(Perca Flavescens) Diets: Responses of Nutritional Programming on Grow

Home , Perch, Yellow perch

Fish meal replacement with soybean meal in yellow perch (Perca flavescens) diets:

responses of nutritional programming on growth, transcriptome and isoflavone

accumulation

DISSERTATION

Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy

in the Graduate School of The Ohio State University

Megan Marie Kemski

Graduate Program in Food Science and Technology

The Ohio State University

2018

Dissertation Committee

Dr. Macdonald Wick, Advisor

Dr. Konrad Dabrowski, Co-Advisor

Dr. Chad Rappleye

Dr. Yael Vodovotz 1

Copyrighted by

Megan Marie Kemski

2018

Abstract

One of the most economical and sustainable ways to reduce the consumption of wild fish resources, is to replace fish meal protein in aquafeeds with plant-based proteins.

Aquaculturists, however, have found that they are not able to entirely replace fish meal with plant-based proteins in diets due to amino acid profile differences and possible anti- nutritional factors present, which have caused fish to have hindered growth, survival and reproductive quality. Thus, more research needs to be done for plant-based proteins to be fully integrated into fish diets that ultimately allow for optimal fish growth and health.

The current study addresses a significant gap in the literature by focusing on fish meal replacement in yellow perch (Perca flavescens). The central hypothesis of this project is that there is a phenotype-diet interaction that occurs through nutritionally programming fish to a soybean meal-based diet. Nutritional programming is described as early dietary events that occur during critical developmental windows that can result in permanent changes later in life such as, growth potential health and metabolic status. To test this hypothesis, a series of experiments were designed. In the first experiment (Chapter 2), yellow perch were nutritionally programmed over 4 phases; growth performance (growth, survival and specific growth rate (SGR) was measured, and subsequently, the fish were reproduced to determine if soybean meal in the diet affected reproductive quality. In the ii second experiment (Chapter 3), the nutritionally programmed adult yellow perch from the previous study were reproduced annually over the course of three years (2015, 2016 and

2017) to determine if the offspring had improved growth performance when given a fish meal (FM) or soybean meal-based (SBM) diet as their first formulated feed. The aim of this project was to determine if possible parental inheritance of nutritional programming occurred, and if it had an effect on the growth of offspring when they were fed a SBM- based diet. It was determined that parental inheritance and initial offspring diets, independent of one another, had a significant effect on percent weight gain on the progeny. In Chapter 4, a more mechanistic approach was taken, in which RNA- sequencing (RNA-seq) was done to determine transcriptional differences in the mid- intestine of juvenile yellow perch, after being fed either a FM or SBM-based diet as their first formulated feed. In this experiment, the goal was to examine the gene expression differences between juveniles when fed FM and SBM-based diets as their first formulated fed. RNA-seq analysis revealed that nine of the genes up-regulated in the SBM-fed fish were directly involved in the cholesterol biosynthesis pathway, and these fish were found to have similar cholesterol levels compared to the FM-fed fish.

Because higher inclusion levels of soybean meal in aquafeeds is needed for more sustainable aquaculture production, and ultimately as a food product, it was important to also examine yellow perch after being fed a SBM-based diet from the consumer’s perspective. In Chapter 5, the soybean isoflavone (phytoestrogen) content within the soybean meal-based diets was analyzed, along with the possible accumulation in the yellow perch fillets. Studies of isoflavone accumulation are limited in fish, and this was iii the first such study in yellow perch. It is hypothesized that by feeding yellow perch diets with a high concentration of SBM, isoflavone accumulation will occur within the fillet

(muscle tissue) of these fish in a dose dependent manner. This is important from a consumer’s perspective because isoflavones have been shown to play roles as antioxidants, and have exhibited anticancer, and anti-inflammatory activities within mammals. However, they are also structurally similar to estrogen, and can cause agonistic/antagonistic estrogenic effects. Results of this study discovered that isoflavone accumulation occurred within the muscle tissue of yellow perch at low concentrations (ng isoflavones/g muscle tissue), and was significantly effected by the first feed given to fish as juveniles. Interestingly, there was no effect of sex of fish, weight or dose dependent response seen on isoflavone accumulation in the muscle tissue.

Overall, these outcomes of the presented studies contributed to the knowledge of how nutritional programming can be used to raise carnivorous fish on diets with high levels soybean meal-based protein without compromise to survival, growth or reproductive quality. Transcriptomic results revealed specific alterations to the metabolism of these fish after being fed a SBM-based diet, which will allow for better diet formulation and a possible explanation of why depressed growth is seen when these fish are juveniles.

Finally, the findings that isoflavone accumulation was only affected by initial dietary history are important from a consumer’s perspective, as these fish do not seem to have an increase in accumulation after prolonged exposure to soybean meal.

Dedication

To Jim and my family who have supported me,

believed in me, and encouraged me.

Acknowledgments

I have been lucky to work with some amazing people during my time here at The Ohio

State University, and would not have been able to accomplish my degree without the help and support of the following;

To my advisor Dr. Macdonald Wick, words cannot express the gratitude that I have for your guidance, knowledge, patience and support throughout the years. Without your backing and encouragement, this degree would not have been possible. You have helped me grow as a scientist and as a person, and for that, I thank you.

Thank you to Dr. Konrad Dabrowski, for teaching and sharing your limitless knowledge about fish with me. Through your training and patience I have gained so much, and you have made this entire project possible, for which I am truly grateful.

Thank you to Dr. Chad Rappleye, you have provided knowledge, guidance and tremendous encouragement as I grew from a lab tech to a PhD graduate. Without your help, I would have never been able to navigate the RNAseq data, and I am appreciative of your help and support over the years.

Thank you to Dr. Yael Vodovotz, for your support, trust and providing a new and exciting opportunity for me to advance. You brought a unique perspective to my project, and allowed my passion for food science to thrive.

In addition, thank you to Dr. Bruno who has contributed scientific intellect and allowed me to work in his lab, and his post doc, Dr. Priyankar Dey, who spent many hours teaching me and helping me with protocols.

I also could not have made it through my time without my awesome lab members who have also become friends, and are the only other people that really understood the process. To Mackenzie Miller, John Grayson, Kevin Fisher, Dr. Thomas Delomas,

Kristen Towne and the other members of the aquaculture lab past and present, thank you for your continuous help and support. I could not have completed this project without your many hours of assistance, and it always helped to have some humor and fun along the way!

There are also those that have helped me over the years and I am also so grateful for all of your patience and assistance, but also your friendship. Thank you to Dr. Stephanie

Hutsko and Dr. Jackie Griffin, Tim Vasquez, Kelsey Morris, and to the many others that I have had the pleasure to get to know during my time as a graduate student.

Finally, thank you to my family especially my mother, who has been my strength and rock for the past 5 years, supporting me through all of the highs and lows, and never stopped encouraging me along the way. To my father, who provided love and motivation. And to my brother who has provided true friendship and advice, as well as being good distraction from the stress. Above all, thank you to Jim, who has been by my side during this entire journey, supporting me throughout. Thank you for your continuous patience, understanding, encouragement and love. I could not have done it without you!

vii

Vita

2003-2008 B.A. Zoology (Biology), Miami University, Oxford, OH

2008-2011 Research Assistant, Kao Corporation, Cincinnati, OH

2011-2013 Research Associate, Department of Microbiology, The Ohio State

University, Columbus, OH

2013-Present Graduate Research Associate, Department of Food Science and

Technology, The Ohio State University, Columbus, OH

Publications

Kemski, M., Wick, M., Dabrowski, K. (2018). Effect of Nutritional Programming on Growth, Reproduction and Embryonic Development in Yellow Perch (Perca flavescens) fed Soybean Meal-Based Diets. Aquaculture. In press.

Miller, M., Kemski, M., Grayson, J., Towne, K., Dabrowski, K. (2018). Yellow Perch sperm motility, cryopreservation and viability of resulting larvae/juveniles. North American Journal of Aquaculture. 80(1), 3-12.

Kemski, M. M., Stevens, B., & Rappleye, C. A. (2013). Spectrum of T-DNA integrations for insertional mutagenesis of Histoplasma capsulatum. Fungal Biology, 117(1), 41-51

Edwards J.A., Kemski M.M.,Rappleye C.A. (2013). Identification of an aminothiazole with antifungal activity against intracellular Histoplasma capsulatum. Antimicrobial Agents and Chemotherapy, 57(9), 4349-4

viii

Edwards J.A., Kemski M.M., Rappleye C.A., Chen C., Hu J.,Mitchell T.K. (2013). Histoplasma yeast and mycelial transcriptomes reveal pathogenic-phase and lineage-specific gene expression profiles. BMC Genomics, 14(1) 695

Holbrook, E.D1., Kemski, M.M1., Richer, S., Wheat, L., Rappleye, C.A., (2014).Glycosylation and immunoreactivity of the Histoplasma capsulatum Cfp4 yeast-phase exoantigen. Infection and Immunity. 82(10), 4414-4425.

Fields of Study

Major Field: Food Science and Technology

Table of Contents

Abstract ...... ii Dedication ...... v Acknowledgments ...... vi Vita ...... viii List of Tables ...... xiv List of Figures ...... xv Chapter 1: Introduction ...... 1 1.1 Overview ...... 1 1.2 Yellow Perch ...... 2 1.3 Protein replacement in aquafeeds ...... 4 1.4 Nutritional Programming ...... 7 1.5 RNA-sequencing ...... 11 1.6 Isoflavone Accumulation ...... 18 Chapter 2: Nutritional Programming Effects on Growth and Reproduction of Broodstock and Embryonic Development of Progeny in Yellow Perch (Perca flavescens) fed Soybean Meal-Based Diets ...... 27 2.1 Abstract ...... 27 2.2 Introduction ...... 28 2.3 Methods ...... 32 2.3.1 Fish and experimental conditions: Phase 1 ...... 32 2.3.2 Phase 2 ...... 34 2.3.3 Phase 3 ...... 35 2.3.4 Phase 4 ...... 36 2.3.5 Reproduction and egg quality ...... 36 2.3.6 Statistical analysis ...... 38

2.4 Results ...... 39 2.4.1 Phase 1 ...... 39 2.4.2 Phase 2 ...... 40 2.4.3 Phase 3 ...... 40 2.4.4 Phase 4 ...... 41 2.4.5 Reproductive quality ...... 41 2.5 Discussion ...... 43 2.5.1 Phase 1 ...... 44 2.5.2 Phase 2 ...... 44 2.5.3 Phase 3 ...... 45 2.5.4 Phase 4 ...... 47 2.5.5 Reproduction ...... 48 2.6 Conclusion ...... 50 Chapter 3: Inheritance of parental nutritional programming on offspring growth performance in yellow perch (Perca flavescens) when fed soybean meal as first feed ...... 59 3.1 Abstract ...... 59 3.2 Introduction ...... 60 3.3 Materials and Methods ...... 64 3.3.1 Fish and experimental conditions ...... 64 3.3.2 Breeding of fish ...... 65 3.3.3 Rearing of yellow perch larvae ...... 66 3.3.4 Phase 1 ...... 67 3.3.5 Phase 2 ...... 69 3.3.6 Phase 3 ...... 69 3.3.7 Statistical analysis ...... 70 3.4 Results ...... 70 3.4.1 Phase 1 Results ...... 70 3.4.2 Phase 2 Results ...... 72 3.4.3 Phase 3 Results ...... 73 3.5 Discussion ...... 74 3.5.1 Phase 1 ...... 74 3.5.2 Phase 2 ...... 76 xi

3.5.3 Phase 3 ...... 76 3.6 Conclusion ...... 79 Chapter 4: Transcriptomic response to soybean meal-based diets as the first formulated feed in juvenile yellow perch (Perca flavescens) ...... 87 4.1 Abstract ...... 87 4.2 Introduction ...... 88 4.3 Methods ...... 92 4.3.1 Yellow perch feeding trial and sampling ...... 92 4.3.2 RNA Sequencing ...... 93 4.3.3 Library Preparation and De Novo Transcriptome Assembly ...... 94 4.3.4 Differential expression analysis ...... 95 4.3.5 qPCR Validation ...... 96 4.3.6 Cholesterol Analysis ...... 98 4.3.7 Statistical Analysis ...... 98 4.4 Results ...... 99 4.4.1 Feeding Trial ...... 99 4.4.2 Transcriptomic responses ...... 99 4.4.3 Up-regulation of genes found in SBM fed fish ...... 100 4.4.4 Down-regulation of genes found in SBM fed fish ...... 100 4.4.5 qPCR Validation ...... 101 4.4.6 Whole body lipids and cholesterol analysis ...... 102 4.5 Discussion ...... 102 4.5.1 Feeding Trial ...... 103 4.5.2 Transcriptional changes in the mid-intestine ...... 104 4.5.3 Lipid and cholesterol analysis ...... 108 4.6 Conclusions ...... 110 Chapter 5: Isoflavone accumulation in the muscle tissue of yellow perch (Perca flavescens) after being fed soybean meal based diets ...... 118 5.1 Abstract ...... 118 5.2 Introduction ...... 119 5.3 Materials and Methods ...... 123 5.3.1 Fish and experimental conditions ...... 123 5.3.2 Diets ...... 124 xii

5.3.3 Sampling ...... 125 5.3.4 Chemicals and supplies ...... 126 5.4.5 Isoflavone extraction: Soybean meal and diets ...... 126 5.3.6 Isoflavone extraction: Muscle tissue ...... 127 5.3.7 Calibration Curves and Standards ...... 128 5.3.9 Lipid Analysis ...... 129 5.3.10 Statistical analysis ...... 130 5. 4 Results ...... 130 5.4.1 Fish weight and lipid analysis ...... 130 5.4.2 Isoflavone content in SBM and formulated SBM-based diets ...... 131 5.4.3 Isoflavone concentration in muscle tissue ...... 132 5.5 Discussion ...... 133 5.5.1 Fish weight and lipid analysis ...... 133 5.5.2 Isoflavone content in SBM and formulated SBM-based diets ...... 135 5.5.3 Isoflavone concentration in muscle tissue ...... 136 5.6 Conclusions ...... 141 Chapter 6: Conclusions ...... 151 Chapter 2 ...... 151 Chapter 3 ...... 151 Chapter 4 ...... 152 Chapter 5 ...... 153 References ...... 155

xiii

List of Tables

Table 2.1 Composition of experimental diets……………...……………………….……52 Table 2.2 Growth performance parameters measured at the beginning and end of each phase. ……………...…………………………………………………………….53 Table 3.1 Growth performance parameters of offspring groups (2015, 2016 and 2017) for Phase 1.…………...…………………………………………………….…….80 Table 3.2 Growth performance parameters of offspring groups (2015, 2016 and 2017) for Phase 2. …………...………………………………………………………….81 Table 3.3 Growth performance parameters of offspring groups (2015, 2016 and 2017) for Phase 3...……………..……………………………………………………….82 Table 4.1 Dietary ingredients and composition of FM2 and SBM2 diets…………..….111 Table 4.2 Growth performance results of juvenile yellow perch after being fed either a FM2 or SBM2 diet for 61 days...... ………………………………...………112 Table 4.3 RNAseq results of the top differentially expressed genes in the mid intestine…………………………………………………………………………115 Table 4.4 Primers used for Real-Time qPCR expression analysis……………….…….116 Table 5.1 Dietary ingredients and composition of FM2 and SBM2 diets……………...143 Table 5.2: Lipid concentrations (mg/g) in muscle tissue of yellow perch in each group (2013, 2015 and 2016)…………………………………………………...146

xiv

List of Figures

Figure 1.1 A phylogenetic tree representing the major groups of fishes…..…………….24 Figure 1.2 A phylogenetic relationships among the family Percidae………...………….25 Figure 1.3 Structure of isoflavones…………………..…..………...…………………….26 Figure 2.1 Nutritional programming experimental design.…..………………………….54 Figure 2.2 Temperature data throughout the duration of the experiment….…………….54 Figure 2.3 Weight gain (%) results….………………..………………………………….55 Figure 2.4 Relative fecundity among females fed different initial programming diets.…56 Figure 2.5 Reproductive indices compared among females …………………………….57 Figure 2.6 Correlations among female reproductive parameters.……….……………….58 Figure 3.1 Schematic view of the experimental design..…………..…………………….83 Figure 3.2 Weight gain percent after Phase 1………...………………………………….84 Figure 3.3 Weight gain percent after Phase 2.………..………………………………….85 Figure 3.4 Overall weight gain percent after Phase 3…...……………………………….86 Figure 4.1 Weight gain (%) of juvenile yellow perch after being fed either a FM2 or SBM2 diet for 61 days.…..………………………………………………….….112 Figure 4.2 Cholesterol biosynthesis pathway and top differentially expressed genes….113 Figure 4.3 Quantitative real-time PCR confirmation…………………………..……….114 Figure 4.4 Whole body concentration (mg/g tissue) of total lipid and cholesterol levels…………………………………………………………………………....117 Figure 5.1 Flow chart of feeding regime completed by each generation of fish....…….144 Figure 5.2 Fish weight (g) of each group at the time of sampling.…..…………..……..145 Figure 5.3 Total isoflavone concentration (µg/g) present in soybean meal (SBM)

and formulated SBM-based diets…………………………………..…..……….147 Figure 5.4 Profile of various isoflavone forms present in the SBM and SBM2 diet...... 148 Figure 5.5 Total isoflavone concentration (ng/g) in muscle tissue from each group of fish…………………………………………………….……………….…….149 Figure 5.6 Profiles of the three major isoflavones present in the muscle tissue of yellow perch…………………………………………………..………..……….150

xvi

Chapter 1: Introduction

1.1 Overview

In the United States and throughout the world, natural commercial fisheries stocks are threatened. Many commercial stocks are now fully exploited, overexploited, or depleted

(Food and Agriculture Organization of the United Nations, 2016). With declining harvests from capture fisheries, farm-raised aquaculture production will need to increase five-fold from present levels to supply the global demand for seafood (The Ohio

Department of Agriculture, 2010). According to the FAO (2016), fish production has increased from 1 million tons in 1950 to 81 million tons in 2014, making it one of the fastest growing forms of food production in the world. Aquaculture production is required to increase as natural fish stocks continue to be depleted. However, this growth cannot continue at its current rate due to space limitations and restricted feed availability

(Gentry et al., 2017). Another change needed is to develop more commercial aquaculture industries in the United States. In an annual report by the National Oceanic and

Atmospheric Administration, it was reported that the United States is not a significant producer of aquaculture, and imports 91% of its seafood with over half coming from aquaculture, at a cost of around $10.4 billion annually (Voorhees et al., 2016). If the

United States were able to develop and promote more productive and efficient aquaculture facilities to meet the country's demand, it would result in an enormous economic benefit. 1

One of the main reasons that commercial aquaculture facilities throughout the U.S. are struggling is because of the high price of aquafeeds. In order for aquaculture production to have continued growth and be more successful is to reduce the levels of fish meal in diet formulations. Research has shown that one of the most economical and sustainable ways to reduce the consumption of wild fish resources is to replace fish meal protein in aquaculture fish diets with plant-based protein (Klinger and Naylor, 2012). Therefore significant improvements to the industry are necessary to incorporate plant-based protein sources into fish feed formulations. Aquaculturists have found that they are not able to entirely replace the fish meal proteins with plant-based proteins due to amino acid profile differences and possible anti-nutritional factors present in plants, which have caused defects and slow growth rates (Gatlin et al., 2007; Hardy, 2010). Thus, more research is required for plant-based proteins to be included in higher concentrations into formulated fish diets to allow for optimal fish growth and health.

1.2 Yellow Perch

One of the top species priorities for research throughout the Great Lakes region and Ohio is yellow perch, a predatory fish (Weeks et al., 2014) and our species of interest. Yellow perch belong to the Percidae family, one of the largest families of freshwater fish with over 150 species. They are a cool water fish that display optimum growth potential at temperatures between 22-24°C, yet require cold temperatures of around 6-12°C for spawning. Males become sexually mature after 1-2 years of age, or 68-168mm total length, while females are not sexually mature until at least 2 years of age, or 94-173mm

2 total length (Hart et al., 2006). When spawning, multiple males fertilize one female as eggs are being released into the water and are usually deposited over plants, sand or gravel. Eggs are released in gelatinous ribbon with many folds that ward off predators

(Almeida et al., 2017), and females can release anywhere from 10-40,000 eggs (Hart et al., 2006). Once a female deposits eggs and they are fertilized, they are left to develop on their own, which can be a few days or up to two weeks depending on temperature. At the time of hatching, larvae are between 4.5-7mm long and rely on reserved yolk sac as the nutrient source for the first few days. Afterward, the mouth opens, and the stomach and intestinal tract develop, allowing for exogenous feeding to begin on algae, rotifers, and zooplankton. In the next month, the intestinal tract continues to become more complex, with the development of the liver, pancreas and pyloric caeca and their connection to the digestive tract, along with the separation of the anterior and posterior intestine

(Kestemont et al., 2015). It is at this point that cultured juveniles can be transitioned to formulated feeds, yet this has been found to be difficult due to low acceptance rates (Hart et al., 2006).

A significant reason for choosing yellow perch as our species of interest is because regardless of age, they are notoriously difficult to cultivate and grow effectively on a plant-based diet compared to rainbow trout or salmon (Brown et al., 1996). This poses a problem since yellow perch reach market weight at 150g, which can take anywhere from

15-18 months when grown in ponds, or 9-12 months in recirculation systems (Hart et al.,

2006; Kestemont et al., 2015). Yellow perch aquaculture in the North Central Region

(NCR) has traditionally focused on producing this species for recreational pond stocking. 3

Producers feed fish meal-based diets to yellow perch destined for ponds, lakes, and intensive recirculating systems to maximize their production rates, and produce a robust fish for markets and fish processors. Present estimates forecast that yellow perch are being caught within the Great Lakes region at around 50 million pounds per year, which substantially exceeds the wild supply (Environment and Climate Change Canada and the

U.S. Environmental Protection Agency, 2017), and yellow perch aquaculture is responsible for allowing the supply to meet the demand. Thus, the future success within the yellow perch aquaculture industry in Ohio and the Midwest requires addressing the significant physiological/nutritional challenges of predatory fish that impede sustainable aquaculture. This can be accomplished, in part, by nutritionally programming fish

(Chapter 2) to thrive on dietary plant proteins during their grow-out stage, which solves a significant problem toward becoming an economically successful production.

1.3 Protein replacement in aquafeeds

The overall hypothesis of this work is that there is a phenotype-diet interaction that changes in response to plant-protein-based diets delivered through nutritional programming. The studies found in Chapter 2 and 3 present a way to program yellow perch as juveniles to thrive on a plant-based diet when they are adults. Replacement of fish meal proteins with other plant proteins (barley, corn, wheat, pea, cottonseed, etc.) and animal by-products, have been extensively examined to reduce the cost of aquafeeds

(Gatlin et al., 2007). Of the many plant proteins considered, soybean meal has become one of the most promising fish meal replacements. Soybean, Glycine max Linnaeus, has

4 been widely used due to its high protein content, relatively well-balanced amino acid profile and lower cost than fish meal (Chou et al., 2004). The incorporation of soybean meal in fish diets could alleviate some of the sustainability and cost problems associated with fish meal use, as currently, the cost of fish meal is approximately four times more expensive than soybean meal (Indexmundi.com, 2018). Some of the first studies using soybean meal as a protein replacement for aquafeeds occurred in the late 1960’s and through the 1970’s with limited success (Cho et al., 1974; Kaneko, 1969). Researchers found that by supplementing soybean meal-based diets with additional amino acids, resulted in better fish growth and health (Dabrowska and Wojno, 1977; Rumsey and

Ketola, 1975). However, while research on soybean meal inclusion in fish diets has continued since those initial studies, not much has changed, as commercial aquafeeds still only contain around 30% soybean meal protein or less (Jobling, 2016; National Research

Council (NRC), 2011).

The main issue researchers continue to face is although soybean meal is an inexpensive alternative to fish meal, it contains certain undesirable nutritional characteristics. These include high carbohydrate levels, the presence of isoflavones, low methionine levels and anti-nutritional factors (protease inhibitors, lectins, phytic acid, saponins and trypsin inhibitors) that may impede protein digestion and cause intestinal inflammatory responses that may ultimately hinder the growth of the fish (Francis et al., 2001; Gatlin et al., 2007).

Heat does destroy some of these factors; however, the complete removal of carbohydrates and anti-nutritional factors from soybean meal through various processing methods (soy

5 protein concentrate) makes the product as expensive as fish meal. Therefore, it has a limited practical value in commercial diet formulations (Gatlin et al., 2007).

There is abundant evidence of inflammatory responses in the distal intestine of Atlantic salmon due to the presence of soybean meal antinutritional factors in the diet

(Baeverfjord and Krogdahl, 1996; Bakke-McKellep et al., 2008; Krogdahl et al., 2010;

Moldal et al., 2014). Studies have shown that these antinutritional factors can cause shortening of the intestinal mucosal folds and a widening of the lamina propria (Bakke-

McKellep et al., 2000; Ostaszewska et al., 2005; van den Ingh et al., 1991). These changes are thought to disrupt normal digestive and regulatory mechanisms of intestinal function such as; reduced intestinal absorptive ability (Bakke-Mckellep et al., 2007;

Krogdahl et al., 2000; Urán et al., 2008), changes in trypsin and other digestive enzyme activities (Chikwati et al., 2013; Krogdahl et al., 2003; Lilleeng et al., 2007) and proliferation of the distal intestine enterocytes (Sanden et al., 2005; Venold et al., 2012).

These effects, however, seem to be dose dependent, and other studies have shown that even with the incidence of intestinal epithelial enteritis, some species can efficiently utilize other plant proteins, such as wheat gluten, and grow similarly as their counterparts that have been fed fish meal based diets (Kwasek et al., 2012; Palti et al., 2006). While studies are limited, yellow perch have shown to be much more sensitive to the anti- nutritional factors within soy when compared to other species, and growth and performance is hindered as a result (Kasper et al., 2007). Schaeffer et al. (2011) performed a 126 day feeding trial with various levels of distillers dried grains with solubles (DDGS) and soybean meal (SBM) to juvenile yellow perch (19.1±0.5g mean 6 weight). Fish were fed one of six experimental diets with graded DDGS/SBM levels ranging from 0/31.5% to 50/4%, and growth was measured. Results showed that juveniles fed the 40/9.5% DDGS/SBM diet exhibited the highest absolute weight gain and weight gain percent. Juveniles that were fed the highest amount of SBM (0/31.5% and 10/26%) and the highest amount of DDGS (50/4%) in the diets gained significantly less weight than the other groups. The above studies, to the best of our knowledge, are the few soybean meal based feeding studies conducted on yellow perch, both producing less than optimal results when high inclusion levels of dietary SBM were used.

Experiments performed in this dissertation describe feeding studies done on yellow perch with SBM protein replacing 75% of the FM protein in formulated diets.

1.4 Nutritional Programming

While many studies are focused on altering the soybean meal itself to be better utilized by fish through heat and chemical treatment, our goal was to adapt the fish to the SBM through nutritional programming. It was hypothesized that juvenile yellow perch could be nutritionally programmed by being fed SBM-based diets as their first formulated feed and that this would result in improved growth performance when reintroduced to this diet later in life. Nutritional programming is described as prenatal, or early dietary neonatal events exerted during critical developmental windows that may result in permanent changes later in life, such as, growth potential, health and metabolic status in mammals

(Holland et al., 2016; Lucas, 1998; Patel and Srinivasan, 2002). The underlying

7 mechanism behind nutritional programming is thought to be through epigenetic modification.

Epigenetics refers to the study of heritable patterns (phenotypes) of gene expression that are not caused by changes in the DNA sequence (Niculescu and Haggarty, 2011). The

Greek prefix epi- (meaning over, outside of, around) in epigenetics implies features that are "on top of" or "in addition to" the traditional genetic basis for inheritance (Berger et al., 2009). Conrad Waddington proposed the idea of epigenetics in 1942, in which he explained that changes in the genotype produce correlated changes in the adult phenotype, and subsequently, the heritability of a phenotype passed on through either mitosis or meiosis. Barker and Osmond (1986) also explained this phenomenon by reviewing previous studies that took place in England and Whales in the 1920’s, which discovered a link between infant mortality rates from 1921-1925 and heart disease mortality rates in 1968-1978 caused by poor maternal and infant diets. After the findings from this study were presented, epigenetics also became known as the Barker Hypothesis.

Epigenetic interactions among genes and the environment occur over time and are a result of an adaptation beginning with the embryonic development and fetal stages in mammals (Patel and Srinivasan, 2002), and early larval/juvenile stages in fish (Clarkson et al., 2017). These result in functional changes to the genome but do not involve a change in the nucleotide sequence (Jaenisch and Bird, 2003). Such changes are brought about through DNA methylation and histone modification, each of which alters how genes are expressed without altering the underlying DNA sequence. Mechanisms are

8 complex interactions between DNA and nuclear proteins (mainly histones) that regulate gene expression in a given cell. These DNA-histone interactions, as well as gene expression, are influenced by small non-coding RNA (ncRNA) molecules, which further regulate the pattern of gene expression defining a specific cellular phenotype (Niculescu and Haggarty, 2011). DNA methylation is a biological process that results in the addition of methyl groups to DNA. Methylation is a modification that does not alter the nucleotide sequence (DNA structure), yet results in changes in gene expression (Vanhees et al., 2014). Other nuclear proteins that are critical for epigenetic gene regulation are chromatin remodeling complexes, effector proteins with various binding modules for different modifications, and insulator proteins (Milagro et al., 2013).

It was previously thought that these epigenetic changes are surface level, and were not passed on from parent to offspring, yet opposing ideas have emerged. While most epigenetic changes only occur within the course of an individual organism's lifetime; these epigenetic changes can be transmitted to the organism's offspring through a process called transgenerational epigenetic inheritance. Heard and Martinienssen (2014) explain transgenerational epigenetic inheritance also has the potential to be adaptive, and in some cases, may even respond to environmental challenges, with significant implications for heredity, breeding, and evolution.

Nutritional programming through epigenetics has been studied in mammals for many years (Lucas, 1998; Patel and Srinivasan, 2002), it is a relatively new concept within aquaculture, and has only begun to be explored, starting with Koven et al. (2003)

9 followed by Vagner et al. (2007) and Geurden et al. (2007). Since then, many studies have examined early life exposure of various dietary ingredients such as, plant-based feeds (Geurden et al., 2013; Le Boucher et al., 2012, 2011, Perera and Yúfera, 2017,

2016); carbohydrates (Fang et al., 2014; Geurden et al., 2007; Rocha et al., 2015); and lipids (Koven et al., 2003; Vagner et al., 2009, 2007). The success of inheritability of nutritional programming has also been studied in successive generations of fish fed plant- based feeds (Geurden et al., 2013; Izquierdo et al., 2015; Lazzarotto et al., 2016; Le

Boucher et al., 2012, 2011). Le Boucher et al. (2011) found that after only one generation of breeding, there was already a noticeable improvement in the growth, survival, and variability when rainbow trout were selected for best performance when fed soybean meal.

To the best of our knowledge, only a few studies have investigated the mechanisms behind these nutritional programming changes. Balasubramanian et al. (2016) examined the molecular pathways associated with nutritional programming of plant based feeds through RNA-sequencing (RNA-seq) of the transcriptome. Perera and Yúfera (2017) studied the possible epigenetic changes in Gilthead seabream (Sparus aurata L.) induced by nutritional programming on a SBM-based diet, specifically the modification of histones. In their study, they investigated varying lengths of feeding juveniles either FM or SBM based diets as their first feed immediately after hatching. Histone modifications were examined on the entire larvae, focusing on global methylation and specific histone 3 and histone 4 (H3 and H4, respectively) modifications after 4, 14 and 32 days post- hatching (dph). It was found that SBM fed larvae had a global increase in H3K14ac, 10 which is a well-known activating mark of transcriptional initiation and enhances expression, elongation, and DNA repair. Additionally, it was observed that SBM globally decreased H3K27m3, which is known to occur along with histone activating marks such as H3K14ac at bivalent promoters and its decrease may also indicate the up- regulation of regulatory genes (Perera and Yúfera, 2017).

1.5 RNA-sequencing

Gene expression changes that occur after fish consume plant-based proteins are well documented in other species such as Atlantic salmon and rainbow trout (De Santis et al.,

2015; Martin et al., 2016; Qian et al., 2014; Sahlmann et al., 2015). Such reports are limited in yellow perch, and one the goals of the current research was to examine gene expression changes in juveniles after being fed SBM-based diets as their first formulated feed. However, our initial gene expression experiments have proven to be difficult, providing only limited results, because there is currently no annotated genome for yellow perch and insufficient nucleotide sequences available. Because of these difficulties,

RNA-seq was used to annotate the transcriptome of these fish and quantify the relative abundance of transcribed genes. Both gene expression studies and gene annotation of the yellow perch transcriptome have been challenging due to inadequate existing resources.

Even some of the relatively closely related species to yellow perch such as the European perch (Perca fluviatitis) have limited available nucleotide sequences. In a study by Li et al. (2017) in which the gonad and muscle transcriptomes of yellow perch were analyzed, authors were only able to annotate 17.04% of all assembled sequences as a single gene

11 with a significant hit when searched against six reference protein sequences from

Zebrafish (Danio rerio), Tilapia (Oreochromis niloticus), Cichlid (Haplochromis burtoni), Amazon molly (Poecilia formosa), Guppy (Poecilia reticulata), and Medaka

(Oryzias latipes). Studies such as these prove to be challenging due to such a substantial genetic variation among fish species (Blomme et al., 2006; Volff, 2005), among percids

(Michel et al., 2009) and even yellow perch themselves (Kocovsky and Knight, 2012;

Sullivan and Stepien, 2014). In addition, 96% of all extant fish are teleosts or ray-finned fishes (over 25,000 species), that make up half of vertebrates on earth (Steinke et al.,

2006), some of which were introduced over 500 million years ago (Betancur-R et al.,

2017). To depict the enormous diversity among fish species, a phylogenetic tree (Figure

1.1) that represents the major groups (ordinal and subordinal taxa) is presented. Yellow perch belong to the family Percidae which is found in the Order Perciformes and Class

Actinopterygii and contains 11 genera, with an estimated 266-275 species (Kestemont et al., 2015). In Figure 1.1, Betencur-R et al. (2017) grouped the order Perciformes in with

Eupercaria, found at the bottom of the tree. Figure 1.2 depicts the phylogeny among the family Percidae and its sister group Niphonidae, to even further examine variation solely among the various perch species. These two figures represent the tremendous evolution and variation among fish species, and just how diverse Percidae fish are from each other, let alone from Salmonids or Cyprinids. These differences need to be taken into account when gene predictions in transcriptomic studies are made, thus it is important to not limit searches among what is considered to be a "close" species. Even with the increasing number of computer programs available, effective gene predictions and identification of

12 biological functions in transcriptomic and gene expression studies are still a challenge

(Martin et al., 2016).

One goal of this project was to analyze the intestinal transcriptomic differences in yellow perch juveniles after being fed FM or SBM as their first feed. The transcriptome refers to the entire set of transcripts (messenger RNA (mRNA) and non-coding RNA (ncRNA)) at a given time within a specific cell, tissue or organism that regulate gene expression and maintain cellular homeostasis (Lindberg and Lundeberg, 2009). This varies from the genome, which is made up of all the genetic material or the proteome, consisting of all of the proteins (Cao et al., 2001). While the genome is relatively stable, the transcriptome is quite variable under conditions that alter gene translation, such as developmental stage, physiological states and changes to the external environment. RNA-seq captures a snapshot in time of the total transcripts present in the cell (Martin et al., 2016).

In order to examine the transcriptome, RNA is extracted from a specific sample, converted to complementary DNA (cDNA), and libraries are generated following slight modifications to mRNAs, by either removing the ribosomal RNAs or by selecting the poly A tails at the end of the mRNAs (Brown, 2006; Martin et al., 2016).

Polyadenylation is the addition of a poly A tail which is comprised of multiple adenosine bases at the end of an mRNA sequence. It is part of the process that produces mature mRNA for translation, the precursor to gene expression. The cDNA is then fragmented into pieces of various sizes and sequenced, either from a single direction or in both directions (pair-wise) of the strand, which ensures better coverage. At this point, the

13 transcriptome is assembled by aligning sequences to a reference genome, if present, or the method used in this study, de novo assembly. In de novo assembly, sequence fragments are aligned as close as possible based on base pair overlap, all done without a reference genome. Once assembly of the transcriptome is complete, sequences are

BLASTED against other species for prediction, and the relative abundance of mRNA

(expression level) for each gene is calculated (Martin et al., 2016; Ozsolak and Milos,

2011; Wang et al., 2009).

Gene expression analysis at the mRNA level dates back to 1977 as northern blots

(Alwine et al., 1977) and around the same time as DNA sequencing by the Sanger method (Sanger et al., 1977) was developed. Work continued, and development of cDNA libraries by Sim et al. (1979). Through the 1980’s the Sanger method was used to sequence random transcripts, thus producing expressed sequence tags (ESTs) (Putney et al., 1983; Sutcliffe et al., 1982), which are short, subsequences of a cDNA sequence.

ESTs represent a portion of expressed genes and are used to identify gene transcripts and sequences. Significant progress was made through the 1990s, and a partial human transcriptome was annotated with approximately 600 ESTs (Adams et al., 1991). Other sequencing methods such as DNA microarray and serial analysis of gene expression

(SAGE) also became popular throughout the 1990’s, further advancing the field (Schena et al., 1995; Velculescu et al., 1995). But it was not until the late 1990’s that the word

"transcriptome" was used in such studies (Pietu et al., 1999; Velculescu et al., 1997). The initial fish transcriptome was annotated by Ju et al. (2000) and Cao et al. (2001) from the same group which analyzed the transcriptome of the channel catfish (Ictalurus 14 punctatus). These three sequencing approaches were optimized throughout the 2000’s, and in 2006 the first high-throughput RNA-seq of cDNA libraries was completed

Bainbridge et al. (2006). By 2008, RNA-sequencing gained in popularity with three main publications utilizing the newest short-read technology (Mortazavi et al., 2008; Sultan et al., 2008; Wilhelm et al., 2008) and de novo assembly without a reference genome was developed (Zerbino and Birney, 2008). Since then, new operating systems and platforms for analyzing and interpreting such large quantities of data have been developed and optimized (Mcgettigan, 2013). RNA-seq technology has revolutionized genomic research, especially in the aquaculture field, allowing scientists to analyze gene expression levels of fish species that previously had no genome annotation. Fish transcriptomics has been used to study topics in fish such as developmental biology, immunology, toxicology, physiological processes, evolution, dietary ingredients and more, achieving a much deeper understanding of biological processes (Martin et al.,

2017, 2016; Qian et al., 2014).

While still limited in comparison to rainbow trout or Atlantic salmon analyses, there have been a few transcriptomic studies completed on yellow perch. Goetz et al. (2009) examined the hepatic transcriptome of yellow perch after exposure to 17 β-estradiol and generated approximately 3,600 ESTs, the first for this species. Most studies in yellow perch were toxicology studies on wild-caught fish exposed to contamination or heavy metals. Pierron et al. (2011) were the first to use initial RNA-seq technology to examine the hepatic transcriptome of wild yellow perch that had been exposed to metals. In the years following, other authors used the data generated by Pierron et al. (2011) to further 15 investigate wild yellow perch response to contamination (Bélanger-Deschênes et al.,

2013). Bougas et al. (2013) also used the initial RNA-seq data to generate a 1,000 gene microarray examining the hepatic transcriptome of wild yellow perch after chronic metal exposure. The microarray was then used by Azizishirazi et al. (2014) to determine effects of long-term metal exposure on the olfactory rosette of wild yellow perch, and again by Bougas et al. (2016) who examined effects to the hepatic transcriptome of wild yellow perch after brief metal exposure. To the best of our knowledge, the only other de novo transcriptomic study that uses Illumina high-throughput sequencing on yellow perch was conducted by Li et al. (2017). In this study, gonad and muscle transcriptomes of sex- reversed males (testosterone treated) were compared to those of un-treated male and female yellow perch. Results showed that there were no female or sex-reversed male- biased genes involved in molecular pathways, but male-biased genes were specifically neuroactive ligand receptors and trypsin. In the muscle, it was found that pyruvate kinase was highly expressed, which was shown to play a role in cell proliferation.

In regards to dietary ingredients specifically, there have been a multitude of publications examining changes to the transcriptome after fish meal proteins are replaced with plant- based proteins in diets of rainbow trout or Atlantic salmon in the liver (De Santis et al.,

2015; Jordal et al., 2005; Morais et al., 2011; Panserat et al., 2009), whole body

(Lazzarotto et al., 2016), brain (Balasubramanian et al., 2016) and intestine (Abernathy et al., 2017; Kortner et al., 2012; Król et al., 2016; Morais et al., 2012a; Sahlmann et al.,

2013; Tacchi et al., 2012). Of the many studies that investigated dietary SBM replacement on the transcriptome, Sahlmann et al. (2013) looked at the initial responses 16 after day 1, 2, 3, 5 and 7 in Atlantic salmon. Results showed that the most prominent gene expression changes were seen after day 3 and 5, in the immune-related transcripts with increased expression in GTPase IMAP family members, regulators of T-cell and B- cell function, and NF-κΒ-related genes. Fish sampled on day 5, and 7 showed down- regulation of genes related to metabolic processes, endocytosis, exocytosis, and detoxification, which indicated impairments to digestive and metabolic functions. In a longer-term study, De Santis et al. (2015) fed Atlantic salmon graded levels of SBM (0,

10, 20 and 30%) for 12 weeks. Gene expression in the distal intestine was analyzed thought microarray, and results of the fish fed the 30% SBM showed an increase in expression of immune-related responses (T-cell mediated processes, TNF-α signaling pathways, NF-κΒ mediated responses), anti-inflammatory proteins (annexin A1) and of pathways involving protein synthesis and cell proliferation, which indicate active regeneration to damaged intestinal tissue. Among genes whose transcription was down regulated were those involved with metabolic pathways (lipid, sterol and vitamin metabolism, digestion and absorption), along with those in pathways associated with phagocytosis (cellular organelles, lysosomes and phagosomes). Authors were unsure if down-regulation of these processes were caused by tissue damage, or the presence of anti-nutritional factors in SBM, that interfere with normal absorption of nutrients and are known to cause hypocholesterolemia (De Santis et al., 2015; Kortner et al., 2014).

Within the past few years, transcriptomic analyses had been conducted on nutritionally programmed Atlantic salmon (Clarkson et al., 2017; Vera et al., 2017) and rainbow trout

(Abernathy et al., 2017; Balasubramanian et al., 2016) after being programmed on plant- 17 based diets. There are also studies that investigated transcriptional changes based on genotype in Atlantic salmon and the response to vegetable oil replacement in the diet

(Morais et al., 2012b, 2011). However, to the best of our knowledge, there are no current studies examining intestinal transcriptional changes in yellow perch juveniles and their response to FM or SBM-based diets as the first formulated feed.

1.6 Isoflavone accumulation

The overall goal of this study was to reduce fish meal use in aquaculture feeds and to nutritionally program yellow perch to utilize soybean meal protein more efficiently. It is our long-term goal to translate the results of this study to other species of fish to enhance aquaculture production here in the US. However, with the higher inclusion levels of soybean meal in aquafeeds, it is essential to understand especially from a consumer’s perspective if there are nutritional changes or accumulation of soy bioactives

(isoflavones) in the muscle/meat of these fish. Isoflavones are a class of phytochemicals within the flavonoid family found in legumes, such as soybeans, peas, red clover and alfalfa and other members of the Fabaceae family (Messina, 2010; Vacek et al., 2008).

Like many other phytochemicals, isoflavones are secondary metabolites produced by the plant in response to various stressors. They possess antimicrobial, antifungal and antioxidant properties that aid in the survival of the soybean plant (Messina, 2010).

Because they are secondary metabolites, concentrations within the plant vary significantly from crop to crop and are dependent on genetics, site of cultivar and

18 environmental conditions such as temperature, water availability, UV exposure and pests

(Eldridge and Kwolek, 1983; Lee et al., 2003b).

It was first proposed in the 1940's that plants can exert estrogenic mimicking activity after a study was conducted by Bennetts et al. (1946) in which female ewes displayed various fertility problems after eating red clovers. By the 1950’s isoflavones were identified as the compounds of interest that exhibited estrogenic activity, and studies had begun on specific reproductive effects (Carter et al., 1955; Kendall et al., 1950).

Mechanistic studies of its estrogenic activity continued into the 1960’s when scientists determined isoflavone binding affinities for estrogen receptor alpha and validated its role as both an estrogen agonist and antagonist (Noteboom and Gorski, 1963; Siebert and

Macfarlane, 1969). Studies that continued into the 1980's began to uncover how isoflavones were broken down and digested after ingestion (Akiyama et al., 1987;

Setchell et al., 1984), and it was not until the 1990’s that the role of soybean meal and its role in disease prevention became of interest to health scientists (Barnes et al., 1990;

Messina and Barnes, 1991; Messina, 1997; Moyad, 1999).

Structurally, isoflavones are characterized by two benzene rings linked between a heterocyclic ring. There are three main isoflavones that are classified based on four chemical forms; aglycones without attached sugars (genistein, daidzein and glycitein), β- glucosides with attached sugars (genistin, daidzin, glycitin), acetylglucosides and malonylglucosides (Inbaraj and Chen, 2012; Kao and Chen, 2006) (Figure 1.3). In plants, isoflavones are naturally present as β-glucosides; however, processing methods

19 and storage conditions can influence form as well as content (Lee et al., 2003a; Rostagno et al., 2009). After ingestion, isoflavone glucosides are broken down by glucosidases in the small intestine, converting them aglycones, where they are either absorbed or further metabolized by intestinal microflora in the large intestine (Day et al., 1998).

There is some debate as to whether these flavonoids have beneficial or adverse health effects when consumed. Positively, isoflavones have been shown to play a role in antioxidant, anticancer, and anti-inflammatory activities within mammals (Ravishankar et al., 2016; Wei et al., 1995; Yu et al., 2016). There has also been epidemiological evidence linking isoflavones to the reduction in prostate and breast cancer rates in Asia due to high consumption rates of soy products. In addition, they can be supplemented into foods for use as a delivery system and are deemed “functional foods” for their health benefits (Ahn-Jarvis et al., 2013; Messina, 2010). However, there have also been contradictory studies that discovered that isoflavones can bind to estrogen receptors and cause a pseudohormonal response (Yu et al., 2016), which may increase the percentage of proliferative cells in tumors and can elevate the weight of estrogen-dependent mammary adenocarcinomas in rats (Allred et al., 2004; Kijkuokool et al., 2006).

The two most abundant isoflavones in soybean meal are genistein, daidzein. Genistein has been found be more bioavailable due to the presence of the hydroxyl group than daidzein, yet may cause reproductive effects because it can exert both estrogenic and antiestrogenic activities due to its structural similarities to estrogen (Dixon and Ferreira,

2002). Studies in fish have found this to be true, with findings by Bennetau-Pelissero et

20 al. (2001) showing that that 0.5mg/g genistein in the diet fed to rainbow trout for one year caused reproductive issues, by reducing gamete quality and fertilization. These results were also consistent with those found in a study by Pollack et al. (2003) on striped bass fed diets with varying levels of genistein (0, 2, 4 and 8mg/g) for 6 weeks. There were significant increases in vitellogenin (estrogen-inducible precursor to yolk-protein) found in fish fed diets containing 2 and 8mg/g genistein. However, in a study by Ko et al.

(1999) on yellow perch, authors found no effects on reproductive function after feeding fish diets with levels of 0, 0.75 and 7.5mg/g genistein for 63 days.

Studies in humans conversely, have proven that genistein and other isoflavones have health benefits when consumed (Chacko et al., 2007; Rodríguez-Roque et al., 2013; Yu et al., 2016). After ingestion, isoflavone glucosides are hydrolyzed to form aglycones by glucosidases in the small intestine, where metabolites are either completely absorbed or further metabolized (Yu et al., 2016). Lin et al. (2004) found that when fed a 50 or

100mg genistein supplement for 5 days, deposition was found in the eggs of Japanese quail in a dose-dependent manner, 1.25 and 2.5 µg, respectively. The investigators hypothesized that deposition within the eggs would increase after longer periods of feeding as well as with higher doses, which would increase human consumption of isoflavones, as the eggs could be used as a delivery vehicle. However, genistein deposition studies remain limited in fish. D'souza et al. (2005) failed to find a dose- dependent deposition in rainbow trout fillets after feeding various amounts of supplemented genistein (0, 500, 1,000 and 3,000 ppm) for 6-12 months, yet overall product quality (color and taste of fillet) was not affected. Authors found 5.4-pmol/mg 21 genistein in the fillets of trout fed the 3,000ppm supplemented genistein diet after 6 months, which correlates to only a 40% soybean meal inclusion within the diet. In another study examining the deposition of genistein in both adult rainbow trout and

Siberian sturgeon, Gontier-Latonnelle et al., (2007) fed an oral dose of 200 mg/kg radioactive labeled genistein and fish were sampled 48 and 72 hours following consumption. It was determined that genistein was not largely distributed throughout the body, and remained mostly in the plasma. Distribution occurred briefly, but levels were significantly reduced after 48 hours and were not detectible after 72 hours. During dispersion, the greatest amounts of radioactive genistein present in sturgeon was found to be in the liver, yet in trout was located in the intestinal fat and gonads. For both species, muscle only contained around 0.14% of total radioactivity and levels decreased significantly after 48 hours. Because studies are limited in fish in general, with no such study on yellow perch, the goal of the current research was to determine if a dose dependent response in accumulation of isoflavones was seen within the muscle tissue of yellow perch after being fed SBM-based diets for various lengths of time.

Overall, the expected result of this project was the development of larger, higher quality yellow perch that can be raised on a SBM-based diet, a highly desirable physiological feature that will diminish reliance on expensive animal protein sources. The broader impact is a translatable and reproducible fish model for use by the commercial aquaculture industry. By combining nutritional programming with an understanding of diet-induced intestinal gene expression, the development of a translational model for commercial-scale aquaculture production of fish that performs well on plant-based feed 22 can be developed. The new strain of "herbivorous yellow perch" will become a commodity that will be highly valued by the Ohio aquaculture industry. With these promising developments and the possibility of production on a commercial scale, it was also necessary to determine if there are changes in the meat quality of fish fed soybean meal diets. Therefore, by examining possible isoflavone accumulation in the fillet, we evaluated if feeding soybean meal-based diets added value, due to the human health benefits of isoflavones.

Figure 1.1: A phylogenetic tree representing the major groups of fishes (ordinal or subordinal taxa) as well as the time of their development (million years ago). Yellow perch belong to the Order Perciformes in which authors combined with the Eupercaria group at the bottom of the tree. Presented in Betencur-R et al., (2017).

Figure 1.2: Phylogenetic relationships among the family Percidae and the sister group

Niponidae from Kestemont et al., (2015).

Figure 1.3: Basic isoflavone structure (genistein, daidzein and glycitein) with various moieties that further classify them into four chemical forms; aglycones, glucosides, acetylglucosides and malonylglucosides. Presented in Inbaraj and Chen (2012).

Chapter 2: Nutritional Programming Effects on Growth and Reproduction of Broodstock and Embryonic Development of Progeny in Yellow Perch (Perca flavescens) fed Soybean Meal-Based Diets

2.1 Abstract

The objective of this research was to determine if growth performance of yellow perch

(Perca flavescens) could be enhanced through nutritional programming when fed soybean meal-based (SBM) diets as their first feed. It was hypothesized that juvenile yellow perch could be nutritionally programmed by being fed SBM-based diets as their first formulated feed, and that this would result in improved growth performance when reintroduced to this diet later in life. It is assumed that nutritionally programmed fish transitioned back to fish meal (FM)-based diets during gametogenesis would eliminate reproductive effects from the SBM-based diets, such as reduced fecundity, egg quality or fertilization rate.

The experimental design was carried out over four phases. In Phase 1, juvenile yellow perch were divided into four dietary treatments: 1) FM control, 2) wheat gluten meal

(WG) and 3) and 4) two soybean meal (SBM-A/SBM-B) varieties, all fed to fish in triplicate. Plant proteins replaced 75% of the fish meal protein in the diet formulations, and diets were fed to juveniles for 2 months. In Phase 2, fish were combined in a

“common garden” design, and all fed the FM diet for 9 months. In Phase 3, all fish were

27 transitioned to the SBM-A diet for 7 months. Mean individual weights of fish and survival were measured at the beginning and end of each phase. Growth performance was calculated as weight gain (%) and specific growth rate (SGR, %/day. In Phase 4, fish were fed the FM diet over winter (6 months) when gametogenesis occurs and somatic growth was minimal. When females were ready to ovulate, they were manually stripped of eggs and fertilized with pond-raised males. Mean egg weight, total number of eggs, fecundity, gonadosomatic index (GSI), and fertilization rate were all calculated to determine reproductive quality.

Results from phase 3 show that fish nutritionally programmed with SBM-A, had a higher weight gain percent when returned to SBM than fish from the other three diets, though not significant. Specifically, weight gain of SBM-A fed fish was 150±31%, whereas other groups were 114±24% (FM), 110±26% (WG) and 118±15% (SBM-B). Female reproductive quality was not affected by diet and results showed no significant differences between any of the parameters examined. However, growth during phase 4 was correlated with time to ovulation, increased GSI and mean egg weight.

2.2 Introduction

Yellow perch has been identified as one of the top priorities for aquaculture research throughout the North Central region of the US and demand is exceeding supply of wild caught fish (Weeks et al., 2014). Intensive aquaculture production of this species has proven to be challenging and costly in both the larval rearing phase, as well as later in life when they require the use of relatively expensive fish meal-based diets to achieve market 28 weight (Hart et al., 2006). Therefore, the high dietary protein requirements of yellow perch are achieved by a large fish meal inclusion in the diet (Brown et al., 1996;

Ramseyer and Garling Jr., 1998). These elevated levels of fish meal result in increased feed costs, which can be detrimental to aquaculturists, and have become unsustainable long term.

Fish meal replacement with plant proteins has been extensively studied in many species of fish. However, it has been shown that, depending on the species, only up to 30–50% of fish meal protein can be replaced by a mix of plant proteins before growth and feed efficiency are negatively affected in Atlantic cod, (Gadus morhua L) (Hansen et al.,

2007), Rainbow trout (Oncorhynchus mykiss) (Alami-Durante et al., 2010; Mambrini et al., 1999), Atlantic Salmon (Salmo salar) (Pratoomyot et al., 2010) and Yellow perch

(Kasper et al., 2007). One of the most promising fish meal replacements is soybean meal

(SBM) due to its high protein content, relatively well balanced amino acid profile and significantly lower cost than FM (Chou et al., 2004). However, high inclusion levels of

SBM in aquafeeds may pose problems due to anti-nutritional factors. Yet, the removal of anti-nutritional factors from SBM (soy protein concentrate) makes the product as expensive as fish meal; therefore, it has a limited practical value in commercial diet formulations (Gatlin et al., 2007).

SBM contains phytoestrogens and anti-nutritional factors that include trypsin inhibitors, lectins, saponins, phytic acid, oligosaccharides (Chen et al., 2011; Elangovan and Shim,

2000; Gatlin et al., 2007). They have been shown to have many negative consequences

29 on fish growth, reproduction and overall health with the most notable effect being intestinal enteritis (Gatlin et al., 2007; Krogdahl et al., 2010). Anti-nutritional factors in

SBM have been shown to cause intestinal inflammation and damage, ultimately reducing growth and feed efficiency and have been studied extensively in rainbow trout

(Oncorhynchus mykiss)(Heikkinen et al., 2006; Venold et al., 2012) and Atlantic salmon

(Salmo salar) (Bakke-McKellep et al., 2000; Krogdahl et al., 2015; Marjara et al., 2012;

Moldal et al., 2014).

Yellow perch are quite sensitive to the anti-nutritional factors within SBM, thus, growth and performance is hindered. Kasper et al. (2007) examined the effect of SBM on yellow perch juveniles (27g mean weight), which were fed diets for 8 weeks with increasing amounts (0-730g/kg) of SBM replacement. The authors found an inversely proportional response in weight gain and groups with increasing levels of SBM had reduced weight gain. Specifically, in the diet formulation with 73% SBM, weight gain of fish was roughly one fifth that of the control fish meal (0% SBM) diet, with fish gaining 20% and

108% weight, respectively. Many studies focus on modification of SBM through processing, thermal, chemical treatments, etc. to alleviate negative effects by the anti- nutritional properties. The present study focused on adapting the fish to SBM based diets rather than trying to alter the diets to the fish. This study was designed to test the hypothesis, that nutritional programming during early life exposure of yellow perch to plant-based diets can be expected to improve growth when fish are reintroduced to the same plant-based diet later as adults. This idea was first advanced by Geurden et al.

(2013) with rainbow trout. 30

Nutritional programming is described as early nutritional events exerted during critical developmental periods that may result in changes later in life such as growth potential, health and metabolic status (Patel and Srinivasan, 2002; Vera et al., 2017). The underlying mechanism behind nutritional programming is thought to be epigenetic modification. Epigenetic interactions among genes and the environment occur over time, and are a result of an adaptation beginning with the embryonic development and fetal stages in mammals (Patel and Srinivasan, 2002), and early larval/juvenile stages in fish

(Clarkson et al., 2017). These result in functional changes to the genome but do not involve a change in the nucleotide sequence (Jaenisch and Bird, 2003). Such changes are brought about through DNA methylation and histone modification, each of which alters how genes are regulated without altering the underlying DNA sequence.

While nutritional programming has been employed in mammals for many years, it is a relatively new concept within aquaculture, and has only begun to be explored, starting with Vagner et al. (2007) and Geurden et al. (2007). Since then, many studies have examined early life exposure of various dietary ingredients such as, plant-based feeds

(Geurden et al., 2013; Le Boucher et al., 2012, 2011, Perera and Yúfera, 2017, 2016); carbohydrates (Fang et al., 2014; Geurden et al., 2007; Rocha et al., 2015); and lipids

(Vagner et al., 2009, 2007). The success of inheritability of nutritional programming has also been studied in successive generations of fish fed plant-based feeds (Geurden et al.,

2013; Izquierdo et al., 2015; Lazzarotto et al., 2016; Le Boucher et al., 2012, 2011). To the best of our knowledge, only few studies have investigated the mechanisms behind these nutritional programming changes. Balasubramanian et al. (2016), examined the 31 molecular pathways associated with nutritional programming of plant based feeds through RNA-sequencing of the transcriptome. Perera and Yufera (2017) studied the possible epigenetic changes in gilthead seabream induced by nutritional programming on a SBM-based diet, specifically the modification of histones.

The hypotheses of this study were that yellow perch juveniles could be nutritionally programmed by being fed SBM-based diets as their first formulated feed, and that their early nutritional history would result in improved growth performance when fish were reintroduced to SBM-based diets later in life. It is also assumed that these nutritionally programmed fish could be transitioned back to FM-based diets during gametogenesis with no negative reproductive implications from the SBM diets, and fecundity, egg quality or fertilization rate.

2.3 Methods

2.3.1 Fish and experimental conditions: Phase 1

The experiment was conducted at The Ohio State University’s aquaculture facility and was broken up into 4 phases (Figure 1). All experimental methods and protocols were approved and performed according to The Ohio State University’s Institutional Animal

Care and Use Committee. City water supplied to the tanks was treated with activated charcoal filters and sodium thiosulfate to keep the level of chlorine below 0.01 mg/L.

Dissolved oxygen and ammonia levels were kept at constant and safe levels (Hart et al.,

2006; Kestemont et al., 2015).

In Phase 1, juvenile yellow perch (0.25+0.01 g mean weight) were randomly distributed into twelve, 37 L aquaria with 20 fish per tank. Four diets were formulated: 1) control diet with fish meal as the major protein source (FM), 2) wheat gluten meal replacing 75% of the fish meal protein (WG) (Kwasek et al., 2012), 3) Soybean meal A (SBM-A), was a non-GM, selectively-bred soy variety with high protein (56%) and low oligosaccharide content, processed as a cooked and solvent-extracted meal, and 4) SBM-B was a conventional soybean meal variety. Both SBM varieties were extruded-expelled meals, whose biochemical profiles are explained in detail in Baker et al. (2010) and Baker and

Stein, (2009). The two soybean meal varieties were provided by Schillenger Seeds Inc.

(Des Moines, IA) and replaced 75% of the fish meal protein (Table 2.1). Diets were isonitrogenous and isoenergetic (Table 2.1), were formulated in our laboratory as previously described in Kwasek et al. (2012), and remained consistent throughout the entire study. Lysine was supplemented to the plant based diets because their amino acid levels are less than set requirements for yellow perch (see details in Kwasek et al. (2014,

2012). Methionine, arginine and threonine were added to all diets based on recommendations by Brown et al. (1996). Diets were formulated to be a 3mm pellet and sieved to an appropriate size for gape width, starting with 710-850 µm, and increasing when necessary. Proximate diet analysis was conducted at the Service Testing and

Research Laboratory at The Ohio State University. AOAC official methods were used in determination of total nitrogen (method 990.03), percent ash (method 942.05) and percent moisture (method 934.01) (AOAC, 2005) and total lipid analysis was conducted using the

Folch method (Folch et al., 1956).

Fish were fed each diet in triplicate groups, for 60 days from mid July to September.

Temperatures during this time ranged from 21.9°C-25°C and fish were fed three times daily. Diets were given starting at a rate of 7% tank biomass, and re-adjusted daily, equally between tanks, based on tank with the lowest feed intake (feeding rates ranged from 6-7% tank biomass). Each tank’s biomass was collected every 20 days in order to correct and update the feeding rate. In between weighing’s, biomass increase was estimated daily, assuming an FCR of 1 (Kwasek et al., 2012). At the end of the phase, specific growth rate (SGR = (Ln (weight gain(g))/days)x100) was determined, and the feed conversion ratio (FCR= total food consumed (g)/ weight gain (g)) was calculated only after this phase. During handling and weighing, fish were anesthetized with MS-

222 (50mg/L; Western Chemical, Ferndale, WA). At the end of phase 1, fish were weighed and percent weight gain (weight gain % = final weight (g) - initial weight

(g)/initial weight (g) x100) for each diet treatment was calculated.

2.3.2 Phase 2

Prior to phase 2, fish were fin-clipped to denote initial dietary treatment, and 12 fish

(with the exception of 1 group containing 10 fish due to sampling) from each dietary treatment (n=4) were combined into a 400L tank (48 fish per tank). This “common garden” design remained the same during both phase 2 and phase 3, was done in triplicate, thus all 12 initial tanks from phase 1 were divided among 3, 400L tanks. All fish were transitioned to the same fish meal based-diet (FM) from Phase 1, and fed for 9

34 months. Feeding rate during this phase ranged from 2-5% biomass daily, and was calculated in a similar manner as previously explained from Phase 1, but adjusted to tank biomass (n=3). Fish were weighed every 3 months in order to monitor performance and update and correct feeding rate. Phase 2 spanned from September 2013 to May 2014.

Water was heated over the winter months to allow for continued somatic growth. The average temperature for the total duration of this phase was 19.8±1.8°C (Figure 2.2). At the end of Phase 2, fish were weighed and tagged with passive integrated transponders- tags (PIT- tags; Biomark, Boise, ID, USA. PIT tagging occurred at the end of Phase 2, because at this time all fish were at an acceptable size (Baras et al., 2000). In addition, this allowed for repeated monitoring of individual fish growth throughout the remainder of the study.

2.3.3 Phase 3

In Phase 3, also referred to as the “challenge” phase, all fish were transitioned to the

SBM-A diet and maintained for 7 months. SBM-A was chosen as the challenge diet for

Phase 3 due to this dietary treatment group having lower weight gain in Phase 1, and its unique nutritional profile. Fish were raised using the same “common garden” design from Phase 2 and remained in the same tanks. Feeding was done as previously explained, and adjusted equally among tanks based on daily feed intake and assumed daily biomass increase. Tank weight was measured every 3 months to further correct feeding rate, which ranged from 2-5% tank biomass per day. At the end of the phase, individual fish weights were measured, and gender was determined (December 2014).

2.3.4 Phase 4

In Phase 4 of the experiment, fish were prepared for breeding by alterations of the photothermal regime that mimicked natural seasonal variations (Figure 2.2). Fish were fed the FM-based diets for 6 months during gametogenesis in order to not compromise possible impact on gamete quality affected by soybean phytoestrogens (Bennetau-

Pelissero et al., 2001; Le Boucher et al., 2012). Fish remained in the common garden grouping in triplicate from the previous phases with 46±3 fish in each tank

(approximately 11 fish from each of the 4, phase 1 programming diets). A group of fish that served as control for spawning purposes, termed the commercial control (CC) group, was also included in breeding experiment. These fish were siblings of the nutritionally programmed fish, but were fed solely Aquamax 400-500 commercial diet (Purina, Gray

Summit, MO) throughout their lives, yet had been kept under identical environmental conditions to the nutritionally programmed fish, in a separate 400L tank. Water temperature in the tanks followed seasonal cycles similar to those found in the north central region of the US (Farmer et al., 2015). The photoperiod in the room was also manually altered to mimic natural seasonal light regimes to induce gametogenesis

(Dabrowski et al., 1996).

2.3.5 Reproduction and egg quality

From April 23 to May 31, 2015 females were checked 3 times daily for preparedness for

36 ovulation (i.e. genital papilla extension and purple in color). A record was taken of the date when females were stripped of eggs, at which point she was identified and weighed.

This allowed for the determination of the number of days since the start of Phase 4 until spawning. Egg ribbons were deposited into a dry container, weighed and held at 10°C until sperm was collected. Gonadosomatic index was calculated as follows (GSI =

((gonad weight (g)/body weight (g)) x 100) (Devlaming et al., 1982). Total number of eggs was estimated based on the number in a 1 g portion, and relative fecundity (number of eggs x g-1 total body weight) were calculated for each female (Dabrowski and

Ciereszko, 1996). A single outside control, pond-raised male source (n=4) was used to fertilize all eggs. These males were 1+-year-old from Mill Creek Perch Farm in

Marysville, OH and had only been fed a commercial FM-based diet. Sperm was stripped from males and combined, then added at a volume of 12.5µl per gram of egg; assuming

20 x 1011 spermatozoa/mL (Ciereszko and Dabrowski, 1993). A 0.4% NaCl activator was added at a volume of 0.8mL per gram of egg, immediately following addition of sperm and mixed for approximately 1 minute. Hatchery water (10mL) was then added to the eggs and they were shaken manually for 5 minutes, rinsed twice with fresh water and transferred to California-style trays (20L) with continuous water flow at approximately

14°C. After 2-4 hours following fertilization approximately 100 embryos from different portions of the ribbon were removed and observed under the stereo-microscope

(Olympus, Waltham, MA) to determine the initial fertilization rate. This was calculated by the number of eggs at the two to four-cell stage/total number of eggs. A second observation of embryonic survival was done with different egg ribbon segments at 65

37 degree-days, corresponding to the onset of the pigmented eyed stage. Due to temperature variations throughout the breeding season (14-18°C), embryonic development differs, thus requiring the use of a thermal constant expressed as degree-days (Wallich, 1901).

Embryo survival was calculated as the number of viable embryos/total number of embryos.

2.3.6 Statistical analysis

Results are expressed as mean ± standard deviation (SD). Data were checked for normality, and weight gain and survival in each phase compared using the analysis of variance (ANOVA; PROC GLM SAS v9.3). Survival of fish during each phase was arcsine transformed for normality. A Tukey-Kramer test was also done for multiple comparisons of means. Egg quality parameters (days to spawning, fecundity, fertilization rate and survival of embryos) along with female weight (g) were analyzed by a multivariate analysis of variance (PROC MIXED SAS v9.3) to determine if there was a significant diet effect. Once it was determined that there was no significant diet effect

(P>0.068) among any of the response variables, the mean of each dietary group was calculated and used to determine linear relationships among other factors that may influence egg/embryo quality (Farmer et al., 2015). Regressions were run to evaluate the relationship between the weight change in females over Phase 4 (gametogenesis) as well as female weight to the egg quality parameters listed above. All data was analyzed using

SAS software (SAS Institute Inc., Cary, NC) and a P value of <0.05 was considered statistically significant.

2.4 Results

Juvenile yellow perch were nutritionally programmed with 1 of 4 experimental diets for two months (phase 1), and then all transitioned to FM diets for 9 months (phase 2), followed by being fed SBM-A for 7 months (phase 3). Growth performance, and SGR were measured after each phase. In phase 4, females were spawned and fecundity, GSI and egg size, and were all compared back to initial programming diet from phase 1.

2.4.1 Phase 1

At the end of the 2-month period, survival rates among all dietary groups were not significantly different, with the FM and WG groups having 93±3% and SBM-A and

SBM-B having 87±3% and 88±12% survival, respectively. However, individual weights of fish differed significantly (P<0.001) between SBM groups and FM/WG groups (Table

2.2). The FCR and SGR were also significantly different among the FM/WG groups and

SBM groups. FCR was significantly higher (P<0.0001) (lower efficiency) in the SBM fed groups; SBM-A (2.11±0.18) and SBM-B (1.87±0.03) compared to the FM

(1.12±0.03) and WG (1.04±0.08) groups. The opposite was seen in regards to SGR, where the SBM fed groups had a significantly lower SGR than the other two groups

(Table 2.2). The weight gain percent was significantly lower in the two SBM fed groups

(P<0.0001), with SBM-A and SBM-B having 660±105% and 748±50% weight gain, respectively, while FM had 2028±272% and WG had 2270±126% weight gain (Figure

2.3A).

2.4.2 Phase 2

In phase 2, all fish were fed FM-based diets and growth performance results were calculated based on original diet groups from Phase 1. This phase showed a reverse trend from phase 1 in regard to weight gain (%); both of the SBM groups gained a significantly larger percentage of weight (P=0.0015) than the FM/WG groups; SBM-A (2740±734%),

SBM-B (2421±367%), FM (1152±207%), WG (881±12%) (Figure 2.3B). However, while the SBM groups gained a greater percentage of weight, the final mean weights were still lower than the FM/WG groups (Table 2.2). Specifically, SBM-A had a final mean weight of 53±11g, SBM-B had 54±21g, while FM had a final mean weight of

63±32g and WG was 60±18g and were not significantly different (P=0.735). Survival in this phase was also not significantly different, and was 90-100% in all groups. Table 2.2 shows that SGR was very similar among groups, and was not significantly different.

FCR for this phase could not be calculated based on initial diet treatments since all initial

(phase 1) treatment groups were distributed among three tanks in the common garden design. Feeding rates were calculated based on total biomass of each tank (n=3).

2.4.3 Phase 3

After being challenged with SBM-A for 7 months in Phase 3, all groups had similar final mean weights, with no significant differences among groups (Table 2.2). Mean SGR was highest in phase 3 (Table 2.2) compared to other phases, and yet, was not significantly

40 different among groups. However, while there was no significance (P=0.254) among groups in term of weight gain percent, and encouraging fact was that SBM-A had a numerically higher weight gain percent than the other groups at 150±31%; FM gained

114±24%, WG (110±26%) and SBM-B (118±15%) (Figure 2.3C).

2.4.4 Phase 4

As expected, there was minimal change in total mean weight after fish were fed FM- based diets over the winter months when gametogenesis in yellow perch occurs

(Dabrowski et al., 1996) (Table 2.2). There were no significant differences found among groups in regards to weight gain percent, FM (2.9±8.8), WG (4.7±10.2), SBM-A

(3.4±7.4), SBM-B (2.1±7.3) (Figure 2.3D). Mean SGR among groups also showed no significant differences, and again was lower in comparison to other phases (Table 2.2).

Commercial control diet fed group data are also presented in Table 2.2, which shows that this group lost weight. It is not compared in Figure 2.3 due to the fact that this group did not go through the other nutritional programming phases.

2.4.5 Reproductive quality

Reproductive performance results showed that the SBM diet fed fish during phase 3 did not have any influence on gametogenesis. Initial analyses were based on the number of females that were manually stripped of egg ribbons and successfully fertilized (FM;n=11,

WG;n=8, SBM-A;n=6, SBM-B;n=7 and CC;n=13). Female weight was taken immediately after eggs had been stripped, and there were no significant differences seen

41 among mean weights of each group; FM (203±46g), WG (207±52g), SBM-A (211±43g),

SBM-B (190±39g) and CC (171±24g). Relative fecundity results, similarly, showed no significant differences (P=0.474) among groups (Figure 2.4). Figure 2.5A/B displays the

GSI and egg size (mg) of each group and shows that there were no significant differences between groups (GSI P=0.341; egg size (mg) P=0.949). Fertilization rate taken at the 2- cell stage was 56±26% (FM), 78±14% (WG), 75±22% (SBM-A), 58±40% (SBM-B) and

76±14% (CC) (Figure 2.5C). Survival at eyed-embryo stage (65-degree days) is also presented in Figure 2.5D. There were no significant differences found among any of the groups for fertilization rate (2-cell) (P=0.250) or embryonic survival at 65-degree days

(P=0.516). Once it was determined that there was no significant diet effect (P=0.068) among any of the response variables, the mean of each dietary group was calculated and used to determine linear relationships among other factors that may influence egg/embryo quality. Regressions were run to evaluate the relationship between the weight changes in females (g) over Phase 4 as well as female weight (g) to egg quality parameters. It was found that in females that lost somatic weight over the course of phase 4, it took 173±4 days to spawn compared to those that gained the most weight (153±5 days to spawn)

(linear regression: F=17.39, Adj. R2= 0.80, P=0.025) (Figure 2.6A). Weight change of females also had a significant correlation in regards to GSI (%) (linear regression: F=

5.55, Adj. R2= 0.53, P=0.099), with females that had lost weight having the lowest GSI percentage (Figure 2.6B). In regards to female weight (g), there was a positive correlation between egg size, (Linear Regression: F= 13.65, Adj. R2=0.76, P=0.034) and female size, resulting in larger females producing larger eggs. However, there was no

42 correlation found between female weight and percent fertilization rate at the 2-cell stage

(Linear Regression: F=0.0006, Adj. R2= -0.33, P=0.98) (Figure 2.6D).

2.5 Discussion

Nutritional programming occurs during critical periods of early development, which in fish occur during their major metamorphosis phase, the larval-juvenile transition

(Burggren, 2014; Clarkson et al., 2017). In the present study, different plant protein- based diets were used to nutritionally program juvenile yellow perch to determine if these specific dietary ingredients had an influence on metabolism later in life. After programming, all fish were fed FM diets for 9 months and then challenged with a SBM diet as adults. Growth performance was compared back to initial diets. Results showed that juveniles originally fed SBM-A grew numerically better than other groups when re- introduced to the SBM-A diet as adults (Figure 2.3C, P=0.254). Reproductive performance of females in the present study was also conducted based on the findings by

Bennetau-Pelissero et al. (2001) who showed that reproduction can become impaired or reduced after feeding high levels of SBM-based diet to rainbow trout during maturation.

Our results showed no differences among female reproductive performance (GSI, fecundity, and fertilization rate) regardless of previous diet history if fish are transitioned to a FM diet during gametogenesis.

2.5.1 Phase 1

Growth results from Phase 1 were similar to previous SBM inclusion studies in salmonids and cichlids (Collins et al., 2012; Lin and Luo, 2011) with the SBM groups growing significantly less than the FM/WG diet groups (P<0.0001). These findings were consistent with Kasper et al., (2007) in which yellow perch were studied. Depressed growth in the present study’s SBM-based diet fed groups can possibly be explained by compromised FCR results. These groups had a higher FCR (lower efficacy of food utilization), which resulted in lower SGR than the FM/WG groups, which is likely due to a higher consumption of food, yet hindered digestion/bioavailability, resulting in less growth. A possible explanation for this increase in FCR could be caused by the incidence of intestinal epithelium inflammation that occurs when fish are fed a SBM diet (Chikwati et al., 2013; Perera and Yúfera, 2016; Sahlmann et al., 2013). Bakke-McKellep et al.

(2000) attributed the lower feed efficiency of Atlantic salmon fed a 30% soybean meal containing diet to lower brush border membrane enzyme activities due to the inflammatory response within the intestine.

2.5.2 Phase 2

Trends were reversed in phase 2 with the SBM groups having a significantly higher weight gain percent (P=0.0015). Yet final mean weight, weight gain and SGR at the end of phase 2 were all very similar. This is explained by the fact that the SBM groups had much smaller starting weights (2.01±0.20g) in phase 2 than the other groups (5.56±0.72g)

(Table 2.2), and the impact of compensatory growth during this phase cannot be 44 disregarded. Geurden et al. (2013) saw similar results with significant differences in body weight after the programming phase and ended with very similar mean weights after the acclimation phase in which all fish were fed fishmeal-based feed.

A very important factor to be considered in order to explain the weight gain of all of the groups, in this phase, was the fact that the water was warmed throughout the winter and held at an average of 19.8±1.8°C (Figure 2.2), near the optimum for growth (Hokanson,

1977). With temperatures of 22-24°C being optimal for yellow perch growth (Hart et al.,

2006), this allowed for continued somatic growth throughout the entire phase 2. The average temperature during the winter months is 9.2±3.7°C, which halts somatic growth and promotes gametogenesis to occur (Dabrowski et al., 1996). By keeping the temperature elevated, we were able to have uniform weights among groups prior to the onset of phase 3. This was necessary, so that any changes seen in weight gain during phase 3 were directly related to the SBM diet, and had no carry over effects from phase 2.

2.5.3 Phase 3

Since there were no significant differences in growth performance of SBM-A and SBM-

B after phase 1, SBM-A, with its unique nutritional profile, was chosen to be the SBM used in Phase 3. Promising results were seen after fish were challenged with SBM-A in phase 3. Groups that had initially been fed the SBM-A diet in phase 1 showed a numerically higher weight gain % than the other groups when re-introduced to the diet in phase 3. Even though there were no significant differences between groups in final weight, mean weight gain and SGR, the trend of the SBM-A group performing slightly 45 better than the other groups was beginning to emerge. Because this phase was completed in December due to water temperatures below 10°C and gonad development was increasing, growth differences between groups could not be followed further, yet would be an interesting path for further research. There are many suggestions as to what the mechanism nutritional programming is. One possibility is that metabolic changes are induced through epigenetic modifications with DNA methylation in the CpG islands within promoter regions and or histone modifications due to the first feed (Lillycrop et al., 2014). Milagro et al. (2013) showed that there are complex interactions between food components and epigenetic changes, which occur in mammals, most prevalently during the perinatal period, and play a large role in their adult metabolism. Another possible explanation for the programming effect is the flavor learning, or olfactory response

(Balasubramanian et al., 2016). Fish fed the SBM diets initially, may develop a memory to the flavor, and will be more willing to accept it when re-introduced to it again later in life.

The nutritional programming concept is becoming more widely accepted and within the past few years with many new studies being published. It has been studied in zebrafish with both soybean meal and high carbohydrate diets, with authors investigating tolerance to the inflammatory response and gene expression changes (Fang et al., 2014; Perera and

Yúfera, 2016). Results from both studies showed that gene expression was differently regulated in adult fish that had been nutritionally programmed early on as larvae. Other species that tested the nutritional programming concept include sea bream with plant based diets (Izquierdo et al., 2015) and European sea bass with highly unsaturated fatty

46 acids (Vagner et al., 2007). Rainbow trout have also been used in such studies with investigators examining the effects of fish meal replacement with plant proteins on growth performance and reproduction, respectively (Geurden et al., 2013; Lazzarotto et al., 2015). Geurden et al. (2013) showed that adult fish that had been nutritionally programmed on a plant protein-based diet as alevins (with starter feeds), had a significantly higher feed intake, growth rate and feed utilization as adults. This was seen after fish were fed a plant-based diet again after being fed a FM-based diet for a period of time, when compared to fish that had never been introduced to the plant-based diet. Our results are consistent with Geurden et al. (2013), which reinforces the concept of nutritional programming and that changes occur regardless of the species of fish.

2.5.4 Phase 4

Previous studies have shown that yellow perch exposure to temperatures below 10°C result in reduced somatic growth and an acceleration of gametogenesis (Farmer et al.,

2015; Wang et al., 2010). The present study found similar results during phase 4 in which all groups had reduced somatic growth when fed FM diets (P=0.115). Results showed that over a 6-month period all females, regardless of previous dietary history, had an average of only 3.3±11.5g weight gain. During this period, bodily nutrients are allocated for gametogenesis (Dabrowski et al., 1996), however reproduction efficiency can suffer if plant based diets are fed during this time (Bennetau-Pelissero et al., 2001; Hamoutene et al., 2009).

2.5.5 Reproduction

It has been shown that broodstock nutrition, specifically dietary protein and lipids, is directly related to the quality of fish eggs and larvae and is important for successful reproduction and survival of offspring (Izquierdo et al., 2001; Jonsson and Jonsson,

2014). Nguyen et al. (2012) evaluated various criteria for assessing reproductive performance as differences between species hinder standard assessments. The authors showed that assessing fecundity, egg size, buoyancy, biochemical composition and fertilization rate can be good indicators of egg quality, but may not be predictive of hatching success. Previous studies have shown that reproductive success and egg viability were significantly reduced when broodstock were fed varying levels of a soybean meal in their diets (Bagheri et al., 2013; Callan et al., 2012). It is hypothesized that reproduction is negatively influenced in fish fed SBM containing diets, due to the presence of isoflavones in SBM. Isoflavones are structurally similar to estrogens, and known to be estrogen mimicking by having a binding affinity to estrogen receptors (Ng et al., 2006).

Genistein, one of the most prevalent isoflavones, is known to have the highest estrogenic activity due to its increased binding affinity to estrogen receptors (Ko et al., 1999).

Bennetau-Pelissero et al., (2001) fed rainbow trout various amounts of genistein in their diets (0, 500 and 1000mg/kg) for a year prior to spawning and analyzed reproductive performance in females and sperm motility in males. Results showed that the 500 mg/kg genistein diet group had the most negative effects in regards to fecundity and gamete quality, as these indicators were significantly lower than in the other two groups (0 and

1000 mg/kg genistein). It was hypothesized that the highest inclusion level of geinstein 48

(1000 mg/kg) did not cause negative effects on fecundity due to its agonistic/antagonistic effect on estrogen function. The authors also found that the effects of genistein were accelerated in males, with sperm motility being affected almost immediately, whereas females showed a more delayed response in regards to gamete quality. These findings support our experimental design that used males that were pond raised and not exposed to genistein containing SBM. In addition, the females in the current study were transitioned back to FM during gametogenesis, however, it was important to determine if reproductive quality had been affected due to previous SBM diet exposure due to a possible delayed response. During winter months, when the water temperature is below 10°C, gametogenesis proceeds by means of mobilization of nutrients from the somatic portion within the body of the yellow perch (Brown et al., 1996; Kwasek et al., 2014). Our results were encouraging, showing that after fish were fed SBM-based diets for 7 months, and transitioned back to FM-based diets during gametogenesis, that no reproductive differences were seen among the experimental groups and the commercial diet fed

(control) group. This instills confidence that if yellow perch are fed FM-based diets even during the short period of gametogenesis, then reproductive issues can be avoided.

After determining that there was no effect of dietary history on fecundity and egg quality parameters, it was still important to examine trends in female size and weight change during the period of gametogenesis. Previous studies have shown that temperature increases over winter months can affect egg size and quality (Dabrowski et al., 1996;

Farmer et al., 2015), in part, due to changes in energy allocation (Feiner et al., 2016).

Since our nutritionally programmed females were subjected to a warmer winter their first 49 year of life to allow for increased somatic growth, it was important to compare female fecundity and egg quality to other studies in farmed and wild fish to ensure there were no carry over effects. The GSI of yellow perch females at the time of spawning was on average of 24±5%, which was consistent with previous literature. Wild yellow perch average GSI ranges from 15-25% (Hayes and Taylor, 1994; Henderson et al., 2000;

Jansen, 1996). However, GSI has been found to be up to 30.9% in cultured yellow perch

(Dabrowski and Ciereszko, 1996). Jansen (1996) found that relative fecundity in wild fish is usually in the range of 80-140 (number of eggs x g-1 total body weight) and can even be upwards of 233 (number of eggs x g-1 total body weight) (Hart et al., 2006). The average relative fecundity in our study was 137±38 (number of eggs x g-1 total body weight). It has also been well documented that larger female weight and length are positively correlated with an increase in GSI (%) and reproductive success (Dabrowski et al., 1996; Farmer et al., 2015; Henderson et al., 2000). Reproductive results of the present study are consistent with those obtained in wild yellow perch, thereby supporting that reproductive efficiency was not affected in our cultured yellow perch.

2.6 Conclusion

Results of this study found that nutritional programming to be a successful way to adapt yellow perch to better utilize SBM-based diets as adults. These fish, when fed a diet with

75% SBM replacement for FM protein as their first feed, and then re-introduced to the same diet 9 months later, had numerically higher growth than fish fed other diets initially.

These results are very promising as there are only a limited number of studies that have

50 shown improved growth performance with SBM inclusion above 60% in diet formulations. It was also shown that reproductive function and egg quality are not affected in fish that were previously fed a SBM diet, if transitioned to a FM diet during gametogenesis. Future studies are being conducted to determine if nutritional programming of broodstock fish will pass on the programming changes to their progeny, allowing them to be even further adapted to utilize SBM diets.

Table 2.1 Composition of experimental diets (g 100g-1) for FM, WG (Kwasek et al.,

2012), SBM-A and SBM-B

Ingredients: FM WG SBM-A SBM-B Fish meal1 54.09 12.00 12.00 12.00 SBM Schillinger2 - - 52.86 - SBM conventional3 - - - 61.67 Wheat gluten4 - 37.00 - - Wheat meal 23.41 22.71 8.23 - Fish Oil 4.34 8.00 8.00 8.00 Soybean oil 1.98 1.61 0.24 - Lecithin 4.00 4.00 4.00 4.00 CPSP 4.00 4.00 4.00 4.00 Mineral mix5 3.00 3.00 3.00 3.00 Vitamin mix5 4.00 4.00 4.00 4.00 Calcium phosphate - 1.00 1.00 1.00 Lysine - 1.50 1.50 1.50 Methionine 0.33 0.33 0.33 0.33 Arginine 0.65 0.65 0.65 0.65 Threonine 0.20 0.20 0.20 0.20 TOTAL 100.00 100.00 100.00 100.35 Moisture (%) 3.47 3.66 3.88 3.35 Ash (%) 13.83 13.24 10.50 10.05 Protein (%) 24.3 25.4 23.9 24.4 Lipid (%) 25.5 25.5 23.0 23.3 170.3% protein

256% protein,

348% protein

480% protein 5Mineral and Vitamin mixes are the same as

used in Kwasek et al., 2012

Table 2.2: Growth performance parameters calculated at the beginning and end of each

phase. All measurements were traced back to the initial diet fed during Phase 1. Means

are presented +/- standard deviation and letters denote significant differences.

Phase 1

Initial Diet Initial Mean Weight (g) Final Mean Weight (g) Weight Gain (g) Weight Gain (%) SGR FM 0.23±0.02 5.00±1.83b 4.76±0.53b 2028±272a 2.59±0.19a WG 0.26±0.01 6.13±1.42a 5.87±0.24a 2270±126a 2.95±0.07a SBM-A 0.25±0.03 1.89±0.67c 1.64±0.21c 660±105b 0.82±0.22b SBM-B 0.25±0.01 2.13±1.04c 1.88±0.08c 748±50b 1.08±0.07b

Initial Diet Phase 2- All fed FM FM 5.00±1.83b 62.7±21.9 57.7±12.7 1152±207b 1.58±0.08 WG 6.13±1.42a 60.2±18.1 54.0±2.8 881±12b 1.56±0.02 SBM-A 1.89±0.67c 52.6±10.7 50.7±7.5 2740±734a 1.53±0.06 SBM-B 2.13±1.04c 53.7±21.0 51.6±6.6 2421±367a 1.54±0.05

Initial Diet Phase 3- All fed SBM FM 62.7±21.9 132±43 69.2±1.9ab 114±24 2.05±0.01 WG 60.2±18.1 126±39 65.8±12.7ab 110±26 2.02±0.10 SBM-A 52.6±10.7 131±27 78.5±9.1a 151±31 2.10±0.60 SBM-B 53.7±21.0 116±38 62.5±4.4b 118±15 2.00±0.03

Initial Diet Phase 4- All fed FM (Females Post-Spawn) FM 149±31 152±28 2.9±11.3ab 2.9±8.8a 1.4±0.6 WG 151±33 156±29 5.0±13.7a 4.7±10.2a 1.4±1.0 SBM-A 154±27 159±28 4.9±10.9a 3.4±7.4ab 0.8±1.4 SBM-B 147±41 149±37 1.9±10.6ab 2.1±7.3ab 1.2±0.7 CC 140±25 133±21 -6.9±9.8b -4.5±6.5b

Phase 1: Phase 2: Phase 3: Phase 4: FM, WG, All fed FM All fed SBM All fed FM SBM-A/B 2 Months 9 Months 7 Months 7 Months

Figure 2.1: Nutritional programming experimental design. Flow chart depicting each diet fed to the fish, and the duration of each phase.

25 Mean Daily Temperature Natural Temperature Variation 20

10 Spawning

Water Temperature (° C) (° Temperature Water 5 Phase 1: Phase 2: FM Phase 3: SBM Phase 4: FM Program Diet Diet Diet 0 July Sept Nov Jan March May July Sept Nov Jan March May Month

2013 2014 2015

Figure 2.2: Temperature data throughout the duration of the experiment from 2013-2015.

The dashed line represents normal temperature fluctuations throughout Phase 2, while the solid line denotes the warmed water temperature during this phase and temperatures for the remainder of the experiment.

Phase 1 A Phase 2- FM Diets B 3000 4000 a a 2500 a 3000 a 2000

1500 2000 b 1000 b b b 1000 Weight gain (%) Weight 500 gain (%) Weight 0 0 FM WG SBM-A SBM-B FM WG SBM-A SBM-B Programming (phase 1) diet Programming (phase 1) diet

Phase 3- SBM Diets Phase 4- FM Diets 200 C 20 D

150 15

100 10 50

Weight gain (%) Weight 5 Weight gain (%) Weight

0 0 FM WG SBM-A SBM-B FM WG SBM-A SBM-B Programming (phase 1) diet Programming (phase 1) diet

Figure 2.3: Weight gain (%) results for fish fed different initial programming diets after each phase. A) Phase 1- “programming,” fish fed 4 original diets, B) Phase 2- all fish fed

FM-based diet C) Phase 3- all fed SBM-A based diet, D) Phase 4- all fed FM-based diet over winter months. Results are grouped into original programming diets from phase 1.

Letters denote significant differences among groups (P<0.0015).

200

-1 175 150 125 100 75 50 25 total bodytotal weight) 0 FM WG SBM-A SBM-B CC Fecundity (number of eggs x g (number Fecundity Diet

Figure 2.4: Relative fecundity among females fed different initial programming diets.

Relative fecundity is presented as number of eggs x g-1 total body weight. No significant differences were found among groups (P=0.474).

35 A 3.0 B 30 2.5 25 2.0 20 1.5 15 GSI (%)

10 (mg) Egg Size 1.0 5 0.5

0 0.0 FM WG SBM-A SBM-B CC FM WG SBM-A SBM-B CC

100 C 100 D 90 90 80 80 70 70 60 60 50 50 40 40 30 30

20 Eyed stage survival 20 Fertilization rate (%) rate Fertilization 10 10 0 0 FM WG SBM-A SBM-B CC FM WG SBM-A SBM-B CC Programming (phase 1) diet Programming (phase 1) diet

Figure 2.5: Reproductive indices compared among females from different initial programming diets; A) GSI (%) (P=0.341), B) egg size (mg) (P=0.949), C) initial fertilization rate (%) at the 2 cell division stage (P=0.254) and D) Eyed-stage embryonic survival (%) after 65° days (P=0.516).

180 A B 29 175 y = -1.5094x + 165.9 27 170 Adj. R² = 0.804 25 P = 0.0251 165 23 160 GSI (%) 21

Days to Spawn 155 19 y = 0.3662x + 22.702 Adj. R² = 0.532 150 17 P=0.0998 145 15 -20.0 -10.0 0.0 10.0 20.0 -20.0 -10.0 0.0 10.0 20.0 Weight Change (g) Weight Change (g)

2.90 C 100 D 90 2.70 80 2.50 70

2.30 60 50 2.10

Egg Size (mg) Egg Size 40 1.90 30 y = 0.0076x + 0.8281 (%) at 2 Cell Survival 20 y = 0.0083x + 67.025 1.70 Adj. R² = 0.759 Adj. R² = -0.333 P=0.0344 10 P=0.981 1.50 0 140 190 240 140 160 180 200 220 240 260 Female Weight (g) Female Weight (g)

Figure 2.6: Correlations among female reproductive parameters. All data is presented as means of females from original dietary programming groups. Because diets had no significant effect on response variables (P=0.068), the linear regressions were run comparing; A) Days from start of Phase 4 until spawning vs. weight change (g), B) GSI

(%) vs. weight change (g), C) Egg size (mg) vs. female weight (g), D) Fertilization rate

(%) at the 2-cell stage vs. female weight (g).

Chapter 3: Inheritance of parental nutritional programming on offspring growth performance in yellow perch (Perca flavescens) when fed soybean meal-based diets as first feed

3.1 Abstract

With the need to continually reduce fish meal (FM) in aquafeeds for both cost and sustainability reasons, diets are being formulated with increasingly higher inclusion levels of alternative protein sources. Yet these diets are still expected to yield high quality fish with a similar growth performance to that of those fed FM. Through nutritional programming, fish can develop a metabolic memory response to a soybean meal-based diet when fed as their initial feed, displaying improved growth when reintroduced to it later in life. In addition, it is well known that broodstock diets directly influence reproduction and larval quality, and more studies have begun to examine the inheritance of nutritional programming and its effects on offspring. The present study is a continuation of our previous experiments in which yellow perch were nutritionally programmed as juveniles with various plant based diets as their initial formulated feed

(Phase 1), all transitioned to a FM based diet for 9 months (Phase 2), challenged with a soybean meal (SBM) based diet for 7 months (Phase 3) and transitioned back the FM based diet prior to spawning (Phase 4). Broodstock growth performance was measured after each phase, and reproductive quality parameters after Phase 4 were examined.

The goal of the current study was to use the previously nutritionally programmed broodstock fish (F0) and spawn them successively over the course of 3 years (2015, 2016 and 2017). We determined the inheritability potential of nutritional programming by bringing the offspring (F1) through the same 3 phases, with the same diets (FM2 and

SBM2) in order to compare growth performance. In year 1 (2015), nutritionally programmed females were crossed with pond raised males from an outside source to determine maternal inheritability. In year 2 and 3 (2016/2017) both nutritionally programmed males and females from the same initial diet were crossed in order to compare the combined parental inheritability. Results showed that the offspring groups from 2016/2017 outperformed those from 2015 in terms of overall size, weight gain and specific growth rate (SGR) over all three phases. At the end of phase 1 it was found that both parental diet and offspring diet, individually, had significant effects on weight gain percent when comparing all three years of offspring growth. Additionally, at the end of phase 3, again both the maternal diet and offspring diet, independently, were found to have a significant effect on weight gain percent, yet no significant effects were seen from the interaction of the two factors.

3.2 Introduction

Nutritional programming is described as prenatal or early nutritional neonatal events exerted during critical developmental windows that may result in permanent changes later in life, such as, growth potential, health and metabolic status in mammals (Holland et al., 2016; Lucas, 1998; Patel and Srinivasan, 2002). The underlying mechanism

60 behind nutritional programming is thought to be epigenetic modification. Epigenetic interactions among genes and the environment occur over time, and are a result of an adaptation beginning with the embryonic development and fetal stages in mammals

(Patel and Srinivasan, 2002), and early larval/juvenile metamorphosis stages in fish

(Clarkson et al., 2017). Nutrients and bioactive food components can influence epigenetic events either by directly inhibiting enzymes that catalyze DNA methylation or histone modifications, or by altering the availability of substances necessary for those enzyme reactions (Choi and Friso, 2010).

The concept of nutritional programming has been well documented in mammals (Hanley et al., 2010; Lillycrop et al., 2014; Lucas, 1998), and is becoming increasingly prevalent in aquaculture. Studies have examined early life exposure of various dietary ingredients such as plant protein-based feeds (Geurden et al., 2013; Le Boucher et al., 2012, 2011,

Perera and Yúfera, 2017, 2016); carbohydrates (Fang et al., 2014; Geurden et al., 2007;

Rocha et al., 2015); and lipids (Vagner et al., 2009, 2007). However, the majority of these studies focus on initial nutritional exposure and their effects later in life, with only a few beginning to examine the effect of broodstock nutritional programming on offspring responses such as feed acceptance and growth. In addition, it has been previously disputed that these nutritional programming/epigenetic changes were considered surface level, and were not passed on from parent to offspring, yet opposing ideas have emerged.

While most epigenetic changes only occur within the course of an individual organism's lifetime; these epigenetic changes can be transmitted to the organism's offspring through a process called transgenerational epigenetic inheritance. Heard and Martinienssen (2014) 61 explained that transgenerational epigenetic inheritance also has the potential to be adaptive, and in some cases, might even respond to environmental challenges, with major implications for heredity, breeding and evolution.

It has been well documented that broodstock nutrition influences reproductive performance and larval quality of fish (Fernández-Palacios et al., 2011; Izquierdo et al.,

2001). Thus, more studies have begun to examine the success of inheritability of nutritional programming among successive generations of fish fed plant-based feeds

(Fontagné-Dicharry et al., 2017; Izquierdo et al., 2015; Lazzarotto et al., 2016, 2015;

Seiliez et al., 2017; Turkmen et al., 2017). In a study by Izquierdo et al. (2015), authors fed Gilthead Sea Bream broodstock diets with graded levels of fish oil (FO) replacement with linseed-oil (LO) (0, 60, 80 and 100%), evaluated spawning efficiency, and challenged the progeny with a low-FM and low-FO diet. It was found that feeding the broodstock with high-LO diets (80 and 100% replacement), significantly reduced fecundity, spawn quality and grow out of larvae/juveniles. However, when the progeny were 4 months old and challenged with a low-FM and low-FO diet, offspring from parents fed with the 60% LO replacement diet displayed better growth and feed utilization than offspring from parents fed with 0% LO. In a continuation of that study,

Turkmen et al. (2017) took the progeny after they were challenged with the low-FM, low-

FO diets and transitioned them all to commercial pellets for 11 months. When the fish were 16 months old, they were re-introduced to the same initial challenge diets and fed for 2 months. Results showed that parents that had been fed diets with 60% LO produced offspring that had significantly improved growth and feed utilization at 16 months after 62 being fed the low-FM, low FO diets when those diets were given previously as the first feed.

In aquaculture, a possible application of nutritional programming could be the development of fish better adapted to utilize dietary plant proteins. Of the many plant proteins, soybean meal has become one of the most promising fishmeal replacements.

The incorporation of soybean meal in fish diets can alleviate some of the sustainability and cost problems associated with fish meal use in aquafeeds (Food and Agriculture

Organization of the United Nations, 2016; Oliva-Teles et al., 2015). However, while soybean meal is an inexpensive alternative to fishmeal, it contains certain undesirable nutritional characteristics. These include high carbohydrate levels, the presence of isoflavones, low methionine and lysine levels and anti-nutritional factors (lectins, oligosaccharides, saponins and trypsin inhibitors) that may impede protein digestion and cause intestinal inflammatory responses that may ultimately hinder the growth of the fish

(Francis et al., 2001; Gatlin et al., 2007; National Research Council (NRC), 2011; Zhou et al., 2017). Thus, nutritional programming can be used as a tool to adapt carnivorous fish to higher inclusion levels of SBM within the diet without the incidence of reduced growth.

This study is a continuation of our previous experiment (Kemski et al., 2018) in which yellow perch were nutritionally programmed with various plant-based diets as their initial formulated feed (Phase 1), all transitioned to a FM-based diet for 9 months (Phase 2), challenged with a soybean meal (SBM)-based diet for 7 months (Phase 3) and

63 transitioned back the FM-based diet prior to spawning (Phase 4). Growth performance and reproductive efficiency parameters were examined, and results showed that after

Phase 3, that there were no differences among dietary groups in regards to weight gain percent (P=0.254). Additionally, no significant differences found after Phase 4 in regards to reproductive quality such as fecundity, egg size and fertilization rate (P=0.474). The goal of the present study was to use the nutritionally programmed broodstock fish (F0) from the aforementioned study and spawn them successively over the course of 3 years to determine the heritability potential of nutritional programming by bringing the offspring

(F1) through the same 3 Phases with identical diets (renamed as FM2 and SBM2) and compare growth performance. In year 1 (2015), nutritionally programmed females were crossed with pond-cultured males to determine maternal heritability. In year 2 and 3, both nutritionally programmed males and females from the same initial diet were crossed in order to compare the combined parental heritability. It was hypothesized that crossing both nutritionally programmed males and females, previously fed a SBM-based diet, would produce offspring that would have superior growth performance on the SBM2- based diet, when compared to those not programmed with SBM2.

3.3 Materials and Methods

3.3.1 Fish and experimental conditions

The experiment was conducted at The Ohio State University’s aquaculture facility in the

School of Environment and Natural Resources. All experimental methods and protocols were approved and performed according to The Ohio State University’s Institutional 64

Animal Care and Use Committee. City water supplied to the tanks was treated with activated charcoal filters and sodium thiosulfate was added to keep the level of chlorine below 0.01 mg/L. Dissolved oxygen and ammonia levels were kept at constant and safe levels (Hart et al., 2006; Kestemont et al., 2015). Broodstock fish were held in triplicate,

400L, semi-closed, recirculating tanks with an average of 43 fish per tank. Within each tank, fish were marked by passive integrated transponders-tags (PIT-tags; Biomark,

Boise, ID) to denote their initial dietary history; FM, SBM-A or Commercial Control

(CC).

3.3.2 Breeding of fish

The nutritionally programmed broodstock fish were fed FM-based diets for 6 months during gametogenesis (December-May) in order to not compromise possible impact on gamete quality affected by soybean phytoestrogens (Bennetau-Pelissero et al., 2001; Le

Boucher et al., 2012). A group of fish that served as control, termed the commercial control (CC) group, was also included for spawning comparisons, and fed only the CC diet. These fish were siblings of the nutritionally programmed fish, but were fed solely

Aquamax 400-500 commercial diet (Purina, Gray Summit, MO) throughout their lives.

Additionally, these fish were kept under identical environmental conditions to the nutritionally programmed fish, in a separate 400L tank. Water temperature in the tanks followed an annual seasonal cycle similar to what is found in the north central region of the US (Farmer et al., 2015). The photoperiod in the room was also manually altered to mimic natural seasonal light regimes to induce gametogenesis (Dabrowski et al., 1996).

In 2015, 8 nutritionally programmed females (4 females initially programmed on FM and

4 initially programmed on SBM) and 4 CC females were fertilized with combined sperm from outside control, pond-raised males (n=4). These males were 1+-year-old from Mill

Creek Perch Farm in Marysville, OH and had only been fed a commercial fish meal- based diet. In 2016, 6 nutritionally programmed females (3 FM and 3 SBM) and 3 CC females were fertilized with a mix of sperm from males with matching dietary history (4

FM, 4 SBM or 4 CC). In 2017, 6 nutritionally programmed females (3 FM and 3 SBM) and 3 CC females were fertilized with a mix of sperm from males with the same dietary history (4 FM, 4 SBM or 4 CC). For years 2016 and 2017, sperm from 4 males with the identical dietary history was combined, and the mix was used to fertilize all females from that same nutritional history. Thus, all females with FM as the initial programming diet were fertilized with a mixture of sperm from all 4 males from the FM nutritional history.

This was repeated in the same manner for the SBM and CC groups. Further details regarding the fertilization process and embryo incubation can be found in Kemski et al.

(2018).

3.3.3 Rearing of yellow perch larvae

The major achievement over recent years in our laboratory has been the ability to successfully culture larval yellow perch (Kwasek et al., 2012; Grayson et al., 2014). This is a significant advancement in yellow perch aquaculture because previous studies required the use of natural ponds and limited the possibility of replication of progenies from individual parents, and the number of fish batches available to be examined

(Malison et al., 1993a, b). When larvae in the present study were seen hatching and freely swimming, they were force hatched with the use siphon pressure to induce chorion disruption, and stocked in 10L aquaria at a rate of 100/L. Incoming water to each aquaria was supplemented with 3ppt salt (Instant Ocean, Spectrum Brands, Blacksburg, VA) and a constant supply of algae (Nanno 3600, Reed Mericulture, Campbell, CA), which kept the water turbid in order to reduce cannibalism. After 2 days, when all larvae were at the swim-up stage, and the yolk sack reserves depleted, exogenous feeding began with live rotifers (Brachionus plicatilis) every two hours. Larvae were fed with rotifers for 5 days until they reached an appropriate size for Artemia nauplii. Feeding with Artemia continued for approximately 30-45 days, or until juveniles reached at least 50-120mg

(depending on year) to ensure that their digestive tract was fully developed and they were readily accepting formulated feeds. Once metamorphosed juveniles reached the desired size, each set of offspring was divided into two groups of 50, weighed and stocked into

40L conical tanks to begin the feeding trial and were allowed to acclimate for 3 days prior to beginning feeding with formulated feeds. The feeding trial phases were the same nutritionally programmed phases as their parents (Kemski et al., 2018). Each of the following phases were replicated for all three offspring groups from each year (2015,

2016, and 2017). However, the 2017 offspring group was only taken through phase 1 due to time constraints, thus will only be compared until that point (Figure 1.1).

3.3.4 Phase 1

The only variation that occurred in Phase 1 among the three years were the specific

67 starting weight of the offspring groups. Groups from 2015 and 2016 were stocked at

0.057±0.01g and the groups from 2017 were stocked at 0.133±0.028g to determine if initial size of fish when introduced to formulated feed had a significant impact on growth and survival.

In Phase 1, juvenile yellow perch from each individual dam were divided into two groups of 50, weighed and stocked into 40 L conical tanks (n=32 in 2015 and n=24 in

2016/2017). Two diets were used: 1) a control diet with fish meal as the major protein source (FM2) and 2) soybean meal-A (Kemski et al., 2018) providing 75% replacement of fish meal protein (SBM2). Characteristics of SBM2 are provided in Baker and Stein

(2009) and Baker et al. (2010). Diets were those used in the previous study (Chapter 2), just re-named to differentiate parental diet from offspring diet. Diets were isonitrogenous and isoenergetic, and formulated as previously described (Kemski et al., 2018). During this first phase, fish were fed each diet, for 60 days from mid July to September.

Temperatures during this time ranged from 21.9°C-25°C, and fish were fed three times daily. Diets were given starting at a rate of 7% tank biomass, and re-adjusted daily, equally between tanks, based on tank with the lowest feed intake (feeding rates ranged from 5-7% tank biomass). Each tank’s biomass was measured every 30 days in order to correct and update the feeding rate. In between weighing, biomass increase was estimated daily, assuming a feed coefficient ratio (FCR; dry food(g)/wet weight gain(g)) of 1 (Kwasek et al., 2012). At the end of the phase specific growth rate (SGR = (Ln

(weight gain(g))/days)x100) and percent weight gain (weight gain % = final weight (g) - initial weight (g)/initial weight (g) x100) for each diet treatment was determined. During

68 handling and weighing, fish were anesthetized with MS-222 (Western Chemical,

Ferndale, WA; 50mg/L). Fish were fin-clipped at the end of phase 1 to denote initial diet

(FM2 or SBM2) and also tagged with a visible implant elastomer tags (VIE; Northwest

Marine Technology Inc., Shaw Island, WA) in different patterns over the caudal fin to identify their mother.

3.3.5 Phase 2

Prior to beginning Phase 2, approximately 20 siblings (10 initially fed FM2, and 10 initially fed SBM2) from 3 mothers of each dietary history (FM, SBM, CC) were combined among 3, 400L tanks to be raised in a “common garden” design (n=80 fish/tank). All fish were transitioned to the fish meal based-diet (FM2), and fed for 9 months. Feeding rate during this phase ranged from 2 to 5% biomass daily, and was adjusted in the same manner as previously described for Phase 1. Fish were weighed every 3 months in order to monitor performance and update and correct feeding rate. At the end of phase 2, fish were weighed and tagged with passive integrated transponders- tags (PIT- tags; Biomark, Boise, ID). Fish were PIT-tagged at this point as they were at a more appropriate size (Baras et al., 2000).

3.3.6 Phase 3

In Phase 3, also referred to as the “challenge” phase, all fish were transitioned to the

SBM2-based diet and maintained for 7 months. Fish were raised in an identical manner as Phase 2 and in the same “common garden” design. Feeding was done as previously

69 explained, and adjusted equally among tanks based on daily feed intake and assumed daily biomass increase. Tank weight was measured every 3 months to accurately correct feeding rate, which ranged from 2-5% tank biomass per day. At the end of the phase fish were individually weighed and growth performance was calculated.

3.3.7 Statistical analysis

Results are expressed as mean ± standard deviation (SD). Data were checked for normality, and weight gain and survival after each phase compared using the analysis of variance (ANOVA; PROC GLM SAS v9.3). Survival of fish during each phase was arcsine transformed for normality. A Tukey-Kramer test was also done for multiple comparisons of means. Weight gain (%) in the various phases was assessed by year, F0 diet, and F1 diet, by a multivariate analysis of variance (PROC MIXED SAS) to determine the possible effects and interactions of the variables. All data was analyzed using SAS software (SAS Institute Inc., Cary, NC) and a P value of <0.05 was considered significant.

3.4 Results

3.4.1 Phase 1 Results

2015: Fish were stocked at an average weight of 0.63±0.01g with the commercial control group being larger than average at 0.75±0.004g. At the end of phase 1, all of the SBM2 fed groups were smaller than the FM2 fed groups regardless of the mother diet

(P=0.0001). It is important to note that because of the larger starting weight of the CC 70 offspring group, both the FM2 and SBM2 fed juveniles had a larger final weight and weight gain than any of the other groups (P=0.0001) (Table 3.1). Overall, survival for this phase was not significantly different among any of the groups and was 77±12% (FM-

FM2), 59±19% (FM-SBM2), 76±23% (SBM-FM2), 59±5% (SBM-SBM2), 86±14%

(CC-FM2), and 61±16% (CC-SBM2) (P=0.0675).

2016: When stocking for Phase 1 in 2016, fish were much more uniform in size, with an average weight of 0.51±0.01g. Similar to 2015, at the end of the phase all of the groups that had been fed SBM2 as their initial diet were smaller than the FM2 fed groups, regardless of the mother (P=0.0022) (Table 3.1), which was also seen in weight gain and

SGR among all groups (P=0.0016). Survival for phase 1 was significantly different among the FM2 fed groups from all mothers at 91±7%a (FM-FM2), 85±5%a (SBM-

FM2), 87±7%a (CC-FM2) compared to 60±5%b seen in the SBM-SBM2 group.

Although the other two SBM fed groups were not significantly different among any of the other groups at 62±7%ab (FM-SBM2) and 60±8%ab (CC-SBM2) (P=0.001).

2017: Overall, fish were stocked to begin phase 1 at a much larger size (0.133±0.03g) in

2017 when compared to the previous two years (0.057±0.01g). This resulted in a larger relative weight gain, and SGR among all of the groups (P=0.02). Yet comparable with previous years, all of the SBM2 fed groups were smaller and had gained less weight compared to any of the FM2 fed groups regardless of the mother diet (P=0.0179) (Table

3.1). Survival during this year was improved in comparison to the previous years;

84±14% (FM-FM2), 75±8% (FM-SBM2), 89±11% (SBM-FM2), 76±6% (SBM-SBM2),

98±2% (CC-FM2) and 95±5% (CC-SBM2) and no significant differences were found among the groups (P=0.7216).

Figure 3.2 depicts the weight gain percent among the different offspring groups based on their initial diet in phase 1. All groups from 2015, regardless of the mother diet or initial diet were significantly lower than groups from 2016/2017 (P=0.01). The SBM2 fed groups from 2017 showed a higher weight gain percent than previous years. Mixed model results showed that the mother diet independent of other variables, had a significant effect (P<0.0001) on weight gain percent when compared among all years.

The offspring diet was also found to have a significant effect (P<0.001) on weight gain percent. However, the interaction between mother diet and offspring diet was found to not be significant (P=0.866).

3.4.2 Phase 2 Results

2015: Total weight gain and SGR during phase 2 in 2015 was significantly higher in the offspring group from the SBM mothers, regardless of their initial diet compared to offspring from the CC mothers (P<0.0001) (Table 3.2). Survival over this phase was not significantly different among groups (P=0.460) and was on average 91±5%.

2016: Fish from all groups in the second year had a larger final weight, weight gain and

SGR compared to 2015 (Table 3.2). However, in parallel with 2015, all of the SBM2 fed groups from 2016 were smaller on average than the FM2 fed groups (P=0.044), but no significant differences were found among groups for weight gain or SGR (P=0.120, and

P=0.127, respectively). Survival for this year was also similar to 2015 in that there were no significant differences among groups and the average survival for all groups was

89±5% (P=0.085).

The overall weight gain percent for phase 2 was comparable among all groups from both years with the exception of the CC offspring in 2015 (Figure 3.3). Similar to phase 1 results, the mother diet and F1 diet were found to have individual significant effects on weight gain percent (P<0.0001 and P=0.002, respectively). However, the interaction between mother diet and F1 diet was not found to have any significant effects on weight gain (P=0.884).

3.4.3 Phase 3 Results

2015: There were no significant differences found among any groups at the end of phase

3 in regards to final weight (P=0.773) or weight gain (P=0.065). Yet there were significant differences found among groups for weight gain percent (P<0.0001) and SGR

(P=0.03) (Table 3.3). Survival was similar among all groups with no significant differences found and an average of 87±5% (P=0.141).

2016: At the end of phase 3 in this year, there were no significant differences found for any of the growth parameters among any of the groups (P=0.106) and weight gain

(P=0.116) and SGR were all very similar among all diet treatments (P=0.157) (Table

3.3). Survival again was similar among all groups (94±3%) with no significant differences found (P=0.898).

Weight gain percent among all groups is displayed in Figure 3.4 was comparable between

2015 and 2016. Mixed model results that compared both years showed that only mother diet, individually, had a significant effect on weight gain percent (P=0.004), as well as the

F1 diet (P=0.029). However, the interaction of mother diet (F0) and F1 diet (P=0.773) were found to not have a significant effect on weight gain percent.

3.5 Discussion

3.5.1 Phase 1

Regardless of diet composition, fish from 2015 had the lowest weight gain percent when compared to those from 2016 and 2017 at the end of Phase 1. Offspring in 2015 were produced from only nutritionally programed females and outside control males, whereas in 2016/2017 offspring were from both nutritionally programmed males and females.

Offspring weight gain in 2015 was similar to that of their parents (F0), which was surprising due to the fact that the F0 generation began phase 1 at a much larger individual size (0.25±0.02g). In 2016/2017, while the SBM2 fed groups gained less weight overall, their size was much closer to fish fed FM2. This was a noteworthy improvement compared to growth of parental fish and offspring groups from 2015. Results from

2016/2017 are comparable to previous results found in Michl et al. (2017) who observed similar final body weights among groups after initial feeding of juvenile brown trout

(Salmo trutta) for 60 days with diets of either 50 or 90% fishmeal protein replacement with plant protein. However, it is interesting to note that even with larger starting and ending weights in 2017 in phase 1, overall weight gain percent that year was not 74 significantly higher than weight gain percent in 2016. What was affected between those two years was survival. Fish from 2017 with the larger starting weight had an overall higher survival (86±8%) than those in 2016 (75±7%) or 2015 (70±7%). It has been shown that the size of yellow perch when first introduced to formulated feeds is important, as acceptance of diets can be reduced if fish are too small or too large (Brown et al., 1996; Kestemont et al., 2015), which can affect weight gain and survival.

When comparing weight gain percent from all three years (Figure 3.2), results showed that independently, mother diet and offspring initial diet had a significant influence on weight gain (P=0.0001). However, the interaction of these factors, (mother diet (FO) vs. offspring diet (F1)) were found not to be significant, thus their effects on growth were individual. These results are consistent with Izquierdo et al. (2015), in which Gilthead sea bream (Sparus aurata) broodstock were fed diets with graded levels of fish oil (FO) replacement with linseed-oil (LO) (0, 60, 80 and 100%) and evaluated obtained offspring.

Progeny were challenged with a low-FM and low fish-oil (FO) diet when they were 4 months old, and results showed that offspring from parents fed with the 60% LO replacement diet showed better growth and feed utilization than offspring from parents fed with FO. The authors found that independent of each other, the broodstock diet and the juvenile diet had significant effects on body weight and SGR, yet the interaction of those two factors were not significant.

3.5.2 Phase 2

Weight gain percent for the duration of phase two was comparable between 2015 and

2016 with the SBM2 fed groups having gained more weight, with the exception of the

CC offspring group in 2015 that gained only a fraction of the weight when compared to other offspring groups during phase 2. Results of the SBM2 fed groups gaining a higher percent weight were similar to what were found in Kemski et al. (2018) when compared to the growth of the F0 generation, as well as in the study by Geurden et al. (2013). It should be noted that the SBM2 fed groups (Phase 1) exhibiting more weight gain during phase 2 could be a result of compensatory growth. A main difference to point out was that while these groups had a similar weight gain percent for the duration of phase 2, the final average weight was much higher in offspring from 2016 (16.6±4.6g) versus weight in 2015 (8.3±1.9g). This is interesting especially since starting weights were similar at

1.3±0.6g in 2015 and 1.9±0.6g in 2016 and treatments were the same between years.

3.5.3 Phase 3

Weight gain percent after phase 3 among all groups from both 2015 and 2016 was comparable. However, in continuation from phase 2 the average final weight among all of the fish was much larger in 2016 (69.6±12.5g) than in 2015 (35.0±4.0g). Interestingly, it was found that in 2015 the CC offspring group fed SBM2 showed a significantly higher weight gain percent than any other group. Results that compared both years exhibited that the F0 diet and F1 diet separately had a significant effect on weight gain percent

(P=0.004 and P=0.029, respectively). Consistent with the previous phases however, the 76 interaction of mother diet and F1 diet (P=0.773) were found to not have a significant effect on weight gain percent. These results were similar to Geurden et al. (2013), in which authors found that the initial plant-based programming diet did have a significant effect on growth performance when it was re-introduced to rainbow trout 7 months later.

However, our results were in contrast with Michl et al. (2017) in which brown trout were fed with three diets, a control and two plant protein replacements (50 or 90%) for 61 days. Each dietary group was then divided into three and fed one of the three experimental diets for another 48 days and growth performance followed based on initial diet. Authors did not find a significant effect of initial diet on growth performance among any groups with brown trout.

When the results of the current study are compared to growth results of their parents (F0) from our previous study, growth performance had greatly improved in the offspring groups during the challenge (Phase 3) when all of the fish had been fed SBM2 diets. In addition, results showed that the parental diet and offspring diet did, in fact, have a significant effect on weight gain percent in offspring groups, regardless of year (P=0.004 and P=0.029, respectively). While it is hypothesized that nutritional programming is the cause behind the increased weight gain in phase 3, there have been various ideas as to what may also be occurring. Milagro et al. (2013) showed that there are complex interactions between food components and epigenetic changes, which occur in mammals, most prevalently during the perinatal period, and play a large role in their adult metabolism. Another possible explanation for the programming effect is the flavor learning, or olfactory response (Balasubramanian et al., 2016). Fish fed the SBM diets

77 initially, can develop a memory to the flavor, and may be more willing to accept it when re-introduced to it again later in life.

Yúfera, 2016). Results from both studies showed that gene transcription was differently regulated in adult fish that had been nutritionally programmed early on as larvae.

Rainbow trout have also been used, with investigators examining the effects of fish meal replacement with plant proteins on growth performance and reproduction, respectively

(Geurden et al., 2013; Lazzarotto et al., 2015). Geurden et al. (2013) showed that adult rainbow trout that had been nutritionally programmed on a plant protein-based starter feeds as alevins, had a significantly higher feed intake, growth rate and feed utilization to that same diet as adults. This was seen after fish were fed a plant-based diet again after being fed a FM-based diet for a period of time, when compared to trout that had never been introduced to the plant-based diet. However, results from initial, “imprinting” phase in rainbow trout with plant-protein based diets showed that the fish only negligible growth, which was in contrast to the present study.

Results from the current study after phase 3 are promising, especially because the SBM2 diet replaces 75% of the FM protein with SBM protein. As previous studies have shown, yellow perch are sensitive to the anti-nutritional factors within SBM, thus, growth and performance is hindered. Kasper et al. (2007) examined the effect of SBM on yellow 78 perch juveniles (27g mean weight), which were fed diets for 8 weeks with increasing amounts (0-73%) of SBM replacement. The authors found an inverse response in weight gain, showing that groups with higher levels of SBM had reduced weight gain.

Specifically, in the diet formulation with 73% SBM, weight gain of fish was roughly one fifth that of the control fish meal (0% SBM) diet, with fish gaining 20% and 108% weight, respectively.

3.6 Conclusion

We found nutritional programming to be a significant way to adapt yellow perch to better utilize SBM-based diets as adults, and that their parental nutritional history does have a significant effect on offspring weight gain percent throughout their lives. It was found that offspring from the CC mothers in 2015, when fed a diet with 75% soybean protein replacement for fishmeal protein as their first formulated feed, and then re-introduced to the same diet 9 months later, had significantly higher weight gain percent than those previously fed FM2-based diets. These results are very promising as there are only a limited number of studies that have shown improved growth performance with SBM inclusion above 60% in diet formulations.

Table 3.1 Growth performance parameters of offspring groups (2015, 2016 and 2017) for

Phase 1. Values are presented as mean ± SD, and letters denote significant differences

within columns (P<0.05).

2015- Phase 1 F0 Diet F1 Diet Stocking Weight (g) End Phase 1 Weight (g) Ave. Weight Gain (g) Ave. Weight Gain (%) SGR FM FM2 0.059±0.015b 1.26±0.19b 1.20±0.19b 2138±542ab 0.29±0.25b FM SBM2 0.059±0.015b 0.85±0.19cd 0.79±0.19cd 1407±392bc 0.06±0.07c SBM FM2 0.056±0.006b 1.24±0.22bc 1.19±0.22bc 2175±552ab 0.27±0.29b SBM SBM2 0.056±0.006b 0.70±0.06d 0.64±0.06d 1177±227c 0.04±0.01c CC FM2 0.075±0.004a 2.41±0.33a 2.34±0.33a 3118±397a 1.40±0.22a CC SBM2 0.075±0.004a 1.53±0.20b 1.46±0.20b 1940±211bc 0.62±0.22b 2016-Phase 1 F0 Diet F1 Diet Stocking Weight (g) End Phase 1 Weight (g) Ave. Weight Gain (g) Ave. Weight Gain (%) SGR FM FM2 0.054±0.014 2.41±0.67a 2.36±0.66a 4385±148a 1.38±0.45a FM SBM2 0.054±0.014 1.75±0.70ab 1.70±0.69ab 3122±669ab 0.78±0.66ab SBM FM2 0.051±0.008 2.20±0.41a 2.15±0.40a 4281±683a 1.25±0.31a SBM SBM2 0.051±0.008 1.41±0.41b 1.36±0.40b 2665.9b 0.44±0.52b CC FM2 0.050±0.016 2.55±0.27a 2.50±0.28a 4336±1067a 1.52±0.19a CC SBM2 0.050±0.016 1.74±0.29ab 1.69±0.29ab 3006±667ab 0.85±0.30ab 2017-Phase 1 F0 Diet F1 Diet Stocking Weight (g) End Phase 1 Weight (g) Ave. Weight Gain (g) Ave. Weight Gain (%) SGR FM FM2 0.114±0.027 4.61±0.45a 4.50±0.42a 4032±603 2.50±0.16a FM SBM2 0.114±0.027 4.08±0.38ab 3.97±0.41ab 3641±1238 2.29±0.17ab SBM FM2 0.124±0.013 4.86±0.64ab 4.73±0.63ab 3828±90 2.58±0.22ab SBM SBM2 0.124±0.013 3.22±0.81ab 3.09±0.83ab 2556±948 1.85±0.45ab CC FM2 0.141±0.042 5.62±0.45ab 5.47±0.49ab 3829±1375 2.83±0.15ab CC SBM2 0.141±0.042 4.73±0.41b 4.57±0.38b 3113±553 2.53±0.14b

Table 3.2 Growth performance parameters of offspring groups (2015, 2016 and 2017) for

Phase 2. Values are presented as mean ± SD, and letters denote significant differences

within columns (P<0.05).

2015-Phase 2 F0 Diet F1 Diet End Phase 1 Weight (g) End Phase 2 Weight (g) Ave. Weight Gain (g) Ave. Weight Gain (%) SGR FM FM2 1.26±0.19b 8.77±1.06a 7.51±1.23ab 614±164bc 0.77±0.07a FM SBM2 0.85±0.19cd 7.77±0.99ab 6.91±1.12ab 850±248ab 0.74±0.06ab SBM FM2 1.24±0.22bc 11.06±2.49a 9.82±2.41a 795±191b 0.87±0.10a SBM SBM2 0.70±0.06d 8.88±0.83a 8.18±0.88a 1184±221a 0.81±0.04a

CC FM2 2.41±0.33a 7.55±0.67ab 5.13±0.38bc 214±19d 0.63±0.03bc

CC SBM2 1.53±0.20b 5.92±0.66b 4.39±0.49c 287±24cd 0.57±0.04c

2016-Phase 2 F0 Diet F1 Diet End Phase 1 Weight (g) End Phase 2 Weight (g) Ave. Weight Gain (g) Ave. Weight Gain (%) SGR FM FM2 2.41±0.67a 22.0±8.4a 18.98±6.48 799±205 1.16±0.13 FM SBM2 1.75±0.70ab 19.5±6.23ab 17.92±4.53 1090±295 1.15±0.10 SBM FM2 2.20±0.41a 15.2±2.7ab 13.87±3.28 661±241 1.04±0.09 SBM SBM2 1.41±0.41b 12.9±3.1b 11.95±3.12 945±460 0.98±0.11 CC FM2 2.55±0.27a 17.0±3.0ab 14.51±2.77 578±63 1.07±0.08

CC SBM2 1.74±0.29ab 15.3±3.0ab 13.56±2.76 768±89 1.04±0.08

Table 3.3 Growth performance parameters of offspring groups (2015, 2016 and 2017) for

Phase 3. Values are presented as mean ± SD, and letters denote significant differences

within columns (P<0.05).

2015-Phase 3 F0 Diet F1 Diet End Phase 2 Weight (g) End Phase 3 Weight (g) Ave. Weight Gain (g) Ave. Weight Gain (%) SGR FM FM2 8.77±1.06a 34.35±2.11 25.58±1.12 294±27bc 1.49±0.02 FM SBM2 7.77±0.99ab 36.40±3.82 28.64±2.94 371±29b 1.54±0.05 SBM FM2 11.06±2.49a 32.52±3.36 21.46±2.02 203±57c 1.40±0.04

SBM SBM2 8.88±0.83a 35.00±4.51 26.12±4.05 295±42bc 1.49±0.08 ab b CC FM2 7.55±0.67 36.62±4.46 28.86±3.90 371±42 1.54±0.06 b a CC SBM2 5.92±0.66 35.15±6.11 29.23±5.53 492±54 1.54±0.09 2016-Phase 3 F0 Diet F1 Diet End Phase 2 Weight (g) End Phase 3 Weight (g) Ave. Weight Gain (g) Ave. Weight Gain (%) SGR FM FM2 22.0±8.4a 72.5±15.7 50.5±7.4 243±57 1.80±0.06 FM SBM2 19.5±6.23ab 70.6±6.6 51.1±4.9 280±86 1.80±0.04 SBM FM2 15.2±2.7ab 66.8±7.6 51.6±8.7 355±109 1.80±0.08 SBM SBM2 12.9±3.1b 60.4±12.5 47.4±12.3 384±140 1.76±0.12 CC FM2 17.0±3.0ab 84.7±15.1 67.7±13.6 403±91 1.93±0.09 CC SBM2 15.3±3.0ab 74.8±8.8 59.5±5.8 394±43 1.87±0.04

Figure 3.1: Schematic view of the experimental design: FM, fishmeal-based diet; SBM,

75% FM replacement with SBM; CC, commercial control diet (no programming history).

FM2/SBM2 are identical diets to FM/SBM, yet were re-named to differentiate from parental history.

6000 F0 Diet: P<0.0001 F1 Diet: P<0.0001 2016 2017 F0xF1 Diet: P=0.866 5000

4000 2015

3000 FM2

2000 SBM2 Weight gain (%) Weight 1000

0 FM SBM CC FM SBM CC FM SBM CC Mother (F0) Diet

Figure 3.2: Weight gain percent after Phase 1 among the offspring groups from each year

(2015, 2016 and 2017) based on parental (F0) nutritional history and initial F1 diet fed during phase 1. Results are presented as mean ± SD (n=4) and were analyzed using a mixed model.

1600 2015 2016 F0 Diet: P<0.0001 1400 F1 Diet: P=0.0002 F0xF1 Diet: P=0.884 1200

1000

800 FM2 600 SBM2 Weight gain (%) Weight 400

200

0 FM SBM CC FM SBM CC Mother (F0) Diet

Figure 3.3: Weight gain percent after Phase 2 among the offspring groups from each year

(2015, 2016 and 2017) based on parental (F0) nutritional history and initial F1 diet fed during phase 1. Results are presented as mean ± SD (n=3) and were analyzed using a mixed model.

600 2015 F0 Diet: P=0.004 2016 F1 Diet: P=0.029 500 F0xF1 Diet: P=0.773

400

300 FM2

Weight gain (%) Weight 200 SBM2

100

0 FM SBM CC FM SBM CC Mother (F0) Diet

Figure 3.4: Overall weight gain percent after Phase 3 among the offspring groups from each year (2015, 2016 and 2017) based on mother (F0) nutritional history and initial F1 diet fed during phase 1. Results are presented as mean ± SD (n=3) and were analyzed using a mixed model.

Chapter 4: Transcriptomic response to soybean meal-based diets as the first formulated feed in juvenile yellow perch (Perca flavescens)

4.1 Abstract

With increasing levels of fish meal-based protein (FM) replacements in aquafeeds with soybean meal-based protein (SBM), a major topic of investigation for scientists is to further understand fish responses, and molecular mechanisms involved in response to new diets. Transcriptomic studies have been able to provide tissue specific, gene transcription patterns of the physiological responses of fish to diets with various FM replacements. Yet, to the best of our knowledge, there are no transcriptional studies in yellow perch regarding dietary ingredients and FM replacements. Thus, the goal of this study was to examine transcriptional responses through RNA-sequencing (RNA-seq) in the mid intestine of juvenile yellow perch after their first introduction to a formulated diet with 75% SBM for 60 days compared to those fed a traditional FM-based diet. RNA-seq results revealed that the majority of genes up-regulated in juveniles fed a SBM-based diet were highly involved in the cholesterol biosynthesis pathway, some being rate-limiting factors. Cholesterol levels in the SBM-based diet were significantly less (0.49%) versus the FM-based diet (1.59%), and lipid levels in the SBM diet were 13.6% compared to

15.6% in the FM diet. However, total lipid levels (FM: 11.01mg/g and SBM: 7.80 mg/g tissue) and cholesterol levels (FM: 2.23 mg/g and SBM: 2.72mg/g tissue) were not significantly different from each other. Cholesterol analysis validated the transcriptomic 87 results that yellow perch juveniles up-regulated their cholesterol biosynthesis pathway in response to SBM-based diets.

4.2 Introduction

In the United States and throughout the world, natural commercial fisheries stocks are threatened. Many commercial stocks are now fully exploited, overexploited, or depleted

(Food and Agriculture Organization of the United Nations, 2016). With the reduction in wild caught fish, the major approach for supply to be able to meet demand is through aquaculture production (Little et al., 2016). According to the FAO (2016), aquaculture production has increased from 1 million tons in 1950 to 81 million tons in 2014, making it one of the fastest growing forms of food production in the world. However, this trend cannot continue at its current rate, and significant improvements to the industry are necessary to make aquaculture a more sustainable and profitable production (Gentry et al., 2017; Little et al., 2016). As the growth of aquaculture continues, the industry cannot continue to use diets exclusively formulated with fish meal (FM) and fish oil (FO) due to the high cost and sustainability issues, thus more research is needed to find suitable replacements.

Of the many plant proteins, soybean protein in the form of soybean meal has become one of the most promising fish meal replacements. It has been widely used due to its high protein content, relatively well balanced amino acid profile and lower cost than fish meal

(Chou et al., 2004). The incorporation of soybean meal in fish diets could alleviate some of the sustainability and cost problems associated with fish meal use (Gatlin et al., 2007). 88

However, while soybean meal is an inexpensive alternative to fishmeal, it contains certain undesirable nutritional characteristics. These include high carbohydrate levels, the presence of isoflavones, low methionine levels and anti-nutritional factors (lectins, oligosaccharides, saponins and trypsin inhibitors) that have shown to impede protein digestion, impair immune response and cause intestinal inflammatory responses that may ultimately hinder the growth of the fish (Bakke-McKellep et al., 2008; Francis et al.,

2001; Gatlin et al., 2007). Yellow perch have proven to be quite sensitive to the anti- nutritional factors within soybean meal, and growth performance was hindered as a result of inclusion levels within the diet of 40% or higher (Kasper et al, 2007).

Research has shown that SBM and other plant proteins can be incorporated into adult

European seabass (Dicentrarchus labrax) diets at levels of 50% and higher without negative effects on growth, yet there are still limitations, specifically for juvenile fish

(Kaushik et al., 2004). For example, decreased growth was seen in juvenile Atlantic salmon (Salmo salar) and rainbow trout (Oncorhynchus mykiss) that had been fed diets with more than 50% fish meal replacement with plant proteins (Burr et al., 2012;

Torstensen et al., 2008), as well as in juvenile rainbow trout that were fed a completely plant-based diet (Panserat et al., 2009). Studies in our lab have also shown that juvenile yellow perch fed a diet with 75% FM replacement with soybean meal displayed significantly reduced growth when compared to those fed FM-based diets, yet adults fed the same diets had similar growth between the two groups (Kemski et al., 2018).

A major topic of investigation for scientists in the past few years has been to further understand fish responses, and molecular mechanisms involved in response to new diets

(Martin et al., 2016; Qian et al., 2014). Transcriptomic studies have been able to provide tissue specific, gene transcription patterns of the physiological responses of fish to diets with various FM replacements. There have been several publications examining transcriptional changes after FM is replaced with plant-based proteins in diets of rainbow trout or Atlantic salmon in various tissues; liver (De Santis et al., 2015; Jordal et al.,

2005; Morais et al., 2011; Panserat et al., 2009), whole body (Lazzarotto et al., 2016), brain (Balasubramanian et al., 2016) and intestine (Abernathy et al., 2017; Kortner et al.,

2012; Król et al., 2016; Morais et al., 2012a; Sahlmann et al., 2013; Tacchi et al., 2012).

Of the studies that investigated dietary SBM replacement on the transcriptome, Sahlmann et al. (2013) looked at initial responses in the distal intestine after feeding Atlantic salmon a diet with 20% SBM on day 1, 2, 3, 5 and 7. Results showed that the most prominent gene transcription changes were seen after day 3 and 5, in the immune-related transcripts with increased expression in GTPase IMAP family members, regulators of T- cell and B-cell function, and NF-κΒ-related genes. Fish sampled on day 5 and 7 showed down regulation of transcripts related to metabolic processes, endocytosis, exocytosis and detoxification, which indicated impairments to digestive and metabolic functions. In a longer term study, De Santis et al. (2015) fed Atlantic salmon graded levels of SBM (0,

10, 20 and 30%) for 12 weeks. Gene transcription was analyzed in the liver and distal intestine, and results from the intestine showed that of the fish fed the 30% SBM, they had an increase in expression of immune related responses (T-cell mediated processes,

TNF-α signaling pathways, NF-κΒ mediated responses), anti-inflammatory proteins

(annexin A1) and of pathways involving protein synthesis and cell proliferation, which authors indicated as being involved in regeneration to damaged intestinal tissue. Among those genes whose transcription was down regulated were those involved with metabolic pathways (lipid, sterol and vitamin metabolism, digestion and absorption), along with those in pathways associated with phagocytosis (cellular organelles, lysosomes and phagosomes). The authors were unsure if down regulation of these processes were caused by tissue damage, or the presence of anti-nutritional factors in SBM, that interfere with normal absorption of nutrients and are known to cause hypocholesterolemia (De

Santis et al., 2015; Kortner et al., 2014).

While limited in comparison to studies on other species of fish, there have been a few yellow perch transcriptomic studies, with the majority using microarray and examining exposure to contamination and heavy metals (Azizishirazi et al., 2014; Bélanger-

Deschênes et al., 2013; Bougas et al., 2013; Pierron et al., 2011). To the best of our knowledge, the only other de novo transcriptomic study that uses Illumina high- throughput sequencing on yellow perch was conducted by Li et al. (2017). In this study, gonad and muscle transcriptomes of sex-reversed males (testosterone treated) were compared to those of normal male and female yellow perch. Yet to the best of our knowledge there are no transcriptional studies in yellow perch in regards to dietary ingredients and FM replacements. Thus, the goal of this study was to examine transcriptional responses in the mid intestine of juvenile yellow perch after their first

91 introduction to a formulated diet with 75% SBM for 60 days compared to those fed a traditional FM-based diet.

4.3 Methods

4.3.1 Yellow perch feeding trial and sampling

Broodstock yellow perch were spawned in our laboratory in 2017, as previously explained in Chapter 3. However, we have chosen not to focus on the nutritional programming aspect in regards to transcriptomic response in this experiment, and solely focus on the expressional differences between juvenile yellow perch fed fish meal-based diets or soybean meal-based diets as their first formulated feed. After hatching, larval yellow perch were reared for 45-60 days on live feeds (rotifers then artemia nauplii) until reaching a stocking weight of 0.125±0.023g, at which time they were stocked among 12, conical tanks (40L), with 50 fish per tank. Juveniles were allowed to acclimate for 1 week remaining on live artemia, after which experimental diets were introduced (average body weight 0.133±0.028g). Two diets were used: 1) control diet with fish meal as the major protein source (FM2), 2) soybean meal (3011; Schillenger Genetics Inc.,), providing 75% replacement of fish meal protein (SBM2) (Baker et al., 2010; Baker and

Stein, 2009). Diets were formulated to be isonitrogenous and isoenergentic, and were produced using commercial manufacturing technology at the U.S. Fish &Wildlife Service

Bozeman Fish Technology Center (Bozeman, MT)(Table 4.1). Fish were fed each diet, for 60 days from mid July to September. Temperatures during this time ranged from

21.9°C-25°C and fish were fed three times daily. Diets were given starting at a rate of 92

7% tank biomass, and re-adjusted daily, equally between tanks, based on tank with the lowest feed intake (feeding rates ranged from 5-7% tank biomass). Each tank’s biomass was collected after 30 days in order to correct and update the feeding rate. In between weighing’s, biomass increase was estimated daily, assuming an FCR of 1 (Kwasek et al.,

2012). At the end of the feeding trial, survival, specific growth rate (SGR = (Ln (weight gain(g))/days)x100) and percent weight gain (weight gain % = final weight (g) - initial weight (g)/initial weight (g) x100) for each diet treatment were determined.

On the 61st day, fish were fed twice (8am and 11am) to apparent satiation and sampled 3 hours following the final meal to ensure that food was present throughout the intestinal tract. At the time of sampling, fish were euthanized by an overdose of MS-222;

250mg/L), weighed and tissues were excised and snap frozen in liquid nitrogen and stored at -80°C for RNA analysis or proximate composition/cholesterol analysis. Twelve fish were sampled for proximate composition, 6 FM2 fed juveniles and 6 SBM2 fed juveniles. For RNA analysis, the intestinal tract was removed and a 5mm portion of the mid-intestine from each individual was gently flushed with 5% saline to remove intestinal contents. A total of 12 fish were sampled for RNA; 6 FM2 fed juveniles and 6 SBM2 fed juveniles.

4.3.2 RNA Sequencing

Total RNA was isolated from each sample (50mg) using the Norgen Biotek Animal

Tissue RNA purification kit (Norgen Biotek Corp.,Thorold, ON, Canada) according to the manufacturer’s protocol. Sample concentrations were quantified using an Epoch 93

Microplate Spectrophotometer (BioTek, Winooski, VT) by measuring the concentration at 230, 260 and 280nm and their ratios. 2µl RNA was also run on a 2% agarose TAE gel in order to determine integrity of 18/28S bands and if contaminating DNA was present.

Half of all of the RNA samples (~33ug of RNA) were sent on dry ice to Admera Health,

LLC (South Plainfield, NJ) for further processing, the remaining samples were stored at -

80°C until further qPCR analysis. RNA integrity was determined using a bioanalyzer prior to library preparation. Integrity numbers (RIN) of >8 were considered high-quality, and all samples had a RIN of 8.7 and above.

4.3.3 Library Preparation and De Novo Transcriptome Assembly

Library preparation, mRNA sequencing/annotation and gene expression analysis was completed by Admera Health, LLC (South Plainfield, NJ). For the creation of cDNA libraries, paramagnetic beads coupled with oligo d(T) were combined with total RNA to isolate poly(A)+ transcripts based on the NEBNext® Poly(A) mRNA Magnetic Isolation

Module protocol. Prior to first strand synthesis, samples were randomly primed (5’ d(N6) 3’ [N=A, C, G, T]) and fragmented based on manufacturer’s recommendations

(NEBext®Ultra™ RNA Library Prep Kit for Illumina®). First strand synthesis was completed using the Protoscript II Reverse Transcriptase with a longer extension period

(40 minutes at 42°C). All remaining steps for library construction were used according to the NEBext®Ultra™ RNA Library Prep Kit for Illumina®. Illumina 8-nt dual-indices were used. Samples were pooled and sequenced on a HiSeq with a read length configuration of 150 paired-end reads (PE).

Prior to transcriptome assembly, pre-processing steps were performed, including a quality check (FastQC), removal of adapter content, quality thresholds and removal of poor quality reads (Phred Score < 30). Quality RNAseq data was then assembled de novo without the use of a genome, and the Trinity assembler (Haas et al., 2013) was used to assemble the reads into larger length transcripts (Honaas et al., 2016). The assembly was completed by taking the initial quality reads and constructing a k-mer dictionary with all reads having a k-mer length of 25. As another quality control step, the low-complexity and singleton k-mers, along with potential error-prone k-mers were removed. Within the k-mer dictionary, the most frequent k-mer was used as a seed for the first contig assembly, where it was extended in each direction by finding the next closest k-mer with a k-1 overlap. The extension procedure was repeated in both directions until it could not be taken further and the longest linear contig was reported as one expressed gene. These steps were repeated, continuing with the next most abundant k-mer, creating the rest of the contigs, until the k-mer dictionary was exhausted, and all possible contigs were formed, which created the assembled transcriptome.

4.3.4 Differential expression analysis

Differential expression analysis was calculated using the normalized gene expression

(FPKM) of each annotated gene, and done using the Tuxedo suite workflow (Trapnell et al., 2012), utilizing Tophat, Cuffdiff and Cuffnorm. Repetitive sequences were removed along with any gene/transcript with an average FPKM value of < 2.0 (Tasnim et al.,

2015). Fish that had been fed FM2 were used as the reference transcriptome, and data

95 from the SBM2 fed fish were compared to it. Genes were considered significant if the p- value was less than 0.05 between the FM2 and SBM2 fed fish. Transcripts with significantly different expression between the two conditions were selected for gene annotation and characterization. Up-regulated genes were annotated manually by

BLAST search against the NCBI reference proteins.

4.3.5 qPCR Validation

Primers for qPCR were designed using the NCBI Primer-BLAST tool, with gene sequences taken from the annotated transcriptome. Genes of interest were chosen based on the top results from the significantly different gene expression list that had also shown consistent expression levels among replicates with the same diet history (Table 4.2).

Housekeeping genes were also determined based on consistent expression levels seen among all samples regardless of diet history in RNA-seq results, and previous literature

(Grasset et al., 2014; Gu et al., 2014; Kortner et al., 2011) (Table 4.2).

Total RNA was reverse transcribed using iScript Reverse Transcription Supermix for RT- qPCR (Bio-Rad, Hercules, CA) according to the manufacturer’s protocol. Briefly, 200ng total RNA was combined with the iScript Supermix and incubated in a thermocycler using the following protocol; 5 minutes at 25°C, 20 minutes at 46°C for reverse transcription, and 1 minute at 95°C. Relative PCR and restriction fragment length polymorphism (RFLP) were done to ensure that primers generated a single amplicon of predicted size and that the amplicon was of the correct sequence. For relative PCR, cDNA templates were used at a 1:25 concentration in a standard 50µl PCR master mix

96 with Platinum Taq (Thermo Scientific, Waltham, MA) and 1 µl of each gene-specific primer. PCR products were amplified for 35 cycles starting at 95°C for 3 ½ minutes,

55°C for 30 seconds, and 72°C for 1 minute using a Biorad T100 thermal cycler (Biorad,

Hercules, CA). For RFLP, 10ul of the PCR products were incubated overnight at 37°C with sequence specific NEB restriction enzymes (New England Biolabs Inc., Ipswich,

MA) and compatible buffer, and visualized by agarose gel electrophoresis. Each PCR product with cut with 4 different restriction enzymes, to ensure that the predicted sequence matched the amplicon generated.

PCR efficiency for each primer pair was determined using 10-fold serial dilutions of pooled CDNA. Subsequent RT-qPCR amplifications were carried out in duplicate along with no a template control for each primer pair, using a CFX382 Touch™ Real Time

PCR Detection System (Bio-Rad, Hercules, CA). Each 10µl reaction mixture contained

2µl CDNA at a 1:10 dilution, 2µl PCR-grade water, 0.5µl each primer (500nM final concentration), and 5µl SsoAdvanced ™ Universal SYBR® Green Supermix (Bio-Rad,

Hercules, CA). The following program was used for amplification; 95°C for 2 minutes, followed by 39 cycles of 95°C for 10 sec, 60°C for 30 sec, 72°C for 30 sec and a final extension time at 72°C for 5 minutes. Melt curves for each PCR product were analyzed following the run to ensure that only a single product had been amplified. Housekeeping genes were evaluated by ranking the relative gene expression based on overall CV of each gene as previously described (Novak et al., (2002), de Jonge et al., (2007) and

Kortner 2012), along with the geNorm of each sample according to Vandesompele et al.,

(2002). These methods measured the stability of each reference gene by calculating pair-

97 wise variations among standard error or among other reference genes (McCurley et al.,

2008).

4.3.6 Cholesterol Analysis

A total lipid extraction was done on whole body samples according to the Folch (1956) method. The lipid fraction was then used to determine total cholesterol content among samples using the Cholesterol Reagent Kit (Pointe Scientific, Canton, MI), with slight modifications. Briefly, 500µl of chloroform/Triton X-100 mixture (100:1 v/v) was added to the dried lipid samples, they were mixed and dried again under nitrogen gas. Samples were then suspended in 250µl water, and shaken in a 37°C water bath for 15 minutes.

Subsequent steps were carried out according to manufacturer’s instructions, and absorbance was read in a Synergy H1 Hybrid Multi-Mode spectrophotometer (BioTek,

Winooski, VT) at 37°C and 500nm.

4.3.7 Statistical Analysis

Results are expressed as mean ± standard deviation (SD). Data were checked for normality, and weight gain, survival and lipid/cholesterol data were compared using an analysis of variance with a pooled t-test (ANOVA; PROC GLM SAS v9.3). Survival of fish during each phase was arcsine transformed for normality.

4.4 Results

4.4.1 Feeding Trial

Survival and body weight data for juveniles after the 61-day feeding trial are presented in

Table 4.1 and Figure 4.1, respectively. At the end of the trial, significantly lower weight gain (P=0.022) and SGR (P=0.03) were seen in the SBM2 fed group compared to the

FM2 fed group. However, no significant differences were seen between the groups in terms of survival (P=0.289) and weight gain percent (P=0.106). While fish weight is a convenient measurement of growth, fish growth is a combination of many factors. For example, different tissue accumulation (e.g., fat versus muscle) could result in different sized fish but with similar weights. Therefore, it is worth noting that SBM-fed fish had extensive visceral fat surrounding the intestinal tract as well as more distended stomachs that were not observed in FM-fed fish (data not shown).

4.4.2 Transcriptomic responses

All responses of the intestinal transcriptome to the SBM2 diet were expressed relative to the FM2 diet, with up and down-regulation referring to the changes in gene expression in juveniles fed the soybean meal based protein diets than in fish fed the fish meal based diet, respectively. A total of 15,368 differentially expressed genes were found between the two dietary treatment groups. There were a total of 7,613 genes up-regulated in the

SBM-fed fish and 6,761 down-regulated in the SBM-fed fish in comparison to the FM- fed fish. After the removal of genes that were not found to be significant between treatments (P<0.05), 1,800 significant differentially expressed genes remained. Of those, 99 only 650 genes had expression differences with fold changes of ±3 or greater (log2 (fold change) of ±1.5 or greater). Finally, a group of 65 top genes were selected based on the previous parameters and consistent expression levels among all replicates.

4.4.3 Up-regulation of genes found in SBM fed fish

Of the genes 40 that were found to be up-regulated in juveniles fed a SBM-based diet as their initial feed, those involved in the cholesterol biosynthesis pathway were the most enriched. There were 9 genes that were found to play integral roles in the cholesterol synthesis pathway (Figure 4.2). Two of which are known to be rate limiting in both the cholesterol biosynthesis pathway (3-hydroxy- 3-methylglutaryl reductase; HMGCR) as well as in bile acid production (Cytochrome P4507A1; CYP7A1). Also, up-regulated in fish fed the SBM-based diet were genes involved in lipid metabolism and transport, such as long chain fatty acid elongase and proprotein convertase subtilisin/kexin-type 9, which are both involved in LDL binding, the lipid metabolic process as well as cholesterol homeostasis (Table 4.2). Genes involved in vitamin B12 binding and transport were also found to be up-regulated in the SBM fed fish.

4.4.4 Down-regulation of genes found in SBM fed fish

The majority of genes that were down-regulated in fish fed SBM-based diets were those involved in the negative regulation of the immune response, such as NACHT, LRR and

PYD domains-containing protein 1. There were also genes involved in negative regulation of cell proliferation and the apoptotic process (Table 4.3). Other down-

100 regulated genes were those involved in the regulation of proteins, such as proteasome subunit beta2, alpha 2 macroglobulin and synaptogyrin 1-like, which are involved in protein deubiquitination, negative regulation of peptidase activity and protein targeting, respectively. Additionally, genes involved in lipid regulation such as adenosine receptor

A1-like that plays a role in fatty acid homeostasis, Mid1 interacting protein that regulates the lipid biosynthesis pathway and peroxisomal 2,4 dienoyl CoA reductase that is involved in the unsaturated fatty acid biosynthetic pathway were all down-regulated.

Finally, in the SBM fed fish there were also genes down-regulated that were involved in

DNA/RNA/GTPase binding, such as OTX 1B homeobox protein, DLX3 homeobox protein and DNA binding protein subunit B (Table 4.2).

4.4.5 qPCR Validation

Quantitative real-time PCR analysis was performed on 11 genes, both up and down regulated in the mid intestine of fish fed SBM-based diets compared to those fed FM- based diets for 60 days (Table 4.3). Expression levels were normalized to RNA-

Polymerase (RNApol), as this deemed to be the most stable gene based on both qPCR results as well as RNA-seq results (Figure 4.3). RT-qPCR results, presented as the mean of the Log2 (fold change) were found to be consistent with RNA-seq expression levels.

In contrast to RNA-seq results, the proteasome subunit beta 2 (PSMB2) and Nod-like receptor family CARD domain-containing 3 (NLRC3) genes tested by RT-qPCR were found to not be significantly different between treatments. Yet the remaining 9 genes

101 tested by RT-qPCR were found to be significantly different between dietary treatments, which were in agreement with RNA-seq results.

4.4.6 Whole body lipids and cholesterol analysis

Whole body lipid analysis revealed that the FM fed fish had a total lipid content of

11.01±1.6%, whereas SBM fed fish had a total lipid content of 7.82±1.9%, which were not significantly different from each other (Figure 4.4). Additionally, cholesterol content in the whole body in fish fed the SBM-based diet was 2.72±0.70 mg/g, compared to

2.23±0.89 mg/g cholesterol in fish fed the FM based diet. Again, no significant differences were found between groups.

4.5 Discussion

As growth of the aquaculture industry continues, the need for a cost effective, sustainable fish meal replacement is imperative. By having a better understanding of the molecular mechanisms and gene expression changes that occur when plant-based proteins are introduced into piscivorous fish diets, it allows researchers to create more optimal feeds that are better suited for these fish. There have been many studies that examined the transcriptional response of fish to FM replacements with plant-proteins (Balasubramanian et al., 2016; Morais et al., 2012a; Tacchi et al., 2012), with SBM replacements (De Santis et al., 2015; Marjara et al., 2012), various mixtures of SBM and other plant based protein sources (Król et al., 2016), and even complete replacement of fish meal/fish oil with plant sources (Lazzarotto et al., 2018; Panserat et al., 2009). However, the majority of the

102 aforementioned studies analyzed transcriptional changes within the liver or distal intestine, also reviewed by Martin et al., (Martin et al., 2016), and to the best of our knowledge there have been no previous studies analyzing yellow perch.

4.5.1 Feeding Trial

Results of the feeding trial showed that after 60 days, juveniles fed the SBM2 diet had lower average final weight when compared to fish fed the FM2 diet. However, in terms of weight gain percent, the SBM2 fed group was not significantly different from the FM fed fish. It is important to note that when compared to juvenile growth performance results found in Kemski et al. (2018), as well as results from Chapter 3 (juvenile growth performance from 2015 and 2016), the current overall growth performance results are significantly improved. In terms of final weight after the 60-day feeding trial, there was only a 1g (20%) difference between the two groups (FM2 5.11±0.63g; SBM2

4.11±0.81g), whereas in our earlier study (Kemski et al., 2018), there was a difference of more than 3g (62%) between the two groups (FM 5.00±1.83g; SBM 1.89±0.67g). This indicates that through rearing improvements in our lab, as well as a potential nutritional programming affect through selective breeding (see Chapter 3), we were able to improve the growth performance of juvenile yellow perch when fed a 75% SBM-based diet.

These results are consistent with other studies in which fish meal proteins are replaced at high levels with plant based proteins and fed to juvenile Coho salmon (Oncorhynchus kisutch) (Twibell et al., 2012), Atlantic salmon and rainbow trout (Burr et al., 2012;

Panserat et al., 2009; Torstensen et al., 2008). This was also seen in a study by

103

Lazzarotto et al. (Lazzarotto et al., 2016) in which authors examined rainbow trout broodstock nutritional history on growth of progeny fed diets with 46% fish meal replacement with plant proteins (C), or a diet completely void of fish meal (V) for 3 weeks. Results showed that maternal nutritional history did not have a significant effect on weight gain. In addition, juveniles fed diet V had significantly lower weight gain when compared to diet C, and both diets showed significantly lower weight gain than juveniles fed a control (FM) based diet. However, use of weight gain as a measure of growth and health should be used with caution as weight is the sum of many components and the body composition (e.g., tissue sizes, fat versus muscle, etc.) may be dramatically different but yield a similar overall weight.

4.5.2 Transcriptional changes in the mid-intestine

The intestine plays a vital role in digestion and absorption of nutrients and thus is highly sensitive to dietary modifications, especially in carnivorous fish. In a study by Tacchi et al. (2012), authors examined transcriptional changes in the mid intestine, liver and skeletal muscle of Atlantic salmon after being fed a plant-based diet, and found the greatest transcriptomic response in the intestine of all the tissues studied, reflecting its sensitivity to dietary changes. The intestine is divided into three discrete regions; the proximal region that also includes the pyloric caeca, the mid and distal intestine. While nutrient absorption occurs in all regions, it is highest in the proximal/mid region (Collie and Ferraris, 1995). Previous studies have shown that the distal intestine is the main region for enteritis due to SBM inclusion in the diet (Baeverfjord and Krogdahl, 1996;

104

Francis et al., 2001; Kaushik et al., 1995; Marjara et al., 2012). Therefore, the focus of this experiment was on transcriptional changes in response to a SBM-based diet in the mid intestine, so that metabolic changes could be highlighted rather than solely enteritis.

Cholesterol biosynthesis: The majority of genes up-regulated in juveniles fed the SBM2 diet were found to be involved in the cholesterol biosynthesis pathway (Figure 4.2). It is well known that cholesterol is only found in animal products, and that dietary replacements with plant proteins will have reduced levels of cholesterol. However, with higher inclusion levels of plant proteins, naturally occurring plant phytosterols increase.

Research has shown that cholesterol intake in the intestinal lumen may even be impaired by the presence of phytosterols (Francis et al., 2001; Kortner et al., 2014; Krogdahl et al.,

2010), and hyocholesterolemia can occur (Kortner et al., 2013). Fish are able to synthesize cholesterol de novo from acetate, yet the potential for synthesis can vary and is not well defined in fish (Deng et al., 2013). Cholesterol homeostasis is regulated by cholesterol intake, and achieved by the biosynthesis pathway, a rate limiting reaction, as well as conversion and excretion of bile acids. When intakes of cholesterol are low, the biosynthesis pathway takes over and is regulated by 3-hydroxy- 3-methylglutaryl reductase (HMGR), a rate-controlling enzyme in the mevalonate pathway, which results in the production of cholesterol (Brown and Goldstein, 1997). Another rate limiting enzyme important in the removal of cholesterol is Cytochrome P4507A1 (CYP7A1). Its activity is regulated by the flux of cholesterol through the liver, and when levels are high its activation allows for the conversion of cholesterol to bile acids (Chiang, 2011).

Consequently, circulating cholesterol levels in the body and bile acid metabolism are 105 closely linked, which is not surprising that both of these genes were found to be up- regulated in fish fed the SBM2 diet, along with other genes that are all involved in the cholesterol biosynthesis pathway and homeostasis. This is in agreement with other studies in which HMGCR and CYP7A1 up-regulated in response to non-FM based diets in Atlantic salmon (Kortner et al., 2013), rainbow trout (Lazzarotto et al., 2018, 2016),

European seabass (Geay et al., 2011) and yellowtail (Maita et al., 2006).

It has been hypothesized that these reduced cholesterol levels can cause a reduction in growth performance in fish, and studies have reported significantly improved growth when cholesterol was supplemented in diets that contained high levels of SBM in catfish

(Twibell and Wilson, 2004); turbot (Yun et al., 2012, 2011); and rainbow trout (Deng et al., 2013). However, in a study by Kortner et al., (2014), cholesterol was supplemented

(1.5%) into plant based diets fed to Atlantic salmon for 77 days. Growth and transcriptional responses from both the pyloric caeca and the liver were analyzed and compared among fish fed the same plant based diets without supplementation. Results showed that dietary supplementation of cholesterol did not significantly affect growth between groups. It did however; decrease expression of genes involved in the cholesterol biosynthesis pathway, suggesting that by supplementing dietary cholesterol, metabolic functions are altered since there is less of a need to synthesize cholesterol de novo.

Metabolic processes: Genes involved in lipid metabolism and transport were also up regulated in fish fed the SBM2 diet, even though the diet had similar lipid levels as the

FM2 diet. The up-regulation of these genes could be explained by the fact that the

106 intestine is not only involved in the absorption of lipids, but in significant hydrolysis of lipids to fatty acids and monoglycerides (Kestemont et al., 2015; Morais et al., 2012a). In addition, it is well known that lipid and cholesterol transport are closely related and specific genes that were found up-regulated, such as proprotein convertase subtilisin/kexin-type (PCSK9) and low-density lipoprotein receptor-related protein 2-like

(LRP2) are involved in LDL binding as well as cholesterol homeostasis. This transcriptomic up-regulation of lipid transport/metabolism in connection with cholesterol homeostasis was also seen in studies by Krol et al. (2016) and Tacchi et al. (2012) after

Atlantic salmon had been fed plant-based diets. These results suggest that regulatory mechanisms exist in the intestine, as well as the liver of fish fed plant-based diets in order for them to compensate for reduced cholesterol levels and lipid export rate (Gu et al.,

2014).

Cellular response/defense to damage: While cellular response/defense against DNA damage, interleukins, T-cells, and interferons were up regulated in the SBM-fed fish, it was also discovered that many genes involved in the negative regulation of cellular defense and inflammation were down regulated. These results are consistent with a study by Tacchi et al. (2012), who also found that numerous genes involved in the inflammatory response were both up and down regulated in the mid intestine of Atlantic salmon after being fed a plant based diet for 77 days. The authors identified genes that were up regulated were those involved in the NF-kB pathway, which results in an increase in pro-inflammatory cytokines, as well as SOCS-7 known to aid in inflammation reduction. Both IL-8 and IL-17D were found to be down regulated in fish fed the plant 107 based diet, which are involved in inflammatory/allergic responses, possibly suggesting that the reaction to the plant based diet may not be an allergic response. Results from our study showed that GTPase IMAP family member 8-like (GIMAP8), NACHT, LRR and

PYD domains-containing protein 1 (NLRP1), NOD-like receptor family CARD domain containing 3 (NLRC3) and C1q / TNF-related protein 3 (C1QTNF3) all function in the negative regulation of the inflammatory response, and were found to all be down- regulated in the SBM2 fed fish. GTPase IMAP genes have previously been associated with the differentiation of T-helper 2 cells, which exerts an anti-apoptotic effect in the immune system and are involved in responses to infections (Sahlmann et al., 2013). The down-regulation of these genes indicates that there is an increase in the inflammatory response within these fish and could suggest a dysfunction within the immune response

(Kortner et al., 2012). These findings also overlap with the immune response in inflammatory bowel disease in humans (Maloy and Powrie, 2011). As well as with several previous studies that have examined the transcriptomic changes within the distal intestine of fish in response to plant proteins, and found numerous pro-inflammatory genes up regulated (De Santis et al., 2015; Marjara et al., 2012; Sahlmann et al., 2013).

4.5.3 Lipid and cholesterol analysis

Overall lipid content in the body of these fish was found to be comparable between the

SBM2 fed groups and the FM2 fed groups after 60 days of feeding, with no significant differences found. Additionally, whole body cholesterol content also showed no significant differences in fish fed the SBM2 diets versus the FM2 diets. Results found in

108 the current study are in accordance to those found in studies by Deng et al. (2013) and

Zhu et al. (2018), in which both authors examined the response in rainbow trout after being fed diets with various levels of cholesterol and plant proteins. It should be noted however, that in both of the previously mentioned studies, results were comparable to our study when plant proteins were supplemented up to a certain point in the diet. The authors did find that whole body cholesterol content was significantly reduced when rainbow trout were fed completely marine free diets (Zhu et al., 2018), or those completely void of cholesterol (Deng et al., 2013). Interestingly, there have been numerous studies in which plasma cholesterol content was significantly reduced after FM was replaced with plant proteins in rainbow trout (Kaushik et al., 1995; Zhu et al., 2018),

European seabass (Dias et al., 2005), and Atlantic salmon (Kortner et al., 2014). This occurrence might be explained by the fact that excess cholesterol in the body is stored as cholesterol esters in cytoplasmic lipid droplets (Chang et al., 2009), because of the reduced cholesterol content in the plant based diets, levels in the plasma remain low, causing an increase in cholesterol biosynthesis, thus resulting in cholesterol accumulation through lipid droplets (Gu et al., 2014). On the other hand, results from present study found that cholesterol synthesis was increased to compensate for cholesterol deficiency in

SBM-fed diets, in order to keep fish at homeostatic levels of cholesterol explained by our observed similar lipid and cholesterol levels seen between groups.

109

4.6 Conclusions

In conclusion, results from the present study validates that FM protein can be replaced in juvenile yellow perch diets by up to 75% SBM protein without causing in a significant reduction in weight gain percent, or survival and only a slight reduction in terms of final weight. It is also promising that both total body lipid levels and cholesterol levels were not statistically different between fish fed the SBM2 diet and those fed the FM2 diet, especially since dietary levels were not equal. This is one of the first studies of its kind to examine the transcriptional response by RNA-seq in the mid intestine of juvenile yellow perch after being fed a SBM-based diet. It is also promising that juvenile yellow perch were capable of responding to dietary changes, especially in regards to cholesterol, and that they were able to regulate their cholesterol biosynthesis and other metabolic pathways to adapt within the first 60 days of introduction to a formulated feed.

Overall, findings produced from this study provides a better understanding of the mechanisms and pathways expressed when FM is replaced with SBM and given as the first formulated feed to juvenile yellow perch. These results offer the basis for additional research on improved diets as well as possible transcriptional differences between parental nutritional history and progeny response to SBM based diets. It also brings us one step closer to using SBM and other, more sustainable plant-based protein sources at higher inclusion levels in aquafeeds, thereby continuing to reduce the reliance on FM.

110

Table 4.1: Dietary ingredients and composition of FM2 and SBM2 diets Ingredient FM2 (g) SBM2 (g) Menhaden Fish Meal1 54.09 12.00 Soybean Meal (3011)2 - 52.86 Wheat Flour 30.15 14.96 Fish Oil 4.34 8.00 Soybean oil 1.98 0.24 Lecithin 3.00 3.00 CPSP 4.00 4.00 Vitamin mix3 1.00 1.00 Mineral mix4 0.10 0.10 Vitamin C 0.10 0.10 Choline Chloride 0.06 0.06 Calcium Phosphate - 1.00 Lysine - 1.50 Methionine 0.33 0.33 Arginine 0.65 0.65 Threonine 0.20 0.20 TOTAL 100.00 100.00 Composition (% DM) Dry Matter 91.7 94.35 Crude Protein 42.4 43.9 Crude Lipid 15.6 13.6 Ash 10.7 6.7 Cholesterol 1.59 0.49 159.03% Protein 258.9% Protein 3 Provided by U.S. Fish &Wildlife Service Bozeman Fish Technology Center; Batch 702

4Provided by U.S. Fish &Wildlife Service Bozeman Fish Technology Center; Batch 1520

111

Table 4.2: Growth performance results (± SD) after the 61 day feeding trial in which juvenile yellow perch were fed either a fish meal based diet (FM2) or a soybean meal- based diet (SBM2). Values within a column with * denote significant (P<0.05) differences between groups.

Stocking 60 Days Weight Weight gain Survival Diet SGR Weight Weight Gain (g) (%) (%) FM2 0.13±0.03 5.11±0.63* 4.98±0.62* 3887±838 2.66±0.21* 91±5 SBM2 0.13±0.03 4.11±0.81 3.98±0.79 3105±839 2.27±0.37 84±5

5000 4500 4000 3500 3000 2500 2000 1500 Weight gain (%) Weight 1000 500 0 FM2 SBM2 Diet

Figure 4.1: Weight gain (%) of juvenile yellow perch after being fed either a FM2 or

SBM2 diet for 61 days. Data are presented as mean ± standard deviation. No significant differences were found between the two groups (P=0.106).

112

Gene Fold Change (up-regulation in Acetyl-CoA + SBM fed fish) Acetoacetyl-CoA 4.9 HMG-CoA Synthase

HMG-CoA 5.6

HMG-CoA Reductase

Mevalonate

Mevalonate diphosphase 4.3 decarboxylate

Farnesyl PP

Squalene synthase 6.41

Squalene

Squalene monooxygenase 8.8

Lanosterol Lanosterol 14-alpha 4.1 demethylase 4.0 Emoparnil binding protein Cholesterol

Cholesterol 7-alpha 5.14 monooxygenase 6.35 Methylsterol monooxygenase

Figure 4.2: Cholesterol biosynthesis pathway and top differentially expressed genes.

The major metabolic intermediates are presented in gray boxes next to expressed genes and their expression level. Values are represented as the mean fold change between the

SBM2 fed group and FM2 fed fish identified through RNA-seq. HMG-CoA: 3-hydroxy-

3methylglutaryl-CoA; Farnesyl PP: Farnesyl pyrophosphate.

113

5.0 * 4.0 * 3.0 * * * 2.0 1.0

0.0 -1.0 -2.0

-3.0 * Change)Log2 (Fold S vs. F -4.0

-5.0 *

Figure 4.3: Quantitative real-time PCR confirmation of gene expression in the mid intestine of fish fed SBM-based diets compared to fish fed FM-based diets for 60 days.

Data are presented as the mean Log2 (fold change) ± standard deviation of twelve fish, and asterisks (*) indicate significant differences between treatments (P<0.05).

114

Table 4.3: RNA-seq results of the top differentially expressed genes in the mid intestine

between fish fed a SBM-based diet and fish fed FM-based diet. All of these genes were

found to be significantly different between treatments (P<0.05).

SBM vs. FM Up-Regulated Genes in SBM fed fish Gene Symbol Log2 (Fold Change) Cholesterol biosynthetic pathway squalene monooxygenase SQLE 3.13 farnesyl-diphosphate farnesyltransferase 1 FDFT1 2.68 methylsterol monooxygenase MSMO1 2.67 hydroxymethylglutaryl-CoA reductase HMGCR 2.47 cholesterol 7-alpha monooxygenase CYP7A1 2.36 hydroxymethylglutaryl-CoA synthase HMGCS1 2.28 diphosphomevalonate decarboxylase MVD 2.10 lanosterol 14-alpha demthylase L14AD 2.02 emopamil-binding protein EBP1 2.00 Cellular Defense ubiquitin C-terminal hydrolase 10 USP10 2.46 immunoglobulin mu heavy chain constant region IGHM 2.43 carcinoembryonic antigen-related cell adhesion molecule 5-like CAECAM5 2.03 butyrophilin subfamily 2 member A2 BTN2A2 2.00 macrophage mannose receptor MRC1 1.18 Metabolic Processes N-acetylneuraminate lyase NPL 2.54 long chain fatty acid elongase ELOVL6 1.82 cathepsin B CTSB 1.75 proprotein convertase subtilisin/kexin-type PCSK9 1.75 low-density lipoprotein receptor-related protein 2-like LRP2 1.47 Down-Regulated Genes Gene Symbol Log2(Fold Change) Negative regulation of immune response GTPase IMAP family member 8-like GIMAP8 -2.10 NACHT, LRR and PYD domains-containing protein 1 NLRP1 -2.26 NOD-like receptor family CARD domain containing 3 NLRC3 -2.68 C1q / TNF-related protein 3 C1QTNF3 -3.57 Regulation of proteins proteasome subunit beta 2 PSMB2 -1.26 alpha-2 macroglobin A2ML1 -1.47 synaptogyrin 1-like SYNGR1 -2.47 Regulation of lipids Adenosine Receptor A1-like ADORA1 -2.58 Peroxisomal 2,4 dienoyl CoA Reductase (SPS19 like) DECR2 -2.01 mid1-interacting protein M1IP1 -1.95 DNA binding Ccaat box DNA binding protein subunit B NFYB -2.35 Dlx3 homeobox protein DLX3 -3.41 Transport/Structure chroiolytic enzyme HCEA -3.72 transporter (solute carrier family 13) SLC13A4 -3.40 cytochrome P450 1A1 CYP1A -2.41 lamin LMN -inf lamin-A LMNA -inf 115

Table 4.4: Primers used for Real-Time qPCR expression analysis and restriction enzymes

used for amplicon confirmation.

Gene Product Tm Restriction Gene Name Forward Reverse Name Size (bp) (°C) Enzymes Down-Regulated Cytochrome HaeIII, Hinf1, CYP1A GACAGGCCAGTGGAGATGG TTTGTGCAGAGGATCGTGG 200 58 P450 Sau3A1 HaeIII, Hinf1 MID1 M1IP1 CTGATGAGGTGTGGGACTGC GGCCTTGTGGTCATCTTGGA 128 60 Msc1 ProteasomeB2 PSMB2 ATGGCCCCGGTCTCTACTAC ACACGCATGAGGCATTTCTCT 146 60 Taq1, Sau3A1 NLRC3 NLRC3 AAATGGTCTGACTGGTCGCT CATCAGAGCAGAACAGACTCCA 197 59

Up-Regulated Bsty1, BspH1, Hydroxy-CoA HMGCS1 GATTGGAGATCAGCAGGGAGG CCAAGGTGGCAAAAGTACAACTC 182 60 Sau3A1 Sau3A1, Bsty1, Lanosterol LI4AD TGGGGCGTGTGTTTTGAGAG CACCACGGATGTCAGGTTGT 129 55 Mluc1 Emopamil- Hinf1, PvuII, EBP1 CTCCACGGTCCCAACAAGAG TGGTGTCCGTGCTATCTCTG 188 57 binding HaeIII Sau3A1, AflIII, Squalene SQLE AGCCCCGTCGTCATAGAGAT GGCGATGCGTACAACATGAG 120 55 Hinf1 Hinf1,Sau3A1 Butyrophilin BTN2A2 AGCCTGCTGTTCATCCTCAT TGATCCAGAAGAAGCCGAGG 187 55

Housekeeping

18S 18S TACAGTGAAACTGCGAATGG GCATGGGTTTTGGGTCTG 153 60

Beta-actin ACTB GGCCAACAGGGAAAAGATGA ACCGGAGTCCATGACGATAC 130 59

EF-1alpha EF1α TGACAACGTCGGCTTCAACA GGGTGGTTCAGGATGATGAC 135 60

RNAPOL RNAPOL GCCATGACACCCAGCTAACA GCAACGTGTGTCCGTGTTTT 157 60

RPS20 RPS20 AGCCGCAACGTCAAGTCT GTCTTGGTGGGCATACGG 98 60

116

8 FM2 6 SBM2

tissue) Concentration (mg/g 2

0 Total Lipids Cholesterol

Figure 4.4: Whole body concentration (mg/g tissue) of total lipid and cholesterol levels between fish fed FM2 and SBM2 diets. No significant differences were found between groups.

117

Chapter 5: Isoflavone accumulation in the muscle tissue of yellow perch (Perca flavescens) after being fed soybean meal based diets

5.1 Abstract

In an effort to reduce production costs as well as reliance on wild caught fish used for aquafeeds, a greater concentration of plant-based proteins are being incorporated into diet formulations. One of the most promising fish meal (FM) replacements is soybean meal

(SBM), due to its high protein content, relatively well balanced amino acid profile and lower cost compared to FM. However, as inclusion levels of soybean meal in aquafeeds increase, it is important to understand from a consumer’s perspective, if accumulation of soy bioactives (isoflavones) occurs in the muscle/meat of these fish. Studies of isoflavone accumulation are limited in fish, and this was the first such study in yellow perch, an important fish species in the Great Lakes region. It was hypothesized that by feeding yellow perch diets with a high concentration of SBM, isoflavone accumulation would occur within the fillet (muscle tissue) of these fish in a dose dependent manner.

Fish from three groups were analyzed (2013, 2015 and 2016), based on their initial formulated feed (fish meal-based diets, FM2; or diets with 75% of the fish meal protein replaced with soybean meal protein, SBM2), length of exposure to SBM2-based diets, sex and weight. As juveniles, all groups of fish had been fed either the FM2 diet or the

SBM2 diet as their first formulated feed for 2 months. Afterwards, they were transitioned 118 back and forth between the FM2 diet over the winter/spring months and to the SBM2 diet during the summer/fall months. At the time of sampling, all fish had been fed the SBM2 based diet the 6 months prior. Results of this study discovered that isoflavone accumulation occurred within the muscle tissue of yellow perch at low concentrations (ng isoflavones/g muscle tissue), and was significantly affected by the first feed given to fish as juveniles P=0.0008). Of the fish from the 2013 and 2015 groups, those that had been fed the FM2 diets initially had 2.57±0.98 ng/g, in comparison to the SBM2 initial diets,

26.2±11.4 ng/g isoflavones present in the muscle tissue. Interestingly, fish from the 2016 group were found to have no significant differences in concentrations of total isoflavones regardless of their initial feeding, 19.5±3.0 ng/g in those fed the FM2 diet, and 31±4.3 ng/g in those fed the SBM2 diet.

5.2 Introduction

As the global demand for seafood continues to increase, aquaculture production has grown rapidly to meet this need, while wild caught fish production has remained stagnant since 1985. In 1995, roughly 28 million tons of fish were produced through aquaculture, and in 2014 that amount grew to around 81 million tons (Food and Agriculture

Organization of the United Nations, 2016). However, this growth cannot continue at the current rate due to sustainability issues along with narrow profit margins. Predominantly, feed (fish meal) is one of the highest operating costs for farmers (Gentry et al., 2017;

Little et al., 2016). In an effort to reduce production costs as well as reliance on wild caught fish used for aquafeeds, a greater concentration of plant-based proteins are being

119 incorporated into diet formulations. One of the most promising fish meal (FM) replacements is soybean meal (SBM), due to its high protein content, relatively well balanced amino acid profile and lower cost compared to FM (Chou et al., 2004).

However, as inclusion levels of soybean meal in aquafeeds increase, it is important to understand from a consumer’s perspective, if accumulation of soy bioactives

(isoflavones) occurs in the muscle/meat of these fish. Isoflavones are a class of phytochemicals within the flavonoid family found in legumes, such as soybeans, peas, red clover and alfalfa and other members of the Fabaceae family (Messina, 2010; Vacek et al., 2008). Like many other phytochemicals, isoflavones are secondary metabolites produced by the plant in response to various stressors. They possess antimicrobial, antifungal and antioxidant properties that aid in the survival of the soybean plant

(Messina, 2010). Because of this, concentrations of isoflavones within the plant vary greatly from crop to crop and are dependent on genetics, site of cultivar and environmental conditions such as temperature, water availability, UV exposure and pests

(Eldridge and Kwolek, 1983; Lee et al., 2003b).

There are three main isoflavones that are classified based on four chemical forms; aglycones (genistein, daidzein and glycitein), β-glucosides (genistin, daidzin, glycitin), acetylglucosides and malonylglucosides (Inbaraj and Chen, 2012; Kao and Chen, 2006).

In plants, isoflavones are naturally present as β-glucosides, however, processing methods and storage conditions can influence form and content (Eldridge and Kwolek, 1983; Lee et al., 2003a; Rostagno et al., 2009). After ingestion, isoflavone glucosides are broken

120 down by glucosidases in the small intestine, converting them aglycones, where they are either absorbed or further metabolized by intestinal microflora in the large intestine (Day et al., 1998).

The two most abundant isoflavones in soybean meal are genistein, daidzein, followed by glycitein typically found in a 1.3:1.0:0.5 ratio, respectively (Pastore et al., 2018). While genistin and its derivatives are usually the most prevalent in SBM, it has been found that daidzein is absorbed at higher levels than genistein, but all three forms are bioavailable

(Rowland et al., 2003). There is some debate as to whether isoflavones have beneficial or negative health effects when consumed. Some researchers claim that genistein and daidzein as well as their metabolites may cause reproductive effects because they can exert both estrogenic and anti-estrogenic activities due to structural similarities to estrogen (Dixon and Ferreira, 2002). Studies have discovered that the estrogenic effect of isoflavones may increase the percentage of proliferative cells in tumors and can elevate the weight of estrogen-dependent mammary adenocarcinomas in rats (Allred et al., 2004; Kijkuokool et al., 2006). In fish, they have been shown to cause reproductive issues, reducing gamete quality and fertilization in rainbow trout (Bennetau-Pelissero et al., 2001), and yellow perch (Ko et al., 1999).

There is also epidemiological evidence linking isoflavones to the reduction in prostate and breast cancer rates in Asia due to high consumption rates of soy products. It has been estimated that people in Asian countries consume between 25-50mg isoflavones/day, in comparison to Western diets that contain 0.15-1.7mg isoflavones/day (Messina et al.,

121

2006; Messina and Barnes, 1991; Yu et al., 2016). Importantly, numerous studies have also shown that genistein and other isoflavones have health benefits in humans, even when consumed in small doses (Chacko et al., 2007; Rodríguez-Roque et al., 2013; Yu et al., 2016). Isoflavones have been shown to play roles as antioxidants, and have exhibited anticancer, and anti-inflammatory activities in mammals (Wang et al., 2013; Wei et al.,

1995). In addition, they can be supplemented into foods for use as a delivery system and are deemed “functional foods” for their health benefits (Ahn-Jarvis et al., 2013; Messina,

2010).

Lin et al. (2004) found that when Japanese quail were fed a 50 or 100mg genistein supplement for 5 days, deposition was found in the eggs in a dose dependent manner,

1.25 and 2.5 µg, respectively. Authors predicted that deposition within the eggs would increase after longer periods of feeding as well as with higher doses, which would help increase westerners’ consumption of isoflavones, as the eggs could be delivery vehicles.

Yet genistein deposition studies remain limited in fish. D’souza et al. (2005) failed to find a dose dependent deposition of genistein in rainbow trout fillets after feeding various amounts of supplemented genistein (0, 500, 1,000 and 3,000 ppm) for 6-12 months, and overall product quality (color and taste of fillet) was not affected. Authors found 5.4- pmol/mg genistein in the fillets of trout fed the 3,000ppm genistein diet, which is equivalent to a 40% soybean meal inclusion within the fish diets. Gontier-Latonnelle et al. (2007) found that rainbow trout muscle accumulated only around 0.14% of total ingested radioactively labeled genistein, with the majority found in the viscera (15%) after 48 hours of ingesting of an oral dose of 200mg/kg genistein, yet no accumulation 122 was observed after 72 hours. However, both of these studies supplemented the genistein aglycone into fish meal based diets or oral dose, thus it remains unclear if isoflavone absorption and accumulation occurs when fish are fed soybean meal based diets that contain various forms of isoflavones and to what extent this occurs. In addition, this will be the first such study in yellow perch, a fish that typically has much lower lipid levels in the tissue compared to rainbow trout or salmon (Twibell et al., 2001; Zhou et al., 1996).

The goal of this study was to determine if isoflavone accumulation occurred within yellow perch muscle tissue after being fed a SBM-based diet (replacing 75% of the fish meal) over varying lengths of time, and how deposition varied among different age groups, sex of fish or dietary history. It was hypothesized that isoflavone accumulation within the muscle tissue of yellow perch would occur in a dose dependent manner based on duration of exposure to the SBM2 based diets.

5.3 Materials and Methods

5.3.1 Fish and experimental conditions

Fish were raised at The Ohio State University’s aquaculture facility in the School of

Environmental and Natural Resources. There were three groups (2013, 2015 and 2016) of yellow perch used in this study that were from a previous nutritional programming study in our lab, see Kemski et al. (2018) and Chapter 3 for details. The first group of fish were produced in 2013 (n=50) and are the parents of the other two groups. The second group of fish were produced in 2015 (n= 160), and are the first set of offspring from the 2013 fish. Finally, the third group of fish are those that hatched in 2016 123

(n=160), and are again the offspring of the 2013 group of fish. All groups of fish were housed in individual, 400L recirculation tanks and were tagged with passive integrated transponders-tags (PIT-tags; Biomark, Boise, ID), so that growth and diet history could be monitored.

5.3.2 Diets

Two diets were used in this study: 1) fish meal as the major protein source (FM2) and 2) soybean meal, providing 75% replacement of fish meal protein (SBM2), provided by

Schillenger Seeds Inc. (Des Moines, IA). The soybean meal was a non-GM, selectively- bred soy variety with high protein (56%) and low oligosaccharide content (0.72g/100g), processed as a cooked and solvent-extracted and extruded-expelled meal, whose biochemical profile is explained in detail in Baker et al. (2010) and Baker and Stein,

(2009). Diets were formulated to be isonitrogenous and isoenergetic, and were produced using commercial manufacturing technology at the U.S. Fish &Wildlife Service Bozeman

Fish Technology Center (Bozeman, MT)(Table 5.1). Although groups of fish used in the study were from various years, they all followed the same feeding regime, and were transitioned through phases in which they were fed either the FM2 diet or the SBM2 diet

(Figure 5.1). These phases were repeated in an identical manner with each group of fish.

When fish were juveniles they were divided into two groups and fed either the FM2 or

SBM2 diets as their first formulated feed for 60 days (Phase 1). After which, they were

PIT-tagged to denote initial diet history, and all fed the FM2 diet for 9 months (Phase 2).

Fish were then all transitioned to the SBM2 diet, and fed for 7 months (Phase 3). These

124 three phases were part of the previously described nutritional programming study

(Kemski et al., 2018 and Chapter 3) and fish growth and survival was measured rigorously throughout these phases. After phase 3 was complete, fish were then kept for use in this study as well as for spawning purposes in order to maintain stocks. At this point, all fish were transitioned to the FM2 diet over the winter/spring months

(December-May; Phase 4) and switched back to the SBM2 diet during the summer/fall months (June-November; Phase 5). This was done for spawning purposes as energy in the fish during the winter months is used for gametogenesis (gonad development) and energy in the summer months is allocated to somatic (muscle) growth (Dabrowski et al.,

1996). Transitioning the fish between Phase 4 and Phase 5 was repeated identically each year, until sampling. Thus, the 2013 group of fish repeated these final two phases 3 times from 2015-2017. The 2015 group of fish only went through phase 4 and 5 once, and the

2016 group did not go through phase 4/5, as they had just finished phase 3.

Since the three groups fish all transitioned between FM and SBM-based diets, thus all receiving SBM at some point in their lives, a control (CON) group of fish reared alongside all of the aforementioned groups was included in the study. Control fish were siblings from each of the three groups (2013, 2015 and 2016) and were raised in an identical manner as the other fish, yet were only fed the FM2 diet throughout their lives.

5.3.3 Sampling

All fish were sampled in November 2017, after being fed SBM2 diet for 6 months. Fish were sacrificed using an overdose of anesthesia (tricaine methanesulfonate, MS-222, 125

Western Chemical, Ferndale, WA) at a concentration of 200mg/L. Three females and three males were sampled for analysis from each group, identified sex, initial (phase 1) diet and weight recorded. Muscle tissue was excised, snap frozen in liquid nitrogen, and stored at -80°C until subsequent analyses.

5.3.4 Chemicals and supplies

Daidzin, genistein and glycitein used as standards were purchased from LC Laboratories

(Woburn, MA) and β-glucuronidase/arylsulfatase from Helix pomatia (Type H-1, S9626) was obtained from Sigma-Aldrich (St. Louis, MO). Remaining reagents were purchased from Fischer Scientific Co. (Fairlawn, NJ) and were all HPLC grade.

5.4.5 Isoflavone extraction: Soybean meal and diets

Because of the long duration of the study, diets initially made in 2013 had to be remade with a new batch of soybean meal in 2015 and again in 2017. Thus three batches of SBM were used in the study, and all tested for isoflavone content individually as concentrations can vary from year to year. Isoflavone extraction was done on the pure soybean meal and fish meal, as well as the formulated SBM2 and FM2 diets. All samples were ground through a 500µm sieve prior to extraction. Extractions were performed according to

Achouri et al. (2005), with slight modifications. Briefly, a 0.5g sample was mixed with

3mL aqueous acetonitrile (66% v/v). Sample mixtures were placed in a sonicator (Fisher

Scientific FS30H) for 30 min. with ice, and subsequently centrifuged at 1,000 x g for 30 minutes. Following separation, the aqueous layer was transferred to a new 22mL glass

126 vial, leaving the pellet intact. The entire process was repeated two more times with the remaining pellet to ensure complete extraction. Aqueous layers were pooled and dried under a gentle stream of nitrogen gas. Extracts were stored at -20°C until HPLC analysis.

Immediately prior to HPLC analysis, extracts were reconstituted in 1mL methanol (80%) and filtered through a 0.2µm nylon filter (Grace Davison Discovery Science, Deerfield,

IL) into 2mL HPLC vials.

5.3.6 Isoflavone extraction: Muscle tissue

Sample extraction was performed according to Zuniga et al. (2013), with slight modifications. Two grams of frozen muscle tissue was homogenized over ice in a 15 mL centrifuge tube containing 2 mL of HPLC grade water. Mixtures were stored on ice and sonicated using a probe sonicator (40 kHz, Digital Sonifier SLPt, Branson Ultrasonic

Corporation, Danbury, CT) at 40% amplitude for three, 4 second bursts. 8 mL of chilled acetonitrile (ACN; 100%) was added to induce protein precipitation and extract isoflavone metabolites. Samples were vigorously shaken by hand for 30 seconds, and centrifuged (2500 x g) for 5 minutes. Supernatant was collected in a glass vial (22 mL) using glass Pasteur pipette, making sure not to disrupt the pellet. Chilled ACN (66%, 10.0 mL) was added to suspend pellet and mixture was probe sonicated with same parameters as above. Samples were then centrifuged (2500 x g for 5 min), supernatant collected and pooled into respective glass vial. These steps were repeated once more beginning by adding 10mL chilled ACN (66%) and extracts were continually pooled into respective vials to produce 3 transfers. Pooled supernatants were dried under nitrogen gas until only

127 a sticky film remained at the bottom. Immediately after drying, samples were reconstituted in 1.0 mL acetate buffer (2M, pH 5.5) and 10 µL sulfatase solution (~160 units of β-glucuronidase and ~5 units arylsulfatase), and incubated for 2 hours at 37oC

(Dubnoff Metabolic Shaking incubator; GCA/ Precision Scientific). Extracts were then cooled to room temperature diethyl ether (3.0 ml) was added and vigorously shaken for approximately 1 minute. After sitting for 5 minutes to allow for separation, the supernatant was collected and placed into a new 12ml glass vial. These steps were repeated with a second volume (3.0 mL) of diethyl ether, allowed to separate, and collected supernatant was pooled with previously collected extract. Samples were dried under nitrogen gas, and subsequently stored at –80°C until HPLC analysis. Dried extracts were solubilized in 150µL methanol (80%), bath sonicated for 30 minutes and filtered with a 0.2 µm nylon filter prior to HPLC with PDA analysis.

5.3.7 Calibration Curves and Standards

Stock solutions of each isoflavone analyte (genistein, genistin, daidzein, daidzein, and glycitin) were prepared at a 1mg/ml concentration in 100% MeOH. Absorbance was measured at the lambda max of each isoflavone, and concentrations were determined using the extinction coefficients from Murphy et al., (2002). Once stock solutions were made, a mixture was made with standards combined at an 8:1 ratio of glucosides to aglycones to create a final concentration of 50-150 nM/mL. The standard mixture was serially diluted (2-fold) and run in triplicate to generate the calibration curve. Peak areas versus nominal concentrations were used to generate a calibration curve, which was used

128 to determine isoflavone concentrations in extracts. Calibration standards were run with each HPLC run to ensure correct methodology and retention time. In addition, as an initial determination of the extraction protocols effectiveness, an internal standard of genistein (100µl of 4,000µg/mL in 80% MeOH) was added to tissue samples and water at the start of the extraction procedure, prior to the homogenization step. It was determined that recovery of the standard by HPLCS analysis after the extraction protocol was 94%.

5.3.8 HPLC Analysis

HPLC conditions and protocol were run according to Ahn-Jarvis et al. (2013). Analysis was done in an Agilent 1100 series HPLC with a photodiode array (PDA) detector

(Agilent Technologies, Santa Clara, CA). Briefly, reverse phase separation was done in

Symmetry C18 columns (4.6x75mm, 3µm particle- Waters Associates). The binary mobile phase (A) 0.1% aqueous formic acid, B) 0.1% formic acid in acetonitrile) gradient was as follows, 90:10 to 65:35 for 20 minutes, 65:35 to 0:100 from 20-22 min, followed by 0:100 to 90:10 from 22-25 min. and equilibration at 90:10 for 5 minutes between samples. Injection volume was 10µl per sample, and all samples were analyzed in triplicate. UV detection was used at the lambda max for each specific isoflavone, 249 nm for daidzein, 259 nm for glycitein and 261 nm for genistein. Peaks were integrated, and the area under the peak was used for calculation of concentrations based on the standard curve for that specific isoflavone.

5.3.9 Lipid Analysis

Proximate diet analysis was conducted at the Service Testing and Research Laboratory at

129

The Ohio State University. AOAC official methods were used in determination of total nitrogen (method 990.03), percent ash (method 942.05) and percent moisture (method

934.01) (AOAC, 2005) and total lipid analysis was conducted using the Folch method

(Folch et al., 1956).

5.3.10 Statistical analysis

Results are expressed as mean ± standard deviation (SD). Data were checked for normality, and weight gain, and lipid concentrations were compared among the three phase 1 diets within a single group (year) of fish using the analysis of variance

(ANOVA). A Tukey-Kramer test was also done for multiple comparisons of means. For isoflavone concentrations, an ANOVA was run to compare all groups of fish over the three years, and again, a Tukey-Kramer test was run for multiple comparisons of means.

All data was analyzed using JMP software (SAS Institute Inc., Cary, NC) and a P value of <0.05 was considered statistically significant.

5. 4 Results

5.4.1 Fish weight and lipid analysis

Fish weight from every group, related back to their phase 1 diets is presented in Figure

5.2. Fish from the 2013 group were the largest, with an average overall weight of

281±24g, and weights among the phase 1-diet histories (FM2, 297±76g; SBM2,

292±95g; CON, 253±54g) were not significantly different (P=0.567). Weight of fish from the 2015 group was also not significantly different (P=0.351) among diet treatments 130

(FM2, 95±33g; SBM2, 107±27g; CON, 143±61g), and was an average of 111±25g.

Finally, fish weight from the 2016 group was an average of 93±21g, and weights among the phase 1 diet treatments (FM2, 111±13g; SBM2, 97±19g; CON, 70±15g) were found to be significantly different (P=0.014) (Figure 5.2).

Total lipid levels in fish muscle tissue are presented in Table 5.2, and were compared among fish groups (years) and phase 1 diet history. Overall, total lipid levels were around 16.5±3.7 mg/g muscle tissue, and did not vary significantly among fish from different years (P=0.874), as well as within the 2016 group among the different dietary histories (P=0.088). However, a significant variation was seen in fish within the 2013

(P=0.0314) and 2015 (P=0.004) year groups, among the different dietary histories, with the SBM2 fed groups (from phase 1) having the highest total lipid levels (Table 5.2).

5.4.2 Isoflavone content in SBM and formulated SBM-based diets

Isoflavone concentration was analyzed in both the SBM itself as well as the formulated

SBM2 based diet. A sum of the combined twelve isoflavones found in the SBM and

SBM2 diet for each fish group is presented in Figure 5.3 as total isoflavone concentration

(µg/g). Diets were remade every two years with a new batch of soybean meal, and isoflavone content was compared among each year. A significant difference was found among the three batches of SBM (P<0.001) and total isoflavone content among the three years were 862±18 µg/g (2013), 1914±18 µg/g (2015) and 1953±4 µg/g (2017). This same trend was also seen in the SBM2 diets (P<0.001), and concentrations were 473±27

131

µg/g (2013), 932±43 µg/g (2015) and 879±36 µg/g (2017). While it is stated that the

SBM protein replaces 75% of the fish meal protein in the diets, the SBM as one of the many dietary ingredients is actually present in the overall diet formulation 52.9% (Table

5.1), which is consistent with the isoflavone concentrations, as the content was 49.7±5% reduced in the SBM2 diets compared to what was found in the SBM itself.

Isoflavone profiles were also consistent between the SBM and the SBM2 diets each year, and the relative concentration of each of the twelve forms is shown in Figure 5.4.

Genistin was found to be the most prevalent form present in the samples, both as genistin as well as in the 6”-O-malyonyl genistin form. Daidzin and 6”-O-malonyl daidzein were the next two most prevalent forms found, followed by 6”-O acetyl genistin and 6”-O acetyl daidzin. Glycitin and its derivatives were present in the least amount, as well as the aglycone form of each isoflavone.

5.4.3 Isoflavone concentration in muscle tissue

For isoflavone content in the muscle tissue, the three main aglycones (genistein, glycitein and daidzein) were analyzed, as those are the main forms typically found within the body.

Total isoflavone concentration (ng/g) within muscle tissue was found to be significantly different among the SBM2 fed fish compared to the FM2 fed fish from phase 1, regardless of the group (year) of fish (Figure 5.5, P=0.0008). Fish from the 2013 and

2015 groups that had been fed FM2 diets in phase 1 had 2±0.4 and 3±2 ng/g tissue isoflavone concentration, respectively, compared to the SBM2 fed fish (phase 1) from those same years that had 21±7 ng/g tissue (2013) and 31±15 ng/g tissue (2015). The 132 differences among the phase 1 diets in these two groups of fish was surprising as all fish were being fed SBM2-based diets at the time of sampling. Fish from the 2016 group of fish had similar concentrations of total isoflavones in the muscle 20±3 ng/g tissue in the

FM2 fed fish and 31±4 ng/g tissue in the SBM2 fed fish from phase 1.

The profile of the three isoflavones tested were found to be significantly different among the groups (years) of fish as well as between the FM2 and SBM2 fed diets from phase 1

(P<0.001). Figure 5.6 depicts the average concentration of each group (ng/g tissue) for each isoflavone, compared between the FM2 and SBM based diets. Daidzein was found to be the most prevalent isoflavone present in the muscle tissue followed by glycitein, while genistein was present in the least amounts. This was in contrast to what was seen in the isoflavone profile of the SBM, in which genistin was found to be the most abundant form.

5.5 Discussion

5.5.1 Fish weight and lipid analysis

Final weight of fish at the time of sampling (end of phase 5 for 2013, 2015 groups; and end phase 3 for 2016 group) was all very similar among dietary history treatments, except for those in the 2016 group. In this group, there was a significant difference found among dietary history with the CON group being significantly smaller (P=0.014) than the

FM2 fed group from phase 1. This was unexpected as the CON group was fed with the same FM2 diet throughout their lives, and have never been introduced to SBM, which

133 typically results in reduced weight gain (Kasper et al., 2007). One explanation for this lower final weight could be the fact that yellow perch exhibit dimorphic growth between males and females, with females having a faster growth rate and are able to reach an ultimate larger size than males (Hart et al., 2006; Kestemont et al., 2015; Malison et al.,

1986). When fish were initially stocked into tanks as juveniles, they are sexually indistinguishable (Shepherd et al., 2013), making it impossible to determine ratio of males to females. At the time of sampling for the present study, sex of fish was also recorded, and it was determined that the 2016 CON group had more males than females, thus contributing to the reduced weight of that group.

It is also important to note, that while the phase 1 diets are referred to throughout the study (FM2 or SBM2), all fish from each year (with the exception of the CON group) had just completed a phase in which they had been fed the SBM2 diet for 6 months prior to sampling. Depressed growth was not observed in this trial among the SBM2-fed fish in comparison to the CON-fed fish, unlike previous trials. For example, in a study by

Kasper et al. (2007), authors examined the effect of SBM on yellow perch juveniles (27g mean weight), which were fed diets for 8 weeks with increasing amounts (0-730g/kg) of

SBM replacement. Results showed an inversely proportional response in weight gain, and groups with increasing levels of SBM had significantly reduced weight gain.

Specifically, in the diet formulation with 73% SBM, weight gain of fish was roughly 20% that of the control fish meal (0% SBM) diet, with fish gaining 20% and 108% weight, respectively.

134

Lipid analysis was determined among muscle samples, because yellow perch are known to have low muscle lipid levels (<4g/100g) (Twibell et al., 2001), and it has been predicted that isoflavones may accumulate in lipophilic tissue since they bind to oestrogen receptor cites and become partitioned from the blood (Chang et al., 2000).

Results from the lipid analysis among the phase 1 diets found that in each group (year) the SBM2 fed fish had the highest concentration (mg/g) of lipid present in the muscle, which was found to be significantly different in the 2013 and 2015 groups (P=0.031 and

P=0.004, respectively). This is in contrast to the aforementioned study done by Kasper et al. (2007), in which graded levels of SBM-based diets were fed to yellow perch. Lipid concentrations found in the muscle tissue at the end of the 8-week study showed no significant differences among any of the dietary treatments. Results of this study are also in contrast to a study by D’Souza et al. (2006), in which rainbow trout were fed diets with

0, 20 or 40% SBM for 6 months, and also showed no difference in percent lipids in the muscle tissue.

5.5.2 Isoflavone content in SBM and formulated SBM-based diets

Results of HPLC analysis determined that there was a significantly higher amount of isoflavones present in the 2015 and 2017 batch of soybean meal, in comparison to the

2013 batch (P<0.001). Phytoestrogens, such as isoflavones, are a secondary metabolite within plants, which play a role as a defense mechanism when they become stressed, thus can vary significantly from crop to crop (Lee et al., 2003a; Messina, 2010). It has been reported in Lee et al. (2003a) that total isoflavone content in soybeans can range from

135

458-2317 µg/g from crops within a single year, thus variation among our SBM batches was expected.

Isoflavones are known to be fairly stable compounds, but heating and other treatments that occur during processing can both reduce levels as well as alter the chemical composition, causing conversions from naturally present forms (Ahn-Jarvis et al., 2013;

Erdman et al., 2004; Xu et al., 2002). Therefore, it was important to determine and compare isoflavone concentrations and profiles in both the SBM as well as the SBM2 formulated diets, to determine if there had been changes during processing. The SBM component within the diets comprised 52.9%, which was found to be consistent with the isoflavone concentrations. Total isoflavone content was found to be 49.7±5% less in the

SBM2 diets compared to concentrations present in the SBM itself. This validates that there was minimal loss of total isoflavone content after diet formulation and processing.

Additionally, the profile of the 12 individual isoflavones was the same among the various batches of SBM, as well as in the formulated SBM2 diets, which is promising that diet processing is not causing a loss or alteration of compounds.

5.5.3 Isoflavone concentration in muscle tissue

There have been definitive studies on the how isoflavones affect fish growth and reproductive performance in many species of fish such as, striped bass (Pollack et al.,

2003), salmonids (Bennetau-Pelissero et al., 2001; Ng et al., 2006), sturgeon (Pelissero et al., 1991), goldfish (Bagheri et al., 2013) and even yellow perch (Ko et al., 1999).

However, there is still a large gap in knowledge as to how they are digested and absorbed 136 in fish (Pastore et al., 2018). A significant limitation as well, is that in the aforementioned studies, isoflavones were supplemented into fish diets in their aglycone forms, which may be absorbed and utilized differently within the body than glucosides, the naturally occurring form in SBM (Erdman et al., 2004). Thus, this study addresses the more relevant question of isoflavone metabolism in fish after consumption of a SBM- based diet.

Concentrations of isoflavones found in yellow perch muscle tissue were negligible and only present in the range of ng/g of tissue, yet there was a significant difference

(P=0.008) found among fish fed the SBM2 diet as their initial formulated feed in phase 1 compared to those fed the FM2 diet. This was surprising as all groups of fish, with the exception of the control fish, had just completed phase 3 (2016 group), or phase 5 (2015 and 2013 group) in which they had been fed the SBM2 diet for 6 months prior. These findings suggest that the deposition of isoflavones occur when fish are juveniles, and are dependent on diet fed in the early stages of life when the growth rate is very high. In yellow perch, larval and juvenile fish experience exponential growth during this time, which in the wild help prevent predation from other fish (Post and Evans, 1989), yet also results in extreme muscle growth and development within the first 6 months of life.

Previous studies in our lab (Chapter 3) have shown that juvenile yellow perch at the start of phase 1 are stocked at a mean weight of 0.055±0.017g, weigh 0.496±0.152g after 30 days and weigh 2.0±0.64g after 60 days, which represents an average overall increase of around 3780% weight gain. When this is compared to the later phases after fish were more than a year old, phase 3 for example, individual fish weight went from an average 137 of 16.6±4.6g to 69.6±12.5g over 7 months, and is only a 350% increase in weight gain.

This being said however, results are still surprising, as based on average calculated feed consumption within these fish, adult fish had eaten larger amounts of feed, thus consumed higher concentrations of isoflavones, on average, daily. Specifically, it was calculated that during phase 1, fish consumed on average 21µg total isoflavones per day over 60 days, whereas in phase 3, they were consuming an average of 124µg isoflavones per day, over 7 months (see Chpater 3 for details on feeding rates). Overall, it was determined that the phase 1 diets exclusively, had an effect on muscle accumulation of isoflavones, and that sex, age and duration of exposure to SBM showed no effect on deposition.

Of the few isoflavone accumulation studies done in fish, D’souza et al. (2005) fed juvenile rainbow trout varying levels (0, 500, 1,000 and 3,000ppm) of supplemented genistein in the diet for 12 months and compared muscle tissue concentrations after both

6 and 12 months. The authors found that levels in the diet significantly influenced the concentration of genistein in muscle tissue, yet the duration of the study did not significantly change levels present. Specifically, it was found that after being fed a diet with 3,000ppm genistein, tissue levels at 6 months were 5.48pmol/mg and 5.39pmol/mg at 12 months. These results help support findings from the present study, as possible accumulation occurs during initial exposure to isoflavones in juvenile fish, when growth rate is highest, yet at this time, do not seem to be dependent on duration of feeding. In another study examining the deposition of genistein in both adult rainbow trout and

Siberian sturgeon, Gontier-Latonnelle et al. (2007) fed an oral dose of 200 mg/kg 138 radioactive labeled genistein and fish were sampled 48 and 72 hours following consumption. It was determined that genistein was not largely distributed throughout the body, and remained mostly in the plasma. Distribution occurred briefly, but levels were significantly reduced after 48 hours and were not detectible after 72 hours. During dispersion, the greatest amounts of radioactive genistein present in sturgeon was found to be in the liver, yet in trout was located in the intestinal fat and gonads. For both species, muscle only contained around 0.14% of total radioactivity and levels decreased significantly after 48 hours. The authors also examined metabolites of genistein and it was determined that sturgeon produced sulfate conjugates, whereas trout produce mainly glucuronides. The difference in metabolites, and circulating form of genistein present in the body may affect bioavailability, yet more studies are needed to confirm this.

Subsequently, due to the short duration of that study, it is difficult to compare findings to results of the present study.

For absorption of isoflavones to take place, glucosides must be broken down by β- galactosidases present in the intestine (Danciu et al., 2017; Yu et al., 2016). In fish, it has been reported that there is glucosidase activity within the pyloric caeca (Buddington and

Diamond, 1986), and breakdown products of the aglycone form of genistein have also been shown (Gontier-Latonnelle et al., 2007), indicating that breakdown occurs in a similar manner in fish compared to mammals. Additionally, the extraction method in this study used, β-glucuronidase/arylsulfatase, an enzyme that further breaks down sugar moieties and aids in extraction of isoflavones from the muscle tissue matrix. Therefore, concentrations of the three main aglycone isoflavones (genistein, glycitein and daidzein) 139 were determined, and possible glucosides that may have been present were not examined.

Breakdown metabolites of the aglycones, such as O-desmethylangolensin (ODMA), 5- hydroxy-equol (from genistin), equol (from daidzin) and dihydrodaidzein, were not detected at any level in any of the samples tested in the study (data not shown). Results showed that of the three main aglycones, there was a variation between the profiles found from fish fed the FM2 diet versus the SBM2 diet in phase 1. Fish from the SBM2 diet, accumulated significantly higher levels of daidzein compared to glycitein and genistein

(P=<0.0001). Whereas fish fed the FM2 diet, had surprisingly higher levels of glycitein, followed by daidzein and genistein, which were found to be significantly different

(P<0.0001). Other studies have shown that daidzein is typically absorbed at higher levels than genistein in humans and mammals (Rowland et al., 2003), which confirms the results seen in the SBM2 fed fish, yet results from the FM2 fed fish were found to be different from what is known in mammals. In a recent study by Merlanti et al. (2018), the authors developed a new method for isoflavone extraction from fish muscle tissue, and compared farmed rainbow trout that had been fed diets with various SBM inclusion levels. Total isoflavone levels detected in the study were actually quite similar to concentrations found in the present study. The authors detected 28.9±13.8 µg/kg in muscle from fish fed diets with the highest inclusion levels of SBM, 24.1±19.2 µg/kg in muscle from fish fed diets with low levels of SBM inclusion and only 0.35±0.14 µg/kg in muscle from the FM fed fish. Surprisingly, which they credit their extraction method that lacks β-glucuronidase/arylsulfatase treatment, they report equal if not higher levels of the glucoside forms present in the rainbow trout muscles. This goes against findings in

140 humans and mammals reported for many years (Ganai and Farooqi, 2015; Rowland et al.,

2003). Setchell et al. (2002), even provided evidence against the possibility that isoflavone glucosides could be absorbed in humans, and showed that bioavailability required hydrolysis of the sugar moiety and they are not absorbed intact. However, there are no studies that examine the breakdown and absorption of these compounds in fish, as well as a lack of data as to how these molecules are distributed within the body and in what forms they are subsequently stored. It is also important to note, that the study by

Merlanti et al. (2018), did not present any specific information regarding the diets fed to the fish (i.e. amount of SBM inclusion), the length of feeding, age of fish or sampling procedures prior to extraction process, making it difficult to determine validity of findings, thus it is difficult to compare these studies and more research is required.

5.6 Conclusions

This study presented some of the first evidence of isoflavone accumulation within the muscle tissue of yellow perch after being fed a SBM-based diet. To the best of our knowledge, no other study has been done examining SBM-based diets that have a naturally present isoflavone content on possible deposition within the body, as other studies have supplemented isoflavones within diets. Results of the present study determined that initial diets fed to juveniles (phase 1), had a significant effect on total isoflavone concentration in the muscle, and that sex, weight or even duration of exposure to the SBM2 diet had no influence on accumulation. However, levels present in the muscle were negligible and would not be deemed a significant dietary source of

141 isoflavones if fish fillets were consumed, as levels in the range of mg/g food product are typically used (Messina and Barnes, 1991). More detailed research is needed to determine details of how fish breakdown, absorb and store isoflavones, as well as if other isoflavone metabolites are also present in the muscle tissue. Yet, findings in this study did confirm that isoflavone accumulation occurs when fish are fed a SBM-based diet, and further studies on other species of fish is also needed to understand potential differences in accumulation rates and possible effects on consumers.

142

Table 5.1: Dietary ingredients and composition of FM2 and SBM2 diets

Ingredient FM2 (g) SBM2 (g) Menhaden Fish Meal1 54.09 12.00 Soybean Meal2 - 52.86 Wheat Flour 30.15 14.96 Fish Oil 4.34 8.00 Soybean oil 1.98 0.24 Lecithin 3.00 3.00 CPSP 4.00 4.00 Vitamin mix3 1.00 1.00 Mineral mix4 0.10 0.10 Vitamin C 0.10 0.10 Choline Chloride 0.06 0.06 Calcium Phosphate - 1.00 Lysine - 1.50 Methionine 0.33 0.33 Arginine 0.65 0.65 Threonine 0.20 0.20 TOTAL 100.00 100.00 Composition (% DM) Dry Matter 91.7 94.4 Crude Protein 42.4 43.9 Crude Lipid 15.6 13.6 Ash 10.7 6.7 Cholesterol 1.59 0.49 159.03% Protein 258.9% Protein- Provided by Schillenger Genetics Inc. 3 Provided by U.S. Fish &Wildlife Service Bozeman Fish Technology Center; Batch 702 4Provided by U.S. Fish &Wildlife Service Bozeman Fish Technology Center; Batch 1520

143

Phase 1: Phase 4: 2 Months Phase 2: Phase 3: Phase 5: 9 Months 7 Months 6 Months FM2 or 6 Months FM2 SBM2 SBM2 SBM2 FM2

Figure 5.1: Flow chart of feeding regime followed by each generation of fish. Phase 1, 2 and 3 were part of a previous nutritional programming study. After Phase 3 was complete, fish were transitioned back and forth between Phase 4 and 5 each year until sampling. The control (CON) group was fed only FM throughout the duration of the study, and did not go through any of these phases.

144

450 P=0.567 400

350 300 P=0.351 P=0.014 250 FM2 200 SBM2 Weight (g) Weight a 150 ab CON 100 b 50 0 2013 2015 2016

Group of fish

Figure 5.2: Fish weight (g) of each group at the time of sampling. Diets listed represent diet fish were fed during Phase 1 of the experiment. Values presented are means ± standard deviation, and statistics were done comparing fish from different Phase 1 diets within the same year. Letters denote significant differences (P<005).

145

Table 5.2: Lipid concentrations (mg/g) in muscle tissue of yellow perch in each group

(2013, 2015 and 2016). Values presented are means ± standard deviation and statistics were done comparing lipid concentrations among fish from same year. Letters denote significant differences (P<0.05).

Year Phase 1 Diet Lipid mg/g tissue P Value FM2 13.5±2.7b 2013 SBM2 18.8±4.1a 0.0314 CON 14.2±3.2ab FM2 14.9±2.1b 2015 SBM2 23.9±4.1a 0.004 CON 13.2±2.5b FM2 14.5±4.9 2016 SBM2 20.7±3.0 0.088 CON 14.9±3.7

146

a 2000 a 1800

1600

1400 1200 A A 1000 b SBM 800 600 B SBM Diets

Concentration (µg/g) Concentration (µg/g) 400 200 0 2013 2015 2017 SBM batches by year

Figure 5.3: Total isoflavone concentration (µg/g) present in soybean meal (SBM) and formulated SBM2-based diets made from various batches of SBM from 2013-2017.

Values are presented as the sum of the combined 12 isoflavone conjugates ± standard deviation among three years. Fish were transitioned to the next SBM batch at that specific year. At the time of sampling, all fish were being fed SBM-based diets from

2017. Statistics were run individually among SBM versus year and then among SBM diets versus year. Letters denote significant differences among years (P<0.001).

147

Relative concentration (µg/g) (µg/g) concentration Relative

Isoflavone

Figure 5.4: Profile of various isoflavone forms present in the SBM and SBM2 Diet.

Values presented are relative concentration levels among one another.

148

50 a 45

40 ab 35 bc 30

25 c FM2 20 15 SBM2 10 d Concentration (ng/g tissue) d 5 0 2013 2015 2016 Fish groups

Figure 5.5: Total isoflavone concentration (ng/g) in muscle tissue samples from each group of fish (2013, 2015 and 2016). FM2 and SBM2 correspond to initial diets fed during Phase 1 of the experiment. Values are presented as mean ± SD, and letters denote significant differences among groups (P=0.0008).

149

a 30

15 FM2 b 10 bc SBM2 5 cd d d Concentration (ng/g tissue) 0 Daidzein Glycetein Genistein Isoflavone

Figure 5.6: Profiles of the three major isoflavones present in the muscle tissue of yellow perch. Values presented are means ± SD of the concentration (ng/g) within the FM2 or

SBM2 muscle tissue of all groups of fish. Statistics were run between FM2 and SBM2 values of the same isoflavone, and letters denote significance differences (P<0.0001).

150

Chapter 6: Conclusions

Chapter 2

Results of this study found that nutritional programming to be a successful way to adapt yellow perch to better utilize SBM-based diets as adults. Juvenile fish when fed a diet with 75% SBM replacement for FM protein as their first feed, and then re-introduced to the same diet 9 months later, had numerically higher growth than fish fed other diets initially. These results are very promising as there are only a limited number of studies that have shown improved growth performance with SBM inclusion above 60% in diet formulations. It was also shown that reproductive function and egg quality are not affected in fish that were previously fed a SBM diet, if transitioned to a FM diet during gametogenesis. Future studies are being conducted to determine if nutritional programming of broodstock fish will pass on the programming changes to their progeny, allowing them to be even further adapted to utilize SBM diets.

Chapter 3

It was determined through a second series of experiments that nutritional programming was a successful way to adapt yellow perch to better utilize SBM-based diets as adults.

Results demonstrated that parental nutritional history had a significant effect on offspring percent weight gain throughout their lives. Fish from 2015, when fed a diet with 75%

151 soybean replacement for fishmeal protein as their first feed, and then re-introduced to the same diet 9 months later, were found to have significantly higher weight gain percent than those previously fed FM-based diets. These results are very promising as there are only a limited number of studies that have shown improved growth performance with

SBM inclusion above 60% in diet formulations.

Chapter 4

Transcriptomic results validated that FM protein can be replaced in juvenile yellow perch diets by up to 75% SBM protein without causing in a significant reduction in weight gain percent, or survival and only a slight reduction in terms of final weight. Findings from this study were promising in that both total body lipid levels and cholesterol levels were not significantly different between fish fed the SBM2 diet and those fed the FM2 diet, especially since dietary levels were not equal. This is one of the first studies of its kind that examined the transcriptional response by RNAseq in the mid intestine of juvenile yellow perch after being fed a SBM-based diet. Additionally, it is encouraging to observe that juvenile yellow perch were capable of responding to dietary changes, especially in regards to cholesterol, and that they were able to regulate cholesterol biosynthesis and other metabolic pathways to adapt within the first 60 days of introduction to a formulated feed.

Overall, these data produced from this study provides a better understanding of the mechanisms and pathways expressed when FM is replaced with SBM in diets and given as the first formulated feed to juvenile yellow perch. These results offer the basis for 152 additional research on improved diets as well as possible transcriptional differences between parental nutritional history and progeny response to SBM based diets. It also brings us one step closer to using SBM and other, more sustainable plant-based protein sources at higher inclusion levels in aquafeeds, thereby continuing to reduce the reliance on FM.

Chapter 5

This study presented some of the first evidence of isoflavone accumulation within the muscle tissue of yellow perch after being fed a SBM-based diet. To the best of our knowledge, no other study has been done examining SBM-based diets that have a naturally present isoflavone content on possible deposition within the body, as other studies have supplemented isoflavones within diets. Findings from this study determined that initial diets fed to juveniles (phase 1), had a significant affect on total isoflavone concentration in the muscle, and that sex, weight or even duration of exposure to the

SBM2 diet had no influence on accumulation. However, levels present in the muscle were negligible and would not be deemed a significant dietary source of isoflavones if fish fillets were consumed, as levels in the range of mg/g food product are typically used

(Messina and Barnes, 1991). More detailed research is needed to determine details of how fish breakdown, absorb and store isoflavones, as well as if other isoflavone metabolites are also present in the muscle tissue. Yet, findings in this study did confirm that isoflavone accumulation occurs when fish are fed a SBM-based diet, and further

153 studies on other species of fish is also needed to understand potential differences in accumulation rates and possible effects on consumers.

154

References

Abernathy, J., Brezas, A., Snekvik, K.R., Hardy, R.W., Overturf, K., 2017. Integrative functional analyses using rainbow trout selected for tolerance to plant diets reveal nutrigenomic signatures for soy utilization without the concurrence of enteritis. PLoS One 12, e0180972. doi:10.1371/journal.pone.0180972

Achouri, A., Boye, J.I., Belanger, D., 2005. Soybean isoflavones: Efficacy of extraction conditions and effect of food type on extractability. Food Res. Int. 38, 1199–1204. doi:10.1016/j.foodres.2005.05.005

Adams, M.D., Kelly, J.M., Gocayne, J.D., Dubnick, M., Polymeropoulos, M.H., Xiao, H., Merril, C.R., Wu, A., Olde, B., Moreno, R.F., Kerlavage, A.R., McCombie, W.R., Venter, J.C., 1991. Complementary DNA Sequencing: Expressed Sequence Tags and Human Genome Project. Science (80-. ). 252, 1651–1656.

Ahn-Jarvis, J.H., Riedl, K.M., Schwartz, S.J., Vodovotz, Y., 2013. Design and selection of soy breads used for evaluating isoflavone bioavailability in clinical trials. J. Agric. Food Chem. 61, 3111–3120. doi:10.1021/jf304699k

Akiyama, T., Ishida, J., Nakagawa, S., Ogawara, H., Watanabe, S.-I., Itoh, N., Shibuya, M., Fukami, Y., 1987. Genistein, a specific inhibitor of tyrosine-specific protein kinases. J. Biol. Chem. 262, 5592–5595.

Alami-Durante, H., Médale, F., Cluzeaud, M., Kaushik, S.J., 2010. Skeletal muscle growth dynamics and expression of related genes in white and red muscles of rainbow trout fed diets with graded levels of a mixture of plant protein sources as substitutes for fishmeal. Aquaculture 303, 50–58. doi:10.1016/j.aquaculture.2010.03.012

Allred, C., Allred, K.F., Ju, Y.H., Clausen, L.M., Doerge, D.R., Schantz, S.L., Korol, D.L., Wallig, M.A., William, G.H., 2004. Dietary genistein results in larger MNU- induced, estrogen-dependent mammary tumors following ovariectomy of Sprague- Dawley rats. Carcinogenesis 25, 211–218. doi:10.1093/carcin/bgg198

Almeida, L.Z., Guffey, S.C., Krieg, T.A., Höök, T.O., 2017. Predators Reject Yellow 155

Perch Egg Skeins. Trans. Am. Fish. Soc. 146, 173–180. doi:10.1080/00028487.2016.1249294

Alwine, J.C., Kemp, D.J., Stark, G.R., 1977. Method for detection of specific RNAs in agarose gels by transfer to diazobenzyloxymethyl-paper and hybridization with DNA probes. Biochemistry 74, 5350–5354.

AOAC, 2005. Official Methods of Analysis, 18th ed. AOAC International, Arlington, VA.

Azizishirazi, A., Dew, W.A., Bougas, B., Dashtban, M., Bernatchez, L., Pyle, G.G., 2014. Chemosensory mediated behaviors and gene transcription profiles in wild yellow perch (Perca flavescens) from metal contaminated lakes. Ecotoxicol. Environ. Saf. 106, 239–245. doi:10.1016/j.ecoenv.2014.04.045

Baeverfjord, G., Krogdahl, Å., 1996. Development and regression of soybean meal induced enteritis in Atlantic salmon, Salmo salar L., distal intestine: A comparison with the intestines of fasted fish. J. Fish Dis. 19, 375–387. doi:10.1111/j.1365- 2761.1996.tb00376.x

Bagheri, T., Imanpoor, M.R., Jafari, V., Bennetau-Pelissero, C., 2013. Reproductive impairment and endocrine disruption in goldfish by feeding diets containing soybean meal. Anim. Reprod. Sci. 139, 136–144. doi:10.1016/j.anireprosci.2013.02.003

Bainbridge, M.N., Warren, R.L., Hirst, M., Romanuik, T., Zeng, T., Go, A., Delaney, A., Griffith, M., Hickenbotham, M., Magrini, V., Mardis, E.R., Sadar, M.D., Siddiqui, A.S., Marra, M.A., Jones, S.J.M., 2006. Analysis of the prostate cancer cell line LNCaP transcriptome using a sequencing-by-synthesis approach. BMC Genomics 7. doi:10.1186/1471-2164-7-246

Baker, K.M., Kim, B.G., Stein, H.H., 2010. Amino acid digestibility in conventional, high-protein, or low-oligosaccharide varieties of full-fat soybeans and in soybean meal by weanling pigs. Anim. Feed Sci. Technol. 162, 66–73. doi:10.1016/j.anifeedsci.2010.08.012

Baker, K.M., Stein, H.H., 2009. Amino acid digestibility and concentration of digestible and metabolizable energy in soybean meal produced from conventional, high- protein, or low-oligosaccharide varieties of soybeans and fed to growing pigs. J. Anim. Sci. 87, 2282–2290.

Bakke-McKellep, A.M., McL Press, C., Baeverfjord, G., Krogdahl, Å., Landsverk, T., 2000. Changes in immune and enzyme histochemical phenotypes of cells in the intestinal mucosa of Atlantic salmon, Salmo salar L., with soybean meal-induced enteritis. J. Fish Dis. 23, 115–127. doi:10.1046/j.1365-2761.2000.00218.x 156

Bakke-Mckellep, A.M., Penn, M.H., Salas, P.M., Refstie, S., Sperstad, S., Landsverk, T., Ringø, E., Krogdahl, Å., Shild Krogdahl, Å., 2007. Effects of dietary soyabean meal, inulin and oxytetracycline on intestinal microbiota and epithelial cell stress, apoptosis and proliferation in the teleost Atlantic salmon (Salmo salar L.). Br. J. Nutr. 97, 699. doi:10.1017/S0007114507381397

Bakke-McKellep, A.M., Refstie, S., Cyrino, J.E.P., Bureau, D.P., Kapoor, B.G., 2008. Alternative protein sources and digestive function alterations in teleost fishes. Feed. Dig. Funct. fishes 445–478.

Balasubramanian, M.N., Panserat, S., Dupont-Nivet, M., Quillet, E., Montfort, J., Le Cam, A., Medale, F., Kaushik, S.J., Geurden, I., 2016. Molecular pathways associated with the nutritional programming of plant-based diet acceptance in rainbow trout following an early feeding exposure. BMC Genomics 17, 449. doi:10.1186/s12864-016-2804-1

Baras, E., Malbrouck, C., Houbart, M., Kestemont, P., Melard, C., 2000. The effect of PIT tags on growth and physiology of age-O cultured Eurasian perch Perca fluviatilis of variable size. Aquaculture 185, 159–173.

Barker, D.J., Osmond, C., 1986. Infant mortality, childhood nutrition, and ischaemic heart disease in England and Wales. Lancet (London, England) 1, 1077–81.

Barnes, S., Grubbs, C., Setchell, K.D., Carlson, J., 1990. Soybeans inhibit mammary tumors in models of breast cancer. Prog. Clin. Biol. Res. 347, 239.

Bélanger-Deschênes, S., Couture, P., Campbell, P.G.C., Bernatchez, L., 2013. Evolutionary change driven by metal exposure as revealed by coding SNP genome scan in wild yellow perch (Perca flavescens). Ecotoxicology 22, 938–957. doi:10.1007/s10646-013-1083-8

Bennetau-Pelissero, C., Breton, B., Bennetau, B., Corraze, G., Le Menn, F., Davail- Cuisset, B., Helou, C., Kaushik, S.J., 2001. Effect of genistein-enriched diets on the endocrine process of gametogenesis and on reproduction efficiency of the rainbow trout Oncorhynchus mykiss. Gen. Comp. Endocrinol. 121, 173–187. doi:10.1006/gcen.2000.7585

Bennetts, H.W., Uuderwood, E.J., Shier, F.L., 1946. A Specific Breeding Problem of Sheep on Subterranean Clover Pastures in Western Australia. Aust. Vet. J. 22, 2–12. doi:10.1111/j.1751-0813.1946.tb15473.x

Berger, S.L., Kouzarides, T., Shiekhattar, R., Shilatifard, A., 2009. An operational definition of epigenetics. Genes Dev. 23, 781–3. doi:10.1101/gad.1787609 157

Betancur-R, R., Wiley, E.O., Arratia, G., Acero, A., Bailly, N., Miya, M., Lecointre, G., Ortí, G., 2017. Phylogenetic classification of bony fishes. BMC Evol. Biol. 17, 162. doi:10.1186/s12862-017-0958-3

Blomme, T., Vandepoele, K., De Bodt, S., Simillion, C., Maere, S., Van De Peer, Y., 2006. The gain and loss of genes during 600 million years of vertebrate evolution Gene duplication in vertebrate evolution Phylogenetic analysis of gene gain and loss during vertebrate evolution provides evidence for the importance of early gene or genome duplication events in evolution of complex vertebrates. Genome Biol. 7. doi:10.1186/gb-2006-7-5-r43

Bougas, B., Normandeau, E., Grasset, J., Defo, M.A., Campbell, P.G.C., Couture, P., Bernatchez, L., 2016. Transcriptional response of yellow perch to changes in ambient metal concentrations-A reciprocal field transplantation experiment. Aquat. Toxicol. 173, 132–142. doi:10.1016/j.aquatox.2015.12.014

Bougas, B., Normandeau, E., Pierron, F., Campbell, P.G.C., Bernatchez, L., Couture, P., 2013. How does exposure to nickel and cadmium affect the transcriptome of yellow perch (Perca flavescens) - Results from a 1000 candidate-gene microarray. Aquat. Toxicol. 142–143, 355–364. doi:10.1016/j.aquatox.2013.09.009

Brown, M.S., Goldstein, J.L., 1997. The SREBP pathway: regulation of cholesterol metabolism by proteolysis of a membrane-bound transcription factor. Cell 89, 331– 40. doi:10.1016/S0092-8674(00)80213-5

Brown, P.B., Dabrowski, K., Garling, D.L., 1996. Nutrition and feeding of yellow perch (Perca flavescens). J. Appl. Ichthyol. 12, 171–174. doi:10.1111/j.1439- 0426.1996.tb00084.x

Brown, T.A., 2006. Genomes 3 [Paperback], 3rd ed. Garland Science Pub., New York.

Buddington, R.K., Diamond, J.M., 1986. Aristotle revisited: The function of pyloric caeca in fish (intestine/nutrient uptake/digestion/adaptation). Physiol. Sci. Solut. Compos. 83, 8012–8014.

Burggren, W.W., 2014. Epigenetics as a source of variation in comparative animal physiology - or - Lamarck is lookin’ pretty good these days. J. Exp. Biol. 217, 682– 689. doi:10.1242/jeb.086132

Burr, G.S., Wolters, W.R., Barrows, F.T., Hardy, R.W., 2012. Replacing fishmeal with blends of alternative proteins on growth performance of rainbow trout (Oncorhynchus mykiss), and early or late stage juvenile Atlantic salmon (Salmo salar). Aquaculture 334–337, 110–116. doi:10.1016/j.aquaculture.2011.12.044 158

Callan, C.K., Laidley, C.W., Forster, I.P., Liu, K.M., Kling, L.J., Place, A.R., 2012. Examination of broodstock diet effects on egg production and egg quality in flame angelfish (Centropyge loriculus). Aquac. Res. 43, 696–705. doi:10.1111/j.1365- 2109.2011.02877.x

Cao, D., Kocabas, A., Ju, Z., Karsi, A., Li, P., Patterson, A., Liu, Z., 2001. Transcriptome of channel catfish (Ictalurus punctatus): Initial analysis of genes and expression profiles of the head kidney. Anim. Genet. 32, 169–188. doi:10.1046/j.1365- 2052.2001.00753.x

Carter, M.W., Matrone, G., Smart, W.W.., 1955. Effect of Genistin on Reproduction of the Mouse. J. Nutr. 55, 639–645.

Chacko, B.K., Chandler, R.T., D’Alessandro, T.L., Mundhekar, A., Khoo, N.K.H., Botting, N., Barnes, S., Patel, R.P., 2007. Anti-Inflammatory Effects of Isoflavones are Dependent on Flow and Human Endothelial Cell PPARγ1. J. Nutr. 137, 351– 356.

Chang, H.C., Churchwell, M.I., Delclos, K.B., Newbold, R.R., Doerge, D.R., 2000. Mass spectrometric determination of Genistein tissue distribution in diet-exposed Sprague-Dawley rats. J. Nutr. 130, 1963–1970.

Chang, T.-Y., Li, B.-L., Y Chang, C.C., Urano, Y., 2009. A:cholesterol acyltransferases. Am J Physiol Endocrinol Metab 297, 1–9. doi:10.1152/ajpendo.90926.2008

Chen, W., Ai, Q., Mai, K., Xu, W., Liufu, Z., Zhang, W., Cai, Y., 2011. Effects of dietary soybean saponins on feed intake, growth performance, digestibility and intestinal structure in juvenile Japanese flounder (Paralichthys olivaceus). Aquaculture 318, 95–100. doi:10.1016/j.aquaculture.2011.04.050

Chiang, J.Y.L., 2011. Bile Acid Metabolism, in: Molecular Pathology of Liver Diseases. Springer, Boston, MA, pp. 165–179. doi:10.1007/978-1-4419-7107-4_12

Chikwati, E.M., Sahlmann, C., Holm, H., Penn, M.H., Krogdahl, Å., Bakke, A.M., 2013. Alterations in digestive enzyme activities during the development of diet-induced enteritis in Atlantic salmon, Salmo salar L. Aquaculture 402–403, 28–37. doi:10.1016/j.aquaculture.2013.03.023

Cho, C.Y., Bayley, H.S., Linger, S.J., 1974. Partial Replacement of Herring Meal with Soybean Meal and Other Changes in a Diet for Rainbow Trout ( Salmo gairdneri ). J. Fish. Res. Board Canada 31, 1523–1528. doi:10.1139/f74-184

Choi, S.-W., Friso, S., 2010. Epigenetics: A New Bridge between Nutrition and Health. 159

Adv. Nutr. 1, 8–16. doi:10.3945/an.110.1004

Chou, R.L., Her, B.Y., Su, M.S., Hwang, G., Wu, Y.H., Chen, H.Y., 2004. Substituting fish meal with soybean meal in diets of juvenile cobia Rachycentron canadum. Aquaculture 229, 325–333. doi:10.1016/S0044-8486(03)00395-8

Ciereszko, A., Dabrowski, K., 1993. Estimation of sperm concentration of rainbow trout, whitefish and yellow perch using a spectrophotometric technique. Aquaculture 109, 367–373. doi:10.1016/0044-8486(93)90175-X

Clarkson, M., Migaud, H., Metochis, C., Vera, L.M., Leeming, D., Tocher, D.R., Taylor, J.F., 2017. Early nutritional intervention can improve utilisation of vegetable-based diets in diploid and triploid Atlantic salmon (Salmo salar L.). Br. J. Nutr. 118, 1–13. doi:10.1017/S0007114517001842

Collie, N.L., Ferraris, R.P., 1995. Nutrient fluxes and regulation in fish intestine, in: Hochachka, P.., Mommsen, T.P. (Eds.), Biochemistry and Molecular Biology of Fishes. pp. 221–239.

Collins, S.A., Desai, A.R., Mansfield, G.S., Hill, J.E., Van Kessel, A.G., Drew, M.D., 2012. The effect of increasing inclusion rates of soybean, pea and canola meals and their protein concentrates on the growth of rainbow trout: Concepts in diet formulation and experimental design for ingredient evaluation. Aquaculture 344– 349, 90–99. doi:10.1016/j.aquaculture.2012.02.018

D’Souza, N., Skonberg, D.I., Camire, M.E., Guthrie, K.E., Malison, J., Lima, L., 2005. Influence of dietary genistein levels on tissue genistein deposition and on the physical, chemical, and sensory quality of rainbow trout, Oncorhynchus mykiss. J. Agric. Food Chem. 53, 3631–3636. doi:10.1021/jf0484024

D’Souza, N., Skonberg, D.I., Stone, D.A.J., Brown, P.B., 2006. Effect of soybean metal- based diets on the product quality of rainbow trout fillets. J. Food Sci. 71. doi:10.1111/j.1750-3841.2006.00018.x

Dabrowska, H., Wojno, T., 1977. Studies on the utilization by rainbow trout (Safmo gairdneri Rich.) of feed mixtures containing soya bean meal and an addition of amino acids. Aquaculture 10, 297–310.

Dabrowski, K., Ciereszko, A., 1996. The dynamics of gonad growth and ascorbate status in yellow perch, Perca flavescens (Mitchill). Aquac. Res. 27, 539–542. doi:10.1111/j.1365-2109.1996.tb01284.x

Dabrowski, K., Ciereszko, R.E., Ciereszko, A., Toth, G.P., Christ, S.A., El-Saidy, D., Ottobre, J.S., 1996. Reproductive physiology of yellow perch (Perca flavescens): 160

environmental and endcrinological cues. J. Appl. Ichthyol. 12, 139–148. doi:10.1111/j.1439-0426.1996.tb00079.x

Danciu, C., Antal, D.S., Ardelean, F., Chiş, A.R., Şoica, C., Andrica, F., Dehelean, C., 2017. New Insights Regarding the Potential Health Benefits of Isoflavones, in: Flavonoids - From Biosynthesis to Human Health. InTech. doi:10.5772/67896

Day, A.J., Dupont, M.S., Ridley, S., Rhodes, M., Rhodes, M.J.., Morgan, M.R.., Williamson, G., 1998. Deglycosylation of flavonoid and isoflavonoid glycosides by human small intestine and liver β-glucosidase activity. FEBS Lett. 436, 71–75. doi:10.1016/S0014-5793(98)01101-6

De Santis, C., Bartie, K.L., Olsen, R.E., Taggart, J.B., Tocher, D.R., 2015. Nutrigenomic profiling of transcriptional processes affected in liver and distal intestine in response to a soybean meal-induced nutritional stress in Atlantic salmon (Salmo salar). Comp. Biochem. Physiol. - Part D Genomics Proteomics 15, 1–11. doi:10.1016/j.cbd.2015.04.001

Deng, J., Bi, B., Kang, B., Kong, L., Wang, Q., Zhang, X., 2013. Improving the growth performance and cholesterol metabolism of rainbow trout (Oncorhynchus mykiss) fed soyabean meal-based diets using dietary cholesterol supplementation. Br. J. Nutr. 110, 29–39. doi:10.1017/S0007114512004680

Devlaming, V., Grossman, G., Chapman, F., 1982. On the use of the Gonosomatic Index. Comp. Biochem. Physiol. Part A Physiol. 73A, 31–39. doi:https://doi.org/10.1016/0300-9629(82)90088-3

Dias, J., Alvarez, M.J., Arzel, J., Corraze, G., Diez, A., Bautista, J.M., Kaushik, S.J., 2005. Dietary protein source affects lipid metabolism in the European seabass (Dicentrarchus labrax). Comp. Biochem. Physiol. A. Mol. Integr. Physiol. 142, 19– 31. doi:10.1016/j.cbpb.2005.07.005

Dixon, R.A., Ferreira, D., 2002. Genistein. Phytochemistry 60, 205–211. doi:10.1016/S0031-9422(02)00116-4

Elangovan, A., Shim, K.F., 2000. The influence of replacing fish meal partially in the diet with soybean meal on growth and body composition of juvenile tin foil barb (Barbodes altus). Aquaculture 189, 133–144. doi:10.1016/S0044-8486(00)00365-3

Eldridge, A.C., Kwolek, W.F., 1983. Soybean Isoflavones: Effect of Environment and Variety on Composition. J. Agric. Food Chem. 31, 394–396. doi:10.1021/jf00116a052

Environment and Climate Change Canada and the U.S. Environmental Protection 161

Agency, 2017. State of the Great Lakes 2017-Technical Report.

Erdman, J.W., Badger, T.M., Lampe, J.W., Setchell, K.D.R., Messina, M., 2004. Not all soy products are created equal: caution needed in interpretation of research results. J. Nutr. 134, 1229S–1233S.

Fang, L., Liang, X.-F., Zhou, Y., Guo, X.-Z., He, Y., Yi, T.-L., Liu, L.-W., Yuan, X.-C., Tao, Y.-X., 2014. Programming effects of high-carbohydrate feeding of larvae on adult glucose metabolism in zebrafish, Danio rerio. Br. J. Nutr. 111, 808–18. doi:10.1017/S0007114513003243

Farmer, T.M., Marschall, E.A., Dabrowski, K., Ludsin, S.A., 2015. Short winters threaten temperate fish populations. Nat. Commun. 6, 7724. doi:10.1038/ncomms8724

Feiner, Z.S., Coulter, D.P., Guffey, S.C., Höök, T.O., 2016. Does overwinter temperature affect maternal body composition and egg traits in yellow perch Perca flavescens? J. Fish Biol. 88, 1524–1543. doi:10.1111/jfb.12929

Fernández-Palacios, H., Norberg, B., Izquierdo, M., Hamre, K., 2011. Effects of Broodstock Diet on Eggs and Larvae, in: Larval Fish Nutrition. Wiley-Blackwell, Oxford, UK, pp. 151–181. doi:10.1002/9780470959862.ch5

Folch, J., Lees, M., Sloane Stanley, G.H., 1956. A Simple Method for the isolation and purification of total lipides from animal tissues. J. Biol. Chem. 226, 497–509.

Fontagné-Dicharry, S., Alami-Durante, H., Aragão, C., Kaushik, S.J., Geurden, I., 2017. Parental and early-feeding effects of dietary methionine in rainbow trout (Oncorhynchus mykiss). Aquaculture 469, 16–27. doi:10.1016/j.aquaculture.2016.11.039

Food and Agriculture Organization of the United Nations, 2016. The State of World Fisheries and Aquaculture, Contributing to food security and nutrition for all. Rome, Italy.

Francis, G., Makkar, H.P.. S., Becker, K., 2001. Antinutritional factors present in plant- derived alternate fish feed ingrediants and their effects in fish, Aquaculture. doi:10.1016/S0044-8486(01)00526-9

Ganai, A.A., Farooqi, H., 2015. Bioactivity of genistein: A review of in vitro and in vivo studies. Biomed. Pharmacother. 76, 30–38. doi:10.1016/j.biopha.2015.10.026

Gatlin, D.M., Barrows, F.T., Brown, P., Dabrowski, K., Gaylord, T.G., Hardy, R.W., Herman, E., Hu, G., Krogdahl, Å., Nelson, R., Overturf, K., Rust, M., Sealey, W., Skonberg, D., Souza, E.J., Stone, D., Wilson, R., Wurtele, E., 2007. Expanding the 162

utilization of sustainable plant products in aquafeeds: A review. Aquac. Res. 38, 551–579. doi:10.1111/j.1365-2109.2007.01704.x

Geay, F., Ferraresso, S., Zambonino-Infante, J.L., Bargelloni, L., Quentel, C., Vandeputte, M., Kaushik, S., Cahu, C.L., Mazurais, D., 2011. Effects of the total replacement of fish-based diet with plant-based diet on the hepatic transcriptome of two European sea bass (Dicentrarchus labrax) half-sibfamilies showing different growth rates with the plant-based diet. BMC Genomics 12. doi:10.1186/1471-2164- 12-522

Gentry, R.R., Froehlich, H.E., Grimm, D., Kareiva, P., Parke, M., Rust, M., Gaines, S.D., Halpern, B.S., 2017. Mapping the global potential for marine aquaculture. Nat. Ecol. Evol. doi:10.1038/s41559-017-0257-9

Geurden, I., Aramendi, M., Zambonino-Infante, J., Panserat, S., 2007. Early feeding of carnivorous rainbow trout (Oncorhynchus mykiss) with a hyperglucidic diet during a short period: effect on dietary glucose utilization in juveniles. AJP Regul. Integr. Comp. Physiol. 292, R2275–R2283. doi:10.1152/ajpregu.00444.2006

Geurden, I., Borchert, P., Balasubramanian, M.N., Schrama, J.W., Dupont-Nivet, M., Quillet, E., Kaushik, S.J., Panserat, S., Médale, F., 2013. The positive impact of the early-feeding of a plant-based diet on its future acceptance and utilisation in rainbow trout. PLoS One 8. doi:10.1371/journal.pone.0083162

Goetz, F.W., Rise, M.L.M., Rise, M.L.M., Goetz, G.W., Binkowski, F., Shepherd, B.S., 2009. Stimulation of growth and changes in the hepatic transcriptome by 17beta- estradiol in the yellow perch (Perca flavescens). Physiol. Genomics 38, 261–280. doi:10.1152/physiolgenomics.00069.2009

Gontier-Latonnelle, K., Cravedi, J.P., Laurentie, M., Perdu, E., Lamothe, V., Le Menn, F., Bennetau-Pelissero, C., 2007. Disposition of genistein in rainbow trout (Oncorhynchus mykiss) and siberian sturgeon (Acipenser baeri). Gen. Comp. Endocrinol. 150, 298–308. doi:10.1016/j.ygcen.2006.10.001

Grasset, J., Bougas, B., Campbell, P.G.C., Bernatchez, L., Couture, P., 2014. Temperature , oxygen , and diet modulate gene transcription and metabolic capacities in yellow perch. Can. J. Fish. Aquat. Sci. 71, 1635–1641. doi:10.1139/cjfas-2014-0059

Gu, M., Kortner, T.M., Penn, M., Hansen, A.K., Krogdahl, Å., 2014. Effects of dietary plant meal and soya-saponin supplementation on intestinal and hepatic lipid droplet accumulation and lipoprotein and sterol metabolism in Atlantic salmon (Salmo salar L.). Br. J. Nutr. 111, 432–444. doi:10.1017/S0007114513002717

163

Haas, B.J., Papanicolaou, A., Yassour, M., Grabherr, M., Blood, P.D., Bowden, J., Couger, M.B., Eccles, D., Li, B., Lieber, M., Macmanes, M.D., Ott, M., Orvis, J., Pochet, N., Strozzi, F., Weeks, N., Westerman, R., William, T., Dewey, C.N., Henschel, R., Leduc, R.D., Friedman, N., Regev, A., 2013. De novo transcript sequence reconstruction from RNA-Seq: reference generation and analysis with Trinity. Nat. Protoc. 8, 1494–1512. doi:10.1038/nprot

Hamoutene, D., Lush, L., Drover, D., Walsh, A., 2009. Investigation of the temporal effects of spawning season and maternal and paternal differences on egg quality in Atlantic cod Gadus morhua L. broodstock. Aquac. Res. 40, 1668–1679. doi:10.1111/j.1365-2109.2009.02271.x

Hanley, B., Dijane, J., Fewtrell, M., Grynberg, A., Hummel, S., Junien, C., Koletzko, B., Lewis, S., Renz, H., Symonds, M., Gros, M., Harthoorn, L., Mace, K., Samuels, F., van Der Beek, E.M., 2010. Metabolic imprinting, programming and epigenetics – a review of present priorities and future opportunities. Br. J. Nutr. 104, S1–S25. doi:10.1017/S0007114510003338

Hansen, A.-C., Rosenlund, G., Karlsen, Ø., Koppe, W., Hemre, G.-I., 2007. Total replacement of fish meal with plant proteins in diets for Atlantic cod (Gadus morhua L.) I — Effects on growth and protein retention. Aquaculture 272, 599–611. doi:10.1016/j.aquaculture.2007.08.034

Hardy, R.W., 2010. Utilization of plant proteins in fish diets: Effects of global demand and supplies of fishmeal. Aquac. Res. 41, 770–776. doi:10.1111/j.1365- 2109.2009.02349.x

Hart, S.D., Garling, D.L., Malison, J.A., 2006. Yellow Perch (Perca flavescens) Culture Guide- NCRAC Culture Series #103.

Hayes, D.B., Taylor, W.W., 1994. Changes in the composition of somatic and gonadal tissues of yellow perch following white sucker removal. Trans. Am. Fish. Soc. 123, 204–216. doi:10.1577/1548-8659(1994)123<0204:citcos>2.3.co;2

Heard, E., Martienssen, R.A., 2014. Transgenerational epigenetic inheritance: myths and mechanisms. Cell 157, 95–109. doi:10.1016/j.cell.2014.02.045

Heikkinen, J., Vielma, J., Kemiläinen, O., Tiirola, M., Eskelinen, P., Kiuru, T., Navia- Paldanius, D., von Wright, A., 2006. Effects of soybean meal based diet on growth performance, gut histopathology and intestinal microbiota of juvenile rainbow trout (Oncorhynchus mykiss). Aquaculture 261, 259–268. doi:10.1016/j.aquaculture.2006.07.012

Henderson, B.A., Trivedi, T., Collins, N., 2000. Annual cycle of energy allocation to 164

growth and reproduction of yellow perch. J. Fish Biol. 57, 122–133. doi:10.1111/j.1095-8649.2000.tb00780.x

Hokanson, K.E.F., 1977. Temperature Requirements of Some Percids and Adaptations to the Seasonal Temperature Cycle. J. Fish. Res. Board Canada 34, 1524–1550. doi:10.1139/f77-217

Holland, M.L., Lowe, R., Caton, P.W., Gemma, C., Carbajosa, G., Danson, A.F., Carpenter, A.A.M., Loche, E., Ozanne, S.E., Rakyan, V.K., 2016. Early-life nutrition modulates the epigenetic state of specific rDNA genetic variants in mice. Science (80-. ). 353, 495–498. doi:10.1126/science.aaf7040

Honaas, L.A., Wafula, E.K., Wickett, N.J., Der, J.P., Zhang, Y., Edger, P.P., Altman, N.S., Pires, J.C., Leebens-Mack, J.H., Depamphilis, C.W., 2016. Selecting Superior De Novo Transcriptome Assemblies: Lessons Learned by Leveraging the Best Plant Genome. doi:10.1371/journal.pone.0146062

Inbaraj, B.S., Chen, B.H., 2012. Isoflavones in Foods and Ingestion in the Diet, in: Preedy, V.R. (Ed.), Isoflavones: Chemistry, Analysis, Function and Effects. RSC Publishing.

Indexmundi.com, 2018. Soybean Oil - Monthly Price - Commodity Prices - Price Charts, Data, and News [WWW Document]. Index Mundi. URL https://www.indexmundi.com/commodities/?commodity=soybean-meal (accessed 1.2.18).

Izquierdo, M.S., Fernández-Palacios, H., Tacon, A.G.J., 2001. Effect of broodtsock nutrition on reproductive performance of fish. Aquaculture 197, 25–42.

Izquierdo, M.S., Turkmen, S., Montero, D., Zamorano, M.J., Afonso, J.M., Karalazos, V., Fernández-Palacios, H., 2015. Nutritional programming through broodstock diets to improve utilization of very low fishmeal and fish oil diets in gilthead sea bream. Aquaculture 449, 18–26. doi:10.1016/j.aquaculture.2015.03.032

Jaenisch, R., Bird, A., 2003. Epigenetic regulation of gene expression: how the genome integrates intrinsic and environmental signals. Nat. Genet. 33, 245–254. doi:10.1038/ng1089

Jansen, W.A., 1996. Plasticity in maturity and fecundity of yellow perch , Perca flavescens (Mitchill): comparisons of stunted and normal-growing populations. Ann. Zool. Fennici 33, 403–415.

Jobling, M., 2016. Fish nutrition research: past, present and future. Aquac. Int. 24, 767– 786. doi:10.1007/s10499-014-9875-2 165

Jonsson, B., Jonsson, N., 2014. Early environment influences later performance in fishes. J. Fish Biol. 85, 151–188. doi:10.1111/jfb.12432

Jordal, A.-E.O., Torstensen, B.E., Tsoi, S., Tocher, D.R., Lall, S.P., Douglas, S.E., 2005. Dietary rapeseed oil affects the expression of genes involved in hepatic lipid metabolism in Atlantic salmon (Salmo salar L.). J. Nutr. 135, 2355–61. doi:135/10/2355 [pii]

Ju, Z., Karsi, A., Kocabas, A., Patterson, A., Li, P., Cao, D., Dunham, R., Liu, Z., 2000. Transcriptome analysis of channel catfish (Ictalurus punctatus): Genes and expression profile from the brain. Gene 261, 373–382. doi:10.1016/S0378- 1119(00)00491-1

Kaneko, T.P., 1969. Composition of food for carp and rainbow trout. EIFAC Tech. Pap. No. 9 161–168.

Kao, T.H., Chen, B.H., 2006. Functional components in soybean cake and their effects on antioxidant activity. J. Agric. Food Chem. 54, 7544–7555. doi:10.1021/jf061586x

Kasper, C.S., Watkins, B.A., Brown, P.B., 2007. Evaluation of two soybean meals fed to yellow perch (Perca flavescens). Aquac. Nutr. 13, 431–438. doi:10.1111/j.1365- 2095.2007.00494.x

Kaushik, S.J., Covè, D., Dutto, G., Blanc, D., 2004. Almost total replacement of fish meal by plant protein sources in the diet of a marine teleost, the European seabass, Dicentrarchus labrax. Aquaculture 230, 391–404. doi:10.1016/S0044- 8486(03)00422-8

Kaushik, S.J., Cravedi, J.P., Lalles, J.P., Sumpter, J., Fauconneau, B., Laroche, M., 1995. Partial or total replacement of fish meal by soybean protein on growth, protein utilization, potential estrogenic or antigenic effects, cholesterolemia and flesh quality in rainbow trout, Oncorhynchus mykiss. Aquaculture 133, 257–274. doi:10.1016/0044-8486(94)00403-B

Kemski, M., Wick, M., Dabrowski, K., 2018. Nutritional programming effects on growth and reproduction of broodstock and embryonic development of progeny in yellow perch (Perca flavescens) fed soybean meal-based diets. Aquaculture. doi:10.1016/J.AQUACULTURE.2018.07.001

Kendall, K.A., Salisbury, G.W., Vandemabk, N.L., 1950. Sterility in the Rabbit Associated with Soybean Hay Feeding. J. Nutr. 42, 487–500.

Kestemont, P., Dabrowski, K., Summerfelt, R.C., 2015. Biology and Culture of Percid 166

Fishes Principles and Practices. Springer Berlin Heidelberg.

Kijkuokool, P., Parhar, I.S., Malaivijitnond, S., 2006. Genistein enhances N- nitrosomethylurea-induced rat mammary tumorigenesis. Cancer Lett. 242, 53–59. doi:10.1016/j.canlet.2005.10.033

Klinger, D., Naylor, R., 2012. Searching for Solutions in Aquaculture: Charting a Sustainable Course. Annu. Rev. Environ. Resour. 37, 247–276. doi:10.1146/annurev-environ-021111-161531

Ko, K., Malison, J.A., Reed, J.D., 1999. Effect of Genistein on the Growth and Reproductive Function of Male and Female Yellow Perch Perca flavescens. J. World Aquac. Soc. 30, 73–79. doi:10.1111/j.1749-7345.1999.tb00319.x

Kocovsky, P.M., Knight, C.T., 2012. Morphological evidence of discrete stocks of yellow perch in Lake Erie. J. Great Lakes Res. 38, 534–539. doi:10.1016/j.jglr.2012.04.006

Kortner, T.M., Björkhem, I., Krasnov, A., Timmerhaus, G., Krogdahl, Å., 2014. Dietary cholesterol supplementation to a plant-based diet suppresses the complete pathway of cholesterol synthesis and induces bile acid production in Atlantic salmon (Salmo salar L.). Br. J. Nutr. 111, 2089–2103. doi:10.1017/S0007114514000373

Kortner, T.M., Gu, J., Krogdahl, Å., Bakke, A.M., 2013. Transcriptional regulation of cholesterol and bile acid metabolism after dietary soyabean meal treatment in Atlantic salmon (Salmo salar L.). Br. J. Nutr. 109, 593–604. doi:10.1017/S0007114512002024

Kortner, T.M., Skugor, S., Penn, M.H., Mydland, L.T., Djordjevic, B., Hillestad, M., Krasnov, A., Krogdahl, Å., 2012. Dietary soyasaponin supplementation to pea protein concentrate reveals nutrigenomic interactions underlying enteropathy in Atlantic salmon (Salmo salar). BMC Vet. Res. 8. doi:10.1186/1746-6148-8-101

Kortner, T.M., Valen, E.C., Kortner, H., Marjara, I.S., Krogdahl, S., Bakke, A.M., 2011. Candidate reference genes for quantitative real-time PCR (qPCR) assays during development of a diet-related enteropathy in Atlantic salmon (Salmo salar L.) and the potential pitfalls of uncritical use of normalization software tools. Aquaculture 318, 355–363. doi:10.1016/j.aquaculture.2011.05.038

Koven, W., Van Anholt, R., Lutzky, S., Atia, B., Nixon, O., Ron, B., Tandler, A., 2003. The effect of dietary arachidonic acid on growth, survival, and cortisol levels in different-age gilthead seabream larvae (Sparus auratus) exposed to handling or daily salinity change. Aquaculture 228, 307–320. doi:10.1016/S0044-8486(03)00317-X

167

Krogdahl, Å., Bakke-McKellep, a M., Røed, K.H., Baeverfjord, G., 2000. Feeding Atlantic salmon Salmo salar L. soybean products: effects on disease resistance (furunculosis), and lysozyme and IgM levels in the intestinal mucosa. Aquac. Nutr. 6, 77–84. doi:10.1046/j.1365-2095.2000.00129.x

Krogdahl, Å., Bakke-McKellep, A.M., Baeverfjord, G., 2003. Effects of graded levels of standard soybean meal on intestinal structure, mucosal enzyme activities, and pancreatic response in Atlantic salmon (Salmo solar L.). Aquac. Nutr. 9, 361–371. doi:10.1046/j.1365-2095.2003.00264.x

Krogdahl, Å., Gajardo, K., Kortner, T.M., Penn, M., Gu, M., Berge, G.M., Bakke, A.M., 2015. Soya Saponins Induce Enteritis in Atlantic Salmon (Salmo salar L.). J. Agric. Food Chem. 63, 3887–3902. doi:10.1021/jf506242t

Krogdahl, Å., Penn, M., Thorsen, J., Refstie, S., Bakke, A.M., 2010. Important antinutrients in plant feedstuffs for aquaculture: An update on recent findings regarding responses in salmonids. Aquac. Res. 41, 333–344. doi:10.1111/j.1365- 2109.2009.02426.x

Król, E., Douglas, A., Tocher, D.R., Crampton, V.O., Speakman, J.R., Secombes, C.J., Martin, S.A.M., 2016. Differential responses of the gut transcriptome to plant protein diets in farmed Atlantic salmon. BMC Genomics 17, 156. doi:10.1186/s12864-016-2473-0

Kwasek, K., Dabrowski, K., Nynca, J., Takata, R., Wojno, M., Wick, M., 2014. The Influence of Dietary Lysine on Yellow Perch Female Reproductive Performance and the Quality of Eggs. N. Am. J. Aquac. 76, 351–358. doi:10.1080/15222055.2014.911223

Kwasek, K., Dabrowski, K., Ware, K., Reddish, J.M., Wick, M., 2012. The effect of lysine-supplemented wheat gluten-based diet on yellow perch Perca flavescens (Mitchill) performance. Aquac. Res. 43, 1384–1391. doi:10.1111/j.1365- 2109.2011.02941.x

Lazzarotto, V., Corraze, G., Larroquet, L., Mazurais, D., Médale, F., 2016. Does broodstock nutritional history affect the response of progeny to different first- feeding diets? A whole-body transcriptomic study of rainbow trout alevins. Br. J. Nutr. 115, 2079–2092. doi:10.1017/S0007114516001252

Lazzarotto, V., Corraze, G., Leprevost, A., Quillet, E., Dupont-Nivet, M., Médale, F., 2015. Three-year breeding cycle of rainbow trout (Oncorhynchus mykiss) fed a plant-based diet, totally free of marine resources: Consequences for reproduction, fatty acid composition and progeny survival. PLoS One 10. doi:10.1371/journal.pone.0117609 168

Lazzarotto, V., Médale, F., Larroquet, L., Corraze, G., 2018. Long-term dietary replacement of fishmeal and fish oil in diets for rainbow trout (Oncorhynchus mykiss): Effects on growth, whole body fatty acids and intestinal and hepatic gene expression. PLoS One. doi:10.1371/journal.pone.0190730

Le Boucher, R., Dupont-Nivet, M., Vandeputte, M., Kerneïs, T., Goardon, L., Labbé, L., Chatain, B., Bothaire, M.J., Larroquet, L., Médale, F., Quillet, E., 2012. Selection for Adaptation to Dietary Shifts: Towards Sustainable Breeding of Carnivorous Fish. PLoS One 7, 3–10. doi:10.1371/journal.pone.0044898

Le Boucher, R., Quillet, E., Vandeputte, M., Lecalvez, J.M., Goardon, L., Chatain, B., Médale, F., Dupont-Nivet, M., 2011. Plant-based diet in rainbow trout (Oncorhynchus mykiss Walbaum): Are there genotype-diet interactions for main production traits when fish are fed marine vs. plant-based diets from the first meal? Aquaculture 321, 41–48. doi:10.1016/j.aquaculture.2011.08.010

Lee, S.J., Ahn, J.K., Kim, S.H., Kim, J.T., Han, S.J., Jung, M.Y., Chung, I.M., 2003a. Variation in isoflavone of soybean cultivars with location and storage duration. J. Agric. Food Chem. 51, 3382–3389. doi:10.1021/jf0261405

Lee, S.J., Yan, W., Ahn, J.K., Chung, I.M., 2003b. Effects of year, site, genotype and their interactions on various soybean isoflavones. F. Crop. Res. 81, 181–192. doi:10.1016/S0378-4290(02)00220-4

Li, Y.-H., Wang, H.-P., Yao, H., O’Bryant, P., Rapp, D., Guo, L., Waly, E.A., 2017. De novo transcriptome sequencing and analysis of male, pseudo-male and female yellow perch, Perca flavescens. PLoS One 12, e0171187. doi:10.1371/journal.pone.0171187

Lilleeng, E., Froystad, M.K., Ostby, G.C., Valen, E.C., Krogdahl, A., 2007. Effects of diets containing soybean meal on trypsin mRNA expression and activity in Atlantic salmon (Salmo salar L). Comp. Biochem. Physiol. A. Mol. Integr. Physiol. 147, 25– 36. doi:10.1016/j.cbpa.2006.10.043

Lillycrop, K.A., Hoile, S.P., Grenfell, L., Burdge, G.C., 2014. DNA methylation, ageing and the influence of early life nutrition, in: Proceedings of the Nutrition Society. pp. 413–421. doi:10.1017/S0029665114000081

Lin, F., Wu, J., Abdelnabi, M.A., Ottinger, M.A., Giusti, M.M., 2004. Effects of Dose and Glycosylation on the Transfer of Genistein into the Eggs of the Japanese Quail (Coturnix japonica). J. Agric. Food Chem. 52, 2397–2403. doi:10.1021/jf034921f

Lin, S., Luo, L., 2011. Effects of different levels of soybean meal inclusion in 169

replacement for fish meal on growth, digestive enzymes and transaminase activities in practical diets for juvenile tilapia, Oreochromis niloticus×O. aureus. Anim. Feed Sci. Technol. 168, 80–87. doi:10.1016/j.anifeedsci.2011.03.012

Lindberg, J., Lundeberg, J., 2009. The plasticity of the mammalian transcriptome. Genomics 95, 1–6. doi:10.1016/j.ygeno.2009.08.010

Little, D.C., Newton, R.W., Beveridge, M.C.M., 2016. Aquaculture: A rapidly growing and significant source of sustainable food? Status, transitions and potential, in: Proceedings of the Nutrition Society. pp. 274–286. doi:10.1017/S0029665116000665

Lucas, A., 1998. Symposium: The Effects of Childhood Diet on Adult Health and Disease Programming by Early Nutrition: An Experimental Approach 1. J. Nutr. 128, 401S–406S.

Maita, M., Maekawa, J., Satoh, K.I., Futami, K., Satoh, S., 2006. Disease resistance and hypocholesterolemia in yellowtail Seriola quinqueradiata fed a non-fishmeal diet. Fish. Sci. 72, 513–519. doi:10.1111/j.1444-2906.2006.01179.x

Malison, J.A., Kayes, T.B., Best, C.D., Amundson, C.H., Wentworth, B.C., 1986. Sexual Differentiation and Use of Hormones to Control Sex in Yellow Perch ( Perca flavescens ). Can. J. Fish. Aquat. Sci. 43, 26–35. doi:10.1139/f86-004

Maloy, K.J., Powrie, F., 2011. Intestinal homeostasis and its breakdown in inflammatory bowel disease. Nature 474, 298–306. doi:10.1038/nature10208

Mambrini, M., Roem, A.J., Cravèdi, J.P., Lallès, J.P., Kaushik, S.J., 1999. Effects of replacing fish meal with soy protein concentrate and of DL-methionine supplementation in high-energy, extruded diets on the growth and nutrient utilization of rainbow trout, Oncorhynchus mykiss. J. Anim. Sci. 77, 2990–2999. doi:10.2527/1999.77112990x

Marjara, I.S., Chikwati, E.M., Valen, E.C., Krogdahl, Å., Bakke, A.M., 2012. Transcriptional regulation of IL-17A and other inflammatory markers during the development of soybean meal-induced enteropathy in the distal intestine of Atlantic salmon (Salmo salar L.). Cytokine 60, 186–196. doi:10.1016/j.cyto.2012.05.027

Martin, S.A.M.M., Dehler, C.E., Zbieta, E._, Ol, K., Krol, E., Król, E., 2016. Transcriptomic responses in the fish intestine. Dev. Comp. Immunol. 64, 103–117. doi:10.1016/j.dci.2016.03.014

Martin, S.A.M.M., Zbieta, E._, Ol, K., Król, E., 2017. Nutrigenomics and immune function in fish: new insights from omics technologies. Dev. Comp. Immunol. 75, 170

86–98. doi:10.1016/j.dci.2017.02.024

Mcgettigan, P.A., 2013. Transcriptomics in the RNA-seq era. Curr. Opin. Chem. Biol. 17, 4–11. doi:10.1016/j.cbpa.2012.12.008

Merlanti, R., Lucatello, L., De Jesus Inacio, L., Pastore, M.R., Laverda, S., Capolongo, F., 2018. Isoflavones quantification in rainbow trout muscle by QuEChERS tecnique and liquid chromatography coupled with mass spectrometry. J. Food Compos. Anal. 70, 114–124. doi:10.1016/j.jfca.2018.04.009

Messina, M., 2010. A brief historical overview of the past two decades of soy and isoflavone research. J. Nutr. 140, 1350S–4S. doi:10.3945/jn.109.118315.isoflavone

Messina, M., Barnes, S., 1991. The role of soy products in reducing risk of cancer. J. Natl. Cancer Inst. 83, 541–546. doi:10.1093/jnci/83.8.541

Messina, M., Nagata, C., Wu, A.H., 2006. Estimated Asian adult soy protein and isoflavone intakes. Nutr. Cancer. doi:10.1207/s15327914nc5501_1

Messina, M.J., 1997. Soybeanfoods: their role in disease prevention and treatment. Soybeans Chem. Technol. Util. 442–447. doi:10.1007/978-1-4615-1763-4_10

Michel, C., Bernatchez, L., Behrmann-Godel, J., 2009. Diversity and evolution of MHII β genes in a non-model percid species—The Eurasian perch (Perca fluviatilis L.). Mol. Immunol. 46, 3399–3410. doi:10.1016/j.molimm.2009.07.012

Michl, S.C., Weis, B., Hutchings, J.A., Schulz, C., 2017. Plastic responses by wild brown trout (Salmo trutta) to plant-based diets. Aquaculture 476, 19–28. doi:10.1016/J.AQUACULTURE.2017.04.006

Milagro, F.I.I., Mansego, M.L.L., De Miguel, C., Martínez, J.A., 2013. Dietary factors, epigenetic modifications and obesity outcomes: Progresses and perspectives. Mol. Aspects Med. 34, 782–812. doi:10.1016/j.mam.2012.06.010

Moldal, T., Løkka, G., Wiik-Nielsen, J., Austbø, L., Torstensen, B.E., Rosenlund, G., Dale, O., Kaldhusdal, M., Koppang, E., 2014. Substitution of dietary fish oil with plant oils is associated with shortened mid intestinal folds in Atlantic salmon (Salmo salar). BMC Vet. Res. 10, 60. doi:10.1186/1746-6148-10-60

Morais, S., Edvardsen, R.B., Tocher, D.R., Gordon Bell, J., Bell, J.G., 2012a. Transcriptomic analyses of intestinal gene expression of juvenile Atlantic cod (Gadus morhua) fed diets with Camelina oil as replacement for fish oil. Comp. Biochem. Physiol. Part B Biochem. Mol. Biol. 161, 283–293. doi:10.1016/j.cbpb.2011.12.004 171

Morais, S., Pratoomyot, J., Taggart, J.B., Bron, J.E., Guy, D.R., Bell, J.G., Tocher, D.R., 2011. Genotype-specific responses in Atlantic salmon (Salmo salar) subject to dietary fish oil replacement by vegetable oil: A liver transcriptomic analysis. BMC Genomics 12. doi:10.1186/1471-2164-12-255

Morais, S., Silva, T., Cordeiro, O., Rodrigues, P., Guy, D.R., Bron, J.E., Taggart, J.B., Bell, J.G., Tocher, D.R., 2012b. Effects of genotype and dietary fish oil replacement with vegetable oil on the intestinal transcriptome and proteome of Atlantic salmon (Salmo salar). BMC Genomics 13. doi:10.1186/1471-2164-13-448

Mortazavi, A., Williams, B.A., McCue, K., Schaeffer, L., Wold, B., 2008. Mapping and quantifying mammalian transcriptomes by RNA-Seq. Nat. Methods 5, 621–628. doi:10.1038/nmeth.1226

Moyad, M.A., 1999. Soy, disease prevention, and prostate cancer. Semin Urol Oncol 17, 97–102.

National Research Council (NRC), 2011. Nutrient Requirements of Fish and Shrimp. The National Academies Press, Washington.

Ng, Y., Hanson, S., Malison, J.A., Wentworth, B., Barry, T.P., 2006. Genistein and other isoflavones found in soybeans inhibit estrogen metabolism in salmonid fish. Aquaculture 254, 658–665. doi:10.1016/j.aquaculture.2005.10.039

Nguyen, H.Q., Reinertsen, H., Rustad, T., Tran, T.M., Kjørsvik, E., 2012. Evaluation of egg quality in broodstock cobia Rachycentron canadum L. Aquac. Res. 43, 371– 385. doi:10.1111/j.1365-2109.2011.02840.x

Niculescu, M.D., Haggarty, P., 2011. Nutrition in Epigenetics, Nutrition in Epigenetics. Wiley-Blackwell, Oxford, UK. doi:10.1002/9780470959824

Noteboom, W.D., Gorski, J., 1963. Estrogenic Effect of Genistein and Coumestrol Diacetate. Endocrinology 73, 736–739. doi:10.1210/endo-73-6-736

Oliva-Teles, A., Enes, P., Peres, H., 2015. Replacing fishmeal and fish oil in industrial aquafeeds for carnivorous fish, in: Feed and Feeding Practices in Aquaculture. Elsevier, pp. 203–233. doi:10.1016/B978-0-08-100506-4.00008-8

Ostaszewska, T., Dabrowski, K., Palacios, M.E., Olejniczak, M., Wieczorek, M., 2005. Growth and morphological changes in the digestive tract of rainbow trout (Oncorhynchus mykiss) and pacu (Piaractus mesopotamicus) due to casein replacement with soybean proteins. Aquaculture 245, 273–286. doi:10.1016/j.aquaculture.2004.12.005 172

Ozsolak, F., Milos, P.M., 2011. RNA sequencing: Advances, challenges and opportunities. Nat. Rev. Genet. doi:10.1038/nrg2934

Palti, Y., Silverstein, J.T., Wieman, H., Phillips, J.G., Barrows, F.T., Parsons, J.E., 2006. Evaluation of family growth response to fishmeal and gluten-based diets in rainbow trout (Oncorhynchus mykiss). Aquaculture 255, 548–556. doi:10.1016/j.aquaculture.2005.11.029

Panserat, S., Hortopan, G.A., Plagnes-Juan, E., Kolditz, C., Lansard, M., Skiba-Cassy, S., Esquerré, D., Geurden, I., Médale, F., Kaushik, S., Corraze, G., 2009. Differential gene expression after total replacement of dietary fish meal and fish oil by plant products in rainbow trout (Oncorhynchus mykiss) liver. Aquaculture 294, 123–131. doi:10.1016/j.aquaculture.2009.05.013

Pastore, M.R., Negrato, E., Poltronieri, C., Barion, G., Messina, M., Tulli, F., Ballarin, C., Maccatrozzo, L., Radaelli, G., Bertotto, D., 2018. Effects of dietary soy isoflavones on estrogenic activity, cortisol level, health and growth in rainbow trout, Oncorhynchus mykiss. Aquac. Res. 49, 1469–1479. doi:10.1111/are.13602

Patel, M.S., Srinivasan, M., 2002. Metabolic programming: Causes and consequences. J. Biol. Chem. 277, 1629–1632. doi:10.1074/jbc.R100017200

Pelissero, C., Le Menn, F., Kaushick, S., 1991. Estrogenic effect of dietary soya bean meal on vitellogenesis in cultured siberian sturgeon Acipenser baeri. Gen. Comp. Endocrinol. 83, 447–457. doi:10.1016/0016-6480(91)90151-U

Perera, E., Yúfera, M., 2017. Effects of soybean meal on digestive enzymes activity, expression of inflammation-related genes, and chromatin modifications in marine fish (Sparus aurata L.) larvae. Fish Physiol. Biochem. 43, 563–578. doi:10.1007/s10695-016-0310-7

Perera, E., Yúfera, M., 2016. Soybean Meal and Soy Protein Concentrate in Early Diet Elicit Different Nutritional Programming Effects on Juvenile Zebrafish. Zebrafish 13, 61–69. doi:10.1089/zeb.2015.1131

Pierron, F., Normandeau, E., Defo, M.A., Campbell, P.G.C., Bernatchez, L., Couture, P., 2011. Effects of chronic metal exposure on wild fish populations revealed by high- throughput cDNA sequencing. Ecotoxicology 20, 1388–1399. doi:10.1007/s10646- 011-0696-z

Pietu, G., Mariage-Samson, R., Fayein, N.-A., Matingou, C., Eveno, E., Houlgatte, R., Decraene, C., Vandenbrouck, Y., Tahi, F., Devignes, M.-D., Wirkner, U., Ansorge, W., Cox, D., Nagase, T., Nomura, N., Auffray, C., 1999. The Genexpress IMAGE 173

Knowledgebase of human brain Transcriptome: a prototype of integrated resource for functional and computational genomics. Genome Res. 9, 195–209.

Pollack, S.J., Ottinger, M.A., Sullivan, C. V, Woods, L.C., 2003. The effects of the soy isoflavone genistein on the reproductive development of striped bass. N. Am. J. Aquac. 65, 226–234. doi:10.1577/C02-041

Post, J.R., Evans, D.O., 1989. Experimental evidence of size-dependent predation mortality in juvenile yellow perch. Can. J. Zool. 67, 521–523. doi:10.1139/z89-076

Pratoomyot, J., Bendiksen, E.Å., Bell, J.G., Tocher, D.R., 2010. Effects of increasing replacement of dietary fishmeal with plant protein sources on growth performance and body lipid composition of Atlantic salmon (Salmo salar L.). Aquaculture 305, 124–132. doi:10.1016/j.aquaculture.2010.04.019

Putney, S.D., Herlihy, W.C., Schimmel, P., 1983. A new troponin T and cDNA clones for 13 different muscle proteins, found by shotgun sequencing. Nature 302, 718–721. doi:10.1038/302718a0

Qian, X., Ba, Y., Zhuang, Q., Zhong, G., 2014. RNA-Seq Technology and Its Application in Fish Transcriptomics. Omi. A J. Integr. Biol. 18, 98–110. doi:10.1089/omi.2013.0110

Ramseyer, L.J., Garling Jr., D.L., 1998. Effects of dietary protein to metabolizable energy ratios and total protein concentrations on the performance of yellow perch Perca flavescens. Aquac. Nutr. 4, 217–223. doi:10.1046/j.1365-2095.1998.00061.x

Ravishankar, D., Rajora, A.K., Greco, F., Osborn, H.M.I., 2016. Flavonoids as prospective compounds for anti - cancer therapy. Int. J. Biochem. Cell Biol. 45, 1–5. doi:10.1016/j.biocel.2013.10.004

Rocha, F., Dias, J., Engrola, S., Gavaia, P., Geurden, I., Dinis, M.T., Panserat, S., 2015. Glucose metabolism and gene expression in juvenile zebrafish (Danio rerio) challenged with a high carbohydrate diet: effects of an acute glucose stimulus during late embryonic life. Br. J. Nutr. 113, 403–413. doi:10.1017/S0007114514003869

Rodríguez-Roque, M.J., Rojas-Graü, M.A., Elez-Martínez, P., Martín-Belloso, O., 2013. Soymilk phenolic compounds, isoflavones and antioxidant activity as affected by in vitro gastrointestinal digestion. Food Chem. 136, 206–212. doi:10.1016/j.foodchem.2012.07.115

Rostagno, M.A.A., Villares, A., Guillamón, E., García-Lafuente, A., Martínez, J.A.A., 2009. Sample preparation for the analysis of isoflavones from soybeans and soy foods, Journal of Chromatography A. doi:10.1016/j.chroma.2008.11.035 174

Rowland, I., Faughnan, M., Hoey, L., Wähälä, K., Williamson, G., Cassidy, A., Rowland, I., 2003. Bioavailability of phyto-oestrogens. Br. J. Nutr. 89, 838–852. doi:10.1079/BJN2002796

Rumsey, G.L., Ketola, H.G., 1975. Amino Acid Supplementation of Casein in Diets of Atlantic Salmon ( Salmo salar ) Fry and of Soybean Meal for Rainbow Trout ( Salmo gairdneri ) Fingerlings. J. Fish. Res. Board Canada 32, 422–426. doi:10.1139/f75-052

Sahlmann, C., Gu, J., Kortner, T.M., Lein, I., Krogdahl, Å., Bakke, A.M., 2015. Ontogeny of the digestive system of Atlantic salmon (Salmo salar L.) and effects of soybean meal from start-feeding. PLoS One. doi:10.1371/journal.pone.0124179

Sahlmann, C., Sutherland, B.J.G., Kortner, T.M., Koop, B.F., Krogdahl, Å., Bakke, A.M., 2013. Early response of gene expression in the distal intestine of Atlantic salmon (Salmo salar L.) during the development of soybean meal induced enteritis. Fish Shellfish Immunol. 34, 599–609. doi:10.1016/j.fsi.2012.11.031

Sanden, M., Berntssen, M.H.G., Krogdahl, Å., Hemre, G.I., Bakke-McKellep, A.M., 2005. An examination of the intestinal tract of Atlantic salmon, Salmo salar L., parr fed different varieties of soy and maize. J. Fish Dis. 28, 317–330. doi:10.1111/j.1365-2761.2005.00618.x

Sanger, F., Nicklen, S., Coulson, A.R., 1977. {DNA} sequencing with chain-terminating inhibitors. Proc Natl Acad Sci U S A 74, 5463–5467.

Schaeffer, T.W., Brown, M.L., Rosentrater, K.A., 2011. Effects of dietary distillers dried grains with solubles and soybean meal on extruded pellet characteristics and growth responses of juvenile yellow perch. N. Am. J. Aquac. 73, 270–278. doi:10.1080/15222055.2011.593461

Schena, M., Shalon, D., Davis, R.W., Brown, P.O., 1995. Quantitative monitoring of gene expression patterns with a complementary DNA microarray. Science (80-. ). 270, 467–470.

Seiliez, I., Vélez, E.J., Lutfi, E., Dias, K., Plagnes-Juan, E., Marandel, L., Panserat, S., Geurden, I., Skiba-Cassy, S., 2017. Eating for two: Consequences of parental methionine nutrition on offspring metabolism in rainbow trout (Oncorhynchus mykiss). Aquaculture 471, 80–91. doi:10.1016/J.AQUACULTURE.2017.01.010

Setchell, K.D.R., Borriello, S.P., Hulme, P., 1984. Nonsteroidal estrogens of dietary origin: Possible roles in hormone-dependent disease. Am. J. Clin. Nutr.

175

Setchell, K.D.R., Brown, N.M., Zimmer-Nechemias, L., Brashear, W.T., Wolfe, B.E., Kirschner, A.S., Heubi, J.E., 2002. Evidence for lack of absorption of soy isoflavone glycosides in humans, supporting the crucial role of intestinal metabolism for bioavailability. Am. J. Clin. Nutr. 76, 447–53.

Shepherd, B.S., Rees, C.B., Sepulveda-Villet, O.J., Palmquist, D.E., Binkowski, F.P., 2013. Identification of Gender in Yellow Perch by External Morphology: Validation in Four Geographic Strains and Effects of Estradiol. N. Am. J. Aquac. 75, 361–372. doi:10.1080/15222055.2013.783520

Siebert, B., Macfarlane, W., 1969. Body water content and water turnover of tropical Bos taurus, Bos indicus, Bibos banteng, and Bos bubalus bubalis. Aust. J. Agric. Res. 20, 613. doi:10.1071/AR9690613

Sim, G.K., Kafatos, F.C., Jones, C.W., Koehler, M.D., 1979. Use of a cDNA Library for Studies on Evolution and Developmental Expression of the Chorion Multigene Families. Cell 8, 1303–316.

Steinke, D., Salzburger, W., Meyer, A., 2006. Novel relationships among ten fish model species revealed based on a phylogenomic analysis using ESTs. J. Mol. Evol. 62, 772–784. doi:10.1007/s00239-005-0170-8

Sullivan, T.J., Stepien, C.A., 2014. Genetic diversity and divergence of yellow perch spawning populations across the Huron-Erie Corridor, from Lake Huron through western Lake Erie. J. Great Lakes Res. 40, 101–109. doi:10.1016/j.jglr.2012.12.004

Sultan, M., Schulz, M.H., Richard, H., Magen, A., Klingenhoff, A., Scherf, M., Seifert, M., Borodina, T., Soldatov, A., Parkhomchuk, D., Schmidt, D., O’Keeffe, S., Haas, S., Vingron, M., Lehrach, H., Yaspo, M.L., 2008. A global view of gene activity and alternative splicing by deep sequencing of the human transcriptome. Science (80-. ). 321, 956–960. doi:10.1126/science.1160342

Sutcliffe, J.G., Milnert, R.J., Bloomt, F.E., Lerner, R.A., 1982. Common 82-nucleotide sequence unique to brain RNA (cDNA/Alu sequences/small cytoplasmic RNA/introns). Biochemistry 79, 4942–4946.

Tacchi, L., Secombes, C.J., Bickerdike, R., Adler, M.A., Venegas, C., Takle, H., Martin, S.A.M., 2012. Transcriptomic and physiological responses to fishmeal substitution with plant proteins in formulated feed in farmed Atlantic salmon (Salmo salar). BMC Genomics 13, 363. doi:10.1186/1471-2164-13-363

Tasnim, M., Ma, S., Yang, E.-W., Jiang, T., Li, W., 2015. Accurate inference of isoforms from multiple sample RNA-Seq data. BMC Genomics 16, S15. doi:10.1186/1471- 2164-16-S2-S15 176

The Ohio Department of Agriculture, 2010. Ohio Aquaculture Industry Analysis.

Torstensen, B.E., Espe, M., Sanden, M., Stubhaug, I., Waagbø, R., Hemre, G.I., Fontanillas, R., Nordgarden, U., Hevrøy, E.M., Olsvik, P., Berntssen, M.H.G., 2008. Novel production of Atlantic salmon (Salmo salar) protein based on combined replacement of fish meal and fish oil with plant meal and vegetable oil blends. Aquaculture 285, 193–200. doi:10.1016/j.aquaculture.2008.08.025

Trapnell, C., Roberts, A., Goff, L., Pertea, G., Kim, D., Kelley, D.R., Pimentel, H., Salzberg, S.L., Rinn, J.L., Pachter, L., 2012. Differential gene and transcript expression analysis of RNA-seq experiments with TopHat and Cufflinks. Nat. Protoc. 7, 562–578. doi:10.1038/nprot.2012.016

Turkmen, S., Zamorano, M.J., Fernández-Palacios, H., Hernández-Cruz, C.M., Montero, D., Robaina, L., Izquierdo, M., 2017. Parental nutritional programming and a reminder during juvenile stage affect growth, lipid metabolism and utilisation in later developmental stages of a marine teleost, the gilthead sea bream (Sparus aurata). Br. J. Nutr. 1–13. doi:10.1017/S0007114517002434

Twibell, R.G., Gannam, A.L., Hyde, N.M., Holmes, J.S.A., Poole, J.B., 2012. Effects of fish meal- and fish oil-free diets on growth responses and fatty acid composition of juvenile coho salmon (Oncorhynchus kisutch). Aquaculture 360–361, 69–77. doi:http://dx.doi.org/10.1016/j.aquaculture.2012.07.019

Twibell, R.G., Watkins, B.A., Brown, P.B., 2001. Dietary Conjugated Linoleic Acids and Lipid Source Alter Fatty Acid Composition of Juvenile Yellow Perch , Perca flavescens 1. Nutr. Metab. 2322–2328.

Twibell, R.G., Wilson, R.P., 2004. Preliminary evidence that cholesterol improves growth and feed intake of soybean meal-based diets in aquaria studies with juvenile channel catfish, Ictalurus punctatus. Aquaculture 236, 539–546. doi:10.1016/j.aquaculture.2003.10.028

Urán, P.A., Gonçalves, A.A., Taverne-Thiele, J.J., Schrama, J.W., Verreth, J.A.J., Rombout, J.H.W.M., 2008. Soybean meal induces intestinal inflammation in common carp (Cyprinus carpio L.). Fish Shellfish Immunol. 25, 751–760. doi:10.1016/j.fsi.2008.02.013

Vacek, J., Klejdus, B., Lojková, L., Kubán, V., 2008. Current trends in isolation, separation, determination and identification of isoflavones: A review. J. Sep. Sci. doi:10.1002/jssc.200700569

Vagner, M., Robin, J.H., Zambonino-Infante, J.L., Tocher, D.R., Person-Le Ruyet, J., 177

2009. Ontogenic effects of early feeding of sea bass (Dicentrarchus labrax) larvae with a range of dietary n-3 highly unsaturated fatty acid levels on the functioning of polyunsaturated fatty acid desaturation pathways. Br. J. Nutr. 101, 1452. doi:10.1017/S0007114508088053

Vagner, M., Zambonino Infante, J.L., Robin, J.H., Person-Le Ruyet, J., 2007. Is it possible to influence European sea bass (Dicentrarchus labrax) juvenile metabolism by a nutritional conditioning during larval stage? Aquaculture 267, 165–174. doi:10.1016/j.aquaculture.2007.01.031 van den Ingh, T.S.G.A.M., Krogdahl, Å., Olli, J.J., Hendriks, H.G.C.J.M., Koninkx, J.G.J.F., 1991. Effects of soybean containing diets on the proximal and distal intestine in Atlantic salmon (Salmo salar) - A morphological study. Aquac. 94(4) 297-305. 94, 297–305.

Vanhees, K., Vonhögen, I.G.C., Van Schooten, F.J., Godschalk, R.W.L., 2014. You are what you eat, and so are your children: The impact of micronutrients on the epigenetic programming of offspring. Cell. Mol. Life Sci. 71, 271–285. doi:10.1007/s00018-013-1427-9

Velculescu, V.E., Zhang, L., Vogelstein, B., Kinzler, K.W., 1995. Serial Analysis of Gene Expression. Science (80). 270, 484–487. doi:10.1126/science.270.5235.484

Velculescu, V.E., Zhang, L., Zhou, W., Vogelstein, J., Basrai, M.A., Bassett, D.E., Hieter, P., Vogelstein, B., Kinzler, K.W., 1997. Characterization of the yeast transcriptome. Cell 88, 243–251. doi:10.1016/S0092-8674(00)81845-0

Venold, F.F., Penn, M.H., Krogdahl, Å., Overturf, K., 2012. Severity of soybean meal induced distal intestinal inflammation, enterocyte proliferation rate, and fatty acid binding protein (Fabp2) level differ between strains of rainbow trout (Oncorhynchus mykiss). Aquaculture 364–365, 281–292. doi:10.1016/j.aquaculture.2012.08.035

Vera, L.M., Metochis, C., Taylor, J.F., Clarkson, M., Skjærven, K.H., Migaud, H., Tocher, D.R., 2017. Early nutritional programming affects liver transcriptome in diploid and triploid Atlantic salmon, Salmo salar. BMC Genomics 18, 886. doi:10.1186/s12864-017-4264-7

Volff, J.-N., 2005. Genome evolution and biodiversity in teleost fish. Heredity (Edinb). 94, 280–294. doi:10.1038/sj.hdy.6800635

Voorhees, D. Van, Lowther, A., Liddel, M., 2016. Fisheries of the United States-2015.

Wallich, C., 1901. Method of Recording Egg Development, for use of Fish Culturists. Rep. Comm. Fish Fish. 185–195. 178

Wang, N., Teletchea, F., Kestemont, P., Milla, S., Fontaine, P., 2010. Photothermal control of the reproductive cycle in temperate fishes. Rev. Aquac. 2, 209–222. doi:10.1111/j.1753-5131.2010.01037.x

Wang, Q., Ge, X., Tian, X., Zhang, Y., Zhang, J., Zhang, P., 2013. Soy isoflavone: The multipurpose phytochemical (Review). Biomed. Reports 1, 697–701. doi:10.3892/br.2013.129

Wang, Z., Gerstein, M., Snyder, M., 2009. RNA-seq: A revolutionary tool for transcriptomics. Nat. Rev. Genet. 10, 57–63.

Weeks, C., Lynch, W., Morris, J., 2014. 2014 NCR Aquaculture Needs Survey.

Wei, H., Bowen, R., Cai, Q., Barnes, S., Wang, Y., 1995. Antioxidant and Antipromotional Effects of the Soybean Isoflavone Genistein. Exp. Biol. Med. 208, 124–130. doi:10.3181/00379727-208-43844

Wilhelm, B.T., Marguerat, S., Watt, S., Schubert, F., Wood, V., Goodhead, I., Penkett, C.J., Rogers, J., Bähler, J., 2008. Dynamic repertoire of a eukaryotic transcriptome surveyed at single-nucleotide resolution. Nature 453, 1239–1243. doi:10.1038/nature07002

Xu, Z., Wu, Q., Godber, J.S., 2002. Stabilities of daidzin, glycitin, genistin, and generation of derivatives during heating. J. Agric. Food Chem. 50, 7402–7406. doi:10.1021/jf025626i

Yu, J., Bi, X., Yu, B., Chen, D., 2016. Isoflavones: Anti-inflammatory benefit and possible caveats. Nutrients 8, 1–16. doi:10.3390/nu8060361

Yun, B., Ai, Q., Mai, K., Xu, W., Qi, G., Luo, Y., 2012. Synergistic effects of dietary cholesterol and taurine on growth performance and cholesterol metabolism in juvenile turbot (Scophthalmus maximus L.) fed high plant protein diets. Aquaculture 324–325, 85–91. doi:10.1016/j.aquaculture.2011.10.012 Yun, B., Mai, K., Zhang, W., Xu, W., 2011. Effects of dietary cholesterol on growth performance, feed intake and cholesterol metabolism in juvenile turbot (Scophthalmus maximus L.) fed high plant protein diets. Aquaculture 319, 105–110. doi:10.1016/j.aquaculture.2011.06.028

Zerbino, D.R., Birney, E., 2008. Velvet: Algorithms for de novo short read assembly using de Bruijn graphs. Genome Res. 18, 821–829. doi:10.1101/gr.074492.107

Zhou, S., Ackman, R.G., Morrison, C., 1996. Adipocytes and lipid distribution in the muscle tissue of Atlantic salmon (Salmo salar). Can. J. Fish. Aquat. Sci. 53, 326– 179

332. doi:10.1139/cjfas-53-2-326

Zhou, Z., Ringø, E., Olsen, R.E., Song, S.K., 2017. Dietary effects of soybean products on gut microbiota and immunity of aquatic animals: A review. Aquac. Nutr. 644– 665. doi:10.1111/anu.12532

Zhu, T., Corraze, G., Plagnes-Juan, E., Quillet, E., Dupont-Nivet, M., Skiba-Cassy, S., 2018. Regulation of genes related to cholesterol metabolism in rainbow trout ( Oncorhynchus mykiss ) fed a plant-based diet. Am. J. Physiol. Integr. Comp. Physiol. 314, R58–R70. doi:10.1152/ajpregu.00179.2017

Zuniga, K.E., Clinton, S.K., Erdman, J.W., 2013. The interactions of dietary tomato powder and soy germ on prostate carcinogenesis in the TRAMP model. Cancer Prev. Res. 6, 548–557. doi:10.1158/1940-6207.CAPR-12-0443

180