qwertyuiopasdfghjklzxcvbnmq

wertyuiopasdfghjklzxcvbnmqw

ertyuiopasdfghjklzxcvbnmqwer tyuiopasdfghjklzxcvbnmqwerty uiopasdfghjklzxcvbnmqwertyui Dissertation opasdfghjklzxcvbnmqwertyuiopDoan Thanh Loan asdfghjklzxcvbnmqwertyuiopas dfghjklzxcvbnmqwertyuiopasdf ghjklzxcvbnmqwertyuiopasdfgh jklzxcvbnmqwertyuiopasdfghjkl zxcvbnmqwertyuiopasdfghjklzx cvbnmqwertyuiopasdfghjklzxcv bnmqwertyuiopasdfghjklzxcvbn mqwertyuiopasghjklzxcvbnmq

werty Page 0 of 208 uiopasdfghjklzxcvbnmqwertyui

Microbial composition in biofloc-based shrimp aquaculture systems

A dissertation by

Doan Thanh Loan

2019

Page 1 of 208

STATEMENT OF ORIGINALITY

I, Doan Thanh Loan with this certify that I am the sole author of this thesis and any assistance received is fully acknowledged. To the best of my knowledge, this thesis does not contain any material, which has been previously published or written by others. Any content of other work utilized in this thesis is cited following the standard referencing procedures.

Date:

------

Signed:

------

Page 2 of 208

This thesis was conceived and written at the Leibniz Centre for Tropical Marine Ecology (ZMT) in Bremen and at the Laboratory of Aquaculture & Artemia Reference Centre (ARC) in Belgium between October 2014 and April 2019. This thesis was conducted under the supervision of Dr. Astrid Gärdes, Dr. Andreas Kunzmann, Prof. Dr. Matthias Wolff (ZMT) and Prof. Peter Bossier (ARC), and with non-panel additional supervision of Dr. Christiane Hassenrück (ZMT). Research for this thesis was funded by the Vietnamese Government, ARC, and ZMT. Thesis writing period was supported by Bremen Early Career Researcher Development BYRD (Universität Bremen).

Chairperson: Prof. Dr. Michael Friedrich

Co-examiners: Dr. Astrid Gärdes

Prof. Dr. Matthias Wolff

Prof. Dr. Patrick Sorgeloos

Other non-voting members of the dissertation committee:

Dini Adyasari (ZMT PhD candidate)

Ramón Alejandro Plazas Gómez (ISATEC student)

Date of Defense: 31/07/2019

Page 3 of 208

DEDICATION

I dedicate this work to My parents, Doan Van Lai and Nguyen Thi Lien, My son, Nguyen Bao Kim, All the family members All my teachers

For their endless love and unconditional support

I would hereby like to thank the many people whose help has been invaluable during the completion of this work.

“The Beauty of Science is to Make Things Simple” – Zymo research

Page 4 of 208

ACKNOWLEDGEMENTS

Writing this part has reminded me that I would never have completed this doctoral project without the help and support from my mentors, teachers, colleagues, advocates, family and friends. Many thanks to all!

Specifically, I would first like to express my gratitude to my mentor, Dr Astrid Gärdes, for giving me an excellent chance to work on microbial composition in shrimp aquaculture with the application of biofloc technology at ZMT and ARC. If I had been in Berlin working on Yersinia and small animal diseases at FLI, it would not have been much benefit to the Fisheries Faculty, where I have been working on fish diseases for VNUA. Thanks so much, Astrid!

In addition to my first mentor, I would like to express my deepest gratitude to Prof. Peter Bossier for his insightful comments and encouragement from the MSc thesis and from the moment I gave him the PhD literature review. He was such a patient and understanding advocate, giving me significant opportunities to learn. He listened to all my crazy ideas (e.g. about nitrobacteria, and PHB) whenever his office door was open after 6 pm, and he gave me direction. I will never forget the discussion I had with him in the Challenge Lab about my first findings on shrimp mortality, because he had to take off his shoes and wear the Challenge Lab boots. He slowly informed me that the key finding for my thesis would be TSS, with its role and relationship to operating systems, although I was arguing that it was pH. (I still have his handwritten note on how to calculate NH3 from pH and temperature). Thanks so much, Peter!

I would also like to express my sincere gratitude to Prof. Matthias Wolff, Dr Andreas Kunzmann and Prof. Patrick Sorgeloos, who provided me ideas and comments over the last four years, beginning with key papers provided by Patrick to the end of this thesis with the meetings with panel members. Patrick and Mama Magda are always in my thoughts, such as when I attended a lovely dinner with all their family members, including 11 grandchildren. In addition, Andreas’ office door was always open to my questions and urgent help after 2 pm, and Andreas provided me with many critical papers in forming this thesis story. Thanks so much! Finally, thanks very much to Matthias for being my promotor at Bremen University.

Special gratitude goes out to all the people at Vietnam International Education Development (VIED), ARC, ZMT and BYRD for helping and providing the funding for the work. ZMT is my second home at the moment, where I can learn important knowledge to bring to my university. Page 5 of 208

These benefits should open more opportunities for those at the university, and my future students may be able to come work at ZMT on new projects. There are several people at ZMT whose help during my stay and study I can never forget, namely, Achim, Mercè, Sonja, Constanze, Stefanie,

Christina, Kai, Agostino, Jailson, Jennifer, Inken, Febrina, Nan, Dini, Agustin, and Akuila. As a special mention, Christiane introduced me to R language so that I could understand her guidance in bioinformatics and data analysis. Thank you very much, Christiane! Another special mention goes to Gérman and Anna, who shared their time and unconditional support with me. Mainly, I would like to thank Dr Janine Reinhard, ZMT’s board of directors and members working for ZMT for establishing this second home for many international people, who can benefit from this environment through learning and cooperation.

I thank my lab mates for the stimulating discussions, the sleepless nights we spent working together until 3 am and all the fun we had in the ARC laboratory during the experiments. Special mentions are owed to Tom B, Anita, Brigitte, Caroline, Barbara, Peter S, Jorg, Alex, Geert and Marc, who helped me in this research. It was fantastic to have the opportunity to carry out most of my thesis work in the ARC laboratory. It was an excellent place to work on Litopenaeus vannamei disease and bioflocs, and I gained knowledge and friendships there.

My sincere gratitude is also extended to VNUA’s college, who have been working on my job’s responsibilities. They helped me to have time to stay in ZMT to complete this PhD. Thanks very much!

I am also grateful to the members of my thesis examination committee for their invaluable remarks and suggestions for the improvement of my PhD thesis. I am also thankful for the hard questions which motivated me to widen my research from various perspectives.

Finally, I would like to thank my family members for their prayers and trust, encouraging me to study to the highest level and taking care of my son. Mom, it was hard for me when you broke your wrist while I was away! Many thanks to my son for his understanding of my absence in his important moments at school, for his smiles and laughter, for his kisses and hugs; I love him so much. Thanks to Liem for being with me during the whole time of my studying in Bremen and visiting me in Ghent with jokes and smiles, surprise presents and flowers … Thanks, Liem.

Thank you all for your encouragement!

Page 6 of 208

ABBREVIATIONS

A-Floc/Afloc Autotrophic biofloc system AHL N-acylhomoserine lactone AHPND Acute hepatopancreatic necrosis disease AIC Akaike information criterion BFT Bioflocs Technology C/N carbon/nitrogen CFU Colony forming units DNA Deoxyribonucleic acid DO Dissolved oxygen EMS Early Mortality Syndrome FAO Food and agriculture organization FCR Food conversion ratio FT Flow-through aquaculture system GLMMs General linear mixed models H-Floc/ Hfloc Heterotrophic biofloc system

NH3-N un-ionized ammonia + NH4 -N ammonia-N/ ammonium NMDS Non-metric multidimensional scaling - NO2 -N nitrite-N - NO3 -N nitrate-N OTU Operational taxonomic unit PCA Principal component analysis PCR Polymerase chain reaction PHB Poly-β-hydroxybutyrate pVpAHPND potentially AHPND-causing strain of Vibrio parahaemolyticus RAS Recirculating aquaculture system RDA Redundancy analysis SS Setting solids suc+Vibrio spp. Vibrio bacteria have yellow colour colony suc-Vibrio spp. Vibrio bacteria have green colour colony TAN Total Ammonia Nitrogen TCBS Thisulfate citrate bile salt sucrose agar TIN Total Inorganic Nitrogen TSS Total suspended solids VSS Volatile suspended solids WGR Weight gain rate

Page 7 of 208

TABLE OF CONTENTS

ABSTRACT ------11 CHAPTER 1 ------18 INTRODUCTION ------18

1.1 MAKING A DIFFERENCE THROUGH RESPONSIBLE AQUACULTURE ------19 1.2 MICROBIOTA AND DISEASE IN AQUACULTURE: ISSUES RELATED TO THE VIBRIO GENUS ------21 1.3 STABLE MICROFLORA SYSTEMS ------22 1.4 MICROBIAL FLOCS REPRESENT A STABLE MICROFLORA SYSTEM ------24 1.5 SHRIMP CULTURE WITH BIOFLOC TECHNOLOGY------25 1.6 KNOWLEDGE GAPS AND JUSTIFICATION OF THIS THESIS ------26 1.7 MANUSCRIPT OVERVIEW ------32 CHAPTER 2 ------38 OPPORTUNISTIC PATHOGENS IN INTENSIVE SHRIMP CULTURE SYSTEMS: CONTROL AND MANAGEMENT ------38 ABSTRACT ------38 2.1 INTRODUCTION ------39 2.2 MATERIAL AND METHODS ------42 2.2.1 Experimental design and facilities ------42 2.2.2 Water quality parameters and analyses ------43 2.2.3 Microbiological analysis ------44 2.2.4 Shrimp growth performance analysis ------46 2.2.5 Statistical analysis ------46

2.3 RESULTS ------47 2.3.1 Biofloc performance parameters ------47 2.3.2 Changes in water quality parameters ------47 2.3.3 Changes in microbial components ------48 2.3.4 Shrimp growth parameters ------49 2.3.5 The relation between parameters of environmental conditions, microbial components, and shrimp growth performance ------50

2.4 DISCUSSION ------50 2.4.1 Ecological interactions: r- vs. K- strategists ------50 2.4.2 The role of TSS and pH on microbial control ------51 2.4.3 TSS differences: advantages of heterotrophic bacteria ------52 2.4.4 The H-Floc: the best system for shrimp culture ------53 2.4.5 Selected isolates as indicators ------54 2.4.6 Further studies ------54 CONCLUSIONS------55

Page 8 of 208

CHAPTER 3 ------67 INFLUENCE OF ENVIRONMENTAL FACTORS ON THE BACTERIAL COMMUNITIES OF FOUR DIFFERENT SHRIMP CULTURE SYSTEMS ------67 ABSTRACT ------67 3.1 INTRODUCTION ------68 3.2 MATERIAL AND METHODS ------70 3.2.1 Experimental design and sample collection ------70 3.2.2 Processing of sequencing data ------71 3.2.3 Statistical analysis ------72

3.3 RESULTS ------73 3.3.1 Characteristics of culturing systems ------73 3.3.2 Alpha diversity ------73 3.3.3 Community composition ------74 3.3.4 Healthy and productivity predicted shrimp model ------77

3.4 DISCUSSION ------78 3.4.1 TSS: a primary factor influencing bacterial communities ------78 3.4.2 Bacterial variation by taxonomic sequence abundances ------79 3.4.3 Environmental condition-selected bacteria ------85 3.4.4 Generation of a novel microbial ecological model for Litopenaeus vannamei aquaculture --- 86 3.4.5 Further studies ------87 CONCLUSIONS------87 CHAPTER 4 ------96 CULTIVATION SHRIMP POST-LARVAE CHALLENGED WITH A POTENTIAL PATHOGENIC VIBRIO PARAHAEMOLYTICUS: FINDING AN OPEN GATEWAY TO PREVENT SHRIMP FARMING DISEASE ------96 ABSTRACT ------96 4.1 INTRODUCTION ------97 4.2. MATERIAL AND METHODS ------100 4.2.1 Experimental design------100 4.2.2 Preparation of bacterial strains ------101 4.2.3 Data summary ------102 4.2.4 Statistical analysis of shrimp mortality ------103

4.3 RESULTS ------103 4.3.1 Survival results------103 4.3.2 Microbial distribution in shrimp tissue ------105 4.3.3 Target taxa for Litopenaeus vannamei aquaculture ------107

4.4 DISCUSSION ------107

Page 9 of 208

4.4.1 Strain-specific bacterial concentration as lethal dose ------107 4.4.2 TSS: a vital and challenging factor ------109 4.4.3 Shrimp post larval immune system: An important note ------112 4.4.4 Bio-inoculants as an open gateway for control Vibrio disease ------113 4.4.5 Diagnostic methods ------116 4.4.6 Further studies and applications ------117 CONCLUSIONS------118 CHAPTER 5 ------129 GENERAL DISCUSSION AND CONCLUDING REMARKS ------129

5.1 TSS AND PH INDICATE A SYSTEM’S BIOSECURITY ------129 5.2 ECO-CULTURE SYSTEM: THE IMPROVEMENT OF SEMI HETEROTROPHIC BIOFLOC IMPLEMENTATION IN SHRIMP AQUACULTURE---- 132 5.3 THRESHOLDS: A HOLISTIC APPROACH TO PREVENTING VIBRIO DISEASE ------138 5.4 CONCLUDING REMARKS AND OUTLOOK ------141 REFERENCES ------144 APPENDICES ------179

APPENDIX 1 ------180 APPENDIX 2 ------189 APPENDIX 3 ------202

Page 10 of 208

ABSTRACT

Aquaculture is the fastest growing industry in the animal production sector as one of the most important sustainable protein sources of humans, and shrimp production contributes a significant portion of aquaculture, exceeding that of the capture fishery. However, shrimp disease outbreaks can occur at the early post-stocking stage due to the presence of bacteria. As a result, farmers are fearful of stocking shrimp in many ponds because of the effects of early mortality syndrome (EMS) and/or acute hepatopancreatic necrosis disease (AHPND) caused by opportunistic bacteria, such as Vibrio. EMS can cause up to 100% shrimp mortality and result in a 70% drop in total shrimp production. This bacteria-caused disease has led to the conclusion that sustainable shrimp aquaculture will depend on the development of a bio-secure production system, which needs to be improved to ensure growth and disease resistance to facilitate bio-shrimp productivity. Among the methods that could be applied for shrimp aquaculture systems, I found that biofloc technology (i.e. using suspended aggregates growth) can help in the prevention of opportunistic bacteria invasion based on r/K selection theory (i.e. r-bacteria can multiply rapidly; K-bacteria are slowly multiplying). Therefore, this thesis was designed to comprise three manuscripts, presented in chapters 2, 3 and 4, with the aim of using biofloc technology as an alternative approach to controlling disease in the context of shrimp farming, especially in the early stage after stocking.

This work not only focussed on improving the existing methods of disease prevention, but research was also carried out on the nature of shrimp-pathogen interactions to identify promising applications for shrimp farming, which involve the use of bio-inoculants for application of a potential eco-culture system. To successfully use the microbial composition in bacterial activity management, four different approaches in intensive shrimp aquaculture were applied for comparing the bacterial composition regarding the biological and operational parameters. In chapter 2, I presented the conventional method of examining microbial populations using the culture-dependent quantification of bacteria and Sanger sequencing for looking at some bacterial bio-indicator candidates from the isolates of rearing water. Then, I further classified the bacteria into r- and K-bacteria according to r/K selection theory. The presumptive opportunistic bacteria (e.g. Vibrio spp.) and probiotic bacteria (e.g. Bacillus spp.) were also evaluated along with heterotrophic bacteria by testing with physicochemical water characteristics and shrimp performance parameters (i.e. weight, weight gain rate and feed conversion rate). Remarkably, the Page 11 of 208

pH was displaying negative correlations with Bacillus spp., K-bacteria and volatile suspended solids (VSS). The most important factor shaping the Bacillus population was total suspended solids (TSS). A stable microbiome including beneficial bacteria, such as Bacillus spp., and a low ratio of r/K-bacteria, may protect shrimp from opportunistic pathogens like Vibrio spp. The chapter suggested that the application of the heterotrophic biofloc system is better than using the other three systems in terms of controlling harmful bacteria and lowering the environmental impact. The first gap in biofloc technology for Vibrio control was addressed with the - recommendations for close monitoring of the NO2 -N concentrations, r/K-bacteria ratio and pH level during the first 3 weeks of post-stocking. Therefore, managing the water microbiome according to the r/K selection and pH balance may serve as a suitable approach for shrimp farmers to maintain productive shrimp culture and contribute to more sustainable shrimp aquaculture.

In chapter 3, I presented the independent method for examining the bacterial composition using Illumina Miseq V3 sequencing of 16S rRNA genes. The ecology of the bacterial communities was studied in relation to major environmental factors. Bacterial operational taxonomic units - (OTUs) richness was only influenced by TSS, pH, VSS, temperature and NO2 -N. Moreover, the bacterial community structure was also influenced by these factors, along with dissolved oxygen - (DO), NO3 -N, salinity and alkalinity. The study clearly showed that environmental parameters, and mainly TSS, can exert a strong structuring influence on bacterial communities in shrimp rearing water. Moreover, the shrimp weight gain rates were correlated with some bacterial OTUs (i.e. Ruegeria sq 1 and sq90, Pseudoalteromonas sq91, Catenovulum sq142, Holomonas sq263, Roseovarius sq175, Roseobacter sq81, Pseudooceanicola sq78). These relationships suggest the need for further studies on the effects of single bacterial species on elements that will benefit aquaculture, such as greater shrimp weight and health, and their relationship with the amount and sources of organic matter. Although the heterotrophic biofloc system evidenced the maximum OTUs from the beginning to end of the experiment, the highest OTU richness occurred in the autotrophic biofloc systems. The OTU numbers in the particle-attached samples were much larger than in the free-living samples. The bacterial OTU composition changed with the time; however, the distribution was highly influenced by the type of culture system (i.e. four different operating systems). As another important result, the most dominant single-sequence OTU abundance does not always belong to the relative class abundances of the community. Therefore,

Page 12 of 208

single bacterial species with a high proportion had a greater ecological function in the system than the class relative abundances did in this thesis.

In chapter 4, I presented the findings on bio-inoculants as an open gateway for controlling Vibrio disease based on the results of immersion challenge tests of cultivated shrimp with a potential acute hepatopancreatic necrosis disease bacterium, Vibrio parahaemolyticus

(pVpAHPND). As consistent with isolates of chapter 2 and OTU bio-indicators of chapter 3, the bio-inoculants in chapter 4 were employed as evidence for explaining the differences in shrimp survival probability by different bacterial community patterns in shrimp tissue. Especially the sequence-dominant target taxa (e.g. Pseudoalteromonas sq91, Ruegeria sq1), could have a high potential as presumptive bacterial candidates for further study at the species level to contribute to sustainable shrimp aquaculture protected from AHPND. Several rapid diagnostic methods to apply at the farm level to help prevent AHPND were noted at the end of this chapter.

Overall, the thesis provided baseline information on bacterial ecology, intending to establish healthy super-intensive shrimp aquaculture with good recommendations for practice. As a future outlook, a shrimp eco-culture system is proposed with high potential in terms of contributing to responsible aquaculture, which is needed to develop a quick beneficial bacteria kit, such as for Ruegeria sq1 and Pseudoalteromonas sq91 bacteria as r-matured bacteria, using biofloc technology as an alternative approach to controlling shrimp disease outbreaks via bacterial activity management.

Page 13 of 208

Zusammenfassung

Aquakultur ist die am schnellsten wachsende Industrie im Tierproduktionssektor und eine der am nachhaltigsten Proteinquellen für Menschen. Garnelenzucht trägt einen signifikanten Teil zur Aquakultur bei und übertrifft sogar die Produktion durch Fischfang. Aber während des frühen Garnelenbesatzes können Krankheiten durch Bakterienbefall ausbrechen. Wegen den Auswirkungen der sogenannten „akuten nekrotisierenden Hepatopankreatitis“ (NHP; auf Englisch “early mortality syndrome” (EMS) or “acute hepatopancreatic necrosis disease” (AHPND), welche durch opportunistische Bakterien, wie z.B. Vibrio, hervorgerufen wird, zögern viele Züchter ihre Teiche mit Garnelen zu füllen. NHP kann zu einem 100%igen Garnelensterben und zu einem 70%igen Rückgang der Garnelenproduktion führen. Das mögliche Auftreten dieser Bakterien verursachten Krankheit hat zu der Schlussfolgerung geführt, dass nachhaltige Garnelenzucht von der Entwicklung eines biologisch-sicheren Produktionssystems abhängig ist. Dieses Produktionssystem muss verbessert werden, um Wachstum und Krankheitsresistenz der Garnelen zu gewährleisten und die Produktivität der Garnelenzucht zu fördern. Unter den Methoden, welche für die Garnelenzucht in Aquakultursystemen verwendet werden können, fand ich heraus, dass eine Aggregat-basierte (“biofloc“) Methode vor der Invasion von opportunistischen Bakterien helfen kann. Diese Methode basiert auf der r/K Fortpflanzungstheorie (r-Bakterien können sich schnell multiplizieren; K-Bakterien vermehren sich langsam). Diese Doktorarbeit ist aus drei Manuskripten erstellt, welche in den Kapiteln zwei, drei und vier vorgestellt werden mit dem Ziel, die Aggregat basierte („biofloc“) Methode als einen alternativen Ansatz zu verwenden, um das Garnelensterben (vor allem in den frühen Stadien des Garnelenbesatzes in der Aquakultur) zu kontrollieren.

Der Fokus dieser Arbeit lag nicht nur auf der Verbesserung existierender Methoden zur Krankheitsprävention, sondern auch auf der Untersuchung der Interaktion von Garnelen und ihrer Pathogenen, um vielversprechende Applikationen für die Garnelenzucht zu identifizieren, welche die Benutzung von Bio-Inokulanten von potenziellen Öko-Aquakultur Systemen beinhalten. Um die mikrobielle Zusammensetzung erfolgreich für das Management der bakteriellen Aktivität zu nutzen, wurden vier verschiedene Ansätze in intensiven Garnelen- Aquakulturen angewendet, um die bakterielle Zusammensetzung mit Hinblick auf deren biologische und betriebliche Parameter zu vergleichen. In Kapitel 2 habe ich die konventionelle Methode zur Charakterisierung der Bakteriengemeinschaft verwendet: die kulturbasierte Quantifizierung und das Sanger Sequenzieren von einigen Bakterien als Indikatoren aus den Isolaten des Garnelenwassers. Danach klassifizierte ich die Bakterien in r und K Bakterien, laut der r/K Fortpflanzungstheorie. Die vermutlichen opportunistischen Bakterien (z.B. Vibrio spp.) und probiotischen Bakterien (z.B. Bacillus spp.) wurden zusammen mit heterotrophen Bakterien auch evaluiert, indem sie mit physikalischen und chemischen Wasserparametern sowie der Leistungsfähigkeit der Garnelen (Gewicht, Gewichtszunahme-Rate und Futterverwertung) getestet wurden. Bemerkenswert ist, dass der pH-Wert negative Korrelationen mit Bacillus spp., K-Bakterien und flüchtigen Schwebstoffen (Volatile Suspended Solids VSS) zeigte. Der Page 14 of 208

wichtigste Faktor, der die Bacillus-Population prägte, waren die gesamten Schwebstoffe (Total Soluable Substances, TSS). Ein stabiles Mikrobiom mit nützlichen Bakterien wie Bacillus spp. und ein niedriges r/K-Bakterien Verhältnis kann Garnelen vor opportunistischen Krankheitserregern wie Vibrio spp. schützen. In diesem Kapitel wird gezeigt, dass die Anwendung des heterotrophen Bioflockensystems schädliche Bakterien besser kontrolliert und damit die Umweltbelastung im Vergleich zur Verwendung der drei anderen Systeme verringert. Die fehlende Vibrio-Kontrolle in der Biofloc-Technologie wurde mit den Empfehlungen für eine - genaue Überwachung der NO2 -N-Konzentrationen, des r/K-Bakterienverhältnisses und des pH- Wertes in den ersten 3 Wochen nach der Lagerung gewährleistet. Daher kann die Verwaltung des Wassermikrobioms gemäß der r/K-Auswahl und des pH-Wertes als geeigneter Ansatz für Garnelenzüchter dienen, um eine produktive Garnelenkultur aufrechtzuerhalten und zu einer nachhaltigeren Garnelen-Aquakultur beizutragen.

In Kapitel 3 stellte ich die unabhängige Methode zur Untersuchung der bakteriellen Zusammensetzung mittels Illumina Miseq V3 Sequenzierung von 16S rRNA-Genen vor. Die Ökologie der Bakteriengemeinschaften wurde in Bezug auf die wichtigsten Umweltfaktoren untersucht. Die Anzahl der bakteriellen operativen taxonomischen Einheiten (OTUs) wurde nur - durch TSS, pH, VSS, Temperatur und NO2 -N beeinflusst. Darüber hinaus wurde auch die bakterielle Community-Struktur von diesen Faktoren beeinflusst, zusammen mit dem gelöstem - Sauerstoff (DO), NO3 -N, Salzgehalt und Alkalität. Die Studie zeigte deutlich, dass Umweltparameter, vor allem TSS, einen starken strukturierenden Einfluss auf Bakteriengemeinschaften in Garnelenaufzuchtwasser ausüben können. Darüber hinaus wurden die Gewichtszunahmeraten der Garnelen mit einigen bakteriellen OTUs korreliert (d.h. Ruegeria sq 1 und sq90, Pseudoalteromonas sq91, Catenovulum sq142, Holomonas sq263, Roseovarius sq175, Roseobacter sq81, Pseudooceanicola sq78). Diese Beziehungen deuten darauf hin, dass weitere Studien über die Auswirkungen einzelner Bakterienarten auf Elemente, die der Aquakultur zugute kommen, wie z.B. ein höheres Garnelengewicht und eine bessere Gesundheit, sowie deren Zusammenhang mit der Menge und den Quellen der organischen Substanz erforderlich sind. Obwohl das heterotrophe Bioflockensystem die höchste durchschnittliche Anzahl der OTUs vom Beginn bis zum Ende des Experiments zeigte, trat der höchste OTU- Reichtum in den autotrophen Bioflocksystemen auf. Die OTU-Zahlen in den partikelgebundenen Proben waren dabei viel größer als in den freilebenden Proben. Die bakterielle OTU- Zusammensetzung änderte sich mit der Zeit; die Verteilung wurde jedoch stark von der Art des Kultursystems beeinflusst (d.h. vier verschiedene Kultursysteme). Als weiteres wichtiges Ergebnis wird gezeigt dass die dominante OTU Sequenz nicht immer zur häufigsten Klasse der Gemeinschaft gehört. Daher hatten einzelne häufige Bakterienarten eine größere ökologische Funktion im System als die Bakterienklasse mit der größten relativen Häufigkeit.

In Kapitel 4 stellte ich die Ergebnisse zu Bio-Impfstoffen zur Bekämpfung der Vibrio- Infektionen in Garnelenkulturen vor. Die Ergebnisse basieren auf immersion challenge tests mit Vibrio parahaemolyticus, dem potenziellen Erreger der akuten hepatopankreatischen

Nekroseerkrankung, (pVpAHPND). Im Einklang mit den im Kapitel 2 beschriebenen Isolaten und

Page 15 of 208

den OTU-Bioindikatoren aus Kapitel 3, wurden die Bio-Impfstoffe in Kapitel 4 verwendet, um die Unterschiede in der Überlebenswahrscheinlichkeit von Garnelen durch verschiedene bakterielle Gemeinschaftsmuster im Garnelengewebe nachzuweisen. Besonders die sequenzdominanten Zieltaxa (z.B. Pseudoalteromonas sq91, Ruegeria sq1) könnten ein hohes Potenzial als bakterielle mutmaßliche Kandidaten für weitere Studien auf Artenebene haben, um zu einer nachhaltigen, vor AHPND geschützten Garnelen-Aquakultur beizutragen. Mehrere Schnelldiagnosemethoden für AHPND, werden am Ende des Kapitels zur Anwendung in shrimp Farmen empfohlen.

Insgesamt lieferte die Arbeit grundlegende Informationen über die bakterielle Ökologie, die darauf abzielt, eine gesunde, super-intensive Garnelen-Aquakultur mit Hilfe guter Praxis- Empfehlungen aufzubauen. Als Zukunftsperspektive wird ein Ökokultursystem für Garnelen mit hohem Potenzial für einen Beitrag zur verantwortungsvollen Aquakultur vorgeschlagen, dass für die Entwicklung eines schnellen und nützlichen Bakterienkits erforderlich ist, wie beispielsweise für Ruegeria sq1 und Pseudoalteromonas sq91 Bakterien als r-angereifte Bakterien, wobei die Biofloc-Technologie als alternativer Ansatz zur Bekämpfung von Krankheitenausbrüchen durch bakterielles Aktivitätsmanagement eingesetzt wird.

Page 16 of 208

Chapter 1

Page 17 of 208

Chapter 1

Introduction

The concept of biofloc technology (BFT) goes against the conventional view that water in a pond has to be clear as an indicator of high quality. In addition to bacteria and algae, bioflocs contain a high-density aggregate of organic matter, protozoa, rotifers, nematodes and dead particulate organic matter (Avnimelech and Kochba, 2009; Hargreaves, 2006). Hence, BFT is not yet thoroughly predictable, and thus, it can be risky to implement at the farm level on water quality and disease outbreaks. However, the goal of this work approach is using biofloc as an alternative approach for controlling opportunistic pathogens in shrimp culture systems. Specifically, it aims to control opportunistic pathogens via microbial composition management.

Figure 1: Matured microbial biofloc compositions under microscope at magnification 200x from the water column of heterotrophic biofloc shrimp culture system at Laboratory of Aquaculture & Artemia Reference Centre (ARC), Belgium. The flocs had copepods which are an indicator of a mature ecosystem (De Schryver et al., 2014).

As shown in Figure 1, microbial flocs are formed mainly from heterotrophic bacteria, and other microorganisms, such as microalgae, zooplankton and protozoa attach to a floating surface, resulting in suspended growth (Najdegerami et al., 2016; Martínez-Córdova et al., 2017). The main characteristics of BFT regarding aquaculture systems are high stocking densities, heavy

Page 18 of 208

mixing, intense aeration, low or no water exchange and high organic matter input (Browdy et al., 2001).

In general, flocs are variable in shape and have a broad distribution of particle sizes; moreover, they are plenty of microorganisms, easily compressible and highly spongy, and they can flow easily (Chu and Lee, 2004). While individual bacterial cells are sized in the order of 1 μm in diameter (Madigan et al., 2006), bioflocs are sized in the range of 1.0 - 6.0 mm, depending on different conditions (Mitchell and Kogure, 2006).

1.1 Making a difference through responsible aquaculture

Responsible aquaculture helps to make a difference to the environment through sustainable production; due to the rapid growth of the aquaculture industry, this theme has been expressed in presentations and discussions at many conferences and workshops around the world. Aquaculture products are proliferating as a solution to meet the increasing demand for marine products in the face of overfishing of the world’s waters (FAO, 2018). The market for aquaculture products is high, while the following effects often characterise the current aquaculture practices:

(1) Nutrient enrichment: The dissolved products include ammonia, phosphorus and dissolved organic carbon. These waste products have a variety of sources; they may be excreted directly as faecal particles from a reared animal, suspended from the feed or released from particles. Moreover, lipids released from the diet may form a film on the water surface, causing a lack of dissolved oxygen;

(2) Organic matter enrichment: Sediments are diverse environments supporting a range of microbiota existing in a complex matrix with parameters including particle size and organic carbon availability; and

(3) Disease outbreaks: A significant issue affecting aquaculture production is the loss of stock through diseases. Vibrio spp. are commonly causing disease with substantial mortality (WHO, 1999).

Aquaculture systems are characterised by extensive, semi-intensive, intensive and super intensive systems, and they apply methods like flow-through, recirculation systems, mature

Page 19 of 208

water and bioflocs. Most of today’s world aquaculture production is reared in extensive and semi-intensive systems. Nowadays, the interest in high-technology-intensive aquaculture systems is increasing with the increasing demand for aquaculture products (FAO, 2018).

For an intensive system, BFT could enhance the water quality, natural food availability, growth and production (Rahman et al., 2008). Moreover, for protecting the environment and natural resources, the expansion of aquaculture must take place sustainably. BFT makes this possible by minimising the exchange and water usage in aquaculture systems with adequate water quality in the culture unit (Crab, 2010a). In BFT, bacterial biomass as a source of proteins is generated from the organic nitrogenous waste (e.g. ammonium) stemming from rearing animals with a well-balanced carbon/nitrogen ratio (Schneider et al., 2005). Carbohydrates are recommended as extra additions to the culture systems for promoting heterotrophic bacterial growth, and nitrogen uptake takes place through the production of microbial proteins (Avnimelech, 1999).

The high concentration of total suspended solids (TSS) present in the rearing water of the intensive system is the aspect that most needs to be well considered in terms of its effects on animal performance (Avnimelech and Kochba, 2009). Moreover, consumers now call for guarantees that their food has been produced, dealt with and commercialised to support their health and the environment while addressing various other ethical and social considerations (FAO, 2012). BFT is highly recommended as a technique for meeting such requirements and lowering fish production prices.

In brief, BFT is a production system where dense microbial communities are managed to control ammonia, which is mainly released by the cultivated animals. Microalgae or heterotrophic bacteria can absorb ammonia or be transformed into biomass by nitrifying bacteria (Ebeling et al., 2006). These organisms usually grow in the form of microbial flocs (bioflocs) that can be used as feed for the cultivated animal (De Schryver et al., 2008).

Awareness of responsible aquaculture as part of the general concept of responsibility is increasing among stakeholders and farmers through high-cost investments in the intensive farms by farmers (FAO, 1999). The precautionary approaches in aquaculture have been incorporated into regulations and licenses at the farm level by the councils (FAO, 2018). Among multiple methods, BFT has several advantages for supporting sustainable development goals (Bossier and

Page 20 of 208

Ekasari, 2017); thus, it can help in moving towards responsible aquaculture for the shrimp industry.

1.2 Microbiota and disease in aquaculture: Issues related to the Vibrio genus

hepatopancreas

intestine was empty

Figure 2: The juvenile shrimp (Litopenaeus vanamei) was challenged with a potential acute hepatopancreatic necrosis disease bacterium (Vibrio parahaemolyticus). Shrimp hepatopancreas was destroyed with the dose at 107 CFU mL-1 in the challenge room at Laboratory of Aquaculture & Artemia Reference Centre (ARC), Belgium.

Vibrio bacteria are highly motile, with high metabolic flexibility, and they quickly increase their population size when the environmental conditions are favourable (Stocker, 2012). Vibrio spp. are some of the most important pathogens for reared aquatic animals, such as penaeid shrimps (Lightner, 1993; Stocker, 2012; Liu et al., 2018a), several fish species and molluscs (Austin, 1988) and corals (Ben-Haim et al., 1999). The genus Vibrio comprises approximately 14 members (Sawabe et al., 2013); some luminescent Vibrio spp. include Vibrio cholerae, V. fischeri, V. harveyi, V. logei, V. splendidus, V. mediterranei and V. orientalis (Yang et al., 1983;

Page 21 of 208

Farmer and Hickman-Brenner, 1992). Some pathogenic Vibrio for shrimp at the larval stage include V. harveyi and V. campbellii (Karunasagar et al., 1994; Hameed et al., 1996), whereas V. penaeicida and V. parahaemolyticus had greater effects on juvenile and adult shrimp (Roque et al., 1998). Temperatures in the range of 25-37°C were considered optimum for V. parahaemolyticus growth (Ahmed et al., 2018); the same temperature is also optimal for shrimp growth (Mohanty et al., 2018).

Other bacteria are not a significant problem in shrimp disease. However, Photobacterium leiognathi and P. phosphoreum have generally been involved with disease outbreaks in shrimp larviculture systems (Lavilla-Pitogo et al., 1990), and they have a lesser effect on disease outbreaks in grow-out ponds (Lavilla-Pitogo et al., 1998).

1.3 Stable microflora systems

Microorganisms are essential drivers of all ecosystems; microbial ‘resistance’ and ‘resilience’ are important terms for understanding the microbial community’s stability (Shade et al., 2012). According to Pimm, (1984) and Allison and Martiny, (2008), resistance is the degree to which a community lacks sensitivity when faced with disturbance. In contrast, resilience is the rate of community recovery after being disturbed (Allison and Martiny, 2008). Moreover, disturbances modify a community directly or via environment change (Glasby and Underwood, 1996). The members of a community may die or change in their relative abundance after the disturbance (Rykiel, 1985). Community stability comprises resistance and resilience or the community’s response to disturbance (Rykiel, 1985; Pimm, 1984; Shade et al., 2012), as illustrated in Figure 3.

Page 22 of 208

© Shade et al., (2012)

Figure 3: The authors described the stability of microbial community after disturbance. The three levels at individual, population, and community contribute to community resistance and resilience that making the stability of microbial communities. The individual survival after the disturbances contributes overall to the community stability (Shade et al., 2012).

Moreover, in measuring microbial stability, the successful establishment in the long term of a non-native organism in a community, called invasion (Litchman, 2010), can be used as an indicator of stability for the compositional and functional community (Shade et al., 2012).

In the water column, the rearing animals and their excretion, and feed are all available together.

Therefore, all beneficial, neutral and pathogenic bacteria can generate in a matrix (i.e. the water column). For a long time, research focussed on the effects of pathogenic bacteria on disease outbreaks. However, few pathogens have been identified in the larval stage of cultivated animals (Skjermo and Vadstein, 1999; Sandlund and Bergh, 2008). The intensive aquaculture production Page 23 of 208

method increases the bacterial load by the presence of a massive amount of organic matter and nutrient enrichment; thus, there is high opportunity for the selection of opportunistic microbes, which are r-strategists according to the r/K concept (Andrews and Harris, 1986; Vadstein et al., 2004). The microbial maturation concept, known as the r/K concept or r/K selection theory, was proposed by Vadstein et al. (1993) in terms of the conception of preventing adverse effects by competition among microbes rather than by removing them from the system. Currently, K- strategists are microorganisms with a reasonably slow growth rate that can grow in conditions with a low amount of nutrient per bacterium, with populations filled to carrying capacity; opportunistic pathogens are normally r-strategists that can grow rapidly within hours with plenty of nutrients per bacterium. As a microbial control strategy in aquaculture, it is helpful to have K- strategists in competition with r-strategists (Vadstein et al., 2018b).

The stable microflora in a system comprise mature microbiota, such as algae-rich green waters and microbially matured water systems, including K-strategists and r-matured-strategists (Andrews and Harris, 1986; Vadstein et al., 1993; Defoirdt, 2016). However, the bacterial richness and evenness are lower in the rearing systems after disinfection to remove opportunistic pathogens, causing an unstable microflora system; thus, disinfection causes more harm than good (De Schryver et al., 2014). Hence, there is a need for establishing sustainable microbial management strategies based on ecological principals in aquaculture. For this reason, the ecological theory (i.e. r/K selection) in microbial ecology will help develop a strategy in aquaculture to prevent pathogenic infections regarding microbial stability (De Schryver et al., 2014; Vadstein et al., 2018a).

1.4 Microbial flocs represent a stable microflora system

According to Boyd and Clay, (2002), the water quality of a microbial-based production system containing bacterial flocs is more stable than a phytoplankton-based production system is. In addition, according to Otoshi et al., (2009), bacteria-dominated systems promote more stable water quality than systems dominated by mixed algae do, without the negative effects of algal bloom and algal collapse phages (e.g. algal-dominated systems), resulting in high pH fluctuation and dissolved oxygen.

The total bacterial number in the biofloc is approximately 10 times higher than that in clear water (Kim et al., 2014). These heterotrophic bacteria not only serve as a direct food source for

Page 24 of 208

shrimp but also shorten the food chain in the rearing water, recirculating nutrients and promoting the proliferation of bacteria that can inhibit the reproduction of pathogenic bacteria (Martínez- Córdova et al., 2016).

Primarily, Zhao et al., (2012) evaluated whether adding Bacillus could reduce the number of Vibrio. The probiotic Bacillus attaches to harmful bacteria by producing peptides that are inhibit their growth. They also use molecular language, for example, N-acylhomoserinelactone (AHL) in communication between Vibrio (Li and Nair, 2012). Therefore, Bacillus can help to reduce opportunistic pathogens population such as Vibrio by that two ways, making the biofloc system a stable system (Verschuere et al., 2000). There are fewer antibiotic, antifungal, probiotic and prebiotic applications, although they are still helpful; however, BFT can be a combined strategy in disease prevention (Emerenciano et al., 2013).

1.5 Shrimp culture with biofloc technology

Litopenaeus vannamei (Boone, 1931) is farmed in over 70% of all the world’s shrimp aquaculture operations, and it commands a reasonable price in the world seafood market. In 2016, the farmed crustacean production reached 7.9 million tonnes, and L. vannamei contributed 53% of the global crustaceans aquaculture production (FAO, 2016; FAO, 2018). Penaeid shrimps (L. vannamei) are highly valued seafood commodities in the domestic and international markets. The economics of shrimp farming mostly depends on the feed, which constitutes 40- 60% of the operational expenses (Tan et al., 2005). However, Hari et al. (2006) reported that, in shrimp biofloc systems with extra carbon addition, from 25 to 40% of the required protein level in the feed can be reduced due to the following factors: 1) the increased nitrogen retention in harvested shrimp biomass via bacterial protein production, 2) reduced demand for feed protein without compromising shrimp production, 3) decreasing concentrations of potentially toxic + - nitrogens (NH3-N and NH4 -N) and nitrite concentration (NO2 -N) in the system and 4) reduced water-based nitrogen discharge to the environment (Burford et al., 2004; Wasielesky et al., 2006; Xu and Pan., 2012). Nevertheless, shrimp can eat the bacteria present on the flocs (i.e. particles/aggregates; Xu and Pan, 2013). Moreover, the floc is a source of ‘in situ proteins’ and fatty acids (Crab et al., 2010b).

Several studies have shown that shrimp can be produced successfully in systems that use limited or no water exchange (McIntosh, 2000), such as the biofloc system. Furthermore, BFT can help

Page 25 of 208

minimise the introduction of pathogens with zero water exchange (Crab, 2010a), and the heterotrophic microbial biomass is suspected of having a controlling effect on pathogenic bacteria (Michaud et al., 2009).

In another approach, biofloc was evidenced to have a positive effect on the survival, growth, and digestive enzyme activities of Pacific vannamei, such as L. vannamei (Xu and Pan, 2013; Zheng et al., 2019). The Vibrio population was found in lower numbers in a biofloc system compared with another, non-biofloc system (Souza et al., 2012). In a study about disease resistance, BFT may enhance immune activity (Ekasari et al., 2014), with positive effects on both growth and immune gene expression (Kim et al., 2014). In Germany, green aquaculture farming involves applying a production volume of 15 metric tonnes per year for L. vannamei in Grevesmühlen as a land-based recirculation system with BFT application (Edition, 2014) as a green approach in the direction of sustainable shrimp aquaculture.

1.6 Knowledge gaps and justification of this thesis

Disease is a limiting factor for the aquaculture industry (FAO, 2012). For the shrimp culture industry, outbreaks of bacterial disease (FAO, 2012) have been the primary cause of production loss (Lee et al., 2015; Rurangwa and Verdegem, 2015). The marine shrimp industry is also challenged by shrimp diseases caused by (opportunistic or r-strategists) pathogens, such as Vibrio spp. (Defoirdt et al., 2007). From 2009 to 2015, there were global outbreaks of a serious bacterial shrimp disease called acute hepatopancreatic necrosis disease (AHPND) or early mortality syndrome (EMS; Joshi et al., 2014; Lee et al., 2015; Peña et al., 2015, Hong et al., 2016). This disease causes severe loss of shrimp after stocking in the period of the larval stage until day 30 (approximately), and the mortality can be 100% (Tran, 2013). Tran et al., (2014) illustrated that controlling shrimp disease outbreak using biofloc systems is very difficult and labour-intensive in practice. Moreover, Xu and Pan, (2014) recommended a need for studying biofloc composition and activity under different operational practices and proposing the probability that operating biofloc compositions and activity will achieve the desired nutritional contribution and functionality for cultured shrimp (Bossier and Ekasari, 2017).

Page 26 of 208

Figure 4: The advantages and the gaps of biofloc technology in shrimp aquaculture.

In general, biofloc represents a co-culture of heterotrophic and autotrophic bacteria (Ebeling et al., 2006); therefore, there is a dynamics of the bacterial proportions in biofloc systems. This indicates that using biofloc systems is extremely difficult at the farming level (Tran et al., 2014; Prangnell et al., 2016). As can be seen in detail in Figure 4, in practice, the gaps of control in biofloc’s bacterial community composition for the optimal water quality, shrimp health and productivity remain, especially for shrimp at the post-larval stage. It is necessary to measure the complex microbial biofloc community, which will support the establishment of the desired, stable microbial communities for the aims of optimal shrimp aquaculture (Smith et al., 2003). Ekasari et al., (2014) and Aguilera-Rivera et al., (2014) examined the total heterotrophic bacteria and Vibrio spp. as an evaluation. Especially, Zhao et al., (2012) found that adding Bacillus could reduce the number of Vibrio in the shrimp farming systems. Although the importance of microbial communities in pond water quality has been recognised (Bentzon-Tilia et al., 2016), however, the relationship between biofloc bacteria and cultured shrimp performance (host) has

Page 27 of 208

not yet been thoroughly evaluated. Therefore, it is challenging to determine the mortality of rearing animals due to the interplay of pathogenic and non-pathogenic strains (Lemire et al., 2015), as well as the interaction of r strategists and K strategists (Andrews and Harris, 1986), which is causing a new type of polymicrobial disease. Therefore, the mortality is not yet entirely understood with the complex interactions among microbes, larvae, free and attached bacteria, microbial grazers and planktonic organisms (Ray et al., 2010; Vadstein et al., 2018a), although it is known that the interactions among factors like nutrition, the environment and physical conditions all play a role (Kimes et al., 2012).

It is recommended to conduct more research on the different microbial species in biofloc systems, as aquaculture’s prospects rely on suitable microbes for feeding cultured animals (Martínez-Porchas, 2017; Emerenciano et al., 2013). Worldwide, the current evidence supports the view that the use of microorganisms as a direct food source in aquaculture may bridge the gaps on the road to sustainability (Martínez-Córdova et al., 2016; Nevejan et al., 2018). Bringing the gaps of BFT into this work, with the objective of successfully using microbial compositions in biofloc systems, the focal topics were established for experimental designs; these are depicted in Figure 5.

Figure 5: The matrix of this work to minimizing the gaps of biofloc technology.

Page 28 of 208

As the figure shows, there is a need to strengthen studies on bacterial ecology in aquaculture systems, the relationship between microbial communities, shrimp performances (growth and survivorship) and water quality (Martínez Cruz et al., 2012; Vadstein et al., 2018a).

For well-functioning biofloc systems, water quality must be controlled, and this will depend on the intensity of nitrogen and the organic load of the culture water. The start-up period of BFT may be shortened by adding water from existing well-performing biofloc ponds; alternatively, using specific bacterial inoculants may allow an accelerated start-up. However, to support shrimp health and productivity, more research on the effects of biofloc is required in pursuit of biofloc inoculants with specific bacterial candidates (Hashim et al., 2018).

In BFT systems, reduced water exchange, high organic matter input and high growth rates of heterotrophic bacteria contribute to an increase in the TSS (Van Wyk, 2006). Ebeling et al., (2006) reported that the density of suspended solids is the crucial limiting factor affecting increased productivity. Moreover, Van Wyk, (2006) reported that the rate of solids removal can also determine which component of the biota (which groups of bacteria and/or microalgae) will be more dynamic in monitoring the nitrogen compounds. Therefore, some solids removal is essential for protecting the system’s stability and the health of the cultivated organisms. Many researchers have highlighted the need to set appropriate functional limits for solids in BFT systems. The primary use of settling chambers to reduce suspended solids can remove the nitrifying bacteria from the system due to their slower growth rate (i.e. beneficial bacteria in the control of ammonia concentration; Ebeling et al. 2006); nevertheless, the consequences of high solid concentrations include poor water quality, changes in the composition of the organisms that make up the flocs and adverse effects on health and performance of the cultured animals (Crab et al., 2007; Ebeling et al., 2006; Hargreaves, 2006). Therefore, it is necessary to determine the levels of bioflocs (i.e. TSS) that appear to be suitable for high shrimp performance in different culture systems.

The dietary protein level and carbon/nitrogen ratio (C/N ratio) used can have important consequences for water quality, biofloc composition and shrimp performance in intensive, zero- exchange biofloc-based shrimp aquaculture culture systems (Xu and Pan, 2012). Therefore, the feeding traits (or feed utilisation), shrimp size, biofloc density and biofloc size are factors that should be considered when making efforts to reduce the protein level in feed without causing negative effects to shrimp performance while operating BFT as recommended by Bossier and Page 29 of 208

Ekasari, (2017) to optimise the system. Nevertheless, both dietary protein levels and C/N ratios also affect the activities of extracellular proteases and amylases. These enzymes produced in bioflocs could help break down proteins, complex carbohydrates and other nutritional element availability in the water column into smaller components; in this manner, BFT can help to facilitate feed digestibility and absorption with the specific enzymes by the shrimp host (Xu and Pan, 2012).

Crab et al., (2010b) observed that bioflocs formed from glycerol inoculated with Bacillus have higher protein (58% crude protein) than bioflocs designed solely from glycerol or acetate (42 - 43% crude protein) and glucose (28% crude protein). This suggests that microbial components, unknown growth factors or probiotic microorganisms like Bacillus present in the bioflocs may have led to a considerably higher growth rate and better feed conversion rate (FCR) in shrimp nourished in biofloc systems (Burford et al., 2004; Wasielesky et al., 2006); however, no composition of specific nutrients have been given. Yet, this also suggests that the nutritional composition of the biofloc differs with the type of carbohydrate source, microbial community structure and culture conditions. Thus, it is not easy to convince farmers to apply this technology (i.e. BFT) at the farming level. Many further studies are still needed.

+ In the water column, un-ionised ammonia (NH3-N) and ammonium (NH4 -N) are in equilibrium, depending on the pH and temperature (Timmons et al., 2002). According to Körner et al., (2001), + both NH3-N and NH4 -N concentrations may be harmful to the shrimp host, and un-ionised + ammonia is more toxic than NH4 -N ions are. Figure 6 shows an example of the effect of high

NH3-N concentrations on shrimp performance. The toxicity threshold depends strongly on the species of cultivated animals, animal size, solids situation, intractable organics and other non- bioactive compounds (Colt, 2006).

Page 30 of 208

all pleopods, pereiopods and antennal spine were extruded

Figure 6: The juvenile shrimp (Litopenaeus vanamei) died in the heterotrophic biofloc system -1 when exposed to ammonia concentration at 3.74 mg L (NH3-N) with pH 7.44, and the temperature is 28.4 oC. The lessons learned in studying on generating bioflocs at Laboratory of Aquaculture & Artemia Reference Centre (ARC), Belgium.

The biofilm bacteria showed the ability whereby biofilm cells are more resilient when exposed to disinfectants and antibacterial agents than similar bacteria are when they are in a free-swimming cell state (Kadam et al., 2013). However, more than 80% of bacteria will cause animal diseases to have biofilm ability (Hall et al., 2014). Hence, currently, probiotics cannot help control EMS in shrimp culture or with an ineffective dose. Methods of controlling EMS outbreaks are under research; therefore, the availability of a standard challenge model to establish a fast, reliable and systemic evaluation of potential management for EMS is the final objective of this work.

Bioflocs is a challenging concept to explain for because it regards a complex system with constant physical, chemical and biological interactions. The knowledge gaps concerning BFT Page 31 of 208

have not yet been addressed; therefore, the present work, titled ‘Microbial composition in biofloc-based shrimp aquaculture systems’, was carried out with the final goal of using bioflocs as alternative approaches to controlling opportunistic bacterial pathogens in shrimp aquaculture systems. The work was based on microbial composition management according to the ecological theory perspective, which employs microbial activity management as a new study approach.

1.7 Manuscript overview

The manuscript was divided into three chapters (chapters 2, 3 and 4). In each chapter, the contribution of co-authors, technical staff and scientific staff of the Laboratory of Aquaculture & Artemia Reference Centre and Leibniz Centre for Tropical Marine Research was fully acknowledged for any assistance received. Since the manuscripts are still in preparation, acknowledgements and contributions from prospective co-authors will be updated upon submission. Table 1 is a summary of my contributions to the first draft of each manuscript.

Table 1: The table shows my contribution to the first draft of each manuscript in percentage (100 % for each task) following each chapter.

Task Chapter 2 Chapter 3 Chapter 4

Experimental concept and design 80 85 85

Acquisition of experimental data 100 100 100

Data analysis and interpretation 55 60 90

Preparation of figures and tables 55 65 95

Draft the manuscript 80 100 100

Figure 7 gives an overview of possible steering parameters in studying microbial composition in biofloc-based shrimp aquaculture systems.

Page 32 of 208

Figure 7: An overview of possible roles of biofloc-based shrimp aquaculture system in the enhancement of shrimp aquaculture performance and the coverage of these aspects in the chapter’s studies with manuscripts preparation. C/N ratio is carbon/nitrogen ratio. TSS is total suspended solids.

In the previous part of this work, the investigations on bioflocs for shrimp aquaculture had remaining gaps related to the aims of each chapter in the thesis. However, the overall objective of this thesis is assessing why disease outbreaks occurred regularly in the shrimp post larval stage (i.e. until post-stocking day 30), causing a 70% drop in production due to EMS/AHPND. This disease makes farmers unenthusiastic about stocking shrimp in large numbers until the EMS can be controlled or successful, failsafe harvests can be offered (Thitamadee et al., 2016; Flegel, 2019). Although BFT shows several advantages in shrimp farming and opportunistic control at the laboratory level, it is not yet successful at the farming level because of remaining gaps in the operational systems (Tran et al., 2014; Prangnell et al., 2016; Bossier and Ekasari, 2017). Therefore, the issue of shrimp disease outbreaks at the post larval stage regarding the application

Page 33 of 208

of BFT has been partly addressed in each chapter, with specific aims for study. The following topics are considered: (1) In chapter 2, I aim to find the linkages among the environmental and water-quality conditions; microbial populations, such as presumptive Vibrio spp., Bacillus spp. and heterotrophic bacteria with the composition of r-bacteria and K-bacteria; and shrimp growth performance. This is done by comparing four different shrimp culturing approaches (FT, RAS, H-Floc and A-Floc). These linkages can help generate a better understanding of the operational parameters in heterotrophic biofloc systems, which exhibit labour-intensive operations at the farm level; (2) In chapter 3, I aim at describing the bacterial ecology of the four shrimp culture systems to provide useful baseline information for future studies that seek to understand the factors affecting bacterial diversity and community structure. Therefore, the excellent shrimp performance (i.e. healthy and disease free) and productivity are explained by the relationships with some proposed bacterial OTUs enhancing shrimp weight gain rates; and (3) In chapter 4, I aim to identify ways of preventing shrimp disease at the farming level by proposing several fast diagnostic methods to help avoid EMS/AHPND. Based on the differences in survival hours of shrimp challenged with a potential acute

hepatopancreatic necrosis disease bacterium, Vibrio parahaemolyticus (pVpAHPND), the interactions between shrimp (the cultivated host) and water (including chemical and bio compositions) can be understood. With the intention of elucidating the

virulence factors of pVpAHPND, the shrimp mortality is determined and compared to find a link with the bacterial community composition in water and shrimp tissue to generate a better understanding of EMS outbreaks.

Hypotheses of the thesis study

The goal of this work approach is using bioflocs as an alternative approach to control opportunistic pathogens in shrimp systems. Therefore, I hypothesise that microbial community management via r/K selection theory can have benefits for shrimp production and animal health. In chapter 2, I hypothesise that the heterotrophic biofloc system is better than three other systems for shrimp production and growth of beneficial bacteria. In chapter 3, I hypothesise the

Page 34 of 208

following: 1) there is a strong difference in bacterial composition for each system, 2) TSS as an indicator of organic matter will mostly drive such compositional differences in shrimp aquaculture systems and 3) the excellent shrimp performance and productivity can be explained by the relationships with some bacterial OTUs enhancing shrimp weight gain rates and survivorship. In chapter 4, I hypothesise the following: 1) there is interaction between the shrimp factor and water factor in protecting shrimp during bacterial immersion challenge testing and 2) the virulence factor of pVpAHPND and predictability of shrimp mortality can be determined.

I focus on addressing specific questions for each chapter to answer the overall thesis research question as the goal of this work approach, specifically, ‘Is it possible to use BFT as an alternative approach to microbial activity management?’. These questions are as follows:

(1) Chapter 2: What are the differences between BFT systems and other systems in shrimp aquaculture? What are differences between the heterotrophic biofloc system and autotrophic biofloc system? Which system gives the best shrimp performance and health? What are the differences in the r/K-bacterial populations among the shrimp aquaculture systems? Which factors explain the differences among the shrimp culture systems? Which factors result in microbial control? (2) Chapter 3: Is there a difference in the richness and bacterial composition of the four culture systems at a fraction of the water column sample? Which environmental factors influence the diversity and bacterial composition? Is there a difference in the richness, diversity and bacterial composition of the two biofloc culture systems at two different fractions of free-living and particle-attached samples? Which taxa with high proportional OTUs and significant differences among the four systems and in time could contribute as bio-indicators for intensive shrimp culture systems, especially for the biofloc shrimp-based aquaculture system? Which key taxa are likely associated with shrimp weights in the four systems? Does greater diversity correlate with higher shrimp weight gain? (3) Chapter 4: Which factor (e.g. cultivated shrimp, rearing water) is best for protecting shrimp from pathogenic Vibrio? Which shrimp have longer survival hours after the infection? How do the microbial community compositions in rearing water and reared

Page 35 of 208

shrimp contribute to the shrimp survival probability? What is the most critical factor playing a role in protecting shrimp?

Overall, the new findings of chapters 2, 3 and 4 are discussed in parallel in chapter 5 with the concluding remarks.

Page 36 of 208

Chapter 2

Page 37 of 208

Chapter 2

Opportunistic pathogens in intensive shrimp culture systems: control and management

Loan Thanh Doan1,3*, Christiane Hassenrück1, Andreas Kunzmann1, Patrick Sorgeloos2, Astrid Gärdes1 and Peter Bossier2.

1 Leibniz Centre for Tropical Marine Research (ZMT), Fahrenheitstr.6, 28359 Bremen, Germany 2Laboratory of Aquaculture & Artemia Reference Centre (ARC), Coupure Links 653, B-9000 Gent, Belgium 3Vietnam National University of Agriculture, Trau Quy, Gia Lam, Hanoi, Vietnam

* Correspondence: Corresponding Author [email protected]; [email protected]

Keywords: aquaculture, Litopenaeus vannamei, post-larvae, microbial ecology, bacterial populations, r/K-strategists, biofloc

Running title: Pathogens in intensive shrimp farming

ABSTRACT

Intensive shrimp aquaculture methods tend to increase the carrying capacity for microbes, favouring the growth of opportunistic pathogens that can multiply rapidly (r-strategists). To reduce the risk of diseases, management of shrimp cultures aims to establish a stable microbiome, particularly slowly multiplying K-strategists and probiotic bacteria. We investigated four different approaches for intensive culturing of Litopenaeus vannamei post- larvae by setting up flow-through (FT), recirculation (RAS), heterotrophic biofloc (H-Floc) and autotrophic biofloc (A-Floc) systems. Over 30 days, we monitored water quality, shrimp growth and microbial populations using culture-dependent quantification of bacteria. Bacteria were further classified into r- and K-strategists and several isolates identified by sequencing the 16S ribosomal genes. We detected a clear distinction between the RAS system and the other three

Page 38 of 208

- - culturing systems in all measured parameters (e.g. pH, NH3-N, NO2 -N, NO3 -N, r- and K- strategists, and shrimp weight). Shrimp growth was highest in the H-Floc system, which also exhibited the lowest pH (7.38 ± 0.13). pH was displaying negative correlations with Bacillus spp. (rrm = - 0.37, p = 0.019), K-strategists (rrm = - 0.37, p = 0.020) and volatile suspended solids

(rrm = - 0.46, p = 0.001). The most important factor shaping Bacillus spp. was total suspended solids-TSS (rrm = 0.46, p = 0.002). A stable microbiome including beneficial bacteria, such as Bacillus spp. (≥ 104 CFU mL-1) and also a low ratio of r/K-strategists may protect shrimp from opportunistic pathogens such as Vibrio spp. (< 104 CFU mL-1). Therefore, the application of the H-Floc system in shrimp aquaculture may offer a practical and effective solution to control harmful bacteria and lower the environmental impact. However, we recommend close - monitoring during the first three weeks of culturing, as toxic NO2 -N concentrations may develop and opportunistic pathogens, which may proliferate before the establishment of balance and beneficial conditions for long-term culturing of shrimps. Managing the water microbiome according to the r/K selection and pH balance may serve as a suitable approach for shrimp farmers to maintain productive shrimp culture and contribute to more sustainable shrimp aquaculture.

2.1 INTRODUCTION

According to the Food and Agriculture Organization of the United Nations (FAO, 2017), aquaculture plays an essential role in food security for the growing human population and is estimated to contribute significantly to seafood production worldwide. The culture of penaeid shrimp is economically significant and increases the livelihood of coastal human populations, which thereby gain an additional income from the aquaculture sector (Wang et al., 2008a; FAO, 2012). Especially, Litopenaeus vannamei is the dominant species for shrimp farming and contributes 53% of the global crustacean aquaculture production (FAO, 2016; FAO, 2018). Treating diseases, their transmission and control, is an overarching goal of the aquaculture industry. Disease-causing bacteria are habitually saprophytes or heterotrophs, and are generally detected in the rearing water and sediments of pond systems. They also take an active part in the decomposition of organic matter derived from faeces and uneaten food. Generally, opportunistic pathogens degrade organic matter or colonise their host in a commensal relationship, but can

Page 39 of 208

cause diseases when the host’s resistance is low, for example, due to environmental stress. Particularly diseases caused by Vibrio spp. are likely to occur when the animals are stressed due to over-stocking, low oxygen concentrations, high ammonia concentrations in the water or poor nutrition for shrimp growth (Moriarty, 1997). These opportunistic pathogenic bacteria cause tremendous losses in the shrimp aquaculture, for example, through the acute hepatopancreatic necrosis disease (AHPND) or early mortality syndrome (EMS) (Lee et al., 2015). EMS, caused by Vibrio parahaemolyticus, has emerged to initiate significant acute mortality in L. vannamei post-larvae within approximately 30 days of culture, and it was shown to be a frequent cause of up to 100% production losses (Tran, 2013). EMS is a fast disease and was considered a global outbreak, first reported in southern China in 2009, and afterwards expanding to Southeast Asia, reaching Mexico in 2013 and Philippines in 2015 (Joshi et al., 2014; Peña et al., 2015; Hong et al., 2016). Therefore, attempts on L. vanamei post-larvae cultivation have focused on the prevention of several bacterial diseases.

The ecological r/K selection theory was introduced by Andrews and Harris (1986) and later initially applied in aquaculture by De Schryver and Vadstein (2014) as a strategy to control the invasion of potential opportunistic pathogenic bacteria, especially in larviculture. Although, disinfection and probiotics with Bacillus are commonly used as microbial management methods (Vadstein et al., 2018b). According to r/K selection, the opportunistic pathogens causing bacterial diseases in marine aquaculture worldwide are mostly r-strategists blooming in unstable environments, while K-strategists are known as slow multiplying microorganisms preferring a stable environment (Andrews and Harris., 1986; Vadstein et al., 2013; De Schryver and Vadstein., 2014). The nutrient level per microorganism in the rearing water determines the classification as an unstable environment or a stable environment. In the intensive shrimp aquaculture, nutrients are mostly organic matter derived from faeces and uneaten food. It is challenging to know the mortality of cultured shrimp as reported by Lemire et al., (2015) that the mortality of cultured animals was increased due to the interplay of pathogenic and non- pathogenic strains, or the interaction of r-strategists and K-strategists causing a new type of polymicrobial disease. Therefore, the circumstances that increase mortality are not yet entirely understood, though the interactions among several factors such as nutrition, environment and physical conditions are known to play a role (Kimes et al., 2012).

Page 40 of 208

Conventionality and commonly used shrimp rearing systems are flow-through systems (FT) and recirculating aquaculture systems (RAS). In the FT system, the frequent water exchange removes microalgae or bacteria and also averts the accumulation of harmful chemical compounds. The RAS system promotes denitrification, where oxidized inorganic nitrogen compounds, such as - - nitrite-N (NO2 -N) and nitrate-N (NO3 -N), are reduced to basic nitrogen (N2) and subsequently eliminated from the system to the air. This process is accelerated by facultative anaerobic microorganisms, which play a role in the decomposition of either organic matter called heterotrophic denitrification or decreasing inorganic matter called autotrophic denitrification (van Rijn, 1996; van Rijn et al., 2006; Chen et al., 2016). The recently introduced biofloc technology (BFT) (Becerra et al., 2011; Emerenciano et al., 2013; Avnimelech, 2016; Emerenciano et al., 2017) can be divided into two systems: autotrophic biofloc (A-Floc) and heterotrophic biofloc (H-Floc). This technology optimizes the interaction of microalgae, bacteria, and grazers of bacteria in bioflocs (suspended aggregates) to maintain water quality and promote shrimp growth (Hargreaves, 2006). Essentially, there are two bacterial populations responsible for water quality maintenance in the biofloc systems: heterotrophic ammonia-assimilative and chemoautotrophic nitrifying bacteria (Ebeling et al., 2006; Capone et al., 2008). Hence, the A- Floc system promotes nitrobacters and microalgae as the primary producers, while the H-Floc system benefits heterotrophic bacteria as consumers with the additional carbon source. Alteration in the C/N ratio results in a shift from a heterotrophic to an autotrophic system, or vice versa (Avnimelech, 2015). To date, BFT is considered as a novel strategy for disease management and environmentally friendly aquaculture. By minimizing water discharge and reusing water, BFT can be considered a ‘green’ approach with low environmental impact (Emerenciano et al., 2013). The minimum water exchange maintains heat and prevents temperature fluctuations, permitting the growth of tropical species in cold areas (Crab et al., 2009; Emerenciano et al., 2013). Conversely, controlling EMS/AHPND using biofloc systems is very difficult and labour- intensive in practice (Tran, 2014; Prangnell et al., 2016).

Although the linkages among environmental conditions and bacterial disease in the water column are becoming apparent (Kimes et al., 2012), and the effects of environmental factors on microbial communities in aquaculture systems have been extensively studied, little is known about how shrimp cultures are influenced (Hou et al., 2017). For instance, the differences in bacterial populations among FT, RAS, H-Floc and A-Floc are still unknown. Further, there are

Page 41 of 208

complex interactions among microbe, larvae, free and attached bacteria, microbial grazers, and planktonic organisms (Ray et al., 2010; Vadstein et al., 2018a).

This study aimed to find the linkages among the environmental and water quality conditions, microbial populations such as presumptive Vibrio spp., Bacillus spp., heterotrophic bacteria (HB) with the composition of r-strategists and K-strategists, and shrimp growth performance by comparing four different shrimp culturing approaches (FT, RAS, H-Floc and A-Floc). We hypothesized that the H-Floc system is better for shrimp production and growth of beneficial bacteria.

2.2 MATERIAL AND METHODS

2.2.1 Experimental design and facilities

A setup of 15 experimental plastic tanks (40 L with 35 L of active volume) was monitored over 28 days with four replicates for the FT, RAS, H-Floc systems and three replicates for the A-Floc system (Figure 1). Marine water from the North Sea was stored and aerated in 100 L plastic containers (source water). A water mix composed of source water, raceway water, protein skimmer water, marine microalgae and a microbial inoculant (LT3), which is a quorum- quenching probiotic Bacillus strain, was prepared. The water in the tanks of the FT system consisted of 33 L of source water. Thirty percent of the water volume was exchanged every three days. Water loss by evaporation and sampling was compensated by source water. According to van Rijn et al. (2006) and Ebeling et al. (2006), the RAS system included a 5 L sedimentation chamber and a 10 L nitrifier tank. Bacterial nitrifiers were established on bio-plastic substrates one month in advance by submerging the substrate in protein skimmer water. Two weeks before stocking, the RAS tanks were filled with 43 L of source water and 5 L of the water mix to - -1 achieve stable nitrite concentrations (i.e. NO2 -N ≤ 0.05 mg L ). For the BFT systems, the tanks were filled with 28 L of the source water and 5 L of the water mix, which was prepared two months in advance to create matured floc water. In the A-Floc systems, nitrifiers (e.g. Nitrosomonas, Nitrobacter) were allowed to settle on the provided plastic substrates as described above for the RAS system. In the H-Floc systems, heterotrophic bacteria (e.g. Bacillus, Paracoccus, Pseudomonas, Azoarcus) were promoted by adding glucose (Glucose anhydrous - VWR) as an additional carbon source at a C/N ratio of 15/1 according to Ebeling et al. (2006) Page 42 of 208

and Avnimelech (2015). Every three days, the glucose solution was pumped continuously using a WATSON Marlow 205S peristaltic pump at 1.25 rpm for 24 hours. Immediately prior to shrimp stocking, 2 L of aerated water mix was added to each experimental tank.

One thousand shrimp post-larvae (PL24) were adapted to the experimental conditions for two days with a gradual increase of temperature from 20°C to approximately 30°C, and salinity from 25 ppt to approximately 30 ppt by seawater addition. Then, 65 post-larvae (mean weight of 0.21 ± 0.10 g; n = 60) were randomly transferred to each respective experimental tank, which resulted in a density of 1857.14 larvae per m3 corresponding to 552.72 shrimp per m2. The tanks were covered with mesh to avoid shrimp loss. The light was provided at 649.30 Lux in a 12:12 hours dark:light cycle. Constant aeration was supplied through plastic tubes and air stones. The feeding rations started at 10% of the shrimp body weight per day, and gradually reduced to 7.5% per day at week 1; 6% per day at week 2; 5% per day at week 3 to the end of the experiment. A formulated feed (Crevetec Grower 40% crude protein, 10% crude fat, 7.5% crude ash, 4% crude fiber, 0.84% phosphorus, 8g kg-1 calcium, 7g kg-1 natrium-sodium) was used and supplied twice a day. Water was not exchanged and wastes or solids were not removed for the RAS, H-Floc or A-Floc systems.

2.2.2 Water quality parameters and analyses

Turbidity (water transparency using a Secchi disk), salinity (HI 96822 seawater refractometer), pH (pH 1100L-VWR), temperature and dissolved oxygen (DO; ARC Field-Lab OXI) were measured in situ daily at 10:00 a.m. during the first two weeks and then weekly until the end of the experiment. The size and colour of the setting solids (SS) were recorded weekly following the “Standard methods for the examination of water and wastewater” (APHA, 2005) and adjusted after Avnimelech (2015). Briefly, flocs were sedimented with Imhoff cones and measured in mL L-1 natural size SS (no sieving), sieved through the mesh size of 250 μm and 50 μm.

For measurements, water column samples were filtered through a microporous filtering film of pore size 0.2 μm (Whatman-FP30/0.2CA-S). The filtrate was then measured with an HI 8203 + - - Aquaculture Photometer for NH4 -N (ammonia-N), NO2 -N (nitrite-N), NO3 -N (nitrate-N).

NH3-N (un-ionized ammonia) was determined according to UNESCO (1983), then total + ammonia nitrogen (TAN; a sum of NH3-N and NH4 -N) and total inorganic nitrogen (TIN; a

Page 43 of 208

+ - - sum of NH3-N, NH4 -N, NO2 -N, and NO3 -N) were calculated respectively. Alkalinity (mg - -1 HCO3 L ) was measured weekly using a JBL test kit (JBL Company, Neuhofen, Germany). To quantify total suspended solids (TSS), water samples were filtered using a dried and weighted glass fiber filter with pore size 0.7μm (GF/F, Whatman, Germany). Ammonium formate -1 (CH5NO2; 0.5 mol L ) was added to the filter to eliminate salt, and filters were dried at 103°C to 105°C for 4 hours. TSS was determined as the difference of the filter weights. Then, the filters used for TSS quantification were burned at 550oC for 30 minutes to calculate the weight loss by burning of the total solids to determine the weight of volatile suspended solids (VSS; APHA, 2005).

2.2.3 Microbiological analysis

Sampling procedures

The water column samples (50 mL), whole shrimp larvae samples, and free-living and particle- attached bacteria samples from the two biofloc systems were collected from the experimental tanks every week from the beginning (week 0) to the end of the experiment (week 4). Imhoff cones were used to separate the particles (suspended aggregates or flocs) in 1L of floc water by sedimentation for 15-30 minutes. The sedimented particles constituted the particle-attached bacteria samples, while the water in the upper part of the Imhoff cones was used for the free- living bacteria samples (APHA, 2005). The particle and shrimp samples were homogenized with a Seward Stomacher 80 Lab Blender to detach the bacteria associated with the particles or shrimp tissues. The shrimp samples were washed prior to homogenization to remove the surface bacteria and to select for bacteria inside of the shrimp larvae. For the sterilized washing process, the shrimp larvae were caught on a sterile mesh and successively anaesthetized by immersion in a benzocaine solution for 10 seconds, disinfection in benzalkonium chloride solution for 10 seconds, and repeated rinsing (3x) in sterile, autoclaved source water for 5 seconds. The shrimp larvae were aseptically transferred to a sterile plastic bag containing 10 mL of sterile source water to determine shrimp weight before homogenization. All samples were divided into sub- samples for the enumeration of total (1) HB, which were further classified into r- and K- strategists, (2) Bacillus spp., and (3) presumptive Vibrio spp.. Bacillus and Vibrio were all classified only as r-strategists.

Page 44 of 208

Agar plate counts

Bacteria were quantified as colony-forming units (CFU) mL-1 (for water column and free-living samples) or CFU g-1 (for particle-attached and shrimp samples). Serial sample dilutions (dilution factors 10-0 to 10-6, 10-fold serial dilutions) were prepared in advance for all sub-samples to determine the optimal dilution for counting. Sterile beads (0.1 mm SI- BG01; 0.5 mm SI-BG05; 1.0 BioSpec-11079110) were added to selected dilutions of the shrimp, and particle-attached samples and bacteria were more thoroughly detached by mixing using a Vortex-genie2. Only agar plates with less than 300 colonies were considered for counting to ensure that individual colonies could be identified accurately.

Thiosulphate-citrate-bile salts-sucrose (TCBS) agar was used for quantifying presumptive Vibrio spp. and separated into sucrose-positive (suc+) and sucrose-negative (suc-) Vibrio (Kaysner and Depaola, 2004). Fully visible CFUs of suc-Vibrio spp. were counted after one day of incubation, whereas suc+CFUs were counted after two days of incubation because some colonies were growing more slowly. Prior to plating on marine agar, samples for the quantification of Bacillus spp. were pasteurized in a 70oC water bath (SUB Aqua 5 Plus) for 30 minutes. CFUs of r- and K- strategists were determined according to Leij et al. (1994). Colonies visible after 24h of incubation were considered r-strategists, while colonies visible only after two to ten days of incubation represented the more slowly growing K-strategists. Plating on marine and TCBS agar-plates was performed using an easySpiral Plater (Accelerated Bacterial Colony Enumeration) in duplicate for each sub-sample by applying the forward and backward plating methods of the spiral-plating device.

Identification of isolates

Bacterial isolates from the water mix and tank water with pink flocs, which may represent candidates to develop bio-indicators for sustainable floc systems, were grown on marine and TCBS agar. Seventeen isolates were selected for taxonomic identification using Sanger sequencing. Genomic DNA was extracted following the Cetyltrimethyl Ammonium Bromide (CTAB) protocol according to Doyle and Doyle (1987). The bacterial 16S ribosomal gene was amplified in a standard polymerase chain reaction (PCR) using the primer set developed by Nadkarni et al. (2002). The amplified 466-bp PCR products were Sanger sequenced by Star-SEQ (Mainz, Germany). Sequences were screened by Geneious 11.1.2 for chimeras before submission to NCBI. The generated sequences for each isolate are available in the NCBI SRA repository Page 45 of 208

under the accession numbers MH100960, MH100961, MH100962, MH100963, MH100964, MH100965, MH100966, MH100967, MH100968, MH100969, MH100970, MH100971, MH100972, MH100973, MH100974, MH100975, and MH100976.

2.2.4 Shrimp growth performance analysis

Once a week, 10 post-larvae shrimp were sampled randomly from each tank and weighed on a digital balance (OHAUS PA3202) to calculate the amount of feed applied per day for each week of the experiment and the amount of glucose needed for the H-Floc systems according to Ebeling et al. (2006) and Avnimelech (2015). The average initial weight (g) was determined once immediately before stocking. The weight gain rate (WGR), protein efficiency ratio (PER) and feed conversion rate (FCR) were calculated with the following equations:

ሺˆ‹ƒŽ™‡‹‰Š– െ ‹‹–‹ƒŽ™‡‹‰Š–ሻ  ሺΨ’‡”™‡‡ሻ ൌ šͳͲͲΨ ‹‹–‹ƒŽ™‡‹‰Š–

„‘†›™‡‹‰Š–‰ƒ‹ ሺΨሻ ൌ šͳͲͲΨ ’”‘–‡‹‹–ƒ‡

ˆ‡‡†‹–ƒ‡  ൌ „‘†›™‡‹‰Š–‰ƒ‹

2.2.5 Statistical analysis

We used general linear mixed models (GLMMs) with subsequent posthoc pairwise comparisons using culturing system and sampling week as fixed and experimental tank as the random factors to account for the repeated measurements of each tank over the experiment duration. Parameters were organized in four blocks (water parameters: salinity, pH, temperature, DO, alkalinity, + - - NH3-N, NH4 -N, NO2 -N, NO3 -N, TSS, VSS, TAN and TIN; bacterial parameters: suc-Vibrio, suc+Vibrio, Bacillus spp., r-strategists and K-strategists (heterotrophic bacteria); shrimp parameters: weight, weight gain rate and feed conversion rate; biofloc parameters: transparency and SS), but analyzed individually. For each block, the significance threshold of 0.05 was adjusted using Benjamini-Hochberg correction based on the number of parameters in the block. Prior to testing, the Shapiro-Wilk test and diagnostic plots confirmed normality and homoscedasticity of residuals. If necessary, the parameters were log-transformed to meet the Page 46 of 208

assumptions of the statistical test. Values were presented as the mean and standard deviation, if applicable. Principal component analysis (PCA) was used to summarize water quality parameters in a low-dimensional space to identify patterns between experimental tanks. Repeated measurements correlations of shrimp performance and water parameters, as well as bacterial parameters from the water column samples were determined by as implemented in the R package rmcorr (Bakdash and Marusich, 2017). Only significant correlations were visualized in a correlation network.

All statistical analyses were performed in R (Team, 2016) (R v.3.0.2, R Core Team, 2016, using R Studio v.0.98.1056) with the packages vegan (Oksanen et al., 2016), lme4 (Bates et al., 2015), gplots (Warnes et al., 2016), nlme (Pinheiro et al., 2018), lsmeans (Lenth, 2016), igraph (Csardi, 2006), TeachingDemos (Snow, 2016), and rmcorr (Bakdash and Marusich, 2017).

2.3 RESULTS

2.3.1 Biofloc performance parameters

The two-biofloc systems, H-Floc and A-Floc, were evaluated for floc performance (Table 1). The H-Floc and A-Floc systems showed a clear difference in colour: brown for the H-Floc system and green-brown for the A-Floc system. The natural size’s SS (without sieving) had a maximum of 10.50 ± 1.66 mL L-1 and a minimum of 2.00 ± 0.00 mL L-1. The H-Floc system seemed to have more condensed flocs than the A-Floc system because the transparency was significantly lower, whereas the natural size’s SS was considerably higher (Table 1). The 50 μm pink-floc size’s SS appeared only once, at week 3, in the H-Floc system. In the A-Floc system, the 50 μm size’s SS appeared twice at week 1 and week 2 without pink colour.

2.3.2 Changes in water quality parameters

+ DO and NH4 -N concentrations showed no differences among the four intensive shrimp culture systems. However, Figure 2 and Figure 3 evidenced clear differences in the remaining water parameters between the RAS system and the three other systems (FT, H-Floc and A-Floc). NH3- N concentrations were significantly higher in the RAS system (Supplementary Table S1.2) during the whole experiment, with a peak at week 2 (0.48 ± 0.09 mg L-1). Concentrations of the - -1 less toxic NO2 -N dropped significantly from 24.17 ± 8.36 mg L at week two to 2.42 ± 1.94 mg Page 47 of 208

L-1 towards the end of the experiment in the H-Floc system. The RAS and A-Floc systems had - significantly lower NO2 -N concentrations throughout the experiment. Only in the FT system, - NO2 -N concentrations fluctuated because the water chemistry was regularly disturbed by water exchange, showing remarkably higher concentrations (18.33 ± 12.50 mg L-1) by the end of the experiment with very high variability among replicate tanks. Throughout the experiment, except for week 1, pH values in the BFT and FT systems were significantly lower than in the RAS system (Figure 3 and Supplementary Table S1.2). Among all systems and weekly sampling events, the lowest pH was 7.38 ± 0.13 in the H-Floc system at week 1, while the highest pH was 8.33 ± 0.15 in the RAS system at the beginning of the experiment. In the H-Floc system, the maximum TSS and VSS concentrations were 426.50 ± 65.24 mg L-1 and 331.50 ± 53.59 mg L-1, respectively. The A-Floc, FT and RAS systems showed lower TSS and VSS concentrations (Figure 3). The four systems evidenced significant differences in salinity, starting with rather high salinity ranging from 33 to 37 ppt that decreased to a balance value of approximately 30 ppt at week 2 until the end of the experiment in all four systems. In summary, there were clear differences in water quality parameters among the four different approaches (FT, RAS, H-Floc and A-Floc) to water management.

2.3.3 Changes in microbial components

CFUs of presumptive suc- and suc+Vibrio spp., Bacillus spp., and HB (r- and K-strategists) in the shrimp larvae, the water column, as well as free-living and particle-attached in the BFT systems were quantified with (Figure 4, Supplementary Table S4).

-1 At week 0, the number of suc-Vibrio in the shrimp larvae was 282.50 CFU g , and the number of suc+Vibrio inside the shrimp larvae was 376.50 CFU g-1. Suc+Vibrio in shrimp larvae of week 2 or at the end of the experiment were not observed. The suc-Vibrio in shrimp samples at week 2 had a maximum of 15.47 r 8.38 x 106 CFU g-1 in the H-Floc system, where they decreased toward the end of the experiment (8.61 x 106 CFU g-1) (Supplementary Table S4).

For the water column samples, as well as the free-living and particle fractions, the counts of presumptive Vibrio spp., Bacillus spp., heterotrophic bacteria (r- and K-strategists) are detailed in Figure 4. There was no significant difference in suc+Vibrio numbers in all four systems. However, it should be noted that there were significant differences in suc+Vibrio during the experimental period regardless of the culturing systems (Supplementary Table S2.1). Numbers of

Page 48 of 208

suc-Vibrio were significantly higher in the FT system than in the three remaining systems (A- Floc, H-Floc and RAS). The lowest suc-Vibrio numbers occurred in the RAS system. The remaining bacterial populations were significantly higher in the H-Floc system. In summary, the highest proportions of Bacillus spp., r-strategists, K-strategists (both HB) in the water column samples were detected in the H-Floc system.

No differences in presumptive Vibrio and heterotrophic bacteria in the free-living and particle samples between the two-biofloc systems were evidenced. However, there were higher numbers of HB at week 3 in the two-biofloc systems in both free-living and particle samples (Figure 4; Supplementary Table S2.1 and Table S2.2). Free-living samples showed the highest number of Bacillus spp. in the A-Floc system, although it was not statistically significant difference. Particle samples showed the highest number of Bacillus spp. in the H-Floc system, though this difference was only detected at week 3. In examining only free-living samples, no difference in r-strategists between the two biofloc systems was observed; nevertheless, there were significantly higher numbers of K-strategists in the H-Floc system, again only at week 3.

Sanger sequencing was applied to identify the 17 isolates (Table 2). The isolates consisted of four Vibrio spp., eleven Bacillus spp. and two Pseudoalteromonas spp. according to the most closely related strains reported by NCBI BLAST. The genus Pseudoalteromonas (accession - MH100972) was isolated from the pink floc in the H-Floc system at week 3.

2.3.4 Shrimp growth parameters

In intensive shrimp aquaculture systems, maximizing shrimp growth parameters is the most critical objective. Shrimp weight, WGR, FCR and PER are indicated in Figure 5. Although shrimp weight increased in all four aquaculture system, significant differences in shrimp weight, WGR, and FCR occurred between the RAS system and the other systems at week 3, with the exception of the shrimp weight in the FT system. The RAS system displayed a significantly higher FCR (4.41 ± 3.04) (Supplementary Table S3.2) and a significantly lower WGR (12.94 ± 4.98% per week). The final shrimp weight was significantly higher in the H-Floc system (1.80 ± 0.33 g per shrimp) than in the RAS system (1.14 ± 0.19 g per shrimp) (Supplementary Table S3.2).

Page 49 of 208

2.3.5 The relation between parameters of environmental conditions, microbial components, and shrimp growth performance

Figure 6 and Supplementary Table S5 shows the significantly positive and negative correlations among parameters belonging to environmental conditions, microbial populations (water column samples) and shrimp performance. Only salinity showed a positive correlation with pH (rrm =

0.57, p < 0.001). Many negative correlations with pH were identified, such as Bacillus spp. (rrm =

- 0.37, p = 0.019), K-strategists (rrm = - 0.37, p = 0.020), TSS (rrm = - 0.53, p = 0.001) and VSS

(rrm = - 0.46, p = 0.001). K-strategists were positively correlated with shrimp weight (rrm = 0.46, p = 0.009). r-strategists were strongly negatively correlated with shrimp weight (rrm = - 0.53, p =

0.002). NH3-N concentration and shrimp weight evidenced a negative relation (rrm = - 0.55, p = + 0.001). NH4 -N concentration had a strong negative correlation with shrimp weight (rrm = - 0.69, - p < 0.001). NO3 -N concentration and shrimp weight gain rate (WGR) were negatively correlated. Temperature and shrimp weight were strongly positively correlated (rrm = 0.68, p <

0.001). While suc+Vibrio and WGR were negatively correlated (rrm = - 0.43, p = 0.019), suc+Vibrio was positively correlated with shrimp weight. And last but not least, TSS was positively correlated with Bacillus spp. (rrm = 0.46, p = 0.002).

2.4 DISCUSSION

2.4.1 Ecological interactions: r- vs. K- strategists

Presumptive Vibrio spp. (suc-Vibrio, suc+Vibrio) and Bacillus spp. were (all classified only) r- strategists and r- and K-strategists (both HB) followed different dynamics in the four shrimp culture systems as examples of the complexity of ecological interactions. Our data showed an overview of relationships between parameters of microbes, environment, and shrimp performance. Many studies on reducing the risks of disease in shrimp aquaculture systems have been suggested that there are complex interactions among those parameters in establishing a stable microbiota condition. Microbial maturation, stable microbiota condition, selects for beneficial bacteria (i.e. K-selected strategists) that protect the marine larvae from bacterial diseases (Skjermo et al., 1997). Moreover, De Schryver and Vadstein (2014) reported that the ecological r/K selection theory could be the key in the microbial management to prevent diseases caused by the proliferation of opportunistic pathogens (r-selected strategists). Recently, Vadstein Page 50 of 208

et al. (2018a) supported that the ecological theory is needed to test the interactions among microbe-microbe and microbe-larval host.

As primary finding, we could assess that the interactions between microbe-microbe were most likely to compete with each other. As a consequence of, first, there were no significant correlations among them inferring neither positive nor negative trends on the proportions (suc- Vibrio, suc+Vibrio, Bacillus spp., and HB). Second, the proportions of suc-Vibrio, Bacillus spp., and HB were significantly different among the four systems in the water column. While the microbe-larval host relations were either evidenced through positive (suc+Vibrio - shrimp weight) or negative (suc+Vibrio - shrimp weight gain RATE) correlations. Since the shrimp weight was likely continuously increasing during the experimental periods. Therefore, the WGR is a much more relevant rather than shrimp weight in a relationship with microbes. It is likely that microbial proportions integrated the rate of shrimp weight increase. Finally, according to the overview, the differences in r- and K-strategists proportions were evidenced by the differences in the operations of the four systems with clear differences in water quality parameters and shrimp weight. As a result of the interactions, microbial components differently occupied the four systems and influenced the TSS.

2.4.2 The role of TSS and pH on microbial control

The TSS linearly correlated with pH value. We could infer that a higher TSS induced a lower pH. The changes in TSS influenced the changes in pH in all four systems. Although pH is not a primary factor directly altering the microbial component patterns, we could assess that a high pH decreased Bacillus spp. (r-strategists), K-strategists (mature microbial communities) and also VSS (microbiota biomass). Hence, pH is an integrating variable that provides an integrated index in the shrimp rearing water. More details on pH trend comparisons among the four systems, our study suggested the advantages of the H-Floc systems, in which a significantly lower pH value showed the significant high proportions of beneficial bacteria. Moreover, the decrease in pH values in the H-Floc systems were also reported by several previous studies (Tacon et al., 2002; Wasielesky et al., 2006; Panjaitan, 2010; Huang et al., 2017). However, in our study, there is a - possible drawback in the H-Floc system is the lows in pH regarding the high NO2 -N concentrations. Unlike the H-Floc system, the RAS system showed the highest pH with the highest NH3-N concentration, and an unexpected higher FCR, while WGR was significantly

Page 51 of 208

- lower, even though the system had clear water with a lower NO2 -N concentration, and the lowest suc-Vibrio number.

According to the pH trends (low or high) with benefits and drawbacks could be affected the shrimp healthy and weight gains among all four systems. We recommend a stabilized pH (balance) may indicate the stability of the microbial components (e.g. Vibrio spp., Bacillus spp., and ratio of r/K-strategists) as also suggested by several previous studies, which evidenced that changes in pH refer in variations of the microbial populations (Mikkelsen et al., 1996; Ligi et al., 2014; Liu et al., 2015; Hou et al., 2017). For a ‘green’ approach, we suggest that regularly monitoring pH could help to predict the bacterial populations (e.g. the important proportions of Bacillus spp. and K-strategists) in the shrimp culture systems. We could see that a stable microbial condition including beneficial bacteria, such as Bacillus spp. (≥ 104 CFU mL-1) and a low ratio of r/K-strategists may protect shrimp from opportunistic pathogens such as Vibrio spp. (< 104 CFU mL-1). In combination with minimizing stress for shrimp, a stabilized pH is a great way to get started concerning stable microbial proportions or components to keep shrimp healthy. This factor of pH balance is a significant factor in the microbial activity management without adding any chemical solutions (for disinfection) into the rearing water towards a good aquaculture practice.

2.4.3 TSS differences: advantages of heterotrophic bacteria

The inputs to form TSS of our study were mostly coming from feed and VSS. According to the experimental design for feeding regime that feed was equally supplied into the four systems and by weekly estimated shrimp biomass. On the other hand, the wastes or solids were not removed to the end of the experiment, therefore, the significant differences of TSS among the four systems probably because of (1) the hydraulic retention time in the FT system, (2) the function of biofilter chambers in the RAS system, and (3) the differences in type of flocs which were added carbon into the H-Floc systems to promote HB and added autotrophs into the A-Floc system by inoculated bio-plastic substrates. For that reason, TSS (feed + VSS) was apparently different because of different VSS (e.g. microbial biomass). In our H-Floc system, the maximum TSS concentration was accordingly with previous reports (Ray et al., 2010; Gaona et al., 2011; and Schveitzer et al., 2013), and lower TSS concentrations were observed in the A-Floc, FT and RAS systems. These differences were significantly evidenced in all four water management systems. According to Ebeling et al., (2006) that autotrophic bacteria (abundance in the A-Floc systems) Page 52 of 208

was promoted mostly based on Nitrosomonas and Nitrobacter (Capone et al., 2008) following the equation 1 theoretically for bacterial biomass:

+ - - NH4 + 1.83 O2 + 1.97 HCO3 Æ 0.024 C5H7O2N + 0.976 NO3 + 2.9 H2O + 1.86 CO2 (1)

And, heterotrophic bacteria (abundance in the H-Floc systems) were promoted mostly based on Bacillus, Paracoccus, Pseudomonas, Azoarcus (Capone et al., 2008) following the biosynthesis equation 2 theoretically for bacterial biomass:

+ - NH4 + 1.18 C6H12O6 + HCO3 + 2.06 O2 Æ C5H7O2N + 6.06 H2O + 3.07 CO2 (2)

Some species of heterotrophic nitrification bacteria are highly resistant to ammonia. That benefits of heterotrophic nitrification bacteria have great importance for treating organic wastewater high in salinity and nitrogen (Chen et al., 2016).

2.4.4 The H-Floc: the best system for shrimp culture

In comparisons between two type of flocs following the two equations, in the H-Floc system, bacteria could increase the levels of water inorganic carbon (CO2) and subsequently decrease the pH (higher pH in the A-Floc system), as also reported by Panjaitan., (2010). Similarly with the - most top NO2 -N concentration at week 2 in the H-Floc systems, as by Tacon et al., (2002) - reported that an increase in nitrite (NO2 -N) caused in decreasing the pH in shrimp culture. Alteration in the C/N ratio (our study used C/N at 15/1 and semi-biofloc with carbon adding every three days) results in a shift from heterotrophic dominated biofloc forms to chemoautotrophic dominated biofloc forms, or vice versa (Avnimelech, 2015). Microbial flocs (bioflocs) are formed by heterotrophic bacteria, microalgae, zooplankton, and protozoa attached to a floating surface - suspended growth (Najdegerami et al., 2016; Martínez-Córdova et al., 2017). These heterotrophic bacteria have served not only as a direct food source for shrimp but also to shorten the food chain in the rearing water, recirculating nutrients and promoting the proliferation of bacteria that could inhibit the reproduction of pathogenic bacteria (Martínez- Córdova et al., 2016). While, the chemoautotrophic dominated biofloc forms are considered biofilms microorganisms attached to the bio-plastic surface as our experimental design, thereby, limitation in recirculating nutrients in the rearing water for shrimp.

In general, biofloc is co-culture of heterotrophic bacteria and autotrophic bacteria (Ebeling et al., 2006). Therefore, causing in a dynamic of bacterial proportions in biofloc systems. This Page 53 of 208

indicates that using biofloc systems is very difficult and labour-intensive (Tran, 2014; Prangnell et al., 2016), though the H-Floc system would be the best system in promoting beneficial bacteria (Bacillus spp. and K-strategists) and shrimp weight with a lower pH value if the system was added C/N ratio correctly and by a correct feeding regime. Additionally, the application of BFT through the H-Floc system in intensive shrimp aquaculture offered the possibility of - maintaining good water quality (NO2 -N concentrations decreased from week 3 onwards), producing additional food for shrimp, all as a result of promoting more proportion of heterotrophic bacteria (e.g. Huang et al., 2017). This finding about the advantages (lower pH, higher shrimp weight) of the H-Floc system (with an abundance of HB) was also in line with the results of previous studies, but we would highlight the importance of techniques to add C/N ratio correctly in terms of concentration and adding speed (De Schryver et al., 2008; Xu and Pan, 2014; Avnimelech, 2015; Xu et al., 2016; Huang et al., 2017).

2.4.5 Selected isolates as indicators

The isolates from this study supported the evidence that the presence of potential bacterial diseases (opportunistics) does not always cause the mortality of shrimp. The bacterial components stability regarding proportions according to the different ecological niches is an important implication to control the opportunistic bacteria. It seems that potential opportunistic pathogens (Vibrio spp., for examples about the candidates of MH100967, MH100968, MH100964, MH100974) together with a large number of non-pathogenic species existing in the rearing water, may link to a new case of polymicrobial disease in which non-pathogenic bacteria contribute to building up animal mortality (Lemire et al., 2015). The genus Pseudoaltermonas was identified from the pink flocs, which was proposed by closest relative that may be used as a bio-index because the candidate Pseudoalteromonas arabiensis (e.g. MH100976) is novel species that produces-polysaccharide (Matsuyama et al., 2013; Wu et al., 2017); and has probiotic properties to fight against fish pathogens (Wesseling et al., 2015), which might be functional in aquaculture bioaugmentation as or potential bio-indicator candidates, which was described in our study methods for a sustainable floc system in the future.

2.4.6 Further studies

However, some limitations are worth noting. The presumptive Vibrio spp. proportions at a fraction of shrimp sample need to be more frequent sampling to determine because suc-Vibrio

Page 54 of 208

generally overweighed suc+Vibrio. Presumptive suc-Vibrio and suc+Vibrio in this study should be further studied using qPCR, and full-length sequencing reads for identifying toxic genes and species. The TAN/TIN ratio should be examined in more depth to calculate the rate of energetic costs of intensive shrimp culture systems. Future work should evaluate the 50-μm and 250-μm size SS with more frequent sampling and whether the pink colour 50-μm size SS can be used as a bio-index of floc quality in the H-Floc system. Such further works could help to learn more about the presence of Vibrio spp. in the systems without disease outbreaks and the productive system operations. As recommended by Martínez-Porchas (2017), the different microbial species in the H-Floc and A-Floc systems require more research (Emerenciano et al., 2017), as the prospect of aquaculture relies on suitable microbes to feed cultured animals like shrimp. Worldwide, the current evidence supports the hypothesis that using microorganisms as a direct food source in aquaculture may revolutionize the industry, bridging the gap towards sustainability (Martínez-Córdova et al., 2016).

CONCLUSIONS

We conclude that different water management (four different systems) causes differences in water quality, bacterial proportions of r- and K- strategists (HB, Bacillus, and Vibrio), and shrimp weight. The growth of bacteria within the systems is the most important factor in determining sustainability since all four systems were starting-up with almost the same diversity in the microbial community.

In the complex interactions among microbes, microbes with larvae and finally microbes with larvae and environment, the pH balance was the most important indicator in healthy and sustainable shrimp culture systems. In this study, the pH close to 7.4 in the H-Floc system stabilized the internal microbial components in marine shrimp culture. However, maintaining this pH in the slightly alkaline state was a constant challenge, especially without adding chemical solutions into the rearing water.

The increase of aquaculture production without further increasing the use of the natural resources of water and land, and avoiding negatively influencing the environment can potentially be met

Page 55 of 208

by applying BFT. Such technology includes the H-Floc system described in this study and may be used to contribute to more sustainable and productive marine shrimp aquaculture.

Author Contributions LD, PS, AG and PB conceptualized and designed this work. LD was involved in carrying out the experiment, sample processing and data collection. CH and LD completed the statistical analysis of the data. LD prepared the manuscript with input from all co-authors.

Funding This work was partially funded by three stakeholders (Vietnamese Government-Project Number: VIED-911 for PhD studying, ZMT and ARC-Ghent University-Project Number: VAR02116).

Conflict of Interest Statement The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Acknowledgments

We would like to thank the staff of the ARC and ZMT. We would also like to thank Ass. Prof. Nguyen Quang Linh (Hue University), and Dr. Germán Kopprio (ZMT) for their comments and revisions. We would like to thank the reviewers for their thoughtful comments and suggestions, which improved the overall quality of our manuscript.

Page 56 of 208

Table 1| Biofloc performance parameters (mean and standard deviation).

Time System1 Colour Transparency SS2 (mL L-1) by size (cm) Natural size 250 μm filter 50 μm filter Week 1 H-Floc Brown 10.00 ± 0.00Abc 5.50 ± 0.35Ba 4.00 ± 0.61a 0.00 ± 0.00 A-Floc Green- 17.00 ± 0.00Bb 2.83 ± 0.24Aa 2.83 ± 0.24a 1.33 ± 0.47 Brown Week 2 H-Floc Brown 8.75 ± 0.43ab 10.50 ± 1.66Bb 7.25 ± 4.49b 0.00 ± 0.00 A-Floc Green- 10.00 ± 0.00a 8.33 ± 0.47Ab 9.33 ± 0.94b 0.33 ± 0.47 Brown Week 3 H-Floc Brown 11.00 ± 2.12Ac 3.50 ± 0.35Ba 2.50 ± 1.46a 0.38 ± 0.65 (pink colour) A-Floc Green- 21.33 ± 0.47Bc 2.00 ± 0.00Aa 1.50 ± 0.41a 0.00 ± 0.00 Brown Week 4 H-Floc Brown 7.25 ± 0.43Aa 4.88 ± 1.43Ba 3.00 ± 1.54a 0.00 ± 0.00 A-Floc Green- 20.00 ± 0.00Bc 3.00 ± 1.47Aa 0.50 ± 0.00a 0.00 ± 0.00 Brown

1 A-Floc: autotrophic biofloc system; H-Floc: heterotrophic biofloc system 2 SS: Setting solids A, B Different uppercase superscripts letters indicate a significant difference among systems a, b Different lowercase superscripts letters indicate a significant difference among sampling times

Page 57 of 208

Table 2| GenBank sequence accession numbers of partial 16S rDNA sequences of the isolates and related strains sharing 100% sequence identity with the sequences of the isolates (based on NCBI BLAST, date accessed 01.03.2018).

Time Isolates Closest relatives The possible functionality of isolates according to the closest relative Water mix LT3, M39 Bacillus thuringiensis AHL-degrading bacteria can decrease mortality caused by aquaculture pathogens (Defoirdt et MH100960, MH100969 al., 2011). PV1 Vibrio parahaemolyticus The opportunistic marine pathogen. However, the mechanism of disease-causing was not clear

MH100967 because the toxic gene may be lost via transposition (Lee et al., 2015).

LV11 Vibrio tubiashii “Pathogen of Bivalve Mollusks” (Hada et al., 1984). MH100968

LV21 Vibrio campbellii “Bacterial pathogen (for luminous vibriosis) of Litopenaeus vanname” (Haldar et al., 2010). MH100964

B12 Bacillus tianshenii This strain is a novel bacterium with broad ranges of pH, salinity and temperature (Jiang et al., MH100962 2014).

M41 Bacillus mycoides “A bacterial pathogen of channel catfish.” (Goodwin et al., 1994). MH100971 “Variation in the morphology.” (Stratford et al., 2013).

M7, B29 Bacillus cereus “Causing foodborne disease (Three enterotoxins: hemolysin BL - HBL, nonhemolytic MH100970, MH100963 enterotoxin – NHE, and cytotoxin K)” (Tajkarimi, 2007).

But, “B. cereus var. tonoi is novel bacterium for commercial product TOYOCERIN.” (Jiménez et al., 2013). P17 Pseudoalteromonas arabiensis “A novel bacterium, marine polysaccharide-producing bacterium.” (Matsuyama et al., 2013; Wu MH100976 et al., 2017).

Page 58 of 208

P14 Bacillus wiedmannii “A facultative anaerobe bacterium, aerobic endospore-forming bacteria.” (Miller et al., 2016). MH100973

Experimental PV2 Vibrio alginolyticus “Bacterial pathogen of coral reefs.” (Zhenyu et al., 2013). period MH100974 “Bacterial pathogen of cobia Rachycentron canadum.” (Liu and Lee, 2004).

M16 Bacillus pseudomycoides This strains is an aerobic, spore forming, rod-shaped, non-motile organism (Nakamura, 1998). MH100965 This strains could produce antimicrobial substance with activity against Gram-positive bacteria (Basi-Chipalu et al., 2015). M40 Bacillus mycoides “A bacterial pathogen of channel catfish.” (Goodwin et al., 1994). MH100966 “Variation in the morphology.” (Stratford et al., 2013).

B2, P16 Bacillus wiedmannii “A facultative anaerobe bacterium, aerobic endospore-forming bacteria.” (Miller et al., 2016). (P16 was MH100961, MH100975 isolated from pink-floc) P13 Pseudoalteromonas gelatinilytica “A novel bacterium with biofilm formation can prevent marine fish egg clutches from (isolated MH100972 pathogenic microbes infection.” (Aranda et al., 2012; Wesseling et al., 2015; Yan et al., 2016). from pink- floc)

Page 59 of 208

Source water: Un-filtered aerated marine water

33L 43L 28L 28L

+ 5L water mix + 5L A-Floc + 5L H-Floc matured water matured water

FT (35L) RAS (35 + 15L) A-Floc (35L) H-Floc (35L)

FT (35L) RAS (35 + 15L) A-Floc (35L) H-Floc (35L)

FT (35L) RAS (35 + 15L) A-Floc (35L) H-Floc (35L)

FT (35L) RAS (35 + 15L) H-Floc (35L)

Addition of 2L of water mix prior to stocking

Carbon adding Water exchange Biological filtration principle (C:N 15:1) (Conventional (van Rijn et al., 2006; Ebeling et al., 2006) (Avnimelech, 2015) aquaculture)

Figure 1| Flowchart of the experimental design. FT is the flow-through system, and RAS is recirculation system, H- Floc is the heterotrophic biofloc system, and A-Floc is the autotrophic biofloc system. In the FT system, 35 L consisting of 33 L of source water and 2 L of water mix. In the RAS system, 35 L consisting of 33 L of RAS water mix which was prepared two weeks in advance and 2 L of water mix; 5 L water of sedimentation chamber; 10 L water of nitrifier tank. In the A-Floc systems, 35 L consisting of 28 L of source water, 2 L of water mix, and 5 L of matured A-Floc water. In the H-Floc system, 35 L consisting of 28 L of source water, 2 L of water mix, and 5 L of matured H-Floc water.

Page 60 of 208

● System ● FT ● RAS + ● H−Floc 1.5 NH4−N ● A−Floc NH3−N ● ● Sampling time ● ● ● w0 ● ● 1.0 ● ● w1 ● ● w2 ● ● ● ● ● w3 Alkalinity ● ● w4 0.5 ● DO● ● pH ●● ● ● ● ● ● ● ● ● PC2 (16.41%) PC2 ● ● ● 0.0 ●● ● ● ● TSS ● ● ●●● ● ● ● VSS − ● ● NO2−N ● ● ●● − NO −N ● Salinity 3 ●● ● ● ● −0.5 Temperature● ●● ● ● ● ● ● ● ● ● ●● ● ● ● −1.0 ● ● ●

−1.5 −1.0 −0.5 0.00.51.01.5

PC1 (33.48%) Figure 2| Water quality parameters determined during the experiment, which were plotted by Principal components analysis (PCA). Black colour dots present the variables of the FT system (flow-through system). Red colour dots introduce the variables of the RAS system (recirculation system). Green colour dots present the variables of the H-Floc system (heterotrophic biofloc system). Blue colour dots introduce the variables of the A-Floc system (autotrophic biofloc system). Small dot to bigger dot as time points from beginning to the end of the experiment, the arrow shows the direction of the data from smaller value to bigger value.

Page 61 of 208

(A) (B) 38

● ●● 36 ● ● 8.5 ●● ● ● ● ● ● ● ● ● ● 34 ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 32 ● ● ● 8.0 ● ● ● ● ● ●●● ● ●● ● ● ● ● ● ● ● ● ●●●●● ● ●● ● ●●●●●●●●●● ● ●●●●●●● ● ● ● ● ● ● ● ● ● ● 30 pH ● ● ● ● ●● ● ● ● ● ●● ●● ● ● ● ● ● ● Salinity ● ● ● 28 ● ● ● ● ● ● ● 7.5 ● ● ● ● ● ● 26 ● ● 24

(C) (D) 30.5

● 8 ●

C] 30.0 ●

] ● ° ● 1 ● − ● 7 ● ● ● ● ● ● 29.5 ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ●● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 6 ● ● ●● ●● ● ●● ●● ● ● ● ● ● ● ● 29.0 ●● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● [mg L DO 5 28.5 ● ● ● ●● ● ● ● ● ● ●●● ●● ●● ● ● ● ● Tempe rature [ ● ● ● ● ● ● ● ● ●● ● ● ● 28.0 4

(E) (F) 120 0.6

] ● ] 1 1 − 0.5 ●

100 − ●

0.4 ● 80 ● ● 0.3 ● ●●●●● 60 L N [mg ● ● −

3 0.2 ● ● ●●●●●● ● ● ● 40 ● ● NH 0.1 ● ●

Alkalinity [mg L ● ● ● ● 20 ●●●● ● 0.0

(G) (H) 35 10 ] ] 30 1 1 − − 8 ● 25 ●

20 ● ● ● 6 ● ● ● 15

● N [mgL N [mgL ●● ● − − 4 ● ● − 2 + 4 ● 10 ● ● ● ● ● ● 5 NH

● NO 2 ● ●● ● ● ● ●● ● 0 ● ● ● ●● ●●

(I) (J) 150 500

] ● 1

● ] − 400 1

● − 100 ● ● ● ● ● 300 ● ● ● ● ● ● ● ● N [mg L N ● 50 ● ● ●● − ● 200 ● ● − 3 ● ● ● ● ● ● ● ● ● ● TSS[mg L ● ● ● ●

NO 100 0 ●

(K) (L) 400 1.2

●

] 1.0 1 300 − ● ● 0.8 ● 200 0.6 ● ● ● ● ● ● ● 0.4 ● 100 ● ● ● VSS [mg L [mg VSS ● ● ● ● ● Ratio TAN/TIN ● ● ● ● 0.2 ● ● ● ● ●● ● ● ● ●● ● ● ●● 0 0.0 02468101214 21 28 (M)

0.9 ● System ● ● 0.8 ● ● ●● ● FT ● 0.7 ● ● ● ● ● ● ● RAS ● 0.6 ● ●● ● ● ● H−Floc 0.5 Ratio VSS/TSS Ratio ● A−Floc 0.4

02468101214 21 28 Day of experiment Figure 3| Water quality parameters determined during the experiment among culture systems according to sample times. The details of statistical analysis can be seen in the Supplementary Table S1.1 and Supplementary Table S1.2. Back colour presents the flow-through system (FT), red colour shows the recirculation system (RAS), green colour presents the heterotrophic biofloc system (H-Floc), and blue colour presents the autotrophic biofloc system (A-Floc). DO is dissolved oxygen. TSS is total suspended solids. VSS is volatile solids. TAN is total ammonia nitrogen. TIN is total inorganic nitrogen.

Page 62 of 208

Water column [CFU mL−1] Free−living [CFU mL−1] Particle−attached [CFU g−1]

1e+08 ● ●

● ● ● ● 1e+07 ●● ●

● ● ● ● ● ● HB ●● ● ● ● 1e+06 ●● ● ● ● ● ● ● ● ● ● ● ● ● ● 1e+05

1e+09 1e+08 1e+07 ● 1e+06 ● ● ● 1e+05 ● Vibrio

− 10000 ● ●● ● ● ● 1000 ● ● ● ● ● ● suc ● ● ● 100 ● ● ● ● ● ● ● ● ● 10 ● ● ● ● ● 1 ●●

1e+08 1e+07 ● ● 1e+06 ● ● 1e+05 ● ● ● 10000 ● ● ● ● ● ● 1000 ● ● ● ● ● ● ● ● ●

suc+ Vibrio suc+ ● 100 ● ● ● ● ● ● ● ● ● 10 ● ● ● 1 ●

● 1e+06 ● ● ● ● ● ● 1e+05

● ● ● ● 10000 ● ● ● ● ● ● ● ● ● ● ● ● ● ● Bacillus 1000 ● ● ● ● ● ● ● ● ● 100 ●

w0 w1 w2 w3 w4 1e+07

●● System ● ● ● ● ● 1e+06 ●● FT ● ● ● ● ● ● ● ● ● ● ● ● RAS traits ● ● −

r ● ● ● ● ● H−Floc 1e+05 ● ● ● ● A−Floc

1e+08

1e+07 ●

● ● 1e+06 ● ● traits ● ●● ● ● − ●● ● ● ● K ● ● ● ● ●● ● ●● 1e+05 ● ● ● ●

w0 w1 w2 w3 w4 w0 w1 w2 w3 w4 Sampling time Figure 4| Microbial indicators were determined during the experimental periods according to sample times. The details of statistical analysis can be seen in the Supplementary Table S2.1 and Supplementary Table S2.2 The figure presents the water column sample, free-living sample and particle sample. Black colour presents the flow-through system (FT), red colour shows the recirculation system (RAS), green colour presents the heterotrophic biofloc system (H-Floc), and blue colour presents the autotrophic biofloc system (A-Floc). CFU is a colony forming unit. HB is total heterotrophic bacterial count. suc-Vibrio is presumptive green Vibrio number. suc+Vibrio is presumptive yellow Vibrio number. Bacillus is Bacillus spp. number. r_traits are r-strategists determined via CFU count according to r/K selection theory. K-traits are K-strategists determined via CFU count according to r/K selection theory. w1, w2, w3 and w4 are week 1, week 2, week 3 and week 4.

Page 63 of 208

(A) (B) System 2.0 140 ● FT ● RAS ● H−Floc 120 ● A−Floc 1.5 100

80 1.0

60 Shrimp [g] weight Weight gain rate [%] gain rate Weight 40 0.5

20

0.0 0

(C) (D) 8 100

80 6

60

4

40 Feed conversion rate conversion Feed Protein efficiency ratio 2 20

0 0 w1 w2 w3 w4 w1 w2 w3 w4

Sampling time Figure 5 | Shrimp growth performance parameters. The details of statistical analysis can be seen in the Supplementary Table S3.1 and Supplementary Table S3.2. FT presents the flow-through system. RAS presents recirculation system. H-Floc presents the heterotrophic biofloc system. A-Floc presents the autotrophic biofloc system. w1, w2, w3 and w4 are week 1, week 2, week 3 and week 4.

Page 64 of 208

FCR suc-Vibrio

DO

WGR TAN/TIN

- NO3 -N

suc+Vibrio - NO2 -N

NH3-N + NH4 -N

r-strategists Temperature Weight

Alkalinity HB K-strategists Salinity

Feed

TSS pH

VSS

Bacillus spp. −1 01 Correlation coefficient

Figure 6| The rmcorr correlations network. Edges are the lines connecting the node in the network. The width and colour of the edges correspond to the strength of the correlation. Red edge is a significant positive correlation. Blue edge is a significant negative correlation. The strongly correlated nodes are closer together. TSS represents total suspended solids. VSS presents a volatile suspended solid. DO presents dissolved oxygen. TAN/TIN presents the ratio of Total ammonia/Total inorganic nitrogen. WGR presents shrimp weight gain rate. FCR presents shrimp food conversion rate. HB presents heterotrophic bacteria.

Page 65 of 208

Chapter 3

Page 66 of 208

Chapter 3

Influence of environmental factors on the bacterial communities of four different shrimp culture systems

Loan Thanh Doan1,2*, Andreas Kunzman1, Germán A. Kopprio1, Achim Meyer1, Christiane Hassenrück1 and Astrid Gärdes1.

1Leibniz Centre for Tropical Marine Research (ZMT), Fahrenheitstr.6, 28359 Bremen, Germany 2Vietnam National University of Agriculture, Trau Quy, Gia Lam, Hanoi, Vietnam

* Correspondence: Corresponding Author [email protected]; [email protected]

Keywords: Litopenaeus vannamei, biofloc, post-larvae, microbiota, aquaculture, Illumina sequencing

Running title: microbes in intensive shrimp cultures

ABSTRACT

The ecology of bacterial communities was studied in four intensive aquaculture systems for Litopenaeus vannamei using Illumina Miseq amplicon sequencing for 16S rRNA genes. The biological variables were examined in relation with major environmental factors such as - - temperature, salinity, pH, dissolved oxygen (DO), nitrite-N (NO2 -N), nitrate-N (NO3 -N), + ammonium-N (NH4 -N), un-ionized ammonia (NH3-N), total suspended solids (TSS), and volatile solids (VSS). The aquaculture systems: flow-through (FT), recirculation (RAS), heterotrophic biofloc (H-Floc) and autotrophic biofloc (A-Floc), were set up for the experiment of post-larval shrimp culture until day 30, the most crucial period for the instability of microbial communities after shrimp stocking. Clear differences were observed regarding to the richness of operational taxonomic units (OTUs) and community composition among the four aquaculture - systems. Bacterial OTU richness was only influenced by TSS, pH, VSS, temperature, NO2 -N

Page 67 of 208

(GLMMs, F-ratio = 600.60, p < 0.001). Moreover, bacterial community structure was influenced - - by TSS, pH, VSS, temperature, NO2 -N; but also by other factors such as DO, NO3 -N, salinity, alkalinity (RDA, Adjusted R2 = 0.22, Variance = 7809.50, F-ratio = 4.16, p = 0.001). Our study showed clearly that environmental parameters and mainly TSS, can endeavour a strong 2 structuring influence on bacterial communities (RDA, PC1, R TSS = 0.43, p = 0.001) in shrimp rearing water. Moreover, the shrimp weight gain rates were correlated among the four systems and between weeks, with the following OTUs: Ruegeria (sq90, rrm = 0.64, p = 0.008; and sq1, rrm = 0.63, p = 0.009), Pseudoalteromonas (sq91, rrm = 0.52, p = 0.038), Catenovulum (sq142, rrm

= 0.75, p = 0.001), Holomonas (sq263, rrm = 0.66, p = 0.006), Roseovarius (sq175, rrm = 0.52, p =

0.038), Roseobacter (sq81, rrm = 0.61, p = 0.011), Pseudooceanicola (sq78, rrm = 0.50, p = 0.046). These relations suggest the need for further studies on the impact of individual OTUs (presumptive species) into more benefits for aquaculture such as shrimp weight and health, and their relationship with the amount and source of organic matter. Although the H-Floc system evidenced the maximum of OTUs from beginning to the end of the experiment, the highest OTU richness occurred in the A-Floc systems for water column, free-living and particle-attached samples. The OTU numbers in the particle-attached samples were much larger than in the free- living samples. Bacterial OTU composition changed with the time; however, their distribution was much influenced by the type of the culture systems. Another result, the most dominant single-sequence OTU abundance does not always belong to the relative class abundances of the community. Therefore, the individual sequence OTU (at the presumptive species level) with a high proportion has a greater ecological function in the system than the class relative abundances in this study do.

3.1 INTRODUCTION

In the field of the study of microorganisms, there is a hypothesis quoted by Baas-Becking and Beijerink: ‘everything is everywhere, but the environment selects’ (Wit and Bouvier, 2006). The true meaning of this statement is that while all microbial life is distributed worldwide, in a particular environmental condition, most of the microbial species are cryptic. Since most species occur at abundances below the laboratory limit of detection, it infers the hypothesis that all

Page 68 of 208

microorganisms are universal, but only some microorganisms are observed in their unique environments.

Lafferty et al., (2015) classified 67 examples from temperate water and determined that 34% of the infectious agents for marine fishes were bacteria. Other agents causing diseases had smaller percentages: 25% virus, 19% protists, and 18% metazoans. Bacteria infect a wide range of cultured fishes and crustaceans and can reduce their commercial production due to disease and off-flavours (Lee et al., 2015; Rurangwa and Verdegem, 2015). Therefore, there is a need to strengthen studies of bacterial ecology in aquaculture systems, the relationship between microbial communities, shrimp performances (growth and survivorship) and water quality (Martínez Cruz et al., 2012; Vadstein et al., 2018). Bacteria are characterized as food in aquaculture (microbial flocs); thus, they take an important role in aquaculture systems in positive and negative biological drivers (De Schryver et al., 2008; Crab et al., 2012; Martínez-Córdova et al., 2016; Nevejan et al., 2018). A common challenge of microbial research is predicting how the functions and/or composition of microbial communities respond to specific environmental and the disturbances (Shade et al., 2012). Amplicon-based sequencing is an efficient tool to support the analysis of bacterial diversity and community structure (Salipante et al., 2014; Lahens et al., 2017). Recently, Peters et al., (2018) suggested that environmental DNA (eDNA) with metabarcoding of the 16S rRNA genes can also help to quantify the complex mixture of microbes present in marine samples. The author especially reported that the number of OTUs was much larger in the unfiltered seawater samples (not filter through the 0.22 μm) than the pre- filtered samples. Such unfiltered samples were able to detect the expected changes in relative sequence abundances and were accurated in representing actual (presumptive) species abundances. Accordingly, our study was based on the unfiltered marine water samples from shrimp culture and will contribute to this gap in microbial research.

Not only the physical and chemical water parameters affect marine microbial communities, but also the organic matter influences strongly microbial communities (Austen et al., 2002; Schauer et al., 2010; Zinger et al., 2011; Bienhold et al., 2012; Nguyen and Landfald, 2015; Dang and Lovell, 2016; Learman et al., 2016). Furthermore, to understand clearly how the organic matter influence the bacterial ecology of aquaculture systems, in this work the diversity and community

Page 69 of 208

structure of the water column, free-living and particle-attached bacteria were compared from the four intensive culture systems of whiteleg shrimps.

This study aimed to understand how the physicochemical water characteristics influence the microbial community structure in marine shrimp aquaculture and to provide baseline information of the primary environmental factors that affect bacterial diversity and community structure in the specific culture systems for shrimp. We hypothesize strong differences in bacterial composition at each system and that TSS as an indicator of organic matter (APHA, 2005) will mostly drive such compositional differences in shrimp aquaculture systems.

The high productivity of Penaeus shrimps is a critical element of the growing aquaculture industry. Therefore, we hypothesize that the excellent shrimp performance and productivity could be explained by the relationships with some bacterial OTUs enhancing shrimp weight gain rates and survivorship.

3.2 MATERIAL AND METHODS

3.2.1 Experimental design and sample collection

Four intensive shrimp culture systems were set up with details as can be seen in the experimental design of chapter 2. Briefly, before the experiment, the biofloc culture (both heterotrophic and autotrophic) was established with low shrimp biomass of juveniles for two months. The RAS systems were settled before the experiment 14 days with shrimp juveniles. All four systems used the same source of marine water and water mix with Bacillus sp. LT3 as bio-inoculant. The stocking density of Litopenaeus vannamei PL24, and feeding regime were similarly applying to all four systems. The calculation of C: N ratio (15: 1) for heterotrophic biofloc systems takes into account the nitrogen content in the feed, the quantity of food distributed and the rate of nitrogen excretion by the shrimp as mentioned in the biofloc culture methods (Avnimelech., 2015). Shrimps were fed twice a day with commercial shrimp feed.

A total of 83 DNA samples were collected from 15 tanks of the four systems: 53 water column samples, 14 free-living samples, and 16 particle samples. Of which, free-living and particle samples were separated with an Imhoff Cone (AHPHA, 2015; Avnimelech, 2015). The tanks

Page 70 of 208

presented a similar size and water depth for each system. Samples were collected using sterile 50 mL falcon tubes were immediately placed on ice before further processing. To avoid DNA fragmentation, samples were treated shortly with liquid-N and stored at -80oC until DNA extraction was performed.

A CTAB method (SDS 20% and N-Lauroylsarcosin 10%) was used for DNA extraction following the company’s instructions. This method was performed on the lysis extract to remove polysaccharides and residual proteins (Wilson, 2001; Bossier et al. 2004; Nercessian et al., 2005). Nucleic acids were extracted using Phenol: Chloroform: Isoamyl Alcohol (25: 24: 1). After centrifugation at 17,000 g and 4 oC for 90 min, the pellet of DNA was washed with 70% ethanol to remove salts, dried under vacuum and resuspended in 40 μL TE-water at pH 8.0. The concentrations of the extracted DNA were measured using the NanoDrop (TECAN M200 Infinite Pro Microplate Reader) and their purity assessed by the ratio of absorbance at 260 nm and 280 nm. The DNAs were sent to LGC Genomics (Berlin, Germany) for high throughput Illumina sequencing with 300bp paired-end products. Regarding the 16S rRNA genes amplicon analysis, the V3F (341F - CCTACGGGNGGCWGCAG) and V4R (785R - GACTACHVGGGTATCTAAKCC) were selected because they are highly recommended for microbial diversity determination (Cai et al., 2013; Klindworth et al., 2013).

3.2.2 Processing of sequencing data

The demultiplexed and primer-clipped sequences provided by LGC Genomics were further processed with dada2 version 1.8.0 (Callahan et al., 2016) to generate operational taxonomic units (OTU) based on unique amplicon sequence variants. The learnErrors function from the dada2 package was applied for error rates, sequences were quality filtered at a maximum expected error rate of 4 and a minimum length of 240 and 220 bp for forward and reverse reads, respectively. The complete data set was used for error learning, denoising and pooling all sequences across samples. Merging of forward and reverse reads, as well as, chimera removal were conducted using default settings. Only OTUs between 401 and 429 bp that occurred more than once in the data set were considered for further analyses. These sequences were taxonomically classified using the silvangs service (https://www.arb-silva.de/ngs/, date accessed 27.06.2018) with version 132 of the SILVA reference database (Quast et al., 2013). The identical parameters for OTUs were removed from the 16S sequence data set including: 1) OTUs with a

Page 71 of 208

sequence similarity of < 93% to the reference database, 2) OTUs unclassified on phylum level, and 3) OTUs affiliated with eukaryotic, archaeal, mitochondrial, or chloroplast sequences.

3.2.3 Statistical analysis

Differences in bacterial alpha-diversity among the four systems were tested using the general linear mixed models (GLMMs) on rarefied OTU numbers as univariate analysis for the water column, free-living and particle-attached samples. To compare the bacterial OTU richness between free-living and particle samples we used a paired t-test. From the rarefaction curves were given in Supplementary Figure 1, we observed that there were no saturations even for the most deeply sequenced samples. Therefore, OTU richness assessed based on repeated (100x) rarefying runs to a minimum library size to compare among samples.

The Beta-diversity among the four systems was examined using Bray-Curtis dissimilarity matrix (Bray and Curtis, 1957; Legendre and Legendre, 1998; Legendre and Anderson, 1999) for analysis of dissimilarities (redundancy analysis, RDA) and ordinated by non-metric multidimensional scaling (NMDS) with the fitting of the environmental parameters (Chapter 2_dataset) applying the envfit function from the vegan package (Oksanen et al., 2013) as multivariate analysis on microbial community structure to visualize the differences between samples. The envfit finds vectors with projections of points onto vectors which have a maximum correlation with the corresponding environmental variables. The result is an object containing coordinates of quantitative variables that can be used to project these variables as factor levels into the ordination diagram. Thus, RDA is a multivariate analysis method combining multiple regression and classical ordination (PCA) to model (or explain) multivariate response variables based on community composition data tables (Legendre and Gallagher, 2001). Therefore, we used RDA to look for a possible strong linear correlation among the large set of explanatory variables, which could provide the regression coefficients of the explanatory variables. By application of RDA, we can conclude that the explanatory variables (i.e. TSS, VSS, salinity, - - + temperature, alkalinity, DO, pH, NO2 -N, NO3 -N, NH4 -N, NH3-N) explain the variation of the dependent matrix (i.e. OTUs). Data sets were centered log-ratio transformed before being tested. To avoid autocorrelations in explaining the influence of the environmental factors on bacterial structure, we did RDA for the selected PCs. In which, we applied the forward selection (Akaike information criterion, AIC) of explanatory variables for PCs, thereby a large set of explanatory

Page 72 of 208

variables were reduced afterwards by a selected model with four remaining PCs (PC1, PC3, PC4,

PC6). Comparisons between free-living bacteria and particle-attached bacteria per samples belong to both biofloc systems using pairwise test with the vegdist function (Bray-Curtis dissimilarity indices) in the vegan package.

The heatmap was also created with using RDA model for each OTU to get the differential OTU proportions among the four systems, in which the selected steps was applied for OTUs occurring at least 1% in one sample. We applied the Benjamini-Hochberg method for adjusted p-value. For fraction of water column, the filter steps were used by adjusted rrm (≥ 0.5), and adjusted p-value (≤ 0.05) to get the significant OTU proportion differences among the four systems. For the free- living and particle samples, the filter step was not applied. Repeated measurements correlations of the differential OTU high proportions from the water column samples and shrimp weight gain rates were determined by R package rmcorr (Bakdash and Marusich, 2017). The limitations of this test is only based on taxonomic resolutions (i.e. the relative sequence abundant OTUs) without absolute quantitative data such as bacterial cell count. According to author of the package, we use rrm to state about the coefficients of rmcorr test. All statistical analyses performed in R (R version 3.5.2, R Core Team, 2018, using R studio version1.1.456) with a number of identical R packages.

3.3 RESULTS

3.3.1 Characteristics of culturing systems

The four intensive shrimp culture systems showed the differences in water quality and shrimp performances as described in chapter 2. We summarized critical characteristics of the four shrimp systems and shrimp performance parameters (Table 1). Briefly, the best shrimp performance was in the H-Floc, a system with low pH (~ 7.4). Vice versa, the shrimp performance was the worst in the RAS system with high pH (~ 8.5) and clean water.

3.3.2 Alpha diversity

There were a total of 1,883,779,700 bases in 8,562,635 high-quality sequences from the Illumina MiSeq data. In summary, 25 - 80% of input sequences were retained, 47 - 82% input sequences

Page 73 of 208

were filtered, 90 - 100% filtered sequences was denoised, 78 - 99% of the denoised sequences were merged, chimeras from the 64 - 100% merged sequences were removed, 88 - 100% sequences with no chimeras were tabled into OTUs. Finally, a total of 5553 unique OTUs were generated for diversity analysis of the bacterial community.

The FT and RAS systems showed a decreased trend in alpha diversity to the end of the experiment. The H-Floc system evidenced a marked increment of OTUs from beginning to the end of the experiment. The highest OTU richness occurred in the A-Floc systems. The patterns of increasing OTUs in H-Floc and more OTUs diversity in the A-Floc systems were also recorded at the fractions of free-living and particle samples (Figure 1 and Supplementary Table 1).

An univariate analysis was performed to check the significant differences in OTU richness (Table 2) among the four systems at different fractions and samples. At the fraction of water column, we found strong differences in OTUs among the four shrimp culture systems (with GLMMs, F = 8.28, pairwise, p = 0.004). It should be taken into account that there was an interaction between systems and time points (weeks) on OTUs (GLMMs, F = 6.74, p < 0.001). At a fraction of free-living samples, there were no significant differences in OTU numbers between the two biofloc systems. At a fraction of particle samples, no significant differences in OTUs between the two biofloc systems were found, but there were significant differences in the time points, week 2 presented a less number of OTUs compared to week 4 in the H-Floc system (GLMMs, F =14.99, pairwise, p = 0.031). We observed that the environmental factors such as TSS, VSS and pH statistically shaped the OTU richness in the four shrimp systems (GLMMs, F - = 5.63, p ≤ 0.001). Temperature and NO2 -N influenced the OTU richness in the four shrimp systems (with GLMMs, F = 2.44, p = 0.025).

A high dissimilarity between free-living OTUs and particle-attached OTUs per sample (with Pair t-test, t = -3.71, df = 11, p = 0.003) was found. Moreover, the number of OTUs in the particle samples was much larger in the free-living samples.

3.3.3 Community composition

The relative sequence abundance of OTUs analysed using NMDS (ordination analysis) to study the distribution of the bacterial community in relation with physical and chemical characteristics.

Page 74 of 208

Figure 2 shows a clear distinct pattern between the RAS and three remaining systems (FT, H- Floc, A-Floc) in bacterial community structure. Particularly, it could be seen in panel A that TSS was a possible driving factor in structuring bacterial community. In panel B, there was a clear separation between free-living and particle-attached bacterial communities. Redundancy analysis further confirmed the influence of these environmental parameters within the four systems (RDA, R2 = 0.38, p = 0.001; Table 3). Approximately, the 38% of OTUs variation could be accounted for bacterial compositions by influences of environmental factors. At water column fraction, the full model explained 31% and 9% of the bacterial community considering systems and weeks, respectively. The OTU compositions changed much more to specific environmental parameters than to the time. As described in the methods, the RDA was applied with four PCs to avoid autocorrelation in explaining the influence of environmental factors on the bacterial structure. Among the four PCs, the PC1 explained the maximum 12% of variation in OTU 2 2 compositions among the four systems considering the variables TSS (R PC1 = 0.43), pH (R PC1 = 2 - 0.43), and VSS (R PC1 = 0.43) (Supplementary Table 3).

The dissimilarity between the two biofloc systems on bacterial composition was compared at two fractions. At the fraction of free-living samples, approximately 17% of OTU variation was explained by system factor, whereas 12% of OTU variation was explained by the week factor with adjusted R²-value of 26% on free-living community composition. At the fraction of particle samples, approximately the 34% of OTU variance was explained by factor system, by week was only 6%, with adjusted R²-value of 37% on particle community composition (Table 3). Especially, a very high dissimilarity of bacterial composition was found between free-living and particle samples with a median of 92% dissimilarity based on Bray-Curtis indices among two systems and between weeks (Pairwise, min = 0.76, max = 0.99).

In general, 35 phyla, 77 classes, 212 orders, 363 families, 782 genera were identified from the rarefied OTUs (Supplementary Figure 2). The analysis of relative abundances shows that eight classes approximately accounted for more than 80% of the total amplicons in four systems (Supplementary Table 2). In which, the Alphaproteobacteria was the dominant class in all four systems, varying from 8.22 to 82.82 % of OTU proportions. Secondarily, Gammaproteobacteria (0.61±0.21 - 58.05±20.58%), Bacteroidia (6.12±2.56 - 72.60%), Actinobacteria (0.35±0.01 - 33.58%), Parcubacteria (0.00 - 14.82%), Saccharimonadia (0.01±0.01 - 4.48 %),

Page 75 of 208

Deltaproteobacteria (0.08±0.02 - 9.57 %) and Chlamydiae (0.02 - 9.31±1.05 %) were abundant classes (Figure 3 and Supplementary Table 2). The FT system was characterized with the Saccharimonadia, and the RAS system was characterized by Bacteroidia, Deltaproteobacteria, and Gammaproteobacteria. Two classes of Actinobacteria and Alphaproteobacteria characterized the H-Floc system, while Chlamydiae and Parcubacteria were represented the A-Floc system.

As in Figure 3 details, for the water column fraction, heterotrophic Gammaproteobacteria were common bacteria in the FT system at week 0 and in the RAS system at week 4. Alphaproteobacteria abundance was higher in the H-Floc system at week 0 and week 3. The A- Floc system showed an even community pattern with other three moderately dominant classes: Actinobacteria, Bacteroidia, and Alphaproteobacteria. At the free-living fraction, the different patterns in class composition were observed between the two biofloc systems. Alphaproteobacteria was the dominant class in the free-living H-Floc samples at week 2 and week 4, while Chlamydia was the dominant class in the free-living A-Floc samples at week 4. At the particle fraction, changing relative abundances in OTU proportions were less frequent over time in comparisons with RAS samples. Mainly, Actinobacteria characterized the particle fraction in both biofloc systems.

At the fraction of the water column (Figure 4), Ruegeria sq1 and Marivita sq3 were important taxa at the beginning and from week 2 to the end of the experiment in the H-Floc system. In the FT system, Leisingera sq72 and Ruegeria sq90 showed the higher proportions at the beginning of the experiment and moderately at week 1. Marivita sq3 evidenced significantly higher proportions at week 4. In the RAS system, Catenovulum sq142 and Pseudoalteromonas sq91 were more abundant at week 2 and Mycobacterium sq32 at week 3. Hanstruepera sq5 and Saprospiraceae (uncl.) sq7 were dominant in the FT communities at week 2 and week 4, while Saprospiraceae (uncl.) sq7 was at week 4 in the A-Floc systems, and they were likely particle- attached bacteria. As also can be seen in Figure 4 at two fractions of free-living and particle samples, Ruegeria sq1 was likely a particle-attached bacteria in the H-Floc systems and Marivita sq3 was likely a free-living bacterium and important dominant taxa in the H-Floc systems when Ruegeria sq1 started to reduce its OTU proportions. In the A-Floc systems, Hanstruepera sq5, Saprospiraceae (uncl.) sq7, Lewinella sq23, Gammaproteobacteria (uncl.) sq26, Fulvivirga sq38

Page 76 of 208

and Gammaproteobacteria (uncl.) sq40 were found at high proportions in the water column and particle fractions, but less at the free-living fraction.

Notably, in the Supplementary Figure 3 - panel A, we can see the other trends in OTU changes among the four systems, there are more OTU proportions per samples in the H-Floc system, especially in the free-living fraction. Looking at the panel B of this figure (Supplementary Figure 3 and Figure 4), Marivita sq3 (Figure 4) presented a higher OTU proportions, which contributed up to more than 80 % of free-living OTU proportions (Supplementary Figure 3 - panel A and B).

3.3.4 Healthy and productivity predicted shrimp model

Generally, a moderate correlation between shrimp weight gain rates (WGR) with the number of

OTUs was found among the four systems (rmcorr, rrm = 0.58, p = 0.017). Then, we further tested the correlations between 36 significant differentials high OTU proportions of Figure 4 with shrimp weight gain rates. There were respective correlations as can be seen in Figure 5 and Supplementary Table 4. For the FT system, three candidates showed high positive correlations:

Ruegeria (sq90, rrm = 0.64, p = 0.008) with significant high proportion at week 0,

Pseudooceanicola (sq78, rrm = 0.50, p = 0.046) and Holomonas (sq263, rrm = 0.66, p = 0.006). For the RAS system, the positive correlation with WGR belong to the genus of Catenovulum

(sq142, rrm = 0.75, p = 0.001); Roseovarius (sq175, rrm = 0.52, p = 0.038); Pseudoalteromonas

(sq91, rrm = 0.52, p = 0.038). The bacteria characterizing the H-Floc system considering their positive correlation with WGR were: Ruegeria (sq1, rrm = 0.63, p = 0.009) at significant high proportions during whole period of the experiment and Roseobacter (sq81, rrm = 0.61, p = 0.011) with higher OTU proportions at week 0. For the A-Floc system, the candidate was Pseudooceanicola sq78 (similar to the FT system) with high OTU proportion at week 0. These target OTUs were positively associated with shrimp WGR and may serve as bio-indicators for each particular system in the Supplementary Table 5.

Page 77 of 208

3.4 DISCUSSION

3.4.1 TSS: a primary factor influencing bacterial communities

We focussed on studying the bacterial distribution and composition in rearing water of intensive Litopenaeus vannamei culturing starting from post-larval stocking until day 30 of the culture period, representing the period of sensitivity to disease outbreak (Tran, 2013), according to different environmental conditions, namely, FT, RAS, H-Floc and A-Floc. Differences in the distribution of relative abundances, alpha-diversity (in terms of bacterial richness) and bacterial community structure were found among the four systems, reflecting the physical and chemical water differences. The model explained 38% of the variation in community compositions of the water column samples and 26% and 37% of the bacterial variations for the free-living and particle samples, respectively, although the large fractions remained unexplained. However, the (large) remaining fractions could be understood according to the experimental designs, in which an absence of bias was ensured due to the different stocking density, different water source, different feeds and feeding regimes, different shrimp larvae for stocking and removal of the wastes or solids until the end of the experiment. Therefore, the finding variation percentages of community compositions from fractions of the water column, free-living and particle samples were highly explained by the full RDA model to find the best set of environmental factors responsible for the differences. As a summary, the patterns of the microbial community were understood, where environmental factors shaped the structure and functions (free-living or particle attached bacteria, high OTU proportions) of water microbial communities in intensive shrimp culture systems. Among the significant impact factors, such as temperature, salinity, - - alkalinity, DO, pH, TSS, VSS, NO2 -N, and NO3 -N, we found that TSS was the most influential factor shaping water microbial community structure, as well as strongly affecting to OTU richness among the four systems when other bias effects, such as adding antibiotics and disinfection chemical solutions into the systems, were avoided. Although, as the model showed statically, the VSS and pH factors have equal roles to that of the TSS factor in defining the variations, the VSS was calculated based on the TSS amount, and pH was not a primary factor generating differences in the systems in the method we applied for the experiment (APHA, 2005). Further, the TSS linearly correlated with pH value. We could infer that a higher TSS induced a lower pH (Chapter 2); alternatively, pH was an integrating variable providing an

Page 78 of 208

integrated index in the shrimp rearing water. Therefore, the changes in TSS influenced the changes in bacterial OTUs number and composition. For this reason, to culture shrimp, one must culture the water by regularly monitoring TSS and/or pH for a metacommunity property because changes in pH refer to variations of the microbial community (Mikkelsen et al., 1996; Ligi et al., 2014; Liu et al., 2015; Hou et al., 2017). Ray et al., (2010) and Schveitzer et al., (2013) proposed that a TSS concentration of approximately 460 mg L-1 (400 - 600 mg L-1) should be the goal of control in aquaculture systems for benefitting shrimp culturing. Since the excessive accumulation of organic matter (TSS, APHA, 2005) is reported to stimulate the growth of off- flavour-producing microorganisms (with a foul or faecal smell) and opportunistic pathogens by the favourite ecological niche (Masser et al., 1999; Defoirdt, 2016). This study records the first ecological microbial 16S rRNA sequenced data according to TSS maximum, at 426.50 ± 65.24 mg L-1 (Chapter 2) in shrimp culturing water. These results generally support our hypothesis that TSS is the primary environmental factor directly affecting the bacterial OTU richness and actively shaping the water microbial community structure.

3.4.2 Bacterial variation by taxonomic sequence abundances

Here, we present the major OTUs with significant high proportions, followed by eight relative sequence-abundant classes according to different patterns among the four systems (different environments). Litopenaeus vannamei post-larvae were cultured in the marine water (approximately 30 ppt); we found that the most abundant phylum (Proteobacteria) was presented by three relatively abundant classes Alpha-, Gamma- and Deltaproteobacteria followed by the different OTU proportions among the four systems and between the samples. This result was in line with the previous studies showing that this phylum has a wide phylogenetic and phenotypic diversity in several marine environments (Sogin et al., 2006; Williams et al., 2010; Zinger et al., 2011; Xiong et al., 2014; Hou et al., 2017). Especially, alphaproteobacterial OTUs were characterized (high dominance) by the H-Floc system, with a maximum of 82.82% relative sequence proportion at week 3. Alphaproteobacteria was also evidenced by five out of eight potential species from different systems, which positively correlated with shrimp weight gain rates.

First, according to our knowledge, Ruegeria sq1 (alphaproteobacterial OTU) as ruegerial bacteria in particle samples of the H-Floc systems has not yet been previously reported; in this

Page 79 of 208

study, Ruegeria is proposed as newly identified of the shrimp culture particle component. Notably, these bacteria presented as a stable population in the specific H-Floc environmental conditions from the beginning to the end of the experiment.

Second, Ruegeria sq90 (alphaproteobacterial OTU) was only dominant in the FT system at the beginning. Most species of Ruegeria genus can be switched between free-living or particle attach bacteria (Miura et al., 2018) thereby they were recommended for studying as biofilms candidates (Christie-Oleza and Armengaud, 2010; Miura et al., 2018), have a wide range of salinity and pH (pH 7.5 is optimum for growth) values (Huo et al., 2011), they can degrade phosphate triesters (Yamaguchi et al., 2016), they are ecologically diverse and they are well defined based on 16S rRNA sequenced data (Miura et al., 2018). Gram-negative heterotrophic marine ruegerial bacteria were reported as candidates of using N-acylhomoserine lactones (AHLs), which are quorum sensing signalling molecules for intercellular communication (Christie-Oleza and Armengaud, 2010; Cai et al., 2018). The use of molecular AHLs could inhibit the pathogenicity of many opportunistic pathogens from the Vibrio group (Defoirdt et al., 2011; Cai et al., 2018). Recently, Cai et al., (2018) found that Ruegeria mobilis YJ3, isolated from healthy shrimp, showed AHLs degradative activity via designated Ruegeria mobilis marine lactonase (RmmL). Recombinant RmmL could degrade AHLs in vitro; therefore, that study suggested that the RmmL of Ruegeria mobilis may be used as a therapeutic agent in aquaculture. Moreover, Ruegeria pomeroyi is the first marine heterotrophic Roseobacterium proposed for further study in more complex regulatory strategies using the diverse mixture of organic C molecules available in seawater. This use of Ruegeria pomeroyi could solve the challenges to the growth of heterotrophic bacteria, where they mostly have to rely on a C:N of approximately 15:1 (Singer et al., 2012; Medeiros et al., 2015; Avnimelech., 2015; Rivers et al., 2016). As a summary of ruegerial bacteria, such as Ruegeria sq1, they can exert some soft of biocontrol activity against the opportunistic Vibrio, especially the Vibrio harveyi - a pathogen using AHLs in cell-cell communication (Defoirdt et al., 2008; Crab et al., 2010c; Miura et al., 2018).

Third, this study showed that Marivita sq3 is likely a free-living bacterium of the Alphaproteobacteria. At week 3, this group reached a major population of free-living bacteria. The Marivita group seems to have required 2 weeks to reach the best fit in the specific H-Floc - environment. Notably, from week 3, the NO2 -N concentration was also reduced to the lowest,

Page 80 of 208

- and thus, the NO3 -N concentration was coercion reduced. Given these results, we need to - - perform more experiments on the correlation of Marivita sq3 number with NO2 -N and/or NO3 - N concentrations to determine the advantages of Marivita sq3 may have in the H-Floc system. Moreover, Alphaproteobacteria were mainly represented by Marivita spp. in most of the free- bacterial community (Cruz-lópez and Maske, 2016); these bacteria exhibit optimum growth at pH 7.5 - 8.5 (Zhong et al., 2015), and they are nitrate-reducing bacteria (Hou et al., 2017). Further, the Marivita genus was also closely associated with algae, resistant to grazing and present in several marine environments (Hwang et al., 2009); such characteristics are good for shrimp pond culture. The total sequence of alphaproteobacterial OTUs was the maximum number in the H-Floc system at week 3, while the total rarefied OTUs in the H-Floc system was the lowest at week 3. This means that, for the individual sequence, OTU abundance (OTU evenness) drives the functionality of the bacterial community in the shrimp culture system with an extremely high proportion. In chapter 2, the Vibrio number was smaller than 104 CFU mL-1; thus, disease did not occur in the four systems (Tran, 2013). From another perspective, it may be that the high proportions of Marivita sq3 and Ruegeria sq1 provided more benefits for the H- - Floc systems in terms of reducing the NO2 -N concentrations and shrimp weight gain. However, these suggestions need to be tested via the integrated high-throughput absolute abundance quantification (iHAAQ) method to find definitive quantitative data based on the relative sequence abundance results from this study to establish further potential conclusions (Lou et al., 2018).

Fourth, other potential Alphaproteobacteria that were positively associated with shrimp weight gain rates, such as Pseudoocenicola sq78 (in the FT and A-Floc systems), Roseovarius sq 175 (in the RAS systems), and Roseobacter sq81 (in the H-Floc systems), all had a swim-or-stick lifestyle (Bartling et al., 2018). However, the limitations of the correlation approaches need to be considered, such as that the test only showed moderately significant correlations and was based on taxonomic resolutions (i.e. sequence-abundant OTU proportions), in interpreting the results.

Fifth, Pseudoalteromonas sq91 and Catenovulum sq142 were found as RAS characteristic bacteria. Especially, they were not dominantly indicated in the three other systems. As a minor bacterium in the H-Floc system, but one that should be considered, Pseudoalteromonas sq91 was only present moderately in particle samples from week 3 to the end of the experiment. Recently,

Page 81 of 208

Pseudoalteromonas sp. (SCSE709-6) was shown to exert bioaugmentation effects for saline blackwater treatment (Jiang et al., 2019). Another Pseudoalteromonas bacterium (P. espejiana DSM9414) was reported as an amino-acid-requiring strain from seawater (Rong et al., 2018), and P. phenolica was reported as a novel bacteriophage isolated from the coastal water of Qingdao, China (Liu et al., 2018b). Especially, Pseudoalteromonas could be a novel bacterium with biofilm formation for preventing marine fish egg clutches from being infected by pathogenic microbes (Aranda et al., 2012; Wesseling et al., 2015; Yan et al., 2016), or it could be a novel marine bacterium for producing polysaccharide (Matsuyama et al., 2013; Wu et al., 2017). It has a wide range of pH values, with pH 6-10 and pH 7-7.5 optimum for growth (Isnansetyo and Kamei, 2003; Yan et al., 2016; Ying et al., 2016), and it has an important role in anti-VpAHPND (Vibrio parahaemolyticus-caused acute hepatopancreatic necrosis disease, AHPND) as a probiotic candidates and potential biological control agent against AHPND in shrimp aquaculture (Wang et al., 2018a). Catenovulum sq142 was only visible with a high proportion in the RAS system at week 2. These bacteria were highly correlated with shrimp weight gain rates. Recently, the genus Catenovulum was recognised as having the ability to degrade algal biomass as anti-biofouling bacteria (causing off-flavours) in Penang, Malaysia (Lau et al., 2019); these bacteria have a wide pH range, at 6-9 (Shi et al., 2017), as a very good characteristic enabling them to adapt to a number of different environments/niches.

Sixth, Halomonas sq263, in the class Gammaproteobacteria, was only high dominant in the FT and A-Floc systems at the beginning of the experiment; however, these bacteria are highly correlated with shrimp weight gain rates. Halomonas is a genus of salt-tolerating proteobacteria, where 5 - 25 ppt is optimum for growth. According to Gaboyer et al., (2014), a summary of the genus Halomonas showed that halomonas bacteria are widespread and commonly colonise extreme environments, is hypersaline, non-marine biofilms, human blood and deep oceans (Vreeland et al., 1980; Swings et al., 2002; Reddy et al., 2003; Oueriaghli et al., 2014). However, the bacteria were not specific bacteria in our RAS and H-Floc systems. Therefore, Halomonas sq263 showed a lower response than Catenovulum sq142 and Pseudoalteromonas sq91 did in the RAS conditions, as well as lower adaptation than Ruegeria sq1 and Marivita sq3 in the H-Floc conditions. In contrast, after the shrimp stocking, Halomonas sq263 did not exhibit permanent resistance in the FT and A-Floc systems with high proportions. Halomonas groups seemed to represent good bacteria, but the environmental conditions of this study did not select them.

Page 82 of 208

Saccharimonadia had FT class characteristics with a maximum of 4.48% relative abundance. In the A-Floc systems, the maximum relative abundance of Chlamydiae and Parcubacteria were 9.31±1.05% and 14.82%, respectively. Notably, Bacteroidia and Deltaproteobacteria were RAS characteristics with extremely high relative abundances of 72.60% and 9.57%, respectively. However, the bacteroidial OTUs were not as high in proportion as Catenovulum sq142 and Pseudoalteromonas sq91 (gammaproteobacterial OTUs) in the RAS systems were. As a result, the most dominant single-sequence OTU abundance does not always belong to the relative class abundances of the community. Therefore, the individual sequence OTU (at the presumptive species level) with a high proportion has a greater ecological function in the system than the class relative abundances in this study do. Thus, the RAS environmental conditions are the dimensions in the ecological niche for Catenovulum and Pseudoalteromonas strains. Conversely, the Catenovulum and Pseudoalteromonas species could be potential bacterial bio-indicators for the RAS system according to this study design.

To give more details on the A-Floc system, Parcubacteria were usually observed in exclusively anoxic environments as a group with an anaerobic lifestyle or groundwater environment, although there were many publicly available parcubacterial genome sequences illustrating that the organisms are found in oxic environments (Nelson and Stegen, 2015). The parcubacterial organisms have also been implicated in hydrogen and sulfur cycles in anoxic sediments (Wrighton et al., 2014). Chlamydia bacteria are strict intracellular bacteria characterised by a biphasic developmental cycle, for example, C. trachomatis and C. pneumoniae, which are associated with genital infections and respiratory diseases, respectively (Vouga et al., 2018). Therefore, the A-Floc systems evidenced no special finding on beneficial bacteria for shrimp aquaculture. This result on the A-Floc system may be an answer to the proposed research question on the differences in bacterial populations between the H-Floc and A-Floc systems by Emerenciano et al., (2017) and Martínez-Porchas, (2017). As first clarified in chapter 2 and reiterated in this chapter, the H-Floc system evidenced the advantages of heterotrophic bacteria, for example, Marivita sq3 and Ruegera sq1 (Hou et al., 2017) in comparison with the A-Floc system in shrimp aquaculture.

Seventh, Mycobacterium bacteria (e.g. Mycobacterium sq253) belong to class Actinobacteria, which also have H-Floc characteristics (more likely particle bacteria than free-living bacteria)

Page 83 of 208

with a maximum relative abundance of 33.58% among all four systems. Members of Actinobacteria are a vital source of peptide products; peptides exhibit strong activity against various human pathogens and they represent attractive alternatives to current antibiotics as bioactive molecules (Gao and Gupta, 2012; Kao et al., 2016; Zhao et al., 2018). Examples include Streptomycetes (the largest genus of Actinobacteria), which are the most abundant source of natural antibiotics (Liu et al., 2013) and biopharmaceuticals have an emerging need for Mycobacterium tuberculosis (Zumla et al., 2013; Zhao et al., 2018). Actinobacteria provide future drugs against acute diseases like cancer, HIV, microbial and protozoal infections and severe inflammations (Hassan et al., 2017; Hassan and Shaikh, 2017). In summary, the Actinobacteria have proven to have various important roles as a ‘gold mine’ in biotechnological applications in several industrial and agricultural processes; they serve as bioinoculants for sustainable agriculture (Sharma et al., 2018; Yadav et al., 2018), and future perspectives in nanotechnology suggest that multi-omics analysis will enable the design of new strategies using Actinobacteria (Alvarez et al., 2017). Commonly, Actinobacteria are universal in marine water, and all major novel natural products from Actinobacteria can be found in the research of Karuppiah et al., (2016). In constrast, Mycobacterium tuberculosis is the causes of disease and drug-resistance in humans (Koch and Mizrahi, 2018). Accordingly, the question is how to increase the number of novel species and reduce the harmful species of Actinobacteria taxa, representing an emerging need for ecologists and aquaculturists. This is because Actinobacteria have always been detected as the dominant phyla in enclosed shrimp culture ecosystems (Xiong et al., 2014; Qin et al., 2016). On this account, we should combine culture-dependent and culture-independent technology to find the potential Actinobacteria candidates strains for the acute H-Floc systems to generate urgently needed bio-shrimp products.

Eighth is Marixanthomonas sq382 from the class Bacteroidia; Hahnke and Harder, (2013) reported that Marixanthomonas bacteria need a maximum of about 20 days to form visible colonies as K-strategists bacteria (i.e. bacteria are slow growing and can grow in the condition of limit nutrient per bacterium), although the 16S rRNA gene sequences of Marixanthomonas have only been found in cultivation-independent studies. Ninth, the high abundance of Lewinella genus from Actinobacteria (e.g. Lewinella sq23 in A-Floc systems), correlated with an ester- hydrolyzing ability, may play an essential role in the high efficiency of organics removal (Sun et al., 2018). Tenth, Fulvivirga sq38 is a particle bacterium in the A-Floc systems from the class

Page 84 of 208

Bacteroidia. Fulvivirga bacteria were found in sludge sample under both high and low aeration as potential anti-biofouling or anti-off flavour genera (Zhang et al., 2014b).

Finally, a high level of variability between the eight relative abundant classes and 36 high proportion OTUs was evidenced among the four rearing systems, with significant differences. Here, from only two taxa (Alpha- and Gammaproteobacteria), eight potential bacteria that are likely positively associated with shrimp weight gain rates may account for taxonomic groups that may serve as bioindicators of the FT system (e.g. Ruegeria sq 90, Pseudoocenicola sq78 and Halomonas sq 263); the RAS system (e.g. Roseovarius sq175, Catenovulum sq142 and Pseudoalteromonas sq91); the H-Floc system (e.g. Roseobacter sq81 and Ruegeria sq1); and moderately, the A-Floc system (with Pseudoocenicola sq78, which was also present in the FT system).

3.4.3 Environmental condition-selected bacteria

From a general perspective, the biological attributes of Ruegeria sq1 could be a stable population from the beginning to end of the experiment in H-Floc systems with a high proportion definitely depending on ecosystem drivers. The biological drivers can be stress tolerance (e.g. pH reduced to the lowest level, 7.4, in the H-Floc system; chapter 2; Furtado et al., 2015), adaptability and complex microbe-microbe interactions, such as between Ruegeria sq1 and Marivita sq3 in this study (Ray et al., 2010; Vadstein et al., 2018; Wang et al., 2018b). Hence, this again shows that the growth of bacteria within the systems is the most crucial factor in determining the stability of microbial communities (Shade et al., 2012; Li et al., 2017; Chapter 2). We can see that the shrimp-beneficial Bacillus sp. LT3, which was inoculated into the four systems at the beginning as a potential bio-inoculant (Defoirdt et al., 2011, Chapter 2), did not show permanent resistance with the expected high proportion at the end of the experiment. For this reason, it is challenging to identify the single bacterium that is characteristic for healthy or pathogenic to the rearing water and/or shrimp-host because its activity will ecosystem functioning or depend on the specific system (environmental conditions) designs (Defoirdt, 2016). Moreover, opportunistic pathogenic bacteria will only be harmful to cultured animals in extremely high number (e.g. ≥ 104 - 106 CFU mL-1; Tran, 2013; Joshi et al., 2014; Soto-Rodriguez et al., 2015), where the interplay of pathogenic and non-pathogenic bacteria play a role in the response to the disturbance Page 85 of 208

of the environment as ‘population - cell number’ not ‘clone - gene’ units, and this needs to be considered (Lemire et al., 2015).

3.4.4 Generation of a novel microbial ecological model for Litopenaeus vannamei aquaculture

For specific aims, such as to meet the high demand for Litopenaeus vannamei products, it would be ideal to have a new hybrid system combining the following presumptive potential bacteria as bio-inoculant candidates: (1) Ruegeria sq1 (pH 7.5 is optimum for growth) and Marivita sq3 (pH 7.0 - 7.5 is optimum for growth) from the H-Floc condition and (2) Catenovulum sq142 (pH 6.0 - 9.0, with pH 7-7.5 optimum for growth) and Pseudoalteromonas sq91 (pH 6-10, with pH 7-7.5 optimum for growth) from the RAS condition. To date, Pseudoalteromonas spp. were reported as candidates with very high quality and long resistance in a specific biofloc system (Hashim et al., 2018). In our H-Floc systems, Pseudoalteromonas sp. (e.g. Pseudoalteromonas sp. sq91) was also present from week 3 to the end of the experiment. The potential hybrid system will benefit greatly from the positive characteristics of the four presumptive strains towards a super-intensive shrimp culturing system and sustainable tropical marine water management. This means that we can design a prospective full-biofloc system (adding carbon everyday) instead of semi-biofloc systems like this study design (only adding carbon at every 3 days), without too much consideration of harmful TSS levels, in which excessive TSS amounts always have the potential to cause diseases outbreaks via opportunistic pathogens (Ray et al., 2010; Schveitzer et al., 2013; Avnimelech, 2015). Therefore, the shrimp productivity can be maximised in a biological sustainability because we can add more feed into the systems to greatly increase shrimp biomass without adding antibiotics and disinfectant solutions towards a good aquaculture practice. This potential hybrid aquaculture system as a microbial ecological model and many more fundamental insights can be expected in the coming years. Therefore, the long-term sustainability of the aquaculture industry may be ensured because the environmental effects must be minimised by measures like the flocculation process for aquaculture wastewater treatment. The next step is generating this microbial ecological H-Floc_RAS hybrid system by carrying further experiments for good recommendations for shrimp aquacultures.

Page 86 of 208

3.4.5 Further studies

Understanding the abundance change of certain bacterial taxa is important for shrimp aquaculture. However, the observed differences of relative abundances in the high-throughput techniques of this study may not highly accurately reflect the real taxon abundances. Therefore, we propose application the iHAAQ method for further quantitative bacterial ecology by combining qPCR (with similar bacterial V3V4 primers) with the high-throughput sequencing techniques from this study to generate more beneficial recommendations for bio-shrimp products, which are urgently needed.

CONCLUSIONS

This study explored the distribution of bacterial communities by applying Illumina Miseq V3 sequencing of 16S rRNA genes to different aquaculture culture environments (FT, RAS, H-Floc, A-Floc). The interaction among microbial species, for example, Ruegeria sq1 versus Marivita sq3, led to important findings in the ecology of shrimp culture, namely, that microbial communities are highly complex, nonlinear, evolving systems that can be disordered, and therefore, unpredictable. Thus, individual sequence OTUs with high proportions of significant differences among systems represent more important indicators than the relative class abundances do. Moreover, these taxonomic OTU bio-indicators should be investigated in depth with r/K-strategists’ study for potential beneficial bacteria under the ‘ecological theory point of view’, using culture-dependent technology with the cell count number for monitoring the benefits and harms different species cause for aquaculture.

TSS was the most influential factor shaping the differences in bacterial compositions and sequence OTU proportions among the four systems. Therefore, it should be taken into consideration that TSS is a useful bio-monitoring tool for bacterial ecology or biological threats. Based on our findings on TSS as an influential factor, we could see how purposeful the beginning quote is: ‘everything is everywhere, but, the environment selects’ (Wit and Bouvier, 2006).

Page 87 of 208

Free-living and particle bacteria were highly different in terms of the OTU number and composition between the two biofloc systems. They were also different in the sequence OTU bio-indicators. Therefore, this study provides useful baseline information for future studies that could monitor biofloc communities and correlations with shrimp productivity, elucidate parameters controlling the biofloc community structure and function, develop management tools to guide biofloc communities into their optimal composition and integrate desirable biofloc community composition into protocols for organic shrimp production with more research on immunity and disease resistance.

Acknowledgments This study was sponsored by the Vietnam Government, ZMT, and ARC. We want to thank the staff of ZMT and ARC for their help during running the experiment and Lab analysis and help with the bioinformatics analysis and CH for her help with the scripts in R analysis.

Page 88 of 208

Table 1| The table summarizes the characteristics of the culture systems and the shrimp performance parameters among the four intensive shrimp culture systems. FT is the flow-through system. RAS is the recirculating aquaculture system. H-Floc is the heterotrophic biofloc system. A-Floc is the autotrophic biofloc system. Term FT RAS H-Floc A-Floc Characteristics Solid remove No Sedimentation No No chamber Water exchange Yes No No No Primary producers Not specific Nitrobacters Heterotrophic Nitrobacters, (natural bacteria microalgae microbes) Shrimp growth Initial weight (0.21±0.10 g) parameters Final body weight (g) 2.11±0.19c 2.02±0.07a 2.40±0.21b 2.23±0.11d Protein efficiency ratio 38.78±11.89 28.49±12.28a 59.09±12.79b 34.70±20.71 (per day) Feed conversion rate (FCR) 1.16±0.36 2.06±1.52a 0.73±0.17b 1.59±0.71 a,b,c,dDifferent superscript in the same row indicates significant differences between treatments

Table 2| Contributions of categorical factors (systems, weeks) and environmental factors on the diversity of the rarefied OTUs among the four systems at three differential fractions (water column, free-living and particle samples) based on GLMMs and model selections according to the principal components (PCs).

Source of variations df F-ratio p-value

Water column Complete model (system + week) 1, 22 1,233.33 0.000 System 3, 11 8.28 0.004 Week 2, 22 2.28 0.126 Interaction 6, 22 6.74 0.000

Complete model (PC1 + PC5) 1, 28 600.60 0.000 PC1 1, 28 5.63 0.000 (TSS , pH, VSS,) PC5 1, 28 2.44 0.025 - (temperature, NO2 -N) Free-living Complete model (system + week) 1, 5 50.27 0.129 System 1, 5 0.13 0.001 Week 1, 5 2.17 0.201 Interaction 1, 5 4.08 0.731

Particle Complete model (system + week) 1, 5 142.68 0.201 System 1, 5 4.82 0.100 Week 1, 3 14.99 0.031 Interaction 1, 3 9.63 0.000

Page 89 of 208

Table 3| Contribution and significance of categorical (systems, weeks) and environmental factors to explaining the variation in community composition among the four systems at three differential fractions (water column, free-living and particle samples) based on redundancy analysis (RDA).

Source of variations Adjusted df Variance F-ratio p-value R² Water column Complete model (system + week) 0.38 5 3550.00 6.50 0.001

System 0.31 3 2643.00 8.07 0.001 Week 0.09 2 907.00 4.15 0.001

Complete model (PC1 + PC3 + PC4 + PC6) 0.22 4 7809.50 4.16 0.001 PC1 0.12 1 1036.30 7.51 0.001 PC3 0.02 1 313.90 2.28 0.033 PC4 0.05 1 504.40 3.66 0.012 PC6 0.04 1 438.40 3.18 0.019

Free-living Complete model (system + week) 0.26 2 1529.50 3.29 0.016 System 0.17 1 857.07 3.69 0.016 Week 0.12 1 672.45 2.89 0.016

Particle Complete model (system + week) 0.37 2 2154.30 4.24 0.031 System 0.34 1 1645.10 6.48 0.031 Week 0.06 1 483.49 1.90 0.031

Page 90 of 208

700

600

500

400 Number ofOTUs

300

200

100 Floc.FL Floc.FL Floc.FL Floc.FL Floc.PA Floc.PA Floc.PA Floc.PA Floc.PA Floc.PA Floc.PA Floc.PA − − − − − − − − − − − − Floc.Water Floc.Water Floc.Water Floc.Water Floc.Water Floc.Water Floc.Water Floc.Water Floc.Water Floc.Water − − − − − − − − − − Wo.FT.Water W1.FT.Water W2.FT.Water W3.FT.Water W4.FT.Water W2.A W4.A W1.A W2.A W3.A W4.A W2.H W4.H W1.H W2.H W3.H W4.H Wo.RAS.Water W1.RAS.Water W2.RAS.Water W3.RAS.Water W4.RAS.Water Wo.A Wo.H W1.A W2.A W3.A W4.A W1.H W2.H W3.H W4.H Figure 1| The box plots show the bacterial richness based on the rarefied OTUs among the four systems at a fraction of water column, and between the two biofloc systems at the fractions of free-living and particle samples. Flow-through systems: black, recirculating aquaculture systems: red, heterotrophic biofloc systems: green, autotrophic biofloc systems: blue.

Page 91 of 208

1.5 ● 2 A B

1.0 ● ●

1 NH3DO NO2 0.5 NH4 ●

NO3 0.0 ● 0 pH ● ●

NMDS2 TSS NMDS2 ● VSS −0.5 ● Salinity ● −1 ● −1.0 Temperature ● ● −1.5 −2 ●

−2 −1 01 −1.0 −0.5 0.0 0.5 1.0 1.5 2.0 NMDS1 NMDS1 Figure 2| NMDS: Bacterial community structure of the four shrimp culture system based on a Bray-Curtis dissimilarity matrix of the OTU data for relative sequence OTU abundance. Flow- through systems: black, recirculating aquaculture systems: red, heterotrophic biofloc systems: green, autotrophic biofloc systems: blue. A-left panel: water column samples-square; B-right panel: free-living samples-circle, and particle samples-triangle.

100

80

60 other Saccharimonadia Parcubacteria 40 Gammaproteobacteria Deltaproteobacteria Chlamydiae Bacteroidia 20 Alphaproteobacteria Actinobacteria Relative sequence abundance [%] sequence abundance Relative

0 Floc Wo Floc Floc Wo Floc Floc Wo Floc Floc Wo Floc Floc Wo Floc Wo Floc Wo Floc Floc W4 Floc W2 Floc W2 Floc W2 Floc W2 Floc W4 Floc W4 Floc W4 Floc W4 Floc W2 Floc W2 Floc W2 Floc W4 Floc W4 Floc W4 Floc W1 Floc W2 Floc W2 Floc W2 Floc W2 Floc W3 Floc W4 Floc W4 Floc W4 Floc Floc W1 Floc Floc W2 Floc Floc W2 Floc Floc W3 Floc Floc W4 Floc Floc W4 Floc Floc W4 Floc Floc W1 Floc Floc W2 Floc Floc W2 Floc Floc W2 Floc Floc W2 Floc Floc W3 Floc Floc W4 Floc Floc W4 Floc Floc W4 Floc Floc W4 Floc W1 Floc W2 Floc W2 Floc W2 Floc W3 Floc W4 Floc W4 Floc − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − − Water FT Wo Water FT Wo Water FT Wo Water FT Wo Water Water FT W1 Water FT W2 Water FT W2 Water FT W2 Water FT W2 Water FT W3 Water FT W4 Water FT W4 Water FT W4 Water FT W4 Water FL A FL A FL A FL A FL A FL A PA A PA A PA A PA A PA A PA A PA A FL H FL H FL H FL H FL H FL H FL H FL H FL PA H PA H PA H PA H PA H PA H PA H PA H PA H Water RAS Wo Water RAS Wo Water RAS Wo Water RAS Wo Water RAS W1 Water RAS W2 Water RAS W2 Water RAS W2 Water RAS W2 Water RAS W3 Water RAS W4 Water RAS W4 Water RAS W4 Water RAS W4 Water A Water A Water A Water Water A Water A Water A Water A Water A Water A Water A Water A Water H Water H Water H Water H Water H Water H Water H Water H Water H Water H Water H Water H Water H Water H Figure 3| The bar-plots display the differential sequence proportions of the most eight abundant classes among the four shrimp culture systems at three fractions of the water column, free-living and particle-attached samples. The other is all classified classes with a relative abundance of < 10% in at least one sample.

Page 92 of 208

02468 sqrt rel

Water FT Wo Water FT Wo Water FT Wo Water FT Wo Water FT W1 Water FT W2 Water FT W2 Water FT W2 Water FT W2 Water FT W3 Water FT W4 Water FT W4 Water FT W4 Water FT W4 Water RAS Wo Water RAS Wo Water RAS Wo Water RAS Wo Water RAS W1 Water RAS W2 Water RAS W2 Water RAS W2 Water RAS W2 Water RAS W3 Water RAS W4 Water RAS W4 Water RAS W4 Water RAS W4 Water H−Floc Wo Water H−Floc Wo Water H−Floc Wo Water H−Floc Wo Water H−Floc W1 Water H−Floc W2 Water H−Floc W2 Water H−Floc W2 Water H−Floc W2 Water H−Floc W3 Water H−Floc W4 Water H−Floc W4 Water H−Floc W4 Water H−Floc W4 Water A−Floc Wo Water A−Floc Wo Water A−Floc Wo Water A−Floc W1 Water A−Floc W2 Water A−Floc W2 Water A−Floc W2 Water A−Floc W3 Water A−Floc W4 Water A−Floc W4 Water A−Floc W4 FL H−Floc W2 FL H−Floc W2 FL H−Floc W2 FL H−Floc W2 FL H−Floc W4 FL H−Floc W4 FL H−Floc W4 FL H−Floc W4 FL A−Floc W2 FL A−Floc W2 FL A−Floc W2 FL A−Floc W4 FL A−Floc W4 FL A−Floc W4 PA H−Floc W1 PA H−Floc W2 PA H−Floc W2 PA H−Floc W2 PA H−Floc W2 PA H−Floc W3 PA H−Floc W4 PA H−Floc W4 PA H−Floc W4 PA A−Floc W1 PA A−Floc W2 PA A−Floc W2 PA A−Floc W3 PA A−Floc W4 PA A−Floc W4 PA A−Floc W4 Colwellia sq74 Colwellia ***Marivita sq3 Lewinella sq23 Lewinella Fulvivirga sq38 Fulvivirga Maritalea sq237 ***Ruegeria sq1 Parahaliea sq58 Parahaliea ***Ruegeria sq90 Hanstruepera sq5 Hanstruepera Halomonas sq263 Halomonas Roseobacter sq81 Roseobacter ***Leisingera sq72 ***Leisingera OM43 clade sq176 Roseovarius sq175 B2M28 (uncl.) sq70 (uncl.) B2M28 Catenovulum sq142 Catenovulum Aliiroseovarius sq63 Aliiroseovarius Mycobacterium sq32 Mycobacterium Aliiroseovarius sq137 Mycobacteriumsq253 Oceanospirillum sq104 Oceanospirillum Winogradskyella sq165 Winogradskyella Granulosicoccus sq278 Granulosicoccus NS3a marine sq9 group NS3a Pseudooceanicola sq78 Pseudooceanicola Oceaniserpentillasq115 Oceaniserpentillasq111 Marixanthomonas sq382 Marixanthomonas Saprospiraceae (uncl.) sq7 (uncl.) Saprospiraceae ***Pseudoalteromonas sq91 ***Pseudoalteromonas Chitinophagales (uncl.) sq212 (uncl.) Chitinophagales Gammaproteobacteria (uncl.) sq50 (uncl.) Gammaproteobacteria Gammaproteobacteria (uncl.) sq26 (uncl.) Gammaproteobacteria sq40 (uncl.) Gammaproteobacteria Candidatus Campbellbacteria (uncl.) sq132 (uncl.) Campbellbacteria Candidatus SAR324 clade(Marine group B) (uncl.) sq95 (uncl.) B)SAR324 clade(Marine group

Figure 4| Heatmap showing the differential OTU proportions (abundant OTUs at least 1% in one sample of microbial communities in the four shrimp culture systems at the fractions of the water column, free-living and particle samples based on Bray-Curtis distance. The statistical testing was only for water column samples, although the figure includes both free-living and particle samples. FT is the flow-through system. RAS is the recirculating aquaculture system. H-Floc is the heterotrophic biofloc system. A-Floc is the autotrophic biofloc system. Water is water column samples, FL is free-living samples, PA is particle samples.

Page 93 of 208

Granulosicoccus sq278

Oceaniser pentilla sq115

Oceaniser pentilla sq111

Oceanospir illum sq104 Halomonas sq263 Gammaproteobacteria (uncl.) sq40 Gammaproteobacteria (uncl.) sq26

Gammaproteobacteria (uncl.) sq50

Parahaliea sq58

OM43 clade sq176

B2M28 (uncl.) sq70

***Pseudoalteromonas sq91

Colwellia sq74 Catenovulum sq142

SAR324 clade(Mar ine group B) (uncl.) sq95 ***Ruegeria sq90 ***Ruegeria sq1 Roseovarius sq175 Roseobacter sq81

Pseudooceanicola sq78

***Marivita sq3

***Leisingera sq72

Aliiroseovarius sq63

Aliiroseo varius sq137

Maritalea sq237

Candidatus Campbellbacter ia (uncl.) sq132

Winogradskyella sq165

NS3a marine group sq9

Mar ixanthomonas sq382

Hanstr ueper a sq5

Fulvivirga sq38

Saprospiraceae (uncl.) sq7

Lewinella sq23

Mycobacterium sq32

Mycobacter ium sq253

Figure 5| The network shows the correlations between the differential OTU proportions and the shrimp weight gain rates among the four systems and between weeks at a fraction of water column samples. The size and colour intensity of the tip labels shows the strength of the correlation. Blue is the negative correlation, and red is the positive correlation.

Page 94 of 208

Chapter 4

Page 95 of 208

Chapter 4

Cultivation shrimp post-larvae challenged with a potential pathogenic Vibrio parahaemolyticus: Finding an open gateway to prevent shrimp farming disease

Loan Thanh Doan1,2*, Astrid Gärdes1, Germán A. Kopprio1, Andreas Kunzman1.

1Leibniz centre for Tropical Marine Research (ZMT), Fahrenheitstr.6, 28359 Bremen, Germany 2Vietnam National University of Agriculture, Trau Quy, Gia Lam, Hanoi, Vietnam

* Correspondence: Corresponding Author [email protected]; [email protected]

Keywords: Litopenaeus vanamei, challenge test, microbiota, aquaculture, Illumina sequencing

Running title: Vibrio disease and lethal concentration

ABSTRACT

In this study, a five day bacterial immersion challenge test was used to evaluate survivorship of

Litopenaeus vanamei that had been cultured four weeks (shrimpuse) in flow-through (FT), recirculation (RAS), heterotrophic biofloc (Hfloc), and autotrophic biofloc (Afloc) aquaculture systems. The experiment was comprehensively designed with eight conditions for challenge shrimp with two Vibrio isolates (Vibrio sp. LV11-yellow colony and Vibrio sp. LV21-green colony) from rearing water and one potentially AHPND-causing strain of Vibrio parahaemolyticus (pVpAHPND; AHPND: Acute Hepatopancreatic Necrosis Disease) to identify factors that could be protect shrimp from a pathogenic bacterium. Four rearing wateruse samples

(waterFT, waterRAS, waterHfloc, and waterAfloc) and one non-rearing watersource sample (seawater) were applied in the challenge test to monitor survival probability of the four reared shrimpuse groups (shrimpFT, shrimpRAS, shrimpHfloc, and shrimpAfloc) in a matrix of the challenge with three

Page 96 of 208

Vibrio strains. The Kaplan-Meier method was used to estimate shrimp survival curves. The Cox regression method was applied for hypothesis testing to determine factors that best protect shrimp from a toxic bacterium. The Illumina Miseq amplicon sequencing for 16S rRNA genes was used to examine bacterial community composition in shrimpuse. WaterRAS demonstrated the highest probability of protecting shrimp compared with the other three systems (Cox regression, hazard ratio = 0.40, p = 0.014). ShrimpHfloc showed a higher probability of survivorship, but not statistically significant differences among shrimpuse against pVpAHPND (Kaplan-Meier survival curves). Overall, shrimp from the Hfloc system with water from the RAS system tend to survive longer in the challenge test. Such differences in shrimp survival probability were also evidenced by different bacterial community patterns in shrimpuse, as well as some sequence dominant target taxa (Pseudoalteromonas sq91, Ruegeria sq1, and Marivita sq3), which could be proposed as bacterial presumptive candidates for further study at the species level to contribute to sustainable shrimp aquaculture protected from AHPND. Several quick diagnostic methods to apply at the farm level to help prevent AHPND are noted at the end of the paper.

4.1 INTRODUCTION

In 2016, farmed crustaceans contributed significantly to global aquaculture production with 8.3% share in quantity (live weight) and 23% share in value (FAO, 2018), of which shrimp and prawns accounted for 16.1% shared value; other crustaceans accounted for only 6.9% shared value. Intensive shrimp farming has succeeded in Hawaii, Florida, Texas of America, and other places in the world, where approximately 5-10 kg shrimp per cubic meter of water is produced in a three-month period. The most difficult aspect of crustacean farming is water quality control. Microorganisms within desirable levels for shrimp are often exposed to variable conditions, eventually becoming more susceptible to microbial infections, especially in their larval stage (Smith et al., 2003; Mohanty et al., 2018). Vadstein et al., (2018a) evidenced by the detrimental larvae-microbe interaction causing poor viability and quality in larval rearing. Many research questions have been proposed regarding the stress-strain tolerance of bacteria and its effects on productivity and survival probability of cultivated animals in the aquaculture industry. The answers to such questions based on ecological approaches will extend understanding of animal- bacterial interaction (e.g. with a toxic level causing diseases) have not yet been verified by

Page 97 of 208

aquaculturists and biologists (McFall-Ngai et al., 2013). The pathogenic Vibrio presents critical problems in shrimp farming because of the disease it causes (Ashrafudoulla et al., 2019). The Vibrionaceae family contains the major bacterial pathogens in marine aquaculture. Vibrios are highly motile with high metabolic flexibility, quickly increasing their population size when the environmental conditions are favourable (Stocker, 2012). Members of the genus Vibrio comprise approximately 14 recognized clades (Sawabe et al., 2013). Among the pathogenic Vibrio (e.g. V. harveyi, V. parahaemolyticus, V. vulnificus, V. alginolyticus, V. damsela, and V. penaeicida), Vibrio parahaemolyticus are causative agent of the most issue bacterial diseases in shrimp farming, including acute hepatopancreatic necrosis disease (AHPND), more commonly known as the early mortality syndrome (EMS), because shrimp regularly die within 30 days post stocking (Tran et al., 2013; Abdullah Sani et al., 2013; Zhang et al., 2014a). Vibrio infections occur in the presence of virulence factors (toxins; Zhou et al., 2012; Letchumanan et al., 2014). Toxins secreted by Vibrio parahaemolyticus into both culture media and cell-free broth can cause AHPND (Tran et al., 2013), and the plasmid-located toxin genes pirA and pirB were identified in AHPND-pathogenic strains of Vibrio parahaemolyticus (Han et al., 2015). However, sometimes the presence or absence of virulence genes depends on the differences in geographical regions, testing methodologies, and sample sources (Tran et al., 2013; Raghunath, 2015). Vibrio parahaemolyticus were reported with remarkable biofilm formation ability on the shrimp surface (Mizan et al., 2018; Ashrafudoulla et al., 2019). The biofilm formation ability of Vibrio parahaemolyticus increases the cell’s attachment ability to suspended particles (i.e. total suspended solids), as well as shellfish (Elexson et al., 2014). It is more dangerous that Vibrio parahaemolyticus showed positive amplification to pathogenic gene tdh and biofilm genes with strong genetic relationships (Ashrafudoulla et al., 2019). The temperatures ranging 25 - 37 0C were considered optimum for significant biofilm formation by Vibrio parahaemolyticus (Ahmed et al., 2018), while this temperature range is also optimal for shrimp growth (Mohanty et al., 2018).

Strategies to remedy EMS/AHPND were updated with VpAHPND genomic and proteomic studies, the effect of bacterial concentrations and temperature to the VpAHPND virulence, and toxins detection methods (Prachumwat et al., 2019). However, questions remain on the effects of other proteins on VpAHPND virulence and other bacteria interacting with VpAHPND. As the AHPND causative agent is a bacterium, using antibiotics to control the disease is possible, but Vibrio

Page 98 of 208

parahaemolyticus has proved resistant to a wide range of antibiotics and carry the markers of different virulence genes (Han et al., 2015; Ashrafudoulla et al., 2019); and Vibrio parahaemolyticus is a common inhabitant in shrimp farming environments (Tran et al., 2013). In another approach, biofloc has also been shown to have a positive impact on the survival, growth, and digestive enzyme activities of Litopenaeus vannamei (Xu et al., 2013; Zheng et al., 2019). A reduction in the abundance of Vibrio has been reported in biofloc systems (Souza et al., 2012). In a study about disease resistance, biofloc may enhance immune activity (Ekasari et al., 2014 ); in another study based on mRNA expression of six immune-related genes (ProPO1: prophenoloxidase1, ProPO2: prophenoloxidase 2, PPAE: prophenoloxidase activating enzyme 1, ran: ras-related nuclear protein, mas: masqueradelike serine proteinase, and SP1: serine proteinase1), biofloc had positive effects on both growth and immune gene expression (Kim et al., 2014; Soto-Alcalá et al., 2019). However, to support shrimp health and productivity, more research on the effects of biofloc is required in pursuit of biofloc-inoculants (Hashim et al., 2018).

Biofilm cells are more resistant to disinfectants and antibacterial agents than the same bacteria in a free-swimming cell state; survival, infectivity, and transmission are enhanced due to strong bacterial biofilm formation ability (Kadam et al., 2013). Hence, currently, there is no probiotic that can control EMS/AHPND Vibrio parahaemolyticus in shrimp culture. Moreover, the virulence of isolated VpAHPND (e.g. PV1-ARC: potentially AHPND-causing strain of Vibrio parahaemolyticus) varies for reasons yet unknown (Joshi et al., 2014). One suggested theory is that such variation is related to variation in VpAHPND toxic genes copy number (e.g. 7 to 121 copies per cell; Han et al., 2015). Currently, methods to control EMS/AHPND are under research; thus, the availability of standard challenge models to establish a fast, reliable, and systemic evaluation of possible treatments for EMS/AHPND are crucial.

In our knowledge, this is the first study to evaluate the survival probability of four weeks reared

Litopenaeus vanamei against pVpAHPND and rearing water Vibrio isolates with explanatory data combining shrimp survival hours and shrimp bacterial community composition. We hypothesized that an interaction existed between shrimpuse and wateruse in protecting shrimp

Page 99 of 208

during bacterial immersion challenge testing. The pVpAHPND virulence and predictability of shrimp mortality can be determined.

4.2. MATERIAL AND METHODS

4.2.1 Experimental design

Shrimp from the four systems (i.e. FT, RAS, Hfloc, Afloc, chapter 2) were harvested after four weeks of culturing for shrimpuse in the challenge tests. Shrimp from replicates (tanks) of each system were pooled into one tank with enough oxygen supply and without feeding for overnight incubation to get rid of all remaining faeces inside the shrimp body. Therefore, shrimpuse (1.44 ±

0.37 g) consisted of four groups (shrimpFT, shrimpRAS, shrimpHfloc, and shrimpAfloc). For each system, the rearing water (water column) was also pooled into one tank from different replicates to prepare for wateruse in the challenge tests. Therefore, wateruse also consisted of four groups

(waterFT, waterRAS, waterHfloc, and waterAfloc) and another was watersource (i.e. seawater used for setup of the systems and shrimp cultivation). The rearing water (wateruse) had differences in total -1 -1 suspended solids (TSS): waterFT (194.00 ± 22.89 mg L ), waterRAS (101.50 ± 18.57 mg L ), -1 -1 waterHfloc (333.50 ± 44.33 mg L ), and waterAfloc (194.00 ± 24.71 mg L ).

We used three bacterial strains: Vibrio sp. LV11 (yellow colony, chapter 2), Vibrio sp. LV21

(green colony, chapter 2), and pVpAHPND (green colony, PV1-ARC). Details on these strains can be seen in Supplementary Table 1A.

Shrimpuse were next individually transferred to the challenge rooms. In this room, shrimpuse were incubated 24 hours to eliminate moulting shrimp and shrimp that died due to stress before the challenge test. Therefore, the effect of faeces and feed were eliminated; thus, water quality difference with consideration of TSS and/or bacterial community were isolated as explanatory factors to evaluate shrimp mortality caused by bacterial virulence. Other stress and disturbances to shrimp were also eliminated following restricted regulations defined by the ARC guideline for challenge room usage (e.g. quiet and sterile conditions). Challenge room conditions were 12 hours dark and 12 hours light with a constant temperature at 28 ± 1 oC.

Shrimpuse were not fed during the challenge period according to results of the pilot challenge

Page 100 of 208

data (Supplementary Table 1B). Before transferring to the challenge room, each shrimpuse was divided into one sterilized transparent plastic box with a volume of 1.5 L containing 300 mL wateruse and watersource. Eight conditions were applied per each group of shrimpuse. Finally, we used a total of 480 boxes as the experimental design (Supplementary Figure 1), and one shrimpuse was stocked in one box. For each condition, 15 shrimp were maintained independently as individual replicates, for which the pair design of ‘shrimpuse vs. wateruse’ was only challenged with the pVpAHPND strain. The ‘shrimpuse vs. watersource’ pair was challenged with all three bacterial strains (Vibrio sp. LV11, Vibrio sp. LV21, and pVpAHPND) with one pair not challenged as a control design.

The bacterial immersion challenge tests for Litopenaeus vanamei post-larvae were conducted as described by Tran et al. (2013) with certain modifications. Based on results from the pilot 6 -1 challenge test, the effective dose for pVpAHPND was identified as 10 CFU mL for two weeks 7 -1 reared shrimp; we applied a higher dose of pVpAHPND at 10 CFU mL for the four weeks reared shrimp, as shrimpuse were getting bigger at this point (week 4). For the infection route, lethal bacterial doses were applied three times: the first dose was applied at the starting point of the five days immersion challenge, a second dose at 24 hours post immersion (h.p.i), and a third dose at the end of day 4 at double the amount of challenged bacterial concentration strains with the aim of eliminating all remaining live shrimp as recommended by the pilot data. The shrimp mortality data were recorded every six hours while exposed to the challenge strains.

4.2.2 Preparation of bacterial strains

All three bacterial strains (Vibrio sp. LV11, Vibrio sp. LV21, and pVpAHPND) were stored in marine broth with 20% glycerol at -80 oC for long term storage. Therefore, to prepare bacteria for challenge tests, frozen live bacterial strains were thawed and homogenized using a Vortex- genie2. Continuously, each strain was started with 50 μL of the stock to resuspend into 5 mL of marine broth (40.01 g L-1; Art.-Nr. 8952.3 Carl Roth GmbHt + Co.RG) to culture in 6 hours as the first bacterial batch. The second bacterial batch was the next continuous cultivation in 300 mL marine broth inoculated with 50 μL bacterial strains from the first bacterial batch for 18 hours (OD600 =1.044 for Vibrio sp. LV11, OD600 =1.158 for Vibrio sp. LV21, OD600 =1.304 for pVpAHPND). Finally, from the second bacterial batch, we used again 50 μL bacteria for inoculant

Page 101 of 208

with 300 mL marine broth to culture in 6 hours to get young bacteria for challenge tests. The bacterial culture conditions were 28 oC under constant shaking (150 rpm). The marine broth was autoclaved at 121 oC for 20 minutes and cooled to room temperature before use. We used the marine broth as blank for all measurements and for serial dilutions to get bacterial density approximately 107 CFU mL-1 for the challenge tests. All three bacterial strains were cultured in sterile conical flasks and autoclaved marine broth. Culture density of each strain was determined using the optical density measurements by a spectrophotometer (OD by Thermo Spectronic

Model 4001/4) at 600 nm (approximately 1.0 OD = 109 cells mL-1).

4.2.3 Data summary

Data for shrimp survival hours was divided into two data subsets. Statistical analysis was based on the first data subset. However, it is important to note that the second data subset evidenced that pVpAHPND was potentially a true pathogenic bacterium (Supplementary Table 1; Supplementary Figure 3).

The Illumina Miseq 16S rRNA data used the same amplicon library with sequencing data of wateruse (Chapter 3) for consistency in assessments of rearing water and reared shrimp for the challenge tests. However, the procedure to sample shrimp tissue for DNA extraction was different with rearing water (water column, Chapter 3). For the sterilized washing process, the shrimp larvae were caught on a sterile mesh. Shrimp were successively anaesthetized by immersion in a benzocaine solution for 10 seconds, disinfected in benzalkonium chloride solution for 10 seconds, and repeatedly rinsed (3x) with autoclaved seawater (source water) for 5 seconds. The shrimp larvae were aseptically transferred to a sterile plastic bag containing 10 mL of sterile seawater for homogenization. Shrimp tissue samples after homogenization were applied quick deep with liquid-N before being stored at -80 oC until DNA extraction took place. DNA extraction, processing of sequencing data, and bacterial alpha- and beta-diversity were examined with methods of Chapter 3 with minor differences (i.e. we used Inverse Simpson diversity index to compare the bacterial OTU richness between shrimp samples before and after the challenge test). In a total of 25 samples with no replicates for the hypothesis test; therefore, this sequencing data helped in explaining the results of hypotheses test with descriptive analysis.

Page 102 of 208

4.2.4 Statistical analysis of shrimp mortality

Whereas the Kaplan-Meier method with the log-rank test is useful for comparing survival curves in two or more groups, Cox regression (or proportional hazards regression) allows analysing the effect of several risk factors on survival hours. It identifies stratification variables with interaction function (Cox, 1972; Therneau and Grambsch, 2000; Goel et al., 2010; Zhang et al., 2018).

For descriptive and visualization data analysis, we used the Kaplan_Meier method; Surv() function was used to define the time-to-event outcome, and the survfit() function was used to estimate survival curves. Kaplan-Meier is a useful method to measure the alive subject (e.g. the survival of shrimp) in a certain amount of time after infection with a bacteria (Goel et al., 2010).

In shrimp challenge tests, the effect of the virulent factor caused by pVpAHPND was assessed by measuring the number of shrimps survived after the introduction of pVpAHPND over a period of time.

For hypothesis tests in shrimp survival probability, we used the Cox regression model to examine which factor best protected shrimp from a toxic Vibrio. The Chi-squared statistic tests the relationship between time and all covariates (i.e. shrimpuse and wateruse) in the model. For each covariate, the cox.zph() function correlates the corresponding set of scaled Schoenfeld residuals with time, to test for independence between residuals and time (survminer package; version 0.4.0.99). The Cox model assumption on the influent of observations was fulfilled as can be seen by the Supplementary Figure 2. Additionally, the Cox model performs a global test for the capacity of the model in predictable survival probability (Zhang et al., 2018).

All statistical analyses were performed in R (R version 3.5.2, R Core Team, 2018, using R studio version1.1.456) with a number of identical R packages.

4.3 RESULTS

4.3.1 Survival results

According to Figure 1, survival curves showed the two pairs of ‘shrimpRAS vs. waterHfloc’ (hours, median = 42, p ≤ 0.001) and ‘shrimpHfloc vs. waterRAS’ (hours, median = 42, p ≤ 0.001) had

Page 103 of 208

higher median survival times in comparison with other shrimpuse vs. wateruse pairs. Therefore, we infer that shrimpHfloc and shrimpRAS had a greater chance of survival probability, and waterRAS might better protect all shrimpuse against the pVpAHPND compared to other wateruse (hours, medianmin = 30, p ≤ 0.001). Vice versa, the Supplementary Figure 3 indicated that there was no difference in shrimpuse in using of seawater for the challenge test (p = 0.13); and pVpAHPND showed a strong virulent effect on shrimp survival (i.e. shrimpuse died rapidly within 12-18 h.p.i with 50% of survival probability).

Table 1 shows the results of the hypothesis tests designed to identify factors jointly impacting survival as a multivariable analysis. Chi-Square value of the wateruse was highly significant (Cox, chi-square = 20.41, p ≤ 0.001), thereby indicating that at least some of the independent variables among the four water differences (waterFT, waterRAS, waterHfloc, and waterAfloc) were significantly related to longer shrimp survival hours. Shrimpuse showed no significant difference in survival hours. However, Chi-Square of the interaction between wateruse vs. shrimpuse was highly significant (Cox, chi-square = 24.65, p = 0.003). As a result, wateruse was identified as an important determining factor in the shrimp challenge test. Therefore, it should be considered statistically significant in predicting hazard related to the effects of shrimpuse. Thus, we can find which shrimpuse likely showed a higher survival hour as a consequence of interaction. Further analysis from Table 1 is presented in Table 2. We found no differences between pairs: waterFT and waterAfloc and waterRAS and waterHfloc. Differences in protecting shrimp challenged with pVpAHPND were found between four pairs: (1) waterFT and waterRAS, (2) waterFT and waterHfloc,

(3) waterRAS and waterAfloc, (4) waterHfloc and waterAfloc.

Table 3 presents results of further analysis. The waterRAS at this analysis step showed significant difference compared to the other three wateruse in protecting shrimp survival longer against pVpAHPND (Cox, model with interaction, p = 0.014). Moreover, the hazard ratio for shrimpFT was

1.19 (> 1); therefore, shrimpFT has an increased risk of death as compared to shrimpAfloc.

ShrimpHfloc had the lowest risk of death among the four shrimp groups. Combining findings from

Figure 1 and Table 1, we found that shrimpHfloc likely showed higher survival hours or a lower hazard ratio among shrimpuse. Therefore, we focused on finding the differences of shrimpuse per each application of wateruse in the challenge against pVpAHPND.

Page 104 of 208

Table 4 shows results of subsequent analysis steps following the data presented in Table 3. At this stage, we observed differences of shrimpFT from both shrimpHfloc and shrimpAfloc in survival hours when we used waterFT. When we used waterRAS, we observed no differences among shrimpuse. Table 3 shows the lowest risk for waterRAS (Cox, hazard ratio = 0.40, p = 0.014) among wateruse; therefore, waterRAS can best protect ‘all’ shrimpuse with no differences in hazard probability. When we used waterHfloc, we observed three differences in shrimpuse, including

‘shrimpFT vs. shrimpRAS’ (p = 0.044), ‘shrimpRAS vs. shrimpHfloc’ (p = 0.009), ‘shrimpRAS vs. shrimpAfloc’ (p = 0.009). When we used waterAfloc, we found no differences among shrimpuse, but, as shown in Table 3, waterAfloc had the highest risk. Thus, waterAfloc had highest risk in protecting shrimp against pVpAHPND.

In summary, based on pair comparison, waterRAS was significantly better in protecting all shrimpuse against pVpAHPND with higher survival hours and the lowest hazard ratio. ShrimpHfloc likely showed higher survival hours or lower hazard ratio among shrimpuse. As an additional interpretation for supporting information, when evaluating the hazard probability by using the

Cox model without interaction, as shown in Table 3, both waterRAS and waterHfloc is significantly different from other wateruse. WaterHfloc was also effective in protecting shrimp survival, and this seems in line with the results of Kaplan-Meier curves. However, there was an interaction between wateruse and shrimpuse; therefore, we found the shrimpHfloc survival was better in waterRAS with higher h.p.i against pVpAHPND (h.p.i for shrimpHfloc = 42) and lowest hazard ratio (0.79, Table 3).

4.3.2 Microbial distribution in shrimp tissue

A total of 5,553 unique operational taxonomic units (OTUs) generated from the Illumina MiSeq data for diversity analysis of the bacterial community. Regarding taxonomic levels, there were 25 phylum, 55 classes, 153 orders, 260 families, and 555 genera for all shrimp tissue samples.

As illustrated in Figure 2, according to the Inverse Simpson diversity index, for the fraction before the challenge samples, shrimp from week 0 and week 2 showed a different trend in alpha diversity in comparison with all four shrimpuse from week 4. Of those, shrimpAfloc marked with the highest diversity, and shrimpRAS had the lowest diversity (at week 4, just before the challenge

Page 105 of 208

test). For the fraction after the challenge samples, all four shrimpuse vs. pVpAHPND had lower diversity in comparison with shrimpuse vs. other challenged conditions (i.e. shrimpuse vs. Vibrio sp. LV11, shrimpuse vs. Vibrio sp. LV21, and shrimpuse vs. no challenge bacterial strains).

As shown in Figure 3, the relative sequence abundance of OTUs was analysed for shrimp tissue samples between two fractions of before and after the challenge test using NMDS (ordination analysis, not hypothesis test). The figure shows a clearly separated pattern between before and after challenge samples. Notably, for the fraction of after the challenge, the differences of bacterial composition showed clearly between ‘shrimpuse vs. pVpAHPND’ compared with three remaining shrimpuse groups (i.e. ‘shrimpuse vs. Vibrio sp. LV11,’ ‘shrimpuse vs. Vibrio sp.

LV21,’ and ‘shrimpuse vs. no challenge bacterial strains’).

Figure 4 and Supplementary Table 2, for the fraction before the challenge, at week 0- shrimp before stocking were identified with three families, including Bacillales_Family XII (with 45.08% relative abundance), Vibrionaceae (40.11%), and Carnobacteriaceae (13.23 %). At week 2, of the four shrimpuse (shrimpFT, shrimpRAS, shrimpHfloc, and shrimpAfloc), shrimpRAS marked with the highest relative abundance of Shewanellaceae (70.07 %) but the lowest abundance of Vibrionaceae (29.23 %), although all shrimpuse had only relative abundance of the Vibrionaceae and Shewanellaceae families. The relative percentage of abundance of other bacteria was less than 5%. At week 4, the four shrimpuse groups showed much difference in bacterial distribution with three bacterial groups such as Rhodobacteraceae (shrimpRAS with

51.01%; shrimpHfloc with 47.13%), Flavobacteriaceae, and Gammaproteobacteria_unclassified. In general, Bacillales_Family XII bacteria had no presence in shrimp tissue after stocking; Vibrionaceae and Shewanellaceae had a critical reduction in relative abundance after week 2; Rhodobacteraceae, Flavobacteriaceae, and Gammaproteobacteria_unclassified were observed in shrimp tissue at week 4. In summary, bacterial distribution in shrimp tissue showed much difference and increased diversity from stocking to the end of cultivation period at week 4

(shrimpuse for challenge test).

Figure 4 and Supplementary Table 2, for the fraction after the challenge, show notable results for the pair ‘shrimpuse vs. pVpAHPND,’ which had a lower bacterial relative abundance group number, while the Vibrionaceae were dominant at approximately 70% in all four shrimpuse. In other pairs (‘shrimpuse vs. Vibrio sp. LV11,’ ‘shrimpuse vs. Vibrio sp. LV21,’ and ‘shrimpuse vs. Page 106 of 208

no challenged bacterium’) the bacteria with relative abundance were Flavobacteriaceae, Rhodobacteraceae, and Rubritaleaceae.

In summary, the bacterial composition distribution in shrimp tissue was not clearly different among all four shrimpuse. The differences in bacterial community in shrimp tissue were only significant according to cultivation time (shrimp age/ shrimp size).

4.3.3 Target taxa for Litopenaeus vannamei aquaculture

As shown in Figure 5, based on the results of chapter 3, we examined some bacterial target taxa that might better help explain shrimp survival probability against pVpAHPND from the challenge test. First, Ruegeria sq1, shrimpHfloc showed higher relative abundance of this bacteria at week 4

(just before the challenge), and in all shrimpHfloc after the challenge with different conditions.

Second, Pseudoalteromonas sq91, shrimpRAS had a higher relative abundance OTU at week 4

(before the challenge). Third, Marivita sq3, shrimpHfloc showed especially higher relative abundance OTU before the challenge at week 4, although the proportion of sequence smaller than 1%, but much difference among the shrimpuse. Fourth, Pseudooceanicola sq78 showed only high relative abundance at week 0 (before stocking) in shrimp tissue, but none after stocking and/or after the challenge test in comparision among shrimpuse according to times although the percentage of relative abundance was very low. As a summary, shrimpHfloc showed Ruegeria sq1 at a higher relative abundance both before (> 5%) and after the challenge test (> 5%); shrimpRAS had Pseudoalteromonas sq91 higher relative abundance only before the challenge (> 5%).

4.4 DISCUSSION

4.4.1 Strain-specific bacterial concentration as lethal dose

According to the pilot data, only pVpAHPND caused shrimp post-larvae (0.70 ± 0.07 g, shrimp weight at week 2, chapter 2) to died gradually after six h.p.i at the dose of 106 CFU mL-1. The other isolated strains from rearing water, Vibrio sp. LV11 and Vibrio sp. LV21 showed no risk to shrimp in the pilot or main challenge tests. Moreover, shrimp exposed to 2 such Vibrio isolates could stay alive even until 15 days in starvation conditions, although, they looked gradually smaller in size (stunted animal). In another study testing a different infection method (injection), Page 107 of 208

some Vibrio isolates were not able to cause shrimp mortality even at high doses from 105-107 cells per g of shrimp body weight (Flegel et al., 2005). Shrimpuse for the main challenge test (1.44 ± 0.37 g, at week 4, chapter 2) were also not affected by the two Vibrio isolates (i.e. Vibrio sp. LV11 and Vibrio sp. LV21), while the test of ‘shrimpuse vs. pVpAHPND’ showed statistically significant differences in survival probability at the dose of 107 CFU mL-1 upon the infection route of this study. Therefore, based on our data, the virulence factor could be explained by different bacterial strain (Joshi et al., 2014; Soto-Rodriguez et al., 2015) and by different shrimp size, the smaller shrimp with the pilot data, the bigger shrimp with the main data (Aguirre- Guzman et al., 2004). The older shrimp may develop increased resistance to disease (Aguirre- Guzman et al., 2004); the bacterial community for older shrimp at week 4 changed much more in diversity in shrimp tissue after stocking compared to the begining of cultivation period. Literatures indicated that some bacterial strains caused 100% mortality within 24 h.p.i, whereas other strains did not cause 100% mortality until 96 h.p.i (Joshi et al., 2014). In our experiment, the mortality rate was high within 48 hours after immersion; it then gradually reduced or stayed level until the end of the experiment. Therefore, we applied the last dose at day 4 (96 h.p.i) for the infection route to eliminate all remaining live shrimpuse. Based on this study, the minimal lethal dose applied for immersion challenge tests on L. vannamei could start at the dose of 106 CFU mL-1 for shrimp 2 weeks after stocking. However, to safeguard any challenge analysis against all shrimp dying immediately after immersion, the pilot challenge test should be done 4 -1 before the main test upon a specific experimental design with a starting dose of 10 CFU mL (certain threshold or quorum). This starting dose is the minimum needed to reach a certain bacterial density/biofilm formation to express pathogenicity resulting in toxin production, based on the premise that almost all Vibrio pathogenic bacteria (e.g. V. parahaemolyticus) use the quorum-sensing system by ‘molecular language’ and thereby tend to maintain an infective density above 104 CFU mL-1 (Joshi et al., 2014; Gomez-Gil et al., 2014). Vibrio virulence factors include various enzymes (e.g. proteases, lipases, chitinases, serine proteases, metalloproteases), siderophores and proteinaceous toxins (Zhou et al., 2012); and hemolysin occurrence (Gomez- Gil et al., 2014). These virulence factors are important in biofilm forming, which is regulated by quorum sensing (Zhong et al., 2006, Ashrafudoulla et al., 2019).

Originally, it was thought that a specific strain of V. parahaemolyticus caused AHPND. However, it has by now become clear that the causative agent of AHPND is not a single strain.

Page 108 of 208

There are even other Vibrio species (V. harveyi and V. punensis) found to carry the AHPND toxic genes (Restrepo et al., 2018). That finding suggested the virulent plasmid might be spreading to non-Vibrio parahaemolyticus bacteria (anecdotal data from Ikuo Hirono, [email protected]; Prachumwat et al., 2019). Therefore, the toxic gene could be carried by many different Vibrio clades, a phenomenon that will more seriously impact Vibrio pathogenesis. Moreover, the copy number of virulent plasmids is different in different strains.

Some VpAHPND have more than 30 copies, but some toxic genes have 1 or 2 copies (Han et al., 2015). Added to those considerations is that every species has a different number of (toxic) genes; however, there is no relationship between the copy number of plasmid and virulence of

VpAHPND (Han et al., 2015). The high virulent strains secreted toxin proteins. The toxic genes exist with transposition, which is a common transposon in Vibrio species. During the culture of AHPND Vibrio parahaemolyticus, the toxic genes might be removed entirely or partly from plasmid (Madigan et al., 2006).

Primarily, as stated by Tran et al. (2013), toxins secreted by V. parahaemolyticus into both culture media and cell-free broth, for example extracellular products (Supplementary Figure 4), can cause AHPND, as the disease has a toxin-mediated aetiology. Hence, in other ways, we can argue that a reduced or an increased virulence can depend on the media in which pathogenic bacteria are resuspended, causing different lethal doses. Our data gave evidence that wateruse showed significant differences in protecting shrimp from pVpAHPND and, therefore, was the most important factor in the prediction of hazards caused by pVpAHPND.

We kept some factors constant that significantly affect the lethal dose of a pathogenic bacterium, such as temperature (Prachumwat et al., 2019) and oxygen supply (i.e. we did not supply feed because feed increases the uptake of water and reduces dissolved oxygen), as shrimp tend to die from lack of oxygen more than from the toxic bacteria (Ebeling et al., 2006). Therefore, the different wateruse qualities were likely caused by TSS differences. The different TSS refers to differences in microbial proportions and/or bacterial populations (Chapter 3). Consequently, the interaction between microbes might cause a strain-specific lethal dose of bacterial pathogen.

4.4.2 TSS: a vital and challenging factor

The shrimp-culture systems comprise three key components: environment, host, and microbiota. The microbiota consists of both pathogenic and beneficial microorganisms. In Page 109 of 208

nature, Vibrio has a wide range of niche specializations, such as free-living populations, attached bacteria to biotic and abiotic surfaces, symbionts or pathogens, and biofilm populations (Reen et al., 2006; Chaweepack et al., 2015; Yatip et al., 2018; Soowannayan et al., 2019b). It is estimated over 95% of bacteria existing in nature are in biofilms (Overman, 2006); over 80% of animal-bacterial diseases are known to associate with bacterial biofilms (Hall et al., 2014). Moreover, the short generation time of Vibrio facilitates colonization where large numbers of cells can dominate the host in a short period of time as one of the pioneer groups (r-strategists), giving the bacterium an advantage that allows it to outcompete other bacteria (Huang et al., 2009; Defoirdt, 2016). However, our bioinformatics data evidenced that Vibrio bacteria had to be entirely changed by other bacterial groups four weeks after stocking, although Vibrio bacteria accounted for 40.11% as a second relative dominance group in shrimp tissue before stocking (at week 0). Just before the main challenge test - at week 4 - shrimpuse was relative abundant with three differential bacterial groups: Flavobacteriaceae, Rhodobacteraceae in shrimpRAS and shrimpHfloc; and Gammaproteobacteria_unclassified in shrimpFT and shrimpAfloc.

Biofilm theory could help to explain the complex relationship of environment-microbiota-host. The extracellular material derived both from the cells themselves and from the environment form a slime layer (as a fraction of biofilms, Supplementary Figure 4), creating a surface-attached community of bacterial cells. As production of bioflocs, biofilm formation requires many bacteria and different types of bacteria working together (Madigan et al., 2006). Bacteria are communicate well each other using molecular signalling systems (e.g. AHL, molecular language interacting by both Bacillus and Vibrio; Defoird et al., 2011). The successful communications with enough signals will change in gene regulation. The bacteria now produce pili, which use to attach each other or attach themselves to a substrate making flocs. The pili could be seen clearly in a study by Richards et al., (2017) about mechanisms for Pseudoalteromonas piscicida-induced killing of Vibrios and other bacterial pathogens using scanning electron micrographs. Afterwards, the bacteria secrete a slimy extracellular matrix of proteins, polysaccharides, and nucleic acids in the biofilms (Wesseling et al., 2015). Therefore, biofilm is a complex assembly of bacterial colonies, polymers and organic matters. Ultimately, biofilm may become depleted and degraded because of lacking nutrient and oxygen over time. Thus, disease onset can be facilitated by the presence of nonvirulent strains (i.e. biofilm strains) or a new form of the polymicrobial disease in which nonpathogenic strains contribute to increasing mortality (Lemire

Page 110 of 208

et al., 2015). Therefore, suspended particle (i.e. flocs or TSS) evidenced the interaction between microbes, and thus, the microbes-microbes relationship becomes non-linear. The relationship clearly depends on favourable chemical conditions; subsequently, chemical-physical environment changes will take place after interaction. Therefore, it is more difficult to apply antibiotic treatment to eradicate (pathogenic) bacteria in biofilms, generally requiring higher concentrations than needed for treatment of free bacteria to be effective (Olson et al., 2002), and more difficult to determine the lethal dose for a specific remedy for application to control pathogens.

Our data clearly showed that the post larval shrimpuse was not significantly different in survival hours and therefore was not an important predictor of hazard caused by pVpAHPND in the shrimp challenge test (based on both the pilot and main challenge tests). Hence, survival probability variations in the challenge test occurred not only due to the virulence of the pVpAHPND used, but likely also due to variations in experimental conditions (environment). As our data shows, wateruse proved to be the most important predictor of shrimp survival probability. We suggest possible reasons why differences in wateruse affected shrimp survival probability. First, using watersource for the challenge test, we found no differences in survival hours among shrimpuse groups (Kaplan-Meier, p = 0.13). In this environmental condition (watersource), the pVpAHPND showed the strongest toxic level with 50% mortality within 12 to 18 h.p.i (Supplementary Figure

3). Second, using waterRAS, we observed that pVpAHPND had difficulty reaching a population with bacterial density high enough to cause mortality (disease); waterRAS best protected all shrimpuse against pVpAHPND and showed the lowest hazard ratio. Nevertheless, how do watersource and waterRAS affect shrimp survival hours differently? It may be that waterRAS is significantly characterized by Pseudoalteromonas sq91, Catenovulum sq143, Roseovarius 175

(bacteriaenvironment, chapter 3), and shrimpRAS (host) at week 4 and was characterized with the highest relative sequence abundance of Pseudoalteromonas sq91 (bacteriahost), which might be highly potential good bacterium with a number of species belonging to this genus were reported as good bacterial probiotic (Wang et al., 2018a). Third, without feed supply, the organic matters in wateruse were presented by TSS differences. Therefore, lower opportunities existed for pVpAHPND and for opportunistic r-strategists to invade higher density populations in waterRAS -1 with the lowest TSS amount (waterRAS, at week four, 101.50 ± 18.57 mg L ; Chapter 2). Vice -1 - versa, waterHfloc with the highest TSS (waterHfloc, TSS at week four, 333.50 ± 44.33 mg L mg L

Page 111 of 208

1 ; Chapter 2) likely protected shrimpuse well (Kaplan-Meier curves). Moreover, the lower the TSS, the lower shrimp productivity, and the higher the TSS, the higher shrimp productivity (Chapter 3), although there was a recommended threshold of TSS for shrimp performance (Mohanty et al., 2018).

It was likely that TSS was vital and also challenged factors in the challenge tests as well as in the shrimp culture systems since, at the larval stage, shrimp showed no apparent difference in the bacterial community among the four reared shrimpuse, although the rearing water of shrimp culture systems showed differences in bacterial community composition and water quality due to the differences in TSS (chapters 2 and 3). Therefore, TSS dictates the differences in microbial community and water quality in wateruse, which was the most important factor in predictable shrimp mortality.

4.4.3 Shrimp post larval immune system: An important note

Shrimp (invertebrate animals) lack an adaptive immune system. Despite that, a so-called innate immunity response was developed with a rich amount of innate immune genes and innate defence molecules that respond to invading foreign pathogens (Subramanian et al., 2014), including lectins, protease inhibitors, antimicrobial peptides, and haemocytes. And, the innate molecules together create the innate immune system, protecting shrimp from potential bacterial pathogens (Hoebe et al., 2004; Chen and He, 2019). The defence mechanism in shrimp thereby depends mainly on a non-specific or innate immune response to fight infectious diseases (Kurtz and Franz, 2003). In a specific vaccine study, the authors try to find the receptors of AHPND toxins for selection breeding of AHPND resistant family, that is, the FKC vaccine, which worked very well for larger shrimp (over 5 g); however, the vaccine did not work for small shrimp. The immune system of small shrimp is not well developed, while, AHPND occurred approximately 30 days after stocking, when shrimp usually weigh less than 5 g (anecdotal data from Ikuo Hirono, [email protected]). Thus, immunization therapy may not be useful in preventing and controlling shrimp AHPND disease when they are of small size. However, innate immune response mechanisms and/or environmental adaptation mechanisms may help control shrimp disease based on ecological control theory applied to aquaculture water (Vadstein et al., 2018a, Chen and He, 2019, Chapter 2). The environmental adaptation mechanism in a shrimp system could, via biological and non-biological factors, regulate each other. Indeed, physio-chemical

Page 112 of 208

factors dictate the growth of various microorganismal populations, while microorganismal populations affect non-biological aquatic factors, such as molecular ammonia nitrogen, dissolve oxygen, pH, and TSS (TSS includes: feed, microbial biomass, other organic matters; AHPHA, 2005, chapter 2). Our challenge data also showed a probability that shrimp could be recovered after 42 h.p.i with the toxic pVpAHPND. That evidence from the pilot challenge informed our decision to apply the last lethal dose at day 4. While using watersource, there was a 50% probability that shrimp died before 18 h.p.i, which is a shorter time survival time. The differences in hours post-immersion (42 h.p.i vs. 18 h.p.i) evidenced that the innate immunity in shrimp may take a less important role; vice versa, the wateruse composition played a more critical role in the challenge test. Hence, adaptation, that is, the recovery of shrimp during the challenge test to the environmental stress (e.g. toxin proteins secreted by pVpAHPND) was probably explained by another complex interaction relationship between the innate immune host’s response and microbes in shrimp tissue (gut) and in rearing water. As we argued before (i.e. we did not provide feed during the challenge test, no nutrient source), when nutrients in the biofilm (home of bacterial interaction) deplete, the biofilm may begin to degrade; thus, no more toxins are released by bacteria, explaining why shrimp did not experience 100% mortality if they were still alive after 42 hours.

4.4.4 Bio-inoculants as an open gateway for control Vibrio disease

Bacterial biology and ecology have identified candidates for microbial bio-indicators of ecosystem drivers (chapter 3). Measuring the abundance of some key populations is important for understanding the contributions of the microbial community in rearing water and shrimp tissue, since each bacterial strain belongs to a distinct ecological population (Wit and Bouvier, 2006; Shade et al., 2012).

We observed different patterns in microbial community in wateruse and shrimpuse, providing evidence of the complex relationship between host response and microbes. For example, waterRAS experienced relative sequence abundance with Pseudoalteromonas sq91 (chapter 3), while shrimpHfloc experienced relative abundance with Pseudoalteromonas sq91 (Gammaproteobacteria), Ruegeria sq1, and Marivita sq3 (Rhodobacteraceae bacteria).

Therefore, the pair of ‘waterRAS vs. shrimpHfloc’ in the challenge test had the higher probability of survival against pVpAHPND because of the many members of genus Pseudoalteromonas, Ruegeria

Page 113 of 208

and Marivita reported as beneficial bacterium (Miura et al., 2018; Wang et al., 2018a; Cai et al., 2018; Su et al., 2019). Moreover, the bacterium can be termed as probiotic in aquaculture practices only if their functions are demonstrated as beneficial to the growth and health of cultured shrimp (Fuller, 1989) and survival of the bacteria in the new surrounding and in competition with native bacteria (Ninawe and Selvin, 2009; Skjermo et al., 2015). Some functions of such presumptive bacteria are yet unknown. We highly recommend that a beneficial bacterium is needed to stabilize resistance and resilience in both rearing water and colonized shrimp gut. Ruegeria sq1 and Marivita sq3 are presumptive bacteria belonging to the Rhodobacteraceae family, which provide key examples of biofilm with many significant important roles in marine water (Buchan et al., 2005; Huang et al., 2009; Elifantz et al., 2013). For example, regarding strain-stress tolerance, that is, Ruegeria sq1 stabilized populations in Hfloc rearing water from before stocking to the end of shrimp cultivation at week 4. It also showed relative sequence abundance in shrimp tissue before the challenge test (week 4). Therefore, we would select Ruegeria sq1 for further study on stress-strain behavior as well as other beneficial functions it might have. Additionally, Pseudooceanicola sq78 is also a member of Rhodobacteraceae, but that bacterium was no longer resistant in shrimp tissues after stocking; for this, Pseudooceanicola sq78 is not highly suggested for microbial model study since they had less chance to be colonized in the shrimp gut according to this study design. For another example, the Bacillales_Family XII bacteria was relatively abundant (over 40%) before shrimp stocking; however, this bacterial group showed no resistance or resilience for loading bacteria into the host gut in shrimp tissue. Hence, they play no role against the toxins secreted by pVpAHPND, though many members of Bacillus can use AHL signalling directly from pathogenic Vibrio to lower their quorum-sensing threshold, for example, Bacillus sp. LT3 (Defoidt et al., 2011). Therefore, directions for future research should focus on Pseudoalteromonas sq91, Ruegeria sq1, and Marivita sq3 as presumptive species recommendations for potential bio- inoculants (Hashim et al., 2018) to combat opportunistic bacteria (r-strategists). This approach on bio-inoculants is becoming increasingly important for further development of more sustainable aquaculture practices, that is, they can dominate pathogenic bacteria in the rearing water when host shrimp are young (and possibly not able to eat microbial flocs); and stimulate or trigger an innate immune response when host shrimp are older (and can eat microbial flocs - biofilm products) (Chen and He, 2019). This approach is especially useful for Hfloc systems

Page 114 of 208

designed for higher bio-shrimp productivity without disease outbreaks. It is important to note that problems caused by r-bacteria, that is, pVpAHPND, should be dealt with effectively by other antagonist r-bacteria since pVpAHPND worked quickly in causing mortality within six hours according to this study data. Moreover, nonpathogenic r-strategists in matured water could be used as inoculants for the pre-colonization of intake water (Defoirdt, 2016), although finding methods to identify appropriate r-strategists is still challenging, and also microbial management based on slow growth to favour K-strategists still has effects (Vadstein et al., 2018b).

Wang et al., (2018a) reported on two strains of Pseudoalteromonas spp., namely CDM8 (likely

P. flavipulchra) and CDA22 (P. piscicida): 1) they actively antagonised VpAHPND; 2) they significantly reduced presumptive Vibrio counts, decreased cumulative mortality, and decreased Vp copy numbers of pirA ; 3) they were proposed as anti-VpAHPND probiotic candidates; 4) they were identified as potential biological control agents; 5) the strains, especially the pigmented Pseudoalteromonas species, usually produce bioactive compounds with antibacterial, antifouling and antibiofilm activities, while the nonpigmented tend to have broader environmental tolerances than pigmented ones; 6) if using the co-existed design niche, CDM8 and CDA22 might contribute to offset their probiotic effects (Aranda et al., 2012; Richards et al., 2017; Wang et al., 2018a). Future studies remain on interactions among microbes to determine potential applications of potential probiotic-inoculants in aquaculture and food safety.

At week 2 in shrimp cultivation (the sensitive period of disease outbreak) in shrimp cultivation,

Shewanellaceae bacteria were relatively dominant in shrimpRAS with 70.07% and in three other shrimpuse with approximately 40%. Many members of this family are reported as epibionts, in symbiotic associations with invertebrate hosts, the most common free-living state, and as pathogens (Ivanova et al., 2004; Hau and Gralnick, 2007; Janda and Abbott, 2014; Hong et al., 2017). The ecological functions of this bacterial group are converting heavy metals and toxic substances to the least harmful products (Hau and Gralnick, 2007). Shewanellaceae also showed significant functions in biogeochemical cycles of C, N with the ability to reduce nitrate to nitrite (Brettar et al., 2001; Ivanova et al., 2004). In another study about AHPND Vibrio, an isolated genus Shewanella from the Shewanellaceae family had an interaction causing shrimp mortality as enhancement of VpAHPND virulence, although the Shewanella bacteria themselves have not yet been confirmed to directly affect shrimp mortality (Prachumwat et al., 2019). The genome of a

Page 115 of 208

symbiotic Shewanellaceae bacterium isolated from shrimp might offer insight into molecular strategies for niche-specific lifestyles, and for a better understanding of any existing virulence mechanisms. However, this study evidenced that Shewanellaceae was replaced entirely by other bacterial groups at week 4 of the shrimp cultivation period. In summary, problems caused by bacteria can also be controlled by bacteria to achieve the goal of low Vibrio count and/or replace Vibrio population by other bacterial populations in the rearing water and the shrimp gut as an alternative to antibiotic treatment in shrimp aquaculture, although further investigation is needed, such as on the consequence of interaction among microbes (Vadstein et al., 2018a) and even the interaction between two probiotic bacteria (Wang et al., 2018a).

4.4.5 Diagnostic methods

Monitoring of species harmful and beneficial to aquaculture is equally important according to this challenge data (and chapter 2, chapter 3). The proper diagnostic methods were incorporated with the recognition of pure exotoxins, mutants, bacteriophages, plasmids, and molecular tools help reveal the shrimp pathology process. Such methods include 1) PCR: nested-PCR detection AP3, AP4, and AP5 (Dangtip et al., 2015); 2) LAMP: loop-mediated isothermal amplification for detection of VpAHPND (Arunrut et al., 2016); 3) PAS: printed array-strip, a new technology that can detect max 12 pathogens in one time by PCR for shrimp diseases (Ikuo Hirono, [email protected]); and 4) GeoChips: version-1 was developed in 2004 with 763 gene variants involved in nitrogen cycling, methane oxidation, and sulphite reduction (Rhee et al., 2004). Version-2 was launched in 2007 with 24,243 oligonucleotide probes, targeting approximately 10,000 functional gene variants in the geochemical cycling of carbon, nitrogen, phosphate-P, sulphate reduction, metal reduction and resistance, and organic contaminant degradation (He et al., 2007). Version-3 was provided in 2010 with 28,000 probes and targeted about 57,000 sequences from 292 gene families (He et al., 2010). Version-4 was displaced and improved from 2010 to 2012 with 84,000 probes, targeting more than 152,000 genes from 410 functional families with additional genes from bacteriophage that are involved in stress response and virulence (He et al., 2012). Therefore, the Geochip is a powerful tool for specific, sensitive, and quantitative analysis of microbial communities from variable habitats (He et al., 2013). However, at farm level, disease (AHPND disease) always occurs very quickly, sometimes in hours; thus, we proposed several ‘quick’ diagnostic methods based on this study and also based on ecological control theory. First, we propose suc-Vibrio/suc+Vibrio ≤ 1, since the green Vibrio Page 116 of 208

population (suc-Vibrio: green Vibrio on TCBS agar plate; TCBS: Thiosulfate Citrate Bile Sucrose Agar; suc+Vibrio: yellow Vibrio on TCBS agar plate) can shift rapidly with environmental conditions (Huang et al., 2009). From an assessment of the impact of milkfish (Chanos chanos) farming on potentially pathogenic Vibrios in coastal maricultural environments of the Lingayen Gulf (Philippines), the suc-Vibrio/suc+Vibrio ratio was larger than 1, the peak value of suc-Vibrio in the gut of the fish was 545 million cells per g wet weight. The authors reported suc-Vibrio outnumbered suc+Vibrio as a thread threshold for the contaminated environment (Reichardt et al., 2013). In another study (chapter 2), suc-Vibrio/suc+Vibrio in rearing water at week 0 was smaller than one (CFU g-1, 282.50/ 376.50 = 0.75), at week 2 and week 4 the suc-Vibrio was dominant over suc+Vibrio, and the suc-Vibrio was the highest number of cell count at week 2. Therefore, we need a quick regular check of suc-/suc+Vibrios with population equilibrium for rearing water; Second, we propose Vibrio/Bacillus ≤ 1 for water and shrimp tissue; Third, pH fluctuated less than 0.3 per day is recommended since change in pH will dictate the change in microbial populations and other non-biofactors (Liu et al., 2015; Hou et al., 2017). Fourth, the following kits may be used: API-20E (Biomerieux) for quantifying Vibrio number (Sanjuán et al., 2009), and Compact Dry Nissuni to check pathogenic Vibrio (Nissui pharmaceutical Co., LTD).

4.4.6 Further studies and applications

In the long term, and as a contribution to global food security, effective disease therapies must be developed through an improved and detailed understanding of immune pathways and associated immune surveillance mechanisms within specific host groups vs. specific pathogenic bacterium. Such therapies will lead to an understanding of shrimp disease pathways while assisting the development of novel processes for Vibrio disease control.

We should further study the implementation of some natural products with as bioactive compounds for effective disease therapies towards shrimp culturing without using chemicals (for disinfection): 1) ethanolic extract of ginger inhibits biofilm formation by a Vibrio parahaemolyticus isolate that causes AHPND in shrimp; and that use of the extract in shrimp feed can protect shrimp from AHPND mortality with a survival of 30 - 40% higher than that of shrimp given un-supplemented feed (Soowannayan et al., 2019a); and 6-gingerol was found to be the bioactive compound that interacts with the quorum sensing receptor LasR (Kim et al.,

Page 117 of 208

2015); 2) green tea as a natural antimicrobial agent to inactivate Vibrio parahaemolyticus (Xi et al., 2012); 3) Japanese fermented soybean product by Bacillus subtilis var Natto BSN1 can help inhibit biofilm of Vibrio (Yatip et al., 2018).

Based on the results of this study and chapter 3, we suggest the development a compatible kit (molecular diagnostic tools for beneficial bacteria, that is, may be Ruegeria sq1, Marivita sq3, and Pseudoalteromonas sq91) for the specific RAS and Hfloc shrimp aquaculture systems. That kit would be a key factor in determining the RAS and Hfloc microbial community stability as a biocontrol strategy for shrimp aquaculture health and productivity via approaches of stress-strain resistance and resilience and/or microbial bio-indicators for ecosystem drivers.

CONCLUSIONS

It is impossible to eradicate Vibrios from culture systems totally; therefore, we should focus on ‘join them,’ not ‘beat them’ approaches in managing microbial activity (De Schryver and Vadstein, 2014; Vadstein et al., 2018a). The aim is to lower the population densities of native opportunistic bacteria and a higher number of antagonistic and/or beneficial bacteria populations. This approach is called bacterial activity management via establishing ecological niche differentiation to maximize beneficial effects and minimize adverse effects by manipulating target bacterial species. This means managing single pathogenic dominance of r-bacterium by other bio-inoculant bacteria because beneficial K-bacteria cannot help with immediate impact to prevent the 50% probability for shrimp mortality occurring within six hours, according to this study. The difference in survival rate different water use (with different TSS) showed the important role of antagonistic r-matured bacteria against pVpAHPND under the theory of biofilm formation and quorum sensing interaction. These limited data are expected to be useful to build strong evidence in future research through increasing the number of shrimp replicates for challenge tests, and shrimp replicates for DNA extractions. However, the strong point of this study was the comprehensive coverage of bioinformatics relationships, virulence effects, and as well as cultured shrimp conditions in a challenge test. This study recommends bacterial study community in the different stages of shrimp aquaculture (post-larvae, juvenile, adult) for necessary information to study about shrimp vaccine applied to specific ecological niches (e.g.

Page 118 of 208

Hfloc, and/or RAS) in the shrimp aquaculture industry.

For a specific Litopenaeus vannamei challenge test, the provision of feed during the immersion challenge is not necessary but needed an accurate shrimp incubation to eliminate molting shrimp and died shrimp before the test. The results of this study would open the gateway of further research on identification of possible alternative treatments, such as the use of probiotics and immunostimulants and application of microbially matured water systems (with matured r- bacteria) in preventing of EMS/AHPND disease outbreaks in shrimp farming.

Acknowledgments

We would like to thank the staff of the ARC and ZMT. We would also like to thank Dr Yufeng Niu (ARC scientific staff; [email protected]) for his comments on the experimental design. We thank Dr Christiane Hassenrück (ZMT) for her comments and R-scripts on sequencing data. Especially we would like to thank Prof. Peter Bossier for his instruction in the experimental design.

Page 119 of 208

Table 1| The table shows the analysis of deviance from the Cox regression model. wateruse include waterFT, waterRAS, waterHfloc and waterAfloc; and shrimpuse include shrimpFT, shrimpRAS, shrimpHfloc and shrimpAfloc from the four systems (FT: flow- through system, RAS: recirculating system, Hfloc: heterotrophic biofloc system, Afloc: autotrophic biofloc system)

Log-Likelihood Chi-Square df p-value Wateruse (A) - 1068.80 20.41 3 < 0.001 Shrimpuse (B) - 1065.90 5.77 3 0.123 Interaction (A : B) - 1053.60 24.65 9 0.003

Table 2| The table shows p-value from Pairwise comparisons using Log-Rank test and

Benjamini-Hochberg adjustment method. WaterAfloc is a reference to compare from the first to the last wateruse. waterFT presents the flow-through rearing water. waterRAS presents the recirculating rearing water. waterHfloc presents the heterotrophic biofloc rearing water. waterAfloc presents the autotrophic biofloc rearing water.

Wateruse waterFT waterRAS waterHfloc waterAfloc waterFT - reference waterRAS 0.001 - reference waterHfloc 0.018 0.662 - reference waterAfloc 0.662 < 0.001 0.010 reference

Page 120 of 208

Table 3| The table shows the forest from two Cox models proportional hazards. waterFT presents the flow-through rearing water. waterRAS presents the recirculating rearing water. waterHfloc presents the heterotrophic biofloc rearing water. waterAfloc presents the autotrophic biofloc rearing water. shrimpFT presents the shrimp cultivated from the flow-through systems. shrimpRAS presents the shrimp cultivated from the recirculating systems. shrimpHfloc presents the shrimp cultivated from the heterotrophic biofloc systems. shrimpAfloc presents the shrimp cultivated from the autotrophic biofloc systems. waterAfloc and shrimpAfloc were used as the reference in comparision among different shrimp and water. AIC is the Akaike's Information Criterion.

Model with interaction Model with no interation Explanatory factors Global p_value: 8.8004e-06 Global p_value: 0.00020627 AIC: 2137.19 AIC: 2143.85 Concordance Index: 0.68 Concordance Index: 0.62 Hazard ratio p_value Hazard ratio p_value

Wateruse reference reference waterAfloc (n = 60) waterFT 0.67 (0.33 – 1.38) 0.278 0.98 (0.68 – 1.40) 0.892 (n = 60) ** waterHfloc 0.73 (0.36 – 1.51) 0.397 0.55 (0.38 – 0.79) 0.001 (n = 60) * *** waterRAS 0.40 (0.19 – 0.83) 0.014 0.51 (0.36 – 0.73) < 0.001 (n = 60)

Shrimp reference reference use shrimpAfloc (n = 60) shrimpFT 1.19 (0.58 – 2.46) 0.636 1.41 (0.98 – 2.03) 0.066 (n = 60) shrimpHfloc 0.79 (0.38 – 1.61) 0.513 0.99 (0.69 – 1.42) 0.973 (n = 60) shrimpRAS 0.97 (0.47 – 1.99) 0.926 0.93 (0.64 – 1.35) 0.706 (n = 60)

Page 121 of 208

Table 4| The effect of shrimpuse per each level of wateruse. Table values are p-value from Pairwise comparisons using Log-Rank test and Benjamini-Hochberg adjustment method.

ShrimpAfloc is reference to compare from the first to the last shrimpuse. waterFT presents the flow-through rearing water. waterRAS presents the recirculating rearing water. waterHfloc presents the heterotrophic biofloc rearing water. waterAfloc presents the autotrophic biofloc rearing water. shrimpFT presents the shrimp cultivated from the flow-through systems. shrimpRAS presents the shrimp cultivated from the recirculating systems. shrimpHfloc presents the shrimp cultivated from the heterotrophic biofloc systems. shrimpAfloc presents the shrimp cultivated from the autotrophic biofloc systems.

Wateruse Shrimpuse shrimpFT shrimpRAS shrimpHfloc shrimpAfloc waterFT shrimpFT - - - reference shrimpRAS 0.327 - - reference shrimpHfloc 0.010 0.141 - reference shrimpAfloc 0.009 0.206 0.669 reference waterRAS shrimpFT - - - reference shrimpRAS 0.850 - - reference shrimpHfloc 0.510 0.510 - reference shrimpAfloc 0.510 0.510 0.850 reference waterHfloc shrimpFT - - - reference shrimpRAS 0.044 - - reference shrimpHfloc 0.150 0.009 - reference shrimpAfloc 0.715 0.039 0.715 reference waterAfloc shrimpFT - - - reference shrimpRAS 0.740 - - reference shrimpHfloc 0.740 0.740 - reference shrimpAfloc 0.740 0.740 0.740 reference

Page 122 of 208

int=Afloc.water.Afloc int=Afloc.water.FT int=Afloc.water.Hfloc int=Afloc.water.RAS int=FT.water.Afloc int=FT.water.FT int=FT.water.Hfloc int=FT.water.RAS Strata int=Hfloc.water.Afloc int=Hfloc.water.FT int=Hfloc.water.Hfloc int=Hfloc.water.RAS int=RAS.water.Afloc int=RAS.water.FT int=RAS.water.Hfloc int=RAS.water.RAS

1.00

0.75

0.50 Survival probability

0.25

p < 0.0001

0.00

0 6 12 18 24 30 36 42 48 54 60 66 72 78 84 90 96 102 108 114 120 Time median hours (0.95LCL - 0.95UCL) shrimpFT shrimpRAS shrimpHfloc shrimpAfloc waterFT 18 (12 - 18) 18 (12 - 36) 24 (18 - 48) 24 (18 - 42) waterRAS 30 (18 - 42) 30 (24 - 42) 42 (24 - NA) 36 (30 - NA) waterHfloc 30 (30 - NA) 42 (36 - 72) 12 ( 6 -30) 30 (6 - 48) waterAfloc 18 (12 - 18) 18 (12 - 36) 30 (18 - 36) 18 (12 - 36)

Figure 1| The shrimp survival curves were based on the Kaplan-Meier method for shrimp from the four systems challenge with the four water differences. waterFT presents the flow-through rearing water. waterRAS presents the recirculating rearing water. waterHfloc presents the heterotrophic biofloc rearing water. waterAfloc presents the autotrophic biofloc rearing water. shrimpFT presents the shrimp cultivated from the flow-through systems. shrimpRAS presents the shrimp cultivated from the recirculating systems. shrimpHfloc presents the shrimp cultivated from the heterotrophic biofloc systems. shrimpAfloc presents the shrimp cultivated from the autotrophic biofloc systems. LCL: lower control limit. UCL: Upper control limit.

Page 123 of 208

Relative sequence abundance (%) Before challenge

week 4

After challenge with pVpAHPND

week 2 week 0

Before challenge test After challenge test

Figure 2| Barplot shows bacterial alpha diversity based on Inverse Simpson diversity index for shrimp tissues samples before and after the challenge test. FT is flow-through system. RAS is recirculating system.

H-Floc is heterotrophic biofloc system. A-Floc is autotrophic biofloc system. W0, W2, and W4 are weeks 0, 2 and 4 in the shrimp culture period. LV11 is Vibrio sp. LV11, yellow colony. LV21 is Vibrio sp. LV21, green colony. VP+ is potentially AHPND-causing strain of Vibrio parahaemolyticus.

Page 124 of 208

week 0

week 4 NMDS2 challenged with

pVpAHPND

0.5 0.0week 0.5 2 1.0 1.5 2.0 − challenged with LV11, LV21 and no challenge Vibrio strain 1.0 − 1.5 − −3 −2 −1 01

NMDS1

Figure 3| Non-metric multidimensional scaling (NMDS) plot shows the bacterial beta diversity among challenge conditions and shrimpuse according to pre- and post-stocking time points, and after the challenge test. LV11 is Vibrio sp. LV11, yellow colony. LV21 is Vibrio sp. LV21, green colony. pVpAHPND is potentially AHPND-causing strain of Vibrio parahaemolyticus. The colours presents for shrimp samples from the four systems: Black presents the flow- through system, red presents the recirculating system, green presents the heterotrophic biofloc system, blue present autotrophic biofloc system, grey presents the shrimp samples pre-stocking (therefore, they were pooled together). W0, W2, and W4 are weeks 0, 2 and 4 in the shrimp culture period.

Page 125 of 208

Figure 4| Barplot shows the relative sequence proportional OTU abundance in samples among challenge conditions and shrimpuse according to stocking time points. FT is flow-through system. RAS is recirculating system. HFloc is heterotrophic biofloc system. AFloc is autotrophic biofloc system. W0, W2, and W4 are weeks 0, 2 and 4 in the shrimp culture period. LV11 is Vibrio sp. LV11, yellow colony. LV21 is Vibrio sp. LV21, green colony. pPV+ is potentially AHPND-causing strain of Vibrio parahaemolyticus. pPV- presents no challenge Vibrio strain condition.

Page 126 of 208

Sequence proportions larger than 1% Sequence proportions smaller than 1%

before 0.20

8 A: Ruegeria sq1 B: Marivita sq3

0.15

6 before + pPV 0.10 + pPV 4

0.05 2

0 0.00 FT FT Floc Floc Floc Floc Floc Floc Floc Floc Floc Floc Floc Floc Floc Floc Floc Floc Floc Floc Floc Floc Floc Floc Floc Floc − − RAS RAS − − − − − − − − − − − − − − − − − − − − − − − − − − w4 FT w4 FT A H A H − − − − w4 RAS w4 w4 RAS w4 w4 A w4 H w4 A w4 H FT water w2_T4 FT FT water FT water w2_T4FT water FT source water VP source water source water VP+ FT source water source water VP+ FT source water source water VP source water w0 pooled 4 systems source water LV11 FT LV11 source water source water LV21 FT LV21 source water w0 pooled 4 systems source water LV21 FT LV21 source water source water LV11 FT LV11 source water source water VP source water source water VP+ RAS VP+ water source source water VP+ RAS source water source water VP source water RAS water w2_T6 RAS RAS water RAS water w2_T6 RAS RAS water source water LV11 RAS LV11 source water source water LV21 RAS LV21 source water source water LV21 RAS source LV21 water source water LV11 RAS source LV11 water source water VP source water source water VP+ A source water source water VP source water source water VP+ H source water source water VP+ A source water source water VP source water source water VP+ H source water source water VP source water source water LV11 A LV11 source water source water LV21 A LV21 source water source water LV11 H LV11 source water source water LV21 H LV21 source water source water LV21 A LV21 source water source water LV11 A LV11 source water source water LV21 H LV21 source water source water LV11 H LV11 source water Floc water w2_T14 A Floc water Floc water H w2_T12 Floc water Floc water w2_T14 A Floc water Floc water H w2_T12 Floc water − − − − A A H H

C: Pseudoalteromonas sq91 b 6 e 0.30 f D: Pseudooceanicola sq78 5 o 0.25 r e 4 0.20 before 3 0.15

2 0.10

1 0.05 + pPV before 0 0.00 FT FT Floc Floc Floc Floc Floc Floc Floc Floc Floc Floc Floc Floc Floc Floc Floc Floc Floc Floc Floc Floc Floc Floc Floc Floc − − RAS RAS − − − − − − − − − − − − − − − − − − − − − − − − − − w4 FT w4 A w4 FT w4 H A H − − − w4 RAS w4 − w4 RAS w4 w4 A w4 H w4 A w4 H FT w2_T4 water FT FT w2_T4 water FT source water VP+ source FT water source water VP water source source water VP water source source water VP+ FT VP+ water source w0 pooled 4 systems w0 pooled source water LV21 FT LV21 source water source water LV11 FT LV11 source water pooled 4 systems w0 source water LV21 FT LV21 source water source water LV11 FT LV11 source water source water VP+ RAS source water source water VP source water RAS w2_T6water RAS source water VP source water source water VP+ RAS source water RAS water w2_T6 RAS RAS water source water LV21 RAS LV21 source water source water LV11 RAS LV11 source water source water LV21 RAS LV21 source water source water LV11 RAS LV11 source water source water VP+ A source water source water VP source water source water VP+ H source water source water VP source water source water VP+ A VP+ source water source water VP source water source water VP source water source water VP+ H source water source water LV21 A LV21 source water source water LV11 A LV11 source water source water LV21 H LV21 source water source water LV11 H LV11 source water source water LV21 A LV21 source water source water LV11 A LV11 source water source water LV21 H LV21 source water source water LV11 H LV11 source water Floc water w2_T14 A w2_T14 water Floc Floc water w2_T12 H water Floc Floc water w2_T14Floc water A Floc water w2_T12 H − − − − A H A H

Figure 5| Barplot shows the proportion of sequence target taxa in differential shrimp tissue samples from before and after the challenge test. FT is flow-through system. RAS is recirculating system. H-Floc is heterotrophic biofloc system. A-Floc is autotrophic biofloc system. W0, W2, and W4 are weeks 0, 2 and 4 in the shrimp culture period. LV11 is Vibrio sp. LV11, yellow colony. LV21 is Vibrio sp. LV21, green colony. pPV+ is potentially AHPND-causing strain of Vibrio parahaemolyticus. pPV- presents no challenge Vibrio strain condition.

Page 127 of 208

Chapter 5

Page 128 of 208

Chapter 5

General discussion and concluding remarks

5.1 TSS and pH indicate a system’s biosecurity

In the rearing water column of shrimp culture systems, organic matter contains feed and metabolic waste, which can be split into solids and dissolved organic matter. This thesis focussed on studying the total suspended solids (TSS) as an indicator of organic matter (APHA, 2005). In addition, in the intensive shrimp culture, TSS included both setting solids and volatile suspended solids (VSS); VSS were calculated from the weight loss incurred after burning the TSS at 550°C for 30 minutes. As a result, VSS represents the microorganism biomass (e.g. VSSalgae, VSSbacteria; Ebeling et al., 2006). The major part of TSS in this study was bioflocs of bacteria, since the experiment used artificial light for the 12-h:12-h light-dark cycles. According to this condition of light (i.e. not sunlight), the algae’s optimal growth was not promoted, and thus, the VSSalgae probably represented a limited production. Bioflocs are suspended aggregates or suspended growth of organic matter, including living organisms like protozoa, rotifers and nematodes (Hargreaves, 2006; Najdegerami et al., 2016; Martínez-Córdova et al., 2017). Aggregates can adhere to the shrimp shell; affect the oxygen exchange and feed fish (i.e. aquaculture animals) pathogens (Liltved and Cripps, 1999), as well as consuming dissolved oxygen in the rearing water. According to Lam et al., (2015), there are many functions and methods of controlling the accumulation of organic matter in a preferred range, and shrimp farmers are aware of these approaches. However, this thesis was conducted based on an ecological theory point of view; therefore, there was a lack of removal of the wastes or solids until the end of all the experiments for a precise evaluation of the effects or contributions of TSS in the shrimp culture systems.

In chapter 2, TSS differences were evident, showing the advantages of heterotrophic bacteria with a higher TSS amount in the heterotrophic biofloc system (H-Floc) as a consequence of promoting heterotrophic bacteria by adding extra carbon (Ebeling et al., 2006; Avnimelech, 2015). The higher amount of TSS was associated with a higher amount of shrimp weight and highest proportions of Bacillus spp., r-bacteria and K-bacteria, as well as maintaining good water - quality (NO2 -N concentration decreased from week 3 onwards). Therefore, BFT through the H-

Page 129 of 208

Floc system in intensive shrimp culture showed numerous advantages based on promoting heterotrophic bacteria (Huang et al., 2017). Many authors have reported that organic matter influences microbial communities (Austen et al., 2002; Schauer et al., 2010; Bienhold et al., 2012; Dang and Lovell, 2016; Learman et al., 2016); however, the thresholds at which this occurs are still in proposed ongoing research questions. Thus, in chapter 3, TSS was evidenced as a primary factor influencing bacterial communities. According to the experimental designs, TSS explained major portion of the total 38% of the variation in bacterial community 2 compositions among the four systems (RDA, PC1, R TSS = 0.43, p = 0.001). TSS were directly affecting the bacterial OTU richness and shaping the structure of the rearing water microbial community. However, TSS, VSS and pH had a closed relationship with each other according to the statistical model shown, and they all had equal roles in determining the variation of the microbial community matrix. Therefore, it can be useful to monitor the TSS and/or pH of the water column to quickly check the stability of the microbial community (Mikkelsen et al., 1996; Ligi et al., 2014; Hou et al., 2017) and avoid harmful TSS levels (Ebeling et al. 2006; Van Wyk, 2006). In chapter 4, TSS took a vital role in the challenge test of the shrimp cultivated for 4 weeks with a potentially pathogenic Vibrio. TSS were proposed as an explanatory factor in the differences in survival probability, which were interpreted through the quorum-sensing communication among microbes in the substrates (i.e. TSS) of the biofilm. TSS differences evidenced the possibility of bio-inoculants as an open gateway to controlling the invasion of Vibrio populations. For this concern, TSS was understood as an ecological niche of the abundance of heterotrophic bacteria in the H-Floc systems. In contrast, the heterotrophic bacteria abundance resulted in TSS differences among the four shrimp culture systems (i.e. flow-through, recirculating, heterotrophic biofloc and autotrophic biofloc systems). Such differences in TSS and heterotrophic bacteria among the four systems gave evidence of the abundance of some fundamental bacterial populations belonging to a distinct ecological population (Wit and Bouvier, 2006; Shade et al., 2012). It was resistance and resilience of the presumptive ruegerial bacteria (e.g. Ruegeria sq1) and marivital bacteria (e.g. Marivita sq3) to disturbances in the H- Floc systems that allowed them to reach the level of a stable population. For example, Ruegeria sq1 was resistant to lower pH, while Marivita sq3 was resilient to low pH and the invasion of bacteria in free-living conditions (Litchman, 2010; Shade et al., 2012). However, Marivita sq3 did not exhibit high relative sequence abundance in the host shrimp tissue (< 1%), while

Page 130 of 208

Ruegeria sq1 (likely a particle-attached bacterium) was relatively dominant in the host shrimp (gut) at week 4 in the H-Floc systems, with > 5% of sequence abundance. Both Ruegeria sq1 and Marivita sq3 are presumptive heterotrophic bacteria (Huang et al., 2017); thus, the results evidenced the advantages of these bacteria in higher TSS level conditions.

For the security of shrimp culture systems, we could use solely TSS or pH parameters, or better, combine the TSS and pH parameters in regular monitoring to predict the other harmfull indices, - such as NO2 -N concentration, NH3-N concentration, microbial populations, shrimp weight gain rates (chapter 2, chapter 3) and shrimp survival when facing a pathogenic Vibrio (chapter 4). We could see that good water quality is essential, but it is not the only factor that affects shrimp performance and production.

© Bregnballe, (2015)

+ Figure 1: Equilibrium between un-ionized ammonia (NH3-N) and ammonium (NH4 -N) at 20°C (Bregnballe, 2015).

Page 131 of 208

As shown in Figure 1, toxic ammonia is absent at a pH level close to 7, but it emerges quickly as the pH is increased. Hence, for the optimum water temperature of around 30°C for shrimp culture, pH levels between 7 and 8 have an acceptable nitrification rate, while as chapter 2 concluded, at around pH 7.4, a high rate of bacterial nitrification should occur (Pacheco-Vega et al., 2018). If the temperature increases, the pH increases as a consequence (UNESCO, 1983;

Timmons et al., 2002), thereby resulting in an increasing amount of free ammonia (NH3-N), which will enhance the toxicity level and have a negative effect on the shrimp host. Thus, to create a balance between two different scenarios of the suitable pH (high or low), a threshold of around 7.4 has been proposed. This was established from the results of this thesis and other studies (Bregnballe, 2015; Pacheco-Vega et al., 2018). Therefore, in line with the threshold of pH 7.4, the TSS should be kept in the range delineated in chapter 2 (≤ 426.50 ± 65.24 mg L–1 in the H-Floc systems).

Changes in pH determine the stability of the bioflocs present in the ponds (Mikkelsen et al., 1996). In general, pH is not a generating parameter in the systems; thus, it will change when + temperature and NH4 -N concentration change (Timmons et al., 2002). Nevertheless, pH is easy to check quickly; therefore, it is a fast approach for predicting other factors that need to be controlled, such as the TSS through the feed supply, feeding regime, solids removal, C/N ratio and other additives. Consequently, pH fluctuation must be limited to the optimal range for the cultured animals to avoid mortality and the dysfunction of the culture system. Thus, pH balance is an indicator of a cultured stability system, and pH fluctuations should occur in a range of ~0.3 pH units (Hinners et al., 2015).

5.2 Eco-culture system: The improvement of semi heterotrophic biofloc implementation in shrimp aquaculture

The aim of biofloc implementation is promoting the built-up efficiency of the microbial biomass; thus, start-up inoculants by aggregative bacterial species can be selected, resulting in improved floc formation with good quality and taking less time to form flocs. For this reason, culture conditions like the C/N ratio, temperature and ventilation speed in the system are essential factors for the activity of these bacteria (Salehizadeh and Shojasadati, 2001).

Page 132 of 208

flocs was not setting well (not yet matured flocs) pink-flocs in the heterotrophic bioflocs

Figure 2: Biofloc performance parameters evaluated by sieving according to different sizes (i.e. 50, 250, 500, 1000 μm; the sample without sieving is the original size). An Imhoff cone was used to measure the setting solids (APHA, 2005; Avnimelech, 2015), and the pink colour flocs 50-μm size was determined at week 3 in the heterotrophic biofloc system.

The two 50 and 250 μm floc sizes of setting solids (SS) evidenced the bio-indexes for the biofloc systems. In the first index, the 250-μm size of SS was larger than the original size of SS. This indicated the flocs had not yet matured, since the flocs were still young in the biofilm process; therefore, they had not yet stuck together well, and they were not able to set well to the bottom of the Imhoff cone. For the second index, the 50-μm pink-floc size’s SS appeared in the heterotrophic biofloc system, parallel to the isolates (i.e. Pseudoalteromonas sp. P13; chapter 2) were identified in the 50-μm pink flocs. The presumptive pseudoalteromonas bacterium (i.e. Pseudoalteromonas. sq1) was the representative candidate of the recirculating systems (RAS) Page 133 of 208

and likely particle bacterial candidate in the H-Floc systems (chapter 3). Further, in the host shrimp tissue, presumptive Pseudoalteromonas sp. was successfully resistant and resilient with a relative sequence proportion greater than 1% in the shrimp cultivated from RAS and H-Floc systems (chapter 4). For this reason, as can be seen in Table 1, the potential functions of Pseudoalteromonas spp. candidates were summarised.

Table 1: Potential functions of Pseudoalteromonas spp. candidates

Proposed taxonomic Function Reference name P. sp. SCSE709-6 bioaugmentation effects in marine Jiang et al., 2019 blackwater treatment to prevent marine pollution P. espejiana DSM9414 amino-acid-requiring strain from Rong et al., 2018 seawater P. phenolica a novel bacteriophage Liu et al., 2018b P. phenolica decomposes starch and produces Isnansetyo and Kamei, 2003 phenolic antibiotics. P. sp. DIT09, DIT44, anti-Vibrio parahaemolyticus Aranda et al., 2012 activity DIT46 P. gelatinilytica producing acid from carbohydrates Yan et al., 2016 P. citrea novel biofilm bacteria for preventing Wesseling et al., 2015 P. rubra marine fish from pathogenic microbes P. amylolytica violacein-producing bacterium Wu et al., 2017 P. arabiensis novel bacteria for producing Matsuyama et al., 2013 polysaccharide P. flavipulchra CDM8 have an important role in anti- Vibrio Wang et al., 2018a P. piscicida CDA22 parahaemolyticus-caused acute hepatopancreatic necrosis disese,

AHPND (VpAHPND),

they were proposed as anti-VpAHPND probiotic candidates

Page 134 of 208

As shown in Table 1, Pseudoalteromonas spp. are potentially highly functional in biofloc-based shrimp aquaculture system. They also represent potential bio-indicator candidates with pink flocs, and this may create a sustainable floc system in the future.

Chapters 2 and 3 supported the advantages of heterotrophic bacteria; for example, Marivita sq3 and Ruegeria sq1 are two candidates for this (Hou et al., 2017). According to Chen et al., (2016), some heterotrophic nitrification bacteria are highly resistant to ammonia can have an advantage in treating organic wastewater. Marivital bacteria have not yet been proposed for potential functions in aquaculture as ruegerial bacteria have; therefore, only the possible functions of Ruegeria spp. are summarised in Table 2.

Table 2: Potential functions of Ruegeria spp. candidates

Proposed Function Reference taxonomic name Ruegeria sp. inhibit growth of the pathogen Vibrio Miura et al., 2018 coralliilyticus R. pomeroyi DSS-3 non-pathogenic marine bacterium, Christie-Oleza and Armengaud, 2010 consuming toxin proteins R. pomeroyi DSS-3 is the first marine heterotrophic Rivers et al., 2016 Roseobacterium was proposed as a needed study on using the diverse mixture of organic C molecules available in seawater R. mobilis YJ3 using N-acylhomoserine lactones Cai et al., 2018 (AHLs), proposed as a therapeutic agent in aquaculture R. mobilis are able to degrade of phosphate Yamaguchi et al., 2016 triesters (neurotoxic) R. spp. have a wide range of salinity and pH Huo et al., 2011 (pH 7.5 is optimum for growth)

As shown in Table 2, ruegerial bacteria (e.g. Ruegeria sq1) could engage in biocontrol activity against the Vibrio group, especially V. harveyi-a pathogen using N-acylhomoserine lactones

Page 135 of 208

(AHLs) in cell-cell communication for virulency (Defoirdt et al., 2008; Crab et al., 2010c; Miura et al., 2018). Therefore, directions for future biofloc research should focus on Pseudoalteromonas sq91 and Ruegeria sq1 as presumptive species that are highly recommended for potential biofloc inoculants (Hashim et al., 2018) to combat opportunistic bacteria (r- strategists) available in the water column, as well as in the host shrimp gut. This approach to biofloc inoculants is becoming increasingly important for further development of more sustainable aquaculture practices. These are bacteria that can compete with opportunistic and/or pathogenic bacteria in the rearing water when the host shrimp are young (i.e. small shrimp that cannot eat microbial flocs). These bacteria can stimulate or trigger an innate immune response when host shrimp are older (and can eat microbial flocs, i.e. cultivated shrimp at week 4; Chen and He, 2019).

We could see that the heterotrophic bioflocs formed a system in which the microbial mixed culture was rapidly growing using waste nitrogen, which was recycled to young bacterial cells (i.e. protein). The uptake of bioflocs (i.e. microbial flocs) by host shrimp depends on shrimp size (chapter 4) and floc size (chapter 2). This heterotrophic biofloc method can be applicable to either extensive, intensive shrimp culture systems, or super-intensive shrimp aquaculture as this work. From the laboratory results, the heterotrophic microbial biomass is suspected to have a controlling effect on pathogenic bacteria (Michaud et al., 2006) with the presence of polyhydroxybutyrate (PHB) in bioflocs. PHB-accumulating bacteria may reduce pathogenic bacteria in aquaculture systems (Defoirdt et al., 2007; Halet et al., 2007). This heterotrophic biofloc method was evidenced as a method of reducing the feed supply; as can be seen, the advantages of Ruegeria sq1 particle-attached bacteria were permanent resistance and resilience in both the water column and host shrimp tissue (i.e. shrimp gut) at week 4 (chapter 3, chapter 4). The potential ruegerial bacteria in this thesis study showed that beneficial bacteria mineralise organic matter, and thus, they may help to reduce the disease risk and increase nitrogen retention in harvested shrimp biomass. For example, Ruegeria sq1 could help reduce the demand for feed protein via bacterial protein. Most pathogens relevant to shrimp aquaculture are opportunistic bacteria and constitute part of the normal microbiota. However, organic matter in the culture systems also provides excellent growth conditions for opportunistic pathogens. Therefore, this thesis study was designed to carry out semi-biofloc implementation to avoid the invasion of pathogenic bacteria in the systems because full-biofloc implementation normally causes

Page 136 of 208

excessive amounts of TSS. Nevertheless, we could develop a full-biofloc system model, as can be seen in Figure 3.

Opportunistic bacteria Pseudoalteromonas sq91

shrimp tank at level of high carrying capacity

Catenovulum sq142

Ruegeria sq1

Marivita sq3

bioflocculant-producing bacteria Heterotrophic biofilter

Figure 3: Proposed model of full-biofloc-based shrimp aquaculture system based on the outputs of this thesis with the proposed name shrimp eco-culture system.

As shown in Figure 3, this model may work in terms of both the following components: 1) ecosystem (i.e. microbial ecological model without disinfection) and 2) high shrimp + productivity. The model may help reduce potentially nitrogen toxic (NH3-N and NH4 -N) and - nitrite (NO2 -N) concentrations in the rearing water with the presence of Marivita sq3 and reduce the amount of water-based nitrogen discharged to the surrounding environment, causing aquaculture effluents. The carbohydrate is added to the water column to promote heterotrophic bacterial protein production (e.g. the high proportion of Ruegeria sq1 in the shrimp tissue), and the feeding regime can be reduced from 40% to 25% of the protein level in the shrimp diet (Hari et al., 2006). This proposed microbial ecological model is highly possible, since many members

Page 137 of 208

of Ruegeria, Marivita, Catanovulum and Pseudoalteromonas bacteria have the same approximately pH 7.5 optimum for growth.

In another update, Pseudoalteromonas spp. were reported as candidates with high quality and long resistance in a specific biofloc system, and it was found that they can quickly kick start the formulating bioflocs, called bioflocculant-producing bacteria (Hashim et al., 2018). Hence, this shrimp eco-culture system will have all the benefits from such potential bacterial species in the direction of super-intensive shrimp aquaculture and sustainable tropical marine water management. Shrimp productivity can be maximised using a biological sustainability approach. More feed can be added more frequently, and carbohydrate can be inserted into the systems daily to increase the shrimp biomass because of TSS levels substantially can be controlled by the heterotrophic biofilter chamber. As excessive TSS amounts always have the potential to cause disease outbreaks through opportunistic pathogen invasion (Crab et al., 2007; Ebeling et al., 2006; Hargreaves, 2006; Ray et al., 2010; Schveitzer et al., 2013; Avnimelech, 2015), many fundamentals can be expected based on this bacterial model in the coming years. Therefore, the long-term sustainability of the aquaculture industry may be ensured because the environmental effects can be minimised by measures like the flocculation process for aquaculture wastewater treatment. The model will need to be applied by highly educated or at least highly skilled farmers and shrimp growers (Bossier and Ekasari, 2017). Moreover, it may minimise the disadvantage of the slow growth rate when using nitrifying bacteria in bio-filters (The model will use heterotrophic bacteria in the bio-filter chamber). Indeed, the growth rate of nitrifying bacteria is 10 times lower than that of heterotrophic bacteria present in the bioflocs (Crab et al., 2007). Overall, this model is a proposed idea with high potential for developing a new applied project in accordance with ‘Research2Business’ and ‘Science2Market’ to contribute to the scientific development pursued by the Leibniz Centre for Tropical Marine Research (ZMT). This will involve more in-depth investigations at the species level for the proposed bacterial candidates to identify suitable probiotic bacteria for super-intensive shrimp aquaculture in the future.

5.3 Thresholds: A holistic approach to preventing Vibrio disease

It is essential for the future BFT to understand the microbial mechanisms involved in the process of flocculation, quorum sensing and controlling the effect on pathogenic microbes. Therefore, I

Page 138 of 208

propose to use the bio-index/indicators based on this thesis study and references, as shown in Table 3.

Table 3: Indicators for quick checking of items in intensive shrimp aquaculture

Checking item Indicators Reference a matured system Suc-Vibrio/ Suc+Vibrio ≤ 1 Chapter 2 Huang et al., 2009 Reichardt et al., 2013 a matured system 250-μm size SS < original size SS Chapter 2 (flocs size) healthy shrimp Vibrio/ Bacillus ≤ 1 Chapter 2 healthy shrimp Shewanellaceae/ Vibrionaceace ~ 1 Chapter 4 (in shrimp gut, a symbiotic relationship) healthy shrimp the more bacterial diversity the less dominant Chapter 4 Vibrio population (in shrimp gut) healthy shrimp Setting solids ≤ 10.50 ± 1.66 mL L-1 Chapter 2 - healthy shrimp NO2 -N concentration ≤ 0.05 Chapter 2 healthy shrimp suc-Vibrio ≤ 104 CFU mL-1 Chapter 2 (green colony Vibrio) Chapter 4 Tran, 2013 Joshi et al., 2014 Soto-Rodriguez et al., 2015 high weight gain rates pH ~ 7.4 Chater 2 (for the heterotrophic bioflocs systems Furtado et al., 2015 described in this thesis) Pacheco-Vega et al., 2018 free-living bacterium in Marivita sq3 Chapter 3 heterotrophic biofloc (presumptive alphaproteobacterial species) system particle bacteria in Ruegeria sq1 Chapter 3 heterotrophic biofloc (presumptive alphaproteobacterial species) system Mycobacterium sq253 (presumptive actinobacterial species) Page 139 of 208

RAS’s characteristic Roseovarius sq175 Chapter 3 bacteria (presumptive alphaproteobacterial species)

Catenovulum sq142 (presumptive gammaproteobacterial species)

Pseudoalteromonas sq142 (presumptive gammaproteobacterial species) rearing water ‘system differences’ take important effects on Chapter 3 bacterial community composition

Marivita sq3 is important for water quality index shrimp gut ‘times’ differences take important effects on Chapter 4 bacterial community composition (regarding to shrimp size)

Ruegeria sq1 is important for protein food index (high proportion in shrimp gut)

Table 3 lists the suggested bio-indicators and indicators with thresholds for the intensive shrimp aquaculture that can be applicable for an indoor or outdoor eco-system model. However, a holistic approach to preventing Vibrio diseases cannot rely solely on the threshold of each indicator. Many issues must be incorporated to realise excellent shrimp productivity. Here, some possible tools for EMS prevention can be listed, including the following: 1) hatchery management, 2) farming management, 3) resistant genes and 4) additives. These tools can be clarified in more detail in terms of an EMS prevention approach with the following aims: 1) eliminating the EMS risk, 2) establishing a stable microbiome, 3) promoting shrimp resistance to EMS, 4) using suitable probiotics as bio-inoculants, 5) avoiding excessive feeding, 6) separating between the larval stage and juvenile stage for intensive shrimp aquaculture, 7) improving biosecurity monitoring or attention to the system’s carrying capacity, 8) regularly applying quick

Page 140 of 208

diagnostic methods and 10) using laboratory help for a long-term EMS prevention approach to killing all possible vectors of EMS.

5.4 Concluding remarks and outlook

The aim of this thesis study was using bioflocs as an alternative approach to controlling opportunistic bacteria in shrimp aquaculture. The goal of this work approach was designed by application of r/K selection (or research based on ecological theory) in microbial community composition management. This work’s hypothesis that the heterotrophic biofloc system is better for L. vannamei production and the growth of beneficial bacteria was addressed with conclusions in each chapter. Although critical future work remains in terms of evaluating the interactions among bacteria to determine potential applications of the proposed bio-inoculants in aquaculture and food safety, here, some concluding remarks are given, as the new findings allow recommendations for a biofloc-based shrimp aquaculture system or super-intensive shrimp aquaculture:

First, TSS should be considered as a useful bio-monitoring tool for bacterial ecology or biological threats in the shrimp culture systems.

Second, the heterotrophic biofloc system was demonstrated to be better than the autotrophic biofloc, flow-through and recirculating systems, with several advantages of heterotrophic bacteria for improving the water quality and shrimp weight gain rate, for example, Marivita sq3 and Ruegeria sq1. Marivita sq3 indicated a presumptive free-living bacterium of the Alphaproteobacteria class. Ruegeria sq1 was indicated as a presumptive particle bacterium in members of ruegerial bacteria in the H-Floc systems; this has not been previously reported. From the results of this thesis, a ruegerial bacterium (Ruegeria sq1) is proposed as having a newly identified shrimp culture particle component and shrimp gut component. Notably, this bacterium presented as a stable population in the specific heterotrophic biofloc conditions from the beginning to the end of the shrimp culture period, as well as representatives in the host shrimp gut.

Third, Pseudoalteromonas spp. can be highly applicable to heterotrophic biofloc systems as bio- inoculant or bio-indicator.

Page 141 of 208

Fourth, free-living and particle bacteria were highly different in terms of the OTU numbers and compositions. They were also different in the sequence OTU bio-indicators. At a fraction of free- living samples, there were no significant differences in OTU numbers among the two biofloc systems. Therefore, particle-attached bacteria are a key factor in making differences in bacterial community compositions. This thesis is the first study analyzed bacterial community separated into free-living and particle-attached bacteria in biofloc aquaculture systems.

Fifth, a beneficial bacterium is needed to stabilise resistance and resilience in both rearing water and the colonised host shrimp gut. For example, Ruegeria sq1 bacteria can be introduced.

Sixth, r-matured bacteria (e.g. Ruegeria sq1) were proposed as an alternative approach for preventing Vibrio disease through bacterial ‘activity’ management by using Ruegeria as competing bacteria.

Seventh, the individual dominance of a single bacterium was most important in the culture systems, since the study evidenced that they are functional systems with an extremely high proportion. As another result, the most dominant single-sequence OTU abundance does not always belong to the relative class abundances of that bacterial community. Therefore, the individual sequence OTU (at the presumptive species level) with a high proportion had a greater ecological function in the aquaculture systems than the relative class abundances in this study did. This means that managing single pathogenic dominance is effective with another antagonistic r-bacterium, for example, a bio-inoculant bacterium. They are both r-strategists; therefore, they are equal in speed of growth to combat each other. The interpretation concerning Ruegeria sq1 and Pseudoalteromonas sq91 in this thesis will open the gateway to further research on the identification of possible alternative treatments. These can include the use of probiotics and immunostimulants and application of microbially matured water systems with matured r-probiotic bacteria for preventing EMS disease outbreaks in shrimp farming. It is recommended that managing the microbial activity will be more efficient than eliminating harmful bacteria. Hence, the growth of bacteria in the systems determines their sustainability and security.

Eighth, single bacterial species will only be dominant with an extremely high proportion if the environmental conditions are designed especially for their growth. Thus, cultured shrimp are

Page 142 of 208

needed to culture the water column; for example, the heterotrophic biofloc system described in this study represented the niches of Ruegeria sq1 and Marivita sq3 bacteria.

Ninth, in the complex interactions among microbes, microbes with larvae, and finally, microbes with shrimp host larvae and the environment, the pH balance was the most important indicator of healthy and sustainable shrimp culture systems. For example, a pH level close to 7.4 in the H- Floc system stabilised the internal microbial composition.

Tenth, for a specific L. vannamei challenge assay with toxic bacteria, the provision of feed during the immersion challenge is not necessary. However, accurate shrimp incubation practice is needed to eliminate moulting shrimp and dead shrimp due to stress.

As a future outlook, it should be considered that existing and new infectious agents can arise at any time during the shrimp culture period. Therefore, this thesis provided baseline information on bacterial ecology with a view to establishing healthy super-intensive shrimp aquaculture. The shrimp eco-culture system has high potential in contributing to responsible aquaculture in the future, focussing on developing a beneficial bacteria kit, such as for Ruegeria sq1 and Pseudoalteromonas sq91 bacteria.

Page 143 of 208

References

Abdullah S.N., Ariyawansa, S., Babji, A. S., and Hashim, J. K. (2013). The risk assessment of Vibrio parahaemolyticus in cooked black tiger shrimps (Penaeus monodon) in Malaysia. Food Control. 31, 546-552. doi:10.1016/J.FOODCONT.2012.10.018.

Aguilera-Rivera, D., Prieto-Davó, A., Escalante, K., Chávez, C., Cuzon, G., and Gaxiola, G. (2014). Probiotic effect of floc on Vibrios in the pacific white shrimp Litopenaeus vannamei. Aquaculture. 424-425, 215-19. doi:10.1016/j.aquaculture.2014.01.008.

Aguirre-Guzman, G., Mejia Ruiz, H., and Ascencio, F. (2004). A review of extracellular virulence product of Vibrio species important in diseases of cultivated shrimp. Aquac. Res. 35, 1395-1404. doi:10.1111/j.1365-2109.2004.01165.x.

Ahmed, H. A., El Bayomi, R. M., Hussein, M. A., Khedr, M. H. E., Abo Remela, E. M., and El- Ashram, A. M. M. (2018). Molecular characterization, antibiotic resistance pattern and biofilm formation of Vibrio parahaemolyticus and V. cholerae isolated from crustaceans and humans. Int. J. Food Microbiol. 274, 31-37. doi:10.1016/J.IJFOODMICRO.2018.03.013.

Allison, S. D., and Martiny, J. B. H. (2008). Resistance, resilience, and redundancy in microbial communities. Proc. Natl. Acad. Sci. U. S. A. 105 Suppl, 11512-9. doi:10.1073/pnas.0801925105.

Alvarez, A., Saez, J. M., Davila Costa, J. S., Colin, V. L., Fuentes, M. S., Cuozzo, S. A., Benimeli, C.S., Polti, M.A., and Amoroso, M.J. (2017). Actinobacteria: Current research and perspectives for bioremediation of pesticides and heavy metals. Chemosphere 166, 41- 62. doi:10.1016/J.CHEMOSPHERE.2016.09.070.

Andrews, J. H., and Harris, R. F. (1986). r- and K-Selection and Microbial Ecology, 9, 99-147. Springer, Boston, MA. doi: 10.1007/978-1-4757-0611-6_3

APHA. (2005). Standard Methods for the Examination of Water and Wastewater. 21st Edition/American Public Health Association/American Water Works Association/Water Environment Federation Washington, DC.

Page 144 of 208

Aranda, C. P., Valenzuela, C., Barrientos, J., Paredes, J., Leal, P., Maldonado, M., Godoy, F.A., and Osorio, C. G. (2012). Bacteriostatic anti-Vibrio parahaemolyticus activity of Pseudoalteromonas sp. strains DIT09, DIT44 and DIT46 isolated from Southern Chilean intertidal Perumytilus purpuratus. World Journal of Microbiology and Biotechnology, 28(6), 2365-2374. doi: 10.1007/s11274-012-1044-z

Arunrut, N., Kampeera, J., Sirithammajak, S., Sanguanrut, P., Proespraiwong, P., Suebsing, R., and Kiatpathomchai, W. (2016). Sensitive Visual Detection of AHPND Bacteria Using Loop-Mediated Isothermal Amplification Combined with DNA-Functionalized Gold Nanoparticles as Probes. PLoS One 11, e0151769. doi:10.1371/journal.pone.0151769.

Ashrafudoulla, M., Mizan, M. F. R., Park, H., Byun, K.-H., Lee, N., Park, S. H., Byun, N.L, Park, S.H and Ha, S.D. (2019). Genetic Relationship, Virulence Factors, Drug Resistance Profile and Biofilm Formation Ability of Vibrio parahaemolyticus Isolated From Mussel. Front. Microbiol. 10. doi:10.3389/fmicb.2019.00513.

Austen, M. C., Lambshead, P. J. D., Hutchings, P. A., Boucher, G., Snelgrove, P. V. ., Heip, C., King, G., and Koike, I., Smith, C. (2002). Biodiversity links above and below the marine sediment-water interface that may influence community stability. Biodivers. Conserv. 11, 113-136. doi:10.1023/A:1014098917535.

Austin, B. (1988) Marine Microbiology. Cambridge, UK: Cambridge University Press.

Avnimelech, Y. (1999). Carbon/nitrogen ratio as a control element in aquaculture systems. Aquaculture 176, 227-235. doi:10.1016/S0044-8486(99)00085-X.

Avnimelech, Y. (2015). Biofloc Technology - A Practical Guidebook. World Aquaculture Society. Baton Rouge, United States. ISBN: 9781888807226.

Avnimelech, Y. (2016). Biofloc Technology - A Practical Guide Book Control of microbial activity in aquaculture systems: active suspension ponds. Reprinted from World Aquaculture 34(4):19-21. ©2003 World Aquaculture Society.

Avnimelech, Y., and Kochba, M. (2009). Evaluation of nitrogen uptake and excretion by tilapia in biofloc tanks, using 15N tracing. Aquaculture 287, 163-168. doi:10.1016/j.aquaculture.2008.10.009. Page 145 of 208

Bakdash, J. Z., and Marusich, L. R. (2017). Repeated measures correlation. Frontiers in Psychology, 8(MAR), 1-13. doi: 10.3389/fpsyg.2017.00456

Bartling, P., Vollmers, J., and Petersen, J. (2018). The first world swimming championships of roseobacters -Phylogenomic insights into an exceptional motility phenotype. Syst. Appl. Microbiol. doi:10.1016/j.syapm.2018.08.012.

Basi-Chipalu, S., Dischinger, J., Josten, M., Szekat, C., Zweynert, A., Sahl, H.-G., and Bierbaum, G. (2015). Pseudomycoicidin, a Class II Lantibiotic from Bacillus pseudomycoides. Applied and Environmental Microbiology, 81(10), 3419-29. doi: 10.1128/AEM.00299-15

Bates, D., Maechler, M., Bolker, B., and Walker S. (2015). Fitting Linear Mixed-Effects Models Using lme4. Journal of Statistical Software, 67(1), 1-48. doi:10.18637/jss.v067.i01.

Becerra Dórame, M. J., Martínez Córdova, L. R., Martínez Porchas, M., and Lopez Elías, J. A. (2011). Evaluation of Autotrophic and Heterotrophic Microcosm-based Systems on the Production Response of Litopenaeus vannamei Intensively Nursed without Artemia and with Zero Water Exchange. The Israeli Journal of Aquaculture-Bamidgeh, 620(7), 1-7. doi: 10524/36290

Ben-Haim, Y., Banim, E., Kushmaro, A., Loya, Y., and Rosenberg, E. (1999). Inhibition of photosynthesis and bleaching of zooxanthellae by the coral pathogen Vibrio shiloi. Environ. Microbiol. 1, 223-229. doi:10.1046/j.1462-2920.1999.00027.x.

Bentzon-Tilia, M., Sonnenschein, E. C., and Gram, L. (2016). Monitoring and managing microbes in aquaculture - Towards a sustainable industry. Microbial Biotechnology. 9, 576- 584. doi:10.1111/1751-7915.12392.

Bienhold, C., Boetius, A., and Ramette, A. (2012). The energy-diversity relationship of complex bacterial communities in Arctic deep-sea sediments. ISME J. 6, 724-732. doi:10.1038/ismej.2011.140.

Boone, L. (1931). Anomuran, macruran Crustacea from Panama and Canal Zone. Bull. Am. Mus. Nat. Hist. 63, 137-189.

Page 146 of 208

Bossier, P., and Ekasari, J. (2017). Biofloc technology application in aquaculture to support sustainable development goals. Microbial Biotechnology. doi:10.1111/1751-7915.12836.

Boyd, C. E. and Clay, J. W. (2002) Evaluation of Belize Aquaculture, Ltd: A Super-intensive Shrimp Aquaculture System. In Review report prepared under the World Bank, NACA, WWF and FAO Consortium Program on Shrimp Farming and the Environment, 1-17.

Bray, J. R., and Curtis, J. T. (1957). An Ordination of the Upland Forest Communities of Southern Wisconsin. Ecol. Monogr. 27, 325-349. doi:10.2307/1942268.

Bregnballe, J. (2015). Recirculation Aquaculture: An introduction to the new environmentally friendly and highly productive closed fish farming systems.

Brettar, I., Moore, E. R. B., and Höfle, M. G. (2001). Phylogeny and Abundance of Novel Denitrifying Bacteria Isolated from the Water Column of the Central Baltic Sea. Microb. Ecol. 42, 295-305. doi:10.1007/s00248-001-0011-2.

Browdy, C. L., Bratvold, D., Mcintosh, A. D. S. R. P., and Castañeda, J. (2001). Perspectives on the application of closed shrimp culture systems. https://www.researchgate.net/publication/250309306

Buchan, A., González, J. M., and Moran, M. A. (2005). Overview of the marine roseobacter lineage. Appl. Environ. Microbiol. 71, 5665-77. doi:10.1128/AEM.71.10.5665-5677.2005.

Burford, M. A., Thompson, P. J., McIntosh, R. P., Bauman, R. H., and Pearson, D. C. (2004). The contribution of flocculated material to shrimp (Litopenaeus vannamei) nutrition in a high-intensity, zero-exchange system. Aquaculture 232, 525-537. doi:10.1016/S0044- 8486(03)00541-6.

Cai, L., Ye, L., Tong, A. H. Y., Lok, S., and Zhang, T. (2013). Biased Diversity Metrics Revealed by Bacterial 16S Pyrotags Derived from Different Primer Sets. PLoS One 8, 1–11. doi:10.1371/journal.pone.0053649.

Cai, X., Yu, M., Shan, H., Tian, X., Zheng, Y., Xue, C., and Zhang, X.-H. (2018). Characterization of a Novel N-Acylhomoserine Lactonase RmmL from Ruegeria mobilis YJ3. Mar. Drugs 16, 370. doi:10.3390/md16100370.

Page 147 of 208

Callahan, B. J., McMurdie, P. J., Rosen, M. J., Han, A. W., Johnson, A. J. A., and Holmes, S. P. (2016). DADA2: High-resolution sample inference from Illumina amplicon data. Nat. Methods 13, 581-583. doi:10.1038/nmeth.3869.

Capone, D. G., Bronk, D. A., Mulholland, M. R., and Carpenter, E. J. (2008). Nitrogen in the marine environment. 2nd Edition. ISBN: 978-0-12-372522-6. (1758 pp)

Chaweepack, T., Muenthaisong, B., Chaweepack, S., and Kamei, K. (2015). The Potential of Galangal (Alpinia galanga Linn.) Extract against the Pathogens that Cause White Feces Syndrome and Acute Hepatopancreatic Necrosis Disease (AHPND) in Pacific White Shrimp (Litopenaeus vannamei). Int. J. Biol. 7, p8. doi:10.5539/ijb.v7n3p8.

Chen, J., Han, Y., Wang, Y., Gong, B., Zhou, J., and Qing, X. (2016). Start-up and microbial communities of a simultaneous nitrogen removal system for high salinity and high nitrogen organic wastewater via heterotrophic nitrification. Bioresource Technology. doi: 10.1016/j.biortech.2016.05.064

Chen, Y.-H., and He, J.-G. (2019). Effects of environmental stress on shrimp innate immunity and white spot syndrome virus infection. Fish Shellfish Immunol. 84, 744-755. doi:10.1016/J.FSI.2018.10.069.

Christie-Oleza, J. A., and Armengaud, J. (2010). In-Depth Analysis of Exoproteomes from Marine Bacteria by Shotgun Liquid Chromatography-Tandem Mass Spectrometry: the Ruegeria pomeroyi DSS-3 Case-Study. Mar. Drugs 8, 2223-2239. doi:10.3390/md8082223.

Chu, C. P., and Lee, D. J. (2004). Multiscale structures of biological flocs. Chem. Eng. Sci. 59, 1875-1883. doi:10.1016/j.ces.2004.01.040.

Colt, J. (2006). Water quality requirements for reuse systems. Aquac. Eng. 34, 143-156. doi:10.1016/J.AQUAENG.2005.08.011.

Cox, D.R . (1972). Regression Models and Life-Tables. JR Stat. Soc B; 34 187-220.

Crab, R., Avnimelech, Y., Defoirdt, T., Bossier, P., and Verstraete, W. (2007). Nitrogen removal techniques in aquaculture for a sustainable production. Aquaculture 270, 1-14. doi:10.1016/j.aquaculture.2007.05.006.

Page 148 of 208

Crab, R. (2010a). Bioflocs technology: an integrated system for the removal of nutrients and simultaneous production of feed in aquaculture. PhD thesis, Ghent University. 178 pp.

Crab, R., Chielens, B., Wille, M., Bossier, P., and Verstraete, W. (2010b). The effect of different carbon sources on the nutritional value of bioflocs, a feed for Macrobrachium rosenbergii postlarvae. Aquac. Res. 41, 559-567. doi:10.1111/j.1365-2109.2009.02353.x.

Crab, R., Lambert, A., Defoirdt, T., Bossier, P., and Verstraete, W. (2010c). The application of bioflocs technology to protect brine shrimp (Artemia franciscana) from pathogenic Vibrio harveyi. J. Appl. Microbiol. 109, no-no. doi:10.1111/j.1365-2672.2010.04791.

Crab, R., Defoirdt, T., Bossier, P., and Verstraete, W. (2012). Biofloc technology in aquaculture: Beneficial effects and future challenges. Aquaculture 356-357, 351-356. doi:10.1016/j.aquaculture.2012.04.046.

Crab, R., Kochva,M., Verstraete,W., and Avnimelech, Y. (2009). Bio-flocs technology application in over-wintering of tilapia. Aquaculture. Eng. 40, 105-112. doi: 10.1016/j.aquaeng.2008.12.00

Cruz-lópez, R., and Maske, H. (2016). The Vitamin B1 and B12 required by the marine Dinoflagellate Lingulodinium polyedrum can be provided by its associated bacterial community in culture. 7, 1-13. doi:10.3389/fmicb.2016.00560.

Csardi G, N. T. (2006). The igraph software package for complex network research. InterJournal, Complex Systems 1695. Http://igraph.org.

Dang, H., and Lovell, C. R. (2016). Microbial Surface Colonization and Biofilm Development in Marine Environments. Microbiol. Mol. Biol. Rev. 80, 91-138. doi:10.1128/MMBR.00037- 15.

Dangtip, S., Sirikharin, R., Sanguanrut, P., Thitamadee, S., Sritunyalucksana, K., Taengchaiyaphum, S., Mavichak, R., Proespraiwong, P., and Flegel, T.W. (2015). AP4 method for two-tube nested PCR detection of AHPND isolates of Vibrio parahaemolyticus. Aquac. Reports 2, 158-162. doi:10.1016/j.aqrep.2015.10.002.

Defoirdt, T. (2016). Implications of Ecological Niche Differentiation in Marine Bacteria for

Page 149 of 208

Microbial Management in Aquaculture to Prevent Bacterial Disease. PLoS Pathogens. 12, 1-5. doi:10.1371/journal.ppat.1005843.

Defoirdt, T., Boon, N., Sorgeloos, P., Verstraete, W., and Bossier, P. (2008). Quorum sensing and quorum quenching in Vibrio harveyi: lessons learned from in vivo work. ISME J. 2, 19- 26. doi:10.1038/ismej.2007.92.

Defoirdt, T., Halet, D., Vervaeren, H., Boon, N., Van de Wiele, T., Sorgeloos, P., and Bossier, P., Verstraete. (2007). The bacterial storage compound poly-β-hydroxybutyrate protects Artemia franciscana from pathogenic Vibrio campbellii. Environmental. Microbiology. 9, 445-452. doi:10.1111/j.1462-2920.2006.01161.x.

Defoirdt, T., Doan, T.L., Van Delsen, B., De Schryver, P., Sorgeloos, P., Boon, N., and Bossier, P. (2011). N-acylhomoserine lactone-degrading Bacillus strains isolated from aquaculture animals. Aquaculture, 311(1-4). doi: 10.1016/j.aquaculture.2010.11.046

De Schryver, P., Defoirdt, T., and Sorgeloos, P. (2014). Early Mortality Syndrome Outbreaks: A Microbial Management Issue in Shrimp Farming? PLoS Pathogens. 10, e1003919. doi:10.1371/journal.ppat.1003919.

De Schryver, P., and Vadstein, O. (2014). Ecological theory as a foundation to control pathogenic invasion in aquaculture. The ISME Journal, 8(12), 2360-2368. doi: 10.1038/ismej.2014.84

De Schryver, D.P., Crab, R., Defoirdt, T., Boon, N., and Verstraete, W. (2008). The basics of bio-flocs technology: The added value for aquaculture. Aquaculture 277, 125-137. doi:10.1016/J.AQUACULTURE.2008.02.019.

Doyle, J. J., and Doyle, J. L. (1987). A Rapid DNA Isolation Procedure for Small Quantities of Fresh Leaf Tissue. Phytochemical Bulletin, 19, 11-15.

Ebeling, J. M., Timmons, M. B., and Bisogni, J. J. (2006). Engineering analysis of the stoichiometry of photoautotrophic, autotrophic, and heterotrophic removal of ammonia- nitrogen in aquaculture systems. Aquaculture 257, 346-358. doi:10.1016/j.aquaculture.2006.03.019.

Page 150 of 208

Edition, D. (2014). The future of European shrimp: land-based, sustainable and German. Fish Farming International.

Ekasari, J., Hanif Azhar, M., Surawidjaja, E. H., Nuryati, S., De Schryver, P., and Bossier, P. (2014). Immune response and disease resistance of shrimp fed biofloc grown on different carbon sources. Fish Shellfish Immunology. 41, 332-339. doi:10.1016/j.fsi.2014.09.004.

Elexson, N., Yaya, R., Nor, A. M., Kantilal, H. K., Ubong, A., Yoshitsugu, N., Nishibuchi, M. and Son, R. (2014). Biofilm assessment of Vibrio parahaemolyticus from seafood using random amplified polymorphism DNA-PCR. Int. Food Res. J. doi:10.3945/ajcn.112.042457.Am.

Elifantz, H., Horn, G., Ayon, M., Cohen, Y., and Minz, D. (2013). Rhodobacteraceae are the key members of the microbial community of the initial biofilm formed in Eastern Mediterranean coastal seawater. FEMS Microbiol. Ecol. 85, 348-357. doi:10.1111/1574-6941.12122.

Emerenciano, M. G. C., Martínez-Córdova, L. R., Martínez-Porchas, M., and Miranda-Baeza, A. (2017). Biofloc Technology (BFT): A Tool for Water Quality Management in Aquaculture. In Water Quality. InTech. doi: 10.5772/66416

Emerenciano, M., Gaxiola, G., and Cuzon, G. (2013). Biofloc Technology (BFT): A Review for Aquaculture Application and Animal Food Industry. Intech, 301-328. doi:10.5772/53902.

FAO. (1999). Responsible aquaculture production. Consult. Appl. Artic. FAO Code Conduct Responsible Fish.

FAO. (2012). The State of World Fisheries and Aquaculture 2012. Rome.

FAO. (2016). The State of World Fisheries and Aquaculture 2016. Rome.

FAO. (2017). The State of World Fisheries and Aquaculture 2017. Rome.

FAO. (2018). The State of World Fisheries and Aquaculture 2018-Meeting the sustainable development goals. doi:issn 10.

Farmer J.J., and Hickman-Brenner F.W. (1992). The genera Vibrio and Photobacterium. A Handbook on the Biology of Bacteria: Ecophysiology, Isolation, Identification, and

Page 151 of 208

Applications. 2nd ed. Springer; Berlin, Germany. pp. 2952-3011.

Flegel, T., Pasharawipas, T., and Owens, L., Oakey, H. J. (2005). Evidence for Phage-Induced Virulence in the Shrimp Pathogen Vibrio harveyi. Dis. Asian Aquac., 329-337.

Flegel, T. W. (2019). A future vision for disease control in shrimp aquaculture. J. World Aquac. Soc. 50, 249-266. doi:10.1111/jwas.12589.

Fuller, R. (1989). Probiotics in man and animals. J. Appl. Bacteriol. 66, 365-378. doi:10.1111/j.1365-2672.1989.tb05105.x.

Furtado, P. S., Fugimura, M. M. S., Monserrat, J. M., Souza, D. M., Garcia, L. de O., and Wasielesky, W. (2015). Acute effects of extreme pH and its influences on the survival and biochemical biomarkers of juvenile white shrimp, Litopenaeus vannamei. Mar. Freshw. Behav. Physiol. 48, 417-429. doi:10.1080/10236244.2015.1086539.

Gaboyer, F., Vandenabeele-Trambouze, O., Cao, J., Ciobanu, M.-C., Jebbar, M., Le Romancer, M., and Alain, K. (2014). Physiological features of Halomonas lionensis sp. nov., a novel bacterium isolated from a Mediterranean Sea sediment. Res. Microbiol. 165, 490-500. doi:10.1016/j.resmic.2014.07.009.

Gao, B., and Gupta, R. S. (2012). Phylogenetic framework and molecular signatures for the main clades of the phylum Actinobacteria. Microbiol. Mol. Biol. Rev. 76, 66-112. doi:10.1128/MMBR.05011-11.

Gaona, C. A. P., Poersch, L. H., Krummenauer, D., Foes, G. K., and Wasielesky, W. J. (2011). The Effect of Solids Removal on Water Quality, Growth and Survival of Litopenaeus vannamei in a Biofloc Technology Culture System. International Journal of Recirculating Aquaculture, 12(1). doi: 10.21061/ijra.v12i1.1354

Glasby, T. M., and Underwood, A. J. (1996). Sampling to differentiate between pulse and press perturbations. Environ. Monit. Assess. 42, 241-252. doi:10.1007/BF00414371.

Goel, M. K., Khanna, P., and Kishore, J. (2010). Understanding survival analysis: Kaplan-Meier estimate. Int. J. Ayurveda Res. 1, 274-8. doi:10.4103/0974-7788.76794.

Gomez-Gil, B., Soto-Rodríguez, S., Lozano, R., and Betancourt-Lozano, M. (2014). Draft Page 152 of 208

Genome Sequence of Vibrio parahaemolyticus Strain M0605, Which Causes Severe Mortalities of Shrimps in Mexico. Genome Announc. 2, e00055-14. doi:10.1128/genomeA.00055-14.

Goodwin, A. E., Roy, J. S., Grizzle, J. M., and Goldsby, M. T. (1994). Bacillus mycoides: a bacterial pathogen of channel catfish, 18, 173-179. Diseases of aquatic organisms.

Hada, H. S., West, P. A., Lee, J. V, Stemmler, J., and Colwell1, R. R. (1984). Vibrio tubiashii sp. nov., a Pathogen of Bivalve Mollusks. International journal of systematic bacteriology, 34, 1-4.

Hahnke, R. L., and Harder, J. (2013). Phylogenetic diversity of Flavobacteria isolated from the North Sea on solid media. Syst. Appl. Microbiol. 36, 497–504. doi:10.1016/j.syapm.2013.06.006.

Haldar, S., Chatterjee, S., Sugimoto, N., Das, S., Chowdhury, N., Hinenoya, A., Asakura, M., and Yamasaki, S. (2010). Identification of Vibrio campbellii isolated from diseased farm- shrimps from south India and establishment of its pathogenic potential in an Artemia model. doi: 10.1099/mic.0.041475-0

Halet, D., Defoirdt, T., Van Damme, P., Vervaeren, H., Forrez, I., Van de Wiele, T., Boon, N., and Bossier, P., Verstraete, W. (2007). Poly-β-hydroxybutyrate-accumulating bacteria protect gnotobiotic Artemia franciscana from pathogenic Vibrio campbellii. FEMS Microbiol. Ecol. 60, 363–369. doi:10.1111/j.1574-6941.2007.00305.x.

Hall, M. R., McGillicuddy, E., and Kaplan, L. J. (2014). Biofilm: Basic Principles, Pathophysiology, and Implications for Clinicians. Surg. Infect. (Larchmt). 15, 1-7. doi:10.1089/sur.2012.129.

Hameed, A. S. S., Rao, P. V., Farmer, J. J., Hickman-Brenner, F. W., and Fanning, G. R. (1996). Characteristics and pathogenicity of a Vibrio campbellii-like bacterium affecting hatchery- reared Penaeus indicus (Milne Edwards, 1837) larvae. Aquac. Res. (United Kingdom).

Han, J., Tang, K., Tran, L., and Lightner, D. (2015). Photorhabdus insect-related (Pir) toxin-like genes in a plasmid of Vibrio parahaemolyticus, the causative agent of acute hepatopancreatic necrosis disease (AHPND) of shrimp. Dis. Aquat. Organ. 113, 33-40. Page 153 of 208

doi:10.3354/dao02830.

Hargreaves, J. A. (2006). Photosynthetic suspended-growth systems in aquaculture. Aquac. Eng. 34, 344-363. doi:10.1016/J.AQUAENG.2005.08.009.

Hari, B., Madhusoodana Kurup, B., Varghese, J. T., Schrama, J. W., and Verdegem, M. C. (2006). The effect of carbohydrate addition on water quality and the nitrogen budget in extensive shrimp culture systems. Aquaculture 252, 248-263. doi:10.1016/J.AQUACULTURE.2005.06.044.

Hashim, N. F. C., Ghazali, N. A., Kasan, N. A., Ismail, N., and Amin, N. M. (2018). Description of novel marine bioflocculant-producing bacteria isolated from biofloc of pacific whiteleg shrimp, Litopenaeus vannamei culture ponds. bioRxiv, 402065. doi:10.1101/402065.

Hassan, S. S. ul, and Shaikh, A. L. (2017). Marine actinobacteria as a drug treasure house. Biomed. Pharmacother. 87, 46-57. doi:10.1016/J.BIOPHA.2016.12.086.

Hassan, S. S. ul, Anjum, K., Abbas, S. Q., Akhter, N., Shagufta, B. I., Shah, S. A. A., and Tasneem, U. (2017). Emerging biopharmaceuticals from marine actinobacteria. Environ. Toxicol. Pharmacol. 49, 34-47. doi:10.1016/J.ETAP.2016.11.015.

Hau, H. H., and Gralnick, J. A. (2007). Ecology and Biotechnology of the Genus Shewanella. Annu. Rev. Microbiol. 61, 237-258. doi:10.1146/annurev.micro.61.080706.093257.

He, Z., Deng, Y., and Zhou, J. (2012). Development of functional gene microarrays for microbial community analysis. Curr. Opin. Biotechnol. 23, 49-55. doi:10.1016/j.copbio.2011.11.001.

He, Z., Deng, Y., Van Nostrand, J. D., Tu, Q., Xu, M., Hemme, C. L., Li, X., Wu, L., Gentry, T.J., Yin, Y., Liebich, J., Hazen, T.C., and Zhou, J. (2010). GeoChip 3.0 as a high- throughput tool for analyzing microbial community composition, structure and functional activity. ISME J. 4, 1167-1179. doi:10.1038/ismej.2010.46.

He, Z., Gentry, T. J., Schadt, C. W., Wu, L., Liebich, J., Chong, S. C., Huang, Z., Wu, W., Gu, B., Jardine, P., and Criddle, C., Zhou, J. (2007). GeoChip: a comprehensive microarray for investigating biogeochemical, ecological and environmental processes. ISME J. 1, 67-77.

Page 154 of 208

doi:10.1038/ismej.2007.2.

He, Z., Van Nostrand, J. D., and Zhou, J. (2013). “GeoChip-Based Metagenomic Technologies for Analyzing Microbial Community Functional Structure and Activities” in Encyclopedia of Metagenomics (New York, NY: Springer New York), 1-13. doi:10.1007/978-1-4614-

6418-1-760-4.

Hinners, J., Hofmeister, R., and Hense, I. (2015). Modeling the Role of pH on Baltic Sea Cyanobacteria. Life (Basel, Switzerland) 5, 1204-17. doi:10.3390/life5021204.

Hoebe, K., Janssen, E., and Beutler, B. (2004). The interface between innate and adaptive immunity. Nat. Immunol. 5, 971-974. doi:10.1038/ni1004-971.

Hong, H.-H., Choi, H., Cheon, S., Lee, H.-G., and Park, C. (2017). Genome Sequences of Two Shewanella spp. Isolated from the Gut of the Sea Cucumber Apostichopus japonicus (Selenka, 1867). Genome Announc. 5, e00674-17. doi:10.1128/genomeA.00674-17.

Hong, X., Lu, L., and Xu, D. (2016). Progress in research on acute hepatopancreatic necrosis disease (AHPND). Aquac. Int. 24, 577-593. doi:10.1007/s10499-015-9948-x.

Hou, D., Huang, Z., Zeng, S., Liu, J., Wei, D., Deng, X., Weng, S., He, Z., and He, J. (2017). Environmental Factors Shape Water Microbial Community Structure and Function in Shrimp Cultural Enclosure Ecosystems. Frontiers in Microbiology, 8, 2359. doi: 10.3389/fmicb.2017.02359

Huang, J., Yang, Q., Ma, Z., Zhou, F., Yang, L., Deng, J., and Jiang, S. (2017). Effects of adding sucrose on Penaeus monodon (Fabricius, 1798) growth performance and water quality in a biofloc system. Aquaculture Research. doi: 10.1111/are.13067

Huang, Y.-L., Ki, J.-S., Lee, O. O., and Qian, P.-Y. (2009). Evidence for the dynamics of Acyl homoserine lactone and AHL-producing bacteria during subtidal biofilm formation. ISME J. 3, 296-304. doi:10.1038/ismej.2008.105.

Huo, Y.-Y., Xu, X.-W., Li, X., Liu, C., Cui, H.-L., Wang, C.-S., and Wu, M. (2011). Ruegeria marina sp. nov., isolated from Marine Sediment. Int. J. Syst. Evol. Microbiol. 61, 347-350. doi:10.1099/ijs.0.022400-0.

Page 155 of 208

Hwang, C. Y., Bae, G. D., Yih, W., and Cho, B. C. (2009). Marivita cryptomonadis gen. nov., sp. nov. and Marivita litorea sp. nov., of the family Rhodobacteraceae, isolated from marine habitats. Int. J. Syst. Evol. Microbiol. 59, 1568-1575. doi:10.1099/ijs.0.005462-0.

Isnansetyo, A., and Kamei, Y. (2003). Pseudoalteromonas phenolica sp. nov., a novel marine bacterium that produces phenolic anti-methicillin-resistant Staphylococcus aureus substances. Int. J. Syst. Evol. Microbiol. 53, 583-588. doi:10.1099/ijs.0.02431-0.

Ivanova, E. P., Nedashkovskaya, O. I., Sawabe, T., Zhukova, N. V., Frolova, G. M., Nicolau, D. V., Mikhailov, V.V., and Bowman, J.P. (2004). Shewanella affinis sp. nov., isolated from marine invertebrates. Int. J. Syst. Evol. Microbiol. 54, 1089-1093. doi:10.1099/ijs.0.02992- 0.

Janda, J. M., and Abbott, S. L. (2014). The genus Shewanella: from the briny depths below to human pathogen. Crit. Rev. Microbiol. 40, 293-312. doi:10.3109/1040841X.2012.726209.

Jiang, L., Chen, X., Qin, M., Cheng, S., Wang, Y., and Zhou, W. (2019). On-board saline black water treatment by bioaugmentation original marine bacteria with Pseudoalteromonas sp. SCSE709-6 and the associated microbial community. Bioresour. Technol. 273, 496-505. doi:10.1016/j.biortech.2018.11.043.

Jiang, Z., Zhang, D.-F., Khieu, T.-N., Ky Son, C., Zhang, X.-M., Cheng, J., Tian, X.-P., Zhang, Z., and Li, W.-J. (2014). Bacillus tianshenii sp. nov., isolated from a marine sediment sample. https://doi.org/10.1099/ijs.0.062224-0

Jiménez, G., Urdiain, M., Cifuentes, A., López-López, A., Blanch, A. R., Tamames, J., Kämpfer, P., Kolstø, A.-B., Ramón, D., Martínez, J.F., Codoñer, F.M., and Rosselló-Móra, R. (2013). Description of Bacillus toyonensis sp. nov., a novel species of the Bacillus cereus group, and pairwise genome comparisons of the species of the group by means of ANI calculations. Systematic and Applied Microbiology, 36(6), 383-391. doi: 10.1016/J.SYAPM.2013.04.008

Joshi, J., Srisala, J., Truong, V. H., Chen, I.-T., Nuangsaeng, B., Suthienkul, O., Lo, C.F., Flegel, T.W., Sritunyaalucksana., K., and Thitamadee, S. (2014). Variation in Vibrio parahaemolyticus isolates from a single Thai shrimp farm experiencing an outbreak of

Page 156 of 208

acute hepatopancreatic necrosis disease (AHPND). Aquaculture, 428-429, 297-302. doi: 10.1016/J.AQUACULTURE.2014.03.030

Kadam, S. R., den Besten, H. M. W., van der Veen, S., Zwietering, M. H., Moezelaar, R., and Abee, T. (2013). Diversity assessment of Listeria monocytogenes biofilm formation: Impact of growth condition, serotype and strain origin. Int. J. Food Microbiology. 165, 259-264. doi:10.1016/J.IJFOODMICRO.2013.05.025.

Kao, W.-C., Kleinschroth, T., Nitschke, W., Baymann, F., Neehaul, Y., Hellwig, P., Richers, S., Vonck, J., Bott, M., and Hunte, C. (2016). The obligate respiratory supercomplex from Actinobacteria. Biochim. Biophys. Acta - Bioenerg. 1857, 1705-1714. doi:10.1016/J.BBABIO.2016.07.009.

Karunasagar, I., Pai, R., Malathi, G. R., and Karunasagar, I. (1994). Mass mortality of Penaeus monodon larvae due to antibiotic-resistant Vibrio harveyi infection. Aquaculture 128, 203- 209. doi:10.1016/0044-8486(94)90309-3.

Karuppiah, V., Sun, W., and Li, Z. (2016). Natural Products of Actinobacteria Derived from Marine Organisms. Stud. Nat. Prod. Chem. 48, 417-446. doi:10.1016/B978-0-444-63602- 7.00013-8.

Kaysner C. A., Depaola, A. Jr. (2004). Bacteriological Analytical Manual. Chapter 9. Vibrio. Silver Spring, MD: D.O.H.a.H. Services; Food U.S. and Drug Administration.

Kim, H.-S., Lee, S.-H., Byun, Y., and Park, H.-D. (2015). 6-Gingerol reduces Pseudomonas aeruginosa biofilm formation and virulence via quorum sensing inhibition. Sci. Rep. 5, 8656. doi:10.1038/srep08656.

Kim, S. K., Pang, Z., Seo, H. C., Cho, Y. R., Samocha, T., and Jang, I. K. (2014). Effect of bioflocs on growth and immune activity of Pacific white shrimp, Litopenaeus vannamei postlarvae. Aquac. Res. 45, 362-371. doi:10.1111/are.12319.

Kimes, N. E., Grim, C. J., Johnson, W. R., Hasan, N. A., Tall, B. D., Kothary, M. H., Kiss, H., Munk, A.C., Tapia, R., Green, L., Detter, C., Bruce, D.C., Brettin, T.S., Colwell, R.R., and Morris, P. J. (2012). Temperature regulation of virulence factors in the pathogen Vibrio coralliilyticus. The ISME Journal, 6(4), 835-846. doi: 10.1038/ismej.2011.154 Page 157 of 208

Klindworth, A., Pruesse, E., Schweer, T., Peplies, J., Quast, C., Horn, M., and Glöckner, F. O. (2013). Evaluation of general 16S ribosomal RNA gene PCR primers for classical and next- generation sequencing-based diversity studies. Nucleic Acids Res. 41, e1-e1. doi:10.1093/nar/gks808.

Koch, A., and Mizrahi, V. (2018). Mycobacterium tuberculosis. Trends Microbiol. 26, 555-556. doi:10.1016/J.TIM.2018.02.012.

Körner, S., Das, S. K., Veenstra, S., and Vermaat, J. E. (2001). The effect of pH variation at the ammonium/ammonia equilibrium in wastewater and its toxicity to Lemna gibba. Aquat. Bot. 71, 71-78. doi:10.1016/S0304-3770(01)00158-9.

Kurtz, J., and Franz, K. (2003). Evidence for memory in invertebrate immunity. Nature 425, 37- 38. doi:10.1038/425037a.

Lafferty, K. D., Harvell, C. D., Conrad, J. M., Friedman, C. S., Kent, M. L., Kuris, A. M., Powell, E.N., Rondeau D., and Saksida S.M. (2015). Infectious Diseases Affect Marine Fisheries and Aquaculture Economics. Ann. Rev. Mar. Sci. 7, 471-496. doi:10.1146/annurev-marine-010814-015646.

Lahens, N. F., Ricciotti, E., Smirnova, O., Toorens, E., Kim, E. J., Baruzzo, G., Hayer, K.E., Ganguly, T., and Schug, J., Grant, G.R . (2017). A comparison of Illumina and Ion Torrent sequencing platforms in the context of differential gene expression. BMC Genomics 18, 602. doi:10.1186/s12864-017-4011-0.

Lam, S. S., Ma, N. L., and Ambak, M. A. (2015). Biological nutrient removal by recirculating aquaponic system: Optimization of the dimension ratio between the hydroponic & rearing tank components. Int. Biodeterior. Biodegradation 102, 107-115. doi:10.1016/J.IBIOD.2015.03.012.

Lau, N.-S., Tan, W. R., Furusawa, G., and Amirul, A.-A. A. (2019). Complete genome sequence of the novel agarolytic Catenovulum-like strain CCB-QB4. Mar. Genomics 43, 50-53. doi:10.1016/J.MARGEN.2018.08.009.

Lavilla-Pitogo, C. ., Leaño, E. ., and Paner, M. . (1998). Mortalities of pond-cultured juvenile shrimp, Penaeus monodon, associated with dominance of luminescent vibrios in the rearing Page 158 of 208

environment. Aquaculture 164, 337-349. doi:10.1016/S0044-8486(98)00198-7.

Lavilla-Pitogo, C. R., Baticados, M. C. L., Cruz-Lacierda, E. R., and de la Pena, L. D. (1990). Occurrence of luminous bacterial disease of Penaeus monodon larvae in the Philippines. Aquaculture 91, 1-13. doi:10.1016/0044-8486(90)90173-K.

Learman, D. R., Henson, M. W., Thrash, J. C., Temperton, B., Brannock, P. M., Santos, S. R., Mahon, A.R., and Halanych, K.M. (2016). Biogeochemical and Microbial Variation across 5500 km of Antarctic Surface Sediment Implicates Organic Matter as a Driver of Benthic Community Structure. Front. Microbiol. 7, 284. doi:10.3389/fmicb.2016.00284.

Lee, C.-T., Chen, I.-T., Yang, Y.-T., Ko, T.-P., Huang, Y.-T., Huang, J.-Y., Huang, M.-F., Lin, S.-J., Chen, C.-Y., Lin, S.-S., Lightner, D.V., Wang, H.-C., Wang, A. H.-J., Wang, H.-C., Hor, L.-I., and Lo, C.-F. (2015). The opportunistic marine pathogen Vibrio parahaemolyticus becomes virulent by acquiring a plasmid that expresses a deadly toxin. Proceedings of the National Academy of Sciences, 112 (34), 10798-10803. doi: 10.1073/pnas.1503129112

Legendre and Gallagher (2001). Ecologically meaningful transformations for ordination of species data. Indian J. Tradit. Knowl. 129, 271-280. doi:10.1007/S004420100716.

Legendre, P., and Anderson, M. J. (1999). Distance-based redundancy analysis: Testing multispecies responses in multifactorial ecological experiments. Ecol. Monogr. 69, 1-24. doi:10.1890/0012-9615(1999)069[0001:DBRATM]2.0.CO;2.

Legendre, P., and Legendre, L. (1998). Numerical ecology. Elsevier.

Leij, F. A. A. M., Whipps, J. M. and Lynch, J. M. (1994). The use of colony development for the characterization of bacterial communities in soil and on roots. Microbial Ecology, 27, 81- 97.

Lemire, A., Goudenège, D., Versigny, T., Petton, B., Calteau, A., and Labreuche, Y., Roux, F.L. (2015). Populations, not clones, are the unit of vibrio pathogenesis in naturally infected oysters. ISME J. 9, 1523–1531. doi:10.1038/ismej.2014.233.

Lenth, R. V. (2016). Least-Squares Means: The R Package lsmeans. Journal of Statistical

Page 159 of 208

Software, 69(1), 1-33. doi: 10.1080/00031305.1980.10483031

Letchumanan, V., Chan, K.-G., and Lee, L.-H. (2014). Vibrio parahaemolyticus: a review on the pathogenesis, prevalence, and advance molecular identification techniques. Front. Microbiol. 5, 705. doi:10.3389/fmicb.2014.00705.

Li, T., Li, H., Gatesoupe, F.-J., She, R., Lin, Q., Yan, X., Li, J., and Li, X. (2017). Bacterial Signatures of “Red-Operculum” Disease in the Gut of Crucian Carp (Carassius auratus). Microb. Ecol. 74, 510-521. doi:10.1007/s00248-017-0967-1.

Li, Z., and Nair, S. K. (2012). Quorum sensing: how bacteria can coordinate activity and synchronize their response to external signals? Protein Sci. 21, 1403-17. doi:10.1002/pro.2132.

Lightner, D. V. (1993). Diseases of cultured penaeid shrimp. In CRC. Handbook of Mariculture, Crustacean Aquaculture, pp. 393-486 Edited by J. P. Mcvey. Boca Raton, FL: CRC Press.

Ligi, T., Oopkaup, K., Truu, M., Preem, J.-K., Nõlvak, H., Mitsch, W. J., Mander, Ü., and Truu, J. (2014). Characterization of bacterial communities in soil and sediment of a created riverine wetland complex using high-throughput 16S rRNA amplicon sequencing. Ecological Engineering, 72, 56-66. doi: 10.1016/J.ECOLENG.2013.09.007

Liltved, H., and Cripps, S. J. (1999). Removal of particle-associated bacteria by prefiltration and ultraviolet irradiation. Aquac. Res. 30, 445-450. doi:10.1046/j.1365-2109.1999.00349.x.

Litchman, E. (2010). Invisible invaders: non-pathogenic invasive microbes in aquatic and terrestrial ecosystems. Ecol. Lett. 13, 1560-1572. doi:10.1111/j.1461-0248.2010.01544.x.

Liu, G., Chater, K. F., Chandra, G., Niu, G., and Tan, H. (2013). Molecular Regulation of Antibiotic Biosynthesis in Streptomyces. Microbiol. Mol. Biol. Rev. 77, 112-143. doi:10.1128/MMBR.00054-12.

Liu, L., Xiao, J., Zhang, M., Zhu, W., Xia, X., Dai, X., Pan, Y., Yan, S., and Wang, Y. (2018a). A Vibrio owensii strain as the causative agent of AHPND in cultured shrimp, Litopenaeus vannamei. J. Invertebr. Pathol. doi:10.1016/J.JIP.2018.02.005.

Liu, Q., Han, Y., Wang, D., Wang, Q., Liu, X., Li, Y., Song, X., Wang, M., Jiang, Y., Meng, Z., Page 160 of 208

Shao, H., and McMinn, A. (2018b). Complete genomic sequence of bacteriophage J2-1: A novel Pseudoalteromonas phenolica phage isolated from the coastal water of Qingdao, China. Mar. Genomics 39, 15-18. doi:10.1016/J.MARGEN.2017.12.001.

Liu, P.-C., Lin, J.-Y., Hsiao, P.-T., and Lee, K.-K. (2004). Isolation and characterization of pathogenic Vibrio alginolyticus from diseased cobia Rachycentron canadum. Journal of Basic Microbiology, 44(1), 23-28. doi: 10.1002/jobm.200310316

Liu, S., Ren, H., Shen, L., Lou, L., Tian, G., and Zheng, P., Hu, B. (2015). pH levels drive bacterial community structure in sediments of the Qiantang River as determined by 454 pyrosequencing. Front. Microbiol. 6, 285. doi:10.3389/fmicb.2015.00285.

Lou, J., Yang, L., Wang, H., Wu, L., and Xu, J. (2018). Assessing soil bacterial community and dynamics by integrated high-throughput absolute abundance quantification. PeerJ 6, e4514. doi:10.7717/peerj.4514.

Madigan et al. (2006). Biology of Microorganism. Thirteenth Edition.

Martínez Cruz, P., Ibáñez, A. L., Monroy Hermosillo, O. A., and Ramírez Saad, H. C. (2012). Use of Probiotics in Aquaculture. ISRN Microbiol. 2012, 1-13. doi:10.5402/2012/916845.

Martínez-Córdova, L. R., Martínez-Porchas, M., Emerenciano, M. G. C., Miranda-Baeza, A., and Gollas-Galván, T. (2016). From microbes to fish the next revolution in food production. Crit. Rev. Biotechnol. 8551, 1-9. doi:10.3109/07388551.2016.1144043.

Martínez-Córdova, L. R., Martínez-Porchas, M., Porchas-Cornejo, M. A., Gollas-Galván, T., Scheuren-Acevedo, S., Arvayo, M. A., Lopez-Elias, J.A., and Lopez-Elias, M.A. (2017). Bacterial diversity studied by next-generation sequencing in a mature phototrophic Navicula sp-based biofilm promoted into a shrimp culture system. Aquac. Res. 48, 2047- 2054. doi:10.1111/are.13037.

Martínez-Porchas, M. (2017). What is yet to be known About Microbial-Based Systems for Aquaculture? J. Aquac. Mar. Biol. 5, 2016-2018. doi:10.15406/jamb.2016.05.00108.

Masser, M.P., and Rakocy, J., Losordo, T.M. (1999). Recirculating Aquaculture Tank Production Systems: Management of Recirculating Systems. South. Reg. Aquac. Center,

Page 161 of 208

Publication no. 452, 12pp.

Matsuyama, H., Minami, H., Kasahara, H., Kato, Y., Murayama, M., and Yumoto, I. (2013). Pseudoalteromonas arabiensis sp. nov., a marine polysaccharide-producing bacterium. International journal of systematic and evolutionary microbiology, 63(Pt 5), 1805-1809. doi: 10.1099/ijs.0.043604-0

McFall-Ngai, M., Hadfield, M. G., Bosch, T. C. G., Carey, H. V., Domazet-Lošo, T., Douglas, A. E., Dubilier, N., Eberi, G., Fukami, T., Gilbert, S.F., Hentschel, U., King, N., Kjelleberg, S., Knoll, A.H., Kremer, N., mazmanian, S.K., Metcalf, J.L., Nealson, K., Pierce, N.E., Rawls, J.F., Reid, A., Ruby, E.G., Rumpho, M., Sanders, J.G., and Tautz, D., Wernegreen, J.J. (2013). Animals in a bacterial world, a new imperative for the life sciences. Proc. Natl. Acad. Sci. 110, 3229-3236. doi:10.1073/pnas.1218525110.

McIntosh, R. P. (2000). Changing paradigms in shrimp farming: low protein feeds and feeding strategies. Global Aquaculture. Advocate, 3 44-50.

Medeiros, P. M., Seidel, M., Dittmar, T., Whitman, W. B., and Moran, M. A. (2015). Drought- induced variability in dissolved organic matter composition in a marsh-dominated estuary. Geophys. Res. Lett. 42, 6446-6453. doi:10.1002/2015GL064653.

Michaud, L., Blancheton, J. P., Bruni, V., and Piedrahita, R. (2006). Effect of particulate organic carbon on heterotrophic bacterial populations and nitrification efficiency in biological filters. Aquac. Eng. 34, 224-233. doi:10.1016/J.AQUAENG.2005.07.005.

Michaud, L., Giudice, A. Lo, Troussellier, M., Smedile, F., Bruni, V., and Blancheton, J. P. (2009). Phylogenetic characterization of the heterotrophic bacterial communities inhabiting a marine recirculating aquaculture system. doi:10.1111/j.1365-2672.2009.04378.x.

Mikkelsen, L. H., Gotfredsen, A. K., Agerbxk, M. L., Nielsen, P. H., and Keiding, K. (1996). Effects of colloidal stability on clarification and dewatering of activated sludge. Water Science and Technology, 34(3-4), 449-457. doi: 10.1016/0273-1223(96)00611-7

Miller, R. A., Beno, S. M., Kent, D. J., Carroll, L. M., Martin, N. H., Boor, K. J., and Kovac, J. (2016). Bacillus wiedmannii sp. nov., a psychrotolerant and cytotoxic Bacillus cereus group species isolated from dairy foods and dairy environments. International Journal of Page 162 of 208

Systematic and Evolutionary Microbiology, 66(11), 4744-4753. doi: 10.1099/ijsem.0.001421

Mitchell, J. G., and Kogure, K. (2006). Bacterial motility: links to the environment and a driving force for microbial physics. FEMS Microbiol. Ecol. 55, 3-16. doi:10.1111/j.1574- 6941.2005.00003.x.

Miura, N., Motone, K., Takagi, T., Aburaya, S., Watanabe, S., Aoki, W., and Ueda, M. (2018). Ruegeria sp. Strains Isolated from the Reef-Building Coral Galaxea fascicularis Inhibit Growth of the Temperature-Dependent Pathogen Vibrio coralliilyticus. Mar. Biotechnol. doi:10.1007/s10126-018-9853-1.

Mizan, M. F. R., Ashrafudoulla, M., Sadekuzzaman, M., Kang, I., and Ha, S.-D. (2018). Effects of NaCl, glucose, and their combinations on biofilm formation on black tiger shrimp (Penaeus monodon) surfaces by Vibrio parahaemolyticus. Food Control 89, 203-209. doi:10.1016/J.FOODCONT.2017.12.004.

Mohanty, R. K., Ambast, S. K., Panigrahi, P., and Mandal, K. G. (2018). Water quality suitability and water use indices: Useful management tools in coastal aquaculture of Litopenaeus vannamei. Aquaculture 485, 210-219. doi:10.1016/j.aquaculture.2017.11.048.

Moriarty, D. J. W. (1997). The role of microorganisms in aquaculture ponds. Aquaculture, 151, 333-349.

Nadkarni, M. A., Hunter, N., Jacques, N. A., and Martin, F. E. (2002). Determination of bacterial load by real-time PCR using a broad-range (universal) probe and primers set. Microbiology, 148(1), 257-266. doi: 10.1099/00221287-148-1-257

Najdegerami, E. H., Bakhshi, F., and Lakani, F. B. (2016). Effects of biofloc on growth performance, digestive enzyme activities and liver histology of common carp (Cyprinus carpio L.) fingerlings in zero-water exchange system. Fish Physiol. Biochem. 42, 457-465. doi:10.1007/s10695-015-0151-9.

Nakamura, L. K. (1998). Bacillus pseudomycoides sp. nov. International Journal of Systematic Bacteriology, 48(103), 1-1.

Page 163 of 208

Nelson, W. C., and Stegen, J. C. (2015). The reduced genomes of Parcubacteria (OD1) contain signatures of a symbiotic lifestyle. Front. Microbiol. 6, 713. doi:10.3389/fmicb.2015.00713.

Nercessian, O., Noyes, E., Kalyuzhnaya, M. G., Lidstrom, M. E., and Chistoserdova, L. (2005). Bacterial Populations Active in Metabolism of C1 Compounds in the Sediment of Lake Washington, a Freshwater Lake. Appl. Environ. Microbiol. 71, 6885-6899. doi:10.1128/AEM.71.11.6885-6899.2005.

Nevejan, N., De Schryver, P., Wille, M., Dierckens, K., Baruah, K., and Van Stappen, G. (2018). Bacteria as food in aquaculture: do they make a difference? Rev. Aquac. 10, 180-212. doi:10.1111/raq.12155.

Nguyen, T. T., and Landfald, B. (2015). Polar front associated variation in prokaryotic community structure in Arctic shelf seafloor. Front. Microbiol. doi:10.3389/fmicb.2015.00017.

Ninawe, A. S., and Selvin, J. (2009). Probiotics in shrimp aquaculture: Avenues and challenges. Crit. Rev. Microbiol. 35, 43-66. doi:10.1080/10408410802667202.

Nissui pharmaceutical Co., L. Compact Dry “ Nissui ” VP For Viblio parahaemolyticus. 1-4.

Oksanen, A. J., Blanchet, F. G., Friendly, M., Kindt, R., Legendre, P., Mcglinn, D., Minchin, P.R., O'Hara, R.B., Simpson, G.L., Solymos, P., Stevens, M.H.H., Szoecs, E., and Wagner, H. (2013). Vegan: Community Ecology Package. R package version 2.0-8.

Olson, M. E., Ceri, H., Morck, D. W., Buret, A. G., and Read, R. R. (2002). Biofilm bacteria: formation and comparative susceptibility to antibiotics. Can. J. Vet. Res. 66, 86-92.

Otoshi, C. A., Tang, L. A., Moss, D. R., Arce, S. M., Holl, C. M. and Moss, S. (2009). Performance of pacific white shrimp, Penaeus (Litopenaeus) vannamei, cultured in biosecure, super-intensive, recirculating aquaculture systems. In: Browdy, C.L., Jory, D.E. (Eds.), The Rising Tide, Proceedings of the Special Session on Sustainable Shrimp Farming. World Aquaculture Society, Baton Rouge, LA, USA, pp. 244-254.

Oueriaghli, N., González-Domenech, C. M., Martínez-Checa, F., Muyzer, G., Ventosa, A., Quesada, E., and Béjar, V. (2014). Diversity and distribution of Halomonas in Rambla

Page 164 of 208

Salada, a hypersaline environment in the southeast of Spain. FEMS Microbiol. Ecol. 87, 460-474. doi:10.1111/1574-6941.12237.

Overman, P. R. (2006). Biofilm: a new view of plaque. J Contempt Dent Pr., 1:1-11.

Pacheco-Vega, J. M., Cadena-Roa, M. A., Leyva-Flores, J. A., Zavala-Leal, O. I., Pérez-Bravo, E., and Ruiz-Velazco, J. M. J. (2018). Effect of isolated bacteria and microalgae on the biofloc characteristics in the Pacific white shrimp culture. Aquac. Reports 11, 24-30. doi:10.1016/J.AQREP.2018.05.003.

Panjaitan, P. (2010). Shrimp culture of Penaeus monodon with zero water exchange model (ZWEM) using molasses. Journal of Coastal Development, 14(1), 1410-5217.

Peña, L., Cabillon, N., Catedral, D., Amar, E., Usero, R., Monotilla, W., CAlpe, A.T., Fernandez, D.DG., and Saloma, C.P. (2015). Acute hepatopancreatic necrosis disease (AHPND) outbreaks in Penaeus vannamei and P. monodon cultured in the Philippines. Dis. Aquat. Organ. 116, 251-254. doi:10.3354/dao02919.

Peters, L., Spatharis, S., Dario, M. A., Dwyer, T., Roca, I. J. T., Kintner, A., Hanssen, Ø.K., and Llewellyn, M.S., Praebel, K. (2018). Environmental DNA: A New Low-Cost Monitoring Tool for Pathogens in Salmonid Aquaculture. Front. Microbiol. 9, 3009. doi:10.3389/fmicb.2018.03009.

Pimm, S. L. (1984). The complexity and stability of ecosystems. Nature 307, 321-326. doi:10.1038/307321a0.

Pinheiro J, Bates D, DebRoy S, S. D. and R. C. T. (2018). nlme: Linear and Nonlinear Mixed Effects Model.

Prachumwat, A., Taengchaiyaphum, S., Mungkongwongsiri, N., Aldama-Cano, D. J., Flegel, T. W., and Sritunyalucksana, K. (2019). Update on early mortality syndrome/acute hepatopancreatic necrosis disease by April 2018. J. World Aquac. Soc. 50, 5-17. doi:10.1111/jwas.12559.

Prangnell, D.I., Castro, L.F., Ali, A.S., Browdy, C.L., Zimba, P.V., Laramore, S.E., and Samocha, T.M. (2016). Some limiting factors in superintensive production of juvenile

Page 165 of 208

Pacific white shrimp, Litopenaeus vannamei, in no-water-exchange, biofloc-dominated systems. Journal of the World Aquaculture Society, 47 (3), 396-413. doi: 10.1111/jwas.12275

Qin, Y., Hou, J., Deng, M., Liu, Q., Wu, C., and Ji, Y., He, X. (2016). Bacterial abundance and diversity in pond water supplied with different feeds. Sci. Rep. 6, 1-13. doi:10.1038/srep35232.

Quast, C., Pruesse, E., Yilmaz, P., Gerken, J., Schweer, T., Yarza, P., Peplies, J., and Glöckner, F.O. (2013). The SILVA ribosomal RNA gene database project: improved data processing and web-based tools. Nucleic Acids Res. 41, D590-D596. doi:10.1093/nar/gks1219.

Raghunath, P. (2015). Roles of thermostable direct hemolysin (TDH) and TDH-related hemolysin (TRH) in Vibrio parahaemolyticus. Front. Microbiol. 5, 805. doi:10.3389/fmicb.2014.00805.

Rahman, M. M., Nagelkerke, L. A. J., Verdegem, M. C. J., Wahab, M. A., and Verreth, J. A. J. (2008). Relationships among water quality, food resources, fish diet and fish growth in polyculture ponds: A multivariate approach. Aquaculture 275, 108-115. doi:10.1016/J.AQUACULTURE.2008.01.027.

Ray, A. J., Lewis, B. L., Browdy, C. L., and Leffler, J. W. (2010). Suspended solids removal to improve shrimp (Litopenaeus vannamei) production and an evaluation of a plant-based feed in minimal-exchange, superintensive culture systems. Aquaculture 299, 89-98. doi:10.1016/j.aquaculture.2009.11.021.

Reddy, G., Raghavan, P., Sarita, N., Prakash, J., Nagesh, N., Delille, D., and Shivaji, S. (2003). Halomonas glaciei sp. nov. isolated from fast ice of Adelie Land, Antarctica. Extremophiles 7, 55-61. doi:10.1007/s00792-002-0295-2.

Reen, F. J., Almagro-Moreno, S., Ussery, D., and Boyd, E. F. (2006). The genomic code: inferring Vibrionaceae niche specialization. Nat. Rev. Microbiol. 4, 697-704. doi:10.1038/nrmicro1476.

Reichardt, W. T., Reyes, J. M., Pueblos, M. J., and Lluisma, A. O. (2013). Impact of milk fish farming in the tropics on potentially pathogenic vibrios. Mar. Pollut. Bull. 77, 325-332. Page 166 of 208

doi:10.1016/j.marpolbul.2013.09.018.

Restrepo, L., Bayot, B., Arciniegas, S., Bajaña, L., Betancourt, I., and Panchana, F., Muñoz, A.R. (2018). PirVP genes causing AHPND identified in a new Vibrio species (Vibrio punensis) within the commensal Orientalis clade. Sci. Rep. 8, 1-14. doi:10.1038/s41598- 018-30903-x.

Rhee, S.-K., Liu, X., Wu, L., Chong, S. C., Wan, X., and Zhou, J. (2004). Detection of Genes Involved in Biodegradation and Biotransformation in Microbial Communities by Using 50- Mer Oligonucleotide Microarrays. Appl. Environ. Microbiol. 70, 4303-4317. doi:10.1128/AEM.70.7.4303-4317.2004.

Richards, G. P., Watson, M. A., Needleman, D. S., Uknalis, J., Boyd, E. F., and Fay, J. P. (2017). Mechanisms for Pseudoalteromonas piscicida-Induced Killing of Vibrios and Other Bacterial Pathogens. Appl. Environ. Microbiol. 83, e00175-17. doi:10.1128/AEM.00175- 17.

Rivers, A. R., Burns, A. S., Chan, L.-K., and Moran, M. A. (2016). Experimental Identification of Small Non-Coding RNAs in the Model Marine Bacterium Ruegeria pomeroyi DSS-3. Front. Microbiol. 7, 380. doi:10.3389/fmicb.2016.00380.

Rong, J.-C., Cha, Q.-Q., Ren, X.-B., Xie, B.-B., and Song, X.-Y. (2018). Complete genome sequence of Pseudoalteromonas espejiana DSM9414T, an amino-acid-requiring strain from seawater. Mar. Genomics 38, 21-23. doi:10.1016/J.MARGEN.2017.08.006.

Roque, A., Turnbull, J., Escalante, G., Gomez-Gil, B., and Alday-Sanz, M. (1998). Development of a bath challenge for the marine shrimp Penaeus vannamei Boone, 1931. Aquaculture 169, 283-290. doi:10.1016/S0044-8486(98)00370-6.

Rurangwa, E., and Verdegem, M. C. J. (2015). Microorganisms in recirculating aquaculture systems and their management. doi:10.1111/raq.12057.

Rykiel, E. J. (1985). Towards a definition of ecological disturbance. Austral Ecol. 10, 361-365. doi:10.1111/j.1442-9993.1985.tb00897.x.

Salehizadeh, H., and Shojaosadati, S. (2001). Extracellular biopolymeric flocculants: Recent

Page 167 of 208

trends and biotechnological importance. Biotechnol. Adv. 19, 371-385. doi:10.1016/S0734- 9750(01)00071-4.

Salipante, S. J., Kawashima, T., Rosenthal, C., Hoogestraat, D. R., Cummings, L. A., Sengupta, D. J., Harkins, T.T., Cookson, B.T., and Hoffman, N.G. (2014). Performance Comparison of Illumina and Ion Torrent Next-Generation Sequencing Platforms for 16S rRNA-Based Bacterial Community Profiling. Appl. Environ. Microbiol. 80, 7583-7591. doi:10.1128/AEM.02206-14.

Sandlund, N., and Bergh, Ø. (2008). Screening and characterisation of potentially pathogenic bacteria associated with Atlantic cod Gadus morhua larvae: bath challenge trials using a multidish system. Dis. Aquat. Organ. 81, 203-217. doi:10.3354/dao01934.

Sanjuán, E., Fouz, B., Oliver, J. D., and Amaro, C. (2009). Evaluation of genotypic and phenotypic methods to distinguish clinical from environmental Vibrio vulnificus strains. Appl. Environ. Microbiol. 75, 1604-13. doi:10.1128/AEM.01594-08.

Sawabe, T., Ogura, Y., Matsumura, Y., Feng, G., Amin, A. R., Mino, S., nakagawa, S., Sawabe, T., Kumar, R., Fukui, Y., Satomi, M., Matsushima, R., Thompson, F.L., Gomez-Gil, B., Christen, R., Maruyama, F., and Kurokawa, K., Hayashi, T. (2013). Updating the Vibrio clades defined by multilocus sequence phylogeny: proposal of eight new clades, and the description of Vibrio tritonius sp. nov. Front. Microbiol. 4, 414. doi:10.3389/fmicb.2013.00414.

Schauer, R., Bienhold, C., Ramette, A., and Harder, J. (2010). Bacterial diversity and biogeography in deep-sea surface sediments of the South Atlantic Ocean. ISME J. 4, 159- 170. doi:10.1038/ismej.2009.106.

Schneider, O., Sereti, V., Eding, E. H., and Verreth, J. A. J. (2005). Analysis of nutrient flows in integrated intensive aquaculture systems. Aquac. Eng. 32, 379-401. doi:10.1016/J.AQUAENG.2004.09.001.

Schveitzer, R., Arantes, R., Costódio, P. F. S., do Espírito Santo, C. M., Arana, L. V., Seiffert, W. Q., and Andreatta, E. R. (2013). Effect of different biofloc levels on microbial activity, water quality and performance of Litopenaeus vannamei in a tank system operated with no

Page 168 of 208

water exchange. Aquacultural Engineering, 56, 59-70. doi: 10.1016/J.AQUAENG.2013.04.006

Shade, A., Peter, H., Allison, S. D., Baho, D. L., Berga, M., Bürgmann, H., Huber, D.H., Langenheder, S., Lennon, J.T., Martiny, J.B.H., Matulich, K.L., and Schmidt, T.M., Handelsman, J. (2012). Fundamentals of Microbial Community Resistance and Resilience. Front. Microbiol. 3, 417. doi:10.3389/fmicb.2012.00417.

Sharma, P., Dutta, J., and Thakur, D. (2018). Future Prospects of Actinobacteria in Health and Industry. New Futur. Dev. Microb. Biotechnol. Bioeng., 305-324. doi:10.1016/B978-0-444- 63994-3.00021-7.

Shi, M. J., Du, W. J., Wang, C., Chen, G. J., and Du, Z. J. (2017). Catenovulum sediminis sp. Nov., isolated from coastal sediment. Int. J. Syst. Evol. Microbiol. 67, 3894-3898. doi:10.1099/ijsem.0.002219.

Singer, G. A., Fasching, C., Wilhelm, L., Niggemann, J., Steier, P., and Dittmar, T., Battin, T.J. (2012). Biogeochemically diverse organic matter in Alpine glaciers and its downstream fate. Nat. Geosci. 5, 710-714. doi:10.1038/ngeo1581.

Skjermo, J., and Vadstein, O. (1999). Techniques for microbial control in the intensive rearing of marine larvae. Aquaculture 177, 333-343. doi:10.1016/S0044-8486(99)00096-4.

Skjermo, J., Bakke, I., Dahle, S. W., and Vadstein, O. (2015). Probiotic strains introduced through live feed and rearing water have low colonizing success in developing Atlantic cod larvae. Aquaculture 438, 17-23. doi:10.1016/J.AQUACULTURE.2014.12.027.

Skjermo, J., Salvesen, I., Øie, G., Olsen, Y., and Vadstein, O. (1997). Microbially matured water: a technique for selection of a non-opportunistic bacterial flora in water that may improve performance of marine larvae. Aquaculture International, 5(1), 13-28. https://doi.org/10.1007/BF02764784

Smith, V. J., Brown, J. H., and Hauton, C. (2003). Immunostimulation in crustaceans: does it really protect against infection? Fish Shellfish Immunol. 15, 71-90.

Snow, G. (2016). TeachingDemos: Demonstrations for Teaching and Learning. Available at:

Page 169 of 208

https://cran.r-project.org/web/packages/TeachingDemos/TeachingDemos.pdf

Sogin, M. L., Morrison, H. G., Huber, J. A., Mark Welch, D., Huse, S. M., Neal, P. R., Arrieta, J.M., and Herndl, G.J. (2006). Microbial diversity in the deep sea and the under explored “rare biosphere”. Proc. Natl. Acad. Sci. U. S. A. 103, 12115-20. doi:10.1073/pnas.0605127103.

Soowannayan, C., Boonmee, S., Puckcharoen, S., Anatamsombat, T., Yatip, P., Ng, W.-K., Thitamadee, S., Tuchinda, P., Munyoo, B., Chabang, N., Nuangsaeng, M.S, and Withyachumnarnkul, B. (2019a). Ginger and its component shogaol inhibit Vibrio biofilm formation in vitro and orally protect shrimp against acute hepatopancreatic necrosis disease (AHPND). Aquaculture 504, 139-147. doi:10.1016/J.AQUACULTURE.2019.02.007.

Soowannayan, C., Chandra Teja, D. N., Yatip, P., Mazumder, F. Y., Krataitong, K., Unagul, P., Suetrong, S., Preedanon, S., Klaysuban, A., Sakayaroj, J., and Sangtiean, T. (2019b). Vibrio biofilm inhibitors screened from marine fungi protect shrimp against acute hepatopancreatic necrosis disease (AHPND). Aquaculture 499, 1-8. doi:10.1016/J.AQUACULTURE.2018.09.004.

Soto-Alcalá, J., Álvarez-ruiz, P., Audelo-Naranjo, J.M., Esparza-Leal, H.M., Luis-Villasenor, I.E., Estrada-Godinez, J.A., Luna-Gozalez, A., Gamez-Jmenez, C., and Diarte-Plata, G. (2019). Transcriptional response of immune-related genes in Litopenaeus vannamei post- larvae cultured in recirculating aquaculture systems with and without biofloc. 209-225.

Soto-Rodriguez, S. A., Gomez-Gil, B., Lozano-Olvera, R., Betancourt-Lozano, M., and Morales- Covarrubias, M. S. (2015). Field and experimental evidence of Vibrio parahaemolyticus as the causative agent of acute hepatopancreatic necrosis disease of cultured shrimp (Litopenaeus vannamei) in Northwestern Mexico. Appl. Environ. Microbiol. 81, 1689-99. doi:10.1128/AEM.03610-14.

Souza, D. M., Suita, S. M., Leite, F. P. L., Romano, L. A., Wasielesky, W., and Ballester, E. L. C. (2012). The use of probiotics during the nursery rearing of the pink shrimp Farfantepenaeus brasiliensis (Latreille, 1817) in a zero exchange system. Aquac. Res. 43, 1828-1837. doi:10.1111/j.1365-2109.2011.02992.x.

Page 170 of 208

Stocker, R. (2012). Marine Microbes See a Sea of Gradients. Science (80). 338, 628-633. doi:10.1126/science.1208929.

Stratford, J. P., Woodley, M. A., and Park, S. (2013). Variation in the Morphology of Bacillus mycoides Due to Applied Force and Substrate Structure. PLoS ONE, 8(12), e81549. doi: 10.1371/journal.pone.0081549

Su, Y., Tang, K., Liu, J., Wang, Y., Zheng, Y., and Zhang, X.-H. (2019). Quorum Sensing System of Ruegeria mobilis Rm01 Controls Lipase and Biofilm Formation. Front. Microbiol. 9, 3304. doi:10.3389/fmicb.2018.03304.

Subramanian, K., Balaraman, D., Balachandran, D.N., Thirunavukarasu, S., Gopal, S., Renuka, P.S., and Kumarappan, A. (2014). Immune response of shrimp (Penaeus monodon) against V. furnissii pathogen. J. Coast. Life Med. 2, 281–286. doi:10.12980/JCLM.2.201414J14.

Sun, F., Hu, J., Zhou, Y., Mei, R., Wang, C., He, Y., and Wu, W. (2018). High efficient alternating anaerobic/aerobic process for polyester resin wastewater treatment: Performance and microbial community structure. Biochem. Eng. J. 138, 121-130. doi:10.1016/J.BEJ.2018.07.005.

Swings, J., Balcaen, A., Heyrman, J., and De Vos, P. (2002). Halomonas muralis sp. nov., isolated from microbial biofilms colonizing the walls and murals of the Saint-Catherine chapel (Castle Herberstein, Austria). Int. J. Syst. Evol. Microbiol. 52, 2049-2054. doi:10.1099/00207713-52-6-2049.

Tacon, A. G. J., Cody, J. J., Conquest, L. D., Divakaran, S., Forster, I. P., and Decamp, O. E. (2002). Effect of culture system on the nutrition and growth performance of Pacific white shrimp Litopenaeus vannamei (Boone) fed different diets. Aquaculture Nutrition, 8, 121- 137.

Tajkarimi, M. (2007). Bacillus cereus. Available at: https://www.cdfa.ca.gov/ahfss/Animal-

Health/PHR250/2007/25007BcerMH-2-.pdf

Tan, B., Mai, K., Zheng, S., Zhou, Q., Liu, L., and Yu, Y. (2005). Replacement of fish meal by meat and bone meal in practical diets for the white shrimp Litopenaeus vannamai (Boone). Aquac. Res. 36, 439-444. doi:10.1111/j.1365-2109.2005.01223.x. Page 171 of 208

Team, R. C. (2016). R: The R Project for Statistical Computing. Available at: https://www.r- project.org/

Therneau T.M. and Grambsch, P.M. (2000). Modeling Survival Data: Extending the Cox Model. Springer, New York. ISBN 0-387-98784-3.

Thitamadee, S., Prachumwat, A., Srisala, J., Jaroenlak, P., Salachan, P. V., Sritunyalucksana, K., Flegel, T.W., and Itsathitphaisarn, O. (2016). Review of current disease threats for cultivated penaeid shrimp in Asia. Aquaculture 452, 69-87. doi:10.1016/J.AQUACULTURE.2015.10.028.

Timmons, M.B., Ebeling, J.M., Wheaton, F.W., Summerfelt, S.T., Vinci, B.J., 2002. Recirculating aquaculture systems, 2nd edition. NRAC Publication, vol. 01-002.

Tran, L. H. (2013). Determination, Characterization, and Control Measures of the Agent Causing Early Mortality Syndrome (EMS) also known as Acute Hepatopancreatic Necrosis Syndrome (AHPNS) in Farmed Penaeid Shrimp. PhD thesis.

Tran, L., Nunan, L., Redman, R. M., Mohney, L. L., Pantoja, C. R., Fitzsimmons, K., and Lightner, D.V. (2013). Determination of the infectious nature of the agent of acute hepatopancreatic necrosis syndrome affecting penaeid shrimp. Dis. Aquat. Organ. 105, 45- 55. doi:10.3354/dao02621.

Tran H.L, Kevin. F. and Donald. V. L. (2014). AHPND/EMS: From the academic science perspective to the production point of view. AQUA Culture Asia Pacific Magazine. 10 (2).

UNESCO (United Nations Educational, Scientific and Cultural Organization). (1983). Chemmical methods for use in marine environmental monitoring. In Paris: Intergovernmental Oceanographic Commission (p. 53 pp).

Vadstein O, Øie G, Olsen Y, Salvesen I, Skjermo J and G, S. k.-B. (1993) A strategy to obtain microbial control during larval development of marine fish. In: Reinertsen H, Dahle LA, Jørgensen L, Tvinnereim K (eds). Fish Farming Technology, pp. 69-75. A.A. Balkema Publishers, Rotterdam.

Vadstein, O., Attramadal, K. J. K., Bakke, I., and Olsen, Y. (2018b). K-Selection as Microbial

Page 172 of 208

Community Management Strategy: A Method for Improved Viability of Larvae in Aquaculture. Front. Microbiol. 9, 2730. doi:10.3389/fmicb.2018.02730.

Vadstein, O., Attramadal, K. J. K., Bakke, I., Forberg, T., Olsen, Y., Verdegem, M., Giatsis, C., Skjermo, J., Aasen, M.I., Gatesoupe, J.F., Dierckens, K., Sorgeloos, P., and Bossier, P. (2018a). Managing the Microbial Community of Marine Fish Larvae: A Holistic Perspective for Larviculture. Front. Microbiol. 9, 1820. doi:10.3389/fmicb.2018.01820.

Vadstein, O., Bergh, Ø., Gatesoupe, F.-J., Galindo-Villegas, J., Mulero, V., Picchietti, S., Scapigliati, G., Makridis, P., Olsen, Y., Dierckens, K., Defoirdt, T., Boon, N., De Schryver, P., and Bossier, P. (2013). Microbiology and immunology of fish larvae. Reviews in Aquaculture, 5(s1), S1-S25. doi: 10.1111/j.1753-5131.2012.01082.x

Vadstein, O., Mo, T. A., and Bergh, . (2004). “Microbial Interactions, Prophylaxis and Diseases” in Culture of Cold-Water Marine Fish (Oxford, UK: Blackwell Publishing Ltd), 28-72. doi:10.1002/9780470995617.ch3. van Rijn, J. (1996). The potential for integrated biological treatment systems in recirculating fish culture-A review. Aquaculture, 139(3-4), 181–201. doi: 10.1016/0044-8486(95)01151-X van Rijn, J., Tal, Y., and Schreier, H. J. (2006). Denitrification in recirculating systems: Theory and applications. Aquacultural Engineering, 34(3), 364-376. doi: 10.1016/j.aquaeng.2005.04.004

Van Wyk, P. (2006). Production of Litopenaeus vannamei in Recirculating Aquaculture Systems: Management and Design Considerations. Int. Conf. Recirc. Aquac. 6, 38-47.

Verschuere, L., Rombaut, G., Sorgeloos, P., and Verstraete, W. (2000). Probiotic bacteria as biological control agents in aquaculture. Microbiol. Mol. Biol. Rev. 64, 655-71.

Vouga, M., Kebbi-Beghdadi, C., Liénard, J., Baskin, L., Baud, D., and Greub, G. (2018). What is the true clinical relevance of Simkania negevensis and other emerging Chlamydiales members? New Microbes New Infect. 23, 1-5. doi:10.1016/J.NMNI.2018.01.001.

Vreeland, R. H., Litchfield, C. D., Martin, E. L., and Elliot, E. (1980). Halomonas elongata, a New Genus and Species of Extremely Salt-Tolerant Bacteria. Int. J. Syst. Bacteriol. 30,

Page 173 of 208

485-495. doi:10.1099/00207713-30-2-485.

Wang, H., Wang, C., Tang, Y., Sun, B., Huang, J., and Song, X. (2018a). Pseudoalteromonas probiotics as potential biocontrol agents improve the survival of Penaeus vannamei challenged with acute hepatopancreatic necrosis disease (AHPND)-causing Vibrio parahaemolyticus. Aquaculture 494, 30-36. doi:10.1016/j.aquaculture.2018.05.020.

Wang, A. R., Ran, C., Ringø, E., and Zhou, Z. G. (2018b). Progress in fish gastrointestinal microbiota research. Rev. Aquac. 10, 626-640. doi:10.1111/raq.12191.

Wang, Y.-C., Chang, P.-S., and Chen, H.-Y. (2008). Differential time-series expression of immune-related genes of Pacific white shrimp Litopenaeus vannamei in response to dietary inclusion of β-1,3-glucan. Fish & Shellfish Immunology, 24(1), 113-121. doi: 10.1016/J.FSI.2007.09.008

Warnes, G.R., Bolker, B., Bonebakker, B., Gentleman, R., Liaw, W.H.A., Lumley, T., Maechler, M., Magnusson, A., Moeller, S., Schwartz, M., and Venables, B. (2016). gplots: Various R Programming Tools for Plotting Data. Available at: https://CRAN.R- project.org/package=gplots

Wasielesky, W., Atwood, H., Stokes, A., and Browdy, C. L. (2006). Effect of natural production in a zero exchange suspended microbial floc based super-intensive culture system for white shrimp Litopenaeus vannamei. Aquaculture 258, 396-403. doi:10.1016/J.AQUACULTURE.2006.04.030.

Wesseling, W., Wittka, S., Kroll, S., Soltmann, C., Kegler, P., Kunzmann, A., Riss, H.W., and Lohmeyer, M. (2015). Functionalised ceramic spawning tiles with probiotic Pseudoalteromonas biofilms designed for clownfish aquaculture. Aquaculture, 446, 57-66. doi: 10.1016/J.AQUACULTURE.2015.04.017

WHO. (1999). Food safety issues associated with products from aquaculture. WHO.

Williams, K. P., Gillespie, J. J., Sobral, B. W. S., Nordberg, E. K., Snyder, E. E., Shallom, J. M., and Dickerman, A.W. (2010). Phylogeny of gammaproteobacteria. J. Bacteriol. 192, 2305- 14. doi:10.1128/JB.01480-09.

Page 174 of 208

Wilson, K. (2001). Preparation of Genomic DNA from Bacteria. Curr. Protoc. Mol. Biol. 56, 2.4.1-2.4.5. doi:10.1002/0471142727.mb0204s56.

Wit and Bouvier (2006). Management of ventilatory failure. Environ. Microbiol. 8, 755-758. doi:10.1111/j.1462-2920.2006.01017.

Wrighton, K. C., Castelle, C. J., Wilkins, M. J., Hug, L. A., Sharon, I., Thomas, B. C., Handley, K.M., Mullin, S.W., Nicora, C.D., Singh, A., Lipton, M.S., Long, P.E., Williams, K.H., and Banfield, J.F. (2014). Metabolic interdependencies between phylogenetically novel fermenters and respiratory organisms in an unconfined aquifer. ISME J. 8, 1452-1463. doi:10.1038/ismej.2013.249.

Wu, Y.-H., Cheng, H., Xu, L., Jin, X.-B., Wang, C.-S., and Xu, X.-W. (2017). Physiological and genomic features of a novel violacein-producing bacterium isolated from surface seawater. doi: 10.1371/journal.pone.0179997

Xi, D., Liu, C., and Su, Y.-C. (2012). Effects of green tea extract on reducing Vibrio parahaemolyticus and increasing shelf life of oyster meats. Food Control 25, 368-373. doi:10.1016/J.FOODCONT.2011.11.002.

Xiong, J., Zhu, J., Wang, K., Wang, X., Ye, X., Liu, L., Zhao, Q., Hou, M., Qiuqian, L., and Zhang, D. (2014). The Temporal Scaling of Bacterioplankton Composition: High Turnover and Predictability during Shrimp Cultivation. Microb. Ecol. 67, 256-264. doi:10.1007/s00248-013-0336-7.

Xu, W. J., Morris, T. C., and Samocha, T. M. (2016). Effects of C/N ratio on biofloc development, water quality, and performance of Litopenaeus vannamei juveniles in a biofloc-based, high-density, zero-exchange, outdoor tank system. Aquaculture, 453, 169- 175. doi: 10.1016/j.aquaculture.2015.11.021

Xu, W. J., Pan, L. Q., Sun, X. H., and Huang, J. (2013). Effects of bioflocs on water quality, and survival, growth and digestive enzyme activities of Litopenaeus vannamei (Boone) in zero- water exchange culture tanks. Aquac. Res. 44, 1093–1102. doi:10.1111/j.1365- 2109.2012.03115.x.

Xu, W.-J., and Pan, L.-Q. (2012). Effects of bioflocs on growth performance, digestive enzyme Page 175 of 208

activity and body composition of juvenile Litopenaeus vannamei in zero-water exchange tanks manipulating C/N ratio in feed. Aquaculture 356-357, 147-152. doi:10.1016/J.AQUACULTURE.2012.05.022.

Xu, W.-J., and Pan, L.-Q. (2013). Enhancement of immune response and antioxidant status of Litopenaeus vannamei juvenile in biofloc-based culture tanks manipulating high C/N ratio of feed input. Aquaculture 412-413, 117-124. doi:10.1016/J.AQUACULTURE.2013.07.017.

Xu, W.-J., and Pan, L.-Q. (2014). Evaluation of dietary protein level on selected parameters of immune and antioxidant systems, and growth performance of juvenile Litopenaeus vannamei reared in zero-water exchange biofloc-based culture tanks. Aquaculture 426-427, 181–188. doi:10.1016/J.AQUACULTURE.2014.02.003.

Yadav, A. N., Verma, P., Kumar, S., Kumar, V., Kumar, M., Kumari Sugitha, T. C., Singh, B.P., Saxena, A.K., and Dhaliwal, H.S. (2018). Actinobacteria from Rhizosphere: Molecular Diversity, Distributions, and Potential Biotechnological Applications. New Futur. Dev. Microb. Biotechnol. Bioeng., 13-41. doi:10.1016/B978-0-444-63994-3.00002-3.

Yamaguchi, H., Arisaka, H., Seki, M., Adachi, M., and Kimura, K. (2016). International Biodeterioration & Biodegradation Phosphotriesterase activity in marine bacteria of the genera. Int. Biodeterior. Biodegradation 115, 186-191. doi:10.1016/j.ibiod.2016.08.019.

Yan, J., Wu, Y.-H., Meng, F.-X., Wang, C.-S., Xiong, S.-L., Zhang, X.-Y., Yang, Y.-Y, Xu, X.- W., and Zhang, D.-M. (2016). Pseudoalteromonas gelatinilytica sp. nov., isolated from surface seawater. International Journal of Systematic and Evolutionary Microbiology 66: 3538-3545. doi: 10.1099/ijsem.0.001224

Yang, Y. K., Yeh, L. P., Cao, Y. H., Baumann, L., Baumann, P., Tang, B. J. E. and Beaman (1983). Characterization of marine luminous bacteria isolated off the coast of China and description of Vibrio.

Yatip, P., Nitin Chandra Teja, D., Flegel, T. W., and Soowannayan, C. (2018). Extract from the fermented soybean product Natto inhibits Vibrio biofilm formation and reduces shrimp mortality from Vibrio harveyi infection. Fish Shellfish Immunol. 72, 348-355.

Page 176 of 208

doi:10.1016/j.fsi.2017.11.008.

Ying, Y., Tian, X.-X., Wang, J.-J., Qu, L.-Y., and Li, J. (2016). Pseudoalteromonas fenneropenaei sp. nov., a marine bacterium isolated from sediment of a Fenneropenaeus chinensis pond. Int. J. Syst. Evol. Microbiol. 66, 2754-2759. doi:10.1099/ijsem.0.001128.

Zhang, X.-J., Yan, B.-L., Bai, X.-S., Bi, K.-R., Gao, H., and Qin, G.-M. (2014a). Isolation and Characterization of Vibrio parahaemolyticus and Vibrio rotiferianus Associated with Mass Mortality of Chinese Shrimp (Fenneropenaeus chinensis). J. Shellfish Res. 33, 61-68. doi:10.2983/035.033.0108.

Zhang, Q., Jie, Y. W., Loong, W. L. C., Zhang, J., Fane, A. G., Kjelleberg, S., Rice, S.A., and McDougald, D. (2014b). Characterization of biofouling in a lab-scale forward osmosis membrane bioreactor (FOMBR). Water Res. 58, 141-151. doi:10.1016/j.watres.2014.03.052.

Zhang, Z., Reinikainen, J., Adeleke, K. A., Pieterse, M. E., and Groothuis-Oudshoorn, C. G. M. (2018). Time-varying covariates and coefficients in Cox regression models. Ann. Transl. Med. 6, 121. doi:10.21037/atm.2018.02.12.

Zhao, P., Huang, J., Wang, X.-H., Song, X.-L., Yang, C.-H., Zhang, X.-G., and Wang, G-C. (2012). The application of bioflocs technology in high-intensive, zero exchange farming systems of Marsupenaeus japonicus. Aquaculture 354-355, 97-106. doi:10.1016/J.AQUACULTURE.2012.03.034.

Zhao, P., Xue, Y., Gao, W., Li, J., Zu, X., Fu, D., Feng, S., Bai, X., Zuo, Y., and Li, P. (2018). Actinobacteria-Derived peptide antibiotics since 2000. Peptides 103, 48-59. doi:10.1016/J.PEPTIDES.2018.03.011.

Zheng, X., Wang, Y., and Zhang, D. (2019). Effects of microbial products and extra carbon on water quality and bacterial community in a fish polyculture system. Aquac. Res., 1220- 1229. doi:10.1111/are.13996.

Zhenyu, X., Shaowen, K., Chaoqun, H., Zhixiong, Z., Shifeng, W., and Yongcan, Z. (2013). First Characterization of Bacterial Pathogen, Vibrio alginolyticus, for Porites andrewsi White Syndrome in the South China Sea. PLoS ONE, 8(9), e75425. doi: Page 177 of 208

10.1371/journal.pone.0075425

Zhong, Y., Zhang, X.-H., Chen, J., Chi, Z., Sun, B., Li, Y., and Austin, B. (2006). Overexpression, purification, characterization, and pathogenicity of Vibrio harveyi hemolysin VHH. Infect. Immun. 74, 6001-5. doi:10.1128/IAI.00512-06.

Zhong, Z. P., Liu, Y., Hou, T. T., Liu, H. C., Zhou, Y. G., and Wang, F., Liu, Z.-P. (2015). Marivita lacus sp. nov., isolated from a saline lake. Int. J. Syst. Evol. Microbiol. 65, 1889- 1894. doi:10.1099/ijs.0.000195.

Zhou, J., Fang, W., Yang, X., Zhou, S., Hu, L., Li, X., Qi, X., Su, H., and Xie, L. (2012). A Nonluminescent and Highly Virulent Vibrio harveyi Strain Is Associated with “Bacterial White Tail Disease” of Litopenaeus vannamei Shrimp. PLoS One 7, e29961. doi:10.1371/journal.pone.0029961.

Zinger, L., Amaral-Zettler, L. A., Fuhrman, J. A., Horner-Devine, M. C., Huse, S. M., Welch, D. B. M., Martiny, J.B.H., Sogin, M., Boetius, A., and Ramette, A. (2011). Global patterns of bacterial beta-diversity in seafloor and seawater ecosystems. PLoS One 6, 1-11. doi:10.1371/journal.pone.0024570.

Zumla, A., Nahid, P., and Cole, S. T. (2013). Advances in the development of new tuberculosis drugs and treatment regimens. Nat. Rev. Drug Discov. 12, 388-404. doi:10.1038/nrd4001.

Page 178 of 208

Appendices

Page 179 of 208

Appendix 1

(Supplementary information chapter 2)

Supplementary Table S1.1| Results of GLMMs to test for differences in environmental parameters among culturing systems and sampling times. Insignificant model terms were excluded from the final models shown in the table.

Parameter Term F Degrees of freedomb pc Salinity System 21.6 3, 11 < 0.001 Week 508.3 4, 44 < 0.001 Interaction 9.5 12, 44 < 0.001 pH System 56.0 3, 11 < 0.001 Week 24.7 4, 44 < 0.001 Interaction 4.1 12, 44 < 0.001 Temperature System 9.8 3, 11 0.0020 Week 17.5 4, 44 < 0.001 Interaction 3.4 12, 44 0.0015 DO Week 4.6 4, 56 0.0029 Alkalinity System 175.6 3, 11 < 0.001 Week 153.7 4, 44 < 0.001 Interaction 48.2 12, 44 < 0.001 a NH3-N System 16.1 3, 11 < 0.001 Week 14.9 4, 56 < 0.001 + a NH4 -N Week 15.5 4, 56 < 0.001 - a NO2 -N System 31.4 3, 11 < 0.001 Week 5.6 4, 44 0.0010 Interaction 5.3 12, 44 < 0.001 - NO3 -N System 8.43 3, 11 0.0034 TSS System 27.35 3, 11 < 0.001 Week 23.39 4, 56 < 0.001 Interaction 3.03 12, 44 0.0036 VSS System 54.4 3, 11 < 0.001 Week 31.8 4, 44 < 0.001 Interaction 5.7 12, 44 < 0.001 TAN/TIN System 11.1 3, 11 0.0012 Page 180 of 208

Week 4.3 4, 44 0.0053 Interaction 4.1 12, 44 < 0.001 VSS/TSSa System 9.8 3, 11 0.0020 a log-transformed b given as numerator, denominator c Bonferroni-adjusted significance threshold of 0.0038

Page 181 of 208

Supplementary Table S1.2| Tukey post-hoc multiple pairwise comparisons among culturing systems per sampling time (upper case letters) and among sampling times per culturing system (lower case letters). Letters indicate significantly different groups at a family-wise significance threshold of 0.05.

Parameter System Week 0 Week 1 Week 2 Week 3 Week 4 Salinity FT A c A a NS b NS b NS b RAS C b B a NS a NS a NS a H-Floc B b B a NS a NS a NS a A-Floc B b B a NS a NS a NS a pH FT B b B b A a A a A a RAS C b B ab B a B ab B ab H-Floc AB c A a A ab A b A ab A-Floc A b A ab A a A b A ab Temperature FT A ab NS a A a NS ab A b RAS B c NS a B b NS ab B c H-Floc A b NS a AB ab NS b A ab A-Floc A ns NS ns A ns NS ns A ns DO All b a b ab ab Alkalinity FT NS a A a A a A b NS b RAS NS a B b B b B c NS b H-Floc NS a C c C c B c NS b A-Floc NS a D e D d B c NS b a NH3-N FT B bc B c B bc B b B a RAS C bc C c C bc C b C a H-Floc A bc A c A bc A b A a A-Floc AB bc AB c AB bc AB b AB a + NH4 -N All ab c c b a - NO2 -N FT NS ns BC ns B ns B ns B ns RAS NS b AB ab A ab A ab A a H-Floc NS a C b B b A a A a A-Floc NS ns A ns A ns A ns A ns - NO3 -N FT B RAS AB H-Floc A A-Floc B TSS FT A a NS a AB b NS ab A ab RAS A a NS ab A b NS ab A ab H-Floc AB a NS ab C c NS b B bc A-Floc B ab NS a BC b NS a A a VSS

Page 182 of 208

FT A a NS a A b AB ab A ab RAS A ns NS ns A ns A ns A ns H-Floc AB a NS ab B c B b B c A-Floc B a NS a B b AB a A a TAN/TIN FT AB b NS b NS b NS ab A a RAS AB ab NS b NS b NS ab A a H-Floc B ab NS a NS a NS ab B b A-Floc A ab NS b NS ab NS ab A a VSS/TSS FT B RAS A H-Floc B A-Floc B a no interaction

Supplementary Table S2.1| Results of GLMMs to test for differences in bacterial counts (CFU) among culturing systems and sampling times. The insignificant model term was excluded from the final models shown in the table. All counts were log-transformed prior to model testing.

Fraction Parameter Term F Degrees of freedomb pc Water column HB System 10.16 3, 11 0.0017 suc- Vibrio System 26.42 3, 11 < 0.001 Week 13.87 4, 44 < 0.001 Interaction 12.66 12, 44 < 0.001 suc+ Vibrio Week 4.91 4, 56 0.0018 Bacillus System 52.97 3, 11 < 0.001 Week 9.37 4, 44 < 0.001 Interaction 5.07 12, 44 < 0.001 r-traits System 12.78 3, 11 < 0.001 Week 9.66 4, 56 < 0.001 K-traits System 6.77 3, 11 0.0075 Week 9.30 4, 56 < 0.001 Free-living HB Week 13.22 3, 18 < 0.001 suc- Vibrio All terms ns suc+ Vibrio Week 11.31 3, 18 < 0.001 Bacillus All terms ns r-traits

Page 183 of 208

Week 17.76 3, 18 < 0.001 K-traits System 7.07 1, 5 0.0449 Week 6.46 3, 15 0.0051 Interaction 6.40 3, 15 0.0052 Particle-attached HB Week 23.17 3,18 < 0.001 suc- Vibrio Week 5.25 3, 18 0.0089 Suc+ Vibrio All terms ns Bacillus System 5.49 1, 5 0.0661 Week 5.00 3, 15 0.0134 Interaction 8.29 3, 15 0.0017 c significance threshold: 0.0083 (water column), 0.0083 (free-living), 0.0125 (particle-attached)

Supplementary Table S2.2| Tukey post-hoc multiple pairwise comparisons of bacterial parameters among culturing systems per sampling time (upper case letters) and among sampling times per culturing system (lower case letters). Letters indicate significantly different groups at a family-wise significance threshold of 0.05.

Fraction Parameter System Week 0 Week 1 Week 2 Week 3 Week 4 Water column HB FT A RAS A H-Floc B A-Floc A suc- Vibrio FT B ns C ns B ns NS ns C ns RAS A a A a A a NS b B b H-Floc B c BC b B c NS bc A a A-Floc B c B ab B bc NS c AB a suc+ Vibrio All ab a a ab b Bacillus FT A a AB b B c A b A ab RAS A ns A ns A ns A ns A ns H-Floc B a C ab B bc B ab B c A-Floc B a BC a B ab B ab B b r-traitsa FT A b A b A b A b A a RAS A b A b A b A b A a H-Floc B b B b B b B b B a A-Floc A b A b A b A b A a K-traits FT AB a AB ab AB bc AB b AB c RAS A a A ab A bc A b A c

Page 184 of 208

Fraction Parameter System Week 0 Week 1 Week 2 Week 3 Week 4 Water column H-Floc B a B ab B bc B b B c A-Floc AB a AB ab AB bc AB b AB c Free-livingb HB All a a b a suc+ Vibrio All a a b b r-traits All a a b a K-traits H-Floc NS a NS a B b ns a A-Floc NS ns NS ns A ns ns ns Particle-attachedb HB All a a b a suc-+

Vibrio All b b ab a Bacillus H-Floc NS ab NS a B b NS ab A-Floc NS b NS ab A a NS b a No interaction b No values for week 0 available, only determined for H-Floc and A-Floc

Supplementary Table S3.1| Results of GLMMs to test for differences in shrimp performance parameters among culturing systems and sampling times. The insignificant model term was excluded from the final models shown in the table. All parameters were log-transformed prior to model testing.

Parameter Term F Degrees of freedomb p Weight System 5.15 3, 11 0.0182 Week 673.08 3, 33 < 0.001 Interaction 4.73 9, 33 < 0.001 Weight gain rate System 3.74 3, 11 0.0449 Week 14.10 3, 33 < 0.001 Interaction 3.50 9, 33 0.0039 Feed conversion rate System 3.74 3, 11 0.0449 Week 8.66 3, 33 < 0.001 Interaction 3.50 9, 33 0.0039

Page 185 of 208

Supplementary Table S3.2| Tukey post-hoc multiple pairwise comparisons of shrimp performance parameters among culturing systems per sampling time (upper case letters) and among sampling times per culturing system (lower case letters). Letters indicate significantly different groups at a family-wise significance threshold of 0.05. FT is flow-through system. RAS is recirculating system. H-Floc is heterotrophic biofloc system. A-Floc is autotrophic biofloc system.

Parameter System Week 1 Week 2 Week 3 Week 4 Weight FT NS a NS b AB c AB b RAS NS a NS b A b A c H-Floc NS a NS b B c C d A-Floc NS a NS b B c BC d Weight gain rate FT NS b NS b B ab NS a RAS NS b NS b A a NS b H-Floc NS ns NS ns B ns NS ns A-Floc NS ns NS ns B ns NS ns Feed conversion rate FT NS ab NS a A ab NS b RAS NS bc NS a B c NS ab H-Floc NS ns NS ns A ns NS ns A-Floc NS ns NS ns A ns NS ns

Page 186 of 208

Supplementary Table S4| The presumptive Vibrio spp. number at a fraction of shrimp samples

Presumptive Vibrio Week 4 Week 2 Week 0 Week 0 spp. (pooled sample; (x 106 CFU g-1; (pooled sample; (pooled sample; x 106 CFU g-1; suc-Vibrio) CFU g-1; CFU g-1; only suc-Vibrio ) suc-Vibrio) Suc+Vibrio) Flow-through system 8.09 14.32±7.95

Recirculation system 10.92 8.11±2.34 282.50 376.50 Heterotrophic biofloc 8.61 15.47±8.38 system

Autotrophic 9.19 7.63±1.03 biofloc system

Supplementary Table S5| The repeated measurements correlation coefficients (rrm) are given with their 95% confidence intervals (Lower confidence interval - LCI; Upper confidence interval

- UCI). Variable 1 and Variable 2 present high rrm value of correlated parameters from the dataset applied for the correlation test.

Adjusted Variable 1 Variable 2 rrm df LCI UCI P-value Alkalinity Feed 0.48 44 0.210 0.678 0.007 + 0.42 Alkalinity NH4 -N 59 0.182 0.609 0.007 DO Suc-Vibrio 0.45 59 0.213 0.629 0.003 DO TAN/TIN ratio -0.36 59 -0.561 -0.110 0.026 0.44 NH3-N r-strategists 59 0.210 0.628 0.003 -0.41 NH3-N Suc+Vibrio 59 -0.599 -0.166 0.009 -0.55 NH3-N Shrimp weight 44 -0.726 -0.299 0.001 + 0.43 NH4 -N r-strategists 59 0.196 0.618 0.005 + 0.42 NH4 -N Shrimp weight gain rate 44 0.140 0.637 0.021 + -0.45 NH4 -N Suc+Vibrio 59 -0.629 -0.213 0.003 + -0.69 NH4 -N Shrimp weight 44 -0.821 -0.498 0.000 - 0.38 NO2 -N TSS 59 0.132 0.576 0.019 - - 0.37 NO2 -N NO3 -N 59 0.128 0.573 0.019 - 0.37 NO2 -N VSS 59 0.121 0.568 0.021 - -0.47 NO2 -N TAN/TIN ratio 59 -0.650 -0.246 0.002 - -0.42 NO3 -N Shrimp weight gain rate 44 -0.639 -0.143 0.021 - -0.47 NO3 -N TAN/TIN ratio 59 -0.649 -0.244 0.002 pH K-strategists -0.37 59 -0.569 -0.122 0.021 pH Bacillus spp. -0.37 59 -0.574 -0.129 0.019 pH Feed -0.39 44 -0.616 -0.107 0.036 - -0.40 pH NO2 -N 59 -0.598 -0.165 0.009 pH VSS -0.46 59 -0.643 -0.236 0.002

Page 187 of 208

pH Alkalinity -0.50 59 -0.671 -0.282 0.001 pH TSS -0.53 59 -0.695 -0.322 0.000 Shrimp weight Feed 0.98 29 0.956 0.990 0.000 Shrimp weight gain rate FCR -0.55 44 -0.727 -0.300 0.001 Salinity pH 0.57 59 0.368 0.721 0.000 Salinity Temperature 0.41 59 0.171 0.602 0.008 Salinity TSS -0.35 59 -0.558 -0.106 0.028 Salinity r-strategists -0.47 59 -0.646 -0.240 0.002 Salinity Feed -0.61 44 -0.768 -0.384 0.000 Salinity Alkalinity -0.64 59 -0.770 -0.460 0.000 TAN/TIN ratio Suc-Vibrio -0.57 59 -0.719 -0.364 0.000 Temperature Shrimp weight 0.68 44 0.475 0.811 0.000 Temperature r-strategists -0.42 59 -0.610 -0.183 0.007 -0.51 Temperature NH3-N 59 -0.681 -0.297 0.000 + -0.55 Temperature NH4 -N 59 -0.704 -0.338 0.000 TSS Bacillus spp. 0.46 59 0.227 0.638 0.002 VSS Bacillus spp. 0.45 59 0.219 0.633 0.003 K-strategists Shrimp weight 0.46 44 0.193 0.668 0.009 r-strategists Feed 0.40 44 0.113 0.620 0.033 r-strategists Shrimp weight -0.53 44 -0.713 -0.274 0.002 HB K-strategists 0.67 56 0.492 0.792 0.000 HB r-strategists 0.39 56 0.136 0.588 0.019 Suc+Vibrio Shrimp weight 0.55 44 0.303 0.728 0.001 Suc+Vibrio Shrimp weight gain rate -0.43 44 -0.646 -0.154 0.019

Page 188 of 208

Appendix 2

(Supplementary information chapter 3)

Supplementary Table 1| A: Summary of 16S rRNA gene sequence data and alpha diversities from the four shrimp cultured systems for water column samples. FT is the flow-through system. FT is the flow-through system. RAS is the recirculating aquaculture system. H-Floc is the heterotrophic biofloc system. A-Floc is the autotrophic biofloc system.

Samples Time FT RAS H-Floc A-Floc

Median Min Max Median Min Max Median Min Max Median Min Max Wo 77304 41119 121333 119746.5 57417 140180 71172 13618 92121 67948.5 27455 98065 W1 34817 34817 34817 30424 30424 30424 13789 13789 13789 Raw 32115 32115 32115 sequences W2 114747 37745 154104 54118 29568 104643 66077 10795 69963 35970.5 11678 48583 W3 41144 41144 41144 26125 26125 26125 14014 14014 14014 36630 36630 36630 W4 75280.5 35292 121352 52706 14831 66368 59660 16574 63160 30869 14315 49384

Median Min Max Median Min Max Median Min Max Median Min Max Wo 979.5 800 1164 927.5 580 1223 824.5 475 872 1019 623 1220 W1 862 862 862 801 801 801 628 628 628 760 760 760 Raw OTUs W2 1139.5 833 1352 670.5 470 921 999.5 760 1112 1467 666 1527 W3 783 783 783 867 867 867 506 506 506 785 785 785 W4 1178 828 1183 904 635 1230 410.5 268 542 1021 780 1269

Median Min Max Median Min Max Median Min Max Wo 521.325 378.7 668.83 557.085 309.96 748.6 293.08 213.89 315.38 Median Min Max 492.65 492.65 492.65 454.39 454.39 454.39 364 364 364 529.18 501.55 529.29 Rarefied W1 634.56 634.56 634.56 OTUs W2 494.395 451.89 564.04 443.525 342.5 490.67 477.445 458.06 527.9 738.16 622.25 768.63 W3 390.56 390.56 390.56 465.08 465.08 465.08 296.47 296.47 296.47 405.045 343.09 560.6 273.31 221.31 357.54 571.475 451.37 618.06 647.66 647.66 647.66 W4

Page 189 of 208

578.62 553.91 627.9

rare OTUs raw OTUs rare OTUs raw OTUs rare OTUs raw OTUs rare OTUs raw OTUs 529.3 1490 Wo 612.87 1711 687.56 2001 303.96 1359 634.62 760 Total number W1 493.61 862 454.19 801 364.96 628 OTUs 790.88 1936 W2 607.76 1974 548.27 1520 567.13 1857 647.28 785 W3 389.05 783 465.01 867 295.83 506 619.89 1725 W4 522.78 1639 416.1 1070 630.84 2029

Page 190 of 208

B: Summary of 16S rRNA gene sequence data and alpha diversities from the two biofloc shrimp cultured systems for free-living samples. H-Floc is the heterotrophic biofloc system. A-Floc is the autotrophic biofloc system.

Samples Time H-Floc A-Floc

Median Min Max Median Min Max 21113 11465 23282 Raw sequences W2 52055.5 25307 66310 71501.5 21540 86645 56442 43684 76503 W4

Median Min Max Median Min Max 329 206 751 Raw OTUs W2 276 159 347 W4 841.5 363 1537 403 350 816

Median Min Max Median Min Max 313.59 156.53 577.23 Rarefied OTUs W2 135.825 95.56 225.99 358.405 256.63 694.7 238.65 218.91 374.92 W4

rare OTUs raw OTUs rare OTUs raw OTUs Total number OTUs W2 259.59 711 484.33 879 W4 595.2 1962 384.37 1062

Page 191 of 208

C: Summary of 16S rRNA gene sequence data and alpha diversities from the two biofloc shrimp cultured systems for particle samples

Samples Time H-Floc A-Floc

Median Min Max Median Min Max 16347 16347 16347 W1 19321 19321 19321 Raw 70217 43302 97132 W2 44860.5 9270 131969 sequences 33224 33224 33224 W3 20779 20779 20779 48156 35496 67166 75933 34085 208610 W4

Median Min Max Median Min Max W1 576 576 576 863 863 863 Raw OTUs W2 516 328 890 1270 1207 1333 W3 742 742 742 1115 1115 1115 W4 1062 885 1367 1316 1141 1595

Median Min Max Median Min Max W1 412.51 412.51 412.51 664.85 664.85 664.85 Rarefied OTUs W2 346.785 276.33 366.65 622.26 483.03 761.49 W3 531.47 531.47 531.47 641.14 641.14 641.14 W4 504.76 487.62 670.85 524.46 504.25 760.67

rare OTUs raw OTUs rare OTUs raw OTUs Total W1 411.86 576 665.37 863 number OTUs W2 372.83 1144 628.85 1651 W3 530.5 742 638.96 1115 W4 626.12 1908 584.69 2019

Page 192 of 208

Supplementary Table 2| The percentage of frequency distributions for the fraction of water column with statistical values. FT is the flow-through system. RAS is the recirculating aquaculture system. H-Floc is the heterotrophic biofloc system. A-Floc is the autotrophic biofloc system.

Classes FT RAS H-Floc A-Floc

Mean week0 week1 week 2 week 3 week 4 week 0 week 1 week 2 week3 week 4 week 0 week 1 week 2 week 3 week 4 week 0 week 1 week 2 week 3 week 4 (Min, Max) Actinobacteria 10.56 0.35 7.14 13.63 4.42 5.98 7.11 13.29 3.92 18.77 1.50 20.90 33.58 20.91 2.21 4.60 16.18 10.98 12.51 6.27 6.94 (0.35 -33.58 ) ±0.01 ±4.82 ±1.44 ±3.50 ±1.70 ±1.56 ±7.63 ±12.07 ±5.25 ±1.91 ±3.13 ±1.82 Alphaproteobacteria 32.98 22.00 37.05 32.23 31.32 28.91 41.43 8.22 14.22 24.38 11.10 66.98 42.33 40.21 82.82 50.67 25.83 17.65 29.37 30.95 21.87 (8.22 - 82.82 ) ±9.94 ±2.46 ±7.11 ±2.18 ±4.18 ±6.25 ±3.82 ±12.73 ±13.29 ±0.73 ±16.19 ±5.79 Bacteroidia 29.17 37.73 32.07 34.28 42.58 43.82 31.99 72.60 39.15 22.86 21.67 6.12 12.15 13.32 6.98 21.57 30.89 26.70 28.06 27.51 31.42 (6.12 -72.60) ±9.72 ±8.82 ±4.37 ±4.48 ±3.70 ±17.54 ±2.56 ±3.78 ±11.12 ±2.54 ±8.56 ±5.71 Chlamydiae 0.80 0.07 0.03 0.04 0.02 0.05 0.19 0.21 0.93 0.56 1.59 0.39 0.29 1.41 0.11 0.23 0.06 0.14 0.17 0.14 9.31 (0.02-9.31) ±0.04 ±0.05 ±0.04 ±0.08 ±0.63 ±1.45 ±0.18 ±1.80 ±0.08 ±0.02 ±0.11 ±1.05 Deltaproteobacteria 2.28 0.08 1.36 1.38 1.77 2.27 0.89 0.81 1.03 9.57 0.94 0.51 0.69 5.67 2.68 6.72 1.01 0.88 2.91 2.43 1.96 (0.08 - 9.57) ±0.02 ±0.45 ±1.78 ±0.27 ±0.31 ±0.55 ±0.24 ±2.26 ±4.07 ±0.16 ±1.60 ±0.34 Gammaproteobacteria 13.25 39.20 1.18 7.61 3.95 6.75 12.23 2.41 37.75 17.43 58.05 0.61 2.69 8.43 2.45 7.02 7.62 6.25 12.23 18.05 13.06 (0.61- 58.05 ) 1.98± ±2.20 ±0.57 ±2.82 ±5.55 ±20.58 ±0.21 ±2.77 ±2.10 ±0.35 ±2-37 ±1.94 Parcubacteria 1.34 0.06 0.00 0.32 0.01 1.79 0.73 0.43 1.26 2.95 3.08 0.00 0.00 0.00 0.03 0.07 0.77 14.82 0.34 0.01 0.13 (0.00 - 14.82) ±0.07 ±0.35 ±2.67 ±0.44 ±0.64 ±3.74 ±0.00 ±0.00 ±0.03 ±0.20 ±0.25 ±0.02 Saccharimonadia 0.82 0.01 4.48 2.99 2.15 2.22 0.95 0.15 0.15 0.44 0.02 0.11 0.08 1.09 0.06 0.17 0.28 0.11 0.63 0.14 0.13 (0.01-4.48) ±0.01 ±1.12 ±1.21 ±0.38 ±0.08 ±0.02 ±0.04 ±1.38 ±0.06 ±0.09 ±0.40 ±0.04

Page 193 of 208

Supplementary Table 3| The contributions of the environmental factors based on the forwared selection in the multivariate analysis to look for possible strong effect factor. Value is dissimilarity coefficients from redundancy analysis (RDA).

Environmental The RDA complete model with four PCs (PC1 + PC3 + PC4 + PC6) factors

PC1 PC3 PC4 PC6 TSS 0.43 - 0.34 pH - 0.43 VSS 0.43 - 0.40 DO 0.61 - NO3 -N 0.65 - 0.37 - NO2 -N 0.48 Salinity - 0.67 Temperature - 0.49 Alkalinity - 0.58

Page 194 of 208

Supplementary Table 4| Correlations of shrimp weight gain rates with high proportional OTUs significant differences among systems and between weeks with statistical values from Figure 3, Figure 4 and Figure 5. FT is the flow-through system, and RAS is the recirculating aquaculture system, H-Floc is the heterotrophic biofloc system, A-Floc is the autotrophic biofloc system.

Taxonomic levels Class Genera

The classes (Figure 3) OTU name (Figure 4) Characteristic rrm df p LCI UCI factors Gammaproteobacteria Gammaproteobacteria (uncl.) sq26 A-Floc -0.49 14 0.055 -0.81 0.06 Gammaproteobacteria Gammaproteobacteria (uncl.) sq40 A-Floc -0.64 14 0.007 -0.87 -0.17 Gammaproteobacteria Gammaproteobacteria (uncl.) sq50 -0.26 14 0.335 -0.69 0.32 Gammaproteobacteria Parahaliea sq58 0.39 14 0.137 -0.18 0.76 Gammaproteobacteria Colwellia sq74 -0.50 14 0.050 -0.81 0.04 Gammaproteobacteria Pseudoalteromonas sq91 RAS, week2 0.52 14 0.038 -0.01 0.82

Gammaproteobacteria Oceanospirillum sq104 -0.24 14 0.369 -0.68 0.33 Gammaproteobacteria Oceaniserpentilla sq111 -0.07 14 0.810 -0.58 0.48 Gammaproteobacteria Oceaniserpentilla sq115 -0.15 14 0.576 -0.63 0.41 Gammaproteobacteria Catenovulum sq142 RAS, week2 0.75 14 0.001 0.37 0.92

Gammaproteobacteria Halomonas sq263 FT, week0 0.66 14 0.006 0.19 0.88 Gammaproteobacteria Granulosicoccus sq278 0.58 14 0.018 0.07 0.85 Alphaproteobacteria Ruegeria sq1 H-Floc 0.63 14 0.009 0.15 0.87 Particle bacteria week01234 Alphaproteobacteria Marivita sq3 FT -0.53 14 0.033 -0.83 -0.01 H-Floc Alphaproteobacteria Leisingera sq72 0.37 14 0.159 -0.20 0.75 Alphaproteobacteria Ruegeria sq90 FT, week 0 0.64 14 0.008 0.16 0.87

Page 195 of 208

Alphaproteobacteria Aliiroseovarius sq63 0.44 14 0.092 -0.12 0.78 Alphaproteobacteria Pseudooceanicola sq78 A-Floc, week0 0.50 14 0.046 -0.04 0.82 FT, week0 Alphaproteobacteria Roseobacter sq81 H-Floc, week0 0.61 14 0.011 0.12 0.86 Alphaproteobacteria Roseovarius sq175 RAS 0.52 14 0.038 -0.01 0.82 week2 Alphaproteobacteria Maritalea sq237 0.44 14 0.085 -0.11 0.79 Alphaproteobacteria Aliiroseovarius sq137 0.21 14 0.426 -0.36 0.67 Bacteroidia Hanstruepera sq5 FT -0.20 14 0.461 -0.66 0.37 Bacteroidia Saprospiraceae (uncl.) sq7 FT -0.31 14 0.235 -0.72 0.26 A-Floc Bacteroidia Marixanthomonas sq382 0.10 14 0.716 -0.46 0.60 Bacteroidia Lewinella sq23 -0.28 14 0.293 -0.71 0.29 Bacteroidia Fulvivirga sq38 -0.46 14 0.072 -0.80 0.09 Bacteroidia Winogradskyella sq165 -0.34 14 0.202 -0.74 0.24 Actinobacteria Mycobacterium sq253 -0.21 14 0.424 -0.67 0.36 Actinobacteria Mycobacterium sq32 RAS -0.40 14 0.125 -0.77 0.17 Parcubacteria Candidatus Campbellbacteria (uncl.) sq132 -0.40 14 0.129 -0.77 0.17 unclassified B2M28 (uncl.) sq70 H-Floc -0.23 14 0.398 -0.68 0.35 unclassified NS3a marine group sq9 RAS -0.58 14 0.018 -0.85 -0.08 unclassified SAR324 clade (Marine group B) (uncl.) sq95 0.09 14 0.752 -0.47 0.59 unclassified OM43 clade sq176 -0.51 14 0.043 -0.82 0.03 unclassified Chitinophagales (uncl.) sq212 0.03 14 0.903 -0.51 0.55

Page 196 of 208

Supplementary Table 5| The taxonomic groups that may serve as bio-indicators of the four shrimp culture systems (different environments).

Systems OTUs bio-indicator based Time of OTUs Classes bio-indicator on positive associated with present based on proportion shrimp weight gain rates of relative abundance FT Ruegeria sq90 week 0 Saccharimonadia (flow-through) (Alphaproteobacteria class) (Max = 4.48%)

Pseudooceanicola sq78 week 0 (Alphaproteobacteria class)

Halomonas sq263 week 0 (Gammaproteobacteria class)

RAS Roseovarius sq175 week 0 Bacteroidia (recirculating aquaculture) (Alphaproteobacteria class) (Max = 72.60 %)

Catenovulum sq142 week 2 Gammaproteobacteria (Gammaproteobacteria class) (Max = 58.05 ± 20.58%)

Pseudoalteromonas sq91 week 2 Deltaproteobacteria (Gammaproteobacteria class) (Max = 9.57 %)

H-Floc Roseobacter sq81 week 0 Alphaproteobacteria (heterotrophic biofloc) (Alphaproteobacteria class) (Max = 82.82%)

Ruegeria sq1 week 0 Actinobacteria (Alphaproteobacteria class) week 1 (Max = 33.58 %) week 2 week 3 week 4

A-Floc Pseudooceanicola sq78 week 0 Parcubacteria (autotrophic biofloc) (Alphaproteobacteria class) (Max = 14.82 %)

Chlamydiae (Max = 9.31 ± 1.05%)

Page 197 of 208

Water FL PA Number of of OTUs Number

Wo FT W1 RAS W2 H−Floc W3

0 500A−Floc 10000 1500 500W4 10000 1500 500 1000 1500

0 50000 100000 150000 200000 0 50000 100000 150000 200000 0 50000 100000 150000 200000 Number of sequences

Supplementary Figure 1| Rarefaction analysis for each sample based on best matches against the SILVA database (OTUs) for three factions of the four systems. Three-panel figures: water column, free-living (FL), and particle-attached (PA) samples. Systems and weeks indicate by colour and line type. FT is the flow-through system, and RAS is the recirculating aquaculture system, H-Floc is the heterotrophic biofloc system, A-Floc is the autotrophic biofloc system.

Page 198 of 208

Supplementary Figure 2| The bacterial alpha diversities according to different taxonomic taxa from the four shrimp cultured systems. FT is the flow-through system, and RAS is the recirculating aquaculture system, H-Floc is the heterotrophic biofloc system, A-Floc is the autotrophic biofloc system.

Page 199 of 208

Supplementary Figure 3 A Free-livingFreee -living heterotrophicheterotrophic relative sequencesequence abundance Relative sequence abundance (%)

Page 200 of 208

Supplementary Figure 3 B MarivitaMarivita inin thethe free-livingfree-living heterotrophicheterotrophic bioflocs samples (> 80%) Relative sequence abundance (%)

The heterotrophic biofloc samples at week 2, and 4

Supplementary Figure 3| The boxplot shows proportional sequences of different significant OTUs among the four systems (A) and the genera of specific interest taxa are present over all (B) at fractions of the water column, free-living and particle samples. FT is the flow-through system, and RAS is the recirculating aquaculture system, H-Floc is the heterotrophic biofloc system, A-Floc is the autotrophic

biofloc system. W0, W1, W2, W3, W4 are weeks 0, 1, 2, 3 and 4 in the shrimp culture period.

Page 201 of 208

Appendix 3

(Supplementary information chapter 4)

Supplementary Table 1| A: Bacterial strains used for the challenge tests. B: Summary the results of pilot challenge test. ARC is Laboratory of Aquaculture & Artemia Reference Centre, Coupure Links 653, B-9000 Gent, Belgium

A: Bacterial strains using for the challenge test Code of TCBS growth Source of isolation Sanger sequencing Reference strains LV11 yellow colony water mix (i.e. seawater) for rearing shrimp Vibrio sp. LV11 Chapter 2 LV21 green colony water mix (i.e. seawater) for rearing shrimp Vibrio sp. LV21 Chapter 2 pVpAHPND green colony ARC: potentially AHPND-causing strain of NA PV1-ARC Vibrio parahaemolyticus

B: Sumarry the results of pilot challenge test Bacterial strains Key findings Recommendations LV11 and LV21 + non-toxic bacteria + feed supply for the Litopenaeus vannamei

+ shrimp were still alive after three doses of 106 CFU/mL until day 15 after the challenge test is not necessary infection. + incubation to avoid died shrimp because of stress is the most important thing before application of challenged bacterium. pVpAHPND + toxic bacterium + the shrimp mortality data should be + causing shrimp post-larvae (0.70 ± 0.07 g) died after 6 hours post immersion at the recorded every 6 hours exposing to the

dose ≥ 106 CFU mL-1. Shrimp did not die at the dose of < 106 CFU mL-1. challenge strains.

+ at bacterial concentration of 106 CFU mL-1, shrimp died almost within 24 h.p.i; if + the challenge test needs more than one shrimp were still alive after 48 hours, they were recovering and alive until day 15 (day lethal dose of bacterium. of stopping observations).

Page 202 of 208

Supplementary Table 2| Distribution of bacterial family before (A) and after (B) challenge test in shrimp tissue

A: Before challenge test Taxonomic levels Relative abundance (%) Class Family w0 w2.FT w2.RAS w2.Hfloc w2.Afloc w4.FT w4.RAS w4.Hfloc w4.Afloc Bacilli Bacillales_Family XII 45.08 0.02 0.04 0.01 0.08 0.01 0.01 0.00 0.00 Bacilli Carnobacteriaceae 13.23 0.02 0.01 0.00 0.00 0.00 0.00 0.00 0.00 un_classified Demequinaceae 0.04 0.00 0.00 0.00 0.00 1.17 0.67 0.37 0.63 Bacteroidia Flavobacteriaceae 0.14 0.00 0.00 0.00 0.00 27.27 33.20 16.49 12.84 Gammaproteobacteria Gammaproteobacteria_unclassified 0.00 0.00 0.00 0.00 0.00 7.92 0.07 1.55 13.02

Alphaproteobacteria Rhodobacteraceae 0.91 0.01 0.01 0.01 0.00 29.11 51.01 47.13 26.03

un_classified Rubritaleaceae 0.00 0.00 0.00 0.00 0.00 0.72 0.40 0.50 1.29 Gammaproteobacteria Shewanellaceae 0.14 43.22 70.07 42.27 45.98 0.09 0.04 0.00 0.01 Gammaproteobacteria Vibrionaceae 40.11 55.12 29.23 54.30 52.81 1.70 3.88 1.81 5.68 Other Bacteria 0.34 1.59 0.64 3.40 1.12 32.01 10.73 32.15 40.48

B: After challenge test Relative abundance (%) pPV- pPV- pPV- pPV- LV11.FT LV11.RAS LV11.Hfloc LV11.Afloc LV21.FT LV21.RAS LV21.Hfloc LV21.Afloc pPV+.FT pPV+.RAS pPV+.Hfloc pPV+.Afloc .FT .RAS .Hfloc .Afloc Bacillales_Family XII 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Carnobacteriaceae 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Demequinaceae 4.10 10.80 0.79 0.67 0.32 0.56 0.07 0.05 0.50 1.07 0.75 0.01 0.02 0.00 0.00 0.01 Flavobacteriaceae 44.17 26.12 24.61 34.55 31.68 32.77 19.22 16.74 37.42 33.26 24.03 11.24 1.97 3.24 8.04 7.73 Gammaproteobacteria_unclassified 0.19 0.03 2.24 0.26 0.17 0.15 1.36 0.10 0.09 0.06 1.24 0.07 0.19 0.08 0.06 0.09 Rhodobacteraceae 11.16 17.31 19.08 20.56 43.63 37.12 41.28 60.63 31.66 33.55 37.29 48.02 27.50 20.09 15.92 18.38 Rubritaleaceae 14.34 19.44 28.06 25.12 7.17 19.79 15.42 3.57 16.41 10.93 15.23 3.95 0.06 0.04 0.14 0.08 Shewanellaceae 0.06 1.52 0.10 0.00 0.07 0.01 0.04 0.01 0.14 0.15 0.01 0.48 0.01 0.00 0.00 0.02 Vibrionaceae 1.14 3.11 1.04 0.43 0.05 0.08 0.46 0.72 0.76 1.56 0.98 13.98 67.03 73.60 67.85 68.21 other 24.82 21.66 24.09 18.41 16.91 9.52 22.13 18.19 13.02 19.42 20.46 22.26 3.22 2.95 7.99 5.48

Page 203 of 208

120 shrimpFT 120 shrimpRAS 120 shrimpHfloc 120 shrimpAfloc

480 boxes of shrimp for the challenge tests

1 shrimp per box, each box contain 300 mL water, 15 boxes per condition

waterFT 4 conditions each shrimp using rearing was challenged waterRAS water (first with pVpAHPND 8 conditions data-subset) at dose of 107 waterHfloc (15 shrimp 1 replicates per

condition) waterAfloc

7 -1 pVpAHPND at dose of 10 mL

Vibrio sp. LV11 at dose of 107 mL-1 4 conditions using seawater (second data-subset) Vibrio sp. LV21 at dose of 107 mL-1

without challenged Vibrio strain

Supplementary Figure 1| The experimental design. shrimpFT is shrimp cultivated in flow-through system. shrimpRAS is shrimp cultivated in recirculating system. shrimpHfloc is shrimp cultivated in heterotrophic biofloc system. shrimpAfloc is shrimp cultivated in autotrophic biofloc system. waterFT is rearing water from flow-through system. waterRAS is rearing water from recirculating system. waterHfloc is rearing water from heterotrophic biofloc system. waterAfloc is rearing water from autotrophic biofloc system. Vibrio sp. LV11 is yellow colony Vibrio. Vibrio sp. LV21 is green colony Vibrio. Both LV11 and

LV21 were isolated in the rearing water. pVpAHPND is potentially Acute Hepatopancreatic Necrosis Disease -causing strain of Vibrio parahaemolyticus.

Page 204 of 208

residuals

2

1

0

−1

−2 Residuals (type =deviance)

−3 0 50 100 150 200 250 Time

Supplementary Figure 2| The figure shows the influent of observations for the Cox model’s assumption was fulfilled by residuals as a function of time in examination by deviance.

Page 205 of 208

Supplementary Figure 3 A: Shrimp cultivated from the four systems challenged with pVpseawater.system=seawater.FT,AHPND. There were no differences strain.diff=p.pathogenicV among shrimp in using of seawater. seawatersystem=seawater.Hfloc, for the challenge test s (p = 0.13).

1.00

0.75

0.50

Survival probability Survival 0.25 p = 0.13

0.00 0 6 12 18 24 30 36 42 48 54 60 66 72 78 84 90 96 102108114120 Time(hours post immersion)

shrimpuse shrimpAfloc shrimpFT shrimpHfloc shrimpFT 0.91 - - shrimpHfloc 0.17 0.32 - shrimpRAS 0.91 0.17 0.66 Pairwise comparisons using Log-Rank test/ p-value adjustment method: Benjamini- Hochberg

Page 206 of 208

Supplementary Figure 3 B: The figure shows the time points that 50% shrimp cultivated from the four system=seawater.Aflocsystems challenged strain.diff=p.pathogenicV, with pVpAHPND seawater.system= have died. Theseawater.FT Kaplan- Meier strain.diff=p.pathogenicV, method was seawater.system=seawat applied for the er.Hflocanalysis str of 50% survival probability. Shrimp all have died 50% before 24 hours (at 12 and 18 hours).

1.00

0.75

0.50 Survival probability Survival

0.25

0.00

012 18 25 50 75 100 Time (hours post immersion)

Supplementary Figure 3| The second data-subset shows that pVpAHPND was a true virulent bacterium with causing shrimp died within 12 to 18 hours post immersion with no differences among shrimp cultivated from four culture systems using seawater for challenge condition (Supplementary Figure 3 A and B). pVpAHPND is potentially AHPND-causing strain of Vibrio parahaemolyticus.

Page 207 of 208

extracellular products of pVpAHPND

Supplementary Figure 4: The picture shows the extracellular products of pVpAHPND in the challenge room at the Laboratory of Aquaculture & Artemia Reference Centre (ARC) in Belgium.

“If I knew what I was doing, it wouldn't be called research” - A. Einstein

Page 208 of 208