn, Nathaniel moved to Bergen, Norway on a Fulbright Fellowship to study salmon nutrition. His talk at the 2016 International Symposium of Fish Nutrition and Feed received first prize for student presentations. Nathaniel came to Cornell in 2016, where he joined Telluride House – a self- governing community of undergraduates, graduate students, and faculty. At Telluride, he served in a variety of leadership positions, including house president. For his research on insect-based fish feeds, he was awarded predoctoral fellowships from both NE SARE and USDA. After graduation, Nathaniel will be moving to Belgium where he will continue his work on insect biology and sustainable aquaculture. v For Alicia vi ACKNOWLEDGMENTS I am profoundly grateful to Hélène Marquis, who took a chance on me when I had nowhere else to go, and who has gracefully alternated for all these years between being a mentor, an editor, a boss, an advocate, a lab mate, and a friend. I’m sorry I didn’t finish before you retired; thank you for being patient. My graduate committee members, Vimal Selvaraj and Cliff Kraft, were likewise unflappable and supportive sounding boards throughout my PhD. I’d also like to recognize the contributions of my unofficial committee members, Rod Getchell and Eugene Won, who were far more invested in my progress and growth than they had any reason to be. I had the good fortune to work with excellent colleagues: Jen Ren, Kelly Sams, Rachael Labitt, Loredana Locatelli, Kari Brossard Stoos, and Marvin Ho all made it fun to come to work and gave freely of their knowledge and time. The staff in my home departments of food science and microbiology and immunology were always helpful and supportive; I want to thank Erin Atkins, Karen Hollenbeck, and Doug Haner – who consistently surprised me with not only their ability but also their willingness to help me out again and again. I learned so much from so many people at Cornell. I’m especially grateful to my collaborators, Nicolas Buchon and Min-Ting Lee, who gave me crash courses in insect immunology and microbiome analysis, respectively. To all the brilliant and caring people I had the privilege of coming home to at the Telluride house, thank you for being my Cornell family. And especially to Londell, my friend through it all, thank you for helping me see the best in people. Before I ever came to Cornell, a great many people helped set me on this path. From a young age, my parents have encouraged me to think critically and have vii graciously put up with my critical thinking ever since. My granny shared my enthusiasm for sustainable fish farming from my earliest days of backyard experimentation and has remained a committed believer in my work ever since. And my aunt Sara helped me to imagine myself as someone who could (and should) pursue a PhD. I’ve been fortunate to have inspiring teachers at every stage of my education – I particularly want to recognize the incredible science teachers I had at Horace Greeley High School: Richard Erhardt, Paul Bianchi, Richard Goodman, and Bob Oddo. My classmates at Deep Springs cemented in me the importance of doing meaningful, ethical work, for which I will always both appreciate and resent them. Martin Schreibman rekindled my love of school and taught me how to successfully navigate academic bureaucracy through persistence, wit, and a dash of feigned ignorance. My research mentors in Norway – Nina Liland, Erik-Jan Lock, and Rune Waagbø – gave me a warm welcome to the big leagues of aquaculture research. Finally, this PhD would not have been possible without my remarkable wife Alicia – the Odysseus to my Penelope – who crossed an ocean to make sure I got through it in one piece. viii TABLE OF CONTENTS Biographical Sketch v Dedication vi Acknowledgments vii List of Figures x List of Tables xii Preface xiii Chapter 1: Introduction 1 Chapter 2: Tissue specific differences in detection of Yersinia ruckeri carrier 35 status in rainbow trout (Oncorhynchus mykiss Chapter 3: Longitudinal sampling of the rainbow trout (Oncorhynchus mykiss) 59 microbiome reveals effects of dietary cecropin A and Yersinia ruckeri infection Chapter 4: Use of an immune-overexpressing insect line to investigate the effect of antimicrobial peptides (AMP) on fish health and microbiome 103 Chapter 5: Conclusions and Future Perspectives 143 ix LIST OF FIGURES Figure 2.1. Kaplan-Meier survival curve during acute Y. ruckeri infection 46 Figure 2.2. qPCR threshold count (Ct) for intestine and spleen controls 47 Figure 2.3. Distribution of tissues positive for Y. ruckeri over time 50 Figure S2.1. Standard curves for intestine and spleen samples 51 Figure 2.4. Estimated Y. ruckeri CFU in intestine and spleen over time 52 Figure 3.1. Dose-dependent inhibition of Y. ruckeri by cecropin A in vitro 74 Figure 3.2. Kaplan-Meier curve showing survival of fish fed either the control or the diet supplemented with cecropin A following an immersion challenge with Y. 75 ruckeri Figure 3.3. Microbiome comparison between mock-challenged fish fed either 79 control or +cecropin diet Figure S3.2. Volcano plot showing taxa significantly enriched in microbiomes of 80 mock-challenged fish fed the control or +cecropin diets at the day 30 time point Figure 3.4. Microbiome comparison between fish challenged with Y. ruckeri fed 83 either control or +cecropin diet Figure 3.5. Correlation analysis between ASV_3428 and ASV_1057 identified as 86 Y. ruckeri and elements of the microbiome Figure 4.1. Summary of the major insect immune pathways 108 Figure S4.1. Expression levels of the AMPs cecropin c and drosomycin by 110 Drosophila line and heat-shock treatment Figure 4.2. Difference in expression levels of two AMP genes between the low 120 and high AMP fly lines Figure S4.2. Expression levels of cecropin c over the collection period 121 Figure 4.3. Kaplan-Meier survival curve for fish fed either the high AMP insect 122 diet or the low AMP insect diet following immersion challenge with Y. ruckeri Figure 4.4. Estimated Y. ruckeri burden per 20mg of intestinal tissue 124 x Figure 4.5. Microbiomes of mock-challenged fish fed either the low AMP or high 126 AMP insect diet Figure 4.6. Microbiomes of fish challenged with Y. ruckeri and fed either the low 128 AMP or high AMP insect diet xi LIST OF TABLES Table 2.1. Prevalence of Y. ruckeri detection in tissue samples of fish collected at 49 various time points post-infection Table S3.1. Nutritional analysis of experimental diets 67 Table S3.2. Distribution of retained samples after sequencing quality control 72 Table 3.1 Detection of Y. ruckeri 16S DNA in intestinal extracts by qPCR 77 Table S3.3. Taxonomic assignments of ASVs enriched in microbiomes of fish fed 81 the control and +cecropin diets 30 days after mock-challenge Table S4.1 Primer sequences 112 Table 4.1. Composition of diets (g kg–1) 113 Table 4.2. Proximate composition of experimental insect meals and diets (g kg–1) 113 Table 4.3. Detection of the Y. ruckeri 16S gene by qPCR in intestinal DNA 123 extracts Table S4.2. Samples sequenced and retained after quality control 129 xii PREFACE This dissertation might be described as existing at the intersection between insect-derived feed ingredients and fish health. This is not an immediately obvious place to locate a dissertation; in this preface, I will attempt to briefly trace the intellectual journey that led me here. My hope is that by the end of this thesis the reader will understand my enthusiasm for this intersection and perhaps share some of my optimism for the future research to be done in this area. My undergraduate training was in marine ecology, with an emphasis on the New England coastal environment. In this setting, twenty years removed from the collapse of the Atlantic northwest cod fishery, it was impossible to ignore the connections between human impacts and the natural environment. Humans had pushed the cod fishery to the breaking point, only for the devastation to wash right back onto shore. Cities like Gloucester, Falmouth, and New Bedford that had been built on one of the most productive fisheries in world history appeared to me by then as so many empty shells on the beach. During my studies it became clear to me that while ecologists could explain what had happened in exacting detail, they did not have a path forward from this mutual destruction. (To be fair, I don’t think anyone had a path forward, but I mostly knew ecologists in those days.) It was the desire for a constructive solution to the ravages of overfishing that led me to focus my attention on sustainable aquaculture. I joined the lab of Martin Schreibman at Brooklyn College, where I was able to gain firsthand experience working with fish. My hope at that time was to decouple the xiii production of farmed fish from the use of fishmeal and fish oil derived from wild- caught fish – a topic discussed at greater length in the following introduction chapter. My masters thesis was about the use of duckweed (Lemna minor) as a feed supplement for Nile tilapia (Oreochromis niloticus). Between Brooklyn College and starting my PhD at Cornell, I was fortunate to have the opportunity to spend one year in Norway doing research on insect-based aquaculture feeds. In that project, we explored whether feeding seaweeds to farmed insects could enrich their omega-3 fatty acid content and thereby make them a better replacement for fishmeal. The conclusion was essentially that it was possible to enrich the insects, but not at a sufficient level to compensate for the observed decrease in growth rate. The design (and the outcome) of this study have been important in my own thinking over the course of this PhD. I came to Cornell intending to focus on fish nutrition. I was interested in insect ingredients, plant ingredients, algae ingredients – pretty much any combination that could reduce the use of fishmeal and fish oil in aquaculture feeds. When I found my research footing at Cornell, however, it was in the hybrid space between nutrition and fish health that I now occupy. I inherited a system for producing housefly larvae using dairy cow manure and I proceeded to turn these larvae into ingredients for fish feed. A study had recently been published which showed complete protection from bacterial infection in fish given diets containing housefly pupae. This was an intriguing finding, and we thought we were in a good position to try and reproduce this result and understand mechanistically what was causing the protective effect. It turned out to be quite a challenging study, however, and it took years to identify and develop a suitable infection challenge model. xiv In the course of developing the Yersinia ruckeri infection model that ultimately became central to my PhD, I expanded my thinking about sustainable aquaculture to include disease management. The recognizable dynamics of mutual destruction that I first absorbed thinking about cod fishing are apparent in the growing antimicrobial resistance crisis. Following the sudden closure of the university in response to the COVID-19 pandemic, I developed a somewhat spontaneous experiment detailed in chapter 2. In that study, I adapted a method for sensitive detection of Y. ruckeri for use in fish intestinal tissue and used it to re-examine old findings about carrier status in survivors of this infection. During my PhD, interest in insect-based aquaculture feeds has increased dramatically (from 17 hits for publications under “insect feed aquaculture” in 2016 to 107 in 2021, per Web of Science). One of the striking features of this body of literature is its variability – there are results to support a wide variety of conclusions about the suitability of insects as fish feed ingredients. I became interested in finding a mechanism that could not only explain protective effects of insect ingredients but also why these effects appeared so inconsistently. The central hypothesis of this dissertation coalesced around the insect antimicrobial peptide response, which is variable in response to infection. It is ultimately this malleability of insect biology – first highlighted for me during my time in Norway – that I think is the most exciting and challenging aspect of their use in feeds moving forward. Chapters 3 and 4 focus on the impact of insect antimicrobial peptides in the context of Y. ruckeri challenge. xv CHAPTER 1 Introduction 1 The potential of insects to improve the environmental sustainability of fish feed in the aquaculture industry History of fish feed composition and demand for alternative ingredients Scientific discussion of aquaculture feed ingredients has historically revolved around reducing the use of fishmeal (FM) and fish oil (FO). The excellent nutritional properties and relative availability of these ingredients, combined with the absence of regulation or scrutiny of environmental impacts led to high inclusion levels in feeds for farmed fish in the early days of commercial aquaculture. Indeed, marine ingredients comprised nearly 90% of a typical feed for Atlantic salmon (Salmo salar) farmed in Norway in 1990 (1). As the scale of aquaculture expanded, however, this level of FM and FO inclusion became impractical for a number of reasons. First, the majority of FM and FO are produced from wild fish, with anchoveta from Peru and Chile representing the largest single source. Anchoveta supply is affected dramatically by El Niño years, introducing a significant source of price volatility. A turning point came in 2006, when a severe El Niño effect and other factors caused FM prices to jump to twice their previous 30-year high (2). At the same time, use of FM and FO derived from wild fish has come under heavy criticism from environmental advocacy groups. An often cited (and just as often critiqued) metric of sustainability has been the Fish-In-Fish-Out (FIFO) ratio of aquaculture. At stake in this metric is the question of whether aquaculture is a net producer or net consumer of fish. The twin pressures of rising prices and concern from environmental groups over the impact of aquaculture on wild fish populations have continued to drive significant investment in the development of alternative ingredients and steady reductions in FM and FO have been 2 achieved. By 2012, scientists were proudly announcing that they had achieved FIFO ratios of less than 1 for Atlantic salmon (3) using diets containing less than 25% FM and FO. And by 2016, total inclusion levels of FM and FO in industry settings had also dropped below 25% for Atlantic salmon in Norway (1). There are, however, a number of important caveats to that story. The first is that demand for FM and FO varies greatly depending on the type of aquaculture production. Feeds for salmonid and other marine fish, as well as shrimp, have relatively high requirements for nutrients that are difficult to provide without use of FM and FO. Inclusion levels in feed for these species are thus quite high when compared to feeds for freshwater fish like Tilapia or catfish which typically contain only 2-4% FM and FO (4). While Atlantic salmon (4.5%) and rainbow trout (1.6%) comprised just over 6% of all finfish production in 2018, 17% of finfish production worldwide came from traditional carp pond production systems that rely on fertilization of primary producers and don’t use manufactured feed at all (5). The scientific literature on aquaculture nutrition focuses overwhelmingly on production of species with the highest typical inclusion levels for FM and FO. Another consideration is that these shifts in feed composition have taken place within the context of rapid growth of aquaculture production around the world. Between 1995 and 2015, global production of aquaculture feed increased sixfold (4). This means that even as inclusion rates in feed decreased, the total share of FM used annually for aquaculture feeds grew from 10% of the global supply in 1980 to 73% in 2010 (6). Given that FM and FO production have been harvested at close to their maximum sustainable yields for the past 25 years, stretching a fixed supply of FM and 3 FO across more feed can be seen as a simple requirement of increasing feed production. The majority of FM and FO replacements to date have been accomplished using terrestrial plant ingredients – especially soybeans. While this approach has been successful in reducing use of FM and FO and keeping feed costs down, it is less clear that it is environmentally friendly in other ways. The single-minded focus on reducing FM and FO has tended to treat any alternative ingredient as ipso facto sustainable. And while industry and environmental groups have generally been aligned in their desire to reduce use of FM and FO, it is important to remember that the primary motivation of feed producers has generally been cost reduction rather than sustainability. As we move towards a future where cost and sustainability may well diverge, a more holistic evaluation of feed sustainability than inclusion levels of FM and FO is required. A number of studies have applied life cycle analysis (LCA) methods to evaluate the environmental impact of various aquaculture production systems (7,8). These analyses tend to converge around two major points: 1) feed is the major driver of impacts in most intensive aquaculture systems, and 2) almost any change to a production system represents a tradeoff with mixed impacts. One recent study found that high inclusion levels of terrestrial animal byproducts in farmed salmon diets resulted in more than triple the greenhouse gas emissions when compared to a typical diet using a mix of FM, FO, and plant ingredients (9). Another study demonstrated that replacing FM in shrimp diets with plant-derived ingredients would significantly exacerbate existing environmental impacts of terrestrial industrial agriculture 4 (specifically use of land, freshwater, and phosphorous), though the authors were careful not to argue that this meant FM and FO should be favored (10). The key takeaway from this body of literature is that all feed ingredients have environmental impacts, and systematic evaluation of these ingredients is required to improve the sustainability of aquaculture. Insect production systems and environmental impacts of insect farming Insect meals have generally rated highly in sustainability analyses when compared to other protein sources. A direct comparison of several FM alternatives – microalgae biomass, insect meal made from Black Soldier Fly (BSF) larvae, and poultry byproduct – concluded that insect meal had the lowest overall energy cost when accounting for both natural and artificial processes (11). A similar analysis of BSF insect meal in the context of poultry farming also concluded that insect meal was more sustainable than soybean meal (12). These results do not come as a large surprise; the (still mostly theoretical) environmental sustainability of insect production has in many ways been the driving force behind the explosion of research and investment capital to support insect farming over the past ten years. Insect farming is a cornerstone of many proposals for how industrial agriculture can move towards a circular economy model that minimizes waste and reduces environmental impacts (13,14). Insects can be produced in a small area using a wide variety of organic waste streams. Insects extract nutrition from low- value organic waste, “upcycling” a percentage of that waste into higher value protein 5 and lipid. The remainder after insect production is known as frass, which can be used as fertilizer for plants (15). The amount of water required for mealworms, a model system for farmed insects in both academia and industry, is similar to chickens and less than pigs or cattle on a per weight basis (16). Furthermore, a variety of insect species convert readily available food wastes for growth at least as efficiently as conventional livestock on formulated feeds (17). This is remarkable: insects are efficient converters of waste products even when compared to chickens and pigs raised on optimized diets that represent a major cost to farmers and the environment. The amount of land required to produce a given amount of insect protein is half that of high-intensity poultry, and less than a third that of pigs (18). Space-efficient production means that replacing plant-based ingredients with insect meals can reduce the amount of land in cultivation and help to slow deforestation – one of the major drivers of climate change associated with agriculture. Insects fit strongly into the framework of sustainable intensification (19). Not only are they incredibly well-suited to high-density production with relatively small environmental impacts, insect rearing can also ameliorate negative effects of intensive production of other species. A benefit of the small spatial footprint required is that insect production facilities can be located close to point sources of organic waste, minimizing transportation. One of the major trends across all species of livestock production over the past 50 years has been the concentration of production into fewer farms with more animals. This in turn has created larger point sources of animal manure which can pose a variety of environmental and human hazards if not managed 6 properly (20,21). Raising insects on animal manure has a number of positive effects independent of any value that might be harvested from the resulting insects: insect treatment can reduce biomass, water content, odor, nitrogen, phosphorus, and pathogenic and antibiotic resistant bacterial populations in manure (22,23). Environmental sustainability has thus far been the core appeal of insect production. Their unparalleled ability to rapidly process large volumes of organic waste into protein and lipids has the potential to improve the efficiency and sustainability of existing agricultural supply chains. It is this optimism, more than an established track record of profitability, that has driven large scale investment into insect production systems. Ultimately, however, insect-based feed ingredients will need to provide surplus value in a highly competitive agricultural commodities market if they are to realize these environmental benefits. Nutritional requirements of fish and general nutritional properties of insects Modern feed formulation is premised on the idea that nutritional requirements exist entirely independently from individual ingredients in the diet; that is, fish have requirements for nutrients, not for ingredients. In this paradigm, a feed formulation is a puzzle that must minimize costs while meeting the various dietary requirements for energy, protein, essential amino and fatty acids, minerals, vitamins, etc. Individual ingredients are the puzzle pieces that contribute towards meeting one or more these requirements. Acceptance of at least part of this outlook is apparent in the sheer number of studies touting various ingredients as replacements for FM and FO. However, the idea of any one ingredient directly substituting for another only works if 7 a) the nutritional profiles of the ingredients are very similar, or b) the differences are primarily in nutrients that are extraneous to the requirements of the fish. The difficulty in replacing FM and FO is that these ingredients fail both tests: they are relatively unique – with the minor exception of other wild-caught marine animal products (e.g. krill or squid) with many of the same drawbacks – and they contribute towards more or less every nutritional requirement of farmed fish. Replacement of FM and FO should thus be considered in the framework of a patchwork quilt of ingredients that can act in a complementary fashion to meet nutritional requirements (24). Insect ingredients are not a magic bullet, but they have strong potential to contribute to low- FM/FO diet formulations. Fish have a high apparent requirement for dietary protein. The most intuitive way of explaining this requirement is to flip it around: fish have a relatively low requirement for dietary energy. Owing to their comparative metabolic efficiency (poikilothermy, neutral buoyancy), fish require fewer calories for maintenance energetic costs than endothermic terrestrial livestock. As a consequence, fish are much less reliant than conventional livestock on carbohydrates as an energy source. In formulated aquaculture feeds, a large share of the energetic demands of the fish are met by lipids, with high levels of protein to supply building blocks for tissue growth. Carbohydrates may supply some energy – especially in freshwater species such as catfish, carp, and tilapia – but are a comparatively minor aspect of fish nutrition. This makes insects, which are rich in proteins and lipids and low in digestible carbohydrates, a promising ingredient for aquaculture feeds. Indeed, insects and other arthropods comprise a large part of the natural diet for many fish species. 8 Most farmed insects are harvested at the prepupae stage to minimize the amount of chitin, which for many fish species appears to be largely undigestible (25,26). The concentrations of protein and fat vary considerably depending on the species of insect and the feed substrate, but farmed insects generally fall between 40- 60% crude protein and 10-30% lipid; industrial production of insects typically involves a processing step to separate the protein and lipid fractions. Processing is, for the moment, a relatively opaque part of the insect ingredient picture. A number of collaborative feed trials between industry and academia use differently processed insect ingredients from the same supplier (27) but detailed descriptions of these processing techniques are withheld as proprietary information. The lack of information in the literature should not obscure the importance of this step – the viability of most plant-based ingredients as aquaculture feed ingredients has hinged on the processing step (28,29). Diversity of insects used as feed ingredients and performance in aquafeeds Insects encompass an enormously diverse group of organisms that fill a wide range of ecological niches. Several species are currently being seriously explored for large-scale production, but it is likely that continued research will lead to the utilization of new species and that the emergence of selective breeding programs will build upon existing natural diversity. The key selection criteria for farmed insect species can be split into two categories: rearing and final products. Considerations for rearing include growth rate, generation time, feed conversion efficiency, density of growth, water use, disease resistance, optimal temperature range, native status 9 (potentially invasive species should be avoided), docile behavior (e.g. non-flying, non- jumping, non-biting), and the types, variety, and availability of inputs the insects can consume. Selection based on final product should consider digestibility, protein and lipid content, amino acid profile, fatty acid profile, ease of processing and waste disposal, levels of both beneficial and harmful minerals and chemicals, microbial safety, consumer acceptance, as well as any potential valuable non-feed compounds (e.g. cochineal dye, silk fibers, pharmaceuticals). Prioritization among these considerations is likely to shift based on market forces, and it is easy to imagine that different conditions will favor the production of a variety of different insect species. Aquaculture feeds already include a dizzying array of ingredients, with a large degree of regional variation; feed companies place a high value on flexibility and the ability to adjust formulations to minimize ingredient costs based on local availability. Insect-based ingredients will always exist as one option among many, and their use in aquaculture feed will be based primarily on the tradeoff of cost vs. performance. It is thus difficult to draw strong conclusions about the future economic viability of an ingredient from academic studies, which typically ignore cost in order to measure performance. That said, insect ingredients have shown promise in a number of recent studies. It is notable that though there are many published nutritional studies using insect ingredients (reviewed by (30),(31), and (32)), they must be interpreted critically. Growth rate can only be measured relative to a control diet, and the design of many studies involves direct replacement of a single ingredient in the control diet (usually FM) with an insect meal. While this is consistent with basic principles of experimental 10 design – only alter one variable at a time – the question of whether insect meal is nutritionally equivalent to FM can be confidently answered in the negative without even doing such an experiment. Indeed, the most likely explanation when these studies conclude that FM can be partially replaced by insect ingredients is that the control diet is formulated with wasteful levels of FM that are entirely surplus to both modern commercial diets and the nutritional requirements of the fish. A more informative, though admittedly more complicated, approach is to compare practical formulations that include insect ingredients with control diets that mimic existing commercial diets; in practice this usually means adjusting multiple ingredients. The following targeted review of the literature emphasizes studies with the latter approach. Black soldier flies Larvae of the Black Soldier Fly (BSF) (Hermetia illucens) have rapidly become the model for industrial production of insects. A number of BSF-producing companies have attracted substantial investment in recent years and are expanding production around the world. BSF peak in size at the pre-pupae phase, at which point they are typically about 40% crude protein and 28% crude fat (33), although this can vary significantly based on the feed substrate of the larvae (34). Performance of BSF ingredients in aquaculture feeds has generally been good. Digestibility of feeds with BSF meal generally compares favorably to control diets, though very high inclusion levels (500-600 g/kg feed) can result in reduced digestibility (31). Growth rate compares favorably to control diets at BSF meal inclusion levels of 200 g/kg feed and below for Atlantic salmon (S. salar) (27,35,36), and European seabass (D. labrax) 11 (37). At the high end, a study in freshwater stage Atlantic salmon including 600 g/kg feed of BSF meal performed comparably to the control diet in terms of growth, but while this formulation reduced FM from 350 g/kg feed to 60 g/kg, it also entailed raising the FO content from 46 g/kg to 69 g/kg (38). At the heart of this tradeoff is that many aquaculture species have a dietary requirement for omega-3 fatty acids, of which most insects are a poor source. A similar compromise was necessary in Pacific white shrimp (Litopenaeus vannamei): a diet containing 210g/kg of BSF meal performed similarly to the control diet, allowing reduction of FM from 250g/kg to 100g/kg but requiring an increase of FO from 10g/kg to 25g/kg (39). Cardinaletti et. al used a puzzling formulation in a feeding trial with Rainbow trout (O. mykiss) that decreased both FM (from 420g/kg feed to 210g/kg) and FO (from 70g/kg to 28g/kg) while introducing 210g/kg of full fat ground BSF meal and increasing pea protein concentrate (from 55g/kg to 100g/kg); while the changes in growth were not statistically significant, there was a clear trend towards reduced growth in the BSF diets (40). Reducing FM only to increase FO may seem counterproductive on its face; however, the option to do so if prices or regulations make such a choice advantageous is of practical significance for feed producers. BSF ingredients have clear potential as a bulk protein ingredient, provided they are utilized in well-designed diets that compensate for their low omega-3 fatty acid content. Mealworms Yellow mealworms (Tenebrio molitor) represent the other major farmed insect species. Live mealworms have been produced commercially as a food for exotic pets 12 for many decades, but in recent years there has been significant interest in expanding this production to a commodity scale. Mealworm larvae are typically around 47% crude protein and 30% crude fat, again subject to rearing conditions (41). There is a notable discrepancy in the degree of involvement of large feed companies in mealworm studies: whereas many BSF feed trials involve active industry partners that assist with formulation and diet production, mealworm studies are more likely to have been conducted entirely by academic labs. The lower barriers to production also mean that mealworms have been studied in more parts of the world and more species, but less systematically and with a higher proportion of idiosyncratic experimental designs and contradictory results. Interestingly, despite having a similar level of crude fat to BSF, mealworms have tended until very recently to appear in the literature as full fat meals, rather than partially defatted meals. This may be a consequence of the fact that small- and medium-scale commercial mealworm production pre-dates the recent rise of industrial-scale BSF farming and that mealworm producers are typically less equipped to perform extensive processing. Another possible explanation for the relative lack of processing, however, is that mealworms appear to be highly digestible. Even with minimal processing, mealworms were clearly the most digestible out of five insect meals tested in juvenile Nile tilapia (Oreochromis niloticus) (42). In one of the rare direct comparisons of mealworms and BSF, mealworm meals were more digestible for European sea bass than BSF meals (43). Interestingly, defatted mealworm meal was the most digestible, while full fat mealworm meal performed similarly to defatted BSF meal (43). An intriguing result in rainbow trout is consistent with the idea that while full fat 13 mealworms may be adequate ingredients, there are still gains to be had from processing: defatted mealworm meal produced by Ynsect (the lone industrial-scale mealworm producer) comprehensively outperformed FM in terms of growth and feed conversion (44). Another study using defatted mealworm meal found no difference in growth of rainbow trout with complete replacement of FM up to 200g/kg inclusion (45). Diets including up to 305g/kg feed of full fat mealworm meal showed no difference in growth for Pacific white shrimp, allowing a reduction of FM from 206g/kg feed to 0g/kg and soybean oil from 51g/kg to 0g/kg, although there was an increase in FO from 2.5 g/kg to 9.8g/kg in the highest inclusion diet (46). A study in sea trout (Salmo trutta) found no significant differences in growth rate between the control diet and any of three formulations with differently processed mealworm meals included at 276g/kg feed (47). Mealworms have performed remarkably well in aquaculture feed trials and would seem to have great potential as a feed ingredient. The concern is more on the production side – mealworms grow more slowly and can handle a narrower range of feed inputs than fly larvae. Houseflies Larvae of the housefly (Musca domestica) have also been investigated as a feed ingredient, though there does not appear to be the same push for industrial-scale production. This is likely due to the potential consequences of escaped houseflies; in addition to being pests, their ability to carry pathogens has been extensively documented (48). Given these risks, housefly larvae production would have to provide significant advantages relative to other insect species – either in terms of efficiency of 14 production or in the quality of the final product. One potential advantage of houseflies relative to mealworms is that they can be reared on substrates with very high moisture contents and bacterial loads, including manure from production of swine (49), poultry (50), and dairy (23). BSF, it should be noted, can also be reared on all of these substrates. In both BSF and houseflies, however, manure appears to be a suboptimal feed substrate when compared to diets composed of organic wastes from food production (e.g. wheat bran or spent brewers grains) (51,52). As such, houseflies raised on manure would appear to have limited appeal as an industrial production system. The reason houseflies are still worth considering is that they are already ubiquitous in the context of animal production systems. While transporting manure to a centralized location for production of insects may not make economic sense, small- to medium-scale production on-site at farms could provide added value and help to break down manure (23). Much of the existing literature on housefly ingredients in fish feeds uses freshwater species and control diets with unrealistically high levels of FM for these species, making these results difficult to interpret. A recent result is worth highlighting, however: a study in red seabream (Pagrus major), found that defatted housefly larvae meal performed well at high inclusion levels (400g/kg and 700g/kg) (53). Full fat housefly meals, however, showed reduced growth, even when purified Eicosapentaenoic acid (EPA) and Docosahexaenoic acid (DHA) were supplemented to mimic the levels present in the replaced FO fraction. The authors thus concluded that there are antinutritional factors in the lipid fraction of housefly larvae that can inhibit fish growth even in the presence of apparently adequate nutrients (53). This study 15 highlights the need for systematic research of insect ingredients, as well as the potential importance of processing on ingredient performance. Future prospects as an ingredient The takeaway from this body of literature seems to be that insect meals perform well in rationally-designed diets that take advantage of their highly digestible proteins and lipids and compensate as necessary for their lack of omega-3 fatty acids relative to FM. While that compensation currently requires increasing FO levels, significant progress has been made towards the use of farmed algae as a sustainable source of omega-3 rich oils; these bodies of literature seem certain to intersect in the near future. In the long-term, algae-derived oils are likely to be more significant than insects for the development of nutritionally viable feeds with zero FM or FO. But, as laid out earlier in this chapter, FM and FO levels have always been a reductive framework for thinking about sustainability. Insects have real potential benefits for reducing waste, mitigating emissions, and facilitating feed production using locally abundant supplies. To have any positive effect, however, insect ingredients have to displace existing ingredients and achieve widespread use. There are two major avenues for insect ingredients to gain traction in fish feeds: reduced cost and improved performance. The emphasis of industry thus far appears to have been on reducing cost by increasing the scale of production. There is arguably much more upside to focusing on improved performance. Studies that compare differently processed insect meals (e.g. full fat vs. defatted) have consistently found differences, and one of the great 16 uncontrolled variables across different studies is knowing how similar, for instance, one BSF meal is to another. This makes intuitive sense; the soy protein concentrate that makes it into modern fish feeds by the ton is clearly several industrial processing steps removed from a soybean plant growing in a field. Similarly, animal byproducts result from a hugely complex process involving physical separation, rendering, drying, and milling. Insect larvae are small, but they remain biologically complex and have heterogeneous compositions. Understanding and leveraging that complexity holds tremendous promise for improving their long-term potential utility and value. Over the course of my PhD, I have focused on trying to understand how physiological changes dictated by the insect immune system might alter the properties of these ingredients. Nutritional health management of farmed fish Antibiotics and Antimicrobial Resistance Infectious disease is a major challenge for animal production systems, including aquaculture. Much of the recent growth in food animal production has been achieved by increasing the scale and density of existing farms, amplifying both the probability and the impact of disease outbreaks. On-farm antibiotic use has increased dramatically as a result: as of 2012, consumption of antibiotics by livestock was estimated to be roughly twice that of humans (54) and consumption by farm animals is predicted to rise an additional 67% globally by 2030 (55). Reliable data on agricultural antibiotic use is limited for many parts of the world, and especially so for aquaculture (55,56). Nevertheless, use of antibiotics in aquaculture has been heavily criticized and estimates based on the available information point to clearly irresponsible use in some 17 regions (57–60). Antibiotics in aquaculture are typically administered via inclusion in specially formulated feeds; antibiotic residues from aquaculture diffuse into aquatic environments, which contain a pool of mobile resistance genes (the resistome) that can facilitate rapid emergence of novel multi-drug resistant phenotypes by horizontal gene transfer (61–63). Selection for Antimicrobial Resistance (AMR) has been documented not only in fish pathogens (64,65), but also in environmental isolates (66) and fish microbiota (67). AMR poses a direct threat to aquaculture in the form of increasingly intractable and endemic outbreaks in farm settings. However, perhaps the greater threat is that the danger that widespread AMR poses to human health seems likely to lead to increasingly severe restrictions on the use of antibiotics in the near future. The first wave of regulation targeting agricultural antibiotic use has primarily focused on antibiotic growth promoters (AGP); many countries have now banned AGP. In contrast to terrestrial systems, however, there has been little discussion of growth promoting effects in fish and shellfish; preemptive use of antibiotics in aquaculture is almost always framed as prophylactic (68). The first wave of on-farm antibiotic regulations, targeting AGP, is thus unlikely to necessitate large-scale changes for aquaculture. Where prophylactic antibiotics are still in use however, aquaculture stakeholders would be wise to anticipate the second wave of regulations, targeting all preemptive treatment. Managing infectious disease without antibiotics Managing fish disease without antibiotics is complicated by a variety of factors. First and foremost is the huge number of aquaculture species currently under 18 production. Whereas broad spectrum antimicrobials are at least somewhat applicable against bacterial infections regardless of the fish species involved, other solutions typically require greater understanding of each individual host-pathogen dynamic. Given that there are approximately 600 aquaculture species under cultivation and an ever-growing list of known pathogens, the idea of understanding each potential combination quickly becomes overwhelming. In fact, disease management is one of the major factors pushing aquaculture productions towards a consolidated set of core species (69). Salmonid production on the west coast of North America shifted away from native Pacific salmon species to non-native Atlantic salmon largely because of the difficulties posed by bacterial kidney disease (BKD); an effective vaccine against BKD exists for Atlantic salmon but not for Pacific salmon (70). Similarly, Thai shrimp farms rapidly transitioned from native black tiger shrimp (Penaeus monodon) propagated from wild-caught broodstock to imported Pacific white shrimp (L. vannamei) in the mid 2000s because of the availability of specific-pathogen-free stock reared in indoor facilities by specialized companies (71). Reflection on the sheer scale of the research that has gone into understanding, treating, and preventing diseases of terrestrial livestock – entire departments at universities around the world in the past hundred years alone! – makes clear that intensive production of 600 aquaculture species is almost certainly not feasible. Consolidation of intensive farming around a few core species would thus appear to give the best chance of accumulating sufficient knowledge and expertise to manage and respond to disease without resorting to irresponsible use of antibiotics. 19 This approach, however, is clearly fraught with unintended consequences. Risking introduction of non-native species – as happened repeatedly with Atlantic salmon in the Pacific northwest (72–74) – goes against basic ecological principles and can be very unpopular with consumers; indeed, Washington state banned the production of non-native fish species in 2018 leaving finfish production in the state in a tenuous position. There is also the risk that this will simply consolidate aquaculture production around existing infrastructure and expertise, as in the example of Thai shrimp farming. The existing infrastructure and expertise is at least somewhat attributable to historical accident and it may well be shortsighted to let it continue to dictate future species selection. There is no obvious answer to this dilemma, except to recognize that one of the most exceptional features of antibiotics is their non-specificity and simplicity of use; a future with reduced antibiotic use must necessarily be one with greater understanding of specific disease mechanisms and host-pathogen dynamics. Still, it is important to emphasize that with proper understanding of disease risks and mechanisms, fish farms can come close to eliminating antibiotic use. In Norway, for example, a 99% reduction in the use of antibiotics in aquaculture was achieved between the late 1980s and 2007 despite production volume increasing more than eightfold over the same period (75). This reduction is attributed to stricter regulation and mandatory reporting of all veterinary antibiotic use, along with a proactive commitment to developing alternative strategies for disease control. Among these strategies are a coordinated effort to develop and utilize vaccines against common diseases, mandatory fallow periods without fish between harvest cycles, enforced standards for fish care and welfare, and a rationalized permitting system 20 designed to ensure that outdoor net pens are sited with sufficient distance to prevent farm-to-farm transmission (76). While these specific strategies are not necessarily transferable to other settings – Chilean production of Atlantic salmon remains highly dependent on antibiotic use, for example – they suggest that with sufficient experience, understanding, and regulation it is possible to design aquaculture systems in such a way that large-scale intensive production is viable without antibiotics. Reducing exposure to infection Preventing spread of diseases should always be the first priority. While the field of fish epidemiology is in its infancy, the necessary tools are increasingly available and there is reason to believe that spread of infections can be greatly reduced from current levels. For fish reared outdoors in net pens, limiting the access of potential disease vectors from the environment is an important and often under-studied approach (77). For species reared in tanks or raceways, however, most disease transfer is facilitated by transfer of individuals from one facility to another (78,79). Surveillance testing of fish before they enter a farm (or exit a hatchery) can detect latent infections that may be carried from hatcheries or other farms. In shrimp aquaculture, the development and availability of specific pathogen free (SPF) shrimp stocks has proven transformative for the industry (80). A number of increasingly sensitive qPCR detection assays for a wide variety of diseases have been published in recent years; qPCR-based detection methods can screen for multiple diseases simultaneously either in fish tissues (81) or in water samples (82). While testing of individual samples is not cost-prohibitive for any but the smallest producers, these 21 assays require expensive machinery, clean laboratory space, and trained operators. It is likely not feasible for every farm to run its own tests and it remains an open question how best to make these diagnostic tests available. Dietary approaches to improving disease resistance Animals are highly adaptable. While it is tempting to think of the range of phenotypes in a given species as hard-wired by genetics, each individual genotype in fact retains a great deal of plasticity. A growing organism incorporates a variety of environmental signals to make important physiological tradeoffs. Animals do not simply react to their environment; they constantly attempt to anticipate and adapt to future environmental states. Allocating surplus dietary energy to growth of muscular or adipose tissue can be seen as a tradeoff largely based on anticipation of future food availability. Growing too fast in a resource-poor environment may compromise survival by increasing the cost of metabolic maintenance, while growing too slowly may compromise ability to compete for food or mates in a resource-rich environment. The farm environment is unique: farm animals live with far lower risk of predation or starvation than wild animals, but much higher risk of disease and stress. Tilting animal physiology to better align with this radically altered risk profile is key to optimized farming. While terrestrial farm animals have had hundreds if not thousands of years of careful selective breeding, little attention has been paid to fish genetics until the 1980s (83). Even then, only a few species were initially featured in genetic improvement programs and the selection was almost entirely for increased growth rate. By 2010, one study estimated that less than 10% of global aquaculture 22 production utilized genetically improved stocks (84). As a consequence, most farmed fish species retain a great deal of the genotypic variety and physiological plasticity of their wild counterparts, and disease resistance remains highly variable (85). Over the past 20 years, it has become increasingly evident that the biology of the vertebrate gut extends far beyond nutrient uptake. The gut is intimately involved in chemical sensing of the world and sends signals that interact with almost every part of the body; indeed, the sensory machinery of the gut significantly pre-dates all other sensory organs in an evolutionary sense (86). Understanding how diet can prime metabolism for the challenges a fish will face (and, just as importantly, challenges it won’t face) has potential to improve fish health and growth. A movement within aquaculture nutrition has sought to leverage this insight by designing feeds that interact beneficially with host physiology. Collectively, these are called functional feeds. The array of new discoveries, methods, and technologies around gut physiology is poised to revolutionize this field by allowing scientists to generate and test much more specific hypotheses about potential functional ingredients. In recent years, therefore, there has been a great deal of interest in feed ingredients that can improve fish resistance to disease. There are at least three main pathways to improving disease resistance via the diet, though many of the studies in this area do not differentiate between them. Feeds may 1) directly stimulate increased metabolic investment in the primary and secondary immune response (e.g. increase proliferation or activity of macrophage and other immune effector cells), 2) reduce pressure on these same systems by reducing the infectious burden they are exposed to (via direct killing of pathogens and/or strengthening of barriers: skin, scales, mucus, 23 gut epithelia, etc.), or 3) modify the microbiome in ways that accomplish the previous two ends indirectly. A great many putative functional ingredients have been shown to improve fish health and disease resistance, though these have been studied with varying levels of rigor and reproducibility (87–90). Insect ingredients as functional feeds Insect meals occupy an interesting position. The primary focus has been on their use as bulk protein and lipid ingredients comprising 20% or more of a feed formula. However, purified components of insects including chitin and chitosan (91), polysaccharides (92), lauric acid (93), and antimicrobial peptides (AMPs) (94) have all been shown to be bioactive and protective against disease at much lower (less than 1%) inclusion levels in the diet. How these levels of purified components compare to the levels present in crude insect meals is an interesting comparison that in most cases has yet to be measured. It should be noted, however, that with a few notable exceptions (95–97), most studies that feed crude insect meals have not reported benefits to disease resistance. This suggests that either the levels in crude insect meals are simply not high enough to be bioactive, or that feeding them in combination with other insect components alters their activity in some way. The presence of so many potentially bioactive compounds in insect meals complicates analysis of their effects. All of the components listed above have the potential to vary significantly in concentration between insect meals made from the same species depending on rearing substrate, age at harvest, and ingredient processing. There is a great deal of work still to be done to identify bioactive components of insect 24 ingredients and to assess their functionality (or lack thereof) in practical diet formulations. In some cases, it may be advantageous to maximize bioactive effects in insect meals, while in others these effects may prove deleterious. The goal of this dissertation was to develop a framework for systematic evaluation of the effect of insect AMPs on fish health in the context of infection with Y. ruckeri. 25 References 1. Aas TS, Ytrestøyl T, Åsgård T. Utilization of feed resources in the production of Atlantic salmon (Salmo salar) in Norway: An update for 2016. Aquaculture Reports. 2019 Nov;15:100216. 2. Hardy R. Aquaculture feeds and ingredients: an overview. In: New Technologies in Aquaculture [Internet]. Elsevier; 2009 [cited 2021 Oct 12]. p. 370–86. Available from: https://linkinghub.elsevier.com/retrieve/pii/B9781845693848500129 3. Liland NS, Rosenlund G, Berntssen MHG, Brattelid T, Madsen L, Torstensen BE. Net production of Atlantic salmon (FIFO, Fish in Fish out < 1) with dietary plant proteins and vegetable oils. Aquacult Nutr. 2013 Jun;19(3):289–300. 4. FAO, editor. Meeting the sustainable development goals. Rome; 2018. 210 p. (The state of world fisheries and aquaculture). 5. FAO, editor. Sustainability in action. Rome; 2020. 206 p. (The state of world fisheries and aquaculture). 6. Shepherd CJ, Jackson AJ. Global fishmeal and fish-oil supply: inputs, outputs and markets a: global production of fishmeal and fish-oil. J Fish Biol. 2013 Oct;83(4):1046–66. 7. Cao L, Diana JS, Keoleian GA, Lai Q. Life Cycle Assessment of Chinese Shrimp Farming Systems Targeted for Export and Domestic Sales. Environ Sci Technol. 2011 Aug 1;45(15):6531–8. 8. Wilfart A, Prudhomme J, Blancheton JP, Aubin J. LCA and emergy accounting of aquaculture systems: Towards ecological intensification. Journal of Environmental Management. 2013 May;121:96–109. 9. Parker R. Implications of high animal by-product feed inputs in life cycle assessments of farmed Atlantic salmon. Int J Life Cycle Assess. 2018 May;23(5):982–94. 10. Malcorps W, Kok B, van‘t Land M, Fritz M, van Doren D, Servin K, et al. The Sustainability Conundrum of Fishmeal Substitution by Plant Ingredients in Shrimp Feeds. Sustainability. 2019 Feb 25;11(4):1212. 11. Maiolo S, Cristiano S, Gonella F, Pastres R. Ecological sustainability of aquafeed: An emergy assessment of novel or underexploited ingredients. Journal of Cleaner Production. 2021 Apr;294:126266. 26 12. Allegretti G, Talamini E, Schmidt V, Bogorni PC, Ortega E. Insect as feed: An emergy assessment of insect meal as a sustainable protein source for the Brazilian poultry industry. Journal of Cleaner Production. 2018 Jan;171:403–12. 13. Madau FA, Arru B, Furesi R, Pulina P. Insect Farming for Feed and Food Production from a Circular Business Model Perspective. Sustainability. 2020 Jul 4;12(13):5418. 14. van Huis A. Potential of Insects as Food and Feed in Assuring Food Security. Annual Review of Entomology. 2013 Jan 7;58(1):563–83. 15. Poveda J. Insect frass in the development of sustainable agriculture. A review. Agron Sustain Dev. 2021 Feb;41(1):5. 16. Miglietta P, De Leo F, Ruberti M, Massari S. Mealworms for Food: A Water Footprint Perspective. Water. 2015 Nov 6;7(11):6190–203. 17. Oonincx DGAB, van Broekhoven S, van Huis A, van Loon JJA. Feed Conversion, Survival and Development, and Composition of Four Insect Species on Diets Composed of Food By-Products. Papadopoulos NT, editor. PLoS ONE. 2015 Dec 23;10(12):e0144601. 18. van Huis A. Edible insects: future prospects for food and feed security. Rome: Food and Agriculture Organization of the United Nations; 2013. 187 p. (FAO forestry paper). 19. Pretty J, Bharucha ZP. Sustainable intensification in agricultural systems. Annals of Botany. 2014 Dec 1;114(8):1571–96. 20. Murphy DJ. Manure Storage Hazards. Pensylvania State University; 1991. (Fact Sheet Safety). Report No.: 28. 21. Qian X, Gu J, Sun W, Wang XJ, Su JQ, Stedfeld R. Diversity, abundance, and persistence of antibiotic resistance genes in various types of animal manure following industrial composting. Journal of Hazardous Materials. 2018 Feb;344:716–22. 22. Cammack JA, Miranda CD, Jordan HR, Tomberlin JK. Upcycling of manure with insects: current and future prospects. Journal of Insects as Food and Feed. 2021 Aug 13;7(5):605–19. 23. Hussein M, Pillai VV, Goddard JM, Park HG, Kothapalli KS, Ross DA, et al. Sustainable production of housefly (Musca domestica) larvae as a protein-rich feed ingredient by utilizing cattle manure. PLOS ONE. 2017 Feb 7;12(2):e0171708. 27 24. Turchini GM, Trushenski JT, Glencross BD. Thoughts for the Future of Aquaculture Nutrition: Realigning Perspectives to Reflect Contemporary Issues Related to Judicious Use of Marine Resources in Aquafeeds. North Am J Aquaculture. 2019 Jan;81(1):13–39. 25. Lindsay GJH, Walton MJ, Adron JW, Fletcher TC, Cho CY, Cowey CB. The growth of rainbow trout (Salmo gairdneri) given diets containing chitin and its relationship to chitinolytic enzymes and chitin digestibility. Aquaculture. 1984 Apr 1;37(4):315–34. 26. Krogdahl Å, Hemre GI, Mommsen T p. Carbohydrates in fish nutrition: digestion and absorption in postlarval stages. Aquaculture Nutrition. 2005;11(2):103–22. 27. Lock ER, Arsiwalla T, Waagbø R. Insect larvae meal as an alternative source of nutrients in the diet of Atlantic salmon ( Salmo salar ) postsmolt. Aquacult Nutr. 2016 Dec;22(6):1202–13. 28. Drew MD, Borgeson TL, Thiessen DL. A review of processing of feed ingredients to enhance diet digestibility in finfish. Animal Feed Science and Technology. 2007 Oct;138(2):118–36. 29. Francis G, Makkar HPS, Becker K. Antinutritional factors present in plant- derived alternate fish feed ingredients and their effects in fish. Aquaculture. 2001 Aug;199(3–4):197–227. 30. Henry M, Gasco L, Piccolo G, Fountoulaki E. Review on the use of insects in the diet of farmed fish: Past and future. Animal Feed Science and Technology. 2015 May;203:1–22. 31. Gasco L, Biasato I, Dabbou S, Schiavone A, Gai F. Animals Fed Insect-Based Diets: State-of-the-Art on Digestibility, Performance and Product Quality. Animals. 2019 Apr 16;9(4):170. 32. Gasco L, Biancarosa I, Liland NS. From waste to feed: A review of recent knowledge on insects as producers of protein and fat for animal feeds. Current Opinion in Green and Sustainable Chemistry. 2020 Jun;23:67–79. 33. Liu X, Chen X, Wang H, Yang Q, ur Rehman K, Li W, et al. Dynamic changes of nutrient composition throughout the entire life cycle of black soldier fly. Falabella P, editor. PLoS ONE. 2017 Aug 10;12(8):e0182601. 34. Meneguz M, Schiavone A, Gai F, Dama A, Lussiana C, Renna M, et al. Effect of rearing substrate on growth performance, waste reduction efficiency and chemical composition of black soldier fly ( Hermetia illucens ) larvae: Rearing substrate effects on performance and nutritional composition of black soldier fly. J Sci Food Agric. 2018 Dec;98(15):5776–84. 28 35. Belghit I, Liland NS, Gjesdal P, Biancarosa I, Menchetti E, Li Y, et al. Black soldier fly larvae meal can replace fish meal in diets of sea-water phase Atlantic salmon (Salmo salar). Aquaculture. 2019 Mar;503:609–19. 36. Fisher HJ, Collins SA, Hanson C, Mason B, Colombo SM, Anderson DM. Black soldier fly larvae meal as a protein source in low fish meal diets for Atlantic salmon (Salmo salar). Aquaculture. 2020 May;521:734978. 37. Magalhães R, Sánchez-López A, Leal RS, Martínez-Llorens S, Oliva-Teles A, Peres H. Black soldier fly ( Hermetia illucens ) pre-pupae meal as a fish meal replacement in diets for European seabass ( Dicentrarchus labrax ). Aquaculture. 2017 Jul;476:79–85. 38. Belghit I, Liland NS, Waagbø R, Biancarosa I, Pelusio N, Li Y, et al. Potential of insect-based diets for Atlantic salmon ( Salmo salar ). Aquaculture. 2018 Apr;491:72–81. 39. Cummins VC, Rawles SD, Thompson KR, Velasquez A, Kobayashi Y, Hager J, et al. Evaluation of black soldier fly (Hermetia illucens) larvae meal as partial or total replacement of marine fish meal in practical diets for Pacific white shrimp (Litopenaeus vannamei). Aquaculture. 2017 Apr;473:337–44. 40. Cardinaletti G, Randazzo B, Messina M, Zarantoniello M, Giorgini E, Zimbelli A, et al. Effects of Graded Dietary Inclusion Level of Full-Fat Hermetia illucens Prepupae Meal in Practical Diets for Rainbow Trout (Oncorhynchus mykiss). Animals. 2019 May 17;9(5):251. 41. Hong J, Han T, Kim YY. Mealworm (Tenebrio molitor Larvae) as an Alternative Protein Source for Monogastric Animal: A Review. Animals. 2020 Nov 8;10(11):2068. 42. Fontes TV, de Oliveira KRB de, Gomes Almeida IL, Maria Orlando TM, Rodrigues PB, Costa DV da, et al. Digestibility of Insect Meals for Nile Tilapia Fingerlings. Animals. 2019 Apr 20;9(4):181. 43. Basto A, Matos E, Valente LMP. Nutritional value of different insect larvae meals as protein sources for European sea bass (Dicentrarchus labrax) juveniles. Aquaculture. 2020 May;521:735085. 44. Rema P, Saravanan S, Armenjon B, Motte C, Dias J. Graded Incorporation of Defatted Yellow Mealworm (Tenebrio molitor) in Rainbow Trout (Oncorhynchus mykiss) Diet Improves Growth Performance and Nutrient Retention. Animals. 2019 Apr 23;9(4):187. 45. Chemello G, Renna M, Caimi C, Guerreiro I, Oliva-Teles A, Enes P, et al. Partially Defatted Tenebrio molitor Larva Meal in Diets for Grow-Out Rainbow 29 Trout, Oncorhynchus mykiss (Walbaum): Effects on Growth Performance, Diet Digestibility and Metabolic Responses. Animals. 2020 Jan 31;10(2):229. 46. Panini RL, Freitas LEL, Guimarães AM, Rios C, da Silva MFO, Vieira FN, et al. Potential use of mealworms as an alternative protein source for Pacific white shrimp: Digestibility and performance. Aquaculture. 2017 Apr;473:115–20. 47. Hoffmann L, Rawski M, Nogales-Merida S, Mazurkiewicz J. Dietary Inclusion of Tenebrio Molitor Meal in Sea Trout Larvae Rearing: Effects on Fish Growth Performance, Survival, Condition, and GIT and Liver Enzymatic Activity. Annals of Animal Science. 2020 Apr 1;20(2):579–98. 48. Khamesipour F, Lankarani KB, Honarvar B, Kwenti TE. A systematic review of human pathogens carried by the housefly (Musca domestica L.). BMC Public Health. 2018 Dec;18(1):1049. 49. Čičková H, Newton GL, Lacy RC, Kozánek M. The use of fly larvae for organic waste treatment. Waste Management. 2015 Jan;35:68–80. 50. Barnard DR, Harms RH, Sloan DR. Biodegradation of Poultry Manure by House Fly (Diptera: Muscidae). Environmental Entomology. 1998 Jun 1;27(3):600–5. 51. Miranda C d., Cammack J a., Tomberlin J k. Life-history traits of house fly, Musca domestica L. (Diptera: Muscidae), reared on three manure types. Journal of Insects as Food and Feed. 2020 Feb 6;6(1):81–90. 52. Miranda CD, Cammack JA, Tomberlin JK. Life-History Traits of the Black Soldier Fly, Hermetia illucens (L.) (Diptera: Stratiomyidae), Reared on Three Manure Types. Animals. 2019 May 25;9(5):281. 53. Hashizume A, Ido A, Ohta T, Thiaw ST, Morita R, Nishikawa M, et al. Housefly (Musca domestica) Larvae Preparations after Removing the Hydrophobic Fraction Are Effective Alternatives to Fish Meal in Aquaculture Feed for Red Seabream (Pagrus major). Fishes. 2019 Jun 27;4(3):38. 54. Aarestrup F. Get pigs off antibiotics. Nature. 2012 Jun;486(7404):465–6. 55. Van Boeckel TP, Brower C, Gilbert M, Grenfell BT, Levin SA, Robinson TP, et al. Global trends in antimicrobial use in food animals. PNAS. 2015 May 5;112(18):5649–54. 56. Schar D, Klein EY, Laxminarayan R, Gilbert M, Van Boeckel TP. Global trends in antimicrobial use in aquaculture. Sci Rep. 2020 Dec 14;10(1):21878. 57. Cabello FC. Heavy use of prophylactic antibiotics in aquaculture: a growing problem for human and animal health and for the environment. Environmental Microbiology. 2006;8(7):1137–44. 30 58. Lulijwa R, Rupia EJ, Alfaro AC. Antibiotic use in aquaculture, policies and regulation, health and environmental risks: a review of the top 15 major producers. Reviews in Aquaculture. 2020;12(2):640–63. 59. Millanao AB, Barrientos MH, Gómez CC, Tomova A, Buschmann A, Dölz H, et al. Injudicious and excessive use of antibiotics: public health and salmon aquaculture in Chile. Rev Med Chil. 2011 Jan 1;139(1):107–18. 60. Smith PG. Antimicrobial resistance in aquaculture. Rev Sci Tech OIE. 2008 Apr 1;27(1):243–64. 61. Baquero F, Martínez JL, Cantón R. Antibiotics and antibiotic resistance in water environments. Current Opinion in Biotechnology. 2008 Jun 1;19(3):260–5. 62. Cabello FC, Godfrey HP, Buschmann AH, Dölz HJ. Aquaculture as yet another environmental gateway to the development and globalisation of antimicrobial resistance. The Lancet Infectious Diseases. 2016 Jul 1;16(7):e127–33. 63. Zago V, Veschetti L, Patuzzo C, Malerba G, Lleo MM. Resistome, Mobilome and Virulome Analysis of Shewanella algae and Vibrio spp. Strains Isolated in Italian Aquaculture Centers. Microorganisms. 2020 Apr;8(4):572. 64. Algammal AM, Mabrok M, Ezzat M, Alfifi KJ, Esawy AM, Elmasry N, et al. Prevalence, antimicrobial resistance (AMR) pattern, virulence determinant and AMR genes of emerging multi-drug resistant Edwardsiella tarda in Nile tilapia and African catfish. Aquaculture. 2022 Feb 15;548:737643. 65. Preena PG, Swaminathan TR, Kumar VJR, Singh ISB. Antimicrobial resistance in aquaculture: a crisis for concern. Biologia. 2020 Sep;75(9):1497–517. 66. Schmidt AS, Bruun MS, Dalsgaard I, Larsen JL. Incidence, Distribution, and Spread of Tetracycline Resistance Determinants and Integron-Associated Antibiotic Resistance Genes among Motile Aeromonads from a Fish Farming Environment. Appl Environ Microbiol. 2001 Dec;67(12):5675–82. 67. Labitt RN, Ren J, Marquis H. Emergence of phenotypic and genotypic resistance in the intestinal microbiota of rainbow trout (Oncorhynchus mykiss) exposed long-term to sub-inhibitory concentrations of sulfamethoxazole. Ecotoxicology. 2021 Dec 1;30(10):2043–54. 68. Ibrahim M, Ahmad F, Yaqub B, Ramzan A, Imran A, Afzaal M, et al. Chapter 4 - Current trends of antimicrobials used in food animals and aquaculture. In: Hashmi MZ, editor. Antibiotics and Antimicrobial Resistance Genes in the Environment [Internet]. Elsevier; 2020 [cited 2022 Feb 22]. p. 39–69. (Advances in Environmental Pollution Research series; vol. 1). Available from: https://www.sciencedirect.com/science/article/pii/B9780128188828000048 31 69. Henriksson PJG, Rico A, Troell M, Klinger DH, Buschmann AH, Saksida S, et al. Unpacking factors influencing antimicrobial use in global aquaculture and their implication for management: a review from a systems perspective. Sustain Sci. 2018 Jul;13(4):1105–20. 70. Morrison DB, Saksida S. Trends in antimicrobial use in Marine Harvest Canada farmed salmon production in British Columbia (2003–2011). Can Vet J. 2013 Dec;54(12):1160–3. 71. Lebel L, Mungkung R, Gheewala SH, Lebel P. Innovation cycles, niches and sustainability in the shrimp aquaculture industry in Thailand. Environmental Science & Policy. 2010 Jun 1;13(4):291–302. 72. Kibenge MJT, Wang Y, Gayeski N, Morton A, Beardslee K, McMillan B, et al. Piscine orthoreovirus sequences in escaped farmed Atlantic salmon in Washington and British Columbia. Virol J. 2019 Dec;16(1):1–13. 73. McKinnell S, Thomson AJ. Recent events concerning Atlantic salmon escapees in the Pacific. ICES Journal of Marine Science. 1997 Dec 1;54(6):1221–5. 74. Waknitz FW, Iwamoto RN, Strom MS. Interactions of Atlantic salmon in the Pacific Northwest. Fisheries Research. 2003 Jun;62(3):307–28. 75. Subasinghe R, Soto D, Benzie J. Current Status and Options for Biotechnologies in Aquaculture and Fisheries in Developing Countries. [Internet]. United Nations Food and Agriculture Organization; 2011 [cited 2022 Jan 3]. Available from: https://repository.oceanbestpractices.org/handle/11329/1569 76. Midtlyng PJ, Grave K, Horsberg TE. What has been done to minimize the use of antibacterial and antiparasitic drugs in Norwegian aquaculture? Aquaculture Research. 2011;42(s1):28–34. 77. Barrett LT, Oppedal F, Robinson N, Dempster T. Prevention not cure: a review of methods to avoid sea lice infestations in salmon aquaculture. Reviews in Aquaculture. 2020;12(4):2527–43. 78. Boyd CE, Jescovitch LN. Penaeid shrimp aquaculture. In: Fisheries and Aquaculture: Volume 9. Oxford University Press; 2020. p. 233. (The Natural History of the Crustacea). 79. Troyer RM, Kurath G. Molecular epidemiology of infectious hematopoietic necrosis virus reveals complex virus traffic and evolution within southern Idaho aquaculture. Diseases of Aquatic Organisms. 2003 Aug 4;55(3):175–85. 80. Alday-Sanz V, Brock J, Flegel TW, McIntosh R, Bondad-Reantaso MG, Salazar M, et al. Facts, truths and myths about SPF shrimp in Aquaculture. Reviews in Aquaculture. 2020;12(1):76–84. 32 81. Filipa-Silva A, Parreira R, Nunes M, Crespo MTB. Virological screening of wild and farmed fish species: Application of innovative Taqman multiplex real-time PCR protocols. Aquaculture. 2021 Jul 15;540:736693. 82. Lewin AS, Haugen T, Netzer R, Tøndervik A, Dahle SW, Hageskal G. Multiplex droplet digital PCR assay for detection of Flavobacterium psychrophilum and Yersinia ruckeri in Norwegian aquaculture. Journal of Microbiological Methods. 2020 Oct 1;177:106044. 83. Murray B, editor. An Impact Evaluation of the Development of Genetically Improved Farmed Tilapia and Their Dissemination in Selected Countries [Internet]. Asian Development Bank; 2005 [cited 2022 Jan 2]. Available from: https://www.adb.org/publications/impact-evaluation-development-genetically- improved-farmed-tilapia-and-their 84. Gjedrem T, Robinson N, Rye M. The importance of selective breeding in aquaculture to meet future demands for animal protein: A review. Aquaculture. 2012 Jun 20;350–353:117–29. 85. Houston RD. Future directions in breeding for disease resistance in aquaculture species. R Bras Zootec. 2017 Jun;46:545–51. 86. Kaelberer MM, Bohórquez DV. The now and then of gut-brain signaling. Brain Research. 2018 Aug;1693:192–6. 87. Encarnação P. 5 - Functional feed additives in aquaculture feeds. In: Nates SF, editor. Aquafeed Formulation [Internet]. San Diego: Academic Press; 2016 [cited 2022 Apr 26]. p. 217–37. Available from: https://www.sciencedirect.com/science/article/pii/B9780128008737000051 88. Hayatgheib N, Moreau E, Calvez S, Lepelletier D, Pouliquen H. A review of functional feeds and the control of Aeromonas infections in freshwater fish. Aquacult Int. 2020 Jun 1;28(3):1083–123. 89. Martin SAM, Król E. Nutrigenomics and immune function in fish: new insights from omics technologies. Developmental & Comparative Immunology. 2017 Oct 1;75(Supplement C):86–98. 90. Martinez-Rubio L, Morais S, Evensen Ø, Wadsworth S, Vecino JG, Ruohonen K, et al. Effect of functional feeds on fatty acid and eicosanoid metabolism in liver and head kidney of Atlantic salmon (Salmo salar L.) with experimentally induced heart and skeletal muscle inflammation. Fish Shellfish Immunol. 2013 Jun;34(6):1533–45. 91. Kumaran S, Perianaika Anahas AM, Prasannabalaji N, Karthiga M, Bharathi S, Rajasekar T, et al. Chitin derivatives of NAG and chitosan nanoparticles from 33 marine disposal yards and their use for economically feasible fish feed development. Chemosphere. 2021 Oct 1;281:130746. 92. Ali M f. z., Nakahara S, Otsu Y, Ido A, Miura C, Miura T. Effects of functional polysaccharide from silkworm as an immunostimulant on transcriptional profiling and disease resistance in fish. Journal of Insects as Food and Feed. 2021 Oct 5;1–14. 93. Ullah S, Zhang J, Xu B, Tegomo AF, Sagada G, Zheng L, et al. Effect of dietary supplementation of lauric acid on growth performance, antioxidative capacity, intestinal development and gut microbiota on black sea bream (Acanthopagrus schlegelii). PLOS ONE. 2022 Jan 13;17(1):e0262427. 94. Dai J, Zheng J, Ou W, Xu W, Ai Q, Zhang W, et al. The effect of dietary cecropin AD on intestinal health, immune response and disease resistance of juvenile turbot (Scophthalmus maximus L.). Fish & Shellfish Immunology. 2020 May 1;100:117–25. 95. Ido A, Iwai T, Ito K, Ohta T, Mizushige T, Kishida T, et al. Dietary effects of housefly (Musca domestica) (Diptera: Muscidae) pupae on the growth performance and the resistance against bacterial pathogen in red sea bream (Pagrus major) (Perciformes: Sparidae). Appl Entomol Zool. 2015 May;50(2):213–21. 96. Su J, Gong Y, Cao S, Lu F, Han D, Liu H, et al. Effects of dietary Tenebrio molitor meal on the growth performance, immune response and disease resistance of yellow catfish (Pelteobagrus fulvidraco). Fish & Shellfish Immunology. 2017 Oct 1;69:59–66. 97. Xiang J, Qin L, Zhao D, Xiong F, Wang G, Zou H, et al. Growth performance, immunity and intestinal microbiota of swamp eel (Monopterus albus) fed a diet supplemented with house fly larvae (Musca domestica). Aquaculture Nutrition. 2020;26(3):693–704. 34 CHAPTER 2 Tissue specific differences in detection of Yersinia ruckeri carrier status in rainbow trout (Oncorhynchus mykiss) Sibinga, N. A. & Marquis, H. Tissue‐specific differences in detection of Yersinia ruckeri carrier status in rainbow trout (Oncorhynchus mykiss). Journal of Fish Diseases. (2021) doi:10.1111/jfd.13515. 35 ABSTRACT Effective monitoring for subclinical infections is a cornerstone of proactive disease management in aquaculture. Salmonid fish that survive Enteric Redmouth Disease (ERM) can carry Yersinia ruckeri as a latent infection for several months, potentially facilitating cryptic spread between facilities that exchange fish. In this study, fingerling Rainbow trout (Oncorhynchus mykiss) were infected by immersion and sampled for up to 14 weeks post-infection. Y. ruckeri was cultured from the posterior kidney of more than 89% of fish up to 4 weeks post-infection, but from 2% or fewer of fish sampled at later time points. In contrast, qPCR-based detection of the Y. ruckeri 16s rRNA gene in intestine and spleen extracts was much more sensitive: at 14 weeks post-infection Y. ruckeri was detected in nearly 50% of spleens and 15% of intestines. The discrepancy between spleen and intestine is likely due at least in part to technical limitations of qPCR on intestinal DNA extracts; accordingly, we propose that qPCR of spleen DNA ought to be considered the preferred standard for detection of carriers of Y. ruckeri. 36 INTRODUCTION Yersinia ruckeri is a gram-negative bacterium responsible for enteric redmouth disease (ERM), also known as yersinosis, a widespread fish disease (1,2). In the United States, Y. ruckeri was first isolated from rainbow trout (Oncorhynchus mykiss) in the 1950s. The infection is characterized by petechial haemorrhages in and around the mouth, fins, and belly, as well as in the liver, intestine, swim bladder, and adipose tissue. The infection is most acute and lethal in fish up to fingerling size, whereas older fish typically suffer from a chronic infection (3). This disease is historically most prevalent in farmed rainbow trout but has a broad host range including at least 20 species of fish (4). Outbreaks of ERM continue to occur around the world and are commonly linked to poor water quality and husbandry practices (5). The association of environmental stress with disease outbreaks suggests that Y. ruckeri is widespread and that clinical infections are at least partially opportunistic. In recent years, Y. ruckeri has caused lethal outbreaks in Atlantic salmon (Salmo salar) (6), Nile tilapia (Oreochromis niloticus) (7), and Channel catfish (Ictalurus punctatus) (8,9), suggesting that the threat posed by this disease to aquaculture species is increasing. Signs that Y. ruckeri has evolved and extended its host range (7,10–12) have led to renewed interest in this bacterial species. Phylogenetic and biochemical approaches have revealed strain differences based on geography (1,2) and host species (6) . Virulence-associated loss of motility (biotype 2) appears to have arisen independently in multiple populations and might have been driven by the use of motile biotype 1 strains for vaccination (2). A genomic comparison of five strains found significant differences between serotype O1 isolates 37 from rainbow trout (150, ATCC29473 and CSF007-82), a serotype O2 isolate from Chinook salmon (Oncorhynchus tshawytscha) (Big Creek 74), and an untyped isolate from catfish (SC09) (13). Seventy-five percent of the genome was common between these five strains, but most strains had a unique subset of putative virulence genes. A gene from SC09 was specifically associated with virulence in rainbow trout and intracellular survival in their macrophages (14). In fact, the ability of Y. ruckeri to survive intracellularly has been previously reported (15,16) and might be an important feature of the disease, especially after the initial acute infection is cleared. There is evidence for strain persistence within individual hatcheries and transmission between hatcheries that exchange fish (17), and the existence of asymptomatic carrier fish has been known for many years. Busch and Lingg recovered Y. ruckeri via bacterial culture from the intestine of 25-75% of fish for more than two months after an immersion challenge, while bacterial cultures from kidney, liver, or spleen of these fish were rarely positive (18). This finding gave rise to a broadly accepted model wherein the intestine has been implicated as the primary tissue associated with carrier status (3,19,20). However, other studies suggest that Y. ruckeri persists in multiple tissues (21,22). The aim of this study was to define the kinetics of bacterial clearance in fish intestine and spleen following an immersion challenge with Y. ruckeri, using a sensitive molecular method for bacterial detection. The results show that, as previously reported, the bacterium persists in the intestine of a subset of fish for many weeks after infection. However, our findings suggest that the spleen is an important reservoir in carrier fish and that Y. ruckeri is more likely to be detected in the spleen than the intestine. Accordingly, we propose 38 that the spleen be selected as the tissue of preference to monitor carrier status of Y. ruckeri in fish populations. MATERIALS AND METHODS Fish care All animals were cared for in compliance with the Guide for the Care and Use of Laboratory Animals and American Association of Laboratory Animal Science Position Statements, and all procedures were approved by the Cornell University IACUC. Rainbow trout (Oncorhynchus mykiss) fry (2.5g average weight) were provided by the New York State fish hatchery at Bath, NY. Fish were acclimated in a 700 litre fiberglass tank under flow through conditions with supplemental aeration for a total of three weeks prior to the challenge experiment. After the first week of acclimation, feed was shifted from BioOregon BioVida crumble to BioOregon BioClarks Fry 1.2mm pellets. After the second week of acclimation, water temperature was gradually increased from 9±1°C to 15±1°C over a period of six days. Yersinia ruckeri culture Y. ruckeri strain CSF007-82 (23) was provided by Timothy J. Welch of USDA and stored at -80°C in 50% glycerol. For the challenge, 10mL of tryptone soy broth (TSB) were inoculated from a single colony and grown overnight at 24°C with shaking (200rpm). This stationary phase culture was passaged 1:2000 into 2L flasks each containing 750mL TSB and grown for 22 hours at 24°C with shaking (200rpm). 39 The resulting stationary phase cultures contained ≈ 2x1010 CFU/ml, as measured by plate counts. Infection challenge One day prior to infection with Y. ruckeri, fish were moved from the 700L tank to a zebrafish rack system that had been converted to operate in a flow-through mode at 15±1°C with supplemental aeration in each tank. Fish comprising the experimental group were distributed among twelve 10L tanks (34-35 fish per tank). The control non-infected group (225 fish) was kept in one 700L tank. On the day of the infection, water flow into the zebrafish rack was stopped and water level in each tank was adjusted to 4.2L. Each tank was inoculated with 200mL of stationary phase Y. ruckeri to achieve the target density of ≈ 9x108 CFU/ml. Water samples were collected from each tank and enumerated by plate counts to determine the actual dose. Fish were immersed for one hour before flow was restored to rapidly flush out bacteria. The next day, infected fish were transferred to three 700L tanks maintained at 15±1°C with a continuous water flow. Fish from four 10L infection tanks were randomly assigned to each tank. Fish were observed twice daily for the first month of the experiment and once daily thereafter; all mortalities and morbidities were removed from the tanks as soon as observed. Effluent waters from fish tanks were collected at all times in two 5000-gallon tanks equipped with a chlorine injection system for decontamination prior to being released into the sewer system. Sample collection 40 A subsample of fish was euthanized with an overdose of buffered MS-222, weighed and dissected at each time point to capture the course of infection and recovery. Fish were sampled 6 days, 4 weeks, 9 weeks, and 14 weeks post-infection. At each of these timepoints whole intestine and spleen were collected into separate sterile 2ml screw cap tubes and immediately frozen in liquid nitrogen for later analysis. The posterior kidney was stabbed with a sterile wire loop to inoculate a tryptic soy agar (TSA) plate for bacteriological analysis. Throughout the experiment, any observed moribund fish were euthanized with an overdose of buffered MS-222 and the presence of Y. ruckeri confirmed by kidney stab and subsequent PCR analysis of recovered colonies. DNA extraction Tissue DNA was extracted using an adapted version of the method described by Small et. al (24). Frozen tissues were homogenized in 500uL DNA extraction buffer (200mM NaCl, 200mM Tris-HCl pH 7.5, 20mM EDTA, 10% SDS) in sterile tubes with 3-5 1.3mm chrome steel beads. Tubes were secured in a mini beadbeater (BioSpec) and subjected to 4 cycles of 30 seconds with 2 minutes on ice in between each cycle. Each homogenate was transferred to a new tube containing 100uL 0.1 mm zirconia/silica beads and DNA extraction buffer was added to bring total volume to 650 uL. For the intestine, if the initial sample weight was greater than 20mg, a subfraction of the homogenate corresponding to 20mg of intestine was used for this step; for spleens, the entire homogenate was transferred. Samples were further homogenized with 3 beating cycles of 30 seconds with 2 minutes on ice in between 41 each cycle. Particulates were pelleted at 10,000xg for 2 minutes at 4°C. 600ul of homogenate were mixed by inversion with 600ul of a 25:24:1 (v/v) mixture of phenol:chloroform:isoamyl alcohol in a phase lock gel (PLG) tube, and centrifuged at 18,000xg for 5 minutes. The aqueous layer was transferred to a clean PLG tube, mixed with an additional 500ul of 24:1 (v/v) chloroform:isoamyl alcohol, and centrifuged at 18,000xg for 5 minutes. The aqueous layer was transferred to a new tube and the DNA was precipitated by adding 750uL of cold isopropanol and incubating at -20°C. The DNA was pelleted by centrifugation at 5,800xg, washed one time with 70% ethanol, washed two times with 95% ethanol, dried at 55°C and dissolved in 100uL TE buffer (10 mM Tris-HCl pH 8.0, 1 mM EDTA). DNA concentration was quantified with a Qubit fluorometer (ThermoFisher Scientific). Quantitative polymerase chain reaction (qPCR) detection assay DNA extracted from fish tissues was analysed for presence of Y. ruckeri using a hydrolysis (TaqMan) probe-based method targeting a region of the 16S rRNA gene (Ghosh et al., 2018). Forward and reverse primer sequences were 5’AACCCAGATGGGATTAGCTAGTAA3’ and 5’GTTCAGTGCTATTAACACTTAACCC3’. TaqMan probe sequence was 5’-6- Fam-AGCCACACTGGAACTGAGACACGGTCC-IBFQ-3’ (Integrated DNA Technologies). All qPCR experiments were conducted in 96-well plates on a 7500 real-time PCR instrument (Applied Biosystems). Each 10ul reaction contained 5ul 2x PrimeTime Gene Expression Master Mix (Integrated DNA Technologies), forward and reverse primers (400nM each), TaqMan probe (100nM), and 1ul DNA template. 42 To model the challenge of detecting small amounts of Y. ruckeri 16s sequence in a sample containing high concentrations of non-target DNA, positive and negative controls were generated by spiking tissues from naïve fish with Y. ruckeri or PBS. Bacterial concentration of a log-phase culture was initially estimated by spectrophotometry at OD600. Bacterial cells were washed with sterile PBS, serially diluted, and estimated numbers of cells were added directly to frozen tissue samples from uninfected fish. Specific CFU were determined by plating dilutions of bacteria in PBS on tryptone soy agar (TSA). DNA was extracted from spiked tissues as described above, aliquoted in small volumes, and stored at -80°C. Positive and negative control samples were included in all subsequent qPCR plates. Each experimental sample was tested in triplicate wells twice in two separate reaction plates; samples that tested positive in only one plate were retested with six additional replicates in a third plate. Samples with positive reactions in two independent experiments are reported as positive. All others are reported as negative. Estimation of bacterial load A standard curve was generated to determine the LOD of Y. ruckeri 16s rRNA gene by qPCR, using DNA extracted from a mid-log phase bacterial culture. Bacteria were washed twice and resuspended in sterile PBS, a dilution was used for CFU counts on TSA, and DNA was extracted using the protocol described above. Purified Y. ruckeri DNA was serially diluted and quantified using the Qubit dsDNA high- sensitivity kit. The number of bacterial chromosomes per dilution was calculated as follows: 43 𝐷𝑁𝐴 (𝑛𝑔) × 6.02 × 10!" 𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑐ℎ𝑟𝑜𝑚𝑜𝑠𝑜𝑚𝑒𝑠 = 𝑔𝑒𝑛𝑜𝑚𝑒 𝑠𝑖𝑧𝑒 (𝑏𝑝) × 660 × 1 × 10# Where the genome size is 3.83 × 10$ bp (Nelson et al., 2015), 660 %&'() is the (*+, average mass of 1bp of DNA, and 6.02 × 10!" and 1 × 10# are unit conversion factors. In silico analysis of the Y. ruckeri CSF007-82 genome confirmed 7 copies of the target 16s sequence in each genome (23). Serial dilutions were tested by qPCR in six replicate wells in each of three independent reaction plates, for a total of 18 replicates. A standard curve was generated from these results (Fig. S1) and LOD was calculated according to the method described by Klymus et al. (25). Using a 95% detection cutoff, the modelled LOD cut-off for 6 replicate wells was 0.77 gene copies; that is, detection in at least one of six replicate wells is expected in 95% of experiments for an initial concentration of 0.77 gene copies per well. Statistical analysis of data Survival data during the initial acute infection was analysed using a Kaplan- Meier curve; sampled fish and fish surviving until the end of the experiment were censored at the appropriate time points. A subsequent generalized Wilcoxon test was used to verify statistical significance of differences between groups. Survival analysis was performed using JMP Pro 14 software. Detection data were analysed individually within each tissue type using a Fisher’s exact test for overall significant difference between time points. This was followed by post-hoc pairwise Fisher’s exact test comparisons of all timepoints using a Bonferroni correction for multiple comparisons 44 using RStudio 1.4.1103 software. Detection data were not compared statistically between tissue types due to differences in assay sensitivity. RESULTS Susceptibility of trout to an immersion challenge with Y. ruckeri Severe and lethal infections were first observed 3 days after the immersion challenge with Y. ruckeri and continued until day 13 (Fig. 1). Dark coloration was the most frequently observed sign of infection, though examples of exophthalmia, petechial haemorrhage, splenomegaly, and fin erosion were also noted. Morbid fish were euthanized and tested for the presence of Y. ruckeri in the posterior kidney by inoculating a TSA plate. Positive Y. ruckeri cultures were obtained from all tested morbid fish and confirmed by PCR of 16s rRNA gene (data not shown). After day 13, no mortalities were observed in any of the infected tanks for the remainder of the 14- week experiment. Two mortalities were recorded in the uninfected control tank: one at day 6 and one at day 78. In both of these cases, gross pathology and kidney stabs did not reveal the presence of Y. ruckeri. The final survival rate in challenged fish was 52% on average (57%, 44%, and 56%); survival rate in control fish was 99%. 45 100 Uninfected fish 75 50 3 tanks of 25 infected fish 0 0 3 6 9 12 15 18 Days Post-Infection Figure 2.1. Kaplan-Meier survival curve during acute Y. ruckeri infection. Fish were infected in 10L aquaria via immersion in a suspension of Y. ruckeri before being assigned to one of three tanks (Infected 1-3). A control tank of uninfected fish was maintained in a fourth identical tank. All Y. ruckeri-associated mortality occurred within 13 days post-infection, after which surviving populations were stable. Detection of Y. ruckeri in fish tissues Persistence of Y. ruckeri in fish following the immersion challenge was measured by bacterial culture for detection of Y. ruckeri in the posterior kidney, and by qPCR assay for detection of Y. ruckeri in the intestine and spleen. Sensitivity of the qPCR assay was assessed by spiking intestines from uninfected fish with 10-fold serial dilutions of Y. ruckeri which corresponded to 85, 700, or 6400 CFU per 100 µl of extracted DNA. qPCR analysis revealed a detection rate of 92% in wells containing the equivalent of ~0.85 CFU per well (n=66 wells) 46 Percent survival (Fig. 2) and of 100% of reactions from intestines spiked with 10 and 100 times more Y. ruckeri CFU. False positive reactions in negative controls occurred in 2.9% of wells (n=69 wells across 23 reaction plates). Samples with positive reactions in two independent assays were reported as positive. All others were reported as negative. DNA was extracted from spiked spleens and resuspended in 100 µl of buffer: CFU were 1.1, 11, 111, 1100, or 4400 per qPCR reaction well. qPCR analysis revealed a detection rate of 100% in spleens spiked with bacteria and of 0% in control uninfected tissue samples (Fig. 2). Intestine Spleen 40 40 n=69 35 35 30 n=66 30 25 n=64 25 n=50 n=18 each 20 20 0 0.85 7 64 0 1.1 11 111 1110 4440 CFU per well CFU per well Figure 2.2. qPCR threshold count (Ct) for intestine and spleen controls. Control samples were prepared by spiking known amounts of mid-log phase Y. ruckeri CFU into whole frozen intestines or spleens collected from uninfected fish. DNA was extracted, aliquoted, and stored at -80°C. qPCR was performed using primers specific for the Y. ruckeri 16s rRNA gene. Triangles represent wells from multiple experiments. n indicates the number of wells and the solid blue lines represent the median for each data set. The dashed line represents the maximum number of cycles performed; points above this line were undetected after 40 cycles. 47 Ct Ct Persistence of Y. ruckeri post-infection At 6 days post-infection, 100% of fish were positive for Y. ruckeri by qPCR in the intestine and the spleen (n=15: 5 per tank) (Table 2.1 and Fig. 2.3) and 93% of fish were positive for bacterial growth from the posterior kidney (n=45: 15 per tank). At 4 weeks post-infection, 93% of fish were positive for Y. ruckeri by qPCR in intestine and spleen (n=15: 5 per tank) and 89% were positive for bacterial growth from the posterior kidney (n=45: 15 per tank). By 9 weeks, the proportion of fish positive by qPCR in both intestine and spleen was 44%, with an additional 42% positive in the spleen only, and 14 % negative in both tissues (n=45: 15 per tank); Y. ruckeri was not isolated from any of the kidney stabs. At 14 weeks, 7% of fish were positive by qPCR in intestine and spleen and an additional 40% were positive in the spleen only, whereas 5% were positive in the intestine only, and 48% were negative in both tissues (n=43; 15, 15 and 13 per tank); Y. ruckeri was isolated from the kidney of one fish (2.2%). No fish from the control group tested positive for Y. ruckeri in any tissue at any of the time points (n=5 per time point). 48 Table 2.1. Prevalence of Y. ruckeri detection in tissue samples of fish collected at various time points post-infection Time post-infection† Tissue analysed 6d 4w 9w 14w Kidney culture‡ 93.3% 88.9% 0% 2.2% (n=45§)A ¶ (n=45)A (n=45)B (n=45)B Spleen qPCR 100% 93.3% 86.7% 46.5% (n=15)A (n=15)A (n=45)A (n=43)B Intestine qPCR 100% 93.3% 44.4% 13.3% (n=15)A (n=15)A (n=45)B (n=45)C † days (d) or weeks (w) post-infection ‡ Y. ruckeri cultured on TSA from a kidney stab; a subset of collected samples were tested by qPCR § n=number of fish sampled ¶Each tissue type was analysed independently using Fisher’s exact test followed by post-hoc pairwise comparisons with Bonferroni correction; comparisons are strictly over time for a single tissue. Differing upper-case letters within the same line indicate statistically significant differences in detection rate (P<0.05). 49 n=15 n=15 n=45 n=45 Figure 2.3. Distribution of tissues positive for Y. ruckeri over time. Individual fish were sampled at each time point and analysed for presence or absence of Y. ruckeri in spleen and intestine by qPCR. n=number of fish analysed per time point. Quantification of Y. ruckeri The number of bacteria per positive sample was estimated using the qPCR standard curve generated from purified Y. ruckeri DNA (Fig. S2.1). The number of bacteria per intestine and spleen was 1.63 × 10- ± 6.30 × 10. CFU at 6 days post- infection and dropped thousand-fold between 6 days and 4 weeks post-infection (Fig. 2.4). In the intestine, the estimated number of bacteria remained relatively stable at approximately 100 bacteria per positive intestine between 4- and 14-weeks post- infection. In the spleen, the estimated number of bacteria dropped another 10-fold between 4- and 9-weeks post-infection to approximately 10 bacteria per positive spleen and that number was maintained at 14 weeks post-infection. When analysing these data, it is important to consider that the detection assay was less sensitive for intestinal samples than spleen samples because of the subsampling step in DNA 50 extraction. Extraction of DNA was fixed at 20mg of intestinal tissue: at 6 days post- infection, 20mg represented 35.7% of the median total intestinal weight; by 14 weeks post-infection, 20mg represented only 4.9%. The relative abundance of CFU from kidney stabs was compared at 6 days and 4 weeks post-infection. At 6 days post-infection, 80% of the fish had too many CFU of Y. ruckeri to count. In contrast, at 4 weeks, 73% of the fish had between one and 15 CFU. A single kidney stab was positive at 14 weeks with less than 15 CFU. Semilog line -- X is log, Y is linear 40 Best-fit valuesYintercept 36 Slope -3.294 Std. Error 30 Yintercept 0.2188 Slope 0.07835 95% CI (profile likelihood) Yintercept 35.5 to 36.49 20 Slope -3.472 to -3.117 Goodness of Fit Degrees of Freedom 9 10 R square 0.9949Absolute Sum of Squares 2.762 Sy.x 0.5539 0 Number of points 10-1 100 101 102 103 104 105 106 # of X values 66 Y. ruckeri # Y values analyzed 11 chromosomes per well Figure S2.1. Standard curves for intestine and spleen samples. Y. ruckeri purified DNA was used to generate a standard curve to determine the LOD of the 16s rRNA gene by qPCR. The number of chromosomes per well as a function of threshold counts (Ct) was calculated considering that each Y. ruckeri chromosome contains seven copies of the 16s rRNA gene. 51 Ct 107 Intestine 107 Spleen 106 106 105 105 104 104 103 103 102 102 101 101 100 100 6d 4w 9w 14w 6d 4w 9w 14w Time Post-Infection Time Post-Infection Figure 2.4. Estimated Y. ruckeri CFU in intestine and spleen over time. The number of bacteria per positive sample was estimated using the qPCR standard curve generated from purified Y. ruckeri DNA (Fig. S2.1). n: 15, 14, 20, and 6 for the intestine, and 15, 14, 39, and 20 for the spleen at each consecutive time point. The top and bottom of each box represent the 25th and 75th percentile values, and the whiskers show the maximum and minimum observed values. DISCUSSION The bacterial pathogen Y. ruckeri has been a source of economic loss for the aquaculture industry for many decades despite the development of effective vaccines, in part due to the ability of this bacterial pathogen to persist in tissues of healthy fish. In this study, the carrier status of fish was assessed over a period of 14 weeks following an experimental infection by immersion in which approximately 50% of the fish survived. Using a sensitive molecular method for detection of Y. ruckeri 16s rRNA, we observed that carrier fish persist for at least 14 weeks post-infection. Moreover, by 14 weeks, carrier fish were three times as likely to be detected when using the spleen as compared to the intestine, and the limit of detection was estimated 52 Estimated number of bacteria from positive samples Estimated number of bacteria from positive samples to be approximately 10 bacteria per spleen. These results indicate that the spleen is a long-term reservoir of Y. ruckeri and is a more reliable tissue when monitoring for the presence of Y. ruckeri in populations of trout. Intestinal shedding is likely important in the transmission of Y. ruckeri from carriers to naïve fish, making intestinal content an attractive target for monitoring (19,26,27). However, while improved methods of DNA extraction from intestinal samples have made PCR-based detection assays viable, there remain practical limitations that can only be resolved by compromising sensitivity. Subsampling of intestinal samples improves DNA yield and quality and decreases the concentration of PCR inhibitors present in faeces (24,28), but also presents a conceptual bottleneck to detection of low copy numbers, especially in large fish. Despite use of subsampling, the efficiency of the qPCR was 80.2% for spiked intestinal samples in this study, suggesting continued PCR inhibition. For comparison, qPCR efficiency was 99.1% in spiked spleen samples. A previous study by Ghosh et al. (19) shows a similar pattern, with a fourfold decrease in sensitivity for spiked faecal samples relative to both spiked spleens and an independent bacterial standard. This difference supports the concept that the spleen is a more reliable tissue to monitor the existence of carriers in fish populations. In this study, we observed a steady decrease in abundance and frequency of detection of Y. ruckeri in the intestine over the three months following infection. In contrast, Busch reported that bacterial recovery from intestine is cyclical (29), whereas Ghosh et. al detected Y. ruckeri in 100% of pooled faecal samples several months after infection (19). These overall differences between results from different groups are 53 likely related to variations in bacterial strain, pathogen exposure, rearing conditions, and host genetics. In the present study, fish were maintained at low density in flowthrough tanks with very stable temperatures, minimizing stress and decreasing the likelihood of potential exposure and reinfection from bacteria released from faeces. Our results indicate that at 14 weeks post-infection Y. ruckeri was present in the spleen of nearly fifty percent of the surviving fish. A proteomic study of rainbow trout immune response to infection with Y. ruckeri indicated that the spleen remains active for at least 28 days post-infection, with the phagosome pathway being particularly upregulated (22). This observation is presumably indicative of macrophage activity, as these cells are abundant in the spleen. Although macrophages are important phagocytic cells that have the ability to kill bacteria, there is evidence that Y. ruckeri survives and multiplies intracellularly in these cells (15,16). Other pathogenic Yersinia species, such as Y. enterocolitica and Y. pseudotuberculosis, have been shown to survive in macrophages (30). Similarly to Y. ruckeri, these human pathogens invade through the intestine before spreading to other tissues. Persistence of intracellular bacteria inside of macrophages is a well-documented form of latent infection (31). However, strategies to eliminate these latent infections remain elusive. In conclusion, our findings reinforce the notion that carrier fish are a significant problem in the control of ERM and highlight the potential importance of the spleen as a reservoir of bacteria. Nearly fifty percent of surviving fish in this study still harboured Y. ruckeri in the spleen 14 weeks post-infection. This finding has significant implications for development of monitoring and treatment regimens to monitor and contain the spread of this disease. It is difficult with current tools to 54 determine whether the lower rate of detection of Y. ruckeri in intestinal tissue after 14 weeks in this study is an artifact of differences in detection sensitivity or is indicative of a true difference between the intestine and the spleen in their capacity to completely clear the infection. In either case, our results suggest that qPCR of spleen DNA extracts should be considered the preferred standard for determination of carrier status. 55 References 1. Altinok I, Capkin E, Boran H. Comparison of molecular and biochemical heterogeneity of Yersinia ruckeri strains isolated from Turkey and the USA. Aquaculture. 2016 Jan;450:80–8. 2. Bastardo A, Ravelo C, Romalde JL. Phylogeography of Yersinia ruckeri reveals effects of past evolutionary events on the current strain distribution and explains variations in the global transmission of enteric redmouth (ERM) disease. Front Microbiol [Internet]. 2015 Oct 29 [cited 2021 Jan 20];6. Available from: http://journal.frontiersin.org/Article/10.3389/fmicb.2015.01198/abstract 3. Tobback E, Decostere A, Hermans K, Haesebrouck F, Chiers K. Yersinia ruckeri infections in salmonid fish. J Fish Diseases. 2007 May;30(5):257–68. 4. Pajdak‐Czaus J, Platt‐Samoraj A, Szweda W, Siwicki AK, Terech‐Majewska E. Yersinia ruckeri — A threat not only to rainbow trout. Aquac Res. 2019 Nov;50(11):3083–96. 5. Tinsley JW, Lyndon AR, Austin B. Antigenic and cross-protection studies of biotype 1 and biotype 2 isolates of Yersinia ruckeri in rainbow trout, Oncorhynchus mykiss (Walbaum): Antigenic studies of Yersinia ruckeri. Journal of Applied Microbiology. 2011 Jul;111(1):8–16. 6. Ormsby MJ, Caws T, Burchmore R, Wallis T, Verner-Jeffreys DW, Davies RL. Yersinia ruckeri Isolates Recovered from Diseased Atlantic Salmon (Salmo salar) in Scotland Are More Diverse than Those from Rainbow Trout (Oncorhynchus mykiss) and Represent Distinct Subpopulations. Drake HL, editor. Appl Environ Microbiol. 2016 Oct 1;82(19):5785–94. 7. Abdel-Latif HMR, Khalil RH, Saad TT, El-Bably RY. Identification and molecular characterization of Yersinia ruckeri isolated from mass mortalities of cultured Nile tilapia at Kafr El-sheikh governorate. Global Journal of Fisheries and Aquaculture Researches. 2014;1(2):45–56. 8. Danley ML, Goodwin AE, Killian HS. Epizootics in farm-raised channel catfish, Ictalurus punctatus (Rafinesque), caused by the enteric redmouth bacterium Yersinia ruckeri. J Fish Diseases. 1999 Dec;22(6):451–6. 9. Wang K yu, Liu T, Wang J, Chen D fang, Wu X jing, Jiang J, et al. Complete Genome Sequence of the Fish Pathogen Yersinia ruckeri Strain SC09, Isolated from Diseased Ictalurus punctatus in China. Genome Announc. 2015 Feb 26;3(1):e01327-14, /ga/3/1/e01327-14.atom. 56 10. Bastardo A, Bohle H, Ravelo C, Toranzo A, Romalde J. Serological and molecular heterogeneity among Yersinia ruckeri strains isolated from farmed Atlantic salmon Salmo salar in Chile. Dis Aquat Org. 2011 Feb 22;93(3):207–14. 11. Gulla S. The Health Situation in Norwegian Aquaculture 2016. Norwegian Veterinary Institute; 2017 Jun p. 78–82. 12. Liu T, Wang KY, Wang J, Chen DF, Huang XL, Ouyang P, et al. Genome Sequence of the Fish Pathogen Yersinia ruckeri SC09 Provides Insights into Niche Adaptation and Pathogenic Mechanism. IJMS. 2016 Apr 14;17(4):557. 13. Cascales D, Guijarro JA, García-Torrico AI, Méndez J. Comparative genome analysis reveals important genetic differences among serotype O1 and serotype O2 strains of Y. ruckeri and provides insights into host adaptation and virulence. MicrobiologyOpen. 2017 Aug;6(4):e00460. 14. Liu T, Wang E, Wei W, Wang K, Yang Q, Ai X. TcpA, a novel Yersinia ruckeri TIR-containing virulent protein mediates immune evasion by targeting MyD88 adaptors. Fish & Shellfish Immunology. 2019 Nov;94:58–65. 15. Ryckaert J, Bossier P, D’Herde K, Diez-Fraile A, Sorgeloos P, Haesebrouck F, et al. Persistence of Yersinia ruckeri in trout macrophages. Fish & Shellfish Immunology. 2010 Oct;29(4):648–55. 16. Welch T, Wiens G. Construction of a virulent, green fluorescent protein-tagged Yersinia ruckeri and detection in trout tissues after intraperitoneal and immersion challenge. Dis Aquat Org. 2005;67:267–72. 17. Huang Y, Jung A, Schäfer W, Mock D, Brenner Michael G, Runge M, et al. Analysis of Yersinia ruckeri strains isolated from trout farms in northwest Germany. Dis Aquat Org. 2015 Oct 27;116(3):243–9. 18. Busch RA, Lingg AJ. Establishment of an Asymptomatic Carrier State Infection of Enteric Redmouth Disease in Rainbow Trout ( Salmo gairdneri ). J Fish Res Bd Can. 1975 Dec;32(12):2429–32. 19. Ghosh B, Crosbie PBB, Nowak BF, Bridle AR. A highly sensitive, non-invasive qPCR-based strategy for direct quantification of Yersinia ruckeri in fish faeces. J Fish Dis. 2018 Sep;41(9):1421–8. 20. Kumar G, Menanteau-Ledouble S, Saleh M, El-Matbouli M. Yersinia ruckeri, the causative agent of enteric redmouth disease in fish. Veterinary Research [Internet]. 2015 Dec [cited 2018 Apr 20];46(1). Available from: http://www.veterinaryresearch.org/content/46/1/103 21. Altinok I, Grizzle J, Liu Z. Detection of Yersinia ruckeri in rainbow trout blood by use of the polymerase chain reaction. Dis Aquat Org. 2001;44:29–34. 57 22. Kumar G, Hummel K, Noebauer K, Welch TJ, Razzazi-Fazeli E, El-Matbouli M. Proteome analysis reveals a role of rainbow trout lymphoid organs during Yersinia ruckeri infection process. Sci Rep. 2018 Dec;8(1):13998. 23. Nelson MC, LaPatra SE, Welch TJ, Graf J. Complete Genome Sequence of Yersinia ruckeri Strain CSF007-82, Etiologic Agent of Red Mouth Disease in Salmonid Fish. Genome Announc. 2015 Feb 26;3(1):e01491-14, /ga/3/1/e01491- 14.atom. 24. Small CM, Currey M, Beck EA, Bassham S, Cresko WA. Highly Reproducible 16S Sequencing Facilitates Measurement of Host Genetic Influences on the Stickleback Gut Microbiome. Chu H, editor. mSystems. 2019 Aug 13;4(4):e00331-19, /msystems/4/4/msys.00331-19.atom. 25. Klymus KE, Merkes CM, Allison MJ, Goldberg CS, Helbing CC, Hunter ME, et al. Reporting the limits of detection and quantification for environmental DNA assays. Environmental DNA. 2020 Jul;2(3):271–82. 26. Carson J, Wilson T. Yersiniosis in fish. Canberra: Sub‐committee on Animal Health and Laboratory Standards; 2002. 27. Rodgers CJ. Development of a selective-differential medium for the isolation of Yersinia ruckeri and its application in epidemiological studies. J Fish Diseases. 1992 May;15(3):243–54. 28. Acharya KR, Dhand NK, Whittington RJ, Plain KM. PCR Inhibition of a Quantitative PCR for Detection of Mycobacterium avium Subspecies Paratuberculosis DNA in Feces: Diagnostic Implications and Potential Solutions. Front Microbiol [Internet]. 2017 Feb 2 [cited 2021 May 11];08. Available from: http://journal.frontiersin.org/article/10.3389/fmicb.2017.00115/full 29. Busch RA. Enteric Red Mouth Disease (Hagerman Strain). Marine Fisheries Review. 1978;40(3):42–51. 30. Lemarignier M, Pizarro-Cerdá J. Autophagy and Intracellular Membrane Trafficking Subversion by Pathogenic Yersinia Species. Biomolecules. 2020 Dec;10(12):1637. 31. Bentrup KH zu, Russell DG. Mycobacterial persistence: adaptation to a changing environment. Trends in Microbiology. 2001 Dec 1;9(12):597–605. 58 CHAPTER 3 Longitudinal sampling of the rainbow trout (Oncorhynchus mykiss) microbiome reveals effects of dietary cecropin A and Yersinia ruckeri infection Sibinga, N.A., Lee, M.T., Johnson, E.L., Selvaraj, V., Marquis, H. Longitudinal sampling of the rainbow trout (Oncorhynchus mykiss) microbiome reveals effects of dietary cecropin A and Yersinia ruckeri infection. Frontiers in Marine Science. (2022) doi: 10.3389/fmars.2022.901389 59 60 ABSTRACT The aquaculture industry faces growing pressure to reduce the use of antibiotics for control of bacterial diseases. In this study we tested the effectiveness of dietary cecropin A, an insect-derived antimicrobial peptide, at preventing mortality and reducing incidence of carrier status in rainbow trout (Oncorhynchus mykiss) challenged by immersion with Yersinia ruckeri. Additionally, we conducted longitudinal analyses of microbiome changes to elucidate effects of both cecropin A and bacterial infection. An in vitro experiment indicated that Y. ruckeri is susceptible to cecropin A. However, dietary cecropin A did not improve the survival of fish challenged with Y. ruckeri, nor did it decrease the persistence of Y. ruckeri in the intestine of fish that survived infection. Moreover, levels of intestinal Y. ruckeri as measured by qPCR suggested that cecropin A may have negatively impacted the ability of fish to resist colonization by this bacterial pathogen. Concomitantly with the survival experiments, the microbiomes of challenged and mock-challenged fish were sampled at days 0, 3, 8, and 30. The microbiomes were in general dominated by Mycoplasma sp. at days 0, 3 and 8, independent of diet, and whether fish had been challenged or mock-challenged. At day 30, the microbiomes of mock-challenged fish fed the +cecropin diet were characterized by lower internal (alpha) diversity (p<.01), greater relative abundance of Mycoplasma sp., and a decrease in gram-negative taxa, when compared to the microbiomes of fish fed the control diet. The opposite was observed in the microbiome of challenged fish. Lastly, correlation analysis of amplicon sequence variants (ASVs) revealed a negative correlation between the presence of Y. ruckeri and seven ASVs, including Mycoplasma sp., suggesting 61 possible beneficial effects of these taxa. In addition, six ASVs were positively correlated to Y. ruckeri, including Flavobacterium succinicans – a known opportunistic fish pathogen. In conclusion, this study revealed that dietary cecropin A was bioactive and exerted significant effects on the microbiome but did not improve fish resistance to infection by Y. ruckeri. Based on our observations and other published results, it appears that high relative abundance of Mycoplasma sp. correlates with higher resistance to intestinal colonization by bacterial pathogens. 62 INTRODUCTION Historical use of antibiotics is not well-documented for aquaculture (1), but evidence from regional studies where data is available supports the claim that the rapid growth and intensification of aquaculture production in the past fifty years have been fueled in many cases by heavy antimicrobial use (2–5). As international momentum coalesces around tighter regulation of antibiotic use, it is essential that the aquaculture industry continue to develop alternative strategies for control of infectious disease (6). The primary focus of most research on fish disease has been has rightfully been on prevention rather than treatment of outbreaks. Development of vaccines, better nutrition, increasingly rapid and accurate diagnostic services, improvements in farm siting and biosecurity, and selective breeding for disease resistance have all been instrumental breakthroughs in maintaining fish health (7). In recent years, the role of the fish microbiome has also been increasingly investigated in aquaculture contexts (8–10). The microbiome of fish is influenced by diet (11), and this has been shown to mediate changes in the immune response to infection (12). Changes in fish microbiome are associated with infection states of various pathogens, including bacteria (13,14), viruses (15,16) and parasites (17,18). Furthermore, a number of studies have shown that modification of the microbiome via probiotic bacteria can improve immune parameters and disease resistance in fish (19–21). Despite these findings, however, current understanding of what constitutes a healthy or natural microbiome for most fish species is limited (9). In addition to preventive strategies, it is also worth considering whether it will be possible to one day replace antibiotics for disease intervention. Therapeutic 63 antibiotic treatments remain an important backstop for fish welfare in the event of disease outbreaks, even though resistance has led to diminishing utility in some regions (22). One proposed alternative is to replace or augment the use of conventional antibiotics with natural or synthetic antimicrobial peptides (AMPs), also referred to as host-defense peptides. AMPs are a part of the innate immune response in animals, fungi, and plants, and are also produced by bacteria. They are typically small (10-50 amino acid) cationic peptides with the ability to inhibit bacterial growth in vitro, though many are now known to have additional functions in immune signaling. Structurally, AMPs are characterized by a combination of positively charged and hydrophobic residues which facilitate their interaction with bacterial cell membranes (23). Bacterial resistance to AMPs can develop via selective pressure (24), but there are a number of reasons why AMPs are likely to be refractory to recapitulating the current AMR crisis (25). AMPs have been tested as dietary supplements in feeds for a variety of terrestrial animal production systems, including swine, poultry, and cattle, and a number of positive effects including mitigation of disease symptoms and modest increases in growth and/or feed efficiency have been reported (26). A growing body of research in recent years has also begun to investigate the effects of AMPs in fish, examining their effects on growth (27–30), immune status (31), and disease resistance (32–39). For the most part, these studies have not reported adverse effects of AMPs, and about half have shown protective effects in the context of infection. Of the studies listed above, only one looked at the effect of AMPs on the microbiome, finding that microbial diversity was increased relative to the control group in fish fed the highest 64 tested dose of AMP (30). Furthermore, in studies that have tested AMPs in the context of fish infection, their application has nearly always been prophylactic rather than therapeutic. There remains a knowledge gap in the understanding of how microbiome- based approaches to fish health management interact with other strategies, and in the effectiveness of AMPs administered close to or after the onset of infection. Conventional antibiotics have been shown in mammalian systems to negatively impact resistance to some diseases by disrupting the microbiome (40–42). While the body of literature is much smaller, similar effects have been shown in fish: treatment with streptomycin sulfate prior to challenge with Vibrio anguillarum increased mortality in black molly (Poecilia sphenops) (43), and pre-treatment with olaquindox increased mortality in zebrafish (Danio rerio) challenged with Aeromonas hydrophila (44). In both of these studies, the authors attributed the observed changes in disease resistance to disruption of the microbiome by antibiotic treatment. These findings, and evidence for the importance of the microbiome in general, would appear to be in tension with the fact that historically a significant portion of antibiotic use in aquaculture has been prophylactic (45): presumably this is the case because farmers have concluded that the health benefits of prophylactic antibiotic use outweigh the costs. Understanding the effect that antimicrobials – either conventional antibiotics or AMPs – exert on the microbiome, and the consequences of those effects on fish health, is key to integrating the findings of these developing fields. In this study, we examined the effect of a canonical insect-derived AMP – Hyalaphora cecropia cecropin A – on the microbiome, disease resistance, and carrier 65 status of Rainbow trout (Oncorhynchus mykiss) in the context of bacterial infection. The cecropin family of AMPs is highly expressed in insects when exposed to infection by gram-negative bacteria (46) and has strong in vitro activity against this group of bacteria. Our hypothesis was that introduction of cecropin A in the diet would protect trout against the gram-negative bacterium Yersinia ruckeri, the causative agent of enteric redmouth disease (ERM) and would reduce the prevalence of detectable load of Y. ruckeri in the intestine over the course of infection and recovery. The timing of when to introduce cecropin to the fish was an important consideration. In most previous studies of AMPs involving a disease challenge, fish spend at least a month on experimental feeds before being exposed to pathogens. In the context of this project, we aimed to better evaluate the potential effectiveness of dietary cecropin as a tool for disease intervention. Therefore, we opted to begin feeding 3 days prior to the challenge. This gave a sufficient window to ensure that fish would accept the experimental feed while minimizing changes to the underlying host biology. We also sought to understand how the fish microbiome responds over the course of infection and how, if at all, dietary cecropin might modify this response. Our findings reveal that in this experimental context, cecropin A did not improve resistance to infection or clearance of Y. ruckeri from the intestine. Analyses of intestinal microbiomes indicated that fish fed a control diet displayed a lower alpha-diversity following a challenge with Y. ruckeri compared to those fed the +cecropin diet, whereas in mock- challenged fish an increase in alpha diversity was observed in both diet groups. 66 MATERIALS AND METHODS In vitro activity of cecropin A against Yersinia ruckeri Recombinant cecropin A peptide (Abbexa, Ltd., Houston, TX) was used for this project. Its activity against Y. ruckeri was first tested in vitro in three replicate experiments. Log-phase broth culture of Y. ruckeri strain CSF007-82 (47) was diluted to a density of 1.5–5 x106 CFU/ml and incubated at 16°C with serial dilutions of cecropin A in triplicate wells of a 96 well plate. OD600 was recorded for 48 hours using an Elx808 absorbance microplate reader (BioTek). Samples with no change in OD600 after 48 hours were deemed fully inhibited; the minimum inhibitory concentration (MIC) was defined as the lowest tested concentration of cecropin A for which no change in OD600 was observed. To differentiate between bacteriostatic and bactericidal concentrations, samples with no apparent bacterial growth were plated on tryptic soy agar. The bactericidal concentration was defined as the lowest concentration of cecropin A that killed 99.9% of the initial inoculum. Diet production and analysis Two diets, hereafter referred to as the control and +cecropin diets, were created for this experiment using BioClarks Fry 1.2-mm pellets (Bio-Oregon, Longview, WA) as a basal diet. This diet was selected because, in contrast to many commercial diets for fry, it does not contain immune-stimulating ingredients. The +cecropin diet was generated by vacuum-spray coating the basal diet with aqueous cecropin A: 110 mg of cecropin A per kilogram of basal diet, which is within the range used in previous studies of this AMP in fish diets (48,49). The control diet was generated by vacuum 67 spray-coating the basal diet with sterile water. Both diets were lyophilized and stored at -20°C. Proximate analysis of remaining diet samples was performed by Exact Scientific Services (Ferndale, WA) at the conclusion of the study (Table S1). Table S3.1. Nutritional analysis of experimental diets Proximate compositiona Control diet +Cecropin diet Carbohydrates 15.4 16.8 Protein 50.6 47.8 Fat 21.1 19.9 Ash 6.8 6.4 Moisture 6.2 9.2 a Experimental diets were made using BioClarks Fry 1.2-mm pellets as the base. Pellets were vacuum spray-coated with either distilled water (control diet) or aqueous cecropin A (+cecropin diet) and then lyophilized and stored at -20°C until feeding. Fish care and handling Rainbow trout (Oncorhynchus mykiss) fry were provided by the New York State fish hatchery at Bath, NY where they were kept outdoors in flowthrough concrete raceways. All animals were cared for in compliance with the Guide for the Care and Use of Laboratory Animals and American Association of Laboratory Animal Science Position Statements, and all experimental protocols were approved by the Cornell University Institutional Care and Use Committee (IACUC). Upon arrival at the Cornell aquatic facility, fish were kept in a 700 liter fiberglass tank under flowthrough conditions at 12±1°C and fed BioClarks Fry 1.2-mm pellets at a rate of 3% of total bodyweight per day. Over a third week of acclimation, temperature was 68 gradually increased from 12±1°C to 15±1°C. For the duration of the experiment, effluent from fish tanks was collected and decontaminated with a chlorine injection system prior to being released into the sewer system. For the first trial, 172 fish (average body weight = 2.5g) were randomly assigned to one of four treatments: control diet mock-infection (28 fish), +cecropin diet mock-infection (28 fish), control diet infection (58 fish), or +cecropin diet infection (58 fish). Three days after commencement of feeding the experimental diets, fish were either challenged with Y. ruckeri or mock-challenged as described below. Six fish were sampled at each of the first three timepoints and all remaining fish were sampled at day 30. Upon completion of this trial, all tanks were sterilized and flushed with water for 48 hours in preparation for the second trial. The second trial focused on repeating the infection treatments with larger sample sizes: 138 fish (avg. = 4.0g) were randomly assigned to each diet. Fish were fed the experimental diet for three days before being challenged with Y. ruckeri. Eight fish from each diet group were sampled at each of the first three timepoints and 10 fish per diet group were sampled at day 30. Bacterial challenge design and sampling Fish were challenged by immersion with Y. ruckeri strain CSF007-82 at a concentration of ~8x108 CFU/ml. The mock-challenge group was immersed in water with a corresponding amount of sterile tryptic soy broth (TSB). Challenges took place at a target stocking density of 7.5g/L in randomly assigned 10-L tanks within a zebrafish rack modified to operate as a flowthrough system with supplemental 69 aeration. Water flow was paused during the challenge. Bacterial concentration was verified by plating of water samples from each tank. After 1 hour, flow was restored to rapidly flush out the system; after 24 hours, fish were transferred back to 700-L flowthrough tanks and maintained on the experimental diets for the remainder of the experiment. Mortalities and morbidities were monitored for 30 days. Appetite was notably depressed in the challenged group from day 3 to day 14 post-infection; during this period feeding was reduced to 1% of total estimated bodyweight per day. For all other days, feeding was 3% of total estimated bodyweight. Samples were collected at 4 timepoints: immediately prior to the challenge (day 0), and then on days 3, 8, and 30 post-challenge. Sampled fish were euthanized via overdose with buffered MS-222 and all intestinal tissue (including contents) posterior to the pyloric caeca were aseptically collected, flash frozen, and stored at - 80°C. In addition, posterior kidney samples from challenged fish were streaked on TSA plates for confirmation of infection. At 30 days post-challenge all remaining fish were euthanized with buffered MS-222. DNA extraction from fish intestines DNA was extracted as previously described (50). Briefly, whole intestines were weighed before an initial homogenization step using 1.3mm chrome steel beads and DNA extraction buffer (200 mM NaCl, 200 mM Tris– HCl pH 7.5, 20 mM EDTA, 5% SDS). A volume of the primary homogenate corresponding to 20 mg of intestine was subsequently subjected to a secondary homogenization step with 0.1-mm 70 zirconia/silica beads. DNA was isolated via phenol:chloroform extraction and isopropanol precipitation before resuspension in Tris-EDTA buffer. qPCR detection of the Y. ruckeri 16S rRNA gene Quantitative polymerase chain reaction (qPCR) was performed using a 7500 real-time PCR instrument (Applied Biosystems) for detection of Y. ruckeri in intestinal DNA extracts. We used a previously published hydrolysis (Taqman) primer/probe set targeting the Y. ruckeri 16S rRNA gene (51). The primers target the V2 and V3 regions, while the probe targets the conserved region in between. Reactions of 10μl were performed in triplicate wells of a 96-well plate using 5μl 2X PrimeTime Gene Expression Master Mix (Integrated DNA Technologies, Coralville, IA), primers (400nM each), probe (100nM), and DNA template (1μl). Samples for which the 16S gene was detected in at least two of three wells were considered positive for Y. ruckeri. PCR amplification of 16S rRNA gene for DNA sequencing Intestinal DNA extracts were diluted 1:10 before being amplified by PCR targeting the 16S rRNA gene sequence (region V4) using the universal primers 515F and GoLay-barcoded 806R and the PCR program as previously described (52). Samples were amplified in duplicate with the following thermocycler protocol: hold at 94°C for 3 min; 30 cycles of 94°C for 45 s, 50°C for 1 min, 72°C for 1.5 min; and hold at 72°C for 10 min, and the duplicate final amplified products were pooled. Amplicons were further purified using Mag-Bind® RxnPure Plus (Omega 71 Bio-tek, Inc., GA) and quantified with NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA), and 150 ng of amplicons from each sample were pooled and paired-end sequenced (2x250bp) on an Illumina MiSeq instrument in the Genomic Facility at Cornell Institute of Biotechnology. Analysis of 16S rRNA gene sequences Sequence data processing was performed using the QIIME 2 (v 2020.2) pipeline (53). The Divisive Amplicon Denoising Algorithm 2 (DADA2) (54) method was applied to quality-filter sequences and categorize amplicon sequence variants (ASVs). Taxonomic alignment of all ASVs was performed using Wasabi (55). The resulting ASVs were assigned taxonomy by mapping with the Greengenes 16S rRNA Gene Database (56). Feature tables were generated collapsed at various taxonomic levels including phylum, class, order, family, genus, and species. The QIIME2 diversity plugin was used to compute Shannon’s diversity index and Bray-Curtis distances for alpha and beta diversity evaluation (57). Beta diversity distances were calculated between all samples to facilitate comparisons between and within each group. The QIIME output data was then imported to Rstudio (Version 1.0.136) with several packages including phyloseq (58), microbiome (59), tidyverse (60), EnhancedVolcano (61) and ggplot2 (62) for normalizing and plotting of the input data. Statistical analysis Fish survival was plotted using a Kaplan-Meier curve; sampled fish and those surviving to the end of the experiment were censored at the appropriate time points. 72 Analysis of survival data used JMP Pro 14 software to conduct a Gehan-Breslow- Wilcoxon test and a Cox proportional hazard model accounting for diet and trial effects. Detection data was analyzed using Fisher’s exact test for each timepoint with Bonferroni correction for multiple comparisons. QIIME 2 (v 2020.2) and Rstudio (v 1.0.136) were used to analyze sequence data. In the QIIME workflow, samples with poor quality (as evaluated by over 60% chimeric reads) were ignored in subsequent analysis and graphing (Table S2). Shannon and Bray-Curtis diversity metrics were tested against sample metadata factors using QIIME2 diveristy plugin (57). Shannon index was compared with pairwise Kruskal-Wallis tests. Bray-Curtis distances were compared with pairwise permutational multivariate analysis of variance (PERMANOVA) tests using 999 permutations and Benjamini/Hochberg FDR p-value adjustment for pairwise comparisons. Volcano plots for identifying differentially abundant taxa between two selected groups were calculated via DESeq2 (63). The correlation coefficients plot between Y. ruckeri and taxonomic variables was generated by calculating the Pearson correlation between centered log-ratio transformation of the abundance of taxonomic variables and the Cq value of Y. ruckeri level in the intestine. Table S3.2. Distribution of retained samples after sequencing quality control Challenge treatment Mock-Challenge treatment Samples Diet group Samples sequenced Samples sequenced Samples retained retaineda Control 60 53 28 26 +Cecropin 56 49 28 26 a After sequencing, samples with greater than 60% chimeric reads were excluded from subsequent analysis on the basis of poor quality. Low-quality samples were evenly distributed across diet, time point, and trial. 73 RESULTS Cecropin A inhibits the growth of Y. ruckeri in vitro We first tested the in vitro activity of cecropin A, an insect AMP, against Y. ruckeri, a fish bacterial pathogen. The minimum inhibitory concentration (MIC) of cecropin A against Y. ruckeri was 3.75 μM under the parameters tested in this study, while partial inhibition of growth was observed as low as 0.47μM (Fig. 3.1). Plating revealed that 3.75 μM is bacteriostatic: while there was no change in OD600 after 48 hours, by plate count the bacterial population actually doubled in this time period. By contrast, both the 7.5 μM and 15 μM concentrations were bactericidal over the same timeframe. The 3.75 μM MIC in this study corresponds to approximately 6.6x10-16 moles (4x108 molecules) of cecropin A for each bacterial cell. Assuming equal distribution of peptide among feed particles and equal consumption of feed by each fish, the feed in this study would have supplied each fish with approximately 3.4x10-9 moles (2.1x1015 molecules) of cecropin A per day. 74 F igure 3.1. Dose-dependent inhibition of Y. ruckeri by cecropin A in vitro. Y. rFuicgkuerrei 1 w: Daoss ien-dceupbenadteendt ainth 1ib6it°iCon ionf Yb.r routchk esrui pbpy lceemcroepnitne Ad iwn ivtihtr os.e Yri. arlu cdkielrui twioans sincubated at 16°C of cecropin in broth supplemented with serial dilutions of cecropin A. The experiment was performed in a 96- Aw.e Tll hpela teex apned rOimD6e0n0 wt wasa rse cpoerdrefdo rinm teridpl iicna tae w96el-lws peelrl spamlaptlee afonrd a OpeDriod owf 4a8s hroeucros.r dPleodtt eidn values 600 are the average of three independent experiments. No change in OD600 was observed in wells containing tr3i.p75li,c 7a.5te, awnde 1ll5s µpMer c escarmoppinle A f o(lar ttae rp tewroio cdon ocef n4tr8a thioonus rnso.t Pshloowttne)d. values are the average of three independent experiments. No change in OD600 was observed in wells containing 3.75, 7.5, and 15 µM cecropin A (latter two concentrations not shown). Cecropin A did not protect fish against Y. ruckeri Considering that cecropin A inhibits the growth of Y. ruckeri in vitro, we aimed to assess whether adding cecropin A to diet could improve fish resistance to a challenge with this bacterial pathogen. Fish were fed a diet supplemented or not with cecropin A and challenged with Y. ruckeri by immersion. The infection challenge was performed twice, with the second trial immediately following the first. Fish were monitored for a period of 30 days post-challenge (Fig. 3.2). In the first trial, mortalities were observed between day 5 and 9 for the control diet group, and between day 4 and 9 for the +cecropin diet group. Survival rates were 31% and 20% for the 75 control and +cecropin diet groups, respectively. In the second trial, mortalities were observed between day 5 and 14 for the control diet group, and between day 4 and 22 for the +cecropin diet group. Survival rates were 43% and 32% for the control and +cecropin diet groups, respectively. No mortality was observed in mock-challenged fish for either diet group. A Cox proportional hazards model combining data from both trials revealed a statistically significant difference between the trials (p<0.001) but not the diets (p=0.059); the infection trials were therefore considered separately for analysis. Differences in survival between diet groups were not statistically significant for either trial. FFiiguree 2 3: .K2a.p Klana-pMlaeiner- Mcureviee srh ocwuirnvge s usrhviovwal ionfg f isshu frevdi eviathle or tfh fei csohn tfreodl o er itthhe edrie tt hsuep pcloenmternoteld o r with cecropin A following an immersion challenge with Y. ruckeri. Each trial was analyzed tihnede dpeiendt esnutlpy.p Tlheemree wnetreed n ow sittahtis ctiecaclr doifpfeinre nAc efso blelotwweeinn gth ea ntw iom dimet egrrosuiopsn f ocrh eaitlhleern tgrieal wacictohrd Yin.g rtuo cthkee Grie.h Eana-cBhre tsrlioawl -wWailsc oaxnoanl tyezste: dp =i0n.d15e pfoern Tdreianl t1ly, p. =T0h.1e6r efo wr Terrieal n2o. N sota mtiosrttiacliatyl dwiafsf eorbesenrcveeds in mock-challenged fish fed either diet. b etween the two diet groups for either trial according to the Gehan-Breslow-Wilcoxon t est: p=0.15 for Trial 1, p=0.16 for Trial 2. No mortality was observed in mock- challenged fish fed either diet. 76 Cecropin A did not reduce the prevalence of Y. ruckeri in intestines of fish While cecropin A did not significantly impact survival, we next assessed whether it had on an effect on the ability of fish to clear Y. ruckeri from the intestine following infection challenge. The presence of Y. ruckeri in the intestines of fish was monitored at 3, 8, and 30 days post-challenge using a qPCR assay described in a previous study (50). In the first trial, Y. ruckeri was detected in 83% and 100% of fish at day 3 and 8, respectively, independent of the diet (Table 3.1). At 30 days post- challenge, 88% of fish in the control diet group and 100% of fish in the +cecropin diet group were positive for Y. ruckeri; this difference was not statistically significant. In the second trial, Y. ruckeri was detected in 88% of fish at day 3 in both the control and +cecropin diets, and in 100% of fish at day 8 independent of the diet. At 30 days post- challenge, 80% of fish in the control diet group and 100% of fish in the +cecropin diet group were positive for Y. ruckeri. No statistically significant differences in prevalence of carrier status were observed between diet groups for any of the time points sampled. There were signs, however, that the bacterial load of carrier fish in the +cecropin diet may have been elevated. 10 of the 11 samples with the lowest cq values (which correspond to high abundance of Y. ruckeri) across both trials were found in the +cecropin diet group (Fig. S3.1). Based on a standard curve of purified Y. ruckeri DNA, these fish were all estimated to possess >1.5x104 bacterial cells per 20 mg intestinal tissue (data not shown). 77 Table 3.1 Detection of Y. ruckeri 16S DNA in intestinal extracts by qPCR Time post-infection Diet group 3 days 8 days 30 days Trial 1 Control 5/6 (83%)a,b 6/6 (100%) 7/8 (88%) +Cecropin 5/6 (83%) 6/6 (100%) 4/4 (100%) Trial 2 Control 7/8 (88%) 8/8 (100%) 8/10 (80%) +Cecropin 7/8 (88%) 8/8 (100%) 10/10 (100%) a Fish positive for Y. ruckeri by qPCR over total number of sampled fish with % in parentheses bComparisons were performed separately for each time point and replicate using Fisher’s exact test. No statistically significant differences were found between diet groups for any time point. F igure S1. Correlation of qPCR quantification cycles (Cq) to relative abundance oFaf i b gYu.r er uSc1k: Ceroir.r eRlaetiloant iovfe q PaCbRun qduaanntcifeic watiaosn cycles (Cq) to relativeundance was transformed via a centered log- rtartaion strfaonrsmforemda vtiioan .a A c abundance of Y. ruckeri. Relative gelnotbearl eadna lloysgis- rraetvieoa led two ASVs, tdriafnfesrfinogr min asteiqoune.n cAe bgylo obnae lb aasnea plyaisr,i sth raet vweearele sdtr otnwgoly AneSgVatisv,e dlyi fcfoerrrienlagte idn t os eCqqu deentecrme ibnyed o unsien g a qPCR assay specific for Y. ruckeri. Low cq values correspond to high abundance of Y. ruckeri. Both bAaSsVe _p3a4i2r8, (tAha) ta nwde AreS Vs_tr1o0n57g l(yB )n reeglaatitviev eablyu ncdoanrrceel sahtoewde tdo s tCroqn gd ceoterrremlatiinoend to u Csqin vga laue qs.P VCeRry high abundance – as determined by qPCR or by sequencing – occurred more frequently in fish fed the a+scseacyro psipne dciiefti. c for Y. ruckeri. Low cq values correspond to high abundance of Y. r uckeri. Both ASV_3428 (A) and ASV_1057 (B) relative abundance showed strong c orrelation to Cq values. Ve ry high abundance – as determined by qPCR or by sequencing – occurred more frequently in fish fed the +cecropin diet. 78 Cecropin A modifies the intestinal microbiome of mock-challenged fish We began our examination of the microbiomes in this study by comparing the differences between the control and +cecropin diet groups in mock-challenged fish. At the start of trial 1, the microbiomes were dominated by a single amplicon sequence variant (ASV) from the genus Mycoplasma (Fig. 3.3A). Over time, the internal (alpha) diversity and richness of both microbial communities increased, as measured by Shannon index, though this increase was more modest in fish fed the +cecropin diet: at day 30 the control diet had significantly higher alpha diversity (Fig. 3.3B). Given the initial high relative abundance of Mycoplasma sp., increases in alpha diversity roughly corresponded to decreases in Mycoplasma sp.; accordingly, control diet group fish had lower average relative abundance of Mycoplasma sp.. The rapid increase in alpha diversity observed in the control diet group led to significantly different microbiomes between diet groups at day 30, as evaluated by PERMANOVA (Fig. 3.3C). Using DESeq2, we identified a number of ASVs significantly enriched at day 30 in the microbiomes of mock-challenged fish fed the control diet: 28 of the 31 ASVs identified in this screen were putatively matched by 16S sequence to gram-negative phyla (FIG. S3.2, Table S3.3). 79 Figure 3: Microbiome comparison between mock-challenged fish fed either control or +cecropin diet. (A Figure 3.3. Microbiome comparison between mock-challenged fish fed either and B) Relative abundance of the core ASVs shifted over time. Microbiomes started out dominated by a cosinngtrleo slp oecri e+s coef cMryocpopinla sdmiae.t A. (lpAh aa dnidve Brsi)t yR ineclraetaisveed ainb buonthd adinect eg rooufp tsh oev ecro trime eA, tShVousg hs hbyif dteady 30 ovaleprh at idmive.r sMityi wcraos bsigonmifiecsa nstltya rhtiegdhe or uint fdisohm feidn tahtee cdo nbtyro la d sieitn. g(Cle) Isnpdeivciideusa lo mf icMroybciompelas somf fais.h fed Atlhpe hcao ndtriovl edriseti twye irne ccoremapsaeredd itno mboictrho bdioiemte gs roof ufipshs foedv ethre t +imceecr,o tphino ufogrh e abcyh tdimaye p3o0in ta. lEpahcah point on the graph represents a pairwise Bray-Curtis distance between individual microbiomes from the two diversity was significantly higher in fish fed the control diet. (C) Individual diet groups. Lower distance values represent more similar microbiomes, while higher distances represent mmicorreo bdiivoemrgeenst omfi cfriosbhi ofmedes .t hStea tcisotincatrl osilg dniifeict awnceer we acso emvapluaarteedd b tyo P mERicMroAbNiOomVAe,s i noifti aflilsyh f ofre adl l the +tcimecerpopinitns afnodr beoathc hdi etitsm, aen dp othiennt .f oEr apcosht -phoci npta iorwni steh ceo gmrpaaprihso rnesp breetwsenents d iae tps afoir weaicshe t iBmreapyo-int. CSutrattiisst idcails tsaignncifeic banectew oene tnh ei ngrdaipvhi rdeufearls mto itchero dbififoermenecse fbreotwmee tnh dei ettw goro udpise tfo gr raonu inpdsi.v iLduoawl er timepoint, where q is the p-value after correction for multiple comparisons (*q<.05) distance values represent more similar microbiomes, while higher distances represent more divergent microbiomes. Statistical significance was evaluated by PERMANOVA, initially for all timepoints and both diets, and then for post-hoc pairwise comparisons between diets for each timepoint. Statistical significance on the graph refers to the difference between diet groups for an individual timepoint, where q is the p-value after correction for multiple comparisons (*q<.05) 80 Control diet +Cecropin diet Figure S3.2. VolcFanigou prloet Ssh2o:w Vinogl tcaaxnao s ipgnloifti csahnotlwy iengri cthaexda isni mgnicirfoicbaionmtleys eofn riched in microbiomes of mock-challenged fish fed mock-challengedt hfiesh c foend ttrhoel corn t+rocle ocrr o+pceicnr odpieint sd iaett st hate t dhea yd a3y0 3 0ti tmimee p pooiintt.. ASVs in red on the left side of the plot were ASVs in red on thsei glenfti fsiidcea noft ltyhe e pnlorit cwheerde siingn tihfieca mntliyc reonrbiciohemd eins tohfe mfiischro fbeiodm tehse control diet. Identified taxa are listed in order of significance of enrichment in Table S3, along with taxonomic identities assigned by comparison to the of fish fed the conGtrorel edinegt. eIdneenst i1fi6edS t arxRaN arAe l idstaetda ibna osred.e r of significance of enrichment in Tab le S3.3, along with taxonomic identities assigned by comparison to the Greengenes 16S rRNA database. 81 Table S3.3. Taxonomic assignments of ASVs enriched in microbiomes of fish fed the control and +cecropin diets 30 days after mock-challenge. Up in control diet Phylum Class Order Family Genus Species ASV_77a Proteobacteria Alphaproteobacteria Rhodobacterales Rhodobacteraceae Rhodobacter NA ASV_2309 Proteobacteria Betaproteobacteria NA NA NA NA ASV_2087 Proteobacteria Alphaproteobacteria Rhizobiales Hyphomicrobiaceae Devosia NA ASV_3167 Proteobacteria Alphaproteobacteria Rhizobiales NA NA NA ASV_4321 Proteobacteria Alphaproteobacteria Rhizobiales Rhizobiaceae Shinella granuli ASV_2189 Proteobacteria Alphaproteobacteria Rhizobiales Bradyrhizobiaceae Bosea genosp. ASV_2764 Proteobacteria Gammaproteobacteria Legionellales Legionellaceae NA NA ASV_1718 Proteobacteria Alphaproteobacteria Rhizobiales NA NA NA ASV_2977 Chlamydiae Chlamydiia Chlamydiales Parachlamydiaceae NA NA ASV_2829 Proteobacteria Alphaproteobacteria Rhizobiales NA NA NA ASV_2692 Proteobacteria Gammaproteobacteria Enterobacteriales Enterobacteriaceae NA NA ASV_3608 Proteobacteria Betaproteobacteria Neisseriales Neisseriaceae Deefgea NA ASV_3050 Firmicutes Clostridia Clostridiales Clostridiaceae NA NA ASV_3665 Proteobacteria Alphaproteobacteria Rhodobacterales Rhodobacteraceae Rhodobacter NA ASV_1238 Proteobacteria Alphaproteobacteria Rhizobiales NA NA NA ASV_2613 Proteobacteria Alphaproteobacteria Rhodospirillales Rhodospirillaceae Reyranella massiliensis ASV_3315 Actinobacteria Actinobacteria Actinomycetales Nocardiaceae Rhodococcus fascians ASV_4311 Chlamydiae Chlamydiia Chlamydiales Parachlamydiaceae NA NA ASV_4079 Proteobacteria Betaproteobacteria Burkholderiales Comamonadaceae Pelomonas NA ASV_2090 Proteobacteria Alphaproteobacteria Rhodobacterales Rhodobacteraceae Rhodobacter NA ASV_704 Bacteroidetes Sphingobacteriia Sphingobacteriales NA NA NA ASV_1585 Proteobacteria Gammaproteobacteria Aeromonadales Aeromonadaceae NA NA ASV_3380 Proteobacteria Alphaproteobacteria Rhodobacterales Rhodobacteraceae Rhodobacter NA ASV_208 Actinobacteria Actinobacteria Actinomycetales Microbacteriaceae Salinibacterium NA ASV_3273 Proteobacteria Gammaproteobacteria Legionellales NA NA NA ASV_3135 Proteobacteria Alphaproteobacteria Rhodospirillales Rhodospirillaceae Reyranella massiliensis ASV_501 Tenericutes Mollicutes Mycoplasmatales Mycoplasmataceae Mycoplasma NA ASV_4126 Proteobacteria Gammaproteobacteria Legionellales Coxiellaceae NA NA ASV_1593 Proteobacteria Alphaproteobacteria NA NA NA NA ASV_605 Bacteroidetes Sphingobacteriia Sphingobacteriales Sphingobacteriaceae Pedobacter NA ASV_3380 Proteobacteria Alphaproteobacteria Rhodobacterales Rhodobacteraceae Rhodobacter NA Up in +cecropin diet Phylum Class Order Family Genus Species ASV_4681 Firmicutes Bacilli Lactobacillales Lactobacillaceae NA NA ASV_2436 TM6 SJA-4 NA NA NA NA a ASVs highlighted in green are from gram-negative phyla 82 Cecropin A influences the alpha-diversity of the intestinal microbiome of fish challenged by bacterial infection We performed two challenge trials to investigate the effect of dietary cecropin on the fish microbiome in the context of infection. Fish in trial 2 were from the same origin and age cohort as those in trial 1 but had been maintained in a separate room in the Cornell aquatic facility for the duration of trial 1. Initial microbial community structure differed between the two trials, with fish at the start of trial 1 displaying higher average relative abundance of Mycoplasma sp. than those at the start of trial 2 (Fig. 3.4A). The relative abundance of Mycoplasma sp. appeared to increase following the infection challenge in both trials but this effect was more pronounced in trial 2, owing to lower initial relative abundance (Fig. 3.4A). Reflecting the observed patterns in Mycoplasma sp. relative abundance, alpha diversity was higher for most fish in the second trial than the first, but the same pattern was observed in both cases (Fig. 3.4B). In fish challenged with Y. ruckeri, a drop in alpha diversity was observed from day 0 to day 3 independent of diet (Fig. 3.4B). Alpha diversity subsequently appeared to increase in both groups at day 8 and again at day 30 post-challenge. The magnitude of these increases, however, was notably smaller for fish fed the control diet than for those fed the +cecropin diet. At day 30 post-challenge, alpha diversity was higher for fish fed the +cecropin diet in trial 1 (p=0.012); the same trend was apparent in trial 2, though in this case the difference between diet groups was not statistically significant (p=0.054) (Fig. 3.4B). This suggests that dietary supplementation of cecropin A can improve the recovery of alpha diversity of the microbiome following Y. ruckeri infection. 83 Figure 3.4. Microbiome comparison between fish challenged with Y. ruckeri fed either control or +cecropin diet. (A) Relative abundance of the 9 most prevalent taxa are shown for individual fish sampled before challenge (day 0) and post challenge (days 3, 8, and 30). Taxonomic assignments represent the most specific level of resolution assigned by Greengenes, with the exception of ASV_3428 which was identified by multiple methods (see text for details). Fish marked with an * had high absolute abundance of Y. ruckeri as measured by qPCR (detection after 25 or fewer cycles, corresponding to ~1.5x104 bacteria/20 mg of intestinal tissue). Baseline microbiomes at the start of trial 1 were markedly different from those at the start of trial 2. (B) Alpha diversity as measured by Shannon index for trial 1 and trial 2; p- values are for comparisons of diet groups at each timepoint, as assessed by Wilcoxon signed rank test with correction for multiple comparisons. 84 Identification of ASVs corresponding to Y. ruckeri and correlation analysis We next aimed to determine what microbiome states were associated with detection of Y. ruckeri. There was a strong correlation between Y. ruckeri qPCR threshold count, which depends on the abundance of copies of the target sequence at the beginning of the reaction, and the relative abundance of ASV_3428 and ASV_1057 (Fig. S3.1). Using BLAST, both candidate ASVs’ sequences aligned to the published 16S sequence of the Y. ruckeri CSF007-82 strain used in this study (47). Both sequence variants, which differed by one base pair, appeared in each challenge trial, suggesting that the frozen stock of Y. ruckeri used in this study may have contained both genotypes. Phylogenetic alignment of ASVs identified a total of 5 additional ASVs with at least 99.5% sequence identity to Y. ruckeri CSF007-82. All of these additional ASVs were extremely rare, however, occurring in fewer than 2% of fish sampled and only at very low abundance. A comparison of qPCR and sequencing data confirmed that while identification of carriers was possible using sequence data, this method was less sensitive than qPCR: only 30% of the fish identified as carriers by qPCR had sequence reads corresponding to Y. ruckeri after quality control and filtering. ASV_3428 and ASV_1057 were then used to conduct a global correlation analysis in the microbiomes of challenged fish. Statistically significant correlations were identified for 14 ASVs (Fig. 3.5). Six ASVs were found to have positive correlations to Y. ruckeri, including ASVs from the genera Flavobacterium, Achromobacter, and Sphingomonas. Seven ASVs were negatively correlated with Y. ruckeri, suggesting possible benefits for fish health. These ASVs included two strains 85 identified as belonging to the family Lactobacillaceae, as well as ASV_1159, which was characterized only to the class level as a Betaproteobacteria in the GreenGenes database. ASV_1159 was detected in all but one fish sampled in this study, though the average relative abundance was less than 4%. Also negatively correlated with Y. ruckeri was the Mycoplasma sp. species that was dominant in most fish. The negative correlation of Y. ruckeri and Mycoplasma sp. at the individual level contrasts with the observation that at the fish population level exposure to Y. ruckeri led to an increase in Mycoplasma sp. relative abundance at (Fig. 3.4A). 86 Negative correlation with Y. ruckeri Phylum Class Order Family Genus Species ASV_3932 Firmicutes Bacilli Lactobacillales Lactobacillaceae NA NA ASV_2891 Bacteroidetes [Saprospirae] [Saprospirales] Chitinophagaceae Sediminibacterium NA ASV_501 Tenericutes Mollicutes Mycoplasmatales Mycoplasmataceae Mycoplasma NA ASV_1159 Proteobacteria Betaproteobacteria NA NA NA NA ASV_2086 Proteobacteria Betaproteobacteria Neisseriales Neisseriaceae Deefgea NA ASV_2076 Proteobacteria Gammaproteobacteria Xanthomonadales Sinobacteraceae NA NA ASV_4681 Firmicutes Bacilli Lactobacillales Lactobacillaceae NA NA Positive correlation with Y. ruckeri Phylum Class Order Family Genus Species ASV_3213 Bacteroidetes Flavobacteriia Flavobacteriales Flavobacteriaceae Flavobacterium succinicans ASV_4425 Proteobacteria Betaproteobacteria Burkholderiales Alcaligenaceae Achromobacter NA ASV_3173 Proteobacteria Gammaproteobacteria Enterobacteriales Enterobacteriaceae Serratia NA ASV_1095 Proteobacteria Alphaproteobacteria Sphingomonadales Sphingomonadaceae Sphingomonas NA ASV_550 Proteobacteria Alphaproteobacteria Caulobacterales Caulobacteraceae NA NA ASV_4147 Firmicutes Bacilli Lactobacillales Streptococcaceae Lactococcus NA FFiguigreu 5r: Ceo r3re.l5at.io Cn aonarlyrseisl baettwioeenn A aSnV_a3l4y28s iasnd b AeStVw_1e05e7n id eAntiSfiVed _as3 Y4. r2uc8k eari nandd eAlemSeVnts_ o1f t0he5 m7i cirdobeionmtei. fAilel tdax aa ssh own have a statistically significant correlation (p<0.05) to Y. ruckeri after correction. The size of the circles represents the significance of the correlation (lYarg. er ucircclkese hraiv ea lnowder ep-lveamluese) nantds t hoe fc otlohr ede mnotiecs trhoe dbirieoctmione a.n dA stlreln tgathx oaf thseh coorwrelnat iohna. Fvoer eaxa smtpalet,i AsStiVc_a44ll2y5 is strongly positively correlated to ASV_3428 and ASV_1057 with high statistical significant. Taxonomic assignments for each ASV were generated by cosmigpanriisfoinc toa nthte Gcroeernrgeelnaest idoatnab a(spe. <0.05) to Y. ruckeri after correction. The size of the circles represents the significance of the correlation (larger circles have lower p-values) and the color denotes the direction and strength of the correlation. For example, ASV_4425 is strongly positively correlated to ASV_3428 and ASV_1057 with high statistical significance. Taxonomic assignments for each ASV were generated by comparison to the Greengenes database. 87 DISCUSSION This project aimed to evaluate the impact of dietary cecropin A, an insect- derived AMP, on fish health. Specifically, this project aimed (i) to determine if dietary cecropin A aids in fighting colonization of the fish intestines by Y. ruckeri, the cause of enteric redmouth disease, and (ii) to evaluate the effects of dietary AMPs, such as cecropin A, on the integrity of the intestinal microbiota. First, we observed that dietary cecropin A had a slight negative effect on fish health, with trends towards decreased survival and increased pathogen load in fish that survived infection at the 30-day time point. Second, we observed that Mycoplasma sp. dominated for the most part the intestinal microbiome of fish independent of diet in the early stages of the experiment. By day 30, the microbiome of mock-challenged fish showed an increase in alpha diversity independent of diet. However, fish fed the +cecropin diet had higher relative abundance of the dominant strain of Mycoplasma sp. compared to fish on control diet. On the contrary, following a challenge with Y. ruckeri, the microbiome of surviving fish fed the +cecropin diet group showed higher alpha diversity and lower relative abundance of Mycoplasma sp. than the control diet group at the 30-day time point. Overall, these observations indicated that cecropin A was biologically active, although not beneficial to fish health when given as a feed ingredient. Nevertheless, the observed changes in intestinal microbiota of challenged and mock-challenged fish over a period of 30 days should contribute to a better understanding of the relationship between AMPs, microbiome composition, and fish health. As expected, cecropin A proved to be inhibitory of Y. ruckeri in vitro. This supported our experimental approach, though it was of limited use in determining an 88 effective oral dose for in vivo experiments. To set our experimental dose of cecropin A, we relied on previously published studies using cecropin as a feed additive in nile tilapia [75, 150, 225 mg/kg](48) and crucian carp [100, 150, 200, 250 mg/kg](49). Both of these studies reported increased growth and survival in the context of an infection challenge, as well as measurable changes to physiological parameters including immune status and tissue composition at higher doses. A study in turbot using a higher dose range (250, 500, 750, 1000 mg/kg) reported negative effects on the microbiome diversity and disease resistance at 1000 mg/kg (64). Accordingly, we set our experimental dose at 110 mg of cecropin A per kg of feed. The moisture content of the +cecropin diet was 9.2% compared to 6.2% for the control diet. This could be the result of suboptimal storage – the diets were kept at -20° C and daily opening of the cold container in the humid conditions of the aquatic facility may have resulted in accumulation of condensation; because there was only a small amount of +cecropin feed remaining at the end of the study, this effect would have been more pronounced for this diet. While we cannot rule out the possibility that this had an effect on the nutritional quality of the +cecropin feed, both diets were comfortably above the published requirement for rainbow trout of this size (65). We observed a trend towards lower survival in the +cecropin diet following a challenge with Y. ruckeri, though this difference was not statistically significant. Overall survival was lower than anticipated, resulting in fewer samples for microbiome analysis. We subsequently repeated the challenge in a second trial using fish from the same cohort and larger sample sizes. In the second trial, survival in both groups was somewhat improved relative to the first trial, which could be explained by 89 the fact that younger fish are more susceptible to infection with Y. ruckeri (66). The trend, however, was similar to the first trial with a lower survival rate in the group of fish fed a diet supplemented with cecropin, though again not statistically significant. Since dietary cecropin did not improve survival, we next tried to determine if it could reduce the prevalence of Y. ruckeri in the trout intestine over the course of infection and recovery. We found no significant differences in the likelihood of detection between diet groups for any timepoint, suggesting that cecropin did not help eliminate Y. ruckeri from the gut. We did, however, observe that most of the fish with high absolute abundance of Y. ruckeri as measured by qPCR of intestinal DNA extracts came from the +cecropin diet group, possibly suggesting that this group experienced more severe infections. The microbiomes of fish in this study started out, and to some extent remained, overwhelmingly dominated by a single taxon identified as belonging to the genus Mycoplasma. At the beginning of the first trial, Mycoplasma sp. accounted for at least 75% of the sequence reads in all sampled fish. For fish not exposed to the infection challenge, however, the proportion of Mycoplasma sp. tended to decrease over time, especially in fish fed the control diet (Fig. 3.3A). Though the larger fish in the second trial began with lower mean relative abundances of Mycoplasma sp. compared to those in the first trial, in both trials we observed that exposure to Y. ruckeri was linked to higher mean Mycoplasma sp. relative abundance by day three post-challenge (Fig. 3.4A). It should be noted that in challenged fish, decreased feed intake from day 3-14 introduces an additional variable. We extracted DNA from whole intestinal samples, so lower feed intake would be expected to alter the ratio of digesta to mucosa in these 90 samples. In fish, digesta samples tend to have higher microbial diversity than mucosa samples (67); decreased feed intake post-challenge could therefore help explain the observed decrease in alpha diversity in this period. Interestingly, while exposure to Y. ruckeri led to higher Mycoplasma sp. levels at the population level, for individual fish higher detected levels of Y. ruckeri were associated with lower relative abundance of Mycoplasma sp. (Fig. 3.5, Fig. S3.1). This suggests that relative abundance of Mycoplasma sp. may be an indicator of the ability of the fish to maintain low pathogen loads. In addition to Mycoplasma sp., negative correlations were observed between Y. ruckeri and six other ASVs, including two from the family Lactobacillaceae (Fig. 3.5). These are strong candidates to be considered health-promoting members of the microbiome as certain Lactobacillus spp. are well-established probiotics, and members of this family have been shown to improve growth and resistance to bacterial infection in rainbow trout (68). On the other hand, positive correlations to Y. ruckeri included Flavobacterium succinicans, which has previously been identified as a likely opportunistic pathogen in the context of bacterial gill disease (69). Additional positive correlations included genus-level identifications that are consistent with previously reported fish pathogens: Serratia (70,71) and Lactococcus (72). Furthermore, a member of the Sphingomonas genus has recently been reported as an opportunistic pathogen of rainbow trout (73). This suggests that Y. ruckeri infection created a niche for the opportunistic expansion of other potential pathogens. Mycoplasma sp. relative abundance in the intestine has recently been proposed as a biomarker of health in rainbow trout (74). One of the earliest sequence-based 91 analyses of the salmonid gut revealed that Mycoplasma sp. comprised more than 95% of the microbiome in wild Atlantic salmon (Salmo salar) (75). More recently, Mycoplasma sp. has been recognized as a natural and likely beneficial component of the salmonid microbiome (76–79). Increased relative abundance of Mycoplasma sp. in the intestine has been shown to correlate with greater resistance to disease (13,14,80,81). Furthermore, in systems where Mycoplasma sp. is present, it frequently comprises a majority of all sequence reads; this is consistent with the observation that for salmonids, lower alpha diversity is at least in some cases associated with microbiome health rather than dysbiosis (17,18,80,82,83). Our study offers a rare longitudinal view of Mycoplasma sp. abundance in the trout gut before, during, and after an acute infection challenge. Our data show that average Mycoplasma sp. relative abundance was elevated in DNA extracted from whole intestines in the time points immediately following the bacterial challenge. This is in contrast to the findings of Rasmussen et. al who saw marked decreases in average Mycoplasma sp. relative abundance in the digesta of trout fed a control diet following Y. ruckeri infection (74). However, in keeping with that study, we also observed that during infection, high Mycoplasma sp. relative abundance at the individual fish level was associated with lower relative abundance of Y. ruckeri (Fig. 3.5). A recent survey of wild and hatchery-reared Atlantic salmon from the United Kingdom and France identified Mycoplasma sp. as more abundant in hatchery fish than wild fish and as the sole core element of the microbiome across three separate hatcheries (present in >80% of individuals), though the relative abundance of Mycoplasma sp. varied substantially by hatchery (84). The fish in this study were 92 received from a hatchery where they were kept at high stocking density. Over the experimental period, we observed a trend towards reduction in Mycoplasma sp. relative abundance, and an increase in alpha diversity (Fig. 3.3A). Brown et. al observed differences in Mycoplasma sp. relative abundance and alpha diversity of rainbow trout held at different densities (81), suggesting that the lower stocking density in our facility may have played a role in this shift. This highlights a potential larger issue in the design of microbiome-oriented studies of fish disease: research settings are unlikely to replicate the stocking density and immune stress experienced by fish in farm and hatchery settings. It should be noted that Mycoplasma sp. has not been found in all studies of the salmonid microbiome, including two high-profile papers that sought to define reference microbiomes for Atlantic salmon (67) and rainbow trout (85). The factors that favor proliferation of Mycoplasma sp. in the microbiome are a topic worthy of further study. Interestingly, the population-level decline in Mycoplasma sp. relative abundance over time that took place in mock-challenged fish was less pronounced in this study for fish fed the +cecropin diet (Fig. 3.3A), though the mechanism underlying this difference remains unclear. The difference in microbiomes between the two diet groups could potentially help explain the reported effectiveness of longer- term prophylactic treatment of fish with AMPs prior to infection challenge (34–36). However, when fish had been challenged by exposure to a bacterial pathogen, the opposite was true: infection challenged fish fed the +cecropin diet had a more pronounced drop off in Mycoplasma sp. relative abundance at day 30 post-challenge 93 than fish fed the control diet, suggesting that cecropin A may have interfered with innate resistance mechanisms to infection. This study highlights the complexity of the interactions between AMPs, the microbiome, and the host response to infection. Cecropin did not prove to be effective at preventing mortality, and if anything appeared to exacerbate the infectious burden in this context. That said, there are many variables that we did not test that could affect this outcome – for example, the dose of cecropin and the timing of the challenge relative to the introduction of the AMP. Examination of the fish microbiome over the course of infection supports the idea that Mycoplasma sp. is intimately involved in the fish resistance to colonization by Y. ruckeri. Any future use of AMP-based therapeutics in aquaculture settings will likely have to take into account the effects of AMPs on the host microbiome and immune system. 94 References 1. Schar D, Klein EY, Laxminarayan R, Gilbert M, Van Boeckel TP. Global trends in antimicrobial use in aquaculture. Sci Rep. 2020 Dec 14;10(1):21878. 2. Millanao AB, Barrientos MH, Gómez CC, Tomova A, Buschmann A, Dölz H, et al. Injudicious and excessive use of antibiotics: public health and salmon aquaculture in Chile. Rev Med Chil. 2011 Jan 1;139(1):107–18. 3. Rico A, Phu TM, Satapornvanit K, Min J, Shahabuddin AM, Henriksson PJG, et al. Use of veterinary medicines, feed additives and probiotics in four major internationally traded aquaculture species farmed in Asia. Aquaculture. 2013 Nov 1;412–413:231–43. 4. Cabello FC, Godfrey HP, Buschmann AH, Dölz HJ. Aquaculture as yet another environmental gateway to the development and globalisation of antimicrobial resistance. The Lancet Infectious Diseases. 2016 Jul 1;16(7):e127–33. 5. Chuah LO, Effarizah ME, Goni AM, Rusul G. Antibiotic Application and Emergence of Multiple Antibiotic Resistance (MAR) in Global Catfish Aquaculture. Curr Envir Health Rpt. 2016 Jun 1;3(2):118–27. 6. Stentiford GD, Sritunyalucksana K, Flegel TW, Williams BAP, Withyachumnarnkul B, Itsathitphaisarn O, et al. New Paradigms to Help Solve the Global Aquaculture Disease Crisis. PLOS Pathogens. 2017 Feb 2;13(2):e1006160. 7. Austin B. Infectious disease in aquaculture prevention and control. Oxford; Philadelphia: Woodhead Pub. Ltd.; 2012. 8. de Bruijn I, Liu Y, Wiegertjes GF, Raaijmakers JM. Exploring fish microbial communities to mitigate emerging diseases in aquaculture. FEMS Microbiology Ecology [Internet]. 2018 Jan 1 [cited 2019 Nov 11];94(1). Available from: https://academic.oup.com/femsec/article/doi/10.1093/femsec/fix161/4675208 9. Legrand TPRA, Wynne JW, Weyrich LS, Oxley APA. A microbial sea of possibilities: current knowledge and prospects for an improved understanding of the fish microbiome. Reviews in Aquaculture. 2020;12(2):1101–34. 10. Infante-Villamil S, Huerlimann R, Jerry DR. Microbiome diversity and dysbiosis in aquaculture. Reviews in Aquaculture. 2021;13(2):1077–96. 11. Ringø E, Zhou Z, Vecino JLG, Wadsworth S, Romero J, Krogdahl Å, et al. Effect of dietary components on the gut microbiota of aquatic animals. A never- ending story? Aquacult Nutr. 2016 Apr;22(2):219–82. 95 12. Ingerslev HC, Strube ML, Jørgensen L von G, Dalsgaard I, Boye M, Madsen L. Diet type dictates the gut microbiota and the immune response against Yersinia ruckeri in rainbow trout (Oncorhynchus mykiss). Fish & Shellfish Immunology. 2014 Oct 1;40(2):624–33. 13. Li T, Long M, Ji C, Shen Z, Gatesoupe FJ, Zhang X, et al. Alterations of the gut microbiome of largemouth bronze gudgeon (Coreius guichenoti) suffering from furunculosis. Sci Rep. 2016 Jul 28;6(1):30606. 14. Bozzi D, Rasmussen JA, Carøe C, Sveier H, Nordøy K, Gilbert MTP, et al. Salmon gut microbiota correlates with disease infection status: potential for monitoring health in farmed animals. Anim Microbiome. 2021 Apr 20;3:30. 15. She R, Li TT, Luo D, Li JB, Yin LY, Li H, et al. Changes in the Intestinal Microbiota of Gibel Carp (Carassius gibelio) Associated with Cyprinid herpesvirus 2 (CyHV-2) Infection. Curr Microbiol. 2017 Oct 1;74(10):1130–6. 16. Reid KM, Patel S, Robinson AJ, Bu L, Jarungsriapisit J, Moore LJ, et al. Salmonid alphavirus infection causes skin dysbiosis in Atlantic salmon (Salmo salar L.) post-smolts. Boudinot P, editor. PLoS ONE. 2017 Mar 6;12(3):e0172856. 17. Llewellyn MS, Leadbeater S, Garcia C, Sylvain FE, Custodio M, Ang KP, et al. Parasitism perturbs the mucosal microbiome of Atlantic Salmon. Sci Rep. 2017 Mar 7;7(1):43465. 18. Vasemägi A, Visse M, Kisand V. Effect of Environmental Factors and an Emerging Parasitic Disease on Gut Microbiome of Wild Salmonid Fish. Blader IJ, editor. mSphere. 2017 Dec 27;2(6):e00418-17. 19. Chen SW, Liu CH, Hu SY. Dietary administration of probiotic Paenibacillus ehimensis NPUST1 with bacteriocin-like activity improves growth performance and immunity against Aeromonas hydrophila and Streptococcus iniae in Nile tilapia (Oreochromis niloticus). Fish & Shellfish Immunology. 2019 Jan 1;84:695–703. 20. Yi Y, Zhang Z, Zhao F, Liu H, Yu L, Zha J, et al. Probiotic potential of Bacillus velezensis JW: Antimicrobial activity against fish pathogenic bacteria and immune enhancement effects on Carassius auratus. Fish & Shellfish Immunology. 2018 Jul 1;78:322–30. 21. Gauthier J, Rouleau-Breton S, J. Charette S, Derome N. Stimulated Growth and Innate Immunity in Brook Charr (Salvelinus fontinalis) Treated with a General Probiotic (Bactocell®) and Two Endogenous Probiotics That Inhibit Aeromonas salmonicida In Vitro. Microorganisms. 2019 Jul;7(7):193. 96 22. Cabello FC, Godfrey HP. Salmon aquaculture, Piscirickettsia salmonis virulence, and One Health: Dealing with harmful synergies between heavy antimicrobial use and piscine and human health. Aquaculture. 2019 May;507:451–6. 23. Fjell CD, Hiss JA, Hancock REW, Schneider G. Designing antimicrobial peptides: form follows function. Nat Rev Drug Discov. 2012 Jan;11(1):37–51. 24. Joo HS, Fu CI, Otto M. Bacterial strategies of resistance to antimicrobial peptides. Philos Trans R Soc Lond B Biol Sci. 2016 May 26;371(1695):20150292. 25. Lazzaro BP, Zasloff M, Rolff J. Antimicrobial peptides: Application informed by evolution. Science. 2020 May;368(6490):eaau5480. 26. Silveira RF, Roque-Borda CA, Vicente EF. Antimicrobial peptides as a feed additive alternative to animal production, food safety and public health implications: An overview. Animal Nutrition. 2021 Sep 1;7(3):896–904. 27. Zhou X xia, Wang Y bo, Li W fen. Effect of feeding apidaecin on common carp (Cyprinus carpio) growth performances and immune function. Aquaculture. 2008 Jul 2;279(1):108–12. 28. Hu QY, Wu P, Feng L, Jiang WD, Liu Y, Kuang SY, et al. Antimicrobial peptide Isalo scorpion cytotoxic peptide (IsCT) enhanced growth performance and improved intestinal immune function associated with janus kinases (JAKs)/signal transducers and activators of transcription (STATs) signalling pathways in on- growing grass carp (Ctenopharyngodon idella). Aquaculture. 2021 Jun 30;539:736585. 29. Dong XQ, Zhang DM, Chen YK, Wang QJ, Yang YY. Effects of antimicrobial peptides (AMPs) on blood biochemical parameters, antioxidase activity, and immune function in the common carp (Cyprinus carpio). Fish & Shellfish Immunology. 2015 Nov;47(1):429–34. 30. Ge H, Wang Q, Chen H, Liu G, Pan Y, Chen J, et al. Effects of antimicrobial peptide APSH-07 on the growth performance, anti-oxidation responses, stress resistance and intestine microbiota in large yellow croaker Larimichthys crocea. Aquaculture Nutrition. 2020;26(3):715–26. 31. Wang S, Xie S, Zhou A, Zhang C, Wen L, Xu G, et al. Effects of mixed antimicrobial peptide on the growth performance, antioxidant and immune responses and disease resistance of Pengze crucian carp (Carassius auratus var. Pengze). Fish & Shellfish Immunology. 2021 Jul 1;114:112–8. 32. Lygren B, Sveier H, Hjeltnes B, Waagbø R. Examination of the immunomodulatory properties and the effect on disease resistance of dietary 97 bovine lactoferrin and vitamin C fed to Atlantic salmon (Salmo salar) for a short- term period. Fish & Shellfish Immunology. 1999 Feb;9(2):95–107. 33. de Sousa EL, Assane IM, Santos-Filho NA, Cilli EM, de Jesus RB, Pilarski F. Haematological, biochemical and immunological biomarkers, antibacterial activity, and survival in Nile tilapia Oreochromis niloticus after treatment using antimicrobial peptide LL-37 against Streptococcus agalactiae. Aquaculture. 2021 Feb 25;533:736181. 34. Lee BC, Hung CW, Lin CY, Shih CH, Tsai HJ. Oral administration of transgenic biosafe microorganism containing antimicrobial peptide enhances the survival of tilapia fry infected bacterial pathogen. Fish & Shellfish Immunology. 2019 Dec 1;95:606–16. 35. Rashidian G, Moosazadeh Moghaddam M, Mirnejad R, Mohammadi Azad Z. Supplementation of zebrafish (Danio rerio) diet using a short antimicrobial peptide: Evaluation of growth performance, immunomodulatory function, antioxidant activity, and disease resistance. Fish & Shellfish Immunology. 2021 Dec 1;119:42–50. 36. Simora RMC, Wang W, Coogan M, El Husseini N, Terhune JS, Dunham RA. Effectiveness of Cathelicidin Antimicrobial Peptide against Ictalurid Catfish Bacterial Pathogens. J Aquat Anim Health. 2021 Sep;33(3):178–89. 37. Liu M, Wang B, Jiang K, Gong K, Sun S, Wang L, et al. Rice Bran Expressing a Shrimp Antimicrobial Peptide Confers Delayed Spoilage of Fish Feed and Resistance of Tilapia to Aeromonas hydrophila. Journal of the World Aquaculture Society. 2014;45(3):269–78. 38. Chettri JK, Mehrdana F, Hansen EB, Ebbensgaard A, Overgaard MT, Lauritsen AH, et al. Antimicrobial peptide CAP18 and its effect on Yersinia ruckeri infections in rainbow trout Oncorhynchus mykiss (Walbaum): comparing administration by injection and oral routes. Journal of Fish Diseases. 2017;40(1):97–104. 39. Ting CH, Chen YC, Chen JY. Nile tilapia fry fed on antimicrobial peptide Epinecidin-1-expressing Artemia cyst exhibit enhanced immunity against acute bacterial infection. Fish & Shellfish Immunology. 2018 Oct 1;81:37–48. 40. Theriot CM, Koenigsknecht MJ, Carlson PE, Hatton GE, Nelson AM, Li B, et al. Antibiotic-induced shifts in the mouse gut microbiome and metabolome increase susceptibility to Clostridium difficile infection. Nat Commun. 2014 Jan 20;5(1):3114. 41. Langdon A, Crook N, Dantas G. The effects of antibiotics on the microbiome throughout development and alternative approaches for therapeutic modulation. Genome Med. 2016 Dec;8(1):1–16. 98 42. Gutierrez D, Weinstock A, Antharam VC, Gu H, Jasbi P, Shi X, et al. Antibiotic- induced gut metabolome and microbiome alterations increase the susceptibility to Candida albicans colonization in the gastrointestinal tract. FEMS Microbiology Ecology. 2020 Feb 1;96(1):fiz187. 43. Schmidt V, Gomez-Chiarri M, Roy C, Smith K, Amaral-Zettler L. Subtle Microbiome Manipulation Using Probiotics Reduces Antibiotic-Associated Mortality in Fish. Knight R, editor. mSystems. 2017 Dec 26;2(6):e00133-17. 44. He S, Wang Q, Li S, Ran C, Guo X, Zhang Z, et al. Antibiotic growth promoter olaquindox increases pathogen susceptibility in fish by inducing gut microbiota dysbiosis. Sci China Life Sci. 2017 Nov 1;60(11):1260–70. 45. Ibrahim M, Ahmad F, Yaqub B, Ramzan A, Imran A, Afzaal M, et al. Chapter 4 - Current trends of antimicrobials used in food animals and aquaculture. In: Hashmi MZ, editor. Antibiotics and Antimicrobial Resistance Genes in the Environment [Internet]. Elsevier; 2020 [cited 2022 Feb 22]. p. 39–69. (Advances in Environmental Pollution Research series; vol. 1). Available from: https://www.sciencedirect.com/science/article/pii/B9780128188828000048 46. Lemaitre B, Reichhart JM, Hoffmann JA. Drosophila host defense: Differential induction of antimicrobial peptide genes after infection by various classes of microorganisms. Proceedings of the National Academy of Sciences. 1997 Dec 23;94(26):14614–9. 47. Nelson MC, LaPatra SE, Welch TJ, Graf J. Complete Genome Sequence of Yersinia ruckeri Strain CSF007-82, Etiologic Agent of Red Mouth Disease in Salmonid Fish. Genome Announc. 2015 Feb 26;3(1):e01491-14, /ga/3/1/e01491- 14.atom. 48. Lin X, Chen W, Lin S, Luo L. Effects of dietary cecropin on growth, non- specific immunity and disease resistance of tilapia (Oreochromis niloticus × O. aureus). Aquaculture Research. 2015;46(12):2999–3007. 49. ZiRui W, ShiYan Q, Bo L, JiMing R, Lili W, ZhuQing Y. Cecropin: effects on growth performance, non-specific immunity and disease resistance of triploid crucian carp. Chinese Journal of Animal Nutrition. 2014;26(7):1856–63. 50. Sibinga NA, Marquis H. Tissue‐specific differences in detection of Yersinia ruckeri carrier status in rainbow trout ( Oncorhynchus mykiss ). Journal of Fish Diseases [Internet]. 2021 Aug 25 [cited 2021 Oct 28]; Available from: https://onlinelibrary.wiley.com/doi/10.1111/jfd.13515 51. Ghosh B, Crosbie PBB, Nowak BF, Bridle AR. A highly sensitive, non-invasive qPCR-based strategy for direct quantification of Yersinia ruckeri in fish faeces. J Fish Dis. 2018 Sep;41(9):1421–8. 99 52. Caporaso JG, Lauber CL, Walters WA, Berg-Lyons D, Lozupone CA, Turnbaugh PJ, et al. Global patterns of 16S rRNA diversity at a depth of millions of sequences per sample. Proceedings of the National Academy of Sciences. 2011 Mar 15;108(supplement_1):4516–22. 53. Caporaso JG, Kuczynski J, Stombaugh J, Bittinger K, Bushman FD, Costello EK, et al. QIIME allows analysis of high-throughput community sequencing data. Nat Methods. 2010 May;7(5):335–6. 54. Callahan BJ, McMurdie PJ, Rosen MJ, Han AW, Johnson AJA, Holmes SP. DADA2: High-resolution sample inference from Illumina amplicon data. Nat Methods. 2016 Jul;13(7):581–3. 55. Veidenberg A, Medlar A, Löytynoja A. Wasabi: An Integrated Platform for Evolutionary Sequence Analysis and Data Visualization. Molecular Biology and Evolution. 2016 Apr 1;33(4):1126–30. 56. McDonald D, Price MN, Goodrich J, Nawrocki EP, DeSantis TZ, Probst A, et al. An improved Greengenes taxonomy with explicit ranks for ecological and evolutionary analyses of bacteria and archaea. ISME J. 2012 Mar;6(3):610–8. 57. Bolyen E, Rideout JR, Dillon MR, Bokulich NA, Abnet CC, Al-Ghalith GA, et al. Reproducible, interactive, scalable and extensible microbiome data science using QIIME 2. Nat Biotechnol. 2019 Aug;37(8):852–7. 58. McMurdie PJ, Holmes S. phyloseq: An R Package for Reproducible Interactive Analysis and Graphics of Microbiome Census Data. PLOS ONE. 2013 Apr 22;8(4):e61217. 59. Lahti L, Shetty S. Microbiome R package [Internet]. microbiome; 2012 [cited 2022 Mar 18]. Available from: https://github.com/microbiome/microbiome/blob/1db2fdb9c1f27e01deb158f72eb 89548bddc710f/inst/CITATION 60. Wickham H, Averick M, Bryan J, Chang W, McGowan LD, François R, et al. Welcome to the Tidyverse. Journal of Open Source Software. 2019 Nov 21;4(43):1686. 61. Blighe K, Rana S, Turkes E, Ostendorf B, Lewis M. EnhancedVolcano: publication-ready volcano plots with enhanced colouring and labeling [internet]. 2021. 62. Wickham H. ggplot2: Elegant Graphics for Data Analysis. Springer; 2016. 266 p. 63. Love MI, Huber W, Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 2014 Dec;15(12):550. 100 64. Dai J, Zheng J, Ou W, Xu W, Ai Q, Zhang W, et al. The effect of dietary cecropin AD on intestinal health, immune response and disease resistance of juvenile turbot (Scophthalmus maximus L.). Fish & Shellfish Immunology. 2020 May 1;100:117–25. 65. National Research Council. Nutrient Requirements of Fish and Shrimp [Internet]. 2011 [cited 2017 May 2]. Available from: https://www.nap.edu/catalog/13039/nutrient-requirements-of-fish-and-shrimp 66. Ohtani M, Villumsen KR, Strøm HK, Lauritsen AH, Aalbæk B, Dalsgaard I, et al. Effects of fish size and route of infection on virulence of a Danish Yersinia ruckeri O1 biotype 2 strain in rainbow trout (Oncorhynchus mykiss). Aquaculture. 2019 Mar 30;503:519–26. 67. Gajardo K, Rodiles A, Kortner TM, Krogdahl Å, Bakke AM, Merrifield DL, et al. A high-resolution map of the gut microbiota in Atlantic salmon (Salmo salar): A basis for comparative gut microbial research. Sci Rep. 2016 Aug 3;6(1):30893. 68. Mohammadian T, Nasirpour M, Tabandeh MR, Heidary AA, Ghanei-Motlagh R, Hosseini SS. Administrations of autochthonous probiotics altered juvenile rainbow trout Oncorhynchus mykiss health status, growth performance and resistance to Lactococcus garvieae, an experimental infection. Fish & Shellfish Immunology. 2019 Mar 1;86:269–79. 69. Good C, Davidson J, Wiens GD, Welch TJ, Summerfelt S. Flavobacterium branchiophilum and F. succinicans associated with bacterial gill disease in rainbow trout Oncorhynchus mykiss (Walbaum) in water recirculation aquaculture systems. J Fish Dis. 2015 Apr;38(4):409–13. 70. Baya AM, Toranzo AE, Lupiani B, Santos Y, Hetrick FM. Serratia marcescens: a potential pathogen for fish. Journal of Fish Diseases. 1992;15(1):15–26. 71. McIntosh D, Austin B. Recovery of an extremely proteolytic form of Serratia liquefaciens as a pathogen of Atlantic salmon, Salmo solar, in Scotland. Journal of Fish Biology. 1990;36(5):765–72. 72. Karami E, Alishahi M, Molayemraftar T, Ghorbanpour M, Tabandeh MR, Mohammadian T. Study of pathogenicity and severity of Lactococcus garvieae isolated from rainbow trout (Oncorhynchus mykiss) farms in Kohkilooieh and Boyerahmad province. Fisheries and Aquatic Sciences. 2019 Oct 1;22(1):21. 73. Çanak Ö, Akayli T, Ürkü Ç. A mixed Bacillus gibsonii and Sphingomonas echinoides infection in cultured rainbow trout (Oncorhynchus mykiss). Marine and Life Sciences. 2021 Dec 31;3(2):71–9. 74. Rasmussen JA, Villumsen KR, von Gersdorff Jørgensen L, Forberg T, Zuo S, Kania PW, et al. Integrative analyses of probiotics, pathogenic infections and 101 host immune response highlight the importance of gut microbiota in understanding disease recovery in rainbow trout (Oncorhynchus mykiss). Journal of Applied Microbiology [Internet]. 2022 Jan [cited 2022 Feb 24];n/a(n/a). Available from: http://onlinelibrary.wiley.com/doi/abs/10.1111/jam.15433 75. Holben WE, Williams P, Saarinen M, Särkilahti LK, Apajalahti JHA. Phylogenetic Analysis of Intestinal Microflora Indicates a Novel Mycoplasma Phylotype in Farmed and Wild Salmon. Microbial Ecology. 2002 Sep 1;44(2):175–85. 76. Llewellyn MS, McGinnity P, Dionne M, Letourneau J, Thonier F, Carvalho GR, et al. The biogeography of the atlantic salmon (Salmo salar) gut microbiome. ISME J. 2016 May;10(5):1280–4. 77. Huyben D, Sun L, Moccia R, Kiessling A, Dicksved J, Lundh T. Dietary live yeast and increased water temperature influence the gut microbiota of rainbow trout. Journal of Applied Microbiology. 2018;124(6):1377–92. 78. Heys C, Cheaib B, Busetti A, Kazlauskaite R, Maier L, Sloan WT, et al. Neutral Processes Dominate Microbial Community Assembly in Atlantic Salmon, Salmo salar. McBain AJ, editor. Appl Environ Microbiol. 2020 Apr;86(8):e02283-19. 79. Rasmussen JA, Villumsen KR, Duchêne DA, Puetz LC, Delmont TO, Sveier H, et al. Genome-resolved metagenomics suggests a mutualistic relationship between Mycoplasma and salmonid hosts. Commun Biol. 2021 May 14;4(1):1– 10. 80. Karlsen C, Ottem KF, Brevik ØJ, Davey M, Sørum H, Winther-Larsen HC. The environmental and host-associated bacterial microbiota of Arctic seawater- farmed Atlantic salmon with ulcerative disorders. Journal of Fish Diseases. 2017;40(11):1645–63. 81. Brown RM, Wiens GD, Salinas I. Analysis of the gut and gill microbiome of resistant and susceptible lines of rainbow trout (Oncorhynchus mykiss). Fish & Shellfish Immunology. 2019 Mar 1;86:497–506. 82. Lavoie C, Courcelle M, Redivo B, Derome N. Structural and compositional mismatch between captive and wild Atlantic salmon (Salmo salar) parrs’ gut microbiota highlights the relevance of integrating molecular ecology for management and conservation methods. Evolutionary Applications. 2018;11(9):1671–85. 83. Zhang X, Ding L, Yu Y, Kong W, Yin Y, Huang Z, et al. The Change of Teleost Skin Commensal Microbiota Is Associated With Skin Mucosal Transcriptomic Responses During Parasitic Infection by Ichthyophthirius multifillis. Frontiers in Immunology [Internet]. 2018 [cited 2022 Mar 16];9. Available from: https://www.frontiersin.org/article/10.3389/fimmu.2018.02972 102 84. Webster TMU, Consuegra S, Hitchings M, Leaniz CG de. Interpopulation Variation in the Atlantic Salmon Microbiome Reflects Environmental and Genetic Diversity. Applied and Environmental Microbiology [Internet]. 2018 Jun 18 [cited 2022 Feb 24]; Available from: https://journals.asm.org/doi/abs/10.1128/AEM.00691-18 85. Wong S, Waldrop T, Summerfelt S, Davidson J, Barrows F, Kenney PB, et al. Aquacultured Rainbow Trout (Oncorhynchus mykiss) Possess a Large Core Intestinal Microbiota That Is Resistant to Variation in Diet and Rearing Density. Applied and Environmental Microbiology [Internet]. 2013 Aug 15 [cited 2022 Jan 14]; Available from: https://journals.asm.org/doi/abs/10.1128/AEM.00924- 13 103 CHAPTER 4 Use of an immune-overexpressing insect line to investigate the effect of antimicrobial peptides (AMP) on fish health and microbiome Sibinga, N.A., Lee, M.T., Buchon, N., Johnson, E.L., Selvaraj, V., Marquis, H. (In preparation). Use of an immune-overexpressing insect line to investigate the effect of antimicrobial peptides (AMP) on fish health and microbiome. 104 ABSTRACT Insect meals are an increasingly common ingredient in pet food and animal feeds. Several studies have reported increased resistance to disease in animals fed insect-based diets, but the mechanisms of this effect are poorly understood and not well reproduced in the literature. We hypothesized that antimicrobial peptides (AMPs) naturally present in insects could explain both the observed protective effects and the variability of these outcomes. AMPs are part of the insect immune response, and as such their levels differ dramatically depending on the state of immune activation. Transgenic fruit flies (Drosophila melanogaster) lines were used to model high and low states of immune activation and then processed these insects into aquaculture feeds for juvenile rainbow trout (Oncorhynchus mykiss). No significant protective effects of the high AMP diet were observed following an immersion challenge with Yersinia ruckeri. However, there were trends towards higher survival and lower Y. ruckeri load in the intestines of fish fed the high AMP diet. In challenged fish, microbiomes of fish differed between diet groups: microbiomes of fish fed the low AMP diet had greater variability, as evidenced by higher within-group beta-diversity at day 8 and day 30 post-challenge. There are multiple limitations of this study, largely owing to technical challenges of rearing sufficient numbers of fruit flies. These are discussed, along with perspectives on how to leverage the wealth of knowledge and experimental tools present in the Drosophila community for research in commonly farmed insects (e.g. Hermetia illucens and Tenebrio molitor). 105 INTRODUCTION Farmed insects have been highlighted as a promising alternative protein source to help meet the nutritional needs of a growing global population (1). In addition to having excellent nutritional profiles, insect ingredients compare favorably to existing protein sources in terms of environmental impact; when compared to conventional livestock, insects have lower requirements for space, food, and water, as well as reduced emissions of greenhouse gases (GHG) and ammonia (2). Insect products have potential for human food applications (3), but in recent years focus has primarily been on their use in animal feeds (4,5). There has been particular interest in the use of insect ingredients in aquaculture feeds, where the levels of dietary protein are generally high and many aquatic species have limited tolerance for plant-derived protein ingredients (6). Current evidence suggests that at moderate inclusion levels (up to ~30% of the diet), insect meals perform well in fish (7). An intriguing finding of some studies of insect-based diets in aquaculture settings has been the observation of increased resistance to disease. These effects do not appear to be consistently repeatable but they have been reported across a variety of different labs, hosts, and pathogens – for a review, see (8). The mechanisms by which any diet, let alone an insect-based diet, can improve disease resistance remain an open topic of study (9,10). In aquaculture, multiple mechanisms appear to be supported by experimental data including effects related to modification of the microbiome (11,12), and regulation of immune effectors (13–15). A diverse set of bioactive components of insect ingredients have been identified and tested in fish diets in purified form: chitin and chitosan (16), lauric acid (17), novel polysaccharides (18), and antimicrobial 106 peptides (19) have all been shown to have the ability to improve fish resistance to disease. While these findings represent meaningful progress, there remains a gap in our understanding of how these components function in the context of non-purified insect ingredients (e.g., insect protein meals). In some notable cases, crude insect ingredients have shown protective effects against fish diseases (20–22), but we currently lack sufficient understanding to predict how and when crude insect ingredients will affect immune health. Other authors have commented on the apparent inconsistency of results between studies of insect-based aquaculture feeds; this is usually attributed to variability between insect ingredients from different sources (23–26). Differences in processing can cause meaningful differences in nutritional composition and digestibility for insect meals derived from the same species (27,28). Furthermore, insect larvae show a great deal of plasticity in their nutrient profile based on rearing conditions and feed substrate – particularly with regard to the lipid fraction (8,29–31). In addition to differences in nutritional composition, we hypothesized that variability in the levels of some or all of the bioactive components listed above could represent an important axis of variation between insect meals. We focused on antimicrobial peptides (AMPs), which have previously been shown to be differentially expressed between black soldier fly (Hermetia illucens) larvae reared on different feed substrates (32). AMPs are small cationic peptides (10-50 amino acids) that are a key part of the innate immune response in multicellular organisms. Insect AMPs have been studied in the context of immunology (33,34), drug development (35,36), and, increasingly, as 107 feed additives (37). Several recent studies have examined the use of purified AMPs to improve fish resistance to disease (38–41), though only a subset have used insect AMPs (19,42,43). Some (though not all) of these studies have reported protective effects, but it is hard to extrapolate results from individual purified AMPs to the use of crude insect meals. For one thing, supplementation of feed with a single AMP is an imperfect model: insects produce many different AMPs and these are known to act synergistically (44,45). Another consideration is that it is difficult to isolate the effect of AMPs in many studies of insect-based diets, since those diets inevitably also contain chitin and lauric acid, among other potentially bioactive compounds. To elucidate the effects of AMPs in the context of insect ingredients, we therefore generated two insect meals differing in their level of AMPs. To do this, we employed a genetic model of immune regulation using fruit flies (Drosophila melanogaster). In fruit flies, activation of the AMP response is stimulated by exposure to pathogen-associated molecular patterns via one of two major pathways: the immune deficiency (Imd) pathway responds primarily to gram-negative bacteria and the Toll pathway to gram-positive bacteria, yeast, and fungi (46)(Fig. 4.1). These pathways are highly conserved, both between insects and vertebrates (47), and among commonly farmed insect species including mealworms (Tenebrio molitor) (48), house flies (Musca domestica) (49), and silkworms (Bombyx mori) (50). We manipulated the Imd pathway to generate fruit flies with either high or low expression of AMPs independent of their exposure to pathogens. These fruit flies were processed into aquaculture diets and fed to fish for one month to evaluate the effect of AMPs on the fish microbiome. In a subsequent experiment, fish were fed the experimental diets and 108 then subjected to a challenge with the gram-negative fish pathogen Yersinia ruckeri. In the absence of infection challenge, no differences between diet groups were observed. In challenged fish, there was not a statistically significant protective effect of the high AMP insect diet, though the trend was towards higher survival, lower pathogen load, and a more stable microbiome in these fish. Figure 4.1. Summary of the major insect immune pathways. Antimicrobial peptides (AMPs) are a major component of the insect immune response and are primarily induced by one of two key pathways. A) In the Imd pathway, peptidoglycan recognition proteins (PRGP) are activated by gram-negative bacteria and in turn activate Imd. This leads to a signal cascade that culminates with the NF-κB transcription factor Relish stimulating expression of AMPs. B) The Toll pathway responds to gram-positive bacteria, yeast, and fungi via proteolytic release of the cytokine Spätzle, which binds to the Toll receptor. The Toll signal cascade results in translocation of Dif (another NF-κB transcription factor) to the nucleus and transcription of AMPs. Individual AMPs may be activated by either or both pathways, but the overall suite of induced AMPs is markedly different depending on the pathogenic stimulus. 109 MATERIALS & METHODS Drosophila care and processing Two lines of transgenic fruit flies (D. melanogaster) were reared in plastic vials on standard fly medium (sucrose, cornmeal, yeast, agar) at room temperature (~22°C). One line, Rel-, was selected to have low expression of antimicrobial peptides (AMP). Rel- is an extensively-studied immune-deficient line that lacks the gene encoding the transcription factor Relish – essential for induction of the humoral immune response via the Imd pathway (51). The second line, HS_imd, was selected to have high expression of AMPs. HS_imd utilizes a heat-shock/GAL4 promoter system. Due to technical limitations, heat shocking the large numbers of flies required for this study proved impractical. However, both the initial experiments using this line (52) and our own preliminary experiments (Fig. S4.1) demonstrate that this is a leaky promoter system that results in constitutive expression of AMPs regulated by the Imd pathway even under stable room temperature conditions. Adult flies from each line were collected once per week over the course of three months directly into liquid nitrogen and stored at -80°C. At the end of the collection period flies were lyophilized and ground into meal using an electric burr grinder. 110 Fig S1FFigig S S11 100 11000 CecropinC CCeeccroroppinin C C 75 Drosomy7755 D cDrionrossoommyyccinin 50 5500 25 2255 5 55 0 00 Rel- HRRSe_imd HS_imd HS_imd HSnoel- lh- s HHSSh__simi m1dxdHHSSh__simi m2dxdHHfSoS_r_i _mimimdddHHSS_R_imiemld-d RReel-l- nnoo h hss hhss 1 1xx hhss 2 d2xixetsfofforo rdr d d(no hs) ieie itesttssfofor rd dieietsts (n(noo h hss)) DrosopDDhriroloass oloipnpheh ialialna ld ilni nteree a aantnmdd et rtnreetaatmtmeenntt Figure S4.1. Expression levels of the AMPs cecropin c and drosomycin by Drosophila line and heat-shock treatment. Preliminary experiments determined that expression of cecropin c, which is regulated by the Imd pathway, was elevated in the HS_imd line relative to the Rel- line when flies were maintained at room temperature without heat shock (hs). A one-hour heat shock of adult flies at 37°C, followed by 24 hour incubation at room temperature resulted in even higher expression of cecropin c for the HS_imd line. A second heat shock 24 hours after the first decreased fly survival and did not result in higher overall expression levels than a single heat shock. Expression levels of drosomycin, which is regulated by the Toll pathway, were slightly upregulated in heat-shocked HS_imd flies relative to the Rel- line but did not differ between lines for flies maintained at room temperature. RNA extraction from Drosophila, reverse transcription, and qPCR Preliminary experiments were conducted to confirm upregulation of AMPs in the high AMP Drosophila line. Three groups of triplicate vials of newly emerged HS_imd flies were maintained at room temperature. After one week, one of the groups Expression fold-difference EExxpprreessssiioonn ffoolldd--ddiiffffeerreennccee 111 was heat-shocked by submerging the vials in a 37°C water bath for one hour. The following day, this group and one of the naïve groups were heat shocked. On the third day, 10 flies from each vial of each of the three groups (naïve, heat-shocked 1x, heat- shocked 2x), as well as an additional group comprised of age-matched Rel- flies, were collected for extraction of RNA using Trizol reagent (Invitrogen). Flies were collected directly into Trizol reagent (Invitrogen) on ice and homogenized for 30 seconds on a bead beater with 5-10 1.3-mm chrome steel beads. Homogenates were incubated at 4°C overnight before RNA extraction. RNA was extracted as per manufacturer instructions with the following modification: prior to precipitation of RNA, the aqueous phase containing nucleic acids was re-extracted with Trizol reagent once. RNA and DNA levels were quantified using a Qubit fluorometer (Thermo Fisher Scientific) to verify extract quality and purity. cDNA libraries were then generated for each sample using the SuperScript IV First-Strand Synthesis System (Invitrogen). Multiplex quantitative polymerase chain reaction (qPCR) was performed on a 7500 real-time PCR instrument (Applied Biosystems) and used to compare expression levels of representative AMP genes (see table S4.1 for primer sequences). The Drosophila cecropin c gene is regulated by the Imd pathway, while drosomycin is regulated by the Toll pathway. Expression of both genes was normalized by comparison to the Rpl32 housekeeping gene (53) and the expression levels in Rel- flies using the double delta Ct method (54). To confirm the stability of AMP gene expression over multiple generations, 10 flies were taken from each of several collection dates spanning the entirety of the three-month collection period and analyzed as described above. 112 Table S4.1 Primer sequences Target gene Suppliera Primer Sequence (5’ → 3’) Y. ruckeri IDT Forward AACCCAGATGGGATTAGCTAGTAA 16S rRNA Reverse GTTCAGTGCTATTAACACTTAACCC Probe Fam-AGCCACACTGGAACTGAGACACGGTCC-IBFQ D. melanogaster IDT Forward GTGAGAACCTTTTCCAATATGATGCA Drosomycin Reverse CGGCATCGGCCTCGTT Probe CY5-CCAGGACCACCAGCATC-MGB D. melanogaster IDT Forward CACCAGTCGGATCGATATGCT Rpl32 Reverse ACGCACTCTGTTGTCGATACC Probe TAMRA-CCATTTGTGCGACAGCTTAGCATATCG-MGB D. melanogaster TF Primer sequences not publicly available, kit: Dm02151846_gH Cecropin C a IDT = Integrated DNA Technologies, Inc. TF = Thermo Fisher Scientific Diet formulation and analysis Two diets were used for this experiment. These diets incorporated insect meal generated from either low AMP (Rel-) or high AMP (HS_imd) Drosophila lines, and were formulated to meet the published nutritional requirements of juvenile rainbow trout (55). Limitations on the scale of production of Drosophila meal complicated the formulation of these diets: proximate analysis of Drosophila insect meal and all diets was performed by Exact Scientific Services (Ferndale, WA) at the end of the study (Table 4.1), but formulation relied on published values to estimate the nutritional composition of Drosophila adults. Accordingly, due to a lower than anticipated protein content in the transgenic Drosophila, these diets were below the published protein requirement of 39.5% for rainbow trout of this size (Table 4.2). During 113 production and storage of all diets, care was taken to keep ingredients cold to minimize potential degradation of any AMPs present. Dry ingredients were combined on ice, oils were added, then just enough cold distilled water was added to form a dough. This dough was immediately extruded through a pre-chilled single screw extruder directly into liquid nitrogen and resulting pellets were lyophilized. All diets were stored at -20°C until use. Table 4.1. Composition of diets (g kg–1) Ingredients Insect meal 200 Menhaden fishmeala 150 Wheat glutenb 50 Wheat flourc 311 Cod liver oild 110 Soy protein concentratee 134 Soybean oilf 34 Mineral premixg 1 Vitamin premixg 10 a Omega Protein® Special Select Menhaden b Betta Foods, Inc. c King Arthur Flour Company, Inc. d Piping Rock Health Products, LLC e The Sausage Maker, Inc. f C & S Wholesale Grocers, Inc. g Florida Aqua Farms, Inc. Table 4.2. Proximate composition of experimental insect meals and diets (g kg–1) High AMP Low AMP High AMP Low AMP insect meal insect meal insect diet insect diet Dry matter 946 937 895 889 Lipid 166 118 n.t.a 185 Protein 540 540 342 372 Carbohydrates 198 234 n.t.a 280 Ash 42 46 53 52 a Insufficient sample for analysis – not tested 114 Fish Care and Handling All animals were cared for following the Guide for the Care and Use of Laboratory Animals and American Association of Laboratory Animal Science Position Statements; the Cornell University Institutional Care and Use Committee (IACUC) reviewed and approved all experimental protocols. Rainbow trout (Oncorhynchus mykiss) fry from the New York State fish hatchery at Bath, NY were acclimated to the Cornell aquatic facility in a single 700 liter fiberglass tank under flowthrough conditions for three weeks and fed BioClarks Fry 1.2-mm pellets (Bio- Oregon, Longview, WA). In the third week of acclimation water temperature was incrementally increased from 12±1°C to 15±1°C. After acclimation, fish were assigned to experimental treatments in the first trial or retained in a separate tank on BioClarks Fry pellets for use in the second trial. Y. ruckeri challenge and sampling The challenge and mock-challenge conditions in this study are identical to those described in another study that was conducted concurrently (43). Briefly, challenged fish were immersed for one hour in a suspension of Y. ruckeri strain CSF007-82 at a concentration of ~8x108 CFU/ml in 10-L flowthrough tanks with supplemental aeration. Mock-challenged fish were immersed in an equivalent volume of sterile tryptic soy broth (TSB). After one hour, water flow was restored and after 24 hours fish were returned to their original 700-L tanks and maintained on the experimental diets for one month. Challenged fish displayed decreased appetite between days 3 and 14 post-infection; accordingly, daily feeding was reduced for this 115 period from 3% of total estimated bodyweight to 1% of total estimated bodyweight. Sampling occurred at the following time points: immediately prior to challenge/mock- challenge (day 0), and then at 3, 8, and 30 days post-challenge. Sampled fish were euthanized via an overdose of buffered MS-222 prior to aseptic removal of the whole intestine posterior of the pyloric caeca. Intestine samples were flash frozen and stored at -80 °C. In the first trial, fish (average body weight = 2.5g) were assigned to the low AMP or high AMP insect diet and then, three days after introduction of the experimental diets, either challenged with Y. ruckeri (50 fish per diet) or mock- challenged with TSB (28 fish per diet). Challenged fish in this trial sustained very high mortality, resulting in too few samples for informative analysis of the complete time course. Accordingly, the challenge treatment (but not the mock-challenge) was repeated in a second trial with larger fish from the same cohort (average body weight = 4.0g) and larger sample sizes (76 fish for the high AMP diet and 71 fish for the low AMP diet). Two weeks post-challenge in the second trial, it became apparent that the mortality rate was lower than that observed in the first trial. In order to ensure that there would be sufficient feed to complete the 30-day time course, fish populations were therefore culled to 10 fish per diet group on day 16. All culled and sampled fish were censored from the data at the appropriate time points for survival analysis. DNA extraction from fish intestines Extraction of DNA from trout intestines was performed as previously described (56). Whole intestines (and contents) were weighed prior to addition of 116 DNA extraction buffer (200 mM NaCl, 200 mM Tris– HCl pH 7.5, 20 mM EDTA, 5% SDS). Subsequently, samples were homogenized on a Mini-Beadbeater (BioSpec Products, Inc. Bartlesville, OK) with 1.3-mm chrome steel beads. A secondary homogenization step with 0.1-mm zirconia/silica beads was performed on a subsample of the initial homogenate corresponding to 20mg of intestinal tissue and contents. DNA was extracted from the secondary homogenate via phenol:chloroform extraction and isopropanol precipitation. qPCR detection and quantification of the Y. ruckeri 16S rRNA gene To determine the pathogen load of challenged fish over time, we used a highly sensitive qPCR assay to detect the Y. ruckeri 16S rRNA gene in intestinal DNA extracts of infected fish from the second trial. Primer and probe design followed the work of Ghosh and colleagues (57) (Table S4.1), and targeted the V2-V3 region of the 16S gene. Triplicate 10μl reactions were performed in 96-well plates using the following reaction mix: 2X PrimeTime Gene Expression Master Mix (Integrated DNA Technologies, Coralville, IA), forward and reverse primers (400nM each), TaqMan probe (100nM), DNA template (1μl), and ultrapure water and plates were run on a 7500 real-time PCR instrument (Applied Biosystems). Samples with amplification of the 16S gene in at least two of three replicate qPCR wells were deemed positive for Y. ruckeri. A standard curve generated using genomic DNA from a pure culture of Y. ruckeri was used to estimate bacterial loads as previously described (56). The limit of quantitation (LOQ) was determined using a published R script (58); the LOQ is the lowest threshold count (Ct) value for which the coefficient of variance – i.e. the 117 standard deviation expressed as a percentage of the mean (59) – is less than 35% based on the results of the standard curve. PCR amplification of V4 region of 16S gene for DNA sequencing In preparation for sequencing, DNA extracts were diluted 1:10 in ultrapure water and amplified via PCR using the 515F and GoLay-barcoded 806R universal 16S primers, which target the V4 region of the 16S gene (60). Amplification (3 minutes 94°C; 30 cycles of 45 seconds 94°C, 60 seconds 50°C, 90 seconds 72°C; 10 minutes 72°C) was performed in duplicate, after which duplicates were pooled. The Mag- Bind® RxnPure Plus kit (Omega Bio-tek, Inc., GA) was used to purify amplicons before 150ng of each amplicon sample were pooled for sequencing. Paired-end sequencing (2x250bp) was performed by the Genomic Facility at Cornell Institute of Biotechnology on an Illumina MiSeq sequencer. Analysis of 16S rRNA gene sequences The QIIME 2 (v 2020.2) pipeline was used for analysis and organization of raw sequence data (61). Analysis of sequence quality and assignment of amplicon sequence variants (ASVs) was performed with the Divisive Amplicon Denoising Algorithm 2 (DADA2) (62); samples with greater than 50% chimeric reads were excluded from further analysis. ASVs were mapped to the Greengenes 16S rRNA database for assignment of taxonomic identities (63). ASV_3428 was identified as Y. ruckeri in a previous study (43); briefly, ASV_3428 relative abundance was strongly correlated with qPCR detection and quantification data and ASV_3428 displays 118 perfect identity to the published 16S gene sequence of the Y. ruckeri strain used in this study. The diversity plugin for QIIME2 was used to calculate Bray-Curtis distances between all samples for analysis of beta-diversity (64). Rstudio (v 1.0.136) was used for subsequent analysis of QIIME2 output data with the following packages: phyloseq (65), microbiome (66), tidyverse (67), EnhancedVolcano (Blighe et al., 2021), and ggplot2 (68). Statistical analysis Gene expression data was analyzed using Quantstudio Design and Analysis software (v 2.6.0). Fish survival data was analyzed using a Kaplan-Meier function and Gehan-Breslow-Wilcoxon test in JMP Pro 14. Y. ruckeri detection data was analyzed with Fisher’s exact test; quantification data was analyzed by analysis of variance (ANOVA) including only samples above the LOQ. Sequence data was analyzed using Rstudio and QIIME 2. Shannon index values were analyzed by ANOVA with post-hoc pairwise Tukey tests. Bray-Curtis distances for group and individual pairwise comparisons were evaluated using permutational multivariate analysis of variance (PERMANOVA) with 999 permutations and Benjamini/Hochberg FDR p-value correction. DESeq2 (Love et al., 2014) was used to identify significant differences in the abundance of individual ASVs between groups. 119 RESULTS Expression of AMPs was upregulated in the high AMP Drosophila line Preliminary experiments demonstrated that untreated flies from the high AMP Drosophila line (HS_imd) had approximately 24-fold upregulated expression of cecropin c compared to the low AMP line (Rel-) (Fig. 4.2). Heat-shocking the high AMP line induced an even stronger response, with 105-fold upregulation of cecropin c (Fig. S4.1), though due to technical limitations we were not able to use heat-shocked flies for production of insect meal. Drosomycin, an AMP regulated by the Toll pathway, was upregulated 2.7-fold in heat shocked high AMP flies but was not upregulated in untreated flies (Fig. S4.1). Subsampling of high AMP flies from different harvest dates revealed that the observed gene expression levels were stable across generations, and that the mean expression level of cecropin c across all collection dates sampled was 23-fold upregulated relative to the low AMP line over the same period (Fig. S4.2). 120 30 cecropin c drosomycin 20 10 0 Low AMP High AMP Drosophila line Figure 4.2. Difference in expression levels of two AMP genes between the low and high AMP fly lines. Transgenic Drosophila lines differed in the Imd pathway, a major regulator of AMP expression. The low AMP line has a non-functional Imd pathway, while the high AMP line expresses Imd constitutively. We collected adult flies from large colonies of each line for a period of three months. Shown above are mean gene expression levels of flies from each line over the collection period for cecropin c, a representative AMP regulated by the Imd pathway. For comparison, a representative AMP under control of the Toll pathway, drosomycin, is also shown. Expression levels of cecropin c, but not drosomycin, were elevated in the high AMP line throughout the collection period. Expression fold-difference 121 60 Low AMP line High AMP line 40 20 0 18 30 50 67 87 17 29 43 61 78 Day of Collection Figure S4.2: Expression levels of cecropin c over the collection period. Adult Drosophila from the Low AMP (Rel-) and High AMP (HS_imd) lines maintained at room temperature were collected weekly for three months and stored at -80°C before being processed into insect meals. At the end of the collection period, 10 flies were randomly selected from each of several collection dates to verify that expression levels of cecropin c remained stable across generations of flies. The high AMP diet did not significantly protect against Y. ruckeri challenge A first infection challenge trial was performed concurrently with the mock- challenge trial, however survival rate in this experiment was very low in both the high AMP (1 fish, 6%) and low AMP (2 fish, 7%) diet groups (data not shown). This left insufficient samples for analysis of the final time point, therefore this trial was excluded from further analysis. The challenge experiment was repeated at the conclusion of the first trial using larger sample sizes of trout from the same cohort. These fish were one month older and therefore also larger than the fish in the first trial. Cecropin C relative expression 122 In the second challenge experiment, the survival rates were 48% and 27% for fish fed the high AMP and low AMP diets, respectively (Fig. 4.3). The trend towards higher survival rate of fish fed the high AMP diet in the second trial was not statistically significant (p=0.26). Mock-Challenged 100 High AMP Low AMP Challenged 50 High AMP Low AMP 0 0 5 10 15 20 25 30 Days Post-Challenge Figure 4.3. Kaplan-Meier survival curve for fish fed either the high AMP insect diet or the low AMP insect diet following immersion challenge with Y. ruckeri. The difference in survival between the diet groups was not statistically significant (p=0.26). No mortality was observed in mock-challenged fish fed the experimental diets. High AMP diet did not significantly reduce Y. ruckeri prevalence or abundance To assess whether AMP levels in the insect diets might help to protect fish from colonization by Y. ruckeri, we used qPCR to detect the Y. ruckeri 16S rRNA gene in intestinal DNA extracts. All fish from both diet groups were positive for Y. ruckeri at the day 3 and day 8 time points (Table 4.3). At day 30, 100% (6/6) of % Survival 123 surviving fish from the low AMP insect diet were positive while 70% (7/10) fish from the high AMP diet were positive; this difference was not statistically significant (p=0.25). We next used the qPCR results to estimate the Y. ruckeri load in intestines of challenged fish (Fig. 4.4). The LOQ in this study was 13.86 gene copies, which corresponds to ~198 bacteria per 20mg sample of intestine; no estimate of Y. ruckeri load was generated for samples with detectable loads below this threshold. At day 3, for samples above the LOQ, the median load of Y. ruckeri was ~1.1x104 cells/20mg intestine in the low AMP insect diet, compared to ~7.9 x 102 cells/20mg intestine suggesting potentially increased colonization of the intestine during the early stage of infection in these fish (Fig. 4.4), though this difference was again not statistically significant (p=0.25). This pattern was less pronounced at day 8 and day 30, though the highest observed Y. ruckeri loads for each of these time points were also from fish fed the low AMP diet. Table 4.3. Detection of the Y. ruckeri 16S gene by qPCR in intestinal DNA extracts Time post-infection Diet group 3 days 8 days 30 days Low-AMP insect diet 8/8 (100%)ab 8/8 (100%) 6/6 (100%) High-AMP insect diet 8/8 (100%) 8/8 (100%) 7/10 (70%) a Fish positive for Y. ruckeri by qPCR over total number of sampled fish with % in parentheses b Comparisons between diet groups were performed separately for each time point using Fisher’s exact test. No statistically significant differences were found between diet groups for any time point. 124 106 105 Low AMP 104 High AMP 103 LOQ 3 8 30 Days Post-Challenge Figure 4.4. Estimated Y. ruckeri burden per 20mg of intestinal tissue. The number of Y. ruckeri cells in each intestinal sample was estimated by qPCR using a standard curve. No estimate was made for samples below the limit of quantitation (LOQ) – samples that were positive for Y. ruckeri but below the LOQ are displayed in the gray box. An ANOVA test of all samples above the LOQ revealed no significant effect of diet on Y. ruckeri load (p=0.249). The high AMP diet did not affect the microbiomes of mock-challenged fish To understand how AMPs impacted commensal bacteria in the absence of infection, we compared the microbiomes of mock-challenged fish fed either the low AMP or high AMP insect diet (Fig. 4.5A). The internal (alpha) diversity of the microbiomes did not change over the course of the experiment for fish fed the high AMP diet (Fig. 4.5B). In fish fed the low AMP diet, however, there was a significant Y. ruckeri load (cells) 125 increase in alpha diversity between the day 0 time point (corresponding to 3 days after introduction of the experimental diet) and all subsequent sampling dates (Fig. 4.5B). Multivariate comparison of between-sample (beta) diversity using PERMANOVA showed no significant effect of diet or timepoint on the microbiomes of uninfected fish. DESeq2 analysis revealed a single ASV significantly enriched over the 30 day feeding period in each diet group; both of these strains were mapped to the family Lactobacillaceae. 126 Figure 4.5. Microbiomes of mock-challenged fish fed either the low AMP or high AMP insect diet. A) Relative abundances of the five most prevalent ASVs in fish intestinal DNA samples over time. The microbiomes of mock-challenged fish were broadly similar between diet groups and over time. Each diet group had a unique member of the Lactobacillaceae family; these strains were likely part of the microbiomes of the flies used to generate the diets. B) Internal (alpha) diversity of each diet group is shown over time, as measured by Shannon index. Data for each diet group was analyzed separately by ANOVA, with post-hoc Tukey test: groups with different letters have statistically significant different mean values. 127 AMP levels influenced intestinal microbiomes in fish challenged with Y. ruckeri Immersion challenge with Y. ruckeri. resulted in pronounced shifts in the microbiome over the course of infection and recovery (Fig. 4.6A). Prior to challenge, microbiomes tended to display high levels of Mycoplasma sp. After the challenge, Mycoplasma sp. relative abundance dropped in the high AMP diet group and was largely replaced by Deefgea sp. and, to a lesser extent, an ASV from the class Betaproteobacteria. In the low AMP diet, Mycoplasma sp. levels remained high at day 3, began to decline somewhat by day 8, and were low at day 30. At days 3, 8, and 30, microbiomes of challenged fish were significantly different between the diet groups (PERMANOVA, adjusted-p<.05). The microbiomes of fish fed the low AMP diet were more variable, with higher intragroup beta-diversity at day 8 (mean Bray distance: 0.63) and day 30 (0.75) than fish fed the high AMP diet at the corresponding timepoints (mean Bray distances: 0.32 and 0.41, respectively). Consistent with this finding, a greater number of different ASVs had high relative abundances in the microbiomes of fish in the low AMP diet group. These included Serratia sp., the Lactobacillaceae sp. associated with this diet group, and Y. ruckeri, Despite these shifts in microbial composition, alpha diversity levels were not significantly different between timepoints for either diet group (Fig. 4.6B). In addition to the Lactobacillaceae strains identified in mock-challenged fish, two additional enriched ASVs were identified by DESeq2 – both in the low AMP diet group: Rhodobacter sp., and a member of the family Enterobacteriaceae. A seemingly disproportionate number of low-quality sequencing outputs (greater than 50% chimeric reads) occurred in the first three timepoints of the high AMP diet and were discarded (Table S4.2); it is 128 unknown whether this is a random effect, an issue with sample processing, an artifact of the treatment, or misclassification of a non-chimeric read. Figure 4.6. Microbiomes of fish challenged with Y. ruckeri and fed either the low AMP or high AMP insect diet. A) Relative abundances of the eight most prevalent ASVs over time by diet group. The highest relative abundances of Y. ruckeri occurred in fish fed the Low AMP diet. Microbiomes shifted over time in both diet groups, with Mycoplasma sp. being replaced as the most abundant ASV in most fish by day 30. Microbiomes were significantly different between diet groups at Day 3, Day 8, and Day 30 by PERMANOVA (adjusted p<.05). B) Internal (alpha) diversity of each diet group is shown over time, as measured by Shannon index. Data for each diet group was analyzed separately by ANOVA: no significant differences over time were detected for either diet group. 129 Table S4.2. Samples sequenced and retained after quality control Mock-Challenge treatment Challenge treatment Samples Diet group Samples sequenced Samples sequenced Samples retained retaineda Low AMP 28 23 30 26 High AMP 28 25 34 23 a Sequenced samples with greater than 50% chimeric reads were removed on the basis of poor quality. Low-quality samples occurred in all treatments but were especially common in samples from challenged fish fed the high AMP diet: within this group, omitted samples were evenly distributed over the first three timepoints. Discussion This study aimed to evaluate the effects of insect-based diets with different levels of antimicrobial peptides (AMPs) on fish subjected to a bacterial infection challenge. We utilized a genetic model of immune regulation to vary the transcription of AMPs in Drosophila and then produced feeds using insect meal made from these flies. In this study, the primary focus was on detection of direct antimicrobial effects of these diets in vivo. To that end, the presence and abundance of Y. ruckeri in the fish intestine was monitored over time and the microbiomes of fish in both infection challenge and mock-challenge settings were analyzed. In mock-challenged fish, no differences in overall microbiome structure between the high and low AMP diet groups were observed. In challenged fish, however, there were multiple trends suggesting that the high AMP diet group may have been beneficial: higher survival, lower median loads of Y. ruckeri, and a more stable microbiome composition. It should be noted that all of these observations suffer from a lack of statistical power; our results suggest that elevating AMP levels in insect meals is a promising direction 130 for future research, but no firm conclusions should be drawn from this data. In addition to lack of statistical significance, this experiment has several other limitations that will be discussed in the context of the results. Many of the limitations of this study stem from technical challenges of producing Drosophila in sufficient quantities for production of experimental feed. The first point that should be noted is that the Drosophila insect meals were below previously reported protein levels, and the diets were therefore below the published protein requirement of 39.5% for rainbow trout of this size (55). This could explain why the survival rates in the ill-fated first challenge trial were lower than expected for both diet groups; fish from the same cohort challenged at the same time and fed a nutritionally complete diet as part of a separate study had higher survival rates (43). The nutritional profiles of the Drosophila meals were similar to one another, however, and mock-challenged fish fed both diets were apparently healthy. Therefore, we believe that comparison between the diet groups in the context of this study is feasible, if not ideal. In the second challenge trial, there was a trend towards higher survival in fish fed the high AMP insect diet, but this was not statistically significant. Similarly, the high AMP diet appeared to decrease the load of Y. ruckeri in the intestines of challenged fish, but this again was not significant. These are intriguing results but should be repeated with larger sample sizes of fish and nutritionally complete diets. The final key limitation of this study is the presence of confounding variables for analysis of the fish microbiomes. Microbiomes typically shift in response to changes in diet (69), and this effect would potentially be exacerbated by the use of nutritionally deficient diets. These effects, however, should be consistent between the 131 tested diets; of more concern are the confounders of diet-specific effects. We did not use germ-free (axenic) flies and we did not sequence the fly microbiomes. Because the diets were processed without any heating step, it is probable that viable bacteria remained in the feed and established themselves in the fish gut. Since these fly lines came from different source populations and were reared separately, it is also likely that they had different microbiomes. In assessing the changes to fish microbiomes there are therefore two key experimental factors to consider: the microbes introduced directly in the feed from each line of flies and the selection pressure due to AMPs. While better study design would be necessary to conclusively separate these factors, we can make inferences. Lactobacillus species tend to be abundant in the microbiomes of Drosophila reared on cornmeal media (70). Since ASVs 4681 and 3932 are Lactobacillus spp. that are specific to each diet group, we propose that these were dominant microbial taxa of the low AMP and high AMP Drosophila lines, respectively. In the fish, these were the taxa most significantly enriched by each diet, and the only significantly different taxa between diet groups of mock-challenged fish identified by DESeq2. This suggests minimal effects of the AMP levels present in these diets on the fish microbiome in mock-challenged fish. In challenged fish, the microbiomes of the two diet groups were more divergent. At the day 3 and day 8 time points, challenged fish in the low AMP diet group had higher mean relative abundance of ASV 501 (Mycoplasma sp.), which has been proposed as a biomarker of health (71); by day 30, however, relative abundance of Mycoplasma was similar between the groups. Challenged fish in the low AMP diet group may have been more prone to dysbiosis: there were multiple fish in 132 this group with microbiomes displaying high relative abundance of either Y. ruckeri or ASVs from the genus Serratia, members of which have previously been identified as opportunistic pathogens of fish (72,73). This is limited evidence, however, and there remains insufficient knowledge of fish microbiomes to confidently differentiate between healthy and unhealthy microbiomes (12). Overall, although the microbiomes of challenged fish were significantly different between diet groups at day 30, the effect of dietary AMP levels on the fish microbiomes in this study is difficult to characterize. We have previously demonstrated that the inclusion of a purified insect AMP in the diet can alter the rainbow trout microbiome by reducing alpha diversity and selecting against gram-negative bacteria (43). Other studies have shown protective effects of dietary insect AMPs against bacterial infection in piglets (74), broilers (75), and fish (19). It is, however, exceedingly difficult to extrapolate the doses of purified AMPs used in these studies to the levels present in an insect meal. Quantitative measures of AMP levels in processed insect ingredients are lacking; most published comparisons of AMP levels in insects are inferred either from RNA transcription data or from antimicrobial activity of extracts from a relatively small number of individuals insects. This makes it difficult to assess the relevance of the wide range of purified AMP doses present in the animal feed literature. One of the few studies to attempt to bridge this gap compared activity levels of immune-stimulated Drosophila hemolymph to those of purified solutions of cecropin (76). Based on the findings of that study and literature values for size (77,78) and hemolymph content (79) of Drosophila adults, we estimate the total antimicrobial activity of immune-stimulated Drosophila hemolymph to be very approximately equivalent to 120 mg cecropin per 133 kg (dry weight) of flies. At the 20% inclusion level of insect meal present in this study, this would equate to ~24mg/kg of feed – below the levels reported in experimental diets for fish (75-250 mg), but not hopelessly so. It should be emphasized, however, that this is an incredibly rough estimate based on data cobbled together from several studies and only reflects the contributions of hemolymph AMPs. Increased availability and accuracy of proteomic analyses will hopefully provide more accurate information on the total amount of AMPs present in insect meals, though this kind of analysis is not trivial (80). It is likely that the AMP levels present in the high AMP diet in this study were not as high as those of previously reported diets using purified AMPs. Flies from the high AMP (HS_imd) Drosophila line had approximately 24-fold upregulated expression of cecropin c compared to flies from the low AMP (Rel-) Drosophila line. Because of the promoter system in this line, heat-shocking the high AMP flies further increased expression to over 100-fold upregulation. However, incubation at 37°C to induce the heat shock response weakened flies and led many to perish in the softened media; we therefore chose not to use heat-shocked flies. There remains potential for higher levels of AMPs than those achieved in this study by using either a different promoter system or an improved methodology for heat shocking large numbers of flies. Furthermore, regulation at the level of the entire Imd pathway is a relatively blunt instrument: new Drosophila models are capable of manipulating expression levels of individual AMPs (44). This kind of system could allow for more precise testing of AMP hypotheses, as well as for the long-term possibility of tailoring levels 134 of individual AMPs in insect meals to, for example, maximize activity against known pathogens while minimizing effects on commensal bacteria in the gut. The size and collective expertise of the Drosophila research community is a tremendous resource for the study of farmed insect species. The current study hopefully sheds light on some of the experimental approaches unlocked by these tools and perspectives, while also highlighting the challenges of translating science in Drosophila to farmed insects; the small size of Drosophila makes them an especially challenging species for production of experimental feeds. Fortunately, annotated genomes for mealworms (81) and black soldier flies (82) and new CRISPR/Cas9 tools for gene editing in these species (83) mean that many of the tools developed in Drosophila labs should hopefully be adaptable to production insect species in the coming decade. There is ample opportunity for knowledge transfer and collaboration between these fields. 135 References 1. Kim SW, Less JF, Wang L, Yan T, Kiron V, Kaushik SJ, et al. Meeting Global Feed Protein Demand: Challenge, Opportunity, and Strategy. Annu Rev Anim Biosci. 2019 Feb 15;7(1):221–43. 2. Van Huis A, Van Itterbeeck J, Klunder H, Mertens E, Halloran A, Muir G, et al. Edible insects: future prospects for food and feed security. Food and agriculture organization of the United Nations; 2013. 3. Doi H, Gałęcki R, Mulia RN. The merits of entomophagy in the post COVID-19 world. Trends in Food Science & Technology. 2021 Apr 1;110:849–54. 4. Belghit I, Liland NS, Waagbø R, Biancarosa I, Pelusio N, Li Y, et al. Potential of insect-based diets for Atlantic salmon ( Salmo salar ). Aquaculture. 2018 Apr;491:72–81. 5. Józefiak D, Józefiak A, Kierończyk B, Rawski M, Świątkiewicz S, Długosz J, et al. 1. Insects – A Natural Nutrient Source for Poultry – A Review. Annals of Animal Science. 2016 Apr 1;16(2):297–313. 6. Hardy RW, Kaushik SJ. Fish Nutrition. Academic Press; 2021. 924 p. 7. Hua K. A meta-analysis of the effects of replacing fish meals with insect meals on growth performance of fish. Aquaculture. 2021 Jan 15;530:735732. 8. Gasco L, Józefiak A, Henry M. Beyond the protein concept: health aspects of using edible insects on animals. Journal of Insects as Food and Feed. 2021 Aug 13;7(5):715–41. 9. Martin SAM, Król E. Nutrigenomics and immune function in fish: new insights from omics technologies. Developmental & Comparative Immunology. 2017 Oct 1;75:86–98. 10. Wu D, Lewis ED, Pae M, Meydani SN. Nutritional Modulation of Immune Function: Analysis of Evidence, Mechanisms, and Clinical Relevance. Frontiers in Immunology [Internet]. 2019 [cited 2022 Mar 31];9. Available from: https://www.frontiersin.org/article/10.3389/fimmu.2018.03160 11. Infante-Villamil S, Huerlimann R, Jerry DR. Microbiome diversity and dysbiosis in aquaculture. Reviews in Aquaculture. 2021;13(2):1077–96. 12. Legrand TPRA, Wynne JW, Weyrich LS, Oxley APA. A microbial sea of possibilities: current knowledge and prospects for an improved understanding of the fish microbiome. Reviews in Aquaculture. 2020;12(2):1101–34. 136 13. Ali MFZ, Yasin IA, Ohta T, Hashizume A, Ido A, Takahashi T, et al. The silkrose of Bombyx mori effectively prevents vibriosis in penaeid prawns via the activation of innate immunity. Sci Rep. 2018 Jun 11;8(1):8836. 14. Lallès JP. Biology, environmental and nutritional modulation of skin mucus alkaline phosphatase in fish: A review. Fish & Shellfish Immunology. 2019 Jun;89:179–86. 15. Wan-Mohtar WAAQI, Taufek NM, Thiran JP, Rahman JFP, Yerima G, Subramaniam K, et al. Investigations on the use of exopolysaccharide derived from mycelial extract of Ganoderma lucidum as functional feed ingredient for aquaculture-farmed red hybrid Tilapia (Oreochromis sp.). Future Foods. 2021 Jun 1;3:100018. 16. Riaz Rajoka MS, Mehwish HM, Wu Y, Zhao L, Arfat Y, Majeed K, et al. Chitin/chitosan derivatives and their interactions with microorganisms: a comprehensive review and future perspectives. Critical Reviews in Biotechnology. 2020 Apr 2;40(3):365–79. 17. Borrelli L, Varriale L, Dipineto L, Pace A, Menna LF, Fioretti A. Insect Derived Lauric Acid as Promising Alternative Strategy to Antibiotics in the Antimicrobial Resistance Scenario. Frontiers in Microbiology [Internet]. 2021 [cited 2022 Mar 31];12. Available from: https://www.frontiersin.org/article/10.3389/fmicb.2021.620798 18. Ali M f. z., Nakahara S, Otsu Y, Ido A, Miura C, Miura T. Effects of functional polysaccharide from silkworm as an immunostimulant on transcriptional profiling and disease resistance in fish. Journal of Insects as Food and Feed. 2021 Oct 5;1–14. 19. Dai J, Zheng J, Ou W, Xu W, Ai Q, Zhang W, et al. The effect of dietary cecropin AD on intestinal health, immune response and disease resistance of juvenile turbot (Scophthalmus maximus L.). Fish & Shellfish Immunology. 2020 May 1;100:117–25. 20. Ido A, Iwai T, Ito K, Ohta T, Mizushige T, Kishida T, et al. Dietary effects of housefly (Musca domestica) (Diptera: Muscidae) pupae on the growth performance and the resistance against bacterial pathogen in red sea bream (Pagrus major) (Perciformes: Sparidae). Appl Entomol Zool. 2015 May 1;50(2):213–21. 21. Su J, Gong Y, Cao S, Lu F, Han D, Liu H, et al. Effects of dietary Tenebrio molitor meal on the growth performance, immune response and disease resistance of yellow catfish (Pelteobagrus fulvidraco). Fish & Shellfish Immunology. 2017 Oct 1;69:59–66. 137 22. Xiang J, Qin L, Zhao D, Xiong F, Wang G, Zou H, et al. Growth performance, immunity and intestinal microbiota of swamp eel (Monopterus albus) fed a diet supplemented with house fly larvae (Musca domestica). Aquaculture Nutrition. 2020;26(3):693–704. 23. Abdel-Tawwab M, Khalil RH, Metwally AA, Shakweer MS, Khallaf MA, Abdel-Latif HMR. Effects of black soldier fly (Hermetia illucens L.) larvae meal on growth performance, organs-somatic indices, body composition, and hemato- biochemical variables of European sea bass, Dicentrarchus labrax. Aquaculture. 2020 May 30;522:735136. 24. Agbohessou PS, Mandiki SNM, Gougbédji A, Megido RC, Lima LMW, Cornet V, et al. Efficiency of fatty acid-enriched dipteran-based meal on husbandry, digestive activity and immunological responses of Nile tilapia Oreochromis niloticus juveniles. Aquaculture. 2021 Dec 15;545:737193. 25. Alves AP do C, Paulino RR, Pereira RT, da Costa DV, e Rosa PV. Nile tilapia fed insect meal: Growth and innate immune response in different times under lipopolysaccharide challenge. Aquaculture Research. 2021;52(2):529–40. 26. Bruni L, Pastorelli R, Viti C, Gasco L, Parisi G. Characterisation of the intestinal microbial communities of rainbow trout (Oncorhynchus mykiss) fed with Hermetia illucens (black soldier fly) partially defatted larva meal as partial dietary protein source. Aquaculture. 2018 Feb 25;487:56–63. 27. Basto A, Matos E, Valente LMP. Nutritional value of different insect larvae meals as protein sources for European sea bass (Dicentrarchus labrax) juveniles. Aquaculture. 2020 May;521:735085. 28. Oonincx D g. a. b., Finke M d. Nutritional value of insects and ways to manipulate their composition. Journal of Insects as Food and Feed. 2021 Aug 13;7(5):639–59. 29. Liland NS, Biancarosa I, Araujo P, Biemans D, Bruckner CG, Waagbø R, et al. Modulation of nutrient composition of black soldier fly (Hermetia illucens) larvae by feeding seaweed-enriched media. PLOS ONE. 2017 Aug 24;12(8):e0183188. 30. Oonincx DGAB, van Broekhoven S, van Huis A, van Loon JJA. Feed Conversion, Survival and Development, and Composition of Four Insect Species on Diets Composed of Food By-Products. Papadopoulos NT, editor. PLoS ONE. 2015 Dec 23;10(12):e0144601. 31. Scala A, Cammack JA, Salvia R, Scieuzo C, Franco A, Bufo SA, et al. Rearing substrate impacts growth and macronutrient composition of Hermetia illucens (L.) (Diptera: Stratiomyidae) larvae produced at an industrial scale. Sci Rep. 2020 Nov 10;10(1):19448. 138 32. Vogel H, Müller A, Heckel DG, Gutzeit H, Vilcinskas A. Nutritional immunology: Diversification and diet-dependent expression of antimicrobial peptides in the black soldier fly Hermetia illucens. Developmental & Comparative Immunology. 2018 Jan 1;78(Supplement C):141–8. 33. Hanson MA, Lemaitre B. New insights on Drosophila antimicrobial peptide function in host defense and beyond. Current Opinion in Immunology. 2020 Feb 1;62:22–30. 34. Marra A, Hanson MA, Kondo S, Erkosar B, Lemaitre B. Drosophila Antimicrobial Peptides and Lysozymes Regulate Gut Microbiota Composition and Abundance. Ja WW, McFall-Ngai MJ, editors. mBio. 2021 Aug 31;12(4):e00824-21. 35. Manniello MD, Moretta A, Salvia R, Scieuzo C, Lucchetti D, Vogel H, et al. Insect antimicrobial peptides: potential weapons to counteract the antibiotic resistance. Cell Mol Life Sci. 2021 May 1;78(9):4259–82. 36. Mylonakis E, Podsiadlowski L, Muhammed M, Vilcinskas A. Diversity, evolution and medical applications of insect antimicrobial peptides. Phil Trans R Soc B. 2016 May 26;371(1695):20150290. 37. Józefiak A, Engberg R. Insect proteins as a potential source of antimicrobial peptides in livestock production. A review. Journal of Animal and Feed Sciences. 2017 May 15;26(2):87–99. 38. Chettri JK, Mehrdana F, Hansen EB, Ebbensgaard A, Overgaard MT, Lauritsen AH, et al. Antimicrobial peptide CAP18 and its effect on Yersinia ruckeri infections in rainbow trout Oncorhynchus mykiss (Walbaum): comparing administration by injection and oral routes. Journal of Fish Diseases. 2017;40(1):97–104. 39. Hu QY, Wu P, Feng L, Jiang WD, Liu Y, Kuang SY, et al. Antimicrobial peptide Isalo scorpion cytotoxic peptide (IsCT) enhanced growth performance and improved intestinal immune function associated with janus kinases (JAKs)/signal transducers and activators of transcription (STATs) signalling pathways in on- growing grass carp (Ctenopharyngodon idella). Aquaculture. 2021 Jun 30;539:736585. 40. Rashidian G, Moosazadeh Moghaddam M, Mirnejad R, Mohammadi Azad Z. Supplementation of zebrafish (Danio rerio) diet using a short antimicrobial peptide: Evaluation of growth performance, immunomodulatory function, antioxidant activity, and disease resistance. Fish & Shellfish Immunology. 2021 Dec 1;119:42–50. 41. de Sousa EL, Assane IM, Santos-Filho NA, Cilli EM, de Jesus RB, Pilarski F. Haematological, biochemical and immunological biomarkers, antibacterial 139 activity, and survival in Nile tilapia Oreochromis niloticus after treatment using antimicrobial peptide LL-37 against Streptococcus agalactiae. Aquaculture. 2021 Feb 25;533:736181. 42. Lin X, Chen W, Lin S, Luo L. Effects of dietary cecropin on growth, non- specific immunity and disease resistance of tilapia (Oreochromis niloticus × O. aureus). Aquaculture Research. 2015;46(12):2999–3007. 43. Sibinga NA, Lee MT, Johnson EL, Selvaraj V, Marquis H. Longitudinal sampling of the rainbow trout (Oncorhynchus mykiss) microbiome reveals effects of dietary cecropin A and Yersinia ruckeri infection. Frontiers in Marine Science. 2022;((Forthcoming)). 44. Hanson MA, Dostálová A, Ceroni C, Poidevin M, Kondo S, Lemaitre B. Synergy and remarkable specificity of antimicrobial peptides in vivo using a systematic knockout approach. MacPherson AJ, Garrett WS, Hornef M, Hooper LV, editors. eLife. 2019 Feb 26;8:e44341. 45. Marxer M, Vollenweider V, Schmid-Hempel P. Insect antimicrobial peptides act synergistically to inhibit a trypanosome parasite. Philosophical Transactions of the Royal Society B: Biological Sciences. 2016 May 26;371(1695):20150302. 46. Buchon N, Silverman N, Cherry S. Immunity in Drosophila melanogaster--from microbial recognition to whole-organism physiology. Nature Reviews Immunology. 2014 Dec;14(12):796–810. 47. Dushay MS, Eldon ED. Drosophila Immune Responses as Models for Human Immunity. The American Journal of Human Genetics. 1998 Jan;62(1):10–4. 48. Keshavarz M, Jo YH, Patnaik BB, Park KB, Ko HJ, Kim CE, et al. TmRelish is required for regulating the antimicrobial responses to Escherichia coli and Staphylococcus aureus in Tenebrio molitor. Sci Rep. 2020 Mar 6;10(1):4258. 49. Tang T, Li X, Yang X, Yu X, Wang J, Liu F, et al. Transcriptional Response of Musca domestica Larvae to Bacterial Infection. PLOS ONE. 2014 Aug 19;9(8):e104867. 50. Wang Q, Wang J, Ren M, Ma S, Liu X, Chen K, et al. Peptidoglycan recognition protein-S1 acts as a receptor to activate AMP expression through the IMD pathway in the silkworm Bombyx mori. Developmental & Comparative Immunology. 2021 Feb 1;115:103903. 51. Hedengren M, BengtÅsling, Dushay MS, Ando I, Ekengren S, Wihlborg M, et al. Relish, a Central Factor in the Control of Humoral but Not Cellular Immunity in Drosophila. Molecular Cell. 1999 Nov 1;4(5):827–37. 140 52. Georgel P, Naitza S, Kappler C, Ferrandon D, Zachary D, Swimmer C, et al. Drosophila Immune Deficiency (IMD) Is a Death Domain Protein that Activates Antibacterial Defense and Can Promote Apoptosis. Developmental Cell. 2001 Oct 1;1(4):503–14. 53. Ponton F, Chapuis MP, Pernice M, Sword GA, Simpson SJ. Evaluation of potential reference genes for reverse transcription-qPCR studies of physiological responses in Drosophila melanogaster. Journal of Insect Physiology. 2011 Jun 1;57(6):840–50. 54. Livak KJ, Schmittgen TD. Analysis of Relative Gene Expression Data Using Real-Time Quantitative PCR and the 2−ΔΔCT Method. Methods. 2001 Dec;25(4):402–8. 55. National Research Council. Nutrient Requirements of Fish and Shrimp [Internet]. 2011 [cited 2017 May 2]. Available from: https://www.nap.edu/catalog/13039/nutrient-requirements-of-fish-and-shrimp 56. Sibinga NA, Marquis H. Tissue‐specific differences in detection of Yersinia ruckeri carrier status in rainbow trout ( Oncorhynchus mykiss ). Journal of Fish Diseases [Internet]. 2021 Aug 25 [cited 2021 Oct 28]; Available from: https://onlinelibrary.wiley.com/doi/10.1111/jfd.13515 57. Ghosh B, Crosbie PBB, Nowak BF, Bridle AR. A highly sensitive, non-invasive qPCR-based strategy for direct quantification of Yersinia ruckeri in fish faeces. J Fish Dis. 2018 Sep;41(9):1421–8. 58. Merkes CM, Klymus K, Allison M, Goldberg C, Helbing C, Hunter M, et al. Generic qPCR Limit of Detection (LOD) / Limit of Quantification (LOQ) calculator. 2019 [cited 2021 Mar 29]; Available from: https://www.sciencebase.gov/catalog/item/5d36fd3fe4b01d82ce8a6cd9 59. Forootan A, Sjöback R, Björkman J, Sjögreen B, Linz L, Kubista M. Methods to determine limit of detection and limit of quantification in quantitative real-time PCR (qPCR). Biomolecular Detection and Quantification. 2017 Jun 1;12:1–6. 60. Caporaso JG, Lauber CL, Walters WA, Berg-Lyons D, Lozupone CA, Turnbaugh PJ, et al. Global patterns of 16S rRNA diversity at a depth of millions of sequences per sample. Proceedings of the National Academy of Sciences. 2011 Mar 15;108(supplement_1):4516–22. 61. Caporaso JG, Kuczynski J, Stombaugh J, Bittinger K, Bushman FD, Costello EK, et al. QIIME allows analysis of high-throughput community sequencing data. Nat Methods. 2010 May;7(5):335–6. 141 62. Callahan BJ, McMurdie PJ, Rosen MJ, Han AW, Johnson AJA, Holmes SP. DADA2: High-resolution sample inference from Illumina amplicon data. Nat Methods. 2016 Jul;13(7):581–3. 63. McDonald D, Price MN, Goodrich J, Nawrocki EP, DeSantis TZ, Probst A, et al. An improved Greengenes taxonomy with explicit ranks for ecological and evolutionary analyses of bacteria and archaea. ISME J. 2012 Mar;6(3):610–8. 64. Bolyen E, Rideout JR, Dillon MR, Bokulich NA, Abnet CC, Al-Ghalith GA, et al. Reproducible, interactive, scalable and extensible microbiome data science using QIIME 2. Nat Biotechnol. 2019 Aug;37(8):852–7. 65. McMurdie PJ, Holmes S. phyloseq: An R Package for Reproducible Interactive Analysis and Graphics of Microbiome Census Data. PLOS ONE. 2013 Apr 22;8(4):e61217. 66. Lahti L, Shetty S. Microbiome R package [Internet]. microbiome; 2012 [cited 2022 Mar 18]. Available from: https://github.com/microbiome/microbiome/blob/1db2fdb9c1f27e01deb158f72eb 89548bddc710f/inst/CITATION 67. Wickham H, Averick M, Bryan J, Chang W, McGowan LD, François R, et al. Welcome to the Tidyverse. Journal of Open Source Software. 2019 Nov 21;4(43):1686. 68. Wickham H. ggplot2: Elegant Graphics for Data Analysis. Springer; 2016. 266 p. 69. Ringø E, Zhou Z, Vecino J l. g., Wadsworth S, Romero J, Krogdahl Å, et al. Effect of dietary components on the gut microbiota of aquatic animals. A never- ending story? Aquaculture Nutrition. 2016;22(2):219–82. 70. Douglas AE. The Drosophila model for microbiome research. Lab Anim. 2018 Jun;47(6):157–64. 71. Bozzi D, Rasmussen JA, Carøe C, Sveier H, Nordøy K, Gilbert MTP, et al. Salmon gut microbiota correlates with disease infection status: potential for monitoring health in farmed animals. Anim Microbiome. 2021 Apr 20;3:30. 72. Baya AM, Toranzo AE, Lupiani B, Santos Y, Hetrick FM. Serratia marcescens: a potential pathogen for fish. Journal of Fish Diseases. 1992;15(1):15–26. 73. McIntosh D, Austin B. Recovery of an extremely proteolytic form of Serratia liquefaciens as a pathogen of Atlantic salmon, Salmo solar, in Scotland. Journal of Fish Biology. 1990;36(5):765–72. 74. Wu S, Zhang F, Huang Z, Liu H, Xie C, Zhang J, et al. Effects of the antimicrobial peptide cecropin AD on performance and intestinal health in 142 weaned piglets challenged with Escherichia coli. Peptides. 2012 Jun 1;35(2):225–30. 75. Zhou G, Wang J, Zhu X, Wu Y, Gao M, Shen H. Induction of maggot antimicrobial peptides and treatment effect in Salmonella pullorum-infected chickens. J Appl Poult Res. 2014 Sep 1;23(3):376–83. 76. Samakovlis C, Kimbrell DA, Kylsten P, Engström Å, Hultmark D. The immune response in Drosophila: pattern of cecropin expression and biological activity. The EMBO journal. 1990;9(9):2969–76. 77. Church RB, Robertson FW. A biochemical study of the growth of Drosophila melanogaster. Journal of Experimental Zoology. 1966;162(3):337–51. 78. Katz AJ, Young SSY. Selection for high adult body weight in Drosophila populations with different structures. Genetics. 1975 Sep 1;81(1):163–75. 79. MacMillan HA, Hughson BN. A high-throughput method of hemolymph extraction from adult Drosophila without anesthesia. Journal of Insect Physiology. 2014 Apr;63:27–31. 80. Dong L, Ariëns RMC, America AHP, Paul A, Veldkamp T, Mes JJ, et al. Clostridium perfringens suppressing activity in black soldier fly protein preparations. LWT. 2021 Sep 1;149:111806. 81. Eriksson T, Andere A a., Kelstrup H, Emery V j., Picard C j. The yellow mealworm (Tenebrio molitor) genome: a resource for the emerging insects as food and feed industry. Journal of Insects as Food and Feed. 2020 Oct 26;6(5):445–55. 82. Generalovic TN, McCarthy SA, Warren IA, Wood JMD, Torrance J, Sims Y, et al. A high-quality, chromosome-level genome assembly of the Black Soldier Fly (Hermetia illucens L.). G3 Genes|Genomes|Genetics. 2021 May 1;11(5):jkab085. 83. Zhan S, Fang G, Cai M, Kou Z, Xu J, Cao Y, et al. Genomic landscape and genetic manipulation of the black soldier fly Hermetia illucens, a natural waste recycler. Cell Res. 2020 Jan;30(1):50–60. 143 CHAPTER 5 Conclusions and Future Perspectives 144 Conclusion and Future Directions There are inescapable parallels between the trajectory of aquaculture over the past fifty years and the ambitions of farmed insect producers for the next fifty. The relative ignorance about the nutritional requirements of insects, the cavalier use of wild brood stock, the ability (and willingness) to ignore infectious diseases, and the precarious uncertainty of boom or bust production cycles all harken back to the early days of industrial aquaculture. Nowhere in my work was this more apparent than in my study of Y. ruckeri carrier status. Y. ruckeri was first isolated in farmed trout in Idaho in the 1950s, and while there remains some ambiguity, it is likely that it was present in other areas of the world before that (1,2). However, many of the Y. ruckeri strains isolated from outbreaks around the world bear genetic resemblance to those associated with the early days of trout farming in the United States (3), suggesting that transportation of fish carrying the bacteria have contributed significantly to its spread. Now there are signs of waning effectiveness of vaccines against this disease and an expanded host range (4). Clearly, effective surveillance testing before transporting fish – especially before transporting fish between countries – is a key to containing newly virulent pathogen strains. Given the ubiquity of infectious diseases, however, there is also a need for a next generation of prophylactic and therapeutic strategies to maintain fish welfare; functional feeds may offer the best hope for non-antibiotic treatments. Seen through this lens, there is perhaps a sort of incidental prescience in studying components of the insect immune system. It seems inevitable that insect pathogens will become a challenge for insect farmers, though for the moment there is 145 virtually no discussion of this topic in the literature. When insect diseases do become a problem for the industry, the existing literature on immunity and infectious disease in fruit flies and honeybees will be a great resource. However, it is important to recognize that the answers will not all be found in existing literature: in addition to having their own unique evolutionary history and host-pathogen biology, farmed insects are reared under different and likely more stressful conditions. Replicating the chronic stress of high-density animal production systems in the lab remains a challenge for researchers, regardless of species. Future work on regulation of the insect immune system should also address the implications of immunity for the health of farmed insects. Furthermore, proactive identification of insect pathogens and development of surveillance testing methods and treatment strategies before widespread outbreaks occur will hopefully minimize dissemination of regionally specific pathogens. The malleable biology of insects is at the core of their utility as feed ingredients as it allows them to thrive in a wide variety of conditions. However, it also presents significant challenges for industrial use and experimental design. Feed formulators model ingredients as having specific chemical compositions that will be consistent from batch to batch; this saves producers from having to analyze ingredients and adjust formulations every time they make feed. This is an approximation since most ingredients do exhibit variability – there is ample evidence, for example, that FM and FO are variable depending on their source (5), processing (6), and storage (7), and that these differences can have measurable dietary effects on palatability (8), digestibility (9), and physiology (10) – but the model remains useful to 146 the extent that ingredient variability occurs within a tolerable range that can be anticipated and accounted for. In the case of FM and FO, the clear utility of these ingredients has made them worth using despite some level of batch-to-batch variability. Comparison of insect meals from different sources or rearing methods consistently demonstrates differences in composition (11,12). Rearing substrates have also been shown to alter expression of AMPs (13). Rather than ignoring these differences, future studies should seek to carefully document both the production conditions and processing steps of insect meals and, in as much detail as possible, their chemical compositions. While insect ingredients will likely always exhibit a relatively high degree of variability, the goals of the industry should be 1) to understand the drivers of variability in order to produce more consistent ingredient compositions, and 2) to develop a sufficiently useful product that feed producers will tolerate some level of inconsistency. Documenting the range of composition and then striving to understand the drivers of variability within that range is a key step to the practical utility of farmed insects. It is important, however, to recognize that simply minimizing variability may limit the potential utility of insect ingredients: reproducibility of a suboptimal product may be a dead end. There has been relatively little effort made thus far to define the characteristics of high (and low) quality insect meals, though it is clear that differences exist (14). There is ample room for experiments that introduce and compare targeted variations in insect ingredients. To aid in such experiments, there is a wealth of genetic tools available in fruit fly models, and many of these can be employed for 147 controlled study of farmed insects – either directly, as in chapter 4 of this dissertation, or by adaptation to species with direct relevance for insect farming. The entwined futures of aquaculture and insect farming will have a critical part to play in navigating the challenges of the coming decades. In theory, their low demands for space and fresh water should help both to be resilient in the face of climate change. And while there is clearly urgency to start implementing these systems to solve real world problems, it is important not to lose sight of how much research still needs to be done. If there is one lesson that the fish and insect farming industries should take from conventional livestock, it is that understanding the biology underlying nutrition and health is a project that will continue to pay dividends for decades to come. 148 1. Barnes AC, Delamare-Deboutteville J, Gudkovs N, Brosnahan C, Morrison R, Carson J. Whole genome analysis of Yersinia ruckeri isolated over 27 years in Australia and New Zealand reveals geographical endemism over multiple lineages and recent evolution under host selection. Microbial Genomics [Internet]. 2016 Nov 30 [cited 2020 May 4];2(11). Available from: https://www.microbiologyresearch.org/content/journal/mgen/10.1099/mgen.0.00 0095 2. Bullock GL, Stuckey HM, Shotts EB. Enteric redmouth bacterium: comparison of isolates from different geographic areas. J Fish Diseases. 1978 Oct;1(4):351– 6. 3. Bastardo A, Ravelo C, Romalde JL. Phylogeography of Yersinia ruckeri reveals effects of past evolutionary events on the current strain distribution and explains variations in the global transmission of enteric redmouth (ERM) disease. Front Microbiol [Internet]. 2015 Oct 29 [cited 2021 Jan 20];6. Available from: http://journal.frontiersin.org/Article/10.3389/fmicb.2015.01198/abstract 4. Wrobel A, Leo JC, Linke D. Overcoming Fish Defences: The Virulence Factors of Yersinia ruckeri. Genes. 2019 Sep 11;10(9):700. 5. Moffat CF, McGill AS. Variability of the composition of fish oils: significance for the diet. Proc Nutr Soc. 1993 Oct;52(3):441–56. 6. Jasour MS, Wagner L, Sundekilde UK, Larsen BK, Rasmussen HT, Hjermitslev NH, et al. Fishmeal with different levels of biogenic amines in aquafeed: Comparison of feed protein quality, fish growth performance, and metabolism. Aquaculture. 2018 Mar 10;488:80–9. 7. Tapia-Salazar M, Cruz-Suárez LE, Ricque-Marie D, Pike IH, Smith TK, Harris A, et al. Effect of fishmeal made from stale versus fresh herring and of added crystalline biogenic amines on growth and survival of blue shrimp Litopenaeus stylirostris fed practical diets. Aquaculture. 2004 Dec 20;242(1):437–53. 8. Berg LR, Martinson RD. Comparative Free Choice Selection of Various Fish Meal Diets by the Chick. Poultry Science. 1972 Sep 1;51(5):1684–9. 9. Sticker LS, Hembry FG, White TW, Essimg HW, Guthrie LD. In Situ Nitrogen and Dry Matter Disappearance of Soybean Meal, Mexican Anchovy, Maine Herring, and Menhaden Fish Meals1. The Professional Animal Scientist. 1990 Jun 1;6(1):1–4. 10. Ghasemi Fard S, Wang F, Sinclair AJ, Elliott G, Turchini GM. How does high DHA fish oil affect health? A systematic review of evidence. Critical Reviews in Food Science and Nutrition. 2019 Jun 17;59(11):1684–727. 149 11. Gasco L, Biancarosa I, Liland NS. From waste to feed: A review of recent knowledge on insects as producers of protein and fat for animal feeds. Current Opinion in Green and Sustainable Chemistry. 2020 Jun;23:67–79. 12. Scala A, Cammack JA, Salvia R, Scieuzo C, Franco A, Bufo SA, et al. Rearing substrate impacts growth and macronutrient composition of Hermetia illucens (L.) (Diptera: Stratiomyidae) larvae produced at an industrial scale. Sci Rep. 2020 Nov 10;10(1):19448. 13. Vogel H, Müller A, Heckel DG, Gutzeit H, Vilcinskas A. Nutritional immunology: Diversification and diet-dependent expression of antimicrobial peptides in the black soldier fly Hermetia illucens. Developmental & Comparative Immunology. 2018 Jan 1;78(Supplement C):141–8. 14. Basto A, Matos E, Valente LMP. Nutritional value of different insect larvae meals as protein sources for European sea bass (Dicentrarchus labrax) juveniles. Aquaculture. 2020 May;521:735085.