INTRASPECIFIC FLAVOBACTERIUM PSYCHROPHILUM DIVERSITY AS A FACTOR IN BACTERIAL COLDWATER DISEASE ECOLOGY AND MANAGEMENT By Christopher Kay Knupp A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Fisheries and Wildlife – Doctor of Philosophy 2023 ABSTRACT Flavobacterium psychrophilum, causative agent of bacterial coldwater disease (BCWD) and rainbow trout fry syndrome (RTFS), causes substantial economic losses worldwide, particularly in salmonid species (Family Salmonidae) such as rainbow trout (Oncorhynchus mykiss), coho salmon (O. kisutch), and Atlantic salmon (Salmo salar). Current challenges in managing, preventing, and controlling BCWD outbreaks may partially relate to the considerable intraspecific diversity present within this species, including that revealed via multilocus sequence typing (MLST). Indeed, MLST-based epidemiological studies and others suggest F. psychrophilum diversity may influence the ecology and behavior of some variants, affect their detection and diagnosis, and play a role in host specificity, transmission, and environmental persistence. However, controlled studies examining these aspects of BCWD ecology in relation to F. psychrophilum genetic diversity are lacking. My dissertation addressed these critical knowledge gaps, with the goal of contributing to enhanced strategies for the management, prevention, and control of BCWD. To improve F. psychrophilum recovery and detection, I initially compared colony yields of geographically, temporally, and genetically diverse F. psychrophilum isolates on three previously established F. psychrophilum culture media. The selected media included the current gold-standard medium, tryptone yeast extract salts (TYES) agar, which yielded the most colonies. With TYES as a foundation, I employed a Plackett- Burman experimental design, culminating in the development of two new culture media (F. psychrophilum medium-A and -B). These optimized media significantly improved F. psychrophilum recovery in the laboratory and from naturally infected salmonids when compared to TYES, and thus will enhance BCWD research and diagnostic efforts. To assess F. psychrophilum host specificity, variants belonging to MLST clonal complexes most associated with Atlantic salmon, coho salmon, or rainbow trout were cross challenged against each of these salmonid species via immersion. Resultingly, some variants were host specific, as evidenced by only causing disease and mortality in one species, whereas others caused disease and mortality in all three species, although to varying degrees. Variation in molecular serotype and proteolytic activity were also observed among variants. Collectively, findings highlighted the complexities of host-pathogen interactions and may guide the development of BCWD prevent strategies, such as vaccines. In a separate experiment, the shedding dynamics of live and dead Atlantic salmon, coho salmon, and rainbow trout were assessed, marking the first study to evaluate F. psychrophilum shedding dynamics in Atlantic salmon and coho salmon. Although both live and dead fish of all species shed F. psychrophilum, dead fish shed substantially more bacterial cells and for a longer duration. Furthermore, shedding dynamics varied by F. psychrophilum variant and/or host species, a matter that may complicate BCWD management. The persistence of predominating F. psychrophilum variants in microcosms composed of sterile well water only, sterile well water with commercial trout feed, and sterile well water with raceway detritus was measured via culture over 13 weeks. All variants remained culturable in each microcosm for at least eight weeks, with bacterial concentrations significantly higher in the presence of raceway detritus. However, significant differences in culturability were observed within and between microcosms, suggesting potential variability in environmental persistence strategies among specific variants. In total, the findings of my dissertation supported my overarching hypothesis that F. psychrophilum intraspecific diversity plays an important role in shaping our understanding of BCWD ecology. Copyright by CHRISTOPHER KAY KNUPP 2023 ACKNOWLEDGMENTS I would like to express my gratitude to my advisor, Dr. Thomas Loch, for his unwavering support and guidance throughout my graduate career. He has provided me with many research and professional development opportunities, all of which have helped me to become a well- rounded scientist. I would also like to thank my committee members, Dr. Travis Brenden, Dr. Matti Kiupel, and Dr. Jean Tsao for their guidance and assistance over the past five years. I would also like to sincerely thank the United States Department of Agriculture – National Institute of Food and Agriculture for funding most my dissertation research (Grant number: 2019-70007-30417). I would also like to express my gratitude to Schrems West Michigan Trout Unlimited and Hal and Jean Glassen for recognizing the importance of my dissertation research by awarding me with generous fellowships. I would also like to thank Dr. Douglas Call, Dr. Esteban Soto, Dr. Jesse Trushenski, and many Michigan Department of Natural Resources Fisheries Division personnel for their collaboration on this research. v TABLE OF CONTENTS LIST OF TABLES ....................................................................................................................... viii LIST OF FIGURES ....................................................................................................................... xi LIST OF ABBREVIATIONS ...................................................................................................... xiv INTRODUCTION ...........................................................................................................................1 0.1. Study objectives ..................................................................................................................3 REFERENCES ..........................................................................................................................6 Chapter 1: Literature Review ...........................................................................................................8 1.1. Bacterial coldwater disease (BCWD) and rainbow trout fry syndrome (RTFS) ................9 1.2. Flavobacterium psychrophilum, causative agent of BCWD and RTFS ...........................14 REFERENCES ........................................................................................................................36 APPENDIX ..............................................................................................................................49 Chapter 2: Enhanced culture media for the recovery of Flavobacterium psychrophilum, causative agent of bacterial coldwater disease and rainbow trout fry syndrome.........................52 2.1. Abstract .............................................................................................................................53 2.2. Introduction .......................................................................................................................54 2.3. Materials and Methods......................................................................................................57 2.4. Results ...............................................................................................................................65 2.5. Discussion .........................................................................................................................68 REFERENCES ........................................................................................................................74 APPENDIX ..............................................................................................................................80 Chapter 3: Immersion challenge of three salmonid species (Family Salmonidae) with three heterologous Flavobacterium psychrophilum multilocus sequence typing variants provides further evidence of differential host specificity ..........................................107 3.1. Abstract ...........................................................................................................................108 3.2. Introduction .....................................................................................................................109 3.3. Materials and Methods ....................................................................................................111 3.4. Results .............................................................................................................................116 3.5. Discussion .......................................................................................................................122 REFERENCES ......................................................................................................................128 APPENDIX ............................................................................................................................133 Chapter 4: Varying Flavobacterium psychrophilum shedding dynamics in three bacterial coldwater disease-susceptible salmonid (Family Salmonidae) species .....................141 4.1. Abstract ...........................................................................................................................142 4.2. Introduction .....................................................................................................................143 4.3. Materials and Methods ....................................................................................................145 4.4. Results .............................................................................................................................153 4.5. Discussion .......................................................................................................................158 vi REFERENCES ......................................................................................................................165 APPENDIX ............................................................................................................................171 Chapter 5: Culturability of heterologous Flavobacterium psychrophilum multilocus sequence typing variants in three microcosms that simulate common fish farm and hatchery environments ..............................................................................................................177 5.1. Abstract ...........................................................................................................................178 5.2. Introduction .....................................................................................................................179 5.3. Materials and Methods ....................................................................................................181 5.4. Results .............................................................................................................................184 5.5. Discussion .......................................................................................................................188 REFERENCES ......................................................................................................................195 APPENDIX ............................................................................................................................199 Chapter 6: Conclusions and future research ................................................................................209 6.1. Conclusions .....................................................................................................................210 6.2. Future research ................................................................................................................214 vii LIST OF TABLES Table 1.1. Formulation (per 1L H2O) of Flavobacterium psychrophilum recovery medium Anacker and Ordal (AO; Anacker and Ordal, 1959) and its derivatives, as reported in the literature. ........................................................................................................................................49 Table 1.2. Formulation (per 1L H2O) of Flavobacterium psychrophilum recovery medium tryptone yeast extract salts (TYES; Holt, 1987) agar and its derivatives, as reported in the literature. ........................................................................................................................................50 Table 1.3. Formulation (per 1L H2O) of less frequently used (i.e., reported) Flavobacterium psychrophilum recovery media. .....................................................................................................51 Table 2.1. Metadata for the 165 Flavobacterium psychrophilum isolates used in this study, including multilocus sequence typing sequence type (ST), clonal complex (CC), isolation location, and host of origin. Information is presented in order by CC then ST. ...........................80 Table 2.2. Additional meta data of the 165 Flavobacterium psychrophilum isolates used in this study, including isolate identifier (ID), year of isolation, location of isolation, host of origin, tissue of origin, life history (e.g., wild/feral or captive), life stage, and multilocus sequencing typing sequence type (ST) and clonal complex (CC). All 165 F. psychrophilum isolates were used to select the best basal medium (e.g., tryptone yeast extract agar; TYES) and in the comparison of the new media (F. psychrophilum medium-A and -B) to TYES. Fifty F. psychrophilum isolates were used in the Plackett-Burman experiment (denoted here with h). The information is presented in order by multi-locus sequence typing (MLST) clonal complex (CC), then MLST sequence type (ST). ....................................................................................................83 Table 2.3. Formulations (per 1L H2O) of media used in basal medium experiment, including tryptone yeast extract salts agar (TYES; Holt 1987), Oplinger and Wagner medium (OW; Oplinger and Wagner, 2012), enriched Anacker and Ordal medium supplemented with activated charcoal and aromatic compounds (EAOCa; Alvarez and Guijarro, 2007), and neutral medium (NM; this study). ............................................................................................................................88 Table 2.4. Medium components tested as part of Plackett-Burman experimental design, including tested quantities per 1L H2O. .........................................................................................89 Table 2.5. Plackett-Burman experimental design matrix, consisting of 11 tested medium components (e.g., X1-X11) resulting in 12 media formulations. Medium components incorporated at their low or high concentration are denoted with -1 and +1, respectively............90 Table 2.6. Differences of least square mean estimates ± standard error (SE) for the interaction between medium (e.g., EAOCa, OW, and TYES) and 165 Flavobacterium psychrophilum isolates. Pairwise comparisons between media for each isolate are provided. Tukey-Kramer adjusted P-values for multiple comparisons are shown (α = 0.05). Table is ordered by isolate. ............................................................................................................................................91 viii Table 2.7. Medium component effect ± standard error (SE) on Flavobacterium psychrophilum recovery resulting from Plackett-Burman experiment as well as t-value and p-value. The symbol, *, indicates a medium component had a significant effect on F. psychrophilum recovery (a = 0.05). ....................................................................................................................................102 Table 3.1. Flavobacterium psychrophilum isolates used in this study for in vivo challenge experiments against Atlantic salmon (Salmo salar), coho salmon (Oncorhynchus kisutch), and rainbow trout (O. mykiss) and for the assessment of proteolytic activity, which is presented as a ratio of the median clear zone diameter to the colony diameter (in mm) ± standard error (SE). Median clear zone ratios for a particular enzyme (e.g., caseinase, gelatinase, or elastinase) containing an identical symbol (e.g., *, **, ***) are not significantly different (a = 0.05) and a ratio of 1.00 ± 0.00 indicates no protease activity. ......................................................................133 Table 3.2. Proportion of dead Atlantic salmon (Salmo salar), coho salmon (Oncorhynchus kisutch), and rainbow trout (O. mykiss) with a range of gross external and internal bacterial coldwater disease signs following exposure to Flavobacterium psychrophilum isolates US19- COS, US62-ATS, and US87-RBT. ..............................................................................................134 Table 4.1. Repeatability of Flavobacterium psychrophilum Marancik and Wiens (2013) qPCR assay. ............................................................................................................................................171 Table 4.2. Evaluation of Flavobacterium psychrophilum DNA extraction procedure from water using quantitative PCR (qPCR). Key to abbreviations: SD, standard deviation; CV, coefficient of variation; IPC, internal positive control; Cq, cycle threshold. .....................................................172 Table 4.3. Mean Flavobacterium psychrophilum shedding rates (cells/fish/hour) ± standard deviation of dead Atlantic salmon, coho salmon, and rainbow trout on each sampling day. Number of fish sampled (N) is to the left of the host species name. ...........................................173 Table 5.1. Flavobacterium psychrophilum isolates selected for this study. The multilocus sequence typing clonal complex (CC), sequence type (ST; i.e., genetic variant), recovery location, host of origin, year of isolation, and starting concentration (i.e., at week 0) in the microcosm experiment are presented for each isolate. ................................................................199 Table 5.2. Weekly mean concentrations (in log10 colony forming units) ± (standard error) of the ten Flavobacterium psychrophilum sequence types (i.e., genetic variants) used in this study, as measured in the well water only microcosm................................................................................200 Table 5.3. Weekly mean concentrations (in log10 colony forming units) ± (standard error) of the three experimental microcosms (e.g., well water, well water with raceway detritus or commercial trout feed). ....................................................................................................................................201 Table 5.4. Differences of least square mean estimates ± standard error (SE) for the interaction between treatment group (e.g., raceway “detritus”, commercial trout “feed”, or “water” only) and ten Flavobacterium psychrophilum variants (e.g., ST13, ST78, ST277, ST10, ST286, ST275, ST353, ST253, ST342, ST267). Pairwise comparisons between variants within the same ix treatment are provided. Tukey-Kramer adjusted P-values for multiple comparisons are shown (α = 0.05). Table is ordered by treatment. ........................................................................................202 Table 5.5. Weekly mean concentrations (in log10 colony forming units) ± (standard error) of the ten Flavobacterium psychrophilum sequence types (i.e., genetic variants) used in this study, as measured in the well water with raceway detritus microcosm. ...................................................205 Table 5.6. Weekly mean concentrations (in log10 colony forming units) ± (standard error) of the ten Flavobacterium psychrophilum sequence types (i.e., genetic variants) used in this study, as measured in the well water with commercial trout feed microcosm. ..........................................206 Table 5.7. Differences of least square mean estimates ± standard error (SE) for the interaction between treatment group (e.g., raceway “detritus”, commercial trout “feed”, or “water” only) and ten Flavobacterium psychrophilum variants (e.g., ST13, ST78, ST277, ST10, ST286, ST275, ST353, ST253, ST342, ST267). Pairwise comparisons between treatments for each variant are provided. Tukey-Kramer adjusted P-values for multiple comparisons are shown (α = 0.05). ....207 x LIST OF FIGURES Figure 2.1. Map of Michigan with Flavobacterium psychrophilum surveillance locations and salmonid species sampled (ATS = Atlantic salmon, Salmo salar; CHS = Chinook salmon, Oncorhynchus tshawytscha; COS = Coho salmon, O. kisutch; STT = Steelhead trout, O. mykiss).....................................................................................................................................103 Figure 2.2. Normal probability plot of the effect of different factors (i.e., medium components) on Flavobacterium psychrophilum growth (measured in colony forming units) in the Plackett- Burman design. Factors with a square symbol significantly affected F. psychrophilum recovery, whereas those with a circular symbol did not (a = 0.05). Factors to the left of the line improved F. psychrophilum recovery when incorporated into media at their low concentration, whereas factors to the right of the line improved recovery when incorporated at their high concentration. ...............................................................................................................................104 Figure 2.3. Pareto chart of the effect of different factors (i.e., medium components) on Flavobacterium psychrophilum growth (measured in colony forming units) in the Plackett- Burman design. The horizontal bars represent the ratio between the effects of the medium components and their standard error. The bars are ordered according to effect size, with the greatest effect on F. psychrophilum growth on top. The vertical dashed line illustrates the critical t value (a = 0.05), meaning medium components crossing the line had a significant effect on F. psychrophilum growth. ............................................................................................................105 Figure 2.4. Flavobacterium psychrophilum infection prevalence among spawning phase salmonid (CHS, Chinook salmon, Oncorhynchus tshawytscha; ATS, Atlantic salmon, Salmo salar; COS, Coho salmon, O. kisutch; STT, Steelhead trout, O. mykiss) broodstock sampled at four Michigan gamete collection locations (Little Manistee River Weir, LMRW; St. Mary’s River, SMR; Platte River Weir, PRW; Swan River Weir, SRW) according to three F. psychrophilum media (F. psychrophilum medium-A, FPM-A; F. psychrophilum medium-B, FPM-B; tryptone yeast extract salts medium, TYES). ................................................................106 Figure 3.1. Gross external lesions in rainbow trout (O. mykiss) following immersion challenge with F. psychrophilum isolate US87-RBT. (A) Focally extensive ulceration of the caudal peduncle that penetrated the underlying musculature. (B) Focally extensive ulceration of the caudal peduncle that penetrated the underlying musculature and exposed the vertebral column. (C) Bilateral exophthalmia. (D) Unilateral exophthalmia with intraocular ecchymosis (arrow). (E) Gill pallor with ecchymoses and petechiae (arrow)...............................................................135 Figure 3.2. Gross internal lesions in rainbow trout (O. mykiss) following immersion challenge with F. psychrophilum isolate US87-RBT. (A) Pale liver with multifocal ecchymoses. (B) Severe splenic swelling with perisplenic hemorrhage. (C) Severe intestinal hemorrhage with accompanying peri-intestinal hemorrhage. ..................................................................................136 Figure 3.3. Kaplan-Meier survival probability curves of Atlantic salmon (Salmo salar; ATS), coho salmon (Oncorhynchus kisutch; COS), and rainbow trout (O. mykiss; RBT) over a 25-day period following immersion challenge with Flavobacterium psychrophilum isolates (A) US87- xi RBT, (B) US19-COS, and (C) US62-ATS. Shaded regions depict 95% confidence intervals. Lines with different symbols (e.g., *, **) indicate significant differences in survival (α = 0.05). ....................................................................................................................................137 Figure 3.4. Gross external lesions in Atlantic salmon (Salmo salar), coho salmon (Oncorhynchus kisutch), and rainbow trout (O. mykiss) following immersion challenge with F. psychrophilum isolate US19-COS. (A) Atlantic salmon with shallow focally extensive dermal ulceration of the caudal peduncle. (B) Coho salmon with focally extensive dermal ulceration of the caudal peduncle that penetrated the underlying musculature. (C) Rainbow trout with focally extensive dermal ulceration of the caudal peduncle that penetrated the underlying musculature. Diffuse ecchymoses and petechiae surround the ulcer. (D) Atlantic salmon with focally extensive dermal ulceration of the caudal peduncle, exposing the vertebral column. (E) and (F) Peri-oral ulceration in coho salmon (E) and rainbow trout (F). ..................................................................138 Figure 3.5. Gross internal lesions in Atlantic salmon (Salmo salar), coho salmon (Oncorhynchus kisutch), and rainbow trout (O. mykiss) following immersion challenge with F. psychrophilum isolate US19-COS. (A) Atlantic salmon with liver pallor. (B) Coho salmon with pale liver and multifocal ecchymoses, splenic swelling, and intestinal hemorrhage. (C) Rainbow trout with liver pallor and splenic swelling. (D) Coho salmon with renal pallor. (E) Rainbow trout with renal pallor and diffuse ecchymoses. (F) Rainbow trout with intestinal hemorrhage. .........................139 Figure 3.6. Gross external lesions in Atlantic salmon (Salmo salar), coho salmon (Oncorhynchus kisutch), and rainbow trout (O. mykiss) following immersion challenge with F. psychrophilum isolate US62-ATS. (A) Atlantic salmon with focally extensive dermal ulceration of the caudal peduncle that penetrated the underlying musculature. (B) Rainbow trout with multiple, focally extensive dermal ulcerations of the caudal peduncle that penetrated the underlying musculature. Ulceration is surrounded by severe diffuse ecchymoses that extends posteriorly into the caudal fin. (C) Coho salmon with focally extensive ulceration of the caudal peduncle that penetrated the underlying musculature. (D) Atlantic salmon with focal ecchymosis of the caudal fin. (E) Coho salmon with severe peri-oral ulceration. ..........................................140 Figure 4.1. Mean Flavobacterium psychrophilum shedding rates (black bars with standard deviation) from and cumulative percent mortality (gray lines with standard error) of live A) Atlantic salmon (Salmo salar), B) rainbow trout (Oncorhynchus mykiss), and C) coho salmon (O. kisutch). *, water sampling occasion, which continued four weeks past the last detection of F. psychrophilum. //, indicates 6-day gap in time. ...........................................................................174 Figure 4.2. Flavobacterium psychrophilum shedding rates of dead individual (A) Atlantic salmon (Salmo salar), (B) rainbow trout (Oncorhynchus mykiss), and (C) coho salmon (O. kisutch). Legend explanation: ≤ 2 fish (F) per replicate (R) aquarium were maintained. Therefore, in the legend and as an example, R1.F1 corresponds to replicate one, fish one. .......175 Figure 4.3. Representative images of dead fish used in Flavobacterium psychrophilum shedding experiment. (A) Coho salmon (Oncorhynchus kisutch) 12 days post death. (B) Rainbow trout (O. mykiss) 25 days post death. (C) Coho salmon 98 days post death. Note the yellowish discoloration present on fish. .......................................................................................................176 xii Figure 5.1. 13-week culturability of Flavobacterium psychrophilum sequence type (ST) 10, ST13, ST78, ST253, ST267, ST275, ST277, ST286, ST342, and ST353 in microcosms containing (A) sterilized well water only, (B) sterilized well water with raceway detritus, or (C) sterilized well water with commercial trout feed. Standard error between replicate flasks on each sampling week (e.g., 0 – 8 and 13) are shown. ............................................................................208 xiii LIST OF ABBREVIATIONS AFS-FHS American Fisheries Society – Fish Health Section AK Alaska AO Anacker and Ordal medium ATCC American Type Culture Collection atpA Adenosine triphosphate synthetase, α subunit ATS Atlantic salmon AUF Animal Use Form BC British Columbia BCWD Bacterial Coldwater Disease BKT Brook trout BNT Brown trout bp Base pair C Celsius CA California CaCl2 Calcium chloride CACM Anacker and Ordal medium with carbohydrates and skimmed milk CC Clonal complex cfu Colony forming unit CHS Chinook salmon CO Colorado COS Coho salmon Cq Quantification cycle xiv CUT Cutthroat trout CV Coefficient of variation CZR Clear zone ratio DECF DNA extraction correction factor df Degrees of freedom DNA Deoxyribose nucleic acid dnak Chaperone heat-shock protein 70 EAO Enriched Anacker and Ordal medium EAOC Enriched Anacker and Ordal medium with charcoal EAOCa Enriched Anacker and Ordal medium with charcoal and aromatic compounds EAOS Enriched Anacker and Ordal medium with horse serum FBS Fetal bovine serum FCS Fetal calf serum FLPA Flavobacterium psychrophilum agar FPM-A Flavobacterium psychrophilum medium-A FPM-B Flavobacterium psychrophilum medium-B fumC Fumarate hydratase class II g Gram g Gravity G+C Guanine plus Cytosine gDNA Genomic DNA gyrb Topoisomerase II, β subunit H2 O Water xv HS Horse serum ID Idaho IPC Internal positive control L Liter LAT Lake trout LMRW Little Manistee River weir LOD Limit of detection LOQ Limit of quantification LWF Lake whitefish mAO Modified Anacker and Ordal medium MAOG Modified Anacker and Ordal medium with glucose MAT TYES with 1% maltose MD Maryland mg Milligram MgSO4 Magnesium sulfate MI Michigan ml Milliliter MLST Multilocus sequence typing mm Millimeter MS-222 Methanesulfonate MSU Michigan State University MT Montana mTYES Modified TYES xvi murg Glycosyltransferase murein G n Number N Number NA Nutrient agar NAC Nutrient agar with charcoal NC North Carolina NCIMB National Collection of Industrial Food and Marine Bacteria ng Nanogram NJ New Jersey NM Neutral medium OD Optical density OR Oregon OW Oplinger and Wagner medium PA Pennsylvania PCR Polymerase chain reaction pH Potential of hydrogen ppm Parts per million PRW Platte River weir qPCR Quantitative PCR R2 Coefficient of determination RBT Rainbow trout rpm Revolutions per minute RTFS Rainbow Trout Fry Syndrome xvii SD Standard deviation SE Standard error SMR St. Mary’s River SOC Sockeye salmon SPL Splake SRW Swan River weir ST Sequence type STT Steelhead trout tRNA Transfer ribonucleic acid trpB Tryptophane synthetase β subunit tuf Elongation factor Tu TX Texas TYES Tryptone yeast extract salts agar U.S. United States µl Microliter µm Micrometer µM Micromolar USA United States of America UT Utah VA Virginia w/v Weight per volume WA Washington WHT White sturgeon xviii WI Wisconsin WV West Virginia WY Wyoming xix INTRODUCTION 1 Fish currently comprise one of the most important animal protein sources for the ever- expanding human population, much of which is supplied through aquaculture (FAO, 2022). Concurrently, many fish species are the focus of substantial conservation and stock enhancement efforts by fishery resource agencies throughout the world. A significant threat to the productivity and health of hatchery and farm-reared salmon and trout (Family Salmonidae) is bacterial coldwater disease (BCWD) and rainbow trout fry syndrome (RTFS), caused by Flavobacterium psychrophilum (Holt, 1987; Loch and Faisal, 2017). This bacterial fish pathogen has been recovered in >20 countries distributed over five continents (e.g., Asia, Australia, Europe, North America, and South America; Starliper, 2011) and continues to devastate both wild and farmed fish species, especially salmonids, by causing up to 90% mortality in affected populations (Barnes and Brown, 2011). Despite nearly a century of research and various disease management efforts having been undertaken, the efficacy of currently available BCWD prevention and control methods remains inadequate. In this context, recent studies have identified some predominant F. psychrophilum genetic variants responsible for BCWD epizootics in the USA and abroad, and made significant advancements towards untangling how such intraspecific diversity relates to bacterial virulence (Sundell et al. 2019; Knupp et al. 2021; Li et al. 2021), mounting antimicrobial resistance (Van Vliet et al. 2017; Li et al. 2021), host species predilections (Knupp et al. 2021), and geographical distribution (Nicolas et al. 2008; Van Vliet et al. 2016; Knupp et al. 2019; Sebastião et al. 2020; Harrison et al. 2021). Notably, these and other studies have highlighted a range of knowledge gaps and raised questions regarding why some of the most damaging USA F. psychrophilum variants are so successful through the lens of disease ecology. For example, do some F. psychrophilum variants truly have an affinity for a particular host species and if so, what 2 mechanism(s) may contribute to these associations? Have artificial rearing environments contributed to the persistence of the predominating and seemingly highly successful F. psychrophilum variants? Do shedding dynamics (e.g., F. psychrophilum shedding loads and duration) vary among F. psychrophilum variants and/or salmonid species? Are current F. psychrophilum culture media, including the gold-standard medium, tryptone yeast extract salts (TYES), sufficient for BCWD diagnosis and research purposes? At present, these gaps in knowledge contribute to productivity losses for salmonid farms and impede hatchery-based conservation of salmonid populations. To fill in these knowledge gaps and improve BCWD diagnosis, management, prevention, and control, I reviewed a range of F. psychrophilum literature, and then conceptualized, planned, and conducted a series of experiments that collectively comprise my dissertation research. 0.1. Study objectives The overarching goal of my dissertation research is to elucidate how F. psychrophilum intraspecific diversity may affect: a) detection in fish tissues and diagnose BCWD, b) BCWD research findings, c) management of BCWD epizootics, and d) development of effective BCWD prevention and control strategies. To accomplish this goal, I reviewed the literature on F. psychrophilum and then performed a series of in vitro and in vivo experiments as described below. In Chapter 1, I reviewed the literature on F. psychrophilum and bacterial coldwater disease/RTFS, focusing on: a) the use and development of culture media for F. psychrophilum isolation and propagation in vitro; b) F. psychrophilum-host associations and contributing mechanisms; c) F. psychrophilum shedding dynamics, with a primary focus on horizontal 3 transmission; and d) the persistence of F. psychrophilum outside of its host in various aquatic environments. In Chapter 2, I compared the recovery of 165 genetically diverse F. psychrophilum isolates on three previously published F. psychrophilum culture media and selected the medium (e.g., TYES) that achieved the best overall recovery for subsequent comparisons. Next, I used a Plackett-Burman experimental design to quantify the effect 11 medium components had on F. psychrophilum recovery to subsequently develop two new culture media (e.g., F. psychrophilum medium A and -B; FPM-A and FPM-B). I then compared F. psychrophilum recovery on FPM-A and FPM-B to TYES under laboratory conditions, finding both new media significantly improved F. psychrophilum recovery by >141%, although FPM-A performed slightly better than FPM-B. Likewise, I compared all three media under field conditions while performing routine health surveillance on wild/feral Michigan salmonid broodstock populations and again found the new media outperformed TYES for F. psychrophilum primary isolation. In Chapter 3, I examined the hypothesis that some F. psychrophilum variants had host- associations following close examination of multilocus sequence typing metadata and the completion of a preliminary F. psychrophilum challenge study using coho salmon (Oncorhynchus kisutch). I tested this hypothesis by exposing Atlantic salmon (Salmo salar), coho salmon, and rainbow trout (O. mykiss) under controlled laboratory conditions to three distinct F. psychrophilum variants, each with putative associations to one of the salmonid species. Study results demonstrated some F. psychrophilum variants have strong host associations, whereas others appear to have a wider host range. In Chapter 4, I assessed whether horizontal transmission strategies differed amongst F. psychrophilum variants and/or salmonid species. To do so, I intramuscularly injected rainbow 4 trout, coho salmon, and Atlantic salmon with one putatively host specific variant per species, and then quantified (via quantitative PCR) F. psychrophilum loads that were shed into the water by live and dead fish up to 98 days post-exposure. Findings collectively showed that dead fish were more efficient F. psychrophilum shedders than live fish (i.e., F. psychrophilum was shed at higher loads for a longer duration), but that other aspects of shedding dynamics may vary depending on F. psychrophilum variant and/or affected fish species. In Chapter 5, I compared the survival of 10 distinct F. psychrophilum genetic variants, including some that are most widespread in the USA, in three microcosms that mimic environments common to aquaculture and hatchery facilities, including one with water only, a second with water and trout feed, and the third with water and raceway detritus (i.e., uneaten food, fish byproducts). Overall, all tested isolates persisted in each environment for at least 8 weeks, but survival was best in water with detritus. Notably, some isolates did survive better in water only compared to water with feed, highlighting potential differences in environmental persistence strategies by F. psychrophilum variant. In Chapter 6, I discussed how the new knowledge described herein has not only led to improved detection of F. psychrophilum and diagnosis of BCWD across the USA, but also has informed best management practices and the future development of BCWD prevention and control strategies. Ultimately, this research will minimize losses, and reduce the risk of BCWD epizootics in salmonid farms and hatcheries across the USA. 5 REFERENCES Barnes, M.E., Brown, M.L. 2011. A review of Flavobacterium psychrophilum biology, clinical signs, and bacterial coldwater disease prevention and treatment. The Open Fish Science Journal. 4, 40-48. Food and Agriculture Organization of the United Nations. 2022. State of the world fisheries and aquaculture 2022. Towards Blue Transformation. Rome, FAO. Harrison, C.E., Knupp, C.K., Brenden, T.O., Ebener, M.P., Loch, T.P. 2022. First isolation of Flavobacterium psychrophilum from wild adult Great lakes lake whitefish (Coregonus clupeaformisi). Journal of Fish Diseases. 45(7), 1023-1032. Holt, R.A. 1987. PhD thesis. Cytophaga psychrophila, the causative agent of bacterial cold water disease in salmonid fish. Oregon State University, Corvallis, OR. Knupp, C., Wiens, G.D., Faisal, M., Call, D.R., Cain, K.D., Nicolas, P., Van Vliet, D., Yamashita, C., Ferguson, J.A., Meuninck, D., Hsu, H-M., Baker, B.B., Shen, L., Loch, T.P. 2019. Large-scale analysis of Flavobacterium psychrophilum multilocus sequence typing genotypes indicates that both newly identified and recurrent clonal complexes are associated with disease. Applied and Environmental Microbiology. 86, e02305-18. Knupp, C., Kiupel, M., Brenden, T.O., Loch, T.P. 2021a. Host-specific preference of some Flavobacterium psychrophilum multilocus sequence typing genotypes determines their ability to cause bacterial coldwater disease in coho salmon (Oncorhynchus kisutch). Journal of Fish Diseases. 44(5), 521-531. Li, S., Chai, J., Cao, Y., Knupp, C., Wang, D., Nicolas, P., Chen, F., Liu, H., Lu, T., Loch, T.P. 2021. Characterization and molecular epidemiological analysis of Flavobacterium psychrophilum recovered from diseased salmonids in China. Microbiology Spectrum. 9(2), e00033021. Loch, T.P., Faisal, M. Flavobacterium spp. In Fish viruses and bacteria: Pathobiology and protection., Woo, P.T.K., Cipriano, R.C., Eds., CABI: 2017, pp. 211-232. Nicolas, P., Mondot, S., Achaz, G., Bouchenot, C., Bernardet, J-F., Duchaud, E. 2008. Population structure of the fish-pathogenic bacterium Flavobacterium psychrophilum. Applied and Environmental Microbiology. 74, 3701-3709. Sebastiao, F., Loch, T.P., Knupp, C., Mukkatira, K., Veek, T., Richey, C., Adkison, M., Griffin, M.J., Soto, E. 2020. Multilocus sequence typing (MLST) analysis of California Flavobacterium psychrophilum reveals novel genotypes and predominance of CC-ST10 in California salmonid hatcheries. Aquaculture Research. 51, 2349-2358. Starliper, C.E. 2011. Bacterial coldwater disease of fishes caused by Flavobacterium psychrophilum. Journal of Advanced Research. 2, 97-108. 6 Sundell, K., Landor, L., Nicolas, P., Jørgensen, J., Castillo, D., Middelboe, M., Dalsgaard, I., Donati, V.L., Madsen, L., Wiklund, T. 2019. Phenotypic and genetic predictors of pathogenicity and virulence in Flavobacterium psychrophilum. Frontiers in Microbiology. 10, 1711. Van Vliet, D., Wiens, G.D., Loch, T.P., Nicolas, P., Faisal, M. 2016. Genetic diversity of Flavobacterium psychrophilum isolates from three Oncorhynchus spp. in the United States, as revealed by multilocus sequence typing. Applied and Environmental Microbiology. 82, 3246- 3255. Van Vliet, D., Loch, T.P., Smith, P., Faisal, M. 2017. Antimicrobial susceptibilities of Flavobacterium psychrophilum isolates form the Great Lakes basin, Michigan. Microbial Drug Resistance. 6, 791-798. 7 Chapter 1: Literature Review 8 1.1. Bacterial coldwater disease (BCWD) and rainbow trout fry syndrome (RTFS) 1.1.1. Losses and economic impact Flavobacterium psychrophilum is considered one of the most economically devastating flavobacterial fish pathogens, resulting in mortality rates as high as 90% in affected salmonid populations (Cipriano and Holt, 2005; Bernardet and Bowman, 2006; Starliper, 2011; Nilsen et al. 2011a,b). Although many reference the economic impact of F. psychrophilum-related losses, an exact monetary value to these losses has not yet been calculated. Considering rainbow trout (Oncorhynchus mykiss) and Atlantic salmon (Salmo salar) are the most valuable and farmed salmonid species worldwide and continue to sustain losses on multiple continents (e.g., Asia, Europe, North America, and South America; Van Vliet et al. 2016; Söderlund et al. 2018; Duchaud et al. 2018; Knupp et al. 2019; Avendaño-Herrera et al. 2020; Li et al. 2021), the economic impact of BCWD likely equates to millions of dollars in losses annually (Duchaud et al. 2018). 1.1.2. Host range Flavobacterium psychrophilum primarily affects fish within the family Salmonidae. Although coho salmon (O. kisutch) and rainbow trout are considered most susceptible (Holt, 1987), reports of BCWD epizootics in Atlantic salmon are also frequent (Nilsen et al. 2011b; Avendaño-Herrera et al. 2020; Macchia et al. 2022). F. psychrophilum has also been recovered from many other salmonids, including amago (O. rhodurus), Apache trout (O. apache), Arctic char (S. alpinus), Arctic grayling (Thymallus arcticus), brown trout (S. trutta), brook trout (Salvelinus fontinalis), Chinook salmon (O. tshawytscha), chum salmon (O. keta), coruh trout (S. coruhensis), cutthroat trout (O. clarkii), European grayling (T. thymallus), lake trout (S. namaycush), lake whitefish (Coregonus clupeaformis), masou salmon (O. masou), 9 pink salmon (O. gorbuscha), sockeye salmon (O. nerka), splake (S. namaycush x S. fontinalis), and white spotted char (S. leucomaenis; Davis 1946, Rucker et al. 1953, Borg 1948, Schachte 1983, Holt 1987, Iida and Mizokami 1996, Ekman et al. 1999, Madetoja et al. 2001, Cipriano and Holt 2005, Fujiwara-Nagata et al. 2013; Saticioglu et al. 2018; Knupp et al. 2019; Harrison et al. 2022; Loch and Knupp, unpublished). F. psychrophilum has also been recovered from non- salmonid species, including ayu (Plecoglossus altivelis), common carp (Cyprinnus carpio), crucian carp (Carassius carassius), European eel (Anguilla Anguilla), European flounder (Platichthys flesus), far eastern brook lamprey (Lethenteron reissneri), forktongue goby (Chaenogobius urotaenia), goldfish (C. auratus), Indian catfish (Clarias batrachus), Japanese eel (A. japonica), Japanese crucian carp funbuna (C. auratus langsdorfii), Japanese dace (Trybolodon hakonensis), Japanese smelt (Hypomesus nipponensis), lake goby (Rhinogobius brunneus), numachichibu (Tridentiger brevispinis), pale chub (Zacco platypus), perch (Perca fluviatilis), roach (Rutilus rutilus), sea lamprey (Petromyzon marinus L.), takahaya (Phoxinus jouyi), tench (Tinca tinca), three-spined stickleback (Gasteroosteus aculeatus), and white sturgeon (Acipenser transmontanus; Lehmann et al. 1991; Wakabayashi et al. 1994; Amita et al. 2000; Iida and Mizokami 1996; Madetoja et al. 2002; Starliper 2011; Fujiwara-Nagata et al. 2013; Harrison et al. 2021). Flavobacterium psychrophilum has also been recovered from other aquatic/semi-aquatic hosts, such as freshwater leeches (Myzobdella lugubris; Schulz and Faisal, 2010), benthic diatoms (Izumi et al. 2005); newts (Brown et al. 1997), aquatic plants (e.g., algae; Amita et al. 2000) and insects (e.g., caddisfly; Fujiwara-Nagata et al. 2013). 1.1.3. Geographic distribution Flavobacterium psychrophilum is geographically widespread, having been recovered from fishes inhabiting at least 24 countries on five continents (e.g., Asia, Australia, Europe, 10 North America, and South America). In Asia, F. psychrophilum has been reported in the Korean Peninsula (Lee and Heo, 1998), Japan (Fujiwara-Nagata et al. 2013), Turkey (Kum et al. 2008; Saticioglu et al. 2018), China (Li et al. 2021), and Russia (Sundell et al. 2019). In Europe, F. psychrophilum has been reported in Belgium (Nematollahi et al. 2003), Denmark (Lorenzen et al. 1991; Nilsen et al. 2014), Estonia (Madetoja et al. 2001), Finland (Dalsgaard and Madsen, 2000; Madetoja et al. 2001), France (Siekoula-Nguedia et al. 2012), Germany (Nilz et al. 2009), Italy (Sarti et al. 1992), Ireland (Lorenzen et al. 1997), Norway (Nilsen et al. 2014), Scotland (Starliper, 2011), Spain (Toranzo and Barja, 1993), Sweden (Ekman et al. 1999), Switzerland (Strepparava et al. 2013), and the United Kingdom (Austin and Stobie, 1991). In South America, F. psychrophilum has been reported in Chile (Avendaño-Herrera et al. 2014; Avendaño-Herrera et al. 2020) and Peru (Leon et al. 2009). Flavobacterium psychrophilum has been reported in Australia (Schmidtke and Carson, 1995). In North America, F. psychrophilum has been reported in multiple states of the USA, including Alaska, California, Colorado, Idaho, Indiana, Maryland, Michigan, Minnesota, Missouri, Montana, North Carolina, New Jersey, New Mexico, New York, Oregon, Pennsylvania, South Dakota, Utah, Vermont, Virginia, Washington, Wisconsin, West Virginia, Wyoming (Holt, 1987; Van Vliet et al. 2016; Knupp et al. 2019; Loch and Knupp, unpublished, https://pubmlst.org/fpsychrophilum), as well as in the Canadian provinces of British Columbia and Ontario (Hesami et al. 2008; Knupp et al. 2019). 1.1.4. Modes of transmission 1.1.4.1. Vertical Flavobacterium psychrophilum is known to be present in the reproductive fluids (e.g., ovarian fluid and milt) and egg surfaces of spawning-aged salmonids (Rangdale et al. 1996; Brown et al. 1997; Taylor 2004; Madsen et al. 2005; Cipriano, 2005; Van Vliet et al. 2016). In 11 addition, there is strong evidence that F. psychrophilum is also present intra-ova, occupying the perivitelline space, which may allow it to circumvent the current and widely utilized egg-surface disinfectant, iodophor (Brown et al. 1997; Cipriano, 2005). Importantly, vertical transmission appears important for the infiltration of fish hatcheries by some F. psychrophilum multilocus sequence typing variants (Knupp et al. 2019) and may account for this bacterium’s intercontinental dispersion via the trade of infected eggs (Nicolas et al. 2008; Knupp et al. 2019; Fujiwara-Nagata et al. 2013; Avendaño-Herrera et al. 2014; Li et al. 2021). 1.1.4.2. Horizontal Flavobacterium psychrophilum has been recovered from and detected in multiple freshwater sources (Wiklund et al. 2000; Strepparava et al. 2013; Nilsen et al. 2014; Nguyen et al. 2018). In vivo cohabitation/contact studies involving infected and naive fish (Madsen and Dalsgaard, 1999; Madetoja et al. 2000) have provided evidence for this bacterium’s horizontal transmission. Seemingly key to F. psychrophilum transmission is shedding from both live and dead infected fish (Madetoja et al. 2000; Madetoja et al. 2002). Indeed, Madetoja et al. (2000) found live infected rainbow trout shed ~103 – 106 cells/fish/hour and for 10 – 21 days, depending on water temperature. Notably, the same authors also found dead rainbow trout shed ~104 – 108 cells/fish/hour and for at least 80 days. Brenden et al. (2023) used this data to develop a model predicting the transmission risk of F. psychrophilum in rainbow trout, and highlighted the importance of removing dead fish from rearing units. However, the applicability of these findings to other F. psychrophilum variants and other BCWD-susceptible salmonid species remains an open question, as these studies were conducted solely in rainbow trout and with one F. psychrophilum variant. 12 1.1.5. Prevention and control Various methods for preventing and controlling BCWD have been thoroughly reviewed elsewhere (Nematollahi et al. 2003; Starliper, 2011; Barnes and Brown, 2011; Loch and Faisal 2017). In brief, BCWD prevention largely depends on comprehensive management strategies. These not only aim to reduce the likelihood of pathogen exposure (e.g., stringent biosecurity measures, disinfection of eggs with iodophor, health monitoring of fish, and the use of water sources free from F. psychrophilum; Starliper, 2011; Van Vliet et al. 2015; Loch and Faisal 2017), but also strive to minimize stress on the fish (Starliper, 2011). Additionally, maintaining optimal fish nutrition, water quality, and rearing conditions is critical (Starliper, 2011). These comprehensive best management practices are essential, particularly given that effective, licensed vaccines for BCWD remain unavailable in the USA at this time (Gomez et al. 2014). Alternative prevention methods are also being explored, including the use of genetically resistant fish (Wiens et al. 2013; Lee et al. 2023), probiotics (Burbank et al. 2012), and phage therapy (Christiansen et al. 2014). Although these strategies have shown promise in reducing the risk of BCWD and minimizing related losses, BCWD outbreaks continue to cause substantial losses in salmonid farming across the globe. Control of BCWD epizootics relies heavily upon antibiotic and chemotherapeutic treatments (e.g., Chloramine-T and hydrogen peroxide). In the USA, Terramycin® (e.g., oxytetracycline dihydrate) and Aquaflor® (florfenicol) are the only two Food and Drug Administration-approved antibiotics for use in food fish. Concerningly, reports of mounting resistance to oxytetracycline have been reported in salmonid farms worldwide, including in the USA (Bruun et al. 2000; Schmidt et al. 2000; Van Vliet et al. 2017; Ngo et al. 2018). 13 1.2. Flavobacterium psychrophilum, causative agent of BCWD and RTFS 1.2.1. Taxonomy Flavobacterium psychrophilum was originally placed within the genus Cytophaga by Borg (1948), and because of this species’ affinity for low temperatures, it was given the name Cytophaga psychrophila. Because F. psychrophilum does not produce fruiting bodies or degrade polysaccharides, C. psychrophila was re-classified as Flexibacter psychrophilus (Bernardet and Grimont 1989), but subsequent analysis of DNA G+C content indicated that F. psychrophilus was most like bacteria within the genus Flavobacterium. In 1996, Bernardet et al. (1996) emended the description of the genus Flavobacterium and renamed Flexibacter psychrophilus as Flavobacterium psychrophilum (Phylum Bacteroidota; Class Flavobacteriia; Order Flavobacteriales; Family Flavobacteriaceae). 1.2.2. Phenotypic characteristics Flavobacterium psychrophilum is a Gram-negative, rod-shaped (0.5-µm wide × 1 – 5- µm long), and weakly refractile bacterium (Pacha 1968; Bernardet and Kerouault 1989). In addition, F. psychrophilum is strictly aerobic and produces yellow-pigmented (due to the presence of a flexirubin-type pigment) colonies with a raised center and a smooth or thinly spreading edge (Bernardet and Kerouault 1989, Holt et al. 1987), although colony morphology varies (Hogfors-Ronnholm and Wiklund, 2010). This bacterium grows in the presence of up to 1.0% NaCl and at a pH range of 4.0 - 8.0 (Bernardet and Kerouault 1989). The bacterium degrades multiple substrates (e.g., tributyrin, tyrosine, lecithin, tween 20, and tween 80; Bernardet and Kerouault 1989) and proteolyzes many host tissue constituents (e.g., gelatin, casein, elastin, collagen, fibrinogen, chondroitin sulphate, and fish muscle extract; Holt, 1987; Lorenzen et al. 1997; Soule et al. 2005). In contrast, F. psychrophilum cannot hydrolyze 14 xanthine, chitin, starch, agar, carboxymethylcellulose, or esculin. Some F. psychrophilum variants reportedly use simple or complex carbohydrates, but this trait appears uncommon (Bernardet and Kerouault 1989; Cepeda et al. 2004; Cipriano and Holt 2005). This bacterium is weakly positive for both cytochrome oxidase and catalase, does not produce hydrogen sulfide, indole, arginine dihydrolase, lysine decarboxylase, or ornithine decarboxylase (Bernardet and Kerouault 1989; Cipriano and Holt 2005). Flavobacterium psychrophilum cannot absorb congo red nor reduce nitrate to nitrite (Bernardet and Kerouault 1989). 1.2.3. Proteolytic activity Although F. psychrophilum reportedly proteolyzes multiple protein substrates, including albumin, casein, chondroitin sulfate, collagen, elastin, fibrinogen, and fish muscle extract (Pacha, 1968; Otis, 1984; Holt, 1987; Bertolini et al. 1994; Dalsgaard and Madsen, 2000; Madetoja et al. 2001; Madetoja et al. 2002; Soule et al. 2005; Sundell and Wiklund, 2015; Rochat et al. 2019; Sundell et al. 2019; Knupp et al. 2021), most studies assessing this bacterium’s proteolytic activity have focused on the degradation of casein, elastin, and gelatin (Pacha, 1968; Otis, 1984; Holt, 1987; Bertolini et al. 1994; Dalsgaard and Madsen, 2000; Madetoja et al. 2001; Madetoja et al. 2002; Soule et al. 2005; Sundell and Wiklund, 2015; Rochat et al. 2019; Sundell et al. 2019; Knupp et al. 2021). In this context, the casein, elastin, and gelatin-degrading activity of >260 F. psychrophilum isolates recovered from numerous host species (e.g., Arctic char, Atlantic salmon, ayu, brook trout, brown trout, chinook salmon, coho salmon, flounder, perch, rainbow trout, sea trout, tench, and white sturgeon) in Asia, Europe, North America, and South America has been evaluated, although most isolates were recovered from rainbow trout. Nevertheless, of these F. psychrophilum isolates, only two did not degrade casein and/or gelatin (e.g., WB-1 recovered from ayu and FPS-S11B from rainbow trout; Nakayama et al. 2016; Sundell et al. 15 2019) and interestingly, were avirulent to their respective host species, possibly suggesting at least some proteolytic activity is required for virulence. In contrast to the nearly universal observation that F. psychrophilum degrades casein and gelatin, the ability of F. psychrophilum to degrade elastin is far more variable and may relate to the possession of a recently described elastase gene, FP0506 (Rochat et al. 2019). However, elastase activity does not dictate the capacity to cause mortality in rainbow trout, as many elastase negative F. psychrophilum isolates have been recovered from diseased trout suffering mortality events (Dalsgaard and Madsen, 2000; Madetoja et al. 2002; Soule et al. 2005; Sundell and Wiklund, 2015; Rochat et al. 2019; Sundell et al. 2019). Likewise, multiple F. psychrophilum elastase negative isolates were virulent to rainbow trout following in vivo challenge, though fish were challenged via injection (Sundell et al. 2019), a route that bypasses some immune defenses (Fast et al. 2002). Besides virulence, elastase activity may also be associated with genetic lineage (Soule et al. 2005) and/or MLST variant, considering most F. psychrophilum isolates belonging to CC-ST10 degrade elastin (Sundell et al. 2019; Rochat et al. 2019; Knupp et al. 2021). 1.2.4. Genetic diversity and epidemiology The diversity and epidemiology of F. psychrophilum has been studied for decades using multiple approaches, including pulsed field gel electrophoresis (Arai et al. 2007; del Cerro et al. 2010), plasmid profiling (Holt, 1987; Lorenzen et al. 1997; Chakroun et al. 1998; Izumi 2004; Madsen and Dalsgaard, 2000), randomly amplified polymorphic DNA analysis (Chakroun et al. 1997, Valdebenito and Avendaño-Herrera 2009), restriction fragment length polymorphism analysis (Izumi et al. 2003; Soule et al. 2005; Hesami et al. 2008), and ribotyping (Cipriano et al. 1996; Chakroun et al. 1998; Madsen and Dalsgaard, 2000). Currently, the most widespread method for characterizing F. psychrophilum intraspecific diversity is multilocus sequence typing 16 (Nicolas et al. 2008), which has been applied to >1500 F. psychrophilum isolates recovered across Asia, Australia, Europe, North America, and South America, revealing the existence of >260 different sequence types (STs; i.e., genetic variants; https://pubmlst.org/fpsychrophilum). Notably, MLST-based findings are largely consistent with whole genome-MLST-based findings (Duchaud et al. 2018). Most F. psychrophilum genetic variants in North America differ from those found in other continents (Nicolas et al. 2008; Siekoula-Nguedia et al. 2012; Fujiwara- Nagata et al. 2013; Strepparava et al. 2013; Nilsen et al. 2014; Avendaño-Herrera et al. 2014; Van Vliet et al. 2016; Knupp et al. 2019; Li et al. 2021) and some variants also appear to differ in host species association (Nicolas et al. 2008; Knupp et al. 2019; Knupp et al. 2021) and recovery environment (Van Vliet et al. 2016; Sebastião et al. 2020; Knupp et al. 2019). However, whether some F. psychrophilum variants are host specific, better suited to the artificial rearing environment, differ in transmission strategies, and/or are circumventing current gold-standard detection methods (e.g., culture), remains to be determined. 1.2.4.1. Multilocus sequence typing Multilocus sequence typing (MLST) is a highly discriminatory and reproducible technique for characterizing the intraspecific diversity of bacteria. Originally developed for Neisseria meningitidis, a cause of meningitis in humans (Maiden et al. 1998), MLST has since been applied to multiple bacterial pathogens affecting multiple taxa, including fish (Nicolas et al. 2008). Nicolas et al. (2008) developed a MLST scheme for F. psychrophilum in 2008, relying upon partial sequences of seven conserved housekeeping genes, including trpB, gyrB, dnaK, tuf, fumC, murG, and atpA. In that pioneering study, 50 F. psychrophilum isolates recovered from 10 fish species in Asia, Europe, North America, and South America were genotyped, revealing the existence of 30 distinct sequence types (STs; i.e., genetic variants), some of which (e.g. ST2, 17 ST10, and ST13) were seemingly associated with a specific host species (e.g., rainbow trout, Oncorhynchus mykiss or coho salmon, O. kisutch; Nicolas et al. 2008). Later, Siekoula-Nguedia et al. (2012) examined 66 F. psychrophilum isolates from farmed rainbow trout in France, uncovering previously identified ST2, as well as 14 new genetic variants, most of which belonged to clonal complex (CC; i.e., a group of closely related variants)-ST2. Fujiwara-Nagata et al. (2013) studied 114 F. psychrophilum isolates recovered from 15 different fish species in Japan, finding 32 new genetic variants, some of which (e.g., ST40, ST45, ST54, and ST55) appeared associated with ayu (Plecoglossus altivelis) or masou salmon (O. masou). Moreover, some previously detected genetic variants (e.g., ST10, in CC-ST2 and ST13, in CC-ST9) were discovered, and again infecting either rainbow trout or coho salmon, continuing to provide observational evidence these STs/CCs may have host associations. In the same year, Strepparava et al. (2013) genotyped 112 F. psychrophilum isolates from Switzerland, finding CC-ST2 variants to be responsible for BCWD in their rainbow trout farms. Nilsen et al. (2014) analyzed 560 F. psychrophilum isolates recovered from 10 fish species in Denmark, Finland, Norway, and Sweden, revealing 81 different F. psychrophilum genetic variants. At this time, CC-ST2 was renamed CC-ST10 as ST10 had the most single locus variants. Moreover, the variants belonging to CC-ST10 and described in this study were almost exclusively recovered from rainbow trout. In the same year, Avendaño-Herrera et al. (2014) examined 91 F. psychrophilum isolates recovered from Atlantic salmon (Salmo salar), coho salmon and rainbow trout in Chile, and found most CC-ST10 variants were recovered from rainbow trout, whereas those belonging to CC-ST9 and CC-ST21 were mostly recovered from coho salmon and Atlantic salmon, respectively. 18 Despite the extensive MLST-based analyses conducted to this point in Europe, Asia, and South America, only 10 F. psychrophilum isolates originating from North America had been typed using the same method (Nicolas et al. 2008). Then, Van Vliet et al. (2016) studied 96 F. psychrophilum isolates recovered from Chinook salmon (O. tshawytscha), coho salmon, and rainbow trout in nine different USA states, and identified 34 STs. Like previous studies, most BCWD epizootics in rainbow trout were caused by CC-ST10 variants, whereas CC-ST9 was most prevalent among isolates recovered from coho salmon. Next, Knupp et al. (2019) genotyped 314 F. psychrophilum isolates recovered from 10 fish species in 20 states in the USA, as well as one Canadian province, and found 66 genetic variants, 47 of which were newly described. Most isolates belonging to a CC-ST10 variant were recovered from rainbow trout. Similarly, CC-ST9 variants were recovered from coho salmon. Additional CCs with apparent host associations for Atlantic salmon (e.g., CC-ST232) and rainbow trout (e.g., CC-ST191, CC- ST281, and CC-ST310) were also described, as was at least one generalist CC (e.g., CC-ST256) that was isolated from naturally infected rainbow trout and chinook salmon. Moreover, it was proposed that some genetic variants (e.g., ST253) may be better suited to hatchery environments and fish farms, as evidenced by their repeated recovery from the same facility over multiple years. This contrasted with other F. psychrophilum variants, such as ST256 and ST257, that were only recovered from wild/feral fish. Sebastião et al. (2020) genotyped 49 F. psychrophilum isolates recovered from Chinook salmon and rainbow trout in California, USA and found most isolates recovered from rainbow trout belonged to CC-ST10. Li et al. (2021) used MLST to assess genetic diversity among 31 F. psychrophilum isolates recovered from brook trout (Salvelinus fontinalis), rainbow trout, and masou salmon in China, and described five genetic variants, including two that belonged to CC-ST10, both of which were recovered from rainbow 19 trout. Calvez et al. (2021) genotyped 31 F. psychrophilum isolates recovered from rainbow trout in France, and most belonged to CC-ST10. Flavobacterium psychrophilum genetic diversity according to MLST has been studied for nearly two decades and observations suggest some variants may be host specific or generalists. Although a few studies have directly or indirectly investigated such associations under in vivo laboratory conditions (Holt, 1987; Ekman and Norrgren, 2003; Fredriksen et al. 2016), most have used a less natural exposure route (e.g., injection) that bypasses important immune defenses (Fast et al. 2002; Dash et al. 2018). In contrast, at least one study reported the virulence of two host-associated F. psychrophilum variants (e.g., US19-coho salmon in ST9 and US53-rainbow trout in ST78) in coho salmon following laboratory immersion exposure (Knupp et al. 2021), and found US19-COS caused disease and mortality in coho salmon but US87-RBT did not. Although this study provided evidence some F. psychrophilum variants are host specific, a study has yet to simultaneously cross-challenge multiple salmonid species of a similar age with multiple putatively host specific F. psychrophilum variants. Besides host specificity, some F. psychrophilum variants appear to be associated with the fish farming environment; however, the long-term survival of different F. psychrophilum variants within microenvironments simulating fish farm/hatchery conditions has not been attempted. 1.2.5. Serotypic diversity The serotypic diversity of F. psychrophilum has also been studied for decades (Pacha, 1968; Cipriano and Holt, 2005), although varying methodologies (e.g., slide agglutination and enzyme-linked immunosorbent assay; Wakabayashi et al. 1994; Lorenzen and Olesen, 1997; Izumi et al. 2003; Mata et al. 2002) and reagents have precluded many interlaboratory comparisons. Most recently, however, a reproducible PCR-based serotyping assay, which is 20 based on a widely accepted serotyping scheme (e.g., FpT, Fd, and Th) of Lorenzen and Olesen (1997) and targets putative O-polysaccharide genes, was developed by Rochat et al. (2017) to detect four serotypes (e.g., Type-0 – Type-3). Notably, Type-0 has a less conserved genomic structure and thus encompasses isolates not belonging to Types 1 – 3 (i.e., could contain several to many yet to be defined molecular serotypes). In this pioneering study, which examined 244 F. psychrophilum isolates from Canada, Chile, Denmark, Finland, France, Germany, Israel, Italy, Japan, Norway, Oregon, Scotland, Spain, Switzerland, Tasmania, and the USA, it was apparent that some serotypes were associated with specific host species. For example, most (n = 22/23 = 95.6%) isolates recovered from coho salmon belonged to Type-0, most isolates recovered from rainbow trout belonged to Type-1 (n = 59/151, 39.1%) or Type-2 (n = 60/151, 39.7%), and all (n = 35/35) isolates recovered from ayu belonged to Type-3 (Rochat et al. 2017). Saticioglu et al. (2018) serotyped 25 F. psychrophilum isolates recovered from rainbow trout in Turkey, finding most isolates belonged to Type-1 (n = 8/25, 32%) or Type-2 (n = 10, 40%), but a few belonged to Type-0 (n = 3, 12%) or Type-3 (n = 3, 12%), while one was untypable. Avendaño-Herrera et al. (2020) serotyped 118 F. psychrophilum isolates recovered from rainbow trout (n = 85), Atlantic salmon (n = 32), and coho salmon (n = 1) from Chile, and identified another O-antigen gene conserved among some F. psychrophilum isolates (e.g., Type-4). Using this updated molecular serotyping scheme, most Chilean isolates recovered from rainbow trout belonged to Type-2 (n = 50/85, 58.8%) followed by Type-4 (n = 17/85, 20.0%), Type-1 (n = 14/85, 16.5%), and Type-0 (n = 4/85, 4.7%), whereas most isolates recovered from Atlantic salmon belonged to Type-4 (n = 29/32, 90.6%), followed by Type-0 (n = 2, 6.3%), and Type-1 (n = 1/32, 3.1%). The isolate recovered from coho salmon belonged to Type-2. In China, Li et al. (2021) serotyped eight F. psychrophilum isolates recovered from rainbow trout, finding all belonged to Type-1. In 21 the same year, Calvez et al. (2021) serotyped 31 F. psychrophilum isolates recovered from rainbow trout in France, revealing most belonged to Type-2 (n = 14/31, 45.2%), followed by Type-0 (n = 9, 29.0%) and Type-1 (n = 8, 25.8%). In the USA, Knupp et al. 2021 serotyped one F. psychrophilum isolate recovered from rainbow trout and one isolate recovered from coho salmon, finding they belonged to Type-2 and Type-0, respectively. Collectively, these studies provide evidence that rainbow trout are seemingly most affected by Type-1 and Type-2 F. psychrophilum isolates, and to a lesser extent, Type-0, Type-3, and Type-4. At present, Atlantic salmon seem to be most affected by Type-2 and Type-4, and sometimes Type-0, Type-1, and Type-3. Coho salmon are most affected by Type-0 and, infrequently, Type-2. Although observations suggest molecular serotypes may have host associations, in vivo evidence for such associations are lacking. Indeed, if some F. psychrophilum serotypes are associated with specific host species this could have important implications, such as for strain selection for BCWD vaccines and selective breeding programs. 1.2.6. Detection methods from fish Pathogen detection is essential to disease diagnosis, which in turn informs management and treatment strategies. These strategies are designed to mitigate associated losses and prevent further transmission of the pathogen. In this context, multiple molecular assays have been developed to detect F. psychrophilum in fish tissue, including conventional, nested, multiplex, and quantitative PCRs targeting single (e.g., rpoC, gyrB, parE, RFPS00910; Izumi et al. 2005; Marancik and Wiens, 2013; Strepparava et al. 2014) and multicopy genes (e.g., 16S rRNA Nakagawa and Yamasota, 1993; Toyama et al. 1994; Urdaci et al. 1998). Likewise, multiple serological assays, including agglutination (Nagai and Nakai, 2011), immunofluorescent antibody technique (Aoki et al. 2005; Vatsos et al. 2006), immunohistochemistry (Lorenzen and 22 Karas, 1992; Evensen and Lorenzen, 1997; Madetoja et al. 2000), and enzyme-linked immunosorbent assay (Mata and Santos, 2001; Crump et al. 2003; Lindstrom et al. 2009) have been developed to directly detect F. psychrophilum in fish tissue. Although these assays are invaluable tools for detecting this fish pathogen and facilitate a timely diagnosis, a primary limitation is their inability to recover viable F. psychrophilum isolates, which are needed for antibiotic susceptibility testing, conducting molecular typing assays, developing vaccines, and for further research (Van Vliet et al. 2017; Ma et al. 2019; Knupp et al. 2019). These needs, coupled with the fact that most of these assays do not readily differentiate between live and dead F. psychrophilum cells, contribute to the reason why culture-based F. psychrophilum diagnostics are considered the gold-standard for detecting and identifying F. psychrophilum. In fact, the American Fisheries Society – Fish Health Section (AFS-FHS) Blue Book identifies culture as a prerequisite step to confirmatory diagnosis of BCWD (AFS-FHS 2020). In this context, many culture media have been used for F. psychrophilum isolation and identification, and although variation in employed medium occurs by country/region, tryptone yeast extract salts medium (TYES; Holt, 1987) is among the most widely used and thus could be considered the current gold-standard medium. However, TYES was developed before we had a more complete understanding of F. psychrophilum diversity, and ongoing research suggests it may not be ideally-suited for the recovery of all F. psychrophilum variants. 1.2.7. Bacterial recovery on solid media 1.2.7.1. Names and formulations Anacker and Ordal (AO) medium, also referred to as Cytophaga agar, was developed by Anacker and Ordal (1959) for F. psychrophilum recovery and is formulated with tryptone, beef extract, yeast extract, and sodium acetate (Table 1.1). Since its development, at least 9 23 derivations of AO, typically referred to as modified or enriched AO (mAO or EAO), have been used for F. psychrophilum recovery (Table 1.1). Most derivations have retained tryptone, beef extract, yeast extract, and sodium acetate but have modified their concentration and/or added medium components, including CaCl2, fetal bovine serum, horse serum, skimmed milk, fish or horse blood, carbohydrates (e.g., glucose, galactose, rhamnose), activated charcoal, and/or aromatic compounds (e.g., L-Tyrosine, L-Phenylalanine, L-Tryptophan, 4-Aminobenzoic acid, 4-Hydroxybenzoic acid, and 2,3-Dihydroxybenzoic acid). In addition, some derivations removed tryptone and beef extract and replaced them with a different protein source (e.g., peptone) or removed sodium acetate (Table 1.1). Tryptone yeast extract salts medium (TYES) was developed by Holt (1987) for F. psychrophilum recovery and is formulated with tryptone, yeast extract, CaCl2, and MgSO4 (Table 1.2). Since its development, at least three derivations of TYES have been used for F. psychrophilum recovery, which have adjusted the concentration of one (e.g., CaCl2; F. psychrophilum agar, FLPA; Cepeda et al. 2004) or two components (e.g., yeast extract and agar; no change to medium name; Madetoja et al. 2000) and/or added new components, including horse serum (modified TYES, mTYES; Knupp et al. 2021) or glucose (FLPA; Table 1.2). Besides TYES, AO, and their derivations, other media have been used to recover F. psychrophilum. These media include nutrient agar (Secades et al. 2001), Shieh’s medium (Holt, 1987), Hsu-Shotts agar (Cipriano and Holt, 2005), and modified veggietone (Ngo et al. 2017; Table 1.3). 24 1.2.7.2. Comparative studies 1.2.7.2.1. Michel et al. (1999) In a study by Michel et al. (1999), five F. psychrophilum isolates recovered from rainbow trout in France (n = 3), Idaho (n = 1), and Denmark (n = 1) were grown in four different broth solutions. Then, bacterial yields (i.e., measured in colony forming unit; cfu) on EAO [(Bernardet and Kerouault (1989) formulation; Table 1.1] and TYES [Holt (1987) formulation with 1% skimmed milk; Table 1.2) with and without 10% fetal calf serum (FCS) were compared. These comparisons showed that TYES alone did not recover F. psychrophilum, but TYES+FCS recovered <1 cfu, on average, from 1/4 broth solutions. In contrast, EAO recovered 0.66 – 10.3 mean F. psychrophilum cfus across all broth solutions, and EAO+FCS improved recovery further by obtaining 16.3 – 84 mean cfus. In subsequent trials, where FCS was replaced by horse serum (HS, 5 or 10% concentration), it was observed that HS improved recovery over FCS by 18.3%. Thus, the authors suggested that EAO should be reformulated to include 5% HS. Michel et al. (1999) presented an improved F. psychrophilum recovery medium; however, it is unclear if all five isolates were recovered similarly (i.e., cfus were reported as means without standard deviations). Furthermore, replication and statistical analyses were not reported, thus it is difficult to determine if this modified EAO yielded significantly more cfus in comparison to the original base media. 1.2.7.2.2. Crump et al. (2001) Although not the primary goal of Crump et al. (2001), F. psychrophilum recovery was compared between mAO (Bernardet and Kerouault, 1989; Table 1.1) and mAO with 3% fish or horse blood using four F. psychrophilum isolates recovered from the USA (n = 2), England (n = 25 1), and Denmark (n = 1). As a result, mAO with blood was observed to enhance F. psychrophilum recovery, though to what extent was not reported. 1.2.7.2.3. Cepeda et al. (2004) In a study by Cepeda et al. (2004), recovery of 13 F. psychrophilum isolates recovered from rainbow trout (n = 8), coho salmon (n = 3), European eel (n = 1), and common carp (n = 1) in Spain (n = 5), France (n = 3), the USA (n = 2), the U.K. (n =1), and Japan (n = 1) were compared on five solid media, including MAO (Toranzo and Barja, 1993; Table 1.1), MAO with glucose (MAOG), TYES (Cepeda et al. 2004 formulation), FLPA (this study; Table 1.2), and AO (Anacker and Ordal, 1959; Table 1.1) agar with carbohydrates (e.g., glucose, galactose, and rhamnose) and skimmed milk (CACM). Following these comparisons, FLPA recovered the most F. psychrophilum cfus overall, followed by TYES and CACM (i.e., tryptone-containing media), and recovery was reportedly worse in the peptone-containing media (e.g., mAO and MAOG). Moreover, and as noted in this study, recovery varied according to F. psychrophilum isolate, although the extent to which this occurred was not discussed. In addition to comparing F. psychrophilum recovery in the laboratory, Cepeda et al. (2004) compared media (e.g., MAO, MAOG, TYES, FLPA, and CACM) as to their ability to recover F. psychrophilum from the kidney and spleen of naturally infected, farm-reared “trout” via streak plating (n = 17 cultures over two years). Here too, FLPA was most effective at F. psychrophilum recovery, and recovered the bacterium in pure and mixed (e.g., with Aeromonas and Enterobacter spp.) cultures. 1.2.7.2.4. Álvarez and Guijarro (2007) In a study by Álvarez and Guijarro (2007), the recovery of one F. psychrophilum isolate recovered in the USA (e.g., THC02-90, MLST ST9 in CC-ST9; Nicolas et al. 2008) was 26 compared on six media, including nutrient agar (NA), NA with activated charcoal (NAC), EAO with 5% horse serum (EAOS; Michel et al. 1999), EAOS with aromatic compounds (e.g., L- Tyrosine, L-Phenylalanine, L-Tryptophan, 4-Aminobenzoic acid, 4-Hydroxybenzoic acid, and 2,3-Dihydroxybenzoic acid; EAOSa), EAOS with activated charcoal (EAOC), and EAOC with aromatic compounds (EAOCa). Results showed that the addition of charcoal significantly increased F. psychrophilum recovery across media. Moreover, and although NAC yielded the most colonies overall, F. psychrophilum recovery on EAOCa was less variable and thus ultimately the recommended medium for future studies. 1.2.7.2.5. Oplinger and Wagner (2012) In a study by Oplinger and Wagner (2012), F. psychrophilum growth was compared in broth, but because the newly optimized medium developed by the authors was also shown to support F. psychrophilum recovery, it has been included in this section. The growth of one F. psychrophilum isolate recovered in France (e.g., ATCC FP49510) was initially compared in six iterations of EAO (Bernardet and Kerouault, 1989; Table 1.1). Within iterations, individual medium component (tryptone, yeast extract, beef extract, sodium acetate, skimmed milk, maltose, horse serum, CaCl2, and MgSO4) concentrations were varied (2 – 4 concentrations per medium component). Overall, tryptone, yeast extract, beef extract, skimmed milk, and horse serum were reported to benefit F. psychrophilum growth, and thus were included in a newly developed medium. The newly optimized medium was compared to EAO, TYES, and MAT (TYES with 1% maltose) using the same F. psychrophilum isolate ATCC FP49510, as well as CSF259-93 (MLST ST10 in CC-ST10; Nicolas et al. 2008), revealing that growth was best in the newly developed medium (OW), followed by MAT, EAO, and TYES. 27 Although these studies each reported an improved F. psychrophilum recovery medium, most only tested 1 – 5 isolates. Notably, however, in the one study (Cepeda et al. 2004) that used 13 F. psychrophilum isolates recovered from multiple host species, noticeable variation in cfu yields were described. Thus, future research should prioritize incorporating a diversity of F. psychrophilum isolates to produce an optimized recovery medium. 1.2.8. Detection and quantification from water Detection and quantification of F. psychrophilum from water containing fish has been conducted via culture and immunofluorescence antibody technique (Madetoja et al. 2000; Madetoja and Wiklund, 2002; Madetoja et al. 2003), both of which are sensitive (e.g., detection limit of ~101 - 102 cfus/mL) but time consuming and vary in specificity. One quantitative PCR (qPCR) has been reported in the published literature for detecting and quantifying F. psychrophilum from water (Strepparava et al. 2014; Nguyen et al. 2018) and although sensitive (detection limit of ~101 gene copies) and specific, its quantification limit was relatively high (e.g., ~103 F. psychrophilum cells/mL; Strepparava et al. 2014). Of note, some laboratories have noticed non-specific amplification with this assay (Loch and Soto, unpublished). Indeed, the development or optimization of a qPCR assay to detect F. psychrophilum from water could be useful for deepening our understanding of transmission dynamics and may be useful for early F. psychrophilum detection in fish farms and hatcheries, thereby potentially mitigating losses. 1.2.9. Survival outside of fish The widespread distribution and apparent success of F. psychrophilum may be partially related to its ability to survive outside its host. For example, Vatsos et al. (2003) found F. psychrophilum could survive for 133 days in sterile stream water. A lengthy survival time in water was also noted by Madetoja et al. (2003), whereby F. psychrophilum remained viable in 28 sterile lake water for 300 days. Although these studies are invaluable for clarifying F. psychrophilum persistence outside of a fish host, they were conducted using a total of three F. psychrophilum isolates (e.g., NCIMB 1947T, ST9 in CC-ST9; V9/93, MLST variant unknown; B97026, MLST variant unknown) and represent the totality of our knowledge on this subject. Thus, it is still unknown whether the ability to persist long term in environments with limited nutrients is a widespread trait among F. psychrophilum variants. Madetoja et al. (2003) also found that F. psychrophilum survival could be improved (e.g., cells were recovered over longer period and at higher concentrations) if lake water also contained natural beach sand. Although beach sand is not commonly found in fish farm and hatchery rearing units, other sediments are, like detritus and uneaten fish food (Schumann, 2021). However, F. psychrophilum survival in these rearing unit microenvironments has yet to be explored and could partially explain why some MLST variants (e.g., ST10, ST253) are repeatedly recovered (i.e., over multiple years) from the same facility (Knupp et al. 2019). Indeed, by improving our understanding of F. psychrophilum environmental persistence strategies within fish farm/hatchery environments, we may be able to disrupt F. psychrophilum transmission pathways and mitigate the spread and losses caused by this bacterium. 1.2.10. Bacterial inoculum preparations Various liquid culture media, such as TYES, iron-limited TYES, TYES with maltose or horse serum, Shieh (with or without iron), EAO, and modified EAO have been employed for the in vitro growth of F. psychrophilum prior to experimental challenge (Madsen and Dalsgaard, 1999; Decostere et al. 2000; Garcia et al. 2000; Madetoja et al. 2000; Aoki et al. 2005; Long et al. 2014; Sundell et al. 2019; Knupp et al. 2021). Similarly, a range of incubation conditions, including varying temperatures (e.g., 15 – 18 °C), durations (e.g., 18, 24, 48, or 72 hours), with 29 or without agitation, and different methods of estimating inoculum size prior to infection (e.g., correlating incubation time to growth phase or using optical density to correlate with viable cell counts) have been used (Madsen and Dalsgaard, 1999; Decostere et al. 2000; Garcia et al. 2000; Madetoja et al. 2000; Aoki et al. 2005; Long et al. 2014; Sundell et al. 2019; Knupp et al. 2021). Importantly, this variation in F. psychrophilum inoculum preparation affects study outcomes, thereby impeding comparability of findings among laboratories. For example, Aoki et al. (2005) found F. psychrophilum was more virulent when used in logarithmic phase in comparison to other culture phases. Thus, additional studies comparing the virulence of different F. psychrophilum variants following different culture conditions are warranted and may result in a reliable and standardized approach to inoculum preparation. 1.2.11. Experimental challenge models Experimental challenge models are not only crucial for fulfilling Koch’s postulates (Walker, 2006), but are invaluable for investigating host-pathogen interactions, including pathogenesis, virulence mechanisms, host immune response and susceptibility, and for assessing the efficacy of BCWD prevention (e.g., prophylactics and vaccines) and control (e.g., bacteriophages and antimicrobials) measures (Ekman and Norrgren, 2003; Van Vliet et al. 2017; Perez-Pascual et al. 2017; Rochat et al. 2019; Ma et al. 2019; Sundell et al. 2020; Semple et al. 2020; Deng et al. 2022; Huyben et al. 2023). To this end, various infection routes, including subcutaneous, intraperitoneal, and intramuscular injection (Holt, 1987; Obach and Laurencin, 1991; Fredriksen et al. 2016), oral and anal intubation, cohabitation of naïve with infected fish, and bath immersion methods have been used with varying success and reproducibility (Lorenzen et al. 1991; Decostere et al. 2000; Madetoja et al. 2000; Knupp et al. 2021). 30 Injection methods (e.g., subcutaneous, intraperitoneal, intramuscular) are generally considered to be the most reproducible exposure route as they generate consistent mortality among experimentally challenged fish (Madsen and Dalsgaard, 1999; Garcia et al. 2000). Moreover, some studies suggest intramuscular injection may be most advantageous, especially among larger (e.g., >5g) fish (Garcia et al. 2000; Fredriksen et al. 2013). For example, Fredriksen et al. (2013) used intramuscular injection to cause mortality among ~36.6g rainbow trout but was unable to produce mortality via intraperitoneal injection, even at a ~10-fold higher dose. Likewise, Holt (1987) found that to elicit comparable mortality in coho salmon, a substantially higher dose was required for intraperitoneal injection in comparison to intramuscular injection. Although injection methods may be preferred under specific circumstances, they inherently bypass natural barriers to infection, such as skin, mucus, lysozyme, complement, heat shock proteins, and immunoglobulins (Fraslin et al. 2018). For these reasons, injection methods may confound study findings, especially those influenced by host defenses. Indeed, injection could partially explain why some F. psychrophilum isolates belonging to MLST variants with apparent host-associations were able to cause mortality in multiple “unnatural” host species. For example, Holt (1987) found F. psychrophilum isolate SH3-81, which was recovered from coho salmon and belongs to CC-ST9 (i.e., a widespread, coho salmon-associated CC; Fujiwara-Nagata et al. 2013; Avendaño-Herrera et al. 2014; Knupp et al. 2019) was virulent via injection to not only coho salmon but chinook salmon and rainbow trout. Likewise, Bruce et al. (2021) found F. psychrophilum isolate CSF259-93, which was recovered from rainbow trout and belongs to CC-ST10 (i.e., the largest and most widespread rainbow trout-associated CC; Nicolas et al. 2008; Fujiwara-Nagata et al. 2013; Avendaño- Herrera et al. 2014; Nilsen et al. 2014; Knupp et al. 2019) was virulent via injection to Atlantic 31 salmon and brook trout. However, it is also important to consider the true host specific nature of these isolates are unknown and may also contribute to these findings. The prevalent use of injection methods in experimental challenges may inadvertently be shaping our understanding of F. psychrophilum-host interactions and consequently could significantly impact the development of BCWD prevention and control strategies. If our understanding of F. psychrophilum host associations and host response is primarily informed by injection studies, it may lead to narrow focus on specific F. psychrophilum variants while overlooking the substantial intraspecific diversity present in this species. Consequently, we risk developing narrow-spectrum vaccines or salmonid genetic lines resistant to only a limited set of F. psychrophilum variants. Thus, it is important to further evaluate F. psychrophilum host associations and host responses using more natural exposure routes that better capture the intricacies of host-pathogen interactions. In this context, immersion studies, which involve immersing fish in suspensions of F. psychrophilum, more accurately simulate natural exposure. However, achieving reproducibility often necessitates disturbing a portion of the skin/mucus layer of the fish using mechanical or chemical methods (Madsen and Dalsgaard 1999, Garcia et al. 2000, Madetoja et al. 2000, Henriksen et al. 2013, Long et al. 2013; Macchia et al. 2022). The ideal approach would minimize this disruption, thereby maintaining the integrity of the fish’s innate defenses. In this context, Madetoja et al. (2000) found that creating a 1 – 2-mm incision below the dorsal fin was sufficient for F. psychrophilum to invade and cause >80% mortality in rainbow trout. Long et al. (2013) described nearly identical findings in rainbow trout by using the same method. However, this wound type is not representative of those naturally incurred by fish in farms/hatcheries. Thus, Long et al. (2013) also attempted to induce mortality in rainbow trout by removing the 32 adipose fin, which is a common practice used in salmonid hatcheries for delineating hatchery fish from wild fish (Auld et al. 2019). In this context, adipose fin clipping prior to F. psychrophilum exposure led to successful bacterial invasion and resulted in >80% mortality. Besides rainbow trout, adipose fin-clipping prior to immersion has been successful in causing mortality among coho salmon (Holt, 1987; Knupp et al. 2021). Whether this exposure method is suitable for other intensively reared and economically important salmonid species (e.g., Atlantic salmon) remains to be determined but warrants further investigation. Indeed, an effective immersion challenge model for multiple salmonid species would be instrumental in examining F. psychrophilum-host interactions and testing the efficacy of BCWD vaccines. 1.2.12. Gross signs of BCWD in rainbow trout (Oncorhynchus mykiss), coho salmon (O. kisutch), and Atlantic salmon (Salmo salar) 1.2.12.1. Rainbow trout External BCWD signs in naturally infected rainbow trout include lethargy, anorexia, distended abdomen, deep, focally extensive hemorrhagic ulceration on the dorsal aspect of the caudal peduncle, complete erosion of the caudal fin/peduncle, exophthalmia, pale gills, and dark skin pigmentation (Nematollahi et al. 2003; Nilsen et al. 2011a; Starliper et al. 2011; Li et al. 2021). Internally, naturally infected rainbow trout present with signs of anemia (e.g., pale liver and kidney), and have a swollen spleen and/or enteritis (Nematollahi et al. 2003; Nilsen et al. 2011a; Starliper et al. 2011; Li et al. 2021). Following experimental immersion challenge, identical external and internal signs of BCWD have been observed, along with an additional sign - erosion of the mouth (Aoki et al. 2005; Henriksen et al. 2013; Hoare et al. 2017). 33 1.2.12.2. Coho salmon External BCWD signs in naturally infected coho salmon include erratic swimming, ulcerations on the caudal peduncle, trunk, jaw, and/or skull, dorsal swelling posterior to the skull, and dark skin pigmentation (Davis, 1946; Wood, 1974; Holt, 1987; Cipriano and Holt, 2005). Internally, acute septicemia has been described (Davis, 1946; Wood, 1974; Holt, 1987; Cipriano and Holt, 2005). Despite the known susceptibility of coho salmon to BCWD, their critical importance in conservation and stock enhancement efforts, and their prominent role in aquaculture (Nematollahi et al. 2003; Fujiwara-Nagata et al. 2013; Avendano-Herrera et al. 2014; Van Vliet et al. 2016; Knupp et al. 2019; FAO, 2022), few studies have examined F. psychrophilum-coho salmon interactions via injection or immersion studies. In this context, Holt (1987) successfully caused mortality in juvenile coho salmon via immersion; however, the resultant gross disease signs were not reported. Knupp et al. (2021) immersion challenged coho salmon and reported multifocal ulcerations on the caudal peduncle that eventually exposed the spinal processes. Moreover, coho salmon had focally extensive ulcerations of the rostrum, in addition to diffuse ecchymoses and petechiae of the gills and intraocular focal ecchymosis. Internally, coho salmon had organ pallor, splenic swelling, perisplenic hemorrhage, multifocal hepatic ecchymoses, and hemorrhage within the pyloric caeca and the surrounding adipose tissue. 1.2.12.3. Atlantic salmon Published reports of gross BCWD signs in Atlantic salmon are less frequent, but appear to be similar to rainbow trout. Externally, lethargy, erratic swimming or resting at the bottom of tanks, exophthalmia, swelling and/or ulceration with hemorrhage on the caudal peduncle are present (Nilsen et al. 2011b). Petechiae on the abdomen have also been reported (Nilsen et al. 34 2011b). Internally, Atlantic salmon display mottled liver, swollen and dark spleen, serohemorrhagic ascites, hemorrhage in the abdominal fat and skeletal muscle (Nilsen et al. 2011b). Although immersion challenges have used Atlantic salmon and caused mortality, gross disease signs have not always manifested or haven’t been reported (Martinez et al. 2004; Macchia et al. 2022). 35 REFERENCES Álvarez, B., Guijarro, J.A. 2007. Recovery of Flavobacterium psychrophilum viable cells using a charcoal-based solid medium. Letters in Applied Microbiology. 44(5), 569-572. Amita, K., Hoshimo, M., Honma, T., Wakabayashi, H. 2000. An investigation on distribution of Flavobacterium psychrophilum in the Umikava River. Fish Pathology. 35, 193-197. Anacker, R.L., Ordal, E.J. 1959. Studies on the myxobacterium Chondrococcus columnaris: 1. Serological typing. Journal of Bacteriology. 78, 25. Aoki, M., Kondo, M., Kawai, K., Oshima, S.I. 2005. Experimental bath infection with Flavobacterium psychrophilum, inducing typical signs of rainbow trout Oncorhynchus mykiss fry syndrome. Disease of Aquatic Organisms. 67, 73-79. Arai, H., Morita, Y., Izumi, S., Katagiri, T., Kimura, H. 2007. Molecular typing by pulsed-field gel electrophoresis of Flavobacterium psychrophilum isolates derived from Japanese fish. Journal of Fish Diseases. 30, 345-355. Auld, H.L., Noakes, D.L.G., Banks, M.A. 2019. Advancing mate choice studies in salmonids. Reviews in Fish Biology and Fisheries. 29, 249–276 Austin, B., Stobie, M. 1991. Recovery of yellow pigmented bacteria from dead and moribund fish during outbreaks of rainbow trout, Oncorhynchus mykiss (Walbaum), fry syndrome in England. Journal of Fish Diseases. 14, 677-682. Avendaño-Herrera, R., Houel, A., Irgang, R., Bernardet J-F., Godoy, M., Nicolas, P., Duchaud, E. 2014. Introduction, expansion, and coexistence of epidemic Flavobacterium psychrophilum lineages in Chilean fish farms. Veterinary Microbiology. 170, 298-306. Avendano-Herrera, R., Tapia-Cammas, D., Duchaud, E., Irgang, R. 2020. Serological diversity in Flavobacterium psychrophilum: A critical update using isolates retrieved from Chilean salmon farms. Journal of Fish Diseases, 43, 877–888. Barnes, M.E., Brown, M.L. 2011. A review of Flavobacterium psychrophilum biology, clinical signs, and bacterial coldwater disease prevention and treatment. The Open Fish Science Journal. 4, 40-48. Bernardet, J-F., Bowman, J.P. 2006. The genus Flavobacterium. In The Prokaryotes: A Handbook on the Biology of Bacteria (Dworking, M., Falkow, S., Rosenberg, E., Schleifer, K. & Stackebrandt, E., eds.), pp.481-531. New York, NY: Springer-Verlag. Bernardet, J-F. Kerouault, B. 1989. Phenotypic and genomic studies of Cytophaga psychrophila isolated from diseased rainbow trout (Oncorhynchus mykiss) in France. Applied and Environmental Microbiology. 55, 1796-1800. 36 Bernardet, J.F., Sergers, P., Vancanneyt, M., Berthe, F., Kersters, K., Vandamme, P. 1996. Cutting a Gordian Knot: emended classification and description of the genus Flavobacterium, emended description of the family Flavobacteriaceae, and proposal of Flavobacterium hydatis nom. nov. (basonym, Cytophaga aquatilis Strohl & Tait 1978). International Journal of Systematic Bacteriology. 46, 128-148. Bertolini, J.M., Wakabayashi, H., Watral, V.G., Whipple, M.J., Rohovec, J.S. 1994. Electrophoretic detection of proteases from selected strains of Flexibacter psychrophilus and assessment of their variability. Journal of Aquatic Animal Health. 6, 224-233. Borg, A.F. 1948. Ph.D. thesis. University of Washington, Seattle, Washington. Studies on myxobacteria associated with diseases in salmonid fishes. Brenden, T.O., Ivan, L.N., Loch, T.P. 2023. Reducing Flavobacterium psychrophilum transmission risk via hatchery-rearing practices: An individual-based modeling evaluation. Aquaculture. 563, 738868. Brown, L., Cox, W., Levine, R. 1997. Evidence that the causal agent of bacterial coldwater disease Flavobacterium psychrophilum is transmitted within salmonid eggs. Diseases of Aquatic Organisms. 29(3), 213-218. Bruce, T.J., Ma, J., Jones, E.M., Vuglar, B.M., Oliver, L.P., Knupp, C., Loch, T.P., Cain, K.D. 2021. Assessment of Flavobacterium psychrophilum-associated mortality in Atlantic salmon (Salmo salar) and brook trout (Salvelinus fontinalis). Journal of Fish Diseases. 44(5), 645-653. Bruun, M.S., Schmidt, A.S., Madsen, L., Dalsgaard, I. 2000. Antimicrobial resistance patterns in Danish isolates of Flavobacterium psychrophilum. Aquaculture. 187(3-4), 201-212. Calvez, S., Navarro-Gonzalez, N., Siekoula-Nguedia, C., Fournel, C., Duchaud, E. High genetic diversity in Flavobacterium psychrophilum isolates from healthy rainbow trout (Oncorhynchus mykiss) farmed in the same watershed, revealed by two typing methods. Applied and Environmental Microbiology. 87(2): e01398-20. Cepeda, C., García-Márquez, S., Santos, Y. 2004. Improved growth of Flavobacterium psychrophilum using a new culture medium. Aquaculture. 238, 75-82. Cipriano, R.C., Schill, W.B., Teska, J.D., Ford, L.A. 1996. Epizootiological study of bacterial cold-water disease in Pacific salmon and further characterization of the etiologic agent, Flexibacter psychrophila. Journal of Aquatic Animal Health. 8, 28-36. Cipriano, R.C., Holt, R.A. 2005. Flavobacterium psychrophilum, cause of bacterial cold-water disease and rainbow trout fry syndrome Fish Disease Leaflet, no. 86. United States Department of the Interior. U.S. Geological Service, National Fish Health Research Laboratory, Kearneysville, WV. 37 Chakroun, C., Urdaci, M.C., Faure, D., Grimont, F., Bernardet, J-F. 1997. Random amplified polymorphic DNA analysis provides rapid differentiation among isolates of the fish pathogen Flavobacterium psychrophilum and among Flavobacterium species. Diseases of Aquatic Organisms. 31, 187-196. Chakroun, C., Grimont, F., Urdaci, M.C., Bernardet, J-F. 1998. Fingerprinting of Flavobacterium psychrophilum isolates by ribotyping and plasmid profiling. Diseases of Aquatic Organisms. 33, 167-177. Christiansen, R.H., Dalsgaard, I., Middelboe, M., Lauritsen, A.H., Madsen, L. 2014. Detection and quantification of Flavobacterium psychrophilum-specific bacteriophages in vivo in rainbow trout upon oral administration: implications for disease control in aquaculture. Applied and Environmental Microbiology. 80(24), 7683-7693. Crump, E.M., Perry, M.B., Gale, S., Crawford, E., Kay, W.W. 2003. Lipopolysaccharide O- antigen antibody-based detection of the fish pathogen Flavobacterium psychrophilum. Journal of Molecular Microbiology and Biotechnology. 6, 182–190. Dalsgaard, I., Madsen, L. 2000. Bacterial pathogens in rainbow trout, Oncorhynchus mykiss (Walbaum), reared at Danish freshwater farms. Journal of Fish Diseases. 23(3), 199-209. Dash, S., Das, S.K., Samal, J., Thatoi, H.N. 2018. Epidermal mucus, a major determinant in fish health: a review. Iranian Journal of Veterinary Research. 19, 72-81. Daskalov, H., Austin, D.A., Austin, B. 1999. An improved growth medium for Flavobacterium psychrophilum. Letters in Applied Microbiology. 28, 297-299. Davis, H.S. 1946. Care and diseases of trout. United States Fish and Wildlife Service Research Report 12. (98pp). Decostere, A., D’Haese, E., Lammens, M., Nelis, H., Haesebrouck, F. 2001. In vivo study of phagocytosis, intracellular survival and multiplication of Flavobacterium psychrophilum in rainbow trout, Oncorhynchus mykiss (Walbaum), spleen phagocytes. Journal of Fish Diseases. 24, 481–487. del Cerro, A., Marquez, I., Prieto, J.M. 2010. Genetic diversity and antimicrobial resistance of Flavobacterium psychrophilum isolated from cultured rainbow trout, Oncorhynchus mykiss (Walbaum), in Spain. Journal of Fish Diseases. 33, 285-291. Deng, F., Wang, D., Loch, T.P., Chen, F., Lu, T., Cao, Y., Fan, D., Li, S. 2022. Time-course transcriptome analyses of spleen in rainbow trout (Oncorhynchus mykiss) post-Flavobacterium psychrophilum infection. Frontiers in Immunology. 38 Duchaud, E., Rochat, T., Habib, C., Barbier, P., Loux, V., Guerin, C., Dalsgaard, I., Madsen, L., Nilsen, H., Sundell, K., Wiklund, T., Strepparava, N., Wahli, T., Caburlotto, G., Manfrin, A., Wiens, G.D., Fujiwara-Nagata, E., Avendaño-Herrera, R., Bernardet, J-F., Nicolas, P. 2018. Genomic diversity and evolution of the fish pathogen Flavobacterium psychrophilum. Frontiers in Microbiology. 9, 138. Ekman, E., Borjeson, H., Johansson, N. 1999. Flavobacterium psychrophilum in Baltic salmon Salmo salar brood fish and their offspring. Diseases of Aquatic Organisms. 37, 159–163. Ekman, E., Norrgren, L. 2003. Pathology and immunohistochemistry in three species of salmonids after experimental infection with Flavobacterium psychrophilum. Journal of Fish Diseases. 26, 529-538. Evensen, O., Lorenzen, E. 1997. Simultaneous demonstration of infectious pancreatic necrosis virus (IPNV) and Flavobacterium psychrophilum in paraffin-embedded specimens of rainbow trout Oncorhynchus mykiss fry by use of paired immunohistochemistry. Diseases of Aquatic Organisms. 29, 227-232. Fast, M.D., Sims, D.E., Burka, J.F., Mustafa, A., Ross, N.W. 2002. Skin morphology and humoral non-specific defense parameters of mucus and plasma in rainbow trout, coho, and Atlantic salmon. Comparative Biochemistry and Physiology Part A: Molecular & Integrative Physiology. 132, 645-657. FAO. 2022. The State of World Fisheries and Aquaculture 2022. Towards Blue Transformation. Rome, FAO. Fraslin, C., Dechamp, N., Bernard, M., Krig, F., Hervet, C., Guyomard, R., Esquerre, D., Barbieri, J., Kuchly, C., Duchaud, E., Boudinot, P., Rochat, T., Bernardet, J-F., Quillet, E. 2018. Quantitative trait loci for resistance to Flavobacterium psychrophilum in rainbow trout: effect of the mode of infection and evidence of epistatic interactions. Genetics, Selection, and Evolution. 50(1), 60. Fredriksen, B.N., Furevik, A., Gauthier, D., Egenberg, M., Paulsen, E.D., Brudeseth, B. 2013. Intramuscular challenge of rainbow trout (Oncorhynchus mykiss) with two Norwegian field strains of Flavobacterium psychrophilum. Fish and Shellfish Immunology. 35, 595-598. Fredriksen, B.N., Furevik, A., Olesen, R.H., Gauthier, D., Mendoza, J., Norderhus, E.A. 2016. Virulence of Chilean field isolates of Flavobacterium psychrophilum in Atlantic salmon (Salmo salar L.) parr. Bulletin of the European Association of Fish Pathologists. 36, 67. Fujiwara-Nagata, E., Chantry-Darmon, C., Bernardet, J-F., Eguchi, M., Duchaud, E., Nicolas, P. 2013. Population structure of the fish pathogen Flavobacterium psychrophilum at whole-country and model river levels in Japan. Veterinary Research. 44, 34. 39 Garcia, C., Pozet, F., Michel, C. 2000. Standardization of experimental infection with Flavobacterium psychrophilum, the agent of rainbow trout Oncorhynchus mykiss fry syndrome. Diseases of Aquatic Organisms. 42:191-197. Gomez, E., Mendez, J., Cascales, D., Guijarro, J.A. 2014. Flavobacterium psychrophilum vaccine development: a difficult task. Microbial Biotechnology. 7(5), 414-423. Guijarro, J.A. 2001. Purification and characterization of a psychrophilic, calcium-induced, growth-phase-dependent metalloprotease from the fish pathogen Flavobacterium psychrophilum. Applied and Environmental Microbiology. 67(6), 2436-2444 Harrison, C.E., Knupp, C.K., Brenden, T.O., Ebener, M.P., Loch, T.P. 2022. First isolation of Flavobacterium psychrophilum from wild adult Great lakes lake whitefish (Coregonus clupeaformisi). Journal of Fish Diseases. 45(7), 1023-1032. Henriksen, M.M.M., Madsen, L., Dalsgaard, I. 2013. Effect of hydrogen peroxide on immersion challenge of rainbow trout fry with Flavobacterium psychrophilum. PLoS ONE. 8, e62590. Hesami, S., Allen, K.J., Metcalf, D., Ostland, V.E., MacInnes, J.I., Lumsden, J.S. 2008. Phenotypic and genotypic analysis of Flavobacterium psychrophilum isolates from Ontario salmonids with bacterial coldwater disease. Canadian Journal of Microbiology. 54, 619-629. Hoare, R., Ngo, T.P.H., Bartie, K.L., Adams, A. 2017. Efficacy of a polyvalent immersion vaccine against Flavobacterium psychrophilum and evaluation of immune response to vaccination in rainbow trout fry (Oncorhynchus mykiss L.). Veterinary Research. 48, 43. Högfors-Rönnholm, E., Wiklund, T. 2010. Phase variation in Flavobacterium psychrophilum: characterization of two distinct colony phenotypes. Diseases of Aquatic Organisms. 90, 43-53. Holt, R.A. 1987. PhD thesis. Cytophaga psychrophila, the causative agent of bacterial cold water disease in salmonid fish. Oregon State University, Corvallis, OR. Huyben, D., Jarau, M., MacInnes, J., Stevenson, R., Lumsden, J. 2023. Impact of infection with Flavobacterium psychrophilum and antimicrobial treatment on the intestinal microbiota of rainbow trout. Pathogens. 12(3), 454. Iida, Y. Mizokami, A. 1996. Outbreaks of coldwater disease in wild ayu and pale chub. Fish Pathology. 31,157–164. Izumi, S., Liu, H., Aranishi, F., Wakabayashi, H. 2003. A novel serotype of Flavobacterium psychrophilum detected using antiserum against an isolate from amago, Oncorhynchus masou rhodurus Jordan & Gilbert, in Japan. Journal of Fish Diseases. 26, 677-680. Izumi, S., Aranishi, F. 2004. Relationship between gyrA mutations and quinolone resistance in Flavobacterium psychrophilum isolates. Applied and Environmental Microbiology. 70, 3968- 3972. 40 Izumi, S., Fujii, H., Aranishi, F. 2005. Detection and identification of Flavobacterium psychrophilum from gill washings and benthic diatoms by PCR-based sequencing analysis. Journal of Fish Diseases. 28(9), 559-564. Knupp, C., Wiens, G.D., Faisal, M., Call, D.R., Cain, K.D., Nicolas, P., Van Vliet, D., Yamashita, C., Ferguson, J.A., Meuninck, D., Hsu, H-M., Baker, B.B., Shen, L., Loch, T.P. 2019. Large-scale analysis of Flavobacterium psychrophilum multilocus sequence typing genotypes indicates that both newly identified and recurrent clonal complexes are associated with disease. Applied and Environmental Microbiology. 86, e02305-18. Knupp, C., Kiupel, M., Brenden, T.O., Loch, T.P. 2021a. Host-specific preference of some Flavobacterium psychrophilum multilocus sequence typing genotypes determines their ability to cause bacterial coldwater disease in coho salmon (Oncorhynchus kisutch). Journal of Fish Diseases. 44(5), 521-531. Kum C., Kirkan S., Sekkin S., Akar F., Boyacioglu, M. 2008. Comparison of in vitro antimicrobial susceptibility in Flavobacterium psychrophilum isolated from rainbow trout fry. Journal of Aquatic Animal Health. 20, 245-51. Lee, K.B., Heo, G.J. 1998. First isolation and identification of Cytophaga psychrophila from cultured ayu Korea. Fish Pathology. 33, 37-38. Leeds, T.D., Silverstein, J.T., Weber, G.M., Vallejo, R.L., Palti, Y., Rexroad III., C.E., Evenhuis, J., Hadidi, S., Welch, T.J., Wiens, G.D. 2010. Response to selection for bacterial cold water disease resistance in rainbow trout. Journal of Animal Science. 88, 1936-1946. Lehmann J., D. Mock, F.J., Sturenberg, Bernardet, J-F. 1991. First isolation of Cytophaga psychrophila from a systemic disease in eel and cyprinids. Diseases of Aquatic Organisms. 10, 217-220. Leon, J., Avalos, R., Ponce, M. 2009. Flavobacterium psychrophilum and its pathology of alevins from the El Ingenio fish farm, Huancayo. Reviews Peru Biology. 15, 117-124. Levipan, H.A., Avendaño-Herrera, R. 2017. Different phenotypes of mature biofilm in Flavobacterium psychrophilum share a potential for virulence that differs from planktonic state. Frontiers in Cellular and Infection Microbiology. 7, 76. Li, S., Chai, J., Cao, Y., Knupp, C., Wang, D., Nicolas, P., Chen, F., Liu, H., Lu, T., Loch, T.P. 2021. Characterization and molecular epidemiological analysis of Flavobacterium psychrophilum recovered from diseased salmonids in China. Microbiology Spectrum. 9(2), e00033021. Lindstrom, N.M., Call, D.R., House, M.L., Moffitt, C.M., Cain, K.D. 2009. A quantitative enzyme-linked immunosorbent assay and filtration-based fluorescent antibody test as potential tools to screen broodstock for infection with Flavobacterium psychrophilum. Journal of Aquatic Animal Health. 21(1), 43-56. 41 Loch, T.P., Faisal, M. Flavobacterium spp. In Fish viruses and bacteria: Pathobiology and protection., Woo, P.T.K., Cipriano, R.C., Eds., CABI: 2017, pp. 211-232. Long, A., Fehringer, T.R., LaFrentz, B.R., Call, D.R., Cain, K.D. 2014. Development of a waterborne challenge model for Flavobacterium psychrophilum. FEMS Microbiology Letters. 359, 154–160. Long A., Fehringer T.R., Swain M.A., LaFrentz B.R., Call, D.R., Cain, K.D. 2013. Enhanced efficacy of an attenuated Flavobacterium psychrophilum strain cultured under iron-limited conditions. Fish Shellfish Immunology. 35, 1477–1482. Lorenzen, E., Dalsgaard, I., From, J., Hansen, E.M., Horlyck, V., Korsholm, H., Mellergaard, S., Olesen, N.J. 1991. Preliminary investigation of fry mortality syndrome in rainbow trout. Bulletin European Association of Fish Pathologists. 11, 77-79. Lorenzen, E., Karas, N. 1992. Detection of Flexibacter psychrophilus by immunofluorescence in fish suffering from fry mortality syndrome: a rapid diagnostic method. Diseases of Aquatic Organisms. 13, 231-234. Lorenzen, E., Olesen, N.J. 1997. Characterization of isolates of Flavobacterium psychrophilum associated with coldwater disease or rainbow trout fry syndrome II: serological studies. Diseases of Aquatic Organisms. 31, 209-220. Ma, J., Bruce, T.J., Sudheesh, P.S., Knupp, C., Loch, T.P., Faisal, M., Cain, K.D. 2019a. Assessment of cross-protection to heterologous strains of Flavobacterium psychrophilum following vaccination with a live-attenuated coldwater disease immersion vaccine. Journal of Fish Diseases. 42(1), 75-84. Macchia, V., Inami, M., Ramstad, A., Grammes, F., Reeve, A., Moen, T., Torgersen, T.S., Adams, A., Desbois, A.P., Hoare, R. 2022. Immersion challenge model for Flavobacterium psychrophilum infection of Atlantic salmon (Salmo salar L.) fry. Journal of Fish Diseases. 45(11), 1781-1788. Madetoja, J., Nyman, P., Wiklund, T. 2000. Flavobacterium psychrophilum, invasion into and shedding by rainbow trout Oncorhynchus mykiss. Diseases of Aquatic Organisms. 43, 27-38. Madetoja, J., Hänninen, M-N., Hirvelä-Koski, V., Dalsgaard, I., Wiklund, T. 2001. Phenotypic and genotypic characterization of Flavobacterium psychrophilum from Finnish fish farms. Journal of Fish Diseases. 24, 469–479. Madetoja, J., Dalsgaard, I., Wiklund, T. 2002. Occurrence of Flavobacterium psychrophilum in fish-farming environments. Diseases of Aquatic Organisms. 52, 109-118. Madetoja, J., Nystedt, S., Wiklund, T. 2003. Survival and virulence of Flavobacterium psychrophilum in water microcosms. FEMS Microbiology Ecology. 43(2), 217-223. 42 Madetoja, J., Wiklund, T. 2002. Detection of the fish pathogen Flavobacterium psychrophilum in water form fish farms. Systematic and Applied Microbiology. 25(2), 259-266. Madsen, L., Moller, J.D., Dalsgaard, I. 2005. Flavobacterium psychrophilum in rainbow trout, Oncorhynchus mykiss (Walbaum), hatcheries: studies on broodstock, eggs, fry, and environment. Journal of Fish Diseases. 28(1), 39-47. Madsen, L., Dalsgaard, I. 1999. Reproducible methods for experimental infection with Flavobacterium psychrophilum in rainbow trout Oncorhynchus mykiss. Diseases of Aquatic Organisms. 36, 169-199. Madsen, L., Dalsgaard, I. 2000. Comparative studies of Danish Flavobacterium psychrophilum isolates: Ribotypes, plasmid profiles, serotypes, and virulence. Journal of Fish Diseases. 23, 211–218. Maiden, M.C., Bygraves, J.A., Feil, E., Morelli, G., Russell, J.E., Urwin, R., Zhang, Q., Zhou, J., Zurth, K., Caugant, D.A., Feavers, I.M., Actman, M., Spratt, B.G. 1998. Multilocus sequence typing: A portable approach to the identification of clones within populations of pathogenic microorganisms. Proceedings of the National Academy of Sciences of the United States of America. 95, 3140-3145. Marancik, D., Wiens, G.D. 2013. A real-time polymerase chain reaction assay for identification and quantification of Flavobacterium psychrophilum and application to disease resistance studies in selectively bred rainbow trout Oncorhynchus mykiss. FEMS Microbiology Letters. 339(2), 122-129. Martinez, J.L., Casado, A., Enriquez, R. 2004. Experimental infection of Flavobacterium psychrophilum in fins of Atlantic salmon Salmo salar revealed by scanning electron microscopy. Diseases of Aquatic Organisms. 59, 79-84. Mata, M., Skarmeta, A., and Y. Santos. 2002. A proposed serotyping system for Flavobacterium psychrophilum. Letters in Applied Microbiology. 35, 166-170. Mata, M., Santos, Y. 2001. An improved enzyme-linked immunosorbent assay (ELISA) for detection of Flavobacterium psychrophilum isolated from salmon and trout. Bulletin of the European Association of Fish Pathologists. 21, 5. Michel, C., Antonio, D., Hedrick, R.P. 1999. Production of viable cultures of Flavobacterium psychrophilum: approach and control. Research in Microbiology. 150, 351-358. Nagai, T., Nakai, T. 2011. Growth of Flavobacterium psychrophilum in fish serum correlates with pathogenicity. Journal of Fish Diseases. 34, 303–310. Nakayama, H., Tanaka, K., Teramura, N., Hattori, S. 2016. Expression of collagenase in Flavobacterium psychrophilum isolated from cold-water disease-affected ayu (Plecoglossus altivelis). Bioscience, Biotechnology, and Biochemistry. 80(1), 135-144. 43 Nakagawa, Y., Yamasoto, K. Phylogenetic diversity of the genus Cytophaga revealed by 16S rRNA sequencing and menaquinone analysis. Journal of General Microbiology. 139, 1155-1161. Nematollahi, A., Decostere, A., Pasmans, F., Haesebrouck, F. 2003. Flavobacterium psychrophilum infections in salmonid fish. Journal of Fish Diseases. 26, 563-574. Ngo, T.P.H., Bartie, K.L., Thompson, K.D., Verner-Jeffreys, D.W., Hoare, R., Adams, A. 2017. Genetic and serological diversity of Flavobacterium psychrophilum isolates form salmonids in United Kingdom. Veterinary Microbiology. 201, 216-224. Ngo, T.P.H., Smith, P., Bartie, K.L., Thompson, K.D., Verner-Jeffreys, D.W., Hoare, R., Adams, A. 2018. Antimicrobial susceptibility of Flavobacterium psychrophilum isolates from the United Kingdom. Journal of Fish Diseases. 41(2), 309-320. Nguyen, P.L., Sudheesh, P.S., Thomas, A.C., Sinnesael, M., Haman, K., Cain, K.D. 2018. Rapid detection and monitoring of Flavobacterium psychrophilum in water by using a handheld, field- portable quantitative PCR system. Journal of Aquatic Animal Health. 30(4), 302-311. Nicolas, P., Mondot, S., Achaz, G., Bouchenot, C., Bernardet, J-F., Duchaud, E. 2008. Population structure of the fish-pathogenic bacterium Flavobacterium psychrophilum. Applied and Environmental Microbiology. 74, 3701-3709. Nilsen, H., Olsen, A. B., Vaagnes, ø., Hellberg, H., Bottolfsen, K., Skjelstad, H., Colquhoun, D. J. 2011a. Systemic Flavobacterium psychrophilum infection in rainbow trout, Oncorhynchus mykiss (Walbaum), farmed in fresh and brackish water in Norway. Journal of Fish Diseases. 34, 403-408. Nilsen, H., Johansen, R., Colquhoun, D.J., Kaada, I., Bottolfsen, K., Vagnes, O., Olsen, A.B. 2011b. Flavobacterium psychrophilum associated septicaemia and necrotic myositis in Atlantic salmon Salmo salar: a case report. Disease of Aquatic Organisms. 97(1), 37-46. Nilsen, H., Sundell, K., Duchaud, E., Nicolas, P., Dalsgaard, I., Madsen, L., Aspan, A., Jansson, E., Colquhoun, D.J., Wiklund, T. 2014. Multilocus sequence typing identifies epidemic clones of Flavobacterium psychrophilum in Nordic countries. Applied and Environmental Microbiology. 80, 2728 –2736. Nilz, J., Schaumburg, M., Eskens, U., Hamann, H.P., Zschock, M. 2009. Flavobacterium psychrophilum in hessian fish farms. Tieraerztliche Umschau. 64, 191-199. Obach, A., Baudin-Laurencin, F. 1991. Vaccination of rainbow trout (Oncorhynchus mykiss) against the visceral form of cold-water disease. Diseases of Aquatic Organisms. 12, 13-15. Oplinger, R. W., Wagner, E.J. (2012). Effects of media ingredient substitution and comparison of growth of Flavobacterium psychrophilum among four media. Journal of Aquatic Animal Health. 24, 49-57. 44 Otis, E.J. 1984. Lesions of coldwater disease in steelhead trout (Salmo gairdneri): the role of Cytophaga psychrophila extracellular products. M.S. Thesis, University of Rhode Island, Kingston, Rhode Island. Pacha, R.E. 1968. Characteristics of Cytophaga psychrophila (Borg) isolated during outbreaks of bacterial cold-water disease. Applied Microbiology. 16, 97-101. Perez-Pascual, D., Rochat, T., Kerouault, B., Gomez, E., Neulat-Ripoll, F., Henry, C., Quillet, E., Guijarro, J.A., Bernardet, J-F., Duchaud, E. More than gliding: involvement of GldD and GldG in the virulence of Flavobacterium psychrophilum. Frontiers in Microbiology. 8. Rangdale, R.E., Richards, R.H., Alderman, D.J. 1996. Isolation of Cytophaga psychrophila, causal agent of rainbow trout fry syndrome (RTFS) from reproductive fluids and egg surfaces of rainbow trout (Oncorhynchus mykiss). Bulletin of the European Association of Fish Pathologists. 16, 97-101. Rochat, T., Fujiwara-Nagata, E., Calvez, S., Dalsgaard, I., Madsen, L., Calteau, A., Lunazzi, A., Nicolas, P., Wiklund, T., Bernardet, J-F., Duchaud, E. 2017. Genomic characterization of Flavobacterium psychrophilum serotypes and development of a multiplex PCR-based serotyping scheme. Frontiers in Microbiology. 8, 1752. Rochat, T., Pérez-Pascual, Nilsen, H., Carpentier, M., Bridel, S., Bernardet, J-F, Duchaud, E. 2019. Identification of a Novel Elastin-Degrading Enzyme from the Fish Pathogen Flavobacterium psychrophilum. Applied and Environmental Microbiology. 85, e02535-18. Rucker, R.R., Earp, B.J., Ordal, E.J. 1953. Infectious diseases of Pacific salmon. Transactions of the American Fisheries Society. 83, 297-312. Salazar, S., Oliver, C., Yanez, A.J., Avendano-Herrera, R. 2016. Comparative analysis of innate immune responses to Streptococcus phocae strains in Atlantic salmon (Salmo salar) and rainbow trout (Oncorhynchus mykiss). Fish and Shellfish Immunology. 51, 97-103. Sarti, M., Giorgetti, G., Manfrin, A. 1992. Method for rapid diagnosis of visceral myxobacteriosis in reared trout in Italy. Bulletin of the European Association of Fish Pathologists. 12, 209-210. Saticioglu, I.B., Duman, M., Smith, P., Wiklund, T., Altun, S. 2019. Antimicrobial resistance and resistance genes in Flavobacterium psychrophilum isolates from Turkey. Aquaculture. 512, 734293. Schachte, J.H. 1983. Coldwater disease. Pages 193-197 in F.P. Meyer, J.W. Warren, and T.G. Carey, editors. A Guide to Integrated Fish Health Management in the Great Lakes Basin. Special Publication No. 83-2, Great Lakes Fishery Commission, MI, USA. 45 Sebastiao, F., Loch, T.P., Knupp, C., Mukkatira, K., Veek, T., Richey, C., Adkison, M., Griffin, M.J., Soto, E. 2020. Multilocus sequence typing (MLST) analysis of California Flavobacterium psychrophilum reveals novel genotypes and predominance of CC-ST10 in California salmonid hatcheries. Aquaculture Research. 51, 2349-2358. Schmidt, A.S., Bruun, M.S., Dalsgaard, I., Pedersen, K., Larsen, J.L. 2000. Occurrence of antimicrobial resistance in fish pathogenic and environmental bacteria associated with 4 Danish rainbow trout farms. Applied and Environmental Microbiology. 66, 4908-4915. Schmidtke, L.M., Carson, J. 1995. Characteristics of Flexibacter psychrophilus isolated from Atlantic salmon in Australia. Diseases of Aquatic Organisms. 21, 157-161. Schulz, C., Faisal, M. 2010. The bacterial community associated with the leech Myzobdella lugubris Leidy 1851 (Hirudinea:Piscicolidae) from Lake Erie, Michigan, USA. Parasite. 17(2), 113-121. Schumann, M.F. 2021. PhD thesis. Suspended solids in salmonid aquaculture: Extended insights and new approaches for control. University of Konstanz, Konstanz, Germany. Sebastiao, F., Loch, T.P., Knupp, C., Mukkatira, K., Veek, T., Richey, C., Adkison, M., Griffin, M.J., Soto, E. 2020. Multilocus sequence typing (MLST) analysis of California Flavobacterium psychrophilum reveals novel genotypes and predominance of CC-ST10 in California salmonid hatcheries. Aquaculture Research. 51, 2349-2358. Secades, P., Alvarez, B., Guijarro, J.A. 2001. Purification and characterization of a psychrophilic calcium induced, growth-phase-dependent metalloprotease from the fish pathogen Flavobacterium psychrophilum. Applied and Environmental Microbiology. 67, 2436-2444. Siekoula-Nguedia, C., Blanc, G., Duchaud, E., Calvez, S. 2012. Genetic diversity of Flavobacterium psychrophilum isolated from rainbow trout in France: predominance of a clonal complex. Veterinary Microbiology. 161, 169-178. Soderlund, R., Hakhverdyan, M., Aspan, A., Jansson, E. 2018. Genome analysis provides insights into the epidemiology of infection with Flavobacterium psychrophilum among farmed salmonid fish in Sweden. Microbial Genomics. 4(12), e000241. Soule, M., LaFrentz, S., Cain, K.D., LaPatra, S., Call, D.R. 2005. Polymorphisms in 16S rRNA genes of Flavobacterium psychrophilum correlate with elastin hydrolysis and tetracycline resistance. Diseases of Aquatic Organisms. 65, 209-216. Starliper, C.E. 2011. Bacterial coldwater disease of fishes caused by Flavobacterium psychrophilum. Journal of Advanced Research. 2, 97-108. Strepparava, N., Nicolas, P., Wahl, T., Segner, H., Petrini, O. 2013. Molecular epidemiology of Flavobacterium psychrophilum from Swiss fish farms. Diseases of Aquatic Organisms. 105, 203-210. 46 Strepparava, N., Wahli, T., Segner, H., Petrini, O. 2014. Detection and quantification of Flavobacterium psychrophilum in water and fish tissue samples by quantitative real time PCR. BMC Microbiology. 105. Sundell, K., Landor, L., Nicolas, P., Jørgensen, J., Castillo, D., Middelboe, M., Dalsgaard, I., Donati, V. L., Madsen, L., Wiklund, T. 2019. Phenotypic and genetic predictors of pathogenicity and virulence in Flavobacterium psychrophilum. Frontiers in Microbiology. 10, 1711. Sundell, K., Landor, L., Castillo, D., Middelboe, M., Wiklund, T. 2020. Bacteriophages as biocontrol agents for Flavobacterium psychrophilum biofilms and rainbow trout infections. Phage (New Rochelle). 1(4), 198-204. Sundell, K., Wiklund, T. 2015. Characteristics of epidemic and sporadic Flavobacterium psychrophilum sequence types. Aquaculture. 441, 51-56. Taylor, P.W. 2004. Detection of Flavobacterium psychrophilum in eggs and sexual fluids of pacific salmonids by a polymerase chain reaction assay: implications for vertical transmission of bacterial coldwater disease. Journal of Aquatic Animal Health. 16, 104-108. Toranzo, A.E., Barja, J.L. 1993. Fry mortality syndrome (FMS) in Spain. Isolation of the causative bacterium Flexibacter psychrophilus. Bulletin European Association of Fish Pathologists. 13, 30-32. Toyama, T., Kita-Tsukamoto, K., Wakabayashi, H. 1994. Identification of Cytophaga psychrophila by PCR targeted 16S ribosomal RNA. Fish Pathology. 29, 271-275. Valdebenito, S., Avendaño-Herrera, R. 2009. Phenotypical, serological and genetic characterization of Flavobacterium psychrophilum strains isolated from salmonids in Chile. Journal of Fish Diseases. 32, 321-333. Van Vliet, D., Loch, T.P., Faisal, M. 2015. Flavobacterium psychrophilum infections in salmonid broodstock and hatchery-propagated stocks of the Great Lakes Basin. Journal of Aquatic Animal Health. 27, 192–202. Van Vliet, D., Wiens, G.D., Loch, T.P., Nicolas, P., Faisal, M. 2016. Genetic diversity of Flavobacterium psychrophilum isolates from three Oncorhynchus spp. in the United States, as revealed by multilocus sequence typing. Applied and Environmental Microbiology. 82, 3246- 3255. Van Vliet, D., Loch T.P., Smith, P., Faisal, M. 2017. Antimicrobial susceptibilities of Flavobacterium psychrophilum isolates from the Great Lakes basin, Michigan. Microbial Drug Resistance. 23, 791-798. Vatsos, I.N., Thompson, K.D., Adams, A. 2003. Starvation of Flavobacterium psychrophilum in broth, stream water and distilled water. Diseases of Aquatic Organisms. 56, 115-126. 47 Vatsos, I.N., Thompson, K.D., Adams, A. 2006. Colonization of rainbow trout, Oncorhynchus mykiss (Walbaum), eggs by Flavobacterium psychrophilum, the causative agent of rainbow trout fry syndrome. Journal of Fish Diseases. 29, 441– 444. Urdaci, M.C., Chakroun, C., Faure, D., Bernardet, J-F. 1998. Development of a polymerase chain reaction assay for identification and detection of the fish pathogen Flavobacterium psychrophilum. Research in Microbiology. 149(7), 519-530. Wakabayashi, H., Egusa, S. 1974. Characteristics of myxobacteria associated with some freshwater fish diseases in Japan. Bulletin of the Japanese Society of Scientific Fisheries. 40(8), 751-757. Wakabayashi, H., Toyama, T., Iidia, T. 1994. A study on serotyping of Cytophaga psychrophila isolated form fishes in Japan. Fish Pathology. 29, 101-104 Wiens, G.D., Vallejo, R.L., Leeds, T.D., Palti, Y., Hadidi, S., Liu, S., Evenhuis, J.P., Welch, T.J., Rexroad III., C.E. 2013. Assessment of genetic correlation between bacterial cold water disease resistance and spleen index in a domesticated population of rainbow trout: identification of QTL on chromosome Omy19. PLoS One. 8, e75749. Wiklund, T., Madsen, L., Bruun, M.S., Dalsgaard, I. 2000. Detection of Flavobacterium psychrophilum from fish tissue and water samples by PCR amplification. Journal of Applied Microbiology. 88(2), 299-307. Walker, L., LeVine, H., Jucker, M. 2006. Koch’s postulates and infectious proteins. Acta Neuropathologica. 112(1), 1-4. Wood, J.W. 1974. Diseases of Pacific salmon, their prevention and treatment, 2nd edition. Washington Department of Fisheries, Olympia. 48 APPENDIX Table 1.1. Formulation (per 1L H2O) of Flavobacterium psychrophilum recovery medium Anacker and Ordal (AO; Anacker and Ordal, 1959) and its derivatives, as reported in the literature. Medium Medium component Unit AO mAOa mAOb mAOc mAOd mAOe mAOf mAOg EAOCah OWi Agar g 11.0 11.0 9.0 15.0 11.0 15.0 10.0 15.0 15.0 10.0 Tryptone g 0.5 5.0 0.5 0.5 2.0 0.5 5.0 5.0 5.0 Peptone g 5.0 Beef extract g 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2 Yeast extract g 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 Sodium acetate g 0.2 0.2 0.2 0.01 0.2 0.2 0.2 0.2 0.2 CaCl2 g 0.2 Fetal bovine serum % 10.0 Horse serum % 5.0 5.0 1.0 Skimmed milk % 5.0 0.2 Fish blood % 5.0 Horse blood % 5.0 Glucose g 0.5 Galactose g 0.5 Rhamnose g 0.5 Activated charcoal g 0.5 L-Tyrosine μmol 10.0 L-Phenylalanine μmol 10.0 L-Tryptophan μmol 10.0 4-Aminobenzoic acid μmol 10.0 4-Hydroxybenzoic acid μmol 10.0 2,3-Dihydroxybenzoic acid μmol 10.0 a Modified AO (mAO), reported in Bernardet and Kerouault (1989). b Reported in Daskalov et al. (1999). c Reported in Toranzo and Barja (1993). d Reported in Obach and Laurencin (1991). e Reported in Wakabayashi and Egusa (1974). f Reported in Michel et al. (1999). g Reported in Crump et al. (2001). h Enriched AO supplemented with charcoal and aromatic compounds, reported in Alvarez and Guijarro (2007). i Oplinger and Wagner (OW) medium, reported in Oplinger and Wagner (2012). 49 Table 1.2. Formulation (per 1L H2O) of Flavobacterium psychrophilum recovery medium tryptone yeast extract salts (TYES; Holt, 1987) agar and its derivatives, as reported in the literature. Medium Medium component Unit TYES mTYESa TYESb FLPAc Agar G 10.0 10.0 15.0 10.0 Tryptone G 4.0 4.0 4.0 4.0 Yeast extract G 0.4 0.4 0.5 0.4 CaCl2 G 0.5 0.5 0.5 0.2 MgSO4 G 0.5 0.5 0.5 0.5 Glucose G 0.5 Skimmed milk % Horse serum % 5.0 a Modified TYES (mTYES), reported in Knupp et al. (2021). b Reported in Madetoja et al. (2000). c Flavobacterium psychrophilum agar (FLPA), reported in Cepeda et al. (2004). Later referred to as TYESG (i.e., TYES + glucose) in Barbier et al. (2020). 50 Table 1.3. Formulation (per 1L H2O) of less frequently used (i.e., reported) Flavobacterium psychrophilum recovery media. Medium Medium component Unit Nutrient agara Shieh’sb Modified veggietonec Hsu-Shottsd Agar G Not reported 15.0 15.0 9.0 -15.0 Tryptone G 0.20 Casitone G 0.30 Veggitones soya peptone G 5.0 Beef extract G 3.0 Peptone G 5.0 5.0 Yeast extract G 0.5 0.5 0.05 Sodium acetate G 0.01 CaCl2 G 0.0067 0.2 0.03 MgSO4 G 0.3 0.5 Glucose G 1.0 2.0 KH2PO4 G 0.05 NaHCO3 G 0.05 Sodium pyruvate G 0.1 K2HPO4 G 0.1 FeSO4 G 0.001 BaCl2 G 0.1 a Reported in Secades et al. (2001). b Reported in Holt (1987). c Reported in Ngo et al. (2017). d Reported in Cipriano and Holt (2005). 51 Chapter 2: Enhanced culture media for the recovery of Flavobacterium psychrophilum, causative agent of bacterial coldwater disease and rainbow trout fry syndrome 52 2.1. Abstract Flavobacterium psychrophilum causes bacterial coldwater disease (BCWD) in salmon and trout, resulting in significant economic losses worldwide. Bacterial culture remains a gold- standard method for detecting F. psychrophilum and is a core component of many BCWD research studies; however, observations suggest some variants may require an improved recovery medium. Therefore, this study sought to develop a culture medium that enhanced the recovery of a wide diversity of F. psychrophilum multilocus sequence typing (MLST) variants. Initially, the recovery of 165 geographically, temporally, and genetically diverse F. psychrophilum isolates was compared on three published F. psychrophilum culture media (e.g., tryptone yeast extract salts agar; TYES, Oplinger and Wagner medium, and enriched Anacker and Ordal medium supplemented with charcoal and aromatic compounds). The medium recovering the most colony forming units for the greatest number of isolates (e.g., TYES) was then modified following a Plackett-Burman experimental design, in which eleven nitrogen and two salt sources were assayed for recovery effects. Five compounds (e.g., CaCl2, MgSO4, casamino acids, tryptose, and fetal bovine serum) significantly influenced F. psychrophilum recovery. Guided by these findings, two new culture media (Flavobacterium psychrophilum medium-A and -B; FPM-A and FPM-B) were formulated, both of which significantly increased recovery (e.g., >141% increase in yielded colony forming units) compared to TYES. Next, the F. psychrophilum detection capabilities of the new media were compared to TYES during surveillance at four Michigan salmonid spawning-sites. In total, 300 spawning-age salmonids belonging to four different species (e.g., Atlantic salmon, Salmo salar; Chinook salmon, Oncorhynchus tshawytscha; coho salmon, O. kisutch; and rainbow trout, O. mykiss) were sampled and overall, the new media recovered F. psychrophilum from 7.9 - 8.7% more fish than TYES. Collectively, results 53 demonstrate that FPM-A and FPM-B support the recovery of many F. psychrophilum genetic variants from cryostock, improve F. psychrophilum recovery from naturally infected fish, and will therefore serve as a resource for enhancing ongoing and future BCWD research and diagnostic efforts. 2.2. Introduction Flavobacterium psychrophilum (Family Flavobacteriaceae; Phylum Bacteroidetes), causative agent of bacterial coldwater disease (BCWD) and rainbow trout fry syndrome (RTFS), is a Gram-negative fish-pathogenic bacterium that causes considerable economic losses in hatcheries and fish farms, particularly those raising salmonids (Family Salmonidae; Loch and Faisal, 2017). The current gold standard for detecting F. psychrophilum in infected fish is via in vitro isolation on nutrient deplete culture media (Nematollahi et al. 2003; Alvarez and Guijarro, 2007; Barnes and Brown, 2011). In contrast to most molecular and serological techniques, culture-based methods provide evidence for active infections via recovery of live cells, and give diagnosticians access to isolates from disease outbreaks, enabling ancillary tests necessary for guiding outbreak-specific treatment recommendations and/or to produce bacterins. Moreover, live isolates are required for many different BCWD experiments, including those leading to the development of efficacious BCWD prevention and control strategies. For instance, live F. psychrophilum has been used to test the efficacy of fish egg disinfectants (Brown et al. 1997; Kumagai et al. 1998), antimicrobials (Bruun, et al. 2000; Van Vliet et al. 2016; Li et al. 2021), and vaccines (Gliniewicz et al. 2012; Ma et al. 2019a). However, F. psychrophilum culture-based detection methods are not without challenges. Flavobacterium psychrophilum is fastidious and relatively slow growing, with a reported generation time of ~2 hours at 15ºC (Holt, 1987). Likewise, the bacterium does not grow well on 54 commercially available culture media routinely used for growing other bacterial fish pathogens (Nematollahi et al. 2003). In this context, several culture media have been developed for the primary isolation and cultivation of F. psychrophilum (Anacker and Ordal, 1959; Bernardet and Kerouault, 1989; Bernardet and Bowman, 2006; Austin and Austin, 2007; Starliper, 2011), which are formulated with low nutrients and contain a protein source and salts. Some of the most common culture media for F. psychrophilum recovery are Cytophaga agar (Anacker and Ordal, 1959), tryptone enriched Cytophaga agar (Bernardet and Kerouault, 1989), and tryptone yeast extract salts medium (TYES; Holt, 1987); however, several studies have attempted to improve upon these formulations by adding different carbohydrates (Daskalov et al. 1999; MacLean et al. 2001; Cepeda et al. 2004; Hoare et al. 2019), proteins (Obach and Baudin-Laurencin, 1991; Lorenzen and Olesen, 1997; Daskalov et al. 1999; Michel et al. 1999; MacLean et al. 2001; Cepeda et al. 2004) and/or a detoxifier (Álvarez and Guijarro, 2007). Although some studies provided evidence these formulations improved F. psychrophilum growth in the laboratory and/or its recovery from infected fishes compared to other media (Daskalov et al. 1999; Michel et al. 1999; Cepeda et al. 2004; Álvarez and Guijarro, 2007), others found some incorporated medium components had neutral or negative effects on F. psychrophilum growth (Oplinger and Wagner, 2012). For instance, although Holt (1987) found the addition of CaCl2 and MgSO4 improved F. psychrophilum growth, Oplinger and Wagner (2012) reported these salts stunted growth. The variability in formulations and performance among different F. psychrophilum culture media may be partially caused by this species’ intraspecific diversity. Indeed, substantial genetic, serotypic, and phenotypic heterogeneity has been unearthed among different F. psychrophilum isolates (Wakabayashi et al. 1994; Lorenzen and Olesen, 1997; Madsen and 55 Dalsgaard, 1998; Madetoja et al. 2001; Nicolas et al. 2008; Högsfors-Rönnholm and Wiklund, 2010). Currently, the most widely-adopted technique for characterizing F. psychrophilum genetic diversity is multilocus sequence typing (MLST), which has been applied to >1500 isolates from 19 countries, resulting in the generation of >260 sequence types (STs; https://pubmlst.org/organisms/flavobacterium-psychrophilum). Most F. psychrophilum STs (i.e., variants) found in the U.S. are genetically distinct from those reported in other countries, and some also appear to vary in virulence, apparent host-associations, antibiotic susceptibility, and proteolytic activity (Van Vliet et al. 2016; Van Vliet et al. 2017; Sebastião et al. 2020; Li et al. 2021; Knupp et al. 2021a; Knupp et al. 2021b; Harrison et al. 2022). Likewise, ongoing studies suggest some U.S. F. psychrophilum variants may also differ in nutritional requirements, as evidenced by difficulties in primary isolation (Loch and Knupp, unpublished). In reviewing published studies that aimed to improve F. psychrophilum recovery, most studies used ≤5 isolates and found universal improvement. (Daskalov et al. 1999; Michel et al. 1999; Alvarez and Guijarro, 2007; Oplinger and Wagner, 2012). In contrast, another study using a greater number of F. psychrophilum isolates (e.g., 13) found recovery differed among isolates (Cepeda et al. 2004). Given the increasing recognition of F. psychrophilum intraspecific diversity and the inconsistencies among in vitro culture findings, this study was designed to develop a culture medium that would enhance recovery and in vitro culture of a range of F. psychrophilum variants, thereby supporting future BCWD research and diagnostic efforts. To accomplish this goal, the recovery of a large collection of diverse F. psychrophilum isolates from cryostock was initially compared on three published media, including TYES (Holt, 1987), Oplinger and Wagner medium (OW; Oplinger and Wagner, 2012), and enriched Anacker and Ordal medium 56 supplemented with activated charcoal and aromatic compounds (EAOCa; Alvarez and Guijarro, 2007), which are formulated with different proteins, salts, and other nutrients. Next, findings were used to select a basal medium and formulate two new media via a Plackett-Burman experimental design (Plackett and Burman, 1944). The newly developed culture media and the original medium were then compared for their ability to recover F. psychrophilum from cryostock and naturally infected fish at several Michigan gamete collection sites. 2.3. Materials and Methods 2.3.1. Flavobacterium psychrophilum isolates A total of 165 F. psychrophilum isolates, recovered from 1981 to 2020 and originating from 19 U.S. states, one Canadian province, and the countries of Chile and Denmark, were used in this study (Table 2.1, Table 2.2). The isolates were recovered from external and internal organs of captive or wild/feral fish belonging to five genera (within the families Salmonidae and Acipenseridae) and 12 species, including rainbow/steelhead trout (Oncorhynchus mykiss; n = 96), Chinook salmon (O. tshawytscha; n = 25), coho salmon (O. kisutch; n = 17), brown trout (Salmo trutta; n = 9), Atlantic salmon (S. salar; n = 7), lake whitefish (Coregonus clupeaformis; n = 3), brook trout (S. fontinalis; n = 2), lake trout (S. namaycush; n = 2), cutthroat trout (O. clarkii; n = 1), sockeye salmon (O. nerka; n = 1), splake (S. fontinalis x S. namaycush; n = 1), and white sturgeon (Acipenser transmontanus; n =1) at various life-stages (Table 2.1, Table 2.2). The isolates were genetically diverse according to multilocus sequence typing (MLST), whereby most (n = 153) had been previously genotyped (Van Vliet et al. 2016; Knupp et al. 2019; Ma et al. 2019a; Li et al. 2021; Harrison et al. 2022) but several (n = 12) were newly genotyped in the current study following published protocols (Knupp et al. 2019; Table 2.2). In total, isolates 57 belonged to 105 MLST sequence types (STs) that were either assigned to one of eighteen clonal complexes (CCs) or were singletons (Table 2.1, Table 2.2). 2.3.2. Basal medium experiment 2.3.2.1. Media selection A total of three previously reported media, including tryptone yeast extract salts agar (TYES; Holt, 1987), Oplinger and Wagner medium (OW; Oplinger and Wagner, 2012), and enriched Anacker and Ordal medium supplemented with activated charcoal and aromatic compounds (EAOCa; Alvarez and Guijarro, 2007) were compared for F. psychrophilum recovery [e.g., colony forming unit (cfu) yield] from cryostock. The three media (e.g., TYES, OW, and EAOCa) were selected for their widespread geographical and/or recommended use (Barnes and Brown, 2011), differing formulations (Table 2.3), and/or reported ability to improve F. psychrophilum growth in comparison to other media recommended for F. psychrophilum culture (Holt, 1987; Alvarez and Guijarro, 2007; Oplinger and Wagner, 2012). 2.3.2.2. Bacterial preparation For this experiment, 165 cryopreserved F. psychrophilum isolates were revived onto a medium formulated herein that would not favor one medium over the others (termed “neutral medium”), which consisted of (per liter): 10 g bacteriological agar (ThermoFisher Scientific), 4.5 g tryptone (ThermoFisher Scientific), and 0.45 g yeast extract (ThermoFisher Scientific; Table 2.3). Following inoculation of cryostock onto the neutral medium, plates were incubated at 15 ºC for 72 hours, after which cultures were visually inspected for purity. A 1-µl loopful of each isolate was inoculated into 40 mL of analogous neutral medium broth and incubated at 15 ºC with constant shaking at 180 rpm for 48 hours. Bacteria were harvested from broth via centrifugation (2,571 × g, 10 min) and then adjusted to a standardized optical density at 600-nm 58 (OD600 = 2.0) using a Biowave CO8000 Cell Density Meter (i.e., spectrophotometer; WPA Inc.) and sterile 0.65% saline. To compare F. psychrophilum recovery, serial dilutions in 10-fold increments (diluted up to 100,000,000-fold) were plated using 10-µl drops on TYES, OW, and EAOCa, in duplicate, and then incubated at 15 ºC for seven days, after which final colony counts were performed. 2.3.2.3. Data analysis The goal of this experiment was to determine which medium (e.g., TYES, OW, or EAOCa) would be used as the base formulation for the Plackett-Burman experiment. To do so, the medium that recovered the most cfus of F. psychrophilum for the greatest number of isolates was selected. In this context, a linear mixed model was used to quantify the effect (log10 cfu yield) of each medium on the 165 F. psychrophilum isolates. The model included medium, isolate, and the interaction between medium and isolate as fixed effects. Replicates nested within isolates were treated as random effects to account for the variability among replicates within each isolate and allow for the inference about the variability of the response variable (e.g., log10 cfu yield) for different isolates not specifically measured in this study. Degrees of freedom for fixed effects were calculated using the Kenward-Roger method. Custom hypothesis tests as to differences in overall mean log10 cfu yield between media (e.g., EAOCa vs. OW, EAOCa vs. TYES, and EAOCa vs. OW) and between media for each isolate, were evaluated through pairwise comparisons of least-square means and adjusted for multiple comparisons using the Tukey-Kramer method (α = 0.05). Analyses were performed using PROC MIXED in SAS® Version 9.4; custom hypothesis testing was performed using the LSMEANS statement and pdiff option. 59 2.3.3. Plackett-Burman experiment 2.3.3.1. Experimental design A Plackett-Burman experimental design (Plackett and Burman, 1944) was used to screen 11 independent variables (i.e., medium components) at a low (-1) and high (+1) concentration for their main effect on F. psychrophilum recovery (Table 2.4, Table 2.5). Of the 11 tested medium components, eight were nitrogen sources, as metabolic reconstruction of F. psychrophilum following whole-genome sequencing indicated that proteins and amino acids are likely the main energy sources for F. psychrophilum (Duchaud et al. 2007). In this context, tryptone, yeast extract, beef extract, casamino acids, tryptose, L-aspartic acid, horse serum, and fetal bovine serum (FBS) were evaluated (Table 2.4). Additionally, two salts (e.g., calcium chloride dihydrate, CaCl2 and magnesium sulfate heptahydrate, MgSO4) that were reported to improve F. psychrophilum growth (Holt, 1987), and rainbow trout blood, which is a source of multiple host factors, including iron, an essential element for bacterial growth and virulence (Ratledge and Dover, 2000; Table 2.4), were also evaluated. All medium components were obtained from ThermoFisher Scientific or Millipore Sigma as detailed in Table 2.4, except rainbow trout blood, which was obtained from yearling rainbow trout reared under quarantine at the Michigan State University (MSU) – University Research Containment Facility in accordance with the MSU – Institutional Animal Care and Use Committee (AUF:201900312). Briefly, rainbow trout were euthanized with an overdose (250 mg/L) of sodium bicarbonate-buffered tricaine methanesulfonate (MS-222; Syndel) and then blood was collected by caudal venipuncture, which was immediately combined with an equal volume of Alsever’s solution (Millipore Sigma) and transported on ice for immediate use. In total, 12 media types were formulated according to the Plackett-Burman design matrix (Table 2.5), whereby each media type incorporated six medium 60 components at the high concentration and five at the low concentration, except for the first medium, which incorporated all medium components at the low concentration. Across the 12 media types, each medium component was incorporated six times at the high and low concentrations (see columns in Table 2.5). Medium component concentrations were selected to be at least 25% below and 50% above those used in EAOCa, OW, and TYES, while medium components not used EAOCa, OW, and TYES were modeled similarly. 2.3.3.2. Media preparation Each media type was prepared identically as follows: bacteriological agar, CaCl2 and MgSO4 were weighed and added to ultrapure water that was subsequently brought to a boil, and then sterilized by autoclaving for 15 min at 121 ºC. Meanwhile, a proteinaceous solution composed of tryptone, yeast extract, beef extract, casamino acids, tryptose, and L-aspartic acid was prepared and adjusted to a pH of 7.20 ± 0.01 using potassium hydroxide. The proteinaceous solution was then filter sterilized using a 0.22-µm filter flask with a polyethersulfone filter membrane (Santa Cruz Biotechnology, Dallas, TX, USA) to prevent protein denaturation that occurs with autoclave sterilization (Taha and Mohamed, 2003). The sterilized proteinaceous solution, horse serum, FBS, and rainbow trout blood were added to the autoclaved solution (cooled to 55 ºC), and then continuously mixed via a stir plate while 24mL of media was dispensed into Petri dishes (100-mm in diameter x 15-mm in height; VWR, Radnor, PA, USA). Once the media solidified, the petri dishes were immediately stored at 4 ºC and used within seven days. 61 2.3.3.3. Bacteria preparation For this experiment, 50 cryo-preserved F. psychrophilum isolates representing 49 different MLST STs (Table 2.2) were revived, cultured in broth, adjusted to a standardized OD600, and then inoculated onto the 12 media types as described in section 2.3.2.2. 2.3.3.4. Data analysis The Plackett-Burman experimental design consisted of 11 factors (i.e., medium components), 12 runs (i.e., media types), and 50 replicates (i.e., F. psychrophilum isolates), equating to 600 total runs. The effect of each medium component was calculated with the following equation: 2(∑ 𝐶!" − 𝐶!# ) 𝐸(𝑥! ) = 𝑁 where E(xi) is the concentration effect of the tested medium component, 𝐶!" and 𝐶!# are the F. psychrophilum cfus recovered from the runs where the medium component (xi) was at the high and low concentrations, respectively, and N is the number of runs (600). The significance of each medium component’s concentration was determined using a one-sample, two-tailed Student’s t test (H0 = 0, a = 0.05): 𝐸(𝑥! ) 𝑡(𝑥! ) = 𝑆𝐸 where E(xi) is the concentration effect of the tested medium component xi. The development of the Plackett-Burman experimental design and generation of the normal plot of the standardized effects and Pareto chart was completed in Minitab 21 Trial (Minitab Inc., USA). 62 2.3.4. Comparison of newly formulated media to TYES 2.3.4.1. Formulation of new media Based upon results from the Plackett-Burman experiment, two new media types (e.g., F. psychrophilum medium-A, FPM-A; F. psychrophilum medium-B, FPM-B), were formulated and then compared to TYES for F. psychrophilum growth. Flavobacterium psychrophilum medium- A consisted of (per L): 0.75 g tryptone, 0.075 g yeast extract, 0.75 g CaCl2, 1.11 g MgSO4, 0.075 g beef extract, 0.75 g L-aspartic acid, and 7.5 mL horse serum. Flavobacterium psychrophilum-B consisted of (per L): 0.75 g Tryptone, 0.075 g yeast extract, 0.75 g CaCl2, 1.11 g MgSO4, 0.075 g beef extract, 0.75 g casamino acids, 0.75 g tryptose, 0.75 g L-aspartic acid, 7.5 mL horse serum, and 7.5 mL FBS. Both FPM-A and FPM-B were prepared as described in section 2.3.3.2. 2.3.4.2. Bacterial preparation In this experiment, 165 F. psychrophilum isolates (Table 2.1, Table 2.2) were revived, cultured in broth, adjusted to a standardized OD600, and then inoculated onto TYES, FPM-A, and FPM-B as described in section 2.3.2.2. 2.3.4.3. Data analysis A linear mixed model was used to quantify the effect (cfu yield) of each medium on the 165 F. psychrophilum isolates. The model included medium, isolate, and the interaction between medium and isolate as fixed effects. Replicates nested within isolates were treated as random effects to account for the variability among replicates within each isolate and allows for the inference about the variability of the response variable (e.g., cfu yield) for different isolates not specifically measured in this study. Degrees of freedom for fixed effects were calculated using the Kenward-Roger method. Custom hypothesis tests as to differences in overall mean cfu yield among media (e.g., FPM-A vs. TYES, FPM-B vs. TYES, and FPM-A vs. FPM-B) were 63 evaluated through pairwise comparisons of least-square means and adjusted for multiple comparisons using the Tukey-Kramer method (α = 0.05). Analyses were performed using PROC MIXED in SAS® Version 9.4; custom hypothesis testing was performed using the LSMEANS statement and pdiff option. In addition, overall percent change in F. psychrophilum cfus was calculated for each of the three comparisons using the following equation: 𝑀𝑒𝑎𝑛 𝑐𝑓𝑢𝑠 𝑜𝑛 𝑀𝑒𝑑𝑖𝑢𝑚 𝑋 − 𝑀𝑒𝑎𝑛 𝑐𝑓𝑢𝑠 𝑜𝑛 𝑀𝑒𝑑𝑖𝑢𝑚 𝑌 𝑃𝑒𝑟𝑐𝑒𝑛𝑡 𝑐ℎ𝑎𝑛𝑔𝑒 = , 8 ∗ 100 𝑀𝑒𝑎𝑛 𝑐𝑓𝑢𝑠 𝑜𝑛 𝑀𝑒𝑑𝑖𝑢𝑚 𝑋 All computations were performed in SAS® Version 9.4. 2.3.5. Comparison of new media to TYES for the recovery of Flavobacterium psychrophilum from naturally infected fish 2.3.5.1. Flavobacterium psychrophilum surveillance in Michigan salmonid broodstock To compare the F. psychrophilum detection capabilities of FPM-A, FPM-B, and TYES, each media type was used in the fall of 2021 to test for F. psychrophilum infections at four salmonid spawning sites (i.e., weirs) in Michigan (Figure 2.1), where systemic infections are prevalent in most years (Van Vliet et al. 2015; Knupp et al. 2019). Spawning phase Chinook salmon (n = 30 males and n = 30 females per site) were collected at the Little Manistee River Weir (LMRW) and Swan River Weir (SRW), whereas coho salmon and Atlantic salmon (n = 30 males and n = 30 females per species) were obtained at Platte River Weir (PRW) and St. Mary’s River (SMR), respectively. Similarly, in the spring of 2022, FPM-A was used alongside TYES at LMRW, to test for F. psychrophilum infections in 30 male and 30 female spawning-phase steelhead trout that were captured. All fish were collected and processed for bacterial isolation as previously described (Loch et al. 2012; Van Vliet et al. 2015). Briefly, fish were euthanized on site using a pneumatic stunner (Seafood Innovations International, Queensland, Australia) and 64 then necropsied, whereby fish were surfaced disinfected with 70% ethanol and then the coelomic cavities and renal capsule were opened using sterile forceps and scissors. A sterile 10-µl loop was passed through the entire length of the kidney multiple times and then tissues streaked directly onto each media type, all of which contained 4 mg/L of neomycin sulfate (Millipore Sigma). All media types were subsequently incubated at 15 ºC for seven days, after which plates were examined for yellow-pigmented bacterial growth. All yellow-pigmented bacteria recovered on each media type were sub-cultured, and following verification of culture purity, bacterial genomic DNA was extracted, quantified, and diluted to 20 ng/µl as described previously (Knupp et al. 2019). To determine if the identity of the recovered bacterial isolate was F. psychrophilum, the conventional PCR assay of Toyama et al. (1994) was employed as detailed previously (Van Vliet et al. 2015). The proportion of fish with F. psychrophilum (i.e., infection prevalence) is reported. 2.4. Results 2.4.1. Selection of basal medium Of the three previously published and tested media (e.g., TYES, OW, and EAOCa), TYES recovered, on average, significantly more F. psychrophilum log10 cfus (e.g., 8.73 ± 0.013) than OW (e.g., 8.32 ± 0.013; t = 23.49, df = 330, P-value < 0.0001) and EAOCa (7.36 ± 0.013; t = 79, df = 330, P-value < 0.0001), and OW recovered significantly more F. psychrophilum cfus than EAOCa (e.g., 8.32 ± 0.013 vs. 7.36 ± 0.013; t = 55.51, df = 330, P-value < 0.0001). Additionally, pairwise comparisons between media types for individual F. psychrophilum isolates showed TYES recovered the most cfus for the greatest number of isolates (n = 7/165 isolates, 4.24%; P-values < 0.05; Table 2.6) or recovered a similar number of cfus (i.e., no significant difference; P-values > 0.05) as OW (n = 84/165 isolates, 50.91%), EAOCa (n = 8/165 65 isolates, 4.85%), or OW and EAOCa (n = 64/165 isolates, 38.79%). In contrast to TYES, EAOCa achieved the most growth for one isolate (e.g., US151), or OW and EAOCa recovered a similar (i.e., no significant difference) number of cfus for one isolate (e.g., US487; Table 2.6). Based on these results, TYES was selected as the medium to improve upon as guided by the Plackett-Burman experiment. 2.4.2. Plackett-Burman experiment The Plackett-Burman experiment was designed to screen 11 medium components for their effect on F. psychrophilum recovery. Of these, five medium components (e.g., CaCl2, MgSO4, casamino acids, tryptose, and FBS) had a significant effect on F. psychrophilum recovery (P-values < 0.0001), whereas six medium components did not (P-values > 0.05; Table 2.7). Normal plot of the standardized effects showed that of the five statistically significant medium components, two (e.g., CaCl2 and MgSO4) had a significant positive effect on F. psychrophilum recovery (i.e., increase of component concentration yielded more cfus), whereas the remaining three medium components (e.g., casamino acids, tryptose, and FBS) had a significant negative effect on F. psychrophilum recovery (i.e., increase of component concentration yielded less cfus; Figure 2.2). Overall, casamino acids affected F. psychrophilum recovery the most, followed by tryptose, CaCl2, MgSO4, FBS, L-aspartic acid, tryptone, beef extract, O. mykiss blood, horse serum, and yeast extract (Figure 2.3). As a result of the Plackett-Burman experiment, two new media were formulated (see section 2.3.4) and named FPM-A and FPM-B (F. psychrophilum medium-A and -B). FPM-A contained the statistically significant medium components that had a positive effect on F. psychrophilum recovery with increasing concentration (e.g., CaCl2 and MgSO4) in addition to most of the statistically insignificant medium components (e.g., tryptone, yeast extract, beef 66 extract, L-aspartic acid, and horse serum), but omitted the statistically significant medium components that had a negative effect on F. psychrophilum recovery with increasing concentration (e.g., casamino acids, tryptose, and FBS) in addition to O. mykiss blood, as it did not significantly affect F. psychrophilum recovery, has a short shelf-life, and is difficult to source. FPM-B was formulated with all tested medium components except O. mykiss blood for the same rationale stated previously. The medium components for both new media types were incorporated according to their concentration effect in Table 2.7. More specifically, tryptone, yeast extract, beef extract, casamino acids, tryptose, horse serum, and FBS were incorporated at the low concentration, whereas CaCl2, MgSO4, and L-aspartic acid were incorporated at the high concentration (Table 2.4). 2.4.3. Comparisons of Flavobacterium psychrophilum growth on FPM-A, FPM-B, and TYES Of the three media (e.g., FPM-A, FPM-B, and TYES) compared for F. psychrophilum recovery, FPM-A recovered, on average, significantly more F. psychrophilum cfus than TYES (e.g., 2.67 × 109 ± 1.10 × 108 vs. 1.07 × 109 ± 1.10 × 108; 149.5% increase; t = 10.27, df = 495, P- value < 0.0001) but not FPM-B (e.g., 2.67 × 109 ± 1.10 × 108 vs. 2.58 × 109 ± 1.10 × 108; 3.5% increase; t = 0.59, df = 495, P-value = 0.8236). Similarly, FPM-B recovered significantly more F. psychrophilum cfus than TYES (2.58 × 109 ± 1.10 × 108 vs. 1.07 × 109 ± 1.10 × 108; 141.1% increase; t = 9.67, df = 495, P-value < 0.0001). 2.4.4. FPM-A, FPM-B, and TYES Flavobacterium psychrophilum detection comparison Flavobacterium psychrophilum was recovered from each of the four sampling locations and salmonid species (Figure 2.4), where Chinook salmon from SRW had the lowest infection prevalence (0 - 1.7%) and Atlantic salmon from SMR had the highest (73.3 - 86.7%; Figure 2.4). 67 When comparing the three media across the four sampling events during which they were used simultaneously, FPM-A recovered F. psychrophilum at the highest prevalence overall (86/240 fish, 35.8%), followed by FPM-B (84/240 fish, 35.0%) and then TYES (65/240 fish, 27.1%). When comparing systemic F. psychrophilum infection prevalence as detected on each medium and at each individual site, prevalence ranged from 11.7% with TYES to 25.0% with FPM-A at LMRW with Chinook salmon; 73.3% (TYES) - 86.7% (FPM-B) at SMR with Atlantic salmon; 23.3% (TYES) - 35.0% (FPM-A) at PRW with coho salmon, and 0.0% (TYES and FPM-A) - 1.7% (FPM-B) at SRW with Chinook salmon (Figure 2.4). For steelhead collected from the LMRW, F. psychrophilum infection prevalence on FPM-A and TYES was 35.0% and 26.7%, respectively (Figure 2.4). 2.5. Discussion Although several previously published F. psychrophilum culture media exist, ongoing research, deepening insight into the potential implications of F. psychrophilum intraspecific diversity, and the need for enhanced culture media to support BCWD vaccine development collectively necessitated the need for an evaluation and comparison of contemporary culture media, with the intent of further improvement. Accordingly, two new F. psychrophilum culture media (e.g., FPM-A and FPM-B) were developed, both of which significantly increased (e.g., by 141.1 – 149.5%) F. psychrophilum recovery in the laboratory when compared to the widely used and original basal medium, TYES. Indeed, the robust performance of FPM-A and FPM-B suggests that these media will be instrumental to future BCWD research studies with a culture component, which is a prevalent feature across multiple study areas. For example, culture is used to study F. psychrophilum virulence mechanisms, such as proteolytic activity, biofilm, and motility (Levipan and Avendaño-Herrera, 2017; Perez-Pascual et al. 2017; Rochat et al. 2019), 68 and host-pathogen interactions including virulence, host specificity, comorbidity, and transmission (Madetoja et al. 2000; Ma et al. 2019b; Knupp et al. 2021a; Knupp et al. 2021b; Li et al. 2021; Bruce et al., 2021). Likewise, culture is necessary for the development and testing of BCWD prevention (e.g., vaccines, phage-therapy; ultraviolet light susceptibility; Christiansen et al. 2014; Ma et al. 2019a; Donati et al. 2021; Knupp et al. 2023, in press) and control (e.g., antimicrobial susceptibility; Miranda et al. 2016; Saticioglu et al. 2019; Sebastiao et al. 2020) strategies. Thus, FPM-A and FPM-B are promising tools for improving the efficiency of many important facets of BCWD research designed to reduce economic losses. Another important outcome of this study is the increased recovery of F. psychrophilum from naturally infected wild/feral salmonid broodstock populations on the new media when compared to TYES. Notably, the recovery of non-target bacteria was minimal (data not shown), although whether this reflects the specificity of the new media, or the infection status of the host requires further investigation. The improved detection capability not only underscores the usefulness of the new media under field conditions, but also promises more effective F. psychrophilum surveillance in these populations, which can guide the development of improved management strategies potentially leading to healthier broodstock and their progeny. Likewise, and although not studied herein, early detection in captive salmonid broodstock or their progeny could also prevent future losses. Similarly, if F. psychrophilum infections go undetected during BCWD epizootics, treatments may be delayed subsequently resulting in avoidable losses. In this context, FPM-A and TYES were deployed side-by-side to detect F. psychrophilum in hatchery- reared Atlantic salmon with gross signs of BCWD (data not shown). Consequently, only FPM-A recovered F. psychrophilum, which led to management and treatment actions that prevented additional losses. An additional outcome of the improved F. psychrophilum recovery from 69 naturally infected fish is that these new media provided isolates that will aid in deepening our understanding of F. psychrophilum molecular and serological diversity, the results of which can continue to inform BCWD vaccine development and testing (Hoare et al. 2017; Ma et al. 2019a). Collectively, results provide evidence that the new media are effective at detecting F. psychrophilum from a range of naturally infected salmonids and thus will serve as a critical resource for fisheries personnel and diagnosticians alike. The improved recovery of F. psychrophilum by FPM-A and FPM-B likely resulted from the inclusion of essential medium components at favorable concentrations. For instance, calcium has been previously shown to contribute to several microbial cellular processes, including growth (Herbaud et al. 1998; Wang et al. 2019). In the present study, calcium was found to significantly enhance bacterial recovery with increasing concentration, which may have resulted from increased metalloprotease activity, considering some F. psychrophilum metalloproteases (e.g., Fpp1) are calcium-dependent (Secades et al. 2001). Like calcium, magnesium also significantly improved F. psychrophilum recovery and although results imply a role in cell growth, its function in F. psychrophilum cellular processes remains to be determined. In contrast to the improvements made by calcium and magnesium and despite representing primary F. psychrophilum energy sources (e.g., amino acids/proteins; Duchaud et al. 2007), tryptose, casamino acids, and FBS negatively impacted F. psychrophilum recovery. In view of this, tryptose may be associated with a higher metabolic cost, considering this product is comprised of mostly polypeptides (i.e., long amino acid chains with many peptide bonds). Casamino acids, while mostly free amino acids (i.e., a seemingly low metabolic cost), also contains low levels of cystine, maltose, iron, and sodium chloride, some, or all of which may have reduced F. psychrophilum recovery. Lastly, FBS, which has been used to enrich F. psychrophilum culture 70 media (Obach & Baudin-Laurencin, 1991; Lorenzen, 1993; Michel et al. 1999), appears to have an inhibitory effect on this species under some conditions, possibly due to a deficiency of essential nutrients or presence of inhibitory host factors. In this context, future studies assessing F. psychrophilum metabolism may lead to further culture media improvements. To develop FPM-A and FPM-B, F. psychrophilum recovery was originally compared between three previously published F. psychrophilum culture media (e.g., EAOCa, OW, and TYES). This comparison revealed that TYES outperformed both EAOCa and OW, yielding the most F. psychrophilum cfus overall. However, and despite TYES being the best basal medium, EAOCa and/or OW collectively recovered two F. psychrophilum isolates (e.g., US151 and US487) at significantly higher yields compared to TYES. Similarly, FPM-A and FPM-B outperformed TYES in terms of overall cfu yield; however, 11.5 – 13.9% of isolates (in 26 STs and recovered from eight host species) yielded more cfus on TYES, although these yields were not significantly greater (data not shown). Although current observations suggest that the new media are not selective for specific F. psychrophilum genetic variants or host species, further investigation with a larger sample size per variant and host is warranted to validate this finding. Collectively, these observations not only demonstrate that F. psychrophilum diversity affects its recovery but also underscores the importance of considering a wide array of isolates in future studies. Indeed, incorporating such diversity into study design is crucial for ensuring robust results and a comprehensive understanding of F. psychrophilum behavior. This study employed a Plackett-Burman experimental design (Plackett and Burman, 1944), an approach to culture medium development that has been used to improve bacterial recovery (Stevens, 1995; Bhattacharjee and Joshi, 2016), increase bacterial biomass in vitro (Waśko et al. 2010), and enhance the production of bacterial products for biotechnological 71 purposes (Zeinab et al. 2015; Ekpenyong et al. 2017; El-Shanshoury et al. 2018). However, and prior to this study, a Plackett-Burman design approach had not been used for the development of F. psychrophilum culture media. When developing the OW medium, Oplinger and Wagner (2012) began with a single base medium and took an iterative approach to optimization by adding or subtracting a few medium components, keeping those that improved F. psychrophilum growth. In contrast, Cepeda et al. (2004) and Álvarez and Guijarro (2007) both began with two or three basal media, added one or more medium components at a single concentration, and then compared F. psychrophilum growth and/or recovery. Similarly, Michel et al. (1999) began with two basal media and added three medium components at one or more concentrations and in different combinations. The Plackett-Burman approach to culture medium development used herein is an improvement over previous methods as it allows for the simultaneous screening of many medium components at two concentrations and is inherently designed so that the concentration effect of each component can be assessed, which ultimately guided the development of the two new and improved F. psychrophilum media. In conclusion, BCWD and RTFS, although not regulatory diseases, remain substantial impediments to the production and health of salmonids being raised for food and/or conservation purposes, and bacterial culture remains the gold-standard for its detection and diagnosis. Crucially, culture-derived outbreak isolates not only guide the immediate treatment strategies to minimize losses but also serve as invaluable resources for future research targeting the development of BCWD prevention and control strategies. Herein, two new culture media (e.g., FPM-A and FPM-B) were developed, both of which not only yielded significantly more F. psychrophilum for a wide diversity of isolates, but were also more capable of recovering F. psychrophilum from multiple naturally infected salmonid species in comparison to the current 72 gold standard medium, TYES. Thus, these media set the foundation for more effective BCWD detection and provide an improved platform for subsequent studies, ultimately paving the way towards more robust prevention and control measures against this devastating fish pathogen. 73 REFERENCES Álvarez, B., Guijarro, J.A. 2007. Recovery of Flavobacterium psychrophilum viable cells using a charcoal-based solid medium. Letters in Applied Microbiology. 44(5), 569-572. Anacker, R.L., Ordal, E.J. 1959. Studies on the myxobacterium Chondrococcus columnaris: 1. Serological typing. Journal of Bacteriology. 78, 25. Barnes, M.E., Brown, M.L. 2011. A review of Flavobacterium psychrophilum biology, clinical signs, and bacterial coldwater disease prevention and treatment. The Open Fish Science Journal. 4, 40-48. Bernardet, J-F., Bowman, J.P. 2006. The genus Flavobacterium. In The Prokaryotes: A Handbook on the Biology of Bacteria (Dworking, M., Falkow, S., Rosenberg, E., Schleifer, K. & Stackebrandt, E., eds.), pp.481-531. New York, NY: Springer-Verlag. Bernardet, J-F. Kerouault, B. 1989. Phenotypic and genomic studies of Cytophaga psychrophila isolated from diseased rainbow trout (Oncorhynchus mykiss) in France. Applied and Environmental Microbiology. 55, 1796-1800. Bhattacharjee, K., Joshi S.R. 2016. A selective medium for recovery and enumeration of endolithic bacteria. Journal of Microbiological Methods. 129, 44-54. Brown, L., Cox, W., Levine, R. 1997. Evidence that the causal agent of bacterial coldwater disease Flavobacterium psychrophilum is transmitted within salmonid eggs. Diseases of Aquatic Organisms. 29(3), 213-218. Bruce, T.J., Ma, J., Jones, E.M., Vuglar, B.M., Oliver, L.P., Knupp, C., Loch, T.P., Cain, K.D. 2021. Assessment of Flavobacterium psychrophilum-associated mortality in Atlantic salmon (Salmo salar) and brook trout (Salvelinus fontinalis). Journal of Fish Diseases. 44(5), 645-653. Bruun, M.S., Schmidt, A.S., Madsen, L., Dalsgaard, I. 2000. Antimicrobial resistance patterns in Danish isolates of Flavobacterium psychrophilum. Aquaculture. 187(3-4), 201-212. Cepeda, C., García-Márquez, S., Santos, Y. 2004. Improved growth of Flavobacterium psychrophilum using a new culture medium. Aquaculture. 238, 75-82. Christiansen, R.H., Dalsgaard, I., Middelboe, M., Lauritsen, A.H., Madsen, L. 2014. Detection and quantification of Flavobacterium psychrophilum-specific bacteriophages in vivo in rainbow trout upon oral administration: implications for disease control in aquaculture. Applied and Environmental Microbiology. 80(24), 7683-7693. Daskalov, H., Austin, D.A., Austin, B. 1999. An improved growth medium for Flavobacterium psychrophilum. Letters in Applied Microbiology. 28, 297-299. 74 Donati, V.L., Dalsgaard, I., Sundell, K., Castillo, D., Er-Rafik, M., Clark, J., Wiklund, T., Middelboe, M., Madsen, L. 2021. Phage-mediated control of Flavobacterium psychrophilum in Aquaculture: in vivo experiments to compare delivery methods. Frontiers in Microbiology. 12:628309. Duchaud, E., Boussaha, M., Loux, V., Bernardet, J-F., Michel, C., Kerouault, B., Stanislas Mondot, S., Nicolas, P., Bossy, R., Caron, C., Bessières, P., Gibrat, J-F., Claverol, S., Dumetz, F., Hénaff, M., Benmansour, A. 2007. Complete genome sequence of the fish pathogen Flavobacterium psychrophilum. Nature Biotechnology. 25, 763-769. Ekpenyong, M.G., Antai, S.P., Asitok, A.D., Ekpo, B.O. 2017. Plackett-Burman design and response surface optimization of medium trace nutrients for glycolipopeptide biosurfactant production. Iranian Biomedical Journal. 21(4), 249-260. El-Shanshoury, A.E-R.R., Kenawy, E-R., Amara, A.A., Mohamed, S.F., Salama, A.F., Kishk, S. BioTechnologia. 99(3), 185-196. Gliniewicz, K., Plant, K.P., LaPatra, S.E., LaFrentz, B.R., Cain, K., Snekvik, K.R., Call, D.R. 2012. Comparative proteomic analysis of virulent and rifampicin-attenuated Flavobacterium psychrophilum. Journal of Fish Diseases. 35:529–539. Harrison, C.E., Knupp, C.K., Brenden, T.O., Ebener, M.P., Loch, T.P. 2022. First isolation of Flavobacterium psychrophilum from wild adult Great lakes lake whitefish (Coregonus clupeaformisi). Journal of Fish Diseases. 45(7), 1023-1032. Herbaud, M.L., Guiseppi, A., Denizot, F., Haiech, J., Kilhoffer, M.C. 1998. Calcium signaling in Bacillus subtilis. Biochimica Biophysica Acta. 1448(2), 212-226. Hoare, R., Ngo, T.P.H., Bartie, K.L., Adams, A. 2017. Efficacy of a polyvalent immersion vaccine against Flavobacterium psychrophilum and evaluation of immune response to vaccination in rainbow trout fry (Oncorhynchus mykiss L.). Veterinary Research. 48, 43. Hoare, R., Jung, S-J., Ngo, T.P.H., Bartie, K.L., Thompson, K.D., Adams, A. 2019. Efficacy of a polyvalent injectable vaccine against Flavobacterium psychrophilum administered to rainbow trout (Oncorhynchus mykiss L.). Journal of Fish Diseases. 42, 229-236. Högfors-Rönnholm, E., Wiklund, T. 2010. Phase variation in Flavobacterium psychrophilum: characterization of two distinct colony phenotypes. Diseases of Aquatic Organisms. 90, 43-53. Holt, R.A. 1987. PhD thesis. Cytophaga psychrophila, the causative agent of bacterial cold water disease in salmonid fish. Oregon State University, Corvallis, OR. 75 Knupp, C., Wiens, G.D., Faisal, M., Call, D.R., Cain, K.D., Nicolas, P., Van Vliet, D., Yamashita, C., Ferguson, J.A., Meuninck, D., Hsu, H-M., Baker, B.B., Shen, L., Loch, T.P. 2019. Large-scale analysis of Flavobacterium psychrophilum multilocus sequence typing genotypes indicates that both newly identified and recurrent clonal complexes are associated with disease. Applied and Environmental Microbiology. 86, e02305-18. Knupp, C., Kiupel, M., Brenden, T.O., Loch, T.P. 2021a. Host-specific preference of some Flavobacterium psychrophilum multilocus sequence typing genotypes determines their ability to cause bacterial coldwater disease in coho salmon (Oncorhynchus kisutch). Journal of Fish Diseases. 44(5), 521-531. Knupp, C., Faisal, M., Brenden, T.O., Wiens, G.D., Loch, T.P. 2021b. In vivo experiments provide evidence that Flavobacterium psychrophilum strains belonging to multilocus sequence typing clonal complex ST191 are virulent to rainbow trout (Oncorhynchus mykiss). Journal of Aquatic Animal Health. 33(3), 190-195. Knupp, C., Faisal, M., Brenden, T.O., Soto, E., LaFrentz, B.R., Griffin, M.J., Wiens, G.D., Cavender, W., Van Vliet, D., Loch, T.P. 2023. Ultraviolet light differentially reduces viability of fish- and fish-farm associated flavobacteria (Family Flavobacteriaceae and Weeksellaceae). North American Journal of Aquaculture. In press. Kumagai, A., Takahashi, S., Yamaoka, S., Wakabayashi, H. 1998. Ineffectiveness of iodophor treatment in disinfecting salmonid eggs carrying Cytophaga psychrophila. Fish Pathology. 33, 123-128. Levipan, H.A., Avendaño-Herrera, R. 2017. Different phenotypes of mature biofilm in Flavobacterium psychrophilum share a potential for virulence that differs from planktonic state. Frontiers in Cellular and Infection Microbiology. 7, 76. Li, S., Chai, J., Cao, Y., Knupp, C., Wang, D., Nicolas, P., Chen, F., Liu, H., Lu, T., Loch, T.P. 2021. Characterization and molecular epidemiological analysis of Flavobacterium psychrophilum recovered from diseased salmonids in China. Microbiology Spectrum. 9(2), e00033021. Loch, T.P., Scribner, K., Tempelman, R., Whelan, G., Faisal, M. 2012. Bacterial infections of Chinook salmon, Oncorhynchus tshawytscha (Walbaum), returning to gamete collecting weirs in Michigan. Journal of Fish Diseases. 35, 39–50. Loch, T.P., Faisal, M. Flavobacterium spp. In Fish viruses and bacteria: Pathobiology and protection., Woo, P.T.K., Cipriano, R.C., Eds., CABI: 2017, pp. 211-232. Lorenzen, E. 1993. The importance of the brand of the beef extract in relation to the growth of Flexibacter psychrophilus in Anacker and Ordals medium. Bulletin of the European Association of Fish Pathologists. 13(2), 64-65. 76 Lorenzen, E., Olesen, N.J. 1997. Characterization of isolates of Flavobacterium psychrophilum associated with coldwater disease or rainbow trout fry syndrome II: serological studies. Diseases of Aquatic Organisms. 31, 209-220. Ma, J., Bruce, T.J., Sudheesh, P.S., Knupp, C., Loch, T.P., Faisal, M., Cain, K.D. 2019a. Assessment of cross-protection to heterologous strains of Flavobacterium psychrophilum following vaccination with a live-attenuated coldwater disease immersion vaccine. Journal of Fish Diseases. 42(1), 75-84. Ma, J., Bruce, T.J., Oliver, L.P., Cain, K.D. 2019b. Co-infection of rainbow trout (Oncorhynchus mykiss) with infectious hematopoietic necrosis virus and Flavobacterium psychrophilum. Journal of Fish Diseases. 42(7), 1065-1076. MacLean, L.L., Vinogradov, E., Crump, E.M., Perry, M.B., Kay, W.W. 2001. The structure of the lipopolysaccharide O-antigen produced by Flavobacterium psychrophilum (259-93). European Journal of Biochemistry. 268, 2710-2716. Madetoja, J., Nyman, P., Wiklund, T. 2000. Flavobacterium psychrophilum, invasion into and shedding by rainbow trout Oncorhynchus mykiss. Diseases of Aquatic Organisms. 43, 27-38. Madetoja, J., Dalsgaard, I., Wiklund, T. 2002. Occurrence of Flavobacterium psychrophilum in fish-farming environments. Diseases of Aquatic Organisms. 52, 109-118. Madsen, L., Dalsgaard, I. Characterization of Flavobacterium psychrophilum; comparison of proteolytic activity and virulence of strains isolated from trout (Oncorhynchus mykiss) In: Barnes A.C., Davidson, G.A., Hiney, M.P., McIntosh, D., editors. Methodology in fish disease research. Aberdeen, Scotland: Fisheries Research Service, 1998. pp. 45-52. Madsen, L., Dalsgaard, I. 1999. Reproducible methods for experimental infection with Flavobacterium psychrophilum in rainbow trout Oncorhynchus mykiss. Diseases of Aquatic Organisms. 36, 169-199. Michel, C., Antonio, D., Hedrick, R.P. 1999. Production of viable cultures of Flavobacterium psychrophilum: approach and control. Research in Microbiology. 150, 351-358. Miranda, C.D., Smith, P., Rojas, R., Contreras-Lynch, S., Alonso Vega, J.M. 2016. Antimicrobial susceptibility of Flavobacterium psychrophilum from Chilean salmon farms and their epidemiological cut-off values using agar dilution and disk diffusion methods. Frontiers in Microbiology. 7:1880. Nematollahi, A., Decostere, A., Pasmans, F., Haesebrouck, F. 2003. Flavobacterium psychrophilum infections in salmonid fish. Journal of Fish Diseases. 26, 563-574. Nicolas, P., Mondot, S., Achaz, G., Bouchenot, C., Bernardet, J-F., Duchaud, E. 2008. Population structure of the fish-pathogenic bacterium Flavobacterium psychrophilum. Applied and Environmental Microbiology. 74, 3701-3709. 77 Obach, A., Baudin-Laurencin, F. 1991. Vaccination of rainbow trout (Oncorhynchus mykiss) against the visceral form of cold-water disease. Diseases of Aquatic Organisms. 12, 13-15. Oplinger, R. W., Wagner, E.J. (2012). Effects of media ingredient substitution and comparison of growth of Flavobacterium psychrophilum among four media. Journal of Aquatic Animal Health. 24, 49-57. Perez-Pascual, D., Rochat, T., Kerouault, B., Gomez, E., Neulat-Ripoll, F., Henry, C., Quillet, E., Guijarro, J.A., Bernardet, J-F., Duchaud, E. 2017. More than gliding: involvement of GldD and GldG in the virulence of Flavobacterium psychrophilum. Frontiers in Microbiology. 8:2168. Plackett, R.L., Burman, J. P. 1944. The design of optimum multifactorial experiments. Biometrica. 33, 305-325. Ratledge, C., Dover, L.G. 2000. Iron metabolism in pathogenic bacteria. Annual Reviews in Microbiology. 54, 881-941. Rochat, T., Pérez-Pascual, Nilsen, H., Carpentier, M., Bridel, S., Bernardet, J-F, Duchaud, E. 2019. Identification of a Novel Elastin-Degrading Enzyme from the Fish Pathogen Flavobacterium psychrophilum. Applied and Environmental Microbiology. 85, e02535-18. Saticioglu, I.B., Duman, M., Smith, P., Wiklund, T., Altun, S. 2019. Antimicrobial resistance and resistance genes in Flavobacterium psychrophilum isolates from Turkey. Aquaculture. 512, 734293. Sebastiao, F., Loch, T.P., Knupp, C., Mukkatira, K., Veek, T., Richey, C., Adkison, M., Griffin, M.J., Soto, E. 2020. Multilocus sequence typing (MLST) analysis of California Flavobacterium psychrophilum reveals novel genotypes and predominance of CC-ST10 in California salmonid hatcheries. Aquaculture Research. 51, 2349-2358. Secades, P., Alvarez, B., Guijarro, J. A. 2001. Purification and characterization of a psychrophilic calcium induced, growth-phase-dependent metalloprotease from the fish pathogen Flavobacterium psychrophilum. Applied and Environmental Microbiology. 67, 2436-2444. Starliper, C.E. 2011. Bacterial coldwater disease of fishes caused by Flavobacterium psychrophilum. Journal of Advanced Research. 2, 97-108. Stevens, T.O. 1995. Optimization of media for enumeration and isolation of aerobic heterotrophic bacteria from the deep terrestrial subsurface. Journal of Microbiological Methods. 21(3), 293-303. Taha, F.S., Mohamed, S.S. 2004. Effect of different denaturating methods on lipid-protein complex formation. LWT – Food Science and Technology. 37(1), 99-104. Toyama, T., Kita-Tsukamoto, K., Wakabayashi, H. 1994. Identification of Cytophaga psychrophila by PCR targeted 16S ribosomal RNA. Fish Pathology. 29, 271-275. 78 Van Vliet, D., Loch, T.P., Faisal, M. 2015. Flavobacterium psychrophilum infections in salmonid broodstock and hatchery-propagated stocks of the Great Lakes Basin. Journal of Aquatic Animal Health. 27, 192–202. Van Vliet, D., Wiens, G.D., Loch, T.P., Nicolas, P., Faisal, M. 2016. Genetic diversity of Flavobacterium psychrophilum isolates from three Oncorhynchus spp. in the United States, as revealed by multilocus sequence typing. Applied and Environmental Microbiology. 82, 3246- 3255. Van Vliet, D., Loch T.P., Smith, P., Faisal, M. 2017. Antimicrobial susceptibilities of Flavobacterium psychrophilum isolates from the Great Lakes basin, Michigan. Microbial Drug Resistance. 23, 791-798. Wakabayashi, H., Toyama, T., Iidia, T. 1994. A study on serotyping of Cytophaga psychrophila isolated form fishes in Japan. Fish Pathology. 29, 101-104 Wang, T., Flint, S., Palmer, J. 2019. Magnesium and calcium ions: roles in bacterial cell attachment and biofilm structure maturation. Biofouling. 35(9), 959-974. Zeinab, A.K., Aly, M.S., Faiza, A.K., and Fatma, E.M. 2015. The Plackett-Burman design to evaluate significant media components for antimicrobial production of Lactobacillus rhamnosus. International Journal of Current Microbiology and Applied Sciences. 4(4), 1082-1096. 79 APPENDIX Table 2.1. Metadata for the 165 Flavobacterium psychrophilum isolates used in this study, including multilocus sequence typing sequence type (ST), clonal complex (CC), isolation location, and host of origin. Information is presented in order by CC then ST. U.S. state, Canadian province, or country of ST (CC) No. of isolates isolationa Host of originb ST9 (CC-ST9)c 4 BC (Canada), OR, UT, COS, RBT WA ST13 (CC-ST9)c, d, e 3 MI, WA COS ST2 (CC-ST10)f 1 Denmark RBT ST10 (CC-ST10)c 14 CA, CO, ID, MD, MI, CHS, RBT, WHT MT, NC, NM, OR, PA, SD, UT, WA ST78 (CC-ST10)c, d, g 10 CA, CO, ID, MI, NC, RBT NM, UT, WA, WY ST79 (CC-ST10)f 1 Chile ATS ST84 (CC-ST10)c 1 NC RBT ST85 (CC-ST10)c, e 2 CA, WA RBT ST86 (CC-ST10)c 2 UT, WA RBT ST275 (CC-ST10)c 6 MI, NC, PA, VA, WV RBT ST294 (CC-ST10)c 1 WA RBT ST300 (CC-ST10)c 2 ID, MT RBT ST303 (CC-ST10)c 1 ID RBT ST304 (CC-ST10)c 1 ID RBT ST305 (CC-ST10)c 1 ID RBT ST306 (CC-ST10)c 1 ID RBT ST316 (CC-ST10)c 1 WA RBT ST317 (CC-ST10)c 1 ID RBT ST319 (CC-ST10)e 1 CA RBT ST341 (CC-ST10)e 1 MI RBT ST342 (CC-ST10)g 2 MI RBT ST11 (CC-ST11)c 1 OR RBT ST262 (CC-ST11)d 1 MI CHS ST29 (CC-ST29)d 1 MI CHS ST308 (CC-ST29)c 1 OR CHS ST28 (CC-ST31)c 1 OR CUT ST31 (CC-ST31)d 1 MI RBT ST70 (CC-ST124)d 1 WA ATS ST267 (CC-ST191)c, d 2 MI RBT ST301 (CC-ST191)c 3 ID, PA, WV RBT ST277 (CC-ST232)c 1 MI ATS ST252 (CC-ST256)d 1 MI COS ST256 (CC-ST256)c, d, g 4 MI, WI COS, CHS, RBT ST257 (CC-ST276)c, d 2 MI RBT ST276 (CC-ST276)c 2 PA RBT ST279 (CC-ST281)c 1 PA RBT ST281 (CC-ST281)c 1 PA RBT ST331 (CC-ST281)d 1 PA RBT ST332 (CC-ST281)d 1 PA RBT ST286 (CC-ST286)g 1 MI BNT ST349 (CC-ST286)g 1 MI RBT ST371 (CC-ST286)h 1 MI ATS ST375 (CC-ST286)h 2 MI ATS, RBT 80 Table 2.1. (cont’d) U.S. state, Canadian province, or country of ST (CC) No. of isolates isolationa Host of originb ST254 (CC-ST287)d 1 MI CHS ST280 (CC-ST287)c 1 WI SPL ST287 (CC-ST287)c 1 MI COS ST354 (CC-ST287)g 1 MI RBT ST288 (CC-ST288)c 1 MI CHS ST289 (CC-ST288)c 1 MI CHS ST347 (CC-ST288)g 1 MI CHS ST265 (CC-ST296)d 1 MI CHS ST296 (CC-ST296)c 1 WA SOC ST299 (CC-ST296)c 1 WA RBT ST291 (CC-ST310)c 3 CA, ID, NC RBT ST310 (CC-ST310)c 1 ID RBT ST311 (CC-ST310)c 1 ID RBT ST318 (CC-ST318)c 1 CA RBT ST330 (CC-ST318)c 1 Unknown RBT ST343 (CC-ST343)g 1 NJ RBT ST344 (CC-ST344)g 1 NJ RBT ST27c 1 CA RBT ST30c 1 OR COS ST74c 1 OR COS ST76c 1 WA CHS ST250d 1 MI CHS ST251d 1 MI CHS ST253c 2 MI BNT ST255d 2 MI CHS ST258c, g 3 MI COS, RBT ST259d 1 MI CHS ST260d 1 MI CHS ST261d, h 2 MI CHS, RBT ST263d 1 MI CHS ST264d 1 MI COS ST266d 1 MI CHS ST278c, h 2 MI ATS, LAT ST282c 1 MI ATS ST284c 1 PA BNT ST290c 1 MI CHS ST292c 5 NC RBT ST293c 1 NC RBT ST295c 1 WA CHS ST297c 1 OR COS ST298c 1 ID RBT ST302c 1 ID RBT ST307c 1 OR RBT ST309c 1 ID RBT ST312c 1 ID RBT ST313c 1 MT RBT ST314c 1 NC RBT ST315c 1 NC RBT ST320c 1 AK COS ST333e 1 CA RBT ST345g 1 MI CHS 81 Table 2.1. (cont’d) U.S. state, Canadian province, or country of ST (CC) No. of isolates isolationa Host of originb ST346g 1 MI CHS ST350g 2 MI ATS, COS ST351g 1 WI BKT ST352g 1 WI BNT ST353g 2 MI BKT, LAT ST369h 1 MI RBT ST373h 1 MI BNT ST374i 1 MI LWF ST376h 1 MI BNT ST377i 1 MI LWF ST378i 1 WI LWF Unknownh 1 MI BNT a Isolates were recovered from 19 U.S. states, one Canadian province, and the countries of Chile and Denmark. AK, Alaska; BC, British Columbia; CA, California; CO, Colorado; ID, Idaho; MD, Maryland; MI, Michigan; MT, Montana; NC, North Carolina; NJ, New Jersey; NM, New Mexico; OR, Oregon; PA, Pennsylvania; SD, South Dakota; UT, Utah; VA, Virginia; WA, Washington; WI, Wisconsin; WV, West Virginia; WY, Wyoming. b Isolates were recovered from 12 fish species. ATS, Atlantic salmon (Salmo salar); BKT, brook trout (Salvelinus fontinalis); BNT, brown trout (S. trutta); CHS, Chinook salmon (Oncorhynchus tshawytscha); COS, coho salmon (O. kisutch); CUT, cutthroat trout (O. clarkii); LAT, lake trout (S. namaycush); LWF, lake whitefish (Coregonus clupeaformis); RBT, rainbow trout (O. mykiss); SOC, sockeye salmon (O. nerka); SPL, splake (S. namaycush x S. fontinalis); WHT, white sturgeon (Acipenser transmontanus). c Published in Knupp et al. (2019). d Published in Van Vliet et al. (2016). e Published at https://pubmlst.org/fpsychrophilum/ f Published in Ma et al. (2019) and Madsen and Dalsgaard (1999). g Published in Li et al. (2021). h Published in this study. i Published in Harrison et al. (2022). 82 Table 2.2. Additional meta data of the 165 Flavobacterium psychrophilum isolates used in this study, including isolate identifier (ID), year of isolation, location of isolation, host of origin, tissue of origin, life history (e.g., wild/feral or captive), life stage, and multilocus sequencing typing sequence type (ST) and clonal complex (CC). All 165 F. psychrophilum isolates were used to select the best basal medium (e.g., tryptone yeast extract agar; TYES) and in the comparison of the new media (F. psychrophilum medium-A and -B) to TYES. Fifty F. psychrophilum isolates were used in the Plackett-Burman experiment (denoted here with h). The information is presented in order by multi-locus sequence typing (MLST) clonal complex (CC), then MLST sequence type (ST). Year of Location of Wild/feral or Isolate ID isolation isolationi Host of origin Isolation tissue captive Life stage ST CC US161a 1990 OR O. kisutch Unknown Unknown Unknown ST9 CC-ST9 US165a 1984 BC O. kisutch Unknown Unknown Unknown ST9 CC-ST9 US254a,h Unknown UT O. mykiss Spleen Unknown Unknown ST9 CC-ST9 US256a Unknown WA O. kisutch Unknown Unknown Unknown ST9 CC-ST9 US019b,h 2010 MI O. kisutch Kidney Wild/feral Adult ST13 CC-ST9 US155a Unknown WA O. kisutch Unknown Unknown Unknown ST13 CC-ST9 US458e 2019 MI O. kisutch Kidney Captive Juvenile ST13 CC-ST9 900406-1/3f,h 1990 Denmark O. mykiss Kidney Unknown Juvenile ST2 CC-ST10 US075a,h 2016 PA O. mykiss Kidney Captive Broodstock ST10 CC-ST10 US133a 2012 SD O. mykiss Pooled tissue Captive Unknown ST10 CC-ST10 US148a Unknown CA A. transmontanus Kidney Unknown Unknown ST10 CC-ST10 US162a 1990 OR O. mykiss Unknown Unknown Unknown ST10 CC-ST10 US188a 2016 MD O. mykiss Unknown Captive Fingerling ST10 CC-ST10 US226a 2013 NC O. mykiss Lesion Captive Fingerling ST10 CC-ST10 US236a 2011 UT O. mykiss Spleen Captive Fingerling ST10 CC-ST10 US250a 2000 WA O. tshawytscha Ovarian fluid Unknown Unknown ST10 CC-ST10 US253a Unknown CO O. mykiss Unknown Unknown Unknown ST10 CC-ST10 US305a,h 2011 UT O. mykiss Spleen Captive Fry ST10 CC-ST10 US312a 2010 MT O. mykiss Kidney Captive Fingerling ST10 CC-ST10 US351a 2015 ID O. mykiss Spleen Captive Juvenile ST10 CC-ST10 US352a 2010 MI O. mykiss Kidney Captive Fingerling ST10 CC-ST10 US368a 2002 NM O. mykiss Kidney Unknown Unknown ST10 CC-ST10 US042b 2011 MI O. mykiss Kidney Captive Juvenile ST78 CC-ST10 US053b,h 2011 MI O. mykiss ext. lesion Captive Juvenile ST78 CC-ST10 US118a 2013 NC O. mykiss Unknown Captive Fingerling ST78 CC-ST10 US172a 2014 WA O. mykiss Kidney Captive Yearling ST78 CC-ST10 US241a 2013 ID O. mykiss Unknown Captive Juvenile ST78 CC-ST10 US251a 2002 NM O. mykiss Kidney Unknown Unknown ST78 CC-ST10 US252a Unknown CO O. mykiss Kidney Unknown Unknown ST78 CC-ST10 US361a 2017 CA O. mykiss Unknown Unknown Unknown ST78 CC-ST10 US367a 2008 UT O. mykiss Spleen Captive Fry ST78 CC-ST10 83 Table 2.2. (cont’d) Year of Location of Wild/feral or Isolate ID isolation isolationi Host of origin Isolation tissue captive Life stage ST CC US459c Unknown WY Unknown Unknown Unknown Unknown ST78 CC-ST10 622-97f Unknown Chile S. salar Unknown Unknown Unknown ST79 CC-ST10 US167a 2009 NC O. mykiss Spleen Captive Fingerling ST84 CC-ST10 US364a 2014 WA O. mykiss Heart Captive Yearling ST85 CC-ST10 US392e,h 2017 CA O. mykiss Unknown Unknown Unknown ST85 CC-ST10 US073a,h 2008 UT O. mykiss Spleen Captive Fry ST86 CC-ST10 US354a 2015 WA O. mykiss Skin Captive Yearling ST86 CC-ST10 US057a 2014 PA O. mykiss Kidney Captive Fingerling ST275 CC-ST10 US087a 2016 MI O. mykiss Ext. lesion Captive Fingerling ST275 CC-ST10 US104a,h 2016 MI O. mykiss Kidney Wild/feral Adult ST275 CC-ST10 US116a 2013 NC O. mykiss Spleen Captive Fingerling ST275 CC-ST10 US119a 2013 WV O. mykiss Spleen Captive Fingerling ST275 CC-ST10 US171a 2014 VA O. mykiss Spleen Captive Fingerling ST275 CC-ST10 US151a,h Unknown WA O. mykiss Unknown Unknown Unknown ST294 CC-ST10 US178a,h 2015 ID O. mykiss Spleen Captive Unknown ST300 CC-ST10 US314a 2011 MT O. mykiss Kidney Captive Fingerling ST300 CC-ST10 US323a 2013 ID O. mykiss Spleen Captive Juvenile ST303 CC-ST10 US205a 2015 ID O. mykiss Kidney Captive Juvenile ST304 CC-ST10 US245a 2011 ID O. mykiss Kidney Captive Juvenile ST305 CC-ST10 US259a 1990 ID O. mykiss Unknown Unknown Unknown ST306 CC-ST10 US355a 2015 WA O. mykiss Unknown Captive Yearling ST316 CC-ST10 US356a 2016 ID O. mykiss Spleen Captive Juvenile ST317 CC-ST10 US390e 2017 CA O. mykiss Unknown Unknown Unknown ST319 CC-ST10 US399e 2018 MI O. mykiss Kidney Wild/feral Adult ST341 CC-ST10 US400c 2018 MI O. mykiss Kidney Wild/feral Adult ST342 CC-ST10 US465g 2019 MI O. mykiss Kidney Captive Juvenile ST342 CC-ST10 US163a,h 1984 OR O. mykiss Unknown Unknown Unknown ST11 CC-ST11 US033b,h 2010 MI O. tshawytscha Kidney Wild/feral Adult ST262 CC-ST11 US025b,h 2011 MI O. tshawytscha Kidney Wild/feral Adult ST29 CC-ST29 US261a,h 1981 OR O. tshawytscha Unknown Unknown Unknown ST308 CC-ST29 US164a 1986 OR O. clarkii Unknown Unknown Unknown ST28 CC-ST31 US028b,h 2010 MI O. mykiss Kidney Wild/feral Adult ST31 CC-ST31 US149a,h Unknown WA S. salar Unknown Unknown Unknown ST70 CC-ST124 US054b 2013 MI O. mykiss Kidney Captive Juvenile ST267 CC-ST191 US215a,h 2017 MI O. mykiss Kidney Wild/feral Adult ST267 CC-ST191 US181a,h 2015 PA O. mykiss Unknown Captive Unknown ST301 CC-ST191 US277a 2015 ID O. mykiss Spleen Captive Juvenile ST301 CC-ST191 84 Table 2.2. (cont’d) Year of Location of Wild/feral or Isolate ID isolation isolationi Host of origin Isolation tissue captive Life stage ST CC US343a 2014 WV O. mykiss Caudal Fin Captive Fingerling ST301 CC-ST191 US062a,h 2012 MI S. salar Ext. lesion Wild/feral Adult ST277 CC-ST232 US008b,h 2011 MI O. kisutch Kidney Wild/feral Adult ST252 CC-ST256 US047b 2011 MI O. tshawytscha egg Wild/feral NA ST256 CC-ST256 US101a 2014 MI O. tshawytscha Kidney Wild/feral Adult ST256 CC-ST256 US217a,h 2017 MI O. mykiss Kidney Wild/feral Adult ST256 CC-ST256 US445c 2018 WI O. kisutch Unknown Unknown Unknown ST256 CC-ST256 US016b 2013 MI O. mykiss Kidney Captive Juvenile ST257 CC-ST276 US095a,h 2016 MI O. mykiss Kidney Wild/feral Adult ST257 CC-ST276 US061a,h 2014 PA O. mykiss Kidney Captive Fingerling ST276 CC-ST276 US200a 2016 PA O. mykiss Ulcer Captive Broodstock ST276 CC-ST276 US065a,h 2016 PA O. mykiss Kidney Wild/feral Adult ST279 CC-ST281 US067a,h 2016 PA O. mykiss Kidney Captive Fry ST281 CC-ST281 US379a,h 2017 PA O. mykiss Kidney Wild/feral Adult ST331 CC-ST281 US380a,h 2017 PA O. mykiss Kidney Wild/feral Adult ST332 CC-ST281 US461c 2019 MI S. trutta Gill Captive Juvenile ST286 CC-ST286 US439c,h 2018 MI O. mykiss Kidney Captive Juvenile ST349 CC-ST286 US478g,h 2020 MI S. salar Kidney Captive Juvenile ST371 CC-ST286 US493g 2020 MI S. salar Pectoral fin Captive Juvenile ST375 CC-ST286 US502g 2020 MI O. mykiss Caudal ped. Captive Broodstock ST375 CC-ST286 US012b 2013 MI O. tshawytscha Kidney Wild/feral Adult ST254 CC-ST287 S. namaycus x US066a,h 2016 WI Unknown Captive Unknown ST280 CC-ST287 S. fontinalis a,h US096 2013 MI O. kisutch Dorsal fin Captive Yearling ST287 CC-ST287 US455c 2019 MI O. mykiss Caudal ped. Captive Juvenile ST354 CC-ST287 US108a,h 2016 MI O. tshawytscha Kidney Wild/feral Adult ST288 CC-ST288 US109a,h 2016 MI O. tshawytscha Kidney Wild/feral Adult ST289 CC-ST288 US414c,h 2018 MI O. tshawytscha Kidney Wild/feral Adult ST347 CC-ST288 US037b 2012 MI O. tshawytscha Kidney Wild/feral Adult ST265 CC-ST296 US153a,h Unknown WA O. nerka Unknown Unknown Unknown ST296 CC-ST296 US176a 2014 WA O. mykiss Kidney Captive Yearling ST299 CC-ST296 US122a 2011 NC O. mykiss Brain Captive Fingerling ST291 CC-ST310 US331a 2014 ID O. mykiss Kidney Captive Juvenile ST291 CC-ST310 US359a 2017 CA O. mykiss Unknown Unknown Unknown ST291 CC-ST310 US329a,h 2014 ID O. mykiss Kidney Captive Juvenile ST310 CC-ST310 US328a 2014 ID O. mykiss Spleen Captive Juvenile ST311 CC-ST310 US357a 2016 CA O. mykiss Unknown Unknown Unknown ST318 CC-ST318 85 Table 2.2. (cont’d) Year of Location of Wild/feral or Isolate ID isolation isolationi Host of origin Isolation tissue captive Life stage ST CC US374a,h Unknown Unknown O. mykiss Unknown Unknown Unknown ST330 CC-ST318 US403c,h 2018 NJ O. mykiss Kidney Captive Juvenile ST343 CC-ST343 US404c,h 2018 NJ O. mykiss Kidney Captive Juvenile ST344 CC-ST343 US191a,h 2016 CA O. mykiss Spleen Captive Unknown ST27 US156a 1981 OR O. kisutch Unknown Unknown Unknown ST30 US255a 1981 OR O. kisutch Unknown Unknown Unknown ST74 US249a,h 2000 WA O. tshawytscha Ovarian fluid Unknown Unknown ST76 US005b 2011 MI O. tshawytscha Egg Wild/feral Unknown ST250 US006b 2011 MI O. tshawytscha Egg Wild/feral Unknown ST251 US060a 2010 MI S. trutta Ascites Captive Adult ST253 US098a 2013 MI S. trutta Spleen Captive Fingerling ST253 US013b 2012 MI O. tshawytscha Kidney Wild/feral Adult ST255 US031b 2012 MI O. tshawytscha Kidney Wild/feral Adult ST255 US099a 2013 MI O. kisutch Gill Captive Yearling ST258 US218a 2017 MI O. mykiss Kidney Wild/feral Adult ST258 US454c 2018 MI O. kisutch Kidney Wild/feral Adult ST258 US023b 2013 MI O. tshawytscha Kidney Wild/feral Adult ST259 US056b 2010 MI O. tshawytscha Kidney Wild/feral Adult ST260 US024b 2009 MI O. tshawytscha Kidney Wild/feral Adult ST261 US479g 2020 MI O. mykiss Eye Captive Juvenile ST261 US034b 2012 MI O. tshawytscha Kidney Wild/feral Adult ST263 US035b 2011 MI O. kisutch Kidney Wild/feral Adult ST264 US036b 2008 MI O. tshawytscha Kidney Wild/feral Adult ST266 US063a 2012 MI S. namaycus Fin Captive Fingerling ST278 US508g 2020 MI S. salar Kidney Wild/feral Adult ST278 US083a 2016 MI S. salar Eye Wild/feral Adult ST282 US088a,h 2016 PA S. trutta Lesion Captive Adult ST284 US110a 2016 MI O. tshawytscha Kidney Wild/feral Adult ST290 US139a 2013 NC O. mykiss Spleen Captive Fingerling ST292 US141a 2011 NC O. mykiss Kidney Captive Fingerling ST292 US225a 2013 NC O. mykiss Spleen Captive Fingerling ST292 US232a 2011 NC O. mykiss Kidney Captive Fingerling ST292 US244a,h 2011 NC O. mykiss Spleen Captive Fingerling ST292 US142a 2011 NC O. mykiss Brain Captive Fingerling ST293 US152a Unknown WA O. tshawytscha Unknown Unknown Unknown ST295 US158a,h 1989 OR O. kisutch Unknown Unknown Unknown ST297 US168a,h 2014 ID O. mykiss Kidney Captive Juvenile ST298 86 Table 2.2. (cont’d) Year of Location of Wild/feral or Isolate ID isolation isolationi Host of origin Isolation tissue captive Life stage ST CC US187a 2014 ID O. mykiss Spleen Captive Unknown ST302 US260a 1985 OR O. mykiss Unknown Unknown Unknown ST307 US265a 2014 ID O. mykiss Kidney Captive Juvenile ST309 US283a 2014 ID O. mykiss Spleen Captive Juvenile ST312 US310a,h 2010 MT O. mykiss Kidney Captive Fingerling ST313 US324a 2014 NC O. mykiss Spleen Captive Fry ST314 US325a 2014 NC O. mykiss Spleen Captive Fry ST315 US372a,h 2011 AK O. kisutch Kidney Captive Fingerling ST320 US394e 2017 CA O. mykiss Unknown Unknown Unknown ST333 US411c 2018 MI O. tshawytscha Kidney Wild/feral Adult ST345 US413c 2018 MI O. tshawytscha Kidney Wild/feral Adult ST346 US442c 2018 MI S. salar Kidney Wild/feral Adult ST350 US443c 2019 MI O. kisutch Kidney Captive Juvenile ST350 US444c,h 2018 WI S. fontinalis Unknown Unknown Unknown ST351 US449c 2018 WI S. trutta Unknown Unknown Unknown ST352 US450c 2019 MI S. namaycus Dorsal fin Captive Juvenile ST353 US487g 2020 MI S. fontinalis Dorsal fin Captive Broodstock ST353 US476g 2020 MI O. mykiss Pectoral fin Captive Juvenile ST369 US485g 2020 MI S. trutta Kidney Captive Juvenile ST373 US490d,h 2019 MI C. clupeaformis Kidney Wild/feral Adult ST374 US505g 2020 MI S. trutta Caudal ped. Captive Adult ST376 US488d 2019 MI C. clupeaformis Kidney Wild/feral Adult ST377 US489d 2019 WI C. clupeaformis Kidney Wild/feral Adult ST378 US460g 2019 MI S. trutta Kidney Captive Juvenile NA a Published in Knupp et al. (2019) b Published in Van Vliet et al. (2016) c Published in Li et al. (2021) d Published in Harrison et al. (2022) e Published at https://pubmlst.org/fpsychrophilum/ f Published in Ma et al. (2019) and Madsen and Dalsgaard (1999) g Published in this study h Isolate used in Plackett-Burman experiment i Key to location abbreviations: AK=Alaska; BC=British Columbia; CA=California; CO=Colorado; ID=Idaho; MD=Maryland; MI=Michigan; MT=Montana; NC=North Carolina; NJ=New Jersey; NM=New Mexico; OR=Oregon; PA=Pennsylvania; SD=South Dakota; UT=Utah; VA=Virginia; WA=Washington; WI=Wisconsin; WV=West Virginia; WY=Wyoming 87 Table 2.3. Formulations (per 1L H2O) of media used in basal medium experiment, including tryptone yeast extract salts agar (TYES; Holt 1987), Oplinger and Wagner medium (OW; Oplinger and Wagner, 2012), enriched Anacker and Ordal medium supplemented with activated charcoal and aromatic compounds (EAOCa; Alvarez and Guijarro, 2007), and neutral medium (NM; this study). Medium Medium component Unit TYES OW EAOCa NM Vendor Bacteriological agar g 10 10 15 10 Thermofisher Scientific Tryptone g 4.0 5.0 5.0 4.5 Thermofisher Scientific Yeast extract g 0.4 0.5 0.45 Thermofisher Scientific CaCl2 • 2H2O g 0.5 Thermofisher Scientific MgSO4 • 7H2O g 0.5 Thermofisher Scientific Beef extract g 0.2 0.2 Millipore Sigma Horse serum % 1.0 5.0 Thermofisher Scientific Skim milk powder % 0.2 Thermofisher Scientific Activated charcoal g 0.5 Millipore Sigma Sodium acetate g 0.2 Millipore Sigma L-tyrosine µmol 10 Millipore Sigma L-phenylalanine µmol 10 Millipore Sigma L-tryptophan µmol 10 Millipore Sigma p-aminobenzoic acid µmol 10 Millipore Sigma p-hydroxybenzoic acid µmol 10 Millipore Sigma 2,3-di-hydroxybenzoic acid µmol 10 Millipore Sigma 88 Table 2.4. Medium components tested as part of Plackett-Burman experimental design, including tested quantities per 1L H2O. Medium Component Low value High value Catalog component Unit symbol (-1) (+1) Brief rationale Vendor number Tryptone g X1 0.750 7.50 Amino acid source ThermoFisher Scientific LP0042B Yeast extract g X2 0.075 0.75 Vitamin and nitrogen source ThermoFisher Scientific Y1625 CaCl2 • 2H2O g X3 0.075 0.75 Ca++ source Millipore Sigma C3306 MgSO4 • 7H2O g X4 0.111 1.11 Mg++ source Millipore Sigma 63138 Beef extract g X5 0.075 0.75 Amino acid source Millipore Sigma B4888 Casamino acids g X6 0.750 7.50 Amino acid source ThermoFisher Scientific 223050 Tryptose g X7 0.750 7.50 Amino acid source Millipore Sigma 70937 L-aspartic acid g X8 0.075 0.75 A main amino acid in O. mykiss muscle tissue Millipore Sigma A9256 Horse serum % X9 0.750 7.50 Source of proteins and other growth factors ThermoFisher Scientific 26050088 Fetal bovine serum % X10 0.750 7.50 Source of proteins and other growth factors ThermoFisher Scientific 100500 O. mykiss blood % X11 0.750 7.50 Source of iron and other host factors This study 89 Table 2.5. Plackett-Burman experimental design matrix, consisting of 11 tested medium components (e.g., X1-X11) resulting in 12 media formulations. Medium components incorporated at their low or high concentration are denoted with -1 and +1, respectively. Medium Component Medium Formulation X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 1 -1 -1 -1 -1 -1 -1 -1 -1 -1 -1 -1 2 +1 -1 -1 -1 +1 +1 +1 -1 +1 +1 -1 3 +1 +1 -1 +1 -1 -1 -1 +1 +1 +1 -1 4 -1 -1 +1 +1 +1 -1 +1 +1 -1 +1 -1 5 -1 +1 -1 -1 -1 +1 +1 +1 -1 +1 +1 6 -1 +1 +1 -1 +1 -1 -1 -1 +1 +1 +1 7 +1 -1 +1 -1 -1 -1 +1 +1 +1 -1 +1 8 +1 +1 +1 -1 +1 +1 -1 +1 -1 -1 -1 9 +1 +1 -1 +1 +1 -1 +1 -1 -1 -1 +1 10 -1 -1 -1 +1 +1 +1 -1 +1 +1 -1 +1 11 +1 -1 +1 +1 -1 +1 -1 -1 -1 +1 +1 12 -1 +1 +1 +1 -1 +1 +1 -1 +1 -1 -1 90 Table 2.6. Differences of least square mean estimates ± standard error (SE) for the interaction between medium (e.g., EAOCa, OW, and TYES) and 165 Flavobacterium psychrophilum isolates. Pairwise comparisons between media for each isolate are provided. Tukey-Kramer adjusted P-values for multiple comparisons are shown (α = 0.05). Table is ordered by isolate. Medium 1 Isolate 1 Medium 2 Isolate 2 Estimate SE DF t-value P-value EAOCa FP622-97 OW FP622-97 -0.6505 0.2224 330 -2.93 1 EAOCa FP622-97 TYES FP622-97 -0.9515 0.2224 330 -4.28 0.4881 OW FP622-97 TYES FP622-97 -0.301 0.2224 330 -1.35 1 EAOCa FP900406 OW FP900406 -6.88E-15 0.2224 330 0 1 EAOCa FP900406 TYES FP900406 -1.2386 0.2224 330 -5.57 0.0046 OW FP900406 TYES FP900406 -1.2386 0.2224 330 -5.57 0.0046 EAOCa US005 OW US005 -1.787 0.2224 330 -8.04 <.0001 EAOCa US005 TYES US005 -2.088 0.2224 330 -9.39 <.0001 OW US005 TYES US005 -0.301 0.2224 330 -1.35 1 EAOCa US006 OW US006 -1.0945 0.2224 330 -4.92 0.0734 EAOCa US006 TYES US006 -1.5104 0.2224 330 -6.79 <.0001 OW US006 TYES US006 -0.4158 0.2224 330 -1.87 1 EAOCa US008 OW US008 -0.3099 0.2224 330 -1.39 1 EAOCa US008 TYES US008 -0.7955 0.2224 330 -3.58 0.9786 OW US008 TYES US008 -0.4856 0.2224 330 -2.18 1 EAOCa US012 OW US012 -1.3288 0.2224 330 -5.98 0.0006 EAOCa US012 TYES US012 -1.2407 0.2224 330 -5.58 0.0043 OW US012 TYES US012 0.08805 0.2224 330 0.4 1 EAOCa US013 OW US013 -2.2218 0.2224 330 -9.99 <.0001 EAOCa US013 TYES US013 -3.0713 0.2224 330 -13.81 <.0001 OW US013 TYES US013 -0.8495 0.2224 330 -3.82 0.8924 EAOCa US016 OW US016 -1.4225 0.2224 330 -6.4 <.0001 EAOCa US016 TYES US016 -1.6109 0.2224 330 -7.24 <.0001 OW US016 TYES US016 -0.1884 0.2224 330 -0.85 1 EAOCa US019 OW US019 -1.5812 0.2224 330 -7.11 <.0001 EAOCa US019 TYES US019 -2.2211 0.2224 330 -9.99 <.0001 OW US019 TYES US019 -0.6399 0.2224 330 -2.88 1 EAOCa US023 OW US023 1.1505 0.2224 330 5.17 0.0271 EAOCa US023 TYES US023 -1.0256 0.2224 330 -4.61 0.2095 OW US023 TYES US023 -2.1761 0.2224 330 -9.79 <.0001 EAOCa US024 OW US024 0.1823 0.2224 330 0.82 1 EAOCa US024 TYES US024 -0.3741 0.2224 330 -1.68 1 OW US024 TYES US024 -0.5564 0.2224 330 -2.5 1 EAOCa US025 OW US025 -0.3656 0.2224 330 -1.64 1 EAOCa US025 TYES US025 -0.554 0.2224 330 -2.49 1 OW US025 TYES US025 -0.1884 0.2224 330 -0.85 1 EAOCa US028 OW US028 -0.699 0.2224 330 -3.14 0.9999 EAOCa US028 TYES US028 -1.9375 0.2224 330 -8.71 <.0001 OW US028 TYES US028 -1.2386 0.2224 330 -5.57 0.0046 EAOCa US031 OW US031 -1.2614 0.2224 330 -5.67 0.0028 EAOCa US031 TYES US031 -1.588 0.2224 330 -7.14 <.0001 OW US031 TYES US031 -0.3266 0.2224 330 -1.47 1 EAOCa US033 OW US033 -0.7992 0.2224 330 -3.59 0.9755 EAOCa US033 TYES US033 -1.6109 0.2224 330 -7.24 <.0001 OW US033 TYES US033 -0.8117 0.2224 330 -3.65 0.9625 EAOCa US034 OW US034 -2.713 0.2224 330 -12.2 <.0001 91 Table 2.6. (cont’d) Medium 1 Isolate 1 Medium 2 Isolate 2 Estimate SE DF t-value P-value OW US035 TYES US035 0.03959 0.2224 330 0.18 1 EAOCa US036 OW US036 -2 0.2224 330 -8.99 <.0001 EAOCa US036 TYES US036 -2.199 0.2224 330 -9.89 <.0001 OW US036 TYES US036 -0.199 0.2224 330 -0.89 1 EAOCa US037 OW US037 -0.6109 0.2224 330 -2.75 1 EAOCa US037 TYES US037 -0.6109 0.2224 330 -2.75 1 OW US037 TYES US037 -1.17E-15 0.2224 330 0 1 EAOCa US042 OW US042 -0.1505 0.2224 330 -0.68 1 EAOCa US042 TYES US042 -0.5 0.2224 330 -2.25 1 OW US042 TYES US042 -0.3495 0.2224 330 -1.57 1 EAOCa US047 OW US047 -0.9621 0.2224 330 -4.33 0.4413 EAOCa US047 TYES US047 -0.8751 0.2224 330 -3.93 0.815 OW US047 TYES US047 0.08708 0.2224 330 0.39 1 EAOCa US053 OW US053 -2.1505 0.2224 330 -9.67 <.0001 EAOCa US053 TYES US053 -2.9375 0.2224 330 -13.21 <.0001 OW US053 TYES US053 -0.787 0.2224 330 -3.54 0.9845 EAOCa US054 OW US054 -4.1109 0.2224 330 -18.49 <.0001 EAOCa US054 TYES US054 -5.0089 0.2224 330 -22.52 <.0001 OW US054 TYES US054 -0.8979 0.2224 330 -4.04 0.7273 EAOCa US056 OW US056 -0.4543 0.2224 330 -2.04 1 EAOCa US056 TYES US056 0.1672 0.2224 330 0.75 1 OW US056 TYES US056 0.6215 0.2224 330 2.79 1 EAOCa US057 OW US057 -1.412 0.2224 330 -6.35 <.0001 EAOCa US057 TYES US057 -2.301 0.2224 330 -10.35 <.0001 OW US057 TYES US057 -0.8891 0.2224 330 -4 0.763 EAOCa US060 OW US060 -3.273 0.2224 330 -14.72 <.0001 EAOCa US060 TYES US060 -2.7959 0.2224 330 -12.57 <.0001 OW US060 TYES US060 0.4771 0.2224 330 2.15 1 EAOCa US061 OW US061 -2.7461 0.2224 330 -12.35 <.0001 EAOCa US061 TYES US061 -2.7461 0.2224 330 -12.35 <.0001 OW US061 TYES US061 -1.78E-15 0.2224 330 0 1 EAOCa US062 OW US062 -0.4771 0.2224 330 -2.15 1 EAOCa US062 TYES US062 -0.3891 0.2224 330 -1.75 1 OW US062 TYES US062 0.08805 0.2224 330 0.4 1 EAOCa US063 OW US063 -2.4098 0.2224 330 -10.84 <.0001 EAOCa US063 TYES US063 -2.301 0.2224 330 -10.35 <.0001 OW US063 TYES US063 0.1087 0.2224 330 0.49 1 EAOCa US065 OW US065 -1 0.2224 330 -4.5 0.2909 EAOCa US065 TYES US065 -2.2386 0.2224 330 -10.07 <.0001 OW US065 TYES US065 -1.2386 0.2224 330 -5.57 0.0046 EAOCa US066 OW US066 -2.772 0.2224 330 -12.46 <.0001 EAOCa US066 TYES US066 -4.1505 0.2224 330 -18.66 <.0001 OW US066 TYES US066 -1.3785 0.2224 330 -6.2 0.0002 EAOCa US067 OW US067 -2.8406 0.2224 330 -12.77 <.0001 EAOCa US067 TYES US067 -3.1751 0.2224 330 -14.28 <.0001 OW US067 TYES US067 -0.3345 0.2224 330 -1.5 1 EAOCa US073 OW US073 -3.4375 0.2224 330 -15.46 <.0001 EAOCa US073 TYES US073 -3.7526 0.2224 330 -16.87 <.0001 OW US073 TYES US073 -0.315 0.2224 330 -1.42 1 EAOCa US075 OW US075 -0.2849 0.2224 330 -1.28 1 92 Table 2.6. (cont’d) Medium 1 Isolate 1 Medium 2 Isolate 2 Estimate SE DF t-value P-value EAOCa US075 TYES US075 -0.728 0.2224 330 -3.27 0.9992 OW US075 TYES US075 -0.443 0.2224 330 -1.99 1 EAOCa US083 OW US083 -0.8391 0.2224 330 -3.77 0.9168 EAOCa US083 TYES US083 -0.801 0.2224 330 -3.6 0.9739 OW US083 TYES US083 0.03808 0.2224 330 0.17 1 EAOCa US087 OW US087 -2.2218 0.2224 330 -9.99 <.0001 EAOCa US087 TYES US087 -2.5625 0.2224 330 -11.52 <.0001 OW US087 TYES US087 -0.3406 0.2224 330 -1.53 1 EAOCa US088 OW US088 -1.7614 0.2224 330 -7.92 <.0001 EAOCa US088 TYES US088 -2.014 0.2224 330 -9.06 <.0001 OW US088 TYES US088 -0.2526 0.2224 330 -1.14 1 EAOCa US095 OW US095 -1.8495 0.2224 330 -8.32 <.0001 EAOCa US095 TYES US095 -2.088 0.2224 330 -9.39 <.0001 OW US095 TYES US095 -0.2386 0.2224 330 -1.07 1 EAOCa US096 OW US096 -0.6972 0.2224 330 -3.13 0.9999 EAOCa US096 TYES US096 -0.8081 0.2224 330 -3.63 0.9667 OW US096 TYES US096 -0.1109 0.2224 330 -0.5 1 EAOCa US098 OW US098 -1.9208 0.2224 330 -8.64 <.0001 EAOCa US098 TYES US098 -2.2218 0.2224 330 -9.99 <.0001 OW US098 TYES US098 -0.301 0.2224 330 -1.35 1 EAOCa US099 OW US099 -0.8406 0.2224 330 -3.78 0.9135 EAOCa US099 TYES US099 -0.9031 0.2224 330 -4.06 0.7056 OW US099 TYES US099 -0.06247 0.2224 330 -0.28 1 EAOCa US101 OW US101 0.3702 0.2224 330 1.66 1 EAOCa US101 TYES US101 -0.2803 0.2224 330 -1.26 1 OW US101 TYES US101 -0.6505 0.2224 330 -2.93 1 EAOCa US104 OW US104 -0.301 0.2224 330 -1.35 1 EAOCa US104 TYES US104 -0.7197 0.2224 330 -3.24 0.9995 OW US104 TYES US104 -0.4186 0.2224 330 -1.88 1 EAOCa US108 OW US108 -0.7614 0.2224 330 -3.42 0.9949 EAOCa US108 TYES US108 -0.772 0.2224 330 -3.47 0.9917 OW US108 TYES US108 -0.01059 0.2224 330 -0.05 1 EAOCa US109 OW US109 -0.7261 0.2224 330 -3.27 0.9993 EAOCa US109 TYES US109 -1.184 0.2224 330 -5.32 0.0141 OW US109 TYES US109 -0.4578 0.2224 330 -2.06 1 EAOCa US110 OW US110 -2.6215 0.2224 330 -11.79 <.0001 EAOCa US110 TYES US110 -2.588 0.2224 330 -11.64 <.0001 OW US110 TYES US110 0.03347 0.2224 330 0.15 1 EAOCa US116 OW US116 -0.6422 0.2224 330 -2.89 1 EAOCa US116 TYES US116 -0.6611 0.2224 330 -2.97 1 OW US116 TYES US116 -0.01889 0.2224 330 -0.08 1 EAOCa US118 OW US118 -1.5379 0.2224 330 -6.92 <.0001 EAOCa US118 TYES US118 -1.3645 0.2224 330 -6.14 0.0003 OW US118 TYES US118 0.1734 0.2224 330 0.78 1 EAOCa US119 OW US119 -1.8724 0.2224 330 -8.42 <.0001 EAOCa US119 TYES US119 -2.6109 0.2224 330 -11.74 <.0001 OW US119 TYES US119 -0.7386 0.2224 330 -3.32 0.9985 EAOCa US122 OW US122 2.0256 0.2224 330 9.11 <.0001 EAOCa US122 TYES US122 0.2641 0.2224 330 1.19 1 OW US122 TYES US122 -1.7614 0.2224 330 -7.92 <.0001 93 Table 2.6. (cont’d) Medium 1 Isolate 1 Medium 2 Isolate 2 Estimate SE DF t-value P-value EAOCa US133 OW US133 -0.9269 0.2224 330 -4.17 0.6 EAOCa US133 TYES US133 -0.5775 0.2224 330 -2.6 1 OW US133 TYES US133 0.3495 0.2224 330 1.57 1 EAOCa US139 OW US139 -2.8335 0.2224 330 -12.74 <.0001 EAOCa US139 TYES US139 -3.5229 0.2224 330 -15.84 <.0001 OW US139 TYES US139 -0.6894 0.2224 330 -3.1 0.9999 EAOCa US141 OW US141 -1.3099 0.2224 330 -5.89 0.0009 EAOCa US141 TYES US141 -2.4604 0.2224 330 -11.06 <.0001 OW US141 TYES US141 -1.1505 0.2224 330 -5.17 0.0271 EAOCa US142 OW US142 -1.912 0.2224 330 -8.6 <.0001 EAOCa US142 TYES US142 -3.7386 0.2224 330 -16.81 <.0001 OW US142 TYES US142 -1.8266 0.2224 330 -8.21 <.0001 EAOCa US148 OW US148 1.057 0.2224 330 4.75 0.1336 EAOCa US148 TYES US148 0.01109 0.2224 330 0.05 1 OW US148 TYES US148 -1.0459 0.2224 330 -4.7 0.1575 EAOCa US149 OW US149 0.07306 0.2224 330 0.33 1 EAOCa US149 TYES US149 0.03347 0.2224 330 0.15 1 OW US149 TYES US149 -0.03959 0.2224 330 -0.18 1 EAOCa US151 OW US151 3.3684 0.2224 330 15.15 <.0001 EAOCa US151 TYES US151 1.787 0.2224 330 8.04 <.0001 OW US151 TYES US151 -1.5814 0.2224 330 -7.11 <.0001 EAOCa US152 OW US152 -2.2614 0.2224 330 -10.17 <.0001 EAOCa US152 TYES US152 -2.7386 0.2224 330 -12.31 <.0001 OW US152 TYES US152 -0.4771 0.2224 330 -2.15 1 EAOCa US153 OW US153 1.3891 0.2224 330 6.25 0.0001 EAOCa US153 TYES US153 0.1505 0.2224 330 0.68 1 OW US153 TYES US153 -1.2386 0.2224 330 -5.57 0.0046 EAOCa US155 OW US155 0.7245 0.2224 330 3.26 0.9993 EAOCa US155 TYES US155 0.1225 0.2224 330 0.55 1 OW US155 TYES US155 -0.6021 0.2224 330 -2.71 1 EAOCa US156 OW US156 -0.2218 0.2224 330 -1 1 EAOCa US156 TYES US156 -0.9208 0.2224 330 -4.14 0.6277 OW US156 TYES US156 -0.699 0.2224 330 -3.14 0.9999 EAOCa US158 OW US158 1.0986 0.2224 330 4.94 0.0685 EAOCa US158 TYES US158 -0.9894 0.2224 330 -4.45 0.3297 OW US158 TYES US158 -2.088 0.2224 330 -9.39 <.0001 EAOCa US161 OW US161 -3.5542 0.2224 330 -15.98 <.0001 EAOCa US161 TYES US161 -3.3891 0.2224 330 -15.24 <.0001 OW US161 TYES US161 0.1651 0.2224 330 0.74 1 EAOCa US162 OW US162 -0.4515 0.2224 330 -2.03 1 EAOCa US162 TYES US162 -0.7386 0.2224 330 -3.32 0.9985 OW US162 TYES US162 -0.287 0.2224 330 -1.29 1 EAOCa US163 OW US163 -1.199 0.2224 330 -5.39 0.0104 EAOCa US163 TYES US163 -2.1505 0.2224 330 -9.67 <.0001 OW US163 TYES US163 -0.9515 0.2224 330 -4.28 0.4881 EAOCa US164 OW US164 1.3266 0.2224 330 5.97 0.0006 EAOCa US164 TYES US164 0.2386 0.2224 330 1.07 1 OW US164 TYES US164 -1.088 0.2224 330 -4.89 0.0818 EAOCa US165 OW US165 -0.1722 0.2224 330 -0.77 1 EAOCa US165 TYES US165 -1.1228 0.2224 330 -5.05 0.045 94 Table 2.6. (cont’d) Medium 1 Isolate 1 Medium 2 Isolate 2 Estimate SE DF t-value P-value EAOCa US133 OW US133 -0.9269 0.2224 330 -4.17 0.6 EAOCa US133 TYES US133 -0.5775 0.2224 330 -2.6 1 OW US133 TYES US133 0.3495 0.2224 330 1.57 1 EAOCa US139 OW US139 -2.8335 0.2224 330 -12.74 <.0001 EAOCa US139 TYES US139 -3.5229 0.2224 330 -15.84 <.0001 OW US139 TYES US139 -0.6894 0.2224 330 -3.1 0.9999 EAOCa US141 OW US141 -1.3099 0.2224 330 -5.89 0.0009 EAOCa US141 TYES US141 -2.4604 0.2224 330 -11.06 <.0001 OW US141 TYES US141 -1.1505 0.2224 330 -5.17 0.0271 EAOCa US142 OW US142 -1.912 0.2224 330 -8.6 <.0001 EAOCa US142 TYES US142 -3.7386 0.2224 330 -16.81 <.0001 OW US142 TYES US142 -1.8266 0.2224 330 -8.21 <.0001 EAOCa US148 OW US148 1.057 0.2224 330 4.75 0.1336 EAOCa US148 TYES US148 0.01109 0.2224 330 0.05 1 OW US148 TYES US148 -1.0459 0.2224 330 -4.7 0.1575 EAOCa US149 OW US149 0.07306 0.2224 330 0.33 1 EAOCa US149 TYES US149 0.03347 0.2224 330 0.15 1 OW US149 TYES US149 -0.03959 0.2224 330 -0.18 1 EAOCa US151 OW US151 3.3684 0.2224 330 15.15 <.0001 EAOCa US151 TYES US151 1.787 0.2224 330 8.04 <.0001 OW US151 TYES US151 -1.5814 0.2224 330 -7.11 <.0001 EAOCa US152 OW US152 -2.2614 0.2224 330 -10.17 <.0001 EAOCa US152 TYES US152 -2.7386 0.2224 330 -12.31 <.0001 OW US152 TYES US152 -0.4771 0.2224 330 -2.15 1 EAOCa US153 OW US153 1.3891 0.2224 330 6.25 0.0001 EAOCa US153 TYES US153 0.1505 0.2224 330 0.68 1 OW US153 TYES US153 -1.2386 0.2224 330 -5.57 0.0046 EAOCa US155 OW US155 0.7245 0.2224 330 3.26 0.9993 EAOCa US155 TYES US155 0.1225 0.2224 330 0.55 1 OW US155 TYES US155 -0.6021 0.2224 330 -2.71 1 EAOCa US156 OW US156 -0.2218 0.2224 330 -1 1 EAOCa US156 TYES US156 -0.9208 0.2224 330 -4.14 0.6277 OW US156 TYES US156 -0.699 0.2224 330 -3.14 0.9999 EAOCa US158 OW US158 1.0986 0.2224 330 4.94 0.0685 EAOCa US158 TYES US158 -0.9894 0.2224 330 -4.45 0.3297 OW US158 TYES US158 -2.088 0.2224 330 -9.39 <.0001 EAOCa US161 OW US161 -3.5542 0.2224 330 -15.98 <.0001 EAOCa US161 TYES US161 -3.3891 0.2224 330 -15.24 <.0001 OW US161 TYES US161 0.1651 0.2224 330 0.74 1 EAOCa US162 OW US162 -0.4515 0.2224 330 -2.03 1 EAOCa US162 TYES US162 -0.7386 0.2224 330 -3.32 0.9985 OW US162 TYES US162 -0.287 0.2224 330 -1.29 1 EAOCa US163 OW US163 -1.199 0.2224 330 -5.39 0.0104 EAOCa US163 TYES US163 -2.1505 0.2224 330 -9.67 <.0001 OW US163 TYES US163 -0.9515 0.2224 330 -4.28 0.4881 EAOCa US164 OW US164 1.3266 0.2224 330 5.97 0.0006 EAOCa US164 TYES US164 0.2386 0.2224 330 1.07 1 OW US164 TYES US164 -1.088 0.2224 330 -4.89 0.0818 EAOCa US165 OW US165 -0.1722 0.2224 330 -0.77 1 EAOCa US165 TYES US165 -1.1228 0.2224 330 -5.05 0.045 95 Table 2.6. (cont’d) Medium 1 Isolate 1 Medium 2 Isolate 2 Estimate SE DF t-value P-value OW US165 TYES US165 -0.9506 0.2224 330 -4.27 0.4925 EAOCa US167 OW US167 -1.1365 0.2224 330 -5.11 0.0351 EAOCa US167 TYES US167 -1.3601 0.2224 330 -6.12 0.0003 OW US167 TYES US167 -0.2236 0.2224 330 -1.01 1 EAOCa US168 OW US168 -1.145 0.2224 330 -5.15 0.03 EAOCa US168 TYES US168 -1.1422 0.2224 330 -5.14 0.0316 OW US168 TYES US168 0.002802 0.2224 330 0.01 1 EAOCa US171 OW US171 0.2526 0.2224 330 1.14 1 EAOCa US171 TYES US171 -0.3601 0.2224 330 -1.62 1 OW US171 TYES US171 -0.6127 0.2224 330 -2.75 1 EAOCa US172 OW US172 -0.456 0.2224 330 -2.05 1 EAOCa US172 TYES US172 0.03959 0.2224 330 0.18 1 OW US172 TYES US172 0.4956 0.2224 330 2.23 1 EAOCa US176 OW US176 -2.912 0.2224 330 -13.09 <.0001 EAOCa US176 TYES US176 -3.2218 0.2224 330 -14.49 <.0001 OW US176 TYES US176 -0.3099 0.2224 330 -1.39 1 EAOCa US178 OW US178 0.1505 0.2224 330 0.68 1 EAOCa US178 TYES US178 -6.22E-15 0.2224 330 0 1 OW US178 TYES US178 -0.1505 0.2224 330 -0.68 1 EAOCa US181 OW US181 -1.316 0.2224 330 -5.92 0.0008 EAOCa US181 TYES US181 -1.2636 0.2224 330 -5.68 0.0026 OW US181 TYES US181 0.05237 0.2224 330 0.24 1 EAOCa US187 OW US187 -2.4604 0.2224 330 -11.06 <.0001 EAOCa US187 TYES US187 -2.4604 0.2224 330 -11.06 <.0001 OW US187 TYES US187 1.28E-15 0.2224 330 0 1 EAOCa US188 OW US188 -0.617 0.2224 330 -2.77 1 EAOCa US188 TYES US188 -0.8785 0.2224 330 -3.95 0.8029 OW US188 TYES US188 -0.2614 0.2224 330 -1.18 1 EAOCa US191 OW US191 -1.301 0.2224 330 -5.85 0.0011 EAOCa US191 TYES US191 -1.7782 0.2224 330 -8 <.0001 OW US191 TYES US191 -0.4771 0.2224 330 -2.15 1 EAOCa US200 OW US200 -2.23 0.2224 330 -10.03 <.0001 EAOCa US200 TYES US200 -2.5311 0.2224 330 -11.38 <.0001 OW US200 TYES US200 -0.301 0.2224 330 -1.35 1 EAOCa US205 OW US205 -1.213 0.2224 330 -5.45 0.0078 EAOCa US205 TYES US205 -1.6021 0.2224 330 -7.2 <.0001 OW US205 TYES US205 -0.3891 0.2224 330 -1.75 1 EAOCa US215 OW US215 -0.8495 0.2224 330 -3.82 0.8924 EAOCa US215 TYES US215 -0.8495 0.2224 330 -3.82 0.8924 OW US215 TYES US215 -3.55E-15 0.2224 330 0 1 EAOCa US217 OW US217 -1.6109 0.2224 330 -7.24 <.0001 EAOCa US217 TYES US217 -1.8495 0.2224 330 -8.32 <.0001 OW US217 TYES US217 -0.2386 0.2224 330 -1.07 1 EAOCa US218 OW US218 -0.7557 0.2224 330 -3.4 0.9962 EAOCa US218 TYES US218 -0.4062 0.2224 330 -1.83 1 OW US218 TYES US218 0.3495 0.2224 330 1.57 1 EAOCa US225 OW US225 -2.2297 0.2224 330 -10.03 <.0001 EAOCa US225 TYES US225 -2.3495 0.2224 330 -10.56 <.0001 OW US225 TYES US225 -0.1198 0.2224 330 -0.54 1 EAOCa US226 OW US226 -0.787 0.2224 330 -3.54 0.9845 96 Table 2.6. (cont’d) Medium 1 Isolate 1 Medium 2 Isolate 2 Estimate SE DF t-value P-value EAOCa US226 TYES US226 -1.1594 0.2224 330 -5.21 0.0229 OW US226 TYES US226 -0.3724 0.2224 330 -1.67 1 EAOCa US232 OW US232 -2.1233 0.2224 330 -9.55 <.0001 EAOCa US232 TYES US232 -2.3363 0.2224 330 -10.51 <.0001 OW US232 TYES US232 -0.213 0.2224 330 -0.96 1 EAOCa US236 OW US236 0.3141 0.2224 330 1.41 1 EAOCa US236 TYES US236 -0.3717 0.2224 330 -1.67 1 OW US236 TYES US236 -0.6858 0.2224 330 -3.08 1 EAOCa US241 OW US241 -2.6184 0.2224 330 -11.77 <.0001 EAOCa US241 TYES US241 -2.588 0.2224 330 -11.64 <.0001 OW US241 TYES US241 0.03035 0.2224 330 0.14 1 EAOCa US244 OW US244 -1.0625 0.2224 330 -4.78 0.1229 EAOCa US244 TYES US244 -1.6734 0.2224 330 -7.52 <.0001 OW US244 TYES US244 -0.6109 0.2224 330 -2.75 1 EAOCa US245 OW US245 0.4287 0.2224 330 1.93 1 EAOCa US245 TYES US245 -0.04358 0.2224 330 -0.2 1 OW US245 TYES US245 -0.4722 0.2224 330 -2.12 1 EAOCa US249 OW US249 0.8662 0.2224 330 3.89 0.8445 EAOCa US249 TYES US249 -0.106 0.2224 330 -0.48 1 OW US249 TYES US249 -0.9722 0.2224 330 -4.37 0.3982 EAOCa US250 OW US250 -3.8116 0.2224 330 -17.14 <.0001 EAOCa US250 TYES US250 -3.9621 0.2224 330 -17.82 <.0001 OW US250 TYES US250 -0.1505 0.2224 330 -0.68 1 EAOCa US251 OW US251 -0.3251 0.2224 330 -1.46 1 EAOCa US251 TYES US251 -1.0485 0.2224 330 -4.71 0.1517 OW US251 TYES US251 -0.7234 0.2224 330 -3.25 0.9994 EAOCa US252 OW US252 0.5195 0.2224 330 2.34 1 EAOCa US252 TYES US252 0.4225 0.2224 330 1.9 1 OW US252 TYES US252 -0.09691 0.2224 330 -0.44 1 EAOCa US253 OW US253 -0.08805 0.2224 330 -0.4 1 EAOCa US253 TYES US253 -0.199 0.2224 330 -0.89 1 OW US253 TYES US253 -0.1109 0.2224 330 -0.5 1 EAOCa US254 OW US254 2 0.2224 330 8.99 <.0001 EAOCa US254 TYES US254 0.6109 0.2224 330 2.75 1 OW US254 TYES US254 -1.3891 0.2224 330 -6.25 0.0001 EAOCa US255 OW US255 0.04846 0.2224 330 0.22 1 EAOCa US255 TYES US255 -0.1672 0.2224 330 -0.75 1 OW US255 TYES US255 -0.2157 0.2224 330 -0.97 1 EAOCa US256 OW US256 -0.8099 0.2224 330 -3.64 0.9646 EAOCa US256 TYES US256 -0.5713 0.2224 330 -2.57 1 OW US256 TYES US256 0.2386 0.2224 330 1.07 1 EAOCa US259 OW US259 -1.1505 0.2224 330 -5.17 0.0271 EAOCa US259 TYES US259 -1.4771 0.2224 330 -6.64 <.0001 OW US259 TYES US259 -0.3266 0.2224 330 -1.47 1 EAOCa US260 OW US260 -2.0619 0.2224 330 -9.27 <.0001 EAOCa US260 TYES US260 -2.485 0.2224 330 -11.17 <.0001 OW US260 TYES US260 -0.4231 0.2224 330 -1.9 1 EAOCa US261 OW US261 0.7157 0.2224 330 3.22 0.9996 EAOCa US261 TYES US261 2.22E-15 0.2224 330 0 1 OW US261 TYES US261 -0.7157 0.2224 330 -3.22 0.9996 97 Table 2.6. (cont’d) Medium 1 Isolate 1 Medium 2 Isolate 2 Estimate SE DF t-value P-value EAOCa US265 OW US265 -0.713 0.2224 330 -3.21 0.9997 EAOCa US265 TYES US265 -1.2386 0.2224 330 -5.57 0.0046 OW US265 TYES US265 -0.5256 0.2224 330 -2.36 1 EAOCa US277 OW US277 -0.02558 0.2224 330 -0.12 1 EAOCa US277 TYES US277 -0.3205 0.2224 330 -1.44 1 OW US277 TYES US277 -0.2949 0.2224 330 -1.33 1 EAOCa US283 OW US283 -2.1505 0.2224 330 -9.67 <.0001 EAOCa US283 TYES US283 -3.1505 0.2224 330 -14.17 <.0001 OW US283 TYES US283 -1 0.2224 330 -4.5 0.2909 EAOCa US305 OW US305 -0.1127 0.2224 330 -0.51 1 EAOCa US305 TYES US305 -0.3513 0.2224 330 -1.58 1 OW US305 TYES US305 -0.2386 0.2224 330 -1.07 1 EAOCa US310 OW US310 0.08805 0.2224 330 0.4 1 EAOCa US310 TYES US310 -0.6109 0.2224 330 -2.75 1 OW US310 TYES US310 -0.699 0.2224 330 -3.14 0.9999 EAOCa US312 OW US312 -0.7793 0.2224 330 -3.5 0.9887 EAOCa US312 TYES US312 -1.8155 0.2224 330 -8.16 <.0001 OW US312 TYES US312 -1.0363 0.2224 330 -4.66 0.1807 EAOCa US314 OW US314 -1.5396 0.2224 330 -6.92 <.0001 EAOCa US314 TYES US314 -1.9031 0.2224 330 -8.56 <.0001 OW US314 TYES US314 -0.3635 0.2224 330 -1.63 1 EAOCa US323 OW US323 -1.0485 0.2224 330 -4.71 0.1517 EAOCa US323 TYES US323 -0.8979 0.2224 330 -4.04 0.7273 OW US323 TYES US323 0.1505 0.2224 330 0.68 1 EAOCa US324 OW US324 -1.3495 0.2224 330 -6.07 0.0004 EAOCa US324 TYES US324 -1.3495 0.2224 330 -6.07 0.0004 OW US324 TYES US324 -1.50E-15 0.2224 330 0 1 EAOCa US325 OW US325 -0.4287 0.2224 330 -1.93 1 EAOCa US325 TYES US325 -0.7157 0.2224 330 -3.22 0.9996 OW US325 TYES US325 -0.287 0.2224 330 -1.29 1 EAOCa US328 OW US328 -0.6761 0.2224 330 -3.04 1 EAOCa US328 TYES US328 -1.699 0.2224 330 -7.64 <.0001 OW US328 TYES US328 -1.0229 0.2224 330 -4.6 0.2173 EAOCa US329 OW US329 0.3099 0.2224 330 1.39 1 EAOCa US329 TYES US329 -0.6505 0.2224 330 -2.93 1 OW US329 TYES US329 -0.9604 0.2224 330 -4.32 0.4488 EAOCa US331 OW US331 -0.1505 0.2224 330 -0.68 1 EAOCa US331 TYES US331 -1 0.2224 330 -4.5 0.2909 OW US331 TYES US331 -0.8495 0.2224 330 -3.82 0.8924 EAOCa US343 OW US343 -2.8099 0.2224 330 -12.63 <.0001 EAOCa US343 TYES US343 -3.0485 0.2224 330 -13.71 <.0001 OW US343 TYES US343 -0.2386 0.2224 330 -1.07 1 EAOCa US351 OW US351 0.1505 0.2224 330 0.68 1 EAOCa US351 TYES US351 -0.2386 0.2224 330 -1.07 1 OW US351 TYES US351 -0.3891 0.2224 330 -1.75 1 EAOCa US352 OW US352 -0.9659 0.2224 330 -4.34 0.425 EAOCa US352 TYES US352 -0.8779 0.2224 330 -3.95 0.8051 OW US352 TYES US352 0.08805 0.2224 330 0.4 1 EAOCa US354 OW US354 -1.659 0.2224 330 -7.46 <.0001 EAOCa US354 TYES US354 -1.5938 0.2224 330 -7.17 <.0001 98 Table 2.6. (cont’d) Medium 1 Isolate 1 Medium 2 Isolate 2 Estimate SE DF t-value P-value OW US354 TYES US354 0.06527 0.2224 330 0.29 1 EAOCa US355 OW US355 -2.3406 0.2224 330 -10.52 <.0001 EAOCa US355 TYES US355 -2.272 0.2224 330 -10.22 <.0001 OW US355 TYES US355 0.06859 0.2224 330 0.31 1 EAOCa US356 OW US356 0.699 0.2224 330 3.14 0.9999 EAOCa US356 TYES US356 -0.301 0.2224 330 -1.35 1 OW US356 TYES US356 -1 0.2224 330 -4.5 0.2909 EAOCa US357 OW US357 -1.1901 0.2224 330 -5.35 0.0125 EAOCa US357 TYES US357 -1.5 0.2224 330 -6.74 <.0001 OW US357 TYES US357 -0.3099 0.2224 330 -1.39 1 EAOCa US359 OW US359 -0.5853 0.2224 330 -2.63 1 EAOCa US359 TYES US359 -0.4348 0.2224 330 -1.96 1 OW US359 TYES US359 0.1505 0.2224 330 0.68 1 EAOCa US361 OW US361 -0.4793 0.2224 330 -2.16 1 EAOCa US361 TYES US361 -0.6694 0.2224 330 -3.01 1 OW US361 TYES US361 -0.1901 0.2224 330 -0.85 1 EAOCa US364 OW US364 -1.199 0.2224 330 -5.39 0.0104 EAOCa US364 TYES US364 -2.1505 0.2224 330 -9.67 <.0001 OW US364 TYES US364 -0.9515 0.2224 330 -4.28 0.4881 EAOCa US367 OW US367 1 0.2224 330 4.5 0.2909 EAOCa US367 TYES US367 -0.199 0.2224 330 -0.89 1 OW US367 TYES US367 -1.199 0.2224 330 -5.39 0.0104 EAOCa US368 OW US368 -0.897 0.2224 330 -4.03 0.7313 EAOCa US368 TYES US368 -1.699 0.2224 330 -7.64 <.0001 OW US368 TYES US368 -0.802 0.2224 330 -3.61 0.973 EAOCa US372 OW US372 1.9272 0.2224 330 8.67 <.0001 EAOCa US372 TYES US372 0.6797 0.2224 330 3.06 1 OW US372 TYES US372 -1.2474 0.2224 330 -5.61 0.0038 EAOCa US374 OW US374 0.6901 0.2224 330 3.1 0.9999 EAOCa US374 TYES US374 0.03959 0.2224 330 0.18 1 OW US374 TYES US374 -0.6505 0.2224 330 -2.93 1 EAOCa US379 OW US379 -0.3995 0.2224 330 -1.8 1 EAOCa US379 TYES US379 -1.0728 0.2224 330 -4.82 0.1045 OW US379 TYES US379 -0.6734 0.2224 330 -3.03 1 EAOCa US380 OW US380 -1.1065 0.2224 330 -4.98 0.0599 EAOCa US380 TYES US380 -1.412 0.2224 330 -6.35 <.0001 OW US380 TYES US380 -0.3054 0.2224 330 -1.37 1 EAOCa US390 OW US390 -0.6734 0.2224 330 -3.03 1 EAOCa US390 TYES US390 -0.912 0.2224 330 -4.1 0.6672 OW US390 TYES US390 -0.2386 0.2224 330 -1.07 1 EAOCa US392 OW US392 0.2474 0.2224 330 1.11 1 EAOCa US392 TYES US392 -0.1505 0.2224 330 -0.68 1 OW US392 TYES US392 -0.3979 0.2224 330 -1.79 1 EAOCa US394 OW US394 0.3116 0.2224 330 1.4 1 EAOCa US394 TYES US394 -0.07745 0.2224 330 -0.35 1 OW US394 TYES US394 -0.3891 0.2224 330 -1.75 1 EAOCa US399 OW US399 -0.01609 0.2224 330 -0.07 1 EAOCa US399 TYES US399 -0.443 0.2224 330 -1.99 1 OW US399 TYES US399 -0.4269 0.2224 330 -1.92 1 EAOCa US400 OW US400 1.1505 0.2224 330 5.17 0.0271 99 Table 2.6. (cont’d) Medium 1 Isolate 1 Medium 2 Isolate 2 Estimate SE DF t-value P-value EAOCa US400 TYES US400 0.2441 0.2224 330 1.1 1 OW US400 TYES US400 -0.9065 0.2224 330 -4.08 0.6912 EAOCa US403 OW US403 -3 0.2224 330 -13.49 <.0001 EAOCa US403 TYES US403 -3 0.2224 330 -13.49 <.0001 OW US403 TYES US403 1.44E-15 0.2224 330 0 1 EAOCa US404 OW US404 -1.3495 0.2224 330 -6.07 0.0004 EAOCa US404 TYES US404 -3.9515 0.2224 330 -17.77 <.0001 OW US404 TYES US404 -2.6021 0.2224 330 -11.7 <.0001 EAOCa US411 OW US411 -2.5018 0.2224 330 -11.25 <.0001 EAOCa US411 TYES US411 -2.4604 0.2224 330 -11.06 <.0001 OW US411 TYES US411 0.04139 0.2224 330 0.19 1 EAOCa US413 OW US413 -0.199 0.2224 330 -0.89 1 EAOCa US413 TYES US413 -0.9756 0.2224 330 -4.39 0.3843 OW US413 TYES US413 -0.7766 0.2224 330 -3.49 0.9899 EAOCa US414 OW US414 -1.228 0.2224 330 -5.52 0.0057 EAOCa US414 TYES US414 -1.529 0.2224 330 -6.88 <.0001 OW US414 TYES US414 -0.301 0.2224 330 -1.35 1 EAOCa US439 OW US439 -0.6795 0.2224 330 -3.06 1 EAOCa US439 TYES US439 -1.228 0.2224 330 -5.52 0.0057 OW US439 TYES US439 -0.5485 0.2224 330 -2.47 1 EAOCa US442 OW US442 -1.23 0.2224 330 -5.53 0.0055 EAOCa US442 TYES US442 -1.23 0.2224 330 -5.53 0.0055 OW US442 TYES US442 3.89E-16 0.2224 330 0 1 EAOCa US443 OW US443 -3.4096 0.2224 330 -15.33 <.0001 EAOCa US443 TYES US443 -2.9324 0.2224 330 -13.19 <.0001 OW US443 TYES US443 0.4771 0.2224 330 2.15 1 EAOCa US444 OW US444 -0.156 0.2224 330 -0.7 1 EAOCa US444 TYES US444 -0.5935 0.2224 330 -2.67 1 OW US444 TYES US444 -0.4375 0.2224 330 -1.97 1 EAOCa US445 OW US445 -0.3724 0.2224 330 -1.67 1 EAOCa US445 TYES US445 -0.383 0.2224 330 -1.72 1 OW US445 TYES US445 -0.01059 0.2224 330 -0.05 1 EAOCa US449 OW US449 -2.2386 0.2224 330 -10.07 <.0001 EAOCa US449 TYES US449 -2.7157 0.2224 330 -12.21 <.0001 OW US449 TYES US449 -0.4771 0.2224 330 -2.15 1 EAOCa US450 OW US450 -0.699 0.2224 330 -3.14 0.9999 EAOCa US450 TYES US450 -1.0485 0.2224 330 -4.71 0.1517 OW US450 TYES US450 -0.3495 0.2224 330 -1.57 1 EAOCa US454 OW US454 -0.4287 0.2224 330 -1.93 1 EAOCa US454 TYES US454 -0.5 0.2224 330 -2.25 1 OW US454 TYES US454 -0.07133 0.2224 330 -0.32 1 EAOCa US455 OW US455 -1.6734 0.2224 330 -7.52 <.0001 EAOCa US455 TYES US455 -1.5207 0.2224 330 -6.84 <.0001 OW US455 TYES US455 0.1527 0.2224 330 0.69 1 EAOCa US458 OW US458 -1.3363 0.2224 330 -6.01 0.0005 EAOCa US458 TYES US458 -1.9472 0.2224 330 -8.76 <.0001 OW US458 TYES US458 -0.6109 0.2224 330 -2.75 1 EAOCa US459 OW US459 -0.699 0.2224 330 -3.14 0.9999 EAOCa US459 TYES US459 -1.1215 0.2224 330 -5.04 0.046 OW US459 TYES US459 -0.4225 0.2224 330 -1.9 1 100 Table 2.6. (cont’d) Medium 1 Isolate 1 Medium 2 Isolate 2 Estimate SE DF t-value P-value EAOCa US460 OW US460 -0.9921 0.2224 330 -4.46 0.3196 EAOCa US460 TYES US460 -1.4147 0.2224 330 -6.36 <.0001 OW US460 TYES US460 -0.4225 0.2224 330 -1.9 1 EAOCa US461 OW US461 -1 0.2224 330 -4.5 0.2909 EAOCa US461 TYES US461 -1.1505 0.2224 330 -5.17 0.0271 OW US461 TYES US461 -0.1505 0.2224 330 -0.68 1 EAOCa US465 OW US465 -1.642 0.2224 330 -7.38 <.0001 EAOCa US465 TYES US465 -1.841 0.2224 330 -8.28 <.0001 OW US465 TYES US465 -0.199 0.2224 330 -0.89 1 EAOCa US476 OW US476 -2 0.2224 330 -8.99 <.0001 EAOCa US476 TYES US476 -3 0.2224 330 -13.49 <.0001 OW US476 TYES US476 -1 0.2224 330 -4.5 0.2909 EAOCa US478 OW US478 -0.6505 0.2224 330 -2.93 1 EAOCa US478 TYES US478 -1.1505 0.2224 330 -5.17 0.0271 OW US478 TYES US478 -0.5 0.2224 330 -2.25 1 EAOCa US479 OW US479 -0.3239 0.2224 330 -1.46 1 EAOCa US479 TYES US479 -0.5 0.2224 330 -2.25 1 OW US479 TYES US479 -0.1761 0.2224 330 -0.79 1 EAOCa US485 OW US485 -2.1901 0.2224 330 -9.85 <.0001 EAOCa US485 TYES US485 -1.8891 0.2224 330 -8.49 <.0001 OW US485 TYES US485 0.301 0.2224 330 1.35 1 EAOCa US487 OW US487 -0.199 0.2224 330 -0.89 1 EAOCa US487 TYES US487 2.801 0.2224 330 12.6 <.0001 OW US487 TYES US487 3 0.2224 330 13.49 <.0001 EAOCa US488 OW US488 -1.8266 0.2224 330 -8.21 <.0001 EAOCa US488 TYES US488 -2.6505 0.2224 330 -11.92 <.0001 OW US488 TYES US488 -0.8239 0.2224 330 -3.7 0.9453 EAOCa US489 OW US489 -1.1858 0.2224 330 -5.33 0.0136 EAOCa US489 TYES US489 -0.857 0.2224 330 -3.85 0.8721 OW US489 TYES US489 0.3288 0.2224 330 1.48 1 EAOCa US490 OW US490 -3.088 0.2224 330 -13.89 <.0001 EAOCa US490 TYES US490 -3.2849 0.2224 330 -14.77 <.0001 OW US490 TYES US490 -0.1969 0.2224 330 -0.89 1 EAOCa US493 OW US493 -1.1338 0.2224 330 -5.1 0.0369 EAOCa US493 TYES US493 -1.1338 0.2224 330 -5.1 0.0369 OW US493 TYES US493 4.00E-15 0.2224 330 0 1 EAOCa US502 OW US502 -0.8363 0.2224 330 -3.76 0.9226 EAOCa US502 TYES US502 -0.7343 0.2224 330 -3.3 0.9988 OW US502 TYES US502 0.1021 0.2224 330 0.46 1 EAOCa US505 OW US505 -2.0669 0.2224 330 -9.29 <.0001 EAOCa US505 TYES US505 -1.8239 0.2224 330 -8.2 <.0001 OW US505 TYES US505 0.243 0.2224 330 1.09 1 EAOCa US508 OW US508 -2.0969 0.2224 330 -9.43 <.0001 EAOCa US508 TYES US508 -2.3979 0.2224 330 -10.78 <.0001 OW US508 TYES US508 -0.301 0.2224 330 -1.35 1 101 Table 2.7. Medium component effect ± standard error (SE) on Flavobacterium psychrophilum recovery resulting from Plackett-Burman experiment as well as t-value and p-value. The symbol, *, indicates a medium component had a significant effect on F. psychrophilum recovery (a = 0.05). Medium Component Effect SE t-value P-value Tryptone -0.0595 -0.0298 1.658 0.0978 Yeast extract -0.0073 -0.0036 0.203 0.8392 CaCl2 • 2H2O 0.3199 0.1600 8.916 <0.0001* MgSO4 • 7H2O 0.3051 0.1525 8.503 <0.0001* Beef extract -0.0399 -0.0199 1.112 0.2666 Casamino acids -0.4181 -0.2091 11.65 <0.0001* Tryptose -0.3488 -0.1744 9.721 <0.0001* L-aspartic acid 0.0618 0.0309 1.722 0.0856 Horse serum -0.0121 -0.0061 0.337 0.7362 Fetal bovine serum -0.2933 -0.1466 8.174 <0.0001* O. mykiss blood -0.0352 -0.0176 0.981 0.3270 102 Figure 2.1. Map of Michigan with Flavobacterium psychrophilum surveillance locations and salmonid species sampled (ATS = Atlantic salmon, Salmo salar; CHS = Chinook salmon, Oncorhynchus tshawytscha; COS = Coho salmon, O. kisutch; STT = Steelhead trout, O. mykiss). 103 Figure 2.2. Normal probability plot of the effect of different factors (i.e., medium components) on Flavobacterium psychrophilum growth (measured in colony forming units) in the Plackett- Burman design. Factors with a square symbol significantly affected F. psychrophilum recovery, whereas those with a circular symbol did not (a = 0.05). Factors to the left of the line improved F. psychrophilum recovery when incorporated into media at their low concentration, whereas factors to the right of the line improved recovery when incorporated at their high concentration. 104 Figure 2.3. Pareto chart of the effect of different factors (i.e., medium components) on Flavobacterium psychrophilum growth (measured in colony forming units) in the Plackett- Burman design. The horizontal bars represent the ratio between the effects of the medium components and their standard error. The bars are ordered according to effect size, with the greatest effect on F. psychrophilum growth on top. The vertical dashed line illustrates the critical t value (a = 0.05), meaning medium components crossing the line had a significant effect on F. psychrophilum growth. 105 Figure 2.4. Flavobacterium psychrophilum infection prevalence among spawning phase salmonid (CHS, Chinook salmon, Oncorhynchus tshawytscha; ATS, Atlantic salmon, Salmo salar; COS, Coho salmon, O. kisutch; STT, Steelhead trout, O. mykiss) broodstock sampled at four Michigan gamete collection locations (Little Manistee River Weir, LMRW; St. Mary’s River, SMR; Platte River Weir, PRW; Swan River Weir, SRW) according to three F. psychrophilum media (F. psychrophilum medium-A, FPM-A; F. psychrophilum medium-B, FPM-B; tryptone yeast extract salts medium, TYES). 106 Chapter 3: Immersion challenge of three salmonid species (Family Salmonidae) with three heterologous Flavobacterium psychrophilum multilocus sequence typing variants provides further evidence of differential host-specificity 107 3.1. Abstract Bacterial coldwater disease (BCWD), caused by Flavobacterium psychrophilum, results in globally significant losses amongst multiple salmonid (Family Salmonidae) species. Molecular epidemiology and serotyping studies have suggested that some variants are host specific; however, these associations have not been evaluated by cross-challenging fish species with putatively host-associated F. psychrophilum isolates via more natural (i.e., immersion) exposure routes. To this end, F. psychrophilum isolates US19-COS, US62-ATS, and US87-RBT, each originally recovered from diseased coho salmon (Oncorhynchus kisutch), Atlantic salmon (Salmo salar), or rainbow trout (O. mykiss), and belonging to a host-associated multilocus sequence typing clonal complex (CC; e.g., CC-ST9, CC-ST232, or CC-ST10), were PCR- serotyped, evaluated for proteolytic activity, and challenged against 4-month old Atlantic salmon, coho salmon, and rainbow trout via immersion. Findings showed US87-RBT caused disease and mortality only in rainbow trout (e.g., 56.7% survival probability vs. 100% in coho salmon and Atlantic salmon). US19-COS and US62-ATS caused more mortality in coho salmon and Atlantic salmon (i.e., their hosts of origin) but were also capable of causing disease in both other host species, albeit to a lesser extent. Observed differences in survival may be due to variant antigenic/virulence determinants as differences in serotype and proteolytic activity were discovered. Collectively, results highlight the intricacies of F. psychrophilum-host interactions and provide further in vivo evidence that some F. psychrophilum MLST variants are host specific, which may have implications for the development of BCWD prevention and control strategies. 108 3.2. Introduction Flavobacterium psychrophilum, causative agent of bacterial coldwater disease (BCWD) and rainbow trout fry syndrome (RTFS), causes substantial mortality and economic losses in farm and hatchery-reared salmonids (Family Salmonidae) worldwide (Loch and Faisal, 2017). Rainbow trout (Oncorhynchus mykiss) and coho salmon (O. kisutch) are considered most susceptible (Holt, 1987), particularly at early life stages when mortality is highest (e.g., 50-90%; Barnes and Brown, 2011). Likewise, BCWD epizootics in farmed Atlantic salmon (Salmo salar) are also common (Nilsen et al. 2011; Avendaño-Herrera et al. 2020; Macchia et al. 2022). Multilocus sequence typing (MLST) has become a widespread tool for molecular epidemiological studies on F. psychrophilum (Nicolas et al. 2008; Fujiwara-Nagata et al. 2013; Nilsen et al. 2014; Avendaño-Herrera et al. 2014; Li et al. 2021). To date, >1500 isolates recovered in 18 countries on five continents have been genotyped via MLST, revealing the existence of >260 distinct sequence types (STs; https://pubmlst.org/fpsychrophilum). A common finding amongst MLST-based epidemiological studies is that some F. psychrophilum MLST clonal complexes (CCs) are most commonly associated with a single host species (Fujiwara- Nagata et al. 2013; Avendaño-Herrera et al. 2014; Nilsen et al. 2014; Van Vliet et al. 2016; Knupp et al. 2019; Sebastiao et al. 2020; Li et al. 2021). For example, most F. psychrophilum isolates belonging to CC-ST10, which is the largest and most reported CC worldwide, are recovered from rainbow trout (Van Vliet et al. 2016; Nilsen et al. 2014; Avendaño-Herrera et al. 2014; Knupp et al. 2019; Li et al. 2021). Similarly, most F. psychrophilum isolates belonging to CC-ST9 are recovered from coho salmon, whereas all isolates in CC-ST232 have been recovered from Atlantic salmon (Fujiwara-Nagata et al. 2013; Avendaño-Herrera et al. 2014; Nilsen et al. 2014; Van Vliet et al. 2016; Knupp et al. 2019). 109 In addition to being genetically diverse, F. psychrophilum also exhibits serotypic diversity (Pacha, 1968; Holt, 1987; Lorenzen and Olesen, 1997). Although classical serotyping methods remain in use, a multiplex PCR-based serotyping assay was recently developed by Rochat et al. (2017), which detects four molecular serotypes (e.g., Type-0 – Type-3) and that was later amended by Avendaño-Herrera et al. (2020) to detect a fifth molecular serotype (e.g., Type- 4). This molecular serotyping assay has been applied to >350 F. psychrophilum isolates, and findings suggest some serotypes have host associations. For instance, F. psychrophilum isolates recovered from rainbow trout most commonly belong to Type-1 and Type-2, whereas isolates recovered from Atlantic salmon and coho salmon are most frequently associated with Type-2 and Type-4, or Type-0, respectively (Rochat et al. 2017; Sundell et al. 2019; Avendaño-Herrera et al. 2020; Li et al. 2021; Calvez et al. 2021; Knupp et al. 2021a). Likewise, F. psychrophilum isolates recovered from ayu (Plecoglossus altivelis) have been exclusively reported as Type-3 (Rochat et al. 2017). The putative host-associations of some F. psychrophilum geno- and sero-variants are largely based on observational associations in naturally infected fish. However, several studies have directly or indirectly investigated such associations under in vivo laboratory conditions (Holt, 1987; Ekman and Norrgren, 2003; Fredriksen et al. 2016), though most have used a less natural exposure route (e.g., injection) that bypasses important immune defenses (Fast et al. 2002; Dash et al. 2018). In contrast, at least one study reported the virulence of two host- associated F. psychrophilum variants (e.g., US19-coho salmon in ST9 and US53-rainbow trout in ST78) in coho salmon following laboratory immersion exposure (Knupp et al. 2021a). This study revealed that only the coho salmon-recovered isolate caused disease and mortality in coho salmon, whereas the rainbow trout-recovered isolate did not, despite proving virulent to rainbow 110 trout in a previous study (Knupp et al. 2021b). However, a study has yet to simultaneously cross- challenge multiple salmonid species of a similar age with multiple putatively host specific F. psychrophilum variants, thereby leaving a gap of knowledge in BCWD ecology with potentially important implications for future prevention and control strategies. To further investigate F. psychrophilum host specificity, in vitro and in vivo experiments were designed to elucidate the interactions between F. psychrophilum and three salmonid species. 3.3. Materials and Methods 3.3.1. Flavobacterium psychrophilum isolates Three F. psychrophilum isolates (e.g., US19, US62, and US87) belonging to three MLST STs (e.g., ST13, ST277, and ST275) within three MLST CCs (e.g., CC-ST9, CC-ST232, and CC-ST10) that are geographically widespread (i.e., detected on two to four continents; Fujiwara- Nagata et al. 2013; Nilsen et al. 2014; Avendaño-Herrera et al. 2014; Knupp et al. 2019; Li et al. 2021) and exclusively or nearly exclusively recovered from one of three economically important salmonid species (e.g., coho salmon, rainbow trout, or Atlantic salmon) were selected from this study. In addition, recent studies have demonstrated US19 is virulent to coho salmon via immersion, and US87 to rainbow trout via injection (Knupp et al. 2021b). 3.3.2. Molecular serotyping Like MLST ST, molecular serotype may also be indicative of some F. psychrophilum host associations (Rochat et al. 2017; Sundell et al. 2019; Avendaño-Herrera et al. 2020); therefore, the molecular serotypes of F. psychrophilum isolates US62-ATS and US87-RBT were determined using an adapted version of the mPCR-based serotyping approach as described by Knupp et al. (2021a). In brief, each 50-μl mPCR reaction contained 25 μl of 2X GoTaq® Green Master Mix (Promega), 20 ng of DNA template, 0.1 μM of each control primer, and 0.5 μM of 111 each primer targeting the molecular serotypes, with the remaining volume consisting of nuclease-free water. Sterile nuclease-free water functioned as a negative control, while F. psychrophilum isolates US19-COS, FP900406, CSF259-93, US104, and US515 acted as positive controls for Type-0, Type-1, Type-2, Type-3, and Type-4 respectively (Knupp et al. 2021; Loch and Knupp, unpublished). The mPCR cycling conditions outlined by Rochat et al. (2017) were employed using an Eppendorf Mastercycler pro thermal cycler. A 1.5% agarose gel containing 1X SYBR Safe DNA gel stain was used to separate 5 μl of the amplified PCR product via electrophoresis for 35 minutes at 100 V. A 1-Kb Plus DNA Ladder (ThermoFisher Scientific) served as the molecular size standard. The gel was examined under UV transillumination to estimate amplicon size and assign mPCR serotypes (e.g., Type-0, 188 bp; Type-1, 188 and 549 bp; Type-2, 188 and 841 bp; Type-3, 188 and 361 bp; and Type-4, 188 and 992 bp; Rochat et al., 2017; Avendaño-Herrera et al., 2020). 3.3.3. Characterization of proteolytic activity Flavobacterium psychrophilum can proteolyze multiple components (e.g., casein, gelatin, elastin, and collagen) representative of host connective and muscle tissue, and therefore proteolytic activity has been suggested as a virulence determinant (Madsen and Dalsgaard, 1998; Nakayama et al. 2016; Rochat et al. 2019; Knupp et al. 2021a). Moreover, proteolysis of some components (e.g., elastin and collagen) may be more commonly associated with F. psychrophilum isolates with different host associations (e.g., Nakayama et al. 2016; Rochat et al. 2019; Knupp et al. 2021a). Therefore, the proteolytic activities of US62-ATS and US87-RBT were assessed on tryptone yeast extract salt medium (TYES; Holt, 1987) supplemented with casein, elastin, or gelatin as described previously (Sundell et al. 2019; Knupp et al. 2021a). The proteolytic activity of US19-COS has been previously assessed on these substrates (Knupp et al. 112 2021) but was included for comparison purposes. Additionally, collagenase activity was assessed by supplementing TYES with 5% (w/v) collagen from bovine Achilles tendon (ThermoFisher). Briefly, F. psychrophilum was revived from cryostock (maintained at -80 °C) on TYES, which was modified according to Michel et al. (1999), incubated for 48 hours at 15 °C, and then visually inspected for purity. Each F. psychrophilum isolate was inoculated into 80 mL of analogous broth and incubated with constant shaking (180 rpm) for 48 hours at 15 °C. Bacteria were harvested via centrifugation (2,571 × g, 15 min) and adjusted to an optical density at 600 nm (OD600) corresponding to 1 × 109 cfu/mL using a spectrophotometer (WPA, Inc.) and sterile 0.65% saline. To quantify F. psychrophilum concentrations, serial dilutions in ten-fold increments (up to 100,000,000-fold) were plated on modified TYES in duplicate, and then incubated for seven days, after which final colony counts were performed. To determine proteolytic activity, 10 μL of each F. psychrophilum isolate was spotted in triplicate on the surface of the four media, allowed to dry, and then incubated for seven days at 15 °C. The colony and clear zone diameters were summed and then divided by the colony diameter to yield the clear zone ratio (CZR; Sundell et al. 2019). 3.3.4. In vivo virulence assessment of Flavobacterium psychrophilum isolates US19-COS, US62-ATS, and US87-RBT to Atlantic salmon, coho salmon, and rainbow trout 3.3.4.1. Origin of fish for challenge experiments Embryonated Atlantic salmon and rainbow trout eggs were sourced from a commercial egg distributor, while embryonated coho salmon eggs were procured from Platte River State Fish Hatchery. Coordination occurred so that all eggs from the three species arrived at the Michigan State University – University Research Containment Facility on the same day. In brief and upon receipt, eggs were disinfected with 100 ppm iodophor solution (pH 7.30) for 10 minutes before 113 being placed in a vertical incubator supplied with UV-treated, sand-filtered well water maintained at 12 ºC ± 1 ºC until hatching. Sac-fry were then moved to aerated flow-through tanks (40 L; 12 ºC ± 1 ºC) and, once exogenous feeding commenced, were given a continuous supply of appropriately sized commercial trout food (Skretting, the Netherlands) via an automatic feeder. After eight weeks, fish were hand-fed twice daily and the water volume in the tanks was increased (400 L; 12 ºC ± 1 ºC). Tanks were cleaned and siphoned daily to remove waste and any uneaten food. Before the challenge experiment, a sample of fish from each species were cultured to screen for bacterial infections (Knupp et al. 2021a), including those caused by F. psychrophilum, and confirmed to be bacterial infection-free. 3.3.4.2. Flavobacterium psychrophilum inoculum preparation for immersion challenge Flavobacterium psychrophilum isolates US19-COS, US62-ATS, and US87-RBT were revived from cryostock (maintained at -80 ºC) on Flavobacterium psychrophilum medium-A (FPM-A; Chapter 2), incubated for 48 hours at 15 ºC, and then visually inspected for purity. Each isolate was inoculated into 3 L of FPM-A broth, incubated, harvested, and adjusted to 109 cfu/mL using sterile 0.65% saline as described in section 3.3.3. F. psychrophilum concentrations were verified via plate counts. 3.3.4.3. Immersion challenge experiment The ability of Flavobacterium psychrophilum isolates US19-COS, US62-ATS, and US87-RBT to infect and cause disease in four-month-old Atlantic salmon (mean weight 1.1g), coho salmon (mean weight 5.0g), and rainbow trout (mean weight 7.8g) was assesed via immersion exposure. One-hundred and twenty Atlantic salmon, coho salmon, and rainbow trout were anesthetized in sodium bicarbonate-buffered (200 mg/L) tricaine methanesulfonate (MS- 222; Syndel) at a concentration of 100 mg/L, adipose fin-clipped using sharp sterile scissors 114 (Holt, 1987), and then allowed to recover in aerated water. Fish (n = 15, in duplicate) of each species (n = 3) were immersed for 30 min in aerated water (12 ± 1 ºC) containing 107 cfu/mL of US19-COS, US62-ATS, or US87-RBT. Control fish (n = 15, in duplicate) of each species (n = 3) were immersed in an identical volume of water only. After bacterial exposure, fish were transferred into aerated flow-through glass aquaria (37.8 L; n =15 fish per aquarium, in duplicate) supplied with ultraviolet light-treated, sand-filtered, pathogen-free well water (12 ± 1 ºC). Fish were monitored daily for 25 days and cared for as described in section 3.3.4.1; mortalities were necropsied and clinically examined, and multiple tissues (e.g., external ulcers and kidney) were bacteriologically analyzed for F. psychrophilum on FPM-A. Surviving fish (e.g., 25 days post-exposure) were euthanized via MS-222 overdose (250 mg/L) and analyzed similarly. All challenge experiments were conducted in accordance with the MSU-Institutional Animal Care and Use Committee (AUF:201900312). Representative isolates recovered from dead and surviving fish were confirmed as F. psychrophilum via endpoint PCR (Toyama et al. 1994; Van Vliet et al. 2015). Likewise, F. psychrophilum MLST STs of representative isolates were confirmed via PCR amplification and Sanger-sequencing of all seven F. psychrophilum-specific MLST loci as previously described (Knupp et al. 2019). 3.3.5. Data analysis The Kruskal-Wallis test was used to examine median CZR differences among isolates (e.g., US19-COS, US62-ATS, and US87-RBT) for the tested media (e.g., caseinase, collagenase, elastase, and gelatinase). If the null hypothesis of no difference in median CZR among isolates was rejected, pairwise comparisons of median CZR between isolates were carried out using 115 Dunn’s test and applying the Bonferroni correction for multiple comparisons (α = 0.05). The Kruskal-Wallis tests and Dunn’s tests were conducted using PROC npar1way and custom SAS code, respectively. Kaplan-Meier plots (Kaplan and Meier, 1958) with 95% confidence intervals were generated using PROC LIFETEST and SGPLOT to visualize Atlantic salmon, coho salmon, and rainbow trout survival probabilities over time after exposure to either US19-COS, US62-ATS, or US87-RBT. Relative risk of death among fish species exposed to each isolate was assessed using Cox proportional hazards regression models. Fish species and fish weight were treated as factor and continuous variables, respectively. Fish weight was included as a covariate as it can affect F. psychrophilum-induced mortality (Madsen and Dalsgaard, 1999). Comparisons of hazard ratios (i.e., risk of death) between fish species were evaluated for each isolate while accounting for the effect of fish weight. If mortality did not occur among experimental units (i.e., aquaria) for one or more fish species, the Cox proportional hazards regression model was replaced by pairwise comparisons of survival rate between fish species on day 25 (i.e., the end of the experiment) using two-sample z-tests. The Cox proportional hazards regression models and pairwise comparisons were conducted using PROC PHREG and custom code. All statistical analyses were performed using SAS® Version 9.4; (α = 0.05). 3.4. Results 3.4.1. Molecular serotype The three F. psychrophilum isolates each belonged to a different molecular serotype, whereby US19-COS, US62-ATS, and US87-RBT were identified as Type-0 (Knupp et al. 2021a), Type-1, and Type-2, respectively. 116 3.4.2. Proteolytic activity Flavobacterium psychrophilum isolates US19-COS, US62-ATS, and US87-RBT all proteolyzed casein, collagen, and gelatin; however, only US87-RBT proteolyzed elastin (Table 3.1). The Kruskal-Wallis test indicated that there were overall significant differences among the isolates in median CZR for elastase (χ2 = 8.00; df = 2; P-value = 0.0183) and gelatinase (χ2 = 7.71; df = 2; P-value = 0.0211). However, the null hypothesis of no difference in median CZR among the isolates for caseinase (χ2 = 5.84; df = 2; P-value = 0.0538) and collagenase (χ2 = 5.73; df = 2; P-value = 0.0571) could not be rejected (i.e., no significant difference in caseinase and collagenase activity among the isolates). For elastase, the median CZR produced by US87- RBT was significantly greater than US19-COS (Z-value = 3.4641; P-value = 0.0008) and US62- ATS (Z-value = 3.4641; P-value = 0.0008; Table 3.1). For gelatinase, the median CZR produced by US19-COS and US87-RBT were both significantly greater than US62-ATS (US19-COS vs. US62-ATS: Z-value = 2.3094, P-value = 0.0314; US87-RBT vs. US62-ATS: Z-value = 3.4641, P-value = 0.0008; Table 3.1). 3.4.3. Virulence of Flavobacterium psychrophilum isolates US19-COS, US62-ATS, and US87-RBT to Atlantic salmon, coho salmon, and rainbow trout 3.4.3.1. Negative control Atlantic salmon, coho salmon, and rainbow trout Throughout the course of the in vivo challenge experiments, no negative control fish died in any of the three fish species. 3.4.3.2. Flavobacterium psychrophilum isolate US87-RBT Following immersion exposure to F. psychrophilum isolate US87-RBT, rainbow trout was the only fish species to develop gross signs of BCWD. Disease signs were evident as early as nine days post-exposure in the form of focally extensive dermal ulceration of the caudal 117 peduncle that penetrated the underlying musculature (Figure 3.1A). As disease progressed, the caudal peduncle ulceration deepened further into the underlying musculature, exposing the vertebral column (Figure 3.1B). Rainbow trout also exhibited uni- or bilateral exophthalmia with or without intraocular ecchymosis (Figure 3.1C-D; Table 3.2) and/or gill pallor with or without ecchymosis and/or petechiae (Figure 3.1E). Internally, rainbow trout infected with US87-RBT presented with visceral organ (e.g., heart, liver, and kidney) pallor, multifocal hepatic ecchymoses (Figure 3.2A), severe splenic swelling with perisplenic hemorrhage (Figure 3.2B), and severe intestinal hemorrhage with accompanying peri-intestinal hemorrhage (Figure 3.2C; Table 3.2). In a subset of surviving rainbow trout (i.e., euthanized 25 days post-exposure), unilateral exophthalmia and caudal peduncle ulceration was present. Internally, most surviving rainbow trout had moderate to severe splenic swelling. Surviving Atlantic salmon and coho salmon remained apparently healthy. Overall survival was 56.7% in rainbow trout and 100% in Atlantic salmon and coho salmon (i.e., no Atlantic salmon or coho salmon died; Figure 3.3A). Rainbow trout mortality began nine days post-exposure and peaked on day 18 (Figure 3.3A). Because Atlantic salmon and coho salmon experienced no mortality, the Cox proportional hazards regression model could not be used to compare risk of death; however, the model did indicate that fish weight contributed significantly (P-value = 0.0036) to the risk of death among rainbow trout, whereby for each gram increase in weight, risk of death was reduced by 34.8% (hazard ratio = 0.652). Rainbow trout survival was significantly lower than both Atlantic salmon and coho salmon (Z- scores = 4.7897; P-values < 0.0001). 118 3.4.3.3. Flavobacterium psychrophilum isolate US19-COS Atlantic salmon, coho salmon, and rainbow trout exposed to US19-COS (ST13, CC-ST9) developed similar gross signs of BCWD as early as four days post-exposure in the form of focally extensive dermal ulceration of the caudal peduncle that penetrated into the underlying muscle; however, and early in the disease course, ulcerations were shallower (i.e., not exposing the vertebral column) in Atlantic salmon compared to coho salmon and rainbow trout (Figure 3.4A-C). In rainbow trout only, the tissues surrounding the ulcer frequently had diffuse ecchymoses and petechiae (Figure 3.4C; Table 3.2). As disease progressed, caudal peduncle ulcerations deepened in all species, including Atlantic salmon (Figure 3.4D). Another external BCWD sign common to all species was gill pallor with or without ecchymoses and petechiae. In contrast, peri-oral ulceration was apparent in coho salmon and rainbow trout only (Figure 3.4E- F). Internally, most gross disease signs caused by US19-COS were similar among the three species, which included visceral organ (e.g., heart, liver, and kidney) pallor and hepatic and renal ecchymoses (Figure 3.5A-E). However, splenic swelling and intestinal hemorrhage was present only among coho salmon and rainbow trout (Figure 3.5B-C, F; Table 3.2). Surviving Atlantic salmon, coho salmon, and rainbow trout did not exhibit gross external or internal BCWD signs. Overall, survival ranged from 6.7% in coho salmon to 50.0% to 53.3% in Atlantic salmon, and rainbow trout, respectively (Figure 3.3B). Mortality began three days post-exposure in all species, and peaked on days six, seven, and eight in Atlantic salmon, rainbow trout, and coho salmon, respectively (Figure 3.3B). The Cox proportional hazards regression model indicated fish species and weight significantly affected survival (fish species: Wald χ2 = 44.78, df = 2, P-value < 0.0001; fish weight: Wald χ2 = 31.03, df = 1, P-value < 0.0001); therefore, both variables were included in the model. As a result, coho salmon and rainbow trout were 119 significantly less likely (P-values < 0.0001) to survive than Atlantic salmon, whereas the risk of death among rainbow trout and coho salmon was not significantly different (P-value = 0.6102). For each gram increase in fish weight, the risk of death among all species decreased by 45.1% (hazard ratio = 0.549; P-value < 0.0001). 3.4.3.4. Flavobacterium psychrophilum isolate US62-ATS Atlantic salmon, rainbow trout, and coho salmon exposed to US62-ATS (ST277, CC- ST232) developed gross BCWD signs within two, three, or four days, respectively. All species exhibited focally extensive dermal ulceration of the caudal peduncle (Figure 3.6A-C). Like rainbow trout exposed to US19-COS, rainbow trout ulcers were often accompanied by surrounding and severe diffuse ecchymotic hemorrhage that extended posteriorly into the caudal fin (Figure 3.6B; Table 3.2). Atlantic salmon also had caudal fin ecchymoses (Figure 3.6D). In contrast, gross hemorrhage of the caudal peduncle/fin was not present in any coho salmon. Other external BCWD signs common to all fish species included gill pallor with or without ecchymoses and petechiae. A disease sign unique to coho salmon was severe peri-oral ulceration (Figure 3.6E). Internal BCWD signs caused by US62-ATS were identical among all species and included visceral organ (e.g., heart, liver, and kidney) pallor and mild splenic swelling (Table 3.2). Some surviving Atlantic salmon, coho salmon, and rainbow trout showed evidence of a prior caudal peduncle ulceration, evidenced by incomplete or complete healing. Internally, surviving fish of all species appeared grossly normal. Overall survival ranged from 3.3% to 13.3% in Atlantic salmon and rainbow trout, respectively, to 66.7% in coho salmon (Figure 3.3C). Mortality began one (Atlantic salmon), three (rainbow trout), and six (coho salmon) day(s) post-exposure and peaked on days four, eight, and 10 in Atlantic salmon, rainbow trout, and coho salmon, respectively (Figure 3.3C). 120 The Cox proportional hazards regression model indicated fish species and weight significantly affected survival (fish species: Wald χ2 = 42.73, df = 2, P-value < 0.0001; fish weight: Wald χ2 = 23.27, df = 1, P-value < 0.0001); therefore, both variables were included in the model. Atlantic salmon and rainbow trout were significantly less likely (P-values < 0.0001) to survive US62- ATS infection when compared to coho salmon. There was no significant difference in the risk of death among Atlantic salmon and rainbow trout (P-value = 0.5127). For each gram increase in fish weight, the risk of death among all species decreased by 34.1% (hazard ratio = 0.659; P- value < 0.0001). 3.4.4. Infection status in Atlantic salmon, coho salmon, and rainbow trout following immersion exposure to Flavobacterium psychrophilum isolates US19-COS, US62-ATS, and US87-RBT 3.4.4.1. Negative control Atlantic salmon, coho salmon, and rainbow trout No bacteria were recovered from any negative control fish throughout these experiments. 3.4.4.2. Flavobacterium psychrophilum isolate US87-RBT F. psychrophilum isolate US87-RBT was recovered in a pure form and as perfuse lawns (i.e., colony forming units, cfus, too numerous to count) from the caudal peduncle and kidney of all dead rainbow trout (n = 13). In surviving rainbow trout, pure cultures of US87-RBT were recovered from ~35.3% (n = 6/17) of the kidney cultures at intensities ranging from 100 – 101 cfu/g (as determined by calibrated inoculating loops and colony counts) of tissue. Molecular analyses confirmed the recovered bacteria were F. psychrophilum and belonged to ST275 (data not shown). 121 3.4.4.3. Flavobacterium psychrophilum isolate US19-COS F. psychrophilum isolate US19-COS was recovered in a pure form and as perfuse lawns from the caudal peduncle and kidney of all dead Atlantic salmon (n = 15), coho salmon (n = 28), and rainbow trout (n = 14). F. psychrophilum isolate US19-COS was not recovered from the external or internal tissues of any surviving fish. Molecular analyses confirmed the recovered bacteria were F. psychrophilum and belonged to ST13 (data not shown). 3.4.4.4. Flavobacterium psychrophilum isolate US62-ATS F. psychrophilum isolate US62-ATS was recovered in a pure form and as perfuse lawns from the caudal peduncle ulcerations of all dead Atlantic salmon (n = 29) and rainbow trout (n = 26). Similarly, US62-ATS was recovered at intensities ranging from 103 cfu/g to perfuse lawns from the kidney of all dead rainbow trout and most (e.g., n = 23/29, 79%) dead Atlantic salmon and in a pure form. In contrast, pure cultures of US62-ATS were obtained from the caudal peduncle ulcer of 40% (n = 4/10) of dead coho salmon at intensities ranging from 103 cfu/g to perfuse lawns. Internally, US62-ATS was recovered in a pure form from the kidney of all dead coho salmon (n = 10), with intensities ranging from 103 cfu/g to perfuse lawns. F. psychrophilum isolate US62-ATS was not recovered from external or internal tissues of any surviving fish. Molecular analyses confirmed the recovered bacteria were F. psychrophilum and belonged to ST277 (data not shown). 3.5. Discussion Herein, results provide evidence that some F. psychrophilum MLST variants are host specific, a matter that may affect the development of targeted BCWD prevention and control strategies. Indeed, US87-RBT (ST275, in CC-ST10) showed strong infection and disease-fidelity to rainbow trout, as evidenced by causing disease and subsequent mortality only in rainbow 122 trout, despite coho salmon and Atlantic salmon being at a significantly greater risk of death because of their smaller size. Knupp et al. (2021a) also suggested at least one other CC-ST10 variant (e.g., ST78) was rainbow trout-specific after proving avirulent to coho salmon following immersion challenge; however, this study did not also assess the virulence of this variant in rainbow trout. Likewise, Fredriksen et al. (2016) reported that a F. psychrophilum isolate highly virulent to rainbow trout was avirulent to Atlantic salmon via injection; however, the MLST genotype of this isolate was not reported. The in vivo findings from this study support the MLST-based observations that CC-ST10 appears to be rainbow trout-specific (Nicolas et al. 2008; Nilsen et al. 2014; Van Vliet et al. 2016; Knupp et al. 2019). In fact, of the 851 F. psychrophilum isolates recovered from fish and belonging to CC-ST10 (https://pubmlst.org/fpsychrophilum), >95% were recovered from O. mykiss. Interestingly, the CC-ST10 isolates that were recovered from other fish species (e.g., white sturgeon, Acipenser transmontanus; chinook salmon, O. tshawytscha; brown trout, S. trutta; coho salmon, Atlantic salmon, and Salvelinus sp.) may be the result of tightly interconnected fish farming practices. For example, most fish farms in Chile simultaneously rear Atlantic salmon, coho salmon, and rainbow trout (Avendaño-Herrera et al. 2020) and thus may partially explain why some CC- ST10 isolates were recovered from species other than rainbow trout. In addition to being host specific, US87-RBT was the only isolate recovered 25 days post infection (i.e., the end of the experiment), possibly suggesting it has evolved to circumvent the rainbow trout immune response. Like US87-RBT, US19-COS (ST13, in CC-ST9) and US62-ATS (ST277, in CC-ST232) also caused the most mortality in their host of origin (e.g., coho salmon and Atlantic salmon, respectively). However, and in contrast to US87-RBT, these F. psychrophilum isolates also 123 proved capable of causing disease and mortality in other salmonids, albeit to a lesser degree. Ekman and Norrgren (2003) noted similar findings, whereby an F. psychrophilum isolate (MLST variant unknown) recovered from Atlantic salmon caused the most mortality in Atlantic salmon but also caused mortality in rainbow trout and sea trout (S. trutta L.). Likewise, Holt (1987) found that although an F. psychrophilum isolate (e.g., SH3-81, in the same CC as US19-COS; Van Vliet et al. 2016) recovered from coho salmon caused the most mortality in coho salmon, it also caused mortality in Chinook salmon (O. tshawytscha) and rainbow trout. However, and notably, these previous studies were conducted via injection, which bypasses some host defenses (Fast et al. 2002; Dash et al. 2018) and thereby complicates assessment of the host specificity of the tested variants. Nevertheless, findings herein show that some F. psychrophilum variants may have a broader host range, which could have substantial implications for fish farms and hatcheries rearing multiple salmonid species. For instance, these facilities may be at greater risk for widespread losses in the event of a BCWD outbreak in comparison to an outbreak caused by a host specific variant (e.g., ST275). In this context, future studies assessing the transmission dynamics of these F. psychrophilum variants and salmonid species are warranted. Beyond the immediate risks, these findings may also affect the development of vaccines intended to protect multiple salmonid species. Although the primary focus of this study was not to extensively examine the mechanisms underlying F. psychrophilum host specificity, observations suggest O-polysaccharide antigenic determinants may play a role. In this context, US87-RBT belonged to Type-2 (i.e., serotype Th; Lorenzen and Olesen, 1997; Rochat et al. 2017) and was strongly host specific to rainbow trout. Knupp et al. (2021a) suggested a Type-2 F. psychrophilum variant (e.g., ST78) was rainbow trout specific after proving avirulent to coho salmon via immersion. Indeed, many Type-2/Th F. 124 psychrophilum isolates are virulent to rainbow trout and/or recovered from this species (Lorenzen and Olesen, 1997; Sundell et al. 2019; Avendaño-Herrera et al. 2020). US19-COS belonged to Type-0 (i.e., serotype FpT; Lorenzen and Olesen, 1997; Rochat et al. 2017; Knupp et al. 2021a) and caused the most mortality in its host of origin (e.g., coho salmon). These findings are consistent with previous studies, whereby most Type-0/FpT F. psychrophilum isolates are recovered from coho salmon (Lorenzen and Olesen, 1997; Rochat et al. 2017). Notably, the F. psychrophilum type strain (NCIMB 1947T) also belongs to Type-0/FpT (Lorenzen and Olesen, 1997; Rochat et al. 2017) and is considered avirulent to rainbow trout (Jarau et al. 2018; Madsen and Dalsgaard, 2000; Sundell et al. 2019); however, and given our findings, it seems NCIMB 1947T may not be well-suited to infect rainbow trout. US62-ATS belonged to Type-1 (i.e., serotype Fd; Lorenzen and Olesen, 1997), which contrasts with most previous studies showing Atlantic salmon isolates most often belong to Type-2 or Type-4 (Rochat et al. 2017; Avendaño- Herrera et al. 2020). Indeed, most F. psychrophilum isolates belonging to Type-1/Fd are recovered from rainbow trout (Lorenzen and Olesen, 1997; Rochat et al. 2017; Saticioglu et al. 2018; Avendaño-Herrera et al. 2020), and many are virulent to this species via injection (Sundell et al. 2019). Thus, our finding that US62-ATS belonged to Type-1 may partially explain the virulence of this isolate to not only Atlantic salmon but rainbow trout. Cisar et al. (2019) reported that serogroup, rather than serotype, more accurately defines Fd, Th, and FpT, and thus also applies to the molecular serotypes. Collectively, previous findings and observations herein suggest additional studies characterizing the serotypes of F. psychrophilum are needed, a matter that could impact BCWD vaccine development and selective breeding programs. Another mechanism potentially contributing to F. psychrophilum host specificity is proteolytic activity. Rainbow trout-specific isolate US87-RBT was the only tested isolate to 125 degrade elastin. Indeed, elastinolytic activity is common among isolates recovered from rainbow trout and belonging to CC-ST10 (Sundell et al. 2019). However, not all rainbow trout-recovered F. psychrophilum isolates possess this capability (Dalsgaard and Madsen, 2000; Soule et al. 2005; Sundell and Wiklund, 2015; Rochat et al. 2019). Thus, elastinolytic activity may only provide an advantage for some rainbow trout-associated isolates. The observation that US19- COS lacks elastinolytic activity is unsurprising given most F. psychrophilum isolates recovered from coho salmon and/or belonging to MLST CC-ST9 and/or serotype FpT lack this ability (Dalsgaard and Madsen, 2000; Soule et al. 2005; Rochat et al. 2019). Thus, other virulence determinants and/or proteases are sufficient for causing mortality in coho salmon. Indeed, Barbier et al. (2020) reported F. psychrophilum isolate OSU THCO2-90, which was recovered from diseased coho salmon and belongs to CC-ST9 (i.e., same CC as US19-COS; Nicolas et al. 2008), secretes at least 49 proteins, including multiple undescribed proteases. Studies assessing the elastinolytic activity of Atlantic salmon-recovered F. psychrophilum isolates have produced mixed results to date (Soule et al. 2005; Sundell and Wiklund, 2015; Rochat et al. 2019). Findings herein clearly show some Atlantic salmon-recovered F. psychrophilum isolates can cause high mortality without this ability. Whether the lack of this trait is common to most Atlantic salmon-associated F. psychrophilum isolates remains to be determined. This study emphasized the potential role of pathogen specificity. However, disease outcomes result from complex interactions between the pathogen, its host, and its environment (Casadevall and Pirofski, 1999), and previous studies have shown host genetics play a role in BCWD resistance. For example, Leeds et al. (2010) demonstrated via laboratory challenges that selective breeding was effective at increasing BCWD resistance among rainbow trout. Moreover, multiple quantitative trait loci associated with BCWD resistance in rainbow trout have been 126 identified (Wiens et al. 2013; Vallejo et al. 2014; Palti et al. 2015). Host genetics may contribute to BCWD resistance via differences in immune response. For example, Lee et al. (2023) found that in comparison to a BCWD-susceptible rainbow trout line, a BCWD-resistant rainbow trout line had increased expression of M2 macrophages involved anti-inflammatory responses and tissue repair, and two Toll-like receptors responsible for pathogen detection and inflammatory response. Nagai and Nakai (2011) found ayu-recovered F. psychrophilum isolates could survive and grow in ayu serum, whereas isolates recovered from salmonids and cyprinids could not. Herein, rainbow trout was the only species with severe diffuse ecchymotic hemorrhage surrounding the caudal peduncle and exophthalmia. Whether the observed differences in diseases signs among species following exposure to US19-COS, US62-ATS, and US87-RBT is pathogen and/or host-derived remains to be determined. In conclusion, we confirmed the MLST-based observations that some F. psychrophilum variants are host-specific, whereas others appear more generalistic. We posit the mechanisms driving these disparities are multifaceted, potentially influenced by not only F. psychrophilum serotype and secreted proteases but also host genetics and corresponding immune response. The implications of these findings are broad and may affect F. psychrophilum transmission dynamics, and the development of effective BCWD vaccines and BCWD-resistant salmonid lines. Thus, future studies evaluating F. psychrophilum host specificity, transmission, and the underlying mechanisms are warranted. 127 REFERENCES Avendaño-Herrera, R., Houel, A., Irgang, R., Bernardet J-F., Godoy, M., Nicolas, P., Duchaud, E. 2014. Introduction, expansion, and coexistence of epidemic Flavobacterium psychrophilum lineages in Chilean fish farms. Veterinary Microbiology. 170, 298-306. Avendano-Herrera, R., Tapia-Cammas, D., Duchaud, E., Irgang, R. 2020. Serological diversity in Flavobacterium psychrophilum: A critical update using isolates retrieved from Chilean salmon farms. Journal of Fish Diseases, 43, 877–888. Barbier, P., Rochat, T., Mohammed, H.H., Wiens, G.D., Bernardet, J.-F., Halpern, D., Duchaud, E., McBride, M.J. 2020. The type IX secretion system is required for virulence of the fish pathogen Flavobacterium psychrophilum. Applied and Environmental Microbiology. 86, e00799-20. Barnes, M.E., Brown, M.L. 2011. A review of Flavobacterium psychrophilum biology, clinical signs, and bacterial coldwater disease prevention and treatment. The Open Fish Science Journal. 4, 40-48. Calvez, S., Navarro-Gonzalez, N., Siekoula-Nguedia, C., Fournel, C., Duchaud, E. High genetic diversity in Flavobacterium psychrophilum isolates from healthy rainbow trout (Oncorhynchus mykiss) farmed in the same watershed, revealed by two typing methods. Applied and Environmental Microbiology. 87(2): e01398-20. Cisar, J.O., Bush, C.A., Wiens, G.D. 2019. Comparative structural and antigenic characterization of genetically distinct Flavobacterium psychrophilum O-polysaccharides. Frontiers in Microbiology. 10:1041. Casadevall, A., Pirofski, L.-A. 1999. Host-Pathogen interactions: redefining the basic concepts of virulence and pathogenicity. Infection and Immunity. 67(8), 3703-3713. Dalsgaard, I., Madsen, L. 2000. Bacterial pathogens in rainbow trout, Oncorhynchus mykiss (Walbaum), reared at Danish freshwater farms. Journal of Fish Diseases. 23(3), 199-209. Dash, S., Das, S.K., Samal, J., Thatoi, H.N. 2018. Epidermal mucus, a major determinant in fish health: a review. Iranian Journal of Veterinary Research. 19, 72-81. Ekman, E., Norrgren, L. 2003. Pathology and immunohistochemistry in three species of salmonids after experimental infection with Flavobacterium psychrophilum. Journal of Fish Diseases. 26, 529-538. Fast, M.D., Sims, D.E., Burka, J.F., Mustafa, A., Ross, N.W. 2002. Skin morphology and humoral non-specific defense parameters of mucus and plasma in rainbow trout, coho, and Atlantic salmon. Comparative Biochemistry and Physiology Part A: Molecular & Integrative Physiology. 132, 645-657. 128 Fredriksen, B.N., Furevik, A., Olesen, R.H., Gauthier, D., Mendoza, J., Norderhus, E.A. 2016. Virulence of Chilean field isolates of Flavobacterium psychrophilum in Atlantic salmon (Salmo salar L.) parr. Bulletin of the European Association of Fish Pathologists. 36, 67. Fujiwara-Nagata, E., Chantry-Darmon, C., Bernardet, J-F., Eguchi, M., Duchaud, E., Nicolas, P. 2013. Population structure of the fish pathogen Flavobacterium psychrophilum at whole-country and model river levels in Japan. Veterinary Research. 44, 34. Holt, R.A. 1987. PhD thesis. Cytophaga psychrophila, the causative agent of bacterial cold water disease in salmonid fish. Oregon State University, Corvallis, OR. Jarau, M., Natale, A.D., Huber, P.E., MacInnes, J.I., Lumsden, J.S. 2018. Virulence of Flavobacterium psychrophilum isolates in rainbow trout Oncorhynchus mykiss (Walbaum). Journal of Fish Diseases. 41, 1505–1514. Kaplan, E.L., Meier, P. 1958. Nonparametric estimation from incomplete observations. Journal of the American Statistical Association. 53(282), 457-481. Knupp, C., Wiens, G.D., Faisal, M., Call, D.R., Cain, K.D., Nicolas, P., Van Vliet, D., Yamashita, C., Ferguson, J.A., Meuninck, D., Hsu, H-M., Baker, B.B., Shen, L., Loch, T.P. 2019. Large-scale analysis of Flavobacterium psychrophilum multilocus sequence typing genotypes indicates that both newly identified and recurrent clonal complexes are associated with disease. Applied and Environmental Microbiology. 86, e02305-18. Knupp, C., Kiupel, M., Brenden, T.O., Loch, T.P. 2021a. Host-specific preference of some Flavobacterium psychrophilum multilocus sequence typing genotypes determines their ability to cause bacterial coldwater disease in coho salmon (Oncorhynchus kisutch). Journal of Fish Diseases. 44(5), 521-531. Knupp, C., Faisal, M., Brenden, T.O., Wiens, G.D., Loch, T.P. 2021b. In vivo experiments provide evidence that Flavobacterium psychrophilum strains belonging to multilocus sequence typing clonal complex ST191 are virulent to rainbow trout (Oncorhynchus mykiss). Journal of Aquatic Animal Health. 33(3), 190-195. Leeds, T.D., Silverstein, J.T., Weber, G.M., Vallejo, R.L., Palti, Y., Rexroad III., C.E., Evenhuis, J., Hadidi, S., Welch, T.J., Wiens, G.D. 2010. Response to selection for bacterial cold water disease resistance in rainbow trout. Journal of Animal Science. 88, 1936-1946. Li, S., Chai, J., Cao, Y., Knupp, C., Wang, D., Nicolas, P., Chen, F., Liu, H., Lu, T., Loch, T.P. 2021. Characterization and molecular epidemiological analysis of Flavobacterium psychrophilum recovered from diseased salmonids in China. Microbiology Spectrum. 9(2), e00033021. Loch, T.P., Faisal, M. Flavobacterium spp. In Fish viruses and bacteria: Pathobiology and protection., Woo, P.T.K., Cipriano, R.C., Eds., CABI: 2017, pp. 211-232. 129 Lorenzen, E., Olesen, N.J. 1997. Characterization of isolates of Flavobacterium psychrophilum associated with coldwater disease or rainbow trout fry syndrome II: serological studies. Diseases of Aquatic Organisms. 31, 209-220. Macchia, V., Inami, M., Ramstad, A., Grammes, F., Reeve, A., Moen, T., Torgersen, T.S., Adams, A., Desbois, A.P., Hoare, R. 2022. Immersion challenge model for Flavobacterium psychrophilum infection of Atlantic salmon (Salmo salar L.) fry. Journal of Fish Diseases. 45(11), 1781-1788. Madsen, L., Dalsgaard, I. Characterization of Flavobacterium psychrophilum; comparison of proteolytic activity and virulence of strains isolated from trout (Oncorhynchus mykiss) In: Barnes A.C., Davidson, G.A., Hiney, M.P., McIntosh, D., editors. Methodology in fish disease research. Aberdeen, Scotland: Fisheries Research Service, 1998. pp. 45-52. Madsen, L., Dalsgaard, I. 1999. Reproducible methods for experimental infection with Flavobacterium psychrophilum in rainbow trout Oncorhynchus mykiss. Diseases of Aquatic Organisms. 36, 169-176. Madsen, L., Dalsgaard, I. 2000. Comparative studies of Danish Flavobacterium psychrophilum isolates: Ribotypes, plasmid profiles, serotypes, and virulence. Journal of Fish Diseases. 23, 211–218. Michel, C., Antonio, D., & Hedrick, R. P. (1999). Production of viable cultures of Flavobacterium psychrophilum: approach and control. Research in Microbiology. 150, 351-358. Nagai, T., Nakai, T. 2011. Growth of Flavobacterium psychrophilum in fish serum correlates with pathogenicity. Journal of Fish Diseases. 34, 303–310. Nakayama, H., Tanaka, K., Teramura, N., Hattori, S. 2016. Expression of collagenase in Flavobacterium psychrophilum isolated from cold-water disease-affected ayu (Plecoglossus altivelis). Bioscience, Biotechnology, and Biochemistry. 80(1), 135-144. Nicolas, P., Mondot, S., Achaz, G., Bouchenot, C., Bernardet, J-F., Duchaud, E. 2008. Population structure of the fish-pathogenic bacterium Flavobacterium psychrophilum. Applied and Environmental Microbiology. 74, 3701-3709. Nilsen, H., Johansen, R., Colquhoun, D.J., Kaada, I., Bottolfsen, K., Vagnes, O., Olsen, A.B. 2011. Flavobacterium psychrophilum associated septicaemia and necrotic myositis in Atlantic salmon Salmo salar: a case report. Disease of Aquatic Organisms. 97(1), 37-46. Nilsen, H., Sundell, K., Duchaud, E., Nicolas, P., Dalsgaard, I., Madsen, L., Aspan, A., Jansson, E., Colquhoun, D.J., Wiklund, T. 2014. Multilocus sequence typing identifies epidemic clones of Flavobacterium psychrophilum in Nordic countries. Applied and Environmental Microbiology. 80, 2728 –2736. 130 Pacha, R.E. 1968. Characteristics of Cytophaga psychrophila (Borg) isolated during outbreaks of bacterial cold-water disease. Applied Microbiology. 16, 97-101. Palti, Y., Gao, G., Liu, S., Kent, M.P., Lien, S., Miller, M.R., Rexroad III, C.E., Moen, T. 2015. The development and characterization of a 57K single nucleotide polymorphism array for rainbow trout. Molecular Ecology Resources. 15(3). 662-672. Rochat, T., Fujiwara-Nagata, E., Calvez, S., Dalsgaard, I., Madsen, L., Calteau, A., Lunazzi, A., Nicolas, P., Wiklund, T., Bernardet, J-F., Duchaud, E. 2017. Genomic characterization of Flavobacterium psychrophilum serotypes and development of a multiplex PCR-based serotyping scheme. Frontiers in Microbiology. 8, 1752. Rochat, T., Pérez-Pascual, Nilsen, H., Carpentier, M., Bridel, S., Bernardet, J-F, Duchaud, E. 2019. Identification of a Novel Elastin-Degrading Enzyme from the Fish Pathogen Flavobacterium psychrophilum. Applied and Environmental Microbiology. 85, e02535-18. Saticioglu, I.B., Duman, M., Wiklund, T., Altun, S. 2018. Serological and genetic characterization of Flavobacterium psychrophilum isolated from farmed salmonids in Turkey. Journal of Fish Diseases. 41, 1899–1908. Sebastiao, F., Loch, T.P., Knupp, C., Mukkatira, K., Veek, T., Richey, C., Adkison, M., Griffin, M.J., Soto, E. 2020. Multilocus sequence typing (MLST) analysis of California Flavobacterium psychrophilum reveals novel genotypes and predominance of CC-ST10 in California salmonid hatcheries. Aquaculture Research. 51, 2349-2358. Soule, M., LaFrentz, S., Cain, K.D., LaPatra, S., Call, D.R. 2005. Polymorphisms in 16S rRNA genes of Flavobacterium psychrophilum correlate with elastin hydrolysis and tetracycline resistance. Diseases of Aquatic Organisms. 65, 209-216. Sundell, K., Landor, L., Nicolas, P., Jørgensen, J., Castillo, D., Middelboe, M., Dalsgaard, I., Donati, V. L., Madsen, L., Wiklund, T. 2019. Phenotypic and genetic predictors of pathogenicity and virulence in Flavobacterium psychrophilum. Frontiers in Microbiology. 10, 1711. Sundell, K., Wiklund, T. 2015. Characteristics of epidemic and sporadic Flavobacterium psychrophilum sequence types. Aquaculture. 441, 51-56. Toyama, T., Kita-Tsukamoto, K., Wakabayashi, H. 1994. Identification of Cytophaga psychrophila by PCR targeted 16S ribosomal RNA. Fish Pathology. 29, 271-275. Vallejo, R.L., Palti, Y., Liu, S., Marancik, D.P., Wiens, G.D. 2014. Validation of linked QTL for bacterial coldwater disease resistance and spleen size on rainbow trout chromosome Omy19. Aquaculture. 432, 139-143. Van Vliet, D., Loch, T.P., Faisal, M. 2015. Flavobacterium psychrophilum infections in salmonid broodstock and hatchery-propagated stocks of the Great Lakes Basin. Journal of Aquatic Animal Health. 27, 192–202. 131 Van Vliet, D., Wiens, G.D., Loch, T.P., Nicolas, P., Faisal, M. 2016. Genetic diversity of Flavobacterium psychrophilum isolates from three Oncorhynchus spp. in the United States, as revealed by multilocus sequence typing. Applied and Environmental Microbiology. 82, 3246- 3255. Wiens, G.D., Vallejo, R.L., Leeds, T.D., Palti, Y., Hadidi, S., Liu, S., Evenhuis, J.P., Welch, T.J., Rexroad III., C.E. 2013. Assessment of genetic correlation between bacterial cold water disease resistance and spleen index in a domesticated population of rainbow trout: identification of QTL on chromosome Omy19. PLoS One. 8, e75749. 132 APPENDIX Table 3.1. Flavobacterium psychrophilum isolates used in this study for in vivo challenge experiments against Atlantic salmon (Salmo salar), coho salmon (Oncorhynchus kisutch), and rainbow trout (O. mykiss) and for the assessment of proteolytic activity, which is presented as a ratio of the median clear zone diameter to the colony diameter (in mm) ± standard error (SE). Median clear zone ratios for a particular enzyme (e.g., caseinase, gelatinase, or elastinase) containing an identical symbol (e.g., *, **, ***) are not significantly different (a = 0.05) and a ratio of 1.00 ± 0.00 indicates no protease activity. Protease clear zone ratio ± SE a b Isolate ID Host of origin ST CC Caseinase Collagenase Gelatinase Elastinase US19 Coho salmon ST13 CC-ST9 3.00 (0.13)* 2.00 (0.13)* 2.31 (0.05)* 1.00 (0.00)* US62 Atlantic salmon ST277 CC-ST232 3.88 (0.30)* 2.00 (0.12)* 1.79 (0.00)** 1.00 (0.00)* US87 Rainbow trout ST275 CC-ST10 3.44 (0.00)* 3.44 (0.00)* 3.50 (0.00)*** 2.00 (0.00)** a Sequence type. b Clonal complex. 133 Table 3.2. Proportion of dead Atlantic salmon (Salmo salar), coho salmon (Oncorhynchus kisutch), and rainbow trout (O. mykiss) with a range of gross external and internal bacterial coldwater disease signs following exposure to Flavobacterium psychrophilum isolates US19-COS, US62-ATS, and US87-RBT. External disease signs Internal disease signs Hemorrhage Visceral surrounding caudal organ Splenic Intestinal Isolate Host speciesa peduncle ulcer Exophthalmia hemorrhage swelling hemorrhage Atlantic salmon 0/15 0/15 1/15 0/15 0/15 US19-COS Coho salmon 0/28 0/28 2/28 25/28 3/28 Rainbow trout 7/14 0/14 1/14 14/14 3/14 Atlantic salmon 0/29 0/29 0/29 27/29 0/29 US62-ATS Coho salmon 0/10 0/10 0/10 10/10 0/10 Rainbow trout 12/26 0/26 3/26 26/26 0/26 Atlantic salmon - - - - - US87-RBT Coho salmon - - - - - Rainbow trout 0/13 3/13 6/13 13/13 3/13 a All dead fish with caudal peduncle ulceration, gill pallor, and visceral organ pallor. 134 Figure 3.1. Gross external lesions in rainbow trout (O. mykiss) following immersion challenge with F. psychrophilum isolate US87-RBT. (A) Focally extensive ulceration of the caudal peduncle that penetrated the underlying musculature. (B) Focally extensive ulceration of the caudal peduncle that penetrated the underlying musculature and exposed the vertebral column. (C) Bilateral exophthalmia. (D) Unilateral exophthalmia with intraocular ecchymosis (arrow). (E) Gill pallor with ecchymoses and petechiae (arrow). 135 Figure 3.2. Gross internal lesions in rainbow trout (O. mykiss) following immersion challenge with F. psychrophilum isolate US87-RBT. (A) Pale liver with multifocal ecchymoses. (B) Severe splenic swelling with perisplenic hemorrhage. (C) Severe intestinal hemorrhage with accompanying peri-intestinal hemorrhage. 136 Figure 3.3. Kaplan-Meier survival probability curves of Atlantic salmon (Salmo salar; ATS), coho salmon (Oncorhynchus kisutch; COS), and rainbow trout (O. mykiss; RBT) over a 25-day period following immersion challenge with Flavobacterium psychrophilum isolates (A) US87- RBT, (B) US19-COS, and (C) US62-ATS. Shaded regions depict 95% confidence intervals. Lines with different symbols (e.g., *, **) indicate significant differences in survival (α = 0.05). 137 Figure 3.4. Gross external lesions in Atlantic salmon (Salmo salar), coho salmon (Oncorhynchus kisutch), and rainbow trout (O. mykiss) following immersion challenge with F. psychrophilum isolate US19-COS. (A) Atlantic salmon with shallow focally extensive dermal ulceration of the caudal peduncle. (B) Coho salmon with focally extensive dermal ulceration of the caudal peduncle that penetrated the underlying musculature. (C) Rainbow trout with focally extensive dermal ulceration of the caudal peduncle that penetrated the underlying musculature. Diffuse ecchymoses and petechiae surround the ulcer. (D) Atlantic salmon with focally extensive dermal ulceration of the caudal peduncle, exposing the vertebral column. (E) and (F) Peri-oral ulceration in coho salmon (E) and rainbow trout (F). 138 Figure 3.5. Gross internal lesions in Atlantic salmon (Salmo salar), coho salmon (Oncorhynchus kisutch), and rainbow trout (O. mykiss) following immersion challenge with F. psychrophilum isolate US19-COS. (A) Atlantic salmon with liver pallor. (B) Coho salmon with pale liver and multifocal ecchymoses, splenic swelling, and intestinal hemorrhage. (C) Rainbow trout with liver pallor and splenic swelling. (D) Coho salmon with renal pallor. (E) Rainbow trout with renal pallor and diffuse ecchymoses. (F) Rainbow trout with intestinal hemorrhage. 139 Figure 3.6. Gross external lesions in Atlantic salmon (Salmo salar), coho salmon (Oncorhynchus kisutch), and rainbow trout (O. mykiss) following immersion challenge with F. psychrophilum isolate US62-ATS. (A) Atlantic salmon with focally extensive dermal ulceration of the caudal peduncle that penetrated the underlying musculature. (B) Rainbow trout with multiple, focally extensive dermal ulcerations of the caudal peduncle that penetrated the underlying musculature. Ulceration is surrounded by severe diffuse ecchymoses that extends posteriorly into the caudal fin. (C) Coho salmon with focally extensive ulceration of the caudal peduncle that penetrated the underlying musculature. (D) Atlantic salmon with focal ecchymosis of the caudal fin. (E) Coho salmon with severe peri-oral ulceration. 140 Chapter 4: Varying Flavobacterium psychrophilum shedding dynamics in three bacterial coldwater disease-susceptible salmonid (Family Salmonidae) species 141 4.1. Abstract Flavobacterium psychrophilum causes bacterial coldwater disease (BCWD) and is responsible for substantial losses in farm and hatchery-reared salmonids (Family Salmonidae). Although F. psychrophilum infects multiple economically important salmonid species and is efficiently transmitted horizontally, the extent of knowledge regarding F. psychrophilum shedding rates and duration is limited to rainbow trout (Oncorhynchus mykiss). Concurrently, hundreds of F. psychrophilum sequence types (STs; i.e., genetic variants) have been described using multilocus sequence typing (MLST) and evidence suggests that some variants have distinct phenotypes, including differences in host associations. Whether shedding strategies differ amongst F. psychrophilum variants and/or salmonid hosts remains unknown. To this end, three F. psychrophilum isolates (e.g., US19, US62, and US87) in three MLST STs (e.g., ST13, ST277, and ST275) with apparent host associations for either coho salmon (O. kisutch), Atlantic salmon (Salmo salar), or rainbow trout, were intramuscularly injected into each respective fish species, thereby ensuring a consistent initial dosage of bacteria. Shedding rates of live and dead fish were determined at regular intervals by quantifying F. psychrophilum loads in water via quantitative PCR. Both live and dead Atlantic and coho salmon shed F. psychrophilum, as did live and dead rainbow trout. Regardless of salmonid species, dead fish shed F. psychrophilum at higher rates (e.g., up to ~108 – 1010 cells/fish/hour) compared to live fish (up to ~107 - 109 cells/fish/hour), and for a longer duration (5 – 35 days vs. 98 days); however, shedding dynamics varied by F. psychrophilum variant and/or host species, a matter that may complicate BCWD management. Findings herein expand knowledge on F. psychrophilum shedding dynamics across multiple salmonid species and can be used to inform future BCWD management strategies. 142 4.2. Introduction Flavobacterium psychrophilum, causative agent of bacterial coldwater disease (BCWD) and rainbow trout fry syndrome (RTFS), causes substantial losses in farmed and hatchery-reared salmonid species (Family Salmonidae; Loch and Faisal, 2017). Many salmonid species are BCWD-susceptible (Starliper, 2011), and although rainbow trout (Oncorhynchus mykiss) and coho salmon (O. kisutch) are generally considered most at risk (Holt, 1987), BCWD in Atlantic salmon (Salmo salar) is also common (Nilsen et al. 2011; Avendaño-Herrera et al. 2020; Macchia et al. 2022). Trout and salmon naturally infected by F. psychrophilum in their early life stages sustain substantial losses, with mortality rates ranging from 20 - 90% (Wood et al. 1974; Barnes and Brown, 2011). Although vertical transmission of F. psychrophilum plays an important role in the perpetuation of BCWD (Rangdale et al. 1996; Brown et al. 1997, Ekman 1999; Kumagai et al. 2001), horizontal transmission is also problematic (Starliper, 2011), facilitating the spread of F. psychrophilum within a population and exacerbating disease outbreaks. Shedding of F. psychrophilum from infected fish into the water column is a primary factor driving horizontal transmission (Madetoja et al. 2000; Madetoja et al. 2002). Madetoja et al. (2000) showed that live rainbow trout shed up to ~106 cells/fish/hour for up to 21 days. Notably, the same authors also showed that dead rainbow trout shed F. psychrophilum at even higher rates (e.g., up to ~108 cells/fish/hour), and for a longer duration (e.g., up to 59 days longer; Madetoja et al. 2000). Although these studies provided invaluable data related to F. psychrophilum shedding dynamics (e.g., time to shedding, shedding rate and duration) in rainbow trout, they represent the totality of knowledge on this subject. Thus, it remains to be 143 determined if F. psychrophilum shedding dynamics in other BCWD-susceptible host species, such as Atlantic salmon and coho salmon, are similar. In a similar context, whether F. psychrophilum shedding dynamics vary according to the F. psychrophilum multilocus sequence typing (MLST) variant causing the epizootic has yet to be determined. Multilocus sequencing typing of >1500 F. psychrophilum isolates has revealed >260 different sequence types (STs) worldwide (https://pubmlst.org/organisms/flavobacterium- psychrophilum), some of which differ in host species association (Nicolas et al. 2008; Knupp et al. 2019; Knupp et al. 2021a), virulence (Sundell et al. 2019; Knupp et al. 2021b), and prevalence in either hatchery-reared or wild/feral fish populations (Van Vliet et al. 2016; Knupp et al. 2019). Notably, most F. psychrophilum STs in North America differ from those that have been reported from other continents (e.g., Asia, Europe, South America; Nicolas et al. 2008; Siekoula-Nguedia et al. 2012; Fujiwara-Nagata et al. 2013; Strepparava et al. 2013; Nilsen et al. 2014; Avendaño-Herrera et al. 2014; Van Vliet et al. 2016; Knupp et al. 2019; Li et al. 2021), suggesting that the F. psychrophilum variant evaluated by Madetoja et al. (2000) is likely distinct from many causing losses in the USA. The detection and quantification of F. psychrophilum from water containing fish has been attempted via culture and immunofluorescence antibody technique (Madetoja et al. 2000; Madetoja and Wiklund, 2002; Madetoja et al. 2003), both of which are sensitive (e.g., detection limit of ~101 - 102 cfus/mL) but time consuming and vary in specificity. An alternative that is highly sensitive, specific, and well-suited for F. psychrophilum quantification in water is quantitative PCR (qPCR). Strepparava et al. (2014) developed an F. psychrophilum qPCR for this purpose, and although sensitive (detection limit of ~101 gene copies) and specific, its quantification limit was reportedly high (e.g., ~103 F. psychrophilum cells/mL). A more sensitive 144 F. psychrophilum-specific qPCR exists (detection limit of ~100 gene copies; Marancik and Wiens, 2013), but in the published literature, this assay has not been used to quantify F. psychrophilum loads in water. Towards improving our understanding of F. psychrophilum shedding dynamics in rainbow trout and, for the first time, Atlantic salmon and coho salmon, a series of experiments were devised to first optimize a previously developed F. psychrophilum-specific qPCR (Marancik and Wiens, 2013) for detection and quantification in water. Next, in vivo experiments were designed to elucidate F. psychrophilum shedding rates and durations in live and dead Atlantic salmon, coho salmon, and rainbow trout. Clarifying these aspects of BCWD ecology will offer insights into improved BCWD management strategies for multiple valuable salmonid species. 4.3. Materials and Methods 4.3.1. Marancik and Wiens (2013) qPCR 4.3.1.1. Reaction mixture and cycling parameters The F. psychrophilum-specific qPCR developed by Marancik and Wiens (2013) was performed according to protocol with some modification. Briefly, each 15-µL reaction mixture contained 7.5-µL of TaqMan® Environmental Master Mix 2.0, 0.67 µM of forward and reverse primers, 0.17 µM of TaqMan® probe, 0.60-µL of VetMAX™ Xeno™ Internal Positive Control (IPC) Assay, and 1 µL of template, with nuclease-free water comprising the remainder. Reactions were run in MicroAmp™ Optical 96-Well Fast Reaction Plates (0.1-mL) covered with MicroAmp™ Optical Adhesive Film. A QuantStudio™ 3 real-time thermal cycler (ThermoFisher Scientific) was used to amplify the 77-bp target amplicon according to the cycling program of Marancik and Wiens (2013). All consumables were purchased through 145 ThermoFisher Scientific except the primers, which were obtained from Integrated DNA Technologies (IDT). 4.3.1.2. Preparation of standards The qPCR target gene sequence was PCR amplified using previously extracted gDNA from F. psychrophilum isolate US53 (Van Vliet et al. 2016) and the same primers used for qPCR. Briefly, a 50-µL reaction mixture was prepared with 25-µL 2X GoTaq® Green Master Mix (Promega), 0.25 µM forward and reverse primers, 40 ng of US53 gDNA, with nuclease-free water comprising the remainder. A touchdown protocol consisting of initial denaturation for 2 min at 94 °C followed by 30 cycles of 94 °C for 1 min, 60 °C for 1 min, 72 °C for 1 min, and a final extension step at 72 °C for 7 min was performed using an Eppendorf™ Mastercycler™ pro conventional thermal cycler (ThermoFisher Scientific). The PCR product was run through a 1.5% agarose gel prepared with 1X SYBR™ Safe DNA gel stain (ThermoFisher Scientific) for 40 min at 100 V, after which the gel was viewed under UV transillumination to confirm the presence of an appropriately sized band. The PCR product (57.1% guanine-cytosine content; molecular weight of 37,497.9 g mol-1; Marancik and Wiens, 2013) was purified using the QIAquick® PCR Purification Kit (Qiagen) and then quantified using a Qubit™ fluorometer and the broad range Quant-iT™ dsDNA Assay Kit (Thermofisher Scientific). Gene copy concentration was calculated using a previously published formula (Standish et al. 2018): (ngDNA µL#$ ) × (6.022 × 10%& copies mol#$ ) Gene copies/µL = (37497.9 g mol#$ ) × (1 × 10' ng g #$ ) Serial 10-fold dilutions of purified amplicon were made over nine orders of magnitude (e.g., 108 copies/µL – 100 copies/µL) using 1X IDTE solution (pH 8.0; Integrated DNA Technologies) supplemented with 100 ng/µL tRNA (Yeast tRNA; Thermofisher Scientific). Two standard curve assays were performed, each on individual 96-well plates using eight replicate reactions of 146 each dilution (i.e., qPCR standard); intra- and inter-assay variation was determined using mean quantification cycle (Cq), Cq standard deviation, and coefficient of variation (CV). Assay efficiency, slope, and the correlation coefficient (R2) of each assay were also calculated; efficiency estimates between 90% and 110% were considered acceptable (Griffin et al. 2009; Griffin et al. 2011). These newly generated standards were used to determine F. psychrophilum DNA extraction efficiency, the limit of F. psychrophilum quantification and detection from spiked water (section 4.3.2.3), and to quantify F. psychrophilum loads in 50-mL water samples obtained during the in vivo shedding experiments (section 4.3.3.6). 4.3.2. Optimization of water sampling method and qPCR for Flavobacterium psychrophilum quantification from water 4.3.2.1. Preparation of mock water samples containing Flavobacterium psychrophilum Flavobacterium psychrophilum isolate US53 was revived from cryostock (maintained at - 80 ºC) on F. psychrophilum medium A (FPM-A; Chapter 2), incubated for 48 hours at 15 ºC, visually inspected for purity, inoculated into 250-mL of analogous broth, and incubated with constant shaking (180 rpm) for 48 hours at 15 ºC. Bacteria were harvested via centrifugation (2,571 × g, 15 minutes) and resuspended in 50-mL of ultraviolet light-treated, sand-filtered well water (i.e., the same water supplying the shedding experimental aquaria in section 4.3.3), which was then serially diluted up to 100,000,000-fold in ten-fold increments to create nine total mock water samples. To quantify bacteria in the most concentrated mock water sample, a 1-mL aliquot was serially diluted 100,000,000-fold in ten-fold increments and plated on FPM-A in duplicate, and then incubated for seven days, after which final colony counts were performed. All mock water samples were brought to a final volume of 50 mL to replicate the water sampling volume used during the shedding experiment. 147 4.3.2.2. Bacterial DNA extraction from water Each mock water sample (n = 9; section 4.3.2.1) was vacuum filtered through a single sterile piece of Whatman® qualitative filter paper (grade 4, 70 mm in diameter; Millipore Sigma) that had been placed in a 70-mm diameter Büchner funnel. The filter paper was removed from the Büchner funnel and placed into a sterile Petri dish; sterile forceps were used to first fold the paper in half (the side receiving the bacterial suspension was facing inward) and then loosely roll it into a cylindrical shape. The rolled paper was placed inside a PowerBead Pro tube of the DNeasy® PowerSoil® Pro Kit (Qiagen) along with 20,000 copies of Xeno™ IPC to monitor inhibition; DNA was then extracted according to the manufacturers’ protocol, resulting in 50-µL of eluted DNA per dilution. 4.3.2.3. DNA extraction efficiency and limit of quantification and detection The qPCR standards (section 4.3.1.2) were used to simultaneously determine (i.e., on one 96-well plate) F. psychrophilum DNA extraction efficiency and the limit of detection (LOD) and quantification (LOQ). DNA extraction efficiency was measured as it is highly variable (e.g., 0.2% – 108.9%; Lebuhn et al. 2004; Slana et al. 2008; Stoeckel et al. 2009; Kralik et al. 2011; Van Tongeren et al. 2011; Ricchi et al. 2016), thereby potentially affecting estimation of the target microorganism in a sample (e.g., F. psychrophilum in water) and therefore the LOD/LOQ (Kralik et al. 2011). Flavobacterium psychrophilum DNA extraction efficiency was calculated as the quotient of the mean qPCR-determined concentration of F. psychrophilum (in cells/mL) and the mean theoretical input (i.e., pre-DNA extraction) concentration of F. psychrophilum (Kralik and Ricchi et al. 2017). The input concentration of F. psychrophilum was considered theoretical, given that the most concentrated mock water sample alone was quantified via plate counts (section 4.3.2.1). 148 The median of the mean F. psychrophilum DNA extraction efficiency values was used as the universal DNA extraction efficiency (Kralik et al. 2011; Kralik and Ricchi et al. 2017); this number was used to apply a DNA extraction correction factor (DECF) to all qPCR-derived F. psychrophilum concentrations according to the formula: 100 DECF = Median of mean DNA extraction efficiency The LOD was determined by running 10 replicate reactions theoretically containing 100, 10, and 1 F. psychrophilum cell(s)/mL and defined as the lowest mean qPCR-derived concentration of F. psychrophilum that could be detected in ≥95% of qPCR replicate reactions; this concentration was multiplied by the DECF to yield the LOD (Kralik and Ricchi et al. 2017). The LOQ was determined by running triplicate reactions of eight qPCR standards (108 – 101 gene copies/reaction), triplicate reactions of six F. psychrophilum dilutions (1.00 × 108 cells/mL – 1.00 × 103 cells/mL), and 10 replicate reactions of each remaining dilution (e.g., 1 × 102 cells/mL – 1 cell/mL) and defined as the lowest mean qPCR-derived concentration of F. psychrophilum with a CV <25%; this concentration was multiplied by the DECF to yield the LOQ (Kralik and Ricchi et al. 2017). 4.3.3. In vivo assessment of shedding dynamics in Atlantic salmon, coho salmon, and rainbow trout 4.3.3.1. Flavobacterium psychrophilum isolate selection Three F. psychrophilum variants, including US19 (ST13, CC-ST9; Van Vliet et al. 2016), US62 (ST277, CC-ST232; Knupp et al. 2019), and US87 (ST275, CC-ST10; Knupp et al. 2019), were selected for this study. Each CC is present across a wide geographic range, whereby CC- ST232 has been detected in two continents (Europe and North America; Nilsen et al. 2014; Knupp et al. 2019), and CC-ST9 and CC-ST10 have both been detected in four continents (e.g., 149 Asia, Europe, North America, and South America; Fujiwara-Nagata et al. 2013; Nilsen et al. 2014; Avendaño-Herrera et al. 2014; Knupp et al. 2019; Li et al. 2021). Moreover, each CC has been recovered almost exclusively from one salmonid species, including Atlantic salmon (CC- ST232), coho salmon (CC-ST9), and rainbow trout (CC-ST10). 4.3.3.2. Origin of fish for shedding experiment Embryonated Atlantic salmon and rainbow trout eggs were sourced from a commercial egg distributor, while embryonated coho salmon eggs were procured from Platte River State Fish Hatchery. Coordination occurred so that all eggs from the three species arrived at the Michigan State University – University Research Containment Facility on the same day. In brief and upon receipt, eggs were disinfected with 100-ppm iodophor solution (pH 7.30) for 10 minutes before being placed in a vertical incubator supplied with UV-treated, sand-filtered well water maintained at 12 ºC ± 1 ºC until hatching. Sac-fry were then moved to aerated flow-through tanks (40 L; 12 ºC ± 1 ºC) and, once exogenous feeding commenced, were given a continuous supply of appropriately sized commercial trout food (Skretting, the Netherlands) via an automatic feeder. After eight weeks, fish were hand-fed twice daily and the water volume in the tanks was increased (400 L; 12 ºC ± 1 ºC). Tanks were cleaned and siphoned daily to remove waste and any uneaten food. Before the challenge experiment, a subset of fish from each species were cultured to screen for bacterial infections (Knupp et al. 2021a), including those caused by F. psychrophilum, and confirmed to be bacterial infection-free. 4.3.3.3. Flavobacterium psychrophilum inoculum preparation for shedding experiment Flavobacterium psychrophilum variants US19, US62, and US87 were revived from cryostock and verified to be pure cultures as detailed in section 4.3.2.1. Each F. psychrophilum variant was inoculated into 250-mL of FPM-A broth and incubated with constant shaking (180 150 rpm) for 48 hours at 15 ºC. Bacteria were harvested via centrifugation (2,571 × g, 15 minutes) and adjusted to an optical density at 600-nm (OD600) corresponding to 2.0 ×108 colony forming units (cfu)/mL using sterile 0.65% saline. Concentrations of each F. psychrophilum variant were determined as detailed in section 4.3.2.1. 4.3.3.4. Intramuscular challenge of fish Each F. psychrophilum variant was inoculated into the salmonid species it is associated with, according to MLST. Thus, US19 was inoculated into coho salmon, whereas US62 and US87 were inoculated into Atlantic salmon and rainbow trout, respectively. Atlantic salmon (n = 5, in duplicate; 8 months old; mean weight 18.1g), coho salmon (n = 5, in duplicate; 8 months old; mean weight 20.5g), and rainbow trout (n = 5, in duplicate; 8 months old; mean weight 25.1g) were anesthetized in sodium bicarbonate-buffered (200 mg/L) tricaine methanesulfonate (MS-222; Syndel) at a concentration of 100 mg/L. Then, fish were intramuscularly injected at the base of the dorsal fin with a 50-µL volume of F. psychrophilum, equating to a dose of 105 cfu/g, and then placed into aerated flow-through glass experimental aquaria (37.85 L; n = 2 aquaria per species, n = 5 fish per aquarium) supplied with ultraviolet light-treated, sand-filtered well water (12 ºC ± 1 ºC). Control fish (n = 5 per species, in duplicate aquaria) were treated identically except they were intramuscularly injected with an equal volume of 0.65% saline. The challenge experiment was conducted in accordance with the MSU- Institutional Animal Care and Use Committee (AUF:201900312). 4.3.3.5. Sampling of water containing live and dead fish All live fish in a replicate experimental aquarium (i.e., main aquarium) were net transferred to a clean non-flow-through glass aquarium (i.e., shedding aquarium; 9.46 L) containing 3000-mL of fresh, ultraviolet light-treated, sand-filtered well water (12 ºC ± 1 ºC) 151 with constant aeration. The shedding aquarium was placed inside a larger, opaque, plastic aquarium that had flow-through water to maintain a water temperature of 12 ºC ± 1 ºC in the shedding aquarium; the plastic aquarium also had an opaque cover to minimize light-induced stress. After one hour, all fish were removed from the shedding aquarium and transferred back to the main aquarium. A 50-mL water sample was collected using a sterile 50-mL conical tube and then processed immediately as detailed in section 4.3.2.2. Water sampling of live fish (including negative control fish) occurred on every other day for the first week, twice per week during the second and third weeks, and then once a week until the end of the experiment (i.e., four weeks without detection of F. psychrophilum). Up to two dead fish per replicate aquarium (i.e., ≤ 4 dead fish per species) were net transferred into a new flow-through glass aquarium (37.85 L, n = 1 dead fish per aquarium) supplied with ultraviolet light-treated, sand-filtered well water (12 ºC ± 1 ºC). After 1, 3, 5, 7, 14, 63, and 98-days post-death, fish were transferred to individual shedding aquaria and treated/sampled identically to the live fish. Fish that died that were not used for determining F. psychrophilum shedding rates were necropsied, clinically examined, and multiple tissues (e.g., external ulcers and kidney) were bacteriologically analyzed for F. psychrophilum on FPM-A. To disinfect shedding aquaria between samplings, aquaria were completely immersed in a 10% (v/v) bleach solution for ≥ 10 min, rinsed thoroughly with pathogen-free water, and then air dried. 4.3.3.6. Determination of Flavobacterium psychrophilum shedding rate from live and dead fish via qPCR Flavobacterium psychrophilum shedding rates were determined via qPCR (section 4.3.1.1). Briefly, each 96-well qPCR plate consisted of triplicate reactions of eight qPCR standards (108 gene copies – 101 gene copies), duplicate reactions of template DNA, triplicate 152 reactions of no-template control (e.g., sterile, nuclease free water), and triplicate reactions of IPC amplification control (e.g., 1,000 copies of Xeno™ IPC). The shedding rate of F. psychrophilum from infected fish is reported as F. psychrophilum cells shed per fish per hour in 1-mL of water (i.e., F. psychrophilum cells/fish/hour; Madetoja et al. 2000) and was calculated using the following formula: Mean qPCR gene copies × DECF × 3000 mL 𝐹. 𝑝𝑠𝑦𝑐ℎ𝑟𝑜𝑝ℎ𝑖𝑙𝑢𝑚 cells/fish/hour = Number of fish in aquarium 4.4. Results 4.4.1. qPCR standards All qPCR standards, which spanned nine orders of magnitude (108 gene copies/reaction – 1 gene copy/reaction), successfully amplified. Linear regression of Cq values versus the log10 target gene demonstrated acceptable correlation (R2 = 0.999 ± 0.001) with a slope and efficiency of -3.23 ± 0.01 and 104.01% ± 0.53%, respectively. The assay was repeatable within and between runs, as CV values were <2.5% within the linear range (108 gene copies/reaction – 1 gene copy/reaction) of the standard curve (Table 4.1). Consequently, and in accordance with Standish et al. (2018), these standards were used in all subsequent qPCR assays. 4.4.2. Flavobacterium psychrophilum DNA extraction efficiency and qPCR limit of detection and quantification Mean F. psychrophilum concentrations of the nine mock water samples ranged from 1.00 × 108 to 1.00 × 100 cell(s)/mL and were compared to the mean qPCR-determined concentrations (Table 4.2). Mean DNA extraction efficiency ranged from 3.63% - 99.70%, with a median of 14.03% (Table 4.2); therefore, the DECF applied to all qPCR-derived concentrations was 7.128 (=100/14.03; Kralik et al. 2011). 153 The assay LOD and LOQ were calculated as 30.94 ± 6.84 cells/mL, whereby the DECF (7.128) was multiplied by the lowest mean qPCR-derived F. psychrophilum concentration meeting the qualifications for the LOD and LOQ (e.g., both 4.34 ± 0.96 cells/mL; Table 4.2). Additionally, amplification was achieved in 40% of the replicates at the next lowest concentration, which corresponded to 13.68 ± 7.09 cells/mL after correction for DNA extraction efficiency (=1.92 ± 0.99 cells/mL × 7.128). The IPC DNA, which was added at the beginning of the DNA extraction process, successfully amplified in all replicate reactions of all but the most concentrated mock water sample (Table 4.2), which according to the VetMAX™ Xeno™ user guide was most likely caused by high target concentration. Nonetheless, of the eight mock water samples that had successful IPC amplification, Cq values ranged from 30.12 to 35.29 and all replicates were highly consistent (e.g., CV values ranged from 0.36% to 1.17%; Table 4.2). 4.4.3. Flavobacterium psychrophilum shedding dynamics in three salmonid species 4.4.3.1. Shedding rates of Flavobacterium psychrophilum from live fish Following intramuscular injection of 105 cfu/g of F. psychrophilum, all tested species (e.g., Atlantic salmon, coho salmon, and rainbow trout) shed F. psychrophilum into the water over multiple days and/or prior to the occurrence of any mortality (Figure 4.1A-C). Atlantic salmon shed 1.59 × 104 cells/fish/hour one day after inoculation; this shedding rate increased to 4.34 × 107 cells/fish/hour (i.e., >2,700-fold increase) by day five, which was two days prior to the first mortality (Figure 4.1A). Atlantic salmon shed the most F. psychrophilum (e.g., 6.37 × 107 cells/fish/hour) on day seven (i.e., the day of the first mortality; Figure 4.1A), with shedding rates decreasing thereafter on each subsequent sampling day (e.g., from 1.43 × 106 cells/fish/hour on day 11 to 4.30 × 103 cells/fish/hour on day 21; Figure 4.1A). 154 Cumulative mortality in Atlantic salmon peaked at 20% on day 9. Surviving Atlantic salmon were euthanized on day 49 (i.e., four weeks past the last detection of F. psychrophilum; Figure 4.1A), at which time F. psychrophilum infection status was examined (section 4.4.4). Live rainbow trout began shedding F. psychrophilum into the water at a rate of 1.65 × 104 cells/fish/hour one day after inoculation; this shedding rate increased to 4.16 × 108 cells/fish/hour (i.e., >25,000-fold increase) by day seven, which was two days prior to the first mortality (Figure 4.1B). Rainbow trout shedding rates then increased to 4.18 × 109 cells/fish/hour on day 11 (50% cumulative mortality), decreased to 4.32 × 108 cells/fish/hour on day 14 (60% cumulative mortality) before increasing to 4.23 × 109 cells/fish/hour on day 18 (60% cumulative mortality; Figure 4.1B). On day 19, cumulative mortality increased to 80% and remained constant through day 35; during this time, rainbow trout shedding rates decreased from 5.98 × 107 cells/fish/hour (day 21) to 2.86 × 104 cells/fish/hour (day 35), after which F. psychrophilum was not detected in the water (Figure 4.1B). Cumulative mortality in rainbow trout reached its peak (e.g., 90%) on day 42 and the experiment ended on day 63 (Figure 4.1B). Live coho salmon began shedding F. psychrophilum into the water at a rate of 1.04 × 105 cells/fish/hour one day after inoculation. Mortality began quickly in this species (e.g., on day two) and rapidly reached 100% by day seven. The highest detected F. psychrophilum shedding rate of live coho salmon (e.g., 7.39 × 107 cells/fish/hour) occurred on day five (Figure 4.1C). Control fish did not shed F. psychrophilum nor did they experience any mortality (data not shown). Likewise, inhibition of gene target amplification was not observed among water samples originating from the negative control or F. psychrophilum-exposed fish, as evidenced by all IPC Cq values falling within 28 – 34. 155 4.4.3.2. Shedding rates of Flavobacterium psychrophilum from dead fish Atlantic salmon, coho salmon, and rainbow trout shed F. psychrophilum into the water from one day post-death (PD) to the end of the 98-day experiment (Figure 4.2A-C). Throughout the experiment, fish underwent post-mortem decomposition (Figure 4.3A-C). Because of extensive decomposition after day 98, sampling ceased. Two Atlantic salmon died during the experiment, one from each replicate aquarium (Figure 4.2A). Initially (i.e., one day PD), F. psychrophilum shedding rates differed by ~15-fold, whereby one fish shed 2.20 × 107 cells/fish/hour and the other shed 3.45 × 108 cells/fish/hour (mean of 1.8 ± 1.6 × 108 cells/fish/hour; Figure 4.2A; Table 4.3); however, shedding rates became more consistent between 3 - 14 days PD (e.g., differed by 1 - 3-fold). Over the next 12 weeks, the shedding rate of one Atlantic salmon decreased to 1.63 × 104 cells/fish/hour and then F. psychrophilum became undetectable, whereas the shedding rate of the other Atlantic salmon decreased but was still detectable (e.g., 4.51 × 106 cells/fish/hour; Figure 4.2A) 98 days PD. Nine of the ten rainbow trout died during the experiment; therefore, individual F. psychrophilum shedding rates of two rainbow trout per replicate aquarium were measured (Figure 4.2B). Initially (i.e., one day PD), most (i.e., 3/4) rainbow trout were shedding ~108 cells/fish/hour (range of 4.91 – 8.20 × 108 cells/fish/hour), whereas one rainbow trout was shedding 8.73 × 107 cells/fish/hour (overall mean of 5.0 ± 2.7 × 108 cells/fish/hour; Figure 4.2B; Table 4.3). By 5 days PD, all rainbow trout shed at substantially higher rates, whereby individual rainbow trout were shedding 0.28 - 3.59 × 1010 cells/fish/hour (mean of 1.6 ± 1.2 × 1010 cells/fish/hour; Figure 4.2B; Table 4.3). Between 7 - 14 days PD, rainbow trout shedding rates decreased ~1 – 12,000-fold, and the shedding rate of one rainbow trout continued to decrease through the end of the experiment (final shedding rate of 1.36 × 108 cells/fish/hour; Figure 4.2B). 156 For the other three rainbow trout, shedding rates increased ~4 – 1,200-fold (e.g., to 0.68 – 3.5 × 109 cells/fish/hour) by 63 days PD. At the end of the 98-day experiment, these three rainbow trout were shedding 1.77 × 106 to 6.09 × 109 cells/fish/hour (mean of 1.6 ± 1.2 × 109 cells/fish/hour; Figure 4.2B; Table 4.3). All ten coho salmon died during the experiment; therefore, individual F. psychrophilum shedding rates of two coho salmon per replicate aquarium were measured (Figure 4.2C). Initially (i.e., one day PD), most (i.e., 3/4) coho salmon were shedding ~108 cells/fish/hour (range of 2.71 – 9.26 × 108 cells/fish/hour), whereas one coho salmon was shedding 3.24 × 107 cells/fish/hour (overall mean of 4.8 ± 3.5 × 108 cells/fish/hour; Figure 4.2C; Table 4.3). By 3 days PD, shedding increased by ~1 – 100-fold for most (i.e., 3/4) coho salmon, whereby individuals were shedding between 0.38 - 8.45 × 109 cells/fish/hour (Figure 4.2C). In contrast, F. psychrophilum shedding rate decreased by ~100-fold (e.g., to 5.59 × 106 cells/fish/hour) for the other coho salmon. By 5 days PD shedding rates had increased for all coho salmon (range of 0.21 × 1010 cells/fish/hour; mean of 5.0 ± 3.8 × 109 cells/fish/hour; Figure 4.2C; Table 4.3) and continued to increase for 2/4 coho salmon through 7 days PD (range of 2.12 – 6.94 × 109 cells/fish/hour). In the other two coho salmon, a decrease in shedding rates was appreciated (range of 1.34 – 9.28 × 109 cells/fish/hour; Figure 4.2C). Over the next 13 weeks (i.e., through the end of the 98-day experiment), shedding rates generally decreased, and most (i.e., 3/4) coho salmon were still shedding F. psychrophilum (range of 4.95 × 105 to 1.28 × 107 cells/fish/hour; mean of 1.6 ± 2.6 × 109 cells/fish/hour; Figure 4.2C; Table 4.3). 4.4.4. Infection status in salmonids challenged with Flavobacterium psychrophilum Flavobacterium psychrophilum was recovered in a pure form and as perfuse lawns (i.e., colony forming units were too numerous to count) from external lesions (e.g., muscle 157 ulcerations) and the kidneys of all dead Atlantic salmon, coho salmon, and rainbow trout. Flavobacterium psychrophilum was not recovered from the kidneys of any surviving Atlantic salmon (n = 8 fish; 49 days post-inoculation) or rainbow trout (n = 1 fish; 63 days post- inoculation). However, F. psychrophilum was recovered in a pure form and as a perfuse lawn from the eye of the only surviving rainbow trout. Because all coho salmon died after seven days, no survivors could be cultured. 4.5. Discussion For the first time, F. psychrophilum shedding dynamics (e.g., time to shedding, shedding rate, and duration in live and dead fish) have been elucidated in Atlantic salmon and coho salmon, two of the most BCWD-susceptible salmonid species (Holt, 1987; Barnes and Brown, 2011; Nilsen et al. 2011). Addressing this knowledge gap was essential, as several studies showed that BCWD epizootics in Atlantic salmon and coho salmon are instigated by phenotypically distinct F. psychrophilum variants (Madetoja et al. 2001; Sundell and Wiklund, 2015; Sundell et al. 2019). Moreover, these F. psychrophilum variants differ from those affecting rainbow trout (Nilsen et al. 2014; Avendaño-Herrera et al. 2014; Knupp et al. 2019), the sole species our entire understanding of F. psychrophilum shedding dynamics is based upon. Although some commonalities in shedding dynamics were observed among Atlantic salmon, coho salmon, and rainbow trout, multiple differences were also apparent. Dead Atlantic salmon, rainbow trout, and coho salmon were found to shed F. psychrophilum cells at rates up to 5.4-, 8.5-, and 171.8-fold greater than their living counterparts, and did so for extended periods (e.g., up to 77, 63, and 93 days longer). This the first time these comparisons have been made among Atlantic salmon and coho salmon, demonstrating these species, in addition to rainbow trout, are efficient F. psychrophilum shedders. Madetoja et al. 158 (2000) also found dead rainbow trout shed at higher (e.g., up to ~100-fold) rates and for a longer duration (e.g., 59 days longer) than living rainbow trout. The difference in shedding rate among live and dead rainbow trout observed by Madetoja et al. (2000) is substantially greater than the 8.5-fold difference found herein for rainbow trout, while differences in shedding duration were similar (59 vs. 63 days; Madetoja et al. 2000). Taken together with the findings in Atlantic salmon and coho salmon, results suggest that some F. psychrophilum variants may be more efficiently shed by dead fish than others. Nevertheless, dead F. psychrophilum-infected fish clearly pose a significant risk for disease perpetuation within fish farms and hatcheries, underscoring the importance of implementing management strategies that aim to remove dead fish from rearing units quickly. Like F. psychrophilum, higher shedding rates in dead salmonids versus live salmonids have been noted in at least one other fish pathogenic Flavobacterium species (e.g., F. columnare, a cause of columnaris disease; Kunttu et al. 2009; LaFrentz et al. 2022), as well as Aeromonas salmonicida subspecies salmonicida (e.g., the cause of furunculosis; Rose et al. 1989), and therefore dead fish shedding may be an important transmission strategy among some bacterial fish pathogens. Despite these similarities in shedding, multiple shedding differences with likely implications for F. psychrophilum transmission and BCWD management were also observed. For example, on day 11, live rainbow trout were shedding 2,923-fold more F. psychrophilum cells/hour compared to live Atlantic salmon. This difference continued to increase through day 21, which was the last day of F. psychrophilum detection among Atlantic salmon. Meanwhile (i.e., on day 21), F. psychrophilum shedding rates remained high among rainbow trout (e.g., 5.98 × 107 cells/fish/hour) and this species continued to shed for another two weeks. Adding a further layer of complexity, F. psychrophilum shedding rates among live Atlantic salmon and rainbow 159 trout tended to peak on and/or near days with mortality, and could possibly suggest that risk of transmission may vary by F. psychrophilum variant. For example, Atlantic salmon infected with US62 (ST277, in CC-ST232) died over three days, whereas rainbow trout infected with US87 (ST275) mostly died over 11 days (i.e., a >3-fold longer period). Madetoja et al. (2000) infected rainbow trout with a different F. psychrophilum variant and observed that fish died over a four- day period. Therefore, it appears some F. psychrophilum variants pose a greater transmission risk among live fish compared to others. Differences in F. psychrophilum infection status were also apparent among rainbow trout and Atlantic salmon survivors, as the lone surviving rainbow trout was still infected with F. psychrophilum 63 days post-inoculation, in contrast to the surviving Atlantic salmon. Indeed, a common finding among previous in vivo F. psychrophilum challenge (Madetoja et al. 2000; Knupp et al. 2021a; Macchia et al. 2022) and surveillance studies (Madsen et al. 2005; Marancik and Wiens, 2013; Van Vliet et al. 2016; Knupp et al. 2019) is the existence of F. psychrophilum carriers. Where coho salmon fall in this regard is less clear, as US19 proved highly virulent (i.e., caused rapid and fulminant mortality) to exposed fish in the current study, despite being inoculated with an identical dose. Nevertheless, our findings suggest rainbow trout may be a longer-term infection reservoir and pose greater F. psychrophilum transmission risk compared to at least Atlantic salmon. With these findings in mind, the results of Chapter 3 add an additional layer of complexity to our understanding of F. psychrophilum transmission dynamics. The coho salmon-associated isolate US19 and Atlantic salmon-associated isolate US62 proved capable of infecting and causing disease and mortality in Atlantic salmon, rainbow trout, and coho salmon, albeit to varying degrees. Within the context of a facility rearing multiple salmonid species, generalist F. psychrophilum variants may pose a higher risk of cross-species transmission. Thus, 160 such facilities may need to implement spatial and/or physical barriers to limit cross-species transmission. In this context, future studies comparing transmission capabilities of host specific and generalist F. psychrophilum variants are warranted. The higher F. psychrophilum shedding rates in live Atlantic salmon and rainbow trout coupled with delayed death [e.g., onset at 7 (Atlantic salmon) or 9 (rainbow trout) days] could affect F. psychrophilum transmission dynamics and increase the number of subsequent infections (i.e., the basic reproduction number, R0; Delamater et al. 2019) as higher shedding rates over a longer period may provide F. psychrophilum with more opportunities to infect new hosts. On the other hand, coho salmon exhibited a different shedding patten – dying rapidly (e.g., ≤ 7 days) and potentially before peak shedding, which could potentially limit F. psychrophilum transmission if death occurred prior to maximum pathogen release. However, the initial shedding rate in coho salmon was ~10-fold higher than that of Atlantic salmon and rainbow trout, possibly compensating for the shorter shedding period. In the context of a fish farm or hatchery rearing unit, this could potentially enhance transmission despite the quick death. Moreover, if high early shedding followed by rapid death is a strategy for F. psychrophilum, it could be an effective way to maximize transmission while limiting the time for host immune response. Ogut et al. (2005) found Aeromonas salmonicida subsp. salmonicida transmission was host density dependent, whereby transmission increased with fish density. However, Ogut et al. (2005) also found R0 was similar (e.g., 1.17 – 1.45), even over a wide range of densities (e.g., 0.36 – 9.13 fish/L), possibly indicating fish behavior changed with density. Whether these findings are emulated by F. psychrophilum remains to be determined but warrants investigation as they may have implications for F. psychrophilum transmission. 161 The observed differences in shedding dynamics across F. psychrophilum variants is likely influenced by pathogen and/or host factors. In this study, US87, identified in Chapter 3 as molecular serotype 2 (nearly equivalent to conventional serotype Th; Lorenzen and Olesen, 1997; Rochat et al. 2017), showed a protracted shedding period in rainbow trout. This variant was also observed to persist systemically in one (e.g., the lone surviving rainbow trout in this study) or more (e.g., survivors in Chapter 3) individuals. Collectively, these findings suggest that the rainbow trout immune system struggles to eliminate this variant, which likely contributed to its prolonged shedding. Considering the heightened transmission risk of this variant (ST275), it should be incorporated into the development of BCWD prevention and control strategies. US62 also exhibited prolonged shedding herein, albeit to a lesser extent than US87, but appeared to be successfully cleared by Atlantic salmon at the end of the experiment. This F. psychrophilum variant belonged to molecular serotype 1 (nearly equivalent to conventional serotype Fd; Lorenzen and Olesen, 1997; Rochat et al. 2017; Chapter 3) and was less host specific (Chapter 3). Moreover, these serotypes are less common in Atlantic salmon (Avendaño-Herrera et al. 2020) and thus possibly less capable of circumventing this species immune response, possibly helping to explain the shorter duration of shedding and lack of recovery from survivors herein and Chapter 3. US19 was highly virulent herein and caused mortality before a rigorous adaptive immune response could likely be mounted. Whether the shedding dynamics of these variants are representative of other F. psychrophilum variants affecting Atlantic salmon, coho salmon, and rainbow trout should be further investigated. The prolonged duration and high rate of F. psychrophilum shedding from dead fish may be influenced by multiple factors. In the absence of a host immune response, the conditions may become ideal for F. psychrophilum proliferation, possibly owing to the fish carcass serving as an 162 abundant protein source (i.e., F. psychrophilum main energy source; Duchaud et al. 2007; Chapter 2), this species’ inherent proteolytic capabilities (Barbier et al. 2020), and water temperatures that not only slow fish tissue decomposition (Nobre et al. 2019) but are conducive to F. psychrophilum growth. Another possible contributing factor is the formation of biofilm by F. psychrophilum, which can form on various surfaces (Levipan and Avendaño-Herrera, 2017; Vidal et al. 2020), and is often comprised of multiple species (Schoina et al. 2022). In this context, F. psychrophilum has been shown to lyse cells of multiple bacterial species and use them as a growth source (Pacha and Porter, 1968). In addition to elucidating important aspects of F. psychrophilum shedding dynamics in three salmonid species, this study also built upon the previous work of Marancik and Wiens (2013) to optimize their qPCR assay for detection and quantification of F. psychrophilum in water. Originally, Marancik and Wiens (2013) described this assay for the detection/quantification of F. psychrophilum in spleen tissue, with a LOD of 3.1 genome units per reaction and LOQ of ~486 colony forming units. Herein, this assay was further optimized to quantify F. psychrophilum from filtered water, with a LOD/LOQ of 30.94 ± 6.84 cells/mL. The sensitivity of the assay could potentially be improved by filtering a larger water volume and/or by increasing sample volume per reaction. Indeed, even a 50-mL water volume containing 108 F. psychrophilum cells/mL filtered quickly (i.e., no indication of filter fouling). Strepparava et al. (2014) also designed a F. psychrophilum-specific qPCR to quantify F. psychrophilum from water samples, and although the LOD was similar (e.g., 66 cells/mL), the LOQ was >100-fold higher (e.g., 3,300 cells/mL). Moreover, and in the hands of the study authors and at least two other laboratories, specificity issues were observed with this qPCR assay (unpublished data) and so alternatives were sought. An additional improvement for the qPCR assay herein was the addition 163 of an internal positive control, which allowed for monitoring of PCR-inhibition, a known source of “false negatives” with molecular assays (Kavlick, 2018). Moving forward, this qPCR assay will be instrumental to future studies assessing F. psychrophilum shedding dynamics and could have application to F. psychrophilum detection in environmental field settings. Prior to this study, knowledge of F. psychrophilum shedding dynamics were limited only to rainbow trout that were infected with a single F. psychrophilum variant. Given that F. psychrophilum is now well-known as a genetically diverse pathogen of multiple farm and hatchery-reared salmonid species with varying phenotypes, coupled with the differential shedding results by variant/host species herein, future studies assessing shedding dynamics and underlying pathogen/host mechanisms are warranted and may lead to improved BCWD prevention and control. Overall, study results demonstrated dead Atlantic salmon, coho salmon, and rainbow trout shed F. psychrophilum at higher rates than their liver counterparts and for at least several months, thereby potentially posing a greater transmission risk. Therefore, and with the insight provided by a recent F. psychrophilum transmission model (Brenden et al. 2023), it remains critical for personnel raising fish to remove dead, F. psychrophilum-infected fish from rearing units quickly and frequently. 164 REFERENCES Avendaño-Herrera, R., Houel, A., Irgang, R., Bernardet J-F., Godoy, M., Nicolas, P., Duchaud, E. 2014. Introduction, expansion, and coexistence of epidemic Flavobacterium psychrophilum lineages in Chilean fish farms. Veterinary Microbiology. 170, 298-306. Avendano-Herrera, R., Tapia-Cammas, D., Duchaud, E., Irgang, R. 2020. Serological diversity in Flavobacterium psychrophilum: A critical update using isolates retrieved from Chilean salmon farms. Journal of Fish Diseases, 43, 877–888. Barnes, M.E., Brown, M.L. 2011. A review of Flavobacterium psychrophilum biology, clinical signs, and bacterial coldwater disease prevention and treatment. The Open Fish Science Journal. 4, 40-48. Barbier, P., Rochat, T., Mohammed, H.H., Wiens, G.D., Bernardet, J.-F., Halpern, D., Duchaud, E., McBride, M.J. 2020. The type IX secretion system is required for virulence of the fish pathogen Flavobacterium psychrophilum. Applied and Environmental Microbiology. 86, e00799-20. Brenden, T.O., Ivan, L.N., Loch, T.P. 2023. Reducing Flavobacterium psychrophilum transmission risk via hatchery-rearing practices: An individual-based modeling evaluation. Aquaculture. 563, 738868. Brown, L.L., Cox, W.T., Levine, R.P. 1997. Evidence that the causal agent of bacterial coldwater disease Flavobacterium psychrophilum is transmitted within salmonid eggs. Diseases of Aquatic Organisms. 29(3), 213-218. Delamater, P.L., Street, E.J., Leslie, T.F., Yang, Y.T., Jacobsen, K.H. 2019. Complexity of the basic reproduction number (R0). Emerging Infectious Diseases. 25(1), 1-4. Duchaud, E., Boussaha, M., Loux, V., Bernardet, J-F., Michel, C., Kerouault, B., Stanislas Mondot, S., Nicolas, P., Bossy, R., Caron, C., Bessières, P., Gibrat, J-F., Claverol, S., Dumetz, F., Hénaff, M., Benmansour, A. 2007. Complete genome sequence of the fish pathogen Flavobacterium psychrophilum. Nature Biotechnology. 25, 763-769. Ekman, E., Borjeson, H., Johansson, N. 1999. Flavobacterium psychrophilum in Baltic salmon Salmo salar brood fish and their offspring. Diseases of Aquatic Organisms. 37, 159–163. Fujiwara-Nagata, E., Chantry-Darmon, C., Bernardet, J-F., Eguchi, M., Duchaud, E., Nicolas, P. 2013. Population structure of the fish pathogen Flavobacterium psychrophilum at whole-country and model river levels in Japan. Veterinary Research. 44, 34. Griffin, M.J., Mauel, M.J., Greenway, T.E., Khoo, L.H., Wise, D.J. 2011. A real-time polymerase chain reaction assay for quantification of Edwardsiella ictaluri in catfish pond water and genetic homogeneity of diagnostic case isolates from Mississippi. Journal of Aquatic Animal Health. 23(4), 178-188. 165 Griffin, M.J., Pote, L.M., Camus, A.C., Mauel, M.J., Greenway, T.E., Wise, D.J. 2009. Application of a real-time PCR assay for the detection of Henneguya ictaluri in commercial channel catfish ponds. Diseases of Aquatic Organisms. 86, 223-233. Holt, R.A. 1987. PhD thesis. Cytophaga psychrophila, the causative agent of bacterial cold water disease in salmonid fish. Oregon State University, Corvallis, OR. Kavlick, M.F. 2018. Development of a universal internal positive control. BioTechniques. 65(5). Knupp, C., Wiens, G.D., Faisal, M., Call, D.R., Cain, K.D., Nicolas, P., Van Vliet, D., Yamashita, C., Ferguson, J.A., Meuninck, D., Hsu, H-M., Baker, B.B., Shen, L., Loch, T.P. 2019. Large-scale analysis of Flavobacterium psychrophilum multilocus sequence typing genotypes indicates that both newly identified and recurrent clonal complexes are associated with disease. Applied and Environmental Microbiology. 86, e02305-18. Knupp, C., Kiupel, M., Brenden, T.O., Loch, T.P. 2021a. Host-specific preference of some Flavobacterium psychrophilum multilocus sequence typing genotypes determines their ability to cause bacterial coldwater disease in coho salmon (Oncorhynchus kisutch). Journal of Fish Diseases. 44(5), 521-531. Knupp, C., Faisal, M., Brenden, T.O., Wiens, G.D., Loch, T.P. 2021b. In vivo experiments provide evidence that Flavobacterium psychrophilum strains belonging to multilocus sequence typing clonal complex ST191 are virulent to rainbow trout (Oncorhynchus mykiss). Journal of Aquatic Animal Health. 33(3), 190-195. Kralik, P., Slana, I., Kralova, A., Babak, V., Whitlock, R.H., Pavlik, I. 2011. Development of a predictive model for detection of Mycobacterium avium subsp. paratuberculosis in faeces by quantitative real time PCR. Veterinary Microbiology. 149, 133–138. Kralik, P., Ricchi, M. 2017. A basic guide to real time PCR in microbial diagnostics: Definitions, parameters, and everything. Frontiers in Microbiology. 8, 108. Kumagai, A., Yamaoka, S., Takahashi, K., Fukuda, H., Wakabayashi, H.. 2000. Waterborne Transmission of Flavobacterium psychrophilum in Coho Salmon Eggs. Fish Pathology. 35, 25- 28. Kunttu, H.M., Valtonen, E.T., Jokinen, E.I., Suomalainen, L-R. 2009. Saprophytism of a fish pathogen as a transmission strategy. Epidemics. 1(2), 96-100. LaFrentz, B.R., Králová, S., Burbick, C.R., Alexander, T.L., Phillips, C.W., Griffin, M.J., Waldbieser, G.C., García, J.C., Alexandre Sebastião, F., Soto, E., Loch, T.P., Liles, M.R., Snekvik, K.R. 2022. The fish pathogen Flavobacterium columnare represents four distinct species: Flavobacterium columnare, Flavobacterium covae sp. nov., Flavobacterium davisii sp. nov. and Flavobacterium oreochromis sp. nov., and emended description of Flavobacterium columnare. Systematic and Applied Microbiology. 45(2), 126293. 166 Lebuhn, M., Effenberger, M., Garces, G., Gronauer, A., Wilderer, P.A. 2004. Evaluating real- time PCR for the quantification of distinct pathogens and indicator organisms in environmental samples. Water Science & Technology. 50(1), 263-270. Levipan, H.A., Avendaño-Herrera, R. 2017. Different Phenotypes of Mature Biofilm in Flavobacterium psychrophilum Share a Potential for Virulence That Differs from Planktonic State. Frontiers in Cellular and Infection Microbiology. 7, 76. Li, S., Chai, J., Cao, Y., Knupp, C., Wang, D., Nicolas, P., Chen, F., Liu, H., Lu, T., Loch, T.P. 2021. Characterization and molecular epidemiological analysis of Flavobacterium psychrophilum recovered from diseased salmonids in China. Microbiology Spectrum. 9(2), e00033021. Loch, T.P., Faisal, M. Flavobacterium spp. In Fish viruses and bacteria: Pathobiology and protection., Woo, P.T.K., Cipriano, R.C., Eds., CABI: 2017, pp. 211-232. Lorenzen, E., Olesen, N.J. 1997. Characterization of isolates of Flavobacterium psychrophilum associated with coldwater disease or rainbow trout fry syndrome II: serological studies. Diseases of Aquatic Organisms. 31, 209-220. Macchia, V., Inami, M., Ramstad, A., Grammes, F., Reeve, A., Moen, T., Torgersen, T.S., Adams, A., Desbois, A.P., Hoare, R. 2022. Immersion challenge model for Flavobacterium psychrophilum infection of Atlantic salmon (Salmo salar L.) fry. Journal of Fish Diseases. 45(11). 1781-1788. Madetoja, J., Dalsgaard, I., Wiklund, T. 2002. Occurrence of Flavobacterium psychrophilum in fish-farming environments. Diseases of Aquatic Organisms. 52, 109-118. Madetoja, J., Nyman, P., Wiklund, T. 2000. Flavobacterium psychrophilum, invasion into and shedding by rainbow trout Oncorhynchus mykiss. Diseases of Aquatic Organisms, 43, 27-38. Madetoja, J., Nystedt, S., Wiklund, T. 2003. Survival and virulence of Flavobacterium psychrophilum in water microcosms. FEMS Microbiology Ecology. 43(2), 217-223. Madetoja, J., Wiklund, T. 2002. Detection of the fish pathogen Flavobacterium psychrophilum in water from fish farms. Systematic and Applied Microbiology. 25(2), 259-266. Madsen, L., Moller, J.D., Dalsgaard, I. 2005. Flavobacterium psychrophilum in rainbow trout, Oncorhynchus mykiss (Walbaum), hatcheries: studies on broodstock, eggs, fry, and environment. Journal of Fish Diseases. 28(1), 39-47. Marancik, D., Wiens, G.D. 2013. A real-time polymerase chain reaction assay for identification and quantification of Flavobacterium psychrophilum and application to disease resistance studies in selectively bred rainbow trout Oncorhynchus mykiss. FEMS Microbiology Letters. 339(2), 122-129. 167 Nicolas, P., Mondot, S., Achaz, G., Bouchenot, C., Bernardet, J-F., Duchaud, E. 2008. Population structure of the fish-pathogenic bacterium Flavobacterium psychrophilum. Applied and Environmental Microbiology. 74, 3701-3709. Nilsen, H., Johansen, R., Colquhoun, D.J., Kaada, I., Bottolfsen, K., Vagnes, O., Olsen, A.B. 2011. Flavobacterium psychrophilum associated septicaemia and necrotic myositis in Atlantic salmon Salmo salar: a case report. Disease of Aquatic Organisms. 97(1), 37-46. Nilsen, H., Sundell, K., Duchaud, E., Nicolas, P., Dalsgaard, I., Madsen, L., Aspan, A., Jansson, E., Colquhoun, D.J., Wiklund, T. 2014. Multilocus sequence typing identifies epidemic clones of Flavobacterium psychrophilum in Nordic countries. Applied and Environmental Microbiology. 80, 2728 –2736. Nobre, R.L.G., Carneiro, L.S., Panek, S.E., Gonzalex, M.J., Vanni, M.J. 2019. Fish, including their carcasses, are net nutrient sources to the water column of a eutrophic lake. Frontiers in Ecology and Evolution. 7, 340. Ogut, H., LaPatra, S.E., Reno, P.W. 2005. Effects of host density on furunculosis epidemics determined by the simple SIR model. Preventive Veterinary Medicine. 71(1-2), 83-90. Pacha, R.E., Porter, S. 1970. Myxobacterial disease of salmonids. In: Snieszko SF (ed) A symposium on disease of fishes and shellfishes, Special publication No. 5. Washington D.C, American Fisheries Society, pp 243-257. Rangdale, R.E., Richards, R.H., Alderman, D.J. 1996. Isolation of Cytophaga psychrophila, causal agent of rainbow trout fry syndrome (RTFS) from reproductive fluids and egg surfaces of rainbow trout (Oncorhynchus mykiss). Bulletin of the European Association of Fish Pathologists. 16, 97-101. Ricchi, M., Savi, R., Bolzoni, L., Pongolini, S., Grant, I.R., De Cicco, C., Cerutti, G., Cammi, G., Garbarino, C.A., Arrigoni, N. 2016. Estimation of Mycobacterium avium subsp. Paratuberculosis load in raw bulk tank milk in Emilia-Romagna Region (Italy) by qPCR. Microbiology Open. 5(4), 551-559. Rochat, T., Fujiwara-Nagata, E., Calvez, S., Dalsgaard, I., Madsen, L., Calteau, A., Lunazzi, A., Nicolas, P., Wiklund, T., Bernardet, J-F., Duchaud, E. 2017. Genomic characterization of Flavobacterium psychrophilum serotypes and development of a multiplex PCR-based serotyping scheme. Frontiers in Microbiology. 8, 1752. Rose, A.S., Ellis, A.E., Munro, A.L.S. 1989. The infectivity by different routes of exposure and shedding rates of Aeromonas salmonicida subsp. Salmonicida in Atlantic salmon, Salmo salar L., held in sea water. Journal of Fish Diseases. 12(6), 573-578. Schoina, E., Doulgeraki, A.I., Miliou, H., Nychas, G-J.E. 2022. Dynamics of water and biofilm bacterial community composition in a Mediterranean recirculation aquaculture system. Aquaculture Journal. 2(2), 164-179. 168 Siekoula-Nguedia, C., Blanc, G., Duchaud, E., Calvez, S. 2012. Genetic diversity of Flavobacterium psychrophilum isolated from rainbow trout in France: predominance of a clonal complex. Veterinary Microbiology. 161, 169-178. Slana, I., Kralik, P., Kralova, A., Pavlik, I. 2008. On-farm spread of Mycobacterium avium subsp. paratuberculosis in raw milk studied by IS900 and F57 competitive real time quantitative PCR and culture examination. International Journal of Food Microbiology. 128, 250-257. Standish, I., Leis, E., Schmitz, N., Credico, J., Erickson, S., Bailey, J., Kerby, J., Phillips, K., Lewis, T. 2018. Optimizing, validating, and field testing a multiplex qPCR for the detection of amphibian pathogens. Diseases of Aquatic Organisms. 129(1), 1-13. Starliper, C.E. 2011. Bacterial coldwater disease of fishes caused by Flavobacterium psychrophilum. Journal of Advanced Research. 2, 97-108. Stoeckel, D.M., Stelzer, E.A., Dick, L.K. 2009. Evaluation of two spike-and-recovery controls for assessment of extraction efficiency in microbial source tracking studies. Water Research. 43(19), 4820-4827. Strepparava, N., Nicolas, P., Wahl, T., Segner, H., Petrini, O. 2013. Molecular epidemiology of Flavobacterium psychrophilum from Swiss fish farms. Diseases of Aquatic Organisms. 105, 203-210. Strepparava, N., Wahli, T., Segner, H., Petrini, O. 2014. Detection and quantification of Flavobacterium psychrophilum in water and fish tissue samples by quantitative real time PCR. BMC Microbiology. 105. Sundell, K., Landor, L., Nicolas, P., Jørgensen, J., Castillo, D., Middelboe, M., Dalsgaard, I., Donati, V. L., Madsen, L., Wiklund, T. 2019. Phenotypic and genetic predictors of pathogenicity and virulence in Flavobacterium psychrophilum. Frontiers in Microbiology. 10, 1711. Sundell, K., Wiklund, T. 2015. Characteristics of epidemic and sporadic Flavobacterium psychrophilum sequence types. Aquaculture. 441, 51-56. Vidal, J.M., Miranda, C.D., De la Fuente, M., Alarcon, M., Aroca, G., Sossa, K., Ruiz, P., Urrutia, H. 2020. Formation of biofilms of the salmon pathogen Flavobacterium psychrophilum in different surfaces using the CDC biofilm reactor. Aquaculture. 514, 734459. Van Vliet, D., Wiens, G.D., Loch, T.P., Nicolas, P., Faisal, M. 2016. Genetic diversity of Flavobacterium psychrophilum isolates from three Oncorhynchus spp. in the United States, as revealed by multilocus sequence typing. Applied and Environmental Microbiology. 82, 3246- 3255. Van Tongeren, S.P., Degener, J.E., Harmsen, H.J.M. 2011. Comparison of three rapid and easy bacterial DNA extraction methods for use with quantitative real-time PCR. European Journal of Clinical Microbiology and Infectious Diseases. 9, 1053-1061. 169 Wood, J.W. 1974. Diseases of Pacific salmon, their prevention and treatment, 2nd edition. Washington Department of Fisheries, Olympia. 170 APPENDIX Table 4.1. Repeatability of Flavobacterium psychrophilum Marancik and Wiens (2013) qPCR assay. Gene copies Cq Mean Cq SD Cq CV (%) Intra-assay 100000000 11.34 0.12 1.06 10000000 14.63 0.09 0.65 1000000 17.83 0.09 0.48 100000 21.78 0.11 0.52 10000 24.61 0.12 0.48 1000 27.60 0.11 0.40 100 30.76 0.16 0.51 10 34.05 0.36 1.06 1 36.96 0.86 2.32 Inter-assay 100000000 11.22 0.12 1.09 10000000 14.65 0.17 1.17 1000000 17.79 0.10 0.57 100000 21.58 0.13 0.61 10000 24.49 0.20 0.83 1000 27.58 0.12 0.42 100 30.69 0.20 0.65 10 34.01 0.31 0.90 1 37.05 0.83 2.25 Note: Mean, standard deviation, and coefficient of variation (CV) of the cycle threshold (Cq) values are shown. To determine intra-assay variation, the target gene was serially diluted in ten- fold increments over nine orders of magnitude (i.e., 108 gene copies/reaction – 1 gene copy/reaction) and eight replicate reactions of each dilution were PCR amplified and quantified. Inter-assay variation was determined by repeating the qPCR run on a separate plate using an identical number of replicate reactions. The assay was repeatable within and between runs, as CV percentages were <2.5 across the linear range (Standish et al. 2018). 171 Table 4.2. Evaluation of Flavobacterium psychrophilum DNA extraction procedure from water using quantitative PCR (qPCR). Key to abbreviations: SD, standard deviation; CV, coefficient of variation; IPC, internal positive control; Cq, cycle threshold. Theoretical concentration of F. psychrophilum F. psychrophilum qPCR-derived concentration of F. psychrophilum DNA extraction Mean Mean SD Signal ratioa CV (%) Cq IPC (SD) IPC CV (%) efficiency (%)b 1.00 × 108 9.97 × 107 5.21 × 106 3/3 5.23 - - 99.70 1.00 × 107 7.85 × 106 7.08 × 104 3/3 0.90 35.29 (0.32) 0.91 78.49 1.00 × 106 3.19 × 105 2.82 × 104 3/3 8.83 31.08 (0.26) 0.82 31.91 1.00 × 105 8.87 × 103 4.35 × 102 3/3 4.90 30.12 (0.21) 0.70 8.87 1.00 × 104 6.73 × 102 3.30 × 101 3/3 4.90 31.17 (0.15) 0.49 6.73 1.00 × 103 3.63 × 101 6.07 × 100 3/3 16.74 32.31 (0.11) 0.36 3.63 1.00 × 102 4.34 × 100 9.63 × 10-1 10/10 22.19 31.56 (0.37) 1.17 4.34 1.00 × 101 1.92 × 100 9.95 × 10-1 4/10 51.81 31.04 (0.24) 0.76 19.20 1.00 × 100 - - 0/10 - 32.12 (0.36) 1.11 - Median 14.03 a Number of amplified replicate reactions/total number of replicate reactions. b Calculated as the quotient of the mean qPCR-derived concentration of F. psychrophilum and the mean theoretical concentration of F. psychrophilum cells, multiplied by 100. 172 Table 4.3. Mean Flavobacterium psychrophilum shedding rates (cells/fish/hour) ± standard deviation of dead Atlantic salmon, coho salmon, and rainbow trout on each sampling day. Number of fish sampled (N) is to the left of the host species name. Host species Sampling Day N Atlantic salmon N Coho salmon N Rainbow trout 1 2 1.8 ± 1.6 × 108 4 4.8 ± 3.5 × 108 4 5.0 ± 2.7 × 108 3 2 5.0 ± 0.8 × 107 4 3.1 ± 3.4 × 109 4 1.9 ± 0.5 × 109 5 2 6.7 ± 2.1 × 107 4 5.0 ± 3.8 × 109 4 1.6 ± 1.2 × 1010 7 2 1.7 ± 0.7 × 108 4 4.9 ± 3.3 × 109 4 8.1 ± 6.3 × 109 14 2 1.2 ± 0.6 × 108 4 4.4 ± 5.1 × 109 4 1.6 ± 2.6 × 109 63 2 6.9 ± 6.9 × 106 4 9.1 ± 8.9 × 108 4 1.6 ± 1.2 × 109 98 1 4.5 ± 0.0 × 106 3 4.9 ± 5.6 × 106 4 1.6 ± 2.6 × 109 173 Figure 4.1. Mean Flavobacterium psychrophilum shedding rates (black bars with standard deviation) from and cumulative percent mortality (gray lines with standard error) of live A) Atlantic salmon (Salmo salar), B) rainbow trout (Oncorhynchus mykiss), and C) coho salmon (O. kisutch). *, water sampling occasion, which continued four weeks past the last detection of F. psychrophilum. //, indicates 6-day gap in time. 174 Figure 4.2. Flavobacterium psychrophilum shedding rates of dead individual (A) Atlantic salmon (Salmo salar), (B) rainbow trout (Oncorhynchus mykiss), and (C) coho salmon (O. kisutch). Legend explanation: ≤ 2 fish (F) per replicate (R) aquarium were maintained. Therefore, in the legend and as an example, R1.F1 corresponds to replicate one, fish one. 175 Figure 4.3. Representative images of dead fish used in Flavobacterium psychrophilum shedding experiment. (A) Coho salmon (Oncorhynchus kisutch) 12 days post death. (B) Rainbow trout (O. mykiss) 25 days post death. (C) Coho salmon 98 days post death. Note the yellowish discoloration present on fish. 176 Chapter 5: Culturability of heterologous Flavobacterium psychrophilum multilocus sequence typing variants in three microcosms that simulate common fish farm and hatchery environments 177 5.1. Abstract Flavobacterium psychrophilum, causative agent of bacterial coldwater disease (BCWD), is a widespread threat to the success of salmonid farms and hatcheries, where it causes substantial economic losses. In these environments, fish are fed frequently and reared at elevated densities, leading to the accumulation of uneaten food and/or raceway detritus. Such substrates are potential nutrient sources for F. psychrophilum when outside its host, possibly contributing to BCWD epizootic risk. Concurrently, hundreds of F. psychrophilum sequence types (STs; i.e., genetic variants) have been described using multilocus sequence typing (MLST) and findings show that some variants are more frequently associated with fish farms and hatcheries. Whether persistence strategies in these environments differ amongst F. psychrophilum variants remains unknown. To this end, the culturability of ten distinct F. psychrophilum MLST variants was evaluated for 13 weeks in three microcosms comprised of sterilized well water, sterilized well water with commercial trout feed, or sterilized well water with raceway detritus. These variants belonged to globally distributed clonal complexes (CCs) and/or are repeatedly recovered from fish farms and hatcheries. All ten F. psychrophilum variants remained culturable in each of the three microcosms for at least eight weeks, with bacterial concentrations often highest in the presence of raceway detritus. In addition, most (e.g., 90%) F. psychrophilum variants remained culturable for at least 13 weeks. Significant differences in culturability were observed both within and between microcosms, suggesting potential variability in environmental persistence strategies among specific variants. Collectively, results highlight the remarkable ability of F. psychrophilum to not only persist for weeks under nutrient limited conditions, but also thrive in the presence of organic substrates common in fish farm and hatchery rearing units. 178 5.2. Introduction Flavobacterium psychrophilum, causative agent of bacterial coldwater disease (BCWD), poses a significant economic challenge by affecting multiple salmonid (Family Salmonidae) species reared in farms and hatcheries around the world (Loch and Faisal, 2017). The widespread success of this pathogen is multifactorial, in part relating to vertical transmission, circumvention of egg-surface disinfectants such as iodophor (Brown et al. 1997; Cipriano, 2005; Starliper 2011), and efficient horizontal transmission (Madetoja et al. 2000). Moreover, there are reports of reduced susceptibility to the few antibiotics that are approved to treat F. psychrophilum infections in food fish (Bruun et al. 2000; Van Vliet et al. 2017), and development of efficacious, licensed BCWD vaccines in the USA remain elusive to date (Gomez et al. 2014). Another factor likely perpetuating losses caused by F. psychrophilum is its ability to survive outside its host for extended periods. Vatsos et al. (2003) found F. psychrophilum survived for 133 days in sterile stream water, while Madetoja et al. (2003) showed F. psychrophilum remained culturable in sterile lake water for 300 days. Although some fish farms and hatcheries use surface water sources (e.g., stream or river water; Strepparava et al. 2014), others use groundwater sources (e.g., well or spring water; Van Vliet et al. 2015), which differ in quality (e.g., hardness; Summerfelt, 2000) and may affect F. psychrophilum survival. Although not demonstrated for F. psychrophilum, water hardness has been shown to affect the biofilm formation of at least one other flavobacterial fish pathogen, F. columnare (Cai et al. 2013). Madetoja et al. (2003) found that F. psychrophilum could survive longer and proliferate more extensively in lake water microcosms that also contained natural beach sand. Other substrates, such as detritus and uneaten food (Schumann, 2021), may also affect F. psychrophilum survival but have yet to be examined. Given the common presence of detritus and uneaten food in fish 179 farms and hatcheries, investigating their role in F. psychrophilum survival could be pivotal for informing improved BCWD management strategies. The considerable intraspecific diversity within F. psychrophilum may also affect its survival within fish farm and hatchery environments. In this context, multilocus sequence typing (MLST) has been applied to >1500 F. psychrophilum isolates worldwide, revealing over 260 distinct sequence types (STs; https://pubmlst.org/organisms/flavobacterium-psychrophilum). Importantly, the wealth of diversity within this species has been linked to phenotypic variation with respect to host associations, virulence, ultraviolet light susceptibility, and antimicrobial susceptibility (Van Vliet et al. 2017; Knupp et al. 2019; Sundell et al. 2019; Knupp et al. 2021a; Knupp et al. 2021b; Knupp et al. 2023, in press). However, it remains unclear if such intraspecific diversity also affects F. psychrophilum environmental persistence strategies. Indeed, previous studies examining F. psychrophilum environmental persistence strategies only used one to two isolates yet represent the totality of our knowledge (Vatsos et al. 2003; Madetoja et al. 2003). Several MLST-based studies suggested that certain F. psychrophilum variants (e.g., ST10, ST13, ST78, ST267, ST275, and ST253) are common in fish farms and hatcheries, as evidenced by their repeated detections in these environments (Fujiwara-Nagata et al. 2013; Nilsen et al. 2014; Van Vliet et al. 2016; Knupp et al. 2019; Li et al. 2021). In contrast to these reports, other variants, such as ST252 and ST256 (Van Vliet et al. 2016; Knupp et al. 2019), appear to be more common in wild and/or feral fish populations. Whether these disparities in recovery environment reflect differences in environmental persistence strategies or abilities to survive outside a host has yet to be thoroughly investigated. To improve our understanding of F. psychrophilum environmental persistence strategies in fish farms and hatcheries, the culturability of ten F. psychrophilum variants, representing eight 180 MLST clonal complexes and two singletons, was measured for 13 weeks in three microcosms containing well water only, well water with detritus, and well water with commercial trout feed, to simulate environments commonly found in salmon and trout rearing facilities. 5.3. Materials and Methods 5.3.1. Flavobacterium psychrophilum isolate selection A total of ten F. psychrophilum isolates, recovered from 2010 – 2021 from two gamete collection locations and six hatcheries in two U.S. states (e.g., Michigan and Pennsylvania) were selected for this study (Table 5.1). The isolates were recovered from three salmonid (Family Salmonidae) genera and five species, including Atlantic salmon (Salmo salar, n = 1), brown trout (S. trutta, n = 2), brook trout (Salvelinus fontinalis, n = 1), coho salmon (Oncorhynchus kisutch, n = 1), and steelhead trout (O. mykiss, n = 5; Table 5.1). The isolates were genetically diverse according to MLST, whereby most (n = 6) had been previously genotyped (Van Vliet et al. 2016; Knupp et al. 2019; https://pubmlst.org/fpsychrophilum/) and four were newly genotyped in this study following published protocols (Knupp et al. 2019). In total, eight isolates belonged to eight STs within five CCs, two isolates belonged to two distinct singleton STs, and all ten comprised globally relevant and/or widespread STs (Fujiwara-Nagata et al. 2013; Nilsen et al. 2014; Avendaño-Herrera et al. 2014; Knupp et al. 2019; Li et al. 2021; Table 5.1). 5.3.2. Bacterial inoculum preparation The ten cryogenically preserved (maintained at -80 ºC) F. psychrophilum variants were revived on F. psychrophilum medium-A (FPM-A; Chapter 2), incubated for 48 hours at 15 ºC, and then visually inspected for purity. Each variant was inoculated into 150-mL of FPM-A broth and incubated with constant shaking (180 rpm) for 48 hours at 15 ºC. Bacteria were harvested via centrifugation (2,571 × g, 15 minutes), washed once using 0.22 µm filtered and autoclaved well 181 water (hereafter referred to as well water) that originated from a salmonid-rearing quarantine facility, and then resuspended in 15-mL of analogous water. 5.3.3. Microcosms Three microcosms (i.e., treatment groups) that simulate environments common to fish- farms and hatchery facilities were prepared using 125-mL glass Erlenmeyer flasks with partially loosened screw caps. Each flask (n = 2 per isolate per treatment) contained 67.5-mL of either well water alone (treatment 1) or well water with 12.5 g (dry weight) raceway detritus (treatment 2) or commercial trout feed (treatment 3; see microcosm preparations below). A 2.5-mL aliquot of bacterial suspension was added to each flask and gently swirled ten times. To quantify starting bacterial concentrations (i.e., week 0), a 1-mL aliquot was collected from each flask, serially diluted 100,000,000-fold in ten-fold increments, plated on FPM-A in duplicate, and then incubated for seven days, after which final colony counts were performed. This sampling procedure was performed once per week for eight consecutive weeks, as well as on week 13. Prior to removing a sample for colony enumeration and on weeks 9 – 12 (i.e., when no sampling occurred), flasks were gently swirled ten times. Between samplings, flasks were incubated statically in the dark at 10 ºC (i.e., common temperature for BCWD epizootics; Barnes and Brown, 2011). 5.3.3.1. Fish-rearing facility well water Ultraviolet light-treated, sand-filtered well water (pH 7.3) was obtained from the aquatics section of the Michigan State University – University Research Containment Facility (MSU- URCF), which is routinely used to rear salmonids. For this experiment, well water was filtered using 0.22-µm filter flasks with a polyethersulfone filter membrane (Santa Cruz Biotechnology, Dallas, TX, USA) and then aliquoted into 60 flasks (67.5 mL/flask); a subset of which (n = 20) 182 were immediately autoclaved at 121 ºC for 15 min. The remaining flasks were autoclaved after raceway detritus or commercial trout feed was added (see sections 5.3.3.2 and 5.3.3.3). 5.3.3.2. Raceway detritus Detritus was selected for inclusion as a component of an experimental microcosm, as it represents an accumulation of uneaten food and fish byproducts (e.g., mucus and feces) commonly found in fish rearing units, thereby representing a potential nutrient source for F. psychrophilum while outside its host. To simulate this environment, detritus was siphoned from a single raceway of rainbow trout hatched and reared under quarantine conditions at MSU-URCF and collected on a mesh screen. Rainbow trout had been screened for F. psychrophilum via culture on FPM-A and verified free of bacterial infection (see Chapter 3). Amassed detritus was air-dried for 24 hours, finely crushed via mortar and pestle, and then stored in the dark at 4 ºC in a sterile glass bottle until use. Dried detritus was weighed (12.5 g/flask), added to flasks (n = 20) containing 67.5-mL filtered well water, and then autoclaved at 121 ºC for 15 min. 5.3.3.3. Commercial trout feed One potential environmental reservoir for F. psychrophilum is uneaten fish feed, which can act as an organic substrate supporting bacterial proliferation. Furthermore, when fish are infected with F. psychrophilum, one of the first observable disease signs is often a decrease in feeding. This could result in an accumulation of uneaten feed in the rearing units, potentially promoting the persistence and proliferation of F. psychrophilum. For this reason, commercial trout feed Skretting #2 (Skretting USA, Tooele, Utah) was included as a treatment. This feed is specifically formulated for coldwater fish species, including rainbow trout, which are particularly vulnerable to BCWD during their early life-stages (Holt, 1987). Fresh feed was weighed (12.5 183 g/flask), added to flasks (n = 20) containing 67.5-mL filtered well water, and then autoclaved at 121 ºC for 15 min. 5.3.4. Data analysis A linear mixed model was used to quantify the effect (log10 concentration in cfu/mL) of each microcosm (i.e., treatment) on the ten F. psychrophilum variants. The model included treatment, week, variant, the interaction between treatment and week, and the interaction between treatment and variant as fixed effects. Degrees of freedom for fixed effects were calculated using the Kenward-Roger method. Custom hypothesis tests as to differences in collective (i.e., all weeks combined) mean concentration among variants within the same treatment group and between treatment groups for the same variant were evaluated through pairwise comparisons of least-square means and adjusted for multiple comparisons using the Tukey-Kramer method (α = 0.05). Analyses were performed using PROC MIXED in SAS® Version 9.4; custom hypothesis testing was performed using the LSMEANS statement and pdiff option. 5.4. Results 5.4.1. General model analyses Based on the fitted linear mixed-effects model, treatment group (e.g., raceway detritus, commercial trout feed, and water only) effects differed significantly depending on week (df = 18, 543; F = 14.21; P-value < 0.0001) and F. psychrophilum variant (df = 27, 543; F = 17.70; P- value < 0.0001). Averaging observations across all ten variants and nine sampling weeks, mean F. psychrophilum concentrations were significantly higher in the raceway detritus treatment group compared to both the commercial trout feed (LSmeans estimate 1.64 ± 0.08; df = 543; t- value = 20.72; P-value < 0.0001) and well water only treatment groups (LSmeans estimate 1.22 184 ± 0.08; df = 543; t-value = 15.46; P-value <0.0001). Likewise, mean F. psychrophilum concentrations were significantly higher in well water only compared to commercial trout feed (LSmeans estimate 0.42 ± 0.08; df = 543; t-value = 5.26; P-value < 0.0001). 5.4.2. Flavobacterium psychrophilum culturability in well water only Initial F. psychrophilum concentrations within each well water only microcosm flask ranged from 2.14 × 107 cfu/mL to 1.78 × 108 cfu/mL (Figure 5.1A; Table 5.2) compared to 1.00 × 105 cfu/mL – 9.33 × 105 cfu/mL at week 13 (Figure 5.1A; Table 5.2). Averaging observations across all ten variants and nine sampling weeks, mean F. psychrophilum concentrations decreased between weeks zero and seven but increased on week eight before decreasing through the end of the study (e.g., week 13; Figure 5.1A; Table 5.3). However, mean concentrations of ST78 and ST267 increased between weeks zero and one (Figure 5.1A; Table 5.2). Also of note, the mean concentration of ST277 initially declined more rapidly than all other variants over the first two weeks, but by week five, its mean concentration aligned more closely with the other variants (Figure 5.1A; Table 5.2). Averaging observations for each variant across all nine sampling weeks, ST342 maintained the highest concentration, whereas ST277 had the lowest concentration, with the mean concentration difference between these two variants being significant (df = 543; t-value = 4.55; P-value = 0.0025; Table 5.4). Likewise, collective mean concentrations of ST286 and ST275 were significantly higher than ST277 (df = 543; t-values = 4.11 and 4.25; P-values = 0.0085 and 0.0148, respectively). For the remaining variant comparisons, collective mean bacterial concentrations were not significantly different from one another (df = 30; t-values = 0.01 – 3.97; P-values = 0.0707 – 1.0000; Table 5.4). 185 5.4.3. Flavobacterium psychrophilum culturability in well water with detritus Initial bacterial concentrations within each well water with detritus microcosm flask ranged from 2.14 × 107 cfu/mL to 1.78 × 108 cfu/mL (Figure 5.1B; Table 5.5) compared to 0.00 cfu/ml – 5.25 × 107 cfu/mL at week 13 (Figure 5.1B; Table 5.5). Averaging observations across all ten variants and nine sampling weeks, mean F. psychrophilum concentrations increased between weeks zero and two, decreased on week three, increased on week four, and then decreased over the remainder of the study (Figure 5.1B; Table 5.3). Of note, larger differences in culturability were observed beginning week seven, whereby mean concentrations of ST353 (2.24 × 104 cfu/mL) and ST286 (6.76 × 105 cfu/mL) were lower than the other variants (9.77 × 106 – 3.72 × 108 cfu/mL; Figure 5.1B; Table 5.5). Averaging observations for each variant across all nine sampling weeks, ST253 maintained the highest mean concentration, whereas ST353 had the lowest mean concentration. In this context, most (e.g., 8/9 ≈ 88.9%) F. psychrophilum variants had a significantly higher collective mean concentration compared to ST353 (df = 543; t-values = 4.17 – 7.77; P-values < 0.0001 – 0.0003), and ST253 had a significantly higher collective mean concentration compared to ST286 (ST286; df = 543; t-value = 4.74; P-value = 0.0011; Table 5.4). Similarly, ST277 had a significantly higher collective mean concentration compared to ST286 (df = 543; t-value = 4.17; P-value = 0.0119). In examining the remaining variant comparisons, collective mean concentrations were not significantly different (df = 543; t-values = 0.11 – 3.69; P-values = 0.0654 – 1.0000; Table 5.4). 5.4.4. Flavobacterium psychrophilum culturability in well water with commercial trout feed Initial bacterial concentrations within each well water with commercial trout feed microcosm flask ranged from 2.14 × 107 cfu/mL to 1.78 × 108 cfu/mL (Figure 5.1C; Table 5.6) compared to 10 cfu/mL – 3.89 × 106 cfu/mL at week 13 (Figure 5.1B; Table 5.6). Averaging 186 observations across all ten variants and nine sampling weeks, mean F. psychrophilum concentrations decreased between weeks zero and two, increased between weeks three and five, and then steadily decreased over the remaining eight weeks (Figure 1C; Table 5.3). However, the mean concentration of ST267 increased between weeks zero and one before decreasing sharply over the next two weeks (Figure 5.1C; Table 5.6). Averaging observations for each variant across all nine sampling weeks, ST275 maintained the highest mean concentration, whereas ST267 had the lowest mean concentration and the mean concentration difference between these two variants was significant (df = 543; t-value = 14.67; P-value < 0.0001; Table 5.4). Likewise, collective mean concentrations of all other F. psychrophilum variants were significantly higher than ST267 (df = 30; t-values = 8.19 – 14.06; P-values < 0.0001; Table 5.4). Similarly, 66.7% (e.g., 6/9 variants) and 55.6% (e.g., 5/9 variants) of F. psychrophilum variants had collective mean concentrations that were significantly higher than ST13 or ST10, respectively (Table 5.4). Overall, collective mean F. psychrophilum concentrations were most variable in this microcosm, whereby each variant differed significantly from one to nine other variants (Table 5.4). 5.4.5. Flavobacterium psychrophilum culturability comparisons across microcosms Overall, most F. psychrophilum variants (e.g., 8/10 = 80.0%) maintained higher collective (i.e., all weeks combined) mean concentrations in raceway detritus compared to commercial trout feed (df = 543; t-values = 3.93 – 18.84; P-values < 0.0001 – 0.0284; Table 5.7). Likewise, most F. psychrophilum variants (e.g., 8/10 = 80.0%) maintained higher collective mean concentrations in raceway detritus compared to well water only (df = 543; t-values = 3.89 – 9.63; P-values < 0.0001 – 0.0337; Table 5.7). When comparing collective mean concentrations in commercial trout feed to well water only, no significant differences (df = 543; t-values = 0.18 – 3.16; P-values = 0.2758 – 1.0000) 187 were observed for most F. psychrophilum variants (e.g., 7/10 = 70.0%; Table 5.7). However, one variant (ST277) had a significantly higher collective mean concentration in the presence of commercial trout feed (df = 543; t-value = 3.82; P-value = 0.0422; Table 5.7), whereas the two other variants (ST13 and ST267) each had a significantly higher collective mean concentration in well water only (df = 543; t-values = 4.26 and 12.11; P-values = 0.0081 and < 0.0001, respectively; Table 5.7). Of note, one F. psychrophilum variant (ST353) had no significant differences in collective mean concentrations amongst all three microcosms (df = 543; t-values = 0.09 – 1.83; P-values = 0.9937 – 1.0000; Table 5.7). 5.5. Discussion Although F. psychrophilum has been recovered from wild/feral fish (Van Vliet et al. 2016; Knupp et al. 2019; Harrison et al. 2021) and has caused mortalities in these populations (Davis, 1946), most BCWD epizootics occur in fish farms and hatcheries (Kum et al. 2008; Nilsen et al. 2011; Avendaño-Herrera et al. 2014; Knupp et al. 2019; Li et al. 2021). Indeed, the artificial rearing environment is ideal for F. psychrophilum, as fish are reared at elevated densities, providing enhanced opportunity for horizontal transmission (Kennedy et al. 2015). In this context, F. psychrophilum is efficiently shed from live and dead fish into the water column during BCWD epizootics (Madetoja et al. 2000; Chapter 4). Although some F. psychrophilum cells may be transmitted directly to a new host, others may encounter suspended solids that will eventually settle in the rearing unit. Two settable solids commonly found in fish farms and hatcheries are uneaten feed and detritus (Schumann, 2021). Despite being natural and common organic sources in fish rearing facilities, where they represent substrates for microbial proliferation (Dauda et al. 2019), controlled experiments evaluating F. psychrophilum survival in these microenvironments has not been reported. Herein, the 13-week culturability of ten 188 genetically diverse F. psychrophilum variants was evaluated in laboratory microcosms containing sterile well water, sterile well water with commercial trout feed, and sterile well water with raceway detritus. Overall, cfu yields across weeks and in well water only were similar among F. psychrophilum variants. Importantly, however, well water with raceway detritus led to significant increases in cfu yields amongst most F. psychrophilum variants. Moreover, well water with commercial trout feed also improved yields among some F. psychrophilum variants. These findings provide evidence that these common and natural settable solids have the potential to increase F. psychrophilum loads in fish farms and hatcheries, thereby potentially increasing the risk of BCWD. In addition to these findings, it was also apparent that some F. psychrophilum variants persisted well in all tested microcosms, whereas some were better suited to specific microcosms. Collectively, these findings not only improve our understanding of BCWD ecology and F. psychrophilum persistence outside its host but also have potential to improve BCWD management strategies. All F. psychrophilum variants remained culturable in filtered well water for 13 weeks, and at similar concentrations over most weeks, possibly suggesting metabolic responses among most variants are comparable in this nutrient-limited environment. Vatsos et al. (2003) reported one F. psychrophilum isolate (MLST variant unknown) remained culturable for 19 weeks in filtered stream water. Likewise, Madetoja et al. (2003) found a different F. psychrophilum isolate (MLST variant unknown) remained culturable for 300 days (~43 weeks) in filtered lake water. In contrast to the slow decrease in bacterial concentrations observed amongst most F. psychrophilum variants herein and in previous studies (Vatsos et al. 2003; Madetoja et al. 2003), concentrations of F. psychrophilum variant ST277 decreased rapidly over the first two weeks but then consistently increased over the following three weeks. These results possibly suggest that F. 189 psychrophilum variant ST277 may be less equipped to transition from a nutrient-rich environment to a nutrient-limited one. Despite this, F. psychrophilum variant ST277 demonstrated a capacity to recover, possibly by exploiting the substantial number of cells that initially died as a growth source. Indeed, F. psychrophilum has been demonstrated to lyse and utilize bacteria for its proliferation (Pacha and Porter, 1968). However, these findings may also mean under natural conditions, where a significant mass of bacterial cells is less likely to be available, the survival potential of this specific variant might be compromised. Nonetheless, results provide further evidence F. psychrophilum is capable of persisting for weeks under nutrient-limited conditions. Most of the analyzed F. psychrophilum variants persisted at high concentrations (e.g., ~107 – 108 cfu/mL) in the presence of raceway detritus for at least seven weeks. The apparent success of F. psychrophilum in the presence of this common fish farm and hatchery substrate may be related to the composition of raceway detritus, which is primarily composed of fish feces and uneaten food, both of which are rich in nutrients, including protein (Hardy and Gatlin, 2022; Schumann, 2021; Skretting USA). Genomic studies have indeed indicated that proteins are the primary energy source for F. psychrophilum (Duchaud et al. 2007; Castillo et al. 2021), findings further corroborated by in vitro growth studies (Holt, 1987; Alvarez and Guijarro, 2007; Oplinger and Wagner, 2012; Chapter 2). Interestingly, these detrital compounds only appear to better support the persistence of most studied F. psychrophilum variants in the first eight weeks in comparison to raw trout feed. Indeed, by week 13, concentrations of most F. psychrophilum variants were higher in the presence of commercial trout feed compared to detritus. Cautiously extrapolating these in vitro findings to field conditions suggests that raceway detritus may serve as a greater “short-term” substrate for F. psychrophilum proliferation, whereas commercial trout 190 feed may be a better “long-term” substrate, a matter highlighting the importance of removing detritus and feed from fish farm and hatchery rearing units. Although similar trends in culturability were apparent in each microcosm, potential differences in environmental persistence strategies and survival responses among F. psychrophilum variants within a specific microcosm were also apparent. For instance, the culturability of F. psychrophilum variant ST267 was significantly less than all other variants in the presence of commercial trout feed. Interestingly, ST267 was the only variant that grew between weeks zero and one in the presence of commercial trout feed, while the concentrations of all other variants decreased by ≥10-fold. A similar growth pattern for ST267 was observed in the other microcosms. Thus, results may suggest that certain F. psychrophilum variants, when faced with sub-optimal nutrient conditions, initially respond by mobilizing energy reserves rather than adapting to a new nutrient source. Lending support to this theory, one genomic study (Duchaud et al. 2007) identified genes in one F. psychrophilum variant (ST20 in CC-ST10) that encode cyanophycinase and cyanophycin synthetase – enzymes necessary for cyanophycin production. Although the biological function of cyanophycin has not been explored in F. psychrophilum, it is known that other bacterial species use cyanophycin as a carbon/nitrogen (i.e., energy) storage mechanism, employing it for energy in sub-optimal nutrient conditions (Krehenbrink et al. 2002). Whether cyanophycin is common amongst F. psychrophilum variants and/or is used in this capacity remains to be investigated. In contrast to ST267, overall culturability of F. psychrophilum variant ST353 in the presence of commercial trout feed was like most other variants; however, this variant’s culturability in the presence of detritus was significantly worse than most other variants, and was the only non-culturable variant at 13- 191 weeks. Thus, some F. psychrophilum variants may be better suited to different hatchery microenvironments. Findings may lend support to previous molecular epidemiology research studies that indicated certain F. psychrophilum MLST variants, such as ST253, are more adapted to the fish hatchery environment, a claim based on the frequent recovery of ST253 from the same facility over several years (e.g., 2010, 2013, 2017, and 2020; Knupp et al. 2019). For instance, F. psychrophilum variant ST253 had the highest collective mean concentration in the presence of raceway detritus and was the only variant to persist a high concentration (>107 cfu/mL) up to the 13th week, indicating a possible superior adaptation to this environment compared to other variants. Moreover, ST253 was culturable for 13 weeks in both other microcosms, and at moderately high concentrations (e.g., ~105 – 106 cfu/mL). Recognizing that some F. psychrophilum variants may thrive in common fish farm and hatchery environments may lead to improved BCWD management strategies, including altering feeding to minimize waste accumulation, adjusting rearing unit disinfection protocols, or even considering selective breeding programs for fish resistant to these highly successful F. psychrophilum variants. Although this research provides evidence that F. psychrophilum can survive for extended periods in microenvironments commonly found in fish farms and hatcheries, further studies are needed to assess the virulence of variants originating from these settings, particularly under conditions where the bacterium has not been nutrient-deprived (e.g., in the presence of commercial trout feed and raceway detritus). Madetoja et al. (2003) exposed rainbow trout via immersion to one F. psychrophilum isolate that had been in sterile lake water for 14 days, finding it was moderately virulent (e.g., approximately 50% cumulative mortality). In addition, this same F. psychrophilum isolate retained most (e.g., 50% reduction in cumulative mortality) of its 192 virulence after subcutaneous injection into rainbow trout after being in sterile lake water for 49 days. In this context, if F. psychrophilum can remain virulent after extended periods of starvation, then perhaps under conditions when nutrients are available, F. psychrophilum will retain its virulence or potentially become more virulent. Indeed, Kinnula et al. (2017) found F. columnare (a causative agent of columnaris disease; LaFrentz et al. 2022) virulence increased proportionally with nutrient availability and possibly promoted virulence factor activation. In conclusion, it is well-recognized that F. psychrophilum is a substantial threat to both aquaculture farms raising fish for consumption and hatcheries dedicated to conservation and stock enhancement efforts. Results herein provide evidence for the first time that multiple F. psychrophilum variants can persist for weeks in well water, a common groundwater source for fish farms and hatcheries, and at moderate concentrations. Moreover, it is apparent that raceway detritus is a growth source for most F. psychrophilum variants, elevating bacterial concentrations for weeks. Thus, efforts to remove raceway detritus fish farm and hatchery rearing units may be very beneficial to reducing overall F. psychrophilum loads, and possibly a source of BCWD outbreaks. Notably, however, some variants demonstrated an ability to persist and thrive using not only raceway detritus but commercial trout feed as well, which may potentially confer a fitness advantage in the fish farm and hatchery environment. In contrast, other variants appear to be only suited to one type of nutrient environment. This distinction underscores the complexity of BCWD ecology, and highlights the need for management strategies that consider the varying survival strategies and growth preferences of different F. psychrophilum variants. Nevertheless and in the interim, it is recommended that personnel raising fish enhance rearing unit hygiene practices, particularly focusing on removing detritus and especially during BCWD outbreaks as 193 removing this environmental reservoir could pay dividends in reducing bacterial loads and infection risk. 194 REFERENCES Álvarez, B., Guijarro, J.A. 2007. Recovery of Flavobacterium psychrophilum viable cells using a charcoal-based solid medium. Letters in Applied Microbiology. 44(5), 569-572. Avendaño-Herrera, R., Houel, A., Irgang, R., Bernardet J-F., Godoy, M., Nicolas, P., Duchaud, E. 2014. Introduction, expansion, and coexistence of epidemic Flavobacterium psychrophilum lineages in Chilean fish farms. Veterinary Microbiology. 170, 298-306. Barnes, M.E., Brown, M.L. 2011. A review of Flavobacterium psychrophilum biology, clinical signs, and bacterial coldwater disease prevention and treatment. The Open Fish Science Journal. 4, 40-48. Brown, L., Cox, W., Levine, R. 1997. Evidence that the causal agent of bacterial coldwater disease Flavobacterium psychrophilum is transmitted within salmonid eggs. Diseases of Aquatic Organisms. 29(3), 213-218. Bruun, M.S., Schmidt, A.S., Madsen, L., Dalsgaard, I. 2000. Antimicrobial resistance patterns in Danish isolates of Flavobacterium psychrophilum. Aquaculture. 187(3-4), 201-212. Cai, W., De La Fuente, L., Arias, C.R. 2013. Biofilm formation by the fish pathogen Flavobacterium columnare: development and parameters affecting surface attachment. Applied and Environmental Microbiology. 79(18), 5633-5642. Castillo, D., Donati, V.L., Jorgensen, J., Sundell, K., Dalsgaard, I., Madsen, L., Wiklund, T., Middelboe, M. 2021. Comparative genomic analyses of Flavobacterium psychrophilum isolates reveals new putative genetic determinants of virulence traits. Microorganisms. 9(8), 1658. Cipriano, R.C., Holt, R.A. 2005. Flavobacterium psychrophilum, cause of bacterial cold-water disease and rainbow trout fry syndrome Fish Disease Leaflet, no. 86. United States Department of the Interior. U.S. Geological Service, National Fish Health Research Laboratory, Kearneysville, WV. Dauda, A.B., Ajadi, A., Tola-Fabunmi, A.S., Akinwole, A.O. 2019. Waste production in aquaculture: sources, components, and managements in different culture systems. Aquaculture and Fisheries. 4(3), 81-88. Davis, H.S. 1946. Care and diseases of trout. United States Fish and Wildlife Service Research Report 12. (98pp). Duchaud, E., Boussaha, M., Loux, V., Bernardet, J-F., Michel, C., Kerouault, B., Stanislas Mondot, S., Nicolas, P., Bossy, R., Caron, C., Bessières, P., Gibrat, J-F., Claverol, S., Dumetz, F., Hénaff, M., Benmansour, A. 2007. Complete genome sequence of the fish pathogen Flavobacterium psychrophilum. Nature Biotechnology. 25, 763-769. 195 Fujiwara-Nagata, E., Chantry-Darmon, C., Bernardet, J-F., Eguchi, M., Duchaud, E., & Nicolas, P. 2013. Population structure of the fish pathogen Flavobacterium psychrophilum at whole- country and model river levels in Japan. Veterinary Research. 44, 34. Gomez, E., Mendez, J., Cascales, D., Guijarro, J.A. 2014. Flavobacterium psychrophilum vaccine development: a difficult task. Microbial biotechnology. 7(5), 414-423. Hardy, R.W., Gatlin, D. 2002. Nutritional strategies to reduce nutrient losses in intensive aquaculture. In: Cruz-Suarez, L.E., Ricque-Marie, D., Tapia-Salazar, M., Gaxiola-Cortes, M.G., Simoes, N. (Eds.). Cancun, Quintana Roo, Mexico. Harrison, C.E., Knupp, C.K., Brenden, T.O., Ebener, M.P., Loch, T.P. 2022. First isolation of Flavobacterium psychrophilum from wild adult Great lakes lake whitefish (Coregonus clupeaformisi). Journal of Fish Diseases. 45(7), 1023-1032. Holt, R.A. 1987. PhD thesis. Cytophaga psychrophila, the causative agent of bacterial cold water disease in salmonid fish. Oregon State University, Corvallis, OR. Kennedy, D.A., Kurath, G., Brito, I.L., Purcell, M.K., Read, A.F., Winton, J.R., Wargo, A.R. 2015. Potential drivers of virulence evolution in aquaculture. Evolutionary Applications. 9(2), 344-354. Kinnula, H., Mappes, J., Valkonen, J.K., Pulkkinen, K., Sundberg, L-R. 2017. Higher resource level promotes virulence in an environmentally transmitted bacterial fish pathogen. Evolutionary Applications. 10(5), 462-470. Knupp, C., Wiens, G.D., Faisal, M., Call, D.R., Cain, K.D., Nicolas, P., Van Vliet, D., Yamashita, C., Ferguson, J.A., Meuninck, D., Hsu, H-M., Baker, B.B., Shen, L., Loch, T.P. 2019. Large-scale analysis of Flavobacterium psychrophilum multilocus sequence typing genotypes indicates that both newly identified and recurrent clonal complexes are associated with disease. Applied and Environmental Microbiology. 86, e02305-18. Knupp, C., Kiupel, M., Brenden, T.O., Loch, T.P. 2021a. Host-specific preference of some Flavobacterium psychrophilum multilocus sequence typing genotypes determines their ability to cause bacterial coldwater disease in coho salmon (Oncorhynchus kisutch). Journal of Fish Diseases. 44(5), 521-531. Knupp, C., Faisal, M., Brenden, T.O., Wiens, G.D., Loch, T.P. 2021b. In vivo experiments provide evidence that Flavobacterium psychrophilum strains belonging to multilocus sequence typing clonal complex ST191 are virulent to rainbow trout (Oncorhynchus mykiss). Journal of Aquatic Animal Health. 33(3), 190-195. Knupp, C., Faisal, M., Brenden, T.O., Soto, E., LaFrentz, B.R., Griffin, M.J., Wiens, G.D., Cavender, W., Van Vliet, D., Loch, T.P. 2023. Ultraviolet light differentially reduces viability of fish- and fish-farm associated flavobacteria (Family Flavobacteriaceae and Weeksellaceae). North American Journal of Aquaculture. In press. 196 Krehenbrink, M., Oppermann-Sanio, F.B., Steinbüchel, A. 2002. Evaluation of non- cyanobacterial genome sequences for occurrence of genes encoding proteins homologous to cyanophycin synthetase and cloning of an active cyanophycin synthetase from Acinetobacter sp. strain DSM 587. Archives of Microbiology. 177, 371–380. Kum C., Kirkan S., Sekkin S., Akar F., Boyacioglu, M. 2008. Comparison of in vitro antimicrobial susceptibility in Flavobacterium psychrophilum isolated from rainbow trout fry. Journal of Aquatic Animal Health. 20, 245-51. LaFrentz, B.R., Králová, S., Burbick, C.R., Alexander, T.L., Phillips, C.W., Griffin, M.J., Waldbieser, G.C., García, J.C., Alexandre Sebastião, F., Soto, E., Loch, T.P., Liles, M.R., Snekvik, K.R. 2022. The fish pathogen Flavobacterium columnare represents four distinct species: Flavobacterium columnare, Flavobacterium covae sp. nov., Flavobacterium davisii sp. nov. and Flavobacterium oreochromis sp. nov., and emended description of Flavobacterium columnare. Systematic and Applied Microbiology. 45(2),126293. Li, S., Chai, J., Cao, Y., Knupp, C., Wang, D., Nicolas, P., Chen, F., Liu, H., Lu, T., Loch, T.P. 2021. Characterization and molecular epidemiological analysis of Flavobacterium psychrophilum recovered from diseased salmonids in China. Microbiology Spectrum. 9(2), e00033021. Loch, T.P., Faisal, M. Flavobacterium spp. In Fish viruses and bacteria: Pathobiology and protection., Woo, P.T.K., Cipriano, R.C., Eds., CABI: 2017, pp. 211-232. Madetoja, J., Nyman, P., Wiklund, T. 2000. Flavobacterium psychrophilum, invasion into and shedding by rainbow trout Oncorhynchus mykiss. Diseases of Aquatic Organisms, 43, 27-38. Madetoja, J., Nystedt, S., Wiklund, T. (2003). Survival and virulence of Flavobacterium psychrophilum in water microcosms. FEMS Microbiology Ecology. 32, 217-223. Nilsen H., Olsen A.B., Vaagnes Ø., Hellberg, H., Bottolfsen, K., Skjelstad, H., Colquhoun, D.J. 2011. Systemic Flavobacterium psychrophilum infection in rainbow trout Oncorhynchus mykiss (Walbaum) farmed in fresh and brackish water in Norway. Journal of Fish Diseases. 34(5), 403- 408. Nilsen, H., Sundell, K., Duchaud, E., Nicolas, P., Dalsgaard, I., Madsen, L., Aspan, A., Jansson, E., Colquhoun, D.J., Wiklund, T. 2014. Multilocus sequence typing identifies epidemic clones of Flavobacterium psychrophilum in Nordic countries. Applied and Environmental Microbiology. 80, 2728 –2736. Oplinger, R.W., Wagner, E.J. 2012. Effects of media ingredient substitution and comparison of growth of Flavobacterium psychrophilum among four media. Journal of Aquatic Animal Health. 24, 49-57. 197 Pacha, R.E., Porter, S. 1970. Myxobacterial disease of salmonids. In: Snieszko SF (ed) A symposium on disease of fishes and shellfishes, Special publication No. 5. Washington D.C, American Fisheries Society, pp 243-257. Schumann, M.F. 2021. PhD thesis. Suspended solids in salmonid aquaculture: Extended insights and new approaches for control. University of Konstanz, Konstanz, Germany. Starliper, C.E. 2011. Bacterial coldwater disease of fishes caused by Flavobacterium psychrophilum. Journal of Advanced Research. 2, 97-108. Strepparava, N., Wahli, T., Segner, H., Petrini, O. 2014. Detection and quantification of Flavobacterium psychrophilum in water and fish tissue samples by quantitative real time PCR. BMC Microbiology. 14, 105. Summerfelt, R.C. 2000. Water quality considerations for aquaculture. Department of Animal Ecology, pp2-7. Iowa State university, Ames, Iowa. Sundell, K., Landor, L., Nicolas, P., Jørgensen, J., Castillo, D., Middelboe, M., Dalsgaard, I., Donati, V. L., Madsen, L., Wiklund, T. 2019. Phenotypic and genetic predictors of pathogenicity and virulence in Flavobacterium psychrophilum. Frontiers in Microbiology. 10, 1711. Van Vliet, D., Loch, T.P., Faisal, M. 2015. Flavobacterium psychrophilum infections in salmonid broodstock and hatchery-propagated stocks of the Great Lakes Basin. Journal of Aquatic Animal Health. 27,192–202. Van Vliet, D., Wiens, G.D., Loch, T.P., Nicolas, P., M. Faisal. 2016. Genetic diversity of Flavobacterium psychrophilum isolates from three Oncorhynchus spp. in the United States, as revealed by multilocus sequence typing. Applied and Environmental Microbiology. 82, 3246– 3255. Van Vliet, D., Loch T.P., Smith, P., Faisal, M. 2017. Antimicrobial susceptibilities of Flavobacterium psychrophilum isolates from the Great Lakes basin, Michigan. Microbial Drug Resistance. 23, 791-798. Vatsos, I.N., Thompson, K.D., Adams, A. 2003. Starvation of Flavobacterium psychrophilum in broth, stream water and distilled water. Diseases of Aquatic Organisms. 56, 115-126. 198 APPENDIX Table 5.1. Flavobacterium psychrophilum isolates selected for this study. The multilocus sequence typing clonal complex (CC), sequence type (ST; i.e., genetic variant), recovery location, host of origin, year of isolation, and starting concentration (i.e., at week 0) in the microcosm experiment are presented for each isolate. Recovery Host Isolate CC ST Year Referencec locationa speciesb US019 9 13 MI-W-1 COS 2010 Van Vliet et al. (2016) US075 10 10 PA-H-1 STT 2016 Knupp et al. (2019) US053 10 78 MI-H-2 STT 2011 Van Vliet et al. (2016) US464 10 275 MI-H-3 STT 2019 This study US524 10 342 MI-H-2 STT 2021 This study US531 191 267 MI-H-2 STT 2021 This study US062 232 277 MI-W-2 ATS 2012 Knupp et al. (2019) US461 286 286 MI-H-4 BNT 2019 This study US503 253 MI-H-5 BNT 2020 This study US487 353 MI-H-6 BKT 2020 This study a Key to recovery location: U.S. state (MI, Michigan; PA, Pennsylvania) – Facility (H, hatchery; W, weir) – Site code (1 – 6). b Key to host species: ATS = Atlantic salmon (Salmo salar); BKT = Brook trout (Salvelinus fontinalis); BNT = Brown trout (S. trutta); COS = Coho salmon (Oncorhynchus kisutch); STT = Steelhead trout (O. mykiss). c Link to pubMLST database for F. psychrophilum: https://pubmlst.org/fpsychrophilum/. 199 Table 5.2. Weekly mean concentrations (in log10 colony forming units) ± (standard error) of the ten Flavobacterium psychrophilum sequence types (i.e., genetic variants) used in this study, as measured in the well water only microcosm. Week ST 0 1 2 3 4 5 6 7 8 13 13 8.10 (0.00) 7.37 (0.37) 6.30 (0.30) 6.50 (0.20) 6.09 (0.09) 6.20 (0.20) 5.97 (0.57) 5.68 (0.50) 5.89 (0.41) 5.74 (0.56) 10 8.15 (0.00) 7.84 (0.06) 7.00 (0.00) 6.76 (0.11) 6.59 (0.29) 6.09 (0.09) 5.48 (0.00) 4.97 (0.57) 5.39 (0.21) 5.27 (0.13) 78 8.25 (0.00) 8.29 (0.11) 7.24 (0.06) 6.39 (0.39) 6.20 (0.20) 6.27 (0.27) 5.63 (0.03) 5.54 (0.06) 6.00 (0.00) 5.09 (0.39) 275 8.15 (0.00) 8.09 (0.09) 7.63 (0.15) 6.72 (0.02) 6.59 (0.11) 6.30 (0.00) 6.09 (0.09) 6.00 (0.00) 6.18 (0.00) 5.74 (0.44) 342 7.85 (0.00) 7.79 (0.25) 7.18 (0.00) 7.18 (0.00) 6.94 (0.01) 6.85 (0.00) 6.58 (0.40) 6.13 (0.35) 6.12 (0.19) 5.97 (0.07) 267 7.33 (0.00) 7.39 (0.09) 7.72 (0.12) 6.24 (0.24) 6.14 (0.40) 6.00 (0.00) 5.50 (0.20) 5.35 (0.350 5.79 (0.39) 5.50 (0.24) 277 8.15 (0.00) 6.30 (0.00) 4.50 (0.20) 5.00 (0.00) 5.35 (0.05) 6.09 (0.09) 5.50 (0.10) 5.57 (0.09) 5.57 (0.09) 5.18 (0.00) 286 8.15 (0.00) 7.67 (0.07) 7.48 (0.00) 7.39 (0.09) 7.15 (0.15) 6.54 (0.00) 5.64 (0.10) 5.77 (0.23) 6.15 (0.15) 5.89 (0.41) 253 8.15 (0.00) 6.91 (0.07) 7.06 (0.00) 6.53 (0.35) 6.35 (0.35) 6.40 (0.00) 6.24 (0.24) 5.89 (0.41) 6.00 (0.30) 5.44 (0.44) 353 8.15 (0.00) 7.62 (0.08) 6.65 (0.00) 6.45 (0.15) 6.00 (0.00) 6.00 (0.00) 6.00 (0.00) 5.64 (0.10) 6.00 (0.00) 5.00 (0.00) 200 Table 5.3. Weekly mean concentrations (in log10 colony forming units) ± (standard error) of the three experimental microcosms (e.g., well water, well water with raceway detritus or commercial trout feed). Week Microcosm 0 1 2 3 4 5 6 7 8 13 8.04 7.53 6.88 6.52 6.34 6.27 5.86 5.65 5.91 5.48 Well water (0.06) (0.13) (0.21) (0.15) (0.12) (0.06) (0.10) (0.10) (0.08) (0.11) Raceway 8.04 8.47 8.54 8.43 8.52 8.33 7.91 7.06 6.84 4.56 detritus (0.06) (0.04) (0.04) (0.08) (0.05) (0.08) (0.16) (0.28) (0.25) (0.47) Commercial 8.04 6.54 5.66 5.72 5.98 6.18 5.82 5.51 5.44 5.41 trout feed (0.06) (0.13) (0.22) (0.42) (0.41) (0.33) (0.36) (0.41) (0.37) (0.42) 201 Table 5.4. Differences of least square mean estimates ± standard error (SE) for the interaction between treatment group (e.g., raceway “detritus”, commercial trout “feed”, or “water” only) and ten Flavobacterium psychrophilum variants (e.g., ST13, ST78, ST277, ST10, ST286, ST275, ST353, ST253, ST342, ST267). Pairwise comparisons between variants within the same treatment are provided. Tukey-Kramer adjusted P-values for multiple comparisons are shown (α = 0.05). Table is ordered by treatment. Treatment 1 Isolate 1 Treatment 2 Isolate 2 Estimate SE DF t-value P-value Detritus ST13 Detritus ST78 0.2894 0.2502 543 1.16 1.0000 Detritus ST13 Detritus ST277 -0.1201 0.2502 543 -0.48 1.0000 Detritus ST13 Detritus ST10 0.4208 0.2502 543 1.68 0.9983 Detritus ST13 Detritus ST286 0.9224 0.2502 543 3.69 0.0654 Detritus ST13 Detritus ST275 0.2385 0.2502 543 0.95 1.0000 Detritus ST13 Detritus ST353 1.6799 0.2502 543 6.71 <.0001 Detritus ST13 Detritus ST253 -0.2631 0.2502 543 -1.05 1.0000 Detritus ST13 Detritus ST342 0.1787 0.2502 543 0.71 1.0000 Detritus ST13 Detritus ST267 0.02690 0.2502 543 0.11 1.0000 Detritus ST78 Detritus ST277 -0.4095 0.2502 543 -1.64 0.9989 Detritus ST78 Detritus ST10 0.1315 0.2502 543 0.53 1.0000 Detritus ST78 Detritus ST286 0.6330 0.2502 543 2.53 0.7582 Detritus ST78 Detritus ST275 -0.05086 0.2502 543 -0.20 1.0000 Detritus ST78 Detritus ST353 1.3906 0.2502 543 5.56 <.0001 Detritus ST78 Detritus ST253 -0.5525 0.2502 543 -2.21 0.9288 Detritus ST78 Detritus ST342 -0.1107 0.2502 543 -0.44 1.0000 Detritus ST78 Detritus ST267 -0.2625 0.2502 543 -1.05 1.0000 Detritus ST277 Detritus ST10 0.5410 0.2502 543 2.16 0.9433 Detritus ST277 Detritus ST286 1.0425 0.2502 543 4.17 0.0119 Detritus ST277 Detritus ST275 0.3587 0.2502 543 1.43 0.9999 Detritus ST277 Detritus ST353 1.8001 0.2502 543 7.20 <.0001 Detritus ST277 Detritus ST253 -0.1430 0.2502 543 -0.57 1.0000 Detritus ST277 Detritus ST342 0.2988 0.2502 543 1.19 1.0000 Detritus ST277 Detritus ST267 0.1471 0.2502 543 0.59 1.0000 Detritus ST10 Detritus ST286 0.5015 0.2502 543 2.00 0.9771 Detritus ST10 Detritus ST275 -0.1823 0.2502 543 -0.73 1.0000 Detritus ST10 Detritus ST353 1.2591 0.2502 543 5.03 0.0003 Detritus ST10 Detritus ST253 -0.6840 0.2502 543 -2.73 0.6009 Detritus ST10 Detritus ST342 -0.2422 0.2502 543 -0.97 1.0000 Detritus ST10 Detritus ST267 -0.3939 0.2502 543 -1.57 0.9995 Detritus ST286 Detritus ST275 -0.6839 0.2502 543 -2.73 0.6013 Detritus ST286 Detritus ST353 0.7576 0.2502 543 3.03 0.3673 Detritus ST286 Detritus ST342 -0.7437 0.2502 543 -2.97 0.4086 Detritus ST286 Detritus ST267 -0.8955 0.2502 543 -3.58 0.0914 Detritus ST275 Detritus ST353 1.4414 0.2502 543 5.76 <.0001 Detritus ST275 Detritus ST253 -0.5016 0.2502 543 -2.01 0.9771 Detritus ST275 Detritus ST342 -0.05985 0.2502 543 -0.24 1.0000 Detritus ST275 Detritus ST267 -0.2116 0.2502 543 -0.85 1.0000 Detritus ST353 Detritus ST253 -1.9431 0.2502 543 -7.77 <.0001 Detritus ST353 Detritus ST342 -1.5013 0.2502 543 -6.00 <.0001 Detritus ST353 Detritus ST267 -1.6530 0.2502 543 -6.61 <.0001 Detritus ST253 Detritus ST342 0.4418 0.2502 543 1.77 0.9963 Detritus ST253 Detritus ST267 0.2900 0.2502 543 1.16 1.0000 Detritus ST342 Detritus ST267 -0.1518 0.2502 543 -0.61 1.0000 Feed ST13 Feed ST78 -1.1281 0.2502 543 -4.51 0.0029 Feed ST13 Feed ST277 -1.3593 0.2502 543 -5.43 <.0001 202 Table 5.4. (cont’d) Treatment 1 Isolate 1 Treatment 2 Isolate 2 Estimate SE DF t-value P-value Feed ST13 Feed ST10 -0.2469 0.2502 543 -0.99 1.0000 Feed ST13 Feed ST286 -0.7860 0.2502 543 -3.14 0.2891 Feed ST13 Feed ST275 -1.6210 0.2502 543 -6.48 <.0001 Feed ST13 Feed ST353 -1.4694 0.2502 543 -5.87 <.0001 Feed ST13 Feed ST253 -1.2197 0.2502 543 -4.88 0.0006 Feed ST13 Feed ST342 -1.3749 0.2502 543 -5.50 <.0001 Feed ST13 Feed ST267 2.0485 0.2502 543 8.19 <.0001 Feed ST78 Feed ST277 -0.2312 0.2502 543 -0.92 1.0000 Feed ST78 Feed ST10 0.8812 0.2502 543 3.52 0.1083 Feed ST78 Feed ST286 0.3421 0.2502 543 1.37 1.0000 Feed ST78 Feed ST275 -0.4929 0.2502 543 -1.97 0.9818 Feed ST78 Feed ST353 -0.3412 0.2502 543 -1.36 1.0000 Feed ST78 Feed ST253 -0.09153 0.2502 543 -0.37 1.0000 Feed ST78 Feed ST342 -0.2468 0.2502 543 -0.99 1.0000 Feed ST78 Feed ST267 3.1766 0.2502 543 12.70 <.0001 Feed ST277 Feed ST10 1.1124 0.2502 543 4.45 0.0038 Feed ST277 Feed ST286 0.5733 0.2502 543 2.29 0.8963 Feed ST277 Feed ST275 -0.2617 0.2502 543 -1.05 1.0000 Feed ST277 Feed ST353 -0.1101 0.2502 543 -0.44 1.0000 Feed ST277 Feed ST253 0.1396 0.2502 543 0.56 1.0000 Feed ST277 Feed ST342 -0.01564 0.2502 543 -0.06 1.0000 Feed ST277 Feed ST267 3.4078 0.2502 543 13.62 <.0001 Feed ST10 Feed ST286 -0.5391 0.2502 543 -2.15 0.9455 Feed ST10 Feed ST275 -1.3741 0.2502 543 -5.49 <.0001 Feed ST10 Feed ST353 -1.2224 0.2502 543 -4.89 0.0005 Feed ST10 Feed ST253 -0.9727 0.2502 543 -3.89 0.0333 Feed ST10 Feed ST342 -1.1280 0.2502 543 -4.51 0.0029 Feed ST10 Feed ST267 2.2954 0.2502 543 9.18 <.0001 Feed ST286 Feed ST275 -0.8350 0.2502 543 -3.34 0.1804 Feed ST286 Feed ST353 -0.6834 0.2502 543 -2.73 0.6030 Feed ST286 Feed ST253 -0.4336 0.2502 543 -1.73 0.9972 Feed ST286 Feed ST342 -0.5889 0.2502 543 -2.35 0.8664 Feed ST286 Feed ST267 2.8345 0.2502 543 11.33 <.0001 Feed ST275 Feed ST353 0.1516 0.2502 543 0.61 1.0000 Feed ST275 Feed ST253 0.4013 0.2502 543 1.60 0.9992 Feed ST275 Feed ST342 0.2461 0.2502 543 0.98 1.0000 Feed ST275 Feed ST267 3.6695 0.2502 543 14.67 <.0001 Feed ST353 Feed ST253 0.2497 0.2502 543 1.00 1.0000 Feed ST353 Feed ST342 0.09443 0.2502 543 0.38 1.0000 Feed ST353 Feed ST267 3.5179 0.2502 543 14.06 <.0001 Feed ST253 Feed ST342 -0.1553 0.2502 543 -0.62 1.0000 Feed ST253 Feed ST267 3.2682 0.2502 543 13.06 <.0001 Feed ST342 Feed ST267 3.4234 0.2502 543 13.68 <.0001 Water ST13 Water ST78 -0.1061 0.2502 543 -0.42 1.0000 Water ST13 Water ST277 0.6633 0.2502 543 2.65 0.6676 Water ST13 Water ST10 0.02875 0.2502 543 0.11 1.0000 Water ST13 Water ST286 -0.4008 0.2502 543 -1.60 0.9993 Water ST13 Water ST275 -0.3650 0.2502 543 -1.46 0.9999 Water ST13 Water ST353 0.03091 0.2502 543 0.12 1.0000 Water ST13 Water ST253 -0.1133 0.2502 543 -0.45 1.0000 Water ST13 Water ST342 -0.4748 0.2502 543 -1.90 0.9891 Water ST13 Water ST267 0.08658 0.2502 543 0.35 1.0000 Water ST78 Water ST277 0.7694 0.2502 543 3.08 0.3335 203 Table 5.4. (cont’d) Treatment 1 Isolate 1 Treatment 2 Isolate 2 Estimate SE DF t-value P-value Water ST78 Water ST10 0.1349 0.2502 543 0.54 1.0000 Water ST78 Water ST286 -0.2947 0.2502 543 -1.18 1.0000 Water ST78 Water ST275 -0.2588 0.2502 543 -1.03 1.0000 Water ST78 Water ST353 0.1371 0.2502 543 0.55 1.0000 Water ST78 Water ST253 -0.00712 0.2502 543 -0.03 1.0000 Water ST78 Water ST342 -0.3686 0.2502 543 -1.47 0.9998 Water ST78 Water ST267 0.1927 0.2502 543 0.77 1.0000 Water ST277 Water ST10 -0.6345 0.2502 543 -2.54 0.7540 Water ST277 Water ST286 -1.0641 0.2502 543 -4.25 0.0085 Water ST277 Water ST275 -1.0282 0.2502 543 -4.11 0.0148 Water ST277 Water ST353 -0.6323 0.2502 543 -2.53 0.7601 Water ST277 Water ST253 -0.7765 0.2502 543 -3.10 0.3140 Water ST277 Water ST342 -1.1380 0.2502 543 -4.55 0.0025 Water ST277 Water ST267 -0.5767 0.2502 543 -2.31 0.8902 Water ST10 Water ST286 -0.4296 0.2502 543 -1.72 0.9976 Water ST10 Water ST275 -0.3937 0.2502 543 -1.57 0.9995 Water ST10 Water ST353 0.002160 0.2502 543 0.01 1.0000 Water ST10 Water ST253 -0.1420 0.2502 543 -0.57 1.0000 Water ST10 Water ST342 -0.5035 0.2502 543 -2.01 0.9760 Water ST10 Water ST267 0.05783 0.2502 543 0.23 1.0000 Water ST286 Water ST275 0.03587 0.2502 543 0.14 1.0000 Water ST286 Water ST353 0.4317 0.2502 543 1.73 0.9974 Water ST286 Water ST253 0.2876 0.2502 543 1.15 1.0000 Water ST286 Water ST342 -0.07394 0.2502 543 -0.30 1.0000 Water ST286 Water ST267 0.4874 0.2502 543 1.95 0.9843 Water ST275 Water ST353 0.3959 0.2502 543 1.58 0.9994 Water ST275 Water ST253 0.2517 0.2502 543 1.01 1.0000 Water ST275 Water ST342 -0.1098 0.2502 543 -0.44 1.0000 Water ST275 Water ST267 0.4515 0.2502 543 1.80 0.9948 Water ST353 Water ST253 -0.1442 0.2502 543 -0.58 1.0000 Water ST353 Water ST342 -0.5057 0.2502 543 -2.02 0.9746 Water ST353 Water ST267 0.05567 0.2502 543 0.22 1.0000 Water ST253 Water ST342 -0.3615 0.2502 543 -1.44 0.9999 Water ST253 Water ST267 0.1998 0.2502 543 0.80 1.0000 Water ST342 Water ST267 0.5613 0.2502 543 2.24 0.9160 204 Table 5.5. Weekly mean concentrations (in log10 colony forming units) ± (standard error) of the ten Flavobacterium psychrophilum sequence types (i.e., genetic variants) used in this study, as measured in the well water with raceway detritus microcosm. Week ST 0 1 2 3 4 5 6 7 8 13 13 8.10 (0.00) 8.60 (0.05) 8.74 (0.04) 8.87 (0.06) 8.70 (0.04) 8.74 (0.14) 8.57 (0.09) 7.89 (0.41) 7.24 (0.24) 4.65 (0.25) 10 8.15 (0.00) 8.39 (0.21) 8.35 (0.05) 8.20 (0.20) 8.44 (0.26) 8.00 (0.00) 7.70 (0.04) 7.22 (0.74) 6.70 (0.00) 4.74 (0.19) 78 8.25 (0.00) 8.65 (0.05) 8.69 (0.09) 8.65 (0.05) 8.70 (0.04) 8.51 (0.03) 8.18 (0.00) 7.18 (0.00) 6.50 (0.50) 3.89 (0.41) 275 8.15 (0.00) 8.57 (0.03) 8.72 (0.02) 8.54 (0.06) 8.53 (0.13) 7.90 (0.00) 7.54 (0.06) 6.99 (0.18) 6.91 (0.09) 5.85 (0.00) 342 7.85 (0.00) 8.57 (0.09) 8.70 (0.04) 8.30 (0.00) 8.60 (0.00) 8.48 (0.00) 8.33 (0.15) 7.24 (0.06) 7.30 (0.00) 4.94 (0.19) 267 7.33 (0.00) 8.33 (0.15) 8.54 (0.06) 8.46 (0.28) 8.69 (0.09) 8.60 (0.00) 8.41 (0.24) 7.59 (0.59) 7.87 (0.03) 5.99 (0.21) 277 8.15 (0.00) 8.60 (0.05) 8.55 (0.15) 8.63 (0.03) 8.63 (0.03) 8.68 (0.28) 8.42 (0.042) 8.57 (0.09) 7.70 (0.04) 5.37 (0.37) 286 8.15 (0.00) 8.44 (0.04) 8.42 (0.12) 8.39 (0.09) 8.30 (0.00) 8.09 (0.09) 7.30 (0.00) 5.83 (0.83) 5.50 (0.50) 2.44 (0.73) 253 8.15 (0.00) 8.45 (0.15) 8.54 (0.00) 8.60 (0.30) 8.59 (0.11) 8.48 (0.00) 8.35 (0.05) 7.73 (0.08) 8.10 (0.20) 7.72 (0.41) 353 8.15 (0.00) 8.15 (0.15) 8.20 (0.20) 7.70 (0.00) 8.00 (0.00) 7.85 (0.00) 6.30 (0.00) 4.35 (0.95) 4.59 (0.59) 0.00 (0.00) 205 Table 5.6. Weekly mean concentrations (in log10 colony forming units) ± (standard error) of the ten Flavobacterium psychrophilum sequence types (i.e., genetic variants) used in this study, as measured in the well water with commercial trout feed microcosm. Week ST 0 1 2 3 4 5 6 7 8 13 13 8.10 (0.00) 6.09 (0.09) 4.00 (0.00) 4.24 (0.24) 6.20 (0.20) 6.64 (0.10) 6.09 (0.09) 4.00 (2.00) 4.15 (2.15) 3.65 (1.65) 10 8.15 (0.00) 6.91 (0.07) 4.09 (0.09) 6.00 (0.00) 5.30 (0.30) 6.00 (0.00) 3.00 (0.00) 4.94 (1.54) 4.65 (1.65) 6.59 (0.29) 78 8.25 (0.00) 6.72 (0.02) 6.18 (0.00) 5.59 (0.41) 6.35 (0.35) 6.27 (0.27) 6.30 (0.30) 6.30 (0.30) 6.70 (0.04) 5.79 (0.61) 275 8.15 (0.00) 7.00 (0.00) 6.57 (0.03) 7.29 (0.11) 7.50 (0.20) 6.90 (0.00) 6.72 (0.02) 6.53 (0.13) 6.29 (0.11) 6.42 (0.12) 342 7.85 (0.00) 6.78 (0.18) 5.24 (0.24) 6.51 (0.03) 6.60 (0.3) 6.85 (0.00) 7.05 (0.35) 6.83 (0.02) 6.86 (0.01) 6.34 (0.05) 267 7.33 (0.00) 7.50 (0.10) 5.09 (0.09) 1.00 (1.00) 1.00 (1.00) 2.00 (0.00) 3.04 (1.04) 2.00 (0.00) 2.70 (0.00) 1.00 (1.00) 277 8.15 (0.00) 6.00 (0.00) 6.27 (0.27) 6.61 (0.13) 7.00 (0.00) 7.37 (0.37) 6.65 (0.48) 6.37 (0.37) 6.09 (0.09) 6.24 (0.06) 286 8.15 (0.00) 6.39 (0.09) 6.35 (0.05) 6.42 (0.42) 6.30 (0.30) 6.30 (0.00) 5.65 (0.65) 4.74 (1.56) 4.54 (1.06) 6.18 (0.00) 253 8.15 (0.00) 6.49 (0.32) 6.20 (0.20) 6.33 (0.15) 7.00 (0.00) 6.48 (0.00) 6.09 (0.09) 6.44 (0.26) 5.80 (0.10) 6.38 (0.36) 353 8.15 (0.00) 5.50 (0.20) 6.65 (0.05) 7.24 (0.06) 6.59 (0.11) 7.00 (0.00) 7.59 (0.11) 7.00 (0.60) 6.59 (0.59) 5.55 (1.15) 206 Table 5.7. Differences of least square mean estimates ± standard error (SE) for the interaction between treatment group (e.g., raceway “detritus”, commercial trout “feed”, or “water” only) and ten Flavobacterium psychrophilum variants (e.g., ST13, ST78, ST277, ST10, ST286, ST275, ST353, ST253, ST342, ST267). Pairwise comparisons between treatments for each variant are provided. Tukey-Kramer adjusted P-values for multiple comparisons are shown (α = 0.05). Treatment 1 Isolate 1 Treatment 2 Isolate 2 Estimate SE DF t-value P-value Detritus ST13 Feed ST13 2.6924 0.2502 543 10.76 <.0001 Detritus ST13 Water ST13 1.6254 0.2502 543 6.50 <.0001 Feed ST13 Water ST13 -1.0670 0.2502 543 -4.26 0.0081 Detritus ST78 Feed ST78 1.2749 0.2502 543 5.10 0.0002 Detritus ST78 Water ST78 1.2299 0.2502 543 4.92 0.0005 Feed ST78 Water ST78 -0.04497 0.2502 543 -0.18 1.0000 Detritus ST277 Feed ST277 1.4532 0.2502 543 5.81 <.0001 Detritus ST277 Water ST277 2.4088 0.2502 543 9.63 <.0001 Feed ST277 Water ST277 0.9556 0.2502 543 3.82 0.0422 Detritus ST10 Feed ST10 2.0246 0.2502 543 8.09 <.0001 Detritus ST10 Water ST10 1.2333 0.2502 543 4.93 0.0004 Feed ST10 Water ST10 -0.7913 0.2502 543 -3.16 0.2758 Detritus ST286 Feed ST286 0.9840 0.2502 543 3.93 0.0284 Detritus ST286 Water ST286 0.3022 0.2502 543 1.21 1.0000 Feed ST286 Water ST286 -0.6818 0.2502 543 -2.73 0.6082 Detritus ST275 Feed ST275 0.8329 0.2502 543 3.33 0.1844 Detritus ST275 Water ST275 1.0220 0.2502 543 4.08 0.0163 Feed ST275 Water ST275 0.1891 0.2502 543 0.76 1.0000 Detritus ST353 Feed ST353 -0.4569 0.2502 543 -1.83 0.9937 Detritus ST353 Water ST353 -0.02363 0.2502 543 -0.09 1.0000 Feed ST353 Water ST353 0.4333 0.2502 543 1.73 0.9972 Detritus ST253 Feed ST253 1.7359 0.2502 543 6.94 <.0001 Detritus ST253 Water ST253 1.7753 0.2502 543 7.10 <.0001 Feed ST253 Water ST253 0.03943 0.2502 543 0.16 1.0000 Detritus ST342 Feed ST342 1.1388 0.2502 543 4.55 0.0024 Detritus ST342 Water ST342 0.9720 0.2502 543 3.89 0.0337 Feed ST342 Water ST342 -0.1668 0.2502 543 -0.67 1.0000 Detritus ST267 Feed ST267 4.7140 0.2502 543 18.84 <.0001 Detritus ST267 Water ST267 1.6851 0.2502 543 6.74 <.0001 Feed ST267 Water ST267 -3.0289 0.2502 543 -12.11 <.0001 207 Figure 5.1. 13-week culturability of Flavobacterium psychrophilum sequence type (ST) 10, ST13, ST78, ST253, ST267, ST275, ST277, ST286, ST342, and ST353 in microcosms containing (A) sterilized well water only, (B) sterilized well water with raceway detritus, or (C) sterilized well water with commercial trout feed. Standard error between replicate flasks on each sampling week (e.g., 0 – 8 and 13) are shown. 208 Chapter 6: Conclusions and future research 209 6.1. Conclusions Despite nearly a century of research, effective prevention and control of bacterial coldwater disease (BCWD) and rainbow trout fry syndrome (RTFS) remains elusive. Recent studies have identified multiple predominating Flavobacterium psychrophilum multilocus sequence typing (MLST) clonal complexes (CCs) and sequence types (STs; i.e., genetic variants) causing BCWD epizootics globally, including in the United States of America (USA). Studies have also begun to advance our understanding of how F. psychrophilum diversity relates to virulence, susceptibility to antimicrobials, host species association, and geographical distribution. Resulting from these advancements, multiple unanswered questions emerged regarding the potential factors associated with the apparent success of some of these variants, such as their affinity for certain host species, the influence of the artificial rearing environment, differences in shedding dynamics among variants and salmonid species, and the adequacy of current gold-standard research/diagnostic culture media. These knowledge gaps are likely contributing to productivity losses for salmonid farms and challenges in hatchery-based conservation efforts by impeding BCWD diagnosis, and the development of effective BCWD management, prevention, and control strategies. Towards filling these knowledge gaps, I conceptualized, designed, and executed a series of in vitro and in vivo experiments. In Chapter 2, I developed two new culture media, F. psychrophilum medium-A and -B (FPM-A and FPM-B), that significantly improved the recovery of a wide diversity of F. psychrophilum variants in the laboratory. Considering most BCWD research projects have a culture component, these new media will be instrumental in helping researchers achieve their study goals, and ultimately will contribute to the mitigation of BCWD-associated losses. Indeed, FPM-A could be considered the backbone of Chapters 3 – 5, as findings of those studies may 210 have been obscured by a less sensitive culture medium. Apart from demonstrating the utility of FPM-A and FPM-B under controlled laboratory conditions, both media were highly successful under field conditions, as evidenced by their recovery of F. psychrophilum from naturally infected Atlantic salmon (Salmo salar), coho salmon (Oncorhynchus kisutch), chinook salmon (O. tshawytscha), and steelhead trout (O. mykiss) broodstock when compared to the current gold- standard medium for F. psychrophilum detection, tryptone yeast extract salts agar. Since the development of FPM-A, it has also been used to recover F. psychrophilum from multiple salmonid species experiencing BCWD in Michigan state fish hatcheries, and thus has been crucial for not only BCWD diagnosis but guiding timely treatment recommendations. Similarly, FPM-A is now also being used by other laboratories and with reportedly good success. Thus far, these new media have shown tremendous promise for not only BCWD research efforts but the widespread diagnosis of BCWD. In Chapter 3, I investigated the host specificity of F. psychrophilum variants US19-COS, US62-ATS, and US87-RBT, which belong to MLST CCs (e.g., CC-ST9, CC-ST232, and CC- ST10) commonly associated with a single salmonid species - coho salmon, Atlantic salmon, and rainbow trout, respectively. This was the first study to simultaneously cross-challenge multiple salmonid species with multiple putatively host specific F. psychrophilum variants via immersion. Such an experimental design was necessary to appropriately assess the host specificity of each variant as it not only controlled for differences in inoculum preparation, age-related susceptibility, and environmental factors but also avoided bypassing external immune defenses. Study results supported MLST-based observations that some F. psychrophilum variants (e.g., US87-RBT) are host specific, whereas others (e.g., US19-COS and US62-ATS) have a wider host range. The underlying pathogen/host mechanisms contributing to these findings require 211 further investigation, but given that each variant belonged to a different serogroup, which may affect how the host “views” the pathogen, it is plausible sero-variation played a role. Nonetheless, assessing the host specificity of F. psychrophilum was crucial as it can guide the selection of variants for inclusion in BCWD vaccines, support the evaluation of vaccine efficacy, provide insight into BCWD transmission dynamics, and aid in the further development of BCWD-resistant salmonids. In Chapter 4, I assessed F. psychrophilum shedding dynamics (e.g., time to shedding, shedding rate and duration) in rainbow trout and, for the first time, in coho salmon and Atlantic salmon. Live fish from all examined species were capable of shedding F. psychrophilum, with notable differences in shedding patterns and durations being observed across host species and F. psychrophilum variants. These differences may imply the existence of diverse transmission strategies, potentially varying based on the specific interaction between the host species and the infecting F. psychrophilum variant. Bridging the findings from Chapters 3 and 4, it seems that host specificity of F. psychrophilum variants may provide distinct advantages. In Chapter 3, US87-RBT was specific to rainbow trout and found to persist in survivors, whereas the other two, more generalist variants, were not. In Chapter 4, US87-RBT displayed the longest shedding duration in live fish and, once again, was the only variant recovered from survivors. Although survivor data for coho salmon was inconclusive due to total mortality, these findings suggest a possible correlation between host specificity, shedding duration, and survival in hosts. Therefore, host specificity may not only influence survival but also play a role in shaping transmission dynamics by affecting the shedding behavior of F. psychrophilum. Dead fish from all examined species were also found capable of shedding F. psychrophilum, and did so at higher rates and for a longer duration compared to live fish. This finding underscores the necessity of efficient 212 management strategies aimed at prompt removal of dead fish from rearing units as dead fish shedding also appears to be an important transmission strategy for F. psychrophilum. Whether the shedding dynamics observed herein are emulated by other F. psychrophilum variants remains to be determined but warrants further investigation as it will lead to a more comprehensive understanding of F. psychrophilum transmission dynamics, and thus potentially inform more widely adoptable BCWD management strategies. In Chapter 5, I assessed the culturability of ten diverse F. psychrophilum variants in microcosms that simulate fish farm/hatchery rearing unit environments. All F. psychrophilum variants demonstrated remarkable adaptability by persisting under nutrient-limited (e.g., well water only) and nutrient-rich (e.g., raceway detritus and commercial trout feed) conditions, suggesting potential environmental reservoirs for this bacterium. However, and for the first time, it was apparent that most F. psychrophilum variants proliferated best in the presence of raceway detritus compared to the other microcosms, as evidenced by maintaining significantly higher concentrations (e.g., ~107 – 108 colony forming units/mL) for several weeks. This finding is particularly significant, as it suggests F. psychrophilum could not only survive but thrive in the presence of raceway detritus, possibly posing a greater transmission risk, thereby underscoring the importance of regularly removing detritus from rearing units. Besides these similarities, it was also evident that some F. psychrophilum variants may be better suited to these environments, which could partially explain their repeated recovery from hatcheries over multiple years. Overall, examining this important aspect of BCWD ecology offered valuable insight into additional factors potentially affecting transmission dynamics. Applying these findings to fish farm and hatchery facilities, it may be beneficial to invest more time and resources into thoroughly cleaning rearing units during disease outbreaks. At the same time, 213 avoiding overfeeding by observing feeding behavior and adjusting as needed may lower F. psychrophilum loads in diseased rearing units and therefore disease spread. These measures could significantly aid in controlling BCWD-associated losses, and enhance the overall health and productivity of fish farms and hatcheries. In total, the findings of my dissertation support the hypothesis that F. psychrophilum intraspecific diversity plays an important role in shaping our understanding of BCWD ecology. In this context, I found F. psychrophilum diversity influences our ability to detect and diagnose BCWD, and affects its interaction with its host and environment. Thus, it is crucial that future studies also consider intraspecific diversity when selecting study isolates to ensure a comprehensive understanding of BCWD ecology, which shapes how we develop and implement BCWD management, prevention, and control strategies. Indeed, this approach holds significant promise for mitigating losses of multiple salmonid species, thereby improving fish farm productivity and hatchery conservation and stock enhancement efforts. 6.2. Future research Although the findings of my dissertation made significant strides in optimizing BCWD research and diagnostic tools, elucidating host-pathogen interactions, unraveling F. psychrophilum transmission dynamics, and exploring its persistence in fish farm and hatchery environments (i.e,. BCWD ecology), they have also brought to light numerous new and unexplored questions. These knowledge gaps, if pursued, hold the potential to further deepen our understanding in each of these facets and, ultimately, enhance our ability to diagnose, manage, and control BCWD more effectively. Building upon Chapter 2, a critical avenue for future research lies in the evaluation of the diagnostic sensitivity and specificity of the newly developed FPM-A and FPM-B culture media 214 via controlled laboratory experiments. The observed limited to non-existent non-target bacterial growth during my field experiments provides promising preliminary evidence of high specificity. However, to validate this, more extensive testing is necessary. Furthermore, refining the specificity of these media by incorporating antibiotics, such as the neomycin sulfate used during field testing, could further suppress non-target bacterial growth. However, the potential impact of this antibiotic (or others) on the recovery of F. psychrophilum variants needs to be assessed to ensure any enhancement in specificity doesn't compromise the sensitivity of the medium. In the context of sensitivity, the new media successfully detected F. psychrophilum from multiple naturally infected salmonids under field conditions and from more fish than TYES. However, controlled laboratory experiments are critical for definitively assessing diagnostic sensitivity, as the true infection prevalence among the fish examined herein was unknown. Another avenue of future research is testing the ability of the new media to recover F. psychrophilum from other tissues, such as reproductive fluids, or environmental samples. Considering disease dynamics are complex in natural and aquaculture settings, the ability to detect and isolate the pathogen from various sources could greatly aid in early detection and effective disease control measures. Finally, comparing the sensitivity of the new culture media to molecular assays (e.g., conventional PCR, qPCR, and LAMP) could offer a more comprehensive evaluation of their performance. This comparison would provide valuable insights into the capability of the media to detect low-level infections and contribute to a broader understanding of the BCWD prevalence and distribution. Chapter 3 marked a significant step forward in the understanding of F. psychrophilum disease ecology by cross-challenging multiple salmonid species via immersion with putatively host specific variants. Yet, to gain a more comprehensive understanding, it would be beneficial 215 to evaluate additional isolates that belong to the same variants tested in this Chapter, as well as variants that have not yet been tested. It could also be insightful to investigate the ability of these variants to cause disease in closely related salmonid species that appear less susceptible to F. psychrophilum infection, such as Chinook salmon. Such an experiment coupled with transcriptomics may help to identify markers that make Chinook salmon less susceptible to BCWD. These findings could then possibly be leveraged to selectively breed other salmonid species to have similar disease-resistance traits. Our study also highlighted that F. psychrophilum variant US62-ATS, belonging to molecular serogroup 1 (seemingly equivalent to serogroup Fd), was highly virulent to rainbow trout and Atlantic salmon. Although serogroup Fd is known to affect rainbow trout, it is not often that this serogroup has been reported in Atlantic salmon. In this context, knowledge regarding the serogroups and serotypes affecting salmonids in the USA is scant but is of utmost importance considering this information will likely be useful in developing a cross-protective BCWD vaccine. Thus, additional studies addressing the serodiversity of salmonids in the USA are warranted. In Chapter 4, I found that both live and dead Atlantic salmon, coho salmon, and rainbow trout shed F. psychrophilum, but dead fish did so at higher rates and for a longer duration. However, the implications of these findings for disease transmission are unclear partly due to the contrasting behaviors of live and dead fish. Live fish, despite shedding less bacteria, might pose a higher transmission risk due to their mobility. In contrast, although dead fish shed higher bacterial loads, they are stationary, which might limit transmission to a localized area. Together, this highlights the need to understand host behavior in conjunction with shedding dynamics. In a similar context, cohabitation studies have proven valuable for understanding disease transmission in rainbow trout, but similar research has yet to be conducted with Atlantic salmon 216 and coho salmon. Hence, future studies should focus on confirming the importance of horizontal transmission in these species. Moreover, comparing the infectivity of F. psychrophilum shed from live and dead fish could significantly influence management strategies. Synthesizing findings from Chapters 3 and 4, the consequences of generalist variants like US62-ATS, which was highly virulent to both Atlantic salmon and rainbow trout should be examined. In this context, future research elucidating whether US62-ATS is more likely to be shed from and successfully infect the same-species (e.g., Atlantic salmon shedding and transmitting to Atlantic salmon) or different-species (e.g., Atlantic salmon shedding and transmitting to rainbow trout) should be investigated as it could have significant implications for transmission in facilities rearing multiple salmonid species. Also observed in Chapters 3 and 4 was the ability of US87- RBT (i.e., the rainbow trout specific variant) to persist in survivors, suggesting some fish become carriers. In this context, understanding the potential role of carriers in perpetuating BCWD epizootics is worth further investigation. Notably, investigating how long carriers harbor F. psychrophilum, whether they shed the bacterium and under what conditions (e.g., times of increased stress), could further contribute to our understanding of BCWD transmission dynamics. Expanding on the insights from Chapter 5, the role of water chemistry in F. psychrophilum proliferation and survival should be considered for future studies. Like the known effect of water hardness on biofilm formation in F. columnare, alterations in water chemistry could potentially influence F. psychrophilum biofilm formation and consequently, this bacterium’s persistence in fish farms and hatcheries. By manipulating water conditions, it may be possible to create an environment less conducive for F. psychrophilum, thereby reducing bacterial loads in fish farms and hatcheries. My findings indicated that the presence of 217 commercial trout feed and raceway detritus enhanced the proliferation of certain F. psychrophilum variants. However, the virulence of F. psychrophilum after being exposed to these conditions for various time periods, and the likelihood of transmission from these sources to fish, remains to be elucidated. Consequently, future studies should focus on these areas to improve our understanding of the complex disease dynamics in fish farms and hatcheries. Furthermore, understanding the mechanisms that facilitate F. psychrophilum persistence outside its host is crucial. Genomic studies focusing on variants exhibiting differential survival capacities could provide key insights into these mechanisms. For instance, exploring the role or existence of cyanophycin, a putative energy storing polymer, in F. psychrophilum variants may shed light on potential strategies this bacterium employs for persistence outside its host. Understanding these aspects would significantly enhance our capacity to mitigate BCWD outbreaks and improve fish farm and hatchery disease management strategies. 218