Microbiome and Disease Associated with Ostreid Herpesvirus-1 (Oshv-1)

Environmental influences on the Pacific oyster (Crassostrea gigas) microbiome and disease associated with Ostreid herpesvirus-1 (OsHV-1)

Bhagini Erandi PATHIRANA

A thesis submitted in fulfilment of the requirements for the degree of Doctor of Philosophy

THE UNIVERSITY OF SYDNEY

Farm Animal Health Sydney School of Veterinary Science Faculty of Science

February 2020

Declaration of Authorship

Apart from the assistance stated in the acknowledgements section, this thesis represents the original work of the author. To the best of my knowledge the results from this study have not been presented for award for any other degree or diploma at this or any other university.

Bhagini Erandi PATHIRANA MSc, BVSc (Hons) February 2020

Acknowledgements

First, I would like to extend my sincere thanks to Dr. Paul Hick for his expert supervision as the primary supervisor of my doctoral research study. His patience and guidance helped me to develop skills and competence in research and scientific communication. I would like to take this opportunity to thank him for supervising and guiding me to achieve valuable outcomes from my PhD research and for providing me with great opportunities to reach new horizons in molecular diagnostics, bioinformatics and aquatic animal health, through professional training.

The role of Emeritus Professor Richard Whittington as my auxiliary supervisor is no way smaller. I would like to take this opportunity to thank him again for accepting me to the diverse and amazing Farm Animal Health Research group of the University of Sydney, as a PhD candidate. Without his acceptance, I would not have been able to pursue a doctoral research at the University of Sydney. Further, I would like to extend my gratitude for his expert supervision in helping me to develop my research and scientific communication skills.

The financial assistance provided by the Australian Government through the Endeavour Leadership Program, supporting my doctoral studies and my stay in Australia as an international student is greatly appreciated. Securing an Australian Government Scholarship to pursue doctoral studies at the University of Sydney is one of the greatest achievements of my academic career and a turning point of my career as a university academic and as a researcher. Further, I would like to acknowledge the financial assistance provided by the University of Sydney to support my doctoral studies.

It has been a wonderful privilege to work with a highly qualified, friendly and a supportive team of people in the Farm Animal Health Research group. In particular, I would like to extend my sincere appreciation to Associate Professor Jenny-Ann Toribio, Dr. Om Dhungyel, Dr. Karren Plain, Dr. Auriol Purdie and Dr. Kumudika de Silva, for their support and expert views provided during research group meetings and research presentations which positively contributed to improve my doctoral research. My sincere thanks also go to Associate Professor Joy Becker for providing me with the opportunity to continue my teaching career and to widen my teaching experience as a graduate teaching assistant. Further, I would like to thank Anna Waldron, Alison Tweedie, Ann-Michele Whittington, Slavicka Patten, Natalie Schiller, Nicole

Carter, Rebecca Maurer, Stuart Glover and Craig Kristo for their technical expertise, support and guidance in carrying out various laboratory procedures and for creating a pleasant laboratory environment. The support from late Pabitra Dhungyel in the laboratory, during the early years of my candidature is fondly remembered. The laboratory skills and competence that I developed by working in NATA-accredited laboratories will undoubtedly be a great asset in my future career. The support and thoughtful assistance extended by Marion Saddington, was an invaluable strength during my entire PhD candidature. I would also like to take this opportunity to thank Associate Professor Navneet Dhand and Dr. Kathrin Schemann for their expert statistical advice in refining the statistical analyses of my research.

I am very much thankful for Dr. Olivia Evans and Dr. Marine Fuhrmann for their great support in carrying out various research activities during my doctoral programme. My sincere thanks also go to Dr. Andrew McPherson for helping me to get started with microbiome analysis using the QIIME software and for his invaluable contribution for the first research publication during my PhD candidature. It has been a pleasure to work alongside friendly and competent PhD candidates, Dr. Matt Johansen, Dr. Kamal Aacharya and Dr. Hannah Pooley (as former PhD students) and Maximillian de Kantzow, Karen Smith, Anna Ly, Katherine Wright, Cahya Fusianto and Ed Annand.

I would also like to thank the bioinformatics staff at the Sydney Informatics Hub, especially Dr. Rosemarie Sadsad and Dr. Tracy Chew, for their expert guidance in bioinformatic analyses and for their prompt support whenever needed. Dr. Neil Horadagoda and Kanchana Ekanayake are sincerely acknowledged for their great helping hand whenever needed. They were a great support from the time of our establishment at Camden and throughout the PhD candidature.

Last but in no way least, I am very much grateful and indebted to my loving husband Indunil for his continuous encouragement and immense moral and physical support throughout my PhD candidature. Without him this PhD would not be a reality. I would also like to thank my loving son Nisanga for spending long, lonely hours as a little boy and for his support as a teenager to make his mum’s PhD a reality. This thesis is the output of the sacrifice made by all three of us!

iii

Summary

Pacific oyster mortality syndrome (POMS) is a high mortality disease in Pacific oysters (Crassostrea gigas), which has negatively impacted oyster farming. Despite the causal relationship between Ostreid herpesvirus-1 (OsHV-1) and oyster mortality, the incidence and severity of disease is determined by complex interactions between the physiology of oysters, the environmental conditions and secondary pathogens. Understanding the multifactorial nature of this disease is required to develop effective management strategies. Recent investigations revealed a polymicrobial pathogenesis of this disease, with primary infection by OsHV-1 followed by the involvement of opportunistic bacteria present in the microbiome completing the disease expression. Apart from the implications of various environmental risk factors on POMS, the oyster microbiome is also affected by environmental disturbances. It is important to determine if differences in the microbiome such as dysbiosis, are an outcome of the pathogenesis of the disease or whether the features of the microbiome predispose or contribute to the severity of this disease. Although comprehensive studies concerning polymicrobial pathogenesis of POMS were conducted after this thesis was commenced, the influence of environmental factors on the polymicrobial pathogenesis of POMS and how it is mediated by the oyster microbiome remained a knowledge gap in this regard. This thesis aimed at addressing different aspects of this knowledge gap by investigating the impact of different environmental factors on the Pacific oyster microbiome and their subsequent effects on disease associated with OsHV-1.

Chapter 3 investigated how the microbiome of genetically related Pacific oysters with a common hatchery origin differed, when grown in different estuaries. Results indicated an influence of farming environment in shaping the microbiome. Bacterial diversity as determined by 16S rRNA gene sequence analysis, indicated that different estuarine environments generated unique microbiomes which were also associated with a differential response to OsHV-1 infection. The quantitative dynamics of total bacteria and Vibrio spp. during an OsHV-1 infection was assessed using qPCR assays. The microbiome changed with the environment and after an experimental OsHV-1 challenge. A strong correlation was observed between the OsHV- 1 and Vibrio quantities in OsHV-1 infected oysters. Different microbiomes prior to infection were associated with altered responses to OsHV-1 challenge and different disease outcomes.

Optimizing the quality and quantity of bacterial DNA that was purified from Pacific oyster tissues proved critical for accurate characterisation of the Pacific oyster microbiome. This is particularly important in assessing the potential polymicrobial pathogenesis of Pacific oyster mortality diseases. Chapter 4 focused on evaluating methods to sample tissues, extract and appropriately store bacterial nucleic acids from Pacific oyster, to accurately determine the microbiome. Both intrinsic oyster factors and experimental factors influenced the results of oyster microbiome studies with potential biases being introduced due to sampling method and approaches to nucleic acid extraction. A tissue compartmentalization of the Pacific oyster microbiome was confirmed with identification of distinct, tissue-specific microbiomes in the haemolymph, gill and gut. The different experimental procedures substantially impacted the quantity and diversity of bacteria identified.

The controlled environment of a laboratory challenge system for OsHV-1 was utilized to investigate the pathogenesis of POMS. A period of acclimation to this environment before an experimental study can induce changes to the microbiome that may confound findings attributed to the disease being studied. Chapter 5 focused on assessing changes in the Pacific oyster microbiome during acclimation to a laboratory environment when oysters are maintained with either constant immersion in water or a simulated tide. While the overall diversity (alpha diversity) of the microbiome was not affected by acclimation to the laboratory environment, the abundance of phyla such as Cyanobacteria and Tenericutes were reduced. Although the dominant phylum Proteobacteria was not affected at the phyla level, changes occurred at lower taxonomic levels, with increases in abundance of the genera Vibrio and Arcobacter, both of which are known opportunistic bacteria in POMS. However, changes in the microbiome were not affected by the immersion regime (constant immersion and tidal emersion) of oysters.

Although the microbiome of oysters maintained under different immersion regimes in the laboratory remained similar, the oyster microbiome responded differently when challenged with OsHV-1, along with differential oyster mortality. In Chapter 6, increased susceptibility to OsHV- 1 was identified for oysters in a simulated tidal environment when exposure to OsHV-1 was controlled by injection. In the event of an adequate exposure to OsHV-1, periodic emersion of oysters in an intertidal environment predisposed oysters to POMS concurrent with an impact on the microbiome indicated by an increase in alpha diversity and increase in rare bacterial genera such as Polaribacter, Marinicella and Sediminibacterium. On the other hand, the oysters under

v constant immersion showed a decreased alpha diversity in the microbiome and a lower oyster mortality.

Seawater temperature plays a major role in triggering diseases of ectothermic organisms in the marine environment. Chapter 7 examined the oyster microbiome under different water temperature profiles consistent with the diurnal and longer-term variation that can be experienced in intertidal and estuarine surface waters. The seawater temperature altered the microbiome with differences in the abundance at the level of bacterial genera depending on the degree and nature of changes within the temperature profile. Increased seawater temperatures were associated with decreased abundance of Arcobacter and increased abundance of Vibrio, strengthening the potential role of Vibrio in Pacific oyster mortality. Together with a higher concentration of OsHV-1 and higher oyster mortality associated with increased water temperature, there was also a higher concentration of Vibrio in oyster tissues.

To conclude, this thesis demonstrated that the environment plays a significant role in shaping the Pacific oyster microbiome. Pacific oysters with a common genetic background may respond differently to disease associated with OsHV-1 owing to the influence of the farming environment on their microbiome. Fluctuating conditions in the intertidal environment negatively impacted the oyster microbiome in the event of an adequate exposure to OsHV-1, predisposing oysters to disease associated with OsHV-1. Higher seawater temperatures (21°C, 22°C and 26°C) were positively linked with an increase of the Vibrio fraction in the oyster microbiome together with higher OsHV-1 content and increased oyster mortality, confirming knowledge of the opportunistic role of Vibrio in POMS. Further studies are needed to evaluate the functional significance of bacteria such as Arcobacter in POMS. However, potential changes in the microbiome during laboratory acclimation should also be taken into consideration in future experimental studies.

The knowledge generated in this thesis, contributes to a growing understanding of the resilience of the oyster microbiome under the influence of different environmental conditions and its potential to contribute to a polymicrobial pathogenesis of disease. Improved knowledge of the role of the microbiome and the environment on disease expression can help direct advice on oyster farm management strategies. Changes to growing infrastructure and farm management, particularly in the face of harsh conditions and a changing environment can provide conditions which favour a healthy and resilient community of commensal bacteria. vi

List of publications and conference presentations

Refereed first author publications (included in the thesis): 1. Pathirana E., Fuhrmann, M., Whittington, R., Hick, P. (2019). Influence of environment on the pathogenesis of Ostreid herpesvirus-1 (OsHV-1) infections in Pacific oysters (Crassostrea gigas) through differential microbiome responses. Heliyon 5. doi: 10.1016/j.heliyon.2019.e02101

This publication constitutes Chapter 3 of this thesis. I am the first author of this paper. I co- designed the study with the co-authors, performed the experiments, collected and analysed the data, and wrote drafts of the manuscript. Dr. Paul Hick is the corresponding author.

2. Pathirana, E., McPherson, A., Whittington, R., Hick, P. 2019. The role of tissue type, sampling and nucleic acid purification methodology on the inferred composition of Pacific oyster (Crassostrea gigas) microbiome. Journal of Applied Microbiology 127; 429-444. doi: 10.1111/jam.14326

This publication constitutes Chapter 4 of this thesis. I am the first author of this paper. I co- designed the study with the co-authors, performed the experiments, collected and analysed the data, and wrote drafts of the manuscript. Dr. Paul Hick is the corresponding author.

Refereed co-author publications: 3. Evans, O., Kan, J.Z.F., Pathirana E., Whittington, R.J., Dhand, N., Hick, P. 2019. Effect of emersion on the mortality of Pacific oysters (Crassostrea gigas) infected with Ostreid herpesvirus-1 (OsHV-1). Aquaculture 505; 157-166. doi: 10.1016/j.aquaculture.2019.02.041

This publication was generated from a concurrent study that was undertaken for Chapter 6 of this thesis. I am a co-author of this paper. I co-designed the study and performed parts of the experiment that are reported in the publication, with the co-authors. Dr. Paul Hick is the corresponding author. I co-designed the study, conducted the experiments and analysed and drafted the report of the study in Chapter 6.

vii

Authorship confirmation for publications:

Student: In addition to the statement above, where I am not listed as the corresponding author, permission to include the material has been granted by the corresponding author.

BHAGINI ERANDI PATHIRANA 26.02.2020

Supervisor: As supervisor for the candidature upon which this thesis is based, I can confirm that the authorship attribution statement above is correct.

PAUL HICK 26.02.2020

viii

Conference presentations: 1. Pathirana, E., Whittington, R., Hick, P. (2019). The role of tissue type, sampling and nucleic acid purification methodology on the inferred composition of Pacific oyster (Crassostrea gigas) microbiome. Sydney School of Veterinary Science (SSVS) Postgraduate Conference 2019, The University of Sydney, Camden, Australia. December 2019. 2. Pathirana, E., Fuhrmann, M., Whittington, R., Hick, P. (2018). The role of Pacific Oyster (Crassostrea gigas) microbiome in the pathogenesis of Ostreid herpesvirus-1 infections. 8th International Symposium for Aquatic Animal Health (ISAAH), Charlottetown, Prince Edward Island, Canada. September 2018. 3. Pathirana, E., Whittington, R., Tweedie, A., Hick, P. (2018). Tissue preparation and nucleic acid purification influences the outcome of Pacific oyster (Crassostrea gigas) microbiome studies. Annual meeting of the Australian Society for Microbiology (ASM) 2018, Brisbane, Queensland, Australia. July 2018. 4. Pathirana, E., Whittington, R., Hick, P. (2017). Pacific oyster (Crassostrea gigas) microbiome: influence of environment and infection. Marie Bashir Institute Colloquium 2017, The Marie Bashir Institute, Sydney Medical School, The University of Sydney, Sydney, Australia. November 2017. 5. Pathirana, E., Whittington, R., Hick, P. (2017). Pathirana, E., Whittington, R., Hick, P. (2017). Pacific oyster (Crassostrea gigas) microbiome in an experimental Ostreid herpesvirus-1 (OsHV-1) infection. Sydney School of Veterinary Science (SSVS) Postgraduate Conference 2017, The University of Sydney, Camden, Australia. November 2017. 6. Pathirana, B. E., Fuhrmann, M., Whittington, R., Hick, P. (2016). Assessing the role of Vibrio spp. in the pathogenesis of Pacific oyster (Crassostrea gigas) mortality associated with Ostreid herpesvirus-1 (OsHV-1). Postgraduate Research Conference 2016, Faculty of Veterinary Science, The University of Sydney, Sydney, Australia. November 2016.

List of Tables

Table 1.1: Fluctuation of respiratory variables in oyster haemolymph during immersion and emersion………………………………………………………………………………………….10 Table 1.2: Molecular techniques other than high throughput sequencing of 16S rRNA gene regions, in bivalve microbiome studies and their results…………………………….…………..22 Table 1.3: Different hypervariable regions of the 16S rRNA gene targeted in bivalve microbiome studies………………………………………………………………………………23 Table 2.1: Primers and probes used in this thesis……………………………………………….51 Table 2.2: Thermocycling conditions for the OsHV-1 qPCR assay…………………………….51 Table 2.3: Thermocycling conditions for the Vibrio qPCR assay………………………………54 Table 2.4: Thermocycling conditions for the total bacteria qPCR assay………………………..56 Table 3.1: Cohorts of oysters used for microbiome analysis……………………………………71 Table 3.2: Odds of mortality for oysters injected with OsHV-1 depending on the batch of oysters……………………………………………………………………………………………73 Table 3.3: Mean OsHV-1 concentration in oyster tissues during the infection………..………..75 Table 3.4: Results of Generalised Linear Model analysis of total bacterial load and total Vibrio load……………………………………………………………………………………………….78 Table 3.5: Total bacterial load and total Vibrio load in oysters during OsHV-1 infection……...78 Table 4.1: Workflow carried out for optimizing tissue sampling, tissue disruption, microbiome enrichment and nucleic acid extraction for analysis of Crassostrea gigas microbiome………...91 Table 4.2: Mean bacterial DNA yields for different Crassostrea gigas tissues, sampling methods used for the same tissue, and nucleic acid extraction methods………………………………...101 Table 4.3: Comparison of the detection efficiency for the internal positive control (IPC) in nucleic acid extracts obtained from different tissues, sampling techniques, and nucleic acid extraction methods……………………………………………………………………....……...103 Table 4.4: Bray-Curtis dissimilarity between bacterial communities in the same tissue type for different individual oysters and between different tissue types from the same individual oyster……………………………………………………………………………………………106 Table 5.1: Total cultivable bacterial count and total cultivable Vibrio count in the whole soft- tissue mass of oysters……………………………………...... 127 Table 5.2: Total bacterial DNA quantity in gill and gut tissues of oyster……...………...... 128

Table 5.3: Changes in the absolute abundance of dominating bacterial phyla in gill and gut microbiota during acclimation to the laboratory environment…………………………………136 Table 5.4: Alpha diversity of gill and gut microbiota of oysters during acclimation to the laboratory environment…………………………………………………………………….…..138 Table 5.5: Beta diversity of gill and gut microbiota of oysters during acclimation to the laboratory environment………………….……………………………………………………..139 Table 6.1: OsHV-1 DNA concentration in gill and mantle tissues of oysters challenged with OsHV-1 ……………………………………………………………..……...... 159 Table 6.2: Results of Generalized Linear Mixed Model analysis of total bacterial DNA quantity in gill and gut tissues of oysters………………………………………...……...... 161 Table 6.3: Total bacterial DNA quantity in oysters challenged with OsHV-1 and oysters injected with an OsHV-1 negative tissue homogenate………………………………………….………162 Table 6.4: Total cultivable bacterial count (TCBC) in oysters challenged with OsHV-1 and oysters injected with an OsHV-1 negative tissue homogenate ………………………………...163 Table 6.5: Total cultivable Vibrio count (TCVC) in oysters challenged with OsHV-1 and oysters injected with an OsHV-1 negative tissue homogenate …….…………………………………..163 Table 6.6: Total Vibrio count in oysters challenged with OsHV-1 and in oysters injected with an OsHV-1 negative tissue homogenate…………………………………………………………..164 Table 6.7: Alpha diversity of gill microbiota following OsHV-1 challenge in oysters maintained in tidal emersion……………………………………………...... 171 Table 6.8: Temporal changes in the mean relative abundance of selected dominant genera of the gill microbiota of oysters under constant immersion.…………………………..……………...171 Table 6.9: Temporal changes in the mean relative abundance of selected, dominant genera of gill microbiota oysters under tidal emersion.…………………………………………………...... 172 Table 6.10: Summary of differential response to the OsHV-1 challenge under different immersion regimes ……………………………………………………..……………………...178 Table 7.1: OsHV-1 DNA concentration in gill and mantle tissues of live and dead oysters challenged with OsHV-1 and maintained with different water temperature profiles…...……..190 Table 7.2: Quantity of total bacterial DNA associated with the gill of oysters, before and after acclimation to the laboratory…………………………………………………………………...192 Table 7.3: Alpha diversity of gill microbiota in oysters challenged with OsHV-1 and controls and maintained at different temperature regimes……………………………………………….195

Table 7.4: Average relative abundance of dominant phyla in the gill microbiota of oysters, before and after the temperature treatments and in live oysters sampled 48h after OsHV-1 challenge………………………………………………………………………………………..196 Table 7.5: Average relative abundance of the dominant genera in the gill microbiota of oysters, before and after the temperature treatments and 48h post-OsHV-1 challenge…………………197

xii

List of Figures

Figure 2.1: Technique for dissecting larger oysters (shell length > 4 cm) using an oyster knife……………………………………………………………………………………………...40 Figure 2.2: Colony morphology of bacteria isolated on Marine salt agar and TCBS agar……..44 Figure 3.1: Kaplan-Meier survival curves for each batch (Clyde River, Patonga Creek, Shoalhaven River) of oysters challenged with OsHV-1………………………………………....74 Figure 3.2: Temporal distribution of OsHV-1 viral load, total Vibrio load and total bacterial load in live oysters…………………………………………………………………………………….77 Figure 3.3: Relative abundance of bacterial families in oysters farmed in different estuaries….80 Figure 3.4: Non-metric multidimensional scaling based on two-dimensional Bray-Curtis dissimilarity………………………………………………………………………………………81 Figure 4.1: Tissue compartmentalization of Crassostrea gigas microbiota…………………...105 Figure 4.2: Taxa plots summarizing the relative abundance of bacterial phyla in nucleic acid extracts obtained by different tissue-types using different extraction methods..……………….108 Figure 4.3: Bacterial diversity with different nucleic acid extraction methods………………..111 Figure 4.4: Effects of sampling and storage methods on bacterial community composition….112 Figure 5.1: Temporal changes in (A) total cultivable bacterial count and (B) total cultivable Vibrio count during acclimation to the laboratory environment………….…………………….130 Figure 5.2: Rarefaction curves for (A) gill microbiota (B) gut microbiota of oysters………...131 Figure 5.3: Taxa plots summarizing the changes in the relative abundance of bacterial phyla during acclimation to the laboratory environment ……………………………….…...………..135 Figure 5.4: Principal coordinate plots based on Bray-Curtis distances between the oyster microbiomes acclimated to the laboratory environment for different durations…………...…..140 Figure 6.1: Schematic representation of the experimental design showing the allocation of oysters (C. gigas) across the 4 recirculation systems and between treatment groups………….153 Figure 6.2: Kaplan-Meier survival curves for Pacific oysters challenged with OsHV-1 and maintained under constant immersion and under tidal emersion ………………………………158 Figure 6.3: Taxa bar plots indicating the relative abundance of bacterial phyla in the gill and gut microbiota of oysters under different regimes and after the OsHV-1 challenge ………………169 Figure 6.4: Box and whisker plots indicating the number of observed operational taxonomic units in oysters under different immersion regimes and after the OsHV-1 challenge…………170

xiii

Figure 7.1: Schematic representation of the experimental design showing the allocation of oysters (C. gigas) across the 4 temperature profiles……………………………………….…..184 Figure 7.2: Mean water temperature in oyster tanks recorded using temperature probes immersed in the water…………………………………………………………………………..188 Figure 7.3: Kaplan-Meier survival curves for Pacific oysters challenged with OsHV-1……...189 Figure 7.4: Principal coordinate plot based on Bray-Curtis distances between the gill microbiome of Pacific oysters………………………………………………………………….198 Figure 7.5: Taxa bar plots indicating the relative abundance of bacterial phyla in gill microbiota of oysters before and after acclimation………………………...………………………………199 Figure 7.6: Taxa bar plots indicating the relative abundance of bacterial phyla associated with the gill of oysters before and after OsHV-1 challenge and in negative control oysters………..203 Figure 7.7: Absolute abundance of the genus Vibrio in the gill microbiota of oysters challenged with OsHV-1 and maintained at different temperatures………………………………………..204

xiv

Abbreviations

ASW artificial seawater AVNV Acute viral necrosis virus BLOQ below the limit of quantification bp base pair CFU colony forming unit C. gigas Crassostrea gigas C. virginica Crassostrea virginica CTAB cetyltrimethylammonium bromide Ct cycle threshold DADA2 Divisive Amplicon Denoising Algorithm 2 g grams g (italicized) gravitational force h hour(s) kbp kilo-basepair kPa kilo Pascal L litres

MgCl2 Magnesium chloride mg milligrams min minutes mL millilitres

MSA-B Marine Salt Agar – Blood NaOCl Sodium hypochlorite OIE Office International des Epizooties (World Organisation for Animal Health) OsHV-1 ref Ostreid herpesvirus-1 reference strain OTU Operational taxonomic unit PCR Polymerase chain reaction

PO2 partial pressure of oxygen

PCO2 partial pressure of carbondioxide QIIME Quantitative Insights Into Microbial Ecology

xv qPCR quantitative PCR rRNA ribosomal RNA s seconds SPF specific pathogen free TCBS Thiosulfate-citrate-bile salts-sucrose v volume w weight µl microlitre µVar microvariant °C degree Celcius

xvi

Table of Contents Declaration of Authorship………………………………………………………………………..i Acknowledgements………………………………………………………………………………ii Summary…………………………………………………………………………………………iv List of Publications and Conference Presentations……………………………………………...vii List of Tables……………………………………………………………………………………..x List of Figures…………………………………………………………………………………...xiii Abbreviations…………………………………………………………………………………….xv CHAPTER 1 ...... 1 Literature Review ...... 1 1.1 Introduction ...... 1 1.2 Pacific oyster disease and mass mortality ...... 2 1.3 Ostreid herpesvirus-1 ...... 4 1.3.1 Genotypes of OsHV-1 ...... 4 1.3.2 Methods for detection of OsHV-1 ...... 6 1.4 Interactions between host, pathogen/s and the environment ...... 6 1.4.1 Host factors ...... 7 1.4.2 Environmental factors ...... 9 1.4.2.1 Seawater temperature ...... 9 1.4.2.2 Salinity ...... 10 1.4.2.3 Oyster farm management practices...... 11 1.4.2.4 Nutrient levels in seawater ...... 12 1.4.3 Pathogen factors...... 13 1.5 The Pacific oyster microbiome ...... 15 1.5.1 Role of the microbiome ...... 15 1.5.2 Microbiome analysis ...... 16 1.5.2.1 Bacterial 16S rRNA gene in microbial diversity profiling ...... 16 1.5.3 Pacific oyster microbiome studies ...... 17 1.5.3.1 Methods to characterize the oyster microbiome ...... 17 1.5.3.2 16S rRNA gene diversity profiling of the Pacific oyster microbiome ...... 19 1.5.3.3 Factors that influence the Pacific oyster microbiome ...... 19 1.5.3.4 Impact of the environment on the Pacific oyster microbiome ...... 30 1.5.4 Role of the Pacific oyster microbiome in mortality events ...... 32 1.6 Aims of this study ...... 33

xvii

CHAPTER 2 ...... 35 General Materials and Methods ...... 35 2.1 Laboratory facilities and oysters ...... 35 2.1.1 Experimental procedures in the PC2 Aquatic Animal Facility ...... 36 2.2 Reagents ...... 36 2.2.1 General Reagents ...... 36 2.2.2 Disinfectants ...... 38 2.3 Oysters ...... 38 2.3.1.1 Live oysters ...... 39 2.3.1.2 Dead oysters ...... 39 2.3.2 Dissection of oyster tissues ...... 39 2.4 Sampling oyster tissues ...... 40 2.4.1 Gill and mantle tissues for OsHV-1 DNA quantification ...... 40 2.4.2 Sampling gill and gut tissues for total bacteria and Vibrio quantification ...... 41 2.4.3 Soft tissue homogenization for bacteriology ...... 41 2.5 Conventional bacterial culture ...... 42 2.5.1 Preparation of bacterial culture media ...... 42 2.5.2 Isolation of bacteria from oyster tissue homogenates ...... 42 2.5.3 Identification and quantification of cultivable Vibrio and total bacteria ...... 43 2.5.4 Preparation and quantification of genomic DNA standards ...... 44 2.5.4.1 Vibrio genomic DNA ...... 44 2.5.4.2 Total bacterial genomic DNA ...... 45 2.6 Tissue homogenization for DNA purification ...... 46 2.6.1 Tissue homogenization for OsHV-1 qPCR assays ...... 46 2.6.2 Tissue homogenization for bacterial DNA extraction ...... 47 2.7 Nucleic acid purification ...... 48 2.7.1 Purification of nucleic acids for OsHV-1 assays ...... 48 2.7.2 Bacterial DNA extraction ...... 48 2.8 Real-time quantitative PCR for the detection of OsHV-1...... 49 2.8.1 OsHV-1 quantitative plasmid standard ...... 49 2.8.2 OsHV-1 qPCR assay ...... 50 2.8.3 PCR controls ...... 50

xviii

2.8.4 Quality control criteria ...... 52 2.9 Real-time quantitative PCR for the detection of Vibrio...... 53 2.9.1 Preparation of Vibrio genomic DNA standards ...... 53 2.9.2 Fast SYBR® Green real-time qPCR assay ...... 54 2.9.3 PCR controls ...... 54 2.9.4 Quality control criteria ...... 55 2.10 Real-time quantitative PCR for the detection of total bacteria ...... 55 2.10.1 Preparation of total bacteria genomic standards ...... 55 2.10.2 TaqMan® real-time qPCR assay ...... 56 2.10.3 PCR controls ...... 56 2.10.4 Quality control criteria ...... 56 2.11 Oyster microbiome analysis by high throughput 16S rRNA gene sequencing ...... 57 2.11.1 Bioinformatics ...... 58 2.11.2 Amplification of V1-V3 hypervariable regions of 16S rRNA gene ...... 58 2.11.3 Quality control of sequence reads ...... 58 2.11.4 Generation and validation of metadata ...... 59 2.11.5 Importing paired-end DNA sequence reads into the QIIME2 pipeline ...... 59 2.11.6 Sequence quality control (denoising) ...... 59 2.11.7 Visualizing the feature table and feature data summaries ...... 60 2.11.8 Alpha and beta diversity analysis ...... 60 2.11.9 Taxonomic analysis ...... 60 2.12 Sequence data management ...... 61

CHAPTER 3 ...... 62 Influence of the environment on the pathogenesis of Ostreid herpesvirus-1 (OsHV-1) infections in Pacific oysters (Crassostrea gigas) through differential microbiome responses ...... 62 3.1 Abstract ...... 62 3.2 Introduction ...... 63 3.3 Materials & Methods ...... 66 3.3.1 Oysters and aquarium management ...... 66 3.3.2 Source of infective OsHV-1 ...... 67 3.3.3 Challenge with OsHV-1 ...... 67

xix

3.3.4 Sampling...... 68 3.3.5 OsHV-1 DNA quantification ...... 68 3.3.6 Bacterial studies ...... 68 3.3.6.1 Isolation, identification and quantification of cultivable Vibrio and total bacteria 69 3.3.6.2 Quantification of total Vibrio spp. DNA by qPCR...... 69 3.3.6.3 Quantification of total bacterial DNA by qPCR ...... 69 3.3.7 Microbiome analysis by high throughput 16S rRNA gene sequencing ...... 70 3.3.8 Bioinformatic analyses ...... 71 3.3.9 Data analyses ...... 71 3.4 Results ...... 73 3.4.1 Mortality of oysters ...... 73 3.4.2 Quantification of OsHV-1 DNA by qPCR ...... 74 3.4.3 Quantification of bacteria ...... 75 3.4.4 Bacterial community structure in different batches of oysters ...... 79 3.4.5 Changes in bacterial community during OsHV-1 infection ...... 79 3.5 Discussion...... 81 3.6 Conclusion ...... 86

CHAPTER 4 ...... 87 The role of tissue type, sampling and nucleic acid purification methodology on the inferred composition of Pacific oyster (Crassostrea gigas) microbiome...... 87 4.1 Abstract ...... 87 4.2 Introduction ...... 88 4.3 Materials and Methods ...... 90 4.3.1 Oysters...... 90 4.3.2 Tissue sampling ...... 92 4.3.3 Tissue disruption and nucleic acid extraction ...... 93 4.3.4 Spiking with internal positive control (IPC) DNA ...... 95 4.3.5 Bacterial DNA quantification ...... 95 4.3.6 Analysis of bacterial DNA yields ...... 97 4.3.7 Analysis of extraction efficiency and PCR inhibition ...... 97 4.3.8 Microbiome analysis by high throughput 16S rRNA gene sequencing ...... 98

4.4 Results ...... 100 4.4.1 Bacterial DNA yield ...... 100 4.4.2 PCR inhibition ...... 102 4.4.3 Bacteria in nucleic acid extracts ...... 103 4.4.4 Bacterial diversity in different tissue types ...... 103 4.4.5 Bacterial diversity with different extraction and sampling methods ...... 109 4.4.6 Effects of frozen storage on bacterial DNA ...... 109 4.5 Discussion...... 113 4.6 Conclusion ...... 117

CHAPTER 5 ...... 118 Impact of the laboratory environment on the Pacific oyster (Crassostrea gigas) microbiome ...... 118 5.1 Abstract ...... 118 5.2 Introduction ...... 119 5.3 Materials and Method ...... 121 5.3.1 Oysters...... 121 5.3.2 Experimental design and oyster management ...... 122 5.3.3 Sampling and tissue processing ...... 122 5.3.4 Isolation and quantification of cultivable Vibrio and total bacteria ...... 123 5.3.5 Quantification of total Vibrio spp. DNA by qPCR ...... 123 5.3.6 Quantification of total bacteria DNA by qPCR ...... 124 5.3.7 Microbiome analysis by high throughput 16S rRNA gene sequencing ...... 124 5.3.8 Statistical analysis ...... 125 5.4 Results ...... 126 5.4.1 Total bacteria and total Vibrio quantity ...... 126 5.4.2 Changes in microbiome during laboratory acclimation ...... 130 5.4.2.1 High throughput 16S rRNA gene sequencing ...... 130 5.4.2.2 Bacterial community composition of oysters from the field ...... 130 5.4.2.3 Microbiome of oysters during acclimation ...... 137 5.5 Discussion...... 141 5.6 Conclusion ...... 146

xxi

CHAPTER 6 ...... 147 Impact of constant immersion compared to tidal emersion on the microbiome of Pacific oysters (Crassostrea gigas) challenged with Ostreid herpesvirus-1 (OsHV-1) ...... 147 6.1 Abstract ...... 147 6.2 Introduction ...... 148 6.3 Materials and Method ...... 151 6.3.1 Oysters...... 151 6.3.2 Experiment design and oyster management ...... 152 6.3.3 Challenge with OsHV-1 ...... 154 6.3.4 Sampling...... 155 6.3.5 OsHV-1 DNA quantification ...... 155 6.3.6 Identification and quantification of cultivable Vibrio and total bacteria ...... 155 6.3.7 Molecular quantification of bacteria ...... 156 6.3.8 Microbiome analysis by high throughput 16S rRNA gene sequencing ...... 156 6.3.9 Statistical analysis ...... 156 6.4 Results ...... 158 6.4.1 Mortality ...... 158 6.4.2 Quantification of OsHV-1 DNA ...... 159 6.4.3 Quantification of total bacteria and total Vibrio ...... 160 6.4.4 Changes in bacterial community structure after OsHV-1 challenge ...... 165 6.4.4.1 High throughput 16S rRNA gene sequencing ...... 165 6.4.4.2 Changes in the bacterial community composition ...... 165 6.5 Discussion...... 172 6.6 Conclusion ...... 177

CHAPTER 7 ...... 179 Influence of seawater temperature on the Pacific oyster (Crassostrea gigas) microbiome and its impact on Ostreid herpesvirus-1 (OsHV-1) infection ...... 179 7.1 Abstract ...... 179 7.2 Introduction ...... 180 7.3 Materials and Methods ...... 182 7.3.1 Oysters...... 182 7.3.2 Experiment design and aquarium management ...... 183

xxii

7.3.3 Challenge with OsHV-1 ...... 185 7.3.4 Sampling...... 185 7.3.5 Molecular quantification of OsHV-1 ...... 186 7.3.6 Molecular quantification of bacteria ...... 186 7.3.7 Microbiome analysis by high throughput 16S rRNA gene sequencing ...... 186 7.3.8 Statistical analyses ...... 187 7.4 Results ...... 188 7.4.1 Water temperature ...... 188 7.4.2 Mortality ...... 188 7.4.3 OsHV-1 detection and quantity ...... 190 7.4.4 Total bacteria and total Vibrio quantity ...... 191 7.4.5 Bacterial community composition ...... 192 7.4.5.1 High throughput 16S rRNA gene sequencing ...... 192 7.4.5.2 Changes in the bacterial community composition ...... 193 7.5 Discussion...... 205 7.6 Conclusion ...... 209

CHAPTER 8 ...... 210 General Discussion ...... 210 8.1 Introduction ...... 210 8.2 Accurate identification of the oyster microbiome ...... 213 8.3 Dynamic changes in the Pacific oyster microbiome ...... 214 8.4 Microbiome studies in a controlled environment ...... 216 8.5 Oyster microbiome dynamics during acclimation to the laboratory environment ...... 217 8.6 Environmental influence on the microbiome and OsHV-1 associated oyster mortality ………………………………………………………………………………………...219 8.7 Vibrio as a commensal organism and as a pathogen in oysters...... 222 8.8 Conclusion ...... 225 REFERENCES ...... 227

xxiii

CHAPTER 1

Literature Review

Environmental influences on the Pacific oyster (Crassostrea gigas) microbiome and disease associated with Ostreid herpesvirus-1 (OsHV-1)

1.1 Introduction

Pacific oysters (Crassostrea gigas) make the greatest contribution to global oyster production, amounting to 625,925 tonnes, worth US$ 1.3 billion out of a total oyster production of 5.2 million tonnes worth US$ 4.2 billion (FAO, 2014). At present, the Pacific oyster is cultivated in more than 40 countries across all continents except Antartica (Alfaro et al., 2019). While 84% of the global Pacific oyster production occurs in China, there is also over 100,000 tonnes per year produced in Japan, Korea and France (FAO, 2014). The origin of Pacific oyster farming was in Japan, where it is endemic and has been cultivated for centuries. The species was subjected to a widespread introduction across the globe, including to Australia in 1940 and France in 1966, owing to its potential for rapid growth and wide range of tolerance to environmental conditions (FAO, 2005).

In the recent past Pacific oyster farming was greatly impacted in Australia (de Kantzow et al., 2017; Jenkins et al., 2013; Paul-Pont et al., 2014), New Zealand (Keeling et al., 2014) and in Europe (Clegg et al., 2014; Garcia et al., 2011; Peeler et al., 2012; Roque et al., 2012; Segarra et al., 2010) by outbreaks of severe, widespread mortality. Among the multiple aetiologies of Pacific oyster mortality including pathogens such as viruses, bacteria and parasites (Bower et al., 1994; Paul-Pont et al., 2013b), viral pathogens are one of the most important groups, particularly OsHV-1. However, as with all organisms, diseases in Pacific oysters are multifactorial, involving the environment, host factors, pathogens (EFSA, 2010; Samain and McCombie, 2008) and different combinations of risk factors triggered similar mass mortality events (de Kantzow et al., 2017; Go et al., 2017; Jenkins et al., 2013; Vezzulli et al., 2010). As diagnostic modalities have improved over time, the ability to differentiate primary and secondary aetiologies of the diseases has improved (Bruto et al., 2017; Petton et al., 2015b; Segarra et al., 2010). The diseases also follow a complex pathogenesis in which damage from adverse environmental conditions and primary pathogens can contribute and be exacerbated by secondary environmental bacteria and

commensal bacteria that can overgrow or come out of balance (Alfaro et al., 2019; de Lorgeril et al., 2018). Recent advances in molecular biology improved our understanding of the oyster microbiome, facilitating the emerging view of the polymicrobial pathogenesis of oyster mortality disease (de Lorgeril et al., 2018; Lasa et al., 2019; Petton et al., 2019).

The objective of this literature review was to document the impact of mass mortality disease on the productivity of Pacific oyster aquaculture. Further, as technology to study the etiology of Pacific oyster diseases has improved, to identify how an understanding of the complex pathogenesis of the diseases might help to devise disease control programs that mitigate the severity of disease on oyster production.

1.2 Pacific oyster disease and mass mortality

Natural rates of mortality in Pacific oysters are typically <10% per year, whereas 30–70% of oysters may die during a mass mortality event (Chaney and Gracey, 2011; Soletchnik et al., 2005). Mass mortalities of Pacific oysters are associated with multiple factors including environmental stress factors, oyster factors and pathogens (OIE, 2014; Samain and McCombie, 2008). Multifactorial diseases result from a combination of host, environment and pathogen factors that interact in determining the incidence and severity of disease. These would be the seasonal disease syndromes which can have different combinations of disease risk factors for each occurrence.While environmental factors such as elevated seawater temperatures (>16°C), freshwater inputs, trophic resources and the sediment in the environment were directly or indirectly implicated to mass mortalities of Pacific oysters (Samain and McCombie, 2008), oyster factors such as physiological stress associated with reproductive maturation along with aquaculture practices, pathogens and pollutants were also associated with summer mortality (Goulletquer et al., 1998; Samain et al., 2007; Soletchnik et al., 1997). As most mass mortality events of Pacific oysters were linked with elevated seawater temperatures the mass mortality events of Pacific oysters were termed summer mortality events (Chaney, 2011).

Disease outbreaks that result in mass mortality are a significant constraint to commercial mollusc production and trade (Paillard et al., 2004; Renault, 2011). Throughout the development of oyster culture, infectious diseases were among the major limiting factors, owing to the massive losses caused during outbreaks and thus, continuous need for effective disease control

and management exists (Paul-Pont et al., 2013b; Pernet et al., 2016). Infectious diseases of molluscs can be caused by a wide variety of pathogens including protozoans and metazoans (genera Perkinsus, Marteilia, Bonamia, Microcytos, Mytilicola), bacteria (Vibrio species, Nocardia) and viruses (Iridoviridae, Papoviridae, Togaviridae, Reoviridae, Birnaviridae, Picornaviridae and Malacoherpesviridae) (Bower et al., 1994; Elston, 1993; Garnier et al., 2007; Paul-Pont et al., 2013b; Travers et al., 2015). Nocardiosis caused by the bacteria Nocardia crassostreae affected adult Pacific oysters causing green pustules in oyster tissues and mortality (Friedman and Hedrick, 1991). Several pathogenic Vibrio species have been reported as causing mass mortality diseases of Pacific oysters (Alfaro et al., 2019; Lasa et al., 2019; Petton et al., 2015b; Saulnier et al., 2010). Vibriosis caused by diverse Vibrio species such as Vibrio splendidus and Vibrio tubiashii affected oyster larvae and spat (defined as oysters <12 months old), and caused mortalities at hatchery level (Elston et al., 2008; Sugumar et al., 1998). Among protozoans, Haplosporidium species affected both spat and adults, causing mild to severe infestations in spat (Elston, 1993). Irido-like viruses are thought to have caused the destruction of Portuguese oyster (Crassostrea angulata) stocks between 1967 and 1973 along the Atlantic coast in France (Renault, 2011). A similar virus was then associated with disease of Pacific oysters during a summer mortality event in France in 1977 (Elston, 1993). In the recent past, Ostreid herpesvirus 1 (OsHV-1) and their microvariant genotypes affected both spat and adult oysters causing mass mortality (Garcia et al., 2011; Jenkins et al., 2013; Renault, 2011; Renault et al., 1995). Out of all infectious agents, viruses and particularly herpes-like viruses are of particular concern due to their commercial and ecological impact on marine molluscs over the past 20 years (Paul-Pont et al., 2013b; Renault, 2011).

Outbreaks of summer mortality in Pacific oysters were first reported in Japan, the native habitat of C. gigas in 1950 (Takeuchi et al., 1960). The disease impacted both wild populations and harvested beds of adult oysters and was attributed to elevated seawater temperature (Chaney, 2011; Delaporte et al., 2007; Imai et al., 1965; Koganezawa, 1975). This syndrome has since affected adult Pacific oysters in many parts of the world, including Italy (Takeuchi et al., 1960), the USA (Glude, 1974) and France (Maurer et al., 1986). Since 1993, mass mortality events were reported from France, in C. gigas that were less than 1 year of age (spat) (Renault et al., 1994a). The disease of oysters less than one year old (spat) was associated with Ostreid herpesvirus-1 (OsHV-1) infection (Davison et al., 2005; Garcia et al., 2011). The reference genotype of OsHV- 3

1 (Davison et al., 2005; Le Deuff and Renault, 1999) and related genotypes have been identified as prominent pathogens which caused mortality of C. gigas in France from 1991 to 2008 (Martenot et al., 2011; Renault et al., 1994b; Renault et al., 2012) and in the USA (Burge et al., 2006; Friedman et al., 2005). However, at the end of spring 2008, widespread mortalities were reported in France which killed young oysters and a genomic variant of OsHV-1, called µVar, was identified from these outbreaks (Martenot et al., 2011; Renault et al., 2012; Segarra et al., 2010). Since then, severe disease of Pacific oysters associated with the OsHV-1 µVar and closely related microvariant genotypes was observed in different parts of the world, including Australia (Jenkins et al., 2013; Paul-Pont et al., 2015; Whittington et al., 2015b), New Zealand (Renault et al., 2012), Ireland (Clegg et al., 2014; Peeler et al., 2012), Spain (Roque et al., 2012) and Scandinavia (Mortensen et al., 2016). The acronym POMS (Pacific Oyster Mortality Syndrome) was first used in Australia to refer to mass mortality disease caused by microvariant genotypes of OsHV-1(Paul-Pont et al., 2013b). Up until now, infections in Pacific oysters caused by any of the OsHV-1 genotypes are not listed as notifiable diseases by the World Organization for Animal Health (OIE, 2014).

1.3 Ostreid herpesvirus-1

Ostreid herpesvirus-1 belongs to the family Malacoherpesviridae, order Herpesvirales, and is one of only two known herpesviruses to infect invertebrates, the other being a neurotropic herpesvirus infecting the gastropod, abalone (Haliotis species) (Davison et al., 2009; Savin et al., 2010). The virions of OsHV-1 have a linear, double-stranded DNA genome contained within an icosahedral capsid and surrounded by a proteinaceous tegument and a lipid envelope (Le Deuff and Renault, 1999). The OsHV-1 virion is approximately 116 nm in diameter (Le Deuff and Renault, 1999).

1.3.1 Genotypes of OsHV-1

Initially sequenced from infected C. gigas larvae collected from France in 1995, the OsHV-1 genome consists of 207,439 base pairs (Davison et al., 2005). The genome structure has similarities to mammalian herpesviruses, with two invertible unique regions (Unique Long (UL), 167.8 kbp; Unique Short (US), 3.4 kbp) each flanked by inverted repeats Terminal Repeat Long (TRL)/ Inverted Repeat Long (IRL), 7.6 kbp; Terminal Repeat Short (TRS)/Inverted Repeat Short (IRS), 9.8 kbp), with an additional unique sequence (X, 1.5 kbp) between IRL and IRS

(Davison et al., 2005). Of the 124 unique genes predicted from the genome, 38 encoded proteins relate to helicases, inhibitors of apoptosis and deoxyuridine triphosphatase (Davison et al., 2005).

This sequence was the first complete genome of OsHV-1 to be described and thus is considered as the reference genotype (OsHV-1 ref; GenBank accession number AY509253) (OIE, 2014). The reference genotype of OsHV-1 was identified as the main pathogen in Pacific oyster mortality that occurred in France from 1995 to 2008 (Martenot et al., 2011). A genotypic variant of OsHV-1 (OsHV-1 var) was detected in 2001 which was also associated with mortalities of C. gigas in French hatcheries (Arzul et al., 2001a). The OsHV-1 var was different to the reference genotype by a 2800 bp deletion in the TRL region and was also detected in Manila clams (Ruditapes philippinarum) (Arzul et al., 2001a).

Since 2008, severe mortalities of Pacific oysters associated with OsHV-1 μVar were detected in different parts in France (Segarra et al., 2010). The variant OsHV-1 μVar was characterized mainly by 12 consecutive deletions followed by one deletion of an adenine in the C region of the viral genome (Martenot et al., 2011; Segarra et al., 2010). Since the mass mortality outbreaks in France in 2008, OsHV-1 μVar has become predominant and the majority of the Pacific oyster mortality events reported in France were associated with OsHV-1 μVar and not OsHV-1 ref (Dégremont et al., 2015; Martenot et al., 2011). Phylogenetic studies have suggested that OsHV-1 ref and OsHV-1 μVar share a common ancestor (Renault et al., 2012). However, the same study states that it is unlikely that μVar genotype descended directly from OsHV-1 ref. OsHV-1 μVar has also been reported across Europe since 2008, in Ireland (Peeler et al., 2012), Spain (Roque et al., 2012), Scandinavia (Mortensen et al., 2016) and in other parts of the world, including Australia (Jenkins et al., 2013; Paul-Pont et al., 2015; Whittington et al., 2015b) and New Zealand (Renault et al., 2012).

Other microvariant genotypes of OsHV-1 (OsHV-1 μVar 9, OsHV-1 μVar 15 and undescribed variants) have also been reported in France (Martenot et al., 2011), Italy (Burioli et al., 2016), Ireland (Lynch et al., 2012), Australia (Jenkins et al., 2013) and New Zealand (Keeling et al., 2014; Renault et al., 2012). In summary, OsHV-1 ref and a variant and OsHV-1 microvariant genotypes have been recorded to be responsible for mass mortalities in Pacific oysters in different parts of the world.

Apart from the reports of mass mortalities in Pacific oysters, OsHV-1 and its variants have also been reported in other bivalve species including Eastern oyster (Crassostrea virginica), Hong Kong oyster (Crassostrea hongkongensis), European flat oyster (Ostrea edulis), Sydney rock oyster (Saccostrea glomerata), Zhikong scallop (Chlamys farreri) and various other species of scallops, clams, cockles and mussels (Alfaro et al., 2019; Arzul et al., 2001b; Bai et al., 2015; Evans et al., 2017). In 1990, mass mortalities were reported in commercially farmed Zhikong scallop in China (Yu, 1998). Although the causative agent of this mortality outbreak was initially identified as Acute viral necrosis virus (AVNV), this virus was later discovered to be a genotypic variant of OsHV-1 (Bai et al., 2015; Ren et al., 2013). While this variant had a 97% similarity to OsHV-1 ref, several insertion and deletions were present in the non-coding region (Ren et al., 2013). In 2000, a sporadic mortality event with 100% mortality was reported in great scallop (Pecten maximus) larvae in France (Arzul et al., 2001b). The cause of mortality was identified to be a variant of OsHV-1 and was the first OsHV-1 outbreak in scallops in France (Arzul et al., 2001b). Since then, OsHV-1 and its variants have been reported in European flat oyster in France (Arzul et al., 2001b), cultured blood clams (Scapharca broughtonii) in China (Xia et al., 2015) and various other species of bivalves (Batista et al., 2015; Guichard et al., 2011).

1.3.2 Methods for detection of OsHV-1

In situ hybridization (ISH) (Arzul et al., 2002), polymerase chain reaction (PCR) and transmission electron microscopy (TEM) (OIE, 2014) have been used in OsHV-1 detection in C. gigas tissues. Out of these methods, PCR methods have a higher specificity and sensitivity (Martenot et al., 2010; OIE, 2014). ISH and TEM are not used as standalone methods in OsHV-1 detection (OIE, 2014). While ISH is used alongside PCR to confirm the presence of OsHV-1 in tissues, TEM is used to identify the location of OsHV-1 viral particles in tissues after confirmatory diagnosis with PCR (OIE, 2014). Different PCR methods can be used to detect OsHV-1 in oyster tissues which includes conventional PCR, nested PCR and real-time quantitative PCR methods. A major limitation of PCR methods is that these methods cannot differentiate between live and dead OsHV-1 viruses.

1.4 Interactions between host, pathogen/s and the environment

Despite the causal relationship of OsHV-1 with Pacific oyster mortality, it is interesting to note that OsHV-1 microvariants has been reported in Pacific oysters without any associated

mortality (Dundon et al., 2011; Shimahara et al., 2012). Historically, mass mortality episodes of Pacific oysters were determined to be associated with elevated seawater temperature, high input of nutrients resulting from eutrophication in water, and physiological stress associated with maturation or spawning (Le Roux et al., 2016; Samain et al., 2007). It is now believed that the cause of mortality events in Pacific oysters is multifactorial, with OsHV-1 infection (mainly microvariant genotypes of OsHV-1) considered a necessary but not a sufficient cause (EFSA, 2010; Samain and McCombie, 2008). It is most often considered to be a result of complex interactions between the physiological status of the oysters, the environment and multiple pathogens (Alfaro et al., 2019; de Lorgeril et al., 2018; Pernet et al., 2018; Samain et al., 2007).

1.4.1 Host factors

The scale of OsHV-1-induced mortality is potentially influenced by host factors such as age, genetic background, life-history traits and physiological status of oysters (Pernet et al., 2016). Pacific oysters of all ages are susceptible to infection with OsHV-1. However, spat and juvenile oysters experience higher mortality (de Kantzow et al., 2017; Dégremont, 2013; Paul- Pont et al., 2014; Renault et al., 1994b; Schikorski et al., 2011) as is also the case with other herpes viral infections affecting other aquatic organisms (Lepa and Siwicki, 2013). Spat mortality ranging from 60 to 100% has been recorded in both Australia and Europe (Garcia et al., 2011; Martenot et al., 2011; Segarra et al., 2010; Whittington et al., 2015a). Comparatively lower mortality was reported in adult oysters in Europe, while significant mortalities have been reported in adult oysters in Australia (40 - 75%) and New Zealand (Alfaro et al., 2019; de Kantzow et al., 2017; Paul-Pont et al., 2014; Segarra et al., 2010; Whittington et al., 2015a). Field experiments conducted in Australia and France have shown an impact of oyster age on OsHV-1 associated mortalities, with a 1.4 - 4.4 times higher hazard of death in young oysters (approximately 10 months old) compared to adult oysters (approximately 18 months old) (Hick et al., 2018; Paul-Pont et al., 2013b; Petton et al., 2015a). An interaction between the age and the environment was observed in a study where batches of oysters responded differently to successive summer mortality events, with different mortality patterns in different environments (Dégremont et al., 2010).

In addition to age, the size of oysters also plays a role in the susceptibility to OsHV-1 infections (Hick et al., 2018). In a study conducted in France, Dégremont (2013) showed that the

resistance to mortality caused by OsHV-1 increased as the size and the age of the oyster increased. Various studies conducted in the field and in laboratories have reported better survival in larger and older oysters compared to smaller and younger oysters (Azéma et al., 2017a; de Kantzow et al., 2017; Dégremont, 2013; Oden et al., 2011; Paul-Pont et al., 2013b; Peeler et al., 2012; Whittington et al., 2015b). In this regard, it is important to note that size of oysters of a particular age can be variable and can be manipulated via growth by changing the environment and farm management (Hick et al., 2018). However, size of the oysters alone was not significant when different age classes were considered (Paul-Pont et al., 2014). Hick et al. (2018) demonstrated that size alone was not protective against OsHV-1 associated mortality, by manipulating the size of oysters through growth restriction.

There is also a role of oyster genotype in susceptibility to viral infections and mortality events (Azéma et al., 2017b; de Lorgeril et al., 2018; Dégremont, 2011). Genetic variation in oysters can be manipulated through selective breeding programmes aiming to produce offspring with improved disease resistance (Alfaro et al., 2019; Dégremont et al., 2007; Dégremont et al., 2005). Dégremont et al. (2005) studied the relative importance of a genetic basis for resilience to disease along with rearing site and timing of field placement, growth, and yield, using three successive groups of bi-parental families. Among the significant differences observed in survival after OsHV-1 disease, growth and yield of oysters, among each site and between families, the differences were greater between families. The largest variation in survival (46%) was attributed to family, indicating a significant influence of genetics on survival of oyster spat (Dégremont et al., 2005). Azéma et al. (2017a) studied the influence of oyster genetics with different farm management practices. When small and large oyster spat from 40 families of Pacific oysters were tested for two years in the field, the mean cumulative mortality was lower (54%) in large oysters compared to smaller oysters (75%), when first exposed to OsHV-1. There was a large variation between families in susceptibility to OsHV-1, regardless of size, with total cumulative mortality ranging from <32% to >80%. Azéma et al. (2017a) also studied how different family lines respond to growing height and observed that oysters can be susceptible or resistant to disease caused by OsHV-1 infection, regardless of the growing height but under the influence of genetics. During early stages of life, oysters could be selectively bred to improve resistance to OsHV-1 infection but not for V. aestuarianus infections (Azéma et al., 2017b).

1.4.2 Environmental factors

Various environmental factors have been implicated with Pacific oyster mortality, including seawater temperature, freshwater inputs salinity, pH, turbidity of water, oyster farm management practices, nutrient levels in seawater and dissolved oxygen concentration in water (Malham et al., 2009; Soletchnik et al., 2007).

1.4.2.1 Seawater temperature

Seawater temperature is a key factor associated with Pacific oyster mortality (Clegg et al., 2014; Petton et al., 2013; Renault et al., 1994b). An increase or a sudden change in the temperature of the water around oysters has been shown to be an important risk factor predisposing to disease (EFSA, 2010). Being poikilotherms, the physiological processes of oysters such as assimilation of food, maintenance and growth depend on seawater temperature (Petton et al., 2013). The rates of physiological processes increased exponentially with increases in temperature, within a species-specific temperature tolerance range (Petton et al., 2013). The general temperature tolerance range for Pacific oysters is from 3°C to 32°C (Bourlès et al., 2009). Application of this temperature tolerance range in an oyster dynamic energy budget model indicated that oyster growth was retarded at temperatures >25°C (Bourlès et al., 2009; Petton et al., 2013).

Elevated seawater temperatures have long been associated with Pacific oyster mortality events in France (Goulletquer et al., 1998; Samain and McCombie, 2008; Samain et al., 2007). In Australia, elevated seawater temperature (>20°C) and low salinity (< 20 ppt) were attributed to be causative for an outbreak of Pacific oyster mass mortality that occurred in Port Stephens, NSW, in the absence of a specific infectious pathogen (Go et al., 2017). Considering OsHV-1 related oyster mortality, the threshold temperature above which mortality was recorded was 16°C in Europe (Clegg et al., 2014; Pernet et al., 2012; Renault et al., 2014). However, in Australia, mortality outbreaks occurred at temperatures that were 4-5°C warmer than that recorded in France and other European countries, even when OsHV-1 was present at lower water temperatures earlier in the season (Paul-Pont et al., 2014; Whittington et al., 2019). Pernet et al. (2012) suggest 24°C as the upper threshold temperature for oyster mortality associated with OsHV-1, based on observation of disease outbreaks in the field. However, laboratory studies carried out in Australia (de Kantzow et al., 2016) and in France (Delisle et al., 2018) have

demonstrated OsHV-1 infection can cause mortality at 26°C. When taken together studies suggest OsHV-1 outbreaks are most likely to occur at water temperatures between 16-26°C. It is important to note that oysters farmed in intertidal estuarine environments can be exposed to relatively extreme environmental temperatures when emersed in air compared to a relatively stable seawater temperature (Allen and Burnett, 2008). Moreover, the patterns of immersion and emersion resulted in fluctuation of respiratory variables of the hemolymph of oysters (Table 1.1).

Table 1.1 Fluctuation of respiratory variables in oyster haemolymph during immersion and emersion (Allen and Burnett, 2008).

Immersion status pH PO2 (kPa) PCO2 (kPa) Submerged 18 °C 7.5 7.1 0.2 Emersed for 4 h 22°C 7.1 3.8 0.4 Emersed for 4 h 30 °C 6.8 3.1 1.3

Laboratory studies in Australia have identified that the dose of exposure to OsHV-1 interacts with the seawater temperature to influence the severity of disease (de Kantzow et al., 2016; Whittington et al., 2019). Experimental exposure of Pacific oysters to OsHV-1 in water below 18°C did not result in mortality. While some mortality occurred at 18°C with a higher dose of virus (106 genome copies/100 µl), the lower dose (103 genome copies/100 µl) which did not cause mortality at 18°C led to high mortality at 22°C and 26°C indicating that the dose response was influenced by the water temperature (de Kantzow et al., 2016). In another laboratory study Petton et al. (2013) found that OsHV-1 transmission was optimal in the temperature range of 16 °C −22°C. The stability of OsHV-1 in seawater at different temperatures also influences the transmission of OsHV-1. Infectivity of OsHV-1 declined below detection after 48 h at 20°C (Hick et al., 2016). OsHV-1 RNA associated with oyster mortality was recorded even after 54 h at 16°C compared to 33 h at 25°C (Martenot et al., 2015).

1.4.2.2 Salinity

As oysters are frequently grown in estuaries and bays, they are often exposed to wide fluctuations in salinity compared to oceanic water; this is caused by evaporation, rainfall and freshwater inflow from rivers (Fuhrmann et al., 2016; Soletchnik et al., 2007). In a controlled, laboratory experiment, Nell and Holliday (1988) identified a wide salinity range between 15 –

45 ‰ which did not impact the survival and permitted optimum growth of C. gigas spat. In a field experiment conducted in the USA, there was a higher mortality (63 %) in 10 month old Pacific oyster spat compared to eastern oysters (C. virginica) (10 %), at a low salinity (<15 ‰) (Calvo et al., 1999). However, the same experiment reported a low mortality (<11 %) in both species at a salinity >25 ‰, with a comparatively faster growth in Pacific oyster spat within a period of one year when salinity exposure was controlled by deploying oysters in different sites. Soletchnik et al. (2007) also observed a low autumn–winter salinity (9-3 ‰) associated with rainfall and increased mortality of Pacific oysters in a field study in France. In the same study, the oysters were less affected by mortality events at sites that were less impacted by freshwater input and had lower variations in temperature.

In general, bivalves experiencing changes in salinity have altered immune function and become more susceptible to infection and disease (Matozzo and Marin, 2011). Contrary to the higher mortality observed in Pacific oysters at low salinity, Fuhrmann et al. (2016) observed a low mortality (<5%) in oysters exposed to OsHV-1 when they were acclimated to a low salinity (10‰). The OsHV-1 challenge resulted in a higher mortality (27-57%) for oysters acclimated to higher salinities (15-35 ‰). Further investigation indicated increased adenosine monophosphate- activated protein kinase (AMPK) and lower hexokinase (HK) activity leading to increased energetic reserves of the oysters at low salinity and provided an explanation for the increased survival in OsHV-1 infection (Fuhrmann et al., 2018) .

1.4.2.3 Oyster farm management practices

Various factors related to aquaculture practices, such as choice of farming site, type of growing infrastructure and stocking density affect disease incidence and severity and can be used for management of oyster diseases (Alfaro et al., 2019; de Kantzow et al., 2017; Lightner, 2011; Murray and Peeler, 2005). At the farm level, modification of growing practices resulted in a reduction of OsHV-1 associated mortality (Paul-Pont et al., 2013b; Pernet et al., 2012; Whittington et al., 2015b). Oysters that were not exposed to OsHV-1 and maintained in the Mediterranean sea exhibited mass mortalities (80%) during the first 2 years of age, when transferred to baskets in the Thau lagoon, France, where OsHV-1 associated mortalities have been reported every year since 2008 (Pernet et al., 2012). At the same time the oysters that were deployed outside the farming area of the Thau lagoon showed no mortality, and OsHV-1 was not

detected. However, the mortality was approximately 30% for oysters grown cemented onto ropes, compared to 80 % when grown in baskets, suggesting an impact of the rearing structures (Pernet et al., 2012). It should be noted that the baskets were maintained outside the farming area where oysters were cemented on to ropes. By increasing the intertidal growing height of oysters by 300 mm above the standard rack height used by farmers, Paul-Pont et al. (2013b) demonstrated a reduction in adult oyster mortality at the higher height (<50%) compared to the standard height (80%), during an OsHV-1 outbreak in Woolooware Bay, New South Wales, Australia. It was suggested that improved survival with lower immersion time was a consequence of reduced exposure to OsHV-1 (Paul-Pont et al., 2013b). However, there was no effect of increased growing height on mortality of oyster spat (Paul-Pont et al., 2013b; Whittington et al., 2015b). Further investigation by Whittington et al. (2015b) revealed that increases in growing height (particularly, 600mm above the standard rack height) resulted in a growth penalty due to reduced feeding opportunities with limited tidal immersion. In addition to the role of growing infrastructure, manual handling of oysters during routine production such as grading into size categories was associated with increased the mortality during OsHV-1 disease if undertaken a short period prior to the risk period for disease (de Kantzow et al., 2017; Peeler et al., 2012).

1.4.2.4 Nutrient levels in seawater

Eutrophication occurs when an excessive richness of nutrients in water, frequently due to run-off from the land causes excessive algal growth and is one of the most severe and widespread forms of disturbance affecting coastal marine systems (Gray et al., 2002). Causes include catchment disturbance (Paterson et al., 2003) due to farming, application of fertilizers, deforestation and discharge of domestic wastewaters (Wu, 2002). Eutrophication has been associated with Pacific oyster mortality (Malham et al., 2009; Mori, 1979). Peaks of dissolved organic matter (cDOM) in seawater at two study sites preceded oyster mortality events by 2 weeks while another site with elevated cDOM levels was not followed by a mass mortality event in oysters (Malham et al., 2009). In some cases the mortality events were associated with over- maturation of the gonads (Mori, 1979). It was proposed that the high level of nutrient input resulted in excessive investment of energy from trophic resources into sexual maturation (Soletchnik et al., 2007). In this context, Mori (1979) showed that over-maturation of oocytes due to high trophic levels in oysters in Matsushima Bay (Miyagi Prefecture, Japan), led to 12

physiological and metabolic disorders inducing high mortality. Moreover, changes in food quality and oyster energy reserves have been thought to affect OsHV-1 expression (Pernet et al., 2014; Tamayo et al., 2014).

1.4.3 Pathogen factors

Vibrio species are ubiquitous marine bacteria in planktonic and animal-associated microbial communities (Le Roux et al., 2016; Thompson et al., 2004a). They are Gram-negative bacteria belonging to the phylum Proteobacteria and family Vibrionaceae. An increasing number of Vibrio species and strains have been found to be pathogenic for fish and shellfish, so they have become a source of concern in aquaculture (Saulnier et al., 2010). Vibrio species such as V. coralliilyticus, V. splendidus, V. crassostreae, V. tasmaniensis can be considered non- pathogenic, commensal bacteria in healthy oysters that can become opportunistic with compromised health caused by infection or occurrence of stressful events such as temperature stress (Lasa et al., 2019; Vezzulli et al., 2010). In this circumstance, it is often not clear whether Vibrio species isolated from diseased oysters are the causative agent, secondary opportunistic pathogens, or commensals (Le Roux et al., 2016).

The role of bacteria in summer mortality of oysters was first questioned after demonstration of high loads of Vibrio spp. in the haemolymph of moribund oysters (Lipp et al., 1976). Since then, a number of pathogenic Vibrio species have been isolated during mass mortality outbreaks of Pacific oysters, both in the presence of OsHV-1 (Keeling et al., 2014), and in the absence of OsHV-1 (De Decker et al., 2011; Garnier et al., 2007; King et al., 2018a). The virulence of several isolates was demonstrated using challenge studies that caused mortality, including: Vibrio splendidus (Lacoste et al., 2001; Le Roux et al., 2002; Pernet et al., 2012), Vibrio aestuarianus (Garnier et al., 2008; Saulnier et al., 2009) and Vibrio harveyi (Saulnier et al., 2010). Both V. splendidus and V. aestuarianus were isolated from C. gigas, in a longitudinal study carried out during the 2010-2011 summer mortality outbreak associated with OsHV-1 infection in New Zealand (Keeling et al., 2014). Further, Gay et al. (2004) experimentally demonstrated the pathogenicity of two strains belonging to the Splendidus super-clade of the genus Vibrio, that were isolated from summer mortality outbreaks. Recently, a novel species (Vibrio crassostreae) was identified from the haemolymph of diseased oysters (Bruto et al., 2017). This species was considered to be a benign oyster colonizer that developed virulence by

introgression of a virulence plasmid (Bruto et al., 2017). In addition, Vibrio tubiashii isolated from shellfish hatcheries in North America has been reported to cause disease of larval and juvenile C. gigas (Elston et al., 2008; Estes et al., 2004). Vibrio coralliilyticus re-emerged in the USA in 2005, causing catastrophic losses in larvae and juveniles in Pacific oyster hatcheries (Elston et al., 2008). It is considered that oysters are usually colonized by a diverse assemblage of vibrios (Le Roux et al., 2016; Petton et al., 2015b). In diseased oysters these resident vibrios can be mutualistic, opportunistic or pathogenic (Lemire et al., 2015).

Metalloprotease is a virulence factor in pathogenic Vibrio strains of Pacific oysters (Saulnier et al., 2010). It has been shown to be an essential determinant of lethality in Vibrio extracellular products injected into C. gigas (Le Roux et al., 2007). Metalloprotease-like activity in culture supernatant fluid has been used as a laboratory test for virulence of Vibrio spp. isolates (Saulnier et al., 2010). Metalloproteases have been reported in some strains of V. splendidus, V. aestuarianus and V. tubiashii that are pathogenic to Pacific oysters (Labreuche et al., 2006; Le Roux et al., 2007; Saulnier et al., 2010).

Petton et al. (2015b) observed rapid colonization of diverse vibrios preceding OsHV-1 replication in a natural mortality outbreak of Pacific oysters. Further, the same authors showed that a high load of OsHV-1 alone was insufficient to induce the full expression of the disease, by demonstrating a potential role of bacteria in the pathogenesis of the disease. In this study, there was reduced mortality after exposure to OsHV-1 for oysters treated with the broad-spectrum antibiotic chloramphenicol, compared to untreated oysters. An 8 mg/L dose of chloramphenicol added to the oysters tanks every 2 days was sufficient to prevent cultivation of microbiota including Vibrio from the oyster tissues (Petton et al., 2015b). The complex aetiology of disease caused by OsHV-1 was recently investigated by de Lorgeril et al. (2018). A primary infection with OsHV-1 (µVar) in oyster haemocytes resulted in an immune-compromised state which was followed by a bacteraemia with opportunistic bacteria. This secondary bacterial involvement was necessary for complete disease expression demonstrated by histological studies which revealed bacterial accumulation and infiltrating haemocytes both inside and outside the gill tissues of oysters infected with OsHV-1 (de Lorgeril et al., 2018). Further support of the emerging view of a polymicrobial pathogenesis in Pacific oyster diseases was provided by Lasa et al. (2019) using a molecular approach to identifying microbial communities associated with C. gigas samples

collected during mortality episodes at different European sites. The study revealed signs of disruption to a balanced microbiota and the presence of previously undetected, potentially pathogenic bacteria, mostly belonging to the genera Vibrio and Arcobacter.

1.5 The Pacific oyster microbiome

1.5.1 Role of the microbiome

One of the earliest definitions for the term ‘microbiome’ was provided by Whipps and Cooke (1988). It was consistent with its current use in microbiology (Prescott, 2017). Here, the microbiome was defined as a characteristic microbial community occupying a reasonably well- defined habitat which has distinct physico-chemical properties. Based on fossil record, microorganisms formed spatially organized communities as early as 3.25 billion years ago (Allwood et al., 2006). Today, microbial life is found in diverse habitats all over the biosphere (Ley et al., 2008), including the ocean (Burgsdorf et al., 2014; Gonzalez-Acosta et al., 2006; Simister et al., 2012), human body (Costello et al., 2009; Singh and Manning, 2016; Yatsunenko et al., 2012) and in animals (Henderson et al., 2013; Niu et al., 2015). Key ecosystem processes such as nutrient cycling in this diverse range of habitats are driven by microbial communities (Shade et al., 2012). Analyses of 16S rRNA gene sequences from environmental DNA revealed the unprecedented diversity of microbial communities found in many habitats (Ley et al., 2006). In this study, the analysis of a hypersaline microbial mat present in an aquatic ecosystem revealed a microbial richness of 1336 unique 16S rRNA sequences out of 1586 that represented 752 species.

The microbial community that is present in most individuals of a particular animal or plant species, without causing harm to the host despite their continuous presence in different tissues, has long been considered the normal microbiota of that species (Berg, 1996; Romero et al., 2002). This microbial community contributes essential and diverse physiological functions to the host including digestion, nutrient absorption, maintaining the energy balance, immune response and disease resistance (Costello et al., 2009; Gopalakrishnan et al., 2018; Hoffmann et al., 2016; Littman and Pamer, 2011). Characterization of the host-associated microbiome is considered to be an important part of understanding the dynamics of overall host health (Pierce et al., 2016). For example, the colonization resistance (limiting the integration of exogenous bacteria into the

existing population) created by the intestinal microbiota, contributes to resistance against infection by intestinal pathogens (Littman and Pamer, 2011).

1.5.2 Microbiome analysis

Introduction of high-throughput sequencing platforms revolutionized the analysis of microbial communities by revealing a greater microbial diversity compared to culture-dependent analysis of microbes (Pollock et al., 2018; Streit and Schmitz, 2004). High-throughput sequencing combined with the development of advanced computational tools facilitated the analysis of a vast number of samples within a short time period (Caporaso et al., 2011). The approach of analyzing marker genes such as 16S rRNA and 18S rRNA gene is common. 16S rRNA gene analysis is the commonly used approach in bacterial diversity profiling (Clarridge, 2004; Pollock et al., 2018) (Pollock et al., 2018).

1.5.2.1 Bacterial 16S rRNA gene in microbial diversity profiling

Analysis of universal genes, especially the small-subunit ribosomal RNA (SSU rRNA), provide phylogenetic portraits of microbial communities, including micro-organisms that have not yet been cultivated (Frank et al., 2008; Lane et al., 1985; Pavlova et al., 2002). Of the universal genes used for phylogenetic analysis, the 16S rRNA gene, has proven to be the most useful for establishing distant relationships because of the universal distribution and high information content within a conserved area (Lane et al., 1985). Sequence analysis of the 16S rRNA gene is widely used in bacterial taxonomic studies and in the identification of bacterial species (Clarridge, 2004; Munson et al., 2004). The 16S rRNA gene sequence is about 1,550 bp long and consists of both variable and conserved regions (Clarridge, 2004). One of the most important features of the 16S rRNA gene that facilitate its use in bacterial diversity profiling is the interspersion of more or less conserved sequences together with sufficient interspecific polymorphisms within the rRNA genes, to provide distinguishing and statistically valid assignment to bacterial genera (Clarridge, 2004; Frank et al., 2008). However, it should also be noted that bacterial 16S rRNA gene diversity profiling is also not without limitations. For instance, 16S rRNA gene analysis is inadequate in identification of strains within a bacterial species (Clarridge, 2004). Moreover, better genes can be employed for differentiation of species within a particular genus (Clarridge, 2004).

The bacterial 16S rRNA gene contains nine hypervariable regions (V1-V9) that exhibit a considerable sequence diversity between bacterial species (Chakravorty et al., 2007; Van de Peer et al., 1996), facilitating the clustering of DNA sequences into operational taxonomic units (OTUs) (Blaxter et al., 2005). The choice of hypervariable region plays an important role in 16S rRNA gene-based microbial diversity profiling (Chakravorty et al., 2007; Tremblay et al., 2015; Yang et al., 2016). A study that analysed 110 bacterial species, demonstrated that V2 (nucleotides 137–242), V3 (nucleotides 433–497) and V6 (nucleotides 986–1043) regions contain the maximum nucleotide heterogeneity and the maximum discriminatory power (Chakravorty et al., 2007). The same study demonstrated that the V1 region best differentiated among Staphylococcus aureus and coagulase negative Staphylococcus species and V2 and V3 regions were most suitable for distinguishing all bacterial species to the genus level except for closely related Enterobacteriaceae. It was also shown that V4, V5, V7 and V8 were less useful targets for genus or species-specific probes. In another study, V4-V6 regions were identified as optimal regions with superior phylogenetic resolution for bacterial phyla (Yang et al., 2016). In addition to the choice of hypervariable region, amplification primers, sequencing primers and sequencing technologies may impact the results of bacterial diversity profiling (Tremblay et al., 2015). Tremblay et al. (2015) compared results obtained using different primer sets (targeting V4, V6–V8, and V7–V8 regions) and two sequencing technologies (454 pyrosequencing and Illumina MiSeq). The study revealed that the MiSeq reads were of higher quality and that primer choice influenced the abundance estimations.

1.5.3 Pacific oyster microbiome studies

Considering the impact of diseases on Pacific oysters, and the emerging understanding of the polymicrobial pathogenesis of these diseases, it is important to study factors that affect the microbiome.

1.5.3.1 Methods to characterize the oyster microbiome

Conventional bacteriological studies that involved bacterial culture in both basal media and selective media identified Pseudomonas and Vibrio spp. dominating the Pacific oyster microbiome (Colwell and Liston, 1960). These genera were followed in abundance by members of the family Achromobacteriaceae, which are presently classified as family Alcaligenaceae (De Ley et al., 1986). Using plate-culturing and broth-culturing, Colwell and Liston (1960) carried

out total viable counts and estimated most probable number (MPN) values to quantify bacteria in the oyster microbiome. They found that coliforms (Escherichia coli and related genera) never constituted more than 0.5 % of the microbiome. Another study conducted in Tokyo Bay, Japan found Vibrio, Aeromonas, Pseudomonas, Moraxella, Micrococcus and Enterobacteriaceae to dominate the Pacific oyster microbiome (Prieur et al., 1990; Sugita et al., 1981). The bacteria in the sediment and seawater at the study-site had a similar composition to the oyster microbiome, but the bacteria were present in lower number in the environment. On the other hand, Kueh and Chan (1985) identified Pseudomonas followed by Vibrio, Acinetobacter and Aeromonas where both the abundance and composition of the bacteria in the oyster microbiome were different from that of surrounding seawater. The differences between the oyster microbiome and the surrounding seawater and sediment, observed in these different studies can be attributed to the differences between the uptake of bacteria by filter-feeding which can vary with the abundance of bacteria, presence of bacteria that are bound to particles and season (Prieur et al., 1990).

Romero and Espejo (2001) provided a new insight to bivalve microbiome studies through microscopic observations of Chilean oyster (Tiostrea chilensis) tissue homogenates, after staining with diaminophenylindole/acridine orange (DAPI/AO). They showed that oyster tissues carry 105 times more bacteria than can be determined by agar plate counts, indicating that a great proportion of the oyster microbiome cannot be cultivated by standard procedures. The advent of molecular methods has subsequently enabled identification of bacterial diversity in the bivalve microbiome, including the Pacific oysters (Table 1.2). Using terminal restriction fragment length polymorphism (T-RFLP), Fernandez-Piquer et al. (2012) determined differences between the microbiome of live Pacific oysters and post-harvest tissues at different temperatures. This technique carries out microbial community profiling by digesting a mixture of PCR amplified variants of a single gene, using restriction enzymes and detecting the size of each of the terminal fragments using DNA sequencing. The study revealed that phylum Proteobacteria dominated the live oysters before storage while phylum Fusobacteria (Psychrilyobacter spp.) dominated the oyster microbiome after storage at 4°C. Meanwhile, Green and Barnes (2010) studied the digestive gland microbiome of Sydney rock oysters (Saccostrea glomerata) infected with the paramyxean parasite, Marteilia sydneyi, using restriction fragment length polymorphism (RFLP) followed by sequencing of 16S rRNA gene (27F-1492R). They found that the digestive glands of uninfected Sydney rock oysters were 18

dominated by phylum Firmicutes followed by phylum Proteobacteria. In contrast, the digestive gland microbiome of the QX-infected oysters was dominated by α-Proteobacteria.

1.5.3.2 16S rRNA gene diversity profiling of the Pacific oyster microbiome

During the past decade, a range of studies has been carried out on the microbiome of Pacific oysters and other marine bivalves, targeting different hypervariable regions of the 16S rRNA gene (Table 1.3). Using 16S rRNA gene sequencing coupled with RFLP, Romero et al. (2002) showed that Arcobacter spp. (phylum Proteobacteria) were abundantly present in the Chilean oyster (Tiostrea chilensis). This is one of the earliest studies to suggest Arcobacter as a member of the oyster microbiota. The study also showed that Arcobacter could not be grown in vitro justifying its absence in conventional bacteriological studies of the oyster microbiome. Meanwhile, King et al. (2012) used the V3-V4 hypervariable regions of the 16S rRNA gene to profile the microbiomes associated with the stomach and gut of the Eastern oyster (Crassostrea virginica) (Table 1.2). In this species, the gut microbiome was different and more diverse compared to that of the stomach in oysters from two different geographic locations. This study differentiated a putative core microbiome for both the stomach and gut, separately, in addition to the bacteria from the environment that were considered to be present as transient microorganisms.

Based on the findings of molecular studies, the Pacific oyster microbiome is dominated by Proteobacteria (Fernandez-Piquer et al., 2012; Hernández‐Zárate and Olmos‐Soto, 2006; Trabal et al., 2012). The results of recent 16S rRNA gene studies suggest that Pseudomonas and Vibrio are not abundant members of the oyster microbiome (King et al., 2018a; Lokmer and Wegner, 2015; Trabal Fernández et al., 2014). Instead, members of the phylum Bacteroidetes and phylum Firmicutes (Trabal Fernández et al., 2014) are abundant in the Pacific oyster microbiome.

1.5.3.3 Factors that influence the Pacific oyster microbiome

Despite the different methods that were employed to characterize the Pacific oyster microbiome, some common bacterial phyla including Proteobacteria, Bacteroidetes, Cyanobacteria, Spirochaetes, Fusobacteria, Firmicutes, Tenericutes, Planctomycetes and Verruomicrobia have been identified (Fernandez-Piquer et al., 2012; Hernandez-Zarate and

Olmos-Soto, 2006; Trabal Fernández et al., 2014; Trabal et al., 2012; Wegner et al., 2013). The microbiome composition can be influenced by oyster genetics (Wegner et al., 2013) and can change under various conditions that cause stress in oysters, including change in temperature (Lokmer and Wegner, 2015), translocation (Lokmer et al., 2016a) and pathogenic infection (Lasa et al., 2019).

The oyster microbiome demonstrated substantial changes in its composition during growth from post-larvae to adults (Trabal Fernández et al., 2014). One study which looked at the growth phases of C. gigas and C. corteziensis (mangrove oyster), identified Burkholderia cepacia (class β-Proteobacteria) as the most abundant bacteria in both species of oysters (Trabal et al., 2012). The study also revealed that B. cepacia colonized both species during early growth, showing a possible symbiotic host–bacteria relationship that was considered to protect oyster larvae against V. alginolyticus and V. harveyi (Trabal et al., 2012). This association was not altered by environmental conditions or the management of the oysters at the grow-out site (Trabal et al., 2012). Trabal Fernández et al. (2014) looked at the changes in the oyster microbiome of C. gigas, C. corteziensis and C. sikamea (Kumamoto oyster) during commercial production and found that Proteobacteria was the most abundant phylum in all three species of oysters studied. However, variations were observed at the genus level between post-larvae and adult oysters. Phylum Bacteroidetes was the second most common phylum but a higher abundance was seen in the post-larvae compared to adults. Post-larvae of each species had a higher bacterial diversity compared to adults of the same species (Trabal Fernández et al., 2014).

Dysbiosis, or the loss of bacterial diversity and proliferation of a few OTUs, is a phenomenon that has been observed in many microbiome studies of Pacific oysters that were subjected to stress and disease (Lasa et al., 2019; Lokmer and Wegner, 2015; Wegner et al., 2013). Dysbiosis can result in inflammatory diseases in humans and it can be caused by usually non-pathogenic commensal microbes (Littman and Pamer, 2011). Both internal (host) factors and external (environmental) factors can cause shifts in the microbiome, which might result in dysbiosis. A major shift in the composition of the commensal community can also result in an increase of invasive pathogenic microbes, as in infections following treatment with antibiotics (Littman and Pamer, 2011). Antibiotics impact the microbiome of an organism and will open up the niche for invasion by less compatible species of bacteria. This might stop the protective

function provided to the host by the healthy microbiome and remove the self-regulating balance of microbiota (Littman and Pamer, 2011). Antibiotic therapy provides an alternative approach to characterizing the impact of commensal bacteria (Littman and Pamer, 2011). As antibiotics differ in terms of antimicrobial spectrum, specific antibiotics may be used to target specific microbes in the microbiome (Sekirov et al., 2008).

Table 1.2 Molecular techniques other than high throughput sequencing of 16S rRNA gene regions, used in bivalve microbiome studies and their results Molecular method Species Tissue-types Bacteria/bacterial groups Reference used identified Fluorescent in situ hybridization Pacific oyster Digestive glands Class ɤ-Proteobacteria Hernandez-Zarate (FISH) (Crassostrea gigas) Gonads Gram-positive bacteria with and Olmos-Soto PCR amplification of 16S rDNA a low G+C content (2006) gene fragments with genus and group-specific oligonucleotides Gills Gram-positive bacteria with low and high G+C content Members of the Cytophaga/Flavobacterium cluster Class ɤ-Proteobacteria Class α-Proteobacteria Class β-Proteobacteria

Terminal restriction fragment Pacific oyster Whole soft- Proteobacteria length polymorphism (T-RFLP) (Crassostrea gigas) tissue Spirochaetes Fernandez-Piquer et 16S rRNA gene sequencing homogenates Planctomycetes al. (2012) (10F-907R) Verrucomicrobia Cyanobacteria Fusobacteria

Restriction fragment length Sydney rock oyster Firmicutes polymorphism (RFLP) (Saccostrea Digestive gland Proteobacteria Green and Barnes 16S rRNA gene sequencing glomerata) Cyanobacteria (2010) (27F-1492R) Spirochaetes

Table 1.3 Different hypervariable regions of the 16S rRNA gene targeted in bivalve microbiome studies

Bivalve species Tissue-type Hypervariable region/s Bacterial groups References described Chilean oyster Whole soft-tissue V1-V9 Data not available Romero and Espejo (Tiostrea chilensis) homogenates 16S-23S intergenic (a quantitative study) (2001) region

Chilean oyster Whole soft-tissue V1-V9 Arcobacter Romero et al. (2002) (Tiostrea chilensis) homogenates 16S-23S intergenic Staphylococcus region

Indo-Pacific oyster Gill V1-V9 γ-Proteobacteria: Zurel et al. (2011) (Chama pacifica, Order Oceanospirillales Chama savignyi)

Eastern oyster Stomach and gut V3-V4 Stomach content: King et al. (2012) (Crassostrea virginica) content Phylum Tenericutes: Mycoplasma Phylum Planctomycetes

Gut content: Phylum Proteobacteria: Class γ-Proteobacteria Shewanella Chloroflexi Phylum Tenericutes: Mycoplasma Phylum Firmicutes Phylum Verrucomicrobia

Bivalve species Tissue-type Hypervariable region/s Bacterial groups References described Pacific oyster Post-larvae: whole soft- V3-V5 Crassostrea corteziensis: Trabal et al. (2012) (Crassostrea gigas) tissue homogenates Class β-Proteobacteria: Mangrove oyster Juveniles and adults: Burkholderia (Crassostrea gut Phylum Spirochaetes: corteziensis) Borrelia Phylum Actinobacteria Phylum Propionibacterium

Crasssotrea gigas: Class α-Proteobacteria: Ruegeria Paracoccus Methylobacterium Class β-Proteobacteria: Burkholderia Class γ-Proteobacteria: Vibrio Phylum Firmicutes: Bacillus

Pacific oyster Gut V3-V5 Proteobacteria Trabal Fernández et al. (Crassostrea gigas) Bacteroidetes (2014) Mangrove oyster Actinobacteria (Crassostrea Firmicutes corteziensis) Crassostrea sikamea (Kumamoto oyster)

Bivalve species Tissue-type Hypervariable region/s Bacterial groups References described Pacific oyster Gill V3-V4 Phylum Proteobacteria: Wegner et al. (2013) (Crassostrea gigas) Sphingomonas Phylum Bacteroidetes: Class Flavobacteria In disturbed microbiome: Phylum Tenericutes: Mycoplasma Phylum Planctomycetes Phylum Actinobacteria

Pacific oyster Haemolymph V1-V2 Phylum Proteobacteria: Lokmer and Wegner (Crassostrea gigas) In moribund oysters: (2015) Arcobacter Infected oysters: Photobacterium Shewanella Phylum Bacteroidetes

Spiny oyster Gill V2-V3 Class γ-Proteobacteria: Roterman et al. (2015) (Spondylus spinosus, Oceanospirillales Spondylus avramsingeri Phylum Spirochaetes Spondylus pickeringae)

Pacific oyster Haemolymph V1-V2 *Haemolymph: Lokmer et al. (2016a) (Crassostrea gigas) Gut Class ε-Proteobacteria: Gill Arcobacter Mantle Class β-Proteobacteria Phylum Bacteroidetes: Class Flavobacteria Phylum Fusobacteria: Psychrilyobacter 25

Bivalve species Tissue-type Hypervariable region/s Bacterial groups References described Class γ-Proteobacteria: Oceanospirillaceae Vibrionaceae Phylum Spirochaetes: Brachyspiraceae Gut: Phylum Tenericutes: Mycoplasma

Pacific oyster Haemolymph V1-V2 **Laboratory samples: Lokmer et al. (2016b) (Crassostrea gigas) Phylum Fusobacteria Class ε-Proteobacteria: Arcobacter Class γ-Proteobacteria: Vibrionaceae Field samples: Class α-Proteobacteria Phylum Tenericutes

Eastern oyster Whole soft-tissue V1-V3 Class α-Proteobacteria: Ossai et al. (2017) (Crassostrea virginica) homogenates Pelagibacteraceae Class γ-Proteobacteria: Enterobacteriaceae Vibrio Phylum Cyanobacteria: Synechococcus

Pacific oyster Haemolymph V6 Haemolymph: Vezzulli et al. (2018) (Crassostrea gigas) Digestive gland Pseudoalteromonas Vibrio

Bivalve species Tissue-type Hypervariable region/s Bacterial groups References described Mediterranean mussel (Mytilus galloprovincialis)

Pacific oyster Adductor-muscle V1-V3 Phylum Spirochaetes: King et al. (2018a) (Crassostrea gigas) Brachyspiraceae Spirochaetaceae: Treponema Phylum Tenericutes: Mycoplasma Class α-Proteobacteria Class γ-Proteobacteria: Vibrio Class ε-Proteobacteria: Arcobacter Pseudoalteromonadaceae Phylum Bacteroidetes: Order Bacteroidales Paludibacter Bacteroides Phylum Firmicutes: Order Clostridiales

Pacific oyster Oysters spats V3-V4 Rhodobacteraceae Green et al. (2018) (Crassostrea gigas) Erythrobacteraceae Flavobacteriaceae Vibrionaceae: Vibrio Arcobacter Alteromonadaceae

Bivalve species Tissue-type Hypervariable region/s Bacterial groups References described Pacific oyster Oysters spats V3-V4 Vibrio de Lorgeril et al. (2018) (Crassostrea gigas) Arcobacter

Pacific oyster Whole soft-tissue V4 Class α-Proteobacteria Lasa et al. (2019) (Crassostrea gigas) homogenates of adult Class γ-Proteobacteria: oysters and spats Vibrio Haliea Marinicella Class ε-Proteobacteria: Arcobacter Phylum Tenericutes: Mycoplasma Phylum Bacteroidetes: Flavobacteria Winogradskyella Polaribacter Phylum Spirochaetes: Borrelia

Eastern oyster Gut V4 Proteobacteria Pierce and Ward (2019) (Crassostrea virginica) Tenericutes Blue mussel Verrucomicrobia (Mytilus edulis) Bacteroidetes Cyanobacteria Plantomycetes Actinobacteria Firmicutes Cladithrix Fusobacteria

Bivalve species Tissue-type Hypervariable region/s Bacterial groups References described Pacific oyster Whole soft-tissue V3-V4 Proteobacteria: Flores-Higuera et al. (Crassostrea gigas) homogenates of larvae Class α-Proteobacteria (2019) and Class γ-Proteobacteria spat Bacteroidetes Firmicutes Planctomycetes *The bacterial groups that were in higher abundance in the tissue compared to other tissues tested in the study, are listed. **The bacterial groups that were in higher abundance in the group compared to other groups are listed.

1.5.3.4 Impact of the environment on the Pacific oyster microbiome

Bacterial 16S rRNA gene diversity profiling was used to analyse the short-term responses of the oyster microbiome to disturbances caused by abiotic and biotic factors. The gill microbiome of Pacific oysters was impacted towards a reduced diversity and removal of rare OTUs reflecting bacteria at low-abundance after a heat shock treatment (Wegner et al., 2013). Initially, these authors observed an association between the oyster genotype and the microbiome composition which was reflected by the individual variations in the gill microbiome. It was suggested that microbial communities might assemble according to individual genotype of their hosts. The disturbance created by heat shock broke apart this association by removing rare OTUs (low-abundant OTUs) in the microbiome and thereby reducing overall bacterial diversity. The heat stress-related shifts in the microbiome composition did not increase the abundance of genera with potentially pathogenic strains. In another study temperature-acclimated oysters (one group at 8°C and the other at 22°C) were exposed to a temperature stress and experimentally challenged with a virulent strain of Vibrio (D29w) affiliated to Vibrio orientalis/tubiashii clade (Lokmer and Wegner, 2015). The haemolymph microbiome was affected by temperature and temperature stress, but not by the pathogenic infection challenge. Changes in abundance were seen at species level but not at higher taxonomic levels. In the moribund and dead oysters, the bacterial community structure was disrupted, resulting in very low bacterial diversity and proliferation of a few OTUs. Overall, these disturbance studies revealed that the oyster microbiome is shaped by both host factors (oyster genotype) and environmental factors.

The role of tissue-specific microbiota associated with the haemolymph, gill, mantle and gut has been studied in the establishment of oysters in new environments (Lokmer et al., 2016a). Pacific oysters were translocated to a new environment where the microbiota of one group of oysters were treated with antibiotics, to control the effects of the oyster microbiome on establishment. It was revealed that the environmental microbiota of the new environment not only interacted with the translocated oysters but also interact with their microbiome. Of the different types of tissues, the haemolymph microbiota demonstrated the highest connectivity to the bacteria present in the seawater

(Lokmer et al., 2016a). Moreover, the increase of the Vibrio fraction that was already present in the haemolymph microbiota resulted in systemic disease and high mortality, only when there was a spill-over of Vibrio to the solid tissues of the oyster (Lokmer et al., 2016a). The interaction of host and environmental factors on the oyster microbiome was further studied by translocating oysters from different genetic backgrounds (Lokmer et al., 2016b). Oysters from genetically different backgrounds were reciprocally transplanted in two different environments. It was found that a greater proportion of the microbiome variation occurred as a result of differences in the environment compared to the genetic background. However, the oyster genotype, physiology and health condition were also thought to play a role in determining the microbiome composition, owing to the individual differences between the oysters (Lokmer et al., 2016b; Wegner et al., 2013).

Ocean acidification from climate change is also responsible for impacting the physiology and disease susceptibility of marine organisms (Burge et al., 2014; Travers et al., 2009). Flores-Higuera et al. (2019) studied the impact of changes in the pH of seawater on the Pacific oyster microbiome. The microbiome of Pacific oyster larvae and spat maintained in a laboratory environment changed in response to reduced pH. In particular, the families Rhodobacteraceae and Campylobacteraceae were most affected with increases in family Vibrionaceae in larvae.

Artificial seawater environments such as that provided in the depuration of oysters also affect the Pacific oyster microbiome. The depuration process is usually carried out using clean seawater that was treated in a sequence of steps including sand filtration, UV treatment, ozonation (infusion of ozone into water), and biofiltration, to remove contaminants (Vezzulli et al., 2018). Although depuration is targeted at removing pathogenic microbes from oyster tissues, it has not always been effective in removing naturally occurring marine vibrios such as V. parahaemolyticus and V. vulnificus (Lee et al., 2008) that are pathogenic to humans and V. aestuarianus and V. splendidus clade that are pathogenic to oysters (Lacoste et al., 2001; Vezzulli et al., 2018). In contrary, Vezzulli et al. (2018) demonstrated an increase in the Vibrio fraction in the haemolymph of C. gigas, after depuration. Vezzulli et al. (2018) interpreted the persistence of Vibrio during depuration as a feature of the possibly long co-evolutionary history of Vibrio with their

invertebrate hosts, representing permanent residents of the microbiota which were able to increase in abundance under the changed environmental conditions. Similarly, the abundance of Vibrio in the tissues of C. gigas maintained in artificial seawater (Lokmer et al., 2016a) was thought to be a result of the static conditions in the tanks (Lokmer et al., 2016b). Strengthening the concept of a co-evolutionary role of Vibrio, Le Roux et al. (2016) also suggest that the pathogenesis of oyster disease involves a coevolutionary interplay between vibrios, oysters, and their microbiota. In this regard, a secondary opportunistic role for Vibrio spp. as pathogens was demonstrated in OsHV-1 infections (de Lorgeril et al., 2018b).

1.5.4 Role of the Pacific oyster microbiome in mortality events

In addition to the role of OsHV-1 and pathogenic Vibrio spp. in Pacific oyster mortality disease, recent molecular studies indicate a potential role of the oyster microbiome in disease pathogenesis (de Lorgeril et al., 2018; King et al., 2018a; Lasa et al., 2019). The oyster microbiome is the collection of all commensal and transient microbial species and strains that are inhabiting and are associated with different tissues of the oyster. However, the current understanding of the Pacific oyster microbiome is limited with respect to how it is affected by environmental factors, and by the host and how this can influence OsHV-1 infections. Thus, the role of host and environmental factors on the Pacific oyster microbiome and how it affects disease associated with OsHV-1 requires further investigation.

The emerging view of mass mortality diseases of Pacific oysters considers pathogenesis by a group of microbial species and strains that act as a consortium of pathogens (de Lorgeril et al., 2018; Lasa et al., 2019; Lemire et al., 2015). This gives rise to the concept of polymicrobial pathogenesis in bivalve infectious diseases. The term polymicrobial diseases was coined to describe diseases in animals and humans resulting from concurrent infection with multiple infectious agents (Brogden, 2002).

In a disease outbreak in Port Stephens, NSW, Australia, the oyster mortality was attributed to adverse environmental conditions such as elevated seawater temperature (>20°C) and low salinity (< 20 ppt) without the involvement of a specific infectious

pathogen (Go et al., 2017). However, in a subsequent analysis of the same disease outbreak, King et al. (2018a) identified a significant increase in rare OTUs belonging to Vibrio harveyi and an unidentified member of the Vibrio genus in the Pacific oyster tissues in this disease outbreak. These findings which are consistent with previous studies that have implicated vibrios for their role in oyster disease outbreaks, suggested an involvement of the microbiome in the disease pathogenesis (Garnier et al., 2007; Lemire et al., 2015). de Lorgeril et al. (2018) investigated the complex aetiology of Pacific oyster mortality disease using a holistic approach. They used an experimental cohabitation of specific pathogen free (SPF) oysters with oysters exposed to a natural mortality outbreak associated with OsHV-1 as donors of OsHV-1. In this study, primary infection with OsHV-1 µVar in oyster haemocytes resulted in an immune-compromised state which was followed by a bacteraemia with opportunistic bacteria already present in the oyster microbiome (de Lorgeril et al., 2018). The secondary bacterial involvement was shown to be necessary for complete disease expression. The immunosuppression in oysters was demonstrated by the significant reduction of expression of genes encoding antimicrobial peptides (AMPs), regularly over time. Histological studies revealed bacterial accumulation and infiltrating haemocytes both inside and outside the gill tissues of oysters infected with OsHV-1 (de Lorgeril et al., 2018). There were signs of bacterial community disruption and the presence of other potentially pathogenic Vibrio species in 525 C. gigas samples from recurrent Pacific oyster mortality outbreaks in Europe that were associated with infection with OsHV-1 or Vibrio aestuarianus (Lasa et al., 2019). The authors of the same study also used a new target enrichment next-generation sequencing protocol for capturing 884 phylogenetic and virulence markers, in order to identify the potential pathogenic microbial species from these genera. Opportunistically pathogenic Vibrio species and Arcobacter increased in abundance during the disease.

1.6 Aims of this study

To date, numerous and diverse research studies have been carried out to better understand the interactions between host, pathogen and environment that impact the incidence and severity of Pacific oyster diseases. This has contributed to controlling the spread of pathogens during mass mortality outbreaks and to identify disease control options to reduce the severity of disease. Nevertheless, diseases continue to pose a 33

significant threat to both farmed and natural stocks of Pacific oysters (Alfaro et al., 2019). There is an emerging view of a polymicrobial pathogenesis in which the oyster microbiome is thought to play a role (de Lorgeril et al., 2018; Lemire et al., 2015). Dysbiosis of the Pacific oyster microbiome has been identified and reported in both natural and experimental Pacific oyster mortality events associated with OsHV-1 and in infections caused by Vibrio aestuarianus (de Lorgeril et al., 2018; Lasa et al., 2019). It is important to determine if microbial dysbiosis is a sequel to the pathogenesis of the primary diseases or alternatively whether predisposing risk factors that directly lead to dysbiosis of the oyster microbiome contribute directly to the pathogenesis and severity of the disease.

Overall, this thesis is focused on identifying the degree to which changes in environmental factors can influence the Pacific oyster microbiome. The impact of these factors in changing the microbiome and the subsequent outcome of OsHV-1 infections in Pacific oysters is initially determined by: 1) evaluating methods to sample different tissues and extract nucleic acids from Pacific oysters to accurately determine the Pacific oyster microbiome and assessing the microbiome in different compartments associated with the oyster; 2) comparing the microbiome of genetically related Pacific oysters from a common hatchery origin after being grown in different estuaries; 3) assessing changes in the Pacific oyster microbiome during acclimation to a laboratory environment under different simulated tidal immersion conditions; 4) testing the impact of different immersion environments on the microbiome of oysters when challenged using a laboratory OsHV-1 infection model and 5) assessing the impact of seawater temperature on the oyster microbiome and how the changes can affect the outcome of an OsHV-1 infection. New knowledge about the influence of environmental factors on the oyster microbiome and how this can impact disease expression will inform advice for management of oyster farms to minimize the impact of disease. Further knowledge about the dynamics of the microbiome during disease will help direct future studies to determine the pathogenic mechanisms of polymicrobial diseases.

CHAPTER 2

General Materials and Methods

2.1 Laboratory facilities and oysters

All experimental methods described in this chapter were conducted at the Farm Animal Health Infectious Disease Laboratories of the Sydney School of Veterinary Science, The University of Sydney, Camden, NSW, Australia. This facility was certified as compliant with physical containment level 2 (PC2) according to the ‘Guidelines for certification of a physical containment level 2 laboratory (version 3.2 effective March 2013). Triploid Pacific oysters were obtained from commercial Pacific oyster farms at Hawkesbury River, Shoalhaven River, Clyde River and Georges River estuaries. The Pacific oyster is not considered by the NSW Animal Research Act 1985 nor the National Health and Medical Research Council (NHMRC) Australian code for the care and use of animals for scientific purposes, 8th edition (2013). Therefore, approval from the Animal Ethics Committee, University of Sydney was not required for the experimental trials described in this thesis.

Oyster tissues were either obtained from fresh or frozen whole Pacific oysters, for PCR quantification of bacteria and microbiome analysis. Fresh oysters were maintained at 4°C until they were brought into the laboratory and shucked, and tissue samples were collected immediately. At the laboratory, fresh oysters were maintained at 4°C until tissue dissection and the frozen oysters were kept at -80oC. While fresh oysters were kept at 4oC for a maximum of an overnight period (<15 h), frozen oysters were thawed at 4oC for no longer than this time before shucking and tissue dissection. For bacterial cultures, fresh oysters were maintained at 4°C for < 10 min (tissue-processing was done as soon as possible). In order to maintain the original and accurate microbiome composition, few oysters were removed from the experimental tanks at a time. The tissue samples were frozen immediately at -80°C until molecular studies were conducted. In Chapter 3, whole soft-tissue (except the gut tissue) was homogenized immediately and a coarsely clarified homogenate was stored at -80°C, until used. In Chapter 4, swab samples were collected in addition to tissue samples, and were also stored at -80°C.

2.1.1 Experimental procedures in the PC2 Aquatic Animal Facility

This facility contained 4 separate recirculation systems and each incorporated 6 tanks (each 20 L) and a sump (250 L). Each system was connected to a biofilter (Fluval 406 canister filter) and a chiller unit (HC-300a, Hailea Aquarium chiller). The tanks were used in this format for Chapters 5 and 6, when exchange of seawater was required to simulate natural environmental conditions. In Chapter 7, the recirculation systems were used as a water bath to maintain the water temperature of water-tubs (each 12 L) inserted into the 20 L tanks to increase the number of replicate tanks. For either set-up, each tank was individually aerated, and oysters were maintained on a perforated plastic shelf insert positioned approximately 10 cm above the tank floor. The biofilters were seeded with nitrifying bacteria 6 weeks prior to the experiment by including filter material harvested from established biofilters on a recirculating aquaculture system containing juvenile barramundi (Lates calcarifer) in artificial seawater at 30 ppt salinity. Each tank in each system was individually aerated.

The artificial seawater (Red Sea; 30 ppt ± 1 ppt salinity) in the systems were prepared with unfiltered water from the municipal supply and API® Tapwater Conditioner™). Water quality was tested daily using an API® Marine Saltwater Master Test kit. The quality of water was maintained at the target levels (pH range 8.0–8.2 and ammonia, nitrite and nitrate levels<0.25 ppm) by means of water exchange and/or the addition of the appropriate amount of sodium bicarbonate to the system.

During each live oyster experiment, the oysters were fed with a commercial microalgae diet (Shellfish Diet 1800, Reed Mariculture) containing a mix of Isochrysis spp., Pavlova spp., Thalassiosira weissflogii and Tetra selmis. The feeding rate per tank was calculated according to the recommendation of the manufacturer (https://reedmariculture.com/support_feeding_shellfish.php) based on the total wet weight the oysters in each tank.

2.2 Reagents

2.2.1 General Reagents

Milli-Q® water

Milli-Q® water was produced using a Milli-Q® Biocel Ultrapure Water Purification System (Millipore, Billerica, U.S.A.). The properties of the Milli-Q® water that was generated were as follows:

o Resistivity at 25°C: 18.2 Ω⋅m o Total organic carbon: 5 to 10 parts per billion o Pyrogens < 0.001 EU/mL o Bacteria < 1 colony forming unit (CFU)/mL

Saline (0.85% w/v)

NaCl 8.5 g

MilliQ® water 1 L

Dissolved NaCl (ACS grade; Astral Scientific, NSW, Australia) in a glass bottle (Schott) and autoclaved at 121oC for 15 min. Stored at room temperature.

Artificial Seawater (ASW)

Sea salt (RedSea®) 33.4 g Standard tap water (de-chlorinated) 1000 mL

Prepared with a salinity of 30-31 ppt. This solution was stored at room temperature, in an 1880 L reservoir, until use. The bulk preparation was used for housing oysters in experimental tanks. Sterile ASW was obtained by filtering the pre-prepared ASW solution through a 0.22 μm cellulose-acetate syringe filter (Minisart® NML, Sartorius, Göttingen, Germany) and UV-sterilization. Milli-Q® water was used to dissolve sea salt when ASW was prepared to be used as a reagent.

MgCl2 solution:

MgCl2.6 H20 (Sigma Aldrich) 250 g

ASW 1 L

Standard tap water (de-chlorinated) 4 L

A 50 g L-1 solution was prepared for immersion of oysters as a muscle relaxant.

2.2.2 Disinfectants

Sodium hypochlorite (NaOCl) solution (bleach solution):

NaOCl solution (125 g L-1) (Formula Chemicals, Australia) 40 mL standard tap water 960 mL A 500 mg L-1 solution was prepared and stored at room temperature for single day use.

Trigene™ Advance:

The disinfectant was diluted 1:100 for surface disinfection purposes that used Trigene™ Advance in procedures indicated in this thesis. It was diluted as follows:

Trigene™ Advance stock solution (MediChem International) 10 mL

Standard tap water 990 mL

The diluted solution was stored at room temperature for up to 6 months before being used for disinfection.

Virkon® S (1% solution):

Virkon® stock powder (Virkon® Antec International Ltd.) 10 g

Standard tap water 1000 mL

The powder was measured under a fume-hood to avoid any health hazards. The diluted solution was stored at room temperature for no more than 7 days before being used for disinfection.

2.3 Oysters

2.3.1 Sampling methods for oysters

2.3.1.1 Live oysters

A random sampling procedure was followed for live oysters on each occasion according to the following: each oyster in a tank was assigned a number based on its position in the tank. The number corresponded to a sequence from left to right, back to front (reading position). Random numbers were generated between 1 and the number of oysters in each tank using Microsoft Excel®, and the corresponding oysters were sampled.

2.3.1.2 Dead oysters

Mortality was assessed twice daily by visual inspection and dead oysters were sampled. Mortality was identified by non-responsive gaping on exposure to external stimuli including handling and air exposure for 3 min.

2.3.2 Dissection of oyster tissues

The live oyster experiments described in this thesis used oysters from a wide age range (5 months - 44 months) and shell length range (20 mm - 146 mm), depending on the objectives of the experiment. In instances where samples were obtained for bacteriology, the external shell surfaces of the oysters were thoroughly cleaned, first by scrubbing and rinsing with tap-water, followed by wiping with 70% ethanol to minimize contamination of internal tissues by surface bacteria. The oysters that were less than 4 cm in length were shucked using sterile, disposable scalpel-blades (Livingstone International). The blade was inserted into the hinge between the two valves of the oyster and was twisted carefully. The blade was then slid along the inner surface of the lower valve to transect the adductor muscle. The lower valve was removed to expose the soft tissues. In bacteriological sampling, the shucking procedure was carried as quickly as possible to minimize changes to the oyster microbiome after the time of sample collection. Shucked oysters and their soft-tissue samples were maintained on ice until the soft-tissue samples were transferred to – 80°C. A new blade was used for each oyster.

The oysters that were greater than 4 cm in length were opened using an oyster knife. In this method, the oyster was placed in a metal cradle with the hinge directed towards the operator and the lower valve uppermost (Fig. 2.1). The blade of the knife was inserted into the hinge of the oyster. A hammer was used if necessary, to move the blade 39

into the oyster and to open the valves. The blade was then slid forward to transect the adductor muscle at its insertion on lower valve which was excised from the rest of the oyster. The blade of the oyster knife was submerged in NaOCl solution (500 mg L-1) for 3 s followed by rinsing with hot water once and cold water twice, between each oyster. The gloves of the operator were also changed between different treatment groups or different sites from which oysters were obtained.

Figure 2.1 Technique for dissecting larger oysters (shell length > 4 cm) using an oyster knife. The oyster is placed in a metal cradle with the hinge directed towards the operator with the lower valve on top. The blade of the knife was inserted into the hinge of the oyster.

2.4 Sampling oyster tissues

2.4.1 Gill and mantle tissues for OsHV-1 DNA quantification

The whole oyster that remained in the upper shell after shucking was placed on a clean, disposable bench mat. A new sterile disposable scalpel blade was used for each oyster to excise equal portions of gill and mantle tissue (0.08-0.12 g total) which were

placed in nuclease-free 2 ml tubes (SSIBio, USA) containing 1 ml of distilled water (Ultrapure™) and 0.4 g of 0.1 mm zirconia-silica beads (Biospec Products, Daintree Scientific). The tubes were stored at -80°C until homogenization of gill and mantle tissues.

2.4.2 Sampling gill and gut tissues for total bacteria and Vibrio quantification

Gill and gut tissue sampling for bacteriology preceded gill and mantle tissue sampling for OsHV-1 DNA quantification, unless otherwise stated. The whole oyster soft tissue was collected onto a sterile Petri-dish lid when gill and gut tissues were sampled for bacteriology. Approximately 30 mg (wet weight) sample from the gill tissue was collected from the central region of the gill lamellae, using a sterile, disposable scalpel blade, avoiding the tissue at the extremities of the gill. The tissue sample was then collected aseptically into a labelled, sterile 1.5 ml tube and stored immediately at -80°C. The digestive gland (hereafter referred to as gut) of the same oyster was then incised along a sagittal plane into halves. A 30 mg portion of tissue was excised that included the luminal surface of the gut. The gut tissue sample was also collected aseptically into a separate, labelled, sterile tube and stored immediately at -80°C. Each tissue sample from the same oyster was dissected using separate sterile scalpel-blades to reduce cross contamination between tissues as well as between oysters.

2.4.3 Soft tissue homogenization for bacteriology

Homogenization of whole soft tissues by stomaching was carried out in experiments using conventional bacterial culture. The whole oyster soft tissue was collected onto a sterile Petri-dish lid and the gut was excised to reduce bacterial contamination of the other tissues from gut contents. The tissues were transferred into a stomacher bag (Interscience, France) and weighed. The bags were maintained on wet ice until stomaching. Oyster tissues were homogenized either individually or pooled (n=4) with either 4× or 10× (w/v) sterile ASW for different experiments. Stomaching was carried out in a bag mixer (MiniMix, Interscience, France) inside a Class 2 biosafety cabinet, for 1 min, at maximum speed (9). A coarsely clarified homogenate (5 ml) was obtained by collecting the material filtered through the inner mesh (porosity <250 µm) of

the stomaching bag (BagPage®, Interscience, France). These homogenates were used directly for bacterial culture and the remainder was stored at -80°C for molecular studies of bacterial content.

2.5 Conventional bacterial culture

2.5.1 Preparation of bacterial culture media

Marine Salt Agar – Blood (MSA-B)

Becton Dickinson (BD)® tryptone soy agar 20 g

NaCl 7.5 g

Milli-Q® water (deionized, ultrapure water) 500 ml

Sterile, anticoagulated sheep blood 15 ml

The agar-salt base was prepared and autoclaved (121°C for 15 min). After cooling to 50°C the sheep blood was added and mixed by gentle swirling. The MSA-B was poured into Petri-dishes, using a paddle-operated pump. The agar plates were allowed to cool to room temperature and stored at 4°C until use. Each plate of MSA-B carried approximately 0.6 ml of blood.

Thiosulfate-citrate-bile salts-sucrose agar (TCBS agar)

TCBS cholera medium (Oxoid, UK) 132g

Milli-Q® water 1.5L

The TCBS agar was prepared and autoclaved (121°C for 15 min) and the medium was poured into Petri-dishes, using a paddle-operated pump. The agar plates were allowed to cool to room temperature and stored at 4°C until use.

2.5.2 Isolation of bacteria from oyster tissue homogenates

Ten microliters each from each fresh tissue homogenate (described in Section 2.4.3) was spread on an MSA-B plate and a TCBS agar plate separately, using sterile

disposable spreaders (Copan, Brescia, Italy). Bacterial culture procedures were carried out inside a Class 2 biosafety cabinet with tissue- homogenates maintained on ice. The inoculated culture plates were incubated at 23°C for 24h (MSA-B) and 48hr (TCBS agar) in a refrigerated incubator (LMS Ltd, UK).

2.5.3 Identification and quantification of cultivable Vibrio and total bacteria

The colony morphology of bacterial colonies on both MSA-B plates and TCBS plates were studied separately. Records were made based on visual assessment of colonies considering the size, shape, colour, colony margins, elevation, moistness and haemolytic properties on MSA-B. The number of bacterial colonies on plates were counted manually was expressed as the number of colony forming units (CFU) per gram of oyster tissue (total cultivable bacterial count, TCBC; total cultivable Vibrio count, TCVC).

Single colonies of dominant morphotypes on TCBS agar were directly inoculated into nutrient broth with 2% NaCl and incubated at 23°C overnight. Broth cultures (0.85 ml) were mixed with glycerol (0.15 ml) and stored at -80°C. Species identification of select cryopreserved Vibrio cultures was performed at the Animal Health Laboratories, Department of Agriculture, Western Australia, using biochemical methods, for Chapter 3 of this thesis (Buller, 2014). However, it is important to note that Vibrio have been reported to demonstrate a greater phenotypic variability both within and between species (Fabbro et al., 2012).

Figure 2.2 A. Morphologically diverse bacterial colonies isolated on marine salt agar- blood (MSA-B) including β-haemolytic colonies (black arrows). B. Morphologically diverse bacterial colonies isolated on TCBS agar.

2.5.4 Preparation and quantification of genomic DNA standards

2.5.4.1 Vibrio genomic DNA

Quantitative standards and positive control samples were prepared from genomic DNA of Vibrio alginolyticus isolated from oysters used in an infection trial (Section 3.3.6.1). Prior to preparation of Vibrio standards for the qPCR assay, the cryopreserved V. alginolyticus cultures were tested for viability by sub-culturing on TCBS agar. The V. alginolyticus was then sub-cultured in nutrient broth with 2% NaCl and was incubated at 23°C for 24h. One ml of this broth culture was obtained after vortexing and was mixed with 9 ml of sterile, 0.85% saline and subsequently, this process was repeated to prepare a serial 10-fold dilutions (up to a maximum dilution of 10-6) of the V. alginolyticus broth culture. The spread plate method was used to estimate the number of CFU/ml of the original broth culture, by spreading the serial dilutions on TCBS agar plates (Buck and

Cleverdon, 1960; Herbert, 1990). Three replicate plates were prepared for each dilution by spreading 100 µl and the plates were incubated at 23°C for 24h. The dilutions that gave rise to a total number of Vibrio colonies ranging between 30-300 were selected and the average number of Vibrio colonies per dilution were calculated. The Vibrio gene copy numbers in the qPCR standards were then estimated by the calculated CFU/ml of the V. alginolyticus culture at stationary phase. The calculation of CFU/ml in the neat broth culture is as follows:

Based on the number of Vibrio colonies per plate, 10-4 dilution was selected. Average number of Vibrio colonies at this dilution = 232

Based on the assumption that one bacterial cell gives rise to one bacterial colony: The number of V. alginolyticus cells in 100 µl of the neat culture = 232 × 104 = 2.32 × 107 cells/ml

One ml aliquots were prepared from the neat broth culture of V. alginolyticus.

2.5.4.2 Total bacterial genomic DNA

The quantitative standards and positive control were prepared from nucleic acids purified from a lyophilized, non-pathogenic Escherichia coli strain 2038 culture available in the laboratory. Initially, the bacterial pellet was dissolved in 500 µl of Milli-Q® water and was swirled on nutrient agar. The plate was then incubated at 37°C overnight. The resulting bacterial culture was sub-cultured in nutrient broth, at 37°C for 24h. One ml of this broth culture was obtained after vortexing and was mixed with 9 ml of sterile 0.85% saline. This process was then proceeded to prepare a decimal serial dilution (upto a maximum dilution of 10-7) of the E. coli broth culture. Spread plate method was carried out to estimate the CFU/ml of the original broth culture, by spreading the serial dilutions on nutrient agar plates (Buck and Cleverdon, 1960; Herbert, 1990). Hundred µl each of the 7 ten-fold dilutions was spread on three replicate plates and the plates were incubated at 37°C for 24h. The dilutions that gave rise to a total number of E. coli colonies ranging between 30-300 were selected and the average number of E. coli colonies per dilution

were calculated. The E. coli gene copy numbers in the qPCR standards were then estimated by the calculated CFU/ml of the E. coli culture at stationary phase.

Based on the number of E. coli colonies per plate, 10-6 dilution was selected. Average number of E. coli colonies at this dilution = 237

Based on the assumption that one bacterial cell gives rise to one bacterial colony: The number of E. coli cells in 100 µl of the neat culture = 237 × 106 = 237 × 107 cells/ml

One ml aliquots were prepared from the neat broth culture of E. coli.

2.6 Tissue homogenization for DNA purification

2.6.1 Tissue homogenization for OsHV-1 qPCR assays

Gill and mantle tissues stored in bead-tubes (described in Section 2.4.1) were thawed and subjected to tissue lysis by bead-beating with a TissueLyser II (Qiagen). Previously prepared oyster tissue homogenates with known OsHV-1 infection status were included for each tissue-homogenization session: OsHV-1 positive control and two OsHV-1- free oyster tissue negative controls. These control homogenates had been created using the stomaching method described in section 2.4.3 using pooled whole oyster soft-tissues (n = 5) and stored in multiple single use aliquots. The negative control tubes contained nuclease free, distilled water (Ultrapure™) and were included at the start and end of each sample set (n = 48) in one run. The tissue-lysis by bead-beating was performed at a frequency of 30 Hz for 2 x 2 min cycles with 180° rotation of the tubes between cycles. The tubes were then centrifuged using a microcentrifuge (Heraeus® Biofuge® Pico, Thermo Electron Corporation, USA), at 900 g for 10 min and 200 μl of

the supernatant was transferred into 200 μl PCR tubes. These supernatants were stored at -20°C until nucleic acid extraction.

2.6.2 Tissue homogenization for bacterial DNA extraction

Two tissue homogenization methods were used for the experiments described in this thesis that extracted bacterial DNA from oyster tissues. In Chapter 3, a tissue lysis method was utilized to obtain bacterial DNA from coarse tissue-homogenates. After optimizing methods used to extract bacterial DNA from oyster tissues (Chapter 4), a different tissue lysis method was practiced which included subjecting oyster tissues to Proteinase K digestion.

A) Tissue lysis of coarse tissue-homogenates

Whole tissue homogenates obtained by stomaching method described in Section 2.4.3 were thawed. An aliquot of the tissue homogenate (150 µl) was mixed with 390 µl lysis/binding solution (MagMax™-96 Viral RNA Isolation Kit), in a 2 ml microcentrifuge tube containing 0.4 g of silica-zirconia beads. This mixture was homogenised by bead- beating using the TissueLyser II, as described in section 2.6.1. Each tissue-lysis session carried a positive control and two negative controls. The positive control was prepared from an Escherichia coli (strain 2038) broth culture, initially by centrifuging 1 ml aliquots at 1010 g (3700 rpm) for 15 min. Each bacterial pellet was mixed by vortexing with 150 µl of sterile 0.85% saline and 390 µl of lysis/binding solution (MagMax™ - 96 Viral RNA Isolation Kit, Ambion, Thermo Fisher Scientific). The negative control tubes contained nuclease free, distilled water (Ultrapure™) and were included at the start and end of each sample set in each session (n = 48). After bead-beating the tubes were centrifuged at 900 g for 10 min and 200 μl of the supernatant was transferred into 200 μl PCR tubes. These supernatants were stored at -20°C until nucleic acid extraction.

B) Tissue lysis method after Proteinase K digestion

A 30 mg portion of a gill or gut tissue sample collected as described in Section 2.4.2, was added to screwcap 2 ml tubes containing 350 μl ML1 buffer (E.Z.N.A.® Mollusc DNA kit, Omega Bio-Tek, USA) and 0.4 g of silica-zirconia beads. These tubes were then

subjected to bead-beating using the TissueLyser II at 30 Hz for 5 min. The tissue was then further digested by incubating overnight at 37°C with 25 µl of Proteinase K solution (Proteinase K 20 mg ml-1; E.Z.N.A.® Mollusc DNA kit). The samples were then transferred into a sterile 1.5 ml tube. The ML1 buffer contains the cationic detergent, cetyltrimethylammonium bromide (CTAB) which selectively binds the polysaccharides and proteins that remain after digestion with Proteinase K. These were then removed by extraction with chloroform:isoamylalcohol (24:1) (Winnepenninckx et al., 1993) followed by nucleic acid extraction and purification. Adding chloroform:isoamylalcohol to the digesta, separated the mucopolysaccharides and the DNA into two phases. While mucopolysaccharides were present in the chloroform phase, the DNA was in the aqueous phase which facilitated the removal of mucopolysaccharides after centrifugation.

2.7 Nucleic acid purification

2.7.1 Purification of nucleic acids for OsHV-1 assays

Nucleic acids were extracted from supernatants described in Section 2.6.1, using the Ambion MagMax™-96 Viral RNA Isolation Kit with a BioSprint-96 ™ magnetic particle processor (Qiagen) using the AM-1836 deep-well standard program (Thermo Fisher Scientific). According to the directions of the manufacturer, a 50 μL volume of supernatant was used to purify nucleic acids and this was eluted in 75 μl elution buffer. The purified nucleic acid extract was stored at -20°C until OsHV-1 DNA was quantified. An OsHV-1 negative tissue homogenate (negative control) and a control containing only the DNA extraction reagents, were used as the negative DNA extraction controls while a known OsHV-1 positive tissue homogenate was used as the positive DNA extraction control.

2.7.2 Bacterial DNA extraction

Two different commercial kits were used to extract bacterial DNA from oyster tissue homogenates. For Method A, the Ambion MagMax™-96 Viral RNA Isolation Kit was used with a BioSprint-96 ™ magnetic particle processor running the AM1836 deep- well standard program (Thermo Fisher Scientific). A 180 μL volume of supernatant was used for purification of nucleic acids with elution in 75 μl elution buffer. The purified nucleic acid extract was stored at -20°C until total bacteria and Vibrio were quantified.

For Method B, the E.Z.N.A.® Mollusc DNA kit was used to extract and purify nucleic acids using spin columns from a starting volume of 350 μL tissue homogenate supernatant. The nucleic acids were eluted in 100 μl elution buffer. The purified nucleic acid extract was stored at -20°C until total bacteria and Vibrio were quantified using real- time, quantitative PCR assays. The E.Z.N.A.® Mollusc DNA kit was selected for its use of CTAB to remove mucopolysaccharides present in mollusc tissues. The idea was to overcome PCR inhibition in downstream applications and to assess its impact on the microbiome composition. The high salt ML1 buffer of this kit contains the cationic detergent, cetyltrimethylammonium bromide (CTAB) which lyses host cells and bacterial cells and selectively binds polysaccharides and proteins (Wilson, 2001).

2.8 Real-time quantitative PCR for the detection of OsHV-1

In this thesis, a TaqMan® hydrolysis probe-based real-time qPCR assay was used to detect and quantify OsHV-1 DNA in oyster tissues based on a previously described assay (Martenot et al., 2010). The number of copies of the B-region of OsHV-1 genome was determined relative to a plasmid DNA standard according to a method described by (Paul-Pont et al., 2013b).

2.8.1 OsHV-1 quantitative plasmid standard

A plasmid carrying the B region of the OsHV-1 genome (pOsHV1-Breg), prepared by Alison Tweedie (University of Sydney) was used as the quantitative standard of this qPCR assay. The plasmid was prepared using a PCR amplicon generated from OsHV-1 B region and according to the method described by Evans (2016) using nucleic acids from juvenile C. gigas that was experimentally infected with OsHV-1 . The target region of the OsHV-1 genome was amplified using Expand High Fidelity PCR system (Roche). This contained a proof-reading DNA polymerase mix suitable for TA cloning, a subcloning technique that does not use restriction enzymes. A 10-fold dilution series of pOsHV1-Breg in water containing between 107-100 copies/5 μL (assay) was prepared prior to each qPCR run by diluting the in nuclease-free water and was used in duplicate to generate the standard curves of the qPCR assay.

2.8.2 OsHV-1 qPCR assay

Samples were tested in duplicate 25 μL PCR reactions prepared with Path-ID qPCR master mix (Life Technologies) using a Mx3000P Multiplex Quantitative PCR System (Stratagene, Agilent Technologies). Each reaction contained: 12.5 μL of 2x real- time qPCR buffer, 5.425 μL of nuclease free water, 900nM each of forward (OsHV1BF; Table 1) and reverse primer (OsHV1B4), 250nM OsHV1ProbeB, 1 μL of 25x real-time qPCR enzyme mix, and 5 μL of nucleic acid extract described in Section 2.7.1 (template DNA). Each PCR plate of 96 reactions contained a 10-fold dilution series of pOsHV1- Breg (106-101 copies/μL), a purified nucleic acid extract from an oyster confirmed to be infected with OsHV-1 (PCR positive control), a PCR reaction with only the PCR reaction mixture without any template DNA (PCR negative control) and the DNA extraction controls. The thermocycling conditions for the qPCR assay are detailed in Table 2.2.

2.8.3 PCR controls

Nucleic acids purified from oysters confirmed to be infected with OsHV-1 were used as the positive control. The nucleic acid preparation contained an adequate amount of OsHV-1 genomic DNA to produce a Ct value of approximately 24, when used undiluted (Evans, 2016). Ct value is the cycle number when the fluorescence of the PCR product can be detected above the background signal (cycle threshold). A stock solution prepared by diluting these nucleic acid extracts in TBE buffer was available in the laboratory and was used in the OsHV-1 quantification assays. The diluted solution had a Ct value in the range of 30-34 and was stored a 4oC. The PCR negative control carried only the PCR mixture described in section 2.8.2, with water used in place of template DNA.

Table 2.1 Primers and probes used in this thesis

Code of Sequence (5’→3’) Reference primer/probe OsHV1BF 5’- GTCGCATCTTTGGATTTAACAA-3’ Martenot et al. (2010) OsHV1B4 5’-ACTGGGATCCGACTGACA AC-3’ Martenot et al. (2010) OsHV1ProbeB 5’-6FAM- Martenot et al. (2010) TGCCCCTGTCATCTTGAGGTATAGACAATC- TAMRA-3’

Vib1 5’-GGCGTAAAGCGCATGCAGGT-3’ Vezzulli et al. (2012) Vib2 5’-GAAATT CTACCCCCCTCTACAG-3’ Vezzulli et al. (2012)

Nadak_rRNAF 5’-TCCTACGGGAGGCAGCAGT-3’ Nadkarni et al. (2002) Nadak_rRNAR 5’-GGACTACCAGGGTATCTAATCCTGTT-3’ Nadkarni et al. (2002) Nadak_rRNA 5’-6FAM-CGTATTACCGCGGCTGCTGGCAC- Nadkarni et al. (2002) probe BHQ1-3’

27F 5’-AGAGTTTGATCMTGGCTCAG-3’ Lane (1991) 519R 5’-GWATTACCGCGGCKGCTG-3’ Lane (1991) 1492R 5’-TACCTTGTTACGACTT-3’ Weisburg et al. (1991)

Table 2.2 Thermocycling conditions for the OsHV-1 qPCR assay Phase Temperature (°C) Time Cycles Hot start activation 95 10 min 1 Denaturation 95 15 s 40 Annealing and 60 45 s 40 extension

2.8.4 Quality control criteria

A PCR run was considered valid when there was no amplification of negative controls; amplification of both replicates of the positive control with a Ct value within the range of the standard curve; and standard curve with r2 > 0.95 and efficiency between 90 and 110%. The standard curve was obtained by plotting the log value of the template copy number of the plasmid standard against the Ct generated for each dilution. The fluorescence threshold for each PCR run was determined using the amplification-based threshold algorithm from Stratagene, for the standard curves. Samples exhibiting an exponential increase in the ROX normalised, baseline corrected FAM fluorescence signal in both replicates with a valid cycle threshold was considered positive. Samples that had a valid cycle threshold in one replicate were considered inconclusive and were re-tested. The quantification limit of the assay was 12 OsHV-1 genome copies per PCR reaction (Evans et al., 2017; Evans, 2016). Positive PCR results below the quantification limit were designated as below the limit of quantification (BLOQ).

Demonstration of OsHV-1 DNA calculation in oyster tissues used in this thesis (using an example):

Quantity of tissue = A g Volume of homogenising media and tissue = B mL Tissue concentration = A g/B mL = C g/mL

Volume used for DNA extraction = 50 μL Quantity used for DNA extraction = C × 50 = D mg

Volume of eluate = 75 μL Concentration of eluate = D mg/75 μL

= E mg/μL

Volume of eluate per PCR reaction = 5 μL Quantity of tissue represented per PCR reaction = E x 5 μL = F mg

OsHV-1 gene copy number in the PCR reaction = Average quantity of copies for each replicate well (G)

DNA copies per mg tissue = G/F = H OsHV-1 genomic DNA copies/mg

2.9 Real-time quantitative PCR for the detection of Vibrio

The qPCR assay described by Vezzulli et al. (2012) for quantification of Vibrio, was adapted for quantification of Vibrio spp. DNA in oyster tissues. A Fast SYBR® Green real-time qPCR assay was used. The primers Vib1 and Vib2 (Table 2.1) have an amplification target between position 567 and 680 (based on the Escherichia coli numbering system) of the 16S rRNA gene which is specific for the genus Vibrio (Thompson et al., 2004b; Vezzulli et al., 2012) .

2.9.1 Preparation of Vibrio genomic DNA standards

The aliquots of Vibrio broth cultures described Section 2.5.4.1 were centrifuged at 1010 g (3700 rpm) for 15 min, using a microcentrifuge (Heraeus® Biofuge® Pico). Each bacterial pellet was dissolved in sterile saline, after discarding the supernatants. In Chapter 3, Vibrio DNA was extracted using the method described in section 2.6.2 A. After optimization of methods the Vibrio DNA used for qPCR standards were extracted using MagMAXTM CORE Nucleic Acid Purification Kit (ThermoFisher Scientific, Applied Biosystems, CA) and MagMAXTM Express 96 magnetic particle processor (Applied Biosystems, Foster City, CA) with the MagMAX_Core_50μl program,

according to manufacturer guidelines. In either method, the neat DNA extracts were stored at -80°C. A 10-fold dilution series (106-101 Vibrio gene copies/μL) was prepared prior to each qPCR run by diluting the neat DNA extract in nuclease-free water and was used to generate the standard curves of the qPCR assay.

2.9.2 Fast SYBR® Green real-time qPCR assay

Each reaction contained 10 µl of Fast SYBR® Green Master Mix (Applied Biosystems), 0.2 µM of each primer (Vib1 and Vib2; Table 1), 5 µl of neat template DNA (as described in Section 2.7.1) and sterile, nuclease free water, using a 7500 Fast RT-PCR system (Applied Biosystems, Foster City, CA). The thermocycling conditions for the qPCR assay are detailed in Table 2.3.

Table 2.3 Thermocycling conditions for the Vibrio qPCR assay Phase Temperature (°C) Time (s) Cycles Initial denaturation 95 20 1 Denaturation 95 3 40 Annealing and extension 58 30

Each PCR plate included a positive control sample, a negative control sample and duplicate reactions prepared with a 10-fold serial dilution starting from 2.2 × 104 copies of Vibrio 16S rRNA gene as a quantitative standard. Samples were tested in duplicate 20 μL PCR reactions. Melt curves of the final PCR products were generated and analysed from 60°C to 95°C at 1°C intervals. Primer Melting Temperature (Tm) is the temperature at which half of the DNA duplex will dissociate to become single stranded.

2.9.3 PCR controls

Positive control reactions were prepared with genomic DNA of Vibrio alginolyticus isolated from oysters used in Chapter 3 of this thesis. The negative control was prepared with water in place of DNA template in the PCR. Quantification of Vibrio spp. DNA in samples was determined by standard curve quantitation method, using 7500

software v2.3 (Applied Biosystems) for a 10-fold serial dilution (101 - 106 copies) of Vibrio rRNA gene.

2.9.4 Quality control criteria

A PCR run was considered valid when there was no amplification of negative controls, amplification of both replicates of the positive control with a cycle threshold (Ct) within the range of the standard curve, and a standard curve with r2 > 0.99 and efficiency between 90 and 110%. Samples exhibiting an exponential increase in SYBR

fluorescence signal in both replicates with a Ct value >15 and <35 and a dissociation

curve (Tm) ranging from 77.7 to 79.1°C) that conformed to that of the positive control, were considered for quantification of total Vibrio DNA. The total Vibrio count per PCR sample was calculated by dividing the total Vibrio 16S rDNA copy number by the presumptive number of 16S rDNA copies per Vibrio genome (n = 9) (Klappenbach et al., 2001; Thompson et al., 2004b; Vezzulli et al., 2012). The total Vibrio count per oyster tissue homogenate was expressed as number of Vibrio genome equivalents per mg of oyster tissue while that in gill and gut tissues were expressed as number of Vibrio genome equivalents per g of gill or gut tissue.

2.10 Real-time quantitative PCR for the detection of total bacteria

The qPCR assay based on TaqMan® chemistry described by Nadkarni et al. (2002) was adapted to quantify total bacteria in oyster tissues (Pathirana et al., 2019a). The primers Nadak_rRNAF and Nadak_rRNAR and probe, Nadak_rRNA (Table 2.1) target positions 331 to 797 of 16S rRNA gene (Escherichia coli numbering system) which is highly conserved across domain Bacteria.

2.10.1 Preparation of total bacteria genomic standards

The aliquots of the E. coli broth culture described in Section 2.5.4.2 were centrifuged at 1010 g (3700 rpm) for 15 min, using a microcentrifuge (Heraeus® Biofuge® Pico). Each bacterial pellet was dissolved in 1 ml sterile saline, after discarding the supernatants. In Chapter 3, E. coli DNA was extracted using the method described in Section 2.6.2 A. After optimization of methods the E. coli DNA used for qPCR standards were extracted using MagMAXTM CORE Nucleic Acid Purification Kit, as described in

Section 2.9.1. For both methods, the neat DNA extracts were stored at -80°C. A 10-fold dilution series (108-101 E. coli gene copies/μL) was prepared prior to each qPCR run by diluting the neat DNA extract in nuclease-free water and was used to generate the standard curves of the qPCR assay.

2.10.2 TaqMan® real-time qPCR assay

Samples were tested in duplicate 25 µl PCR reactions prepared in TaqMan™ Fast Universal PCR Master Mix (2x) (Thermofisher Scientific), using a 7500 Fast Real-time PCR system (Applied Biosystems). Each reaction contained 12.5 µl of 2× Master Mix, 0.9 µM of each primer (Nadak_rRNAF and Nadak_rRNAR; Table 1) and 0.1 µM probe (Nadak_rRNA probe), 5 µl of neat template DNA and sterile, nuclease free water. The thermocycling conditions for the qPCR assay are detailed in Table 2.4.

Table 2.4 Thermocycling conditions for the total bacteria qPCR assay

Phase Temperature (°C) Time (s) Cycles Initial denaturation 95 10 min 1 Denaturation 95 15 s 40 Annealing and extension 60 1 min 40

2.10.3 PCR controls

Positive control samples were prepared from genomic DNA of E. coli used for the PCR standards. The negative control contained the PCR mixture with water used in place of template DNA. Quantification of total bacterial DNA in samples was determined by standard curve quantitation method, using ABI 7500 software v2.3 (Applied Biosystems) for a 10-fold serial dilution (101 – 108 copies) of E. coli rRNA gene.

2.10.4 Quality control criteria

PCR runs were analyzed by the standard curve quantitation method (Applied Biosytems) based on amplification of a 10-fold serial dilution containing between 101 - 108 copies/μL of E. coli rRNA gene. A PCR run was considered valid when there was no amplification of negative controls; amplification of both replicates of the positive control

with a cycle threshold (Ct) within the range of the standard curve; and standard curve with r2 > 0.99 and efficiency between 90 and 110%. Samples exhibiting an exponential

increase in the fluorescence signal in both replicates with a Ct value >15 and <35 were considered for quantification of total bacteria DNA. The total bacteria count per PCR sample was calculated by dividing 16S rDNA gene copies by the presumptive number of 16S rDNA copies per Proteobacteria genome [n = 3.5, Kormas (2011); Vezzulli et al. (2012)], based on the genuses of bacteria that dominate the oyster microbiota (Lokmer et al., 2016a; Prieur et al., 1990; Wegner et al., 2013). Total bacteria quantity per oyster tissue homogenate was expressed as the number of bacteria genome equivalents per mg oyster tissue while that in gill and gut tissues were expressed as number of bacteria genome equivalents per g of gill or gut tissue.

2.11 Oyster microbiome analysis by high throughput 16S rRNA gene sequencing

High-throughput sequencing (MiSeq, Illumina) of the hypervariable region V1-V3 of the 16S rRNA gene was used in this thesis to identify the microbial community composition of Pacific oysters. The Illumina Miseq Next Generation Sequencer produced paired sequence reads up to 300 bp long (http://www.illumina.com). For experiments detailed in Chapters 3 and 4, DNA sequencing was performed through the Australian Genome Research Facility (AGRF), Queensland while for that of Chapters 5-7 the service was provided by the Ramaciotti Centre for Genomics, University of New South Wales, Australia. At the sequencing facility, PCR amplicons were generated using primers 27F and 519 R (Table 2.1), using AmpliTaq Gold 360 Master mix (Life Technologies, Australia). Indexing the amplicons was performed with TaKaRa Taq DNA Polymerase (Clontech) and the resulting amplicons were measured by fluorometry (Invitrogen™ Picogreen) and normalised to provide an equimolar pool of amplicons. These were quantified by qPCR and sequenced with Illumina MiSeq 300bp paired end chemistry. Two bioinformatic pipelines were used microbiome analysis in the analysis of these high- throughput DNA sequence data. While the Chapters 3 and 4 of this thesis used the version 1 of QIIME™ (Quantitative Insights In to Microbial Ecology), the Chapters 5-7 of this thesis used QIIME2, after its introduction in 2018 (http://www.qiime.org).

2.11.1 Bioinformatics

Quantitative Insights Into Microbial Ecology (QIIME) is an open-source bioinformatics pipeline for microbiome analysis using raw DNA sequencing data (http://www.qiime.org). The analysis involves demultiplexing of raw sequence data and quality filtering, OTU picking (selecting OTUs), taxonomic assignment, phylogenetic reconstruction, diversity analyses and visualizations (Caporaso et al., 2010). Demultiplexing is the use of barcode information to identify and track the origin of sequences to their samples after they had all be sequenced together. Quality filtering involves reducing noise, removing replication and chimera-filtering the sequence reads (Callahan et al., 2016). QIIME2 is an improved version of the QIIME1 software package which includes automatic tracking of data provenance, plugins for extended microbiome analysis while supporting multiple types of user interfaces (both command line and graphical user interface) (Bolyen et al., 2019). As of now, plugins are available to compositional data analysis, demultiplexing, sequence quality control with DADA2, sequence quality control with Deblur etc. The plugins that were used in the microbiome analysis of this thesis are described in the next sections.

2.11.2 Amplification of V1-V3 hypervariable regions of 16S rRNA gene

The amplification of V1-V3 hypervariable regions of 16S rRNA gene was carried out as described in Section 2.11.

2.11.3 Quality control of sequence reads

The DNA sequence reads were provided in FASTQ format. Unlike the FASTA format, the FASTQ format files provided a quality score for the sequence reads in addition to the read itself. The quality scores provide a measure about the probability that a base is called incorrectly during sequencing (Illumina, 2020). Bases with lower quality scores may suggest considerable sections of the sequencing reads to be removed before the proper bioinformatic analysis is carried out.

In this thesis, the quality of the raw sequence data was assessed using the software programme FASTQC (Andrews, 2010).

2.11.4 Generation and validation of metadata

Metadata includes descriptions of the samples that are being used in the microbiome analysis. In this thesis it included individual oyster identification number, batch of oyster, age, the days post-OsHV-1 injection at the time of sample collection, immersion regime and tissue-type. Metadata were tabulated and validated at the beginning of microbiome analysis. In this thesis, metadata files were generated in Microsoft® Excel software and was validated using the browser-based metadata validation tool, Keemei, to check for correct formatting and errors in the metadata file (Rideout et al., 2016).

2.11.5 Importing paired-end DNA sequence reads into the QIIME2 pipeline

The demultiplexed, paired-end DNA sequence reads that were obtained from the sequencing facility was initially imported into the QIIME2 pipeline. Various data formats are available for the raw sequencing reads depending upon the type of sequencer used. In this thesis, the sequence reads were obtained in the Casava 1.8 paired-end demultiplexed fastq format (https://docs.qiime2.org/2019.10/tutorials/importing/) and thus the following script was used to import the sequence reads into the QIIME2 platform. qiime tools import \

--type 'SampleData[PairedEndSequencesWithQuality]' \

--input-path casava-18-paired-end-demultiplexed \

--input-format CasavaOneEightSingleLanePerSampleDirFmt \

--output-path demux-paired-end.qza

2.11.6 Sequence quality control (denoising)

In this thesis, the Divisive Amplicon Denoising Algorithm 2 (DADA2) was used to reduce noise, remove replication and chimera-filter the reads (Callahan et al., 2016). Depending upon the base quality that was determined using FastQC at the beginning of

the DNA sequence analysis, the sequence reads were truncated (eg. at 120 bp) to improve the quality of sequence data.

2.11.7 Visualizing the feature table and feature data summaries

A feature table was created which contained the number of reads of each unique sequence in each sample in the dataset. Each unique sequence was identified by a feature identifier and the feature table mapped feature identifiers to the sequences they represent (McDonald et al., 2012b). The feature table was the equivalent to OTU table (BIOM table) of QIIME1 pipeline. The feature-table summarize command was executed to obtain information about the number of sequences associated with each sample in the microbiome analysis process and related summary statistics. The feature-table tabulate- seqs command provided a mapping of feature IDs to sequences.

2.11.8 Alpha and beta diversity analysis

The microbial diversity (alpha and beta diversity) analyses were performed using the q2-diversity plugin of QIIME2 (Bokulich et al., 2018). This computed alpha and beta diversity metrics and generated interactive visualizations with statistical analysis. Using the core-metrics-phylogenetic method, samples were rarefied to a required read depth prior to computing alpha and beta diversity metrics and generating principle coordinates analysis (PCoA) plots using Emperor (Vázquez-Baeza et al., 2013). In this thesis, the number of observed OTUs and Shannon’s diversity index were used as the parameters to assess the alpha diversity of samples. Beta diversity was visualized by generating principal coordinate plots (PCoA) based on the two-dimensional Bray-Curtis (BC) dissimilarity index. The beta-group-significance command used one-way permutational multivariate analysis of variance (PERMANOVA) for the statistical analysis of beta diversity. Significance was set at p < 0.05 for statistical analyses.

2.11.9 Taxonomic analysis

The taxonomy was assigned using a naïve-bayes classifier trained using Greengenes v.13_8, clustered at 99% sequence similarity, where the sequences were trimmed to only include the V1-V3 hypervariable region of 16S rRNA gene (McDonald et al., 2012a). The relative abundance of bacteria (phyla) in each sample was initially

visualized using interactive taxonomic bar plots. Graphical representation of relative abundance was further performed by 100% stacked 2-D column graphs in Microsoft® Excel.

2.12 Sequence data management

All high-throughput sequence data generated in this thesis were stored in the Research Data Storage (RDS) facility of the Sydney Informatics Hub of the University of Sydney. A backup of the data was also maintained in a cloud storage.

CHAPTER 3

Influence of the environment on the pathogenesis of Ostreid herpesvirus-1 (OsHV-1) infections in Pacific oysters (Crassostrea gigas) through differential microbiome responses

3.1 Abstract

The oyster microbiome is thought to contribute to the pathogenesis of mass mortality disease in Pacific oysters, associated with OsHV-1. As filter-feeders, oysters host a microbiota that can be influenced by the estuarine environment. This may alter susceptibility to OsHV-1 infections, causing variable mortality. This study aimed at: (1) differences in the microbiome of Pacific oysters with a common origin but grown in geographically distinct estuaries; (2) evaluating changes occurring in the microbiota, especially in Vibrio, and (3) differential responses of the oyster microbiome, in response to an OsHV-1 infection. Pacific oysters sourced from a single hatchery but raised separately in Patonga Creek, Shoalhaven River and Clyde River of NSW, Australia, were used and challenged with OsHV-1. The initial microbiome composition was different in the three batches and changed further, post-injection (p<0.05). The Patonga oysters with the highest mortality also had higher OsHV-1 and Vibrio quantities compared to the other two batches (p<0.05). The higher initial bacterial diversity in Patonga oysters decreased in moribund oysters which was not observed in the other two batches (p<0.05). The microbiome of survivors of OsHV-1 infection and negative control oysters of two batches, did not show any changes with the relevant pre-challenged microbiome. A strong correlation was observed between the OsHV-1 and Vibrio quantities in OsHV-1 infected oysters (r = 0.6; p < 0.001).

NOTE: This chapter is a previously published paper reformatted to the specifications of this thesis (refer to page vii for reference). I am the first author of this publication. The text contained within this chapter is identical to the aforementioned publication except for an additional table in the paper which is included in the Chapter 2 of this thesis (Table 2.1). Survival analyses and hazard ratios presented in this Chapter were calculated by Dr. Marine Fuhrmann.

In conclusion, the Pacific oyster microbiome differed in different batches despite a common hatchery origin. Different microbiomes responded differently with a differential outcome of OsHV-1 challenge. The higher Vibrio load in oysters with higher OsHV-1 content and higher mortality, suggests a role in Vibrio in the pathogenesis of this mortality disease. This study provided insights of the potential of different estuarine environments to shape the Pacific oyster microbiome and how different microbiomes are associated with different outcomes of OsHV-1 infection.

Keywords: Pacific oyster, Crassostrea gigas, microbiome, environment, Ostreid herpesvirus-1

3.2 Introduction

Pacific oysters (Crassostrea gigas) make the greatest contribution to global oyster production with 625,925 tonnes per annum worth US$ 1.3 billion out of a total oyster production of 5.2 million tonnes worth US$ 4.2 billion (FAO, 2014). In Australia, edible oyster farming contributes approximately 38% to the total marine mollusc production, with an estimated 11.3 thousand tonnes produced in 2015-2016 (Mobsby and Koudah, 2017). However, outbreaks of severe and widespread oyster mortality have greatly impacted Pacific oyster farming in Australia (de Kantzow et al., 2017; Jenkins et al., 2013; Paul-Pont et al., 2014), New Zealand (Keeling et al., 2014) and in Europe (Clegg et al., 2014; Garcia et al., 2011; Peeler et al., 2012; Roque et al., 2012; Segarra et al., 2010).

Outbreaks of summer mortalities in Pacific oysters were first reported in Japan in 1950 (Takeuchi et al., 1960) and since then this syndrome affecting adult oysters was reported in many parts of the world, including Italy (Takeuchi et al., 1960), USA (Glude, 1974) and France (Maurer et al., 1986). In France, mass mortality events have also been reported in C. gigas spats since 1993 (Renault et al., 1994a), which were later identified to be associated with Ostreid herpesvirus-1 (OsHV-1) (Davison et al., 2005; Dégremont et al., 2015). The reference genotype of OsHV-1 (Davison et al., 2005; Le Deuff and Renault, 1999) and related genotypes have been identified as the prominent pathogens which caused C. gigas mortalities in France from 1991 to 2008 (Martenot et al., 2011; Renault et al., 2012). However, at the end of spring 2008, widespread mortalities were

reported in France which killed billions of young oysters and a genomic variant of OsHV- 1, called µVar, was identified from these outbreaks (Renault et al., 2012; Segarra et al., 2010). Since then, severe disease of Pacific oysters associated with the variant and closely related microvariant genotypes were associated with disease events observed in different parts of the world, including Australia (Jenkins et al., 2013; Paul-Pont et al., 2015; Whittington et al., 2015b), New Zealand (Renault et al., 2012), Ireland (Clegg et al., 2014; Peeler et al., 2012), Spain (Roque et al., 2012) and Scandinavia (Mortensen et al., 2016).

Despite the different genotypes of OsHV-1 and their causal relationship with oyster mortality, Pacific oyster mortality events are most often considered a result of complex interactions between the physiological status of the oysters, the environment and multiple pathogens (Alfaro et al., 2019; de Lorgeril et al., 2018; Pernet et al., 2018; Samain et al., 2007). Considering the age of oysters, spat and juvenile oysters are generally more susceptible to OsHV-1 infection and related mortality events (Renault et al., 1994b; Schikorski et al., 2011), while processes such as spawning which causes physiological stress also predispose oysters to mortality (Garcia et al., 2011; Samain et al., 2007). Furthermore, research has revealed a role of oyster genotype in susceptibility to the viral infections and mortality events (Azéma et al., 2017b; de Lorgeril et al., 2018; Dégremont, 2011). Meanwhile, environmental factors such as elevated seawater temperature (Petton et al., 2013; Renault et al., 2014), alterations in salinity (Fuhrmann et al., 2016; Soletchnik et al., 2007) and pH of water (Fuhrmann et al., 2019) have also been reported to affect the outcome of OsHV-1 infections.

Reinforcing the complex interactions between oysters, the environment and pathogen factors, it is notable that microvariant genotypes of OsHV-1 can be present in Pacific oysters without associated mortality (Dundon et al., 2011; Hick et al., 2016; Shimahara et al., 2012). Not all host and environmental factors that contribute to the pathogenesis related to infectious processes and subsequent oyster mortality are clearly understood. The role of bacteria in summer mortality of oysters was questioned after demonstration of high loads of Vibrio spp. in the haemolymph of moribund oysters (Lipp et al., 1976). Several subsequent studies report isolation/detection of pathogenic Vibrio 64

spp. from Pacific oysters during mass mortality outbreaks such as Vibrio splendidus (Lacoste et al., 2001; Le Roux et al., 2002; Pernet et al., 2012), Vibrio aestuarianus (Garnier et al., 2008) and Vibrio harveyi (Saulnier et al., 2010). Recently, Pacific oysters exposed to a mass mortality outbreak and subsequently treated with antibiotics demonstrated delayed and reduced mortality compared to oysters that were not treated with antibiotics (Petton et al., 2015b). The suggestion of a role for bacteria present in the oyster microbiome or environment in the pathogenesis was supported by a rapid increase in the cultivable Vibrio count preceding OsHV-1 replication in the same outbreak (Petton et al., 2015b). Recently, de Lorgeril et al. (2018) demonstrated that infections caused by the variant μVar result in immunosuppression which shifts the oyster microbiome to allow opportunistic infections from bacteria such as Vibrio.

The Pacific oyster microbiome is dominated by Proteobacteria (Fernandez-Piquer et al., 2012; Hernandez-Zarate and Olmos-Soto, 2006; Trabal Fernández et al., 2014). The results of recent 16S rRNA gene studies suggest that Pseudomonas and Vibrio are not abundant members of the oyster microbiome (King et al., 2018a; Lokmer and Wegner, 2015; Trabal Fernández et al., 2014). Instead, members of the phylum Bacteroidetes, phylum Firmicutes (Trabal Fernández et al., 2014) and Arcobacter spp. (Fernandez- Piquer et al., 2012) have been shown to be abundant in the oyster microbiome. The microbiome composition can be influenced by oyster genetics (Wegner et al., 2013) and can change under various conditions that cause stress in oysters, including temperature (Lokmer and Wegner, 2015), translocation (Lokmer et al., 2016a), infection (Green and Barnes, 2010) and antibiotic stress (Lokmer et al., 2016b). The immediate environment can influence the microbiome composition of oysters, owing in part to their filter-feeding (Le Roux et al., 2016; Lokmer et al., 2016a). During times of stress, many of the environmental bacteria that are taken in may opportunistically colonize the oyster (Green and Barnes, 2010).

It is interesting to investigate how the microbiome composition of Pacific oysters with a common genetic background is affected by the environment they live in, and whether alterations in the microbiome can contribute to a differential response to the infection caused by OsHV-1. The aim of the current study was to evaluate changes in the

Pacific oyster microbiome with changes in their environment and during progression of an experimental OsHV-1 infection. Specifically, studies were conducted to: (1) assess differences in the bacterial communities in Pacific oysters with a common origin but grown in three geographically distinct estuaries; (2) to evaluate changes occurring in the microbiota, with special reference to Vibrio populations and in response to OsHV-1 infection; and (3) to evaluate the responses of different batches of oysters with distinctly different life history, to an experimental OsHV-1 infection, conducted in a controlled laboratory environment.

3.3 Materials & Methods

3.3.1 Oysters and aquarium management

Healthy triploid Pacific oysters (n=348; 30-80 mm shell length) sourced from a commercial hatchery (Shellfish Culture Ltd., Tasmania) were used in this study. Different batches were grown under commercial farming conditions for 7 months in Patonga Creek, Hawkesbury River (juveniles, 10 months old, n = 116), or for 11-12 months in Shoalhaven River (adults, 14 months old, Goodnight Oysters, Greenwell Point, n = 116) or the Clyde River (adults,16 months old, Signature Oysters, n = 116). These are 3 geographically separated estuaries in NSW, Australia which are impacted by a variety of human activities in addition to oyster farming. None of these oysters had been previously exposed to OsHV-1 based on surveillance for freedom from OsHV-1 in these estuaries and all oysters were certified negative for OsHV-1 by the competent government authority at the time of interstate transport from the hatchery to NSW. Further, the oysters were confirmed negative by testing a sample (n=12) of each batch using a real-time quantitative PCR (qPCR) assay. The oysters were transported to a physical containment level 2 aquatic facility at the University of Sydney and acclimated for 3 days before experimental infection.

The oysters from each batch were randomly allocated to eight replicate 15 L tanks (n = 14 oysters per tank), making 24 tanks in total. A static water system was maintained with aerated, artificial seawater (ASW; Red Sea® salt, 30−31 ppt salinity) at 22°C. Six of

the tanks for each batch were designated for challenge with OsHV-1 while 2 tanks were designated for oysters challenged with the negative control.

Temperature data-loggers (Thermocron®) recorded the temperature every 30 min in selected tanks. The pH of water was maintained at 8.2 (range: 8.0−8.2) and ammonia, nitrite and nitrate levels were maintained <0.25 ppm. These water quality parameters were monitored daily, and adjustments were performed as required, by water exchange. Oysters were fed with a maintenance ration of commercial algae diet (Shellfish Diet 1800, Reed Mariculture) daily in the morning. Mortality was recorded with twice daily monitoring for 7 days. Dead oysters were removed and stored at −80°C until molecular studies were conducted.

3.3.2 Source of infective OsHV-1

A cryopreserved, filtered oyster tissue homogenate with confirmed OsHV-1 infection was used as the inoculum. The tissue homogenate was stored as multiple aliquots, at -80°C with 10% v/w fetal bovine serum (Gibco) and 10% v/v glycerol. The inoculum was prepared from diseased oysters from a previous OsHV-1 related mortality event and had been used in previous experimental infections (Paul-Pont et al., 2015). The negative control was prepared as a cryopreserved, filtered tissue homogenate prepared from apparently healthy Pacific oysters from a disease-free population (Evans et al., 2015). The cryopreserved inoculums were thawed at 4°C and diluted by 1:100 in distilled water (Ultrapure™) to provide the dose described below.

3.3.3 Challenge with OsHV-1

Oysters were immersed in a solution of MgCl2 (50g/l) for 4-6h until relaxation of the adductor muscles caused the valve to open. A 100 μl aliquot of the diluted OsHV-1 inoculum (6.5 × 105 OsHV-1 DNA copies μl−1) was injected into the adductor muscle, using a 1 ml syringe and a 25-gauge needle. The control oysters were injected with a diluted, filtered oyster tissue homogenate that was negative for OsHV-1. After injection, oysters were transferred to the same designated tanks without feed for 8 hours. Physical separation and procedural care were taken to prevent cross contamination between tanks.

3.3.4 Sampling

Twelve oysters (n=4 from each batch) were sampled on Day 0 prior to injection. Thereafter, 2 live oysters per tank were sampled on Days 1, 2, 3, 5 and 7. Samples were maintained on ice until processing. Dead oysters were sampled if identified by non- responsive gaping when exposed to external stimuli including air exposure.

3.3.5 OsHV-1 DNA quantification

Oysters were carefully shucked, and gill and mantle were collected from each oyster and stored, as described in Section 2.4.1. These tissue samples were later homogenized, and the final supernatant was stored, as described in Section 2.6.1. Nucleic acids were extracted from these supernatants, purified and stored, as described in Section 2.7.1. The number of copies of the B-region of the OsHV-1 genome was determined by a qPCR assay, as described in Section 2.8.

3.3.6 Bacterial studies

Each sample was prepared from soft tissues remaining after the gill and mantle tissue sampling, as described in Section 2.4.3, and after removal of the digestive gland. The tissues were transferred into a pre-weighed stomacher bag (Interscience, France), weighed and disrupted by stomaching with 4× (w/v) sterile ASW, in a bag mixer (MiniMix, Interscience, France) for 1 min, at maximum (9) speed. On Day 0, individual oysters were homogenized and thereafter, tissues from 4 oysters from each treatment and the same batch were pooled and homogenized. A coarsely clarified homogenate (5 ml) was collected, used for bacterial culture and the remainder was stored, as described in Section 2.4.3.

Preparation of homogenates for bacterial molecular studies was by bead beating as previously described, with 150 µl of homogenates, 390 µl lysis/binding solution (MagMax™ - 96 Viral RNA Isolation Kit) in a 2 ml microcentrifuge tube containing 0.4 g of silica-zirconia beads. Homogenates were centrifuged at 900 g for 10 min and 180 μl of the supernatant was used for the nucleic acid purification with the MagMax™-96 Viral Isolation Kit. Nucleic acids were stored at -20°C until bacterial DNA quantification.

3.3.6.1 Isolation, identification and quantification of cultivable Vibrio and total bacteria

Ten microliters each from each fresh tissue homogenate was spread on a marine salt agar-blood (MSA-B) plate and a thiosulphate-citrate-bile salts-sucrose (TCBS) agar (Oxoid, UK) plate separately, as described in Section 2.5.2. MSA-B and TCBS agar were prepared as described in Section 2.5.1 (Buller, 2014). The inoculated culture plates were incubated at 23°C for 24h (MSA-B) or 48hr (TCBS agar) in a refrigerated incubator (LMS Ltd, UK). Bacterial colonies on MSA-B plates and TCBS plates were counted and the number of colony forming units (CFU)/g of oyster tissue were calculated (total cultivable bacterial count, TCBC; total cultivable Vibrio count, TCVC). Colony morphology on TCBS plates was studied and selected Vibrio colonies were prepared for cryopreservation and species identification, as described in Section 2.5.3.

3.3.6.2 Quantification of total Vibrio spp. DNA by qPCR

The qPCR assay described by Vezzulli et al. (2012) was adapted for quantification of Vibrio spp., as described in Section 2.9. Quantitative standards and positive control samples for the assay were prepared as described in Section 2.9.1. The Vibrio gene copy numbers in the standards were estimated, as described in Section 2.5.4.1. The quantification of Vibrio spp. DNA in samples was determined by standard curve quantitation method, using 7500 software v2.3 (Applied Biosystems) for a 10-fold serial dilution containing between 4.65 × 101 and 4.65 × 106 copies of Vibrio rRNA gene. 3.3.6.3 Quantification of total bacterial DNA by qPCR

The qPCR assay based on TaqMan® chemistry and described by Nadkarni et al. (2002) was used to quantify total bacteria. This qPCR assay is described in Section 2.10. The quantitative standards and the positive control samples were prepared as described in Sections 2.5.4.2 and 2.10.1. PCR runs were analyzed by the standard curve quantitation method (Applied Biosytems) based on amplification of a 10-fold serial dilution containing between 1.84 × 101 and 1.84 × 108 copies of E. coli rRNA gene. Total bacteria quantity of samples was expressed as the number of bacteria genome equivalents per mg oyster tissue. This was calculated by dividing the total 16S rDNA gene copies by the presumptive number of 16S rDNA copies per Proteobacteria genome [n = 3.5, Kormas

(2011); Vezzulli et al. (2012)], based on the species of bacteria of the phyla that dominate the oyster microbiota (Lokmer et al., 2016a; Prieur et al., 1990; Wegner et al., 2013). 3.3.7 Microbiome analysis by high throughput 16S rRNA gene sequencing

The microbial community composition was identified by high-throughput sequencing (MiSeq, Illumina) of the hypervariable region V1-V3 of the 16S rRNA gene. Twenty-six nucleic acid extracts were selected for analysis, to represent 5 different OsHV-1 infection and disease states for each of the 3 batches of oysters (Table 3.1).

Initially, the V1-V8 hypervariable region of the 16S rRNA gene was amplified using the primers 27F and 1492R [Table 2.1; (Weisburg et al., 1991)]. Reactions were prepared with the Expand High Fidelity PCR System (Roche®, Germany) and used a BioRad T100™ thermal cycler (Applied Biosystems). The PCR amplification was carried out in two stages to overcome inhibition and minimize the impact of PCR bias on diversity profiling. The first stage of amplification was carried out with 8 replicates per nucleic acid sample in a 50 µl reaction containing 200 µM of each dNTP, 0.3 µM of each primer, 5 µl of template DNA, 0.75 µl of Expand High Fidelity enzyme mix, 5 µl of

Expand High Fidelity buffer (10x, 15mM MgCl2) and sterile, nuclease free water. The PCR program was: initial denaturation of 95°C for 5 min; 30 cycles of denaturation at 94°C for 15 s, annealing at 50°C for 30 s and elongation at 72°C for 1 min; final elongation of 72°C for 7 min. Replicate PCR products were pooled, mixed by vortexing and 5 µl from the amplicon pool was used as template in stage 2 amplification using the same PCR conditions. Products were visualised by gel electrophoresis using 2% agarose gels containing RedSafe (Intron, Biotechnology).

Appropriate size PCR products were purified using the QIAamp DNA minikit (Qiagen) and submitted to the Australian Genome Research Facility (AGRF). PCR amplicons generated using primers 27F (Lane, 1991) and 519R (Lane et al. (1985); Table 2.1), using AmpliTaq Gold 360 Master mix (Life Technologies, Australia). Indexing the amplicons was performed with TaKaRa Taq DNA Polymerase (Clontech) and the resulting amplicons were measured by fluorometry (Invitrogen™ Picogreen) and

normalised to provide an equimolar pool of amplicons. These were quantified by qPCR and sequenced with Illumina MiSeq 300bp paired end chemistry.

Table 3.1 Cohorts of oysters used for microbiome analysis (On average, n=8 oysters were evaluated per batch/cohort).

Cohort Description

A Pacific oysters before OsHV-1 challenge

B Apparently healthy oysters soon after OsHV-1 challenge

C Challenged oysters at or near death

D Surviving oysters at the completion of a 7-day OsHV-1 challenge period

E Unchallenged control oysters at the end of a 7-day trial period

3.3.8 Bioinformatic analyses

Paired-ends reads were assembled by aligning the forward and reverse reads using PEAR v0.9.5 (Zhang et al., 2014), primer sequences were trimmed and USEARCH ver. 8.0.1623 (Edgar, 2010; Edgar et al., 2011) was used for quality filtering and to remove full-length duplicate sequences and singletons. Chimeric sequences were detected against the reference database (RDP Gold database) and discarded. Sequences were clustered into operational taxonomic units (OTUs) at 97% similarity and taxonomy was assigned using the RDP classifier (Liu and Wong, 2013) as implemented in QIIME, against the Greengenes database ver. 13_8 (Aug 2013) (DeSantis et al., 2006). An OTU table was generated by calculating the absolute abundance of each identified OTU for each sample, based on the number of sequenced reads. OTUs with an abundance of less than 10 reads per sample were discarded.

3.3.9 Data analyses

Survival analyses were conducted using the Kaplan-Meier estimate (Kaplan and Meier, 1958), notwithstanding an experiment design focused on longitudinal sampling. Oysters collected during random sampling and surviving oysters at the end of the

experiment were censored at the respective observation time. Survival time was measured as days from the onset of the injection when oysters were injected with the OsHV-1 inoculum. Each batch of oysters was compared using a mixed Cox regression model (Cox, 1972), taking into account the random effect of tank, using RStudio ver. 3.4.4 (R Core Team, 2018). The proportionality of hazards (PH) was checked based on Schoenfeld residuals (Schoenfeld, 1982).

The OsHV-1 DNA quantity, total bacteria, total Vibrio, TCBC and TCVC in oysters were compared between batches, between live and dead oysters of the same batch, between days after OsHV-1 challenge in a batch and with relevant control groups. Data were log10 transformed for normal distribution. Kruskal-Wallis test was used for OsHV-1 data analysis and the analyses of bacterial data in oysters before OsHV-1 challenge. The bacterial data (total bacteria, total Vibrio, TCBC and TCVC) in oysters during the viral challenge period were done using separate generalized linear models (GzLM, SPSS Statistics ver. 22; IBM SPSS Cooperation, Somers, NY, USA). The main effects of batch including the geographic origin of oysters (Patonga Creek, Shoalhaven River, Clyde River), treatment (OsHV-1 injected or negative control), outcome of infection (live or dead) and the days post-injection (Day 1, 2, 3, 5 and 7) were tested together with possible interactions. Post-hoc pairwise mean comparisons were made using the least significant difference method. The results were presented as geometric means and their corresponding 95% confidence intervals of OsHV-1 genome equivalents, bacterial genome equivalents or Vibrio genome equivalents/mg, CFU values/g (for TCBC and TCVC), respectively. Finally, the correlation between OsHV-1 load and Vibrio gene copies in oyster tissues was tested using Spearman’s correlation coefficient (SPSS Statistics ver. 22.0). Significance was set at p < 0.05 for all statistical analyses.

The relative abundance of bacterial families in each sample was graphically represented using 100% stacked 2-D column graphs (Excel, Microsoft). Diversity analyses were performed with Paleontological Statistics (PAST) software ver. 3.16 (Hammer et al., 2001). Alpha diversity was assessed by means of Shannon’s diversity index and Simpson’s evenness index. Dissimilarity of bacterial community structure between samples from different batches and different cohorts (beta diversity) was

visualized by non-metric multidimensional scaling (n-MDS) based on the two- dimensional Bray-Curtis dissimilarity index (BC). One-way permutational multivariate analysis of variance (PERMANOVA) was used for statistical analysis of beta diversity. Significance was set at p < 0.05 for all statistical analyses.

3.4 Results

3.4.1 Mortality of oysters

The onset of mortality occurred 2 days post-OsHV-1 injection and the cumulative mortality was different for the three batches of oysters (Table 3.2; Figure 3.1, p<0.05). The highest cumulative mortality was seen in the batch from Patonga Creek (63.1%) compared to oysters from Clyde River (38.1%) and Shoalhaven River (17.7%). No mortality was observed in the negative control oysters.

Table 3.2 Odds of mortality for oysters injected with OsHV-1 depending on the batch of oysters (Odd’s ratio, standard error (SE), the 95% confidence interval of the risk ratio and p-value).

Contrasts Odd’s SE 95% Confidence Interval p-value ratio (CI) Patonga vs Clyde 3.00 0.32 1.61 - 5.54 <0.001 Shoalhaven vs Clyde 0.48 0.42 0.15 - 1.53 0.21 Shoalhaven vs Patonga 0.16 0.41 0.06 - 0.46 <0.001 Patonga, Patonga Creek; Shoalhaven, Shoalhaven River; Clyde, Clyde River

Figure 3.1 Kaplan-Meier survival curves for each batch (Clyde River, Patonga Creek, Shoalhaven River) of oysters challenged with OsHV-1 (solid line) and 95% confidence intervals (dashed line). Each batch comprised n=73-86 oysters at the onset of the viral challenge and n=12 live oysters were sampled from each batch on days 1, 2, 3, 5, and 7. Different letters (a and b) indicate significance (p<0.05) identified by a mixed Cox regression model.

3.4.2 Quantification of OsHV-1 DNA by qPCR

The live oysters from Patonga Creek that were sampled during the infection trial had higher OsHV-1 DNA concentrations compared to their counterparts from the other two locations (Table 3.3; p<0.05). This batch also had the highest mortality and a higher OsHV-1 DNA concentration at the time of death compared to dead/moribund oysters from the Clyde River (Table 3.3; p<0.05). In all three batches, longitudinal random sampling of live oysters indicated that the quantity of OsHV-1 DNA increased on Day 2 compared to Day 1 post-injection (Figure 3.2A), which coincided with the onset of mortality. All dead oysters were positive for OsHV-1 DNA and OsHV-1 DNA was not detected in oysters injected with the negative control inoculum. The quantity of OsHV-1

DNA in moribund/dead oysters was higher than in live, apparently healthy oysters sampled at random on Days 1, 2, 3, 5, and 7 post-injection (Table 3.3; p<0.001).

Table 3.3 Mean OsHV-1 concentrations in oyster tissues (geometric means and their corresponding 95% confidence intervals) during the infection.

Status of oysters Mean OsHV-1 concentration 95% Confidence Interval

(OsHV-1 genome (CI) equivalents/mg)

Live oysters: 1.61 × 102 8.24 × 101 – 3.14 × 102

Patonga Creek 1.22 × 103 * 2.95 × 102 – 5.04 × 103

Shoalhaven River 4.05 × 101 1.61 × 101 – 1.02 × 102

Clyde River 1.31 × 102 4.45 × 101 – 3.88 × 102

Dead oysters: 1.36 × 105 8.48 × 104 – 2.20 × 105

Patonga Creek 2.25 × 105 * 1.20 × 105 – 4.22 × 105 Shoalhaven River 5.55 × 104 8.15 × 103 – 3.77 × 105

Clyde River 9.69 × 104 4.75 × 104 – 1.98 × 105

* OsHV-1 concentration in live Patonga oysters was higher than the other two batches. It was also higher in dead Patonga oysters compared to dead oysters from Clyde River.

3.4.3 Quantification of bacteria

There was no difference in any bacteriological parameter (total bacteria load, total Vibrio load, TCBC and TCVC) between the three batches of oysters, at the beginning of the infection trial (p > 0.05 for each parameter). Considering temporal dynamics during the infection, the total Vibrio load and TCVC in live oysters from Patonga were higher on Day 1 compared to Day 2 post-viral injection (Figure 3.2B; Table 3.4; p < 0.05). There

was no such trend in the other two batches (p > 0.05). Overall, the highest Vibrio quantity was seen in oysters from Patonga during the viral infection (Table 3.5; p < 0.05). The batches from Shoalhaven and Clyde Rivers demonstrated increases in the total bacterial load in live oysters on Day 2 as opposed to Day 1 post-injection (Figure 3.2C; p < 0.05). The total bacterial concentration and the total Vibrio concentration were higher in the tissues of dead or moribund oysters compared to apparently healthy oysters sampled on Days 1, 2, 3, 5, and 7 post-injection (Table 3.5; p < 0.05). Interestingly, no change in total bacteria and total Vibrio was observed in the negative control oysters from all 3 batches, during the trial (p > 0.05).

Although changes were noted with the Vibrio load in oysters challenged with OsHV-1, the average relative percentage of Vibrio spp. from the total bacterial population, remained low with 0.26% in pre-challenged oysters and reaching a maximum of 5.25% on Day 1 post-injection. Overall, a strong correlation was observed between the OsHV-1 DNA load and the total Vibrio load in OsHV-1 infected oyster tissues (r = 0.6; p < 0.001). Representative Vibrio isolates from the OsHV-1 infected oysters were identified as Vibrio alginolyticus (3/6) and Vibrio mediterranei (2/6), by biochemical methods, as described in Section 2.5.3 of this thesis.

Figure 3.2 Temporal distribution of (A) OsHV-1 viral load (log10 OsHV-1 genome equivalents +1/mg), (B) Total Vibrio load (Vibrio genome equivalents/mg tissue) and (C) Total bacterial load (bacterial genome equivalents/mg) in live oysters that were sampled longitudinally on Days 0, 1, 2, 3, 5 and 7 post-OsHV-1 injection. All three batches showed a significant/nearly significant increase in OsHV-1 concentration on Day 2 compared to Day 1 (Patonga, p < 0.05; Shoalhaven, p = 0.053; Clyde, p = 0.052). The Patonga oysters also showed an increase in total Vibrio load on Day 1 compared to Day 2 post-viral injection (p < 0.05). All data are represented as log10 means ± SE. 77

Table 3.4 Results of Generalised Linear Model analysis of total bacterial load and total Vibrio load with the effects of geographic origin of oysters (batch), OsHV-1 challenge (treatment), days post-OsHV-1 injection (day) and outcome of infection (status) accounted for in the model.

Total bacterial load Total Vibrio load

df X2 p-value df X2 p-value

Intercept 1 2609.97 <0.001 1 644.97 <0.001 Batch 2 1.05 0.593 2 52.69 <0.001 Treatment 1 1.85 0.174 1 2.40 0.121 Day 6 28.61 <0.001 6 19.39 0.004 Status 1 30.78 <0.001 1 36.41 <0.001 Batch × Group × Day × 28 108.62 <0.001 24 118.97 <0.001 Status Note: df, degree of freedom; X2, Wald Chi Square

Table 3.5 Total bacterial load and total Vibrio load (geometric means† and their corresponding 95% confidence intervals, CI) in oysters during OsHV-1 infection.

Status of oysters Mean bacterial load a (95% CI) Mean Vibrio load b (95% CI)

Live oysters: 1.08 × 104 (7.89 × 103 –1.47 × 104) 4.24 × 101 (3.14 × 101 – 5.73× 101) Patonga Creek 2.72 × 104 (1.75 × 104 – 4.20 × 104) 2.88 × 102 (1.91 × 102 – 4.36 × 102)* Shoalhaven River 1.48 × 104 (9.17 × 103 – 2.38 × 104) 1.89 × 101 (1.16 × 101 – 3.07 × 101) Clyde River 1.86 × 104 (1.20 × 104 – 2.89 × 104) 1.15 × 102 (7.74 × 101 – 1.72 × 102) Dead oysters: 9.24 × 104 (5.66 × 104 –1.51 × 105) 4.19 × 102 (2.69 × 102 – 6.54 × 102) a Bacterial genome equivalents/mg; b Vibrio genome equivalents/mg; †Geometric means were derived from back-transformed model means. *The mean Vibrio load in Patonga oysters were higher than in the other two batches (p<0.05).

3.4.4 Bacterial community structure in different batches of oysters

Both Shannon’s diversity index (H) and Simpson’s evenness index (S) were different between the oysters grown in the three different estuaries, before acclimation to the laboratory conditions and OsHV-1 challenge (p<0.05). Both H and S were highest in oysters from Patonga Creek (H=1.77, S= 0.74) compared to those from the Clyde River (H=0.77, S=0.33) and Shoalhaven River (H=1.14, S=0.54). While the relative abundance of family Brachyspiraceae was quite low (0.8%) in the batch from Patonga Creek, a comparatively higher relative abundance was seen in the other two batches (Shoalhaven, 24.3%; Clyde 10.3%; Figure 3.3). In contrast, the relative abundance of family Moraxellaceae was higher in oysters from Patonga Creek (13.5%) compared to the other two batches (Shoalhaven, 2.9%; Clyde 1.7%). Further, the relative abundance of family Deinococcaceae was observed to be 16.1 % in the oysters from Patonga Creek while it was negligible in the other two batches (Figure 3.3).

3.4.5 Changes in bacterial community during OsHV-1 infection

The relative abundance of bacterial families changed in all three batches of oysters, 24 h post-OsHV-1 injection. However, the direction of changes was different between batches. For example, the relative percentage of family Brachyspiraceae increased up to 60.4 % in Patonga oysters while it decreased to 3.2% and 2.4% in oysters from Shoalhaven and Clyde Rivers, respectively (see section 3.4 for initial values; Figure 3.3). Moreover, the family Hyphomicrobiaceae which was either absent or present in negligible amounts in the pre-challenge oysters was seen to increase in oysters from Shoalhaven River (2.1%) and Clyde River (18.1%). This was not observed in Patonga oysters. During the OsHV-1 infection, the dead/moribund oysters from Patonga Creek showed a decrease in bacterial diversity (H=1.66, S=0.68) compared to pre-challenge (H=1.77, S=0.74; p<0.05). However, the dead or moribund oysters from the other two locations, with an originally lower diversity, did not demonstrate such a decrease (p>0.05). Furthermore, the bacterial community structure in Patonga and Shoalhaven oysters that survived the OsHV-1 challenge at the end of the 7-day trial, were similar to their counterparts in the pre-challenged cohort (Figure 3.4; p>0.05). Moreover, the bacterial community structure in Patonga and Clyde oysters that were challenged with the

negative control, were not different from the pre-challenge counterparts, at the end of 7 days in the laboratory.

100%

90%

80%

70%

60%

50%

40%

Relative abundanceRelative 30%

20%

10%

Clyde

Clyde Clyde

Clyde

Patonga

Shoalhaven

Shoalhaven Shoalhaven Shoalhaven A B C D E

Enterobacteriaceae Brachyspiraceae Deinococcaceae Moraxellaceae Verrucomicrobiaceae Hyphomicrobiaceae Dermacoccaceae Patulibacteraceae Propionibacteriaceae Rhizobiaceae Other

Figure 3.3 Relative abundance of bacterial families in oysters farmed in different estuaries: before OsHV-1 challenge (A); apparently healthy oysters 24h after OsHV-1 challenge (B); challenged oysters at or near death (C); surviving oysters at the completion of a 7-day OsHV-1 challenge period (D); unchallenged control oysters at the end of the trial period (E). Oysters (n = 26) were used in the bacterial diversity profiling. Patonga, oysters from Patonga Creek; Shoalhaven, oysters from Shoalhaven River; Clyde, oysters from Clyde River. The bacterial families that constituted >5% of an individual sample and were found in at least two samples are presented, and the rest are indicated as ‘others’.

Figure 3.4 Non-metric multidimensional scaling based on two-dimensional Bray-Curtis

dissimilarity. Colours indicate different cohorts: black (A), before OsHV-1 challenge; green (B), 24h after viral challenge; red (C), dead/moribund oysters; blue (D), survived oysters; yellow (E), control oysters at the end of the experiment. Shapes indicate different locations of origin of the oysters: circle, Patonga Creek; square, Shoalhaven River; star, Clyde River.

3.5 Discussion

The present study discusses how the bacterial community composition of Pacific oysters with a common hatchery origin can be influenced by the geographic site at which they are grown. It also discusses how these influences are related to the changes in the Pacific oyster microbiome and OsHV-1 load during an experimental OsHV-1 infection conducted in a controlled environment. Further, the study the provides insights as to how these differences in microbiome composition affect the outcome of the OsHV-1 infection, in addition to the impact of age factor.

Our results showed that the bacterial community composition of the Pacific oyster microbiome was different between the oysters grown in the three geographically distinct estuaries despite their common hatchery origin. The filter-feeding pattern of oysters leading to the uptake of environmental microbes with their feed can be attributed to these

differences in microbiomes of oysters with a common genetic origin. It should be noted that the microbiota in estuarine water was not analysed in this study. The study involved three batches of oysters from two broad age categories (juvenile oysters (n=1) and adult (n=2). The adult batches were of approximately equal ages (14 and 16 months). In addition to the difference of age between the juveniles and adults, we also observed a difference between the microbiome composition of the two batches of adult oysters, indicating an influence of farming environment in shaping the microbiome composition. Further studies are required to focus on controlling other factors such as age to further evaluate the influence of environment on the microbiome. A metagenomic study in Crassostrea virginica revealed differences in microbiomes in oysters reared in different sites (Ossai et al., 2017) while conventional bacterial studies conducted using C. virginica have also shown that the bacterial composition in oysters varies with their environment (Prieur et al., 1990). However, it is also important to note that the microbiome composition is also influenced by host-related factors as well as by interactions within the microbiome, in the long run (Lokmer et al., 2016b).

In the present study, I observed changes in the oyster microbiome soon after the viral challenge. These changes may primarily reflect the initial response of the microbiome to the viral infection. Having said that, it is interesting to note that the microbiome of Patonga oysters behaved differently to the other two batches along with a higher OsHV-1 DNA concentration in tissues and higher mortality. The difference in the mortality pattern was consistent with previous studies (Hick et al., 2018; Renault et al., 1995; Whittington et al., 2015b) whereby juvenile oysters (<1 year) had higher mortality than adult oysters (>1 year). However, it was interesting to note that there was no difference in mortality between the two batches of adults despite a higher concentration of OsHV-1 in Clyde oysters which can be corresponded to differences noted in the initial microbiome composition. Moreover, the oysters from Patonga Creek demonstrated the highest initial bacterial diversity and showed a decrease in bacterial diversity during the moribund stage. This was not seen with the other two batches, indicating a differential response of the microbiome. In a study conducted by Wegner et al. (2013), Pacific oysters with initially high microbial diversity had shown a decrease in diversity following disturbance due to heat stress, whereas oysters with originally lower diversity did not 82

show such a decrease after exposure to heat. Low bacterial diversity has repeatedly been found with impaired health in oysters (Garnier et al., 2007; Green and Barnes, 2010). Thus, the decrease in bacterial diversity which precedes mortality can be considered a good indicator in declining health in oysters (Lokmer and Wegner, 2015; Wegner et al., 2013). Furthermore, studies involving pathogenic Vibrio infections in Pacific oysters have shown bacterial community structure disruption in dead/moribund oysters, characterized by very low diversity (Lokmer and Wegner, 2015).

The changes in the microbiome that were observed in oysters challenged with the virus, were not seen in the control oysters from two batches, in our study. The bacterial community composition of those control oysters was not different from their pre- challenged counterparts, at the end of 7 days in the laboratory. As the control oysters were also subjected to muscle relaxation and was given an injection, the disturbance to the microbiome prior to the viral challenge, appear to be transient. In contrast, the changes induced by OsHV-1 infection seem to remain and can be considered as an indication of a disease state. Oyster families susceptible to OsHV-1 showed disruption of the bacterial community structure when cohabitated with OsHV-1-injected oysters (de Lorgeril et al., 2018).

Despite the inter-batch differences observed in bacterial community composition after 24 h of the viral challenge, the total bacterial load remained similar. The persistent bacterial community composition at higher taxonomic levels regardless of the changes in lower taxonomic levels (Lokmer et al., 2016a) can be attributed to this static bacterial load. In the present study, the Patonga oysters displayed the highest total Vibrio load with a peak on Day 1 post-injection. This peak may be associated with the higher and early (on Day 2 post-injection) mortality in Patonga oysters which was not seen in oysters from the other two batches. While the Shoalhaven oysters reported few mortalities, no mortalities were observed in Clyde oysters as early as Day 2 post-injection. Although I could not test the pathogenicity of the Vibrio isolates of our study, it is possible that this increase in Vibrio, to have a role in the final cumulative mortality. Stress-related shifts in the oyster microbiome resulting in increase in the potentially pathogenic bacteria are thought to contribute to oyster mortality (Lokmer and Wegner, 2015).

Although differences in temporal Vibrio dynamics and the total Vibrio load were noted between batches, I observed a low average relative abundance of Vibrio spp. in all batches, throughout the course of the viral infection. This was consistent with previous studies reporting oyster mortality events in the field (King et al., 2018a) as well as experimental conditions (Lokmer and Wegner, 2015). Trabal et al. (2012) also reports a very low abundance of Vibrio in gut tissues obtained from healthy Pacific oysters. Contrary to the conventional knowledge that identifies Vibrio spp. as one of the dominating bacterial genera in oysters (Colwell and Liston, 1960; Prieur et al., 1990), a large proportion of the oyster microbiome cannot be cultivated by standard procedures (Fernandez-Piquer et al., 2012; Romero and Espejo, 2001). Previous literature on Eastern oysters (C. virginica) and Sydney rock oysters (Saccostrea glomerata) also report low relative abundances of Vibrio spp. (Green and Barnes, 2010; Ossai et al., 2017). Nevertheless, Vibrio spp. can have an impact on the host health, despite its low relative abundance (Thurber et al., 2009).

I also observed a wider range of OsHV-1 DNA concentrations (102 - 106 viral copies/mg oyster tissue) in oysters that died during this infection trial. Previous field and lab studies have demonstrated viral concentrations exceeding 106-107 copies mg/tissue to be associated with moribund/dead oysters infected with OsHV-1 (Oden et al., 2011; Paul- Pont et al., 2014; Pepin et al., 2008). As highlighted by Paul-Pont et al. (2015), variations in the OsHV-1 DNA concentration in moribund oyster tissues can be expected due to various reasons such as genetic background, age, physiology and life history. Despite the highest OsHV-1 DNA concentration in the youngest batch (10 months), the lowest OsHV-1 concentration was not reported from the oldest batch (16 months) but a younger batch (14 months), in this study. This indicated a potential role of factors other than age in determining the physiological response of oysters to OsHV-1. Apart from the role played by age in OsHV-1 infections (Hick et al., 2018; Paul-Pont et al., 2013b), the differences in the microbiome composition among the batches may have differentially influenced the viral multiplication in these organisms. Recent studies on OsHV-1 disease have highlighted the role of size and other factors such as life history and genetics, on oyster mortality (Azéma et al., 2017a; Azéma et al., 2017b; Petton et al., 2015a). Interactions between the age and size of oysters in determining the susceptibility of Pacific oysters to 84

OsHV-1-related mortality (Hick et al., 2018) suggest complex interactions of a range of factors other than age, in causing mortality.

There was a strong correlation between the OsHV-1 DNA load and Vibrio load in infected oysters, in the present study. Conventional culture methods used in Vibrio quantification have not seen such a correlation between the Vibrio concentration and the OsHV-1 DNA load in oysters exposed to a field mortality outbreak (Petton et al., 2015b). Lemire et al. (2015) have shown that the non-virulent strains of Vibrio spp. inhabiting healthy oysters get progressively replaced by phylogenetically coherent, virulent strains of Vibrio during the onset of a natural infection. These non-virulent strains are also expected to facilitate the disease caused by the virulent strains. Thus, the difference in correlation of Vibrio gene copy number and cultivable Vibrio load, to OsHV-1 DNA load and can be attributed to the ability of the PCR assay in detecting non-cultivable Vibrio strains (Cai et al., 2006), or the non-virulent strains of Vibrio that were dead and were replaced by pathogenic Vibrio strains.

The apparently healthy oysters used in this study were injected with a bacteria-free, partially purified tissue homogenate carrying OsHV-1 and were maintained in artificial seawater throughout the experimental period. Therefore, all Vibrio strains quantified and isolated in this study should originate from the oyster microbiome. Vibrio alginolyticus that was isolated in this study have been previously reported in mass mortality events of Pacific oysters (Go et al., 2017). As mentioned above, I could not test the pathogenicity of the Vibrio isolates of this study nor could I identify the same species with 16S rRNA metabarcoding as the genetic resolution did not permit identification of Vibrio at species level. Testing the pathogenicity of Vibrio isolates may have provided valuable clues to identify the pathogenic role of Vibrio spp. present in the Pacific oyster microbiome, during OsHV-1 infections. In an experiment where oysters exposed to a field mortality outbreak and subsequently treated with chloramphenicol, Petton et al. (2015b) demonstrated that a high load of OsHV-1 DNA (≥ 108 viral copies/mg tissue) is not sufficient for a full expression of the disease by demonstrating a delayed and reduced oyster mortality in the absence of bacteria. In this regard, the indication that the pathogenic species or strains of Vibrio in healthy oysters become pathogenic when there

is a primary viral infection, needs to be studied further. Moreover, it is important to investigate whether the mass mortality events reported among Pacific oysters are caused by several different infectious diseases or by a single polymicrobial disease, by targeting Vibrio spp. alone.

3.6 Conclusion

The composition of the bacterial community in Pacific oysters grown in different geographic sites was different despite their common hatchery origin. In addition to the differences related to age, different outcomes of OsHV-1 challenge were seen between same-aged batches, with differences in the initial microbiome. The initial bacterial community composition did not change in some batches of oysters that survived OsHV-1 challenge and in oysters challenged with the negative control. The higher Vibrio load in oysters with higher OsHV-1 DNA content and higher mortality suggests a role of Vibrio in mortality associated with OsHV-1 infections. Further studies are needed to evaluate the direct and indirect influences of Vibrio and other members of the resident microbial community of oysters during OsHV-1 infection.

CHAPTER 4

The role of tissue type, sampling and nucleic acid purification methodology on the inferred composition of Pacific oyster (Crassostrea gigas) microbiome.

4.1 Abstract

This study evaluated methods to sample and extract nucleic acids from Pacific oysters to accurately determine the microbiome associated with different tissues. Samples were collected from haemolymph, gill, gut and adductor-muscle, using swabs and homogenates of solid tissues. Nucleic acids were extracted from fresh and frozen samples using three different commercial kits. The bacterial DNA yield varied between methods (P < 0.05) and each tissue harboured a unique microbiota, except for gill and muscle. Higher bacterial DNA yields were obtained by swabbing compared to tissue homogenates and from fresh tissues compared to frozen tissues, without impacting the bacterial community composition estimated by 16S rRNA gene (V1-V3 region) sequencing. Despite the higher bacterial DNA yields with QIAamp® DNA microbiome kit, the E.Z.N.A.® Mollusc DNA kit identified twice as many operational taxonomic units (OTUs) and eliminated PCR inhibition from gut tissues. Sampling and nucleic acid purification substantially affected the quantity and diversity of bacteria identified in Pacific oyster microbiome studies and a fit-for-purpose strategy is recommended. Accurate identification of Pacific oyster microbial diversity is instrumental for understanding the polymicrobial aetiology of Pacific oyster mortality diseases which greatly impact oyster production.

Keywords: Pacific oyster, Crassostrea gigas, microbiome, bacterial DNA extraction, oyster tissue sampling, nucleic acid extraction

NOTE: This chapter is a previously published paper reformatted to the specifications of this thesis (refer to page vii for reference). I am the first author of this publication. The text contained within this chapter is identical to the aforementioned publication except for an additional table in the paper which is included in the Chapter 2 of this thesis (Table 2.1). The QIIME analysis of this Chapter was carried out by Dr. Andrew McPherson.

4.2 Introduction

Pacific oysters (Crassostrea gigas) harbour a diverse bacterial community (Lokmer and Wegner, 2015; Prieur et al., 1990) with functional specializations (Lokmer et al., 2016a; Lokmer et al., 2016b). For example, components of the haemolymph microbiota of Pacific oysters produce antimicrobial compounds that prevent colonization by external pathogens and subsequent disease (Defer et al., 2013; Lokmer et al., 2016b). The gill microbiota is considered stable and is relatively enriched with symbiotic bacteria which are integrated into the gill tissues (Prieur et al., 1990; Roterman et al., 2015; Zurel et al., 2011). These symbiotic bacteria are typically found within cells of the gill lamellae (Prieur et al., 1990). In contrast, the microbiota of the gonads and digestive system generally consists of transient bacteria (Roterman et al., 2015; Zurel et al., 2011). These transient gut microbiota are concentrated in the lumen of the hind gut, comprising of bacteria that withstand the digestive enzymes of the oyster (Prieur et al., 1990). Thus, studies that evaluate the nature and role of commensal microbiota require consideration of the differences in the microbiota and the relationship with different oyster tissues. Current research that focuses on understanding the polymicrobial aetiology of Pacific oyster mortality for example, Petton et al. (2015b) and de Lorgeril et al. (2018b), relies on accurate identification of Pacific oyster microbiota. In this context, it is important to accurately identify the diversity of bacteria to understand their dynamics in Pacific oyster mortality.

Microscopic observations of oyster tissue homogenates reveal 105 times more bacteria per gram of tissue than can be determined by quantification from culture on agar plates, indicating that a great proportion of the oyster microbiome cannot be cultivated by standard procedures (Fernandez-Piquer et al., 2012; Romero and Espejo, 2001). Molecular techniques are providing new insights into the diversity of microbiota associated with oysters (Fernandez-Piquer et al., 2012; Green and Barnes, 2010; Hernandez-Zarate and Olmos-Soto, 2006; Pierce and Ward, 2018; Romero and Espejo, 2001). In this regard, sequencing of the bacterial 16S rRNA gene is a common approach which requires bacterial DNA extraction from oyster tissues (Hernandez-Zarate and Olmos-Soto, 2006; King et al., 2012; Lokmer et al., 2016a; Mardis, 2008; Romero et al.,

2002). Microbial DNA extraction is a critical step in metagenomics which introduces potential limitations (Morita et al., 2007). The quantity of host DNA greatly outweighs microbial DNA (Feehery et al., 2013) which limits the sensitivity of nucleic acid-based microbial diagnostic systems (Horz et al., 2010; Kostić et al., 2007). Differences in CpG methylation density can be used to enrich microbial DNA in the presence of vertebrate host DNA (Feehery et al., 2013). However, the comparatively low level of genome methylation in invertebrates reduces the effectiveness of enrichment for microbial DNA using this method (Zemach et al., 2010). Therefore it is necessary to try alternative approaches for the selective enrichment of microbial DNA present in Pacific oyster tissues, such as the removal of host DNA prior to the disruption of bacterial cells and sampling specifically for bacteria (Horz et al., 2010).

Further to oyster DNA, other constituents of mollusc tissues should also be taken into consideration when optimizing DNA extraction. Molluscs secrete mucopolysaccharides and polyphenolic proteins which co-purify with DNA and interfere with enzymatic processing of nucleic acids (Pereira et al., 2011; Winnepenninckx et al., 1993). Inhibitors of PCR have been reported in oyster tissues severely limiting downstream PCR applications (Abolmaaty et al., 2007; Kaufman et al., 2004). Thus, removal of inhibiting substances may improve the quality and detection of extracted DNA. Furthermore, significant differences in bacterial community structure have been observed in studies involving live oysters from retail outlets after storage at various temperatures including chilling (Fernandez-Piquer et al., 2012). Therefore, identification of appropriate storage and preservation methods for Pacific oyster tissues are also needed to minimise potential post-collection bias.

The objective of this study was to identify preferred methods to collect and prepare samples from C. gigas to accurately determine the microbiome associated with different tissues. Different tissue-types, sampling methods, and commercial nucleic acid purification kits were evaluated. The effects of frozen storage were also evaluated. Physical methods to enrich bacterial cells in the samples prior to extraction of microbial DNA were tested and compared to microbiome enrichment by a commercial kit which

removes host genomic DNA. Methods were identified to maximise the yield of bacterial DNA and to provide an unbiased DNA template for bacterial diversity profiling.

4.3 Materials and Methods

4.3.1 Oysters

Triploid Pacific oysters produced in a hatchery in Tasmania (Shellfish Culture Pty Ltd., Batch SPL14B) were collected from a commercial farm at Woolooware Bay (Georges River), NSW, Australia (n = 10). These oysters were three years and eight months of age (length 14.6 ± 1.4 cm), had been grown on intertidal trays and were survivors of disease caused by Ostreid herpesvirus-1 infection in the same cohort (Hick et al., 2018). The use of triploid Pacific oysters in this study is reflective of commercial farming of this species in Australia. The external valves of the oysters were thoroughly cleaned, first by scrubbing and rinsing with water and finally by wiping with 70% ethanol to minimise contamination of internal tissues by surface bacteria. The oysters were then carefully shucked and held on ice for < 10 min until tissue samples were collected. The details of tissue sampling, disruption, microbial enrichment and nucleic acid extraction methods tested in this study are summarized in Table 4.1. The Pacific oyster is not considered by the NSW Animal Research Act 1985, thus approval from the University of Sydney Animal Ethics committee was not required for this study.

Table 4.1 Workflow carried out for optimizing tissue sampling, tissue disruption, microbiome enrichment and nucleic acid extraction for analysis of Crassostrea gigas microbiome Tissue type n Sampling Tissue disruption method/s Sampled Microbiome enrichment Extraction method amount method Haemolymph 10 Aspiration PK 1 ml Host cell lysis SC* BB Host nucleic-acid degradation Muscle 10 Swab PK NA Host cell lysis SC* BB Host nucleic-acid degradation Gill 10 Swab PK NA Host cell lysis SC* BB Host nucleic-acid degradation Gill 10 Swab PK NA No enrichment SC† Binding and removal of MPL Gill 10 Tissue PK 30 mg No enrichment MB‡ Gill 10 Tissue PK, BB, 30 mg No enrichment SC† Binding and removal of MPL Gill 10 Tissue Stomaching 120 mg Low speed centrifugation MB‡ Gill 10 Tissue Stomaching 120 mg Low speed centrifugation SC* Gill 10 Tissue Stomaching 120 mg Filtration MB‡ Gut 10 Swab PK NA Host cell lysis SC* BB Host nucleic-acid degradation Gut 10 Swab PK NA No enrichment SC† Binding and removal of MPL Gut 10 Tissue PK 30 mg No enrichment MB‡ Gut 10 Tissue PK, BB, 30 mg No enrichment SC† Binding and removal of MPL *QIAamp® DNA microbiome kit, †E.Z.N.A.® Mollusc DNA kit, ‡MagMAX™ CORE Nucleic Acid Purification Kit. BB, Bead-beating; BZ, Benzonase; MPL, mucopolysaccharides; NA, not applicable; PK, Proteinase K digestion; SC, spin column extraction; MB, magnetic beads (MagMax™ Express-96)

4.3.2 Tissue sampling

Samples were collected from haemolymph, gills, adductor-muscle and gut of each oyster according to the procedures detailed below. Each tissue sample from the same oyster was dissected using separate sterile scalpel-blades and forceps to reduce cross contamination between tissues as well as between oysters. Both swab and tissue samples were collected. Swab samples were always rotated in either sterile, artificial seawater (ASW; Red sea Salt®, 30 ppt) or ML1 buffer (E.Z.N.A.® Mollusc DNA kit, Omega Bio-Tek, USA). The methods A, B and C were applied only for tissue samples.

Haemolymph samples were collected from the pericardial sac using sterile 1 ml syringes with 18G needles. Care was taken to avoid contaminating the samples with fluid from the pallial cavity. Approximately 1 ml of haemolymph from each oyster was collected into sterile 2 ml tubes and stored on ice until nucleic acids were extracted on the same day. The samples were briefly mixed by repeated inversion of the tubes prior to nucleic-acid extraction.

Two gill-swab samples were collected from each oyster from the central gill-surface of the second gill-lamella after excision of the first gill-lamella. Sterile flocked nylon swabs (FLOQSwabs™, Copan®, Italy) were passed once across a surface of approximately 2 cm2 with firm pressure. A different area of the central gill surface was targeted for each swab sample. One swab was rotated in 1 ml of ASW while the other was rotated in 350 µl of ML1 buffer for at least 20 s and the swabs were discarded. Liquid samples were stored on ice and nucleic-acid extraction was conducted on the same day. The high salt ML1 buffer contains the cationic detergent, cetyltrimethylammonium bromide (CTAB) which lyses host cells and bacterial cells and selectively binds polysaccharides and proteins (Wilson, 2001).

Additional gill tissue samples from each oyster were collected aseptically onto sterile Petri-dishes. The target tissues were central on the second gill lamella, avoiding the tissue at the extremities and surfaces which had been swabbed. Approximately 100 mg of gill tissue from each oyster was dissected and collected into separate, sterile tubes and stored immediately at -80°C. The remaining gill tissue from the target location was excised and divided into approximately 30 mg pieces aligned to a grid marked on the outside of a sterile Petri-dish lid. This facilitated random selection of two 30 mg portions of gill-tissue which were used fresh for tissue disruption on the same day, as described in methods A and B

below. Another selection of 120 mg of remaining gill tissue was homogenized separately on the same day to facilitate nucleic acid extraction from gill tissues, after physically enriching the samples for bacterial DNA by centrifugation and sedimentation or filtration of host cells, as described in method C below. Samples were maintained on ice during the selection procedures.

Samples from adductor muscle were collected by making a transverse incision into the central adductor muscle and subsequently collecting a swab sample from the interior of the muscle, using a sterile, flocked nylon swab applied with firm pressure over an approximately 2 cm2 area. The swab was rotated in sterile ASW in a tube, as described for gill-swab samples above.

The digestive gland (hereafter referred to as gut) of each oyster was incised along a sagittal plane into halves. Two swab samples were collected, each from an area of approximately 2 cm2 of the inner mucosal surface using sterile, flocked nylon swabs. A different area of the inner mucosal surface was targeted for each swab sample. The swabs were swirled in tubes containing sterile ASW or ML1 buffer as described for gill-swab samples. Two 30 mg portions of tissue were excised from a randomly selected position on the remaining half of the gut, weighed aseptically into separate sterile tubes and stored on ice until tissue disruption was carried out on the same day, as described in methods A and B below.

4.3.3 Tissue disruption and nucleic acid extraction

Different tissue disruption, selective bacterial DNA enrichment and nucleic acid extraction methods were evaluated, as summarised in Table 4.1. Three commercial nucleic acid extraction kits were tested for efficiency of bacterial DNA extraction and for suitability of resulting extracts for microbiome studies. Low-speed centrifugation and filtration were also trialled for selective enrichment of bacterial cells. The QIAamp® DNA microbiome kit (Qiagen, Germany) was selected to evaluate the efficiency of bacterial DNA extraction from oyster tissues after selective lysis of host cells and digestion of host genomic DNA. Meanwhile the E.Z.N.A.® Mollusc DNA kit was selected for its use of CTAB to remove mucopolysaccharides present in mollusc tissues. The idea was to overcome PCR inhibition in downstream applications and to assess its impact on the microbiome composition. The MagMAX™ CORE Nucleic Acid Purification Kit (ThermoFisher Scientific) was chosen based on the manufacturer data indicating bacterial DNA extraction from both swab and

tissue samples from animal tissues using magnetic bead technology, as opposed to spin column method used in the other two kits.

Haemolymph and swabs of gill, gut and muscle collected in ASW were processed for nucleic acid extraction using the QIAamp® DNA microbiome kit according to manufacturer’s directions. According to the manufacturer this method selectively enriches bacterial nucleic acid samples by differential lysis of host cells followed by degradation of host DNA with benzonase. Benzonase is a genetically engineered endonuclease which acts on all forms of DNA and RNA, but has no impact on intact bacterial cells (Wen et al., 2016). Bacterial cell disruption is then achieved by mechanical and chemical methods which involve bead-beating in lysis-buffer containing guanidinium thiocyanate. A spin column was used to adsorb nucleic acids onto a silica membrane and bound nucleic acids were then washed twice and eluted in the buffer recommended by the manufacturer.

Gill and gut-swab samples in ML1 buffer were processed according to the directions for the E.Z.N.A.® Mollusc DNA kit. According to the manufacturer, digestion was achieved by incubating overnight at 37°C with 25 µl of Proteinase K solution (Proteinase K 20 mg ml- 1; E.Z.N.A.® Mollusc DNA kit). The CTAB in ML1 buffer selectively binds polysaccharides and proteins that remain after digestion with Proteinase K. These were then removed by extraction with chloroform:isoamylalcohol (24:1) (Winnepenninckx et al., 1993). The DNA was further purified with the spin columns provided.

The gill and gut tissue samples were processed, and nucleic acids were extracted by methods A, B and C, respectively. In method A, a 30 mg sample of gill or gut was digested separately with Proteinase K (MagMAX™ CORE Nucleic Acid Purification Kit) at 55°C for 2 h in a dry heat block. Nucleic acid extraction was then conducted using a MagMAX™ CORE Nucleic Acid Purification Kit according to the manufacturer’s instructions with a MagMAX™ Express 96 magnetic particle processor (Applied Biosystems, USA) and the MagMAX_Core_50µl program.

With method B, a tissue homogenate of gill or gut (30 mg) was prepared by bead- beating in 350 µl ML1 buffer (E.Z.N.A.® Mollusc DNA kit) with 0.4 g of 0.1 mm zirconia- silica beads (Biospec Products, Daintree Scientific) using a TissueLyser II (Qiagen) at a frequency of 30 Hz for 5 min. The samples were enzymatically digested with 25 µl of Proteinase K solution (E.Z.N.A.® Mollusc DNA kit) by overnight incubation at 37°C. DNA

extraction was conducted using E.Z.N.A.® Mollusc DNA kit, as described above. This method was also used to extract DNA from an additional 30 mg gill tissue sample that had been frozen at -80°C for 30 days.

In method C, a tissue suspension was prepared from 120 mg samples of gill with 10 ml of ASW using a stomaching machine (MiniMix, Interscience, France). Each sample was homogenized for 1 min at maximum speed. A coarsely clarified homogenate was obtained by collecting the material filtered through the inner mesh (porosity <250 µm) of the stomaching bag (BagPage®, Interscience, France). Two samples of each homogenate (1 ml each) were centrifuged at 100 g for 1 min in a microcentrifuge (Heraeus® Biofuge® Pico, Thermo Electron Corporation). The supernatants were harvested separately, and one was subjected to nucleic acid extraction using the QIAamp® DNA microbiome kit. The other supernatant (200 µl) was processed for nucleic acid extraction using the MagMAX™ CORE Nucleic Acid Purification Kit. The remainder of the original clarified homogenate was initially filtered through a 5 µm cellulose-acetate syringe filter (Minisart® NML, Sartorius) followed by a 0.8 µm syringe filters (Minisart® NML, Sartorius). Nucleic acids were then extracted from the filtrate using the MagMAX™ CORE Nucleic Acid Purification Kit.

4.3.4 Spiking with internal positive control (IPC) DNA

Each sample was spiked with 20,000 copies of VetMAX™ Xeno™ Internal Positive Control (IPC) DNA (Thermofisher Scientific) per extraction volume, during nucleic acid extraction. This provided a control for PCR inhibitors and was used to evaluate the efficiency of each nucleic acid extraction method. The IPC DNA was incorporated into the lysis/binding solution in the MagMAX™ CORE Nucleic Acid Purification Kit (according to the instructions of the manufacturer), whereas the same quantity of IPC DNA was added to the nucleic acid/DNA solution just before spin column extraction for the other two nucleic acid extraction methods.

4.3.5 Bacterial DNA quantification

A real-time PCR assay that targets a highly conserved region of the 16S rRNA gene of domain Bacteria (position 331 to 797; Escherichia coli numbering system) was used to quantify bacterial DNA purified through each sampling and extraction method. The PCR assay described by Nadkarni et al. (2002) was used according to an optimized protocol after duplexing with the VetMAX™ Xeno™ IPC Assay (ThermoFisher Scientific). The 16S rRNA target was detected by a FAM-labelled probe (Nadak_rRNA probe; Table 2.1) and the

IPC assay used a VIC-labelled probe. Each nucleic acid extract was tested in duplicate 25 µl PCR reactions, using a 7500 Fast Real-time PCR system (Applied Biosystems Foster City, CA). Each reaction contained 12.5 µl of 2× TaqMan™ Fast Universal PCR Master Mix (Thermofisher Scientific), 0.9 µmol l-1 of the primers Nadak_rRNAF and Nadak_rRNAR, 0.1 µmol l-1 of FAM-labelled probe, 1 µl of 25× Xeno™ IPC assay, 5 µl of sample template DNA and 1.75 µl of sterile, nuclease free water. The PCR programme was as follows: initial denaturation of 95°C for 10 min, subsequent 40 cycles of denaturation at 95°C for 15 s, and annealing at 60°C for 1 min.

PCR runs were analysed by the standard curve quantitation method, using 7500 software v.2.3 (Applied Biosystems). Quantitative standards, positive PCR controls and no template controls were included in each run. The PCR run was considered valid when there was no amplification of negative controls; amplification of both replicates of the positive control with a cycle threshold (Ct) within the range of the standard curve; and standard curve with r2 > 0.99 and efficiency of 90-110%. Samples exhibiting an exponential increase in the fluorescence signal in both replicates with a Ct value >15 and <35, were considered for quantification of bacterial DNA (Applied Biosystems). Bacterial DNA concentration in the

PCR reaction was calculated by comparing the Ct value to the standard curve. The bacterial DNA in each nucleic acid extract was expressed as the number of bacterial 16S rRNA gene copies per sample.

The quantitation standards and positive control DNA were prepared as described in Section 2.5.4.2. Initially, the number of 16S rRNA gene copies was estimated after determining the number of colony forming units (CFU) ml-1 in an E. coli broth culture at stationary phase. Quantification of CFU was done by the spread plate method using a decimal serial dilution of the E. coli broth culture, with each dilution on three replicates of sterile, nutrient agar plates (Buck and Cleverdon, 1960; Kaper et al., 1978). Nucleic acids were extracted from the broth culture using the MagMAX™ CORE Nucleic Acid Purification Kit. The DNA concentration of the neat extract was measured using a Nanodrop™ spectrophotometer (ND-1000, ThermoFisher Scientific, USA). Standard curves for the qPCR assay were generated by using a 5-step decimal dilution series of extracted E. coli genomic DNA starting with 6.61 × 107 copies of 16S rRNA gene copies in the PCR reaction mixture to 6.61 × 103 copies.

4.3.6 Analysis of bacterial DNA yields

The number of 16S rRNA gene copies per mg of tissue or per swab sample showed a skewed distribution. Thus, the data were logarithmic-transformed on the log10 scale for normal distribution. The bacterial DNA yield was then compared between (1) different types of tissues, (2) different sampling methods used for the same tissue, and (3) different nucleic acid extraction methods, using a generalized linear mixed model (GLMM, SPSS v.22; IBM SPSS Cooperation, Somers, NY, USA). Tissue type, sampling method and the nucleic acid extraction method were used as fixed effects in the model, significant interactions were retained in the model and oyster identification number was a random effect. Post-hoc pairwise mean comparisons were made using the least significant difference method. Paired t-tests (SPSS v.22) were used to compare the bacterial DNA yield for comparisons that could not be made with the GLMM due to the experimental design. Accordingly, bacterial DNA yield from fresh and frozen gill tissue using EZNA extraction, and bacterial DNA yield from gill tissue extracts with and without physical pre-enrichment methods (Method C), were compared using paired t-tests. Significance was set at P < 0.05. The results of bacterial DNA yields were presented as geometric mean of bacterial gene copy numbers and their corresponding 95% confidence intervals.

4.3.7 Analysis of extraction efficiency and PCR inhibition

Extraction efficiency and PCR inhibition for nucleic acid extracts was assessed using the VetMAX™ Xeno™ IPC assay. The IPC DNA was added to all samples at a concentration of 20,000 copies per extraction volume immediately prior to nucleic acid purification. Depending on the elution volume recommended for each nucleic acid extraction method, this resulted in 2000, 1000, 1111 copies of IPC DNA template per PCR assay, for QIAamp® kit extracts, E.Z.N.A. ® kit extracts and MagMAX™ CORE kit extracts, respectively. Additionally, a calibrator sample was prepared by directly adding 5000 copies of IPC DNA to duplicate PCR reactions on each plate. Nucleic acid extraction efficiency and

PCR inhibition was indicated by the Ct values for the Xeno DNA assay. For samples in which the Ct for the IPC was >32 cycles, this evidence of inhibition or inefficient extraction prompted retesting after 2-fold dilution of the nucleic acids in nuclease-free water. Detection efficiency (DE) for IPC DNA in each sample was determined by comparative quantitation of - (C sample – C calibrator) Ct values for IPC DNA, according to the formula, DE = 2 t t . Detection efficiency provided a measure of PCR inhibition as well as the relative proportion of control nucleic acid that was recovered by the extraction method to purify nucleic acids. Detection

efficiency of the IPC DNA for different nucleic acid extraction methods were compared using a non-parametric Kruskal-Wallis test (SPSS v.22).

4.3.8 Microbiome analysis by high throughput 16S rRNA gene sequencing

A subset of the oyster samples was selected randomly, and nucleic acid extracts obtained from haemolymph, muscle-swab samples, gill swab samples (extracted by both QIAamp® method and E.Z.N.A.® method), gill tissue samples (extracted by E.Z.N.A.® method from both fresh and frozen samples) and from gut swab samples (extracted by E.Z.N.A.® method) were used in the microbiome analysis (n = 33). The bacterial community composition in each extract was identified by high-throughput sequencing of the hypervariable V1-V3 region of the 16S rRNA gene. The region was selected based on its maximum nucleotide heterogeneity to provide maximum discriminatory power of bacteria to the genus level (Chakravorty et al., 2007). Sequencing was performed through the Australian Genome Research Facility (AGRF). Initially, PCR amplicons were generated at AGRF using primers 27F and 519 R (Handl et al., 2011) followed by sequencing on the Illumina MiSeq System with 300-bp paired end chemistry.

The quality of the raw sequence data was assessed with FASTQC (Andrews, 2010). Paired-end reads were merged, and quality filtered using USEARCH v.10.0 (Edgar, 2010; Edgar and Flyvbjerg, 2015). Paired-end reads were merged using the default parameters, and primers were trimmed from the merged reads. The reads were quality filtered using a maximum expected error of 1.0 erroneous base per sequenced region (Edgar and Flyvbjerg, 2015). De novo OTUs were generated with USEARCH, as implemented in Quantitative Insights into Microbial Ecology (QIIME) v.2.7.10 (Caporaso et al., 2010), based on a minimum sequence similarity cut-off of 97%. De novo and reference-based chimera filtering was performed against the RDP Gold database and an operational taxonomic unit (OTU) table was generated. OTUs with an abundance of less than 10 reads per sample were filtered. The absolute abundance of each identified OTU was obtained for each sample, based on the number of sequencing reads. Taxonomy assignment was performed using the RDP classifier in QIIME (Liu and Wong, 2013), against the GreenGenes v.13_8 OTU database (McDonald et al., 2012a). The DNA sequences were classified from phylum to genus level. A representative set of sequences for each OTU was generated and aligned with PyNAST (Caporaso et al., 2010). Gaps and variable positions were filtered from the alignment with

Lane masking (Lane, 1991) and a phylogenetic tree was constructed using FastTree (Price et al., 2009).

The relative abundance of bacterial phyla was graphically represented using 100% stacked 2-D column graphs in Microsoft Excel 365. Further, diversity analyses were performed with QIIME. Alpha diversity analyses were performed on a rarefied dataset, with minimum and maximum rarefaction depths of 10 and 9,000 reads per sample, respectively, in steps of 1,000 reads with 10 iterations per step. Alpha diversity was assessed using the following parameters: number of observed OTUs, Shannon’s diversity index and Simpson’s index and was compared between selected groups of nucleic acid extracts by means of a non- parametric t-test with 10,000 Monte Carlo permutations and Bonferroni-corrected P-values.

Beta diversity analyses were performed on a dataset rarefied to 5,000 reads per sample, to assess the bacterial community differences in: (1) different types of tissues, (2) different sampling methods for the same tissue, (3) nucleic acids obtained from different nucleic acid extraction methods and (4) fresh and frozen samples of the same tissue. To visualize differences in bacterial community structure Bray-Curtis dissimilarity index was calculated and PCoA plots were generated using Paleontological Statistics (PAST) software v.3.16 (Hammer et al., 2001). The beta diversity was statistically analysed between the groups of interest using permutational multivariate analysis of variance (PERMANOVA) (Anderson, 2001), to test the null hypothesis of no difference amongst a priori defined groups using PERMANOVA for PRIMER (v.7, PRIMER-E, Quest Research Ltd., Auckland, New- Zealand). The similarity matrices based on Bray-Curtis distance were constructed with 999 permutations, using the square root transformed relative OTU abundance. Post-hoc pairwise comparisons were used to determine significant differences in bacterial community composition across different tissue types, sampling methods, nucleic acid extraction methods and between fresh and frozen tissues.

Bray-Curtis dissimilarity between bacterial communities in the same tissue type of different oysters was used to calculate average individual dissimilarity within a tissue type. The average individual dissimilarity between bacterial communities in one tissue type was then statistically compared with average dissimilarity of bacterial communities between two tissue types of the same oysters, using Kruskal-Wallis test.

4.4 Results

4.4.1 Bacterial DNA yield

The bacterial DNA yield varied in different tissues (P < 0.05) and the highest concentration of bacterial DNA was obtained from adductor muscle with 1.43 × 107 bacterial rRNA gene copies per sample (95% CI: 1.07 × 107 - 1.91 × 107). The yield of bacterial DNA was 16-fold lower in gill tissues and nearly 100-fold lower in haemolymph (Table 4.2). The method of sampling the tissues also affected the bacterial DNA yield (p < 0.05). More bacterial DNA was obtained from swabbing compared to homogenised tissue (Table 4.2; p < 0.05). Across sample types, the QIAamp® DNA microbiome kit (Qiagen, Germany) provided the highest bacterial DNA yield of 2.62 × 106 16S rRNA gene copies/sample (95% CI: 2.57 × 105 - 2.68 × 107, P < 0.05), out of the three, commercial nucleic acid extraction kits tested in this study. Of the two physical methods tested for bacterial enrichment, nucleic acid extracts from the supernatant of centrifuged samples showed no difference in the bacterial DNA yield compared to direct gill tissue extracts while filtrates of filtered samples had a lower bacterial DNA yield compared to direct gill tissue extracts (p < 0.05).

100

Table 4.2 Mean bacterial DNA yields (geometric mean and their corresponding 95% confidence intervals) for Crassostrea gigas tissues, sampling methods used for the same tissue, and nucleic acid extraction method

Type/Method Samples Mean bacterial DNA yields Confidence Interval (CI) tested (n) (bacterial 16S rRNA gene copies/sample) Tissue type: Haemolymph 10 1.51 × 105 2.44 × 104 - 9.34 × 105 Muscle 10 1.43 × 107 1.07 × 107 - 1.91 × 107 Gill 40 8.84 × 105 1.12 × 105 - 6.93 × 106 Gut 22 4.37 × 105 9.17 × 103 - 2.09 × 107

Sampling method: Tissue-swabs 44 5.49 × 106 1.17 × 106 - 2.58 × 107 4 3 5 Tissue samples 28 7.79 × 10 7.17 × 10 - 8.47 × 10 Nucleic acid extraction method: QIAamp® DNA microbiome kit 30 2.62 × 106 2.57 × 105 - 2.68 × 107 ® 5 4 7 E.Z.N.A. Mollusc DNA kit 38 6.33 × 10 2.76 × 10 - 1.46 × 10 MagMAX™ CORE nucleic acid 20 9.20 × 104 2.98 × 104 - 2.84 × 105 purification kit Storage method: Fresh tissues 10 3.81 × 105 8.75 × 104 - 1.66 × 106

Frozen tissues 10 2.91 × 104 1.19 × 104 - 7.13 ×104

101

4.4.2 PCR inhibition

All gut samples other than samples obtained with the E.Z.N.A.® Mollusc DNA kit, demonstrated PCR inhibition when tested without dilution. Thus, the gut nucleic acid extracts from other extraction kits were diluted 2-fold prior to addition as PCR template. This dilution step improved the detection efficiency of the IPC in QIAamp® kit extracts (Table 4.3). However, the IPC DNA could not be detected with MagMAX™ CORE Nucleic Acid Purification Kit extracts, even after dilution. Considering all types of nucleic acid extracts from gut, the highest detection efficiency of IPC DNA was observed with E.Z.N.A.® extracts of gut tissues compared to both E.Z.N.A.® kit extracts of gut swabs and QIAamp® kit extracts of gut swabs (p < 0.05; Table 4.3). Amplification of IPC was not inhibited in all other tissue and swab samples extracted with all three types of kits. Considering the gill tissue, the highest detection efficiency was seen with QIAamp® kit extracts compared to E.Z.N.A.® kit extracts (p < 0.05; Table 4.3).

102

Table 4.3 Comparison of the detection efficiency for the internal positive control (IPC) in nucleic acid extracts obtained from different tissues, sampling techniques, and nucleic acid extraction methods. Tissue type Nucleic acid purification method n* Detection efficiency (median (range)) Gut QIAamp® kit extracts undiluted 0/10 NA diluted 7/10 0.04 (0.03-0.08)A MagMAX™ kit extracts undiluted 0/10 NA diluted 0/10 NA E.Z.N.A.® kit extracts swab 10/10 0.06 (0.03-0.10)A tissue homogenates 10/10 0.35 (0.07-0.58)B

Gill QIAamp® kit extracts 9/10 0.19 (0.14-0.24)A MagMAX™ kit extracts 10/10 0.03 (0.01-0.04)B E.Z.N.A.® kit extracts swabs 10/10 0.14 (0.07-0.29)AC tissue homogenates 10/10 0.10 (0.05-0.16)BC

Haemolymph QIAamp® kit extracts 10/10 0.16 (0.08-0.18)

Muscle QIAamp® kit extracts 10/10 0.17 (0.12-0.21) *Number of nucleic acid extracts (n) showing detectable amplification of IPC, out of the total number of extracts. NA, not applicable. Superscript A, B or C next to data in the same tissue indicates that significant differences exist among different nucleic acid extracts (p < 0.05).

4.4.3 Bacteria in nucleic acid extracts

Targeting the hypervariable V1-V3 region of the 16S rRNA gene, a total of 2,994,290 paired-end raw reads were obtained initially from the samples analysed (n = 33), leaving 1,293,823 reads after quality filtering. Rarefaction curves showed saturation for most of the samples, indicative of a good coverage of diversity (data not shown).

4.4.4 Bacterial diversity in different tissue types

The alpha diversity of the bacterial communities in different tissue types was not significantly different (p > 0.05 for all parameters tested). However, each tissue type harbored

a unique bacterial community composition (PERMANOVA, pseudo-F(df=2,10) = 3.49, p < 0.05) except for the microbiota in gill and adductor muscle tissues. Post-hoc pairwise tests for

103

PERMANOVA revealed significant differences between the microbiota of haemolymph and gut with that of gill (gill vs haemolymph: t = 2.26, P = 0.02; gill vs gut: t = 1.60, p = 0.01; Figure 4.1A and 4.1B) and muscle (muscle vs haemolymph: t =1.85, P = 0.02). Microbiota of gill and muscle were not distinct (t = 1.03; P = 0.41; Figure 4.1A). Moreover, the mean Bray- Curtis dissimilarity between tissue types from the same oysters was greater than the mean Bray-Curtis dissimilarity between individuals for the same tissue type, except for gill and muscle microbiota (Table 4.4). There were 47 phyla of bacteria identified in this population of oysters. In general, phylum Proteobacteria (69%) dominated in all tissue types except gut tissue where phylum Fusobacteria (39%) was most abundant (Figure 4.2A and 4.2B). Phylum Spirochaetes (26%) and phylum Bacteroidetes (16%) were among the next most abundant phyla in tissues, considering the nucleic acid extracts which demonstrated a greater bacterial diversity. The families Chromatiaceae (27.4%) and Rhodobacteraceae (21.6%) dominated the haemolymph at family level, whilst families Rhodobacteraceae (28.6%) and Rhodospirillaceae (23.6%) accounted for most of the OTUs observed in muscle tissue. Meanwhile, the gill tissue extracts were dominated by family Brachyspiraceae (20%) and the gut tissues were mainly comprised of family Fusobacteriaceae (37.8%).

104

Figure 4.1 Tissue compartmentalization of Crassostrea gigas microbiota. Principal coordinate plots generated for tissue microbiota, based on Bray-Curtis dissimilarity index. (A) Samples extracted with the QIAamp® DNA microbiome kit showed clustering of the microbiota of gill (red) with that of muscle (green) and a distinct tissue compartmentalization with haemolymph microbiota (blue) (p < 0.05). (B) Samples extracted with the E.Z.N.A.® Mollusc DNA kit showed compartmentalization of gill (red) and gut (black) microbiota (p < 0.05).

105

Table 4.4 Bray-Curtis dissimilarity (mean pairwise Bray-Curtis distance ± SD) between bacterial communities in the same tissue type for different individual oysters (within tissue type) and between different tissue types from the same individual oyster (between tissues). Mean Bray-Curtis Dissimilarity Tissue type Within tissue type Between tissues Hemolymph 0.89 ± 0.04 Muscle 0.36 ± 0.09 Gill 0.41 ± 0.08 Gut 0.56 ± 0.10 Hemolymph and muscle 0.97 ± 0.01 Hemolymph and gill 0.99 ± 0.01 Muscle and gill 0.36 ± 0.11* Gill and gut 0.79 ± 0.04 *The Bray-Curtis dissimilarity between bacterial communities of different tissues from the same oysters exceeded the inter-individual Bray-Curtis dissimilarity for each type of tissue (p < 0.05), except for the gill and muscle tissues (P > 0.05).

106

0.9

0.8 Proteobacteria

Actinobacteria 0.7 OD1 0.6 Chloroflexi

Firmicutes 0.5 Other

Spirochaetes

Relative Relative (%) abundance 0.4 Planctomycetes 0.3 Bacteroidetes Verrucomicrobia 0.2 Fusobacteria 0.1

0 1 2 3 4 5 1 2 3 1 2 3 4 5 Gill Haemolymph Muscle Oyster ID

0.9

0.8 Proteobacteria Fusobacteria 0.7 Bacteroidetes Tenericutes 0.6 Spirochaetes

Planctomycetes 0.5 GN02 Firmicutes

0.4 Thermotogae Relative (%) abundance Chloroflexi ZB3 0.3 Actinobacteria Verrucomicrobia 0.2 Other

0.1

0 1 2 3 4 5 1 2 3 4 5 Gut Gill Oyster ID

107

Figure 4.2 Taxa plots summarizing the relative abundance of bacterial phyla in nucleic acid extracts of: (A) gill, haemolymph and muscle tissues using QIAamp® DNA microbiome kit; and (B) gill and gut using E.Z.N.A.® Mollusc DNA kit, from the same oysters. The different colours stand for different phyla. Genera with a relative abundance of less than 1% and unclassified bacteria were grouped in the ‘other’ category.

108

4.4.5 Bacterial diversity with different extraction and sampling methods

Despite the higher bacterial DNA yields obtained with the QIAamp® kit, the alpha diversity was lower (observed OTUs: 184.9 ± 32.9) compared to the E.Z.N.A.® Mollusc DNA kit extracts (observed OTUs: 360.2 ± 95.6) obtained from the same tissues of the same oysters (p < 0.05). The bacterial community composition was also different when inferred from the two types of extracts (PERMANOVA, pseudo-F(df=1,8) = 9.59, p < 0.05); Figure 4.3A). Phylum Proteobacteria dominated in nucleic acid extracts from gill swabs obtained by both methods, but the E.Z.N.A.® kit also generated DNA sequences from phyla Bacteroidetes, Spirochaetes and Firmicutes, which were not seen or were very-low abundant in the QIAamp® kit extracts (Figure 4.3B). Interestingly, the method of sampling (swabs vs. tissue samples) did not affect the bacterial community composition in extracts obtained from the gill tissue, both in terms of alpha diversity (p > 0.05 for all parameters tested) and beta diversity (PERMANOVA, pseudo-F(df=1,8) = 0.8, p = 0.75); Figure 4.4A). The bacterial community composition in extracts from tissue samples from the gut could not be analysed in this study.

4.4.6 Effects of frozen storage on bacterial DNA

Storage of gill tissues at -80°C resulted in more than a 10-fold reduction of bacterial DNA yield (p < 0.05; Table 4.2). However, the bacterial community composition in tissue samples demonstrated no difference with regard to alpha diversity (P > 0.05 for all parameters tested) as well as with beta diversity (PERMANOVA, pseudo-F(df=1,8) = 0.69, P = 0.86); Figure 4.4B), after freezing.

109

0.9

Proteobacteria 0.8 Fusobacteria Bacteroidetes 0.7 Tenericutes Spirochaetes 0.6 Planctomycetes GN02 0.5 Firmicutes ZB3 0.4 Thermotogae OD1 Chloroflexi

Relative abundance (%) abundance Relative 0.3 Actinobacteria Other 0.2

0.1

0 1 2 3 4 5 1 2 3 4 5

EZNA QiaAmp

110

Figure 4.3 Bacterial diversity with different extraction methods. (A) Difference in bacterial community composition. GL6-GL10 and GL1-GL5 correspond to gill swab samples extracted from E.Z.N.A.® mollusc DNA kit and QIAamp® DNA microbiome kit, respectively. (B) Relative abundance of bacterial phyla in different nucleic acid extracts

(E.Z.N.A.® and QIAamp®) from gill swab samples of the same oysters. The different colours stand for different phyla. Genera with a relative abundance of less than 1% and unclassified bacteria were grouped in the ‘other’ category.

111

Figure 4.4 Effects of sampling and storage methods on bacterial community composition.

Principal coordinate plots based on Bray-Curtis dissimilarity index showing clustering of microbiota from: (A) swabs (S) and tissue samples (T) from gill tissue of the same oysters (1

– 5). (B) Fresh gill tissue (S) and gill tissue samples stored at -80°C for 30 days (Z), obtained from the same oysters (1-5).

112

4.5 Discussion

This study draws attention to the importance of selecting the appropriate tissue type, sampling method, nucleic acid extraction method and storage, for studies of the oyster microbiome. Each of these had profound effects on the results of downstream molecular studies either in terms of bacterial community composition or bacterial DNA yield, or both. Although there was no difference between the alpha diversity of tissue types assessed with the same nucleic acid purification method, our results clearly indicated a tissue compartmentalization of microbial communities in Pacific oysters. A distinct bacterial community composition was observed in haemolymph compared to the microbiota of gill and muscle. Moreover, the gill microbiota was distinct from gut microbiota extracted using the same nucleic acid extraction method. The similarity between gill and muscle microbiota may reflect transient bacteria from pallial fluid. On the other hand, the gill and muscle microbiota may be more stable in the face of environmental influence, compared to gut and haemolymph microbiota. In either scenario, gill and muscle microbiota seem to respond differently to the environment as opposed to haemolymph and gut microbiota. Consideration of these tissue- specific differences is required to better understand the polymicrobial pathogenesis of Pacific oyster mortality.

Lokmer et al. (2016a) have previously shown the close connection of haemolymph microbiota to the environment of oysters, while numerous other studies provide evidence for the relationship between haemolymph bacterial communities and the host condition (Garnier et al., 2007; Lipp et al., 1976). The latter showed increased loads of vibrios in moribund oysters suffering from summer mortality. It has also been suggested that changes in bacterial community structure in haemolymph could be used as a disease indicator (Lokmer and Wegner, 2015). The microbiota of the digestive system generally consists of transient bacteria (Roterman et al., 2015; Zurel et al., 2011) while the gill microbiota in oysters are relatively stable (Roterman et al., 2015; Zurel et al., 2011). Considering the large surface area of gills which is in direct contact with the surrounding water, there is opportunity for exposure to environmental bacteria in addition to harbouring commensal bacteria. However, according to Wegner et al. (2013) the resident population of gill bacteria is minimally disturbed by the filter-feeding behaviour of the oysters. The resident gill microbiota is intimately bound to the gill tissue (Hernandez-Zarate and Olmos-Soto, 2006) while the filtration that occurs in the gills selectively excludes some bacteria from reaching the gut (Meisterhans et al., 2016). These mechanisms may also contribute to the different microbial

113

communities in the gills and gut. Thus, a study of microbial dysbiosis should consider the use of specific tissues rather than haemolymph or mixed tissue samples. The present study did not compare environmental samples with tissue samples to assess potentially transient bacteria. Samples from the gut lumen would be more suitable if the target is transient contaminant microbiota (e.g., as in public health surveillance of food-borne pathogens). The concept of tissue-specific microbiota has also been shown in studies involving eastern oyster (Arfken et al., 2017; King et al., 2012) and other marine invertebrates such as the Manila clam (Ruditapes philppinarum) (Meisterhans et al., 2016) and corals (Sweet et al., 2010).

The dominance of phylum Proteobacteria in haemolymph, gill and muscle tissues in the present study was comparable to the results of previous conventional bacteriological studies (Colwell and Liston, 1960; Fernandez-Piquer et al., 2012; Olafsen et al., 1993; Prieur et al., 1990) where Vibrio, Pseudomonas, Aeromonas and Enterobacteriaceae dominated in Pacific oyster microflora. This dominance of Proteobacteria was also demonstrated in recent studies using other molecular methods such as fluorescent in situ hybridization (Hernandez- Zarate and Olmos-Soto, 2006), terminal restriction fragment length polymorphism (T-RFLP) (Fernandez-Piquer et al., 2012) and 16S rRNA gene temperature gradient gel electrophoresis (TGGE) method (Trabal et al., 2012). Considering 16S rRNA gene studies, the choice of hypervariable region of 16S rRNA gene on determining the phylogenetic resolution cannot be overlooked (Tremblay et al., 2015; Yang et al., 2016). Studies that involve high throughput sequencing of V1-V2 region (haemolymph) (Lokmer et al., 2016b), V3-V4 region (gill) (Wegner et al., 2013), V3-V5 region (gut) (Trabal Fernández et al., 2014), V6 region (haemolymph and digestive gland) (Vezzulli et al., 2018) of the 16S rRNA gene, also reported the dominance of phylum Proteobacteria. However, at a higher resolution Wegner et al. (2013) demonstrated dominance of genus Sphingomonas (family Sphingomonadaceae) in gill tissues as opposed to family Brachyspiraceae, in the present study. Furthermore, Vezzulli et al. (2018) report genus Pseudoalteromonas (family Pseudoalteromonadaceae) to be dominating the haemolymph microbiota while families Chromatiaceae and Rhodobacteraceae dominated haemolymph microbiota in the present study. It should also be noted that the above studies have used different nucleic acid extraction kits compared to the present study. Being strict anaerobes, the dominance of phylum Fusobacteria in oyster gut shown in the present study, had not been identified in conventional studies where bacterial isolation was limited to aerobic cultures.

114

Bacterial DNA yields varied in different tissues of the Pacific oysters used in this study, and after accounting for the impact of sample preparation methods, was highest in adductor muscle. Higher bacterial DNA yield from gut compared to gill tissues of the same oysters was demonstrated based on uninhibited measurement using E.Z.N.A.® kit. However, higher bacterial DNA yields in QIAamp® kit extracts coupled with the PCR inhibition in gut tissue extracts may have contributed to the overall highest bacterial DNA yield from muscle tissue. This contradicted previous studies of bacterial counts from aerobic, heterotrophic bacterial cultures from a range of different tissues which included muscle, gill and gut from the Pacific oyster (Kueh and Chan, 1985) and molecular quantification of bacterial 16S rRNA gene (Wang et al., 2014). Both types of studies showed that the number of bacteria was higher in gut tissues compared to gill and adductor muscle.

Extraction of bacterial DNA from gut tissues was identified as a challenge in this study, unless suitable procedures were carried out to eliminate PCR inhibition. The CTAB method used by the E.Z.N.A.® Mollusc DNA kit for removal of mucopolysaccharides, was identified as a critical step in bacterial DNA extraction from oyster gut tissues. The use of IPC DNA as an internal control helped identify PCR inhibition in gut tissue extracts other than with E.Z.N.A.® kit. Oyster tissue extracts have exhibited a high level of PCR inhibition which was not effectively eliminated by various resins and commercial spin columns used for purification of DNA (Abolmaaty et al., 2007). Nucleic acid extraction methods employing CTAB, which specifically binds proteins and polysaccharides have been successfully used in extracting inhibitor free genomic DNA from bivalves, gastropods, cephalopods and other invertebrates belonging to phyla such as Echinodermata and Platyhelminthes (Winnepenninckx et al., 1993; Yap and Thompson, 1987). Abolmaaty et al. (2007) have used activated charcoal in oyster tissue homogenates with low bacterial load to effectively adsorb the PCR inhibitors thereby increasing the sensitivity of downstream qPCR assays. Moreover, the inclusion of IPC DNA as an internal control has been practised in previous studies which involved bacterial DNA extraction from soil, food etc. (Al-Soud and Rådström, 1998; Dineen et al., 2010; Fricker et al., 2007).

Considering nucleic acid extraction from haemolymph, the present study demonstrated a lower bacterial DNA yield compared to muscle, gill and gut tissues, supporting the previous findings of low concentrations of bacteria in haemolymph in healthy oysters (Garnier et al., 2007). Moreover, I observed higher bacterial DNA yields through swabbing compared to

115

tissue samples, suggesting a more selective approach to obtain bacterial DNA relative to oyster genomic DNA. This study also investigated the impact of physical methods in enriching bacterial DNA. Low speed centrifugation method which removed host cells by sedimentation, and filtration method which removed host cells by retention, prior to bacterial lysis and nucleic acid extraction, were tested as physical bacterial enrichment methods. However, we could demonstrate that physical methods for bacterial enrichment did not increase the bacterial DNA yield. Considering all sampling approaches tested in this study, swabbing of tissues appear to be the most suitable method in obtaining higher bacterial DNA yields.

Although the QIAamp® kit extracts achieved a higher bacterial DNA yield in this study, the alpha diversity of the bacterial community was lower. In contrast, the extracts from E.Z.N.A.® kit preserved a more diverse bacterial community structure in addition to eliminating PCR inhibition in gut tissues. The amplification of the bacterial 16S gene from neat gut extracts was successful only with the E.Z.N.A.® kits, in this study. In summary, this study demonstrated the suitability of the E.Z.N.A.® Mollusc DNA kit (CTAB method) in effectively extracting diverse bacterial DNA from oyster tissues. Further, it proved to be a better option for purifying bacterial DNA from gut (digestive glands) for PCR-related applications, over the other two commercial nucleic acid extraction methods tested in this study. Both the presence of PCR inhibitors and bacterial cells that are resistant to lysis, can have an impact on the DNA extraction efficiency. Alpha diversity in bacterial community structure have previously been reported to be affected by the DNA extraction methodology based on the presence/absence of a bead-beating step and quality of resulting DNA (Guo and Zhang, 2013; Henderson et al., 2013; Wagner-Mackenzie et al., 2015).

The storage of tissue at -80°C for 30 days resulted in a reduction of bacterial DNA yield, but surprisingly, the diversity of the bacterial community was preserved, both in terms of alpha and beta diversity. This result may find application in reducing the logistical challenges of microbiome studies. Choo et al. (Choo et al., 2015) have shown the absence of significant alterations in the alpha diversity and beta diversity of faecal microbiome of humans after storage of faecal samples at -80°C and recommended freezing as a storage method for faecal microbiome analyses. Rapid preservation of tissues is critical for microbiome studies. Fernandez-Piquer et al. (2012) recognised a significant difference in bacterial communities in freshly harvested live oysters compared to oysters stored live at 4°C

116

for more than 24h where members of the phylum Proteobacteria predominated in fresh oysters while phylum Fusobacteria (Psychrilyobacter spp. in particular) predominated at 4°C. As genus Psychrilyobacter was identified as the dominant bacterial genus in gut tissues in the same study, the dominance of genus Psychrilyobacter in refrigerated oysters may be a result of migration and multiplication of gut bacteria during storage at 4°C. Although freezing causes a decrease in bacterial DNA yield by DNA-shearing (Seutin et al., 1991; Tegelström, 1989), this preservation method has proven convenient to preserve the original bacterial community composition in samples. Characterisation of the microbiome using amplicon sequencing of the bacterial 16S rRNA gene is desirable as a relatively rapid and inexpensive procedure that is inclusive of species which cannot be cultured (Choo et al., 2015; Prakash et al., 2011). Nevertheless, this approach can only yield accurate results if the analysed nucleic acids are purified without biases and accurately represent the bacterial community at the time of sample collection (Choo et al., 2015). Although Lauber et al. (2010) indicate no changes to the bacterial community composition with short-term storage at different temperatures (2 weeks), prolonged storage (2 years) at -80°C showed a reduction of alpha diversity by means of the observed number of OTUs.

4.6 Conclusion

A fit-for-purpose sampling strategy is required for analyses of the oyster microbiome, owing to tissue-specific differences in bacterial communities and potential for biases during nucleic acid extraction. The distinct tissue compartmentalization of oyster microbiota provides insights for future microbiome studies. The CTAB method provided in the E.Z.N.A.® Mollusc DNA kit was preferred over other nucleic acid extraction methods tested in this study as it clearly preserved the original bacterial community structure in tissues while alleviating PCR inhibition in oyster gut tissues. The sampling method (swabbing vs. tissue homogenisation) provided a different bacterial DNA yield but did not impact bacterial community structure. Preservation of samples by freezing was suitable for determining the microbiome, however, it reduced the bacterial DNA yield, which may create a negative impact on downstream molecular analyses.

117

CHAPTER 5

Impact of the laboratory environment on the Pacific oyster (Crassostrea gigas) microbiome

5.1 Abstract

Oysters are usually acclimated to the laboratory environment before an experimental study. However, exposure to novel environmental conditions may affect the oyster and its microbiome. The microbiome has become a focus of recent research on Pacific oyster disease. Any changes induced specifically by the laboratory environment may confound changes in the microbiome attributed to a disease studied in an experiment. The objectives of the present study were to: 1) study the temporal variation in the microbiome of oysters with a common origin and from a common environment and; 2) to assess changes in the Pacific oyster microbiome during acclimation to a laboratory environment with both constant immersion in water and in a simulated tidal environment. Pacific oysters were sourced from a commercial farm in Patonga Creek, NSW. Oysters were sampled from the field on Day 0 and 14 and from the laboratory on Days 3, 7 and 14 after acclimation. Whole tissue homogenates were cultured on marine salt agar-blood and TCBS agar after sampling gill and gut tissues for molecular studies. Total bacteria and Vibrio in tissues were quantified using qPCR and bacterial community composition was estimated by 16S rRNA gene (V1–V3 region) sequencing. Oysters obtained from the same environment two weeks apart had different total cultivable bacterial count (TCBC) and total cultivable Vibrio count (TCVC) (p < 0.05). Irrespective of the laboratory management system (constant immersion and simulated tide) the TCBC and total bacterial DNA in oysters remained similar after laboratory acclimation for 14 days. The TCVC increased with laboratory acclimation (p < 0.05). The Vibrio count was initially low in gill and gut tissues but increased in gut tissue after acclimating to the laboratory for 14 days. While the gill microbiome in oysters from the field demonstrated temporal stability the gut microbiota showed temporal variation. The alpha diversity of both gill and gut microbiota did not differ with laboratory acclimation irrespective of the management system (p > 0.05). Beta diversity of both gill and gut microbiota increased with acclimation to the laboratory environment (p < 0.05). The abundance of the dominant phyla Proteobacteria in both gill and gut microbiota did not change after acclimation, in both management systems (p > 0.05). However, the genus Arcobacter (phylum Proteobacteria) increased in gill and genus Vibrio increased in gut (p < 0.05). The laboratory acclimation

118

reduced the abundance of phylum Cyanobacteria and phylum Tenericutes in the gill and gut. In conclusion, the microbiome composition of Pacific oysters with a common origin and living in a common environment had individual variations. The oyster microbiome changed during acclimation to a controlled laboratory environment. There was no impact of maintaining oysters in constant immersion in water as opposed to maintaining in a simulated tidal environment. Although the total quantity of bacteria did not change in the laboratory, the community composition changed with time owing to the changes in the relative abundance of bacterial genera. Changes in the microbiome because of changes in the environment should be considered in future experimental studies and in commercial depuration of oysters.

Keywords: Pacific oyster, Crassostrea gigas, microbiome, laboratory acclimation, constant immersion, simulated tide

5.2 Introduction

The Pacific oyster (Crassostrea gigas) is the most important commercial oyster species globally, with an annual global production of 625,925 tonnes worth US$ 1.3 billion out of a total oyster production of 5.2 million tonnes worth US$ 4.2 billion (FAO, 2014). In the recent past, severe mortality diseases have greatly impacted Pacific oyster production in Europe (Martenot et al., 2011; Renault et al., 2012; Segarra et al., 2010), Australia (Jenkins et al., 2013; Paul-Pont et al., 2014; Whittington et al., 2018) and New Zealand (Keeling et al., 2014). The reference genotype of Ostreid herpes virus-1 (OsHV-1) (Davison et al., 2005; Le Deuff and Renault, 1999) and related genotypes have been identified as the prominent pathogens which caused C. gigas mortalities in France from 1991 to 2008 (Martenot et al., 2011; Renault et al., 2012). However, at the end of spring 2008, widespread mortalities were reported in France which killed billions of young oysters and a genomic variant of OsHV-1, called µVar, was identified from these outbreaks (Renault et al., 2012; Segarra et al., 2010). Since the first detection of Ostreid herpesvirus-1 (OsHV-1) in mass mortality events of Pacific oysters in France (Arzul et al., 2002; Le Deuff and Renault, 1999; Renault et al., 1994b), a considerable amount of research has been carried out to understand the mode of infection, transmission and host susceptibility to OsHV-1 (Alfaro et al., 2019). In this context, experimental OsHV-1 challenge studies conducted in laboratory aquaria are common for Pacific oyster research (Evans et al., 2019; Hick et al., 2016; Paul-Pont et al., 2015; Petton et al., 2019; Schikorski et al., 2011).

119

In laboratory experiments, oysters are usually acclimated to the laboratory environment before the experimental study. Exposure to novel environmental conditions may not only affect the oysters but can also affect the associated microbiota (Lokmer et al., 2016a). Excessive stress can destabilize microbial communities and may shift the microbiome towards pathogenic states (Lokmer and Wegner, 2015; Pita et al., 2013). Recent studies of Pacific oyster disease associated with OsHV-1 have highlighted greater changes in the microbiome of oysters susceptible to the disease compared to their resistant counterparts (de Lorgeril et al., 2018). Bacteria belonging to the genera Vibrio and Arcobacter were associated with some of these changes (de Lorgeril et al., 2018). Interestingly, both Vibrio and Arcobacter (Lokmer and Wegner, 2015) have been associated with oyster mortalities in laboratory infections (Lokmer and Wegner, 2015; Pathirana et al., 2019b) as well as in field infections (Faury et al., 2004; Petton et al., 2015b). As the microbiome has become a focus of Pacific oyster disease research (Green et al., 2018; King et al., 2018a), changes induced specifically by the laboratory environment may mask or confound changes attributed to the microbiome from the disease of interest.

Several environmental factors have been identified as risk factors for disease caused by OsHV-1 such as elevated seawater temperature (de Kantzow et al., 2016; Petton et al., 2013; Renault et al., 2014), alterations in salinity (Fuhrmann et al., 2016; Soletchnik et al., 2007) and pH of water (Fuhrmann et al., 2019). Meanwhile, increased emersion time demonstrated a significant protective effect for oysters farmed in intertidal regions of estuaries (Paul-Pont et al., 2013b; Whittington et al., 2015b). Laboratory studies have also been conducted to further study this protective effect for OsHV-1 (Evans et al., 2019). Apart from direct effects of environmental factors on OsHV-1, the outcome of disease can also be influenced by the response of the oyster microbiome to changes in the environment (Pathirana et al., 2019b).

It is important to differentiate/understand the changes in the Pacific oyster microbiome that may arise in a disease as opposed to the changes that may occur, simply by maintaining oysters in a laboratory environment. Data regarding the optimal periods to acclimate bivalve species prior to laboratory experimentation are lacking (Thompson et al., 2012). There has been no standard period for acclimation of Pacific oysters in laboratory- based experiments with a range of days reported including: 2d (Jo et al., 2008), 3d (Pathirana et al., 2019b), 7d (Boutet et al., 2004; de Kantzow et al., 2019a; Zhang and Li, 2006), 9d

120

(Petton et al., 2019) and up to 10 days (Evans et al., 2019; Medeiros et al., 2008). Acclimation is important to prevent any stress-related changes due to transportation from their original habitat and allow adjustment to the new environment to focus on the factor of interest in the experiment (Hedge and Johnston, 2014). Furthermore, Harding et al. (2004) have shown that the physical stress of handling causes changes in several cellular parameters in bivalves including lysosomal stability. Minor mechanical stress such as shaking for 15 min in a plastic container down-regulated the immune functions of oysters by reducing the number of circulating haemocytes and their phagocytic activity (Lacoste et al., 2002). These stress-related changes may alter the course of the disease of interest in an experimental disease challenge.

The normal variability of microbial communities in the absence of any disturbance event is often unreported and thus, the knowledge of baseline microbial community stability as well as post-disturbance dynamics remains limited (Shade et al., 2012). However, Shade et al. (2013) found that temporal variability in community diversity was consistent among microbial communities from similar environments. On the other hand, understanding the temporal stability of oyster microbial communities is vital to understand and differentiate their response to host or environmental disturbances. The purpose of the present study was to gain insight into the influence of the management conditions and acclimation time on the microbiome composition of Pacific oysters in laboratory studies. Specifically, the objectives were to: 1) determine the temporal variation in the microbiome over 2 weeks for a batch of Pacific oysters in a commercial farm environment in NSW, Australia; 2) to measure changes in the microbiome of this batch of oysters over the same time frame during acclimation to a laboratory environment with constant immersion in water and in a simulated tidal environment.

5.3 Materials and Method

5.3.1 Oysters

Hatchery produced, single seed, triploid Pacific oysters (Batch SPL16A; Shellfish Culture Ltd., Tasmania; n = 224) were used in this study. They were grown under commercial farming conditions using intertidal baskets in Patonga Creek, Hawkesbury River, NSW, until mid-autumn of 2017. This waterway had a sandy bottom with sea grass, low turbid water and a water temperature ranging from 14-26°C (Whittington et al., 2019). For the purpose of this experiment oysters were recruited at 5 months of age (20–40mm shell

121

length) and were transported to a physical containment level 2 aquatic animal facility at the University of Sydney, Camden, NSW.

5.3.2 Experimental design and oyster management

Oysters were randomly allocated to individually aerated, 20 L tanks with artificial seawater (ASW; Red Sea® salt) at 30 ± 1 ppt salinity (n=20 per tank). There were 5 replicate tanks each for the two treatment groups: the first maintained oysters with constant immersion in ASW; the other used a simulated tidal system by removing the standpipes from each tank at 11:00 am and 11:00 pm and replacing the standpipes at 6:00 am and 6:00 pm daily resulting in the oysters being alternately emersed and immersed in ASW ( immersion: 2 × 5 h/day; emersion: 2 x 7 h/day). The 10 tanks were randomly chosen from four separate recirculation systems which comprised 6 tanks and a 250 L-sump. A chiller (HC-300a, Hailea Aquarium chiller) was used to maintain the water temperature of each recirculation system at 22 °C while a biofilter (Fluval 406 canister filter) was also connected.

Water quality of each recirculation system was tested daily using an API ® Marine Saltwater Master Test kit. The pH of water was maintained at 8.2 (range: 8.0−8.2) and total ammonia nitrogen, nitrite and nitrate were maintained <0.25 ppm by exchange of water when required and by adding sodium bicarbonate to the sump. Oysters were held on a perforated plastic insert approximately 10 cm above the bottom of the tank and were fed with a commercial algae diet (Instant Algae® Shellfish Diet 1800; Reed Mariculture). Two millilitres of the algae diet were added to each tank daily and the oysters were allowed to feed for 1 h with the recirculation system disconnected from the treatment tanks to prevent removal of the feed.

Oysters are not considered by the NSW Animal Research Act 1985 nor the NHMRC Australian code for the care and use of animals for scientific purposes, 8th edition (2013). Therefore, approval from the Animal Ethics Committee, University of Sydney was not required for this experimental trial.

5.3.3 Sampling and tissue processing

Oysters were sampled directly from the field at the beginning of the experiment (n=12). Thereafter, two oysters were randomly sampled from each tank on Days 3, 7 and 14 of the acclimation period. Additional oysters were collected directly from the field on Day 14 (n=12). All sampled oysters were maintained on ice until processing.

122

Oysters were shucked and tissues were sampled for culture and molecular assessment of bacterial content. Thirty mg (wet weight) of gill and gut tissues were collected separately (from oysters collected on Days 0, 7 and 14) into sterile 1.5 ml tubes. Separate sterile scalpel- blades were used for each tissue sample from the same oyster to reduce cross contamination between tissues as well as between oysters. These tissue samples were frozen immediately at -80°C until molecular studies were conducted. Tissue homogenates were prepared for direct bacterial culture from the remaining soft tissues after removal of the digestive gland from each oyster (collected on Days 0, 3, 7 and 14). The tissues were transferred separately into stomacher bags, weighed and disrupted by stomaching with 10× (w/v) sterile ASW for 1 min, at maximum (9) speed (MiniMix, Interscience, France). A coarsely clarified homogenate (5 ml) was obtained by collecting the material filtered through the inner mesh (porosity <250 µm) of the stomaching bag (BagPage®, Interscience, France).

5.3.4 Isolation and quantification of cultivable Vibrio and total bacteria

Ten microliters each of each fresh tissue homogenate was spread on separate marine salt agar-blood (MSA-B) plate and thiosulphate-citrate-bile salts-sucrose (TCBS) agar plates, as described in Section 2.5.2. MSA-B and TCBS agar were prepared as described in Section 2.5.1 (Buller, 2014). The inoculated culture plates were incubated at 23°C for 24h (MSA-B) or 48h (TCBS agar), in a refrigerated incubator (LMS Ltd, UK). The total cultivable bacterial count (TCBC) and the total cultivable Vibrio count (TCVC) was calculated as the number of colony forming units (CFU)/g of oyster tissue based on the count of bacterial colonies on MSA-B plates and TCBS plates, respectively.

5.3.5 Quantification of total Vibrio spp. DNA by qPCR

Nucleic acids were purified from gill and gut tissues using the E.Z.N.A.® Mollusc DNA kit (Omega Bio-Tek, USA), as described in Pathirana et al. (2019a). The qPCR assay described by Vezzulli et al. (2012) was adapted for quantification of Vibrio spp. in oysters, as described in Section 2.9. Each PCR plate included a positive control sample, a negative control sample and duplicate reactions prepared with a 10-fold serial dilution ranging from 2.2 × 104 copies of Vibrio 16S rRNA gene as a quantitative standard. The positive control sample and the quantitative standards were prepared as described in Section 2.9.1. Quantification of Vibrio spp. DNA in samples was determined by standard curve quantitation method, using 7500 software v2.3 (Applied Biosystems).

123

The total Vibrio in samples was calculated by dividing the total Vibrio 16S rRNA gene copy number determined by qPCR by 9, as the presumptive number of copies per Vibrio genome (Klappenbach et al., 2001; Thompson et al., 2004b; Vezzulli et al., 2012)]. The total Vibrio count was expressed as number of Vibrio genome equivalents per gram of oyster tissue.

5.3.6 Quantification of total bacteria DNA by qPCR

The qPCR assay based on TaqMan® chemistry described by Nadkarni et al. (2002) was used to quantify total bacteria. This qPCR assay is described in Section 2.10. The quantitative standards and the positive control samples were prepared as described in Sections 2.5.4.2 and 2.10.1. PCR runs were analyzed by the standard curve quantitation method (Applied Biosytems) based on amplification of a 10-fold serial dilution containing between 6.61 × 103 and 6.61 × 107 copies of E. coli rRNA gene. Total bacteria quantity of samples was expressed as the number of bacteria genome equivalents per gram oyster tissue. This was calculated by dividing the total 16S rDNA gene copies by the presumptive number of 16S rDNA copies per Proteobacteria genome [n = 3.5, Kormas (2011); Vezzulli et al. (2012)], based on the species of bacteria Proteobacteria that dominate the oyster microbiota (Lokmer et al., 2016a; Wegner et al., 2013).

5.3.7 Microbiome analysis by high throughput 16S rRNA gene sequencing

Bacterial community compositions of gill and gut tissue samples (n = 30 each) were identified by high-throughput sequencing of the hypervariable V1-V3 region of the 16S rRNA gene. Sequencing was performed at the Ramaciotti Centre for Genomics, University of New South Wales, Australia. PCR amplicons were generated using primers 27F and 519 R (Handl et al., 2011) followed by sequencing on the Illumina MiSeq System with 300-bp paired end chemistry.

Initially, the quality of the raw sequence data was assessed with FASTQC (Andrews, 2010). A sample metadata file was generated in Microsoft Excel software and was validated using the browser-based metadata validation tool, Keemei (Rideout et al., 2016), to check for correct formatting and errors in the metadata file. The sequence data in FASTQ format were analysed using Quantitative Insights into Microbial Ecology Version 2 (QIIME2) Software Suite (2018.11 release). In this process, the sequence reads were initially imported into QIIME2 and demultiplexed. The Divisive Amplicon Denoising Algorithm 2 (DADA2) was used to reduce noise, remove replication and chimera-filter the reads (Callahan et al., 2016).

124

Using the DADA2 pipeline, the sequence reads were truncated at 120 bp based on the quality of sequence data. A feature table was created which contained the number of reads of each unique sequence in each sample in the dataset and mapped feature identifiers to the sequences they represent. The taxonomic analysis was done using a naïve-bayes classifier trained using Greengenes v.13_8 99% operational taxonomic units (OTUs), where the sequences were trimmed to only include the V1-V3 hypervariable region of 16S rRNA gene (McDonald et al., 2012a). The relative abundance of bacteria in each sample was initially visualized using interactive taxonomic bar plots. The relative abundance of bacterial phyla in each sample was then graphically represented using 100% stacked 2-D column graphs (Excel, Microsoft).

5.3.8 Statistical analysis

The TCBC and TCVC for soft tissues and the total bacteria qPCR data in gill and gut tissues were compared separately between oysters. Comparisons were made between field oysters and laboratory-acclimated oysters at different time points during acclimation, and in both management systems (constantly immersed and in a simulated tide). With TCBC, TCVC and total bacteria, the data were log10 transformed for normal distribution and the comparisons were done using separate generalized linear mixed models (GLMM, SPSS Statistics ver. 22; IBM SPSS Cooperation, Somers, NY, USA). The duration of acclimation to the laboratory environment (Day 0, 3, 7 and 14 for TCBC and TCVC; Day 0, 7 and 14 for total bacteria) and the environment (laboratory or field) were used as fixed effects in the model and the significant interactions were retained. Post-hoc pairwise mean comparisons were made using the least significant difference method. The estimated means obtained from the model were transformed to geometric means and their corresponding 95% confidence intervals of CFU/g (for TCBC and TCVC) and bacterial genome equivalents (for total bacteria), respectively. Significance was set at p < 0.05 for all statistical analyses.

Microbial diversity analyses were performed using the q2-diversity plugin of QIIME2 (Bokulich et al., 2018) to compute alpha and beta diversity metrics and generate interactive visualizations with statistical analysis. The number of observed OTUs and Shannon’s diversity index were used as the parameters to assess the alpha diversity of samples while the dissimilarity of bacterial community structure between samples (beta diversity) was visualized by principal coordinate plots (PCoA) based on the two-dimensional Bray-Curtis (BC) dissimilarity index. One-way permutational multivariate analysis of variance

125

(PERMANOVA) was used for statistical analysis of beta diversity. Significance was set at p < 0.05 for all statistical analyses.

5.4 Results

5.4.1 Total bacteria and total Vibrio quantity

A striking feature was the difference between TCBC as well as TCVC in oysters obtained from the field, i.e. from the same intertidal estuarine farm environment on Days 0 and 14 (Table 5.1; p <0.05). While the TCBC demonstrated a 10× increase on Day 14, the TCVC showed a 100× increase. In the laboratory environment, the TCBC in oysters maintained with constant immersion in water decreased after 7 days (Figure 5.1; p <0.05). However, the TCBC was similar to the pre-acclimation value after 14 days in the laboratory (p <0.05). Interestingly, these changes were not seen in the oysters maintained with a simulated tidal environment (Table 5.1; p > 0.05). Overall, the TCBC remained the same after acclimating to the laboratory environment for 14 days, irrespective of the laboratory management system (constantly immersed and simulated tide) of oysters.

The TCVC, on the other hand, increased compared to the field values on Day 0, during the laboratory acclimation in both management systems (Figure 5.1; Table 5.1; p <0.05). Meanwhile, the total bacteria (in gill and gut tissues separately) quantified by qPCR did not differ between oysters collected directly from the field with laboratory-acclimated oysters, under both management systems (Table 5.2). As the Vibrio counts quantified by qPCR in both gill and gut tissues were below the limit of quantification of the qPCR assay with the exception of gut tissue after 14 days of laboratory acclimation (1.82 – 57.74 Vibrio gene copies/mg tissue), it was not employed in the statistical analysis.

126

Table 5.1 Total cultivable bacterial count (TCBC) and total cultivable Vibrio count (TCVC) in the whole soft-tissue mass (except the digestive gland) of oysters (n = 15-20) collected directly from the field at different times and during acclimation to the laboratory environment. Data are presented as the geometric meansa and their corresponding 95% confidence intervals of the number of colony forming units per gram (CFU/g of tissue). aGeometric means were derived from back-transformed model means. *Both TCBC and

Environment Day TCBC (95% CI) TCVC (95% CI)

Field 0 7.66 × 103 (6.12 × 102 – 9.55 × 104) 3.56 × 103 (5.35 × 102 – 2.37 × 104)

14 3.52 × 104 (3.32 × 103 – 3.75 × 105)* 1.08 × 105 (7.26 × 103 – 1.61 × 106)* Laboratory 3 1.99 × 104 (1.96 × 103 – 2.01 × 105) 1.45 × 104 (2.62 × 103 – 7.96 × 104)$ (constantly immersed in water) 7 8.69 × 103 (8.57 × 102 – 8.81 × 104)# 1.33 × 104 (2.40 × 103 – 7.36 × 104)$

14 2.29 ×104 (2.20 × 103 – 2.37 ×105) 2.61 × 104 (4.81 × 103 – 1.41 × 105)$ Laboratory 3 1.59 × 104 (6.08 × 103 – 4.17 × 104) 1.63 × 104 (8.09 × 103 – 3.27 × 104)$ (tidal environment) 7 9.57 × 103 (3.66 × 103 – 2.50 × 104) 1.75 × 104 (8.71 × 103 – 3.52 × 104)$

14 1.24 × 104 (3.92 × 103 – 3.91 × 104) 1.54 × 104 (6.04 × 103 – 3.93 × 104)$ TCVC in field oysters collected on Day 14 were different from that of Day 0 (p <0.05). #TCBC was lower in Day 7 compared to Day 3 and Day 14 in oysters constantly immersed in water (p <0.05). $TCVC was higher on Days 3, 7 and 14 compared to Day 0 in both systems (constant and tidal) (p <0.05).

127

Table 5.2 Total bacterial DNA quantity in gill and gut tissues of oysters in this study

Total bacteria (Geometric mean and 95% CI*) Environment Day Gill tissue Gut tissue Field 0 4.60 × 103 7.43 × 102 - 2.84 × 104 3.67 × 106 2.27 × 105 - 5.94 × 107

14 2.54 × 103 3.31 × 102 - 1.95 × 104 1.49 × 106 9.23 × 104 - 2.42 × 107

Laboratory 7 2.32 × 103 3.74 × 102 - 1.43 × 104 8.45 × 106 7.59 × 105 - 9.40 × 107 (constantly immersed in water) 14 5.05 × 106 2.99 × 103- 1.15 × 105 5.05 × 106 5.85 × 105 - 4.36 × 107

Laboratory 7 8.49 × 105 8.45 × 104- 8.55 × 106 4.10 × 105 3.37 × 104 - 5.0 × 106 (tidal environment) 14 3.40 × 105 3.38 × 104 - 3.41 × 106 3.40 × 105 3.38 × 104 - 3.41 × 106

Total bacteria measured using qPCR, in gill and gut tissues of oysters collected directly from the field on Day 0 and 14 of the experiment and oysters that were acclimated to the laboratory environment (constantly immersed in seawater/maintained in a simulated tide) for 7 and 14 days. *Geometric means and their corresponding 95% confidence intervals were derived from the back-transformed model means of generalised linear mixed models (GLMM), for bacterial genomes per g of tissue (n = 5-10/day/tissue)

128

129

Figure 5.1 Temporal changes in (A) total cultivable bacterial count (TCBC; log10 colony forming units/g) and (B) total cultivable Vibrio count (TCVC; log10 colony forming units/g) in whole soft-tissue mass of oysters that were acclimated to the laboratory environment under constant immersion in seawater (constant; C) and under a simulated tide (tidal; T), for 3, 7 and 14 days. TCBC was lower in Day 7 compared to Day 3 and Day 14 in oysters constantly immersed in water (p <0.05). TCVC was higher on Days 3, 7 and 14 compared to Day 0 in both systems (constant and tidal) (p <0.05). All data are represented as log10 means ± SE.

5.4.2 Changes in microbiome during laboratory acclimation

5.4.2.1 High throughput 16S rRNA gene sequencing

Targeting the hypervariable V1-V3 region of the 16S rRNA gene, a total of 1,459,329 paired-end raw reads were obtained initially from the samples analysed (n = 60), leaving 1,322,581 reads after quality control and bioinformatic processing. The number of reads per gill tissue sample (median: 7402; maximum: 38476; minimum: 1926) was lower compared to that of gut tissue samples obtained from the same oyster (median: 31906; maximum: 95925; minimum: 2991; p < 0.05). The reads were rarefied to 4900 per sample for both gill and gut samples. Rarefaction curves showed saturation for most of the samples, indicative of a good coverage of diversity (Figure 5.2).

5.4.2.2 Bacterial community composition of oysters from the field

Inter-oyster heterogeneity in bacterial community composition was observed in both the gill and gut microbiota of oysters despite them being collected from a single batch of oysters with a common origin, from the same environment and on the same day. Despite the differences in TCBC and TCVC, this inter-oyster heterogeneity in both gill and gut microbiota of field oysters collected on Days 0 and 14 was not different (p > 0.05). The alpha diversity between gill microbiota of field oysters collected on Day 0 and Day 14 was also similar (observed OTUs ± SE: Day 0, 227.2 ± 32.25; Day 14, 223.8 ± 35.80; p > 0.05). Interestingly, the alpha diversity of gut microbiota from oysters in the field was higher on Day 14 compared to Day 0 (observed OTUs ± SE: Day 0, 245 ± 72.6; Day 14, 516 ± 69.8; p < 0.05). The bacterial community composition in gill tissues was dominated by phylum Proteobacteria (Figure 5.3A and C; Table 5.3) followed by the phyla Cyanobacteria, Bacteroidetes and Tenericutes, on both days. The gut microbiota of field oysters was dominated by phylum Tenericutes (class Mollicutes) on Day 0 while Proteobacteria was

130

dominant on Day 14 (Figure 5.3B and D; Table 5.3). Cyanobacteria and Bacteroidetes were the second and third most abundant phyla, on both days.

300

200

Observed OTUs Observed 100 Day 0 Day 7 Day 14 Day 14 Field

0 0 1000 2000 3000 4000 5000 Sequencing Depth

500

400

300

200

Observed OTUs Observed 100 Day 0 Day 7 Day 14 Day 14_field 0 0 1000 2000 3000 4000 5000 Sequencing Depth

Figure 5.2 Rarefaction curves for (A) gill microbiota (B) gut microbiota of oysters collected from the field and post-acclimation to the laboratory environment. Colours represent different samples collected on different days from different groups of oysters of this study.

131

132

133

134

Figure 5.3 Taxa plots summarizing the relative abundance of bacterial phyla in: A) gill tissue B) gut tissue of oysters obtained from the field (on Day 0; D0 and Day 14; D14f) and in oysters acclimated to the laboratory environment for 7 (D7) and 14 (D14) days, constantly immersed in water; C) gill tissue D) gut tissue of oysters obtained from the field (on Day 0; D0 and Day 14; D14f) and in oysters acclimated to the laboratory environment for 7 (D7) and 14 (D14) days, in a simulated tidal environment. Each column represents the bacterial community composition of an individual oyster. Bacteria that could not be assigned to a particular phylum are categorized under ‘unassigned’ and phyla with a relative abundance of less than 5% and were not present in at least two samples are categorized as ‘others’.

135

Table 5.3. Changes in the absolute abundance of dominating bacterial phyla in gill and gut microbiota during acclimation to the laboratory environment

Phylum Day 0 Day 7 Day 14 Significance* Gill microbiota: Proteobacteria 3.48 × 103 (2.88 × 103 - 7.81 × 103) 3.93 × 103 (5.34 × 102 - 2.38 × 104) 2.2× 103 (1.51× 103 - 3.96 × 103) p > 0.05

Cyanobacteria 1.05 × 103 (7.16 × 102 - 1.89 × 103) 2.4 × 101 (1.1 × 101 - 3.97× 102) 4.0 × 101 (0 - 4.9 × 101) p < 0.05

Bacteroidetes 1.02 × 103 (7.02 × 102 - 3.15 × 103) 9.8 × 102 (2.97 × 102 - 1.14 × 104) 1.54 × 103 (7.02 × 102 - 3.58 × 103) p > 0.05

Tenericutes 2.11 × 102 (8.3 × 101 - 9.35 × 103) 5.0 × 101 (0 - 3.52 × 102) 4.0 × 101 (0 - 4.9 × 101) p < 0.05

Gut microbiota: Tenericutes 6.6 × 103 (3.45 × 102 - 1.51 × 104) 8.89 × 102 (6.0 × 101 - 2.33× 103) 5.3 × 101 (4.0 × 101 - 4.24× 102) p < 0.05

Proteobacteria 1.5 × 103 (3.9 × 102 - 1.97 × 104) 1.21 × 104 (5.26 × 103 - 1.7 × 104) 1.4 × 104 (7.99 × 103 - 3.21 × 104) p > 0.05

Cyanobacteria 8.92 × 102 (2.67 × 102 - 1.38× 104) 9.0 × 101 (3.4 × 101 - 2.48× 102) 2.78 × 102 (6.8 × 101 - 4.46 × 102) p < 0.05

Bacteroidetes 3.12 × 102 (2.5 × 101 - 3.7 × 103) 7.55 × 103 (1.68 × 103 - 1.24 × 104) 1.97 × 103 (8.55× 102 - 1.3 × 104) p < 0.05

Absolute abundance (median (range)) of dominating bacterial phyla in gill and gut microbiota of oysters during laboratory acclimation for 7 and 14 days. The absolute abundance values obtained from QIIME2 analysis of 16S (V1-V3) rRNA gene diversity profiling were statistically analysed using non-parametric Kruskal-Wallis tests.

* Significance was set at p < 0.05.

136

5.4.2.3 Microbiome of oysters during acclimation

The alpha diversity in both gill and gut microbiota of oysters did not differ during 14 days of acclimation to the laboratory environment irrespective of the management system with respect to immersion time (Table 5.4; p > 0.05). However, changes in the beta diversity of both gill and gut microbiota were observed with acclimation to the laboratory environment, under both management systems (Table 5.5; Figure 5.4). The Bray-Curtis dissimilarity between bacterial communities of field oysters and laboratory oysters, increased during acclimation to the laboratory environment in both gill and gut microbiota (Figure 5.4; p < 0.05). It is important to note that this difference exceeded the inter-oyster heterogeneity in bacterial community composition that was observed before laboratory acclimation (mean Bray-Curtis distance ± SE: Day 0, 0.76 ± 0.03; Day 0 vs Day 7, 0.86 ± 0.02; Table 5.5; p < 0.05). Irrespective of the management system (constant immersion vs. simulated tide), the abundance of the dominant phyla Proteobacteria and Bacteroidetes in gill tissue did not change after acclimation (Table 5.3; p > 0.05). In contrast, the abundance of phylum Cyanobacteria and phylum Tenericutes reduced after 14 days in the laboratory (p < 0.05; Table 5.3). In addition to the changes observed at phylum level, an interesting feature was the increase in abundance of genus Arcobacter (phylum Proteobacteria), after 14 days in the laboratory (p < 0.05). Considering the gut microbiota, the abundance Phylum Proteobacteria did not change after acclimation (Table 5.3; p > 0.05). However, the fraction of Vibrio that demonstrated a very low number of reads in the gut microbiota of field oysters, increased significantly after acclimation to the laboratory environment for 14 days (p < 0.05). This result agreed with the increase in TCVC on Day 14 post-acclimation in both management systems and was supported by qPCR quantification of Vibrio genome copies. The abundance of phylum Bacteroidetes in gut microbiota increased after acclimating to the laboratory environment for 7 days (p < 0.05). Meanwhile, phylum Cyanobacteria and phylum Tenericutes decreased in abundance, after 7 days in the laboratory (p < 0.05). It is interesting to note that the abundance of Mycoplasma in both gill and gut microbiota did not reduce with laboratory acclimation despite the significant reduction in the abundance of phylum Tenericutes (p > 0.05).

137

Table 5.4 Alpha diversity (mean observed OTUs ± SE) of gill and gut microbiota of oysters during acclimation to the laboratory environment for 0, 7 and 14 days, under two different management systems (constant immersion in water and in a simulated tidal system).

Oyster Management System Observed OTUs* Day 0 Day 7 Day 14 Significance Constant Immersion: Gill microbiota 227.2 ± 32.2 167 ± 39.4 165 ± 16.0 p > 0.05 Gut microbiota 245 ± 72.6 249.4 ± 55.9 341.4 ± 46.3 p > 0.05

Simulated tide: Gill microbiota 227.2 ± 32.2 156.3 ± 65.9 160.2 ± 18.4 p > 0.05 Gut microbiota 245 ± 72.6 138.7 ± 27.7 267.6 ± 76.5 p > 0.05

138

Table 5.5 Beta diversity (mean pairwise Bray-Curtis dissimilarity ± SE) of gill and gut microbiota of oysters during acclimation to the laboratory environment for 0, 7 and 14 days, under two different management systems (constant immersion in water and in a simulated tidal system). Oyster Management Days 0 vs. 7 Days 7 vs. 14 Overall significance System

Constant Immersion: Gill microbiota 0.86 ± 0.02* 0.86 ± 0.01* p < 0.05 Gut microbiota 0.96 ± 0.01* 0.82 ± 0.02 p < 0.05

Simulated tide: Gill microbiota 0.87 ± 0.02* 0.79 ± 0.03* p < 0.05 Gut microbiota 0.94 ± 0.01 0.72 ± 0.03 p < 0.05 *Bray-Curtis dissimilarity between the bacterial communities of the two days of concern, exceeded the Bray-Curtis dissimilarity between bacterial communities of each individual day.

139

Figure 5.4 Principal coordinate plots based on Bray-Curtis distances: A) Gill microbiota in Pacific oysters; Gill microbiota in field oysters collected on Day 0 (red) and Day 14 (green) showed clustering (p > 0.05) while gill microbiota of oysters acclimated to the laboratory environment for 7 days (blue) and 14 days (purple) were different (p < 0.05) from the field oysters and with each other (p < 0.05). A similar pattern was observed for gill microbiota in oysters maintained in a simulated tidal environment. B) Gut microbiota in Pacific oysters; Gut microbiota in field oysters collected on Day 0 (red) and Day 14 (green) showed clustering (p < 0.05). Further, the gut microbiota of oysters acclimated to the laboratory environment for 7 days (blue) and 14 days (purple) showed clustering (p > 0.05). A similar pattern was observed for gut microbiota in oysters maintained in a simulated tidal environment.

140

5.5 Discussion

This study focuses on the microbiome of Pacific oysters inhabiting a specific farming environment over a period of 2 weeks in autumn, and the changes that occurred in the bacterial community composition during acclimation to controlled laboratory environments. Although the bacterial community composition was not static in the farm environment, there were consistent systematic changes observed during laboratory acclimation that increased from week 1 to week 2. These changes should be considered in the interpretation of results of experimental studies of oysters in a laboratory environment.

A striking feature of the present study was the difference between the alpha diversity in gut microbiota of field oysters collected from the same environment on two different days. The same oysters did not show such a difference in gill microbiota. Further, the gill microbiota of field oysters was dominated by phylum Proteobacteria on both days except in one oyster where phylum Tenericutes dominated the gill microbiota. On the other hand, phylum Proteobacteria and phylum Tenericutes were found to dominate the gut microbiota on the two different days. Despite the large surface area of gills exposed to environmental bacteria, the gill microbiota seems to demonstrate some degree of stability. This result is consistent with the findings of Wegner et al. (2013) where the resident population of gill bacteria was minimally disturbed by filter-feeding. According to Zurel et al. (2011) and Roterman et al. (2015), the gill microbiota in oysters are relatively stable while the microbiota of the digestive system generally consists of transient bacteria. The findings of the study by Trabal et al. (2012) further strengthened this idea. They attributed the reduced bacterial diversity in gut microbiota of oysters to depuration conducted before sample collection. Depuration has been previously employed by Romero et al. (2002) to remove the allochthonous (transient) bacteria while retaining the autochthonous (resident) bacteria in oyster tissues. The gut microbiota were dependent upon the external environment, diet, the life stage and the physiological status of marine invertebrates (Harris, 1993; Trabal Fernández et al., 2014). Trabal et al. (2012) also highlight a potential influence on gut microbiota of cultured marine organisms by the changes in environment such as availability of food, operational design in grow-out and management methods. According to Colwell and Liston (1960) all microorganisms in the immediate environment of an oyster may actively or passively enter the oyster, however, only those that are well adapted to the microenvironment will establish themselves as a significant component of the oyster microflora.

141

Consistent with other studies (Pathirana et al., 2019a; Wegner et al., 2013), phylum Proteobacteria was dominant in gill tissue of Pacific oysters in this study. Gut tissue was found to be dominated by different phyla in different studies, including phylum Proteobacteria and phylum Fusobacteria (Pathirana et al., 2019a; Trabal Fernández et al., 2014; Vezzulli et al., 2018). Hernandez-Zarate and Olmos-Soto (2006) identified members of the phylum Proteobacteria and phylum Firmicutes as the dominant, metabolically active bacteria within the digestive gland of healthy Pacific oysters, using culture-independent techniques. Meanwhile Green and Barnes (2010) identified members of the phylum Cyanobacteria, phylum Spirochaetes and phylum Firmicutes (Clostridium spp.) in the gut microbiota of Sydney rock oyster (Saccostrea glomerata). The Cyanobacteria spp. were associated with filter-feeding while members of the other two phyla were involved in digestion (Green and Barnes, 2010). The transient role of Cyanobacteria spp. was also observed in the present study where the abundance of phylum Cyanobacteria reduced with acclimation to the laboratory environment, suggesting a possible unfavourable condition in the laboratory environment, for these species.

The different relative abundances in bacterial phyla between gill and gut tissues in this study, supports the concept of tissue-specificity in oyster microbiota (Lokmer et al., 2016a; Pathirana et al., 2019a). Furthermore, the present study demonstrated an abundance of class Mollicutes (phylum Tenericutes) in the gut, in line with studies that observed similar abundances in the digestive gland of the Sydney rock oyster (Saccostrea glomerata) (Green and Barnes, 2010) and in the stomach microbiome of eastern oyster (Crassostrea virginica) (King et al., 2012). Nevertheless, the functional role of class Mollicutes in the oyster gut is not clear (King et al., 2012). As the bivalve digestive gland plays a major role in the digestion and storage of nutrients, Vezzulli et al. (2018) suggest that the bacteria associated with the digestive gland may also contribute to this role. Studies have reported both commensalism of Mollicutes (Bano et al., 2007; Kellogg et al., 2009) and pathogenesis of Mycoplasma and Mycoplasma-like organisms in other aquatic invertebrates (Azevedo, 1993; Krol et al., 1991) and in fish (Kirchhoff et al., 1987). Furthermore, it is interesting to highlight that Vibrio was not a major component of phylum Proteobacteria in both gill and gut microbiota. Trabal et al. (2012) have also reported a very low abundance of Vibrio in gut tissues obtained from healthy Pacific oysters. As opposed to the conventional knowledge that identified Vibrio spp. as one of the dominating bacterial genera in oysters (Colwell and Liston, 1960; Prieur et al., 1990), a large fraction of the oyster microbiome cannot be cultivated by standard procedures

142

(Fernandez-Piquer et al., 2012; Romero and Espejo, 2001). Previous literature on Eastern oysters (C. virginica) and Sydney rock oysters (Saccostrea glomerata) also report low relative abundances of Vibrio spp. (Green and Barnes, 2010; Ossai et al., 2017). However, Vibrio spp. can have an impact on the oyster health despite its low relative abundance (Thurber et al., 2009). A strong correlation was observed between the OsHV-1 and Vibrio quantities in OsHV-1 infected oysters (Pathirana et al., 2019b). Furthermore, de Lorgeril et al. (2018) reported an approximately 30× increase in Vibrio spp. following an OsHV-1 infection.

In the present study, a difference in bacterial community composition was observed in both gill and gut tissue after acclimation to the laboratory environment. The different microbiome that was observed after a 7 day-laboratory acclimation was further differentiated on acclimation for 14 days. Rather than reducing the bacterial diversity (observed OTUs and Shannon’s diversity index) in oyster tissues, the acclimation process reduced the abundance of selected bacterial phyla such as phylum Tenericutes and phylum Cyanobacteria, in both gill and gut microbiota while increasing the abundance of others. In this context, the phylum Proteobacteria that initially dominated the gill microbiota, further increased in abundance, followed by phylum Bacteroidetes. In particular, an increase in Arcobacter (belonging to phylum Proteobacteria) was noted. Arcobacter was not abundant in studies conducted on stomach (King et al., 2012), gut (Trabal Fernández et al., 2014) and gill microbiota (Wegner et al., 2013) of oysters. However, it is interesting to note that an increase in Arcobacter abundance has been observed in the haemolymph (Lokmer et al., 2016b) and whole tissue homogenates (Romero et al., 2002) of oysters, after laboratory pre-treatment and depuration, respectively. As the oysters were maintained in clean, ASW in this study, with no access to external bacteria other than the nitrifying bacteria in the biofilters and the potential presence of bacteria from the commercial algal feed (this feed was not tested for bacteria), the laboratory acclimation process simulated the depuration process of commercial oyster production. It is a positive trend to observe the reduction of potentially transient bacteria such as that belonging to phylum Cyanobacteria, in the context of depuration. Deviation from the natural, undisturbed bacterial community composition will have an impact on laboratory- based OsHV-1 challenge studies and will not essentially provide the true picture of a natural field infection. Infections caused by OsHV-1 μ Var have resulted in immunosuppression and shifts in the oyster microbiome to allow opportunistic infections from bacteria such as Vibrio (de Lorgeril et al., 2018). In addition to the strong correlation between OsHV-1 and Vibrio

143

loads in laboratory-based OsHV-1 infections (Pathirana et al., 2019b), polymicrobial pathogenesis of Pacific oyster mortality associated with OsHV-1 μ var infections (Petton et al., 2015b), has been demonstrated. Thus, it is important to identify an optimum duration for laboratory acclimation with minimum impact to the commensal microbiome of Pacific oysters. However, it was interesting to note that the method of laboratory acclimation (constant immersion vs. tidal emersion) did not create any differential impact on the oyster microbiome composition.

Changes in the relative abundance of different bacterial phyla during laboratory acclimation, may influence the oyster in different ways. Being a photosynthetic group, phylum Cyanobacteria may not play an important role in any disease pathogenesis. Green and Barnes (2010) have demonstrated Cyanobacteria in the digestive gland of healthy Saccostrea glomerata as opposed to those affected by QX disease and have ceased feeding, suggesting a transient role of Cyanobacteria. On the other hand, the steady abundance of the Mycoplasma fraction (in the face of decline in abundance of phylum Tenericutes) in both gill and gut microbiota and that of phylum Bacteroidetes in gill microbiota, are positive findings of this study. In contrast to laboratory acclimation, an increase in the abundance of Mycoplasma and decrease in the abundance of phylum Bacteroidetes have been reported in disturbance (heat stress, starvation and transportation stress) in a study that analysed gill microbiota (Wegner et al., 2013). Nevertheless, the increase of the Vibrio fraction with laboratory acclimation cannot be overlooked. This increase was evident in the TCVC, quantification by qPCR and also through bacterial diversity profiling using 16S rRNA gene sequencing. It may create an impact on the potential for Vibrio infections that can occur secondary to OsHV-1 infections (de Lorgeril et al., 2018). Further, Lemire et al. (2015) have shown that the non-virulent strains of Vibrio spp. inhabiting healthy oysters get progressively replaced by phylogenetically coherent, virulent strains of Vibrio during the onset of an infection. The non-virulent strains are expected to facilitate the disease caused by the virulent strains. An increase of Vibrio during acclimation to the laboratory environment may play a supportive role in experimental OsHV-1 infections. However, further studies are required to confirm that the increase in Vibrio during laboratory acclimation is a maladaptation which supports the pathogenesis of OsHV-1 infections.

Inter-oyster heterogeneity in bacterial community composition was observed in both the gill and gut microbiota of oysters despite them being collected from a single batch of

144

oysters with a common origin, from the same environment and on the same day. Inter-oyster heterogeneity of microbiota has previously been demonstrated in gill microbiota (Wegner et al., 2013), adductor-muscle microbiota (King et al., 2018a) and multiple tissue microbiota (Pathirana et al., 2019a). Wegner et al. (2013) suggested an association between individual oyster genotypes and oyster microbial communities beyond the influence of environment and genetic origin. They have further demonstrated a significant correlation between individual pairwise genetic distances and microbial community distances, for oysters. Based on this correlation, Wegner et al. (2013) suggested that the bacterial communities in oysters are assembled according to individual oyster genotypes. Similarly, studies have shown that the human gut microbiota is unique for individuals and that similarities exist between related individuals (Vanhoutte et al., 2004).

The increase of Vibrio in gut tissue should also be considered in the context of food safety. As mentioned above, the laboratory acclimation procedure can be compared to the depuration process in commercial oyster production. While UV sterilized seawater is being used in oyster depuration, artificial seawater prepared from chlorinated and conditioned tapwater was used in the laboratory acclimation process. The primary objective of commercial depuration is to expel the biological contaminants in oyster tissues, in particular the faecal bacterial contaminants such as Escherichia coli (Lee et al., 2008). Ideally, commercial depuration should also remove common shellfish-borne pathogens belonging to the genus Vibrio. However, depuration has not always been effective in removing naturally occurring marine vibrios such as V. parahaemolyticus and V. vulnificus (Lee et al., 2008) that are pathogenic to humans and V. aestuarianus and V. splendidus clade that are pathogenic to oysters (Lacoste et al., 2001; Vezzulli et al., 2018). In line with this, the present study also did not observe a reduction of the Vibrio fraction but rather observed an increase of Vibrio in the gut, at the end of a 14-day acclimation period. Similarly, Lokmer et al. (2016b) observed an increase in Vibrionaceae in haemolymph microbiota of oysters that underwent pre- treatment in a laboratory. Vezzulli et al. (2018) also demonstrated an increase in the Vibrio fraction after depuration, in haemolymph of C. gigas but not in the digestive gland. However, the same study showed an increase in Vibrio in the digestive gland and haemolymph of Mediterranean mussel (Mytilus galloprovincialis), after depuration. Vezzulli et al. (2018) sees this resistance of Vibrio to depuration as a feature of the possibly long co-evolutionary history of Vibrio with their invertebrate hosts, representing permanent residents of the microbiota. Abundance of Vibrio has also been observed in solid tissues of C. gigas

145

maintained in ASW (Lokmer et al., 2016a) but not in haemolymph (Lokmer and Wegner, 2015). This is thought to be a result of the stationary conditions in the tanks (Lokmer et al., 2016b). Thus, it is important to understand the potential risk of food-borne Vibrio infections even after commercial depuration of oysters, in the event of the presence of pathogenic Vibrio spp. such as V. parahaemolyticus and V. vulnificus.

5.6 Conclusion

The microbiome of Pacific oysters with a common origin and living in a common environment has individual variations. While the gill microbiota was relatively stable, the gut microbiota was more variable. This feature should be taken into consideration when selecting tissues for temporal studies of the oyster microbiome. The oyster microbiome changed during acclimation to a controlled laboratory environment. When considering the microbiome of oysters in this laboratory setting, there was no impact of maintaining oysters in constant immersion in water as opposed to maintaining them in a simulated tidal environment. Although the total quantity of bacteria did not change after 14 days in the laboratory, the bacterial community composition changed reflecting the relative abundance at different taxonomic levels including the increase in Vibrio in gut microbiota and Arcobacter in gill microbiota. The microbiome changed after acclimating to the laboratory for 7 days and further differentiated on acclimation for 14 days. Similar changes in the microbiome should be considered in the design and analysis of experimental studies using Pacific oysters and might be relevant for understanding commercial depuration of oysters.

146

CHAPTER 6

Impact of constant immersion compared to tidal emersion on the microbiome of Pacific oysters (Crassostrea gigas) challenged with Ostreid herpesvirus-1 (OsHV-1)

6.1 Abstract

Microvariant genotypes of Ostreid herpesvirus-1 (OsHV-1) are responsible for mass mortality outbreaks of Pacific oysters. Increasing the height of growing structures in intertidal farm environments decreased mortality of adult oysters as the periods of emersion from water were increased. The present study focused on the microbiome of Pacific oysters challenged with OsHV-1 and maintained in different immersion regimes, to better understand the impact of tidal emersion. Two groups of oysters (n = 212 in total) were maintained under constant immersion and in periodic emersion. A random selection of live oysters on Days 0,1, 2, 3, 4 and 7 post-injection and oysters at the time of death were sampled for quantification of OsHV-1, total bacteria and total Vibrio and for bacterial 16S rRNA gene (V1-V3) diversity profiling in gill and gut. Using this laboratory challenge model, I demonstrated higher mortality in oysters injected with OsHV-1 in a tidal immersion-emersion pattern compared to those in constant immersion. This contrasted with the outcome in prior field trials and indicated that the protective effect of elevated growing height in the field was not a result of beneficial altered physiology of oysters during emersion. Higher mortality in the tidal emersion was accompanied by a lower total bacterial count (p < 0.05). Reduced alpha diversity was noted in the gill microbiome of constantly immersed oysters, at the onset of mortality (observed operational taxonomic units [OTUs]: Day 1, 192.7 ± 20.7, Day 2, 143.4 ± 8.1; p < 0.05). This was not observed in tidal oysters (OTUs: Day 1, 155 ± 23.5, Day 2, 192 ± 14.8). A reduction in Vibrio spp. (Day 0: 11.09%; Day 1: 2.9%) and an increase in Arcobacter spp. (Day 0: 0.18%; Day 1: 5.93%) in the gill microbiota was seen in constant immersion (p < 0.05) but did not occur in tidal emersion. Bacterial genera that were rare in the gill microbiota including Polaribacter, Marinicella and Sediminibacterium increased in abundance in tidal oysters (p < 0.05). The alpha diversity of gut microbiota did not change after the OsHV-1 challenge in both systems. The OsHV-1 challenge seemed to initiate a differential response in the microbiome of oysters under different immersion regimes. Further studies are required to evaluate the potential role of Arcobacter in oyster mortality associated with OsHV-1.

147

Keywords: Laboratory infection model, intertidal environment, microbiome, Pacific oyster, Crassostrea gigas, Ostreid herpesvirus-1

6.2 Introduction

The Pacific oyster (Crassostrea gigas) is the most important commercial oyster species globally, with an annual global production of 625,925 tonnes worth US$1.3 billion out of a total edible oyster production of 5.2 million tonnes worth US$ 4.2 billion (FAO, 2014). Since 1950, outbreaks of summer mortalities in adult Pacific oysters have been reported in many parts of the world (Glude, 1974; Maurer et al., 1986; Takeuchi et al., 1960). In France, mass mortality events have also been reported in C. gigas spat since 1993 (Renault et al., 1994a), which were later identified to be associated with Ostreid herpesvirus-1 (OsHV- 1) (Davison et al., 2005; Garcia et al., 2011). The reference genotype of Ostreid herpesvirus 1 (OsHV-1) (Davison et al., 2005; Le Deuff and Renault, 1999) and related genotypes have been identified as the prominent pathogens which caused C. gigas mortalities in France, predominantly in oyster larvae, from 1991 to 2008 (Martenot et al., 2011; Renault et al., 2012). However, at the end of spring 2008, much more severe and widespread mortalities in France killed billions of young oysters and a variant of OsHV-1, called the microvariant (µVar) genotype, was identified from these outbreaks (Renault et al., 2012; Segarra et al., 2010). Since then, severe mortality disease outbreaks have greatly impacted Pacific oyster production in Europe (Martenot et al., 2011; Renault et al., 2012; Segarra et al., 2010), Australia (Jenkins et al., 2013; Paul-Pont et al., 2014; Whittington et al., 2018) and New Zealand (Keeling et al., 2014). Considerable research has been directed at understanding the mode of infection, transmission and host susceptibility to OsHV-1 (Alfaro et al., 2019) with the need to identify more effective strategies to control disease outbreaks associated with OsHV-1 and to prevent the spread of OsHV-1 to uninfected estuaries.

Pacific oyster mortality events are most often considered a result of complex interactions which are influenced by the physiological status of the oysters, many environmental parameters and the actions of multiple potential pathogens (Pernet et al., 2012; Petton et al., 2019; Samain et al., 2007). In particular, the disease associated with OsHV-1 requires not only the presence of OsHV-1, but also environmental conditions that are favourable for the disease and suitable host factors (de Kantzow et al., 2017; Evans et al., 2015; Evans et al., 2019; Petton et al., 2015a). The severity of disease is influenced by host factors such as genetics (Dégremont, 2011; Samain et al., 2007), size and age of oysters

148

(Azéma et al., 2017a; Hick et al., 2018; Renault et al., 1994b), and environmental factors such as seawater temperature (de Kantzow et al., 2016; Delisle et al., 2018; Petton et al., 2013). Understanding the complex relationships that exist between the host, pathogens and the environment is required to develop control strategies for mass mortality in Pacific oysters.

The orientation of oysters in relation to the water current, the degree of water circulation around each oyster and the intertidal height are all considered to influence the mortality patterns of oysters in natural mortality events in the field (Paul-Pont et al., 2013b; Pernet et al., 2019; Pernet et al., 2012; Whittington et al., 2015b). Adult oysters that were grown at increased height (300 mm above the standard intertidal growing height) demonstrated a 25-50% decrease in mortality compared to those grown at the standard height during outbreaks of disease caused by OsHV-1 (Evans et al., 2019; Paul-Pont et al., 2013b; Whittington et al., 2015b). Changes in feeding opportunity and physiology that accompany increased periods of intertidal emersion may have contributed to the altered mortality of these oysters in addition to a proposed protective effect of reduced exposure to OsHV-1 (Evans et al., 2019; Whittington et al., 2015b). Increased emersion time in the intertidal estuarine environment exposes oysters to increased temperature and hypoxia during valve closure, leading to hypercapnia and reduction in pH of tissues (Allen and Burnett, 2008). Heat stress, hypoxia, hypercapnia and immunosuppression are direct effects of emersion on oyster physiology and have been identified previously (Allen and Burnett, 2008; Boyd and Burnett, 1999). As with many other bivalves, Pacific oysters maintain closed valves during tidal emersion to survive the harsh, intertidal environment (Allen and Burnett, 2008; Evans et al., 2019). The same study (Allen and Burnett, 2008) showed that emersion at a higher temperature resulted in suppressed bactericidal activity in haemocytes of oysters maintained in a simulated intertidal environment. The latter result further indicates an altered immune response as a result of altered emersion time. Moreover, exposure to environmental hypoxia has resulted in the production of 66% less reactive oxygen intermediates (ROI) in the haemocytes of C. virginica (Boyd and Burnett, 1999). These ROIs are important in the oyster defence system against viral and bacterial infections (Adema et al., 1991; Boyd and Burnett, 1999).

Contrary to the field studies (Paul-Pont et al., 2013b; Whittington et al., 2015b) there was higher mortality in oysters maintained in a simulated tide in a laboratory environment (Evans et al., 2019). This laboratory study used injection of oysters with a measured dose of

149

OsHV-1 to compare the outcome of disease in constant immersion with that in a tidal environment (Evans et al., 2019). The results suggested that the higher growing height with increased tidal emersion was beneficial only through reduced exposure to OsHV-1. In a circumstance where the OsHV-1 infection was established by injection of oysters with a measured dose, tidal emersion was detrimental to survival. The experiment by Evans et al. (2019) maintained an air temperature which was similar to that in water, to prevent the effects of heat stress on the oyster and subsequent mortality. Nevertheless, the protective valve closure mechanism is impaired in OsHV-1 infections resulting in ‘gaping’ (Evans et al., 2019) which exposes oysters to desiccation during tidal emersion (Evans et al., 2019; Soliman et al., 2015). Given the apparent discrepancy between the outcome of OsHV-1 infections in tidal oysters that were naturally exposed to the virus and those that were injected with the virus, the potential role of microbiome was questioned.

The role of bacteria in oyster mortality was first questioned after demonstration of high loads of Vibrio spp. in the haemolymph of moribund oysters (Lipp et al., 1976). The oyster microbiome is considered to play a key role in oyster health by its contribution to the pathogenesis of diseases (King et al., 2018b; Lasa et al., 2019), in addition to its beneficial effects such as denitrification in eutrophic aquatic environments (Arfken et al., 2017). It is the collection of all microbial communities that are associated with different tissues of the oyster. The composition of the Pacific oyster microbiome was correlated with the host oyster genotype and the resulting individual variations in the microbiome was particularly mediated by the assembly of rare taxa in the microbiome (Wegner et al., 2013). A recent study identified significant differences between the microbiomes of oysters from families with different levels of susceptibility to OsHV-1 µVar disease (King et al., 2019). The immediate environment can also influence the microbiome of oysters (Lokmer et al., 2016b; Pathirana et al., 2019b), owing in part to their filter-feeding behaviour (Le Roux et al., 2016; Lokmer et al., 2016a). As with other organisms, the oyster microbiome is considered dynamic and responds to changes and disturbances in the external environment including translocation to a different environment (Lokmer et al., 2016a; Wegner et al., 2013), unavailability of food (Wegner et al., 2013), and changes in environmental temperature and heat shock (Lokmer and Wegner, 2015; Wegner et al., 2013). Moreover, changes in the oyster microbiome has been reported in diseases such as QX (Green and Barnes, 2010) and summer mortality (King et al., 2018a). In an experimental study, Lokmer et al. (2016b) observed persistent effects of antibiotic treatment on the haemolymph microbiome in addition to those by the immediate

150

environmental conditions and translocation. The advent of molecular bacteriological methods enabled identification of a greater bacterial diversity in the Pacific oyster microbiome (Fernandez-Piquer et al., 2012; Hernandez-Zarate and Olmos-Soto, 2006) and the diseased oyster microbiome could be better studied using 16S rRNA gene diversity profiling (Green et al., 2018; Lokmer and Wegner, 2015). In a natural Pacific oyster mortality event, King et al. (2018a) observed changes in the microbiome composition of disease-affected oysters, in particular, an increase in rare operational taxonomic units (OTUs). OTUs are clusters of bacteria grouped by the DNA sequence similarity of their 16S rRNA gene. Another study suggested a role of bacteria in mortality associated with OsHV-1 infection, by studying the mortality pattern of oysters that were administered antibiotics, following exposure to a natural mortality outbreak (Petton et al., 2015b). In this, a 2-fold and 4-fold reduction of mortality were observed in donor oysters and recipient oysters, respectively, after the administration of antibiotics. Further, it was demonstrated that the initial microbiome composition was associated with the differential outcome of an experimental OsHV-1 infection in Pacific oysters (Pathirana et al., 2019b). Meanwhile, de Lorgeril et al. (2018) demonstrated that OsHV-1 infection in Pacific oysters results in immunosuppression which leads to opportunistic infections by bacteria such as Vibrio spp. that are present in the microbiome.

In Chapter 5 of this thesis, I observed that the Pacific oyster microbiome changed during acclimation to a controlled laboratory environment. However, the changes did not differ between the different immersion regimes of oysters. With this background, investigating the microbiome dynamics of oysters under tidal emersion, during an OsHV-1 infection may provide insights about the potential adverse effects of emersion which lead to higher mortality. The present study focused on evaluating the microbiome of oysters in a laboratory after injection of a measured dose of OsHV-1, comparing constant immersion in water with a simulated intertidal environment. Differences in the composition of the microbiome due to the intertidal environment were proposed as an explanation for differences in mortality in OsHV-1 challenged oysters.

6.3 Materials and Method

6.3.1 Oysters

151

Hatchery produced, single seed, triploid Pacific oysters were used in this study (Batch SPL15AT; Shellfish Culture Ltd., Tasmania; n = 212). The parent oysters of this batch had not been exposed to OsHV-1 and the batch was certified free from OsHV-1 by real-time quantitative PCR (qPCR) by the competent government authority at the time of interstate transport. A large number were grown under commercial farming conditions using floating baskets in Patonga Creek, Hawkesbury River, NSW. This waterway had a sandy bottom with sea grass, low turbidity water and the water temperature ranged from 14-26°C during the period from 2014-2017 (Whittington et al., 2019). The selection recruited for the study in mid-autumn of 2017 were 24 months of age and 70–90mm shell height, when they were transported to a physical containment level 2 aquatic animal facility at the University of Sydney, Camden, NSW.

Mortality due to OsHV-1 has never been reported and OsHV-1 was rarely detected during active surveillance at Patonga Creek (Whittington et al., 2019). To assess freedom from infection, qPCR tests for OsHV-1 were conducted for a random selection of the oysters prior to the study (n=52). This sample size provided evidence of freedom from infection in the oysters recruited in this study (n=212), at a minimum expected prevalence of <5 % assuming the sensitivity of the qPCR was 0.95 and the specificity was 0.99 (http://epitools.ausvet.com.au/freecalcone).

6.3.2 Experiment design and oyster management

Oysters were randomly allocated to 10 individually aerated, 20 L tanks with artificial seawater (ASW; Red Sea® salt) at 30 ± 1 ppt salinity. The tanks belonged to four recirculation systems and each comprised 6 tanks and a 250 L-sump. A heater-chiller (HC-300a, Hailea Aquarium chiller) was used to maintain the water temperature of each recirculation system at 22 °C and a biofilter (Fluval 406 canister filter) was connected to maintain water quality. Eight of the tanks were chosen at random from three recirculation systems for the experimental challenge while 2 tanks were designated from a fourth recirculation system to be negative controls in which oysters were injected with an OsHV-1 free tissue homogenate (Figure 6.1). Each tank designated for this experiment contained 16 oysters at the time of OsHV-1 challenge. Half of the tanks were maintained with oysters constantly immersed in ASW and a simulated tidal system was applied to the others. The immersion-emersion pattern required removal of the standpipes from each tank at 11:00 am and 11:00 pm and replacing the standpipes at 6:00 am and 6:00 pm each day resulting in the oysters being alternately emersed and immersed in

152

ASW (immersion: 2 × 5 h/day; emersion: 2 x 7 h/day). The air temperature was maintained at 24 °C ± 2 °C using an air conditioning unit.

TIDAL CONSTANT n=16 n=16 System 1 oysters oysters REPLICATE 1 REPLICATE 1

CONSTANT CONSTANT TIDAL TIDAL n=16 n=16 n=16 n=16 System 2 oysters oysters oysters oysters REPLICATE 2 REPLICATE 3 REPLICATE 2 REPLICATE 3

CONSTANT TIDAL n=16 n=16 System 3 oysters oysters REPLICATE 4 REPLICATE 4

TIDAL CONSTANT n=15 n=16 System 4 oysters oysters NC NC

Figure 6.1 Schematic representation of the experimental design showing the allocation of oysters (C. gigas) across the 4 recirculation systems and between treatment groups, before the OsHV-1 challenge. Each square indicates one tank. The oysters in the tidal treatment experienced 5 h of immersion and 7 h of emersion twice per day. NC=negative control (oysters injected with an OsHV-1 free tissue homogenate). Note: the tank ‘tidal NC’ in system 4 had 15 oysters immediately before the OsHV-1 challenge as 1 oyster died during acclimation (OsHV-1 negative qPCR).

153

Water quality was tested daily using an API ® Marine Saltwater Master Test kit. The pH of water was maintained in the range of 8.0−8.2 and total ammonia nitrogen, nitrite and nitrate were maintained <0.25 ppm by exchange of water when required and by adding sodium bicarbonate to the sump. Oysters were held on a perforated plastic insert approximately 10 cm above the bottom of the tank and were fed with a commercial algae diet (Instant Algae® Shellfish Diet 1800; Reed Mariculture). Two millilitres of the algae diet were added to each tank daily at 10.00 and the oysters were allowed to feed for 1 h with the recirculation system disconnected from the treatment tanks to prevent removal of the feed. Additionally, the oysters in the constant immersion treatment were fed at 14.00, to reflect the greater feeding opportunity with constant immersion. Oysters were acclimated to the experimental system for 10 days before challenging with OsHV-1.

This study was conducted in parallel to another study which analysed the OsHV-1 quantity in the same oysters, after injection (Evans et al., 2019). The OsHV-1 quantity data was used in the present study to interpret the changes observed in the microbiome.

6.3.3 Challenge with OsHV-1

The oysters were challenged with OsHV-1 by injection according to the design of the experiment conducted parallel to the present study (Evans et al., 2019). Cryopreserved stocks of 0.2 µm filtered oyster tissue homogenate was used as the inoculum. This stock was prepared from OsHV-1 infected oysters collected during a disease event in the Georges River, NSW, Australia in 2011 (Paul-Pont et al., 2015) and had been used in previous experimental infections (Pathirana et al., 2019b). The cryopreserved inoculum was thawed at 4°C and was diluted in sterile artificial seawater to obtain a further dilution of 1/100. The oysters were immersed in a solution of MgCl2 (50 g/l) for 4–6h until relaxation of the adductor muscle caused the valves to open. Each oyster was injected into the adductor muscle with 100 μl of the diluted OsHV-1 inoculum (5.84×103 viral copies per 100 μL) using a 1 ml syringe and a 25-gauge needle. The control oysters were injected with a diluted, cryopreserved, filtered tissue homogenate prepared from apparently healthy Pacific oysters from an OsHV-1-free population (Evans et al., 2015). Physical separation and care during the procedures focused on preventing cross contamination between OsHV-1 challenged tanks and negative control tanks such as carrying out the routine procedures (feeding, water-quality testing) in the control tanks first.

154

6.3.4 Sampling

Mortality was assessed twice daily at 9:00 am and 3:00 pm by visual inspection and dead oysters were sampled. Mortality was identified by non-responsive gaping on exposure to external stimuli including handling and air exposure for 3 min. Sampling of live oysters was performed on Days 0, 1, 2, 3, 4 and 7 post-injection. Two live oysters were randomly sampled from each tank on the designated days. This was achieved by assigning each oyster a number based on its position in the tank. Two random numbers between 1 and the number of oysters in the tank at that time were generated for each tank using Microsoft Excel®. All remaining oysters that survived the infection challenge were sampled on Day 10. Sampled oysters were stored at 4 °C prior to dissection. The cumulative mortality for each tank and treatment group was calculated according to the method described by Whittington et al. (2015b) which takes into account the number of live oysters sampled on each day (Evans et al., 2019).

6.3.5 OsHV-1 DNA quantification

Quantification of OsHV-1 DNA in oysters was determined in a concurrent experiment (Evans et al., 2019), as mentioned in Section 6.2.2. Gill and mantle tissue were collected from each oyster and stored, as described in Section 2.4.1. These tissue samples were then homogenized, and the final supernatant was stored, as described in Section 2.6.1. Nucleic acids were extracted from these supernatants, purified and stored, as described in Section 2.7.1. The number of copies of the B-region of the OsHV-1 genome was determined by a qPCR assay as described in Section 2.8.

6.3.6 Identification and quantification of cultivable Vibrio and total bacteria

After the gill and mantle tissue sampling described in Section 6.2.5, the remaining soft tissues of each oyster was sampled for bacterial culture and molecular assessment of bacterial content. Gill and gut tissue samples were collected and stored as described in Section 2.4.2. Tissue homogenates were prepared from the remaining soft tissues for direct bacterial culture, as described in Section 2.4.3. Cultivable Vibrio and total bacteria were isolated and quantified, as described in Section 2.5.2 and 2.5.3. The marine salt agar – blood (MSA-B) and thiosulfate-citrate-bile salts-sucrose (TCBS) agar was prepared as described in Section 2.5.1 of this thesis.

155

6.3.7 Molecular quantification of bacteria

Nucleic acids were purified from gill and gut tissue samples collected on Section 6.2.6, as described in method B of Section 2.7.2 total Vibrio spp. DNA was quantified by qPCR as described in Section 2.9. Total bacteria DNA in gill tissue was quantified by qPCR as described in Section 2.10.

6.3.8 Microbiome analysis by high throughput 16S rRNA gene sequencing

Nucleic acid extracts were selected from both gill and gut tissue samples from all treatment groups on all sampling days (n=5/day/immersion regime). The bacterial community composition of each extract was identified by high-throughput sequencing of the hypervariable V1-V3 region of 16S rRNA gene, as described in Section 2.11. The quality of the raw sequence data was assessed as described in Section 2.11.3 and formatted for QIME2 analysis according to the methods described in Sections 2.11.4 to 2.11.6. The processed sequence data were analysed using Quantitative Insights into Microbial Ecology Version 2 (QIIME2) Software Suite (2018.11 release), as described in Sections 2.11.7 to 2.11.9. The number of observed OTUs was used as the parameter to assess alpha diversity of samples. The dissimilarity of bacterial community structure between samples (beta diversity) was visualized by principal coordinate plots (PCoA) based on the two-dimensional Bray-Curtis (BC) dissimilarity index. The relative abundance of bacteria in each sample was initially visualized using interactive taxonomic bar plots and then the relative abundance of bacterial phyla in samples from different treatment groups was graphically presented using 100% stacked 2-D column graphs (Excel, Microsoft). The quality of the raw sequence data was assessed and subsequently processed and analysed using Quantitative Insights into Microbial Ecology Version 2 (QIIME2) Software Suite (2018.11 release), as described in Sections 2.11.3 to 2.11.9.

6.3.9 Statistical analysis

The mortality of oysters under different immersion regimes were evaluated by plotting Kaplan-Maier curves (SPSS Statistics ver. 22; IBM SPSS Cooperation, Somers, NY, USA) and by Cox proportional hazard models using SAS (SAS Institute Inc., 2002–2012). Oysters that were sampled live and surviving oysters at the end of the experiment were censored at the respective observation times, in the Kaplan-Maier curves. The quantity of OsHV-1 DNA, TCBC, TCVC, total bacteria DNA and total Vibrio spp. DNA in oysters were

156

compared between different treatment groups. The data were log10 transformed and normal distribution was assessed using the Shapiro-Wilk test (SPSS Statistics). Separate generalized linear mixed models were used for analysis of OsHV-1 DNA, TCBC, TCVC, total bacteria and total Vibrio (GLMM, SPSS Statistics). The immersion regime (constant immersion or tidal emersion), outcome of infection (live or dead), treatment (OsHV-1 injected or negative control) and the days post-injection (Day 0, 1, 2, 3, 4 and 7) were used as fixed effects together with possible interactions while replicate tank identification number was used as a random effect, in the models. As all negative control oysters were negative for OsHV-1 throughout the experiment, the treatment factor was not taken into consideration in the model concerning OsHV-1 DNA. No dead oysters were sampled for conventional bacterial cultures and hence the outcome of infection was not taken as a factor in models concerning TCBC and TCVC. Meanwhile, the tissue-type (gill or gut) was also used a fixed effect in models concerning total bacterial DNA and total Vibrio DNA. Post-hoc pairwise mean comparisons were made using the least significant difference method. The results were presented as geometric means and their corresponding 95% confidence intervals, for OsHV-1 genome copies/mg, CFU/g (for TCBC and TCVC), bacterial genome equivalents or Vibrio spp. genome equivalents/g, respectively. Significance was set at p < 0.05 for all statistical analyses.

The number of DNA sequence reads were compared between gill and gut tissue microbiota using a non-parametric Kruskal-Wallis test (SPSS ver. 22). Microbial diversity analyses were performed using the q2-diversity plugin of QIIME2 to compute alpha and beta diversity metrics and generate interactive visualizations with statistical analysis (Bokulich et al., 2018). The number of observed OTUs and Shannon’s diversity index were used as parameters to assess alpha diversity of samples and were analysed using Kruskal-Wallis test. The dissimilarity of bacterial community structure between samples (beta diversity) was visualized by principal coordinate plots (PCoA) based on the two-dimensional Bray-Curtis (BC) dissimilarity index. One-way permutational multivariate analysis of variance (PERMANOVA) was used for statistical analysis of beta diversity. Temporal variation in the absolute abundance of selected phyla and genera was assessed using Generalized Linear Models (GzLM; SPSS). Significance was set at p < 0.05 for all statistical analyses.

157

6.4 Results

6.4.1 Mortality

The survival analysis indicated a higher survival in OsHV-1 challenged oysters maintained in constant immersion compared to the oysters in tidal emersion (Figure 6.2; p < 0.05). Mortality commenced on Day 2 post-injection in oysters challenged with OsHV-1 in both constant immersion and in tidal emersion. None of the oysters that were constantly immersed in water died after Day 3 post-injection when the total cumulative mortality was 11.3%. However, mortality continued until Day 5 post-injection and the cumulative mortality was much higher (67.2%) for oysters maintained in the simulated tide (Evans et al., 2019). The hazard of death from OsHV-1 was 4.02 times higher for oysters in the tidal environment compared to constant immersion (Evans et al., 2019). One oyster in negative control group died during the acclimation to the tidal treatment. This oyster was negative for OsHV-1 when tested by qPCR (Evans et al., 2019).

Figure 6.2 Kaplan-Meier survival curves for Pacific oysters challenged with OsHV -1 and maintained under constant immersion (constant) and under tidal emersion (tidal). A higher survival was observed in OsHV-1 challenged oysters maintained under constant immersion (p < 0.05).

158

6.4.2 Quantification of OsHV-1 DNA

Overall, the oysters that died after the OsHV-1 challenge had a higher OsHV-1 DNA load compared to the randomly sampled live oysters after the OsHV-1 challenge (Table 6.1; p < 0.05). After allowing for variation due to outcome of infection (live or dead), the OsHV-1 concentration was higher for oysters in tidal emersion compared to those under constant immersion (Table 6.1; p < 0.05). For oysters that were sampled live, the peak OsHV-1 DNA concentration occurred on Day 2 post-injection for both constant immersion (3.76×104 OsHV-1 genome equivalents/mg) and tidal emersion (3.71×105 OsHV-1 genome equivalents/mg) (Evans et al., 2019).

Table 6.1 OsHV-1 DNA concentrations in gill and mantle tissues of oysters challenged with OsHV-1. OsHV-1 DNA was quantified by qPCR from both live oysters and dead oysters (outcome of infection), from both immersion regimes (constant immersion or tidal emersion), and on Days 0, 1, 2, 3, 4 and 7 post-OsHV-1 injection. As all oysters collected on Day 0 were OsHV-1 negative, the model does not include data for Day 0. Predicted means and their corresponding 95% confidence intervals from a generalised linear mixed model (GLMM) for OsHV-1 DNA, were back transformed to obtain mean OsHV-1 genomes per mg of tissue (n = 6-11 status/day/immersion regime).

Mean OsHV-1 concentration Factor (genome equivalents/mg) Geometric mean CI; lower-upper Immersion regime: Constant immersion 5.96×103 2.01×103–1.77×104 Tidal emersion 3.89×104* 1.14×104–1.33×105

Status: 2.09×103 9.31×102 – 4.72×103 Live 1.10×106** 2.22×105– 5.47×106 Dead *The OsHV-1 DNA quantity was higher in oysters under tidal emersion compared those under constant immersion during the experiment (p < 0.05).

** The OsHV-1 DNA quantity was higher in dead oysters compared to the live oysters that were sampled during the experiment (p < 0.05).

159

6.4.3 Quantification of total bacteria and total Vibrio

Differences in the total bacterial DNA quantity were observed between live and dead oysters and between gill and gut tissues of oysters (Tables 6.2, 3; p < 0.05). After allowing for variation due to immersion regime, time of sampling and tissue-type, approximately 10- fold higher load of total bacteria DNA was seen in oysters at the time of death compared to the oysters that were sampled alive (Tables 6.2, 6.3; p < 0.05). After allowing for variation due to the other factors, the total bacterial DNA was higher in gill tissue compared to gut tissue, after the OsHV-1 challenge (Table 6.3; p < 0.05). Considering the oysters that were challenged with OsHV-1, the quantity of total bacteria was lower for oysters maintained in tidal emersion, compared to those that were under constant immersion (Table 6.3; p < 0.05). However, the lower bacterial load in tidal emersion was not reflected in the cultivable fraction. There was no difference between the TCBC in OsHV-1 challenged oysters in the two immersion regimes (Table 6.4). The tidal system which resulted in a lower load of total bacterial DNA also had a higher OsHV-1 concentration and higher mortality of oysters. Moreover, the gill tissue of unchallenged oysters under tidal emersion had higher total bacterial load compared to that of OsHV-1 challenged oysters (p < 0.05). There was no difference in the TCVC (Table 6.5) and the quantity of total Vibrio spp. DNA (Table 6.6) between oysters maintained in the two different management systems. Further, there were no temporal patterns or differences in total bacterial DNA or total Vibrio DNA quantity, between the immersion regimes on any single day, across the sampling periods after OsHV-1 challenge.

160

Table 6.2 Results of Generalized Linear Mixed Model (GLMM) analysis of total bacterial DNA quantity in gill and gut tissues of oysters challenged with OsHV-1 and oysters injected with an OsHV-1 negative tissue homogenate (challenge), maintained under two different immersion regimes (immersion; constant immersion and tidal emersion). Tissue samples were collected on days 0, 1, 2, 3, 4 and 7 post-injection (days post-injection).

df1 df2 F-value p-value Fixed effects: Immersion 1 34 1.21 0.28 Challenged 1 34 2.58 0.12 Days post-injection 4 34 1.93 0.13 Tissue type 1 34 7.39 0.01 Estimate t-value df Adjusted p-value Pairwise contrast: Gill-gut 1.17 3.22 34 0.00

Outcome of infection 1 34 12 0.00 Estimate t-value df Adjusted p-value Pairwise contrast: Live-dead 1.03 3.46 34 0.00

Interactions: Immersion × Challenged 1 34 5.26 0.028 Estimate t-value df Adjusted p-value Pairwise contrast: Constant-challenged – tidal-challenged 1.39 3.78 34 0.00

Immersion × Tissue × Challenged 1 34 17.85 0.00 Estimate t-value df Adjusted p-value Pairwise contrast: Tidal-gill-challenged – tidal-gill-unchallenged 1.311 2.086 34 0.045 df, degree of freedom

161

Table 6.3 Total bacterial DNA quantity in oysters challenged with OsHV-1 and oysters injected with an OsHV-1 negative tissue homogenate. Total bacterial DNA was quantified by qPCR in gill and gut tissues from live oysters collected on Days 0, 1, 2, 3, 4 and 7 post- injection and in dead oysters sampled after the OsHV-1 challenge. Predicted means and their corresponding 95% confidence intervals from a generalised linear mixed model (GLMM) were back transformed for mean bacterial genomes per g of tissue (n = 5- 10/tissue/day/immersion regime). Factor Total bacteria (bacterial genomes per g of tissue) Geometric mean CI; lower-upper Immersion regime: Constant immersion 5.11×106* 1.17×106 – 2.24×107 Tidal emersion 1.08×106 2.19×105 – 5.32×106

Outcome of infection: Live 7.18×105 2.28×105 – 2.26×106 Dead 7.67×106** 1.44×106 – 4.09×107

Infection status: OsHV-1 challenged 5.75×107 2.35×107 – 1.31×108 Unchallenged 3.60×108 7.33×107 – 1.42×109

Tissue type: Gill 5.74×106*** 1.34×106 – 2.47×107 Gut 3.92×105 8.34×104 – 1.84×106

Interactions:

Gill OsHV-1 challenged 1.76×106 4.61×105– 6.71×106 Unchallenged 1.39×107 7.78×105– 2.49×108 Gut OsHV-1 challenged 1.60×106 2.84×105– 9.04×106 Unchallenged 2.92×106 6.62×105– 1.29×107

Constant immersion OsHV-1 challenged 6.17×106 1.94×105– 1.96×108 Unchallenged 4.65×106 1.46×106– 1.49×107 Tidal emersion OsHV-1 challenged 3.57×107 2.10×106– 6.05×108 Unchallenged 1.87×105 3.40×104– 1.04×106 * The total bacterial DNA quantity was higher in oysters maintained in constant immersion in water as opposed to those maintained in a tidal environment (p < 0.05). ** The total bacteria quantity was higher in dead oysters compared to the live oysters that were sampled during the experiment (p < 0.05). *** Overall, the total bacteria quantity was higher in gill tissue compared to gut tissue of oysters after the OsHV-1 challenge (p < 0.05).

162

Table 6.4 Total cultivable bacterial count (TCBC) in oysters challenged with OsHV-1 and in oysters injected with an OsHV-1 negative tissue homogenate and maintained in different immersion regimes (constant immersion or tidal emersion). TCBC was quantified using the whole soft-tissue mass (except the digestive gland) from live oysters (n = 10-15 day/immersion regime) collected on Days 0, 1, 2, 3, 4 and 7 post-injection. Predicted means and their corresponding 95% confidence intervals from a generalised linear mixed model (GLMM) were back transformed for number of colony forming units per gram (CFU/g).

Factor TCBC (CFU/g) Geometric mean CI; lower-upper Immersion regime: Constant immersion 1.77×104 1.17×104– 2.67×104 Tidal emersion 1.94×104 1.27×104– 2.96×104

Infection status: OsHV-1 challenged 2.60×104 1.94×104– 3.49×104 Unchallenged 1.32×104 7.94×103–2.20×104

Table 6.5 Total cultivable Vibrio count (TCVC) in oysters challenged with OsHV-1 and in oysters injected with an OsHV-1 negative tissue homogenate, maintained in different immersion regimes (constant immersion or tidal emersion). TCVC in whole soft tissue homogenates (except the digestive gland) from live oysters (n = 10-15 day/immersion regime) collected on Days 0, 1, 2, 3, 4 and 7 post-injection. Predicted means and their corresponding 95% confidence intervals from a generalised linear mixed model (GLMM) were back transformed for number of colony forming units per gram (CFU/g). Factor Geometric mean (CFU/g) CI; lower-upper Immersion regime: Constant immersion 3.15×103 1.02×103 – 9.17×103 Tidal emersion 7.62×103 2.58×103 – 2.25×104 Infection status: OsHV-1 challenged 4.65×105 1.83×105 – 1.10×106 Unchallenged 4.71×105 8.14×104 – 2.09×106

163

Table 6.6 Total Vibrio count in oysters challenged with OsHV-1 and in oysters injected with an OsHV-1 negative tissue homogenate, maintained in different immersion regimes (constant immersion or tidal emersion). Total Vibrio count was quantified by qPCR in gill and gut tissues from live oysters collected on Days 0, 1, 2, 3, 4 and 7 post-injection and in dead oysters sampled after the OsHV-1 challenge. Predicted means and their corresponding 95% confidence intervals from a generalised linear mixed model (GLMM) were back transformed for number of Vibrio genome equivalents per gram of tissue (n=5-10 day/immersion regime). Factor Geometric mean CI; lower-upper Immersion regime: Constant immersion 3.87×102 1.65×102 – 9.10×102 Tidal 7.96×102 3.10×102 – 2.05×103

Infection status: OsHV-1 challenged 2.25×104 6.11×103 – 7.13×104 Unchallenged 3.50×104 2.61×103 – 2.74×105

Outcome of infection Live 2.77×104 1.04×104 – 6.74×104 Dead 2.01×104 1.92×103 – 1.34×105 Tissue type: Gill 3.59×102 1.40×102 – 9.18×102 Gut 8.57×102 3.82×102 – 1.93×103

164

6.4.4 Changes in bacterial community structure after OsHV-1 challenge 6.4.4.1 High throughput 16S rRNA gene sequencing

Targeting the hypervariable V1-V3 region of the 16S rRNA gene, a total of 3,622,620 paired-end raw reads were obtained initially from the samples analysed, leaving 3,352,640 reads after quality control and bioinformatic processing. The number of reads per gill tissue sample (median: 6978; maximum: 67,425; minimum: 2640) was lower compared to that of gut tissue samples obtained from the same oyster (median: 40,007; maximum: 142,082; minimum: 6712) (p < 0.05). The reads were rarefied to 5584 per sample for gill tissue and 22143 per sample for gut tissue. Rarefaction curves showed saturation for most of the samples, indicative of a good coverage of diversity.

6.4.4.2 Changes in the bacterial community composition

The results of Chapter 5 indicated that similar changes to the oyster microbiome occurred during acclimation to the laboratory environment, under both immersion regimes. However, subsequent to the OsHV-1 challenge, the microbiome responded differently in the two immersion regimes. Further, this differential response was seen at the level of tissue type.

Gill microbiome

The gill microbiota of constantly immersed oysters challenged with OsHV-1 was dominated by phylum Proteobacteria throughout the experimental period. In contrast, phylum Spirochaetes dominated throughout the same period in oysters maintained in tidal emersion (Figure 3A and 3C). Moreover, phylum Spirochaetes dominated the gill microbiota of negative control oysters of both immersion regimes. Concurrent with the onset of mortality on Day 2 post-OsHV-1 injection, a decline in the alpha diversity occurred in the gill microbiota of live, constantly immersed oysters (observed OTUs: Day 1, 192.7 ± 20.7, Day 2, 143.4 ± 8.1; Figure 4A; p < 0.05). This decline in alpha diversity did not occur in the negative control oysters maintained with constant immersion. By contrast, there was an increase in alpha diversity of gill microbiota in live oysters kept in tidal conditions after challenge with OsHV-1 (Figure 4B; Table 6.7; p < 0.05). Such an increase in the alpha diversity was not observed in unchallenged oysters in the tidal conditions.

Despite the differences at OTU level, there were no significant temporal differences at phyla level in the gill microbiota of oysters challenged with OsHV-1 and kept in constant

165

immersion. Differences in abundance were observed at genus level for Vibrio (phylum Proteobacteria), Arcobacter (phylum Proteobacteria) and Sediminibacterium (phylum Bacteroidetes) under this immersion regime (Table 6.8). While Vibrio and Sediminibacterium declined in abundance, the number of reads for genus Arcobacter increased soon after the OsHV-1 challenge (p < 0.05). No such temporal differences in abundance were noted in negative control oysters at phyla or genera level. Unlike in constantly immersed oysters, the abundance of bacterial phyla in the gill microbiome of tidal oysters, changed following OsHV-1 challenge (p < 0.05). In general, phylum Proteobacteria, phylum Bacteroidetes, phylum TM6 and phylum Actinobacteria increased in relative abundance in this group (p < 0.05). For the negative control oysters in tidal system, only the phylum TM6 demonstrated an increase, during this period (p < 0.05). Moreover, increase in the abundance was also observed at genus level after OsHV-1 challenge, for the genera Sediminibacterium (phylum Bacteroidetes), Polaribacter (phylum Bacteroidetes) and Marinicella (phylum Proteobacteria) (Table 6.9). This increase did not occur in the negative control oysters.

Gut microbiome

The gut microbiota of oysters challenged with OsHV-1 was dominated by phylum Proteobacteria, in both immersion regimes (Figure 3B and 3D). Unlike the gill microbiota of the same oysters, the alpha diversity of gut microbiota did not change after the OsHV-1 challenge in both systems. Phylum level changes in abundance after OsHV-1 challenge reflected an increase in Proteobacteria and Spirochaetes and decrease in Bacteroidetes in the gut of oysters sampled alive from constant immersion (Figure 3B; p < 0.05). By contrast, phyla Bacteroidetes, Actinobacteria and Firmicutes increased in abundance in the negative control oysters maintained in constant immersion (p < 0.05). The gut microbiota of live oysters in tidal emersion, exhibited an increased abundance in phylum Fusobacteria and phylum TM6, after the OsHV-1 challenge (Figure 3D; p < 0.05). No significant changes in the abundance of bacterial phyla were observed in the gut microbiota of negative control oysters maintained in tidal emersion.

At genus level, the relative abundance of Shewanella (phylum Proteobacteria) increased in the gut microbiota of oysters maintained under constant immersion after OsHV-1 challenge (p < 0.05). Meanwhile the genus Shewanella and an unidentified genus of phylum TM6 increased in abundance in oysters challenged with OsHV-1 and were maintained in tidal emersion (p < 0.05). However, it is interesting to note that neither Vibrio spp. nor

166

Mycoplasma spp. changed in their abundance in the gut of either groups of oysters. In particular, the dominance of genus Mycoplasma during acclimation to the laboratory environment (Chapter 5) continued after the OsHV-1 challenge in tidal oysters. Moreover, genus Shewanella did not change in abundance in negative control oysters in both treatment groups. Instead, the genera Sediminibacterium (phylum Bacteroidetes) and Clostridium (phylum Firmicutes) increased in abundance in the negative control oysters under constant immersion, during this same period (p < 0.05). No such change in abundance was noted in the gut microbiota of negative control oysters in the tidal environment.

167

A 1

0.9

0.8 Proteobacteria Spirochaetes 0.7 Bacteroidetes Tenericutes 0.6 Unassigned TM6 0.5 Firmicutes Planctomycetes 0.4

Verrucomicrobia Relative Relative abundance (%) Other 0.3

0.2

0.1

0 D0 D0 D0 D0 D0 D1 D1 D1 D2 D2 D2 D2 D2 D2 D2 D3 D3 D3 D3 D3 D3 D3 D4 D4 D4 D4 D4 Days post OsHV-1 injection

B 1

0.9

0.8 Proteobacteria Bacteroidetes 0.7 Spirochaetes Tenericutes 0.6 Planctomycetes Actinobacteria 0.5 TM6 Unassigned 0.4 Firmicutes Relative Relative (%) abundance Cyanobacteria 0.3 Fusobacteria Other 0.2

0.1

0 D0 D0 D0 D0 D0 D1 D1 D1 D2 D2 D2 D2 D2 D2 D2 D3 D3 D3 D3 D3 D3 D3 D4 D4 D4 D4 D4 Days post OsHV-1 injection

168

0.9

0.8 Proteobacteria 0.7 Spirochaetes Bacteroidetes 0.6 Tenericutes Unassigned 0.5 TM6 0.4 Firmicutes

Relative Relative abundance (%) Fusobacteria 0.3 Other

0.2

0.1

0 D0 D0 D0 D0 D0 D1 D1 D1 D2 D2 D2 D2 D3 D3 D3 D3 D3 D4 D4 Days post OsHV-1 injection

D) 1

0.9 Proteobacteria 0.8 Spirochaetes Bacteroidetes 0.7 Tenericutes TM6 0.6 Firmicutes 0.5 Planctomycetes Fusobacteria 0.4 Actinobacteria Verrucomicrobia

Relative (%) Relative abundance 0.3 Unidentified 0.2 Unassigned Other 0.1

0 D0 D0 D0 D0 D0 D1 D1 D1 D2 D3 D3 D3 D3 D3 D4 D4 D4 Days post OsHV-1 injection

Figure 6.3 Taxa bar plots indicating the relative abundance of bacterial phyla in: A) gill microbiota; B) gut microbiota of oysters maintained with constant immersion in water, and C) gill microbiota; D) gut microbiota of oysters maintained in a simulated tidal environment in the laboratory. All oysters were challenged with OsHV-1 and were sampled on Days 0,1,2,3 and 4 post-viral injection. Results were analysed using QIIME2. Bacteria that could not be assigned to a particular phylum are categorized under ‘unassigned’, DNA sequences that could not identified under the kingdom Bacteria are categorized under ‘unidentified’ and phyla with a relative abundance of less than 5% and were not present in at least two samples are categorized as ‘other’.

169

Figure 6.4 Box and whisker plots representing the observed operational taxonomic units (OTUs) in gill microbiota of oysters maintained in: A) constant immersion in water and B) tidal emersion, after challenging with OsHV-1. The gill microbiota in constantly immersed oysters showed a drop in observed OTUS on Day 2 post-injection (p < 0.05) while that of tidal oysters showed a continuous increase in the number of observed OTUs, during the course of the viral infection (p < 0.05).

170

Table 6.7 Alpha diversity of gill microbiota (mean observed OTUs* ± SE) following OsHV-1 challenge in oysters maintained in tidal emersion. Microbiota were analysed from samples collected on Days 0, 1, 2, 3, and 4 post-injection (n=5/day). Days post-OsHV-1 injection Observed OTUs (mean ± SE) Challenged Unchallenged Day 0 63.8 ± 13.1 91 ± 16 Day 1 155 ± 23.5 172.8 ± 70 Day 2 192 ± 14.8 141 ± 9 Day 3 230.4 ± 38.1 121 ± 7 Day 4 243.5 ± 40.5 111.5 ± 33.5 * OTU, operational taxonomic units

Table 6.8 Temporal changes in the mean relative abundance of selected dominant genera of the gill microbiota of oysters maintained with constant immersion in water, after challenging with either OsHV-1 or OsHV-1 negative tissue homogenate. Microbiota were analysed from gill tissue samples (n=5/infection status/day) collected on Days 0, 1, 2, 3, and 4 post- injection. Relative abundance between days were analysed using a generalized linear model (GzLM). Genus Mean relative abundance (%) Day 0 Day 1 Day 2 Day 3 Day 4 Significance OsHV-1: Vibrio 11.09 2.90 2.93 4.77 2.31 p = 0.02 Mycoplasma 0.70 1.93 0.33 2.33 0.39 p > 0.05 Arcobacter 0.18 5.93 0.50 0.76 0.42 p = 0.01 Spirochaeta 0.36 0.13 15.38 11.64 9.89 p > 0.05 Sediminibacterium 6.41 4.21 1.95 0.30 0.50 p = 0.03

OsHV-1 negative: Vibrio 2.18 3.19 4.51 2.06 1.31 p > 0.05 Mycoplasma 0.53 0.16 0.10 0.00 0.31 p > 0.05 Arcobacter 0.23 1.06 0.64 0.08 0.13 p > 0.05 Spirochaeta 0.48 0.00 0.00 28.48 0.31 p > 0.05 Sediminibacterium 1.76 6.71 0.67 1.67 0.27 p > 0.05

171

Table 6.9 Temporal changes in the mean relative abundance of selected, dominant genera of gill microbiota oysters maintained in tidal emersion, following challenge with either OsHV-1 or OsHV-1 negative tissue homogenate. Microbiota were analysed from gill tissue samples (n=5/infection status/day) collected on Days 0, 1, 2, 3, and 4 post-injection. Genus Mean relative abundance (%) Day 0 Day 1 Day 2 Day 3 Day 4 Significance OsHV-1: Vibrio 22.09 14.96 29.06 14.8 7.0 p > 0.05 Polaribacter 0.12 0.46 9.54 1.32 9.24 p = 0.00 Arcobacter 0.09 0.78 0.44 0.77 0.08 p > 0.05 Marinicella 0.10 0.01 0.67 0.40 0.91 p = 0.00 Sediminibacterium 1.59 4.38 0.31 0.51 0.23 p = 0.00

OsHV-1 negative: Vibrio 1.49 1.58 1.07 1.27 7.0 p > 0.05 Polaribacter 1.07 0.49 0.17 0.11 0.23 p > 0.05 Arcobacter 0.0 0.72 1.29 0.54 0.73 p > 0.05 Marinicella 0.03 0.07 0.22 0.02 0.25 p > 0.05 Sediminibacterium 0.23 1.36 2.25 0.24 0.09 p > 0.05

6.5 Discussion

Oysters farmed in intertidal estuarine environments can be exposed to extreme environmental conditions such as extreme temperature and decreased ambient oxygen (Allen and Burnett, 2008). The microbiome associated with oysters is influenced by these environmental conditions and the changes may contribute or predispose oysters to disease. In Chapter 5 of this thesis, I observed changes in the Pacific oyster microbiome, during acclimation to the laboratory environment, and the changes that occurred were not different between the immersion regimes (constant immersion or tidal). This study focused on the microbiome dynamics of Pacific oysters after an OsHV-1 challenge in a laboratory environment, under two different immersion regimes. In line with the findings of de Lorgeril et al. (2018), we observed changes in the composition and the quantity of bacteria in the microbiome, after the OsHV-1 challenge. However, unlike the changes observed with laboratory acclimation in Chapter 5 of the thesis, the oysters under different immersion regimes responded differently to the OsHV-1 challenge. Moreover, the changes were generally not observed in oysters that were not challenged with the virus and maintained in the same conditions. The dominant bacterial phyla of the gill and gut microbiome were common between oysters maintained in either immersion regime, but to varying degrees. These phyla and the constituent genera responded differently to the OsHV-1 infection.

172

Furthermore, increased abundance was noted for rare taxonomic groups in response to the OsHV-1 challenge.

The oysters maintained with tidal emersion had a much higher mortality coupled with a higher OsHV-1 concentration, compared to those under constant immersion. There was also a decrease in the total bacteria quantity in the period after the OsHV-1 challenge, compared to the bacterial load in OsHV-1 challenged oysters under constant immersion where fewer oysters showed evidence of disease. de Lorgeril et al. (2018) reported an increase in the total bacterial load in oysters, in a laboratory experiment involving cohabitation with OsHV-1 donor oysters. Similar to the TCBC data of this study, Petton et al. (2015b) reported that there was no change in the TCBC in oysters deployed in the Bay of Brest, France, during an oyster mortality outbreak. The bacterial concentration is usually higher in moribund or dead oysters compared to live oysters challenged with OsHV-1 (Pathirana et al., 2019b). Prolonged periods of valve closure during tidal emersion resulted in hypercapnia and reduction in pH in oyster tissues (Allen and Burnett, 2008). Although the pH changes in oyster tissues were not measured in this study, a similar pH reduction may have occurred in this study, affecting the multiplication and growth of microbiota of tidal oysters. However, a reduction of total bacterial load was not recorded during acclimation of oysters to the laboratory environment using tidal emersion (Chapter 5). Moreover, such a reduction was not seen in negative control oysters maintained in tidal emersion, in the present study. Thus, the lower total bacterial count in OsHV-1 challenged, tidal oysters can be considered a combined effect of tidal emersion and OsHV-1 challenge, resulting in a reduction of the non-cultivable fraction of the microbiome. OsHV-1 infections are followed by opportunistic bacterial infections causing dysbiosis (loss of bacterial diversity and proliferation of a few OTUs) in the Pacific oyster microbiome (de Lorgeril et al., 2018). Although not measured in this study, the hypercapnia and reduction of pH in oyster tissues has been reported to occur during tidal emersion (Allen and Burnett, 2008); perhaps this occurred in the present study and limited the growth of fastidious bacteria present in the microbiome.

Similar to the observations of Chapter 5 of this thesis, inter-oyster heterogeneity in bacterial community composition was observed in both the gill and gut microbiota of oysters, at the beginning of the experiment. Subsequently, the gill and gut microbiome reacted to the OsHV-1 challenge and this reaction was different between the oysters from the two different immersion regimes. A low bacterial diversity in gill microbiota was seen at the onset of

173

mortality in the constantly immersed oysters. This can be a reflection of dysbiosis in sub- clinically infected oysters that were sampled at this point. Loss of bacterial diversity and proliferation of bacteria representing a few OTUs (dysbiosis) has repeatedly been associated with impaired health in oysters secondary to viral, bacterial and parasitic diseases (de Lorgeril et al., 2018; Garnier et al., 2007; Green and Barnes, 2010; Lokmer and Wegner, 2015). In the present study, the decline in bacterial diversity in the gill of constantly immersed oysters was coupled with a reduction in the relative abundance of Vibrio spp. This was preceded by an increase in the abundance of Arcobacter on Day 1 post-OSHV-1 injection. Arcobacter spp. were reported to be abundant in the normal microbiota of Tiostrea chilensis (Chilean oyster) (Romero et al., 2002). However, the Ɛ-Proteobacteria that includes Arcobacter, are usually rare in coastal seawater (Campbell et al., 2011), and in the oyster stomach (King et al., 2012), gut (Trabal Fernández et al., 2014) and gill microbiota (Wegner et al., 2013). Arcobacter being microaerophilic (Vandamme and De Ley, 1991), the abundance seen in the present study could possibly be the result of hypoxia due to disease- induced reduction of filtration activity (Lokmer and Wegner, 2015; McHenery and Birkbeck, 1986). However, increased abundance of Arcobacter spp. in diseased or stressed marine organisms other than bivalves also suggests a role for Arcobacter as an opportunistic pathogen. The increased abundance of Arcobacter in starved abalones (Tanaka et al., 2004) and necrotic sponges (Fan et al., 2013) are two examples. Meanwhile, Lokmer and Wegner (2015) have also highlighted the dominance of Arcobacter in moribund oysters, contrary to their expectation of an abundance of Vibrio after an experimental Vibrio challenge in Pacific oysters. A recent study which analysed the microbiome of Pacific oysters from recurrent mortality episodes in Europe, identified a clear loss of alpha diversity with a concomitant increase of Arcobacter spp. and Vibrio spp. in Vibrio aestuarianus infections in adult oysters (Lasa et al., 2019). The same study reports a reduction in alpha diversity associated with abundance of only Vibrio, in OsHV-1 infected oysters. Dominance of Arcobacter spp. in unhealthy animals was accompanied by low microbial diversity, in contrast to the diverse microbiomes of oysters that survived infection and negative controls (Lokmer and Wegner, 2015). In an experimental OsHV-1 challenge an initial OTU fraction of 2% (involving Vibrio and Arcobacter) increased to 59% in the oyster microbiome, with the onset of mortality (de Lorgeril et al., 2018). With this background, and in light of the present findings, the potential role of Arcobacter in oyster disease requires further investigation. The pathogenicity of Arcobacter per se, and its role in polymicrobial infections, merits study.

174

The changes that were observed in the gill microbiome of oysters maintained under the two different immersion regimes were same, during acclimation to the laboratory environment (Chapter 5). However, despite originating from the same batch of oysters and being maintained in artificial seawater in a laboratory environment, the bacterial community composition of the gill tissues was different in oysters maintained in constant immersion compared to tidal emersion, after the OsHV-1 challenge. While the gill microbiota in constant immersion was dominated by phylum Proteobacteria, it was dominated by phylum Spirochaetes in the tidal environment. Unlike the reduction in bacterial diversity with constant immersion (in response to the viral infection), a continuous increase in the bacterial diversity was observed in tidal oysters, during the experiment. As dysbiosis is commonly observed before a disease (Lasa et al., 2019; Pathirana et al., 2019b), this increase may be an indication of an additional response of the microbiota to the environmental changes induced by periodic emersion. Microbial communities exposed to repeated fluctuations have been identified to be more diverse and more active than communities exposed to a constant condition (Shade et al., 2012). Moreover, rare microbial taxa may rapidly respond to changes in the environment and become more abundant (Shade et al., 2012). In line with this, the present study observed an increase in the relative abundance of rare bacterial genera (Sediminibacterium, Polaribacter and Marinicella) and members of the phylum Actinobacteria, in the gill microbiota of tidal oysters which was not noted in constant immersion. Although the genus Arcobacter was also expected to increase in abundance owing to its microaerophilic nature, it did not increase in diseased tidal oysters.

The fitness of the host is strongly associated with the stability of its microbiome and a large proportion of microbiome dynamics can be attributed to immediate environmental conditions (Lokmer et al., 2016b). Thus, the greater mortality seen in tidal oysters challenged with OsHV-1 of this study, can also be related to the instability of the oyster microbiome caused by prolonged periods of emersion. Stress-induced changes in the microbiome composition, linked to immunosuppression were considered responsible for a polymicrobial disease (Lasa et al., 2019). It is important to note that oysters under tidal emersion in the natural environment are exposed to additional stress factors such as desiccation with the sunlight (Allen and Burnett, 2008; Evans et al., 2019). However, as a result of the reduced exposure to OsHV-1, oysters subjected to prolonged emersion demonstrated a reduced mortality in the field (Paul-Pont et al., 2013b; Whittington et al., 2015b).

175

Further to the differences seen in the microbiome after the OsHV-1 challenge, phylum Fusobacteria increased in the gut microbiota of tidal oysters. Being strict anaerobes, increase in phylum Fusobacteria, supports the notion that the tidal oysters experienced anaerobic environments inside their shells owing to the prolonged periods of valve closure. The response was different in the gut microbiota of oysters in constant immersion where the relative abundances changed in the dominant phyla Proteobacteria, Bacteroidetes and Spirochaetes, with the viral infection including a reduction in the abundance of phylum Bacteroidetes. The latter is usually a dominant phylum in Pacific oyster gut microbiota (Pathirana et al., 2019a) and has been reported to decrease in abundance in the face of disturbance (Wegner et al., 2013). At the genus level, Shewanella (phylum Proteobacteria) increased in abundance in the gut of both constantly immersed oysters and tidal oysters. Genus Shewanella comprises facultatively anaerobic bacteria and it has been reported to dominate the gut microbiome of oysters (King et al., 2012) while being part of the gut microbiome of abalones (Kim et al., 2007). Further, Shewanella are considered as opportunistic pathogens in aquatic animals and humans (Aguirre et al., 1994; Brink et al., 1995). In this context, Shewanella can be considered an opportunistic pathogen in the present study where the relative abundance increased in the gut microbiota in both treatment groups. Recently, de Lorgeril et al. (2018) demonstrated that oyster mortality is a polymicrobial disease with an initial step of OsHV-1 infection with a subsequent bacteraemia caused by opportunistic bacteria. Thus, one or many species/strains of bacteria that increase in abundance following the viral challenge, may be a component of this polymicrobial disease. Despite the increase of Shewanella, gut microbiota of tidal oysters was dominated by genus Mycoplasma at the genus level, throughout the experimental period. I observed the abundance of Mycoplasma in the gut microbiota of oysters in a previous trial (Chapter 5). Although the physiological and ecological significance is not exactly known, the stomach microbiome in Eastern oysters (C. virginica) and the digestive gland of Sydney rock oyster (Saccostrea glomerata) have also been reported to be dominated by Mycoplasma (King et al., 2012). According to Zurel et al. (2011) and Roterman et al. (2015), the gut microbiota in oysters generally consists of transient bacteria which can also become opportunistic pathogens (King et al., 2012). Thus, in the present study, the increased Shewanella population may act alongside Mycoplasma, in the disease process.

In the present study, a reduction of the relative abundance of Vibrio was noted in the gill microbiota of constantly immersed oysters, at the onset of mortality. As previously

176

mentioned, Lokmer and Wegner (2015) identified reduced abundance of Vibrio spp. in the haemolymph after an experimental Vibrio challenge. Pathirana et al. (2019b) demonstrated an increase in Vibrio spp. in oyster tissues in an experimental OsHV-1 challenge which analysed whole soft tissue homogenates. While the commensal Vibrio spp. of Pacific oysters seem to play an opportunistic role in mortality associated with young oysters, pathogenic Vibrio spp. such as Vibrio aestuarianus are directly associated with adult Pacific oyster mortality (Lasa et al., 2019). Moreover, co-detection of OsHV-1 with V. aestuarianus and V. splendidus has also been reported in mass mortality outbreaks in Pacific oysters in Europe (EFSA, 2010; Solomieu et al., 2015).

The oysters were maintained in ASW in the present study. As such, the oysters had a minimal chance of acquiring external bacteria apart from the nitrifying bacteria that were present in the biological filters and bacteria that may be acquired from the commercial algae feed (if any). However, it should be noted that the microbiome composition of oysters maintained in constant immersion can be further differentiated compared to the oysters under tidal emersion, by prolonged exposure and increased acquisition of bacteria present in seawater, under natural conditions. This mechanism may add to the differences in microbiome composition in oysters maintained under constant immersion in natural, field conditions producing a different disease outcome.

6.6 Conclusion

The oyster microbiome responded differently to the OsHV-1 challenge, depending on the immersion regime. The differences in oyster physiology that were induced by periodic emersion may have facilitated this differential response. The differential microbial response may have contributed to the polymicrobial pathogenesis, leading to differential mortality between constantly immersed and tidally emersed oysters.

Although the gill microbiome was more stable than the gut microbiome in healthy oysters, the gill microbiome responded to the OsHV-1 challenge differently, between the two immersion regimes. The gut microbiota also changed with the OsHV-1 infection with differences between the two immersion regimes. While the overall bacterial diversity reduced in the gill microbiome of constantly immersed oysters, the diversity increased in tidal oysters, with the OsHV-1 challenge. Unlike with laboratory acclimation, the effects of prolonged emersion periods on the oyster microbiome was evident with the OsHV-1 challenge. Further

177

studies are required to assess the pathogenicity of Arcobacter as well as to assess the role of rare taxa in the oyster microbiome during mortality associated with OsHV-1. The differential response to the OsHV-1 challenge can be summarized as follows (Table 6.10):

Table 6.10 Summary of differential response to the OsHV-1 challenge under different immersion regimes Immersion Mortality OsHV-1 Bacterial -diversity Abundance of regime load count dominant genera Constant immersion Low Low High Decreased Increasing: Arcobacter Decreasing: Vibrio Sediminibacterium Tidal emersion High High Low Increased Increasing: Polaribacter Marinicella Sediminibacterium

178

CHAPTER 7

Influence of seawater temperature on the Pacific oyster (Crassostrea gigas) microbiome and its impact on Ostreid herpesvirus-1 (OsHV-1) infection

7.1 Abstract

Temperature is a strong determinant of microbiome composition associated with marine bivalves while Pacific oyster mortality syndromes are associated with elevated seawater temperature. Oysters living in intertidal estuarine environments are exposed to higher temperatures during the daytime, leading to heat stress. This may cause shifts in the oyster microbiome composition towards pathogen-dominated communities. Understanding the impact of seawater temperature on the oyster microbiome will present avenues to improve oyster health. The aims of the present study were to: 1) analyse the impact of different seawater temperature profiles (constant at 21°C, 22°C, or 26°C and with diurnal fluctuation between 22°C and 26°C) on the microbiome of Pacific oysters; 2) to investigate the association of temperature and microbiome on the response of oysters to an OsHV-1 challenge carried out in a controlled laboratory environment. Triploid Pacific oysters at 15-16 months of age (n = 450) were acclimated to tanks with the 4 different temperature profiles. They were challenged with OsHV-1 by injection and samples of live oysters were taken before and after acclimation and at 2 and 10 days after OsHV-1 exposure. In addition, dead oysters were sampled at the time of death. OsHV-1 DNA, total bacterial DNA and Vibrio DNA were quantified by qPCR. Bacterial 16S rRNA gene (V1-V3) was sequenced from gill microbiota. The highest mortality (84.4%) occurred in oysters at 26°C and these and the oysters with the 22/26°C dynamic temperature had higher OsHV-1 DNA loads compared to those at the lower constant temperatures (p < 0.05). The total bacterial quantity in gill did not change after acclimation to any of the temperature profiles but changed after the OsHV-1 challenge in oysters with a diurnal fluctuation of water temperature. The alpha diversity in the gill microbiome did not change after acclimation to different temperatures but increased after the OsHV-1 challenge in oysters with constant 21°C and dynamic 22/26°C water temperature profiles. The beta diversity changed both after acclimation to different temperatures and after the OsHV-1 challenge. There was a reduction in the genus Arcobacter (phylum Proteobacteria) after acclimation to all temperature profiles (p < 0.05). Phylum Proteobacteria and phylum Bacteroidetes increased in abundance after OsHV-1 challenge. The highest abundance of Vibrio was seen in OsHV-1 challenged oysters at 26°C and was

179

associated with higher OsHV-1 DNA (p < 0.05). In conclusion, water temperature profiles altered the oyster microbiome. The degree and nature of these changes varied with the temperature profile and were reflected in differences between bacterial genera. The opportunistic role of Vibrio in OsHV-1-associated oyster mortality appeared to be facilitated by higher seawater temperature. Higher oyster mortality was associated with a higher water temperature, higher OsHV-1 load and the highest Vibrio concentration.

Keywords: Pacific oyster, Crassostrea gigas, microbiome, temperature, laboratory acclimation, Ostreid herpesvirus 1

7.2 Introduction

Host-associated microbiomes are generally diverse and differ from the microbiomes of the surrounding environment (Adair and Douglas, 2017; Ley et al., 2008). Despite considerable research considering the microbiomes of various vertebrates and invertebrates (Erwin et al., 2012; Flores-Higuera et al., 2019; Littman and Pamer, 2011; Wagner- Mackenzie et al., 2015), the processes that shape the microbiome composition are not completely understood. Nevertheless, it is known that host-associated microbiomes are dynamic and respond to various internal factors such as the age of host, within-microbiome interactions as well as external factors (Adair and Douglas, 2017; Mancuso et al., 2016; Meyer et al., 2016; Shade et al., 2013; Yatsunenko et al., 2012). The Pacific oyster microbiome is influenced by the environment as these animals are ectothermic and do not possess an acquired immune system (Green et al., 2014; Lokmer et al., 2016b). Microbiome is the collection of all microbial communities that are associated with different tissues of the oyster. Studies have shown that the Pacific oyster microbiome changes in response to changes in the pH (Flores-Higuera et al., 2019), temperature (Lokmer and Wegner, 2015) of seawater, stress (Lokmer and Wegner, 2015), immediate environment (Pathirana et al., 2019b), translocation (Lokmer et al., 2016a; Lokmer et al., 2016b), and antibiotic treatment (Lokmer et al., 2016b).

Temperature is a key factor that influences bacterial growth and persistence (Zwietering et al., 1994). Moreover, temperature can influence the expression of virulence factors in bacteria (Kimes et al., 2012). Seawater temperature is a strong determinant of bacterial populations inhabiting seawater (Lokmer and Wegner, 2015) and an important external factor influencing host associated microbiomes (Lokmer and Wegner, 2015; Roterman et al., 2015). For

180

example in sponges, the symbiotic microbial community can be replaced by pathogenic microbes in the event of elevated seawater temperature (Cebrian et al., 2011; Maldonado et al., 2010). Disruption of core microbial communities can lead to an increase in rare microbial taxa, resulting in increased heterogeneity in the microbiome composition with increased seawater temperatures (Erwin et al., 2012). The core microbiome comprise of microbial taxa that are found in high prevalence in a host population or species (Adair and Douglas, 2017). Higher heterogeneity of the microbiome was also observed in spondylus oysters (Spondylus spinosus) in summer when the seawater temperature was higher than 30°C (Roterman et al., 2015). On the other hand, several studies have observed loss of bacterial diversity in the oyster microbiome in the face of diverse stress factors, indicating the complex nature of the microbiome (Lasa et al., 2019; Lokmer and Wegner, 2015; Lokmer et al., 2016a). In addition to the microbiome dynamics, temperature increases have also been linked to enhanced disease expression by increasing pathogen development and host susceptibility to disease (Burge et al., 2014; Harvell et al., 2002). In particular, OsHV-1 disease outbreaks in Pacific oysters have been associated with elevated seawater temperatures (Garcia et al., 2011; Paul- Pont et al., 2014; Whittington et al., 2019).

Pacific oyster mortality events associated with OsHV-1 infection require environmental conditions that are favourable for the development of disease (de Kantzow et al., 2017; Evans et al., 2019; Petton et al., 2015a). The severity of disease is influenced by host factors such as genetics (Dégremont, 2011; Samain et al., 2007), the size and age of oysters (Azéma et al., 2017a; Hick et al., 2018; Renault et al., 1994b), and environmental factors such as elevated seawater temperature, salinity, pH and nutrient levels in seawater (de Kantzow et al., 2016; Delisle et al., 2018; Petton et al., 2013; Soletchnik et al., 2007). In Europe, OsHV-1 disease outbreaks usually occurred at seawater temperature at or above 16°C (Clegg et al., 2014; Pernet et al., 2012; Petton et al., 2015a; Renault et al., 2014) and oyster mortality did not occur at temperatures above 26°C (Pernet et al., 2012). However, this disease in Australia occurred when the water temperature was 4-5°C warmer than that recorded in Europe (Paul- Pont et al., 2014; Whittington et al., 2019).

During summer, oysters living in intertidal estuarine environments are exposed to diurnal temperature fluctuations which can include extremely high temperatures. Heat stress can favour shifts in the microbiome towards pathogen-dominated communities (Boutin et al., 2013). Moreover, it can facilitate the increase of opportunistic pathogens in the microbiome

181

(Lokmer and Wegner, 2015). Increased seawater temperatures (>20°C) have been associated with Pacific oyster mortality outbreaks in Port Stephens, NSW in the absence of OsHV-1 (Go et al., 2017). Microbiome analysis of affected oysters from this outbreak revealed an increase of rare microbiota (low-abundant bacteria in the microbiome) belonging to the genus Vibrio (King et al., 2018a). Green et al. (2018) demonstrated that mortality triggered initially by increasing seawater temperature to 25°C, was reduced (from 77.4% to 4.3%) when the oysters were treated with a broad-spectrum antibiotic (combination of penicillin and streptomycin), indicating a role of bacteria in temperature-associated mortality. While the increase of temperature was associated with a 324-fold of Vibrio harveyi, a reduction from 40.5% to 2.2% was reported after the antibiotic treatment. There is an emerging view of polymicrobial pathogenesis causing Pacific oyster mortality disease which can be attributed to the microbiome disruption seen in these mortality events(King et al., 2018b; Lasa et al., 2019). Studies carried out on recurrent Pacific oyster mortality outbreaks that were associated with either OsHV-1 or Vibrio aestuarianus, revealed signs of bacterial community disruption with the presence of potentially pathogenic bacteria (Lasa et al., 2019). Understanding the impact of seawater temperature on the oyster microbiome may lead to new ways of controlling oyster mortality.

The aims of the current study were to: 1) determine the impact of different seawater temperature profiles on the microbiome of Pacific oysters, and; 2) evaluate the course of OsHV-1 associated disease under different temperature profiles in a controlled laboratory environment and determine a potential role of the microbiome in the disease outcome.

7.3 Materials and Methods

7.3.1 Oysters

Hatchery-reared, single seed, triploid Pacific oysters were used in this study (Batch SPL17C; Shellfish Culture Ltd., Tasmania; n = 450). These oysters were grown under commercial farming conditions in Upper Patonga Creek, Hawkesbury River, NSW (Broken Bay Oysters). For the purpose of this experiment, oysters were recruited at 15-16 months of age (shell length 50-80 mm; weight 17.5 – 44.5 g) and were transported to a physical containment level 2 aquatic animal facility at the University of Sydney, Camden, NSW. These oysters were not considered to have been previously exposed to OsHV-1 based on the following criteria: the batch were certified negative for OsHV-1 by the competent government

182

authority at the time of interstate transport from the hatchery in TAS to NSW; long-term active surveillance indicated freedom from disease caused by OsHV-1 in this water body (Whittington et al., 2019); and a random sample of oysters tested negative for OsHV-1 by qPCR prior to the trial (n=30).

7.3.2 Experiment design and aquarium management

Initially, the oysters were purged overnight in artificial seawater (ASW; Red Sea® salt), after arrival at the experimental facility. The oysters were then randomly allocated to 24 individually aerated tanks containing 12 L of ASW at 30 ± 1 ppt salinity. The water of each tank was not allowed to mix with each other. Each tank contained a biofilter and an aeration unit. The biofilters had been seeded with nitrifying bacteria obtained from a marine aquarium housing juvenile barramundi (Lates calcarifer) prior to this study. Initially, 20 oysters were maintained per tank and on perforated plastic racks that ensured adequate circulation and access to food. The water temperature in tanks was maintained by using an external water bath adapted from a recirculation system described in Section 2.1.1. The temperature of the water bath was maintained by the combined actions of thermostatically controlled aquarium heaters (AquaOne®, NSW, Australia); aquarium heater chiller units (HC-300a, Hailea Aquarium chiller); and air temperature maintained at 24°C ± 2°C by air conditioning.

Six tanks were allocated to each temperature profile: constant at 21°C, 22°C, or 26°C and diurnal variation between 20 and 26°C (Figure 7.1). The temperature 21°C was selected to represent a threshold temperature for OsHV-1 infections in Australia (Paul-Pont et al., 2014; Whittington et al., 2019) while 22°C was chosen to investigate any potential changes in the microbiome response with a subtle increase in temperature. Comparatively higher oyster mortality in OsHV-1 infection studies and the observation as an upper threshold temperature led to the selection of 26°C (de Kantzow et al., 2016; Delisle et al., 2018). The 20 and 26°C treatment was selected to analyse the response of the microbiome to diurnal temperature dynamics in the field. This dynamic temperature profile was achieved by setting the heater and chiller units to 26°C at 9am each morning and 22°C at 6pm each evening. During acclimation the water temperature in all tanks was set at 20°C and for 2 days. For the static 21°C, 22°C and 26°C treatments, the temperature was increased by 1°C.d−1 until reaching the target temperature. For the 22/26°C dynamic temperature, the highest temperature in the cycle was increased by 1°C.d−1, until it reached 26°C.Temperature data loggers (Thermocron)

183

were placed in two randomly selected tanks of each temperature treatment, to record the water temperature every 30 min.

Oysters were fed with a maintenance ration (2 mL.tank.day-1) of commercial algae concentrate (Shellfish Diet 1800, Reed Mariculture) and were provided with 12 h of artificial light each day. Total ammonia nitrogen (TAN) and pH in tank water was measured in opportunistically selected tanks of each temperature treatment, every day, using an API® Marine Saltwater Master Test kit. The water quality was maintained at target levels (TAN<2 ppm, pH range 8.0–8.2) by water exchange and the addition of sodium bicarbonate to the system to adjust the pH when it was ≤7.8. Oysters were acclimated to the tank environment for a total of 14 days, including the period of temperature adjustment.

CONTROL CONTROL OsHV-1 OsHV-1 OsHV-1 OsHV-1 (21 0C) (21 0C) 0 0 0 0 System 1 (21 C) (21 C) (21 C) (21 C) n=20 n=20 n=20 n=20 n=20 n=20 REPLICATE 1 REPLICATE 2 REPLICATE 1 REPLICATE 2 REPLICATE 3 REPLICATE 4

CONTROL CONTROL OsHV-1 OsHV-1 OsHV-1 OsHV-1 (22 0C) (22 0C) (22 0C) (22 0C) (22 0C) (22 0C) System 2 n=20 n=20 n=20 n=20 n=20 n=20 REPLICATE 1 REPLICATE 2 REPLICATE 1 REPLICATE 2 REPLICATE 3 REPLICATE 4

CONTROL CONTROL OsHV-1 OsHV-1 OsHV-1 OsHV-1 (26 0C) (26 0C) (26 0C) (26 0C) (26 0C) (26 0C) System 3 n=20 n=20 n=20 n=20 n=20 n=20 REPLICATE 1 REPLICATE 2 REPLICATE 1 REPLICATE 2 REPLICATE 3 REPLICATE 4

CONTROL CONTROL OsHV-1 OsHV-1 OsHV-1 OsHV-1 0 0 0 0 0 0 (22/26 C) (22/26 C) (22/26 C) (22/26 C) (22/26 C) (22/26 C) System 4 n=20 n=20 n=20 n=20 n=20 n=20 REPLICATE 1 REPLICATE 2 REPLICATE 1 REPLICATE 2 REPLICATE 3 REPLICATE 4

Figure 7.1 Schematic representation of the experiment design showing the allocation of oysters (C. gigas) across the 4 temperature profiles (constant at 21 ºC, 22 ºC, 26ºC and diurnal fluctuation between 22 ºC and 26ºC). Each square indicates an individual replicate tank which housed 20 oysters at the beginning of the experiment. Each system consisted of two control tanks in which the oysters were injected with an OsHV-1 negative tissue homogenate when the oysters in the other tanks were injected with OsHV-1.

184

7.3.3 Challenge with OsHV-1

The oysters were challenged with OsHV-1 using a cryopreserved, 0.2 µm filtered oyster 1/10 w/v tissue homogenate (Evans et al., 2015). The tissue homogenate was prepared from diseased oysters with confirmed OsHV-1 infection sampled during a mortality event in 2011 in the Georges River, NSW, Australia (Paul-Pont et al., 2013a). This inoculum was stored as multiple aliquots at -80°C with 10% v/v fetal bovine serum (Gibco) and 10% v/v glycerol and has been used in previous experimental infections (Evans et al., 2015). The negative control was prepared from a cryopreserved, filtered tissue homogenate prepared from apparently healthy Pacific oysters from an OsHV-1 and disease-free population (Evans et al., 2015). Inocula were thawed at 4°C and diluted 1:100 in sterile, artificial seawater to obtain the viral dose stated below. For injection, the oysters were first immersed in a solution of MgCl2 (50 g/l) for 4–6h until relaxation of the adductor muscles caused the valves to open. A 100 μl aliquot of the diluted OsHV-1 inoculum (1.34 ×106 viral copies per 100 μL) was injected into the adductor muscle, using a 1 ml syringe and a 25-gauge needle. After injection, oysters were rinsed in ASW and transferred to their tanks. Two tanks from each temperature treatment were allocated for negative control oysters. Physical separation and procedural care were taken to prevent cross contamination between OsHV-1 challenged tanks and negative control tanks. It should be noted that the water temperature was at 26°C in the tanks with the dynamic temperature profile, when OsHV-1 injections were made, in the middle of daytime.

7.3.4 Sampling

Oysters were sampled before (n=12) and after the 14-day acclimation (n=48; 2 oysters from each tank). This sampling was repeated 48 h after challenging the oysters with OsHV-1 (n=12 per temperature profile). These samples included 4 control oysters from each treatment that were injected with the OsHV-1 negative tissue homogenate. The random sampling procedure described in Section 2.3.1.1 was followed for sampling live oysters on each occasion. Oysters were inspected twice daily at 9:00 am and 3:00 pm and any dead oysters according to the definition in Section 2.3.1.2 were immediately sampled. All oysters that survived the infection challenge and the negative control oysters were sampled on Day 10 post-injection. All sampled oysters were maintained at 4°C until tissue dissection within 1 h. The cumulative mortality for each temperature treatment group was calculated according to

185

the method described by Whittington et al. (2015b) which takes into account the number of live oysters sampled during the trial.

7.3.5 Molecular quantification of OsHV-1

Each oyster was carefully shucked and gill and mantle tissue were collected and stored, as described in Section 2.4.1, for quantification of OsHV-1. These tissue samples were then homogenised and the final supernatant was stored, as described in Section 2.6.1. Nucleic acids were extracted from these supernatants, purified and stored, as described in Section 2.7.1. The number of copies of the B-region of the OsHV-1 genome was determined by a qPCR assay as described in Section 2.8.

7.3.6 Molecular quantification of bacteria

Gill tissue samples were collected and stored as described in Section 2.4.2. Nucleic acids were purified from gill tissue samples, as described in method B of Section 2.7.2 and total Vibrio spp. DNA was quantified by qPCR as described in Section 2.9. Total bacteria DNA in gill tissue was quantified by qPCR as described in Section 2.10.

7.3.7 Microbiome analysis by high throughput 16S rRNA gene sequencing

Nucleic acid extracts from gill tissues (n=94) were selected to represent oysters before acclimation, from all 4 temperature treatments and representing each tank in each treatment and, OsHV-1 challenged oysters and negative control oysters from each temperature treatment. The bacterial community composition of each extract was identified by high- throughput sequencing of the hypervariable V1-V3 region of 16S rRNA gene as described in Section 2.11.

The quality of the raw sequence data was assessed as described in Section 2.11.3 and formatted for QIME2 analysis according to the methods described in Sections 2.11.4 to 2.11.6. The processed sequence data were analysed using Quantitative Insights into Microbial Ecology Version 2 (QIIME2) Software Suite (2018.11 release), as described in Sections 2.11.7 to 2.11.9. The number of observed OTUs was used as the parameter to assess alpha diversity of samples. The dissimilarity of bacterial community structure between samples (beta diversity) was visualized by principal coordinate plots (PCoA) based on the two-

186

dimensional Bray-Curtis (BC) dissimilarity index. The relative abundance of bacteria in each sample was initially visualized using interactive taxonomic bar plots and then the relative abundance of bacterial phyla in samples from different treatment groups was graphically presented using 100% stacked 2-D column graphs (Excel, Microsoft).

7.3.8 Statistical analyses

Kaplan-Meier survival curves and Cox regression analyses were used to investigate the differences between survival of oysters with different temperature profiles after the OsHV-1 challenge (SPSS Statistics ver. 22; IBM SPSS Cooperation, Somers, NY, USA). The OsHV-1 DNA, and total bacteria DNA were compared between oysters, before acclimation to the tank environment, in different temperature treatment groups and after the OsHV-1 challenge. All data were log10 transformed and normal distribution of data was checked with the descriptive statistics function (SPSS Statistics). Univariate general linear models were used for separate analysis of OsHV-1 DNA and total bacteria DNA (GLM, SPSS Statistics). Fixed factors considered for OsHV-1 DNA were, temperature profile; days after OsHV-1 challenge (Day 2 to 10); outcome of infection (live or dead); together with possible interactions and tank identification number was included as a random factor. Meanwhile, for total bacteria, time of sampling (before and after acclimating to a temperature profile and after OsHV-1 challenge) and temperature profile were tested as fixed factors together with possible interactions and tank was included as a random factor. For both models, post-hoc pairwise mean comparisons were made using the least significant difference method. The results were presented as geometric means and their corresponding 95% confidence intervals for OsHV-1 genome equivalents and bacterial genome equivalents/mg, respectively. As the total Vibrio count was below the limit of quantification for some treatment groups, it was not statistically analysed.

Alpha diversity of bacterial communities was analysed using the Kruskal-Wallis test and one-way permutational multivariate analysis of variance (PERMANOVA) was used for analysis of beta diversity. Variation in the absolute abundance of selected phyla and genera between different treatment groups were evaluated using Generalized Linear Models (GzLM; SPSS). Significance was set at p < 0.05 for all statistical analyses.

187

7.4 Results

7.4.1 Water temperature

The water temperature profiles were achieved according to requirements of the experiment (Figure 7.2). The peak temperature (26°C) of the diurnal temperature fluctuation was reached after 8 h 56 min ± 16 min while the minimum target temperature (22oC) was reached at night after 3h 46min ± 15 (Figure 7.2).

( 26

25 22/26 °C

24 21 °C

23 22 °C

Water Temperature WaterTemperature 22 26 °C 21 20 0 0.5 3.0 4.0 5.0 6.5 7.5 9.0/0 1.0 2.0 3.5 4.5 6.0 7.0 8.0 9.5 10.5 12 13 14

Phase I Phase II

Time after temperature alteration in each cycle (h)

Figure 7.2 Mean water temperatures in oyster tanks recorded using temperature probes immersed in the water.

7.4.2 Mortality

Mortality began on Day 2 post-injection in the OsHV-1 challenged oysters in all treatment groups except at 21°C constant water temperature, where mortality commenced on Day 3 post-injection. While the mortality ceased after Day 7 post-injection at 26°C, it continued until Day 9 at 22°C and Day 10 at both constant 21°C and the 22/26°C dynamic temperature profiles. The total cumulative mortality aggregated across all replicate tanks was 48.4 %, 68.8%, 84.4% and 78.1 %, at 21°C, 22°C, 26°C and with a diurnal fluctuation between 22°C and 26°C, respectively. No mortality was recorded in any of the negative control groups during the trial. Survival analysis indicated a significantly higher survival in OsHV-1 challenged oysters at a constant temperature of 21°C compared to the oysters at other temperatures (Figure 7.3; p < 0.05). The oysters at 22°C, 26°C and 22/26 °C had a hazard of death 1.70, 3.11 and 2.43 times

188

higher than those at 21°C, respectively (Cox hazard ratios). Interestingly, the mortality of oysters at a constant temperature of 26°C was not significantly different from this at 22°C or in the 22/26°C dynamic treatment.

Figure 7.3 Kaplan-Meier survival curves for Pacific oysters challenged with OsHV-1 and maintained with water temperature profiles of constant 21°C, 22°C, or 26°C and with a diurnal fluctuation between 20 and 26°C.

189

7.4.3 OsHV-1 detection and quantity

OsHV-1 was not detected in any of the oysters tested before acclimation to the laboratory environment, after acclimating to different temperature treatments or in oysters injected with an OsHV-1-negative tissue homogenate during and at the completion of the trial. The concentration of OsHV-1 DNA was lowest in oysters maintained at a constant 21°C water temperature and was lower in oysters at 22°C compared to oysters at 26°C or with a diurnal fluctuation between 20 and 26°C (Table 7.1; p < 0.05). The OsHV-1 quantity was higher in dead oysters compared to those that were sampled alive on days 2 and 10, accounting for the water temperature profile (p < 0.05).

Table 7.1 OsHV-1 DNA concentration in gill and mantle tissues of live and dead oysters challenged with OsHV-1 and maintained with different water temperature profiles. The live oysters were sampled on Day 2 and Day 10 post-injection while the dead oysters were sampled at the time of mortality between days 2 to 10 post-injection. Predicted means and their corresponding 95% confidence intervals from a general linear model (GLM) for OsHV- 1 DNA, were back transformed to obtain mean OsHV-1 genomes per mg of tissue. Mean OsHV-1 concentration (genome equivalents/mg) Number of Parameter Geometric 95% Confidence oysters mean* Interval (lower – upper) Temperature profile: 21°C constant 6.21×103 a 2.27×103 – 1.70×104 70 22°C constant 6.30×103 b 2.28×103 – 1.74×104 70 26°C constant 7.60×103 c 2.53×103 –2.29×104 69 22/26°C dynamic 2.12×104 c 8.06×103 – 5.55×104 71 Status of oysters: Live 4.97×102 1.02×102 – 2.42×103 104 Dead 2.69×105 A 8.14×104 – 3.53×105 176 *Geometric means and their corresponding 95% confidence intervals were derived by back- transforming the estimated model means of a general linear model (GLM). a-cMean OsHV-1 concentrations with different superscripts were significantly different (p < 0.05). AThe OsHV-1 concentration was higher in dead oysters than in live oysters (p < 0.05).

190

7.4.4 Total bacteria and total Vibrio quantity

The total quantity of bacterial DNA in gills after acclimation was similar for all water temperature profiles and was not different from that before acclimation to temperature treatments (Table 7.2). Following the OsHV-1 challenge, the total bacteria quantity did not change, except for an increase in oysters with a diurnal fluctuation of water temperature (22/26°C) (p < 0.05). This increase was not observed in oysters that were injected with the OsHV-1 negative tissue homogenate.

The total Vibrio DNA quantified by qPCR in gill were below the limit of quantification (BLOQ) of the qPCR assay for 6/10 oysters that were sampled before acclimation. After acclimation, Vibrio DNA associated with gills increased to quantified amounts (1.87 × 104 – 3.66 × 106 Vibrio genome equivalents per g of tissue) in oysters maintained at 26°C. However, as the Vibrio counts for approximately 90% of the samples from the other treatments were BLOQ, a statistical analysis could not be undertaken (The limit of detection for this qPCR assay was 2 Vibrio gene copies/PCR reaction). An alternative measure of Vibrio quantity abundance was provided by high throughput 16S rRNA gene sequence reads, which is presented in Section 3.5.2.

191

Table 7.2 Quantity of total bacterial DNA associated with the gill of oysters, before and after acclimation to the laboratory at different water temperature profiles and 48h post-OsHV-1 challenge. The number of bacterial genomes per g of tissue was measured using qPCR and

the log10 transformed data were analysed using a general linear model (GLM). The predicted means and their corresponding 95% confidence intervals were back-transformed to obtain mean bacterial genomes/mg tissue, for different groups.

Treatment group Total bacteria concentration Number Geometric mean CI; lower-upper of oysters Before acclimation 3.23×104 1.19×104 -8.79×104 10

After acclimation: 21°C constant temperature 8.51×104 3.41×104-2.12×105 12 22°C constant temperature 3.60×104 1.44×104-8.99×104 12 26°C constant temperature 3.18×104 1.27×104-7.93×104 12 22/26°C dynamic temperature 2.21×104 8.85×103 -5.51×104 12

After OsHV-1 challenge: 21°C constant temperature 3.18×104 1.14×104-8.87×104 8 22°C constant temperature 2.34×104 5.19×103-1.06×105 8 26°C constant temperature 2.73×104 1.08×104-6.92×104 8 22/26°C dynamic temperature 2.63×105* 6.30×104-1.10×106 8

OsHV-1 negative control: 21°C constant temperature 2.93×104 6.88×103 - 1.25×105 4 22°C constant temperature 2.44×104 2.90×103 -2.06×105 4 26°C constant temperature 3.56×104 7.13×103 -1.79×105 3 22/26°C dynamic temperature 2.09×104 2.77×103-1.58×105 4 *The total bacteria concentration increased after the OsHV-1 challenge, in the 22/26°C dynamic temperature treatment group.

7.4.5 Bacterial community composition

7.4.5.1 High throughput 16S rRNA gene sequencing

Targeting the hypervariable V1-V3 region of the 16S rRNA gene, a total of 3,779,301 paired-end raw reads were obtained initially from the 94 samples analysed, leaving 3,219,256 after quality control and bioinformatic processing. The median number of reads per gill tissue sample was 28,343 (maximum: 111,722; minimum: 2,772). The reads were rarefied to 19730 per sample. Rarefaction curves showed saturation for most of the samples, indicative of a good coverage of diversity.

192

7.4.5.2 Changes in the bacterial community composition

The bacteria associated with the gill of oysters from all treatment groups was dominated by phylum Proteobacteria throughout the experiment (Figure 7.6 A-D). However, changes in the bacterial communities were observed after acclimation to the laboratory with different water temperature profiles and subsequent to OsHV-1 challenge. Different changes to the microbiome were observed for each water temperature profile and reflected variation in in the phyla that were impacted. Moreover, inter-oyster heterogeneity in bacterial community composition was observed in the gill microbiota before and after acclimation to the temperature treatment and after the OsHV-1 challenge.

The alpha diversity (number of observed OTUs) in the gill microbiome did not change after acclimation to the tank environment, irrespective of the temperature profile (observed OTUs: before acclimation, 297.7 ± 18.6; after acclimation, Table 7.3). However, the bacterial community composition (beta diversity) changed after acclimation at each temperature (Figure 7.4; p < 0.05). The absolute abundance of phylum Tenericutes increased in oysters that were acclimated at 21°C (Table 7.4; Figure 7.5; p < 0.05). No such changes were observed at phylum level after acclimation at 26°C. At the genus level, it was particularly the genus Mycoplasma (phylum Tenericutes) that increased in abundance at 21°C (p < 0.05). Meanwhile, there was a reduction in the abundance of genus Arcobacter (phylum Proteobacteria) in all treatment groups after acclimation (p < 0.05).

It is important to note that the Vibrio fraction present in gills did not change after acclimation to different temperatures. Nevertheless, differences were observed between the different temperature treatments for other bacterial groups. For instance, higher abundance of phylum Tenericutes at 21°C compared to all other temperature profiles (Table 7.4; Figure 7.4; p < 0.05).

Alpha diversity of bacteria associated with the gill of oysters injected with OsHV-1 was higher compared to the negative control counterparts for the constant 21°C and dynamic 22/26°C water temperature profiles (Table 7.3; p < 0.05). Amongst the OsHV-1 challenged oysters, the number of observed OTUs was the same for different temperatures, except for the lower number in oysters at a constant 26°C (Table 7.3; p < 0.05). The beta diversity of gill

193

microbiota changed in all temperature treatment groups (Figure 7.4; Figure 7.6 A-D; p < 0.05).

It is interesting to note that the phyla Proteobacteria and Bacteroidetes did not change in absolute abundance after acclimating at different temperatures (Fig. 7.6 A-D). However, compared to the negative control oysters, both phyla increased in abundance after OsHV-1 challenge, except for phylum Bacteroidetes at the constant 22°C and 26°C temperature (Table 7.4; Figure 7.5; Figure 7.6 A and D); p < 0.05). Phylum Tenericutes reduced in abundance after the OsHV-1 challenge at 21°C and the 22/26°C dynamic temperature (Table 7.4; Figure 7.5; Figure 7.6 A and D); p < 0.05). This reduction also occurred for the negative control oysters at 21°C.

After the OsHV-1 challenge the changes at the genus level were increase in abundance of Polaribacter (phylum Bacteroidetes) at 21°C and the genus Vibrio (phylum Proteobacteria) in all treatment groups (p < 0.05). These changes did not occur in oysters injected with the negative control. A decrease of Mycoplasma (phylum Tenericutes) was noted in OsHV-1 injected and control oysters when the water temperature was a constant 21°C. It is interesting to note that the abundance of Vibrio in the gill microbiota of OsHV-1 challenged oysters that were maintained at 26°C was higher than for oysters maintained at other temperatures (Table 7.5; Figure 7.7; p < 0.05).

194

Table 7.3 Alpha diversity (number of observed OTUs ± SE) of gill microbiota in oysters challenged with OsHV-1 and controls and maintained at different temperature regimes.

Temperature treatment Number of observed OTUs (mean ± SE) Pre-challenge OsHV-1 challenged Negative control 21°C static temperature 323.3 ± 29.4 (n=4) 603.9 ± 103.0 (n=8) 251.8 ± 7.1 (n=4) 22°C static temperature 312 ± 40.8 (n=10) 418.5 ± 24.2 (n=8) 424 ± 90.1 (n=4) 26°C static temperature 340 ± 55.3 (n=7) 309 ± 38.3 (n=8) 299 ± 78.2 (n=3) 22/26°C dynamic temperature 435.6 ± 68.5 (n=8) 516 ± 51.4 (n=8) 357.8 ± 27.2 (n=4)

195

Table 7.4 Average relative abundance of dominant phyla in the gill microbiota of oysters, before and after the temperature treatments and in live oysters sampled 48h after OsHV-1 challenge. Microbiomes were analysed from gill tissue samples (n=6–10) collected from each treatment group. Treatment group Average relative abundance (%) Before 21°C _static 22°C _static 26°C _static 22/26°C _dynamic Significance

After acclimation: Proteobacteria 17.49 30.28 35.99 30.15 28.67 p > 0.05 Bacteroidetes 2.18 7.53 2.70 3.77 3.45 p > 0.05 Tenericutes 0.03 5.91* 0.86 0.24 0.48 p = 0.00 Spirochaetes 25.57 27.42 29.66 34.79 33.19 p = 0.04 Planctomycetes 0.20 0.90 0.71 1.4 0.77 p > 0.05 48h after OsHV-1 challenge: Proteobacteria - 48.90** 51.49** 69.13** 60.02** p = 0.00 Bacteroidetes - 12.55† 7.69 10.75 14.69† p = 0.01 Tenericutes - 0.20 0.06 0.03 0.12 p > 0.05 Spirochaetes - 15.90 15.10 11.20 12.01 p > 0.05 Planctomycetes - 1.52 0.82 0.70 0.99 p < 0.00 Negative control: Proteobacteria - 35.06 42.16 46.11 43.46 p > 0.05 Bacteroidetes - 1.99 7.93 5.73 4.40 p > 0.05 Tenericutes - 0.21 0.03 0.04 0.11 p > 0.05 Spirochaetes - 39.42 21.95 24.87 23.06 p > 0.05 Planctomycetes - 0.39 2.00 0.84 1.17 p > 0.05 * The abundance of phylum Tenericutes was higher in the gill microbiota of oysters that were acclimated to 21°C static treatment compared to that in oysters sampled before acclimation and those acclimated to other temperatures (p<0.05). ** The phylum Proteobacteria increased in abundance after OsHV-1 challenge in oysters maintained in 21°C, 22°C and 26°C static treatments (p<0.05). † The phylum Bacteroidetes increased in abundance after OsHV-1 challenge in oysters maintained in 21°C static and 20/26 °C dynamic treatments (p<0.05).

196

Table 7.5 Average relative abundance of the dominant genera in the gill microbiota of oysters, before and after the temperature treatments and 48h post-OsHV-1 challenge. Microbiomes were analysed from gill tissue samples (n = 6–10) collected from each treatment group. Treatment group Average relative abundance (%) Before 21°C_static 22°C _static 26°C _static 22/26°C Significance _dynamic After acclimation: Vibrio 0.11 0.26 0.62 1.50 0.41 p > 0.05 Polaribacter 0.02 0.16 0.23 0.04 0.06 p > 0.05 Mycoplasma 0.00 5.91* 0.86 0.24 0.48 p = 0.00 Arcobacter 1.18** 0.03 0.08 0.03 0.04 p = 0.01 After OsHV-1 challenge: Vibrio - 2.11 5.50 10.89† 2.70 p = 0.00 Polaribacter - 1.97 1.99 0.19 1.08 p = 0.00 Mycoplasma - 0.20†† 0.06 0.03 0.12 p = 0.00 Arcobacter - 0.01 0.02 0.04 0.04 p > 0.05 Negative control: Vibrio - 1.72 1.96 6.35 3.58 p > 0.05 Polaribacter - 0.35 1.71 0.28 0.21 p > 0.05 Mycoplasma - 0.21†† 0.03 0.04 0.11 p = 0.00 Arcobacter - 1.55 0.01 0.01 0.37 p =0.01 * The genus Mycoplasma increased in abundance after acclimating to 21°C static treatment and this abundance was higher than that in oysters acclimated to 22°C and 26°C static treatments (p<0.05). ** The genus Arcobacter decreased in abundance after acclimating to all temperature treatments (p<0.05). † The genus Vibrio increased in abundance after OsHV-1 challenge in oysters maintained in all temperature treatments (p<0.05). The Vibrio abundance in OsHV-1 challenged oysters at 26°C static treatment were higher than that in OsHV-1 challenged oysters of all other temperature treatments (p<0.05). †† The genus Mycoplasma decreased in abundance after OsHV-1 challenge and in negative control oysters maintained in 21°C static treatment (p<0.05).

197

Figure 7.4 Principal coordinate plot based on Bray-Curtis distances between the gill microbiome of Pacific oysters before and after acclimation to different seawater temperatures. The gill microbiota of oysters before acclimation (red; 20°C) were distinct from those after acclimation to temperatures, constant 21°C (orange), 22°C (green), 26°C (purple) and diurnal fluctuation to 22/26°C (blue) (p > 0.05).

198

0.9

0.8

Proteobacteria 0.7 Spirochaetes

0.6 Other

Bacteroidetes 0.5 Unassigned

Verrucomicrobia

0.4 Relative (%) Relative abundance Tenericutes

0.3 Plactomycetes

GN02 0.2 Fusobacteria

0.1 Firmicutes

21_s

22_s

26_s

Before

22/26_d

22/26_d 22/26_d Pre-acclimation Post-acclimation

Temperature treatment

Figure 7.5 Taxa bar plots indicating the relative abundance of bacterial phyla in gill microbiota of oysters before and after acclimation to the laboratory environment with different water temperature profiles over 14 days: constant 21°C (21_s), 22°C (22_s), 26°C (26_s) and diurnal fluctuation between 22°C and 26°C (22/26_d) temperatures. Results were analysed using QIIME2. Bacteria that could not be assigned to a particular phylum are categorized under ‘unassigned’ and phyla with a relative abundance of less than 5% and were not present in at least two samples are categorized as ‘other’.

199

A 1

0.9

0.8

0.7 Proteobacteria

Spirochaetes 0.6 Other

Bacteroidetes 0.5 Unassigned

Relative (%) Relative abundance 0.4 Verrucomicrobia

Tenericutes 0.3 Planctomycetes

0.2 GN02

Fusobacteria 0.1 Firmicutes

21_s 21_s 21_s 21_s 21_s 21_s 21_s 21_s 21_s 21_s 21_c 21_c 21_c 21_c

21_ch 21_ch 21_ch 21_ch 21_ch 21_ch 21_ch 21_ch Pre-challenge OsHV-1 challenged Control

Temperature treatment

200

201

C 1

0.9

0.8

Proteobacteria 0.7 Spirochaetes

Other 0.6 Bacteroidetes

0.5 Unassigned

Verrucomicrobia 0.4

Tenericutes Relative (%) Relative abundance 0.3 Planctomycetes

GN02 0.2 Fusobacteria

0.1 Firmicutes

26_c

26_s

26_ch

26_ch 26_ch Pre-challenge OsHV-1 challenged Control

Temperature treatment

202

D 1

0.9

0.8 Proteobacteria

Spirochaetes 0.7 Other

0.6 Bacteroidetes

Unassigned 0.5 Verrucomicrobia

0.4 Tenericutes Relative (%) abundance Relative Planctomycetes 0.3 GN02 0.2 Fusobacteria

Firmicutes 0.1

22/26_c

22/26_d

22/26_ch

22/26_ch 22/26_ch 22/26_ch Pre-challenge Challenged Control

Temperature treatment

Figure 7.6 Taxa bar plots indicating the relative abundance of bacterial phyla associated with the gill of oysters before and after OsHV-1 challenge and in negative control oysters maintained with the same temperature profiles. A) After acclimation at constant 21°C (21_s), 21°C OsHV-1 challenged (21_ch) and negative control 21_c; B) after acclimation at constant 22°C (22_s), 22°C OsHV-1 challenged (22_ch) and negative control 22_c; C) after acclimation at constant 26°C (26_s), 26°C OsHV-1 challenged (26_ch) and negative control 26_c; and D) after acclimation at diurnal fluctuation between 22°C and 26°C (20/26d), 22/26°C OsHV-1 challenged (26/20_ch) and negative control (26/20_c). Live oyster samples were collected 14 days after acclimation to the relevant temperatures, and 48 h after the OsHV-1 challenge. Results were analysed using QIIME2. Bacteria that could not be assigned to a particular phylum are categorized under ‘unassigned’ and phyla with a relative abundance of less than 5% and were not present in at least two samples are categorized as ‘other’.

203

4000 *$

3000 * Unchallenged Challenged

* sequencereads 2000 *

Vibrio

1000

Number of of Number

0 20 22 26_20 26

Temperature profile

Figure 7.7 Absolute abundance (number of DNA sequence reads) of the genus Vibrio in the gill microbiota of oysters challenged with OsHV-1 and maintained at different temperatures. *Higher abundance of Vibrio in oysters challenged with OsHV-1 compared to unchallenged oysters at the same temperature (p < 0.05); $Higher abundance of Vibrio in OsHV-1 challenged oysters at 26°C compared to those at other temperatures of the study (p < 0.05).

204

7.5 Discussion

This study evaluated the influence of seawater temperature on the Pacific oyster microbiome, with different water temperature profiles in laboratory aquaria and the impact of it on OsHV-1 infection. Significant changes in the bacterial community associated with the gills occurred after acclimation to different water temperatures. These changes were different depending on the water temperature profile within the range at which OsHV-1 causes disease outbreaks (Clegg et al., 2014; Paul-Pont et al., 2014; Renault et al., 2014; Whittington et al., 2019).

The phylum Proteobacteria dominated the gill microbiome of oysters in all treatment groups, throughout the experimental period. The dominance of phylum Proteobacteria has previously been demonstrated in gill microbiota (Pathirana et al., 2019a; Wegner et al., 2013) as well as in other tissues of oysters (Fernandez-Piquer et al., 2012; Hernandez-Zarate and Olmos-Soto, 2006; Lokmer et al., 2016b; Trabal et al., 2012). Together with the results of Chapter 5 the present study demonstrates that the phyla Proteobacteria and Bacteroidetes do not change in abundance after acclimation to the laboratory environment, irrespective of their immersion regime and the water temperature, under the tested condition.

Despite the picture of the microbiome at higher taxonomic level, changes in the gill microbiome composition was observed at genus level. In the present study, an increase in the abundance of the genus Mycoplasma (phylum Tenericutes) was observed in the 21°C static treatment group. Mycoplasma has been reported to naturally dominate the oyster microbiome in warmer seawater temperatures (King et al., 2012). Wegner et al. (2013) observed an increase in the Mycoplasma fraction in the gill microbiome, after subjecting the oysters to a disturbance treatment which mainly involved an increase of temperature (2°C to 26°C) along with transfer to the laboratory environment and absence of feed. Mycoplasma species represent a temperature-sensitive part of oyster microbiota and may selectively proliferate at higher temperatures (Wegner et al., 2013). It has been associated with disease in shellfish (Paillard et al., 2004). Studies have also reported pathogenesis of Mycoplasma and Mycoplasma-like organisms in other aquatic invertebrates (Azevedo, 1993; Krol et al., 1991) and in fish (Kirchhoff et al., 1987). Thus, the increase of Mycoplasma at warmer seawater temperatures may be an indication of an opportunistically pathogenic role of Mycoplasma. However, an increase in the Mycoplasma fraction was not noted at 26°C, in this study.

205

The abundance of genus Arcobacter (phylum Proteobacteria) decreased in the gill microbiome of oysters used in this study, during acclimation to different temperature profiles. groups. This observation was different from what was observed in Chapter 5 where an increased abundance of Arcobacter was seen in the gill microbiome, after acclimation to the laboratory environment. While young oysters (5 months old) were used in that study, the present study employed adult oysters (15-16 months of age), resulting in a contribution of age factor, to this differential microbial response. On the other hand, Arcobacter being microaerophilic, the periodic valve closing in oysters may have facilitated the growth of Arcobacter in the gill (Vandamme and De Ley, 1991). As the oysters were maintained constantly immersed in water, in the laboratory environment, the access to the microaerophilic environment might have reduced, reducing the multiplication of Arcobacter. Further, the increase in seawater temperature may have disturbed the microbiome, leading to a shift in the microbiome composition with fewer numbers of Arcobacter (Wegner et al., 2013). The high abundance of Arcobacter in the haemolymph, previously led to the understanding that they were symbionts of the oyster haemolymph (Lokmer and Wegner, 2015). However, they have also been shown to increase in moribund oysters suggesting an opportunistic role of these bacteria in disease (Lokmer and Wegner, 2015). The increase of phylum Spirochaetes in the present study was in line with the results of previous studies in other oyster species. Seasonal temperature variations in seawater have been related to variations in the microbiome composition in the gill of spondylus oysters which included an increase of phylum Spirochaetes, throughout summer, with temperatures ranging from 23°C to 31°C (Roterman et al., 2015).

The OsHV-1 challenge of the present study resulted in mortality in oysters from all treatment groups. However, the response was delayed by one day in those at the lowest, constant water temperature of 21°C. I observed the highest total cumulative mortality in oysters in the 26°C static treatment group coupled with a higher OsHV-1 content. Further a graded response was observed with mortality, down to 21°C. Mortality related to OsHV-1 usually occurs at water temperatures between 16°C to 24°C, under field conditions (Pernet et al., 2012; Petton et al., 2015a; Renault et al., 2014; Whittington et al., 2019). Laboratory models have also shown mortality related to OsHV-1 at 26°C (de Kantzow et al., 2016; Delisle et al., 2018). Higher mortality at higher water temperatures has also been observed in other laboratory studies (de Kantzow et al., 2016; Delisle et al., 2018). Based on studies conducted in oysters in the field and in the laboratory, the OsHV-1 DNA content has a

206

positive relationship with the seawater temperature (de Kantzow et al., 2016; Petton et al., 2013). Following a standard OsHV-1 exposure, the quantity of OsHV-1 DNA in moribund oysters at 26°C has been shown to be approximately 6-fold greater than that at 18°C and 1.2- fold greater than that at 22°C (de Kantzow et al., 2016). A recent laboratory study conducted by Delisle et al. (2018) also showed a higher OsHV-1 DNA content in oysters at 26°C as opposed to 21°C and 29°C. Apart from increasing the OsHV-1 load, the oyster microbiome was disturbed in elevated seawater temperatures, favouring shifts in the composition towards pathogen-dominated communities (Le Roux et al., 2016; Lokmer and Wegner, 2015).

Concurrent with differences in the OsHV-1 content and mortality in different temperature profiles, increase in abundance of the Vibrio fraction was observed in oysters maintained with all the temperature profiles (constant 21°C, 22°C, 26°C and 22/26°C dynamic temperature profile) tested in this study. These increases were not seen in their negative control counterparts indicating an association of the OsHV-1 infection with the increase of Vibrio. The higher increase of Vibrio observed in OsHV-1 challenged oysters at 26°C indicates a role of increased seawater temperature in facilitating this increase of Vibrio. Moreover, the concurrent increase of Vibrio with the OsHV-1 also strengthens the potential role of Vibrio in the proposed polymicrobial pathogenesis in OsHV-1 associated oyster mortality events. The complex aetiology of disease caused by OsHV-1 was recently investigated by de Lorgeril et al. (2018). They showed that a primary infection with OsHV-1 in oyster haemocytes resulted in bacteraemia with opportunistic bacteria. The same study demonstrated that Vibrio was one of the two main genera associated with this opportunistic role, in completing the disease pathogenesis initiated by OsHV-1. Increase of the genera Bacteroides, Vibrio and Arcobacter were reported in natural mortality events of Pacific oysters that were associated with OsHV-1, Vibrio aestuarianus and Vibrio harveyi (King et al., 2018a; Lasa et al., 2019). With a comparative analysis, Lasa et al. (2019) showed that both Vibrio and Arcobacter increase in abundance in oysters infected with Vibrio aestuarianus whereas, the Vibrio fraction increased in OsHV-1 infections. The emerging view of the polymicrobial pathogenesis in Pacific oyster mortality (in both the presence and absence of OsHV-1) has been strengthened by an array of studies which analysed oyster microbiota both from natural field outbreaks (King et al., 2018a; Lasa et al., 2019) and from laboratory infection models (Pathirana et al., 2019b; Petton et al., 2019).

207

In the present study, increase in the Vibrio fraction was only seen after the OsHV-1 challenge and not after the temperature acclimation phase. The OsHV-1 challenge resulted in an increased abundance of Vibrio and a mortality rate of 84.4%. Moreover, this increase was not seen in the negative control oysters of our study. Gill tissue being a predilection site for OsHV-1, this increase of Vibrio in the gill microbiome is in line with the recent findings of de Lorgeril et al. (2018). With a histopathological analysis they showed that bacteria accumulated in the gill tissue at the onset of the viral infection which was supported by an increase of Vibrio counts. Lokmer et al. (2016a) demonstrated a higher Vibrio load in solid tissues of oysters during mortality that occurred after translocation into a new environment. In addition to OsHV-1, several members of the genus Vibrio are considered to play a primary role in oyster mortality outbreaks (Vezzulli et al., 2015). Unlike the higher mortality reported in oyster spat and juveniles due to OsHV-1 (Renault et al., 1994b; Schikorski et al., 2011), natural mortality events due to Vibrio infections are mostly reported in adult oysters (Travers et al., 2015). Although we could not isolate and test the pathogenicity of Vibrio in the present study, the higher quantity of Vibrio reported at 26°C may have played a secondary opportunistic role in oyster mortality. Fluctuating water temperatures in the range 22/26°C did not alter OsHV-1 associated mortality in this study compared to 26°C, but there was a higher total bacteria quantity. Other environmental factors such as tidal emersion may interact with changing temperatures in the natural estuarine environment to produce a different picture. Contrary to the results of the present study and several other studies, Green et al. (2018) showed an increase of the Vibrio fraction in the whole tissue homogenates of oyster spat when the water temperature was elevated to 25°C. Further this temperature increase was associated with a mortality of 77.4%, which occurred in the absence of OsHV-1.

Unlike the increase of Vibrio, the present study did not observe any increase in Arcobacter in the gill microbiome, after the OsHV-1 challenge. The study by Lasa et al. (2019) also report increase of Vibrio alone in OsHV-1 infections while both Vibrio and Arcobacter species increased in Vibrio aestuarianus infections. Dominance of Arcobacter in moribund oysters has also been reported in the haemolymph microbiome, in the absence of an increase in Vibrio (Lokmer and Wegner, 2015). At higher taxonomic levels, the abundance of phylum Proteobacteria and phylum Bacteroidetes increased in the gill microbiome, after OsHV-1 challenge.

208

One of the main objectives of this study was to investigate the impact of different seawater temperatures on the microbiome in relation to an OsHV-1 challenge. In this context, we also observed a lower alpha diversity in the gill of oysters maintained at a constant 26°C. Dysbiosis (loss of bacterial diversity and proliferation of few OTUs) has repeatedly been associated with impaired oyster health (Garnier et al., 2007; Green and Barnes, 2010; Lokmer and Wegner, 2015; Pathirana et al., 2019b). The shorter clinical course of infection (5 days as compared to 7 and 8 days in other treatments) in oysters maintained in the 26°C static treatment, may have increased the probability of getting dysbiotic oysters, during the live oyster sampling, 48h after the OsHV-1 challenge.

The dynamic temperature treatment (22/26°C) employed in this study did not result in any drastic changes in the oyster mortality or in the oyster microbiome. Although the required temperature changes were reached in seawater, the time taken to reach the peak temperature was longer and the experimental design did not allow the peak temperature to remain at the level for a very long period. This may have affected this temperature treatment resulting in an outcome more or less similar to other treatments.

7.6 Conclusion

The water temperature profiles provided in the present laboratory aquaria did not affect the quantity of bacteria associated with oysters but did alter the bacterial community composition. The degree and nature of these changes varied with the water temperature profile and reflected differences between bacterial genera. The opportunistic role of Vibrio in OsHV-1-associated oyster mortality appeared to be further facilitated by the seawater temperature. Higher oyster mortality was not only associated with a higher water temperature and a higher OsHV-1 load but was also associated with the highest Vibrio concentration. Except for the increase in total bacterial quantity after OsHV-1 challenge, the diurnal temperature fluctuations between 22 and 26°C did not decrease the stability of the oyster microbiome compared to constant temperatures.

209

CHAPTER 8

General Discussion

8.1 Introduction

Since 2008, mass mortality outbreaks associated with microvariant genotypes of Ostreid herpesvirus-1 (OsHV-1) have been reported in commercially produced Pacific oysters, initially in France (Segarra et al., 2010). This emerging pathogen subsequently impacted Pacific oysters in different parts of the world including Australia (Jenkins et al., 2013; Paul-Pont et al., 2015; Whittington et al., 2015b), New Zealand (Renault et al., 2012), Ireland (Clegg et al., 2014; Peeler et al., 2012), Spain (Roque et al., 2012) and Scandinavia (Mortensen et al., 2016). Meanwhile, a number of pathogenic Vibrio species have also been isolated from mass mortality disease outbreaks in Pacific oysters and the isolates have been tested for virulence, causing mortality in oysters. These species included Vibrio splendidus (Garnier et al., 2007; Lacoste et al., 2001; Pernet et al., 2012), Vibrio aestuarianus (Garnier et al., 2008; Le Roux et al., 2016; Saulnier et al., 2010), Vibrio harveyi (Saulnier et al., 2010) and Vibrio crassostreae (Bruto et al., 2017). Further, some outbreaks of mass mortality of Pacific oysters have been attributed to adverse environmental conditions such as elevated seawater temperature (>20°C) and low salinity (< 20 ppt) without the involvement of a specific infectious pathogen (Go et al., 2017).

Mass mortalities of Pacific oysters are considered multifactorial, which involves environmental stress factors, oyster factors and pathogens (OIE, 2014; Samain and McCombie, 2008). The incidence and severity of the disease varied with oyster physiology (Green et al., 2016; Samain and McCombie, 2008; Solomieu et al., 2015), pathogen (Dundon et al., 2011; Segarra et al., 2010) and environmental stress factors (Clegg et al., 2014; Petton et al., 2013; Samain and McCombie, 2008). The complex aetiology of disease caused by OsHV-1 was recently investigated by de Lorgeril et al. (2018). A primary infection with OsHV-1(µVar) in oyster haemocytes resulted in an immune-compromised state in oysters which was followed by a bacteraemia with opportunistic bacteria present in the oyster microbiome. This secondary bacterial involvement was necessary for complete disease expression. Histological studies revealed bacterial accumulation and infiltrating haemocytes both inside and outside the gill tissues of oysters infected with OsHV-1 (de Lorgeril et al., 2018). Petton et al. (2015b) also demonstrated that a high load of OsHV-1 alone is

210

insufficient to induce the full expression of the disease, demonstrating a role of bacteria in disease transmission and development. In this study, delayed and reduced mortality of Pacific oysters exposed to an OsHV-1 outbreak was demonstrated in oysters treated with a broad- spectrum antibiotic (chloramphenicol), compared to oysters that were not treated with chloramphenicol. An 8 mg/L dose of chloramphenicol was added to tanks every 2 days to remove the cultivable microbiota including Vibrio in oyster tissues (Petton et al., 2015b). Green et al. (2018) observed low mortality (4.3%) in spat exposed to a combination of penicillin (100 units/ml) and streptomycin (0.1 mg/ml) in seawater when the water temperature was at 25 °C. A higher spat mortality (77.4%) was seen at the same water temperature in the absence of antibiotics.

With this background, recent research sheds light on a potential role for the oyster microbiome in the pathogenesis of a range of Pacific oyster diseases. The oyster microbiome can be considered the collection of all microbial communities that are associated with different tissues of the oyster. However, a large proportion of the oyster microbiome cannot be cultivated by standard procedures (Romero and Espejo, 2001). Introduction of high throughput sequencing methods (Caporaso et al., 2011; Streit and Schmitz, 2004) facilitated the analysis of bacterial 16S rRNA genes providing phylogenetic portraits of microbial communities, including micro-organisms that had not yet been cultivated (Frank et al., 2008; Lane et al., 1985; Pavlova et al., 2002). Thus, 16S rRNA gene diversity profiling enabled Pacific oyster microbiome studies, providing new insights to the pathogenesis of Pacific oyster mortality events (de Lorgeril et al., 2018; King et al., 2018a; Pathirana et al., 2019b).

The oyster microbiome carries out a range of functions including defence against external pathogens (Desriac et al., 2014), providing the host with growth factors such as vitamins and amino acids and assisting in digestion of food (Prieur et al., 1990). To date, the dynamics of the Pacific oyster microbiome has been studied in a vast range of situations including that in oyster translocation to new environments (Lokmer et al., 2016a; Lokmer et al., 2016b), temperature and temperature stress (Lokmer and Wegner, 2015), pH of seawater (Flores-Higuera et al., 2019), during growth from post-larvae to adults (Trabal Fernández et al., 2014), depuration at farm level (Vezzulli et al., 2018), experimental infection with pathogenic Vibrio (Lokmer and Wegner, 2015) and in natural oyster mortality events (King et al., 2018a). The studies revealed that the oyster microbiome can be affected by environmental disturbances such as temperature stress (Lokmer and Wegner, 2015) and translocation

211

(Lokmer et al., 2016b) with destabilization of the microbiome, resulting in dysbiosis (loss of bacterial diversity and proliferation of few OTUs) and shifts in the microbiome composition, favouring pathogenic bacteria (Lasa et al., 2019; Wegner et al., 2013). Both natural and experimental infections were identified to affect the oyster microbiome with increases in bacteria such as Vibrio and Arcobacter, in addition to the primary infectious agent (de Lorgeril et al., 2018; Lasa et al., 2019).

There is an emerging view of a polymicrobial pathogenesis underlying infectious diseases in marine bivalves. Outbreaks of high mortality disease in Pacific oysters associated with OsHV-1, are thought to be caused by a group of microbial species, or multiple strains of the same microbial species, that act as a community of pathogens (de Lorgeril et al., 2018; Lasa et al., 2019; Lemire et al., 2015). However, the detailed mechanisms for polymicrobial pathogenesis in oyster mortality are yet to be understood. In addition to the emerging view of polymicrobial pathogenesis, various environmental risk factors have also been implicated as contributing to Pacific oyster mortality associated with OsHV-1. Seawater temperature above a threshold (16°C in Europe and 4-5°C warmer in Australia) has been reported with oyster mortality associated with OsHV-1 (Paul-Pont et al., 2014; Pernet et al., 2012; Renault et al., 2014; Whittington et al., 2019). Modification of oyster farming practices in commercial oyster farming, such as increasing the height of growing structures above the standard intertidal position resulted in reduced mortality of adult Pacific oysters during OsHV-1 disease outbreaks (Paul-Pont et al., 2013b; Whittington et al., 2015b).

Dysbiosis in the oyster microbiome is commonly observed when there is a disturbance to the microbiome and usually precedes mortality events of oysters (de Lorgeril et al., 2018; Lasa et al., 2019). It is important to determine if this is an outcome of a primary disease or whether various host and environmental factors predispose the oyster microbiome for dysbiosis which contributes to the severity of disease. This thesis aimed at addressing the impact of environment factors on the Pacific oyster microbiome and the subsequent effects on disease associated with OsHV-1. Initially, methods to sample different tissues of Pacific oysters and to extract nucleic acids for the accurate identification of the microbiome, were optimized. The microbiome of oysters with a common hatchery origin but then were grown in geographically distinct estuaries was then studied. Thereafter, a laboratory model was used to determine how controlled changes in specific environmental factors could alter the microbiome in the short-term. These changes in the microbiome were assessed concurrently

212

with the direct effects of the environmental changes in determining the impact on the severity of disease caused by OsHV-1.

8.2 Accurate identification of the oyster microbiome

Accurate identification of Pacific oyster microbial diversity is instrumental for understanding the polymicrobial pathogenesis of Pacific oyster mortality diseases. Both intrinsic oyster factors as well as experimental factors can influence the results of oyster microbiome studies with potential biases being introduced from the level of sampling up to the level of nucleic acid extraction (Chapter 4). Healthy Pacific oysters usually harbour a diverse microbial community (Lokmer and Wegner, 2015; Prieur et al., 1990). The results of Chapter 4 indicated a tissue compartmentalization in the Pacific oyster microbiome, with distinct, tissue-specific microbiomes in the haemolymph, gill and gut tissues. The gill microbiome is considered generally stable and relatively enriched with symbiotic bacteria which are integrated into the gill tissues (autochthonous bacteria) (Prieur et al., 1990; Roterman et al., 2015; Zurel et al., 2011). On the other hand, the microbiota of the digestive system generally consisted of transient bacteria (allochthonous bacteria) (Zurel et al. 2011; Roterman et al. 2015). The tissue compartmentalization of the Pacific oyster microbiome and their functional differences indicated the importance of studying tissue-specific microbiomes, that fit the purpose of the study. Chapters 5 and 6 considered both gill and gut microbiomes of the same oysters to assess the dynamics of two different microbiomes under controlled changes in the environment and OsHV-1 challenge.

Nucleic acid extracts from oyster tissues have demonstrated a high level of PCR inhibition (Abolmaaty et al., 2007). In Chapter 4, extraction of bacterial DNA from oyster tissues that were free from inhibitors, particularly from gut tissues, was identified as a challenge. Molluscs secrete mucopolysaccharides that copurify with DNA and interfere with downstream nucleic-acid processing (Winnepenninckx et al. 1993; Pereira et al. 2011). The cetyltrimethylammonium bromide (CTAB method) used in the E.Z.N.A. Mollusc DNA Kit was effective in removing mucopolysaccharides and improved the quality of bacterial DNA extraction from oyster tissues (Chapter 4; Pathirana et al. (2019a)). A selection of general- purpose and microbiome specific DNA extraction kits were less effective for representing the

bacterial diversity in oyster tissue samples (Chapter 4).

213

A fit-for-purpose sampling strategy is required for analyses of the oyster microbiome, owing to tissue-specific differences in bacterial communities and potential for biases during nucleic acid extraction. The CTAB method is preferred as it preserves the original bacterial community structure in tissues while alleviating PCR inhibition in oyster gut tissues. Swabbing provides a different bacterial DNA yield but does not impact bacterial community structure. Although freezing is suitable for sample preservation the bacterial DNA yield may

be reduced which may create a negative impact on downstream molecular analyses.

8.3 Dynamic changes in the Pacific oyster microbiome

Changes in the Pacific oyster microbiome have been observed with both natural (Lasa et al., 2019) and experimental OsHV-1 infections (de Lorgeril et al., 2018; Pathirana et al., 2019b). de Lorgeril et al. (2018) studied the oyster microbiome in an OsHV-1 infection, using oyster families with highly contrasting resistance phenotypes for oyster mortality associated with OsHV-1 (susceptible vs. resistant). They demonstrated that changes in the microbiome were greater in oysters that were susceptible, with significant changes in the relative abundance of 105 operational taxonomic units (OTUs). In contrast, no changes in OTU abundance were seen in the resistant group. Some of the OTUs that changed in their relative abundance belonged to the genera Arcobacter and Vibrio (de Lorgeril et al., 2018). Interestingly, increases in the same genera have been reported in other studies of Pacific oyster mortality, both with OsHV-1 infection (Lasa et al., 2019) and without (Green et al., 2018; Lemire et al., 2015; Lokmer and Wegner, 2015). An increase of Arcobacter in the haemolymph of moribund oysters, following an experimental infection with a pathogenic Vibrio species and increase of pathogenic Vibrio species and Arcobacter in oyster spat after an increase of seawater temperature, were not associated with OsHV-1 (Green et al., 2018; Lokmer and Wegner, 2015). These findings can be considered indications of the involvement of Arcobacter and Vibrio in oyster mortality associated with environmental disturbances and with infections other than OsHV-1 infections.

In Chapter 3 of this thesis, the influence of different estuarine environments on the Pacific oyster microbiome was evaluated for oysters with a common hatchery origin that were subsequently grown in three geographically distinct estuaries. The microbiome of the three batches were different both in terms of alpha diversity and beta diversity. Similar results have been observed previously in Crassostrea virginica, both with metagenomic analysis (Ossai et al., 2017) and with conventional bacteriological studies (Prieur et al., 1990), in

214

oysters grown in distinct estuaries. Transient and opportunistic microbial taxa are thought to dominate the oyster microbiome while the core microbial communities were limited to only a few resident bacteria (Lasa et al., 2019). Accordingly, the differences in the microbiome that was seen with distinct estuaries in this study, can be associated with the filter-feeding behaviour of oysters. This feeding style is thought to expose the oysters to colonization by complex and highly variable microbial communities found in seawater (Lasa et al., 2019). However, the microbiome is also influenced by host-related factors as well as by interactions within the microbiome. In this regard, Lokmer et al. (2016b) demonstrated persistent effects of the microbiome in oysters translocated to a new environment and treated with antibiotics together with recognizable individual temporal dynamics. The Pacific oyster microbiome was dominated by classes Gammaproteobacteria and Alphaproteobacteria and was highly variable in relation to health status, geographic location, season and oyster age (Lasa et al., 2019).

In Chapter 3, a differential microbiome response was observed for the three batches of oysters with initially distinct microbiomes, after challenging with OsHV-1. Oysters from Patonga Creek which had a higher OsHV-1 DNA concentration and higher mortality, also had a different microbiome response to the OsHV-1 challenge. Whilst these oysters had the highest initial bacterial diversity, there was a decrease during the moribund stage which did not occur for oysters from the other two locations. The microbiome response of Pacific oysters exposed to heat stress was consistent with this observation (Wegner et al., 2013). Oysters with a high microbial diversity had a decrease following the disturbance due to heat stress, whereas oysters with originally lower diversity did not show such a decrease after exposure to heat (Wegner et al., 2013). Dysbiosis is a common outcome of factors that disturb the oyster microbiome. Low bacterial diversity has been associated with impaired health in Pacific oysters after infection with pathogenic Vibrio species as well as in Sydney rock oysters affected by QX disease (Garnier et al., 2007; Green and Barnes, 2010). The decrease in bacterial diversity which preceded mortality has been considered an indicator of declining health in oysters (Lokmer and Wegner, 2015; Wegner et al., 2013).

The Chapter 3 of this thesis strengthens the concept that environment plays an important role in shaping the oyster microbiome. In addition to genetics of oysters, factors such as age of oysters may also influence the composition of the microbiome. These changes may subsequently result in different outcomes when the oysters are exposed to disease.

215

8.4 Microbiome studies in a controlled environment

Simplified experimental systems used in the study of infectious disease are thought to minimize the influences of host, pathogen diversity and environmental factors on the disease, making it difficult to characterize diseases of complex aetiologies (de Lorgeril et al., 2018; Petton et al., 2015b; Petton et al., 2019). However, the effect of one specific environmental factor can be better studied in a controlled environment by minimizing the variations of other environmental factors as well as by controlling oyster factors and the diversity of pathogen exposure. Chapter 5 of this thesis looked at the changes that may occur in the Pacific oyster microbiome simply by maintaining oysters in a controlled laboratory environment. Meanwhile, Chapters 6 and 7 studies specifically looked at the effect of immersion regime and seawater temperature on the oyster microbiome in disease associated with OsHV-1, using a controlled laboratory environment.

Using a controlled laboratory environment, Lokmer and Wegner (2015) studied the hemolymph microbiota of temperature-acclimated (8 or 22°C) Pacific oysters exposed to temperature stress and an experimental injection challenge with a virulent strain of a Vibrio species. The research revealed that the microbiome of oysters was affected by temperature and temperature stress. However, the low abundance of the injection strain in DNA sequence analysis of diseased oyster tissues made it difficult to link the injection strain to oyster mortality. As exogenous bacteria can be cleared quickly from the oyster haemolymph (Parisi et al., 2008), the commensal Vibrio species were thought to contribute to the increase of cultivable Vibrio count and the oyster mortality, questioning the use of Vibrio injection challenges in laboratory experiments (Lokmer and Wegner, 2015). Direct injection of pathogens of interest is not considered to reflect the natural route of infection and is also thought to prevent the study of disease transmission (Petton et al., 2019). Although cohabitation experiments are considered more realistic over experiments involving injection challenges (de Lorgeril et al., 2018; Green et al., 2018), cohabitation experiments are not considered to properly address the polymicrobial pathogenesis of oyster mortality associated with OsHV-1 (Petton et al., 2019).

Laboratory experiments conducted in controlled environments have previously being used to study the pathogenesis of OsHV-1 infection enabling the study of the impact of individual factors (Fuhrmann et al., 2016; Paul-Pont et al., 2015; Schikorski et al., 2011). This approach has also been taken in other studies which evaluated genetic resistance of 216

Pacific oysters to OsHV-1 (Azéma et al., 2017b), emersion patterns on OsHV-1 associated mortality (Evans et al., 2019) and the influence of water temperature during the initial replication of OsHV-1 (de Kantzow et al., 2019b).

8.5 Oyster microbiome dynamics during acclimation to the laboratory environment

In Chapter 5 of this thesis, the oyster microbiome was not subjected to any type of seeding from environmental reservoirs except for the nitrifying bacteria that were present in the biofilters of the aquarium and any potential presence of bacteria in the commercial oyster feed. Two groups of oysters from the same hatchery run were grown under one farming condition, then were maintained under two different immersion regimes, i.e. under constant immersion in water or in a simulated tidal environment. Inter-oyster heterogeneity in the microbiome was noted between oysters with a common origin and living in a common environment (Chapter 5). Inter-oyster heterogeneity in the microbiome of individuals has previously been demonstrated in the gill (Wegner et al., 2013), adductor-muscle (King et al., 2018a) and in multiple tissue types (Chapter 4; (Pathirana et al., 2019a)). However, it was demonstrated that the difference (Bray-Curtis dissimilarity) between the microbiomes of different tissue-types of the same oyster was 2.5 times higher than inter-oyster heterogeneity in microbiomes of one tissue-type of individual oysters (Chapter 4).

Temporal microbiome studies are considered to be challenging owing to the changes that occur in the microbiome during acclimation. There has been no standard period for acclimation of Pacific oysters in laboratory-based experiments. Various periods of acclimation are reported including: 2d (Jo et al., 2008), 3d (Pathirana et al., 2019b), 7d (Boutet et al., 2004; de Kantzow et al., 2019a; Zhang and Li, 2006), (Boutet et al., 2004; de Kantzow et al., 2019a; Zhang and Li, 2006), 9d (Petton et al., 2019) and up to 10 days (Evans et al., 2019; Medeiros et al., 2008). Harding et al. (2004) have shown that the physical stress of handling causes changes in several cellular parameters in bivalves including lysosomal stability. Minor mechanical stress such as shaking for 15 min in a plastic container down- regulated the immune functions of oysters by reducing the number of circulating haemocytes and their phagocytic activity (Lacoste et al., 2002). These stress-related changes may alter the course of the disease of interest in an experimental disease challenge.

In Chapter 5 of this thesis, both the gill and gut microbiome of Pacific oysters were subject to change during acclimation to a controlled laboratory environment. The alpha

217

diversity of both gill and gut microbiota did not differ with laboratory acclimation irrespective of the management system. However, the beta diversity of both gill and gut microbiota increased with acclimation to the laboratory environment. Rather than reducing the bacterial diversity in oyster tissues, the acclimation process reduced the abundance of selected bacterial phyla such as phylum Tenericutes and phylum Cyanobacteria, in both gill and gut microbiota (Chapter 5). Meanwhile, the microbiome diversity during acclimation was maintained with increased representation from the phylum Proteobacteria which initially dominated the gill microbiota. This was followed by phylum Bacteroidetes. There was increased abundance of Proteobacteria after laboratory acclimation in studies considering the microbiome of Pacific oysters measured in haemolymph (Lokmer et al., 2016b) and whole tissue homogenates (Romero et al., 2002). Changes in the relative abundance of different bacterial phyla during laboratory acclimation may impact the oyster in different ways. Being a photosynthetic group, phylum Cyanobacteria may not play an important role in any disease pathogenesis. Green and Barnes (2010) demonstrated a reduction in Cyanobacteria in the digestive gland of Saccostrea glomerata affected by QX disease that had ceased feeding, indicating a transient role of Cyanobacteria that might have entered the oysters with feeding. The origin of Cyanobacteria in oyster tissues is considered to be the natural environment (Lasa et al., 2019). Thus, in the absence of seeding from the natural environment, the reduction of Cyanobacteria in the microbiome is self-explanatory. It is, however, interesting to note that Synechococcus elongatus (phylum Cyanobacteria) in the oyster microbiome has a well-developed host–endobiont relationship with oysters (Avila-Poveda et al., 2014).

The steady abundance of the Mycoplasma fraction and phylum Bacteroidetes in the gill microbiome, are positive findings of this study (Chapter 5). Family Mycoplasmataceae has been reported to dominate the microbiota in healthy (non-infected) oysters suggesting a beneficial role in oyster fitness and health status (King et al., 2019). However, an increase in the abundance of Mycoplasma and a decrease in the abundance of phylum Bacteroidetes in the gill microbiome, were a consequence of disturbance by heat stress, starvation and transportation of Pacific oysters (Wegner et al., 2013). The increase in the Vibrio fraction with laboratory acclimation (Chapter 5) cannot be overlooked as Vibrio is considered an important secondary pathogen in OsHV-1 infections (de Lorgeril et al., 2018). The former study showed that bacterial colonization that occurs secondary to OsHV-1 infection was associated with Vibrio along with other opportunistic bacteria in the microbiome of oysters. Lemire et al. (2015) have shown that the non-virulent strains of Vibrio spp. in the oyster

218

microbiome get progressively replaced by virulent Vibrio strains during the onset of an infection, by characterizing the Vibrio strains using partial sequencing of a protein-coding gene (gyrB). These strains were subsequently tested for virulence using an injection model (Lemire et al., 2015). The non-virulent Vibrio strains were expected to facilitate the disease caused by the virulent strains. Injection challenges that combined the virulent strains with non-virulent strains produced a greater mortality (90%) compared to those that had the virulent strains alone (5% mortality), even if the infection was allowed to progress for a longer time (Lemire et al., 2015).

Interestingly, the microbiome changed in the same way over 14 days of acclimation, for oysters that were constantly immersed and those in a tidal system (Chapter 5). Rather, the changes reflected decreases in the relative abundance of specific taxonomic groups, at different taxonomic levels including the increase in Vibrio in gut microbiota and Arcobacter in gill microbiota. The impact of laboratory acclimation on the microbiome should be considered in the design and analysis of experimental studies using Pacific oysters and may be relevant in commercial depuration of oysters. Depuration is a process in which oysters are held in clean seawater under conditions which maximize the natural filtering activity resulting in expulsion of intestinal contents from the oysters and prevents recontamination (Lee et al., 2008).

8.6 Environmental influence on the microbiome and OsHV-1 associated oyster mortality

Understanding the factors and processes that shape the microbiome is essential to understand the stability of the microbiome and thereby the host health (Lokmer et al., 2016b). Living in a highly variable, intertidal estuarine environment the oyster microbiome is subjected to a range of environmental influences and these influences may affect its response to mass mortality associated with OsHV-1. Adult oysters that were grown at an increased height (300 mm above the standard intertidal growing height) demonstrated a 25-50% decrease in mortality compared to those grown at the standard height during outbreaks of disease caused by OsHV-1 (Paul-Pont et al., 2013b; Whittington et al., 2015b). Contrary to the field studies above, a higher mortality was observed in adult oysters maintained in tidal emersion in the laboratory compared to full time immersion (Evans et al., 2019). This suggested that the higher growing height with increased tidal emersion was beneficial only through reduced exposure to OsHV-1. Increased emersion time in the intertidal estuarine environment exposes oysters to increased temperature and hypoxia during valve closure, and

219

alters the immune response of oysters (Allen and Burnett, 2008). Higher temperatures suppressing the bactericidal activity in oyster haemocytes (Allen and Burnett, 2008) and environmental hypoxia reducing the production of reactive oxygen intermediates (ROI) in the haemocytes of C. virginica (Boyd and Burnett, 1999) have been reported in this regard. ROIs are important in the oyster defence system against viral and bacterial infections (Adema et al., 1991; Boyd and Burnett, 1999). Similar to the microbiome of other organisms, the oyster microbiome is considered dynamic and responds to changes and disturbances in the external environment including translocation to a different environment (Lokmer et al., 2016a; Wegner et al., 2013), availability of food (Wegner et al., 2013), changes in environmental temperature and heat shock (Lokmer and Wegner, 2015; Wegner et al., 2013).

With the addition of OsHV-1 infection, the microbiome under the different immersion regimes was differentiated (Chapter 6). Higher concentrations of OsHV-1 DNA and higher mortality in oysters kept in a simulated tide was accompanied by reduced total bacterial quantity. Prolonged periods of valve closure during tidal emersion resulting in hypercapnia and reduction in pH in oyster tissues (Allen and Burnett, 2008) may have occurred in this study, affecting the multiplication and growth of microbiota of tidal oysters. Experimental OsHV-1 infections in oysters under constant immersion have previously seen increases in the total bacterial load (de Lorgeril et al., 2018; Pathirana et al., 2019b). Taken together, the OsHV-1 quantity in oyster tissues and their mortality pattern in Chapter 6, suggested an increased susceptibility to OsHV-1 in tidal oysters, when a variation in exposure to OsHV-1 was introduced in the laboratory, by an infection model. Haemocytes being the primary site of infection for OsHV-1 (de Lorgeril et al., 2018), the reduction of haemocyte killing activity associated with emersion (Allen and Burnett, 2008) may have predisposed the tidal oysters to this increased susceptibility to OsHV-1. The hypercapnia and resulting reduction of pH in oyster tissues during tidal emersion (Allen and Burnett, 2008) may limit the growth of bacteria present in the microbiome that are sensitive to low pH (Flores-Higuera et al., 2019). For example, the bacterial families Rhodobacteraceae and Campylobacteraceae that are found in the Pacific oyster microbiome (King et al., 2019; Pathirana et al., 2019a) are sensitive to low pH (Krause et al., 2012).

Despite the reduction in the total quantity of bacteria, the laboratory tidal emersion supported an increase in bacterial diversity (Chapter 6). This increase may be a potential factor in the discrepancy between the mortality observed in tidal oysters in the field compared

220

to that in the laboratory. Frequently, a loss of diversity is seen in the microbiome during OsHV-1 infection (de Lorgeril et al., 2018; Lasa et al., 2019; Pathirana et al., 2019b). The high mortality disease associated with OsHV-1 that was linked with increasing alpha diversity and reduced total bacterial load indicates that these factors do influence the pathogenesis. Microbiomes exposed to repeated fluctuations in their environment were more diverse and more metabolically active than communities exposed to a constant condition (Shade et al., 2012). Moreover, rare microbial taxa may rapidly respond to changes in the environment and will become abundant (Shade et al., 2012). In line with this concept, increases in the relative abundance of rare bacterial genera (Sediminibacterium, Polaribacter and Marinicella) and members of the phylum Actinobacteria were evident in tidal oysters, indicating a disturbed microbiome (Chapter 6). Although the genus Marinicella was identified to be a low-abundant member of the gill microbiota of tidal oysters (Chapter 6), it dominated the microbiome of OsHV-1 infected oyster spat (Lasa et al., 2019). The same study reported differences in the microbiome of adult oysters and spat where the Vibrio- infected adult oyster microbiome was dominated by Vibrio and Arcobacter, indicating an influence of age on the microbiome composition.

Increased seawater temperatures (>20°C) have been associated with Pacific oyster mortality outbreaks in the absence of OsHV-1 (Go et al., 2017). Microbiome analysis of the oyster tissues of this outbreak revealed an increase in rare OTUs belonging to the genus Vibrio (King et al., 2018a). A disease of oyster spat that was triggered by increasing seawater temperature from 20°C to 25°C resulted in 77.4% mortality in an experimental system (Green et al., 2018). However, mortality was much lower (4.3%) in spat that were administered penicillin-streptomycin (a broad-spectrum antibiotic) through water, under the same conditions. The temperature increase was associated with a 324-fold increase in Vibrio harveyi quantity (along with an increase of Arcobacter) with a concurrent decrease in families Rhodobacteraceae and Flavobacteriaceae (Green et al., 2018). While the relative abundance of Vibrio was reduced from 40.5% to 2.2% by the antibiotic treatment, Rhodobacteraceae and Flavobacteriaceae dominated the microbiome after the antibiotic treatment. These studies indicated a role of Vibrio in temperature-associated oyster mortality.

An interesting consideration in the role of temperature as a risk factor for disease is the dynamic nature in the intertidal and estuarine environments inhabited by oysters (Allen and Burnett, 2008; Lokmer et al., 2016b). In Chapter 7, the impact of different seawater

221

temperature profiles including diurnal fluctuation were considered with respect to changes in the Pacific oyster microbiome and the response to OsHV-1 infection associated disease, in a controlled laboratory environment. Although the Vibrio fraction in the gill microbiome did not change with different temperature profiles of this study, there was an increase in Vibrio abundance for all temperature profiles, after OsHV-1 challenge. Similar to the comparatively higher OsHV-1 DNA content seen at higher temperatures in other studies involving experimental infections (de Kantzow et al., 2016; Delisle et al., 2018), the oysters at 26°C in this study, also had the highest OsHV-1 DNA content with a higher cumulative mortality (Chapter 7). The quantity of OsHV-1 DNA in moribund oysters at 26°C has been shown to be approximately 6-fold greater than that at 18°C and 1.2-fold greater than that at 22°C (de Kantzow et al., 2016). This increase of OsHV-1 can be associated with immunosuppression related to the increased seawater temperature (Gagnaire et al., 2006; Malham et al., 2009). On the other hand, shifts in the oyster microbiome and an increase of the Vibrio fraction after the OsHV-1 challenge (highest at 26°C), may have driven the pathogenesis resulting in a higher oyster mortality (Chapter 7). In corals, high temperatures have been shown to be selective for strains of potentially pathogenic Vibrio spp. that were in low abundance in the environment (Thurber et al., 2009). Green et al. (2018) demonstrated the likely involvement of Vibrio in temperature-induced mortality using an antibiotic treatment for Vibrio while maintaining the other members of the oyster microbiome. In a study conducted in Germany using oysters collected from the Wadden Sea and acclimated at 8°C, the oyster microbiome was shown to be disturbed in elevated seawater temperatures (22°C), favouring shifts in the composition towards pathogen-dominated communities (Le Roux et al., 2016; Lokmer and Wegner, 2015). Shifts in the gill microbiome were seen after acclimating to higher temperatures (Chapter 7). While Mycoplasma increased in abundance at a constant temperature of 20°C, an increase in the Vibrio fraction was evident at all temperatures of this study, 48 hours after the OsHV-1 challenge, with the highest tissue concentration at 26°C. This observation re- confirmed the secondary, opportunistic role of Vibrio.

8.7 Vibrio as a commensal organism and as a pathogen in oysters

Pathogenic Vibrio species have been isolated from natural mass mortality events associated with OsHV-1 (Lasa et al., 2019; Petton et al., 2015b) and changes in the Vibrio dynamics have been observed in experimental infections (Pathirana et al., 2019b; Petton et al., 2019). Further, Vibrio have been identified to act as secondary opportunistic pathogens in OsHV-1 infections (de Lorgeril et al., 2018). A strong correlation was observed between 222

OsHV-1 and Vibrio in oyster tissues infected with OsHV-1 (Chapter 3; (Pathirana et al., 2019b)). The same organisms peaked in the seawater during the first 48 h of an experimental OsHV-1 infection (Petton et al., 2019). Strengthening these observations, Chapter 7 of this thesis demonstrated increase in abundance of the Vibrio fraction in oysters, concurrent with differences in the OsHV-1 content and mortality in different temperature profiles. With this background, further studies are required to identify the exact role of Vibrio in pathogenesis of mass mortality associated with OsHV-1. Various pathogenic Vibrio spp. have been isolated from Pacific oysters during mass mortality outbreaks and the isolates have been tested for virulence, causing mortality in oysters. This includes disease records in which OsHV-1 was present (Keeling et al., 2014) and when it was not (De Decker and Saulnier, 2011; Garnier et al., 2007; King et al., 2018a). A retrospective study of Pacific oysters infected by Vibrio aestuarianus also showed the presence of other potential pathogenic Vibrio species including V. coralliilyticus, V. splendidus, V. crassostreae, V.tasmaniensis at a lower relative abundance (Lasa et al., 2019). These species were thought to be commensal Vibrio in the oyster microbiome that acquired a role as opportunistic pathogens.

The proliferation of Vibrio was observed to be different in different batches of oysters with a different microbiome, following an experimental OsHV-1 infection (Chapter 3). The mortality and the OsHV-1 quantity were also different between these batches of oysters. While the oysters that displayed the highest Vibrio load also demonstrated a higher and early mortality, it was not observed in oysters from the other two batches. In a recent study which analysed the microbiome of Pacific oysters from recurrent mortality episodes in Europe, OsHV-1-infected oyster spat showed an increase in the Vibrio fraction compared to non- infected oysters (Lasa et al., 2019).

A low average relative abundance of Vibrio spp. was a common observation in the microbiome analyses carried out in this thesis. This was consistent with previous studies reporting natural and experimental oyster mortality events (King et al., 2018a; Lokmer and Wegner, 2015). Contrary to the conventional knowledge that identifies Vibrio as one of the dominating bacterial genera in oysters (Colwell and Liston, 1960; Prieur et al., 1990), a large proportion of the oyster microbiome cannot be cultivated by standard procedures (Fernandez- Piquer et al., 2012; Romero and Espejo, 2001). Low relative abundance of Vibrio is a common finding in other oyster species such as Eastern oysters (C. virginica) and Sydney rock oysters (Saccostrea glomerata) (Green and Barnes, 2010; Ossai et al., 2017).

223

Nevertheless, Vibrio spp. have been consistently identified and highlighted as potential contributors to poor health of oysters, despite the low relative abundance (Thurber et al., 2009).

The increase of Vibrio observed with laboratory acclimation (Chapter 5) can be compared to the increase of Vibrio seen with depuration of oysters (Vezzulli et al., 2018). Depuration has not always been effective in removing naturally occurring marine vibrios such as V. parahaemolyticus and V. vulnificus (Lee et al., 2008) that are pathogenic to humans and V. aestuarianus and V. splendidus clade that are pathogenic to oysters (Lacoste et al., 2001; Vezzulli et al., 2018). Vezzulli et al. (2018) demonstrated an increase in the Vibrio fraction after depuration, in the haemolymph of C. gigas, but not in the digestive gland. The depuration process was carried out using clean seawater that was treated in a sequence of steps including sand filtration, UV treatment, ozonation (infusion of ozone into water), and biofiltration, to remove contaminants (Vezzulli et al., 2018). The same study showed an increase in Vibrio in the digestive gland and haemolymph of the Mediterranean mussel (Mytilus galloprovincialis), after depuration. Vezzulli et al. (2018) interpreted the persistence of Vibrio during depuration as a feature of the possibly long co- evolutionary history of Vibrio with their invertebrate hosts, representing permanent residents of the microbiota. The abundance of Vibrio in the tissues of C. gigas maintained in artificial seawater (Lokmer et al., 2016a) was thought to be a result of the static conditions in the tanks (Lokmer et al., 2016b). Strengthening the concept of a co-evolutionary role of Vibrio, Le Roux et al. (2016) also suggest that the pathogenesis of oyster disease involves a coevolutionary interplay between vibrios, oysters, and their microbiota. In this regard, a secondary opportunistic role has already been demonstrated in OsHV-1 infections (de Lorgeril et al., 2018). Lemire et al. (2015) showed that the onset of oyster mortality is associated with progressive replacement of non-pathogenic, commensal Vibrio by taxonomically related pathogenic strains. Further, they showed that the commensal Vibrio are instrumental in the pathogenesis of the disease by positively contributing to the mortality.

It is also important to investigate the factors that contribute to the multiplication of Vibrio and Arcobacter over the other members of the microbiome, in the event of an immunosuppression caused by a primary infection by OsHV-1.

224

8.8 Conclusion

Pacific oyster mortality outbreaks are a major concern for oyster farming throughout the world. The oyster microbiome is thought to play a role in the pathogenesis Pacific oyster mortality associated with OsHV-1 and other diseases. A tissue-specific microbiome was identified in the Pacific oyster. The oyster microbiome is influenced by the surrounding environment, differentiating the microbiomes of oysters with a common origin. This influence further determined the outcome of an OsHV-1 infection, resulting in a differential response, based on the microbiome composition before an OsHV-1 challenge. The increase of Vibrio following an OsHV-1 challenge appeared to vary with the microbiome composition before the viral challenge indicating an involvement of Vibrio in the polymicrobial pathogenesis of OsHV-1 associated mortality. The increase of Vibrio was also associated with the outcome of infection. Acclimation of oysters to a laboratory environment reduced the abundance of phyla including Tenericutes and Cyanobacteria while increasing the abundance of dominating phyla (Proteobacteria and Bacteroidetes) rather than reducing the bacterial diversity in oyster tissues. Being considered as a secondary, opportunistic pathogen, the increase of the Vibrio fraction in the gut microbiota with acclimation, was important. Further studies are required to confirm that the increase in Vibrio during laboratory acclimation is a maladaptation which supports the pathogenesis of OsHV-1 infections.

The immersion regimes used in this thesis, did not create in any difference in oyster microbiome dynamics between constantly immersed oysters and tidal oysters, during acclimation. However, the response of the microbiome was differentiated by the immersion regime, when oysters were challenged with OsHV-1, indicating a polymicrobial pathogenesis driven by OsHV-1. This differential microbial response may have contributed to the differential mortality between constantly immersed and tidally emersed oysters. Although the gill microbiome was more stable than the gut microbiome in healthy oysters, it responded to the OsHV-1 challenge differently between the two immersion regimes. Unlike with different regimes, increased seawater temperatures affected the microbiome even before a disease process was initiated by OsHV-1. However, further changes in the microbiome were observed after the OsHV-1 challenge which included a differential increase in the Vibrio fraction. The increased abundance of Vibrio at higher seawater temperatures seemed supportive of the existing role of Vibrio as a secondary opportunistic pathogen in OsHV-1 infections. The insights provided in this thesis regarding the microbiome dynamics under

225

different environmental conditions were in line with the current view that OsHV-1 causes the primary infection which is followed by the involvement of bacteria present in the microbiome that take up the role of secondary, opportunistic pathogens. The microbial dynamics observed in this thesis, particularly those of Vibrio and Arcobacter provide supportive evidence for this secondary opportunistic role of bacteria.

The knowledge generated in this thesis, can be used to understand how external environmental factors such as seawater temperature and the immersion regime can shape the Pacific oyster microbiome and how environmental factors can be controlled/adjusted to maintain a healthy composition in the microbiome particularly in the face of harsh environmental conditions. A knowledge of the shifts in the microbiome particularly those that favour mortality associated with OsHV-1 and the surviving microbiome, may provide insights into the manipulation of the microbiome by means probiotic treatments as a disease management strategy. Understanding the influence of the surrounding environment on the microbiome and the seawater microbial dynamics may help early identification of adverse conditions in seawater. Furthermore, oyster microbiome studies need to be extended beyond this level to understand and identify the functional significance of the microbiome dynamics observed under various environmental conditions and how they will contribute to the pathogenesis of oyster mortality associated with OsHV-1.

226

REFERENCES

Abolmaaty, A., Gu, W., Witkowsky, R., Levin, R.E., 2007. The use of activated charcoal for the removal of PCR inhibitors from oyster samples. Journal of Microbiological Methods. 68, 349-352. Adair, K.L., Douglas, A.E., 2017. Making a microbiome: the many determinants of host- associated microbial community composition. Current Opinion in Microbiology. 35, 23-29. Adema, C., Van der Knaap, W., Sminia, T., 1991. Ediated Cytotoxicity: the Role of Reactive Oxygen Intermediates. Reviews in Aquatic Sciences. 4, 201-223. Aguirre, A.A., Balazs, G.H., Zimmerman, B., Spraker, T.R., 1994. Evaluation of Hawaiian green turtles (Chelonia mydas) for potential pathogens associated with fibropapillomas. Journal of Wildlife Diseases. 30, 8-15. Al-Soud, W.A., Rådström, P., 1998. Capacity of nine thermostable DNA polymerases to mediate DNA amplification in the presence of PCR-inhibiting samples. Applied and Environmental Microbiology. 64, 3748-3753. Alfaro, A.C., Nguyen, T.V., Merien, F., 2019. The complex interactions of Ostreid herpesvirus 1, Vibrio bacteria, environment and host factors in mass mortality outbreaks of Crassostrea gigas. Reviews in Aquaculture. 11, 1148-1168. Allen, S.M., Burnett, L.E., 2008. The effects of intertidal air exposure on the respiratory physiology and the killing activity of hemocytes in the pacific oyster, Crassostrea gigas (Thunberg). Journal of Experimental Marine Biology and Ecology. 357, 165-171. Allwood, A.C., Walter, M.R., Kamber, B.S., Marshall, C.P., Burch, I.W., 2006. Stromatolite reef from the Early Archaean era of Australia. Nature. 441, 714-718. Anderson, M.J., 2001. A new method for non‐parametric multivariate analysis of variance. Austral Ecology. 26, 32-46. Andrews, S., 2010. FastQC: a quality control tool for high throughput sequence data. August 17, 2018. http://www.bioinformatics.babraham.ac.uk/projects/fastqc/ Arfken, A., Song, B., Bowman, J.S., Piehler, M., 2017. Denitrification potential of the eastern oyster microbiome using a 16S rRNA gene based metabolic inference approach. PLoS One. 12, e0185071. Arzul, I., Renault, T., Lipart, C., Davison, A.J., 2001a. Evidence for interspecies transmission of oyster herpesvirus in marine bivalves. Journal of General Virology. 82, 865-870. Arzul, I., Nicolas, J.-L., Davison, A.J., Renault, T., 2001b. French scallops: a new host for Ostreid herpesvirus-1. Virology. 290, 342-349. Arzul, I., Renault, T., Thébault, A., Gérard, A., 2002. Detection of oyster herpesvirus DNA and proteins in asymptomatic Crassostrea gigas adults. Virus Research. 84, 151-160. Avila-Poveda, O.H., Torres-Arino, A., Girón-Cruz, D.A., Cuevas-Aguirre, A., 2014. Evidence for accumulation of Synechococcus elongatus (Cyanobacteria:Cyanophyceae) in the tissues of the oyster Crassostrea gigas (Mollusca:Bivalvia). Tissue and Cell. 46, 379- 387. Azéma, P., Maurouard, E., Lamy, J.-B., Dégremont, L., 2017a. The use of size and growing height to improve Crassostrea gigas farming and breeding techniques against OsHV- 1. Aquaculture. 471, 121-129. Azéma, P., Lamy, J.-B., Boudry, P., Renault, T., Travers, M.-A., Dégremont, L., 2017b. Genetic parameters of resistance to Vibrio aestuarianus, and OsHV-1 infections in the Pacific oyster, Crassostrea gigas, at three different life stages. Genetics Selection Evolution. 49, 23. Azevedo, C., 1993. Cerastoderma edule (Mollusca, Bivalvia). Diseases of Aquatic Organisms. 16, 55-59.

227

Bai, C., Wang, C., Xia, J., Sun, H., Zhang, S., Huang, J., 2015. Emerging and endemic types of Ostreid herpesvirus 1 were detected in bivalves in China. Journal of Invertebrate Pathology. 124, 98-106. Bano, N., deRae Smith, A., Bennett, W., Vasquez, L., Hollibaugh, J.T., 2007. Dominance of Mycoplasma in the guts of the Long‐Jawed Mudsucker, Gillichthys mirabilis, from five California salt marshes. Environmental Microbiology. 9, 2636-2641. Batista, F.M., López-Sanmartín, M., Grade, A., Morgado, I., Valente, M., Navas, J.I., Power, D.M., Ruano, F., 2015. Sequence variation in ostreid herpesvirus 1 microvar isolates detected in dying and asymptomatic Crassostrea angulata adults in the Iberian Peninsula: Insights into viral origin and spread. Aquaculture. 435, 43-51. Berg, R.D., 1996. The indigenous gastrointestinal microflora. Trends in Microbiology. 4, 430- 435. Blaxter, M., Mann, J., Chapman, T., Thomas, F., Whitton, C., Floyd, R., Abebe, E., 2005. Defining operational taxonomic units using DNA barcode data. Philosophical Transactions of the Royal Society B: Biological Sciences. 360, 1935-1943. Bokulich, N.A., Dillon, M.R., Bolyen, E., Kaehler, B.D., Huttley, G.A., Caporaso, J.G., 2018. q2-sample-classifier: machine-learning tools for microbiome classification and regression. Journal of Open Research Software. 3, 30. Bolyen, E., Rideout, J.R., Dillon, M.R., Bokulich, N.A., Abnet, C.C., Al-Ghalith, G.A., Alexander, H., Alm, E.J., Arumugam, M., Asnicar, F., 2019. Reproducible, interactive, scalable and extensible microbiome data science using QIIME 2. Nature Biotechnology. 37, 852-857. Bourlès, Y., Alunno-Bruscia, M., Pouvreau, S., Tollu, G., Leguay, D., Arnaud, C., Goulletquer, P., Kooijman, S., 2009. Modelling growth and reproduction of the Pacific oyster Crassostrea gigas: advances in the oyster-DEB model through application to a coastal pond. Journal of Sea Research. 62, 62-71. Boutet, I., Tanguy, A., Moraga, D., 2004. Response of the Pacific oyster Crassostrea gigas to hydrocarbon contamination under experimental conditions. Gene. 329, 147-157. Boutin, S., Bernatchez, L., Audet, C., Derôme, N., 2013. Network analysis highlights complex interactions between pathogen, host and commensal microbiota. PLoS One. 8, e84772. Bower, S.M., McGladdery, S.E., Price, I.M., 1994. Synopsis of infectious diseases and parasites of commercially exploited shellfish. Annual Review of Fish Diseases. 4, 1- 199. Boyd, J.N., Burnett, L.E., 1999. Reactive oxygen intermediate production by oyster hemocytes exposed to hypoxia. Journal of Experimental Biology. 202, 3135-3143. Brink, A., Van Straten, A., Van Rensburg, A., 1995. Shewanella (Pseudomonas) putrefaciens bacteremia. Clinical Infectious Diseases. 20, 1327-1332. Brogden, K.A., 2002. Polymicrobial diseases of animals and humans. in: Brogden KA, Guthmiller JM (Eds.), Polymicrobial diseases. ASM press, Washington (DC). Bruto, M., James, A., Petton, B., Labreuche, Y., Chenivesse, S., Alunno-Bruscia, M., Polz, M.F., Le Roux, F., 2017. Vibrio crassostreae, a benign oyster colonizer turned into a pathogen after plasmid acquisition. ISME Journal 11, 1043-1052. Buck, J.D., Cleverdon, R.C., 1960. The spread plate as a method for the enumeration of marine bacteria. Limnology and Oceanography. 5, 78-80. Buller, N.B., 2014. Preparation of media for culture and identification. in: Buller, N.B. (Ed.), Bacteria and fungi from fish and other aquatic animals: a practical identification manual. CAB International, Oxfordshire, UK, pp. 725-772. Burge, C.A., Griffin, F.J., Friedman, C.S., 2006. Mortality and herpesvirus infections of the Pacific oyster Crassostrea gigas in Tomales Bay, California, USA. Diseases of aquatic organisms. 72, 31-43.

228

Burge, C.A., Mark Eakin, C., Friedman, C.S., Froelich, B., Hershberger, P.K., Hofmann, E.E., Petes, L.E., Prager, K.C., Weil, E., Willis, B.L., 2014. Climate change influences on marine infectious diseases: implications for management and society. Annual Review of Marine Science. 6, 249-277. Burgsdorf, I., Erwin, P.M., López-Legentil, S., Cerrano, C., Haber, M., Frenk, S., Steindler, L., 2014. Biogeography rather than association with cyanobacteria structures symbiotic microbial communities in the marine sponge Petrosia ficiformis. Frontiers in Microbiology. 5, 529 Burioli, E., Prearo, M., Riina, M., Bona, M., Fioravanti, M., Arcangeli, G., Houssin, M., 2016. Ostreid herpesvirus type 1 genomic diversity in wild populations of Pacific oyster Crassostrea gigas from Italian coasts. Journal of Invertebrate Pathology. 137, 71-83. Cai, T., Jiang, L., Yang, C., Huang, K., 2006. Application of real-time PCR for quantitative detection of Vibrio parahaemolyticus from seafood in Eastern China. FEMS Immunology and Medical Microbiology. 46, 180-186. Callahan, B.J., McMurdie, P.J., Rosen, M.J., Han, A.W., Johnson, A.J.A., Holmes, S.P., 2016. DADA2: high-resolution sample inference from Illumina amplicon data. Nature methods. 13, 581. Calvo, G.W., Luckenbach, M.W., Burreson, E.M., 1999. A comparative field study of Crassostrea gigas and Crassostrea virginica in relation to salinity in Virginia. Campbell, B.J., Yu, L., Heidelberg, J.F., Kirchman, D.L., 2011. Activity of abundant and rare bacteria in a coastal ocean. Proceedings of the National Academy of Sciences. 108, 12776-12781. Caporaso, J.G., Lauber, C.L., Walters, W.A., Berg-Lyons, D., Lozupone, C.A., Turnbaugh, P.J., Fierer, N., Knight, R., 2011. Global patterns of 16S rRNA diversity at a depth of millions of sequences per sample. Proceedings of the National Academy of Sciences. 108, 4516-4522. Caporaso, J.G., Kuczynski, J., Stombaugh, J., Bittinger, K., Bushman, F.D., Costello, E.K., Fierer, N., Pena, A.G., Goodrich, J.K., Gordon, J.I., 2010. QIIME allows analysis of high-throughput community sequencing data. Nature Methods. 7, 335-336. Cebrian, E., Uriz, M.J., Garrabou, J., Ballesteros, E., 2011. Sponge mass mortalities in a warming Mediterranean Sea: are cyanobacteria-harboring species worse off? PLoS One. 6, e20211. Chakravorty, S., Helb, D., Burday, M., Connell, N., Alland, D., 2007. A detailed analysis of 16S ribosomal RNA gene segments for the diagnosis of pathogenic bacteria. Journal of Microbiological Methods. 69, 330-339. Chaney, M.L., Gracey, A.Y., 2011. Mass mortality in Pacific oysters is associated with a specific gene expression signature. Molecular Ecology. 20, 2942-2954. Choo, J.M., Leong, L.E., Rogers, G.B., 2015. Sample storage conditions significantly influence faecal microbiome profiles. Scientific Reports. 5, 16350. Clarridge, J.E., 2004. Impact of 16S rRNA gene sequence analysis for identification of bacteria on clinical microbiology and infectious diseases. Clinical Microbiology Reviews. 17, 840-862. Clegg, T.A., Morrissey, T., Geoghegan, F., Martin, S.W., Lyons, K., Ashe, S., More, S.J., 2014. Risk factors associated with increased mortality of farmed Pacific oysters in Ireland during 2011. Preventive Veterinary Medicine. 113, 257-267. Colwell, R., Liston, J., 1960. Microbiology of shellfish. Bacteriological study of the natural flora of Pacific oysters (Crassostrea gigas). Applied Microbiology. 8, 104-109. Costello, E.K., Lauber, C.L., Hamady, M., Fierer, N., Gordon, J.I., Knight, R., 2009. Bacterial community variation in human body habitats across space and time. Science. 326, 1694- 1697.

229

Cox, D.R., 1972. Regression models and life‐tables. Journal of the Royal Statistical Society: Series B (Methodological). 34, 187-202. Davison, A.J., Trus, B.L., Cheng, N., Steven, A.C., Watson, M.S., Cunningham, C., Le Deuff, R.-M., Renault, T., 2005. A novel class of herpesvirus with bivalve hosts. Journal of General Virology. 86, 41-53. Davison, A.J., Eberle, R., Ehlers, B., Hayard, G.S., McGeoch, D.J., Minson, A.M., Pellett, P.E., Roizman, B., Studdert, M.J., Thiry, E., 2009. The order Herpesvirales. Archives in Virology. 154. De Decker, S., Saulnier, D., 2011. Vibriosis induced by experimental cohabitation in Crassostrea gigas: evidence of early infection and down-expression of immune-related genes. Fish and Shellfish Immunology. 30, 691-699. De Decker, S., Normand, J., Saulnier, D., Pernet, F., Castagnet, S., Boudry, P., 2011. Responses of diploid and triploid Pacific oysters Crassostrea gigas to Vibrio infection in relation to their reproductive status. Journal of Invertebrate Pathology. 106, 179-191. de Kantzow, M., Hick, P., Becker, J.A., Whittington, R.J., 2016. Effect of water temperature on mortality of Pacific oysters Crassostrea gigas associated with microvariant Ostreid herpesvirus 1 (OsHV-1 µVar). Aquaculture Environment Interactions. 8, 419-428. de Kantzow, M.C., Hick, P.M., Dhand, N.K., Whittington, R.J., 2017. Risk factors for mortality during the first occurrence of Pacific Oyster Mortality Syndrome due to Ostreid herpesvirus–1 in Tasmania, 2016. Aquaculture. 468, 328-336. de Kantzow, M.C., Whittington, R.J., Hick, P., 2019a. Prior exposure to Ostreid herpesvirus 1 (OsHV-1) at 18° C is associated with improved survival of juvenile Pacific oysters (Crassostrea gigas) following challenge at 22° C. Aquaculture. 507, 443-450. de Kantzow, M.C., Whittington, R.J., Hick, P.M., 2019b. Different in vivo growth of Ostreid herpesvirus 1 at 18° C and 22° C alters mortality of Pacific oysters (Crassostrea gigas). Archives of Virology, 1-9. De Ley, J., Segers, P., Kersters, K., Mannheim, W., Lievens, A., 1986. Intra-and intergeneric similarities of the Bordetella ribosomal ribonucleic acid cistrons: proposal for a new family, Alcaligenaceae. International Journal of Systematic and Evolutionary Microbiology. 36, 405-414. de Lorgeril, J., Lucasson, A., Petton, B., Toulza, E., Montagnani, C., Clerissi, C., Vidal-Dupiol, J., Chaparro, C., Galinier, R., Escoubas, J.-M., 2018. Immune-suppression by OsHV-1 viral infection causes fatal bacteraemia in Pacific oysters. Nature Communications. 9, 4215. Defer, D., Desriac, F., Henry, J., Bourgougnon, N., Baudy-Floc'h, M., Brillet, B., Le Chevalier, P., Fleury, Y., 2013. Antimicrobial peptides in oyster hemolymph: the bacterial connection. Fish and Shellfish Immunology. 34, 1439-1447. Dégremont, L., 2011. Evidence of herpesvirus (OsHV-1) resistance in juvenile Crassostrea gigas selected for high resistance to the summer mortality phenomenon. Aquaculture. 317, 94-98. Dégremont, L., 2013. Size and genotype affect resistance to mortality caused by OsHV-1 in Crassostrea gigas. Aquaculture. 416, 129-134. Dégremont, L., Garcia, C., Allen Jr, S.K., 2015. Genetic improvement for disease resistance in oysters: a review. Journal of Invertebrate Pathology. 131, 226-241. Dégremont, L., Ernande, B., Bédier, E., Boudry, P., 2007. Summer mortality of hatchery- produced Pacific oyster spat (Crassostrea gigas). I. Estimation of genetic parameters for survival and growth. Aquaculture. 262, 41-53. Dégremont, L., Boudry, P., Ropert, M., Samain, J.-F., Bédier, E., Soletchnik, P., 2010. Effects of age and environment on survival of summer mortality by two selected groups of the Pacific oyster Crassostrea gigas. Aquaculture. 299, 44-50.

230

Dégremont, L., Bédier, E., Soletchnik, P., Ropert, M., Huvet, A., Moal, J., Samain, J.-F., Boudry, P., 2005. Relative importance of family, site, and field placement timing on survival, growth, and yield of hatchery-produced Pacific oyster spat (Crassostrea gigas). Aquaculture. 249, 213-229. Delaporte, M., Soudant, P., Lambert, C., Jegaden, M., Moal, J., Pouvreau, S., Dégremont, L., Boudry, P., Samain, J.-F., 2007. Characterisation of physiological and immunological differences between Pacific oysters (Crassostrea gigas) genetically selected for high or low survival to summer mortalities and fed different rations under controlled conditions. Journal of Experimental Marine Biology and Ecology. 353, 45-57. Delisle, L., Petton, B., Burguin, J.F., Morga, B., Corporeau, C., Pernet, F., 2018. Temperature modulate disease susceptibility of the Pacific oyster Crassostrea gigas and virulence of the Ostreid herpesvirus type 1. Fish & Shellfish immunology. 80, 71-79. DeSantis, T.Z., Hugenholtz, P., Larsen, N., Rojas, M., Brodie, E.L., Keller, K., Huber, T., Dalevi, D., Hu, P., Andersen, G.L., 2006. Greengenes, a chimera-checked 16S rRNA gene database and workbench compatible with ARB. Applied Environmental Microbiology. 72, 5069-5072. Desriac, F., Le Chevalier, P., Brillet, B., Leguerinel, I., Thuillier, B., Paillard, C., Fleury, Y., 2014. Exploring the hologenome concept in marine bivalvia: haemolymph microbiota as a pertinent source of probiotics for aquaculture. FEMS Microbiology Letters. 350, 107-116. Dineen, S., Aranda, R., Anders, D., Robertson, J., 2010. An evaluation of commercial DNA extraction kits for the isolation of bacterial spore DNA from soil. Journal of Applied Microbiology. 109, 1886-1896. Dundon, W.G., Arzul, I., Omnes, E., Robert, M., Magnabosco, C., Zambon, M., Gennari, L., Toffan, A., Terregino, C., Capua, I., Arcangeli, G., 2011. Detection of Type 1 Ostreid Herpes variant (OsHV-1 μvar) with no associated mortality in French-origin Pacific cupped oyster Crassostrea gigas farmed in Italy. Aquaculture. 314, 49-52. Edgar, R.C., 2010. Search and clustering orders of magnitude faster than BLAST. Bioinformatics. 26, 2460-2461. Edgar, R.C., Flyvbjerg, H., 2015. Error filtering, pair assembly and error correction for next- generation sequencing reads. Bioinformatics. 31, 3476-3482. Edgar, R.C., Haas, B.J., Clemente, J.C., Quince, C., Knight, R., 2011. UCHIME improves sensitivity and speed of chimera detection. Bioinformatics. 27, 2194-2200. EFSA, 2010. Scientific Opinion on the increased mortality events in Pacific oysters, Crassostrea gigas. EFSA Journal. 8, 1894. Elston, R.A., 1993. Infectious diseases of the Pacific oyster, Crassostrea gigas. Annual Review of Fish Diseases. 3, 259-276. Elston, R.A., Hasegawa, H., Humphrey, K.L., Polyak, I.K., Häse, C.C., 2008. Re-emergence of Vibrio tubiashii in bivalve shellfish aquaculture: severity, environmental drivers, geographic extent and management. Diseases of Aquatic Organisms. 82, 119-134. Erwin, P.M., Pita, L., Lopez-Legentil, S., Turon, X., 2012. Stability of sponge-associated bacteria over large seasonal shifts in temperature and irradiance. Applied Environmental Microbiology. 78, 7358-7368. Estes, R.M., Friedman, C.S., Elston, R.A., Herwig, R.P., 2004. Pathogenicity testing of shellfish hatchery bacterial isolates on Pacific oyster Crassostrea gigas larvae. Diseases of Aquatic Organisms. 58, 223-230. Evans, O.M., 2016. Transmission of Ostreid herpesvirus-1 in the Pacific oyster (Crassostrea gigas). PhD Thesis. Sydney School of Veterinary Science, Faculty of Science. The University of Sydney. Sydney, Australia.

231

Evans, O., Hick, P., Dhand, N., Whittington, R.J., 2015. Transmission of Ostreid herpesvirus- 1 in Crassostrea gigas by cohabitation: effects of food and number of infected donor oysters. Aquaculture Environment Interactions. 7, 281-295. Evans, O., Kan, J.Z., Pathirana, E., Whittington, R.J., Dhand, N., Hick, P., 2019. Effect of emersion on the mortality of Pacific oysters (Crassostrea gigas) infected with Ostreid herpesvirus-1 (OsHV-1). Aquaculture. 505, 157-166. Evans, O., Paul-Pont, I., Whittington, R.J., 2017. Detection of Ostreid herpesvirus 1 microvariant DNA in aquatic invertebrate species, sediment and other samples collected from the Georges River estuary, New South Wales, Australia. Diseases of Aquatic Organisms. 122, 247-255. Fabbro, C., Celussi, M., Russell, H., Del Negro, P., 2012. Phenotypic and genetic diversity of coexisting Listonella anguillarum, Vibrio harveyi and Vibrio chagassi recovered from skin haemorrhages of diseased sand smelt, Atherina boyeri, in the Gulf of Trieste (NE Adriatic Sea). Letters in Applied Microbiology. 54, 153-159. Fan, L., Liu, M., Simister, R., Webster, N.S., Thomas, T., 2013. Marine microbial symbiosis heats up: the phylogenetic and functional response of a sponge holobiont to thermal stress. The ISME Journal. 7, 991-1002. FAO, 2005. Cultured Aquatic Species Information Programme. Crassostrea gigas. Cultured Aquatic Species Information Programme. February 8, 2020. http://www.fao.org/fishery/culturedspecies/Crassostrea_gigas/en FAO, 2014. FAO Yearbook - Fishery & Aquaculture Statistics 2012. Nations, Food and Agriculture Organization of the United Nations, Rome, Italy. 2014. Faury, N., Saulnier, D., Thompson, F., Gay, M., Swings, J., Le Roux, F., 2004. Vibrio crassostreae sp. nov., isolated from the haemolymph of oysters (Crassostrea gigas). International Journal of Systematic and Evolutionary Microbiology. 54, 2137-2140. Feehery, G.R., Yigit, E., Oyola, S.O., Langhorst, B.W., Schmidt, V.T., Stewart, F.J., Dimalanta, E.T., Amaral-Zettler, L.A., Davis, T., Quail, M.A., Pradhan, S., 2013. A method for selectively enriching microbial DNA from contaminating vertebrate host DNA. PLoS One. 8, e76096. Fernandez-Piquer, J., Bowman, J.P., Ross, T., Tamplin, M.L., 2012. Molecular analysis of the bacterial communities in the live Pacific oyster (Crassostrea gigas) and the influence of postharvest temperature on its structure. Journal of Applied Microbiology. 112, 1134-1143. Flores-Higuera, F.A., Luis-Villaseñor, I.E., Rochin-Arenas, J.A., Gómez-Gil, B., Mazón- Suástegui, J.M., Voltolina, D., Medina-Hernández, D., 2019. Effect of pH on the bacterial community present in larvae and spat of Crassostrea gigas. Latin American Journal of Aquatic Research. 47, 513-523. Frank, J.A., Reich, C.I., Sharma, S., Weisbaum, J.S., Wilson, B.A., Olsen, G.J., 2008. Critical evaluation of two primers commonly used for amplification of bacterial 16S rRNA genes. Applied Environmental Microbiology. 74, 2461-2470. Fricker, M., Messelhäußer, U., Busch, U., Scherer, S., Ehling-Schulz, M., 2007. Diagnostic real-time PCR assays for the detection of emetic Bacillus cereus strains in foods and recent food-borne outbreaks. Applied and Environmental Microbiology. 73, 1892-1898. Friedman, C.S., Hedrick, R.P., 1991. Pacific oyster nocardiosis: isolation of the bacterium and induction of laboratory infections. Journal of Invertebrate Pathology. 57, 109-120. Friedman, C.S., Estes, R.M., Stokes, N.A., Burge, C.A., Hargove, J.S., Barber, B.J., Elston, R.A., Burreson, E.M., Reece, K.S., 2005. Herpes virus in juvenile Pacific oysters Crassostrea gigas from Tomales Bay, California, coincides with summer mortality episodes. Diseases of Aquatic Organisms. 63, 33-41.

232

Fuhrmann, M., Delisle, L., Petton, B., Corporeau, C., Pernet, F., 2018. Metabolism of the Pacific oyster, Crassostrea gigas, is influenced by salinity and modulates survival to the Ostreid herpesvirus OsHV-1. Biology Open. 7, bio028134. Fuhrmann, M., Richard, G., Quéré, C., Petton, B., Pernet, F., 2019. Low pH reduced survival of the oyster Crassostrea gigas exposed to the Ostreid herpesvirus 1 by altering the metabolic response of the host. Aquaculture. 503, 167-174. Fuhrmann, M., Petton, B., Quillien, V., Faury, N., Morga, B., Pernet, F., 2016. Salinity influences disease-induced mortality of the oyster Crassostrea gigas and infectivity of the Ostreid herpesvirus 1 (OsHV-1). Aquaculture Environment Interactions. 8, 543- 552. Gagnaire, B., Frouin, H., Moreau, K., Thomas-Guyon, H., Renault, T., 2006. Effects of temperature and salinity on haemocyte activities of the Pacific oyster, Crassostrea gigas (Thunberg). Fish & Shellfish Immunology. 20, 536-547. Garcia, C., Thébault, A., Dégremont, L., Arzul, I., Miossec, L., Robert, M., Chollet, B., François, C., Joly, J.-P., Ferrand, S., 2011. Ostreid herpesvirus 1 detection and relationship with Crassostrea gigas spat mortality in France between 1998 and 2006. Veterinary Research. 42, 73. Garnier, M., Labreuche, Y., Nicolas, J.L., 2008. Molecular and phenotypic characterization of Vibrio aestuarianus subsp. francensis subsp. nov., a pathogen of the oyster Crassostrea gigas. Systematic and Applied Microbiology. 31, 358-365. Garnier, M., Labreuche, Y., Garcia, C., Robert, M., Nicolas, J.-L., 2007. Evidence for the involvement of pathogenic bacteria in summer mortalities of the Pacific oyster Crassostrea gigas. Microbial Ecology. 53, 187-196. Gay, M., Renault, T., Pons, A.-M., Le Roux, F., 2004. Two Vibrio splendidus related strains collaborate to kill Crassostrea gigas: taxonomy and host alterations. Diseases of Aquatic Organisms. 62, 65-74. Glude, J., 1974. A summary report of Pacific coast oyster mortality investigations 1965–1972, Proceedings of the Third US–Japan Meeting on Aquaculture, Tokyo, Japan, pp. 28. Go, J., Deutscher, A.T., Spiers, Z.B., Dahle, K., Kirkland, P.D., Jenkins, C., 2017. Mass mortalities of unknown aetiology in Pacific oysters Crassostrea gigas in Port Stephens, New South Wales, Australia. Diseases of Aquatic Organisms. 125, 227-242. Gonzalez-Acosta, B., Bashan, Y., Hernandez-Saavedra, N.Y., Ascencio, F., De la Cruz-Agüero, G., 2006. Seasonal seawater temperature as the major determinant for populations of culturable bacteria in the sediments of an intact mangrove in an arid region. FEMS Microbiology Ecology. 55, 311-321. Gopalakrishnan, V., Spencer, C., Nezi, L., Reuben, A., Andrews, M., Karpinets, T., Prieto, P., Vicente, D., Hoffman, K., Wei, S., 2018. Gut microbiome modulates response to anti– PD-1 immunotherapy in melanoma patients. Science. 359, 97-103. Goulletquer, P., Soletchnik, P., Le Moine, O., Razet, D., Geairon, P., Faury, N., Taillade, S., 1998. Summer mortality of the Pacific cupped oyster Crassostrea gigas in the Bay of Marennes-Oléron (France), ICES Statutory Meeting, Lisbon, pp. CC(14):11-20. Gray, J.S., Wu, R.S.-s., Or, Y.Y., 2002. Effects of hypoxia and organic enrichment on the coastal marine environment. Marine Ecology Progress Series. 238, 249-279. Green, T., Barnes, A., 2010. Bacterial diversity of the digestive gland of Sydney rock oysters, Saccostrea glomerata infected with the paramyxean parasite, Marteilia sydneyi. Journal of Applied Microbiology. 109, 613-622. Green, T.J., Vergnes, A., Montagnani, C., de Lorgeril, J., 2016. Distinct immune responses of juvenile and adult oysters (Crassostrea gigas) to viral and bacterial infections. Veterinary Research. 47, 72.

233

Green, T.J., Montagnani, C., Benkendorff, K., Robinson, N., Speck, P., 2014. Ontogeny and water temperature influences the antiviral response of the Pacific oyster, Crassostrea gigas. Fish & Shellfish Immunology. 36, 151-157. Green, T.J., Siboni, N., King, W.L., Labbate, M., Seymour, J.R., Raftos, D., 2018. Simulated Marine Heat Wave Alters Abundance and Structure of Vibrio Populations Associated with the Pacific Oyster Resulting in a Mass Mortality Event. Microbial Ecology, 1-12. Guichard, B., Garcia, C., Joly, J.-P., Lupo, C., Chollet, B., Robert, M., Omnes, E., Arzul, I., 2011. REPAMO: a French network for the surveillance of molluscs health, Proceedings of the International Conference on Animal Health Surveillance (ICAHS), 17-20 May 2011, Lyon. Guo, F., Zhang, T., 2013. Biases during DNA extraction of activated sludge samples revealed by high throughput sequencing. Applied Microbiology and Biotechnology. 97, 4607- 4616. Hammer, Ø., Harper, D., Ryan, P., 2001. PAST: Paleontological statistics software package for education and data analysis. Palaeontologia Electronica. 4, 1-9. Handl, S., Dowd, S.E., Garcia-Mazcorro, J.F., Steiner, J.M., Suchodolski, J.S., 2011. Massive parallel 16S rRNA gene pyrosequencing reveals highly diverse fecal bacterial and fungal communities in healthy dogs and cats. FEMS Microbiology Ecology. 76, 301- 310. Harding, J.M., Couturier, C., Parsons, G.J., Ross, N.W., 2004. Evaluation of the neutral red retention assay as a stress response indicator in cultivated mussels (Mytilus spp.) in relation to seasonal and environmental conditions. Journal of Shellfish Research. 23, 745-752. Harris, J.M., 1993. The presence, nature, and role of gut microflora in aquatic invertebrates: a synthesis. Microbial Ecology. 25, 195-231. Harvell, C.D., Mitchell, C.E., Ward, J.R., Altizer, S., Dobson, A.P., Ostfeld, R.S., Samuel, M.D., 2002. Climate warming and disease risks for terrestrial and marine biota. Science. 296, 2158-2162. Hedge, L.H., Johnston, E.L., 2014. Colonisation of the non-indigenous Pacific oyster Crassostrea gigas determined by predation, size and initial settlement densities. PLoS One. 9, e90621. Henderson, G., Cox, F., Kittelmann, S., Miri, V.H., Zethof, M., Noel, S.J., Waghorn, G.C., Janssen, P.H., 2013. Effect of DNA extraction methods and sampling techniques on the apparent structure of cow and sheep rumen microbial communities. PLoS One. 8, e74787. Herbert, R., 1990. Methods for Enumerating Microorganisms and Determining Biomass in Natural Environments. in: Grigorova, R., Norris, J. (Eds.), Methods in Microbiology. Elsevier Science Publishing Co. Inc., San Diego, United States, pp. 1-39. Hernandez-Zarate, G., Olmos-Soto, J., 2006. Identification of bacterial diversity in the oyster Crassostrea gigas by fluorescent in situ hybridization and polymerase chain reaction. Journal of Applied Microbiology. 100, 664-672. Hick, P., Evans, O., Looi, R., English, C., Whittington, R.J., 2016. Stability of Ostreid herpesvirus-1 (OsHV-1) and assessment of disinfection of seawater and oyster tissues using a bioassay. Aquaculture. 450, 412-421. Hick, P.M., Evans, O., Rubio, A., Dhand, N.K., Whittington, R.J., 2018. Both age and size influence susceptibility of Pacific oysters (Crassostrea gigas) to disease caused by Ostreid herpesvirus-1 (OsHV-1) in replicated field and laboratory experiments. Aquaculture. 489, 110-120.

234

Hoffmann, A.R., Proctor, L., Surette, M., Suchodolski, J., 2016. The microbiome: the trillions of microorganisms that maintain health and cause disease in humans and companion animals. Veterinary Pathology. 53, 10-21. Horz, H.P., Scheer, S., Vianna, M.E., Conrads, G., 2010. New methods for selective isolation of bacterial DNA from human clinical specimens. Anaerobe. 16, 47-53. Illumina, 2020. Measuring sequencing accuracy. Experiments and protocol. February 22, 2020. https://sapac.illumina.com/science/technology/next-generation-sequencing/plan- experiments/quality-scores.html Imai, T., Numachi, K.-i., Oizumi, J., Sato, S., 1965. Studies on the Mass Mortality of the Oyster in Matsushima Bay II. Search for the Cause of Mass Mortality and the Possibility to Prevent it by Transplantation Experiment. Bulletin of the Tohoku National Fisheries Research Institute. 25, 27-38. Jenkins, C., Hick, P., Gabor, M., Spiers, Z., Fell, S.A., Gu, X., Read, A., Go, J., Dove, M., O'Connor, W., Kirkland, P.D., Frances, J., 2013. Identification and characterisation of an Ostreid herpesvirus-1 microvariant (OsHV-1 µ-var) in Crassostrea gigas (Pacific oysters) in Australia. Diseases of Aquatic Organisms. 105, 109-126. Jo, P.G., Choi, Y.K., Choi, C.Y., 2008. Cloning and mRNA expression of antioxidant enzymes in the Pacific oyster, Crassostrea gigas in response to cadmium exposure. Comparative Biochemistry and Physiology Part C: Toxicology & Pharmacology. 147, 460-469. Kaper, J.B., Mills, A.L., Colwell, R.R., 1978. Evaluation of the accuracy and precision of enumerating aerobic heterotrophs in water samples by the spread plate method. Applied and Environmental Microbiology. 35, 756-761. Kaplan, E.L., Meier, P., 1958. Nonparametric estimation from incomplete observations. Journal of the American Statistical Association. 53, 457-481. Kaufman, G., Blackstone, G., Vickery, M., Bej, A., Bowers, J., Bowen, M.D., Meyer, R.F., DePaola, A., 2004. Real-time PCR quantification of Vibrio parahaemolyticus in oysters using an alternative matrix. Journal of Food Protection. 67, 2424-2429. Keeling, S., Brosnahan, C., Williams, R., Gias, E., Hannah, M., Bueno, R., McDonald, W., Johnston, C., 2014. New Zealand juvenile oyster mortality associated with Ostreid herpesvirus-1 an opportunistic longitudinal study. Diseases of Aquatic Organisms. 109, 231-239. Kellogg, C.A., Lisle, J.T., Galkiewicz, J.P., 2009. Culture-independent characterization of bacterial communities associated with the cold-water coral Lophelia pertusa in the northeastern Gulf of Mexico. Applied Environmental Microbiology. 75, 2294-2303. Kim, D., Baik, K.S., Kim, M.S., Jung, B.-M., Shin, T.-S., Chung, G.-H., Rhee, M.S., Seong, C.N., 2007. Shewanella haliotis sp. nov., isolated from the gut microflora of abalone, Haliotis discus hannai. International Journal of Systematic and Evolutionary Microbiology. 57, 2926-2931. 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., 2012. Temperature regulation of virulence factors in the pathogen Vibrio coralliilyticus. The ISME Journal. 6, 835-846. King, G.M., Judd, C., Kuske, C.R., Smith, C., 2012. Analysis of stomach and gut microbiomes of the Eastern oyster (Crassostrea virginica) from coastal Louisiana, USA. PLoS One. 7, e51475. King, W.L., Jenkins, C., Seymour, J.R., Labbate, M., 2018b. Oyster disease in a changing environment: Decrypting the link between pathogen, microbiome and environment. Marine Environmental Research. 143, 124-140. King, W.L., Jenkins, C., Go, J., Siboni, N., Seymour, J.R., Labbate, M., 2018a. Characterisation of the Pacific oyster microbiome during a summer mortality event. Microbial Ecology. 77, 502-512.

235

King, W.L., Siboni, N., Williams, N.L.R., Kahlke, T., Nguyen, K.V., Jenkins, C., Dove, M., O'Connor, W., Seymour, J.R., Labbate, M., 2019. Variability in the composition of Pacific oyster microbiomes across oyster families exhibiting different levels of susceptibility to OsHV-1 μvar disease. Frontiers in Microbiology. 10, 473. Kirchhoff, H., Beyene, P., Fischer, M., Flossdorf, J., Heitmann, J., Khattab, B., Lopatta, D., Rosengarten, R., Seidel, G., Yousef, C., 1987. Mycoplasma mobile sp. nov., a new species from fish. International Journal of Systematic and Evolutionary Microbiology. 37, 192-197. Klappenbach, J.A., Saxman, P.R., Cole, J.R., Schmidt, T.M., 2001. rrndb: the ribosomal RNA operon copy number database. Nucleic Acids Research 29, 181-184. Koganezawa, A., 1975. Present status of studies on the mass mortality of cultured oyster in Japan and its prevention, Third US-Japan Meeting on Aquaculture, Fishery Agency. Tokyo, Japan, pp. 29-34. Kormas, K.A., 2011. Interpreting diversity of Proteobacteria based on 16S rRNA gene copy number. in: Sezenna, M. (Ed.), Proteobacteria: Phylogeny, Metabolic Diversity and Ecological Effects. Nova Publishers, NY, USA, pp. 73-89. Kostić, T., Weilharter, A., Rubino, S., Delogu, G., Uzzau, S., Rudi, K., Sessitsch, A., Bodrossy, L., 2007. A microbial diagnostic microarray technique for the sensitive detection and identification of pathogenic bacteria in a background of nonpathogens. Annals of Biochemistry. 360, 244-254. Krause, E., Wichels, A., Giménez, L., Lunau, M., Schilhabel, M.B., Gerdts, G., 2012. Small changes in pH have direct effects on marine bacterial community composition: a microcosm approach. PLoS One. 7, e47035. Krol, R.M., Hawkins, W.E., Overstreet, R.M., 1991. Rickettsial and mollicute infections in hepatopancreatic cells of cultured Pacific white shrimp (Penaeus vannamei). Journal of Invertebrate Pathology. 57, 362-370. Kueh, C.S., Chan, K.y., 1985. Bacteria in bivalve shellfish with special reference to the oyster. Journal of Applied Microbiology. 59, 41-47. Labreuche, Y., Soudant, P., Gonçalves, M., Lambert, C., Nicolas, J.-L., 2006. Effects of extracellular products from the pathogenic Vibrio aestuarianus strain 01/32 on lethality and cellular immune responses of the oyster Crassostrea gigas. Developmental and Comparative Immunology. 30, 367-379. Lacoste, A., Malham, S.K., Gélébart, F., Cueff, A., Poulet, S.A., 2002. Stress-induced immune changes in the oyster Crassostrea gigas. Developmental and Comparative Immunology. 26, 1-9. Lacoste, A., Jalabert, F., Malham, S., Cueff, A., Gelebart, F., Cordevant, C., Lange, M., Poulet, S., 2001. A Vibrio splendidus strain is associated with summer mortality of juvenile oysters Crassostrea gigas in the Bay of Morlaix (North Brittany, France). Diseases of Aquatic Organisms. 46, 139-145. Lane, D., 1991. 16S/23S rRNA sequencing. in: Stackebrandt, E., Goodfellow, M. (Eds.), Nucleic Acid Techniques in Bacterial Systematics. John Wiley and Sons, Chichester, UK, pp. 115-175. Lane, D.J., Pace, B., Olsen, G.J., Stahl, D.A., Sogin, M.L., Pace, N.R., 1985. Rapid determination of 16S ribosomal RNA sequences for phylogenetic analyses. Proceedings of National Academy of Sciences. U.S.A. 82, 6955-6959. Lasa, A., di Cesare, A., Tassistro, G., Borello, A., Gualdi, S., Furones, D., Carrasco, N., Cheslett, D., Brechon, A., Paillard, C., Bidault, A., Pernet, F., Canesi, L., Edomi, P., Pallavicini, A., Pruzzo, C., Vezzulli, L., 2019. Dynamics of the Pacific oyster pathobiota during mortality episodes in Europe assessed by 16S rRNA gene profiling

236

and a new target enrichment next-generation sequencing strategy. Environmental Microbiology. 21, 4548-4562. Lauber, C.L., Zhou, N., Gordon, J.I., Knight, R., Fierer, N., 2010. Effect of storage conditions on the assessment of bacterial community structure in soil and human-associated samples. FEMS Microbiology Letters. 307, 80-86. Le Deuff, R.-M., Renault, T., 1999. Purification and partial genome characterization of a herpes-like virus infecting the Japanese oyster, Crassostrea gigas. Journal of General Virology. 80, 1317-1322. Le Roux, F., Wegner, K.M., Polz, M.F., 2016. Oysters and vibrios as a model for disease dynamics in wild animals. Trends in Microbiology. 24, 568-580. Le Roux, F., Binesse, J., Saulnier, D., Mazel, D., 2007. Construction of a Vibrio splendidus mutant lacking the metalloprotease gene vsm by use of a novel counterselectable suicide vector. Applied Environmental Microbiology. 73, 777-784. Le Roux, F., Gay, M., Lambert, C., Waechter, M., Poubalanne, S., Chollet, B., Nicolas, J.-L., Berthe, F., 2002. Comparative analysis of Vibrio splendidus-related strains isolated during Crassostrea gigas mortality events. Aquatic Living Resources. 15, 251-258. Lee, R., Lovatelli, A., Ababouch, L., 2008. Bivalve depuration: fundamental and practical aspects:FAO Fisheries Technical Paper No. 511. Nations, Food and Agriculture Organization of the United Nations, Rome, Italy. 2008. 139. Lemire, A., Goudenege, D., Versigny, T., Petton, B., Calteau, A., Labreuche, Y., Le Roux, F., 2015. Populations, not clones, are the unit of Vibrio pathogenesis in naturally infected oysters. ISME Journal. 9, 1523-1531. Lepa, A., Siwicki, A.K., 2013. Fish herpesvirus diseases: a short review of current knowledge. Acta Veterinaria Brunensis. 81, 383-389. Ley, R.E., Lozupone, C.A., Hamady, M., Knight, R., Gordon, J.I., 2008. Worlds within worlds: evolution of the vertebrate gut microbiota. Nature Reviews Microbiology. 6, 776-788. Ley, R.E., Harris, J.K., Wilcox, J., Spear, J.R., Miller, S.R., Bebout, B.M., Maresca, J.A., Bryant, D.A., Sogin, M.L., Pace, N.R., 2006. Unexpected diversity and complexity of the Guerrero Negro hypersaline microbial mat. Applied Environmental Microbiology. 72, 3685-3695. Lightner, D.V., 2011. Virus diseases of farmed shrimp in the Western hemisphere (the Americas): a review. Journal of Invertebrate Pathology. 106, 110-130. Lipp, P.R., Brown, B., Liston, J., Chew, K., 1976. Recent Findings on the Summer Diseases of Pacific Oysters. Proceedings of National Shellfish Association. 65, 9-10. Littman, D.R., Pamer, E.G., 2011. Role of the commensal microbiota in normal and pathogenic host immune responses. Cell Host Microbe. 10, 311-323. Liu, K.-L., Wong, T.-T., 2013. Naïve Bayesian classifiers with multinomial models for rRNA taxonomic assignment. IEEE/ACM Transactions on Computational Biology and Bioinformatics. 10, 1334-1339. Lokmer, A., Wegner, K.M., 2015. Hemolymph microbiome of Pacific oysters in response to temperature, temperature stress and infection. The ISME Journal. 9, 670-682. Lokmer, A., Kuenzel, S., Baines, J.F., Wegner, K.M., 2016a. The role of tissue-specific microbiota in initial establishment success of Pacific oysters. Environmental Microbiology. 18, 970-987. Lokmer, A., Goedknegt, M.A., Thieltges, D.W., Fiorentino, D., Kuenzel, S., Baines, J.F., Wegner, K.M., 2016b. Spatial and temporal dynamics of Pacific oyster hemolymph microbiota across multiple scales. Frontiers in Microbiology. 7, 1367. Lynch, S., Carlsson, J., Reilly, A., Cotter, E., Culloty, S., 2012. A previously undescribed Ostreid herpesvirus 1 (OsHV-1) genotype detected in the Pacific oyster, Crassostrea gigas, in Ireland. Parasitology. 139, 1526-1532.

237

Maldonado, M., Sánchez-Tocino, L., Navarro, C., 2010. Recurrent disease outbreaks in corneous demosponges of the genus Ircinia: epidemic incidence and defense mechanisms. Marine Biology. 157, 1577-1590. Malham, S.K., Cotter, E., O'Keeffe, S., Lynch, S., Culloty, S.C., King, J.W., Latchford, J.W., Beaumont, A.R., 2009. Summer mortality of the Pacific oyster, Crassostrea gigas, in the Irish Sea: the influence of temperature and nutrients on health and survival. Aquaculture. 287, 128-138. Mancuso, F.P., D'Hondt, S., Willems, A., Airoldi, L., De Clerck, O., 2016. Diversity and temporal dynamics of the epiphytic bacterial communities associated with the canopy- forming seaweed Cystoseira compressa (Esper) Gerloff and Nizamuddin. Frontiers in Microbiology. 7, 476. Mardis, E.R., 2008. The impact of next-generation sequencing technology on genetics. Trends in Genetics. 24, 133-141. Martenot, C., Oden, E., Travaille, E., Malas, J.P., Houssin, M., 2010. Comparison of two real- time PCR methods for detection of Ostreid herpesvirus 1 in the Pacific oyster Crassostrea gigas. Journal of Virological Methods. 170, 86-89. Martenot, C., Oden, E., Travaillé, E., Malas, J.-P., Houssin, M., 2011. Detection of different variants of Ostreid herpesvirus 1 in the Pacific oyster, Crassostrea gigas between 2008 and 2010. Virus Research. 160, 25-31. Martenot, C., Denechère, L., Hubert, P., Metayer, L., Oden, E., Trancart, S., Travaillé, E., Houssin, M., 2015. Virulence of Ostreid herpesvirus 1 μVar in sea water at 16 C and 25 C. Aquaculture. 439, 1-6. Matozzo, V., Marin, M., 2011. Bivalve immune responses and climate changes: is there a relationship? Invertebrate Survival Journal. 8, 70-77. Maurer, D., Comps, M., His, E., 1986. Characteristics of the of the summer mortality of the Crassostrea gigas oyster in the Arcachon basin. Haliotis. 15, 309-317. McDonald, D., Price, M.N., Goodrich, J., Nawrocki, E.P., DeSantis, T.Z., Probst, A., Andersen, G.L., Knight, R., Hugenholtz, P., 2012a. An improved Greengenes taxonomy with explicit ranks for ecological and evolutionary analyses of bacteria and archaea. The ISME journal. 6, 610-618. McDonald, D., Clemente, J.C., Kuczynski, J., Rideout, J.R., Stombaugh, J., Wendel, D., Wilke, A., Huse, S., Hufnagle, J., Meyer, F., 2012b. The Biological Observation Matrix (BIOM) format or: how I learned to stop worrying and love the ome-ome. GigaScience. 1, 7. McHenery, J., Birkbeck, T., 1986. Inhibition of filtration in Mytilus edulis L. by marine vibrios. Journal of Fish Diseases. 9, 257-261. Medeiros, I.D., Siebert, M.N., Toledo-Silva, G., Rodrigues, T.B., Marques, M.R., Bainy, A.C., 2008. Induced gene expression in oyster Crassostrea gigas exposed to sewage. Environmental Toxicology and Pharmacology. 26, 362-365. Meisterhans, G., Raymond, N., Girault, E., Lambert, C., Bourrasseau, L., de Montaudouin, X., Garabetian, F., Jude-Lemeilleur, F., 2016. Structure of Manila Clam (Ruditapes philippinarum) microbiota at the organ scale in contrasting sets of individuals. Microbial Ecology. 71, 194-206. Meyer, J.L., Gunasekera, S.P., Scott, R.M., Paul, V.J., Teplitski, M., 2016. Microbiome shifts and the inhibition of quorum sensing by Black Band Disease Cyanobacteria. The ISME Journal. 10, 1204-1216. Mobsby, D., Koudah, A., 2017. Australian fisheries and aquaculture statistics 2016. Fisheries Research and Development Corporation project 2017-095, ABARES, Department of Agriculture and Water Resources, Canberra, Australia. https://data.gov.au/dataset/ds- dga-c12a0597-b535-48b3-a602-1f8382d8dae6/details?q=

238

Mori, K., 1979. Effects of artificial eutrophication on the metabolism of the Japanese oyster Crassostrea gigas. Marine Biology. 53, 361-369. Morita, H., Kuwahara, T., Ohshima, K., Sasamoto, H., Itoh, K., Hattori, M., Hayashi, T., Takami, H., 2007. An improved DNA isolation method for metagenomic analysis of the microbial flora of the human intestine. Microbes and Environments. 22, 214-222. Mortensen, S., Strand, Å., Bodvin, T., Alfjorden, A., Skår, C.K., Jelmert, A., Aspán, A., Sælemyr, L., Naustvoll, L.-J., Albretsen, J., 2016. Summer mortalities and detection of Ostreid herpesvirus microvariant in Pacific oyster Crassostrea gigas in Sweden and Norway. Diseases of Aquatic Organisms. 117, 171-176. Munson, M., Banerjee, A., Watson, T., Wade, W., 2004. Molecular analysis of the microflora associated with dental caries. Journal of Clinical Microbiology. 42, 3023-3029. Murray, A.G., Peeler, E.J., 2005. A framework for understanding the potential for emerging diseases in aquaculture. Preventive Veterinary Medicine. 67, 223-235. Nadkarni, M.A., Martin, F.E., Jacques, N.A., Hunter, N., 2002. Determination of bacterial load by real-time PCR using a broad-range (universal) probe and primers set. Microbiology. 148, 257-266. Nell, J.A., Holliday, J.E., 1988. Effects of salinity on the growth and survival of Sydney rock oyster (Saccostrea commercialis) and Pacific oyster (Crassostrea gigas) larvae and spat. Aquaculture. 68, 39-44. Niu, Q., Li, P., Hao, S., Zhang, Y., Kim, S.W., Li, H., Ma, X., Gao, S., He, L., Wu, W., 2015. Dynamic distribution of the gut microbiota and the relationship with apparent crude fiber digestibility and growth stages in pigs. Scientific Reports. 5, 9938. Oden, E., Martenot, C., Berthaux, M., Travaillé, E., Malas, J.P., Houssin, M., 2011. Quantification of Ostreid herpesvirus-1 (OsHV-1) in Crassostrea gigas by real-time PCR: Determination of a viral load threshold to prevent summer mortalities. Aquaculture. 317, 27-31. OIE, 2014. Infection with Ostreid herpesvirus-1 microvariants (May 2014 version). in: (OIE), World Organization for Animal Health. (Ed.), Manual of Diagnostic Tests for Aquatic Animals 2015. World Assembly of Delegates of the OIE, Paris, France. Olafsen, J.A., Mikkelsen, H.V., Giæver, H.M., Hansen, G.H., 1993. Indigenous bacteria in hemolymph and tissues of marine bivalves at low temperatures. Applied and Environmental Microbiology. 59, 1848-1854. Ossai, S., Ramachandran, P., Ottesen, A., Reed, E., DePaola, A., Parveen, S., 2017. Microbiomes of American Oysters (Crassostrea virginica) harvested from two sites in the Chesapeake Bay. Genome Announcements. 5, e00729-00717. Paillard, C., Le Roux, F., Borrego, J.J., 2004. Bacterial disease in marine bivalves, a review of recent studies: trends and evolution. Aquatic Living Resources. 17, 477-498. Paterson, K., Schreider, M., Zimmerman, K., 2003. Effect of catchment disturbance on seston quality and quantity in Brisbane Water, NSW. Estuarine, Coastal and Shelf Science. 56, 355-368. Pathirana, E., McPherson, A., Whittington, R., Hick, P., 2019a. The role of tissue-type, sampling and nucleic acid purification methodology on the inferred composition of Pacific oyster (Crassostrea gigas) microbiome. Journal of Applied Microbiology. 127, 429-444. Pathirana, E., Fuhrmann, M., Whittington, R., Hick, P., 2019b. Influence of environment on the pathogenesis of Ostreid herpesvirus-1 (OsHV-1) infections in Pacific oysters (Crassostrea gigas) through differential microbiome responses. Heliyon. 5, e02101. Paul-Pont, I., Dhand, N.K., Whittington, R.J., 2013a. Spatial distribution of mortality in Pacific oysters Crassostrea gigas: reflection on mechanisms of OsHV-1 transmission. Diseases of Aquatic Organisms. 105, 127-138.

239

Paul-Pont, I., Dhand, N.K., Whittington, R.J., 2013b. Influence of husbandry practices on OsHV-1 associated mortality of Pacific oysters Crassostrea gigas. Aquaculture. 412- 413, 202-214. Paul-Pont, I., Evans, O., Dhand, N.K., Rubio, A., Coad, P., Whittington, R., 2014. Descriptive epidemiology of mass mortality due to Ostreid herpesvirus-1 (OsHV-1) in commercially farmed Pacific oysters (Crassostrea gigas) in the Hawkesbury River estuary, Australia. Aquaculture. 422-423, 146-159. Paul-Pont, I., Evans, O., Dhand, N.K., Whittington, R.J., 2015. Experimental infections of Pacific oyster Crassostrea gigas using the Australian Ostreid herpesvirus-1 (OsHV-1) μVar strain. Diseases of Aquatic Organisms. 113, 137-147. Pavlova, S., Kilic, A., Kilic, S., So, J.S., Nader‐Macias, M., Simoes, J., Tao, L., 2002. Genetic diversity of vaginal lactobacilli from women in different countries based on 16S rRNA gene sequences. Journal of Applied Microbiology. 92, 451-459. Peeler, E.J., Reese, R.A., Cheslett, D.L., Geoghegan, F., Power, A., Thrush, M.A., 2012. Investigation of mortality in Pacific oysters associated with Ostreid herpesvirus-1 μVar in the Republic of Ireland in 2009. Preventive Veterinary Medicine. 105, 136-143. Pepin, J.-F., Riou, A., Renault, T., 2008. Rapid and sensitive detection of Ostreid herpesvirus 1 in oyster samples by real-time PCR. Journal of Virological Methods. 149, 269-276. Pereira, J.C., Chaves, R., Bastos, E., Leitao, A., Guedes-Pinto, H., 2011. An efficient method for genomic DNA extraction from different molluscs species. International Journal of Molecular Sciences. 12, 8086-8095. Pernet, F., Lupo, C., Bacher, C., Whittington, R.J., 2016. Infectious diseases in oyster aquaculture require a new integrated approach. Philosophical Transactions of the Royal Society of London B Biological Sciences. 371, 20150213. Pernet, F., Gachelin, S., Stanisière, J.-Y., Petton, B., Fleury, E., Mazurié, J., 2019. Farmer monitoring reveals the effect of tidal height on mortality risk of oysters during a herpesvirus outbreak. ICES Journal of Marine Science. 76, 1816-1824. Pernet, F., Barret, J., Le Gall, P., Corporeau, C., Dégremont, L., Lagarde, F., Pépin, J.-F., Keck, N., 2012. Mass mortalities of Pacific oysters Crassostrea gigas reflect infectious diseases and vary with farming practices in the Mediterranean Thau lagoon, France. Aquaculture Environment Interactions. 2, 215-237. Pernet, F., Lagarde, F., Jeannée, N., Daigle, G., Barret, J., Le Gall, P., Quere, C., D’orbcastel, E.R., 2014. Spatial and temporal dynamics of mass mortalities in oysters is influenced by energetic reserves and food quality. PLoS One. 9, e88469. Pernet, F., Fuhrmann, M., Petton, B., Mazurié, J., Bouget, J.-F., Fleury, E., Daigle, G., Gernez, P., 2018. Determination of risk factors for herpesvirus outbreak in oysters using a broad-scale spatial epidemiology framework. Scientific Reports. 8, 10869. Petton, B., Pernet, F., Robert, R., Boudry, P., 2013. Temperature influence on pathogen transmission and subsequent mortalities in juvenile Pacific oysters Crassostrea gigas. Aquaculture Environment Interactions. 3, 257-273. Petton, B., Boudry, P., Alunno-Bruscia, M., Pernet, F., 2015a. Factors influencing disease- induced mortality of Pacific oysters Crassostrea gigas. Aquaculture Environment Interactions. 6, 205-222. Petton, B., Bruto, M., James, A., Labreuche, Y., Alunno-Bruscia, M., Le Roux, F., 2015b. Crassostrea gigas mortality in France: the usual suspect, a herpes virus, may not be the killer in this polymicrobial opportunistic disease. Frontiers in Microbiology. 6, 686. Petton, B., de Lorgeril, J., Mitta, G., Daigle, G., Pernet, F., Alunno-Bruscia, M., 2019. Fine- scale temporal dynamics of herpes virus and vibrios in seawater during a polymicrobial infection in the Pacific oyster Crassostrea gigas. Diseases of Aquatic Organisms. 135, 97-106.

240

Pierce, M.L., Ward, J.E., 2018. Microbial Ecology of the Bivalvia, with an emphasis on the Family Ostreidae. Journal of Shellfish Research. 37, 793-806. Pierce, M.L., Ward, J.E., 2019. Gut microbiomes of the Eastern Oyster (Crassostrea virginica) and the Blue Mussel (Mytilus edulis): Temporal variation and the influence of marine aggregate-associated microbial communities. mSphere. 4, e00730-00719. Pierce, M.L., Ward, J.E., Holohan, B.A., Zhao, X., Hicks, R.E., 2016. The influence of site and season on the gut and pallial fluid microbial communities of the Eastern oyster, Crassostrea virginica (Bivalvia, Ostreidae): community-level physiological profiling and genetic structure. Hydrobiologia. 765, 97-113. Pita, L., Erwin, P.M., Turon, X., Lopez-Legentil, S., 2013. Till death do us part: stable sponge- bacteria associations under thermal and food shortage stresses. PLoS One. 8, e80307. Pollock, J., Glendinning, L., Wisedchanwet, T., Watson, M., 2018. The madness of microbiome: attempting to find consensus “best practice” for 16S microbiome studies. Applied Environmental Microbiology. 84, e02627-02617. Prakash, S., Rodes, L., Coussa-Charley, M., Tomaro-Duchesneau, C., 2011. Gut microbiota: next frontier in understanding human health and development of biotherapeutics. Biologics: Targets and Therapy. 5, 71–86. Prescott, S.L., 2017. History of medicine: Origin of the term microbiome and why it matters. Human Microbiome Journal. 4, 24-25. Price, M.N., Dehal, P.S., Arkin, A.P., 2009. FastTree: computing large minimum evolution trees with profiles instead of a distance matrix. Molecular Biology and Evolution. 26, 1641-1650. Prieur, D., Nicolas, J., Plusquellec, A., Vigneulle, M., 1990. Interactions between bivalve mollusks and bacteria in the marine-environment. Oceanography and Marine Biology. 28, 277-352. Ren, W., Chen, H., Renault, T., Cai, Y., Bai, C., Wang, C., Huang, J., 2013. Complete genome sequence of acute Viral necrosis virus associated with massive mortality outbreaks in the Chinese scallop, Chlamys farreri. Virology Journal. 10, 110. Renault, T., 2011. Viruses infecting marine molluscs. Studies in Viral Ecology. 2, 153-176. Renault, T., Cochennec, N., Le Deuff, R.-M., Chollet, B., 1994a. Herpes-like virus infecting Japanese oyster (Crassostrea gigas) spat. Bulletin of the European Association of Fish Pathologists. 14, 64-66. Renault, T., Le Deuff, R.-M., Cochennec, N., Maffart, P., 1994b. Herpesviruses associated with mortalities among Pacific oyster, Crassostrea gigas, in France-Comparative study. Revue de Médecine Vétérinaire. 145, 735-742. Renault, T., Le Deuff, R., Cochennec, N., Chollet, B., Maffart, P., 1995. Herpes-like viruses associated with high mortality levels in larvae and spat of Pacific oysters, Crassostrea gigas: a comparative study, the thermal effects on virus detection in hatchery-reared larvae, reproduction of the disease in axenic larvae. Veterinary Research. 26, 539-543. Renault, T., Bouquet, A.L., Maurice, J.-T., Lupo, C., Blachier, P., 2014. Ostreid herpesvirus 1 infection among Pacific oyster (Crassostrea gigas) spat: relevance of water temperature to virus replication and circulation prior to the onset of mortality. Applied Environmental Microbiology. 80, 5419-5426. Renault, T., Moreau, P., Faury, N., Pepin, J.-F., Segarra, A., Webb, S., 2012. Analysis of clinical Ostreid herpesvirus 1 (Malacoherpesviridae) specimens by sequencing amplified fragments from three virus genome areas. 86, 5942-5947. Rideout, J.R., Chase, J.H., Bolyen, E., Ackermann, G., Journal of Virology. González, A., Knight, R., Caporaso, J.G., 2016. Keemei: cloud-based validation of tabular bioinformatics file formats in Google Sheets. GigaScience. 5, 27.

241

Romero, J., Espejo, R.T., 2001. The prevalence of noncultivable bacteria in oysters (Tiostrea chilenensis Philippi, 1845). Journal Shellfish Research. 20, 1235–1240. Romero, J., Garcia-Varela, M., Laclette, J., Espejo, R., 2002. Bacterial 16S rRNA gene analysis revealed that bacteria related to Arcobacter spp. constitute an abundant and common component of the oyster microbiota (Tiostrea chilensis). Microbial Ecology. 44, 365- 371. Roque, A., Carrasco, N., Andree, K.B., Lacuesta, B., Elandaloussi, L., Gairin, I., Rodgers, C.J., Furones, M.D., 2012. First report of OsHV-1 microvar in Pacific oyster (Crassostrea gigas) cultured in Spain. Aquaculture. 324-325, 303-306. Roterman, Y.R., Benayahu, Y., Reshef, L., Gophna, U., 2015. The gill microbiota of invasive and indigenous Spondylus oysters from the Mediterranean Sea and northern Red Sea. Environmental Microbiology Reports. 7, 860-867. Samain, J.F., McCombie, H., 2008. Summer mortality of Pacific oyster Crassostrea gigas: The Morest Project. Edition Quae, INRA, Paris, France. Samain, J.F., Dégremont, L., Soletchnik, P., Haure, J., Bédier, E., Ropert, M., Moal, J., Huvet, A., Bacca, H., Van Wormhoudt, A., Delaporte, M., Costil, K., Pouvreau, S., Lambert, C., Boulo, V., Soudant, P., Nicolas, J.L., Le Roux, F., Renault, T., Gagnaire, B., Geret, F., Boutet, I., Burgeot, T., Boudry, P., 2007. Genetically based resistance to summer mortality in the Pacific oyster (Crassostrea gigas) and its relationship with physiological, immunological characteristics and infection processes. Aquaculture. 268, 227-243. Saulnier, D., De Decker, S., Haffner, P., 2009. Real-time PCR assay for rapid detection and quantification of Vibrio aestuarianus in oyster and seawater: a useful tool for epidemiologic studies. Journal of Microbiological Methods. 77, 191-197. Saulnier, D., De Decker, S., Haffner, P., Cobret, L., Robert, M., Garcia, C., 2010. A large-scale epidemiological study to identify bacteria pathogenic to Pacific oyster Crassostrea gigas and correlation between virulence and metalloprotease-like activity. Microbial Ecology. 59, 787-798. Savin, K.W., Cocks, B.G., Wong, F., Sawbridge, T., Cogan, N., Savage, D., Warner, S., 2010. A neurotropic herpesvirus infecting the gastropod, abalone, shares ancestry with oyster herpesvirus and a herpesvirus associated with the amphioxus genome. Virology Journal. 7, 308. Schikorski, D., Faury, N., Pepin, J.F., Saulnier, D., Tourbiez, D., Renault, T., 2011. Experimental Ostreid herpesvirus 1 infection of the Pacific oyster Crassostrea gigas: kinetics of virus DNA detection by q-PCR in seawater and in oyster samples. Virus Research. 155, 28-34. Schoenfeld, D., 1982. Partial residuals for the proportional hazards regression model. Biometrika. 69, 239-241. Segarra, A., Pepin, J.F., Arzul, I., Morga, B., Faury, N., Renault, T., 2010. Detection and description of a particular Ostreid herpesvirus 1 genotype associated with massive mortality outbreaks of Pacific oysters, Crassostrea gigas, in France in 2008. Virus Research. 153, 92-99. Sekirov, I., Tam, N.M., Jogova, M., Robertson, M.L., Li, Y., Lupp, C., Finlay, B.B., 2008. Antibiotic-induced perturbations of the intestinal microbiota alter host susceptibility to enteric infection. Infection and Immunity. 76, 4726-4736. Seutin, G., White, B.N., Boag, P.T., 1991. Preservation of avian blood and tissue samples for DNA analyses. Canadian Journal of Zoology. 69, 82-90. Shade, A., Caporaso, J.G., Handelsman, J., Knight, R., Fierer, N., 2013. A meta-analysis of changes in bacterial and archaeal communities with time. The ISME Journal. 7, 1493- 1506.

242

Shade, A., Peter, H., Allison, S.D., Baho, D.L., Berga, M., Burgmann, H., Huber, D.H., Langenheder, S., Lennon, J.T., Martiny, J.B., Matulich, K.L., Schmidt, T.M., Handelsman, J., 2012. Fundamentals of microbial community resistance and resilience. Frontiers in Microbiology. 3, 417. Shimahara, Y., Kurita, J., Kiryu, I., Nishioka, T., Yuasa, K., Kawana, M., Kamaishi, T., Oseko, N., 2012. Surveillance of type 1 Ostreid herpesvirus (OsHV-1) variants in Japan. Fish Pathology. 47, 129-136. Simister, R., Taylor, M.W., Tsai, P., Fan, L., Bruxner, T.J., Crowe, M.L., Webster, N., 2012. Thermal stress responses in the bacterial biosphere of the Great barrier reef sponge, Rhopaloeides odorabile. Environmental Microbiology. 14, 3232-3246. Singh, P., Manning, S.D., 2016. Impact of age and sex on the composition and abundance of the intestinal microbiota in individuals with and without enteric infections. Annals of Epidemiology. 26, 380-385. Soletchnik, P., Razet, D., Geairon, P., 1997. Ecophysiology of maturation and spawning in oyster (Crassostrea gigas): metabolic (respiration) and feeding (clearance and absorption rates) responses at different maturation stages. Oceanographic Literature Review. 12, 1545. Soletchnik, P., Lambert, C., Costil, K., 2005. Summer mortality of Crassostrea gigas (Thunberg) in relation to environmental rearing conditions. Journal of Shellfish Research. 24, 197-207. Soletchnik, P., Ropert, M., Mazurié, J., Fleury, P.G., Le Coz, F., 2007. Relationships between oyster mortality patterns and environmental data from monitoring databases along the coasts of France. Aquaculture. 271, 384-400. Soliman, M.F., El-Shenawy, N.S., Tadros, M.M., Abd El-Azeez, A.A., 2015. Impaired behavior and changes in some biochemical markers of bivalve (Ruditapes decussatus) due to zinc toxicity. Toxicological and Environmental Chemistry. 97, 674-686. Solomieu, V.B., Renault, T., Travers, M.-A., 2015. Mass mortality in bivalves and the intricate case of the Pacific oyster, Crassostrea gigas. Journal of Invertebrate Pathology. 131, 2-10. Streit, W.R., Schmitz, R.A., 2004. Metagenomics–the key to the uncultured microbes. Current Opinions in Microbiology. 7, 492-498. Sugita, H., Tanaami, H., Kobashi, T., Deguchi, Y., 1981. Bacterial flora of coastal bivalves. Bulletin of the Japanese Society of Scientific Fisheries. 47, 655-661. Sugumar, G., Nakai, T., Hirata, Y., Matsubara, D., Muroga, K., 1998. Vibrio splendidus biovar II as the causative agent of bacillary necrosis of Japanese oyster Crassostrea gigas larvae. Diseases of Aquatic Organisms. 33, 111-118. Sweet, M.J., Croquer, A., Bythell, J.C., 2010. Bacterial assemblages differ between compartments within the coral holobiont. Coral Reefs. 30, 39-52. Takeuchi, T., Takemoto, Y., Matsubara, T., 1960. Haematological study of bacteria affected oysters. Hiroshima-ken Suisan Shikenjo Hokoku (Hiroshima Prefectural Fisheries Experimental Station). 22, 1-7. Tamayo, D., Corporeau, C., Petton, B., Quere, C., Pernet, F., 2014. Physiological changes in Pacific oyster (Crassostrea gigas) exposed to the herpesvirus OsHV-1 μVar. Aquaculture. 432, 304-310. Tanaka, R., Ootsubo, M., Sawabe, T., Ezura, Y., Tajima, K., 2004. Biodiversity and in situ abundance of gut microflora of abalone (Haliotis discus hannai) determined by culture- independent techniques. Aquaculture. 241, 453-463. Tegelström, H., 1989. Cold room storage of blood may be better than storage at -70o C. Fingerprint News. 4, 5-6.

243

Thompson, E.L., Taylor, D.A., Nair, S.V., Birch, G., Coleman, R., Raftos, D.A., 2012. Optimal acclimation periods for oysters in laboratory-based experiments. Journal of Molluscan Studies. 78, 304-307. Thompson, F.L., Iida, T., Swings, J., 2004a. Biodiversity of vibrios. Microbiology and Molecular Biology Reviews. 68, 403-431. Thompson, J.R., Randa, M.A., Marcelino, L.A., Tomita-Mitchell, A., Lim, E., Polz, M.F., 2004b. Diversity and dynamics of a north atlantic coastal Vibrio community. Applied Environmental Microbiology. 70, 4103-4110. Thurber, R.V., Willner‐Hall, D., Rodriguez‐Mueller, B., Desnues, C., Edwards, R.A., Angly, F., Dinsdale, E., Kelly, L., Rohwer, F., 2009. Metagenomic analysis of stressed coral holobionts. Applied Environmental Microbiology. 11, 2148-2163. Trabal Fernández, N., Mazón-Suástegui, J.M., Vázquez-Juárez, R., Ascencio-Valle, F., Romero, J., 2014. Changes in the composition and diversity of the bacterial microbiota associated with oysters (Crassostrea corteziensis, Crassostrea gigas and Crassostrea sikamea) during commercial production. FEMS Microbiology Ecology 88, 69-83. Trabal, N., Mazon-Suastegui, J.M., Vazquez-Juarez, R., Asencio-Valle, F., Morales-Bojorquez, E., Romero, J., 2012. Molecular analysis of bacterial microbiota associated with oysters (Crassostrea gigas and Crassostrea corteziensis) in different growth phases at two cultivation sites. Microbial Ecology. 64, 555-569. Travers, M.-A., Miller, K.B., Roque, A., Friedman, C.S., 2015. Bacterial diseases in marine bivalves. Journal of Invertebrate Pathology. 131, 11-31. Travers, M.A., Basuyaux, O., Le Goïc, N., Huchette, S., Nicolas, J.L., Koken, M., Paillard, C., 2009. Influence of temperature and spawning effort on Haliotis tuberculata mortalities caused by Vibrio harveyi: an example of emerging vibriosis linked to global warming. Global Change Biology. 15, 1365-1376. Tremblay, J., Singh, K., Fern, A., Kirton, E.S., He, S., Woyke, T., Lee, J., Chen, F., Dangl, J.L., Tringe, S.G., 2015. Primer and platform effects on 16S rRNA tag sequencing. Frontiers in Microbiology. 6, 771. Van de Peer, Y., Chapelle, S., De Wachter, R., 1996. A quantitative map of nucleotide substitution rates in bacterial rRNA. Nucleic Acids Research. 24, 3381-3391. Vandamme, P., De Ley, J., 1991. Proposal for a new family, Campylobacteraceae. International Journal of Systematic and Evolutionary Microbiology. 41, 451-455. Vanhoutte, T., Huys, G., De Brandt, E., Swings, J., 2004. Temporal stability analysis of the microbiota in human feces by denaturing gradient gel electrophoresis using universal and group-specific 16S rRNA gene primers. FEMS Microbiology Ecology. 48, 437- 446. Vázquez-Baeza, Y., Pirrung, M., Gonzalez, A., Knight, R., 2013. EMPeror: a tool for visualizing high-throughput microbial community data. Gigascience. 2, 16. Vezzulli, L., Pezzati, E., Stauder, M., Stagnaro, L., Venier, P., Pruzzo, C., 2015. Aquatic ecology of the oyster pathogens Vibrio splendidus and Vibrio aestuarianus. Environmental Microbiology. 17, 1065-1080. Vezzulli, L., Stagnaro, L., Grande, C., Tassistro, G., Canesi, L., Pruzzo, C., 2018. Comparative 16SrDNA gene-based microbiota profiles of the Pacific oyster (Crassostrea gigas) and the Mediterranean mussel (Mytilus galloprovincialis) from a shellfish farm (Ligurian Sea, Italy). Microbial Ecology. 75, 495-504. Vezzulli, L., Previati, M., Pruzzo, C., Marchese, A., Bourne, D.G., Cerrano, C., Consortium, V., 2010. Vibrio infections triggering mass mortality events in a warming Mediterranean Sea. Environmental microbiology. 12, 2007-2019.

244

Vezzulli, L., Brettar, I., Pezzati, E., Reid, P.C., Colwell, R.R., Hofle, M.G., Pruzzo, C., 2012. Long-term effects of ocean warming on the prokaryotic community: evidence from the vibrios. ISME Journal. 6, 21-30. Wagner-Mackenzie, B., Waite, D.W., Taylor, M.W., 2015. Evaluating variation in human gut microbiota profiles due to DNA extraction method and inter-subject differences. Frontiers in Microbiology. 6, 130. Wang, D., Zhang, Q., Cui, Y., Shi, X., 2014. Seasonal dynamics and diversity of bacteria in retail oyster tissues. International Journal of Food Microbiology 173, 14-20. Wegner, K.M., Volkenborn, N., Peter, H., Eiler, A., 2013. Disturbance induced decoupling between host genetics and composition of the associated microbiome. BMC Microbiology. 13, 252. Weisburg, W.G., Barns, S.M., Pelletier, D.A., Lane, D.J., 1991. 16S ribosomal DNA amplification for phylogenetic study. Journal of Bacteriology. 173, 697-703. Wen, Y., Xiao, F., Wang, C., Wang, Z., 2016. The impact of different methods of DNA extraction on microbial community measures of BALF samples based on metagenomic data. American Journal of Translational Research. 8, 1412-1425. Whipps, J., Cooke, R., 1988. Mycoparasitism and plant disease control in: Burge, N.M. (Ed.), Fungi in biological control systems. Manchester University Press, Manchester, United Kingdom, pp. 161-187. Whittington, R.J., Dhand, N.K., Evans, O., Paul-Pont, I., 2015b. Further observations on the influence of husbandry practices on OsHV-1 μVar mortality in Pacific oysters Crassostrea gigas: Age, cultivation structures and growing height. Aquaculture. 438, 82-97. Whittington, R.J., Paul-Pont, I., Evans, O., Hick, P., Dhand, N.K., 2018. Counting the dead to determine the source and transmission of the marine herpesvirus OsHV-1 in Crassostrea gigas. Veterinary Research. 49, 34. Whittington, R.J., Liu, O., Hick, P.M., Dhand, N., Rubio, A., 2019. Long-term temporal and spatial patterns of Ostreid herpesvirus 1 (OsHV-1) infection and mortality in sentinel Pacific oyster spat (Crassostrea gigas) inform farm management. Aquaculture, 734395. Whittington, R.J., Hick, P.M., Evans, O., Rubio, A., Alford, B., Dhand, N., Paul-Pont, I., 2015a. Protection of Pacific oyster (Crassostrea gigas) spat from mortality due to Ostreid herpesvirus 1 (OsHV-1 μVar) using simple treatments of incoming seawater in land- based upwellers. Aquaculture. 437, 10-20. Wilson, K., 2001. Preparation of genomic DNA from bacteria, Current Protocols in Molecular Biology. 00:I:2.4:2.4.1–2.4.5. Winnepenninckx, B., Backeljau, T., Dewachter, R., 1993. Extraction of high molecular weight DNA from molluscs. Trends in Genetics. 9, 407. Wu, R.S., 2002. Hypoxia: from molecular responses to ecosystem responses. Marine Pollution Bulletin. 45, 35-45. Xia, J., Bai, C., Wang, C., Song, X., Huang, J., 2015. Complete genome sequence of Ostreid herpesvirus-1 associated with mortalities of Scapharca broughtonii broodstocks. Virology Journal. 12, 110. Yang, B., Wang, Y., Qian, P.-Y., 2016. Sensitivity and correlation of hypervariable regions in 16S rRNA genes in phylogenetic analysis. BMC Bioinformatics. 17, 135. Yap, K., Thompson, R., 1987. CTAB precipitation of cestode DNA. Parasitology Today. 3, 220-222. Yatsunenko, T., Rey, F.E., Manary, M.J., Trehan, I., Dominguez-Bello, M.G., Contreras, M., Magris, M., Hidalgo, G., Baldassano, R.N., Anokhin, A.P., 2012. Human gut microbiome viewed across age and geography. Nature. 486, 222.

245

Yu, R., 1998. Discussion on the high mortality and its prevention in Scallop, Chlamys farreri. Transactions of Oceanology and Limnology. 71, 69-72. Zemach, A., McDaniel, I.E., Silva, P., Zilberman, D., 2010. Genome-wide evolutionary analysis of eukaryotic DNA methylation. Science. 328, 916-919. Zhang, J., Kobert, K., Flouri, T., Stamatakis, A., 2014. PEAR: a fast and accurate Illumina Paired-End reAd mergeR. Bioinformatics. 30, 614-620. Zhang, Z., Li, X., 2006. Evaluation of the effects of grading and starvation on the lysosomal membrane stability in Pacific oysters, Crassostrea gigas (Thunberg) by using neutral red retention assay. Aquaculture. 256, 537-541. Zurel, D., Benayahu, Y., Or, A., Kovacs, A., Gophna, U., 2011. Composition and dynamics of the gill microbiota of an invasive Indo‐Pacific oyster in the Eastern Mediterranean Sea. Environmental Microbiology. 13, 1467-1476. Zwietering, M., De Wit, J., Cuppers, H., Van't Riet, K., 1994. Modeling of bacterial growth with shifts in temperature. Applied and Environmental Microbiology. 60, 204-213.

246