THE MICROBIOME of the EASTERN OYSTER, Crassostrea

THE MICROBIOME OF THE EASTERN OYSTER,

Crassostrea virginica, IN HEALTH AND DISEASE

Eric G. Sakowski

A dissertation submitted to the Faculty of the University of Delaware in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Biological Sciences

Fall 2015

INFORMATION TO ALL USERS The quality of this reproduction is dependent upon the quality of the copy submitted.

In the unlikely event that the author did not send a complete manuscript and there are missing pages, these will be noted. Also, if material had to be removed, a note will indicate the deletion.

ProQuest 10014741

Published by ProQuest LLC (2016). Copyright of the Dissertation is held by the Author.

ProQuest LLC. 789 East Eisenhower Parkway P.O. Box 1346 Ann Arbor, MI 48106 - 1346

THE MICROBIOME OF THE EASTERN OYSTER,

Crassostrea virginica, IN HEALTH AND DISEASE

Eric G. Sakowski

Approved: ______Robin W. Morgan, Ph.D. Chair of the Department of Biological Sciences

Approved: ______George H. Watson, Ph.D. Dean of the College of Arts and Sciences

Approved: ______Ann L. Ardis, Ph.D. Interim Vice Provost for Graduate and Professional Education

I certify that I have read this dissertation and that in my opinion it meets the academic and professional standard required by the University as a dissertation for the degree of Doctor of Philosophy.

Signed: ______K. Eric Wommack, Ph.D. Professor in charge of dissertation

Signed: ______Shawn W. Polson, Ph.D. Professor in charge of dissertation

Signed: ______E. Fidelma Boyd, Ph.D. Member of dissertation committee

Signed: ______John R. Jungck, Ph.D. Member of dissertation committee

Signed: ______M. Ramona Neunuebel, Ph.D. Member of dissertation committee

Signed: ______Carl J. Schmidt, Ph.D. Member of dissertation committee ACKNOWLEDGMENTS

There are many people I need to thank, without whom this would not have been possible. I must begin by thanking my advisors, Dr. Eric Wommack and Dr. Shawn Polson. They have been a constant source of support and guidance. Eric has an infectious enthusiasm for viral ecology that has not diminished since we first met. His enthusiasm did not waiver even when the rest of ours did during those late nights on the research cruises. Eric has always encouraged me to think ambitiously and push boundaries, for which I am extremely grateful. I wish him the best of luck as he takes on new challenges as deputy dean. Shawn has been an invaluable source of knowledge on a seemingly inexhaustible number of topics (seriously, is there anything he doesn’t have at least a passing knowledge of?). No matter how busy he is – and he is always busy – his door has always been open. I truly appreciate the many thought-provoking impromptu conversations we shared. I must also thank my committee members who have been there to help guide me along the way. Dr. Fidelma Boyd has been a part of this journey through nearly every major milestone, beginning with my Preliminary exam. I thank her for her guidance and insights. Thank you to Dr. Diane Herson, who was always encouraging. I wish her all the best as she prepares for a well-deserved retirement. I must also thank Dr. John Jungck and Dr. Ramona Neunuebel who both graciously agreed to join my committee last minute. Their support is greatly appreciated. Thank you to Dr. Carl Schmidt, whose novel perspective has led me to answer questions I had not thought to

v ask. Finally, I must thank Dr. David Smith, who was always teaching whether he was inside a classroom or outside of it. I hope he is enjoying his retirement. There are several members of the community to whom I am grateful. Eric Weissberger and Carol McCollough of the Maryland Dept. of Natural Resources were instrumental in getting this project started. They were there to aid in sample collection early in the morning every month and helped put me in contact with the wider oyster community. I must also thank Mitch Tarnowski and David White of the Maryland Dept. of Natural Resources for allowing me to tag along the annual MD fall oyster survey and collect my own samples. Thank you, as well, to John Ewart of the Delaware Aquaculture Resource Center, E.J. Chalabala of the DE Center for the Inland Bays, and Jeff Harrison for their aid with sample collection. I must also express my gratitude to current and past members of the Wommack Lab. This would not have been possible without the aid of undergraduate interns and graduate students alike, and their support helped me to persevere when it seemed like there was no light at the end of the tunnel. For this I am truly grateful. Finally, I must thank my friends and family for their love and support. They were unwavering through the good times and bad and never once complained that I was talking about grad school AGAIN.

vi To my loving wife, Lauren. For everything.

vii TABLE OF CONTENTS

LIST OF TABLES ...... xiii LIST OF FIGURES...... xv ABSTRACT ...... xix

Chapter

1 INTRODUCTION...... 1

1.1 General Biology of Crassostrea virginica ...... 1 1.2 Ecological Impacts of C. virginica...... 2 1.3 Threats to Current and Future C. virginica Populations ...... 5

1.3.1 Habitat degradation ...... 5 1.3.2 Temperature and pH...... 6 1.3.3 Bacterial pathogens ...... 8 1.3.4 Viral pathogens...... 10 1.3.5 Protozoan pathogens...... 11

1.4 The Oyster Microbiome ...... 14

1.4.1 Microbiomes and host health...... 14 1.4.2 Commensal bacteria in oysters...... 16 1.4.3 The oyster microbiome and bacterial pathogens...... 19 1.4.4 Protozoan-bacterial interactions ...... 20

1.5 Natural Viral Assemblages in Oysters ...... 21

2 SEASONALLY DYNAMIC AND DIFFERENTIALLY REGULATED BACTERIAL COMMUNITIES IN OYSTER EXTRAPALLIAL FLUID MAY PLAY A ROLE IN SHELL FORMATION...... 25

2.1 Abstract ...... 25 2.2 Introduction ...... 26 2.3 Materials and Methods ...... 29

2.3.1 Annual survey sample collection...... 29 2.3.2 Bacterial abundance and correlations...... 30 2.3.3 Bacterial DNA isolation ...... 31

viii 2.3.4 16S amplification, barcoding, and sequencing...... 31 2.3.5 Denoising and taxonomic assignment...... 32 2.3.6 Rarefaction of oyster and water samples...... 32 2.3.7 Bacterial community composition of oyster and water samples. 33 2.3.8 Bacterial abundance correlations with environmental parameters...... 33 2.3.9 Core microbiome and oyster-associated OTUs...... 34

2.4 Results ...... 35

2.4.1 Oyster extrapallial fluid and water-associated bacterial abundances over one annual cycle ...... 35 2.4.2 Seasonal variability in alpha diversity...... 35 2.4.3 Bacterial community composition over one annual cycle...... 35 2.4.4 Microbiome dynamics and oyster-associated OTUs...... 36 2.4.5 Seasonal variability in bacterial community composition...... 38 2.4.6 Temporal differences in community responses to environmental parameters ...... 39

2.5 Discussion ...... 40

2.5.1 Bacterial community composition of C. virginica extrapallial fluid ...... 40 2.5.2 Seasonal variability of bacterial communities...... 44 2.5.3 A possible function of extrapallial fluid bacterial communities in shell formation...... 46 2.5.4 Potential impacts of warming and acidification on EF communities ...... 48

FIGURES ...... 51 TABLES...... 63

3 LOCATION, MORE THAN DISEASE, INFLUENCES OYSTER EXTRAPALLIAL FLUID MICROBIAL COMMUNITIES...... 95

3.1 Abstract ...... 95 3.2 Introduction ...... 96 3.3 Materials and Methods ...... 99

3.3.1 Oyster sample collection ...... 99 3.3.2 Screening for P. marinus infection...... 99 3.3.3 Bacterial DNA isolation ...... 99 3.3.4 16S amplification, barcoding, and sequencing...... 100 3.3.5 Adapter ligation and sequencing ...... 100

ix 3.3.6 Amplicon filtering and taxonomic assignment...... 101 3.3.7 Alpha diversity of EF communities...... 101 3.3.8 Beta diversity of EF communities ...... 102 3.3.9 Identification of enriched bacterial taxa...... 102

3.4 Results ...... 103

3.4.1 Alpha diversity of oyster EF bacterial communities ...... 103 3.4.2 Differences in oyster EF bacterial community composition by geography ...... 103 3.4.3 Differences in oyster EF bacterial community composition by P. marinus infection status ...... 104

3.5 Discussion ...... 105

FIGURES ...... 113 TABLES...... 121

4 RIBONUCLEOTIDE REDUCTASES REVEAL NOVEL VIRAL DIVERSITY AND PREDICT BIOLOGICAL AND ECOLOGICAL FEATURES OF UNKNOWN MARINE VIRUSES1 ...... 129

4.1 Abstract ...... 129 4.2 Introduction ...... 130 4.3 Materials and Methods ...... 132

4.3.1 Metagenomic libraries...... 132 4.3.2 Identification and distribution of putative virioplankton RNRs 134 4.3.3 Prevalence of RNR genes in DNA virome libraries...... 135 4.3.4 Inter-library RNR frequency normalization ...... 135 4.3.5 Intra-library RNR frequency normalization ...... 136 4.3.6 Assembly of putative virioplankton RNR sequences...... 136 4.3.7 Alignments and phylogenetic trees ...... 137 4.3.8 Predicted RNR group abundances in the Chesapeake Bay ...... 138 4.3.9 Predicted cyanophage population biology...... 138 4.3.10 Identification of Redoxins ...... 139 4.3.11 Contig assembly and annotation from the Rhode River...... 140

4.4 Results ...... 140

4.4.1 Diversity of putative virioplankton RNR sequences...... 140 4.4.2 Virioplankton RNR population distributions and ecology...... 141

4.5 Discussion ...... 143

x 4.5.1 The identity of phages carrying novel virioplankton RNR sequences...... 143 4.5.2 RNR as a proxy of phage population biology ...... 144 4.5.3 RNR class and phage biology...... 146

4.6 Conclusions ...... 150

FIGURES ...... 151 TABLES...... 163

5 VIRAL DIVERSITY IN OYSTERS IS ENDEMIC AND SEASONALLY DYNAMIC...... 166

5.1 Abstract ...... 166 5.2 Introduction ...... 167 5.3 Materials and Methods ...... 171

5.3.1 Annual survey sample collection...... 171 5.3.2 Biogeographic survey sample collection...... 172 5.3.3 Isolation and concentration of viruses from extrapallial fluid... 172 5.3.4 Isolation of viral DNA...... 173 5.3.5 PCR amplification of RTPR RNR sequences ...... 173 5.3.6 Adapter ligation and sequencing ...... 174 5.3.7 Amplicon filtering and error correction ...... 174 5.3.8 Identification of virome RTPR sequences...... 175 5.3.9 Clustering of RTPR peptides...... 176 5.3.10 Jackknifing of virome sequences...... 176 5.3.11 Identification of reference sequences ...... 177 5.3.12 Phylogenetic analyses of amplicon and virome sequences ...... 177 5.3.13 Hierarchical clustering of RTPR virome contigs ...... 178 5.3.14 Alpha diversity of extrapallial fluid samples...... 178 5.3.15 RTPR assemblage composition of extrapallial fluid samples ... 178 5.3.16 Relative abundance correlations with environmental parameters...... 179

5.4 Results ...... 179

5.4.1 Diversity of viral RTPR RNR sequences in oyster EF...... 179 5.4.2 Geographic distribution of RTPR sequences in oyster EF...... 180 5.4.3 Alpha diversity of RTPR viral assemblages in oyster EF ...... 182 5.4.4 Persistence and relative abundance of RTPR viral populations over one year ...... 182 5.4.5 Viral assemblage dynamics over time...... 182

xi 5.5 Discussion ...... 184

5.5.1 Identity of RTPR RNR phage...... 184 5.5.2 Assemblage structure and dynamics ...... 187 5.5.3 Geographic distribution of phage populations ...... 189 5.5.4 Comparison of viral diversity from PCR and viromes...... 191

FIGURES ...... 194 TABLES...... 209

6 CONCLUSIONS AND FUTURE DIRECTIONS ...... 210

REFERENCES...... 217

Appendix

A COPYRIGHT PERMISSION FOR CHAPTER 4 ...... 243

xii LIST OF TABLES

Table 2.1: Correlations between total bacterial abundance and measured water parameters...... 63

Table 2.2: Bacterial classes with significantly different relative and absolute abundances over the annual study...... 64

Table 2.3: Oyster EF bacterial families ranked by mean abundance over one annual cycle...... 64

Table 2.4: Water bacterial families ranked by mean abundance over one annual cycle...... 75

Table 2.5: Spearman’s rank R values of correlations between bacterial classes and water parameters...... 85

Table 2.6: Taxa observed in oyster EF and water samples grouped by environmental parameter correlations with relative abundance...... 86

Table 2.7: Taxa observed in oyster EF and water samples grouped by environmental parameter correlations with absolute abundance...... 90

Table 3.1: Oysters sampled from Choptank River and Tangier Sound locations. . 121

Table 3.2: Statistical comparisons of bacterial communities by sample group...... 121

Table 3.3: Bacterial taxa significantly associated with Choptank River or Tangier Sound oysters...... 122

Table 3.4: Bacterial taxa significantly associated with healthy or diseased oysters...... 126

Table 4.1: Distribution frequency of putative virioplankton RNR sequences among designated groups by environment...... 163

Table 4.2: Frequency of sampled genomes and RNR alpha subunits in cyanophage and pelagiphage populations...... 163

xiii Table 4.3: Distribution frequencies of RNR alpha subunit sequences among designated groups...... 164

Table 4.4: Characterization of viral metagenomic libraries queried...... 164

Table 5.1: Proteobacteria bacterium JGI 0000113-P07 contig 59.59 BLAST results...... 209

xiv LIST OF FIGURES

Figure 1.1: Ecological impacts of C. virginica. Differences in estuarine ecosystems with (left) and without (right) oysters are depicted...... 4

Figure 2.1: Bacterial abundance of oyster extrapallial fluid (EF) and water samples from the Smithsonian Environmental Research Center collected monthly from October 2010 to September 2011...... 51

Figure 2.3: Chao1 estimates of richness for oyster EF and water samples by season...... 53

Figure 2.4: The most abundant bacterial taxa (RDP-designated class level) identified in oyster EF and water samples over the annual study...... 54

Figure 2.5: Core and overrepresented OTUs identified in oyster EF and water samples...... 55

Figure 2.6: Unweighted and weighted diversity and variability of oyster EF and water samples by season...... 56

Figure 2.7: Hierarchical clustering of oyster EF and water samples by season...... 57

Figure 2.8: Bacterial classes that demonstrated significant seasonal change in oyster EF and water samples over one annual cycle...... 58

Figure 2.9: Hierarchical clustering of bacterial class relative abundance correlations with measured water parameters...... 59

Figure 2.10: Hierarchical clustering of bacterial class relative abundance correlations with measured water parameters and all shared taxa listed...... 60

Figure 2.11: Hierarchical clustering of bacterial class absolute abundance correlations with measured water parameters...... 61

xv Figure 2.12: Strength of Spearman’s rank correlations (R values) for bacterial classes in oyster EF or water samples with measured environmental parameters...... 62

Figure 3.1: Chao1 estimates of richness for oyster EF bacterial communities from two sites in the Choptank River (Royston and Tilghman Wharf) and two sites in Tangier Sound (Marumsco and Old Woman’s Leg)...... 113

Figure 3.5: Weighted UniFrac distances between oyster EF bacterial communities...... 117

Figure 3.6: The most abundant bacterial taxa (RDP-designated class level) in oyster EF from the Choptank River and Tangier Sound...... 117

Figure 3.7: Rank-abundance curves of oyster EF bacterial taxa (RDP-designated class level)...... 118

Figure 3.8: The number of oyster EF bacterial taxa significantly associated with site, geographic location, or Perkinus marinus infection status (health)...... 119

Figure 3.9: Hierarchical clustering of oyster EF bacterial communities by a combination of sample site and P. marinus infection status...... 120

Figure 4.1: Unrooted maximum likelihood tree with 100 bootstrap replicates of Class I and Class II alpha subunit RNR sequences from reference genomes and virioplankton metagenomic libraries...... 151

Figure 4.2: Unrooted maximum likelihood tree with 100 bootstrap replicates of Class II RNR reference and putative metagenomic ‘RTPR’ sequences.152

xvi Figure 4.3: Unrooted maximum likelihood tree with 100 bootstrap replicates of Class I alpha and Class II RNR reference and putative metagenomic ‘Cyano’ sequences...... 153

Figure 4.4: Cyanophage-like RNR diversity and distribution in the Gulf of Maine, Chesapeake Bay, and Dry Tortugas dsDNA libraries...... 154

Figure 4.5: Unrooted maximum likelihood tree with 100 bootstrap replicates of Class I alpha (left) and Class II (right) RNR reference and putative metagenomic ‘Other’ sequences...... 155

Figure 4.6: Predicted RNR frequency among dsDNA virioplankton populations from the Gulf of Maine, Chesapeake Bay, and Dry Tortugas...... 156

Figure 4.7: Distribution and dynamics of cyanophage populations...... 157

Figure 4.8: Predicted ORFs on Rhode River contigs 12643, 5585, 8066, and 12399...... 159

Figure 4.9: Unrooted maximum likelihood trees with 100 bootstrap replicates of Rhode River RNRs from contigs >5kb...... 160

Figure 4.10: Taxonomic distribution and alignment of ssDNA virome sequences... 162

Figure 5.1: Distribution of RTPR RNR sequences from aquatic virome and oyster extrapallial fluid amplicon libraries...... 194

Figure 5.2: Maximum likelihood tree with 100 bootstrap replicates of viral and reference RTPR RNR peptide sequences...... 195

Figure 5.3: Reference sequences with closest homology to RTPR RNR amplicons from oyster EF...... 196

Figure 5.4: UPGMA hierarchical clustering of Rhode River virome contigs >10kb with RTPR RNR sequences...... 197

Figure 5.5: Oyster EF RTPR RNR amplicon clusters shared with local and global virome sequences...... 198

Figure 5.6: Distribution of RTPR RNR peptide clusters from oyster EF samples by sample site...... 199

Figure 5.7: Abundance and distribution of RTPR phage populations (95% AA ID) in oyster EF...... 200

xvii Figure 5.8: Alpha diversity of oyster EF RTPR RNR amplicon libraries...... 201

Figure 5.9: Seasonal trends in RTPR RNR viral assemblages from oyster EF. Libraries were compared by weighted UniFrac distance...... 202

Figure 5.10: Weighted UniFrac principle coordinates analysis of oyster EF RTPR RNR libraries from the Rhode River (annual survey), Choptank River, Indian River, and DE Bay...... 203

Figure 5.11: Strength of correlation (Spearman’s Rank R) between RTPR RNR peptide cluster (95% ID) relative abundance and salinity, temperature, pH, and dissolved oxygen (DO) over one year with a one month lag. 204

Figure 5.12: Strength of correlation (Spearman’s Rank R) between RTPR RNR peptide cluster (95% ID) relative abundance and temperature, pH, salinity, and dissolved oxygen (DO) over one year...... 205

Figure 5.13: Maximum likelihood tree with 100 bootstrap replicates of oyster EF RTPR RNR amplicon clusters and references...... 206

Figure 5.14: Proportion of persistent oyster EF phage populations sampled from Dec. 2011 to Dec. 2012 in the Rhode River...... 207

Figure 5.15: The number of RTPR RNR peptide clusters shared between oyster sample sites in the Delmarva region...... 208

Figure 6.1: Known and potential factors influencing the composition of the C. virginica extrapallial fluid microbiome...... 210

Figure 6.2: Potential interaction between calcifying deltaproteobacteria and denitrifying bacteria...... 215

xviii ABSTRACT

The eastern oyster, Crassostrea virginica, is a keystone species in estuarine environments along the east coast of North America. Oysters are ecosystem engineers that provide habitat for native species and increase biodiversity, improve water quality, influence biogeochemical cycling, and structure trophic levels. To the detriment of estuarine environments, oyster populations today are a fraction of their historic levels due to anthropogenic and disease pressures, and climate change and pathogens pose serious threats to future populations. Next-generation, high-throughput sequencing technologies have led to an appreciation for the diversity of host- associated microbial communities (microbiomes) and their role in the health and physiology of their host; yet, little is currently known about the diversity of the oyster microbiome, the factors that shape the microbiome community, or the role of the microbiome in oyster health and fitness. In the work described herein, bacterial and viral populations within C. virginica extrapallial fluid (EF) – fluid produced by the oyster that initiates shell formation – were investigated using high-throughput, cultivation-independent approaches. The EF microbiome was characterized over time and between locations to identify temporal and geographic differences in community composition. In addition, the bacterial communities of healthy and diseased individuals were compared to identify changes in the normal microbiota in response to infection by Perkinsus marinus, a major oyster pathogen and the etiological agent of oyster disease Dermo.

xix The EF bacterial community was distinct from the surrounding water and seasonally dynamic. Water and EF bacterial communities were both influenced by temperature, but the same taxa in water and EF often responded differently to changing environmental conditions. The EF bacterial community was also influenced by geographic location, and oysters from the same sample site had more similar EF communities to each other than to oysters from other sites. In contrast, the EF community was largely unchanged by infection with P. marinus, though several individual taxa were associated with health status. A number of bacterial taxa were enriched in EF, notably Deltaproteobacteria, which are capable of catalyzing calcium precipitation and are found in caves and lithifying microbial mats where calcification is common. Given that the EF is the site of shell formation, which occurs via calcium precipitation, it is possible that the EF microbiome plays an unappreciated role in oyster shell formation. The abundance of Deltaproteobacteria in oyster EF decreased with higher water temperatures and may be an important consideration as average water temperatures increase in the future. Lytic phage populations were investigated in EF via ribonucleotide reductase (RNR), a diverse, abundant, and ecologically important marker gene of viral diversity in aquatic environments. The RNR assemblage as a whole was seasonally dynamic, although many of the most abundant populations were persistent over one year. The most abundant populations were also endemic to the Delmarva region. These endemic populations belonged to the dsDNA tailed bacteriophage family Podoviridae and likely infected members of the Proteobacteria, the most abundant bacterial phylum in oyster EF. Thus, these abundant and persistent lytic phage populations may play an important role in the top-down regulation of EF bacterial communities. Ultimately, a

xx better understanding of the EF microbial community, its role in oyster health, and how that role might be impacted by various threats will enable more targeted and effective conservation efforts in the future.

xxi Chapter 1

INTRODUCTION

1.1 General Biology of Crassostrea virginica

The eastern oyster, Crassostrea virginica (Gmelin, 1791), is a bivalve mollusc native to estuarine environments along the east coast of North America with a natural range extending from the Gulf of St. Lawrence, Canada to the coast of Brazil (Buroker 1983). C. virginica is tolerant of a wide range of environmental conditions, capable of surviving temperatures as high as 46°C during low tide in summer months (Galtsoff

1964) and freezing solid in the winter (Loosanoff 1965). C. virginica can also survive at salinities as low as 6ppt and as high as 35ppt (Kennedy, Newell et al. 1996). However, the optimum range for growth and reproduction is 10-28ppt (Wilson, Scotto et al. 2005). Sexual maturity is a function of size. C. virginica can reach sexual maturity in as little as four months in southern climates where more favorable conditions result in longer growing seasons (Hayes and Menzel 1981). In the Chesapeake Bay, C. virginica reaches sexual maturity at approximately one year of age when the oyster is at least 31mm in length (Rothschild, Ault et al. 1994). Spawning occurs in the summer months and is regulated by temperature. The minimal temperature for mass spawning is associated with latitude and increases moving from north to south (Shumway 1996). In the Chesapeake Bay, spawning begins to occur at 18°C (64°F) and continues above 20°C (68°F) (Kennedy and Krantz 1982). The early stages of C. virginica’s lifecycle are planktonic. Fertilized eggs develop into trochophore larvae and then shelled

1 veliger larvae within 24 hours (Gosling 2015). After approximately two weeks the larva develops a foot, becoming a pediveliger, and settles to the bottom of the water column, where it cements itself to a hard substrate (Kennedy, Newell et al. 1996). Cementation triggers a metamorphosis from the juvenile to adult form (Baker and Mann 1994). C. virginica is known to live up to 20 years (Buroker 1983); however, adults reach market size in approximately three years. Oysters are filter feeders that ingest particulate matter from the water column. However, they are not indiscriminate feeders. Size plays a role in the particles that are ingested or rejected as pseudofeces (Baldwin 1995). More important than size, however, in determining whether a particle is ingested or rejected is its nutritional value. C. virginica has been shown to preferentially ingest living organic material while rejecting silt or detritus (Newell and Jordan 1983; Ward, Levinton et al. 1997). The mechanism for this selection is not entirely understood. It appears that sorting occurs in the gill where retained particles are transported towards the dorsal tract and rejected particles towards the ventral tract (Ward, Levinton et al. 1997). The selection for ingestion may be due to the presence of lectins in oyster mucus that bind carbohydrate moieties on the cell surface of microalgal species (Espinosa, Perrigault et al. 2009).

1.2 Ecological Impacts of C. virginica

Oysters are keystone species in estuarine environments and provide numerous services vital to ecosystem health (Fig. 1.1). Many of these services are primary or secondary effects of water filtration. A single adult oyster is capable of filtering up to 50 gallons of water in 24 hours while feeding (NOAA). This has the primary effect of decreasing turbidity by removing phytoplankton and bacterial biomass, as well as

2 suspended sediments from the water column (Grabowski and Peterson 2007). Secondary effects of this process include: 1) the promoted growth of submerged aquatic vegetation (SAV); 2) stimulation of important biogeochemical cycles; 3) shifting primary production and trophic level structure; 4) reduced impacts of eutrophication; and 5) sequestration of carbon. By decreasing turbidity, oysters enable more light to penetrate the water column and improve conditions for SAV growth (Grabowski and Peterson 2007). This in turn provides habitat for many species, including nurseries for coastal fish (Thayer, Stuart et al. 1978). The concentration of inorganic and organic matter in oyster feces and pseudofeces stimulates denitrification in the sediment (Newell, Cornwell et al. 2002), which promotes a shift in primary production from phytoplankton to phytobenthos communities (Porter, Cornwell et al. 2004). In contrast, removing oysters from the ecosystem results in trophic restructuring dominated by phytoplankton and bacteria, which in combination with excess nutrient loads result in eutrophication and hypoxia (Newell 1988; Ulanowicz and Tuttle 1992; Paerl, Pinckney et al. 1998; Baird, Christian et al. 2004). Oyster reefs also serve as carbon sinks. Oysters turn ingested organic material into biomass, and shell deposition of growing oysters removes dissolved inorganic carbon from the water column. Thus, oyster reefs may help reduce atmospheric concentrations of greenhouse gases (Peterson and Lipcius 2003).

Additional ecosystem services provided by oysters are related to the physical structure of the oyster reefs. As occupants of sub-tidal mudflats, oyster reefs often provide the only hard substrate available for the attachment and growth of other sessile invertebrates (Kennedy, Newell et al. 1996). Oyster reefs also serve as habitat for polychaetes and crustaceans (Bahr and Lanier 1981), and are sanctuaries or feeding

3 grounds for crustaceans and juvenile fish species (Tolley and Volety 2005). Thus, oyster reefs increase the local biodiversity and productivity of the ecosystem (Rodney and Paynter 2006; Stunz, Minello et al. 2010). Finally, oyster reefs provide important structural functions. Reefs act as erosional barriers and help protect other habitats (Meyer, Townsend et al. 1997). Oyster reefs may also serve as important corridors between sheltering and foraging grounds (Peterson and Lipcius 2003).

CO2

CO2 O2 Phytoplankton

CO2 O2 Phytobenthos - NO3 N2

Figure 1.1: Ecological impacts of C. virginica. Differences in estuarine ecosystems with (left) and without (right) oysters are depicted. Oysters decrease turbidity and promote phytobenthos communities and submerged aquatic vegetation growth. They also stimulate biogeochemical cycles, act as carbon sinks, promote biodiversity, and provide erosional barriers.

4 1.3 Threats to Current and Future C. virginica Populations

1.3.1 Habitat degradation

Oyster populations are in decline worldwide. Numerous factors have played a role in this decline, including overfishing and disease (discussed below). One of the biggest factors in oyster decline is habitat degradation, which can be largely attributed to destructive fishing practices like dredging and hydraulic patent tongs (Rothschild, Ault et al. 1994). These practices reduce oyster reef height and disrupt the shells on the bottom. Lenihan et al. (1998) observed that one season’s equivalent of dredging reduced reef height by approximately 30%. Reef height was associated with hypoxia- induced mortality, as a greater proportion of oysters at a depth of 6m died than oysters found at 3m (Lenihan and Peterson 1998). Across the United States, oyster habitat was estimated to have declined by 64% since the late 19th century (Zu Ermgassen, Spalding et al. 2012). In the Chesapeake Bay, habitat degradation may be even more severe. In less than half a century (1907- 1952), oyster bar acreage declined by >50% (Rothschild, Ault et al. 1994). A survey of three Chesapeake Bay tributaries identified reductions in oyster habitat between 74% and 95% (Seliger and Boggs 1988). Degraded oyster habitat is associated with sandy or muddy bottoms, which negatively impacts spat settlement. Efforts to seed oyster habitat with oyster shells have had only transient success, and shells became highly sedimented after only 5.5 years (Smith, Bruce et al. 2005). Over 90% of historically productive oyster habitat was degraded in a recent survey, leading the authors to conclude that reasonable increases in management practices would not reverse habitat decline (Smith, Bruce et al. 2005).

5 Perhaps even more alarming than the area of oyster habitat lost is the resulting decrease in oyster biomass. Rothschild et al. (1994) estimated there was a 96% reduction in oyster abundance per habitable area in the Chesapeake Bay compared to historic levels. Worldwide, oyster biomass has decreased an estimated 88% (Zu Ermgassen, Spalding et al. 2012). Thus, areal measurements of oyster habitat mask the true extent of habitat degradation (Zu Ermgassen, Spalding et al. 2012).

1.3.2 Temperature and pH

The average surface temperature of the ocean is expected to rise from 19.7°C to 22.7°C by the turn of the century. Concurrently, the pH is predicted to decrease 0.2-

0.4 units (Mackenzie, Lynch et al. 2014). These changing environmental conditions pose a challenge to marine organisms. Calcifying organisms will be particularly affected by acidification, and acidification may be more pronounced in estuarine waters due to lower buffering capacities (Waldbusser, Voigt et al. 2011). Indeed, the average pH in some tributaries is already similar to values found to negatively impact shell formation in laboratory experiments (Waldbusser, Voigt et al. 2011). Oysters precipitate CaCO3 for shell formation and deposition from the surrounding environment, a process that is reversed (shell dissolution) under acidic conditions (Mackenzie, Ormondroyd et al. 2014). While seawater is often supersaturated with

Ca2+ (Braissant, Decho et al. 2007), biodeposition is an energetically expensive process that occurs at a much higher rate than abiotic CaCO3 precipitation (Waldbusser, Brunner et al. 2013). The mechanism of shell formation in oysters remains poorly understood; however, the process involves Ca2+ transport by granulocytes to the mineralization front in the extrapallial cavity (Mount, Wheeler et

6 al. 2004). Once at the mineralization front, granulocytes release the crystals, which are incorporated into an organic matrix (Wang, Li et al. 2013). Oyster larvae appear particularly susceptible to acidic conditions. In the Pacific oyster (Crassostrea gigas), larvae precipitate approximately 90% of their body weight as CaCO3 within the first 48 hours of life (Waldbusser, Brunner et al. 2013). This requires substantial energy input during a stage when energy is limited (Waldbusser,

Brunner et al. 2013). As a result, elevated partial pressure CO2 levels in the Pacific Northwest have been linked to failed oyster seed production (Barton, Hales et al. 2012). Numerous laboratory studies have also demonstrated the deleterious effects of acidification on marine calcifying organisms. Calcification by C. virginica juveniles decreased with a pH reduction of 0.5 units (Waldbusser, Voigt et al. 2011), a level that will likely be approached in the environment by the year 2100. Acidic conditions also impacted shell formation in the blue mussel (Mytilus edulis) and reduced shell flex (Mackenzie, Ormondroyd et al. 2014). Acidification has negative impacts on biodeposition absent other stressors; however, the combined impact of acidification and other stressors appears to be additive. Larval bay scallops (Argopecten irradians) displayed increased mortality under acidic conditions that was exacerbated by hypoxia (Gobler, DePasquale et al. 2014). Acidified waters were also shown to reduce the growth in early life stage hard clams (Mercenaria mercenaria). Later stage clams were resistant to acidified waters and hypoxic conditions individually but had reduced growth rates when the two stressors were combined (Gobler, DePasquale et al. 2014). Increased water temperature also negatively impacts shell formation, growth, and immune response in molluscs. Increased water temperature significantly reduced

7 shell strength in M. edulis, likely due to a re-allocation of energy from shell formation to stress response (Mackenzie, Ormondroyd et al. 2014). In fact, warmer water temperatures not only decreased shell strength, but also shell surface area and condition indices (Mackenzie, Ormondroyd et al. 2014). Higher temperatures also resulted in lower hemocyte counts and a greater proportion of phagocytosed cells in M. edulis (Mackenzie, Lynch et al. 2014). Finally, warmer temperatures are associated with the increased range and intensity of many bivalve pathogens.

1.3.3 Bacterial pathogens

A number of bacterial pathogens are recognized among various oyster species. Much of this knowledge comes from aquaculture where bacterial diseases are a major concern in commercial shellfish hatcheries. Molluscan aquaculture has doubled in production worldwide over the past few decades (Romero, Novoa et al. 2012) and is currently the fastest growing source of animal food production (Food and Agriculture Organization). However, the high mortality rates of larvae and juveniles in commercial hatcheries, facilitated by high larval density, increased water temperature and poor food quality, is a limiting factor in mollusc production (Bachere 2003). Less is known about the impact of bacterial and viral disease on natural oyster populations, due in part to the fact that the larvae and juveniles most susceptible to these pathogens leave behind little evidence of their existence in the environment. On the east coast, C. virginica juveniles are susceptible to Roseovarius Oyster Disease (ROD, formerly known as Juvenile Oyster Disease). This disease is caused by Roseovarius crassostreae, an Alphaproteobacteria species of the Roseobacter clade (Maloy, Ford et al. 2007). R. crassostreae invades the pallial cavity of juvenile oysters and colonizes the inner valve surface (Boardman, Maloy et al. 2008). In response to

8 this colonization, the oyster deposits conchiolin – an organic protein matrix – on the inner valve surface to wall off R. crassostreae and prevent further colonization (Boardman, Maloy et al. 2008). Conchiolin deposition resembles that of Brown Ring Disease (BRD) in clams, although the unrelated organism Vibrio tapetis, to which C. virginica is not susceptible (Allam and Ford 2006), causes BRD (Paillard, Le Roux et al. 2004). Outbreaks of ROD are largely confined to the northeastern United States from New York to Maine and occur during the summer when water temperatures rise above 20°C (Maloy, Ford et al. 2007). Symptoms include decreased growth rate, extreme cupping of the lower valve and breaking of the growing edge on the upper valve, conchiolin deposition, mantle retraction, and mantle lesions (Maloy, Ford et al. 2007; Boardman, Maloy et al. 2008). Mortality can exceed 90% in individuals <25mm in length and occurs 1-2 weeks after the onset of symptoms. Larger juveniles can also show signs of ROD (conchiolin deposits) but have much lower mortality rates (Maloy, Ford et al. 2007). Mortality of C. virginica has also been observed in the laboratory setting when challenged with several species of the genus Vibrio. Vibrio spp. are the most well documented bacterial pathogens of oysters (particularly C. gigas) and pose a serious threat to hatchery production. Members of the Vibrio genus have been implicated in a number of mortality events, and disease outbreaks caused by Vibrio spp. have increased with rising ocean temperatures (Schmitt, Rosa et al. 2012). In 2006, the re- emergence of Vibrio tubiashii in hatcheries on the U.S. west coast led to production declines of C. gigas in excess of 50% (Elston, Hasegawa et al. 2008). Brown and Losee (1978) isolated a Vibrio sp. similar to Vibrio anguillarum that caused spontaneous infection in C. virginica larvae. The pathogen was almost undetectable in

9 the water but caused near complete mortality 12-18 days after infection. Other species, including Vibrio corallilyticus and V. tubiashii, have been shown to induce high mortality rates in challenged oyster larvae in as little as 24 hours (Lim, Kapareiko et al. 2011; Karim, Zhao et al. 2013). The impact of bacterial pathogens, including Vibrio spp., on C. virginica mortality decreases as oysters mature. Three different C. virginica lines challenged with Vibrio spp. RE22 and RE101 showed mean survival time increases from 24 hours as larvae to over 6 weeks as juveniles (Gomez-Leon, Villamill et al. 2008). Tubiash (1965) noted that adult oysters remained healthy despite ingesting large numbers of larval pathogens.

1.3.4 Viral pathogens

Viral pathogens of oysters are far less understood than bacterial pathogens; nevertheless, several viruses have been implicated in mortalities of various oyster species. Most notable among viral pathogens of bivalves are the Iridoviridae, dsDNA nucleo-cytoplasmic viruses with linear genomes. Iridoviruses infecting bivalves include Gill Necrosis Virus, which caused mass mortalities of Portuguese oysters (Crassostrea angulata); Hemocyte Infection Virus, also identified in C. angulata during an epizootic in 1970; and the etiological agent of oyster velar virus disease, which caused mortality in C. gigas larvae (reviewed in Renault and Novoa 2004).

Ostreid herpesvirus 1 was responsible for summer mortalities in C. gigas larvae in 1991 and 2006 (Renault and Novoa 2004; Sauvage, Pepin et al. 2009). Similar to bacterial pathogens, larvae and juveniles appear more susceptible to viral pathogens than adults, and genetics appears to play a role in resistance/susceptibility (Renault and Novoa 2004).

10 1.3.5 Protozoan pathogens

Most bacteria pathogenic to juvenile oysters have little impact on adults (Tubiash, Chanley et al. 1965; Paillard, Le Roux et al. 2004). Instead, disease and mortality in adult oysters is most greatly influenced by protozoans (Dorrington and Gomez-Chiarri 2008). During the last half-century, two protozoan parasites have emerged as important sources of mortality among wild C. virginica populations: Perkinsus marinus, the etiological agent of the oyster disease Dermo; and Haplosporidium nelsoni, which causes the disease MSX. Both of these pathogens thrive in warm waters and higher salinities, and each of these diseases can result in high oyster mortalities during the summer and early fall months (Ewart and Ford 1993). For example, an MSX epizootic in 1985 resulted in a 47% mortality rate of oysters in the Delaware Bay (Powell, Ashton-Alcox et al. 2008). Since 1990 – the first year Dermo was detected in the Delaware Bay – oyster mortality rates in the Delaware Bay have remained above 15% and frequently surpassed 20%, compared to mortality rates of 5-10% in years prior to the emergence of these pathogens (Powell, Ashton- Alcox et al. 2008). In 2002, the survival rate of oysters sampled in the Chesapeake Bay was only 42% due to the combined effects of MSX and Dermo (Tarnowski 2012). Because these diseases are more prevalent in higher salinity regions, they have begun to alter the range of oyster populations (Powell, Ashton-Alcox et al. 2008). In this respect, P. marinus is considered to be the more important pathogen of oysters. Unlike H. nelsoni, P. marinus can persist in low salinity regions for years (Andrews 1996) and has a much wider distribution than H. nelsoni (Burreson and Ragone Calvo 1996). In 2013, P. marinus was identified at 98% of the Chesapeake Bay oyster bars surveyed (Tarnowski 2014). Its range has also extended north into traditionally colder bodies of water in recent years, likely due in part to rising water temperatures (Burge,

11 Mark Eakin et al. 2014). Despite annual infections, resistance to Dermo in wild oyster populations appears limited, and naïve oysters placed 5 km from diseased bars have acquired Dermo in as little as 10 days (McCollough, Albright et al. 2007). Simulations of genetically inherited resistance to Dermo in oysters suggest the rate of selection is too slow at mortality rates ≤ 25% for oyster populations to rebound through natural emergence of disease resistance (Powell, Klinck et al. 2011). P. marinus transmission occurs readily from oyster to oyster, peaking in late summer during periods of maximum host mortality when water temperature and salinity are high (Ragone Calvo, Calvo et al. 2003; Audemard, Ragone Calvo et al. 2006). Host death results in release of P. marinus cells into the water column and is likely the major source of disease transmission (Andrews 1996; Audemard, Ragone Calvo et al. 2006). Laboratory studies have demonstrated a dose-dependent relationship between P. marinus cellular abundance and rates of infection and mortality (Bushek, Allen et al. 1994; Chu and Volety 1997; Chintala, Bushek et al. 2002). However, this link has not been directly observed in situ, as infections tend to decrease in early autumn even when P. marinus abundance in the water column remains high (Audemard, Ragone Calvo et al. 2006). Conversely, transmission has also been observed in the environment between oysters during periods of low mortality (Ragone Calvo, Calvo et al. 2003; Audemard, Ragone Calvo et al. 2006), likely due to factors such as fecal release of P. marinus cells (Bushek, Allen et al. 1994; Scanlon 1997; Bushek, Ford et al. 2002). In fact, oysters with moderate or heavy P. marinus infections can shed thousands of cells each day via feces (Bushek, Ford et al. 2002). Other modes of transmission have also been described, including feeding by scavengers and parasitic snails (White, Powell et al. 1987; Diamond 2012).

12 P. marinus has three life stages, all of which have been determined experimentally to successfully initiate infection (Volety and Chu 1994; Andrews 1996); however, the main infective stage appears to be vegetative trophozoites (Volety and Chu 1994), while the role of swimming zoospores in transmission remains unknown. Infection was historically believed to initiate upon ingestion of viable P. marinus cells due to histological observations of P. marinus in the gut (Mackin 1951). However, an immunohistological study of P. marinus infection identified proliferating P. marinus cells in the gill, mantle, and labial palp, but not the digestive tract (Dungan, Hamilton et al. 1996). A comparison of exposure routes also found the heaviest infections occurred in oysters challenged with P. marinus via adductor muscle and shell cavity injections, while feeding produced the lightest infections (Chintala, Bushek et al. 2002). Recently, a second model of primary P. marinus infection has been proposed based on these observations. According to the proposed model, infection is acquired when C. virginica rejects P. marinus cells while feeding, sending them to the pseudofeces discharge area of the mantle (Allam, Carden et al. 2013). At this location, high parasite loads initiate infection. This process is likely aided by changes in P. marinus gene expression induced by oyster mucus. Oyster mucus caused a rapid increase in P. marinus replication rate (Allam, Carden et al. 2013; Espinosa, Winnicki et al. 2013) and was shown to increase P. marinus expression of virulence factors (Espinosa, Corre et al. 2014). Thus, the interaction of P. marinus with oyster mucus and fluids at the mantle likely plays an important role in subsequent infection. Distribution and disease progression occurs through hemocyte phagocytosis via binding of P. marinus trophozoites to the hemocyte surface receptor CvGal

13 (Tasumi and Vasta 2007). The trophozoite then replicates within the phagosome, inhibiting respiratory burst by superoxide dismutase production (Volety and Chu 1995) and suppressing cellular apoptosis (Hughes, Foster et al. 2010). Lysis of the hemocyte then occurs, releasing 8-32 daughter trophozoite cells. Emaciation of C. virginica occurs due to loss of hemocytes, tissue damage, and decreased nutrient availability (reviewed in Smolowitz 2014), although mortality may result from aggregate formation and impeded hemolymph flow resulting in embolism and circulatory failure (Mackin 1951).

1.4 The Oyster Microbiome

1.4.1 Microbiomes and host health

Bacteria, archaea, and viruses are by far the most successful biological entities on the planet. Their global abundances – estimated at 1030 bacteria (Kallmeyer, Pockalny et al. 2012) and 1031 viruses (Hendrix, Smith et al. 1999) – dwarf all other forms of life, and they are the planet’s greatest source of genetic diversity (Kristensen, Mushegian et al. 2010). This has enabled these microorganisms to thrive in even the harshest environments (Rothschild and Mancinelli 2001). It is not surprising, then, that microbes should be adapted to living on and within multicellular organisms. Yet until recently, the nature of these microbial communities (i.e. microbiomes) has been largely unknown. This has been attributable to the inability to cultivate and study the vast majority of bacterial species (the great plate count anomaly). High throughput, cultivation-independent technologies have overcome this hurdle to examining whole microbial communities. For example, high throughput sequencing of 16S rRNA amplicon libraries has enabled detailed characterizations of bacterial community

14 compositions, while more recent advances in throughput, read length, and assembly have enabled metabolic characterizations of abundant community members from shotgun bacterial metagenomes (see, for example Hug, Thomas et al. 2015). Emerging from these cultivation-independent studies of microbiome communities is the realization that microbiomes do not consist of random distributions of bacteria. Rather, bacterial communities have evolved with their host (Backhed, Ley et al. 2005) and interact in a complex relationship we are only beginning to understand. What is clear is that these microbial communities are vital to host health and physiology (see, for example Turnbaugh, Ley et al. 2007). Bacterial communities in the gut influence nutrient acquisition (Turnbaugh, Ley et al. 2006) and produce essential vitamins for their hosts (Gill, Pop et al. 2006). Research also indicates that bacterial communities play a key role in training the immune system (Mazmanian, Liu et al. 2005), and disruption of these communities may lead to diseases like irritable bowel syndrome (IBS) (reviewed in Bennet, Ohman et al. 2015). The mechanism of protection provided by microbiome communities is multifaceted (Khosravi and Mazmanian 2013). One way these communities can provide protection against pathogen colonization is through competitive exclusion (Reid, Howard et al. 2001). Diverse bacterial communities occupy environmental niches that might otherwise be colonized by pathogens, and decreased diversity may be associated with disease onset or progression. For example, respiratory microbiome diversity of cystic fibrosis patients decreased with disease progression despite an observed long-term resilience of community composition (Zhao, Schloss et al. 2012). Protection against pathogens can also come from production of inhibitory compounds (Kaewsrichan, Peeyananjarassri et al. 2006).

15 For many organisms, the nature of microbiome communities remains unknown. Characterizing the microbiome includes understanding which taxa are normally present in healthy individuals, the factors that shape microbiome communities and community dynamics, and the role of the microbiome in host fitness and physiology. This is particularly true for organisms like oysters that are exposed to trillions of bacteria and viruses from the water column daily.

1.4.2 Commensal bacteria in oysters

In general, the relationship between bivalves and their associated microbial communities is poorly understood; however, bivalves appear to select for commensal microbial communities that are unique from those found in the surrounding aquatic environment. Comparisons of bacteria cultivated from bivalves and the surrounding water have identified differences in taxonomic composition between the two (Lovelace, Tubiash et al. 1967; Kueh and Chan 1985; Pujalte, Ortigosa et al. 1999; Thomas, Wafula et al. 2014), indicating that the bacterial communities of bivalves and the water are semi-independent (Vasconcelos and Lee 1972). The presence of bacterial species in bivalves not found in the water suggests selective enrichment of species adapted for the bivalve microenvironment (Colwell and Liston 1960; Kueh and Chan 1985). Notably, biochemical analyses of cultivated bacteria from bivalves indicate similar metabolic capabilities between geographic locations and species (Colwell and Liston 1960; Murchelano and Brown 1968). The majority of bacteria cultivated from bivalves are proteolytic, capable of anaerobic fermentation, and able to utilize nitrate as an alternate electron acceptor (Colwell and Liston 1960; Murchelano and Brown 1968; Pujalte, Ortigosa et al. 1999). A number of bacterial isolates from C. virginica populations sampled in the Gulf of Mexico after the 2010 Deepwater Horizon oil spill

16 were able to degrade hydrocarbons and may be important components of the post-spill ecosystem (Chauhan, Green et al. 2013; Thomas, Wafula et al. 2014). Bivalves also appear to actively maintain a selected community. The Pacific oyster (Crassostrea gigas) can eliminate Escherichia coli from its tissues while retaining marine bacteria (Vasconcelos and Lee 1972), and the freshwater mussel (Anodonta cygnea) was observed to phagocytose E. coli introduced from polluted waters (Antunes, Hinzmann et al. 2010). Furthermore, the most commonly cultivated bacteria from the tissues and fluids of various oysters, clams, and mussels worldwide are similar and include Vibrio, Pseudomonas, Acinetobacter, Aeromonas, and members of the Flavobacteria/Cytophaga/Bacteroides (Colwell and Liston 1960; Lovelace, Tubiash et al. 1967; Murchelano and Brown 1968; Kueh and Chan 1985; Olafsen, Mikkelsen et al. 1993; Pujalte, Ortigosa et al. 1999; Beleneva, Zhukova et al. 2003; Antunes, Hinzmann et al. 2010). It is well established that the majority of marine bacteria are not readily cultivable (Ward, Weller et al. 1990), and this appears especially true for oysters. Direct counts of bacterial abundances within the Chilean oyster (Tiostrea chilensis) indicate that <0.01% of bacteria within oysters can be cultivated (Romero and Espejo 2001; Romero, Garcia-Varela et al. 2002). Furthermore, molecular fingerprinting approaches based on the 16S rRNA gene demonstrated that traditionally cultivated bacteria do not represent the most abundant members within the commensal bacterial community (Romero, Garcia-Varela et al. 2002; La Valley, Jones et al. 2009). Consistent with this observation, only 6% of the bacterial community in the digestive gland of Sydney rock oysters (Saccostrea glomerata) was identified as belonging to the family Vibrionaceae using cultivation-independent methods (Green and Barnes

17 2010). Similarly, Vibrio spp. were not identified in either C. gigas or Crassostrea corteziensis using cultivation-independent methods (Trabal, Mazon-Suastegui et al. 2012). However, Vibrio spp. are among the predominant bacterial taxa detected within oysters using cultivation-based approaches. Therefore, cultivation-based approaches do not accurately depict the composition of the oyster commensal bacterial community or the dynamics of its most abundant members. The few available cultivation- independent analyses of oyster-associated bacterial communities suggest seasonal differences in community composition (La Valley, Jones et al. 2009) and differences in community variability between oyster species (Trabal, Mazon-Suastegui et al. 2012). More recently, high-throughput, cultivation-independent characterizations of oyster gut, stomach and gill microbiomes have been reported. C. virginica gut and stomach microbiomes were compared at two different sites in Louisiana (King, Judd et al. 2012). Taxonomic differences were observed between sites and between stomach and gut. Surprisingly low diversity was found in the stomach, with communities dominated by Mollicutes at one site and Planctomycetes at the second. Neither stomach nor gut tissue displayed a distinct core microbiome – just five and 44 OTUs respectively – demonstrating the variability from oyster to oyster (King, Judd et al. 2012). This variability was also observed in C. gigas gill communities (Wegner,

Volkenborn et al. 2013). Proteobacteria, most notably Sphingomonas (Alphaproteobacteria), dominated gill communities in oysters from three different sites. However, no difference in community composition between sites was observed due to high variability between oyster gill microbiomes within a site. Interestingly, the authors noted the distribution of rare bacterial groups correlated with the genetic

18 relatedness between oysters. Gill-associated bacterial communities displayed a shift in composition after heat stress, with overall bacterial diversity decreasing in stressed oysters (Wegner, Volkenborn et al. 2013). Temperature-induced shifts in the bacterial communities of C. gigas hemolymph have also been reported (Lokmer and Mathias Wegner 2015). The impact of shifting community composition and diversity on oyster health remains unknown but deserves further attention as water temperatures rise and decreased diversity provides an opportunity for pathogen infection.

1.4.3 The oyster microbiome and bacterial pathogens

Bacterial diseases are a major concern in commercial shellfish hatcheries. One emerging approach for preventing bivalve disease in hatcheries is treatment of larvae with probiotic bacteria. From this work it is now apparent that autochthonous bacteria can promote survival through a number of mechanisms, including competitive exclusion of pathogens and production of inhibitory compounds (Bachere 2003). Currently, there are very few studies of probiotic bacteria in bivalve molluscs – almost none in C. virginica – and the protective role played by bacteria is poorly understood. The majority of these studies, however, have focused on the protective role of putative probiotic bacteria against Vibrio pathogens. For example, Alteromonas haloplanktis isolated from scallops displayed inhibition against pathogens Vibrio anguillarum and

Vibrio alginolyticus in vitro (Riquelme, Hayashida et al. 1996). Similarly, an unidentified bacterium S21 isolated from rearing seawater of C. gigas increased larval survival from 8.4% to 78% after 24 hours of exposure to V. alginolyticus (Nakamura, Takahashi et al. 1999), and Aeromonas media A199 protected against Vibrio tubiashii (Gibson, Woodworth et al. 1998) through indole production (Lategan, Booth et al. 2006). This is of note since 20% of bacterial isolates from C. virginica were capable of

19 producing indole (Murchelano and Brown 1968), and bacteria isolated from C. virginica significantly improved survival of larvae challenged with Vibrio corallilyticus (Lim, Kapareiko et al. 2011). C. virginica inner shell bacterial isolate Phaeobacter sp. S4 also protected C. virginica larvae and juveniles challenged with R. crassostreae and V. tubiashii (Karim, Zhao et al. 2013). However, the protection provided by these probiotics was short-lived (Karim, Zhao et al. 2013), and more work is necessary to understand the protective potential of the commensal community in C. virginica larvae and adults.

1.4.4 Protozoan-bacterial interactions

A few studies have examined the role of bacteria in the severity and mortality of protozoan diseases in fish and shellfish. An examination of the bacterial community in the gills of Atlantic salmon (Salmo salar L.) identified a Psychroserpens sp. (Flavobacteriaceae) phylotype specifically associated with Neoparamoeba pemaquidensis, the etiological agent of amoebic gill disease (Bowman and Nowak 2004). Co-infection of N. pemaquidensis and the related Flavobacterium Winogradskyella sp. enhanced infectivity and lesion formation in vivo (Embar- Gopinath, Butler et al. 2005). Likewise, increased virulence of the protist Entamoeba histolytica was observed in vitro upon co-infection with Gram-negative bacteria

(Bracha and Mirelman 1984). The digestive gland bacterial flora of Sydney rock oysters (Saccostrea glomerata) infected by Marteilia sydneyi was associated with the dominance of a single Alphaproteobacterium (Green and Barnes 2010). Since infection of Sydney rock oysters with M. sydneyi is not always fatal, it is possible that mortality results from a secondary infection by bacteria (Green and Barnes 2010). Similarly, it has been proposed that high parasite burdens in fish species like the

20 common dentex (Dentex dentex) (Company, Sitja-Bobadilla et al. 1999) and gilthead seabream (Sparus aurata) (Sitja-Bobadilla, Pujalte et al. 2006) impair immunocompetence and increase susceptibility to bacterial pathogens. It is unknown whether bacterial groups are associated with P. marinus infection in C. virginica that may synergistically increase mortality. However, the serine protease produced by P. marinus compromises the oyster’s immune defense and has been shown to suppress the ability of oyster hemocytes to eliminate Vibrio vulnificus (Volety and Chu 1995; Garreis, La Peyre et al. 1996; Tall, La Peyre et al. 1999). Therefore, P. marinus infection may promote gross changes to the commensal bacterial community and increase the risk of secondary infection by bacterial pathogens.

1.5 Natural Viral Assemblages in Oysters

Viruses are the most abundant biological entities on Earth and influence bacterial communities through cell lysis, mediation of nutrient and biogeochemical cycling in the microbial loop, and promoting horizontal gene transfer (Wommack and Colwell 2000; Suttle 2005). Diversity studies of viral assemblages have also indicated that viruses are the greatest reservoir of genetic diversity on the planet (Kristensen, Mushegian et al. 2010). However, characterizing natural viral assemblages remains a methodological challenge. Like bacteria, the majority (>99%) of viruses in environmental samples cannot be readily cultivated (Kennedy, Flemer et al. 2010). Unlike bacteria, viruses are polyphyletic and lack any universal marker gene equivalent to the 16S rRNA gene, making PCR-based taxonomic characterizations of all viruses impossible. In light of these limitations, viral ecology has relied upon two main approaches for characterizing natural viral assemblages: shotgun viral

21 metagenomics and PCR amplification of marker genes targeted at specific viral groups. Shotgun viral metagenomes (viromes) provide a powerful, unbiased approach to viral assemblage characterization (reviewed in Edwards and Rohwer 2005; Sakowski, Kress et al. 2011). The technique involves shotgun sequencing all dsDNA (or ssDNA or cDNA) from a viral sample and subsequently re-assembling viral genomes from DNA sequence data. Thus, the method requires no a priori knowledge of the types of viruses present in any given sample. However, cost; DNA sequencing technologies, including read length (Wommack, Bhavsar et al. 2008) and throughput; and bioinformatic tools such as error correction and assembly algorithms limit the method (Edwards and Rohwer 2005). Second-and third-generation sequencing technologies can now provide enormous amounts of data, but the amount of biologically meaningful information that can be extracted from viral metagenomic sequence data varies. This is partly due to the genetic diversity in viral genomes. Up to 95% of viral metagenomic sequences may show no similarity to database sequences (Breitbart, Thompson et al. 2007). This diversity poses challenges for assembling and annotating even single genes from viral shotgun metagenomic data, particularly with the short read length of sequences typical of most next-generation sequencers (Schoenfeld, Liles et al. 2010; Luo, Tsementzi et al. 2012). As the number of virome libraries has increased, many of these open reading frames (ORFs) have been observed across libraries, and comparisons between samples is now possible (Cesar Ignacio- Espinoza, Solonenko et al. 2013); yet, the function of these ORFs and oftentimes the type of viruses from which they originated remains a mystery.

22 A second approach to characterizing viral diversity relies upon the amplification of marker genes. Genetic markers of viral (bacteriophage) diversity include the structural proteins g20 (Zhong, Chen et al. 2002) and g23 (Comeau and Krisch 2008) for characterizing tailed Myoviruses (dsDNA contractile-tailed phages); functional genes such as DNA polymerase A (Breitbart, Miyake et al. 2004; Schmidt, Sakowski et al. 2014) for examining T7-like Podoviruses (dsDNA short, non- contractile tailed phages); and photosystem genes psbA and psbD (Sullivan, Lindell et al. 2006) for cyanophages. Nucleotide metabolism genes appear to be particularly useful for characterizing a broad range of viral diversity and may provide biological and ecological insights about the virus (Wommack, Nasko et al. 2015). For example, a single amino acid substitution in DNA Polymerase I (encoded by the DNA Pol A gene) is predictive of lytic or lysogenic phage lifestyles (Schmidt, Sakowski et al. 2014). Marker gene approaches to characterizing viral diversity have their own set of strengths and limitations. Unlike viral metagenomics, marker gene approaches limit the diversity that can be characterized. Amplifying protein-coding genes is fundamentally different than stable RNA genes, and primer bias can fail to amplify alleles even with degenerate primers. For example, a recent comparison of DNA Pol A genes from environmental virome and amplicon libraries revealed that PCR failed to amplify the most abundant Pol A clade in aquatic environments (Schmidt, Sakowski et al. 2014). Within a subset of diversity, however, marker genes provide unmatched insight into viral diversity. Even the most rare populations can be amplified by PCR and sequenced. In contrast, only the more abundant populations in virome libraries often have sufficient depth of sequencing to assemble a region long enough for

23 phylogenetic analysis (Sullivan 2015). This can be particularly important when considering the Bank Model of viral diversity, which predicts that the majority of viral diversity for any given environment lies in the low abundance members of the community (Breitbart and Rohwer 2005). Despite the methodological challenges of characterizing viral diversity, knowledge about the natural diversity of viruses in oysters is surprisingly limited. A few viruses have been related to oyster disease (Farley 1978; Munn 2006; Sauvage, Pepin et al. 2009) (see above), and the accumulation of viral pathogens of humans in oyster tissues has been examined from a public health perspective (Dubois, Merle et al. 2004; Le Guyader, Loisy et al. 2006; Bosch, Pinto et al. 2009). Viruses infecting Rickettsia-like organisms (RLOs) have been observed in several bivalve species, including Crassostrea ariakensis (Sun and Wu 2004; Friedman and Crosson 2012). In at least one instance, viral infection coincided with changes in RLO morphology and may be responsible for altering RLO virulence and pathogenicity (Friedman and Crosson 2012). Diverse and persistent assemblages of Vibrio phages were identified within oysters by cultivation techniques, even when host bacteria were undetectable (Baross, Liston et al. 1978; DePaola, McLeroy et al. 1997; DePaola, Motes et al. 1998; Comeau, Buenaventura et al. 2005). However, the diversity and dynamics of natural viral assemblages within oysters has yet to be examined in depth by either marker gene or viral metagenomic approaches.

24 Chapter 2

SEASONALLY DYNAMIC AND DIFFERENTIALLY REGULATED BACTERIAL COMMUNITIES IN OYSTER EXTRAPALLIAL FLUID MAY PLAY A ROLE IN SHELL FORMATION

2.1 Abstract

The eastern oyster is a key member of estuarine ecosystems but faces threats of habitat degradation, disease, and ocean warming and acidification. Commensal bacterial communities play important roles in the fitness of their hosts and have been reported for various oyster tissues. However, the bacterial community composition within oyster extrapallial fluid, located in the extrapallial cavity where shell mineralization occurs, remains poorly characterized. In particular, the influence of naturally changing environmental conditions on commensal communities in the extrapallial fluid remains unknown despite the negative impacts of warming and acidification on oyster shell growth and integrity. This study examined monthly changes in the composition of extrapallial fluid bacterial communities over one year within the context of seasonal environmental parameters. Total bacterial abundance was nearly two times greater in extrapallial fluid than water and was correlated with temperature, dissolved oxygen, and pH. Extrapallial fluid bacterial communities changed seasonally and were distinct from those within the surrounding water, and neither environment hosted bacterial populations that were present year round. At the phylum level, Proteobacteria were the most abundant taxa throughout the year, and Delta-and Gammaproteobacteria were significantly more abundant in extrapallial

25 fluid. OTUs from these bacterial classes were typical seasonal “core OTUs” unique to extrapallial fluid. Deltaproteobacteria are capable of calcium precipitation, and these taxa may play a role in calcium precipitation and shell mineralization within the extrapallial cavity. The relative abundance of Delta-and Gammaproteobacteria decreased at lower pH, indicating these potentially critical communities and their contributions to oyster fitness may be threatened by ocean acidification.

2.2 Introduction

The eastern or American oyster (Crassostrea virginica) is an economically and ecologically important species in estuarine environments along the east coast of North America. C. virginica landings in the United States exceeded 23 million pounds and $104 million in 2012 (NOAA NMFS). Beyond their importance as a fishery, the services oysters provide are invaluable to estuarine and coastal ecosystems. As filter feeders, oysters reduce turbidity and improve water quality (Grizzle, Greene et al. 2008). Oyster reefs provide hard substrate for the attachment and growth of other sessile invertebrate species and serve as sanctuaries for fish and invertebrates. As such, oyster reefs increase the abundance and diversity of other native species (Rodney and Paynter 2006; Stunz, Minello et al. 2010). Oyster reefs also protect the shoreline from erosion (Coen, Brumbaugh et al. 2007) and play an important role in nutrient cycling, particularly the removal of nitrogen (Kellogg, Cornwell et al. 2013). Since the late 19th century, oyster populations have dramatically decreased, causing a corresponding reduction in the ecosystem services they provide. C. virginica populations in the Chesapeake Bay are estimated to be 1% of their historical abundance (Newell 1988), largely due to anthropogenic factors, including overharvesting and habitat destruction (Rothschild, Ault et al. 1994). Populations have

26 been further decimated by the introduction and spread of protozoan parasites Haplosporidium nelsoni and Perkinsus marinus, the etiological agents of the oyster diseases MSX and Dermo, respectively. P. marinus is particularly prevalent and was identified in 98% of oyster bars surveyed in the Chesapeake Bay and its tributaries in 2013 (Tarnowski 2012). Beyond these issues lies the looming threat of ocean warming and acidification, factors which have been shown to negatively impact the growth rate, survival, and shell and immunological integrity of oysters and other calcifying bivalves (Waldbusser, Brunner et al. 2013; Gobler, DePasquale et al. 2014; Mackenzie, Lynch et al. 2014; Mackenzie, Ormondroyd et al. 2014). Warming temperatures have also been linked to the increased range, prevalence, and severity of H. nelsoni and P. marinus infections (Andrews 1996; Burreson and Ragone Calvo 1996), as well as the spread of bacterial pathogens like Vibrio spp. (Garnier, Labreuche et al. 2007; Elston, Hasegawa et al. 2008; Vezzulli, Brettar et al. 2012). Thus, future attempts to restore local C. virginica populations face a number of challenges. Key to meeting these challenges will be a better understanding of C. virginica’s ability to respond to these threats. The emerging field of microbiome research has implicated commensal microbial communities in a number of roles vital to their metazoan host. Like other filter-feeding bivalves, C. virginica interacts with a multitude of microbes from its environment. Several of these are known oyster pathogens, notably various Vibrio species among oyster larvae and juveniles (Bachere 2003; Paillard, Le Roux et al. 2004; Romero, Novoa et al. 2012). Also of concern are protozoan parasites P. marinus and H. nelsoni that infect adult oysters (Powell, Klinck et al. 2011; Soniat, Klinck et al. 2012). However, the majority of microbes within oysters are believed to be

27 commensal (Colwell and Liston 1960). Commensal communities can influence the acquisition of nutrients (Turnbaugh, Ley et al. 2006) and produce essential compounds for their hosts (Hill 1997; Shimada, Kinoshita et al. 2013). They can also provide protection against pathogens through competitive exclusion and the production of inhibitory molecules (Bachere 2003). The importance of commensal bacterial communities has been of particular interest in aquaculture, and bacteria administered as a probiotic supplement have increased the survival of marine invertebrate species challenged with known pathogens (Riquelme, Hayashida et al. 1996; Gibson, Woodworth et al. 1998; Nakamura, Takahashi et al. 1999; Lim, Kapareiko et al. 2011; Kesarcodi-Watson, Miner et al. 2012). However, such work in C. virginica has been hampered by limited knowledge regarding its normal commensal microbiome (Bachere 2003). Most studies examining oyster-associated bacteria have focused on cultivable bacteria, particularly those relevant to human health (e.g. Vibrio spp.). Comparisons of bacteria cultivated from bivalves and the surrounding water have identified differences in taxonomic composition between the two (Lovelace, Tubiash et al. 1967; Kueh and Chan 1985; Pujalte, Ortigosa et al. 1999), indicating that the bacterial communities of bivalves and the water are semi-independent (Vasconcelos and Lee 1972). Recently, cultivation-independent characterizations of bacterial communities associated with the stomach, gut, and gills have been reported and indicate distinct community compositions between tissues and between oysters and water (King, Judd et al. 2012; Wegner, Volkenborn et al. 2013; Chauhan, Wafula et al. 2014). Low-resolution cultivation-independent methods have also shown that oyster-associated communities are seasonally dynamic (La Valley, Jones et al. 2009). However, it remains unclear

28 what impact temporal (i.e. seasonal) changes in environmental conditions have on the commensal bacterial community composition and the potential consequences for C. virginica fitness. In this study, we examined the bacterial community composition of oyster extrapallial fluid (EF) over one year. Extrapallial fluid is produced in the extrapallial cavity, the site of shell mineralization. Thus, seasonal differences observed in EF-associated communities may provide insights into unappreciated impacts of changing environmental conditions on commensal communities and the potential roles of EF-bacterial communities in shell mineralization.

2.3 Materials and Methods

2.3.1 Annual survey sample collection

Cultured oysters (C. virginica) were obtained from Marinetics, Inc. (Cambridge, MD) and placed in wire cages on September 16, 2010. Cages contained 35 oysters each and were suspended approximately one meter below the surface of the Rhode River at the Smithsonian Environmental Research Center (SERC) in Edgewater, MD. Oysters were allowed to acclimate for 39 days prior to sampling. Five oysters were harvested each month from October 2010 to September 2011 and each oyster was rinsed with DI water and scrubbed with 70% ethanol prior to extrapallial fluid (EF) extraction. A hole was drilled into the posterior end of the oyster at the interface between valves with a 3/32-inch drill bit. EF was extracted using a 5mL syringe with a 23G needle. Samples were placed on ice and transported to Newark, DE for further processing. A 10L water sample was collected at the same site and time as the oysters. Temperature, pH, salinity, fluorescence, and dissolved oxygen (DO) were obtained from the long-term monitoring station located on site

29 (courtesy of Charles Gallegos, Smithsonian Environmental Research Center). The water sample was placed in a cooler filled with ambient water from the Rhode River to maintain temperature for transport back to Newark, DE.

2.3.2 Bacterial abundance and correlations

A 200µL aliquot of each oyster sample was combined with 37% formalin to a final concentration of 1% (v/v), snap-frozen in liquid nitrogen (LN2), and stored prior to use for bacterial enumeration at -80°C. Similarly, 4.5mL of the sample water was fixed with 37% formalin to 1% (v/v) final concentration and snap-frozen for bacterial direct counts. Samples were thawed on ice and combined with 0.22µm-filtered 1x PBS in the following ratios: 10µL EF in 990µL of PBS for oyster samples; 70µL water in 930µL PBS for water samples. The solutions were rocked at 30°C for twenty minutes to dissociate bacterial cells from mucopolysaccharides and improve filtration and then vacuum filtered onto a 13mm, 0.02µm Anodisc filter (Whatman). Filters were stained with 400µL of 2.5x SYBR Gold (Life Technologies) in the dark for 15 minutes prior to being mounted onto slides. Bacteria were visualized using epifluorescence microscopy with a 100X oil-immersion objective and images were collected at 15 randomly selected sites on each filter. Bacterial cells on the images were counted manually. Bacterial abundance was calculated based on the following equation: Bt =

Bc ÷ Fc × At ÷ Af ÷ S, where Bt = bacterial abundance mL-1, Bc = total number of bacteria counted, Fc = total number of fields counted, At = surface area of the filter (µm2), Af = area of each field (µm2), and S = volume of sample filtered (mL) (adapted from (Suttle and Fuhrman 2010). Bacterial abundances between oyster and water samples were compared over the annual survey by a two-tailed Mann-Whitney test. Significantly different (p<0.05) bacterial abundances between oyster and water

30 samples were identified by month by unpaired, two-tailed Student T-Tests with unequal variance. Mean bacterial abundances were compared to water conditions for all months except January and February, when water conditions were unavailable, to identify significant correlations between bacterial abundance and water parameters.

2.3.3 Bacterial DNA isolation

Oyster EF samples were combined with sterile 1x PBS buffer at a 1:25 ratio and rocked for 1 hour at 30°C to improve filtration. Samples were then filtered through a Millipore Sterivex 0.22µm filter unit. Approximately 500mL of water was filtered through a Millipore Sterivex 0.22µm filter. DNA was extracted from each of the filters as previously described (Crump, Kling et al. 2003) with amendments. Briefly, proteinase-K (20mg/mL) and lysozyme (100mg/mL) were combined with DNA Extraction Buffer (DEB: 100 mM Tris buffer (pH 8), 100 mM NaEDTA (pH 8), 100mM phosphate buffer (pH 8), 1.5 M NaCl, 1% CTAB) and added to the filter. Filters were incubated at 37°C for 30 minutes and subjected to three freeze-thaw cycles of -80°C and 37°C, followed by incubation at 37°C for 30 minutes. DEB was removed from the filter, combined with 10% (w/v) SDS and incubated for 2 hours at 65°C. DNA was washed twice with buffered phenol:chloroform:isoamyl alcohol (25:24:1) and once with chloroform. DNA was precipitated with 0.6 volumes of 100% isopropanol and resuspended in sterile 1x TE buffer.

2.3.4 16S amplification, barcoding, and sequencing

Universal 16S primers 27F (5’ – GCCTTGCCAGCCCGCTCAGTCAGAG – 3’) and 334R (5’ – GCCTCCCTCGCGCCATCAGAGTGAC – 3’) with 8-mer barcodes on the reverse primer were used to amplify ~300bp of the 16S rRNA gene

31 (V1-V2 region) from 3 oyster and 2 water samples each month. Bacterial DNA (1- 3µL) was combined with 10X buffer (1X final concentration), dNTPs (0.25mM each), forward primer mix (0.1µM final concentration), reverse primer mix (0.1µM final concentration), and TaKaRa Ex Taq DNA Polymerase (1.25 U) to a final volume of 25µL. PCR amplification of samples was performed using the following conditions:

95°C for 5 minutes; 30-34 cycles of 95°C for 30 seconds, 52°C for 30 seconds, 72°C for one minute; 72°C for 7 minutes. The entire PCR volume was run on a 1.8% agarose gel. Amplicon bands were excised and DNA gel purified (Qiagen QIAquick Gel Extraction kit). One hundred nanograms of DNA per sample were used for sequencing. Samples were sequenced on the Roche 454 Genome Sequencer with FLX Titanium technology.

2.3.5 Denoising and taxonomic assignment

Sequence reads were denoised with AmpliconNoise (Quince, Lanzen et al. 2011). Operational taxonomic units (OTUs) were picked in QIIME (Caporaso, Kuczynski et al. 2010) using Uclust (Edgar 2010) with a similarity cutoff of 97%. Representative sequence reads were assigned taxonomy using the Ribosomal Database Project (RDP) classifier (Wang, Garrity et al. 2007) with the Greengenes reference database (McDonald, Price et al. 2011). Sequences assigned the taxonomy

‘Chloroplast’ were excluded.

2.3.6 Rarefaction of oyster and water samples

Estimates of alpha diversity were calculated for oyster and water samples in QIIME (Caporaso, Kuczynski et al. 2010). Sample libraries from the same season were analyzed together to create seasonal libraries. Seasonal libraries were sub-

32 sampled at a depth of 35,000 sequences with 100 jackknifed replicates. Seasons were delineated as follows: Autumn (October-December); Winter (January-March); Spring (April-June); Summer (July-September).

2.3.7 Bacterial community composition of oyster and water samples

Bacterial community composition was identified over the annual cycle as the mean proportion of bacterial taxa in all oyster (n=36) and water (n=24) samples. Unweighted and weighted EF and water communities were compared over the annual cycle by analysis of similarities (ANOSIM) (Clarke 1993) with 999 permutations. Significantly different (p<0.05) bacterial taxa (class level) were identified by two- tailed Mann-Whitney tests. Seasonal bacterial community compositions were identified for oyster and water samples. Bacterial community compositions were identified by season as the mean proportion of bacterial taxa in all oyster (n=9) and water (n=6) samples for that season. Bacterial taxa (class level) that significantly differed (p<0.05) between seasons were identified by unpaired, two-tailed Student T- Tests with unequal variance. Oyster and water samples were grouped by season for Principal Coordinate Analyses (PCoAs) and Unweighted Pair Group Method with Arithmetic Mean (UPGMA) hierarchical clustering. Each group (e.g. Autumn oysters) was sub-sampled at a depth of 35,000 sequences by jackknifing with 100 replicates.

PCoAs and UPGMA clustering were performed using unweighted and weighted UniFrac distances (Lozupone and Knight 2005).

2.3.8 Bacterial abundance correlations with environmental parameters

Bacterial taxa (RDP-designated class level) observed in both oyster and water samples were used in correlation analyses. The mean monthly abundance of each

33 bacterial class was correlated with monthly salinity, DO, pH, temperature, and fluorescence values by Spearman’s rank correlation. Spearman’s rank R-values for each correlation were used to create a profile for each bacterial taxon. Taxa were clustered in R using the Heatmap.2 package with default settings. Bacterial taxa from oyster and water samples that grouped together were identified.

2.3.9 Core microbiome and oyster-associated OTUs

Core OTUs were identified in oyster and water samples using QIIME (Caporaso, Kuczynski et al. 2010). OTUs present in at least 80% of samples over the course of the annual cycle or season were considered to be part of the core microbiome for that time period. Seasonal core OTUs were compared between oyster and water and binned into two groups: Unique to Oyster and Shared Between Oyster and Water. Any OTU represented solely by oyster libraries in a given season was classified as ‘Unique to Oyster.’ Any OTU represented by oyster libraries and at least one water library in a given season was classified as ‘Shared Between Oyster and Water.’ OTUs that were represented solely by water libraries in a given season were not included in subsequent analyses. OTUs in each group were then characterized by taxonomy. The proportion of core OTUs unique to oyster samples was calculated by combining core OTUs from all four seasons. Non-core OTUs significantly associated with oyster samples (p<0.05, FDR-corrected) were identified by the Kruskal-Wallis test after removing OTUs present in fewer than 25% of samples.

34 2.4 Results

2.4.1 Oyster extrapallial fluid and water-associated bacterial abundances over one annual cycle

Over the course of the study, the bacterial abundance in EF was nearly twice that of the water (Fig. 2.1C), and EF and water-associated bacterial abundances were not correlated over time (Fig. 2.1D, r2 = 0.15). In contrast, when EF bacterial abundances lagged water-associated abundances by one month the two were highly correlated (Fig. 2.1D, r2 = 0.75, p<0.001). Extrapallial fluid bacterial abundances also correlated with water temperature when a one-month lag was introduced (Table 2.1). Bacterial abundances displayed similar seasonal trends in EF and water. Abundances decreased during the winter months in both environments before increasing in the late spring and peaking in summer (Fig. 2.1A-B).

2.4.2 Seasonal variability in alpha diversity

C. virginica EF and water bacterial communities displayed different trends in alpha diversity (community richness and evenness). Alpha diversity in oysters was highest in the summer and was lowest in the winter (p<0.05). In contrast, the alpha diversity of water communities was highest in the winter and spring and lowest during the summer and autumn (p<0.05) (Figs. 2.2 and 2.3).

2.4.3 Bacterial community composition over one annual cycle

Unweighted (p<0.001) and weighted (p<0.05) EF and water-associated bacterial communities were significantly different from each other. Over the course of one annual cycle, 143 different bacterial taxa (RDP-designated class level) were identified in oyster and/or water communities. However, the majority of OTUs belonged to the 12 most abundant classes between the two environments and

35 comprised on average 93% of the EF community and 97% of the water community (Fig. 2.4). Among these, the relative abundance of eight classes significantly differed (p<0.05) between EF and water communities, most notably among members of the Proteobacteria and Bacteroidetes (Fig. 2.4). Using direct count data it was possible to make estimates of the absolute abundances of bacterial groups within EF samples. Seven of the eight most abundant bacterial classes with significantly different relative abundances also showed significantly different (p<0.05) absolute abundances between sample types, while no significant difference in Betaproteobacteria absolute abundance was observed (Table 2.2). In total, 23 bacterial classes had significantly different (p<0.05) mean relative abundances in EF and water over the annual cycle, and 14 of these were more abundant in EF (Table 2.2). Three bacterial classes with significantly different mean relative abundances between EF and water did not have different absolute abundances between sample types (Betaproteobacteria, Oscillatoriophycideae, and Verrucomicrobiae). For all three of these classes the ratios of oyster to water relative abundance values were less than 1 (Table 2.2).

2.4.4 Microbiome dynamics and oyster-associated OTUs

Nineteen OTUs were present in at least 80% of oyster EF samples, while 23 OTUs were present in at least 80% of water samples (Fig. 2.5A). Among the 19 core

OTUs identified in EF samples, 15 were also core OTUs in water samples. Of the four core OTUs specific to EF samples, two showed 100% sequence identity with various Vibrio species. OTU 750 was identical to Vibrio brasiliensis and Vibrio neptunius, both of which have been isolated from Nodipecten nodosus (Lion’s Paw Scallop) larvae (Thompson, Li et al. 2003); coral pathogen Vibrio coralliilyticus; and nitrogen- fixing Vibrio diazotrophicus. OTU 13172 was identical to Vibrio aestuarianus, a

36 pathogen of Crassostrea gigas larvae (Pacific Oyster); fish pathogens Vibrio ordalii and Vibrio anguillarum; and Vibrio jasicida, a marine invertebrate-associated Vibrio species. OTU 337 shared 99% sequence identity with Desulfobacterium catecholicum, an anaerobe capable of sulfate and nitrate reduction and catechol oxidation (Szewzyk and Pfennig 1987). OTU 13841 shared 92% sequence identity with several species of the genus Rickettsia. A greater number of OTUs were present in at least 80% of samples within a given season than over the entire annual cycle (i.e. seasonal core OTUs) (Fig. 2.5B). A total of 16 bacterial taxa (RDP-designated class level) were represented by seasonal core OTUs (Fig. 2.5C), of which 12 were taxa with significantly different abundances between EF and water samples (9 higher in EF, 3 in water) (Table 2.2). In addition, 11 of the 16 taxa were comprised of seasonal core OTUs that were predominantly unique to EF samples (Fig. 2.5C). Among the 11 bacterial taxa with >50% seasonal core OTUs unique to EF samples, only Flavobacteria were significantly more abundant in water samples over the annual survey (Table 2.2). Over the annual study, 61 non-core OTUs from 17 bacterial taxa (RDP- designated class level) were significantly (p<0.05) associated with EF samples (Fig. 2.5D); 12 of these classes were also taxa represented by seasonal core OTUs (Figs. 2.5C, D). The greatest numbers of EF-associated OTUs were from unidentifiable bacteria (Bacteria; Other) and Alphaproteobacteria (Fig. 2.5D) despite Alphaproteobacteria contributing few seasonal core OTUs unique to EF (Fig. 2.5C). A similar trend was also observed with the Betaproteobacteria (Figs. 2.5C, D). In contrast, Delta-, Epsilon-, and Gammaproteobacteria were among the most abundant

37 classes contributing EF-associated OTUs that were also predominantly comprised of seasonal core OTUs unique to EF (Figs. 2.5C, D).

2.4.5 Seasonal variability in bacterial community composition

Extrapallial fluid and water bacterial communities were dynamic over time and displayed similar patterns of change (Fig. 2.6). However, in no instance did the EF and water communities demonstrate overlap along the first principle coordinate axis, which accounted for 22 and 36% of the variability in community composition in the unweighted and weighted analyses, respectively (Fig. 2.6B, E). In each environment, autumn and winter communities clustered together, as did spring and summer communities (Figs. 2.6B, 2.7A). Spring, summer, and autumn weighted EF communities also clustered together, while winter communities were more different (Figs. 2.6E, 2.7B). In contrast, winter and spring weighted water communities formed one cluster and summer and autumn communities formed another cluster (Figs. 2.6E, 2.7B). Sample-to-sample variability was similar across seasons for both EF and water communities. However, UniFrac distances were often significantly higher (i.e. communities were less similar) between EF communities than between water- associated communities (P<0.05) (Figs. 2.6C, F). Perhaps due to greater sample-to-sample variability within a season, the EF contained fewer bacterial classes that differed significantly between seasons than water (Fig. 2.8), and few trends were shared between the two environments. Epsilonproteobacteria relative abundance increased in winter before decreasing in spring, while Actinobacteria absolute abundance followed the same pattern. The autumn-winter and winter-spring transitions coincided with the greatest number of

38 bacterial classes with significant changes in abundance in both EF and water (Fig. 2.8).

2.4.6 Temporal differences in community responses to environmental parameters

To identify bacterial taxa (RDP-designated class level) with similar responses to changing environmental conditions over the year, taxon abundance (absolute and relative) was correlated with five environmental parameters: salinity, dissolved oxygen (DO), pH, temperature, and fluorescence. In total, 87 of the 143 bacterial classes identified in the annual study were present in both EF and water samples. Of these, 11 taxa had similar relative abundance correlation profiles (i.e. grouped together) between oyster and water samples (Figs. 2.9, 2.10). Likewise, 11 taxa shared similar absolute abundance correlation profiles between oyster and water samples (Fig. 2.11), although only three of these also showed similar correlations when considered in the relative abundance analyses (Figs. 2.9, 2,10). While few taxa observed in oyster and water samples had similar correlation profiles (i.e. grouped together) between treatments, most groups were comprised of taxa from both EF and water samples. In contrast, group VII was dominated by taxa from EF and included several members of the Acidobacteria and Chlorobi, as well as Flavobacteria. Group X included Alpha-, Epsilon-, and Gammaproteobacteria from EF samples (Fig. 2.10, Table 2.6). The taxa in group X comprised on average 44% of the oyster-associated bacterial community, peaking at 60% of the community in the winter. These same taxa were present in groups VIII (Epsilonproteobacteria) and IX (Alpha- and Gammaproteobacteria) for water samples (Fig. 2.10, Table 2.6). Group IX also

39 included water-associated Flavobacteria, and the taxa in this clade comprised on average 56% of the water-associated bacterial community. Oyster and water bacterial communities responded to changing environmental conditions on different time scales. Relative abundances of water-associated taxa were more strongly correlated with temperature, salinity, and dissolved oxygen than EF bacterial taxa (Fig. 2.12). Introducing a one-month lag of EF bacterial taxa relative abundances behind environmental parameters within the water column resulted in stronger correlations of EF taxa with temperature, salinity, dissolved oxygen, and fluorescence (Fig. 2.12). In contrast, introducing a one-month lag between water- associated bacterial community abundances and environmental parameters had no impact on the strength of correlations (data not shown). The strength of correlations of EF taxa relative abundances with a one-month lag and water-associated bacterial taxa abundances with no lag were similar for all parameters except DO, which was greater in EF bacterial taxa (Fig. 2.12). Temperature and DO had the strongest correlations with bacterial abundance in both EF (with lag) and water samples (Fig. 2.12).

2.5 Discussion

2.5.1 Bacterial community composition of C. virginica extrapallial fluid

Oyster-associated bacterial community members can be categorized as autochthonous (i.e. resident bacteria) or allochthonous (i.e. transient bacteria) (Romero and Espejo 2001). Extrapallial fluid sampled over the course of one year lacked a substantive core microbiome, with only 19 OTUs present in 80% or more of oysters sampled (Fig. 2.5A). This is not surprising considering the variability between EF communities (Fig. 2.6C, F), which may be attributable to genetic variability between

40 individual oysters (Wegner, Volkenborn et al. 2013). King et al. (2012) concluded many C. virginica gut and stomach bacterial community members were transient or opportunists, as core stomach and gut microbiomes were limited to just 5 and 44 OTUs, respectively. Similarly, low bacterial diversity in depurated C. gigas and Crassostrea corteziensis (Trabal, Mazon-Suastegui et al. 2012) suggested transient bacteria comprise a large proportion of oyster-associated microbial communities. The individual variability of oyster-associated microbial communities observed in this and other studies makes drawing conclusions about the temporal stability of commensal community members difficult, particularly in the absence of longitudinal samples following individual oysters. However, compositionally different bacterial communities may function similarly due to metabolic redundancy between taxa (Huttenhower 2012). Bacteria cultivated from oysters in different locations and even between different species have exhibited similar metabolic activities (Colwell and Liston 1960; Murchelano and Brown 1968), and this may apply to the EF community as a whole. Extrapallial fluid bacterial communities differed from the water column as determined by ANOSIM analyses. This agrees with cultivation-dependent and independent studies of various bivalve species and oyster tissues (Kueh and Chan 1985; Pujalte, Ortigosa et al. 1999; La Valley, Jones et al. 2009; Thomas, Wafula et al.

2014). Extrapallial fluid and water communities were both dominated by the phylum Proteobacteria, with the Alphaproteobacteria being most common in each sample type (Fig. 2.4). In contrast, Pseudomonas (Gammaproteobacteria), Vibrio (Gammaproteobacteria), and members of the Flavobacteria (Bacteroidetes) were dominant in cultivation-based characterizations of oyster-associated bacterial

41 microbiota (Colwell and Liston 1960; Lovelace, Tubiash et al. 1967; Murchelano and Brown 1968; Kueh and Chan 1985; Olafsen, Mikkelsen et al. 1993; Pujalte, Ortigosa et al. 1999; Chauhan, Green et al. 2013; Thomas, Wafula et al. 2014). It is well established that the majority of marine bacteria are not readily cultivable (Ward, Weller et al. 1990), including <0.01% of bacterial cells from oysters (Romero and Espejo 2001; Romero, Garcia-Varela et al. 2002). Furthermore, traditionally cultivated bacteria do not represent the most abundant members within oyster bacterial communities (Romero, Garcia-Varela et al. 2002; La Valley, Jones et al. 2009). In this study, the mean relative abundances of Pseudomonadaceae, Vibrionaceae, and members of the Flavobacteria in EF were 0.09 ± 0.03%, 2.8 ± 0.8%, and 5.4 ± 0.8%, respectively (Table 2.3). Extrapallial fluid communities also differed from microbial communities reported for other oyster tissues using cultivation-independent approaches. For example, Sphingomonas (Gammaproteobacteria) and Mycoplasma (Mollicutes) dominated the gill microbiome of Crassostrea gigas (Wegner, Volkenborn et al. 2013), but the Sphingomonadaceae and Mycoplasmataceae were the 94th (0.07 ± 0.02%) and 125th (0.03 ± 0.03%) most abundant bacterial families in C. virginica EF

(Table 2.3). C. virginica stomach communities were also dominated by Mollicutes and Planctomycetes, while Proteobacteria comprised only 20% of the gut community (King, Judd et al. 2012). In contrast, Proteobacteria comprised 52 ± 4% of the EF community (Fig. 2.4). Gut and stomach communities differed from each other (King, Judd et al. 2012), and the differences between these and EF suggests C. virginica tissues and fluids select for distinct microbial communities. RFLP banding patterns from Chilean oyster (Tiostria chilensis) homogenates indicated Arcobacter

42 (Epsilonproteobacteria) were common and abundant (Romero, Garcia-Varela et al. 2002). Epsilonproteobacteria were the eighth most abundant bacterial class in EF (3.9 ± 0.2%) (Table 2.3), and the Campylobacteraceae, which includes Arcobacter spp., comprised over 90% of the Epsilonproteobacteria OTUs. The relative abundance of Epsilonproteobacteria was also significantly greater in EF than water (Fig. 2.4, Table 2.2), indicating these bacteria may play an important role in the microbial communities of multiple oyster species, particularly considering the high oyster: water ratio observed for this taxa (Table 2.2). Alphaproteobacteria were the most abundant bacterial class in both EF and water samples and displayed similar relative abundances between the two (Fig. 2.4). However, ANOSIM analysis indicated the communities comprising this class differed between samples (p<0.05). In particular, the Phyllobacteriaceae (order Rhizobiales) was the second-most abundant bacterial family in EF behind unidentifiable Bacteria (Bacteria; Other) (Table 2.3), but was the 61st most abundant family in water (Table 2.4). This family includes the Mesorhizobia, which were also abundant in one study of C. gigas microbial communities (Fernandez-Piquer, Bowman et al. 2012). In that study, Mesorhizobia were only found in oysters prior to storage, suggesting they may be allochthonous. However, in this study OTUs comprising the Phyllobacteriaceae were identified in 30/36 oyster samples but only 15/24 water samples, and the difference in relative abundance between oysters (11.8 ± 3.9%) (Table 2.3) and water (0.08 ± 0.02%) (Table 2.4) suggests these may be resident members of the EF community. The relative abundance of Alphaproteobacteria taxa within EF and water also correlated differently with environmental parameters. In fact, very few bacterial

43 classes in EF and water grouped together according to their correlative profiles (Figs. 2.9, 2.10). Therefore, comparing the relative abundance of taxa between environments alone was insufficient for identifying differences in communities since only 23 bacterial classes had significantly different relative abundances but 76 bacterial classes displayed different temporal dynamics. This indicates that EF communities responded to changing environmental parameters (e.g. temperature, salinity, etc.) differently than water-associated communities. In contrast, those bacterial classes that grouped together between oyster and water samples were likely allochthonous bacteria. Not surprisingly, several of the putative allochthonous taxa were Cyanobacteria (Figs. 2.9, 2.10), a photoautotrohic phylum that would not be expected to survive in the dark extrapallial cavity. Thus, though only 19 of the bacterial classes sampled had relative abundances that differed between EF and water samples, most of the corresponding EF and water taxa were likely members of distinct communities existing under different selective pressures.

2.5.2 Seasonal variability of bacterial communities

Extrapallial fluid and water communities displayed seasonal patterns of abundance and composition. Overall, bacterial abundance was greater in EF than water (Fig. 2.1C) and water abundances agreed with previously reported values for the

Chesapeake Bay (Shiah and Ducklow 1994; Smith 1998; Heidelberg, Heidelberg et al. 2002). Cultivation-based bacterial abundance within bivalves has been reported to be greater than the water by an order of magnitude or more (Lovelace, Tubiash et al. 1967; Kueh and Chan 1985). In contrast, direct counts from this study indicate that bacterial abundance within the EF was approximately twice that of the water. This discrepancy is likely due to a greater number of readily cultivable cells in oysters, such

44 as Vibrio spp., which are considered to be among the most dominant cultivable bacterioplankton (Vezzulli, Brettar et al. 2012). Bacterial abundance increased during the summer in both sample types (Figs. 2.1A, B) and was correlated with water temperature (Table 2.1). Temperature has been identified as the limiting factor of bacterial abundance in the Chesapeake Bay (Shiah and Ducklow 1994; Smith 1998), and this also appears to be the case for EF communities. Bacterial abundance in EF samples responded more slowly to changing water temperature (Fig. 2.1D), and introducing a one month lag behind measured water parameters significantly increased correlations between bacterial abundance and temperature, salinity, DO, and fluorescence (Fig. 2.12). This suggests that the extrapallial cavity microenvironment is partially buffered against changing environmental conditions, including temperature, a surprise given that C. virginica is ectothermic. Nevertheless, temperature and DO were most strongly correlated with bacterial abundance in both EF and water samples (Fig. 2.12). Oyster-associated bacterial diversity also displayed seasonal trends. Seasonal differences in oyster-associated bacterial diversity have been previously reported, as RFLP fingerprinting of C. virginica bacterial communities indicated a greater number of phylotypes as water temperatures increased (La Valley, Jones et al. 2009). Likewise, the alpha diversity of EF communities increased during the summer and autumn (Figs. 2.2, 2.3). Interestingly, while alpha diversity decreased in EF during the winter and spring, alpha diversity in the water increased during this time (Figs. 2.2, 2.3). Thus, bacterial diversity in the EF was influenced very differently by water temperature than simultaneous water column bacterial diversity. Whether this is a direct response to changing temperatures or related to a physiological response of

45 oysters to water temperature (e.g. lower respiration and feeding rates during colder temperatures) remains unclear.

2.5.3 A possible function of extrapallial fluid bacterial communities in shell formation

Despite lacking a true core microbiome, several taxa were consistently overrepresented in seasonal core EF communities, including Gammaproteobacteria and the sulfate-reducing Deltaproteobacteria (Fig. 2.5C). Interestingly, the extrapallial cavity is the site of calcium precipitation and shell formation, and the EF community shared similarities with the microbial communities of rhodoliths (Cavalcanti, Gregoracci et al. 2014), coralline algae that form calcareous structures. Both organisms had communities distinct from the water and enriched for Gamma- and Deltaproteobacteria. The observance of Deltaproteobacteria in both of these organisms and functional groups related to organomineralization in rhodoliths suggests a potential role in biomineralization and shell formation (Cavalcanti, Gregoracci et al. 2014). Although seawater is supersaturated with calcium carbonate, precipitation does not spontaneously occur (Mount, Wheeler et al. 2004; Braissant, Decho et al. 2007). In contrast, rates of bivalve shell mineralization are much faster (Waldbusser, Brunner et al. 2013). Current models of oyster shell formation posit that Ca2+ precipitation occurs in granulocytes that are transported to the mineralization front where the crystals are released and interact with the organic matrix (Mount, Wheeler et al. 2004; Zhang, Fang et al. 2012; Wang, Li et al. 2013). In these models, the role of the EF bacteria community has not been considered. Sulfate-reducing bacteria within the Deltaproteobacteria have been implicated in a number of calcification processes. They are key members of lithifying microbial mat communities (Braissant, Decho et al.

46 2007) and play a role in the formation of pool fingers, stalactites, and stalagmites (Cacchio, Ercole et al. 2012). In fact, hypogean environments appear to select for calcifying microbes (Cacchio, Ercole et al. 2012), and our data indicate that the same appears true for the extrapallial cavity. Deltaproteobacteria were significantly more abundant in EF, and Deltaproteobacteria OTUs were overrepresented in seasonal core EF communities. One of the four core OTUs unique to EF over the annual survey were most closely related to Desulfobacterium catecholicum. In vitro experiments have demonstrated similar calcifying abilities between different members of the sulfate-reducing Deltaproteobacteria, including the Desulfovibrionaceae and Desulfobacteraceae (Braissant, Decho et al. 2007). These bacteria produce substantial amounts of exopolymeric substances (EPS), which can bind and act as nucleation sites for the precipitation of CaCO3 (Braissant, Decho et al. 2007; Cacchio, Ercole et al. 2012). It has been suggested that microbes may play an unappreciated role in carbonate precipitation in rhodoliths (Cavalcanti, Gregoracci et al. 2014). A similar role of SRB in remote calcification of oyster shells has been proposed but experimental evidence is lacking (Chinzei and Seilacher 1993; Vermeij 2013). Supporting these hypotheses is the discovery of endosymbiotic bacteria in archaeocyte-like sponge cells that mediate the calcification process (Uriz, Agell et al. 2012). This relationship is more intimate than that proposed for SRB-mediated calcification in oyster shells but shows bacterial-mediated calcification does occur in marine invertebrates. The potential role of non-calcifying bacteria in the shell calcification process should not be overlooked. Many heterotrophic bacteria can degrade EPS, which may serve to release calcium (Braissant, Decho et al. 2007). Biochemical characterizations

47 of bacteria cultivated from oysters indicate many isolates are capable of reducing nitrate (Murchelano and Brown 1968). This may be important in shell formation because denitrification and ammonification increase alkalinity and may aid indirectly in CaCO3 precipitation (Cacchio, Ercole et al. 2012). A significant proportion of these nitrate-reducing isolates were Vibrio spp., and we observed that two of the four EF- unique OTUs were related to nitrate-reducing bacterial taxa. While several Vibrio spp. have been implicated in shellfish diseases (reviewed in Bachere 2003), others are believed to be commensal microbial community members (Sawabe, Setoguchi et al. 2003; Antunes, Hinzmann et al. 2010). This appears to be the case for the two core Vibrio OTUs, as all sampled oysters in this study appeared healthy. Furthermore, Vibrio spp. are among the most abundant dissimilatory nitrate reducers in marine sediments (Bonin 1996), and dissimilatory nitrate reduction in anaerobic aquatic sediments is enhanced when S2- is available as an electron donor (Brunet and Garcia- Gil 1996). Threading these two lines of evidence together we posit that the persistence of denitrifying Gammaproteobacteria populations and sulfate-reducing Deltaproteobacteria populations in the EF are the result of their coordinated role in

CaCO3 precipitation and shell formation.

2.5.4 Potential impacts of warming and acidification on EF communities

The average surface temperature of seawater is expected to increase from 19.7 to 22.9°C by 2100, and pH is predicted to decrease 0.2-0.4 units during that time

(Mackenzie, Lynch et al. 2014). Oysters are ectothermic and poor regulators of pH (Mackenzie, Lynch et al. 2014), and studies have demonstrated the susceptibility of molluscs to rising temperatures and acidification, particularly in regards to shell formation and maintenance. In light of changing environmental conditions, addressing

48 the potential role of EF-associated communities in oyster shell mineralization is particularly important. Warmer temperatures reduced shell strength in the blue mussel (Mytilus edulis), while acidification reduced shell flex (Mackenzie, Ormondroyd et al. 2014). Acidification reduced survivorship in scallops and oyster (C. gigas) larvae, while a combination of acidification and hypoxia reduced growth rates in clams (Gobler, DePasquale et al. 2014). In addition, warming and/or acidification may result in decreased hemocyte abundance due to reallocation of resources during stress (Mackenzie, Lynch et al. 2014), an important consideration based on the role of granulocytes in shell formation. Warming temperatures and acidification have also been shown to influence microbial communities, including those commensal with oysters. Oysters incubated under heat stress had decreased bacterial diversity (Wegner, Volkenborn et al. 2013), and bacterioplankton incubated at lower pH display reduced diversity and altered community structure and metabolism (Siu, Apple et al. 2014). Notably, the relative abundance of Alpha-and Gammaproteobacteria had the strongest negative correlations with water temperature among EF taxa, while Bacteroidia, Epsilonproteobacteria, and Flavobacteria also had negative R values (Table 2.5). Thus, six of the nine most abundant taxa in EF displayed decreased relative abundances to some degree as temperature increased.

Interestingly, bacterial abundance in EF was inversely correlated with pH (Table 2.1), and the strength of this correlation did not increase with the introduction of a one-month lag (Fig. 2.12). In contrast, no correlation between bacterial abundance and pH was observed in water samples (Table 2.1). This suggests that EF bacterial communities were more susceptible to changes in pH, and that the response to

49 environmental pH changes was more rapid than to other measured environmental parameters. Given increasing trends of ocean acidification, susceptibility to pH changes could impact the abundance and community structure of oyster-associated bacteria. In fact, the Deltaproteobacteria, Actinobacteria, and Gammaproteobacteria had among the largest differences in Spearman rank R value for oyster and water taxa relative abundances correlated with pH (Table 2.5). In each instance, the relative abundance of these taxa was positively correlated with pH in EF, but inversely correlated with pH in water samples (Fig. 2.10, Table 2.5). Thus, ocean acidification could have a disproportionately large impact on oyster-associated microbial communities by suppressing the relative abundance of key taxa. In the case of the Deltaproteobacteria and Gammaproteobacteria, decreasing pH may indirectly impact shell formation by reducing the abundance of calcifying SRB and suppressing nitrate reduction and subsequent calcium precipitation.

50 FIGURES

A. B. 108 10 8 ) ) -1 -1 * 107 10 7 * * * * * * * * * 106 10 6

Bacterial Abundance (mL Bacterial 105 Abundance (mL Bacterial 10 5 OND J FMAM J J AS O N D J F M A M J J A S Sample Month Sample Month

1.2E+07 C. D. ) -1 1E+07 R = 0.75

108 ) *

-1 8E+06

R = 0.15 107 6E+06 No Lag Lag

4E+06 106

2E+06 Bacterial Abundance (mL Bacterial 105 Water Bacterial Abundance (mL Oyster EF Water 0.E+00 Treatment 0.0E+00 5.0E+06 1.0E+07 1.5E+07 2.0E+07 2.5E+07 -1 Oyster EF Bacterial Abundance (mL ) Figure 2.1: Bacterial abundance of oyster extrapallial fluid (EF) and water samples from the Smithsonian Environmental Research Center collected monthly from October 2010 to September 2011. Bacterial abundance was determined by direct counts with epifluorescence microscopy. A) & B) Mean monthly bacterial abundances of EF (A) and water samples (B). An asterisk indicates significantly different (p<0.05) abundances between EF and water. Error bars are SD. C) Mean bacterial abundance of oyster EF (n=52) and water (n=48) samples collected during the annual study (Mann- Whitney, p<0.001). D) Correlation between mean monthly EF and water bacterial abundances without a lag and with EF bacterial abundances lagging water bacterial abundances by one month.

51 A. B. 3500 7000

3000 6000

2500 (OTUs) 5000

2000 4000 Index Oyster Oyster Species 1500 3000 Water Water Chao1 1000 2000

Observed 500 1000

0 0 Autumn Winter Spring Summer Autumn Winter Spring Summer

C. D. 9 300 8 250 7

Diversity 6 200 5 Oyster 150 Oyster 4 Water Water 3 100 Shannon Index Phylogenetic

2 50

1 Faith's 0 0 Autumn Winter Spring Summer Autumn Winter Spring Summer

Figure 2.2: Alpha diversity estimates for oyster EF and water samples by season. Estimates were performed by combining sample libraries by season (oyster EF n=9 per season; water n=6 per season). Combined seasonal libraries were sub-sampled at a depth of 35,000 sequences with 100 jackknife replicates. Seasons were: Autumn (October-December); Winter (January-March); Spring (April-June); Summer (July- September). Oyster EF (black); Water (gray); Error bars (Standard Deviation of jackknifed replicates). A) Observed species (OTUs, 97% similarity). B) Chao1 index of predicted richness. C) Shannon diversity index. D) Faith’s phylogenetic diversity. All oyster-water pair wise comparisons were significantly different (T Test, p < 0.001).

52 6000

5000

4000 !"#"$%&'()#*+& ,-%#*+&'()#*+& 3000 ./+-%0&'()#*+& ."$$*+&'()#*+& !"#$%&'()*+& !"#"$%&,1#*+& 2000 ,-%#*+&,1#*+& ./+-%0&,1#*+& ."$$*+&,1#*+& 1000

0 0 5000 10000 15000 20000 25000 30000 35000 40000 45000

,&$-&'()./.)0#12&345627&

Figure 2.3: Chao1 estimates of richness for oyster EF and water samples by season. Estimates of EF (n=9 per season) and water (n=6 per season) diversity were performed with 100 randomizations without replacement. Seasons were: Autumn (October- December); Winter (January-March); Spring (April-June); Summer (July-September). Oyster EF (solid line); Water (dotted line); Error bars (Standard Deviation).

53 Phylum Class Actinobacteria Actinobacteria Bacteroidia * Bacteroidetes; Other Bacteroidetes Flavobacteria * Sphingobacteria * Cyanobacteria Synechococcophycideae Alphaproteobacteria Oyster EF Water Betaproteobacteria * Proteobacteria Deltaproteobacteria * Epsilonproteobacteria * Gammaproteobacteria * Unassigned Bacteria; Other * 0 0.1 0.2 0.3 0.4 Mean Proportion of Community Figure 2.4: The most abundant bacterial taxa (RDP-designated class level) identified in oyster EF and water samples over the annual study. The mean proportion of the community comprised by each bacterial class was identified for all EF samples (n=36) and water samples (n=24). Depicted taxa comprised 93% of EF communities and 97% of water communities. Taxa with significantly different (p<0.05) mean relative abundances between oyster EF and water are noted. Oyster EF (Black); Water (Grey); Error bars (SE).

54 A. 120 B. 350

300 100

250 80 Autumn

200 Winter OTUs 60 OTUS

Spring

# Oyster EF # 150 Water Summer 40 100

20 50

0 0 50% 55% 60% 65% 70% 75% 80% 85% 90% 95% 100% 50% 55% 60% 65% 70% 75% 80% 85% 90% 95% 100% % of Samples Where OTU Was Observed % of Samples Where OTU Was Observed C. Proteobacteria; Other D. Alphaproteobacteria Nitrospira * Bacteria; Other Clostridia * Deltaproteobacteria Bacteroidetes * Betaproteobacteria Bacteroidetes; Other Bacteroidia

Gammaproteobacteria Holophagae * axa axa

T Epsilonproteobacteria T Deltaproteobacteria * Flavobacteria Epsilonproteobacteria * Clostridia Gammaproteobacteria * Sphingobacteria Bacterial Bacterial Bacteria; Other * * Actinobacteria Flavobacteria * TM7; Other Betaproteobacteria * Proteobacteria; Other Alphaproteobacteria Nitrospira Actinobacteria Bacilli Synechococcophycideae OPB56 Sphingobacteria * Ignavibacteria 0 10 20 30 40 50 60 70 80 90 100 0 1 2 3 4 5 6 7 8 9 10 % of Each Taxa # OTUs Associated with Oysters Oyster and Water Unique to Oyster Figure 2.5: Core and overrepresented OTUs identified in oyster EF and water samples. A) Number of OTUs identified in a given proportion of samples over the course of the annual study. B) Number of OTUs identified in a given proportion of samples within a given season. Seasons were: Autumn (October-December); Winter (January-March); Spring (April-June); Summer (July-September). Oyster EF (solid line); Water (dotted line). C) Proportion of seasonal core OTUs grouped by bacterial taxa (RDP-designated class level) that were unique to oysters or shared with water samples. OTUs identified in at least 80% of EF samples were included. Taxa with significantly different (p<0.05) relative abundances over the annual study are noted. Black (Unique to oyster EF); Grey (Shared). D) Census of OTUs that were not part of a seasonal core in EF samples but were significantly associated with oyster samples (Kruskal-Wallis, p<0.05).

55 A. B. C.

0.3 0.9 0.9 Winter * * a * 0.8 a 0.85 0.2 Autumn ce n

a 0.7 t

0.8 s i D 0.1 0.6 0.75

Winter rac f 0.5 ni

Autumn U 0.7 0 d

Oyster e 0.4 Oyster 0.65 t Water Spring Water igh

-0.1 Spring e 0.3 0.6 nw

U 0.2

Unweighted Unifrac Distance Unweighted Summer 0.55 -0.2 Summer

Percent VariationPercent 18.68% Explained: 0.1 0.5 0 0 2 4 6 8 10 12 -0.3 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 Autumn Winter Spring Summer Months Between Samples Season Percent Variation Explained: 21.62%

D. E. F. 0.3 0.8 0.7 * * * *

0.7 0.2 Spring ce 0.6 Summer n Autumn a t

0.6 s b 0.1 i 0.5 D

0.5 Winter rac 0 Spring f

ni 0.4 Summer 0.4 U d

Oyster e Oyster -0.1 t 0.3 0.3 Water b Water igh e

0.2 -0.2 W 0.2 Autumn Weighted Unifrac Distance Weighted 0.1 Winter

Percent VariationPercent 20.95% Explained: -0.3 0.1 0 0 2 4 6 8 10 12 -0.4 0 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 Autum Winter Spring Summer Months Between Samples Percent Variation Explained: 35.92% Season Figure 2.6: Unweighted and weighted diversity and variability of oyster EF and water samples by season. A) Unweighted UniFrac distance between samples over time. B) Unweighted principle coordinate analysis of EF and water samples by season. C) Mean unweighted UniFrac distance between samples by season. D) Weighted UniFrac distance between samples over time. E) Weighted principle coordinate analysis of EF and water samples by season. F) Mean weighted UniFrac distance between samples by season. Error bars (Standard Deviation). Significant differences (p<0.05) between treatments (oyster vs. water; *) and within treatments (a, b) are noted.

56 A. Autumn Oysters

100 Winter Oysters

100 Summer Oysters

80 77 Spring Oysters

Winter Water

100 Autumn Water 100

Summer Water

100 Spring Water 0.05

B. Winter Water

100 Spring Water

100 Autumn Water

100 Summer Water

100 Autumn Oysters

100 Summer Oysters

100 100 Spring Oysters

Winter Oysters 0.03 Figure 2.7: Hierarchical clustering of oyster EF and water samples by season. Each group (e.g. Autumn oysters) was sub-sampled at a depth of 35,000 sequences by jackknifing with 100 replicates. A) Hierarchical clustering of unweighted samples. B) Hierarchical clustering of weighted samples. Node labels are jackknifing support values. Scale bar represents UniFrac distance between samples.

57 A. B.

120 Opitutae

15 20 4c0d-2 Actinobacteria Bacteria; Other Bacteria; WS6; Other WS6; Sphingobacteria Epsilonproteobacteria 15 Spartobacteria Bacteroidetes; Other Bacteroidetes; 7.5 Other; Other Other; Betaproteobacteria

AB16 10 Elusimicrobia Phycisphaerae Epsilonproteobacteria ZB2; Other ZB2; Other; Other Other; Actinobacteria 0 Other Proteobacteria; Betaproteobacteria AB16 Epsilonproteobacteria

5 Oscillatoriophycideae Betaproteobacteria Alphaproteobacteria Change in relative abundance GKS2-174 Actinobacteria Deltaproteobacteria Proteobacteria; Other Proteobacteria; Other Bacteroidetes; Alphaproteobacteria Other Tenericutes; Other ZB2; Epsilonproteobacteria Elusimicrobia 4c0d-2 Actinobacteria VC12-cl04 Bacteria; Other Bacteria; Actinobacteria -7.5 0 Flavobacteria Change in absolute abundance Mean Fold Change Mean Fold Change Change in absolute & relative -5 -15 abundance

-10

-22.5 -15

-30 -20 Autumn-Winter Winter-Spring Spring-Summer Autumn-Winter Winter-Spring Spring-Summer Season Season Figure 2.8: Bacterial classes that demonstrated significant seasonal change in oyster EF and water samples over one annual cycle. A) Bacterial classes whose absolute and/or relative abundance changed between seasons in oyster EF. B) Bacterial classes whose absolute and/or relative abundance changed between seasons in water. Change in relative abundance (Black); Change in absolute abundance (White); Change in relative and absolute abundance (Grey). Italicized taxa were not detected in one season and fold change was estimated.

58 0.5 0.5 R Value Taxa from Oyster EF and Water with Similar Correlation Profiles

Acidobacteria; Other OP8-1 I Betaproteobacteria

Verrucomicrobiae II

Verrucomicrobia: Other III Synechococcophycideae Oscillatoriophycideae IV Spartobacteria

GN15 V

VII

TM7-3 VIII

Bacteroidia IX X y e e H O p D atur r Salinit empe T Fluorescenc Environmental Parameters Figure 2.9: Hierarchical clustering of bacterial class relative abundance correlations with measured water parameters. Eighty seven bacterial taxa (RDP- defined classes) occurred in both oyster EF and water samples and were compared. A correlation profile of relative abundance was created for each taxon from the R values of Spearman’s rank correlations with five environmental parameters. Taxa were clustered by similar correlation profiles. Bacterial classes belonged to one of ten groups according to their correlation profile. Bacterial classes from EF and water that had similar correlation profiles (i.e. were part of the same group) are noted. EF bacterial classes are noted on the y-axis dendrogram in red, water in blue.

59 0.5 0.5 R Value Taxa Water_Acidobacteria;c__Sva0725 Water_Acidobacteria;c__ Water_Chlorobi;c__SJA 28 Water_OP3;c__BD4 9 Water_OP8;c__OP8_1 Water_Tenericutes;c__Erysipelotrichi Oyster_Proteobacteria;c__Betaproteobacteria Water_Other;Other;Other Water_Chlorobi;c__BSV19 Oyster_Chloroflexi;Other Oyster_Acidobacteria;c__RB25 Oyster_Acidobacteria;c__ Oyster_Acidobacteria;c__Solibacteres Oyster_OP8;c__OP8_1 Oyster_Proteobacteria;c__Zetaproteobacteria Oyster_SC3;c__ Oyster_TM6;c__SBRH58 I Oyster_Verrucomicrobia;c__Verruco 5 Water_Firmicutes;Other Oyster_Actinobacteria;c__Actinobacteria Water_Cyanobacteria;c__Gloeobacterophycideae Oyster_Proteobacteria;c__Deltaproteobacteria Oyster_Planctomycetes;c__agg27 Water_Bacteroidetes;c__Sphingobacteria Water_Chloroflexi;c__SOGA31 Water_TG3;c__TG3 2 Water_Proteobacteria;c__Betaproteobacteria Oyster_Nitrospirae;c__Nitrospira Water_Acidobacteria;c__Acidobacteria Oyster_OP8;c__OP8_2 Water_Cyanobacteria;Other Water_Verrucomicrobia;c__Opitutae Oyster_SM2F11;c__ Oyster_Planctomycetes;c__Phycisphaerae Water_GN02;Other Water_Caldithrix;c__Caldithrixae Oyster_WS6;c__WM1006 Oyster_OP3;c__koll11 Oyster_OP3;c__BD4 9 Oyster_MVP 15;c__ Oyster_Chlorobi;c__OPB56 II Water_Acidobacteria;c__Chloracidobacteria Oyster_Tenericutes;Other Oyster_WS3;c__PRR 12 Water_Chloroflexi;c__Anaerolineae Oyster_Fusobacteria;c__Fusobacteria Oyster_Chloroflexi;c__Chloroflexi Water_Verrucomicrobia;c__Verrucomicrobiae Oyster_Verrucomicrobia;c__Verrucomicrobiae Water_WS3;c__PRR 12 Water_Lentisphaerae;c__Lentisphaerae Water_WS6;c__SC72 Water_Verrucomicrobia;c__R76 B18 Oyster_Bacteroidetes;c__Sphingobacteria Oyster_Verrucomicrobia;Other Oyster_Verrucomicrobia;c__Opitutae Oyster_Chloroflexi;c__SOGA31 III Water_Bacteroidetes;Other Water_Verrucomicrobia;Other Oyster_Cyanobacteria;c__Synechococcophycideae Oyster_Spirochaetes;c__Spirochaetes Water_Bacteria;Other;Other Oyster_OP11;c__WCHB1 64 Oyster_Elusimicrobia;c__Elusimicrobia Water_Cyanobacteria;c__Synechococcophycideae Water_Gemmatimonadetes;c__Gemmatimonadetes Water_Cyanobacteria;c__Oscillatoriophycideae Oyster_Verrucomicrobia;c__Spartobacteria Water_Chlorobi;c__Ignavibacteria Water_Verrucomicrobia;c__Spartobacteria IV Oyster_Cyanobacteria;c__Oscillatoriophycideae Oyster_Cyanobacteria;c__S15B MN24 Oyster_Cyanobacteria;c__4C0d 2 Oyster_Acidobacteria;c__Acidobacteria Water_TM7;Other Water_Proteobacteria;Other Oyster_Chlorobi;c__Ignavibacteria Water_OP8;c__OP8_2 Water_Acidobacteria;c__OS K Water_Proteobacteria;c__Deltaproteobacteria Oyster_Tenericutes;c__Mollicutes Oyster_Bacteria;Other;Other Water_Planctomycetes;c__Phycisphaerae Oyster_Other;Other;Other Oyster_GN02;Other Water_Planctomycetes;c__agg27 Water_Chloroflexi;c__Chloroflexi Oyster_GN02;c__BB34 Oyster_WS6;c__SC72 Oyster_SR1;c__ V Water_Actinobacteria;c__Actinobacteria Water_Chlorobi;c__OPB56 See Table 2.6 Water_GN04;c__GN15 Oyster_GN04;c__GN15 Oyster_Cyanobacteria;Other Water_Cyanobacteria;c__S15B MN24 Oyster_Verrucomicrobia;c__Methylacidiphilae Oyster_Cyanobacteria;c__Gloeobacterophycideae Water_SAR406;c__AB16 Oyster_TG3;c__TG3 2 Oyster_GN02;c__VC12 cl04 Oyster_Acidobacteria;c__Holophagae Water_Nitrospirae;c__Nitrospira Oyster_Tenericutes;c__ML615J 28 Water_Fusobacteria;c__Fusobacteria Oyster_Firmicutes;Other Water_MVP 15;c__ Water_OP11;c__WCHB1 64 Oyster_TM7;Other Oyster_Proteobacteria;Other Water_Firmicutes;c__Clostridia Oyster_Chloroflexi;c__Anaerolineae VI Water_Verrucomicrobia;c__Methylacidiphilae Water_Verrucomicrobia;c__Verruco 5 Water_SM2F11;c__ Water_Acidobacteria;c__iii1 8 Water_GN02;c__GKS2 174 Water_TM6;c__SBRH58 Oyster_Chlorobi;c__SJA 28 Oyster_Chlorobi;c__BSV19 Oyster_Firmicutes;c__Bacilli Oyster_Verrucomicrobia;c__R76 B18 Oyster_Chloroflexi;c__Dehalococcoidetes Oyster_Acidobacteria;c__iii1 8 Oyster_TM6;c__SJA 4 Oyster_Gemmatimonadetes;c__Gemmatimonadetes Oyster_TM7;c__TM7 1 Oyster_Acidobacteria;c__OS K Oyster_Acidobacteria;c__Sva0725 Oyster_Bacteroidetes;c__Flavobacteria Oyster_Firmicutes;c__Clostridia VII Oyster_Bacteroidetes;Other Water_OP3;c__koll11 Oyster_Acidobacteria;c__Chloracidobacteria Water_Acidobacteria;c__Solibacteres Oyster_GN02;c__GKS2 174 Water_SR1;c__ Water_GN02;c__VC12 cl04 Water_GN02;c__BB34 Water_Spirochaetes;c__Spirochaetes Water_Chloroflexi;c__Dehalococcoidetes Water_TM7;c__TM7 3 Water_Tenericutes;c__Mollicutes Oyster_TM7;c__TM7 3 Water_Acidobacteria;c__RB25 Water_Elusimicrobia;c__Elusimicrobia Water_Proteobacteria;c__Epsilonproteobacteria VIII Water_Chloroflexi;Other Water_Tenericutes;c__ML615J 28 Water_ZB2;c__ Water_Proteobacteria;c__Zetaproteobacteria Oyster_Caldithrix;c__Caldithrixae Water_WS6;c__WM1006 Oyster_Bacteroidetes;c__Bacteroidia Water_TM7;c__TM7 1 Water_Proteobacteria;c__Alphaproteobacteria Oyster_Tenericutes;c__Erysipelotrichi Oyster_Acidobacteria;Other Water_Firmicutes;c__Bacilli IX Water_Bacteroidetes;c__Flavobacteria Water_Bacteroidetes;c__Bacteroidia Water_Proteobacteria;c__Gammaproteobacteria Oyster_Proteobacteria;c__Gammaproteobacteria Oyster_Lentisphaerae;c__Lentisphaerae Oyster_Proteobacteria;c__Epsilonproteobacteria Oyster_ZB2;c__ Water_Tenericutes;Other Water_TM6;c__SJA 4 Water_Cyanobacteria;c__4C0d 2 Water_Acidobacteria;c__Holophagae Oyster_Proteobacteria;c__Alphaproteobacteria Water_SC3;c__ Water_Acidobacteria;Other X Oyster_SAR406;c__AB16 y e e H O p D atur r Salinit empe T Fluorescenc Environmental Parameters Figure 2.10: Hierarchical clustering of bacterial class relative abundance correlations with measured water parameters and all shared taxa listed. Eighty-seven bacterial taxa (RDP-designated classes) were identified in both oyster EF and water samples and were compared. A correlation profile was created for each taxon from the R values of Spearman’s rank correlations between taxon relative abundance and each measured water parameter. Taxa were clustered by similar correlation profiles. Bacterial classes belonged to one of ten groups according to their correlation profile. Bacterial classes from EF and water that were in the same group are highlighted in red. EF bacterial classes are noted on the dendrogram in red, water in blue.

60 Color Key

0.5 0.5 Value Taxa Water_Acidobacteria;c__Sva0725 Water_Acidobacteria;c__ Water_Chlorobi;c__SJA 28 Water_OP3;c__BD4 9 Water_OP8;c__OP8_1 Water_Tenericutes;c__Erysipelotrichi Water;Other Water_Chlorobi;c__BSV19 Water_Cyanobacteria;c__Gloeobacterophycideae Oyster_Acidobacteria;c__RB25 Oyster_Acidobacteria;c__ Oyster_Acidobacteria;c__Solibacteres Oyster_OP8;c__OP8_1 Oyster_Proteobacteria;c__Zetaproteobacteria Oyster_SC3;c__ Oyster_TM6;c__SBRH58 Oyster_Verrucomicrobia;c__Verruco 5 Water_Firmicutes;Other Oyster_Chloroflexi;Other Water_TG3;c__TG3 2 Water_Chloroflexi;c__SOGA31 Water_Chloroflexi;c__Anaerolineae Water_Verrucomicrobia;c__Verrucomicrobiae Water_Acidobacteria;c__Chloracidobacteria Water_Caldithrix;c__Caldithrixae Oyster_WS6;c__WM1006 Oyster_OP3;c__koll11 Oyster_OP3;c__BD4 9 Water_Spirochaetes;c__Spirochaetes Water_Chloroflexi;c__Dehalococcoidetes Water_Tenericutes;c__Mollicutes Water_TM7;c__TM7 3 Water_GN02;c__BB34 Water_GN02;c__VC12 cl04 Water_OP3;c__koll11 Water_Acidobacteria;c__RB25 Oyster_Acidobacteria;c__Chloracidobacteria Water_Tenericutes;c__ML615J 28 Water_ZB2;c__ Water_Chloroflexi;Other Water_Proteobacteria;c__Zetaproteobacteria Oyster_Caldithrix;c__Caldithrixae Water_WS6;c__WM1006 Water_TM7;c__TM7 1 Oyster_Verrucomicrobia;c__R76 B18 Oyster_Chloroflexi;c__Dehalococcoidetes Water_Firmicutes;c__Bacilli Oyster_ZB2;c__ Oyster_Tenericutes;c__Erysipelotrichi Oyster_Acidobacteria;Other Oyster_Firmicutes;c__Bacilli Oyster_Firmicutes;c__Clostridia Oyster_TM7;Other Oyster_Gemmatimonadetes;c__Gemmatimonadetes Oyster_Proteobacteria;Other Oyster_Bacteroidetes;c__Flavobacteria Oyster_Bacteroidetes;c__Bacteroidia Oyster_Proteobacteria;c__Epsilonproteobacteria Oyster_Chlorobi;c__SJA 28 Oyster_Chlorobi;c__BSV19 Oyster_NKB19;Other Oyster_TM7;c__TM7 3 Oyster_Acidobacteria;c__OS K Oyster_Acidobacteria;c__Sva0725 Water_Elusimicrobia;c__Elusimicrobia Water_Acidobacteria;c__Solibacteres Water_Tenericutes;Other Oyster_GN02;c__GKS2 174 Oyster_Planctomycetes;c__agg27 Water_SR1;c__ Oyster_Acidobacteria;c__iii1 8 Oyster_TM6;c__SJA 4 Oyster_TM7;c__TM7 1 Water_Bacteroidetes;c__Bacteroidia Water_Proteobacteria;c__Epsilonproteobacteria Water_Verrucomicrobia;c__Methylacidiphilae Water_Verrucomicrobia;c__Verruco 5 Water_SM2F11;c__ Water_Acidobacteria;c__iii1 8 Water_TM6;c__SBRH58 Water_GN02;c__GKS2 174 Water_TM6;c__SJA 4 Oyster_Lentisphaerae;c__Lentisphaerae Water_Cyanobacteria;c__4C0d 2 Water_Acidobacteria;c__Holophagae Water_SC3;c__ Water_Acidobacteria;Other Oyster_SAR406;c__AB16 Oyster_Proteobacteria;c__Gammaproteobacteria Water_SAR406;c__AB16 Oyster_GN02;c__VC12 cl04 Water_Proteobacteria;c__Gammaproteobacteria Oyster_Bacteroidetes;Other See Table 2.7 Oyster_Chloroflexi;c__Anaerolineae Oyster_TG3;c__TG3 2 Oyster_Proteobacteria;c__Alphaproteobacteria Water_Verrucomicrobia;c__R76 B18 Water_GN02;Other Oyster_Chloroflexi;c__SOGA31 Water_Bacteroidetes;c__Sphingobacteria Water_Bacteroidetes;Other Water_Proteobacteria;c__Betaproteobacteria Water_Verrucomicrobia;Other Water_Cyanobacteria;Other Water_Cyanobacteria;c__Synechococcophycideae Oyster_Verrucomicrobia;c__Opitutae Water_Gemmatimonadetes;c__Gemmatimonadetes Water_Cyanobacteria;c__Oscillatoriophycideae Water_Chlorobi;c__Ignavibacteria Water_Verrucomicrobia;c__Spartobacteria Water_Proteobacteria;Other Water_Planctomycetes;c__Phycisphaerae Water_Chloroflexi;c__Chloroflexi Water_Bacteria;Other Water_Actinobacteria;c__Actinobacteria Oyster_Cyanobacteria;c__Oscillatoriophycideae Oyster_Cyanobacteria;c__S15B MN24 Oyster_Cyanobacteria;c__4C0d 2 Water_TM7;Other Oyster_Acidobacteria;c__Acidobacteria Water_WS3;c__PRR 12 Water_Lentisphaerae;c__Lentisphaerae Water_WS6;c__SC72 Oyster_Verrucomicrobia;Other Oyster_SM2F11;c__ Water_Verrucomicrobia;c__Opitutae Oyster_Actinobacteria;c__Actinobacteria Water_Acidobacteria;c__Acidobacteria Water_Planctomycetes;c__agg27 Oyster_OP8;c__OP8_2 Oyster_Fusobacteria;c__Fusobacteria Water_Bacteroidetes;c__Flavobacteria Oyster_Proteobacteria;c__Betaproteobacteria Water_Firmicutes;c__Clostridia Oyster_Verrucomicrobia;c__Verrucomicrobiae Oyster_Chloroflexi;c__Chloroflexi Water_Fusobacteria;c__Fusobacteria Oyster_Tenericutes;c__ML615J 28 Oyster_Firmicutes;Other Water_Nitrospirae;c__Nitrospira Oyster_Tenericutes;Other Oyster_Acidobacteria;c__Holophagae Oyster_MVP 15;c__ Oyster_WS3;c__PRR 12 Water_MVP 15;c__ Water_OP11;c__WCHB1 64 Oyster_Planctomycetes;c__Phycisphaerae Oyster_Verrucomicrobia;c__Spartobacteria Oyster_Cyanobacteria;c__Synechococcophycideae Oyster_OP11;c__WCHB1 64 Oyster_Elusimicrobia;c__Elusimicrobia Oyster_Spirochaetes;c__Spirochaetes Oyster_Bacteria;Other Water_Acidobacteria;c__OS K Oyster;Other Water_Chlorobi;c__OPB56 Oyster_GN02;c__BB34 Oyster_Chlorobi;c__Ignavibacteria Oyster_WS6;c__SC72 Water_Proteobacteria;c__Alphaproteobacteria Oyster_Nitrospirae;c__Nitrospira Oyster_Proteobacteria;c__Deltaproteobacteria Water_OP8;c__OP8_2 Water_NKB19;c__ Water_Cyanobacteria;c__S15B MN24 Oyster_Verrucomicrobia;c__Methylacidiphilae Oyster_Cyanobacteria;c__Gloeobacterophycideae Water_Proteobacteria;c__Deltaproteobacteria Oyster_Bacteroidetes;c__Sphingobacteria Oyster_Chlorobi;c__OPB56 Water_GN04;c__GN15 Oyster_GN04;c__GN15 Oyster_GN02;Other Oyster_Cyanobacteria;Other Oyster_SR1;c__ Oyster_Tenericutes;c__Mollicutes y e p H O p m D e T Salinit Fluorescenc Environmental Parameters Figure 2.11: Hierarchical clustering of bacterial class absolute abundance correlations with measured water parameters. Eighty-seven bacterial taxa (RDP- designated classes) were identified in both oyster EF and water samples and were compared. A correlation profile was created for each taxon from the R values of Spearman’s rank correlations between taxon relative abundance and each measured water parameter. Taxa were clustered by similar correlation profiles. Bacterial classes belonged to one of ten groups according to their correlation profile. Bacterial classes from EF and water that were in the same group are highlighted in red. EF bacterial classes are noted on the dendrogram in red, water in blue.

61 A. 1.00 * *

0.75 *

0.50 Oyster Water 0.25

) 0.00 value B. 1.00 * * *

(absolute 0.75 * Oyster No Lag 0.50 value

Oyster With Lag R 0.25

0.00 Rank

Spearman 1.00 * 0.75

0.50 Oyster With Lag

0.25 Water No Lag

0.00 Temperature DO Salinity pH Fluorescence Figure 2.12: Strength of Spearman’s rank correlations (R values) for bacterial classes in oyster EF or water samples with measured environmental parameters. A) Comparison of the R values between EF and water taxa for each measured parameter. B) Comparison of the R values between EF taxa before and after introducing a one- month lag. C) Comparison of the R values between EF taxa with a one-month lag and water taxa without a lag.

62 TABLES

Table 2.1: Correlations between total bacterial abundance and measured water parameters.

Bacterial Abundance Oyster Association Water Association

Temperature (°C) p <0.05* + p <0.05 + Conductivity (mS/cm) NS NS Salinity (PPT) NS NS Dissolved Oxygen (mg/L) p <0.05 - p <0.05 - pH p <0.01 - NS Chlorophyll (μg/L) NS NS Turbidity+ (NTU) NS NS

*After introducing a one-month lag behind measured water parameters.

63 Table 2.2: Bacterial classes with significantly different relative and absolute abundances over the annual study.

Mean Mean Mean Mean Proportion Proportion Taxonomy (Phylum) Taxonomy (Class) Oyster:Water Significance Abundance in Abundance in Oyster:Water Significance Oyster Water Oysters Water Community Community Acidobacteria Holophagae 0.10% <0.01% 53.53 *** 1.16E+04 4.01E+01 289.86 *** Unclassified Bacteria Other 17.21% 4.23% 4.07 *** 2.20E+06 2.70E+05 8.13 *** Bacteroidia 3.40% 0.27% 12.76 *** 3.43E+05 4.99E+03 68.65 *** Bacteroidetes Flavobacteria 5.41% 17.04% 0.32 *** 5.05E+05 6.26E+05 0.81 * Sphingobacteria 1.26% 9.11% 0.14 *** 1.35E+05 4.10E+05 0.33 *** Chloroflexi Other 0.01% 0.05% 0.16 ** 7.68E+02 1.15E+03 0.67 * Oscillatoriophycideae 0.01% 0.05% 0.12 * 7.64E+02 3.19E+03 0.24 NS Cyanobacteria Other 0.06% 0.53% 0.12 *** 1.05E+04 2.79E+04 0.37 *** Bacilli 1.75% 0.05% 38.42 ** 1.97E+05 1.64E+03 120.56 *** Firmicutes Clostridia 1.21% 0.14% 8.88 ** 1.27E+05 4.68E+03 27.13 *** Fusobacteria Fusobacteria 0.65% 0.02% 26.91 ** 3.96E+04 1.29E+03 30.71 *** GN02 Other 0.12% 0.02% 5.51 ** 1.18E+04 1.41E+03 8.38 ** Nitrospirae Nitrospira 0.21% 0.01% 23.78 *** 2.14E+04 3.78E+02 56.60 *** OP8 OP8_2 0.15% <0.01% 293.21 * 7.61E+03 2.11E+01 360.87 * Betaproteobacteria 5.00% 8.09% 0.62 ** 4.60E+05 3.57E+05 1.29 NS Deltaproteobacteria 2.59% 1.09% 2.38 ** 2.10E+05 5.89E+04 3.57 *** Proteobacteria Epsilonproteobacteria 3.87% 0.23% 17.03 *** 2.35E+05 6.25E+03 37.57 *** Gammaproteobacteria 11.23% 6.39% 1.76 ** 1.21E+06 2.35E+05 5.15 *** Other 0.53% 0.15% 3.63 ** 5.75E+04 7.69E+03 7.47 ** Tenericutes Other <0.01% 0.45% 0.01 *** 1.85E+02 1.00E+04 0.02 *** TM7 Other 0.09% 0.01% 15.42 *** 1.02E+04 3.03E+02 33.56 *** Verrucomicrobia Verrucomicrobiae 0.02% 0.05% 0.36 ** 2.46E+03 3.29E+03 0.75 NS WS6 Other 0.02% 0.00% * 1.48E+03 0 * p<0.05 ** p<0.01 *** p<0.001

Table 2.3: Oyster EF bacterial families ranked by mean abundance over one annual cycle.

Rank Taxonomy Mean SE 1 Bacteria;Other 17.21% 3.40% 2 Phyllobacteriaceae (Alphaproteobacteria) 11.79% 3.87% 3 Rickettsiales;Other (Alphaproteobacteria) 9.54% 1.51% 4 Synechococcaceae 4.70% 1.56% (Synechococcophycideae) 5 Flavobacteriaceae (Flavobacteria) 3.81% 0.58% 6 Campylobacteraceae 3.45% 1.41% (Epsilonproteobacteria) 7 Microbacteriaceae (Actinobacteria) 2.80% 0.55% 8 Vibrionaceae (Gammaproteobacteria) 2.61% 0.77% 9 Rhodobacteraceae (Alphaproteobacteria) 2.61% 0.53% 10 CL500-29 (Actinobacteria) 2.56% 0.47% 11 Bacteroidetes;Other 1.70% 0.23% 12 Rickettsiaceae (Alphaproteobacteria) 1.70% 0.82%

64 13 Methylophilaceae (Betaproteobacteria) 1.68% 0.27% 14 Porphyromonadaceae (Bacteroidia) 1.54% 0.67% 15 Bacteroidales;Other (Bacteroidia) 1.44% 0.74% 16 Erythrobacteraceae (Alphaproteobacteria) 1.37% 0.37% 17 Halomonadaceae (Gammaproteobacteria) 1.35% 0.35% 18 Comamonadaceae (Betaproteobacteria) 1.33% 0.23% 19 ACK-M1 (Actinobacteria) 1.22% 0.26% 20 Desulfobacteraceae (Deltaproteobacteria) 1.02% 0.21% 21 Lactobacillaceae (Bacilli) 1.01% 0.97% 22 Shewanellaceae (Gammaproteobacteria) 0.95% 0.38% 23 Alteromonadaceae (Gammaproteobacteria) 0.84% 0.23% 24 Alphaproteobacteria;Other; 0.81% 0.16% 25 Flavobacteria;Other 0.79% 0.27% 26 Desulfobulbaceae (Deltaproteobacteria) 0.67% 0.12% 27 Fusobacteriaceae (Fusobacteria) 0.65% 0.50% 28 Gammaproteobacteria;Other 0.61% 0.11% 29 Clostridiales;Other (Clostridia) 0.59% 0.27% 30 Pseudoalteromonadaceae 0.57% 0.12% (Gammaproteobacteria) 31 Proteobacteria;Other; 0.53% 0.10% 32 Oceanospirillales;Other 0.49% 0.08% (Gammaproteobacteria) 33 Psychromonadaceae (Gammaproteobacteria) 0.47% 0.17% 34 Burkholderiales;Other (Betaproteobacteria) 0.46% 0.21% 35 Flavobacteria;Other 0.45% 0.10% 36 Francisellaceae (Gammaproteobacteria) 0.44% 0.10% 37 Rhodospirillaceae (Alphaproteobacteria) 0.42% 0.10% 38 Burkholderiaceae (Betaproteobacteria) 0.39% 0.11% 39 Betaproteobacteria;Other 0.38% 0.08% 40 Helicobacteraceae (Epsilonproteobacteria) 0.35% 0.15% 41 Idiomarinaceae (Gammaproteobacteria) 0.34% 0.10% 42 Cryomorphaceae (Flavobacteria) 0.32% 0.08% 43 Saprospiraceae (Sphingobacteria) 0.32% 0.09% 44 Actinomycetales;Other (Actinobacteria) 0.32% 0.08% 45 Bacillaceae (Bacilli) 0.30% 0.19% 46 Aeromonadaceae (Gammaproteobacteria) 0.30% 0.14% 47 Desulfovibrionaceae (Deltaproteobacteria) 0.30% 0.09% 48 Micrococcaceae (Actinobacteria) 0.27% 0.25% 49 Clostridiales Family XI. Incertae Sedis 0.27% 0.27% (Clostridia) 50 OM60 (Gammaproteobacteria) 0.26% 0.06% 51 Rhodocyclaceae (Betaproteobacteria) 0.25% 0.08%

65 52 Corynebacteriaceae (Actinobacteria) 0.25% 0.22% 53 Oceanospirillaceae (Gammaproteobacteria) 0.25% 0.07% 54 HTCC2188 (Gammaproteobacteria) 0.25% 0.07% 55 Sinobacteraceae (Gammaproteobacteria) 0.24% 0.07% 56 Sphingobacteriales;Other (Sphingobacteria) 0.24% 0.08% 57 Actinomycetales;Other (Actinobacteria) 0.24% 0.08% 58 Clostridiales Family XII. Incertae Sedis 0.23% 0.09% (Clostridia) 59 Flammeovirgaceae (Sphingobacteria) 0.22% 0.04% 60 Prevotellaceae (Bacteroidia) 0.22% 0.21% 61 Nitrospiraceae (Nitrospira) 0.20% 0.05% 62 Cyclobacteriaceae (Sphingobacteria) 0.20% 0.08% 63 OPB56;Other 0.20% 0.07% 64 ZA3412c (Gammaproteobacteria) 0.20% 0.06% 65 Burkholderiales;Other (Betaproteobacteria) 0.18% 0.07% 66 Flexibacteraceae (Sphingobacteria) 0.17% 0.06% 67 Rhizobiaceae (Alphaproteobacteria) 0.17% 0.08% 68 OP8_2;Other 0.15% 0.11% 69 VC12-cl04;Other 0.15% 0.04% 70 Rhizobiales;Other (Alphaproteobacteria) 0.14% 0.03% 71 Aerococcaceae (Bacilli) 0.14% 0.14% 72 Other 0.13% 0.03% 73 Bacteriovoracaceae (Deltaproteobacteria) 0.12% 0.02% 74 GN02;Other 0.12% 0.02% 75 Gemellaceae (Bacilli) 0.12% 0.12% 76 Xanthomonadaceae (Gammaproteobacteria) 0.12% 0.03% 77 ZB2;Other 0.11% 0.02% 78 Bacteroidaceae (Bacteroidia) 0.11% 0.08% 79 Deltaproteobacteria;Other 0.11% 0.03% 80 Kiloniellaceae (Alphaproteobacteria) 0.10% 0.02% 81 Staphylococcaceae (Bacilli) 0.10% 0.07% 82 Legionellaceae (Gammaproteobacteria) 0.10% 0.05% 83 Rhodobacteraceae (Alphaproteobacteria) 0.10% 0.03% 84 Alphaproteobacteria;Other 0.10% 0.02% 85 Alcaligenaceae (Betaproteobacteria) 0.10% 0.03% 86 Pseudomonadaceae (Gammaproteobacteria) 0.09% 0.03% 87 Chromatiales;Other (Gammaproteobacteria) 0.09% 0.02% 88 TM7;Other 0.09% 0.02% 89 Nitrospinaceae (Deltaproteobacteria) 0.08% 0.02% 90 Oxalobacteraceae (Betaproteobacteria) 0.08% 0.02% 91 Chromatiaceae (Gammaproteobacteria) 0.08% 0.03% 92 Gammaproteobacteria;Other 0.08% 0.02%

66 93 Litincolaceae (Gammaproteobacteria) 0.08% 0.02% 94 Sphingomonadaceae (Alphaproteobacteria) 0.07% 0.02% 95 MIZ46;Other (Deltaproteobacteria) 0.07% 0.04% 96 Enterobacteriaceae (Gammaproteobacteria) 0.07% 0.04% 97 TM7-1;Other 0.06% 0.02% 98 Marinilabiaceae (Bacteroidia) 0.06% 0.03% 99 SOGA31;Other 0.06% 0.03% 100 Cyanobacteria;Other 0.06% 0.02% 101 Desulfuromonadales;Other 0.06% 0.02% (Deltaproteobacteria) 102 Phycisphaerales;Other (Phycisphaerae) 0.06% 0.03% 103 Holophagales;Other (Holophagae) 0.06% 0.02% 104 SC72;Other 0.06% 0.03% 105 Puniceicoccaceae (Opitutae) 0.06% 0.03% 106 Balneolaceae (Sphingobacteria) 0.05% 0.01% 107 Acetobacteraceae (Alphaproteobacteria) 0.05% 0.02% 108 Deltaproteobacteria;Other 0.05% 0.01% 109 ZB1;Other (Ignavibacteria) 0.05% 0.02% 110 Kordiimonadaceae (Alphaproteobacteria) 0.05% 0.01% 111 Veillonellaceae (Clostridia) 0.04% 0.04% 112 Nitrosomonadaceae (Betaproteobacteria) 0.04% 0.01% 113 SJA-4;Other 0.04% 0.01% 114 Moraxellaceae (Gammaproteobacteria) 0.04% 0.02% 115 Alteromonadales;Other 0.04% 0.01% (Gammaproteobacteria) 116 Flavobacteriales;Other (Flavobacteria) 0.04% 0.01% 117 Holophagaceae (Holophagae) 0.04% 0.01% 118 Coriobacteriaceae (Actinobacteria) 0.04% 0.04% 119 Colwelliaceae (Gammaproteobacteria) 0.04% 0.01% 120 Sphingobacteriales;Other (Sphingobacteria) 0.04% 0.01% 121 Gemmatimonadetes;Other 0.04% 0.01% 122 Rhodocyclales;Other (Betaproteobacteria) 0.04% 0.01% 123 Chlorobiaceae (Chlorobia) 0.04% 0.02% 124 ML615J-28;Other 0.03% 0.02% 125 Mycoplasmataceae (Mollicutes) 0.03% 0.03% 126 Epsilonproteobacteria;Other 0.03% 0.02% 127 Roseiflexales;Other (Chloroflexi) 0.03% 0.01% 128 Saccharospirillaceae 0.03% 0.01% (Gammaproteobacteria) 129 Spirochaetes;Other; 0.03% 0.03% 130 Bacteroidales;Other (Bacteroidia) 0.03% 0.01% 131 Peptostreptococcaceae (Clostridia) 0.03% 0.03%

67 132 Caulobacteraceae (Alphaproteobacteria) 0.03% 0.01% 133 S15B-MN24;Other 0.03% 0.01% 134 Chromatiales;Other (Gammaproteobacteria) 0.03% 0.01% 135 125ds10 (Gammaproteobacteria) 0.03% 0.01% 136 Campylobacterales;Other 0.03% 0.01% (Epsilonproteobacteria) 137 Methylobacteriaceae (Alphaproteobacteria) 0.02% 0.01% 138 Spirochaetaceae (Spirochaetes) 0.02% 0.01% 139 Propionibacteriaceae (Actinobacteria) 0.02% 0.01% 140 Moritellaceae (Gammaproteobacteria) 0.02% 0.02% 141 Spartobacteriaceae (Spartobacteria) 0.02% 0.01% 142 Actinobacteria;Other 0.02% 0.01% 143 WS6;Other 0.02% 0.02% 144 Lactobacillales;Other (Bacilli) 0.02% 0.02% 145 Sphingomonadales;Other 0.02% 0.01% (Alphaproteobacteria) 146 DS-18;Other (iii1-8) 0.02% 0.01% 147 Alteromonadaceae (Gammaproteobacteria) 0.02% 0.01% 148 Desulfuromonadaceae (Deltaproteobacteria) 0.02% 0.02% 149 Paenibacillaceae (Bacilli) 0.02% 0.01% 150 Chlamydiales;Other (Chlamydiae) 0.02% 0.02% 151 Hyphomonadaceae (Alphaproteobacteria) 0.02% 0.01% 152 BB34;Other 0.02% 0.01% 153 Acidimicrobiales;Other (Actinobacteria) 0.02% 0.01% 154 Hydrogenophilaceae (Betaproteobacteria) 0.02% 0.01% 155 Campylobacterales;Other 0.02% 0.01% (Epsilonproteobacteria) 156 Streptococcaceae (Bacilli) 0.02% 0.01% 157 Rhodocyclales;Other (Betaproteobacteria) 0.02% 0.01% 158 Acidobacteriales;Other (Acidobacteria) 0.02% 0.01% 159 Coxiellaceae (Gammaproteobacteria) 0.02% 0.00% 160 Clostridiaceae (Clostridia) 0.01% 0.01% 161 koll13;Other (Actinobacteria) 0.01% 0.00% 162 TM7;Other 0.01% 0.01% 163 Bradyrhizobiaceae (Alphaproteobacteria) 0.01% 0.01% 164 NKB15;Other (Deltaproteobacteria) 0.01% 0.01% 165 Rhizobiales;Other (Alphaproteobacteria) 0.01% 0.00% 166 TG3-2;Other 0.01% 0.01% 167 Ignavibacteria;Other 0.01% 0.00% 168 MVP-15;Other 0.01% 0.01% 169 SR1;Other 0.01% 0.00% 170 GKS2-174;Other 0.01% 0.00%

68 171 Neisseriaceae (Betaproteobacteria) 0.01% 0.00% 172 CTD005-82B-02;Other 0.01% 0.00% (Deltaproteobacteria) 173 GN07;Other 0.01% 0.01% 174 Pasteurellaceae (Gammaproteobacteria) 0.01% 0.01% 175 Opitutaceae (Opitutae) 0.01% 0.00% 176 Clostridiales;Other (Clostridia) 0.01% 0.01% 177 Enterococcaceae (Bacilli) 0.01% 0.01% 178 Piscirickettsiaceae (Gammaproteobacteria) 0.01% 0.00% 179 WCHB1-64;Other 0.01% 0.00% 180 211ds20 (Gammaproteobacteria) 0.01% 0.00% 181 Bdellovibrionaceae (Deltaproteobacteria) 0.01% 0.00% 182 Acholeplasmataceae (Mollicutes) 0.01% 0.00% 183 Actinomycetaceae (Actinobacteria) 0.01% 0.01% 184 OS-K;Other 0.01% 0.01% 185 Syntrophobacteraceae (Deltaproteobacteria) 0.01% 0.00% 186 Verrucomicrobiales;Other 0.01% 0.00% (Verrucomicrobiae) 187 ABY1_OD1;Other 0.01% 0.00% 188 Brevibacteriaceae (Actinobacteria) 0.01% 0.00% 189 Syntrophobacterales;Other 0.01% 0.00% (Deltaproteobacteria) 190 Dehalococcoidaceae (Dehalococcoidetes) 0.01% 0.00% 191 Sphingobacteriaceae (Sphingobacteria) 0.01% 0.00% 192 Euzebiaceae (Actinobacteria) 0.01% 0.00% 193 Lachnospiraceae (Clostridia) 0.01% 0.00% 194 Chloroflexi;Other 0.01% 0.01% 195 HTCC2089 (Gammaproteobacteria) 0.01% 0.00% 196 Solirubrobacterales;Other (Actinobacteria) 0.01% 0.01% 197 Microthrixaceae (Actinobacteria) 0.01% 0.00% 198 Verrucomicrobiaceae (Verrucomicrobiae) 0.01% 0.00% 199 Thermodesulfovibrionaceae (Nitrospira) 0.01% 0.00% 200 Methylococcaceae (Gammaproteobacteria) 0.01% 0.00% 201 TM7-3;Other 0.01% 0.01% 202 Nannocystaceae (Deltaproteobacteria) 0.01% 0.00% 203 SM2F11;Other 0.01% 0.00% 204 Gemmatimonadaceae (Gemmatimonadetes) 0.01% 0.00% 205 Syntrophaceae (Deltaproteobacteria) 0.01% 0.00% 206 GN03;Other (PRR-12) 0.01% 0.00% 207 Legionellales;Other (Gammaproteobacteria) 0.01% 0.00% 208 SC72;Other 0.01% 0.00% 209 Clostridia;Other 0.01% 0.00%

69 210 Myxococcales;Other (Deltaproteobacteria) 0.01% 0.00% 211 K2-4-19;Other (WCHB1-64) 0.01% 0.00% 212 Thiotrichales;Other (Gammaproteobacteria) 0.01% 0.00% 213 Rhodobacterales;Other 0.01% 0.00% (Alphaproteobacteria) 214 Carnobacteriaceae (Bacilli) 0.01% 0.01% 215 Sva0725;Other (Sva0725) 0.01% 0.00% 216 EW055;Other (TM7-3) 0.01% 0.00% 217 Clostridiales Family XIII. Incertae Sedis 0.01% 0.00% (Clostridia) 218 Ruminococcaceae (Clostridia) 0.01% 0.00% 219 Actinobacteria;Other 0.01% 0.00% 220 Desulfobacterales;Other 0.01% 0.00% (Deltaproteobacteria) 221 Phycisphaeraceae (Phycisphaerae) 0.00% 0.00% 222 Desulfomicrobiaceae (Deltaproteobacteria) 0.005% 0.004% 223 Rs-045 (TM7-3) 0.005% 0.003% 224 Gemmatimonadales;Other 0.005% 0.002% (Gemmatimonadetes) 225 Chroococcales;Other 0.005% 0.003% (Oscillatoriophycideae) 226 Pseudomonadales;Other 0.005% 0.002% (Gammaproteobacteria) 227 Hyphomicrobiaceae (Alphaproteobacteria) 0.005% 0.002% 228 Gallionellaceae (Betaproteobacteria) 0.004% 0.002% 229 Alcanivoracaceae (Gammaproteobacteria) 0.004% 0.003% 230 Rhodobacterales;Other 0.004% 0.002% (Alphaproteobacteria) 231 R76-B18;Other 0.004% 0.002% 232 Sphingomonadales;Other 0.004% 0.002% (Alphaproteobacteria) 233 SJA-28;Other 0.004% 0.003% 234 GN08;Other 0.004% 0.002% 235 Legionellales;Other (Gammaproteobacteria) 0.004% 0.002% 236 Brucellaceae (Alphaproteobacteria) 0.004% 0.004% 237 Erysipelotrichaceae (Erysipelotrichi) 0.004% 0.003% 238 Chlorobi;Other 0.004% 0.002% 239 Betaproteobacteria;Other 0.004% 0.002% 240 GN02;Other 0.004% 0.003% 241 RB384;Other 0.004% 0.002% 242 OP8_1;Other 0.003% 0.003% 243 Methylocystaceae (Alphaproteobacteria) 0.003% 0.002%

70 244 Caldilineaceae (Anaerolineae) 0.003% 0.003% 245 Pseudanabaenaceae 0.003% 0.003% (Synechococcophycideae) 246 Rhodothermaceae (Sphingobacteria) 0.003% 0.002% 247 Pelobacteraceae (Deltaproteobacteria) 0.003% 0.002% 248 Myxococcales;Other (Deltaproteobacteria) 0.003% 0.002% 249 Rickettsiales;Other (Alphaproteobacteria) 0.003% 0.001% 250 Anaerolineae;Other 0.003% 0.002% 251 Elusimicrobiales;Other (Elusimicrobia) 0.003% 0.002% 252 Cohaesibacteraceae (Alphaproteobacteria) 0.003% 0.002% 253 Crenotrichaceae (Gammaproteobacteria) 0.003% 0.001% 254 Verrucomicrobia;Other 0.003% 0.001% 255 Acidobacteria;c__Chloracidobacteria;Other 0.003% 0.002% 256 HDBW-WB69;Other 0.003% 0.002% 257 Desulfovibrionales;Other 0.002% 0.001% (Deltaproteobacteria) 258 Tenericutes;Other 0.002% 0.001% 259 WCHB1-64;Other 0.002% 0.002% 260 envOPS12;Other (Anaerolineae) 0.002% 0.001% 261 A4b;Other (Anaerolineae) 0.002% 0.002% 262 OP11;Other 0.002% 0.001% 263 Phycisphaerales;Other (Phycisphaerae) 0.002% 0.002% 264 Methylococcales;Other 0.002% 0.002% (Gammaproteobacteria) 265 Dehalococcoidetes;Other 0.002% 0.002% 266 Bacillales;Other (Bacilli) 0.002% 0.001% 267 Proteobacteria;Other 0.002% 0.002% 268 Oleiphilaceae (Gammaproteobacteria) 0.002% 0.001% 269 PW285;Other 0.002% 0.002% 270 PRR-12;Other 0.002% 0.001% 271 Chromatiaceae (Gammaproteobacteria) 0.002% 0.001% 272 WCHB1-07;Other (OP11-2) 0.002% 0.002% 273 TG3-2;Other 0.002% 0.001% 274 Sinobacteraceae (Gammaproteobacteria) 0.002% 0.002% 275 Ignavibacteriaceae (Ignavibacteria) 0.002% 0.001% 276 Verrucomicrobiales;Other 0.002% 0.001% (Verrucomicrobiae) 277 WM1006;Other 0.002% 0.002% 278 BSV19;Other 0.002% 0.002% 279 Acidobacteria;Other 0.002% 0.001% 280 Beijerinckiaceae (Alphaproteobacteria) 0.002% 0.001% 281 Coriobacteriales;Other (Actinobacteria) 0.002% 0.001%

71 282 Bifidobacteriaceae (Actinobacteria) 0.002% 0.002% 283 OP11-2;Other 0.002% 0.001% 284 NT-B4;Other (Dehalococcoidetes) 0.002% 0.001% 285 Methylacidiphilales;f__LD19 0.001% 0.001% (Methylacidiphilae) 286 Phycisphaerae;Other 0.001% 0.001% 287 S0208;Other (Anaerolineae) 0.001% 0.001% 288 Mycobacteriaceae (Actinobacteria) 0.001% 0.001% 289 Firmicutes;Other 0.001% 0.001% 290 PBS-III-9;Other (PRR-12) 0.001% 0.001% 291 YS2;Other (4C0d-2) 0.001% 0.001% 292 Elusimicrobiaceae (Elusimicrobia) 0.001% 0.001% 293 Thermoanaerobacteraceae (Clostridia) 0.001% 0.001% 294 Kordiimonadales;Other 0.001% 0.001% (Alphaproteobacteria) 295 Gloeobacteraceae (Gloeobacterophycideae) 0.001% 0.001% 296 Desulfuromonadales;Other 0.001% 0.001% (Deltaproteobacteria) 297 HOC21 (Gammaproteobacteria) 0.001% 0.001% 298 SHA-20;Other (Anaerolineae) 0.001% 0.001% 299 VHS-B5-50;Other 0.001% 0.001% 300 Synechococcophycideae;Other 0.001% 0.001% 301 Streptomycetaceae (Actinobacteria) 0.001% 0.001% 302 Catabacteriaceae (Clostridia) 0.001% 0.001% 303 NKB19;Other 0.001% 0.001% 304 kpj58rc (koll11) 0.001% 0.001% 305 SUP05 (Gammaproteobacteria) 0.001% 0.001% 306 JdFBGBact (Actinobacteria) 0.001% 0.001% 307 Gomphosphaeriaceae 0.001% 0.001% (Oscillatoriophycideae) 308 SSW63Au;Other (AB16) 0.001% 0.001% 309 Caldithrixales;Other (Caldithrixae) 0.001% 0.001% 310 Victivallales;Other (Lentisphaerae) 0.001% 0.001% 311 Anaerolineae;Other 0.001% 0.001% 312 mle1-12;Other (4C0d-2) 0.001% 0.001% 313 Simkaniaceae (Chlamydiae) 0.001% 0.001% 314 MC47;Other (Actinobacteria) 0.001% 0.001% 315 Rubrobacteraceae (Actinobacteria) 0.001% 0.001% 316 JTB38 (Deltaproteobacteria) 0.001% 0.001% 317 HMMVPog-54;Other (OP8_1) 0.001% 0.001% 318 Syntrophomonadaceae (Clostridia) 0.001% 0.001% 319 CH21;Other 0.001% 0.001%

72 320 MVP-88;Other (Elusimicrobia) 0.001% 0.000% 321 Hahellaceae (Gammaproteobacteria) 0.001% 0.001% 322 A29 (TM7-3) 0.001% 0.001% 323 BD4-9;Other 0.001% 0.000% 324 Piscirickettsiaceae (Gammaproteobacteria) 0.001% 0.000% 325 GIF10;Other (koll11) 0.001% 0.000% 326 Thiotrichales;Other (Gammaproteobacteria) 0.001% 0.000% 327 Bdellovibrionales;Other 0.001% 0.000% (Deltaproteobacteria) 328 Syntrophorhabdaceae (Deltaproteobacteria) 0.0005% 0.0005% 329 Endoecteinascidiaceae 0.0005% 0.0003% (Gammaproteobacteria) 330 Nocardiaceae (Actinobacteria) 0.0005% 0.0005% 331 CL500-15;Other (agg27) 0.0005% 0.0003% 332 Sphaerochaetaceae (Spirochaetes) 0.0004% 0.0003% 333 Bacteroidetes;Other 0.0004% 0.0003% 334 RCP1-27;Other 0.0004% 0.0003% 335 Xanthobacteraceae (Alphaproteobacteria) 0.0004% 0.0004% 336 NB1-i (Deltaproteobacteria) 0.0004% 0.0004% 337 SHAS460 (AB16) 0.0004% 0.0004% 338 ntu14 (Actinobacteria) 0.0003% 0.0003% 339 Endomicrobia;Other 0.0003% 0.0003% 340 KD3-145;Other 0.0003% 0.0003% 341 Enterobacteriales;Other 0.0003% 0.0003% (Gammaproteobacteria) 342 Leptospiraceae (Leptospirae) 0.0003% 0.0003% 343 Acidobacteria;Other 0.0003% 0.0003% 344 Desulfitobacteraceae (Clostridia) 0.0003% 0.0003% 345 Dehalobacteriaceae (Clostridia) 0.0003% 0.0003% 346 GN13;Other 0.0003% 0.0003% 347 Oceanospirillales;Other 0.0003% 0.0003% (Gammaproteobacteria) 348 A714017 (AB16) 0.0003% 0.0003% 349 OP11-3;Other 0.0003% 0.0002% 350 mle1-48;Other (TK17) 0.0003% 0.0002% 351 Synergistales;Other (Synergistia) 0.0003% 0.0003% 352 A-2AF;Other (SC72) 0.0003% 0.0003% 353 Solirubrobacterales;Other (Actinobacteria) 0.0003% 0.0003% 354 GCA004;Other (Anaerolineae) 0.0003% 0.0003% 355 SHA-124;Other (OP8_1) 0.0003% 0.0003% 356 SBRH58;Other 0.0003% 0.0003% 357 Aurantimonadaceae (Alphaproteobacteria) 0.0002% 0.0002%

73 358 Spirochaetes;Other 0.0002% 0.0002% 359 Other 0.0002% 0.0002% 360 Pseudomonadales;Other 0.0002% 0.0002% (Gammaproteobacteria) 361 GN09;Other (TG3-1) 0.0002% 0.0002% 362 SJA-101;Other (Anaerolineae) 0.0002% 0.0002% 363 Anaerolinaceae (Anaerolineae) 0.0002% 0.0002% 364 Polyangiaceae (Deltaproteobacteria) 0.0002% 0.0002% 365 Desulfobacterales;Other 0.0002% 0.0002% (Deltaproteobacteria) 366 Methylophilales;Other (Betaproteobacteria) 0.0002% 0.0002% 367 GN15;Other 0.0002% 0.0002% 368 Entotheonellaceae (Deltaproteobacteria) 0.0002% 0.0002% 369 Nocardioidaceae (Actinobacteria) 0.0002% 0.0002% 370 SC3;Other 0.0002% 0.0002% 371 GN05;Other 0.0002% 0.0002% 372 RB25;Other 0.0002% 0.0002% 373 Gemmatimonadetes;Other 0.0002% 0.0002% 374 Gemmatimonadales;Other 0.0002% 0.0002% (Gemmatimonadetes) 375 WPS-2;Other 0.0002% 0.0002% 376 Solibacteraceae (Solibacteres) 0.0001% 0.0001% 377 Armatimonadetes;Other 0.0001% 0.0001% 378 SHA-31 (Anaerolineae) 0.0001% 0.0001% 379 Mariprofundaceae (Zetaproteobacteria) 0.0001% 0.0001% 380 CV106;Other (PRR-12) 0.0001% 0.0001% 381 wb1_P06 (Actinobacteria) 0.0001% 0.0001% 382 Chloroflexaceae (Chloroflexi) 0.0001% 0.0001% 383 SM1D11;Other (4C0d-2) 0.0001% 0.0001% 384 Listeriaceae (Bacilli) 0.0001% 0.0001% 385 KSB1;Other 0.0001% 0.0001% 386 OP11-1;Other 0.0001% 0.0001% 387 GIF10;Other (koll11) 0.0001% 0.0001% 388 OM190;Other (agg27) 0.0001% 0.0001% 389 Methylophilales;Other (Betaproteobacteria) 0.0001% 0.0001% 390 Myxococcaceae (Deltaproteobacteria) 0.0001% 0.0001% 391 Ectothiorhodospiraceae 0.0001% 0.0001% (Gammaproteobacteria) 392 SPAM;Other 0.0001% 0.0001% 393 0319-6G9;Other 0.0001% 0.0001% 394 TM6;Other 0.0001% 0.0001% 395 Verruco-5;Other 0.0001% 0.0001%

74 396 PRR-10 (PRR-12) 0.0001% 0.0001%

Table 2.4: Water bacterial families ranked by mean abundance over one annual cycle.

Rank Taxonomy Mean SE 1 Rickettsiales;Other (Alphaproteobacteria) 22.82% 3.44% 2 Flavobacteriaceae (Flavobacteria) 11.61% 2.79% 3 Rhodobacteraceae (Alphaproteobacteria) 7.41% 1.35% 4 Synechococcaceae 7.08% 2.66% (Synechococcophycideae) 5 Saprospiraceae (Sphingobacteria) 4.33% 1.03% 6 Bacteria;Other 4.23% 0.60% 7 Microbacteriaceae (Actinobacteria) 4.16% 0.55% 8 Cyclobacteriaceae (Sphingobacteria) 3.27% 1.29% 9 Flavobacteria;Other 3.19% 0.85% 10 ACK-M1 (Actinobacteria) 3.01% 0.53% 11 Comamonadaceae (Betaproteobacteria) 2.62% 0.35% 12 Methylophilaceae (Betaproteobacteria) 2.54% 0.30% 13 Bacteroidetes;Other 1.83% 0.48% 14 CL500-29 (Actinobacteria) 1.23% 0.31% 15 ZA3412c (Gammaproteobacteria) 1.22% 0.54% 16 HTCC2188 (Gammaproteobacteria) 1.19% 0.35% 17 Cryomorphaceae (Flavobacteria) 1.19% 0.24% 18 Sphingobacteriales;Other (Sphingobacteria) 0.94% 0.44% 19 Oceanospirillales;Other 0.93% 0.21% (Gammaproteobacteria) 20 Flavobacteria;Other 0.92% 0.27% 21 Oxalobacteraceae (Betaproteobacteria) 0.86% 0.45% 22 Burkholderiaceae (Betaproteobacteria) 0.71% 0.14% 23 Alphaproteobacteria;Other 0.64% 0.17% 24 Deltaproteobacteria;Other 0.58% 0.21% 25 Cyanobacteria;Other 0.53% 0.15% 26 Burkholderiales;Other (Betaproteobacteria) 0.50% 0.17% 27 OM60 (Gammaproteobacteria) 0.48% 0.10% 28 Tenericutes;Other 0.45% 0.30% 29 Alcaligenaceae (Betaproteobacteria) 0.42% 0.21% 30 Halomonadaceae (Gammaproteobacteria) 0.41% 0.19% 31 Actinomycetales;Other (Actinobacteria) 0.40% 0.05%

75 32 Roseiflexales;Other (Chloroflexi) 0.39% 0.21% 33 Rhodospirillaceae (Alphaproteobacteria) 0.36% 0.13% 34 Gammaproteobacteria;Other 0.33% 0.06% 35 Sinobacteraceae (Gammaproteobacteria) 0.32% 0.09% 36 OPB56;Other 0.29% 0.14% 37 Saccharospirillaceae (Gammaproteobacteria) 0.26% 0.16% 38 Rickettsiaceae (Alphaproteobacteria) 0.26% 0.06% 39 Balneolaceae (Sphingobacteria) 0.20% 0.07% 40 Betaproteobacteria;Other 0.20% 0.04% 41 Burkholderiales;Other (Betaproteobacteria) 0.20% 0.10% 42 Idiomarinaceae (Gammaproteobacteria) 0.19% 0.13% 43 Flexibacteraceae (Sphingobacteria) 0.19% 0.06% 44 Xanthomonadaceae (Gammaproteobacteria) 0.16% 0.06% 45 MIZ46;Other (Deltaproteobacteria) 0.15% 0.10% 46 Sphingomonadaceae (Alphaproteobacteria) 0.15% 0.04% 47 Proteobacteria;Other 0.15% 0.01% 48 Erythrobacteraceae (Alphaproteobacteria) 0.14% 0.03% 49 Flavobacteriales;Other (Flavobacteria) 0.14% 0.05% 50 Actinomycetales;Other (Actinobacteria) 0.14% 0.04% 51 Puniceicoccaceae (Opitutae) 0.13% 0.05% 52 Piscirickettsiaceae (Gammaproteobacteria) 0.13% 0.07% 53 Bacteroidales;Other (Bacteroidia) 0.12% 0.06% 54 Hyphomonadaceae (Alphaproteobacteria) 0.12% 0.03% 55 SOGA31;Other 0.11% 0.07% 56 ZB2;Other 0.11% 0.03% 57 Helicobacteraceae (Epsilonproteobacteria) 0.11% 0.04% 58 Rhizobiales;Other (Alphaproteobacteria) 0.10% 0.02% 59 Campylobacteraceae (Epsilonproteobacteria) 0.09% 0.03% 60 Desulfobacteraceae (Deltaproteobacteria) 0.09% 0.03% 61 Phyllobacteriaceae (Alphaproteobacteria) 0.08% 0.02% 62 Alphaproteobacteria;Other 0.08% 0.05% 63 Phycisphaerales;Other (Phycisphaerae) 0.08% 0.03% 64 Francisellaceae (Gammaproteobacteria) 0.08% 0.03% 65 Rhizobiales;Other (Alphaproteobacteria) 0.07% 0.02% 66 Litincolaceae (Gammaproteobacteria) 0.07% 0.02% 67 Sphingobacteriales;Other (Sphingobacteria) 0.07% 0.01% 68 Other 0.07% 0.02% 69 Gammaproteobacteria;Other 0.06% 0.02% 70 Sphingobacteriaceae (Sphingobacteria) 0.06% 0.02% 71 Alteromonadaceae (Gammaproteobacteria) 0.06% 0.03% 72 Chromatiaceae (Gammaproteobacteria) 0.06% 0.04% 73 Shewanellaceae (Gammaproteobacteria) 0.05% 0.02%

76 74 Vibrionaceae (Gammaproteobacteria) 0.05% 0.03% 75 Chloroflexi;Other 0.05% 0.02% 76 Pseudanabaenaceae 0.05% 0.02% (Synechococcophycideae) 77 VC12-cl04;Other 0.05% 0.02% 78 Desulfobulbaceae (Deltaproteobacteria) 0.05% 0.01% 79 S15B-MN24;Other 0.04% 0.03% 80 koll13;Other (Actinobacteria) 0.04% 0.02% 81 Flammeovirgaceae (Sphingobacteria) 0.04% 0.01% 82 Verrucomicrobiaceae (Verrucomicrobiae) 0.04% 0.01% 83 Lactobacillaceae (Bacilli) 0.04% 0.02% 84 Bacteroidaceae (Bacteroidia) 0.04% 0.02% 85 Prevotellaceae (Bacteroidia) 0.04% 0.02% 86 Bacteriovoracaceae (Deltaproteobacteria) 0.04% 0.01% 87 CTD005-82B-02;Other 0.03% 0.02% (Deltaproteobacteria) 88 Chromatiales;Other (Gammaproteobacteria) 0.03% 0.01% 89 Porphyromonadaceae (Bacteroidia) 0.03% 0.02% 90 Bacteroidales;Other (Bacteroidia) 0.03% 0.01% 91 Alteromonadaceae (Gammaproteobacteria) 0.03% 0.01% 92 211ds20 (Gammaproteobacteria) 0.03% 0.01% 93 Sphingomonadales;Other 0.03% 0.01% (Alphaproteobacteria) 94 Gemmatimonadetes;Other 0.03% 0.01% 95 Caulobacteraceae (Alphaproteobacteria) 0.03% 0.01% 96 Gemmatimonadaceae (Gemmatimonadetes) 0.03% 0.01% 97 Clostridiales;Other (Clostridia) 0.03% 0.01% 98 125ds10 (Gammaproteobacteria) 0.03% 0.01% 99 Chroococcales;Other (Oscillatoriophycideae) 0.03% 0.01% 100 Deltaproteobacteria;Other 0.03% 0.01% 101 Psychromonadaceae (Gammaproteobacteria) 0.03% 0.02% 102 Chromatiales;Other (Gammaproteobacteria) 0.03% 0.01% 103 Euzebiaceae (Actinobacteria) 0.03% 0.01% 104 Fusobacteriaceae (Fusobacteria) 0.02% 0.01% 105 Hyphomicrobiaceae (Alphaproteobacteria) 0.02% 0.01% 106 Veillonellaceae (Clostridia) 0.02% 0.01% 107 Campylobacterales;Other 0.02% 0.01% (Epsilonproteobacteria) 108 Acetobacteraceae (Alphaproteobacteria) 0.02% 0.01% 109 Syntrophobacteraceae (Deltaproteobacteria) 0.02% 0.01% 110 SJA-4;Other 0.02% 0.01% 111 GN02;Other;Other;Other 0.02% 0.01%

77 112 Pseudomonadaceae (Gammaproteobacteria) 0.02% 0.01% 113 Clostridiaceae (Clostridia) 0.02% 0.01% 114 Lachnospiraceae (Clostridia) 0.02% 0.01% 115 Clostridia;Other 0.02% 0.01% 116 Colwelliaceae (Gammaproteobacteria) 0.02% 0.01% 117 Rhizobiaceae (Alphaproteobacteria) 0.02% 0.00% 118 Alteromonadales;Other 0.02% 0.01% (Gammaproteobacteria) 119 Oceanospirillaceae (Gammaproteobacteria) 0.02% 0.01% 120 Chloracidobacteria;Other 0.02% 0.01% 121 Gomphosphaeriaceae 0.02% 0.01% (Oscillatoriophycideae) 122 Myxococcales;Other (Deltaproteobacteria) 0.02% 0.01% 123 Aeromonadaceae (Gammaproteobacteria) 0.02% 0.01% 124 TM7-1;Other 0.02% 0.01% 125 Desulfobacterales;Other 0.02% 0.01% (Deltaproteobacteria) 126 Desulfuromonadales;Other 0.01% 0.01% (Deltaproteobacteria) 127 ZB1;Other (Ignavibacteria) 0.01% 0.01% 128 Rhodobacterales;Other 0.01% 0.00% (Alphaproteobacteria) 129 Desulfovibrionaceae (Deltaproteobacteria) 0.01% 0.01% 130 Spartobacteriaceae (Spartobacteria) 0.01% 0.00% 131 Sphingomonadales;Other 0.01% 0.01% (Alphaproteobacteria) 132 Pseudoalteromonadaceae 0.01% 0.00% (Gammaproteobacteria) 133 EW055;Other (TM7-3) 0.01% 0.01% 134 Clostridiales Family XII. Incertae Sedis 0.01% 0.01% (Clostridia) 135 Verrucomicrobia;Other 0.01% 0.01% 136 Gallionellaceae (Betaproteobacteria) 0.01% 0.01% 137 Rhodobacteraceae (Alphaproteobacteria) 0.01% 0.01% 138 Enterobacteriaceae (Gammaproteobacteria) 0.01% 0.00% 139 Ignavibacteriaceae (Ignavibacteria) 0.01% 0.00% 140 Bradyrhizobiaceae (Alphaproteobacteria) 0.01% 0.01% 141 Coxiellaceae (Gammaproteobacteria) 0.01% 0.00% 142 Myxococcales;Other (Deltaproteobacteria) 0.01% 0.01% 143 Opitutaceae (Opitutae) 0.01% 0.00% 144 Verrucomicrobiales;Other 0.01% 0.00% (Verrucomicrobiae)

78 145 Legionellaceae (Gammaproteobacteria) 0.01% 0.00% 146 SUP05 (Gammaproteobacteria) 0.01% 0.01% 147 O4D2Z37;Other (Anaerolineae) 0.01% 0.01% 148 Anaerolineae;Other 0.01% 0.00% 149 Rhodocyclaceae (Betaproteobacteria) 0.01% 0.00% 150 SR1;Other 0.01% 0.00% 151 Clostridiales Family XIII. Incertae Sedis 0.01% 0.01% (Clostridia) 152 GN04;Other 0.01% 0.01% 153 Acidimicrobiales;Other (Actinobacteria) 0.01% 0.00% 154 Ruminococcaceae (Clostridia) 0.01% 0.00% 155 Rhodocyclales;Other (Betaproteobacteria) 0.01% 0.00% 156 R76-B18;Other 0.01% 0.00% 157 Caldilineaceae (Anaerolineae) 0.01% 0.00% 158 Syntrophaceae (Deltaproteobacteria) 0.01% 0.00% 159 TM7;Other 0.01% 0.00% 160 Thermodesulfovibrionaceae (Nitrospira) 0.01% 0.00% 161 Nitrospinaceae (Deltaproteobacteria) 0.01% 0.00% 162 Actinobacteria;Other 0.01% 0.00% 163 Acholeplasmataceae (Mollicutes) 0.01% 0.00% 164 Legionellales;Other (Gammaproteobacteria) 0.01% 0.00% 165 Piscirickettsiaceae (Gammaproteobacteria) 0.01% 0.00% 166 Bdellovibrionaceae (Deltaproteobacteria) 0.01% 0.00% 167 Moraxellaceae (Gammaproteobacteria) 0.01% 0.00% 168 Rhodocyclales;Other (Betaproteobacteria) 0.005% 0.002% 169 Synechococcophycideae;Other 0.005% 0.003% 170 Coriobacteriaceae (Actinobacteria) 0.005% 0.002% 171 TM7-3;Other 0.005% 0.002% 172 YS2;Other (4C0d-2) 0.004% 0.003% 173 Neisseriaceae (Betaproteobacteria) 0.004% 0.002% 174 Caldilineales;Other (Anaerolineae) 0.004% 0.003% 175 BSV19;Other 0.004% 0.002% 176 Spirochaetaceae (Spirochaetes) 0.004% 0.002% 177 Kordiimonadaceae (Alphaproteobacteria) 0.004% 0.002% 178 Chromatiaceae (Gammaproteobacteria) 0.004% 0.002% 179 DS-18;Other (iii1-8) 0.003% 0.002% 180 Hydrogenophilaceae (Betaproteobacteria) 0.003% 0.002% 181 Nitrospiraceae (Nitrospira) 0.003% 0.002% 182 Opitutae;Other 0.003% 0.002% 183 ML615J-28;Other 0.003% 0.002% 184 BB34;Other 0.003% 0.001% 185 NT-B4;Other (Dehalococcoidetes) 0.003% 0.002%

79 186 Acidobacteriales;Other (Acidobacteria) 0.003% 0.002% 187 Nitrosomonadaceae (Betaproteobacteria) 0.003% 0.002% 188 Syntrophobacterales;Other 0.003% 0.001% (Deltaproteobacteria) 189 Armatimonadaceae (Armatimonadia) 0.003% 0.001% 190 Clostridiales Family XI. Incertae Sedis 0.003% 0.002% (Clostridia) 191 Thiotrichaceae (Gammaproteobacteria) 0.003% 0.002% 192 Methylocystaceae (Alphaproteobacteria) 0.003% 0.002% 193 Gloeobacteraceae (Gloeobacterophycideae) 0.003% 0.002% 194 mle1-12;Other (4C0d-2) 0.003% 0.001% 195 Mycobacteriaceae (Actinobacteria) 0.003% 0.002% 196 Blgi18;Other (TM7-3) 0.003% 0.001% 197 Elusimicrobiaceae (Elusimicrobia) 0.002% 0.002% 198 GKS2-174;Other 0.002% 0.001% 199 Rhodothermaceae (Sphingobacteria) 0.002% 0.002% 200 Dehalococcoidaceae (Dehalococcoidetes) 0.002% 0.002% 201 OS-K;Other 0.002% 0.001% 202 Elusimicrobiales;Other (Elusimicrobia) 0.002% 0.001% 203 envOPS12;Other (Anaerolineae) 0.002% 0.001% 204 Methylophilales;Other (Betaproteobacteria) 0.002% 0.002% 205 Kiloniellaceae (Alphaproteobacteria) 0.002% 0.001% 206 Sva0725;Other (Sva0725) 0.002% 0.002% 207 MVP-15;Other 0.002% 0.002% 208 FAC88;Other (Elusimicrobia) 0.002% 0.001% 209 Beijerinckiaceae (Alphaproteobacteria) 0.002% 0.002% 210 SBRH58;Other 0.002% 0.002% 211 Mariprofundaceae (Zetaproteobacteria) 0.002% 0.001% 212 Caldilineales;Other (Anaerolineae) 0.002% 0.002% 213 SM2F11;Other 0.002% 0.001% 214 CL500-15;Other (agg27) 0.002% 0.001% 215 Caldithrixales;Other (Caldithrixae) 0.002% 0.001% 216 A714017 (AB16) 0.002% 0.002% 217 Planctomycetaceae (Planctomycea) 0.002% 0.002% 218 Gemmatimonadales;Other 0.002% 0.001% (Gemmatimonadetes) 219 PRR-12;Other 0.002% 0.001% 220 SC72;Other 0.002% 0.001% 221 S0208;Other (Anaerolineae) 0.002% 0.001% 222 TG3-2;Other 0.002% 0.001% 223 Nannocystaceae (Deltaproteobacteria) 0.002% 0.002% 224 SC3;Other 0.002% 0.001%

80 225 wb1_P06 (Actinobacteria) 0.002% 0.001% 226 Bifidobacteriaceae (Actinobacteria) 0.002% 0.002% 227 NKB15;Other (Deltaproteobacteria) 0.002% 0.002% 228 Microcystaceae (Oscillatoriophycideae) 0.002% 0.001% 229 RB25;Other 0.002% 0.001% 230 Methylophilales;Other (Betaproteobacteria) 0.002% 0.001% 231 Aurantimonadaceae (Alphaproteobacteria) 0.002% 0.001% 232 d153;Other (WCHB1-64) 0.002% 0.001% 233 Campylobacterales;Other 0.002% 0.001% (Epsilonproteobacteria) 234 BD4-9;Other 0.002% 0.001% 235 SHA-124;Other (OP8_1) 0.002% 0.001% 236 SJA-28;Other 0.002% 0.001% 237 Crenotrichaceae (Gammaproteobacteria) 0.002% 0.001% 238 Geobacteraceae (Deltaproteobacteria) 0.001% 0.001% 239 GCA004;Other (Anaerolineae) 0.001% 0.001% 240 A4b;Other (Anaerolineae) 0.001% 0.001% 241 Rikenellaceae (Bacteroidia) 0.001% 0.001% 242 BB36;Other (OP11-1) 0.001% 0.001% 243 Streptococcaceae (Bacilli) 0.001% 0.001% 244 GIF10;Other (koll11) 0.001% 0.001% 245 Dehalococcoidetes;Other 0.001% 0.001% 246 Carnobacteriaceae (Bacilli) 0.001% 0.001% 247 Oleiphilaceae (Gammaproteobacteria) 0.001% 0.001% 248 Holophagales;Other (Holophagae) 0.001% 0.001% 249 WM1006;Other 0.001% 0.001% 250 Chloroflexaceae (Chloroflexi) 0.001% 0.001% 251 Nostocaceae (Nostocophycideae) 0.001% 0.001% 252 Haliangiaceae (Deltaproteobacteria) 0.001% 0.001% 253 Firmicutes;Other 0.001% 0.001% 254 LD19 (Methylacidiphilae) 0.001% 0.001% 255 Moritellaceae (Gammaproteobacteria) 0.001% 0.001% 256 Methylobacteriaceae (Alphaproteobacteria) 0.001% 0.001% 257 Thiotrichales;Other (Gammaproteobacteria) 0.001% 0.001% 258 Solibacteraceae (Solibacteres) 0.001% 0.001% 259 OM190;Other (agg27) 0.001% 0.001% 260 Oceanospirillales;Other 0.001% 0.001% (Gammaproteobacteria) 261 Clostridiales;Other (Clostridia) 0.001% 0.001% 262 OP3;Other 0.001% 0.001% 263 GN03;Other (PRR-12) 0.001% 0.001% 264 BPC102;Other 0.001% 0.001%

81 265 SHA-20;Other (Anaerolineae) 0.001% 0.001% 266 Marinilabiaceae (Bacteroidia) 0.001% 0.001% 267 J115 (Gammaproteobacteria) 0.001% 0.001% 268 Bdellovibrionales;Other 0.001% 0.001% (Deltaproteobacteria) 269 HTCC2089 (Gammaproteobacteria) 0.001% 0.001% 270 GN08;Other 0.001% 0.001% 271 HMMVPog-54;Other (OP8_1) 0.001% 0.001% 272 S085;Other (TK17) 0.001% 0.001% 273 Turicibacteraceae (Bacilli) 0.001% 0.001% 274 Betaproteobacteria;Other 0.001% 0.001% 275 NB1-j;Other (Deltaproteobacteria) 0.001% 0.001% 276 Epsilonproteobacteria;Other 0.001% 0.001% 277 Thiotrichales;Other (Gammaproteobacteria) 0.001% 0.001% 278 _J-1;Other (OP11-4) 0.001% 0.001% 279 Desulfuromonadaceae (Deltaproteobacteria) 0.001% 0.001% 280 SHA-116 (WWE1) 0.001% 0.001% 281 LD1-PA13;Other (PRR-12) 0.001% 0.001% 282 Acidobacteriaceae (Acidobacteria) 0.001% 0.001% 283 WD294 (Armatimonadia) 0.001% 0.001% 284 HN1-15;Other (Thermomicrobia) 0.001% 0.001% 285 Lactobacillales;Other (Bacilli) 0.001% 0.001% 286 AT425_EubG1;Other 0.001% 0.001% 287 kpj58rc (koll11) 0.001% 0.001% 288 Micrococcaceae (Actinobacteria) 0.001% 0.001% 289 Chloroflexales;Other (Chloroflexi) 0.001% 0.001% 290 Oscillatoriophycideae;Other 0.001% 0.001% 291 Phycisphaeraceae (Phycisphaerae) 0.001% 0.001% 292 WS1;Other 0.001% 0.001% 293 SHA-114;Other 0.001% 0.001% 294 Nocardioidaceae (Actinobacteria) 0.001% 0.001% 295 Erysipelotrichaceae (Erysipelotrichi) 0.001% 0.001% 296 TG3;Other 0.001% 0.001% 297 TM7-3;Other 0.001% 0.001% 298 Verruco-5;Other 0.001% 0.001% 299 Holophagaceae (Holophagae) 0.001% 0.001% 300 Corynebacteriaceae (Actinobacteria) 0.001% 0.001% 301 MC47;Other (Actinobacteria) 0.001% 0.001% 302 Brucellaceae (Alphaproteobacteria) 0.001% 0.001% 303 Methylococcaceae (Gammaproteobacteria) 0.001% 0.001% 304 ABY1_OD1;Other 0.001% 0.001% 305 HDBW-WB69;Other 0.001% 0.001%

82 306 OP11-3;Other 0.001% 0.001% 307 RCP1-27;Other (OP11-4) 0.001% 0.001% 308 Proteobacteria;Other 0.001% 0.001% 309 Mycoplasmataceae (Mollicutes) 0.001% 0.001% 310 Acidobacteria;Other 0.001% 0.001% 311 Chloroflexi;Other 0.001% 0.001% 312 Dehalococcoidetes;Other 0.001% 0.001% 313 Lentisphaerales;Other (Lentisphaerae) 0.001% 0.001% 314 Xanthomonadales;Other 0.001% 0.001% (Gammaproteobacteria) 315 SJA-176;Other 0.001% 0.001% 316 Lentisphaerae;Other 0.001% 0.001% 317 OP8_2;Other 0.001% 0.001% 318 JTB38 (Deltaproteobacteria) 0.001% 0.001% 319 Methylococcales;Other 0.001% 0.001% (Gammaproteobacteria) 320 Hahellaceae (Gammaproteobacteria) 0.001% 0.001% 321 Microthrixaceae (Actinobacteria) 0.001% 0.001% 322 Enterococcaceae (Bacilli) 0.001% 0.001% 323 NKB19;Other 0.001% 0.001% 324 Rickettsiales;Other (Alphaproteobacteria) 0.001% 0.001% 325 SAW1_B6;Other (PRR-12) 0.001% 0.001% 326 MVP-88;Other (Elusimicrobia) 0.0005% 0.0005% 327 GN02;Other;Other;Other 0.0005% 0.0005% 328 GOUTA4;Other 0.0005% 0.0005% 329 OP11-1;Other 0.0005% 0.0005% 330 SB-45;Other (JS1) 0.0005% 0.0005% 331 Epsilonproteobacteria;Other 0.0005% 0.0005% 332 Spirochaetes;Other 0.0005% 0.0005% 333 Synergistales;Other (Synergistia) 0.0005% 0.0005% 334 MVS-40;Other 0.0005% 0.0005% 335 Pirellulales;Other (Planctomycea) 0.0005% 0.0005% 336 Verrucomicrobiales;Other 0.0005% 0.0005% (Verrucomicrobiae) 337 Acidobacteria;Other 0.0004% 0.0004% 338 Acidimicrobiales;Other (Actinobacteria) 0.0004% 0.0004% 339 BRC1;Other 0.0004% 0.0004% 340 Synechococcales;Other 0.0004% 0.0004% (Synechococcophycideae) 341 Isosphaeraceae (Planctomycea) 0.0004% 0.0004% 342 Alcanivoracaceae (Gammaproteobacteria) 0.0004% 0.0004% 343 Actinobacteria;Other 0.0004% 0.0004%

83 344 GN15;Other 0.0004% 0.0004%

84 Table 2.5: Spearman’s rank R values of correlations between bacterial classes and water parameters.

Taxonomy Temperature Salinity Dissolved Oxygen pH Fluorescence Oyster Water Oyster Water Oyster Water Oyster Water Oyster Water Other 0.503 0.05455 0.01818 -0.5394 -0.5273 -0.07879 -0.304 0.4377 0.4424 0.5758 Bacteria;Other 0.5879 0.8909 0.1394 -0.2727 -0.6121 -0.8788 -0.1702 -0.4195 0.5515 0.6364 Acidobacteria;Other -0.05803 -0.4062 0.5222 -0.05803 0.05803 0.4062 -0.1164 0.0582 -0.5222 -0.4062 Acidobacteria;Unclassified 0.05803 0.1741 -0.4062 -0.5222 -0.05803 -0.1741 0.4074 0.5238 0.4062 0.5222 Acidobacteria 0.3755 0.1099 -0.4165 -0.1228 -0.437 -0.1487 -0.202 0.1102 0.478 0.3814 Chloracidobacteria -0.3892 0.04526 -0.1297 -0.4332 0.3892 -0.08405 0.5119 0.01297 0.1297 0.3426 Holophagae 0.1288 -0.5276 0.1779 0.2249 -0.09203 0.5276 -0.2646 0.1649 0.1043 -0.5276 OS-K -0.09694 0.6833 -0.09694 0.2768 0.1566 -0.7006 0.2767 -0.7028 0.2162 0.3546 RB25 0.05803 -0.4847 -0.4062 -0.1864 -0.05803 0.4847 0.4074 0.359 0.4062 0.03728 Solibacteres 0.05803 -0.2941 -0.4062 -0.32 -0.05803 0.2941 0.4074 0.321 0.4062 -0.06055 Sva0725 -0.03728 0.1741 -0.09694 -0.5222 0.03728 -0.1741 0.2468 0.5238 0.06711 0.5222 iii1-8 -0.2025 -0.1741 0.1411 -0.1741 0.2147 0.1741 0.1539 -0.1164 0.006135 -0.05803 Actinobacteria 0.2 0.503 -0.3697 -0.05455 -0.1515 -0.5152 0.5106 -0.2553 0.5273 0.1273 Bacteroidetes;Other -0.103 0.5515 -0.2364 -0.6727 0.1152 -0.5758 0.4316 0.1033 0.0303 0.7212 Bacteroidia -0.1758 -0.3697 0.3091 0.4061 0.2121 0.4182 0.1641 0 -0.4182 -0.6 Flavobacteria -0.1273 -0.4182 -0.01818 0.3455 0.2 0.4788 0.5957 0.1398 0.09091 -0.6 Sphingobacteria 0.4788 -0.1879 -0.5273 -0.6121 -0.4061 0.2 0.3647 0.5046 0.6 0.2606 Caldithrixae -0.2901 0.2422 0.4062 -0.5276 0.2901 -0.2422 0.1746 0.1388 -0.2901 0.3719 BSV19 -0.1989 0.0478 0.06055 -0.437 0.1989 -0.09559 0.4165 0.589 0.0346 0.676 Ignavibacteria 0.4061 0.8283 0.1636 -0.5951 -0.4182 -0.7915 -0.3343 -0.2523 0.3091 0.7547 OPB56 0.2242 0.5836 -0.297 0.08511 -0.1515 -0.5593 -0.01824 -0.497 0.2364 0.1824 SJA-28 -0.1989 0.1741 0.06055 -0.5222 0.1989 -0.1741 0.4165 0.5238 0.0346 0.5222 Chloroflexi;Other -0.006465 -0.5366 -0.5366 -0.472 0.006465 0.5366 0.6031 0.642 0.4332 0.1616 Anaerolineae 0.01216 0.1099 0.2857 -0.07112 -0.04863 -0.1616 -0.1951 0.3956 -0.2006 0.3297 Chloroflexi 0.0938 0.6748 -0.2189 -0.2067 -0.05002 -0.6626 0.3732 -0.07012 0.1938 0.4499 Dehalococcoidetes -0.2663 -0.4698 0.3687 -0.2908 0.1912 0.4698 0.1438 0.531 0.06828 -0.06711 SOGA31 0.459 -0.1487 -0.64 -0.5883 -0.3944 0.1487 0.1491 0.4474 0.5883 0.2521 Cyanobacteria;Other 0.5106 0.3455 0.06687 -0.2 -0.4499 -0.3212 -0.4512 0.09726 0.1094 0.503 4C0d-2 0.4643 -0.5276 -0.4984 0.2249 -0.3892 0.5276 -0.1541 0.1649 0.4302 -0.5276 Gloeobacterophycideae 0.5882 0.1816 -0.1989 -0.7006 -0.5103 -0.1816 -0.5119 0.7028 0.1643 0.7006 Oscillatoriophycideae 0.5891 0.7441 -0.3654 -0.3064 -0.4996 -0.7691 -0.2019 -0.2446 0.425 0.8004 S15B-MN24 0.5891 0.5882 -0.3952 -0.1989 -0.5145 -0.5103 -0.2767 -0.5119 0.3654 0.1643 Synechococcophycideae 0.6848 0.697 -0.1273 -0.5152 -0.6606 -0.7091 -0.3526 0.01824 0.4667 0.7697 Elusimicrobia 0.8125 -0.3892 -0.1024 -0.08876 -0.7989 0.3414 -0.5547 0.2945 0.5599 -0.02731 Firmicutes;Other 0.2162 0.05803 -0.09694 -0.4062 -0.2162 -0.05803 -0.08975 0.4074 -0.08202 0.4062 Bacilli -0.2121 -0.3743 0.2242 0.6074 0.1394 0.3374 0.4438 0.08 -0.0303 -0.5583 Clostridia -0.006061 -0.0303 -0.1394 0.0303 0.04242 0.1273 0.5167 0.1094 0.006061 -0.2364 Fusobacteria 0.2848 0.1875 -0.2606 -0.03233 -0.2364 -0.2263 0.4195 -0.05188 0.2485 -0.1099 GN02;Other 0.4788 0.2848 -0.07879 -0.3697 -0.4545 -0.2242 -0.1702 -0.09119 0.503 0.3818 BB34 0.5753 -0.5742 -0.05628 -0.1566 -0.6128 0.5742 -0.4265 0.5161 0.419 -0.08202 GKS2-174 -0.2263 -0.1384 -0.2909 -0.2422 0.2263 0.1384 0.2594 -0.1475 -0.04526 -0.2941 VC12-cl04 0.3091 -0.6065 0.3576 -0.1688 -0.2364 0.6065 -0.5532 0.5927 -0.4061 -0.1376 GN15 0.5222 0.5222 0.05803 0.05803 -0.4062 -0.4062 -0.4074 -0.4074 0.1741 0.1741 Gemmatimonadetes -0.1273 0.7697 0.103 -0.4303 0.1394 -0.7455 0.1216 -0.07903 0.0303 0.7212 Lentisphaerae -0.5222 0.4152 0.1741 -0.2335 0.5222 -0.493 0.291 0.07809 -0.1741 0.6228 MVP-15;Other 0.08754 0.2901 -0.1938 -0.2901 -0.07504 -0.2901 -0.03136 -0.291 0.2626 0.05803 Nitrospira 0.1152 0.01251 -0.103 -0.06878 -0.103 -0.0938 0.1277 -0.1819 0.3212 0.04377 WCHB1-64 0.7442 0.0519 -0.1297 -0.3373 -0.7716 -0.0519 -0.4726 -0.2907 0.6009 -0.00865 BD4-9 0.2768 0.1741 -0.5103 -0.5222 -0.2768 -0.1741 0.03471 0.5238 0.32 0.5222 koll11 0.2768 -0.3027 -0.5103 -0.2162 -0.2768 0.3027 0.03471 0.5293 0.32 0.2162 OP8_1 0.05803 0.1741 -0.4062 -0.5222 -0.05803 -0.1741 0.4074 0.5238 0.4062 0.5222 OP8_2 0.2516 0.4062 -0.04295 0.2901 -0.2638 -0.5222 0.1662 -0.5238 0.4847 0.2901 Phycisphaerae 0.3526 0.6261 -0.3465 0.07295 -0.304 -0.6383 0.08537 -0.314 0.4438 0.4377 agg27 -0.0692 0.3719 -0.4498 -0.0346 0.0692 -0.4671 0.256 -0.1562 0.2941 0.5103 Proteobacteria;Other 0.06667 0.3697 0.1273 0.5636 -0.006061 -0.3576 0.01216 -0.2796 -0.006061 0.1273 Alphaproteobacteria -0.4303 -0.1273 -0.04242 0.7455 0.4788 0.1152 -0.1763 -0.304 -0.3697 -0.297 Betaproteobacteria 0.1394 -0.3212 -0.5273 -0.6848 -0.103 0.2848 0.5957 0.5532 0.5273 0.3212 Deltaproteobacteria -0.1152 0.5152 -0.297 0.3818 0.09091 -0.5758 0.4985 -0.4255 0.3455 0.09091 Epsilonproteobacteria -0.503 -0.3939 0.0303 -0.06667 0.5273 0.3455 0.2492 0.383 -0.2121 -0.2121 Gammaproteobacteria -0.7818 -0.4424 0.4667 0.697 0.8061 0.3818 0.4255 -0.231 -0.7333 -0.6121 Zetaproteobacteria 0.05803 -0.2901 -0.4062 0.4062 -0.05803 0.2901 0.4074 0.1746 0.4062 -0.2901 AB16 -0.4062 0.2249 -0.05803 0.6228 0.4062 -0.3027 0.0582 -0.4468 -0.4062 -0.2335 SC3;Other 0.05803 -0.4498 -0.4062 -0.1643 -0.05803 0.4498 0.4074 -0.03037 0.4062 -0.3719 SM2F11;Other 0.3979 -0.1741 -0.2941 -0.1741 -0.32 0.1741 0.06073 -0.1164 0.4498 -0.05803 SR1;Other 0.4502 -0.2595 -0.04377 -0.3551 -0.4127 0.2595 -0.3011 0.2123 0.1376 -0.06828 Spirochaetes 0.653 -0.4996 -0.3168 -0.2312 -0.6788 0.4996 -0.415 0.5909 0.6013 -0.007457 TG3-2 0.444 -0.1297 0.4002 -0.4671 -0.4315 0.1297 -0.4986 0.4685 -0.1876 0.1557 SBRH58 0.05803 -0.0346 -0.4062 -0.2768 -0.05803 0.0346 0.4074 -0.1996 0.4062 -0.2249 SJA-4 -0.2128 -0.6271 0.1337 0.3685 0.2006 0.5754 0.08841 0.2788 0.04863 -0.2909 TM7;Other -0.05455 0.4028 0.2 -0.4028 0.04242 -0.4643 0.06079 -0.2842 0.05455 0.437 TM7-1 -0.1398 -0.1875 0.1459 0.5883 0.1277 0.1228 0.07622 0.1038 0.1581 -0.2263 TM7-3 -0.5471 -0.6828 -0.0304 -0.2117 0.5046 0.6828 0.375 0.4143 -0.0304 -0.2117 Tenericutes;Other 0.1566 -0.5152 -0.1566 0.1636 -0.2312 0.4788 -0.2169 0.2371 0.3505 -0.3333 Erysipelotrichi -0.05803 0.1741 0.5222 -0.5222 0.05803 -0.1741 -0.1164 0.5238 -0.5222 0.5222 ML615J-28 0.2133 -0.5462 0.09698 -0.3619 -0.1616 0.5462 -0.05188 0.6643 -0.05819 0.006828 Mollicutes 0.5628 -0.6691 0.1876 -0.239 -0.5253 0.6691 -0.5363 0.387 -0.06878 -0.239 Verrucomicrobia;Other 0.2651 0.5667 -0.3814 -0.7374 -0.3426 -0.5189 0.227 0.1507 0.6659 0.6555 Methylacidiphilae 0.5882 -0.2335 -0.1989 -0.0519 -0.5103 0.2335 -0.5119 0.04338 0.1643 -0.1038 Opitutae 0.4909 0.3526 -0.5879 -0.2371 -0.4545 -0.3161 0.1641 -0.02744 0.7455 0.4681 R76-B18 -0.2868 0.4549 0.2595 -0.3206 0.3482 -0.5443 0.2876 0.1421 -0.06145 0.768 Spartobacteria 0.8037 0.8129 -0.5215 -0.6003 -0.7669 -0.8004 -0.2031 -0.2979 0.7301 0.7441 Verruco-5 0.05803 -0.1741 -0.4062 -0.1741 -0.05803 0.1741 0.4074 -0.1164 0.4062 -0.05803 Verrucomicrobiae 0.1288 0.06667 -0.1411 -0.3818 -0.1043 -0.1152 0.2523 0.4073 0.04295 0.2 PRR-12 0.1376 0.4152 -0.08129 -0.2335 -0.1626 -0.493 -0.2195 0.07809 0.2251 0.6228 SC72 0.5829 0.4152 0.04295 -0.2335 -0.5706 -0.493 -0.5108 0.07809 0.3497 0.6228 WM1006 0.2422 -0.2901 -0.5276 0.4062 -0.2422 0.2901 0.1388 0.1746 0.3719 -0.2901 ZB2;Other -0.3939 -0.5879 0.1152 -0.4061 0.4182 0.5515 0.2918 0.4316 -0.2364 0.0303

85 Table 2.6: Taxa observed in oyster EF and water samples grouped by environmental parameter correlations with relative abundance.

Taxa (Phylum; Class) Treatment Group Acidobacteria; Sva0725 Water I Acidobacteria; Unassigned Water Chlorobi; SJA 28 Water OP3; BD4 9 Water OP8; OP8_1 Water Tenericutes; Erysipelotrichi Water Proteobacteria; Betaproteobacteria Oyster Other; Unassigned Water Chlorobi; BSV19 Water Chloroflexi; Other Oyster Acidobacteria; RB25 Oyster Acidobacteria; Unassigned Oyster Acidobacteria; Solibacteres Oyster OP8; OP8_1 Oyster Proteobacteria; Zetaproteobacteria Oyster SC3; Unassigned Oyster TM6; SBRH58 Oyster Verrucomicrobia; Verruco 5 Oyster Firmicutes; Other Water Actinobacteria; Actinobacteria Oyster Cyanobacteria; Gloeobacterophycideae Water Proteobacteria; Deltaproteobacteria Oyster Planctomycetes; agg27 Oyster Bacteroidetes; Sphingobacteria Water Chloroflexi; SOGA31 Water TG3; TG3 2 Water Proteobacteria; Betaproteobacteria Water Nitrospirae; Nitrospira Oyster II Acidobacteria; Acidobacteria Water OP8; OP8_2 Oyster Cyanobacteria; Other Water Verrucomicrobia; Opitutae Water SM2F11; Unassigned Oyster Planctoymcetes; Phycisphaerae Oyster GN02; Other Water Caldithrix; Caldithrixae Water WS6; WM1006 Oyster OP3; koll11 Oyster OP3; BD4 9 Oyster

86 MVP 15; Unassigned Oyster Chlorobi; OPB56 Oyster Acidobacteria; Chloracidobacteria Water Tenericutes; Other Oyster WS3; PRR 12 Oyster Chloroflexi; Anaerolineae Water Fusobacteria; Fusobacteria Oyster Chloroflexi; Chloroflexi Oyster Verrucomicrobia; Verrucomicrobiae Water Verrucomicrobia; Verrucomicrobiae Oyster WS3; PRR 12 Water III Lentisphaerae; Lentisphaerae Water WS6; SC72 Water Verrucomicrobia; R76 B18 Water Bacteroidetes; Sphingobacteria Oyster Verrucomicrobia; Other Oyster Verrucomicrobia; Opitutae Oyster Chloroflexi; SOGA31 Oyster Bacteroidetes; Other Water Verrucomicrobia; Other Water Cyanobacteria; Synechococcophycideae Oyster IV Spirochaetes; Spirochaetes Oyster Bacteria; Other Water OP11; WCHB1 64 Oyster Elusimicrobia; Elusimicrobia Oyster Cyanobacteria; Synechococcophycideae Water Gemmatimonadetes; Gemmatimonadetes Water Cyanobacteria; Oscillatoriophycideae Water Verrucomicrobia; Spartobacteria Oyster Chlorobi; Ignavibacteria Water Verrucomicrobia; Spartobacteria Water Cyanobacteria; Oscillatoriophycideae Oyster Cyanobacteria; S15B MN24 Oyster Cyanobacteria; 4c0d 2 Oyster Acidobacteria; Acidobacteria Oyster TM7; Other Water Proteobacteria; Other Water V Chlorobi; Ignavibacteria Oyster OP8; OP8_2 Water Acidobacteria; OS K Water Proteobacteria; Deltaproteobacteria Water Tenericutes; Mollicutes Oyster Bacteria; Other Oyster

87 Planctoymcetes; Phycisphaerae Water Other; Unassigned Oyster GN02; Other Oyster Planctomycetes; agg27 Water Chloroflexi; Chloroflexi Water GN02; BB34 Oyster WS6; SC72 Oyster SR1; Unassigned Oyster Actinobacteria; Actinobacteria Water Chlorobi; OPB56 Water GN04; GN15 Water GN04; GN15 Oyster Cyanobacteria; Other Oyster Cyanobacteria; S15B MN24 Water Verrucomicrobia; Methylacidiphilae Oyster Cyanobacteria; Gloeobacterophycideae Oyster SAR406; AB16 Water VI TG3; TG3 2 Oyster GN02; VC12 cl04 Oyster Acidobacteria; Holophagae Oyster Nitrospirae; Nitrospira Water Tenericutes; ML615J 28 Oyster Fusobacteria; Fusobacteria Water Firmicutes; Other Oyster MVP 15; Unassigned Water OP11; WCHB1 64 Water TM7; Other Oyster Proteobacteria; Other Oyster Firmicutes; Clostridia Water Chloroflexi; Anaerolineae Oyster Verrucomicrobia; Methylacidiphilae Water Verrucomicrobia; Verruco 5 Water SM2F11; Unassigned Water Acidobacteria; iii1 8 Water GN02; GKS2 174 Water TM6; SBRH58 Water Chlorobi; SJA 28 Oyster VII Chlorobi; BSV19 Oyster Firmicutes; Bacilli Oyster Verrucomicrobia; R76 B18 Oyster Chloroflexi; Dehalococcoidetes Oyster Acidobacteria; iii1 8 Oyster TM6; SJA 4 Oyster

88 Gemmatimonadetes; Gemmatimonadetes Oyster TM7; TM7 1 Oyster Acidobacteria; OS K Oyster Acidobacteria; Sva0725 Oyster Bacteroidetes; Flavobacteria Oyster Firmicutes; Clostridia Oyster Bacteroidetes; Other Oyster OP3; koll11 Water Acidobacteria; Chloracidobacteria Oyster Acidobacteria; Solibacteres Water GN02; GKS2 174 Oyster SR1; Unassigned Water GN02; VC12 cl04 Water VIII GN02; BB34 Water Spirochaetes; Spirochaetes Water Chloroflexi; Dehalococcoidetes Water TM7; TM7 3 Water Tenericutes; Mollicutes Water TM7; TM7 3 Oyster Acidobacteria; RB25 Water Elusimicrobia; Elusimicrobia Water Proteobacteria; Epsilonproteobacteria Water Chloroflexi; Other Water Tenericutes; ML615J 28 Water ZB2; Unassigned Water Proteobacteria; Zetaproteobacteria Water IX Caldithrix; Caldithrixae Oyster WS6; WM1006 Water Bacteroidetes; Bacteroidia Oyster TM7; TM7 1 Water Proteobacteria; Alphaproteobacteria Water Tenericutes; Erysipelotrichi Oyster Acidobacteria; Other Oyster Firmicutes; Bacilli Water Bacteroidetes; Flavobacteria Water Bacteroidetes; Bacteroidia Water Proteobacteria; Gammaproteobacteria Water Proteobacteria; Gammaproteobacteria Oyster X Lentisphaerae; Lentisphaerae Oyster Proteobacteria; Epsilonproteobacteria Oyster ZB2; Unassigned Oyster Tenericutes; Other Water TM6; SJA 4 Water

89 Cyanobacteria; 4c0d 2 Water Acidobacteria; Holophagae Water Proteobacteria; Alphaproteobacteria Oyster SC3; Unassigned Water Acidobacteria; Other Water SAR406; AB16 Oyster

Table 2.7: Taxa observed in oyster EF and water samples grouped by environmental parameter correlations with absolute abundance.

Taxa (Phylum; Class) Treatment Group Acidobacteria; Sva0725 Water I Acidobacteria; Unassigned Water Chlorobi; SJA 28 Water OP3; BD4 9 Water OP8; OP8_1 Water Tenericutes; Erysipelotrichi Water Other Water Chlorobi; BSV19 Water Cyanobacteria; Gloeobacterophycideae Water Acidobacteria; RB25 Oyster Acidobacteria; Unassigned Oyster Acidobacteria; Solibacteres Oyster OP8; OP8_1 Oyster Proteobacteria; Zetaproteobacteria Oyster SC3; Unassigned Oyster TM6; SBRH58 Oyster Verrucomicrobia; Verruco 5 Oyster Firmicutes; Other Water Chloroflexi; Other Oyster TG3; TG3 2 Water Chloroflexi; SOGA31 Water Chloroflexi; Anaerolineae Water Verrucomicrobia; Verrucomicrobiae Water Acidobacteria; Chloracidobacteria Water Caldithrix; Caldithrixae Water WS6; WM1006 Oyster OP3; koll11 Oyster OP3; BD4 9 Oyster

90 Spirochaetes; Spirochaetes Water II Chloroflexi; Dehalococcoidetes Water Tenericutes; Mollicutes Water TM7; TM7 3 Water GN02; BB34 Water GN02; VC12 cl04 Water OP3; koll11 Water Acidobacteria; RB25 Water Acidobacteria; Chloracidobacteria Oyster Tenericutes; ML615J 28 Water ZB2; Unassigned Water Chloroflexi; Other Water Proteobacteria; Zetaproteobacteria Water III Caldithrix; Caldithrixae Oyster WS6; WM1006 Water TM7; TM7 1 Water Verrucomicrobia; R76 B18 Oyster Chloroflexi; Dehalococcoidetes Oyster Firmicutes; Bacilli Water ZB2; Unassigned Oyster Tenericutes; Erysipelotrichi Oyster Acidobacteria; Other Oyster Firmicutes; Bacilli Oyster Firmicutes; Clostridia Oyster TM7; Other Oyster Gemmatimonadetes; Gemmatimonadetes Oyster Proteobacteria; Other Oyster Bacteroidetes; Flavobacteria Oyster Bacteroidetes; Bacteroidia Oyster Proteobacteria; Epsilonproteobacteria Oyster Chlorobi; SJA 28 Oyster IV Chlorobi; BSV19 Oyster NKB19; Other Oyster TM7; TM7 3 Oyster Acidobacteria; OS K Oyster Acidobacteria; Sva0725 Oyster Elusimicrobia; Elusimicrobia Water Acidobacteria; Solibacteres Water Tenericutes; Other Water GN02; GKS2 174 Oyster Planctomycetes; agg27 Oyster SR1; Unassigned Water

91 Acidobacteria; iii1 8 Oyster TM6; SJA 4 Oyster TM7; TM7 1 Oyster Bacteroidetes; Bacteroidia Water Proteobacteria; Epsilonproteobacteria Water Verrucomicrobia; Methylacidiphilae Water Verrucomicrobia; Verruco 5 Water SM2F11; Unassigned Water Acidobacteria; iii1 8 Water TM6; SBRH58 Water GN02; GKS2 174 Water TM6; SJA 4 Water Lentisphaerae; Lentisphaerae Oyster Cyanobacteria; 4c0d 2 Water Acidobacteria; Holophagae Water SC3; Unassigned Water Acidobacteria; Other Water SAR406; AB16 Oyster Proteobacteria; Gammaproteobacteria Oyster V SAR406; AB16 Water GN02; VC12 cl04 Oyster Proteobacteria; Gammaproteobacteria Water Bacteroidetes; Other Oyster Chloroflexi; Anaerolineae Oyster TG3; TG3 2 Oyster Proteobacteria; Alphaproteobacteria Oyster Verrucomicrobia; R76 B18 Water VI GN02; Other Water Chloroflexi; SOGA31 Oyster Bacteroidetes; Sphingobacteria Water Bacteroidetes; Other Water Proteobacteria; Betaproteobacteria Water Verrucomicrobia; Other Water Cyanobacteria; Other Water Cyanobacteria; Synechococcophycideae Water Verrucomicrobia; Opitutae Oyster Gemmatimonadetes; Gemmatimonadetes Water Cyanobacteria; Oscillatoriophycideae Water Chlorobi; Ignavibacteria Water Verrucomicrobia; Spartobacteria Water Proteobacteria; Other Water Planctoymcetes; Phycisphaerae Water

92 Chloroflexi; Chloroflexi Water Bacteria; Other Water Actinobacteria; Actinobacteria Water Cyanobacteria; Oscillatoriophycideae Oyster Cyanobacteria; S15B MN24 Oyster Cyanobacteria; 4c0d 2 Oyster TM7; Other Water Acidobacteria; Acidobacteria Oyster WS3; PRR 12 Water VII Lentisphaerae; Lentisphaerae Water WS6; SC72 Water Verrucomicrobia; Other Oyster SM2F11; Unassigned Oyster Verrucomicrobia; Opitutae Water Actinobacteria; Actinobacteria Oyster Acidobacteria; Acidobacteria Water Planctomycetes; agg27 Water OP8; OP8_2 Oyster Fusobacteria; Fusobacteria Oyster Bacteroidetes; Flavobacteria Water Proteobacteria; Betaproteobacteria Oyster Firmicutes; Clostridia Water Verrucomicrobia; Verrucomicrobiae Oyster Chloroflexi; Chloroflexi Oyster Fusobacteria; Fusobacteria Water Tenericutes; ML615J 28 Oyster Firmicutes; Other Oyster Nitrospirae; Nitrospira Water Tenericutes; Other Oyster Acidobacteria; Holophagae Oyster MVP 15; Unassigned Oyster WS3; PRR 12 Oyster MVP 15; Unassigned Water OP11; WCHB1 64 Water Planctoymcetes; Phycisphaerae Oyster VIII Verrucomicrobia; Spartobacteria Oyster Cyanobacteria; Synechococcophycideae Oyster OP11; WCHB1 64 Oyster Elusimicrobia; Elusimicrobia Oyster Spirochaetes; Spirochaetes Oyster Bacteria; Other Oyster Acidobacteria; OS K Water Other; Unassigned Oyster

93 Chlorobi; OPB56 Water GN02; BB34 Oyster Chlorobi; Ignavibacteria Oyster WS6; SC72 Oyster Proteobacteria; Alphaproteobacteria Water Nitrospirae; Nitrospira Oyster Proteobacteria; Deltaproteobacteria Oyster OP8; OP8_2 Water NKB19; Other Water Cyanobacteria; S15B MN24 Water Verrucomicrobia; Methylacidiphilae Oyster Cyanobacteria; Gloeobacterophycideae Oyster Proteobacteria; Deltaproteobacteria Water Bacteroidetes; Sphingobacteria Oyster Chlorobi; OPB56 Oyster GN04; GN15 Water GN04; GN15 Oyster GN02; Other Oyster Cyanobacteria; Other Oyster SR1; Unassigned Oyster Tenericutes; Mollicutes Oyster

94 Chapter 3

LOCATION, MORE THAN DISEASE, INFLUENCES OYSTER EXTRAPALLIAL FLUID MICROBIAL COMMUNITIES

3.1 Abstract

The eastern oyster, Crassostrea virginica, is a keystone species in estuarine environments along the east coast of North America. In addition to improving water quality as filter feeders, oysters are ecosystem engineers that provide important habitat and increase local biodiversity. However, C. virginica populations have been decimated since the last half-century by Perkinsus marinus, a non-native protozoan parasite and etiological agent of Dermo. Microbial communities play an important role in protecting their metazoan hosts from pathogens; yet, the interaction between the oyster microbiome and P. marinus is largely unknown. In this study, bacterial communities were characterized from the extrapallial fluid of healthy (P. marinus- negative) and diseased (P. marinus-positive) oysters across four sites in the Chesapeake Bay. Bacterial community composition was most strongly shaped by location. However, healthy and diseased oyster EF communities were significantly different, and ten taxonomic lineages were significantly associated with health status independent of sample location. Surprisingly, these taxonomic lineages did not include conventional opportunistic pathogens (e.g. Vibrio spp.), and no difference in the alpha diversity between healthy and diseased oysters was observed. The potential role of these taxa in disease resistance/susceptibility requires further investigation in order to improve oyster conservation efforts in the 21st century.

95 3.2 Introduction

The eastern oyster (Crassostrea virginica) is a keystone species in estuarine environments along the east coast of North America. As prolific filter feeders, an individual adult oyster can improve the quality of up to 50 gallons of water daily (NOAA), thus enhancing conditions for growth of submerged aquatic vegetation (Grabowski and Peterson 2007). Oysters also serve as “ecosystem engineers” since oyster reefs provide key habitat for estuarine species (Lenihan and Peterson 1998) and promote increased biodiversity (Rodney and Paynter 2006; Stunz, Minello et al. 2010). In addition to their environmental value, oysters are part of the fastest growing source of animal food production (Mathiesen 2012), and molluscan aquaculture production worldwide has doubled over the past few decades (Romero, Novoa et al. 2012). Despite their environmental and economic importance, oyster habitat and biomass have declined on average 64% and 88%, respectively, since the late 19th century (Zu Ermgassen, Spalding et al. 2012). This decline, largely due to historical anthropogenic factors like overharvesting and habitat degradation (Rothschild, Ault et al. 1994), has contributed to impaired ecosystems (Newell 1988). Recently, C. virginica has faced new challenges to its survival, including the looming threat of ocean warming/acidification and the proliferation of pathogens. Pathogens are a major cause of mortality among oysters. Bacterial pathogens, most notably members of the genus Vibrio, impact larvae and juveniles (Bachere 2003) and can result in production declines in excess of 50% (Maloy, Ford et al. 2007; Elston, Hasegawa et al. 2008). Adult oysters are often not impacted by these bacterial pathogens (Paillard, Le Roux et al. 2004); instead, disease and mortality in adult oysters are most greatly influenced by protozoans (Dorrington and Gomez-Chiarri 2008). Since the 1950s, two protozoan pathogens – Haplosporidium nelsoni (the

96 etiological agent of oyster disease MSX) and Perkinsus marinus (the etiological agent of oyster disease Dermo) – have decimated C. virginica populations. For example, oyster mortality rates reached 47% during an MSX epizootic in the Delaware Bay in 1985 (Powell, Ashton-Alcox et al. 2008). In 2002 the survival rate in the Chesapeake Bay declined to only 42%, largely due to the combined effects of MSX and Dermo (Tarnowski 2012). P. marinus has also had a measurable impact on mortality rates in the Delaware Bay since its detection there in 1990. Oyster mortality rates were on average 5-10% annually prior to detection; since detection, mortality rates have risen to >15% annually and have frequently surpassed 20% (Powell, Ashton-Alcox et al. 2008). Key to improving oyster restoration efforts is a better understanding of the factors that influence oyster fitness, including disease susceptibility/resistance. This is particularly important in the context of global warming since disease outbreaks due to Vibrio spp. have increased with rising ocean temperatures (Vezzulli, Colwell et al. 2013). Likewise, warmer water temperatures are believed to be at least partially responsible for the observed increase in P. marinus distribution along the North American east coast (Burge, Mark Eakin et al. 2014). One factor that may play a role in oyster disease susceptibility/resistance is the oyster microbiome and the interaction of commensal bacteria with potential pathogens. Bacteria isolated from scallops, C. gigas rearing water, and the inner shell of C. virginica have all been experimentally shown to increase survival against Vibrio pathogens (Riquelme, Hayashida et al. 1996) (Nakamura, Takahashi et al. 1999) (Gibson, Woodworth et al. 1998) (Lim, Kapareiko et al. 2011).

97 Less well understood is the interaction between protozoan pathogens and the oyster microbiome in adult oysters. However, a few studies have identified links between protozoan pathogens and bacteria. A Psychroserpens sp. (Flavobacteriaceae) was specifically associated with Neoparamoeba pemaquidensis, the etiological agent of amoebic gill disease, in Atlantic salmon (Salmo salar L.) (Bowman and Nowak 2004). Enhanced infectivity of N. pemaquidensis was observed in vivo when co- infected with a related Winogradskyella sp. (Flavobacteria) (Embar-Gopinath, Butler et al. 2005). Likewise, co-infection with Gram-negative bacteria increased the virulence of Entamoeba histolytica in vitro (Bracha and Mirelman 1984), and bacterial communities in the digestive of Sydney rock oysters (Saccostrea glomerata) infected by Marteilia sydneyi were dominated by a single Alphaproteobacteria phylotype (Green and Barnes 2010). In the latter case, the authors concluded that mortality may be attributable to secondary infection by bacteria in at least some cases of M. sydneyi infection (Green and Barnes 2010). High parasite burdens may increase susceptibility to bacterial pathogens through impaired immunocompetence in fish species like the common dentex (Dentex dentex) (Company, Sitja-Bobadilla et al. 1999) and gilthead seabream (Sparus aurata) (Sitja-Bobadilla, Pujalte et al. 2006). It is unknown whether bacterial groups are associated with P. marinus infection in C. virginica that may synergistically increase mortality. In this study, the extrapallial fluid (EF, i.e. the fluid between the oyster mantle and interior shell) microbiome was characterized from oysters in the Chesapeake Bay. Oysters from sites with a high historical prevalence (Tangier Sound) and low historical prevalence (Choptank River) of P. marinus were compared to identify differences in community composition between locations and between healthy and infected individuals.

98 3.3 Materials and Methods

3.3.1 Oyster sample collection

Oysters were collected from four Maryland Department of Natural Resources (DNR) historical disease survey sites in the Chesapeake Bay in October 2012 (Table 3.1). Oysters were sampled from Tangier Sound locations Marumsco (MA) and Old Woman’s Leg (OWL) on Oct. 9, 2012. Oysters from Choptank River locations Royston (RO) and Tilghman Wharf (TW) were sampled on Oct. 24, 2012. Oysters were sampled by dredge as described in (Tarnowski 2014). Each oyster was scrubbed with 70% ethanol prior to extrapallial fluid (EF) extraction. A hole was drilled into the posterior end of the oyster at the interface between valves with a 3/32-inch drill bit. EF was extracted using a 5mL syringe with 23G needle, snap-frozen in liquid nitrogen (LN2), and stored at −80C until further processing.

3.3.2 Screening for P. marinus infection

Prevalence and intensity of P. marinus infection was determined using the Ray’s fluid thioglycollate medium assay on tissue assays (Ray 1966). Intensity was scored on a 0-7 Cooperative Oxford Laboratory (COL) scale.

3.3.3 Bacterial DNA isolation

Oyster EF samples were thawed on ice with RNAlater-Ice (Ambion). Thawed samples were combined with sterile 1x PBS buffer at a 1:10 ratio and rocked for 30 minutes at 30°C to improve filtration. Samples were then filtered through a Millipore Sterivex 0.22µm filter unit. DNA was extracted from each of the filters as previously described (Crump, Kling et al. 2003) with amendments. Briefly, proteinase-K (20mg/mL) and lysozyme (100mg/mL) were combined with DNA Extraction Buffer

99 (DEB: 100 mM Tris buffer (pH 8), 100 mM NaEDTA (pH 8), 100mM phosphate buffer (pH 8), 1.5 M NaCl, 1% CTAB) and added to the filter. Filters were incubated at 37°C for 30 minutes and subjected to three freeze-thaw cycles of -80°C and 37°C, followed by incubation at 37°C for 30 minutes. DNA was extracted with Trizol (Ambion) following the manufacturer’s protocol and re-suspended in EB buffer.

3.3.4 16S amplification, barcoding, and sequencing

The 16S rRNA gene was PCR amplified using universal primers 16S rRNA For (5’ - AGAGTTTGATCCTGGCTCAG - 3’) and 16S rRNA Rev (5’ - ACGGCTACCTTGTTACGACTT - 3’) (Integrated DNA Technologies). Bacterial DNA (1µL) was combined with 10X buffer (1X final concentration), dNTPs (0.25mM each), forward and reverse primers (0.1M final concentration each), and TaKaRa Ex Taq DNA Polymerase (1.25 U) to a final volume of 25µL. PCR amplification of samples was performed using the following conditions: 95°C for 5 minutes; 33 cycles of 95°C for 30 seconds, 52°C for 30 seconds, 72°C for one minute; 72°C for 7 minutes.

3.3.5 Adapter ligation and sequencing

16-mer adapter sequences with 3’-T overhangs were AT ligated to 16S amplicons. Briefly, amplicons were cleaned and concentrated (Zymo Clean and Concentrator-5). DNA was quantified with the Qubit Fluorometer High Sensitivity assay (Life Technologies). 50ng of DNA per sample was combined with annealed dsDNA adapters (15µM) and 3 units Ligase (Promega) and incubated for 12 hours at 15°C. Ligated amplicons were column cleaned (Zymo Clean and Concentrator-5). Subsequently, the entire volume (10µL) of ligated amplicon product was PCR

100 amplified for 5 cycles using the barcode as the primer to enrich for ligated products prior to sequencing. Finally, amplified products were gel-purified (QIAquick Gel Extraction kit) and 50ng DNA from each sample was pooled for sequencing on the PacBio RS.

3.3.6 Amplicon filtering and taxonomic assignment

PacBio circular consensus (CCS) reads with >3x coverage were screened for the 16S rRNA primers at 85% identity. Amplicons were subsequently deconvoluted by barcode. Barcodes were iteratively screened at 100%, 95%, and 90% identity. Amplicons were clustered with Uclust (Edgar 2010) at 97% identity. OTU taxonomies were assigned in QIIME (Caporaso, Kuczynski et al. 2010) using the Ribosomal Database Project (RDP) classifier (Wang, Garrity et al. 2007) against the Greengenes reference database (McDonald, Price et al. 2011). One healthy oyster from OWL was considered an outlier based on Principle Coordinates Analysis and UniFrac distance and was excluded from analyses.

3.3.7 Alpha diversity of EF communities

Alpha diversity metrics Chao1, observed species, phylogenetic diversity, and Shannon diversity were calculated in QIIME (Caporaso, Kuczynski et al. 2010) with ten rarefactions at 25 sequence intervals. Samples were grouped by P. marinus infection status (healthy vs. diseased; Fig. 3.1A), site (MA, OWL, RO, TW; Fig. 3.1B), general geography (Tangier Sound vs. Choptank River; Fig. 3.1C), and a combination of location and infection status (e.g. OWL_Diseased; Fig. 3.1D). The mean for each group at a given sampling depth was plotted. Rarefaction curves were plotted by fitting logarithmic curves to the mean values at each sample depth (Fig.

101 3.1). Statistical differences in alpha diversity (α = 0.05) were queried in QIIME

(Caporaso, Kuczynski et al. 2010) by 999 Monte Carlo permutations at a sample depth of 325 sequences.

3.3.8 Beta diversity of EF communities

Bacterial communities were queried for statistical differences by site (MA, OWL, RO, TW), P. marinus infection status (positive vs. negative), and general geography (Tangier Sound vs. Choptank River) via analysis of similarity (ANOSIM) (Clarke 1993) and adonis tests with 999 permutations. Samples were grouped by a combination of location and infection status (e.g. OWL_Diseased) and clustered by Unweighted Pair Group Method with Arithmetic Mean (UPGMA) hierarchical clustering with 100 jackknife replicates at a sampling depth of 1,150 sequences.

3.3.9 Identification of enriched bacterial taxa

Bacterial taxa significantly (p < 0.05) associated with P. marinus infection status (Healthy vs. Diseased), site, or general geography (Tangier Sound vs. Choptank River) were identified in QIIME (Caporaso, Kuczynski et al. 2010) by Kruskal-Wallis tests. Bacterial taxa that were significantly associated with sample site or general geography were discarded in order to identify taxa significantly associated with P. marinus infection status independent of geographic influence.

102 3.4 Results

3.4.1 Alpha diversity of oyster EF bacterial communities

The Chao1 estimate of richness was similar between healthy (P. marinus- negative) and diseased (P. marinus-positive) oysters from EF from Tangier Sound and the Choptank River (Fig. 3.1A). Tangier Sound site Old Woman’s Leg (OWL) had the greatest observed Chao1 diversity (Fig. 3.1B), though the difference was not significant. No difference was observed in the richness of EF communities between sites (Fig. 3.1B) or geographic locations (Fig. 3.1C). Likewise, there was no difference in community richness between healthy or diseased oysters within a given site (Fig. 3.1D). Observed species, phylogenetic diversity, and Shannon diversity were also similar for all groups (Figs. 3.2-4).

3.4.2 Differences in oyster EF bacterial community composition by geography

Oyster EF bacterial communities were significantly different (p < 0.05) between sites and between geographic locations by both adonis and ANOSIM analyses (Table 3.2). Oysters from a given site were also significantly (p < 0.05) more similar to each other than to oysters from other sites by weighted UniFrac distance (Fig. 3.5A). When grouped by geography, oysters from the Choptank River were more similar to each other than to Tangier Sound oysters (Fig. 3.5A). However, Tangier

Sound oysters were as different from each other as they were to Choptank River oysters (Fig. 3.5A). A total of 89 taxa (RDP-designated class level) were identified in oyster EF samples. However, the majority of the bacterial community was comprised of only a few abundant taxa (Fig. 3.6). The ten most abundant taxa comprised 87% of the community for Choptank River samples and 83% for Tangier Sound samples (Figs.

103 3.6A, 3.7A). At this broad taxonomic level, there were very few differences between oysters from different geographic locations (Fig. 3.6A, Table 3.3), and only six of the 89 identified taxa (class level) were significantly (p < 0.05) different between geographic locations (Table 3.3). The number of taxa significantly associated with one geographic location increased at finer taxonomic resolution (Fig. 3.8, Table 3.3), and more taxa were significantly associated with Tangier Sound samples than Choptank River samples (Table 3.3). Despite few differences at the class level, most taxa significantly associated with one geographic location belonged to one of the ten most abundant RDP-designated classes (Fig. 3.6A, Table 3.3). In particular, members of the Actinobacteria, Planctomycetes, and Betaproteobacteria were significantly associated with Choptank River samples, while Tangier Sound samples were associated with a broader range of taxonomic diversity, including members of the Acidobacteria, Cyanobacteria, Deltaproteobacteria, and Gammaproteobacteria (Table 3.3). A total of 35 OTUs differed between geographic locations (Fig. 3.8, Table 3.3).

3.4.3 Differences in oyster EF bacterial community composition by P. marinus infection status

No significant difference in unweighted bacterial community composition was observed based on P. marinus infection status (Table 3.2). In contrast, weighted bacterial communities were significantly different based on P. marinus infection status, though this grouping was less strong than geography or site (Table 3.2). Overall, diseased oyster bacterial communities were marginally (p = 0.10) more similar to each other (i.e. had lower UniFrac distances) than healthy oysters were to each other (Fig. 3.5B, left). Likewise, diseased oysters were more similar to each other than to healthy oysters across all sample sites (Fig. 3.5B, left). In contrast, healthy

104 oysters across all sample sites were no more similar to each other than to diseased oysters (Fig. 3.5B, left). Within a given site, bacterial communities of healthy oysters were as similar to each other as diseased oysters were to each other (Fig. 3.5B, right). However, UniFrac distances between diseased oysters from different sites were significantly lower than UniFrac distances between healthy and diseased oysters from different sites (Fig. 3.5B, right). The ten most abundant taxa among healthy and diseased oysters comprised 82% and 86% of the community, respectively (Figs. 3.6B, 3.7B). Very few taxa were different between healthy and diseased oyster samples, particularly at broader taxonomic levels (Figs. 3.6B, 3.8, Table 3.4). The number of taxa associated with healthy or diseased oysters increased with taxonomic resolution but was never as high as taxa associated with site or geography (Fig. 3.8). In total, 22 bacterial taxa representing ten taxonomic lineages were significantly associated with healthy or diseased oysters independent of site or geography (Table 3.4). Of the ten bacterial lineages, six were associated with healthy oysters and four with diseased oysters. Lineages associated with healthy oysters included the Proteobacteria (Alphaproteobacteria and Gammaproteobacteria) and Spirochaetes (Table 3.4). Diseased oysters were associated with Planctomycetes, Spirochaetes, and Mollicutes (Table 3.4).

3.5 Discussion

Bacterial communities of marine bivalves are distinct from the surrounding water (Colwell and Liston 1960; Kueh and Chan 1985; Pujalte, Ortigosa et al. 1999; La Valley, Jones et al. 2009; Thomas, Wafula et al. 2014), but little is known about the impact of geography and parasite infection on the oyster microbiome. Studies

105 examining the impact of geography on the oyster microbiome have yielded mixed results. Stomach and gut microbiomes of C. virginica displayed site-specific profiles (King, Judd et al. 2012), as did bacterial communities from C. virginica whole oyster homogenates (La Valley, Jones et al. 2009). In contrast, no difference was observed between sites in the gill microbiome of C. gigas (Wegner, Volkenborn et al. 2013). An independent study also found no difference in C. gigas bacterial communities, but site- specific differences were observed for C. corteziensis (Trabal, Mazon-Suastegui et al. 2012). In this study, C. virginica EF bacterial communities significantly differed by both sample site and local geography, with sample site being the stronger grouping (Table 3.2). Therefore, it appears that geographic-related microbiome variability in oysters may be at least partially species-specific. Most of the taxonomic differences between oysters from different locations occurred at lower taxonomic levels (finer taxonomic resolution) (Fig. 3.8), while higher taxonomic levels were similar between oysters from different locations (Fig. 3.6). This suggests that the microenvironment of C. virginica EF selects for similar (broad) taxonomic – and likely functional – diversity even against different backdrops of environmental diversity. Non- environmental factors like host genetics may also play a role. Interestingly, rare bacterial groups correlated with genetic relatedness in C. gigas (Wegner, Volkenborn et al. 2013). Although the genetic relatedness of oysters at each site was not compared in this study, it is possible that oysters from the same sample site were more closely related to each other than to oysters from other sites. Oyster EF bacterial communities also differed between healthy oysters and those infected with P. marinus, but P. marinus infection status was the weakest grouping of samples (Table 3.2, Fig. 3.9). Within a sample site, diseased oyster

106 bacterial communities and healthy oyster bacterial communities were as similar to each other as they were amongst themselves (Fig. 3.5B). However, the bacterial communities of diseased oysters from different sites were more similar to each other than they were to healthy oysters from different sites, and diseased oysters overall were more similar to each other than to healthy oysters (Fig. 3.5B). In contrast, healthy oysters were no more similar to each other than to diseased oysters (Fig. 3.5B). Thus, it appears that a greater variation in bacterial diversity exists among healthy oysters, while P. marinus infection alters bacterial diversity in a similar manner regardless of the initial diversity present. The extent of this alteration remains unclear. P. marinus induces changes in the oyster immune response, including the reduction of apoptosis (Sunila and LaBanca 2003; Goedken, Morsey et al. 2005), which is an important defense against P. marinus infection (Hughes, Foster et al. 2010). Changes in lectin expression, oxidative stress regulation, and up-regulation of histone components with antimicrobial activity were also observed in P. marinus-infected oysters (Wang, Peatman et al. 2010). The serine protease of P. marinus also reduced the ability of oyster hemocytes to remove Vibrio vulnificus in vitro (Tall, La Peyre et al. 1999). Furthermore, P. marinus infection is believed to initiate at the mantle epithelium where high parasite burdens are observed (Allam, Carden et al. 2013), and virulence may be increased through contact with the pallial mucus (Espinosa, Corre et al. 2014). Therefore, we hypothesized that the proximity and possible interaction of P. marinus cells with EF coupled with changes in the oyster immune response as a result of P. marinus infection would alter bacterial diversity in the EF of diseased oysters. Specifically, we expected copiotrophic organisms and opportunistic pathogens (e.g. Vibrio spp.) to be more prevalent in

107 diseased oysters. Surprisingly, there was no difference in the alpha diversity (richness and evenness) of healthy and diseased oyster EF bacterial communities (Figs. 3.1-4), and very few differences in taxonomic composition were observed, particularly at higher taxonomic levels (Fig. 3.8). By comparison, changes in bacterial diversity have been observed in other diseased marine invertebrates, particularly corals. Corals with White Plague Disease (Sunagawa, DeSantis et al. 2009; Roder, Arif et al. 2014; Roder, Arif et al. 2014), White Band Disease (Pantos and Bythell 2006), and Yellow Band Disease (Closek, Sunagawa et al. 2014) displayed increased bacterial diversity and a greater abundance of opportunistic bacteria relative to their healthy counterparts. In contrast, diversity in the digestive gland of Sydney Rock oysters (Saccostrea glomerata) infected with protozoan parasite Martellia sydneyi was much lower than in healthy oysters and was dominated by a single Alphaproteobacteria OTU (Green and Barnes 2010). Host response to pathogen exposure may play a role in the increased bacterial diversity observed in diseased corals and the lack of change in bacterial diversity observed in diseased oysters. Whereas antimicrobial peptides (AMPs) were up- regulated in P. marinus-infected oysters (Wang, Peatman et al. 2010), transcriptome analysis of corals with Yellow Band Disease revealed reduced expression of defense- related genes (Closek, Sunagawa et al. 2014). Expression of these AMPs may help to compensate for suppressed hemocyte activity. The nature of the pathogen (bacterial vs. protozoan) likely had little influence on the differences in bacterial communities between healthy and diseased oysters and coral, as a recent analysis of bacterial communities in oyster (C. gigas) hemolymph found little change in community composition at higher taxonomic levels when oysters were challenged with the

108 virulent Vibrio sp. strain D29w (Lokmer and Mathias Wegner 2015). A lack of gross taxonomic differences between healthy and diseased oysters may have also been related to the virulence of P. marinus in oyster samples. Although highly virulent P. marinus strains suppressed apoptosis 24h post-infection, apoptosis was less inhibited by intermediate and low-virulence strains (Hughes, Foster et al. 2010). A greater difference between healthy and diseased oysters may also have been observed if samples had been collected later in the season. P. marinus-induced mortality typically occurs in late summer (August and September) (Burreson and Ragone Calvo 1996), with surviving oysters harboring sub-lethal infections over winter (Andrews 1996). It is possible that the bacterial diversity of healthy and diseased oysters would become less similar over time as sub-lethal infections persisted. Despite the lack of drastic differences between healthy and diseased oysters, there were several taxa associated with either healthy or diseased individuals independent of sample site or geography. Perhaps most intriguing among these were two different members of the Spirochaetes. Spirochetes were detected in the crystalline styles of bivalves over a century ago, and these bacteria appear to form species-specific symbioses with their hosts (Husmann, Gerdts et al. 2010). Phylotypes isolated from marine bivalves include members of the Spirochaetaceae, (Cristispira spp. and Spirochaeta spp.) or Brachyspira spp. (Husmann, Gerdts et al. 2010). In this study, the genus Spirochaeta was significantly (p < 0.05) associated with healthy oysters. This group was present in 6/8 healthy oysters but only 1/9 diseased oysters. A second genus – assigned by the RDP classifier to the Brachyspiraceae – was comprised of OTUs that shared greatest homology (97-98% nucleotide identity) with an uncultured spirochete from the digestive gland of Sydney Rock oysters (Saccostrea

109 glomerata) (Green and Barnes 2010). This genus was found in both healthy (4/8) and diseased (7/9) oysters but was on average eight-fold more abundant in diseased oysters. Thus, it appears that P. marinus infection disturbed the balance of this symbiosis. It is unclear what role spirochetes play in the health of their bivalve host, but it has been hypothesized that they may play a role in extracellular enzymatic digestion (Husmann, Gerdts et al. 2010), and a disruption in this balance may alter the availability of energy for the oyster and/or P. marinus. Other taxa associated with P. marinus-infected oysters included the family Pirellulaceae; an OTU most similar to Microcella spp. and uncultured marine isolates; and the family Entomoplasmataceae (Table 3.4). The latter consist of insect or plant- associated mycoplasma. Mycoplasma cause disease in a number of hosts, including mammals, insects, and plants (Kumar 2012). Interestingly, a symbiotic relationship between the parasitic protozoan Trichomonas vaginalis and Mycoplasma huminis has been described, and other instances of endosymbionts in free-living protozoa are known (Dessi, Delogu et al. 2005). Given that members of the Entomoplasmataceae were only detected in P. marinus-infected individuals (4/9), it is likely that these mycoplasma are not normally members of the EF community and may be associated with P. marinus. No mycoplasma-P. marinus relationship has been reported; however, P. marinus and other Chromalveolates contain remnant plastids (Matsuzaki, Kuroiwa et al. 2008), indicating that there is a precedent for symbiosis in P. marinus. Further investigation is required to determine whether the mycoplasma associated with diseased oysters were simply opportunists or more intimately linked with P. marinus infection.

110 The most intriguing taxon among those associated with healthy oysters was an OTU most similar by BLASTn homology to Phaeobacter gallaeciensis (96% nucleotide identity). Phaeobacter spp. have been noted as probiotics in aquaculture and have been shown to protect scallop (Pecten maximus), flat oyster (Ostrea edulis), Pacific oyster (C. gigas), and C. virginica larvae against Vibrio pathogens (Kesarcodi- Watson, Miner et al. 2012; Karim, Zhao et al. 2013). The protection provided by Phaeobacter spp. is due in part to the production of tropodithietic acid (TDA) (Porsby, Webber et al. 2011), a broad-spectrum antibacterial produced by several Roseobacter lineages (Brinkhoff, Bach et al. 2004). The mechanism by which TDA inhibits or kills bacteria is not known but may involve targeting the cell envelope (Porsby, Webber et al. 2011). However, TDA displayed no antagonistic activity against model eukaryotic organisms (Rabe, Klapschinski et al. 2014) and would not be expected to directly protect oysters against P. marinus. It is possible that oysters protected against bacterial pathogens would be less susceptible to P. marinus. The order Pseudomonadales was also associated with healthy oysters. Pseudomonads have long been known to inhabit oysters and are among the most dominant bacteria cultivated from oysters and other bivalves, comprising up to 50% of all cultivated phylotypes (Colwell and Liston 1960; Lovelace, Tubiash et al. 1967; Murchelano and Brown 1968; Kueh and Chan 1985; Olafsen, Mikkelsen et al. 1993;

Beleneva, Zhukova et al. 2003). Recently, Pseudomonas isolates from oyster mantle (extrapallial) fluid collected in the Gulf of Mexico indicated that these bacteria were crucial to hydrocarbon degradation and are likely playing a key role in the post Deepwater Horizon oil spill ecosystem (Chauhan, Green et al. 2013; Thomas, Wafula et al. 2014). In this study, members of the Pseudomonadales were on average twelve-

111 fold more abundant in healthy oysters than oysters infected with P. marinus; yet, the Pseudomonadales comprised just 1.6% of the community in healthy oysters. This highlights the disparity between cultivation-dependent and independent characterizations of bacterial communities. A previous examination of bacterial communities in the Chilean oyster (C. corteziensis) estimated that only 0.01% of bacteria in oysters were cultivable (Romero, Garcia-Varela et al. 2002). Similarly, Vibrio spp. are among the most abundant bacteria cultivated from oysters but accounted for an average of just 0.8% of the EF community. Vibrio spp. abundance was greater in the higher salinity TS samples, but was not associated with P. marinus infection. Thus, the most dominant bacterial phylotypes identified by cultivation, which are often regarded as opportunistic pathogens, comprised a small fraction of the EF microbiome and did not increase in abundance in immunocompromised individuals. Interestingly, while heat stress decreased overall microbiome diversity in C. gigas, no net increase in potential pathogens was observed (Wegner, Volkenborn et al. 2013). It is unclear whether the same response that prevented pathogen proliferation during heat stress was also important in preventing opportunists from increasing in immunocompromised, P. marinus-infected oysters, but this response could include AMP production. Alternatively, opportunistic pathogens could be limited even in immunocompromised individuals by TDA or other antimicrobial compounds produced by other bacterial community members, and bacteriophage predation could also keep opportunistic pathogen populations at low levels.

112 FIGURES

A. 200 2 B. 250 R = 0.76 2 180 R = 0.84 R2 = 0.97 160 200 R2 = 0.79 140 2 R = 0.96 120 150 R2 = 0.49 100 MA 80 Healthy 100 Chao1 Index Chao1 Index Chao1 OWL 60 Diseased RO 40 50 TW 20 0 0 0 200 400 600 800 1000 1200 0 200 400 600 800 1000 1200 # Sequences # Sequences

C. 250 D. 250 R2 = 0.65 2 0.83 200 2 200 R = R = 0.82 2 R = 0.79 2 2 R = 0.96 R = 0.91 R2 = 0.38 150 150 R2 = 0.89 MA Diseased 100 CR 100 OWL Diseased Chao1 Index Chao1 Index Chao1 TS OWL Healthy 50 50 RO Diseased RO Healthy TW Healthy 0 0 0 200 400 600 800 1000 1200 0 200 400 600 800 1000 1200 # Sequences # Sequences Figure 3.1: Chao1 estimates of richness for oyster EF bacterial communities from two sites in the Choptank River (Royston and Tilghman Wharf) and two sites in Tangier Sound (Marumsco and Old Woman’s Leg). Rarefaction curves are logarithmic regressions of the mean Chao1 value at each sample depth from 10 rarefactions. A) Richness of bacterial communities from healthy (P. marinus-negative) or diseased (P. marinus-positive) oysters. B) Richness of bacterial communities from oysters by site. C) Richness of bacterial communities from oysters sampled at the Choptank River (CR) or Tangier Sound (TS). D) Richness of bacterial communities from oysters according to sample site and P. marinus infection status. Marumsco (MA); Old Woman’s Leg (OWL); Royston (RO); Tilghman Wharf (TW).

113 A. 20 B. 20 18 18 R! = 0.84 R! = 0.79 R! = 0.76 16 16 R! = 0.97 R! = 0.96 14 14 12 12 R! = 0.49 10 10 HealLog.(Healthy)thy MA 8 8 Log.(MA Mean) OWL 6 DiLog.(Diseased)seased 6 Log.(OW Mean) RO

Phylogenetic Diversity Phylogenetic Diversity Phylogenetic Log.(RO Mean) 4 4 TWLog.(TW Mean) 2 2 0 0 0 200 400 600 800 1000 1200 0 200 400 600 800 1000 1200 # Sequences # Sequences D. C. 20 20 R! = 0.65 18 18 R! = 0.82 R! = 0.79 R! = 0.83 16 16 R! = 0.96 R! = 0.91 14 14 R! = 0.89 12 12 R! = 0.38 MA Diseased 10 10 OWL Diseased 8 CR Log.(CR) 8 Log.(MA_Diseased) OWL Healthy TS RO Diseased 6 Log.(TS) 6 Log.(OWL_Diseased) RO Healthy Phylogenetic Diversity Phylogenetic 4 Diversity Phylogenetic 4 Log.(OWL_Healthy) Log.(RO_Diseased) TW Healthy 2 2 Log.(RO_Healthy) 0 0 Log.(TW_Healthy) 0 200 400 600 800 1000 1200 0 500 1000 1500 # Sequences # Sequences

Figure 3.2: Phylogenetic diversity (PD) for oyster EF bacterial communities from two sites in the Choptank River (Royston and Tilghman Wharf) and two sites in Tangier Sound (Marumsco and Old Woman’s Leg). Rarefaction curves are logarithmic regressions of the mean PD value at each sample depth from 10 rarefactions. A) PD of bacterial communities from healthy (P. marinus-negative) or diseased (P. marinus-positive) oysters. B) PD of bacterial communities from oysters by site. C) PD of bacterial communities from oysters sampled at the Choptank River (CR) or Tangier Sound (TS). D) PD of bacterial communities from oysters according to sample site and P. marinus infection status. Marumsco (MA); Old Woman’s Leg (OWL); Royston (RO); Tilghman Wharf (TW).

114 A. 7 B. 8

R! = 0.76 6 7 R! = 0.84 R! = 0.97 6 5 R! = 0.79 R! = 0.96 5 R! = 0.49 4 4 3 HealLog.(Healthy)thy MALog.(MA Mean) 3

Shannon Index Index Shannon DiLog.(Diseased)seased Index Shannon Log.(OW Mean) 2 2 ROLog.(RO Mean) TW 1 1 Log.(TW Mean)

0 0 0 200 400 600 800 1000 1200 0 200 400 600 800 1000 1200 # Sequences # Sequences D. C. 7 8 R! = 0.65 R! = 0.82 7 6 R! = 0.96 R! = 0.91 6 R! = 0.79 5 R! = 0.83 R! = 0.89 5 MA Diseased 4 R! = 0.38 4 OWL Diseased 3 CRLog.(Series1) Log.(MA_Diseased) OWL Healthy 3

Shannon Index Index Shannon TSLog.(Series2) Index Shannon Log.(OWL_Diseased) RO Diseased 2 2 Log.(OWL_Healthy) RO Healthy TW Healthy 1 1 Log.(RO_Diseased) Log.(RO_Healthy) 0 0 Log.(TW_Healthy) 0 200 400 600 800 1000 1200 0 500 1000 1500 # Sequences # Sequences

Figure 3.3: Shannon index of diversity for oyster EF bacterial communities from two sites in the Choptank River (Royston and Tilghman Wharf) and two sites in Tangier Sound (Marumsco and Old Woman’s Leg). Rarefaction curves are logarithmic regressions of the mean Shannon index value at each sample depth from 10 rarefactions. A) Diversity of bacterial communities from healthy (P. marinus- negative) or diseased (P. marinus-positive) oysters. B) Diversity of bacterial communities from oysters by site. C) Diversity of bacterial communities from oysters sampled at the Choptank River (CR) or Tangier Sound (TS). D) Diversity of bacterial communities from oysters according to sample site and P. marinus infection status. Marumsco (MA); Old Woman’s Leg (OWL); Royston (RO); Tilghman Wharf (TW).

115 A. 180 B. 250

160 R! = 0.76 R! = 0.97 200 R! = 0.84 140

120 R! = 0.79 150 100 R! = 0.96 R! = 0.49 80 HealLog.(Healthy)thy 100 MALog.(MA Mean) 60 DiLog.(Diseased)seased OWLLog.(OW Mean) 50 RO

ObservedSpecies Log.(RO Mean) 40 ObservedSpecies TWLog.(TW Mean) 20 0 0 0 200 400 600 800 1000 1200 0 200 400 600 800 1000 1200 -20 -50 # Sequences # Sequences D. C. 200 250

R! = 0.82 ! = 0.65 200 R 150 R! = 0.91 R! = 0.96 R! = 0.79 150 R! = 0.83 MA Diseased 100 R! = 0.38 R! = 0.89 OWL Diseased CRLog.(Series1) 100 Log.(MA_Diseased) OWL Healthy 50 TSLog.(Series2) Log.(OWL_Diseased) RO Diseased 50 Log.(OWL_Healthy) RO Healthy ObservedSpecies ObservedSpecies Log.(RO_Diseased) TW Healthy 0 0 Log.(RO_Healthy) 0 200 400 600 800 1000 1200 0 500 1000 1500 Log.(TW_Healthy) -50 -50 # Sequences # Sequences

Figure 3.4: Observed species (OTUs) for oyster EF bacterial communities from two sites in the Choptank River (Royston and Tilghman Wharf) and two sites in Tangier Sound (Marumsco and Old Woman’s Leg). Rarefaction curves are logarithmic regressions of the mean number of observed species at each sample depth from 10 rarefactions. A) Observed species from healthy (P. marinus-negative) or diseased (P. marinus-positive) oysters. B) Observed species from oysters by site. C) Observed species from oysters sampled at the Choptank River (CR) or Tangier Sound (TS). D) Observed species from oysters according to sample site and P. marinus infection status. Marumsco (MA); Old Woman’s Leg (OWL); Royston (RO); Tilghman Wharf (TW).

116 A. B. 0.35 * * 0.33

0.31 0.35 * * 0.29 0.33 0.27 0.31 0.29 0.25 0.27 0.23 0.25 0.21 0.23 0.21 0.19

0.19 Unifrac Distance PairwiseWeighted 0.17 0.17 0.15 0.15 Pairwise Weighted Unifrac Distance PairwiseWeighted

Within TS CR vs. TS Healthy Within SiteBetween Sites Within CR Diseased Group Comparisons Healthy vs. Diseased Healthy (WithinDiseased Site) (Within Site) Healthy (BetweenDiseased Sites) (Between Sites)

Healthy vs. Diseased (Within Site) Healthy vs. Diseased (Between Sites) Group Comparisons Figure 3.5: Weighted UniFrac distances between oyster EF bacterial communities. A) UniFrac distances between oysters by site (left) and geography (right). B) UniFrac distances between healthy (Perkinsus marinus-negative) and diseased (P. marinus- positive) oysters. Choptank River (CR); Tangier Sound (TS). Error bars are SE.

A. Phylum Class B. Phylum Class

Spirochaetes Spirochaetes Spirochaetes Spirochaetes

Gammaproteobacteria Gammaproteobacteria

Proteobacteria Deltaproteobacteria Proteobacteria Deltaproteobacteria

Alphaproteobacteria Alphaproteobacteria Planctomycetes Planctomycetia Planctomycetes Planctomycetia OP8 OP8_2 TS OP8 OP8_2 Diseased Firmicutes Clostridia CR Healthy Cyanobacteria Synechococcophycideae Cyanobacteria Synechococcophycideae Flavobacteriia Flavobacteriia Bacteroidetes Bacteroidetes Bacteroidia Bacteroidia

Actinobacteria Actinobacteria Actinobacteria Actinobacteria Acidimicrobiia * Acidimicrobiia

0% 5% 10% 15% 20% 25% 30% 35% 0% 5% 10% 15% 20% 25% 30% 35% % of Community % of Community Figure 3.6: The most abundant bacterial taxa (RDP-designated class level) in oyster EF from the Choptank River and Tangier Sound. A) Abundance of bacterial taxa between geographic locations (Choptank River vs. Tangier Sound). B) Abundance of bacterial taxa between healthy (Perkinsus marinus-negative) and diseased (P. marinus-positive) oysters. Error bars are SE.

117 A. 30%

25%

20%

15% CR Abundance 10% TS

0% 1 11 21 31 41 51 61 71 Rank

B. 30%

25%

20%

15% Healthy Abundance 10% Diseased

0% 1 11 21 31 41 51 61 71 Rank

Figure 3.7: Rank-abundance curves of oyster EF bacterial taxa (RDP-designated class level). A) Rank-abundance of bacterial taxa in Choptank River and Tangier Sound oyster EF samples. B) Rank-abundance of bacterial taxa in healthy (Perkinsus marinus-negative) and diseased (P. marinus-positive) oyster EF samples.

118 40

20 Geography Site 15 Health 10 Health (Site & Geography Independent)

5 # Significantly Associated Taxa AssociatedTaxa #Significantly

0 Class Order Family Genus OTUs Taxonomic Level

Figure 3.8: The number of oyster EF bacterial taxa significantly associated with site, geographic location, or Perkinsus marinus infection status (health). Significantly (p < 0.05) different taxa were identified by Kruskal-Wallis. Taxa that differed between P. marinus-negative and P. marinus-positive oysters that also differed by site and/or geographic location were excluded from the site and geography-independent health group.

119 OWL Healthy

100

OWL Diseased

100 MA Diseased

100 TW Healthy

90 RO Diseased

RO Healthy 0.02 Figure 3.9: Hierarchical clustering of oyster EF bacterial communities by a combination of sample site and P. marinus infection status. Individual samples with the same site-health combination were combined to create a single categorical group. Groups were sampled at a depth of 1,150 sequences with 100 jackknife replicates. Node values are jackknife support values. Scale bar represents weighted UniFrac distance between groups.

120 TABLES

Table 3.1: Oysters sampled from Choptank River and Tangier Sound locations.

P. marinus-negative P. marinus-positive Choptank River Royston 2 3 Tilghman Wharf 5 0

Tangier Sound Marumsco 0 4 Old Woman’s Leg 1* 2

Total 8 9

*A second OWL healthy sample was deemed an outlier and not used in subsequent analyses.

Table 3.2: Statistical comparisons of bacterial communities by sample group.

Unweighted Weighted adonis ANOSIM adonis ANOSIM Geography p = 0.001 (R2 = 0.11) p = 0.001 (R = 0.52) p = 0.005 (R2 = 0.13) p = 0.007 (R = 0.26) (Tangier Sound vs. Choptank River)

Site p = 0.001 (R2 = 0.24) p = 0.003 (R = 0.38) p = 0.002 (R2 = 0.29) p = 0.004 (R = 0.28)

P. marinus infection status p = 0.06 (R2 = 0.08) p = 0.07 (R = 0.13) p = 0.02 (R2 = 0.12) p = 0.006 (R = 0.21) (Healthy vs. Diseased)

121 Table 3.3: Bacterial taxa significantly associated with Choptank River or Tangier Sound oysters.

Choptank Tangier River Sound Acidobacteria Sva0725 X Sva0725 X Unassigned X Unassigned X

Actinobacteria Acidimicrobiia X Acidimicrobiales X C111 X Unassigned X Unassigned X Unassigned X Unassigned X Unassigned X Unassigned X Unassigned X Unassigned X Actinobacteria Actinomycetales Microbacteriaceae X Unassigned X Unassigned X Unassigned X Unassigned X Unassigned X Unassigned X

Bacteroidetes Flavobacteriia Flavobacteriales Flavobacteriaceae Unassigned Unassigned X

Cyanobacteria Oscillatoriophycideae X

122 Chroococcales X Gomphosphaeriaceae X Unassigned X Unassigned X Synechococcophycideae Synechococcales Synechococcaceae Synechococcus Unassigned X Unassigned X Unassigned X

OP8 OP8_2 Unassigned Unassigned Unassigned Unassigned X

Planctomycetes Phycisphaerae X Phycisphaerales X Unassigned X Unassigned X Planctomycetia Pirellulales Pirellulaceae Unassigned Unassigned X

Proteobacteria Alphaproteobacteria Rhizobiales Methylobacteriaceae X Methylobacterium X Unassigned X Unassigned X Unassigned X Unassigned X Rhodobacterales Rhodobacteraceae

123 Unassigned Unassigned X Rickettsiales Pelagibacteraceae Unassigned Unassigned X Unassigned X Other X Other X Other X Unassigned X Betaproteobacteria X Unassigned X Unassigned X Unassigned X Unassigned X Deltaproteobacteria Myxococcales X Unassigned X Unassigned X Epsilonproteobacteria Campylobacterales Campylobacteraceae Arcobacter Unassigned X Gammaproteobacteria Alteromonadales X Alteromonadaceae X Umboniibacter X Unassigned X OM60 X Unassigned X Unassigned X Unassigned X Unassigned X Other X Other X Unassigned X Legionellales Coxiellaceae X Unassigned X

124 [Marinicellales] [Marinicellaceae] Unassigned Unassigned X Unassigned X Pseudomonadales Moraxellaceae X Thiohalorhabdales X Unassigned X Unassigned X Unassigned X Vibrionales X Vibrionaceae X Vibrio X Unassigned X

SAR406 AB16 X Spirochaetes [Brachyspirae] [Brachyspirales] Brachyspiraceae Unassigned Unassigned X Phylum Class Order Family Genus OTU

125 Table 3.4: Bacterial taxa significantly associated with healthy or diseased oysters.

Healthy Diseased (P. (P. marinus- Geography & Site marinus- positive) Independent negative) Actinobacteria Acidimicrobiia Acidimicrobiales C111 Unassigned Unassigned X Actinobacteria Actinomycetales Microbacteriaceae Unassigned X Unassigned X X Unassigned X

Cyanobacteria Oscillatoriophycideae X Chroococcales X Gomphosphaeriaceae X Unassigned X Unassigned X Synechococcophycideae Synechococcales Synechococcaceae Synechococcus Unassigned X

Planctomycetes Planctomycetia Pirellulales X X Pirellulaceae X X Unassigned X X Unassigned X

Proteobacteria Alphaproteobacteria Rhizobiales

126 Methylobacteriaceae X Methylobacterium X Unassigned X Rhodobacterales Rhodobacteraceae Loktanella X X Unassigned X X Betaproteobacteria X Unassigned X Unassigned X Unassigned X Unassigned X Deltaproteobacteria Myxococcales Unassigned X Unassigned X Gammaproteobacteria Alteromonadales Alteromonadaceae Umboniibacter X Unassigned X Other X Other X Unassigned X Shewanellaceae X Shewanella X Oceanospirillales Endozoicimonaceae X X Other X X Unassigned X X Halomonadaceae Candidatus Portiera X X Pseudomonadales X X Moraxellaceae X Alkanindiges X Unassigned X Other X Unassigned X Vibrionales X Vibrionaceae X

127 Vibrio X Unassigned X

Spirochaetes [Brachyspirae] X X [Brachyspirales] X X Brachyspiraceae X X Unassigned X X Spirochaetes Spirochaetales X X Spirochaetaceae X X Spirochaeta X X

Tenericutes Mollicutes Entomoplasmatales X X Entomoplasmataceae X X Other X X

TM7 TM7-1 Unassigned Unassigned Unassigned Unassigned X X Phylum Class Order Family Genus OTU

128 Chapter 4

RIBONUCLEOTIDE REDUCTASES REVEAL NOVEL VIRAL DIVERSITY AND PREDICT BIOLOGICAL AND ECOLOGICAL FEATURES OF UNKNOWN MARINE VIRUSES1

4.1 Abstract

Virioplankton play a crucial role in aquatic ecosystems as top-down regulators of bacterial populations and agents of horizontal gene transfer and nutrient cycling. However, the biology and ecology of virioplankton populations in the environment remain poorly understood. Ribonucleotide reductases (RNRs) are ancient enzymes that reduce ribonucleotides to deoxyribonucleotides and thus prime DNA synthesis.

Comprised of three classes according to O2 reactivity, RNRs can be predictive of the physiological conditions surrounding DNA synthesis. RNRs are universal among cellular life, common within viral genomes and virioplankton shotgun metagenomes (viromes), and estimated to occur within >90% of the dsDNA virioplankton sampled in this study. RNRs occur across diverse viral groups, including all three morphological families of tailed phages, making these genes attractive for studies of viral diversity. Differing patterns in virioplankton diversity were clear from RNRs sampled across a broad oceanic transect. The most abundant RNRs belonged to novel lineages of podoviruses infecting Alphaproteobacteria, a bacterial class critical to oceanic carbon cycling. RNR class was predictive of phage morphology among cyanophages and RNR distribution frequencies among cyanophages were largely consistent with the predictions of the 'Kill the Winner-Cost of Resistance' model.

129 RNRs were also identified for the first time within ssDNA viromes. These data indicate that RNR polymorphism provides a novel means of connecting the biological and ecological features of virioplankton populations.

4.2 Introduction

Viruses are key players in biogeochemical cycling and energy flow, and help shape the composition of aquatic microbial communities (Suttle 2005; Sandaa, Gomez-Consarnau et al. 2009; Winget, Helton et al. 2011). Additionally, viruses influence microbial metabolism through horizontal gene transfer and expression of auxiliary metabolic genes during infection (Breitbart, Thompson et al. 2007). Despite their impact, we understand little about the specific biological features and ecological strategies of viral populations within natural ecosystems. Constraining these second order issues is critical to building better quantitative models of how viral processes affect ecosystems (Brussaard, Wilhelm et al. 2008). Methodological limitations have hindered efforts to understand viral ecology. Viruses lack a universally conserved phylogenetic marker, akin to the 16S rRNA gene in cells, which can broadly assay viral distributions and diversity. Marker genes used as proxies of environmental viral diversity are typically limited to specific viral taxa. Furthermore, PCR-based approaches can fail to detect prominent and biologically important viral populations due to the potential for low nucleotide similarity between homologous genes. Recent work examining the diversity of viral DNA polymerase A genes within virioplankton metagenomic (virome) sequence data revealed that low efficiency DNA polymerases, undetected by PCR, were predominant within virioplankton (Schmidt, Sakowski et al. 2014). That work also highlighted the unique ability of DNA polA sequences to provide insights into the biological features of

130 unknown phages within the virioplankton. In general, the ability to connect biological features with sequence diversity in marker genes – including those widely used in ecological studies, such as the 16S rRNA gene – can be tenuous (Jaspers and Overmann 2004). Ideally, a marker gene of viral diversity should: 1) be widely distributed among diverse viral lineages and therefore, evolutionarily ancient; 2) be abundant within environmental viral assemblages; 3) play an important role in viral biology; 4) have a single evolutionary origin and not be replaceable through non-orthologous gene displacement; 5) be phylogenetically informative; and 6) be well represented in reference databases. Ribonucleotide reductase (RNR) gene products fulfill these criteria. Nucleotide metabolism pathways, including biosynthesis, are among the most represented within the virioplankton (Enav, Mandel-Gutfreund et al. 2014; Hurwitz, Westveld et al. 2014). RNRs are the only known enzymes capable of reducing ribonucleotides to deoxyribonucleotides (Lundin, Gribaldo et al. 2010), an essential step for DNA synthesis. As such, RNRs are key to nucleotide biosynthesis, under stringent evolutionary selection pressure and among the most abundant annotated genes in marine virome libraries (Dwivedi, Xue et al. 2013). Importantly, RNR genes are present in all three families of tailed phages in the order Caudovirales and have been identified in viruses infecting hosts within all three domains of life (Lundin,

Gribaldo et al. 2010). RNRs are strongly tied to lytic marine phages (Sullivan, Coleman et al. 2005), which significantly influence nutrient cycles within the global ocean (Brussaard, Wilhelm et al. 2008). Therefore, RNRs easily fit the criteria of being functionally non-redundant, abundant, and widely distributed.

131 In addition, RNRs are biologically informative and form three physiological classes according to reactivity with O2. Class I RNRs are O2-dependent. Class II RNRs are O2-independent and rely upon adenosylcobalamin (vitamin B12). Class III RNRs are sensitive to O2. All three classes share a common catalytic center and utilize similar radical-based chemistry (Nordlund and Reichard 2006). Therefore, all three modern classes of RNR likely evolved from a single common ancestor (Logan, Andersson et al. 1999). This study focused on the catalytic (alpha) subunit of the holoenzyme identified in virome libraries spanning a broad oceanic transect. Subsequently these data were used to examine the biological and ecological features of lytic phage populations within the Caudovirales. The outcomes of these analyses were interpreted within the context of known viral diversity and the ‘Kill-the-Winner, Cost of Resistance’ model for viral-host interactions (Vage, Storesund et al. 2013). Overall, these data show that RNR sequence diversity within the virioplankton connects broadly with phage morphological groups and can be predictive of the ecological strategies within the virioplankton.

4.3 Materials and Methods

4.3.1 Metagenomic libraries

Publically available virioplankton metagenomic libraries (viromes) from the Gulf of Maine, Chesapeake Bay, and Dry Tortugas (Table 4.4) were downloaded from the VIROME database (http://virome.dbi.udel.edu) (Wommack, Bhavsar et al. 2012). Filtration procedures used for the concentration of virus particles from these water samples are described in (Wommack, Sime-Ngando et al. 2010). Subsequently, viral concentrates were processed for shotgun sequencing using the Linker-Amplified

132 Shotgun Library (LASL) method (Breitbart, Salamon et al. 2002). Total viral nucleic acids were separated by hydroxyapatite chromatography into dsDNA, ssDNA, and RNA fractions for Chesapeake Bay and Dry Tortugas samples. The ssDNA and RNA fractions were transformed into dsDNA, ultimately yielding virome libraries CBS and CBR from the Chesapeake Bay; and DTS and DTR from the Dry Tortugas (Andrews- Pfannkoch, Fadrosh et al. 2010). Chesapeake Bay dsDNA virioplankton libraries CFA through CFD were collected over 24 hours at station CB 858 (38°58’N, 76°23’W).

Chesapeake Bay library CIA was an induced dsDNA library where a 1L water sample was incubated for 24h with 0.05µg mL-1 Mitomycin C to induce the production of temperate phages. Only dsDNA viruses were processed from the Gulf of Maine water samples. Random clones were selected from each library and Sanger sequenced; however, the dsDNA Chesapeake Bay time series libraries were also sequenced without cloning using 454 pyrosequencing technology (libraries CFE through CFG). Chesapeake Bay library CBB was sampled in September 2002 and amplified using the Linker-Amplified Shotgun Library (LASL) method prior to transformation and picking random colonies for Sanger sequencing (Bench, Hanson et al. 2007). Chesapeake Bay library CBJ was sampled in October 2004 as part of the first Global Ocean Survey (Rusch, Halpern et al. 2007). DNA was inserted in a medium copy plasmid and randomly selected clones were Sanger sequenced (Rusch, Halpern et al.

2007). Dry Tortugas libraries were sampled from surface seawater near the Dry Tortugas, FL USA (24°29’N, 83°4’W) in January 2004. Chesapeake Bay libraries CFA through CFD were collected over 24 hours at station CB 858 (38°58’N, 76°23’W) in July 2007. Water samples were collected on July 30, 2007 at 0600

(CFA), 1130 (CFB), 1630 (CFC), and July 31, 2007 at 0600 (CFD). Total viral nucleic

133 acids from the time series samples were separated into dsDNA, ssDNA, and RNA fractions using hydroxyapatite chromatography. The ssDNA and RNA fractions from each time point were pooled and transformed into dsDNA to provide libraries CBS and CBR, respectively. Dry Tortugas and Chesapeake Bay libraries were amplified by the LASL method prior to transformation and sequencing (Andrews-Pfannkoch, Fadrosh et al. 2010). After the induction treatment virus particles were concentrated by tangential-flow filtration (Wommack, Sime-Ngando et al. 2010). The Gulf of Maine Library GMF was sampled at station GOM04 (44°07’5”N, 67°58’3”W) in

January 2006.

4.3.2 Identification and distribution of putative virioplankton RNRs

Open reading frames (ORFs) were predicted for sequence reads in each virome using MetageneAnnotator (Noguchi, Taniguchi et al. 2008) and translated. Translated ORFs were queried against a reference protein database of RNR representative sequences from UniRef90 RNR clusters (Suzek, Huang et al. 2007) using BLASTp (Altschul, Gish et al. 1990) with an E value cutoff of 1e-10. All sequences were screened by a Conserved Domain BLAST search (Marchler-Bauer, Lu et al. 2011) to confirm homology to RNR. Each read was queried against the CDD database (v3.10) using Conserved Domain BLAST (Marchler-Bauer, Lu et al. 2011) and BLASTx

(Altschul, Gish et al. 1990) and sorted by the top results. Nucleotide reads containing ORFs with sequence homology to RNR α- subunits were sorted into five groups by Conserved Domain BLAST and top BLASTx hits: Class I ‘Cyano’, Class I ‘Other’, Class II ‘Cyano’, Class II ‘Other’, and Class II ‘RTPR’ (Tables 4.1 & 4.3; Fig. 4.1). Sequences in the Class I ‘Cyano’ group were most similar to known T4-like cyanomyoviruses. Class I RNR sequences not sharing

134 homology with cyanophages were placed in the Class I ‘Other’ group. This group included sequences that were most similar to Pelagibacter phage HTVC008M even though this phage contains a cyanophage-like RNR (Fig. 4.3). Sequences sharing closest homology with Class II cyanosiphovirus and cyanopodovirus RNRs were placed in the Class II ‘Cyano’ group. Sequences that were identified in the Conserved Domain BLAST as belonging to TIGR02505 – identified in the BLAST search as RTPR (Ribonucleotide TriPhosphate Reductase) – were placed in the group Class II ‘RTPR’. Class II RNR sequences that were not placed in either the Class II ‘Cyano’ or Class II ‘RTPR’ groups were placed in the Class II ‘Other’ group. Group frequencies were determined by library (Table 4.3). Each putative virome ORF was normalized by the average read length of the library (Table 4.4) and the gene length of the closest reference sequence identified by BLASTx (Altschul, Gish et al. 1990). The short read lengths of the 454 libraries complicated this analysis; thus, RNR distribution frequencies were not calculated for these libraries. Chesapeake Bay dsDNA library distribution frequencies were combined to compare group distributions between environments (Table 4.1).

4.3.3 Prevalence of RNR genes in DNA virome libraries

DNA sequence reads from the Gulf of Maine, Chesapeake Bay, and Dry

Tortugas libraries were recruited to 3,942 RefSeq DNA viral genomes (tBLASTx (Altschul, Gish et al. 1990); E value < 1e-20).

4.3.4 Inter-library RNR frequency normalization

Differences in read length between libraries were corrected as follows to allow for inter-library RNR frequency comparisons:

135 % RNR of library corrected =

where is the mean read length of the individual library; is the mean read length of all libraries being compared;

is the number of reads with homology to RNR within the individual library; is the number of reads in the individual library.

4.3.5 Intra-library RNR frequency normalization

The number of genomes sampled within the viral metagenome libraries was predicted by: Σ(#Bases recruited to reference/Reference genome length). Identified RNR α-subunit sequences (see above) sharing homology with viruses (E value < 1e- 20) were retained to predict the number of RNR genes sampled by: Σ(#Bases recruited to reference RNR/Reference RNR gene length). The prevalence of RNRs within dsDNA viral genomes in the virome libraries was estimated as: (#RNR Genes/#Predicted dsDNA viral genomes sampled). Sampled genomes were assumed to be of similar size as the references to which they recruited. Class I/ Class II RNR combinations in phage are rare (Dwivedi, Xue et al. 2013); thus, RNR was assumed to be single copy per genome.

4.3.6 Assembly of putative virioplankton RNR sequences

Reads containing putative RNR ORFs were assembled by library into contigs using Geneious v.5.6.2 (Drummond, Ashton et al. 2011) with the following parameters: maximum 2% gaps per read, minimum 50bp overlap, and maximum 3%

136 mismatches per read. Ambiguous bases within contigs were corrected manually. Putative RNR-containing reads from the paired Chesapeake Bay Sanger and 454 libraries CFA-CFD and CFE-CFH, respectively, were co-assembled. Assembled contigs and unassembled metagenomic sequence reads of at least 600bp were retained for another round of ORF prediction using MetaGeneAnnotator (Noguchi, Taniguchi et al. 2008) and translated.

4.3.7 Alignments and phylogenetic trees

Class I and II α-subunits of reference and putative virioplankton RNRs were aligned with MAFFT using the FFT-NS-i x1000 algorithm (Katoh, Misawa et al. 2002). The majority of predicted peptides from assembled contigs and unassembled sequence reads did not span the full length of reference peptides. Therefore, a 189 amino acid region of interest (N437 to S625 in the E. coli nrdA gene) containing key catalytic and binding residues was identified and used in subsequent phylogenetic analyses. Sequences that did not span the entire extracted region or lacked the key catalytic residues C439, E441, and C462 were removed. Gaps introduced by intein sequences were manually deleted. Retained putative RNR sequences identified in this study were deposited in GenBank (Sub. #1693175). A maximum likelihood tree with 100 bootstrap replicates was made with PhyML (Guindon and Gascuel 2003) in

Geneious v.5.6.2 (Drummond, Ashton et al. 2011) (Fig. 4.1). Subsequent alignments and trees were made by extracting sequences from the original alignment (Figs. 4.2, 4.3 & 4.5). Because of the high number of reference sequences within the Class I ‘Other’, Class II ‘Other’, and Class II ‘RTPR’ groups (Fig. 4.1) it was necessary to cluster these sequences at 80% identity using the furthest neighbor algorithm in mothur (Schloss, Westcott et al. 2009). Representative sequences from each cluster

137 were aligned with the metagenomic sequences. Metagenomic sequences belonging to the Class II ‘RTPR’ group were also clustered at 80% identity to reduce the number of sequences on the tree (Fig. 4.2).

4.3.8 Predicted RNR group abundances in the Chesapeake Bay

The abundance (viruses mL-1) of identified RNR groups was predicted for libraries CFA-D using direct count values obtained from epifluorescence microscopy (Chen, Lu et al. 2001) and recruitment to RefSeq viral genomes. The abundance of each group was calculated as follows: (VA) x (BR/TB) x (RNR/G) where VA is the observed viral abundance (mL-1) for a given library; BR is the # of bases in the library that recruited to reference viral genomes; TB is the # of total bases in the library; RNR is the # of predicted RNR genes sampled in a given group in the library; G is the # of total predicted genomes sampled in the library.

4.3.9 Predicted cyanophage population biology

Metagenomic nucleotide reads were binned by closest homology to reference sequences according to BLASTx (Altschul, Gish et al. 1990) (Fig. 4.4B). Sequences were normalized by library average read length and reference gene length. Morphology of putative cyanophage RNR sequences was predicted by the taxonomy of the closest reference sequence (Fig. 4.4C). Virioplankton cyanophage populations were examined by clustering of cyanophage-like RNR peptides. Translated ORFs were aligned using MAFFT as described above (Katoh, Misawa et al. 2002). For rarefaction analysis, peptides containing the extracted region of interest (N437 to S625 in the E. coli nrdA gene)

138 were extracted. All peptides beginning at N437 that contained key catalytic residues C439 and E441 and were ≥ 65 amino acids were clustered at 90% using the furthest neighbor algorithm in mothur (Schloss, Westcott et al. 2009). Rarefaction was performed by RNR group in EstimateS v. 9.10 (Colwell 2013) with 100 iterations and extrapolated to 120 sequences. To examine population dynamics, a 75 amino acid region (Y655 to N729 in the Synechococcus phage S-SM2 nrdA gene) in the C- terminal region of RNR was extracted. This region was chosen to provide the greatest number of environmental sequences for subsequent clustering analyses. ORF frequencies from the extracted region matched those predicted for each environment from the entire data set (Fig. 4.7B); thus, the extracted region did not bias RNR class distributions. Peptide sequences spanning the extracted region were clustered at 98% identity using the furthest neighbor algorithm in mothur (Schloss, Westcott et al. 2009). Peptide clusters were plotted on a rank-abundance curve (Fig. 4.7A). Reference phages with the greatest homology to the translated ORFs in each cluster were identified by BLASTp (Altschul, Gish et al. 1990). Cyanophage dynamics were investigated by determining the proportion that each cluster comprised of the cyanophage assemblage in libraries CFA-CFD (Fig. 4.7C).

4.3.10 Identification of Redoxins

Thioredoxins and glutaredoxins from RNR-encoding phages within the order Caudovirales were obtained from the GenBank nr database. These sequences were compiled to create a reference database. Phage genomes were queried against this reference database using BLASTx with an E value cutoff of 1e-01 to identify hypothetical or unannotated proteins that may be putative redoxins.

139 4.3.11 Contig assembly and annotation from the Rhode River

50L of surface water from the Rhode River was sampled at the Smithsonian Environmental Research Center in Edgewater, MD. The < 0.2µm fraction was concentrated by the Fe (III) chloride method (John, Mendez et al. 2011). 2 x150bp paired-end reads were sequenced with the Illumina HiSeq at the University of Delaware Sequencing and Genotyping Center at the Delaware Biotechnology Institute. Contigs were assembled from ~50 million paired-end reads using MetaVelvet (kmer = 67). Contigs over 5kb were queried against the UniProt90 RNR database (BLASTx, E value ≤ 1e-05) and retained for ORF identification using MetaGene Annotator

(Noguchi, Taniguchi et al. 2008). ORFs were annotated by homology to reference sequences identified by BLAST (Altschul, Gish et al. 1990).

4.4 Results

4.4.1 Diversity of putative virioplankton RNR sequences

The Class II ‘RTPR’ group was the single most abundant and diverse group of RNRs in dsDNA viromes (Figs. 4.1 & 4.4A; Table 4.1). RNRs in this group formed four distinct clades (Figs. 4.1 & 4.2). The majority of virioplankton sequences in this group fell in ‘RTPR’ Clades I and II and were more similar to one another than to reference sequences. Most of these RNRs claded with a single reference phage:

Puniceispirillum phage HMO-2011, a podovirus infecting a member of the SAR116 Alphaproteobacteria (Figs. 4.1 & 4.2). Cyanophage-like RNRs were ubiquitous among viromes and divided according to morphology. T4-like cyanomyoviruses carry Class I RNRs and cyanosiphoviruses and cyanopodoviruses carry Class II RNRs (Figs. 4.1 & 4.3). Class I RNRs homologous with cyanomyoviruses were less abundant than Class II cyanophage-like

140 RNRs (Fig. 4.4C; Table 4.1). Virioplankton RNRs from the Gulf of Maine, Chesapeake Bay, and Dry Tortugas sharing homology with Class II cyano-siphoviral and –podoviral RNRs formed several environment-specific clades (Fig. 4.3), and RNRs from each environment shared the closest homology with different cyanophage references (Fig. 4.4B). Class I ‘Other’ RNRs were distributed across three main clades (Figs. 4.1 & 4.5). Sequences in clade I were most similar to RNRs from invertebrate Iridoviruses, Caulobacter phages, and a monophyletic clade of Alphaproteobacteria. Clade II was the largest and was exclusively comprised of virome sequences. RNRs in this clade were distantly homologous with the RNR from Pelagibacter phage HTVC019P, a podovirus infecting the oceanic SAR11 group of the Alphaproteobacteria. Clade III was dominated by sequences from the Gulf of Maine distantly homologous to cyanophage S-TIM5, a non T4-like cyanophage (Figs. 4.1 & 4.5). Class II ‘Other’ RNRs were almost exclusively part of a clade containing Alphaproteobacterial phages PhiJL001 and Roseobacter phage RDJL phi 1, Pseudomonas phages YuA and M6, and the Alphaproteobacterium “Candidatus Puniceispirillum marinum” (Figs. 4.1 & 4.5).

4.4.2 Virioplankton RNR population distributions and ecology

RNR α-subunit genes were predicted to be present in 93% of sampled dsDNA virioplankton (Fig. 4.6). Subsequently, the abundance (viruses mL-1) of each group in the Chesapeake Bay was estimated as follows: Class I ‘Cyano’ (2.9 ± 1.2 x 105); Class II ‘Cyano’ (1.6 ± 1.3 x 106); Class I ‘Other’ (1.4 x 106 ± 4.5 x 105); Class II ‘Other’ (3.2 ± 2.5 x 105); Class II ‘RTPR’ (3.3 ± 1.8 x 106). Among identified RNRs, Class II,

B12-dependent RNR sequences were more abundant than O2-dependent Class I RNRs

141 in all three environments (Table 4.1). Overall, the ratio of Class I: Class II RNRs decreased with latitude from North to South among dsDNA libraries. Putative Class II RNR sequences were predominant in ssDNA and RNA libraries from the Chesapeake

Bay and Dry Tortugas. A single O2-sensitive Class III RNR sequence was identified in the Gulf of Maine library. RNR groups displayed environment-specific profiles across dsDNA viromes. The Gulf of Maine had the highest proportions of Class I and Class II ‘Other’ RNRs. Cyanophage-like RNRs were most abundant in the Chesapeake Bay, and Class II ‘RTPR’ RNRs were most abundant in the Dry Tortugas. ‘RTPR’ was the single most abundant group of RNRs identified, accounting for 34% of virioplankton RNRs in the Gulf of Maine and Chesapeake Bay and > 50% in the Dry Tortugas (Table 4.1). Putative cyanophage virioplankton RNRs were used as a model for assessing phage population biology in situ. The majority of sampled cyanophage particles were predicted to have RNRs (Table 4.2). Class II cyanophage RNRs (Siphoviruses and Podoviruses) were more diverse than Class I RNRs (Myoviruses) (Fig. 4.4A). Cyanophage populations within the Chesapeake Bay were dynamic over 24 hours (libraries CFA-CFD). Individual cyanophage-like RNR peptide clusters fluctuated over this 24h period, ranging from 0 to >75% of all cyanophage-like RNRs at any given time point (Fig. 4.7C). Within the RNR-carrying cyanophage assemblage, the predicted proportion of Class I cyanomyoviral RNRs increased from 14% at T0 to 45% at T24 (Fig. 4.4C).

142 4.5 Discussion

4.5.1 The identity of phages carrying novel virioplankton RNR sequences

Few phage RNR sequences claded with the abundant virioplankton RNRs within the Class II ‘RTPR’ group and Class I ‘Other’, clade II (Figs. 4.1, 4.2 & 4.5); thus, to better characterize the origin of these RNRs we examined the genomic context of similar RNRs on contigs >5kb from a deeply sequenced virome library. Three contigs > 5kb with an ‘RTPR’ RNR (Figs. 4.8 & 4.9A) and one contig > 5kb with a Class I ‘Other’, clade II RNR (Figs. 4.8 & 4.9B) were identified. The majority of identifiable ORFs on all three ‘RTPR’ contigs were most similar to genes of known podoviruses (Fig. 4.8). In particular, the most common top BLAST hits were to podoviruses that infect Alphaproteobacteria. This provides circumstantial evidence that virioplankton sequences within the ‘RTPR’ group belonged to podoviruses. The Class I ‘Other’, clade II RNR occurred on a 12kb contig (Figs. 4.8 & 4.9B). Seven of the 10 predicted ORFs on this contig shared homology with genes of podoviruses infecting Alphaproteobacteria, such as Pelagibacter phages HTVC011P and HTVC019P, and Roseobacter phage SIO1, though these were not always the top BLAST hit (Fig. 4.8). A bacterial glutaredoxin sharing homology with a SAR11 Alphaproteobacterium was also present (Fig. 4.8), and may have been introduced by horizontal gene transfer from the host. Thus, the Class I ‘Other’, clade II RNRs represent a substantial increase in the diversity of known phages infecting marine Alphaproteobacteria like the SAR11 group. Interestingly, many of the virioplankton ‘RTPR’ sequences were similar to RNRs from green algal species Volvox carteri f. nagariensis and Chlamydomonas reinhardtii, as well as Chlorella viruses like Acanthocystis turfacea Chlorella Virus 1

143 (ATCV-1) (Figs. 4.1 & 4.2). Class II (B12-dependent) RNRs are rarely observed in eukaryotes but have been identified in several microalgal species and are the result of horizontal gene transfer from bacteria (Lundin, Gribaldo et al. 2010). Many microalgal species, including V. carteri, are B12-dependent auxotrophs and rely upon bacteria for this essential co-factor (Croft, Lawrence et al. 2005), and several algal-associated bacteria are capable of de novo B12 synthesis and supporting growth of auxotrophic algal cultures in minimal media (Wagner-Dobler, Ballhausen et al. 2010; Kazamia, Czesnick et al. 2012). The presence of two CRISPR (clustered regularly interspaced short palindromic repeats) arrays in the B12-producing, dinoflagellate-associated bacterium Dinoroseobacteri shibae DFL12T indicates that phage infections occur in these bacteria (Wagner-Dobler, Ballhausen et al. 2010). Based on this information, we hypothesize that these Class II ‘RTPR’ sequences are from podoviruses whose hosts are B12-producing heterotrophic bacteria associated with microalgal species. Furthermore, the hosts for ‘RTPR’ podoviruses may be Alphaproteobacteria as many microalgal-associated bacteria belong to this sub-phyla (Wagner-Dobler, Ballhausen et al. 2010; Helliwell, Wheeler et al. 2011; Kazamia, Czesnick et al. 2012), and many of the ORFs on the ‘RTPR’ contigs (Fig. 4.8) shared homology with Roseobacter phage SIO1 and Puniceispirillum phage HMO-2011.

4.5.2 RNR as a proxy of phage population biology

RNR can provide insight into ecological features of lytic phage populations because of its wide distribution among viral lineages. There is a growing consensus that morphological groups broadly correspond with the ecological strategy of tailed cyanophages. Myoviruses are believed to be generalists, capable of infecting a broader range of hosts; whereas, podo- and siphoviruses are specialists, infecting a narrow

144 range of host species (Suttle and Chan 1993; Sullivan, Waterbury et al. 2003; Wang and Chen 2008; Dekel-Bird, Avrani et al. 2013; Labrie, Frois‚ÄêMoniz et al. 2013). To date no genetic marker has been clearly associated with phage morphological groups. However, the split in RNR class between cyanophage morphological groups (Figs. 4.1, 4.3) and the high proportion of cyanophages with RNR genes observed in this study (Table 4.2) open a window to investigating the population dynamics of generalist and specialist phage groups. Consistent with previous reports of siphoviruses and podoviruses, many Class II RNR cyanophage clades were comprised of sequences from a single environment (Fig. 4.3), indicating endemic populations of specialist phages (Angly, Felts et al. 2006; Williamson, Rusch et al. 2008). Localized diversity of specialist cyanophages may be a bellwether of distinct cyanobacterial host populations between environments. Class II cyanophage RNRs were more abundant than Class I cyanomyoviral RNRs in every library sampled except CBJ and CIA (Fig. 4.4C; Table 4.3), and the top cyanophage peptide clusters in all three environments shared closest homology with putative specialist phages (Fig. 4.4 & 4.7A). The high proportion of Class II cyanophage-like RNRs contradicts previous surveys of cyanophage distributions within microbial metagenomic datasets that identified high proportions of cyanomyoviruses (Williamson, Rusch et al. 2008; Ghai, Martin-Cuadrado et al. 2010;

Labrie, Frois‚ÄêMoniz et al. 2013). However, these observations could have been biased by sequencing the > 0.1 µm microbial fraction, which would have selectively eliminated diminutive podoviruses (Labrie, Frois‚ÄêMoniz et al. 2013). Typically, generalist myoviruses have larger genomes and are larger in size as compared to podoviruses. In marine viromes targeting the <0.2 µm viral fraction, cyanopodoviruses

145 accounted for ≥ 50% of all cyanophage sequences (Labrie, Frois‚ÄêMoniz et al.

2013). Likewise, cyanopodoviruses were previously predicted to dominate cyanophage populations in the Chesapeake Bay (Bench, Hanson et al. 2007), a finding that was corroborated in this study (Fig. 4.4C & 4.7A). The high proportion of Class II RNRs identified in this study, which we believe to be predictive of abundant podoviral populations, agrees with predictions made by a revised model of KtW that incorporates a cost of resistance to viral infection (the KtW-COR model)(Vage, Storesund et al. 2013). A key prediction of the KTW-COR model is that specialist phages, such as podoviruses, dominate within virioplankton assemblages and limit the abundance of competition specialist hosts through phage predation (Vage, Storesund et al. 2013). Interestingly, the Class II:Class I ratio of cyanophage RNRs decreased over 24 hours in the Chesapeake Bay (CFA-CFD) (Fig. 4.4C). Diel patterns in cyanophage abundance have been observed previously (Clokie, Millard et al. 2006). However, the dynamics of generalist and specialist phage populations over these cycles is unknown. Longitudinal sampling over a longer time period may help determine whether there are predictable patterns in the abundance of generalist (Class I) and specialist (Class II) cyanophages within the virioplankton.

4.5.3 RNR class and phage biology

Frequencies of Class I and Class II virioplankton RNR groups differed across marine environments (Table 4.1) and may indicate differences in the abundance of specific bacterial host groups between sample sites. Even in the Gulf of Maine (which had the highest proportion of Class I RNRs), 62% of RNRs identified were Class II (Table 4.1). In contrast, the majority of RNRs within known phages and short-read

146 marine viromes are aerobic, Class I RNRs (Dwivedi, Xue et al. 2013). This discrepancy may be due to the predominance of myoviral genomes in sequence databases and the fact that short sequence reads can make identifying RNR class designations difficult (Wommack, Bhavsar et al. 2008; Dwivedi, Xue et al. 2013). Moreover, the short-read aquatic viromes showing a high abundance of Class I RNRs were amplified by multiple displacement amplification (MDA) prior to sequencing (Angly, Felts et al. 2006; Desnues, Rodriguez-Brito et al. 2008; Lopez-Bueno, Tamames et al. 2009; Rodriguez-Brito, Li et al. 2010). MDA is known to introduce sequencing biases (Polson, Wilhelm et al. 2011; Marine, McCarren et al. 2014) and can result in large coverage discrepancies across loci (Raghunathan, Ferguson et al. 2005). Thus, assessing gene frequency data from MDA libraries is particularly difficult. In contrast, the viromes examined in this study were amplified using relatively unbiased approaches (i.e., the LASL approach) (Henn, Sullivan et al. 2010). Though less abundant, Class I RNRs provide key insights into phage biology and diversity. Most phages have the same RNR class as their host; yet, T4-like cyanophages have Class I RNRs that are similar to those of enterobacteria and T4-like phages rather than Class II RNRs like the majority of marine cyanobacteria (Figs. 4.1 & 4.3). In fact, Class I RNRs are part of the ‘core genome’ among T4-like phages (Petrov, Ratnayaka et al. 2010) and likely play a vital role in T4-like phage biology, which may explain the constrained RNR diversity observed in these phages (Fig. 4.4A). One possible constraint may be related to the small redox protein glutaredoxin, which provides RNR with its reducing potential (Arner and Holmgren 2000). Bacteriophage T4 encodes a glutaredoxin (the protein product of gene nrdC (Nikkola, Engstrom et al. 1993)) that is compatible with only the phage RNR; likewise, the T4-

147 encoded RNR can only be reduced by its own glutaredoxin and not by E. coli’s thioredoxin (Berglund 1969). Upon infection, T4’s glutaredoxin is preferentially reduced over E. coli’s thioredoxin, effectively outcompeting the host’s RNR system (Berglund and Holmgren 1975). The phage-specific RNR-redoxin combination may preclude the acquisition of a host-like RNR in T4-like phages unless both are simultaneously lost. Supporting this hypothesis, Aeromonas phages 25, 31, phiAS4, and 44RR2.8t were the only T4-like phages without T4-like Class I RNRs (Filee, Bapteste et al. 2006)(Fig. 4.5). They were also the only T4-like phages lacking the nrdC gene (Petrov, Nolan et al. 2006). Likewise, Bacteriophage RM378, a T4-like phage with a Class II RNR, has no identifiable redoxins. In contrast, host thioredoxins act as processivity factors in the phage T7 DNA polymerase holoenzyme (Tabor, Huber et al. 1987); thus, T7-like podoviruses may be under increased pressure to carry RNRs compatible with thioredoxins of their hosts. Interestingly, not all related phages carry RNR. Among T7-like podoviruses, only marine phages are known to carry RNR genes (Sullivan, Coleman et al. 2005), and two of the three identified Pelagibacter podoviruses (phages HTVC010P and HTVC011P) have no identifiable RNR genes in their genomes (Zhao, Temperton et al. 2013). It is unclear why some related phages carry RNR genes and others do not, particularly lytic phages in marine environments where RNR may provide a selective advantage in phosphate-limited environments (Sullivan, Coleman et al. 2005; Lindell, Jaffe et al. 2007). Nevertheless, >90% of dsDNA virioplankton sampled in these environments were estimated to carry RNR (Fig. 4.6). As expected, Class I RNRs were ubiquitous among T4-like cyanophages, while >70% of cyano-siphoviruses –and podoviruses had Class II RNRs (Table 4.2). In the case of the Pelagibacter

148 podoviruses, over-recruitment to the HTVC019P RNR in comparison to its genome implies RNRs are more prevalent than cultivated references indicate (Table 4.2). In addition, distinct RNRs were observed between reference myoviral and podoviral Pelagiphages (Figs. 4.1, 4.3 & 4.5), suggesting RNR genes can provide a cultivation- independent means to query the abundance, distribution, and ecological strategies of this poorly studied group. Another unexpected finding was the occurrence of RNRs in ssDNA and RNA viromes. To our knowledge, RNR has not been observed in ssDNA or RNA viral genomes. A significant proportion of sequences in the ssDNA viromes shared homology with cyanopodoviruses (Fig. 4.10A), and most of these sequences mapped to a single genomic region of cyanopodoviruses containing genes involved in DNA replication, including RNR and DNA polymerase A (Fig. 4.10B). The number of reads in the ssDNA viromes mapping to DNA replication regions was approximately 3x that of dsDNA viromes from the same location (Fig. 4.10B). Given these findings, we believe that the most parsimonious explanation is that these DNA replication gene cassettes may have been horizontally transferred to ssDNA viruses from a cyanopodovirus or vice versa. The latter possibility is intriguing because cyanopodoviral RNRs are distinct (Fig. 4.1), as they were among the shortest α- subunit genes identified. It is conceivable that RNR could provide a selective advantage to ssDNA viruses, resulting in its maintenance within a streamlined genome. It seems unlikely RNRs would be present in RNA viruses as most known RNA viruses do not use a DNA intermediate stage, and there are no documented RNRs among known retroviral genomes. These sequences may have originated from ssDNA contaminating the RNA fraction as these nucleic acids elute successively from

149 the hydroxyapatite column. Moreover, the RNRs from the RNA libraries were most similar to cyanopodoviral RNRs (Table 4.1), and nearly 50% of viral reads in Chesapeake Bay RNA library CBR (Table 4.4) shared homology with ssDNA phages (http://virome.dbi.udel.edu).

4.6 Conclusions

Ribonucleotide reductase contains a number of attractive qualities as a marker of viral diversity in aquatic environments. Its occurrence across viral taxa, including the most abundant observed phages in marine environments, captures a cross-section of virioplankton diversity beyond that previously attainable with a single gene. Additionally, RNRs provide biologically relevant information regarding the physiological conditions of DNA replication and provide a means of investigating phage population biology in situ. RNR sequences from the Gulf of Maine, Chesapeake Bay, and Dry Tortugas revealed novel groups of abundant lytic viruses in the ocean and the frequency of RNR sequence types within the virioplankton agreed with phage- host interactions predicted by the KtW-COR model.

150 FIGURES

Class I ‘Other’

Myoviruses Cyanophage S-TIM5 Iridoviruses Clade III T5-like Clade I siphoviruses

T4-like phages Pelagibacter phage Hosts and associated viruses HTCV019P

T4-like cyanophages Clade II Actinobacteria

Class I ‘Cyano’ Alphaproteobacteria

Gammaproteobacteria Archaea

Bacteroidetes Phage

PhiJL001 Class II ‘Other’ Betaproteobacteria

Class I Chloroflexi

Class II Cyanobacteria ‘RTPR’ Clade IV Deinococcus-Thermus Synechococcus & Prochlorococcus Eukaryote

Clade III Firmicutes

Gammaproteobacteria Mycobacterium phages Virome seqs (this study)

Clade II Spirochetae Class II ‘RTPR’ Cyanosiphoviruses & Clade I Cyanopodoviruses Phage RM378 Puniceispirillum phage Class II ‘Cyano’ HMO-2011 0.3 Figure 4.1: Unrooted maximum likelihood tree with 100 bootstrap replicates of Class I and Class II alpha subunit RNR sequences from reference genomes and virioplankton metagenomic libraries. Viral reference RNR sequences are colored the same as their hosts. Major groups of RNR sequences identified in this study were Class I ‘Other’ and Class I ‘Cyano’; and Class II ‘Other’, Class II ‘RTPR’, and Class II ‘Cyano’. Aligned region corresponds to residues N437 to S645 in the E. coli Class I alpha RNR peptide. Scale bar represents amino acid substitutions per site.

151 ATCV-1 Clade I

Clade II

Phage Acanthocystis turfacea Chlorella virus 1 (7) RM378 GMF Contig 14 (3); DT (2); GM (9) DTF Contig 5 (3); CB (1); DT (25); GM (2) Clade IV Puniceispirillum phage HMO-2011 Chlamydomonas reinhardtii Clade I Volvox carteri f. nagariensis Clade III CFA Contig 1 (6); CB (21); DT (12); GM (13) Mycobacterium DTF Contig 15 (2) phage Nitrosococcus halophilus Nc4 Roseiflexus castenholzii DSM 13941 (4) Bacteriophage RM378 CBJ1098127015314_915_1_1 GMF1061042970635f_1_839_1 Diplosphaera colitermitum TAV2 (2) Salpingoeca sp. ATCC 50818 Clade II Monosiga brevicollis MX1 GMF1061042661264f_1_935_1 GMF Contig 3 (8); DT (6); GM (21) Thermus phage P23-45 Thermus phage P74-26 Alkaliphilus metalliredigens QYMF (2) Celeribacter phage P12053L Roseophage SIO1 CFA1061053822165r_5_739_1 Candidatus Protochlamydia amoebophila UWE25 GMF1061042942020f_103_880_1; CB (1); DT (1); GM (2) Clade III CFA1061053817669f_1_691_1 Mycobacterium phage Tiger (22) Mycobacterium phage Gladiator (16) Rhodococcus phage RGL3 (2) DTF1061043203475r_1_800_1; DT (1); GM (9) Clade IV 0.4 Ectocarpus siliculosus (9) Figure 4.2: Unrooted maximum likelihood tree with 100 bootstrap replicates of Class II RNR reference and putative metagenomic ‘RTPR’ sequences. Metagenomic sequences from the large tree (inset) were clustered at 80% identity. Representative metagenomic sequences were placed on the tree, with the number of reads from each environment within that cluster listed. CB (Chesapeake Bay); DT (Dry Tortugas); GM (Gulf of Maine). Bacterial references from the large tree (inset) were clustered at 80% identity. Representative sequences were placed on each tree. Numbers in parentheses following bacterial references indicate the number of reference sequences within that cluster. Scale bar represents amino acid substitutions per site. Bacteria (purple); Eukaryotes and eukaryotic viruses (orange); Myoviruses (red); Siphoviruses (blue); Podoviruses (green); Metagenomic sequences (black). Celeribacter phage P12053L was colored as a podovirus based on its T7-like DNA polymerase even though it is officially listed as an unclassified dsDNA phage. Black, grey, and white circles represent bootstrap support ≥ 100%, 75%, and 50%, respectively.

152 Class I DTF1061043225462r_796_1_1 Cyanobacteria P-RSP5 P-SSP10 P-HP1 P-SSP11 P-SSP5 CBB017C04.y01_102_812_1 Cyanosiphoviruses DTF1061043197464f_808_1_1 & DTF1061043192230f_735_1_1 T4-like CBB008H08.y01_1_669_1 cyanopodoviruses CBJ1098214050374_886_1_1 cyanomyoviruses S-CBP3 S-CBP4 GMF Contig 39 (2) GMF Contig 11 (4) GMF1061042934309f_687_1_1 GMF1061042937625r_817_1_1 GMF1061042970087r_1_912_1 KBS-S-1A GMF Contig 23 (3) S-RIP2 DTS Contig 4 (2) GMF Contig 5 (7) DTR1061059708668r_1_902_1 Cyanosiphoviruses DTS Contig 3 (3) & CBB019C09.y01_1_709_1 S-RIP1 cyanopodoviruses GMF1061042926875f_1_685_1 CFD Contig 14 (3) CFA Contig 10 (2) Cyanobacteria CFB Contig 5 (4) CBJ1098127013765_232_910_2 Class II S-CBS4 P-GSP1 P-SSP7 P-SSP3 P-SSP2 DTF Contig 8 (3) DTF1061042876837_990_1_1 DTR1061059708722r_64_828_2 Syn5 P-SSP9 DTS Contig 2 (4) DTS1061059628551r_1_780_1 CBR1061057323945r_825_1_1 CBJ1098101635164_885_1_1 CFB1061053832039f_822_1_1 S-CBS2 P60 DTF1061043186908f_854_1_1 Synechococcus sp. PCC 7335 Synechococcus elongatus PCC 7942 Thermosynechococcus elongatus BP-1 Synechococcus sp. RCC307 Synechococcus sp. WH 5701 Prochlorococcus marinus str. MIT 9303 Prochlorococcus marinus str. MIT 9313 Synechococcus sp. CC9605 Synechococcus sp. CC9902 Synechococcus sp. BL107 Synechococcus sp. WH 8102 Synechococcus sp. WH 7803 Synechococcus sp. WH 7805 Synechococcus sp. RS9916 Synechococcus sp. CC9311 Synechococcus sp. RS9917 Prochlorococcus marinus subsp. marinus str. CCMP1375 Prochlorococcus marinus str. NATL1A Prochlorococcus marinus str. MIT 9515 Prochlorococcus marinus subsp. pastoris str. CCMP1986 Prochlorococcus marinus str. MIT 9312 Prochlorococcus marinus str. MIT 9301 Prochlorococcus marinus str. MIT 9215 Synechococcus sp. JA-3-3Ab Synechococcus sp. JA-2-3B'a(2-13) Nostoc sp. PCC 7120 Anabaena variabilis ATCC 29413 P-SS2 P-SSM2 S-SSM7 GMF Contig 7 (6) CBJ Contig 2 (2) S-SM2 Syn9 Syn33 P-RSM4 S-SSM5 Syn19 T4-like S-SM1 cyanomyoviruses P-SSM7 P-SSM4 Syn1 S-PM2 P-HM2 P-HM1 S-ShM2 Pelagibacter phage HTVC008M CBJ1098101650305_23_923_1 Cyanothece sp. CCY 0110 NrdA2 Synechococcus sp. PCC7002 Synechocystis sp. PCC6803 Cyanothece sp. PCC 7424 Microcystis aeruginosa PCC 7806 Cyanothece sp. CCY 0110 NrdA1 Cyanothece sp. ATCC 51142 Cyanobacterium UCYN-A Arthrospira platensis NIES-39 Arthrospira maxima CS-328 0.3 Cyanophage Ma-LMM03 Cyanophage Ma-LMM02 Cyanophage Ma-LMM01 Cyanophage Ma-HPM05 Figure 4.3: Unrooted maximum likelihood tree with 100 bootstrap replicates of Class I alpha and Class II RNR reference and putative metagenomic ‘Cyano’ sequences. Numbers in parentheses indicate the number of reads assembled in each contig. Scale bar represents amino acid substitutions per site. Bacteria (purple); Myoviruses (red); Siphoviruses (blue); Podoviruses (green); Metagenomic sequences (black). Black, grey, and white circles represent bootstrap support ≥ 100%, 75%, and 50%, respectively.

153 A. B. 20 C. 70

CB Time Series 100 60 15

80 50 10 Cyano I 60 40 Dry Tortugas Cyano II Chesapeake Bay Gulf of Maine 5 % Cyanophage Sequences

# Clusters 40 30 Other I

Other II 20 20 0 RTPR P60 Syn5 Syn1 Proportion Cyanophage Population (%) Syn10 Syn33 P-HP1 S-SM2 S-PM2 S-RIP1 S-RIP2 0 S-RIM8 P-SSP7 10 P-SSP2 P-SSP5 S-CBP3 S-CBP4 P-RSP5 S-CBS2 S-CBS4 P-GSP1 S-SSM2 S-SSM7 P-SSM2 S-SSM4 S-RSM4 S-CAM8 P-RSM1 S-CAM1 S-CBM2 P-RSM6

P-SSP10 CFA CFB CFC CFDCBB CBJ DTF GM S-MbCm6 KBS-S-1A 0600 1130 1630 0600+1d

0 MetaG-MbCm1 Myoviridae Siphoviridae Podoviridae 0 20 40 60 80 100 120 140 Cyanophage Reference Top BLAST Hits SPMPPPPPP P MMSMSPPMMMMMMPPPMMMMMMMM dsDNA Viromes RNR Sequences Tailed Phage Morphological Family Figure 4.4: Cyanophage-like RNR diversity and distribution in the Gulf of Maine, Chesapeake Bay, and Dry Tortugas dsDNA libraries. Panel A) Rarefaction curve of RNR peptide clusters by group at 90% identity. Groups were extrapolated to a total of 120 RNR sequences (dotted lines). Error bars are SD. Panels B and C) Distribution of cyanophage-like RNR sequences identified by top BLASTx hit. Sequences were normalized for gene length and library read length prior to analyses. Panel B) Rank abundance of cyanophages across all three environments. Myovirus (M); Siphovirus (S); Podovirus (P). Panel C) Relative abundance of cyanophage-like RNR sequences across the Chesapeake time series (CFA-CFD), other Chesapeake viromes (CBB, CBJ), the Dry Tortugas (DTF) and the Gulf of Maine (GM).

154 GMF1061042925962f_784_1_1 Pelagibacter phage HTVC019P Clade I Clade II GMF1061042949966r_830_1_1 Bacteroidetes Candidatus_Pelagibacter_ubique_HTCC1002 (27) CFB Contig 1 (13) α-proteobacteria CBJ1098101648625_866_21_1 DTF1061043129216r_923_1_1 GMF Contig 24 (3) Clade I DTF1061042874584r_837_1_1 γ-proteobacteria DTF1061043202875r_1_816_1 GMF1061039282610r_1_967_1 GMF1061042969816r_1_904_1 Invertebrate iridescent virus 6 Invertebrate iridescent virus 3 Aedes taeniorhynchus iridescent virus Clade III Caulobacter phage CcrColossus PhiJL-like Caulobacter phage CcrRogue Caulobacter phage phiCbK clade Caulobacter phage CcrKarma Caulobacter phage CcrSwift Caulobacter phage CcrMagneto α-proteobacteria Pseudomonas phage KPP10 Pseudomonas phage PAK P3 Pseudomonas phage P3 CHA CBS1061057326800r_735_1_1 β-proteobacteria GMF1061042971560f_1_840_1 Firmicutes CFA1061053812415f_675_1_1 GMF1061042926426f_648_1_1 GMF1061042943969r_1_765_1 Class I Class II GMF Contig 25 (3) GMF1061043214682f_1_747_1 GMF Contig 2 (8) Roseobacter phage RDJL Phi 1 CBJ1098101801525_1_814_1 Pseudomonas phage YuA GMF1061042943894f_1_624_1 GMF1061042957534r_1_794_1 Pseudomonas phage M6 DTF Contig 3 (4) Clade II GMF1061042906920f_834_1_1 GMF Contig 9 (6) GMF Contig 49 (2) GMF1061042906719f_1_672_1 GMF1061042925690f_1_710_1 Phage phiJL001 CFC Contig 1 (6) Candidatus Puniceispirillum marinum IMCC1322 CFA Contig 12 (2) CFD Contig 3 (5) GMF1061042661109r_1_940_1 PhiJL-like DTF1061043183606r_811_1_1 GMF1061042934442r_760_1_1 GMF1061042924671r_733_1_1 GMF Contig 22 (3) GMF1061042970573r_630_1_1 GMF1061042968032r_1_866_1 GMF1061043206965f_804_1_1 CFA1061053829314f_1_770_1 Francisella novicida FTG (6) GMF Contig 10 (5) Deftia phage phiW-14 GMF1061042967907r_717_1_1 Grouper iridovirus CFD Contig 11 (3) Singapore grouper iridovirus Rana tigrina ranavirus DTF Contig 26 (2) Regina ranavirus Clostridium difficile QCD-32g58 (4) Ambystoma tigrinum stebbensi virus Frog virus 3 DTF1061043181921_1782_938_2 Soft-shelled turtle iridovirus Salinibacter ruber DSM 13855 Lymphocystis disease virus-isolate China Zunongwangia profunda SM-A87 (10) Lymphocystis disease virus 1 GMF Contig 4 (7) GMF Contig 40 (2) Halophage AAJ-2005 GMF1061042963326f_1_823_1 Thermus aquaticus Y51MC23 (7) Cyanophage S-TIM5 Rhodospirillum rubrum ATCC 11170 GMF1061042927759r_1_807_1 Burkholderia multivorans ATCC 17616 (3) DTF1061042915561r_791_1_1 GMF1061043206175r_1_828_1 Oxalobacter formigenes HOxBLS GMF1061043245027r_933_1_1 Clade III Burkholderia pseudomallei DM98 (7) GMF1061042968114f_929_1_1 Ralstonia eutropha H16 (7) Class I GMF Contig 30 (3) GMF1061042926741f_1_727_1 Methylococcus capsulatus str. Bath GMF Contig 42 (2) Bacillus coagulans 36D1 (8) GMF Contig 56 (2) Blattabacterium sp. (Blattella germanica) str. Bge (7) Paenibacillus sp. JDR-2 (2) Vibrio phage pVp-1 Bacillus tusciae DSM 2912 (2) CFC Contig 5 (2) Thermobaculum terrenum ATCC BAA-798 Aeromonas phage phiAS4 Aeromonas salmonicida bacteriophage 25 Phenylobacterium zucineum HLK1 (3) Aeromonas phage 44RR2.8t Roseobacter sp. CCS2 (14) Aeromonas phage 31 Class II Magnetospirillum magneticum AMB-1 (3) Roseophage DSS3P2 0.3 Roseophage EE36P1 Roseovarius sp. 217 phage 1 Aeromonas salmonicida subsp. salmonicida A449 (15) Thioalkalivibrio sp. K90mix Thermus phage phiYS40 Thermus phage TMA 0.4 Figure 4.5: Unrooted maximum likelihood tree with 100 bootstrap replicates of Class I alpha (left) and Class II (right) RNR reference and putative metagenomic ‘Other’ sequences. Numbers in parentheses following metagenomic contigs indicate the number of reads assembled in each contig. Bacterial references from the large tree (inset) were clustered at 80% identity. Representative sequences were placed on each tree. Numbers in parentheses following bacterial references indicate the number of reference sequences within that cluster. Scale bar represents amino acid substitutions per site. Bacteria (purple); Eukaryotic viruses (orange); Myoviruses (red); Siphoviruses (blue); Podoviruses (green); Metagenomic sequences (black). Black, grey, and white circles represent bootstrap support ≥ 100%, 75%, and 50%, respectively.

155 Recruitment Predicted Predicted to RefSeq Virioplankton dsDNA Viral Nucleic Acid Virioplankton Genomes Distribution RNR Frequency

7% RNA 82% 6% Virome Bases Not Recruited 18% 24% Virome Bases ssDNA 70% to Recruited dsDNA RefSeq Viral 93% Predicted Proportion References of dsDNA Virioplankton with RNR

Figure 4.6: Predicted RNR frequency among dsDNA virioplankton populations from the Gulf of Maine, Chesapeake Bay, and Dry Tortugas. DNA virome reads were queried against RefSeq viral genomes (tBLASTx; E value <1e-20). Total genomes sampled was predicted by: Σ(#Bases recruited to reference/Reference genome length). Total α-subunit RNRs sampled was predicted from recruitment to reference viral RNR α-subunit genes (E value < 1e-20) by: Σ(#Bases recruited to reference RNR/Reference RNR gene length). The frequency of RNR in dsDNA virioplankton was predicted as: (#Predicted RNR genes/#Predicted dsDNA genomes sampled). RNR was assumed to be single copy per genome. No RNR sequences recruited to ssDNA or RNA reference genomes.

156 A. 30 Chesapeake Bay B.

18 Putative Cyanophage RNR Sequences 12 Class I Class II Predicted Observed Predicted Observed 6 % of Population GM 28% 24% 72% 76%

0 CB 33% 32% 67% 68% 123456789101112131415 DT 10% 10% 90% 90%

Gulf of Maine 15

12 Myoviridae

9 Siphoviridae 6 100%100%

C. n 3 % of Population Podoviridae 75%75% 0 123456789101112131415161718192021 50%50%

25%25% yanophage Populatio yanophage C

30 Dry Tortugas % Cyanophage Population 0%0%

ClusterCluste 12 18 ClusterCluster 110 ClusterCluster 8 12 ClusterCluster 5

6 CFACFA ClusterCluster 4 % of Population ClusterCluster 3 CCFBFB 0 ClusterCluster 2 1234567 CFCCFC ClusterCluster 1 Protein Clusters Ranked by Abundance CFD Figure 4.7: Distribution and dynamics of cyanophage populations. ORFs with a top BLASTx hit to a cyanophage were translated, aligned and clustered at 98% identity. Panel A) Rank-abundance of cyanophage-like RNR clusters in the Chesapeake Bay, Gulf of Maine, and Dry Tortugas. The morphology of the reference phage with the closest homology to sequences in each cluster is identified. Myoviruses (red); Siphoviruses (blue); Podoviruses (green). Panel B) Comparison of RNR distribution frequency in extracted region for peptide cluster analysis and those predicted by normalization of RNR sequences. Gulf of Maine (GM); Chesapeake Bay (CB); Dry Tortugas (DT). Panel C) Dynamics of phage populations by cluster in the Chesapeake Bay time series. Myoviruses (red); Siphoviruses (blue); Podoviruses (green).

157 Class I ‘Other’ Contig 12643

ORF10 ORF9 ORF8 ORF7 ORF6 ORF5 ORF4 ORF3 ORF2 ORF1

DNA Primase DNA Polymerase A Exonuclease Ribonucleotide Reductase Class I alpha

500bp

Protein Protein Hypothetical Hypothetical Protein Glutaredoxin Hypothetical Endonuclease I

Class I beta

ORF # Annotation Homoloogous Sequeences Top BLAST hit (E value) Ribonucleotide Reductase HTVC011P HTVC019P SIO1 1 RNR beta Novosphingobium sp. PP1Y (3e-14) 2 RNR alpha Pelagibacter phage HTVC019P (0) 3 Hypothetical protein Clostridium leptum (4e-16) 4 Hypothetical protein NA 5 Endonuclease I Celeribacter phage P12053L (1e-27) 6 Glutaredoxin alpha proteobacterium HIMB59 (1e-20) 7 Hypothetical protein NA 8 Exonuclease Celeribacter phage P12053L (1e-77) 9 DNA Polymerase A Roseobacter phage SIO1 (2e-125) 10 DNA Primase Azorhizobium caulinodans ORS 571 (3e-151)

Contig 5585 Class II ‘RTPR’

ORF1 ORF2 ORF3 ORF4 ORF5 ORF6 ORF7 ORF8 ORF9 ORF10

Hypothetical Primase/Helicase DNA Polymerase A Methyltransferase Exonuclease Ribonucleotide Reductase Protein

500bp 3310 !

Protein Du Protein

Hypothetical Hypothetical Endonuclease I Top BLAST Hit ORF # Top BLAST hit Viral Sequence with Greatest Homolgy 1 gamma proteobactgerium SCGC AAA160-D02 (1e-165) Vibrio phage CHOED (7e-142) 2 gamma proteobactgerium SCGC AAA160-D02 (2e-157) Podovirus GOM (7e-133) Myoviridae 3 Puniceispirillum phage HMO-2011 (8e-80) Puniceispirillum phage HMO-2011 (8e-80) 4 gamma proteobactgerium SCGC AAA160-D02 (2e-06) Vibrio phage CHOED (1e-03) 5 Rhizobium leguminosarum (43-31) Tetraselmis viridis virus S20 (1e-18) 6 Puniceispirillum phage HMO-2011 (6e-23) Puniceispirillum phage HMO-2011 (6e-23) Podoviridae 7 Celeribacter phage P12053L (1e-73) Celeribacter phage P12053L (1e-73) 8 Puniceispirillum phage HMO-2011 (9e-32) Puniceispirillum phage HMO-2011 (9e-32) 9 Rickettsia felis URRWXCal2 (2e-04) N/A 10 Puniceispirillum phage HMO-2011 (0) Puniceispirillum phage HMO-2011 (0) Eukaryotic virus

Contig 8066 Non-viral/ No hit

ORF1 ORF2 ORF3 ORF4 ORF5 ORF6 ORF7 ORF8 ORF9 ORF10 ORF11 ORF12 ORF13

Primase/ DNA Polymerase A Thymidylate DNA Polymerase A Exonuclease Ribonucleotide Reductase Helicase Exonuclease Domain Synthase

3310 500bp ! Protein Protein Protein Protein Du Protein Nucleotide Hypothetical Hypothetical Hypothetical Hypothetical Hypothetical

Pyrophosphohydrolase

ORF # Top BLAST hit Viral Sequence with Greatest Homolgy 1 Enterobacteria phage phi92 (2e-06) Enterobacteria phage phi92 (2e-06) 2 gamma proteobactgerium SCGC AAA160-D02 (1e-27) Roseobacter phage SIO1 (6e-26) 3 Roseobacter phage SIO1 (7e-77) Roseobacter phage SIO1 (7e-77) 4 Dialister micraerophilus (9e-2) N/A 5 Roseobacter phage SIO1 (2e-101) Roseobacter phage SIO1 (2e-101) 6 N/A N/A 7 Vibrio cholerae (1e-19) Yersinia phage phiA1122 (9e-18) 8 Celeribacter phage P12053L (3e-50) Celeribacter phage P12053L (3e-50) 9 Synechococcus phage S-CRM01 (1e-29) Synechococcus phage S-CRM01 (1e-29) 10 Odoribacter laneus (6e-2) N/A 11 N/A N/A 12 Celeribacter phage P12053L (8e-29) Celeribacter phage P12053L (8e-29) 13 Puniceispirillum phage HMO-2011 (0) Puniceispirillum phage HMO-2011 (0)

Contig 12399

ORF1 ORF2 ORF3 ORF4 ORF5 ORF6 ORF7 ORF8 ORF9 ORF10 ORF11 ORF12 ORF13 ORF14 ORF15

Thymidylate Ribonucleotide DNA Primase DNA Helicase DNA Polymerase A Hypothetical Protein Exonuclease Synthase Reductase

500bp 3310 Protein ! Protein Protein Protein Protein Du Hypothetical Protein Hypothetical Hypothetical Hypothetical Hypothetical Hypothetical Endonuclease I

ORF # Top BLAST hit Viral Sequence with Greatest Homolgy 1 Puniceispirillum phage HMO-2011 (4e-103) Puniceispirillum phage HMO-2011 (4e-103) 2 Puniceispirillum phage HMO-2011 (0) Puniceispirillum phage HMO-2011 (0) 3 Polymorphum gilvum SL003B-26A1 (2e-09) N/A 4 Roseobacter phage SIO1 (6e-97) Roseobacter phage SIO1 (6e-97) 5 Celeribacter phage P12053L (1e-09) Celeribacter phage P12053L (1e-09) 6 N/A N/A 7 Puniceispirillum phage HMO-2011 (0) Puniceispirillum phage HMO-2011 (0) 8 Puniceispirillum phage HMO-2011 (1e-115) Puniceispirillum phage HMO-2011 (1e-115) 9 Puniceispirillum phage HMO-2011 (1e-134) Puniceispirillum phage HMO-2011 (1e-134) 10 N/A N/A 11 Puniceispirillum phage HMO-2011 (3e-56) Puniceispirillum phage HMO-2011 (3e-56) 12 Puniceispirillum phage HMO-2011 (1e-08) Puniceispirillum phage HMO-2011 (1e-08) 13 Olsenella profusa (1e-15) Salicola phage CGphi29 (2e-15) 14 N/A N/A 15 Paramecium bursaria Chlorella Virus OR0704.2.2 (7e-95) Paramecium bursaria Chlorella Virus OR0704.2.2 (7e-95)

158 Figure 4.8: Predicted ORFs on Rhode River contigs 12643, 5585, 8066, and 12399. These contigs contained Class I ‘Other’, clade II (contig 12643) and ‘RTPR’ (contigs 5585, 8066, 12399) RNR sequences. The contigs were assembled from ~50 million Illumina 2x150bp reads using MetaVelvet. ORFs were predicted using MetageneAnnotator. Annotations were assigned by consensus BLASTx results. ORFs without hits less than 1e-3 or lacking hits to definitive genes were annotated as hypothetical protein. ORFs with homology to phage sequences in the Caudovirales were colored by the viral family of the top BLASTx representatives. Scale bar represents 500 nucleotides.

159 Class II ‘RTPR’ Class I ‘Other’

Acanthocystis turfacea Chlorella virus 1 (7) 99 GMF1061042925962f_784_1_1 A. 31 B. 95 Pelagibacter phage HTVC019P GMF1061042949966r_830_1_1 33 GMF Contig 14 (3); DT (2); GM (9) Candidatus Pelagibacter ubique HTCC1002 (27) 100 CBJ1098101648625_866_21_1 Rhode River Contig 8066 CFB Contig 1 (13) 33 36 60 DTF1061043129216r_923_1_1 100 84 GMF Contig 24 (3) Clade I DTF Contig 5 (3); CB (1); DT (25); GM (2) 7 27 DTF1061042874584r_837_1_1 66 31 DTF1061043202875r_1_816_1 24 67 Chlamydomonas reinhardtii 42 GMF1061039282610r_1_967_1 100 GMF1061042969816r_1_904_1 58 84 Invertebrate iridescent virus 6 33 Volvox carteri f. nagariensis 100 Aedes taeniorhynchus iridescent virus Clade I Invertebrate iridescent virus 3 Rhode River Contig 5585 97 Caulobacter phage CcrColossus Caulobacter phage CcrRogue 90 6 100 Caulobacter phage phiCbK CFA Contig 1 (6); CB (21); DT (12); GM (13) 100 70 Caulobacter phage CcrMagneto 23 28 Caulobacter phage CcrSwift DTF Contig 15 (2) Caulobacter phage CcrKarma 65 100 Pseudomonas phage P3 CHA 29 Pseudomonas phage PAK P3 Nitrosococcus halophilus Nc4 46 56 Pseudomonas phage KPP10 98 CBS1061057326800r_735_1_1 Roseiflexus castenholzii DSM 13941 (4) 38 Rhode River Contig 12643 44 GMF1061042971560f_1_840_1 CFA1061053812415f_675_1_1 Bacteriophage RM378 98 GMF1061042926426f_648_1_1 18 GMF1061042943969r_1_765_1 CBJ1098127015314915_1_1 100 GMF Contig 25 (3) 97 GMF1061043214682f_1_747_1 58 78 60 GMF Contig 2 (8) GMF1061042970635f_1_839_1 23 57 75 CBJ1098101801525_1_814_1 87 100 GMF1061042943894f_1_624_1 Diplosphaera colitermitum TAV2 (2) 49 GMF1061042957534r_1_794_1 Clade II 88 100 DTF Contig 3 (4) GMF Contig 9 (6) Salpingoeca sp. ATCC 50818 50 97 Clade II 70 40 GMF Contig 49 (2) 82 70 GMF1061042925690f_1_710_1 Monosiga brevicollis MX1 97 CFC Contig 1 (6) 88 38 95 CFA Contig 12 (2) CFD Contig 3 (5) GMF1061042661264f_1_935_1 96 100 DTF1061043183606r_811_1_1 32 GMF1061042924671r_733_1_1 GMF Contig 3 (8); DT (6); GM (21) 99 GMF Contig 22 (3) GMF1061042968032r_1_866_1 98 CFA1061053829314f_1_770_1 Thermus phage P23-45 100 100 Francisella novicida FTG (6) Deftia phage phiW-14 89 Thermus phage P74-26 100 Grouper iridovirus Singapore grouper iridovirus 86 55 Rana tigrina ranavirus Alkaliphilus metalliredigens QYMF (2) 73 Regina ranavirus 100 61 100 61 95 Ambystoma tigrinum stebbensi virus Roseobacter phage SIO1 67 Frog virus 3 58 Soft-shelled turtle iridovirus 99 Lymphocystis disease virus 1 100 CFA1061053822165r_5_739_1 Lymphocystis disease virus-isolate China 53 GMF Contig 4 (7) 20 100 Candidatus Protochlamydia amoebophila UWE25 Halophage AAJ-2005 Thermus aquaticus Y51MC23 (7) 56 Cyanophage S-TIM5 GMF1061042942020f_103_880_1; CB (1); DT (1); GM (2) GMF1061042927759r_1_807_1 100 Clade III 100 DTF1061042915561r_791_1_1 CFA1061053817669f_1_691_1 71 45 GMF1061043206175r_1_828_1 80 57 GMF1061043245027r_933_1_1 95 43 GMF1061042968114f_929_1_1 Clade III Mycobacterium phage Tiger (22) 51 62 GMF Contig 30 (3) 100 99 100 GMF1061042926741f_1_727_1 100 Mycobacterium phage Gladiator (16) GMF Contig 42 (2) GMF Contig 56 (2) 100 Blattabacterium sp. (Blattella germanica) str. Bge (7) Rhodococcus phage RGL3 (2) 100 Vibrio phage pVp-1 CFC Contig 5 (2) 98 Aeromonas phage phiAS4 Ectocarpus siliculosus (9) Aeromonas salmonicida bacteriophage 25 Aeromonas phage 31 DTF1061043203475r_1_800_1; DT (1); GM (9) Clade IV Aeromonas phage 44RR2.8t 0.4 0.2 Figure 4.9: Unrooted maximum likelihood trees with 100 bootstrap replicates of Rhode River RNRs from contigs >5kb. Panel A) ‘RTPR’ RNRs from contigs 5585 and 8066 (bold) with Class II RNR reference and putative metagenomic ‘RTPR’ sequences. Metagenomic sequences were clustered at 80% identity. Representative metagenomic sequences were placed on the tree, with the number of reads from each environment within that cluster listed. CB (Chesapeake Bay); DT (Dry Tortugas); GM (Gulf of Maine). Bacterial references were clustered at 80% identity. Representative sequences were placed on each tree. Panel B) Rhode River contig 12643 (bold) on the Class I alpha ‘Other’ tree. Numbers in parentheses following metagenomic contigs indicate the number of reads assembled in each contig. Bacterial references were clustered at 80% identity. Representative sequences were placed on the tree. Numbers in parentheses following bacterial references indicate the number of reference sequences within that cluster. Scale bar represents amino acid substitutions per site. Bacteria (purple); Eukaryotic viruses (orange); Myoviruses (red); Siphoviruses (blue); Podoviruses (green); Metagenomic sequences (black). Integer values are bootstrap support values.

160 A. CB (dsDNA) CB (ssDNA) DT (dsDNA) DT (ssDNA) Bacteria Archaea/Eukaryote/Unclassified dsDNA virus ssDNA virus

Domain Other/Unclassified virus 18% 48% 16% 25%

Myoviridae Siphoviridae Podoviridae Other/Unclassified

dsDNA viruses dsDNA 37% 84% 31% 73%

38% 94% 42% 93%

Cyanophage-like Other Podoviridae

B. RNR DNA Pol DNA primase/ helicase Endonuclease polymerase RNA 2000 4000 6000 8000 10000 12000 14000 16000 18000 20000 22000 24000 26000 28000 30000 32000 34000 36000 38000 40000 42000 44000

24 33

Prochlorococcus phage P-GSP1 2000 4000 6000 8000 10000 12000 14000 16000 18000 20000 22000 24000 26000 28000 30000 32000 34000 36000 38000 40000 CB (dsDNA) CB (ssDNA) 50 270 DT (dsDNA) DT (ssDNA)

Synechococcus phage S-CBP4 2000 4000 6000 8000 10000 12000 14000 16000 18000 20000 22000 24000 26000 28000 30000 32000 34000 36000 38000 40000 42000 44000

47 225 25 26

Synechococcus phage S-RIP1 2000 4000 6000 8000 10000 12000 14000 16000 18000 20000 22000 24000 26000 28000 30000 32000 34000 36000 38000 40000 42000 44000

52 183

Synechococcus phage S-RIP2

161 Figure 4.10: Taxonomic distribution and alignment of ssDNA virome sequences. Panel A) Taxonomic distribution of translated ORFs in Chesapeake Bay and Dry Tortugas dsDNA and ssDNA virome libraries. Chesapeake Bay libraries CFA-CFD were combined for taxonomic composition analysis. Podoviral sequences with a top BLASTp hit to known cyanopodoviral sequences were categorized as cyanophage- like. Panel B) Recruitment of dsDNA and ssDNA virome library reads to cyanopodoviral genomes. Reads from Chesapeake Bay dsDNA libraries CFA-CFD were combined prior to mapping. Reads were mapped to each genome independently. Maximal coverage values of reads from dsDNA libraries and ssDNA libraries against each genome are listed on the left and right sides of each plot, respectively. Genomes were aligned with Mauve. Colors on horizontal axes are aligned regions.

162 TABLES

Table 4.1: Distribution frequency of putative virioplankton RNR sequences among designated groups by environment.

Viral Metagenome Libraries dsDNA ssDNA RNA GMa CBb DTc CBb DTc CBb DTc Cyano Class I 6% 12% 2% - - - - Class II 17% 24% 22% 74% 73% 70% 69% Totald 24% (47) 36% (36) 24% (21) 74% (6) 73% (10) 70% (3) 69% (3) Other Class I 31% 21% 16% 5% 5% 12% - Class II 11% 9% 7% 11% 3% - 31% Totald 42% (83) 30% (30) 23% (20) 16% (1) 8% (1) 12% (<1) 31% (1) RTPR (Class II)d 34% (68) 34% (34) 53% (46) 9% (1) 19% (3) 18% (1) - All Groups Class I : Class II 0.60 0.49 0.23 0.06 0.06 0.13 -

Table 4.2: Frequency of sampled genomes and RNR alpha subunits in cyanophage and pelagiphage populations.

# Predicted RNR Genes # Predicted Genomes Predicted RNR Frequency Cyano I 22 16 138%

Cyano II 63 85 74%

Pelagiphage HTVC008M 4 6 HTVC010P - 44 HTVC011P - 15 HTVC019P 36 11 Pelagiphage Total 40 76 53%

163 Table 4.3: Distribution frequencies of RNR alpha subunit sequences among designated groups.

Library CBB CBJ CBR CBS CFA CFB CFC CFD CIA DTF DTR DTS GMF Nucleic acid dsDN dsDN dsDN dsDNA dsDNA RNA ssDNA dsDNA dsDNA dsDNA RNA ssDNA dsDNA type A A A Cyano 3% 20% Class Ia 25% (5) 16% (4) - - 4% (1) 5% (1) 8% (1) 2% (2) - - 6% (13) (<1) (<1) 50% 74% 25% 34% 16% 22% 69% 73% 17% a 10% (2) 70% (3) 9% (2) - Class II (10) (6) (4) (4) (2) (19) (3) (10) (34) 75% 74% 28% 39% 19% 17% 20% 24% 69% 73% 24% a 26% (6) 70% (3) Total (15) (6) (5) (5) (2) (3) (<1) (21) (3) (10) (47)

Other 5% 19% 18% 27% 29% 27% 16% 31% a 5% (1) 30% (7) 12% (<1) - 5% (1) Class I (<1) (3) (2) (3) (5) (<1) (14) (62) 11% 11% 31% 11% Class IIa 4% (1) 11% (3) - 8% (1) 9% (1) 9% (1) - 7% (6) 3% (<1) (1) (2) (1) (21) 41% 12% 16% 27% 27% 36% 40% 27% 23% 31% 42% a 9% (2) 8% (1) Total (10) (<1) (1) (4) (3) (4) (7) (<1) (20) (1) (83)

RTPR (Class 45% 34% 45% 43% 53% 53% 19% 34% 16% (3) 33% (7) 18% (1) 9% (1) - II)a (7) (4) (4) (7) (1) (46) (3) (68)

All RNRs Class I : 0.43 0.84 0.13 0.06 0.30 0.30 0.44 0.57 0.90 0.23 - 0.06 0.60 Class II

Table 4.4: Characterization of viral metagenomic libraries queried.

Viral Avg. Read Sequencing Location Library* Sample Date Genome Mbp Reads Length Citations Technology Type (bp) Gulf of GMF 01/27/2006 dsDNA Sanger 49.9 63,950 780 This study Maine

(Bench, Chesapeake CBB 09/27/2002 dsDNA Sanger 3.90 5,615 695 Hanson et al. Bay 2007) (Rusch, CBJ 10/26/2004 dsDNA Sanger 11.65 11,484 1,014 Halpern et al. 2007) (Andrews- Pfannkoch, CBR 07/30/2007 RNA Sanger 4.27 5,729 746 Fadrosh et al. 2010) CBS 07/30/2007 ssDNA Sanger 4.36 5,712 763 (Andrews-

164 Pfannkoch, Fadrosh et al. 2010) (Andrews- Pfannkoch, CFA 07/30/2007 dsDNA Sanger 12.96 20,283 715 Fadrosh et al. 2010) (Andrews- Pfannkoch, CFB 07/30/2007 dsDNA Sanger 12.81 20,133 711 Fadrosh et al. 2010) (Andrews- Pfannkoch, CFC 07/30/2007 dsDNA Sanger 13.79 21,394 714 Fadrosh et al. 2010) (Andrews- Pfannkoch, CFD 07/31/2007 dsDNA Sanger 13.78 20,989 740 Fadrosh et al. 2010) CFE 07/30/2007 dsDNA 454 34.65 83,197 416 This study CFF 07/30/2007 dsDNA 454 89.24 221,530 403 This study CFG 07/30/2007 dsDNA 454 43.42 122,887 353 This study CFH 07/31/2007 dsDNA 454 73.40 193,829 379 This study CIA 07/30/2007 dsDNA Sanger 1.62 3,245 501 This study

(Andrews- Dry Pfannkoch, DTF 01/08/2004 dsDNA Sanger 49.9 64,833 775 Tortugas Fadrosh et al. 2010) (Andrews- Pfannkoch, DTR 01/08/2004 RNA Sanger 4.22 5,470 771 Fadrosh et al. 2010) (Andrews- Pfannkoch, DTS 01/08/2004 ssDNA Sanger 4.23 5,729 739 Fadrosh et al. 2010)

165 Chapter 5

VIRAL DIVERSITY IN OYSTERS IS ENDEMIC AND SEASONALLY DYNAMIC

5.1 Abstract

The eastern oyster, Crassostrea virginica, is a keystone species in productive estuarine environments that contain abundant virio- and bacterioplankton populations. Despite direct interaction between oysters and the external environment, little is known about the diversity, persistence, and distributions of non-pathogenic viruses in oysters. In this study, viral populations in oyster extrapallial fluid were characterized by the diversity of translated ribonucleotide reductase (RNR) α-subunit gene sequences. Viral RNR genes were PCR-amplified from extrapallial fluid (EF) collected from oysters at different times and locations around the Delmarva Peninsula. Subsequently, amplicons were sequenced using the PacBio RS. Viral RNR translated peptide sequences were clustered with RNRs from 59 globally distributed aquatic viral metagenome (virome) libraries and one aquatic RNR amplicon library. In total, 380 RNR peptide clusters were identified in oyster EF, and 66% (249) of these clusters lacked any virome representative sequences. RNR peptides amplified from oysters were identified as belonging to T7-like podoviruses but shared homology with few viral references in GenBank. Thus, the sequences amplified by RNR primers in oyster EF comprised a novel and underrepresented realm of podoviral diversity. Abundant and persistent viral populations in oysters were largely endemic to the Delmarva region, while the majority of sampled viral diversity appeared to be allochthonous. In

166 monthly samples from a single site spanning 12 months, 52% of clusters were identified in only a single month, and only 12% of RNR clusters contained sequences from ≥ 5 time points. However, these persistent clusters comprised 17 of the 20 most abundant clusters and 71 ± 14% of the sampled assemblage over 12 months. Within the Delmarva, viral populations in oyster EF were widely distributed. More than two- thirds of RNR peptide sequences from oysters sampled concurrently in the Choptank River (MD), Indian River (DE), and Delaware Bay were identified in at least two sites but displayed site-specific abundance profiles. The abundance and distribution of viral diversity identified in oysters is consistent with the Bank model of viral diversity and indicates environmental selection for specific phage populations in this keystone species.

5.2 Introduction

The eastern oyster, Crassostrea virginica, plays a vital role in ecosystem health along the east coast of North America. Highly efficient filter feeders, an individual oyster can remove suspended particulate matter from up to 50 gallons of water daily (NOAA), resulting in decreased turbidity and improved conditions for submerged aquatic vegetation (Grabowski and Peterson 2007; Grizzle, Greene et al. 2008). Beyond water filtration, oysters serve as “ecosystem engineers” by providing hard substrate for the attachment of other sessile invertebrates and habitat for fish and invertebrate species (Lenihan and Peterson 1998), thus increasing local biodiversity in estuarine environments (Rodney and Paynter 2006; Stunz, Minello et al. 2010). However, declining C. virginica populations have contributed to impaired ecosystems (Newell 1988). A recent survey of historical and current oyster populations across the United States found oyster habitat and biomass have declined 64% and 88%,

167 respectively, since the late 19th century (Zu Ermgassen, Spalding et al. 2012). In the Chesapeake Bay, C. virginica biomass may be as low as 1% of historic levels (Newell 1988). Initially due to overharvesting and habitat degradation (Rothschild, Ault et al. 1994), C. virginica populations have been largely impacted the past few decades by disease (Andrews 1996). In particular, diseases MSX and Dermo – caused by protozoan parasites Haplosporidium nelsoni and Perkinsus marinus, respectively – can result in high oyster mortality rates and altered range (Powell, Ashton-Alcox et al. 2008). In a 2013 survey of the Chesapeake Bay, 98% of sites sampled contained detectable levels of P. marinus (Tarnowski 2014). Wider geographic distribution of both pathogens has also been observed along the east coast of the United States in recent years (Burge, Mark Eakin et al. 2014). Improving the future outlook for C. virginica populations will require a detailed understanding of the factors influencing oyster health. Increasingly, microbial communities commensal with metazoan hosts (i.e. microbiome communities) have been implicated in various factors surrounding host health, including nutrient acquisition (Turnbaugh, Ley et al. 2006), molecule production (Hill 1997; Shimada, Kinoshita et al. 2013), and resistance to pathogens (Bachere 2003). In the Chesapeake Bay, bacterial abundances range from 106 – 107 cells mL-1 (Shiah and Ducklow 1994; Heidelberg, Heidelberg et al. 2002). As filter feeders, oysters are intimately associated with environmental bacteria, and filtering 50 gallons of water would place an oyster in contact with 1011 – 1012 bacterial cells daily. Cultivation-dependent (Colwell and Liston 1960; Kueh and Chan 1985; Pujalte, Ortigosa et al. 1999; Thomas, Wafula et al. 2014) and -independent (La Valley, Jones et al. 2009) (Sakowski et al., in prep.) studies have indicated that bacterial communities in oysters are distinct from the

168 surrounding water. Biotic and abiotic factors, including geography (King, Judd et al. 2012; Trabal, Mazon-Suastegui et al. 2012), genotype (Wegner, Volkenborn et al. 2013), stress (Wegner, Volkenborn et al. 2013; Lokmer and Mathias Wegner 2015), and temperature (Pujalte, Ortigosa et al. 1999; La Valley, Jones et al. 2009) can shape the oyster microbiome. However, the role of viruses in microbiome communities remains largely unexplored. In terms of sheer numbers, viruses are the most successful biological entities on the planet, outnumbering bacteria 10:1 in marine environments (Wommack and Colwell 2000) and by even more in terrestrial systems (Srinivasiah, Bhavsar et al. 2008). Viruses were also more abundant than bacteria in oyster extrapallial fluid by a factor of ≥30 (Sakowski, unpublished data). As such, viruses have an ecological impact disproportionate to their size. Through cell lysis, viruses regulate microbial community composition (Sandaa, Gomez-Consarnau et al. 2009). According to the Kill-the-Winner model, top-down regulation by viral predation increases bacterial diversity by reducing the abundance of the most competitive community members (Vage, Storesund et al. 2012). Additionally, viral lysis influences biogeochemical cycles and nutrient availability by increasing the pool of dissolved organic and inorganic matter available to community members in the microbial loop (Suttle 2005), and viruses are key facilitators of horizontal gene transfer (Rohwer and Thurber 2009).

Despite their abundance and impact on microbial communities, little is known about the diversity of non-pathogenic viruses in oysters. An examination of Vibrio phage revealed seasonally dynamic populations (Comeau, Buenaventura et al. 2005), and several studies of Vibrio phage in oysters detected the presence of phages even in the absence of detectable hosts (DePaola, McLeroy et al. 1997; DePaola, Motes et al.

169 1998; Comeau, Buenaventura et al. 2005). Incredibly long retention times – on the order of weeks – have been observed for oysters inoculated with coliphage T2 (Acton and Evans 1968) and Staphylococcus aureus phage 80 (Feng 1966). In contrast, the turnover of viral assemblages in the Chesapeake Bay is approximately 24 hours (Winget and Wommack 2009). Thus, the oyster microenvironment appears to increase the stability of viral particles, which largely degrade due to UV radiation (Noble and Fuhrman 1997), and oyster samples may serve as more accurate indicators of total viral diversity in aquatic environments where rapid turnover and heterogeneous distributions may influence water sample analyses. One of the challenges to characterizing viral diversity remains the absence of any universal marker gene equivalent to the 16S or 18S rRNA gene in cellular organisms. However, genome metabolism genes show promise as abundant, phylogenetically diverse, and biologically informative markers of viral diversity (Wommack, Nasko et al. 2015). In particular, ribonucleotide reductase (RNR) - responsible for converting ribonucleotides to deoxyribonucleotides (Nordlund and Reichard 2006) - is abundant in aquatic viral metagenomes (viromes) (Angly, Felts et al. 2006; Dwivedi, Xue et al. 2013; Sakowski, Munsell et al. 2014) and widely distributed among lytic dsDNA tailed phages (the Caudovirales). In the Chesapeake Bay the most abundant viral RNRs identified were Class II (oxygen-independent,

B12-dependent), monomeric enzymes classified as ‘RTPR’ RNRs (belonging to TIGRfam02505) (Sakowski, Munsell et al. 2014). These virome sequences shared homology with few known references, but were hypothesized to be from T7-like podoviruses infecting Alphaproteobacteria (Sakowski, Munsell et al. 2014), a class of bacteria abundant in both oyster and water samples (Pujalte, Ortigosa et al. 1999;

170 Morris, Rappe et al. 2002; Green and Barnes 2010; Fernandez-Piquer, Bowman et al. 2012; Wegner, Volkenborn et al. 2013; Chauhan, Wafula et al. 2014). In this study, viral diversity was characterized in C. virginica extrapallial fluid (EF; fluid found between the inner shell surface and the mantle) over time and between locations by PCR of the RTPR RNR gene.

5.3 Materials and Methods

5.3.1 Annual survey sample collection

Cultured oysters (C. virginica) were obtained from Marinetics, Inc. (Cambridge, MD) and placed in wire cages on September 23, 2011. Cages contained 33 oysters each and were suspended approximately one meter below the surface of the Rhode River at the Smithsonian Environmental Research Center (SERC) in Edgewater, MD. Oysters were allowed to acclimate prior to sampling. Three oysters were harvested monthly between December 2011 and December 2012 with the exception of June, August, and November 2012. Oysters were transported 20 minutes to Annapolis, MD for processing. Each oyster was rinsed with DI water and scrubbed with 70 percent ethanol prior to extrapallial fluid (EF) extraction. A hole was drilled into the posterior end of the oyster at the interface between valves with a 3/32-inch drill bit. Extrapallial fluid was extracted using a 5mL syringe. Samples were placed on ice and transported back to Newark, DE where they were snap-frozen in liquid nitrogen (LN2) and stored at −80°C until further processing. Upon thawing, EF samples from the three oysters were combined for viral isolation and concentration.

171 5.3.2 Biogeographic survey sample collection

Oysters were harvested from the Choptank River (MD), Indian River (DE), and Delaware Bay in June 2013. Thirty-five oysters were collected from the Choptank River on June 24, 2013; 29 oysters from the Delaware Bay on June 26, 2013; and 39 oysters from the Indian River on June 26, 2013. Oysters were transported on ice back to Newark, DE and subsequently rinsed with DI water and scrubbed with 70 percent ethanol. Extrapallial fluid was extracted as described above and pooled by site.

5.3.3 Isolation and concentration of viruses from extrapallial fluid

Viruses were isolated and concentrated slightly differently for annual survey samples and biogeographic survey samples due to volume differences (3 combined oysters vs. ≥29 oysters). Extrapallial fluid samples were diluted 1:10 with 1x PBS buffer and rocked for 30 minutes at 30°C to dissociate viruses from mucopolysaccharides. Samples were then filtered through a Millipore Sterivex

0.22µm filter unit to remove bacterial cells. Viruses were concentrated using the FeCl3 method as previously described (John, Mendez et al. 2011) with amendments. Briefly,

-1 FeCl3 (4.83g L ) was added to the viral retentate at a concentration of 1µL per 1mL of sample (annual survey samples) or 1µL per 5mL of sample (biogeographic survey oysters). Samples were incubated for one hour and then filtered onto a 25mm, 1.0µm

(annual survey) or 140mm, 0.8µm (biogeographic survey) polycarbonate filter (Whatman). Filters were stored in the dark at 4°C until processing. Viruses were resuspended as described (John, Mendez et al. 2011) in ascorbic acid buffer. Resuspended viral particles were concentrated by ultracentrifugation (Millipore Amicon Ultra) to 200-250µL and dialyzed (Thermo Slide-A-Lyzer 10K MWCO) overnight at 4°C in 14mL sterile SM buffer to remove downstream PCR inhibition

172 from excess iron. Following dialysis, viral concentrates were DNased (Ambion) to remove free DNA. The retentate was 0.22µm-filtered again to remove any remaining or introduced cells. Viral particles were added directly to PCR reactions (see, for example Winget and Wommack 2008) for annual survey samples due to low volumes. DNA was isolated from biogeographic survey viral concentrates by phenol:chloroform:isoamyl alcohol (25:24:1) extraction and ethanol precipitation (see below).

5.3.4 Isolation of viral DNA

Viral DNA was isolated for biogeographic survey samples by phenol:chloroform:isoamyl alcohol (25:24:1) addition followed by chloroform addition with centrifugation at 14,000 rpm for 10 minutes. DNA was precipitated by incubation with 100% ethanol at 4°C followed by centrifugation at 14,000 rpm for 60 minutes and two subsequent washes with 70% ethanol. DNA was resuspended in EB buffer.

5.3.5 PCR amplification of RTPR RNR sequences

Primers with minimal degeneracy were developed to amplify an approximately 750bp fragment from RTPR RNR sequences identified in marine viromes (Sakowski, Munsell et al. 2014). Forward primers F1 (5’ – CCAAGWCKGGYSARTGGTGGG – 3’), F2 (5’ – CTAAATCAGGTGAKTGGTGGG – 3’), F3 (5’ – CTAAGWCAGGWSANTGGTGGR – 3’), and F4 (5’ – GTAARTCWGGYAAYTGGTGGG– 3’) were mixed together to a final concentration of 10µM each. Reverse primers R1 (5’ – GTAACMGAKGGCTTGTGYTC – 3’), R2 (5’ – GTCAYAGAWGGCTTRTGYTC – 3’), R3 (5’ –

173 GTGRCAGAARSWTTGTGWKY – 3’), and R4 (5’ – GTTAYASWAGGTTTRTGTTC – 3’) were mixed together to a final concentration of 10µM each. One µL of viral concentrate (annual survey) or isolated DNA

(biogeographic survey) was combined with 10X buffer (1X final concentration), dNTPs (0.25mM each), forward primer mix (0.1µM final concentration), reverse primer mix (0.1µM final concentration), and TaKaRa Ex Taq DNA Polymerase (1.25 U). RTPR RNR sequences were amplified as follows: 10 minutes at 95°C followed by 35 cycles at 95°C (45 seconds), 52°C (45 seconds), 72°C (90 seconds).

5.3.6 Adapter ligation and sequencing

16-mer adapter sequences with 3’-T overhangs were AT ligated to RTPR amplicons. Briefly, amplicons were cleaned and concentrated (Zymo Clean and Concentrator-5). DNA was quantified with the Qubit Fluorometer High Sensitivity assay (Life Technologies). Up to 10ng DNA per sample was combined with annealed adapters (15µM) and 3 units Ligase (Promega) and incubated for 12 hours at 15°C. Ligated amplicons were column cleaned (Zymo Clean and Concentrator-5). Subsequently, 1µL of ligated amplicon product was PCR amplified for 20 cycles using the barcode as the primer to enrich for ligated products prior to sequencing. Finally, amplified products were gel-purified (QIAquick Gel Extraction kit) and 30ng of DNA per sample was pooled for sequencing on the PacBio RS.

5.3.7 Amplicon filtering and error correction

PacBio sequence reads were filtered by circular consensus sequence (CCS) coverage. Only reads with at least 3x CCS coverage were screened for primer and barcode sequences. Reads with at least 85% identity to primer sequences within the

174 first 50bp were retained for barcode deconvolution. Retained reads were iteratively screened without replacement for 16-mer barcode sequences within the first 20bp at 100%, 95%, and 90% identity. Barcodes were trimmed from retained reads. Reads were screened for indels and corrected with LAST (Kielbasa, Wan et al. 2011). Sequences were aligned with MAFFT using the FFT-NS-i x 1000 algorithm (Katoh, Misawa et al. 2002). Amplicons were queried for the presence of key catalytic residues C439, E441, and C462 (E. coli numbering) and trimmed to the region of interest (N437 to S625 in E. coli). In total, 8,221 amplicon RTPR sequences from 14 amplicon libraries spanned the region of interest and contained all key residues.

5.3.8 Identification of virome RTPR sequences

Publicly available aquatic DNA viral metagenomic libraries from VIROME (http://virome.dbi..udel.edu) (Wommack, Bhavsar et al. 2012) and MetagenomesOnline (http://metagenomesonline.org) (Wommack, Bhavsar et al. 2012) were independently assembled with the Celera assembler (Myers, Sutton et al. 2000). In addition, a deeply sampled virome library from the Rhode River was assembled with the Celera assembler (Myers, Sutton et al. 2000). ORFs were predicted for all virome libraries from contigs and unassembled reads using MetageneMark (Zhu, Lomsadze et al. 2010) and then translated and queried (BLASTp; Altschul, Gish et al.

1990) by library against a UniRef90 RNR database (Suzek, Huang et al. 2007). RNR sequences were grouped by Conserved Domain BLAST (Marchler-Bauer, Lu et al. 2011), and sequences identified as RTPR RNRs (belonging to TIGRfam02505) were retained. The RTPR RNR sequences were aligned with reference sequences in MAFFT and trimmed as described above. Retained sequences were subsequently re- aligned with previously identified RTPR virome sequences from the Chesapeake Bay,

175 Dry Tortugas, and Gulf of Maine (Sakowski, Munsell et al. 2014). Peptide sequences that did not span this region were discarded. In total, 839 sequences from 59 virome libraries spanned the region of interest and contained all key residues.

5.3.9 Clustering of RTPR peptides

9,060 virome and amplicon RTPR peptide sequences spanned the region of interest and contained key catalytic residues. All retained sequences were aligned with MAFFT using the FFT-NS-i x 1000 algorithm (Katoh, Misawa et al. 2002). Peptide clusters were picked using the furthest-neighbor algorithm in mothur (Schloss, Westcott et al. 2009) with a similarity cutoff of 95%. Representative sequence reads were picked for each cluster from the cluster seed.

5.3.10 Jackknifing of virome sequences

469 virome sequences were from local libraries sampled at the Rhode River and Chesapeake Bay, while 370 virome sequences were from all other virome libraries. Local virome sequences were randomly sub-sampled to assess whether sampling depth impacted the virome libraries with which amplicons clustered. Virome sequences were categorized as Rhode River/ Chesapeake Bay (RR/CB) or all other virome libraries (global). RR/CB sequences were randomly sub-sampled 10 times with replacement at a sampling depth equal to that of the global sequences (370 sequences). Each random RR/CB group of sequences was independently clustered with amplicon and virome sequences at 95% identity in mothur (Schloss, Westcott et al. 2009) as described above.

176 5.3.11 Identification of reference sequences

Representatives from the 50 most abundant peptide clusters identified in the annual survey and biogeographic survey were queried against the Genbank nr database (BLASTp on 5-3-15) excluding uncultured/environmental sequences. The top 10 hits for each sequence were identified, and those references hit by at least 5 clusters were included in phylogenetic analyses. Bacterial references identified as top hits were queried for possible phage elements. Assembled contigs from single amplified genomes (SAGs) with RTPR RNRs homologous to amplicon peptides were identified by BLASTp (Altschul, Gish et al. 1990). Subsequently, SAG sequence data was re- assembled with the SPAdes assembler (Bankevich, Nurk et al. 2012), and phage-like contigs were manually annotated by Conserved Domain BLAST (Marchler-Bauer, Lu et al. 2011) and top BLASTx (Altschul, Gish et al. 1990) homology. Putative phage- like elements in bacterial references were manually aligned with the genome of Puniceispirillum phage HMO-2011.

5.3.12 Phylogenetic analyses of amplicon and virome sequences

Amplicon and virome peptide sequences were aligned with MAFFT (Katoh, Misawa et al. 2002) as described above. Sequences were clustered with the furthest neighbor algorithm in mothur (Schloss, Westcott et al. 2009) at 80% identity. Representative sequences from peptide clusters were re-aligned with references as described above. A maximum likelihood tree with 100 bootstrap replicates was built in Geneious (v.5.9.7) (Drummond, Ashton et al. 2011) using PhyML (Guindon, Dufayard et al. 2010) with the LG substitution model.

177 5.3.13 Hierarchical clustering of RTPR virome contigs

Viromes were queried for contigs >10kb with RTPR RNRs. Only the Rhode River virome contained contigs (n = 91) >10kb with the RTPR RNR gene. ORFs were predicted for virome contigs and representative reference sequences by GeneMarkS (Besemer, Lomsadze et al. 2001) and translated. Translated virome contig and reference ORFs were clustered by reciprocal BLASTp (1 e -30). Thirty clusters were randomly sampled per genome/contig in QIIME (Caporaso, Kuczynski et al. 2010). Distances were calculated by Bray-Curtis similarity. Contigs and reference genomes were clustered by UPGMA hierarchical clustering from 1,000 jackknifed replicates.

5.3.14 Alpha diversity of extrapallial fluid samples

Estimates of within-sample (alpha) diversity were calculated by observed species (clusters). Libraries were analyzed using EstimateS v. 9.10 (Colwell 2013) with 100 iterations. The number of observed clusters was extrapolated for all libraries to the sampling depth of the largest library (Choptank River; 1,588 amplicon sequences). Chao1 and Shannon indices, and Faith’s Phylogenetic Diversity were calculated for extrapallial fluid samples in QIIME (Caporaso, Kuczynski et al. 2010) at a sampling depth of 112 sequences per library. Mean values were calculated by library for each metric from 100 iterations.

5.3.15 RTPR assemblage composition of extrapallial fluid samples

RTPR assemblage composition was defined as the relative abundance of each peptide cluster observed in a given sample. UniFrac distances (Lozupone and Knight 2005) were calculated in QIIME (Caporaso, Kuczynski et al. 2010). The change in UniFrac distance over time was calculated by grouping pair-wise sample distances according to the number of months between samples. For example, the mean distance

178 between all samples one month apart included the following pair-wise comparisons: Dec. 2011-Jan.; Jan.-Feb.; Feb.-March; March-April; April-May; July-Aug.; Aug.- Sept.; Sept.-Oct.

5.3.16 Relative abundance correlations with environmental parameters

Environmental parameters were recorded from the Smithsonian Environmental Research Center long-term monitoring station (courtesy of Charles Gallegos, Smithsonian Environmental Research Center). RTPR peptide clusters observed in at least five sample months were retained for correlation analyses. The mean monthly abundance of each peptide cluster was correlated with monthly salinity, dissolved oxygen (DO), pH, and temperature values by Spearman’s Rank correlation. R values for each correlation were used to create a profile for each peptide cluster. Peptide clusters were hierarchically clustered using the Heatmap.2 package with default settings.

5.4 Results

5.4.1 Diversity of viral RTPR RNR sequences in oyster EF

Amplicon and virome RTPR RNR sequences were clustered at 95% amino acid identity and formed 700 peptide clusters (Fig. 5.1B). Amplicons from oyster EF largely formed unique clusters that lacked representatives from water-based virome and amplicon libraries. Viral RTPR amplicon sequences from oyster EF were collectively distributed among 380 peptide clusters, and 66% of these clusters (249 clusters) were unique to oyster EF samples. Likewise, 71% of peptide clusters with water sequences (320 clusters) lacked oyster amplicon representatives (Fig. 5.1B). Amplicon sequences were almost exclusively part of a single diverse clade (Fig. 5.2).

179 This clade was also well-represented by virome sequences and comprised 74% of contig/singleton RTPR sequences from Rhode River/ Chesapeake Bay virome libraries (RR/CB group) and 69% of contig/singleton RTPR sequences from all other virome libraries (Global virome group) (Fig. 5.2). Oyster EF RTPR populations (clusters) were consistently most similar to twelve reference sequences (Fig. 5.2, 5.3A). Of these, only four were viruses, and only two were phage (Puniceispirillum phage HMO-2011 and Streptomyces phage Jay2Jay). Two of the references belonged to green algal species (Volvox carteri f. Nagariensis and Chlamydomonas reinhardtii). The remaining six references were bacteria, including single amplified genomes (SAGs) gamma proteobacterium SCGC AAA076-P13, alpha proteobacterium SCGC AAA536-G10, Verrucomicrobia bacterium SCGC 164-O14, and Verrucomicrobia bacterium SCGC AAA300-N18. Three distinct RTPR peptides that shared homology with amplicon sequences were identified in SAG Proteobacteria bacterium JGI 0000113-P07 (Fig. 5.2). These RNRs were part of phage-like scaffolds that shared synteny and homology with Puniceispirillum phage HMO-2011 and an uncultured Mediterranean phage (Figs. 5.3B, 5.4; Table 5.1). Virome contigs with RTPR RNRs contained ORFs with homology to known podoviruses. In particular, contigs that were part of the same RNR peptide clusters as oyster EF amplicons were most similar to podoviruses Puniceispirillum phage HMO-

2011, Roseobacter phage SIO1, and Celeribacter phage P12053L (Fig. 5.4).

5.4.2 Geographic distribution of RTPR sequences in oyster EF

RTPR RNR sequences were globally distributed in aquatic virome libraries (Fig. 5.1). Viral populations identified in oyster EF were largely endemic to the Delmarva region (Fig. 5.5). In particular, virome sequences that clustered with

180 persistent (≥7 months) and seasonal (4-6 months) oyster EF phage populations were most often exclusively from Rhode River/ Chesapeake Bay virome libraries (RR/CB) (Fig. 5.5B). There was a significant (r2= 0.66, p < 0.01) correlation between the persistence of phage populations and the difference in the proportion of populations shared with local (RR/CB) and global virome sequences (% amplicon clusters shared with RR/CB sequences - % amplicon clusters shared with global sequences; Fig. 5.5C). Viral populations were widely distributed between oysters sampled around the Delmarva Peninsula. In a comparison of all three sites from the biogeographic survey (see Materials and Methods), only the Choptank River contained >50% unique clusters (Fig. 5.6A). More than half of these unique clusters from the Choptank River contained sequences from oysters sampled at the Rhode River (annual survey, Fig. 5.6B), and >66% of the Choptank River oyster EF sequences belonged to populations present in at least one other biogeographic survey site (Fig. 5.7A). Similarly, >78% of Indian River and DE Bay oyster EF viral assemblages were comprised of populations identified at multiple sites (Fig. 5.7A). However, the relative abundance of these shared populations were not evenly distributed geographically (Fig. 5.7B). Overall, the ten most abundant clusters at each site comprised a total of 23 clusters (Fig. 5.7B). Only one of the nine peptide clusters common to all sites (Fig. 5.6A) was among the ten most abundant clusters at each site, and only six clusters overall were one of the ten most abundant clusters in two or more sites (Fig. 5.7B). In contrast, 17 clusters were among the ten most abundant clusters at a single site, though only seven were unique to that site (3 Choptank River, 2 Indian River, and 2 DE Bay samples) (Fig. 5.7B).

181 5.4.3 Alpha diversity of RTPR viral assemblages in oyster EF

Alpha diversity of RTPR viral assemblages was compared by library. On average, most libraries were predicted to have approximately 100 RTPR populations (95% amino acid ID) amplified by the RTPR primers (Fig. 5.8). The richness, evenness, and phylogenetic diversity peaked in April and May and were lowest in July and September (Fig. 5.8B-D).

5.4.4 Persistence and relative abundance of RTPR viral populations over one year

The diversity of RTPR-carrying viral populations in oyster extrapallial fluid was largely transient. In total, 294 protein clusters were observed from Dec. 2011 to Dec. 2012 (Fig. 5.1B, 5.6B). Over half (152) of observed clusters were identified in a single sample month, and 70% (205) of all clusters were observed in two or fewer months (Fig. 5.7C). In contrast, 36 clusters (12% of clusters) were identified in five or more of the 10 months sampled (Fig. 5.7C). These persistent clusters accounted for 71 ± 14% of the RTPR-carrying assemblage each month (Fig. 5.7D) and 17 of the 20 most abundant clusters. The 50 most abundant peptide clusters accounted for 84 ± 7% of the assemblage, while the remaining 244 clusters accounted for 16 ± 7% of the assemblage.

5.4.5 Viral assemblage dynamics over time

Viral assemblage diversity displayed temporal patterns. Samples collected between January and April 2012 from the Rhode River clustered together by weighted UniFrac distances (Fig. 5.9A, 5.10). Similarly, samples from the Rhode River collected between July and December 2012 clustered together (Fig. 5.9A, 5.10). Choptank River, Indian River, and DE Bay libraries were more closely related to each

182 other than to Rhode River samples (Fig. 5.9A, 5.10). These libraries – sampled in June 2013 – were part of a larger group that included the Rhode River July - Dec. 2012 libraries (Fig. 5.9A) and fell along the same plane of the first coordinate as the July - September 2012 Rhode River samples (Fig. 5.10). Overall, the EF viral assemblage became less similar over time (Fig. 5.9B, C). Weighted UniFrac distances between samples increased as the amount of time between samples grew, peaking for samples collected 4-6 months apart (Fig. 5.9B). The addition of biogeographic survey samples to this analysis suggested a periodicity of approximately 12 months in oyster EF RTPR viral diversity (Fig. 5.9C). Among the 36 clusters observed in at least five months, relative abundance was most strongly correlated with temperature (Fig. 5.11, 5.12), and 11 of 36 clusters were significantly correlated (p< 0.05) with changes in temperature (9 of 11 negatively correlated). Salinity, the parameter with the second strongest overall correlation with relative abundance (Fig. 5.11, 5.12), was correlated (p<0.05) with 13 of 36 clusters (10 of 13 negatively correlated). Clusters were part of three different groups based on changes in their relative abundance over time (Fig. 5.11). The smallest group (Group III) consisted of clusters that were significantly negatively correlated with salinity and/or temperature (Fig. 5.11). Eight of the ten clusters in this group were significantly correlated with temperature (p<0.05) and all ten clusters were significantly correlated with salinity (p<0.05). Group I contained clusters that were positively associated with temperature and salinity. This group included all five clusters present in every sample month (persistence = 10) and 50% of clusters present in ≥7 sample months (Fig. 5.11). The second group (Group II) was comprised of clusters that showed no strong association with any of the measured physical

183 parameters, although they were weakly negatively associated with temperature (Fig. 5.11). Peptide clusters from the three physical parameter correlation groups did not display any clear phylogenetic relationship (Fig. 5.13).

5.5 Discussion

5.5.1 Identity of RTPR RNR phage

The RTPR RNR group is abundant in virome libraries but poorly represented in reference databases, and RNR sequences amplified from oyster EF were most similar to the same few reference sequences (Figs. 5.2, 5.3A). Interestingly, many of the top BLAST hits for these populations were to bacteria, most notably SAG Proteobacteria bacterium JGI 0000113-P07 (Fig. 5.3A). Further investigation revealed three distinct RTPR sequences on three separate contigs for P07 (29.29, 55.75, and 59.59). Two of these contigs (29.29 and 55.75) shared homology and synteny with Puniceispirillum phage HMO-2011, a podovirus that infects the abundant SAR116 group of Alphaproteobacteria (Kang, Oh et al. 2013) (Fig. 5.3B). Given their homology to HMO-2011, their assembly into distinct contigs with no apparent flanking bacterial ORFs, and their size (46kb and 41kb), we believe these contigs to represent nearly complete genomes of free (non-integrated) HMO-like phages. The third contig identified in P07 (59.59) was distantly related to HMO-2011 and contigs 29.29 and 55.75 by UPGMA clustering of ORF composition (Fig. 5.4). This contig shared homology with several uncultured Mediterranean phage sequences and lacked flanking bacterial ORFs (Table 5.1), making it likely that this contig also represented a non-integrated phage.

184 Numerous phage genomes have been identified in bacterial SAGs (Roux, Hawley et al. 2014; Labonte, Field et al. 2015; Labonte, Swan et al. 2015). Other SAGs with RTPRs homologous with oyster EF amplicons included gamma proteobacterium SCGC AAA076-P13, alpha proteobacterium SCGC AAA536-G10, Verrucomicrobia bacterium SCGC AAA300-N18, and Verrucomicrobia bacterium SCGC AAA164-O14. RNR sequences from SAGs P13, G10 and O14 were the only ORF identified on their respective contig. An incomplete phage contig (39,559bp) classified as a member of the Podoviridae was previously identified in SAG O14 (Labonte, Swan et al. 2015). We hypothesize that the RNRs in these SAGs are also of viral origin. In the case of SAG O14, the RTPR RNR gene is likely part of the podoviral genome identified and failed to assemble with the larger phage contig. Phages with RTPR RNRs were previously hypothesized to be T7-like podoviruses based on the homology of a few virome contigs with podoviruses Puniceispirillum phage HMO-2011, Roseobacter phage SIO1, and Celeribacter phage P12053L (Sakowski, Munsell et al. 2014). In this study, the genomic context of virome contigs with RTPR RNRs was further examined. In agreement with previous observation, contigs with RTPR RNRs most often clustered with podoviruses HMO- 2011, SIO1, and P12053L by UPGMA hierarchical clustering of translated ORFs (Fig. 5.4). This clustering method also placed Proteobacterium bacteria JGI 0000113-P07 contigs 55.75 and 29.29 with HMO-2011, while contig 59.59 was more distantly related (Fig. 5.4). Interestingly, a small proportion of virome contigs clustered with myoviruses (Fig. 5.4). Reference myoviruses typically contain class I (oxygen- dependent) and/or class III (oxygen-sensitive) RNRs, the lone exception being Rhodothermus phage RM378, a Far-T4-like phage with an RTPR (class II) RNR.

185 Three of these virome contigs contained RNRs that clustered with oyster EF amplicons; yet, oyster amplicons in these clusters were exceedingly rare (Fig. 5.4). Thus, the vast majority of RTPR diversity in both oyster and water samples likely belonged to T7-like podoviral populations. RTPR podoviruses were also hypothesized to infect Alphaproteobacteria based on the hosts of HMO-2011 (Candidatus Puniceispirillum marinum), SIO1 (Roseobacter SIO67), and P12053L (Celeribacter marinus), as well as possible evidence of horizontal gene transfer from Alphaproteobacteria on virome contigs (Sakowski, Munsell et al. 2014). The dominance of Alphaproteobacteria in marine environments (Morris, Rappe et al. 2002) combined with the lytic lifestyle of T7-like phages offers a possible explanation for the abundance of the RTPR group in virome libraries. Likewise, Alphaproteobacteria are among the numerically dominant taxa detected in oysters (Pujalte, Ortigosa et al. 1999; Wegner, Volkenborn et al. 2013; Chauhan, Wafula et al. 2014; Sakowski et al., in prep.), making it likely that phages infecting this taxa would be present in oyster samples. However, difficulties surrounding cultivation make definitively linking viruses to their hosts difficult, although emerging methodologies like viral tagging (Deng, Ignacio-Espinoza et al. 2014) and digital PCR (Tadmor, Ottesen et al. 2011) should aid in this endeavor for certain viral-host groups. Additionally, single-cell genomics offers the opportunity to identify actively infected cells. Although non-specific phage binding and subsequent amplification remains a possibility (Labonte, Swan et al. 2015), the RTPR RNRs identified in SAGs suggest the taxonomic scope of bacterial diversity infected by RTPR podoviruses is underappreciated.

186 5.5.2 Assemblage structure and dynamics

Oyster EF viral assemblages displayed seasonal variability over one year. The average viral assemblage over one year was well represented by persistent phage populations observed in ≥7 months (Fig. 5.7D), but this was largely due to a strong spike in the relative abundance of these populations in one or two months (Fig. 5.14A). Seasonal phage populations observed in 4-6 sample months were also abundant and comprised on average 43% of the monthly assemblage as compared to 39% for populations observed in 7-10 months (Figs. 5.7D, 5.14A). A 4-6 month seasonality was also observed in the viral assemblage composition (Fig. 5.9B). This trend was similar to that observed for bacterial communities in oyster EF (Sakowski, In prep.), as well as in the San Pedro ocean time series over a ten year period (Chow, Sachdeva et al. 2013). In the latter case, bacterial communities displayed seasonal shifts in community composition, slowly drifting apart over time until a stable community average was observed after approximately four years (Chow, Sachdeva et al. 2013). To examine whether the periodic trend in viral assemblage composition persisted beyond one year, oyster EF samples from the Choptank River, Indian River, and Delaware Bay collected in June 2013 were included in the analysis (Fig. 5.9C). These samples were included because of the similarity they displayed by Principle

Coordinates Analysis (Fig. 5.10) and UPGMA hierarchical clustering (Fig. 5.9A) to Rhode River oyster EF samples collected between July and Dec. 2012. Thus, even though they were from different sites, similar diversity appeared to be selected for in these samples. Inclusion of these samples extended the observed phage assemblage dynamics to 18 months and continued the observed pattern of assemblage composition dynamics (Fig. 5.9C). Similar to the bacterial communities in the San Pedro ocean

187 time series, the assemblage composition slowly drifted from the initial sample over time (Fig. 5.9C). It is therefore likely that a long-term stable “average assemblage” would be observed in oyster EF if sampled for several years. Seasonal trends in oyster-associated bacterial diversity have been previously reported (La Valley, Jones et al. 2009), and bacterial abundance in oyster EF was shown to most strongly correlate with temperature after introducing a one-month lag (Sakowski et al., In prep.). Introducing a similar lag to account for the response of bacterial hosts and their respective phage to changing environmental conditions increased the strength of correlation for all four measured environmental parameters (temperature, salinity, pH, and dissolved oxygen) with the relative abundance of individual phage populations (Fig. 5.12). After introducing a one-month lag, the relative abundance of phage populations was also most strongly correlated with temperature (Fig. 5.12). UPGMA hierarchical clustering of phage populations according to their correlation profile with temperature, salinity, pH, and dissolved oxygen (DO) resulted in three distinct groups defined by their correlation with temperature and salinity (Fig. 5.11). Group I, which contained populations that were most strongly associated (positively) with temperature, also contained all five populations observed in each of the ten months sampled and 50% of populations present in ≥7 sample months (Fig. 5.11). Interestingly, the relative abundance of persistent phage populations was marginally negatively correlated (p = 0.07) with Chao1 richness (Fig. 5.8, 5.14B), and overall diversity of the viral assemblage was greatest in colder months (Fig. 5.8). This corresponds with previous reports of higher viral lysis rates in the Chesapeake Bay during the winter (Winget, Helton et al. 2011) and suggests that the viral production of these persistent phage populations remained

188 stable during a period where the production of more seasonal populations decreased, resulting in increased relative abundance of persistent populations. However, the possibility that viral production of these persistent populations increased simultaneously with the decline in total diversity during the warm summer months cannot be ruled out. There was no clear phylogenetic relationship between phage populations based on their UPGMA hierarchical clustering. Several well-supported clades contained populations from only a single environmental parameter correlation group, and all three groups were represented in these exclusive clades (Fig. 5.13). However, other clades contained two or all three groups (Fig. 5.13), indicating that closely related phage populations inhabited different ecological niches and comprised different ecotypes. This has been observed in bacterial species with identical 16S rRNA genes (Hahn and Pockl 2005), and it is possible that these phage populations infect very similar hosts with different ecotypes.

5.5.3 Geographic distribution of phage populations

RTPR RNR sequences were identified in virome libraries spanning the globe (Fig. 5.1A); yet, populations observed in oyster EF were largely endemic. Oyster EF amplicon sequences clustered more frequently with local virome sequences (RR/CB) than global virome sequences even after accounting for greater representation among local viromes (Fig. 5.5A, B). Interestingly, the gap between clusters with local vs. global representatives was positively correlated with persistence (p < 0.01; Fig. 5.5C). Thus, the most persistent phage populations were far more likely to be observed only in local viromes, while the more transient phage populations were observed in global and local viromes at a more similar rate (Fig. 5.5B, C). Transient phage populations

189 observed in oyster EF libraries were also always found in low abundance (Fig. 5.7D). Therefore, transient populations were likely allochthonous populations poorly suited to the estuarine environment, while the more persistent and abundant populations belonged to autochthonous phage. A recent study of 43 dsDNA viromes from the Tara Oceans expedition tracked the movement of viral populations and identified passive transport of phage populations facilitated by ocean currents (Brum, Ignacio-Espinoza et al. 2015). The authors concluded that viral communities were shaped by local biological and environmental conditions (Brum, Ignacio-Espinoza et al. 2015). This interpretation agrees with the correlation of phage population relative abundance with specific environmental parameters observed in this study (Fig. 5.11). Furthermore, the high proportion of diversity in low-abundance, allochthonous populations suggests the local environment shapes the viral assemblage against a background of viral diversity, an observation that agrees with the Bank model (Breitbart and Rohwer 2005). Sample location and passive transport appeared to play a role in viral diversity in oysters as well. The number of populations shared between the Rhode River and the three other sites decreased with distance (Fig. 5.15). The same was observed for the Choptank River with the Indian River and Delaware Bay samples (Fig. 5.15). In addition, these three samples displayed environmental selection for a distinct phage assemblage against a background of diversity as predicted. These samples were more similar to each other than to annual survey oyster samples from the Rhode River (Fig. 5.9A, 5.10), and the majority of the phage assemblage was comprised of populations identified in ≥2 sites for these samples (Fig. 5.7A). However, the distribution of these shared populations differed between sites. Only a single population was one of the ten most abundant populations at all three sites, while 17/23 populations were part of the

190 top ten at a single site (Fig. 5.7B). This suggests that autochthonous populations at one site may have been allochthonous elsewhere, indicating phage ecology can differ even across short geographic differences and between samples with similar diversity. This may be particularly true for the populations identified in this study given that RTPR phages are T7-like podoviruses with (presumably) very narrow host ranges.

5.5.4 Comparison of viral diversity from PCR and viromes

Viruses lack any conserved marker gene for assessing diversity akin to the 16S rRNA gene for bacteria. As such, viral ecologists have relied upon PCR amplification of marker genes, such as major capsid protein (Filee, Tetart et al. 2005) and DNA Pol A (Labonte, Reid et al. 2009), or shotgun viral metagenomics to assess viral diversity. In this study, the diversity and distribution of phage populations in oyster EF were examined by PCR amplification of monomeric, class II (oxygen-independent, cobalamin-dependent) RNR genes belonging to the RTPR group (Sakowski, Munsell et al. 2014). One of the limitations to marker gene studies using PCR-based approaches is the exclusion of phage diversity due to primer bias. For example, a recent study of the DNA Polymerase A gene in virome libraries found the greatest diversity belonged to a novel clade that not been previously observed by PCR (Schmidt, Sakowski et al. 2014). In an attempt to limit similar bias to that observed in

DNA Pol A, RTPR primers were designed from virome sequence data with the goal of limiting primer degeneracy while simultaneously amplifying the broadest range of diversity and the most ecologically relevant phage populations. To that end, amplicon clusters were part of only a single well-supported clade, but this clade also comprised 74% of clusters from local (CB/RR) viromes and 69% of clusters from all other virome libraries (Global) (Fig. 5.2).

191 Due to their inherent limitations, PCR-based marker gene studies have been criticized in favor of shotgun viral metagenomics for assessing spatiotemporal diversity of abundant viruses (Sullivan 2015). Assuming non-chimeric assemblies of complex communities, shotgun metagenomics would ideally provide sufficient coverage of all viral populations present to make biological and ecological inferences about those populations. However, the majority of viral diversity in oyster EF lies in the less abundant populations (Fig. 5.7C-D), an observation predicted by the Bank model (Breitbart and Rohwer 2005), and assessing the spatiotemporal diversity of these populations by viral metagenomics remains problematic. Consider, for example, that the 58 virome libraries (excluding the Rhode River) combined contributed 418 RTPR RNRs that spanned the region of interest and contained key catalytic residues, and these RNRs formed 196 peptide clusters. By comparison, 14 amplicon libraries resulted in 8,221 RNRs and 382 clusters. Only the deeply sequenced Rhode River sample (155 million Illumina 2 x 150 reads) provided similar levels of diversity: 421 RNRs and 307 clusters. Thus, this PCR-based approach not only shed light on previously unidentified viral diversity, but also enabled detection of even rare and/or transient phage populations at a level not possible without very deep viral shotgun metagenomic sequencing. Furthermore, not a single contig with an RTPR RNR from the 58 viromes examined in this study was >10kb, limiting the biological context that could be gleaned from these virome sequences. The much deeper sequencing of the Rhode River virome provided 91 contigs >10kb with RTPR sequences used in this analysis. The Rhode River virome demonstrates the potential for viral metagenomes to provide meaningful biological and ecological information of targeted groups,

192 particularly rare populations. In the future, DNA sequencing technologies may provide the necessary depth of sequencing required at a more reasonable cost, and there is no denying the advantages of viral metagenomics in viral ecology, including the ability to make genome-based hypotheses about phage biology and ecology (Sullivan 2015). For the present, however, we suggest that virome-informed PCR of marker genes continues to provide valuable insight into phage diversity and can be particularly powerful when combined with virome sequence data.

193 FIGURES

A. B.

Annual Biogeographic Survey Survey Oysters Oysters 26 Clusters AS (9%) BS (15%) 160 Clusters 63 Clusters (54%) (37%) 58 Clusters AS (20%) BS (34%) W (13%) 50 Clusters 23 Clusters AS (17%) BS (14%) W (11%) W (5%)

320 Clusters (71%)

Water Virome & Amplicon Figure 5.1: Distribution of RTPR RNR sequences from aquatic virome and oyster extrapallial fluid amplicon libraries. A) Global distribution of virome and amplicon libraries with RTPR RNR peptide sequences used in this study. B) Distribution of RTPR RNR peptide clusters between oyster EF and water. RTPR RNR sequences were clustered at 95% amino acid identity. Percentages refer to the proportion of clusters within each sample group (e.g. Annual survey oyster samples). AS (Annual Survey Oysters, n = 294 clusters); BS (Biogeographic Survey Oysters, n = 170 clusters); W (Water virome and amplicon sequences, n = 451 clusters).

194 Thurmus phage P2345 Thermus phage P7426 Alkaliphilus metalliredigens QYMF Cluster 49 (RR/CB: 1.9%) Celeribacter phage P12053L Roseobacter phage SIO1 Cluster 51 (RR/CB: 0.9%) Cluster 50 (RR/CB: 0.2%) Candidatus Protochlamydia amoebophila Cluster 48 (RR/CB: 1.5%) Cluster 59 (G: 0.8%) Cluster 57 (G: 0.3%) Cluster 58 (G: 1.1%) Nonomuraea candida Mycobacterium phage Gladiator Rhodococcus phage RGL3 Weissella confusa LBAE C39-2 Cluster 56 (RR/CB: 0.4%) Cluster 54 (RR/CB: 6.2%; G: 1.6%) Cluster 55 (RR/CB: 0.6%) Nitratiruptor sp. SB155-2 Nitratifractor salsuginis Mycobacterium phage Gaia archaeon Loki Cluster 52 (RR/CB: 0.6%; G: 3.5%) Ectocarpus siliculosus Nannochloropsis gaditana Cluster 53 (RR/CB: 0.4%; G: 1.1%) Dictyostelium fasciculatum Naegleria gruberi strain NEG-M Rhodothermus phage RM378 Cluster 45 (RR/CB: 5.1%; G: 18%) Cluster 47 (G: 0.3%) Cluster 44 (Amp: 0.01%; RR/CB: 3.4%; G: 1.4%) Cluster 46 (RR/CB: 0.2%) Monosiga brevicollis MX1 Salpingoeca rosetta Diplosphaera colitermitum Cluster 39 (RR/CB: 0.4%) Cluster 38 (RR/CB: 1.7%; G: 1.9%) Cluster 40 (RR/CB: 0.4%; G: 0.3%) Cluster 41 (RR/CB: 1.7%) Cluster Key Cluster 43 (RR/CB: 0.4%) Cluster 42 (RR/CB: 0.2%) Amplicon Only Acanthocystis turfacea Chlorella virus Paramecium bursaria Chlorella virus Cluster 16 (Amp: 0.02%; RR/CB: 0.6%) Cluster 34 (G: 0.3%) Amplicon & RR/CB Viromes Only Cluster 23 (Amp: 21%; RR/CB: 17%; G: 14%) Cluster 20 (Amp: 6.4%; RR/CB: 7.0%; G: 3.5%) Cluster 26 (RR/CB: 0.2%) Amplicon,RR/CB, and Global Viromes Cluster 12 (Amp: 0.3%; RR/CB: 0.2%; G: 0.3%) Cluster 7 (Amp: 0.5%; RR/CB: 0.4%) Cluster 14 (Amp: 0.07%; RR/CB: 0.6%) RR/CB Viromes Only Cluster 24 (Amp: 4.6%) Cluster 18 (Amp: 0.02%) Cluster 17 (Amp: 6.5%; RR/CB: 1.3%) Global Viromes Only Proteobacteria bacterium JGI 0000113-P07_3 Cluster 13 (Amp: 1.9%; RR/CB: 4.1%; G: 5.4%) Chlamydomonas reinhardtii RR/CB & Global Viromes Only Volvox carteri f. nagariensis Cluster 28 (RR/CB: 0.2%) Cluster 9 (Amp: 0.2%; RR/CB: 1.1%) Puniciespirillum phage HMO-2011 Cluster 21 (Amp: 2.2%; RR/CB: 2.6%; G: 1.1%) Cluster 10 (Amp: 0.3%; RR/CB: 1.3%; G: 1.9%) Proteobacteria bacterium JGI 0000113-P07_2 Cluster 22 (Amp: 0.2%; RR/CB: 1.7%; G: 1.9%) Cluster 2 (Amp: 2.0%; RR/CB: 6.0%; G: 2.2%) alpha proteobacterium SCGC AAA536-G10 Verrucomicrobia bacterium SCGC 164-O14 Verrucomicrobia bacterium SCGC AAA300-N18 Streptomyces phage Jay2Jay Cluster 29 (Amp: 0.07%) Solirubrobacter sp. URHD0082 Cluster 11 (Amp: 0.02%; RR/CB: 0.4%; G: 15%) Cluster 15 (RR/CB: 0.4%; G: 2.4%) Cluster 25 (RR/CB: 0.8%) Cluster 19 (Amp: 3.6%; RR/CB: 1.9%; G: 0.5%) Cluster 30 (Amp: 0.06%; RR/CB: 0.2%) Cluster 37 (G: 1.1%) Cluster 36 (RR/CB: 0.2%) Roseiflexus castenholzii Cluster 33 (RR/CB: 0.2%) Methylocaldum szegediense Chlorobium sp. GBChlB Nitrosococcus halophilus Nc 4 Cluster 4 (Amp: 0.01%; RR/CB: 1.3%; G: 0.8%) Cluster 31 (RR/CB: 0.2%) Cluster 27 (Amp: 0.1%) gamma proteobacterium SCGC AAA076-P13 Cluster 1 (Amp: 4.9%; RR/CB: 8.5%; G: 12%) Cluster 6 (Amp: 0.07%; RR/CB: 0.6%) Cluster 35 (Amp: 0.04%; RR/CB: 0.4%) Cluster 8 (Amp: 3.3%; RR/CB: 1.3%) Cluster 32 (RR/CB: 0.3%) Cluster 3 (Amp: 1.6%; RR/CB: 2.1%; G: 2.2%) Cluster 0 (Amp: 40%; RR/CB: 11%; G: 4%) Proteobacteria bacterium JGI 0000113-P07_1 Cluster 5 (G: 0.3%) 0.9

Figure 5.2: Maximum likelihood tree with 100 bootstrap replicates of viral and reference RTPR RNR peptide sequences. Amplicon and virome sequences were aligned in MAFFT and trimmed to the region of interest (N437 to S625 in E. coli). Sequences lacking key catalytic residues C439, E441, and C462 were removed. All amplicon and virome sequences were clustered at 80% identity. The proportion of amplicons/virome contigs contained in each cluster is noted in parentheses: amplicons (Amp); Rhode River or Chesapeake Bay viromes (RR/CB); all other viromes (i.e. Global, G). Circles represent bootstrap support: Black (100%); Gray (≥75%); White (≥50%). Scale bar represents amino acid substitutions per site.

195 A.

Verrucomicrobia bacterium SCGC AAA300-N18

Solirubrobacter sp. URHD0082

Paramecium bursaria chlorella virus

Volvox carteri f. nagariensis

Verrucomicrobia bacterium SCGC AAA164-O14

Streptomyces phage Jay2Jay

alpha proteobacterium SCGC AAA536-G10 Top BLAST Hits BLAST Top Acanthocystis turfacea chlorella virus

Chlamydomonas reinhardtii

gamma proteobacterium SCGC AAA076-P13

Puniceispirillum phage HMO-2011

Proteobacteria bacterium JGI 0000113-P07

0 20 40 60 80 100 120 # Observations B. Proteobacterium bacteria JGI 0000113-P07 55.55_C

Puniceispirillum phage HMO-2011

Proteobacterium bacteria JGI 0000113-P07 29.29_C

5kb Figure 5.3: Reference sequences with closest homology to RTPR RNR amplicons from oyster EF. A) The top BLAST hits for the 50 most abundant RTPR RNR peptide clusters from the annual survey and biogeographic survey. The representative sequence for each cluster was queried against the nr database (BLASTp, excluding uncultured/environmental sequences). The top ten BLAST hits by bit score were recorded. Only references observed for at least ten different clusters are shown. B) Alignment of phage-like contigs from single amplified genome Proteobacteria bacterium JGI 0000113-P07 with Puniceispirillum phage HMO-2011. ORFs are outlined as alternating gray and white boxes. Homologous ORFs are connected by blue rectangles. Homologous ORFs were identified by BLASTx.

196 Streptomyces phage Jay2Jay Acanthocystis turfacea Chlorella virus Paramecium bursaria Chlorella virus Cyanophage PSS2 ctg7180007386136 (0.3%) ctg7180007385665 ctg7180007354327 ctg7180007386049 (0.01%) ctg7180007380058 (0.02%) ctg7180007327934 Cyanophage S-TIM5 Pelagibacter phage HTVC008M Myoviridae Prochlorococcus phage P-SSM2 Prochlorococcus phage P-RSM4 Cyanophage S-SSM2 Prochlorococcus phage P-HM1 Cyanophage MED4-213 Prochlorococcus phage P-HM2 ctg7180007353986 ctg7180007386577 ctg7180007379964 (4.9%) ctg7180007386510 Pelagibacter phage HTVC010P Pelagibacter phage HTVC019P Prochlorococcus phage P-RSP2 ctg7180007358425 Prochlorococcus phage P-SSP7 Synechococcus phage P60 Cyanophage Syn5 ctg7180007361661 (2.0%) ctg7180007375468 (6.5%) ctg7180007368337 (4.9%) ctg7180007386494 (3.6%) ctg7180007325902 (0.2%) Podoviridae ctg7180007367949 (2.0%) ctg7180007356326 (0.3%) ctg7180007385966 (21%) ctg7180007386451 (21%) Roseobacter phage SIO1 Celeribacter phage P12053L Proteobacteria bacterium JGI 0000113-P07 59.59 deg7180006856616 ctg7180007385887 (21%) Cluster Key (80% AA ID) ctg7180007362538 (21%) Amplicon & RR/CB Viromes Only ctg7180007386066 (21%) Puniceispirillum phage HMO-2011 Amplicon,RR/CB, and Global Viromes Proteobacteria bacterium JGI 0000113-P07 55.55 RR/CB Viromes Only Proteobacteria bacterium JGI 0000113-P07 29.29 ctg7180007380072 (0.2%) RR/CB & Global Viromes Only ctg7180007381253 (2.0%) Figure 5.4: UPGMA hierarchical clustering of Rhode River virome contigs >10kb with RTPR RNR sequences. Translated virome contig and reference ORFs were clustered by reciprocal BLASTp (1 e -30). Thirty clusters were randomly sampled per genome/contig 1,000 times. Distances were calculated by Bray-Curtis similarity. Numbers in parentheses indicate the proportion of amplicon sequences in the RNR peptide cluster (80% AA ID) to which the virome contig belonged. Note that some 80% RNR clusters were represented more than once, while others were not represented at all.

197 A. B.

45 80 80 me o

r 40 i 70 70 V h

t 35 60 60

d wi 30

are

h 50 50

ces 25 S n

e 20 ers qu

t 40 40 e s

S RR/CB RR/CBRR/CB lu 15 Sequences Sequences C Global 30 30 GlobalGlobal

on 10 c

pli 5 20 20 Am

% 0 10 10

Virome Clusters Sharedwith Amplicon % Virome Clusters Sharedwith Amplicon % il y y r r y 11 ry ry h r a 12 a rc p Jul be k R. B 20 rua a A M m tobe M e c an E 0 0 c. anua eb t c. 20 t D J F O Indian R. De ep De 10 109 89 78 67 56 45 34 23 12 31 23 12 1 S Chop Sample # Amplicon# Amplicon Libraries Libraries in Clusters in Clusters AS BS C. # Amplicon Libraries in Clusters 70

R! = 0.66 20 CB and Global Viromes (%) Viromes Global CB and

10 DifferenceRR/Between Clusters Sharedwith 0 10 9 8 7 6 5 4 3 2 1 # Amplicon Libraries in Clusters Figure 5.5: Oyster EF RTPR RNR amplicon clusters shared with local and global virome sequences. 370 virome sequences from the Rhode River and Chesapeake Bay (RR/CB) were randomly sub-sampled with replacement and clustered with all global virome sequences (n = 370) and amplicon sequences (n = 8,221) at 95% amino acid identity. A) Oyster EF amplicon clusters shared with virome sequences by amplicon library. B) Oyster EF amplicon clusters shared with virome sequences by persistence (# amplicon library months represented in cluster) for the annual survey (AS, Dec. 2011 - Dec. 2012) and biogeographic survey (BS, June 2013). C) The difference between local (RR/CB) and global clusters shared with oyster EF amplicon sequences from the annual survey (% amplicon clusters shared with RR/CB sequences - % amplicon clusters shared with global sequences). Cluster comparisons were grouped by the persistence of amplicon clusters. Error bars represent standard error from 10 jackknife replicates.

198 A. B. CR Choptank Indian River IR River 33 32 7 25 Clusters CR (25%) 33 8 66 Clusters IR (27%) 47 Clusters DB (66%) (50%) RR 18 2 9 Clusters 4 210 7 CR (9%) 15 0 IR (10%) 5 0 DB (28%) 13 Clusters 0 Clusters IR (14%) DB (41%) 6

10 Clusters (31%)

DE Bay

Figure 5.6: Distribution of RTPR RNR peptide clusters from oyster EF samples by sample site. RTPR RNR sequences were clustered at 95% amino acid identity. A) Distribution of RTPR RNR peptide clusters between oyster extrapallial fluid samples collected during the biogeographic survey from June 24-26, 2013. B) The number of unique and shared RTPR RNR peptide clusters between oyster extrapallial fluid samples by site. RR (Rhode River, i.e. Annual Survey oysters); CR (Choptank River oysters); IR (Indian River oysters); DB (Delaware Bay oysters).

199 A. B. 100% 12

90% s on 80% c 10 pli

70%

60% rvey 8 u S

# of Sites c 50% 1

phi DE Bay Assemblage ra f 2 6 Indian River Survey 40% og o e % 3 Choptank River

30% iog

B Top 10 Cluster

f Biogeographic

20% o 4 % d e

10% z li 2 0% rma o

Choptank R. Indian R. DE Bay N Sample Site 0 1 10 20 30 40 C. D. Cluster Rank 60 35

30 50

20 30 Assemblage

f 15 Annual Survey o

20 % 10 % of Clusters (95% AA ID) AA Clusters %of (95% 10 5

0 0 10 9 8 7 6 5 4 3 2 1 10 9 8 7 6 5 4 3 2 1 # Months Observed # Months Observed Figure 5.7: Abundance and distribution of RTPR phage populations (95% AA ID) in oyster EF. A and B) Composition of the viral assemblage in oyster EF samples from the Choptank River, Indian River, and DE Bay. C and D) Composition and persistence of the viral assemblage in oyster EF samples from the Rhode River collected at ten months between Dec. 2011 and Dec. 2012. A) Composition of the viral assemblage in biogeographic survey oyster samples according to the distribution of clusters across three locations. B) Rank-abundance of the top 50 RTPR clusters from biogeographic oyster EF samples. Proportions were normalized by library size. The ten most abundant peptide clusters from each sample site are noted. Choptank River (red); Indian River (blue); DE Bay (green). C) Distribution of RTPR RNR peptide clusters according to persistence (# months observed) in oyster EF samples from the annual survey. D) Relative abundance of RTPR RNR peptide clusters according to persistence (# months observed) in oyster EF samples from the annual survey. Error bars are SE.

200 A. B.

200 120 ) D

I 180 100 95%

( 160

ers 140 t 80 s

lu 120 C

e 100 60 id t p

e 80 Chao1 Index Chao1 P 40 d 60 e t 40 ma i 20 t s 20 E

# 0 0

1 h il y y r r . y 1 h il y y r r . y 1 ry ry r a R a 1 ry ry r a R a rc p Jul be k B rc p Jul be k B 20 rua a A M m tobe 2012 20 rua a A M m tobe 2012 M e c an E M e c an E c. Janua eb t O c. t D c. Janua eb t O c. t D F ep Indian R. F ep Indian R. De S De De S De Chop Chop Sample Sample C. 6 D. 6

y t

i 5 5 vers i

4 4 ex c i t e n

e 3 3 nnon Ind log a y h h S P 2 2

s h' t i

a 1 1 F

0 0 . . 1 h ril y y r r y 1 h ril y y r r y 1 ry ry a ul R a 1 ry ry a ul R a rc p J be k B rc p J be k B 20 rua a A M m tobe 2012 20 rua a A M m tobe 2012 anua M e c an E anua M e c an E c. J eb t O c. t D c. J eb t O c. t D F ep Indian R. F ep Indian R. De S De De S De Chop Chop Sample Sample Figure 5.8: Alpha diversity of oyster EF RTPR RNR amplicon libraries. RTPR RNR sequences were clustered at 95% amino acid identity. A) Estimated number of peptide clusters (95% ID) for each oyster EF RTPR RNR amplicon library. The number of observed peptide clusters was extrapolated for each library to the sampling depth of the largest library (Choptank River, 1,588 sequences) in EstimateS v9.1.0 (Colwell 2013). Error bars are 95% confidence intervals. B-D) Chao1, Faith’s Phylogenetic Diversity, and Shannon index of diversity for oyster EF amplicon libraries. Libraries were sub-sampled to the depth of the smallest library (DE Bay; 112 sequences) with 100 iterations with replacement. Error bars (SE). The mean value for all oyster amplicon libraries is shown (dotted line).

201 A. B. !#'"

!#&$" May !#&" January

!#%$" February

!#%" March !"#$%&"'()*#+,-.(/#0&-*."(

April !#!$"

Dec. 2012 !" !" &" (" )" *" %!" %&" Choptank R. 12*&%0(3"&4""*(5-678"0(

DE Bay C. !#'" Indian R. !#&$" Dec. 2011 !#&" July !#%$" September !#%"

October !"#$%&"'()*#+,-.(/#0&-*."( 0.02 !#!$"

!" !" &" (" )" *" %!" %&" %(" %)" %*" 12*&%0(3"&4""*(5-678"0( Figure 5.9: Seasonal trends in RTPR RNR viral assemblages from oyster EF. Libraries were compared by weighted UniFrac distance. A) UPGMA hierarchical clustering of oyster EF RTPR RNR libraries from the Rhode River (annual survey), Choptank River, Indian River, and DE Bay. Libraries were sub-sampled with replacement to the depth of the smallest library (DE Bay; 112 sequences). Circles represent jackknife support: Black (100%); Gray (≥75%); White (≥50%). Scale bar represents UniFrac distance. B and C) Similarity of oyster EF RTPR assemblage over time. B) Annual survey oyster EF samples collected between Dec. 2011 and Dec. 2012 from the Rhode River. C) Annual survey oyster EF samples collected between Dec. 2011 and Dec. 2012 and biogeographic survey samples collected in June 2013. Error bars are SD.

202 18% July

September

Dec. 2011 February October January April Dec. 2012 March DE Bay Choptank R. Indian R. May 26% 14%

Figure 5.10: Weighted UniFrac principle coordinates analysis of oyster EF RTPR RNR libraries from the Rhode River (annual survey), Choptank River, Indian River, and DE Bay. Rhode River samples were collected between Dec. 2011 and Dec. 2012. Samples from the Choptank River, Indian River, and DE Bay were sampled in June 2013. Libraries were sub-sampled with replacement to the depth of the smallest library (DE Bay; 112 sequences). Confidence halos from 100 jackknife replicates.

203 − 0.5 0.5 Spearman's Rank R

Cluster Persistence 40 9 6 7 24 5 305 10 18 5 1 10 11 10 42 5 Group I 141 8 49 8 48 10 3 10 497 5 7 5 394 7 392 5 456 6 383 7 55 9 23 8 Group II 306 7 94 5 2 6 398 6 64 7 33 7 310 8 106 6 469 6 256 5 373 5 251 5 Group III 324 5 468 5 28 5 139 8 Salinity Temperature pH DO Parameters Figure 5.11: Strength of correlation (Spearman’s Rank R) between RTPR RNR peptide cluster (95% ID) relative abundance and salinity, temperature, pH, and dissolved oxygen (DO) over one year with a one month lag. Only peptide clusters present in at least five oyster EF samples collected between Dec. 2011 and Dec. 2012 from the Rhode River (n = 36) were included. Clusters were grouped by UPGMA hierarchical clustering according to their environmental parameter correlation profile.

204 0.6 * 0.5 *

0.4

0.3 No Lag Lag 0.2 Spearman'sRank |R|

0.1

0 Temperature pH Salinity DO

Parameter Figure 5.12: Strength of correlation (Spearman’s Rank R) between RTPR RNR peptide cluster (95% ID) relative abundance and temperature, pH, salinity, and dissolved oxygen (DO) over one year. Only peptide clusters present in at least 5 oyster EF libraries collected between Dec. 2011 and Dec. 2012 (n = 36) were included. Relative abundance was correlated without any lag and with a one-month lag behind measured water parameters. Relative abundance was significantly more correlated with temperature and DO after introducing a one-month lag (T Test, p < 0.01). Error bars are SE.

205 Cluster 256 Cluster 251 Cluster 324 Physical Parameter Cluster 305 Cluster 310 Correlation Group Cluster 306 Cluster 468 Cluster 469 Group I Cluster 456 Cluster 497 Cluster 394 Group II Cluster 392 Cluster 398 Group III

Cluster 373 Cluster 383

Cluster 106 Cluster 64

Cluster 94

Cluster 48 Cluster 49 Cluster 55 Cluster 33

Cluster 7 Cluster 11 Cluster 1 Cluster 2 Cluster 3 Cluster 18 Cluster 6 Cluster 40 Cluster 42 Cluster 139 Cluster 141 Cluster 23 Cluster 24 Cluster 28 0.9 Figure 5.13: Maximum likelihood tree with 100 bootstrap replicates of oyster EF RTPR RNR amplicon clusters and references. Amplicon and virome sequences were aligned in MAFFT and trimmed to the region of interest (N437 to S625 in E. coli). Sequences lacking key catalytic residues C439, E441, and C462 were removed. All amplicon and virome sequences were clustered at 95% identity. Only peptide clusters present in at least five oyster EF libraries collected between Dec. 2011 and Dec. 2012 from the Rhode River were included (n = 36). Clusters were colored according to the environmental parameter correlation group to which they belonged. Circles represent bootstrap support: Black (100%); Gray (≥75%); White (≥50%). Scale bar represents amino acid substitutions per site.

206 A. 100% B. 160 90% 140 80% 70% 120

60% 100 50% R! = 0.28231 80 40% 1-3 months Chao1 Index % of Assemblage %of 30% 4-6 months 60 7-10 months 20% 40 10% 20 0% 0 April May July March 0 10 20 30 40 50 60 70 80 90 100 JanuaryFebruary October Dec. 2011 September Dec. 2012 Persistent Populations (7-10 months) Mean Combined Relative Sample Month Abundance Figure 5.14: Proportion of persistent oyster EF phage populations sampled from Dec. 2011 to Dec. 2012 in the Rhode River. A) Relative abundance of phage populations by persistence for each amplicon library. B) Correlation between the relative abundance of persistent phage populations and Chao1 estimate of diversity for the viral assemblage.

207 18 RR DB 45 9 17 58 CR IR 34

Figure 5.15: The number of RTPR RNR peptide clusters shared between oyster sample sites in the Delmarva region.

208 TABLES

Table 5.1: Proteobacteria bacterium JGI 0000113-P07 contig 59.59 BLAST results.

ORF Top Hit Top Viral Hit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

209 Chapter 6

CONCLUSIONS AND FUTURE DIRECTIONS

The eastern oyster, Crassostrea virginica, remains a key component of estuarine environments along the east coast of North America but faces numerous abiotic and biotic threats. In the work contained herein, the microbiome of C. virginica extrapallial fluid (EF) was characterized using high-throughput, cultivation- independent approaches. As a primarily explorative effort, this work raised as many questions as it answered, and much about the EF microbiome and its role in oyster health and physiology remains to be discovered. For example, EF bacterial communities were influenced by environmental factors like temperature and sample site. However, it remains unclear whether other factors, such as individual genetics and the innate immune system play a role in microbiome composition (Fig. 6.1).

Bacteriophage

SURVIVAL?

Bacteria

Geographic Inhibitory Compounds/ AMPs Host Genetics Location Immune Response Phage Lysis

Nutrient availability Competition with other Environmental Conditions bacteria Figure 6.1: Known and potential factors influencing the composition of the C. virginica extrapallial fluid microbiome.

210 It is possible that EF bacterial communities were more similar within a site than between sites because the oysters within a given site were more closely related to each other than to oysters from other locations. This could be examined in future experiments using a three-way experimental design. In addition to characterizing EF bacterial communities from several locations, individual oysters could be genotyped. The genetic relatedness of oysters could then be compared to the similarity of EF bacterial communities within and between sites, and the influences of genetics and sample site on EF community composition could be discerned. It is also clear that the oyster EF microenvironment selects for a unique community since the EF microbiome was distinct from the surrounding water independent of the sample location or time of year samples were collected. Interestingly, differences between EF communities from different sample sites were largely confined to lower taxonomic levels (OTUs), while the proportions of higher taxonomic levels were similar. This could be the result of certain bacterial lineages being broadly adapted to the EF microenvironment. However, if this were the case then it would be expected that EF communities would be dominated by copiotrophic organisms like some Vibrio and Pseudomonas species, particularly since these organisms are so readily cultivated from oysters. However, these organisms made up a relatively small proportion of the community. Therefore, it remains likely that other factors may influence EF community composition, including the oyster’s innate immune system. Oyster hemocytes have been experimentally shown to selectively remove bacteria from the environment, including Escherichia coli. Studies in the future could expand upon this work by examining the scope of native bacterial diversity phagocytosed by C. virginica hemocytes in EF. Interestingly, few differences

211 were observed in EF community compositions between healthy oysters and those infected by Perkinsus marinus despite the fact that P. marinus compromises the oyster immune response. The production of antimicrobial peptides (AMPs) may explain the stability of the EF microbiome even in immunocompromised individuals, and the potential impacts of these AMPs on putative opportunistic pathogens and normal EF taxa could be explored. Sequencing the genome of C. virginica would augment this work (currently only the genome of C. gigas has been sequenced). Despite a lack of large-scale changes in EF communities between healthy and P. marinus-infected individuals, several differences between the two were observed and warrant further investigation. One of the more intriguing differences was the observed shift in Spirochaete taxa between healthy and diseased oysters. It would be interesting to see whether this change impacted the efficiency with which C. virginica extracted energy from its diet and if growth rate would be subsequently impacted. The association of mycoplasma (Mollicutes) with P. marinus infection is also worthy of further consideration. It is currently unclear whether mycoplasma were opportunists or more intimately associated with P. marinus, and their potential to exacerbate P. marinus infection through secondary infection or increase individual oyster susceptibility to P. marinus could be examined. In contrast, the bacterial taxa associated with healthy oysters include relatives of putative probiotic bacteria and could be investigated for any potential probiotic benefits they may provide. One other factor influencing bacterial communities is the top-down pressure of viral predation. Lytic viruses infect and lyse the most competitive (i.e. the fastest growing) bacterial populations according to the Kill-the-Winner model. It is possible that copiotrophic bacterial populations comprised little of the EF bacterial community

212 because they were under greater pressure of viral predation, and future studies could examine the rate at which viral production occurs in oyster EF. However, outside of Vibrio phages the nature of phage diversity and dynamics in oysters remains largely unknown, and this work was the first reported to characterize a wide breadth of phage diversity in oyster EF across time and space using cultivation-independent, high- throughput methods. The identification of ribonucleotide reductase (RNR) as an abundant, diverse, and ecologically important marker of viral diversity greatly aided this approach. Developing primers for viral RNR genes from viral metagenome libraries enabled amplification of the most abundant and ecologically relevant phage populations that may have been missed by either cultivation-dependent methods or by developing primers based on reference sequences. Indeed, the most abundant phages with RNRs in both aquatic metagenomes and oyster EF were novel and only distantly related to a single reference phage (Puniceispirillum phage HMO-2011). From this work it was clear that the majority of phage diversity in oyster EF resided in low-abundance, transient populations that were likely not actively infecting bacterial hosts. In contrast, the most abundant populations were seasonally or persistently abundant and were endemic to the Delmarva region where oysters were sampled. The presence of persistent phage populations is particularly interesting since genomic context indicated these phages were likely members of the Podoviridae, a family of dsDNA phages that have narrow host ranges in cultivation-based studies. The persistence of these populations could be due to several factors: 1) persistent phages infect bacterial hosts that are abundant throughout the year; 2) these phages have a broader host range in the environment than cultivation-based approaches suggest; or 3) EF stabilizes viral particles against decay and long retention times

213 enabled phage detection even in the absence of their host. Key to solving this question will be linking viral populations to their hosts. This is a difficult task for environmental samples where cultivation is not a viable option, but emerging technologies should make future endeavors more successful. For example, digital PCR has been used to successfully link a viral gene of interest to a host, enabling the interaction to be detected. Similarly, active phage infections can be observed in single amplified bacterial genomes, which have become more feasible as sequencing technologies have increased and associated costs decreased. Ultimately, characterizing the EF microbiome promotes hypotheses surrounding the role of this community in oyster health and physiology. This work observed an enrichment of Deltaproteobacteria in EF, a particularly interesting observation since EF mediates Ca2+ precipitation and shell formation and Deltaproteobacteria are calcifying bacteria found in other environments where Ca2+ precipitation occurs like caves and lithifying microbial mats. It was proposed that the EF microbiome, particularly the Deltaproteobacteria, plays a role in oyster shell formation and may be responsible for remote calcification away from the shell mineralization front. This proposal deserves further investigation given the threats to shell formation posed by ocean warming and acidification and the decrease in the abundance of Deltaproteobacteria observed at higher temperatures. Among the first questions to address is whether the EF microbiome influences shell deposition and integrity, which could be tested using filtration and antibiotic-based approaches. Although presence alone is enough for Deltaproteobacteria to promote Ca2+ precipitation, it is currently unknown whether the Deltaproteobacteria are active members of the EF community. Activity could be indirectly measured for the

214 Deltaproteobacteria and other community members by comparing ratios of RNA to DNA. Activity of the Deltaproteobacteria could also be specifically investigated by quantitative PCR of dissimilatory sulfite reductase (Dsr). In this way, potentially important interactions between community members may be elucidated (Fig. 6.2).

SO 2- - NH + NO - 4 S 4 3

Sulfate reducing bacteria (Deltaproteobacteria) Denitrifying bacteria

2+ Ca CaCO3

Figure 6.2: Potential interaction between calcifying Deltaproteobacteria and denitrifying bacteria. Deltaproteobacteria reduce sulfate to sulfide, which in turn stimulates the reduction of nitrate by denitrifying bacteria. The reduction of nitrate results in a slight increase in pH, which could further promote calcium precipitation by the Deltaproteobacteria.

Finally, the hypothesis that Deltaproteobacteria are involved in remote calcification could be investigated by determining the localization of these bacteria. Remote

215 calcification would likely occur if bacteria inhabited the space in between shell layers and could be examined by microscopy. Additionally, Deltaproteobacteria may be associated with the shell mineralization front and their localization could be observed within the extrapallial cavity using fluorescence in situ hybridization.

216 REFERENCES

Acton, R. T. and E. E. Evans (1968). "Bacteriophage clearance in the oyster (Crassotrea virginica)." J Bacteriol 95(4): 1260-6.

Allam, B., W. E. Carden, et al. (2013). "Early host-pathogen interactions in marine bivalves: Evidence that the alveolate parasite Perkinsus marinus infects through the oyster mantle during rejection of pseudofeces." Journal of invertebrate pathology 113(1): 26-34.

Altschul, S. F., W. Gish, et al. (1990). "Basic local alignment search tool." Journal of molecular biology 215(3): 403-410.

Andrews, J. D. (1996). "History of Perkinsus marinus, a pathogen of oysters in Chesapeake Bay 1950-1984." Journal of Shellfish Research 15(1): 13-16.

Andrews-Pfannkoch, C., D. W. Fadrosh, et al. (2010). "Hydroxyapatite-mediated separation of double-stranded DNA, single-stranded DNA, and RNA genomes from natural viral assemblages." Appl Environ Microbiol 76(15): 5039-45.

Angly, F. E., B. Felts, et al. (2006). "The marine viromes of four oceanic regions." PLoS Biol 4(11): e368.

Antunes, F., M. Hinzmann, et al. (2010). "Association between environmental microbiota and indigenous bacteria found in hemolymph, extrapallial fluid and mucus of Anodonta cygnea (Linnaeus, 1758)." Microbial Ecology 60: 304-309.

Arner, E. S. and A. Holmgren (2000). "Physiological functions of thioredoxin and thioredoxin reductase." Eur J Biochem 267(20): 6102-9.

Audemard, C., L. Ragone Calvo, et al. (2006). "Real-time PCR investigation of parasite ecology: in situ determination of oyster parasite Perkinsus marinus transmission dynamics in lower Chesapeake Bay." Parasitology 132(06): 827- 842.

Bachere, E. (2003). "Anti-infectious immune effectors in marine invertebrates: potential tools for disease control in larviculture." Aquaculture 227: 427-438.

Backhed, F., R. E. Ley, et al. (2005). "Host-bacterial mutualism in the human intestine." science 307(5717): 1915-1920.

217 Bahr, L. M. and W. P. Lanier (1981). The ecology of intertidal oyster reefs of the South Atlantic coast: a community profile, US Fish and Wildlife Service.

Baird, D., R. R. Christian, et al. (2004). "Consequences of hypoxia on estuarine ecosystem function: energy diversion from consumers to microbes." Ecological Applications 14(3): 805-822.

Baker, S. and R. Mann (1994). "Description of metamorphic phases in the oyster Crassostrea virginica and effects of hypoxia on metamorphosis." Marine Ecology-Progress Series 104: 91-91.

Baldwin, B. (1995). "Selective particle ingestion by oyster larvae (Crassostrea virginica) feeding on natural seston and cultured algae." Marine Biology 123(1): 95-107.

Bankevich, A., S. Nurk, et al. (2012). "SPAdes: a new genome assembly algorithm and its applications to single-cell sequencing." J Comput Biol 19(5): 455-77.

Baross, J. A., J. Liston, et al. (1978). "Incidence of Vibrio parahaemolyticus bacteriophages and other Vibrio bacteriophages in marine samples." Appl Environ Microbiol 36(3): 492-9.

Barton, A., B. Hales, et al. (2012). "The Pacific oyster, Crassostrea gigas, shows negative correlation to naturally elevated carbon dioxide levels: Implications for near-term ocean acidification effects."

Beleneva, I. A., N. V. Zhukova, et al. (2003). "[Comparative study of microbial communities from cultured and natural population of the mussel Mytilus trossulus in Peter the Great bay]." Mikrobiologiia 72(4): 528-34.

Bench, S. R., T. E. Hanson, et al. (2007). "Metagenomic characterization of Chesapeake Bay virioplankton." Appl Environ Microbiol 73(23): 7629-41.

Bennet, S. M., L. Ohman, et al. (2015). "Gut Microbiota as Potential Orchestrators of Irritable Bowel Syndrome." Gut and liver 9(3): 318.

Berglund, O. (1969). "Identification of a thioredoxin induced by bacteriophage T4." J Biol Chem 244(22): 6306-8.

Berglund, O. and A. Holmgren (1975). "Thioredoxin reductase-mediated hydrogen transfer from Escherichia coli thioredoxin-(SH)2 to phage T4 thioredoxin-S2." J Biol Chem 250(8): 2778-82.

218 Besemer, J., A. Lomsadze, et al. (2001). "GeneMarkS: a self-training method for prediction of gene starts in microbial genomes. Implications for finding sequence motifs in regulatory regions." Nucleic Acids Res 29(12): 2607-18.

Boardman, C. L., A. P. Maloy, et al. (2008). "Localization of the bacterial agent of juvenile oyster disease (Roseovarius crassostreae) within affected eastern oysters (Crassostrea virginica)." J Invertebr Pathol 97(2): 150-8.

Bonin, P. (1996). "Anaerobic nitrate reduction to ammonium in two strains isolated from coastal marine sediment: a dissimilatory pathway." FEMS microbiology ecology 19(1): 27-38.

Bosch, A., R. M. Pinto, et al. (2009). Viral contaminants of molluscan shellfish: detection and characterisation, Woodhead Publishing Limited.

Bowman, J. P. and B. Nowak (2004). "Salmonid gill bacteria and their relationship to amoebic gill disease." Journal of Fish Diseases 27: 483-492.

Bracha, R. and D. Mirelman (1984). "Virulence of Entamoeba histolytica trophozoites. Effects of bacteria, microaerobic conditions, and metronidazole." J Exp Med 160(2): 353-68.

Braissant, O., A. W. Decho, et al. (2007). "Exopolymeric substances of sulfate- reducing bacteria: Interactions with calcium at alkaline pH and implication for formation of carbonate minerals." Geobiology 5(4): 401-411.

Breitbart, M., J. H. Miyake, et al. (2004). "Global distribution of nearly identical phage-encoded DNA sequences." FEMS Microbiol Lett 236(2): 249-56.

Breitbart, M. and F. Rohwer (2005). "Here a virus, there a virus, everywhere the same virus?" Trends Microbiol 13(6): 278-84.

Breitbart, M., P. Salamon, et al. (2002). "Genomic analysis of uncultured marine viral communities." Proc Natl Acad Sci U S A 99(22): 14250-5.

Breitbart, M., L. R. Thompson, et al. (2007). "Exploring the vast diversity of marine viruses." Oceanography 20(2): 135-139.

Brinkhoff, T., G. Bach, et al. (2004). "Antibiotic production by a Roseobacter clade- affiliated species from the German Wadden Sea and its antagonistic effects on indigenous isolates." Appl Environ Microbiol 70(4): 2560-5.

Brum, J. R., J. C. Ignacio-Espinoza, et al. (2015). "Ocean plankton. Patterns and ecological drivers of ocean viral communities." Science 348(6237): 1261498.

219 Brunet, R. and L. Garcia-Gil (1996). "Sulfide-induced dissimilatory nitrate reduction to ammonia in anaerobic freshwater sediments." FEMS microbiology ecology 21(2): 131-138.

Brussaard, C. P., S. W. Wilhelm, et al. (2008). "Global-scale processes with a nanoscale drive: the role of marine viruses." ISME J 2(6): 575-8.

Burge, C. A., C. Mark Eakin, et al. (2014). "Climate change influences on marine infectious diseases: implications for management and society." Annual review of marine science 6: 249-277.

Buroker, N. E. (1983). "Population genetics of the American oyster Crassostrea virginica along the Atlantic coast and the Gulf of Mexico." Marine Biology 75(1): 99-112.

Burreson, E. M. and L. M. Ragone Calvo (1996). "Epizootiology of Perkinsus marinus disease of oysters in the Chesapeake Bay, with emphasis on data since 1985." Journal of Shellfish Research 15(1): 17-34.

Bushek, D., S. Allen, et al. (1994). "Dose response of the eastern oyster, Crassostrea virginica, to cultured cells of Perkinsus marinus, the agent of Dermo disease." Journal of Shellfish Research 13: 313.

Bushek, D., S. E. Ford, et al. (2002). "Comparison of in vitro-cultured and wild-type Perkinsus marinus. III. Fecal elimination and its role in transmission." Diseases of aquatic organisms 51(3): 217-225.

Cacchio, P., C. Ercole, et al. (2012). "INVOLVEMENT OF BACTERIA IN THE ORIGIN OF A NEWLY DESCRIBED SPELEOTHEM IN THE GYPSUM CAVE OF GRAVE GRUBBO(CROTONE, ITALY)." Journal of Cave and Karst Studies 74(1): 7-18.

Caporaso, J. G., J. Kuczynski, et al. (2010). "QIIME allows analysis of high- throughput community sequencing data." Nat Methods 7(5): 335-6.

Cavalcanti, G. S., G. B. Gregoracci, et al. (2014). "Physiologic and metagenomic attributes of the rhodoliths forming the largest CaCO3 bed in the South Atlantic Ocean." ISME J 8(1): 52-62.

Cesar Ignacio-Espinoza, J., S. A. Solonenko, et al. (2013). "The global virome: not as big as we thought?" Curr Opin Virol.

220 Chauhan, A., S. Green, et al. (2013). "Whole-genome sequences of five oyster- associated bacteria show potential for crude oil hydrocarbon degradation." Genome Announc 1(5).

Chauhan, A., D. Wafula, et al. (2014). "Metagenomic assessment of the eastern oyster-associated microbiota." Genome Announc 2(5).

Chen, F., J. R. Lu, et al. (2001). "Application of digital image analysis and flow cytometry to enumerate marine viruses stained with SYBR gold." Appl Environ Microbiol 67(2): 539-45.

Chintala, M. M., D. Bushek, et al. (2002). "Comparison of in vitro-cultured and wild- type Perkinsus marinus. II. Dosing methods and host response." Diseases of aquatic organisms 51(3): 203-216.

Chinzei, K. and A. Seilacher (1993). "Remote Biomineralization I: Fill skeletons in vesicular oyster shells.(With 7 figures in the text)." Neues Jahrbuch fur Geologie und Palaontologie-Abhandlungen 190(2): 349-362.

Chow, C. E., R. Sachdeva, et al. (2013). "Temporal variability and coherence of euphotic zone bacterial communities over a decade in the Southern California Bight." ISME J 7(12): 2259-73.

Chu, F.-L. E. and A. K. Volety (1997). "Disease processes of the parasite Perkinsus marinus in eastern oyster Crassostrea virginica: minimum dose for infection initiation, and interaction of temperature, salinity and infective cell does." Oceanographic Literature Review 7(44): 745.

Clarke, K. R. (1993). "Non-parametric multivariate analyses of changes in community structure." Australian journal of ecology 18: 117-117.

Clokie, M. R., A. D. Millard, et al. (2006). "Virus isolation studies suggest short-term variations in abundance in natural cyanophage populations of the Indian Ocean." JMBA-Journal of the Marine Biological Association of the United Kingdom 86(3): 499-506.

Closek, C. J., S. Sunagawa, et al. (2014). "Coral transcriptome and bacterial community profiles reveal distinct Yellow Band Disease states in Orbicella faveolata." ISME J 8(12): 2411-22.

Coen, L. D., R. D. Brumbaugh, et al. (2007). "Ecosystem services related to oyster restoration." Marine Ecology Progress Series 341: 303-307.

221 Colwell, R. K. (2013). EstimateS: Statistical estimation of species richness and shared species from samples.

Colwell, R. R. and J. Liston (1960). "Microbiology of shellfish. Bacteriological study of the natural flora of Pacific oysters (Crassostrea gigas)." Appl Microbiol 8: 104-9.

Comeau, A. M., E. Buenaventura, et al. (2005). "A persistent, productive, and seasonally dynamic vibriophage population within Pacific oysters (Crassostrea gigas)." Appl Environ Microbiol 71(9): 5324-31.

Comeau, A. M. and H. M. Krisch (2008). "The capsid of the T4 phage superfamily: the evolution, diversity, and structure of some of the most prevalent proteins in the biosphere." Mol Biol Evol 25(7): 1321-32.

Company, R., A. Sitja-Bobadilla, et al. (1999). "Bacterial and parasitic pathogens in cultured common dentex, Dentex dentex L." Journal of Fish Diseases 22: 299- 309.

Croft, M. T., A. D. Lawrence, et al. (2005). "Algae acquire vitamin B12 through a symbiotic relationship with bacteria." Nature 438(7064): 90-3.

Crump, B. C., G. W. Kling, et al. (2003). "Bacterioplankton community shifts in an arctic lake correlate with seasonal changes in organic matter source." Applied and Environmental Microbiology 69(4): 2253-2268.

Deng, L., J. C. Ignacio-Espinoza, et al. (2014). "Viral tagging reveals discrete populations in Synechococcus viral genome sequence space." Nature 513(7517): 242-5.

DePaola, A., S. McLeroy, et al. (1997). "Distribution of Vibrio vulnificus phage in oyster tissues and other estuarine habitats." Appl Environ Microbiol 63(6): 2464-7.

DePaola, A., M. L. Motes, et al. (1998). "Phages infecting Vibrio vulnificus are abundant and diverse in oysters (Crassostrea virginica) collected from the Gulf of Mexico." Appl Environ Microbiol 64(1): 346-51.

Desnues, C., B. Rodriguez-Brito, et al. (2008). "Biodiversity and biogeography of phages in modern stromatolites and thrombolites." Nature 452(7185): 340-3.

222 Dessi, D., G. Delogu, et al. (2005). "Long-term survival and intracellular replication of Mycoplasma hominis in Trichomonas vaginalis cells: potential role of the protozoon in transmitting bacterial infection." Infection and immunity 73(2): 1180-1186.

Diamond, E. A. (2012). Do scavengers influence dermo disease (Perkinsus marinus) transmission?: experiments in oyster parasite trophic interactions, Rutgers University-Graduate School-New Brunswick.

Dorrington, T. and M. Gomez-Chiarri (2008). "Antimicrobial Peptides for Use in Oyster Aquaculture: Effect on Pathogens, Commensals, and Eukaryotic Expression Systems." Journal of Shellfish Research 27(2): 365-373.

Drummond, A., B. Ashton, et al. (2011). Geneious v5. 6. 7.

Drummond, A., B. Ashton, et al. (2011). Geneious.

Dubois, E., G. Merle, et al. (2004). "Diversity of enterovirus sequences detected in oysters by RT-heminested PCR." Int J Food Microbiol 92(1): 35-43.

Dungan, C., R. Hamilton, et al. (1996). "Identification of Perkinsus marinus portals of entry by histochemical immunoassays of challenged oysters." J Shellfish Res 15: 500.

Dwivedi, B., B. Xue, et al. (2013). "A bioinformatic analysis of ribonucleotide reductase genes in phage genomes and metagenomes." BMC Evol Biol 13(1): 33.

Edgar, R. C. (2010). "Search and clustering orders of magnitude faster than BLAST." Bioinformatics 26(19): 2460-2461.

Edwards, R. A. and F. Rohwer (2005). "Viral metagenomics." Nat Rev Microbiol 3(6): 504-10.

Elston, R. A., H. Hasegawa, et al. (2008). "Re-emergence of Vibrio tubiashii in bivalve shellfish aquaculture: severity, environmental drivers, geographic extent and management." Dis Aquat Organ 82(2): 119-34.

Embar-Gopinath, S., R. Butler, et al. (2005). "Influence of salmonid gill bacteria on development and severity of amoebic gill disease." Dis Aquat Organ 67(1-2): 55-60.

223 Enav, H., Y. Mandel-Gutfreund, et al. (2014). "Comparative metagenomic analyses reveal viral-induced shifts of host metabolism towards nucleotide biosynthesis." Microbiome 2(1): 1-12.

Espinosa, E. P., E. Corre, et al. (2014). "Pallial mucus of the oyster Crassostrea virginica regulates the expression of putative virulence genes of its pathogen Perkinsus marinus." International journal for parasitology 44(5): 305-317.

Espinosa, E. P., M. Perrigault, et al. (2009). "Lectins associated with the feeding organs of the oyster Crassostrea virginica can mediate particle selection." The Biological Bulletin 217(2): 130-141.

Espinosa, E. P., S. Winnicki, et al. (2013). "Early host-pathogen interactions in a marine bivalve: Crassostrea virginica pallial mucus modulates Perkinsus marinus growth and virulence." Dis Aquat Org 104: 237-247.

Ewart, J. W. and S. E. Ford (1993). History and Impact of MSX and Dermo Diseases on Oyster Stocks in the Northeast Region, Northeastern Regional Aquaculture Center.

Farley, C. A. (1978). "Viruses and viruslike lesions in marine mollusks." Marine Fisheries Review 40(10): 18-20.

Feng, J. S. (1966). "The fate of a virus, Staphylococcus aureus Phage 80, injected into the oyster, Crassostrea virginica." Journal of Invertebrate Pathology 8: 496- 504.

Fernandez-Piquer, J., J. P. Bowman, et al. (2012). "Molecular analysis of the bacterial communities in the live Pacific oyster (Crassostrea gigas) and the influence of postharvest temperature on its structure." J Appl Microbiol 112(6): 1134-43.

Filee, J., E. Bapteste, et al. (2006). "A selective barrier to horizontal gene transfer in the T4-type bacteriophages that has preserved a core genome with the viral replication and structural genes." Mol Biol Evol 23(9): 1688-96.

Filee, J., F. Tetart, et al. (2005). "Marine T4-type bacteriophages, a ubiquitous component of the dark matter of the biosphere." Proc Natl Acad Sci U S A 102(35): 12471-6.

Friedman, C. S. and L. M. Crosson (2012). "Putative phage hyperparasite in the rickettsial pathogen of abalone, "Candidatus Xenohaliotis californiensis"." Microb Ecol 64(4): 1064-72.

224 Galtsoff, P. S. (1964). The american oyster Crassostrea virginica Gmelin, United States Department of the Interior, Fish and Wildlife Service, Bureau of Commercial Fisheries.

Garnier, M., Y. Labreuche, et al. (2007). "Evidence for the involvement of pathogenic bacteria in summer mortalities of the Pacific oyster Crassostrea gigas." Microb Ecol 53(2): 187-96.

Garreis, K. A., J. F. La Peyre, et al. (1996). "The effects of Perkinsus marinus extracellular products and purified proteases on oyster defence parameters in vitro." Fish and Shellfish Immunology 6: 581-597.

Gibson, L. F., J. Woodworth, et al. (1998). "Probiotic activity of Aeromonas media on the Pacific oyster, Crassostrea gigas, when challenged with Vibrio tubiashii." Aquaculture 169: 111-120.

Gill, S. R., M. Pop, et al. (2006). "Metagenomic analysis of the human distal gut microbiome." science 312(5778): 1355-1359.

Gobler, C. J., E. L. DePasquale, et al. (2014). "Hypoxia and acidification have additive and synergistic negative effects on the growth, survival, and metamorphosis of early life stage bivalves." PLoS One 9(1): e83648.

Goedken, M., B. Morsey, et al. (2005). "Immunomodulation of Crassostrea gigas and Crassostrea virginica cellular defense mechanisms by Perkinsus marinus." Journal of Shellfish Research 24(2): 487-496.

Gomez-Leon, J., L. Villamill, et al. (2008). "Survival of eastern oysters Crassostrea virginica from three lines following experimental challenge with bacterial pathogens." Dis Aquat Organ 79(2): 95-105.

Gosling, E. (2015). Marine Bivalve Molluscs, John Wiley & Sons.

Grabowski, J. H. and C. H. Peterson (2007). "Restoring oyster reefs to recover ecosystem services." Theoretical ecology series 4: 281-298.

Green, T. J. and A. C. Barnes (2010). "Bacterial diversity of the digestive gland of Sydney rock oysters, Saccostrea glomerata infected with the paramyxean parasite, Marteilia sydneyi." J Appl Microbiol 109(2): 613-22.

Grizzle, R. E., J. K. Greene, et al. (2008). "Seston removal by natural and constructed intertidal eastern oyster (Crassostrea virginica) reefs: a comparison with previous laboratory studies, and the value of in situ methods." Estuaries and Coasts 31(6): 1208-1220.

225 Guindon, S., J. F. Dufayard, et al. (2010). "New algorithms and methods to estimate maximum-likelihood phylogenies: assessing the performance of PhyML 3.0." Syst Biol 59(3): 307-21.

Guindon, S. and O. Gascuel (2003). "A simple, fast, and accurate algorithm to estimate large phylogenies by maximum likelihood." Syst Biol 52(5): 696-704.

Hahn, M. W. and M. Pockl (2005). "Ecotypes of planktonic actinobacteria with identical 16S rRNA genes adapted to thermal niches in temperate, subtropical, and tropical freshwater habitats." Appl Environ Microbiol 71(2): 766-73.

Hayes, P. F. and R. W. Menzel (1981). "The reproductive cycle of early setting Crassostrea virginica (Gmelin) in the northern Gulf of Mexico, and its implications for population recruitment." The Biological Bulletin 160(1): 80- 88.

Heidelberg, J. F., K. B. Heidelberg, et al. (2002). "Seasonality of Chesapeake Bay bacterioplankton species." Appl Environ Microbiol 68(11): 5488-97.

Helliwell, K. E., G. L. Wheeler, et al. (2011). "Insights into the evolution of vitamin B12 auxotrophy from sequenced algal genomes." Mol Biol Evol 28(10): 2921- 33.

Hendrix, R. W., M. C. Smith, et al. (1999). "Evolutionary relationships among diverse bacteriophages and prophages: all the world's a phage." Proc Natl Acad Sci U S A 96(5): 2192-7.

Henn, M. R., M. B. Sullivan, et al. (2010). "Analysis of high-throughput sequencing and annotation strategies for phage genomes." PLoS One 5(2): e9083.

Hill, M. (1997). "Intestinal flora and endogenous vitamin synthesis." European Journal of Cancer Prevention 6(2): S43-S45.

Hug, L. A., B. C. Thomas, et al. (2015). "Aquifer environment selects for microbial species cohorts in sediment and groundwater." The ISME journal.

Hughes, F. M., B. Foster, et al. (2010). "Apoptosis as a host defense mechanism in Crassostrea virginica and its modulation by Perkinsus marinus." Fish & shellfish immunology 29(2): 247-257.

Hurwitz, B. L., A. H. Westveld, et al. (2014). "Modeling ecological drivers in marine viral communities using comparative metagenomics and network analyses." Proceedings of the National Academy of Sciences 111(29): 10714-10719.

226 Husmann, G., G. Gerdts, et al. (2010). "Spirochetes in crystalline styles of marine bivalves: Group-specific PCR detection and 16S rRNA sequence analysis." Journal of Shellfish Research 29(4): 1069-1075.

Huttenhower, C. (2012). "Structure, function and diversity of the healthy human microbiome." Nature 486(7402): 207-214.

Jaspers, E. and J. Overmann (2004). "Ecological significance of microdiversity: identical 16S rRNA gene sequences can be found in bacteria with highly divergent genomes and ecophysiologies." Appl Environ Microbiol 70(8): 4831-9.

John, S. G., C. B. Mendez, et al. (2011). "A simple and efficient method for concentration of ocean viruses by chemical flocculation." Environ Microbiol Rep 3(2): 195-202.

Kaewsrichan, J., K. Peeyananjarassri, et al. (2006). "Selection and identification of anaerobic lactobacilli producing inhibitory compounds against vaginal pathogens." FEMS Immunology & Medical Microbiology 48(1): 75-83.

Kallmeyer, J., R. Pockalny, et al. (2012). "Global distribution of microbial abundance and biomass in subseafloor sediment." Proceedings of the National Academy of Sciences 109(40): 16213-16216.

Kang, I., H. M. Oh, et al. (2013). "Genome of a SAR116 bacteriophage shows the prevalence of this phage type in the oceans." Proc Natl Acad Sci U S A 110(30): 12343-8.

Karim, M., W. Zhao, et al. (2013). "Probiotic Strains for Shellfish Aquaculture: Protection of Eastern Oyster, Crassostrea virginica, Larvae and Juveniles Againsl Bacterial Challenge." Journal of Shellfish Research 32(2): 401-408.

Katoh, K., K. Misawa, et al. (2002). "MAFFT: a novel method for rapid multiple sequence alignment based on fast Fourier transform." Nucleic Acids Res 30(14): 3059-66.

Kazamia, E., H. Czesnick, et al. (2012). "Mutualistic interactions between vitamin B(12) -dependent algae and heterotrophic bacteria exhibit regulation." Environ Microbiol 14(6): 1466-76.

Kellogg, M. L., J. C. Cornwell, et al. (2013). "Denitrification and nutrient assimilation on a restored oyster reef." Mar Ecol Prog Ser 480: 1-19.

227 Kennedy, J., B. Flemer, et al. (2010). "Marine metagenomics: new tools for the study and exploitation of marine microbial metabolism." Mar Drugs 8(3): 608-28.

Kennedy, V. and L. Krantz (1982). "Comparative gametogenic and spawning patterns of the oyster Crassostrea virginica (Gmelin) in central Chesapeake Bay [Maryland]." Journal of Shellfish Research.

Kennedy, V. S., R. I. Newell, et al. (1996). The eastern oyster: Crassostrea virginica, University of Maryland Sea Grant College.

Kesarcodi-Watson, A., P. Miner, et al. (2012). "Protective effect of four potential probiotics against pathogen-challenge of the larvae of three bivalves: Pacific oyster (Crassostrea gigas), flat oyster (Ostrea edulis) and scallop (Pecten maximus)." Aquaculture 344-349: 29-34.

Khosravi, A. and S. K. Mazmanian (2013). "Disruption of the gut microbiome as a risk factor for microbial infections." Current opinion in microbiology 16(2): 221-227.

Kielbasa, S. M., R. Wan, et al. (2011). "Adaptive seeds tame genomic sequence comparison." Genome Res 21(3): 487-93.

King, G. M., C. Judd, et al. (2012). "Analysis of stomach and gut microbiomes of the eastern oyster (Crassostrea virginica) from coastal Louisiana, USA." PLoS One 7(12): e51475.

Kristensen, D. M., A. R. Mushegian, et al. (2010). "New dimensions of the virus world discovered through metagenomics." Trends Microbiol 18(1): 11-9.

Kueh, C. S. and K. Y. Chan (1985). "Bacteria in bivalve shellfish with special reference to the oyster." J Appl Bacteriol 59(1): 41-7.

Kumar, S. (2012). Textbook of Microbiology, Jaypee Brothers,Medical Publishers Pvt. Limited.

La Valley, K. J., S. Jones, et al. (2009). "Bacterial Community Profiling of the Eastern Oyster (Crassostrea virginica): Comparison of Culture-Dependent and Culture-Independent Outcomes." Journal of Shellfish Research 28(4): 827-835.

Labonte, J. M., E. K. Field, et al. (2015). "Single cell genomics indicates horizontal gene transfer and viral infections in a deep subsurface Firmicutes population." Front Microbiol 6: 349.

228 Labonte, J. M., K. E. Reid, et al. (2009). "Phylogenetic analysis indicates evolutionary diversity and environmental segregation of marine podovirus DNA polymerase gene sequences." Appl Environ Microbiol 75(11): 3634-40.

Labonte, J. M., B. K. Swan, et al. (2015). "Single-cell genomics-based analysis of virus-host interactions in marine surface bacterioplankton." ISME J.

Labrie, S., K. Frois‚ÄêMoniz, et al. (2013). "Genomes of marine cyanopodoviruses reveal multiple origins of diversity." Environmental microbiology 15(5): 1356- 1376.

Lategan, M. J., W. Booth, et al. (2006). "An inhibitory substance produced by Aeromonas media A199, an aquatic probiotic." Aquaculture 254: 115-124.

Le Guyader, F., F. Loisy, et al. (2006). "Norwalk virus-specific binding to oyster digestive tissues." Emerg Infect Dis 12(6): 931-6.

Lenihan, H. S. and C. H. Peterson (1998). "How habitat degradation through fishery disturbance enhances impacts of hypoxia on oyster reefs." Ecological Applications 8(1): 128-140.

Lim, H. J., D. Kapareiko, et al. (2011). "Isolation and evaluation of new probiotic bacteria for use in shellfish hatcheries: I. Isolation and screening for bioactivity." Journal of Shellfish Research 30(3): 609-615.

Lindell, D., J. D. Jaffe, et al. (2007). "Genome-wide expression dynamics of a marine virus and host reveal features of co-evolution." Nature 449(7158): 83-6.

Logan, D. T., J. Andersson, et al. (1999). "A glycyl radical site in the crystal structure of a class III ribonucleotide reductase." Science 283(5407): 1499-504.

Lokmer, A. and K. Mathias Wegner (2015). "Hemolymph microbiome of Pacific oysters in response to temperature, temperature stress and infection." ISME J 9(3): 670-82.

Loosanoff, V. L. (1965). The American or eastern oyster, US Bureau of Commercial Fisheries.

Lopez-Bueno, A., J. Tamames, et al. (2009). "High diversity of the viral community from an Antarctic lake." Science 326(5954): 858-61.

Lovelace, T. E., H. Tubiash, et al. (1967). "Quantitative and Qualitative Commensal Bacterial Flora of Crassostrea virginica in Chesapeake Bay." Proceedings of the National Shellfisheries Association 58: 82-87.

229 Lozupone, C. and R. Knight (2005). "UniFrac: a new phylogenetic method for comparing microbial communities." Applied and Environmental Microbiology 71(12): 8228-8235.

Lundin, D., S. Gribaldo, et al. (2010). "Ribonucleotide reduction - horizontal transfer of a required function spans all three domains." BMC Evol Biol 10: 383.

Luo, C., D. Tsementzi, et al. (2012). "Direct comparisons of Illumina vs. Roche 454 sequencing technologies on the same microbial community DNA sample." PLoS One 7(2): e30087.

Mackenzie, C. L., S. A. Lynch, et al. (2014). "Future Oceanic Warming and Acidification Alter Immune Response and Disease Status in a Commercial Shellfish Species, Mytilus edulis L." PLoS One 9(6): e99712.

Mackenzie, C. L., G. A. Ormondroyd, et al. (2014). "Ocean warming, more than acidification, reduces shell strength in a commercial shellfish species during food limitation." PLoS One 9(1): e86764.

Mackin, J. (1951). "Histopathology of infection of Crassostrea virginica (Gmelin) by Dermocystidium marinum Mackin, Owen, and Collier." Bulletin of Marine Science 1(1): 72-87.

Maloy, A., S. E. Ford, et al. (2007). "Roseovarius crassostreae, the etiological agent of Juvenile Oyster Disease (now to be known as Roseovarius Oyster Disease) in Crassostrea virginica." Aquaculture 269: 71-83.

Marchler-Bauer, A., S. Lu, et al. (2011). "CDD: a Conserved Domain Database for the functional annotation of proteins." Nucleic Acids Res 39(Database issue): D225-9.

Marine, R., C. McCarren, et al. (2014). "Caught in the middle with multiple displacement amplification: the myth of pooling for avoiding multiple displacement amplification bias in a metagenome." Microbiome 2(1): 3.

Mathiesen, A. (2012). "The State of the World Fisheries and Aquaculture 2012."

Matsuzaki, M., H. Kuroiwa, et al. (2008). "A cryptic algal group unveiled: a plastid biosynthesis pathway in the oyster parasite Perkinsus marinus." Molecular biology and evolution 25(6): 1167-1179.

Mazmanian, S. K., C. H. Liu, et al. (2005). "An immunomodulatory molecule of symbiotic bacteria directs maturation of the host immune system." Cell 122(1): 107-118.

230 McCollough, C. B., B. W. Albright, et al. (2007). "Acquisition and progression of Perkinsus marinus infections by Specific-Pathogen-Free juvenile oysters (Crassostrea virginica GMELIN) in a mesohaline Chesapeake Bay tributary." Journal of Shellfish Research 26(2): 465-477.

McDonald, D., M. N. Price, et al. (2011). "An improved Greengenes taxonomy with explicit ranks for ecological and evolutionary analyses of bacteria and archaea." The ISME journal 6(3): 610-618.

Meyer, D. L., E. C. Townsend, et al. (1997). "Stabilization and erosion control value of oyster cultch for intertidal marsh." Restoration Ecology 5(1): 93-99.

Morris, R. M., M. S. Rappe, et al. (2002). "SAR11 clade dominates ocean surface bacterioplankton communities." Nature 420(6917): 806-10.

Mount, A. S., A. P. Wheeler, et al. (2004). "Hemocyte-mediated shell mineralization in the eastern oyster." Science 304(5668): 297-300.

Munn, C. B. (2006). "Viruses as pathogens of marine organisms-- from bacteria to whales." Journal of the Marine Biological Association of the United Kingdom 86: 453-467.

Murchelano, R. A. and C. Brown (1968). "Bacteriological study of the natural flora of the Eastern oyster, Crassostrea virginica." Journal of Invertebrate Pathology 11: 520-521.

Myers, E. W., G. G. Sutton, et al. (2000). "A whole-genome assembly of Drosophila." Science 287(5461): 2196-204.

Nakamura, A., K. G. Takahashi, et al. (1999). "Vibriostatic bacteria isolated from rearing seawater of oyster brood stock: Potentiality as biocontrol agents for vibriosis in oyster larvae." Fish Pathology 34(3): 139-144.

Newell, R. and S. J. Jordan (1983). "Preferential ingestion of organic material by the American oyster Crassostrea virginica." Marine ecology progress series. Oldendorf 13(1): 47-53.

Newell, R. I. (1988). "Ecological changes in Chesapeake Bay: Are they the result of overharvesting the American oyster, Crassostrea virginica." Understanding the estuary: advances in Chesapeake Bay research 129: 536-546.

Newell, R. I., J. C. Cornwell, et al. (2002). "Influence of simulated bivalve biodeposition and microphytobenthos on sediment nitrogen dynamics: A laboratory study." Limnology and Oceanography 47(5): 1367-1379.

231 Nikkola, M., A. Engstrom, et al. (1993). "An elongated form of T4 glutaredoxin with four extra residues." Biochemistry 32(28): 7133-5.

Noble, R. T. and J. A. Fuhrman (1997). "Virus decay and its causes in coastal waters." Appl Environ Microbiol 63(1): 77-83.

Noguchi, H., T. Taniguchi, et al. (2008). "MetaGeneAnnotator: detecting species- specific patterns of ribosomal binding site for precise gene prediction in anonymous prokaryotic and phage genomes." DNA Res 15(6): 387-96.

Nordlund, P. and P. Reichard (2006). "Ribonucleotide reductases." Annu Rev Biochem 75: 681-706.

Olafsen, J. A., H. V. Mikkelsen, et al. (1993). "Indigenous bacteria in hemolymph and tissues of marine bivalves at low temperatures." Appl Environ Microbiol 59(6): 1848-54.

Paerl, H. W., J. L. Pinckney, et al. (1998). "Ecosystem responses to internal and watershed organic matter loading: consequences for hypoxia in the eutrophying Neuse River Estuary, North Carolina, USA." Marine Ecology Progress Series 166: 17.

Paillard, C., F. Le Roux, et al. (2004). "Bacterial disease in marine bivalves, a review of recent studies: Trends and evolution." Aquatic Living Resources 17: 477- 498.

Pantos, O. and J. C. Bythell (2006). "Bacterial community structure associated with white band disease in the elkhorn coral Acropora palmata determined using culture-independent 16S rRNA techniques." Dis Aquat Organ 69(1): 79-88.

Peterson, C. H. and R. N. Lipcius (2003). "Conceptual progress towards predicting quantitative ecosystem benefits of ecological restorations." Marine Ecology Progress Series 264(11).

Petrov, V. M., J. M. Nolan, et al. (2006). "Plasticity of the gene functions for DNA replication in the T4-like phages." J Mol Biol 361(1): 46-68.

Petrov, V. M., S. Ratnayaka, et al. (2010). "Genomes of the T4-related bacteriophages as windows on microbial genome evolution." Virol J 7: 292.

Polson, S. W., S. W. Wilhelm, et al. (2011). "Unraveling the viral tapestry (from inside the capsid out)." ISME J 5(2): 165-8.

232 Porsby, C. H., M. A. Webber, et al. (2011). "Resistance and tolerance to tropodithietic acid, an antimicrobial in aquaculture, is hard to select." Antimicrob Agents Chemother 55(4): 1332-7.

Porter, E. T., J. C. Cornwell, et al. (2004). "Effect of oysters Crassostrea virginica and bottom shear velocity on benthic-pelagic coupling and estuarine water quality." Marine Ecology Progress Series 271: 61-75.

Powell, E. N., K. A. Ashton-Alcox, et al. (2008). "Long-term trends in oyster population dynamics in Delaware Bay: Regime shifts and response to disease." Journal of Shellfish Research 27(4): 729-755.

Powell, E. N., J. M. Klinck, et al. (2011). "The potential for oysters, Crassostrea virginica, to develop resistance to Dermo disease in the field: Evaluation using a gene-based population dynamics model." Journal of Shellfish Research 30(3): 685-712.

Pujalte, M. J., M. Ortigosa, et al. (1999). "Aerobic and facultative anaerobic heterotrophic bacteria associated to Mediterranean oysters and seawater." Int Microbiol 2(4): 259-66.

Quince, C., A. Lanzen, et al. (2011). "Removing noise from pyrosequenced amplicons." BMC Bioinformatics 12: 38.

Rabe, P., T. A. Klapschinski, et al. (2014). "Synthesis and bioactivity of analogues of the marine antibiotic tropodithietic acid." Beilstein J Org Chem 10: 1796-801.

Raghunathan, A., H. R. Ferguson, Jr., et al. (2005). "Genomic DNA amplification from a single bacterium." Appl Environ Microbiol 71(6): 3342-7.

Ragone Calvo, L. M., G. W. Calvo, et al. (2003). "Dual disease resistance in a selectively bred eastern oyster, Crassostrea virginica, strain tested in Chesapeake Bay." Aquaculture(220): 69-87.

Ray, S. M. (1966). A review of the culture method for detecting Dermocystidium marinum, with suggested modifications and precautions. Proceedings of the National Shellfisheries Association.

Reid, G., J. Howard, et al. (2001). "Can bacterial interference prevent infection?" Trends in microbiology 9(9): 424-428.

Renault, T. and B. Novoa (2004). "Viruses infecting bivalve molluscs." Aquatic Living Resources 17: 397-409.

233 Riquelme, C., G. Hayashida, et al. (1996). "Isolation of a native bacterial strain from the scallop Argopecten purpuratus with inhibitory effects against pathogenic Vibrios." Journal of Shellfish Research 15(2): 369-374.

Roder, C., C. Arif, et al. (2014). "Bacterial profiling of White Plague Disease in a comparative coral species framework." ISME J 8(1): 31-9.

Roder, C., C. Arif, et al. (2014). "Bacterial profiling of White Plague Disease across corals and oceans indicates a conserved and distinct disease microbiome." Mol Ecol 23(4): 965-74.

Rodney, W. S. and K. T. Paynter (2006). "Comparisons of macrofaunal assemblages on restored and non-restored oyster reefs in mesohaline regions of Chesapeake Bay in Maryland." Journal of Experimental Marine Biology and Ecology 335(1): 39-51.

Rodriguez-Brito, B., L. Li, et al. (2010). "Viral and microbial community dynamics in four aquatic environments." ISME J 4(6): 739-51.

Rohwer, F. and R. V. Thurber (2009). "Viruses manipulate the marine environment." Nature 459(7244): 207-12.

Romero, A., B. Novoa, et al. (2012). "Genomics, immune studies and diseases in bivalve aquaculture." Invertebrate Survival Journal 9: 110-121.

Romero, J. and R. Espejo (2001). "The prevalence of noncultivable bacteria in oysters (Tiostrea chilensis, Philippi, 1845)." Journal of Shellfish Research 20(3): 1235-1240.

Romero, J., M. Garcia-Varela, et al. (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)." Microb Ecol 44(4): 365-71.

Rothschild, B. J., J. S. Ault, et al. (1994). "Decline of the Chesapeake Bay oyster population: a century of habitat destruction and overfishing." Marine Ecology Progress Series 111: 29-39.

Rothschild, L. J. and R. L. Mancinelli (2001). "Life in extreme environments." Nature 409(6823): 1092-1101.

Roux, S., A. K. Hawley, et al. (2014). "Ecology and evolution of viruses infecting uncultivated SUP05 bacteria as revealed by single-cell- and meta-genomics." Elife 3: e03125.

234 Rusch, D. B., A. L. Halpern, et al. (2007). "The Sorcerer II Global Ocean Sampling expedition: northwest Atlantic through eastern tropical Pacific." PLoS Biol 5(3): e77.

Sakowski, E., W. Kress, et al. (2011). Bacteriophage and Viral Ecology as Seen Through the Lens of Nucleic Acid Sequence Data. Studies in Viral Ecology: Microbial and Botanical Host Systems. C. J. Hurst, John Wiley & Sons, Inc. 1: 145-168.

Sakowski, E. G., E. V. Munsell, et al. (2014). "Ribonucleotide reductases reveal novel viral diversity and predict biological and ecological features of unknown marine viruses." Proc Natl Acad Sci U S A 111(44): 15786-91.

Sandaa, R. A., L. Gomez-Consarnau, et al. (2009). "Viral control of bacterial biodiversity--evidence from a nutrient-enriched marine mesocosm experiment." Environ Microbiol 11(10): 2585-97.

Sauvage, C., J. F. Pepin, et al. (2009). "Ostreid herpes virus 1 infection in families of the Pacific oyster, Crassostrea gigas, during a summer mortality outbreak: differences in viral DNA detection and quantification using real-time PCR." Virus Res 142(1-2): 181-7.

Sawabe, T., N. Setoguchi, et al. (2003). "Acetic acid production of< i> Vibrio halioticoli from alginate: a possible role for establishment of abalone‚Äì< i> V. halioticoli association." Aquaculture 219(1): 671-679.

Scanlon, C. H. (1997). "The potential for transmission of Perkinsus marinus by fecal matter from the eastern oyster, Crassostrea virginica."

Schloss, P. D., S. L. Westcott, et al. (2009). "Introducing mothur: open-source, platform-independent, community-supported software for describing and comparing microbial communities." Appl Environ Microbiol 75(23): 7537-41.

Schmidt, H. F., E. G. Sakowski, et al. (2014). "Shotgun metagenomics indicates novel family A DNA polymerases predominate within marine virioplankton." ISME J 8(1): 103-14.

Schmitt, P., R. D. Rosa, et al. (2012). "The Antimicrobial Defense of the Pacific Oyster, Crassostrea gigas. How Diversity may Compensate for Scarcity in the Regulation of Resident/Pathogenic Microflora." Front Microbiol 3: 160.

Schoenfeld, T., M. Liles, et al. (2010). "Functional viral metagenomics and the next generation of molecular tools." Trends Microbiol 18(1): 20-9.

235 Seliger, H. and J. Boggs (1988). "Evidence for loss of suitable habitats for oysters in tributaries of the Chesapeake Bay." Understanding the estuary, advances in Chesapeake Bay research. Publication(120): 111-127.

Shiah, F. and H. W. Ducklow (1994). "Temperature regulation of heterotrophic bacterioplankton abundance, production, and specific growth rate in Chesapeake Bay." Oceanography 39(6).

Shimada, Y., M. Kinoshita, et al. (2013). "Commensal bacteria-dependent indole production enhances epithelial barrier function in the colon." PLoS One 8(11): e80604.

Shumway, S. E. (1996). "Natural environmental factors." The Eastern Oyster Crassostrea virginica. Maryland Sea Grant College, University of Maryland, College Park, Maryland: 467-513.

Sitja-Bobadilla, A., M. J. Pujalte, et al. (2006). "Interactions between bacteria and Cryptosporidium molnari in gilthead sea bream (Sparus aurata) under farm and laboratory conditions." Veterinary Parasitology 142: 248-259.

Siu, N., J. K. Apple, et al. (2014). "The Effects of Ocean Acidity and Elevated Temperature on Bacterioplankton Community Structure and Metabolism." Open Journal of Ecology 4(08): 434.

Smith, E. M. (1998). "Coherence of microbial respiration rate and cell-specific bacterial activity in a coastal planktonic community." Aquatic microbial ecology 16(1): 27-35.

Smith, G., D. Bruce, et al. (2005). "Assessment of recent habitat conditions of eastern oyster Crassostrea virginica bars in mesohaline Chesapeake Bay." North American Journal of Fisheries Management 25(4): 1569-1590.

Soniat, T. M., J. M. Klinck, et al. (2012). "Understanding the success and failure of oyster populations: Periodicities of Perkinsus marinus, and oyster recruitment, mortality, and size." Journal of Shellfish Research 31(3): 635-646.

Srinivasiah, S., J. Bhavsar, et al. (2008). "Phages across the biosphere: contrasts of viruses in soil and aquatic environments." Res Microbiol 159(5): 349-57.

Stunz, G. W., T. J. Minello, et al. (2010). "Relative value of oyster reef as habitat for estuarine nekton in Galveston Bay, Texas." Marine Ecology Progress Series 406: 147-159.

236 Sullivan, M. B. (2015). "Viromes, not gene markers, for studying double-stranded DNA virus communities." J Virol 89(5): 2459-61.

Sullivan, M. B., M. L. Coleman, et al. (2005). "Three Prochlorococcus cyanophage genomes: signature features and ecological interpretations." PLoS Biol 3(5): e144.

Sullivan, M. B., D. Lindell, et al. (2006). "Prevalence and evolution of core photosystem II genes in marine cyanobacterial viruses and their hosts." PLoS Biol 4(8): e234.

Sun, J. and X. Wu (2004). "Histology, ultrastructure, and morphogenesis of a rickettsia-like organism causing disease in the oyster, Crassostrea ariakensis Gould." J Invertebr Pathol 86(3): 77-86.

Sunagawa, S., T. Z. DeSantis, et al. (2009). "Bacterial diversity and White Plague Disease-associated community changes in the Caribbean coral Montastraea faveolata." ISME J 3(5): 512-21.

Sunila, I. and J. LaBanca (2003). "Apoptosis in the pathogenesis of infectious diseases of the eastern oyster Crassostrea virginica." Dis Aquat Organ 56(2): 163-70.

Suttle, C. A. (2005). "Viruses in the sea." Nature 437(7057): 356-61.

Suttle, C. A. and J. A. Fuhrman (2010). Enumeration of virus particles in aquatic or sediment samples by epifluorescence microscopy. Manual of aquatic viralecology. S. W. Wilhelm, M. G. Weinbauer and C. A. Suttle, American Society of Limnology and Oceanography, Inc.: 145-153.

Suzek, B. E., H. Huang, et al. (2007). "UniRef: comprehensive and non-redundant UniProt reference clusters." Bioinformatics 23(10): 1282-8.

Szewzyk, R. and N. Pfennig (1987). "Complete oxidation of catechol by the strictly anaerobic sulfate-reducing Desulfobacterium catecholicum sp. nov." Archives of microbiology 147(2): 163-168.

Tabor, S., H. Huber, et al. (1987). "Escherichia coli thioredoxin confers processivity on the DNA polymerase activity of the gene 5 protein of bacteriophage T7." Journal of Biological Chemistry 262(33): 16212-16223.

Tadmor, A. D., E. A. Ottesen, et al. (2011). "Probing individual environmental bacteria for viruses by using microfluidic digital PCR." Science 333(6038): 58- 62.

237 Tall, B. D., J. F. La Peyre, et al. (1999). "Perkinsus marinus extracellular protease modulates survival of Vibrio vulnificus in Eastern oyster (Crassostrea virginica) hemocytes." Appl Environ Microbiol 65(9): 4261-3.

Tarnowski, M. (2012). Maryland Oyster Population Status Report: 2011 Fall Survey, Maryland Department of Natural Resources.

Tarnowski, M. (2014). Maryland Oyster Population Status Report: 2013 Fall Survey. M. D. o. N. Resources.

Tasumi, S. and G. R. Vasta (2007). "A galectin of unique domain organization from hemocytes of the eastern oyster (Crassostrea virginica) is a receptor for the protistan parasite Perkinsus marinus." The Journal of Immunology 179(5): 3086-3098.

Thayer, G. W., H. H. Stuart, et al. (1978). "Habitat values of salt marshes, mangroves, and seagrasses for aquatic organisms." Wetland functions and values: the state of our understanding. American Water Resources Association: 235-247.

Thomas, J. C. t., D. Wafula, et al. (2014). "A survey of deepwater horizon (DWH) oil- degrading bacteria from the Eastern oyster biome and its surrounding environment." Front Microbiol 5: 149.

Thompson, F., Y. Li, et al. (2003). "Vibrio neptunius sp. nov., Vibrio brasiliensis sp. nov. and Vibrio xuii sp. nov., isolated from the marine aquaculture environment (bivalves, fish, rotifers and shrimps)." International journal of systematic and evolutionary microbiology 53(1): 245-252.

Tolley, S. G. and A. K. Volety (2005). "The role of oysters in habitat use of oyster reefs by resident fishes and decapod crustaceans." Journal of Shellfish Research 24(4): 1007-1012.

Trabal, N., J. M. Mazon-Suastegui, et al. (2012). "Molecular analysis of bacterial microbiota associated with oysters (Crassostrea gigas and Crassostrea corteziensis) in different growth phases at two cultivation sites." Microb Ecol 64(2): 555-69.

Tubiash, H. S., P. E. Chanley, et al. (1965). "Bacillary necrosis, a disease of larval and juvenile bivalve mollusks I. Etiology and epizootiology." Journal of bacteriology 90(4): 1036-1044.

Turnbaugh, P. J., R. E. Ley, et al. (2007). "The human microbiome project: exploring the microbial part of ourselves in a changing world." Nature 449(7164): 804.

238 Turnbaugh, P. J., R. E. Ley, et al. (2006). "An obesity-associated gut microbiome with increased capacity for energy harvest." Nature 444(7122): 1027-31.

Ulanowicz, R. E. and J. H. Tuttle (1992). "The trophic consequences of oyster stock rehabilitation in Chesapeake Bay." Estuaries 15(3): 298-306.

Uriz, M. J., G. Agell, et al. (2012). "Endosymbiotic calcifying bacteria: a new cue to the origin of calcification in metazoa?" Evolution 66(10): 2993-9.

Vage, S., J. E. Storesund, et al. (2012). "Adding a cost of resistance description extends the ability of virus-host model to explain observed patterns in structure and function of pelagic microbial communities." Environ Microbiol.

Vage, S., J. E. Storesund, et al. (2013). "Adding a cost of resistance description extends the ability of virus-host model to explain observed patterns in structure and function of pelagic microbial communities." Environmental microbiology 15(6): 1842-1852.

Vasconcelos, G. J. and J. S. Lee (1972). "Microbial flora of Pacific oysters (Crassostrea gigas) subjected to ultraviolet-irradiated seawater." Appl Microbiol 23(1): 11-6.

Vermeij, G. J. (2013). "The oyster enigma variations: a hypothesis of microbial calcification." Paleobiology 40(1): 1-13.

Vezzulli, L., I. Brettar, et al. (2012). "Long-term effects of ocean warming on the prokaryotic community: evidence from the vibrios." ISME J 6(1): 21-30.

Vezzulli, L., R. R. Colwell, et al. (2013). "Ocean warming and spread of pathogenic vibrios in the aquatic environment." Microbial Ecology 65(4): 817-825.

Volety, A. K. and F.-L. E. Chu (1994). "Comparison of infectivity and pathogenicity of meront (trophozoite) and prezoosporangiae stages of the oyster pathogen Perkinsus marinus in eastern oysters, Crassostrea virginica (Gmelin, 1791)." Journal of Shellfish Research 13(2): 521-528.

Volety, A. K. and F. L. Chu (1995). "Suppression of chemiluminescence of eastern oyster (Crassostrea virginica) hemocytes by the protozoan parasite Perkinsus marinus." Dev Comp Immunol 19(2): 135-42.

Wagner-Dobler, I., B. Ballhausen, et al. (2010). "The complete genome sequence of the algal symbiont Dinoroseobacter shibae: a hitchhiker's guide to life in the sea." ISME J 4(1): 61-77.

239 Waldbusser, G. G., E. L. Brunner, et al. (2013). "A developmental and energetic basis linking larval oyster shell formation to acidification sensitivity." Geophysical Research Letters 40(10): 2171-2176.

Waldbusser, G. G., E. P. Voigt, et al. (2011). "Biocalcification in the eastern oyster (Crassostrea virginica) in relation to long-term trends in Chesapeake Bay pH." Estuaries and Coasts 34(2): 221-231.

Wang, Q., G. M. Garrity, et al. (2007). "Naive Bayesian classifier for rapid assignment of rRNA sequences into the new bacterial taxonomy." Applied and Environmental Microbiology 73(16): 5261-5267.

Wang, S., E. Peatman, et al. (2010). "Microarray analysis of gene expression in eastern oyster (Crassostrea virginica) reveals a novel combination of antimicrobial and oxidative stress host responses after dermo (Perkinsus marinus) challenge." Fish Shellfish Immunol 29(6): 921-9.

Wang, X., L. Li, et al. (2013). "Oyster Shell Proteins Originate from Multiple Organs and Their Probable Transport Pathway to the Shell Formation Front." PLoS One 8(6): e66522.

Ward, D. M., R. Weller, et al. (1990). "16S rRNA sequences reveal numerous uncultured microorganisms in a natural community." Nature 345(6270): 63-5.

Ward, J. E., J. S. Levinton, et al. (1997). "Site of particle selection in a bivalve mollusc." Nature 390(6656): 131-132.

Wegner, K. M., N. Volkenborn, et al. (2013). "Disturbance induced decoupling between host genetics and composition of the associated microbiome." BMC Microbiol 13: 252.

White, M. E., E. Powell, et al. (1987). Host-to-host transmission of Perkinsus marinus in oyster (Crassostrea virginica) populations by the ectoparasitic snail Boonea impressa (Pyramidellidae), National Shellfisheries Association.

Williamson, S. J., D. B. Rusch, et al. (2008). "The Sorcerer II Global Ocean Sampling Expedition: metagenomic characterization of viruses within aquatic microbial samples." PLoS One 3(1): e1456.

Wilson, C., L. Scotto, et al. (2005). "Survey of water quality, oyster reproduction and oyster health status in the St. Lucie Estuary." Journal of Shellfish Research 24(1): 157-165.

240 Winget, D. M., R. R. Helton, et al. (2011). "Repeating patterns of virioplankton production within an estuarine ecosystem." Proc Natl Acad Sci U S A 108(28): 11506-11.

Winget, D. M. and K. E. Wommack (2008). "Randomly amplified polymorphic DNA PCR as a tool for assessment of marine viral richness." Appl Environ Microbiol 74(9): 2612-8.

Winget, D. M. and K. E. Wommack (2009). "Diel and daily fluctuations in virioplankton production in coastal ecosystems." Environ Microbiol 11(11): 2904-14.

Wommack, K. E., J. Bhavsar, et al. (2012). "VIROME: a standard operating procedure for analysis of viral metagenome sequences." Standards in Genomic Sciences 6(3): 421-433.

Wommack, K. E., J. Bhavsar, et al. (2008). "Metagenomics: read length matters." Appl Environ Microbiol 74(5): 1453-63.

Wommack, K. E. and R. R. Colwell (2000). "Virioplankton: viruses in aquatic ecosystems." Microbiol Mol Biol Rev 64(1): 69-114.

Wommack, K. E., D. J. Nasko, et al. (2015). "Counts and sequences, observations that continue to change our understanding of viruses in nature." J Microbiol 53(3): 181-92.

Wommack, K. E., T. l. Sime-Ngando, et al. (2010). "Filtration-based methods for the collection of viral concentrates from large water samples." Manual of Aquatic Viral Ecology (MAVE): 110-117.

Zhang, G., X. Fang, et al. (2012). "The oyster genome reveals stress adaptation and complexity of shell formation." Nature 490(7418): 49-54.

Zhao, J., P. D. Schloss, et al. (2012). "Decade-long bacterial community dynamics in cystic fibrosis airways." Proceedings of the National Academy of Sciences 109(15): 5809-5814.

Zhao, Y., B. Temperton, et al. (2013). "Abundant SAR11 viruses in the ocean." Nature 494(7437): 357-60.

Zhong, Y., F. Chen, et al. (2002). "Phylogenetic diversity of marine cyanophage isolates and natural virus communities as revealed by sequences of viral capsid assembly protein gene g20." Appl Environ Microbiol 68(4): 1576-84.

241 Zhu, W., A. Lomsadze, et al. (2010). "Ab initio gene identification in metagenomic sequences." Nucleic Acids Res 38(12): e132.

Zu Ermgassen, P. S., M. D. Spalding, et al. (2012). "Historical ecology with real numbers: past and present extent and biomass of an imperilled estuarine habitat." Proc Biol Sci 279(1742): 3393-400.

242 Appendix

Chapter 4 is modified from journal article, “RIBONUCLEOTIDE REDUCTASES REVEAL NOVEL VIRAL DIVERSITY AND PREDICT BIOLOGICAL AND ECOLOGICAL FEATURES OF UNKNOWN MARINE VIRUSES” published in PNAS as an open-access article. As the author of this article, I hold the copyright and agree to it being modified for use in this dissertation.

243